Serum albumin as indicator of clinical evolution in patients on parenteralnutrition. Multivariate study

Background and aims: The aim of this study is to establish whether serum al bumin concentration at the beginning of parenteral nutrition is related to morbidity and mortality. Methods: In this cohort study spanning four years, a number of patients wer e classified into twelve groups, depending on their clinical status at the beginning of parenteral nutrition. Their serum albumin concentration and ot her clinical parameters were then measured and twelve multiple logistic reg ression models were thus generated in order to model the relationship betwe en initial albumin concentration and risk of morbidity/mortality. Results: 1953 (84%) of the 2321 patients studied were hypoalbuminemic. In s ix models, this condition was associated with a significant increase in the risk of nosocomial infection. However, no model could be associated to sig nificant risk of renal failure, and only patients with previous hepatopathy were at risk for hepatic failure. In seven models, there was a significant increase in mortality. Conclusions: Serum albumin concentration at the beginning of parenteral nut rition is related to mortality and morbidity associated with nosocomial inf ection in some groups of the study. (C) 2001 Harcourt Publishers Ltd.