To investigate whether the Edmonton Frail Scale (EFS), a multidimensional frailty assessment tool, improves the prediction of 30-day or in-hospital mortality over the use of the European System for Cardiac Outcome Risk Evaluation (EuroSCORE) II alone. Single-center prospective observational study. University hospital. Patients aged 75 years or older undergoing cardiac surgery between February 2014 and May 2017. No intervention was performed. The EFS was administered the day before surgery. The primary endpoint was 30-day or in-hospital mortality. Secondary endpoints were times to discharge from the intensive care unit (ICU) and from the hospital, discharge to a health care facility, and ability to return home by postoperative day 30. The EFS had a good discriminative ability for 30-day mortality (area under the receiver operating characteristic curve = 0.69; 95% confidence interval [CI], 0.56-0.82). Adding frailty, defined by an EFS ≥8, to the EuroSCORE II significantly improved the prediction of 30-day (p = 0.04) mortality. The integrated discrimination index was 0.03 (95% CI, 0.01-0.06, p = 0.01), meaning that the difference in predicted risk between patients who died and those who survived increased by 3% due to the addition of frailty determined by the EFS to the EuroSCORE II. Frailty also was associated significantly with a decreased cumulative probability of discharge from the ICU (p = 0.02) and an increased incidence of discharge to a health care facility (p = 0.01). The EFS has a good predictive ability for 30-day mortality after cardiac surgery in elderly patients and improves the prediction of 30-day mortality over the use of the EuroSCORE II.