Impact of Nutritional Status on Morbidity in a Population of Geriatric Rehabilitation Patients

Abstract

The primary objective was to confirm the results of a prior study that demonstrated a strong independent correlation between the severity of protein-energy undernutrition and the risk of subsequent morbidity in a population of elderly rehabilitation patients. A second objective was to determine whether inadequate in-hospital nutrient intake is a co-contributor to the risk of subsequent morbidity. Cohort study. Geriatric Rehabilitation Unit (GRU) of a Veterans Administration hospital. Three hundred fifty randomly selected admissions to the GRU, of whom 99% were male, and 75% were white. The average age of the study patients was 76 years. At admission, each patient completed a comprehensive medical, functional, neuro-psychological, socio-economic, and nutritional assessment. While remaining in the hospital, each subject was monitored on a daily basis for the development of complications. Complete calorie counts were obtained at least every other day, and the average pre-complication daily nutrient intake was expressed as a percent of predicted requirements as determined using the Harris-Benedict equation. Of the 96 variables evaluated, the strongest predictor of subsequent complications was the Katz Index of ADL score, followed by serum albumin, usual weight percent, number of prescription medications, presence of renal disease, individual income, the presence of decubiti, dysphagia, and mid-arm muscle circumference. When all nine of these variables were included in the logistic regression analysis, the final model had a sensitivity of 70%, a specificity of 71%, and an overall predictive accuracy of 71%. When tested using the data from the previous study, the model differentiated the patients who developed a complication from those who had not a sensitivity of 76.7, a specificity of 76.1, and an overall predictive accuracy of 76.3. There was no difference in the pre-complication average daily nutrient intake between the complication and the no-complication groups (79% vs 75% of predicted requirements, P > 0.2). Protein-energy undernutrition appears to be a strong independent risk factor for in-hospital morbidity. However, in-hospital nutrient intake may not be a significant determinant of risk.