Clinical features and predictors of delayed neurological syndrome in carbon monoxide poisoning: the AMICO study.

OBJECTIVES

To identify predictors for developing delayed neurological syndrome (DNS) after an initial episode of carbon monoxide (CO) poisoning in the interest of detecting patients most likely to develop DNS so that they can be followed.

MATERIAL AND METHODS

Retrospective review of cases of CO poisoning treated in the past 10 years in the emergency departments of 4 hospitals in the AMICO study (Spanish acronym for the multicenter analysis of CO poisoning). We analyzed demographic characteristics of the patients and the clinical characteristics of the initial episode. The records of the cohort of patients with available follow-up information were reviewed to find cases of DNS. Data were analyzed by multivariant analysis to determine the relationship to characteristics of the initial exposure to CO.

RESULTS

A total of 240 cases were identified. The median (interquartile range) age of the patients was 36.2 years (17.6-49.6 years); 108 patients (45.0%) were men, and the poisoning was accidental in 223 cases (92.9%). The median carboxyhemoglobin concentration on presentation was 12.7% (6.2%-18.7%). Follow-up details were available for 44 patients (18.3%). Eleven of those patients (25%) developed DNS. A low initial Glasgow Coma Scale score predicted the development of DNS with an odds ratio (OR) of 0.61 (95% CI, 0.41-0.92) and an area under the receiver operating characteristic curve of 0.876 (95% CI, 0.761-0.990) (P .001).

CONCLUSION

The initial Glasgow Coma Scale score seems to be a clinical predictor of DNS after CO poisoning. We consider it important to establish follow-up protocols for patients with CO poisoning treated in hospital EDs.