BACKGROUND The AHA 2010 guidelines for cardiopulmonary resuscitation (CPR) recommended that dispatchers should assertively provide compression-only CPR instructions to untrained rescuers. However, it is unclear whether dispatcher-instructed CPR recommended in the AHA 2005 guidelines can improve survival. METHODS We compared the effects of dispatcher-instructed CPR recommended in the 2005 CPR guidelines with those recommended in the 2000 CPR guidelines from the data of the All-Japan Utstein Registry, a prospective, nationwide, observational registry. The study endpoints were return of spontaneous circulation (ROSC) on hospital arrival and 30-day favorable neurological outcome after cardiac arrest. RESULTS Of the 58,448 adult patients with bystander-witnessed out-of-hospital cardiac arrest due to presumed cardiac etiology in the year 2005, 2008 and 2009, 23,165 who received instructions in CPR given by dispatchers over the telephone were included; 17,282 received the instructions of CPR recommended in the 2005 guidelines, and 5,883 received in the 2000 guidelines. The performance of dispatcher-instructed CPR (chest compression plus mouth-to-mouth ventilation or chest compression only) of the 2005 guidelines group showed a higher proportion than that of the 2000 guidelines group (42% vs.33%, p<0.0001), especially those with compression-only CPR (33% vs.18%, p<0.0001). The 2005 guidelines group had significantly higher rates of ROSC and favorable neurological outcome than the 2000 guidelines group in the whole cohort (ROSC; 12.2% vs. 8.4%, p<0.001, favorable neurological outcome; 5.3% vs. 2.9%, p<0.001), and in the subgroups of patients who received compression-only CPR (ROSC; 13.0% vs. 8.6%, p<0.001, favorable neurological outcome; 6.0% vs. 2.6%, p<0.001). The adjusted odds ratios after dispatcher-instructed CPR recommended in the 2005 guidelines was 1.56 (95% CI, 1.37-1.72) for ROSC and 2.17 (95% CI, 1.79-2.65) for favorable neurological outcome. CONCLUSION Dispatcher-instructed CPR recommended in the 2005 guidelines improved ROSC and favorable neurological outcome in adult patients with bystander-witnessed out-of-hospital cardiac arrest due to presumed cardiac etiology.