Comparing bridging drug therapy approaches prior to balloon pulmonary angioplasty in patients with chronic thromboembolic pulmonary hypertension: a prospective analysis from the national UK CURATE registry Authors: Oliver M Li, Hakim Ghani, Paula Appenzeller, Liliana Amaral-Almeida, Paul Bambrough, John E Cannon, Dolores Taboada, Katherine L Bunclark, Mark Toshner, Karen Sheares, John G Coghlan, Joanna Pepke-Zaba, Stephen P Hoole Published: 2026 ABSTRACT AbstractBackgroundPulmonary hypertension (PH) medical therapies are commonly used as a bridge to balloon pulmonary angioplasty (BPA) in patients with chronic thromboembolic PH (CTEPH) but their effectiveness remains unclear. We investigated the impact of different bridging therapies in patients undergoing BPA.MethodsThis secondary analysis of the prospective CTEPD UK Registry to Assess Treatments and OutcomEs registry evaluated the impact of different PH medical therapies on haemodynamic and physiological parameters in the UK’s national CTEPH cohort. Consecutive patients undergoing first BPA at Royal Papworth Hospital between October 2015 and October 2024 were included. Baseline parameters were assessed at CTEPH diagnosis and after initiating bridging PH medical therapy prior to BPA. Periprocedural complications at first BPA were compared between groups.ResultsHaemodynamic and functional parameters of 158 patients were analysed. Patients managed with phosphodiesterase-5 inhibitor (PDE5i) and endothelin receptor antagonist (ERA) dual therapy had higher baseline mean pulmonary artery pressure (mPAP, p=0.019), right atrial pressure (RAP, p=0.038) and pulmonary vascular resistance (PVR, p=0.004) and lower cardiac output (CO, p=0.017) and cardiac index (CI, p=0.029) versus PDE5i or riociguat monotherapy. Dual therapy was associated with a greater percentage increase in CI (40% vs 13% vs 16%, p<0.001) and CO (34% vs 8% vs 14%, p<0.001) and a greater percentage decrease in PVR (−39% vs −30% vs −25%, p=0.008) and N-terminal pro-B-type natriuretic peptide (−67% vs −21% vs −32%, p=0.004) compared with PDE5i monotherapy and riociguat monotherapy. These differences persisted after multivariable adjustment for baseline haemodynamic severity. Lung injury occurred more frequently in patients with higher PVR (>540 dynes·s·cm⁻⁵, p=0.049) and mPAP (>40 mm Hg, p=0.001) at first BPA. Periprocedural complication rates at first BPA were similar between bridging strategies.ConclusionsBridging with PH medical therapies to first BPA was associated with improved haemodynamic parameters in this real-world observational study. PDE5i and ERA dual therapy was associated with greater haemodynamic improvement than monotherapy with PDE5i or riociguat, without differences in complication rates.Keywords: Hypertension, Pulmonary; Pulmonary Embolism; Cardiac Catheterization; Endovascular Procedures WHAT IS ALREADY KNOWN ON THIS TOPIC Pulmonary hypertension (PH) medical therapies are frequently used as bridging treatment before balloon pulmonary angioplasty (BPA) in patients with inoperable chronic thromboembolic PH (CTEPH). Although PH medical therapies improve pulmonary haemodynamics, comparative data between different bridging strategies and their association with procedural complications remain limited. Pulmonary hypertension (PH) medical therapies are frequently used as bridging treatment before balloon pulmonary angioplasty (BPA) in patients with inoperable chronic thromboembolic PH (CTEPH). Although PH medical therapies improve pulmonary haemodynamics, comparative data between different bridging strategies and their association with procedural complications remain limited. WHAT THIS STUDY ADDS In a national cohort, bridging with phosphodiesterase-5 inhibitor (PDE5i) and endothelin receptor antagonist (ERA) dual therapy was associated with greater haemodynamic improvement before first BPA than riociguat and PDE5i monotherapies. Lung injury following BPA was associated with greater haemodynamic severity at the time of intervention. In a national cohort, bridging with phosphodiesterase-5 inhibitor (PDE5i) and endothelin receptor antagonist (ERA) dual therapy was associated with greater haemodynamic improvement before first BPA than riociguat and PDE5i monotherapies. Lung injury following BPA was associated with greater haemodynamic severity at the time of intervention. