Abstract Background Despite overweight/obesity being a well-known cardiovascular (CV) risk factor, high body mass index (BMI) has been suggested to be associated with longer survival in patients with heart failure (HF). Data on large populations are lacking, as well as whether this advantage might be consistent across the ejection fraction (EF) spectrum, i.e. reduced (HFrEF), mildly reduced (HFmrEF) and preserved (HFpEF, where obesity may have a greater causative role). Purpose To assess the association between BMI and outcomes in a large and unselected contemporary cohort of patients with HF across the EF spectrum. Methods Patients with HF registered in the Swedish HF Registry (SwedeHF) between May 2000 and December 2021, with no missing values for EF and BMI, were selected. Linkage with the Cause of Death Registry and the National Patient Registry provided data on mortality, hospitalizations and comorbidities. Patients were categorised according to BMI class (underweight <18.5 kg/m2; normal weight 18.5-24.9 kg/m2; overweight 25.0-29.9 kg/m2; obesity>30-34.9 kg/m2). The primary outcome was all-cause death, secondary outcomes were CV death and HF hospitalization (HHF). The associations between BMI categories and outcomes were analysed by univariable and multivariable Cox proportional hazards models, both in the whole cohort and according to EF classes; this association was also explored by modeling BMI as a continuous variable through restricted cubic splines. Missing data were handled by chained equation multiple imputation. Results 69,874 patients (mean age 74 years, 36% females) were included, of whom 3% underweight, 35% with normal weight, 35 % overweight, obese 27%. Figure 1 reports the proportions of the different BMI categories according to EF: obesity was significantly more common in HFpEF. Regardless of EF, obese patients were younger, with lower N-terminal B-type natriuretic peptide, higher New York Heart Association class, and with diabetes and hypertension as compared with the other three BMI classes. Compared with normal weight, both overweight and obesity were independently associated with lower mortality risk, while underweight with higher mortality, with consistent results in HFrEF, HFmrEF and HFpEF; the same associations were found for CV death or all three EF classes, while underweight only was associated with HHF in HFrEF. When BMI was modelled as restricted cubic splines, a reverse J-shaped adjusted association was found between BMI and risk of all-cause death (Figure 2). Conclusion HF patients with overweight and obesity had a lower mortality risk compared with patients with normal or underweight patients across all EF classes, at least up to BMI>40 kg/m2. Our results are consistent with the previously suggested potential survival benefit in overweight patients, calling for future research to understand this apparent paradox.Figure 1