Treatment of tuberculosis in HIV-infected persons in the era of highly active antiretroviral therapy

Objective: To assess the risks and benefits of administering highly active antiretroviral therapy (HAART) during the treatment of tuberculosis (TB) in HIV-infected patients. Design and methods: HIV-1 patients presenting to 12 HIV centres in Greater London and south-east England with culture-proven TB were identified from January 1996 to June 1999. Case-notes were reviewed retrospectively. Results: Patients (n = 188) were severely immunocompromised with a median CD4 cell count at TB diagnosis of 90 x 10(6) cells/l (IQR: 30-180). At presentation, 85% (n = 159) were not taking antiretrovirals. A total of 45% commenced HAART during TB treatment, which was associated with significant reductions in viral load, AIDS-defining illness (ADI) [3.5 versus 24.5%; relative risk (RR) = 0.14] and mortality. Only nine of 91 (10%) patients with a CD4 count > 100 x 10(6) cells/l at TB diagnosis experienced a further ADI, whereas 18 of 92 (20%) patients with a CD4 count < 100 x 10(6) cells/l developed this complication. Adverse events (AE) occurred in 99 (54%) of 183 patients, one-third of whom changed or interrupted HIV and/or TB medication. The majority of AE occurred within the first 2 months, with peripheral neuropathy (21%), rash (17%) and gastrointestinal upset (10%) occurring most commonly. Conclusions: Many physicians delay HAART in patients presenting with TB because of pill burden, drug/drug interactions and toxicity. Although the use of HAART led to significant reductions in viral load, ADI and mortality, co-infected patients commonly experienced AE leading to interruptions in TB/HIV therapy. We therefore recommend starting HAART early for patients with advanced HIV disease (CD4 < 100 x 10(6) cells/l) and deferring HAART until the continuation phase of TB therapy (i.e. after 2 months) for patients who are clinically stable (CD4 > 100 x 10(6) cells/l). (C) 2002 Lippincott Williams Wilkins.