FormalPara Key Summary Points

Why carry out this study?

Portopulmonary hypertension (PoPH), a rare and severe form of pulmonary arterial hypertension (PAH), carries the worst prognosis; however, data regarding the use of PAH-specific therapies in patients with PoPH are sparse.

The combined OPUS and OrPHeUS studies provide a unique insight into contemporary real-world clinical practice for the management of patients with PoPH newly treated with macitentan, including patient characteristics, treatment patterns, outcomes, and safety profiles.

What was learned from the study?

At baseline, patients with PoPH had high levels of liver cirrhosis, liver comorbidities, and liver test abnormalities. Over the course of the study, a large proportion of patients with PoPH were in medical need of liver transplantation but did not meet waitlist registration criteria. Additionally, their hemodynamics were indicative of severe PoPH disease (mean pulmonary artery pressure ≥ 45 mmHg in the presence of increased pulmonary vascular resistance).

The majority of patients with PoPH initiated macitentan as a combination therapy: Given the underlying liver disease in the PoPH population, there were no unexpected safety findings and tolerability, hospitalizations, and survival were in line with previous expectations.

These data add to previous evidence supporting the general safety and tolerability of macitentan in patients with PoPH, including as a combination therapy, and highlight the need for improved management and stabilization of this severely ill population to facilitate eligibility for liver transplantation.

Digital Features

This article is published with digital features, including a graphical abstract, to facilitate understanding of the article. To view digital features for this article, go to https://doi.org/10.6084/m9.figshare.24807438.

Introduction

Portopulmonary hypertension (PoPH) is a rare and life-threatening disease defined as pulmonary arterial hypertension (PAH) associated with portal hypertension, with or without cirrhosis [1]. PoPH represents a considerable proportion (5–18%) of patients with PAH [2,3,4,5,6,7] and carries the worst prognosis of all PAH etiologies. This is due not only to the pathogenesis of PAH but also to the underlying liver disease, with 3-year survival estimates of 52–65% [1, 8, 9], compared to 77–80% in idiopathic or heritable PAH (I/HPAH) [8, 10].

Current European Society of Cardiology/European Respiratory Society (ESC/ERS) pulmonary hypertension (PH) Guidelines recommend the use of approved PAH drugs to treat patients with PoPH [6, 7]. However, evidence supporting use of these therapies in this population remains limited [11, 12], as clinical trials usually exclude such patients because of their underlying liver disease [6, 7, 12] and real-world data are scarce. The only randomized controlled trial (RCT) for a targeted PAH therapy in PoPH is PORTICO (N = 85), which investigated the safety and efficacy of the orally administered endothelin receptor antagonist (ERA) macitentan in these patients [13, 14]. In PORTICO, macitentan significantly improved pulmonary vascular resistance (PVR) by 35% versus placebo, with no observed worsening of hepatic function or portal hypertension, or unexpected hepatic safety concerns [13]. In PoPH, improvement in hemodynamics may not only be essential to delay disease progression but also to increase eligibility for a liver transplant, which offers the best chance of survival [14].

The prospective OPsumit® USers (OPUS) registry and the retrospective OPsumit® Historical USers (OrPHeUS) medical chart review evaluated the safety profile of macitentan, with a focus on hepatic safety. Combined, the dataset comprises the largest PH drug-specific registry to date, with 5654 patients with follow-up data [15], including 206 patients with PoPH. This analysis describes the patient characteristics, treatment patterns, overall and hepatic safety profiles, and outcomes of patients with PoPH newly treated with macitentan in a real-world clinical setting, alongside results for patients with I/HPAH.

Methods

OPUS and OrPHeUS Study Design

The OPUS and OrPHeUS studies have been previously described [15]. Briefly, OPUS was a prospective, multicenter, US, observational drug registry (April 2014–June 2020; NCT02126943) and OrPHeUS a retrospective, multicenter, US, medical chart review (October 2013–March 2017; NCT03197688); both studies enrolled patients newly initiating macitentan and were mutually exclusive.

