REVIEW URRENT C OPINION
Hepatopulmonary syndrome David G. Koch a and Michael B. Fallon b
Purpose of review To discuss the advances in the understanding of the pathophysiology of experimental and human hepatopulmonary syndrome (HPS) and in the management of HPS, particularly regarding liver transplantation. Recent findings Advances have been made in defining the pathophysiology of HPS in experimental models as well as in human disease, including the role of endothelin-1, pulmonary monocytes, and angiogenesis. Additionally, the implications of the presence of HPS as it relates to prioritizing patients for liver transplantation and posttransplant outcomes will also be reviewed. Summary Mechanisms of disease continue to be defined in HPS, providing potential targets for pharmacologic intervention. Outcomes after liver transplantation are also becoming clearer, including the management of HPS with severe hypoxemia. Keywords hepatopulmonary syndrome, hypoxemia, intrapulmonary vasodilatation, liver transplantation, portal hypertension
INTRODUCTION The hepatopulmonary syndrome (HPS) is a pulmonary complication that occurs in patients with chronic liver disease and/or portal hypertension, whereby hypoxemia results from alterations in the pulmonary microvasculature, including vasodilatation and angiogenesis. HPS is relatively common, occurring in up to 30% of patients with cirrhosis [1,2]. Although the degree of hypoxemia in HPS does not reliably correlate with the severity of liver disease, patients with HPS have increased mortality relative to cirrhosis patients without HPS. The diagnostic criteria for HPS include documentation of impaired oxygenation and intrapulmonary vasodilatation (IPVD) in the setting of chronic liver disease or portal hypertension [3] (see list below). Oxygenation in HPS is determined by calculating the alveolar–arterial oxygen gradient (AapO2) from an arterial blood gas (ABG); a threshold of greater than 15 mmHg is used except for patients older than age 64, in whom greater than 20 mmHg is appropriate. Therefore, patients with mild HPS may not have hypoxemia. The syndrome may occur in patients with comorbid primary lung diseases [4,5], and having cirrhosis is not imperative as HPS has been described in acute and chronic hepatitis without cirrhosis or portal hypertension www.co-gastroenterology.com
[6,7] as well as in noncirrhotic portal hypertension without chronic liver disease [8–11]. Much of our knowledge of the pathophysiology of HPS is based on a unique experimental model, common bile duct ligation (CBDL), which recreates the features of human HPS. Although progress has been made in defining the mechanisms underlying the pulmonary vascular changes in HPS, to date, there are no approved therapeutic options aside from liver transplantation. Diagnostic criteria for the HPS are as follows: (1) increased age-corrected alveolar–arterial oxygen gradient (>15 mmHg or >20 mmHg if age >64 years) while breathing ambient air: a
Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of South Carolina, Charleston, South Carolina and bDivision of Gastroenterology, Hepatology and Nutrition, Department of Internal Medicine, The University of Texas Medical School at Houston, Houston, Texas, USA Correspondence to David G. Koch, MD, MSCR, Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of South Carolina, 25 Courtenay Drive, ART 7100A, MSC 290, Charleston, SC 29425, USA. Tel: +1 843 792 6901; fax: +1 843 876 4301; e-mail: [email protected] Curr Opin Gastroenterol 2014, 30:260–264 DOI:10.1097/MOG.0000000000000067 Volume 30 Number 3 May 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Hepatopulmonary syndrome Koch and Fallon
AapO2 ¼ [FiO2(Patm –pH2O)–( pCO2/0.8)] – PaO2, in which PaO2 is the partial pressure of arterial oxygen; FiO2 is the fraction of inspired oxygen; Patm is the atmospheric pressure; pH2O is the partial pressure of water vapor at body temperature; and paCO2 is the partial pressure of arterial carbon dioxide (0.8 corresponds to the standard gas-exchange respiratory ratio at rest); (2) positive contrast-enhanced echocardiography or lung-perfusion scanning (brain shunt fraction >6%); (3) cirrhosis and/or portal hypertension.
