RF_WH_3_A_SUDHA.pdf
Media
- extracted text
-
RF_WH_3_A_SUDHA
Effects of Husband Involvement in Egypt
o
2
17
Despite the documented role of husbands as decisionmakers in
matters of reproductive health (Abdel-Aziz, El-Zanaty, and Cross 1992;
Reproduced with permission from
Population Council
Effects of Husband Involvement
on Postabortion Patients’
Recovery and Use of
Contraception in Egypt
Nahla Abdel-Tawab, Dale Huntington.
Ezzeldin Osman Hassan, Hala Youssef,
and Laila Nawar
Ali 1995), very little has been done to formally involve husbands in mat
ters related to the health of their wives, particularly in relation to
postabortion care. Current health services for postabortion patients sel
dom give any information to the patients themselves, let alone to their
husbands. Studies that measure the effects of husbands’ involvement have
commonly been conducted in family planning settings. Moreover, most
of the past research focused on a self-selected group of husbands who
received counseling. Prior to the present study, it was not known in Egypt
whether involving husbands would be feasible and acceptable, or whether
it would have a positive impact on the health of their spouses.
The study examines the effects of counseling the husbands of
postabortion patients on husbands’ level of involvement in their spouses’
Population-based studies on the incidence of abortion (spontaneous or in
recovery and on patients’ recovery and subsequent use of contraception.
duced) are lacking in Egypt, even though evidence suggests that abortion is a
common medical procedure. A recent study of the postabortion caseload in
Study Hypotheses
Egyptian public-sector hospitals found that one out of every five admissions
The following hypotheses were tested in this study:
to the OB/GYN ward was to a postabortion patient (Huntington et al. 1998).
Although the proportion of postabortion patients admitted with
serious complications is relatively small, their ability to recuperate after
receiving emergency medical treatment is of concern, particularly because
• Levels of positive husband involvement in spouses’ postabortion
recovery will be higher among husbands who receive counseling
than among husbands who do not.
• Postabortion patients whose husbands receive counseling will re
of evidence that postabortion patients in Egypt may not be receiving ad
port better physical and psychological recovery one month after
equate care after hospital discharge.
A study examining psychosocial stress in the postabortion period
discharge from the hospital than patients whose husbands do not
found that postabortion patients (especially those in rural areas) resume
• Contraceptive use or intent to use in the near future, measured one
receive counseling.
their domestic labor shortly after the procedure (Huntington, Nawar, and
month after discharge, will be higher among postabortion patients
Abdel-Hady 1997). Early resumption of hard work interferes with recov
whose husbands receive counseling than among patients whose
ery and places a burden on postabortion patients’ already compromised
husbands do nor.
health. Moreover, many patients often have to deal with questions and
doubts from their husbands or in-laws regarding their ability to conceive
again or to carry a pregnancy to term. As a result, some patients may give
in to family pressure for another pregnancy soon after their abortion in
order to prove their fecundity.
16
Study Intervention
The study was conducted in six hospitals in Menia Governorate in south
ern Egypt. The component on husband counseling was added to an inter
vention that included improving postabortion clinical services through using
18
Nahla Abdel-Tawab el al.
Effects of Husband Involvement In Egypt
19
manual vacuum aspiration (MVA), improving case management procedures,
Nawar 1997). Counseling was given only to husbands of consenting pa
and enhancing counseling for postabortion patients (Nawar et al. 1997a).
tients. The following enrollment procedure was developed through ex
Content of Counseling for Husbands
the number of husbands who would receive the intervention.
tensive pretesting to ensure protection ofpatient privacy while maximizing
The counseling of husbands emphasized the important role they can play
Procedures on patient admission. If the patient was brought to the
in their wives’ recovery and in the adoption of a family planning method
hospital by her husband, the nurse in charge asked him to return at dis
during the postabortion period. Although the manner in which this in
charge to pick up and assist his wife. Husbands did not receive counseling
formation was transmitted varied, each husband received, at the mini
at this point; they were counseled only after the patient’s consent was
mum, a “health education” message (although physicians were encouraged
obtained.
to have more lengthy exchanges with husbands). These messages were on
Obtainingpatient’s informed consent. The nurse in charge sought the
the following five topics: (1) patient’s need for rest and adequate nutrition
patient’s permission to have the doctor speak to her husband about her
(emphasizing inexpensive nutritious food); (2) postabortion warning signs
medical condition. The nurse read a standardized consent statement to
indicating need for follow-up care; (3) the possibility of a return to fertil
the patient after she had received complete medical treatment (including
ity within two weeks; (4) the need for family planning to avoid unwanted
counseling) and only after it was determined that she was in stable physi
or poorly timed pregnancy; and (5) cause of the miscarriage (if appropri
cal and emotional condition.
ate) and source of referral care (if necessary). Because abortion is illegal in
Contacting the husband. The pilot test of the informed consent pro
Egypt, no attempt was made to differentiate between cases of spontane
cedures showed that family members often go back and forth to the patient’s
ous or induced abortion.
home while she is in the hospital. Thus it was relatively easy to contact
those husbands who did not accompany the patient at admission. Pa
Procedures for Counseling Husbands
tients who agreed to have their husbands counseled were asked to tell
All postabortion patients received improved medical care and complete
their relatives or companions, who were already at the hospital, to send
counseling about their health condition and family planning. Counseling
for the husband to pick her up at discharge.
of patients and husbands was done separately. This decision was based on
Procedures at discharge. The attending physician commonly signs
findings from previous studies that showed women felt less free to ask
the discharge forms and gives the patient any remaining instructions for
questions or express concerns in the presence of their husbands (Kim and
postabortion care. If the patient had agreed to have her husband coun
Awasum 1996). Counseling of husbands was done by the attending phy
seled and if the husband was present at discharge, the nurse would ensure
sician when the patient was ready to leave the hospital. Because of the
that he met with the physician. In all cases husbands received counseling
sensitive nature of abortion and the possibility that by involving the hus
by the attending physician in private, away from their wives.
band he might gain information about his wife’s medical condition that
could pose a risk to the woman (e.g., a husband might learn that his wife
induced the abortion without his knowledge or that she concealed her
Training Providers on Procedures for Counseling Husbands
Thirty physicians (five from each hospital) received one day of orienta
pregnancy from him), considerable attention was devoted to establishing
tion on the content and procedures for counseling the husbands of
strict ethical guidelines and procedures (Abdel-Tawab, Huntington, and
postabortion patients. The orientation was done in conjunction with train
Nahla Abdcl-Tawab ct al.
20
ing for the “scaling-up" intervention already in place, which included train
ing on use of MVA and basic patient counseling (Nawar et al. 1997b).
The orientation strategy focused on creating a cadre of supervisors who
would then train their colleagues and provide daily on-site supervision
and follow-up for a period of three months.
Research Methodology
Effects of Husband Involvement in Egypt
21
Variables and Measurement
Independent variable. The principal independent variable is whether or
not the husband received counseling regarding his wife’s medical condi
tion. This was determined by the physician’s signature on the patient dis
charge form attesting that counseling was provided.
Intervening or mediating variables. The effects of counseling hus
bands on their wives’ recovery is assumed to be mediated by the degree to
Study Design
which counseling translates into husbands’ support. The study measures
A true experimental, posttest-only control group design (Campbell and
husband support as perceived by wives (i.e., based on patients’ self-re
Stanley 1963) was used to measure the effects of counseling husbands on
ports), which may or may not be the same as that perceived by the hus
their level of involvement in matters related to their spouses’ health and,
bands or a third person.
by consequence, on their wives’ recovery and use of contraceptive meth
A review of relevant literature, consultations with experts in the field,
ods. Consenting postabortion patients were randomly assigned to either
and exploratory studies revealed two broad dimensions of improved hus
an intervention or a control group. Husbands in the intervention group
band support: instrumental and emotional support. Preliminary analysis
received counseling about the medical condition of their wives. Husbands
of the two indexes used to measure these two dimensions, however, re
in the control group did not receive such counseling (although their wives
vealed that they were moderately correlated (r=0.29). Confirmatory fac
did receive improved postabortion medical care and counseling). A fol
tor analysis was conducted in which items from the two indexes were
low-up interview was conducted with both groups of patients one month
pooled together to explore patterning of the data. Rotated factor analysis
after hospital discharge to assess their physical and psychological recovery
revealed a third factor, family planning support, that had been initially
and to determine whether they were currently using or intended to use
contained within the emotional support index. Separated out, these three
contraception in the near future. The 30-day period was deemed suffi
indexes of husband support (instrumental, emotional, and family plan
cient to measure the effects of counseling on recovery and immediate
ning) were less correlated with one another (r=0.143) (i.e., more indepen
postabortion contraceptive use, although it was considered too short to
dent and with greater internal consistency reliability for each index).1 Items
investigate medium- and long-term morbidity or the attainment of other
included in each index are shown in Table 2.1.
reproductive intentions.
The index for instrumental support measured the degree of tangible
or material support provided by husbands. The index tapped indicators
Study Sample
of help provided by husbands in buying and preparing food for the pa
The sample size was initially estimated at 220 couples from each group
tient and assisting with various household activities. The index had five
(intervention and control) to account for losses to follow-up, statistical
items and an internal consistency reliability of 0.77.
significance tests, and budget considerations. All consenting postabortion
The emotional support index measured the husband’s reaction to
patients who were admitted to the study hospitals during a rwo-month
the lost pregnancy, as well as his understanding of his wife’s physical and
enrollment period were eligible for the study; this length of time was antici
emotional needs as reported by the patient herself. The index had eight
pated to be sufficient for the enrollment of the desired sample of patients.
items and an internal consistency reliability of 0.86.
Nahla Abdel-Tawab et al.
22
TABLE 2.1
Items on Instrumental, emotional, and family planning support Indexes
Instrumental support index
• My husband buys nutritious food for me
• My husband helps prepare food for me
• My husband encourages me to stay in bed
• My husband prepares hot drinks for me
• My husband helps with cleaning the house
Emotional support index
• My husband was more concerned about my health than about the lost pregnancy
• My husband was willing to skip his work and stay with me
• My husband encouraged me to make a follow-up visit
• My husband reassured me about my health
• My husband was angry at me because I lost this pregnancy
• My husband was upset because we had to postpone sex
• My husband could not understand how exhausted I was
• I could not tell my husband how I was feeling
Family planning support index
• My husband would not let me use a family planning method
• I can talk freely with my husband about family planning
• My husband would let me choose any family planning method that I want
• My husband would agree to go with me to the family planning clinic
• My husband and I are in agreement with regard to family planning
• My husband encourages me to take a break from pregnancy and childbirth
Effects of Husband Involvement In Egypt
23
work, and so forth. The index had eight items and an internal consistency
reliability of 0.88. A high score on the weakness index was indicative of
poor physical recovery.
Postabortion patients’ emotional health status was assessed through
the Psychological Distress Index (PDI), adapted from the Grief Scale de
veloped by Peppers and Knapp (1980). Patients were presented with a list
of emotional symptoms such as sleep difficulty, loss of appetite, anxiety,
and so forth, and were asked to indicate whether they had experienced
any of those symptoms in the past three days. The index had eight items
and an internal consistency reliability of 0.88. A high PDI score indicated
poor emotional recovery.
An index was constructed that included items on the patient’s re
sumption of social activities such as talking with her husband about his
work, visiting friends, resumption of sexual relations, and relations with
her children. The index had a very low internal consistency reliability and
was therefore dropped from subsequent analysis.
The family planning support index measured aspects of family plan
The second class of dependent variables that were measured in this
ning support by the husband, such as approval of his wife’s use of family
study concerned contraceptive use. Respondents were asked whether they
planning, openness to discussions about family planning with his wife,
were currently using a family planning method. Those who said they were
and willingness to accompany his wife to the family planning clinic. The
not were asked whether and when they intended to begin using a family
index had six items and an internal consistency reliability of 0.85.
planning method.
Dependent variables. The two principal dependent variables for the
study are improved patient recovery and contraceptive use.
Three dimensions of the patient’s health were measured to examine
patient recovery: physical, emotional, and social health. Separate indexes
Control variables. Because of the potential effect of a great number
of factors on husband involvement, patient recovery, or use of contracep
tion, the study identified several control variables and utilized them in the
analysis. The variables measured included patient’s and husband’s charac
were constructed for each dimension, using the same development pro
teristics, husband-wife relationship characteristics, patient’s medical char
cess described above to identify the items on each index, followed by con
acteristics, and characteristics of the index pregnancy. An autonomy index
firmatory factor analysis.
Physical health was measured through a “weakness index,” which
was used to measure husband-wife relations, composed of five items that
measured the wife’s ability to make a solitary decision on each of the fol
identified physical health problems that the patient might be having one
lowing: visiting a neighbor, visiting her parents, going out with a friend,
month after her discharge from the hospital. Patients were asked whether
visiting the doctor, and choosing meals for the family. The patient’s medi-.
they were having any of the following symptoms: fatigue, shortness of
cal characteristics included presence of abortion complications as well as
breath, or headache/dizziness. They were also asked whether they were
any preexisting medical conditions such as hypertension, kidney disease,
having difficulty performing activities such as walking, standing, house
parasitic disease, and so forth. With the exception of information on the
Nahla Abdel-Tawab et al.
24
patients medical characteristics, which was collected by the discharging
physician, information about all other control variables was collected at
the home interview with the patient.
Characteristics of the hospital where the patient was treated was
another control variable measured in the study. The six study hospitals
were divided into two groups: Group A and Group B hospitals (three in
each group). Group A hospitals were generally smaller (i.e., they had fewer
Effects of Husband Involvement in Egypt
Patients
Completed the study
Husband did not show up at discharge
Husband away for entire month
Lost to follow up
Total no. of patients agreeing Io participate
physicians) and therefore a larger proportion of physicians in those hospi
tals were exposed to the training and orientation. It was expected that
25
TABLE 2.2
Patients who participated In study: Enrollment, attrition, and completion
Intervention
group
Control
group
136
45
i
157
—
14
6
188
178
(n)
293
45
15
13
366
Percentage
~80
12
4
100
Follow-up interviews. Follow-up interviews were conducted with
even with the study’s monitoring and supervision visits, performance of
patients in the intervention and control groups one month after discharge
physicians from Group A hospitals would be somewhat better than that
from the hospital. Unless the patient indicated a preference to have the
of those from Group B hospitals.
interview at the hospital, follow-up interviews were conducted at patients’
homes. To the extent that it was feasible, patients were visited 30 days
Data Collection Activities
after their discharge. If the patient was not available, revisits were made to
Patient enrollment and randomization. A patient who agreed to participate
her home, with a maximum of three callback visits within a four-day
in the study was asked her full name and address for the follow-up inter
period. Five trained female interviewers conducted the follow-up inter
view. Patients who preferred to have their follow-up interview outside
views with patients. Interviewers were blinded as to the research group
their homes were asked to come back to the hospital on an assigned date
(i.e., intervention or control) to which the respondent belonged.
and time. A special logbook was used for randomization purposes. Access
to this logbook was strictly controlled. Patients with an even log number
Results
were assigned to the intervention group while those with an odd number
were assigned to the control group (i.e., every second consenting patient
Patients and Husbands Recruited for the Study
was assigned to the intervention group). Three or four nurses in each of
Of 381 postabortion patients who were admitted to the postabortion wards
the participating hospitals were in charge of patient enrollment and ran
during the two months of data collection and who were approached for
domization procedures.
At discharge, an attending physician gave the patient instructions
the study, 366 patients agreed to participate, yielding a refusal rate of 3.9
percent.
regarding her health condition and information on family planning and
Table 2.2 shows the proportion of patients who completed the study
also documented on the patients discharge form any complications or
and patients who were lost at different stages of the study; 12 percent of
preexisting medical conditions. If the patient was in the intervention group,
the patients who agreed to participate were excluded because their hus
a nurse arranged for the meeting to take place between the physician and
bands did not show up at discharge and therefore did not receive counsel
the husband. The physician then indicated that the counseling had taken
ing. Only patients in the intervention group were excluded from the study ■
place and provided basic sociodemographic information about him/her
if the husband did not show up at discharge; 24 husbands of patients in
self, including age, medical degrees, and position.
the control group also did not show up at discharge (data not shown).
Nahla Abdel-Tawab et al.
26
The following were reasons given for. the husband’s absence as explained
by patients in the control group during the follow-up interview: husband
out of town when the abortion happened (20 cases), husband arrived at
hospital after patient’s discharge (3 cases), husband was too busy at work
(1 case). Unfortunately, similar information about husbands in the inter
vention group is lacking.
Very few patients were lost to follow-up (4 percent) and hence did
not complete the follow-up interview. None of the patients who requested
to have their follow-up interview at the hospital showed up on the as
signed date and they were therefore lost to follow-up. Fifteen patients
completed the follow-up interview but were later excluded from the analysis
because their husbands were absent for the entire month of recovery. Thus
the final sample was composed of 293 patients, 136 in the intervention
group and 157 in the control group.
Patients ’ and Husbands' Characteristics
Table 2.3 shows that postabortion patients who participated in the study
were more likely to be younger than 25 years or in their 30s, had little or
no schooling, and had three or more living children. Forty-five percent of
the patients reported previous use of contraception and 61 percent indi
cated a desire for more children in the future. According to 39 percent of
the patients, the lost pregnancy was planned. Forty-three percent of pa
tients said they lost their pregnancy at a gestational age of 12 weeks or
higher. Seventy-seven percent of patients reported experiencing at least
one complication such as excessive bleeding, fever, or offensive discharge
during the 30-day recovery period. Sixty-five percent of patients were
Effects of Husband Involvement in Egypt
27
TABLE 2.3
Percent distribution of selected sociodemographic and medical
characteristics of postabortion patients and their husbands
Characteristic
Patient
Age (years)
<25
25-29
£30
Education
No sctiool/less than primary
Primary or higher
No. of living children
0
1-2
£3
Preexisting medical conditions
Yes
Gestational age of lost pregnancy*
<8 weeks
8-11 weeks
£12 weeks
Not known
Index pregnancy was planned
Yes
Ever use of contraception
Yes
Desire for more children
Yes
Complications during recovery
One or more
Hospital
Group A
Group B
Autonomy leveP
Low
Medium
High
Intervention
(n=136)
Control
(n=157)
Total
(n=293)
37
28
35
38
26
36
38
23
39
79
21
83
17
81
19
24
32
44
22
27
51
23
29
48
2
7
5
14
48
38
0
20
30
48
2
17
39
43
37
40
39
46
45
45
62
60
61
74
79
77
34
66
36
64
35
65
21
57
22
13
64
23
17
61
22
larger hospital with a smaller proportion of physicians who received di
*p<0.01.
a The autonomy index range was 0-100 Scores were classified into < 20. 20-59,260 to denote low.
medium, and high autonomy, respectively
(continued)
rect training and orientation). On the patient autonomy index, the ma
cent) while the remainder lived with parents or in-laws. The majority of
treated in a hospital that was classified in this study as Group B (i.e., a
jority of patients scored a medium score (20-59 points out of 100).
couples who participated in the study lived in rural areas (76 percent).
Fifty-three percent of husbands of postabortion patients were 35
Table 2.3 shows that patients in the intervention and control groups
years or older and 63 percent had little or no schooling. Slightly more
were comparable on most of the above characteristics. However, patients
than half of the studied couples lived in a nuclear family setting (55 per
in the control group were more likely to have lost the index pregnancy at
Nahla Abdel-Tawab ct al.
28
Effects of Husband Involvement in Egypt
29
TABLE 2.3 (continued)
Characteristic
Husband
Age
<30
30-34
>35
Education
No school/less than primary
Primary/preparatory
Secondary/college
Household composition
Nuclear
Extended
Residence
Urban
Rural
Intervention
(n=136)
Control
(n=157)
Total
(n=293)
28
20
52
27
19
54
28
19
53
62
25
64
16
20
63
15
22
50
50
59
41
55
45
21
79
27
73
24
76
12 or more weeks than patients in the intervention group, who were more
likely to have lost it between 8 and 11 weeks. This difference is statisti
cally significant.
FIGURE 2.1 Mean score on three Indexes of support provided by
husbands of postabortion patients
Because of the relatively small sample size and skewness of observa
Support Given by Husbands During Recovery
tions, respondents were classified into two groups depending on the level
The study posits that counseling will lead to greater involvement by hus
of support provided by husbands (i.e., below mean corresponds to low
bands, which, in turn, will positively affect their spouses’ recovery. The
support; above mean corresponds to high support). Logistic regression
intervening variable, husband involvement, is operationalized through an
analysis was conducted to measure the association between counseling of
expression of greater instrumental, emotional, and family planning sup
husbands and each type of support provided, after controlling for poten
port by the husband. This section presents the results of the intervening
tial confoundcrs. A hierarchical modeling technique was used to enter
variable—the indicator of how the study’s intervention was translated into
variables into the model. Counseling of husbands was entered first and
action by the husbands. It also examines the effects of the control vari
ables on these demonstrable actions as a prelude to investigating the im
pact of the intervention on the two dependent variables.
Figure 2.1 shows mean levels of the different types of support pro
then other variables were entered for forward selection with alpha set at 1.0.
Table 2.4 shows that after adjusting for other patient characteristics,
counseling of husbands was associated with an increase in the likelihood
that the husband will provide high instrumental, emotional, or family
vided by husbands of postabortion patients. On average, husbands in the
planning support to his wife during recovery. Counseling was associated
two groups provided a higher level of emotional and family planningsup
with a greater increase in family planning and instrumental support than
port than instrumental support to their wives (mean 83, 65, and 17, re
emotional support. Other factors associated with increased instrumental
spectively). Bivariate analysis showed no differences between the
support by the husband were: living in a nuclear family arrangement,
intervention and the control with regard to the degree of support pro
patient’s experience of complications during recovery, and higher educa
vided by husbands.
tion of the husband (i.e., secondary or above). Interestingly, receiving coun-
Nahla Abdel-Tawab et al.
30
TABLE 2.4
Logistic regression analysis of variables associated with high level
of Instrumental, emotional, and family planning support
provided by husbands1 (n=293)
Variable
Instrumental support6
Household composition (nuclear)
Complications during recovery (yes)
Husband's education (secondary or above)
Husband counseled (yes)
Group A hospital and husband counseled (yes)
Emotional support0
Index pregnancy planned (yes)
No. of living children (>3)
Husband and wife blood relatives (yes)
Husband counseled (yes)
Family planning support*1
Desire for more children (no)
Husband’s education (secondary or above)
No. of living children (>3)
Group A hospital
Husband counseled (yes)
Adjusted
odds ratio
90% confidence
interval
2.6
1.5
0.5
1.7, 4.0
1.1,3.2
1.0,2.5
0.9, 2.4
0.2.0.9
2.8
2.0
1.9
1.3
1.8,5.0
1.2,3.3
1.2,2 9
0.8,1.9
5.0
2.5
2.5
2.5, 10.0
1.4,5.0
1.4,5.0
1.1,2.7
1 1,2.6
Effects of Husband Involvement In Egypt
TABLE 2.5
Logistic regression analysis of variables associated with
good physical recovery of postabortion patients1 (n=293)
Variable
Emotional support by husband (high)
Husband counseled (yes)
Instrumental support by husband (high)
Complications during recovery (yes)
Adjusted
odds ratio
1.3
0.2
90% confidence
interval
1.1,2.6
08,2.0
0.7, 1.7
0.1,0.3
a Reference categories are low level ol emotional support by husband, husband did not receive
counseling, low level ol instrumental support by husband, palient did not experience complications
during recovery
more children in the future, the husband was five times more likely to
provide family planning support to his wife than if the couple wanted
1.6
a Logistic regression analysis was used to measure association between counseling of husbands
and each type ol support after controlling for other patient characteristics Dependent variable in each
model was high level of that dimension of husband support.
b Reference categories are. couple lives with in-laws/parents. patient did not experience complica
tions during recovery, husband completed less than secondary education, husband did not receive
counseling, patient was admitted to Group B hospital and husband did not receive counseling
c Reference categories are. index pregnancy was not planned, palient has fewer than 3 living
children, husband and wife are not blood relatives, husband did not receive counseling.
d Reference categories are: palient wants more children in the future, husband completed less than
secondary education, patient has fewer than 3 living children, palient was admitted to Group 8
hospital, husband did not receive counseling.
seling in a Group A hospital (i.e., a smaller hospital with more trained phy
more children. Also, if the couple had three or more living children, the
husband was 2.5 times more likely to provide family planning support to
his wife. Husbands with secondary or higher education were more likely
to provide family planning support than husbands with fewer years of
schooling. Patients who were treated in smaller Group A hospitals were
1.7 times more likely to receive family planning support from their hus
bands than patients who were treated in larger Group B hospitals.
Outcomes of Counseling
This section reviews the results of the two outcome indicators: patient’s
recovery and use of contraception.
sicians) was associated with decreased instrumental support by husbands.
Physical recovery. The aforementioned weakness index measured pa
Emotional support by husbands was less likely to be affected by coun
tients’ physical recovery one month after discharge. Patients scoring less
seling than by patient and husband characteristics. Husbands were more
than 7 out of 14 on the weakness index were considered as having experi
likely to provide emotional support to their wives if the index pregnancy
enced good physical recovery while those with a score of 7 or above were
was planned. Patients who had three or more living children were more
considered to have poor physical recovery.
likely to receive emotional support from their husbands than patients with
Table 2.5 shows results of logistic regression analysis of variables
fewer children. Also, when patients and their husbands were blood relatives
associated with good physical recovery. The variables husband counseled,
husbands were more likely to provide emotional support to their wives.
instrumental support, and emotional support were entered first, then the
Husbands’ support of family planning was mostly determined by
following variables were entered for forward selection: gestational age,
the couple’s demographic characteristics. When a couple did not want
hospital group, experience of complications during recovery, number of
Nahla Abdel-Tawab et al.
32
Variable
Patient's education (no schooMess than primary)
Emotional support by husband (high)
Index pregnancy planned (yes)
Husband counseled (yes)
Complications during recovery (yes)
Patient's autonomy index score
Low
Medium
Effects of Husband Involvement in Egypt
reported receiving a high level of emotional support from their husbands
TABLE 2.6
Logistic regression analysis of variables associated with
good emotional recovery of postabortion patients* (n=293)
were twice as likely to experience good emotional recovery. Two addi
Adjusted
odds ratio
90% confidence
interval
2.5
2.0
1.7
1.0
0.4
1.4,4.3
1.3,3.0
1.1.2.6
0.6.1.4
0.2,0.7
0.4
1.0
0.2,0.8
0.6,1.7
tional variables were associated with improved emotional recovery: pa
tients with little or no education and patients reporting that the index
pregnancy was planned. Two variables were negatively associated with
good emotional recovery: low level of patient autonomy and patient’s ex
perience of complications during recovery.
Family planning use. The second outcome indicator examined in
this study was family planning practice. By the end of the one-month
follow-up period, 14.7 percent of women reported having started using
a Reference categories are: patient completed primary or higher education, low level ol emotional
support by husband, index pregnancy was not planned, husband did not receive counseling, patient
did not experience complications during recovery, patient scored high (260) on autonomy index.
contraception. An additional 34.5 percent of patients said they were plan
living children, and interaction between counseling of husband and hos
combined the two groups into one, for a total of 49.2 percent of
pital group. Interestingly, a high level of emotional support by the hus
postabortion patients.
ning to use contraception within one month. Because of the small num
ber of observations of patients in the first category, subsequent analysis
band was significantly associated with improved physical recovery by the
Logistic regression analysis was conducted to determine association
patient, whereas instrumental support by the husband was not. The stron
between patient’s use of family planning and counseling of husbands and
gest predictor of physical recovery, however, was absence of any complica
degree of family planning support, while controlling for potential con-
tions during convalescence. Patients who experienced complications were
founders. The following variables were entered into the model for for
five times less likely to report physical recovery than patients who did not
ward selection: number ofliving children, desire for more children, patient's
experience complications.
Emotional recovery. The aforementioned Psychological Distress In
age, previous use of contraception, whether index pregnancy was planned,
dex showed that slightly more than half of the study patients (53 percent)
interaction between hospital group and counseling of husband.
patient’s experience of complications during recovery, hospital group, and
did not report any emotional symptoms, hence were considered as having
Table 2.7 shows that after controlling for other variables, husband
experienced good emotional recovery. Table 2.6 shows results of logistic
support of family planning is highly predictive of contraceptive use or
regression analysis of variables associated with good emotional recovery.
intention to use. Other variables that were predictive of use of contracep
The variables husband counseled and emotional support were entered
tion were previous use of family planning, husband receiving counseling
first. The following variables were entered for forward selection (alpha=0.1):
at a Group A hospital, and couples with three or more children. On the
number of living children, desire for more children, whether index preg
other hand, if the index pregnancy was planned, the patient was less likely
nancy was planned, patients educational level, patients autonomy level,
to use or to intend to use a family planning method in the near future.
experience of complications during recovery, hospital group, and interac
tion between hospital group and husband counseled.
Discussion
As shown in the table, a positive association exists between emo
The study was the first to test an intervention for counseling tl.e hus
tional support by the husband and good emotional recovery. Patients who
bands of postabortion patients in Egypt. Its primary aim was to examine
34
Nahla Abdel-Tawab et al.
1 ABLt i.l
Logistic regression analysis of variables associated with use of
or Intention to use family planning by postabortion patients* (n=293)
Adjusted
Variable_______________________________ odds ratio
90% confidence
Interval
Family planning support by husband (high)
5.9
Group A hospital and husband counseled (yes)
3.8
Previous use of family planning (yes)
35
No. of living children (23)
2,5
Husband counseled (yes)
06
Index pregnancy planned (yes)0 4
3.5,10.0
1.5.9.4
1.7,5 6
1.4,5.0
0.3.1.2
0 2| 0 7
Effects of Husband Involvement in Egypt
35
ciation between husbands’ support of family planning and adoption and
continuation of contraception (Joesoef, Baughman, and Utomo 1988;
Terefe and Larson 1993).
Also supported is the primary hypothesis of the study that counsel
ing of husbands will result in positive changes in their behavior, which in
turn will lead to better patient outcomes. Counseling of husbands is most
likely to influence patient outcomes through changes in husbands’ level
of involvement.
Program managers in Egypt should consider including counseling
a Reference categories are: low level ol family planning support by husband, patient was admitted Io
Group B hospital and husband did not receive counseling, patient has not used family planning
before, patienl has lewer than 3 living children, husband did nol receive counseling, index pregnancy
was nol planned
of husbands as part of the medical services provided to postabortion pa
the effects of counseling on husbands’ behavior and on postabortion pa
support were associated with better patient outcomes. Counseling should
tients’ recovery and use of contraception. Other issues addressed included
be more vigorous with husbands of patients who are otherwise less likely
tients. The content of counseling should highlight the importance of
emotional and family planning support, because these two dimensions of
the feasibility of counseling husbands and of conducting follow-up as
to receive support from their husbands. According to the findings of this
sessments of postabortion patients while protecting their rights to privacy.
study these include patients with fewer living children and those with
Overall the study findings show a positive association between coun
husbands who have less education. It is important that providers establish
seling husbands and their involvement in their spouses’ recovery. On av
two-way communication during counseling in order to identify husband
erage, patients in the two study groups received more emotional and family
groups who are less likely to provide support to their wives.
planning support than instrumental support from their husbands. Be
The study also shows that counseling of husbands is associated with
cause the study was conducted in a traditional society, where men are
better outcomes in some hospitals than in others. This clearly suggests
often discouraged from participating in what are perceived as women’s
that having senior and better-trained staff provide the counseling leads to
activities, this result is expected. It is interesting to note, however, that
a stronger impact on husbands’ behavior. It also suggests that the two-
husbands were more likely to provide instrumental support to their wives
tiered system (training of trainers) may not be as effective as direct train
when the couple lived within a nuclear family, implying that husbands are
ing of all service providers, although, obviously, the former is less expensive.
willing to help if no other women are available to provide such support.
The results show that support by husbands is crucial for patients’
recovery and use of contraception. Emotional support by husbands is
It would have been helpful if counseling sessions were observed directly
so that variations in quality of counseling given in different hospital groups
could be ascertained.
particularly important for women’s physical and emotional recovery. Family
Some administrative and logistic changes may be needed in each
planning support by husbands is the strongest predictor of family plan
hospital in order to make counseling of husbands more feasible and effec
ning use by the patient. These findings are in agreement with those docu
tive. The current set-up in OB/GYN wards may need to be changed to
mented in the literature on the health effects of social support (Cohen
make the hospital environment more husband-friendly and to encourage
and Syme 1985). The family planning literature also documents an asso
husbands to accompany their wives to the hospital. Providing family plan-
Reproduced with permission from Population Council
Improving the Quality of
Services and Contraceptive
Acceptance in the Postabortion
Period in Three Public-Sector
Hospitals in Bolivia
Juan Diaz, Mariel Loayza,
Yamile Torres de Yfipez, Oscar Lora,
Fernando Alvarez, and Virginia Camacho
Maternal mortality is one of the most important public health problems
in Bolivia; although the Ministry of Health (MOH) has declared that
maternal mortality is one of its first priorities and several campaigns aimed
at reducing it have been implemented, the rate is 390 maternal deaths per
100,000 live births, one of the highest in the region. Based on the latest
available data it is estimated that 27-35 percent of maternal deaths are
due to complications from abortion (Encuesta Nacional de Demografla y
Salud 1994; UNICEF 1992). In 1994 the Ministry of Health (MOH)
declared that maternal mortality was one of the most important womens
health problems and instituted a plan for its accelerated reduction. This
included the goal of reducing induced abortions and their complications.
Analysis of abortion in Bolivia is difficult because, as in other coun
tries of Latin America and the Caribbean (LAC), statistical information is
scant and estimates of its incidence are not reliable. There is evidence,
however, that due to very restrictive legislation, most women try to solve
problems resulting from unwanted pregnancy and induced abortion with
out consulting public-sector medical services. This explains why 95 per
cent of deaths due to induced abortion occur at the woman’s home
61
Juan Diaz et al.
Improving Services in Public-Sector Hospitals in Bolivia
63
62
(Ministerio de Relaciones Exteriores y Culto 1994). Women usually present
for services only after they have suffered long periods of pain and bleed
declared that either they did not believe in the efficacy of contraceptive
methods or they were not informed about them.
ing; frequently complications are severe and lead to serious sequelae or
Data from an unpublished survey of women with abortion compli
even to death (Rance 1993). In addition, women do not admit to inducing
cations, conducted by the Bolivian Society of Obstetrics and Gynecology
their abortions because of the risk of being mistreated, denounced, or sued.
(Taborga, Pooley, and Rada 1988), showed that lack of information was
Between 1966 and 1976, abortion complications represented 46
the main reason women did not use contraception (48 percent). Eighteen
percent of all obstetric hospitalizations; in the next decade they made up
percent said that they thought they would not get pregnant or that they
51 percent of all such hospitalizations (Rance 1990; UNFPA 1990). A
were not able to get pregnant. Almost 10 percent gave religious reasons
report published by USAID in 1990 stated that, in the Bolivian social secu
for not using contraception. Only 2 percent mentioned lack of access as
rity system, by far the most important cause of hospitalization of women of
the reason for not using contraceptive methods. The remaining 22 per
reproductive age was complications of illegal abortions (USAID 1990).
cent did not use contraception because they wanted to get pregnant, were
The Caja Nacional de Salud (Social Security Institute) found that
already pregnant, or were lactating. The study also showed that most
the unit cost of treating induced abortion is several times higher than that
women did not receive contraceptive counseling after resolution of the
for deliveries because most women present with complications at more
than one time after the induction of their abortion, which, in most cases,
has been done by a person without formal training in gynecology or even
in medicine. The most frequently used methods to induce abortion are
introducing objects, curettage, oral abortifacienrs, and injections.
Women presenting to hospitals with postabortion complications are
intensively interrogated to discover whether the abortion was illegally in
duced. It is a common practice to charge a higher fee to women present
ing clear evidence of induced abortion. In some cases the fee charged is so
abortion, confirming that postabortion contraceptive services have not
received priority.
A qualitative assessment conducted by the MOH in collaboration
with the World Health Organization showed that access to abortion ser
vices is limited. Some hospitals, mainly Catholic ones, do not accept
women if they are suspected of having induced their abortions. In addi
tion, counseling during the postabortion period is not given, and contra
ceptive methods are not offered. As a consequence, many women get
pregnant soon after their abortions and resort to another induced abor
high that it prevents hospitalization or leads to a significant negative im
tion, again placing their lives at risk (Camacho et al. 1996).
pact on the family budget.
The rationale given by providers for this excessive fee is that it might
egy to improve women’s access to and the quality of postabortion services,
discourage repeated abortions. Sometimes women are mistreated or sim
including postabortion contraception as a means of decreasing mortality
Based on that assessment, the MOH decided to implement a strat
ply not accepted in the hospital. On the other hand, women who suppos
edly have spontaneous abortions and are unable to pay the entire fee are
interviewed by a social worker who is then authorized to lower or even
due to repeated abortions. Moreover, postabortion contraception is con
sidered the best means to break the vicious cycle of unwanted pregnancy
followed by induced abortion, lack of counseling, and new unwanted
pregnancy (Benson 1996).
waive the fee.
It is worth noting that a small proportion of women (17 percent)
who resort to abortion had at one time used a modern contraceptive
method (Bailey et al. 1988). The same study showed that most women
The MOH requested technical assistance from the Population Coun
cil in order to implement a pilot project aimed at assessing the feasibility
of carrying out a program to improve the quality of services for postabor
Juan Diaz et al.
Improving Services In Public-Sector Hospitals In Bolivia
65
64
tion complications, including contraceptive counseling and services, and
the impact on client satisfaction and contraceptive acceptance. This chapter
presents the results of the pilot project undertaken in three public-sector
reluctant to initiate project activities. They were apprehensive about pos
sible negative reaction of the church, which exerts strong influence on the
health authorities. This fact explains, at least in part, the lower contracep
tive acceptance in La Paz, compared with Santa Cruz.
hospitals from October 1995 through December 1997.
The third hospital is in Sucre, a smaller city with only 176,000 in
habitants. Sucre is the official capital of the country and the base of the
Materials and Methods
Judiciary. It is home to a well-known university, with one of the best law
schools in the country and several other highly regarded faculties. Be
Study Sites
The study was undertaken in three maternity hospitals in three cities:
Hospital de la Mujer, La Paz; Institute de Maternidad Percy Boland
Rodrfguez, Santa Cruz de la Sierra; and the Hospital Gineco-Obst^trico,
Jaime Sanchez Porcel, Sucre. The three hospitals are the most important
of their respective departments and are women’s health referral centers.
They attend primarily people from middle and low socioeconomic levels.
The largest, with 165 beds for obstetrics and gynecology, is in Santa
Cruz, the largest of the three cities in our study, with approximately
870,000 inhabitants. Before the initiation of the project, the hospital had
already implemented some family planning activities. In addition, some
physicians were implementing selected activities in the postabortion pe
riod, including providing contraceptive services. More than 1,400 cases
of complicated abortions were attended in the hospital during the year
cause the university receives students from all over the country, Sucre has
a greater percentage of adolescents, most of whom are living away from
their families. Their abortion rate is high, and it is for this reason the
abortion rate in the city is far higher than that in La Paz and Santa Cruz.
During the year before the project, the hospital attended 414 abortions
(2.4 per 1,000 inhabitants) compared to 1.6 per 1,000 in Santa Cruz and
0.3 per 1,000 in La Paz. Implementation of the project was difficult be
cause the Judiciary insisted that all women who have induced abortions
be denounced and prosecuted. In 1994 and 1995, judges had sued and
condemned at least 10 women for having induced abortions. For this
reason, physicians were reluctant to participate in a program designed to
give greater attention to all women having abortions, induced or sponta
neous. It was not easy to convince physicians to stop denouncing women,
before the initiation of the project. Two or three months after project
but agreement was obtained during the second year. In turn, judges began
initiation, the hospital opened an outpatient clinic specifically devoted to
simply ignoring the fact that induced abortions were no longer denounced.
adolescents.
The hospital in La Paz has fewer beds devoted to gynecology (58),
postabortion facilities and equipment significantly, making the hospitals
At the beginning of the project, the three hospitals improved their
although the population of the city (830,000) is comparable to that of
appropriate for attending the demand for services. The technical level of
Santa Cruz. The social and economic status of the population attended is
the personnel was similar in the three hospitals. La Paz had more prob
similar to that attended in the Santa Cruz hospital, but the proportion of
lems implementing good counseling, mainly because of the initial reluc
indigens is higher. Indigens have a cultural resistance to seeking hospital
tance of physicians to authorize postabortion counseling and family
based care, which may explain why the number of abortions attended in
planning services.
the La Paz hospital the year before the project (253) was far lower than
None of the three hospitals previously had a specific program for
the number attended in Santa Cruz. During the first months of the project
postabortion care. Women with postabortion complications were attended
after the training course, hospital personnel, especially physicians, were
as emergencies and discharged from the hospital as soon as possible with
Juan Diaz et al.
66
out receiving postabortion counseling. In addition, before the initiation
Improvlng Services in Public-Sector Hospitals In Bolivia
67
nator of the project maintained continuous contact with the hospitals
of the project, women were frequently mistreated because physicians and
and visited them at several times after training. Monitoring and supervi
most other health providers considered them to be criminals whom they
sion visits showed that the quality of care, specifically counseling, im
should punish for having initiated abortion. The situation was worse in
proved significantly after training. These visits were useful in reinforcing
Sucre where, due to the legal situation, physicians acted more as prosecu
tors than as physicians.
For all these reasons, women with postabortion complications tried
to avoid attending these hospitals. When they finally consulted, it was
usually with severe complications, after long periods of suffering at home.
the training and providing opportunities to retrain some health workers
who still were not providing considerate treatment to women.
Improvement of Facilities and Equipment
Two rooms, one to perform curettage and another for individual counsel
The situation was somewhat better in the hospital in Santa Cruz, where
ing, were equipped and placed in service in each of the three hospitals.
some physicians had been trained in the technique of manual vacuum
This allowed the hospitals to give priority and privacy to women present
aspiration (MVA) and some actions aimed at improving the quality of
ing with postabortion complications. It also served to define a patient
care had been implemented previously.
flow for these cases that was independent of and not interrupted by the
other activities of the hospitals.
Training and Monitoring
Physicians and counselors at the three participating hospitals attended a
Clinical Activities and Data Processing
training course on elements of postabortion care, including contraceptive
After the training course, the participating hospitals implemented a coun
technology. The course emphasized the importance of appropriate coun
seling program that included activities at admission, in the surgical room,
seling during the postabortion period, related both to contraception and
and after curettage. At each point, women were offered the opportunity
to other health issues, such as resumption of sexual activity and infection
to decide to use contraception and to choose a method. Those who did
prevention. The main objective of the training was to change the attitude
not reach a decision had an additional opportunity during the postabortion
of health providers, emphasizing the need to treat women as human be
follow-up visit scheduled one month after discharge from the hospital.
ings in need of health care rather than as criminals. Training also empha
All data were registered in a logbook and also in the clinical records of the
sized that the staffshould use all opportunities available during the womens
hospitals.
stay at the hospital to provide information and counseling on postabortion
care. Particular emphasis was given to avoiding another abortion.
The initial assessment of conditions at the three hospitals showed
After discharge, the women received a card with instructions and an
appointment for the postabortion follow-up visit. After this visit, women
were then scheduled for a six-month visit.
that physicians and other personnel involved in the treatment of abortion
The project evaluation consisted of a pre- and postintervention as
complications were technically well prepared and did not need training
sessment of the quality of postabortion services and the level of contra
ceptive acceptance over the duration of the study. Qualitative analysis was
on technical issues.
At each of the hospitals one of the trainees acted as coordinator of
performed through structured questionnaires and in-depth interviews with
the activities, trained the rest of the personnel, and supervised them in
women and health providers. Quantitative analysis was based on service
order to maintain high-quality services. In addition, the general coordi
statistics.
Juan Diaz et al.
68
Results
Evaluation of Training and Quality of Services
Evaluation of the pre- and posttraining questionnaires showed an impor
tant improvement in the level of providers’ knowledge of postabortion
care, including contraception. In addition, monitoring visits showed that
providers changed the way they treated women with postabortion com
plications. They became more considerate and compassionate, treating
Improving Services In Public-Sector Hospitals In Bolivia
TABLE 4.1
Monthly average number of women with postabortion complications
Before training
After training
Hospital
Jul-Dec 1995
Jan-Jun 1996
Jul-Dec 1996
Jan-Jun 1997
La Paz
Santa Cruz
Sucre
35
136
38
36
138
32
38
135
33
47
146
33
TABLE 4.2
Percent distribution of women with postabortion complications, by age
women as patients in need of assistance rather than as criminals who should
be punished for their actions. Counseling was implemented and offered
Age (years)
La Paz
1996
1997
Santa Cruz
1996
1997
1996
1997
over the course of women’s hospital stays.
Interviews with women in the three hospitals showed that a great
£19
20-34
2*35
N
15.9
65.8
18.4
441
15.5
73.5
11.1
1.635
11.9
72.4
15.8
387
29.6
63.6
6.8
294
majority mentioned that they had been well treated and emphasized the
15.6
639
20.6
379
30.1
57.0
12.9
1,300
Sucre
supportive attitudes of the nurses and auxiliaries. Most women stated that
the counseling they had received had been very appropriate and useful.
The rate of moderate to severe complications (e.g., severe anemia,
Several women mentioned spontaneously that the quality of the counsel
infection, and uterine perforation) in women admitted to the hospitals
ing was perhaps the most important reason they had evaluated the atten
varied from 3.4 percent to 8.4 percent in different periods (data not shown).
tion they received as good.
The changes in provider attitudes and improvement of quality of
The rare of complications was somewhat higher in La Paz, due to higher
services became apparent in Santa Cruz soon after training. In Sucre, it
other hospitals, died as a consequence of their abortions. The four cases
took longer but changes were also very positive. In La Paz, however, im
arrived at the hospital in very bad condition, after a long period of illness
provements in quality, especially in counseling, were very slow and did
at home or in another health facility.
not reach the same level as at the other hospitals.
rates of infection. Four women, two in Santa Cruz and one in each of the
Table 4.2 shows the age distribution of women over the two years of
the project. In Santa Cruz and Sucre, the percentage of adolescents with
Quantitative Results
postabortion complications sharply increased in the second year. In 1997
Table 4.1 shows that the monthly average number of women coming to
roughly 30 percent of all women in both hospitals were less than 20 years
the three hospitals for postabortion complications remained fairly stable
old. In La Paz, the proportion of women less than 20 years old remained
during the study period, maintaining levels comparable with the prestudy
fairly stable during the study period.
period. With the exception of the hospital in Sucre, a slight increase was
Other characteristics of women coming to the three hospitals re
observed in 1997. During the study period, all three participating hospi
mained the same during the study period and were very similar to the ones
tals attended a population very similar to the one attended during the five
observed in the prestudy period. Most women were poor, and between 60
years before project initiation. Therefore the changes in the number or
and 70 percent were married or in a consensual union (data not shown).
characteristics of women coming to the hospitals for treatment of abor
tion complications may be due to the intervention.
The history of previous abortions among women in the three hos
pitals is shown in Table 4.3. The percentage of women who had no previ-
Juan Diaz et al.
70
TABLE 4.3
Percent distribution of women with postabortion complications,
by number of previous abortions_______________ __________ ___
Sucre
Santa Cruz
1997
1996
1997
54.3
26.2
19.5
1,300
73.7
17.8
8.5
387
73.8
18.0
8.2
294
Previous abortions
1996
a Paz
1997
1996
None
One
Two or more
N
51.3
30.6
18.1
441
52.3
31.9
15.8
379
65.9
22.2
11.9
1,635
TABLE 4.4
Percentage of women receiving contraceptivecounseHng
After training
Before training
Improving Services in Public-Sector Hospitals In Bolivia
71
TABLE 4.5
Percentage of women accepting contraception following treatment
for postabortlon complications, by method
Hospital
Jul-Dec 1995
Jan-Jun 1996
Jul-Dec 1996
Jan-Jun 1997
LaPaz
Santa Cruz
Sucre
2.8
12.2
3.6
77.3
90.1
73.2
74.7
100.0
92.4
84.6
100.0
97.0
Hospital
and method
La Paz
IUD
Hormonal
Other
None
Santa Cruz
IUD
Hormonal
Other
None
Sucre
IUD
Hormonal
Other
None
Before training
Jul-Dec 1995
Jan-Jun 1996
After training
Jul-Dec 1996
Jan-Jun 1997
2.9
5.7
5.7
85.7
25.0
5.6
30.6
38.9
13.2
10.5
28.9
47.4
36.2
2.9
4.4
2.9
89.7
31.2
37.7
4.3
26.8
41.5
39.3
36.3
39.0
11.9
12.3
0.0
2.6
10.5
86.8
12.5
6.3
28.1
53.1
27.3
9.1
27.3
36.4
18.2
36.4
18.2
25.5
All women who received counseling declared that counselors were
very friendly, gave complete information about methods, and left the de
ous abortions was higher in Sucre, while the percentage of women who
cision on method choice up to the woman. In addition, counselors also
had two or more abortions was lower. The proportion of women with a
gave good and complete information on when and where to return for
history of previous abortions remained fairly stable during the study pe
postabortion follow-up.
riod in La Paz and Sucre. In Santa Cruz, the proportion of women with
Table 4.5 shows total contraceptive acceptance. As expected, con
two or more abortions was higher in 1997.
Contraceptive counseling was almost nonexistent in two of the three
traceptive acceptance in the first period (before training of personnel) was
low, between 10 and 15 percent. Women who accepted contraception in
hospitals before the study. Nevertheless, interviews with women during
this period were mostly those who initiated a follow-up visit and requested
the six months before the study showed that some counseling activities
a contraceptive method.
had been developed, mainly in Santa Cruz. Table 4.4 shows that counsel
Contraceptive acceptance in the three hospitals increased soon after
ing sharply increased after training of personnel. The increase was very
training and the implementation of the counseling program. In La Paz,
rapid in Santa Cruz, where all women received counseling on contracep
contraceptive acceptance in the first six-month period after training was
tion starting with the second half of 1996. Sucre saw a substantial in
61.2 percent. Acceptance decreased to 52.6 percent in the next period
crease in counseling as well, reaching almost 100 percent.
In La Paz the percentage was somewhat lower because some physi
and then increased to 63.8 percent in the last period. The acceptance of
cians refused to participate and influenced personnel during the periods
IUDs followed the same pattern, with a sharp initial increase, a decrease
in the next period, and a recovery in the third period.
they were in charge of the emergency room. Although the general coordi
In Sucre the process was slower; h 3 wever, a steady increase was ob
nator of the study made several monitoring visits to this hospital, by July
served during the period evaluated. At the end of the study period total
1997 some physicians still remained reluctant to participate.
acceptance reached a higher level than it did in La Paz.
Juan Diaz et al.
72
Improving Services in Public-Sector Hospitals in Bolivia
73
TABLE 4.6
Percentage of womenattendlng postabortion follow-up visit
Before training
La Paz
Santa Cruz
Sucre
After training
Jul-Dec 1995
Jan-Jun 1996
Jul-Dec 1996
Jan-Jun 1997
2.0
3.6
2.3
25.1
17.1
14.0
29.5
29.8
22.2
48.8
20.8
28.4
All women, whether or not they accepted a contraceptive method,
were scheduled for a follow-up visit approximately one month after dis
charge. Before the initiation of the project this was not standard procedure,
which explains why the percentage of women attending a follow-up visit in
FIGURE 4.1
Percentage of women accepting modern contraception
the pretraining period was very low in the three hospitals (Table 4.6).
After training, the percentage increased in the three hospitals, with
Contraceptive acceptance in Santa Cruz was the highest, 73.2 per
cent in the first six-month period after training and 87.6 percent in the
last period evaluated. In the last period observed, acceptance of hormonal
methods stood at 39.0 percent and IUDs at 36.3 percent. It is also worth
highlighting that in La Paz and Sucre a substantial increase was observed
La Paz showing the highest return rates, reaching 48.8 percent during the
first half of 1997. Sucre presented a lower but steady increase, reaching
28.4 percent in the same period. In Santa Cruz the rate also increased but
was variable. During the first half of 1997 the percentage was lower than
it was in the second half of 1996.
in the acceptance of other methods such as condoms (condom data in
cluded as part of other methods).
Figure 4.1 illustrates the rise in acceptance of modern contracep
tion (IUDs and hormonal methods) in the three hospitals during the
periods evaluated. Santa Cruz exhibited the highest rates and the steepest
increase after training.
Sucre had a slower takeoff but showed a steady increase, attaining
moderately high rates of acceptance. In third place, La Paz had a fairly
Discussion
It is widely accepted that successful family planning programs that in
crease contraceptive prevalence are effective in preventing unwanted preg
nancies and, consequently, induced abortions. The impressive increase of
contraceptive prevalence observed worldwide in the last three decades has
had an important effect on maternal mortality, but there is still room for
improvement. This was clearly demonstrated in Chile where the intro
quick response, but the acceptance rate presented great variability over
duction of family planning led to a significant drop in hospitalizations
time due to several changes in the composition of the staff'during the
due to abortions and in maternal mortality attributed to abortion
project. Surprisingly, hormonal methods had high acceptance, mainly in
Santa Cruz. The services at all three hospitals had a reputation of being
IUD clinics, and it was expected that this method would be favored by
providers. However, the increase in acceptance of pills and injectables
clearly demonstrates that the concept of free informed choice truly was
incorporated by counselors and physicians.
(Barzelatto 1986; Viel 1986). Conversely, in countries where access to
contraception is difficult and antiabortion policies are actively imple
mented, maternal mortality due to abortion increased (Hordet al. 1991).
It has also been demonstrated that targeting family planning to
groups at high risk for unwanted pregnancy can decrease maternal mor
tality (Ickis 1987; Ross, Rich, and Molzan 1989). Unfortunately, although
Juan Diaz et al.
74
Improving Services in Public-Sector Hospitals in Bolivia
75
it is widely recognized that women in the postabortion period constitute
a high-risk group, only a few programs have focused on them.
There are many reasons for the lack of focus on postabortion coun
seling and contraception acceptance. Hospitals and health providers spend
minimal time on counseling, and women often do not return for follow
up. In addition, postabortion care is usually given in wards where family
planning is not routinely offered, and the approach to care is usually strictly
and the taboos regarding abortion combine to make the health needs of
women who have abortions a very low priority. Finally, the United States
government’s Mexico City Policy (promulgated at the 1984 International
Conference on Population and rescinded in 1993) prohibited foreign non
governmental organizations that receive US funds from engaging in most
abortion-related activities.
Improving access to services for women with postabortion compli
curative. From the clinical point of view, sometimes a womans health is
precarious after an abortion, and she is not in a situation conducive to
receiving counseling about contraception; usually her only concern is re
covering and going back home. From the psychosocial and cultural per
spective, the main difficulties are lack of adequate support in the
postabortion period and little knowledge about contraceptive use after abor
tion. Clinics are also not prepared to talk about sexuality, and there is com
cations is the first step to improving the quality of services. Services offer
ing compassionate, humanized, and technically adequate care will attract
women who will seek care before serious complications arise. It is also
clear that quality of services, including better interpersonal communica
tion, is the main factor that will create a favorable environment for effec
tive counseling and acceptance of contraceptive methods.
monly a lack of knowledge about the return of fertility after an abortion.
The results of this study show that it is feasible to increase quality of
For these and other reasons, postabortion family planning is a ser
services with very modest material input and an emphasis on training,
vice almost nonexistent in the LAC region, although in the 1970s some
oriented mainly to obtaining a change in providers’ attitudes. The results
authors showed that it was feasible to provide postabortion family plan
also show that hospitals where staff were highly motivated and remained
ning and that this service was widely accepted by women (Goldsmith et
in their positions after training showed more improvement than those
al. 1972; Hardy and Hend 1975). In 1990 Profamilia and the Instituto
where personnel left the hospital soon after they were trained. In addi
Peruano del Seguro Social published the results of a program of
tion, the commitment and support of authorities, program monitoring,
postabortion contraception, coordinated by the Population Council, that
and active supervision are crucial in order to obtain and maintain im
showed training and motivation of personnel was critical for the success
provements in quality.
of the program (Profamilia and Instituto Peruano del Seguro Social 1990).
In Santa Cruz, where almost all the personnel trained remained in
These successful experiences with postabortion contraception, show
their jobs through the entire study period, eventually 100 percent ofwomen
ing that modern contraceptive methods are perfectly suitable for use dur
were discharged from the hospital having received counseling on contra
ing this period, suggest that the lack of effective implementation of these
ception. Conversely in La Paz, where most of the trained professionals
programs derives mostly from social and political rather than technical
were replaced at various times after training, the percentage of women
reasons. The main problems stem from three overarching factors that have
receiving counseling increased less than in Santa Cruz and varied over
contributed to isolating abortion from family planning and other repro
time.
ductive health services.
