RESUMEN
BACKGROUND AND AIM: The prevalence of hospital malnutrition (HM) is variable, explained by the variability of patients, the nutritional evaluation method used among others. The aim is to determine the frequency of malnutrition in hospitals in Latin America, and estimate its association with mortality and length of hospital stay. METHODS: This is an analytical, observational cohort study that included 7,973 patients of both genders, 18 and older, who provided their consent. The survey was administered during the first three days of admission. The nutritional status was estimated using Subjective Global Assessment (SGA) and the Nutrition Risk Screening (NRS), body mass index (BMI), percentage of change of weight (PCW) and co-morbidities. Serum albumin was obtained from the clinical chart. Length of stay (LOS) and the survival status at discharge (dead or alive) were also recorded. RESULTS: By SGA: 10.9% had severe malnutrition and 34% moderate malnutrition. By NRS: 36.9% had nutritional risk. Univariate analysis showed that NRS score and serum albumin were prognostic factors for mortality: NRS 3-4 (OR: 2.3, 95% CI: 1.9-2.8), NRS 5-7 (OR: 5.8, 95% CI: 4.9-6.9), serum albumin < 2.5 g/dl, (OR: 2.9, 95% CI: 2.2-3.8). These results were consistent and similar to a multivariate analysis. Both NRS and serum albumin were also independently and clinically associated to LOS. CONCLUSIONS: The prevalence of hospital malnutrition in Latin America is high. Our results show that screening with NRS and serum albumin can identify hospital malnutrition as well as providing clinically relevant prognostic value.