Hepatopulmonary syndrome in adults: Natural history, treatment, and outcomes

INTRODUCTION — Hepatopulmonary syndrome (HPS) is characterized by the triad of abnormal arterial oxygenation caused by intrapulmonary vascular dilatations in the setting of liver disease, portal hypertension, or congenital portosystemic shunts [1,2].

The natural history, treatment, and outcomes of HPS are reviewed here. The epidemiology, pathophysiology, clinical manifestations, and diagnostic evaluation are discussed separately. (See "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis".)

NATURAL HISTORY — HPS is typically a progressive disorder, the presence of which worsens the prognosis of patients with cirrhosis and probably other liver diseases [1,3-7]. The cause of death among patients with HPS tends to be multifactorial and related to complications of underlying liver disease (eg, hepatic failure, multisystem organ failure due to sepsis, hepatocellular cancer, gastrointestinal bleeding) rather than from HPS-related hypoxemic respiratory failure [1,3,4,8]. Spontaneous resolution of HPS without treatment is unlikely. These conclusions are derived from the following studies:

●One observational study of 138 liver transplant candidates reported a higher mortality in patients with HPS compared with case-matched controls (78 versus 43 percent) [4]. Similarly, patients with HPS had a lower median survival (24 versus 87 months) and lower five-year survival (23 versus 63 percent).

●A single center prospective study of 316 patients referred for liver transplantation reported that compared with patients without HPS, patients with HPS had a nonsignificant increase in waitlist mortality (24 versus 16 percent) and a slightly lower pre-liver transplant survival (35 versus 42 months) [5].

●In a prospective study of 111 patient with cirrhosis, those with HPS had a lower median survival than those without HPS (11 versus 41 months), even when adjusted for severity of underlying liver disease [3]. HPS was an independent predictor of survival.

●A retrospective series of 22 patients with HPS, most of whom had severe HPS, reported an overall mortality of 41 percent, occurring, on average, 2.5 years after diagnosis [6].

●A multicenter, prospective cohort study (85 patients with HPS and 146 patients without HPS) of adults undergoing their first liver transplantation evaluation found that HPS was associated with worse exercise and functional capacity and an overall increased risk of death [8]. Patients with HPS had a lower probability of being alive than control patients with liver disease at one year (87 versus 92 percent), two years (73 versus 83 percent), and three years (63 versus 81 percent). The differences in overall risk of death for HPS did not differ based on partial arterial pressure of oxygen or alveolar-arterial oxygen gradient, suggesting that the relationship between HPS and worse outcomes was not dependent on the severity of HPS.

Although oxygenation usually worsens over time (mean decline of 5.2 mmHg per year [range 0.4 to 8.3 mmHg per year]) [4], it is rare for progressive hypoxemic respiratory failure to be the primary cause of death.

TREATMENT AND PROGNOSIS — The only definitive therapy for patients with HPS is liver transplantation (LT), which is reserved for those with severe or very severe HPS (table 1). Other than long-term supplemental oxygen, there are no effective medical therapies for HPS, although many approaches have been attempted to improve gas exchange and decrease hypoxemia (table 2) [1,9]. Rarely, resolution of HPS in response to treating the underlying acute liver disease has been reported (eg, steroids for patients with granulomatous hepatitis) [10]. Grading disease severity is discussed separately. (See "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Grading disease severity'.)

Mild to moderate hepatopulmonary syndrome — Most patients in this category require clinical monitoring for disease progression, and some may need supplemental oxygen, when indicated.

Observation — Most patients with mild to moderate HPS require monitoring every 6 to 12 months with pulse oximetry (ideally including measurements done in a standing position) and, when indicated, arterial blood gas analysis to determine worsening HPS that may prompt a more aggressive treatment strategy with LT and/or oxygen supplementation, early in the course of disease progression. In addition, a six‑minute walk test (6MWT) may also aid in the early detection of HPS and assessment of functional capacity [11]. Transthoracic contrast echocardiography does not need to be repeated to monitor shunt. (See "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Diagnostic evaluation'.)

