AIM: To determine the feasibility and safety of establishing a porcine hepatic cirrhosis and portal hypertension model by hepatic arterial perfusion with 80% alcohol.

METHODS: Twenty-one healthy Guizhou miniature pigs were randomly divided into three experimental groups and three control groups. The pigs in the three experimental groups were subjected to hepatic arterial perfusion with 7, 12 and 17 mL of 80% alcohol, respectively, while those in the three control groups underwent hepatic arterial perfusion with 7, 12 and 17 mL of saline, respectively. Hepatic arteriography and direct portal phlebography were performed on all animals before and after perfusion, and the portal venous pressure and diameter were measured before perfusion, immediately after perfusion, and at 2, 4 and 6 wk after perfusion. The following procedures were performed at different time points: routine blood sampling, blood biochemistry, blood coagulation and blood ammonia tests before surgery, and at 2, 4 and 6 wk after surgery; hepatic biopsy before surgery, within 6 h after surgery, and at 1, 2, 3, 4 and 5 wk after surgery; abdominal enhanced computed tomography examination before surgery and at 6 wk after surgery; autopsy and multi-point sampling of various liver lobes for histological examination at 6 wk after surgery.

RESULTS: In experimental group 1, different degrees of hepatic fibrosis were observed, and one pig developed hepatic cirrhosis. In experimental group 2, there were cases of hepatic cirrhosis, different degrees of increased portal venous pressure, and intrahepatic portal venous bypass, but neither extrahepatic portal-systemic bypass circulation nor death occurred. In experimental group 3, two animals died and three animals developed hepatic cirrhosis, and different degrees of increased portal venous pressure and intrahepatic portal venous bypass were also observed, but there was no extrahepatic portal-systemic bypass circulation.

CONCLUSION: It is feasible to establish an animal model of hepatic cirrhosis and portal hypertension by hepatic arterial perfusion with 80% alcohol, however, the safety of this model depends on a suitable perfusion dose.

Hepatic cirrhosis is most commonly caused by viral hepatitis and alcohol abuse[1]; the 5- and 10-year survival rates of patients with compensated hepatic cirrhosis are 84% and 68%, respectively, while the 5-year survival rate of decompensated hepatic cirrhosis is only 14%[2]. According to the World Health Organization (WHO), about 800000 patients die of hepatic cirrhosis each year worldwide[3]. Thus, in order to develop new strategies for the diagnosis and treatment of hepatic cirrhosis, it is essential, despite being a major challenge, to establish large animal models similar to humans, which are applicable in all-round research, especially models of hepatic cirrhosis for application in surgery, intervention and pharmaceutical research. Currently, the following methods are widely used to establish animal models of hepatic cirrhosis or portal hypertension: carbon tetrachloride (CCl₄); thioacetamide; surgical prehepatic portal ligation; and combinations of these methods. However, no approach is capable of fully simulating the entire pathophysiological process most frequently involved in the occurrence and development of posthepatitic and alcoholic cirrhosis[4]. In the present study, we successfully established a porcine hepatic cirrhosis and portal hypertension model through hepatic arterial perfusion with alcohol, which accurately reproduced the pathophysiological development process similar to that occurring in humans. The study results are reported below.

Many methods have been used to establish a hepatic cirrhosis model, such as CCl₄[5,6], thioacetamide[7], chemical carcinogen induction[8], alcohol[9], biliary obstruction[6,10,11], dystrophy induction[12], schistosome induction[13], and immunological mediation[14]. These modeling methods are mainly applicable to the following administration modes: oral administration, inhalation, subcutaneous injection, and intraperitoneal injection. They all require drug absorption followed by entry into various body systems, which tends to damage not only the liver, but also other tissues and organs, and more importantly, may even cause animal death. The dead animals may further result in environmental pollution. In addition, rats or mice are used as experimental animals in most cases, which have a small blood volume and therefore cannot undergo several serological tests and imaging examinations. Rats or mice have a small body size and thus relatively small blood vessels, and when used as experimental animals, it is difficult to determine the changes in the portal vein. In addition, surgical procedures are complicated on such small animals, along with intervention, mini-invasive surgery, and biopsy for collecting pathological tissues. It is also not possible to observe pathophysiological changes in the liver during the modeling process. More importantly, there is a large difference in the anatomy, normal physiological and biochemical values, and in the development process of disease between rat or mouse models and humans[15]. Some researchers have also used large animals such as dogs and baboons to establish specific hepatic cirrhosis or portal hypertension models[16,17], but these models are also significantly different from humans in terms of the developmental process of common posthepatitic or alcoholic cirrhosis and portal hypertension caused by hepatic cirrhosis. In this study, using a hepatic cirrhosis model established by perfusing porcine hepatic arteries with a high concentration of alcohol (80%), we were able to perform serological and imaging examinations, evaluate changes in the portal vein at different time points, and observe physiological changes in the pigs as well. In particular, we were able to investigate the pathological developmental process of hepatic cirrhosis by means of hepatic biopsy at different sites and at different time periods.

