The pharmacokinetics of vancomycin in patients undergoing dialysis with cuprophane membranes are well known, however little has been reported of the use of polyacrylonitrile membranes in dialysis. We studied, in a crossover design, eight dialysis patients (7 men, 1 woman) aged 30 to 66 years who prospectively received 1 gram of vancomycin i.v. before first dialysis and were subsequently hemodialyzed with cuprophane every second day for a total of three times. A month later trial was repeated using polyacrylonitrile. A mono-compartment model was used to calculated pharmacokinetic parameters. Mean +/- standard deviation of vancomycin clearance varied from 5.2 +/- 2.1 ml/min in the interdialysis period to 9.7 +/- 2.7 ml/min during dialysis with cuprophane and to 58.4 +/- 15.6 ml/min during dialysis with polyacrylonitrile (p less than 0.001). Vancomycin half-life varied from 71.5 +/- 23.0 to 35.9 +/- 9.8 and to 6.1 +/- 1.4 hours, respectively (p less than 0.001). Fractional removal of vancomycin increased from 4% using the cuprophane dialyzer to 34% using the polyacrylonitrile dialyzer (p less than 0.001). Serum vancomycin levels at 100 and 168 hours were higher with cuprophane than with polyacrylonitrile (7.0 +/- 2.2 vs 3.9 +/- 1.2 micrograms/ml) (p less than 0.001). Moreover, the mean levels at 100 hours were suboptimal on polyacrylonitrile. Approximately 208 +/- 53 mg of vancomycin were removed during one polyacrylonitrile dialysis. Thus, those patients who undergo dialysis with polyacrylonitrile and are treated with vancomycin may need supplementary doses post dialysis or to lessen dosage intervals than those traditionally used for dialysis patients since clearance of the drug is significantly higher than with cuprophane dialyzers. Continuous monitoring of vancomycin levels is also recommended.