variceal hemorrhage; however, these medications may also impair cardiac function and increase pulmonary vascular resistance in PoPH, and as such are generally avoided in this patient population [4, 5, 64–66]. The same concerns apply to the use of calcium channel blocker medications, and consequently patients with PoPH do not typically undergo vasodilator testing during diagnostic RHC as these medications are generally contraindicated. Trans-jugular intrahepatic portosystemic shunting (TIPS) procedures have been shown to relieve portal pressure, decrease the risk of variceal bleeding in patients with advanced portal hypertension and cirrhosis, and decrease mortality in these patients [67, 68]. In PoPH, however, TIPS procedures may result in a marked increase in cardiac output, leading to right ventricular overload and decompensated pulmonary vascular disease, and the decision to proceed with TIPS should be approached with extreme caution in these patients [5, 11, 69]. Although the exact relationship between HPS and PoPH is unclear, administration of targeted pulmonary vasodilator therapy in PoPH has unmasked or exacerbated underlying HPS, both yielding clues regarding the fascinating relationship between these two pulmonary vascular complications of chronic liver disease and offering a cautionary note when utilizing targeted pulmonary vasodilator medication in PoPH [70, 71].

Hepatopulmonary Syndrome (HPS)

Epidemiology and Risk Factors

Hepatopulmonary syndrome is defned by the triad of arterial deoxygenation and intrapulmonary vascular dilation that occurs in the setting of chronic liver disease [5, 72–74]. Arterial deoxygenation is typically confrmed by a resting arterial blood gas analysis while breathing ambient air that demonstrates an alveolar-arterial oxygen gradient of ≥15 mmHg (or, for those over age 65, greater than the expected age-adjusted values) or a partial pressure of oxygen (PaO2) of <80 mmHg (or, in those over age 65, less than the expected age-adjusted values). Intrapulmonary vascular dilation manifests as intrapulmonary shunting, which is typically identifed using contrast TTE imaging.

HPS is typically diagnosed in the sixth decade of life. It is estimated that 1–4% of all chronic liver disease patients have HPS. Among those undergoing evaluation for liver transplantation, between 10% and 47% of patients have some degree of HPS, depending on which diagnostic criteria are used (absolute values or age-adjusted values of arterial blood gas analysis). There is no association between the presence of HPS and either the etiology or severity of chronic liver disease, or with age, gender, or ethnicity. The presence of HPS has been associated with a markedly higher risk of mor-

tality in cirrhotic patients, even after adjustment for severity of liver disease (MELD score) [73, 74].

Molecular Pathogenesis

Although the mechanism(s) of disease pathogenesis in HPS is incompletely understood, it is believed to involve vasoactive mediators that bypass hepatic metabolization and directly affect the pulmonary vasculature (Fig. 11.5). A combination of bacterial translocation into systemic circulation, hepatic endothelin-1 (ET-1) release, and resulting increases pulmonary nitric oxide synthesis and macrophage activation are believed to play key roles [75–82]. Animal models of liver cirrhosis generated by common bile duct ligation have demonstrated increased nitric oxide synthesis in the lungs, increased enteric bacterial translocation into the systemic circulation, and elevated macrophage recruitment into the pulmonary circulation. Additionally, antibiotic treatment or macrophage depletion both serve to reduce intrapulmonary vascular dilations and vascular remodeling characteristic of HPS, and animals with HPS were more likely to have evidence of bacterial invasion of lymph nodes as compared to non-HPS animals, further supporting a potential microbial mechanism for the disease [75–77, 79]. Evidence from common bile duct ligation animal models of liver cirrhosis suggests that ET-1 and pulmonary nitric oxide likely contribute to intrapulmonary vascular dilations and HPS, in that blockage of the endothelin B receptor diminished macrophage accumulation, pulmonary nitric oxide synthesis, and vascular abnormalities [80–82]. The potential role that ET-1 and activated pulmonary macrophages play in both PoPH and HPS pathogenesis also suggest shared mechanisms, however to date the evidence linking these two vascular complications of chronic liver disease remains rudimentary.

The study of the bone morphogenic protein axis appears to be a promising new direction in HPS research [20, 83]. A multicenter case-control study of in cirrhotic liver transplant patient genetics identifed a number of gene polymorphisms associated with the bone morphogenic protein system that were signifcantly associated with HPS, including those involved in regulating angiogenesis (COL18A1), levels of specifc bone morphogenic proteins (endoglin, the circulating ligand trap for BMP-9), and the vascular remodeling response induced by hypoxemic conditions (NOX4 and RUNX1). Building upon this, an investigation of the pulmonary vascular complications of liver disease study, a multicenter prospective cohort study of 454 adult patients with portal hypertension undergoing evaluation for liver transplant between 2013 and 2017, identifed signifcantly lower BMP-9 and bone morphogenic protein 10 levels in HPS patients compared to non-HPS control cirrhotic patients, and

Fig. 11.5  Working hypothesis of molecular pathogenesis of HPS. Release of Endothelin-1 (ET-1) from activated hepatic endothelial cells (a), combined with enteric bacterial translocation into systemic circulation (b), gene polymorphisms in Endoglin, COL18A1, NOX4,

RUNX1 (c), and low levels of circulating BMP-9 and BMP-10, result in macrophage recruitment, intrapulmonary vascular dilation (d), increased pulmonary nitric oxide synthesis, and vascular remodeling characteristic of HPS (e)

an association between BMP-9 levels and widened alveolar-­ arterial gradients and worse functional capacity in HPS patients. Taken together, these data not only implicate the bone morphogenic protein system in driving the vascular malformations characteristic of HPS but also offer another intriguing link between PoPH and HPS disease pathogenesis. Given their ability to resolve gene expression profles at the cellular level, single-cell RNA-sequencing techniques may shed new light on the complex interactions between PoPH and HPS [24].

