Thwarting Tuberculosis Prevention in High-Burden Settings

The promise of chemoprophylaxis for tuberculosis has yet to be fully realized. Until recently, testing for and treating latent tuberculosis infection have been limited largely to low-burden settings, where active tuberculosis has been under good control. In the United States and several other low-burden countries, treatment of latent tuberculosis infection has been an alternative prevention strategy to early childhood immunization with bacille Calmette–Guérin (BCG). In high-burden settings, vaccination and the treatment of patients with positive smears have been the main public health strategies, with chemoprophylaxis recommended primarily for childhood contacts of patients with active tuberculosis, albeit it has been administered inconsistently. However, the emergence of HIV as the primary factor driving high tuberculosis rates in many parts of the world has led to recommendations by the World Health Organization (WHO) to administer isoniazid preventive therapy in coinfected adults and children as part of the three I's strategy — the other I's being intensified case finding and infection control.               Even healthy newborns are highly susceptible to tuberculosis infection and progression to disseminated disease, including meningitis. The added risk of HIV coinfection leads to substantial morbidity and mortality during the first 2 years of life. Although immunization with BCG has been shown to reduce serious extra pulmonary complications, protection is incomplete, as indicated by the rates of illness and death from tuberculosis, despite almost universal vaccine access in high-risk settings. It is in this context that the failure of isoniazid primary chemoprophylaxis to reduce the risk of either tuberculosis or death for 2 years after randomization is especially disappointing. The authors carefully review possible explanations for this failure. In theory, continuous isoniazid treatment at a proper dose, in patients who adhere to the treatment regimen, with appropriate blood levels achieved, should prevent infection and disease from isoniazid-susceptible Mycobacterium tuberculosis, even in immunocompromised persons, and this has been shown in other studies.4 Among the explanations offered for the lack of protection in the study by Madhi et al., the difficulty of diagnosing tuberculosis in young children, with resulting overdiagnosis on the basis of diagnostic algorithms, seems the most probable source of error, but as noted by the authors, the absence of a difference between treated and untreated bacteriologically proven cases raises doubts about that explanation as well. Whatever the cause of the apparent failure of isoniazid to prevent tuberculosis in this study, one thing is certain: the cases that occurred among these children in their first 2 years of life must have been transmitted recently, most likely from the community. The study by Martinson et al. is a more conventional comparison of four secondary prophylaxis regimens among persons with tuberculosis and HIV infection in a high-risk South African setting. Here the new regimens of rifapentine–isoniazid weekly for 3 months, rifampin–isoniazid twice weekly for 3 months, and continuous isoniazid therapy for up to 6 years were compared with 6 months of conventional therapy. On the basis of expected rates of tuberculosis in this population, all four regimens were effective, but rates of active tuberculosis or death were no different with the two new, supervised, rifamycin-containing regimens than with the conventional 6-month isoniazid regimen. However, in a post hoc, as-treated analysis, patients in the continuous-isoniazid group had a 58% lower rate of tuberculosis or death than those receiving the 6-month control regimen of isoniazid, but the rates of tuberculosis in the continuous-isoniazid group markedly increased when therapy was discontinued, which was more common than with the other regimens, probably because of more severe side effects. These findings are consistent with those of the Botswana trial of continuous isoniazid and suggest ongoing transmission and reinfection in this high-prevalence setting, a phenomenon that is likely to compromise the long-term benefit of any chemoprophylactic regimen, regardless of short-term efficacy.5 What is thwarting tuberculosis prevention in high-burden settings? There are probably several factors, but a fundamental one, not fully appreciated, is ongoing transmission and reinfection. Although exogenous reinfection can be assumed to have occurred in patients after continuous isoniazid therapy has been stopped, proving reinfection is difficult, because genotyping of initial and subsequent M. tuberculosis isolates is rarely possible. Despite immunization at birth, by early adulthood most young adults in high-risk settings have been exposed to tuberculosis and infected. Epidemiologic considerations suggest that reinfection routinely occurs in these settings, even in HIV-negative populations.6-8 Reinfection has been postulated as an essential pathogenic pathway to lung cavitation and pathogen propagation in populations where there is partial immunity from prior vaccination or tuberculosis infection early in life.9,10 If this is true, the long-term benefits of chemoprophylaxis are likely to be limited, and if natural infection is not protective, the development of a more effective vaccine will be challenging. Both the risk of tuberculosis in young children and the risk of reinfection when chemoprophylaxis ends in adults point to the need for better control of transmission in the community and in congregate settings, such as clinics, hospitals, and prisons. This can best be achieved by intensified case finding, rapid diagnosis, and prompt institution of effective therapy — fully supervised and based on rapid drug-susceptibility testing. The feasibility and effectiveness of intensified case finding in the community have been shown.11 The Xpert MTB/RIF technology recently endorsed by the WHO will provide an opportunity to screen patients with cough and other risk factors, to establish a precise diagnosis within hours, and to institute effective treatment within days, assuming that treatment programs are in place.12 The impact of effective treatment of tuberculosis transmission is both rapid and profound, preceding sputum-smear and culture conversion, and is the cornerstone of tuberculosis infection control.13 In the near future, it should not be necessary, as it often is today, to wait for weeks or months to diagnose tuberculosis or identify drug resistance, during which time transmission continues. Unless the force of transmission can be reduced by intensified case finding and the use of new rapid diagnostics, resulting in more effective treatment, durable benefits from prevention strategies, either chemoprophylaxis or immunization, are likely to be elusive.

Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.

Source Information

From the Division of Global Health Equity, Brigham and Women's Hospital, and Harvard Medical School — both in Boston (E.N.); the Aurum Institute for Health Research, Johannesburg (G.C.); and the London School of Hygiene and Tropical Medicine, London (G.C.).