Oxygen supplementation — Patients with mild to moderate HPS with resting partial arterial pressure of oxygen (PaO₂) >55 mmHg (7.3 kPa) or peripheral arterial saturation >88 percent do not fulfill the criteria for oxygen supplementation (table 3) unless exercise-induced or nocturnal hypoxemia is present. Thus, in this population, we typically perform periodic 6MWT and nocturnal oximetry. (See "Long-term supplemental oxygen therapy" and "Overview of pulmonary function testing in adults", section on 'Six-minute walk test'.)

Severe or very severe hepatopulmonary syndrome — Most patients in this category require supplemental oxygen therapy and should be evaluated for LT.

Oxygen supplementation — Long-term supplemental oxygen therapy (LTOT) is the most frequently recommended therapy for patients with severe or very severe HPS [1]. Indications for oxygen supplementation are similar to those used for patients with severe hypoxemia due to chronic pulmonary disease (table 3), the details of which are provided separately. (See "Long-term supplemental oxygen therapy".)

LTOT is not a cure for HPS. LTOT only improves features related to intrapulmonary vascular shunts (eg, dyspnea, fatigue, desaturation) in patients who are hypoxemic due to HPS, although the only evidence that supports this conclusion is indirect evidence extrapolated from patients with hypoxemia from other lung diseases (eg, chronic obstructive pulmonary disease) and anecdotal clinical observations. Controlled clinical trials comparing LTOT to no therapy in HPS have not been performed.

Liver transplantation — Patients with HPS and a PaO₂ <60 mmHg (8 kPa; ie, severe and very severe HPS (table 1)) should be evaluated for LT. This strategy is based upon observational studies that demonstrate complete or near complete resolution of HPS with improved oxygenation and shunt in the majority (about 80 percent) of patients within the initial 6 to 12 months [4,12-23] and studies that report a similar mortality in those with and without HPS who undergo LT [5,24-27]. In support of this approach, the 2013 American Association for the Study of Liver Disease and American Society of Transplantation Practice Guidelines and the International Liver Transplant Society practice guidelines recommend an expedited referral and evaluation for LT for patients with severe HPS [2,28]. (See "Model for End-stage Liver Disease (MELD)", section on 'Standard MELD exceptions in liver transplantation' and "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)

While LT may be curative, patients with HPS may have significant postoperative challenges. In a single LT center experience, HPS diagnosis was associated with longer intensive care unit (ICU) stay, longer hospital stay, and increased hospital cost, together with higher odds of being discharged to an extended care facility compared with non-HPS patients [29]. Similarly, a case-control study of 71 HPS patients from an LT center in France reported a higher number of postoperative complications in patients with HPS, with greater frequency of cardiac, infectious, and surgical complications than in control patients. There were also more ICU readmissions at one month among HPS patients [30], an effect not seen in another study [29].

Since some patients with mild liver disease who have advanced HPS may not be eligible for LT (based upon the severity of their underlying liver disease as assessed by Model for End-Stage Liver Disease (MELD) points [31]), the Organ Procurement Transplant Network/United Network for organ sharing (UNOS) policy assigns a standard MELD exception score of 22 for patients with evidence of intrapulmonary shunting and a room air PaO₂ <60 mmHg (8 kPa), with a 10 percent mortality equivalent increase in points every three months if the PaO₂ remains <60 mmHg. Some also allocate increased exception scores in those with a PaO₂ of 50 mmHg (6.7 kPa) [2]. (See "Model for End-stage Liver Disease (MELD)", section on 'Standard MELD exceptions in liver transplantation' and "Liver transplantation in adults: Patient selection and pretransplantation evaluation".)

Although successful outcomes have been reported in patients with HPS who have received living donor LT, cadaveric LT remains the primary graft type for this population [32-34].

