beta-Lactam antibiotics have been shown to cause platelet dysfunction and bleeding in some patients. However, relative antiplatelet activity of various beta-lactams has remained controversial. Results of clinical studies have been variable because of the presence of underlying disease in the study patients, in addition to inherent difficulties of in vivo experimentation such as individual variations of drug metabolism and drug kinetics. Thus, we designed in vitro experiments to study the direct effect of penicillin G, carbenicillin, ticarcillin, mezlocillin, piperacillin, nafcillin, and azlocillin on platelets. Platelets obtained from normal volunteers were exposed in vitro for 15 minutes to increasing concentrations of the test penicillins (10.0, 12.5, 15.0, and 20.0 mmol/L), and the platelet aggregation response determined after the additional of adenosine diphosphate (2.5 to 5.0 mumol/L), epinephrine (0.1 X 10(-3) mol/L), thrombin (0.01 to 0.02 U/ml), and collagen (11.62 micrograms/ml). All tested penicillins inhibited platelet aggregation in a saturable dose-dependent manner that was reversible by platelet washing. Biostatistical comparison of inhibition of platelet aggregation demonstrated nafcillin to cause significantly more inhibition, followed by azlocillin, mezlocillin, and piperacillin as a group. Penicillin G, carbenicillin, and ticarcillin were the least inhibitory. The mean percent inhibition (epinephrine) at 20 mmol/L concentration was nafcillin 86.4%, mezlocillin 83.2%, piperacillin 80.3%, azlocillin 76.4%, ticarcillin 73.2%, carbenicillin 66.4%, and penicillin G 58.4% (overall P less than 0.001). We conclude that all penicillins tested in vitro inhibit platelet aggregation in normal individuals, but to varying degrees. The inhibitory response, which is most likely a membrane-related phenomenon, is dose dependent and reversible.