10626.pdf
Media
- extracted text
-
u
UlSi H. : GENERAUE)
^fiSs WORLD HEALTH ORGANIZATION
V ORGANISATION MONDIALE DE LA SANTE
WHO/TB/93. 173
ENGLISH ONLY
A REVIEW OF CURRENT EPIDEMIOLOGICAL DATA AND ESTIMATION OF
FUTURE TUBERCULOSIS INCIDENCE AND MORTALITY
PAUL JOHN DOLIN
Imperial Cancer Research Fund
Cancer Epidemiology Unit
Radcliffe Infirmary
University of Oxford
Oxford 0X2 6HE, U.K.
MARIO C. RAVIGLIONE and ARATA KOCHI
Tuberculosis Programme
World Health Organization
Geneva, Switzerland
This documeni 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, repro
duced and translated, in part or in whole, but not for sale nor for
use in conjunction with commercial purposes.
Ce document n'est pas une publication officielle de I'Organisation
mondiale de la Sante (QMS) et tous les droits y afferents sont reserves
par I’Organisation. S’il peut etre comment^, resume, reproduit ou
traduit, partiellement ou en totality, il ne saurail cepeiidant I'etre pour
la venle ou a des fins cornmerciales.
The views expressed in documents by named authors are solely the
responsibility of those authors.
les opinions exprim^cs dans les documents par (It'S auteurs cil6s
noinmcment n'engagent que Icsdiis auteurs.
AUu
WHO/TB/93.173
page 1
11
28
SUMMARY
The incidence of tuberculosis in 1990 is estimated al 7.5 gution cases of which 3 J million were in t#e
South-East Asian region of WHO, 1.8 nnHion in the Western Pacific region, 1.0 million in sub-Saharan
Africa and 1.6 million in the remainder of the world. The estimates for South-East Asian and Eastern
Mediterranean regions are higher than previously thought, while the estimates for the Western Pacific and
African regions are lower. Forty-four percent of all cases occurred in China and India.
Annual incidence is predicted to increase to 8.8 million cases by 1995, 10.2 million by 2000 and 11.9
million by 2005, an increase of 58% compared with 1990 incidence. Demographic factors, such as population
growth and changes in age structureof populations, will account for 77% of the predicted increases in
incidence. Increasing incidence rates, particularly in Africa, will account for 23% of the increase in new
cases. In the Eastern Mediterranean region and Central and South America, age-specific incidence rates are
expected to fall during 1990-2005 but the total number of new cases will increase because of population
growth.
It is estimated 315,000 (4.2%) of the 7.5 million incident cases of tuberculosis in 1990 were
attributable to HIV infection. Over half of such cases occurred in sub-Saharan Africa where 23.8% of new
rases of tuberculosis in adults aged 15-59 were attributable to HIV infection. By the year 2000, it is
estimated that over 1.4 million (13.8%) of the forecast 10.2 million incident cases occurring annually will be
attributable to HIV infection.
During the 10-year period 1990-1999 it is estimated that 88.2 million people will develop tuberculosis,
of which 8.0 million cases will be attributable to HIV infection.
In 1990, 2.5 million persons are estimated to have died of tuberculosis. Assuming availability of
treatment remains at its 1990 level, it is predicted 3.0 million tuberculosis deaths will occur annually by 1995
and 3.5 million deaths annually by 2000. Of the 3.5 million tuberculosis deaths predicted to occur annually
by the year 2000, 0.5 million (14%) will be attributable to HIV infection.
During the decade 1990-1999 it is estimated that 30.0 million tuberculosis deaths will occur, of which
2.9 million will be attributable to HIV infection. Around 6.0 million tuberculosis deaths are expected in subSaharan Africa during the decade, 15 million (25%) of which will be attributed to HIV infection.
Community Health Cell
Library and Information Centre
# 359, "Srinivasa Nilaya"
Jakkasandra 1st Main,
1st Block, Koramangala,
BANGALORE - 560 034.
Ph : 2553 15 18/2552 5372
e-mail: chc@sochara.org
mn
m
TTT
f
WHO/TB/93.173
page 2
1. Introduction
c
WHO has declared tuberculosis a global emergency and warned that the disease claims millions of
lives each year and that the situation will rapidly worsen unless immediate action is taken to curb its spread.
This document presents estimates of the present global situation and forecasts of increasing morbidity
and mortality in the near future. Estimates are given for the world, WHO regions, and for selected
countries.
Specific areas addressed in these forecasts include the impact of world population growth, the impact
of demographic aging of the world’s population, and the impact of the HIV epidemic.
The information presented here are forecasts of morbidity and mortality based on current levels of
intervention. Tuberculosis is curable, and treatment is inexpensive. Much of the forecast burden of disease
could be prevented if sufficient global resources were directed immediately towards the control of this
disease.
2. Review of previous estimates of 1990 incidence
Tuberculosis incidence in 1990 was estimated for developing countries by Murray (7) and extended by
Sudre (2) to include industrialised countries. Both authors based their estimates on a comprehensive review
of survey data on annual risk of infection (ARI) in developing countries (3). Annual tuberculosis incidence
was estimated by applying Dr Styblo’s conversion factor, a 1% increase in ARI correlates to 50 additional
smear positive cases per 100,000 population per year(4), to regional ARI estimates.
However, estimates based on such a methodology need to be interpreted with caution for a number of
reasons. First, the conversion factor is based on measurements of certain populations and the validity of
applying this to other populations is unknown. Second, the review of ARI only covered survey data up to
1985 (3) and few comprehensive surveys have been undertaken since 1985. Third, for most regions of the
world there is sparse ARI data. For example, the estimates for the Caribbeans, Central and South America
(46 countries) were based on survey data from two countries only. Fourth, previous immunisation with BCG
excludes survey subjects from the calculation of ARI. In many areas where recent surveys were undertaken,
most of the sample population had previous BCG immunisation and thus was excluded from the calculation.
For example at least 80% of the survey population was previously immunised in Botswana and Malaysia,
and at least 50% in a range of other countries, including Algeria, Argentina, Burundi, Libya, Republic of
Korea, Samoa and Tanzania (3). Thus, subjects included in the ARI calculation in regions with high BCG
immunisation coverage may not be representative of the population from which they were drawn.
On the basis of the above concerns about calculating global incidence of tuberculosis from survey data
on ARI, a new set of estimates of tuberculosis incidence in 1990 have been produced using an independent
methodology.
J. Tuberculosis incidence in 1990
WHO routinely collects data on the number of new tuberculosis cases in Member States each year
(5). Within each WHO region, except the African region, an overall regional crude incidence rate was
calculated by estimating the incidence in the most populated countries. Reliable notification data were
preferentially used for these estimates. Notification data were considered reliable when provided by
programmes with an established surveillance system. For countries with unreliable notification data, ARI
was used to estimate incidence. Notification data arc relatively poor for the African region, so a slightly
different approach was used. The African region was divided into four geographic areas and within each
area, a crude incidence rate was estimated based on the most reliable notification data (e.g., United
Republic of Tanzania for East Africa, Cole d’Ivoire lor West Africa).
WHO/TB/93.173
page 3
Estimates of incidence based on notification data will partly reflect under reporting, and thus must be
considered as conservative estimates. While notification data are of poor quality for many countries, and any
estimates based on such data will seriously risk underestimating incidence, reliable data are available from
other countries, particularly those where good tuberculosis control programmes are established.
Shown in Table 1 are the new 1990 estimates. Based on notification data, the number of incident cases
of tuberculosis in 1990 is estimated at 7,537,000. While these estimates must be considered conservative,
they are consistent with the ARI-based estimate of 8.0 million incident cases in 1990 by Sudre (2).
In the following sections, more detailed information is given on the calculation of tuberculosis
incidence for each WHO region.
3.1 South-East Asian Region
The notification rate for India, the most populated country in the region, was 153 per 100,000
population in 1990. Not included in (he Indian notification system were patients treated privately and
patients hospitalised in the more than 40,000 tuberculosis hospital beds. Of the 438 administrative districts
in India, only 378 had district tuberculosis programmes, and of these only 278 (63% of all districts) notified
cases in 1991. It has been estimated that 57% of identified cases were included in notification reports (6). In
many countries, relapse cases are usually included with incident cases in reports to WHO. For India it is not
known what proportion of notifications were relapse cases, but other countries in the region have reported
around 5% were relapse cases. This suggests tuberculosis incidence in India during 1990 was approximately
242 per 100,000 population.
In Indonesia, the second most populated country in the region, only 2,700 of the 6,000 health centres
are currently included in the national tuberculosis programme. Notification data only includes smear positive
cases and thus underestimates total incidence. Tuberculin surveys undertaken during 1983-87 in various
regions including Java, Sumatra and Kalimantan found an ARI of 1.7-4.1%. In at least five districts the ARI
was above 2.3% (7). ARI of this magnitude suggests an incidence rate at least as high as that for India
where ARI is around 1.5-2%. For other countries of the region (e.g. North Korea, Mongolia, Myanmar,
Thailand) notification data suggest an incidence rate half that of India.
The 1990 incidence rate for the South-East Asian region is estimated at 237 per 100,000 population.
This is higher than the previous estimates of 165-200 per 100,(XX) given by Murray (J) and 194 per 100,000
population given by Sudre (2). Based on these calculations, it is estimated 3.1 million new cases occurred in
the South-East Asia region during 1990, including 2.1 million new cases in India and 0.4 million in
Indonesia.
3.2. Western Pacific Region
Thirty two countries were included in the Western Pacific region. The three industrialised countries of
the region (Japan, Australia and New Zealand) have been excluded and grouped with Europe and North
America.
In 1990, the population of the Western Pacific region was around 1,350 million, the most populated of
all the WHO regions. Approximately 85% of the population live in China, with the remaining 15% spread
across the other 31 countries. Thus, the incidence of tuberculosis in China strongly influences the overall
regional incidence.
Murray (J) estimated an ARI of 1-2% for Asia while Sudre (2) estimated 1-2.25% for the Western
Pacific region. However, data from the 1984-5 and 1990 national tuberculosis surveys of China suggest an
ARI of around 1.0% (8,9). This suggests that previous estimates of tuberculosis in China, and thus the
whole of the Western Pacific region, may have overestimated the situation.
The average notification rate for tuberculosis in China during 1987-90 was 49 per 100,000 population.
Data from the 1984-85 national survey of China suggest that approximately 40% of cases had been
registered (<S’). Recent notification data from China show that relapse cases account for approximately 7.5%
of cases notified to WHO. Applying the 40% coverage rate to notification data and allowing for inclusion of
9
WHO/TB/93.173
page 4
relapse cases, gives an estimated 1990 incidence rale of 113 per 100,000 population for China. This estimate
is approximately 40% lower than the estimate of 192 per 100,000 population by Sudre (2).
4
Of the 15% of the region’s population that live in countries other than China, half reside in the
Philippines and Republic of Korea. During 1988-90, the Philippines reported 312 incident cases per 100,000
population and Republic of Korea reported 194 per 100,000 population. Assuming 7.5% of notified cases
were relapses and a case-finding rate of 80% (a conservative estimate), the 1990 incidence estimate for the
Philippines is 360 per 100,000 population and for the Republic of Korea is 225 per 100,000.
Only 8% of the population live in the other 29 countries of the region? For these countries,
notification rates ranged from 30 to 300 cases per 100,000 population with a weighted average of 100 per
100,000 population. Allowing for 45% case-finding rate (reported by various countries within the region)
and inclusion of relapse cases (7.5%), it is estimated an incidence rate of 205 per 100,000 could be applied
to these countries.
For the entire Western Pacific region, it is estimated that there were 1.8 million new cases (136 per
100,000 population) in 1990, 28% lower than estimated previously. Around 1.3 million cases occurred in
China.
3.3. African Region
Africa was divided into four geographical areas and within each area estimates of incidence were
based on the most reliable notification data. For Eastern Africa, recent notifications ranged from 50 to 220
cases per 100,000 with a mean of 110 cases per 100,000. Tanzania, with one of the best tuberculosis
programmes, reported approximately 105 cases per 100,000 population in 1990. Highest notification rates in
East Africa were reported by countries with a high incidence of AIDS (e.g. Malawi, Zambia). Reports from
national tuberculosis programmes suggest that relapse cases account for 5% of reported incident cases.
Allowing for around 50% case-finding for all countries in the area, the 1990 incidence rate for Eastern
Africa is estimated at 200 cases per 100,000 population.
In Central Africa, recent notifications range from 50 to 200 cases per 100,000 with a mean of 80 per
100,000 population. These rates are similar to those reported by Eastern Africa countries, suggesting an
incidence rate similar to that estimated for Eastern Africa.
Notifications from countries in Southern Africa ranged from 170 to 225 cases per 100,000 population
with a mean of 205 per 100,000. The Republic of South Africa reported 220 cases per 100,000 population.
Allowing for inclusion ot relapse cases (5%) and 70% case-finding, the underlying incidence rate for the
Republic of South Africa was approximately 300 cases per 100,000 population. This rale was applied to
other countries in Southern Africa.
West African notification rates are lower than other areas of sub-Saharan Africa. Notifications ranged
from 20 to 140 cases per 100,000 population, with a mean of 50 per 100,000. Cote D’Ivoire, with one of the
best surveillance systems in West Africa, reported 65 cases per 100,000 population in 1991. Relapse cases
accounted for 5% of notifications and case-finding was estimated at 50%. This together with data from
other countries in (he area suggests an incidence rate of around 150 cases per 100,000 population in West
Africa during 1990.
Il is estimated that the 1990 incidence of tuberculosis for all Sub-Saharan Africa was 992,000 cases
(191 per 100,000 population), 28% lower than previously estimated.
3.4. Eastern Mediterranean Region
Pakistan, the most populated country in the region, reported 169 cases per
|
100,000 population in ;;;
1988.
No data is available on.the proportion of relapse cases included in notifications from Pakistan. Data from
other countries in the region suggest that around 5% of reported incident cases arc relapses and a case
finding rate of around 65%. The 1990 incidence rale for Pakistan is estimated al 250 cases per 100,000
population.
WHO/TB/93.173
page 5
Notification rates for other countries in the region, trend to fall into three categories. Allowing for
50% case-finding and inclusion of relapse cases, incidence rates for Djibouti, Somalia, Sudan and Yemen
are estimated at 200 cases per 100,000 population. For Saudi Arabia, Syria, Oman, Qatar and Iran incidence
is estimated at 100 cases per 100,000 population while for the remaining countries of the region, incidence is
estimated at 50 cases per 100,000 population.
The 1990 incidence of tuberculosis for the entire Eastern Mediterranean region is estimated at 641,000
cases (165 per 100,000 population), similar to previous estimates. It is estimated 306,000 cases occurred in
Pakistan and 335,000 cases in the other 22 countries of the region.
3.5. American Region, excluding USA and Canada
Included in this region were countries of the Caribbean, Central America and South America. Canada
and the USA have been excluded and grouped with Europe and other industrialised countries. Of the 46
countries included in the region, 23 have populations of less than one million. Each of these countries
reports only a small number of cases annually, and any estimates based on these numbers would be subject
to Poisson variation. Of the other 23 countries in the region, notifications ranged from 20 to 180 cases per
100,000 population with a weighted average of 50 cases per 100,000. Brazil, the most populated country in
the region reported around 50 cases per 100,000 population, although notifications only covered Rio de
Janeiro.
Allowing for incomplete reporting, a case-finding rate of around 70%, and data from other countries
in the region, the 1990 incidence for the region is estimated at 127 cases per 100,000 population (569,000
cases), similar to the estimate of Sudre (2). It is estimated 191,000 (34%) of these cases occurred in Brazil
and 378,000 (66%) in the other 45 countries of the region.
3.6. European Region and USA, Canada, Japan, Australia and New Zealand
The region was considered in two section: (a) Eastern Europe and independent states of the former
USSR and (b) Western Europe and the five other industrialised countries.
For Eastern Europe and independent states of the former USSR, the weighted average of 1990
notification rates was 42 cases per 100,000 population. Notification rates for Western European and other
industrialised countries ranged from less than 10 per 100,000 in Canada, Denmark, the Netherlands, Norway
and Sweden to more than 60 per 100,000 in Portugal. The weighted average of notifications for all Western
European and other industrialised countries was 21 per 100,000 population.
If an 85% registration rate is assumed and allowing for 5% of reported cases to be relapses, the
resulting estimates of 1990 incidence are 47 per 100,000 population for Eastern Europe and independent
states of former USSR and 23 per 100,000 for Western Europe and the five other industrialised countries.
For the entire region, it is estimated that 390,000 incident cases occurred in 1990 (31 cases per 100,000
population), similar to previous estimates.
3.7. Summary
It is estimated there were 7,537,000 incident cases of tuberculosis in 1990. Over 4.9 million cases
(65%) occurred in the South-East Asian and Western Pacific regions and particularity in India (2.1 million),
China (1.3 million) and Indonesia (0.4 million).
One million cases are estimated to have occurred in sub-Saharan Africa during 1990, 0.6 million cases
in the Eastern Mediterranean region (including 0.3 million cases in Pakistan) and 0.6 million in the Central
and South America (including 0.2 million cases in Brazil).
Around 0.2 million cases occurred in Eastern Europe and independent states of the former USSR and
0.2 million cases in Western European and other industrialised countries.
The global estimate based on notification data is similar to that estimated by Sudre (2) based on ARI
data. This should, allay some of the concerns expressed about the use of ARI for estimating incidence rates.
>
WHO/TB/93.173
page 6
However, lhe two methods do yield different estimates for some regions. In particular, the notification
based estimates suggest a substantially greater incidence of tuberculosis in South-East Asia than previously
thought. Conversely, a lower incidence in China and hence lhe entire Western Pacific Region was estimated
when based on notifications.
4. Tuberculosis incidence in 1995, 2000 and 2005
In estimating future disease incidence, allowances were made for demographic factors (changes in the
size and age structure of populations) and epidemiological factors (changes in age-specific incidence rates).
To accurately allow for both demographic and epidemiological factors, age-specific incidence rates were
estimated for the years 1995, 2000 and 2005 and then applied to age-specific population projections for
these years.
This was undertaken at the regional level in two steps. First, data available at WHO on the age
distribution of notified cases in each region during 1990 were applied to the 1990 regional crude incidence
rates to derive 1990 regional age-specific incidence rates. Second, trends in regional notification rates during
1985-1990 were applied to the 1990 regional age-specific incidence rates to derive age-specific incidence
rates for the years 1995, 2000 and 2005. This approach assumes that age-specific trends will remain
unchanged. The derived rates were then applied to regional age-specific population projections to calculate
the number of incident cases expected in 1995, 2000 and 2005.
In forecasting future burden of disease, it has been assumed that intervention (i.e. the activities of
national programmes) remains al the 1990 level.
4.1. 1990 age-specific incidence rates
Age was categorised into four groups: 0-14 years, 15-34 years, 35-59 years and 60+ years. Table II
shows the age structure of incident cases for each region in 1990, based on data supplied to WHO by
national tuberculosis programmes. The majority of cases are aged between 15 and 59 years in each region.
Variation between regions partly reflect differences in age structure of lhe underlying populations.
The number of incident cases aged 0-14, 15-34, 35-59 and 60 and older in each region was estimated
by applying the age distribution of cases shown in Table II to lhe revised 1990 incidence estimates shown in
Table I. Regional age-specific incidence rates for 1990 were then calculated by dividing these estimates by
regional age-specific population estimates (10).
4.2. Trends in incidence rates
Trends in notification rates during 1985-90 were analysed to predict trends during 1990-2005. While
notification-based data are subject to variation in completeness of reporting, they do provide a useful source
of information and, when used with data on trends in ARI, trends in incidence rales can be estimated.
South-East Asian Region
In India, notifications increased by approximately eight additional cases per 100,000 population per
year during 1985-90. This probably reflects improvements in completeness of reporiing. Il is also possibly
the underlying incidence rale may be increasing. The extent of dual HlV-iubcrculosis infection in lhe Indian
population is unclear. However, it seems reasonably to assume there is a pool of persons with dual
infection, and without intervention, the pool could be expected to increase. Il is estimated the incidence of
tuberculosis in India will increase during 1990-2005 by 0.4 additional case per 100,000 population per year,
largely due to increasing prevalence of HIV infection.
For other countries in lhe region, notification data suggest annual incidence is decreasing by
approximately 1.0 cases per 100,000 population per year. For lhe region as a whole, lhe weighted average
trend is a reduction in incidence by 0.09 cases per 100,000 population per year. However, this trend could
WHO/TB/93.173
page 7
be reversed if a marked increase in the prevalence of HIV infection among persons infected with
tuberculosis occurs.
Western Pacific Region
Data from the 1979, 1984-85 and 1990 national surveys of China suggest the ARI among seven year
old children was similar in each survey period (8,9). Data from these surveys also suggest the prevalence of
infection among children aged 0-5 was higher in 1990 than in 1979. Analysis of registration data shows
notifications have been increasing by approximately 0.5 additional cases per 100,000 population each year
since 1986. Based on the available data, it is estimated that the annual incidence rale in China during 19902005 is likely to remain at the 1990 level, in the absence of intervention.
For most other countries in this region, notification rates have generally been falling in recent years by
one case per 100,000 population each year. There arc a few exceptions to this trend but mostly among the
less populated countries.
The underlying annual incidence rate for the Western Pacific region as a whole is predicted to
decrease by 0.13 cases per 100,000 population per year.
African Region
Around 1985 some African countries, particularly those in Eastern and Central Africa started
reporting increases in tuberculosis rates. Although an increase in notifications for one country may reflect
improved completeness of reporting, increasing notification rales across an entire region suggests that
changes in reporting alone may not account for the increase.
For countries of Eastern and Central Africa with annual notification rates for AIDS of at least 10
cases per 100,000 population (e.g. Burundi, Congo, Malawi, Rwanda, Uganda and Zambia), tuberculosis
notifications rates have been increasing, on average, by 10 additional cases per 100,000 population per year
since 1985. The increase in notifications in Eastern and Central Africa correlates with the increasing
prevalence of dual HIV-tuberculosis infections.
Increases of 5-10 additional tuberculosis notifications per 100,000 population have also occurred in
other parts of Africa during 1985-90, including countries currently reporting less than 10 new cases of AIDS
per 100,000 population per year (e.g. Cameroon, Equatorial Guinea, Mozambique, Senegal and South
Africa). The interpretation of these trends is that the high prevalence of HIV infection seen initially in
Eastern and Central Africa is now occurring in other parts of sub-Saharan Africa. In parallel with the
spread of the AIDS epidemic, dual HIV-tuberculosis infection is increasing throughout the region. It is
estimated, based on increases in notification during 1985-90, the underlying incidence rates across SubSaharan Africa will continue to increase by 10.0 additional new cases per 100,000 population per year during
1990-2005.
Eastern Mediterranean Region
Murray (7) estimated ARI to be decreasing during 1985-90 in the Eastern Mediterranean region.
Notification data for this period confirm this trend. Notification rates fell in 13 of the 17 countries that have
consistently reported cases to WHO.
However, the notification rale for Pakistan, the most populated country in the region, has been
increasing in recent years by 4.0 additional cases per 100,000 population each year. It is thought the increase
in the notification rale for Pakistan results from improved reporting. It is also possible that some increase in
nolifications could result from increases in underlying incidence rales. The prevalence of dual infection with
HIV-tuberculosis in Pakistan is largely unknown, but thought to be relatively low. It is estimated, based on
available data, that incidence rates in Pakistan will continue at their 1990 level during 1990-2005.
The weighted average of notification rates for other countries in the region is decreasing by around 1.0
cases per l(K),000 population each year. Based on (his, the underlying incidence rales for these other
countries arc predicted to decrease by 1.0 case pur 100,(XX) population per year during 1990-2005.
For the Eastern Mediterranean region as a whole, underlying incidence rales are predicted to
decrease by O.OX cases per 100,(XX) population per year during 1990-2005.
1
WHO/TB/93.173
page 8
American Region
It has been estimated that ARI was falling in Central and South America during 1985-90 (/). Analyses
of notification data confirm this trend and suggest underlying incidence rates are falling by approximately
1.5 case per 100,000 population per year in the region.
Notification rates for some countries have risen in recent years, possibly due to the AIDS epidemic
(e.g. Haiti, Honduras) but the impact of this on regional notification rates to date has been minimal
because the population of (he countries showing an upward trend are relatively small compared with other
countries within the region.
Underlying regional incidence rales are predicted to continue falling by around 1.5 cases per 100,000
population per year during 1990-2005, unless the prevalence of dual HIV-tuberculosis infection increases
appreciably.
European Region and USA. Canada, Japan, Australia and New Zealand
Tuberculosis notification rates increased during 1985-90 in several countries (including The
Netherlands and USA) by approximately 0.5 additional cases per 100,000 population per year. However, for
most other countries in Western Europe, notifications decreased during 1985-90 by 0.5-1.0 cases per 100,000
population per year. It seems reasonable to assume that rates in other countries in Western Europe will
also increase in the near future, partly because of increasing migration of people to Western Europe from
regions of the world with higher incidence of tuberculosis, and partly because of the increasing number of
persons with dual HIV-tuberculosis infection. It is estimated that tuberculosis incidence in Western Europe
and the other industrialised countries will either remain at their current levels or increase slightly during
1990-2005.
In Eastern Europe and states of the former USSR, notifications decreased during 1985-90 by
approximately one case per 100,000 population per year. However, it is not clear that the decline in
incidence rates can be maintained in these countries. It is assumed underlying incidence rates in Eastern
Europe and states of the former USSR will remain at their 1990 level in the near future.
4.3. Incidence estimates in 1995, 2000 and 2005
The forecast trends in underlying regional incidence rales during 1990-2005 (section 4.2) were applied
to 1990 age-specific incidence rates (section 4.1) to produce estimates of age-specific incidence rales for the
years 1995, 2000 and 2005. These rales were then applied to age-specific population projections (70) to
derive the number of incident cases expected in each year.
Table III shows the number of incident cases of tuberculosis predicted to occur in 1990, 1995, 2000
and 2005. The number of incident cases is expected to increase from 7.5 million new cases a year in 1990 to
8.8 million in 1995, 10.2 million in 2000, and 11.9 million new cases a year in 2005, an increase of^SZO^
over 15 years. Around 4.5 million new cases annually can be expected in the South-East Asian region, 2.8
million in the African region and 2.5 million in the Western Pacific region by the year 2005.
The age-specific incidence rales for lhe Eastern Mediterranean region and Central and South America
were predicted to fall during 1990-2005, but the actual numbers of incident cases are expected to increase.
This indicates that demographic factors (changes in size and age structure of populations) are stronger than
epidemiological factors (changes in age-specific incidence rales). The impact of demographic and
epidemiological factors on the forecasts is examined below.
4.4. Impact of demouraphic and upidcmiolotiical factors
The effects of demographic factors (population growth ami changes in lhe age structure of
populations) were examined by fixing age-specific incidence rales al their 1990 levels (i.c., no change in age
specific incidence rales during 1990-2(N)5) and applying these rales to age-specific population estimates for
the years 1995, 2000 and 2^)5.
WHO/TB/93.173
page 9
Table IV shows the expected increase in incidence due solely to demographic factors With age
specific incidence rates fixed al their 1990 level, the number of incident cases of tuberculosis occurring each
year increases from 7.5 million cases annually in 1990 to 10.9 million cases annually by 2005, an increase of
45%. This increase is due solely to demographic factors.
The influence of changing age structure is reflected in changes to the crude incidence rate. The global
crude incidence rale (Table IV) is predicted to increase from 143 cases per 100,000 population in 1990 to
161 cases per 100,000 population by 2005, although age-specific incidence rates were fixed at their 1990
level. This occurs because the proportion of the world’s population in the middle and older age groups is
increasing-and tuberculosis incidence rales are highest among these age groups.
While incidence is predicted to increase from 7.5 million cases annually in 1990 to 10.9 million in 2005
due to demographic factors only, it was forecast in Table III that incidence would rise to 11.9 million by
2005 when allowing for both demographic and epidemiological factors. This indicates that demographic
factors will account for most of the predicted increase in annual incidence.
Shown in Table V are the predicted additional cases of tuberculosis, compared with 1990 incidence,
due to demographic and epidemiological factors. Over three-quarters of the predicted increase will result
from demographic factors, such as population growth and changing age structure of the population.
Less than 25% of the predicted increase will result from changes in age-specific incidence rates. These
forecast changes due to epidemiological factors represents the balance between increasing rates due, largely,
to the HIV epidemic versus falling rates due to effective intervention strategies. For example, by the year
1995, epidemiological factors are expected to account for 234,000 additional cases expected annually (Table
V). This results from the 423,000 additional cases expected due to the HIV epidemic (see section 4.5),
minus 189,000 additional prevented cases due to intervention strategies. Clearly the impact of prevention
programmes, assuming they remain at their 1990 level, will be overshadowed by both population growth and
the HIV epidemic.
While the number of new cases per year in each region is forecast to increase dramatically, the data in
Table V indicate that prevention programmes are working. In South-East Asia, Western Pacific, Eastern
Mediterranean and particularly the Americas, epidemiological factors (falling age-specific incidence rates)
are preventing tens of thousands of additional cases from occurring each year. Unfortunately, at the global
level, these achievements are overshadowed by the increasing age-specific rates in Africa.
The interpretation of above findings is that advances are being made in preventing new cases of
tuberculosis, as seen by the falling age-specific rales. However, these efforts are insufficient to counter the
strong effects of population growth, the demographic aging of populations, and the HIV epidemic.
Tables VI-XII show forecast annual incidence of tuberculosis in 1995, 2000 and 2005 at the regional
level. The top half of each table shows predicted incidence due solely to demographic factors (age-specific
rates fixed at 1990 levels), while the lower half of each table shows predicted incidence when allowances are
made for both demographic and epidemiological factors.
In the South-East Asian region (Table VI), the benefits of a small decrease in age-specific rates will
be overshadowed by the strong effects of demographic factors. By the year 2005, an additional 1.3 million
cases will occur each year in South-East Asia. The effect of changing age structure of the population is
clearly shown in the top half of Table VI. With age-specific rales fixed at their 1990 levels, the crude
incident rale would increase from 237 per KM),000 population in IW() to 257 per 100,000 in 2005. This
results from the increasing proportion ol people in the middle and older age groups.
The Western Pacific (Table VII), Eastern Mediterranean (Table IX) and American regions (Table X)
all show a similar pattern: age-specific incidence rates are predicted to fall during 1990-2005 but the number
of new cases each year will continue to increase due to population growth, and the crude incidence rates
(all ages) will continue to increase because of the demographic aging of regional populations.
In sub-Saharan Alrica (Table VIII), the number ol incident cases each year is predicted to almost
triple during 1990-2005. This is the only region where epidemiological factors arc stronger than
k
WHO/TB/93.173
page 11
By the year 2000, 1.4 million (13.8%) of the 10.2 million new cases expected to occur each year will be
attributable to HIV infection. Substantial changes in the prevalence of HIV infection among persons
infected with M.tubcrculosis would alter (he proportion of cases attributable to HIV infection.
4.6. Total incidence during 1990-1999
During the 10-year period 1990-1999 it is estimated that 88.2 million people will develop tuberculosis,
of which 8.0 million cases will be attributable to HIV infection.
In South-Est Asia 35.1 million new cases of tuberculosis will occur during (he decade, of which 2.8
million will be attributable to HIV infection. Around 20.5 million new cases will occur in the Western
Pacific of which 0.4 million will be attributable to HIV infection, while 15.0 million cases will occur in subSaharan Africa of which 3.9 million will be attributable to HIV infection.
At the end of the decade, an extra 2.7 million tuberculosis cases will
>
'..oeach year, compared
be occurring
with 1990. Around 1.1 of these additional cases (40%) will be attributable to the HIV epidemic.
5. Tuberculosis Mortality Estimates
Notification data from Western Europe and other industrialised countries suggest that around 7% of
tuberculosis cases die of the disease (77). These deaths probably result Irom late presentation for treatment
and failure to diagnose tuberculosis as the underlying disease. Notification data from Eastern Europe
su8Ses^ a case-fatality rate of around 15%, while for Central and South America it was assumed that around
20% of tuberculosis cases die of the disease. For other regions, case-fatality rales were estimated using the
approach of Murray and colleagues (1991). A case-fatality rate of 55% is assumed for cases not receiving
treatment, and 15% for those receiving treatment {18-20). It is also assumed that all notified cases have
been treated and that around 5% of treated patients are not notiGed (due to factors such as incomplete
reporting or exclusion of patients treated privately). By applying these assumptions to notification data, it is
estimated the case-fatality rate is around 35% in South-East Asia and the Western Pacific and slightly
higher in the Eastern Mediterranean and sub-Saharan Africa (Table XIX).
Annual tuberculosis mortality was calculated by applying the regional case-fatality rates to the
estimates of annual incidence. The number of tuberculosis deaths attributable to HIV infection was
estimated by applying case-fatality rales to the estimates of HlV^attribulablc cases. These rates were
conservatively assumed to be the same as those among HIV-uninfected cases.
Table XX shows the estimated number of tuberculosis deaths in each region of the world for the years
1990, 1995 and 2000. It is estimated that 2,530,000 tuberculosis deaths occurred in 1990 of which 116,000
were attributable to HIV infection. Around 1.1 million tuberculosis deaths occurred in South-East Asian in
By the year 2000 it is estimated, assuming the proportion of cases treated remains at (he 1990 level,
3.5 million tuberculosis deaths will occur annually, almost 40% more than in 1990. Half a million of these
deaths will be attributable to HIV infection, almost half of which will occur in Sub-Saharan Africa. In
South-East Asia, 1.4 million deaths per year are anticipated by 2000.
The annual number of incident cases of tuberculosis is predicted to increase by 35.6% during 19902000 (Table HI) while annual mortality is expected to increase by 38.7%. A higher proportional increase in
deaths is expected because incidence is predicted to increase most in Africa where treatment rates are
lowest and therefore risk of dying highest.
5.1. Total deaths during 1990-1999
During the 10-year period 1990-1999 it is estimated that 30.0 million people will die of tuberculosis, of
which 2.9 million will be attributable to HIV infection.
V
WH0/TB./9 3 . 173
page 10
demographic factors. Age-specific incidence rates are forecast to increase by around 10 additional cases per
100,000 population per year through to the year 2005. This increase is almost entirely due to the HIV
epidemic.
In Eastern Europe (including independent stales of the former USSR), Western Europe and other
industrialised countries (Tables XI-XII), age-specific incidence rales are forecast to stay at their 1990 level
during 1990-2005, but increases in the number of new cases are expected due to demographic factors.
Tables XIII-XVII show predicted incidence in 1995, 2000 and 2005 for selected countries: China,
India, Indonesia, Pakistan and Brazil.
4.5. Impact of HIV epidemic
The interaction between HIV infection and tuberculosis infection has been reviewed elsewhere (11).
HIV infection in persons with a prior tuberculosis infection dramatically increases their risk of developing
tuberculosis.
Shown in Table XVIII is the prevalence of HIV infection by age among tuberculosis cases in 1990.
Prevalence was estimated from WHO data on HIV seroprevalence in tuberculosis patients from 75
countries in 1989-1992. The data suggest that in 1990 around 25% of tuberculosis cases aged 15-59 were
HIV seropositive in sub-Saharan Africa, 5% in Western Europe and industrialised countries and Central
and South America, 2-3% in South-East Asia and 1-2% in the Western Pacific, Eastern Mediterranean and
Eastern Europe. There was little HIV infection among cases aged more than 60. For children, there were
limited data on the prevalence of HIV infection among those with tuberculosis. Data from Zambia suggest
that 40% of children with tuberculosis were HIV seropositive (77). In Eastern Africa it is estimated that
around 20% of children with tuberculosis were HIV seropositive in 1990. For other regions of the world, it
is assumed few children with tuberculosis were HIV seropositive.
Estimates of the prevalence of HIV infection among tuberculosis cases in 1995 and 2000 (Table
XVIII) were based on previous trends of HIV seroprevalence in tuberculosis patients and on consensus
opinion of programme staff at WHO.
Among adults eo-infeclcd with HIV and M.tiihercidosis, approximately 5-10% develop tuberculosis
each year on average (11-14). By comparison, al most 0.2% of adults with M.tuberculosis infection only
develop tuberculosis each year on average (75). Based on these data, the risk of developing tuberculosis for
persons infected with both HIV and M.tuberculosis is al least 30 limes higher than for persons infected with
M.tuberculosis only. Using a relative risk of 30, the attributable fraction (76) (i.e., (relative risk - l)/relative
risk) for co-infeclion with HIV is 95%. Thus, 95% of tuberculosis cases co-infecled for HIV are attributable
to the HIV infection. The remaining 5% of co-infecled cases would have developed tuberculosis irrespective
of their HIV status and reflect the occurrence of disease in the absence of the HIV epidemic.
The number of incident tuberculosis cases in each region attributable to co-infection with HIV (Table
XVIII) was calculated by applying the attributable fraction and the regional estimates of HIV
seroprevalence among tuberculosis cases to the estimates of regional tuberculosis incidence in 1990, 1995
and 2000.
Of the 7.5 million incident cases of tuberculosis in 1990, 315,000 (4.2%) were attributable to coinfeclion with HIV. These cases would not have occurred if the persons were not co-infected with HIV. In
sub-Saharan Africa, 25% of tuberculosis cases among persons aged 15-59 are estimated to have been HIV
seropositive and 95% of such cases were attributable to HIV. Thus in sub-Saharan Africa, 23.5% of all
lul>erculosis cases among adults aged 15-59 during 1990 were attributable to HIV infection. Less than 2%
of tuberculosis cases in the Western Pacific, Eastern Mediterranean, Europe and other industrialised
countries were attributable to HIV infection.
By the vear 1995 it is estimated 738.1)00 (8.4%) ol the predicted 8.8 million new cases occurring each
year will be attributable to HIV infection. These estimates assume (he prevalence ol HIV inlection among
tuberculosis cases will increase as shown in Table XVIII.
WHO/TB/93.173
page 12
In South-East Asia 12.3 million tuberculosis deaths arc forecast for the decade, of which 1.0 million
will be attributable to HIV infection. Around 6.0 million tuberculosis deaths are expected in Sub-Saharan
Africa, of which 1.5 million will be attributable to HIV infection.
6. Discussion
Nearly 90 million new tuberculosis cases and 30 million tuberculosis deaths are expected to occur
through to the end of this decade without more effective intervention. For a disease where intervention is
cost-effective (21), this is truly staggering.
These estimates are based on notification data and, because of under reporting of tuberculosis cases,
must be considered conservative. This is reflected in our estimated 1990 incidence of 7.5 million new cases
being slightly lower than previous estimates which were based on annual risk of infection data (2). Similarly,
the estimation of tuberculosis mortality should be considered conservative. This study estimated 2.5 million
tuberculosis deaths in 1990, compared with previous estimates of 2.9 million deaths (2). While the exact
number of new cases and deaths is not known, the current and previous estimates are consistent in
suggesting that between 7.5 and 8.0 million new cases and 2.5-3.0 million tuberculosis deaths occurred in
1990.
Current intervention strategies are expected to result in substantial reductions in age-specific incidence
rates in the Eastern Mediterranean region and Central and South America, and to a lesser degree in the
Western Pacific and South-East Asian regions. However, the total number of new cases in these regions is
predicted to increase in the near future because of population growth.
The impact of the HIV epidemic is most evident in sub-Saharan Africa where the number of new
cases per year is forecast to double by the end of the decade. In South-East Asia and other regions there
has been little impact, to date, of the HIV epidemic on tuberculosis. However, by the year 2000 over
500,000 new cases and 200,000 deaths in South-East Asia will be attributable to HIV infection.
A number of assumptions were made in these analyses. It was estimated that 5% of all treated cases
are not reported to WHO and that 100% of reported cases were treated. Limited global data are available
on the completeness and quality of notifications. These levels were chosen as conservative estimates of the
global situation. Earlier mortality estimates (2) used a case-fatality rale of 50% for HIV-positive
tuberculosis cases, whereas the current estimates did not assume mortality was different between HIV
positive and HIV-ncgative cases.
Forecasting future incidence and mortality is difficult and can only be based on data available at the
time of the modelling. Substantial changes in epidemiological factors, such as greater than expected
increases in the seroprevalence of HIV among persons infected with M. tuberculosis, would increase the
future burden of disease. Conversely, increased availability of treatment, would reduce the forecast number
of future cases and deaths.
It has been demonstrated that effective application of short-course chemotherapy in well-managed
national tuberculosis programmes produces excellent results, even under the most adverse conditions (27).
Short-term chemotherapy of smear-positive tuberculosis cases is one of the most cost-effective health
interventions available (22). A higher priority must be given to this disease, both by the countries most
severely affected and by donor countries which invest in health care programmes in those countries.
WHO/TB/93.173
page 13
REFERENCES
1.
Murray, CJ.. Health sector priorities review: tuberculosis. In: Jamison DT & Mosley WH (ed.).
Disease control priorities in developing countries. Oxford University Press, New York, 1993.
2.
Sudre, P., ten Dam, H.G., Kochi, A. Tuberculosis: a global overview of the situation today. Bulletin of
the World Health Organization, 1992; 70:149-159.
3.
Cauthen, G.M., Pio, A., ten Dam, H.G.. Annual risk of tuberculosis infection. WHO/TB/88.L54.
WHO, Geneva, 1988 (this document is available upon request to the Tuberculosis Programme, WHO,
Geneva, Switzerland).
4.
Styblo, K. The relationship between annual risk of tuberculosis infection and the risk of developing
infectious tuberculosis. Bulletin of the International Union against Tuberculosis and Lung Disease, 1985;
60:117-119.
5.
World Health Organization. Tuberculosis notification update, July 1992. WHO/TUB/92.169. Geneva,
World Health Organization, 1992 (this document is available upon request to the Tuberculosis
Programme, WHO, Geneva, Switzerland).
6.
Murray, CJ. Opportunities for tuberculosis operational research in India. Trip report to WHO/TUB
(1992).
7.
Aditama, T.Y. Prevalence of tuberculosis in Indonesia, Singapore, Brunei Darussalam and the
Philippines. Tubercle, 1992; 72: 255-260.
8.
Ministry of Public Health of the People’s Republic of China. Nationwide random survey for the
epidemiology of tuberculosis in 1984-85. Ministry of Public Health, Beijing, 1988.
9.
Ministry of Public Health of the People’s Republic of China. Nationwide random survey for the
epidemiology of tuberculosis in 1990. Ministry of Public Health, Beijing, 1992.
10.
United Nations. Global estimates and projections of population by sex and age, the 1988 revision.
ST/ESA/SER.R/93. New York, United Nations, 1989.
11.
Narain, J.P., Raviglione, M.C. & Kochi, A. HIV-associated tuberculosis in developing countries:
epidemiology and strategies for prevention. Tubercle and Lung Disease, 1992; 73:311-321.
12.
Selwyn, P.A., Hartel, D., Lewis, V.A. ct al. A prospective study of the risk of tuberculosis among
intravenous drug users with human immunodeficiency virus infection. New England Journal of
Medicine, 1989; 320: 545-550.
13.
Braun, M.M-., Badi, N., Ryder, R.W. ct al. A retrospective cohort study of the risk of tuberculosis
among women of childbearing age with HIV infection in Zaire. American Review of Respiratory
Diseases, 1991; 143: 501-504.
14.
Allen, S., Batungwanayo, J., Kerlikowske, K. et al. Two-year incidence of tuberculosis in cohorts of
HIV-infected and uninfected urban Rwandan women. American Review of Respiratory Diseases, 1992;
146: 1439-1444.
15.
Sutherland, I. Recent studies in the epidemiology of tuberculosis, based on I he risk of being infected
with tubercle bacilli. Advances in Tuberculosis Research, 1976; 19: 1-63.
16.
Hennekens, C.H.-, Buring, J.E. Epidemiology in medicine. Little, Brown & Co., Boston, 1987.
17.
Raviglione, M.C., Sudre, P., Rieder, H.L. ct al. Secular trends of tuberculosis in Western Europe.
Bulletin of the World Health Organization, 1993; 71:297-306.
WHO/TB/93.173
page 14
18.
Lindhart, M. The statistics of pulmonary tuberculosis in Denmark, 1925-1934. A statistical
investigation on the occurrence of pulmonary tuberculosis in (he period 1925-1934, worked out on the
basis of the Danish National Health Service file of notified cases and deaths. E. Munksgaard,
Copenhagen 1939.
19.
Humphries, MJ., Byfield, S.P., Derbyshire, J.H. et al. Deaths occurring in newly notified patients with
pulmonary tuberculosis in England and Wales. British Journal of Diseases of the Chest, 1984; 78:149158.
20.
Springett, V.H..Ten-ycar results during the introduction of chemotherapy for tuberculosis. Tubercle,
1971; 52:73-87.
21.
Murray, CJ., dejonghe, E., Chum, HJ. et al. Cost effectiveness of chemotherapy for pulmonary
tuberculosis in three sub-Saharan African countries. Lancet, 1991; 338:1305-1308.
22.
Styblo, K. The impact of HIV infection on the global epidemiology of tuberculosis. Bulletin of the
International Union against Tuberculosis and Lung Disease, 1991; 66:27-32 1.
ACKNOWLEDGEMENTS
The authors would like to (hank Mr R. Bumgarner, Dr M. Felten, Dr P. Graf, Dr P. Nunn, Dr R.
O’Brien, Dr S. Spinaci, Dr B. Vareklzis, and Mrs D. Weil from (he WHO’s Tuberculosis Programme,
Geneva, Switzerland, Dr Peter Smith from (he London School of Hygiene and Tropical Medicine, London,
United Kingdom, and Dr K. Slyblo from the Tuberculosis Surveillance Research Unit, the Hague, the
Netherlands, for (heir useful suggestions and comments on (his paper.
WHO/TB/93.173
page 15
Table I. Estimated tuberculosis incidence in 1990.
Region
Sudre (2)
New
estimates
Cases
Rate1
Cases
Rate1
South-East Asia
Western Pacific2
Africa
Eastern Mediterranean
Americas3
Europe and others4
2,480,000
2,557,000
1,398,000
594,000
564,000
409,000
195
192
265
155
127
34
3,106,000
1,839,000
992,000
641,000
569,000
390,000
237
136
191
165
127
31
All regions
8,002,000
152
7,537,000
143
1Crude incidence rate per 100,000 population
Excludes Japan, Australia and New Zealand
3Excludes USA and Canada
^Independent states of former USSR, USA, Canada, Japan, Australia and New
Zealand
Table II. Age distribution of incident tuberculosis cases
programmes.
Region
South-East Asia
Western Pacific1
China
Africa
Eastern Mediterranean
Americas2
Europe and others3
in 1990/1991,
%
AGE 15-34
%
AGE 35-59
%
5
10
15
10
10
10
10
30
30
20
45
35
35
30
45
35
30
35
40
40
35
AGf 0-14
based
on data from national
tuberculosis
AGE 60+
%
20
25
35
10
15
15
25
Excludes China, Japan, Australia and New Zealand
2Excludes USA and Canada
■^Independent states of the former USSR, USA, Canada, Japan, Australia and New Zealand
s:
x
o
H
W
Cb
OQ
CJ
rt)
O' UJ
so
Table III. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005.
no
2005
2000
1995
1990
cn
n>
Rate1
Cases
Rate1
Cases
Rate1
Cases
Rate1
3,106,000
South-East Asia
1,839,000
Western Pacific2
992,000
Africa
641,000
Eastern Mediterranean
569,000
.Americas
194,000
Eastern Europe*
Western Europe & others5 196,000
237
136
191
165
127
47
23
3,499,000
2,045,000
1,467,000
745,000
606,000
202,000
204,000
241
140
242
166
123
47
23
3,952,000
2,255,000
2,079,000
870,000
645,000
210,000
211,000
247
144
293
167
120
48
24
4,454,000
2,469,000
.2,849,000
987,000
681,000
218,000
217,000
256
151
345
170
117
49
24
7,537,000
143
8,768,000
152
10,222,000
163
11,875,000
176
Region
All regions
Increase since 1990
Cases
16.3%
^rude incidence rate per 100,000 population
2Excludes Japan, Australia and New Zealand
Excludes USA and Canada
-Includes independent states of former USSR
5USA, Canada, Japan, Australia and New Zealand
35.6%
57.6%
cc
Table IV. Estimated change in tuberculosis incidence due to demographic factors only.
2000
1995
1990
2005
Rate1
Cases
Rate1
Cases
Rate1
Cases
Rate1
South-East Asia
' 3,106,000
Western Pacific2
' 1,839,000
992,000
Africa
641,000
Eastern Mediterranean
569,000
Americas3
194,000
Eastern Europe4
Western Europe & others;5 196,000
237
136
191
165
127
47
23
3,505,000
2,057,000
1,163,000
760,000
643,000
202,000
204,000
241
141
192
170
131
47
23
3,966,000
2,281,000
1,370,000
907,000
726,000
210,000
211,000
248
146
193
173
135
48
24
4,477,000
2,512,000
1,612,000
1,047,000
812,000
218,000
217,000
257
153
195
180
139
49
24
7,537,000
143
8,534,000
148
9,671,000.
155
10,895,000
161
Region
All regions
Increase since 1990
Cases
13.2%
28.3%
44.6%
^rude incidence rate per 100,000 population
Excludes Japan, Australia and New Zealand
3Excludes USA and Canada
^Includes independent states of former USSR
5USA, Canada, Japan, Australia and New Zealand
x
o
H
OJ
TJ • '*•
<0
(TQ GJ
ro •
co co
§
o
Table V. Estimated additional cases of tuberculosis, compared with 1990, attributable to changes in demographic and
H
CO
epidemiological factors
Region
2005
.
.... 1995
-............... 2000---demographic epidemiologic demographic epidemiologic demographic epidemiologic
399,000
South-East Asia
218,000
Western Pacific1
171,000
Africa
119,000
East. Mediterranean
74,000
Americas2
8,000
Eastern Europe3
Western Europe & others4 8,000
- 6,000
-12,000
304,000
-15,000
-37,000
0
0
860,000
442,000
378,000
266,000
157,000
16,000
15,000
-14,000
-26,000
709,000
-37,000
-81,000
0
0
1,371,000
673,000
620,000
406,000
243,000
24,000
21,000
-23,000
-43,000
1,237,000
-60,000
-131,000
0
0
997,000
81,07.
234,000
19.0%
2,134,000
79.5%
551,000
20.5%
3,358,000
77.4%
980,000
22.6%
All regions
Excludes Japan, Australia and New Zealand
Excludes USA and Canada
includes independent states of former USSR
\USA, Canada, Japan, Australia and New Zealand
oo
a>
Table VI. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
South-East Asian region.
AGE 0-14
Cases
AGE 15-34
Rate
Rate
Cases
AGE 35-59
Cases
Rate
AGE 60+
ALL AGES
Cases
Rate
Cases
Rate
Demographic factors only
1990
155,000
31
932,000
207
1,398,000
515
621,000
740
3,106,000
237
1995
170,000
31
1,019,000
207
1,595,000
515
721,000
740
3,505,000
241
2000
180,000
31
1,131,000
207
1,838,000
515
817,000
740
3,966,000
248
2005
186,000
31
1,243,000
207
2,119,000
515
929,000
740
4,477,000
257
Demographic and epidemiological factors
1990
155,000
31
932,000
207
1,398,000
515
621,000
740
3,106,000
237
1995
168,000
30
1,016,000
207
1,595,000
514
720,000
740
3,499,000
241
2000
175,000
30
1,126,000
206
514
816,000
739
3,952,000
247
2005
178,000
29
1,235,000
206
1,835;000
h
2,113.,000
514
928,000
739
4,454,000
256
§
o
H
W
tu
00 UJ
(ft
NJ
o
§o
2000 and 2005:
Table VII. Estimated tuberculosis incidence in 1990, 1995,
Western Pacific region, excluding Japan, Australia and New Zealand
AGE 0-14
Cases
AGE 35-59
AGE 15-34
Rate
Rate
Cases
Cases
Rate
H
to
X)
P
OC
ALL AGES
AGE 60+
Cases
Rate
Cases
Rate
Demographic factors only
423,000
80
579,000
166
589,000
487
136
248,000
70
1,839,000
1990
70
438,000
80
662,000
166
693,000
487
141
264,000
2,057,000
199 5
429,000
80
784,000
166
787,000
487
146
281,000
70
2,281,000
2000
70
414,000
80
947,000
166
877,000
487
153
274,000
2,512,000
2005
Demographic and epidemiological factors
70
423,000
80
579,000
166
589,000
487
136
248,000
1,839,000
1990
69
433,000
79
659,000
165
692,000
487
140
261,000
2,045,000
1995
274,000
68
420,000
79
776,000
165
785,000
486
2,255,000
144
2000
264,000
68
401,000
79
932,000
164
872,000
486
2,469,000
151
2005
-*
U)
Table VIII. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
African region
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
ALL AGES
AGE 60+
Cases
Rate
Cases
Rate
Demographic factors only
1990
99,000
42
447,000
266
347,000
380
99,000
398
992,000
191
1995
115,000
42
526,000
266
406,000
380
116,000
398
1,163,000
192
2000
133,000
42
623.000
266
479,000
380
135,000
398
1,370,000
193
2005
150,000
42
736,000
266
569,000
380
157,000
398
1,612,000
195
Demographic and epidemiological factors
99,000
42
447,000
266
347,000
380
99,000
398
992,000
191
251.000
92
626.000
316
460,000
430
130,000
448
1,467,000
242
2000
447,000
142
857,000
366
606,000
480
169,000
498
2,079,000
293
2005
686,000
192
1,152,000
416
794,000
530
217,000
548
2,849,000
345
1990
§
o
H
TJ
CJ
0O
a
w
vO
CJ
NJ
NJ CJ
§o
Table IX. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
Eastern Mediterranean region.
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
no
cu
QO
fl)
ALL AGES
AGE 60+
Cases
Rate
Cases
H
to
N)
UJ U)
Rate
Demographic factors only
1990
64,000
40
224,000
170
257,000
346
96,000
435
641,000
165
1995
73,000
40
261,000
170
311,000
346
115,000
435
760,000
170
2000
81,000
40
307,000
170
380,000
346
139,000
435
907,000
173
86,000
40
350,000
170
452,000
346
159,000
435
1,047,000
180
2005
Demographic and epidemiological factors
1990
64,000
41
224,000
170
257,000
346
96,000
435
641,000
165
67,000
37
256,000
167
308,000
343
114,000
432
745,000
166
1995
2000
67,000
33
294,000
164
372,000
340
137,000
429
870,000
167
2005
64,000
30
329,000
161
439,000
337
155,000
426
987,000
170
Table X. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
American region, excluding USA and Canada.
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
ALL AGES
AGE 60+
Cases
Rate
Cases
Rate
Demographic factors only
1990
57,000
35
199,000
125
228,000
240
85,000
272
569,000
127
1995
60,000
35
217,000
125
267,000
240
99,000
272
643,000
131
2000
62,000
35
235,000
125
315,000
240
114,000
272
726,000
135
2005
64,000
35
251,000
125
365,000
240
132,000
272
812,000
139
Demographic and epidemiological factors
1990
57,000
35
199,000
125
228,000
240
85,000
272
569,000
127
1995
47,000
28
204,000
117
259,000
232
96,000
265
606,000
123
2000
36,000
20
206,000
110
295,000
225
108,000
257
645,000
120
2005
23,000
13
206,000
102
331,000
217
121,000
250
681,000
117
§o
H
UJ
Ca
GO
(D
VO
N)
GJ
§o
Table XI. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
Eastern Europe and independent states of the former USSR.
TJ
00
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
Cases
ALL AGES
AGE 60+
AGE 35-59
Rate
to
kD
GJ
NJ "xl
Ui GJ
Cases
Rate
Cases
Rate
Demographic factors only
19,000
20
58,000
45
68,000
56
49,000
75
194,000
47
1990
20,000
20
56,000
45
75,000
56
51,000
75
202,000
47
1995
19,000
20
55,000
45
81,000
56
55,000
75
210,000
48
2000
19,000
20
56,000
45
86,000
56
57,000
75
218,000
49
2005
Demographic and epidemiological factors
1990
19,000
20
58,000
45
68,000
56
49,000
75
194,000
47
1995
20,000
20
56,000
45
75,000
56
51,000
75
202,000
47
19,000
20
55,000
45
81,000
56
55,000
75
210,000
48
2000
2005
19,000
20
56,000
45
86,000
56
57,000
75
218,000
49
Table XII. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005:
Western Europe, USA, Canada, Japan, Australia and New Zealand.
