Experience of carbon monoxide poisoning and the outcome predicting score: A multicenter retrospective study.

BACKGROUND

Carbon monoxide poisoning (COP), resulting from accidental and intentional exposure, is a leading cause of fatal poisoning worldwide. Except for early death, neurological sequelae are common and impose a large burden on patients, caregivers, and the society.

MATERIALS AND METHODS

This retrospective study included patients who visited the emergency departments (EDs) of the medical institutes of Chang Gung Memorial Hospital after COP with a carboxyhemoglobin level > 10% between January 2009 and October 2018. Patients who experienced out-of-hospital cardiac arrest (OHCA) were excluded. Poor outcome was defined as mortality or a Glasgow coma scale (GCS) <13 at discharge. Stepwise regression analysis was performed, and a receiver operating characteristic (ROC) curve was applied to analyze our newly created scoring system for prognosis prediction.

RESULTS

This study enrolled 1171 patients. Fire scene (F) (aOR, 20.635; 95% CI, 8.345-51.023), intentional CO exposure (I) (aOR, 2.634; 95% CI, 1.335-5.196), respiratory failure (R) (aOR, 9.944; 95% CI, 5.533-17.873), every point of reduced GCS (E) (aOR, 1.253; 95% CI, 1.186-1.323), and diabetes mellitus (D) (aOR, 2.749; 95% CI, 1.201-6.292) were identified as predictors of poor outcomes. The FIRED score was created.

CONCLUSION

The FIRED score could predict the outcomes of non-OHCA patients with a carboxyhemoglobin level > 10% after COP using five factors that can be obtained by history taking and basic examination. An FIRED score ≥ 10 was associated with a poor outcome (sensitivity, 89.6%; specificity, 82.4%; AUC0.930).