OBJECTIVE The purpose of this study was to investigate the pH-dependent passive and active transport of weakly basic drugs across the human intestinal epithelium. METHODS The bidirectional pH-dependent transport of weak bases was studied in Caco-2 cell monolayers in the physiologic pH range of the gastrointestinal tract. RESULTS A net secretion of atenolol and metoprolol was observed when a pH gradient was applied. However, the bidirectional transport of both compounds was equal in the nongradient system. Hence, at lower apical than basolateral pH a change in passive transport caused by an imbalance in the concentration of the uncharged drug species resulted in a "false" asymmetry (efflux ratio). Furthermore, a mixture of pH-dependent passive and active efflux was found for the P-glycoprotein (P-gp, MDR1, ABCB1) substrates, talinolol and quinidine, but not for the neutral drug, digoxin. However, the clinically important digoxin-quinidine interaction depended on the presence of a pH gradient. Hence, the degree of interaction depends on the amount of quinidine available at the binding site of the P-gp. CONCLUSIONS Active efflux of weak bases can only be accounted for when the fraction of unionized drug species is equal in all compartments because the transport is biased by a pH-dependent passive component. However, this component may take part in vivo and contribute to drug-drug interactions involving P-gp.