There is an increased use of extracorporeal membrane oxygenation (ECMO) in the last 15 years for critically ill neonates. While receiving ECMO therapy, the critically ill infant needs various medications. We performed an in vitro study to evaluate the potential effect of the membrane oxygenator on drug extraction. Two closed ECMO circuits were set up at rates of 320 ml/min. One circuit was new and the other was used clinically for 5 days. Morphine at 8 ng/ml, gentamicin 10 micrograms/ml, vancomycin 40 micrograms/ml, phenobarbital 20 micrograms/ml, and phenytoin 20 micrograms/ml were injected into the circuit at 1-h intervals. Blood samples were drawn from the circuit at 10, 30, 60, and 240 minutes after injection. In the new circuit, drugs were eliminated as follows: vancomycin 36%, gentamicin 10%, phenobarbital 17%, phenytoin 43%, morphine 36%. In the used system, levels fell to a much smaller extent: vancomycin 11%, phenobarbital 6%, gentamicin 0%, phenytoin 0%, and morphine 16%. In a child receiving 20 micrograms/kg/h infusion of morphine, steady-state concentrations of 68.2 ng/ml fell to 11.6 ng/ml after changing the membrane. Our data indicate that the ECMO is associated with lowering of the concentrations of commonly used medications and that this process may depend partially on how new the membrane is. Before these changes may lead to new dosing guidelines for small children receiving ECMO, more experiments with new and used systems are warranted, as well as with different types of ECMO.