• 
• 

Three Part Question

In [critically ill children with low serum albumin] does [giving albumin infusion] [improve COP and hence morbidity and mortality]?

Clinical Scenario

A 16-month-old boy with stage IV neuroblastoma and hypoalbuminaemia presented with a left sided haemorrhagic pleural effusion. He subsequently developed generalised oedema. You wonder if there was a role of albumin infusion in correcting hypoalbuminaemia and colloid osmotic pressure (COP), in order to treat the extravasation of fluid into tissue spaces.

Search Strategy

Cochrane Library (Issue 4, 2000), PubMed. LIMIT to English.
Cochrane, "hypoalbuminaemia" 7 systematic reviews (one relevant). Pubmed, "albumin" AND "critical illness"- 159 references (1 meta-analysis of 55 studies, 4 of which dealt with hypoalbuminaemia); "colloid osmotic pressure" AND "critical illness"- 4 references (one relevant to question); "hypoalbuminaemia" AND "critical illness"-6 references (one relevant to question).

Search Outcome

7 systematic reviews, one relevant (Cochrane) 169 articles, 3 relevant (Pubmed)

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Grundmann R and Heistermann S, 1985, USA	220 patients on adult ITU randomised to receive albumin when COP fell < 24 cm H2O (group 1) or COP<29 cm H2O (group 2)	PRCT(1b)	Post-op complications, COP, duration of intensive care and mortality	Albumin replacement did not influence the final outcome. The 95% CI of risk difference for mortality includes zero (-5.4%, -16.6 to 5.8%). The absolute risk increase of lower COP (<20 cm H2O) for mortality was 50.5% (95% CI, 20.5 to 80.5%)	Both groups received albumin. All patients were post-operative
Blunt MC et al, 1998, UK	145 survivors and nonsurvivors of prolonged critical illness	Retrospective review of practice(4)	COP and mortality	Non-survivors had significantly lower mean serum albumin compared with survivors; p<0.05. Albumin only contributed to 17% of the COP in critically ill patients. There was no relationship between death and COP	Adult pattern disease: one half of this population were post-operative patients, e.g., aortic aneurysm, gastrointestinal and renal patients
Alderson P et al, 2000,	30 randomised controlled trials comparing albumin with other interventions in critically ill patients	Systematic Review(1a)	Mortality	For hypoalbuminaemia relative risk of death was 1.69 (95% CI, 1.07 to 2.67). Pooled relative risk of death with albumin was 1.68 (95% CI, 1.26 to 2.23). The risk of death in the albumin treated group was higher than in the comparison group	Large peer response-(BMJ 1998;317:882,1999;318:464 and 1214). Small trial bias, lack of enough trials in the paediatric population and concerns over homogeneity through the trials
Wilkes MW and Navickis RJ, 2001, USA	55 randomised controlled trials comparing albumin therapy with other interventions	Systematic Review(1a)	Relative risk of death	Pooled relative risk for death was 1.11 (95% CI, 0.95 to 1.28). Relative risk for death in hypoalbuminaemia group was 1.59 (95% CI, 0.91 to 2.78). Overall, no effect of albumin on mortality detected	Relative risk was lower in trials with blinding, mortality as end point, no crossover, and 100 or more patients. Only 5 trials dealt with patients with hypoalbuminaemia

Comment(s)

There is a paucity of data in children. However, in critically ill adults a decrease in serum albumin is associated with increased morbidity and mortality. This may represent a disease-related alteration in hepatic synthetic function. Albumin contributes up to 80% of COP in healthy subjects; however, its contribution towards COP is only 17% in critically ill individuals (Blunt et al). In adults, the studies suggest that albumin administration has no effect on mortality. In addition, its contribution towards COP is questionable. In fact, there appears to be no significant difference in COP of survivors compared with non-survivors of critical illness. Taken together, this information suggests that low serum albumin may merely be a surrogate marker of disease severity rather than an indicator of low COP. Hence, when treating patients with hypoalbuminaemia, efforts must be focused on correcting the underlying disorder, rather than reversal of hypoalbuminaemia; or, alternatively, on measuring COP directly. There are no such studies in children, but in the systematic reviews the relationship between mortality and albumin administration was similar to that described in adults. We therefore speculate that in children with protracted or critical illness, such as seen in oncological or life-threatening disease, the adult relationship between albumin, COP and outcome may also hold. However, this idea should be tested by prospective biochemical study.

Clinical Bottom Line

Little published research addresses the question of albumin use in oncological hypoalbuminaemia in children. Critically ill adults with hypoalbuminaemia do not have better outcomes when treated with albumin.

Level of Evidence

Level 3 - Small numbers of small studies or great heterogeneity or very different population.

References

1. Grundmann R, Heistermann S. Postoperative albumin infusion therapy based on colloid osmotic pressure. A prospectively randomized trial. Archives of Surgery 1985;120(8):911-15.
2. Blunt MC, Nicholson JP, Park GR. Serum albumin and colloid oncotic pressure in survivors and nonsurvivors of prolonged critical illness. Anaesthesia 1998;53(8):755-761.
3. Alderson P, Bunn F, Lefebvre C, et al. Human albumin solution for resuscitation and volume expansion in critically ill patients. In The Cochrane Library Issue 3, 2000. Oxford: Update Software.
4. Wilkes MM, Navickis RJ. Patient survival after human albumin administration. A meta-analysis of randomized controlled trials. Ann Intern Med 2001;135(3):149-164.