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY These findings suggest that haemodynamic severity at the time of BPA may be more closely associated with procedural complications than the bridging therapy used. Prospective studies are needed to clarify optimal bridging approaches to BPA. These findings suggest that haemodynamic severity at the time of BPA may be more closely associated with procedural complications than the bridging therapy used. Prospective studies are needed to clarify optimal bridging approaches to BPA. Introduction Chronic thromboembolic pulmonary hypertension (CTEPH) is a rare disease caused by the incomplete resolution of thromboemboli, resulting in the formation of fibrotic lesions within the pulmonary vasculature. These lesions drive pulmonary vascular remodelling leading to microvasculopathy and haemodynamic compromise.1 CTEPH may eventually progress to right heart failure if left untreated.2 The incidence of CTEPH in adults ranges from 0.9 to 39 cases per million, with a prevalence of 14.5 to 144 cases per million.3 Historically, 5-year survival rates were below 70%; however, advances in its management have improved this to over 90%.4 Currently, the management of CTEPH consists of a three-pronged approach, with two mechanical interventions and pharmacological treatment.5 Pulmonary endarterectomy (PEA) remains the cornerstone for patients with operable disease, with contemporary perioperative mortality rates of approximately 2% at experienced centres.6 However, surgery is not technically feasible in over one-third of patients.7 For these patients, balloon pulmonary angioplasty (BPA) may be an option, introduced more recently for those with surgically inaccessible disease or for whom the risk of surgery is adversely high. This uses balloon-mounted catheters to dilate constrictive lesions, thus restoring flow. Its refinement over the past 20 years has led to the development of a safe, efficacious procedure when performed at expert centres, with contemporary registry data reporting a procedural complication rate of 11.3% and 30-day mortality of 0%.8 9 Pharmacological agents comprise the non-interventional component of CTEPH management. Riociguat, an oral soluble guanylyl cyclase inhibitor, is the only licensed targeted medical therapy for CTEPH in many countries. However, many patients are treated off-label with other PH medical therapies due to shared pathophysiological mechanisms.10 11 Phosphodiesterase-5 inhibitor (PDE5i) and riociguat augment nitric oxide signalling, promoting smooth muscle relaxation. Endothelin receptor antagonists (ERA) can either selectively block the type A endothelin receptor or indiscriminately block type A and B endothelin receptors, abrogating vasoconstriction mediated by endothelin-1. Patients eligible for mechanical intervention who receive PH medical therapy alone are known to have worse outcomes than those undergoing mechanical intervention, perhaps owing to heterogeneous baseline characteristics including comorbidities or inaccessible lesions.12 In patients not undergoing mechanical intervention, the use of PH medical therapy is associated with improved survival.13 PH medical therapy is also used as bridging therapy prior to PEA or BPA.14 Evidence supporting the effectiveness of such a strategy prior to PEA is conflicting, and guidelines do not routinely recommend the use of riociguat or other PH medical therapy in patients referred for surgery.5 13 15 However, the use of these agents as bridging therapy prior to BPA is more commonplace—worldwide, 63% of BPA patients received medical therapy before intervention, compared with 25% of PEA patients, in line with current guidelines.5 13 While observational studies support the use of riociguat to bridge patients to BPA, data directly comparing different classes of bridging therapies are limited.16 17 Here, we conducted a secondary analysis of the prospective CTEPD UK Registry