The regression of cirrhosis is associated with increased intrahepatic collagenolytic enzyme activity. We investigated whether collagenase supplementation via portal vein infusion can retard cirrhosis development and/or reverse cirrhosis. In all, 35 rabbits were initially assigned to study. However, because of high surgical mortality and infection, only 15 animals completed study. Four normal controls (group I) received olive oil subcutaneously (SC) for 12 weeks followed by normal saline portal perfusion for 12 weeks. Four (group II) received CCl(4) SC for 6 weeks followed by portal vein collagenase, 6 mg twice weekly, plus SC CCl(4) for 6 additional weeks and then killed. Four rabbits (group III) received CCl(4) SC for 12 weeks and then 6 mg of collagenase portally for 12 weeks, while three control rabbits (group IV) received CCl(4) for 12 weeks followed by saline for 12 weeks. After 12 weeks of CCl(4), liver hydroxyproline content of collagenase-treated group II (361.1+/-106.6 microg/g) was significantly reduced compared with group III+IV that had not yet received collagenase (589.0+/-162.9 microg/g; P<0.05). In the main comparison, hydroxyproline content of collagenase-treated group III (177.5+/-35.6 microg/g) was significantly decreased compared with saline-treated controls (446.3+/-150.1 microg/g; P<0.01). Further, liver histology showed complete regression of cirrhosis in the collagenase-treated animals. No toxicity of liver, kidney, lung, brain or heart was observed histologically. Anaphylaxis occurred in 2/35 original animals (one fatal). In conclusion, this study provides 'proof of principle' that collagenase portal administration can retard cirrhosis development and speed regression of established cirrhosis in the rabbit CCl(4) model. Potential application to humans is premature, but feasible, if these findings are confirmed in additional animal studies.