The first report about HPS dates back to 1884 when Fluckiger ¹ described a 37-year-old woman with no cardio-respiratory dysfunction, but cyanosis and digital clubbing occurred after several years cirrhosis caused by syphilis, suggesting hypoxemia occurs in hepatic cirrhosis.In 1895, 1901 and 1937, Gilbert ², Hijmans van den Bergh ³ and Evans ⁴ respectively further confirmed that cyanosis and finger clubbing did occur in hepatic cirrhosis. However, the first literature reported hypoxemia in cirrhosis patients was in 1935 by Snell ⁵ who described three patients had liver disease with hemoglobin desaturation. In the fowing 30 years a large number of studies ^{4, 6-20} found the association of hepatic cirrhosis with arterial oxygen desaturation.However, the term “hepatopulmonary syndrome” was put forward until 1977 by Kennedy and Knudson ²¹. They, using an illustrative case, described a syndrome, which is characterized by alcoholic cirrhosis, hypoxemia and hyperdynamic circulation but is normal in the indices of pulmonary air flow, volume, and distribution of ventilation. Moreover, exercise, orthodeoxia and hypocapnia would aggravate the hypoxemia. The features of this syndrome analogous to the hepatorenal syndrome, which is defned as renal vasoconstriction, reduced renal plasma fow and the rate of glomerular fltration, cortical oligemia, and progressive renal failure occurring in histologically normal kidneys. Sometimes, such kind of kidneys can be served as donor organs with normal function and the azotemia can be improved after improvement in hepatic function ²².These results indicating some extrinsic factors stemming from hepatic failure are responsible for renal dysfunction, rather than intrinsic renal pathologic changes. These extrinsic factors may affect pulmonary circulation in HPS, which is hinted by the report of hepatic function improvement reduced intrapulmonary shunting. The word “hepatopulmonary syndrome”, which describes the relationship between hypoxemia and pulmonary vascular abnormalities caused by hepatic dysfunction, was offcially named in the eighth edition (1989) of Disorders of the Liver and Biliary System by Dame Sheila Sherlock ²³.

Why HPS caused so much attention among clinical physicians. Initially, severe hypoxemia(PaO ₂＜50 mmHg) on room air in patients with HPS was considered to be an “absolute”contraindication to liver transplantation in 1984 ²⁴. However, nowadays such hypoxemia is now considered to be a “relative” contraindication to transplantation with the success in liver transplantation success ^{25, 26}. So how to recognize HPS? Are there any medical strategies to improve HPS? Is hypoxemia associated with HPS a contraindication to liver transplantation?Does hypoxemia resolve following successful transplantation? These questions drove physicians to research HPS. In 1999, Mokhashi ²⁷ summarized there are several compelling reasons why internists need to be aware of the HPS. First, the relatively high prevalence of the syndrome: 47%of patients investigated with end stage liver disease and 38% with less severe liver disease, were shown to have intrapulmonary shunting. Second, commonly encountered liver diseases, such as alcoholic cirrhosis and chronic active hepatitis, can present with HPS. Third, clinical features of HPS (dyspnoea and clubbing) are non-specifc. Although the majority presents with hepatic manifestations, up to 18% of patients with HPS have presented with predominantly pulmonary symptoms. Fourth, the oxygenation can be aggressively declined despite the hepatic function is table and according to a report 41% mortality occurred in a mean of 2.5 years after the onset of dyspnea, so failure to recognize HPS is serious. Fifth, clinical diagnosis of HPS at early stage facilitates the treatment decision with either selective embolization of the dilated pulmonary vasculature or liver transplantation. Last, diagnosis of HPS is easy and requires no fancy equipment. In addition, Molleston et al ²⁸ in 1999 reported brain abscess developed one year after HPS was diagnosed in a teenager with well-compensated cirrhosis induced by biliary atresia.They proposed brain abscess maybe resulted from intrapulmonary shunting of contaminated venous blood. Moreover, Alonso et al ²⁹ found the coexistence of HPS worsened the prognosis in liver cirrhosis.

The pathophysiological alteration of HPS underwent a long-term research. At first, Keys and Snell in 1938 ⁶ suggested that HPS is associated with abnormal oxygen tension-saturation relations in cirrhotic patients, but Rodman and Heinemann's results ^11,12 did not consistent with this hypothesis. Afterwards, Caldwell, Fritts and Cournand ²⁰ focused on abnormal oxygen tensionsaturation relations again but concluded that the contribution of these toward hypoxemia was probably small. Then, the possibility of anastomoses between the pulmonary arteries and veins has been stressed ^8-19. Although Mellemgaard and his colleagues ³⁰ in 1963 found intrapulmonary shunting of blood in portal hypertension patients using radioactive tritium, Fritts et al ³¹ in 1960 using krypton gas did not demonstrate this. Functioning portapulmonary anastomoses have been demonstrated ^{30, 32, 33}, but in theory, such shunts to produce marked arterial oxygen desaturation is impossible thanks to the two reasons, one is the high oxygen content of portal venous blood and another is that it is probably small for the blood fow through these pathways in life ³⁴. These results suggest that there are some other pulmonary alterations, which contribute to hypoxemia induced by liver diseases.

