COVID-19 is a severe respiratory disease affecting millions worldwide, causing significant morbidity and mortality. Adrenomedullin (bio-ADM) is a vasoactive hormone regulating the endothelial barrier and has been associated with COVID-19 mortality and other adverse events. This prospective cohort study included 119 consecutive patients with verified SARS-CoV-2infection admitted to the intensive care unit (ICU). Bio-ADM was retrospectively analysed from plasma on ICU admission, day2, and day 7. Information on comorbidities, adverse events and mortality was collected. The primary outcome was 90-daymortality, and secondary outcomes were markers of disease severity. The association between bio-ADM and outcomes was analysed using survival analysis, logistic regression and receiver operating characteristics curves. Bio-ADM predicts 90-daymortality with an area under the receiver operating Curve (AUC) of 0.66, 0.68, and 0.70 for sampling on admission, day 2 and day 7. Age alone predicted 90-day mortality with an AUC of 0.78. The AUCs for age combined with bio-ADM were 0.81, 0.81,and 0.82 (n.s.). Bio-ADM predicted the need for renal replacement therapy with AUCs of 0.77, 0.82, and 0.81. Bio-ADM on ICU admission, day 2 and day 7 predicted 90-day mortality and dialysis needs, which highlights the importance of bio-ADM inCOVID-19 pathophysiology. Bio-ADM could be used to triage patients with a risk of adverse outcomes and as a potential target for clinical interventions