The influence of antituberculosis drugs on diazepam disposition was assessed in a series of volunteers and patients who received single intravenous doses of diazepam. In study 1, nine healthy subjects received diazepam in the drug-free control state and again during treatment with isoniazid (INH), 180 mg/day. INH did not alter diazepam volume of distribution (Vd) or protein binding, but prolonged mean elimination half-life (t1/2) from 34 to 45 hr (p less than 0.02), and reduced total clearance from 0.54 to 0.40 ml/min/kg (p less than 0.02). In study 2, diazepam disposition in a group of seven tuberculous patients on triple therapy with INH, ethambutol (EMB), and rifampin (RIF) was compared with that in healty drug-free controls matched for age and sex. Diazepam Vd and protein binding were nearly identical between groups, but mean t1/2 among patients (14 hr) was significantly shorter than in controls (58 hr, p less than 0.01) and total clearance correspondingly increased (to 1.50 from 0.37 ml/min/kg, p less than 0.01). Study 3 compared six newly diagnoses tuberculous patients receiving initial therapy with EMB alone with age- and sex-matched controls. Diazepam unbound fraction in patients tended to be higher than in controls, and diazepam Vd and clearance tended to be lower but the differences were not statistically significant. Thus, diazepam clearance is impaired and t1/2 prolonged by administration of INH alone. Markedly increased clearance and shortened t1/2 in triple-therapy patients is probably due to enzyme-inducing effects of RIF. Dosage of diazepam may require adjustment in patients with tuberculosis on chemotherapy.