Setting: Studies on tuberculosis (TB) relapse in HIV-infected patients show contradictory results regarding the optimal duration of treatment.
Objective: To assess the incidence of TB relapse and associated factors in HIV-infected patients receiving a 9-month tuberculostatic regimen and concomitant HAART.
Patients and methods: Observational prospective study recording 156 episodes of TB in 137 patients, most of whom were on a 9-month regimen of daily isoniazid and rifampicin-based TB treatment. The primary outcome measure was relapse after completion of therapy.
Results: Forty episodes were excluded due to death or loss to follow-up. The median follow-up was 24 months. Twenty-seven episodes of TB relapse were observed in 22 patients, yielding a relapse rate of 1.9/100 patient-years in those on a regimen of > or = 9 months. A high recurrence rate was observed in those who had prematurely suspended treatment. Treatment duration > or = 9 months and achieving both an undetectable viral load and increasing CD4-cell counts with HAART were associated with the absence of TB relapses.
Conclusions: Considering its safety and tolerance, our results suggest that a 9-month regimen would be recommendable in patients with severe immunosuppression until the optimal duration of TB treatment in HIV-infected patients has been defined in a randomised clinical trial including HAART.