A multivariate analysis of mortality and hospital admissions with high-flux dialysis

The use of high-flux dialysis in clinical practice increased rapidly despite an absence of reports on the clinical effectiveness of the technique. Mortality and hospital admission rates of patients treated with high-flux dialysis were evaluated and compared with those of patients treated with conventional dialysis in a hospital-based renal dialysis unit in northern California. By use of a retrospective, cross-over design, 253 patients enrolled in the dialysis unit from January 1987 to January 1991 were studied. During this period, 107 patients were treated with high-flux dialysis for at least 1 month, and all but 17 of them had received conventional dialysis before switching to high-flux dialysis. The remaining 146 patients were treated with only conventional dialysis. Of the 80 patients who died during the study period, 69 were receiving conventional dialysis and 11 were receiving high-flux dialysis. The multivariate analyses, adjusted for age, gender, ethnic background, type of renal failure, comorbid conditions, and duration of ESRD, showed that annual mortality was substantially less for patients treated with high-flux dialysis compared with that for patients treated with conventional dialysis (7 versus 20%; P < 0.001). The difference in the rate of hospital admissions was not statistically significant. In this nonexperimental study, methods were applied to control for selectivity bias and other factors that might confound the apparent treatment effect. The findings suggest that the potential benefits of high-flux dialysis are sufficient to justify further confirmation in a randomized, controlled trial.