Introduction, demography, and general methods

The International Cooperative Biocompatibility Study was planned to analyse the symptomatic and laboratory response to seven different dialysers studied in five centres in four countries. The dialysers used were the G10-3N, G120 M, CD 4000, T 150, Duo-Flux, F 60, and Filtral. A total of 37 patients in the Veterans Administration Lakeside Medical Center, Chicago; Henry Ford Hospital, Detroit; Osaka City University Hospital, Osaka; Wilhelm Pieck University Hospital, Rostock; and Huddinge University Hospital, Stockholm were studied. All patients had been dialysed for a minimum of 6 months, were non-diabetic, stable, and compliant; and most were middle-aged and male. Patients were treated three times per week for 2 weeks with each new dialyser with random assignment to one of four orders of dialyser use. The same manufacturing lot of each dialyser, blood line sets and needles were used by all centres. Delivery systems were volumetric controlled except for some patients in Osaka treated with negative-pressure equipment. Individual patient prescriptions (Kt/V for urea), in use prior to the study, were continued. Kt/Vs for all treatments were derived from reported blood flows and blood water corrected mass transfer coefficients multiplied by dialyser surface area. Clinical data were measure pre, intra- and postdialysis. Hourly signs, symptoms, drugs, and nursing interventions were recorded using the identical treatment record at all centres. In addition, patients completed a questionnaire form (translated into the appropriate language) at least weekly, relating symptoms experienced with each dialyser. Laboratory investigations were performed during the sixth consecutive treatment. During blood sampling, ultrafiltration rate was maintained at less than 50 ml/h and blood flow established at 200 ml/min. Changes in differential white blood cell count, leukocyte electrophoretic mobility platelet count, platelet aggregation, β-thromboglobulin release, C(3a), C₃ and C(3d), plasma β₂-microglobulin, elastate, and blood gases were routinely measured. All data were collected in Detroit for analysis using an IBM 3090 computer and the SAS programs (Statistical Analysis System, N. Carolina). Parametric and non-parametric tests were used for statistical analysis. Laboratory and blood pressure data were examined as differences from baseline. Questionnaire and staff-recorded symptom data were ranked prior to nonparametric analysis to adjust for the distribution of responses. The Duncan multiple range test and analysis of variance were used for statistical analysis. The ranked data were tested using Friedman's two-way analysis of variance for ranks and the correlations between laboratory data, staff and patient reported symptom rankings were measured using Spearman's test.