Aim Evidence on healthcare resource utilization (HCRU) for hospitalized patients with heart failure (HF) and reduced (HFrEF), mildly reduced (HFmrEF) and preserved (HFpEF) ejection fraction is limited.Methods and results We analysed HCRU in relation to left ventricular ejection fraction (LVEF) phenotypes, clinical features and in-hospital and 12-month outcomes in 16 943 patients hospitalized for HF in a worldwide registry. HFrEF was more prevalent (53%) than HFmrEF (17%) or HFpEF (30%). Patients with HFmrEF and HFpEF were older, more often women, with milder symptoms and more comorbidities, but differences were not pronounced. HCRU was high in all three groups; two or more in- and out-of-hospital services were required by 51%, 49% and 52% of patients with HFrEF, HFmrEF and HFpEF, respectively, and intensive care unit by 41%, 41% and 37%, respectively. Hospitalization length was similar (median, 8 days). Discharge prescription of neurohormonal inhibitors was <80% for each agent in HFrEF and only slightly lower in HFmrEF and HFpEF (74% and 67%, respectively, for beta-blockers). Compared to HFrEF, 12-month all-cause and cardiovascular mortality were lower for HFmrEF (adjusted hazard ratios 0.78 [95% confidence interval 0.59-0.71] and 0.80 [0.70-0.92]) and HFpEF (0.64 [0.59-0.87] and 0.63 [0.56-0.71]); 12-month HF hospitalization was also lower for HFpEF and HFmrEF (21% and 20% vs. 25% for HFrEF). In-hospital mortality, 12-month non-cardiovascular mortality and 12-month all-cause hospitalization were similar among groups.Conclusions In patients hospitalized for HF, overall HCRU was similarly high across LVEF spectrum, reflecting the subtle clinical differences among LVEF phenotypes during hospitalization. Discharge prescription of neurohormonal inhibitors was suboptimal in HFrEF and lower but significant in patients with HFpEF and HFmrEF, who had better long-term cardiovascular outcomes than HFrEF, but similar risk for non-cardiovascular events.