In this prospective, multi-center observational study of HIV-infected TB patients, we found that patients who took ART had one-fifth the risk of dying than those who did not take ART, and that the risk of dying was further reduced with early ART initiation and concomitant use of fluconazole. The choice of non-nucleoside reverse transcriptase inhibitor (NNRTI) in the ART regimen did not alter the risk of adverse events or death.
Our study adds to previous observational data about the benefit of ART on survival during TB treatment, because it collected data prospectively, enrolled a diverse group of patients from multiple centers, and adjusted for a large number of factors associated with severity of illness, medication use, and treatment outcome. In addition to confirming the benefits of ART, we also found an independent survival benefit to using fluconazole prophylaxis during TB treatment, likely because fungal opportunistic infections are a major problem in Southeast Asia . Because over 80% of patients in this study met immunological criteria for AIDS, we expected that co-trimoxazole, which is internationally recommended as prophylaxis against Pneumocystis jirovecii pneumonia, toxoplasmosis, and other infections, would increase survival, but the point estimate was only statistically significant in the all patients analysis, not the subset of bacteriologic-confirmed patients . In Africa, some of co-trimoxazole's benefit has been ascribed to reducing malaria deaths . It is possible that the magnitude of benefit from co-trimoxazole is not as great among HIV-infected TB patients that are receiving ART and living in settings with less malaria, such as Thailand. It is also possible that patients did not initiate co-trimoxazole early enough to reduce HIV-associated deaths that occurred early during TB treatment, that antimicrobial resistance mitigated the benefit of co-trimoxazole, or that our study was under-powered to show a strong beneficial effect. Further prospective studies would help define the added value of co-trimoxazole and fluconazole to ART in reducing mortality during TB treatment.
International guidelines recommend that HIV-infected persons initiate treatment with nevirapine, zidovudine, and lamivudine; because rifampin can alter drug levels of nevirapine, guidelines recommend that efavirenz replace nevirapine in patients receiving rifampin . The high cost of switching from nevirapine to efavirenz has led some physicians in resource-limited settings to continue using nevirapine in HIV-infected TB patients, a decision backed by a pharmacokinetic study in Thailand that showed good HIV and TB treatment outcomes in patients receiving nevirapine and rifampin at the same time . Our study is unique, because it compared the frequency of adverse events and TB treatment outcomes in patients receiving either efavirenz or nevirapine containing regimens during TB treatment. In our study, adverse events and successful TB outcome occurred with similar frequency in both groups. Although nevirapine is frequently associated with rash, we found that rash was infrequent in both ART groups; in fact, patients not treated with ART reported rash more frequently, likely because of progressive cutaneous disease associated with declining CD4 count . We did not, however, serially evaluate HIV RNA viral load or CD4 to assess the benefits of nevirapine vs. efavirenz in achieving HIV treatment success. Current international guidelines also recommend that patients with advanced immune-suppression begin ART within two weeks to two months of beginning TB treatment . Although our study supports this recommendation, we were only able to demonstrate statistical significance for ART initiated within four months of TB treatment. The non-randomised design, relatively small sample size, lack of pharmacokinetic data, and lack of detailed efficacy and safety endpoints make our data about choice of NNRTI and timing of ART initiation hypothesis generating. Studies with larger sample sizes, controlled designs, and more extensive clinical monitoring data are needed to further assess the appropriateness of using nevirapine and rifampin concomitantly and to define the best time to initiate ART.
Our study is subject to important limitations. First, even though our study was derived from a diverse population base, some patients refused HIV testing, and our analysis was limited to a subset of HIV-infected TB patients, due to strict eligibility criteria and specific research questions. It is possible that patients included in our analysis differ in important ways from patients not included, such as those previously treated for TB, and that separate studies of these populations are needed. Second, patients only underwent three study visits, even though treatment was for a minimum of six months, and patients had limited laboratory monitoring. It is possible, therefore, that ascertainment of adverse events, e.g., hepatitis, was incomplete. Third, because this was an observational study, patients who developed adverse events underwent evaluation by their individual physicians. We do not have sufficient information to assign a specific toxicity grade to each adverse event.