to Assess Treatments and OutcomEs (CURATE) registry to evaluate the impact of different bridging strategies on haemodynamic and physiological parameters prior to BPA and to assess periprocedural complication risk at first BPA. Methods Study design and population This secondary analysis of the prospective CURATE registry included consecutive UK-wide CTEPH patients undergoing BPA at the National CTEPH Service at Royal Papworth Hospital.18 Adults (≥18 years) with CTEPH who underwent their first BPA between 1 October 2015 and 1 October 2024 were analysed. Diagnosis of CTEPH was based on clinical guidelines at the time of inclusion and confirmed by right heart catheterisation (RHC) findings.5 19 For patients who had previously undergone PEA, the RHC diagnosing residual PH post-PEA served as the baseline haemodynamic assessment for this analysis. CTEPH patients are initially evaluated by the national CTEPH multidisciplinary team for consideration of PEA and BPA. Inoperable patients or those who refused PEA with disease amenable to BPA were accepted following established eligibility criteria.9 Practice of the UK’s National CTEPH Service and the established BPA protocol, with RHC performed at the beginning of each procedure as standard, have previously been described.9 The choice of PH medical therapy prior to BPA was at the discretion of the referring PH centre.5 Some patients underwent a second RHC, after having started bridging therapy and prior to their first BPA, if it was clinically indicated. Patients who received PH medical therapy prior to the diagnostic RHC were excluded from the study. Patients whose bridging therapy was used by fewer than eight patients in total were excluded due to small numbers. The primary outcome was change in haemodynamic parameters between CTEPH diagnosis and first BPA, after receiving bridging PH medical therapy. To address the potential confounding influence of forthcoming BPA on haemodynamic measurements measured immediately prior to first BPA, the agreement between haemodynamics obtained under identical conditions to those measured during the first RHC, during a subsequent ‘on treatment’ second RHC remote from the BPA, was also evaluated for patients with available data at these timepoints. Secondary outcomes were changes in functional capacity, N-terminal pro-B-type natriuretic peptide (NT-proBNP) and Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR) score after bridging PH medical therapy, prior to first BPA.20 Data at diagnosis and prior to mechanical intervention Data were extracted from the prospective CURATE registry. Patient demographics, WHO functional class (WHO-FC), haemodynamic parameters obtained via RHC, CAMPHOR score, 6 min walk distance (6MWD) and NT-proBNP were recorded at diagnosis, second RHC (after bridging therapy initiation) and immediately prior to first BPA. In cases where 6MWD, NT-proBNP or CAMPHOR were not recorded immediately prior to BPA but were present after having started bridging therapy, the latest available values were collected. Bridging therapy was defined as the latest PH medical therapy administered prior to first BPA. Absolute and percentage changes in parameters were investigated to assess the effects of bridging therapies while accounting for individual baseline characteristics. NT-proBNP concentration was transformed to logarithmic to assess absolute change from diagnosis to first BPA. Complications Complications were defined as thoracic, access site and other. Thoracic complications included haemoptysis, pulmonary artery perforation and dissection and lung injury, which itself comprised radiography-confirmed lung reperfusion injury and parenchymal haemorrhage. Access site injury included haematoma formation, arteriovenous fistula development or rebleeding. Statistical analysis Statistical analyses were performed using RStudio (V.2024.12.0+467). Variables are presented as the median with IQRs. Differences between drug groups with respect to time from diagnosis to intervention, baseline data and absolute and percentage changes in parameters were analysed using the Kruskal-Wallis test. Dunn’s tests were