RESUMO
INTRODUCTION: Obesity is considered a risk factor in severe cases of COVID-19, which has been analysed using body mass index (BMI), an estimator that does not correlate adequately with body fat (BF) percentage. The aim of this study was to analyse the population attributable fraction to BF in severe forms of COVID-19 based on BMI and CUN-BAE. MATERIAL AND METHODS: Multicentre observational prevalence study. Sociodemographic information, personal history, BMI and CUN-BAE were collected in SARS-CoV-2 positive cases from the provinces of León and La Rioja. Logistic regression models were used to calculate odds ratios with their respective 95% confidence intervals adjusting for age and personal history, as well as the population attributable fraction to BF. RESULTS: Seven hundred eighty-five patients participated, 123 (15.7%) were severe. Age, obesity (both by BMI and CUN-BAE) and personal history were detected as risk factors. 51.6% of severe cases could be attributed to excess BMI and 61.4% to excess BF estimated according to CUN-BAE, with a higher underestimation of risk in women. CONCLUSIONS: Excess BF is a risk factor for severe forms of COVID-19 together with advanced age and the presence of cardiovascular, chronic respiratory or oncohematological diseases. BMI underestimates the risk especially in women, being CUN-BAE the predictor selected for its better estimation of the percentage of BF.