First, the legal and administrative constraints to abortion force many
The importance of motivating personnel was also clearly illustrated
in Santa Cruz. At the initiation of the study, counseling activities were
women to seek care in clandestine environments, where family planning
performed by the registered nurse trained for that activity. She was a highly
is usually not available. In addition, womens low socioeconomic status
qualified professional, but she was not motivated to implement the project
because she had several other responsibilities and counseling was an addi
76
Juan Diaz ct al.
Improving Services in Public-Sector hospitals in Bolivia
77
tional burden. As a consequence, counseling rates were very low in the
and continuous supervision. This translated to a slow but steady increase
first two months. The supervisor detected the situation and replaced her
in contraceptive acceptance.
with a nurse auxiliary, who was trained by the nurse and was highly mo
Those who oppose the promotion of postabortion contraception
tivated. The results were impressive and immediate: Counseling rates in
argue that in this period women are extremely vulnerable and will accept
creased to 100 percent in a short period. This person was such a good
whatever is offered, without making an informed choice. Coercion to
counselor and so popular in the hospital that women in the postpartum
accept a method may occur, as has been demonstrated in selected case
period also began to atteno the postabortion counseling sessions.
studies. However, our data show that, when training is adequate and coun
From the point of view of users, counseling is the key issue in qual
ity of services. Users’ evaluation of quality showed a clear positive rela
seling is properly given, women can indeed freely choose the method that
best fits their current needs, or choose not to use contraception at all.
tionship with the quality ofcounseling and with providers’ attitudes. Warm
This is evident in Santa Cruz, where IUDs are traditionally the most
and compassionate providers were almost always evaluated as excellent
prevalent method, in part because it is the method strongly promoted by
providers.
Contraceptive acceptance increased dramatically in the three hospi
widely accepted. To our surprise, in the last period evaluated, hormonal
the hospital. We expected that IUDs would be by far the method most
tals soon after initiation of the project. In the first six-month period after
methods had a higher acceptance than IUDs. Condoms also had a signifi
training, contraceptive acceptance reached 73.2 percent in Santa Cruz,
cant acceptance (condom data included as part of other methods).
61.2 percent in La Paz, and 46.9 percent in Sucre. With the exception of
It has also been argued that there is no need to offer contraception
La Paz, where various trained professionals were replaced, contraceptive
immediately after abortion. Rather, women can come to a follow-up visit
acceptance continued increasing, reaching 87.6 percent in Santa Cruz
one month later, at which point they can arrive at a decision on method
and 81.9 percent in Sucre. It is clear that contraceptive acceptance de
choice and initiate its use. This can be risky for women for two reasons,
pends more on the quality of services, particularly counseling, than on
however. First, it has been demonstrated that ovulation frequently occurs
the characteristics of the population, as demonstrated by the negative
during the first four weeks after an abortion, and most women resume
impacton acceptance of the replacement of trained personnel observed in
sexual activities before the first menses. This places them at risk of becom
La Paz.
The differences between the three hospitals were due more to the
ing pregnant. Offering contraception one month after abortion is effec
tive only ifwomen do not resume sexual activities for 30 days. Also, despite
influence of external factors. In La Paz, the turnover of hospital authori
improvement in the quality ofcounseling, the percentage of women who
ties, in addition to high staff turnover, led to changes in the policy of
returned for a scheduled follow-up visit was not very high in our study. In
service delivery. One of the directors of the hospital, who was in charge
the hospital in La Paz, which had the highest return rate, it was only in
for only a short period of time, attempted to suspend the program and
the third period evaluated that the return rate came close to 30 percent.
removed the trained staff.
In Sucre, the difficulties were mainly with members of the Judi
throughout the study period, confirming the fact that it is very difficult to
ciary, who insisted that physicians denounce women inducing abortion.
In the other two hospitals the return rate remained under 30 percent
reach high levels of return. It is still more difficult to maintain high levels
It was not an easy process, but physicians gradually stopped denouncing
of return in routine work, when the stimulus of the research project has
women and adopted a caring attitude, which was stimulated by training
ended. Moreover, a great proportion of women live far from the hospital
Juan Diaz et al.
78
and do not have access to services designed to deliver IUDs and other
clinic-based methods to them. While women are provided information
on what they should do and where they should go if they experience com
plications, the health facilities located closer to their homes often are not
able to provide contraceptive methods; therefore, it is important to pro
vide contraceptive services while these women are in the hospital.
In summary, our findings indicate that it is feasible to implement a
postabortion program that significantly improves the quality of care, in
cluding contraception acceptance, in the poscabortion period with scarce
material inputs, even in resource-constrained facilities.
The key elements for obtaining these results are:
• training providers on technical issues but, more importantly, on
how to give considerate and compassionate care;
• improving providers’ communications skills;
• encouraging continuous support from the administrative and tech
nical authorities; and
• providing refresher training and continuous active supervision.
The best indicator of the success of this project has been the request
from the MOH and the directors of the womens health division, who are
not those who authorized and initiated the project, to maintain the ac
tivities in the hospitals already involved and to replicate the project in
other regions of the country.
The current challenge for the three hospitals is to maintain and
improve the quality of services they now provide after the resources of the
project have been exhausted and the service activities become part of the
normal routine.
References
Bailey, P.E., L. I Janos, L. Kushner, M. Welsh, and B. Janowitz. 1988. “A hospital study of
illegal abortion in Bolivia," PAHO Bulletin 22(1): 27-41.
Barzelatto, J. 1986. “Abortion and its related problems.” in Infertility Male and Female,
eds. S.S. Ratnam, Eng-Soon Teoh, and C. Anandakumar. Carnforth, United King
dom: The Parthenon Publishing Group, pp. 1—6.
Benson, J.V. 1996. “Meeting women’s needs for post-abortion family planning. Framing
the questions,” Issues in Abortion Care, no. 2. Carrboro, N.C.: Ipas.
Improving Services in Public-Sector Hospitals in Bolivia
79
Camacho, V., A. Galvez Murillo, M. Paz, R. Simmons, A. Young, X. Machicao, L.
Bahamondcs, M.Y. Mackuch, M.D. Castro, J. Diaz, and J. Skibiak. 1996.
Expandiendo opciones de Planificacidn Familiar: Diagndstico cualitativo de la Atencidn
en Salud Reproductiva en Bolivia [Expanding family planning options: qualitative
diagnosis of reproductive health care in Bolivia]. Geneva: WHO.
Encucsta Nacional de Demografia y Salud (ENDSA). 1994. Demographic and Health
National Survey, Bolivia. La Paz, Bolivia: National Institute of Statistics.
Goldsmith, A., R. Goldberg, H. Eyzaguirre, and C. Lizana. 1972. “Immediate post
abortal intrauterine contraceptive device insertion: A double-blind study,” Ameri
can Journal of Obstetrics and Gynecology 112(7): 957-962.
Hardy. E. and K. Hend. 1975. "Effectiveness of a contraceptive education program for
post-abortion patients in Chile," Studies in Family Planning 6(7): 188-191.
Hord, C„ H.P. David. E Donnay, and M. Wolf. 1991. “Reproductive health in Roma
nia: Reversing the Ceausescu legacy," Studies tn Family Planning 22(4): 231-140.
Ickis, J. 1987. "Structural issues related to delivery systems" in Organizing for Effective
Family PlanningPrograms, R.J. Lapham and G.B. Simmons, eds. Washington, D.C.:
National Academy Press, pp. 145-159.
Ministerio de Relaciones Exteriores y Culto (MRF.C). 1994. Informe Sobre el Avance de
las Mujeres en Bolivia [Report on the advancement of women in Bolivia]. La Paz,
Bolivia: Ministerio de Desarrollo Humano.
Profamilia and Instituto Peruano del Scguro Social. 1990. Immediate Post-partum and
Post-abortion Family Planning Program. New York: Population Council.
Rance, Susanna. 1990. Planificacidn Familiar: Se Abre el Debate [Family planning: open
ing the debate]. La Paz, Bolivia: Consejo Nacional de Poblacidn.
---------- . 1993. “Necesidades de informacidn sobre el aborto: Rcflexiones a partir de un
estudio hospitalario" [Information needs on abortion: reflections based on a hospi
tal study], presented at a seminar sponsored by UMSA, Ipas, and the faculty of
medicine of La Paz.
Ross, J. A., M. Rich, and J. Molzan. 1989. Management Strategies for Family Planning
Program. New York: ('enter for Population and Family Health, School of Public
Health, Columbia University.
Taborga, Celia, Bertha Pooley, and Luisa Rada. 1988. "El aborto en Bolivia" (Abortion in
Bolivia], Discussion Document no 2. La Paz: Centro de Investigacion, Educacion
y Servicios.
UNFPA. 1990. Programme Review and Strategy Development Report, Bolivia. New York:
UNFPA.
UNICEF. 1992. Analysis of the Situation of Children and Women in Bolivia. La Paz:
UNICEF.
USAID. 1990. Bolivia project paper, reproductive health services. AID/LAC/P-540.
Washington, D.C.: USAID.
Viel, B. 1986. “Induced abortion in Latin America: Impact on health," in Prevention and
Treatment of Contraceptive Failure, U. Landy and S.S. Ratnam, eds. New York:
Plenum Press, pp. 67-72.
Advances in Contraception 5 (1989) 101-119
© 1989 Kluwer Academic Publishers. Printed in the Netherlands
Reproduced with permission from Kluwer Academic Publishers
Insertional pain and other IUD
insertion-related rare events for breastfeeding
and non-breastfeeding women - a decade’s
experience in developing countries
C. CHI, L.R. WILKENS, C.B. CHAMPION, R.E. MACHEMER and R. RIVERA
I.
Family Health International, PO Box 13950, Research Triangle Park Branch, Durham,
North Carolina 27709, USA
Abstract
The possible effect of breastfeeding on intrauterine device (IUD) insertion
events was investigated. Analysis included a total of 6493 women who enrolled
in multicentcr IUD clinical trials over a tcn-ycar period. Findings indicate that
breastfeeding exerts a protective effect on the incidence of moderate to severe
insertional pain and reduces the need for cervical dilatation to facilitate
insertion. The pain protection effect was most evident in breastfeeding women
who were still in lactational amenorrhea. Subjects with amenorrhea, both
breastfeeding and non-breastfeeding, had a significantly lower incidence of
pain at IUD insertion than the corresponding menstruating subjects. This
effect may be related to a higher secretion of beta-endorphin in the
breastfeeding and lactational amenorrheic subjects.
Introduction
Breastfeeding provides health benefits to infants. It is a natural, convenient and
cost-effective contraceptive method; and by lengthening the birth interval, it also
provides health benefits to mothers (1). The health benefits of breastfeeding arc
especially important in developing countries.
The duration of the contraceptive effect of breastfeeding is, however, variable.
Accordingly, women who want an extended birth interval or no more children need to
switch to another contraceptive method before ovulation resumes. The intrauterine
device (IUD) is a good contraceptive choice for these women, particularly those who
are not yet ready for a permanent and generally irreversible sterilization. The efficacy
of IUDs is far superior to that of barrier methods and, unlike combination oral
contraceptives (OCs), IUDs do not exert a systemic effect that adversely affects
101
y
2(jJ
102
Chi et al
lactation [2-4]. Equally important, studies have shown that breastfeeding during IUD
use does not adversely affect the IUD’s performance [5,6]. For women in developing
countries, the fact that one insertion of an IUD can provide a considerably
long-lasting contraceptive effect makes it a much more desirable contraceptive
method than barriers or orals, which demand regular supply and compliance, and arc
generally more costly. In the most recent World Health Organization (WHO) report,
the IUD was recommended as the best contraceptive method for lactating women [7].
Attention has recently been directed to the relationship between breastfeeding and
IUD insertion-related problems. A frequently cited case-control study conducted on
US data revealed a 10-times higher risk of IUD-associated uterine perforation for
lactating women as opposed to non-lactating women [8]. Two recent case-control
studies using an international data set, on the other hand, suggested a rather
unexpected protective effect of breastfeeding on severe pain occurring at IUD
insertion [9] and a reduction in the need for cervical dilatation to facilitate insertion
[10]. Similar case-control analyses on this international data set did not delineate
breastfeeding at IUD insertion as cither a deleterious or beneficial factor for other
IUD insertion-related events such as failed insertion [11], syncope and other
vasovagal reactions [12], uterine perforation [13] and cervical laceration (Chi et al.t in
preparation).
The relationship between breastfeeding and IUD insertion is programmatically
important and needs clarification. Both breastfeeding and IUD use are prevalent in
developing countries, and many breastfeeding women in these countries use IUDs.
Women having characteristics leading to a smooth IUD insertion are an ideal group
to encourage others to use this method, and, conversely, any distressing events
experienced during insertion can directly affect the acceptability of a family planning
program offering IUDs. The fall of the Singapore IUD Action Program following a
rumor of a high incidence of uterine perforation is one unfortunate example [14J.
With more than 80 million women worldwide using IUDs (60 million of them residing
in China) [15], even rare IUD insertion-related events can be translated into a
significant public health problem*.
Although the case-control analysis approach is generally the most useful method,
and sometimes the only feasible one, for the delineation of risk factors for rare events,
an important limitation is that complete elimination of bias cannot be assured [17].
Findings from this study approach usually need to be replicated by other case-control
or prospective studies before they can be accepted. We believe the international IUD
database developed by Family Health International (FHI) during the last ten years is
sufficiently large to examine these rare events 'prospectively* and to determine if the
results thereof generally agree with the findings of case-control studies.
'According to Irving Sivin, if the risk of pregnancy (also a rare event) can be reduced from five to two per
100 at two years of IUD use in the People’s Republic of China, it could mean a reduction of about 600 000
unwanted pregnancies each year [16]
Insertional Pain and Other IUD Insertion related Rare Events
Methods and materials
The IUD data used in this analysis were collected from international multi-center
clinical trials coordinated by FHI between 1977 and 1986. All participating centers
used identical case record forms and similar protocols.
Our study population was defined as women who were:
1. Parous and whose last pregnancy was a vaginally delivered term live birth,
2.
Users of a common IUD type (i.e., Loops and variants, Copper T, Cu-7 and
Multiload devices) during the interval period (>42 days since last delivery), and
3.
From centers where one of the above IUD types was inserted in al least 100
women during the study period, breastfeeding status at insertion was known for at
least 75% of the acceptors, and at least 10% of these subjects were breastfeeding
al IUD insertion.
Altogether 6493 women were included for study, 3450 of whom were not
breastfeeding (NBF) and 3043 of whom were breastfeeding (BF, including partially
breastfeeding, defined as breastfeeding with supplementary food) al the lime of IUD
insertion. Eighteen international centers were included, seven located in Asia, seven
in Latin America and four in the Middle East.
In this analysis, the IUD types were pooled into three categories according to
configuration, namely: Loops, T-shaped devices (including the Cu-7) and Multiloads,
which arc horseshoe-shaped devices. Our study population comprised the following
number of women, grouped by type of IUD inserted:
IUD type
Number of users
Loops
LLD
Photo-reduced’
Tapered *
LLC
LLD with copper’
Subtotal
T-shaped devices
TCu200
TCu380A
TCu380Ag
Cu-7
90-1
316
237
134
130
1721
1616
1362
363
106
Subtotal
Multiloads
250
375
Subtotal
3447
945
380
1325
' These experimental IUD types were developed by Fill and used for a short time only. They were of the
same shape as the Lippes Loop and were included for study to increase the sample size.
Chi et al
104
IUD insertion-related events (the outcome variables) examined in this analysis
included: moderatc/sevcre insertional pain, cervical dilatation required to facilitate
insertion, cervical laceration, syncope, insertion failure and uterine perforation.
Analysis focuses upon insertional pain because of its relatively high incidence, its close
relationship with other insertion events [18], as well as its strong negative association
with breastfeeding as revealed from our previous case-control study [9],
Patient characteristics and characteristics of the situation surrounding IUD
insertion were first examined between the NBF and BF groups. Incidences of IUD
insertion-related events were then compared between groups by univariate analysis,
stratification and logistic regression [19]. Degree of association was computed using
relative risks (RRs) or odds ratios (ORs) derived from logistic regression. A relative
risk or odds ratio with 95% confidence limits (CLs, two-tailed) excluding unity was
considered statistically significant. The NBF women were used as the reference group.
Results
1. Characteristics of BF and NBF women (Table 1)
Compared to the NBF women, the BF women were, in general, two years
younger; less likely to have used contraceptives, especially oral contraceptives, in
the month prior to this IUD insertion; and more likely to be in the
postpartum/lactational amenorrhea period. The BF women, as would be expected,
had a much shorter open interval (months between ending of last pregnancy and
IUD insertion). Also, the BF women were somewhat more likely to have a
retroverted uterus. Other variables such as number of live births, educational level,
proportion living in urban areas, proportion wanting more children, type of IUD
inserted and type of inserting personnel (obstetrician/gynecologist vs other types
of insertors) were generally similarly distributed between the two groups. Among
women who had resumed menses, the liming of IUD insertion in relation to
menstrual cycle was also similar.
2. Incidences of insertion-related events, univariate analysis (Table 2)
The BF women were about two times less likely to suffer any degree of insertional
pain, and three times less likely to suffer severe pain as compared to the NBF
women. Per 100 women, 3.5 in the BF group and 7.5 in the NBF group
experienced moderate to severe insertional pain. The incidence of severe pain was
0.6 and 2.0 per 100 women in the respective groups. Similarly, cervical dilatation
was also less likely to have been necessary for BF women than for NBF women.
The respective incidences per 100 women were 0.4 and 1.9. Using NBF as the
reference group, the relative risk (RR) for moderate/severe pain for BF women
was 0.47 (for severe pain only, RR = 0.30) and that for cervical dilatation was 0.21.
All of the 95% CLs for these relative risks excluded unity.
Insertional Pain and Other IUD Insertion-related Rare Events
Table 1 Selected characteristics of the non-breastfeeding and breastfeeding groups
Non-breastfeeding
women (n = 3450)
Breastfeeding
women (n - 3043)
A. Patient characteristics
Mean age in years (SD)
Mean live births (SD)
Mean education in years (SD)
% Living in urban area
Contraceptive method used in last month (%)
IUD*
Orals
Injectablcs
Uterine position (%)
Antcvertcd
Retroverted
Midpositioned
Not determined
2.5(16)
6.8 (73)
13.1
53
10.2
29.0
61.0
14.7
25
B. IUD insertion-related characteristics
% Timing of insertion in relation to
menstrual cycle (in days)
18 +
Lactational/postpartum amenorrhea
Unspecified
Open interval (%)
<6 months
6-11.9 months
12-23.9 months
24+ months
Type of insertor
Obstetrician/Gynccologist
Other physician
Nurse/midwifc
Others
Type of IUD inserted*** (%)
T-shaped devices
Multiloads
42.8 (68.3)
17.0(27.1)
2 9(4.6)
36.2
61.3
227
12.8
3.2
73.3
11.0
13.0
24.3
53.6
22.1
The current insertion was thus a reinsertion tor these subjects. The exact length of the interval
between lenninalion of last IUD use and the current insertion is unknown except that the interval
should not be longer than one month
Percentage distribution in parentheses is limited Io women who resumed menses at insertion
Loops include Loops C and D, loop D with copper, the Tapered loop and the Photo-reduced loop.
Copper devices include Cu-7, TCu200, TCu 380A and TCuMOAg. Mulliloads include Mulliload
Cu250and375
Chi el al
106
Table 2 Crude incidences* of various IUD insertion-related rare events by breastfeeding status at
insertion
Non-breastfeeding
*(n=3450)
No.™
Breastfeeding
women (n-3043)
No.
%
Relative risk***
(95% CLs)
Pelvic pain
Moderate or severe
Severe only
259
69
731
2.00
107
18
332
039
0.47 (0.37-039)
0.30 (0.17-031)
Dilatation required
65
1.93
12
0.40
0.21 (0 11-039)
Cervical laceration
22
0.64
14
0.46
0.72 (035-1.47)
Syncope
12
0.35
13
0.43
1.23 (033-2 86)
Events
Insertion failure
3
0.09
3
0.10
1.13(0.18-6.99)
Immediate uterine perforation
2
0.06
5
0.16
2.83 (0.49-21.04)
Any events except immediate
uterine perforation"”
339
9.83
139
437
0.46 (0.38-036)
Incidences were based on the number of women for whom the event status was known
Non-breaslfeeding women were used as the reference group
Multiple events may be reported for the same woman
Incidences of insertion failure, syncope and cervical laceration were generally
similar between the two groups of women. The incidence of immediate uterine
perforations was very low in both groups, but was slightly higher in the BF women
than in the NBF women. The difference, however, was not statistically significant.
Considering all women with one or more IUD insertion-related events except
uterine perforation, the BF women were half as likely as the NBF women to have
an insertion event (RR = 0.46,95% CLs = 0.38-0.56).
3. Stratification
The observed greater risk for NBF women of incurring insertional pain and
cervical dilatation could have been biased because of differences between BF and
NBF women in patient characteristics, in the situational factors surrounding IUD
insertion, or in the type of IUD used. We therefore conducted an analysis of the
study events stratified by these factors.
Insertional Pain and Other IUD Insertion-related Rare Events
107
Table 3 Incidence of moderate/severe insertional pain by patient characteristics IUD insertion
situational factors and breastfeeding status at insertion
Non-breastfeeding women *
(n ==3450)
Total No. with
women pain
%**
Age
1139
83
<25
97
25-29
1149
88
6.4
30+
74
1158
Parity
1-2
2097
156
3+
7.6
1351
103
Education
0-3 years
6.2
1185
73
4-9 years
141
10.2
1386
10+ years
45
5.1
875
Wanting additional children
1859
No
137
7.4
Yes
1587
7.7
122
Contraception used in month prior to insertion
IUD
19
4.2
450
OC
1466
139
93
Others
(including none)
101
6.6
1531
Uterine position
Anteverted
179
2105
83
Retrovcrtcd
748
36
4.8
Midpositioned
509
41
8.1
Timing of insertion in relation Io menstrual cycle
<5 days
2269
218
9.6
6-17 days
24
751
3.2
18+ days
164
6.1
10
Amcnorrheic
219
3
1.4
Open interval
<6 months
700
43
6-12 months
550
38
6.9
13-24 months
659
50
7.6
25+ months
1539
128
8.3
IUD type at this insertion
Loops
981
190
19.4
T-shaped
1814
66
3.6
Multiloads
653
3
03
Insertor type
OB/GYN
2341
54
2.3
Others
1106
204
18.4
Breastfeeding women
(n-=3043)
Total No. with
women pam
%**
Relative risks*
(95% CLs)
1547
911
585
70
18
19
43
2.0
3.2
0.53 (0.39-0.72)
0.26(0.15-0.43)
031 (0.30-0.85)
1978
1065
69
38
33
3.6
0.47 (0.35-0.62)
0.47 (0.32-0.68)
1112
1220
701
46
54
7
1.0
0.67 (0.46-0.98)
0 44 (0.32-0.60)
0.19(0.08-0.44)
1526
1515
57
50
3.7
3.3
0.51 (0.37-0.69)
0.43 (031-0.60)
249
530
7
19
2.8
3.6
0.66(0.26-1.64)
0 38(0.23-0.61)
2263
81
3.6
0.54 (0.40-0.73)
1729
832
316
78
14
14
4.5
17
4.4
0.53 (0.40-0.69)
0.35 (0.18-0.66)
055 (0.29-1.02)
1303
518
87
1103
68
11
7
19
5.2
71
8.0
1.7
0.54 (0.41-0.71)
0 66(0.31-1.40)
1.32(0.46-364)
1.26 (0.36-5.32)
1865
690
388
98
48
30
24
5
2.6
43
6.2
5.1
0.42 (0.27-0.64)
0.63 (0.38-1.03)
0.82 (0.49-1.33)
0.61 (0.22-1.48)
739
1632
672
86
18
3
11.6
ii
0.4
0 60 (0.47-0.78)
0.30 (0.17-032)
0.97 (0.16-5.99)
2231
811
23
84
1.0
10.4
0.45 (0.27-0.74)
036 (0.44-0.72)
Non-breastfeeding women were used as the reference group
The percentages are based on the number of subjects with valid values. Due to some subjects with
unknown values, the totals may not add up to 3450 for non-breastfeeding women and 3(M3 for
breastfeeding women
Insertional Pain and Other IUD Insertion-related Rare Events
108
a. Moderate/severe insertional pain (Table 3)
In most cases, no matter how the data were divided, the BF women were
consistently associated with a lower incidence of moderatc/scvere pain than the
NBF women. It is especially important that this was the case when stratification
was performed by age, contraceptive method used in month prior to this IUD
insertion, length of open interval, and uterine position, because these variables
were (1) differently distributed between the two groups of women and (2) known
to be simultaneously related to both the study variable (breastfeeding) and the
outcome variable (the incidence of insertional pain). Most impressive is the
finding illustrated by Figure 1, that in 12 of the 18* study centers, BF women
consistently had a lower incidence of pain than NBF women. When the incidence
of moderate/severe pain is stratified by breastfeeding status and menstrual status
at insertion, however, the consistent pattern was noticeable only in those women
who had resumed menses. For women resuming menses, the incidences of
moderatc/scvere pain were respectively 4.51 for the BF and 7.91 for the NBF
women (RR = 0.57, and 95% CLs = 0.44-0.73). Among those women who were
still amenorrheic, there was no significant difference in insertional pain between
those who were and those who were not breastfeeding. Amenorrheic women, both
in the BF and in the NBF groups, had a lower incidence of insertional pain than
the corresponding menstruating women. Most of the non-menstruating and
non-breastfeeding women were still in the period of lactational amenorrhea.
b. Need for cervical dilatation (Table 4)
Consistent with our findings on insertional pain, IUD insertions requiring cervical
dilatation were, in general, also found to be less frequent in BF than NBF women.
This finding was unaffected by age, parity, educational level, whether the IUD was
inserted for family-spacing or limiting purposes, previous contraceptive method
used, type of inserting personnel, IUD type used, or length of open interval. Also,
the incidence of cervical dilatation was lower in the BF than the NBF group,
irrespective of whether the woman was still amenorrheic or had resumed menses
at the time of insertion. Seven of the 18 centers* * had a lower proportion of
cervical dilatation for the BF women than for the NBF women (Figure 2).
•Among the six centers not exhibiting the consistent patterns (Figure 1), three did not report any cases of
moderate/severe insertional pain in either group. One center reported identical incidences between the
two groups. In another center, there were only 35 NBF women, none of whom reported pain. One of the
122 BF women in this center complained of pain. Only the last center had adequate numbers of NBF
(n « 396) and BF (n «369) women; one of the NBF women and three of the BF women reported pain.
•*In another eight centers, there were no dilatations in either group. One center reported identical
incidences in the two groups of women. Only in the remaining two centers was the incidence of cervical
dilatation higher for BF than for NBF women (one center had <80 women in each group).
Table 4 Incidences of cervical dilatation performed to facilitate IUD insertion by patient characteristics,
IUD insertion situational factors and breastfeeding status at insertion
Non-breastfeeding women •
= 3450)
Total No. with
women dilatation %••
Breastfeeding women
(n '■ = 3043)
Total No. with
women dilatation %•'
Relative risks*
(957c Cl
1095
1131
1145
26
25
14
2.2
1.2
1518
90-1
577
2043
1330
46
19
22
I4
1944
1055
Age
<25
25-29
30 +
Parity
Education
0-3 years
0.6
1122
4-9 years
1377
22
10+ years
3^i
4.1
872
Wanting additional children
No
1833
31
Yes
34
1538
2.2
Contraception used in month prior to insertion
IUD
447
12
OC
1428
28
2.0
Others
(including none)
1497
25
1.7
Uterine position
Antevertcd
2071
42
2.0
Ret rove rted
733
12
1.6
Midposition
487
9
18
riming of insertion in relation to menstrual cycle
<5 days
2198
39
1-8
6-17 days
747
18+ days
165
8
4.8
I
Amenorrheic
218
0.5
Open interval
<6 mos.
671
1.3
6-12 mos.
1s
534
8
12-23 mos.
646
14
2.2
24 +
1522
34
22
IUD type at this insertion
Loops
966
10
1.0
T-shaped
1755
33
Multiloads
652
22
3.4
Insertor type
OB/GYN
2274
51
2.2
1098
14
1.3
•
3
A
3
0.2
0.7
05
0 08 (0 02-0.29)
0.30 (0.11-0.76)
0.42 (0.10-156)
0.14 (0.05-0.33)
0.40 (0.14-1.05)
6
03
0.6
1074
1214
701
5
0.4
0.2
0.7
0.60 (0.15-2.25)
0.15 (0.04-0 54)
0.17 (0.06-0.46)
1510
1487
(y
6
0.4
0.4
0 23 (009-0.59)
0.18 (0.07-0.45)
246
515
1
1
0.4
0.2
0.15 (0.01-1 10)
0.10 (0.01-0.67)
2237
10
0.4
0.27 (0.12-058)
1708
828
300
g
3
1
05
0.4
0.3
0 23 (0.10-0 51)
0 22 (0.05-0 83)
0.18 (0.01-136)
1263
515
87
1102
8
3
0
0.6
0.6
0.0
0.1
0.36 (0.15-0.79)
0.26 (0.06-0.91)
1847
674
380
96
A
s
]
0
0.3
0.7
0.3
0.0
0.24 (0.08-0.74)
050 (0.14-1.65)
0.12 (0.01-0.87)
727
1600
672
7
c
0
1.0
0.3
0.0
0.93 (0.32-2.63)
0.17 (0.06-0.44)
2192
806
c
4
0.4
05
0.16 (0.07-0.35)
0.39 (0.11-1.26)
0.20 (0.01-7.22)
Non-breastfeeding women were used as the reference group
The percentages arc based on the number of subjects with valid values. Due to sorrtc subjects with
unknown values, the totals may not add up to 3450 for non-breastfeeding women and 3043 for
breastfeeding women
•'• The relative risk cannot be calculated
Chi et al
110
c. Other IUD insertion-related events (Table 5)
Inscrtional Pain and Other IUD Insertion-related Rare Events
Table 5 Incidences of syncope and immediate uterine perforation at IUD insertion by breastfeeding
status and IUD type
Incidences of other events were too low for stratified analysis. Stratification by
IUD type showed a somewhat higher risk of immediate uterine perforation and a
lower risk of syncope for BF as compared to NBF women, both for Loop users
only. Neither of these differences was statistically significant.
Non-breastfeeding women'
Total
No. with
Total No. with
women event
Device type
4. Multivariate analysis, logistic regression (Table 6)
Stepwise logistic regression analysis was used to examine the independent effect of
breastfeeding on the two events, modcratc/scvere inscrtional pain and need for
cervical dilatation. Breastfeeding status (yes vs no) was forced into each model,
and then other covariatcs were allowed to enter. The variables with the
opportunity to enter the model were: age (<25 vs >25 years), IUD use, OC use,
menstrual status (amenorrhea vs resumed menses), open interval (< 12 months vs
>12 months), and Center (Center O vs other centers for pain, Center A vs other
centers for dilatation). For models of inscrtional pain and need for dilatation,
center was the first covariatc to enter the model. Two other covariates were
important in modelling pain: open interval and OC use; two other covariates also
entered the dilatation model: menstrua! status and age.
When center was the only additional covariate in the inscrtional pain model,
the adjusted odds ratio for breastfeeding was 0.55 (95% CLs = 0.43-0.71). After
all three of the additional important covariatcs had entered this model, the effect
of breastfeeding was reduced (odds ratio = 0.76, 95% CLs = 0.57-1.02). For the
dilatation model, the adjusted odds ratio for breastfeeding was 0.18
(95% CLs = 0.09-0.34) when center was the only additional covariatc, and was
0.51 (95% CLs = 0.25-1.01) when all three of the additional important covariatcs
were in the model.
5. The effect of degree of breastfeeding (Table 7)
We further divided the BF women into full and partial breastfeeding to sec if there
was a ‘dose-response’ in the breastfeeding effect on pain and dilatation. Table 7
shows that both breastfeeding groups had lower incidences for both events than
the non-breastfeeding group, but full breastfeeding offered no extra protective
effect as compared to partial breastfeeding. This is a crude breakdown of
breastfeeding since we have no duration or frequency information.
Breastfeeding women *
(n=3O43)
A. Event ofsyncope
T-shaped devices
Multiloads
981
1815
653
0.618
000
672
B. Event ofimmet 'iatc uterine perforation
*----982
0
T-shaped devices
1815
1
Multiloads
653
I
0.00
0.06
0.06
0.07)
I.(XX)
I.(XX)
•Due to some unknown values, the totals may not add up to 3450 for non-brcasifccding women and 3013
for breastfeeding women
Table 6 Adjusted odds ratio and 95% confidence limits (CLs) for the effect of breastfeeding at IUD
insertion on (A) moderate/severe insertional pain and (B) need for cervical dilatation
Odds ratio*
for breastfeeding women
95% (7.<
A. Moderate/severe inscrtional patn
Controlling for
0.43-0.71
b. Center and two other
confounding variables:
0.76
Open interval (<12 months vs > 12 months)
and OC use in month prior to insertion
B. Cervical dilatation
Controlling for
a. Center (A vs others)
0.18
b. Center and two other
confounding variables:
051
Menstrual status (amenorrhea ys resumed menses)
and age (< 25 vs >25)
057-1.02
0.059
0.09-0.34
: 0.001
0.25-1 01
0.043
•Non-breastfeeding women were used as the reference group. The odds ratios were derived from a
stepwise logistic regression model with the insertion event (pain or dilatation) as the dependent variable
and with breastfeeding status forced to enter the model
Chi et al
112
Table 7 Incidence of (A) moderate/severe insertional pain and (B) cervical dilatation performed to
facilitate IUD insertion by extent of breastfeeding at insertion
Total
women •
No. with
A. Event ofmoderate/severe insertional pain
Non-breastfeeding
Partial breastfeeding
Full breastfeeding
3448
1362
1681
B. Event ofcervical dilatation
Non-breastfeeding
Partial breastfeeding
Full breastfeeding
3373
1328
1671
%
Relative risk
(95% CLs)
259
29
78
731
2.13
4.64
1.00
0.30(0.20-0.44)
0.63(0.49-0.82)
65
193
0.30
0.48
1.00
0.16(0.05-0.45)
0.25(0.11-034)
8
’Due to some unknown values, the totals may not add up to 3450 for non-breastfeeding women and 3043
for breastfeeding women
Discussion
Our findings from this ‘prospective’ study generally agree with findings from previous
case-control studies. Breastfeeding exerts a protective effect on the incidence of
moderate/severe insertional pain, and also a beneficial effect of reducing the need for
cervical dilatation to facilitate IUD insertion. An association between breastfeeding
and other IUD insertion-related rare events, namely insertion failure, syncope and
cervical laceration, was, however, not delected. Our analysis docs suggest that uterine
perforation at IUD insertion may be more likely to occur in a BF woman receiving a
Loop device.
The study designs of the case-control approach and the prospective approach
have individual strengths and weaknesses. They arc, however, methodologically
complementary and when they produce similar findings, the validity of the findings is
greatly enhanced. Two additional aspects strengthen the validity of our findings: (1)
the outcome variables under study are those that occurred during and were recorded
immediately after IUD insertion, so the reporting of these events is likely to be
complete and not subject to recall bias (for the case-control approach) or subject to
bias due to loss of follow-up (for the prospective approach) and (2) the potential
beneficial effect of breastfeeding on reduction of insertional pain and the need for
cervical dilatation had not been suspected previously. Therefore, bias due to selective
reporting by the women or prejudice of the insertors is not likely. That similar results
were derived when controlling for potentially confounding variables, either one at a
time through stratification or simultaneously through multivariate analysis, further
strengthened the validity of our findings.
These findings are also supported by results from an experimental study which
actually measured the IUD insertion force in 103 parous women [20]. The study found
Insertional Pain and Other IUD Insertion-related Rare Events
113
that significantly less insertion force is needed to insert an IUD in breastfeeding,
recently delivered women (1.75 newtons) compared to non-breastfeeding, long
delivery interval women (2.8 newtons). In that study, breastfeeding and
non-breastfeeding women were matched by IUD type and parity. Due to few
non-breastfeeding women with recently delivered infants, the independent effect of
breastfeeding controlling for the length of delivery interval (i.c., open interval) could
not be evaluated.
One limitation of our data set is the lack of information regarding the length of
time women had breastfed in the past (for NBF women) or had been breastfeeding
(for BF women) at the time the IUD was inserted. It is possible that, in our study, an
NBF woman might have breastfed for quite some lime but stopped immediately
before this IUD insertion. While it is also possible that some women might have just
started breastfeeding before this IUD insertion and were duly categorized as
breastfeeding cases, it is thought highly unlikely. If the effect of breastfeeding on IUD
insertion takes some time to appear (and to disappear), this ‘misclassification’ effect
could produce bias; this bias would, however, underestimate rather than overestimate
the differences between the BF and NBF groups. Frequency and intensity of
breastfeeding could only be crudely measured by whether the breastfeeding was full
or partial.
We are also cognizant of the intcrccntcr variation and center clustering pattern in
the reporting of insertional pain and cervical dilatation. The results from our analyses
were probably heavily influenced by one center’s data (Center O for pain and Center
A for dilatation) and may lack representativeness. However, the most powerful
evidence supporting the asserted breastfeeding and insertional pain association was
the lower incidence of insertional pain for BF women across virtually all 18 study
centers (Figure 1). To some extent, this was also true for the Finding on cervical
dilatation (Figure 2). No bias could produce such a consistent finding since most
service providers and IUD receivers would not have expected this effect. The
beneficial effect of breastfeeding on both events remained when the center effect was
adjusted in the multivariate analysis. Further adjusting for other confoundcrs, namely
previous OC use and open interval for insertional pain, and menstrual status and age
for dilatation, somewhat mediated the breastfeeding effect, probably due to the close
correlation of the variables with breastfeeding.
Insertional pain appears to be associated with all other insertion-related events,
whether as an effect (pain may be caused by cervical laceration or cervical dilatation)
or as a cause (pain probably is an immediate cause for syncope and insertion failure)
(Table 8). One possible reason that we did not detect a similar protective relationship
between breastfeeding and these other events is that their incidences in both the BF
and NBF groups were too low and the differences too minimal for a study of this size
to detect.
Among the rare events under study, uterine perforation is the one with the most
potentially serious medical consequences. Our suggestive finding that its risk is higher
in breastfeeding women inserted with a Loop device, although in agreement with
Heartwell and Schlcssclman’s study results (8|, was based on a very small number of
occurrences. An experimental in vitro study reported by Goldstuck, however, gives
38.0
Figure 1 Incidences of mode rate/severe insertional pain by breastfeeding status at insertion and by center
Figure 2 Incidences of cervical dilatation required to facilitate IUD insertion by breastfeeding status at insertion and by center
Chi ct al
116
Table 8 Interrelationship between insertional pain and other insertion-related events by breastfeeding
status at IUD insertion
A. Events leading to insertional pain
Cervical dilatation
(+)
(-)
Cervical laceration
(+)
(-)
Women with modcrate/severe insertional pain
Total*
No.
%
77
14
18.2
6295
345
53
36
6456
5
361
139
5.6
(+)
(-)
Women with syncope
Total*
366
6125
No.
12
13
%
(*)
(-)
Women with insertion failure
Total*
No.
i
366
6125
5
D. Insertional pain leading to other events
Moderate/severc pain
Modcrate/severe pain
0.2
%
0.3
0.1
•Due to some unknown values, the total may not add up to 6493 study subjects
support and provides an explanation for why the Lippcs Loop*, as compared to other
devices, is more likely to be associated with uterine perforation [21], If this association
is true, it is possible that the lack of insertional pain in breastfeeding women makes
inserters less careful during insertion. Less careful attention during insertion, coupled
with the possible biological changes of the uterine wall due to lactation, could
contribute to a greater risk of immediate uterine perforation at IUD insertion for
breastfeeding women.
While not directly relevant, other risk factors delineated in this study were also
delineated in our previous case-control studies [18]. For example, from both types of
study approaches, we found that women who were younger than 25 (vs older women),
had used OCs prior to this insertion (vs those using no or other methods), had an
open interval of one or more years (vs those with a shorter interval) and/or were
inserted with a Loop (vs another IUD type) were more likely to suffer
moderate/severc insertional pain. More relevant, however, is the general finding that
the relative effect of these other risk factors seems to be somewhat diminished in BF
women.
•According to Goldsluck, the inserter tube of the Lippcs Loop is much more rigid than that of other
devices, and the forces produced by the Lippcs Loop are close to the lower range from uterine
perforation experiments in vitro (about 12 newtons). The push mechanism of the insertion may also
contribute to its higher risk of uterine perforation [21].
Insertional Pain and Other IUD Insertion-related Rare Events
117
The finding of a protective effect on the incidence of pain at the lime of IUD
insertion associated with breastfeeding, in particular with breastfeeding and
lactational amenorrhea, and with amenorrhea in breastfeeding and non-breastfeeding
women, is of particular clinical importance. It has been shown that suckling episodes
during lactation stimulate the release of /?-cndorphin in the hypothalamus of the ewe
[22], and that the peripheral levels of ^-endorphin rise after suckling in the rat [23].
This suckling-induced secretion of f?-cndorphin plays an important role in the
suppression of ovarian activity during lactation. Endorphins affect several behavioral
and physiological measures, and ^-endorphin is clearly the most potent of these
substances. It has an analgesic effect when applied centrally that is markedly more
potent than morphine [24], It is possible that a rise in /?-cndorphin secretion resulting
from suckling in the breastfeeding women accounts for the decrease in
insertion-related pain observed in these subjects. It is also reasonable to assume that
breastfeeding women who arc still amcnorrhcic maintained higher levels of
endorphins as a result of more frequent suckling episodes and as a consequence this
group had a higher degree of pain protection.
Estrogens promote uterine contractile activity and prostaglandin formation in the
myometrium. In the vast majority of women, both breastfeeding and non-breast
feeding, who were already menstruating at the time of IUD insertion, the IUD
insertion took place on days 1-17 of the menstrual cycle (Table 1). The estrogen
levels present at this time of the cycle would promote uterine contractibility and
prostaglandin formation in the myometrium. Conversely, amcnorrhcic women, breast
feeding or non-breastfeeding, would be free of this estrogenic effect. Il is possible that
this factor also played a role in the reduced incidence of pain at the time of insertion
observed in the amcnorrhcic women.
With the growing cognizance of the health benefits of breastfeeding and of the
superior use- and cost-cffcctivcncss of IUDs, the number of IUD insertions in
breastfeeding women will probably increase. Our study results indicate that, except
for uterine perforation, interval insertions in breastfeeding women do not seem to be
associated with increased incidences of insertion-related rare events. In fact,
breastfeeding seems to have the beneficial effect of reducing the occurrence of
insertional pain and the need for cervical dilatation. Programmatically, it appears that
IUD insertion during breastfeeding should be encouraged and that compared to
breastfeeding women, non-breastfeeding women may need more intensive counseling.
Further epidemiological studies are definitely needed to clarify whether there is a
causal association between breastfeeding and IUD-associated uterine perforation,
although intuitively, uterine perforation as well as cervical laceration would seem to
be more related to inserter factors than to patient characteristics. Accordingly, the
cardinal rule, as asserted by Hatcher et al. [25], that ‘everything done at the time of
IUD insertion should be done slowly and gently” should be conscientiously observed
for all women. Also, more basic research on the physiological and anatomical changes
of the cervix and uterine wall in breastfeeding women is needed.
118
Chi el al
Acknowledgment
Partial support of this work was provided by Family Health International (FHI) with
funds from the United States Agency for International Development (USAID). The
views expressed in this article, however, do not necessarily reflect those of USAID.
FHI is an international not-for-profit organization which conducts research and
provides technical assistance in health, family planning, STDs and AIDS. It is based in
Research Triangle Park, NC, USA.
The authors wish to express their gratitude to the FHI network of collaborators
for contributing data used in the preparation of this paper. The assistance from Nancy
Robinson in data processing is also gratefully acknowledged.
References
McCann, M.F., Liskin, L.S., Piotrow, P.T., Rinehart, W. and Fox, G. (1984). Breastfeeding, fertility
and family planning. Popul. Rep., (J) 24, 527-567
2.
Hingorani, V. and Uma Bai, G.K. (1970). lactation and lactational amcnorrhoca with postpartum
1UCD insertion. Reprod. Fertil., 23,513-515
3.
Hcfnawi, F., Dannoil, O., Badraoui, M.H.II. et al. ({QIS'). Effect of inert IUDs on lactation. In:
Analysis of Intrauterine Contraception. F. Hcfnawi and S. Segal, eds., North Holland Publishing
Company, Amsterdam, pp. 431-438
4.
Prema, K. (1982). Duration of lactation and return of menstruation in lactating women using
hormonal contraception and IUDs. Conrraccpt. Deliv. Syst., 3,39-46
5.
Cole, LP„ McCann, M F . Higgins, J E and Waszak. C.S. (1983). Effects of breast-feeding on IUD
performance. Am. J. Pub. Health, 73, 384-388
6.
Chi, I-c., Wilkens, L. and Rogers. S. (1985). Expulsion in immediate post-partum insertion of Lippcs
Loop D and Copper T IUDs and their counterpart Delta devices - an epidemiological analysis
Contraception, 32, 119-134
7.
World Health Organization (WHO) (1987). Mechanism ofAction, Safety and Efficacy of Intrauterine
Devices. Geneva, WHO. (Technical Report Scries 753). 1-91
8.
Heartwell, S.F. and Schlcssclman. S. (1983). Risk of uterine perforation among users of intrauterine
devices. Obstet Gynecol., 61. 31-36
9.
Chi, I-c., Galich, L.F. and Tauber, P.F., et al. (1986). Severe pain at interval IUD insertion - A
case-control analysis of patient risk factors. Contraception, 34,483-495
10.
Chi, I-c, Champion, CB. and Wilkens, L.R. (1987). Dilatation of cervix in interval insertions of the
IUD - Who requires it and docs it lead to a high expulsion rate? Contraception, 36. 403-415
11.
Chi, I-c. and Rogers, S. (1983). Failure to insert an intrauterine device. Conrraccpt. Deliv. Syst., 4.
207-211
12.
Chi, I-c., Wilkens, L, Siemens, AJ. and Lippcs, J. (1986). Syncope and other vasovagal reactions at
interval insertion of Lippcs Ireop D. Contraception, 33, 179-187
13.
Chi, I-c., Feldblum, PJ. and Rogers, S.M. (1984). IUD-related uterine perforations - An
epidemiologic analysis of a rare event using an international dataset. Contracept. Deliv. Syst., 5,
123-130
14.
Ratnam, S.S. and Tow, S.H. (1970). Translocation of the loop. In: Postpartum Family Planning: A
Report on the International Program. G. Zatuchni, cd., McGraw-Hill, New York, pp. 371-384
15.
Liskin^LS. and Fox, G. (1982). IUDs: An appropriate contraceptive for many women. Popul. Rep.,
1.
Sivin, I. (1985). Recent studies of more effective Copper T devices. In: Intrauterine Contraception:
Advances and Future Prospects. G.I. Zatuchni, A. Goldsmith and JJ. Sciarra, eds.. Harper and Row,
Hagerstown, pp. 70-78
Sackett, D.L (1979). Bias in analytic research. In: The Case-control Study - Consensus and
Controversy. M.A. Ibrahim, cd., Pergamon Press, Oxford, pp. 51-63
18.
Chi, I-c., Wilkens, L, Siemens, AJ. and Potts, M. (1987). Rare events occurring at insertion of an
intrauterine device - A review of an international experience. Adv. Contracept., 3,49-61
16.
17.
Inscrtional Pain and Other IUD Insertion-related Rare Events
KJcinbaum, D.G., Kuppcr, L.L. and Morgenstern, II. (1982). Epidemiologic Research, Lifetime
Learning Publications, Belmont, California, 447
Goldstuck, N.D. (1988). IUD insertion forces: effect of recent childbirth and lactation. Adv.
Contracept., 4,159-164
Goldstuck, N.D. (1987). 'Bowing' forces with IUD insertions in vitro: Relevance to different IUD
insertions. Clin. Reprod. Fertil., 5,173-176
22.
Gordon, K., Rcnfrcc, M.B., Short, R.V. and Clarke, I.J. (1987). Hypothalamo-piluilary portal blood
concentrations of /t-endorphin during suckling in the ewe. J. Reprod. Fertil., 79, 397-408
23.
Riskind, P.N., Millard, WJ. and Martin, J.B. (1984). Evidence that thyrotropin-releasing hormone is
not a major prolactin-releasing factor during suckling in (he rat. Endocrinology, 115, 312-316
24.
Loh, H.H. and Li, Ch.II. (1977). Biological activities of ^-endorphin and its related peptides. Ann. NY
Acad. Sci., 297, 115-130
25.
Hatcher, R.A., Guest, F., Stewart, F. et al. (1986). Contraceptive Technology, 1986-1987. 13th cdn.
Irvington Publishers, New York. 85
19.
20.
21.
MS received 15 Feb. 89.
Accepted for publication 15 Mar. 89.
Resuml
Cette dtudc presente les effets possibles de I’allailcmcnt au scin sur des phdnomcncs lids h (’insertion de
dispositifs intra-utdrins (DIU). l,a recherche portait 6493 femmes inscritcs dans plusicurs centres pour
participcr 3 des cssais cliniqucs pendant unc pdnode de dix ans. On a constatd que I’allailcmcnt au scin a
un effet protccteur contra (’apparition de doulcurs modcrces a fortes au moment de ('insertion ct qu'il
rdduit la ndccssitd de dilator Ic col utdrin cn vuc de facilitcr I’inscrtion. Cct effet de protection contrc les
douleurs prddominait chcz. les femmes qui allaitaicnt ct sc trouvaient encore cn amdnorrhdc de lactation.
Les doulcurs au moment de ('insertion dtaient significativcmcnt moins frdqucntcs au moment de la pose
du DIU chcz ccllcs qui, allaitant ou non, dtaient encore en pdriodc d’amdnorrhdc. que chcz. ccllcs dont Ic
cycle mcnstrucl avait repris. Cct effet pcul ctrc lie 3 unc sderdtion plus abondantc de fl-cndorphinc chcz.
les femmes cn pdriodc d'allaitcmcnt ou d’amdnorrhdc de lactation.
Resumen
Se investigaron cn cstc csludio los posiblcs cfcctos del amamantamiento sobre la inscrcidn de dispositivos
intrauterinos (DIU). El csludio comprcndi6 6493 mujcrcs que partici pa ron en ensayos clinicos cn
divereos centres durante un periodo de dicz anos. Ixjs rcsultados indican que cl amamantamiento cjcrce
un efecto protector contra la aparicidn de dolorcs modcrados a fuertes en el momento de la inscrcidn y
reduce la neccsidad de dilatar el cuello del utcro para facilitar la inscrcidn. Eslc cfccto de protcccidn
contra cl dolor predomind entre las mujeres que amamantaban y se hallaban aun cn amcnorrca de
lactacidn. Las mujcrcs con amcnorrca, tamo las que amamantaban como las que no Io hacian, scnalaron
un nivcl de dolor significativamcnte mcnor cn cl momento de inscrcidn del DIU que aqudllas cuyo ciclo
menstrual se habia rcanudado. Eslc cfccto puede cstar rclacionado con una mayor sccrccidn de
£-cndorfina cn las mujcrcs cn periodo de amamantamiento o de amcnorrca de lactacidn.
Recommendations for
Updating Selected Practices
in Contraceptive Use
Volume II
Produced by the Technical Guidance/Competence Working Group (TG/CWG)
Results of a Technical Meeting and Other Technical Review
Secretariat:
Marcia A. Angle, INTRAH (Program for International Training in Health), School
of Medicine, University of North Carolina at Chapel Hill
Douglas H. Huber, Pathfinder International
James D. Shelton, United States Agency for International Development
Roberto Rivera, Family Health International
Editor:
Monica Gaines, INTRAH and the Department of Epidemiology
University of North Carolina at Chapel Hill
Volunteer Resource Persons:
See Appendix D, Acknowledgements
Esnn
FTTTTTn
September 1997
This publication was written by the Technical Guidance/Competence Working
Group (TG/CWG). It was produced with support from the United States
Agency for International Development (USAID). The views expressed in this
document are the responsibility of the TG/CWG and do not represent the policy
of the USAID.
Any pan of this document may be reproduced or adapted to meet local needs
without prior permission from the TG/CWG Secretariat, provided the TG/CWG
is acknowledged and the material is made available free of charge or at cost.
Any commercial reproduction requires prior permission from the TG/CWG.
The TG/CWG Secretariat would appreciate receiving a copy of any materials in
which text or illustrations from this document are used.
Published 1997
1.8 Progestin-Only. Pills During Breastfeeding*
This section outlines recommendations on the following selected procedural
questions for Progestin-Only Pills (POPs) during breastfeeding:
1.
When can POPs be started for breastfeeding women?
2.
Are there special considerations when a breastfeeding woman is switching
from POPs to other hormonal methods?
If a woman is using POPs during breastfeeding, when should she be advised
to switch to another method?
4.
Can POPs be used when not breastfeeding?
5.
How many POP cycles should be given at the first visit for a new user? At
subsequent visits?
6.
When breastfeeding, is there a best time of day to take POPs?
7.
Are back-up methods advisable in the following situations:
a)
If a breastfeeding client is taking antibiotics, including anti-tuberculosis
medications?
b)
If a breastfeeding client is taking anticonvulsants?
c)
If a breastfeeding client is taking anti-malarial medication?
d)
If it is a breastfeeding client's first cycle of POPs?
e)
If a breastfeeding client has missed pills?
f)
If a breastfeeding client has severe diarrhea and/or vomiting?
* Because the vast majority of POP users are breastfeeding
women, this chapter focuses on POPs during breastfeeding.
However, POPs are an acceptable contraceptive method for
use by women who are not breastfeeding (See Question 4).
September 1997
Progesiin-Only Pills Durin" Breastfeeding
85
Q.1.
When can POPs be started for breastfeeding
women?
Recommendations
Rationales
a) If breastfeeding, POPs may be started
after six weeks postpartum.
a) For breastfeeding women, delaying POP
initiation for six weeks after delivery avoids
exposing the newborn to exogenous steroids
during the time of greatest neuroendocrine
development. In breastfeeding women, the risk
of ovulating in the first six weeks postpartum is
very low. The timing of postpartum initiation
of POPs should be dependent on the woman’s
preference, her previous experience with
breastfeeding and her intentions regarding the
duration of breastfeeding.
POPs are not usually recommended in
the first six weeks postpartum in
breastfeeding women. The timing of
postpartum initiation of POPs should
consider a woman’s breastfeeding
intentions.
’) Howie PW. McNeiily AS. Houston MJ. Cook A. Boyle H.
Fertility after childbirth: postpartum ovulation and menstruation in
bottle and breast feeding mothers. Clinical Endocrinology
1982:17:323-32.
2) Diaz S. Rtxlriguez G. Peralta O, .Miranda P. Casado ME.
Salvatierra AM. et al. Lactational amenorrhea and die recovery of
ovulation and fertility in fully nursing Cllilean women.
Contraception 1988;38(l);53-67.
3) Visness C. Rivera R Progestin-only pill use and pill switching
during breastfeeding. Contraception 1995:51:279-81.
b) A woman who initially chooses to rely
on the Lactational Amenorrhea Method
(LAM) is advised to begin POPs, or
whichever method she chooses to
switch to when one of the following
takes place:
• her menses return, or
• she is no longer fully or nearly fully
breastfeeding, or
• six months postpartum.
Preferably, POP packets are given to
the woman before her intended start
date to ensure that she is able to begin
the method when she needs to.
However, if she prefers, POPs can also
be started when a woman is still relying !
on LAM (providing her with dual
~ i
protection).
I
b) While relying on LAM, a postpartum woman
has at least 98% protection from pregnancy for
six months if she remains amenorrheic and
fully or nearly fully breastfeeds (perfect use
effectiveness rate). Programs sometimes
encourage waiting to initiate POPs until
reliance on LAM ends, because it may be more
programmatically affordable and because using
POPs while breastfeeding may potentially
prolong lactational S'lbfertility.
1) Kennedy K. Rivera R. McNeilly A. Consensus statement on the
use of breastfeeding as a family planning method. Contraception
1989;39(5):477-96
2) Ciiaudhury RR, Choinpootaweep S. Dusitsin N, Friesen H.
Tankeyoon M. The release of prolactin by medroxyprogesterone
acetate in human subjects. British Journal of Pharmacology
1977^9:433-4.
(continued on next page)
September 1997
Progestin-Only Pills During Breastfeeding
87
Q.1. When breastfeeding? (continued)
Rationale
Recommendations
c) After the first six weeks postpartum. > c) Based on current literature, including studies
POPs can be initiated any time you can
with other progestin-only methods, it is
be reasonably sure a woman is not
unlikely that there is a significant effect on the
pregnant (see Appendix A and POP
growth of breastfeeding infants whose mothers
Question 7d).
initiate POPs after the sixth postpartum week.
