Rifampin and rifabutin induce the metabolism of many drugs, which may result in subtherapeutic concentrations and failure of therapy. However, differences between rifabutin and rifampin in potency of induction, and the specific enzymes which are altered, are not clear. This study, involving 12 adult male volunteers, compared the effects of 14-day courses of rifampin and rifabutin on clearance of theophylline, a substrate for the hepatic microsomal enzyme CYP1A2. Subjects were given oral theophylline solution (5 mg/kg of body weight) on day 1 and then randomized to receive daily rifampin (300 mg) or rifabutin (300 mg) on days 3 to 16. Theophylline was readministered as described above on day 15. The first treatment sequence was followed by a 2-week washout period; subjects then received the alternative treatment. Theophylline concentrations were determined for 46 h after each dose, and pharmacokinetic parameters were determined. One subject developed flu-like symptoms while taking rifabutin and withdrew voluntarily. Results from the remaining 11 subjects are reported. Compared with the baseline, the mean area under the concentration-time curve (AUC) (+/- standard deviation) for theophylline declined significantly following rifampin treatment (from 140 +/- 37 to 100 +/- 24 micrograms . h/ml, P <0.001); there was no significant change following rifabutin treatment (136 +/- 48 to 128 +/- 45 micrograms.h/ml). Baseline theophylline AUCs before each treatment phase were not different. A comparison of equal doses of rifampin and rifabutin administered to healthy volunteers for 2 weeks indicates that induction of CYP1A2, as measured by theophylline clearance, is significantly less following rifabutin treatment than it is following rifampin treatment. However, the relative induction potency for other metabolic enzymes remains to be investigated.