Prognostic nutritional index in relation to hospital stay in women with gynecologic cancer

Objective: To estimate the prevalence of malnutrition, correlate it with le ngth of hospital stay, and evaluate laboratory tools to define it in gyneco logic oncology. Methods: Sixty-seven consecutive hospitalized gynecologic oncology patients were evaluated prospectively using the standardized Prognostic Nutritional Index method, based on serum albumin, transferrin, triceps skin fold and s kin sensitivity tests, which defines criteria for malnourished and nourishe d patients. It was correlated with length of hospital stay. The Mann-Whitne y test and Pearson's correlation coefficient were used to evaluate statisti cal relationships. Results: Cancer distribution among study subjects was 39 cervical (58%), 16 uterine (24%), 11 ovarian (16%), and one vulvar (2%). Malnutrition was fou nd in 36 of 67 women (54%; 95% confidence interval [CI] 41%, 66%). The medi an (interquartile range) hospital stays of nourished women (n = 31) and mal nourished women (n = 36) were 6 (range 4-7) days and 8 (range 6-16) days, r espectively (two-sided P = .004). That difference remained after controllin g for age, extent of metastases, and cancer sites. Albumin correlated well with Prognostic Nutritional Index (R = -.78; 95% CI -.86, -.66; P < .001). Albumin also correlated with length of hospital stay R = -.41; 95% CI -.56, -.25; P < .001). Conclusion: Malnutrition is common in gynecologic oncology patients and con tributes to longer hospital stays. Albumin is a good substitute for the Pro gnostic Nutritional Index laboratory test for assessing malnutrition. (Obst et Gynecol 2000;95:844-6. (C) 2000 by The American College of Obstetricians and Gynecologists).