WHO Task Force on Oral Contraceptives. Effects of hormonal
contraceptives on milk volume and infant growth. Contraception
1984;30(6):505-21.
Shaaban M, Salem H, Abdullah K. Influence of levonorgestrel
contraceptive implants, Norplant, initiated early postpartum upon
lactation and infant growth. Contraception 1985;32(6);623-35.
3)
Pardthaisong T, Yenchit C. Gray R. The long-term growth and
development of children exposed to Depo-Provera during
pregnancy or lactation. Contraception 1992:45:313-24.
4)
McCann MF. Moggia AV. Higgins IE. Potts M. Beeker C. The
effects of a progestin-only oral contraceptive (levonorgestrel 0.03
mg) on breastfeeding. Contraception 1989;40(6):635-48.
1)
2)
I
I
d) Even if POPs are inadvertently initiated d) Epidemiologic studies have found no
during pregnancy, there is no known
risk to the fetus.
significant effect on fetal development or
malformations due to taking hormonal methods
in early pregnancy.
1) Bracken MB. Oral contraception and congenital malformations IQ
offspring, a review and meta-analysis of die prospective studies.
Obstetrics and Gynecology 1990;76:552-7.
2) Wiseman RA. Dodds-Smith IC. Cardiovascular birth defects and
antenatal exposure to female sex hormones: a re-evaluation of
some base ilata. Teratology 1984;30(3):359-70.
3)
Simpson JL, Phillips OP. Spermicides. hormonal contraception
and congeiutal malformations. Advances in Contraception
1990;6:141-67.
e) Non-hormonal methods are preferable
to hormonal methods during
breastfeeding because they have no
effect on breastfeeding and the infant is
not exposed to exogenous steroids.
However, WHO lists POPs as Category
1 after six weeks postpartum, and
women should be given a choice of
contraceptive methods.
88
e) Although the amount of exogenous progestins
in breastmilk is extremely low, it is prudent to
try to minimize infant exposures to any drug.
I) Institute of Reproductive Health Guidelines for breastfeeding in
family planning and child survival programs. Washington, DC:
IRH, 1992.
2) World Health Organization, improving access to quality care in
family planning: medical eligibility criteria for contraceptive use.
Geneva: WHO, 1996.
Progestin-Only Pills During Breastfeeding
September 1997
Q.2.
Are there special considerations when a
breastfeeding woman is switching from POPs to
other hormonal methods?
Recommendations
No. A breastfeeding woman can
switch from POPs to another hormonal
method any time the new method is
appropriate.
No back-up method is necessary when
the new method is initiated if the
woman has been breastfeeding, and
has been taking the POPs fairly
consistently. Estrogen-containing
methods should generally not be used
by breastfeeding women prior to six
1
months postpartum or preferably any |
time during long-term breastfeeding. |
Rationale
As long as the woman is breastfeeding and
taking the POPs fairly consistently, she is fully
protected through the transition to the new
hormonal method.
1) McCann MF. Potter LS. Progestin-only oral contraception: a
comprehensive review. Contraception 1994^0(6).
Clinical trial data indicate that the pregnancy
protection conferred by POP use during
breastfeeding is high, indicating a synergistic
pregnancy prevention effect for breastfeeding
while using POPs. In addition, women in
lactational amenorrhea have additional
protection due to their lowered fecundity.
1)
Dunson T. McLaurin V. Grubb G. Rosman A, A multicenter
clinical trial of a progesuu-only oral contraceptive in lactating
women. Contraception 1993;47:23-35.
2) Kennedy KI. Visness C. Contraceptive efficacy of lactational
amenocrlioea. Lancet 1992;339:227-30.
September 1997
Progestin-Only Pills During Breastfeeding
89
Q.3.
If a woman is using POPs during breastfeeding,
when should she be advised to switch to
another method?
Rationale
Recommendations
a) Women can rely on POPs after the first i a) In general, POPs are highly effective, and safe,
six weeks postpartum, and safely use
during breastfeeding.
them during the entire duration of
1) McCann MF. Potter LS. Progestin-only oral contraception a
comprehensive review Contraception 1994:50(6).
breastfeeding.
2) Dunson TR. McLaurin VL. Grubb G. Rosman A. A mulucenter
clinical trial of a progesun-only oral contraceptive id lactating
women. Contraception 1993;47:23-35.
b) Women can continue using POPs after j b) POPs are an effective contraceptive method,
they stop breastfeeding, provided that , even when not breastfeeding, if used correctly
they have been informed of the
and consistently. However, all women should
advantages and disadvantages of the
i
be informed of the advantages and
method and are willing to use the POPs
disadvantages of POPs in the absence of
correctly and consistently.
i
breastfeeding, especially that POPs need to be
used consistently and correctly to provide
It is not mandatory for a woman to
effective pregnancy protection (e.g., the pill
switch from POPs to another family
should be taken at the same time each day), and
planning (FP) method after she stops
that they often cause irregular menstrual
breastfeeding or at six months
bleeding.
postpartum.
j
1) Visness C. Rivera R. Progestin-ouly pill use and pill switdiing
during breastfeeding. Contraception 1995;51:279-81.
2) McCann MF. Potter LS. Progestin-only oral contraception: a
comprehensive review. Contraception 1994:50(6).
3) World Health Organization Improving access to quality care io
family planning: medical eligibility criteria for contraceptive use.
Geneva: WHO. 1996.
c) Breastfeeding women using POPs
: c) Even low-dose (30 mcg) COCs decrease
should be advised not to switch to
breastmilk production and alter its composition.
combined oral contraceptives (COCs). i
I) WHO Task Force on Oral Contraceptives. Effects of hormonal
contraceptives on nulk volume and infant growth. Contraception
or other methods containing estrogen,
19X4:30:505-21.
until at least six months postpartum.
,
2) McCann MF. Potter LS Progestin-only oral contraception: a
]
d) Breastfeeding women can switch to
non-hormonal methods at any time, as
appropriate.
comprehensive review Contraception 1994:50(6).
i d) If not inserted with 48 hours of delivery,
■
postpartum IUDs are usually not inserted until
I
uterine involution is complete. Progestin|
releasing IUDs are not inserted until six weeks
: postpartum, even if involution is complete
i
before six weeks, to avoid the theoretical risks
of infant steroid exposure. Diaphragms are not
fitted until involution is complete.
1) O'Hanley K. Huber D. Postpartum IUDs: keys for success.
Contraception 1992:45:351-61.
2) Wiley A. The Diaphragm. In: Corson S. Dennan R. Tyrer L,
editors. Fertility Control. Boston: Little. Brown & Company.
1985:223-32.
3) World Health Organization. Improving access to quality care in
family planning: medical eligibility criteria for contraceptive use.
Geneva: WHO. 1996.
90
Progestin-Only Pills During Breastfeeding
September 1997
Q.4.
Can POPs be used when not breastfeeding?
Recommendations
Rationale
Yes, if taken consistently and correctly.
Many women gain experience with and
confidence in POPs during
breastfeeding and should be allowed to
continue POPs after breastfeeding if
I
POPs are the woman’s method of
I
choice.
I
POPs are an effective contraceptive method
even when the woman is not breastfeeding if
taken consistently and correctly.
I
I
POPs are a useful alternative for many women
who want to use oral contraceptives (OCs) but
for whom COCs are not appropriate.
Women should be informed of the advantages
and disadvantages of POPs, especially that
POPs need to be used consistently and
correctly to provide effective pregnancy
protection (e.g.. the pill should be taken at the
same time each day) and that POPs often cause
irregular menstrual bleeding. Unless a woman
is breastfeeding, a back-up method of
contraception should be used if a POP is taken
more than three hours after her regularly
scheduled time (See Question 7e).
1)
September 1997
McCann MF. Palter LS. Progestin-only oral contraception. a
comprehensive review. Contraception 1994^0(6).
Progestin-Only Pills During Breastfeeding
91
Q.5.
How many POP cycles should be given at the
first visit for a new user? At subsequent visits?
Recommendations
Rationale
a)
New user?
a) The woman’s convenience is important. To
avoid running out of pills, the woman should
Postpartum women who plan to use
have ready access to more POP cycles.
LAM can be given their pill cycles
i
Ideally,
she should be able to obtain plenty of
immediately postpartum, with
POP cycles at her visit.
instructions to begin taking them (see .
Question 1) when any of the LAM
;
While some providers suspect that clients who
criteria no longer apply. Women who
receive multiple pill cycles may “share” these
plan to rely on LAM for six months can
with friends, such "sharing" is likely to be as
be given at least a six month supply (to i
safe and effective as over-the-counter
begin when the LAM criteria no longer '
distribution systems.
apply), so they will have contraceptive ■
protection for at least one year.
Up to 13 cycles (a full year’s supply)
■
can be given, although only three or
four may be programmatically feasible, i
The greatest need is to guarantee
continuous, ready access.
b) Subsequent visits?
b) The extremely low dose of progestins in POPs
make them a very safe method of
There is no compelling medical reason
contraception. The greatest health risk from
for a routine return visit concerning
POPs is pregnancy due to method failure,
POP use, but clients should be
which is preventable by assuring adequate
encouraged to return at any time with
POP supply and correct, consistent method
concerns, problems or questions.
use.
For first-time users of POPs, programs
1) McCann MF. Potter LS Progestin-oiily oral contraception: a
may encourage a three-month follow-up
comprehensive review. Contraception 1994-^0(6).
2) Harlap S. Kost K. Forrest JD Preventing pregnancy, protecting
visit for counseling to assess whether
health: a new look at birth control choices in die United States.
the client is satisfied with the method
Waxlungton. D C Hie Ala.. Guttmacher Institute, 1991.
and is correctly using the method, to
reinforce instructions, and to help
clients with the management of side
effects.
92
Progestin-Only Pills During Breastfeeding
September 1997
Q.6.
When breastfeeding, is there a best time of day
to take POPs?
Recommendations
Rationale
a) POPs may be taken at any time of the
day for effective use during
breastfeeding. The client may wish to
select a certain time to help her
remember to take a pill every day; it
may help to link this time to a daily
event.
a) Breastfeeding women have additional
protection due to their lower fecundity.
Clinical trial data indicate that the pregnancy
protection conferred by POP use during
breastfeeding is extremely high. The
■
synergistic pregnancy protection by POP use in
combination with breastfeeding should
sufficiently eliminate a client’s risk of
conception, even if she takes POPs at different
times of the day.
1)
2)
Dunson T. McLaunn V. Grubb G. Rosman A. A multicenter
clinical trial of a progestin-only oral contraceptive in lactating
women. Contraception 1993:47:23-35.
Wright SW. Fodierby K. Fairweather F. Effect of daily small
doses of norgestrel on ovarian function Journal of Obstetrics and
Gynecology'of die British Commonwealth 1970;77:65-8.
I
b) However, if a woman continues taking ; b) The most immediate contraceptive effect of
POPs is the alteration of cervical mucus. The
POPs after breastfeeding cessation, then :
POP’s effect on cervical mucus peaks
it is important to take the POP at the
i
same time every day, preferably late
approximately four to five hours after ingestion
of the pill, and is essentially gone by 24 hours
afternoon or four to five hours before
the usual time of sexual activity, so that
after taking one POP.
the pill’s effect on the cervical mucus is
1) McCann MF, Poller LS. Progestin-only oral contraception: a
comprehensive review. Contraception 1994*^0(6).
at its maximum by the time sexual
2) Ciirctieu FC, Sureau C. Neau C. Experimental study of cervical
activity occurs.
blockage induced by continuous low-dose oral progestogens.
Contraception 1980:22.445-56
September 1997
Progestin-Only Pills During Breastfeeding
93
Q.7.
Are back-up methods advisable in the following
situations?
Recommendations
Rationale
a) If a breastfeeding client is taking
antibiotics, including anti
tuberculosis medications?
Back-up methods are not usually
required, unless the woman is taking
rifampin/rifampicin.
With the exception of rifampin/
rifampicin, antibiotics are unlikely to
j
significantly reduce the effectiveness of ,
POPs in breastfeeding women.
If the breastfeeding woman is taking
'
rifampin/rifampicin, she should know
that rifampin/rifampicin:
• passes through breastmilk (with
potential infant side effects),
• may increase breakthrough bleeding,
and
• lowers progestin levels, possibly
significantly reducing the
effectiveness of POPs.
a) Broad-spectrum antibiotics such as ampicillin,
erythromycin and tetracycline have not been
shown to decrease effectiveness of POPs in
careful clinical studies.
Rifampin/rifampicin, which is used primarily
for treating tuberculosis, induces hepatic
enzymes and increases the liver metabolism of
progestins, thus decreasing the effectiveness of
POPs. The enzyme-inducing effects of
rifampin/rifampicin last about four weeks after
short-term use and eight weeks after long-term
use.
Griseofulvin, an anti-fungal antibiotic and
another hepatic enzyme inducer, has not been
proven to reduce POP effectiveness in humans.
but may increase menstrual irregularities.
Rifampin/rifampicin is passed in breastmilk
(milkiplasma ratio of 0.2 to 0.6). Griseofulvin
may also be passed in breastmilk. Infant
exposure to rifampin/rifampicin or griseofulvin
is appropriate only when the maternal benefits
outweigh the potential risks to the infant.
Back DJ. Onne ML Drug interactions. In: Goldaeher JW,
Fotherby K (editors.). Pharmacology of die Contraceptive Steroids.
New York. Raven Press. 1994:407-25.
Fodterby K. Interactions with oral contraceptives. American
Journal of Obstetrics and Gynecology 1990;163:2153-9.
3)
Drug Facts and Comparisons. St. Louis: Facts and Comparisons.
June 1996. p. 35 X and October 1990. p.387a
4)
World Health Organization. Improving access to quality care in
family planning: medical eligibility criteria for contraceptive use.
1)
2)
5)
Baciewicz AM. Self TH. Bekemeer WB. Update on rifampin drug
interactions. Archives of Interna] Medicine 1987;147(3):565-8.
(continued on next page)
94
Progestin-Only Pills During Breastfeeding
September 1997
Q.7. Back-up methods (continued)
Recommendations
Rationale
b) If a breastfeeding client is taking
anticonvulsants?
Yes, usually. The common
anticonvulsants, hydantoins (e.g.,
phenytoin), barbiturates (e.g.,
phenobarbital, primidone), and
probably carbamazepine significantly
decrease the effectiveness of oral
contraceptives. POPs are not
recommended if using these enzyme
inducing anticonvulsants.
b) The hepatic enzyme-inducing effects of most
anticonvulsants probably decrease pregnancy
protection and increase rates of irregular
bleeding among some POP users. It should be
noted however that POPs may decrease the
probability of seizures among users of
anticonvulsants.
Additionally, because anticonvulsants
are excreted in breastmilk, and because
there is a potential for serious adverse
reactions in nursing infants, women
taking hydantoins, barbiturates, or
carbamazepine for chronic seizure
control may be advised to explore safe
alternatives to breastfeeding.
Injectable contraceptives, such as Depo
Provera®, will be effective despite
anticonvulsant use, but infant exposure
to the anticonvulsants will continue.
Non-hormonal methods will continue to
be effective despite anticonvulsant use.
c) If a breastfeeding client is taking
anti-malarial medication?
No back-up is needed.
There is no evidence that anti-malarial
medications reduce the effectiveness of
OCs.
Chloroquine and related anti-malarials
are excreted in breastmilk.
Because of the dangers of fetal exposure to
most anticonvulsants, full protection against
pregnancy is essential. Although increased
doses of POPs might be effective, they might
also further increase bleeding irregularities.
1) Mattson RH. Rebar RW. Contraceptive methcxls for women with
neurologic disorders. American Journal of Obstetrics and
Gynecology 1993;168:2027-32
If a woman ingests hydantoins, barbiturates, or
carbamazepine, her breastmilk will contain
significant quantities of these substances. In
areas where safe alternatives to breastfeeding
exist, and where maternal seizures cannot
otherwise be controlled, women on long-term
anti-seizure medications may be advised to
consider safe alternatives to breastfeeding, to
avoid chronic infant drug exposure.
1) Drug Facts and Comparisons. Si. Louis: Facts and Comparisons,
July 1996. pp. 282-4
2) World Health Organization. Improving access to quality care in
family planning: medical eligibility criteria for contraceptive use.
Geneva. WHO. 1996
3) Anderson GD. Graves NM Drug interactions with antiepileptic
agents. CN’S Drugs 1994;2M):268-79.
c) Chloroquine, primaquine and tetracycline have
not shown any effect on OC hormonal levels,
and are not known to reduce the effectiveness
of POPs.
A nursing infant may consume about half of a
mother’s 3(X) mg chloroquine dose over 24
hours; the maternal milk: blood ratio may be
about 0.36. Children are especially sensitive to
chloroquine and primaquine.
(continued on next page)
September 1997
Progestin-Only Pills During Breastfeeding
95
Q.7. Back-up methods (continued)
Rationale
Recommendations
c) Weighing the nutritional value of the milk to
the child against the effects of the chloroquine,
clients are usually not advised to stop
breastfeeding while on anti-malarial treatment,
unless safe alternatives to breastmilk are
available.
1) Drug Facts and Companions St. Louis: Facts and Comparisons
June 1996. pp. 358 and 387a.
d) If it is a breastfeeding client’s_first
cycle of POPs?
No back-up is needed.
However, if a breastfeeding woman has
resumed menstruating and is beginning
the pills later than the first seven days
of her cycle, some programs
recommend that she use a back-up
method for seven days after beginning
POPs.
d) The cervical mucus thickens enough to prevent
sperm penetration within 24 hours. Also, the .
synergistic protection against pregnancy
conferred by concurrent POP use and
breastfeeding should sufficiently eliminate a
client’s risk of conception. Thus, a back-up
method for a full seven days may not be
necessary.
e) If a breastfeeding client has missed
pills?
e) After missing one pill, breastfeeding women
previously taking POPs are estimated to be
sufficiently subfertile that the probability of the
woman becoming pregnant is extremely low.
If the breastfeeding woman is still
amenorrheic, missed pills are of
minimal consequence.
For a breastfeeding woman who has
already returned to menses, if two or
more pills are missed, the woman
should:
• resume taking a pill as soon as she
remembers,
• take the next pill at the regular time
that day (for added protection), and
• use a back-up method or abstinence
for 48 hours (some programs
recommend use of a back-up method
for up to seven days).
1) Clirelien FC. Surcau C. Ncau C. Experimental study of cervical
blockage induced by continuous low-dose oral progestogen*.
Contraception 1980;22:445-56,
2) Kesseru-Koos E. Influence of various hormonal contraceptives on
spenn migration in vivo. Fertility and Sterility 1971;22:584-603.
3) Moghixsi KS. Syner FN. McBride LC. Contraceptive mechanism
of nucrcxlose norethindrone. Obstetrics and Gynecology
1973;41:585-94.
The most immediate effect of POPs is on
cervical mucus, each tablet offering protection
for approximately 24 hours. Clinical trial data
indicate that the pregnancy protection
conferred by POP use during breastfeeding is
high, indicating a synergistic pregnancy
prevention effect for breastfeeding while using
POPs. In addition, women in lactational
amenorrhea have additional protection due to
their lowered fecundity.
1) Kcsseru-Kocv: E Influence of various hormonal contracepuvex on
sperm migration in vivo. Fertility and Sterility 1971;22:584-603.
2) Dunson f. McLaurin V. Grubb G. Rosman A. A raulucenter
clinical trial of a progesun-ouly oral contraceptive in lactating
women. Contraception 1993;47:23-35.
3) Kennedy KI. Visness C. Contraceptive efficacy of lactational
amenocrhocx Lancet 1992:339:227-30.
(continued on next page)
96
Progestin-Only Pills During Breastfeeding
September 1997
Q.7. Back-up methods (continued)
f)
Recommendations
Rationale
If a breastfeeding client has severe
diarrhea and/or vomiting?
i f)The synergistic protection conferred by POP
use and breastfeeding should sufficiently
eliminate a client’s risk of conception, because
women in lactational amenorrhea have
additional protection due to their lowered
fecundity.
If a woman is breastfeeding and
amenorrheic, no back-up method is
needed since the synergistic effect of
both breastfeeding and POP use should
provide sufficient pregnancy protection.
If a breastfeeding woman has resumed
menstruating, some programs
recommend use of a back-up method
for 48 hours or for 7 days after the
severe vomiting or diarrhea stops.
September 1997
1) Dunson T. McLaurin V. Grubb G. Rosman A. A multicenter
clinical trial of a progestin-only oral contraceptive in laccaung
women. Contraception 1993;47:23-35
2) Orme M. Back DJ. Breckenridge AM. Clinical pliarmacokineucs
of oral contraceptive steroids. Clinical Pliannacolanetjcs 1983,
8:95-136.
3) Kennedy KI. Vixne*s C. Contraceptive efficacy of lactational
ameiiorrhoea Lancet 1992.339.227-30,
Progestin-Only Pills During Breastfeeding
97
Classification of Selected Procedures for
Progestin-only Pills (POPs) during Breastfeeding
Procedure
Class
Rationale
Pelvic examination
(speculum and bimanual)
C
• A pelvic exam is not necessary to ensure safe use of
POPs as a contraceptive method'.
• In some cases, a pelvic exam may help evaluate the
question of pregnancy if a menstrual history suggests
the possibility beyond six weeks duration. In this case
it is Class A.
• Conditions which would restrict use of POPs should be
identified by the client’s history before method
initiation.
Blood pressure
C
Current evidence does not demonstrate any notable effect
of POPs on blood pressure2 ’.
C
1 POPs do not cause breast cancer1"'. Lumps that are
suspicious as cancer should be evaluated. While any
hormonal treatment may in theory cause such lumps to
grow, pregnancy causes much higher hormonal levels;
therefore, potential malignancies of the breast should not
be a reason to delay a woman’s access to the use of this
contraceptive method.
STD screening by lab tests
(for asymptomatic persons)
c
The presence of an STD will not affect the safe use of
POPs. Clients at risk of STDs (by personal history or
socio-demographic risk factors) should be offered STD
screening where possible.
Cervical cancer screening
c
POPs have no known relation to risk of cervical cancer6.
Routine, mandatory lab
tests (e.g„ cholesterol,
glucose, liver function tests)
D
The effect of POPs on cholesterol, blood glucose and
normal liver function are slight, and of no demonstrated
clinical significance6’8.
Proper infection prevention
procedures
C
Proper infection prevention procedures are not applicable
to POP use.
Breast examination
!
(continued on next page)
KEY:
Class A
Class B
Class C
Class D
98
= essential and mandatory or otherwise iinporuint in all circumstances, for sale and effective use of the
contraceptive method
= medica'.ly/epidemiologically rational in some Circumstances to optimize the safe and effective use of
the contraceptive method, but may not be appropriate for all clients in all settings
= may be appropriate for good preventive health care, but not materially related to safe and effective use
of the contraceptive method
= not materially related to either gtxxl routine preventive health care or to the safe and effective use of
the contraceptive method
Progestin-Only Pills During Breastfeeding
September 1997
Classification of Selected Procedures for
Progestin-only Pills (POPs) during Breastfeeding
(continued)
Procedure
Specific counseling points
for POP use:
• efficacy
• common side effects,
including alterations in
bleeding patterns (e.g.
frequent or irregular
bleeding, extended
amenorrhea)
• correct use of method
(including instructions
for missed pills)
• signs and symptoms for
which to see a health
provider
• STD protection (when/as
appropriate)
| Class
I
Rationale
Accurate client education is essential for maximum
quality of FP services.
Appropriate counseling about common contraceptive
side effects at the time of method selection can lead to
improved client satisfaction and contraceptive
continuation.
Irregular or absent menstrual bleeding is the single
most common side effect of POPs, and the chief
complaint leading to discontinuation9.
POPs are highly effective if taken correctly and
consistently. However. POPs are less effective than
COCs after weaning.
The woman should be encouraged to return if she has
any problems or at any time she has questions or
concerns.
(continued on next page)
KEY:
Class A
Class B
Class C
Class D
= essential and mandator,’ or otherwise important in all circumstances, lor safe and effective use of the
contraceptive method
= medically/epidemiologically rational in some circumstances to optimize the safe and effective use of
the contraceptive method, but may not be appropriate for all clients in all settings
= may be appropriate for good preventive health care, but not materially related to safe and effective use
of die contraceptive method
= not materially related to either good routine preventive health care or to the safe and effective use of
the contraceptive method
September 1997
Progestin-Only Pills During Breastfeeding
99
Classification of Selected Procedures for
Progestin-only Pills (POPs) during Breastfeeding
(continued)
Citations for Procedures Table:
1)
2
3)
4)
5)
6)
7)
8)
9)
Huber DH, Huber SC. Screening oral contraceptive candidates and inconsequential pelvic examinations.
Studies in Family Planning 1975;6(2):49-51.
Ball MJ, Ashwell E, Gillmer MDG. Progestagen-only oral contraceptives: comparison of the metabolic effects
of levonorgestrel and norethisterone. Contraception 1991;44(3):223-33.
Wilson ESB, Cniickshank J, McMaster M, Weir RJ. A prospective controlled study of the effect on blood
pressure of contraceptive preparations containing different types and dosages of progestogen. British Journal of
Obstetrics and Gynaecology 1984;91:1254-60.
Stanford JL, Thomas DB. Exogenous progestins and breast cancer. Epidemiologic Reviews 1993;15(l):98-107.
UK National Case-Control Study Group. Oral contraceptive use and breast cancer risk in young women.
Lancet 1989:1:973-82.
World Health Organization. Improving access to quality care in family planning: medical eligibility criteria for
contraceptive use. Geneva: WHO. 1996.
Miale JB, Kent JW. The effects of oral contraceptives on die results of laboratory tests. American Journal of
Obstetrics and Gynecology 1974;120(2):264-72.
Korba VD, Paulson SR. Five years of fertility control with inicrodose norgestrel: an updated clinical review.
Journal of Reproductive Medicine 1974;13(2):71-5
Belsey EM. WHO Task Force on Long-acting Systemic Agents for Fertility Regulation. The association
between vaginal bleeding patterns and reasons for discontinuation of contraceptive use. Contraception
1988;38(2):207-25.
For further information see McCann MF, Potter LS. Progestin-only oral contraception: a comprehensive
review. Contraception 1994;50(6).
100
Progestin-Only Pills During Breastfeeding
September 1997
ARTICLES
Effect of breastfeeding on infant and child mortality due to
infectious diseases in less developed countries: a pooled analysis
WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality"
Summary
Background The debate on breastfeeding in areas of high
HIV prevalence has led to the development of simulation
models that attempt to assess the risks and benefits
associated with breastfeeding. An essential element of
these simulations is the extent to which breastfeeding
protects against infant and child mortality; however, few
studies are available on this topic. We did a pooled analysis
of studies that assessed the effect of not breastfeeding on
the risk of death due to infectious diseases.
Methods Studies were identified through consultations with
experts in international health, and from a MEDLINE search
for 1980-98. Using meta-analytical techniques, we
assessed the protective effect of breastfeeding according to
the age and sex of the infant, the cause of death, and the
educational status of the mother.
Findings We identified eight studies, data from six of which
were available (from Brazil, The Gambia, Ghana. Pakistan.
the Philippines, and Senegal). These studies provided
information on 1223 deaths of children under two years of
age. In the African studies, virtually all babies were breastfed
well into the second year of life, making it impossible to
include them in the analyses of infant mortality. On the basis
of the other three studies, protection provided by breastmilk
declined steadily with age during infancy (pooled odds ratios:
5-8 (95% Cl 3-4-9-8J for infants <2 months of age, 41
(2-7-6-4J for 2-3-month-olds, 2 6 [1-6-3-9] for 4-5-montholds, 1-8 [1-2-2-8] for 6-8-month-olds. and 1-4 [0-8-2-6] for
9-11-month-olds). In the first 6 months of life, protection
against diarrhoea was substantially greater (odds ratio 6-1
[4-1—9-0]) than against deaths due to acute respiratory
infections (2-4 (1-6—3-5]). However, for infants aged 6-11
months, similar levels of protection were observed (1-9
(1-2-31) and 2-5 [1-4-4-6], respectively). For second-year
deaths, the pooled odds ratios from five studies ranged
between 16 and 2-1. Protection was highest when maternal
education was low.
Interpretation These results may help shape policy
decisions about feeding choices in the face of the HIV
epidemic. Of particular relevance is the need to account for
declining levels of protection with age in infancy, the
continued protection afforded during the second year of life.
and the question of the safety of breastmilk substitutes in
families of low socioeconomic status.
Lancet 2000; 355: 451-55
■Members listed at end of paper
Correspondence to: Prof Cesar G Victora. Faculdade de Medicina.
Umversidade Federal de Pelotas. CP 464. 96001-970 Pelotas. RS.
Brazil
(email: cvictora@zaz.com.br)
THE LANCET • Vol 355 • February 5, 2000
Introduction
The recognition that HIV is transmitted through
breastmilk raised the important question of whether
strategies that promote breastfeeding in areas of high HIV
prevalence should be changed. 1 This policy decision is
particularly difficult, given that breastfeeding provides
important protection against infectious diseases,“ which
account for over two-thirds of the 12 million annual
deaths in children younger than 5 years in less developed
countries.’
Theoretical models have been developed to assess the
advantages and disadvantages of breastfeeding for HIV
positive women, as well as for women of unknown HIV
status who live in areas of high HIV prevalence.'-1’ These
models have taken into account the risk of transmission
through breastfeeding with regard to the age of ±e infant,
the protection afforded by breastmilk against infectious
disease mortality, the underlying HIV prevalence, and the
rate of infant and child mortality. However, according to a
recent WHO/UNAIDS review," an important limitation
of these models was the poor quantification of the relative
risks for mortality associated with lack of breastfeeding.
Such models used relative risk estimates ranging from 1 -3
to 7-9, and no model allowed for variable levels of
protection within the first year of life. We did a
comprehensive pooled analysis of existing studies of the
effect of not breastfeeding on risk of infant and child
mortality due to infectious diseases.
Methods
Through consultations with experts in international health, the
Division of Child and Adolescent Health (WHO, Geneva,
Switzerland) obtained a list of published and unpublished
datasets that might provide information on the risk of infant and
childhood mortality according to feeding practices. The search
was limited to studies carried out in the 1980s and 1990s, to
those that provided data on deaths in the first year of life (and if
possible on second-year deaths as well), and to those in which the
cause of death was ascertained. This list was complemented by a
MEDLINE search from 1980 to December, 1998, with the
keywords “breastfeeding” and “mortality”. These sources
revealed a total of 13 research groups who might have had
datasets containing the information required.
The datasets were revised to make them compatible in terms of
the names of variables, and of coding. All deaths that occurred in
the first week of life were excluded, since breastfeeding is unlikely
to have had a marked impact on these deaths (which were mainly
due to perinatal causes and congenital malformations). Also,
deaths not attributed to infectious diseases and deaths for which
the cause was not known, were excluded from most analyses. For
the Ghana study, two independent raters defined the cause of
death, and there was no attempt to reach a consensus. The k
statistic for inter-rater agreement was 0-65 (p<0 001). We
arbitrarily chose to use the opinion of the first rater, but since the
study contributed only 33 deaths to the analysis, we assumed that
the pooled estimate would not be markedly affected. For the
Philippines study, deaths classified as due to “diarrhoea and
acute respiratory infections” were included in both categories.
Information on AIDS-related deaths was not specifically
collected in the studies—all but one were carried out in the
451
ARTICLES
Number of Children
Sue
Design
Setting
Year
Brasi”
1070
Peiotas
Case<ontroi
Urban
1984
The Gambia"
431
Upper River Division
Case-control
Mostly rural
1988-90
Ghana*
1099
Upper East Region
Conort
Rural
1990-91
Pakistan"
2166
Lanore
Conor.
Rural and urban slums
1984-87
Philippines11
9682
Cebu
Cohort
Urban
1988-91
Senegal"
3534
Niakhar
Cohort
Rural
1983-86
Table 1: Characteristics of studies Included In pooled analysis
Ghana
Pakistan
The Gambia
Brazil
Philippines
_ Senegal________
41
33
165
357
202
Total deaths
_ 425__________
Deaths by age (months)’
1 (3-0%»
17 (41-5%)
61 (301%i
51(14.3%)
19 (11-5%)
0-1
30 (9-3*1
5(15-2*1
8 (19-5%>
16 (9-7%)
113(31-7%)
16 (7-9*.1
2-3
29(8-9%>
10(30-3%,
8 (19-5*1
29 (14-3%i
22 (13 3%)
84 (23-5*i
34 (10-5%.
4-5
7 (21-2%.
3 (7-3%,
80122-4*.26 (12-8%i
48 (29-1%)
41112-6%)
6-8
0
27 (19 4%)
3(9-l%i
28 (13-81
29 (8-1%.
43 (13-2*‘.
9-11
5(12-2*1
42 <20-7%:
7 (21-2%.
33 (20-0%>
148 (45 5%)
12-23
Deaths by cause
30 (69-8*!
10 (30-3%)
39 115-1*)
127 (35-6%>
131 (49-3*>
140(13-5%)
Acute respiratory infections
9 (10-9%)
5(15-2*1
62 (24-0%)
27 (10-2%i
288 (27 8%)
170 i47-6%i
Diarrhoea
44 (17-1%)
Diarrhoea and acute respiratory infections
13(39-4%.
2(4 7%i
20 (7-8*1
60(16-8%.
45(16-9%)
Other infections
5(15 2%.
2 <4-7*1
63 (23 7%;
93 (36 1%i
607 (58-7%)f
Other causes
Breastfeeding prevalence by age (months) among controls
100%
84*
100%
98*
80%
0-1
100%
60*
100%
100%
95%
77%
100*
2-3
39*
100%
100%
92*
72%
4-5
100%
64%
28*
100%
100%
88%
6-8
100%
96%
100%
80%
57%
9-11
26%
99%
75%
99%
46*
12-23
26*
91%
•Excluding deaths occumng in tne first week of life, and deaths from noiwnfectious causes, fincluding deaths due to infections otheir than acute respiratory infections or diarrhoea.
Table 2: Distribution of deaths by age and cause, and breastfeeding prevalence among controls
1980s—but HIV is unlikely to have caused more than a few
deaths at any of the sites.
Since maternal education varied markedly across countries,
this variable was classified separately for each study, according to
approximate terciles. In Brazil, the classifications used were 0-2
years, 3-5 years, and 6-17 years of education; in Pakistan,
women were classed as either illiterate, able to read, or able to
read and write; and in the Philippines, women were grouped into
those having 0-5 years, 6-10 years, or 11-16 years of education.
For the African studies, stratification by maternal education was
not possible.
To allow pooling of the data from case-control studies and
cohort studies, results from the latter were stratified by age
groups (0-1, 2-3, 4-5, 6-8, 9-11, 12-15, 16-19, and 20-23
months); the numbers of deaths and of survivors (ie, those who
remained in the study until the end of each age stratum) were
used in the analyses. An 11-month-old child from a cohort study,
therefore, provided information on breastfeeding frequency for
the first four age strata. Stratification was not required for the
preparation of the data files for case-control studies, since each
control child only provided information on breastfeeding at a
single age. This pooling of case-control and cohort data, which
has been used in other collaborative reanalyzes,” is analogous to
treating cohort studies as nested case-control studies.
In the analysis of cases, we assessed breastfeeding status before
the onset of the fatal disease to avoid reverse causality—ie,
feeding changes as a result of illness. When this information was
not available, we used breastfeeding status 7 days before death.
Ag^group Brazil
(months) Cases
For controls, we assessed breastfeeding in the middle of each of
the age intervals described. In the Brazilian study, most deaths
took place in a hospital, and although hospital-acquired
infections may have contributed to mortality,” the underlying
cause of death was diarrhoea severe enough to warrant hospital
admission. In Brazilian children who had one or more admissions
due to diarrhoea in the 2 months before their death,
breastfeeding status was classified before the first of these
admissions. For the analyses of infant mortality stratified by sex,
maternal education, and cause, the same strategy was used, but
due to smaller sample sizes, only two age groups were used (0-5
and 6-11 months).
After preparing the data, we did the analyses using standard
meta-analytic procedures—ie, Mantel-Haenszel pooled estimates
and Cornfield confidence intervals.’’ We added 0*5 to all table
cells containing a value of zero, so that calculation of the odds
ratio was possible. The analyses were done with Stata 5.0 (Stata
Corporation, College Station, TX, USA) and the “metan”
procedure.
Results
We identified eight studies that met our criteria; data were
obtainable from six of them.1’'-' Of those we could not
obtain, one dataset had been destroyed, and the author of
the other failed to respond to repeated attempts to contact
him. The studies were all done between 1983 and 1991,
and covered both urban and rural areas (table 1). Table 2
Pakistan
Philippines
Odds ratio
Weight Cases Odds ratio
Weight Cases Odds ratio
Weight
(95* Cl)
(%)•
(95* Cl)
(*)•
(95% Cl)
(%)•
O-l
51
7-2(3-3-15-9) 43-7
21-3 (7-9-57-7) 6-2
18
2-5 (10-6-3)
501
113
2-3
3-8J2-3-6-1)
83-3
8
11-8 (3-1-45-4)
15
51 U-9-13-7) 142
4-5
84
2-5 (1-4-4-5)
8
67-2
1-6 (0-0-10-3) 5-0
22
2-6 (1-1-5 8)
27-8
80
2-4 (1-2-47)
6-8
37-0
3-5 (0-0-27-1)
47
1-5 (08-2-6)
61-4
9-11
29
1-9 (0-7—5-3)
31-9
0
0
27
1-2 (0-6-2-5)
68-1
AH
357
3-2 (2-84-2)
36
7-9 (3-8-16-3)
129
1-9 (1-3-2-7)
Deaths due to nooinfectxxis causes, and deaths in the first week of life were excluded.
•Weight of data in pooled estimate, fincluding deaths due to non-mfectious causes. tHeterogeneity test p=0-009.
Table 3: Infant mortality associated with not breastfeeding, by country and age-group
Pooled
Cases Odds ratio
(95% Cl)
86
5-8$ (3-4-9-S)
136
4-1 (2-7-64)
114
2-5 (1-6-3-9)
130
1-8 (1-2-2-8)
56
1-4 (08-2-6)
522
Pooled!
Cases Odds ratio
(95% Cl)
4-2(2-8-6-3)
125
144
3-6(2-4-5-5)
2-5 (1-6-40)
119
1-7 (1-1-2-5)
149
1-4 (0-8-24)
68
605
THE LANCET • Vol 355 • February 5,2000
ARTICLES
Age (months) and sex
Brazil
Cases
Pakistan
Cases Odds ratio
(95% Ch
Philippines
Cases Odos ratio
(95% Ch
Odds ratio
Weight
weight
(95% Cl)
(%)•
<%)•
0-5
Boys
140
10 7 |3 9-29 4)
2-4 (12-1-9)
3-5 (2-3-5-4J
70-2
16
32
Girls
108
17
6-8(2 2-20-7) 4-9
25
73-8
3-8(2 4-6-2)
4 5t2-l-9-7i
6-11
Boys
58-7
26
63
1-5(0-8-31)
5-7* 2 0i0-9-4-3.
Giris
24
46
3
11-9+ 3-9 (1-5-10-3) 360
22
Deaths due to non-mfectious causes, and deaths in the first week of life were excluded.
•Weight of data tn pooled estimate. fConfidence interval undefined owing to small sample size
Table 4: Infant mortality associated with not breastfeeding, by country, age-group. and sex
Weight
(%)•
Pooled
Cases Odds ratio
(95% Cl)
26-4
213
188
150
3 5 (2-4-5-01
4 1(2-8-61)
400
61-9
91
73
18(1-1-3-0)
2-9(1-6-52)
Senegal
Gambia
Philippines
Pakistan
Pooiedj
Ghana
Weight Cases Odds ratio Weight
Odds ratio Weight Cases OR
(months) Cases Odds ratio Weight Cases Odds ratio Weight Cases Odds ratio
(95% Ch
(95% Ch (%)■
(95% CD
___ (95% Cl) (%)•
___ (95% Cl> (%)•
<95% CD
540
12-15 15
10-3
18
51 30
15-4 91
07
3
1-6
20-1 4
23 8{§ 0-3
lOO-SOi
(0-6-4-01
<01-40i
11-0-9-3.
(0-8-3-2)
4-4
16-19 16
1-0
362
35 2-9
40-6 66
07
4-5§
11
18-0 2
110*
2
d-3-6-4)
(0-3-6-7)
10-3-3-5)
(1-1-40)
(0-138-7)
20-23
7
3-3
62
0
1 7*§
86-0 74
08
98 1
09
0
3-9*8
(10-2-7)
<01-4-5)
il 1-2-9)
All
38
0-9
7
2-0
33
148 20
231
5
(0-6-2-9.
(03-2-6)
(1-2-53-21
(0-4-1151
H 4-3-1)
•Weight of data in pooled estimate. fAll cases were breastfed. ^Contains one or more zero cells. Odds ratios and 95% Ci calculated by adding 0-5 to all table cells. §95% Cl
undefined owing to small sample size.
Table 5; Mortality at 12-23 months of age associated with not breastfeeding, by country and age-group
Age-
shows the number of deaths by age, excluding deaths in
the first week of life and non-infectious deaths. All studies
except those from Ghana and Pakistan included more
than 160 deaths. Taking into account the different widths
of the age strata, the highest mortality rates were observed
in infants younger than 2 months of age m The Gambia
and Pakistan, in those 2-0-3-9 months old in Brazil, in
those 4-0-5'9 months old in Ghana and Senegal, and in
those 60-8-9 months old in the Philippines
Acute respiratory infections or diarrhoea were the
leading cause of death from infectious diseases in three
studies. However, some differences in design and the
classification of the cause of death affected
the
comparability of the studies. The Brazilian study was
restricted to deaths due to infections. The study from the
Philippines had a category of “diarrhoea and acute
respiratory infections” for deaths of children with
symptoms of both disorders, whereas in the other studies
the referees assigned such deaths to the single cause that
initiated the chain of events leading to death. Deaths due
to infections other than acute respiratory infections or
diarrhoea in Senegal were grouped with non-infectious
deaths.
Breastfeeding patterns varied markedly between studies
(table 2). In the three African sites, virtually every child
was breastfed during the whole of their first year of life.
Therefore, these studies could not be included in the
analyses of lack of breastfeeding and infant mortality.
In the Gambian study, breastfeeding status was ignored
because mothers were unable to provide information
on breastfeeding duration in 14 cases (6-9%) and 19
controls (8-3%); in the study in the Philippines, it was
ignored in three cases (1-8%) and 113 controls (1-2%).
Table 3 shows the odds ratios for infant mortality
associated with lack of breastfeeding in each group in each
study. Pooled data are also shown. The protection against
mortality provided by breastmilk tended to decline with
age in all three studies. When deaths due to noninfectious causes were included in the pooled estimate,
odds ratios for the youngest age ranges were affected the
most (table 3).
THE LANCET • Vol 355 • February 5, 2000
The protecuon provided by breastfeeding against all
infant deaths (ie, in any age group) was greatest in
Pakistan, intermediate in Brazil, and lowest in ±e
Philippines
(table
3).
Protection
provided
by
breastfeeding was slightly greater for girls than for boys,
but the fact that the confidence intervals overlap should
be noted (table 4). In the first 6 months of life, protection
against diarrhoea was substantially greater (odds ratio 6-1
(95% CI 4-1—9-0]) than against acute respiratory
infections (2-4 [1-6-3-5]), but all 6-11 months of age,
similar levels of protection were observed against both
causes of death (T9 1-2—3-1] vs 2-5 (1-4—4-6),
respectively). Whereas protection against diarrhoea
declined markedly with age, protection against acute
respiratory infections was constant.
Lack of breastfeeding was associated with a
substantially larger risk in 0-5-month-old children of
poorly educated mothers (1st tercile; odds ratio 7-6
[4-7—12-3]) than in those with more educated mothers
(2-7 (1-8-4-1) and 3-5 [2-0-6-1] for 2nd and 3rd terciles,
respectively). This effect was also seen in 6-11-month-old
children (5-1 [2-8-9-3J, 2-0 [11-3-8], and 1-1 [0-5-2-6]
for 1st, 2nd, and 3rd terciles, respectively).
All studies except the Brazilian one provided data on
deaths in the second year of life. Except for two studies
(Ghana and Senegal), there was no clear pattern of
decreasing risk with age. There was no significant
heterogeneity. The overall odds ratio for the second year
was greatest in Ghana (7-9, based on only seven deaths).
The Senegal and Pakistan studies showed odds ratios of
2-0, but the latter was not significant. No significant effect
was observed in the Gambian and Filipino studies
(table 5).
Discussion
This paper presents the findings from a comprehensive
attempt to obtain available data on all studies on cause
specific mortality according to breastfeeding. A large
number of investigators were contacted by WHO, and
datasets from those willing and able to participate were
453
ARTICLES
included. Nevertheless, only three datasets provided
information on deaths in the first year of life, and six
provided data on deaths in the second year. Also, we
have not been able to rule out publication bias,
which is likely to result in the exclusion of small studies
with negative findings. Of the two datasets that were
identified but could not be included, one showed a
protective effect of breastfeeding, and the other was not
published.
Observational studies of breastfeeding and infant health
may be affected by a number of methodological problems,
including
self-selection,
reverse
causality,
and
confounding, which have been reviewed in detail else
where.-’4"’7 In this analysis, an attempt was made to avoid
reverse causality by recording breastfeeding before rhe
fatal illness episode. Also, stratification by maternal
education—a potential confounder and effect modifier—
showed a consistent increase in risk for weaned infants in
most education categories.
An extensive review of the literature on breastfeeding
and all-cause mortality revealed one meta-analysis
published in 1984-’’ (which covered eight pre-1950 studies
and one 1979 study), and 13 other studies.-"" Most of
these consisted of reanalyses of existing datasets, which
were mainly retrospective demographic surveys. These
studies were not included in the present analysis because
they did not provide information on cause of death. The
studies vary substantially in terms of definitions of
breastfeeding variables, in the types of analyses used, and
on how the results were presented. Most investigators did
not attempt to rule out reverse causality. Owing to these
limitations, most of the results could not be compared
with our findings. However, all studies showed significant
protective effects of breastfeeding, at least for some age
groups. The review confirmed that our reanalysis included
most of the well-designed, published studies of
breastfeeding and mortality.
Even studies that made specific attempts to control for
bias have shown a protective effect of breastfeeding? 4--4
and several reviews agree that the overall evidence is
compelling?4 Randomised trials of breastfeeding promotion
and infant mortality would be prohibitively difficult and
ethically questionable, but one trial in India compared the
effects of different types of milk given by bottle to high-risk
newborns.41 Raw human milk provided substantial
protection against the incidence of septicaemia,
conjunctivitis, diarrhoea, and umbilical sepsis. Our decision
to exclude first-week deaths from this analysis—in order to
avoid reverse causality—will underestimate the protection
provided by breastmilk against neonatal infections. Another
issue to bear in mind in this reanalysis is that studies did not
employ a uniform protocol for defining causes of death.
Most studies used verbal autopsy methods which are
substantially more accurate for some causes (such as
measles and diarrhoea) than for others (such as acute
respiratory infections or malaria).4-4' Nevertheless, a higher
level of protection was provided against diarrhoea than
against acute respiratory infections in the first six months of
life within each study. Unfortunately, most studies did not
provide sufficient information on breastfeeding partem
(exclusive, predominant, or partial) to allow a pooled
analysis of this variable. Breakdown by sex suggested a
slightly larger protection for girls, but, as noted, the
confidence intervals overlap.
Our reanalysis showed that infants who are not
breastfed have a six-fold greater risk of dying from
454
infectious diseases in the first 2 months of life than those
who are breastfed, but that protection decreases steadily
with age, and is probably due to lower intakes by older
children who also receive complementary feeding. There
was significant heterogeneity within the pooled estimate
for the first age group, so this estimate should be regarded
with caution. The heterogeneity was mainly due to the
very high odds ratio observed in Pakistan (21-3, based on
17 deaths); when Pakistan was excluded, there was no
significant heterogeneity. Since 98% of infants under
2 months of age in Pakistan were reportedly breastfed,
those who were not breastfed may have had underlying
morbidity. In the first year of life, protection was
observed in three studies from three different continents.
For the second year, results were less consistent, but still
suggested a protective effect.
We suggest that our data should be used in future
simulations of the impact of withholding breastfeeding in
HIV-positive mothers. Since our estimates refer to
mortality due to infectious diseases, the relative weight of
infections in overall infant and child mortality should be
taken into account in the simulations. Our results will also
help assess the risks and benefits associated with
breastfeeding for children of different ages, since the risk
of HIV transmission also depends on duration of
breastfeeding?4” Of direct relevance to the debate on HIV
and breastfeeding are the higher levels of protection seen
among less educated women, particularly for deaths at
ages 6-11 months. Such results are consistent with the
finding that infant-mortality differentials according to
breastfeeding status are virtually non-existent in more
developed countries where maternal education is high?"
The main policy issue is whether or not HIV-positive
mothers with low levels of schooling and income will be
able to feed their infants with safe breastmilk substitutes.
Earlier research
from Malaysia showed that the
association between breastfeeding and mortality is twice
as strong in households without piped water or a toilet
than in those with such facilities?’ This question exceeds
the scope of our analysis, but our results suggest that it
will be difficult, if not impossible, to provide safe
breastmilk substitutes to children from underpriviledged
populations.
There is a wide agreement that the final decision
on whether or not to breastfeed if a woman is HIV
positive should reside with the mother and the family? We
hope that our results will provide a firmer basis for policy
decisions on how to help families in this difficult situation.
WHO Collaborative Study Team on the Role of Breastfeeding on
the Prevention of Infant Mortality
Writing committee—C G Victors, AJD Barros.
Brazil— C G Victors, AJD Barros, S C Fuchs (Federal University- of
Pelotas and Federal University of Rio Grande do Sul, Brazil).
Gambia—A de Francisco, J Morris, A J Hall, J R M A Schellenberg,
B M Greenwood (London School of Hygiene and Tropical Medicine,
UK; and Global Forum for Health Research, Switzerland).
Ghana—B R Kirkwood, P Arthur, D A Ross, S S Morris (London School
of Hygiene and Tropical Medicine, UK; and Ministry of Health, Ghana).
Pakistan—S Zaman, F Jalil, J P Karlberg, L A Hanson, R Ashraf (King
Edward Medical College, Pakistan; University of Hong Kong, Hong
Kong; and University- of Gbteborg, Sweden).
Philippines—R E Black, S Becker, P Yoon, R Gray, S A Gultiano (Johns
Hopkins University School of Public Health, USA; and University of San
Carlos, Philippines).
Senegal—M Garenne, O Fontaine, O Leroy, J P Beau, I Sene (Centre
Francais sue la Population er le Developpement, France; and ORSTOM,
UR Population et Same, Senegal).
WHO Secretariat—Jose C Martines.
Acknowledgments
This study was supported by the WHO Division of Child and Adolescent
Health.
THE LANCET • Vol 355 • February 5, 2000
ARTICLES
References
1 WHO UNAIDS "UNICEF HIV and infant feeding. Guidelines for
decision-makers. Geneva: WHO, 1998.
2 Jelliffe DB, Jelliffe EFP. Human milk in the modem world. Oxford.
OUT, 1978
3 Feachem RG, Koblinski MA. Interventions for the control of
diarrhoeal diseases among young children: promotion of breastfeeding.
Bull World Health Organ 1984; 62: 271-91.
4 Jason JM, Nieburg P, Marks JS. Mortality and infectious disease
associated with infant-feeding practices in developing countries.
Pediatrics 1984; 74: 702-27.
5 Cunningham AS, Jelliffe DB, Jelliffe EFP. Breastfeeding and health in
the 1980s; a global epidemiological review. J Pcdiatr 1991; 118:
659-66.
6 Victoria CG. Infection and disease: the impact of early weaning
Food Nutr Bull 1996; 17: 390-96.
7 Murray CJL, Lopez AD. The global burden of disease, a comprehensive
assessment of mortality and disability from diseases, injuries and risk
factors in 1990 and projected to 2020. Global Burden of Disease and
Injury Series, vol 1. Cambndge* Harvard School of Public Health. 199o.
8 He-man SJ. Modelling the impact of breastfeeding by HIV-infected
women on child survival. Am J Public Health 1990; 80: 1305-09.
9 Hu DJ, Heyward WL, Byers RH, et al. HIV infection and
breastfeeding: policy implications through a decision analysis model.
AIDS 1992, 6: 1505-13.
10 Del Fame P. Jenniskens F, Lush L, et al. HIV. breastfeeding and
under-5 mortality: modelling the impact of policy decisions for or
against breastfeeding. J Prop Med Hyg 1993, 96: 203-11
12 Nicoll A, Newell M-L, Van Praag E, Van de Perre P. Peckham C.
Infant feeding policy and practice in the presence of HIV-1 infection.
AIDS 1995; 9: 107-19.
13 Khun L. Infant survival, HIV infection and feeding alternatives in less
developed countries. Am J Public Health 1997; 87: 926- 31
14 Nduati R. HIV and infant feeding: a review of HIV transmission
through breastfeeding. Geneva: WHOUNAIDS, 1998
15 Collaborative Group on Hormonal Factors in Breast Cancer. Breast
cancer and hormonal contraceptives: collaborative reanalysis of
individual data on 53 297 women with breast cancer and 100 239
women without breast cancer from 54 epidemiological studies. Lancet
1996; 347: 1713-27.
16 Victora CG, Huttly SRA, Fuchs SC, et al. International differences in
clinical patterns of diarrhoeal deaths: a comparison of children from
Brazil, Senegal, Bangladesh and India. J Diarrhoeal Dis Res 1993; 11:
25-29.
17 Breslow NE, Day NE. Statistical methods in cancer research, vol 1.
Lyon: International Agency for Research on Cancer, 1980.
18 Victora CG, Smith PG, Vaughan JP, et al. Evidence for a strong
protective effect of breast-feeding against infant deaths due to
infectious diseases in Brazil. Lancet 1987; 2: 319-22.
19 de Francisco A, Morris J, Hall AJ, Schellenberg JRMS,
Greenwood BM. Risk factors for mortality from acute lower
respirators' tract infections in young Gambian children, bit J Epidemiol
1993;22: 1174-82.
20 Ghana VAST Study Team. Vitamin A supplementation in northern
Ghana: effects on clinic attendances, hospital admissions, and child
mortality. Lancet 1992; 342: 7-12.
21 Hanson LS, Ashraf R, Zaman S. Karl ber J, Lindblad BS. Jalil F.
Breastfeeding is a natural contraceptive and prevents disease and
death in infants, linking infant mortality and birth rates. Acta Pacdiatr
1994;83:3-6.
22 Yoon P, Black RE. Moulton LH, Becker S. Effects of not
breastfeeding on the risk of diarrhoeal and respiratory mortality in
children under 2 years of age in Metro Cebu, the Philippines. Am J
Epidemiol 1996; 143: 1142-48.
23 Garenne M, Fontaine O. Assessing probable causes of deaths using a
standardized questionnaire, a study in rural Senegal. In Vallin J,
D'Souza S, Palloni A (eds). Measurement and analysis of mortality.
Oxford: Clarendon Press, 1986; 123-42.
THE LANCET • Vol 355 • February 5, 2000
24 Habicht JP, DaVanzo J, Butz WP. Does breastfeeding really save lives,
or are potential benefits due to biases? Am J Epidemiol 1986; 123:
279-90.
25 Sauls HS. Potential effect of demographic and other variables in
studies comparing morbidity of breast-fed and bottle-fed infants.
Pediatrics 1979; 64: 523-27.
26 Victora CG Case-control studies of the influence of breastfeeding on
child morbidity and mortality, methodological issues. In. Atkinson SA,
Hanson LA, Chandra RK. eds. Breastfeeding, nutntion, infection and
infant growth in developed and developing countries. St John’s,
Canada: ARTS Biomedical Publishers and Distributors, 1990:
405-18.
27 Bauchner H. Leventhal JM. Shapiro ED. Studies of breast-feeding
and infection* how good is the evidence? JAMA 1986;
256: 887-92.
28 Feachem RG, Koblinski ALA. Interventions for the control of diarrheal
diseases among young children: promotion of breastfeeding. Bull
World Health Organ 1<>8 !; 62: 271-91.
29 Habicht JP. DaVanzo J. Butz WP. Mother's milk and sewage; their
interactive effects on infant mortality. Pediatrics 1988; 81: 456-61.
30 Goldberg HI, Rodrigues W. Thome AMT, Janowitz B Moms L.
Infant mortality and breast-feeding in Northeastern Brazil. Popul Stud
1984.38: 105-15
31 Plank SJ, Milanesi ML. Infant feeding and infant mortality in rural
Chile. Bull World Health Organ 1973, 48: 203-10.
