WHO report 2008
Global tuberculosis control
1.8 Progress towards impact targets included in the Millennium Development Goals
1.8.1 Trends in incidence, prevalence and mortality
With the 9.2 million new incident TB cases in 2006, there were an estimated 14.4 million prevalent cases (219/100 000) on average (Table 1.2). An estimated 1.7 million people (25/100 000) died from TB in 2006, including those coinfected with HIV (231 000). The sequence of annual estimates up to 2006 suggests (as in the data up to 2005) that all three major indicators of impact – incidence, prevalence and mortality per 100 000 population – are falling globally. In our assessment, prevalence was already in decline by 1990, mortality peaked before the year 2000 and incidence began to fall in 2003 (Figure 1.20). TB prevalence continued to fall globally between 1990 and 2006 because, in Africa, the HIV epidemic caused a smaller increase in prevalence than in incidence or mortality.
The fall in the global incidence rate reinforces data presented in Global Tuberculosis Control 2007. If verified by further monitoring, the data show that MDG 6 Target 6.C was met by 2004, well ahead of the target date of 2015 (though as noted above, the total number of new cases continues to rise, due to population growth in the African, Eastern Mediterranean, European and South-East Asia regions). This turnover of the global epidemic is largely explained by stable or falling HIV prevalence in Africa and by the stabilization of TB incidence in the independent states that emerged from the dissolution of the Union of Soviet Socialist Republics. It is unlikely that either of these two phenomena is due primarily to the implementation of HIV/AIDS or TB control programmes (see next section 1.8.2 on determinants of TB dynamics), and there is little evidence, from regional trends in case notifications, that DOTS is accelerating the decline of the incidence of TB on a large scale in Asia.
The targets related to reductions in prevalence and deaths that have been set by the Stop TB Partnership – to halve 1990 prevalence and death rates by 2015 – are more demanding. If the estimated changes between 2001 and 2006 are correct, and if the average rates of change over this period persist, then prevalence and deaths per capita will fall quickly enough to meet the 2015 targets in the Region of the Americas and in the Eastern Mediterranean, South-East Asia and Western Pacific regions (Figure 1.21). They will not, however, be met in the African and European regions. In line with the trends in incidence (Figure 1.6), prevalence and death rates increased in the African and European regions between 1990 and 2006, most dramatically in Africa. For this reason, estimates for these two regions in 2006 are very much larger than the 2015 target values.
Based on progress between 2001 and 2006, and combining the results for all regions, the mortality and prevalence targets are unlikely to be met worldwide by 2015 (Figure 1.21).
1.8.2 Determinants of TB dynamics: comparisons among countries
A further assessment of the scale of the impact of DOTS around the world can be made by examining the national statistics that lie behind the regional and global summaries. The series of cases reported by 134 countries between 1997 and 2006 indicate that TB incidence rates per capita in most countries were changing at between –10% and +10% annually between 1997 and 2006, and falling slowly in the majority of these countries (Figure 1.6 and 1.7). It is possible that these variable rates of decline are attributable to the uneven success of TB control programmes. Alternatively, the differences among countries might be explained by other factors that affect transmission of and susceptibility to disease.
One way to distinguish between possible explanations is to identify, by comparing countries, which factors are more or less closely associated with changes in TB incidence. In an preliminary ecological analysis1 of 30 possible explanatory variables (for methods, see Annex 2), trends in incidence per 100 000 population in the Latin America and Caribbean subregion are associated (p <0.05) with HIV prevalence (r2 = 0.41, Figure 1.22a), with under-5 mortality (r2 = 0.32), and with access to clean water (r2 = 0.43) and adequate sanitation (r2 = 0.50), among other variables. In the high-income countries of Western Europe and the United States of America, immigration is the single most important factor associated with TB dynamics (Figure 1.22b). In Central and Eastern Europe and in the Eastern Mediterranean Region, TB trends are linked to a variety of economic indicators including health expenditure per capita (Figure 1.22c) and expenditure in relation to GDP (Figure 1.22d). Only three of seven direct measures of TB control were significantly associated with trends in TB incidence, and the form of the association does not suggest any causal link. For example, smear-positive treatment success under DOTS (r2=0.29), and the product of case detction (all forms of TB) and treatment success (r2=0.32), were inversely correlated with TB decline in high-income countries.
In multivariate analyses of this kind, the numerous explanatory variables tend to be inter-related, and some are more obviously linked to TB trends as covariates, rather than as primary epidemiological determinants. For example, in the African Region incidence was increasing more quickly in countries that spent more on TB control (r2 = 0.49, Figure 1.22e). The likely explanation lies in the association between expenditure on TB per capita and HIV prevalence, with richer African countries that can spend more on health care also having higher HIV prevalence (r2 = 0.53). Similarly, the decline in TB incidence in Central and Eastern Europe tends to be faster in countries where a higher proportion of women smoke (r2 = 0.67, Figure 1.22f). The likely explanation is that smoking among women reflects affluence, which is linked to health and health services in ways that outweigh the importance of smoking as a risk factor for TB (correlation with GDP, r2 = 0.67).
In brief, this ecological analysis provides no evidence that the standard, direct measures of DOTS implementation - case detection and treatment success in various combinations - can yet explain the variation in incidence trends among countries, despite the wide variation in DOTS implementation among countries. This observation suggests - subject to further investigation - that DOTS programmes have not yet had a major impact on TB transmission and incidence around the world.
All of the caveats attached to this proposition must be carefully examined before drawing firm conclusions. Key assumptions to be tested are that trends in case notifications reflect trends in TB incidence, and that there is measurable and meaningful variation among countries in incidence trends and their determinants. It is also possible that DOTS programmes have significantly cut transmission, but it is too soon to see the effects on incidence, or that the effects have been offset by the rise of other risk factors, such as diabetes. In addition, it is crucial to distinguish the well established effects of DOTS on treatment outcome and mortality from the possible effects on transmission (under investigation here).
1 A fuller analysis is in: Dye C et al, Determinants of trends in tuberculosis incidence: an ecologic analysis for 134 countries. Unpublished paper available from the authors.