Acetaldehyde, the first metabolite of ethanol oxidation, has been proposed as a major initiating factor in ethanol-induced liver injury. The aims of this study were to examine whether acetaldehyde is absorbable from the digestive tract and whether, when delivered chronically in drinking water, it is capable of inducing liver injury in rats. Acetaldehyde concentrations in the rat portal and peripheral blood were measured by head space gas chromatography after intragastric (5 ml) and intracolonic (3 ml) administration of 20 mM acetaldehyde solution. In the hepatotoxicity study, rats were exposed to acetaldehyde (20 and 120 mM) delivered in drinking water for 11 weeks and histopathological changes in the liver were morphometrically assessed. Peak blood acetaldehyde levels were found a t 5 min after acetaldehyde infusion and were 235 & 11 pM (mean * SE) after intragastric and 344 i~ 83 p M after intracolonic infusion of 20 mM acetaldehyde solution. The exposure of rats to 120 mM acetaldehyde solution for 11 weeks resulted in the development of fatty liver and inflammatory changes. Morphometric analysis showed significantly more fat accumulation in rats receiving 120 mM acetaldehyde solution (85 & 2 per cent of hepatocytes occupied by fat) than in rats receiving 20 mM acetaldehyde solution (38 * 11 per cent) or in controls (36 k 10 per cent). The dose of extrahepatic acetaldehyde (500 mgkg per day) producing liver injury corresponds to only around 3 per cent of that derived from hepatic ethanol oxidation in animals receiving an ethanol-containing totally liquid diet (15 glkg per day). These results indicate that acetaldehyde delivered via the digestive tract can reach the liver by the portal circulation and that acetaldehyde of extrahepatic origin appears to be more hepatotoxic than acetaldehyde formed during ethanol oxidation within the liver.