( Pa ge 80 )The use of albumin in the critical care setting is a very controversial issue.Although not a new situation, we have seen a renewed interest in this question, especially after the publication of one particular meta-analysis. Much has been said about its publication, but it merits a brief review. Hypoalbuminemia, no longer considered to be "acceptable" or even inevitable in critically ill patients, has also been revisited. We will examine the issue by reviewing studies that have tried to correct the condition, the mechanisms by which it appears and its prognostic value. Finally, a great deal of work has been done to gain a better understanding of albumin and its properties beyond the most commonly accepted, namely maintenance of plasma oncotic pressure. A review of the research will help us to recognize that more randomized controlled trials are warranted.
S U M M A R Y
DEPARTMENT OF INTENSIVE CARE ERASME HOSPITAL FREE UNIVERSITY OF BRUSSELS BRUSSELS, BELGIUM• Albumin, • Hypoalbuminemia• Meta-analysis• Critical care• Intensive care unit ublication of the Cochrane meta-analysis on the administration of albumin solutions 1 in 1998 generated much debate. The authors concluded that overall relative risk of death in patients receiving albumin solutions was 1.67 compared with non-albumin recipients (Table 1). The study was criticized for many reasons. Not insignificant was the inclusion of studies that did not have mortality as an endpoint, not to mention the heterogeneity of the studies included (different patient populations, different goals in administering albumin, etc). Some studies had 0% mortality in the control group (despite it being highly improbable that a study in an intensive care unit [ICU] setting would result in no mortality) whereas others failed to consider important factors (age, hemodynamic data, etc.).Recently, another meta-analysis on the same issue was published by Wilkes and Navickis 2 in which the authors found no increased risk in mortality when albumin was administered (Table 1). Methodological considerations and inclusion of a larger number of studies could account for this discrepancy. However, the clinician