PATHOPHYSIOLOGY
correlated with the degree of bile duct proliferation in the corresponding liver biopsy specimens [26 ]. However, whether proliferating cholangiocytes in human cirrhosis produce and secrete ET-1 thereby raising the blood levels is still not known. A causal relationship between ET-1 and impairment in oxygenation in HPS that would support pharmacologic manipulation of the ET-1 and ETB receptor axis has also not been established. In addition to the role of ET-1 in experimental and human HPS, recent studies in experimental HPS have furthered our understanding of other pathways that contribute to the underlying pulmonary vascular phenomena. These include the roles of pulmonary intravascular monocytes [27 ], TNF-a [28], and angiogenesis [27 ,29 ] as well as the therapeutic utility in manipulating these pathways in order to improve oxygenation. The ability of serum from CBDL rats with HPS to induce a proliferative effect on pulmonary artery smooth muscle cells (PASMCs) and pulmonary microvascular endothelial cells (PMVECs) is also being understood. With respect to intrapulmonary monocytes, it is known that these cells accumulate in the lung vasculature after CBDL and that they contribute to IPVD by increasing the levels of nitric oxide [via inducible nitric oxide synthase (iNOS)] and carbon monoxide [via heme oxygenase-1 (HO-1)]. More recently, it has been demonstrated that the adherence of monocytes is regulated through ET-1 and ETB receptor activation which increases the levels of the monocyte chemokine, fractalkine (FKN/CX3CL1–CX3CR1). FKN induces pulmonary monocyte adhesion and also promotes angiogenesis directly and through monocyte secretion of vascular endothelial growth factor-A (VEGF-A) [27 ]. VEGF-A is also produced in the CBDL liver and induces cholangiocyte proliferation through extracellular-signal-regulated kinase (ERK) activation [30 ]. Furthermore, inhibiting the effects of VEGF-A with the receptor tyrosine kinase inhibitor, sorafenib, improves oxygenation in experimental HPS [29 ,30 ] by reducing pulmonary monocyte adhesion and subsequent VEGF-A production as well as by blocking cholangiocyte proliferation in the liver through inhibition of ERK activation [30 ]. Whether sorafenib has similar effects on cholangiocytes in human HPS and may improve hypoxemia is not known. However, the potential to manipulate this pathway pharmacologically in patients with HPS is attractive as sorafenib is approved for use in humans with cirrhosis. At the cellular level, serum from CBDL rats induces the proliferation of both PMVECs and PASMCs, effects that could contribute to lung angiogenesis and remodeling in HPS. CBDL serum &
&&
The CBDL experimental model recreates the features of human HPS [12–20] (Fig. 1). In this model, proliferating cholangiocytes in the liver produce and secrete endothelin-1 (ET-1) [12–16], whereas other models of portal hypertension that do not result in bile duct proliferation do not develop HPS [17]. Increased shear stress in the pulmonary microvasculature, as a result of a hyperdynamic circulation of cirrhosis, results in upregulation of the endothelin B (ETB) receptor that subsequently binds the increased circulating ET-1 and augments pulmonary nitric oxide production via endothelial nitric oxide synthase (eNOS) [12,15,19,21–25]. The potential role of biliary ET-1 in human HPS was shown in a single study, in which hepatic venous blood levels of ET-1 were higher in cirrhosis patients with IPVD and HPS compared with those without IPVD. In addition, the hepatic venous ET-1 levels in these patients
&&
&&
&&
Liver Cholangiocyte ET-1
Monocyte iNOS
&&
HO-1
ETBR
VEGF-A
&&
CO p-AKT/p-ERK
p-eNOS/NO
Endothelial Cell Vasodilatation
Angiogenesis
HPS
FIGURE 1. Mechanisms of hypoxemia in the common bile duct ligation model of HPS. CO, carbon monoxide; ET-1, endothelin-1; ETBR, endothelin B receptor; HO-1, heme oxygenase-1; HPS, hepatopulmonary syndrome; iNOS, inducible nitric oxide synthase; NO, nitric oxide; p-eNOS, phosphorylated endothelial nitric oxide synthase; VEGF-A, vascular endothelial growth factor A.
