Severely ill patients in the intensive care unit (ICU) often require fluid resuscitation to restore intravascular volume. The choice of fluid to use for resuscitation varies widely, with some centers preferring crystalloid solutions while others use albumin or other colloid solutions. Smaller comparison studies of the effectiveness of these fluids have produced conflicting results, and some larger meta-analyses have come to differing conclusions. The Saline versus Albumin Fluid Evaluation (SAFE) study investigators designed a large randomized study to compare saline with albumin for fluid resuscitation.
This multicenter trial was conducted at 16 academic hospital ICUs, enrolling 7,000 patients who were randomized to intravenous saline or 4 percent albumin when fluid resuscitation was required. Special masking bottles were used so physicians would be blinded to the assigned fluid. Use of blood products, enteral or parenteral nutrition, and specific electrolyte replacement fluids were left to the discretion of the treating physicians. Resuscitation fluids outside of the randomized assignment were used in 8.8 percent of the albumin group and 10.7 percent of the saline group. Twenty-eight days after randomization, follow-up survival data were missing for 1 percent of enrolled patients.
Patients randomized to albumin received a smaller total volume of resuscitation fluids during the first several days of ICU treatment, had a lower mean heart rate, the same mean arterial pressure, and a significantly higher central venous pressure, compared with patients who received intravenous saline.
The mortality rate after 28 days was not significantly different between the albumin group (20.9 percent) and the saline group (21.1 percent). Slightly more patients assigned to albumin (3.2 percent) had an ICU stay prolonged beyond the 28 days compared with the saline group (2.5 percent). The mean length of stay in the ICU and overall stay in the hospital were not significantly different between the two groups. In the subgroup of ICU patients admitted because of trauma, there was a higher mortality rate among those randomized to albumin (13.6 percent) compared with the saline group (10 percent), which was of borderline statistical significance.
The authors conclude that use of albumin or saline for fluid resuscitation in ICU patients does not affect the overall mortality rate at 28 days after randomization.