performed for post-hoc pairwise comparisons. Multivariable linear regression with estimated marginal means was performed for change in mean pulmonary artery pressure (mPAP), cardiac output (CO) and pulmonary vascular resistance (PVR), adjusting for baseline mPAP, PVR, CO, cardiac index (CI) and right atrial pressure (RAP), to estimate the independent contribution of bridging therapy strategy to haemodynamic improvement. Covariates were chosen as routinely measured parameters at RHC as markers of severe disease, and variance inflation factors were verified to ensure no significant multicollinearity. P values were adjusted with the Benjamini-Hochberg method. Analysis of differences in baseline versus first BPA data were performed using the Wilcoxon signed-rank test. Categorical variables, including complication rates, were compared using Fisher’s exact test, both across bridging therapy groups and following stratification by baseline mPAP and PVR. A sensitivity analysis was performed for the association between haemodynamic severity and lung injury across a range of mPAP and PVR thresholds. Spearman’s correlation coefficients assessed temporal associations between PH medical therapy and haemodynamic and physiological parameters, and the change in parameters from diagnosis to second RHC or first BPA. Intraclass correlation coefficients (ICC) were used to assess agreement between parameters measured at second RHC and first BPA. ICC estimates were calculated based on a two-way random effects absolute agreement model. Bland-Altman analysis was further used to assess agreement with bias, SD and 95% limits of agreement reported. Significance was defined as p value <0.05. Patient and public involvement There was no patient involvement in this research. Results Patient characteristics 195 patients underwent BPA and were included in this study (figure 1). Of these, 97% (n=189) received bridging therapy before first BPA. 24 had initiated treatment prior to diagnostic RHC, including 10 patients post-PEA, and were excluded. 158 patients who commenced bridging therapy after diagnostic RHC were treated with a PH therapy used by at least eight patients. CONSORT diagram. Patient flow through inclusion and exclusion criteria. BPA, balloon pulmonary angioplasty; CONSORT, Consolidated Standards of Reporting Trials; PH, pulmonary hypertension; RHC, right heart catheterisation. Bridging therapy to BPA The 158 patients had a median age of 66 years, median time from diagnosis to first BPA of 352 days and 54% were male. The bridging groups were PDE5i monotherapy (n=42), PDE5i and ERA dual therapy (n=67) and riociguat monotherapy (n=49). The time from diagnostic RHC to first BPA did not differ significantly between the bridging groups, with median durations of 324, 352 and 362 days for PDE5i monotherapy, PDE5i and ERA dual therapy and riociguat monotherapy groups, respectively. Table 1 illustrates baseline data comparing the three main bridging therapy approaches. Table 1Baseline data at diagnosis for patients with chronic thromboembolic pulmonary hypertension receiving the three main bridging therapy approaches prior to first BPABridging therapy prior to BPAP valuePDE5iPDE5i+ERARiociguatPatients, n (%)42 (27)67 (42)49 (31) Age at first BPA, years65 (59, 72)67 (57, 72)67 (59, 73)0.570Male sex, %6951470.078BMI (kg/m2)25.1 (23.2, 27.9)26 (23.1, 29)26.1 (23, 30.7)0.370Time from RHC to first BPA, days324 (233, 604)352 (270, 862)362 (212, 510)0.714WHO-FC, % <0.001*I–II431220 III578280 IV060 Baseline RHC Data mPAP, mm Hg46 (36, 50)50 (43, 56)45 (40, 51)0.019*RAP, mm Hg7 (5, 9)10 (6, 13)8 (6, 11)0.038*Cardiac output, L/min4.3 (3.7, 5.2)3.5 (2.8, 4.5)3.9 (3.3, 4.9)0.017*Cardiac index, L/(min·m²)2.3 (2.0, 2.7)1.9 (1.7, 2.4)2.2 (1.9, 2.6)0.029*PVR, dynes·s·cm⁻⁵680 (487, 942)893 (689, 1144)716 (514, 892)0.004*Baseline physiological data 6MWD (m)423 (275, 460)294 (227, 376)336 (218, 416)0.031*NT-proBNP (ng/L)536 (246, 2045)2126 (488, 4041)651 (262, 1410)0.006*CAMPHOR Symptom, score10.5 (6.5, 17)13 (8, 14)16 (10, 17)0.578Activity, score9.5 (5.5, 16.5)10 (7, 14)12.5 (8.5, 14.5)0.749QoL, score10 (6, 15.5)7 (5, 14)11.5 (5.5, 