BMI, sex and outcomes in hospitalised patients in western Sweden during the COVID-19 pandemic.

High body mass index (BMI) is associated with severe COVID-19 but findings regarding the need of intensive care (IC) and mortality are mixed. Using electronic health records, we identified all patients in western Sweden hospitalised with COVID-19 to evaluate 30-day mortality or assignment to IC. Adjusted logistic regression models were used to estimate odds ratios (OR) and 95% confidence intervals (CI) for outcomes. Of totally 9761 patients, BMI was available in 7325 (75%), included in the study. There was a marked inverse association between BMI and age (underweight and normal weight patients were on average 78 and 75 years, whereas overweight and obese were 68 and 62 years). While older age, male sex and several comorbidities associated with higher mortality after multivariable adjustment, BMI did not. However, BMI ≥ 30 kg/m2 (OR 1.46, 95% CI 1.21-1.75) was associated with need of IC; this association was restricted to women (BMI ≥ 30; OR 1.96 (95% CI 1.41-2.73), and not significant in men; OR 1.22 (95% CI 0.97-1.54). In this comprehensive hospital population with COVID-19, BMI was not associated with 30-day mortality risk. Among the obese, women, but not men, had a higher risk of assignment to IC.