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
ALL AGES
AGE 60+
Cases
Rate
Cases
Rate
Demographic factors only
1990
20,000
10
59,000
22
68,000
27
49,000
37
196,000
23
1995
20,000
10
56,000
22
76,000
27
52,000
37
204,000
23
2000
19,000
10
55,000
22
82,000
27
55,000
37
211,000
24
2005
19,000
10
56,000
22
85,000
27
57,000
37
217,000
24
Demographic and epidemiological factors
1990
20,000
10
59,000
22
68,000
27
49,000
37
196,000
23
1995
20,000
10
56,000
22
76,000
27
52,000
37
204,000
23
2000
19,000
10
55,000
22
82,000
27
55,000
37
211,000
24
2005
19,000
10
56,000
22
85,000
27
57,000
37
217,000
24
§
o
H
W
TJ
P
0Q
ro
UD
UJ
NJ
GN GJ
§o
Table XIII. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005: China.
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
H
CD
X)
Cu
ALL AGES
AGE 60+
Cases
Rate
Pases
OQ
fD
Rate
Demographic factors only
1990
193,000
64
257,000
58
386,000
132
451,000
444
1,287,000
113
1995
204,000
64
264,000
58
437,000
132
527,000
444
1,432,000
117 '
2000
216,000
64
255,000
58
512,000
132
595,000
444
1,578,000
122
2005
208,000
64
243,000
58
611,000
132
656,000
444
1,718,000
127
Demographic and epidemiological factors
1990
193,000
64
257,000
58
386,000
132
451,000
444
1,287,000
113
1995
204,000
64
264,000
58
437,000
132
527,000
444
1,432,000
117
2000
216,000
64
255,000
58
512,000
132
595,000
444
1,578,000
122
2005
208,000
64
243,000
58
611,000
132
656,000
444
1,718,000
127
GJ
NJ
--J
GJ
Table XIV. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005: India.
AGE 0-16
Cases
AGE 15-36
Rate
Cases
:
Rate
AGE 35-59
Cases
Rate
AGE 60 I
ALL AGES
Cases
Rate
Cases
Rate
Demographic factors only
1990
103,000
31
619,000
212
929,000
526
613,000
756
2,066,000
262
1995
113,000
31
677,000
212
1,061,000
526
680,000
756
2,331,000
266
2000
120,000
31
752,000
212
1,222,000
526
563,000
756
2,637,000
253
2005
126,000
31
826,000
212
1,609,000
526
618,000
756
2,977,000
262
Demographic and epidemiological factors
1990
103,000
31
619,000
212
929,000
526
613,000
756
2,066,000
262
1995
120,000
33
686,000
216
1,065,000
528
681,000
758
2,350,000
268
2000
135,000
35
766,000
216
1,231,000
530
566,000
760
2,678,000
257
2005
167,000
37
850,000
218
1,625,000
532
623,000
762
3,065,000
268
§
o
H
co
cu o
OG
CD
NJ
Table XV
s:
5:
o
H
to
Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005: Indonesia.
AGE 15-34
AGE 0-14
Cases
Rate
Rate
Cases
AGE 35-59
Cases
Rate
T3
P kO
00 U)
Q
ALL AGES
AGE 60+
Cases
Rate
Cases
Rate
Demographic factors only
22,000
31
134,000
212
201,000
526
89,000
756
446,000
242
1990
24,000
31
145,000
212
226,000
526
102,000
756
497,000
246
1995
25,000
31
158,000
212
257,000
526
114,000
756
554,000
253
2000
25,000
31
170,000
212
290,000
526
127,000
756
612,000
262
2005
Demographic and epidemiological factors
22,000
31
134,000
212
201,000
526
89,000
756
446,000
242
1990
1995
20,000
26
141,000
206
225,000
521
101,000
753
487,000
241
17,000
21
150,000
201
252,000
516
113,000
746
532,000
243
2000
2005
13,000
16
158,000
197
282,000
511
124,000
741
577,000
247
ro
Table XVI. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005: Pakistan.
AGE 0-14
Cases
AGE 15-34
Rate
Cases
Rate
AGE 35-59
Cases
Rate
AGE 60+
ALL AGES
Cases
Rate
Cases
Rate
Demographic factors only
1990
31,000
61
107,000
257
122,000
523
46,000
658
306,000
250
1995
35,000
61
124,000
257
147,000
523
55,000
658
361,000
254
2000
38,000
61
144,000
257
178,000
523
65,000
658
425,000
262
2005
41,000
61
165,000
257
213,000
523
75,000
658
494,000
269
Demographic and epidemiological factors
1990
31,000
61
107,000
257
122,000
523
46,000
658
306,000
250
1995
35,000
61
124,000
257
147,000
523
55,000
658
361,000
254
2000
38,000
61
144,000
257
178,000
523
65,000
658
425,000
262
2005
41,000
61
165,000
257
213,000
523
75,000
658
494,000
269
§
o
UJ
U
CJ
0O
(D
bJ
O CO
x
c
Table JCVII. Estimated tuberculosis incidence in 1990, 1995, 2000 and 2005: Brazil.
H
OJ
AGE 0-14
AGE 15-34
AGE 35-59
ALL AGES
AGE 60+
p UD
00
Cases
Rate
Rate
Cases
Cases
Rate
Cases
Rate
Cases
Rate
— uj
Demographic factors only
1990
19,000
35
67,000
125
76,000
240
29,000
272
191,000
127
1995
20,000
35
73,000
125
90,000
240
33,000
272
216,000
130
2000
21,000
35
78,000
125
105,000
240
38,000
272
242,000
135
2005
21,000
35
83,000
125
121,000
240
44,000
272
269,000
139
Demographic and epidemiological factors
1990
19,000
35
67,000
125
76,000
240
29,000
272
191,000
127
1995
16,000
28
68,000
117
87,000
233
32,000
264
203,000
123
2000
12,000
20
69,000
110
98,000
224
36,000
257
215,000
120
2005
8,000
13
68,000
102
110,000
218
40,000
248
226,000
117
Table XVIII. Estimated number of HIV-attributable tuberculosis cases in 1990,
1990
region
ADULTS AGED 60+
All regions
HIV +ve TB cases
TB cases
1995 and 2000.
2000
1995
HIV attributed
TB cases (X)
HIV +ve TB cases
TB cases
HIV attributed
TB cases (X)
HIV +ve TB cases
TB cases
HIV attributed
TB cases (X)
OX
1,588,000
0
OX
1,855,000
0
OX
2,125,000
0
ADULTS AGED 15-59
3X
South-East Asia
Western Pacific^
2X
25X
Africa
Eastern Mediterranean 2X
5X
Americas*9Eastern ^urope^
IX
Western Europe4
5X
2,330.000
1,002,000
794,000
481,000
427,000
126,000
127.000
66,000 (2.8X)
19,000 (1.9X)
189,000 (23.8X)
9,000 (1.9X)
20,000 (4.8X)
1,000 (0.9X)
6,000 (4.8X)
10X
3X
35X
3X
10X
2X
10X
2,611,000
1,092,000
1,086,000
564,000
463,000
131,000
132,000
248,000 (9.5X)
31,000 (2.6X)
361,000 (33.3X)
16,000 (2.8X)
44,000 (9.5X)
2,000 (1.9X)
13,000 (9.5X)
20 X
6X
40X
6X
20X
5X
20X
2,961,000
1,196,000
1.463,000
666,000
501,000
136,000
137,000
563,000
68,000
556,000
38,000
95,000
6,000
26,000
All regions
5,287,000
310,000
6,079,000
715,000 (U.8X)
63,000
188,000
168,000
47,000
368,000
15,000 (23.7X)
4,000 (1.9X)
3,000 (1.9X)
1,000 (1.9X)
0
CHILDREN AGED 0-14
East Africa^
Other African
South-East Asia
A-mericas^
Other regions
All regions
TOTAL
20X
OX
OX
OX
OX
(6.OX)
(19.OX)
(5.7X)
(38.OX)
(5.7X)
(19.OX)
(4.8X)
(19.OX)
7,060,000 1,352,000 (19.2Z)
113,000
334,000
175,000
36,000
379,000
32,000 (28.OX)
16,000 (4.8X)
8,000 (4.8X)
2,000 (4,8X)
0
25,000
74,000
155,000
57,000
351,000
5,000 (19.OX)
0
0
0
0
662.000
5,000
(0.8X)
834,000
23,000
(2.8X)
1,037,000
7.537,000
315,000
(4.2X)
8,768,000
738,000
(8.4X)
10,222,000 1,410,000 (13.81)
25X
2X
2X
2X
OX
30X
5X
5X
5X
OX
58,000
(5.6X)
^Excludes Japan. Australia and New Zealand
^Excludes USA and Canada
^Includes independent states of the former USSR
’Includes USA, Canada. Japan, Australia and New Zealand
"Burundi. Kenya, Malawi, Mozambique, Rwanda, Tanzania, Uganda, Zambia, Zimbabwe
§
o
H
to
TJ
OJ
(JQ
kO
GJ
bJ Co
x
o
H
Table XIX. Estimated proportion of incident cases receiving treatment in 1990.
W
Ci)
Region
South-East Asia
Western Pacific*
Africa
Eastern Mediterranean
Notified
cases1
Expected
cases
Percent
notified
1,510,311
875,098
370,359
252,413
3,176,000
1,839,000
1,003,000
660,000
47.6%
47.6%
36.7%
38.2%
Case-Fatality
treated2
Rate3
50.1%
50.1%
38.6%
40.2%
35.0%
35.0%
39.6%
38.9%
Notified cases based on most recent report (WHO/TUB/92.169)
Calculated assuming 5% of treated patients are not included in notification data
Calculated assuming 15% mortality among treated cases and 55% mortality among untreated cases
’Excludes Japan, Australia and New Zealand
Q
GJ
Table XX. Estimated tuberculosis deaths in 1990, 1995 and 2000, assuming regional treatment coverage rates remain at
their 1990 level.
Region
1990 Deaths
1995 Deaths
2000 Deaths
HIV
attributed
Total
HIV
attributed
Total
HIV
attributed
South-East Asia
1,087,000
Western Pacific1
664,000
Africa
393,000
Eastern Mediterranean
249,000
Americas2
114,000
Eastern Europe3
29,000
Western Europe 6c others4 14,000
23,000
7,000
77,000
4,000
4,000
<200
<500
1,225,000
716,000
581,000
290,000
121,000
30,000
14,000
88,000
11,000
150,000
6,000
9,000
<600
1,000
1,383,000
789,000
823,000
338,000
129,000
32,000
15,000
200,000
24,000
239,000
15,000
19,000
<900
2,000
All regions
116,000
(4.6%)
2,977,000
266,000
(8.9%)
3,509,000
500,000
(14.2%)
Total
2,530,000
Increase since 1990
17.7%
38.7%
Excludes Japan, Australia and New Zealand
"Excludes USA and Canada
"Includes independent states of former USSR
'‘USA, Canada, Japan, Australia and New Zealand
s:
tn
o
to
TJ
0)
(TO
c
vO
I
I,
C4 I
I
5
ORIGINAL ARTICLES
TUBERCULOSIS
Over the past two years, the World Bank has been undertaking, with a number of collaborators, a "Health
Sector Priorities Review”. The core of this Review is a series ofstudies on the public health significance
of major clusters of diseases in the developing world and on the costs and effectiveness of currently
available technologies for (heir prevention and case management. This analysis oftuberculosis supported
as one of these studies has proven to be one of the most surprising; because of the tremendous burden of
tuberculosis and the existence of interventions ofproven efficacy that are some of the most cost-effective
in the international public health armamentarium.
The present document (some clinical parts have been shortened for our readers who specialize in
tuberculosis - the figures and content are final) comprises Chapter 11 ofEvolving Health Sector Priorities
in Developing Countries edited by Dean T. Jamison and W. Henry Mosley (see below in Introduction).
Tuberculosis in developing countries :
burden, intervention and cost
C.J.L. MURRAY*, K. STYBLO** and A. ROUILLON**
• Harvard School of Public Health, 9 Bow Street, Cambridge, Ma. 02138, USA.
•• International Union Against Tuberculosis and Lung Disease, 68 boulevard Salnt-Mkhel, 75006 Paris, France.
Introduction
Drafts of other components of the Review are now avail
able from the Population, Health and Nutrition Division,
Room S-6141, the The World Bank, Washington DC 20433.
The conclusions of these studies, including the one on
tuberculosis, do not necessarily reflect the policies of the
World Bank.
Tuberculosis is an ancient disease that has long been a
major public health challenge in the world, and remains a
major health problem in developing countries. In the last
century, tuberculosis was responsible for nearly one in ten
deaths in Europe (Preston et al., 1972). There is reliable
evidence that irrespective of its magnitude, the tuberculosis
problem in developed countries has been decreasing at least
for the last40 years, after the introduction of antituberculosis
chemotherapy. In many developed countries, a steady de
crease in mortality from tuberculosis in the pre-chemother
apy era was observed from the turn of this century if not
before (Frost, 1937 ; Styblo, 1980). The elimination of
tuberculosis in most developed countries will not be substan
tially influenced by AIDS because of the low prevalence of
tuberculous infection in subjects aged 20 to 50 years in whom
HIV infection is most frequent (Styblo, 1986, 1989).
On the other hand, in developing countries tuberculosis
continues to be a major problem and there appears to have
been virtually no tendency for tuberculosis to eliminate
6
itself, in the absence of intensive control measures. Unlike in
developed countries, HIV infection will result in a consider
able increase of tuberculosis cases in those developing coun
tries where both tuberculous and HIV infections are preva
lent Tuberculosis remains, therefore, one of the top priori
ties for action in developing countries, since tools exist to
diagnose and cure infectious cases of tuberculosis and thus to
decrease transmission of tuberculous infection.
The purpose of this article is to review the present status
of tuberculosis in the world, with emphasis on the situation in
developing countries, and to examine various policy options
concerning the prevention and treatment of tuberculosis,
with attention to their cost-effectiveness.
Tuberculosis incidence and mortality
Tuberculosis incidence
To put tuberculosis in the proper perspective we need to
know the number and the age-distribution of new cases of
tuberculosis which develop in a community each year, as
well as the number and the age-distribution of patients who
die from tuberculosis each year. Health information systems
in developing countries are too incomplete to provide mean
ingful information on the incidence or mortality of tubercu
losis (Styblo and Rouillon, 1981). We are forced to estimate
the burden of tuberculosis indirectly using several epidemi
ological parameters. These include the average annual risk of
tuberculous infection and the incidence of smear-positive
pulmonary tuberculosis, the proportion of all cases of tuber
culosis that are smear-positive and case-fatality rates for
smear-positive tuberculosis and other tuberculosis.
Annual average risk of tuberculous infection
Tuberculosis epidemiologists have used the ability to
detect the presence of infection using skin tests to measure
the prevalence of infection in communities. A technique has
been developed for converting this information on preva
lence of tuberculous infection into a series of annual risks of
tuberculous infection (Styblo et al., 1969a ; Sutherland,
1976). If several tuberculin surveys of the same population
have been made at different times (using similar techniques
and testing a representative sample of non-BCG-vaccinated
subjects of the same age) the level of and percentage decrease
n the risk of infection can be estimated. The annual risk of
.nfection tells us the probability that any individual will be
infected or reinfected with M. tuberculosis in one year. This
measure has become the standard indicator of the tuberculo
sis burden in a community (Leowski, 1988).
Since the 1950’s a variety of tuberculin sensitivity surveys
in developing countries provide us with an approximate
picture of the annual risk of infection in different regions of
the developing world. Table 1 presents our best estimates
based on a recent review of survey data on the annual risk of
infection (Cauthen et al., 1988). The annual risk of tubercu
lous infection is probably highest in sub-Saharan Africa,
followed closely by South and East Asia. For comparison the
annual risk of infection in the Netherlands in 1985 was
0.012% (Styblo, 1989).
Incidence of smear-positive tuberculosis
The natural history' of tuberculosis illustrates that the main
source of infection are patients suffering from smear-posi
tive pulmonary tuberculosis. For the rest of this paper,
therefore, tuberculosis will be divided into two categories: i)
'.putum smear-positive tuberculosis, which will be referred
by the shorthand smear-positive tuberculosis, and ii) other
tuberculosis, which includes those cases of pulmonary tuber-
600
Incidence Smear ♦ TB Per 100,000
600
400
300
200
100
o
0
2
4
6
8
10
Annual Risk of Infection
Source: Styblo. unpubllehed.
Figure I. The relationship between the annual risk of infection
and the incidence of smear-positive tuberculosis.
TabU 1. Estimated risks oftuberculous infection and their trends
in developing countries, 1985-90
Area
Estimated risk
of tuberculous
infection (%)
Estimated annual
decrease in risk
of infection (%)
Sub-Saharan Africa
North Africa and
Western Asia
Asia
South America
Central America and
Caribbean
1.50-2.50
1-2
0.50 -1.50
1.00-2.00
0.50 -1.50
5-6
1- 3
2- 5
0.50 -1.50
1 -3
Source : Based on data presented in Cauthen el al. (1988).
culosis that are sputum smear-negative and extra-pulmonary
tuberculosis. As children rarely suffer from sputum smear
positive tuberculosis, most cases of tuberculosis in children
will be included in the category other tuberculosis. (If chil
dren are smear-positive, they are, as adult patients, highly
infectious sources of infection. If they are smear-negative/
culture-positive or smear-negative/culture-negative they are
much less infectious.) The above two categories are some
times labelled infectious tuberculosis and non-infectious
tuberculosis, respectively (e.g. Ministry of Health and Fam
ily Welfare, India 1986). The distinction between sputum
smear-positive tuberculosis and other tuberculosis is particu
larly important when considering the policy options for
tuberculosis control and prevention (Rouillon et al., 1976).
Lack of data on smear-positive tuberculosis cases in
developing countries makes it difficult to convey the enor
mity of the tuberculosis problem to the public health commu
nity. It is not possible to readily obtain reliable information
on incidence of smear-positive tuberculosis in developing
countries because case-detection rates can be only a fraction
of the respective true incidence rates.
The relationship between the annual risk of infection and
the incidence of smear-positive tuberculosis can provide one
of the only means of estimating the incidence of smear
positive tuberculosis (Styblo, 1985,1988). Styblo examined
the relationship between the annual risk of infection and the
incidence of smear-positive pulmonary tuberculosis using a
variety of data sources from the developing and developed
world. We have recomputed this relationship using only the
results of a series of WHO sponsored surveys in developing
countries and data from the Netherlands before chemother
apy was widely available. We must note that for some of
these surveys data are available on the prevalence of smear
positive tuberculosis, not the incidence. In such cases, the
incidence rates were derived using the historical observation
that the prevalence of smear-positive tuberculosis was usu
ally twice the incidence in the communities without wide
spread institution of chemotherapy (Holm, 1970). In these
developing countries, the relationship beween the annual
risk of infection and incidence of pulmonary smear-positive
tuberculosis was linear. A least squared regression line
(Figure I) gives an estimate of 49 cases of smear-positive tu
berculosis per 100,000 for every 1 % annual risk of infection.
The 95 % confidence interval for the coefficient is 39 to 59.
7
Table 2. Estimated incidence of smear-positive tuberculosis in
developing countries, 1990
Area
Estimated number of cases Incidence
Low
Midpoint
High
rate
Sub-Saharan
Africa
296,000
North Africa and
Western Asia
53,000
Asia
1,142,000
South America
57,000
Central America
and Caribbean 30,000
Total
1,578,000
521,000
745,000
103
146,000
2,298,000
160,000
239,000
3,455,000
263,000
54
79
54
83,000
136,000
54
3,208,000
4,838,000
77
Note : Based on the annual risk of infection for each region
presented in Table 1, the 1990 population and an incidence of
39 to 59 cases per 100,000 for each one percent annual risk of
infection.
Using the estimates of the risk of infection for different
regions in Table 1 and the confidence interval for the rela
tionship between incidence of pulmonary smear-positive
tuberculosis and the risk of infection, the low and high
estimates of the incidence of smear-positive tuberculosis for
different regions in Table 2 have been calculated. The
midpoint of the confidence interval of the estimates of
smear-positive incidence is 3,208,000 cases or an incidence
of 77 per 100,000 in the developing world. These must be
viewed as only crude estimates, which nevertheless illustrate
the continuing magnitude of the tuberculosis problem.
Age distribution of smear-positive tuberculosis
The age-distribution of incidence is important in deter
mining the public health impact of smear-positive tuberculo
sis and the most appropriate means of preventing or control
ling tuberculosis. From the historical record of developed
countries and epidemiological models, the age and sex
distribution of incidence appears to change as the annual risk
of infection declines. As the majority of developing countries
have an annual risk of tuberculous infection between 1.0 %
and 2.0 %, we propose to use the age-distribution of the
incidence of smear-positive tuberculosis from a developing
country with an annual risk of infection in this range (Tanza
nia). There is no reason to believe that the epidemiology and
thus the age distribution of incidence for a given annual risk
of infection will vary substantially between communities.
Because the tuberculosis control programme in Tanzania is
well organized and captures the majority of tuberculosis
cases, the age-distribution from Tanzania will be used as
representative of the developing world where the risk of
tuberculous infection is between 1 and 2 %. Figure II shows
the age-distribution of smear-positive tuberculosis in Tanza
nia for 1985-87 (Chum el al., 1987 ; Chum, 1989). The
pattern is similar to other developing countries such as
Mozambique, Malawi and Benin that have good reporting
systems (IUATLD 1988). It is important to note that BCG
coverage in Tanzania was roughly 50 % in 1983-87 (Bleiker
et al., 1987) - based on scar examination in the National
8
Tuberculin Survey in Tanzania carried out from 1983 to 1987
on 80,000 schoolchildren from 20 regions selected at random
which is below the officially reported average for the devel
oping world (UNICEF, 1988). Thus, any effect such BCG
coverage may have on preventing tuberculosis in children is
partially represented in the age-distribution ; as world BCG
coverage is probably higher than in Tanzania, the estimate
for the incidence of smear-positive tuberculosis in children
based on this age-distribution may be slightly high. Clearly,
smear-positive cases are relatively rare in children ; smear
positive tuberculosis is concentrated in adults - more than
80 % of cases occur between the ages of 15 and 54, according
to the data from Tanzania.
Incidence of other forms of tuberculosis
Estimates of the incidence of smear-negative pulmonary
and extra-pulmonary tuberculosis are also needed. These
forms of tuberculosis are particularly difficult to quantify as
the major diagnostic tool used in developing countries,
sputum microscopy, does not detect these cases. Because the
diagnosis of extra-pulmonary tuberculosis is often based on
clinical criteria, no survey data are available to estimate the
relationship between the risk of infection and other tubercu
losis. In the past, estimates of smear-positive tuberculosis
have simply been doubled to provide a figure for other
tuberculosis (Styblo and Rouillon, 1981 ; Leowski, 1988).
The distribution of total cases between the categories sputum
smear-positive and other tuberculosis cannot be accurately
established. Whereas smear-positive tuberculosis and tuber
culosis positive by culture only can be objectively deter
mined, the number of culture-negative cases detected de
pends on various factors, such as active case-finding by Mass
Miniature Radiography (MMR) extensively used in Europe
in the 1950’s, 1960’s and 1970’s, criteria for activity in
asymptomatic cases detected by active case-finding, agegroups, etc. However, we will assume that within each agegroup using the same diagnostic approach the percentage of
cases that are sputum smear-positive and other should be the
same relatively independent of the overall annual risk of
Percent of Cases
30 |------------------------------
26
20
66*
Figure ll. The age distrihu;i'.r. of smear-positive tuberculosis
detected in Tanzania, 1985-1987.
infection. Using data for the United States and Norway,
Figure III illustrates the proportion of all tuberculosis cases
that are smear-positive by age (Gaining, 1955; CDC, 1989).
Because the data set for the US is larger and no MMR was
used on a large scale, we will use the ratio of cases of other
tuberculosis to smear-positive tuberculosis within each agegroup in the United States. Using the age-distribution of the
incidence of smear-positive tuberculosis in Tanzania and the
age-specific ratios of other to smear-positive in the United
States, we have derived a rough estimate of the age-distribu
tion of other tuberculosis shown in Figure IV. While the
assumptions underlying these estimates of other tuberculosis
may be challenged on many grounds, we feel it is preferable
to make some objective attempt to estimate the age-distribu
tion of smear-negative and extra-pulmonary tuberculosis in
developing countries because it is an important input to
policy decisions.
Our estimations imply that there are 1.22 cases of smear
negative and extra-pulmonary tuberculosis for every case of
smear-positive tuberculosis in developing countries with an
annual risk of infection between 1 and 2 % and an overall
age-distribution similar to Tanzania. Table 3 provides low
and high estimates of the number of new cases of smear
negative and extra-pulmonary tuberculosis for each region in
the developing world. For all types of tuberculosis combined,
Table 4 indicates that the incidence of tuberculosis exceeds
220 per 100,000 in sub-Saharan Africa.
Percent Smear Poeltlve
70-f
60-
60-
fi
■
4030-
I
2010-
0
0-14
26-34
16-24
36-44
46-64
55-64
66*
Age-Groups
■ USA 1986-87
Norway 1951-72
Bourcaa: CDC and T8RU.
Figure III. The percent ofaU cases oftuberculosis that are smear
positive, USA 1985-1933 and Norway, 1951-1972.
Percent of Cases
30 |--------------------------
20
Tuberculosis mortality
Case-fatality rates, untreated
In order to calculate tuberculosis mortality from the
estimates of incidence derived above, we need to estimate
the case-fatality rate. Without appropriate chemotherapy,
tuberculosis is highly fatal. The results of several studies in
developed countries before chemotherapy became available
demonstrated mortality rates consistently in the 50 % to 60 %
range (Drolet, 1938 ; Lindhart, 1939 ; Gaining Hansen,
1955). These observations were confirmed in the five-year
study of the natural history of tuberculosis in Bangalore,
Table 3. Estimated incidence of other tuberculosis in developing
countries, 1990
Area
Estimated number of cases Incidence
Low
Midpoint
High per 100,000
Sub-Saharan
Africa
361,000
North Africa and
Western Asia
64,000
Asia
1,393,000
South America
71,000
Central America
and Caribbean
37,000
909,000
126
178,000 291,000
2,804.000 4,215.000
Total
1,926,000
635,000
196,000
321,000
66
96
66
101,000
166,000
66
3.914,000 5,902,000
94
Note : The incidence of other tuberculosis has been based on
USA data showing the relationship between smear-positive
tulxzrculosis and other tuberculosis by age combined with the
age-distribution of smear-positive tuberculosis in Tanzania.
16-24
26-34
36-44
45-64
56-64
65*
Age-Groups
Figure IV. The estimated incidence of smear-positive and other
tuberculosis by age for the developing world in 1990.
India: 49 % of bacteriologically confirmed cases (smear- and
culture-positive cases or smear-negative and culture-posi
tive) died within five years (National Tuberculosis Institute,
Table 4. Estimated incidence of all forms of tuberculosis in
developing countries, 1990
Area
Estimated number
Incidence
of cases
per 100,000
Low
Midpoint
High
Sub-Saharan
Africa
656,000 1,156,000
North Africa and
Western Asia
117,000
323,000
Asia
2,535,000 5.102,000
South America
129,000
356,000
Central America
and Caribbean
66,000
185,000
Total
1,655,000
229
530,000 120
7,670,000 174
584,000 120
302,000
120
3,503,000 7,122,000 10,741,000
171
9
Bangalore, 1974). As expected, the case-fatality rate for
smear-positive tuberculosis is even higher; Rutledge and
Crouch (1919) and Lindhart (1939) reported 66 % mortality
in these cases (no information is available from the Banga
lore study). For the rest of this paper, we will assume that the
case fatality rate for smear-positive tuberculosis is 60 % to
70 %, for other tuberculosis as a whole it is 40 % to 50 % and
for all forms combined it is 50 % to 60 %.
Tuberculosis death rates in developing countries
The tuberculosis death rates in developing countries can- .
not be as high as implied by the incidence rates and a case
fatality rate of 50-60 % because a significant proportion of
cases are detected and treated, which lowers the tuberculosis
death rate. Accordingly, we adjusted our estimates for this
effect by using estimates of the number of cases that receive
treatment to derive the likely range of tuberculosis death
rates in developing countries. For all those cases that are
detected and receive treatment, we assume the case-fatality
rate is reduced to 20 % after 5 years. For example, in the East
African and British Medical Research Council surveys in
Tanzania and Kenya the case-fatality rates for patients re
ceiving standard chemotherapy were 12 % and 16 % respec
tively after 12 months (EAMRC 1977,1979). In many coun
tries, however, the case-fatality rate may be over 20 % for
those receiving chemotherapy, after five years of follow-up,
making the following estimates of mortality conservative.
Estimates of the percentage of new cases that are detected
and treated are based on the number of cases of tuberculosis
detected that are reported by countries to the World Health
Organization (Table 5) (WHO, 1988). Because reporting is
extremely variable, these estimates are based on the highest
number of cases reported by each country for any year in the
last decade. This is justified by the assumption that year to
year variation in the number of cases reported, which can be
greater than an order of magnitude, is due more to incomplete
reporting of health service activities than to change in the
epidemiology of tuberculosis. In addition, the highest num
ber of cases reported in the last ten years has been adjusted
Table 5. Estimated tuberculosis cases detected and case fatality
rates in developing countries, 1990
Area
Estimated
cases
detected
Percent
Estimated case
all cases fatality rates (%)
detected
Low
High
Sub-Saharan Africa 325,000
North Africa and
223,000
Western Asia
3,087,000
Asia
222,000
South America
Central America
50,000
and Caribbean
28
41
48
69
61
62
28
30
32
31
34
36
27
41
49
3,907,000
55
32
37
Total
Note : The calculations of effective case-fatality rates are
based on the assumption that 15 % of those patients receiving
standard chemotherapy die. As discussed in detail in the
section on chemotherapy this is a conservative assumption.
10
upwards by 20 % to try and account for those cases that are
detected in the private sector that do not report to the
government; in Asia where data for some large countries
may include a large number of retreatment cases we have not
adjusted the figures by 20 %.
Separate estimates for the percent of smear-positive and
other cases are needed. As the primary means of detecting
tuberculosis in developing countries is sputum microscopy,
the detection rate of smear-positive tuberculosis is higher
than for smear-negative or extra-pulmonary tuberculosis.
Based on data from the National Tuberculosis and Leprosy
Programme in Tanzania, we will assume that 60 % of
detected cases are smear-positive and 40 % are other tuber
culosis. The detection rate of the various forms of tuberculo
sis and the likely range of case fatality rates discussed above
can be combined to estimate the tuberculosis death rates
from smear-positive tuberculosis and other tuberculosis.
Because the detection rate for smear-positive tuberculosis is
higher despite a higher case fatality rate, the overall death
rate from smear-positive tuberculosis is similar to the death
rate from other tuberculosis.
Table 6. Estimated deaths from all forms of tuberculosis in
developing countries, 1990
Area
Estimated number of cases Deaths
per
Low
Midpoint
High
100,000
528,000
790,000
104
99,000
1,709,000
125,000
166,000
2,646,000
211,000
37
58
42
88,000
148,000
57
1,139,000 2,549,000
3,961,000
61
Sub-Saharan Africa 266,000
North Africa and
Western Asia
33,000
Asia
771,000
South America
41,000
Central America
and Caribbean
28,000
Total
Table 6 shows estimated deaths each year from all forms
of tuberculosis for regions based on the calculations of the
tuberculosis death rates discussed above. The wide confi
dence intervals reflect the cumulative uncertainty in the
paremeters of the estimation procedure. Using the midpoints
of the confidence intervals, the total number of deaths from
tuberculosis in the developing world comes to 2,549,000.
Tuberculosis, therefore, accounts for approximately 6.7 %
of al! deaths in the developing world in 1990 (United
Nations, 1989).
Age-distribution of tuberculosis deaths
To estimate the age-distribution of tuberculosis deaths,
we must take into consideration the age-distribution of new
cases and the relationship between case-fatality rates and
age. Clearly, the relationship is complex ; for example, the
death rates may also vary by age because certain age-groups
may be more likely to seek treatment and be cured. With
hesitation, we will apply the age-specific case-fatality rates
from London 1933-1934 to the age-distribution of total
tuberculosis incidence derived above (Styblo, 1984). Tuber
culosis case-fatality rates tend to increase steadily at older
Tablt 7. Distribution of tuberculosis deaths by age
Agegroup
Czechoslovakia
1940
1931
Norway
1941
1951
1931
Netherlands
1941
1951
0-14
15-24
25-34
35-44
45-54
55-64
65+
11.7
22.0
18.7
14.0
125
11.4
9.7
11.8
30.6
25.9
14.5
7.7
5.0
45
103
25.4
25.4
16.1
9.6
6.6
6.5
8.0
10.8
24.4
195
13.2
10.7
13.4
24.0
22.4
20.8
11.7
7.7
63
7.1
19.4
203
20.7
13.1
9.6
8.2
8.7
13.6
12.8
16.9
12.8
11.6
13.4
18.9
3.7
1.8
05
Risk of infection
5.5
(1938)
Age-distribution of deaths have been adjusted using the population age-structure of the developing world in 1990.
ages. Figure V provides the crude estimates of the age
pattern of tuberculosis deaths in a country with an annual risk
of infection of 1 % to 2 % where the probability of detection
is equal for smear-positive tuberculosis across all age-groups
and equal for other tuberculosis across all age-groups.
This estimated pattern can be compared to the age-distri
bution of tuberculosis deaths in Western countries when the
annual risk of infection was similar to that now seen in the
developing world. Table 7 illustrates the age-distribution of
tuberculosis deaths adjusted to the age-structure of the devel
oping world in Czechoslovakia, Norway and the Netherlands
(TSRU, 1966). The percentage of deaths in children under 15
ranged from approximately 10 % to 20 %. In the Netherlands,
the tuberculosis death rates in children were considerably
higher than in Czechoslovakia even at lower risks of infec
tion. Clearly, there are other variables that are major determi
nants of the reported age-distribution of tuberculosis death
rates. One explanation may be the high rates of M. bovis
infection in the Netherlands at the time. According to our
estimates for Tanzania, less than 8 % of tuberculosis deaths
occur in children under age 15, which is less than in the three
70-'
60 60 40 -
30 -
fex
20 ■
w
100 ■
1937
1942
1962
1947
1957
Year
Figure VI. The shifting age-structure of tuberculosis deaths as
the annual risk of infection declines, USA 1937-1957.
Percent of Cases
26 |--------------------------
20
15
16-24
25-34
35-44
45-64
56-64
66*
Age-Groups
Figure V. Estimated age-distribution oftuberculosis deaths in the
developing world in 1990.
developed countries in Table 7. This may be due to the higher
BCG coverage in Tanzania now than in these countries at the
time. Variation in the age pattern of tuberculosis deaths
highlights the tentative nature of the estimates presented
here. The basic conclusion, however, that tuberculosis is
concentrated in the adult age-groups, appears to be robust
As the discussion above implies, the age pattern of tuber
culosis deaths shifts towards higher ages as the annual risk of
infection declines. Using data from the US which has been
adjusted to the 1990 age-structure of the developing world.
Figure VI demonstrates how the mean age of death increases
as the risk of infection declines. The number of deaths in
children declines faster than the annual risk of infection; this
relationship will become important in considering the cost
effectiveness of BCG.
Trends in incidence and mortality
Using the midpoints of the ranges of the annual risk of
infection in Table 1, population projections, and the rates of
decline in the annual risk of infection also reported in Table
11
1, cases and rates of tuberculosis in 2015 have been estimated
(Table 8). These estimates are based on the assumption that
the rates of decline in the annual risk of infection observed
between 1970 and 1985 will continue into the future. In other
words, the projections are based on the assumption that the
socio-economic changes and tuberculosis control activities
that caused the decline in the risk of infection in the last two
decades will continue at the same rate. Such projections
suffer from all the same limitations that any projection of
current trends does.
According to these assumptions, tuberculosis will remain
a major problem in all developing world regions referred
to in Table 8. In Africa, population growth will probably
exceed the projected decline in the annual risk of infection,
so that the absolute number of cases will increase. These
projections for Africa have not taken into consideration
*he interactions between HIV infection and tuberculosis.
. is discussed below, the annual risk of infection in Central
and East Africa may stop declining or even increase in the
next decades. The figures in Table 8, therefore, may be
significant underestimates for Africa. The relationship be
tween HIV infection and tuberculosis will be explored more
fully below.
Social and economic costs
There are few if any studies of the actual costs or conse
quences of tuberculosis on the family, community or econ
omy in developing countries. The special burden of ill-health
and death caused by tuberculosis, however, follows from the
age-distribution of its incidence. While morbidity and mor
tality in any age-groups has significant social and economic
costs, deaths in prime aged adults who are the parents,
community leaders and producers in most societies have a
particularly onerous burden. The incidence of tuberculosis is
concentrated in adults 15-64. For example, while the overall
incidence of tuberculosis in Africa is estimated to be 230, in
adults it is approximately 360 per 100,000.
One of the greatest costs to society and the economy from
.uberculosis is mortality. It has been estimated that there are
just under 10.6 million deaths in adults 15-59 in the develop
ing world (Murray and Feachem, 1990). Of these, our figures
suggest approximately 18.5 % are due to tuberculosis. Not all
these deaths are preventable. Of avoidable adult deaths,
26 % are probably due to tuberculosis.
The consequences of adult death from tuberculosis on
children and other dependents can also be great Studies have
shown that when a mother dies her children suffer higher
rates of mortality (Greenwood et al., 1987). One can specu
late that similar relationships may exist for paternal death.
Several studies from developed countries have shown that
tuberculosis is concentrated in lower socio-economic groups,
those households least able to cope with the burden of
tuberculosis. Pryer (1989) found that children in households
where one parent suffers from a serious debilitating disease
such as tuberculosis, are two and half times more likely to be
severely malnourished. As tuberculosis deaths are concen
trated in the segment of the population that is economi
cally most productive, the economic cost of tuberculosis,
in terms of lost production, must be greater than a disease
that affects exclusively children or the elderly.
Prevention
There are three major strategies for preventing tuberculo
sis : BCG vaccination, chemoprophylaxis and decreasing
sources of infection through case-finding-treatment. Each
will be discussed in turn.
BCG
The bacillus of Calmette and Gu6rin (BCG) was devel
oped in 1921. Since that time, it has become one of the most
widely used yet controversial vaccines. While BCG cover
age has been up to now on average quite high compared to
other immunizations, the effectiveness of BCG in prevent
ing tuberculosis in adults remains controversial. Clinical
trials in the United Kingdom and in the USA found that BCG
was up to 80 % effective (Aronson et al., 1958 ; Medical Re
search Council of Great Britain, 1972). Major vaccine trials
in South India, however, found no effectiveness of BCG (Tu
berculosis Prevention Trial, 1979). A variety of prospective
trials in the developed world and more recent case-control
studies in developing countries have reported effectiveness
ranging from Oto 80% (Smith, 1987 ; Clemens etal., 1983).
Many explanations and theories have been advanced to
explain this variance including differences in strains of BCG,
infections with other mycobacteria and differences in sus-
Table 8. Cases and deaths from all forms of tuberculosis, 2015
Area
Smear +
Estimated number of cases
Other
Total
Deaths
Sub-Saharan Africa
North Africa and
Western Asia
Asia
South America
Central America and
Caribbean
766,000
934,000
1,701,000
777,000
98,000
1,871,000
98,000
120,000
2,283,000
120,000
218,000
4,154,000
218,000
66,000
1,391,000
77,000
80,000
97,000
177,000
84,000
Total
2,913,000
3,554,000
6,468,000
2,395,000
Note: These projections are based on the following assumptions : 1) the current rate of decline in the annual risk of infection will continue
over the next 25 years ; 2) the percent of cases detected will remain the same in each region; and 3) the cure rate and implicitly the percent
of cases treated with standard chemotherapy will remain the same for those cases that arc detected and treated.
12
ceptibility due to factors such as nutritional status (Fine,
1989). While there is no consensus on the effectiveness of
BCG, we will assume that BCG is between 40 % and 70 %
effective in preventing tuberculosis in children 0-14 when
given at birth. Some would argue that BCG given at birth
may protect beyond 15 years; there is, however, no evidence
of this especially in developing countries.
BCG is given as early as possible in life, preferably at
birth, in the vast majority of developing'countries. In addi
tion, one could give serious consideration to “indiscriminate
(re)vaccination” (i.e. without prior tuberculin testing) at
older ages, irrespective of vaccination at birth. Depending on
the feasibility of coverage, BCG (re)vaccination could be
given to children entering school, leaving school, pregnant
women attending for prenatal care of other routine contacts
of the population with health workers. For example, tetanus
toxoid is now considered by many to be an integral compo
nent of prenatal care ; BCG could be delivered at the same
time for only a small increase in the total cost. The actual
effect of BCG (re)vaccination at older ages has not bden
thoroughly studied but there seems little reason that it would
be harmful and it may have some beneficial impact
We must realize, however, that vaccination of newboms
with BCG is a problem in those developing countries where
there is a high prevalence of HIV infection among mothers.
The WHO Expanded Programme on Immunization, which is
responsible for the programme of vaccination against six
selected childhood diseases in the world, has been continuing
BCG-vaccination of newboms and small children including
when the mother is known to be or suspected of being HIVinfected. As of the time of writing, evidence remains incon
clusive regarding the rate of adverse reactions after BCG
immunization among symptomatic HIV-infected individu
als. BCG should be withheld from individuals with sympto
matic HIV infection (WHO, 1987). The current recommen
dations on HIV and BCG-vaccination will be reviewed in the
fall of 1990.
The impact of mass BCG-vaccination on the epidemiol
ogical situation of tuberculosis was overestimated until the
mid- 1970’s (Styblo and Meijer, 1976). As mentioned earlier,
tuberculosis is largely transmitted by sputum smear-positive
cases of pulmonary tuberculosis. From the age-distribution
of smear-positive cases, it is clear that even complete BCG
coverage can have little effect on the annual risk of infection.
Total coverage with BCG, however, will have a major
impact on tuberculosis mortality in children, if BCG is 40 %
to 70 % effective as we have assumed. Based on the assump
tions discussed above, complete coverage could reduce total
tuberculosis mortality by approximately 6 %. BCG will most
likely have very limited effect on the remaining 90 plus
percent of tuberculosis mortality. Evidently, the expansion
of BCG coverage alone cannot or should not be the sole
means employed to control tuberculosis in any community.
Cost-effectiveness of BCG
For two principal reasons, generalizable estimates of the
cost-effectiveness of BCG cannot be made. First, there may
be substantial differences in the computed average and
marginal costs of BCG programmes depending on the pro
gramme considered. Second, the cost-effectiveness of BCG
is inversely proportional to the annual risk of infection.
When more than one vaccine is given at the same time.
average costs for delivering each particular immunization
are often calculated by dividing the cost per client contact by
the number of vaccinations received. Thus the difference
between marginal costs and average costs for a BCG pro
gramme will depend on whether BCG is delivered in an
independent campaign or contact with mother and child or
along with other immunizations such as the first DPT. The
Expanded Programme on Immunization was, unfortunately,
unable to indicate how BCG is delivered in each country. We
conclude that the marginal cost-effectiveness of expanding
BCG will necessarily depend on the location and timing of
vaccination in a particular country.
As the annual risk of infection declines, ceterus paribus,
the cost of vaccinating all newboms does not change. The
benefits of BCG vaccination in terms of cases or deaths
averted, however, will decline inversely to the risk of infec
tion. For example, as the risk of infection declines from 2 %
to 1 %, the cost per death averted will more than double. The
increase in the cost per death averted is greater than the
decline in the risk of infection because the age-distribution of
deaths also shifts away from children as the risk of infection
declines - see Figure VI. The expected relationship between
the risk of infection and the cost per death averted by BCG is
illustrated in Figure VII.
3600
Coat Par Death Averted
3000
2500
2000
1600
1000
600
o
o
0.6
1.5
1
2
2.6
3
3.6
Risk of Infection
| — BCG
I---------------
Case Treatment
Figure VII. A hypothetical comparison of the cost-effectiveness
of BCG immunization and case-treatment.
Only one study has attempted to cost a BCG programme
and estimate its effect in a developing country. Barnum et al.
(1980) estimated the cost of operating a BCG programme
alone and also the marginal cost of adding a BCG pro
gramme to an existing DPT programme. His estimates of
deaths averted were based on local incidence and case
fatality rates of tuberculosis and an assumed effectiveness for
BCG of 50 %. We have recalculated using his original data
the cost per discounted death averted in 1986 dollars. Deaths
prevented by BCG vaccination now occur over the next 14
years, these arc discounted to present value for comparison
with interventions that avert deaths in the current time
period. The cost per death discounted at 3 % was $ 644 for the
BCG programme alone and $ 144 for the marginal BCG
13
programme - both prices are in 1986 US dollars. At the time
in Indonesia survey data suggest the risk of infection was
approximately 3 % ; regional surveys report annual risks of
infection between 2 % and 4 % (Cauthen et al., 1987). It must
be stressed that these estimates of cost-effectiveness do not
take into consideration the potential benefits of BCG in
reducing leprosy (Fine et al., 1986).
Chemoprophylaxis
Clinical tuberculosis can be secondarily prevented by
treating patients with tuberculous infection. Chemoprophy
laxis is applied either to freshly infected so-called tuberculin
converters or to those who have been infected with virulent
tubercle bacilli in the more distant past The latter either do
or do not have abnormalities in the lungs on X-ray.
Tuberculin converters undoubtedly represent a very re
warding group in terms of chemoprophylaxis results and thus
chemoprophylaxis policy has been adopted as a routine
procedure in a number of low prevalence countries. How
ever, mass chemoproplylaxis of converters is impossible,
since their identification depends on repeated tuberculin
tests of the population. On the other hand, a selective search
for converters in high risk groups, such as close family
contacts of smear-positive sources, is a feasible alternative.
As discussed below 6-8 % of recent infections evolve into
clinical tuberculosis. In developing countries, where large
percentages of the population have been infected, the IUATLD
recommends chemoprophylaxis (in HIV low prevalence
countries) only for all non-BCG-vaccinated children aged 5
years or under, with no symptoms suspicious of tuberculosis.
In children with symptoms chemotherapy should, of course,
be given.
Chemoprophylaxis in tuberculin-positive subjects but who
have not developed clinical tuberculosis would reduce the
number of sources of infection, if given for 6 to 12 months.
In most developing countries, this group is very large and
resources would be far better directed to case-detection and
treatment However, chemoprophylaxis might play a very
important role both in developed and developing countries in
subjects with the dual HIV and tuberculous infections with
out clinical and bacteriological signs of tuberculosis. Re
search in this field is urgently needed.
Studies in developed countries have found cost-effective
ness rates per case averted on a 24 week regimen to be greater
than S 17,000 (Snider et al., 1986).
Without accurate data to review the cost-effectiveness of
chemoprophylaxis in developing countries, we can only
make some comparisons to the costs per case treated. Since
only 6-8 % of recent converters evolve into clinical tubercu
losis, 12.5-16.7 recent tuberculin positive patients must be
given chemoprophylaxis to prevent one case of tuberculosis
assuming prophylaxis is 100 % effective. In tuberculin
positive subjects as opposed to new converters, the ratio
would be one or two orders of magnitude higher. The drug
costs for chemoprophylaxis are lower than for treatment, but
the costs of administration, screening, transport, delivery and
monitoring would be similar. Thus, chemoprophylaxis is
unlikely to be more cost-effective in developing countries
than case-finding/trcatment of patients presenting with
symptoms suspicious of tuberculosis as discussed below.
One exception may be in children under 5 exposed to an adult
with active smear-positive pulmonary tuberculosis.
14
Decreasing sources of infection
It has already been suggested that the best way to reduce
transmission of tuberculous infection and thus the number of
tuberculosis cases is to cure patients with smear-positive
tuberculosis. This was stated by Crofton (1962) already in the
mid 1950’s. A variety of epidemiological studies can be used
to quantify this transmission effect. The number of new in
fections caused each year by a case of smear-positive tuber
culosis can be estimated from survey data on the number of
new infections and the prevalence of smear-positive tubercu
losis. It has been estimated using data from developing and
developed countries, that an undiagnosed and untreated
smear-positive source of tuberculous infection would infect
on average between 10 and 14 persons per year (Styblo,
1984 ; Sutherland and Payers, 1975). Breakdown of primary
infection with tubercle bacilli to clinical tuberculosis is the
next link in the chain of transmission. Reference is made to
three reports of newly infected subjects to determine the per
centage that developed clinical tuberculosis: the MRC study
(Sutherland, 1968) found 8.1 % of converters developed
clinical tuberculosis within 15 years ; in Saskatchewan of
recently infected individuals, 6.4 % developed clinical tuber
culosis within a few years after primary infection (Barnet et
al., 1971); and aTSRU study of European data found 6.0 %
of converters developed bacillary tuberculosis in five years
(Sutherland, 1976). For the purposes of modelling transmis
sion, we will assume that from 6 % to 8 % of new infections
will eventually develop some form of clinical tuberculosis.
The studies cited above refer to the risk of developing
clinical tuberculosis soon after primary infection. What about
the risk of developing clinical tuberculosis in cases previ
ously infected with tubercle bacilli, without or with a fresh
reinfection ? As it is not possible to detect reinfection with
tubercle bacilli by tuberculin testing, it cannot be discovered
directly whether or not exogenous reinfection is important in
the development of tuberculosis in an adult. It is evident that
in countries with low annual risks of infection, tuberculosis
in elderly and old persons is predominantly due to en
dogenous exacerbation among those remotely infected with
tubercle bacilli. In developing countries, exogenous reinfec
tion seems to play an important role in developing active
tuberculosis in the adult population, since 0.5 %to 2.5 % or
more of previously infected individuals are annually rein
fected with tubercle bacilli as was the case in developed
countries some 2 to 4 decades ago (Canetti, 1972; Jancik and
Styblo, 1976). Strong evidence forthe latteristhe rapid decline
in tuberculosis incidence in Eskimos over the space of 20
years, not only in children and young adults but also in elderly
and old people, when aggressive case detection and adequate
chemotherapy was introduced (Grzybowski et al., 1976).
We will assume, therefore, that each undiagnosed and
untreated smear-positive case will cause 10 to 14 infections
per year of infectivity. These 10 to 14 infections will subse
quently cause over the next few years 0.6 to 1.2 cases of
tuberculosis. These cases will be approximately equally dis
tributed between smear-positive and other tuberculosis. The
mean period from infection to onset can be estimated more
precisely from data reported by Sutherland (1968) as 1.4
years. These parameters highlight the importance of case
treatment in preventing further cases and will be used in
estimating the cost-cffcciivcncss of treating smear-positive
tuberculosis in the pages that follow.
Curative care
The subject of curative care can naturally be divided into
tuberculosis detection and chemotherapy. Each of these will
be addressed in turn, highlighting the policy options.
Case detection
There are two major issues in detecting cases of clinically
significant tuberculosis : active versus passive detection
strategies and the choice of diagnostic technology. Active
detection means attempts to screen the population at large, or
target populations such as military recruits, for evidence of
tuberculosis. Passive case detection means screening and
diagnosing only those patiepts who present to a health
service provider because of symptoms suspicious of tubercu
losis. In the 1950s and 1960’s, the choice between active and
passive detection in developed and developing countries was
a controversial topic (Styblo et al., 1967 ; Meijer et al.,
1971 ; WHO Expert Committee on Tuberculosis, 1974 ;
Styblo and Meijer, 1980 ; Toman, 1979). In the last two
decades, a consensus for passive case detection of tuberculo
sis in all countries has developed, and both the WHO and
IUATLD advocate this policy.
There are three assumptions that underlie the wide accep
tance of passive case detection as the primary strategy in
tuberculosis control. First, 90 % of patients with smear
positive pulmonary tuberculosis have objective symptoms,
such as cough (with or without sputum), fever, loss of weight
or hemoptysis. These symptoms develop quite soon after the
onset of the disease prompting the patient to seek medical
advice. Second, the great majority of sputum smear-positive
tuberculosis cases develop in a shorter period of time than the
shortest feasible interval between two mass radiography
survey rounds. That is why smear-positive tuberculosis cases
were detected outside (usually earlier than) the periodic case
finding campaigns by the regular health services that the
patient can consult whenever he feels ill. Third, appropriate
diagnostic and curative care ought to be physically, socially
and economically available. Most infections, before chemo
therapy is instituted, would therefore occur within the fam
ily. Whereas in developed countries, it is estimated that 2-3
persons would be infected by a smear-positive case before its
detection, this number may be 4-5 in developing countries,
because of higher number of close contacts. No contacts will
be infected after the start and completion of adequate chemo
therapy.
The validity of these assumptions depends on local condi
tions, cultural perception of disease, access to care and the
efficacy of health services.
Regardless of the technology used, active case detection is
more expensive per case detected because the yield of
tuberculosis per patient screened is lower. For example, if the
incidence of smear-positive tuberculosis is 100/100,000 then
more than 1,000 people’s sputum would have to be screened
to detect one case of smear-positive tuberculosis provided
that the general population is screened. If specific high risk
groups can be identified, the yield would clearly be higher.
For comparison, screening patients who present with cough
in Tanzania using sputum microscopy yields one patient in
10 cases with smear-positive tuberculosis. The second argu
ment against active case detection is that cases actively
identified may be less likely to comply with long drug
regimens. Clearly, they did not yet consider their health to be
impaired enough to seek treatment to start with. Moreover, a
proportion of smear-negative cases with few or no clinical
symptoms cure spontaneously and in a number of cases the
disease is in regression (National Tuberculosis Institute in
Bangalore 1974 ; Meijer etal., 1971, Styblo et al., 1967). In
developing countries, active case-finding was studied by the
Kenyan and British Research Councils in the late 1970’s and
early 1980’s. There are seven reports on these studies and the
conclusion in the last study is that a patient suffering from
symptoms suggestive of pulmonary tuberculosis nearly always
attends, usually several times, a health unit seeking medical
advice (EAMRC, 1987). However, health workers at the
peripheral level do not think in many instances of tuberculo
sis and do not send the sputum or do not refer the patient to
the nearest microscopy centre for sputum examination for
tubercle bacilli. In many developing countries, public trans
port is very rudimentary; even if available, it is not always
affordable to poor people. Moreover, the Kenyan studies
have shown that active case-finding, except in health units, is
not feasible.
The second issue in case-detection is the choice of tech
nology. At present, the major options are sputum micros
copy, sputum culture and radiology. To illustrate the yield
and likely cost of case-detection using microscopy (ZiehlNeelsen), we shall examine data from the National Tubercu
losis and Leprosy Programme in Tanzania. In that country,
one in ten tuberculosis suspects screened by smear examina
tion is identified as a smear-positive case. Normally, three
smears are conducted on each patient The costs of supplies
and reagents alone for these thirty smears is $ 4.05. A
microscopist can examine about 20 sputa per day and is paid
USS 45 per month. The effective cost per case detected in
Tanzania is USS 7.30. This is a high estimate since each case
has three sputa examined to increase sensitivity ; the in
creased sensitivity achieved with the third smear is in fact
small and could be sacrificed to reduce the cost
Sputum culture is used to diagnose pulmonary tuberculo
sis in those patients that produced too few bacilli to be
detected on a smear, to confirm sputum microscopy, and to
characterize the type of mycobacterium. (Finally, culture is
a prerequisite to sensitivity test examination.) As culture
takes several weeks to yield results, it is not useful as a
primary diagnostic tool in developing countries.
The third diagnostic tool is radiography. While it can be an
effective tool particularly for diagnosing smear-negative
pulmonary tuberculosis, the capital cost of an X-ray machine
limits its use to those facilities with a high case load. In
Tanzania, we can estimate that the cost per case detected for
both smear-positive and smear-negative pulmonary tubercu
losis combined to be around USS 10. This calculation,
however, attributes only one sixth of the depreciated capital
cost of an X-ray machine to diagnosing tuberculosis, because
we assume an X-ray machine in a district hospital would be
used for many other purposes. An X-ray facility exclusively
for diagnosing tuberculosis would be much more expensive.
