A non-randomized cross-over study was performed with bupivacaine HCl (5 mg x ml(-1)) aqueous solution and bupivacaine free base (4.44 mg x ml(-1)) in Viscoleo/castor oil 2:1 (v/v) administered s.c. to male Wistar rats. Plasma levels were analyzed by LC-MS. Plasma profiles obtained after administration of oily solution showed a prolonged bupivacaine release with lower peak plasma levels as compared to administration of an aqueous formulation applied in the same compartment. t(1/2), t(max), C(max) and AUC(0-infinity) for the aqueous solution were 63+/-8 min, 19+/-16 min, 194+/-46 ng x ml(-1) and 25,000+/-3000 ng min x ml(-1), respectively, while the corresponding data for the oil solution were 368+/-89 min, 334+/-186 min, 36+/-25 ng x ml(-1) and 25,000+/-6000 ng x min x ml(-1). The present data indicate the potential of designing an oil formulation of bupivacaine with a prolonged local analgetic effect exhibiting a minimum of systemic toxicity. In vivo release of bupivacaine from the oil solution was evaluated by a numerical deconvolution method. In vivo release kinetics was found to be first-order and corresponded well with in vitro release kinetics found using a rotating dialysis cell. This led to establishment of an in vitro/in vivo correlation for this particular formulation.