Pathologic changes in the lungs after LT are poorly studied. One older post-mortem study reported improvement or resolution of the pathologic changes and collagen tissue deposition in capillary and venule walls [35]. Further studies are required to understand the mechanism of resolution.

Mortality — LT appears to improve survival among patients with HPS and is similar to those who undergo LT who do not have HPS:

●In one observational study of 61 patients with HPS, mortality was higher in those who did not undergo LT compared with those who had LT (78 versus 21 percent) [4]. Among patients who underwent LT, those with HPS had a five-year survival rate of 76 percent, which was not significantly different from the five-year survival rate of patients without HPS [4].

●In another observational study of 59 transplant recipients, similar short-term survival (77 versus 68 percent) and long-term survival (64 versus 60 percent) was reported in those with and without HPS; survival was independent of the severity of HPS [24].

●In a series of 106 patients with HPS that compared outcomes in patients who underwent LT in the pre- and post-MELD exception eras, posttransplant survival was consistently better in the post-MELD era: 92 versus 71 percent at one year, 88 versus 67 percent at three years, and 88 versus 67 percent at five years [25].

●Results of a meta-analysis of patients with HPS undergoing LT [27] confirm that survival post-LT for HPS was higher in the time after a MELD exception was implemented for HPS than before, at least in adults (but not in children).

The early postoperative period tends to be the most dangerous, and patients who survive this period tend to do well [12,13]. In one observational study, patients with HPS who survived the initial 10 weeks following transplantation subsequently survived for at least one year.

The etiologies associated with postoperative death are unknown but may be due to refractory postoperative hypoxemia, portal venous thrombosis, intracranial events, infection [36], and multiorgan failure [37]. (See 'Refractory hepatopulmonary syndrome' below and "Liver transplantation in adults: Long-term management of transplant recipients" and "Infectious complications in liver transplantation" and "Liver transplantation in adults: Clinical manifestations and diagnosis of acute T-cell mediated (cellular) rejection of the liver allograft".)

Prognostic factors — Severe hypoxemia may predict a worse outcome in patients with HPS who undergo LT.

●Pre-transplantation hypoxemia – Despite conflicting data, we and others consider that patients with severe deficits in oxygenation are high risk candidates for LT, but that this should not preclude them from undergoing LT [2]. Conflicting data [27] may relate to altered transplantation strategies due to MELD exception criteria or to differences in practice between centers of excellence. (See 'Liver transplantation' above and "Model for End-stage Liver Disease (MELD)", section on 'Standard MELD exceptions in liver transplantation'.)

Several studies have reported that severe hypoxemia (45 to 50 mmHg; 6.0 to 6.7 kPa) pre-LT is associated with an increased posttransplant mortality [12,18,26,38,39], although the exact cut-off is unclear:

•In a prospective evaluation of 24 patients with cirrhosis and LT, a preoperative room air arterial oxygen tension (PaO₂) ≤50 mmHg alone or in combination with a shunt fraction ≥20 percent on macroaggregated albumin scanning (MAA) were strong predictors of postoperative mortality [12]. (See "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Macroaggregated albumin scanning'.)

•In a retrospective cohort study, patients on the LT wait list were analyzed using data derived from the United Network for Organ Sharing (UNOS) from 2002 to 2012 [26]. Rates of three-year unadjusted posttransplantation survival were lower in those with a PaO₂ ≤44 mmHg compared with those who had a PaO₂ >44 mmHg (68 versus 84 percent).

•In a retrospective cohort study from the UNOS/Organ Procurement and Transplant Network database of all adult patients undergoing LT between January 1, 2006, and January 12, 2020, HPS patients with a pre-LT PaO₂ <54 mmHg demonstrated increased mortality following LT as compared with matched non-HPS patients with cirrhosis [38].