The calculation also assumes a case-load of at least 1,000
patients per year. The optimal use of radiography in passive
case-detection clearly depends on the health system infra
structure, population distribution, and possibilities for refer
ral. Usually, in 25-30 % of all pulmonary cases detected in a
developing country, the diagnosis of smear-negative tuber15
culosis is based on X-ray of the chest with a pathology
suspicious of active tuberculosis and on clinical examina
tion.
New diagnostic technologies based on the enzyme linked
immunoabsorbant assay or DNA probes for mycobacterial
DNA or RNA are currently being investigated (Daniel,
1989 ; Bloom, 1989). If these approaches yield new tools
that can be cheaply applied in developing countries, passive
case detection may be improved especially for smear-nega
tive and extra-pulmonary tuberculosis which are not diag
nosed using sputum microscopy. Active case detection in
some high risk groups would perhaps become feasible.
A limited number of interventions are available to im
prove the effectiveness of passive case detection. The most
effective factor for improvement in case-finding is a high
cure rate of diagnosed cases and a friendly relationship
between the treating health staff and the patient. Public
education can increase general awareness of the symptoms of
tuberculosis and encourage suspects to seek medical advice
resulting in diagnosis of tuberculosis and its treatment
Improved diagnostic skills of primary health care providers,
transport of sputum or a patient to a microscopy center, and
X-ray facilities, if available, can also improve the detection
of both smear-positive tuberculosis and other tuberculosis.
Finally, if diagnosis and adequate treatment are free, as
recommended by WHO and IUATLD, more patients will
seek care earlier.
Table 9. Examples of tuberculosis chemotherapy regimens used
in developing countries
Regimen.
New smear-positive cases
Long-course (12 months)
2SH/10JH
2SH/10EH
2SH/10S2H2
Short-course
2SHRZ/6TH
2SHRZ/4HR or 2EHRZ/4HR
2HRZ/4HR
2HRZ/4H3R,
New smear-negative cases
2STH/10TH
2SHRZ/6TH
Retreatment cases
2SHRZE/1HRZE/5HJRJE,
2SHRZE/1HRZE/5TH
Duration
12
12
12
8
6
6
6
12
8
8
8
S : streptomycin ; H : isoniazid ; R : rifampicin ;
Z : pyrazinamide ; E : ethambutol ; T : thiacetazone.
Subscripts refer to intermittent therapy where drugs are given
a limited number of times each week.
effective antituberculosis drugs, tuberculosis treatment pro
grammes in most developing countries have not been very
successful. Overall cure rates for most national programmes
in poor developing countries are below 50 %. Evidently, the
“standard” 12-month chemotherapy (isoniazid, streptomy
cin and thiacetazone) recommended by the WHO Expert
Committee on Tuberculosis (1974) for use in developing
countries is presendy, and probably will be in the future,
beyond the organizational resources of many of them.
While there are many interesting issues in tuberculosis
treatment, this discussion will stress the choice between
standard 12-month chemotherapy regimens that use fewer
and cheaper drugs (isoniazid, streptomycin and thiaceta
zone), and short-course chemotherapy that lasts from 6 to 8
months and uses multiple and more expensive drugs (ri
fampicin and pyrazinamide). To compare these two ap
proaches to chemotherapy, we must examine the relative
effectiveness of each and the relative costs of each. Because
of the great diversity in effectiveness and costs between
countries, the emphasis will be on the key determinants of the
effectiveness and costs of the two regimens. It should be
stressed that the regimen with a higher cure rate leads to a
more rapid reduction in the risk of tuberculosis infection and
the incidence of active tuberculosis.
EfTectiveness of chemotherapy
The effectiveness of standard and short-course chemo
therapy depends on three major factors: (i) the cure rate; (ii)
acquired drug resistance ; and (iii) the impact on the trend of
the risk of tuberculous infection. Without question, the most
important of these factors today in nearly all contexts is the
cure rate which decisively influences the remaining two
factors.
The first determinant of the cure rate is the biological
effectiveness of standard 12-month and short-course chemo
therapy given under ideal conditions of 100 % compliance.
With short-course chemotherapy after 2 months of treatment
85-95 % of smear-positive pulmonary cases will have con
verted to sputum negative status. Under standard 12-month
therapy after 2 months 50 % will remain smear-positive. The
“permanent” cure rale is a more important aspect of the
treatment regimens. The schematic (Figure VIII) shows the
Percent Failures
10 0%-fc-------------80% -
80% -
40%
20%
o%L
0
Treatment
The six drugs recommended by WHO and the IUATLD
and most commonly used in developing countries for tuber
culosis are isoniazid, streptomycin, thiacetazone, cthambutol, rifampicin and pyrazinamide. These drugs are used in a
host of combinations for different durations. See Table 9.
However, despite the availability of powerful and potentially.
4
2
6
8
10
12
14
Months ol Treatment Received
Short Course
-+— Standard
Hypothetical Value*
Figure VIII. The percent ofpatients failing therapy after 2 years
offollow-up as afunction ofthe number of months ofchemother
apy completed hypothetical values.
16
rJ ?
1U 0 U
percent of patients who will remain or become smear
positive, say, 2 years after the start of the (first) treatment
(with no retreatment during the first 2 years) provided that
chemotherapy is discontinued at each point in time. We shall
refer to them as “failures”. (If a patient remained or became
smear-positive and died during the first 2 years, he/she will
also be referred to as a “failure”) (the upper part of Figure
VIII above the interrupted lines). Under short-course chemo
therapy (e.g. 2SHRZ/6TH), about 40 % may be “failures” at
2 months if they discontinue chemotherapy at that time com
pared with approximately 8-10 % if they complete 6 months
of short-course chemotherapy. Under standard 12-month
chemotherapy (e.g. 2STH/10TH), the “failure” rate in pa
tients who discontinue after 2 months may reach 65-70 %,
and in those who complete 6 months it might be approxi
mately 50 %. The “failure” rate only begins to drop signifi
cantly on standard 12-month chemotherapy after 6 months. If
treatment is stopped at 12 months, under ideal conditions of
100 % compliance approximately 10-15 % including deaths
will become failures at two years.
Since standard 12-month and short-course chemotherapy
both give high cure rates and do not lead to secondary resis
tance in controlled clinical trials compliance is the most
important determinant of the cure rate in national tuberculo
sis programmes. There is a vast and detailed literature on
compliance in general and on tuberculosis in particular
(Haynes etal., 1979; Fox, 1983,1985 ;Chaulet, 1987; WHO
Tuberculosis Chemotherapy Centre, 1963 ; Reichmann, 1987).
Many of the factors that one might expect would influence
patients compliance with antituberculosis drugs regimens,
such as the severity of side effects, have not been empirically
observed. There is a clear consensus, however, that the
duration of treatment adversely affects compliance (Haynes,
1979). Moodie (1967) in unusual circumstances in Hong
Kong found that most non-compliers dropped out in the first
three weeks ; but all other studies have observed a steady
drop out over time (EAMRC 1977, 1979). Improved net
compliance due to a shorter regimen is a major advantage of
short-course chemotherapy over standard chemotherapy.
Given the relapse rate as a function of months of treatment
discussed above, in a situation where patients continue to
drop out over time, short-course chemotherapy will have a
higher total cure rate.
The second major determinant of tuberculosis chemother
apy compliance is the degree of supervision of treatment. A
spectrum exists from giving supplies of drugs for multiple
months to patients all the way to hospitalization for the entire
duration of treatment. Between these extremes, a wide vari
ety of supervision strategies are possible, including daily
visits to health centres, health visitors contacting patients in
the home, periodic urine tests to monitor compliance and
hospitalizations for the first 2 months of treatment. While
increased supervision increases compliance in most settings
(Haynes, 1979); increased supervision also means increased
cost. The balance of this trade-off will depend on the specific
institutional and cultural characteristics of each community.
For example, in Madras, in areas where most of the popula
tion has ready access to health centres, entirely ambulatory
care has been successful (Tuberculosis Chemotherapy Centre,
Madras 1959, Dawson et al., 1966). On the other hand, in
many parts of rural sub-Saharan Africa, the only way to
guarantee daily supervision of chemotherapy may be to
hospitalize patients for the first 2 months of chemotherapy;
this has been the experience in 7 African countries (Tanza
nia, Kenya, Mozambique, Malawi, Benin, Senegal and Mali)
(Styblo and Chum, 1987).
The rationale for hospitalizing patients to ensure close
supervision of the initial intensive phase is much greater in
short-course chemotherapy than in standard 12-month che
motherapy because 2 months of short-course chemotherapy
will convert smear-positive sputum into smear-negative in
about 90 % of patients and in the remaining 10 % in a further
2-4 weeks. Even if they stop taking drugs one or 2 months
after they leave hospital many will not relapse. In Tanzania,
approximately 50 % of smear-positive patients enrolled on
standard 12-month chemotherapy remain smear and culture
positive at 2 months. For standard 12-month chemotherapy,
it is crucial to continue to take regularly daily isoniazid and
thiacetazone combined tablets for at least another 2-3 months
to achieve 90 % sputum conversion.
In all probability, the patient’s perception of the effective
ness of treatment and the balance between discounted future
costs and benefits of treatment are also important determi
nants of compliance. In Tanzania and other IUATLD-assisted National Tuberculosis Programmes, it has been ob
served that both the perceived effectiveness of treatment and
individual and group education of patients during the initial
intensive phase of short-course chemotherapy positively
affected compliance during the continuation phase.
Other possible determinants of compliance include the
number of medications taken at each time, the number of
doses per week and the costs to the patients of therapy.
Combination tablets of isoniazid and thiacetazone and isoniazid
and rifampicin have been in use in National Tuberculosis
Programmes of many developing countries for several years.
On the other hand, intermittent standard chemotherapy (strep
tomycin and isoniazid) has never been used on a large scale
in developing countries. In India, it has been shown thatintermittancy leads to increased irregularity (Pamra and
Mathur, 1973). Also Blackwell (1979) could not validate the
expected relationship between reduced number of doses and
improved compliance. The advantages and disadvantages of
intermittent standard chemotherapy will not be addressed
further here. The common sense notion that increasing costs
both in terms of time and money will decrease compliance
has been confirmed in most studies (Haynes, 1979). To
maximize compliance, tuberculosis chemotherapy should be
free and the spatial and temporal ease of access to treatment
should be improved. When alternative treatments are avail
able in the private and public sector, patients may initially
prefer to pay for therapy perceived as belter, but when funds
run out they may switch to the public sector (Uplekar,
personal communication). This mixing of different drug
regimens will tend to increase the failure rate and the
probability of secondary resistance.
The second factor determining the effectiveness is the
development of resistance. Under ideal conditions, such as in
many clinical trials in patients with sensitive bacilli, the cure
rates for both standard and short-course chemotherapy are
over 95 %. In patients infected with tubercle bacilli that are
isoniazid resistant, the cure rate with total compliance is
greatly reduced (Shimao, 1987). Isoniazid resistance is al
ready a major problem in many developing countries (KIcoberg
et al., 1980). A systematic application of short-course che17
motherapy referred to above (2SHRZ/6TH) in new smear
positive cases makes it virtually impossible to select for a
bacillus resistant to all four drugs, provided that the 2-month
initial intensive phase is closely supervised. Decreased de
velopment of resistance means that short-course chemother
apy is a substantially more effective long-term strategy for
tuberculosis control than standard 12-month chemotherapy.
It has to be stressed that acquired (and in contacts of the index
cases, primary) resistance to both isoniazid and rifampicin
results in incurability of the majority of such cases in devel
oping countries with serious consequences for elimination of
tuberculosis.
Costs of chemotherapy
The costs of any tuberculosis control programme com
prise many components including drugs, staff costs, trans
port, training and the cost of hospitalization. While drugs
form a considerable portion of the budget, probably from
20 % to 40 %, they are not the only cost. Cost differences
between short-course and standard 12-month chemotherapy,
however, center on drug costs and hospitalization costs.
Table 10 shows prices for the five major drugs from three
sources : the UNICEF (UNIPAC), and bulk purchase rates
given to the IUATLD by major drug companies. Using
different suppliers will substantially alter the relative costs of
standard and short-course chemotherapy. Resolution of this
question may pave the way for cheaper tuberculosis drug
regimens -both standard and short-course. In general, the
short-course regimen used in IUATLD National Tuberculo
sis Programme is approximately S 25 more per patient than
Table 10. Costs of antituberculosis drugs
Drug
Isoniazid
Isoniazid
Isoniazid
Isoniazid/
thiacetazone
Isoniazid/
thiacetazone
Streptomycin
Streptomycin
100 mg x 1000
300 mg x 1000
Unit cost
IUATLD UNIPAC
S2.95
$2.70
$3.80
300 mg/150 mg x 1000 S10.50
S9.ll
100 mg/50 mg x 1000
$4.80
$4.93
1 gm vial
1 gm vial
1 gm x 50
$0.07
$0.09
$3.55
Pyrazinamide
Pyrazinamide
(GIF)
Pyrazinamide
500 mg x 1000
500 mg x 100
$31.58
Ethambutol
Ethambutol
Ethambutol
400 mg x 1000
400 mg x 500
$17.00
Rifampicin
Rifampicin
Rifampicin
Rifampicin/
isoniazid
Rifampicin/
isoniazid
18
Unit
$8.34
150 mg x 1000
300 mg x 1000
150 mg/100 mg x 1000 $79.50
300 mg/150 mg x 1000 $159.00
$17.13
$40.90
$58.00
standard depending on the supplier.
The second potential source of cost differentials between
short-course and standard 12-month chemotherapy is the
level and intensity of supervision. Both standard 12-month
and short-course chemotherapy should be given whenever
possible on an entirely ambulatory basis. However, in rural
areas, where 80-85 % of the African population live and
where patients do not have easy access to a health centre,
daily regimens may only be delivered in district hospitals.
Not only will this improve compliance, but expensive and
valuable drugs can be better accounted for in these condi
tions. For these reasons, short-course chemotherapy may
justify higher hospitalization rates and more expense in some
countries.
This discussion has thus far been implicitly restricted to
the treatment of smear-positive tuberculosis. Once other
forms of tuberculosis have been identified, treatment costs
for other tuberculosis should be similar to standard 12-month
chemotherapy except for serious forms of smear-negative tu
berculosis, such as miliary tuberculosis, tuberculous menin
gitis, Pott’s disease, etc., who should be enrolled on short
course chemotherapy (in cases with tuberculous meningitis,
rimactazid should also be given in the continuation phase).
For treatment of cases that failed to sputum or culture convert
in the first round of treatment, the drug costs are particularly
high because these patients harbour tubercle bacilli fre
quently resistant, in developing countries, to isoniazid and/or
streptomycin. Many of them have to be treated with short
course chemotherapy for retreatment cases which should
ideally contain 3 drugs to which the bacilli are sensitive. A
retreatment regimen includes, as a rule, rifampicin and
pyrazinamide. In the lUATLD-assisted National Tuberculo
sis Programmes the following regimen is used : 2SHRZE/
1HRZE/5H3R?E3 in patients resistant to isoniazid or 2SHRZE/
1HRZE/5TH in patients sensitive to isoniazid. In programmes
that are committed to treating all patients that present for
care, retreatment must also be considered in examining
short-course and standard 12-month chemotherapy. As fail
ure rates are higher for standard 12-month chemotherapy,
more resources would have to be devoted to retreatment of
these patients.
Cost-effectiveness
The cost-effectiveness of tuberculosis treatment will vary
depending on the type of tuberculosis case that is treated. In
general, the cost per death averted directly and indirectly will
be lowest for smear-positive tuberculosis, higher for other
tuberculosis and highest for retreatment cases. While this
statement may run counter to intuitive notions of the clinical
costs of treating each type of tuberculosis, the rationale is
based on the effect of interrupting transmission as explained
more fully below. Treating smear-positive tuberculosis may
be as much as twice as cost-effective as treating other
tuberculosis because of the benefits of reduced transmission
in the former. Whether a country with limited resources
would choose to target treatment in adult patients to exclu
sively smear-positive tuberculosis is a difficult ethical, medical
and economic choice. The calculations presented here per
tain largely to treatment of smear-positive tuberculosis be
cause little information is available on the results of treating
other tuberculosis; this emphasis should not be interpreted as
a recommendation to treat only smear-positive tuberculosis.
Few studies have examined the cost-effectiveness of tu
berculosis treatment in developing countries (Barnum, 1986;
Joesoef et al., 1989 ; Feldstein et al., 1973). Two of these
investigations reported that per case cured short-course
chemotherapy was more cost-effective. They did not, how
ever, report figures on the cost per death averted. To fill the
gap in information on the cost-effectiveness of short-course
and standard 12-month chemotherapy, we have analyzed the
tuberculosis control activities over the last seven years of the
National Tuberculosis and Leprosy Programme (NTLP) of
Tanzania.
Our assumption and calculations are summarized in Tables
Table 11. Assumptions used in estimating the effectiveness of
standard 12-month and short-course chemotherapy in the Na
tional Tuberculosis and Leprosy Programme of Tanzania
Fate of 100 cases smear-positive tuberculosis with
no treatment based on 5 year epidemiological study
in South India (1974)
Year
Number of cases
Cured
Excreting
Dead
bacilli
Cured
per
year
Died
per
year
0
1
1.5
2
3
4
5
0.0
18.5
27.8
27.8
27.8
30.3
32.5
100.0
61.3
42.0
38.6
31.7
24.9
18.0
0.0
18.5
0.0
20.1
9.3
0.0
2.5
2.2
13.6
6.9
4.4
4.4
0.0
20.1
30.2
33.6
40.5
44.9
49.2
Fate of 100 cases smear-positive tuberculosis receiving
standard 12-month chemotherapy based on treatment
results in Tanzania*
Year
Cured
0
1
1.5
2
3
4
5
0.0
61.1
61.1
61.1
61.1
62.4
63.7
Number of cases
Excreting
Dead
bacilli
Cured
per
year
Died
per
year
0.0
15.8
15.8
17.7
21.4
23.8
26.2
0.0
61.1
0.0
15.8
0.0
0.0
1.4
1.2
1.9
3.8
2.4
2.4
100.0
23.2
23.2
21.3
17.5
13.7
9.9
Fate of 100 cases smear-positive tuberculosis receiving
short-course chemotherapy based on treatment results
in Tanzania*
Year
Cured
Number of cases
Excreting
Dead
bacilli
Cured
per
year
Died
per
year
0.0
79.6
1.8
0.0
0.0
0.5
0.4
0.0
8.8
1.9
0.6
1.3
0.8
0.8
Table 12. Budgeted costsfor tuberculosis chemotherapy, Tanza
nia, 1986
Cost
category
STD
Short
$2.10
$1.95
$2.10
$1.95
Retreat
ment
Diagnosis
Slides/reagents
Sputum container
Bacteriological
Monitoring
Culture
Sensitivity
Drugs
Administration
Labor costs and
hospitalization
Transport
Training
Supervision
Capital 20 %
depreciation
$0.81
$1.50
$0.81
$1.50
$17.00
$2.85
$40.00
$2.85
$1.50
$3.90
$65.00
$2.85
$67.65
$20.35
$2.37
$2.90
$90.20
$20.35
$2.37
$2.90
$169.12
$20.35
$2.37
$2.90
$3.22
$3.22
$3.22
Total cost per case
$122.70
$168.25
$271.21
Notes: 1) Labour costs and hospitalization include the salaries
and wages of all staff working on tuberculosis control and the
cost of hospitalizing all tuberculosis patients. As more disag
gregated information was not available, the costs in this
category have been distributed according to the hospitalization
rate for each type of treatment (60 % standard, 80 % short
course, 100 % retreatment). This underestimates the cost per
case treated for standard chemotherapy and overestimates the
cost per case treated for short-course and retreatment because
all staff costs are included in this category.
2) Drug costs are based on 1986 prices and 1986 exchange
rates. Since that date the price has declined but the value of the
US dollar has also declined.
3) Transport costs include 25 % of the entire purchase cost of
all vehicles and operating expenses. The assumption that'
vehicles will last only 4 years on average may be overly
conservative.
Table 13. Cost-effectiveness of standard 12-monih and short
course chemotherapyfor smear positive pulmonary tuberculosis
in Tanzania, 1986
Standard
Short-course
chemotherapy chemotherapy
Cost per case treated
Cost per case cured
at 18 months
Cost per death averted
Cost per death averted including
one round of transmission
$123
$168
$368
$569
$314
$514
$275
$243
Note : Based on a 3 % discount rate.
0
1
1.5
2
3
4
5
0.0
79.6
81.3
81.3
81.3
81.8
82.2
100.0
11.6
7.9
7.3
6.0
4.7
3.4
0.0
8.8
10.7
11.4
12.7
13.5
14.3
♦ Styblo and Chum (1987); Chum et al. (1989).
11 and 12. The NTLP has been assisted by the IUATLD since
1979. Excellent data are available on the results of both
standard 12-month and short-course chemotherapy (Styblo
and Chum, 1987, Chum et al., 1988). Comparing the results
of standard and short-course chemotherapy with the natural
history of tuberculosis as documented in South India (Na
tional Tuberculosis Institute, 1974), we can estimate (Table
13) the net improvement in cure rates and death rates. For
19
I
standard 12-month chemotherapy and short-course, the cost
per patient treated in 1986 US dollar was $ 122 and $ 168
respectively. In a programme where all patients that continue
to excrete bacilli receive retreatment, this would raise the
cost per case treated under standard 12-month chemotherapy
to $ 186 and $ 190 for short-course chemotherapy.
Per case cured at the end of 18 months in excess of
spontaneous cure, it was $ 368 for standard 12-month chemo
therapy and S 314 for short-course. Short-course chemother
apy is more cost-effective per patient cured. The difference
would be greater except that in Tanzania approximately 80 %
of short-course patients are hospitalized for the first 2 months
and only 60 % of standard chemotherapy are hospitalized.
For standard 12-month chemotherapy, the cost per death
averted was $569 and $ 514 for short-course, based on a three
percent discount rate. These estimates are serious overesti
mates for two reasons. First, the benefits of treating patients
relative to no treatment have only been examined for five
years after treatment. In the South Indian epidemiological
study tuberculosis patients were only followed for five years
(National Tuberculosis Institute, Bangalore, 1974). Thus the
improvement in the death rate due to standard 12-month or
short-course chemotherapy has been underestimated by only
examining the effect for five years. Second, treatment also
reduced the number of new infections of tuberculosis. Using
the lowest estimates of the number of cases that each case of
smear-positive tuberculosis causes in one year derived in the
section above, we can calculate the number of deaths averted
by reduced transmission in one cycle. The resulting estimates
show standard 12-month chemotherapy to cost $ 274 per
death averted and for short-course chemotherapy only $ 242
per death averted.’
These estimates of the cost per death averted through
tuberculosis chemotherapy are specific to Tanzania but proba
bly represent the higher end of the range for most other
developing countries for four reasons. First, the hospitaliza
tion rate in Tanzania for short-course chemotherapy is espe
cially high. Any country with a more developed peripheral
health system could deliver short-course chemotherapy with
a lower hospitalization rate. This observation must be tem
pered by the fact that the cost per bed-day in many develop
ing countries is considerably higher than the $ 1.50 reported
by the Tanzania Government. In a country where the cost per
bed-day is S 5, the cost of hospitalizating patients for 2
months would be increased by S 210. Second, the benefits of
chemotherapy over no treatment have only been considered
for the first five years following treatment. Consideration of
years 6,7 and 8 if such data were available would reduce the
cost per death averted. Third, the assumption on transmission
were the lowest reasonable assumptions, not the midpoint of
the expected range of cases transmitted per excretor. Fourth,
the rale of capital depreciation was assumed to be extremely
high 20-25 % per year. In other developing countries, the true
cost per death averted may be considerably lower than
$ 265 ; although in some countries where staff costs and hos
pitalization is more expensive, they may be higher.
Taken together the studies on the cost-effectiveness of
both standard 12-month and short-course chemotherapy
show that tuberculosis chemotherapy is an excellent in
vestment relative to virtually any health intervention.
Most interventions including immunizations and oral rchydration therapy yield estimates per death averted in the same
20
Table 14. Comparison of standard 12-month chemotherapy and
short-course chemotherapyfor smear-positive cases of tubercu
losis based on data from the NTLP of Tanzania
Cost per death averted
including one round of
transmission
Percent of patients cured
Percent of cases requiring
retreatment
Death rate after 5 years
assuming no retreatment
(%)
Standard
chemotherapy
Short-course
chemotherapy
$275
63.7
$243
822
23.2
7.9
26.2
143
range (Haaga, 1982). Second, the analysis clearly indicates
that short-course chemotherapy is preferable to standard 12month chemotherapy. The most significant advantages of
short-course chemotherapy over standard 12-month chemo
therapy are summarized in Table 14. Not only is it more costeffective per death averted, as calculated above, but it
provides other advantages not included in the estimates. The
cure rate taking into consideration the natural history of
tuberculosis for standard 12-month chemotherapy under ex
cellent conditions is 63.7 % while under short-course it is
82.2 %. The percentage of cases requiring expensive retreat
ment is nearly three times greater with standard 12-month
chemotherapy as compared to short-course chemotherapy.
Finally the death rate with standard 12-month chemotherapy
is nearly twice as high as with short-course chemotherapy.
Short-course chemotherapy, because of a higher cure rate,
will also accelerate the decline in the risk of infection by
reducing transmission. Short-course chemotherapy will also
limit the development of resistance to isoniazid and ri
fampicin in the long term. In summary, short-course chemo
therapy for tuberculosis is an excellent cost-effective health
intervention.
All these computations of cost-effectiveness have been
based on the assumption that all patients treated for smear
positive tuberculosis do indeed have smear-positive tubercu
losis. As the cost per case treated is high, false positives are
a particularly onerous burden on the health system. In Tan
zania, where a reference laboratory is available to monitor
diagnoses made in the periphery, the false positive rate is
low. The most important parameter from a cost point of view
is the predictive value positive or the percent of cases
diagnosed with tuberculosis who actually have the disease.
In Tanzania, this is greater than 97 %. The cost per death
averted must be divided by the predictive value positive to
get the true cost-effectiveness in a particular situation. In
countries with poorly trained microscopists or frequent atypi
cal mycobacteria infections, the predictive value positive
could be much lower than 95 %. The potential of wasting
scarce resources on patients without tuberculosis puls a high
premium on training health workers and microscopists to
diagnose tuberculosis correctly.
BCG and case treatment
One would like to compare the two major interventions for
tuberculosis control : BCG and case treatment. They are,
however, not truly comparable because even complete BCG
coverage at birth will only affect 10 % of mortality. Case
treatment is absolutely necessary to reduce the other 90 % of
mortality. How does the cost-effectiveness of expanding
BCG coverage compare to expanding case treatment activi
ties ? The cost per death averted can be compared directly
using the studies mentioned in the text above. Some may
object that a death between the ages of 0 and 14 represents a
greater loss of years of life than a death at age 35. However,
if we choose to examine discounted years of life lost it will
not significantly alter the comparison. A death at age 7, the
midpoint for deaths averted by BCG, represents at a 3 %
discount rate 29.7 years of life lost; while a death at age 34
the average age of a tuberculosis death represents 23.4 years
at a similar discount rate. Therefore, we can examine the
cost-effectiveness of the two interventions using the cost per
death averted bearing in mind that discounted years of life
lost would change the relationship by less than 20 percent
The cost per death averted through tuberculosis chemo
therapy should change little as the risk of infection in a
community declines. Ceterusparibus, the only change would
be the slight increase in the cost of detection as more cases or
cough would have to be screened per case of tuberculosis
detected. This does not hold true for any immunization
including BCG. The costs of vaccinating all infants will not
change as the risk of infection declines, but the benefits in
terms of deaths averted will decline proportionately to the
risk of infection. In other words, the cost per death averted
through BCG must be inversely proportional to the risk of
infection. Figure VII shows two hypothetical curves for the
cost per death averted as a function of the risk of infection.
The curves are fitted to the single data point on BCG for
Indonesia and the single point on case treatment for Tanza
nia. While the data are clearly weak, the principle is clear. At
low annual risks of infection case treatment is substantially
more cost-effective than expanding BCG coverage. At higher
risks of infection, the costs of both interventions are of the
same order of magnitude. This curve should not be inter
preted to mean that countries with low risks of infection
should curtail BCG vaccination activities. The discussion so
far provides no insight into the savings from cutting back an
existing activity versus the potential reduction in benefits.
This discussion does not imply that the policy choice in
tuberculosis control is between BCG and case-treatment
Some combination of the two is likely to be desirable in many
countries. It does, however, indicate that BCG becomes
relatively less attractive as the risk of infection declines.
Research priorities
This discussion of tuberculosis leads naturally to some
general recommendations for tuberculosis research. These
can be divided into six areas :
- Epidemiology. The wide confidence intervals in the esti
mates of incidence, prevalence and mortality highlight the
need for epidemiological research. Many countries require
basic information on incidence and mortality rates and their
distribution by age and socio-economic status, in order to
establish the importance of tuberculosis as a health sector
priority. For those countries without vital registration, new
survey techniques based on the verbal autopsy may provide
the tools with which tuberculosis mortality can be quantified.
- Prevention. Because of the uncertain and variable effec
tiveness of BCG, a new effective vaccine would be a major
tool, especially if it would also prevent tuberculosis in
already infected individuals. Fine (1989), however, has pointed
out that it will be difficult to test appropriately the effective
ness of any new vaccines for moral and technical reasons.
Research is also needed to explore the most appropriate role
for chemoprophylaxis in developing countries, especially in
subjects infected with tubercle bacilli and HIV.
- Diagnosis. Development of new tools for the rapid and
early diagnosis of tuberculosis would substantially improve
case detection. Research into serological or sputum diagno
sis that can be deployed at reasonable cost in peripheral
health facilities in developing countries should be a priority.
- Chemotherapy. Development of new shorter acting, cheap
drugs would help address two major issues in tuberculosis
control: compliance and cost. While opportunities exist for
developing new drugs (Sensi, 1989), relatively little research
is underway. Another possibility that seems worth exploring
is the use of depot preparations which would solve many of
the compliance problems.
- Programme design. There is an urgent need for opera
tional and health economics research on strategies for tuber
culosis control.
Some key issues have been highlighted in this piece: what
is the trade-off between the cost of supervision and the
improvement in compliance, taking existing infrastructure
into consideration ? What is the cost-effectiveness of alterna
tive diagnosis strategies ? These and many other issues need
to be addressed in an organised fashion.
- HIV and tuberculosis interactions. The interaction be
tween HIV and tuberculosis has not been fully addressed in
this piece. It is evident that immune suppressed patients with
HIV and tuberculous infections have a high probability of
developing clinical tuberculosis. In Central East Africa,
tuberculosis programmes are already reporting a consider
able increase in the number of cases of tuberculosis. The
impact of any HIV-tuberculosis interaction in developing
countries with a high prevalence of tuberculous infection on
the annual risk of infection for the rest of the population is not
yet known. Epidemiological study of these relationship has
just begun and should be considered a priority for research.
Major operational conclusions
This review of tuberculosis can be summarized in five
major points :
- The magnitude of the tuberculosis problem is simply stag
gering. Our estimates suggest that more than 25 million
people die from tuberculosis each year. This is probably
more than any other single pathogen ; the only disease that
come close in terms of total deaths is measles, estimated to
kill 2.5 million per year (Walsh, 1988). The burden of
tuberculosis extents beyond mortality ; the annual incidence
of new cases of all forms of tuberculosis is over 7.1 million
cases in the developing world. Tuberculosis is unique amongst
the major killers of the developing world in that it afflicts
nearly all age-groups. Many children die from tuberculous
meningitis and miliary tuberculosis. But the greatest burden
of tuberculosis incidence and mortality is concentrated in
adults 15 to 59. These are the parents, workers and leaders
of society. This heavy toll of the care givers for the rest of
21
society makes tuberculosis a truly unique problem.
- In at least the last decade, tuberculosis has been ignored by
much ofthe international health community. Shimao (1989)
has outlined the decline of the human and institutional
capacity to address the tuberculosis problem over the last
decades which is but one symptom of a general lack of
priority attached to tuberculosis action and research. An
other example is the Institute of Medicine (1986) study of
vaccine development priorities for the developing world.
They classified diseases into three levels of priority for
research on vaccines. While leprosy received significant
attention tuberculosis was not even mentioned in the lowest
priority group. Clearly,focussing international attention on
tuberculosis is necessaryfirst step ifmore resources are to be
directed to combating tuberculosis.
-Existing diagnostic technology and chemotherapeutic agents
can be used effectively in developing countries to cure
tuberculosis. The lUATLD-assisted National Tuberculosis
Programmes (e.g. Tanzania, Malawi) have shown that short
course chemotherapy can be applied on a national scale with
excellent results. Cure rates approaching 90 % even taking
into consideration compliance can be achieved in even the
most difficult circumstances.
-Tuberculosis chemotherapy and BCG vaccination (in
countries with high risks of infection) are some of the most
cost-effective health interventions available in the health
armamentarium. Our analysis of the National Tuberculosis
and Leprosy Programme in Tanzania has shown that treating
smear-positive tuberculosis costs less than $ 250 per death
averted. The cost per discounted year of life saved is there
fore substantially less than $ 10. There are few interventions
that are as cost-effective as tuberculosis case-treatment.
Given our estimates that slightly more than half of all new
cases oftuberculosis receive some form of treatment that in
most cases is not highly effective, we estimate that the total
increased cost of treating all new cases of tuberculosis
through a well managed chemotherapy programme to be less
than 700 million US dollars per year.
- Evidence has accumulated that the HIV!tuberculosis inter
action may significntly exacerbate the epidemiological situ
ation of tuberculosis. The potential rise, due to this interac
tion, in the risk of infection in Africa and other regions
depending on the spread of HTV makes all our operational
conclusions about tuberculosis all the more pressing.
The combination ofan enormous burden,years ofneglect,
the existence of effective interventions, the demonstrated
interaction between tuberculous and HIV infections and one
of the most cost-effective interventions available must make
tuberculosis one of the highest priorities of action and re
search in international health.
REFERENCES
ARONSON J., ARONSON C. & TAYLOR H. A 20-year appraisal
of BCG vaccination in the control of tuberculosis. Arch.Int.Med.,
101 : 881-893 (1958)
BARNET G.D., GRZYBOWSKIS. & STYBLO K. Present risk of
developing active tuberculosis in Saskatchewan according to pre
vious tuberculin and X-ray status. BulLlni.Un.Tuberc., 45 : 51-74
(1971)
BARNUM H.N., TARANTOLA D. & SETIADY I.F. Cost-effec
tiveness of an immunization programme in Indonesia. BulLWHO,
58 (3): 499-503 (1980)
BARNUM H.N. Cost savings from alternative treatments for
tuberculosis. Soc.Sci.Med., 23 (9) : 847-850 (1986)
BLACKWELL B. The drug regimen and treatment compliance. In
Haynes RJ3., Taylor D.W. & Sackett D.L. (eds.). Compliance in
health care. Baltimore : Johns Hopkins University Press (1979)
BLEIKER M.A., CHUM H.J., NKINDA S.J. & STYBLO K.
Tanzania National Tuberculin Survey, 1983-1986. In : XXVIth
IUAT World Conference on Tuberculosis and Respiratory Dis
eases. Tokyo : Professional Post-graduate Services, p. 117-121
(1987)
BLOOM B. An ordinary mortal’s guide to the molecular biology of
tuberculosis. BuUdni.Un.Tubcrc.Respir.Dis., 64 (3): 50-58 (1989)
CANETTI G. Endogenous reactivation and exogenous reinfection.
Their relative importance with regard to development of non
primary tuberculosis. BuU.Int.Un.Tuberc., 47 : 116-122 (1972)
CAUTHEN G.M., PIO A. & TEN DAM H.G. Annual risk of
tuberculous infection. Geneva : WHO/TB/88.154 (1988)
CDC unpublished data. Atlanta, CDC (1989)
CHAULET P. Compliance with antituberculosis chemotherapy in
developing countries. Tubercle, 68 : 19-24 (1987)
CHUM H.J., STYBLO K. & VAN CLEEF M R.A. Eight-years’
experience of the National Tuberculosis and Leprosy Programme in
Tanzania. In : XXVIth IUAT World Conference on Tuberculosis
and Respiratory Diseases. Tokyo: Professional Postgraduate Serv
ices, p. 111-116(1987)
??
CHUM H J. Ten years of the National Tuberculosis/Leprosy Pro
gramme in Tanzania. BuU.Int.Un.Tuberc.Respir.Dis., 64 (3): 3436(1989)
CLEMENS J.D., CHUONG J.J.H. & FEINSTEIN A.R. The BCG
controversy. A methodological and statistical reappraisal. JAMA,
249 (17): 2362-2369 (1983)
CROFTON J. The contribution of treatment to the prevention of
tuberculosis. BuU.Int.Un.Tuberc., 32 (2) : 643-653 (1962)
DANIEL T.M. Rapid diagnosis of tuberculosis : laboratory tech
niques applicable in developing countries. Rev.InfDis., II (suppl
2): S471-S478 (1989)
DAWSON J.J.Y. ET AL. A 5 year study of patients with pulmonary
tuberculosis in a concurrent comparison of home and sanatorium
treatment for one year with isoniazid plus PAS. BulLWHO, 34 :
533-551 (1966)
DROLET G.J. Present trend of case fatality rates in tuberculosis.
Am.Rev.Tuberc. 37 : 125-151 (1938)
East African and British Medical Research Council. Tuberculosis
in Tanzania : a follow-up of a National Sampling Survey of drug
resistance and other factors. Tubercle, 58 : 55-78 (1977)
East African and British Medical Research Council. Tuberculosis
in Kenya: follow-up of the Second National Sampling Survey and
a comparison with the follow-up data from the First (1964) National
Sampling Survey. Tubercle, 60 : 125-149 (1979)
East African and British Medical Research Council. A study of the
use of maternity and child welfare clinics in case-finding for
pulmonary tuberculosis in Kenya. Tubercle, 68 : 93-103 (1987)
FELDSTEIN M.S., PIOTM.A. & SUNDARESAN T.K. Resource
allocation model for public health planning. A case study of
tuberculosis control. BulLWHO, Supp. : 1-110 (1973)
FINEP.E.M., PONNIGHAUS J.M.. MAINE N., CLARKSON J.A.
& BLISS L. Protective efficacy of BCG against leprosy in Northern
Malawi. Lancet, 1: 499-504 (1986)
FINE P.E.M. The BCG story : lessons from the past and implica
tions for the future. Rev.InfDis., II (suppl. 2): S353-S359 (1989)
3
t***0 ■
J
Pl
un
4:
sis.
Dsis
irug
osis
• and
onal
)fthc
g for
>87)
ource
dy of
N J.A.
irthcrn
nplica(1989)
FOX W. Compliance of patients and physicians : experience and
lessons from tuberculosis. I. Br.MedJ, 287 : 33-35 (1983)
FOX W. Compliance of patients and physicians : experience and
lessons from tuberculosis. II. Br.MedJ., 287 : 101-105 (1983)
FOX W. Short-course chemotherapy for pulmonary tuberculosis
and some problems of its programme application with particular
reference to India. BulLInt.Un.Tuberc^ 60 (1-2): 40-49 (1985)
FROST W.H. How much control of tuberculosis ? AmJ.PubLHlth
27 : 759-766 (1937)
GALTUNG HANSEN O. Tuberculosis mortality and morbidity
and tuberculin sensitivity in Norway. Copenhagen: WHO Euro-84/
15 (1955)
GREENWOOD A., GREENWOOD B.M., BRADLEY A.K.,
WILLIAMS K., SHENTON F., TULLOCH S., BYASS P. &
OLDFIELD F.S.J. A prospective survey of the outcome of preg
nancy in a rural area of the Gambia, West Africa. BulLWHO, 65 :
636-643 (1987)
GRZYBOWSKI S.» STYBLO K. & DORKEN E. Tuberculosis in
Eskimos. Tubercle, 57 (Supplement) : 1-58 (1976)
HAAGA J.G. Cost effectiveness and cost benefit analysis of
immunization programs in developing countries : a review of the
literature. Washington DC : Pharmaceutical Manufacturers Asso
ciation (1982)
HAYNES R.B. Determinants of compliance : the disease and the
mechanics of treatment In Haynes R.B., Taylor D.W., Sackett D.L.
(eds.) Compliance in health care. Baltimore : Johns Hopkins
University Press (1979)
HOLM J. Our enemy, the tubercle bacillus. Int.Tuberc. Digest, S
(1970), IUATLD special publication
Institute of Medicine. New vaccine development establishing
priorities. Volume II. Diseases of importance in developing coun
tries. Washington DC : National Academy Press (1986)
IUATLD. Unpublished documents. Paris (1988)
JANCIK E.H. &. STYBLO K. Die Problematik der postprimaren
mykobacteriellen Superinfektion- Versuch einer epidemiologischklinischer Sicht In: Brecke : Foribildung in Thoraxkrankheiten,
Vol. 7, p. 160-178. Hippokrates Verlag, Stuttgart (1976)
JOESOEF M.R., REMINGTON P.L. & TJIPTOHERIJANTO P.
Epidemiological model and cost-effectiveness analysis of tubercu
losis treatment programmes in Indonesia. IntJ.Epidem., 18 (1) :
174-179(1989)
KLEEBERG H.H. & BOSHOFF M.S. A world atlas of initial drug
resistance. Report to the Scientific Committee on Bacteriology and
Immunology of the IUAT (1980)
LEOWSKI J. Global status of tuberculosis control and its pros
pects. Geneva : WHO (1988) (paper presented at the Regional
Scientific Meeting on Tuberculosis Control organized by the WHO
. Regional Office for the Eastern Mediterranean, Sanaa, Yemen Arab
Republic, 4-7 September 1988. See also document WHO/TB/
88.156 of the WHO's Offset Series)
LINDHART M. The statistics of pulmonary tuberculosis in Den
mark, 1925-1934. A statistical investigation on the occurrence of
pulmonary tuberculosis in the period 1925-1934, worked out on the
basis of the Danish National Health Service file of notified cases
and of deaths. Copenhagen : Ejnar Munkshaard (1939)
Medical Research Council of Great Britain. BCG and vole bacillus
vaccines in the prevention of tuberculosis in adolescence and early
life. BulLWHO, 46 : 371-385 (1972)
MEUER J., BARNETT G.D., KUBIK A. & STYBLO K. Identifi
cation of sources of infection. Bull.Int.Un.Tuberc., 45 : 5-50
(I971)
Ministry of Health and Family Welfare, India. Health atlas of India,
1986. New Delhi, Directorate General of Health services (1986)
MOODIE A.S. Mass ambulatory chemotherapy in the treatment of
tuberculosis in a predominantly urban community. AmJlev.Respir.Dis,
95 : 384-397 (1967)
MURRAY C.J.L. & FEACHEM R.G. Adult mortality in the
developing world. Transactions of the Royal Society of Tropical
Medicine (in press)
National Tuberculosis Institute, B angalore. Tuberculosis in a rural
population of South India : a five-year epidemiological study.
BulLWHO, 51 : 473-488 (1974)
PAMRA S.P. & MATHUR G.P. IndJ.Tub, 20 : 108 (1973)
PRESTON S.H., KEYFITZ N. & SCHOEN R. Causes of death; life
tables for national populations. New York : Seminar Press (1972)
PRYER J. When breadwinners fall ill: preliminary findings from
a case study in Bangladesh. IDS Bull, 20 (2): 49-57 (1989)
REICHMAN L.B. Compliance in developed nations. Tubercle,
68 : 25-29 (1987)
ROUILLON A., PERDRIZET S. & PARROT R. Transmission of
tubercle bacilli: the effects of chemotherapy. Tubercle, SI: 275299 (1976)
RUTLEDGE J.A. & CROUCH J3. The ultimate results in 1,654
cases of tuberculosis treated at the Modem Woodmen of America
Sanatorium. Am.Rev.Tuberc., 2 : 755-763 (1919)
SENSIP. Approaches to the development of new antituberculosis
drugs. Rev.Inf.Dis, 11 (suppl 2) : S471-S478 (1989)
SHIMAO T. Drug resistance in tuberculosis control. Tubercle, 68
(Supplement): 5-15 (1987)
SHIMAO T. Institutional capacity for disease research and con
trol : tuberculosis. In Reich M.,MaruiE. (eds.) International coop
eration for health ; problems, prospects and priorities. Dover, Massachussets : Auburn House Publishing Company (1989)
SMITH P.G. Case-control studies of the efficacy of BCG against
tuberculosis. In: XXVIthlUAT World Conference on Tuberculosis
and Respiratory Diseases ; Singapore, November 1986 ; Profes
sional Postgraduate Services, 4-7 p. 73-79 (1987)
SNIDER D.E., CARAS GJ. & KOPLAN J.P. Preventive therapy
with isoniazid. Cost-effectiveness of different durations of therapy.
JAMA, 255 (12): 1579-1583 (1986)
STYBLO K., DANKOVA D.. DRAPELA J., GALLIOVA J.,
JEZEK Z, KRIVANEK J. ET AL. Epidemiological and clinical
study of tuberculosis in the district of Kolin, Czechoslovakia.
BulLWHO, 37 : 819-874 (1967)
STYBLO K., MEUER J. & SUTHERLAND I. The transmission of
tubercle bacilli, its trend in a human population. Tuberculosis
Surveillance Research Unit, Report No. 1, BuU.Int.Un.Tuberc,
42 : 5-104 (1969)
STYB LO K. & MEUER J. Imp act of BCG vaccination programmes
in children and young adults on the tuberculosis problem. Tubercle,
57 : 17-43 (1976)
STYBLO K. Recent advances in epidemiological research in
tuberculosis. Adv.Tuberc.Res., 20 : 1-63, Karger (1980)
STYBLO K. & ROUILLON A. Estimated global incidence of
smear-positive pulmonary tuberculosis. Unreliability of officially
reported figures on tuberculosis. BulLInt.Un.Tuberc., 56 (3-4) :
118-126(1981)
STYBLO K. Epidemiology of tuberculosis. In : Meissner G. and
other editors. Infektionskrankheiten und ihre Erreger. Mykobakteria
und mykobakteriellen Krankheiten. Vol. 4, Jena, Gustav Fischer
Verlag (1984)
STYBLO K. The relationship between the risk of tuberculosis
infection and the risk of developing infectious tuberculosis.
BulLInt.Un.Tuberc., 60 (3-4) : 117-119 (1985)
STYBLO K. Tuberculosis control and surveillance. In : Flenley
DC. and Petty T.L. Recent Advances in Respiratory Medicine. No.
4, 77-108. Churchill Livingstone, Edinburgh (1986)
STYBLO K. & CHUM H.J. Treatment results of smear-positive
tuberculosis in the Tanzania National Tuberculosis and Leprosy
Programme : standard and short-course chemotherapy. Proceed
ings of the XXVIth World Conference on Tuberculosis and Respi
ratory Diseases, Singapore, 4-7 November 1986. Tokyo : Profes
sional Postgraduate Services, p. 122-126 (1987)
STYBLO K. The relationship between the annual risk of tubercu
lous infection and the incidence of smear-positive pulmonary
tuberculosis. Unpublished manuscript (1988)
23
STYBLO K. Overview and epidemiologic assessment of the cur
rent global tuberculosis situation with an emphasis on control in de
veloping countries. Rev.Inf.Dis., Vol. II, suppl. 2, March-April,
S339-346 (1989)
SUTHERLAND I. The ten-year incidence of clinical tuberculosis
following "conversion” in 2,550 individuals aged 14 to 19 years.
The Hague: TSRU Progress Report (1968)
SUTHERLAND I. & PAYERS P.M. The association of the risk of
tuberculosis infection with age. BulLInt.Vn.Tuberc., 50 : 70-81
(1975)
SUTHERLAND I. Recent studies in the epidemiology of tubercu
losis, based on the risk of being infected with tubercle bacill.
Adv.TubercRes19 : 1-63 (1976).
TOMAN K. Tuberculosis case-finding and chemotherapy. Ques
tions and answers. World Health Organization, Geneva (1979)
TUBERCULOSIS CHEMOTHERAPY CENTRE, Madras. A con
current comparison of home and sanatorium treatment of pul
monary tuberculosis in South India. BulLWHO, 21 : 51-144
(1959)
TUBERCULOSIS PREVENTION TRIAL. Trial of BCG vaccines
24
in south India for tuberculosis prevention: first report. BulLWHO,
57 (5): 819-827 (1979)
TUBERCULOSIS SURVEILLANCE RESEARCH UNIT: TSRU
Progress Report, 1966; The Hague, P.O. Box 146, The Netherlands
UNITED NATIONS. World population prospects. Estimates and
projections as assessed in 1988. New York: United Nations (1989)
UNICEF. State of the World’s Children. Oxford : Oxford Univer
sity Press (1988)
WALSH J. Establishing health priorities in the developing world.
New York : United Nations development Programme (1988)
WHO consultation. Statement from Consultation on human immu
nodeficiency virus (HIV) and routine childhood immunization.
Weekly epidemiological record, 62 : 297-299 (1987)
World Health Organization. WHO Expert Committee on Tubercu
losis, Ninth Report. Tech.Rep.Ser., 552 (1974)
WHO. Reported annual incidence of tuberculosis, 1974-1987.
Geneva : WHO, Mimeo (1988)
WHO Tuberculosis Chemotherapy Centre, Nairobi. Drug accepta
bility in domiciliary tuberculosis control programmes. Bull WHO,
29: 627-639 (1963)
Bulletin o/ rne iniernauonai Lmon Against Iuoercuiosis. ^oi. bO.
3-4. bepiemoer-December 1985
1
■I The relationship between the risk of tuberculous infection
and the risk of developing infectious tuberculosis
J
4
K. STYBLO*
• Director of Scientific ActiritiM. International Union Against Tubercuiosia, 3. r»e Gcactes Ville, 75116 Paris, France.
Introduction
The aim of this report is to study
whether a uniform empirical relationship
can be established between the annual
risk of tuberculous infection in a com
munity and the incidence of infectious
(smear-positive) tuberculosis and, if so,
to estimate this parameter*. Such infor
mation would be especially important for
developing countries because it would
enable them to estimate the incidence of
smear-positive cases of pulmonary tuber
culosis from the annual risk of tubercu
lous infection, without any further
investigation.
Summary Table. Relaaomkiy between the risk of tuberculous infection and the incidence of
smear-positive tuberculosis
Ratio
based on
Risk of
infection (%)
Ratio between the risk
of infection (%) and
mortality incidence prevalence
(rate)1 (rate)1
(rate)1
Netherlands 1921-1938 mortality
Netherlands 1951-1976 incidence
Developing countries
prevalence
1956-1951
Alaska Eskimos
mortality
1948-1951
India long, study
prevalence
1961-1968
India prevention trial
1969-1971
prevalence
2. 7- 6.0
0.038-0.4
19
40-60
2.0 -8.0
25
38
37
26
8O-12O2
52
1.5
53
106
4.1
51
102
Methodology
The relationship between the annual risk
of tuberculous infection can be studied in com
munities where reliable information on both
these variables is available.
We now have information on the annual
risk of tuberculous infection in a number of
developed and developing countries. Unfortu
nately, the incidence of smear-positive tuber
culosis has only been reported in a few devel
oped countries, mostly after World War II. In
formation on tuberculosis incidence in devel
oping countries is very deficient and cannot be
used for the purpose of this study, since it re
flects the detection rather than the incidence
rate.
However, there is reasonable evidence of
the relationship between the incidence of
tuberculosis and death from tuberculosis in
the pre-chemotherapy era in developed coun
tries. It has been shown that about half of the
reported cases of tuberculosis resulted in
death from the disease before the introduction
x-^f chemotherapy to tuberculosis treatment ;
! in other words, incidence was twice the mor
tality (I, 2, 3). It is also assumed that the
prevalence of smear-positive tuberculosis is
twice the incidence, or that the incidence is
half the prevalence (4).
• A '‘parameter’’ is defined as a “constant” indicating
the numerical value which links two variables togeth
er. Ln this case, we are aiming at esuWishing the rela
tionship between the annual risk of tuberculous infec
tion and the incidence of smear-positive tuberculosis
in the general population.
1
2
Per 100,000.
In Nigeria 53 ; in Somaliland 124 ; in Libya 283 ; in Zanzibar 283.
►Thus we shall adhere to the relationship:
Incidence* • = mortality x 2
Incidence — prevalence’• /2
” of smear-positive pulmonary tuberculosis.
While the relationship between incidence
of and mortality from tuberculosis is well es
tablished, the evidence on the relationship be
tween incidence and prevalence seems more
doubtful, although fairly reasonable.
Material
As already mentioned, we are aiming at es
tablishing the relationship between the annual
risk of tuberculous infection and the incidence
of smear-positive tuberculosis in the general
population. For some countries we have data
on mortality from tuberculosis, while for
others information is available on prevalence
or incidence of smear-positive pulmonary
tuberculosis. The relationship between the
two variables was studied on the basis of data
from the following countries:
— Netherlands
1) the annual risks of infection with tubercle
bacilli for the periods 1910 to 1966 and 1967 to
1979 were published by TSRU, (5) and (6) re
spectively ;
2) the death rates from tuberculosis from
1921 to 1938 and the incidence of smear-
positive pulmonary tuberculosis from 1951 to
1968 were used for the present study.
— WHO tuberculosis prevalence surveys in
various developing countries carried out in
the 1950s and 1960s
1) the annual risk of infection was derived
from data on tuberculin testing carried out by
specialized WHO teams;
2) there were estimates of the prevalence of
smear-positive tuberculosis from the original
WHO surveys.
— Alaska (Eskimos)
1) the risk of tuberculous infection for
Eskimo children was estimated as 25 % during
the years 1948-1951 (7);
2) the mortality rate from tuberculosis
among Eskimos in Alaska as reported in 1950
was used for this study (7).
— India
Two studies are considered in this report :
a) Tuberculosis in a rural population of South
India : a five-year epidemiological study car
ried out in the 1960s (8) (hereafter referred to
as “the longitudinal study”) and b) Trial of
BCG vaccines in South India for Tuberculosis
Prevention carried out in the 1970s (9)
(hereafter referred to as '‘tuberculosis preven
tion trial").
1) the risks of infection in the above two trials
have been derived from the results of tubercu
lin testing among children and young adults ;
2) the prevalence of smear-positive pulmo-
117
a
nary tuberculosis cases in the longitudinal
study and the prevalence of smear-positive
cases in the prevention trial were used for this
study.
Table 1. Relationship between mortality from tuberculosis (all forms) and the annual risk of
tuberculous infection. The Netherlands, 1921-1938
Year
Results
Death rate from
tuberculosis
(per 100,000)1
Risk of tuberculous
infection (%)
Ratio of death
to risk2
The relationship between the risk of
19
6.02
115.1
1922
infection and the incidence or prevalence
20
5.13
100.3
1925
of smear-positive pulmonary tuberculosis
20
4.37
87.9
1928
or mortality from tuberculosis concerning
19
3.72
70.5
1931
18
3.16
the data refened to in the previous sec
55.5
1934
18
2.69
47.7
1937
tion cannot be discussed in detail in this
paper. A full report will be published in
19
1921-1938
the Bulletin of the World Health Organi
zation in 1986. A summary table shows
1 Average for 1921-1923, 1924-1926, etc.
the relationship between the risk of
2 Expressed as mortality from tuberculosis per 100,000 general population for each 1 % risk
tuberculous infection and the incidence
of tuberculous infection.
of smear-positive tuberculosis (the last
but one column).
The summary table shows two areas
(the Netherlands and Alaska with an
Eskimo population) with tuberculosis culosis and the risk of infection in the Ne standardized, and examination methods
mortality ; we have doubled the ratio to therlands in the 1950s and 1960s was 36 were entirely independent of one another.
The data presented in the summary
i arrive at the relationship between the risk smear-positive cases per 100,000 for each
! ofmfection and the incidence of smear- 1 % risk of infection. It is seen that the table refer to 13 African countries. In 10
rates of the risk of infection and the inci of them the risk of infection was between
\ positive tuberculosis.
2 % and 4 %. In Nigeria, the rate was
"iable 1 shows the relationship be dence rates were very low.
