To determine the effect of random assignment to fluid resuscitation with albumin or saline on organ function and mortality in patients with severe sepsis. Methods: Pre-defined subgroup analysis of a randomized controlled trial conducted in the intensive care units of 16 hospitals in Australia and New Zealand. Results: Of 1,218 patients with severe sepsis at baseline, 603 and 615 were assigned to receive albumin and saline, respectively. The two groups had similar baseline characteristics. During the first 7 days mean arterial pressure was similar in the two groups, but patients assigned albumin had a lower heart rate on days 1 and 3 (p = 0.002 and p = 0.03, respectively) and a higher central venous pressure on days 1-3 (p <0.005 each day). There was no difference in the renal or total Sequential Organ Failure Assessment score of the two groups; 113/603 (18.7 ) of patients assigned albumin were treated with renal replacement therapy compared to 112/615 (18.2 ) assigned saline (p = 0.98). The unadjusted relative risk of death for albumin versus saline was 0.87 [95 confidence interval (CI) 0.74-1.02] for patients with severe sepsis and 1.05 (0.94-1.17) for patients without severe sepsis (p = 0.06 for heterogeneity). From multivariate logistic regression analysis adjusting for baseline factors in patients with complete baseline data (919/1,218, 75.5 ), the adjusted odds ratio for death for albumin versus saline was 0.71 (95 CI: 0.52-0.97; p = 0.03). Conclusions: Administration of albumin compared to saline did not impair renal or other organ function and may have decreased the risk of death.