Aim Recent data suggest that guideline‐directed medical therapy of patients with heart failure (HF) with reduced ejection fraction (HFrEF) might improve clinical outcomes in patients with HF up to a left ventricular ejection fraction (LVEF) of 55–65%, whereas patients with higher LVEF do not seem to benefit. Recent data have shown that LVEF may have a U‐shaped relation with outcome, with poorer outcome also in patients with supranormal values. This suggests that patients with supranormal LVEF may be a distinctive group of patients. Methods and results RELAX‐AHF‐2 was a multicentre, placebo‐controlled trial on the effects of serelaxin on 180‐day cardiovascular (CV) mortality and worsening HF at day 5 in patients with acute HF. Echocardiograms were performed at hospital admission in 6128 patients: 155 (2.5%) patients were classified as HF with supranormal ejection fraction (HFsnEF; LVEF >65%), 1440 (23.5%) as HF with preserved ejection fraction (HFpEF; LVEF 50–65%), 1353 (22.1%) as HF with mildly reduced ejection fraction (HFmrEF; LVEF 41–49%) and 3180 (51.9%) as HFrEF (LVEF <40%). Patients with HFsnEF compared to HFpEF were more often women, had higher prevalence of non‐ischaemic HF, had lower levels of natriuretic peptides, were less likely to be treated with beta‐blockers and had higher blood urea nitrogen plasma levels. All‐cause mortality was not statistically different between groups, although patients with HFsnEF had the highest numerical rate. A declining trend was seen in the proportion of 180‐day deaths due to CV causes from HFrEF (290/359, 80.8%) to HFsnEF (14/24, 58.3%). The reverse was observed with death from non‐CV causes. No treatment effect of serelaxin was observed in any of the subgroups. Conclusions In this study, only 2.5% of patients were classified as HFsnEF. HFsnEF was primarily characterized by female sex, lower natriuretic peptides and a higher risk of non‐CV death.