Obesity is associated with reduced mortality in some patients hospitalized for heart failure (HF). In this analysis, we determine if this nonlinear relation, referred to as the obesity paradox, extends to secondary outcomes in patients diagnosed with severe obesity. This is a retrospective cohort study using the 2017 and 2018 National Inpatient Sample that includes adults hospitalized for HF. Patients with diagnosis codes specifying severe obesity, nonsevere obesity, or without obesity are compared. The primary outcome is mortality. Secondary outcomes include the length of stay (LOS), total charges, and cardiogenic shock (CS). Multivariate regression is used to adjust for demographics and co-morbidities. A total of 2,439,845 hospitalizations are included. A decreased mortality is found in nonsevere obesity (odds ratio 0.74, 95% confidence interval 0.69 to 0.80, p = 0.000), affirming the obesity paradox. However, this decreased mortality is not found in severe obesity (odds ratio 1.01, 95% confidence interval 0.94 to 1.08, p = 0.766). Severe obesity and nonsevere obesity are also associated with less CS and increased LOS compared with non-obese patients. Severe obesity is associated with increased total charges. In conclusion, a nonlinear, U-shaped relation between obesity and mortality in patients hospitalized for HF is demonstrated, where those not obese and those severely obese experience greater mortality compared with the nonseverely obese. However, for secondary outcomes of CS, LOS, and total charges, the relation is linear and therefore not interpreted as paradoxical. More information is needed using the adiposity-based chronic disease model to characterize complex relations between obesity and mortality.