We used 80% alcohol for porcine hepatic arterial perfusion to establish a hepatic cirrhosis and portal hypertension model, and we observed and compared liver characteristics according to different perfusion doses. Our findings showed that hepatic fibrosis and cirrhosis occurred rapidly; hepatic fibrosis was observed as early as 1 wk after perfusion, and cirrhosis was found in some animals from experimental groups 2 and 3 as early as 2-3 wk. According to the literature, the occurrence of cirrhosis is usually reported 8 wk after perfusion[9,18]. In addition, our study demonstrated only a minor influence on liver function as a result of alcohol perfusion, and no injury to other important organs and tissues was observed during the development of hepatic cirrhosis, while previous reports have shown high hepatotoxicity[4]. The pathophysiological characteristics of hepatic cirrhosis and portal hypertension in our animal model were similar to those of early human posthepatitic and alcoholic cirrhosis, and were mainly demonstrated by the following results: (1) the serological examination was nearly normal; (2) the imaging examination showed an increase in liver size and spleen size, the formation of bypass circulation in the liver, and a widened main portal venous diameter; (3) some of the animals had increased portal venous pressure; and (4) hepatic biopsies at different sites and at different time periods confirmed the typical pathological development of hepatic cirrhosis. There was extensive hepatocyte generation and necrosis, as well as collapse of the liver lobule fibrous framework. The residual hepatocytes regenerated to form irregular, nodular hepatocyte masses (regenerative nodules). A large number of fibrous connective tissues proliferated to form fiber bundles and fibrous septa. The newly-formed fibrous septa connected with each other to enclose regenerative nodules or re-divide the residual liver lobules and remodel them into pseudo-lobules. The hepatic vascular bed was reduced, occluded, or distorted. The blood vessels were compressed by regenerative nodules. The normal relationship between intrahepatic portal veins, hepatic veins and small hepatic arterial branches was no longer present and communicating anastomotic branches were formed. Relatively uniform cirrhotic particles were observed on the liver surface, and the cirrhosis affected each liver lobe. Apart from the observation of splenic congestion caused by hepatic cirrhosis and mild varicose veins on the stomach surface of a few animals, other important organs and tissues were unaffected.

In experimental group 1, there was a long duration and a low success rate of modeling by perfusion with 80% alcohol, but the animals all survived. In experimental group 2, there was a moderate duration and a high success rate of modeling, and the animals all survived. In experimental group 3, there was a short duration and a high success rate of modeling, but the rate of animal death was significantly increased.

