In vitro placental perfusion is widely used to investigate the placental transfer of endogenous compounds and, to a lesser extent, that of drugs. The aim of this study was to assess the suitability and reliability of such in vitro systems for application on drug placental transfer studies. We investigated the time course of theophylline (TH) transfer, a drug frequently used in the perinatal period. Eight experiments were performed with maternal and fetal circuits maintained in an open system, perfusing placentas for 160 min with Earle's enriched bicarbonate buffer containing two test substances, antipyrine (AP), (80 mg/L) and creatinine (CR), (150 mg/L), and the tool drug TH (15 mg/L). All substances equilibrated in our system with times proportional to the chemical-physical characteristics of each compound, being the time required to reach the steady state 5 to 12 min for AP, 12 to 31 min for CR and 10 to 35 min for TH. AP and CR clearances were 2.94 +/- 0.33 and 0.83 +/- 0.26 mL/min, respectively. The transfer profile of TH was similar to that of AP and its clearance was 2.39 +/- 0.37 mL/min, with a clearance index of 0.80 +/- 0.11. Transfer percentages of TH are in agreement with in vivo values for both humans and animals, and with results obtained during in situ placental perfusion in the rabbit. Physiological conditions and biochemical properties of the tissue were well maintained throughout perfusion. Glucose consumption and lactate release were, respectively, 0.65 +/- 0.16 and 0.73 +/- 0.11 mumoles/min/g. Oxygen consumption was 5.29 +/- 1.32 mL/min/kg and oxygen transfer from the maternal to fetal circuit was 0.99 +/- 0.43 mL/min/kg. The findings support the reliability of this technique to study transplacental passage of drugs, and the relevance of such a model to obtain information concerning potential therapeutic or toxicologic effects of drugs during the last trimester of pregnancy.