32 Molbak K. Gottschau A. Aaby P, Hojlyng N, Ingholt L. da Silva AP.
Prolonged breast feeding, diarrhoeal disease, and survival of children
in Guinea-Bissau. BMJ 1994. 308: 1403-06
33 Palloni A, Aguirre GP. Lastin S. The effects of breastfeeding and the
pace of childbearing on early childhood mortality in Mexico Bull Pan
Am Health Organ 1991. 28: 93- 111.
34 Tekce B. Oral rehydration therapy: an assessment of mortality effects
in rural Egypt. Stud Fam Plaim 1982; 13: 315-27.
35 Janowitz B, Lewis JH. Pamell A. Hefnawi F, Younis MN, Serour GA.
Breast-feeding and child survival in Egypt. J Biosoe Sei 1981; 13:
287-97.
30 Nath DC, Land KC. Singh KK. Birth spacing, breastfeeding, and
earls child monalits* in a traditional Indian societv: a hazards model
analysis. S«k Biol 1994; 41: 168-80.
37 Shahidullah M. Breast-feeding and child survival in Matlab,
Bangladesh. J Biosoe Set 1994; 26: 142-54.
38 Ahmad OB, Eberstem IW, Sly DF. Proximate determinants of child
mortality in Liberia. J Biosoe Sci 1991; 23: 313-26.
.39 Majumder AK. Breast-feeding, birth interval and child mortality in
Bangladesh. J Biosoe Sci 1991; 23: 297-312.
40 Shamebo D, Muhe L. Sandstrom A, Freij L, Krantz A, Wall S. The
Butijara Rural Health Project in Ethiopia: a nested case-referent
(control) study of under-5 mortality and its health and behavioural
determinants. Ann Trop Pcdiatr 1994,14: 201-09.
41 Narayanan I, Prakash K. Murthy NS, Gujral W. Randomized
controlled trial of effect of raw and pasteurized human milk and of
formula in the incidence of neonatal infection. Lancet 1984; 2:
42 Gray RH, Simit G, Barss P The use of verbal autopsy methods to
determine selected causes of death in children. Baltimore*. Johns
Hopkins University. 1990.
43 Kalter HD. Gray RH. Black RE. Guiliano SA. Validation of
postmortem interviews to ascertain selected causes of death in
children, bit J Epidemiol 1990; 19: 380-86.
44 Simonon A. Lepage P. Kama E, et al. An assessment of the timing of
mother-to-child transmission of HIV-1 bv means of polymerase chain
reaction. J AIDS 1994; 7: 952-57.
45 Bertolii J, St Louis ME. Simonds RJ, et al. Estimating the timing of
mother-to-child transmission of HIV in a breastfeeding population in
Kinshasa, Zaire. J Infect Dis 1996; 174: 722-26.
16 Kovar MG. Seerdula MK. Marks JS, Fraser DW. Review of the
epidemiologic evidence for an association between infant feeding and
infant health. Pediatrics 1984; 74: 615-38.
455
How to be Reasonably Sure a Client is Not Pregnant
If the client answers YES to any question, proceed to
the box at the bottom of the YES column
NO
1.
Are you less than 6 months postpartum and
fully breastfeeding and free from menstrual
bleeding since you had your child?
2.
Have you abstained from sexual intercourse
since your last menses?
Have you given birth in the last 4 weeks?
Did your last menstrual period start within
the past 7 days?
Have you had a miscarriage or abortion in
the past 7 days?
Have you been using a reliable contraceptive
method consistently and correctly?
3.
4.
5.
6.
YES
I
Client answered NO to all
of the questions
Client answered YES to at
least 1 question
f
Pregnancy cannot be ruled
out
i
Client should await
menses or use pregnancy
test
Client is free of signs or
symptoms of pregnancy
_______ t_____
Provide client with desired
method
WHO/FRH/FPP/96.9
Original: English
Distr.: General
Reproduced with permission from WHO
Improving Access to Quality Care
in Family Planning
Medical Eligibility Criteria for Initiating
and Continuing Use of Contraceptive Methods
Low-dose combined oral contraceptives (COCs)
Emergency contraceptive pills (ECPs)
Combined injectable contraceptives (CICs)
Progestogen-only pills (POPs)
Depot Medroxyprogesterone acetate (DMPA)
Norethisterone enanthate (NET-EN)
Norplant implants (NOR)
Copper IUDs (Cu-IUDs)
Levonorgestrel-releasing IUD (LNG-IUD)
IUD for emergency contraception
Male sterilization
Female sterilization
Natural family planning (NFP)
Barrier methods
Lactational amenorrhea
® World Health Organization, 1996
This document is not a formal publication of the World Health Organization (WHO), and
all rights are reserved by the Organization. The document may, however, be freely
reviewed, abstracted, reproduced and translated, in part or in whole, but not for sale
nor for use in conjunction with commercial purposes.
The views expressed in documents by named authors are solely the responsibility of
those authors.
Summary Tables
Page 125
SUMMARY TABLES
CONDITION
coc
CIC
POP
NET-EN
DMPA
NOR
Cu-IUD
LNG-IUD
4
4
3
I = Initiation, C = Continuation
PREGNANCY
4
4
4
4
4
BREAST-FEEDING
a)
<6 weeks postpartum
4
4
3
3
3
b)
6 weeks to 6 months:
(primarily breast-feeding)
3
3
1
1
1
c)
>6 months postpartum
2
2
1
1
1
< 21 days
3
3
1
1
1
S 21 days
1
1
1
1
1
POSTPARTUM
(in non-breast-feeding women)
POSTPARTUM
(breast-feeding or non-breast-feeding)
including post-caesarian section
a)
< 48 hours
2
b)
48 hours to 4 weeks
3
3
c)
£ 4 weeks
1
1'”
d)
Puerperal sepsis
4
4
If the woman is breast-feeding, LNG-IUD becomes a category 3 until 6 weeks postpartum.
Summary Tables
Page 126
SUMMARY TABLES
coc
CONDITION
CIC
POP
NET-EN
DMPA
NOR
Cu-IUD
LNG-IUD
1 = Initiation, C = Continuation
POST-ABORTION
a|
First trimester
1
1
1
1
1
1
1
b)
Second trimester
1
1
1
1
1
2
2
c)
Immediate post-septic abortion
AGE
1
1
1
1
1
4
4
<40= 1
<40= 1
<16 = 2
<16 = 2
<16 = 2
<20 = 2
<20 = 2
>40 = 2
>40 = 2
>16 = 1
>16 = 1
>16=1
>20 = 1
>20= 1
2
2
1
1
1
1
1
SMOKING
a)
Age < 35
b)
Age > 35
light
3
2
1
1
1
1
1
heavy
(> 20 cigarettes/day)
4
3
1
1
1
1
1
3
2
2
2
1
2
ESSENTIAL HYPERTENSION
a)
History of hypertension where blood
pressure cannot be evaluated
(excluding hypertension in pregnancy)
1
C
3
3
CJYV- 3 0
ARTICLES
Reproduced with the permission of The Alan Guttmacher Institute
from: Tahpa S et al., Contraceptive Use Among Postpartum Women:
Recent Patterns and Prgrammatic Implications. International
Family Planning Perspectives, 1992. 18(3):83-92.
Contraceptive Use Among Postpartum Women:
Recent Patterns and Programmatic Implications
By Shyam Thapa, Sushil Kumar, Jeanne Cushing and Kathy Kennedy
This study of postpartum women, based on Demographic and Health Surveys in 25 develop
ing countries, reveals that the proportion of women who are exposed to the risk of pregnancy
within two years after childbirth ranges from one-third in Sub-Saharan Africa to nearly two-thirds
in Latin America and the Caribbean. More than half of postpartum women are current contra
ceptive users. Women exposed to the risk ofpregnancy are more likely than unexposed women
to be using reversible methods, usually the pill. Among women who are unexposed to the risk
ofpregnancy as a result of abstinence or amenorrhea associated with breastfeeding, 19% are
using a contraceptive method, usually sterilization. The proportion of contraceptive users who
initiate use of a modem method before menses returns ranges from 27-57% among countries
in Latin America and the Caribbean and Asia, and from 24-46% among African countries. Small
er proportions of hormonal contraceptive users initiate use before the return of menses. About
one-fifth of exposed women are not using any contraceptive method. Of this group, more than
one-third want no more children and another one-third want to space their next pregnancy.
(International Family Planning Perspectives, 18:83,1992)
he concept of implementing special
family planning programs for post
partum women is not new. As early
as 1966, The Population Council intro
duced the International Postpartum Pro
gram in a few countries to evaluate the
feasibility of providing family planning
services for women immediately after
childbirth or abortion.1 The program was
gradually expanded to 21 countries, with
services offered in 135 urban areas and
four large rural areas (in the Philippines,
Indonesia, Nigeria and Turkey).2
T
Shyam Thapa and Kathy Kennedy are senior research
associates at Family Health International, Research Tri
angle Park, N. C. Sushil Kumar is consulting associate
and Jeanne Cushing is computer training coordinator at
the Institute for Resource Development. Columbia, Md.
vided financial support for this study. The authors thank
Nancy Williamson, I-Cheng Chi and John Ross for their
helpful comments on an earlier draft of the paper. The
views expressed in this article are those of the authors;
they do not necessarily represent those of the funding
agency, the reviewers or the organizations with which
the authors are associated. Preliminary results of this
study were presented at the International Postpartum
Conference, Mexico City, Mexico, Sept. 17-19.1990, and
at the Demographic and Health Surveys World Confer
ence. Washington, D.C., Aug. 5-7.1991.
Volume 18. Number 3. September 1992
International policymakers, program
managers and health care providers have
recently expressed renewed interest in
postpartum family planning programs.
The following factors may have con
tributed to the recent attention: modest im
provements in contraceptive technology
or its application;3 increased appreciation
of the role of breastfeeding in regelating fer
tility;4 expansion of the service delivery in
frastructure; increased support from donor
agencies; and, perhaps most important, the
realization that sen ice delivery during the
postpartum period, which is thought to be
a desirable and effective time to introduce
contraceptives, has not been promoted in
family planning programs.
Strategies for the delivery of postpar
tum family planning services have been
based partly on the demographic rationale
that the sooner contraceptives are adopt
ed after childbirth, the greater the impact
on fertility, through longer birth intervals.
These strategies have also been based on
the behavioral assumptions that women
tend to be more motivated to accept con
traceptives immediately after childbirth
than at other times, and that their contact
with medical facilities at the time of de
livery is an ideal opportunity for health
care providers to offer family planning
counseling and services.
In a recent article, Beverly Winikoff and
Barbara Mensch critically examined these
issues and questioned both the demo
graphic rationale and the behavioral as
sumptions. They concluded that there is
"little conceptualization of the postpartum
period and the needs of the [postpartum]
women." They also pointed out that, from
the women's perspective, "it is necessary
to understand more about women's needs,
their preferences for an array of health ser
vices and their desires as to the timing and
type of contraceptives available to them."5
This study examines contraceptive be
havior and needs among postpartum
women in developing countries, and ad
dresses three interrelated issues. First,
what is the prevalence of contraceptive
use among postpartum women and when
do they begin using contraceptives after
childbirth? Second, what are their prefer
ences for future childbearing? Third, to
what extent do women have contact with
health care personnel before, during and
after the delivery, who might help them
initiate contraceptive use?
Data and Definitions
Data on postpartum women in 25 devel
oping countries were obtained from the first
phase of the Demographic and Health Sur
veys (DHS), completed between 1986 and
1989. The countries included 10 in Sub-Sa
haran Africa, three in North Africa, nine in
Latin America and the Caribbean, and three
in Asia. They represented approximately
35% of the total population of developing
countries, excluding China and India.
The term "postpartum period" is con
ventionally defined in medical circles as
the time from parturition until six weeks
after childbirth. This definition reflects
clinical considerations, mainly, the usual
83
Contraceptive Use Among Postpartum Women
time for involution of the uterus to have
been completed. Although this definition
is generally accepted in clinical contra
ceptive research, family planning pro
grams have not established a consistent
definition of the postpartum period. Winikoff and Mensch have suggested that in
family planning programs, postpartum is
commonly considered an "approach" to
the delivery of contraceptives immedi
ately after childbirth or abortion.6 The lack
of a time-bound definition for family plan
ning providers probably reflects the need
to consider programmatic aspects of con
traceptive service delivery as well as phys
iologic circumstances.
Previous analyses on postpartum con
traception in developing countries have
been based on various lengths of time after
childbirth, including up to 24 months.7 For
several reasons, we defined the postpar
tum period in this analysis as up to two
years (23 completed months) since the
birth of the last child. First, the two-year
time period encompasses a substantial
proportion of women of reproductive age,
from an average of 25% of women in lower
fertility countries to an average of about
45% in higher fertility countries. Overall,
the postpartum period used in this analy
sis also allows adequate time for the initi
ation of contraceptive use by a substantial
proportion of women who had had a re
cent birth; thus, postpartum contraceptive
behavior can be examined thoroughly.
Second, breastfeeding for long durations
is the rule rather than the exception in most
developing countries. According to the
DHS data, the median duration of breast
feeding in the regions surveyed ranges
from 12 months in Latin America and the
Caribbean to 20 months in Sub-Saharan
Africa. In some countries, weaning serves
as a proximate marker for the initiation of
contraceptive use after childbirth.8
Third, the longer the duration of breast
feeding, the longer the duration of amen
orrhea, at least at the aggregate level.”
Applying John Bongaarts's model of prox
imate determinants to DHS data con
firmed this relationship (Figure 1), sug
gesting that each increase in duration of
breastfeeding by one month increases the
duration of postpartum amenorrhea by
one-half to three-fourths of a month.
In addition to having given birth with
in 24 months prior to the stud v, women in
cluded in the sample had to have been
currently married or in a consensual union
and not pregnant at the time of the survey.
A small proportion of women whose last
pregnancy was ended by abortion was ex
cluded from the study.
Figure 1. Median duration of postpartum amenorrhea, by median duration of breastfeeding, 25
DHS countries
Months of amenorrhea
20______________________________ :_________________________________ _______
0
5
Months of breastfeeding
10
15
20
25
Note. YnMedian curation of postpartum amenorrnea X»Medtan duration of breastfeeding
We classified the postpartum women
into two categories of exposure to the risk
of pregnancy: unexposed and exposed.
Unexposed women included those who
were amenorrheic or abstaining from sex
ual intercourse at the time of the survey;
exposed women included those who had
resumed menstruating since their last
birth and who were not abstaining from
sexual intercourse. In the analysis of sur
vey data, the resumption of menses is gen
erally used as an indicator of a woman's
return to fertility. Contraceptive use in this
analysis refers to use of a method at the
time of the survey.
The classification of the postpartum
women into two categories of exposure
addresses behavioral and programmatic
considerations. Behaviorally, evidence
suggests that the return of menses after
childbirth is an important signal for wom
en to start using contraceptives’”—a more
important signal, in fact, than breast
feeding status.” Recent studies have
shown that the probability of pregnancy
during amenorrhea is low.12 One studv
found that 6% of amenorrheic women were
likely to be pregnant at the end of a year,
while 177 of all breastfeeding women
were likely to conceive.n
Programmatically, the initiation of use
of reversible contraceptive methods, such
as the pill, immediately after delivery may
have either no impact or an adverse im
pact on fertility. Evidence from rural
Bangladesh showed that the initiation of
pill use immediately after childbirth can
lead to shorter, not longer, birth intervals
as a result of early abandonment of both
contraceptive practice and breastfeeding.14
A related consideration is the phenom
enon referred to as the "double protection
dilemma": If the duration of lactational
Table 1. Percentage distribution of postpartum
women, by exposure to the risk of pregnancy,
according to country, 1986-1989
Country
Exposed
Unexposed
Total
Asia
Indonesia
Sri Lanka
Thailand
57.5
48.6
58.1
65.9
42.5
51.4
41.9
34.1
100.0
100.0
100.0
100.0
Latin America/
Caribbean
Bolivia
Brazil
Colombia
Dominican
Republic
Ecuador
Guatemala
Mexico
Peru
Trinidad &
Tobago
62.2
44.7
78.6
73.2
100.0
100.0
100.0
100.0
68.8
55.8
37.3
68.5
51.8
100.0
100 0
100.0
100.0
100.0
80 9
100.0
North Africa
Egypt
Morocco
Tunisia
59.4
58.0
57.4
62.8
100.0
100.0
100.0
100.0
Sub-Saharan
Africa
Botswana
Burundi
Ghana
Kenya
Liberia
Mali
Senegal
Togo
Uganda
Zimbabwe
32.6
45.0
19.1
276
46.2
31.2
29.0
24.9
21.8
36.3
45.0
100.0
100.0
100.0
100.0
1000
100.0
100.0
100.0
100.0
100.0
100.0
Note: In this and subsequent tables, postpartum refers to women
who are currently married, are not pregnant and have had a birth
Within 24 months before the survey, exposed refers to postparnot abstaining Because of rounding, figures may not add to 100.
Regional averages are unweighted
International Family Planning Perspectives
Table 2. Percentage distribution of postpartum women, by exposure to the risk of pregnancy and by method use, according to country
Country
Asia
Indonesia
Sn Lanka
Thailand
Unexposed
Exposed
Total
All
Pill
methods
IUD
Injectables
Stemization'
Other
NO
method
All
methods
Pill
IUD
Inject
ables
Steril
ization*
Other
No
method
64 0
71 0
82 0
18.5
122
24 7
136
24
12.9
196
8.0
186
16
16.5
20 8
107
31.9
5.1
36.0
29.0
180
27 9
32.8
38.5
3.8
07
2.5
6.3
0.5
6.6
10 1
10.4
14
12.6
17.3
6.3
17.7
1.5
72.1
67.2
61 5
100.0
100.0
100.0
41 2
78 3
81 5
35
43.0
276
5.3
0.9
15 1
i T
16
5.9
2.9
20.3
12.2
28.2
12.6
20.8
58.8
21.7
18.5
7.8
30 5
25.3
0.0
4.1
0.7
0.0
3.0
01
0.4
0.0
1.0
203
5.8
5.9
5.7
12.4
92.2
69.5
74.7
100.0
100.0
100 0
73.6
87.0
94.9
83.6
90.8
. 100 0
r 100.0
100.0
100.0
100.0
Bolivia
Brazil
Colombia
Dominican
Republic
Ecuador
Guatemala
Mexico
Peru
Trimdad &
Tobago
65 7
55.0
28.8
68 1
63.1
22 3
12.5
8.7
19.0
11.7
6.4
156
2.2
15.8
10.2
0.0
1.8
0.9
5.5
3.0
26.3
10.7
5.8
13.4
3.0
10.6
14.3
11 2
14 4
35.2
34.3
45.0
71.2
31.9
36 9
26.4
13.0
16.4
9.2
40
09
03
0.2
0.0
0.5
2.5
0.1
26
1 i
0.0
02
03
01
0.2
17.2
5.9
2.3
8.4
16
4.8
3.4
20
75.7
23.1
5.2
1.7
6.6
39.1
24.3
36.5
2.1
1.0
31
8.3
21.9
63.5
100.0
North Africa
Egypt
Morocco
Tunisia
57.7
57.1
653
25.7
42 3
15.8
24.3
3.1
27 7
0.1
0.9
1.0
0.8
0.8
4.3
6.7
10.0
16.5
42 3
42 9
34 7
12.2
65
9.7
1.9
07
4.6
0.6
2.2
0.0
0.1
0.2
04
0.5
2.8
5.8
3.5
37
87.8
93.5
90.3
. 100.0
100.0
■ 100.0
Sub-Saharan Africa
Botswana
45.8
Burundi
26 3
Ghana
26.4
Kenya
36.4
Liberia
11.6
Mali
5.8
Senegal
17.6
Togo
50.6
Uganda
Zimbabwe
68 2
24.2
0.9
4.9
10.6
8.6
1.7
5.2
1.2
27
51.1
8.0
i1
0.5
4.0
1.3
0.3
2.8
21
0.3
1.1
7J
3.2
0.5
53
0.0
0.0
0.6
0.0
0.6
0.0
1.0
0.1
00
28
0.0
0.0
0.0
0.4
0.4
0.6
5.6
21.0
20.5
13.7
1.8
37
9.1
46.9
4.2
15.5
54.2
73.7
73.6
63.6
88.4
94.2
82.4
49.4
91 9
31.8
14.5
8.7
9.4
15.2
0.7
S7
203
468
19
50.7
57
0.0
0.4
22
0.3
0.1
0.2
0.5
0.3
37.4
0.0
0.1
1.1
0.0
0.0
0.3
0.2
0.1
0.5
45
0.0
0.1
0.8
0.0
0.0
0.0
0.0
0.2
0.2
14
0.1
0.5
17
0.2
0.0
0.1
0.2
0.3
0.4
16
8.6
8.3
9.3
0.2
5.5
198
45.9
1 1
123
85.5
91.3
90.6
84.8
99.3
94.3
79.7
53.2
98.1
49.3
I 100.0
! 100.0
. 100.0
100.0
, 100.0
i 100.0
100.0
iI 100.0
; 100.0
100 0
6.2
'Sterilization includes male and female.
amenorrhea is very long, the introduction
of contraceptives early in the postpartum
period may have virtually no demo
graphic impact, and a considerable pro
portion of the potential effect of contra
ceptive practice will be "wasted."15
Results
The proportion of exposed women ranged
from an average of about one-third in SubSaharan Africa to almost two-thirds in
Latin America and the Caribbean (Table
1). In the Sub-Saharan African countries
studied, at least 55%—and in some coun
tries more than 75%—of women who had
given birth within the last two years were
still unexposed to the risk of pregnancy.
We found considerable variation in ex
posure to pregnancy risk among women
in Sub-Saharan Africa and in Latin Amer
ica and the Caribbean. Exposure status
was associated largely with breastfeeding
and to some extent with practices of post
partum abstinence.
Contraceptive Use
Contraceptive use was considerably lower
among unexposed women than among
exposed women. In the 25 countries sur
veyed, an average of only 19% of unex
Volume 18, Number 3, September 1992
posed women were using a method of
contraception, while more than 50% of ex
posed women were using a method. The
ratio of contraceptive use of the unex
posed group to the exposed was 0.38.
Twelve percent of unexposed women and
41% of exposed women were using mod
ern contraceptive methods (the pill, the
IUD, injectables, implants, barrier meth
ods or sterilization).
Not only was overall contraceptive use
considerably lower among unexposed
women compared with exposed women,
but the methods used were markedly dif
ferent. When we analyzed only those
countries with at least 5% prevalence of
each modem method, sterilization (female
or male) appeared to be the most com
monly used method by unexposed wom
en. The pill was the predominant method
among exposed women, followed by the
IUD, sterilization and injectables.
Among exposed women (Table 2), the
pill was used by at least 5% of the women
in 19 of 25 countries. The proportions of
women who were using the pill varied
considerably among countries. Use of the
pill was particularly high in Zimbabwe
(51%), Brazil (43%) and Morocco (42%).
The IUD was used by at least 5% of the
women in 12 countries, most of which are
in Asia, North Africa, and Latin America
and the Caribbean. Use of the IUD by ex
posed women was more than 5% in only
one country (Botswana) in Sub-Saharan
Africa. Among the countries surveyed, the
use of sterilization was limited almost ex
clusively to Asia and Latin America and
the Caribbean; in Indonesia and Bolivia,
however, it was not used at all. Injectables
were used by 8-20% of exposed women
in the three Asian countries surveyed, and
by 6-8% of women in Colombia, Mexico,
Botswana and Kenya.
Among unexposed women, steriliza
tion was used by at least 5% of the women
in eight of the 25 countries, and was par
ticularly popular—used by 13-20%—in
Brazil, the Dominican Republic, Thailand
and Sri Lanka. Reversible methods were
used by unexposed women in only a few
countries. The IUD was used by an aver
age of 6-7% of the women in Thailand and
Indonesia, while injectables were used by
an average of 10% of the women in these
two countries. Botswana was the only
other country where injectables were used
by about 5% of unexposed women. The
pill was used by at least 5%> of unexposed
women in only two countries—Zimbab85
Contraceptive Use Among Postpartum IVohiwi
we (37“><) and Botswana (6!'<). Use of the
pill in all other countries was negligible.
To examine when during the postpar
tum period women were more likely to use
contraceptives, we compared women who
were using the pill, the IUD, in jectables or
sterilization, at six different postpartum pe
riods, by country. The proportion of unex
posed women using reversible methods in
each postpartum period was usually less
than 59i, and so with few exceptions, use
bv exposed women was consistently and
considerably higher than that of unexposed
women in each time period. The data for
exposed women are shown in Table 3.
Use of the pill by exposed women in
creased in Trinidad and Tobago, and
Colombia after the third or the sixth
month postpartum, and stabilized in
subsequent postpartum months In
Liberia and Senegal, use of the pill in
creased after the sixth month postpar
tum, but decreased at later months In a
few countries—Zimbabwe, Brazil, Mex
ico, Peru and Thailand—the proportion
of exposed women using the pill was at
a high point in the first three months
after birth, then declined at 4-6 months
postpartum. Zimbabwe appeared to be
a special case, in that the pill was used
by as many as three-fourths of exposed
women at 0-3 months postpartum.
Where IUD use was measurable, in
about half of the 12 countries, its use
among exposed women increased with
time, particularly at 4-6 months; in the
other countries, IUD use declined after 12
months. In Tunisia and Mexico, the IUD
was used by a substantial proportion ol
exposed women. Its use among ’exposed
women stabilized in Tunisia at 7-9 months
postpartum.
Injectables were used by considerable
proportions of exposed women in Thai
land and Indonesia. In both countries, use
of injectables was fairly constant during
the postpartum period.
Sterilization among exposed women
generally increased over time in all the
countries where it was used by at least 5‘ <
of postpartum women. Sterilization was
particularly high after nine months post
partum in Brazil, the Dominican Repub
lic, Thailand and Sri Lanka.
The timing patterns for reversible meth
ods used by more than 5ci of unexposed
women are shown in Table 4. The pro
portion of unexposed women in Zim
babwe who were using the pill increased
with time postpartum. In Botswana, the
only other country where the pill was
used by at least 591 of the unexposed
women, the trend was less distinct.
86
Table 3. Among postpartum women exposed to the risk of pregnancy, percentage using selected contraceptive methods, by country and duration since last birth
Country and
months since
last birth
ASIA
Indonesia
0-3
4—6
7-9
10-12
13-18
19-23
Sri Lanka
0-3
4-6
10-12
13-18
19-23
Thailand
0-3
. - '4
Pill
58
13.0
12.8
19.8
23 6
IUD
13.0
11.81
7.6
13 11
15.21
0.0
10 6
11 2
91
157
28 2
170
25 8
23 7
5.5
102
17 11
10.51
23 0
14 31
•_ATIN AMERICA'CARIBBEAN
Bolivia
156
2.5
4—t
2.8
9.1
10-12
13-18
Inject
ables
Steril
ization
20 8
19.9
299
17.9
19*2
109
59
94
96
7.3
80
9.6
47
7.9
95
22.0
21.7
20.2
23 3
172
13.8
18.2
9.6
10.5
19.6
25 0
23 5
21 4
Country and
months since
last birth
Pill
Peru (cont.)
91
10-12
13-18
136
19-23
135
Trinidad & Tobago
0-3
107
12.0
7-9
26.6
10-12
25.8
13-18
24 1
19-23
25 8
NORTH AFRICA
Egypt
0-3
234
4-6
26.6
7-9
27 3
10-12
24.4
13-18
25.8
19-23
26.1
Morocco
25.0
4-6
32.1
7-9
46.5
10-12
38.4
13-18
44.8
19-23
49.8
Tunisia
0-3
4-6
14.3
13.1
10-12
20.2
16.7
13-18
19-23
16.4
SUB-SAHARAN AFRICA
Botswana
9
4-6
13.2
7-9
38.0
10-12
14.4
13-18
19-23
315
Ghana
IUD
Inject
ables
Steril
ization
13.6
9.5
87
3.6
75
1.6
45
6.3
6.2
0.0
0.0
3.1
54
14.4
19.9
22.3
21.6
23.2
237
29.7
19.0
135
65
19-23
32.3
Brazil
325
14.4
0-3
49.0
27.2
20.4
35.4
40.4
4-6
44.3
16 1
7-9
43.4
17.3
10-12
37 1
23.2
13-18
t
t
19-23
49.0
25.0
Colombia
2.6
9.2
13
87
85
55
11.9
10.8
0-3
6.8
4—6
65
11.2
285
15.1
6.9
4.1
24 9
10.7
9.3
7.6
7-9
18.9
28 3
13.2
4.6
14.6
10-12
29 2
18 1
13-18
29.7
16.6
12.1
45
19-23
4-6
63
7.7
Dominican Republic
7-9
0-3
29.2
68
13.1
10-12
6.3
4-r21 2
13-18
31 7
31
2.2
“..Cl
5.4
189
6.8
21.3
19-23
6 7”
26 7
230
Kenya
20 4
9.3”
26.8
13-18
0-3
2.6
5.0
39 3
19-23
16 1
7.9
4-6
109
6.7
3.1
Ecuador
7-9
71
71
5.4
3.6
10-12
9.7
0-3
5.8
4-6
105
15.2
12.1
12.3
13-18
4H
7-9
11.8
21 1
11.8
19-23
10.8
10 1
10-12
11 6
5.8
Liberia
10.5
18.5
12.3
13-18
0-3
78
14.0
19-23
17.3
5.2
4-6
11 7
Guatemala
7-9
36.4
147
0-3
0.0
10-12
4.9
7.2
9.8
13-18
14.1
47
6.6
7-9
19-23
10-12
6.3
3.1
Senegal
9.6
13-18
6.9
0.0
0-3
5.2
19-23
5.2
0.0
4-6
11.1
Mexico
7-9
10.6
6.5
0-3
23.8
2.0
6.8
10-12
4.9
13.9
25.0
6.0
4-6
13.0
13-18
47
13.1
97
19-23
7-9
15.3
10.1
20.7
11 9
3.1
10-12
125
Zimbabwe
13-18
14.8
5.3
14.5
15.3
75.0
0-3
4.3
19-23
11.0
17.3
28.6
62.2
51.3
Peru
7-9
52.5
0-3
10.5
10.5
10-12
44.5
4-6
3.5
105
13-18
52.3
7-9 _. ----- 14.8 ----------11.1
19-23
'Not calculated because less than 5% of women were using the
method t Includes 1.1%, 0.9%. 0.3% and 1.0% of implant users in 4-6.10-12.13-18 and 19-23 month intervals, respectively, tineludes 0.7,0.8 and 3.2% of implant users iin 7-9.13-18 and 19-23 month intervals, respectively. §Number of cases fewer tnan to
"Includes 1 1% and 0.3% of implant users for 10-12 and 13-18 month intervals, respectively.
International Family Planning Perspectives
a modem method before ed hormonal methods during the same
menses returned was month that menses returned was about
higher in Latin America one-fourth in Latin America and the
and the Caribbean and Caribbean and Asia, but was considerably
Method and country
Months since last birth
Asia <27—57'7, average of lower in the three Sub-Saharan African
0-3
4-6
7-9 10-12 13-18 19-23
42'1), than in Sub-Saha countries. The data in Table 5 clearly in
ran and \orth African dicate that, on average, at least half of the
Pill
7g
Botswana
27
5.1
27
90
countries (24—46'7. aver women initiated use of hormonal contra
53 7
Zimbabwe
38 5 46 1
22 8
33 3
48 0
age of 33'7). The per ceptives after the return of menses.
IUD
centage ot women who
34
82
7.9t 8 2t
9 9t
9 61
Indonesia
initiated contraceptive Reproductive Preferences
48
7.1
11.5
35
3 21
Thailand
10 6T
use during the same We classified the women's postpartum
Injectables
12.4
8.2 16.0
27 7
1.9
15.8
Indonesia
month menses returned preferences for future childbearing into
41 8
3.8 16.5
6.2
9.9
22.0
Thailand
(average of 18% > varied four categories: want no more children,
Sterilization
little among countries * want to space next pregnancy, want to
29 4 29.1
22 7
0.0
Brazil
26.1
16.3
74
In Latin America and have another child as soon as possible,
14.4
0.0
4.5
0.0
Colombia
39
11.4 32 3
Dominican Republic
16.1
9.3
15.3
226
the Caribbean and Asia. and undecided. The first two categories
Ecuador
64
79
0.0
70
39
2.3
the higher percentage of include women with the potential need for
73
5.4
00
Mexico
12 1
59
17 4
Thailand
82
14.6
13.1 28.3
253
women who initiated contraceptives.t
Trinidad & Tobago
69
182
0.0
contraceptive use before
Among women in need of contraceptives
184 17.9
Sri Lanka
81
8.1
23 0
21.6
the return of menses because they wanted no more children or
•Number ol cases fewer than110 f includes imoiant users
may be related to the wanted to space the next pregnancy, the
high use of sterilization. highest proportion of women were unex
Among unexposed women, IUD use was When the analysis was restricted to hor posed and not using a method (38%); the
negligible in all countries studied except monal methods—the pill, injectables and lowest proportion of women were unex
Thailand (3-12'1) and Indonesia (3-10'1). implants—onlv an average of one-fourth posed but using a contraceptive (991). Ap
Injectables were used by considerable of the women in these regions initiated proximately 32% of the women were ex
proportions of unexposed women in In contraceptive use before the return of posed and using contraceptives, and 21%
donesia and Thailand, and use increased menses. In Thailand and Mexico, only were exposed but not using a method.
with time. The proportion of unexposed about 1091 of the women initiated hor
Among women who were exposed to
women using injectables increased sharp monal contraceptive use before the first the risk of pregnancy but were not using
ly after nine months postpartum; at 19-23 menses. Of all the countries for which data a method, about 36% wanted no more
months postpartum, use of injectables by were available, Zimbabwe had the high children, about 36% wanted to space their
unexposed women exceeded proportions est proportion of women (46%) who ini next pregnancy, 16% wanted to have a
among exposed women.
tiated the use of hormonal methods before child as soon as possible, and 12% were
In the eight countries where sizable pro the return of menses.
undecided about their next pregnancy.
portions of unexposed women used ster
The proportion of women who initial Latin America and the Caribbean had the
ilization, they did so with equal (or great
er) frequency than exposed women.
Table 5. Among postpartum women exposed to the risk of pregnancy, percentage distribution
Table 5 shows the percentage of current
of current contraceptive users, by type of method and when use began, according to country
users of modern methods and hormonal
Country
Any modern method
Hormonal method onlj
Total
methods by whether they initiated use be
Before
Menses After
Before
Menses After
fore or after the return of menses. In those
menses
month
menses montn
menses
menses
countries for which data were available, the
Asia
42.4
24.8
24.4
18.1
39.2
50.8
100.0
percentage of tvomen who initiated using
Table 4. Among postpartum women unexposed to the risk ol preg
nancy, percentage using selected contraceptive methods, by coun
try, according to months since last birth
‘Among women who initialed contraceptive use in the
same month as menses resumed, >t was not possible to
determine which event occurred first. Some women may
have initiated contraceptive use before the resumption
.vhile others may have begun using contratCharles Westoff has made several refinements to the
ienti.il Demand for Family Planning: A New .Measure of
Lumet Need and Estimates for Five Latin American
14 45. IMSS). Because we defined the sample as not preg
nant. currently in union and having recently given birth.
manv ot the refinements suggested by Wesloh are imphritlv incorporated in our analysis of the contraceptive
no-dsol postpartum women. Two elements of Iherehned
measure that we did not include in our analysis are mislimed and unwanted last births among women who are
•hithth undereshm rted unmet needs
Mume 18. Number J September 1992
Indonesia
Sn Lanka
Thailand
Latin America'
Caribbean
Bolivia
Brazil
Colombia
Dominican
Republic
Ecuador
Guatemala
Mexico
Peru
Trinidad
& Tobago
Africa
Botswana
Kenya
Tunisia
Zimbabwe
38.0
52.3
36 8
22 0
8.4
24 0
39.2
39 3
39.2
29 6
35 3
96
28 4
94
35.3
42.0
55.3
55.0
100.0
100.0
100.0
41.5
39.9
57 4
34.6
17.9
11.2
14 7
24.4
40.6
48 9
27 9
41 0
22.9
33 2
40 3
189
23.8
152
20.6
26.4
53.4
51.7
39 1
54.7
100.0
100.0
1000
100.0
54.3
40.8
42.9
45 4
31.5
160
14.3
21.8
19.6
21 5
29 7
44 8
35 3
34.9
47 0
28 6
153
193
109
184
21 6
22.4
29 8
32 9
27 6
49.8
62.4
509
56.2
53.9
100.0
100.0
100.0
100.0
100.0
26.9
32.9
34.0
24.2
27.6
45.6
17.6
13.9
55 5
53.3
58.5
59 0
52.9
42 6
20.9
29.7
28.3
19.0
25 1
46.4
17.3
15.0
7.8
18.9
20.5
127
61.9
55.4
64.0
62 1
54.4
40.9
100.0
100.0
100.0
100.0
100.0
100.0
16.8
19.5
11.8
Note Any modern metnoa induces the Dili the IUD injectables imoiants bamer mernoas ana sterilization; hormonal methods include
tne Dili, injectables and .rnpiams Regional averages are unweighted, because of rouncmg. figures may not add to 100
Contraceptive Use Among Postpartum Women
Table 6. Percentage distribution ol postpartum women exposed to the risk of pregnancy and not using contraceptives, by fertility preferences.
according to country and months since last birth
Want
no more
children
Want to
space next
birth
Want
Chilo
soon
Undecided
Total
0-23
0 3
7-9
13-18
Sri Lanka
0-23
0-3
7-9
13-18
Thailand
0-23
0-3
7-9
13-18
28.0
17.6
31.1
24.0
48.6
48 6
45.2
51 8
96
W2
5.9
8.6
13 8
23 5
100.0
100 0
1000
100.0
34 0
51.7
9.2
5■
100.0
33.1
40.8
56.2
45 3
7.2
63
67
1000
100.0
41 4
35 9
11 5
11 2
100 0
29.9
40.5
53.8
37.2
178
45
100.0
100.0
0-23
74 1
143
8C
", p
100 0
7-9
57.9
73.7
13-18
Ri-azil
56 6
0-23
0-3
61 2
62.0
7-9
59.4
13-18
Colombia
60.7
0-23
0-3
67.5
7-9
48.2
13-18
Dominican Republic
39.7
0-23
28.6
O-3
57.9
7-9
430
13-18
Ecuador
56.0
57.9
0-3
56.7
7-9
13-18
51.5
Guatemala
35.1
0-23
0-3
7-9
31.6
33 1
13-18
Mexico
48.8
0-23
51.4
0-3
7-9
36.2
50.9
13-18
Peru
71.7
0-23
0-3
76.2
7-9
73.8
13-18
Trinidad & Tobago
0-23
51 5
24 7
16 1
7.9
28
100.0
100 0
186
23.5
11.3
15.3
20 4
14.2
19.9
76
125
54
100.0
100 0
100 0
100.0
26.0
11 1
2.2
100.0
27.8
26.4
47
195
00
59
100.0
100.0
35.4
54.2
28 6
31.5
20 3
7.6
126
22.5
96
09
30
100.0
100.0
100.0
100.0
30.2
42.1
33 3
29 4
0.0
67
10.3
61
00
3.3
88
100.0
100.0
100.0
100.0
344
137
168
100.0
50 0
35.9
158
14 5
26
’65
100.0
100.0
20.6
13.6
27 5
21.4
9.5
6.2
24 6
33
21 •
28 8
•i
24 4
100.0
1000
1000
100.0
100.0
Country and months
since last birth
ASIA
7-9
13-18
NORTH AFRICA
Egypt
0-23
0-3
15.6
17.3
8i
14.3
15.4
4.8
9.2
29
A7
16
100.0
100.0
26.3
12,1
10 1
100.0
56.3
40.7
25.0
33.3
12.5
11 1
6.2
149
100.0
100.0
42.1
39.2
21.6
38.1
19.2
12.4
17 1
10J
100 0
100 0
Country and months
smce last birth
Want
no more
children
Egypt (cont.)
40.3
7—9
43.1
13-18
Morocco
0-23
303
29 3
0-3
7-9
30.0
36 9
13-18
Tunisia
36.0
0-23
29.4
0-3
40.9
7-9
13-18
33.3
SUB SAHARAN AFRICA
Botswana
0-23
28.0
0-3
7-9
24.5
13-18
30.2
Burundi
227
0-23
92
0-3
9.1
13-18
24.0
Ghana
0-23
16.7
0-3
8.0
7-9
17.0
13-18
Want to
space next
birth
Want
child
soon
Undecided
Total
25.2
17.7
23.3
22.4
11.2
168
100.0
100.0
33.1
44.8
50.0
22.8
24.7
15.5
13.3
26.2
11.9
104
6.7
14 1
100.0
1000
100.0
100 0
34 7
500
386
34 4
91
25 8
59
6.5
100.0
100.0
100.0
100.0
21 3
40.6
195
11.9
100.0
37.7
43 4
37 7
160
0.0
10.4
100 0
100.0
51 4
90 8
84.5
48.0
20.2
0.0
5.5
20.4
00
0.9
7.6
100.0
100.0
100.0
100.0
54.7
21.3
7.3
100.0
80.0
56.6
8.0
22.6
4.0
3.8
100.0
100.0
Kenya
0-23
0-3
7-9
13-18
37.1
28.2
35.1
37 7
40.7
57.9
43.5
40.7
12.6
8.5
70
14.6
9.6
54
70
100.0
100.0
100.0
100.0
Liberia
0-23
0-3
7-9
13-18
12.9
16.9
7.3
16.6
31.7
33 7
33.8
31.6
38.3
39.8
39.4
30 6
17 1
9.6
19.5
21.2
100.0
100.0
100.0
100.0
Mali
0-23
0-3
7-9
13-18
14.4
0.0
18.8
13.4
31.7
46.2
40 0
32.9
40.7
39.6
32.0
37.1
13.2
14.2
9.2
16.6
100.0
100.0
100.0
100.0
Senegal
0-23
0-3
7-9
13-18
13.0
28.8
18.2
39.8
100.0
25.0
14.3
417
25.5
167
19.4
0.0
40.8
100.0
100.0
12.6
51.3
29.4
6.7
100.0
15.4
7.5
61 5
62.5
15.4
20.0
10’0
100.0
1000
12.4
22.6
17.3
7.7
318
38.7
28.9
49.3
35.2
40.7
56.3
3.9
8.4
3.3
7.1
100.0
100.0
100.0
100.0
27 0
37.8
27.0
8.2
100.0
32.8
36.1
19.7
9.8
100.0
Togo
0-23
0-3
7-9
13-18
Uganda
0-23
0-3
7—9
13-18
Zimbabwe
0-23
0-3
13-18
•Number of cases fewer than 10.
In contrast to the proportions of women
highest proportion of women who want
ed no more children (55%), and Sub-Sa who wanted no more children, the pro
haran Africa had the lowest proportion portions of women who wanted to space
(20%), with wide variation within each re their next pregnancy varied little among
gion. Among all countries surveyed, Peru regions. In Asia and Sub-Saharan Africa,
and Bolivia had the largest proportions an average of 40% of the women wanted
(over 70%) of women who wanted no to space their next pregnancy, while in
■ North Africa and Latin America and the
more children.
88
Caribbean the average was 28%. In all
countries except Egypt, Mexico, Peru and
Bolivia, at least one-fourth of the women
wanted to space their next pregnancy.
Regarding women who wanted to have
another child as soon as possible, there
was less intraregional variation in North
Africa and Asia than there was in the other
International Family Planning Perspectives
two regions. In Sub-Saharan Africa, the
proportions ranged from 12% to as high
as 49%, and in Latin America and the
Caribbean, the proportions ranged from
8% to 20%
Table 6 shows the potential demand for
contraceptives among women who were
exposed to the risk of pregnancy but were
not using a method, at three different
postpartum periods (0-3,7-9 and 13-18
months) as well as for the entire post
partum period (0-23 months). For space
considerations, the alternate time inter
vals (4-6,10-12 and 18-23) are not shown
in Table 6.
The difference between women's future
childbearing preferences in the immedi
ate postpartum period (0-3 months) and
in subsequent periods is of particular in
terest. Data in the first column of Table 6
indicate that in the majority of the coun
tries, women in the immediate postpar
tum period were not necessarily more
likely to want no more children than
women interviewed during later post
partum months. In fact, the proportion of
women in subsequent periods (7-9 and
13-18 months postpartum) who indicat
ed that they wanted no more children was
higher than or equal to that of women in
the immediate postpartum period among
nine of the 13 countries for which data
were available. Further, even in the two
countries, Liberia and Mexico, where the
proportion of women who wanted no
more children decreased in the 7-9 month
period, the proportion increased in later
months. In only two countries, Uganda
and Ecuador, were the proportions of
women who wanted no more children
consistently lower in later months than in
the immediate postpartum period
These findings challenge the notion that
women's motivation to control fertility is
highest during the immediate postpartum
period, at least among postpartum women
who wanted no more children. However,
the differences observed in women's mo
tivation at different points postpartum
could reflect differences in the composi
tions of the populations. For example,
women who wanted to have manv chil
dren might not have been included in the
13-18 month postpartum cohort because
they had already become pregnant again.
A contrasting pattern emerged among
women who wanted to space the next
pregnancy (Table 6, second column). In
nine of the 13 countries for which data
were available, the proportion of women
in the immediate postpartum period who
indicated that they wanted to space their
pregnancy was consistently higher than
Volume 18, Number 3, September 1992
postpartum family planning services is
whether the woman gave birth in a hos
pital: The higher the proportion of women
delivering in hospitals, the greater the
number of potential family planning ac
Delivery
Postnatal’
Country
Prenatal
ceptors. However, if a hospital-based pro
88.1
66.6
83.8
Asia
gram does not give women any informa
j
u
38
7
Indonesia
tion or counseling on family planning
94 1
88.4
34 1
Sri Lanka
options before delivery, a postpartum pro
72 1
82J
83.5
Thailand
gram may not be acceptable or effective.
Latin America
If information or counseling on postpar
58.5
64.7
69.1
Caribbean
46.4
55.9
43.9
Bolivia
tum contraception were provided prena□
77.7
39.5
Brazil
tally, a hospital-based program could be
72 4
80.7
83 1
Colombia
very effective. Therefore, the frequency of
Dominican
15.9
90.3
97.3
Republic
contact between a woman and a health
u
63.9
Ecuador
75.5
care provider during the prenatal period
48.5
275
39 9
Guatemala
u
70 8
74.2
Mexico
is a second indicator of the opportunity to
63.9
49.9
580
Peru
offer postpartum contraceptive services.
Trinidad
The third opportunity to provide contra
76.2
98.9
99 1
& Tobago
ceptives is during the postnatal period,
75.3
53.2
45.1
North Africa
when child immunization can be used as
62 0
62 2
37.5
Egypt
74 2
27.5
26.5
Morocco
a proxy for a woman's access to health
70 4
89 8
70.9
Tunisia
care providers.
Sub-Saharan
Table 7 includes data on women's con
56.7
43.1
75.8
Africa
tact with health care providers during
76.0
91.5
93.1
Botswana
584
three phases—the prenatal period, at de
17.0
66.9
Burundi
50.9
40.0
Ghana
86 3
livery and the postnatal period. On aver
83.0
Kenya
92 8
50.3
age, about 70% of the women reported
423
84.0
55.8
Liberia
16.7
18.1
25.6
Mali
having had contact with health care
20.7
35.3
44.5
Senegal
providers during the prenatal period. In
u
44.1
Togo
82.0
12 countries, at least 80% of the women
43.2
Uganda
90 9
37.8
71.4
89.3
92.0
Zimbabwe
had had contact prenatally. In only a few
•Refers to at least one immunization (tuberculosis, polo, diphtheria countries—Senegal, Mali, Brazil, Moroc
or measles) for children G-23 months old at tne time of the sur co and Guatemala—the proportion of
vey. Note u=unavailable. Regional averages are unweighted
women who had had prenatal contact with
health care providers was less than 50%.
The proportions of women who had had
or equal to the proportion of women in
subsequent months who wanted to space contact with health care providers during
their next pregnancy. Mexico was the only delivery were high in the Asian and Latin
country in which the proportion of wom American samples. In these two regions,
en who wanted to space their pregnancy about two-thirds of the women had de
was lower at 0-3 months than in subse livered in hospitals, and in only a few
quent months. In two countries, Uganda countries—Bolivia, Guatemala and In
and Morocco, the proportion of women donesia—less than 50% of the women had
at 7-9 months who wanted to space their done so. Although the overall percentage
next pregnancy was higher than the pro of women who had had contact with
portion of women at 0-3 months, but health care providers during delivery was
even in these countries the proportion de lower in Sub-Saharan Africa and North
creased at 13-18 months. Indonesia was Africa, a few countries—Zimbabwe, Bot
the only country in which the proportion swana, and Tunisia—had percentages sim
of women who wanted to space their next ilar to those in the other two regions. At the
pregnancy was the highest at 13-18 same time, two countries in the Sub-Sa
months. The data suggest that women haran region—Burundi and Mali—had the
who wanted to space their next preg smallest proportion of women (less than
nancy were more motivated to control 20%) reporting hospital deliveries.
In a majority of the countries, the pro
their fertility in the immediate postpar
tum period than were women who want portion of women who had had contact
with health care providers during the
ed no more children.
postnatal period was generally lower than
that of the prenatal period. The proportion
Contact with Health Care Providers
One indicator of a postpartum woman's of women who had had postnatal contact
contact with trained health care providers was particularly low (less than 20%) in
that may be used to develop or expand Mali and the Dominican Republic.
Table 7. Percentage ol postpartum women
who received perinatal care from trained
health care providers, by type of care, ac
cording to country
89
Contraceptive Use Among Postpartum Women
The number of prenatal and postnatal
contacts with trained health care providers
for each woman was unknown. Moreover,
the same women were not necessarily rep
resented in all three contact categories; it
is quite probable that some women had
had only one of the three types of contact,
while others may have had two or three.
In general, however, the data in Table 7
show that the proportions of postpartum
women who had had contact with trained
health care personnel in the prenatal or
postnatal periods were not necessarily
lower than the proportions who had had
contact during delivery. In fact, in sever
al countries, the proportions of women
who had had prenatal and postnatal con
tact were higher than the proportion of
women who had had contact during de
livery. These results challenge the notion
that women are more likely to have con
tact with health care providers at the time
of delivery than before or after it.
Discussion
The DHS data analyzed in this paper offer
a comparative perspective on the contra
ceptive behavior of postpartum women
from a number of socioeconomically and
culturally diverse countries. These data
support some previous research findings
and provide empirical evidence for pre
vious hypotheses.
Although a relatively large number of
countries included in the sample are from
Latin America and the Caribbean and
Sub-Saharan African regions, onlv three
countries each are from North Africa and
from Asia. The countries in the latter two
regions do not necessarily represent their
regions; for example, all three Asian coun
tries included have experienced rapid fer
tility decline. The sample excludes large
countries, such as India and Pakistan, that
have lower rates of contraceptive preva
lence. Hence, the regional averages, par
ticularly for these two regions, should not
be considered representative.
Very little direct information on attitudes
regarding postpartum contraception was
available in the data analyzed. No direct
information was available on when dur
ing the postpartum period women pre
ferred to use contraceptives and what
methods they preferred using. Instead,
most of the analysis focused on behavior.
The contraceptive behavior observed prob
ably reflects both the family planning pro
gram characteristics and the women's ac
tual preferences. However, because of a
lack of attitudinal data, we cannot deter
mine to what extent behavior was influ
enced by the program characteristics and
90
by women's choices. Winikoff and Mensch
cautioned against making inferences about
"underlying desires" regarding postpar
tum contraception based on the "analysis
of.behavior "'"The World Health Organi
zation's Human Reproduction Program
hopes to fill this gap by supporting post
partum modules in several future Demo
graphic and Health Surveys."'
The focus on current contraceptive use
mav have resulted in misclassification of
some women. The use of injectables, for
example, leads to amenorrhea among a
significant proportion of the users.1''Clas
sifying as amenorrheic women who were
using injectables would exaggerate the
proportion of unexposed women using
contraceptives. However, the potential ef
fect of this kind of misclassification would
be small, since injectable contraceptives
are used by a significant proportion of
women in only a few countries.
Another potential source of misclassifi
cation involves the separation of women
into exposed and unexposed categories.
Current menstrual status is used as one of
the criteria for this classification, but it is
not always a reliable indicator of fertility,
especially when women have difficulty dis
tinguishing between the end of lochia
(bleeding after deliver)') and the return of
menstrual bleeding. This confusion could
result in an overestimation of the propor
tion of exposed women. On the other hand,
many women who experience resumption
of menses, especially within six months
after delivery, have anovulatory cycles.'"
Hence, the resumption of menses does not
necessarily imply fertility-, especially when
menses returns within six months post
partum. However, the average duration of
amenorrhea was generally twice that
length in several of the countries studied.
Endocrinological studies have shown
that ovulation can occur prior to men
struation in both breastfeeding and bottle
feeding mothers.20 Generally, though, the
earlier a woman experiences menses, the
less likely that the bleeding has been pre
ceded by ovulation, and the earlier the first
ovulation occurs, the less likely that the
woman will have had sufficient proges
terone production to sustain a pregnancy.2'
If, as suggested, menstrual status after
the first six months postpartum is a more
reliable indicator of fertility than before
the first six months postpartum,22 some
women, especially those in the immedi
ate postpartum periods, may not really be
exposed to the risk of pregnancy. One way
to refine the definition of exposed women
might be to include only those who have
resumed menses, are at least six months
postpartum and are not abstaining. This
classification method would have reduced
the proportions of exposed women shown
in Tables 1 and 2; in view of this, the pro
portions reported in the tables should be
considered upper-bound estimates. For
this study, such estimates were desirable,
in order to determine the maximum pro
portion of unexposed women who were
using contraceptives.
The status of postpartum women with
regard to exposure to the risk of pregnan
cy is affected by the time since childbirth:
Lnexposed women tend to have shorter
average times since delivery than exposed
women. Contraceptive use and method
choice mav be affected by this time factor.
The potential effect of any confounding
may be minimized by the application of
life-table techniques; however, previous
research has shown that method choice is
not affected by the length of time since
delivery.2' Further, in our analysis, data
on contraceptive use were examined by
time since childbirth, comparing exposed
groups with unexposed groups. The re
sults showed that the patterns of contra
ceptive use were different between ex
posed women and unexposed women,
regardless of the time since childbirth.
The determinants of postpartum con
traceptive behavior were not examined in
our analysis. Women's socioeconomic sta
tus and family life-cycle stage are no
doubt important covariates affecting con
traceptive choice and behavior. Equally
important are the effects of the family
planning program's characteristics and
policies, which require analysis of differ
ent types of information than standard
survey data.
Conclusions
Despite some limitations of the data and
analysis, our findings about the patterns
of contraceptive use among postpartum
women in developing countries have im
plications for program design and imple
mentation. First and foremost, the data
clearly show that reversible modem meth
ods of contraception are not used indis
criminately by unexposed women. By and
large, we found that the women using
postpartum contraceptives were those
who were exposed to the risk of preg
nancy. At the same time, the results also
suggest that in some countries refinements
in the delivery of postpartum contracep
tive services may be desirable. It is not
generally well recognized that relatively
few women need contraceptive protection
soon after a childbirth. In Zimbabwe and
Botswana, for example, service delivery
International Family Planning Perspectives
could be refined to minimize double pro
tection among unexposed women.
At least half of the postpartum women
appeared to initiate hormonal contracep
tion after the resumption of menses. Our re
sults suggest that the previous findings
from Mexico24 and Thailand25 are general
ly applicable to a large number of countnes.
In a few countries, however, program re
finements may be needed with respect to
distribution of the pill to unexposed
women. At least one-third ot the postpar
tum women in Zimbabwe, Brazil, Bolivia
and Sri Lanka initiated use of hormonal
methods long before the resumption of
menses. Use of the pill by unexposed
women results in double contraceptive pro
tection during lactational intertill tv and may
produce a shorter interval between preg
nancies than if the two kinds of protection
were sequenced instead of simultaneous.
An average of only about one-fifth of
the postpartum women in the countries
studied were exposed to the risk of preg
nancy within two vears of giving birth as
a result of not using contraceptives. Our
results on women who want no more chil
dren challenge the notion that motivation
to use contraceptives is highest in the im
mediate postpartum period. However, a
larger proportion of women in the imme
diate postpartum period than in the later
postpartum period wanted to space their
next pregnancy, presumably because
some women in the later postpartum
months would like to become pregnant.
Our results also question the notion that
delivery is the only time women come in
contact with health care providers. The
proportion of women who had been in
contact with providers at least once in a
prenatal or postnatal period was equal to
or higher than the proportion who had
had contact during hospital deliveries.