Establishing a large animal model of hepatic cirrhosis via a mini-invasive intervention is still in the exploratory phase. One study[19] has shown that when a porcine hepatic cirrhosis model is established by injecting an embolic agent containing 3:1 iodinated oil-alcohol into porcine hepatic arteries, the complex surgical procedure is challenging. Specifically, a metal spring loop is used to embolize non-target arteries, a micro-catheter is used to inject the embolic agent, and the portal venous pressure is measured indirectly by the hepatic venous pressure gradient (HVPG). As a result, total costs are markedly increased, the operation time is prolonged, tissue ischemia and necrosis within the embolization range of non-target blood vessels may occur, and large hepatic arteries may be embolized during injection of the embolic agent due to the properties of iodinated oil. In this experiment, 80% alcohol was slowly perfused into the left and right common hepatic arteries (away from the cystic and gastroduodenal arteries) via a 4F catheter, and then uniformly distributed into the liver by blood flow, based on the size and blood supply of the left and right hepatic arteries; and 1 mL 1% lidocaine was immediately injected after injection of 1 mL 80% alcohol until the dose of 80% alcohol was completely injected, to prevent spasm of large hepatic arteries. Only < level 3 hepatic arteries were embolized to preserve maximum blood and oxygen supply. After percutaneous hepatic portal vein puncture was successfully performed in all animals, a 6F multi-sidehole pigtail catheter was applied and lodged into the portal vein for phlebography and direct measurement of portal venous pressure and diameter. Therefore, this modeling method has many advantages, such as a simple surgical procedure, low costs, and the possibility of direct observation of changes in the portal vein and in portal venous pressure, as well as bypass branches and their distribution. Portal phlebography immediately after hepatic arterial perfusion with alcohol confirmed that alcohol passes through small arteries and enters into the hepatic sinus and subsequently small portal venous branches to produce combined hepatic arterial-portal venous embolization. Pavcinik et al[20] injected polyvinyl alcohol particles into porcine portal veins by percutaneous hepatic puncture to embolize portal venous branches to produce portal hypertension, however, 1 wk after the experiment they found that the portal venous pressure had decreased to the normal level, and hepatic cirrhosis did not develop. In a study of opposite hepatic decompensated hypertrophy after hepatic arterial embolization and portal venous embolization, Madoff et al[21] observed hepatocyte and endothelial injury, along with erythrocyte stagnation after injecting an embolic agent containing 3:1 iodinated oil-alcohol into porcine hepatic arteries, and 28 d after injecting polyvinyl alcohol particles into the portal veins and linearly embolizing the left and left-middle liver lobes, the histopathology showed atrophy of the portal vein region and interlobular septa in the embolized segments, without significant fibrosis. In the present study, hepatic biopsies were taken at different time points and pathological examination confirmed that hepatic arterial perfusion with different doses of 80% alcohol caused thrombosis in miniature and small hepatic arteries, portal venous branches and veins, as well as the hepatic sinus, and endothelial injury. This was followed by hepatocyte degeneration and necrosis, as well as inflammatory cell infiltration, which eventually developed into the typical pathophysiological process of hepatic cirrhosis. As reported in the literature, at 6-8 wk after surgery, hepatic histopathology showed fibrous septa and pseudo-lobules, but localized cirrhosis, which was not uniformly distributed, and fibrosis at different levels; and portal venous pressure increased in some animals, as measured indirectly by HVPG[18]. At 2 and 6 wk after perfusion, the local hepatic biopsy pathology and gross pathology in all animals in experimental group 2, and in the surviving animals in experimental group 3, confirmed that a hepatic cirrhosis model was successfully established, with the majority of fibrosis reaching level 4. Samples collected from each site were observed to have cirrhosis, and relatively uniform cirrhotic nodules were found on the surface of various liver lobes. These results suggest that the injuring effect of alcohol on the liver is uniform. In addition, the physiological changes in early hepatic cirrhosis and portal hypertension were also directly confirmed by portal phlebography and portal venous pressure measurement, such as the occlusion of portal venous branches, increased portal venous pressure, widening of portal venous diameters, and the formation of intrahepatic portal venous bypasses. A total of eight animals developed hepatic cirrhosis, including all animals in experimental group 2 and the surviving animals in experimental group 3, of which five animals had concomitant portal hypertension. Among the pathological samples collected at a total of 32 sites in the left and right lateral and medial liver lobes of each animal, only the samples at two sites were observed to have level 3 fibrosis and those at the remaining 30 sites had level 4 fibrosis. This also indicated that the injuring effect of alcohol on the liver was uniform. However, one study[19] using an embolic agent containing 3:1 iodinated oil-alcohol, reported that only 5 of 20 sites in the 16 mL group, and 4 of 20 sites in the 28 mL group had level 4 fibrosis, which suggests that the injuring effect on the liver is not uniform[19].

In our study, the deaths in experimental group 3 were mainly attributed to liver failure caused by extensive hepatic tissue necrosis and to postoperative infection.

In conclusion, at 2-6 wk after hepatic arterial perfusion with 80% alcohol, a porcine hepatic cirrhosis or cirrhotic portal hypertension model can be successfully established, with the typical pathophysiological development process of early hepatic cirrhosis, and with mild or no injury to other organs and tissues. The suitable dose of 80% alcohol for hepatic arterial perfusion in this experiment was the 12 mL dose used in experimental group 2, but the most suitable alcohol concentration or dose for establishing hepatic cirrhosis or cirrhotic portal hypertension models still needs further in-depth research and investigation.