Objective: To determine whether replacement of human albumin will improve a patient's prognosis.
Design: A randomized, double-blind, controlled study in which 25 g of human albumin vs. placebo was administered intravenously daily.
Setting: A university-affiliated hospital.
Patients: Thirty-six patients with hypoalbuminemia (serum albumin of <2.5 g/dL), receiving total parenteral nutrition. None of the patients had known cancer, cirrhosis, or nephrotic syndrome.
Interventions: Each patient received at least 6 days of therapy (6 to 24 days of albumin; 7 to 32 days of placebo). Four subjects were excluded from the study since they received therapy for <6 days. One patient was excluded from the study after nephrotic syndrome was identified. Albumin metabolic rates for those patients receiving albumin were estimated using the formula: Metabolism of albumin = 25 g/day + (albumin 1 - albumin 2)(Vd)/days, where albumin 1 and 2 are the serum albumin concentrations (g/L) at the beginning and end of the serum sampling intervals, respectively; Vd is the volume of distribution (L); and days relates to the number of days of the sampling interval.
Measurements and main results: Sixteen patients received albumin; 15 patients received placebo. One patient receiving placebo and two patients receiving albumin died within 30 days. One patient who received placebo and three patients who received albumin developed sepsis or bacteremia; four patients who received placebo and seven patients who received albumin developed pneumonia during the study (NS). The serum albumin increased in all patients receiving intravenous albumin, but one patient received intravenous albumin for only 6 days. The mean serum albumin concentration increased by 1.42 g/dL in the albumin patients, and increased by 0.29 in the placebo patients (p < .0001 by unpaired t-test). Mean initial albumin metabolism was 17.4 g/day (0.3 g/kg/day). At the end of therapy, albumin metabolism was 20.5 g/day (0.36 g/kg/day) (paired t-test, p = .4, NS).
Conclusions: a) The administration of intravenous albumin to hypoalbuminemic patients receiving total parenteral nutrition does not improve morbidity or mortality. b) Albumin metabolic rates, initially related to the catabolic state, are high; later, these rates are high related to filling of the albumin space and gluconeogenesis. c) On the basis of the high albumin catabolic rates at the end of the infusion, doses of albumin of <25 g/day might be sufficient to replace albumin stores.