Screening andDiagnosis

As HPS is an indication for LT, screening for the disorder is recommended in all chronic liver disease patients undergoing transplant evaluation (Fig. 11.1). Clinical manifestations re ect the underlying physiology of chronic hypoxemia and

liver disease, and include dyspnea, cyanosis, and digital clubbing, as well as the ascites, peripheral edema, jaundice, spider angiomas, and palmar erythema seen in chronic liver disease [5, 74, 84] (Fig. 11.6). Platypnea, dyspnea in the upright position that is relieved by lying supine, and orthodeoxia, hypoxemia in the upright position that is corrected while lying supine, are commonly associated with HPS. None of these fndings are pathognomonic for HPS, however, and the most common presenting symptom is dyspnea. Chest radiography is typically unremarkable, and pulmonary function testing can occasionally identify an impaired diffusion capacity for carbon monoxide, but neither are routinely useful in the screening and diagnosis of HPS.

The mainstay of diagnosis in HPS is twofold: confrmation of arterial deoxygenation and identifcation of intrapulmonary vascular dilation. Arterial deoxygenation can be suggested by pulse oximetry screening, and various cutoffs have demonstrated differing trade-offs between sensitivity

and positive predictive value, but an upright at-rest arterial blood gas is still required for confrmation [85–89]. To enhance the screening value of pulse oximetry, the test can be performed both supine and upright to evaluate for underlying orthodeoxia; however, this screening approach has yet to be fully validated. Arterial deoxygenation requires an elevated alveolar-arterial oxygen gradient of ≥15 mmHg (or, in those over age 65, greater than the expected age-adjusted values) or a PaO2 of <80 mmHg (or, in those over age 65, less than the expected age-adjusted values).

Contrast-enhanced TTE imaging is performed to identify the presence of intrapulmonary vascular dilations in HPS and is recommended as part of the routine evaluation for all liver transplant candidates [3–5]. The late appearance of bubbles in the left atrium, typically after three cardiac cycles, is consistent with the presence of an intrapulmonary shunt and is suffcient to fulfll the diagnostic criteria for HPS [90, 91] (Fig. 11.7). Parenteral macro-aggregated albumin scanning using radiolabeled albumin, as is routinely performed for ventilation perfusions scanning, can also identify intra-

pulmonary shunts, as well as quantify the severity of shunt. A cutoff of >6% shunt, calculated as the percent of extrapulmonary shunt to the cerebrum, is typically used as a diagnostic threshold in HPS (Fig. 11.8) [92–94]. Rarely patients with HPS demonstrate large intrapulmonary shunts, sometimes referred to as Type-II HPS, which may be amenable to therapeutic embolization, and the use of macro-aggregated albumin scanning and pulmonary angiography can help identify this subset of patients. Neither imaging test is routinely performed in the evaluation of HPS, however, and the precise association between shunt severity and outcomes in HPS remains unclear.

HPS Treatment

HPS is defned by arterial hypoxemia, and supplemental oxygen remains the mainstay of therapy. As LT can be curative, debilitating HPS is an indication for an expedited liver transplant evaluation, and the diagnosis of HPS provides MELD exception points during LT listing [3, 4]. Following LT, HPS is expected to resolve in almost all patients, with 85% of patients cured within 12 months [95–97]. HPS patients experience roughly double the mortality rate of non-­ HPS cirrhotic patients while awaiting liver transplantation, but survival following LT is comparable in both groups [5, 74, 98, 99]. Predictors of outcomes in HPS before and after

LT include both severity of liver disease and degree of hypoxemia, as measured by either alveolar-arterial oxygen gradient or PaO2 on arterial blood gas [3, 4, 37, 100, 101].

A number of additional therapeutics have been studied in HPS as adjuncts to LT and supplemental oxygen [102–108]. Given the defning feature of intrapulmonary vascular dilation, somatostatin analogs (octreotide) have been evaluated in HPS but have not demonstrated any benefcial effect on arterial oxygenation or degree of shunting and may actually worsen pulmonary vascular hemodynamics in these patients [102, 103]. Inhibition of nitric oxide via L-NG-Nitro arginine methyl ester has also been studied in HPS and was successful in decreasing nitric oxide levels. Unfortunately, this therapy also decreased cardiac function and increased pulmonary vascular resistance, dampening enthusiasm for this interventional strategy [104]. Treatment with allium sativum (garlic) supplementation did demonstrate improvements in both symptoms of dyspnea and arterial hypoxemia in a small pilot study of HPS patients but has not been validated in larger studies [105]. In an attempt to inhibit angiogenesis, the tyrosine kinase inhibitor sorafenib was studied in HPS patients in a randomized clinical trial, and although it did signifcantly reduce circulating vascular endothelial growth factor levels, it did not affect the degree of shunting on contrast echocardiography or alveolar-arterial oxygen gradient and signifcantly worsened quality of life in HPS patients [106]. TIPS procedures have been effective in improving arterial oxygen-