Deaths have been reported among heroin addicts related to combined buprenorphine and flunitrazepam use. The aim of this study was to determine the existence of a drug-drug interaction during the distribution phase of buprenorphine. Arterial blood gases were measured after intravenous administration of buprenorphine alone (30 mg/kg), flunitrazepam alone (40 mg/kg) or both drugs in rats. Buprenorphine kinetics was studied in plasma and in striatum using cerebral microdialysis, both alone and after rat pretreatment with flunitrazepam. In contrast to buprenorphine or flunitrazepam alone, buprenorphine in combination with flunitrazepam induced a significant, rapid and sustained respiratory depression. Arterial PCO2 was increased at 1.5 min (6.7+/-0.2 versus 5.4+/-0.3 and 5.5+/-0.3 kPa, respectively, P=0.04) (mean+/-S.E.M.), and arterial pH decreased (7.37+/-0.02 versus 7.45+/-0.02 and 7.45+/-0.01, respectively, P=0.03). Plasma buprenorphine kinetics was well described by a three-compartment linear model, with a distribution half-life of 7.4+/-2.7 min and an elimination half-life of 463.9+/-152.3 min. However, neither plasma nor striatal buprenorphine kinetics were significantly altered by pre-administration of flunitrazepam. The adverse interaction between flunitrazepam and buprenorphine cannot be explained by a pharmacokinetic drug-drug interaction during the distribution phase of buprenorphine.