Efficacy of a 6-month versus 9-month intermittent treatment regimen in HIV-infected patients with tuberculosis

Rationale: The outcome of fully intermittent thrice-weekly antituberculosis treatment of various durations in HIV-associated tuberculosis is unclear. Objectives: To compare the efficacy of an intermittent 6-month regimen (Reg6M: 2EHRZ3/4HR3 [ethambutol, 1,200 mg; isoniazid, 600 mg; rifampicin, 450 or 600 mg depending on body weight <60 or ≥60 kg; and pyrazinamide, 1,500 mg for 2 mo; followed by 4 mo of isoniazid and rifampicin at the same doses]) versus a 9-month regimen (Reg9M: 2EHRZ3/7HR3) in HIV/tuberculosis (TB). Methods: HIV-infected patients with newly diagnosed pulmonary or extrapulmonary TB were randomly assigned to Reg6M (n = 167) or Reg9M (n = 160) and monitored by determination of clinical, immunological, and bacteriological parameters for 36 months. Primary outcomes included favorable responses at the end of treatment and recurrences during follow-up, whereas the secondary outcome was death. Intent-to-treat and on-treatment analyses were performed. All patients were antiretroviral treatment–naive during treatment. Measurements and Main Results: Of the patients, 70% had culture-positive pulmonary TB; the median viral load was 155,000 copies/ml and the CD4+ cell count was 160 cells/mm3. Favorable response to antituberculosis treatment was similar by intent to treat (Reg6M, 83% and Reg9M, 76%; P = not significant). Bacteriological recurrences occurred significantly more often in Reg6M than in Reg9M (15 vs. 7%; P < 0.05) although overall recurrences were not significantly different (Reg6M, 19% vs. Reg9M, 13%). By 36 months, 36% of patients undergoing Reg6M and 35% undergoing Reg9M had died, with no significant difference between regimens. All 19 patients who failed treatment developed acquired rifamycin resistance (ARR), the main risk factor being baseline isoniazid resistance. Conclusions: Among antiretroviral treatment–naive HIV-infected patients with TB, a 9-month regimen resulted in a similar outcome at the end of treatment but a significantly lower bacteriological recurrence rate compared with a 6-month thrice-weekly regimen. ARR was high with these intermittent regimens and neither mortality nor ARR was altered by lengthening TB treatment