Discrepancies and Validation of Ethanol Level Determination with Osmolar Gap Formula in Patients with Suspected Acute Poisoning

Jung, Haewon; Lee, Mi Jin; Cho, Jae Wan; Ahn, Jae Yun; Kim, Changho

Abstract

PURPOSE
Osmolar gap (OG) has been used for decades to screen for toxic alcohol levels. However, its reliability may vary due to several reasons. We validated the estimated ethanol concentration formula for patients with suspected poisoning and who visited the emergency department. We examined discrepancies in the ethanol level and patient characteristics by applying this formula when it was used to screen for intoxication due to toxic levels of alcohol.
METHODS
We retrospectively reviewed 153 emergency department cases to determine the measured levels of toxic ethanol ingestion and we calculated alcohol ingestion using a formula based on serum osmolality. Those patients who were subjected to simultaneous measurements of osmolality, sodium, urea, glucose, and ethanol were included in this study. Patients with exposure to other toxic alcohols (methanol, ethylene glycol, or isopropanol) or poisons that affect osmolality were excluded. OG (the measured-calculated serum osmolality) was used to determine the calculated ethanol concentration.
RESULTS
Among the 153 included cases, 114 had normal OGs (OGâ‰¤14 mOsm/kg), and 39 cases had elevated OGs (OG>14). The mean difference between the measured and estimated (calculated ethanol using OG) ethanol concentration was âˆ’9.8 mg/dL. The 95% limits of agreement were âˆ’121.1 and 101.5 mg/dL, and the correlation coefficient R was 0.7037. For the four subgroups stratified by comorbidities and poisoning, the correlation coefficients R were 0.692, 0.588, 0.835, and 0.412, respectively, and the mean differences in measurement between the measured and calculated ethanol levels were âˆ’2.4 mg/dL, âˆ’48.8 mg/dL, 9.4 mg/dL, and âˆ’4.7 mg/dL, respectively. The equation plots had wide limits of agreement.
CONCLUSION
We found that there were some discrepancies between OGs and the calculated ethanol concentrations. Addition of a correction factor for unmeasured osmoles to the equation of the calculated serum osmolality would help mitigate these discrepancies.