The relationship between the risk of higher than 4 %, in Bechuanaland higher
tween tuberculosis mortality and the risk
of infection in the Netherlands from 1921 infection and prevalence of smear than 5 % and in Somaliland it was
to 1938. The table shows that the average positive cases has also been studied, about 8 %. The prevalence^ of smear
death rate from tuberculosis was 115 per based on data taken from WHO tuber positive pulmonary tuberculosis (at all
100,000 for 1921-1923, gradually decreas culosis surveys which sought to deter ages) was mostly between 200 and 300
ing to 48 per 100,000 for 1936-1938. The mine both the infection and disease per 100,000 of the general population ;
annual risk of tuberculous infection was prevalences in a number of African'ahcf however, in Swaziland, Bechuanaland
about 6 % in 1922, decreasing to 2.7 % in Asian countries in the late 1950s and and Somaliland, it was 500 or more per
1937. The ratio of death to risk of infec early 1960s. The material collected by the 100,000 general population.
The ratio of the prevalence of smear
tion was 19 for the whole^ period of 18 WHO teams is unique because the popu
years, ranging from 18 to JO^jgithough lation studied was selected at random, positive cases to 1 % risk of infection in
the- mortality from tuberculosis and the the methods used (for tuberculin testing, 10 of the 13 countries was between 80
risk of infection decreased more than bacteriological examination of sputa for and 120 per 100,000 population ; in one
tubercle bacilli by direct microscopy and country, the ratio was smaller than 80 per
twice during the period under study.
By multiplying the mortality ratio by / X-ray examination of the chest) were 100,000 and in two countries it was
2, we obtain the ratio of incidence to in-1
fection. For the Netherlands, this inci-1
dence/infection ratio is 3S smear-positive
Table 2. Relationship between the incidence of smear-positive tuberculosis and the annual
cases per 100,000 population for each 1 %
risk of tuberculous infection, The Netherlands, 1951-1976
risk of infection (see the summary table).
The second set of information on the
relationship between death from tuber
culosis and risk of infection is on Eskimos
Ratio of inddence
Risk of tuberculoas
Year
Incidence rate of
to risk2
infection (%)
smear-positive tuber
in Alaska from 1948 to 1951. The risk of
culosis (per 100,000)1
infection in Eskimos in the late 1940s and
early 1950s was extremely high ; Com
stock and Philip (7) estimated it to be in
35
0.400
1952
13.9
the order of 25 % (three quarters of
29
0.265
1955
7.8
32
0.176
5.7
\ Eskimo children were fountTlo be 'tposi1958
40
0.116
4.6
1961
11ive ’ at the age of 3-4 years).
0.077
42
1964
3.2
The tuberculosis death rates were also
47
0.051
2.4
1967
extremely high with a figure of 650 per
47
0.038
1973-1976
1.8
100,000 general population. The ratio of
39
1951-1976
incidence of smear-positive tuberculosis
to risk is 2 x 26, i.e. 52 new smear-positive
cases per 100,00frpwpulation for each 1 %
1 Average for 1951-1953, 1954-1956, etc.
2 Expressed as incidence of smear-positive cases per 100,000 general population for each
risk of infection. ,
1 % risk of tuberculous infection.
Table 2 shows that the ratio between
the incidence of smear-positive tuber-
I
118
higher than 120 per 100,000. By dividing
the prevalence ratio by 2, we obtain the
incidence ratio of 40 to 60 smear-positive
cases of tuberculosis per 100,000 general
population for 1 % risk of infection.
The data on the relationship between
the prevalence of smear-positive cases
and infection in two trials in India carried
out in the 1960s show a ratio of 106 and
104 respectively, or an incidence of 53
and 52 smear-positive cases per 100,000
population respectively for each 1 % risk
of infection. The risk of infection in the
longitudinal study was about 1.5 % and in
the prevention trial 4 %. The observed
prevalence rates of smear-positive cases
of pulmonary tuberculosis were about
160 per 100,000 in the former and 420
per 100,000 in the latter study.
REFERENCES
Zi.
2.
3.
4.
5.
6.
7.
The present material suggests a rela
tively constant ratio (lower for developed
and higher for developing countries) be
tween the risk of tuberculous infection
and the incidence of smear-positive
tuberculosis, irrespective of the level of
the risk of tuberculous infection. We
assume that in developing countries, 1 %
risk of infection conesponds to an inci
dence of about 50 to 60 smear-positive
cases of pulmonary tuberculosis.
■
8.
9.
DROLET G.J. Present trend of case fatality rates in tuberculosis.
American Review of Tuberculosis, 37: 125-151 (1938)
LINDHARDT M. The Statistics of Pulmonary Tuberculosis in
Denmark, 1925-1934. A statistical investigation on the occur
rence of pulmonary tuberculosis in the period 1925-1934,
worked out on the basis of the Danish National Health Service
file of notified cases and of deaths. Ejnar Munksgaard, Copenha
gen, 1939
GALTUNG-HANSEN O. Tuberculosis mortality and morbidity
and tuberculin sensitivity in Norway. World Health Organiza
tion, Regional Office for Europe, EURO-84/)5, 28 November
1955
HOLM J. Our enemy. The tubercle bacillus. International
Tuberculosis Digest, 5, IUAT (1970)
STYBLO K., MEIJER J. & SUTHERLAND I. The transmission
of tubercle bacilli. Its trend in a human population. Tuberculosis
Surveillance Research Unit Report N’ 1. Bulletin of the Interna
tional Union Against Tuberculosis, 42: 5-104 (1969)
SUTHERLAND I.. BLEIKER M.A., MEIJER J. & STYBLO K.
The risk of tuberculous infection in the Netherlands, 1967-1980
Tubercle, 64:241-253 (1983)
COMSTOCK G.W. & PHILIP R.N. Decline of the tuberculosis
epidemic in Alaska. Public Health Reports, 76: 19-24 (1961)
National Tuberculosis Institute. Bangalore. Tuberculosis in a
rural population of South India : a five-year epidemiological
study. Bulletin of the World Health Organization, 51: 473-488
(1974)
Tuberculosis Prevention Trial, Madras. Trial of BCG Vaccines
in South India for Tuberculosis Prevention. Indian Journal of
Medical Research, 72 (Suppl.): 1-74 (1980)
119
DISTR
UMIT = D
DISTR. . L1MITEE
WORLD HEXLTH ORGAMZATION
WHO/TB/83.154
ORGANISATION MONDIALE DE LA SAME
ENGLISH ONLY
ANNUAL RISK OF TUBERCULOUS INFECTION
by
3
2
1
G.M. Cauthen , A. Pio and H.G. ten Dam
1
Division of Tuberculosis Control, U.S. Centers for Disease Control, Atlanta, Georgia, USA.
2
Formerly Chief, Tuberculosis and Respiratory Infections Unit, Division of Communicable
Diseases, World Health Organization, Geneva, Switzerland.
3
Scientist, Tuberculosis Unit, Division of Communicable Diseases, World Health Organization
Geneva, Switzerland.
This document is not issued to the general public, and
all rights are reserved by the World Health Organization
(WHO), The document may not be reviewed, abstracted,
quoted. reproduced or translated, in part or in whole,
without the prior written permission of WHO. No part
of this document may be stored in a retrieval system or
transmitted in any form or by any means • electronic,
mechanical or other without the prior written permission
of WHO.
Ce document ii'est pas destine a etre distribue au grand public
et tous les droits y afferents sont reserves par I’Organisatior
mondiale de la Sante (QMS). 11 ne peut etre commente, resume
cite, reproduit ou traduit, partiellement ou en totalite. san
une autorisation prealable ecrite de I'OMS. Aucune parti*
ne doit &tre chargee dans un systeme de recherche documer
taire ou diffusee sous quelque forme ou par quelque moyer
que ce soil • electronique, mecanique, ou autre ■ sans une autc
risation prealable ecrite de I'OMS.
The views expressed m documents by named authors are
solely the responsibility of those authors.
Les opinions exprimees dans les documents par des auteur
cites nommement n'engagent que lesdits auteurs.
WHO/TB/88.154
page 2
TABLE OF CONTENTS
Page
I, INTRODUCTION
3
II. METHODS
3
III. RESULTS
5
IV. DISCUSSION
8
V. CONCLUSIONS
10
REFERENCES
11
FIGURES:
Recent estimates of annual risk of tuberculosis infection
in developing countries
FIGURE 1:
FIGURE 2:
Annual risk of tuberculosis infection
(except Algeria)
19
I. WHO African Region
20
FIGURE 2: (continued) Annual risk of tuberculosis infection
of the Americas
II. WHO Region
FIGURE 2: (continued) Annual risk of tuberculosis infection
South-East Asia region
III. WHO
FIGURE 2: (continued) Annual risk of tuberculosis infection
Mediterranean Region (and Algeria)
IV. WHO Eastern
FIGURE 2: (continued) Annual risk of tuberculosis infection
Pacific Region
V. WHO Western
21
22
23
24
TABLES:
TABLE 1.
ANNUAL RISK OF TUBERCULOSIS INFECTION IN DEVELOPING COUNTRIES
BASED ON TUBERCULIN SURVEYS:
I.
II.
III.
IV.
V.
WHO African Region (except Algeria)
WHO Region of the Americas
WHO South-East Asia Region
WHO Eastern Mediterranean Region (and Algeria)
WHO Western Pacific Region
26
28
29
31
34
NHO/TB/88.154
page 3
I.
INTRODUCTION
Risk of infection in a population is in many respects the most informative index of the
magnitude of the tuberculosis problem (Sutherland, 1976; Pio, 1984). The risk of infection
at a particular time indicates the current magnitude of the incidence and prevalence of
infectious cases (Styblo, 1985) and also indicates the magnitude of the tuberculosis problem
years into the future. An observed decline in the risk of infection would be the earliest
indicator of a decline in the epidemic cycle of tuberculosis, resulting from tuberculosis
control activities or improvements in living standards. A rising risk of infection would be
an early indicator of changes in the other direction, signalling the introduction of new
risk factors, such as the spread of human immunodeficiency virus (HIV) infection.
It is known that risk of infection has been declining for many years in developed
countries, but remains at high levels in many developing countries (Styblo, 1984).
In
developing countries in the last decade, a number of surveys of infection prevalence have
been carried out in national populations as well as in smaller populations.
Therefore a project was carried out to assess the current level and trend in the risk of
infection in developing countries by reviewing and assembling tuberculin skin test survey
data available since 1975.
II. METHODS
Tuberculin skin test data collected since 1975 for populations in developing countries
were assembled from reports to the World Health Organization and from the published
literature. Data from surveys of childhood age groups that were judged of sufficient
quality were selected in order to provide as valid and up to date an assessment as practical
of the magnitude of the risk of infection. Comparable prior data for the same countries
were also selected in order to judge whether the risk of infection is likely to have
declined.
Prevalence of infection observed in childhood age groups was used to derive the average
annual risk of infection that would have resulted in the observed cumulative prevalence
rate. Choosing younger age groups allows the calculated average annual risk to be bracketed
within a relatively narrow period of time between the average birth date of the group and
the date of the survey (Styblo et al, 1969).
Each survey was judged on the basis of available documentation for ability to represent
the population and to detect the proportion infected at a particular time, A sample survey
was judged to have measured the proportion infected in the target population if the
probability sampling design and the estimation method appeared correct in concept and
conduct, and skin testing technique appeared adequate to measure the proportion infected in
the sampling units.
Basis for selection
Surveys were selected which met most of the following criteria;
Specified the sampling design including the sampling frame, staging, stratification,
sampling units, allocation, sampling weights, and estimation formulas and documentation
to support that the design was followed with adequate coverage.
Provided reaction-size distributions for surveyed age groups and for bacteriologically
confirmed cases from the same population in order to assess the definition of infection
and in order to judge technique.
WHO/TB/88.154
page 4
Documented type and strength of antigen used, technique of administration and reading,
and quality control procedures.
Specified the methods used to detect and eliminate persons with a history of BCG
vaccination, and described BCG policy and actual coverage in the population age group
surveyed.
Data assembled from selected surveys:
The area population represented and the survey period.
In the youngest age ranges for which sufficient numbers were tested and read, the
number tested and read, the number counted as infected with M. tuberculosis, and the
mean age.
Information documenting quality of the data.
Criterion of infection
Induration reaction size distributions observed in each survey were used together with
general knowledge about the specific and nonspecific components of observable distributions
to set a uniform criterion for counting infected persons in order to improve comparablility
between surveys.
Because it can be assumed chat nonspecific reactions to intermediate strength purified
protein derivative tuberculins are only infrequently larger than the mode of the
distribution of true reactions, an observed distribution of tuberculin reactions will be
nearly free of contamination by nonspecific reactions above its mode.
Because of this, and because distributions among truly infected persons are symmetric,
the number truly infected in a distribution contaminated by nonspecific reactions would be
closely approximated by adding the number at the mode plus twice the number with reactions
larger than the mode.
This method, described by Nyboe (1960) and by Comstock et al. (1971), was used to
determine the number of infected persons in the selected groups when detailed reaction size
distributions were available and when the mode of reactions among infected persons in the
same general population could be estimated from distributions of either bacteriologically
confirmed cases or older age groups in the population.
When such data were not available, a cutoff criterion was accepted as the second choice
if it appeared reasonably adequate based on distribution data.
Prevalence
Prevalence at a particular calendar time for an age group was calculated as the number
infected divided by the number tested and read. Reported prevalence figures were used when
numerator or denominator counts were not specified.
Risk of infection
For comparability, a single uniform method was chosen to derive the approximate average
annual risk of infection (Styblo et al, 1969, pg. 14). By this method, the annual risk of
infection (R) for a group of average age (A) was derived from the prevalence (P) by
R = 1 - (1 - P)i/A
The slope per year (B) of the trend between two risk estimates R^ and Rj at years
Yi and Yj was approximated by
B ’ 1 - (Rj/Ri)1/1
where T » Yj
*1-
WHO/TB/88.154
page 5
Average age
Average age was calculated by taking the midpoint of a single year (e.g., 6.5 for age
given as 6 at last birthday) or a range of years (e.g., 7.0 for ages given as 6 to 7, or 7.5
for ages given as 5 to 9). When available, the weighted average age was calculated from
year-of-age specific frequencies for an age group.
Time at which average risk occurred
For purposes of calculating and plotting trends, the calendar time of occurrence of a
particular average annual risk (YR) was considered to center approximately at the calendar
time of the midpoint of the average lives of the individuals in the age-group. This was
calculated by YR » Yg - A/2, where A is the average age of the group and Yg is the
midpoint time of the survey. For example, a group of average age 6.1 years surveyed at
1978.9 was taken to contribute information about the annual risk of infection for a time
interval centered at 1978.9 - 6.1/2 » 1975.85.
The midpoint time of the survey (Yg) was calculated from the month, if stated, or the
calendar year of the survey, For example, a survey stated to have taken place in 1975 was
counted as occurring at 1975.5. A survey stated to have taken place from 1981 to 1982 was
counted as occurring at 1982.0. A survey stated to have taken place from May 1952 to
September 1953 was counted as occurring at 1953.04 (the midpoint between 4.5/12 + 1952 and
8.5/12 + 1953).
III.
RESULTS
A variety of tuberculin skin-test surveys were found in the published and unpublished
literature: National sample surveys, sample surveys of other large populations, school
surveys, school BCG campaigns, and mass BCG campaigns.
Twenty-five countries were judged for the purposes of this project to have been
adequately surveyed in whole or significant part since 1975.
Figures 1 and 2 display the results for each country by region. Table 1 contains the
data used in the figures and describes the area population represented, the survey design,
the level of BCG coverage in the age-group surveyed, and the antigen and criterion defining
infection.
An arithmetic scale is used in Figure 1 in order to display the relative magnitudes of
current estimates of annual risk of infection.
A logarithmic scale of annual risks is used in Figure 2 in order to represent constant
percentage decreases or increases between current estimates and prior estimates as straight
line trends.
Dark symbols in the figures indicate risk estimates based on 42 well-conducted national
sample surveys in 16 countries. In Figure 2, dark solid lines are used to connect rates
which are based on more than one large sample survey of the same country.
Light symbols in Figure 2 indicate risk estimates based on data representative of
lesser populations.
Dark dashed lines in Figure 2 are used to connect risk estimates based on national
sample survey data with risk estimates based on other substantial data such as tuberculin
testing in the mass BCG campaigns. A light solid line was used to connect risk estimates
from repeated surveys representative of lesser populations. A light dashed line was used to
connect surveys of approximately the same population done in different years.
'aHO/TB/88.154
page 6-
The symbol + was used in Figure 2 co indicate the average of a set of risk estimates
which together would represent a larger population. Stratum estimates within sample
surveys, plotted as separate points to reveal variation in risk, were displayed as a single
estimate for the whole population in this way. Sets of small area estimates that were less
representative of the whole population than would have been provided by formal probability
sampling, such as the set of completed sampling units in a not yet completed national
survey, or areas selected by judgement from geographic or economic strata in order to be
representative of a large heterogenous population, were also combined in this way.
To indicate the extent of underlying variability, risk estimates from single surveys
chat were representative of only small areas are indicated in Figure 2 by unconnected light
symbols.
Results by region
Africa. The most recent data for the eight countries included indicate current annual
risks of infection around 1 to 2 percent. In the five countries for which comparable
earlier surveys were available, trends over 20 to 25 years appear to be downward at 1 to 6
percent a year.
Data for the nearly completed national survey of Tanzania indicate an annual risk of
infection of 1.1 percent, based on a population weighted average pooling of the samples.
However, comparisons with earlier data available for three of the 18 regions in the sample
do not together suggest a downward trend.
Surveys in Botswana and Lesotho suggest downward trends of about 6 and 1 percent a
year, respectively. In the most recent survey in Botswana, however, BCG coverage in the age
group surveyed had reached 83 percent, leaving only 17 percent who had by whatever
circumstance missed vaccination, upon which to base the estimate.
No national surveys were available for Cameroon, but two surveys of the capital city,
Yaounde, were available. These surveys indicate a level of annual risk of 0.6 percent in
1980 after a 20-year declining trend of 3.6 percent a year.
Survey data collected 30 years apart were available for Addis Ababa, the capital city
of Ethiopia, Th.?-e surveys indicate a 1.3 percent level of annual risk in 1979 succeeding a
downward trend c. 3.7 percent a year from the 4.1 percent risk level in 1949. A survey of
selected village in the Southwest, however, indicated a 3.8 percent annual infection risk
in 1973. This suggests that within Ethiopia there is large variability in the level and
perhaps also in the trend of infection risk.
Data from the 1976 sample survey of Gambia indicates an annual risk in 1971 of 1.9
percent. The trend of risk could not be assessed in Gambia, however, The only prior data
available are from the 1958-59 survey of the capital city, Bathurst, The annual risk of
infection in the urban stratum (4.4 percent) was higher than in the rural stratum (1.7
percent), which was nearly the same as the national level in 1971.
The annual risk of infection derived from skin test data for all military recruits in
Burundi in 1981-82 was 1.2 percent. Data from an earlier survey among a scattered
provincial population were selected in order to provide a rough comparison. The annual risk
of infection in that group was 2.7 percent in 1955.
N’o comprehensive survey of the Nigerian population was available, but data from a
survey of one emirate indicates an annual risk of 2.3 percent in 1973.
_____ . National sample survey data were available only from Argentina, but repeated
Americas
surveys of large local populations within Argentina and also within Brazil were available.
In the latest surveys, infection risks in these local populations varied ten-fold from 0.2
to 2 percent per year. Over time in these local populations, there was evidence of a steep
decline in risk in Argentina, and a lesser decline in Brazil.
WHO/TB/88.154
page 7
The annual risk of infection derived from the 1974-78 survey of the entire population
of Argentina was 0.6 percent. Data from three rsurveys
---- 7--repeated in Santa Fe City and three
surveys repeated in areas around Santa Fe City indicate low
'.-J recent levels in that region of
0.2 to 0.3 percent annual risk, and indicate that the trend
---of2 the previous 12 to 15 years
was steeply downward at a rate of 7 to 10 percent a year, The two trends described by the
two sets of three survey points suggest that the decline was much steeper between the first
i
two surveys but that the trend was nearly
level during the 5 years between’the’second and
third surveys. In the 1960-61 sample survey of another area of Argentina, Resistencia
Province, -the
u'* annual
--- 1 risk -of
e 1 infection
c----- •
was -0.5- percent.
Capitals of Brazilian States and Territories were surveyed in 1970-73 and again in
1980. The annual risk of infection was 0.8 to 1.91 percent in 1976 with evidence of a yearly
downward trend of around 2 to 3 percent. A 1983 survey of the provincial population oi Rio
Grande Do Sol indicated a 0.4 percent annual risk of infection, with a 2.4 fold higher risk
in southern than in northern rural areas.
South-East Asia. Only Thailand had national survey data available,
available but
but crepeated
surveys of local area populations were available for both India and Indonesia, In these
populations, annual risks in the decade of the 1970’s appear to have ranged from about 1 to
5 percent. Annual risks were highest in Indonesia.
Based on the 1977-79 national survey, the annual risk of infection in Thailand was 2.2
percent in 1974. Two earlier surveys, one of Bangkok and Chiengmai Province in 1960-64, and
one of localities of Chiengmai and Kanchanaburi Provinces in 1954, indicate annual risks
only slightly higher than in the national survey 16 to 24 years later.
In India, three large local populations were twice surveyed. Declining trends of 1.6
and 3.5 percent a year were apparent in two of these populations, but an increase of about
1 percent was apparent in the third population. Village surveys in 7 regions throughout
India in 1972 indicate annual risks on the order of 1 to 2 percent. The village survey in
Kashmir, compared to a 1978 sample survey of the Kashmir Valley, suggest a declining annual
trend, but the populations may be only nominally comparable.
Indonesia has followed a policy since 1974 of withholding BCG and conducting 5-yearly
surveys in nine special areas selected to be situated throughout the country. Three surveys
have been completed in five of these areas, two in three areas, and one in another area.
There are also three surveys available from an area with low BCG coverage. The latest
surveys in these areas indicate a median annual risk of 2.3 percent (range; 0.7 to 3.9).
Earlier data from the 1964-65 survey in rural East Java indicate an annual risk of 1.6
percent. Comparing the earliest to the latest survey in each of the nine areas with
repeated surveys, the median downward trend in the annual risk of infection was 2.5 percent
a year (ranging from a decline of 6.3 percent to an apparent increase of 3.4 percent a year).
Eastern Mediterranean (and Algeria). For the purposes of this project, Algeria is
grouped in this region instead of with the African countries. National survey data were
available for six countries, and repeated surveys of local populations were available for a
seventh country in this region. Recent annual risk levels appear to be low to very low in
Bahrain, Libya, Algeria and Kuwait (0.2 to 0.5 percent), but high in Afghanistan and
respectively), Recent
f
*
* of risk■ in the Syrian cities of
Pakistan (1.8 and 3.5 percent, respectively).
levels
respectively), The trend of risk
Aleppo and Homs also appear to be very low (0.3 and 0.14, respectively).
also appears to have decreased steeply in the current low risk countries, but not as steeply
in Pakistan and appears to have increased in Afghanistan.
Four local areas sampled in 1980-84 for the nearly completed national survey of Algeria
had also been canvassed in the 1949-52 mass BCG campaign. Repeated surveys were also
available for the population of Blida. Based on these comparisons, annual risk of infection
appears to have decreased about 6 to 7 percent a year in Algeria. A similar trend is
apparent between risk estimates for the Syrian cities of Aleppo and Homs.
WHO/TB/88.154
page 8
In Kuwait between 1972 and 1981, all school entrants (including non-citizens) were
tuberculin tested prior to BCG vaccination.
--- T national sample survey
’ • CCompared to the 1962-63
estimate, the annual risk of infection appears to have declined by, 10) percent a year, In
Bahrain, a decline of nearly 12 percent a year appears to have occurred between the 1969
national sample survey and mass tuberculin testing of students in 1981. A similarly steep
rate of decline may have occurred in Eastern Libya, if the areas and populations covered in
the surveys are comparable.
The annual risk estimate based on the 1974-78 national sample survey of Pakistan, which
was completed except for Baluchistan, compared to the estimate based on the mass BCG
campaign of 1949-54, suggests that the risk of infection declined by 4.2 percent a year, but
that the risk was still high (1.8 percent) in 1972. The risk estimate based on the
partially completed 1961-62 national survey is consistent with this trend.
The annual risk of infection in Afghanistan appears high (3.5 percent in 1978), without
having much changed since 1958, compared with an estimate based on tuberculin testing
program data collected in 1963 in selected schools around the country.
Western Pacific
__ , Eleven national sample surveys were available for five countries in
this region. IIn the Republic of Korea, the annual risk of infection appears to be
declining, but remains high at about 2 percent,. The risk appears to be about 1.8 percent in
the Philippines, based on the recent national survey. Infection risk appears to be
relatively low in Malaysia (0.4 percent) and Samoa (0.5 percent).
In China the annual risk of Infection appears to have been 1 percent in 1975, based on
areas without BCG coverage in the very large 1979 national survey sample.
Based on infection risk estimates from the five 5“yearly national surveys available
from the Republic of Korea, the trend appears to be downward at 3.8 percent a year to a
level in 1984 of 2 percent.
IV.
DISCUSSION
Interpretation of this collection of risk and risk trend estimates must take into
consideration the many potential sources of error, both in determining absolute levels of
risk and in comparing risk estimates among countries or between periods within the same
country. Some of these sources of error appear to have been adequately controlled, but
uu the
influence of other sources of error (identified and unidentified) must be acknowledged1 as
unknown.
Tuberculin skin testing is a usefully accurate technique for separating groups at
relatively high and low risk of future Mycobacterium tuberculosis disease, but is liable to
varying degrees of error in identifying all the infected and specifying all the uninfected.
Sensitization to environmental mycobacteria is a source of nonspecific reactivity to
tuberculin that is especially prevalent in tropical regions (Edwards and Edwards, 1960).
There was evidence in reaction size distributions of a high prevalence of nonspecific
sensitivity in several of the African and Asian populations surveyed, but distribution data
for older groups or for bacteriologically confirmed cases were frequently available for
locating the mode of the distribution of true infections and deriving their number by
assuming the distribution to be symmetric.
There is underlying variability in tuberculin reactivity as evidenced by the variation
between simultaneously applied identical tests in the same individual (Chaparas et al,
1985). Although this reduces precision of estimation, it is not expected to introduce
systematic error.
WHO/TB/88.154
page 9
A variety of antigen preparations and dosages were used, and these differed between
countries and within countries over time. Estimating the number infected by using the
symmetry of the distribution of true reactions about the mode would adjust for differences
in specific biological potency between antigen preparations, but would not be effective in
controlling for large differences in specificity of antigens.
The Mantoux method of administration was virtually always used, and all the selected
studies measured induration of reaction. Techniques.of measurement varied somewhat, but
almost always consisted of transverse measurement at 48 or 72 hours after administration.
Waning of tuberculin reaction size over time, and complete reversion to zero reactivity
are recognized to occur (Sutherland, 1971). To the extent reversions have occurred in a
population, tuberculin test prevalence at a point in time will underestimate the proportion
who were ever infected. But limiting to groups of about the same low average age may have
minimized and kept comparable the effects of cumulative reversion.
Boosting is a potential problem in survey designs that retest the same individuals
(Styblo, 1984). Boosting is the effect of an initial tuberculin test which, although it
elicits slight or no reactivity itself, apparently acts as an antigenic stimulus able to
recall waned or reverted preexisting mycobacterial sensitivity so that a subsequent test
does elicit reactivity that is indistinguishable from conversion caused by M. tuberculosis
transmission in the interval between the tests. Although some of the data compiled in this
project came from surveys using a repeated testing design, the problem of boosting was
avoided by selecting age groups too young to have been tested in earlier rounds.
BCG vaccination, a major component of tuberculosis control activities in developing
countries, compromises the usefulness of tuberculin testing as a means to survey risk of
infection. Tuberculin skin testing cannot distinguish between sensitivity caused by BCG
vaccination and sensitivity caused by natural infection with M. tuberculosis, because BCG
vaccination, unlike nonspecific sensitization, causes reactions that are distributed with
nearly the same mode and shape as reactions caused by true infections with M. tuberculosis.
For this reason the risk of infection in the whole population can only be inferred from risk
measured in the unvaccinated.
In some of the countries for which survey data were selected, the policy was to
vaccinate indiscriminantly without using tuberculin testing to select the as yet
uninfected. In these countries, surveys limited to children without evidence of vaccination
are subject to bias only to the extent that children selected for vaccination, either by
choice or by local availability, are at a different risk of infection than the
unvaccinated. This is a potentially severe source of bias if BCG coverage is very high.
In some of the surveys selected for this project, none of those surveyed had been
vaccinated, either because BCG had not been introduced into that locale, or because BCG was
withheld by policy until a later age. In other surveys, the unvaccinated were identified by
physical examination to detect the characteristic scar. Levels of coverage by BCG were very
low in some surveys, in which case representation of the whole population by the
unvaccinated would be unaffected. In a few surveys, BCG coverage was so high that
representation was demonstrably affected.
In other countries, the policy was to tuberculin test and vaccinate only those with
weak or no reactions, In this situation, a later survey among children without BCG scars
would be liable to severe overestimation bias. This was avoided by selecting age groups
before the age of eligibility for selective vaccination programs in effect in a particular
country.
Failure to test and measure every person in a survey sample can introduce a bias if
those not reached by the survey are at a different risk of infection than those who were
reached. A severe bias might also result from failure to measure a high proportion of those
tested, because whether a survey participant returned for reading could well depend on
whether there had been a reaction. Surveys with severe problems of this kind were not
selected. Adequate coverage was a criterion for selection by this project, and was high in
most of the surveys selected.
aHO/TB/88.154
page
10
Although the average annual risk for a group lies in the calendar interval from birth
to the time of the survey, it need not be at the interval midpoint.
If the risk is changing
substantially over time, as data from several of the surveys indicate, or varies with age,
as has been suggested by the work of Sutherland (1976), and as might be expected in some of
the diverse cultures surveyed, a complex model is necessary to estimate the risk at a
particular time (Styblo, 1969). Limiting to young, narrow age groups was used as a way to
bracket risk estimates within usefully narrow time intervals.
Average annual risk estimates for the same populations surveyed in different years are
connected with straight lines in Figure 2 to indicate the average trend that would result in
the observed risk estimates. An exponential trend is to be expected in the situation of a
long-term constant ratio of the number of new cases produced by each existing case
(Sutherland, 1976). However, without data for intermediate years, it is difficult to know
the true shape of a trend line.
In some instances where more than 2 surveys were available,
the trend appeared to follow a fairly constant geometric decrease, but in other instances,
the trend 'npeared to have changed over time.
Many
but fo- th
consistent
analysis
V.
ey reports presented more detailed analyses of infection risk and its trend,
eposes of this review, methods were adopted that could be simply and
.pplied to all data. In the few instances of large disagreement, the present
□e in error.
CONCLUSIONS
This attempt to review and compile data available from major surveys carried out in
developing countries since 1975 falls far short of providing a comprehensive picture of
tuberculous infection risk and trend in the entire developing world. The assembled data
cannot safely be generalized beyond the specific target populations surveyed. However, the
assembled data do provide several recent objective examples of apparent progress and lack of
progress in tipping the balance against the tuberculosis epidemic.
The review found that since 1975 many large scale surveys have been successfully
carried out in the developing countries, with impressive attention to correct procedure and
technique.
In most of the countries repeatedly surveyed in whole or in part, risk of infection
does appear to have diminished. In some of these countries, the decline appears to have
been substantial, and current levels of risk have become very low in a few countries.
In other countries, however, rates of infection appear to remain high and there is no
indication of decrease.
It may be speculated that general improvements in living standards and tuberculosis
control programs are having an effect against the tuberculosis epidemic in some parts of the
developing world as it already has in the more developed countries. In the rest of the
developing world, however, very little is known of the current magnitude and trend of the
risk of tuberculous infection.
WHO/TB/88.154
page 11
REFERENCES
Bleiker, M.A. & Styblo, K. (1978) The annual tuberculosis infection rate and its trend in
developing countries. Bulletin of the International Union Against Tuberculosis, 53, 295-298.
Comstock, G.W., Furculow, M.L., Greenburg, R.A., Grzybowski, S., Maclean, R.A., Baer, H. &
Edwards, P.Q. (1971) The tuberculin skin test. American Review of Respiratory Disease, 104,
769-775.
Chaparas, S.D., Vandiviere, H.M., Melvin, I., Koch, G. & Becker, C. (1985) Tuberculin
test. Variability with the Mantoux procedure. American Review of Respiratory Disease, 132,
175-177.
Edwards, P.Q. & Edwards, L.B. (1960) Story of the tuberculin test from an epidemiologic
viewpoint. American Review of Respiratory Disease, 81, suppl., 1-47.
Nyboe, J. (1960) The efficacy of the tuberculin test.
Organization, 22, 5-37.
Pio, A.
Epidemiology of tuberculosis
(1984)
Bulletin of the World Health
Minerva Medica, 75, 507-517.
Styblo K., Meijer, J. & Sutherland, I. (1969) The transmission of tubercle bacilli,
trend in a human population. Tuberculosis Surveillance Research Unit Report No. 1.
Bulletin of the International Union Against Tuberculosis, 62, 1-104.
Styblo K. (1984) Epidemiology of tuberculosis.
Gustav Fischer Verlag, Jena.
Its
Infektionskrankheiten und ihre Erreger, VEB
Styblo K. The relationship between the risk of tuberculous infection and the risk of
developing infectious tuberculosis. Bulletin of the International Union Against
Tuberculosis, 60, 117-119.
Sutherland I. (1971) The effect of tuberculin reversion upon the estimate of the annual
risk of infection. Bulletin of the International Union Against Tuberculosis, 45, 115-118.
Sutherland I. (1976) Recent studies in the epidemiology of tuberculosis, based on the risk
of being infected with tubercle bacilli. Advances in Tuberculosis Research, 19, 1-63.
WHO/TB/88.154
page 12
References to Table 1 by Region and country
WHO African Region - Region Africaine de L'OMS
Algeria
(a) Evaluation of the risk of tuberculous infection in a country with a high BCG-coverage in
newborns. Methodology of the national tuberculin survey in Algeria. R Amrane. ITSC 1985
Progress Report, Vol. 2.
(b) Method for evaluating the infection risk (and its trend) in a country with a high
prevalence of tuberculosis and where generalised BCG is applied at birth to newly born
infants. R. Amrane,. H. Ait-Mesbah, P. Chaulet. Bulletin of the International Union Against
Tuberculosis, 59, 141-143, 1984.
(c) Mass BCG vaccination in Algeria, 1949-52, with special reference to statistics on
tuberculin testing and BCG vaccination. Jorgen Nyboe, Mette Soegaard. Tuberculosis
Research Office, WHO, Copenhagen. Published by The International Tuberculosis Campaign,
December 1953.
(d) Resultas des mesures de surveillance de la Tuberculose en Algeria de 1980 a 1985.
R. Amrane, H. Ait Mesbah, M.T. Hani. Reported at XXVIeme Conference Mondiale de 1* Union
Internationale centre la Tuberculose. Singapore, 2-7 novembre 1986.
Botswana
(a) Prevalence of tuberculosis in Botswana and Lesotho; Results of two random sample
surveys. P.B. Fourie & K. Knoetze. The South African Journal of Epidemiology and
Infection, 1,, 32-37 (1986).
(b) Tuberculosis in Botswana. Results of an epidemiological survey 1981.
Epidemiology Unit, Ministry of Health, P.O. Box 10004, Gabarone.
(c) Tuberculosis survey in Basutoland, Bechuanaland, and Swaziland.
Research Office, Copenhagen, April 1958.
E.T. Maganu,
WHO Tuberculosis
Burundi
(a) Chapitre 1. Ampleur du probleme de la TBC au Burundi - calcul de 1*incidence r£ele.
Service d'Integration de la Lutte Centre la Lepre et la Tuberculose. Lutte Nationale Centre
la Tuberculose, Rapport Annuel 1984. Ministere de la Sante Publique, Republique du Burundi.
(b) A tuberculin sensitivity survey in Burundi,
for Africa, P.O. Box 6, Brazzaville.
June - September 1964.
WHO Regional Office
Cameroon
(a) Interet des enquStes tuberculiniques par sondage. A propos d'une enquSte realis£e dans
les Scoles de Yaounde: MSthodologie et rSsultats prSliminaires. H.G. Delolme, P. Blin,
G. Roscigno, M. Merlin, G. le Mao, L. Sentilhes. XV Conf. Tech. OCEAC 1984, 28-34.
Organisation de Coordination pour la Lutte Centre les Endemies en Afrique Centrale,
Secretariat General, B.P. 288, Yaounde, Republique du Cameroun.
(b) A tuberculin sensitivity survey in Cameroon.
Africa, P.O. Box 6, Brazzaville
March - May 1964.
WHO Regional Office for
WHO/TB/88.154
page 13
Ethiopia
(a) ITSC 1983/84 Report & 9th Progress Report.
ITSC.
M.A. Bleiker, N. Dow, H. Ypma, H. Meesters.
(b) A tuberculin skin test survey in Southwestern Ethiopia., G.K. Fuller, N. Gemeda,
Tropical
and Geographical
Medicine, 31, 365-373 (1979).
D. Fuller, V. Demerest. '
_______________
_____________
(c) Control of tuberculosis in Ethiopia.
Journal, 1^, 128-133, (1963).
P. Chasles, A. Octapodas.
Ethiopian Medical
Gambia
(a) Tuberculin survey in Gambia.
ITSC 5th Report.
(b) Skin sensitivity to human PPD and PPD-marianum in schoolchildren in the Gambia
West-Africa, 1976.
(c) Tuberculosis survey in Gambia.
Brazzaville, February 1960.
MHO/PA/23.60.
WHO Regional Office for Africa,
Lesotho
(a) Prevalence of tuberculosis in Botswana and Lesotho: Results of two random sample
surveys. P.B. Fourie, K. Knoetze. The South African Journal of Epidemiology and Infection.
1, 32-37 (1986).
(b) Tuberculosis survey in Basutoland, Bechuanaland, and Swaziland.
Research Office, Copenhagen, April 1958.
WHO Tuberculosis
(c) Report on the initial examination in the WHO-assisted tuberculosis control project
Basutoland-2. Report to the Regional Director, AFRO from Dr Anton Geser, SMO AFRO-53.
Nigeria
(a) Skin sensitivity to human PPD and to PPD-marianum in schoolchildren in the Kazuare
M.A.
&
Emirate in the Kano State of Nigeria 1977. L.
\ Bleiker,
"
. G.L. Pape,
. - 0. Misljenovic,
’'
I. Blijker. International Tuberculosis Surveillance Center.
Tanzania
(a) Tuberculin survey within the national tuberculosis/leprosy programme in Tanzania.
ITSC 9th Progress Report, 1984.
(b) Tuberculin survey in the Shinyanga Region, the United Republic of Tanzania, 1979.
ITSC 7th Progress Report, 1980.
(c) Draft report. Tuberculin survey and tuberculin training in the Morogoro- and
Arusha-Regions in the United Republic of Tanzania. ITSC 1982/83 Progress Report.
(d) The risk of tuberculosis infection in the Dodoma Region, Tanzania. Results of a
tuberculin survey among schoolchildren carried out in November and December 1978 by a team
of the TB/Leprosy Unit of the Ministry of Health and the Dodoma Regional Leprosy/TB Scheme.
Report to Dr Antonio Pio, WHO TRI Unit. Dr Jaap F. Brockmans, 12 December 1978,
P.O. Box 876, Dodoma.
(e) The national tuberculin survey, Tanzania. M.A. Bleiker, S.J. Nkinda, 0. Misljenovic,
D.W. Mulder & K. Styblo. TSRU Davos Meeting, October 1985.
WHO/TB/88.154
page 14
(f) Skin sensitivity of human PPD in schoolchildren in Tanga in the United Republic of
Tanzania, 1977. M.A. Bleiker, Dr Madundo, 0. Misljenovic, M.h. Milla, I.F. Abdallah. ITSC.
(g) Tuberculosis programme implementation:
22-26 September 1986.
Tanzania.
K. Styblo.
(h) The National Tuberculosis/Leprosy Programme in Tanzania,
visit to Tanzania, February 1987. K. Styblo.
TRI/TB;PHC/86.12
Summary Report no. 17 on Che
(i) Tanzania national tuberculin survey. M.A. Bleiker, H.J. Chum, S.J. Nkinda, K. Styblo.
TSRU/IUAT Progress report 1987, Volume 2, 135-146.
Region of the Americas - Region des Am£riques de I1QMS
Argentina
(a) Determination del Riesgo de Infeccion Tuberculosa en la Republica Argentina.
Resultados de la primera encuesta tubercullnica 1974 - 1978. Ministerio de Bienestar
Social, Secretarfa de Salud Publica.
Instituto Nacional de Tuberculosis, Recreo, Santa FS,
Argentina
(b) The tuberculosis situation in Argentina.
1983, 8-10.
PAHO Epidemiological Bulletin, Vol. 4, No. 3,
(c) Official Communication from Dr Eduardo Balestrino, Director Institute Nacional
Epidemiologica, 10 de marzo de 1986.
(d) Informe sobre la encuesta de prevalencia de la tuberculosis en la Provincia del Chaco..
Ministerio de Asistencia Social y Salud Publica de la Nacion. Organizacion Mundial de la
Salud, UNICEF, Neuqu£n, Marzo 1961.
Brazil
(a) Escudo da prevalencia de infeeqao tuberculosa na populaqao escolar de 6 - 7 anos no Rio
Grande do Sul. Resultados do 1° tempo, 1983 -1988, Volume 1. Estado do Rio Grande do
Sul, Secretaria da Saode e do Meio Ambiente, Porto Alegre, 1984.
(b) Informe final sobre la consultoria en analysis de informacion epidemilogica y
operational del programa nacional de control de tuberculosis. Division Nacional de
Pneumonologia Sanitaria Ministerio de Salud, Rio de Janeiro, Brasil
Del 8 de noviembre al 6 de dicimbre de 1981. Dr E. Balestrino, Consultor, Organizacion
Panamericana de la Salud.
WHO South-East Asia Region - Region de I'Asie du Sud-Est de I1QMS
India
(a) Assignment report on a tuberculosis longitudinal survey, National Tuberculosis
Institute, Bangalore, WHO Project: India 0103. Dr T. Olakowski, Senior Medical Officer (in
cooperation with the members of the Technical Co-ordination Committee at the National
Tuberculosis Institute, Bangalore), January 1969 - March 1972. SEA/TB/129, 12 September
1973 Restricted. SEA-73/2584.
(b) Tuberculosis in a rural population of South India: Report on five surveys.
A.K. Charaborty, H. Singh, K. Srikantan, K.R. Rangaswamy, M.S. Krishnamurthy, J.A. Stephen.
Ind. J. Tub., Vol. XXIX, No.3, 1982, 153-167.
(c) Some aspects of a tuberculosis prevalence survey in a South Indian district, Raj
Narain, A. Geser, M.V.
mbunathan, M. Subramanian. Bull WHO, 1963, _29, 641-664.
WHO/TB/88.154
page 15
(d) Prevalence of infection among unvaccinated children for tuberculosis surveillance.
A.K. Chakraborty, K.T. Ganapathy, G.D. Gothi. Indian J Med Res, 72, July 1980, 7-12.
(e) Changes in the prevalence rates of infection in younger age groups in a rural population
of Bangalore District over a period of 5 years. A.G. Kurthkoti, Hardan Singh.
National Tuberculosis Institute Newsletter (1985) 21/2, 28-40.
(f) Prevalence of tuberculosis in a South Indian District - twelve years after initial
survey. G.D. Gothi, A.K. Chakraborty, S.S. Nair, K.T. Ganapathy, G.C. Banerjee. Ind. J.
Tub. , Vol. XXVI, No. 3, 122-135.
(g) Tuberculosis prevalence survey in Kashmir valley. S. Mayurnath, D.S. Anantharaman,
G.V.J. Baily, M.P. Radhamani, R.S. Vallishayee, P. Venkataraman, S.P. Tripathy. Indian J
Med Res 80, August 1984, 129-140.
(h) Prevalence of non-specific sensitivity in some parts of India. Raj Narain,
M.S. Krishnamurthy, D.S. Anantharaman. Indian J Med Res, 63, 8, August 1975, 1098-1109.
(i) Tuberculosis prevalence survey in Tumkur District. Raj Narain, A. Geser,
M.V. Jambunathan, M. Subramanian. Ind. J. Tub., Vol. X, No. 3, 87-116.
Indonesia
(a) The risk of tuberculosis infection in the District of Tangerang (Indonesia) derived from
the results of tuberculin testing of schoolchildren aged 7 to 10 years, 1972-1983. ITSC
1983/84 Report.
(b) Tuberculin resurveys in two areas in Indonesia.
ITSC 1981/82 Report.
(c) Tuberculosis Research and Control - Control of bacterial and intestinal diseases.
Assignment Report, 1 December 1981 ■" 28 February 1982. Dr Yoshikumi Azuma, WHO Short-Term
Consultant. WHO Project I NO BVM 001. SEA/TB/167, 30 September 1982, Restricted.
(d) Tuberculosis control in Indonesia 1952-65. Report on WHO projects: SEARO 0003 and
Indonesia 0050. Section I: Text, Section II: Data Tables. Prepared by WHO Regional Office
SEA/TB/92 I,II, SEA/VHS/95 I,II, 19 December 1968. Restricted SEA-68/2638.
(e) Annual tuberculosis infection rates between 1972 and 1978 in the district of Tangerang
(Indonesia) based on the results of tuberculin skin testing of schoolchildren aged 7-10
years. ITSC 6th Progress Report.
(f) Skin sensitivity to human PPD and to avian PPD in schoolchildren in Indonesia, 1972.
H. Kusnadi, A.S. Gunardi, M.A. Bleiker. ITSC Confidential Report in WHO files.
(g) 1986 Report of Dr Tripathy.
(h) ITSC working papers.
Thailand
(a) Epidemiology of tuberculosis in Thailand.
Unit files.
Suchart Daramas, MD.
Report in the WHO TRI
(b) Situation of tuberculosis in Thailand. Results of first national epidemiological survey
conducted in Bangkok and the Province of Chiengmai during 1960 - 1964. Proceedings of
Eastern Regional Meeting, Bangkok 1964.
(c) Data for the assessment of naturally acquired tuberculin sensitivity in seven countries
of Asia. WHO Tuberculosis Research Office, Copenhagen, June 1955.
WHO/TB/88.154
page 16
(d) Tuberculosis epidemiology in Thailand.
1969, 157-162.
Dr Boonsong Sunakorn, J Med Assoc Thailand, 32,
WHO Eastern Mediterranean Region - Region de la Mediterranee Qrientale de 1'OMS
Afghanistan
(a) EM/TB/159 EM/AFG/BVM/001 February 1983. Assignment Report. National Tuberculosis
Survey in Afghanistan, 3 August - 2 November 1982, by Dr K.S. Aneja, WHO Consultant.
(b) SEA/TB/96, 10 September 1969. Restricted. /Assignment “Report on Tuberculosis
.
Advisory
Services (WHO Project: Afghanistan 0033 (UNDP/TA) January 1965 - July 1968 by Dr G. Khan,
WHO Medical Officer.
(c) EM/SEM.TB/4.1, 20 August 1975. Regional Seminar on Recent Trends in Tuberculosis
Control, Karachi, 23-30 October 1975. Tuberculosis Control in the Eastern Mediterranean
Region. Review of the Present Situation by Dr J. Kaleta, WHO Regional Adviser on
Tuberculosis.
Bahrain
(a) Survey of tuberculous infection in schoolchildren in Bahrain.
63 (1982), 287-289.
M. Ilyas Khan, Tubercle,
Kuwait
(a) Kuwait National Tuberculosis Control Program. Dr Mohammed Abdel Aty, Head, Tuberculosis
Control Unit. The 18th Middle-East Regional Conference of the IUAT, 5-8 February 1983,
Kuwait.
Libya Arab Jamahiriyah
(a) Assignment Report.
A national tuberculosis prevalence survey in the Socialist People’s
Dr Syed All Husain, WHO Medical
Officer and Mr lb Thorup, WHO Statistician. EM/ST/121, EM/TB/152, EM/LIY/BVD/001, August
1978.
Libyan Arab Jamahiriyah, February 1976 - December 1977.
(b) Regional tuberculosis prevalence survey. Report on Cyrenaica, Libya, 11 July to
22 September 1959. Epidemiological and Statistical Centre, WHO Regional Office for the
Eastern Mediterranean, Alexandria, CAR. EM/TB/58, EM/ST/14, February 1961.
(c) Report on BCG assessment work in Libya.
April 1956.
WHO Tuberculosis Research Office, Copenhagen,
Pakistan
(a) Epidemiological situation of tuberculosis in Pakistan. Results of the National
Tuberculosis Prevalence Survey 1974-1978. Dr Jan Kaleta, Mr Nisar, A. Chaudhry. TB
Seminar, Lahore, April 1982.
(b) Travel Report by Dr Antonio Pio, CDS/TRI, 21.1.85
T9/37O/12PAK.
(c) Report of the tuberculosis survey in Karachi, Rawalpindi & Lahore Division of West
Pakistan. Directorate of Tuberculosis Control (Health Division), Government of Pakistan,
October 1962.
(d) BCG-vaccination programme in Pakistan.
Bull. Wld Hlth Org., 1957, 1_7» 187-202.
E. Roelsgaard, H. Christensen, E. Iversen.
WHO/TB/88.154
page 17
(e) Assignment Report on a visit to Pakistan Tuberculosis Control Programme.
1983.
Dr Jan Kaleta, WHO Consultant.
EM/TB/161, August 1983.
9 may - 6 June
Syrian Arab Republic
(a) The annual risk of tuberculous infection in the Syrian Arab Republic, 1950-1983.
ITSC 1983/84 Report.
(b) Estimates of the risk of infection in schoolchildren in Syria between 1938 and 1978.
ITSC 7th Progress Report 1980.
(c) Tuberculosis survey in the Syrian Arab Republic. Epidemiological and Statistical
Centre, WHO Regional Office for the Eastern Mediterranean, Alexandria, UAR, April 1962.
(d) Mass BCG vaccination in Syria, 1950, with special reference to statistics on BCG
vaccination and pre- and post-vaccination allergy. Published by The International
Tuberculosis Campaign, care of UNICEF, 24 Rue Borghese, Neuilly-sur-Seine, France.
(e) Skin sensitivity to human PPD in schoolchildren in the Governates of Aleppo, Homs, and
Raqua in the Syrian Arab Republic, 1983. M.A. Bleiker, 0. Misljenovic, 0, Post.
I. El-Rifai, B. El-Kateb, W. Sankari, K. Nahawi, N. Nasser, F. Kana, B. Hanifi, S. Trabalsi,
R. Assaf. ITSC Report.
WHO Western Pacific Region - Region du Paciflque Occidental de 1 * QMS
China
(a) Nationwide random survey for pulmonary tuberculosis conducted in 1979. The National
Tuberculosis Control and Research Centtr, September 1985. Draft for possible publication,
submitted to TRI Unit, WHO.
(b) Nation-wide survey for the epidemiology of pulmonary tuberculosis.
Zhonghua Jiehe He Huxixi Jibing Zazhi, 1982, 5_, 67-70 (in Chinese).
Conducted in 1979.
^Malaysia
(a) An analysis on the first community tuberculin survey in peninsular Malaysia (1976-1977).
Dr H.T. Lin and Mr lb Thorup. National Tuberculosis Programme of Malaysia.
ICP/BVD/001-E(Malaysia), 18 February 1979.
Philippines
(a) Report on a national tuberculosis prevalence survey in the Republic of the Philippines
1981-1983. National Institute of Tuberculosis, June 1984.
(b) Second final report from the Philippines BCG mass vaccination programme.
Dr Anton Geser, BCG Consultant. Manila, 1 August 1955. Report in the WHO TRI Unit files.
Republic of Korea
(a) Report on the Sth tuberculosis prevalence survey in Korea, 1985.
Social Affairs, Korean National Tuberculosis Association.
Ministry of Health and
(b) Assignment report. Tuberculosis control, Republic of Korea, 5 February to 3 March 1986.
Dr Tomi Mori, WHO Consultant (WP)CHD/ICP/TUB/001 E, 3 November 1986.
(c) The value of periodical tuberculosis prevalence surveys to assess the epidemiological
trend of the problem in developing countries. Ingela Sjb’gren. ITSC 1984 Progress Report,
Vol. 1.
WHO/TB/88.154
page 18
(d) Report on the 4th Tuberculosis Prevalence Survey in Korea. 1980.
Social Affairs, Korean National Tuberculosis Association.
Ministry of Health and
(e) Prevalence Survey in Korea, B.W. Jin, The Korean National Tuberculosis Association,
Korean Institute of Tuberculosis, 1984. Report in the WHO TRI Unit files.
(f) Summary report on a national tuberculosis prevalence survey in Korea. (June - November
1965). WHO Regional Tuberculosis Advisory Team, WHO Regional Office for the Western
Pacific, Manila, Philippines. WHO/TB/Techn.Information/67.57.
Samoa
(a) Report on the tuberculosis/leprosy prevalence survey in Samoa, 26 June - 24 October 1975.
WHO Regional Tuberculosis Control Team and WHO Leprologist. ICP/BVD/001-E (Western Samoa)
5 December 1977.
(b) Assignment report. Tuberculosis control, Western Samoa, 18 September to 7 July 1974.
Dr Nak Chin Chung, WHO Consultant Western Samoa. 1201-E (WES/MBD/01), 25 October 1974.
(c) Assignment report. Tuberculosis control, Western Samoa, 8 March 1980 to 20 December
1981. Dr Qian Yuan Fu, WHO Medical Officer. SMA/BVM/002-E, 13 April 1982.
RISK (%)
4
LEGEND:
A Afghanlatan
3 Thailand
Gambia
2 -
Oa
Laaotho
Q
□ □
Paklatan
Korea
Philippines
Botswana
□
1
Argentln
Tanzania
Samoa
"IMalayala
2^ Algeria
£1
-Kuwait Xi
A A Bahrain
Libya
0
1970
0 AFRICA
O AMERICAS
Q SOUTH-EAST ASIA.
A EASTERN MEDITERRANEAN
□ WESTERN PACIFIC
T
T
T
1975
1980
1985
YEAR
0)
0Q
CD
X
O
f- tn
oo
00
Cu
OQ
rt)
X
O
H
K) CO
O \
OO
00
RISK (%)
5
■
BOTSWANA (-X-)
LESOTHO (0)
ETHIOPIA (□)
V
NIGERIA (O)
2
■
BURUNDI (X)
TANZANIA (A)
GAMBIA (V)
A
1 '
A
Slope reference:
0.5-
CAMEROON (O)
% decllne/year
---- 1%
5%
0.2-
10%
0.1J
1940
1950
1970
1960
YEAR
1980
RISK (%)
5 •
2
1 '
0.5
BRAZIL (A)
ARGENTINA (Q)
Slope reference:
% decllne/year
A
£
1%
5%
0.2-
A
10%
0.1J
1940
1950
1960
1970
YEAR
1980
■g £
O
CQ
rt>
•—
N) CO
oo
00
Ln
(b
UQ
n>
i
ro us
oo
oo
RISK (%)
5
INDONESIA (□)
o
Q
2
THAILAND (Q)
1 '
INDIA (A)
Slope reference:
0.5-
% decllne/year
A
1%
5%
0.2-
10%
0.1J
1940
1950
1960
1970
YEAR
1980
RISK (%)
5
AFGHANISTAN (0)
2
PAKISTAN (V)
‘
1
Slope reference:
0.6-
% decllne/year
A
1%
5%
0.2-
LIBYA (^)
10%
A
KUWAIT (A)
ALGERIA (Fl)
□
BAHRAIN (Q)
D
SYRIA (Q)
0.1
1940
1950
1960
1970
YEAR
1980
5
OQ
rt)
O
bJ w
u
CO
oo
•u >r
£i>
OQ
pd
O
n> \
b-> w
oo
oo
RISK (%)
5
Ln
■
KOREA ( A)
2
PHILIPPINES (X)
o CHINA (Q)
1
0.5-
Slope reference:
% decllne/year
— O SAMOA (O)
1%
Q MALAYSIA (O)
5%
0.2-
10%
0.1 J
r—
1940
---------1---------
1950
1960
1970
YEAR
1980
LEGEND
VOQXQAO *
Survey estimates of risk
Average of risk estimates
Trend between risk estimated
In different years
Dark symbols and lines are used to display risk estimates and
trends for national populations and large parts of national populations.