Contact during the prenatal period pro
vides an opportunity to inform and coun
sel women about their postpartum con
traceptive options before they encounter
health care providers for delivery. Prena
tal counseling is critical for the adoption
of permanent contraceptives and could
help prevent postoperative regret over
sterilization;2'* it is also important for the
immediate postplacental insertion of
ILDs, in that it allows a woman to make
a well-informed decision prior to labor,
under less stressful circumstances.2’ Sim
ply pul. a good postpartum program
should begin prenatallv.
Although the medical safetv of tubal
sterilization, particularly if performed
within seven days of the delivery, has been
established,■" the extent to which pro:
Volume IK Number 1 September 1992
grams can accommodate immediate tubal
sterilization is not clear. In many devel
oping countries, where the maternity hos
pitals and facilities are limited, the de
mand for this sen-ice can easily exceed the
supplv. Furthermore, it may be medical
ly unsafe to perform tubal sterilization on
the deliven- table. Access to a separate op
erating theater adjacent to the delivery
room is required for tubal sterilization. As
for IUDs, although it is technically safe to
insert the devices during the immediate
postpartum period,2" expulsions are
known to be more frequent in the imme
diate postpartum stage than at later
stages. "' Moreover, as with sterilization,
it may not be feasible to insert IUDs at the
time of deliverv
Hormonal implantsand injectablesare
normally not advised for use in the im
mediate postpartum period; their medi
cal safety, especially- the effect on lactation
and the infant, has not vet been establish
ed. Even if their medical safetv is estab
lished, these methods will have little pro
grammatic impact because of double
protection in countries where lactational
amenorrhea is long. Similarly, even
though the minipill has been established
as medically safe for use while breast
feeding,31 its use in the immediate post
partum period may not be desirable from
a programmatic perspective, as shown by
the data for Zimbabwe and Bangladesh.
The immediate use of barrier methods, al
though known to be medically safe for
women in the immediate postpartum pe
riod, will also have little demographic im
pact if they are providing redundant pro
tection during amenorrhea. Double
protection is not currently a widespread
problem, however.
One important behavioral factor that in
teracts with all aspects of postpartum con
traceptive use, and one that is gaining ap
preciation, is breastfeeding. Efforts to
improve the postpartum delivery of fam
ily planning services may be most suc
cessful if providers fully integrate it with
breastfeeding.’2 Even when the postpar
tum period is considered two years long,
about 8O‘< of the need for postpartum
family planning is met without contra
ceptives, through lactational amenorrhea
or abstinence. In addition, during this
time, women have prenatal and postna
tal contacts with health care providers,
who can use these opportunities to deliv
er family planning information, counsel
ing and services. These findings suggest
that a different programmatic approach
than has been used in previous postpar
tum contraceptive service delivery efforts
may be needed to advise women intelli
gently and sensibly when to initiate con
traceptive use and which method to use.
References
1. R G. Castadot et al., “The International Postpartum
2. B. S. Atkins, ed., Maternal and Child Health Family Plan
ning Program. proceedings from a 1979 technical work
shop, The Population Council. New York. 1980.
3. I-C. Chi. D. Gates and S. Thapa. "Performing Tubal
Sterilization During Women's Postpartum Hospitaliza
tion—The U.S. and International Experience/' Obstet
rics and Gynecological Survey.47:71.1992; l-C. Chi and G.
Farr, “Postpartum IUD Contraception—A Review of an
International Experience," Advances in Contraception.
5:1271989; and J. Guillebaud. "Contraception After Preg
nane)’." The British Journal oiFamily Planning.Supplement.
16:10. 1991.
ontraception , 39:477,1989; and
5.
B. Winikoff and B Mensch, "Rethinking Postpartum
Family Planning." Studies in Family Planning. 22:294,1991.
6.
Ibid.
7.
J Knodel and X. Chayonan, "Contraceptive Initiation
Patterns in Thailand," Health and Population Studies Based
on the 1987 Thailand Demographic and Health Survey. In
stitute of Population Studies, Chulalongkorn Universi
ty, Thailand, 1989; B. Winikoff, M. A Castle and V. H.
Laukaran, Feeding Infants in Four Societies: Causes and Con
sequences of Mothers' Choices. Greenwood Press, New York,
1988, J. Gomez de Leon and J. Potter, "Modelling the In
verse Association Between Breastfeeding and Contra"Breastfeeding and Contraception. Why the Inverse As
sociation?" Studies in Family Planning, 16:61,1985; and
During Lactation in Developing Countries,"
9.
) Bongaarts. "The Proximate Determinants of Natu
ral Fertility." in R. A Bulataoand R.D. Lee, eds.. Deter
minants o' Fertility in Developing Countries. Vol 1, Academic
Press, New York. 1983. p 103.
10.
B. Winikoff. M. A. Castle and V. H. Laukaran, 1988.
op. cit. (see reference 7); and J. Knodel and X. Chayonan.
1989, op. cit. (see reference 7).
11.
Ibid
12.
R. V. Short et al.. "Contraceptive Effects of Extended
Lactational Amenorrhea: Beyond the Bellagio Consen
sus," Lancet., 337:715.1991; and S. Diaz et al.. "Contra
ceptive Efficacy of Breastfeeding in Urban Chilean
Women." Contraception. 43:335 1991.
13.
K. I. Kennedy and C. M. Visness. "Contraceptive Ef
ficacy of Lactational Amenorrhea," Lancet. 339:2271992.
14.
S. Bhatia and Y. Kim, "Oral Contraception in
Bangladesh," Studies in Family Planning. 25:2331984; and
S. Bhatia, S. Becker and Y. Kim, "The Effect of Oral Con
traceptive Acceptance on Fertility in the Postpartum Pe
riod," International Journal of Gynaecology and Obstetrics,
Supplement, 25:11,1987.
15.
D. Adamchak and M. Mbizvo, ‘'The Relationship Be-
Contraceptive Use Among Postpartum Women
m Fertility and Contraceptive Prevalence in Zimbab
we." lHtcnmtiOH.il Fiimilu Planning Per<fkxtr.vf 16:103 W1
31. P. W. Howie. "The Progestogen-Only Pill.' British
lournal of Obstetric* and Gynaecology. 92:1001.1985.
16. B. Winikoff and B. Mensch. 1991, op cit (see refer-
32. K. I. Kennedy. R. Rivera and AS. McNeilly, 1989, op
cit. (see reference4»; M. Labbok et al.. Guidelines for Breast'ceding in Family Planning and Child Survival Programs. In
stitute for International Studies in Natural Family Plan
ning Washington D.C., 1990; and B. Winikoff and B.
Mensch. 1991, op cit. (see reference 5i.
17. X. Williamson, personal communication. 1991.
18. iniectable Hormonal Contraceptives Technical and Safe
ty Aspects. Publication No. 65. World Health Organiza
tion. Geneva. 1984.
25. J. Knodel and 1’. Kamnuansilpa. "Contracepnon After
Childbirth in Thailand," lournal ofBiOfwcial Science. 20:312.
1988.
26. 1-C. Chi, D. Gates and S. Thapa. 1992, op. cit. (see ref
erence 3).
27. J-C. Chi and G Farr, 1989, op. cit. (see reference 3).
28. I-C. Chi, D Gates and S. Thapa, 1992. op. cit. (see ref
erence 3).
29. 1-C. Chi and G. Farr. 1989, op. cit (see reference 3).
92
Resumen
Estc estudio realizado a mujeres durante el
periodo de postparto sc base en datos recogidos por las Encuestos de Demograffa y Salud
(EDS) realizadas en 25 poises en desarroilo y
revela que la proporcidn de mujeres expueslas al riesgo del embarazo durante el periodo
de los dos aiios posteriores al parto, varia entre
tin tercio en el Africa Subsahariana a aproximadamente dos tercios en America Latina y
el Caribe. Mas de la mitad de las mujeres que
sc encuentran en el periodo postparto, actuahnente usan anticonceptivos. Las mujeres
expuestas al riesgo del embarazo son mas pro
clives que las mujeres no expuestas a utilizar
metodos anticonceptivos reversibles, generalmente la ptldora. Entre las mujeres que no
estdn expuestas al riesgo del embarazo debido
a la abstinencia oala amenorrea relacionada
con la lactancia, el 19% utiliza un metodo anticonceptivo, generalmente la esterilizacidn.
La proporcidn de usuarias que comienzan a
utilizar un metodo moderno antes del regreso de la menstruation varia entre el 27% y
57% en los poises de America Latina y el
CaribeyAsia,ydel24% n/46% enlospaises
africanos. Es menor la proporcidn de las usua
rias de anticonceptivos hormonalcs que inician su uso antes de la reanudacidn del perio
do menstrual. Aproxnnadamente una quinta
parte de las mujeres expuestas no utilizan.
ningiin metodo anticoncq’tivo. De estc grupoi
mas de la tcrccra parte no desemi tener mds
hijos y otro tanto quieren espaciar su proxi
mo embarazo.
Resume
Cette etude de femmes post-partum. basee sur
des donnees /trees des Eiiquetcs demographiques et de sonic dans 25 paysen developpement, a rcveleque la proportion de femmes ex
posees an risque degrossesse dans les deux ans
suivant I'accouchement varie entre tut tiers
en Afrique subsaharienne et pres des deux tiers
en Amcriquc latinc et dans les Antilles. Plus
de la moitic des femmes post-partum font couramment usage de contraceptifs. Les femmes
exposees au risque degrossessesont plus susceptibles que les femmes non exposees d'utiliserdes methodes reversibles, generalement la
pilule. Parmi les femmes non exposees au
risque de grossesse par suite d'abstinence on
d'amenorrhee lieed I’allaitement, 19% utilisent une methode contraceptive, generale
ment la sterilisation. La proportion d'utilisatrices de contraceptifs qui commencent a
utiliser une methode moderne avant la reprise
des menstruations varie entre 27% et 57%
parmi les pays d'Amerique latine, des Antilles
et d'Asie, et entre 24% et 46% parmi les pays
africains. Des proportions plus foibles d'utili-1
satrices de contraceptifs hormonaux commencent a utiliser ceux-ci avant la reprise des
menstruations. Environ 20% des femmes ex
posees n'utilisent aucun moyen contraceptif.
Au sein de cegroupe, plus du tiers ne veulent
plus d'enfants et un autre tiers desirent re
tarder la grossesse subsequentc.
International Family Planning Perspectives
Reprints of
Scientific Articles
GOH -3/V.
References
* Abdel-Tawab N, Huntington D, Hassan EO, et al. Effects of husband involvement on postabortion
patients’ recovery and use of contraception in Egypt. In Huntington D, Piet-Pelon NJ, eds. Post
abortion Care: Lessons from Operations Research. New York: Population Council, 1999.
* Chi I, Wilkens L,Champion CB, et al. Insertional pain and other IUD insertion-related rare events
for breastfeeding and non-breastfeeding women - a decade's experience in developing countries.
Adv Conf 1989;5(2): 101-19.
Chi I, Wilkens L, Rogers S. Expulsions in immediate postpartum insertions of Lippes Loop D and
Copper T IUDs and their counterpart delta devices - an epidemiological analysis. Contraception
1985;32(2): 119-34.
* Diaz J, Loayza M, Tones de Yepez Y, et al. Improving the quality of services and contraceptive
acceptance in the postabortion period in three public-sector hospitals in Bolivia. In Huntington D,
Piet-Pelon NJ, eds. Postabortion Care: Lessons from Operations Research. New York: Population
Council, 1999.
Farr G, Rivera R. Interactions between intrauterine contraceptive device use and breast-feeding
status at time of intrauterine contraceptive device insertion: analysis of TCu-380A acceptors in
developing countries. Am J Obstet Gynecol 1992; 167(1): 144-51.
Foreit K, Foreit J, Lagos G, et al. Effectiveness and cost-effectiveness of postpartum IUD insertion
in Lima, Peru. Int Fam Plann Perspect 1993; 19( 1): 19-24.
Hatcher RA, Trussell J, Stewart F, et al. Contraceptive Technology, Seventeenth Edition. (New
York: Ardent Media, Inc., 1998)800-01.
Hobcraft J. Child spacing and child mortality. In Demographic and Health Surveys World Confer
ence Proceedings, Washington, DC, August 5-7,1991, Volume 2. Columbia, MD: Macro Interna
tional, Inc., 1991.
Institute for Reproductive Health. Guidelines for Breastfeeding and the Lactational Amenorrhea
Method, Third Edition. Washington: Georgetown University, 1994.
International Planned Parenthood Federation. New IPPF statement of breastfeeding, fertility and
post-partum contraception. IPPF Med Bull 1990;24(2):3.
Langer A, Garvia-Barrios C, Heimburger A, et al. Improving postabortion care with limited
resources in a public hospital in Oaxaca, Mexico. In Huntington D, Piet-Pelon NJ, eds.
Postabortion Care: Lessons from Operations Research. New York: Population Council, 1999.
Contraception After Pregnancy
Ratnam SS, Tow SH. Translocation of the loop. In Zatuchni, GI, ed. Postpartum Family Planning:
A Report on the International Program. (New York: McGraw-Hill, 1970)371.
Solo J, Billings D, Aloo-Obunga C, et al. Creating linkages between incomplete abortion treatment
and family planning services in Kenya. Stud Fam Plann 1999;30(l): 17-27.
Technical Guidance Working Group. Recommendations for Updating Selected Practices in Con
traceptive Use: Results of a Technical Meeting, Volume I. Washington: U.S. Agency for Interna
tional Development, 1994.
* Technical Guidance/Competence Working Group. Recommendations for Updating Selected
Practices in Contraceptive Use: Results of a Technical Meeting, Volume II. Washington: U.S.
Agency for International Development, 1997.
* Thapa S, Kumar S, Cushing J, et al. Contraceptive use among postpartum women: recent
patterns and programmatic implications. IntFam Plann Perspect 1992; 18(3):83.
* World Health Organization. Improving Access to Quality Care in Family Planning. Medical
Eligibility Criteria for Contraceptive Use. Geneva: World Health Organization, 1996.
* World Health Organization Collaborative Study Team on the Role of Breastfeeding on the Preven
tion of Infant Mortality. Effect of breastfeeding on infant and child mortality due to infectious
diseases in less developed countries: a pooled analysis. Lancet 2000;355(9202):451-55.
World Health Organization. Abortion: A Tabulation ofAvailable Data on the Frequency and
Mortality of Unsafe Abortion, Second Edition. Geneva: World Health Organization, 1993.
World Health Organization. Post-abortion Family Planning: A Practical Guide for Programme
Managers. (Geneva: World Health Organization, 1997)1-72.
Items included in reprints section of this binder.
Contraception After Pregnancy
2
References and
Resources
Evaluation
Evalua
H-5A.G
Pretest/Posttest Questionnaire
and Participant Evaluation
Instructions for Presenters
Pretest and Posttest Questionnaire
This questionnaire should be given before the training session to ascertain knowledge level of the
audience. This same instrument should also be given afterward to assess how much the audience
learned from the presentation.
The following steps are recommended when giving the pretest:
clarify any terms that may not be familiar to the participants
do not tell the participants that there will be a test again after the presentation
(to avoid biasing the results of the posttest)
remain in the room during the test
ask the participants to fill out the questionnaires individually
check responses against answer sheet.
Participant Evaluation
Ask audience members to complete a form after each training session. The comments from the audience
may be useful in planning future presentations.
We would appreciate your sending to us any suggestions for improving the module based on comments
from the participants, or your experience as a presenter. In addition, we are interested in knowing
whether you found the pretest/posttest to be useful. Please return completed evaluation forms, pretest/
posttest results and other comments to:
CTU Modules Project Administrator
Family Health International
P.O. Box 13950
Research Triangle Park, NC 27709 USA
C
(0^?q j
\ f
Contraception after Pregnancy:
Questionnaire
Participant number (if applicable):
Occupation (check one): Physician
Policy-maker
1.
Student
Age:
Other provider
Sex:
Program manager
Other (please specify)
The following statements relate to contraception after pregnancy. Please indicate whether each
statement is True (T) or False (F).
T
F All hormonal contraceptives have a negative effect on milk production.
T
F All nonhormonal contraceptive methods can be initiated immediately after childbirth.
T
F Lactational amenorrhea method (LAM) provides more than 98 percent protection from
pregnancy.
T
F There is no difference in the technique for inserting an IUD immediately after delivery
and inserting it six months later.
T
F Children are more likely to survive if bom two or more years after the preceding birth.
T
F Nonbreastfeeding women postpartum should wait three months before beginning
contraception.
T
F All contraceptive methods can be started immediately after uncomplicated first- or
second-trimester abortions.
T
F IUDs should not be inserted after abortion in cases where infection is suspected.
T
F IUDs should never be inserted earlier than four weeks postpartum.
T
F Sterilization is safe to perform immediately after abortion, even if the woman is
hemorrhaging.
T
F Postabortion counseling increases contraceptive use.
2.
LAM is considered to be a very effective method of contraception for up to:
three months postpartum
six months postpartum
nine months postpartum
12 months postpartum
3.
IUD insertion can be performed... Check (a/) all that apply.
immediately after delivery of the placenta
within the first 48 hours postpartum
within the first 72 hours postpartum
within the first week after delivery
six weeks after delivery
at any time
4.
IUD expulsion rates are:
higher for postpartum insertions
_ higher for interval insertions
the same for both postpartum and interval insertions
5.
The best time to perform sterilization postpartum is:
immediately after delivery
48 hours after delivery
one week after delivery
four weeks after delivery
6.
Male sterilization is considered effective:
immediately after procedure
four weeks after procedure
six weeks after procedure
12 weeks after procedure
Questionnaire, page 2
1.
Contraceptive methods that can be initiated by breastfeeding women immediately after delivery
include... Check (V) all that apply.
condoms
diaphragms
periodic abstinence
DMPA
IUDs
Combined oral contraceptives (COCs)
female sterilization
8.
Contraceptive methods that should not be initiated by nonbreastfeeding women immediately after
delivery include... Check (a/) all that apply.
condoms
diaphragms
periodic abstinence
DMPA
IUDs
Combined oral contraceptives (COCs)
female sterilization
9.
Fertility will most likely return earliest for:
nonbreastfeeding women postpartum
breastfeeding women postpartum
women postabortion
10.
Diaphragms and cervical caps can be fitted immediately for:
nonbreastfeeding women postpartum
breastfeeding women postpartum
women after second-trimester abortion
all of the above
none of the above
Questionnaire, page 3
11.
Please circle A or D to indicate if you agree (A) or disagree (D) with the following statements.
A D During childbirth is a good time to talk to a woman about sterilization.
A
D Prenatal visits provide good opportunities for talking with women about family planning.
ADA woman should be encouraged to discontinue breastfeeding three months after giving
birth to enable her to use a hormonal contraceptive method.
A D Health-care providers should encourage women to fully breastfeed.
A
D HIV-positive women should never breastfeed.
Questionnaire, page 4
Contraception after Pregnancy:
Questionnaire Answer Key
1. The following statements relate to contraception after pregnancy. Please indicate whether each
statement is True (T) or False (F).
T © All hormonal contraceptives have a negative effect on milk production.
T © All nonhormonal contraceptive methods can be initiated immediately after childbirth.
© F Lactational amenorrhea method (LAM) provides more than 98 percent protection from
pregnancy.
T © There is no difference in the technique for inserting an IUD immediately after delivery
and inserting it six months later.
© F Children are more likely to survive if bom two or more years after the preceding birth.
T © Nonbreastfeeding women postpartum should wait three months before beginning
contraception.
T © All contraceptive methods can be started immediately after uncomplicated first- or
second-trimester abortions.
© F IUDs should not be inserted after abortion in cases where infection is suspected.
T © IUDs should never be inserted earlier than four weeks postpartum.
T © Sterilization is safe to perform immediately after abortion, even if the woman is
hemorrhaging.
© F Postabortion counseling increases contraceptive use.
2.
LAM is considered to be a very effective method of contraception for up to:
three months postpartum
T six months postpartum
nine months postpartum
12 months postpartum
3.
IUD insertion can be performed... Check ("V) all that apply.
y
immediately after delivery of the placenta
V
within the first 48 hours postpartum
within the first 72 hours postpartum
within the first week after delivery
V
six weeks after delivery
at any time
4.
IUD expulsion rates are:
\
higher for postpartum insertions
higher for interval insertions
the same for both postpartum and interval insertions
5.
The best time to perform sterilization postpartum is:
.
immediately after delivery
48 hours after delivery
one week after delivery
four weeks after delivery
6.
Male sterilization is considered effective:
immediately after procedure
four weeks after procedure
six weeks after procedure
T 12 weeks after procedure
Questionnaire, page 2
Contraceptive methods that can be initiated by breastfeeding women immediately after delivery
include... Check (^) all that apply.
\
condoms
diaphragms
periodic abstinence
A/
DMPA
IUDs
Combined oral contraceptives (COCs)
3: female sterilization
8. Contraceptive methods that should not be initiated by nonbreastfeeding women immediately after
delivery include... Check ("V) all that apply.
condoms
3: diaphragms
periodic abstinence
DMPA
IUDs
y
Combined oral contraceptives (COCs)
female sterilization
9. Fertility will most likely return earliest for:
nonbreastfeeding women postpartum
breastfeeding women postpartum
y
women postabortion
10. Diaphragms and cervical caps can be fitted immediately for:
nonbreastfeeding women postpartum
breastfeeding women postpartum
women after second-trimester abortion
all of the above
V
none of the above
Questionnaire, page 3
11. Please circle A or D to indicate if you agree (A) or disagree (D) with the following statements.
A (B) During childbirth is a good time to talk to a woman about sterilization.
(a) D Prenatal visits provide good opportunities for talking with women about family planning.
A (d) A woman should be encouraged to discontinue breastfeeding three months after giving
birth to enable her to use a hormonal contraceptive method.
(A) D Health-care providers should encourage women to fully breastfeed.
A ® HIV-positive women should never breastfeed.
Questionnaire, page 4
Participant Evaluation
C
Contraception after Pregnancy
Please answer the questions below after the Contraceptive Technology Update presentation on
Contraception after Pregnancy. Family Health International will use this information to help
improve future presentations.
Please tell us a little about yourself.
Name (optional)
__
Address (optional)
Phone (optional) Fax (optional) E-mail (optional)
What are your current job responsibilities? (Please mark all that apply.)
policy-maker
program manager
practicing doctor
practicing nurse/nurse midwife
nonmedical clinic personnel
other (specify)
community health worker
medical faculty
student (medical, nursing, midwifery)
health educator
pharmacist
Please tell us what you think about the presentation.
Did the presentation address what you consider to be the most important programmatic and
clinical issues on the topic? (Please mark one box.)
Q Yes
0 No
What, if any, additional information do you think should have been covered? (Please specify.)
What, if any, part of the presentation should have been excluded? (Please specify.)
How familiar were you with the information in the module prior to this presentation?
(Please mark one box.)
| | Very familiar
| | Somewhat familiar
| | Not at all familiar
Which two presentation messages do you think will be the most useful to you?
1.
2..________________________________________________________________________________
Please continue on next page.
How did you benefit from attending this presentation? (Please mark all that apply.)
----- learned more about the need for contraception after pregnancy
----- learned more about the timing of initiation of contraceptive methods after pregnancy
verified existing knowledge
----- gained a new perspective related to contraception after pregnancy
----- increased confidence in providing contraception after pregnancy
did not benefit
other (specify)
Based on the information presented today, would you consider making any changes to your professional
practice?
| | Yes
| | No
If No, why would you not consider making any changes?
If Yes, what changes would you consider?
Please respond to each of the following statements by circling the answer that best describes yourfeelings.
The information presented was relevant to my job.
Strongly agree
Agree
Disagree
Strongly disagree
The slides and other visual aids enhanced my understanding of the presentation content.
Strongly agree
Agree
Disagree
Strongly disagree
The training activities enhanced my understanding of the presentation content (if applicable).
Strongly agree
Agree
Disagree
Strongly disagree
Agree
Disagree
Strongly disagree
Disagree
Strongly disagree
The presentation was too technical.
Strongly agree
The duration of the presentation was too long.
Strongly agree
Agree
Please rate the following presentation components as either excellent (A), good (B), fair (C), or poor (D).
If the presenter did not use a particular component, please indicate “not applicable” or “NA”.
Rating
Comments
(A, B, C, D or NA)
35 mm slides/transparencies
fact sheet handout
audience note pages
scientific articles
pretest/posttest
training activities
other (specify)
Please add any additional comments or suggestions.
Thank you for your assistance. Please return completed form to: CTU Modules Project Administrator
Family Health International
P.O. Box 13950
Research Triangle Park, NC 27709 USA
October 2000
Audience
Handouts
UJH- Sfl-S-
Contraception after Pregnancy______________
The initiation and use of a contraceptive method after delivery or abortion.
Method should be initiated before fertility returns.
Contraceptive Options after Childbirth
Breastfeeding Is Important
Breastfeeding Women: Progestin-Only Methods
• Breastfed infants have lower mortality rates
than those who are not breastfed
• International Planned Parenthood
Federation recommends that:
- health-care providers encourage full breastfeeding
- breastfeeding not be discontinued to initiate
a contraceptive method
- the method not adversely affect breastfeeding
or infant health
• When the mother is HIV-positive:
- risk to the infant can be eliminated if a safe, ongoing
and clean method of bottlefeeding is available
- risk to the infant may be reduced if breastfeeding
is limited to the first 6 months
• Progestin-only methods include:
~
- progestin-only pills (POPs)
- progestin-only injectables (DMPA, NET-EN)
- implants (Norplant)
- levonorgestrel intrauterine system (LNg IUS)
• No effect on breastfeeding, breastmilk production
or infant growth and development
• WHO recommends delay of 6 weeks because very
young infants may be at risk to exposure to the
hormone
Breastfeeding women who have unprotected intercourse
can safely use POPs for emergency contraception
Breastfeeding Women:
Nonhormonal Methods
• Combined methods include:
zk
- combined oral contraceptives (COCs)
- monthly injectables (Mesigyna, Cyclofem)
xz
• Not to be used during first 6 weeks postpartum
due to effect on establishment of lactation
• Not recommended during first 6 months
postpartum due to decrease in milk production
• Can be used at 6 months postpartum, but not a
preferred option
Breastfeeding women who have unprotected intercourse
can safely use COCs for emergency contraception
All nonhormonal contraceptive methods can be
used safely by breastfeeding women
• No interference with breastfeeding
• No effect on the quality or quantity of breastmilk
• No effect on infant growth and development
Breastfeeding Women:
Combined Estrogen-Progestin Methods
Method Initiation for
Breastfeeding Women
Method Initiation for
Nonbreastfeeding Women Postpartum
' can be initiated 4 ■ 6 weeks if uterus has returned to normal size
' can be Initiated 4 - 6 weeks if uterus has returned to normal size
Contraceptive Options after Abortion
Women regain fertility 2 - 4 weeks after abortion
Uncomplicated First-Trimester Abortion:
Contraceptive Options
Uncomplicated Second-Trimester Abortion:
Contraceptive Options
• Induced abortion: any method can be
used immediately except.
• Induced and spontaneous abortion: any method
can be used immediately except.
- diaphragm or cervical cap: delay 4 - 6 weeks
• Spontaneous abortion: any method can
be used immediately
- diaphragm or cervical cap: 4 - 6 week delay
may be necessary in some cases
- diaphragm or cervical cap: delay 4 - 6 weeks
- IUD: delay insertion 4 - 6 weeks unless provider
is trained in insertion immediately after abortion
- female sterilization: easier to perform if delayed
until uterus returns to normal position
Complicated Abortion: Infection
Complicated Abortion: Genital Trauma
• If infection is present or suspected,
delay female sterilization and IUD insertion
• Follow guidelines for uncomplicated abortion
for initiation of all other methods
• Advise client to avoid intercourse until
infection is resolved
• If genital trauma exists, delay:
- use condoms if intercourse is not avoided
- female sterilization, unless performed
during required surgery
- IUD insertion
- female barrier methods and spermicides
(depending on extent and location of injury)
• Follow guidelines for uncomplicated abortion
for initiation of all other methods
Complicated Abortion:
Hemorrhage and Severe Anemia
• Delay methods that increase or maintain short-term blood loss (female sterilization, IUD)
• Follow guidelines for uncomplicated abortion for initiation of all other methods
Contraceptive Counseling
of Women after Abortion
Counseling the Partners
of Clients after Abortion
Women who have undergone abortion
should be treated with respect
• Counseling male partners of abortion clients can
increase their support for contraception
• Egyptian study found that husband’s support for
family planning was strongest predictor of
client’s use of contraception
• Counseling male partners may also increase
Counseling can help women:
• identify factors that led to abortion
• make informed choice of contraceptive method
• choose highly effective method when pregnancy
is life threatening
• become aware of emergency contraception
• understand implications for future fertility
© Family Health International, 2000
awareness and use of vasectomy or male
condoms
Source: Abdel-Tawab N„ 1999.
Slide 34 Activity:
Method Initiation for Breastfeeding Women
____________________________ I
Delivery
16 weeks
6 months onward
AAAAAAA
Slide 35 Activity: Method Initiation for
Nonbreastfeeding Women Postpartum
AAAAA
Patient Scenarios
The following patient scenarios are to be used for the Slide 48 Activity in the
narrative. Make copies and cut them so that each question is on an individual strip
ofpaper. Distribute them to the participants. An Answer Key follows.
1.
A 20-year-old woman is two-weeks postpartum after giving birth to her first child.
She is generally healthy and is breastfeeding. She possibly wants more children.
What contraceptive methods can she initiate at this time?
2.
A 32-year-old woman is immediately postpartum after giving birth by cesarean
section. This is her fourth healthy child. She is generally healthy and is planning to
breastfeed. What contraceptive methods can she initiate at this time?
3.
A23-year-old woman is six-weeks postpartum after giving birth to her second
child, who died soon after delivery. She had a difficult pregnancy and is not planning
to have another child soon. What contraceptive methods can she initiate at this time?
4.
A 26-year-old woman is four-months postpartum after giving birth to her first child.
She is breastfeeding, and both mother and child are healthy. She wants to postpone
her next pregnancy for a few years. What contraceptive methods can she initiate at
this time?
5. A 28-year-old woman is two-weeks postpartum with her third child. She is breast
feeding but does not want to rely on LAM. She wants to use another method and feels
strongly about not having more children. What contraceptive methods can she initiate
at this time?
6. A 29-year-old woman experienced a first-trimester uncomplicated miscarriage two
weeks ago. She would like to get pregnant again but wants to delay it for a few
months. What contraceptive methods can she initiate at this time?
7. A 36-year-old woman underwent a second-trimester abortion one week ago,
complicated by an infection which has not yet completely resolved. She does not
want any more children. What contraceptive methods can she initiate at this time?
Patient Scenarios (Answer Key)
1. A 20-year-old woman is two-weeks postpartum after giving birth to her first child.
She is generally healthy and is breastfeeding. She possibly wants more children.
What contraceptive methods can she initiate at this time?
A: LAM, condoms, spermicides or withdrawal.
2. A 32-year-old woman is immediately postpartum after giving birth by cesarean
section. This is her fourth healthy child. She is generally healthy and is planning to
breastfeed. What contraceptive methods can she initiate at this time?
A: LAM, condoms, spermicides, copper IUD, withdrawal,
female sterilization (if discussed and consent given in advance) or
vasectomy (if discussed and consent given in advance).
3.
A 23-year-old woman is six-weeks postpartum after giving birth to her second
child, who died soon after delivery. She had a difficult pregnancy and is not planning
to have another child soon. What contraceptive methods can she initiate at this time?
A: She may be interested in a long-term method such as progestin-only
or combined injectables, implants, copper IUD or LNg IUS. Other
appropriate methods include progestin-only or combined pills,
withdrawal, periodic abstinence, condoms, spermicides, diaphragm
or cervical cap.
4.
A 26-year-old woman is four-months postpartum after giving birth to her first child.
She is breastfeeding, and both mother and child are healthy. She wants to postpone
her next pregnancy for a few years. What contraceptive methods can she initiate at
this time?
A: She may be interested in a long term method such as progestin-only
injectables, implants, copper IUD or LNg IUS. Other appropriate methods
include progestin-only pills, withdrawal, periodic abstinence, condoms,
spermicides, diaphragm or cervical cap.
5. A 28-year-old woman is two-weeks postpartum with her third child. She is breast
feeding but does not want to rely on LAM. She wants to use ano±er me±od and feels
strongly about not having more children. What contraceptive methods can she initiate
at this time?
A: Until she is six-weeks postpartum, the only methods she may use are
condoms, spermicides, withdrawal or vasectomy (if discussed and consent
given in advance). At six weeks, female sterilization may be initiated (if
discussed and consent given in advance), as well as other long-term
methods such as injectables, implants, copper IUD or LNg IUS.
6. A 29-year-old woman experienced a first-trimester uncomplicated miscarriage two
weeks ago. She would like to get pregnant again but wants to delay it for a few
months. What contraceptive methods can she initiate at this time?
A: Condoms, spermicides, progestin-only pills, combined pills, withdrawal
or periodic abstinence.
7. A 36-year-old woman underwent a second-trimester abortion one week ago,
complicated by an infection which has not yet completely resolved. She does not
want any more children. What contraceptive methods can she initiate at this time?
A: Condoms, spermicides, progestin-only pills, injectables or implants,
combined pills or injectables, withdrawal, periodic abstinence or
vasectomy (if discussed and consent given in advance). Once the infection
is resolved and it has been at least six weeks since the procedure, she may
choose to use other long-term methods such as female sterilization (if
discussed and consent given in advance), copper IUD or LNg IUS.
Audience Notes
Contraception
after
Pregnancy
2.
Topics to Be Covered
Contraception after pregnancy: Overview
Contraceptive options after childbirth
Contraceptive options after abortion
3.
Contraception after Pregnancy:
Overview
Unmet need for contraception
Integration with other health services
Characteristics of effective services
Counseling
Contraception After Pregnancy
5.
Birth Spacing Saves Infant Lives
Family Planning Can Reduce
Abortion-Related Deaths
Contraception after Pregnancy:
Definition and Return to Fertility
The initiation and use of a contraceptive method
after childbirth or abortion.
Ideally, initiated before fertility returns.
2
Contraception After Pregnancy
9.
Integration of Services:
Health-Care Contacts___________
The period during and soon after pregnancy can
be a unique opportunity for providers to offer
family planning information and services
‘Abortion
' Childbirth .. "
visits
ivi:’
p..Following d^iveryj''.'..'.|<
Postpartum/.'
gfeinfant healthvisits.tf
10.
t'.: Betoreprocediire Cpt /j.-tife
,7 Following procedure, a <1
;; /.Whenheated (or••.,rcdrnp:lcatlons '. ”
Integration of Services Can
Increase Contraceptive Prevalence
11.
Integration of Services: Facilities
Facilities adding contraceptive services after
pregnancy should have:
private area for counseling
private area where
methods can be
provided
supply of contraceptives
and place to store them
12.
Effective Services after Pregnancy
Effective services include:
• offering a variety
of contraceptive
methods
• training all levels of
providers in clinical
skills and counseling
Contraception After Pregnancy
3
13.
Training of Providers
Providers of contraceptive services can
benefit from training in:
• communication skills
• how to be reasonably sure a woman is not pregnant
• how to help women make voluntary
informed choices
• characteristics of contraceptive methods and their
initiation time following childbirth or abortion
Staff should know where to refer for methods
not provided on-site
Contraceptive Counseling
after Pregnancy
Counseling can help client make an
informed choice
To help client select a method,
providers should:
• help to assess her reproductive goals,
discussing short- and long-term
contraceptive needs
• provide brief, clear, accurate information
about all methods of interest to client
15.
Contraceptive Counseling
after Pregnancy (cont’d)
After client selects a method, provider
should discuss:
• chosen method In detail
• the availability
of emergency
contraception
• condom use, if client
Is at risk for STDs/HIV
Counsel male partners, If appropriate
16.
Contraception for HIV-Positive
Women after Pregnancy
HIV-positive women especially need to know:
• the correct and consistent use of male
and female condoms can prevent
STD/HIV transmission
• using another contraceptive in addition
to a condom (dual method use) reduces
/,i
the chance of pregnancy, thus avoiding
HIV transmission to a child
\ '
4
Contraception After Pregnancy
17.
Contraceptive Options
After Childbirth
/
z
• Importance of breastfeeding
• Contraceptive options for
breastfeeding women
■ Timing of method initiation for
breastfeeding women
• Timing of method initiation for
nonbreastfeeding women
18.
Breastfeeding Is Important
• Breastfed Infants have lower mortality rates
than those who are not breastfed
• International Planned Parenthood
Federation recommends that:
- health-care providers encourage
full breastfeeding
- breastfeeding not be discontinued
to initiate a contraceptive method
- the method not adversely affect
breastfeeding or Infant health
19.
Breastfeeding When the Mother
is HIV-Positive
• Average risk of acquiring HIV through
breastmilk is at least 16%
• According to WHO:
- risk can be eliminated if a safe, ongoing and clean
alternative method of bottlefeeding is available
- risk may be reduced if breastfeeding is limited to the
first 6 months
Providers can help these mothers decide whether the
risks of breastfeeding outweigh the proven benefits
20.
Contraceptive Options for
Breastfeeding Women
• Nonhormonal methods
• Progestin-only methods
• Combined estrogen-progestin
methods
Contraception After Pregnancy
5
21.
Breastfeeding Women:
Nonhormonal Methods
All nonhormonal contraceptive methods can
be used safely by breastfeeding women
• No interference with breastfeeding
• No effect on the quality or quantity of breastmilk
• No effect on infant growth and development
23.
Nonhormonal Methods:
Barrier Methods
• Condoms are highly effective at
preventing pregnancy and STDs/HIV
when used consistently and correctly
Male and female condoms and
.
spermicides can be used
immediately postpartum
'-Sv**
Diaphragm and cervical cap
use must be delayed
until 6 weeks postpartum
&
'
Nonhormonal Methods:
Copper IUDs
• Specially trained providers can safely insert IUDs:
- immediate!/ after delivery of the placenta
- during cesarean section
- within 48 hours of childbirth
• If not inserted within 48 hours, delay
4 - 6 weeks
Clients should be counseled that postpartum
IUD insertions have higher expulsion rates than
do interval insertions
6
Contraception After Pregnancy
Copper IUDs: Experience and
Training Reduce Expulsion Rate
IUD expulsion rates at
6 months postpartum
26.
1
1 fil
Copper IUDs: Removal Rates
Lower for Breastfeeding Women
bleeding or pain
27.
Nonhormonal Methods:
Periodic Abstinence
Abstaining from intercourse
during woman’s fertile time
Can be difficult to use while
breastfeeding because
signs of fertility may
be absent or hard
to interpret
28.
Nonhormonal Methods: Withdrawal
• Presumably more effective when woman
is fully breastfeeding and fertility is reduced
• Pregnancy rates for nonbreastfeeding
women range from 4% - 21%
Contraception After Pregnancy
7
29.
Nonhormonal Methods:
Femalesterilization
• Ideally performed within
48 hours after delivery
• May be performed immediately
following delivery or during
cesarean section
• If not performed within 1 week
of delivery, delay 4 to 6 weeks
Thoroughly counsel clients In advance and
obtain Informed consent
30.
Nonhormonal Methods:
Male Sterilization
• Not effective until after 12 weeks
or 20 ejaculations
• Can be timed to coincide
with the postpartum period
when fertility is reduced or
abstinence may be practiced
Clients should be thoroughly
counseled In advance and
give Informed consent
Breastfeeding Women:
Hormonal Methods
32.
Breastfeeding Women:
Progestin-Only Methods
• Progestin-only methods indude:
- progestin-only pais (POPs)
- progestin-only injectables (DMPA, NET-EN)
- implants (Norplant)
- levonorgestrel Intrauterine system (LNg IUS)
• No effect on breastfeeding, breastmilk
production or infant growth and development
• WHO recommends delay of 6 weeks because very young
infants may be at risk of exposure to the progestin
Breastfeeding women who have unprotected intercourse
can safety use POPs for emergency contraception
8
Contraception After Pregnancy
33.
Breastfeeding Women: Combined
Estrogen-Progestin Methods
• Combined methods include:
- combined oral contraceptives (COCs)
- monthly injectables (Mesigyna, Cyclofem)
• Not to be used during first 6 weeks postpartum
due to effect on establishment of lactation
• Not recommended during first 6 months postpartum
due to decrease In milk production
• Can be used at 6 months postpartum, but not a
preferred option
Breastfeeding women who have unprotected intercourse
can safely use COCs for emergency contraception
36.
Contraceptive Options
After Abortion
\----—- ------------- /
• Types of abortion
• Contraceptive need after abortion
• Uncomplicated abortion
• Complicated abortion
• Provider training
\2
• Counseling
Contraception After Pregnancy
\
9
37.
Types of Abortion
• Contraceptive needs may differ
• Those who want to avoid pregnancy
have the same contraceptive options
38.
Contraceptive Need after Abortion
• Women regain fertility 2-4 weeks after
abortion
• To avoid another pregnancy, contraception
should be initiated as soon as possible after
abortion
Uncomplicated First-Trimester
Abortion: Contraceptive Options
• Induced and spontaneous abortion:
any method can be used immediately
except:
- diaphragm or cervical cap: delay of 4 - 6
weeks may be necessary in some cases
40.
Uncomplicated Second-Trimester
Abortion: Contraceptive Options
Induced and spontaneous abortion: any method
can be used immediately except:
- diaphragm or cervical cap: delay 4 - 6 weeks
- IUD: delay insertion 4 - 6 weeks unless provider
is trained in insertion immediately after abortion
female sterilization: easier to perform if delayed
until uterus returns to normal position
10
Contraception After Pregnancy
41.
Complicated Abortion:
Contraceptive Issues_______________
—
Complications that can affect
contraceptive choice include:
• infection
• trauma to genital tract
• hemorrhage, leading to severe anemia
2 or more complications can occur concurrently
42.
Complicated Abortion: Infection
• If infection is present or suspected, delay:
- female sterilization
- IUD insertion
• Follow guidelines for uncomplicated abortion for
initiation of all other methods
• Advise client to avoid intercourse until infection
is resolved
- use condoms if intercourse is not avoided
43.
—
Complicated Abortion:
Genital Trauma_________________
• If genital trauma exists, delay:
- female sterilization, unless performed during
required surgery
- IUD insertion
- female barrier methods and spermicides
(depending on extent and location of Injury)
• Follow guidelines for uncomplicated abortion for
initiation of all other methods
—
Complicated Abortion:
Hemorrhage and Severe Anemia
• Delay methods that increase or maintain
short-term blood loss
- female sterilization
- IUD
• Follow guidelines for uncomplicated abortion
for initiation of all other methods
Contraception After Pregnancy
11
45.
46.
Contraceptive Counseling
of Women after Abortion
Women who have undergone abortion
should be treated with respect
Counseling can help women:
• identify factors that led to abortion
• make informed choice of contraceptive method
• choose highly effective method when pregnancy
is life threatening
• become aware of emergency contraception
• understand implications for future fertility
47.
Counseling the Partners
of Clients after Abortion
• Counseling male partners of abortion clients
can increase their support for contraception
• Egyptian study found that husband’s support
for family planning was strongest predictor
of clienrs use of contraception
• Counseling male partners may also Increase
awareness and use of vasectomy or male
condoms
48.
Contraception after Pregnancy:
Conclusion
After pregnancy, many women want to delay or
avoid getting pregnant again
I
• Providing family planning to
women for use after pregnancy
can save lives and Improve
health and quality of life
Many contraceptive methods can
be used soon after pregnancy
Integrating contraceptive services
into maternal and child health
services can be highly effective
12
Contraception After Pregnancy
Slides
Contraception
after
Pregnancy
I
J
lx>
3>
Topics to Be Covered
«-t-
.^-
........... .,...
• Contraception after pregnancy: Overview
• Contraceptive options after childbirth
• Contraceptive options after abortion
Contraceptk n after Pregnancy:
Overview
• Unmet need for contraception
• Integration with other health services
• Characteristics of effective services
• Counseling
Unmet Need: Fertility Preferences
of Women Postparfu/n
J
Percentage of postpartam women
Source: Thapa S., 1992.
Birth Spacing Saves infant Lives
Source: Hobcraft J., 1991.
Family Planning Can Reduce
Maternal Mortality
~~..
Source: Koenig M., 1988.
.................................................. i
Parity
Family Planning Can Reduce
Abortion-Related
Deaths
...a
'
i
ZZ.... . . .. .s_
Number of Unsafe
Abortions Annually
Risk of Death
from Abortion
20,000,000
1 in 300
a
I
Asia
9,240,000
tin 250
1
Africa
3,740,000
1 in 150
Worldwide
• _
Source: WHO, 1993.
.
Contraception after Pregnancy:
Definition and Return to Fertility
The initiation and use of a contraceptive method
after childbirth or abortion.
Ideally, initiated before fertility returns.
Return to fertility: BOSOM® ■ z
After childbirth
Not breastfeeding Occurs as early as 4 weeks postpartum
]
Fully breastfeeding May be delayed until 6 months postpartum |
and possibly later
|:
After abortion Occurs as early as 2 weeks after procedure | \
Integration of Services:
Health-Care Contacts
Bifr'W- ->
.......
...............
The period during and soon after pregnancy can
be a unique opportunity for providers to offer
family planning information and services
Childbirth
• Prenatal visits
Abortion
• Following delivery
• Before procedure
• Following procedure
• Postpartum/
infant health visits
• When treated for
complications
Integration of Services Can
Increase Contraceptive Prevalence
Percentage of postpartum women
using contraception;
100
Lima, Peru
Source: Foreit K., 1993.
Received postpartum family planning services
Did not receive postpartum family planning
services
Integration of Services: Facilities
Facilities adding contraceptive services after
pregnancy should have:
private area for counseling
private area where
methods can be
provided
supply of contraceptives
and place to store them
Effective Services after Pregnancy
■
-
_________________________________ I
Effective services include:
• offering a variety
of contraceptive
methods
R. Lord
• training all levels of
providers in clinical
skills and counseling
Training of Providers
WMI^MIW
I
I
..................................................... l
Providers of contraceptive services can
benefit from training in:
• communication skills
• how to be reasonably sure a woman is not pregnant
• how to help women make voluntary
informed choices
• characteristics of contraceptive methods and their
initiation time following childbirth or abortion
Staff should know where to refer for methods
not provided on-site
Contraceptive Counseling
after Pregnancy
WIIHMM*..'.
....... ...............................................
Counseling can help client make an
informed choice
To help client select a method,
providers should:
• help to assess her reproductive goals,
discussing short- and long-term
contraceptive needs
• provide brief, clear, accurate information
about all methods of interest to client
Contraceptive Counseling
after Pregnancy (cont’d)
■Wil "I
... '.......................................................... l
After client selects a method, provider
should discuss:
• chosen method in detail
• the availability
of emergency
contraception
• condom use, if client
is at risk for STDs/HIV
R. Lord
Counsel male partners, if appropriate
Contraception for HIV-Positive
Women after Pregnancy
HIV-positive women especially need to know:
• the correct and consistent use of male
and female condoms can prevent
STD/HIV transmission
w
using another contraceptive in addition
to a condom (dual method use) reduces
the chance of pregnancy, thus avoiding
HIV transmission to a child
'
Contraceptive Options
after Childbirth
------------------- --------------
• Importance of breastfeeding
• Contraceptive options for
breastfeeding women
• Timing of method initiation for
breastfeeding women
• Timing of method initiation for
nonbreastfeeding women
Breastfeeding Is Important
• Breastfed infants have lower mortality rates
than those who are not breastfed
• International Planned Parenthood
Federation recommends that:
- health-care providers encourage
full breastfeeding
- breastfeeding not be discontinued
to initiate a contraceptive method
- the method not adversely affect
breastfeeding or infant health
Sources: WHO, 2000; IPPF, 1990.
R. Lord
Breastfeeding When the Mother
is HIV-Positive
• Average risk of acquiring HIV through
breastmilk is at least 16%
• According to WHO:
- risk can be eliminated if a safe, ongoing and clean
alternative method of bottlefeeding is available
- risk may be reduced it breastfeeding is limited to the
first 6 months
Providers can help these mothers decide whether the
risks of breastfeeding outweigh the proven benefits
Source: Nduati R., 2000.
Contraceptive Options for
Breastfeeding Women
• Nonhormonal methods
• Progestin-only methods
• Combined estrogen-progestin
methods
World Health Organization
Breastfeeding Women:
Nonhormonal Methods
MMM
All nonhormonal contraceptive methods can
be used safely by breastfeeding women
• No interference with breastfeeding
• No effect on the quality or quantity of breastmilk
• No effect on infant growth and development
Nonhormonal Methods: Lactational
Amenorrhea Method (LAM)
■■■■—IIW. . ...... —. -.. -..- - -____________________
Source: Institute for Reproductive Health, 1994.
Nonhormonal Methods:
Barrier Methods
• Condoms are highly effective at
preventing pregnancy and STDs/HIV
when used consistently and correctly W.
• Male and female condoms and
spermicides can be used
■L'-..
immediately postpartum
• Diaphragm and cervical cap
use must be delayed
r
until 6 weeks postpartum
rI
Nonhormonal Methods:
Copper IUDs
• Specially trained providers can safely insert IUDs:
- immediately after delivery of the placenta
- during cesarean section
- within 48 hours of childbirth
• If not inserted within 48 hours, delay
4 - 6 weeks
v
Clients should be counseled that postpartum
IUD insertions have higher expulsion rates than
do interval insertions
Copper IUDs: Experience and
Training Reduce Expulsion Rate
W—Mil mi..
IUD expulsion rates at
Source: Chi I., 1985.
.......... .......................... . ................. I
Copper IUDs: Removal Rates
Lower for Breastfeeding Women
-......... .........................................................
First-year IUD removal for
bleeding or pain
Nonhormonal Methods:
Periodic Abstinence
• Abstaining from intercourse
during woman’s fertile time
• Can be difficult to use while
Nonhormonal Methods: Withdrawal
• Presumably more effective when woman
is fully breastfeeding and fertility is reduced
• Pregnancy rates for nonbreastfeeding
women range from 4% - 21%
Source: Debusschere R., 1980.
Nonhormonal Methods:
Male Sterilization
........
.................................................................. I
• Not effective until after 12 weeks
or 20 ejaculations
• Can be timed to coincide
with the postpartum period
when fertility is reduced, or
abstinence may be practiced
Thoroughly counsel clients
in advance and obtain
informed consent
Breastfeeding Women:
Hormonal Methods
Breastfeeding Women:
Progestin-Only Methods
• Progestin-only methods include:
- progestin-only pills (POPs)
- progestin-only injectables (DMPA, NET-EN)
- implants (Norplant)
- levonorgestrel intrauterine system (LNg I US)
• No effect on breastfeeding, breastmilk
production or infant growth and development
• WHO recommends delay of 6 weeks because very young
infants may be at risk of exposure to the progestin
Breastfeeding women who have unprotected intercourse
can safely use POPs for emergency contraception
Breastfeeding Women: Combined
Estrogen-Progestin Methods
rr -
~
.. ............................ |
• Combined methods include:
- combined oral contraceptives (COCs)
- monthly injectables (Mesigyna, Cyclofem)
• Not to be used during first 6 weeks postpartum
due to effect on establishment of lactation
• Not recommended during first 6 months postpartum
due to decrease in milk production
fj
'
• Can be used at 6 months postpartum, but not a
preferred option
Breastfeeding women who have unprotected intercourse
can safely use COCs for emergency contraception
Method Initiation for Breastfeeding
Women
.. ........... ...................................................... I
Delivery
16 weeks
6 months onward
vwvwv
Lactational Amenorrhea Method
Condom/Spermicides
Copper IUD*
Female Sterilization*
Vasectomy/Withdrawal
Diaphragm / Cervical Cap______________________________
Progestin-only methods (pills, injectables, implants, LNg IUS*)
Periodic Abstinence
Combined Estrogen-Progestin
(pills, injectables)
* can be initiated 4-6 weeks if uterus has returned to normal size
Method Initiation for Non breastfeed! ng
Women Postpartum
I
Delivery
13 weeks
| 6 weeks onward
Vasectomy/Withdrawal
Copper IUD*
Female Sterilization*
AAAAA
Condom/Spermicides
Progestin-only Pills, Progestin-only Injectables, Implants
Periodic Abstinence
Combined Estrogen-Progestin (pills, injectables)
LNgIUS
Diaphragm / Cervical Cap
* can be initiated 4 - 6 weeks if uterus has returned to normal size
Contraceptive Options
after Abortion
■■■MM
• Types of abortion
• Contraceptive need after abortion
• Uncomplicated abortion
• Complicated abortion
• Provider training
• Counseling
Types of Abortion
___________________________________ i
Induced
Spontaneous
(miscarriage)
• Contraceptive needs may differ
• Those who want to avoid pregnancy
have the same contraceptive options
Contraceptive Need after Abortion
• Women regain fertility 2 - 4 weeks after
abortion
• To avoid another pregnancy, contraception
should be initiated as soon as possible after
abortion
Uncomplicated First-Trimester
Abortion: Contraceptive Options
• Induced and spontaneous abortion:
any method can be used immediately
except:
- diaphragm or cervical cap: delay of 4 - 6
weeks may be necessary in some cases
Uncomplicated Second-Trimester
Abortion: Contraceptive Options
—.. "________________________________ I
• Induced and spontaneous abortion: any method
can be used immediately except:
- diaphragm or cervical cap: delay 4 - 6 weeks
- IUD: delay insertion 4 - 6 weeks unless provider
is trained in insertion immediately after abortion
- female sterilization: easier to perform if delayed
until uterus returns to normal position
Complicated Abortion:
Contraceptive Issues
..'
1'
.................... .................... .....................................................................................................................
Complications that can affect
contraceptive choice include:
• infection
• trauma to genital tract
• hemorrhage, leading to severe anemia
2 or more complications can occur concurrently
i
Complicated Abortion: Infection
• If infection is present or suspected, delay:
- female sterilization
- IUD insertion
• Follow guidelines for uncomplicated abortion for
initiation of all other methods
• Advise client to avoid intercourse until infection
is resolved
- use condoms if intercourse is not avoided
Complicated Abortion:
Genital Trauma
_______________________________________ i
• If genital trauma exists, delay:
- female sterilization, unless performed during
required surgery
- IUD insertion
- female barrier methods and spermicides
(depending on extent and location of injury)
• Follow guidelines for uncomplicated abortion for
initiation of all other methods
Complicated Abortion:
Hemorrhage and Severe Anemia
t
- -
.—...... -........... .................................................................... j
• Delay methods that increase or maintain
short-term blood loss
- female sterilization
- IUD
• Follow guidelines for uncomplicated abortion
for initiation of all other methods
Training and Counseling Increase
Contraceptive Acceptance
Percentage of women postabortion
accepting modern contraception at
hospitals where staff receive training
Source: Diaz J., 1999.
Contraceptive Counseling
of Women after Abortion
____________________________________ l
Women who have undergone abortion
should be treated with respect
Counseling can help women:
• identify factors that led to abortion
• make informed choice of contraceptive method
• choose highly effective method when pregnancy
is life threatening
• become aware of emergency contraception
• understand implications for future fertility
Counseling the Partners
of Clients after Abortion
__________________
• Counseling male partners of abortion clients
can increase their support for contraception
• Egyptian study found that husband’s support
for family planning was strongest predictor
of client’s use of contraception
• Counseling male partners may also increase
awareness and use of vasectomy or male
condoms
Source: Abdel-Tawab N., 1999.
Contraception after Pregnancy:
Conclusion
...... ~
■■■Mill
__________________________ I
After pregnancy, many women want to delay or
avoid getting pregnant again
• Providing family planning to
women for use after pregnancy
can save lives and improve
health and quality of life
• Many contraceptive methods can
be used soon after pregnancy
R. Lord
• Integrating contraceptive services
into maternal and child health
services can be highly effective
Narrative
3/;.
Contraception after Pregnancy
Suggested Narrative
Introductory note to the presenter: This presentation will
familiarize health professionals with contraception after
pregnancy; that is, contraception after childbirth or abortion.
It is divided into three sections. The first section contains
general introductory information, including a discussion of
the unmet needfor contraceptive services after pregnancy,
the rationale for integrating these services with other health
services, and characteristics of effective contraceptive
services following pregnancy. The second section contains
information about the contraceptive methods most
appropriate for both breastfeeding and nonbreastfeeding
women postpartum and the timing of initiation of these
various methods. The third andfinal section discusses the
contraceptive methods appropriate for women who have had
spontaneous (miscarriage) or induced abortions (either
uncomplicated or complicated), and ways to improve family
planning servicesfor women following abortion.
Depending upon the background and needs ofyour audience,
you may decide to use some or all of the slides or supplement
them with information relevant to your local situation.
Items found in shaded boxes are suggestions for activities
and discussion questions that you may wish to use with your
audience. Including some or all of these elements can help
your presentation become more interactive, enhancing the
learning process. In some cases when activities are used,
you may find that slides covering the same material may
be omitted.