0267-1379 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
&&
&&
www.co-gastroenterology.com
261
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Liver
increased the cytoskeleton proteins within PMVECs cells and promoted proliferation by decreasing the expression levels of Annexin A1, a Ca[2þ]-dependent phospholipid-binding protein [31]. Conversely, CBDL serum upregulates Annexin A2, promoting PASMC proliferation [32]. Either promoting Annexin A1 expression or inhibiting Annexin A2 was found to ameliorate hypoxemia in experimental HPS [31,32]. Finally, new insights into how bacterial translocation and endotoxemia drive inflammation in experimental HPS have been made. Evidence of endoplasmic reticulum stress as evidenced by increased lung levels of 78 kDa glucose-regulated protein are found after CBDL. This protein, in turn, has a strong, positive correlation with the blood levels of endotoxin and TNF-a. Prophylactic levofloxacin prevented pulmonary airway and vascular remodeling after CBDL, in association with lowering glucose-regulated protein levels, presumably by limiting bacterial translocation and inflammation [33]. The confirmation that TNF-a is implicated in experimental HPS was also achieved by ameliorating hypoxemia with a TNF-a monoclonal antibody [28]. The CBDL model has provided a platform to study the mechanisms that underlie IPVD and HPS. The progress that has been made in defining the importance of inflammation, pulmonary angiogenesis, and vascular remodeling in experimental HPS may provide additional targets that can be manipulated therapeutically in humans, thereby increasing our treatment options beyond liver transplantation.
HUMAN DISEASE Although pharmacologic interventions to improve oxygenation in patients with HPS are currently lacking, two recent reports have shown potential benefit with pentoxifylline [34 ] and mycophenolate mofetil [35]. Although prior studies have shown conflicting results regarding the ability of pentoxifylline to improve oxygenation (attributed to poor tolerability from gastrointestinal side-effects), the current report did show benefit in six of ten children with HPS who were able to tolerate the medication. These six children had an average increase in PaO2 of 26 mmHg, which then diminished after the treatment was held. All children experienced nausea and vomiting, which was the cause for three of the four children who did not tolerate treatment to discontinue its use. In addition to this small pilot study, a case was reported documenting the resolution of HPS with mycophenolate mofetil in a teenage boy with noncirrhotic portal hypertension [35], raising the possibility of its therapeutic benefit. The authors &
262
www.co-gastroenterology.com
postulated that the antiangiogenic effects of mycophenolate mofetil may have been the reason for improved oxygenation. However, the established ability of mycophenolate mofetil to inhibit endothelial cell production of ET-1 and also reduce nitric oxide may have also contributed to the observed therapeutic effect. Despite these small reports of therapeutic options and advancements that have been made in understanding the pathophysiology of HPS in the CBDL model, liver transplantation remains the only therapeutic option that can reliably improve oxygenation and survival in humans [36,37,38 ]. However, there has been continued interest in evaluating the effects of lowering portal pressure with transjugular intrahepatic shunt (TIPS) in patients with HPS [39–47]. Two recent cases of TIPS improving oxygenation in HPS have been published [48,49], increasing the number of reported cases to nine. However, similar to the pharmacologic therapies, TIPS has not been consistently beneficial in HPS, and the short duration of the follow-up reported makes it difficult to know whether TIPS is a reliable treatment option. Given the potential risk for hepatic decompensation and encephalopathy after TIPS placement and the relative lack of published data confirming utility, TIPS is not currently recommended to treat HPS. However, TIPS does not appear to worsen oxygenation in HPS, so it is generally considered as a reasonable option to manage comorbid refractory ascites and variceal bleeding in patients with established HPS. As severe HPS may occur in those with preserved hepatic synthetic function and adversely influence survival, patients with advanced disease (PaO2
