Objective: To identify patients with tuberculosis (TB) showing poor bioavailability for rifampicin and to delineate the role of possible factors such as over-expression of intestinal cytochrome P450 3A4 (CYP3A4) and P-glycoprotein (PGP) in the drug's bioavailability by administering known blockers. Patients, design and setting: 77 clinically proven TB patients were included in this nonblinded, randomised, comparative trial carried out at one centre at the All India Institute of Medical Sciences, New Delhi, India. Interventions: The concentrations of rifampicin and its active metabolite, 25-desacetylrifampicin (DRMP), were measured in blood samples of the 77 TB patients at 0, 1, 2 and 4 hours after their usual morning rifampicin dose. Of these, 19 patients showing the lowest area under the concentration-time curve values from 0 to 4 hours after administration (AUC0-4) were selected and pretreated with a single dose of either verapamil (80mg) or itraconazole (200mg) as both PGP and CYP3A4 blockers 1 hour prior to rifampicin administration. Rifampicin and DRMP concentrations were estimated using high performance liquid chromatography in all serum samples collected at the same timepoints. Main outcome measures and results: A statistically significant increase (p < 0.05) was found both in the serum levels of rifampicin at 2 hours and in the AUC0-4 values (158% and 84%, respectively) after pretreatment with verapamil. However, an increase in the levels of rifampicin was found to be insignificant on pretreatment with itraconazole. The estimated levels of DRMP also supported these results. Conclusions: The increase in rifampicin levels on administration of a PGP/CYP3A4 blocker suggests a pivotal role for PGP/CYP3A4 in the absorption of rifampicin in patients with TB, which may be responsible for lower or variable levels of rifampicin.