Regardless of whether or not you include the participatory
activities, be sure to familiarize yourself with the contents
prior to your presentation. At the beginning of the
presentation, inform your audience about the organization
of your presentation and how you would like to structure ->
Contraception after Pregnancy
the discussion (i.e., take questions during the presentation or
wait until the end). Ifyou plan to include participatory
activities, make sure your audience is aware of this.
Depending on how much information you decide to use, you
may want to break the presentation into two or more sessions.
In its entirety, the presentation takes 11/2 to 2 hours to
complete without the participatory activities and 2 to 2 1/2
hours when they are included.
Throughout the narrative, there are notes to the presenter,
similar to this one, that may enhance your presentation. ■
2
Contraception after Pregnancy
Section I.
Contraception after Pregnancy
Slide 1
Contraception
after
Pregnancy
Contraception after Pregnancy
This presentation is part of a series of training
modules produced by Family Health International.
In this module, we will discuss contraception after
pregnancy, which includes contraception after
childbirth or abortion.
The contraceptive needs of many women remain
unmet after pregnancy. This often has a negative
impact on the health and quality of life of women
and children. ■
Opening Activity -The Question Tree
Before you begin: Draw a picture of a tree on a piece offlip
chart paper and hang it on a nearby wall with pins or tape.
Provide small pieces ofpaper on which participants can write
their questions.
• Tell participants that you would like to create a Question
Tree before you begin the session.
• Point to the image of the tree hanging on the wall. Tell them
that this is the Question Tree. The Question Tree is where
participants will place questions they have about
contraception after pregnancy. Tell participants that they will
answer the questions on the Question Tree together at the end
of the session.
• Distribute the small pieces ofpaper so that everyone has at
least two pieces.
• Ask participants to write down one or two questions they
have about contraception after pregnancy. They do not need to
write their names on the paper. They should use a new piece of
paper for each question. When they are finished, they should
pin or tape their questions to the Question Tree.
• Give participants enough time to finish writing and hanging
their questions.
• Explain that they will return to the Question Tree later to see
if the session helped answer their questions. H
Contraception after Pregnancy
3
Slide 2
Topics to Be Covered
• Contraception after pregnancy: Overview
• Contraceptive options after childbirth
• Contraceptive options after abortion
Topics to Be Covered
This module consists of three sections.
The first section contains general introductory
information about contraception after pregnancy.
The second section describes contraceptive options
after childbirth for women, both breastfeeding and
nonbreastfeeding, and the timing of initiation of
various methods for both situations.
The third section discusses contraceptive options for
women who have had a spontaneous or induced
abortion. ■
Slide3
Contraception after Pregnancy:
Overview
• Unmet need for contraception
• Integration with other health services
• Characteristics of effective services
• Counseling
Contraception after Pregnancy: Overview
In this first section, we will:
• Consider the unmet need for contraception after
pregnancy and the advantages that meeting this need
provides.
• Discuss how integrating family planning services
with other health services gives providers more
opportunity to reach women during and after
pregnancy. During both of these periods, providers
can offer information about contraception and help
improve women’s access to contraceptive methods.
• Describe some characteristics of effective
contraceptive services for women after pregnancy.
• Discuss key messages for counseling after
pregnancy. ■
4
Contraception after Pregnancy
Slide 4
Unmet Need: Fertility Preferences of Women
Postpartum
After childbirth, many women do not want more
children or do not want them soon, yet are at risk for
pregnancy and are not using contraception.
Some of this contraceptive need can be met through
abstinence, which is a common practice for several
months after pregnancy in some cultures, or through
the temporary reduced fertility associated with
breastfeeding, which will be discussed later.
However, this slide shows the results of a study of
married women having unprotected intercourse
interviewed seven to nine months after giving birth.
The number of women who reported that they either
wanted to delay pregnancy for at least two years or
wanted no more children was highest in Latin
American and the Caribbean, reaching 95 percent in
Colombia. The number in Sub-Saharan Africa was
lowest but still substantial.
Women who have had an induced abortion have
demonstrated, by terminating their pregnancy, that the
pregnancy was problematic, unwanted or mistimed.
Many are likely to be motivated to prevent another
pregnancy. However, they seldom leave medical
facilities where abortions are performed or where
abortion-related complications are treated with the
knowledge and means to prevent pregnancy. ■
Slide5
Contraception after Pregnancy
Birth Spacing Saves Infant Lives
One advantage of meeting the need for contraception
after pregnancy is that it can help save infant lives.
Numerous studies have shown that children born after
an interval of at least two years from the preceding
birth are more likely to survive. As this slide shows,
infant deaths are dramatically reduced when births
occur two or more years apart. This pattern of -»
5
reduced infant deaths is evident regardless of the
country’s overall infant mortality rate. ■
Slide 6
Family Planning Can Reduce
Maternal Mortality__________________
uuJJ
Family Planning Can Reduce Maternal
Mortality
Family planning can also reduce maternal mortality by
reducing the total number of pregnancies and births,
both of which are risky. As the number of a woman’s
pregnancies and births increases, so does her lifetime
risk of maternal death.
_
This slide shows the maternal mortality ratio for
women 15 to 44 years old per thousand live births in
rural Bangladesh. Notice that when parity increases to
seven or more, the mortality ratio dramatically
increases. Although not displayed in this chart, the risk
is even greater for older women of high parity whose
bodies may be less capable of meeting the demands of
repeated pregnancy and childbirth. Access to family
planning services and subsequent use of contraception
may help reduce these higher-risk births. ■
Slide?
Family Planning Can Reduce
Abortion-Related Deaths
6
Family Planning Can Reduce AbortionRelated Deaths
Family planning can save the lives of women who
have had an induced abortion by helping them avoid
another unwanted pregnancy and the possibility of an
unsafe abortion that can end either in injury or death.
Based on estimates by the World Health Organization
(or WHO), one of every 300 women worldwide
undergoing an unsafe abortion dies. The danger posed
by unsafe abortion is even greater in certain regions of
the world. One of every 250 women undergoing
unsafe abortion in Asia and one of every 150 in Africa
dies as a result of the procedure. ■
Contraception after Pregnancy
Slide 8 Activity
• Divide the participants into small groups and have each
group answer the following:
- What is contraception after pregnancy?
- How does contraception after pregnancy differfrom
contraception at other times?
• Lead a discussion about the answers that are given and
correct any misconceptions. H
Slide 8
Contraception after Pregnancy:
Definition and Return to Fertility
The initiation and use of a contraceptive method
after childbirth or abortion.
Ideally, initiated before fertility returns.
Contraception after Pregnancy: Definition and
Return to Fertility
Contraception after pregnancy refers to the initiation
and use of a contraceptive method after childbirth or
abortion. Ideally, contraception should be initiated
before fertility returns.
The timing of return to fertility varies:
• For nonbreastfeeding women postpartum, fertility
can return as early as four weeks after childbirth.
These women need effective contraception starting at
this time, although contraceptive methods can be
initiated sooner.
• For breastfeeding women, return to fertility may be
delayed six months (or sometimes longer), depending
upon the intensity of breastfeeding. This will be
further discussed later in this presentation.
• Women are fertile as early as two weeks after an
abortion. Therefore, a contraceptive method should be
started as soon as possible. ■
Contraception after Pregnancy
7
Slide 9
Integration of Services: Health-Care Contacts
Providers have a unique opportunity to meet
women’s contraceptive needs during the prenatal
period and after pregnancy. Not only are women
more likely to be in contact with the health-care
system then, but also they are often highly motivated
to initiate contraception immediately after pregnancy.
For these reasons, integrating family planning
information and services with maternal and child
health services is important.
Family planning information can be naturally
integrated into prenatal visits during which nutrition,
breastfeeding and other issues that enhance the health
of both the mother and baby are commonly
discussed.
If the woman does not receive prenatal care, her
baby’s birth is another - and perhaps the only opportunity providers have to offer contraceptive
information. This information should be provided
only after a woman has recuperated from the
immediate stress of labor and delivery.
Finally, family planning services for new mothers
can be integrated with infant health care, such as
baby checkups and immunizations.
Women who have had an abortion sometimes return
to a medical facility for follow-up. But, unlike
women who have given birth, they will not be
returning for postpartum and infant checkups. Thus,
providers of abortions or those who treat abortion
complications have a responsibility to offer family
planning counseling before an abortion is performed,
following the procedure, or when the woman is
treated for complications. Such services are essential
in order to help women who wish to avoid another
unwanted pregnancy do so. At the very least, women
should be informed after the procedure or after
treatment for complications about their return to -»
8
Contraception after Pregnancy
fertility and options for safe, effective contraception.
If the woman wants to use contraception but is unable
to choose a method at the time of the abortion or
treatment, she should at least be offered a method
such as condoms during the interim. A referral for
family planning services can also be made. ■
Slide 10
Integration of Services Can Increase
Contraceptive Prevalence
One of the many advantages of coordination and full
integration of services is increased efficiency.
Improved access to family planning services after
pregnancy can also be very convenient for the client
and can increase contraceptive prevalence.
The Peruvian Social Security Institute conducted a
study at one of that country’s largest hospitals to
examine the effectiveness of providing postpartum
family planning services. In the study, women on one
maternity ward were offered family planning
counseling and contraceptive methods during their
hospital stay. Women on a second maternity ward
were discharged without receiving similar counseling
or services.
The research demonstrated that the total contraceptive
prevalence at 40 days postpartum was higher for the
group of women receiving family planning services.
At six months postpartum, contraceptive prevalence
was even higher: 82 percent for women receiving
family planning services versus 69 percent for those
who did not receive similar services. In addition, costs
of offering the services at the hospital were
significantly lower than the costs of providing these
services on an outpatient basis. Finally, use of hospital
resources to provide family planning services relieved
overcrowding at outpatient facilities, thus allowing
more women to be served. ■
Contraception after Pregnancy
9
Slide 11-12 Activity
• Divide the participants into small groups and give each
group a piece offlip chart paper and a marker.
• Tell each group to list on the paper the characteristics that
programs should have to offer effective services for
contraception after pregnancy.
• Have a representative from each group present their list to
the larger group. H
Slidell
Integration of Services: Facilities
Facilities adding contraceptive services after
pregnancy should have:
• private area for counseling
• private area where
methods can be
provided
• supply of contraceptives
and place to store them
Integration of Services: Facilities
Integration of services may not be easy to accomplish.
But many hospitals, doctors’ offices, and maternal and
child health facilities can add to existing services some
family planning services for women after pregnancy.
To provide these services, a facility should ideally
have:
• a private area for counseling women about
contraceptive needs,
• a private area where contraceptive methods can be
provided, and
• a supply of contraceptives and a place to store them.
Facilities lacking supplies of contraceptive methods
can make referrals to facilities that do provide them
but should at least be able to offer methods such as
condoms and, possibly, contraceptive pills. ■
Slide 12
Effective Services after Pregnancy
Effective services include:
• offering a variety
of contraceptive
methods
> training all levels of
providers in clinical
skills and counseling
Effective Services after Pregnancy
Policy-makers and providers should consider the
following factors as they develop or modify their
programs to provide effective family planning services
after pregnancy.
Services should offer a variety of contraceptive
methods to clients. Research shows that clients who
have a variety of methods from which to choose are
more likely to find one they like and to begin using it.
Offering a mix of methods results in higher
Contraception after Pregnancy
continuation rates because clients are more likely to
obtain a method that meets their needs.
To be effective, health-care providers at all levels
should be trained. This includes training to develop
new clinical skills, such as IUD insertion or
sterilization techniques after pregnancy. It also
includes training in counseling techniques and in the
special issues of postpartum contraception, such as the
effects some contraceptives have on breastfeeding.
Periodic refresher training should be offered to emphasize
important facts and share new information. ■
Slide 13
Training of Providers
Providers of contraceptive services can
benefit from training in:
• communication skills
• how to be reasonably sure a woman is not pregnant
• how to help women make voluntary
informed choices
• characteristics of contraceptive methods and their
initiation time following childbirth or abortion
Staff should know where to refer for methods
not provided on-site
Training of Providers
Programs that integrate family planning services into
other health services provided after pregnancy can
offer staff additional training in:
• communication skills,
• how to be reasonably sure a woman is not pregnant,
• how to help women make voluntary, informed
choices about contraception, and
• characteristics of contraceptive methods and when
methods can be initiated following childbirth or
abortion.
If certain contraceptive methods cannot be offered on
site, providers should know where in the community
to refer women for these me±ods and for ongoing
family planning care.
Note to presenter: For more information about how to be
reasonably sure a woman is not pregnant, see the pregnancy
checklist in the reprints section of this module. H
Contraception after Pregnancy
11
Slide 14
Contraceptive Counseling
after Pregnancy
Counseling can help client make an
informed choice
To help client select a method,
providers should:
• help to assess her reproductive goals,
discussing short- and long-term
contraceptive needs
• provide brief, clear, accurate information
about all methods of interest to client
Contraceptive Counseling after Pregnancy
Counseling is one of the most important elements of
contraceptive services provided after pregnancy. By
offering clients up-to-date information on available
contraceptives and providing an opportunity to discuss
concerns about reproductive health, providers can help
clients make an informed choice about contraceptive
use.
First, providers should ask clients about their
reproductive goals. Women who have had induced
abortions or have given birth recently may want to
avoid or postpone future pregnancies. Women who
have experienced spontaneous abortions, or
miscarriages, may want to become pregnant again
soon.
Clients should be encouraged to consider both their
short- and long-term contraceptive needs and options.
Proper counseling can help the woman postpartum
determine the method that is best for the short term,
both while she may be breastfeeding and after her
infant is weaned. Providers should also address
particular concerns that clients may have following
childbirth or abortion, such as the effects of
contraceptives on breastfeeding, infant development,
and the resumption of sexual relations.
Once reproductive goals have been discussed, a
woman wishing to postpone pregnancy should receive
brief, clear, accurate information about all appropriate
methods of interest to her, taking into consideration
whether she has given birth, had an induced abortion
or experienced a spontaneous abortion. This
information should include safety, effectiveness,
mechanisms of action, side effects, protection from
sexually transmitted diseases (or STDs), and other
characteristics of the methods. ■
Contraception after Pregnancy
Slide 15
Contraceptive Counseling
after Pregnancy (cont’d)
After client selects a method, provider
should discuss:
chosen method in detail
the availability
of emergency
contraception
condom use, if client
is at risk lor STDs/HIV
Counsel male partners, if appropriate
Contraceptive Counseling after Pregnancy
(continued)
After a client makes an informed choice, the provider
should discuss the chosen method in detail, ensuring
that the client understands its risks, benefits, side
effects and correct use. It is particularly important that
clients choosing female or male sterilization
understand that these methods are permanent,
although - as with all methods - there is a small risk
of contraceptive failure.
Clients should be counseled about the availability of
emergency contraception in case unprotected
intercourse occurs. Those at risk of STDs, including
human immunodeficiency virus (HIV), should be
encouraged to use male or female condoms alone or in
combination with other contraceptive methods.
If appropriate, both partners should be offered
counseling. In general, a male partner’s awareness of
the need for family planning and of available
contraceptive methods can increase contraceptive use
and continuation. It is particularly important that a
woman’s male partner be offered counseling if she is
at high risk for STDs. Some couples may choose to
rely on male methods, such as condoms or male
sterilization. When sterilization is chosen, the male
partner must receive careful counseling. ■
Slide 16
Contraception for HIV-Positive
Women after Pregnancy
HIV-positive women especially need to know:
• the correct and consistent use of male
and female condoms can prevent
STD/HIV transmission
• using another contraceptive in addition
to a condom (dual method use) reduces
the chance of pregnancy, thus avoiding
HIV transmission to a child
Contraception after Pregnancy
Contraception for HIV-Positive Women after
Pregnancy
After pregnancy, clients who are HIV-positive need
specific information, counseling and services, or
appropriate referrals.
At the very least, these clients need to know that,
aside from abstinence, condom use offers the best
protection against STDs. Male or female condoms
should be used every time intercourse occurs to ->
13
avoid HIV transmission to partners and to protect the
woman herself from other STDs, including other
strains of HIV. An HIV-positive woman should be
taught how to use condoms correctly and how to
negotiate condom use with her partner.
HIV-positive women also need to know that children
bom to them may become infected with the virus
during birth or breastfeeding. If they do not wish to
become pregnant, they should consider “dual method”
use; that is, using a condom for disease prevention and
another, more effective method for contraception.
This concludes the first section of the module, which
contained general introductory information about
contraception after pregnancy.
Note to presenter: A breastfeeding mother who is HIV-positive
risks transmitting the infection to her infant through her
breastmilk. In Section II, we will discuss this risk and ways
to reduce it. H
Summary Activity
At this point, essential information from Section I can be
summarized. The presenter can remind participants of the
unmet needfor contraception after pregnancy, discuss
integration offamily planning services with maternal and
child health services, describe some characteristics of
effective contraceptive services for women after pregnancy,
and discuss key messages for counseling after pregnancy. H
Contraception after Pregnancy
Section II.
Contraceptive Options after Childbirth
Contraceptive Options after Childbirth
In this section, we will discuss postpartum
contraceptive options. We will review:
Slide 17
Contraceptive Options
after Childbirth
•\
• Importance of breastfeeding
• Contraceptive options for
breastfeeding women
• Timing of method initiation for
breastfeeding women
• Timing of method initiation for
nonbreastfeeding women
/
• the importance of breastfeeding,
• contraceptive options for breastfeeding women,
• the timing of method initiation for breastfeeding
women, and
• the timing of method initiation for nonbreastfeeding
women. ■
Slide 18
Breastfeeding Is Important
Breastfed infants have lower mortality rates
than those who are not breastfed
International Planned Parenthood
Federation recommends that:
health-care providers encourage
full breastfeeding
- breastfeeding not be discontinued
to initiate a contraceptive method
the method not adversely affect
breastfeeding or infant health
Breastfeeding Is Important
The importance of breastfeeding has been
demonstrated in various studies, including a WHOcoordinated international study. WHO found that
breastfed infants were six times less likely to die of
infectious diseases in the first few months of life and
had a lower mortality rate through the second year of
life compared to those who were not breastfed.
The International Planned Parenthood Federation
recommends that:
• health-care providers encourage full breastfeeding,
• breastfeeding not be discontinued to start the use of
a contraceptive method, and
• the chosen method not adversely affect breastfeeding
or the infant’s health. ■
Contraception after Pregnancy
15
Slide 19
Breastfeeding When the Mother
is HIV-Positive
• Average risk of acquiring HIV through
breastmilk is at least 16%
• According to WHO:
- risk can be eliminated if a safe, ongoing and clean
alternative method of bottlefeeding is available
- risk may be reduced if breastfeeding is limited to the
first 6 months
Providers can help these mothers decide whether the
risks of breastfeeding outweigh the proven benefits
Breastfeeding When the Mother is HIVPositive
Special consideration must be given to the decision
to breastfeed if a mother is HIV-positive, since the
virus may pass to the infant during breastfeeding in
some cases. The average risk of acquiring HIV
infection through breastmilk is at least 16 percent.
According to WHO, an HIV-positive mother can
eliminate the risk of HIV transmission through
breastmilk by using infant formula, modified animal
milks or boiled expressed breastmilk. However, she
must have access to a sufficient, ongoing and clean
supply of this alternative form of milk. If there is no
safe alternative form of milk, an HIV-positive
mother should only give her infant breastmilk. She
should not mix breastfeeding with use of other foods
or milk replacements that are not safe. According to
WHO, limiting breastfeeding to the first six months
may also reduce the risk of HIV transmission.
Health-care providers should counsel HIV-positive
mothers and help them decide whether the risks of
breastfeeding outweigh the proven benefits. ■
Slide 20
Contraceptive Options for Breastfeeding
Women
About 90 percent of all mothers breastfeed their
infants for some period of time, so the impact of
contraceptive methods on breastfeeding, breastmilk
and infant health is a very important consideration.
We will now discuss contraceptive options for
breastfeeding women. These options include
nonhormonal methods, progestin-only methods, and
combined estrogen-progestin methods. Some of
these methods are more appropriate than others for
use during breastfeeding.
16
Contraception after Pregnancy
Keep in mind that the following discussion applies
only to breastfeeding women. ■
Slide 21
Breastfeeding Women:
Nonhormonal Methods
All nonhormonal contraceptive methods can
be used safely by breastfeeding women
• No interference with breastfeeding
• No effect on the quality or quantity of breastmilk
• No effect on infant growth and development
Breastfeeding Women: Nonhormonal
Methods
All nonhormonal methods can be used safely by
breastfeeding women.
They do not:
• interfere with a woman’s ability to breastfeed,
• affect the quality or quantity of breastmilk, or
• affect infant growth and development.
Now, let us discuss types of nonhormonal
contraceptive methods: lactational amenorrhea
method (or LAM), barrier methods, intrauterine
devices (or IUDs), periodic abstinence, withdrawal
and female and male sterilization. ■
Slide 22 Activity
• By way of review, divide the participants into seven groups
and give them each a piece offlip chart paper and a marker.
Have each group discuss one of the nonhormonal methods
(the lactational amenorrhea method, barrier methods, IUDs,
periodic abstinence, withdrawal andfemale and male
sterilization) in terms of:
— How the method works
— Advantages and disadvantages of the method
— Effectiveness of the method
— Protection the method offers against STDs/HTV
• Have a representativefrom each group present their list to
the larger group and correct any misconceptions. H
Contraception after Pregnancy
17
Slide 22
Nonhormonal Methods: Lactational
Amenorrhea Method (LAM)________
Nonhormonal Methods: Lactational
Amenorrhea Method (LAM)
The Lactational Amenorrhea Method, also known as
LAM, is a temporary contraceptive option used for up
to six months postpartum by women who are fully or
nearly fully breastfeeding and remain amenorrheic.
Fully or nearly fully breastfeeding means that there
are no intervals greater than four to six hours between
breastfeeds and no regular dietary supplements have
been introduced. When a woman fully or nearly fully
breastfeeds, the baby’s frequent suckling causes
several biological changes that inhibit ovulation,
resulting in a state of temporarily reduced fertility.
LAM provides more than 98 percent protection from
pregnancy until one of three conditions occurs:
• menses returns, or
• the woman is no longer fully - or nearly fully breastfeeding, or
• the baby reaches six months of age.
When one of these conditions occurs, another
contraceptive method should be started for continued
protection from pregnancy. A woman should continue
breastfeeding for as long as she wishes, since
breastfeeding is good for her infant’s health. Providers
should anticipate the period when LAM will cease in
order to provide protection from pregnancy and
counsel women well in advance about the use of other
methods.
Recent studies suggest that if a woman remains
amenorrheic and continues breastfeeding, LAM may
be effective for some women for up to 12 months.
Research into the duration of LAM’s effectiveness
continues. ->
18
Contraception after Pregnancy
Note to presenter: For more detailed information about this
method, see the "Lactational Amenorrhea Method (LAM)"
module in the Contraceptive Technology Update Series. B
Slide 23
Nonhormonal Methods:
Barrier Methods
Condoms are highly effective at
preventing pregnancy and STDs/HIV
when used consistently and correctly
Male and female condoms and
spermicides can be used
immecfiately postpartum
Diaphragm and cervical cap
use must be delayed
until 6 weeks postpartum
Nonhormonal Methods: Barrier Methods
Barrier methods are another type of nonhormonal
contraception. These methods include male and
female condoms, spermicides, diaphragms and
cervical caps. Barrier methods offer various degrees
of protection against pregnancy. Some barrier
methods also protect against STDs. Condoms are
highly effective at preventing pregnancy and provide
good protection against all STDs, except human
papilloma virus (HPV), when used consistently and
correctly. Spermicides provide modest protection
against some bacterial STDs, and are much less
effective than condoms in preventing pregnancy.
The effectiveness of barrier methods - as is true of
some other methods - depends largely on consistent
and correct use. Thus, it is important to explain to
clients that barrier methods must be used correctly
with each and every act of intercourse to be most
effective. In addition, clients need to know where and
how to obtain supplies of barrier methods.
Typically, male and female condoms and spermicides
can be used immediately postpartum and can be
obtained without a visit to a health-care provider. A
variety of types are usually available. Condoms
should be strongly encouraged for women at risk for
STDs.
Use of a diaphragm or cervical cap must be delayed
until six weeks postpartum. This is because the
effectiveness of diaphragms and cervical caps
depends on proper anatomical fit over the cervix,
which does not return to its normal size until about six
weeks postpartum.
Contraception after Pregnancy
19
Note to presenter: For more detailed information about barrier
contraception, see the "Barrier Methods" module in the
Contraceptive Technology Update Series. For more detailed
information about STDs, see the "Sexually Transmitted
Diseases” module in the Reproductive Health Series. H
Slide 24
Nonhormonal Methods:
Copper IUDs
Specially trained providers can safely insert IUDs:
- Immediately after delivery of the placenta
- within 48 hours of childbirth
If not inserted within 48 hours, delay
4 - 6 weeks
Clients should be counseled that postpartum
IUD insertions have higher expulsion rates than
do interval insertions
Nonhormonal Methods: Copper IUDs
The copper IUD is a long-acting and reversible
method that can be safely inserted postpartum if
providers have proper training. Postpartum insertion
is convenient for the woman and may be her only
opportunity to obtain an IUD. However, like all
women, women postpartum should be carefully
screened to determine if they are at risk of STDs. The
device can be inserted if laboratory testing rules out
the presence of an STD. If women are at risk for
STDs and laboratory testing is not possible, IUD use
is not recommended.
Insertion can be safely performed vaginally
immediately after the delivery of the placenta or
during a cesarean section (through the uterine
incision) as long as the woman has been counseled
and has chosen the method well in advance of
childbirth. If not done immediately, insertion can be
performed safely within 48 hours after childbirth.
Postpartum insertion of copper IUDs poses no greater
risk of infection, bleeding or perforation than
insertion at other times. However, IUD expulsion
rates are higher for postpartum insertions (especially
those taking place after the first 10 minutes after
placental delivery) than for interval insertions (those
taking place any time after six weeks postpartum).
Clients who have IUDs inserted within the 48 hours
ppstpartum should be told that the risk of expulsion is
greater and taught how to check IUD strings to ensure
that the device is still in place.
20
Contraception after Pregnancy
If an IUD insertion is not performed within 48 hours
after delivery, WHO recommends that insertion
be delayed at least four to six weeks. This
recommendation is based on earlier observations that
the risk of perforation is greater for postpartum
insertions of older IUDs, such as Lippes Loop. With
copper IUDs, it has been shown that there is no
increased risk of perforation when inserted during the
first 48 hours postpartum or after four weeks
postpartum. However, there are no data on the risk of
perforation between 48 hours and four weeks. ■
Slide 25
Copper IUDs: Experience and
Training Reduce Expulsion Rate
Copper IUDs: Experience and Training
Reduce Expulsion Rate
Appropriate training of providers in postpartum IUD
insertion techniques can reduce the risk of IUD
expulsion. Training should emphasize placement of
the device high in the uterus.
This slide shows the results of a large study
comparing postpartum expulsion rates of insertions
performed by experienced and inexperienced health
care providers. As you can see, expulsion rates at six
months were significantly lower when the insertion
was performed by an experienced provider trained in
immediate postpartum insertion technique. These
results have been corroborated in another large study
conducted in Belgium. ■
Slide 26
Copper IUDs: Removal Rates
Lower for Breastfeeding Women
Contraception after Pregnancy
Copper IUDs: Removal Rates Lower for
Breastfeeding Women
IUD insertion after childbirth offers certain benefits.
Research shows that breastfeeding women report less
pain during postpartum IUD insertion than women
who are not breastfeeding. Also, breastfeeding women
experience fewer side effects, specifically bleeding
and pain, related to IUD use than nonbreastfeeding
women.
21
As this slide shows, breastfeeding women who are not
amenorrheic at the time of IUD insertion have
significantly higher one-year removal rates for
bleeding or pain than breastfeeding women who are
amenorrheic at the time of insertion. Removal rates
are even higher for women who are not breastfeeding
at the time of insertion.
Note to presenter: For more detailed information about IUDs,
see the “Intrauterine Devices (IUDs)” module in the
Contraceptive Technology Update Series. H
Slide 27
Nonhormonal Methods:
Periodic Abstinence
• Abstaining from intercourse
during woman’s fertile time
• Can be difficult to use while
breastfeeding because
signs of fertility may
be absent or hard
to interpret
Nonhormonal Methods: Periodic Abstinence
Periodic abstinence is another nonhormonal
contraceptive option. To use periodic abstinence, a
woman must learn about the different stages of her
menstrual cycle. She must learn to predict when she
will ovulate and to identify her fertile period, during
which pregnancy is likely to occur. She must avoid
sexual intercourse around the time of expected
ovulation - on what are called “unsafe” or “fertile”
days.
However, this method can be difficult to use after
pregnancy when a woman is breastfeeding because
her signs of fertility may be absent or hard to
interpret. For example, one method of periodic
abstinence involves monitoring morning changes in
basal body temperature. But these changes will not
occur until ovulation has resumed. Even when
ovulation has resumed, these changes may not occur
if the woman gets up several times during the night to
breastfeed her infant. Another method of periodic
abstinence requires watching for patterns in cervical
mucus. However, reduced ovarian function during
breastfeeding may make the cervical mucus pattern
more difficult to interpret. Women who are fully
breastfeeding and amenorrheic may not have signs of
fertility for up to six months after childbirth. ■
22
- Contraception after Pregnancy
Slide 28
Nonhormonal Methods: Withdrawal
Presumably more effective when woman
is fully breastfeeding and fertility is reduced
Pregnancy rates for nonbreastfeeding
women range from 4% - 21%
Slide 29
Nonhormonal Methods:
Female Sterilization
Ideally performed within
48 hours after delivery
May be performed immediately
following delivery or during
cesarean section
If not performed within 1 week
of delivery, delay 4 to 6 weeks
Thoroughly counsel clients In advance and
obtain Informed consent
Nonhormonal Methods: Withdrawal
Withdrawal is not a very effective contraceptive
method, but it may be more reliable when a woman is
breastfeeding and her fertility is reduced.
For nonbreastfeeding women, pregnancy rates
associated with withdrawal range from four percent
to 21 percent during the first year. Even when this
method is used correctly, pregnancy may still be
possible because pre-ejaculatory fluid may contain
sperm. ■
Nonhormonal Methods: Female Sterilization
For women who do not want more children, female
sterilization is a safe and effective nonhormonal
postpartum contraceptive option. Ideally, the
procedure is performed not immediately but within
the first 48 hours postpartum, after the woman has
recovered from delivery, and the health and survival
of the newborn is more certain than immediately after
delivery.
However, surgical sterilization can also be performed
immediately following a vaginal delivery if the
woman has selected this method in advance.
Scheduling sterilization to coincide with delivery may
be appropriate if the woman has limited contact with
the health-care system. Sterilization can also be
performed at the time of cesarean section.
If sterilization is not performed within a week
postpartum, it should be delayed until four to six
weeks postpartum when the uterus returns to its
normal size and the fallopian tubes are easier to
locate.
If sterilization is to be performed at the time of
delivery or during cesarean section, providers should
counsel the client thoroughly and obtain informed
consent during the prenatal period - well before
Contraception after Pregnancy
23
the woman gives birth. Counseling should never be
conducted during the stress of labor or delivery.
Providers must be sure that clients realize that female
sterilization is permanent in order to help avoid regret.
However, they should also inform clients that - as
with all methods - there is a small risk of
contraceptive failure. ■
Slide 30
Nonhormonal Methods:
Male Sterilization
• Not effective until after 12 weeks
or 20 ejaculations
• Can be timed to coincide
.
with the postpartum period
/ /Oa Z
when fertility Is reduced or
/ / /<'; ?. „
abstinence may be practiced
r-ld * •- I
Clients should be thoroughly
\fl :
■I
counseled In advsnce end
'J
give Informed consent
Nonhormonal Methods: Male Sterilization
Male sterilization or vasectomy also should be
considered as a postpartum contraceptive option for
couples who have reached their desired family size.
Vasectomy is highly effective, causes few side effects,
and is safer and typically less expensive than female
sterilization. It is a relatively simple procedure that
can be performed at any time by properly trained
clinicians. Following the procedure, it usually takes
12 weeks or 20 ejaculations until sperm is cleared
from the man’s tubes for the sterilization to become
fully effective. Thus, the postpartum period - when a
breastfeeding woman’s fertility is reduced and
abstinence may be practiced - may be an ideal time
for a man to undergo vasectomy. However, if a
woman’s fertility returns before the vasectomy
becomes effective, another form of contraception
should be used in the interim.
Providers must counsel men that vasectomy, like
female sterilization, is permanent, although - as with
all methods - there is a small risk of contraceptive
failure. Careful counseling and informed consent
before the procedure is performed are essential to help
avoid regret.
Note to presenter: For more detailed information about
sterilization, see the “Female and Male Sterilization ” module
in the Contraceptive Technology Update Series. H
24
Contraception after Pregnancy
Slide 31
Breastfeeding Women: Hormonal Methods
Now that we have discussed nonhormonal method
options for breastfeeding women, let us move on to
hormonal methods, beginning with progestin-only
methods and continuing with combined estrogen
progestin methods. ■
Slide 32
Breastfeeding Women: Progestin-Only
Methods
Progestin-only methods include progestin-only
contraceptive pills (or POPs), injectables such as
DMPA and NET-EN, subdermal implants such as
NORPLANT, and the levonorgestrel intrauterine
system (LNg IUS).
Breastfeeding Women:
Progestin-Only Methods
- progwan-onJy
(DMPA. NET-EN)
..
B
- TotanutNoiptanl)
\\ ///
- Iivonorp«str«i ntrvutarina tyitam (LNg IUS) \\///
No eflod on breastfeeding, breastmilk
X-*/
Studies indicate that the use of progestin-only
contraceptives does not affect breastfeeding,
breastmilk production, or infant growth and
development. Regardless of the progestin-only
method used, studies suggest that only a very small
amount of the progestin reaches the infant through
breastmilk. Long-term follow-up of children exposed
to progestin-only contraceptives through breastmilk
has not shown any adverse effects from this exposure.
However, WHO recommends that breastfeeding
women delay use of progestin-only methods until six
weeks postpartum due to theoretical concerns that
steroid hormones may pose some risk to young
infants.
If a breastfeeding woman has unprotected intercourse,
she can safely use progestin-only emergency
contraceptive pills, although many breastfeeding
women are protected by, and choose to rely on, LAM.
Note to presenter: For more detailed information about
emergency contraceptive pills, see the “Emergency
Contraceptive Pills ” module in the Contraceptive Technology
Update Series. H
Contraception after Pregnancy
25
Slide 33
Breastfeeding Women: Combined
Estrogen-Progestin Methods
• Not to be used during first 6 weeks postpartum
due to effect on establishment of lactation
• Not recommended during first 6 months postpartum
*v A
due to decrease in mik production
'
• Can be used at 6 months postpartum, but not a
preferred option
Breastfeeding women who have unprotected Intercourse
can safely use COCs for emergency contraception
Breastfeeding Women: Combined EstrogenProgestin Methods
Combined estrogen-progestin contraceptives include
combined oral contraceptives (or COCs) and monthly
injectable methods such as Mesigyna and Cyclofem.
Combined contraceptives have been shown to
interfere with the establishment of lactation. For this
reason, they should not be used by breastfeeding
women during the first six weeks postpartum because
lactation is not fully established during this period.
Combined estrogen-progestin contraceptives have
also been shown to decrease milk production. For this
reason, their use is not recommended for
breastfeeding women from six weeks to six months
postpartum.
At six months postpartum, breastfeeding women can
use a combined estrogen-progestin method, but it is
still not a preferred option for women who want to
continue breastfeeding.
If a breastfeeding woman has unprotected intercourse,
she can use combined estrogen-progestin emergency
contraceptive pills, although many breastfeeding
women are protected by, and choose to rely on, LAM.
This short-term exposure to estrogen is thought to
have no effect on milk production.
Note to presenter: For more detailed information about
hormonal contraceptives, see the “Oral Contraceptives, ”
“Injectable Contraceptives, ” “Introduction to Contraceptive
Methods" and “Emergency Contraceptive Pills” modules in
the Contraceptive Technology Update Series. H
26
Contraception after Pregnancy
Slide 34-35 Activity
• Divide the participants into small groups.
• Distribute a blank copy of the charts on slides 34 and 35
(available in the Audience Handouts section of this module) to
each group.
• Tell each group to fill in the charts, determining which
methods go in each category.
• Have a representative from each group present their charts
to the larger group. Lead a discussion and correct any
misconceptions. B
Slide 34
Method Initiation for Breastfeeding
Women
Method Initiation for Breastfeeding Women
Now we will review contraceptive options for
breastfeeding women according to the timing of
initiation.
Women who are fully breastfeeding and want to use
LAM can begin doing so immediately after childbirth.
LAM is considered to be effective for up to six
months. No additional contraceptive method is
needed. However, women who cannot or choose not
to rely on LAM have the following contraceptive
options.
Condoms and spermicides can be initiated at any time
postpartum. IUD insertion can be performed within
the first 48 hours after birth, and female sterilization
can be performed within the first week of birth
(although preferably within 48 hours). Otherwise,
these two procedures should be postponed until four
to six weeks postpartum. Male sterilization, or
vasectomy, can be performed anytime during this
period and withdrawal can be used anytime as well.
At six weeks postpartum, all of the methods
mentioned above can be used. In addition,
diaphragms, cervical caps and progestin-only methods
can be initiated. Periodic abstinence may be difficult
to use because breastfeeding women may not have
signs of fertility or, if they do, the signs may be
Contraception after Pregnancy
27
difficult to interpret. However, women who choose to
use periodic abstinence should begin keeping track of
their menstrual cycles and watching for signs of
fertility at six weeks postpartum.
LAM can be effective for up to 12 months for some
women who are fully breastfeeding; however, it is
generally considered ineffective after six months
postpartum. At this time, any of the other methods
mentioned above may be initiated. In addition,
combined hormonal contraceptives can be initiated at
this time, although they are still not a preferred option
for breastfeeding women. H
Slide 35
Method Initiation for Nonbreastfeeding
Women Postpartum
Method Initiation for Nonbreastfeeding
Women Postpartum
Providers need to offer nonbreastfeeding women
contraceptive options early in the postpartum period
since these women will not benefit from the protection
that LAM offers. Contraceptive use is most effective if
started by three weeks postpartum. For
nonbreastfeeding women postpartum, all contraceptive
methods except LAM are options. However, the
timing of initiation varies according to the method.
As shown in the top three horizontal bars, use of
condoms, spermicides, progestin-only pills, progestinonly injectables, subdermal implants, vasectomy and
withdrawal may be initiated immediately postpartum.
IUD insertion can be performed within the first 48
hours after birth or delayed four to six weeks
postpartum. Female sterilization can be performed
within the first week after giving birth (although
preferably within 48 hours). Otherwise, it should be
delayed until four to six weeks postpartum.
Women who wish to use periodic abstinence should
begin looking for signs of fertility about three weeks
postpartum.
Contraception after Pregnancy
Combined estrogen-progestin contraceptives should
be delayed until three weeks postpartum due to the
increased risk of blood clotting problems during this
period.
LNg IUS insertion should be delayed until the uterus
returns to normal size at six weeks postpartum.
As is true for breastfeeding women, initiation of the
use of a diaphragm or cervical cap must be delayed
until six weeks postpartum when the woman’s uterus
has returned to normal size, at which time the devices
can be fitted by a trained provider.
This concludes the section about contraceptive
options for women postpartum. 13
Contraception after Pregnancy
29
SECTION III.
Contraceptive Options after Abortion
Slide 36
Contraceptive Options
After Abortion
• Types of abortion
• Contraceptive need after abortion
• Uncomplicated abortion
• Complicated abortion
• Provider training
• Counseling
Contraceptive Options after Abortion
In this section, we will discuss contraceptive options
after abortion. We will review:
• types of abortion,
• contraceptive need after abortion,
• contraception after uncomplicated abortion,
• contraception after complicated abortion,
• training of providers working with clients after
abortion, and
• counseling of clients who have had an abortion. ■
Slide 37
Types of Abortion
Induced
'
I|
|
Spontaneous
(ml«carrtaD«)
• Contraceptive needs may differ
• Those who want to avoid pregnancy
have the same contraceptive options
Types of Abortion
There are two types of abortion: induced and
spontaneous. The contraceptive needs of women may
differ, depending upon which type of abortion they
have had.
Women who have had an induced abortion - of which
there are some 26 to 53 million worldwide each year are likely to want to use family planning services to
avoid another pregnancy. Often, unwanted pregnancy
can be prevented with adequate information about,
and access to, a range of contraceptive methods.
Helping such women avoid another unwanted
pregnancy and possibly another induced abortion is
important. A woman’s reproductive health - and
perhaps her life - could be at risk if the abortion is
performed by an untrained individual in unhygienic
conditions. Unsafe abortions are estimated to
Contraception after Pregnancy
31
cause the death of between 50,000 and 100,000
women each year.
In contrast to women who have undergone an induced
abortion, women who have experienced a spontaneous
abortion, or miscarriage, may want to become
pregnant again soon.
Women who want to avoid pregnancy after induced or
spontaneous abortion have the same contraceptive
method options. E3
Slide 38
Contraceptive Need after Abortion
■ Women regain fertility 2-4 weeks after
abortion
• To avoid another pregnancy, contraception
should be initiated as soon as possible after
abortion
Slide 39
Uncomplicated First-Trimester
Abortion: Contraceptive Options
• Induced and spontaneous abortion:
any method can be used immediately
except:
- diaphragm or cervical cap: delay of 4 - 6
weeks may be necessary in some cases
32
Contraceptive Need after Abortion
The contraceptive needs of women who have had
abortions are even more immediate than those of
women who have given birth. This is because women
can conceive soon after undergoing an abortion, with
ovulation typically occurring within two to four
weeks. Therefore, if a woman wants to avoid
pregnancy, she should begin using contraception as
soon as possible after the abortion has occurred. ■
Uncomplicated First-Trimester Abortion:
Contraceptive Options
Women who have an uncomplicated first-trimester
abortion, either induced or spontaneous, can begin
using any method immediately afterwards, except
possibly diaphragms and cervical caps. Fitting these
devices may need to be delayed four to six weeks in
the cases when cervical dilation has occurred. ■
Contraception after Pregnancy
Slide 40
Uncomplicated Second-Trimester
Abortion: Contraceptive Options
Induced and spontaneous abortion: any method
can be used immediately except:
- diaphragm or cervical cap: delay 4 • 6 weeks
- IUD: delay insertion 4 - 6 weeks unless provider
is trained in insertion immediately alter abortion
- female sterilization: easier to perform if delayed
Uncomplicated Second-Trimester Abortion:
Contraceptive Options
Women who have an uncomplicated second-trimester
induced or spontaneous abortion can - like
nonbreastfeeding women who have given birth - use
most methods immediately.
There are a few exceptions:
• As with an uncomplicated first-trimester abortion,
delay fitting or use of a diaphragm or cervical cap for
four to six weeks after abortion to allow the uterus to
return to its normal size. Otherwise, an improper
anatomical fit may reduce the effectiveness of these
barrier methods.
• In addition, delay IUD insertion for four to six
weeks when the uterus returns to normal size, unless
the provider has the skill and experience to insert the
device immediately after abortion.
• Also, be aware that female sterilization may be
easier to perform if delayed until the uterus has
returned to its pre-pregnancy position and size. ■
Slide 41
Complicated Abortion:
Contraceptive Issues
Complications that can affect
contraceptive choice include:
infection
trauma to genital tract
hemorrhage, leading to severe anemia
2 or more complications can occur concurrently
Contraception after Pregnancy
Complicated Abortion: Contraceptive Issues
Complications can occur after abortion - especially
unsafe abortion - that may affect contraceptive
method options. Complications include infection,
trauma to the genital tract and hemorrhage that results
in severe anemia. We will now review these
complications individually. However, it is important
to note that a woman may experience more than one
complication concurrently, in which case the provider
would need to consider the recommendations for all
existing complications. ■
33
Slide 42
Complicated Abortion: Infection
If infection is present or suspected, delay:
- female sterilization
- IUD insertion
Follow guidelines for uncomplicated abortion for i
initiation of all other methods
Advise client to avoid intercourse until infection !
is resolved
- use condoms if intercourse is not avoided
Complicated Abortion: Infection
If an infection is present or suspected, WHO
recommends delaying female sterilization and
insertion of an IUD or levonorgestrel intrauterine
system (LNg IUS) until the infection is either fully
resolved or ruled out. If a woman chooses one of these
methods, a provider should offer an interim
contraceptive method, such as condoms, and make a
follow-up appointment or referral.
All other contraceptive methods can be used as they
would be in the case of an uncomplicated first or
second trimester abortion.
However, until the infection is resolved, the provider
should advise the patient to avoid intercourse. If
intercourse cannot be avoided, the infected woman
and her partner should use a condom as protection
against other infectious agents.
Slide 43
Complicated Abortion:
Genital Trauma
If genital trauma exists, delay:
- female sterilization, unless performed during
required surgery
- IUD insertion
- female barrier methods and spermicides
(depending on extent and location of injury)
Follow guidelines for uncomplicated abortion for
initiation of all other methods
Complicated Abortion: Genital Trauma
Genital trauma includes bums, perforations, cervical
tears or lacerations. If a woman has suffered genital
trauma as the result of an abortion, WHO
recommends delaying sterilization until the trauma is
healed. However, if abdominal surgery is required,
sterilization can be done concurrently. This is
appropriate only if a woman was counseled and gave
her informed consent for sterilization before the
abortion was performed.
WHO also recommends delaying IUD and LNg IUS
insertion until uterine perforation or other serious
trauma has healed. Meanwhile, the client should be
provided with a short-term method that she can use
while awaiting IUD insertion and make a follow-up
appointment or referral. -»
34
Contraception after Pregnancy
Finally, according to WHO, use of female barrier
methods should be delayed, depending on the extent
and location of the injury. These methods should not
be used, for example, if they will come in contact with
a significant cervical or vaginal injury. At the least,
diaphragms and cervical caps should not be fitted
until four to six weeks after abortion, when the
woman’s dilated cervix has returned to its normal size.
All other contraceptive methods can be used as they
would be in the case of an uncomplicated first or
second trimester abortion.
In cases of lower genital tract injury, the provider may
want to advise the woman to avoid intercourse until
the injury has healed. ■
Slide 44
Complicated Abortion:
Hemorrhage and Severe Anemia
> Delay methods that increase or maintain
short-term blood loss
- female sterilization
-IUD
i Follow guidelines for uncomplicated abortion
for initiation of all other methods
Complicated Abortion: Hemorrhage and
Severe Anemia
Some women undergoing unsafe abortions suffer
hemorrhage. Many women in developing countries
are anemic and hemorrhage may worsen their anemia,
making it severe.
Until severe anemia is resolved, any method that
might increase or maintain short-term blood loss
should be avoided. Thus, female sterilization and
insertion of IUDs should be delayed. Meanwhile, the
provider should offer a short-term method and make a
follow-up appointment or referral.
All other contraceptive methods can be used as they
would be in the case of an uncomplicated first or
second trimester abortion. ■
Contraception after Pregnancy
35
Slide 45
Training and Counseling Increase
Contraceptive Acceptance__________
__
Training and Counseling Increase
Contraceptive Acceptance
Having discussed contraceptive options for women
who have experienced abortion, we will now describe
how enhanced training of providers can increase
contraceptive use after abortion.
In a study in three Bolivian hospitals, training hospital
staff about the provision of services following
abortion resulted in marked increases in contraceptive
use by women after abortion. The training emphasized
prevention of future abortions and how to implement
a counseling program.
Physicians and counselors at these hospitals attended
a training course on care and contraception following
abortion. Particular emphasis was placed on training
providers to give compassionate care and to improve
their communication skills. Later, the hospitals
implemented a counseling program that provided
women with three separate opportunities to decide to
use contraception and to choose a method. Those who
were discharged prior to reaching a decision about
contraception had an additional opportunity to discuss
their options during a follow-up visit scheduled one
month after discharge.
Contraceptive counseling was almost nonexistent
before the study. Contraceptive acceptance before the
training of personnel was low. But acceptance in the
three hospitals increased soon after the training and
implementation of the counseling program.
This slide illustrates the rise in acceptance of modem
contraception (IUDs and hormonal methods) in the
three hospitals during the periods evaluated.
A similar intervention program conducted in a large
public hospital in Mexico also resulted in a
statistically significant increase in the percentage of
women who accepted a contraceptive method ->
36
Contraception after Pregnancy
following abortion (from 30 percent to 60 percent),
and in the percentage of women who received the
method at the hospital before discharge (from 34
percent to 56 percent).
Finally, a Kenyan study looked at the provision of
postabortion family planning services on a hospital
gynecological ward. The study found that provision of
these services by trained ward staff was easier, more
effective and more acceptable to clients than provision
by maternal and child health/family planning staff
either on the ward or in the family planning clinic. ■
Slide 46
Contraceptive Counseling
of Women after Abortion
Women who have undergone abortion
should be treated with respect
Counseling can help women:
• identify factors that led to abortion
• make informed choice of contraceptive method
• choose highly effective method when pregnancy
is life threatening
• become aware of emergency contraception
understand implications for future fertility
Contraceptive Counseling of Women after
Abortion
It is important that staff working with women who
have had an abortion treat these clients with respect
and not allow any negative attitudes about abortion to
affect the interaction.
Pregnancies resulting in abortion have many causes. If
an abortion was induced, counseling after the abortion
can help women identify the factors that led to the
pregnancy and, in some cases, prevent a recurrence.
Among the reasons for an unwanted pregnancy is
failure to use contraception. A woman may fail to use
contraception because she lacks access to family
planning services, contraceptive methods are
unavailable, unpleasant side effects prompt her to
discontinue use of a method, or family planning
providers fail to help her find a method that is
acceptable to her. Other unwanted pregnancies occur
when a woman has not used a method consistently and
correctly. Occasionally, an unwanted pregnancy may
occur if the contraceptive method itself fails. Finally,
an unwanted pregnancy may be the result of coerced,
nonconsensual sex.
Contraception after Pregnancy
37
A woman who has had either an induced or
spontaneous abortion and does not wish to become
pregnant again soon should make an informed choice
of a contraceptive method based on information about
method options, method characteristics, and how to
use the methods correctly. If a woman who had an
induced abortion was using a contraceptive method
when the unwanted pregnancy occurred, the provider
can help her choose another method that may be more
effective for her. Or, if the woman wants to continue
using the same method, the provider can ensure that
she knows how to use it correctly and has ready
access to it. The availability and appropriate use of
emergency contraception should also be discussed.
A woman who has had an induced abortion because
pregnancy was life-threatening needs to be counseled
about the use of highly effective methods of
contraception that will protect her from another
pregnancy and its associated health risks. Also,
women treated for abortion complications need to
understand the implications for their future fertility.
Infection resulting from an unsafe, improperly
performed induced abortion, for example, can cause
infertility. ■
Slide 47
Counseling the Partners
of Clients after Abortion
Counseling male partners of abortion clients
can increase their support for contraception
Egyptian study found that husband's support
for family planning was strongest predictor
of client’s use of contraception
Counseling male partners may also increase
awareness and use of vasectomy or male
condoms
Counseling the Partners of Clients after
Abortion
If a woman agrees, her male partner should be
included in counseling when possible. Counseling
male partners of clients can help make these men
more supportive of clients’ family planning decisions,
increasing their subsequent use of contraception. A
study conducted in six Egyptian hospitals, for
example, has shown the importance of counseling the
husbands of women following abortion.
In the study, counseling of husbands was part of an
initiative to improve clinical services for women -»
38
Contraception after Pregnancy
following abortion. The results showed that the
husbands’ support of family planning was the
strongest predictor of family planning use by clients.
Other studies have shown an association between
husbands’ support of family planning and the
initiation and continuation of contraception.
Also, counseling of male partners may increase their
awareness and use of male contraceptive methods,
such as vasectomy and condoms. El
Slide 48 Activity: Patient Scenarios
• Before you begin: Prepare copies of the patient scenarios
located in the Audience Handouts section of this module. Write
a list of all methods on a flip chart.
• Tell participants that they will now consider some common
patient scenarios related to contraception after pregnancy.
Each scenario describes the specific situation of a woman
either postpartum orfollowing abortion. The participants'
task is to develop a list of contraceptive methods appropriate
for each woman to use. It should be assumed that all of these
women are healthy and have no medical conditions (aside
from those mentioned in the scenario) that would affect their
use of contraception. At this point, it is not necessary for them
to cover counseling points.
• Divide participants into small groups. (There are enough
scenarios for seven groups.)
• Give each group a patient scenario.
• Ask participants to list the contraceptive methods
appropriate for the patient in the scenario, based on the
information they received during this session. Point to the
flip chart containing the list of methods and explain that
they should consider all of these methods.
• Allow five to 10 minutes to discuss.
• Each group should then present their scenario and their
list of appropriate methods to the entire group.
• Discuss the lists and correct any misinformation presented.
You may want to develop your own scenarios and include
situations not discussed in this module. Your scenarios might
take into account, for example, advantages or disadvantages
of various contraceptive methods, or specific client preferences. ■
Contraception after Pregnancy
39
Slide 48
Contraception after Pregnancy:
Conclusion
After pregnancy, many women want to delay or
avoid getting pregnant again
heaJh and quality » Ine
services can be highly effective
Contraception after Pregnancy: Conclusion
After pregnancy, many women want to delay or avoid
getting pregnant again. Providing family planning
information and services to women after pregnancy
can improve the health and quality of life of women
and children. It can even save their lives.
Effective contraceptive services provided after
pregnancy help women who want to delay or prevent
subsequent pregnancies choose contraceptive methods
that are safe, effective, convenient, and best meet their
short- and long-term family planning needs.
Many contraceptives can be used successfully soon
after pregnancy. Providers must learn which methods
are appropriate for each woman during this period, the
timing for initiating these methods, and how to
counsel clients accordingly. This is especially
important for clients who are breastfeeding. Because
some contraceptive methods can be initiated
immediately after delivery or abortion, providers
ideally should discuss family planning before labor
and birth, or abortion.
Meeting the contraceptive needs of women following
pregnancy is an important component of maternal and
child health services and services that provide
abortion or treatment for abortion complications.
Integrating family planning services with other health
services can be a highly effective way to reach these
women. Providers, program managers and policy
makers who look for ways to integrate services can
benefit both clients and their health-care systems. ■
40
Contraception after Pregnancy
Concluding Activity
Have two or three participants go to the Question Tree that
was posted on the wall during the Opening Activity.
• Ask them to pull questions off the tree, and read them to the
group. Ask members of the group to volunteer the answers if
they know them. If they do not know the answers you can
answer the questions for them. Be certain to correctfalse
statements.
• Congratulate the group on their new knowledge. ■
Contraception after Pregnancy
41
Contraceptive Wif
"?
———— MxMvua*
Contraception
after Pregnancy
OWaceptlw Technology Update Series produced by:
Family Health International
with financial support from the
U.S. Agency for International Development
ISBN 0-939704-65-X
Library of Congress Catalog Card Number: 00-105463
© Family Health International
October 2000
Any part of the suggested narrative and any slide used in this presentation,
hereafter referred to as material, may be copied, reproduced, distributed or
adapted, without permission from the authors or publisher, provided the
recipient of the material does not copy, reproduce, distribute or adapt this
material for commercial gain, and provided that Family Health International
is credited as the source of such information on all copies, reproductions,
distributions and adaptations of the material.
FHI is a not-for-profit organization dedicated to improving reproductive health
through research and technical assistance in contraceptive development, family
planning and the prevention of sexually transmitted diseases, including AIDS.
Family Health International
P.O. Box 13950
Research Triangle Park, North Carolina 27709 USA
Phone: 919-544-7040
FAX:
919-544-7261
Web site: http://www.fhi.org
UJH
Introduction
This training module is part of a series of slide presentations entitled
Contraceptive Technology Update Series (CTU). The CTU series, as well as
the Reproductive Health Series (RH), are developed and produced by Family
Health International (FHI) with funding from the U.S. Agency for
International Development (USAID) and assistance from other agencies
concerned with international family planning.
Other currently available topics are:
° Barrier Methods
• Client-Provider Interaction
» Emergency Contraceptive Pills
• Female and Male Sterilization
• Injectable Contraceptives
• Intrauterine Devices (IUDs)
• Introduction to Contraceptive Methods
• Lactational Amenorrhea Method
• Oral Contraceptives (OCs)
• Reproductive Health of Young Adults
• Sexually Transmitted Diseases
All modules are available in English, French and Spanish.
The purpose of these training modules is to meet the continuing educational
needs of family planning practitioners and policy-makers in developing
countries by providing the most current information available about
contraceptive technology and reproductive health. The information in the
modules is intended to be used at workshops or seminars for physicians,
nurses, pharmacists, family planning counselors, midwives, or other trained
health-care personnel. The information is appropriate for those with a basic
knowledge of and familiarity with family planning programs, methods of
contraception and reproductive health, but can also be used in preservice
training settings for medical and nursing students.
