Occurrence of liver gas embolism after rapid decompression was assessed in 31 female rats that were decompressed in 12 min after 42 min of compression at 7 ATA ( protocol A). Sixteen rats died after decompression ( group I). Of the surviving rats, seven were killed at 3 h ( group II), and eight at 24 h ( group III). In group I, bubbles were visible in the right heart, aortic arch, liver, and mesenteric veins and on the intestinal surface. Histology showed perilobular microcavities in sinusoids, interstitial spaces, and hepatocytes. In g roup II, liver gas was visible in two rats. Perilobular vacuolization and significant plasma aminotransferase increase were present. In g roup III, liver edema was evident at gross examination in all cases. Histology showed perilobular cell swelling, vacuolization, or hydropic degeneration. Compared with basal, enzymatic markers of liver damage increased significantly. An additional 14 rats were decompressed twice ( protocol B). Overall mortality was 93%. In addition to diffuse hydropic degeneration, centrilobular necrosis was frequently observed after the second decompression. Additionally, 10 rats were exposed to three decompression sessions ( protocol C) with doubled decompression time. Their mortality rate decreased to 20%, but enzymatic markers still increased in surviving rats compared with predecompression, and perilobular cell swelling and vacuolization were present in five rats. Study challenges were 1) liver is not part of the pathophysiology of decompression in the existing paradigm, and 2) although significant cellular necrosis was observed in few animals, zonal or diffuse hepatocellular damage associated with liver dysfunction was frequently demonstrated. Liver participation in human decompression sickness should be looked for and clinically evaluated.