© 2018 Elsevier B.V. Background: Current cardiopulmonary resuscitation (CPR) guidelines do not define the optimal type of CPR (chest compression-only CPR [CCCPR] or conventional CPR with rescue breathing [CCRB]) to be performed by bystanders when they witness someone collapse. Methods: Using a nationwide database of 1.17 million patients who underwent out-of-hospital cardiac arrest (OHCA) in Japan, we enrolled consecutive bystander-witnessed OHCAs of medical origin with resuscitation attempts from January 2005 through December 2014. Multivariable logistic regression analysis was used to assess the association between the type of bystander CPR and the OHCA outcome after one-to-one propensity score matching for CCCPR versus CCRB. The primary outcome measure was one-month survival with a favorable neurological outcome, defined as a cerebral performance category of 1 or 2. Results: Among 143,500 eligible patients with bystander-witnessed OHCAs receiving bystander-initiated CPR, 71.4% received CCCPR and 28.6% received CCRB. In the univariate analysis, the proportion of one-month survival cases with favorable neurological outcome was lower in the CCCPR group than the CCRB group (5.6% [5749/102,487] vs. 6.5% [2682/41,013], odds ratio [OR]; 0.85 [95% confidence interval {CI}; 0.81–0.89]). However, in the multivariate analysis, the CCCPR group showed a more favorable neurological outcome than the CCRB group (adjusted OR 1.12, 95% CI; 1.06–1.19). In the propensity-matched cohort, the CCCPR group also showed a more favorable neurological outcome than the CCRB group (7.2% [2894/40,096] vs. 6.5% [2610/40,096], adjusted OR 1.14, 95% CI; 1.09–1.22). Conclusions: CCCPR is an acceptable resuscitation technique for lay-rescuers responding to bystander witnessed OHCA of presumed medical origin.