Monitoring and Ethical Approval

OPUS and OrPHeUS were executed in accordance with Good Pharmacoepidemiology Practices [16] and the 2008 Declaration of Helsinki ethical principles. Ethical approval was received from independent ethics committees/institutional review boards of participating centers (Supplementary Appendix I). The protocols were reviewed by the US Food and Drug Administration (FDA) with written informed consent obtained from all patients in OPUS, including for publication of anonymized patient data (informed consent was not required in OrPHeUS as an institutional review board [IRB] waiver was obtained). IRB approvals were provided by the Western IRB and Quorum (now Advarra) (OPUS registry; Western IRB approval number 2014-0816, Quorum Review File number 29120/Advarra Pro00035124) and WCG-IRB (OrPHeUS study; IRB numbers 2017-8051 and 2017-2348).

Observations and Assessments

The observation period was from the date of macitentan initiation (which may not have coincided with the date of diagnosis) to study end, or until the first of death, loss to follow-up, withdrawal of consent, or macitentan discontinuation plus 30 days. For both OPUS and OrPHeUS, follow-up data was defined as at least one observation after macitentan initiation. Data collection during macitentan treatment in OPUS and OrPHeUS (including demographics, baseline characteristics, treatment patterns, safety, hospitalizations, and deaths) has been previously described [15]. Information was collected per routine clinical practice and no assessments were mandated. Both studies collected information on hepatic adverse events (HAEs) and HAEs of special interest (HAESIs; Supplementary Appendix II). Additional data collected in OPUS only included non-hepatic adverse events (AEs) and details regarding patients’ liver disease (liver transplantation status, etiology of portal hypertension, Model for End-stage Liver Disease [MELD], and/or Child Pugh score), the latter retrospectively collected via an additional electronic case report form (eCRF) prior to study closure (Supplementary Appendix III). Hepatic safety data were reviewed by an independent liver safety data review board (ILSDRB; Supplementary Appendix IV) as previously described.

Statistical and Other Analyses

Statistical and other analyses in OPUS and OrPHeUS have been previously described [15]. All analyses were descriptive, and no formal statistical comparisons were made between populations. The I/HPAH analysis set comprised all patients with idiopathic or heritable PAH entered as the only reason for macitentan prescription (i.e., patients with multiple PH diagnoses or PAH etiologies reported were excluded). This population is included as it is the most well characterized form of PAH and serves as a reference and anchor for the PoPH data. To be conservative, the PoPH analysis set comprised all patients with PAH associated with portal hypertension entered as either the sole reason for macitentan prescription or in conjunction with another diagnosis (i.e., patients with PoPH and another PH diagnosis or PAH etiology reported were included). Etiologies in OPUS/OrPHeUS were investigator-assessed.

Patient demographics, baseline characteristics, safety information, treatment patterns, and outcomes (hospitalizations and survival) are described for the PoPH and I/HPAH populations. However, as no adjustments were made to account for differences in baseline characteristics, the two populations are not directly compared. Liver disease status during the observation period, defined as 1 year prior to initiation and 30 days after discontinuation of macitentan, is presented for the OPUS PoPH set. All event rates were calculated using the time to first event. All Poisson models included log (exposure time) as an offset to account for varying length of patients’ time on treatment. Confidence limits (95% CL) for rates per person-year were estimated using an unadjusted Poisson model. Time to first events for treatment discontinuation, treatment escalation, hospitalization, and death are presented using Kaplan–Meier (KM) estimates. All analyses were conducted until the end of the observation period. Patients were included in each analysis until first occurrence of the specified event, or until the end of observation period, whichever occurred first. Imputation rules for missing dates are detailed in Supplementary Appendix V; no other data imputations were made.

