Midregional pro-adrenomedullin (MR-proADM) is a vasoactive peptide with key roles in reducing vascular hyperpermeability and thereby improving endothelial stability during infection. While MR-proADM is useful for risk stratification in patients with sepsis, clinical data about prediction accuracy in patients with severe acute respiratory syndrome coronavirus 2 disease (COVID-19) is currently missing. We included consecutively adult patients hospitalized for confirmed COVID-19 at a tertiary care center in Switzerland between February and April 2020. We investigated the association of MR-proADM levels with in-hospital mortality in logistic regression and discrimination analyses. Of 89 included COVID-19 patients, 19% (n=17) died while in the hospital. Median admission MR-proADM levels (nmol/L) were increased almost 1.5-fold increased in non-survivors compared to survivors (1.3 interquartile range IQR 1.1-2.3) vs. 0.8 IQR 0.7-1.1) and showed good discrimination (area under the curve 0.78). An increase of 1 nmol/L of admission MR-proADM was independently associated with a more than fivefold increase in in-hospital mortality (adjusted odds ratio of 5.5, 95% confidence interval 1.4-21.4, p=0.015). An admission MR-proADM threshold of 0.93 nmol/L showed the best prognostic accuracy for in-hospital mortality with a sensitivity of 93%, a specificity of 60% and a negative predictive value of 97%. Kinetics of follow-up MR-proADM provided further prognostic information for in-hospital treatment. Increased levels of MR-proADM on admission and during hospital stay were independently associated with in-hospital mortality and may allow a better risk stratification, and particularly rule-out of fatal outcome, in COVID-19 patients.