Combined toxicity of zidovudine and antituberculosis chemotherapy.

Because antituberculosis agents and zidovudine are commonly used in HIV-infected patients, we performed a cohort study to determine the toxicity of such combined therapy. A group of 24 consecutive human immunodeficiency virus (HIV)-infected patients with tuberculosis who received concomitant antituberculosis therapy and zidovudine (tuberculosis group) were compared with 24 patients who received zidovudine but not antituberculosis medications (comparison group). Comparison patients were matched to tuberculosis patients by age, sex, ethnic group, month of starting zidovudine, and CD4 cell count. Most tuberculosis patients received isoniazid, rifampin, pyrazinamide, and ethambutol initially, followed by isoniazid and rifampin for a mean total duration of 8.4 months. Baseline clinical and laboratory parameters in tuberculosis and comparison patients were similar, except for the mean hemoglobin (11.2 g/dl in tuberculosis patients versus 12.9 g/dl in comparison patients, p = 0.03). The mean zidovudine dose in tuberculosis and comparison patients was approximately 500 mg/day, and the mean duration of zidovudine therapy was 10.3 and 9.6 months, respectively. Symptoms occurred during therapy with similar frequency in both groups. The frequency and severity of leukopenia and granulocytopenia were similar in tuberculosis and comparison patients, but marked anemia (hemoglobin less than 9.5 g/dl) developed in 50% of tuberculosis patients and 17% of comparison patients (p = 0.03). The maximum decrease in hemoglobin during therapy was similar in both groups (mean of 2.0 versus 1.6 g/dl, respectively), suggesting that the higher frequency of marked anemia in tuberculosis patients was due to their lower baseline hemoglobin values. Although transfusions were required in five tuberculosis patients and one comparison patient, zidovudine was not permanently discontinued in any patient because of anemia.(ABSTRACT TRUNCATED AT 250 WORDS)