Light symbols and lines are used to display risk estimates
for subnatlonal units.
Solid lines connect surveys that closely represent the same population.
Dashed lines connect surveys that represent approximately the same population.
(u §
o
rt>
OQ
NJ CO
co
00
u>
CTO
TABLE 1.
ANNUAL RISK OF TUBERCULOUS INFECTION IN DEVELOPING COUNTRIES
BASED ON TUBERCULIN SURVEYS
ro co
00
00
1. WHO African Region (except Algeria) - Region Africaine de 1*OMS (sans Algerle)
Ui
Antigen and
Infection
Criterion
Survey
Year
Mldpt.
Age (ml dpt.)
Tested
N
RT23 2TU,
Mode-13
1982.0
0-14 ( 6.5)
257
RT19-20-21 5 TU,
Mode-16
1957.0
0-14 ( 6.6)
1450
Area Represented
Survey design
BCG
(X)
Whole country
Sample survey
83.3
Whole country
Sample survey
0
1982-84a
Military recruits
All recruits
71.0
RT23 2TU, 9mm+
1983.5
18-22 (20.5)
1964 b
Northern section of
Muramvya Province
Sample survey
56.4
RT23 2TU, 9mm+
1964.6
Sample survey
of schools
58.4
PPD 71 10U, 9mm+
Year
Inter- Preva- Risk/
1 ence year
(X)
(X)
ted
N
Hot swiina
1981-«2ub
1956-57c
8. 17
1.30
472
32.55
5.79
912
195
21.38
1.17
15-19 (17.5)
202
76
37.62
2.66
1984.5
7- 9 ( 8.5)
860
46
5.35
0.64
RT23 2TU, 9mm+
1964.3
5- 9 ( 7.5)
326
31
9.51
1.32
RT23 2TU, 10mm*
1983.9
6-15 ( 8.6)
1251
133
10.63
1.30
... 2TU, 10mm+
1977.5
6-10 ( 8.5)
185
52
28.11
3.81
1954.7
7-14 (11
37
4.1
RT23 2TU,
Mode-15
1976.4
6-16 ( 9.4)
16.65
1.92
RT23 ITU,
Mode-1 7
1959.1
5-14 ( 9.4)
16.38
1.89
34.66
15.25
4.43
1.74
Burundi
Cameroon - Cameroon
1 984a
1 964b
Yaounde
Urban Yaounde
Sample survey
Ethiopia - Ethiople
1983a
1 97 7b
1953-55c
Addis Ababa
Southwest area
Addis Ababa
Sample of
schools
49.4
Selected
villages
BCG school
campaign
0
)
Gambia
1976ab
1958-59c
Whole country
Bathurst In 2 strata
Urban stratum
Rural stratum
Sample survey
Sample survey:
at 50 per 1000
at 7 per 1000
0.7
X
o
2397
176
400
399
61
61
1. WHO African Region (except Algeria) - Region Africaine de I'OMS (sans Algerie) (continued)
Antigen and
Infection
Criterion
Survey
Yea r
Midpt.
Age (mldpt.)
Tested
N
RT23 2TU,
Mode-14
1982.0
0- 9 ( 5.2)
158
Area Represented
Survey design
BCG
(Z)
1981-823
Whole country
Sample survey
70.4
195b-57b
Whole country
Sample survey
0
RT19-20-21 5TU,
Mode-17,18
1957.0
0- 9 ( 4.9)
1101
1962-65c
Rural area
around Maseru
Sample survey
0
RT23 ITU,
Mode-19
1963.8
0- 9 ( 4.9)
10216
Kazaure Emirate of
Kano State
School survey
32.1
RT23 2TU,
Mode-15
1977.4
4-13*( 7* )
RT23 2TU,
Mode-17
1983.6
1984.8
1984.8
1986.4
1986.9
1987.7
1985.4
1984.6
1986.5
1986.6
1984.5
1985.8
1986.4
1984.5
1985.4
1985.7
1987.7
1983.8
1985.6
7-14 (10.0)
( 9.6)
( 9.7)
(10.1)
(10.7)
(10.2)
(10.2)
(10.3)
(10.9)
(11.0)
( 9.7)
(10.2)
(10.8)
(10.4)
( 9.8)
(10.5)
(11.2)
(10.4)
(10.3)
3125
681
1338
1723
1743
1188
2324
2665
1621
865
629
2420
385
1412
3666
985
1 794
2462
30982
RT23 2TU,
lOmm-t-
1979.5
6-14 (10.0)
1817
Year
Infec- Ereva- Risk/
1 ence year
(X)
(Z)
ted
N
Lesotho
10.13
2.03
142
12.90
2.78
1342
13. 14
2.83
15.0
2.29
9.82
10.87
10.24
9.05
20.54
12.04
4.78
9.01
11.91
18.27
19.40
9.5
12.21
9.92
11.6
16.3
8.3
9.67
10.9
1.03
1.19
1.11
0.94
2.13
1.25
0.48
0.91
1.16
1.82
2.20
0.97
1.20
1.00
1.25
1.68
0.77
0.97
1.11
19.59
2.16
Nigeria
1977a
United Republic of Tanzania - Republique-Unle de Tanzania
1983acehlJ
1984
1984
1986
1986
1987
1985
1984
1986
1986
1984
1985
1986
1984
1985
1985
1987
1983
1983-87
Arusha Region
Sample survey
Coast (Pwani) Region
of schools
Dar-es-Salam Region
in progress,
Iringa Region
90Z completed
Dodoma Region
Klgoma Region
Kilimanjaro Region
Llndl Region
Mbeya Region
Morogoro Region
Mtwara Region
Mwanza Region
Rukwa Region
Ruvuma Region
Shinyanga Region
Slnglda Region
Tabora Region
Tanga Region
(18 Area weighted average)
1979b
Shinyanga Region,
except Shinyanga Distr.
Sample survey
1978d
Dodoma Region
Sample survey
63.3
RT23 2TU,
Mode-1 7
1978.9
6-14 ( 10.2)
1329
166
12.49
1.30
19778
Tanga Town and
nearby rural area
Sample survey
of schools
79.0
RT23 2TU,
Mode-1 7
1977.1
7-14 ( 9.7)
383
33
8.62
0.92
52.4
37.6
45.0
57
49
49.1
52.4
31.3
64
67
41.8
67
46.1
50. 1
50.9
307
74
137
156
358
143
111
240
193
158
122
47
140
238
356
•o
a>
oc c
rj cd
Oo
OD
nJ
PJ
UQ
0)
x
o
rj us
11. WHO Region of the Americas - Region des Ameriques de 1’OMS
00 \
CD
CD
Year
Area Represented
Survey design
BCG
(X)
Antigen and
Infect ion
Criterion
Survey
Yea r
Mldpt.
Age (midpt.)
Tested
N
RT23 2TU,
1976.5
6- 7 ( 6.5)
26902
Infec- Preva- Risk/
ted
1 ence year
(X)
N
(X)
Argentina - Argentine
1974-78ab
Whole country
Sample survey
1983c
1979-8Oc
1967-68c
Areas around
Santa Fe City
Sample survey
81.6
63.7
31.6
. . , 10mm*
.., 10mm*
•. , 10mm*
1983.5
1980.0
1968.0
5- 8 ( 7.0)
6- 7 ( 7.0)
5- 8 ( 7.0)
1979-80c
1974-75C
1967-68c
Santa Fe City
All schools
57.7
12.9
.. , 10mm*
.. , 10mm*
.•, 10mm*
1980.0
1975.0
1968.0
1960-6 ld
Resistencia Province
Sample survey
Rare
RT25 ITU, 10<nm+
0
RT23 2TU, 10mm*
Brazi1
1 983a
10mm+
1970-73b
3.61
0.56
325
443
1221
1.8
2.0
8.8
0.26
0.29
1.31
6- 7 ( 7.0)
6- 7 ( 7.0)
6- 7 ( 7.0)
2125
3590
3196
1 .6
1.6
3.9
0.23
0.23
0.57
1961.0
5- 9 ( 7.5)
1259
49
3.89
0. 53
1983.5
7- 8 ( 7.3)
11880
333
2.80
0.39
1983.5
6- 7 ( 7.0)
3507
4163
3670
3.88
3.17
1.31
0.56
0.46
Bres 11
Rio Grande Do Sol
Province
Sample survey of
school enterers
By Provincial zone:
Metropolitan
Southern rural
Northern rural
1980b
970
Capitals of States and
Territories by Region:
North
Northeas t
Southeast
Sao Paulo State
South
East Central
All school
enterers
Capitals of States and
Territories by Region:
North
Nort heas t
Southeast
Sao Paulo State
South
East Central
All school
enterers
1980.5
0.19
... ( 7.7)
1.95
1.56
1.17
0.94
0.78
0.78
RT23 2TU,
10mm*
1972.0
... ( 7.7)
16.9
14.2
8.3
6.7
7.9
2.38
1.97
1. 1
1.2
0.90
1. 1
vn
111. WHO South-East Asia Region - Region de I'Asie du Sud-Est de 1'QMS
Infec- Preva- Risk/
ted
lence year
N
(X)
(X)
BCG
(X)
Antigen and
Infect ion
Cri terion
Survey
Year
Midpt♦
14.1
36
RT23 ITU, 1 2mm+
RT23 ITU, 10wm+.
1979.5
1974.8
0- 4 ( 2.5)
5203
3805
128
98
2.46
2.58
0.99
1.04
10<did+
1977.9
1962.2
0- 4 ( 2.5)
1492
7981
36
166
2.41
2.08
0.97
0.84
RT23 ITU, 10nun+
1973.0
1960.9
0- 4 ( 2.5)
4277
3788
115
155
2.69
4.09
1.08
1.66
12mm+
1978.7
0- 4 ( 2.5)
2448
56
2.29
0.92
1972.8
1- 4 ( 3.0)
679
503
381
290
530
612
76
4.0
5.2
3.1
4.1
5.2
4.4
1.3
1.4
1.8
1.0
1.4
1.8
1.5
0.4
5.1
1.7
Age (midpt.)
Tested
N
Area Represented
Survey deaign
1979®
1974-75de
Dodballapur in
Bangalore District
Sample survey
repeated in same
villages
1977-78b
1961-63®
3 Subdistricts in
Bangalore District
Repeated
sample surveys
1972-73f
1960-61cfl
Tumkur District in
South India
Sample survey
repeated in
same 62 villages
19788
Kashmir Valley
Sample survey
1972h
Villages in:
Kashmir
Kulu Valley
Lohaghat
Plthoragarh
Agra
Haryana
Rajasthan
Selected for
<5
altitude contrast
PPD-S 5TU, 12mm+
1968-7lh
Northern Chlngelput
Selected for
BCG trial
<5
PPD-S 5TU, 12mm+
1969.9
1- 4 ( 3.0)
27520
Indonesia
Indonesle
1983aefgh
1978
1972
Tange rang Regency
in West Java
Sample survey
repeated in
same schools
18.5
22.1
1.9
RT23 2TU, 10mm+
1983.6
1978.4
1972.2
7-10 ( 8.8)
7-10 ( 8.8)
7-10 ( 9.3)
1549
1649
1371
286
440
497
18.46
26.68
36.25
2.29
3.47
4.73
1986bh
1981
1976
Sambas in West
Kalimantan
Selected as
BCG-free area
0
RT23 2TU,
10m in+
1986.8
1981.8
1976.9
7-10 ( 8.8)
7-10 ( 8.6)
7-10 ( 8.6)
3839
2181
1655
576
310
391
15.00
14.21
23.63
1.83
1.77
3.09
1986bh
1981
1976
Padang Pariaman in
West Sumatra
Selected as
BCG-free area
0
RT23 2TU, 10ma>+
1986.9
1981.7
1976.5
7-10 ( 8.7)
7-10 ( 8.9)
7-10 ( 8.6)
1986
2501
1124
472
363
193
23.77
14.51
17.17
3.07
1.75
2.17
I9848h
1979c8h
Malang In East Java
Selected as
BCG-free area
0
RT23 2TU, lOmoj+
1984.5
1979.5
7-10 ( 8.3)
7-10 ( 8.9)
1406
1122
75
86
5.33
7.66
0.66
0.89
1985c8h
1980
197 5
Ogan Kamering 11 ir in
South Sumatra
Selected as
BCG-free area
0
RT23 2TU, 1 Omm-t-
1985.8
1980.2
1975.6
7-10 ( 8.9)
7-10 ( 8.5)
7-10 ( 8.9)
4840
4839
2425
1309
1358
701
27.05
28.06
28.91
3.48
3.80
3.76
Year
Ind 1 a
Inde
RT23 ITU,
1.5
0.2
PPD-S 3TU,
cro
<T>
X
o
k> oa
oo
oo
uo
(D
I
CD
111. WHO South-East Asia Region - Region de 1'Asle du Sud-Est de 1'QMS (continued)
BCG
(X)
Antigen and
Infection
Criterion
Survey
Year
Midpt.
Age (midpt.)
Tested
N
00
oo
Infec- Preva- Risk/
ted
1 ence year
N
(X)
(X)
Year
Area Represented
Survey design
1985CK‘‘
1980
1975
Gowa in South
Sulawesi
Selected as
BCG-free area
0
RT23 2TU, 10nm+
1985.5
1980.5
1975.8
7-10 ( 8.9)
7-10 ( 8.7)
7-10 ( 8.2)
4001
3573
1429
1198
1076
429
29.94
30.11
30.02
3.92
4.03
4.26
1983c8h
1978
Stabat Langkat
In North Suaatra
Selected as
BCG-free area
0
RT23 2TU, 10wn>+
1983.2
1978.4
7-10 ( 8.5)
7-10 ( 8.9)
1894
1125
196
165
10.35
14.67
1.28
1.77
1982c8h
1977
Hulu Sengal Tengah
in South Kallaantan
Selected as
BCG-free area
0
RT23 2TU, 10nm+
1982.8
1977.3
7-10 ( 9.0)
7-10 ( 8.7)
1577
1199
424
352
26.89
29.36
3.42
3.92
198 48h
1979
1974
Patl In
Central Java
Selected as
BCG-free area
0
RT23 2TU, 10nin+
1984.5
1979.5
1974.5
7-10 ( 8.5)
7-10 ( 8.9)
7-10 ( 8.1)
2938
2197
1070
403
381
143
13.72
17.34
13.36
1.72
2.12
1.76
Langsa in
Selected as
BCG-free area
0
RT23 2TU, 10inm+
1983.5
7-10 ( 9
24.4
3.06
RT23 2TU, 10bo+
1965.0
5- 9 ( 7.5)
11.70
1.64
1978.5
0-14 ( 7.5)
15.2
2.17
1962.5
10-14 (12.5)
29
2.7
1954.9
8-12 (10.5)
23.51
2.52
198314
D.I. Aceh
1964-65dn
Thai land
Rural area In
East Java
Saaple survey
)
1633
191
Thailande
1977-79®
Whole country
Saaple survey
1960-64bd
Bangkok and
Chiengaal Province
Saaple survey
1954c
6 localities in
Chiengaal and
Kanchanaburl Provinces
Selected
unvaccinated
areas
39.1
0
... 5TU, 10aa+
1578
371
Ln
IV. WHO Eastern Mediterranean Region (and Algeria) - Region de la Mediterranee Orientale de 1'QMS (et Algerie)
Year
Area Represented
Survey design
BCG
(X)
Antigen and
Infection
Criterion
Survey
Year
Midpt.
Age (midpt.)
Tested
N
Infec- Preva- Risk/
ted
lence year
N
(X)
(X)
Algeria - Algerie
Whole country in 6
economic zones
1985d
Zone 2 Tlzl Ouzou
and Tlemcen
School survey
every 5 years
Second survey
(commencing)
RT23 2TU, lOnm-t-
8 ( 8.5)
67.2
1985.5
2378
55
2.31
0.27
59.1
68.6
66.3
63.3
44.6
1984.5
1983.9
1982.9
1981.9
1980.9
2125
1144
1067
1334
1844
64
51
45
71
71
3.01
4.46
4.22
5.32
3.85
0.36
0.54
0.51
0.64
0.46
1980-84
Zone 566
First survey
Zone 4 incl. Batna
(nearly complete)
Zone 3 incl. Guelma
Zone 1 incl. Mascara
Zone 2 Tlzl Ouzou
and Tlemcen
(5 Area weighted average)
1982.8
7514
302
4.02
0.48
1981ab
1976ab
Blida
Bllda
Survey repeated
In same schools
71.6
1949-52c
Whole country
Hass BCG
Campaign
1984d
1983ab
1982ab
1981ab
1980ab
Selected areas
Bllda
Tlzl Ouzou
Tlemcen
Mascara
Batna
Guelma
IP48 10TU, 6»a+
1981.5
1976.5
8 ( 8.5)
6-10 ( 8.5)
1117
262
6.18
8.4
0.75
1.03
IP48 STU, 6m+
1P48 10TU, 6mm+
Moro Patch,
3+ papules
1951.1
8 ( 8.5)
110547
31.2
4.30
36.5
25.5
24.3
29.7
35.4
29.8
5.20
3.40
3.22
4.06
5.01
4.08
Afghan i st an
1982ac
Whole country
1963bc
Kabul
Selected
Jalalabad
schools
Mazar-e-Sharif
Kandahar
Pulikumry
(5 Area weighted average)
Sample survey
RT23 ITU, 8mm+
1982.5
5- 9 ( 7.5)
881
208
23.61
3.53
RT23 ITU, 8mm+
1963.5
7-12 (10.0)
7-12 (10.0)
7-12 (10.0)
7-12 (10.0)
13-15 (14.5)
7-15 (10.2)
19006
3492
3170
4079
1191
30938
4322
1083
1019
1586
384
8394
22.74
31.01
32.15
38.88
32.24
27.13
2.55
3.64
3.80
4.80
2.65
3.06
3
tb X
CM O
co
03
Bahrain
oo
Bahrein
1981*
Whole country
All schools
I969a
Whole country
Sample survey
0
RT23 ITU, 6ma+
1981.4
6- 7 ( 7.0)
6151
86
1.40
0.20
RT23 ITU, 6mmt
1969.5
6- 7 ( 7.0)
897
55
6.13
0.90
TJ
Cb
00
n>
x
o
iV. WHO Eastern Mediterranean Region (and Algeria) - Region de la Mediterranee Orientale de 1'QMS (et Algerfe) (continued)
CD
00
Area Represented
Year
Kuwait
Survey design
BCG
(Z)
Antigen and
Infection
Criterion
Survey
Year
Mldpt.
Age (mldpt.)
Tested
N
RT23 ...
1981.5
1980.5
1979.5
1978.5
1977.5
1976.5
1975.5
1974.5
1973.5
1972.5
1978.4
4- 6 (
4- 6 (
4- 6 (
4- 6 (
4- 6 (
4- 6 (
4- 5 (
4- 5 (
4- 5 (
4- 5 (
4- 6 (
22710
16149
22674
20843
1 7444
9018
7665
6722
6363
2258
131846
1963.0
5- 9 ( 7.5)
RT23 2TU,
Mode-16,17
1977.0
5- 9 ( 7.5)
1827
951
RT23 ITU,
Mode-18,19
1959.7
5- 9 ( 7.5)
361
Infec- Prtva- Risk/
ted
lence year
N
(X)
(X)
Kowelt
1981a
1980
1979
1978
1977
1976
1975
1974
1973
1972
1972-81
Whole country
i962-63a
Whole country
School campaign
0
(10 Year weighted average)
Sample survey
5.2)
5.1)
5.3)
5.2)
5.3)
5.1)
4.5)
4.2)
4.3)
4.7)
5.1)
1. 15
0.63
1.25
1.62
1.52
2.23
1.30
0.82
1.26
1.68
1.31
0.22
0.12
0.24
0.31
0.29
0.44
0.29
0.20
0.29
0.36
0.26
10.74
1.50
35
1.92
1.9
0.26
0.26
60
16.62
2.39
24.5
3.68
261
101
283
338
266
201
100
55
80
38
1723
Libya Arab Jamahlrlyah - Jamlhlriyah Arabe Libyenne
63.5
1976-77a
Whole country
East Libya
Sample survey
1959ab
East Province
Cyrenalca
Sample survey
1954ac
Ben Walid
Selected area
0
RT19-20-21 5TU,
Mode-18,19
1954.5
5- 9 ( 7.5)
188
1974-7babd
Whole country
except Baluchistan
Sample survey
7.8
RT23 ITU, 10mm+
1976.5
5- 9 ( 7.5)
2289
298
13.02
1.84
1961-62a-d
Karachi, Rawalpindi,
and Lahore
Sample survey
42.7*
RT23 ITU, l0min+
1962.0
5- 9 ( 7.5)
769
178
23.15
3.45
1949-54bt:
Provincial capitals
Other urban areas
Rural areas
(3 Area median)
Mass BCG
campaign
0
RT22 5TU, 5mm+
1954.1
7-14 (11.0)
47.5
44.5
41.0
44.5
5.69
5.21
4.68
5.21
Pakistan
1V. WHO Eastern Mediterranean Region - Region de la Mediterranee Orientale de 1'OMS (continued)
Year
Area Represented
Survey design
BCG
(Z)
Antigen and
Infection
Criterion
Survey
Year
Mldpt.
Age (midpt.)
Tested
N
Infec- Preva- Risk/
ted
1 ence year
N
(Z)
(Z)
Syrian Arab Republic - Republique Arabe Syrlenne
iy83a
Aleppo
Survey repeated
in some 1978
survey schools
14.1
RT23 2TU, 10n>m+
1983.8
6 ( 6.5)
1586
31
1.95
0.30
1978ab
Aleppo
Sample survey
of schools
8.5
RT23 2TU, 10nim+
1978.8
6 ( 6.5)
1845
31
1.68
0.26
1950ad
Aleppo
BCG school
campaign
Moro patch,
3+ papules
1950.3
6 ( 6.5)
15.3
2.52
1 98 3e
Homs
Sample survey
of schools
RT23 2TU, 10n»m+
1983.8
6 ( 6.5)
0.93
0. 14
1 95Od
Homs
Mass BCG
campaign
Moro patch,
3+ papules
1950.6
6 ( 6.5)
10.4
1.68
1960c
Districts of Homa and
Damascus, excluding
City of Damascus
Sample survey
RT23 ITU, 10mm+
1960.8
5- 9 ( 7.5)
5.43
0.74
0
17.2
0
9*
1182
387
11
21
fu 3?
CN O
OJ
00
co
■£
S
TO O
ro \
uj
V. WHO Western Pacific Region - Region du Paciflque Occidental de 1'QMS
Year
China
BCG
(X)
OO
OO
Antigen and
Infection
Cr 1terlon
Survey
Year
Midpt.
Age (midpt.)
Tested
N
Area Represented
Survey design
Areas without BCG,
in whole country
except Taiwan
266 of the 888
clusters in the
national sample
0
RT23 2TU,
1979.5
7 ( 7.5)
10000*
Saaple survey
81-86
RT23 ITU,
Mode-15
1977.0
1- 9 ( 5.5)
1429
Infec- Preva- Risk/
ted
1 ence year
N
(X)
(X)
Chi ne
1979ab
7.3
1.01
29
2.03
0.37
77
5.42
1.97
Malaysia - Malaysle
1976-77®
Whole country
Republic of Korea - Republique de Coree
1985ab
Whole country
Saaple survey
65.9
RT23 ITU, 10bb+
1985.5
0- 4 ( 2.8)
1420
1980cd
Whole country
Saaple survey
50.0
RT23 ITU, 10bb+
1980.5
3«- 4 ( 2.1)
1310
4.89
2.36
1975c
Whole country
Saaple survey
48.2
RT23 ITU, 10bb+
1975.5
3b- 4 ( 2.1)
1871
4.8
2.32
1970c
Whole country
Saaple survey
RT23 ITU, 10bb+
1970.5
3b- 4 ( 2.1)
8.5
4.14
1965cef
Whole country, 2 strata
Saaple survey
RT23 ITU, 10bb+
1965.5
0- 4 ( 2.5)
10.25
4.23
Urban stratua
Rural stratua
6.9
at 1 per 1040
at 1 per 2220
994
1383
9Z
144
9.76
10.41
4.02
4.30
Philippines
1981-83ab
Whole country
Saaple survey
38.9
RT23 ITU,
Mode-14,15
1982.5
2b- 4 ( 2.1)
2038
78
3.83
1.84
1975a
Main Island (Upolu)
Saaple survey
72.6
RT23 ITU, 10bb+
1975.6
3a- 7 ( 2.8)
1824
25
1.37
0.49
1966-68bc
Whole country
BCG Caapaign
RT23 ITU,
10idb+
1967.3
0- 9 ( 5.0)
1710
46
2.69
0.54
1961-63b
Whole country
All persons
RT23 ITU, 10bb+
1962.5
0- 9 ( 5.0)
1284
22
1.71
0.35
Sawoa
Syabols:
* Estiaated
.. Not applicable
... Not available
co
s. d)
11 J- i2l
ptfuf-.n pr
I
•nai Ltd
Tubercle and
Lung Disease
I*
Methodological issues in the estimation of the tuberculosis problem from
tuberculin surveys
I
I
H. L. Rieder
Tuberculosis Section of the International Union Against Tuberculosis and Lung Disease. Paris. France
L
I
5 U M M .4 R Y. Setting: National tuberculin skin test surveys.
Objectives: To review the operating characteristics of the tuberculin skin test, to ascertain the validity of
estimating prevalence and risk of infection from tuberculin skin test surveys under various conditions, and
to review constraints in the estimation of the magnitude of the tuberculosis problem in the community from
such surveys.
Methods: This report utilizes hypothetical and selected real data obtained in regional and national surveys at
various points in time to exemplify methodological issues.
Results: Risk of infection, the essence to be abstracted from tuberculin skin test surveys, theoretically allows
for a comparison of the extent of transmission of tubercle bacilli in various populations. However, the conduct
of tuberculin skin test surveys and the analysis and interpretation of their results are not free from important
technical problems. Accurate estimation of infection prevalence is particularly vulnerable to the great variability
of the test’s specificity under various circumstances. Furthermore, the annual risk of infection has averaging
characteristics that preclude a rapid assessment of changes in transmission patterns. Finally, estimates of infec
tion risk do not necessarily provide a standardized parameter to derive incidence of infectious cases, because
of variations in the quality of intervention and varying risks of progression from latent infection to overt
tuberculosis.
Conclusions: While tuberculin skin test surveys provide the currently most widely used means of assessing
tuberculosis transmission patterns over prolonged periods of time in a community, results from such surveys
must be interpreted with caution when accurate estimates of the tuberculosis problem are sought.
RESUME. Cadre: Enquetes tuberculiniques nationales.
Objet: Examen des caracteristiques operatoires du test tuberculinique cutane, evaluation de la validite du
calcul de la prevalence et du risque d’infection a partir des enquetes tuberculiniques sous diverses conditions,
et finalement revue des contraintes rencontrees dans le calcul de I’ampleur du probleme tuberculeux dans la
communaute a partir de telles enquetes.
Schema: Ce rapport utilise des donnees hypothetiques ainsi que des donnees reelles selectionnees a partir des
resultats des enquetes regionales et nationales effectuees a differentes epoques, afin d’illustrer certains
problemes methodologiques.
Resultats: Le risque d’infection, resultat essentiel des enquetes tuberculiniques, permet en theorie de comparer
I’importance de la transmission des bacilles tuberculeux dans differentes populations. Cependant la realisation
des enquetes tuberculiniques ainsi que I’analyse et 1’interpretation des resultats n’echappent pas a des
problemes techniques importants. Le calcul precis de la prevalence de 1’infection est particulierement soumis
aux variations importantes de la specificite du test dans differentes situations. De plus le risque annuel
d'infection reflete une ‘moyenne’ des evenements passes qui empeche toute evaluation rapide de modifications
dans les donnees de transmission. Finalement, les calculs du risque d’infection ne fournissent pas toujours un
parametre standardise dont on puisse en extraire 1’incidence des cas infectieux, en raison des variations liees a
la qualite de I’intervention et a la variabilite du risque de progression de 1’infection latente vers une
tuberculose maladie.
Conclusion: Bien que les enquetes tuberculiniques fournissent la methode la plus utilisee a I’heure actuelle
pour evaluer les caracteristiques de transmission de la tuberculose a travers de longues periodes dans une
I
r
f
r>
i
I
i
4
r
k
II
I
I
*
I
V
V
<
I,
I
Correspondence to: Dr Hans L. Rieder. Chief. Tuberculosis Section of
ihe IL ATLD. Reichenbachstr. 15. 3004 Bem. Switzerland.
r
Paper receded 11 January 1994. Final version accepted 9 August 1994.
I 14
t
(
------------------------------------------------------ —------- Mcthodologicul issues m the estimation of the tuberculosis problem Irom tubercul.n surveys
115
communaute, on devrait interpreter les resultats de telles enquetes avec prudence quand il s’agit d'obtenir une
estimation precise du probleme tuberculeux.
R E S U M E V. Marco de referenda: Encuestas tuberculinicas nacionales.
Objetivos: Revisar las caracteristicas operativas del test cutaneo de tuberculina, determinar la validez de la
estimacidn de la prevalencia y del riesgo de infection a partir de las encuestas tuberculinicas bajo diversas
condiciones. y revisar las dificultades en la estimacidn de la magnitud del problema de la tuberculosis en la
comunidad a partir de tales encuestas.
Metodo: Este informe utiliza dates hipoteticos y una selectionxie dates reales obtenidos de encuestas regionales
y nacionales en di versos lugares y periodos a fin de ejemplarizar los problemas metodologicos.
Resultados: El riesgo de infeccion, resultado esencial de las encuestas tuberculinicas, tedricamente permite la
compaction de la extension de la transmisidn del bacilo tuberculoso en las diversas poblaciones. Sin embargo,
la conduction de las encuestas tuberculinicas y el analisis e interpretation de sus resultados no estan exentos de
importantes problemas tecnicos.
La estimacidn precisa de la prevalencia de la infeccidn es particularmente vulnerable a la especificidad
del test bajo diversas condiciones. Ademas, el riesgo anual de infeccidn representa un promedio a un momento
dado. Io que impide una rapida evaluation de los cambios en los patrones de transmisidn. Einalmente, la
estimacidn del riesgo de infeccidn no proporciona necesariamente un parametro estandar para calcular
la incidencia de los casos infeccionsos, a causa de las variaciones de la calidad de la intervention y de la
variabilidad del riesgo de progresidn de la infeccidn latente a la enfermedad tuberculosa.
Conclusiones: A pesar que las encuestas tuberculinicas constituyen el medio mas ampliamente utilizado para
evaluar los patrones de transmisidn de la tuberculosis en periodos prolongados en una comunidad, los
resultados de cada encuesta deben ser analizados con precaution cuando se requiere una estimacidn precisa
del problema de la tuberculosis en una comunidad.
INTRODUCTION
TEST CHARACTERISTICS
The most desirable method of ascertaining the current
extent of transmission of tubercle bacilli in a society
would be to measure the incidence of infection with
Mycobacterium tuberculosis in susceptible persons.
Measuring the incidence of infection is, however, a
Herculean task. It requires repeat testing of a large
enough number of persons to identify with reasonable
precision the few who become newly infected over a
specified period of time. Funhermore. the incidence of
infection varies across various subpopulations, e.g. various
age groups in the same calendar year,1 and when esti
mates are based on repeat testing of the same individual
both boosting reactions2 and reversion of initially posi
tive tuberculin skin test reactions3 may greatly affect the
accuracy of the estimates. Some of these problems can
be partially overcome through approximation by calcu
lating the average probability that a person has become
infected over a specified period of time from a tuberculin
skin test prevalence survey in a population which has
not been subjected to previous tuberculin skin testing or
vaccination with BCG. Detailed technical guidelines on
how to conduct such a survey have been published.4-5
The purpose of this paper is to highlight methodolo
gical issues relevant to the analysis and interpretation of
tuberculin skin test survey data and their relation to
the tuberculosis problem in the community. Many of the
problems have been recognized for quite some time?
while other aspects challenge existing concepts and are
addressed here to further the discussion on how to criti
cally appraise the meaning of data generated in tubercu
lin skin test surveys.
Each test has its own intrinsic operating characteristics.
The sensitivity of the test is the proportion accurately
identified by a positive test result among persons with a
characteristic in question, while the specificity repre
sents the proportion with a negative test result among
persons without that characteristic.7 A multitude of tu
berculin skin test techniques has been developed in
this century, but the Mantoux technique is the most
quantifiable and. in conjunction with other information,
theoretically allows the determination of the sensitivity
and specificity of the test at different cut-off points.
In the early 1950s the World Health Organization
(WHO) collected information on tuberculin sensitivity
in over 3600 hospitalized tuberculosis patients in 10
different countries? The combined results closely fit a
normal distribution with a mode at 16-17 mm. The sen
sitivity in identifying infection with M. tuberculosis
calculated from this survey is 93% for an induration
> 10 mm and 78% for an induration > 14 mm. Similar
normal distributions were found in healthy persons in
areas with a very low frequency of small reactions6®-9
and among healthy United States Navy recruits who had
a history of exposure to tuberculosis.10 In the latter sur
vey, a mode was identified at 18-19 mm and the calcu
lated sensitivities for > 10 mm and > 14 mm were 94%
and 75% respectively (Table 1).
Thus, the sensitivity of the tuberculin skin test appears
to vary relatively little when different populations are
compared, except when cellular immunity is seriously
compromised. However, as the tuberculin surveys spon
sored by the WHO demonstrate? sensitization to envi-
116
Tubercle und Lung Disease
Table I. Sensitisity and specificity «>t the tuberculin skin test in
United States Navy recruits, using tuo commonly used cut-off points
as indicating the presence of infection icalculated from Tables I and
3)"
Specificity i%)
Sensitnity 'lTt
Test criterion
98 6
99.5
94.2
’4.9
> 10 mm
> 14 mm
false negative reactions to still obtain the right preva
lence. although individuals themselves are misdiagnosed.
Tuberculin prevalence surveys in numerous countries
have clearly shown that sensitization to environmental
mycobacteria is virtually absent in some, but abundant
in other countries, and that the relative magnitude of the
distribution due to environmental mycobacteria (and/or
sensitization due to BCG vaccination) and the magni
tude of the distribution due to tubercle bacilli may vary
considerably, even within the same country?9-18 Vacci
nation with BCG itself appears to induce cross-reactivity
with tuberculin PPD to a variable extent.I9-2J
Rust and Thomas have developed a model based on
tuberculin skin test data in 700 000 United States Navy
recruits.10 By asking about known history of exposure to
a tuberculosis case in the same household they separated
people into contacts and non-contacts. This allowed a
separation of the influences of reactions caused by envi
ronmental mycobacteria from those resulting from infec
tion with tubercle bacilli and the development of a
model predicting probabilities of tuberculous infection
at various cut-off points. These data thus allow the cal
culation of the specificity of the tuberculin skin test
using various cut-off points in the setting of the United
States (Table 1).
Even if it is assumed that the frequency of cross
sensitivity reactions is constant (which is clearly not the
case),6-91213 and therefore that the specificity of the
tuberculin skin test is unchanged and the information
from Rust and Thomas applicable, the predictive value
of a significant tuberculin test reaction (a/(a+b) in Fig 1)
would be extremely sensitive to variations in prevalence
of infection with tubercle bacilli (Table 2). If the inci-
ronmental mycobacteria apparently varies considerably
in different countries, making the specificity of the test
unpredictable, and thus the underlying normal distribution
attributable to tuberculous infection difficult to ascertain.
Figure 1 shows a hypothetical distribution of tubercu
lin skin test reaction sizes. Three distributions influence
the composite picture that can be ascertained: (1) the
distribution among persons without any mycobacterial
infection or with skin test anergy. usually not exceeding
a few millimeters (thick dotted line); (2) the distribution
among persons with tuberculous infection (thin dotted
line, shown here with a mode at 17 mm); and (3) the
distribution among persons sensitized to mycobacteria
other than tubercle bacilli who cross-react with tuber
culin PPD (dashed line shown here with a mode at
7 mm). It is apparent that in this example at a cut-off
point of 14 mm some true infections will be missed (c)
and some will be falsely counted (b). the latter because
they are attributable to non-specific sensitization. In fact,
as shown in this example, this observation pertains,
albeit to a different degree, to any cut-off point to indi
cate presence or absence of infection with M. tuberculosis
between 4 mm and 21 mm. It is thus apparent that cut
off points are compromises to balance false positive and
f
Tuberculous infection
100- I
Test
O)
c
30 o
1total
pos
a
b
ni
neg
c
d
n2
b+d
N
total
a+c
Teat criterion
14mm
60-
©
n
E
no
f
i,
co
(D
yes
I
b
b
b
f
40-
V
D
z
20-
0
0
5
10
15
20
25
30
Induration (mm)
Fig. I Example of the distribution of diameters in a tuberculin skin test survey. The distribution of non
specific reactions is shown under the dashed curve, the specific distribution under the dotted curve. The
hatched area, labelled b. indicates the number of reactions falsely counted as infection using a criterion of
> 14 mm to indicate infection, the hatched area, labelled c. indicates the infections missed by that
criterion. The number of positive reactions (n,) is found to the left of the test criterion; the number of
negative reactions (n.) is found to its right.
t
i
t
i.
i
Methodological issues m the eMimution of (he tuberculosn problem tn.m luberculin Mines
Table 2. Predictre \alue .>!' a positive teM using > 14 mm as
indicating injection, a^uming (from Table I) a 'pecit'icitv ot 98.6^
and a •.enMti\it> hr ^4 2'< Aith a cut-ott of > 10 mm. and specitlcity
and 'ensiti\it\ or
and '4.9f? respectively using a cut-off of
> 14 mm m a population with a prevalence of infection of 0.28% and
10.0' 'r re^pectivelv
Predictive value of a positive test result (%)
Pre\alence i'; i
0 28
10.0 ’
Criterion > 10 mm
15.9
SX 2
Criterion
29.6
94.3
> 14mm
dence of infection is assumed to be constant at 1% over
calendar time and infection risk independent of age,
then the expected true prevalence of infection with
M. tuberculosis in children aged 10.5 years will be 10%.
In an industrialized country, where the cunent risk of
infection might be 0.015%. and the decline in risk
of infection has been in the order of 10% per year, the
expected prevalence of infection in children of the same
age is 0.28%. Using these examples, in a country with a
low prevalence of infection, a cut-off point of 10 mm
indicating infection with .Vf, tuberculosis would falsely
.classify 84.1% of children thus identified as ‘infected’,
with a reduction in mis-classifications to only 70.4%
using a cut-off point of 14 mm and more. In a country
with a high infection prevalence, the test would falsely
classify only 11.8% or 5.7%, using 10 mm or 14 mm
induration respectively to indicate infection. It is thus
apparent that tuberculin skin test surveys in low infec
tion prevalence countries will almost always preclude
meaningful interpretation, unless environmental myco
bacteria are of such little imponance that the specificity
of the test approximates 100% even at low cut-off points.
It seems nevertheless that information may be useful in
countries with an elevated prevalence of infection.
The commonly used cut-off points of 10 mm and
14 mm. the latter customarily corrected by dividing by
0.82." take the sensitivity into account, assuming that
with a 10 mm cut-off point virtually all true infections
are included, but only about 82% with a cut-off point of
14 mm. The sensitivity of 82% for a cut-off of > 14 mm
to indicate infection is a slight overestimation, because
in the tuberculin survey in Tanganyika that produced
this figure it was assumed that all reactions > 10 mm
were 100% specific in'two areas with relatively few
cross-reactions, which was nevertheless not quite the
case." Applied to areas with any non-specific cross
reactions. these criteria have the disadvantage that they
do not account for the loss of specificity, a loss that
increases the further one moves from the mode of the
distribution from tuberculous infection towards the left
into increasing contamination by non-specific reactions.
Thus, the correction for sensitivity alone will invariably
overestimate the prevalence. This is shown in Figure 1.
where, to enumerate the infected among those with
>14 mm induration, instead of using (a+c). (n,)/[a/
(a+c)l or (n,1/10.82) is erroneously calculated by this
technique. Unfortunately, without a priori knowledge
ot both sensitivity and specificity, the proportion of mis
classifications can not be known.
Another technique to estimate prevalence assumes an
t
underlying normal distribution of reactions due to infec
tion with .Vf. tuberculosis. This mirror technique attempts
to identify the mode, multiplies the number of reactors
above the mode by two and adds the number of reactors
at the mode to arrive at the number of infected persons.
Although the mirror technique partially circumvents the
problem of test specificity, the precision of the estimate
is by necessity poorer and requires the testing of a
much larger number of persons to improve the estimate.
Funhermore. using the mirror technique approach, the
calculation of infection prevalence is very sensitive to
the selection of the location of the mode. If. for example,
in Figure 1 the mode is selected to be at 17 mm then the
calculated prevalence of infection is 13.7%: if the mode
is selected to be at 18 mm. the estimated prevalence of
infection is 10.9%. or 20% less.
If the sensitivity (denoted as .r) and the specificity
(denoted as y) are both known for a certain cut-off point
then the calculation of the prevalence is easily done.
There are four unknowns (a. b. c and d. see Fig. 1) that
can be solved with the four following equations, because
n, (denoted as the number with a positive test result, i.e.
a+b) and n: (denoted as the number with a negative
result, i.e., c+d) are defined:
a = n, - b
eq.(1)
b = d/y - d
eq.(2)
c = a/x - a
eq.(3)
d = n, - c
eq. (4)
These equations can be used to solve, e.g.. for a:
(n, + n,) xy - n.x
a =-------- ---------- —
(x + y - 1)
or any ocher cell.
EXAMPLES OF TUBERCULIN SKIN TEST
SURVEYS
A large tuberculin skin test survey is being carried out in
Tanzania under the auspices of the Tuberculosis Surveil
lance Research Unit of the International Union Against
Tuberculosis and Lung Disease. This survey is con
ducted in 5-year cycles; it encompasses the entire coun
try and is carried out by professional staff, trained by the
International Tuberculosis Surveillance Centre. Figures
2 and 3 show the survey data from 1991 from 3 regions
(chosen for convenience) combined (Dodoma, Mbeya.
and Morogoro) of children with and without BCG scar
respectively (reproduced with the permission of the
Tanzania National Tuberculosis/Leprosy Programme).25
Clearly, neither distribution allows rapid identification
of the proportion of the proponion infected with tubercle
bacilli.
In Figure 3. a mode is proposed (arbitrarily) at 18 mm,
allowing for construction of a mirror image of the sus
pected underlying distribution. Although the mirror
image technique accounts for the loss in sensitivity at this
118
Tubercle and Lung Disease
1500 -i
Q
co
co
1000 -
05
O
o
d
Z
500 -
o
o
0
“t-
0
10
20
30
Diameter (mm)
Fig. 2 Measured distribution of ruberculin skin test reaction size diameters in Tanzania 1991 (Dodoma.
Mbeya. and Morogoro regions) among children with BCG scar. Large circles are recorded data, filled
circles emphasize digit preferences at 10 mm. 15 mm and 20 mm. Points are values linearly adjusted for
digit preference (one neighbour only to each side). Line is fined through adjusted values by inverse
squared distance smoothing. Area under doned line indicates probable tme infection with Mycobacterium
tuberculosis. For better display, the number with 0 mm reaction (10 240 children) has been cut off.
(reproduced with the permission of the Tanzania National Tuberculosis/Leprosy Programme).-
500
400 co
co
05
300 -
o
6 200 Z
o
o
100 -
o
0
0
10
20
30
Diameter (mm)
Fig. 3 Measured distribution of tuberculin skin test reaction size diameters in Tanzania 1991
(Dodoma. Mbeya. and Morogoro regions) among children with no apparent BCG scar. Large circles are
recorded data, filled circles emphasize digit preferences at 10 mm. 15 mm and 20 mm. Points are values
linearly adjusted for digit preference (one neighbour only to each side). Line is fitted through adjusted
values by inverse squared distance smoothing. Area under dotted line indicates probable true infection
with M. tuberculosis. For better display, the number with 0 mm reaction (7210 children) has been cut
off. (reproduced with the permission of the Tanzania National Tuberculosis Leprosy Programme
diameter, the figure suggests that even at this diameter
some reactors are still likely to be falsely classified as
being infected. Here, the mode was assumed to be at
18 mm and the calculated infection prevalence was 7.9%.
Had a mode been selected at 17 mm. the prevalence
calculated with the same technique would be 9.8%. 24%
higher than with a mode at 18 mm.
Figures 2 and 3 also demonstrate that even experi
enced readers clearly have a preference for certain digits
(shown as full circles at 10 mm. 15 mm. and 20 mm).
It would thus be difficult to use a cut-off point of 20 mm
or an immediately neighboring value and account alge
braically for the loss of sensitivity for the sake of gaining
specificity.
0626
Methodological issues in (he estimation of the tuberculosis problem from tuberculm surveys
_______________________________
Figure 4 shows the results of two tuberculin surveys
in Korea conducted 25 years apart with the same tech
nique utilizing 1 TU PPD RT23 in children aged 0-9
years.26 It demonstrates that a large decrease in infection
prevalence fundamentally changes the interpretability of
a tuberculin survey in a country even if it has a relatively
small contribution from sensitization with environmental
mycobacteria. In 1965, the relative contribution of non
specific sensitization was negligible and did not pre
clude a clear dichotomization between those infected
with tubercle bacilli and those not. While the absolute
magnitude of non-specific reactions in 1990 was appar
ently similar to that in 1965. their relative contribution
to the overall distribution had become very important
by 1990. because true prevalence had declined to very
low levels, making it exceedingly difficult to separate
the infected from the non-infected. The ratio of reactions
attributable to environmental mycobacteria to those re
sulting from infection with M. tuberculosis had inverted
to an extent that the mode had shifted to the left (Fig. 4).
CALCULATING THE RISK OF INFECTION
FROM PREVALENCE DATA
Assuming that the prevalence of infection with M. tuber
culosis has been satisfactorily estimated, the essence
to be extracted from the data is the estimation of the
average annual risk of infection. The annual risk of
infection refers to a risk at a specified calendar time
b+x. where b indicates the calendar time at which the
cohort in the survey was bom and x is a number between
0 and a. where a is the age of the cohort at calendar time
b+a, the time when the survey was conducted. It cannot
be known at exactly what calendar time this risk existed
without inferences from serial surveys.1,27 Because the
risk may change over calendar time, x has been approxi-
mated to lie at the midpoint between the year the cohort
was bom and the year the survey was conducted, if data
from a single survey only are available:128
R.
= 1 - (1 -
where
denotes the annual risk of infection at the
midpoint in calendar time between the year the cohort
was borti and the year of the survey, and
the preva
lence of infection at the time of the survey, where both
risk and prevalence are expressed as fractions. Thus, if
the prevalence of infection among 10.5-year-old children
is found to be 10% at the midpoint of the survey
(assumed to be in. for example, the end of June 1993.
i.e. 1993.5) then the risk of infection is:
R 1988.25 = 1 - (1 -0.1),,,°5 = 0.010,
i.e. 1% at the approximated calendar time the end of
March 1988. The estimate of b+x can be improved from
the first approximation of b+a/2 only if serial surveys
are available.27 Serial estimates, only if closely fitting
a calculated regression, allow extrapolation to current
infection risk.
Two sequential surveys alone a few years apart will
not necessarily provide information on the change in
infection risk because of the problems associated with
the comparison of cross-sectional data across time. If it
is assumed that an earlier survey, e.g. conducted in
1988, had also provided an estimated average annual
risk of infection of 1% (approximated at calendar time
1983), then the conclusion is not necessarily warranted
that the risk of infection has remained unchanged over
calendar time up to the time of the second survey in
1993. It may well be that the risk decreased in the first
years after the birth of the second cohort, perhaps because
an efficient programme for identifying and curing cases
spreading infection was implemented, but subsequently
the number of infectious cases began to increase because
1965
1990
O)
= wco
CD
C
CD
O
5 -
O
CL
o 4o
T
“T
4
8
T
I
12
T
To
20
24
28
Induration (mm)
Fig. 4 Frequency distribution of tuberculin skin test reaction sizes among children aged under 10 years.
Korea 1965 and 1990. Reproduced with permission?1'
>
119
1
120
Tubercle and Lung Disease
of increasing prevalence of infection with the human
immunodeficiency virus (HIV) among persons with
tuberculous infection, leading to an increasing number
of transmitters and thus an increased risk of infection in
the community. It may just be that the net effect of an
initial decline, followed by an increase in infection risk,
resulted in no change in infection prevalence.
RISK OF INFECTION AND INFECTIOUS CASES
It is apparent that the rate of transmission of tubercle
bacilli is related to the number of sources of infection
in a society. The number of successful transmissions
by infectious cases to a susceptible population over a
defined period of time (usually one year) at a certain
calendar time defines the risk of infection in the commu
nity during that period. Utilizing the results from a
series of WHO sponsored surveys in low income coun
tries and the results of surveys in the Netherlands before
chemotherapy, Murray et al correlated annual risk of
infection to incidence of tuberculosis by linear regres
sion and obtained an estimate of approximately 50 inci
dent sputum smear-positive cases for each 1% of annual
risk of infection.29 This correlation must be interpreted
w'ith caution. The authors performed least squared re
gression and were thus unable to account for changes in
the variance with sample size. More importantly, the
authors assume that for areas where information on the
incidence was lacking, the incidence was half the preva
lence.29 This is. however, precisely the hypothesis that
has to be proven, if a relation between infection risk and
incidence rather than person-time of infectiousness is to
be demonstrated.
Infection risk is intrinsically coupled to duration of
undiagnosed, untreated transmissible tuberculosis, thus
with person-time of infectiousness in the community.
Intervention with chemotherapy has as its epidemiologi
cal aim to reduce the rate of transmission, and where
(his form of intervention encompasses effectively and
efficiently a large proponion of the population, the aver
age duration of infectiousness connecting prevalence
and incidence becomes fundamentally changed. Preva
lence of infectious tuberculosis thus might correlate
better with infection risk than incidence. Nevertheless, it
has been pointed out that in countries without a struc
tured programme, the number of infectious (sputum
smear-positive) patients remains essentially the same
after 2 years with or without such intervention, because
the main gain with such intervention lies with a reduc
tion of case fatality at the expense of keeping infectious
cases alive.'0 Conversely, in countries where interven
tion effectively cuts the chain of transmission, the number
ot transmissions caused by one case will be reduced.
Thus, to produce a 1% risk of infection, a larger number
of incident cases is required, because (he person-time of
infectiousness is reduced. This has been shown, for ex
ample. for the United States before HIV noticeably af
fected tuberculosis. In that country, extrapolation would
have required some 400 incident cases per 100 000
population to result in 1% risk of infection in the early
1980s,31 some 8 times the number predicted by the model
outlined above.29
Furthermore, the risk of tuberculosis following infec
tion with M. tuberculosis may vary in different popu
lations. It is certainly increased in persons with HIV
infection compared to immunocompetent hosts. Thus, the
epidemiological balance usually observed between host
and bacillus is no longer present under these circum
stances where each case of tuberculosis may produce
more than one new infectious case in the HIV-infected
segment of the population.
SUMMARY AND CONCLUSIONS
Theoretically, the incidence of infection is epidemiologically the most informative parameter, because it
identifies the extent of current transmission in the com
munity. It is usually not feasible to measure infection
incidence, and the derivation of the average annual risk
of infection from a tuberculin prevalence survey as a
proxy has become one of the most cherished tools in
tuberculosis epidemiology. Unfortunately, tuberculin skin
testing is fraught with problems of a technical nature,
including selection of standardized tuberculin, the tech
nique of administration, and reading of the test result.
Even if all of these barriers are overcome, it is in many
circumstances exceedingly difficult to arrive at an esti
mate of the prevalence of infection. Sensitization to
environmental mycobacteria and M. bovis BCG results
in cross-reactions with the standard tuberculin. The
higher this sensitization is and the lower the prevalence
of true infection with tubercle bacilli, the more difficult
it becomes to disentangle the truth from confounding
factors. It is clear that techniques must be developed to
address the problems of interpretation that arise in so
many countries which have completed a tuberculin skin
test survey. It remains to be determined whether simulta
neous testing with different antigens, algebraic manipu
lation or other approaches can help io overcome some
of the apparent shortcomings of tuberculin skin test
surveys. Because sensitivity is already largely known for
various cut-off points, it appears that the most promising
approach would lie w'ith an attempt to determine the
specificity of the tuberculin test at a given cut-off point
in a country planning a tuberculin skin test survey using
different antigens.14-1* This would help greatly in im
proving the validity of the results of a tuberculin skin
test survey.
The calculation of the risk of infection (should the
determination of infection prevalence be successful)
from a single or even two sequential surveys provides
only information on the extent of transmission at some
point in the past, determined by the age of the children
that have been tested. The tool is not sensitive to short
term changes, because of its ‘averaging’ characteristic.
The knowledge of risk of infection cannot precise!}
Methodological issues in the estimation of the tuberculosis problem from tuberculin survevs
121
provide information on the number of expected incident
cases of tuberculosis; it can only state to what extent
such cases are capable of transmitting tubercle bacilli
within the community, which is a function of the number
of infectious cases, the number of case-contact interac
tions. the duration of infectivity, and characteristics of
exposure.
The determination of the risk of infection has never
theless been regularly used to compare the extent of
the tuberculosis problem in various populations. It is. if
technically interpretable, the only available means of
measuring the extent of transmission that has occurred,
on average, over specified periods of time in the past.
The common underlying technical problems will often
undermine the precision in estimating the size of the
tuberculosis problem in a community from a single
survey. It can be useful for global estimates of the level
of the tuberculosis problem in a community and of trends
over relatively long periods of time. The observation of
trends in prevalence or risk of infection over time is by
far more informative than a single survey, because the
change in slope might be freer from bias than the level
of the intercept. A consistent recession of age prevalence
curves, as observed for example in the 6 five-yearly
Korean prevalence surveys,26 is enough to convince that
the risk of infection has been declining. However, any
formal estimation of a change in the risk of infection
within a certain precision may be impossible to assure.
To move forward in gaining a better understanding
of the intricacies in estimating the tuberculosis problem
in a community from tuberculin skin test survey data
requires a concened effort on the part of researchers in
the field to address the various issues and problems
encountered in the conduct of such surveys and the
interpretation of their results.
8. WHO Tuberculosis Research Office. Funher studies of
geographic variation in naturally acquired tuberculin sensitivitv.
Bull World Health Organ 1955: 22: 63-83.
9. Roelsgaard E. Iversen E. Blocher C. Tuberculosis in tropical
Africa. An epidemiological study. Bull World Health Orcan
1964: 30: 459-518.
10. Rust P. Thomas J. A method for estimating the prevalence of
tuberculous infection. Am J Epidemiol 1975; 101: 311-322.
11. World Health Organization. Tuberculosis survey in Tanganyika.
Copenhagen: WHO Research Office. 1958.
12. ten Dam H G. Surveillance of tuberculosis by means of
tuberculin surveys. Geneva: World Health Organization. WHO/
TB/85.145. 1985.
13. Edwards L B. Acquaviva F A. Livesay V T. Cross F W. Palmer
E E. An atlas of sensitivity to tuberculin. PPD-B. and
histoplasmin in the United Stales. Am Rev Respir Dis 1969:
99: 1-132.
14. Lind A. Larsson L O. Bentzon M W et al. Sensitivity to sensitins
and tuberculin in Swedish children. 1. A study of schoolchildren
in an urban area. Tubercle 1991: 72: 29-36.
15. Larsson L O. Skoogh B E. Bentzon M W et al. Sensitivity to
sensitins and tuberculin in Swedish children. 2. A study of
preschool children. Tubercle 1991: 72: 37-42.
16. Larsson L O. Skoogh B E. Bentzon M W. Magnusson M.
Olofson J. Lind A. Sensitivity to sensitins and tuberculin in
Swedish children. 2. Sequential versus simultaneous skin testing.
Tubercle 1991; 72: 187-189.
17. Larsson L O. Magnusson M. Skoogh B E. Lind A. Sensitivity to
sensitins and tuberculin in Swedish children. 4. The influence of
BCG vaccination. Eur Respir J 1992: 5: 584-586.
