The binding of several drugs to an inline i.v. filter that had been treated to inhibit drug-binding was studied. Solutions of mithramycin, vincristine sulfate, digitoxin, insulin, dactinomycin, and nitroglycerin in both 5% dextrose injection and 0.9% sodium chloride injection were allowed to flow through an i.v. administration set containing a 0.22-micron cellulose ester filter that had been treated with a proprietary agent. Actual administration conditions were simulated by using drug concentrations and flow rates commonly employed in clinical practice. The amount of each drug retained by the filter was determined by assaying aliquots of the solutions sampled before and after the solution passed through the filter membrane. In a second experiment, drug binding to the filter membrane was measured by incubating small pieces of the treated membrane in drug solutions and determining concentrations periodically until equilibrium was reached. Untreated filter membrane pieces were used as a control. In the experiment simulating actual i.v. administration, cumulative binding of mithramycin, vincristine sulfate, dactinomycin, and nitroglycerin to the treated filter was less than 6% of the initial dose infused; approximately 8-12% of the initial amounts of digitoxin and insulin were bound to the filter. In the equilibrium binding studies, the untreated filters bound twice as much digitoxin, 5-7 times as much mithramycin, vincristine sulfate, and dactinomycin, and 20 times as much insulin as the treated filters. The amount of nitroglycerin bound to the treated and untreated filters was not substantially different. Insulin had a greater binding tendency in 5% dextrose injection than in 0.9% sodium chloride injection in both experiments regardless of the filter treatment. Treatment of a cellulose ester i.v. filter with the proprietary agent used in this study facilitates drug delivery through this filter.