Increased activation of the central benzodiazepine receptor (BZR) appears to play an important role in hepatic encephalopathy (HE). However, there is controversy regarding whether the density or affinity of BZRs is altered. A previous positron emission tomography (PET) study using the BZR antagonist [11C]flumazenil (FMZ) found two- to threefold greater cerebral cortical tracer uptake in recurrent HE, but did not account for impaired FMZ metabolism due to liver disease or assess the relative contributions of tracer delivery versus BZR binding. We hypothesized that correcting for these factors would affect estimations of BZR binding in HE. Nine patients with recurrent HE and 13 age-comparable controls were studied with [11C]FMZ PET. After intravenous administration of [11C]FMZ, arterial blood samples were collected, and PET images were acquired over 60 minutes. FMZ transport and binding maps were calculated for each subject by using a physiological tracer kinetic model. In agreement with the previous report, we found that FMZ reached a much higher level and was retained longer in the HE cerebral cortex despite similar total blood radioactivity levels in the two groups. However, the patients showed impaired hepatic metabolism of FMZ. After physiological modeling incorporating these data, significant increases in BZR binding were found in the thalamus (13%), cerebellum (20%), and pons (23%). There were minor, statistically insignificant increases in cerebral cortical (10%), putamen (12%), and whole brain (12%) BZR binding in patients with recurrent HE. These findings are in general agreement with results of autopsy studies, confirming a lack of major increases in cortical or basal ganglial BZR binding in HE. They emphasize that physiological tracer modeling should be used and altered peripheral radioligand metabolism considered in future PET studies of HE.