Association of anthropometric and biochemical markers with length of stay and mortality in the hospital.

ABSTRACT

INTRODUCTION: Malnutrition is associated with patient outcome, hospital stay and costs. The objective of our study was evaluated the relationship of different anthropometric and biochemical nutritional markers with hospital stance (days) and mortality. PATIENTS AND METHODS: A sample of 3087 hospitalized patients in a Universitary Hospital was enrolled. Nutritional status was assessed from anthropometric variables; body mass index (BMI), weight, weight loss in previous three months, tricipital skin fold thickness, midarm muscle circumference and midarm muscle area. Biochemical evaluation was assessed with albumin, prealbumin, transferrin and lymphocytes. Length of hospital stance and mortality were recorded. RESULTS: A total of 3087 patients were enrolled, mean age was 67.7±18.3 years, weight 63.3±15,7 kg and BMI 23.7±6.8, with a weight loss (3 months) of 6.3±4.6 kg. Length of stay was 24.7±22.1 days. A total of 2583 patients were discharged. Hospital discharge data showed 87.8% of patients went home, and a 12.2% of patients were discharged to a secondary Hospital. A (n=504) 16.3% of patients died. In the multivariate analysis with a dependent variable [length of stay (days)], only albumin levels remained as an independent predictor in the model (F=2.9; p < 0.05), with an increase of 3.1 days in hospital stay (CI 95% 0.4-5.8) with each decrease of 1 g/dl of albumin. In multivariate analysis, an independent factor that decrease mortality was high albumin levels (hazard ratio 0.41; 95% CI 0.22-0.80), adjusted by age and sex. CONCLUSIONS: Our study shows a significant association among serum albumin levels with length of stay in hospital and mortality.