Results

Study Population

Of the 155 sites that enrolled patients in OPUS/OrPHeUS, 67 contributed to the PoPH and 150 to the I/HPAH population. The OPUS/OrPHeUS PAH population included 4626 patients with follow-up data, including 206 (4.5%) PoPH and 2498 (54.0%) patients with I/HPAH (Fig. 1). Patients with PoPH were younger than patients with I/HPAH (median age, 58 vs 64 years) and fewer were female (52.4% vs 73.1%; Table 1). The median (Q1, Q3) time from PAH diagnosis to macitentan treatment was similar in both populations (7.2 [1.1, 35.4] vs 7.0 [1.3, 36.8] months). Acknowledging the significant amount of missing data (summarized in Table S1), for patients with measurements, the majority in both populations were World Health Organization (WHO) functional class III (52.0% vs 55.0%); however, the median 6-min walk distance (6MWD) was higher (317 vs 289 m) and the proportion of patients in the high-risk category for brain natriuretic peptide/N-terminal pro-brain natriuretic peptide was lower (23.2% vs 35.2%) for patients with PoPH than patients with I/HPAH. Median hemodynamics were broadly similar in both populations.

Fig. 1
figure 1

Patient disposition. aThe remaining patient population of OPUS and OrPHeUS are listed in the “other PAH etiologies group”. bCompared to the analysis sets reported in McLaughlin et al. [15], the PoPH set herein includes 5 additional patients who had PoPH and another reported PH diagnosis or PAH etiology, while the I/HPAH group herein excludes 100 patients who had multiple reported PH diagnoses. AE adverse event, CHD congenital heart disease, CTD connective tissue disease, HIV human immunodeficiency virus, I/HPAH idiopathic/heritable pulmonary arterial hypertension, PAH pulmonary arterial hypertension, PH pulmonary hypertension, PoPH portopulmonary hypertension

Full size image
Table 1 Demographics and disease characteristics at macitentan initiation
Full size table

Hepatic comorbidities were reported for 100% of patients with PoPH and 7.6% of patients with I/HPAH (Table 1). Cirrhosis was reported for 72.8% of patients with PoPH and 2.1% of patients with I/HPAH; liver function test abnormalities were reported in 31.6% of patients with PoPH and 3.0% of patients with I/HPAH. Of the patients with PoPH and patients with I/HPAH with baseline assessments, 10.1% (9/89) and 2.1% (16/766) had alanine aminotransferase (ALT) or aspartate aminotransferase (AST) ≥ 3 × upper limit of normal (ULN), 42.7% (38/89) and 4.6% (34/740) had total bilirubin ≥ 2 × ULN, and 3.4% (7/206) and < 0.1% (1/2498) had ALT or AST ≥ 3 × ULN and total bilirubin ≥ 2 × ULN, respectively. Of the 96 patients with PoPH in OPUS, 56 (58.3%) had a MELD/Child–Pugh score recorded during the observation period (Table S2); 33/56 and 23/56 were categorized as having mild/moderate and severe liver disease, respectively.

Treatment Patterns in PoPH

At macitentan initiation, 25.2% (52/206) of patients with PoPH were on monotherapy, 66.0% (136/206) on double combination therapy, and 8.7% (18/206) on triple combination therapy (Fig. 2). The most common double and triple combinations were macitentan with a phosphodiesterase type 5 inhibitor (PDE5i) (57.3%) and with a PDE5i and prostanoid (8.7%), respectively (Table S3). The proportion of patients with PoPH on double therapy decreased to 54.5% by 24 months, while the proportion on triple therapy increased to 23.6% (Fig. 2). Less than half (44.5%) of patients who initiated macitentan as monotherapy escalated to combination therapy by 24 months, but for the majority of those who did, the time to escalation was less than 6 months (Fig. 3).

