Efficacy of Thrice Weekly Directly Observed Treatment, Short-course (DOTS) in HIV-associated Tuberculosis

Several reports have suggested that the initial response to antituberculosis chemotherapy is comparable among HIV-positive and negative populations. These series generally demonstrate that among "surviving" Patients, the bacteriological, clinical, and radiographic responses are similar between the two groups. However, there are consistent indications of higher rates of early (first month) deaths from tuberculosis as well as excessive deaths from other causes during the course of treatment in the above noted series. These deaths appear related to the advanced stage of tuberculosis at diagnosis as well as the debilitating and underlying diseased from which the patients suffer and not primarily the drug regimens with which they are treated. However, several of the reports from Africa suggested increased early mortality in those who received the less-potent traditional isoniazid, thiacetazone, and streptomycin regimens than the modern short-course regimens featuring isoniazid, rifampin, and pyrazinamide. Excess mortality was seen also among a subset of patients with AIDS and tuberculosis in Uganda receiving a thiacetazone regimen in comparison to those receiving a rifampin regimen: there was both excess mortality and higher rates of drug reactions sequestered among those patients who had elevated levels of neopterin and other markers of cellular immune activation.

Furthermore, several series have shown a modestly greater risk for relapse or recurrence post treatment for persons with AIDS that seems related to the duration of therapy. Perriens and colleagues in a study from Zaire compared the outcome of HIV-positive patients treated with 6-month regimen (2-HRZE daily followed by 10-HR twice weekly) and a 12-month regimen (2-HRZE daily followed by 10-HR twice weekly). Relapse rates were significantly higher among those receiving the 6-months regimen (9%) than 12 months of treatment (1.9%) (p < 0.01). Pulido and colleagues in Spain observed in a non-randomize series that, among patients with AIDS and tuberculosis, 10 of 40 (24%) patients who received less than 9-months or more did so. Multivariate analysis identified duration of therapy as a major element in the disparate relapse rates, with a relative hazard of 9.2 for the shorter-duration therapy. Most recently, a multicenter national trial in the United States compared 6-month and 9-month of treatment for HIV-infected adults with pansusceptible tuberculosis. Relapse rates were 3.9%, two patients, for the 6-month regimen, and 2%, one patient, for the 9-month regimen; because of the limited number, there was no statistically significant difference.

Several other studies contrasted relapse or cure rates among HIV infected and uninfected persons treated simultaneously with identical 6-month regimens. They universally showed somewhat worse outcomes among those with HIV infection.

Hawkens and colleagues described and increased risk of recurrent tuberculosis in a group of patients from Kenya. This report documented that 10 of 58 (17%) HIV Positive patients available for follow-up had recurrence, contrasted with 1 of 138 HIV negative patients, 34-fold apparent relative risk. However, 7 of the 10 who experienced recurrence had major cutaneous drug reactions, interrupting therapy and confounding the issue. But, Elliott in Zambia noted a marked disparity in relapse rates without the confusing association between relapses and drug reactions: HIV-positive patients relapsed at a rate 22-100 patient years of observation versus 6/100 patient years among HIV-negatives. A recent Johns Hopkins study in Haiti found lower cure rates (69% vs. 79%) and slightly higher relapse rates (5.4% vs. 2.8%, p = 0.36) among HIV-infected individuals receiving a 6-month regimen.