Circadian patterns of out-of-hospital cardiac arrest over 23 years in a large urban area of northern Italy

Background

Out-of-hospital cardiac arrest (OHCA) remains a major public health challenge. Temporal patterns may have important implications for patient management and community strategies.

Purpose

To investigate the circadian distribution of OHCA occurrence, outcomes, and initial rhythm over 23 years in a large urban area of northern Italy.

Methods

OHCAs recorded in a population-based registry between 1999 and 2023 were included in the analysis. Time of OHCA was categorized into six 4-hour intervals. Circadian variation in OHCA occurrence was assessed using chi-square tests and Cramer’s V. Two separate multivariable logistic regression models were fitted to evaluate the associations of time intervals with mortality and with shockable rhythm, adjusting for age, sex, and ALS response time. Restricted cubic splines were used to model the non-linear associations between time of day and both mortality and shockable rhythm. Effect modification of the association between time of day and mortality by sex, age, initial rhythm, and ALS response time was assessed using interaction analyses, with age classified at a threshold of 65 years and ALS response time evaluated using cut-offs of 8 minutes (median) and 10 minutes.

Results

A total of 7,398 OHCAs were included (median age 79 [IQR 69-87] years, 43% women). Temporal distribution showed significant circadian variation (chi-square = 649.9, p < 0.001; Cramer’s V = 0.13), with a nocturnal nadir (00:00 - 03:59, 8.9%) and a morning peak (08:00 - 11:59, 23.9%).

In multivariable analyses, mortality was higher during nighttime (00:00–03:59: OR 1.91, 95% CI 1.10–3.54) and early morning hours (04:00–07:59: OR 2.23, 95% CI 1.36–3.79), with late afternoon arrests (16:00–19:59) as the reference category (Table). In the model for rhythm at presentation, early morning arrests (04:00–07:59) were associated with lower odds of a shockable rhythm compared with nighttime arrests (00:00–03:59) (OR 0.55, 95% CI 0.39–0.79). Restricted cubic spline curves showed significant non-linear associations between time of day and the outcomes (both p < 0.001) (Figure). Moreover, in both models, older age, female sex, and longer ALS response time showed independent associations with higher mortality and lower odds of shockable rhythm, respectively.

No significant effect modification by sex, age ≥65 years, or initial rhythm was observed. A borderline interaction was observed for ALS response time ≥8 minutes (p = 0.10), whereas a significant interaction was detected for ALS response time ≥10 minutes (p for interaction = 0.02).

Conclusions

OHCA occurrence shows significant circadian variation. Nighttime and early morning arrests were associated with higher mortality and lower odds of shockable rhythm. Excess nighttime mortality was more pronounced among patients with delayed ALS response, highlighting the importance of timely prehospital advanced care.