Fig. 2
figure 2

Treatment patterns by therapy class from macitentan initiation to 24 months after. Percentages may not add to 100% because of rounding. Double therapy includes macitentan in combination with one other class of PAH therapy; triple therapy includes macitentan in combination with two other classes of PAH therapy. Classes of PAH therapy include PDE5is, prostanoids (oral, inhaled, or intravenous/subcutaneous administration), sGCSs, and investigational drugs. aIncludes 2 patients receiving > 3 classes of PAH therapy; bIncludes 1 patient receiving > 3 classes of PAH therapy; cIncludes 6 patients receiving > 3 classes of PAH therapy. I/HPAH idiopathic/heritable pulmonary arterial hypertension, PDE5i phosphodiesterase type 5 inhibitor, PoPH portopulmonary hypertension, sGCSs soluble guanylate cyclase stimulator

Full size image
Fig. 3
figure 3

Kaplan–Meier estimates of time from macitentan monotherapy to combination therapy. CL confidence limit, I/HPAH idiopathic/heritable pulmonary arterial hypertension, PoPH portopulmonary hypertension

Full size image

Hospitalizations and Survival

Median (Q1, Q3) macitentan exposure was 11.9 (3.1, 26.0) months in the PoPH population and 14.0 (5.1, 29.2) months in the I/HPAH population (Table 2). There were 119 (57.8%) patients with PoPH and 1114 (44.6%) patients with I/HPAH who experienced at least one hospitalization during the exposure period (Table 2). KM estimates (95% CL) showed that 48.6% (40.7, 56.0) and 28.6% (20.9, 36.7) of patients with PoPH and 60.3% (58.1, 62.4) and 45.3% (42.8, 47.8) of patients with I/HPAH were hospitalization free at 1 and 2 years, respectively (Fig. 4a). In total, 46 (22.3%) patients with PoPH and 365 (14.6%) patients with I/HPAH died during the macitentan exposure period (Table 2). The 1- and 2-year KM survival estimates (95% CL) were 82.2% (75.1, 87.4) and 71.7% (62.9, 78.7) and 90.5% (89.1, 91.7) and 82.1% (80.0, 83.9) for the PoPH and I/HPAH populations, respectively (Fig. 4b).

Table 2 Exposure, all-cause hospitalization, and survival
Full size table
Fig. 4
figure 4

Kaplan–Meier estimates of time from macitentan initiation to a first all-cause hospitalization and b death. For dates of first all-cause hospitalization and death, imputation of missing month and/or day was performed for 5 and 5 patients in the PoPH cohort, and 56 and 21 patients in the I/HPAH cohort, respectively. In OPUS, liver and lung transplantation was reported for 5 and 0 patients in the PoPH cohort and 0 and 9 patients in the I/HPAH cohort, respectively. CL confidence limit, I/HPAH idiopathic/heritable pulmonary arterial hypertension, PoPH portopulmonary hypertension

Full size image

PoPH Patient Liver Transplant Status

The liver transplant (LT) status for the 96 patients with PoPH in OPUS was assessed by the treating physician (Fig. S1). During the observation period, 28/96 (29.2%) patients were reported as being in medical need of LT and 65/96 (67.7%) as not being in need of LT (the remaining 3 [3.1%] patients had missing status). Of those in need, 18/28 (64.3%) were not registered on the transplant waitlist (because of PAH severity in half of the cases). Of the 10/28 (35.7%) patients who were registered on the waitlist, five were transplanted, three were still on the LT waitlist at macitentan discontinuation, one was removed because of worsening leading to ineligibility, and one died.

Safety and Tolerability

There were 49.0% (101/206) patients with PoPH and 7.8% (196/2498) patients with I/HPAH who experienced ≥ 1 HAE (Table 3). Macitentan treatment was discontinued in 54.4% (112/206) and 42.9% (1071/2498) of patients with PoPH and patients with I/HPAH; the reason for discontinuation was an HAE in 1.5% (3/206) and 0.2% (5/2498), a non-hepatic AE in 21.8% (45/206) and 17.4% (434/2498), another reason (i.e., not an AE) in 20.9% (43/206) and 18.6% (465/2498), and not reported for 10.2% (21/206) and 6.7% (167/2498) of patients in the PoPH and I/HPAH populations, respectively (Table 4). AEs leading to discontinuation were collected in OPUS only. The most common were septic shock (3.1%), dyspnea (2.1%) and edema (2.1%) (Table S4).

