Abstract
In previous studies, we have found that subjective nutritional assessment in critical-care patients shows short-term prognostic value. In these studies, nutritional assessments were performed by only one observer. The aim of the present study is to ascertain the prognostic value of subjective nutritional assessment of general medical patients when performed by different physicians. Training of the physician team was simple and consisted of two tutorial rounds of about 30 minutes each. The nutritional status of 394 patients (232 males and 162 females) was evaluated at admission by the physician on duty as follows: temporal muscle atrophy, Bichat's fat atrophy, upper and lower extremities muscle atrophy and subcutaneous fat atrophy; each parameter was categorized in three degrees and a nutritional score (the poorest value being 10 and 0 the best one) was obtained. Seventy-three patients died during this admission. Mortality rate was closely related to the nutritional score; 39% (39/99) of the patients with a score higher than 4 died vs only 12% with score values 4 or less. By logistic regression analysis, we found that the high relative risk of mortality (2.38; 95% IC 3.45–1.70) associated with a poor nutritional status is independent of other mortality-related variables such as age, serum albumin, leucocyte count, blood urea nitrogen and type of disease.