In contrast, other observational studies published after instituting the MELD exception criteria have demonstrated that severe hypoxemia or shunt do not predict mortality in this population [14,25,40]. For example, in a retrospective cohort study of patients younger than 18 years with a MELD/pediatric liver disease (PELD) exception request who underwent LT between 2007 and 2018, 124 patients (4 percent) received MELD/PELD exception for HPS. When stratified by pre-LT PaO₂, hypoxemia severity was not associated with differences in one-, three-, or five-year survival rates after LT [40]. Taken together, reports from large-volume, experienced centers tend to show that outcomes post-LT for HPS patients are similar to those following LT for non-HPS individuals [27].

Little is known about whether postoperative improvement in oxygenation can be predicted, although studies suggest that severe hypoxemia or a higher baseline MAA shunt fraction was associated with a lower rate of postoperative improvement in oxygenation and greater need for long term oxygen supplementation [14,15,25,41]. Severe hypoxemia may also predict longer recovery and hospital length of stay.

●Others – Other prognostic factors that predict outcome in this population are poorly studied but some observational studies report that severity of underlying liver disease, age, and elevated BUN may contribute [3].

Oxygenation — The abnormalities of gas exchange improve or resolve in approximately 80 percent of patients following LT, although it is uncertain whether or not gas exchange abnormalities in these studies were solely due to shunt [4,12-14,16-23,42]. In an observational study of 74 consecutive patients with chronic liver disease who underwent a comprehensive cardiopulmonary assessment before and after LT, 50 percent had an abnormal alveolar-arterial (A-a) oxygen gradient (defined as ≥15 mmHg) and 22 percent were hypoxemic (defined as a PaO₂ <80 mmHg) before transplantation [23]. Following transplantation, the mean PaO₂ increased (from 89 to 94 mmHg) and the mean A-a gradient decreased (from 16 to 8 mmHg). However, the diffusion capacity of carbon monoxide (DLCO) may not improve after LT, suggesting that subclinical pulmonary vascular changes persist [43].

The time course by which HPS improves or resolves after LT is variable. Some studies have demonstrated improvement in the shunt fraction and hypoxemia within days after LT while others have demonstrated improvement over 2 to 14 months [4,5,19,44-46]. This wide variation may be due, at least in part, to the studies testing oxygenation at different intervals after transplantation. It may also reflect variability among patients in how much of their hypoxemia is due to ventilation-perfusion mismatch rather than intrapulmonary shunting (it has been hypothesized that the likelihood of improving oxygenation after LT is greater when the principal mechanism of hypoxemia is ventilation-perfusion mismatching) [47] or how much vascular remodeling has occurred (it has been hypothesized that vascular remodeling may be a more important cause of gas exchange abnormalities than NO-dependent vasodilatation) [48]. One case report documents normalization of oxygenation following LT for HPS after 67 days on extracorporeal membrane oxygenation support in a 17-year-old with end-stage liver disease and HPS [49].

Intraoperative and postoperative management should focus on continuous monitoring of peripheral and central oxygenation while the patient is mechanically ventilated in the supine position, preferably using a lung protective strategy. Clinicians should be careful not to over-resuscitate or over-diurese this group. Management strategies for severe postoperative hypoxemia are discussed below. (See "Pulse oximetry" and "Acute respiratory distress syndrome: Ventilator management strategies for adults" and 'Refractory hepatopulmonary syndrome' below.)

Follow-up — There are no clear guidelines regarding follow-up measures in HPS patients after LT other than periodic pulse oximetry and routine transplant protocols; contrast echocardiography is not necessary unless indicated for another reason. Supplemental oxygen may be discontinued when the resting PaO₂ is >55 mmHg (7.3 kPa) and/or there is no sleep- or exercise-induced requirement. (See "Liver transplantation in adults: Long-term management of transplant recipients".)

