The risk of life-threatening complications in a select population of geriatric patients: the impact of nutritional status.

Abstract

The primary objective of this study was to assess the interrelationship between protein-energy nutritional status, disease severity, and life-threatening complications in a population of elderly rehabilitation patients. Three-hundred and fifty randomly selected admissions to Geriatric Rehabilitation Unit of a Veterans Administration hospital were prospectively studied. The average age of the study subjects was 76 years, nearly all (99%) were male, and 75% were white. At admission, each patient completed a comprehensive medical, functional, neuropsychological, socioeconomic, and nutritional assessment. While remaining in the hospital, each subject was monitored on a daily basis for the development of complications. Of the 96 variables evaluated, the best predictors of developing at least one life-threatening complication were serum albumin, body mass index, the presence of renal disease (i.e. blood urea nitrogen > 30 mg/L), the Katz Index of Activities of Daily Living score, and the amount of weight loss in the year prior to admission. When all of these variables were included in the logistic regression analysis, the final model was highly significant by the −2Log Likelihood Chi-square goodness-of-fit criterion (Chi-square of 64.1 with 5 d.f., p < 0.0001) with a sensitivity of 77%, a specificity of 77% and an overall predictive accuracy of 77%. When the predictive accuracy of the logistic model was tested using a second sample of 110 patients, the model differentiated those who developed a life-threatening complication from those who had not with a sensitivity of 88%, a specificity of 61%, and an overall predictive accuracy of 65%. As indicated by the Chi-square test, these results were significant (p < 0.0001). Protein-energy undernutrition appears to be a strong independent risk factor for life-threatening morbidity during hospitalization.