In a randomized trial conducted in 9 Ontario transplant centers, 107 cadaver kidney donors were randomized so that both kidneys were preserved either by simple cold storage (CS) or by pulsatile perfusion (PP). After exclusions and losses to follow-up, there were 90 patients in the CS group and 91 in the PP group. Total preservation times were similar in the two groups (27.7 +/- 12 and 30.5 +/- 10 hr), as was the proportion of kidneys with long (greater than 36 hr) storage times (21% and 26%). Twelve-month graft survival was 70% for CS and 75% for PP (not significant, NS); patient survival was 89% for CS and 95% for PP (NS). The incidence of kidneys that never functioned was also similar in the two groups (14% vs. 15%, respectively). However, the risk of early, reversible graft dysfunction was significantly higher for CS (44%) than for PP (31%). The same effect was observed with two independent methods of assessment of graft function. The chief effect of PP on renal function was in reduction of the mean serum creatinine at 1 week (632 mumol/L for CS vs. 403 mumol/L for PP, P less than .001). There was no significant influence of storage mode on the rapidity of urine output or number of dialyses required. No apparent effect of cyclosporine on the results with either preservation method was seen. We conclude that the case for retention of PP for kidney preservation now rests on its ability to achieve slightly better initial function than CS. However, the mild dysfunction observed with CS did not appear to translate into long-term detriment. When these considerations are weighed against the increased cost of PP, the advantage may lie with cold storage.