Usefulness of a modified poisoning severity score for predicting prognosis in acute carbon monoxide poisoning.

Abstract

The poisoning severity score (PSS) was developed to grade the severity of various types of poisoning. However, in its current form, it requires investigating many variables, some of which have been found not to be associated with carbon monoxide (CO) poisoning severity. Therefore, in this study, we modified the PSS for CO poisoning and compared its usefulness to that of the original PSS, as an early prognostic factor of short-term outcome in CO poisoning patients. This was a retrospective observational study conducted in patients with CO poisoning who visited the emergency department between January 2014 and December 2020. Patients' primary outcome was their Cerebral Performance Category (CPC) scale score at discharge, which classified those with CPC 1-2 as having a favorable outcome and those with CPC 3-5 as having a poor outcome. We calculated the patients' PSS and their CO-modified PSS by replacing blood and metabolic balance category in the original PSS with carboxyhemoglobin (COHb) and lactate levels, respectively. This study included 891 patients, of which 852 (95.6%) and 39 (4.4%) were classified into the favorable and poor outcome groups, respectively. Using multivariate analysis, the PSS (odds ratio [OR], 22.961; 95% confidence interval [CI], 10.641-49.546; p<0.001) and CO-modified PSS (OR, 28.856; 95% CI, 12.874-64.679; p<0.001) were both found to be associated with poor outcomes at hospital discharge. The areas under the receiver operating characteristic curves for the PSS and CO-modified PSS were 0.874 (95% CI, 0.850-0.895) and 0.881 (95% CI, 0.858-0.901), respectively. The CO-modified PSS, with fewer variables than the original PSS, was not inferior to predict poor outcomes, and if COHb level is considered together with other parameters, then it can be used both for predicting prognosis and in diagnosis.