Table 3 Hepatic adverse events, liver enzyme elevations, and adverse events
Full size table
Table 4 Macitentan exposure and discontinuations
Full size table

The most common (≥ 5% of patients) HAEs in the PoPH population were increased bilirubin, ascites, and hepatic encephalopathy; the most common HAESIs were bilirubin increased ≥ 2 × ULN and hepatic encephalopathy (Table S5). The maximum ratio of ALT or AST to the ULN for patients in the PoPH population is shown at baseline and during follow-up in Fig. 5; individual patient data for those with AST or ALT elevations ≥ 3 × ULN reported at any time (n = 29) during the observation period are shown in Fig. S2. The equivalent depictions for bilirubin are shown in Figs. S3 and S4. These results suggest that overall, the magnitude of liver enzyme elevations at baseline remained similar over the course of the observation period. All liver enzyme elevations meeting the biochemical criteria of Hy’s law were reviewed by the ILSDRB and no causal association with macitentan treatment was concluded for events in the PoPH or I/HPAH populations.

Fig. 5
figure 5

Maximum ratio of ALT or AST to the ULN during follow-up in the PoPH population. Boxes represent the interquartile range, whiskers represent the minimum and maximum, horizontal lines indicate the median and crosses indicate the mean. The y-axis has been truncated at 0.1 and 30. Gray dots represent outlying observations (i.e., those > 1.5 above or < 1.5 below the interquartile range); blue diamonds indicate outlying observations falling outside of the displayed range. The dotted line indicates the cutoff for patients with ALT/AST ≥ 3 × ULN. ALT alanine aminotransferase, AST aspartate aminotransferase, PoPH portopulmonary hypertension, ULN upper limits of normal

Full size image

Non-hepatic AEs were collected for patients in OPUS only. During the exposure period, ≥ 1 non-hepatic AE was experienced by 84.4% (81/96) and 79.0% (979/1240) of patients with PoPH and patients with I/HPAH (Tables 3 and 5). The most common serious AEs (SAEs) were dyspnea (17.7%), increased bilirubin (14.6%), anemia, hepatic encephalopathy, and mental status changes (all 10.4%) in the PoPH population, and dyspnea (12.9%), pneumonia (7.4%), and anemia (4.7%) in the I/HPAH population (Table 5). AEs of special interest (AESIs) related to edema and decreased hemoglobin were experienced by 40 (41.7%) and 14 (14.6%) patients with PoPH, and 344 (27.7%) and 122 (9.8%) patients with I/HPAH (Table 3). Hemorrhage of esophageal varices was reported for 1 (1.0%) patient with PoPH and 2 (0.2%) patients with I/HPAH.

Table 5 Most common adverse events (in OPUS only) by Preferred Term
Full size table

Discussion

OPUS and OrPHeUS collected real-world data on patients newly initiating macitentan in the USA and together they comprise the largest PH drug-specific observational cohort to date. Despite being a rare disease, the PAH population included a significant number of patients with PoPH. There were no unexpected safety findings for macitentan use in these patients, including as part of combination therapy.

In addition to being a rare disease, PoPH is usually an exclusion criterion for many PAH clinical trials, making real-world data on these patients particularly valuable. The size of the PoPH cohort in this analysis (n = 206) is comparable to those of the REVEAL (n = 174) [2] and Spanish (n = 237) [8] PAH registries, and is more than double the number of patients enrolled in the dedicated PoPH clinical trial PORTICO (n = 85) [13]. Proportionally, patients with PoPH comprised 4.5% of the OPUS/OrPHeUS PAH population, in line with reports from several other disease registries [2, 3, 6,7,8]. Of note, patients with PoPH represented 18% of incident PAH patients enrolled in the French Pulmonary Hypertension Registry (2007–2017) [17]. This percentage may be especially high owing to the strong network between PAH experts and hepatologists in France, which may result in increased diagnostic rates and indicates that PoPH could constitute a larger proportion of PAH than usually observed.