18. Larsson L O. Bentzon M W. Lind et al. Sensitivity to sensitins
and tuberculin in Swedish children. 5. A study of school children
in an inland rural area. Tubercle Lung Dis 1993; 74: 371-376.
19. Abrahams E W. Tuberculin hypersensitivity following BCG
vaccination in Brisbane school children. Tubercle 1979:
109-113.
20. Tuberculosis Prevention Trial. Madras. Trial of BCG vaccines in
south India for tuberculosis prevention. Indian J Med Res 1980:
72 (suppl): 1-74.
21. Bahr G M, Chugh T D. Behbehani K et al. Unexpected
findings amongst the skin test responses to mycobacteria of
BCG vaccinated Kuwaiti school children. Tubercle 1987;
68:105-112.
22. Menzies R. Vissandjee B. Effect of Bacille Calmette-Guerin
vaccination on tuberculin reactivitv. Am Res Respir Dis 1992;
145:621-625.
23. Menzies R. Vissandjee B. Amyot D. Factors associated with
tuberculin reactivity among the foreign-bom in Montreal. Am
Rev Respir Dis 1992; 146:"752-756.
Acknowledgments
24. Styblo K. The first round of the National Tuberculin Survey in
Tanzania. 1983-1987. Tuberculosis Surveillance Research Unit.
Progress Report 1989. Volume 2. Paris: International Union
Against Tuberculosis and Lung Disease. 1989: pp 101-1 16.
25. International Union Against Tuberculosis and Lung Disease.
Tanzania National Tuberculosis/Leprosy Programme. Progress
Report No. 27. Paris. Apnl 1992.
26. Hong Y P. Kim S J. Kwon D W. Chang S C. Lew W J. Han
Y C. The sixth nationwide tuberculosis survey in Korea. 1990.
Tubercle Lung Dis 1993; 74: 323-331.
27. Styblo K. Meijer J. Sutherland I. The transmission of tubercle
bacilli. Its trend in a human population. Tuberculosis
Surveillance Research Unit Repon No. I. Bull Int Union Tuberc
1969; 42: 1-104:
28. Cauthen G M. Pio A. ten Dam H G. Annual risk of infection.
World Health Organization. WHOCTB/SS. 154. 1988.
29. Murray C J L. Styblo K. Rouillon A. Tuberculosis in developing
countnes: burden, intervention and cost. Bull Int Union Tuberc
Lung Dis 1990: 65: 6-24.
30. Grzybowski S. Enarson D A. The fate of cases of pulmonary
tuberculosis under \anous treatment programmes Bull Int Union
Tuberc 1978: 53: 70-75.
31. Cauthen G M. Rieder H L. Geiter L J A model of the relation
between age-specific prevalence of tuberculous infection and
incidence of infectious tuberculous, implications for screening
policies. Tuberculosis Surveillance Research Unit of the
IUATLD. Progress Report 1991. Volume I The Hague: Royal
Netherlands Tuberculosis Association. IM9I pp 1-20.
Numerous people have provided input into this article. Particularly
acknowledged are the comments made by Nancy Binkin. George M.
Cauthen. Donald A. Enarson. Lawrence J. Geiter. and Dixie E. Snider. Jr.
References
1. Sutherland I. Recent studies in the epidemiology of tuberculosis,
based on the risk of being infected with tubercle bacilli. Adv
Tuber Res 1976; 19: 1-63.
2. Narain R. Interpretation of the repeat tuberculin test. Tubercle
1968:49:92-103.
3. Aronson J D. The fluctuation of the tuberculin reaction in
different geographic areas and its relationship to resistance. Am
Rev Tuberc 1951: 63: 121-139.
4. Deck F. Guld J. Committee on Epidemiology and Statistics. The
WHO tuberculin test. Bull Int Union Tuberc 1964; 34: 53-70.
5. Bleiker M A. Sutherland I. Styblo K. ten Dam H G.
Misljenovic O. Guidelines for estimating the risks of tuberculous
infection from tuberculin test results in a representative sample
of children. Bull Int Union Tuberc Lung Dis 1989: 64; 7-12.
6 Edwards P Q. Edwards L B. Story of the tuberculin test. From an
epidemiologic viewpoint. Am Rev Respir Dis I960: 81: 1-47.
7. Last J M. A dictionary of epidemiology. 2nd edn. New York:
Oxford University Press. 1988.
Re-examining the Annual Risk of Infection as a Monitoring Tool
Christopher JL Murray
Associate Professor of International Health Economics
Harvard School of Public Health
9 Bow Street
Cambridge, MA 02138
i
9/27/96
Draft Not for Quotation
r
Tuberculosis has now been widely appreciated as one of the most important causes of ill health and
mortality in the developing world. WHO is spearheading a global effort to combat tuberculosis using
short-course chemotherapy which has been adopted by an increasing range of countries. Monitoring the
epidemiological trends in tuberculosis has become increasingly essential. While tuberculosis kills more
than 2 million a year, the clinical disease is relatively uncommon in the population. Typical developing
countries experience incidence rates of 100 cases per 100,000 per year. Measuring the prevalence of such
an uncommon disease requires large sample sizes which are perforce costly. Even if the effort is made to
survey prevalence, programs require information or estimates of incidence to monitor the evolution of the
disease because treatment programs can radically alter the relationship between prevalence and incidence.
While disease is rare, infection is extremely common. WHO estimates that 32.8 percent of the world is
infected (Sudre et al. 1992). Skin testing with PPD has been providing information on the prevalence of
infection for nearly 90 years. The interpretation of these prevalences, however, was limited (Sutherland
1976). In the 1960’s, the Tuberculosis Surveillance Research Unit (TSRU) refined methods of estimating
the annual risk of becoming infected with tuberculosis based on the prevalence of past infection from skin
test data (Styblo et al. 1969). This major breakthrough in tuberculosis epidemiology has become the
mainstay of epidemiological monitoring in developing countries and Europe (Styblo 1976, ten Dam 1985,
Bleiker 1991). Recently, a number of individuals and events have challenged the dominance of ARI as a
tool for monitoring the epidemiology of tuberculosis (e.g.Reider 1995).
The objective of this paper is to review the basis for the TSRU method of estimating the ARJ and the
relationship between ARI and other epidemiological indices of tuberculosis. The method of estimating the
ARI from PPD skin test prevalence data and estimating the incidence of tuberculosis on the basis of the
ARI is founded on four key premises. First, one can distinguish the infected from the non-infected at the
population level using induration size. Second, when the ARI is changing, it declines or increases
exponentially overtime. Third, the ARI does not vary with age or at least does not vary over the ages 0-14.
Fourth, there is a close relationship between the ARI and incidence of smear-positive tuberculosis. In this
paper, we will re-examine the evidence supporting each of these four premises, explore some new findings
and speculate on the impact of other changes on this approach to monitoring tuberculosis.
Distinguishing Infected from Non-infected Using PPD Skin Tests
The ARI method depends critically on the ability to estimate the prevalence of past infection in a given
population using PPD skin test results. Measuring the prevalence of past infection is complicated by
several issues: skin test reversion, the sensitivity and specificity of different cutoffs of induration size used
to define infection, and the impact of BCG on skin test interpretation. Each of these issues will be
addressed in turn.
Sutherland (1971; 1976) argued that if skin test positives will revert to being skin test negative at a low rate,
1-2% per year, this would not have a significant effect on the estimation of the risk of infection based on
the prevalence of positive skin test reactions. Felten and Van Der Merwe (1989) showed that children with
skin test reactions greater than 15mm had an average decrease in induration size of 3.6mm on the second
test given an average of 262 days later. Such variability on skin test response may mean that reversion
rates could be substantially higher than 1-2% per year. If skin test reversion occurs at higher rates then
risk estimated from the prevalence of positive skin tests is likely to be underestimated.
For any test, the starting point for discussing the test should be its sensitivity and specificity. Sensitivity of
PPD skin tests for detecting infection with M.tuberculosis has been studied by examining charactericits of
the test in patients with clinical tuberculosis who are therefore knwon to be infected. Depending on what
dose of PPD and type that is used and where it is given, those infected with M.tuberculosis develop an
induration that is normally distributed with a mode between 14 and 20mm. The sensitivity of the PPD skin
2
9/27/96
Draft Not for Quotation
test for detecting M.tuberculosis infection, therefore, depends on the induration size used to define
infection. Table 1 summarizes the sensitivity of a 5TU dose of PPD in ten populations included in a
collaborative WHO study in the 1950s (WHO Tuberculosis Research Unit 1955) and two other large
studies (Edwards and Edwards 1960; Palmer and Bates 1952)using an induration cutoff of 10mm.
In population surveys where exposure to atypicals is low, there is usually a mode somwhere between 14
and 20 mm representing the infected population (Nyboe ). Figure 1 illustrates the distribution of
induration size in Korea in the 1965 prevalence survey for those that are BCG scar-negative. There is a
clear mode at 18 and antimode at 9-11mm. In such cases, the proportion of the population with an
induration greater than the antimode will be a reasonable estimate of the prevalence of infection, even
without making an estimation of the sensitivity of the antimode as a cutoff because the small number of
false negatives are likely to be balanced by a small number of false positives.
In environments such as the South of the United States where individuals are exposed to atypical
mycobacteria, many individuals develop a non-specific reaction to PPD which is larger than the small
reaction that some have to the trauma of the Mantoux test. Unfortunately, the induration size of this non
specific reaction may overlap with the induration size of those with a specific reaction due to past exposure
to M.tuberculosis. Figure 2 shows the distribution of induration size for white male Navy recruits resident
in the United States throughout their life aged 17-21 during the years 1961 to 1968. In this population, the
mode at 4 mm is due to non-specific reactions to PPD and there is no clear mode above 10 mm due to
specific reactions. In essence, this population is a mixture of three groups, those who are uninfected with
’ M.tuberculosis or atypicals, those exposed to atypical mycobacteria and those infected with M.tuberculosis.
The problem of determining the prevalence of infection in a population made up of these three groups is
much more difficult because the proportion with an induration size above a certain cutoff will be a function
of 5 factors: the sensitivity of the cutoff to detecting tuberculosis infection, the sensitivity of the cutoff to
detecting past exposure to atypical mycobacteria, the specificity of the cutoff for those uninfected and
unexposed to aytpical mycobacteria, the prevalence of exposure to atypicals and the prevalence of
M.tuberculosis infection. Even if the sensitivities and specificities were known, the results of the test
cannot be used to simoultaneously estimate the prevalence of atypical exposure and the prevalence of
infection without supplemental information.
Four alternatives have been proposed to try and estimate the prevalence of infection even in settings where
the prevalence of exposure to atypicals is moderate to high and/or BCG coverage is high (see below) so
that false positives may exceed false negatives when using an arbitrary cutoff. Rust and Thomas (1975)
developed a method to estimate the proportion of the Navy recruits skin tested from 1958-1969 that are
infected making use of the information collected on a history of a known contact with a case if
tuberculosis. In their method, they take advantage of having two.populations, contacts and non-contacts,
where the proportion of infected and non-infected are likely to be quite different but the prevalence of
exposure to aytpicals is likely to be the same to estimate the prevalence of infection.1 The only critical
1 Rust and Thomas use the following equations to find the infection prevalences among contacts and
noncontacts:
p
p
3
9/27/96
Draft Not for Quotation
assumption required for their method is the induration size above which everyone is infected with
tuberculosis which they assume is 22 mm. Table 2 shows the results of applying their method to data for
different ethnic groups using induration sizes above which everyone is presumed infected from 16 to
26mm. Estimated prevalence ranges from 2.2% to 1.8% for this range of induration sizes in this
population. The method proposed by Rust and Thomas could be generalized to any situation where two
groups in whom the prevalence of atypical exposure will be similar but the prevalence of M.tuberculosis
infection will differ. Such information whether by contact status or some other differentiating variable
such as socio-economic status is at present rarely available in most skin test surveys.
Second, various investigators have tried to test individuals with other sensitins, PPD like products derived
from other mycobacteria (e.g. Edwards et al. 1969). These efforts, however, have not been successful at
increasing the ability to distinguish true infection with M.tuberculosis with non-specific reactions due to
aytpical mycobacteria (ten Dam 1985).
Third, other efforts at using second PPD tests with or without concomitant BCG have been proposed and
applied in in few situations (Raj Narain et al. 1966, Lotte et al. 1971, Liard et al 1989, ten Dam and Hitze
1980). While these methods are interesting they require repeat surveys at an interval between 2 months to
12 months and thus are much more costly to deploy. These methods have not been widely applied.
Fourth, Bleiker et al. (1989) suggest that the mode of those with a specific reaction will lie between 14 and
18mm. As the distribution of those with a specific reaction to PPD are often normally distributed, a
reasonable estimate of the prevalence of infection can be obtained by doubling the proportion with an
induration over the expected mode of those with a specific reaction. In the USA Navy recruit data, using
18mm instead of 16mm as a cutoff, lowers prevalence for white resident male Navy recruits 1961-1968
from 3.13% to 2.18%. Using results from 30 PPD skin test surveys. Figure 3 compares the prevalence of
infection estimated using the two times mode method with the antimode method in the age-group 0-14. In
this dataset, the two times mode method gives higher estimates of prevalence than the antimode method.
This results may be because the mode of the specific reactions may not be the median of specific reactions
because of small sample sizes.
where
P =
prevalence of infection in contacts,
P' =
prevalence of infection in noncontacts,
ft = proportion of contacts with zero reaction,
fo = proportion of noncontacts with zero reaction.
and
/n
=
A =
proportion of contacts in category n,
proportion of noncontacts in category n,
for reaction size category n in which ail individuals are assumed to be infected.
The combined prevalence for contacts and noncontacts is found by taking the weighted average of P and
P
4
9/27/96
Draft Not for Quotation
In communities where the mode of the specific reactions to PPD cannot be derived from the distribution of
induration size, estimated prevalence will be sensitive to the arbitrarily chosen ‘mode’ (14-20 mm) used.
On the other hand, when serial observations are made of prevalence are made in the same population, it
may be safe to assume that the mode is stable. It can be shown with simulations (not shown here) that the
estimated time trend in the ARJ will not be very sensitive to the arbitrary mode chosen for doubling.
The impact of BCG on the Estimation of the Prevalence of Infection
Because BCG immunization can cause an induration with PPD that is difficult to distinguish from
M.tuberculosis infection, the rising global prevalence of BCG - UNICEF estimates coverage exceeds 80%
in all regions— poses a serious challenge to using PPD skin surveys to estimate the ARI. Faced with a
large proportion of children that are BCG immunized, there are two strategies for estimating the prevalence
of infecteds. The first is to calculate prevalence only in those children that are BCG scar-negative. This
method suffers from four problems. First not all children given BCG may develop a scar. Second, BCG
scars have been found to wane; the National Tuberculosis Insititure of India (1992) found that of 49 3-4
year olds found to have a BCG scar, 11 or 22.4% had no scar on examination 21.6 months later. Thus,
with rising BCG vaccination rates, the BCG scar-negative group may include increasing numbers that have
received BCG and thus are more likely to have an induration greater than some arbitrary cutoff. It remains
to be seen if scar waning rates are this high in other populations. Third, as BCG coverage has risen over
90% in many countries, there is an ever growing likelihood that the BCG scar-negative children are not
representative of the entire population. Fourth, as BCG coverage increases the number of children that
must be screened to find a large enough sample to test will increase raising the costs of surveys. Despite
these difficulties, the method remains the standard approach.
The effect of BCG on subsequent PPD skin tests is a complex function of the age at which BCG was
received, the number of times the individual has been PPD tested in the past, the time since immunization,
the strain of BCG used and other factors (Al-Kassimmi et al. 1991;Capewell and Leitch 1986; Friedland
1990; Guled et al. 1968; Joncas et al. 1975; Menzies and Bissandjee 1992; Shaaban et al. 1990; Snider
1985). Most of these studies show that the effect on skin test positivity can be exceedingly hard to predict
in settings where BCG is given over a wide range of years and individuals are often PPD tested. On the
other hand, Menzies and Vissandjee (1992) show that in Quebec when BCG is given at birth or at least
before 1 and individuals have not been PPD tested before, the effect of BCG on induration size is
negligible after 7 years or less. Because the strain of BCG used in Quebec is local, it is unclear if this result
can be generalized. Despite extensive study, it is difficult at this point to generalize on how skin test
reactions in the BCG scar positive population can be meaninfully interpreted.
The restesting methods discussed above can also be used to try and distinguish true infection from BCG
induced induration but they have not been widely applied. Another possibility is the recent developmental
work has yielded new antigens that may be allow one to distinguish M.tuberculosis infection from BCG
infection. MPB64 is one of the more promising that is now being tested (Nagai et al. 1991, Li et al. 1993).
The development of new or improved skin tests with greater specificity would allay many concerns about
the interpretation of skin test data.
Review of ARI Equations
Styblo et al. (1969) developed a model whereby prevalence of infection in a given age group could be used
to calculate the ARI. In the technical appendix, these equations are re-derived in detail and only the
summary results presented in this section. Styblo et al.(1969) argued based on logic and empirical
observations of the prevalence of infection in Dutch military recruits that the ARI changes exponentially
overtime. To further simplify their model, they assumed that ARI is not a function of age, thus:
X(r) = Re"
(equ. 1)
5
9/27/96
Draft Not for Quotation
where R is the annual risk at time period zero and r is the rate of decline in the annual risk of infection. The
notation used here is not the same as used by Styblo et al. (1969). It has been changed to make the
meaning of the parameters more intuitively clear. In addition, the derivation of the equations presented in
this paper does not depend on the their approximation that X is equivalent to — ln( 1 — X).
Solving the differential equation, the prevalence of infection at each age a is:
7ier'U '■‘'-•J]
P(a)= \-e
, for r
0
(equ 2)
The equation can be rearranged to get the ARI as a function of prevalence:
r * ln(l - /’(a))
X=
e-™-l
,
for r
0
(equ 3)
When r is zero, the equations are much simpler:
P{a) = l-e-Ra
Equ.4
and
X=
-ln(l- P(g))
Equ5.
a
When two surveys are available, we can directly estimate r. Based on equation 3, one can show that:
In
lnS(g,f,)
lnS(a,f2)
r=
equ 6
where tl is the time of the first survey and t2 is the time of the second survey. S(a,tl) is the prevalence of
susceptibles at age a at time tl.
The basic equations for ARI as a function of prevalence of infection at a given age and time are based on
the assumption that ARI is uniform across ages in any given time period. Direct observations of skin test
conversion rates have mostly shown that the ARI increases with age at least in the first two decades of life
(Narain eta. 1966, Nyboe and Christenson 1966, Olakowski 1972) Sutherland and Payers (1975) explored
the possible relationship between ARI and age. They assumed that the ARI might rise exponentially with
age such that:
X(r)
Rer,+va
equ 7
where v is the rate at which the ARI increases with age. Using data from 7 communities where more than
one skin test survey had been undertaken, they used to non-linear regression methods to estimate v from
the data. Their estimates for the percent increase in ARI with each year of age varied from 1% to 18%.
6
9/27/96
Draft Not for Quotation
Confidence intervals for these estimates, however, were not estimated so that they were not able to reject
the null hypothesis that there is no relationship between age and ARI.. Sutherland and Payer also
experimented with other forms of the relationship between ARI and age but found they gave similar fits to
the exponential curve.
While we do not know if the ARI is definitely a function of age, we can derive equations for the
relationship between prevalence of infection and the annual risk of infection for a given v. Based on
equation 7, we can derive formulas for prevalence and ARI with the approach used to derive equations 3
and 4:
(r + v)ln(l- P(a))
e-a{r+v)-\
,
for r
0
equ 8
where R and r are as in equations 1-4, and v is the same as in equation 5. This equation will be used below
in exploring the statistical properties of different estimators of ARI.
Estimating the ARI from Survey Data
Estimating ARI from a Single Survey
There are three distinct analytical situations that we discuss seperately: one survey, two surveys and three
or more surveys. When the results of only one skin test survey are available, the difficulty of estimating
the ARI is estimating the annual trend in ARI or r. There are three options for estimating r. First, we can
assume that is is zero and use equation 5. In effect, we will be calculating the average risk of infection
over a number of years prior to the survey depending on the age-group examined. For example, if the ARI
is calculated based on the prevalence in the population 0-14 assuming no trend in ARI, it will be an
approximation of the ARI 3-5 years before the survey (Table 3). Second, we can try to estimate r based
on circumstantial evidence such as the performance of the tuberculosis control program or trends in the
ARI measured in neighboring or similar communities. While this method may be appropriate in certain
situations, it can always be challenged as non-objective. Third, we can impute r from the pattern of
prevalence in different age-groups - a method first proposed by Mori (1971). To impute r, first calculate
the ARI for the age-groups 0-4, 5-9 and 10-14 using equation 5. If r is negative so that ARI is declining,
the estimated ARI using equation 5 will be higher for the older age-groups. Estimates of r can be derived
by comparing the estimated ARI from equation 5 for two different age-groups. For example, the formula
based on comparing the age-groups 0-4 and 5-9 for r is:
4
-ln<7?75-9 )
25~
equ. 9
where 2.5 represents the average difference in exposure duration between the two age-groups. A more ■
robust estimate of r can be derived from averaging estimates of r based on pairwise comparisons of 0-4 and
5-9, 5-9 and 10-14, and 0-4 and 10-14.
These three methods for estimating r can be combined with using data on the prevalence of infection from
the age-groups 0-4, 5-9, 10-14 or 15-19. Using younger age-groups has the advantage that prevalence is a
7
9/27/96
Draft Not for Quotation
function of recent risk and is less sensitive to time trends in risk. Using older age-groups or pooled data
from several age-groups increases the number of positives and reduces the confidence interval around
estimates of prevalence of infection and therefore the confidence interval around the annual risk of
infection.
To investigate the bias and efficiency of different estimators of ARI we have used a simulation approach.
We have identified the three leading candidate methods for calculating risk of infection: calculate ARI
using the prevalence in the age-group 0-4 and assume no trend in ARI (0-4 Method), calculate the ARI
using the average prevalence in the age group 0-14 and assume no trend in ARI (0-14 Method) and
calculate the ARI by first estimating r from a pairwise comparison of the ARI with no time trend calculated
from 0-4 with 5-9, 5-9 and 10-14 and 0-4 and 10-14, then using this estimated r calculate ARI using the
prevalence 0-14 and equation 3 (Imputed r Method). To evaluate these three methods, we have first
specified the true level and trend in risk. Then we have simulated surveys of different sample sizes
ranging from 250 to 1000 children per age group and applied the three methods for estimating ARI.
Figure 4 illustrates the bias and efficiency of the different methods. The 0-4 Method gives estimates with
very wide confidence intervals because of the small number of positives that will be present in that agegroup but the estimates are only milldy biased, the 0-14 Method gives estimates with very narrow
confidence intervals because many more observations are included but the estimates can be very biased.
The degree of bias is greater if the time trend in risk is greater. Finally, the Imputed r Method is essentially
unbiased but gives unacceptably high confidence intervals.
Estimating ARI from Two Surveys
When two sequential cross-sectional surveys are available, r can be estimated directly from the data.
Using equation 6, there are two estimators. Method A would be to calculate r using the data for 0-14 for
each survey assuming an average age of 7.5 for the interval and Method B would be to first calculate three
estimates of r one for 0-4, one for 5-9 and one for 10-14 and then average the three estimates of r.
Simulations summarized in Figure 5 show that Method A is unbiased and more efficient - in other words,
the confidence interval for r is narrower. Of note, Bleiker et al. (1992) recommend calculating r from two
surveys by estimating the annual rate of change in prevalence. Estimates of r using this method are
inaccurate; for plausible ranges of ARI (0.5-3.0%) and trends in ARI (-0.5 to -6.0%), the estimates will be
3-12% too low.Once r has been estimated from the prevalences 0-14 in the two surveys, ARI can be
calculated using equation 3. Estimates of ARI for the second survey using this method are unbiased and
efficient.
Estimating ARI from Three or More Surveys
When more than two surveys are available, more powerful statistical techniques can be used to estimate r
or ARI directly. With non-linear regression software, a maximum liklihood approach can be used to fit R
and r from equation 1 to the available data using a number of solution methods that are available in
commercial software. If such sophisticated programs are not available, an unbiased estimate of r can be
derived from the coefficient in an ordinary least squares regression of the form:
Ln(-ln(S(a,t)) = rt + C
equ 10
where a is particular age at which the prevalence is measured at each survey. This relationship can be
derived by rearranging equation 3 and noting that overtime holding the age examined constant the complex
expression C will not change only the term rt. Once r is estimated from this equation. ARI can be estimated
for each year using equation 3.
8
9/27/96
Draft Not for Quotation
ARI, Incidence and Prevalence
Styblo (1985) postulated that there was an approximate relationship between ARI and the incidence of
smear-positive tuberculosis such that an ARI of 1% mapped to an incidence of smear-positive tuberculosis
of 50/100,000. This relationship is widely used by many developing countries to estimate case incidence.
In turn these estimates are often used in the denominator to evaluate the performance of the case detection
system. Estimated incidence is also frequently used in planning for case load when a program is being
expanded into a new area or made freely available in an existing area.' ARI based estimates of incidence
are also the basis for the current global estimates of tuberculosis incidence and mortality (Murray et al.
1989, Sudre et al., Raviglione et al etc.). The dominance of ARI as a monitoring tool is closely linked to
this putative relationship between incidence of infection and the incidence of disease.
Styblo derived this relationship using data from three types of situations: first, where ARI and incidence
had been measured, second, where ARI and detected cases are known in communities with high case
detection and third where ARI and prevalence were known. He used the latter data sources by postulating
that prevalence equals 2 times incidence because of an extensive literature showing a case-fatality rate of
tuberculosis of 50% (Styblo 1991). It is interesting to note that the relationship between ARI and
prevalence has also been used to estimate the infection parameter, K, or the number of infections caused
by an infectious source in one year (Styblo 1991). They are all approximately related as follows:
— = p = id
K
equ. 11
where X is ARI, K is the infection parameter, I is incidence and p is prevalence. The relationship shown
should only hold true at the population level if all parameters are constant over age and the population is
not growing. Often cited rules of thumb illustrate the basic relationships: an ARI of 1% and a K of 10 are
consistent with a prevalence of 100 cases of smear-positive tuberculosis per 100,000 which in turn are
consistent with an incidence of 50 cases of smear-positive tuberculosis per 100,000 and a duration of 2
years.
We can rewrite equation 11, to give a formula for incidence as a function of the other parameters:
X
K d
equ 12.
Clearly, if K is roughly constant we must expect the relationship between ARI and incidence to change if
average duration changes. The widespread use of treatment has substantially altered the average duration
of treatment. In developed countries, prompt diagnosis and treatment will have reduced prevalence to be
substantially less than 1 year while in some Asian countries poor treatment may have lowered mortality
and raised the duration of smear-positive cases to be 3-5 years. Table 4 gives the possible range of
incidence rates consistent with a 1% ARI for plausible ranges of the transmission parameter and duration.
Table 5 summarizes the three types of data used by Styblo including his observations and a variety of other
studies. The only two true incidence studies, Chingleput and Bangalore, found an incidence rate of 51 and
53 per 1% risk. In both of these studies, however, smear-positives were not distinguished from culture
positive smear-negatives. Styblo had to assume that in both studies 50% of bacilliary cases were smear
positive. Studies based on case incidence data show that the relationship between ARI and incidence was
37 per 1% risk in the Netherlands 1951-1976 but 175 per 1% risk in the USA 1961-1969. The latter
finding fits with the fact that average duration in the United States is likely to be quite short. The direct
studies of risk and incidence are not convincing for a strong relationship between risk and incidence of the
nature proposed by Styblo.
9
9/27/96
Draft Not for Quotation
Sty bio K. The epidemiology of tuberculosis. The Hague: KNCV, 1992
Styblo K. Surveillance of tuberculosis. Int J Epidem 5( I ):63-68, 1976.
Sudre P, ten Dam G, Kochi A. (1992). Tuberculosis: a global overview of the situation today. Bull WHO
70(2): 149-159.
Sudre P, ten Dam G, Kochi A. Tuberculosis: a global overview of the situation today. Bulletin WHO
70(2): 149-159, 1992.
Sutherland I. Recent studies in the epidemiology of tuberculosis, based on the risk of being infected with
tubercle bacilli. Adv Tuberc Res 19:1-63, 1976.
Sutherland I. The effect of tuberculin reversion upon the estimate of the annual risk of tuberculosis
infection. Bull IUATLD45: 115-118, 1971.
Sutherland I, Payers PM. The association of the risk of tuberculous infection with age. Bull IUAT
50(l):70-81, 1975.
ten Dam HG. Surveillance of tuberculosis by means of tuberculin surveys. Geneva: WHO TB/85.145,
1985.
ten Dam HG, Hitze KL Determining the prevalence of tuberculosis infection in populations with non
specific sensitivity. Bull WHO 58(3):475-483, 1980.
World Health Organization. Regional Office for Africa. A tuberculosis survey in Liberia, 1961.
World Health Organization. Regional Office for Africa. A tuberculosis case-finding programme on Mahe
Island in the Seychelles, 1962.
World Health Organization. Regional Office for Africa. A tuberculosis survey in Ibadan, Nigeria, 1964.
World Health Organization. Regional Office for Africa. A tuberculosis survey in Mozambique, 1962.
World Health Organization. Regional Office for Africa. A tuberculosis survey in Kenya, 1961.
World Health Organization. Regional Office for the Eastern Mediterranean. Assignement report: A
national tuberculosis prevalence survey in the Socialist People’s Libyan Arab jamahiriyah, 1978.
World Health Organization. Regional Office for the Mediterranean. Epidemiological and Statistical
Centre. Tuberculosis Survey in Cyrenaica, Libya, 1961.
World Health Organization. Regional Office for the Mediterranean. Epidemiological and Statistical
Centre. Tuberculosis Survey in Iraq, 1962.
World Health Organization. Regional Office for the Western Pacific. Report on tuberculosis prevalence
survey in Cambodia (1967-1968), 1969
World Health Organization Tuberculosis Research Office. Further studies of geographic variation in
naturally acquired tuberculin sensitivity. Bull WHO 22:63-83, 1955.
14
9/27/96
Draft Not for Quotation
World Health Organization. Tuberculosis Research Office. A tuberculosis survey in Tanganikya, 1958.
World Health Organization. Tuberculosis Research Office. Tuberculosis surveys in Ghana. 1958.
World Health Organization. Tuberculosis Research Office. Tuberculosis survey in Ibadan, Nigeria, 1958.
World Health Organization. Tuberculosis Research Office. Tuberculosis survey in Nigeria, 1957.
World Health Organization. Tuberculosis Research Office. Tuberculosis survey in Uganda, 1959.
World Health Organization. Tuberculosis Research Office. Tuberculosis Survey in Basutoland,
Bechuanaland, and Swaziland, 1958.
World Health Organization. UNICEF-Assisted Tuberculosis Project. Tuberculosis survey of Kiambu,
1960.
World Health Organization. UNICEF-Assisted Tuberculosis Project. Tuberculosis survey of Machakos,
1960.
15
9/27/96
Draft Not for Quotation
Technical Appendix
The basic approach is general to other areas in infectious disease modeling where the relationship between
the proportion of the population susceptible to new infection at each age and the annual risk of infection is:
8S(a,n
6a
8S(q,0
= -X(a,/)S(a,r)
bt
where S(a,t) is the prevalence of susceptible at age a and time t, X(a,t) is the annual risk of infection or
more precisely the instantaneous risk of infection denominated as an annual rate which is a function of time
and age. This formula simply states that the population that is susceptible at each age a to infection
changes with age as a function of the risk of infection. This equation is general, it does not specify the
relationship between the ARI and time or age.
Stybio et al assume that ARI is declining exponentially overtime:
X(O = Re"
Thus:
85(a,Q
8a
- Re"
Time and age are related such that::
t = b 4- a
where t is time, b is year of birth and a is age. This simply means that the current year is equal to the year
of birth for a cohort plus their current age. The formula for the annual risk of infection can then be
rewritten:
R er(^o)
so then we can rewrite the equation for rate of change of susceptibles with age to be:
dS(a)
da
= _Re^-) S(a)
In addition, we know that at birth everyone is susceptible so that S[0]= 1. By integrating both sides of this
equation, and solving for the constant of integration using S[0]= 1 and finally substituting t-a for b, the
proportion of susceptibles at any age and time becomes:
S(a) = erL
16
9/27/96
Draft Not for Quotation
800 -
600 -
52 400 ■3"
200 •-
.Illllll.
0 —
2
4
6
8
10
14 16 18
Induation size (mm)
12
20
22
Figure L
Frequency distributions of tuberculin reactions,
Basutoland, Bechuanaland, and Swaziland, 1958.
WHO Tuberculosis Research Office. Copenhagen. (April 1958).
24
26+
2
5
u
Ilin.
10 12
14
16
18
Induration size (mm)
20
22
24
26+
Figure 2.
Frequency distributions of tuberculin reactions, U.S. Navy
Recruits, 1961-1968.
Rust and Thomas (1975)
50
40
O
o
8
5
*
30 -
o
o
2^
S
20 -
oo<>
(N
$
10 -
0 0
10
20
AM Prevalence (%)
30
40
Figure 3.
Prevalence of infection, ages 0-14, calculatedfrom skin test
results from twenty-three tuberculosis surveys, using two different methods of
determining prevalence. The x-axis shows prevalence calculatedfrom all
reactions above a determined induration size. The y-axis shows prevalence
calculated by doubling the number ofreactions above the mode of the
distribution of induration sizes. The tuberculosis surveys were conducted in:
Basutoland (1956-57), Bechuanaland (1956-57), Cambodia (1967-68),
Ghana (1957), Ibadan, Nigeria (1957-58 and 1962-63), Iraq (1961), Kenya (19
Kiambu, Kenya (1959), Korea (1965 and 80), Liberia (1959-60), Cyrenaica, Li
(1959), Libya (1976-77), Machakos, Kenya (1959), Mahe Island, Seychelles (19
Mozambique (1961), Nigeria (1955-56), Pakistan (1987-88), Swaziland (1956-5
Tanginikya (1957), Tumkur, South India (1960-61), and Uganda (1958).
Imputed r Method
0-14 Method
0-4 Method
2.5 .
2.5 .
2-5 ,
I
I
2 j
: £
2 15
1
i
2
I
g
g I
I
■u
2 15 j
2 1.5
1I *-*03
‘ I
E
i W
u
c5
1
0.5
0.5
I
0
0
-I
-3
-5
True r (%), (sample size)
. .............................. ..
11
I
E
tn
UJ
tn
W
1
i “
<
-I
0 '
-3
-5
-7
True r (%), (sample size)
_ ................ .................... .......................... ....................
-I
.
-
I- -
— ■
-3
-5
-7
True r (%), (sample size)
- ..
Figure 4. Bias and efficiency ofA KI estimates, with true ARI of 1%, based on three different methods of calculating ARI. The first graph shows
the estimate calculatedfor ages 0-4. The second graph shows calculations ofARIfor ages 0-14. The third graph incorporates the trend over lime,
using an imputed value for r. Each diagram shows the results from twelve different simulations, in which the true value of r and the sample size
were varied. Each r-value along the x-axis was tested with sample sizes of250, 500 and 1000 (not shown on axis).
B
5 -
5 -
0 ------
1
LU
I
-10
-15
LU
-1
0- -
£
-------------3
-5
-7
True r (%), (sample size)
-10
-15 ‘--------------------------------------------------.
-1
-3
-5
-7
True r (%), (sample size)
D
2.5
2.5 —
2-
2 -
1.5 -
2 15
I
I3 °-5 L
■3
1 -
2
L L
1 -
. . 1 °-5 -
0--------------------------------------- x----------- :
-1
-3
-5
-7
True r (%), (sample size)
0 i-------------—
-1
-3
-5
-7
True r (%), (sample size)
Figure 5. Graphs show bias and efficiency of estimates of r and ARI based on two
different methods of estimating r. Graph A shows estimates ofr calculatedfor
ages 0-14. Graph B shows estimates ofr calculated as an average of the r-values f
age groups 0-4, 5-9, and 10-14. Graph C shows the ARI estimate using the 0-14
method of calculating r, with a true ARI of 1%. Graph D shows the ARI estimate u
the average r. Each graph presents the results of twelve different simulations, usin
varying true values ofr and varying sample sizes. Each of the r-values along the
x-axis were testedfor sample sizes of250, 500, and 1000 (not shown on axis).
?
1500
O
I
y
y, 1000
5,
E
I
1
i
o
o
o
500 -
0 ---0
o
o
o
0
o
o 0°
o
o<>
o <? o o o
o
o o
o
4
2
o
6
ARI (%)
Figure 6.
ARI and prevalence ofsmear-positive TB, from thirty tuberculosis
surveys: Basutoland (1956-57), Bechuanaland (1956-57), Cambodia (1967-68),
Chingleput, India (1968), Ghana (1957), Ibadan, Nigeria (1957-58 and 1962-6
Iraq (1961), Kenya (1958-59), Kiambu, Kenya (1959), Korea (1965, 70, 75, 80,
90), Liberia (1959-60), Cyrenaica, Libya (1959), Libya (1976-77), Machakos, K
(1959), Mahe Island, Seychelles (1962), Mozambique (1961), Nigeria (1955-56)
Pakistan (1987-88), Philippines (1981-83), Swaziland (1956-57), Tanginikya (1
Transkei, South Africa (1977), Tumkur, South India (1960-61), and Uganda (19
Table 1.
Sensitivity of tuberculin skin testing using varying
induration sizes to define infection.
Induration (mm)
10
12
8
5TU PPP *
Denmark
England
North India
South India
Pakistan
Phillipines
South USA - Negro
South USA - White
Sudan
Vietnam
96.0
100.0
97.6
88.2
98.3
95.1
98.0
98.1
99.4
90.1
99.4
95.9
76.9
97.6
94.3
96.0
96.3
87.1
98.1
92.4
64.6
91.7
91.7
91.0
90.4
98.1
94.7
96.0
98.1
95.3
HURT
Oresund (Denmark)
Kanchrapara (India)
Mehalla (Egypt)
Glen Lake (USA)
90.6
93.8
97.3
89.6
84.1
86.4
93.2
81.0
70.6
1TUPPD-S f
Charity Lake (USA) - white
Charity Lake (USA) - negro
Battley (USA) - white
Battley (USA) - negro
90.9
90.0
85.9
94.0
72.7
83.1
78.0
87.6
53.5
71.5
63.8
75.4
t
WHO Tuberculosis Research Office, 1955.
t Palmer and Bates, 1952
75.7
85.1
72.0
Table 2.
Prevalence ofTB Infection among U. S. Navy Recruits, Ages 17-21. 1960-1969.
Two methods of calculating prevalence of infection are compared. The Rust and Thomas
method is shown using varying values of n, the category in which all individuals are assumed
to be infected. The doubling method is shown for three different modes.
15
17
Rust and Thomas
19
21
23
25
15
Doubling
17
19
White
Nonwhite
2.53
10.88
2.38
9.34
2.35
8.31
2.37
7.35
2.25
7.23
2.58
7.35
3.42
18.69
2.46
13.69
1.59
9.13
ALL RACE
3.26
3.04
2.96
2.94
2.84
3.21
4.52
3.27
2.13
Table 3. The number ofyears prior to the survey at which point the annual risk of infection equals
the average risk of infection calculatedfrom the age group 0-14 using Equation 5.
' ARI at the
Time of the
Survey
-1%
-3%
Trend in ARI
-5%
-7%
-9%
1.0%
3.67
4.62
4.9
5.08
5.23
2.0%
2.27
4.05
4.47
4.7
4.86
Note: The number of years prior to the survey at which point the average risk calculated for the
age-group 0-14 equals ARI was estimated assuming that within the age-group 0-14, the population
was equal at each age. The average prevalence 0-14 was calculated by taking the definite
integral of equation 2. The ARI estimated using this average prevalence and Equ 5 was then used
to calculate the years prior to the survey when this average annual risk equaled annual risk.
Table 4. Incidence oj Smear-Postiive Tuberculosis Expected When the Annual Risk of Infection Equals 1%.
Transmission
Parameter
0.25
0.5
0.75
1
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
800
667
571
500
444
400
364
333
308
286
267
250
235
222
211
200
400
333
286
250
222
200
182
167
154
143
133
125
118
Hl
105
100
267
222
190
167
148
133
121
Hi
103
95
89
83
78
74
70
67
200
167
143
125
111
100
91
83
77
71
67
63
59
56
53
50
Average Duration of a Smear-Positive Case (Years)
1.25
1.5
1.75
2
2.25
160
133
114
100
89
80
73
67
62
57
53
50
47
44
42
40
133
111
95
83
74
67
61
56
51
48
44
42
39
37
35
33
114
95
82
71
63
57
52
48
44
41
38
36
34
32
30
29
100
83
71
63
56
50
45
42
38
36
33
31
29
28
26
25
89
74
63
56
49
44
40
37
34
32
30
28
26
25
23
22
2.5
2.75
3
3.25
3.5
80
67
57
50
44
40
36
33
31
29
27
25
24
22
21
20
73
61
52
45
40
36
33
30
28
26
24
23
21
20
19
18
67
56
48
42
37
33
30
28
26
24
22
21
20
19
18
17
62
51
44
38
34
31
28
26
24
22
21
19
18
17
16
15
57
48
41
36
32
29
26
24
22
20
19
18
17
16
15
14
7 .
57®
World Health Organization
Global Tuberculosis Programme
Report
of the Fourth Meeting of the
Technical Research and Advisory Committee (TRAC)
(Geneva, 2-3 May 1996)
(
The meeting opened with a welcoming overview by Dr. Ralph Henderson (Assistant
Director General). The Chairman of the Co-ordination Advisory and Review Committee
(CARG), Dr. Elzinga, took the chair for the first items of business (agenda attached).
He welcomed the sixteen members of TRAC, who had been selected by a working
sub-group of CARG to represent, in their personal capacities, the range of disciplines relevant
to public policy development for tuberculosis control. In view of the differing backgrounds
and experiences of members a briefing day was held immediately prior to the meeting to
ensure that all members shared a common core knowledge of the principles and mechanisms
of TB control and the structure and functioning of GTB.
<
Dr Elzinga reviewed the recommendations coming from the CARG 1995 meeting and
the environment which CARG 1995 had set for better functioning of the GTB (chart attached)
including adoption of a business-like structure and specific follow up of recommendations.
He linked this to the need for changes in the terms of reference of the TRAC. Specifically.
TRAC was to focus on technical review and advice to the Programme and to the CARG on
the soundness of the Programme’s TB control work and research activities. Functions of
TRAC which had previously duplicated the functions of CARG, in particular the review of
the Programme’s budget and non-technical or non-scientific policy matters, should be
reserved to the CARG. TRAC members reviewed the previous terms of reference for TR.AC
and made several suggestions. These will be incorporated in the overall terms of reference
for the Programme’s advisory bodies and reviewed within WHO and by the CARG.
As the last item of business under Dr. Elzinga’s chair, a chairperson for TRAC was
elected by secret ballot of TRAC members in their first session. Upon election. Dr. Jaap
Broekmans took the chair.
The secretariat explained that the agenda for this meeting was guided by the
recommendations made by CARG 1995. It was agreed that TRAC members would be able
in future to contribute to setting the agenda, either through recommendation to CARG. or by
suggesting items to be included in future meetings.
Presentation offirst agenda zVem. Monitoring and Surveillance, The monitoring and
surveillance project was developed in response to the request of CARG for better information
on the state of the global epidemic and the performance of control programmes in order to
m:\tub\trac\iracrep.doc 3/10/96 8:55
monitor the progress towards achievement of the year 2000 targets. In the six months since
the project’s inception, the data collection system has been strengthened, and interpretation
methods improved. A preliminary analysis of country control programmes showed that 20
countries (11%) are implementing WHO control strategies routinely throughout, while
another 62 (33%) have accepted the WHO strategy and partially implemented it.
Discussion. TRAC members supported the approach that had been developed, and
advocated that the project should be sufficiently resourced and staffed to ensure its routine
functioning. The proposed new data collection forms were endorsed, but several suggestions
were made for possible additions, e.g. outcome of smear negative cases, data on HIV
prevalence, mortality, and country-specific surveillance projects. These additions will be
considered for the data collection forms to be used in subsequent years.
(
The issue of developing new and better ways of estimating incidence stimulated
considerable debate. A practical way to estimate incidence is needed if progress towards the
target of 70% case detection is to be more accurately monitored. However, no consensus
emerged on the method to use and questions on the appropriateness of the target emerged. It
was suggested that, given the high costs and opportunity costs of attempting to measure
incidence by approaches such as prevalence surveys, resources would be better used to
improve effectiveness of TB control programmes.
Recognizing that this will still leave room for confusion, and noting that TRAC needs
to provide advice to the CARG on the year 2000 goals, TRAC members agreed to establish a
task force together with the secretariat. The task force will assist the Programme to determine
whether, with existing knowledge and diagnostic tools, new methods of measurement of
incidence were feasible to develop and, if yes, whether they would be worth the cost and
effort of their application. The task force should also provide advice to the Programme on
whether the case detection target remains appropriate. If not, the task force should
recommend specific alternative targets and provide advice on their use. both within and
outside the WHO system. It was agreed that the task force will complete its draft report in
time for the 1996 CARG meeting.
From a different perspective, however, there was a need for better country-based data
for estimates of the global burden, and the impact of control measures on it. as well as for
effective advocacy and informed debate about the global situation.
A TRAC member noted that the GTB still continues to have two different estimates of
TB deaths and cases circulating in various publications. (Note, this is true within the WHO
system for most causes of death and morbidity - one estimate representing bottom-up
country-specific, cumulative results, without any global constraint on total number of deaths;
the other being the result of the Global Burden of Disease exercise which constrains total
deaths and then apportions their cause. Until the overall policy within WHO is clarified, the
GTB will have to continue to use the two sets of numbers for different purposes. It was noted
that groups outside WHO were developing other estimates and this would add to confusion
among policy makers.
To avoid further confusion in proliferation of estimates TRAC suggested that the
Programme should, for use during the rest of the decade, make a new, estimate
2
m:\tub\trac\tracrep.doc 3/10/96 8:55
incidence and mortality based on the best, currently available, country-by-country
information. The Programme will do this.
National Programme Support Activities,
(
Presentation of agenda item. Overview, GTB’s National Programme Support
Activities. The Programme offers technical support and advice to countries. These activities
are concentrated upon those countries which have the highest case load, and the greatest need
for support. The focus of this work is gradually shifting from Geneva towards a Programme
presence with tuberculosis control advisers in the regions and in high priority countries. In
addition, the Programme is also involved in the development and co-ordination of training,
preparation of guidelines for good practice and development of strategies to prevent drug
resistance. The main problem presently faced by the Programme is the demand for assistance
to an increasing number of countries in an accelerated expansion of the DOTS strategy. The
Programme recognised the need for inclusion of many more partners (non-governmental
organisations, private corporations and private practitioners and their associations) in
adoption and implementation of the DOTS strategy. This required new thinking within the
Programme (to become more client-oriented) and new products and knowledge (more userfriendly guidelines and materials, new forms of drug products which could be simpler to
administer and help to avoid malpractice) and a new approach to expanding the
knowledgeable manpower to deal with TB at all levels. Overall these country support
activities had to help create a public environment favourable to all segments of the health
sector, public and private, to work effectively to control tuberculosis.
The Programme had chosen three topics related to country support to present to
TRAC for its reaction and advice at this meeting, based on CARG’s 1995 recommendations
and areas of interest: revision of the WHO treatment guidelines, rational and more effective
use of tuberculosis drugs and human resource development strategy.
Discussion, In overview discussions TRAC members noted that only 20 countries
had adopted the DOTS strategy fully, and that progress in many others was slow, due to the
difficulty of changing policy, public perceptions and the allocation of adequate resources in
member countries. The strategy therefore proposed by the secretariat of focusing and
decentralising technical assistance to priority countries was endorsed by TRAC members.
They also agreed that quicker expansion of DOTS coverage and maintenance of high cure
rates could be only achieved by involving other partners such as NGOs and the private sector.
One TRAC participant raised a question concerning the priority of smear negative TB
within the overall DOTS strategy. In discussion, another TRAC participant focused on the
increased incidence of smear negative TB among HIV positive persons and asked for
clarification of the DOTS strategy in such settings.
The programme staff clarified that the DOTS strategy was fully articulated in the
document entitled '’Framework for Effective Tuberculosis Control”. Good treatment and
rigorous recording and reponing of the process and the outcomes for all forms of TB was part
of the DOTS strategy, as set out in that document. Some countries gave priority in their
programmes to applying the DOTS approach mainly to smear positive cases. The Global
Tuberculosis Programme’s technical advice and policy orientation was to advocate the more
comprehensive approach as outlined in its policy documents.
- 3 m:\tub\trac\tracrcp.doc 3/10/96 8:55
Presentation 1; Treatment guidelines were published in 1993 (Treatment of
Tuberculosis: Guidelines for National Programmes}^ these were reviewed by a working
group in October 1995. Some of the recommendations of this meeting are being incorporated
into the second edition of the treatment guidelines that will be published in 1996. The new
edition includes more specific case definitions and refinements to the suggested regimens and
thus will help to provide clarity compared to the 1993 edition.
Discussion, One TRAC participant expressed concern about some aspects of the
new guidelines, specifically that some regimens in the continuation phase had not been
proven in clinical trials, another participant noted his worry that regimens advocated for
smear negative cases, especially in HIV prevalent settings, may provide inadequate
treatment.
The Programme noted that the evidence upon which the new guidelines were based
reflected practical experience at country level rather than outcomes of clinical trials. The
comments made by various TRAC members were understandable and accurate in the
context of their particular country and regional experiences, but did not reflect experience
and conditions in other countries and regions. To help TRAC members to gain the
alternative perspectives to provide better advice to the Programme the secretariat undertook
to invite selected members to join in work in countries/regions with which they were less
familiar. The Programme fully agreed on the need for further research on regimens and on a
practical response to the management of smear negative patients in HIV prevalent settings.
It proposed a correspondence group among key TRAC members (knowledgeable about
treatment regimens) to explore these issues further and advise the secretariat appropriately.
The group was established subsequent to the meeting.
Presentation 2. Rational and more effective use of tuberculosis drugs. In response to
CARG guidance, the Programme had succeeded to establish closer co-operation with the
Division of Drug Management and Policies (DMP) in late 1995 and to interest them in the
problems of drug resistance caused by inappropriate use of anti-TB drugs. As a consequence,
the Programme and DMP succeeded to have the WHO Expert Committee on the use of
essential drugs (December 1995) agree to inclusion of combined, fixed dose forms of anti-TB
drugs (double and triple combinations) in the Essential Drugs List. More widespread use of
these drug forms would not only have eventual impact on efficiency of TB programmes and
emergence of drug resistance, but would also begin to alleviate the risks posed by
thiacetazone.
Nonetheless, resistance to the main antituberculosis drugs was seen to be a growing
problem that will compromise the efficacy of TB programmes, and result in even more deaths
from the disease. Good programme management is essential to prevent resistance, but as so
much TB care is carried out in the private sector and in other institutions operating outside the
national TB programmes, the Programme recognised that other innovations are required.
Fixed dose combination drugs can simplify the administration of effective TB treatment and
reduce the risk of developing secondary drug resistance. Such products have the potential to
unify private sector and non-govemmental organisations’ treatment practices if these products
are seen to be the treatment of choice. Ideally, development of a quadruple, combination
product (pill containing all four drugs) for TB treatment would make this approach extremely
attractive.
4
m:\tub\trac\tracrcp.doc 3/10/96 8:55
However, the bioavailability of rifampicin in some combination products now
available in the market is low (and this problem would be equally important in eventual
development of a quadruple combination). GTB thus recognises the need to develop a
system for monitoring of product quality, to promote the use of good quality products and
ultimately establish systems by which TB patients, physicians and national TB programmes
can safely use combination products. Once such a system is working, fixed dose
combinations are likely to become accepted as the treatment of choice in both private and
public sectors and more easily supported in bilateral and multilateral financed projects.
Discussion. TRAC endorsed the importance and feasibility of this approach noting
that where good TB programmes with high cure rates already exist, combination products
need not necessarily be introduced or promoted. TRAC also noted that there will be a limited
continuing use of individual drug dosage forms to meet the needs of some patients. TRAC
agreed that for countries where good TB treatment was not established, or could not easily be
maintained because of health system weaknesses or poor co-ordination with the private or
voluntary sectors, combinations products (especially quadruple) offered a major step forward.
Discussion yielded support for the directions already taken by the Programme in six
areas in order:
1. to finalise and promulgate standardised quality test methodologies;
2. to develop and promulgate specification for combination products suitable for use
in public tenders;
3. to specify the added value of combination products which could be used for the
purpose of deciding when the (higher) prices of such products were a better buy
than single product forms;
4. to develop, in consultation with the World Bank and other development assistance
agencies, the conditions of tender by which bidders would be obliged to include in
their sales price the fees to operate the standardised testing procedure for each
batch delivered;
5. to develop, promulgate and maintain a list of laboratories which could reliably and
independently perform such testing;
6. to advocate private sector development of quadruple combination products as soon
as possible.
The Programme will report on further progress at the 1996 CARG meeting.
Presentation 3. Training and Human Resources. Despite much progress in training
national and district level TB programme staff in the last three years, the Programme
recognizes that additional trained, appropriately skilled people at all levels are needed if the
DOTS strategy is to be adopted and effective tuberculosis control achieved by the year 2000
in much of the world. In many programmes that are in the expansion phase of DOTS,
shortages of qualified staff are perceived to be a major constraint. In order to address these
needs, training materials and normative guidelines have been developed and disseminated by
the Programme, and training courses have been established. These are so far limited in focus
to public sector programmes and do not take into account the needs of non-governmental
organisations or private sector practitioners. The Programme recognised the need to
substantially broaden its efforts but had uncertainties about strategy, techniques and channels.
5
m:\tub\lrac\iracrcp.doc 3/10/96 8:55
Discussion. The Programme was advised to address training needs in their broadest sense,
including not only staff involved in TB control activities, but private physicians and general
health workers, lay people and policy makers. The scarce resources provided by developing
country governments and external donors to invest in human resources in connection with TB
was perceived as a major problem, as was the relatively low appeal to health professionals of
involvement in tuberculosis control. The Programme undertook to discuss the topic more
widely within WHO and to seek outside advice as the first steps in preparing a better
descriptive analysis of the specific issues and alternatives to be considered, which might then
lead to practical next steps to discuss at the CARG 1996 meeting or at the next TRAC
meeting..
Tuberculosis Research Strategy and Activities
Presentations 1 and 2. The research programme of GTB has undergone major
restructuring since 1995, to develop a more comprehensive, unified and less narrowly
biomedical approach, that complements the overall Programme strategy. While links are
being developed with other agencies and institutions, WHO is concentrating its efforts in
areas where it has a comparative advantage.
The three main areas of research are the following:
To develop methods to increase the dissemination of the DOTS strategy
1.
(DOTS More Widely);
To improve the efficiency of the DOTS strategy by developing and testing
2.
innovations (DOTS More Easily);
To promote the development and new tools to facilitate TB control and to
bring TB elimination nearer (Beyond DOTS).
Research was prioritised according to the magnitude of the benefit that would result if the
research were successful, the certainty that it would achieve such a result, and the duration of
the research and its costs. The DOTS strategy is already averting some of the future burden
of death and disease due to TB. Expanding the coverage and efficiency of the DOTS strategy
could avert the majority of the remaining burden, but new tools would be essential for the
complete control of TB.
Discussion. TRAC supported the concept that the future of the Programme depends
on good research and good guidelines for research, but requested that the Programme
articulate better the match between the needs of tuberculosis control programmes and the
research strategy which GTB has adopted a clearer statement of why research is needed.
Based on discussion, four hypotheses (aside from lack of financial resources and
inappropriate TB control strategies) were offered by the TRAC chair as to why better global
TB control was not being achieved: first, current programmes are failing because of a lack of
human resources (insufficient people addressing the problem, and insufficient skills in those
that are) secondly, NTPs are currently failing because patients bypass them and go directly to
the private sector; thirdly, some societies appear willing to tolerate a high burden of disease
from tuberculosis. Lastly, many governmental systems of TB control are dysfunctional.
