Cox regression model of prognostic factors for delayed neuropsychiatric sequelae in patients with acute carbon monoxide poisoning: A prospective observational study.

OBJECTIVE: A major challenge for physicians is to identify patients with acute carbon monoxide (CO) poisoning who are likely to develop delayed neuropsychiatric sequelae (DNS). DNS is defined as neuropsychological sequelae that develops after 2-40 days of lucid interval after CO intoxication. Currently, there is no consensus on factors that predict the prognosis of CO poisoning. Thus, the purpose of this study was to identify factors predicting the development of DNS using a Cox regression model. METHODS: This prospective observational study included 310 CO-poisoned patients admitted to an emergency department in South Korea from July 2017 to February 2020. Demographic, clinical, and laboratory data were analyzed. Kaplan-Meier curves were constructed to estimate the cumulative incidence of DNS. A multivariate Cox regression model was used to identify the main predictors of the development of DNS. RESULTS: The incidence of DNS was 18.8 %, and the median onset time was 23.7 days (interquartile range, 14-30 days). The Kaplan-Meier survival curve showed that a serum creatine kinase (CK) level > 175.5 U/L and initial Glasgow Coma Scale (GCS) score ≤ 9 were associated with a higher cumulative incidence of DNS (log-rank test; p < 0.01 and p = 0.02, respectively). Cox regression analysis showed that a serum CK level > 175.5 U/L (hazard ratio [HR]: 2.862, 95 % confidence interval [CI]: 1.491-5.496; p < 0.01) and an initial GCS ≤ 9 (HR: 2.081, 95 % CI: 1.048-4.131; p = 0.04) were significant prognostic factors. CONCLUSION: In acute CO poisoning, an initial GCS score ≤ 9 and serum CK level > 175.5 U/L are significant predictors of DNS development.