Hepatopulmonary syndrome David G. Koch a and Michael B. Fallon b
Purpose of review To discuss the advances in the understanding of the pathophysiology of experimental and human hepatopulmonary syndrome (HPS) and in the management of HPS, particularly regarding liver transplantation. Recent findings Advances have been made in defining the pathophysiology of HPS in experimental models as well as in human disease, including the role of endothelin-1, pulmonary monocytes, and angiogenesis. Additionally, the implications of the presence of HPS as it relates to prioritizing patients for liver transplantation and posttransplant outcomes will also be reviewed. Summary Mechanisms of disease continue to be defined in HPS, providing potential targets for pharmacologic intervention. Outcomes after liver transplantation are also becoming clearer, including the management of HPS with severe hypoxemia. Keywords hepatopulmonary syndrome, hypoxemia, intrapulmonary vasodilatation, liver transplantation, portal hypertension
INTRODUCTION The hepatopulmonary syndrome (HPS) is a pulmonary complication that occurs in patients with chronic liver disease and/or portal hypertension, whereby hypoxemia results from alterations in the pulmonary microvasculature, including vasodilatation and angiogenesis. HPS is relatively common, occurring in up to 30% of patients with cirrhosis [1,2]. Although the degree of hypoxemia in HPS does not reliably correlate with the severity of liver disease, patients with HPS have increased mortality relative to cirrhosis patients without HPS. The diagnostic criteria for HPS include documentation of impaired oxygenation and intrapulmonary vasodilatation (IPVD) in the setting of chronic liver disease or portal hypertension [3] (see list below). Oxygenation in HPS is determined by calculating the alveolar–arterial oxygen gradient (AapO2) from an arterial blood gas (ABG); a threshold of greater than 15 mmHg is used except for patients older than age 64, in whom greater than 20 mmHg is appropriate. Therefore, patients with mild HPS may not have hypoxemia. The syndrome may occur in patients with comorbid primary lung diseases [4,5], and having cirrhosis is not imperative as HPS has been described in acute and chronic hepatitis without cirrhosis or portal hypertension www.co-gastroenterology.com
[6,7] as well as in noncirrhotic portal hypertension without chronic liver disease [8–11]. Much of our knowledge of the pathophysiology of HPS is based on a unique experimental model, common bile duct ligation (CBDL), which recreates the features of human HPS. Although progress has been made in defining the mechanisms underlying the pulmonary vascular changes in HPS, to date, there are no approved therapeutic options aside from liver transplantation. Diagnostic criteria for the HPS are as follows: (1) increased age-corrected alveolar–arterial oxygen gradient (>15 mmHg or >20 mmHg if age >64 years) while breathing ambient air: a
Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of South Carolina, Charleston, South Carolina and bDivision of Gastroenterology, Hepatology and Nutrition, Department of Internal Medicine, The University of Texas Medical School at Houston, Houston, Texas, USA Correspondence to David G. Koch, MD, MSCR, Division of Gastroenterology and Hepatology, Department of Internal Medicine, Medical University of South Carolina, 25 Courtenay Drive, ART 7100A, MSC 290, Charleston, SC 29425, USA. Tel: +1 843 792 6901; fax: +1 843 876 4301; e-mail: [email protected] Curr Opin Gastroenterol 2014, 30:260–264 DOI:10.1097/MOG.0000000000000067 Volume 30 Number 3 May 2014
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Hepatopulmonary syndrome Koch and Fallon
AapO2 ¼ [FiO2(Patm –pH2O)–( pCO2/0.8)] – PaO2, in which PaO2 is the partial pressure of arterial oxygen; FiO2 is the fraction of inspired oxygen; Patm is the atmospheric pressure; pH2O is the partial pressure of water vapor at body temperature; and paCO2 is the partial pressure of arterial carbon dioxide (0.8 corresponds to the standard gas-exchange respiratory ratio at rest); (2) positive contrast-enhanced echocardiography or lung-perfusion scanning (brain shunt fraction >6%); (3) cirrhosis and/or portal hypertension.
