Recommendations for the use of albumin and immunoglobulins.

Human albumin is a physiological plasma-expander; its limited availability and high cost make it essential to define recommendations for its appropriate use, as an alternative to other therapeutic strategies including solutions of crystalloids and non-protein colloids, and have also stimulated numerous studies, which have sometimes reached contradictory conclusions1–8. In 1998 a meta-analysis of 30 randomised trials suggested that the use of albumin was associated with an increased mortality rate among critically ill patients9,10. This conclusion, also reached by two subsequent Cochrane reviews11,12, was not confirmed by a meta-analysis in 2001 or by more recent studies13–23. A review in 2006 showed that renal damage can be induced by the use of hydroxyethyl starch and gelatine in sepsis and surgery24. The limited power of all these studies could lie in their having combined results from heterogeneous types of patients with different baseline albumin concentrations25,26.