1. Currently available in vitro blood-brain barrier models all have recognized restrictions. In addition to leakiness, inconsistent data about P-glycoprotein mediated efflux limit the attractiveness of the primary bovine brain microvessel endothelial cells (BBMECs). Therefore, we re-evaluated the role of P-glycoprotein mediated efflux with two culture conditions in BBMECs for prediction of drug permeability of potential P-glycoprotein substrates. 2. BBMECs were monocultured on filters on petri dishes and on filter inserts, and expression and localization of P-glycoprotein were compared by using western blot and confocal microscopy, respectively. The functionality of P-glycoprotein was assessed by using cellular uptake, calcein-AM and bidirectional transport assays. 3. P-glycoprotein expression was higher in BBMECs cultured on filter inserts decreasing the permeability of digoxin and paclitaxel, but not the permeability of vinblastine. However, the monocultured BBMECs were not able to demonstrate efflux in the bidirectional transport assays. Under certain culture conditions, occludin may not be correctly located, perhaps explaining in part the leakiness of BBMECs. 4. In conclusion, BBMECs, despite possessing a functional P-glycoprotein, under certain culture conditions may not be a suitable in vitro model for the bidirectional transport assays and for predicting the permeability of drugs and xenobiotics that are potential P-glycoprotein substrates.