PATHOPHYSIOLOGY
correlated with the degree of bile duct proliferation in the corresponding liver biopsy specimens [26 ]. However, whether proliferating cholangiocytes in human cirrhosis produce and secrete ET-1 thereby raising the blood levels is still not known. A causal relationship between ET-1 and impairment in oxygenation in HPS that would support pharmacologic manipulation of the ET-1 and ETB receptor axis has also not been established. In addition to the role of ET-1 in experimental and human HPS, recent studies in experimental HPS have furthered our understanding of other pathways that contribute to the underlying pulmonary vascular phenomena. These include the roles of pulmonary intravascular monocytes [27 ], TNF-a [28], and angiogenesis [27 ,29 ] as well as the therapeutic utility in manipulating these pathways in order to improve oxygenation. The ability of serum from CBDL rats with HPS to induce a proliferative effect on pulmonary artery smooth muscle cells (PASMCs) and pulmonary microvascular endothelial cells (PMVECs) is also being understood. With respect to intrapulmonary monocytes, it is known that these cells accumulate in the lung vasculature after CBDL and that they contribute to IPVD by increasing the levels of nitric oxide [via inducible nitric oxide synthase (iNOS)] and carbon monoxide [via heme oxygenase-1 (HO-1)]. More recently, it has been demonstrated that the adherence of monocytes is regulated through ET-1 and ETB receptor activation which increases the levels of the monocyte chemokine, fractalkine (FKN/CX3CL1–CX3CR1). FKN induces pulmonary monocyte adhesion and also promotes angiogenesis directly and through monocyte secretion of vascular endothelial growth factor-A (VEGF-A) [27 ]. VEGF-A is also produced in the CBDL liver and induces cholangiocyte proliferation through extracellular-signal-regulated kinase (ERK) activation [30 ]. Furthermore, inhibiting the effects of VEGF-A with the receptor tyrosine kinase inhibitor, sorafenib, improves oxygenation in experimental HPS [29 ,30 ] by reducing pulmonary monocyte adhesion and subsequent VEGF-A production as well as by blocking cholangiocyte proliferation in the liver through inhibition of ERK activation [30 ]. Whether sorafenib has similar effects on cholangiocytes in human HPS and may improve hypoxemia is not known. However, the potential to manipulate this pathway pharmacologically in patients with HPS is attractive as sorafenib is approved for use in humans with cirrhosis. At the cellular level, serum from CBDL rats induces the proliferation of both PMVECs and PASMCs, effects that could contribute to lung angiogenesis and remodeling in HPS. CBDL serum &
&&
The CBDL experimental model recreates the features of human HPS [12–20] (Fig. 1). In this model, proliferating cholangiocytes in the liver produce and secrete endothelin-1 (ET-1) [12–16], whereas other models of portal hypertension that do not result in bile duct proliferation do not develop HPS [17]. Increased shear stress in the pulmonary microvasculature, as a result of a hyperdynamic circulation of cirrhosis, results in upregulation of the endothelin B (ETB) receptor that subsequently binds the increased circulating ET-1 and augments pulmonary nitric oxide production via endothelial nitric oxide synthase (eNOS) [12,15,19,21–25]. The potential role of biliary ET-1 in human HPS was shown in a single study, in which hepatic venous blood levels of ET-1 were higher in cirrhosis patients with IPVD and HPS compared with those without IPVD. In addition, the hepatic venous ET-1 levels in these patients
&&
&&
&&
Liver Cholangiocyte ET-1
Monocyte iNOS
&&
HO-1
ETBR
VEGF-A
&&
CO p-AKT/p-ERK
p-eNOS/NO
Endothelial Cell Vasodilatation
Angiogenesis
HPS
FIGURE 1. Mechanisms of hypoxemia in the common bile duct ligation model of HPS. CO, carbon monoxide; ET-1, endothelin-1; ETBR, endothelin B receptor; HO-1, heme oxygenase-1; HPS, hepatopulmonary syndrome; iNOS, inducible nitric oxide synthase; NO, nitric oxide; p-eNOS, phosphorylated endothelial nitric oxide synthase; VEGF-A, vascular endothelial growth factor A.
