Trends in tuberculosis and the influence of HIV infection in northern Malawi, 1988–2001

The HIV epidemic has had a major influence on tuberculosis (TB) epidemiology in sub-Saharan Africa. Notification rates of TB in the general population of most countries have risen sharply during the last two decades and TB is the leading cause of death among HIV-infected Africans [1,2]. The rate of increase in case-notification rates during the 1990s, and the potential for HIV-associated disease to generate secondary epidemics of TB led to suggestions that TB control measures were insufficient, and indeed that TB was out of control [3]. In this issue, however, Glynn and colleagues present an increasingly familiar and somewhat different picture. Their findings are that the increased TB incidence in northern Malawi has been essentially restricted to HIV-infected individuals, with TB incidence rates continuing to decline among non HIV-infected individuals [4]. Similar findings have been reported from Cote d'Ivoire, Northern Thailand, and South Africa [5–7], and there is even more direct evidence from Tanzania that the impact of HIV-infected TB cases on transmission of Mycobacterium tuberculosis may be limited [8]. This scenario is considerably less grave than one in which HIV-associated disease generates secondary epidemics, with increasing TB transmission rates increasing the incidence of TB disease in all sectors of society. Rising TB transmission rates and increasing TB incidence rates among HIV-negative individuals have, however, been reported from Kenya and South Africa [9,10]. What are the implications of these divergent findings for African countries heavily affected by HIV and TB? In the absence of HIV, the major control method for TB, promoted by the World Health Organization (WHO) as the DOTS strategy, is case detection and treatment, in particular of the most infectious, i.e., smear-positive cases. This achieves a dramatic reduction of mortality among cases and is also a highly effective prevention strategy, since prompt detection and cure of infectious patients minimizes onward M. tuberculosis transmission [11,12]. In the absence of case-finding and treatment programmes, most HIV-negative TB patients remain infectious for several years [13]. Additional methods include preventive therapy among infected contacts and Bacille Calmette–Guerin (BCG) vaccination. However, contact investigations tend to identify only a minority of infected contacts [14], and BCG, while reducing the risk of serious childhood tuberculosis, probably has little impact on the incidence of infectious, adult TB in high-burden countries [15]. HIV strongly increases the risk of TB among co-infected individuals. While this increased risk is most pronounced for less infectious types of TB (extrapulmonary and smear-negative pulmonary TB), the risk of smear-positive TB is greatly increased as well [16]. The impact on the point prevalence of infectious TB, however, may not be proportional. HIV-associated TB disease probably becomes symptomatic at a much faster rate than TB in HIV-negative patients and may be detected earlier [17,18]. Even in the absence of effective diagnosis and treatment, high rates of mortality may limit the potential for prolonged infectivity, although this is clearly not a desirable state of affairs [19]. In addition to their direct impact, there are other ways in which the greatly increased numbers of HIV-related TB cases may adversely impact TB control: the sheer volume of patients may overwhelm health systems, and the perceived association of TB and AIDS may demoralize health care workers and discourage health care seeking. Perhaps the increasing risk of M. tuberculosis infection observed in Kenya [9] is attributable to a combination of these effects. The first message for TB control in Africa, however, is one of increased hope. We now have good evidence from a number of diverse settings that application of the WHO DOTS strategy can prevent an increase of M. tuberculosis transmission even in high HIV prevalence areas [4–8]. This is an important message because routinely reported data of National anti-TB programmes in countries severely affected by HIV suggest that DOTS implementation needs to be strengthened in most of these. Of 25 countries with an estimated adult HIV prevalence of 5% or more, only 10 reported national DOTS coverage, and of these only two (Tanzania and Malawi) reported less than 15% of the treatment outcomes that have been particularly associated with suboptimal control of onward TB transmission, namely failure, defaulting, or transferred out [2]. Strengthening DOTS programmes in Africa therefore remains an important priority. Second, even if DOTS can prevent an increase of M. tuberculosis transmission, more is required to reduce TB disease incidence and mortality [20]. Options for achieving these aims with available technologies include: intensifying case finding in order to reduce the average infectious period and tuberculosis-associated mortality; intensifying the uptake of VCT to identify a higher proportion of HIV-infected persons before they present with tuberculosis; provision of effective HIV care once HIV status is known, including: isoniazid preventive therapy to reduce the incidence of TB among HIV-positive persons with latent TB infection [21,22]; isoniazid prophylaxis for HIV-infected individuals after completion of TB treatment to reduce the risk of recurrent TB [23]; cotrimoxazole prophylaxis for HIV-infected individuals to reduce mortality due to other opportunistic infections [24]; highly active anti-retroviral therapy to reduce HIV-associated mortality and reduce TB incidence [25–27]. A choice between these additional options may be guided by their effectiveness, cost-effectiveness and feasibility. Mathematical models agree that TB case finding and treatment is important [26–28], but vary in their assessment of the effectiveness of the additional interventions. Issues to be considered include the time scale (long for HIV prevention, short for TB and HIV treatment), outcome measures (TB incidence or all-cause mortality), and operational aspects of delivering joint HIV/TB interventions. Because evidence is limited, studies to evaluate their impact on TB incidence and all-cause mortality at the population level are of high priority.