Abstract Background [¹¹C]Flumazenil and positron emission tomography (PET) are used clinically to assess gamma-aminobutyric acid (GABA)-ergic function and to localize epileptic foci prior to resective surgery. Enhanced P-glycoprotein (P-gp) activity has been reported in epilepsy and this may confound interpretation of clinical scans if [¹¹C]flumazenil is a P-gp substrate. The purpose of this study was to investigate whether [¹¹C]flumazenil is a P-gp substrate. Methods [¹¹C]Flumazenil PET scans were performed in wild type (WT) (n = 9) and Mdr1a/1b, (the genes that encode for P-gp) double knockout (dKO) (n = 10) mice, and in naive rats (n = 10). In parallel to PET scanning, [¹¹C]flumazenil plasma concentrations were measured in rats. For 6 of the WT and 6 of the dKO mice a second, [¹¹C]flumazenil scan was acquired after administration of the P-gp inhibitor tariquidar. Cerebral [¹¹C]flumazenil concentrations in WT and Mdr1a/1b dKO mice were compared (genetic disruption model). Furthermore, pre and post P-gp-blocking cerebral [¹¹C]flumazenil concentrations were compared in all animals (pharmacological inhibition model). Results Mdr1a/1b dKO mice had approximately 70% higher [¹¹C]flumazenil uptake in the brain than WT mice. After administration of tariquidar, cerebral [¹¹C]flumazenil uptake in WT mice increased by about 80% in WT mice, while it remained the same in Mdr1a/1b dKO mice. In rats, cerebral [¹¹C]flumazenil uptake increased by about 60% after tariquidar administration. Tariquidar had only a small effect on plasma clearance of flumazenil. Conclusions The present study showed that [¹¹C]flumazenil is a P-gp substrate in rodents. Consequently, altered cerebral [¹¹C]flumazenil uptake, as observed in epilepsy, may not reflect solely GABA_A receptor density changes but also changes in P-gp activity.