Refractory hepatopulmonary syndrome — Refractory HPS applies to patients who fail to improve after LT or patients who develop recurrent hypoxemia post-LT. This population should be evaluated for the spontaneous recurrence of HPS using the same strategy outlined prior to LT (see "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Diagnostic evaluation') as well as for the development of portopulmonary hypertension (PPHTN), and/or other diseases that may be contributing to hypoxemia (eg, pulmonary arteriovenous malformations) [50-52]. Options in this population are limited to maximizing oxygen supplementation (eg, high-flow oxygen, transtracheal oxygen) and/or cautious trials of less well validated/investigational therapies. For patients with severe progressive HPS who are not candidates for LT or patients who cannot be adequately oxygenated with supplemental oxygen due to significant shunt, therapies are similarly limited. (See 'Investigational' below.)

Severe posttransplant hypoxemia (defined as 100 percent oxygen required to maintain saturation ≥85 percent) is seen in 6 to 21 percent of patients and is a common cause of mortality following LT [12,53,54]. Management strategies include placing the patient in the Trendelenburg position [55], inhaled epoprostenol [54], inhaled nitric oxide [56], methylene blue [57], embolization of pulmonary arteriovenous malformations (if present) [58], and extracorporeal life support with both venovenous and venoarterial strategies employed [59,60]. A choice among these therapies is institution-dependent and should probably start with the least invasive strategy and escalate to riskier or combination therapies [53].

Uncommonly, post-LT resolution of HPS has been associated with the development of PPHTN, which can be managed in a manner similar to non-HPS-associated PPHTN. (See "Portopulmonary hypertension", section on 'Treatment'.)

Other therapies — Rare case reports of miscellaneous therapies have been published. As examples:

●An isolated case of severe HPS in response to steroids was reported in a patient with granulomatous hepatitis [10], suggesting that monitoring patients with HPS who are undergoing active therapy for their underlying liver disease is prudent.

●Similarly, case reports of HPS resolution have been published in response to inferior vena cava (IVC) stenting (in patients with HPS and a spontaneous IVC-portal vein shunt) and ligation of congenital portosystemic shunts (eg, Abernethy malformations associated with HPS have resulted in resolution of HPS) [61-63].

INVESTIGATIONAL — Various interventions and medications have been tried for patients with HPS, including transjugular intrahepatic portosystemic shunt (TIPS), coil embolization, and vasoactive medications, none of which has clear benefit.

Transjugular intrahepatic portosystemic shunt — The reduction of portal pressures by TIPS placement is an intervention that has been associated with improvement of HPS in several case reports [64-71]. However, we do not advocate routine TIPS placement in patients with HPS, because clinical outcomes have been variable [9,69,72]. In addition, there is a risk that TIPS may worsen HPS by increasing the hyperkinetic state, leading to more pulmonary vasodilatation, shunting, and hypoxemia.

Embolization — Rare patients with HPS have large intrapulmonary vascular dilatations (eg, type II lesions), which are akin to pulmonary arteriovenous malformations (identifiable as contrast-enhanced nodular densities on CT imaging) [73]. Such lesions may be amenable to embolization, with case reports suggesting modest improvements in oxygenation with this procedure (image 1) [74,75]. (See "Hepatopulmonary syndrome in adults: Prevalence, causes, clinical manifestations, and diagnosis", section on 'Contrast pulmonary angiography' and "Therapeutic approach to adult patients with pulmonary arteriovenous malformations" and "Overview of transjugular intrahepatic portosystemic shunts (TIPS)".)

Vasoactive medications

Garlic (containing allium sativum) — Garlic (containing allium sativum) is thought to decrease nitric oxide synthesis, thereby acting as a potential therapy for HPS. As an example, one prospective randomized, placebo-controlled trial evaluated the effects of oral garlic supplementation (1 to 2 g/m²/day) in 41 patients with HPS [76]. After nine months, compared with placebo, garlic supplementation resulted in a three-fold increase in baseline arterial oxygen levels and a decrease in alveolar-arterial oxygen gradient. Reversal of HPS was observed in 14 of 21 patients treated with garlic compared with only 1 of 20 patients in the placebo group (67 versus 5 percent). Mortality was also lower among patients receiving garlic (10 versus 35 percent).

