A regression model applied to gender-specific ethanol elimination rates from blood and breath measurements in non-alcoholics

Abstract

As elimination rates for alcohol are suggested to be gender specific, a novel regression model has been applied to estimate these rates for both men and women using experimentally measured data from 81 female and 96 male volunteers described in previous papers. Breath alcohol measurements were done with the Alcotest 7110 Evidential device and were coupled with concomitant sampling of venous blood. Statistical analyses involved use of a mixed linear model for blood alcohol concentration (BAC) and breath alcohol concentration (BrAC), respectively. The model takes regression lines for each test subject into account with an individual starting value (2 h after the end of drinking) and with an individual alcohol elimination rate per hour (coincidental effects). Further, the data was modeled so that an average alcohol elimination rate per hour could be estimated separately for both genders (constant effects). This enables us to methodically correctly estimate the back calculation. The elimination rates beta (60), which can be used for minimum and maximum back calculations for the BAC, were 0.115 g/kg/h and 0.260 g/kg/h, respectively, for women and 0.096 g/kg/h and 0.241 g/kg/h, respectively, for men. These figures widely deviate from gender-unspecific values commonly used in Germany (0.1 and 0.2 g/kg/h, respectively). The corresponding values for the BrAC were 0.061 mg/l/h and 0.124 mg/l/h for women and 0.049 mg/l/h and 0.112 mg/l/h for men. The probability of an over- or underestimation of the abovementioned extreme values is 0.3% in each case.