Introduction: Carbon monoxide (CO) poisoning is a crucial cause of delayed neuropsychiatric syndrome (DNS). However, most biomarkers are not satisfactory for the prediction of DNS caused by CO poisoning. Thus, we evaluated the adequacy of the serum glucose/potassium (GLU/K) ratio, which may be an easy, quick, and readily available parameter that can be used in the emergency department for predicting DNS. Methods: We evaluated 281 patients who were admitted to our emergency department between January 2012 and December 2018. The patients were divided into two groups: DNS (+) and DNS (-). The GLU/K was compared for the groups. Results: Glucose, blood urea nitrogen, carboxyhemoglobin, and GLU/K ratios of patients in the DNS (+) group were statistically significantly higher than those patients in DNS (-) group (140 +/- 34 vs. 110 +/- 24,p< 0.001; 17.58 +/- 6.14 vs. 14.27 +/- 5.08,p= 0.003; 29 +/- 5.1 vs. 18.9 +/- 7.6,p< 0.001; and 38.35 +/- 10.11 vs. 28.65 +/- 6.53,p< 0.001, respectively). The area under the curve for GLU/K to predict DNS was measured as 0.791, and 35.9 as a cut-off value had 63.6% sensitivity and 89.6% specificity. Conclusions: DNS development in CO poisoning is a serious and feared complication. We suggest that the GLU/K ratio has a high potential as a rapid, easy preliminary marker for the exclusion of patients who will not subsequently develop DNS.