To investigate temporal variations in dispatcher-assisted and bystander-initiated resuscitation efforts and their association with survival after bystander-witnessed out-of-hospital cardiac arrests (OHCAs).

We retrospectively analyzed the neurologically favorable 1-month survival and the parameters related to dispatcher assisted cardiopulmonary resuscitation (DA-CPR) and bystander CPR (BCPR) for 227,524 OHCA patients between 2007 and 2013 in Japan. DA-CPR sensitivity for OHCAs, bystander's compliance to DA-CPR assessed by the proportion of bystanders who follow DA-CPR, and performance of BCPR measured by the rate of bystander-initiated CPR in patients without DA-CPR were calculated as indices of resuscitation efforts.

Performance of BCPR was only similar to temporal variations in the survival (correlation between hourly paired values, R²=0.263, P=0.01): a lower survival rate (3.4% vs 4.2%) and performance of BCPR (23.1% vs 30.8%) during night-time (22:00–5:59) than during non-night-time. In subgroup analyses based on interaction tests, all three indices deteriorated during night-time when OHCAs were witnessed by non-family (adjusted odds ratio, 0.73–0.82), particularly in non-elderly patients. The rate of public access defibrillation for these OHCAs markedly decreased during night-time (adjusted odds ratio, 0.49) with delayed emergency calls and BCPR initiation. Multivariable logistic regression analyses revealed that the survival rate of non-family-witnessed OHCAs was 1.83-fold lower during night-time than during non-night-time.

Dispatcher-assisted and bystander-initiated resuscitation efforts are low during night-time in OHCAs witnessed by non-family. A divisional alert system to recruit well-trained individuals is needed in order to improve the outcomes of night-time OHCAs witnessed by non-family bystanders.