A different view on human albumin.

With great interest we have read the article by Zhang for albumin substitution [9]. Far from being beneficial, the and Frei on the anti-inflammatory effects of human alstudy documented an increased relative risk of death of bumin in the September issue of this journal [1]. They more than 10% overall and up to 76% in subgroups [9,10]. report a downregulation of vascular cell adhesion As mentioned by Zhang and Frei, the maintenance of molecule-1 (VCAM-1) on albumin-treated human aortic capillary permeability seems to be another physiological endothelial cells which supports our previous results of a function of albumin [11]. In contrast to their paper, a more reduced VCAM-1 expression on TNFa-activated venous recent study by Margason and Soni did not confirm a endothelial cells (HUVEC) [2]. Nevertheless, their paper reduction in capillary permeability in septic patients folwarrants further comments since the authors exclusively lowing infusion of human albumin [12]. discuss the beneficial effects of albumin and do not take A striking protection from neuronal injury during isinto account possible risks of albumin supplementation. chemia / reperfusion has been reported in models of stroke Fast infusion of human albumin is well known to cause [13]. Careful review of the experimental design shows that severe hypotension and flushing. Such reactions are susbolus infusion of 25% human albumin was compared to pected to be caused by impurities of clinically used equal volumes of 0.9% saline. The comparison of a albumin preparations involving contamination with proinhyperoncotic volume therapy with a low volume crysflammatory mediators like kinins [3]. Accordingly, altalloid regime results in important confounding effects on bumin preparations do not always downregulate endotheliflow velocity and shear forces which will influence al cell adhesion molecules but can also upregulate Eleukocyte adhesion directly apart from any specific antiselectin and intercellular adhesion molecule 1 (ICAM-1) inflammatory effect [14]. In contrast, when human albumin depending on their manufacture [2]. Upregulation of these was compared to another colloid in a clinical study on molecules affects monocyte adhesion [4] and plays an even severe sepsis, volume resuscitation with human albumin more important role during ischemia / reperfusion and was accompanied by increased plasma levels of soluble systemic inflammation where neutrophils predominate in adhesion molecules and a worsened oxygenation index the microvasculature [5]. Therefore, one should be aware [15]. Manuscripts of basic research are not review papers of such effects before favouring human albumin as an and do not deal with therapeutic means. However, if the anti-inflammatory plasma protein substitute. authors suggest ‘‘ . . . increasing albumin levels may be an Although hypoalbuminemia was associated with ineffective strategy to lower cardiovascular risk’’, they creased morbidity and mortality in some clinical studies should also discuss the possible negative aspects of doing [6], the effects of albumin substitution are much more so. controversially discussed in the literature than it seems from the paper of Zhang and Frei. Every meta-analysis on albumin supplementation has been criticized to a considerR eferences able extent [7–10]. Even the most recent one that defeated the harmful effects of human albumin documented in the [1] Z hang WJ, Frei B. Albumin selectively inhibits TNF alpha-induced expression of vascular cell adhesion molecule-1 in human aortic previous Cochrane analysis [7] could not show any benefit endothelial cells. Cardiovasc Res 2002;55:820–829. [2] N ohe B, Dieterich HJ, Eichner M, Unertl K. Certain batches of *Corresponding author. Tel.: 149-7071-29-86622; fax: 149-7071-29albumin solutions influence the expression of endothelial cell 5533. adhesion molecules. Intensive Care Med 1999;25:1381–1385. ́ E-mail address: [email protected] (B. Nohe). [3] T urner PJ, Young IF, Marley PB, Herrington RW, Schiff P. Albumin