In a previous work, an automated alcohol dehydrogenase method for the quantification of ethanol in whole blood (blood) specimens was presented. In the present work the application of the method to urine specimens has been investigated. Also, method robustness to routine analysis of urine and blood specimens during a period of eight months is shown. The limits of detection and quantification for urine were 0.0012 g/dL and 0.0042 g/dL, respectively. Relative standard deviations for the repeatability and within-laboratory reproducibility were in the ranges 1.4-4.1% and 1.8-4.6%, respectively. The method was compared with two headspace gas chromatography-flame ionization detection methods using authentic forensic urine specimens (n = 305) and blood specimens (n = 3186). Passing-Bablok regression for the concentration range 0.01-0.48 g/dL (urine) and 0.002-0.40 g/dL (blood) showed a statistically significant difference, for urine y = 0.9313 (0.9250 - 0.9377)x + 0.0038 (0.0029-0.0044) and for blood y = 0.9493 (0.9491 - 0.9495)x + 0.0032 (0.00318-0.00323), at 95% confidence level. The results of the external quality control specimens were in accordance with the reported theoretical concentrations.