The anti-N-methyl-D-aspartate (NMDA) effect of dextromethorphan (DEM) seems to be mainly related to the unchanged drug rather than to its more potent metabolite dextrorphan (DOR). The aim of our study was to assess the involvement of P-glycoprotein (P-gp) and pH conditions in the transmembranal transport of these two NMDA antagonists, using a human in vitro Caco-2 cell monolayer model. Transmission electron microscopy, transepithelial electrical resistance, [(3)H]-mannitol permeability, Western blot analysis and the bidirectional transport of the positive controls, rhodamine and digoxine were used to confirm model's integrity and validity. The bidirectional transport of DEM and DOR (1 to 100microM) across the monolayers was investigated in the presence and absence of the P-gp inhibitor cyclosporine A (10microM) at two pH conditions (pH 6.8/7.7-pH 7.4/7.4) and assessed with the specific and more potent P-gp inhibitor GF120918 (4microM). Analytical quantification was achieved using high performance liquid chromatography. At a pH gradient, DEM and DOR were subject to a significant active efflux transport (Papp(B-A) > 2-3x Papp(A-B); p<0.01). However, neither the influx nor the efflux was affected by P-gp inhibitors. At physiological pH, we observed no more efflux of the drugs and no influence of the inhibitors. In conclusion, dextromethorphan and dextrorphan are not P-gp substrates. However, pH-mediated efflux mechanisms seem to be involved in limiting DEM gastrointestinal absorption. The preferential anti-NMDA central effect of DEM appears to be P-gp independent.