| Abstract: | The prevalence of malnutrition in Veterans Administration Hospitals has been well documented. Several methods have been proposed to assess nutritional status including prognostic nutritional index (PNI) and instant nutritional assessment (INA). A prospective study was done to evaluate the currently used nutritional assessments and determine their efficacy based on sensitivity and specificity in predicting surgical morbidity and mortality. Data on 46 patients were evaluated for the multiparameter index of PNI, total lymphocyte count and serum albumin for INA, and weight loss alone. In this analysis, intermediate and high risk PNI were combined as positive predictor of complications. Any abnormal value in INA was considered positive as was weight loss of greater than 6% of usual body weight. Sensitivity and specificity of each assessment method were determined by 2 X 2 contingency table, and significance of observed differences between methods was determined by chi 2 analysis. There were no complications or deaths in patients with less than 6% weight loss. All three patients with abnormal albumin and total lymphocyte count had complications as compared to only three of 32 patients when both of these parameters were normal. The PNI was also able to predict complications with an increasing incidence as the PNI increased. Only the difference between specificity of weight loss alone vs INA was statistically significant, p less than 0.05. Thus, weight loss alone can be used as a rapid, inexpensive assessment of nutritional status for predicting postoperative complications. |