Patients with PoPH in OPUS/OrPHeUS were younger than patients with I/HPAH and a higher proportion were male; however, the median age (58 years) and sex distribution (47.6% male) are broadly similar to what has been reported for other studies in other PoPH populations [8, 9, 13, 17,18,19]. Despite similar or better baseline values for many prognostic measures (time from PoPH/PAH diagnosis, WHO FC, 6MWD, NT-proBNP/BNP, and hemodynamic parameters including mean pulmonary arterial pressure [mPAP], PVR, mean right atrial pressure [mRAP], and cardiac index [6, 7]), the 1-year survival rate for patients with PoPH (82.2%) was lower than for the I/HPAH population (90.5%); additionally, the observed incidence rate for first all-cause hospitalization was higher in the PoPH population. These findings likely reflect the underlying liver disease, which has been found to be a stronger predictor of outcome than hemodynamics [17], and is consistent with other studies where patients with PoPH have worse outcomes than other PAH etiologies [2, 11, 18].

Survival rates for the patients with PoPH in OPUS/OrPHeUS were generally comparable relative to those reported in other real-world PoPH populations, including those in the Spanish, French, and US REVEAL registries [2, 8, 17]. Hospitalization rates are not as well reported, but were higher in OPUS/OrPHeUS than in the REVEAL registry [2]. From a prognostic standpoint, however, the patients with PoPH in this study had a lower baseline 6MWD, a lower CI, and a higher mRAP, though a similar mPAP and PVR, versus patients in other registries and in the PORTICO clinical trial [2, 8, 13, 17]. The cause for these differences is unclear, but could relate to the degree of missing data and/or differences in the way patients were recruited and how macitentan was used in the current study, where most patients with PoPH received macitentan as a part of combination therapy. This treatment pattern is in line with the 2015 ESC/ERS Guideline recommendations [20, 21] in use at the time of the study.

PoPH can develop in the setting of cirrhotic or non-cirrhotic portal hypertension [22]. Consistent with previous literature [8, 19], the majority of patients with PoPH in OPUS/OrPHeUS had underlying liver cirrhosis which is associated with worse survival [22]. Data regarding liver disease severity were only collected for patients with PoPH in OPUS and were frequently incomplete. Nevertheless, among those with data available, 41.1% presented with severe hepatic impairment (MELD ≥ 19 or Child–Pugh C). Despite this, only 28 patients with PoPH (approximately 30%) in OPUS were classified as being “in need of LT” during the observation period, and only 10 were waitlisted and five transplanted. These findings are consistent with previous reports on the large proportion of patients with PoPH in need of LT not meeting the waitlist registration criteria as a result of complications and risk factors associated with their PoPH and other comorbidities [14, 23], and the low proportion of patients with PoPH who receive transplants [24]. Overall, this highlights the need for improved therapy in patients with PoPH to stabilize their PH and achieve acceptable hemodynamic criteria for transplant.

In OPUS, the proportion of patients with PoPH who experienced an AE (84.4%) was similar to that of the I/HPAH population (79.0%). However, as expected for a population with underlying liver disease [2], the incidence rates of HAEs, HAESIs, and abnormal liver function tests in OPUS/OrPHeUS were higher for the PoPH than for the I/HPAH population. The most common HAEs in patients with PoPH were increased bilirubin, ascites, and hepatic encephalopathy. These events, as well as abnormal liver function tests in general, are commonly observed in patients with advanced liver disease [25, 26], making their attribution to macitentan or underlying hepatic condition difficult. However, during the review of hepatic safety data for all patients enrolled in OPUS/OrPHeUS, only two patients had HAESIs deemed by the ILSDRB to be possibly related to macitentan treatment [15] and neither of these was a patient with PoPH. Moreover, the ILSDRB found the clinical features of the events reported for patients with PoPH to be consistent with the natural history of the underlying PoPH disease and its associated comorbidities.