TRAC Members felt the Programme did not have a sufficiently clear picture of the
demand for TB care, and particularly of the profile of TB patients. The financial, physical
- 6 m:\tub\lrac\tracrcp.doc 3/10/96 8:55
and social barriers affecting access to care needed to be more clearly described. Institutions
that already address TB research, especially those in the public sector, need to be in a
dialogue with GTB so that the Programme is aware of all forms of TB research in progress,
and of the networks of research that exist in both biomedical and operational spheres. To do
this the Proaramme needs to make its own research strategy and efforts better known, to
acquire additional staff with strong leadership experience in global research initiatives and
excellent networking skills.
The chairman summarised the following four main deficiencies in high quality TB
control throughout the world, together with the research needed to address them:
1.
Insufficient countries are implementing the DOTS strategy, and the WHO year
2000 targets may therefore not be achievable if the current pace continues.
Research should develop methods for increasing the number of countries
accepting the DOTS strategy.
2.
Operational problems exist even in countries accepting the DOTS strategy,
and operational research should be aimed at resolving these.
Even in countries accepting the DOTS strategy and succeeding with it, there is
a need to increase the coverage of the population. Research should develop
methodologies, such as community-based management of TB, to address this
issue.
4.
Better tools are needed which can only be obtained from research into new
diagnostic agents, new drugs and new vaccines. The wide range of these
activities emphasises the need for prioritisation.
Discussions of this summary emphasised that although the DOTS strategy was only
launched in 1995 to encapsulate the more complex approach outlined in the WHO
■■Framework for TB Control”, it needed much more marketing and dissemination. Adoption
of and support for the DOTS strategy at the regional and country levels required acceleration
and expansion. The need to develop innovative and sustainable methods, particularly with
the private sector involved in delivery of patient care, was emphasised. With regards to
prioritisation of research topics, TRAC suggested that GTB should clearly identify priorities
both for the world at large, and for GTB within and between these four major areas of
research and reporton this to CARG 1996. In addition, GTB should collect information and
experience on working with the private sector, and develop prototypes for ways in which the
DOTS strategy can be adopted by the sector.
The meeting agreed on the necessity to continue to state clearly to the outside world
that DOTS is the best current strategy for TB control and that it works, whilst internally
acknowledging the potential deficiencies of the DOTS strategy and implementing a research
agenda aimed at improving it. This approach will entail considerable tension in the
Programme and staff will be needed who can effectively question established practice, whilst
managers will need to encourage and support both those who question, and those who defend
current strategies.
7
m:'tub\trac\tracrep.doc 3/10/96 8:55
Presentation 3. Community based management of tuberculosis. In many countries
the primary health care system is unable to cope with the burden of TB or to deliver DOTS
effectively. However, existing community structures, such as community-based care
organisations, may be involved in TB care and, in some areas, already are. The aim of this
project is to explore how TB management can be better integrated with community based
organisations to deliver the DOTS strategy effectively. This approach will be piloted through
the community care network for HIV patients in Africa.
Discussion. TRAC members supported the operational research project, but
suggested that the best chances of success are where well developed community networks
already exist and national tuberculosis programmes are well established. It was observed that
the project should still depend upon the health system for diagnosis of patients and for
supervision of community workers. Concern was expressed about the sustainability of such
initiatives, (due to the external support currently provided to most community projects), and
their limited population coverage of community projects relative to the national populations
affected by TB.
Presentation 4. Elimination of TB - concept development. Health systems and
management research concepts were presented to provide a framework for strategic thinking
aimed at the ultimate goal of TB elimination.
Discussion. TRAC warned that, given the current poor state of global TB control,
excessive efforts in this direction would be premature. There is also a risk that the impact of
economic analysis of the benefits of elimination may not be great and may not be credible in
the current environment. Nevertheless, the concept of elimination remains a relevant
eventual Programme pursuit because of the need to a) establish a long-term vision, and b)
use the concept to change the donor environment in order to increase TB research funding.
Presentation 5. Elimination of TB - tools required, The development of new tools
for diagnosis, treatment and prevention is necessary to overcome limitations in the
application, efficiency and sustainability of the current TB control strategy. The Programme
proposed a "‘diagnostics initiative” based on a situation analysis and the establishment of a
coalition (WHO/public and private sector/donors) aimed at developing low cost, easily
applied, robust diagnostic tests. The Programme would manage the co-ordination of
multicentre clinical trials of the new long-acting rifamycin, rifapentine, and a renewed
dialogue with the research community and pharmaceutical companies to stimulate the
development of new drugs. The Programme would also advocate funding for targeted and
more fundamental research aimed at, for example, new vaccine development. Finally, GTB
will collaborate with UNAIDS in the testing and formulation of recommendations for
preventive therapy among HIV-infected people.
Discussion. With respect to research on the development of new tools, TRAC
supported the continuing GTB involvement in new tool development. There was a discussion
concerning the advantages and disadvantages of targeted versus more wide-based
fundamental research. With the current level of understanding, targeted research is likely to
produce new diagnostic tools based on detection of the organism. TRAC endorsed GTB
involvement in an initiative in diagnostics and suggested GTB should make use of the OECD
supported analysis on TB diagnostic agents. Regarding the development of new vaccines,
TRAC suggested GTB should develop advocacy efforts aimed at fundamental research
8
m: tub^rac\tracrep.doc 3/10/96 8:55
needed for vaccine development. TRAC also agreed GTB should encourage the
pharmaceutical industry to develop new TB drugs.
9
m:\iub\trac\tracrep.doc 3/10/96 8:55
t
t
DISTR. : LIMITED
WORLD HEALTH ORGANIZATION
DISTR. : LIMITEE
ORGANISATION MONDIALE DE LA SANTE
WHO/TB/96.197
Original: English
6^?
TUBERCULOSIS - A GLOBAL EMERGENCY:
CASE NOTIFICATION UPDATE
February 1996
Global Tuberculosis Progranune
World Health Organization, Geneva
This document is not issued to the general public, and all rights
are reserved by the World Health Organization (WHO).
The
document may not be reviewed, abstracted, quoted, reproduced or
translated, in part or in whole, without the prior written permission
of WHO. No part of this document may be stored in a retrieval
system or transmitted in any form or by any means - electronic,
mechanical or other ■ without the prior written permission of
WHO.
Ce document n'est pas destine a etre distribue au grand public et tous
les droits y afferents sont reserves par I'Organisation mondiale de la
Sante (OMS). II ne peut etre commente, resume, cite, reproduit ou
traduit, partiellement ou en totalite, sans une autorisation prealable
ecrite de TOMS. Aucune partie ne doit etre chargee dans un systeme de
recherche documentaire ou diffusee sous quelque forme ou par quelque
moyen que ce soit • electronique, mecanique, ou autre - sans une
autorisation prealable ecrite de TOMS.
The views expressed in documents by named authors are solely the
responsibility of those authors.
Les opinions exprimees dans les documents par des auteurs cites
nommement n'engagent que lesdits auteurs.
WHO/TB/96.197
Introduction
Surveillance of tuberculosis is a feasible undertaking if a country has an effective National
Tuberculosis Programme (NTP). In the absence of such a programme, diagnostic skills are
lacking, case definitions are not standardized, and case recording and reporting are weak or
non-existent.
Under a well established NTP, on the other hand, suspected cases are detected, diagnosed,
and recorded and reported, based on standard registers and quarterly reports. In such a
situation, the number of cases notified by district officers to the regional and national levels
represent a reasonable estimate of the incidence of tuberculosis in the country.
The World Health Organization (WHO) has been collecting case notification data since 1984
from all of its Member States and other countries and territories to assess the burden of
tuberculosis and its trend worldwide. The Case Notification Update, a report published
annually by the Global Programme on Tuberculosis, was started in 1992.
Methodology
In mid 1995, WHO sent a data collection form to all Member States and other countries and
territories to obtain information on the burden of tuberculosis in 1994. These forms were
returned to WHO for data analysis. Additional information was obtained from reports by the
International Union Against Tuberculosis and Lung Diseases (IUATLD), the Royal
Netherlands TB Association (KNVC), other organizations involved in tuberculosis control
and published literature. Prior to finalization of the report, the data were sent to all WHO
Regional offices for verification and updating.
For the 1994 case notification update, 141 out of 214 countries (66%) responded to the WHO
inquiry. Data for another 17 countries were obtained from other sources. These 158 countries
make up 90% of the global population.
Data are reported in three tables per WHO Region, listing absolute numbers of cases notified,
case notification rates, and pulmonary smear positive cases respectively (see tables 2.1 to
7.3).
Not all countries provided WHO with data on 1994. In order to obtain a realistic yet up-todate assessment of the global tuberculosis burden, the latest reports available between 1990
and 1994 were therefore used in generating a global picture of the disease.
1
WHO/TB/96.197
Data Interpretation
The data shown in this report should be interpreted with caution for a number of reasons:
1. the information for the most recent years was sometimes missing or
incomplete,
2. in several countries the performance of the NTP, including their reporting
system, is poor, and
3. case definitions vary between countries. Besides non-standard definitions for
pulmonary and extrapulmonary tuberculosis, not all countries distinguish
between new TB cases and relapses. For this reason, the first two tables per
Region show data on all cases (new and relapses) while the third table,
which is being included here for the first time, provides information on
sputum smear positive cases only, i.e., the infectious cases whose trend is of
crucial significance in the TB epidemic.
Global Tuberculosis Situation
Discrepancies between reported case rates and incidence estimates are evident in maps 1 and
2. Map 1 shows the latest available case notification rates for all countries, categorized by
low (<25), medium (25-100), and high (>100) levels of notifications. Map 2 shows
estimates of incidence in 1990 using the same scale, according to the 1993 World Bank
Development Report1 and WHO.
1 World Development Report 1993. Investing in Healthy World Bank. Oxford University Press, New
York, 1993
2
WHO/TB/96.197
(Map 1)
cn
o
©
i
©
£=
ci
p
11.
§3 =
</)
0)
(U
Hl
Ifl
io
fl
0
sh
is.B
o
o
C O
O ■£
(u
co E
O O
_ *■
O (0 Q
"1!
Q
3
o
V)
0)
o
E
g 2
X-
T-
IO
O
0)
A
CM
□>---------
S _!?•-
8-
$ i Sp
5.0 Q
55|
IlSIDiii
iy
C ro
CD "E
(/) re
O
32
O ra
0)
§8*
w <S ®
°s »
C <0 ©
o_ c
sh
pl
ih
iM
® §z
o°o
til
ifl
fi
is
Estimated tuberculosis incidence rates
1990
2:
$
o
§
1
^0
0
0
Range of rates (per 100 000)
>100
25 to 100
WHO i)6054
<25
no report
area or
the ext
its authorities, or concern!
i the delimitation of its frontiers or boundaries,
lull agreement.
WHO/TB/96.197
Table 1.
TB cases notified in the world, latest reports between 1990 and 1994
No. of
cases notified
(per 100,000 pop)
Africa
541,360
96.8
Americas
264,221
34.9
Eastern
Mediterranean
237,937
55.2
Europe
286,608
33.3
South-East Asia
1,298,999
94.4
Western Pacific
725,014
45.5
GLOBAL
3,354,139
60.1
Figure 1
Distribution of notified tuberculosis cases by WHO Region
WHO Region
Rate
WPR
22%
AFR
16%
SEAR
39%
AMR
EMR
7%
EUR
8%
5
8%
WHO/TB/96.197
Regional Tuberculosis Situation
The figures shown below illustrate the trend of tuberculosis over the past 10 years for each
WHO Region.
In the African and European Regions, the disease is on the increase.
In the African Region, HIV, malnutrition, urbanization and to some extent improvement in
NTPs, may be the underlying factors.
In the European Region, population movements, social upheaval, and HIV in some countries
have resulted in the re-emergence of a disease which until a few years ago had virtually
disappeared.
In the American and Western Pacific Regions, case notification rates have remained relatively
stable, indicating no significant improvements in tuberculosis control.
In the Eastern Mediterranean and South-East Asian Regions, an apparent decrease in the case
notification rates occurred during the past four years. This phenomenon is most likely an
artifact, reflecting marked fluctuations in reporting activities of some large countries.
Figure 2
Case notification rates in different WHO Regions, 1985 - 1994
WHO African Region
WHO Americas Region
165,
f 150
8 115
120
-
105
5.
90
1
J
I8 6075
1
|
45
30
3 1
1965
3
1988
1987
1988
1980
1900
1991
19S2
19S3
165 i
150
135
120
105
90
75
60
45
30 ■~
15
0 —
1965
1994
1986
WHO Eastern Mediterranean Region
’50
’35
8
120
5
5>
105
90
I
75
8
60
|
45
30
1980
1000
1901
1992
1903
18M
165 I
150
t 135 ■
§
8
i
®
3
I
a
1965
19M
V\A4O European Region
- 165
8
1987
’966
1967
19M
1969
1990
1991
1992
1993
1994
3
120 •
105
90
75
60
45
30 ~
15
0 —
1985
6
1968
1987
1988
1980
1990
1991
1992
1993
1994
(
WHO/TB/96.197
V\HO South East Asian Region
WHO Western Pacific Region
165
165
150
f 120
f
150
135
§
120
105
8
105
135
I&
i «>
90
Ii 45 ■"
75
i 60
45
30
3
15
0 —
1905
1900
1987
1968
1989
1980
1991
1902
1983
1GM
|
30
J
15
0 —
1986
1988
1987
1988
1989
1990
1991
1«Q
1983
19M
NTP Performance
A new addition to this report is table 3 which provides information on new pulmonary smear
positive cases. “Proportion of new smear positive cases to all cases” is one of the indicators
used for measuring the performance of an NTP, i.e., the ability of the national TB
programme to detect tuberculosis. This indicator allows countries to be classified according to
three criteria:
1.
>50%
good NTP performance or a reporting system which does
not include pulmonary smear negative or extrapulmonary
cases,
2.
25 - 50%
below average NTP performance, and
3.
<25%
inferior NTP performance.
7
Performance status of National Tuberculosis Programmes, 1994
2
o
a
£
s
°
^0
I
Q
o
o'
Proportion of new smear positive cases to all cases
>50% — good NTP performance
Eii 25-50% — below average NTP performance
I
I <25% — inferior NTP performance
I
I no information
The designations employed and the presentation of material on this map do not imply the expression of any opinion whatsoever on the part of the World Health
Oga zation co
g
dotted lines rep'resent approximate border lines for which there may not yet be full agreement.
WHO 96066
Table/Tableau 2.1: African Region - Number of tuberculosis cases notified/RAgion de I'Afrique - Nombre de cas de tuberculose notifies, 1975-1994
COUNTRY/PAYS
Algeria/Algeria
Angola
Benin/Bdnin
Botswana
Burkina Faso
Burundi
Cameroon/Cameroun
Capa Verde/Cap Vert
Cent Afr Rep /Rep. Centrafricai
ChadTFchad
Comoros/Comores
Congo
Ivory Coast/COte d'Ivoire
Equal Gumaa/Guinde Equal.
Eritrea/Eritree
Ethiopia/Ethiopie
Gabon
Gambia/Gambie
Ghana
Guinea/Gumee
Guinea/Guinde-Bissau
Kenya
Lesotho
Liberia/Libdna
Madagascar
Malawi
Mali
Maurrtania/Mauntanie
Mauritius/Maunce
Mozambique
Namibia/Namibie
Niger
Nigeria/Nigdria
Rwanda
SaoTome & Pnn.
Senegal/Sertegal
Seychelles
Sierra Leone
South Africa/Afnque du Sud
Swaziland
Togo
Uganda/Ouganda
U.R. Tanzania/R U. Tanzania
Zaire/Zaire
Zambia/Zambie
Zimbabwe
Other territories:
Reunion/Rdunion
St Helena/Ste Ketene
Total
No. of countnes/No. de pays
% of countries reporting
_ 1978_____ 1979
. .1975
1,950
173
805
2,421
246
599
2,509
9,375
1,613
2,662
1,060
975
512
289
559
2,889
1,026
2,862
725
1,276
2,946
725
1,055
2,936
1,453
3,991
1,542
1,400
3,534
1,062
1,657
1,959
1,576
2,797
1,047
667
6,990
1,597
2,598
2,437
902
_____ 1?81______1982
_ _ 1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
J994
13,916
7,911
1,793
2,705
2,265
13,133
10,153
1,913
3,101
877
13,832
8,653
2,041
2,706
4,547
2,317
3.393
259
520
1,486
12,917
9,363
2,162
2,627
1,018
2,569
2,138
11,325
8,184
2,027
2,740
949
3,745
4,982
276
814
2,977
11,039
9,587
1,941
2,532
1,616
4,608
5,521
210
64
2,572
11,607
10,288
2,074
2,938
1,497
4,575
5,892
221
2,124
2,591
11,487
10,639
2,420
4,179
13,725
7,157
2,119
4,756
4,883
6,814
6,803
13,382
7,318
2.340
4,654
1,443
4,677
7,064
779
1,285
11,212
8,510
1,901
3,173
1,407
2,739
3,878
285
499
1,086
11,332
10,628
2,162
3,274
3,765
393
1,475
127
13,681
6,625
1,804
2,883
3,061
1,053
3,445
230
1,686
1,977
4,156
6,000
2,776
3,120
6,072
1
3,473
6,422
11
139
4.363
6,982
157
140
591
7,841
260
119
618
8,021
331
108
1,179
9,093
262
1,976
9,563
309
115
3,080
6,062
181
2,648
5,729
17
212
3,878
6,556
20
129
3,716
5,000
2,362
7,202
1,688
2,850
480
460
1,037
427
463
512
2,702
10,117
1,835
2,662
2,577
789
2.434
516
651
220
7,501
1,862
2.605
2,391
643
2,236
344
758
286
1,085
3,995
742
4,197
1,214
4,418
42,423
796
58
4,041
1,884
33,174
549
188
8.452
138
26,596
535
294
5,250
65
26,022
699
294
5,000
83
30,880
758
289
4,479
131
40,096
865
239
5,207
715
5,731
1,948
1,454
1,653
4,679
2,037
4,826
198
2,916
599
7,208
2.498
1,366
1,595
4,170
1,408
3,319
191
5,267
802
6,345
2,236
562
770
10,223
2,861
660
3,160
4,652
877
1,997
182
6,561
645
11,049
4,082
774
9,082
4,758
839
7,576
132
7,457
465
10,027
3,830
1,002
7,464
5,033
933
9,427
157
6,984
2,500
15,334
1,759
780
14,750
2,493
3,100
2,527
0
17
57,506
2,605
4
52,502
2
53,710
3,600
13,587
1,653
108
2,654
3
199
54,476
717
9,877
1,495
131
2,014
16
750
55,310
2,871
10,838
1,386
37
2,573
0
847
59,943
143
305
8,478
234
928
4,419
498
720
374
4,726
3,715
4,672
3,832
6,137
3,539
385
957
318
4,715
4,903
3,139
144
163
7,594
3,710
5,243
3,494
208
1,058
5,103
5,122
5,342
4,057
275
244
246
183
217
179.023"_16716b3~’i78;i6d’“ 187,886
30,454
516
343
6,355
5,086
2.795
133
5,320
222
2,142
1,189
18,498
2,691
64,794
802
6,016
4,318
1,164
2,784
195
5,953
14,292
1,447
191
2,967
951
1,904
3,338
285
468
1,430
3,840
7.312
2,045
2,912
356
52.403
761
56,824
752
65,045
654
71,731
855
80,846
769
85,867
864
95,521
721
80,795
912
88,634
917
60,006
906
926
972
99.329
1,034
4,345
1,469
2,651
1,935
1,203
3,235
1,317
3,925
1,128
5,877
1,214
5,297
1,740
6,017
1,869
6,407
1,388
6,732
2,274
7,044
2.867
7,569
3,300
8.894
3,241
205
376
11,966
3,443
885
3,588
4,707
532
2,333
368
530
10,460
2,927
425
3,220
5,334
1,621
4,406
111
5,645
1,310
10,022
21
232
3,717
6,301
1,851
2,257
119
8,263
752
10,515
225
384
4,007
7,581
2,534
3,722
117
10,996
778
10,957
2,346
894
4,393
8,247
2,578
3,928
114
13,863
1,362
12,592
2,463
1,163
11,788
2,525
1,246
12,320
2,994
1,059
14,599
3,327
1,948
8,126
15,183
1,876
4,316
130
15,085
1,558
20,451
3,384
1,766
9,855
17,105
3,309
3,996
159
16,588
1,647
22,930
4,334
1.764
10,671
4,427
570
14,071
2,460
8
927
24
358
55,013
3,640
556
19,723
3,287
55
2,815
631
25,700
4,145
13
5,611
10
120
61,486
1,352
3,703
608
13,342
4,741
2,500
1,756
19,626
3,200
14,802
5,500
626
11,601
10
816
62,717
4,840
698
14,937
1,327
40
1,065
10
865
59,349
68,075
1,394
343
745
1,184
1,464
16,920
6,343
11,525
5,848
1,071
3,066
18,254
940
1,045
19,516
1,324
14,740
22,544
12,876
6,002
14,239
6,822
4,932
835
3,573
4,411
187
2,327
121
5,787
832
152
5,937
2,923
8,673
4,404
1,872
3,977
118
5,204
40
1,612
16
889
64,115
3,059
673
10,212
1,364
59
2,417
16
293
62,556
1,955
126
1,170
6,964
8,929
6,070
4,051
204
497
11,748
174
2,029
11,783
12,089
14,292
6,519
4,577
6,948
3,881
6,500
5,694
6,747
4,759
596
1,392
15,477
5,043
7,909
5,233
153
174
175
167
118
146
173
130
159
"213.5M
224,904
42
88
230^27
30
81
243,856
38
70
269,785
278,493
314,719
41
85
42
88
44
92
343,373
44
92
754
10,949
665
11,439
1,419
49
6,145
14
130
57.406
1,098
5,417
9,431
1,626
4,040
129
15,958
5,965
6
6,261
6,015
12,364
14,322
5,284
119
15,899
3,064
134
16,609
2,671
5,200
20,122
6,387
17
4,977
41
666
80.400
120
5,025 .
5,038
1,466
77,652
1,531
1,664
82.539
22,902
9,132
1,243
19,016
25,449
21,135
23,373
11,710
1,223
20,662
28,711
31,400
25,732
16,237
131
114
122
136
334,469
396,625
41
65
392,929
30
81
352,556
34
71
97
6,487
5
2,691
89,786
1,458
1,005
21,579
31,827
34,319
36,889
20,125
3,076
149
27.885
3,784
8,449
6,913
2,564
90,292
1,137
26,994
34,799
1
173,872
28
58
40
38
40
42
41
83
79
83
88
85
246;i06
41
85
41
85
406,862
402,346
39
81
29
60
X
o
3
o
c*
§
AFRO94 WK4 02/08/96
Table/Tableau 2.2: African Region - Case notification rates (per 100 000 populatlon)/R6gion de I’Afrique - Taux de notification des cas (pour 100 000 habitants), 1975-1994
1975
COUNTRY/PAYS
Algeria/Algene
Angola
Benin/B6nm
Botswana
Burkina Faso
Burundi
Cameroon/Cameroun
Cape Verde/Cap Vert
Cent. Afr. Rep /Rep. Centrafncai
Chad/Tchad
Comoros/Comores
Congo
C6t0 d'Ivoire
Equal. Gumea/Gumee Equal
Enlrea/Eritree
Ethiopia/Elhiopie
Gabon
Gambia/Gambie
Ghana
Gumea/Gumee
Gumea/Gumee-Bissau
Kenya
Lesotho
Liberia/Liberia
Madagascar
Malawi
Mali
Mauntania/Mauritanie
Maurilius/Maurice
Mozambique
O
Namibia/Namibie
Niger
Nigena/Nigdria
Rwanda
Sao Tome & Prin.
Senegal/Sendgal
Seychelles
Sierra Leone
South Africa/Afrique du Sud
Swaziland
Togo
Uganda/Ouganda
U.R. Tanzania/R.U Tanzania
Zaire/Zaire
Zambia/Zambie
Zimbabwe
Other territories
Reunion/Rdunion
St Melena/Ste Helene
Total'-------
AFRO94 WK4 02/08/96
1976
1977
1978
1979
__ 1980
1981
1982
1984
1985
1986
1987
1988____ 1989
1990
1991
1992
1993
1994
13.0
105.8
50 1
325.7
7.1
11.4
12.3
149.3
20.5
11.7
14.4
144.7
53.1
2938
37.0
191
28.1
178.5
28.1
4.9
1044
524
277.7
33.5
15.2
25.1
1178
32.0
6.2
69 7
107.3
49.0
278.6
31 0
21.8
41.1
132.8
60.9
2.7
66.4
87 5
47 9
286.9
409
23.5
366
76.7
680
41.2
61.8
130.7
49 4
298.2
114
41.3
34.5
93.4
18.5
29.1
63.2
108.5
51.2
251.3
57.7
48 8
34.0
83.5
20.0
296
57.4
114.4
526
2356
126
52.5
20.8
47.7
94.7
46.5
229.5
11.1
72.2
45.8
84.7
29.2
56.0
45.4
107.7
43.2
205.0
18.5
863
49.3
63.1
2.2
47.4
46.5
111.9
44.8
230.3
16.7
83.1
51.1
64.8
72.6
46.7
44 4
1116
45 2
248.6
44 0
107.6
49 1
307.5
50.2
67 1
40.4
3296
86 1
57.5
55.8
50 1
71 2
460
332 2
14 8
77.6
56 4
29.3
251
48.5
101.2
44.9
275.0
17.0
54.4
36 7
89.1
18.3
20.8
66.9
50.7
44.5
51.2
70.7
51 9
2104
56.5
228.9
65.2
148.5
63.4
61.6
137.7
57.7
5.4
157.5
58.8
0.3
170.2
59.9
3.3
42.0
184.6
58.9
5.9
26.5
201 5
60.5
45.5
25.8
26.5
65.5
73.9
21.1
26.9
646
91.9
18.5
49.7
70 7
71.0
21 3
80.9
71.8
81.5
183
1224
113.9
94 5
8.8
36.5
41 4
137.3
86.7
145.4
82.2
1623
68.8
174 4
86.8
191 3
75.5
197.6
82.3
2136
66.6
175.5
81.9
186.9
80.0
122 8
76 8
76 3
77 9
185 9
806
37.9
31.6
22.2
17.5
15.6
24.7
25.2
26.4
296
22.0
42.9
23.0
37.5
32.1
41.3
33.5
42.7
24.1
43.5
38.3
44.1
46 9
46.0
52.3
52.5
499
24.6
44.4
64.5
234 4
429
36.0
70.0
7.1
139.1
15.2
45.5
42.8
60.7
52.6
187.3
19.3
303
73.6
20.5
249.5
10.9
41.7
147.4
48.7
1.3
10.2
33.8
82.8
22.7
124.6
11.6
603
83.0
49.4
13.6
16.4
35.2
94.5
30.2
200.3
11.3
79.7
B4.1
49.8
138.1
37.0
374
97.3
29.8
206.2
11.0
99.8
144.3
55 3
141.2
120.6
49.9
140 9
126 6
50.3
162.6
44.5
105.6
182
206.9
12.3
113.9
49.8
132.0
46 3
146 5
149 2
12.6
115.1
151.6
77.5
1742
62.1
71 1
162 6
32 6
184 9
146
109.8
1569
839
217.1
60.0
746
263.8
11.3
112.1
105.3
57.4
175.9
70.8
60 6
149.4
19 1
204 8
12.0
102 4
410.9
10.6
18.0
21.9
37.7
16.7
15.4
24.1
179.6
365.9
8.4
16.5
39 4
7.4
14 1
364
9.8
162.6
292.8
7.9
22.4
51.2
495
91.1
20 9
3.5
165.8
160.1
220.3
8.7
28.4
62.7
11.5
80.9
14.7
3.1
173.6
191.8
282.0
8.1
14 3
69.7
198.0
67.3
20.9
91.4
14.3
67.9
58.6
16.7
216.9
180 5
123.4
19.8
44 6
992
668
14.5
376.5
7.3
11.0
36.8
9.1
72.7
18.7
156.5
65.1
323
18.3
76.0
27.5
6.0
78.7
37.5
82.1
88.1
168.9
64.6
180.5
71.0
28.9
31 3
50.7
3554
16 5
17 8
108.7
500
318.8
37.6
23.6
80.6
48.4
61.8
38 2
58.9
87.2
27.8
59.8
141.7
492
315 0
16.0
24 9
6.3
101.8
25.3
67.4
70.9
42.4
221 0
85.7
41 9
212.6
689
40.2
92.2
52.6
25 2
46 2
4656
100.6
81 9
33.2
84 5
3.3
787
76.1
50.3
51.7
1.5
75.1
95.0
48.8
48.6
1.9
87.0
98 3
46.5
42.7
3.0
1103
107.3
37.3
48.5
108 3
40.2
160 5
87 6
20.6
86.3
32 3
343.5
21.8
27.0
86.1
48.6
201.1
79.9
19.3
74.4
21.9
230 6
20.7
47.5
109.6
41 2
175 9
31 9
70.7
74.4
17.7
1887
208
52.1
1005
63.8
219.6
36.3
36.0
77.6
13.1
132.1
191
55.8
81.1
66.4
304.9
41.3
100.2
77.0
122
488.5
13.7
61.7
569
58 1
277.7
51.8
79.8
79.1
13 2
592.5
16.1
56 2
24 9
29.5
50.9
23.8
38.8
15.4
22.3
53.3
56.0
51.1
51.2
59.3
21.0
29.9
217.0
56.7
0.1
194.2
0.1
193.7
12.8
13.7
29.0
139.4
364
25.4
23.2
189.6
49 7
14.6
26.0
38.5
45 2
0.6
218.3
66.6
194
33.1
118.7
49.3
4.8
6.3
191.5
13.3
75.8
100
8.0
27.0
20.7
6.0
2.2
97.6
60.5
93.5
60.6
118.7
54.3
15.6
7.8
1.8
186
91 7
46.8
5.7
1.3
42.2
142
94.7
50.6
8.0
8.1
27.5
19.0
93.1
56.9
4.7
8.7
36 3
32.0
102.1
55.1
56.9
50.2
20.0
50.2
37.0
43.4
30.2
----------- 17.7
62J
57.2
533
94 5
81.0
62.6
64.6
37.0
235.5
8.3
388 0
24.9
20 4
252.4
166.4
— MJ-------- 58.0
346.6
41.8
37.0
67.5
2.6
142.5
12.2
45.4
84.3
63.3
24.4
231.1
11.7
39.1
10.4
14.2
24.2
47.6
40.4
27.5
25.0
26.4
208.9
514.1
109
13.0
24.0
58 4
40.1
25.0
8.5
198.8
318.9
7.4
3.6
59.3
6.1
14.2
57.7
11.7
24.6
57.3
65.6
1058
60.3
108.8
49.5
98.2
70.2
98.3
56.7
19.1
8.9
68.7
15.4
111.2
60.3
33.9
33.6
31.5
21.9
26.5
30.9
22.8
27.3
6T.7"
67.3
‘65.3
70
763
62.T
68.6
78.5
25.7
200 4
24.8
126
14.4
192.8
15.4
23.3
194.4
83.6
8.7
68.1
51.2
65.4
35.8,
204.8*
2001
39.7
212.9
281.0
92.2
34 1
102.3
96.4
546
277.9
114 9
32 5
107.3
105 5
786
296.7
155.1
22.1
18.9
19.9
21.8
75.3
87.9
80.7
80.0
187.9
192.5
61 8
56 8
76.4
82.1
6.9
62.6
226 4
180.2
91.5
29.4
135
179.6
42 8
7.8
85.3
58.2
2226
259
108.2
113.6
83.2
412.8
187.4
28 4
1309
120.6
85.0
90.6
5
*
Table/Tabieau 2.3: African Region - New Smear Positive Cases/R6gion de I'Afrique • Frottis Positif Nouveaux
COUNTRY/PAYS
Algeria/Algeria
Angola
Benin/Benin
Botswana
Burkina Faso
Burundi
Cameroon/Cameroun
Cape Verde/Cap Vert
Cent. Afr Rep /R6p. Centrafncam
Chad/Tchad
Comoros/Comores
Congo
Cdte d'Ivoire
Equal. Guinaa/Guinde Equal.
Eritrea/Eritree
Ethiopia/Ethiopie
Gabon
Gambia/Gambie
Ghana
Guinea/Guinee
Guinea/Guinee-Bissau
Kenya
Lesotho
Libena/Libena
Madagascar
Malawi
Mali
Mauritania/Mauntanie
Mauntius/Maunce
Mozambique
Namibia/Namibie
Niger
Nigeria/Nig6ria
Rwanda
Sao Tome & Prin.
Senegal/S6n6gal
Seychelles
Sierra Leone
South Africa/Afrique du Sud
Swaziland
Togo
Uganda/Ouganda
U.R. Tanzama/R.U Tanzania
Zaire/Zaire
Zambia/Zambie
Zimbabwe
No. of cases
1994
Rate (per 100,000 pop)
1993 1994
6,793
4,337
1,618
1,668
24.9
40 6
30.8
115.6
4,874
1,653
1,508
1,203
1,861
2,316
1,527
1,883
47.4
32.5
107.6
12.3
30.9
18.5
24.6
14.6
67%
71%
32%
83%
40%
33%
1994
49%
61%
76%
35%
40%
26%
71%
55%
13.0
67.2
82
1,691
73%
52.7
7,012
Proportion of new smear ♦ cases
to all cases
_____ 1993
274
70.4
77%
5,752
395
108
30.8
6%
38%
2,082
2,158
33.0
33 2
63%
67%
1,059
10,149
1,405
1,547
6,881
800
11,324
1,330
76.2
41.4
66.6
51.5
68%
50%
42%
88%
70%
49%
49%
31%
7,366
103.0
38.5
72.3
54.4
49.7
69%
9,526
9,677
63.1
62.3
57%
35%
463
1,723
1,865
5.4
1.6
21.1
74%
15%
49%
40%
2.8
2
14.0
59.9
55 6
36.2
71.6
59 5
54%
55%
49%
43%
55%
49%
26%
49.6
5,331
55%
32.0
1,408
14,763
17,164
67%
56.8
4,599
545
11,949
15,569
14,924
29%
22.0
28
Other territories:
Reunion/Rdunion
St Helena/Ste H6ldne
Total/Total Average
103,610
98,474
2iJ
2X2
25%
24%
o
§
3
Table/Tableau 3.1: American Region - Number of tuberculosis cases notlfied/Rdgion des Am^riques - Nombre de cas de tuberculoso notifies, 1975-1994
l^J
COyNTRY I PAYS
^_1?75
1976
1977
1978
1979
1980
1981
1982
-I??3
1984
1985
1986
.1987
1988
__1?89
1990
1991
1992
Antigua & Barbuda
Argentina/Argentine
Bahamas
Barbados/Barbade
Belize
Bolivia/Bolivie
Brazil/Br&sil
Canada
Chile/Chili
Colombia/Colombie
2
14,311
57
26
29
2,594
53,419
3,551
8,289
12,424
3
18,808
44
16
52
2,965
51,301
3,143
9,482
11,878
6
17,197
33
14
28
3,637
54,552
3,194
9,312
11,569
8
16,267
26
16
13
3,852
56,484
2,940
8,257
10,725
2
16,569
56
23
34
4,062
64,734
2,761
8,105
10,924
8
16,406
70
64
21
4,412
72,608
2,885
8,523
11,589
3
16,693
67
3
33
5,072
86,411
2,554
7,337
11,483
0
17,292
54
30
44
4.777
87,822
2,515
6,941
12,126
1
17,305
58
17
140
5,178
86,617
2,186
6,989
13,716
3
16,359
53
14
35
4,131
88,365
2,345
6,561
12,792
2
15,987
63
12
25
7,679
84,310
1,980
6,644
12,024
7
14,681
52
7
23
6,837
83,731
2,046
6,854
11,639
0
13,368
43
3
41
8,960
81,826
1,952
6,280
11,437
3
13,267
51
4
28
10,664
82,395
2,032
6,324
11,469
3
12,636
52
5
30
12,563
80,048
1,982
6,728
11,329
1
12,309
46
5
57
11,166
74,570
1,964
6,151
12,447
0
12,185
53
5
89
11,223
84,990
2,043
6
12,606
63
6
65
9,520
85,955
2,091
5,304
11,199
Costa Rica
Cuba
Dominica/Domimque
Dominican Rep /R6p. Dornin.
Ecuador/Equateur
El Salvador
Grenada/Grenade
Guatemala
Guyana
Haiti/Haiti
433
1,326
10
1,468
2,790
2,875
8
6,335
137
5,361
389
1,270
14
1,424
2,647
3,181
1
6.208
160
4,897
371
1,257
20
1,287
2,858
2,658
11
6.695
120
1,110
383
1,261
11
1,387
2,617
2,449
0
5,353
110
610
462
1,133
5
2,093
3,149
2,281
10
5,307
71
1,390
396
1,133
20
2,174
3,950
2,255
17
5,624
124
8,306
521
833
26
1,778
3,966
2,091
1
6,641
117
6,550
459
815
18
2,457
3,880
2,171
1
7,277
135
3,337
479
762
16
2,959
3,985
2,053
6
6,013
149
6,839
393
705
5
3,100
4,301
1,564
4
6,586
165
5,803
376
680
8
2,335
4,798
1,461
2
6,570
215
4,959
418
656
35
2,634
5,687
1,659
1
4,806
190
8,583
434
630
27
2,459
5,867
1,647
2
5,700
117
8,514
442
628
7
3,081
5,497
2,378
0
5,739
150
8,054
311
581
13
3,145
5,480
617
4
4,900
120
8,100
230
546
6
2,597
8,243
2,367
0
3,813
168
201
514
14
1,837
6,879
2,304
Honduras
Jamaica/Jamaique
Mexico/Mexique
Nicaragua
Panama
Paraguay
Peru/Pdrou
St Kitts & Nevis
St Luda/Ste Lucie
1,793
348
11,417
1,160
965
1,031
16,668
13
54
1,421
368
11,332
1,937
1,021
1,075
22,257
8
33
1,435
349
10,713
1,741
909
1,087
17,660
3
37
1,323
365
10,158
1,932
888
1,014
15,506
1
50
1,414
255
26,931
1,132
750
1,317
15,616
1
42
1,674
176
31,247
1,300
643
1,354
16,011
7
41
1,696
178
32,572
3,723
580
1,388
21,925
4
39
1,714
153
24,853
3,082
580
1,415
21,579
6
37
1,935
157
22,795
2,773
429
1,800
22,753
2
48
2,120
160
14,531
2,705
413
1,718
22,792
3
55
3,377
130
15,017
2,604
614
1,931
24,438
0
21
4,213
88
13,180
2,617
709
1,628
24,702
0
34
4,227
133
14,631
2,983
765
1,502
30,571
0
25
3,962
65
15,371
2,737
770
1,438
36,908
0
32
4,026
86
15,489
3,106
672
2,270
35,687
0
28
3,647
123
St Vincent
Suriname
Trinidad & Tobago/Tnnitd & Tob.
USA/Etats Ums
Uruguay
Venezuela/Venezuela
54
63
85
33,989
1,706
4,395
33
9
168
32.105
1,635
4.222
21
71
125
28.521
1,709
4,167
19
69
88
27,669
1,850
4,161
78
78
80
27,749
1,874
4,233
11
81
82
27,373
1,699
4,093
14
56
62
25,520
1,450
4,159
4
78
112
23.846
1,359
4,266
23
76
108
22,255
1,389
4,737
14
50
112
22.201
1,201
4,822
9
104
157
30,145
1,654
4,019
119
22,768
1,082
4,974
3
77
122
22,517
1,023
4,954
6
77
108
22.436
951
4,557
6
0
137
46
72
5
6
1
160
49
67
1
3
0
110
28
50
4
0
3
1
71
41
74
1
0
1
0
124
82
40
1
0
2
2
117
32
0
10
1
125
50
61
1
0
3
1
171
70
41
7
1
3
4
219
0
6
1
30
0
2
0
53
0
4
5
0
135
129
63
0
54
9
40
5
45
13
395
0
0
5
310
0
0
4
686
2
0
0
521
0
0
1
473
2
0
1
452
5
1
418
0
0
3
338
4
0
1
363
2
0
1
238^33~ 227J083
227,295
45
96
Other territories:
Anguilla
Bermuda/Bermudes
Cayman Islands/lles Caiman
French Guiana/Guyane Fran
Guadeloupe
Martinique
Montserrat
Neth. Antilles/Antilles N6er
Puerto Rico/Porto Rico
Turks & Caicos Is./lles T & C
Virgin Is./lles Vierges (UK)
Virgin Is./lles Vierges (USA)
Total
No. of countries/No tie pays
% of countries repotting
AMR094 WK4 02/08/96
524
0
9
189,980
43
91
457
3
0
6
196,261
45
96
0
0
120
0
77
363
0
0
2
186,504 “WlB-204,990"
43
91
48
98
48
98
22e;ow—248,299---- 237,643
48
98
45
96
46
98
2
46
98
46
98
12,263
118
410
13
3,490
7,313
2,495
3
3,755
134
10,237
182
2,944
4,560
121
15,216
2,797
4,155
111
8,897
2,885
2,167
37,905
0
13
2,283
40,580
1
25
1,927
52,552
4
26
3
70
124
23,495
987
4,524
2
70
120
25,701
886
5,457
1
33
141
26,283
759
5,216
4
50
142
26.673
699
5,444
0
1
0
52
21
47
6
0
2
2
15
8
0
0
2
0
3
3
0
4
3
11
5
24
1
13
1
0
303
12
275
314
1
2
0
6
1
4
159
0
0
4
241
0
1
4
227377"
233,271
2~42;039
239376--- 2157911"
247,009
45
96
45
98
45
96
60
45
96
41
87
42
89
0
0
1993
1994
13,887
60
13,663
80
8,614
59
9,431
87,280
2,012
4,598
11,043
8,901
790
7
4,033
7,050
3,347
0
2,646
91
325
1,681
12
3,783
9,685
3,901
3
2,976
266
3,745
115
15,145
2,798
1,146
2,037
51,675
6
13
112
25,287
689
5,169
109
16,353
2,750
827
1,850
48,601
2
24
0
53
129
24.361
666
4,877
2
91
31
33
0
257
0
0
244,415 — ISMOT
38
81
4,291
34
72
24M79"
32
68
o
£
5
Table/Tabieau 3.2: American Region - Case notification rates (per 100 000 population)/R6gion des Amiriques - Taux de notification des cas (pour 100 000 habitants), 1975-1994
country/pays
Antigua & Barbuda
Argentina/Argentine
Bahamas
Barbados/Barbade
Belize
Boltvia/Bolivie
Brazil/Bresil
Canada
Chile/Chili
Cokxnbia/Colombie
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
__J?87_
JML
1989
1990
J9?1_____ 199?_____J993
3.4
54 9
30.0
106
21 6
54.5
49.4
15.3
80 2
52.3
100
640
168
5.7
20.3
72.8
48.2
13.4
87.4
46.6
13.3
596
12.9
6.5
9.3
75.2
48.7
12.2
76.4
42.3
3.3
59.8
27.2
9.3
23.8
77.5
54.6
11.3
73.8
42.1
13.1
58.4
33.3
25.7
144
82.4
59.9
11.7
76.5
43.7
4.9
58.5
31.2
1.2
22.1
928
69.7
103
64.8
42.4
59.7
24.5
120
28.8
85.8
69.3
10.0
60.4
43.8
1.6
58 8
25.8
6.8
89.2
91.3
66.9
8.6
59.8
48.5
4.8
54.8
23.1
5.6
21.6
71.5
66.8
91
55.2
44.3
3.2
52.7
26.9
4.7
15.1
130.3
624
7.6
55.0
40.8
11.3
47.7
21.8
2.8
135
113.6
60.8
7.8
55.8
38.7
42.8
17.7
1.2
234
145.8
58.3
7.3
50.3
37.3
4.8
41 9
206
1.6
15.6
169.8
57.6
7.5
49 7
36.8
4.8
39.3
20.6
2.0
16.3
1956
54.9
7.2
52.0
35.7
1.6
37.8
18.0
1.9
30.2
169.9
50.2
7.1
468
38.5
9.2
37.8
23.9
2.3
32.7
138.1
55.9
7.3
39.0
33.5
41 1
22.4
40.0
39.2
1220
28.1
130.3
54.8
17.3
11.7
27.0
382
49.6
49.8
19.1
81.3
163
1552
22 1
8.5
35.1
30 5
48 5
45.7
1.1
93.4
15.3
120.2
189
8.3
24.7
41.2
46.1
47.0
1.1
995
17.5
60.2
19.2
7.7
21.9
48.5
46.1
44.2
6.7
79.9
19.1
121.1
15.3
7.0
6.8
49.7
48.5
33.4
4.4
85.1
21.0
100.9
14.2
6.7
11.1
366
52.7
30.8
2.2
82.5
27.2
84.6
15.4
6.4
48.6
404
61.0
34.5
1.1
58.6
24.0
143.5
15.5
6.1
37.5
36.9
61.4
33.7
2.2
676
148
139.5
15.4
6.0
9.7
45.2
56 1
47.8
7.6
5.2
8.5
36 5
803
45.8
6.5
4.8
19.7
25.3
65.5 •
43.6
1.1
39.7
16 7
154.7
3.7
3.8
18 3
47.2
68 1
46.2
3.3
73
99
53.5
642
607
66.1
18.9
129.3
10.5
5.5
18.1
45.2
546
12.2
4.4
54.8
15.1
127.5
74.7
5.2
90.7
5.1
17.6
73.5
802
4.6
10.1
72.9
42.1
234.1
9.5
19.0
5.1
71 1
22.8
6.5
38.2
60.8
46.4
13.4
904
48.9
37.0
204
1.9
45.9
166.8
56.2
7.3
37.3
Costa Rica
Cuba
Dominica/Dominique
Dominican Rep /R6p. Dornin.
Ecuador/Equateur
El Salvador
Grenada/Grenade
Guatemala
Guyana
Haiti/HaTti
22.0
14.2
139
29.1
404
70.4
8.7
105 2
18.7
109 0
19.2
13.5
19.2
27.5
37.2
76 1
1.1
1003
21 7
97 9
17.8
132
274
24.3
39.1
62.1
12.2
1052
162
21.8
17.8
13.2
15.1
25.5
348
560
81.8
147
11.8
20.8
118
6.8
37.6
40.7
51.2
112
78.9
9.4
26.4
Honduras
Jamaica/Jamaique
Maxtco/Mexique
Nicaragua
Panama
Paraguay
Peru/P6rou
St Kitts & Nevis
St Lucia/Ste Lucie
59 4
17.3
194
47.8
560
384
109.9
28.9
50.0
45 6
18.1
18.7
77.5
57.7
38.9
142 8
17.8
30.0
44.5
169
17.2
677
50.1
38 2
110.3
6.7
33.3
39.6
17.5
15.9
73.0
477
34.5
94 3
2.2
44.6
41.0
12.1
41 2
41.6
39.4
434
92 5
2.3
36.8
46.9
8.3
466
464
330
43.2
92.4
15.9
35.7
460
8.2
474
129.0
29.1
42.8
123.4
9.1
33.3
45.0
6.9
35.3
103.8
28.5
42.3
118.5
13.6
31.1
49.2
7.0
31.6
90.7
20.6
52.0
122.0
4.5
40.0
52.3
7.0
19.7
86.0
19.5
48.1
119.4
7.0
45.1
80.7
5.6
19.9
80.6
28.3
52.3
125.2
97.5
3.8
17.1
79.1
32 1
42.7
123.9
94.9
57
18 5
88.0
33.9
38.2
150.2
86.3
2.8
19.0
78.9
334
35.4
177.8
85.0
3.7
18.7
87.1
28.6
54 2
168.6
50.2
1756
16.9
27.0
195
24.8
21.4
9.8
51.4
184 3
2.4
18.5
29.1
15.2
13.7
57.9
29.4
21.9
19.9
119
12.8
59.5
29.4
196
19.4
8.3
12.3
63 9
28.4
79.6
22.0
7.4
12.2
64.3
280
11.1
22.7
7.5
11.9
57.9
26.4
14.0
15.5
5.6
11.0
49 1
26.1
4.0
21.3
9.9
10.2
45.7
26.1
22.8
20.5
9.4
9.4
46.5
28.3
13.7
13.3
9.7
9.3
39.9
28.1
8.7
15.7
10.1
9.5
35.8
28.3
2.9
19.9
10.2
9.3
33.6
27.4
5.7
19.7
8.9
9.2
31.1
24.6
2.8
17.7
10.1
9.5
32.1
23.8
1.9
17.5
9.7
10.3
286
28.0
0.9
8.1
11.3
10.4
24.4
26.1
3.7
12.2
11.2
10.5
22.3
26.6
1.9
3.7
11.1
162 5
9.7
18.2
5.0
1543
14.7
18 3
9.1
5.5
5.0
198 8
20.1
12.2
63.6
14.3
5.4
19.0
240.7
10.5
4.5
31.3
34
1.7
4.8
11.1
6.5
10.7
52.5
15.8
81.8
11.6
455
12.9
118.2
49.1
5.6
13.4
54.5
3.3
8.0
13.5
2.1
3.1
45.5
7.7
174.6
8.6
15.3
33.3
5.6
5.9
109.2
12.6
22.7
8.3
6.7
9.1
3.6
9.1
12.6
9.8
12.5
10.0
44.4
10.7
20.0
9.0
4.5
6.8
3.0
1.0
1.0
8.8
120.0
6.7
2.0
7.9
1.0
13.7
55.6
7.7
2.0
5.9
6.3
4.0
3.9
5.9
3.9
37.7
37.2
34.8
34.3--------- 33.8
34.1
M8
33.9
34.7
34.7
St Vincent
Suriname
Trintdad & Tobago/Trimtd & Tob.
USA/Etats Unis
Uruguay
Venezueia/Vendzuela
Other Territories:
Anguilla
Bermuda/Bermudes
Cayman Islands/lles Caiman
French Guiana/Guyane Fran.
Guadeloupe
Martinique
Montserrat
Neth. Antilles/Antilles Neer.
Puerto Rico/Porto Rico
Turks & Caicos Is./lles T. & C.
Virgin Is./lles Vierges (UK)
Virgin Is./lles Vierges (USA)
Total
. JI94-
A?L51?76
58.1
17.3
8.4
15.7
60.3
34 5
11 3
240 4
140
21 9
17.5
9.5
34.0
35.1
2.5
16.4
14.7
57.6
32.0
83.3
113
6.7
271.2
14 9
20.4
16.7
5.6
196.7
23.5
15.0
42.9
11.8
6.2
2.0
34.5
32.2
5.1
30.4
182.4
25.1
12.3
8.3
21.4
28.6
1.0
4.1
34.2
16.1
“ 37.4
40.0
57.1
9.1
177.6
385
19.1
14.5
25.0
41.5
21 1
80.1
22.5
7.0
33.3
32.5
26.4
11.2
70.2
4.8
16.8
68.0
45.2
433
225 8
14.3
25.8
9.7
15.3
16.9
49.2
86.3
69.2
3.3
28.8 .
32.2
78.1
4.5
17.8
643
32.0
38.3
208 3
4.9
17.0
11.8
88
9.8
21.9
24.7
12.7
10.0
9.3
21.0
22.8
6.9
67.4
7.5
8.9
7.1
34.1
28.5
3U*
$
o
Cd
§
AMR094 WK4 02/08/96
Table/Tableau 3.3: American Region - New Smear Positive Cases/RAgion des Am^riques - Frottis Posltif Nouveaux
No.of cases
COUNTRY/PAYS
Antigua & Barbuda
Argentina/Argentine
Bahamas
Barbados/Barbade
Belize
Bolivia/Bolivie
Brazil/Brdsil
Canada
Chile/Chili
Colombia/Colombie
5,696
36
6,905 3
542
2,629
2,199 3
6,987
6,532
565
6
2,297
5,325
2,471
0
2,128
51
2,016
83
8,164
1,714
1,046
985
35.646
2
11
16,046
388
2,849
230
914
8
1,762
6,674 3
2,144
3
2,012
61
2,385
61
9,726
1,615
748
873
33,925
2
0
55
14,346
381
2,736 ’
2
27
16
24.5
967
16 7
Proportion of new smear ♦ cases
to all cases
1993
1994
43%
17.1
63%
61%
95.4
79%
73%
51%
28.1
1.9
19.0
20.6
5.2
8.5
30.5
485
44.8
16.9
6.9
8.3
11.3
22.9
59.5
380
3.3
21 2
6.3
19.5
37.8
3.4
9.1
41.7
41.2
74
27%
57%
63%
72%
86%
57%
76%
74%
69%
55%
80%
56%
68%
23%
54%
72%
54%
56%
56%
59%
59%
90%
21.0
1558
145.4
61%
91%
48%
69%
4.8
4.9
33%
47%
70%
100% 4
85%
10.0
43%
4.3
6.2
12.3
13.6
73%
71%
54%
67%
47%
100% 4
43.4
2.5
106
37 8
28.9
18.1
5.5
12.0
12.8
63%
59%
56%
55%
57%
0.5
100% *
8.2
87%
48%
117
0
3.2
46%
104,974
146,716
_1«70
207
63%
X
o
42%
68%
56%
0
3 includes relapses
4 only new smear ♦ cases notified
AMRO94 WK4 02/12/96
17.6
15.3
44,687
St Vincent
Suriname
Trinidad & Tobago/Trimtd & Tob.
USA/Etats Unis
Uruguay
Venezuela/Vendzuela
Total
5,937
50
6,833
Honduras
Jamaica/Jamaique
Mexico/Mexique
Nicaragua
Panama
Paraguay
Peru/P6rou
St Kitts & Nevis
St Lucia/Ste Lucie
Other territories:
Anguilla
Bermuda/Bermudes
Cayman Islands/iles Caiman
French Guiana/Guyane Fran.
Guadeloupe
Martinique
Montserrat
Neth. Antilles/Antilles N6er.
Puerto Rico/Porto Rico
Turks & Caicos Is./lles T & C.
Virgin Is./lles Vierges (UK)
Virgin Is./lles Vierges (USA)
1994
41
Costa Rica
Cuba
Dommica/Dominique
Dominican Rep /Rep. Dornin.
Ecuador/Equateur
El Salvador
Grenada/Grenade
Guatemala
Guyana
Haiti/Haiti
72
I?”
