Treatment of tuberculosis in HIV-infected individuals.

Abstract

We were very interested to read the article by Dean et al. [1] on the treatment of tuberculosis (TB) in HIV-infected patients. Additional data shedding light on this complex and important problem [2] are very welcome. The authors recommend starting highly active antiretroviral therapy (HAART) early (within the first 2 months of TB therapy) in patients with a CD4 cell count of less than 100 × 106/l. However, the basis for such a recommendation is not clear from the data presented. The critical aspect of the study design is the separation of patients. Two major groups are compared: those who received no HAART during TB therapy, and those who had HAART concomitant with TB therapy. The latter group included a number of patients (18%) who were already on HAART at the time of the diagnosis of TB, and all patients starting HAART at any time between 0 and 14 months after starting TB treatment. The comparison is thus between no HAART and HAART at any time after the start of TB therapy. The use of HAART at any time was associated with fewer deaths and AIDS-defining illnesses, but such an observation does not allow one to reach conclusions about the optimum timing of the initiation of HAART. The key question for most clinicians, as the authors’ recommendations indicate, is not whether any HAART at any time is useful, but whether early antiretroviral treatment is better than delayed treatment. Because the median time of the initiation of HAART was in fact 2 months, it would have been interesting to compare patients starting therapy within 2 months with those who started treatment after the induction phase, and to separate out those patients already on antiretroviral therapy when TB developed. The latter group present different issues from patients not known to be HIV infected when TB develops, and are most usefully considered separately. Physicians with experience in the field are only too aware of the problems of HAART in the first 2 months of quadruple TB treatment. Confirmation of the extraordinarily high rate of adverse drug events (54% of patients), many requiring an interruption of therapy (34% of patients), is a valuable contribution of the study. The adverse events occurred predominantly in the first 2 months of TB therapy, and were nearly twice as likely in those on concomitant HAART. A further potential disadvantage of early HAART is the possibility of precipitating more frequent or severe immune reconstitution reactions [3,4]. Finally and importantly, because of drug interactions and the increased pill burden (especially with antiretroviral therapy in the first 2 months of TB treatment) that may affect adherence adversely, concerns remain over the long-term effects of concurrent HIV and TB therapy on viral suppression and TB relapse rates. As stated, such long-term effects could not be assessed in this study. Of the potential benefits of early HAART, clearly an associated reduction in mortality, if causal, overrides all other considerations. However, any association between the use of HAART and improved outcome in an uncontrolled retrospective review needs to be interpreted with caution. There was little discussion of this critical issue, and it would be very useful to see data on the timing and probable causes of death. Given the fact that the median time to the initiation of HAART was 2 months, if a significant number of these deaths occurred early, the apparent association of HAART with survival may have resulted from selection bias, with survivors having a greater chance of going on to receive HAART. Regarding the number of further AIDS-defining illnesses, it would be useful to know what these were. Of note is the fact that two-thirds occurred after the first 2 months and none occurred in the first month. The number of such episodes prevented by early treatment may therefore be small. With an increasing understanding of the drug interactions involved in concomitant TB and HIV therapy [2], and the publication of some outcome data [5], starting antiretroviral therapy in the continuation phase of TB treatment has become easier. The study suggests that such an approach is associated with a reduction in AIDS-defining illnesses in the short term. The development of simpler antiretroviral regimens and TB treatment by directly observed therapy also makes HIV therapy in the first 2 months of TB treatment a realistic option. However, many of the potential problems with concurrent therapy apply particularly to this period. We need additional data, ideally from controlled prospective studies with adequate follow-up, to determine whether HAART in the first 2 months of TB treatment is in our patients’ best interests. Thomas. S. Harrisona Derek C. Macallana Charlotte F. J. Raynerb Mark Wansbrough-Jonesa