Multimodal intervention reduces seasonal carbon monoxide poisoning incidence and myocardial injury: a retrospective cohort study accounting for environmental temperature

Acute carbon monoxide poisoning remains a major global health issue, especially in winter due to unsafe heating. Despite known seasonal trends, few strategies effectively reduce both incidence and severity. This retrospective cohort study analyzed data from patients with non-occupational carbon monoxide poisoning across four consecutive winter seasons, defined as December of one year through February of the following year (e.g., winter 2021: December 2021–February 2022). The historical control cohort included cases from the first three winter seasons (2021–2023), while the intervention cohort consisted of cases from the 2024 winter season, during which an integrated digital and community-based education initiative was implemented. Poisson, linear, and logistic regression models were used, with adjustments for minimum temperature, diurnal temperature range, and baseline covariates. A total of 639 cases were included in the final analysis: 555 in the historical control cohort (winters 2021–2023) and 84 in the Intervention cohort (winter 2024). The intervention was associated with a 55.2% reduction in poisoning incidence (incidence rate ratio [IRR] = 0.448, 95% CI: 0.438–0.459; P < 0.001) and a 3.2% point decrease in median carboxyhemoglobin concentrations (P = 0.022). The incidence of myocardial injury was reduced by 69.2% (odds ratio [OR] = 0.31, 95% CI: 0.11–0.87; P = 0.026). Independent predictors of greater neurological severity, defined as coma, included male sex (aOR = 1.42, 95% CI: 1.03–1.95; P = 0.032) and increasing age (aOR = 1.02/year, 95% CI: 1.01–1.03; P < 0.001). Temperature effects were modest: each 1 °C decrease in minimum temperature was associated with a 0.5% increase in poisoning risk (IRR = 1.005, 95% CI: 1.003–1.007), while the diurnal temperature range had minimal influence (0.1% increase per 1 °C rise; IRR = 1.001, 95% CI: 1.000–1.003). The results demonstrate that the integrated digital and community-based educational intervention significantly reduced the incidence of acute carbon monoxide poisoning, along with carboxyhemoglobin levels and the rate of associated myocardial injury. These findings provide robust evidence to support the targeted implementation of such interventions during high-risk periods among vulnerable populations.