Contraception after Pregnancy
Credits
Design Team:
Instructional Designer:
Nina Frankel, MMA
Writer:
Kim Best, BS
Medical Advisor:
Irina Yacobson, MD
Technical Support:
Technical Monitor:
Nina Frankel, MMA
Graphic Artist:
Salim Khalaf, BA
Cover Designer:
Dick Hill, HillStudio
Copy Editor:
Karen Dickerson, MS
Distribution Coordinator: Hina Wall, BA
Special thanks to the following FHI staff for their assistance with the
development and review of this module:
Patricia E. Bailey, DrPH; David Grimes, MD; Nash Herndon, MA; Karen
Katz, MS, MPA; JoAnn Lewis, MPH; Thulani Magwali, MD; Hally R.
Mahler, MHS; Charles Morrison, PhD; Susan Palmore, MA; Robert Rice,
MA; Roberto Rivera, MD; Jane Schueller, MA; David Sokal, MD
Special thanks to the following collaborators for their assistance in the
review of this module:
James Shelton, MD, MPH, USAID; Jeff Spieler, MSc, USAID; Mihira
Karra, PhD, USAID; Gloria Alvarado, MD, Universidad Juarez del Estado
de Durango, Mexico; Soledad Diaz, MD, Institute Chileno de Medicina
Reproductiva (ICMER); Miriam H. Labbok, MD, MPH, USAID; Jim
McMahan, MPH, Program for International Training in Health (INTRAH/
PRIME); Sharon Rudy, PhD, Program for International Training in Health
(INTRAH/PRIME); Cathy Solter, RN, MPH, Pathfinder International;
Judith Winkler, MEd, IPAS
For invaluable assistance in the field-testing of this module, special thanks to:
Olivia McDonald, MD, FRCOG, MPH, and selected health providers in
Kingston, Jamaica; and Maureen Kuyoh and selected health providers in
Nairobi, Kenya.
This publication is based on the earlier FBI module, Postpartum Contraception.
Contraception after Pregnancy
Module
Contents
Narrative
This is a carefully researched and reviewed script designed to accompany the
slides. It is only a suggested narrative. You are strongly encouraged to use this
information to develop your own presentation. As written, the narration lasts 1 1/2
to 2 hours. If you do elect to read all or part of the suggested narrative, it is best
not to read the topic headings printed in bold type. Suggestions on how to adapt
this presentation are included later in this section.
Activities and Discussion Questions
The activities and discussion questions, found in shaded boxes throughout the
narrative, allow facilitators to create an interactive workshop rather than give a
lecture presentation. Facilitators may use some or all of the activities, as
appropriate. The more activities a facilitator includes, the more likely participants
will retain the information. This is because most people learn better and have a
greater chance of achieving behavior change when actively thinking about, rather
than just receiving, information. As with your selection of slides, conducting such
training activities requires advance preparation.
Slides
Forty-eight 35mm color slides are numbered and labeled for your use. You are
encouraged to consider the usefulness of each slide when preparing a presentation
for your audience. Slides can be deleted, added from another source, or rearranged
to highlight certain information. The slides are followed by black and white paper
copies of the information that appears on the slides. These may be used as a
tabletop flip chart if no slide projector is available. Or, the paper copies can be
used to create overhead transparencies for projection.
Audience Handouts
This section contains handouts that can be copied and distributed to the audience.
The summary fact sheet can help audience members remember key learning points.
There are also pages that the audience can use for note-taking during the presentation.
Pretest and Posttest
There is an evaluation questionnaire that can be used as a pretest and posttest. It
can be given before the session to ascertain the knowledge level of the audience.
This same instrument can also be given afterwards to assess how much the
audience learned and whether the information presented resulted in changed
attitudes. A questionnaire with the answers marked correctly is also included.
Participant Evaluation
There is an evaluation form to be given to each participant in the audience. Please
make copies and have the audience members complete the form after each training
session. You may find that the comments from the audience are useful in planning
future presentations. We would also appreciate your sending to us any suggestions
for improving the module based on your experience as a presenter. Please return
completed forms to FHI to help us revise current modules and improve future ones.
References and Resources
A list of important readings about contraceptives is provided for persons
wanting to learn more about this topic. The information in this presentation was
drawn from these sources.
Reprints of Scientific Articles
Key journal articles and book chapters are included in this section to provide more
detailed scientific information about contraception. You should read these
materials before giving your presentation. If feasible, they may be shared with
audience members.
Contraception after Pregnancy
How to Use
This Module
The following are two ways this module can be presented:
1. Presentation using slides and narrative only (requires approximately 1 1/2
to 2 hours).
2. Presentation using slides, narrative, activities and discussion questions in an
interactive session (requires approximately 2 to 2 1/2 hours).
The needs of the participants, as well as the time available and the size of the
group, are factors to consider when determining which presentation approach to
use. For groups of fewer than 25-30 participants, the interactive approach
works well. For larger groups, a simple slide presentation may be more appro
priate. However, it is important to remember that retention of information and
the impact on participant behavior and attitude will increase with interactive
approaches to learning. Regardless of which presentation method you choose,
the following steps are recommended when preparing for and conducting a
presentation or interactive training session.
Preparing for Your Session
♦ Review the suggested narrative and slides several times to become familiar
with the information.
❖ When choosing an interactive approach, review and rehearse the activities and
discussion questions. Anticipate possible questions and further discussion topics
that may arise.
♦ Read the materials in the “Reprints of Scientific Articles” section and, if
possible, review some of the materials listed in the “References and
Resources” section.
♦ Become familiar with relevant policies and procedures regarding
contraception after pregnancy in the country where you are making the
presentation or training.
♦ Adapt slides, narrative, activities and discussion questions to meet the needs
of your specific audience, considering their level of education, training and
experience. Also, consider the amount of time available for the presentation.
Add or remove slides, information, activities or discussion questions where
appropriate.
♦ Use language that you are comfortable with and that will be easily
understood by the audience. Speaking from your own notes is more
effective than reading directly from the narrative, even if not including
participatory activities. Actively engaging the audience with a combination
of slides, activities and discussion questions is optimal.
♦ Rehearse your session, preferably using the same room and equipment that
you plan to use during your presentation.
Contraception after Pregnancy
Preparing Materials for Your Session
♦ Make photocopies of the pretests, fact sheet, note-taking handouts,
evaluations and World Health Organization (WHO) eligibility criteria for
your audience.
♦
If including the activities, obtain a flip chart and prepare pages for the
activities being used.
♦ Assess which, if any, of the reprints of scientific articles would be
appropriate for your audience. Prepare summaries if necessary.
♦ Reserve a room large enough for your audience that can be darkened for
slide viewing.
♦
Obtain a 35mm slide projector. Select a place to project the image, either a
screen, white sheet or blank wall. If a slide projector is not available, use the
paper copies of the slides to make overhead transparencies or use the pages
as a tabletop flip chart.
Presentation Checklist
__ Before the presentation, give participants the pretest.
__ After the presentation, lead a group discussion. Explore the ideas
presented and examine ways in which each participant thinks he or she
can improve information about and access to contraception after
pregnancy. Identify issues that are especially important for additional
training or research.
. Give participants the posttest.
__ Ask participants to complete an evaluation form.
__ Return the evaluation forms, and any other suggestions for improving
the module, to Family Health International, CTU/RH Modules Project
Administrator.
__ Follow up with participants, if appropriate.
Contraception after Pregnancy
Ideas for
Adapting the
Module for
Different
Audiences
This training module provides basic information about contraception after
pregnancy. Presentations can be made more interesting and valuable by adding
information about local experiences and practices, especially those that apply to
your audience. In addition, some adaptation of the text or slides may be needed
to meet the specific needs, priorities, scope, resources and constraints of
different programs. If method-specific FHI modules are available, the
presentation can be customized and extended by including the more detailed
information contained in these volumes.
The amount of clinical information that is appropriate to present will depend on
the medical background and knowledge of your audience. As designed, all slides
and scripted material are appropriate for clinicians with at least a basic
familiarity with family planning. The best way to select slides for your audience
is to ask yourself: “What information does this audience need to enable them to
provide their clients with information and counseling on contraception after
pregnancy? Is the information provided too basic or too advanced for this
audience?” Answering these questions will help you prepare a presentation
appropriate for your audience.
Learning
Goals and
Objectives
Contraceptive technology is an evolving, dynamic and ever-changing subject.
To offer high-quality services and care, policy-makers, administrators and
providers of contraceptive services need to remain abreast of current
information and practices. The purpose of this module is to provide information
about contraception after pregnancy to policy-makers, managers, health-care
providers and students. The module provides a core of information that can be
adapted as needed for various audiences.
After the presentation, audience members will be able to:
♦ discuss contraception after pregnancy, which contraceptive methods
are appropriate and the timing of their initiation;
♦ discuss the main counseling points for contraception after pregnancy;
♦ discuss the importance of offering contraception after pregnancy.
Contraception after Pregnancy
^ound4z
%
o
IT MATTERS TO
n/rnvnur
imWWNG
z-s FAMILY PLANNING SAVES
fA i WOMEN’S LIVES
• FAMILY PLANNING PREVENTS : CONTRACEPTIVE USE
OOn : canavertmorethanhalfof
vli oUUU maternalhb
O@
MATERNAL DEATHS WORLDWIDE : DEATHS
I
Family Planning Directly Reduces the Number of Maternal Deaths because
it Reduces the Chance of Pregnancy and Associated Complications
X REDUCES
fel n
LOWERS
DELAYS FIRST
® HAZARDS DF
1® 1
PREGNANCY
W® CLOSELY SPACED TIM
INYOONG
IIff PREGNANCIES
Illi
WOMEN
ABORTION
INDIA ACCOUNTED FOR 15% (45,000) OF ALL MATERNAL DEATHS
(303,000) WORLDWIDE IN 2015 (ESTIMATED)*
@ MATERNAL MORTALITY IS
I
•
PROPORTION OF MATERNAL DEATHS
' lO. CM
f■I
r—I* ■ /ML AVERTED BY CONTRACEPTIVE USE'
X r-|
ill <
! P
J
FOR EACH WOMAN
I WHO DIES DURING
: CHILDBIRTH
•rlbc
w5?.3% pn
OC hnri
*
MATERNALDEATHSAVERTED :
OUsUUU BY CONTRACEPTIVE USE
A A FAMILYPLANNINGISTHEPRIMARY
WII INTERVENTION TO PREVENT MATERNAL MORTALITY
MORE SUFFER
FROM INFECTION,
IN JURY AND
DISABILITY
CONNECTED TO
PREGNANCY OR
CHILDBIRTH
ICPD SHAPING FAMILY PLANNING POLICIES
based on
RIGHTS, DIGNITY AND HEALTH
ICPD Commitments -1994
—31 Provide universal access to
family planning and sexual and
reproductive health services
and reproductive rights.
Deliver gender equality,
empowerment of women
and equal access to
education for girls.
Address the individual,
social and economic
impact of urbanization
and migration.
Support sustainable
development and address
environmental issues
associated with population
changes.
Uhf
ICPD Impact on Policies and Programmes in India
*0*^
mMJLI National Population
Policy, 2000- rights
WTlMB based approach and a
|MJJ UUy[] target free approach.
Reproductive and Child
J
1 Health Programme
k (V (phase 1 & 2) 1997,2005 J
\ multiple stakeholders;
f
| lifecycle approach;
'
\ ' quality of care;
informed choice.
National Health Mission for improved health care
delivery in urban and
rural areas.
National Health Policy (2002) equitable access to healthcare;
funding for women's health.
STRENGTHENING FAMILY PLANNING ESSENTIAL TO
ACHIEVE LARGER DEVELOPMENT GOALS
Family planning is a critical element of
social, human, economic and environmental
development.
Linkages
Access to family planning- supports women's social
and economic well-being; enables them to choose
the number and spacing of their children.
Better birth spacing - reduces the incidence of low
birth weight and poor maternal nutrition. Family
planning results in more wealth and less hunger.
Delay Childbearing - Women who are able to delay
childbearing are more likely to meet their educational
goals, obtain productive employment, increase
household income, and thus help reduce extreme poverty.
Using family planning empowers women When women are empowered and are
decision-makers in their families, they spend
more resources on their children's nutrition,
healthcare and education.
STRENGTHENING FAMILY PLANNING ESSENTIAL TO
ACHIEVE LARGER DEVELOPMENT GOALS
Linkages
Birth spacing through family planning Reduces child mortality. Children born three to
five years apart are 2.5 times more likely to
survive than children born two years apart.
Family planning saves lives Family planning allows spacing of pregnancies
and avoids unwanted pregnancies. It can delay
pregnancies in young women at increased risk of
health problems and death from early childbearing.
Family planning and reproductive health services are essential to preventing the spread of H1V/AIDS.
• Improving access to condoms can reduce the
number of infections acquired through sexual
intercourse.
• Increasing contraception use among
HIV-positive women through voluntary family
planning services can avert almost 30%
more cases of mother-to-child-transmission
than anti-retrovirals alone.
THE PATH AHEAD
POST 2015
A
Priorities
Expanding contraceptive
• A cho'ces for delaying and
q “ spacing births
Quality of care
in reproductive
health services
Focus Areas
FAMILY PLANNING - Currently
unmet need for family planning
in India is 21%; unmet need for
spacing methods is highest (26%)
among the youngest age group
(15-19 years)*.
QUALITY OF CARE Emphasis on
ids informed choice
and counselling.
ADOLESCENT HEALTH - Delaying age at marriage and age at first • v •
pregnancy are critical in India - 46% of all women are married
before the legal age of marriage, one in six women aged 15-19
have begun childbearing and 50% of maternal deaths among ||r II
girls in the same age group are due to unsafe abortions. UU II
*
STUDIES show that a large number of risk factors for
■ poor reproductive and child health are rooted in adolescence
i and yet only 15% of youth receive any education on sex or
k family planning.
•
. •
I
COMMUNITY
INVOLVEMENT
to be strengthened
for accountability of
services.
‘Source - DLHS-3 2007-08
Promoting a MULTI STAKEHOLDER
INVOLVEMENT and better
interaction between
Central and State Govt.
ON CONTRACEPTIVE METHOD MIX in
DEVELOPING COUNTRIES : SOUTH/SOUTH-EAST ASIA
evidence
The addition of 1 method available to at least half the population correlates with
an increase of 4-8 percentage points in total use of the 6 modern methods,
for example, from 40% to 44% or 48%?
dj^odern Contraceptive
^Wevalence Rate (mCPRl
WHY METHOD MIX MATTERS
International data over 27 years shows that as each
additional contraceptive method became available to
most of the population, overall modern contraceptive
use rose. In 2009 only 3.5 methods, on average, were
available to at least half the population in surveyed
countries.
Number of Modern
Methods available in
Method Mix
* Source: 'Use of modern contraception increases when more methods
become available: analysis of evidence from 1982-2009', John Ross and
John Stover, 2013
METHOD MIX SCENARIO
Source: 'FP2020 Partnership
in Progress 2013-2014'
LAM
Condom
Pill
Injectable
Implant
mo
Female
Male
Other Modern
Sterilisation Sterilisation
Methods
B®
0
fOUND^
/s M
A %>2
INVESTING IN
vniiTU
THE YOUTH DIVIDEND
0
ADOLESCENT AND YOUTH
POPULATION - INDIA*
PERCENTAGE OF POPULATION
■ Adolescent
(10-19)
AGE GROUP
■ Youth ■ Other
(15-24)
In India, as per Census 2011 ADOLESCENT
population (10-19) is 253.2 MILLION and
that of the YOUTH (15-24) is 231.9 MILLION
constituting 20.9% and 19.2% of the total
population respectively.
There has been a decline in the
proportion of adolescent population
and an increase of youth population
compared to Census 2001.
‘Source - Census 2011
r J’llt
vm ITU FAMILY PLANNING
YUUI Il & UNMET NEED
Addressing their needs will require
Trained family planning
Counsellors
Statutes on Privacy and
Confidentiality
Emphasis on Youth-Friendly
Contraceptive Services
for Informed Choice
Comprehensive Sexuality Education which is
grounded in human rights and gender equality
UH-3>ATHE LANCET
Venous thromboembolic disease and combined oral
contraceptives: results of international multicentre case-control
study
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception *
Summary
Introduction
The composition and use of oral contraceptives (OCs) have
changed since their cardiovascular side-effects were
established 20 years ago. This report describes the risk of
idiopathic venous thromboembolic (VTE) events (deep vein
thrombosis (DVT] and/or pulmonary embolism (PE]) in
association with current use of combined OCs among 1143
cases aged 20-44 and 2998 age-matched controls, as
evaluated in a hospital-based, case-control study in 21
centres in Africa, Asia, Europe, and Latin America.
OC use was associated with an increased risk of VTE in
Europe (odds ratio 4-15 (95% Cl 3-09-5-57]) and in non
European (“developing") countries (3-25 [2-59-4-08]).
Risk estimates were generally higher for DVT than for PE
but no consistent trend by certainty of diagnosis (definite,
probable, possible) was found. Increased risk was apparent
within 4 months of starting OCs, was unaffected by
duration of current episode of OC use, and had disappeared
within 3 months of stopping OCs. Relative risk estimates of
VTE associated with OC use were unaffected by age of
user, by history of hypertension (excluding hypertension in
pregnancy), or in any consistent way by smoking. However,
in both groups of countries increased body mass index
(BMI) was an independent risk factor for VTE, and 0Cassociated odds ratios were higher among those with a
BMI above 25 kg/m2 than among those with smaller BMIs.
OC-associated risk estimates were high among women in
Europe with a history of hypertension in pregnancy.
Odds ratios associated with the use of OCs containing a
third-generation progestagen were higher than those
observed with progestagens of the first (norethindrone
type) and second (norgestrel group) generation. Odds
ratios associated with first and second generation
progestagens tended to be lower, though not significantly,
when used in combination with low (<50 jig oestrogen)
rather than higher oestrogen doses. This study confirms an
association between OC use and VTE in Europe and the
developing countries, although overall risk estimates
associated with use were lower than demonstrated in most
previous studies of non-fatal idiopathic VTE.
Following
the
introduction,
in
1960,
of oral
contraceptives (OCs) into clinical practice, the first
published suggestion of an association with increased risk
of cardiovascular disease was a case-report of a 40-yearold woman who developed a pulmonary embolism (PE)
after taking Enovid.1 Subsequently, a series2-” of case
control studies have evaluated the risk of venous
thromboembolism (VTE) associated with OCs. These
studies differed in the events included—fatal or non-fatal,
idiopathic or secondary to operations, trauma, or
infection, and PE or deep-vein thrombosis (DVT) or a
mixture of both. Relative risk (RR) estimates varied
between 2 and 11 and tended to be higher when only
idiopathic cases were included. Despite being limited in
size, the few cohort studies to have investigated the risks
of V Hi associated with OC use14-” found RRs compatible
with those estimated in case-control studies.
The association between OC use and VTE requires
further evaluation for several reasons. First, almost all
previous studies took place in the 1960s and 1970s, when
OCs contained higher doses of oestrogen and progestagen
than they do today, and before prescribing practice had
changed towards the preferential use of OCs by younger
women who do not have other risk factors for
cardiovascular disease.” Some studies12-20 but not others12
have suggested that such changes in dose and
composition might reduce the risk of V I E. Second, most
studies have been done in northern Europe and the USA,
and it may not be appropriate to extrapolate the results to
populations with different incidence rates of VTE disease
and prevalence of associated risk factors. Third, part of
the apparent variability in OC-associated risk estimates
reported was due to the difficulties in diagnosing DVT
and PE accurately,21-22 particularly when the diagnosis was
based mainly on clinical grounds. Finally, only one study
included more than 100 cases’ and previous attempts to
see if the risk of VTE was modified by or interacted with
other risk factors such as smoking lacked power and
produced inconsistent results.
This paper reports on the relation between idiopathic
VTE and use of combined OCs from a multicentre,
hospital-based, case-control study of stroke, acute
myocardial infarction, and VTE in Africa, Asia, Europe,
and Latin America (including the Caribbean). Findings
on stroke and myocardial infarction will be reported
elsewhere. The principal aim of the VTE component was
to evaluate whether current OC use was associated with
increased risk of a first DVT, PE, or both in women from
Europe and from the other three regions combined.
Secondary aims were to evaluate the risk in each of the
four regions and whether risks differed among subgroups
of women, such as smokers or those overweight, and
according to type, duration, and past use of OCs.
Lancet 1995; 346:1575-82
See Editorial page 1569 and Commentary page 1570
•Writing committee, study organisation, and participants listed at
end of article
Correspondence to: Dr N R Poulter, Department of Epidemiology
and Public Health, University College London Medical School,
1-19 Torrington Place. London WC1E 6BT, UK; and Dr 0 Meirik,
Task Force on Epidemiological Research in Reproductive Health,
UNDP/UNFPA/WHO/World Bank Special Programme of Research.
Development and Research Training in Human Reproduction,
World Health Organization, 1211 Geneva 27, Switzerland
Vol 346 • December 16, 1995
1575
THE LANCET
Certainty
DVT
Definite
Probable
Possible
Other
(^470)
76-4
50-8
15-3
4-9
5-3
Developing countries
All
(n=747)
(n=159)
8T3
250
43-6
14-5
4-2
9L8
13-2
42-1
32-7
3-8
(n=84)
LAm
(n=504)
Combined
(n=1217)
82-2
867
831
488
24-8
350
310
46-2
32-7
10-9
10-8
1-2
4-8
1-2
4-6
23-6
12-7
8-2
17-8
16-9
PE
13-3
Definite
7-2
2-4
00
10-7
1-8
4-3
R-7
7.1
12-4
Probable
9-5
6-9
101
1 -7
1 -1
Possible
1-3
00
1-2
1-3
Other
2-3
0-5
00
0-8
00
1-2
Table 1: Percentage distribution of types of VTE cases by
region
Patients and methods
A more detailed description of study methods will be published
elsewhere.” This hospital-based case-control study was
undertaken in 21 centres in 17 countries subdivided into four
regions. Each centre recruited cases and controls from a variable
number of collaborating hospitals. Women were eligible as cases
if they were aged 20-44 years (15-49 in three centres), had been
admitted to a collaborating hospital between Feb 1, 1989 and Jan
31, 1993, and had a discharge diagnosis of DVT and/or PE.
Those who died within 24 h of admission, who had a history of
stroke, DVT, PE, acute myocardial infarction, or natural or
surgical menopause, or who had a recent history (within 6 weeks)
of pregnancy, major illness causing prolonged bed rest, or
surgery were excluded.
Case definitions
A monitoring system was set up in each centre to identify all
eligible cases. A review of the medical history, examination, and
investigations allowed classification of these cases as definite,
probable, possible, or other in respect of DVT or PE. These four
categories, based on signs and symptoms and the results of
investigations, were developed using published data2”4 and the
opinions of four senior clinicians with a special interest in
thrombotic disease during a pilot review of a sample of cases.
This allowed a standardised classification of DVT and PE
diagnosed with varying degrees of certainty. All cases considered
Characteristic*
Developing countries
Europe
Cases
(n=433)
Cases
(n=710)
All
(n=1044)
Current
(n=356)
Past
(n=484)
Never
(n=204)
340
6-6
32-7
7-3
24 0
4-7
331
69
24-4
50
Controls, by 0C use
All
Current
(n=1954) (n=239)
32-4
30-0
7-3
6-3
25-3$
5-1
23-8$
50
21-3
59-3
19-4
304
60-3
9-3
19-3
39-4
41-3
24-5
37-1
38-5
60-5
Controls, byOCuse
Age (yr) (mean and SO)
32-5
70
32-2
7-0
BMI (kg/nP) (mean and SO)
25-lt
4-8
23-9f
4-5
291
6-5
23-4
40
26-3
59-6
14-1
29-5
56-9
13'6
40-2
51-7
8-2
76-7
Number of live births
1-2
»3
Married/stable union
by their physician to have had a DVT or PE were categorised as:
Definite DVT when confirmed by venography, duplex scanning,
or radioisotope studies.
Probable DVT in the absence of any other likely cause for the
presenting signs and symptoms, when swelling and induration
extended above the knee, and when any two of the following four
indicators were present: increased superficial temperature of
affected limb, tenderness on palpation, superficial venous
engorgement, compatible doppler ultrasound investigation. In
the absence of any other likely cause for the presenting signs or
symptoms, when swelling and induration were only apparent in
the calf, and when three of the four indicators were present, a
case was also accepted as “probable DVT”.
Possible DVT in the absence of any other likely cause for the
presenting signs and symptoms and when swelling occurred
below the knee together with any two of the four clinical
indicators listed under “probable DVT”—or, in the absence of
any other likely cause for the presenting signs and symptoms,
when marked swelling and induration extended above the knee.
Other DVT for all remaining DVT cases.
Definite PE if either ventilation/perfusion (V/Q) scan,
angiography or necropsy findings were so reported.
Probable PE if no good evidence of an alternative diagnosis was
reported and any three of the following signs, symptoms, or
investigations were reported: haemoptysis, pleuritic chest pain of
sudden onset, breathlessness of sudden onset, ECG pattern of
right heart strain, chest X-ray changes compatible with the
diagnosis, arterial gas concentrations compatible with the
diagnosis, syncope with tachypnoea, other compatible clinical
signs (eg, pleural rub, raised jugular venous pressure, added
heart sounds), compatible ventilation (V) or perfusion (Q) scan.
Possible PE if no good evidence of an alternative diagnosis was
reported and any two of the nine signs, symptoms or
investigations above were reported, or where evidence of an
alternative diagnosis was reported and three of the nine were
reported.
Other PE for the remaining PE cases.
Those cases diagnosed as having both DVT and PE were
treated in the analyses under the event diagnosed with more
certainty. When diagnostic certainty was the same the case was
classified as DVT on the basis that PE is usually preceded by
DVT.
76-7
28-6
73-6
35-8
71-9
Education beyond secondary
480
33-7
691
19-6
61-1
20-3
Cunent smoker
40-7
37-1
36-8
41-1
27-9
At least limit of alcohol/wk
Setfroported history of
Hypertension (HBP)
Hypertension (HIP)
Diabetes mellitus
Rheumatic heart disease
30-7
30-8
35-7
29-7
250
15$
3-4
20-2
16-7
4-6
Past
(n=647)
Never
(n=1068)
34-2
6-6
31-8
7-7
23-6
5-4
24-3
4-5
23-5
5-1
18-4
49-4
32-2
65 7
5-3
380
56-7
68-8
37-5
33-7
28-8
52-4
20-5
15-9
22-8
20-9
5-9
15-5
7-3
16-6
2-7____
6-9
40
6-2
0-8
3-5
20
6-6
5-2
9-8
12-5
11-2
11-3
10-6
5-9
14-5
16-2
8-1
17
10
2-8
0-6
1-9
1-2
20
1-9
1-6
0-3
0-4
1-5
0-1
00
0-2
0-6
0-0
5-4
4-4
40
2-5
3-8
3-7
4-5
60
4-6
4-4
FamBy history of CVD
•Notes: current smoker=at least 1 cigarette in 3 months before illness which caused hospital admission (cases) or before admission (controls); hypertension (HBP)=that
diagnosed before current OC use and excluding hypertension in pregnancy: hypertension (HIP)=blood pressure problems in pregnancy including pre-eclampsia and eclampsia:
CVD=stroke and/or AMI in either parent before age 60 and 50, respectively.
f3 cases and 6 controls (2 never. 3 past, 1 current users) with unknown BMI.
$236 cases and 678 controls (346 never. 238 past, 94 current users) with unknown BMI.
Table 2: Characteristics of VTE cases and controls (% unless otherwise shown)
1576
4-6
17-0
1-4
1-2
Vol 346 • December 16, 1995
THE LANCET
Region
Relative to non-users
Europe
Developing countries
Africa
Asia
Latin America
Type of user
Cases
Controls
Crude OR (95% Cl)
Non-user
User
Non-user
User
Non-user
User
Non-user
User
Non-user
User
168
265
505
205
107
46
56
26
342
133
688
356
1715
239
338
69
228
16
1149
154
100
3-95 (2-96-5-28) •
100
3-25 (2-59-4-08)
100
2-14 (1-37-3-36)
100
7-30 (3-39-15-72)
1-00
3-37 (2-53-4’48)
Relative to never users
Europe
Never
53
204
100
Past
484
115
0-84 (0-58-1-24)
Current
265
356
3-53 (2-39-5-21)
Never
320
1068
100
Past
185
647
l-00(0-80-l-24)f
Current
205
3-25 (2-54-4-14)1
239
Adjusted ORs *4-15 (3-09-5-57) adjusted for hypertension in pregnancy (HIP)
excluding 3 cases (users) and 1 control (normser) with unknown HIP; fO-93
(0-74-1-17) and $3-08 (2-40-3-96) adjusted for number of live births (0,1-2, 3 or
more).
Table 3: Odds ratios of VTE in relation to combined OC use by
region, with non-user or never user as reference group
Developing countries
Controls
For each case, an attempt was made to recruit 3 female controls
matched by 5 year age band (20-24, 25-29, and so on).
Controls were subject to the same exclusion criteria as cases and
their date of admission to hospital had to be between 2 weeks
before and 4 months after that of the case. Controls had to be
admitted to the same hospital as the case, with one of 27
diagnoses considered to have no association with OC use.” The
procedure for control selection was defined in a field protocol
produced by each collaborating centre. Methods were adapted to
local conditions but were similar. Controls were recruited by
paying regular visits, in a predefined random order to those
wards which accepted women with any of the control diagnoses.
Interviews
All cases and controls were interviewed in a standard way using
the same questionnaire, which established eligibility and included
self reported details of medical and reproductive history,
estimated height and weight, past and present contraceptive use,
use of other medications, family history of stroke and/or
myocardial infarction (in either parent before age 60 and 50,
respectively), details of smoking habit, use of alcohol and coffee,
level of education and social class (assessed locally on the basis of
housing, income, education and/or occupation). A current
smoker was defined as a woman who had smoked at least 1
cigarette in the 3 months before the illness which had caused her
to be admitted to hospital and a moderate smoker was one who
smoked at least 10 cigarettes per day. Information about
hypertension was obtained from responses to three questions—
had the respondent ever (excluding pregnancy) had high blood
pressure; had she had a blood pressure problem, including pre
eclampsia or eclampsia, during but not necessarily confined to
pregnancy (hypertension in pregnancy); or had medication to
control blood pressure been used in the 3 months before the
illness which had caused hospital admission. Identification of OC
type was assisted by showing the women samples or pictures of
locally available OC packets. Details of cases’ diagnostic
information were recorded.
This report assesses risk of DVT and PE associated with the
current use of combined OCs compared with those not currently
Ate(yr)
Europe*
using OCs (non-users). Women who currently used progestagenonly contraceptives (oral, 24 cases and 68 controls; injectable, 14
cases and 38 control) and women who used combined injectable
contraceptives (7 cases and 12 controls) were not considered to
be current OC users but were classified as never or past OC users
on the basis of any previous use of combined OCs. Current OC
use was defined as having taken an OC at any time during the 3
months before the event (cases) or before hospital admission
(controls). Women who were current OC users but who did not
know the type (3 cases, 4 controls) were considered to be current
users of combined OCs and were included in all analyses, except
those referring to type of preparation. In the analyses non-users
(past and never users combined) were preferred to never users as
the reference group because VTE risk is associated with current
rather than ever use of OCs, and never users, particularly in
Europe, represent a small and possibly atypical subset of nonusers.
11 women (1*0%) eligible as VTE cases were not interviewed
because they were too ill or had died before the questionnaire
could be completed. For them, the closest available relative or
friend was interviewed as a proxy.” Complete participation of
eligible cases and controls was reported by 13 or the 21
collaborating centres and in the remaining 8 centres only 26
cases of stroke, myocardial infarction, or VTE, and 53 controls
refused participation.
Statistics
Conditional logistic regression analyses were done with EGRET
software?6 Adjusted models included confounding variables,
selected in a sequential manner, which resulted in at least a 5%
change in the estimate of the risk associated with OC use.” The
patterns of OC use, risk factors for VTE, and confounding
differed in Europe from those in the developing regions and it
was decided a priori to examine patterns of OC-associated risk
separately in these two groups of countries. Consequently,
separate models were fitted for Europe and for the developing
regions combined. Trends of odds ratios (OR) in stratified
analyses were assessed by a test for linear trend in the log odds
ratios.” 95% confidence intervals were calculated and are shown
after ORs in parentheses.
Results
Of the 1011 cases of DVT and 206 cases of PE recruited,
42% and 25%, respectively, had their diagnoses
confirmed by definitive investigations. The proportion of
VTE cases which were diagnosed as DVTs ranged from
92% in Africa to 76% in Europe. Over 80% of both case
types were classified as definite or probable, and the data
were insufficient to allow classification as definite,
probable, or possible VTE on 71 women (table 1). The
data presented in all subsequent tables (except table 8)
and the analyses exclude these 71 (“other”) cases and
their controls. 1 additional case whose OC status was not
known was also excluded from all subsequent analyses, as
were 2 further cases because no matched controls had
been recruited.
To the remaining 1143 cases a total of 2998 matched
controls were recruited with an average of 2-4 controls per
case in Europe and 2-8 in the developing countries.
56-9% and 56-0% of controls in Europe and the
developing countries, respectively, had a diagnosis of
Developing countries
Non-user
<50 pg
>50 pg
All users_____
Non-user
<50 pg
All users_____
>50 pg
4 32 (2-88-6-49) 3-29(1’95-5-54)
4 00(2-74-5-83)
1-00
3-23 (2-32-4-49) 3-55 (2 36-5-34)
3-35 (2-54-4-41)
(132/198)
(44/75)
(176/273)
(253/942)
(93/117)
(149/180)
(56/63)
3-93 (2-30-6-73) 5-19 (2-78-9-67) 4-39(2-75-7-00)
>35
1-00
3 05 (2-02-4-59)
2-50(1-46-4-28) 3-88 (2-14-7 -05)
(42/37)
(42/43)
(84/80)
t252/773)
(28/35)
(28/24)
(56/59)
•Adjusted for HIP and excluding 2 cases and 3 controls with unknown OC type, and 3 cases (users) and 1 control (non-user) with unknown HIP. fNo of cases/no of controls.
Table 4: Odds ratios of VTE In relation to combined OC use by age group (non-user as reference) and oestrogen dose
<35
1-00
(64/333)1
1-00
(104/354)
V.4 340 ■ December 16, 1995
1577
’
T
OCum
NorHJsers
Users
Non-users
Users
Europe
Noo-HIP_______________
1-00
(128/597)
4-04 (2-97-5-49) (229/335)
Developing countries
Non-HIP
HIP
100
(431/1513)
3-34 (2-63-4 26) (183/211)
BMI S25 kg/m’
1-83 (1-20-2-81) (40/90)
9 24 (4-93-17-31) (33/21)
BMI >25 kg/m’
1-00
(88/445)
3.91 (2-76-5-54) (157/257)
1-52 (1-06-2-19) (80/237)
7-01 (4-65-10-59) (103/98)
Non-smokers
>10 cigarettes per day’
1-00
(186/780)
3-33(2-33-4-75) (85/112)
Non-smokers
BMI ss25 kg/m:
HIP
1-23(0-91-1-67) (68/191)
3-11 (1-72-5-63) (21/27)
BMI >25 kg/m:
1-63(1-21-2-19) (149/351)
9 44 (5-30-16-82) (54/33)
>10 cigarettes per day*
1-00
(100/431)
1-21 (0-82-1-78) (52/185)
1-00
(424/1438)
1-08 (0-66-1-76) (24/76)
Non-users
4-05 (2-83-5-78) (155/225)
5-53 (3-58-8-54) (76/83)
3-46 (2-69-4-44) (173/189)
2-23 (0-98-5 06) (10/19)
Users_______
Exclusions: HIP analyses in Europe 3 cases (users) and 1 control (norsuser) and In developing countries 7 cases (6 non-users. 1 user) and 12 controls (11 non-users. 1 user) with
unknown HIP: BMl analysis in Europe adjusted (or HIP. excluding unknown HIP and 2 cases (users) and 6 controls (5 non-users. 1 user) with unknown BMI: BMI analysis in
developing countries excluding 236 cases (170 non-users. 66 users) and 678 controls (5B4 non-users. 94 users) with unknown BMI: smoking analysis adjusted for HIP in Europe.
risk estimates among smokers of <10 cigarettes/day not shown.
Table 5: Odds ratios of VTE In relation to combined OC use by history of hypertension In pregnancy, BMI, and smoking
)
-x
J
J
trauma, acute infection, appendicitis, or bone or joint
disorders.
Characteristics of cases and controls in Europe and the
developing countries are shown in table 2. Whilst mean
body mass index (BMI) was greater among cases than
controls in both groups of countries, cases and controls
had similar mean age, number of live births and marital
status, and reported similar rates of cigarette and alcohol
consumption. In Europe, but not the developing
countries, cases were less likely than controls to have gone
beyond secondary education and were more likely to have
reported hypertension in pregnancy. Cases in the
developing countries but not in Europe were more likely
than controls to have had at least 1 live birth, to give a
history of high blood pressure (detected before the
current episode of OC use and not during pregnancy),
and rheumatic heart disease and to report a family history
of premature cardiovascular disease. Rates of cigarette
smoking and alcohol intake were higher and mean
number of live births was lower in Europe than in the
developing countries.
In Europe and the developing countries the crude ORs
of having a VTE were 2-32 (0-71-7-63) and 33-0
(4-26-255-6), respectively, for a history of rheumatic
heart disease; 2-70 (1-68^-34) and 4-61 (2-77-7-68) for
a BMI over 30 kg/m! compared with =S2O kg/m1; 2-65
(1-73—4-05) and 3-81 (2-75-5-27) for a history of varicose
veins; 2-59 (0-46-14-55) and 1-22 (0-11-13-53) for
moderate smoking (510 cigarettes per day) compared
with non-smokers; 1-66 (1-20-2-29) and 1-16 (0-89-1-52)
for hypertension in pregnancy; and 0-95 (0-56-1-62) and
1-82 (1-25-2-65) for a history of high blood pressure.
Table 2 also includes the characteristics of controls who
were current users, past users, or had never used OCs.
Never users were less likely than current and past users to
be regular alcohol drinkers. In Europe, never users were
less likely than current and past users to be regular
alcohol drinkers. In Europe, never users were less likely to
be current smokers and to have reached secondary
education than current or past users (the converse trend
was apparent in the developing countries), and were more
likely to have a history of rheumatic heart disease and/or
high blood pressure (outside pregnancy).
The prevalence of current OC use among cases and
controls in Europe (61-2% and 34-1%, respectively) was
higher than in the developing countries (28-9% and
12-2%). The overall ORs of having a VTE associated with
current OC use compared with non-users (never and
past-users combined) and with never users were
significantly raised in all four regions (table 3). In Europe,
HIP confounded the relation whereas in the developing
1578
countries combined no confounders were identified. ORs
among past users compared with never users were less
than unity. Consequently, risk estimates associated with
current OC use were smaller when compared with never
rather than non-users.
No consistent patterns of risk estimates among OC
users were apparent according to age or-oestrogen dose
(table 4). No consistent or important effect of a history of
high blood pressure (excluding pregnancy) on OCassociated risk was apparent either. However, compared
with non-users of OCs with no history of hypertension in
pregnancy, risk estimates associated with OC use in
Europe but not the developing countries were higher
among those who also reported hypertension in
pregnancy (table 5). This effect was similar in younger
and older women and was also apparent among women
whose blood pressure problems were confined to
Months
Duration of episode of
use of current 0C
Non-users
Users
<4
4-12
13-24
25-48
49-72
73-96
>96
(Trend)
Time since last use of
combined OCs
Never user
Past user
>144
109-144
73-108
37-72
25-36
13-24
>3-12
Current user
>2-3
>1-2
Europe)___________
Developing countries
100(168/687)
1-00(505/1715)
4 39 (2-50-7-71) (29/35)
5-63 (3 58-8-84) (61/62)
4-20 (2-41-7-33) (32/52)
3-51 (2-28-5 41) (54/88)
3-88 (2-26-6'64) (33/45)
3 56 (1-78-6-34) (20/32)
4-04 (2-32-7-04) (33/41)
(P=0-25)__________
4-77 (2-96-7-84) (40/33)
3-75 (2-54-5-53) (57/58)
2-23(1-31-3-79) (23/40)
3-31 (2-08-5-25) (38/47)
2-31(1-26-4-25) (18/29)
1-61 (0-67-3-84) (8/18)
6 23(2-78-13-96) (19/9)
(p=0-26)
1-00(53/204)
1-00(320/1068)
0-77 (0-46-1 28) (33/129) 1-17 (0-82-1-65) (63/162)
0-81 (0 42-1-57) (15/59) 0 66 (0-38-1-12) (18/87)
0-72 (0-37-1-42) (14/75) 0-76 (0-45-1-28) (21/85i
128 (0-72-2-27) (25/82)
0-96 (0-62-1-50) (31/108)
1-07 (0-43-2-65) (8/32)
0-81 (0-38-1-72) (9/40)
0-71 (0-28-1-82) (6/37)
0-64 (0-33-1-24) (11/60)
0-70 (0-29-1-67) (7/45)
1-21 (0-72-2-03) (22/69)
0 65 (0 07-5-79) (1/9)
2-67 (1-05-6-81) (8/11)
1-37 (0-57-3-29) (8/25)
1-73 (0-77-3-85) (10/20)
4-05 (2-70-6-07) (256/322) 3-28 (2-53-4-26) (187/208)
(Trend)
(p=0-65)
(p=O-52)
ORs adjusted for HIP for Europe, duration of use. and for live-birth categories (0.1-2.
3 or more) for developing countries, time since last use. Exclusions: duration of use.
unknown HIP and 1 control with unknown duration of current episode of OC use,
Europe, and 2 cases and 5 controls with unknown duration of current episode of OC
use. developing countries: time since last use unknown. 7 cases, and 25 controls.
Europe, and 10 cases and 36 controls, developing countries.
•Including cases and controls taking OCs at time of even and at time of hospital
admission respectively.
Table 6: Odds ratios of VTE In relation to duration and recency
of combined OC use
Vol 346 • December 16, 1995
THE LANCET
Progestagen
Curopet_____________________________________
Developing countries
First
Second
Third
Other
Total
>50 ng
<50 ng_______________
(10/18)
4-05(1-92-8-54)
3-37 (1-44-7-93)
3-61 (2-53-5-13)
(102/163)
3-83(2-44-6-02)
7-36 (4-20-12-90)
(53/51)
15-70(3-90-63-15)
3-83(1-74-8-45)
(9/3)
4-24 (3-07-5-87)
(174/235)
3-96(2-66-5-90)
<s° m
000
(0/8)
(103/137)
2-79 (2-08-3-75)
12-23 (4-76-31-43)
(18/7)
(0/0)
3-02 (2-28-4-00)
(121/152)
(19/27)
(54/73)
(0/0)
(13/18)
(86/118)
>50
3-62 (2-04-6-44)
3-79 (2-47-5-82)
3 44 (1-25-9-46)
3-64 (2-60-5-09)
(26/31)
(50/47)
(0/0)
(8/9)
(84/87)
OR relative to 168/687 and 505/1716 norKisers In Europe and the developing countries, respectively.
•first generation. ethynodiol diacetate, lynoestrenol. norethisterone, norethisterone acetate, and norethynodret; Second generation, norgestrel, levonorgestrel, and norgestrienone;
Third generation, desogestrel, gestodene, and norgestimate; Others, chlormadinone acetate and cyproterone acetate.
tExcluding 2 cases and 3 controls with unknown OC type, and 3 cases (users) and 1 control (nonaiser) with unknown HIP, and adjusted for HIP.
Table 7: Adjusted odds ratios of VTE In relation to current use of combined OCs by progestagen type and oestrogen dose
Certainty
Europe*
_
Definite
Probable
Possible
Other
Developing countries
VTE_________
DVT
PE__________
VTE_________
DVT
PE__________
3-88(2-70-5-57)
4-90 (2-79-8-61)
3-96(1-17-13-34)
2-22 (0-90-5-47)
4-13 (2-79-6-10)
5-61(2-69-11-69)
4-32(1-09-17-15)
2-39 (0-79-7-19)
2-54 (0-95-6-77)
4 02 (1-69-9-54)
2-82 (0-21-38-14)
1-91 (0-40-9-02)
4-75 (3-11-7-26)
3-11 (2-26-4-28)
1-95(1-14-3-31)
6 04 (1-89-19-30)
4 60 (2-97-7-14)
3-47 (2-42-4-99)
1-95(1-14-3-36)
. .t
7-11 (1-38-36-75)
2-09 (1-06-4-15)
1-73 (0-10-30-76)
- -t
Table 8: Odds ratios (95% Cl) of VTE, DVT and PE by certainty of diagnosis In relation to combined OC use
pregnancy. A significant trend of increasing risk of VTE
was apparent with increasing BMI (p<0-001).
In univariate analyses smoking less than 10 cigarettes
per day was not associated with an increased risk of VTE
compared with non-smokers, whereas those smoking at
least 10 cigarettes per day had ORs of 2-59 (0-46-14 55)
in Europe and 1'22 (O-H-13'53) in the developing
countries.
Duration of current episode of use (table 6) and
lifetime duration of use (data not shown) did not affect
the risk estimates of VTE. Increased ORs were fully
realised within 4 months of starting OCs and had resolved
within 3 months of stopping (table 6). However, risk
estimates among those who, although classified by study
definition as current users, had stopped OC use between
1 and 3 months before the event precipitating admission
(cases) or hospital admission (controls) were only
marginally increased in Europe 1 month after stopping,
whereas in the developing countries they remained
significantly increased up to 3 months after stopping OCs.
However, in both groups of countries ORs were highest
among those current users who had continued OC use at
least until 1 month before the event (cases) or admission
(controls).
No appreciable difference in risk estimates was
apparent between users of OCs containing low or higher
oestrogen doses (table 7). However, risk estimates among
users of first and second generation progestagens were
slightly larger when used in combination with a higher
rather than with a low oestrogen dose. Risk estimates
among users of OCs containing third-generation
progestagens, which are only used in com- ination with
lower dose oestrogen, were higher than for users of first
and second generation progestagens, and the highest OR
was observed among users of other progestagens
combined with a low oestrogen dose in Europe.
Table 8 shows OC-associated risks by certainty of
diagnosis. ORs tended to be higher for DVT than PE and
in the developing countries, but not in Europe, tended to
fall with decreasing certainty.
Discussion
In this study, the largest so far to have investigated VTE
and OC use, current use of combined OCs was associated
with a significantly increased risk of a first event of VTE,
with ORs of about 4-2 in Europe and 3-3 in the
Vol 346 • December 16. 1995
developing countries. These estimates are at the lower end
of the range of ORs previously reported for idiopathic
V I E associated with OCs.
Risk factors
The only identifiable confounder was a history of
hypertension in pregnancy, in Europe; there was no
confounding in the developing countries. The only other
risk factors for VTE identified were a history of varicose
veins, a BMI of over 25 kg/m1, and a history of rheumatic
heart disease.
The finding that ORs associated with current OC use
were unaffected by age accords with previous studies.’'’
Similarly, as reported in most41’ but not all studies,10 no
striking association between smoking and risk of VTE
emerged.
In univariate analyses, a history of high blood pressure
detected outside pregnancy and before OC use (so not in
any causal pathway linking OCs and VIE) was a
significant risk factor for VTE in the developing countries
only; hypertension in pregnancy was significantly
associated with VTE in Europe. In Europe OC users who
had a history of hypertension in pregnancy had high ORs
compared with non-OC users without such a history
(table 5). OCs induce adverse effects on platelets and the
coagulation
and
fibrinolytic
systems,"
including
antithrombin III activity,11 which are similar to the
changes seen in pregnancy-induced hypertension. Perhaps
women with a history of pregnancy-induced hypertension
are prone to OC-induced VTE, and these two problems
may share some pathophysiological characteristics.
BMI was calculated from self-reported height and
weight, not routine measurements on admission to
hospital. Nevertheless, compared with lighter women who
did not use OCs, those with a BMI over 25 kg/m1
consistently had higher risk estimates associated with OC
use than those with lower BMIs, a finding which has not
been observed in most previous studies?'" An increasing
risk of VTE associated with increasing levels of BMI over
20 kg/m1 was demonstrated in univariate analyses. Women
with higher BMIs who wish to use OCs should be made
aware of these increased risks. Because BMIs were
reported during the hospital stay this study could not
differentiate women whose BMI was high before OC use
from those whose BMI rose while they were using an OC,
and these data may have implications for those whose
1579
THELAN^hT
body weight rises significantly whilst they are on an OC.
The rapid onset of OC-induced risk, the lack of effect
of duration of OC use, and the rapid reduction of risk
after stopping OCs are in keeping with previous
reports.'111” Table 6 suggests that while there was no
increased risk after stopping OCs for 3 months, very little,
if any, increase in risk of VIE was maintained among
those women in Europe who had stopped OCs for
between 1 and 3 months, whilst in the developing
countries some increase was apparent for up to 3 months
after stopping. These levels of risk were, however, less
than among those who continued OC use at least until 1
month before their illness or admission to hospital. Risk
estimates associated with current OC use, as defined in
this study, therefore tend to be smaller than those among
women who used OCs in the month before the event or
admission.
The lack of any significant oestrogen dose effect on risk
estimates is in keeping with the results of some11 but not
a]po.M.» studies. Whilst risk estimates among users of OCs
containing first and second generation progestagens
tended to be lower when used in combination with low
rather than higher oestrogen doses, the overall impact of
low-dose OCs was affected by the higher risk estimates
associated with third-generation progestagens, which were
always used in combination with low dose oestrogens.
Third-generation progestagens are reported to have less
adverse metabolic effects than with first and second
generation compounds14-" so, despite controversial
pharmacokinetic data14"
(which
have
not been
confirmed),41'41 the high ORs observed with OCs
containing these progestagens were unexpected. These
differences are further evaluated in the accompanying
paper.41
Possible sources of bias
One review44 has questioned all previously published data
on OC use and V l b, pointing to difficulties in accurate
diagnosis1'-11 and the potential for bias. In the WHO study
the proportion of DVT cases definitively investigated was
greater among OC users than non-users in Europe (60%
and 48%, respectively) and in the developing countries
(45% and 34%). This difference may have resulted from
investigation bias; and hence led to an overestimate of
risk. Alternatively, such differences would result if OC use
is associated with more severe cases1-4 and if such cases are
more likely to be investigated. This difference in
investigation rate between OC users and non-users was
not apparent for PE.
Angiography was rarely used at the time of the study in
routine clinical practice to diagnose PE, and V/Q
scanning was the most frequently used ’‘definitive’’
diagnostic tool in this study. V/Q scanning has been
shown to be an insensitive test for detecting
angiographically proven PE41 but the conclusion was that
combining a V/Q scan with a strong clinical suspicion of
PE (as in the present study) was a “sound diagnostic
strategy”.45 Furthermore, the relative stability of the ORs
in various categories of diagnostic certainty of both DVT
and PE in Europe suggests that differing diagnostic
accuracy could not account for the observed association
between V I E and OC use. The pattern of reducing ORs
with decreasing certainty of case diagnosis seen in the
developing countries might be expected because the
association between OC use and risk of VTE was diluted
by the inclusion of more non-VTE case. The ratio of
1580
DVT to PE cases recruited varied in the four regions,
with a very high ratio in Africa. Limited diagnostic
facilities in the African centres may have made PE more
difficult to confirm than DVT.
Hospital-based
controls
have
advantages
and
disadvantages. 44 Hospital controls with one of a range of
permissible diagnoses were preferred to community
controls because of the potential for bias due to
differential use of hospital services by cases and
community controls. In most collaborating centres it was
not feasible to recruit community controls. In the 5
centres where both types of control were recruited ORs
for VTE associated with OC use were similar with
neighbourhood controls (4-07 [2-74-6-04]) and with
hospital controls (3-71 [2-41-5-72]).
Th t possibility of bias introduced by proxy respondents
for cases has been evaluated. Information from proxies
proved reliable for recent or current events, and
specifically for current OC use. Less than 1% of VTE
cases required a proxy, and the impact on risk estimates
for VTE due to misclassification of OC use by these
repondents is less than 0-1 %.2’
Cases were restricted to those who survived for at least
24 hours after hospital admission and, because of the
difficulties of accurate diagnoses, other cases may not
have been included. In addition, subclinical or less
obvious VTE were almost certainly underrepresented in
this study, although one study suggested that OCs may be
associated with more severe VTE events.14 Thus, a
potential limitation of this study is that the cases may not
have been representative of all cases of VTE.
One strength of the study was the detailed information
on OC exposure. Cases and controls were interviewed
under similar conditions in hospital and they were not
aware of the main objective of the study. The use of
samples and pictures of locally available OCs and the
comprehensive history of OC use permitted detailed
evaluation of all types and patterns of OC use. Biased
recall of OC exposure could have occurred as a result of
the relative severity of DVT and PE compared with some
of the control diagnoses and the knowledge of
cardiovascular risk associated with OCs. This potential
bias cannot be explored because contraceptive history was
not validated against medical records or prescriptions. An
exaggerated estimate of risk associated with OC use could
also have occurred if controls stopped taking OCs in
anticipation of hospital admission so that they were
classified as past and not current OC users. However,
only 0-6% and 0-2% of non-user controls from Europe
and the developing countries, respectively, had stopped
OC use more than 3 months before hospital admission for
this reason, and were thus classified as past rather than
current users.
Balancing risks
The risk estimates associated with OC use in this study,
whilst high, are lower than those reported in most
previous studies of idiopathic VTE. In assessing the
public and individual health importance of different
adverse cardiovascular events associated with the use of
OCs, age-specific incidence and morbidity and mortality
need to be taken into account. In the European centres,
VTE was the most common adverse cardiovascular event
associated with OC use while in the developing countries
stroke predominated. However, the case-fatality rate for
VTE events, which occurred in this study with similar
Vol 346 • December 16, 1995
THE LANCET
incidence throughout the reproductive years, is only
1-2% whereas 5-10% of strokes and up to 50% of acute
myocardial infarctions, which occur predominantly in
older women, are fatal.”
Information on these risks of cardiovascular disease
associated with OC use should be made available to and
be considered by women choosing a contraceptive
method. In addition, the estimated risks of VTE in
association with current use of OCs should be considered
in the context of both the risks and benefits of alternative
forms of contraception and the effects of OCs on quality
of life, on protection against certain forms of neoplasia,*7
and ultimately on overall morbidity and mortality.
Study organisation
Writing committee: N R Poulter, C L Chang, T M M Farley, O Meirik,
and M G Marmot.
Centres and principal investigators: Myriam Dcbert-Ribciro, Escola
Paulista de Medicina, Sao Paulo, Brazil; Ernesto Medina, Escuela de
Salud Publica, Universidad de Chile, Santiago, and Jorge Artigas, Escuela
de Medicina, Valparaiso, Chile: Shen He, National Research Institute for
Family Planning, Beijing, Zhong Yu Hui, Sichuan Family Planning
Research Institute, Chengdu, and Zhang De-Wei and Zhao Wcijin,
Shanghai Institute for Planned Parenthood Research, Shanghai, People’s
Republic of China; Oscar Rojas, Facultad de Salud, Universidad del
Valle, Cali, Colombia; Martin Vessey, University of Oxford, UK; Lothar
Heinemann, Zentrum fur Epidemiologie und Gesundheitsforschung,
Berlin, Germany; Stuart Dorman and Suzanne Ho, Chinese University of
Hong Kong, Hong Kong; Gyorgy Bartfai, Albert Szent-Gyorgyi Medical
University, Szeged, Hungary; J Kisjanto, Faculty of Medicine, University
of Indonesia, Jakarta, Indonesia; Rainford Wilks, Tropical Metabolism
Research Unit, University of the West Indies, Kingston, Jamaica; Reuben
Agwanda, Kenya Medical Research Institute, Nairobi, Kenya; Rossario
Ruiz, Grupo Interuniversitario Mexican© de Investigation Epidemiologies
en Salud Reproductiva, Durango, Mexico; Mateja Kozuh-Novak,
University Institute of Public Health, Ljubljana, Slovenia; Nikom
Dusitsin, Pramuan Virutamasen, and Kammant Phanthumchinda,
Chulalongkorn Hospital, and Supom Koetsawang, and Mance PiyaAnant, Siriraj Family Health Research Center, Siriraj Hospital, Bangkok,
Thailand; Jasenka Demirovic and Karen Belkic, School of Medicine,
University of Belgrade, Belgrade, Federal Republic of Yugoslavia;
W S Mwandila and C M Mutale, University Teaching Hospital, Lusaka,
Zambia; and Jonathon Matenga and Adrian Wilson, University of
Zimbabwe, Harare, Zimbabwe.
