Ethanol is the only macronutrient that is thought to be absorbed to any extent in the stomach. Much of the evidence that this occurs in humans is derived indirectly and the fraction of ethanol that is usually absorbed in the stomach from ingested alcoholic drinks is controversial. We compared, with ethics committee approval, the amount of ethanol that was absorbed in the stomach from a 2 % (434 mmol l^-1) ethanol solution (2E), a 6 % (1302 mmol l^-1) ethanol solution (6E) and a solution containing 2 % (434 mmol l^-1) ethanol and 6.83 % (379 mmol l^-1) glucose (EG) over a 60 min period. The calculated energy density of 6E and EG was the same (1.74 kJ ml^-1) while that of 2E (0.58 kJ ml^-1) was lower. Nine healthy subjects (4 males, 5 females), who were assessed as being moderate drinkers, with a mean (S.D.) age of 22 ± 2 years, height of 169 ± 11 cm and body mass of 73 ± 11 kg volunteered for this study. All subjects gave written, informed consent before entering the study. The total and test solution volumes remaining in the stomach were measured at 10 min intervals after ingestion using a double sampling technique (Beckers et al. 1988). The calculated amount of ethanol in the gastric fluid was estimated from the alcohol concentration of the original test solution and the volume of test solution remaining in the stomach at each time point. The actual amount of ethanol in the stomach was determined from the measured alcohol concentration of the gastric aspirates at each time point and the total fluid volume. The amount of ethanol absorbed by the gastric mucosa was estimated as the difference between the calculated and measured ethanol content at each time point. The ethanol concentration of test solutions and aspirate samples was measured spectrophotometrically using an alcohol dehydrogenase assay kit (Sigma, UK). Subjects, who were all familiar with the study procedures, reported at least 6 h fasted to the laboratory on three different occasions, each separated by a minimum of 7 days. Once the gastric aspiration tube was in place, 500 ml of one of the test solutions was rapidly instilled into the stomach. The treatment order was randomised using a Latin square order design. The total and test solution data were not normally distributed, and statistical analysis was carried out using the Kruskal-Wallis test and the Wilcoxon signed rank test where appropriate.

As expected, the volume of test solution remaining in the stomach was similar over the 60 min measurement period between 6E and EG (P = 0.12), and greater than that of 2E (P = 0.001). The median (range) time to empty half the ingested test solution on trial 6E (20.3 (11.4-35.3) min) was similar (P = 0.66) to that on trial EG (21.3 (13.6-33.2) min): 6E and EG were both longer (P < 0.001) than that on trial 2E (9.6 (3.5-18.6) min). The calculated and measured ethanol values matched at the same time points on trial 2E (P = 0.71), 6E (P = 0.70) and EG (P = 0.70). The mean difference (mmol) with 95 % limits of agreement (Bland & Altman, 1986) between the calculated and measured ethanol content was -0.6 (2 to -3) on trial 2E, -1.1 (4 to -6) on trial 6E and -0.4 (2 to -3) on trial EG. These differences are within the coefficient of variance for the ethanol assay method (1.1 %). These data indicate that there is minimal absorption of ethanol across the gastric mucosa from dilute alcohol solutions in fasting humans.

Beckers, E.J., Reher, N.J., Brouns, F., Ten Hoor, F. & Saris, W.H.M. (1988). Gut 29, 1725-1729.

Bland, J.M. & Altman, D.G. (1986). Lancet i, 307-310.