0267-1379 ß 2014 Wolters Kluwer Health | Lippincott Williams & Wilkins
&&
&&
www.co-gastroenterology.com
261
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Liver
increased the cytoskeleton proteins within PMVECs cells and promoted proliferation by decreasing the expression levels of Annexin A1, a Ca[2þ]-dependent phospholipid-binding protein [31]. Conversely, CBDL serum upregulates Annexin A2, promoting PASMC proliferation [32]. Either promoting Annexin A1 expression or inhibiting Annexin A2 was found to ameliorate hypoxemia in experimental HPS [31,32]. Finally, new insights into how bacterial translocation and endotoxemia drive inflammation in experimental HPS have been made. Evidence of endoplasmic reticulum stress as evidenced by increased lung levels of 78 kDa glucose-regulated protein are found after CBDL. This protein, in turn, has a strong, positive correlation with the blood levels of endotoxin and TNF-a. Prophylactic levofloxacin prevented pulmonary airway and vascular remodeling after CBDL, in association with lowering glucose-regulated protein levels, presumably by limiting bacterial translocation and inflammation [33]. The confirmation that TNF-a is implicated in experimental HPS was also achieved by ameliorating hypoxemia with a TNF-a monoclonal antibody [28]. The CBDL model has provided a platform to study the mechanisms that underlie IPVD and HPS. The progress that has been made in defining the importance of inflammation, pulmonary angiogenesis, and vascular remodeling in experimental HPS may provide additional targets that can be manipulated therapeutically in humans, thereby increasing our treatment options beyond liver transplantation.
HUMAN DISEASE Although pharmacologic interventions to improve oxygenation in patients with HPS are currently lacking, two recent reports have shown potential benefit with pentoxifylline [34 ] and mycophenolate mofetil [35]. Although prior studies have shown conflicting results regarding the ability of pentoxifylline to improve oxygenation (attributed to poor tolerability from gastrointestinal side-effects), the current report did show benefit in six of ten children with HPS who were able to tolerate the medication. These six children had an average increase in PaO2 of 26 mmHg, which then diminished after the treatment was held. All children experienced nausea and vomiting, which was the cause for three of the four children who did not tolerate treatment to discontinue its use. In addition to this small pilot study, a case was reported documenting the resolution of HPS with mycophenolate mofetil in a teenage boy with noncirrhotic portal hypertension [35], raising the possibility of its therapeutic benefit. The authors &
262
www.co-gastroenterology.com
postulated that the antiangiogenic effects of mycophenolate mofetil may have been the reason for improved oxygenation. However, the established ability of mycophenolate mofetil to inhibit endothelial cell production of ET-1 and also reduce nitric oxide may have also contributed to the observed therapeutic effect. Despite these small reports of therapeutic options and advancements that have been made in understanding the pathophysiology of HPS in the CBDL model, liver transplantation remains the only therapeutic option that can reliably improve oxygenation and survival in humans [36,37,38 ]. However, there has been continued interest in evaluating the effects of lowering portal pressure with transjugular intrahepatic shunt (TIPS) in patients with HPS [39–47]. Two recent cases of TIPS improving oxygenation in HPS have been published [48,49], increasing the number of reported cases to nine. However, similar to the pharmacologic therapies, TIPS has not been consistently beneficial in HPS, and the short duration of the follow-up reported makes it difficult to know whether TIPS is a reliable treatment option. Given the potential risk for hepatic decompensation and encephalopathy after TIPS placement and the relative lack of published data confirming utility, TIPS is not currently recommended to treat HPS. However, TIPS does not appear to worsen oxygenation in HPS, so it is generally considered as a reasonable option to manage comorbid refractory ascites and variceal bleeding in patients with established HPS. As severe HPS may occur in those with preserved hepatic synthetic function and adversely influence survival, patients with advanced disease (PaO2