Study design and monitoring: N R Poulter and M G Marmot, University
College London Medical School, London, UK; M P Vessey, University of
Oxford, UK; D Petitti, Kaiser Permanente, Pasadena, CA, USA; and
T M M Farley, S Hoick, O Meirik, and Steering Committee of Task Force
on Epidemiological Research in Reproductive Health. UNDP/UNFPA/
WHO/World Bank Special Programme of Research, Development and
Research Training in Human Reproduction, WHO, Geneva, Switzerland.
Study and data coordination: N R Poulter (study coordinator and
principal investigator), M G Marmot (principal investigator), C L Chang
(statistician and data manager), S Lawley (data processor), S Smith (data
processor), M Shipley (statistical adviser), Department of Epidemiology
and Public Health, University College London Medical School, London,
UK.
Publications advisory committee:] Olsen, Danish Epidemiology Science
Centre, University of Aarhus, Denmark; M Thorogood, London School of
Hygiene and Tropical Medicine, London, UK; and J Kelaghan, National
Institutes of Health, Bethesda, MD, USA.
Acknowledgment
UNDP/UNFPA/WHO/World Bank Special Programme of Research,
Development and Research Training in Human Reproduction funded this
study.
References
I Jordan WM. Pulmonary embolism. Lancet 1961; ii: 1146-47.
2 Royal College of General Practitioners. Oral contraception and
thromboembolic disease. J R Coll Gen Pract 1967; 13: 267-79.
3 Vessey MP, Doll R. Investigation of relation between use of oral
contraceptives and thromboembolic disease. BMJ 1968; ii: 199-205.
4 Vessey MP, Doll R. Investigation of relation beween use of oral
contraceptives and thromboembolic disease: a further report. BMJ
1969; ii: 651-57,
5 Sartwell PE, Masi AT, Arthes FDG, Green GR, Smith HE. Thrombo
embolism and oral contraceptives: an epidemiologic case-control studv.
Am J Epidemiol 1969; 90: 365-80.
Xfol 340 • December 16, 1995
6 Vessey MP, Doll R, Fairburn AE, Glober G. Postoperative
thromboembolism and the use of oral contraceptives. BMJ 1970; iii:
123-26.
7 Bottiger LE, Westerholm B. Oral contraceptives and thromboembolic
disease. Acta Med Scand 1971; 190: 455-63.
8 Green GR, Sartwell PE. Oral contraceptive use in patients with
thromboembolism following surgery, trauma or infection. Am J Public
Health 1972; 69: 680-85.
9 Boston Collaborative Drug Surveillance Program. Oral contraceptives
and venous thromboembolic disease, surgically confirmed gallbladder
disease, and breast tumours. Lancet 1973; i: 1399-404.
10 Petitti DB, Winerd J, Pellegrin F, Ramcharan S. Oral contraceptives,
smoking and other factors in relation to risk of venous thromboembolic
disease. Am J Epidemiol 1978; 108: 480-85.
11 Maguire M, Tonascia J, Sartwell PE, Stolley PD, Tockman MS.
Increased risk of thrombosis due to oral contraceptives: a further
report. Am J Epidemiol 1979; 110: 188-95.
12 Helmrich SP, Rosenberg L, Kaufman DW, Strom B, Shapiro S. Venous
thromboembolism in relation to oral contraceptive use. Obstet Gynecol
1987;69:91-95.
13 Thorogood M, Mann J, Murphy M, Vessey M. Risk factors for fatal
venous thromboembolism in young women: a case control study. IntJ
Epidemiol 1992; 21: 48-52.
14 Ramcharan S, Pelligrin FA, Ray R, Hsu JP. The Walnut Creek
Contraceptive Study (Center Popul Res Monog 81-564). Washington,
DC: National Institutes of Health, 1981.
15 Porter JB, Hunter JR, Danielson DA, Jick H, Stergachis A. Oral
contraceptives and non-fatal vascular disease: recent experience. Obstet
Gynecol 1982; 59: 299-302.
16 Porter JB, Hunter JR, Jick H, Stergachis A. Oral contraceptives and
non-fatal vascular disease. Obstet Gynecol 1985; 66: 1-4.
17 Vessey MP, Mant D, Smith A, Yeates D. Oral contraceptives and
venous thromboembolism: findings in a large prospective study. BMJ
1986 ; 292: 526.
18 Royal College of General Practitioners. Oral contraceptives, venous
thrombosis and varicose veins. J R Coll Gen Pract 1978; 28: 393-99.
19 Thorogood M, Vessey MP. Trends in use of oral contraceptives in
Britain. BrJFam Piann 1990; 16: 41-53.
20 Gerstman BB, Piper JM, Freiman JP, et al. Oral contraceptive
oestrogen and progestin potencies and the incidence of deep venous
thromboembolism. Ini J Epidemiol 1990; 4: 931-36.
21 Weils PS, Hirsh J, Anderson DR, et al. Accuracy of clinical assessment
of deep-vein thrombosis. Lancet 1995; 345: 1326-30.
22 Poulose KP, Reba RC, Gilday DL, Deland FH, Wagner HN. Diagnosis
of pulmonary embolism: a correlative study of the clinical, scan, and
angiographic findings. BMJ 1970; iii: 67-71.
23 World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. A multinational case
control study of cardiovascular disease and steroid hormone
contraceptives: description and validation of methods. J Clin Epidemiol
(in press).
24 Nicolaides AN, Irving D. Clinical factors and the risk of deep venous
thrombosis. In: Nicolaides AN, ed. Thromboembolism. London:
MTP, 1995: 193-204.
25 World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. Reliability of data from
proxy respondents in a case-control study of cardiovascular disease and
oral contraceptive use. J Epidemiol Comm Health (in press).
26 Epidemiological graphics, estimation and testing package analysis
module (PECAN). Seattle: SERC & CYTEL, 1991.
27 Maldonado G, Greenland S. Simulation study of confoundcr-selection
strategies. Am J Epidemiol 1993; 138: 923-36.
28 Breslow NE, DAy NE, Statitical methods in cancer research I: analysis
of case-cont rol studies. Lyon: International Agency for Research on
Cancer, 1980: 192-246.
29 Lawson DH, Davidson JF, Jick H. Oral contraceptive use and venous
thromboembolism: absence of an effect of smoking. BMJ 1977; ii: 729-30.
30 Poller L. Oral contraceptives, blood clotting and thrombosis. Br Med
Bull 1978; 34: 151-56.
31 Wessler S, Giter SN, Wans LS, Pasternack BS. Estrogen-containing
oral contraceptive agents: a basis for their thrombogenicity. JAMA
1976; 236:2179-82.
32 Vessey MP. Female hormones and vascular disease: an epidemiological
overview. BrJFam Piann 1980; 6: 1-12.
33 Stadel BV. Oral contraceptives and cardiovascular disease. N Engl J
Med 1981; 305: 612-18, 672-77.
34 Inman WHW, Vessey MP, Westerholm B, Engelund A.
Thromboembolic disease and the steroidal content or oral
contraceptives: a report to the Committee on Safety of Drugs. BMJ
1970; ii: 203-09.
35 Stolley PD, Tonascia JA, Tockman MS, Sartwell PE, Rutledge AH,
Jacobs MP. Thrombosis with low-estrogen oral contraceptives. Am J
Epidemiol 1975; 102: 197-208.
1581
THE LANCET
Perlman JA, Rusel-Briefel R, Ezzati T, Lieberknccht G. Oral glucose
tolerance and the potency of contraceptive progestins. J Chron Dis
1985; 38:857-64.
La Rosa JC. The varying effcts of progestins on lipid levels and
cardiovascular disease. Am J Obstet Gynecol 1988; 158: 1623-29.
38
Godsland IF, Crook D, Simpson R, et al. The effects of different
formulations of oral contraceptive agents on lipid and carbohydrate
metabolism. N EnglJ Med 1990; 323r 1375-81.
39
Kuhl H, Jung-Hoffman C, Heidi F. Alterations in the scrum levels of
gestodene and SHBG during 12 cycles of treatment with 30
micrograms ethinylestradiol and 75 micrograms gestodene.
Contraception 1988; 38: 477-86.
40
Kuhl H, Jung-Hoffman C, Heidt F. Serum levels of 3 keto-desogestrel
and SHBG during 12 cycles of treatment with 30 micrograms
ethinylestradiol and 150 micrograms desogestrel. Contraception 1988;
38: 381-90.
41
Kuhnz W, Humpe! M, Schutt B, Louton T, Steinberg B, Gansau C.
Relative bioavailability of ethinyl estradiol from two different oral
contraceptive formulations after single oral administration to 18
women in an intraindividual cross-over design. Horrn Res 1990; 33:
36
37
Humpel M, Tau ber U, Kuhnz W, et al. Protein binding of active
ingredients and comparison of scrum ethinyl estradiol, sex hormonebinding globulin, corticosteroid-binding globulin, and cortisol levels in
women using a combination of gestodene/ethin yl estradiol (Femovan)
or a combination of desogestrel/ethinyl estradiol (Marvelon) and single
dose ethinyl estradiol bioequivalencc from both oral contraceptives.
Am J Obstet Gynecol 1990; 163 (1 part 2): 329-33.
43 World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. Effect of different
progestagens in low oestrogen oral contraceptives on venous
thromboembolic disease. Lancet 1995; 346: 1582-88.
44
Realini JP, Goldzieher JW. Oral contraceptives and cardiovascular
disease: a critique of the epidemiologic studies. Am J Obstet Gynecol
1985; 152 (6 pan 2): 729-98.
45
The PIOPED Investigators. Value of the ventilation/perfusion scan in
acute pulmonary embolism. JAMA 1990; 263: 2753-59.
46
Wacholder S, McLaughlin JK. Silverman DT, Mandel JS. Selection of
controls in case-control studies. Am J Epidemiol 1992; 135: 1019-50.
47
Vessey MP. The Jephcott lecture 1989: an overview of the benefits and
risks of combined oral contraceptives. In: Mann RD, ed. Oral
contraceptives and breast cancer. Lancaster: Parthenon, 1990: 121-32.
42
Effect of different progestagens in low oestrogen oral
contraceptives on venous thromboembolic disease
World Health Organization Collaborative Study of Cardiovascular Disease and Steroid Hormone Contraception*
Summary
A multinational hospital-based case-control study of the
risk of venous thromboembolic disease associated with
combined oral contraceptives (OCs) done in 1989-93
prompted a separate inquiry comparing the risk of venous
thromboembolism (VTE) associated with low oestrogen
(<35 pg ethinyloestradiol) OCs containing levonorgestrel
with risks in low oestrogen preparations containing the
third-generation progestagens desogestrel or gestodene.
This analysis of data from 9 countries, involved 769 cases
and 1979 age matched hospital controls and. in one
centre, 246 community controls matched on age and
general practice.
137 cases and 203 controls were current users of
levonorgestrel (odds ratio [OR with 95% confidence
interval] 3-5 [2-6—4-7]), with non-users as the reference;
35 cases and 28 controls were current users of
desogestrel (9-1 [4-9-170]), and 36 cases and 28
controls were current users of gestodene (9-1 [4-9-16-7]).
The ratios of these risks, compared with levonorgestrel,
were 2-6 (1-4-4-8) for both products separately. Risk
estimates adjusted for body mass index (BMI) were 3 4.
7-3, and 10-2 for levonorgestrel, desogestrel, and
gestodene, respectively, compared with non-users, and 2-2
and 3-0 for desogestrel and gestodene, respectively,
compared with levonorgestrel. 48 (68%) cases and 48
•Participants and study organisation listed in companion paper
(Lancet 1995; 346:1575-82); writing committee listed at end of
this article
Correspondence to: Dr 0 Meirik, Task Force on Epidemiological
Research in Reproductive Health, UNDP/UNFRA/WHO/World Bank
Special Programme of Research. Development and Research
Training in Human Reproduction, World Health Organization.
1211 Geneva 27, Switzerland and Or N R Poulter, Department of
Epidemiology and Public Health. University College London Medical
School. 1-19 Torrington Place. London WC1E 6BT, UK
1582
(86%) controls exposed to desogestrel or gestodene were
from the UK (Oxford region). In this centre risk estimates
compared with non-users, adjusted for BMI, were 2-6, 5-3,
and 5-7 for levonorgestrel, desogestrel. and gestodene,
respectively.
Current users of low oestrogen dose combined OCs
containing desogestrel or gestodene appear to be at higher
risk of VTE than users of combined OCs containing
levonorgestrel. The possibility that these unexpected
results on a secondary study objective are due to chance,
bias, or residual confounding cannot be excluded entirely
and the results need to be confirmed by independent
studies. They are at variance with the apparently more
favourable metabolic effects of the newer progestagens.
Whether the new progestagens are associated with lower
risk of arterial disease (stroke and myocardial infarction)
must be evaluated further.
Lancet 1995; 346:1582-88
See Editorial page 1569 and Commentary page 1570
Introduction
The WHO Collaborative Study of Cardiovascular Disease
and Steroid Hormone Contraception investigated the
risks of venous thromboembolic disease, myocardial
infarction, and stroke associated with use of combined
oral contraceptives (OCs). The first report to be
published covered venous thromboembolism (VTE).'
This analysis, following a predefined plan, revealed a 3-4fold increased risk among OC users compared with nonusers and a further doubling or tripling of this risk among
users of third-generation OCs. This unexpected
observation, based on 71 exposed cases, prompted a
detailed analysis of risk of VTE associated with the use of
low oestrogen (<35 pg ethinyloestradiol) OCs containing
levonorgestrel or two of the newer third-generation
progestagens, desogestrel and gestodene. The study also
showed that users of low oestrogen OCs containing
Vol 346 • December 16, 1995
■■■
THE LANCET
COOTtt,
Cases
Total
Controls
Non-users
Users
Type of OC*
Ge__
De
25
12
13(52%)
2
2
5
g
28
11(39%)
0
2
215
156
59 (27%)
8
0
16
12?
20
106(84%)
3
2
40
19
3
12
7 (37%)
24
52
28(54%)
0
0
10
133
93
40(30%)
0
11
3
13
6
7(54%)
0
57
158
101(64%)
20
40
28
Total
769
397
372
35
36
137
(100%) (52%)
(48%)
(5%)
(18%)
(5%)
*De=desogestrel. Ge=gestodene. Lng=levonorgestrel. Other=other combined OCs.
Table 1: OC use among VTE cases and hospital controls
Brazil
Chile
Colombia
Germany
Hong Kong
Hungary
Jamaica
Thailand
UK (Oxford region)
progestagens other than first, second, or third generation
had more than 4 times higher risk than users of low
oestrogen dose OCs containing levonorgestrel, though
this observation was based on small numbers. An
assessment of the risk associated with the use of such
products is also presented here.
Patients and methods
This analysis is limited to those participating centres where there
were any cases or controls who were current users of OCs
containing third-generation progestagens. This subset of 10 of
the 21 centres in nine countries provided 829 VTE cases, 2135
age and hospital matched controls, and 506 community controls.
Details of the study methods, case definitions, and selection of
hospital controls and the characteristics of cases and controls are
described elsewhere.M
Community controls
In four centres where this was feasible community controls were
identified in addition to hospital controls. In the Oxford centre
(hospitals in the Oxford region, UK), up to 2 community
controls were selected for each case by referring to the records of
the general practice (GP) with which the case was registered. 6
age-matched women not related to the case were identified as
potential controls by random selection from the practice’s age
and sex register or from the alphabetical list of registered
patients. After the practitioner’s consent had been obtained the
first 2 women selected were contacted by letter, seeking
agreement to an interview. The letter was followed by a
telephone call. If no contact could be made, or there was no
response to a follow-up letter, the next woman identified from
the practice records was contacted in the same way. These
controls were interviewed at home within 4 months of the date of
the case’s admission to hospital and by the same people who did
the case and hospital control interviews. Permission to select
controls from the register was not granted by the general
practitioner for 6 cases (4%), and permission to contact selected
women was not granted for 8 of 924 potential controls (1%). Of
491 women sent letters requesting an interview, 32 had moved,
69 refused, and no reply was received from 110 (non-response
rate 39%). The non-response rates among potential controls
matched to cases aged 20-24, 25-29, 30-34, or 35-44 years were
52%, 44%, 31%, and 27%, respectively. No GP-based controls
were interviewed for 18 cases (11%), representing 8%, 20%,
15%, and 6% of cases in the four age strata, respectively.
Community controls were also interviewed in Colombia,
Hungary, and Jamaica but the identification and interviewing of
such women proved difficult and expensive and not all
subcentres recruited community controls. In Hungary
recruitment of such controls was stopped after the first year. In
Hungary and Jamaica no community control or matched case
used desogestrel or gestodene. In Colombia community controls
were recruited for 98 cases; 65 did not use OCs, and 1,10 and
22 used gestodene, levonorgestrel, or other OCs, respectively.
Vol 346 • December 16, 1995
Total
Non-users
Users
Type of OC*
De__ Ge
74
84
584
251
57
155
375
39
360
56
73
519
115
52
127
326
37
214
18(24%)
11(13%)
65(11%)
136(54%)
5(9%)
28(18%)
49(13%)
2(5%)
146(41%)
2
0
0
0
0
0
0
0
26
1
j
1
0
0
0
0
22
10
9
25
53
O
11
17
0
76
37
82
3
^7
32
2
22
1979
(100%)
1519
(77%)
460
(23%)
28
(1%)
28
(1%)
203
(10%)
201
(10%)
Other
0
35
61
2
18
28
3
13
164
(21%)
Other
5
The corresponding figures among the community controls were
162 and 0, 9, and 25. Thus these community controls
contributed little information on risks associated with desogestrel
or gestodene and they were not included in further analyses.
Classification of OC types and referent groups
Current users of third-generation progestagens were subdivided
according to product (desogestrel, gestodene, and norgestimate)
and desogestrel users were further classified by ethinyloestradiol
dose (20 pg or 30 pg). Non-users (past and never users
combined) were the primary referent group, but risks associated
with use of third-generation OCs were also compared with those
among current users of levonorgestrel OCs in combination with
<35 pg ethinyloestradiol. This type of OC was the most
commonly used by cases (36%) and hospital (44%) and
community (45%) controls, and is the most common combined
OC world wide? Users of levonorgestrel in combination with
more than 35 pg ethinyloestradiol and users of combined OCs
containing other progestagens were classified as a separate group.
1 case and 2 controls used norgestimate (250 pg, with 30 pg
ethinyloestradiol) and these were reclassified with the
levonorgestrel group. Norgestimate is rapidly metabolised to
levonorgestrel or its metabolites.4,5 Also the number of users was
too small to provide useful estimates of risk.
Modelling risks and adjustment for confounding
Conditional logistic regression models were fitted using LogXactTurbo (Cytel Software Corp, Cambridge, MA) that performs
both asymptotic* and exact inference.’ Risk estimates are
presented with 95% confidence intervals in parentheses.
Variables considered a priori as potential confounders were body
mass index (BMI), live births, alcohol consumption, smoking,
and history of hypertension (detected before current episode of
OC use and not during pregnancy), hypertension in pregnancy,
diabetes, and varicose veins. These variables were included
sequentially in the models if they changed the estimate of the
odds ratio associated with the use of levonorgestrel, desogestrel,
or gestodene by more than 5%.’
Results
Distribution of cases by centre
829 VTE cases were identified in 10 collaborating centres
from nine countries, almost half being either in Colombia
(28%) or the UK (21%). 42% were classified as definite
VTE, 42% probable, and 9% possible. In the other 54
(7%) cases the information was insufficient for
classification, and they were excluded from further
analyses. Also excluded were 1 case in the non-Oxford
centres and 2 in the Oxford region whose type of OC was
not known and 3 cases in the non-Oxford centres for
whom no controls were identified. The remaining 769
cases were matched with 1979 hospital controls and 246
GP-based controls with known OC type. Among cases
1583
(
Department of
Epidemiology and
Biostatistics, McGill
University, Montreal,
Canada
Michael A Lewis,
assistant professor
Walter O Spitzer,
Strathcona professor of
preventive medicine
Centre for Epidemiology
and Health Research,
Zepernick/Berlin, Germany
Lothar A J Heinemann,
director
Charing Cross and
Westminster Medical
School, London
Kenneth D MacRae,
reader in medical statistics
regulatory agencies and prominent discussion in the
medical and lay press. The project incorporated three
matched case-control studies with virtually identical
methods 11 for which the exposure factor of particular
interest was use of third generation oral contraceptives.
The outcomes for the three studies were venous
thromboembolism,
myocardial
infarction,
and
ischaemic (thrombotic) stroke. Results of the study
testing an association between the third generation
products and venous thromboembolism are reported
in the accompanying article.’ We report here the
initial results of the case-control study assessing and
contrasting the relation between second and third
generation oral contraceptives and myocardial infarc
tion in young women. Case recruitment continues.
Identification and quantification of publicity bias will
be incorporated in an amended protocol if extended
field work proves feasible in the United Kingdom.
Subjects and methods
Potsdam Institute of
Pharmacoepidemiology
and Technology
Assessment (PIPTA),
Potsdam, Germany
Rudolph Bruppacher,
- senior epidemiologist
London School of Hygiene
^Band Tropical Medicine,
London
Margaret Thorogood,
senior lecturer
The subjects were women aged 16-45 who were
recruited in 16 centres in five countries (Austria,
France, Germany, Switzerland, and the United
Kingdom). In this paper we include results for all five
countries as well as for the United Kingdom and
Germany alone to permit concurrent evaluation of our
results on venous thromboembolism and myocardial
infarction. An average of three controls was matched to
each case; at least one control was from a hospital and
at least one from the community. We matched for age
in five year age bands. The cases of myocardial
infarction (International Classification of Diseases
code 410) met the criteria of the World Health
Organisation.4 Controls were identified and inter
viewed within four months of the myocardial
infarction of the index case. Current use of oral
contraceptives was defined as use within three months
before the event for a case, the date of admission for a
hospital control, and the date of interview for a
community control. The field work, beginning with
feasibility and pilot projects, started in July 1991, and
case recruitment continued until November 1995.
Recruitment of controls for this report continued until
15 November 1995.
We assessed current use of third generation oral
contraceptives containing low doses of ethinyl
oestradiol (usually 30 p.g or 20 p.g) and one of two
progestogens, gestodene or desogestrel. Second
generation oral contraceptives (the main reference
group) are other low dose ethinyl oestradiol prepara
Table I—Odds ratios" for risk of myocardial infarction for current use of different types of
oral contraceptives: principal results of transnational study
Comparison
All cases (n=153)
All controls (n=498)
Third generation vsecond generation products
Third generation products v no current use
Second generation products vno current use
Hospital controls (n=210)
Third generation vsecond generation products
Third generation products vno current use
Second generation products vno current use
Community controls (n=288)
Third generation vsecond generation products
Third generation products vno current use
Second generation products v no current use
Odds ratio
(95% confidence
interval)
P value
No exposed
cases; No
exposed
controls
0*36(0*1 to 1>2)
1-1 (0-4 to 3-4)
3-1 (1-5 to 6-3)
01
0-9
0-003
63'
6; 34
23; 45
0-91 (0-2 to 46)
19 (0-4 to 8-7)
2-0 (0-8 to 4-9)
0-9
0-4
0-1
6; 11
6; 11
23; 26
0-25 (0.1 to 1-0)
0-9 (0-3 to 3-0)
3-5 (1-5 to 8-6)
0-05
0-8
0005
6; 23
6; 23
23; 19
Adjusted for centre, age, body mass index, smoking, alcohol intake, and duration of exposure to oral
contraceptives before current contraceptive.
BMJ
volume 312
13 January 1996
tions (under 50 p.g) with progestogens introduced to
the market earlier. We report unmatched odds ratios
with 95% confidence intervals and their P values. We
combined community controls and hospital controls as
our main reference group. Odds ratios were calculated
by unconditional logistic regression to adjust for the
potential confounders listed in the footnote to the
table, of which current smoking was deemed to be the
most important. We estimated the effect of smoking
adjusted for use of oral contraceptives. Matched
analyses were done as a sensitivity check and to
determine whether overmatching may have occurred.
Further details on methods have been published' ’ or
are available from us. We used a general plan and
operational procedures virtually identical with those of
the WHO’s study group4 on oral contraceptives to
facilitate comparison of the results of that project with
those of the transnational project.
We have outlined the method we used to calculate
the number of lives “saved” by switching from second
to third generation oral contraceptives in the accom
panying paper.'
Results
We enrolled 153 cases of myocardial infarction (11 of
them fatal) and 498 controls. Eighty two cases were
identified in the United Kingdom, 47 in Germany, five
in Switzerland, six in Austria, and 13 in France. When
we compared current use of third generation with
current use of second generation products as risk
factors for myocardial infarction in all 651 women the
odds ratio was 0-36 (0-1 to 1-2) (P=0-1; table 1). When
current use of third generation products was compared
with no previous use of oral contraceptives the odds
ratio was 0-3 (01 to 10) (P=0-06). When we excluded
the three countries with small sample sizes (Austria,
Switzerland, and France) the estimates and confidence
intervals became 0-45 (01 to 1-8) (P=0-26) for use of
third generation v second generation products; 1 -0 (0-3
to 3-5) (P=0-96) for current use of third generation
products v no current use of oral contraception; and
2-2 (1-0 to 5-0) (P=0 07) for current use of second
generation products v no current use.
When we compared women who currently used
third generation oral contraceptives with those who
currently used second generation products we con
sidered the controls as two groups matched for hospital
or community. The odds ratio was 0-91 (0-2 to 4-7)
(P=0-9) for cases and hospital controls and 0-25 (0-06
to 10) (P=0 05) for cases and community controls.
In matched analyses the odds ratio was 0-40 (0-1 to
1-6) (P=0-19) in the comparison of third v second
generation oral contraceptives. The odds ratio for
second generation products v no use, was 3-1 (1-4 to
6-8) (P=0-01) and for third generation products v no
current use 1-2 (0-34 to 4-4) (P=0-76). When current
smoking was adjusted for use of oral contraceptives the
odds ratio for the risk of myocardial infarction was 10-1
(5-7 to 17-9) (PcO-OOl) among our study subjects.
Among women who used third generation products the
crude odds ratio for current smoking was 31 (0-5
to 19-8) (P=0-23). Among those who used second
generation products it was 1T1 (3-0to40-2) (PcO-OOl),
and for women who were not current users of oral
contraceptives the equivalent risk estimate was 7-7 (4-0
to 14-7) (P<0 001).
Given the prevalence of use of third and second
generation preparations in the controls of this study,
the observed odds ratio of 0-36 (0-1 to 1-2) for third
compared with second generation products is con
sistent with a switch from second to third generation
oral contraceptives, resulting in 12 fewer deaths from
acute myocardial infarction per year in England and
Wales. The confidence interval for the number of
80
THE LAN.
Oxford region
Non-Oxford centres
No
Age (yr) (mean [SO])
BMI (kg/m) (mean [SD])
Uve births
0
lor 2
3 or more
Medical history
Hypertension
Hypertension in pregnancy
Varicose veins
Rheumatic heart disease
Smoking
Non-smoker
<10 cigarettes/day
>10 cigarettes/day
_________
Cases
Controls_____
_________
611__________ 1619_____
32-8 (71)
33-2 (7-1)
Cases_____
Hospital controls
GP-based controls
158______
360
30 0(7-1)
246
31-6(70)
23-6 (4-8)
23-1 (3-9)
25-1 (4-9)
23-7 (4-7)
30-7 (7-1)
25-2 (5 0)
201
48-3
31-6
24-3
44-7
31-1
42-4
39-2
18-4
478
38-9
13-3
37-8
45-5
16-7
61
14-4
17-3
1-6
4-6
11-4
4-9
01
5-7
17-1
5-7
0
1-9
10-3
5-3
0
20
17-5
6-5
0
75-2
13-6
111
75-3
55-7
8-2
361
60-3
11-1
286
76-4
8-1
15-4
10-7
Alcohol consumption
8-1
11-2
9-5
24-9
23-3
Never
44-3
41-7
481
67-9
69-3
Occasionally
38-2
37-5
34-8
5-9
6-6
Regular. «13 units/wk
9-3
9-5
7-6
1-3
0-8
Regular. >13 units/wk
3-7________
3-6
1-9
4-7
6-4
Family history of premature CVD
BMI unknown for 76 cases (12%) and 241 controls (15%) in the non-Oxford centres and 1 case, 1 hospital, and 1 GP-based control in the Oxford region: hypertension in pregnancy
not stated for 1 case and 1 control in the nonOxford centres and 1 case in the Oxford region: rheumatic heart disease unknown for 1 control in the non-Oxford centres, smoking
unknown for 1 case in the nonOxford centres; alcohol consumption unknown for 1 case and 4 controls in the non-Oxford centres and 3 hospital controls in the Oxford region; family
history of stroke and/or AMI unknown for 6 cases and 18 controls in the non-Oxford centres and 3 hospital and 4 GP-based controls in the Oxiord region
Table 2: Characteristics of cases and controls
and hospital and GP-based controls the proportions of
OC users were 48-4%, 23-2%, and 32-5%, respectively.
Patterns of OC use among cases and hospital controls
are shown in table 1. The proportion of women who used
desogestrel or gestodene was greater among cases than
controls in all nine countries except Hungary where the
only user of a third-generation OC was a case taking
norgestimate (included with levonorgestrel in table 1).
Among women
using
levonorgestrel
OCs,
the
distributions of OC type were similar in cases and
controls; 10%, 49%, and 2% of controls used
levonorgestrel 125, 150, and 250 pg, respectively, with 30
pg ethinyloestradiol, and 39% used phasic preparations.
Among users of other OC types, 50% used levonorgestrel
with
50
pg
ethinyloestradiol
(including
phasic
preparations),
37%
norethisterone
OCs
with
ethinyloestradiol or mestranol, and 13% other OC
combinations.
Most cases and controls using desogestrel or gestodene
were from the Oxford region and risk estimates were
computed separately for this centre.
Characteristics of cases and controls (table 2)
In both the non-Oxford centres and the Oxford region,
cases had higher BMI and more live births and were more
likely to have reported a history of hypertension outside
or during pregnancy than hospital controls. In the Oxford
region, cases were more likely to be smokers and less
likely to have a family history of premature cardiovascular
disease. GP-based controls in the Oxford region were
older than hospital controls, were less likely to have
extreme BMI values, had more live births, reported more
hypertension in pregnancy, and were less likely to smoke.
In the non-Oxford centres cases were more likely than
controls to have a history of varicose veins and rheumatic
heart disease.
77% of GP-based controls were married or in stable
union compared with 68% of hospital controls in the
] 584
Oxford region (p=0018), with a particular excess of
married women in the 25-29 year age group (90% vs
68%). The prevalence of OC use among GP-based
controls was lower (33%) than among hospital controls
(41%), but among users, 33% of hospital and 38% of
GP-based controls used desogestrel or gestodene.
Risks of VTE according to progestagen
Hospital controls (table 3)—In all centres risks among
users of desogestrel and gestodene were 2-6 times higher
than those for levonorgestrel users, and adjustment for the
main confounder (BMI) did not change risk estimates
substantially; nor did the omission of women with
unknown BMI. Additional adjustment for alcohol
consumption had only a small impact. Among users of
desogestrel in combination with 30 pg ethinyloestradiol
unadjusted risk estimates relative to non-users were 7-6
(3-9-14-7) based on 27 cases and 27 controls, and 38-2
(4-5-325) based on 8 cases and 1 control in combination
with 20 pg ethinyloestradiol. In the non-Oxford centres
risk estimates for desogestrel and gestodene users
compared with non-users were higher than they were in
the Oxford region. After adjustment for BMI and a
history of varicose veins, the risk estimates for desogestrel
and gestodene were similar. In the Oxford region,
adjustment for BMI decreased the risks for both
desogestrel and gestodene, with little change in risk for
levonorgestrel users. Further adjustment for alcohol and
hypertension in pregnancy increased the risk estimates for
all three products, with little impact on the ratio of risk
estimates.
GP-based controls—Risk estimates associated with OC
use for the 140 VTE cases with GP-based controls were
higher than the corresponding risks with hospital controls
for all OC types. Due to lack of controls there were
proportionately more cases omitted from the analysis for
gestodene users (20%) than for levonorgestrel (10%) or
desogestrel (4%) users. After adjustment for BMI and
Vol 346 • December 16, 1995
THE LANCET
Type of user
Cases
Controls
397
35
36
1519
28
28
56
203
201
460
Unadjusted
OR______
Adjusted for BMI*
P____
OR_________ P___
Adjusted for BMI and additional factors!
OR
P____________
7-3(3-9-13-8)
10-2 (5-4-19-2)
8-6 (5-3-14-2)
3-4 (2-5-46)
3-7 (28-5-0)
4-0 (3-1-5 0)
8-3 (4-3-15-9)
10-5 (5-5-20-0)
9-4 (5-6-15-6)
3-4 (2-5-4-7)
3 8 (26-5-2)
4-1 (3-2-52)
AR centres
compared with non-users
Desogestrel
Gestodene
Desogestrel or gestodene
Levonorgestrel
Other
All 0C users
Compared with levonorgestrel users
Desogestrel
Gestodene
Desogestrel or gestodene
Levonorgestrel
137
164
372
91 (4-9-170)
9-1 (4-9-16-7)
9-1 (5-6-14-7)
3-5 (2-6-47)
4-0 (3-0-52)
4-1 (3-3-51)
26(1-4-4*8)
26(1-4-4-8)
26(1-6-4-2)
1________
0-003
0-003
<0-001
2-2(1-2-41)
36(1-8-5-8)
2-6(1-6-4 3)
1
0-015
0-001
<0-001
2 4 (1-3-46)
0-007
3-1 (16-5-9)
0-001
2-7 (1-6-46) <0-001
1_________
d
m
1305
340
251 (2-9-215)
311 (3-7-262)
19-1 (2-1-173)
Desogestrel
18-9(6 6-54 5)
21-0(7-3-60-2)
6
12-5 (4-8-32-9)
16
Gestodene
Q
15-2 (6-3-36-4)
20-1(7-8-51-9)
20-7 (7-9-53-7)
23
Desogestrel or gestodene
127
4-0 (27-5-8)
3-8 (2-7-S-3)
3-7 (2-5-53)
97
Levonorgestrel
3-9 (2-8-S-4)
4-0 (2-9-56)
179
4-1 (31-5-5)
151
Other
314
4-2 (3 3-5-4)
4 4 (3-3-56)
271
4-1 (31-5-5)
All 0C users
Compared with levonorgestrel users
4-8 (0-5-43-4) 016
8-2 (1-0-68-8) 0-054
6-8 (0-8-58-5) 0079
Desogestrel
3-3(l-2-8-8)
0018
5-2 (1-8-15-1) 0-003
5-3 (1-8-15-5) 0002
Gestodene
5-2 (2-0-13-7) 0-001
40(1-6-9-7)
5-5 (2-1-14-4) 0001
0002
Desogestrel or gestodene
1
1
1_________
Levonorgestrel_______________
Oxford region (hospital controls)
compared with non-users
57
214
Nonuser
5-3 (2-5-10-9)
7-3 (3-3-16-0)
6 4 (3-1-13-0)
28
26
Desogestrel
5-7 (2-5-13-2)
6 2 (2-7-14-7)
6-1 (2-7-13-9)
20
22
Gestodene
5 4 (2-9-10-2)
6-8 (3-5-13-4)
6-3 (3-4-11-6)
48
48
Desogestrel or gestodene
3-1 (1-7-5-5)
2-7 (16-4-7)
26(1-5-46)
40
76
Levonorgestrel
46(1-6-9-7)
3-1 (1-3-7-5)
13
22
31(1.3-7.3)
Other
3-7 (2-3-5-9)
3-5(21-5-6)
4-1 (2-56-9)
101
146
All 0C users
Compared with levonorgestrel users
2-0(10-4-1)
0-057
0-027
0-017
2-3(11-4-9)
2-4 (1-2-4-8)
Desogestrel
2-0 (0-8-4-7)
0-055
2-2 (0-9-51)
0-075
0-12
2-3(10-5-2)
Gestodene
0-024
0-018
2-K1-1-3-9)
2-2 (1-1-4-2)
2-3(l-2-4-3)
0008
Desogestrel or gestodene
1
1
1________
Levonorgestrel
Oxford region (GP based controls)!
compared with noousers
Nomiser
49
166
17
15-0 (5-2-436)
12-2 (4-6-32-3)
13-0(5 0-33-5)
27
Desogestrel
7 8 (2-4-25-1)
9-1 (3-0-27-2)
10-7 (3-6-32-0)
16
13
Gestodene
10-9 (4-7-25-4)
11-5 (4-6-28-8)
12-1 (5-2-28-2)
43
30
Desogestrel or gestodene
8-5 (3-5-20-3)
7-4 (3-3-17-0)
40
6-4 (2-9-14-0)
36
Levonorgestrel
8-7 (2-6-28-9)
13-0 (3-547-9)
10
86(2-7-26-7)
12
Other
8-9 (4-3-18-5)
10-0 (4-6-21-9)
80
84 (4-1-17-0)
91
All 0C users
Compared with levonorgestrel users
0-25
0-27
1-6 (0-7-40)
1-8 (0-7-46)
0-10
2-0 (0-9-4-8)
Desogestrel
0-9 (0-3-26)
069
0-70
0-31
1-2 (0-4-3-4)
1-7 (0-6-46)
Gestodene
1-4 (0-6-31)
0-46
0-32
1-5(07-3-1)
1-9 (0-9-4 0)
0088
Desogestrel or gestodene
1
1
1
Levonorgestrel
•BMI categorised as *20. >20 but st25, and >25 kg/m’. Excluding 76 cases (54 non-users. 7 levonorgestrel, and 15 other OC users) and 241 controls (217 non-users,
1 gestodene, 7 levonorgestrel, and 16 other 0C users) in the norvbxford centres, and 1 case (gestodene user). 1 hospital control (gestodene user) and 1 GP-based control
(non-user) in the Oxford region with unknown BMI.
fAdditionai factors were: alcohol consumption (never, occasional, regular light, regular heavy) for all centres comparisons, and Oxford region (hospital controls): varicose veins
(no/yes) for non-Oxford centres: hypertension in pregnancy (nulliparous. none or not stated, yes) for Oxford region (hospital controls); and smoking (non-smoker.
<10 cigarettes/day. >10 cigarettes/day) for Oxford region (GP-based controls).
^Excluding 18 cases with no matched controls (8 non-users and 1 desogestrel. 4 gestodene. 4 levonorgestrel, and 1 other 0C user).
Table 3: ORs (95% Cl) of VTE by type of current OC used (matched risk estimates)
smoking, the risk estimate for gestodene was similar to
that for levonorgestrel users, though the risk for
desogestrel remained high.
Risks according to age, smoking, BMI, and duration of
use
There was no clear pattern of risk estimates in all centres
combined compared with hospital
controls
for
levonorgestrel, desogestrel, or gestodene users according
to age group or smoking status. Risks did increase with
Vol 146 • December 16. 1095
increasing BMI among non-users (1-0, 2-2, and 3-1 for
BMI s20 (referent category], >20 but =625, and
>25 kg/m!, respectively) and among levonorgestrel users
(4-8, 6-0, and 8-7), though risks among desogestrel and
gestodene users combined did not follow this pattern
(32-6, 13-7, and 29-0). There was a slight decline in risk
estimates relative to non-users for users of desogestrel,
gestodene, and levonorgestrel OCs in the first 3 years of
use. However, the ratio of risks for desogestrel and
gestodene relative to levonorgestrel users were higher in
1585
THE LANCET
the second and third years of use (3-4 and 4-4, respec
tively) than in the first year (2-1). The ratio was higher for
women who had not used OCs before the current episode
of use (5-2) than for other current users (2-1).
Type and certainty of VTE diagnosis
The matched, unadjusted risk estimates for deep-vein
thrombosis (DVT) were higher than for pulmonary
embolus (PE) among all OC users (4-3 and 3-4,
respectively), among levonorgestrel users (3-8 and 2-7),
and among users of desogestrel or gestodene (11-6 and
5-5). These patterns of risk remained substantially
unchanged after adjustment for BMI or alcohol
consumption.
Risk estimates for all OC users compared with non
users increased with certainty of diagnosis (4-1 among
definite, probable and possible, 4-4 among definite and
probable, and 6-0 for definite VTE cases) with a similar
pattern of increasing risk estimates among levonorgestrel,
desogestrel, or gestodene users. The ratio of risk estimates
for desogestrel or gestodene compared with levonorgestrel
users remained constant (between 2-6 and 2-8 according
to certainty of diagnosis).
Risk of VTE for users of other progestagens
The 9 cases and 3 controls using combined OCs with
progestagens other than first, second, or third-generation
types in combination with <50 pg ethinyloestradiol all
used a single product (2 mg cyproterone acetate with 35
pg ethinyloestradiol [CPA/35]). 9 cases and 9 controls
who used the same progestagen in combination with 50
pg ethinyloestradiol (CPA/50) were classified with users of
other progestagens and high oestrogen dose. Compared
with non-users, matched, unadjusted risk estimates were
14-9 (3-7-59-4) and 3 8 (1-4-10-7) for CPA/35 and
CPA/50, respectively. The ratios of risks compared with
levonorgestrel users were 5-1 (1-3-20-3) and 1-3
(0-5-3-8), respectively.
Discussion
The WHO Collaborative Study of Cardiovascular Disease
and Steroid Hormone Contraception' demonstrated that
combined OCs were associated with an increased risk of
VTE, although risk estimates were at the lower end of the
range found in earlier case-control studies of non-fatal
idiopathic VTE.’ The analyses reported here indicate that
while users of low oestrogen combined OCs containing
levonorgestrel had 3-5 times the risk of VTE of non-users,
OCs containing the third-generation progestagens
desogestrel or gestodene were both associated with
significantly higher risks (9-1 compared with non-users).
This substantiates the unexpected observation in the main
report, though the discrepancy between the results using
hospital and GP-based controls in the Oxford region
qualifies the interpretation of the data.
Hospital controls were the preferred comparison group
because they provide information in a similar way and
have similar usage of hospital services as the cases.
Community controls were also recruited where feasible.
In areas where there is homogeneous coverage of health
services such controls may be more representative of the
population from which the cases arise. When GP-based
controls in the Oxford region were used for comparison
the risk estimates for all OC types were higher than when
hospital controls were used, and the risks associated with
1586
the use of desogestrel and gestodene were greater than
those for levonorgestrel. However, the ratio of risks for
gestodene compared with levonorgestrel users (1-7) was
less than that observed with the hospital controls (2-3),
and was reduced to 0-9 after adjustment for BMI and
smoking. The risk estimates for desogestrel users
remained high (1-8). The varying results from analyses of
cases compared with hospital and GP-based controls
highlight the importance of appropriate selection and
recruitment of controls and the biases that can occur with
incomplete case or control series. The high non-response
rate, in part a consequence of local ethics committee
restrictions, among women approached as potential GPbased controls, particularly in the younger age groups,
may have resulted in a biased sample of women.
The 18 cases for whom no GP-based controls could be
identified had different patterns of OC use from the
remaining cases, and their exclusion attenuated the higher
risk
associated
with
gestodene
compared
with
levonorgestrel. An analysis of the risks compared with
hospital controls restricted to those cases for whom GPbased controls had been identified resulted in a similar
attenutation of the risk for gestodene users, with little
change in the risk for desogestrel compared with
levonorgestrel. In an unmatched analysis, using all cases
and GP-based controls in the Oxford region, risk
estimates for desogestrel and gestodene were similar and
higher than those for levonorgestrel, both before and after
adjustment for age, BMI and smoking.
Analysis of the prevalence of OC use according to
admission diagnosis for all hospital controls' showed no
systematic differences between groups of admission
diagnoses, other than those expected from the age
distributions of the groups of controls. A detailed
investigation of OC prevalence and distribution of OC
type among hospital controls in the Oxford region showed
no differences within age strata between women admitted
for the two largest groups of diagnoses (ear, nose, and
throat conditions and emergency admissions for trauma
or appendicitis) or other diagnoses. Moreover, hospital
controls were more similar to cases in terms of marital
status, educational level, and social class than were
GP-based controls. Thus, in this study, the former sample
of controls in whom the non-response rate was low
appears better to represent the population from which the
cases arose.
Interpretations
The difference in risk estimates in low oestrogen OCs
according to progestagen type might be due to chance,
subgroup analysis, confounding, or bias, or a combination
of these—or result from a true difference in the effect of
low-dose OCs on the incidence of VTE.
Chance—The p values for the ratio of risk estimates for
desogestrel and gestodene compared with levonorgestrel
were 0-007 and 0 001, respectively, in all centres, even
after adjustment for confounding factors (table 3).
Although the magnitude of the risk differed in the two
main subgroups the p values for both products combined
were 0-001 (non-Oxford centres) and 0-018 (Oxford
region). Similarly, the crude ratios of cases and controls
exposed to desogestrel or gestodene compared with those
exposed to levonorgestrel in each individual centre were
all in the same direction (table 1).
Subgroup analysis—The detailed analyses in this report
were prompted by the observation of an increased risk in
Vol 346 • December 16, 1995
THE LANCET
a subgroup of OC users.1 Although not prespecified in the
study protocol, a secondary objective was to determine
whether the risk of VTE (and also acute myocardial
infarction and stroke) varied with OC composition or
duration of use.2 The layout of the table of risk estimates
by progestagen content and oestrogen dose was defined in
the plan for analysis, prior to the results being computed.
Nevertheless, the estimates reported here may be
exaggerated because the analyses were prompted by
extreme values noted in two of many cells in a table of
ORs. If these results are due to chance, other research will
fail to replicate them. If they represent a true effect, the
magnitude can be expected to be less than that reported
here—an example of regression to the mean, as suggested
for the unexpected observation of an increased risk of
prostate cancer after vasectomy."’'"
Confounding—Although there was evidence of different
prevalence of OC use, as well as brand specific patterns of
use, according to cardiovascular disease risk factors, these
did not appreciably confound the risk estimates for VTE
associated with OC use. The major risk factors for VTE
(trauma, surgery, immobilisation, and pregnancy") were
excluded since the study was designed to investigate
idiopathic V IE. Since differential prescription of OC
types is most likely to be based on known risk factors for
myocardial infarction and stroke, the effect of such
prescription bias on VTE will be limited to risk factors
common to the three conditions, such as increased BMI.
Although controlling for BMI in this study may have been
incomplete because height and weight were self-reported,
adjustment for BMI did not alter risk estimates
sufficiently to suggest that they would change
substantially if exact measurements had been available on
all women. The possibility of other important,
unmeasured confounding factors affecting the risk
estimates is slight, but cannot be excluded.
Bias—Bias due to differential health care seeking
behaviour of cases and controls, or according to exposure
status cannot be excluded. Diagnostic bias in the study of
OC use and VTE is a clear possibility, as is recall and
reporting of contraceptive method. However, such biases
would have to have occurred differentially according to
OC type for them to affect the results presented here.
Exposure to the different types of OCs may have been
related to educational level or social class, as may the
likelihood of being admitted to hospital for VTE, though
adjustment for such variables in the models had no
impact above that of the confounding variables previously
identified. The disparate results with hospital and with
GP-based controls could be accounted for by biased
sampling, directly or indirectly related to the exposures of
interest.
Biological plausibility
The association between OC use and risk of VTE has
been found to be related to dose of oestrogen by some,” ”
although not all,"’ groups. The relation between VTE and
dose and type of progestagen in combined OCs is
uncertain, and evidence for an independent effect of the
progestagen has not been convincing." Although it has
not been clearly shown that the progestagen component
for combined OCs affects the incidence of VTE, some
information is available from studies of the influence of
progestagen-only contraceptives on haemostasis. A 3-year
study of the levonorgestrel subdermal implant Norplant
Vol 346 • December 16, 1995
concluded that changes in haemostatic variables might
increase the risk of thrombosis,” although this was not
found in a similar 6-month study in Egypt.” Changes in
haemostatic indices induced by low-dose OCs and
Norplant are minor, few of the alterations are outside the
normal range, and their clinical significance remains
uncertain.” Recent reviews have concluded that OCs
containing third-generation progestagens do not differ
from earlier low-dose combinations in their impact on
haemostatic variables.” ” These reviews stressed the need
for additional clinical and epidemiological studies. Lowdose OCs containing desogestrel and gestodene seem to
be less androgenic than other low-dose preparations and
have less adverse impact on carbohydrate and lipoprotein
metabolism.21 Lowering the dosages of desogestrel and
gestod ine may reduce the possible excess of venous
thromboembolic side-effects. Such dose reductions may
be achievable without compromising efficacy and cycle
control” since there is evidence that low oestrogen dose
OCs containing these progestagens result in stronger
suppression of ovarian activity.2*
CPA is similar to gestodene and desogestrel in being
less androgenic than levonorgestrel,2’-” and is often
prescribed for the management of acne in young women.”
While the results for CPA-containing OCs are based on
small numbers, the similarity of the observations with the
third-generation progestagens, including the higher risk
associated with use of lower oestrogen dose, as observed
with desogestrel, requires further explanation. It has been
noted from a study of different combined injectable
preparations2' that both the ratio and absolute levels of
progestagen
and
oestrogen
influence
their
pharmacodynamics. However, the observations of
increased risk with lower oestrogen dose with these two
progestagens are both compatible with chance.
Attributable risk and risk of other cardiovascular
endpoints
Any increase in the risk of VTE associated with the use of
different OCs needs to be seen in the context of the
absolute risk of disease, the mortality rates, and the risks
of other cardiovascular endpoints. Cases in the Oxford
region were identified from all hospitals that admitted
acute cases and covered a defined geographical area.
From estimated population size,” total number of VTE
cases identified in the study, patterns of OC use among
the hospital controls by age, and estimated risks from the
Oxford region compared with the hospital controls, the
incidence of idiopathic VTE was estimated” to be 3-9 for
non-users, 10 3 for levonorgestrel users, 21-3 for users of
desogestrel or gestodene, and 12-3 for users of other OCs
per 100 000 woman-years. The excess for users of
desogestrel or gestodene over levonorgestrel OCs was
11-0 per 100 000 woman-years (CI 1-2-22-7). The case
fatality* for VTE is low (l-2%)12 and needs to be
considered in the context of the incidence of other
cardiovascular diseases in women of reproductive age."
Provisional results from the WHO case-control study
show no clear evidence of any difference in the risk of
stroke according to type of progestagen within the class of
low-oestrogen OCs. The cases of acute myocardial
infarction occurred predominantly among older women
within the reproductive age range where the prevalence of
current OC use was low and the number exposed to
third-generation OCs was too small to provide useful
estimates of risk.
1587
THE LANCET
Thorogood M. Oral contraceptives and cardiovascular disease: an
epidemiologic overview. Pharmacocpidemtol Drug Safety 1993; 2: 3-16.
Rosenberg L, Palmer JR, Zaubcr AG, Warshauer ME, Stolley PD,
Shapiro S. Vasectomy and the risk of prostate cancer. Am J Epidemiol
1990; 132: 1051-55.
11
Guess HA. Vasectomy and prostate cancer. Am J Epidemiol 1990; 132:
1062-65.
12
Stadcl BV. Oral contraceptives and cardiovascular disease (part 1).
N Engl J Med 1981; 305: 612-18.
13
Bdttiger LE, Boman G, Eklund G, Westerholm B. Oral contraceptives
and thromboembolic disease: effects of lowering oestrogen content.
Lancet 1980; i: 1097-101.
14
Inman WH, Vessey MP, Westerholm B, Englund A. Thromboembolic
disease and the steroidal content of oral contraceptives: a report to the
Committee on Safety of Drugs. BMJ 1970; ii: 203-09.
15
Gerstman BB, Piper JM, Tomita DK, Fergusson WJ, Stadel BV,
Lundin FE. Oral contraceptive estrogen dose and the risk of deep
venous thromboembolic disease. Am J Epidemiol 1991; 131: 32-37.
16
Helmrich SP, Rosenberg L, Kaufman DW, Strom B, Shapiro S. Venous
thromboembolism in relation to oral contraceptive use. Obstet Gynecol
1987; 69:91-95.
17
Meade TW, Greenberg G, Thompson SG. Progestogens and
cardiovascular reactions associated with oral contraceptives and a
comparison of the safety of 50 and 30 pg oestrogen preparations. BMJ
1980;280: 1157-61.
18
Singh K, Viegas OAC, Ratnam SS. A three-year evaluation of
hemostatic function in Singaporeian Norplant acceptors. Adi'
Contracept 1990; 6: 23-32.
19
Shaaban MM, Elwan SI, El-Kabsh MY, Farghaly SA, Thabet N. Effect
of levonorgestrel contraceptive implants, Norplant, on blood
coagulation. Contraception 1984; 30: 421-30.
20
Robinson GE. Low-dose combined oral contraceptives. BrJ Obstet
Gynaecol 1994; 101: 1036-41.
21
SperoffL, DeChemey A, and the Advisory Board for the New
Progestins. Evaluation of a new generation of oral contraceptives. Obstet
Gynecol 1993; 81: 1034-^7.
22
Stubblefield PG. The effects on hemostasis of oral contraceptives
containing desogestrel. Am J Obstet Gynecol 1993; 168: 1047-52.
23
Newton JR Classification and comparison of oral contraceptives
containing new generation progestogens. Hum Reprod Update 1995; 1:
231-63.
24
Van der Vange N. Ovarian activity during low dose oral contraceptives.
In: Chamberlain I, ed. Contemporary obstetrics and gynaecology.
London: Butterworth Heinemann; 1988: 315-26.
25
Jones EE. Androgenic effects of oral contraceptives: implications for
patient compliance. Am J Med 1995; 98(1 A): 116S-19S.
26
Darney PD. The androgenicity of progestins. Am J Med 1995; 98(1A):
104S-10S.
27
Dieben TO, Vromans L, Theeuwes A, Beanink HJ. The effects of
CTR-24, a biphasic oral contraceptive combination, compared to
Diane-35 in women with acne. Contraception 1994; 50: 373-82.
28
Newton JR, d’Arcangues C, Hall PE. A review of *once-a-month’
combined injectable contraceptives. Obstet Gynecol 1994; 14: S1-S34
(suppl 1).
29
Office of Population Censuses and Surveys. Mortality statistics area,
England and Wales, series DH5 nos 16-19. London: HM Stationery
Office, 1989-92.
30
Schlesselman JJ. Case-control studies: design, conduct, analysis. New
York: Oxford University Press, 1982.
31
Stadel BV. Oral contraceptives and cardiovascular disease (part 2).
N Engl J Med 1981; 305: 672-77.
Conclusion
9
This WHO case-control study, only part of which has
been reported so far, was designed to investigate possible
risk of cardiovascular disease associated with current
patterns of OC use. It is to our knowledge the first
epidemiological study to report specifically on OCs
containing desogestrel and gestodene. These two
products were found to be associated with a similar level
of increased risk of VTE above those for secondgeneration combined OCs with less than 35 pg
ethinyloestradiol. These observations are based on an
analysis of a secondary study objective, and the possibility
that they are due to chance, confounding, or bias or a
combination of these cannot be excluded entirely. They
must be confirmed by independent epidemiological
studies, and the impact of these new OC preparations on
the risk of stroke and myocardial infarction must be
further assessed.
10
Study organisation
Writing committee: T M M Farley, O Meirik, C L Chang, M G Marmot,
and N R Poulter.
Centres and Principal Investigators, Study Design and Monitoring, Study
and Data Coordination, and Publications Advisory Committee as in
companion paper (Lancet 1995; 346:1575-82).
Acknowledgments
Funded by UNDP/UNFPA/WHO/World Bank Special Programme of
Research, Development and Research Training in Human Reproduction.
We thank M J Campbell, P Hannaford, B Hulka, J Kelaghan,
M Tikkanen, M P Vessey and O Viegas who participated in a meeting to
review preliminary results on the risks associated with third-generation
progestagens and advise on their interpretation.
References
1
World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. Venous thromboembolic
disease and combined oral contraceptives: results of international
multicentre case-control study. Lancet 1995; 346: 1575-82.
2
World Health Organization Collaborative Study of Cardiovascular
Disease and Steroid Hormone Contraception. A multinational case
control study of cardiovascular disease and steroid hormone
contraceptives: description and validation of methods. J Clin Epidemiol
(in press).
3
Population Information Program. Popul Rep 1988; ser A, no 7: 1-31.
4
Stanczyk FZ, Roy S. Metabolism of levonorgestrel, norethindrone and
structurally related contraceptive steroids. Contraception 1990; 42:
67-96.
5
Shenfield GM, Griffin JM. Clinical pharmacokinetics of contraceptive
steroids: an update. Clin Pharmacokinet 1991; 20: 15-37.
6
Breslow NE, Day NE. Statistical methods in cancer research: vol I: the
analysis of case-control studies. Lyon: International Agency for
Research on Cancer, 1980.
7
Hirji KF, Mehta CR, Patel NR. Exact inference for matched case
control studies. Biometrics 1988; 44: 803-14.
8
Maldonado G, Greenland S. Simulation study of confounder selection
strategies. Am J Epidemiol 1993; 138: 923-36.
1588
Vol 346 • December 16, 1995
Not viewed