Ethanol elimination rates in an ED population

Knowledge of the rate of ethanol elimination is essential in the assessment of the intoxicated patient. Surprisingly little literature is available regarding ethanol elimination rates in emergency department (ED) patients; prior studies almost exclusively examined populations of alcoholics or normal controls. Consequently, this prospective observational study was undertaken to assess the rate of ethanol elimination in an ED population. Twenty-four consecutive adult ED patients clinically suspected of intoxication who had serum ethanol determinations drawn were enrolled. Patients underwent serial ethanol determinations via breathalyzer (Intoxilyzer 1400, CMI Inc., Owensboro, KY). Linear regression analysis of the plot of decrease in ethanol level over time was performed to determine the rate of ethanol elimination. Initial ethanol levels in the 24 patients ranged from 58 to 447 mg/dL (mean, 249 ± 109 [SD] mg/dL). Patients were observed for a minimum of 2 and a maximum of 9 observations (mean, 3.9 ± 1.7), over a period of 0.5 to 12.1 hours (mean, 4.4 ± 3.5 h). Clinical features of intoxication were poorly correlated with ethanol level (r< .5). The rate of ethanol elimination in the ED population was 19.6 mg/dL/h (r = .83; 95% confidence interval [Cl], 16.9 to 22.3 mg/dL/h). Subgroup analysis found differences that were statistically significant but small. Multiple regression analysis showed that time was the major variable useful in predicting changes in ethanol level (P< .001). The data did not provide sufficient evidence to indicate that a history of chronic use (P = .69), initial level less or greater than 250 mg/dL (P = .83), or level of ethanol (P = .80), when used in a regression equation along with time, are useful variables in predicting the change in ethanol level. In conclusion, the rate of ethanol elimination in an ED population is 19.6 (95% Cl, 16.9 to 22.3) mg/dL/h, corresponding closely to rates reported in the literature for other populations; subgroups have statistically significant but clinically small differences. The prediction of ethanol elimination was not shown to be enhanced by consideration of other parameters; the lack of predictive value of ethanol level implies a zero-order model.