BACKGROUND: Obesity [Body Mass Index (BMI) > 30 kg/m2] is a risk factor for disease conditions enhancing hospitalization and mortality risks, but higher BMI was paradoxically reported to reduce mortality in several acute and chronic diseases. Unintentional weight loss (WL) is conversely associated with disease development and may worsen patient outcome, but the impact of weight loss and its interaction with obesity in modulating risk of death in hospitalized patients remain undefined. METHODS: We investigated the ESPEN nutritionDay database of non-critically ill hospitalized patients to assess the impact of self-reported 3-month WL (WL1:2.5-6.6%; WL2: 6.6-12.6%, WL3: >12.6%) and its interaction with BMI in modulating 30-day in-hospital mortality. Multivariate Cox regression was used to estimate hazard ratios (HR), with stable weight (WL0) as reference category. RESULTS: In 110835 nDay patients, 30-day mortality increased with increasing WL. Male gender, increasing disease severity index PANDORA score (age, nutrient intake, mobility, fluid status, cancer and main patient group) and not having had surgery also predicted 30-day mortality. HR for 30-day mortality remained significantly higher compared to WL0 for WL2 and WL3 after multiple adjustment. Adjusted HR and its increments through increasing weight loss categories were comparable in lean (BMI<25), overweight (BMI 25-30) and obese individuals (BMI>30 kg/m2). Impact of gender, PANDORA score and surgery on 30-day mortality were conversely comparable in the three BMI groups. CONCLUSIONS: These results indicate that self-reported WL could represent a relevant prognostic factor in every hospitalized patient. Overweight and obesity per se have no protective impact against WL-associated mortality.