AEs of edema and anemia are of interest as these are known safety risks associated with macitentan [27]. Liver dysfunction and progressing right heart failure can also increase the risk of fluid retention [28], while anemia can be caused by bleeding esophageal varices or portal hypertensive gastropathy [29, 30]. Peripheral edema and anemia were observed at higher rates in the macitentan treatment arm versus placebo in the PORTICO study [13]. In OPUS, a higher proportion of patients with PoPH experienced an AE of edema (combined terms) than patients with I/HPAH, although presence of edema at baseline was also slightly higher for patients with PoPH. A similar trend was observed for patients treated with riociguat in the PATENT-1 study, with 27% of patients with PoPH versus 17% of the overall PAH population reporting AEs of peripheral edema [18]. The OPUS PoPH population also had a slightly higher proportion of patients reporting an AE of decreased hemoglobin/anemia (combined terms) than the I/HPAH population; however, a similar difference was also observed at baseline. Finally, esophageal varices hemorrhage was reported for only one patient with PoPH in the OPUS dataset.

Overall, the tolerability of macitentan was broadly comparable between the PoPH and I/HPAH populations. Discontinuations due to non-hepatic AEs were similar between the populations (21.8% PoPH; 17.4% I/HPAH), and although there were more HAE-related discontinuations in the PoPH (1.5%) versus I/HPAH population (0.2%), this number is low considering their underlying hepatic conditions. These data do suggest, however, that patients with PoPH may require more care and monitoring. Of the 54.4% patients with PoPH and 42.9% of patients with I/HPAH who discontinued macitentan, nearly half (20.9% and 18.6%, respectively) did so for reasons other than an AE/HAE. As has been previously discussed [15], a number of non-safety-related reasons for discontinuation of PAH therapies have been identified, including healthcare system-related factors, changes in insurance, complexity and cost of the treatment regimen, personal circumstances, and lack of disease education [31, 32]. The overall rates of discontinuation seen in these OPUS/OrPHeUS populations are consistent with other US real-world data, with PAH therapy discontinuation rates of 37.5% in the Optum Research Database [33] and of 42.3% in a meta-analysis of observational studies including PAH therapy adherence and persistence data [34].

Real-world analyses come with limitations. OPUS and OrPHeUS were drug registries set up to evaluate the safety of macitentan-treated patients. There was potential bias in the patient selection, as macitentan is not indicated for the treatment of PoPH in the USA and, given the known safety profile of ERAs, physicians might have been reluctant to initiate macitentan treatment in patients with severe liver disease. Therefore, the results may not be directly comparable with disease registries, which are more representative of the full PoPH population. Several parameters reported in our analysis, including PAH diagnosis and deaths, were investigator assessed and were not adjudicated. Furthermore, as the reason for hospitalization was not collected, the contribution of PAH and the underlying liver disease to the rate of hospitalization is not known. Finally, as a result of the observational nature of the studies, clinical assessments may not be conducted in real-world clinical practice, and thus there was a large amount of missing data, including WHO FC, 6MWD, as well as severity and outcomes of liver disease for the PoPH population. This, in addition to the difference in size between the PoPH and the I/HPAH populations and potential PoPH patient selection bias, limits formal statistical comparisons, and thus the comparisons between the populations are descriptive and of an exploratory nature only.

Conclusion

This analysis of the OPUS/OrPHeUS dataset complements the limited clinical trial data available in the PoPH population by assessing outcomes, treatment patterns, safety, and tolerability of macitentan in a real-world setting. These results highlight the need to identify and stabilize this unique subgroup of patients with PAH, and if possible, facilitate eligibility for liver transplantation. Despite the underlying differences in baseline liver disease in this population, there were no unexpected safety findings. These data add to the evidence supporting the general safety and tolerability of macitentan in patients with PoPH, including as combination therapy.