Others — Agents that have been tested in animal models and appear promising include pentoxifylline (a nonspecific phosphodiesterase inhibitor with inhibitory effects on TNF-alpha synthesis) and quercetin (a flavonoid antioxidant) [77,78]. Two small prospective, open-label studies of pentoxifylline (400 mg every eight hours) in patients with HPS have shown mixed results regarding oxygenation, leaving the role of pentoxifylline unclear [79,80]. In a pilot study of 10 children with HPS, pentoxifylline was given at a dose of 20 mg/kg/day in three divided doses. Four children did not tolerate the drug due to severe vomiting and thrombocytopenia. The six children who completed the three-month trial demonstrated significant improvement in oxygenation; however, these improvements were not maintained after discontinuation of pentoxifylline [81].

Various other medications have been tried, but anecdotal evidence and uncontrolled trials suggest that such medications cause little or no sustained improvement in oxygenation. Examples include methylene blue, N(G)-nitro-L-arginine methyl ester (L-NAME), curcumin, terlipressin, somatostatin analogues (eg, octreotide), nitric oxide synthase inhibitors, cyclooxygenase inhibitors (eg, indomethacin), almitrine bismesylate, antibiotics, chemotherapy (eg, cyclophosphamide), glucocorticoids, beta blockers (eg, propranolol), mycophenolate mofetil, aspirin, inhaled nitric oxide, and sorafenib (table 2) [1,9,81-89].

SUMMARY AND RECOMMENDATIONS

●Introduction – Hepatopulmonary syndrome (HPS) is characterized by the triad of abnormal arterial oxygenation caused by intrapulmonary vascular dilatations in the setting of liver disease, portal hypertension, or congenital portosystemic shunts. (See 'Introduction' above.)

●Natural history – HPS is typically a progressive disorder, the presence of which worsens the prognosis of patients with cirrhosis and probably other liver diseases. It is rare for progressive hypoxemic respiratory failure to be the primary cause of death. Spontaneous resolution of HPS without treatment is unlikely. (See 'Natural history' above.)

●Management – The only definitive therapy for patients with HPS is liver transplantation. Other than long-term supplemental oxygen, there are no effective medical therapies for HPS (table 2) [1,9].

•Mild to moderate HPS – For most patients with mild to moderate HPS (partial arterial pressure of oxygen [PaO₂] >60 mmHg [>8 kPa]) (table 1), we suggest surveillance with pulse oximetry every 6 to 12 months for disease progression. In addition, for those in whom it is indicated (table 3), we suggest supplemental oxygen rather than liver transplantation (Grade 2C). Although liver transplantation is not routinely indicated in this population, we typically refer patients early in the course of their HPS for liver transplant evaluation, so that as HPS progresses, activation on the waitlist can be prompt. (See 'Mild to moderate hepatopulmonary syndrome' above.)

•Severe HPS – For most patients with severe or very severe HPS, we recommend an expedited liver transplant evaluation together with oxygen supplementation (Grade 1B). (See 'Liver transplantation' above.)

•Refractory HPS – Most patients (80 percent) improve following liver transplantation, although the time course for improvement is highly variable (days up to one year). Patients with refractory HPS following liver transplantation should be treated with high flow oxygen and evaluated for the development of portopulmonary hypertension or other conditions that may be contributing to hypoxemia. HPS rarely recurs after LT. (See 'Refractory hepatopulmonary syndrome' above.)

●Investigational therapies – For patients with HPS who are refractory to or not eligible for LT, options are limited and include investigational therapies such as transjugular intrahepatic portosystemic shunt, coil embolization, and garlic. (See 'Investigational' above.)