Rate (per 100,000 pop)
1993
1994
59%
§
Table/Tableau 4.1:
Eastern Mediterranean Region -Number of tuberculosis cases notified/RAgion de la MAditerranAe Orientale- Nombre de cas de tuberculose notifies, 1975-1994
_ 1182_____ 1984
1985
1987
1988
1989
1990
1991
1992
1993
1,932
8,589
6,970
856
855
284
18,784
208
39
1,489
1,572
10,493
6,807
672
812
410
10,742
194
61
2,262
1,308
8,728
6,485
769
717
1,943
14,351
156
48
1,864
1,209
8,032
6,846
592
611
2,257
18,091
120
35
1,978
22,063
10,034
6,517
537
540
2,478
16,051
142
39
2.030
1,378
9,967
6,504
553
480
14,386
122
23
2,040
1,492
12,005
8,032
484
468
4,332
117
29
2,100
2,142
9,255
14,684
23,067
142
43
2,900
3,634
14,246
140
39
2,884
8,876
14,121
439
277
390
330
884
504
282
884
114
37
3,489
3.426
20,569
18,553
427
217
610
26,944
802
117,739
206
7,551
2,838
357
22,279
843
91,572
203
7,163
2,719
325
26,790
861
111,419
250
3,966
2,722
1,509
2,163
2.510
568
7.241
276
27,553
1,265
149,004
220
3,696
3,079
2,460
3,942
2,487
464
8,366
331
27,159
616
179,480
248
3,029
7,322
800
4,290
2,272
818
9,822
416
25,717
477
194,323
223
2,433
2,728
693
4,952
2.309
339
19,993
265
26,756
478
170,562
191
2,583
1,323
701
5,504
2,403
308
36,984
442
27,658
482
156,759
184
2,415
239
27,638
442
194,323
195
2,221
1,164
25,403
367
212
6,018
2,054
285
4,457
16,423
5,651
2,064
234
6,643
19,503
5,437
2,164
227
113
63
82
85
145
64
89
97
72
146
1W4M»~£^841
23
23
100
100
^298,662
23
100
29i;83~2“"287,255— 234,405
22
22
21
94
90
01
3blJ98
21
01
84,108
18
78
200,552
17
74
63,579
20
87
COUNTRY I PAYS
-J975_____ 1976
___ 1977
_J?78
1979
1980
1981
1982
Afghanistan
Bahrain/Bahrein
Cyprus/Chypre
Djibouti
Egypt/Egypte
Iran
Iraq
Jordan/Jordanie
Kuwait/Koweit
Lebanon/Liban
1,788
152
195
1,825
1,076
16,547
20.273
469
905
2.262
189
89
1,949
1,361
4,357
23,462
415
860
36,134
208
85
1,807
1,237
3,712
19,043
352
890
63,364
209
62
1,547
1,963
28,102
17,203
304
900
115,032
214
40
2,297
1,933
37,798
15,525
300
939
71,685
219
69
41,752
156
86
671
1,805
9,509
7,741
860
880
75
52,502
232
73
1,637
42,717
11,809
298
847
71,554
262
69
2,265
1,306
11,728
10,614
646
819
67
Libya/Libye
Morocco/Maroc
Oman
Pakistan
Qatar
Saudi Arabia/Arabie Saoudite
Somalia/Somalie
Sudan/Soudan
Syria/Syrie
Tunisia/Tunisie
Un. Arab Emirales/E. Arab Ums
Yemen/YAmen
1,245
17,768
6,894
92,687
257
38,012
1,136
24,772
3,078
95,930
220
58,506
911
25,889
3,513
66,083
160
31,117
837
28,757
2,528
88.652
147
18,584
1,211
27,754
1,542
2,587
389
14,063
766
24,795
2,396
263.842
162
12,808
718
24,878
1,872
316,340
257
10,956
481
28.637
928
324,576
213
8,263
512
28,095
897
326,492
172
8,529
20,866
1,645
2,959
528
12,641
32,971
1,689
2,504
522
27,627
47,431
1,908
2,316
638
17,088
Other territories:
UNRWA
Total
No. of countnos/No de peyj
% of countnos roporting
1,552
2.810
243
2,705
13,924
1,888
2,736
295
5,173
25,820
1,744
2,585
188
9,744
119
141
163
141
164
191
139
zai^ss-
300,846
517,650
549,806
531,&48
21
01
22
90
21
01
20
87
22
90
207,522
20
87
242,743
21
01
1,838
2,554
597
20,167
1,867
3,062
507
34,634
2,111
2,501
534
34,438
136
136
123
453,524------ 269,189“ ~2Mi.129
22
21
21
90
01
01
2.016
27,626
281
73,175
200
2,386
1994
37
3,311
3.911
443
237
940
30,316
304
2,518
2,023
37,516
2,565
9,899
5,127
2,376
426
11,464
o
£
§
EMRO94.WK4 02/08/96
Table/Tableau 4.2: Eastern Mediterranean Region - Case notification rates (per 100 000 populationyRiglon de la MdditerranAe Oriental® - Taux de notincation des cas (pour 100 000 habitants), 1975-1994
COUNTRY/PAYS
1975
Afghanistan
Bahrain/Bahrem
Cyprus/Chypre
Djibouti
Egypt/Egypte
Iran
Iraq
Jordan/Jordanie
Kuwait/Koweit
Lebanon/Liban
11.6
55 9
32 0
890 2
2.8
496
1840
18.0
89.9
Libya/Libye
Morocco/Maroc
Oman
Pakistan
Qatar
Saudi Arabia/Arabie Saoudite
Somalia/Somalie
Sudan/Soudan
Syria/Syrie
Tunisia/Tumsie
Un Arab Emirates/E Arab Ums
Yemen/Y6men
50 9
102.7
813 9
1240
150.3
524.2
Other territories:
UNRWA
Total
EMRO94.WK4 02/08/96
1977
1978
1979
1980
14 4
66.1
14.6
894.0
3.4
12.7
206.0
15.7
80.1
227.3
69.1
138
7789
3.0
105
161.7
13.1
77.6
394.3
65.9
10.0
626.3
4.7
76.9
141.3
11 1
73.5
713.1
64.5
6.4
870.1
4.5
100.0
123.4
10.7
72.2
446.3
63 1
11.0
44 5
139.9
346 6
125.3
121.5
764.4
34.2
143.0
375.3
842
83.8
384.5
26 3
130.9
229.3
319.0
76.1
141.2
23.6
128.4
170.0
370.9
112.2
114.1
115.1
19.5
47.6
57.6
159.0
176 5
19.4
39 2
51.4
336.1
246.7
21 2
35.3
57.9
201.4
11.4
12.6
8.7
20.9
50 1
48.1
38.7
84.3
246
476
49.4
71.9
151.5
22.0
439
26.8
131.1
30.1
155.3
255.6
110.2
73.5
2170
19.6
157.8
18.9
42.8
48.0
182.9
10.6
11.8
12.7
10.4
91.1
96.9
8979
111.1
190.0
3.7
108.8
90.8
10.2
61.6
196.0
___ .1984.
1985
.1986
1987
1988
1989,
1990
1991
1992
1993
4.1
19.2
48.6
24.6
54.7
106
127.0
52.0
5.9
406.8
3.2
22.4
46.0
18.3
49.7
15.3
740
469
9.2
578 5
2.6
17.8
42.3
20.1
41.7
72.8
100.1
36.4
7.1
447.0
2.4
15.8
432
15.0
33.4
85.5
1268
27.1
5.1
444 5
422
18.9
39 7
134
27.7
95.4
112.0
30.9
5.7
430.1
2.6
18.1
38.3
136
23.2
98.6
25.7
3.3
411 3
2.7
21.0
45.8
11.7
21.9
28.8
23.9
4.1
406.2
3.8
15.7
81.2
10.3
12.9
146.6
28.1
6.1
543 1
6.3
23.4
26.9
5.4
528.2
15.0
226
88
15.9
33.9
10 8
14.6
32.8
21.3
51
626.4
5.7
32.1
954
8.7
12.2
17.5
129.4
63.0
123.9
68.4
66.5
38.1
9.8
1045
63.2
92.8
61.5
59.8
35.5
9.8
110.3
29.8
170.2
50.7
164
32.4
3.0
43.0
29.8
21.7
190.8
6.0
112 3
28.5
144.4
41.2
16.7
15.5
2.9
46.2
30.3
19.0
340.8
23.9
100.0
19.2
21.1
35.3
40.6
369.7
8.1
119.1
40.2
162.8
600
21.5
88.8
3.5
38.6
29.9
54.4
96.6
5.1
1112
24.2
1546
388
13.5
19.4
44.3
40.6
383.3
7.0
123.5
86.5
140.1
57.0
27.7
38.2
11.2
368
33.5
32.2
84.7
9.7
113.7
27.5
128.6
37.9
15.0
19.7
37.9
50.6
230.2
8.6
122 8
61.6
108.7
69.8
31.4
34.6
7.0
20.9
34.6
41.2
75.4
0.9
48.7
25.4
17.1
39.4
65 0
44.2
25.0
13.6
56.0
75.2
41 1
25.7
12.8
8.3
7.8
6.8
6.1
3.3
4.1
4.3
7.0
2.6
3.9
3.7
56.9
68.0
8Z4
78.6
75.0
60.7
7»T
36.0
1982
J88L.
451.3
72.6
108
7525
2.9
28.7
79.0
21.0
56.8
2.5
268 8
41.7
13.4
209.0
39
22.2
55.8
26.3
585
2.8
346.5
59.9
11.2
15.1
144 3
80.1
367.8
85.2
81.3
15.4
138.2
73.8
356.8
62.8
79.4
181.9
160.8
90.2
12.0
1065
14.1
55.0
37.8
139
1994
5.0
585.0
6.3
8.5
145
322
114.5
146
14.4
223
140.8
29.9
75.0
527#
362
27.2
22 9
826
o
3
Table/Tableau 4.3: Eastern Mediterranean Region - New Smear Positive Cases/R6gion de la MdditerranAe Orientals* Frottls Positif Nouveaux
COUNTRY/PAYS
Afghanistan
Bahram/Bahrein
Cyprus/Chypre
Djibouti
Egypl/Egypte
Iran
Iraq
Jordan/Jordanie
Kuwait/Koweit
Lebanon/liban
Libya/Libye
Morocco/Maroc
Oman
Pakistan
Qatar
Saudi Arabia/Arabie Saoudite
Somalia/Somalie
Sudan/Soudan
Syna/Syne
Turusia/Tunisie
Un. Arab Emirates/E. Arab Unis
Yeman/Ydmen
No. of cases
1993
1994
82
10
1,668
5,240
173
148
13,168
123
11,020
10
1,743
1,811
161
155
148
14,650
135
Rate (per 100,000 pop)
1993
1994
15.3
1.4
299 5
26.9
3.5
8.3
50.8
6.2
8.3
1.4
308.0
2.9
3.1
9.5
5.1
55.3
6.5
4.7
800
Proportion of new smear ♦ cases
to all cases
1993
1994
72%
27%
48%
28%
41%
68%
48%
44%
15%
27%
53%
46%
36%
65%
16%
48%
44%
34%
1,168
12.9
58%
1,006
1,175
983
11.7
8.3
11.3
39%
23%
41%
2,896
3,351 3
21.9
24.2
29%
29%
36,334
25,490
9.«
15.3
18%
40%
Other territories:
UNRWA
Total
3 ndudes relapses
X
o
Mb
O'
§
EMR094.WK4 02/08/96
Table/Tableau 5.1: European Region - Number of tuberculosis cases notified/Rdgion de I'Europe - Noinbre de cas de tuberculose notifies, 1975-1995
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
.J?®5
1986
I?87
1988
1989
1990
1991
1992
1993
1994
Albania/Albania
Armenia/Armenia
Austria/Autriche
Azerbaijan/Azerbaijan
Belarus
Belgium/Belgique
Bosnia-Herz /Bosnie-Herz.
Bulgaria/Bulgane
Croatia/Croatie
Czech Republic/Rep tcheque
1,899
952
2,366
3,219
7,147
1,527
868
2,506
3,065
7,002
1,352
851
2,311
2,784
6,514
1,152
826
2,240
2,665
6,059
1,075
829
2,200
3,031
6,361
2,959
1,050
756
2,191
3,080
5,954
2,687
954
924
2,061
3,180
6,198
2,837
978
759
1,942
3,217
5,468
2,652
891
702
1,825
3,176
5,509
2,190
975
774
1,765
3,506
5,065
2,149
916
768
1,442
3,772
4,873
1,956
989
832
1,377
3,804
4,128
1,893
915
766
1,390
3,677
3,911
1,772
759
651
1,402
3,340
3,769
1,588
695
649
1,334
2,989
3,708
1,648
4,273
4,407
6,085
4,179
4,593
5,804
3,745
4,468
5,684
3,575
4,660
5,248
3,396
4,183
4,915
3,280
3,999
4,962
3,007
4,021
4,312
2,999
3,718
4,146
2,892
3,632
4,016
2,856
3,612
3,653
2,555
3,605
3,117
2,530
3,355
2,553
2,352
3,326
2,196
2,387
2,973
2,047
2,301
2,861
1,905
653
590
1,521
2,620
3,039
1,577
600
2,256
2,576
1,937
628
741
1,426
2,771
3,745
1,462
3,546
2,606
2,158
2,079
235
1,354
2,821
2,414
1,335
600
3,096
2,189
1,986
590
1,267
3,036
4,134
1,503
680
3,213
2,279
1,864
707
753
1,264
2,839
4,348
1,521
1,595
5,296
2,217
1,960
Denmark/Danemark
Estoma/E stoma
Fmland/Finlande
France
Georgia/Georgie
Germany/Allemagne
Greece/Grece
Hungary/Hongrie
Iceland/lslande
Ireland/lrlande
619
826
3,497
25,024
2,819
40,233
7,955
6,333
40
1,154
548
777
3,095
22,911
2,700
38,599
8,101
5,790
58
1,061
514
677
3,027
20,087
2,868
36,605
7,981
5,431
23
1,145
438
582
2,757
18,924
2,681
34,334
8,160
5,509
27
1,151
459
608
2,508
17,341
2,388
32,034
8,022
5,120
24
1,099
430
614
2,247
17,199
2,098
29,991
5,412
5,412
25
1,152
394
560
2,204
16,459
2,124
27,083
7,334
5,322
23
1,018
378
563
2,170
15,425
2,168
24,865
5,193
5,181
25
975
348
587
1,882
13,831
1,881
22.977
3,880
5,028
24
924
302
546
1,791
12,302
1,855
20,243
1,956
4,472
26
837
312
541
1,819
11,290
1,822
20,074
1,556
4,852
13
804
299
522
1,546
10,535
1,833
17,906
1,566
4,522
13
602
322
446
1,419
10,241
1.810
17,102
1,193
4,125
12
581
304
471
1,078
9,191
1,598
16.282
907
4,016
16
534
328
422
970
9,027
1,609
15,385
1,068
3,769
18
672
350
332
772
9,030
1,537
14,653
877
3,588
18
624
334
406
771
8,510
13,834
762
3,658
15
640
359
403
700
8,605
2,130
14,113
920
3,960
16
604
411
532
542
9,551
3,741
14,161
4,209
4,163
18
Israel/lsrael
Italy/ltalie
Kazakh stan/Ka sakhstan
Kyrgyzstan/Kyrgisizstan
Latvia/Lettome
Lithuania/Lituame
Luxembourg
Malta/Malte
Moldova
Monaco
416
4,070
16,135
2,037
1,532
2.236
104
53
3,921
0
306
4,782
15,179
1,985
1,427
2,042
226
38
3,946
0
266
4,128
14,914
2,110
1,313
1,916
107
28
3,399
0
239
4,063
14,910
1,964
1,225
1,693
62
24
3,275
1
242
3,936
14,255
1,915
1,167
1,610
98
31
3,033
0
249
3,311
14,442
1,973
1,194
1,636
71
24
2,781
1
227
3,182
13,876
2,085
1,140
1,599
45
26
2,852
0
232
3,850
13,808
2,051
1,077
1,495
41
13
3,197
0
222
4,253
13,357
1,981
1,072
1,477
41
24
2,858
0
257
4,008
12,563
2,022
1,054
1,420
46
15
2,554
0
368
4,136
12,423
2,094
1,223
1,453
42
11
2,732
1
239
4,037
13,090
2,122
982
1,412
45
14
3,022
2
184
3,839
13,286
2,088
948
1,372
48
14
2,810
2
226
3,262
13,501
2,159
938
1,339
16
12
2,510
1
160
4,068
13,307
2,132
857
1,381
45
16
2,281
1
234
4,185
10,969
2,306
906
1,471
48
13
1,728
1
505
4,147
10,821
2,515
943
1,556
48
26
1,910
0
345
4,685
10,920
2,582
955
1,598
25
30
1,835
1
419
4,734
10,425
2,427
994
1,895
35
26
2,426
395
5,816
10,519
2,726
1,470
2,135
33
Netherlands/Pays Bas
Norway/Norvdge
Poland/Pologne
Portugal
Romania/Roumame
Russia/Russie
San Marino/Samt Marin
Slovakia/Slovaquie
Slovenia/SlovGnie
Spain/Espagne
2.230
499
26,255
9,442
23,363
86,779
2,081
654
25,070
7,710
20,078
78,577
1,974
579
26,796
7,498
17,814
80,062
1,911
458
26,801
7,651
14,841
76,267
1,765
489
26,857
6,635
14,385
65,565
1,701
499
25,807
6,873
13,553
74,270
1,734
461
24,087
7,249
13,602
73,369
1,514
448
23,685
7,309
13,588
72,236
1,423
396
23,411
7,052
13,570
73,280
1,400
373
22,527
6,908
12,952
74,597
1,362
374
21,650
6,889
12,677
64,644
1,238
343
20,603
6,624
12,860
71,764
1,227
307
19,757
7,099
13,361
70,132
1,341
294
18,537
6,363
14,137
67,553
1,317
255
16,185
6,664
14,676
62,987
3,035
1,225
3,396
2,761
1,293
3,685
2,614
1,238
3,639
2,511
1,092
4,165
2,465
1,085
4,853
2,304
939
5,552
2,263
982
7,961
2,252
925
8,987
2,152
896
10,078
1,989
923
10,749
2,022
816
13,755
1,830
792
9,468
1,651
760
8,497
1,501
768
8,058
1,345
290
16,496
5,980
15,482
50,407
1
1,620
583
9,007
1,465
288
16,551
5,927
18,097
53,148
3,130
1,161
3,063
1,369
285
16,136
6,214
16,256
50,641
1
1,448
722
7,600
1,587
256
16.828
5,447
20,349
63,591
3
1,799
646
9,441
1,811
242
16,653
5,619
21,422
70,822
2
1.760
526
Sweden/Suede
Switzerland/Suisse
Tajikistan/T adjikistan
TFYRM*
Turkey/Turquie
Turkmenistan
Ukraine
United Kingdom/Royaume-Um
Uzbekistan/Ouzbekistan
Yugoslavia/Yougoslavie
1,446
2,091
2,746
1,307
1,823
3,395
1,105
1,648
2,963
1,127
1,575
2,880
991
1,447
2,910
926
1,160
2,647
875
1,193
2,631
784
1,167
2,628
832
1,097
2,509
754
946
2,427
702
961
2,485
640
881
2,610
545
1,018
2,727
536
1,201
2,474
595
1,104
2,621
557
1,278
2,460
521
1,134
2,116
1,827
29,600
11,781
10,723
13,540
1,918
29,112
11,156
9,555
13,427
100,808
1,795
27,111
11,204
9,518
12,932
39,927
1,713
27,073
10,722
9,767
12,426
36,716
1,677
26,095
10,488
9,163
11,561
39,992
1,625
25,646
9,290
9,682
11,754
26,457
1,559
24,710
8,436
8,697
12,106
616
930
652
1,712
537
924
892
728
20,314
1,922
31,835
12,620
11,195
15,092
28,634
1,541
24,216
7,814
8,817
11,744
27,589
1,604
24,356
7,026
8,544
12,119
30,960
1,607
24,058
6,666
8,717
11,876
31,029
1,614
22,946
6,341
9,427
11,720
30,531
1,956
22,145
5,732
9,794
11,526
27,884
1,904
20,744
5,793
10,134
10,534
26,669
2,169
20,182
6,059
10,632
9,977
24.468
2,325
16,465
5,908
9,414
8,477
25,166
2,358
16.713
6,088
2,751
19,964
6,481
9,774
3,843
20.622
6,196
14,890
4,502
610
987
1,671
1,602
25,455
2,074
18,140
6,411
9,370
3,771
14
14
7
2
7
14
10
23
9
24
2
21
2
24
3
12
4
15
10
292nO8“277,624’
268,041
243,067
232,398
248,482
49
94
51 '
98
49
94
241,779
239,825
50
96
45
87
43
83
COUNTRY/PAYS
GO
Other territories:
Andorra/Andorra
Liechtenstein
Total
No. countnes/No de pays
% countries reporting
EUR094 WK4 02/08/96
404,535
40
t8
358,937
347,594
432.814
355,279
349,774
347,046
325,123
319,990
309,875
45
87
45
87
46
88
47
90
48
92
48
92
48
92
48
92
48
92
299,861“ 303,510
48
92
48
92
48
92
48
92
1,733
640
9,703
494
645
539
9,093
12,982
2,626
1
I
O
3
Ox
Tabls/Tableau 5.2: European Region - Case notification rates (per 100 000 populationpRAgion de ('Europe - Taux de notification annuelles des cas (pour 100 000 habitants), 1975-1994
COUNTRY/PAYS
1975
1976
1977
1978
1®?9
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
Aibania/Albame
Armema/Arrndnie
Austria/Autriche
Azerbaijan/Azecbaidjan
Belarus
Belgium/Belgique
Bosnia-Herz./Bosnie-Herz.
Bulgaria/Bulgarie
Croatia/Croatie
Czech Republic/R6p. (cheque
78.3
33.7
31 2
566
76.3
61.7
30.1
33.1
53.0
74.3
53.6
29.0
30.5
47.4
68.8
44.8
27.7
29.6
44.7
63.7
41 0
27.4
29.1
50.0
66.5
30.0
39.3
24.6
29.0
50.0
61.8
27.3
35.0
29.7
27.3
50.8
64.0
28.8
35.2
24.1
25.8
50.6
56.1
26.9
31.4
22.1
242
49.2
56.1
22.2
33.6
24.1
23.4
53.4
51.2
21.8
30.9
23.8
19.1
566
48.9
198
32.6
25.6
18.2
56.3
41.2
19.2
29.5
23.4
18.3
53.7
38.8
17.9
23.9
19.8
18.4
48.1
37.2
16.0
21.5
19.6
17.4
42.5
36.4
16.6
49.0
103 4
60.9
47.7
107.2
57.7
42.6
103.7
56.2
406
107.6
51.5
38.4
96.1
48.0
37.0
91.4
48.3
338
91.4
41.8
33.7
84 1
40.2
32.4
81.8
38.9
31.9
81.1
35.4
28.5
80.6
30.2
28.2
74.8
24.8
26.2
74.0
21.3
26.5
66.0
19.9
256
634
18.5
19.9
17.6
19.7
36.8
29.8
15.8
13.9
25.1
57.0
18.8
188
21.8
18.4
38.4
36.6
14.6
85.2
29.1
47.8
20.2
6.8
17.3
38 7
23.6
13.3
15.2
34.7
48.5
19.3
16.9
16.1
41 1
40.6
15.0
18.3
36.2
50 5
18.1
20.7
21 2
16.0
38.0
42.8
15.1
45.2
60.1
49.2
19.0
Denmark/Danemark
Estonia/Estonie
Fmland/F inlands
France
Georgia/Gdorgie
Germany/Allemagne
Greece/Gr6ce
Hungary/Hongrie
Iceland/lslande
Ireland/lrlande
12.2
57 7
742
47.5
57.4
51 1
87.9
60.1
18.3
363
10.8
53 9
65.5
43.2
54.7
49.0
88.5
54.7
26.4
32.9
10.1
46.7
63.9
37.7
57.8
46.5
86.0
51.1
10.4
35.0
8.6
39.9
58.0
35.4
53.7
43.7
86.7
51.7
12.1
34.7
9.0
41.4
52.7
32.3
47.6
40.8
84.1
47.9
10.6
32.7
8.4
41.5
47.0
31.9
41.6
38.3
56.1
505
11.0
33.9
7.7
37.6
45.9
30.4
41.8
34.7
75.4
49.7
10.0
29.6
7.4
37.5
45.0
28.4
42.3
31.9
53.1
48.5
10.7
28.0
6.8
38.8
38.8
253
36.4
29.6
39.4
47.2
10.2
26.3
5.9
35.8
36.7
22.4
35.6
26.1
19.8
42.1
10.9
23.7
6.1
35.2
37.1
20.5
34 7
25.8
15.7
45.9
5.4
22.6
5.8
33.7
31.4
19.0
34.6
23.0
15.7
42.9
5.3
16.9
6.3
28.6
28 7
18.4
34 0
21.9
11.9
39.3
4.9
164
5.9
30.0
21 8
16.4
29.8
20.7
9.0
38.4
6.4
15.1
6.4
26.8
195
16.0
29.8
19.5
10.5
36.2
7.1
19.1
6.8
21.1
155
15.9
28.4
18.5
8.6
34.6
7.1
17.8
6.5
258
15.4
14.9
7.0
25.8
13.9
15.0
39.1
17.6
8.9
38.6
6.2
17.2
8.0
34.3
10 7
16.6
68.7
17.5
9.5
41.9
106
15.7
41.2
41.0
68
Israel/lsrafil
Italy/ltalie
Kazakhstan/Kasakhstan
Kyrgyzstan/Kyrgisizstan*
Latwa/Lettonie
Lithuania/Lituanie
Luxembourg
Malta/Malte
Moldova
Monaco
12.0
7.3
114.1
61 7
61 9
676
287
174
102.1
8.6
8.6
106.1
59.0
57.4
61.2
62.1
12.4
101.8
7.3
7.4
103.2
61.6
52.5
57.0
29.3
9.0
86.9
6.4
7.2
102.1
56.3
48.8
50.1
17.0
7.6
83.1
3.8
6.4
7.0
96.6
53.9
46.3
47.3
26.9
9.7
76.3
6.4
5.9
96.9
54.5
47.1
47.7
19.5
7.4
69.3
3.7
5.7
5.6
92.1
56.5
44.7
46.2
124
7.9
70.4
5.8
6.8
90.6
54.5
42.0
42.8
11.3
3.9
78.1
5.4
7.5
86.6
51.6
41.6
41.9
11.2
7.1
69.1
6.2
7.1
80.6
51.7
40.6
39.9
12.6
4.4
61.2
8.7
7.3
78.7
52.5
46.8
40.5
11.4
3.2
64.8
3.6
5.6
7.1
82.0
52.2
37.3
39.0
12.2
4.0
71.1
5.1
5.7
82.6
51.2
35.2
364
4.3
3.4
58.2
3.4
3.5
7.1
80.5
49.7
32.1
37.4
11.9
4.5
52.6
3.3
5.0
7.3
65.8
52.9
33.9
396
12.6
3.7
39.6
3.3
10.5
7.3
64.5
56.7
35.5
41.9
12.5
7.3
43.6
6.9
8.2
64.7
57.2
36.2
43 0
6.4
8.4
41.8
3.2
8.0
8.3
61 5
52.9
38 1
51 1
8.9
7.2
550
7.2
10.2
61.8
58.4
56.9
576
82
7.1
4.2
6.7
82.2
50.4
35.8
37.6
12.9
4.0
65.6
6.9
Netherlands/Pays Bas
Norway/Norvdge
Poland/Pologne
Portugal
Romania/Roumanie
Russia/Russie
San Manno/Samt Mann
Slovakia/Slovaquie*
Stovema/Slovdme
Spam/Espagne
163
12.5
77.2
103.8
110.0
64.6
15.1
16.2
73.0
83.8
93.6
58.2
14.2
143
77.4
80.3
82 3
58.9
13.7
11.3
76.7
80.5
67.9
55.8
12.6
12.0
76.2
68 7
65.3
47.6
12.0
12.2
72.5
70.4
61.0
53.6
12.2
11.2
67.1
73.7
60.9
52.6
10.6
10.9
65.3
73.9
60.5
51.5
9.9
9.6
64.0
71.2
60.2
51.9
9.7
9.0
61.0
69.7
57.2
52.5
9.4
9.0
58.2
69.6
55.8
45.2
8.5
8.2
55.0
66.9
66.3
49.8
8.4
7.3
52.4
71.7
58.2
48.3
9.1
7.0
49.0
64.3
61.2
46.1
8.9
6.0
42.6
67.4
63.3
42.8
66 1
666
86
63.4
69.6
9.4
57.1
72.7
10.1
53.5
68.9
9.9
49.5
59.2
12.9
45.9
50.9
14.7
44.8
52.9
20 9
44.3
49.6
23.5
42 1
47.8
26.3
38.7
49.1
27.9
39.1
43.2
35.6
35 3
41.8
24.4
31.7
39.9
21.8
28.7
40.2
20.6
8.9
6.8
43.2*
60.7
66.8
34.0
4.3
30.7
30 3
22.9
9.7
6.7
43.3
60.2
78.3
35.9
50 9
60.1
11.2
9.2
6.7
42.3
63 0
70.0
34.2
4.3
27.5
37.6
19.4
104
60
43.9
55.4
88.4
43.0
12.5
33 9
33 4
23.9
118
5.6
43.4
57.2
93.5
48.1
8.0 .
33.0
27.1
Sweden/Suede
Switzerland/Suisse
Tajikistan/Tadjikistan
TFYRM*
Turkey/Turquie
Turkmenistan
Ukraine
United Kmgdom/Royaume-Uni
Uzbekistan/Ouzbekistan
Yugoslavia/Yougoslavie*
176
330
79.8
159
288
95.8
13.4
26.0
81.3
13.6
24.9
76.9
11.9
22.9
75.6
11.1
18.4
66.9
10.5
18.8
64.7
9.4
18.3
62.8
10.0
17.1
583
9.0
14.6
54.8
8.4
14.7
54.5
7.6
13.4
55.6
6.5
15.3
56.4
6.3
17.9
49.6
7.0
16.3
51.1
6.5
18.7
46.5
6.1
16.4
38.9
7.1
13.2
11.3
80.8
6.1
13.0
15.0
34.0
508
76 3
649
22 4
80 1
166 1
70.6
60.1
209
746
1476
72.2
589
19.8
647
144.9
236.8
65.8
54.6
19.9
62.8
138.2
91.9
61.3
54 4
19.0
62.9
131.6
82.6
58.6
52.2
18.6
57.5
121.4
87.9
55.4
51.1
16.5
59.2
59.8
50.1
47.9
13.8
51.2
120.7
56.1
51.0
48.0
12.4
48.4
1238
61.5
49.8
47.3
11.8
48.1
120.6
60.2
488
44.9
12.1
50.8
118.4
58.0
57.7
43.2
10.1
51.5
115.9
51.8
54.7
40.3
10.2
52.0
105.5
48.5
60.8
391
10.6
53.3
99.2
43.6
63.6
31.9
10.3
46 1
83.5
44.0
63.0
32.4
10.6
43.8
70.2
38.7
11.2
44.7
362
40 1
10.7
66.6
122.5
56.7
51.9
49.1
15.0
51.8
125.2
7.1
14.1
29.8
76.5
43.6
54.1
35.1
11.1
43.8
36.1
26.9
38.5
11.5
7.4
25.9
50.0
35.7
44.2
31.0
44.4
6.9
36.2
6.9
36.9
56.0
24.0
143
24.6
56.0
44.2
43J-
40.5
39J
38.2
36.8
37.0
36.3
33.4
32.0
28.7
28.2
29.1
30.9
331F
Other territories:
Andorra/Andorre
Liechtenstein
Total
53.6
49.9
48.0
56.1
45.2
17.3
7.4
35.5
5.8
18.3
32.7
33.1
24.6
16.0
59.4
3.2
I
3
o
£
§
EURO94.WK4 02/08/96
Table/Tableau 6.3: European Region - New Smear Positive Cases/R6glon de ('Europe - Frottls Positlf Nouveaux
COUNTRY/PAYS
Albama/Albame
Armenia/Arm6nie
Austna/Autnche
Azerbaijan/Azerbaidjan
Belarus
Belgium/Belgique
Bosnia-Herz/Bosnie-Herz.
Bulgari a/Bulgane
Croatia/Croatie
Czech Republic/R6p tchdque
Denmark/Danemark
Estonia/Estome
Finland/Finlande
Franca
Georgia/Gdorgie
Germany/Allemagne
GreeceTGrdca
Hungary/Hongrie
Iceland/! slande
Ireland/lriande
Israel/! sraei
Italy/! talie
Kazakhstan/Kasakhstan
Kyrgyzstan/Kyrgisizstan
Latvia/Lettonie
Lithuania/Lituanie
Luxembourg
Malta/Malte
Moldova
Monaco
O
Netherlands/Pays Bas
Norway/NorvGge
Poland/Pologne
Portugal
Romania/Roumame
Russia/Russie
San Manno/Samt Marin
Slovakia/Slovaquie
Slovenia/SlovSme
Spain/Espagne
Sweden/Sudde
Switzertand/Suisse
Tajikistan/Tadjikistan
TFYRM*
Turkey/Turquie
Turkmenistan
Ukraine
United Kingdom/Royaume-Uni
Uzbekistan/Ouzbekistan
Y ugoslavia/Yougoslavie
Other territories:
Andorra/Andorre
Liechtenstein
Total'
No. of cases
1993
1994
Rate (per 100,000 pop)
Proportion of new smear ♦ cases
to all cases
1993
1994
499
1.493
484
513
1,775
427
6.8
14 7
4.8
6.9
175
4.2
16%
36%
32%
18%
41%
28%
1.147
548
3.096
1,083
524
25.4
5.3
35.1
24.0
5.1
50%
29%
58%
49%
27%
243
303
123
4,455
347
176 3
3,196
59%
57%
23%
47%
24%
54%
33%
35%
4.730
4.177 ’
58
5.1
33%
32%
1.905
1,357
6
18.7
13.4
2.3
45%
33%
33%
150
129
1,441
3,022
681
451
33%
25%
29%
25%
31%
3.207
470
688
13
615
120
704
9,339
2.3
22.5
3.5
5.5
2.9
2.4
2.5
36%
18.9
17.7
14.6
31%
180
185
17.5
47%
36%
3.6
14.0
15.9
86
4,000
2,072
10.385
30,389
19.9
50%
25%
27%
67%
7.0
1.063
7,606
47
19.5
2.4
7.7
2.0
10.4
45%
21.1
40.6
45.3
20.6
46%
36%
24%
37%
48%
43%
882
361
409
294
16.6
18.6
7.7
15.1
49%
56%
23%
56%
312
528
106
507
3.6
7.5
1.2
7.1
51%
57%
20%
55%
472
8.314
283
8.471
270
7,487
12.0
16.1
0.5
16.5
0.5
33 5
17%
42%
4%
41%
4%
50%
15
24
24.6
36.9
100% 4
100% *
50,248
88,294
6.4
12.1
21%
37%
* The Former Yugoslav Republic of Macedonla/Ex-Rtpublique yougoslave de Macedoine
3 includes relapses
4 only new smear + cases notiSed
EURO94 WK4 02/12^6
35%
42%
7.3
9.0
250
319
o
£
3
Table/Tabloau 6.1: South East Aslan Region - Number of tuberculosis cases notified/R^gion de I'Asle du Sud-Est - Nombre de cas de tuberculosa notifies, 1975-19M
1981 ____ 1982
1984
1985
1986
1987
1988
45,679
904
41,802
1,073
45,599
1,582
45,355
608
44,280
1,126
1990
48,673
1,154
1991
1992
1993
1994
56,052
996
31,400
140
54,001
108
48,276
1,159
1975
1976
J977
_ .1978_
1979
1980
Bangladesh
Bhutan/Bhoutan
DPR Korea/RPD Cor6e
India/lnde
Indonesia/lndonesie
Maldives
Myanmar
Nepal/Nepal
Sri Lanka
Thailand/Thai lands
11.549
6,517
25,943
26,941
976
34,377
1,159
39,774
1,539
42,644
2,657
675,508
17,402
173
10,585
589.768
19,913
185
9,935
1,235
6,823
2,195
610,531
19,374
139
10,156
1,975
5,994
7,066
668.794
19,517
126
10,215
959
6,360
10,584
662.600
21,924
106
11,846
1,264
6,152
13,297
705,600
25,235
73
12,744
1,020
6,212
45,704
Total
No. anntnos/No. da pays
% of countries rapoftng
722,541
e
60
636,571
s
80
681,178
744,472
752,725
837,901
8
80
0
90
0
90
0
90
923,095 1,075,098 1,109,310 1,168,804 1.279,536 1,403,122 1,457,288 1,510,500 1,519,182 1,555,353 1,121,120 1,081,279 1,114,374
769.540
49,647
62,966
98,458
469,832
74,470
97,505
105,516
31,809
32,432
17,681
16.750
32,461
33,000
249
175
92
123
152
115
85
203
143
123
91
111
111
112
19,009
15,583
14,905
17,000
12,416
10,840
11,986
9,348
10,940
11,045
10,506
11,012
12,461
12,069
13,161
15,572
8,983
11,003
1,603
10,142
190
252
1,012
700
52
337
1,459
4,026
6,573
6,372
6,174
6,666
6,596
6,411
6,092
6,429
6,376
5,889
7,334
6,666
6,288
47,767
47,697
49,668
43,858
44,553
51,835
50,021
46,510
69,240
77,611
52,152
48,553
65,413
49,452
815,952 1,076,211 1,244,819 1,275,299 1,323,509 1,413,418 1,520,444 1,667,348 1,735,860 1,719,365 2,156,276 1,319,933 1,286,940 1,298,999
90000080000800
0090909090908090909090809090
COUNTRY/PAYS
7,324
49,870
720
_1W3_
52,961
1,017
45,191
1,525
o
£
§
SEARO94 WK4 02X)8/96
Table/Tableau 6.2: South East Asian Region - Case notification rates (per 100,000 population)R6gion de I'Asie du Sud-East - Taux de notification des cas (pour 100 000 habitants), 1975-1994
1993
27.9
8.8
46.9
6.8
41.0
71.8
181 1
58.7
97 1
26.7
58.6
37.8
81 3
179.3
178 6
252 9
40 7
70 4
55 2
29.7
34 9
527
45.5
35.4
38.7
_837____ 2T9
126.8
522
39 8
389
22 8
83.7
1199
329
73 5
426
63.2
36.7
86 3
121.3
25 5
101 2
34.2
72.9
35.2
82J
136.2
132.4
97.9
93.7
94.4
1986
1?87
1988
1989
45.4
112 2
44.3
420
42.5
75.9
426
1006
45.0
74.7
1462
19.8
82 7
306
4.3
427
.1.82^
147.5
19.8
69 1
30.0
1.1
40.2
137.8
152.2
10.6
49.5
28.0
0.3
36.5
1518
163.1
9.8
58 7
28.3
1.4
40.3
1002
175.2
58 7
30.6
5.7
38.7
97.9
178.3
55.2
42.1
23 3
8.8
36.3
92.8
110.3
110.6
112.4
117.6
124.0
133.4
1??7
1978 .
1979
1980
151
8.3
320
32.2
82.2
40.0
95.6
45.1
124 4
108 8
12 8
126 3
348
930
14 4
131 2
31 9
9.2
49 3
101.2
135
846
31.5
6.8
44.4
23.7
98.2
14 8
688
35.8
8.7
42.2
29.1
102.4
16.7
46 2
37.7
6.9
41.9
97.8
109 3
211
68 7
36 1
2.2
41.7
128.3
210
661
342
9.3
47.8
_.L2_ _
94 3
13 7
95 9
32.0
14.4
426
162
103;8____ ioap_
Total
76.1
65.6
68.7
73.5
72.7
79.2
84.7
97.4
SEARO94 WK4 02/08/96
Jl??2
50.8
636
1985
424
78.0
ffl!
Bangladesh
Bhutan/Bhoutan
DPR Korea/RPD Corfie
India/lnde
Indonesia/lndonesie
Maldives
Myanmar
Nepal/Nepal
Sri Lanka
Thailand/Thai lande
538
__1??1.
.J9J!____ 1??2 .....1983______ 1984
47.2
53.9
56.0
47.3
55.9
77.3
67.2
210.4
COUNTRY / PAYS
162.9
1o
i
O'
§
Tabl«/Tabl*au C.3: South East Aslan Region - New Smear Positive Cases/Rtgion do i'Asie du Sud-Est - Frottis Positif Nouveaux
COUNTRY/PAYS
Bangladesh
Bhutan/Bhoutan
DPR Kwea/RPD Corte
Indiartnde
Indonesia/Indonesie
Maldives
Myanmar
NepaVN6pal
Sri Lanka
ThailancVThailande
Total
No. of cases
1W3
1994
Rata (per 100,000 pop)
1993 1994
Proportion of new smear ♦ cases
to all cases
19M
1993
16,993
1,710
352
16.5
1.5
21.8
35%
4%
30%
225,256
62,966
126
226,543
49,647
125
25.0
32.9
52.9
24.7
25.5
50.8
21%
100% 4
72%
20%
100% 4
50%
6,679
3,335
10,442
3,405
20,260
32.1
18.6
48.9
18.8
34 8
51%
51%
67%
53%
42%
317,355
312,484
23.1
22.7
25%
24%
4 only new smear ♦ cases nottJed
’ data for project areas only
LU
I
*
S
t nnanowv:
Table/Tableau 7.1: Western Pacific Region - Number of tuberculosis cases notified/Rdgion du Pacifique Occidental - Nombre de cas de tuberculose notifies, 1975-1994
COUNTRY / PAYS_____________
1975
Australia/Australia
Brunei Darussalam/Brundi Darus
Cambodia/Cam bodge
China/Chme
Cook Islands/lles Cook
Fiji/Fidji
Japan/Japon
Kiribati
Laos
Malaysia/Malaisie
1,347
Marshall Islands/lles Marshall
Micronesia/Microndsie
Mongolia/Mongolte
Nauru
New Zealand/Nouvelle-Zdlande
Niue
Palau
Papua N.Guinea/Pap.N.-Guin6e
Philippines
Republic of Korea/Rep Cor6e
Samoa
Smgapore/Singapour
Solomon Islands/lles Salomon
Tonga
Tuvalu
Vanuatu
Viet-Nam
Other territories:
American Samoa/Samoa am
French Polynesia/Polyndsie fran
Guam
Hong Kong
Macau/Macao
Mariana Islands/lles Mariannes
New Caledonia/Nouvelle-Caldd
Tokelau
Wallis & Futuna
Total
No countnes/No. de pays
X of countnos reportng
WPRO94 WK4 02/12/96
1976
1.387
336
_1977______ 1978
1,251
1,292
230
1879
1,542
216
1980
1,457
196
2,576
1981
1,386
285
1,980
10
180
65,867
187
55
268
108,088
279
1,806
11.692
97,924
278
1,132
11,098
15
257
89,245
97
999
10,264
36
187
80,629
40
1,028
10,441
30
205
76,455
11,094
37
210
70,916
146
7,630
11,218
0
17
8
6
4
6
7
1,075
7
608
0
7
2,212
107,108
153,334
36
1,103
4
595
0
14
2,446
118,587
107,819
59
1,123
2
542
0
9
2,232
108,813
81,910
58
1,161
0
474
1
17
2,525
112,307
89,803
59
1,094
2
448
0
10
2,508
116,821
98,532
49
20
259
1,116
1,122
663
611
0
5
1,782
148,057
121,735
79
16
20
133,537
121,735
51
3,097
261
90
214
19,514
12
134
49
8,192
1,383
163
0
4
413,786
28
78
2,813
307
67
17
214
56,272
2,760
355
79
5
150
170,878
2,964
411
89
15
131
68,659
12
7
8
110
95
78
64
6,623
1,017
46
7,928
1,088
67
7,191
155
155
0
34
44
454,905
30
83
120
108
0
6
94
2,800
455
71
7
184
11,821
2
81
71
7,903
442
2.710
266
64
33
178
43,062
2
76
55
8,065
1,101
10,970
99
86
12
67
1,340
8
437
2
17
2,742
104,715
100,878
43
15
73
1,512
0
415
3
14
2,955
106,300
91,572
41
12
75
15
66
2,992
0
359
0
26
3,453
151,028
87,169
43
1,651
0
404
1
20
3,505
151,863
85,669
37
6
66
41
7,729
6
65
49
7,527
233
75
120
0
5
8
78
48
7,301
455
74
171
0
17
12
80
54
7,843
671
5
78
37
7,545
571
64
585
26
24
86
127
4.700
11,634
1,088
238
10,145
226,899
36
230
58,567
103
4,258
10,569
1,952
377
49
32
124
46,941
23
356,482
33
92
4,706
11,944
1,299
256
10,241
151,564
20
165
61,521
111
6,528
10,577
2,143
337
54
9
188
43,875
1
308,245
31
193
29
185
62,021
1985
2,065
302
50
23
196
43,185
128
32
98,654
19
163
63.940
1,219
276
7,572
117,557
1984
2,179
324
45
12
173
51,206
108
0
404,689
1,270
245
8,158
1983
2,425
313
49
18
92
43,506
68
10
548,429
31
1982
1
355,345
33
92
461,572
30
100
462,193
30
100
58
144
0
14
541,001
30
100
104
2
14
615,179
30
100
1986
906
1987
1988
1989
952
___ 1990
1,016
143
6,501
375,481
1
226
51,821
68
1,826
11.702
950
954
126
10,691
304,639
20
162
54,357
1,514
10,735
907
189
9,106
251,600
16
173
56,496
110
3,468
11,068
208
7,279
10,944
53,112
121
2,952
10,686
37
60
2,818
8
320
5
13
2,877
153,129
88.789
65
32
98
2,432
6
296
0
38
2,251
163,740
87,419
29
11
77
2,541
8
295
7
68
2,237
0
303
3
17
4,261
183,113
74,460
29
3
3,396
217,272
70,012
37
2,497
317,008
63,904
44
6
3,401
207,371
57,864
44
1,666
372
14
24
118
52,463
1,617
488
36
26
144
52,270
1,591
382
23
23
140
50,203
1,841
309
20
30
230
59,784
13
63
41
7,021
320
27
111
5
73
75
6,704
274
28
128
0
30
9
59
3
49
212
10,325
265,095
17
199
56,690
129
1,760
292
35
27
131
47,557
8
85
49
7,432
420
16
98
0
1,616
334
24
22
90
55,505
9
80
34
7,269
389
56
74
9
34
651,853
35
97
655,019
30
100
1
1
716,450
3fl
100
128
7,906
310,607
1
218
10,903
345.000
8
247
50,612
91
1,951
11,059
26
367
1,577
7
350
1,611
348
0
335
6,510
343
28
143
1
741,916
893,992
35
97
32
89
__ 1992
1993
1994
1,011
180
16,148
320,426
12
240
48,956
100
994
11.420
991
160
1,020
13,270
344,218
6
183
48,461
99
2,093
12,285
15,172
363.804
4
280
44,590
253
1,135
11,708
52
111
61
26
274
1
25
7,451
178,134
46,999
49
173
1,730
4
352
2
41
5,335
180,044
38,155
45
1,778
364
29
30
193
56,594
1,830
367
33
28
114
52,994
1,677
332
23
19
152
51,763
1,502
317
2
4
2,540
236,172
48,070
151
1,433
1
4
4
82
89
70
6,537
285
94
6,319
184
1
22
83
60
6,545
294
67
140
1
4
760,914
3f
86
754,466
35
97
6,283
329
46
11
132
0
11
718,699
32
89
724,508
33
92
o
£
§
Table/Tableau 7.2: Waatern Pacific Region - Case notification rates (per 100 000 population)/R6gion du Pacifique Occidental - Taux de notification des cas (pour 100 000 habitants), 1975-1994
COUNTRY/PAYS
1975
1976
1977
J978
. 1979
1980
Australia/Australie
Brunei Darussalam/Brunei Darus
Cambodia/Cam bodge
Chma/Chine
Cook Islands/lles Cook
Fiji/Fidji
Japan/Japon
Kiribati
Laos
Malaysia/Malaisie
9.7
99
201.2
8.8
9.0
127.8
10.7
116.1
10.0
101 6
39.6
105.3
44 1
86.8
487.7
36.9
88.5
78 9
43.0
78.4
167 2
32 3
800
200.0
30.7
70.2
67 8
32.9
79.5
166.7
33.0
66.0
156.7
82.6
205.6
33.1
60.7
239.3
238 1
81.5
548
250
18.2
11.8
17.1
19.4
77.1
75.4
19.7
70.1
57.1
19.2
69 4
28.6
17.5
69.8
21.5
400.0
181.8
70 3
100.0
19.6
64.0
25.0
14.3
45.5
63 7
336.1
339.2
51.3
63.6
77.1
237.6
420 5
23.1
116.7
83.2
257.2
291 2
37.6
209.8
40.6
122 5
155.8
75.3
242.9
203.8
114.5
118.6
174.0
87.8
71.4
138.9
340.0
40.0
103.1
51.6
186.4
591.0
40.0
82.1
47.4
175.4
463.0
122.6
114.0
44.4
377.8
440.0
147J
158.1
187.6
Marshall Islands/lles Marshall
Micronesia/Microndsie
Mongolia/Mongolie
Nauru
New Zealand/Nouvelle-Zelande
Niue
Palau
Papua N Guinea/Pap.N.-Guinee
Philippines
Republic of Korea/Rep. Coree
Samoa
Singapore/S mgapour
Solomon Islands/lles Salomon
Tonga
Tuvalu
Vanuatu
Viet-Nam
289.5
46 5
969
507.3
59.7
95.4
310.5
3450
33.3
1369
137.4
102.3
... 1?81
9.4
1425
30.1
55.6
27.8
56.0
301.6
77.7
1982
1983
1984
1985______1986_____ 1987
1988
1989
1990
8.5
118.9
120.6
9.6
111.8
24.6
53.9
306.3
141.3
82.5
8.0
129.6
107.9
11.3
170.6
27.4
52.0
198.4
137.9
78.3
8.4
116.4
140.5
14.4
117.6
23.9
51.2
168.2
186.7
69.3
7.0
105.3
134.2
21.2
211.8
32.9
48.5
153.7
118.5
67.4
5.7
91.0
132.1
24.4
100.0
28.1
46.7
1897
40.9
66.7
5.6
79.1
112.8
22.8
94.1
24.3
46.3
1594
90.8
66.9
5.8
51.4
128.5
27.2
111.1
22.7
44.3
297.1
184.7
64.4
5.7
51 0
922
27.3
5.6
30.3
43.1
170.4
72.5
61.3
6.0
55.6
73.5
32.5
5.6
31.1
41.9
94 4
43.5
65.4
39.5
21.2
83.7
30.8
21.3
88.9
37.5
18.2
156.7
12.5
333
153.8
104.0
284.4
212.5
23.1
11.1
76.2
25.3
120.6
66.7
9.0
185.7
100.3
276.3
213.6
26.9
271.4
62.6
286.7
209.8
18.0
25.0
19.2
122.6
88 9
8.9
150.0
113.3
116.0
314.1
177.0
18.0
15.6
16.3
105.2
13.0
100.0
107.7
89.6
204.1
230.1
256
90.2
160
143.6
88.9
9.8
250.0
92.9
81.8
273.9
215.2
40.4
75.0
74.1
230.6
218.0
367
15.2
33.3
141.7
81.8
232.4
235.6
37.1
83.3
79.4
235.9
254.7
30.8
32.4
199
76.2
100.0
13.8
66.7
130.8
85.0
206.2
257.1
26.9
125.8
1948
97.8
214.3
118.0
133 6
1174
2078
78.0
100.0
161.4
22.5
112.2
117.2
69.6
412.5
152.1
80.2
99.3
133.2
53.3
225.0
76.7
79.2
88.1
133.3
49.5
150.0
140.7
91.2
82.6
119.8
54.9
287.5
155.6
75.3
84.7
129.1
59.3
112.5
145.7
74.8
76.3
139.6
53.8
400.0
93.9
78.4
68.0
104.7
38.0
337.5
96.3
77.7
61.8
115.6
25.8
275.0
64.7
88.8
23.3
68.8
67.0
154.6
51.1
258
54.9
62.7
138.3
430.9
6.3
55.1
68.3
160.7
184.9
6.3
50.3
51.4
160.1
455.0
112.3
77.1
75.5
17.1
40.6
43.8
143.8
91.4
468 8
81.6
22.2
47.3
42.1
137.3
172.3
462.5
114.0
31.6
47.3
46.2
145.5
245.8
341.2
94.7
209.1
41.7
141.7
116.7
12.8
44.8
31.1
138.3
201.1
336.8
67.1
100.0
116.7
19.5
47.5
40.2
134.7
142.9
69.6
62.4
60.0
48.2
500.0
9.1
18.2
42.3
37.6
150.3
235.9
162.5
88.3
50.0
218.2
136.3
103.3
114.0
113.1
34.2
33.8
‘39.0
43.7
45:6
_1?91 _ _. J992
5.5
5.8
66.9
171 8
119.6
27.1
29 5
63.2
42.1
33.6
32.2
39.4
40 8
133.3
123.0
22.2
45 0
60.8
60.3
J?93
1994
5.6
584
137.0
28.8
31.6
24.1
389
130.3
45.5
638
5.7
54.2
794
72.4
106 1
24.6
66.1
1220
32.8
61 8
9.9
7.9
50.0
156.3
181 3
274.9
1065
29.3
36.7
73.2
36 4
10.0
100.0
241.2
126.9
272.0
85.6
26.6
152.2
30.1
21.1
36.3
35 7
328.6
239
59.4
9.1
85.5
72.4
70.0
10.4
20.0
90.4
365.0
1649
22.8
65.0
521.6
149.1
27.2
37.5
86.6
334 0
133.7
26.8
9.2
100.0
25.0
63 2
3724
110.0
158
63.0
1244
149
300.0
83.1
82.2
604
1574
37.9
288.9
98.6
80.1
58.8
1194
24.0
255.6
94.0
75.3
67.3
93 4
206
333 3
150.3
87.7
643
106.1
29.9
333.3
122.9
81.2
65.5
103.7
33.7
311.1
70.8
74.3
59.4
90.7
23.5
211.1
92.1
71.0
21.4
43.7
27.2
130.5
127.5
200.0
46.3
450.0
261.5
29.5
33.5
32.0
124.9
100.9
79.4
68.5
50.0
7.7
111
37.8
57.3
118.4
830
71.8
77.6
19.1
29.9
6.3
24.3
114.1
100.3
65.1
85.1
50.0
109.4
92.4
7.8
389
48.6
112.5
74.2
7.5
41.4
639
108.2
230.8
108.2
500
157.1
2.0
40.3
42 9
113.3
79.5
1426
80 9
50.0
28.6
786
78.6
45.1
"48:6
49.6
49.5
48.5
45.7
45.5
to
Other territories:
American Samoa/Samoa am
French Polynesia/Polyndsie (ran.
Guam
Hong Kong
Macau/Macao
Mariana Islands/lles Mariannes
New Caledonia/Nouvelle-Caidd
Tokelau
Wallis & Futuna
Total
58.9
97.9
74.2
X
o
H
CO
£
WPR094 WK4 02/08/96
Table/Tableau 7.3: Western Pacific Region - New Smear Positive Cases/Rdglon du Paclfique Occidental - Frottis Posltif Nouveaux
COUNTRY/PAYS
No. of cases
1993_____ 1994
______
Australia/Australia
Brunei Darussalam/Brundi Darus
Cambodia/Cambodge
China/Chme
Cook Islands/lles Cook
Fiji^idji
Japan/Japon
Kiribati
Laos
Malaysia/Malaisie
Marshall Islands/lles Marshall
Micronesia/Micronesie
Mongolia/Mongolie
Nauru
New Zealand/Nouvelle-Zdlande
Niue
Palau
Papua N.Guinea/Pap.N.-Guinde
Philippines
Republic of Korea/Rep. Coree
Samoa
84,898
4
58
17,890
99
765
6,954
12
8
86
91
0
8
1,653
92,279
16,630
21
513
155
16
2
Singapore/Singapour
Solomon Islands/lles Salomon
Tonga
Tuvalu
Vanuatu
Viet-Nam
to
CA
557
68
36,534
Other territories:
American Samoa/Samoa am.
French Polynesia/Polyndsie fran.
Guam
Hong Kong
Macau/Macao
Manana Islands/lles Mariannes
New Caledonia/Nouvelle-Cal6d
Tokelau
Wallis & Futuna
1
39
11,058
104,729
1
60
16,770
184
752
6,861
15
145
2
61
0
11
573
87,401
13,266
18
7.1
21.1
7.7
14.4
130.3
16.6
36.1
24.0
17
3.7
110.9
8.7
5.3
7.8
13.4
239.0
15.9
34.8
20%
5%
6%
73%
29%
25%
21%
38%
73%
66%
59%
33%
64 7
13 6
132.0
29.8
10.7
32%
22%
52%
35%
43%
27%
11%
49%
35%
40%
30.5
31.1
17.3
11.1
37.6
49.1
28%
42%
48%
7%
51%
34%
74%
5%
41%
69%
2.6
50.0
402
142.4
37.7
126
861
114
17
1
62
35,813
184
43.8
16.3
22.2
4
38
40
2.0
18.5
7.5
17 7
27.2
69%
25%
48%
100% 4
43%
43%
37%
38%
41.8
28.1
2,429
108
25%
67%
32%
37%
100% 4
37%
57%
9%
8%
50%
17%
3.2
6.1
18.2
1.7
51.2
Proportion of new smear ♦ cases
to all cases
19S4
1993
o
56%
43%
3.2
248
48%
32%
468
23.6
2
22
42
0
2
14.3
14.3
18%
18%
"261,880
278;923
16.6
17.5
36%
38%
4 only new smear ♦ cases notified
WPRO94.WK4 02/08/96
Rate (per 100,000 pop)
1993
1994
§
Position: 2631 (2 views)