Volume 10 Supplement 1

26th International Symposium on Intensive Care and Emergency Medicine

Open Access

The Cost of Albumin Sepsis Treatment (COAST) study: efficacy analysis of albumin for severe sepsis. Modelling using the SAFE study and CUB-Rea data

  • B Guidet1,
  • P Aegerter2 and
  • G Jasso-Mosqueda3
Critical Care200610(Suppl 1):P177


Published: 21 March 2006


The use of albumin for the treatment of severe sepsis remains controversial. The SAFE study [1] reported a mortality rate of 35.2% in the group resuscitated with saline compared with 30.7% in the group that received albumin (relative risk 0.74, P = 0.09). This potential beneficial effect could be explained by several properties of albumin (correction of hypovolemia, of low oncotic pressure or of hypoalbuminemia, antioxidative and anti-inflammatory properties). Despite existing recommendations, albumin remains largely underused in this indication. Given its higher cost compared with other volume expansion products, modelling of its cost/efficacy ratio would be justified in light of the product's potential advantages.


Everyday medical practice was compared with systematic use of albumin. The study population was defined as all adult patients having been treated for severe sepsis in one of the 35 units in the CUB-Rea database between 1 January 1998 and 31 December 2002, not including burn patients, patients with mediastinitis, organ transplant recipients and those having received extracorporeal circulation (n = 11,137). Only hospital stays longer than 24 hours and including at least circulatory, kidney or respiratory failure were considered. The costs of intensive care were calculated using the GHS cost (cost of stay) + the daily intensive care cost. Two indicators of efficacy were used: the number of lives saved and the number of years of life gained, using as the baseline case the 4.6% reduction in mortality in the albumin arm of the SAFE study. The cost of albumin was estimated based on the quantities administered in this same study (2.24 l). Life expectancy was determined using the DEALE method with the following factors: age, sex, IGS2 score and MacCabe score.


There were 5980 deaths in intensive care in the CUB-Rea database. The use of albumin induced a 4.6% reduction in the number of deaths (512 deaths avoided). Mean survival of the 5156 patients released alive from the hospital was estimated at 9.78 years (95% CI: 9.64–9.92). The average cost per patient of administration of albumin was evaluated at €78.10. The cost per life saved was €6073 and per year of life saved was €617.


Application of the SAFE results to the CUB-Rea data showed a highly favourable cost/efficacy ratio for the use of albumin in severe sepsis.

Authors’ Affiliations

Hopital Saint-Antoine, Paris Cedex 12
Hopital Ambroise Paré
AREMIS, Neuilly Sur Seine


  1. Cook D: N Engl J Med. 2004, 350: 2247-2256. 10.1056/NEJMe048095View ArticleGoogle Scholar


© BioMed Central Ltd 2006