BACKGROUND: The patient characteristics, pre-hospital interventions, and outcomes of out-of-hospital cardiac arrests (OHCA) occurring during exercise, have not been sufficiently investigated among the general population. METHODS: OHCA data from 2009 to 2015 were obtained from the population-based OHCA registry in Osaka City, Japan. Patients who suffered OHCA, which occurred during exercise before the arrival of emergency medical service personnel, were included. The primary endpoint was one-month survival with a favourable neurological outcome after OHCA, defined using the Glasgow-Pittsburgh cerebral performance category scale 1 or 2. RESULTS: During the 7-year study period, 16,278 OHCAs were observed, and 52 (0.3%) occurred during exercise (male, n=41 [79%]; median age, 62 years). These incidents occurred mainly during running activities (n=14), followed by swimming (n=8), dance/social dance (n=6), tennis (n=4), and weight training (n=3). Within these exercise-related OHCAs, 47 (90%) were of cardiac origin, 45 (87%) were bystander-witnessed cardiac arrests, 49 (94%) received bystander-initiated cardiopulmonary resuscitation, and 30 (57%) received public-access defibrillation (PAD). Overall, 56% (29/52) had one-month survival with a favourable neurological outcome after OHCA, which was significantly higher among OHCAs of cardiac origin with PAD (77%, 23/30) than among those of cardiac origin without PAD (35%, 6/17) and among those of non-cardiac origin (0%, 0/5) (p<0.001). CONCLUSION: In Osaka, OHCAs during exercise represented a small subset of the overall OHCA burden, but occurred during a wide variety of exercise activities. Patients with OHCA of cardiac origin had a good prognosis, and PAD played an important role in improving patient outcomes.