To explore any further anatomic basis for the hypoxemia, Berthelot and his co-workers ³⁵ in 1966 studied the lung vasculature alterations in cirrhosis patients. In all 13 cases of cirrhosis patients without clinical evidence of lung diseases, they found the microvessels in the fine peripheral branches of the respiratory part of the lungs in all cases were markedly dilatated, and in 6 cases there were apparent “spider nevi” on the pleura. This dilatation mainly located at the precapillary levels and affected arterial branches which were below 500 μm in diameters. In fact, Rydell and Hoffbauer in 1956 had described the dilated pulmonary vessels in a 17-year-old boy who presented with dyspnea after 3 years cirrhosis. This study also found only 1 patient had precapillary arteriovenous anastomoses, which further demonstrated intrapulmonary shunting of blood was not the main cause of arterial oxygen unsaturation. Since then, a large number of literatures demonstrated IPVD was the pathophysiological alteration of HPS by using different kinds of tools, such as contrast enhanced echocardiography, perfusion lung scanning, pulmonary angiography and high-resolution computerized tomography (HRCT) scanning, as well as autopsy results. However, the pathogenesis of this vasodilatation was undefned. The majority of patients with HPS had portal hypertension, but it was unlikely to be a simply mechanical result since the pulmonary veins were not dilated and connective-tissue septums had no edema. Thus, a humoral mechanism was possible. Nowadays, many studies have demonstrated abnormal vasodilators play a key role in IPVD.

Several lines of proofs suggest that angiogenesis plays an important role in the pathophysiology of HPS in addition to IPVD. In 1966, autopsy studies to 13 case of cirrhosis patients by Berthelot and his colleagues ³⁵ not only found a marked pulmonary arterial dilatation but also an increased pulmonary capillary density, but at that time the increased vascular density did not show any correlation with hypoxemia. In 1997, Schraufnagel et al ³⁶ found alveolar capillary density was increased in experimental HPS, suggesting angiogenesis may participate in the development of HPS. Recent studies ^37-40 further confirmed angiogenesis and elevated angiogenic factors in experimental and human HPS. However, the roles of angiogenesis in HPS need further demonstration.

PoPH is also an important alteration companied with HPS. In 1999, Mal et al ⁴¹ frst reported a patient with liver cirrhosis who successively developed HPS and PoPH. Initially, the patient presented with severe dyspnea and hypoxemia at rest. HPS was confirmed by right-to-left shunting using perfusion lung scanning and a hyperdynamic state with low pulmonary vascular resistance using hemodynamic study. More than 2 years later, severe pulmonary hypertension was demonstrated by right-sided heart catheterization, which suggested PoPH might occur following HPS in cirrhosis patients. At the same year, a case illustrated the occurrence of HPS and PoPH in the same patient ⁴². Then, a series of studies reported HPS and PoPH occurred in the patients with cirrhosis or portal hypertension. However, the morbidity of PoPH (1%) is lower than that of HPS(4%～10%). Taken together, these results suggest that hypoxemia in HPS may be the incentive to PoPH in patients with cirrhosis or portal hypertension.

A well representative and easily available animal model that can mimic human disease is very important for exploring pathogenic alterations and mechanisms of the disease. Common bile duct ligation (CBDL) rat is the only, until now, accepted animal model of HPS ⁴³. According to the literature, it was frstly proposed in 1992 by Shih-wen Chang and Narumi Ohara. They found that 5 to 6 wk CBDL rats exhibited impaired gas exchange associated with a reversible depression of hypoxic pulmonary vasoconstriction, which was closely parallel with human HPS ⁴⁴. Then in 1997,Michael B. Fallon et al reported that CBDL rats could be served as a HPS model. They found that IPVD was occurred in 2 and 5 wk CBDL rats by using microspheres compared with 3 wk partial portal vein ligation (PVL) rats ⁴⁵. Hence, CBDL rat has been recognized as a HPS model to study the mechanisms of human HPS. Due to the high mortality of this model, Lu et al ⁴⁶ improved this model by ligating the bile duct at the junction of the hepatic ducts which signifcantly decreased rats mortality induced by CBDL. In 2007, Zhang et al ⁴⁷ described multiple pathogenic factors including high fat diet, alcohol, cholesterol and CCl ₄ induced cirrhosis can be a non-invasively induced HPS model. In 2014, Zhang et al ⁴⁸ reported intra-abdominal hypertension (IAP) induced by subcutaneous injection of carbon tetrachloride led to the decrease of arterial partial pressure,the increase of difference of alveolar arterial oxygen pressure, the accumulation of macrophages in the alveolar spaces and alveolar walls widened in mice, suggesting this as a potentially novel mouse model of HPS. However, the latest two models still need to be confrmed.