16.5)0.564Variables are expressed as median (first, third quartile) or numbers (%).BMI, body mass index; BPA, balloon pulmonary angioplasty; CAMPHOR, Cambridge Pulmonary Hypertension Outcome Review; ERA, endothelin receptor antagonist; mPAP, mean pulmonary artery pressure; 6MWD, 6 min walk distance; NT-proBNP, N-terminal pro-B-type natriuretic peptide; PDE5i, phosphodiesterase type-5 inhibitor; PVR, pulmonary vascular resistance; QoL, quality of life; RAP, right atrial pressure; RHC, right heart catheterisation confirming CTEPH diagnosis; WHO-FC, WHO functional class. Baseline data at diagnosis for patients with chronic thromboembolic pulmonary hypertension receiving the three main bridging therapy approaches prior to first BPA Variables are expressed as median (first, third quartile) or numbers (%). BMI, body mass index; BPA, balloon pulmonary angioplasty; CAMPHOR, Cambridge Pulmonary Hypertension Outcome Review; ERA, endothelin receptor antagonist; mPAP, mean pulmonary artery pressure; 6MWD, 6 min walk distance; NT-proBNP, N-terminal pro-B-type natriuretic peptide; PDE5i, phosphodiesterase type-5 inhibitor; PVR, pulmonary vascular resistance; QoL, quality of life; RAP, right atrial pressure; RHC, right heart catheterisation confirming CTEPH diagnosis; WHO-FC, WHO functional class. There were no significant differences in patient demographics between the three main bridging therapy approaches prior to BPA. A larger proportion of patients prescribed PDE5i and ERA dual therapy were classified as WHO-FC III and IV (p<0.001). This was reflected in baseline haemodynamics at diagnosis, with patients who received PDE5i and ERA dual therapy having higher mPAP (p=0.019), PVR (p=0.004) and RAP (p=0.038), with lower CO (p=0.017) and CI (p=0.029), versus those given either PDE5i or riociguat bridging monotherapy (table 1). Poorer baseline physiological parameters were observed in patients who were prescribed PDE5i and ERA dual therapy, with lower 6MWD (p=0.031) and higher NT-proBNP (p=0.006) than patients receiving either PDE5i monotherapy or riociguat monotherapy. All three bridging approaches to BPA were associated with decreased mPAP and PVR and increased CO and CI from diagnostic RHC to immediately before first BPA (online supplemental table 1). PDE5i and ERA dual therapy and riociguat monotherapy were also associated with decreased RAP and increased 6MWD from diagnosis to first BPA, and PDE5i and ERA dual therapy and PDE5i monotherapy were associated with decreased log-NT-proBNP from diagnosis to first BPA. PDE5i and ERA dual therapy was associated with a greater absolute decrease in PVR (p<0.001), RAP (p=0.010) and log-NT-proBNP (p=0.004) and a greater absolute increase in CO (p=0.002) and CI (p=0.002), compared with PDE5i or riociguat monotherapy (table 2). Table 2Absolute and percentage change in haemodynamic and physiological parameters measured at diagnosis and at first BPABridging therapy prior to BPAP valuePDE5iPDE5i+ERARiociguatPatients, n (%)42 (27)67 (42)49 (31) Absolute change mPAP, mm Hg−5 (−11, 0)−7 (−12, −1)−5 (−9, 1)0.234RAP, mm Hg0 (−3, 2)−2 (−6, −1)−2 (−3, 1)0.010*Cardiac output, L/min0.4 (-0.2, 1.3)1.1 (0.4, 1.8)0.5 (0, 0.9)0.002*Cardiac index, L/(min·m²)0.3 (-0.1, 0.7)0.7 (0.3, 1)0.3 (0, 0.5)0.002*PVR, dynes·s·cm⁻⁵−170 (−303, −64)−336 (−516, −132)−137 (−324, −13)<0.001*6MWD, m−1 (−51, 68)32 (2, 86)43 (2, 68)0.159log-NT-proBNP−0.23 (−0.66, 0.07)−1.10 (−1.83, −0.55)−0.38 (−0.83, 0.15)0.004*CAMPHOR symptom, score−1.5 (−3.5, 1)−2 (−6, 0)−2.5 (−6, 1.5)0.742CAMPHOR activity, score−0.5 (−3, 2)−1 (−4, 3)0 (−2.5, 0.5)0.887CAMPHOR QoL, score0 (−3, 3)−1 (−3, 1)−0.5 (−3.5, 2.5)0.896Percentage change mPAP, %−12 (−22, 0)−15 (−23, −3)−11 (−20, 2)0.542RAP, %0 (−34, 53)−33 (−58, −8)−19 (−33, 19)0.005*Cardiac output, %8 (-5, 24)34 (12, 58)14 (0, 26)<0.001*Cardiac index, %13 (−5, 30)40 (15, 60)16 (1, 29)<0.001*PVR, %−30 (−41, −10)−39 (−51, −20)−25 (−40, −3)0.008*6MWD, %0 (-12, 11)13 (1, 39)12 (1, 29)0.084NT-proBNP, %−21 (−48, 7)−67 (−84, −42)−32 (−56, 17)0.004*CAMPHOR symptom, %−17 (−36, 14)−22 (−47, 0)−18 (−33, 3)0.342CAMPHOR activity, %−7 (−21, 25)−20 (−45, 20)0 (−21, 2)0.538CAMPHOR QoL, %−4 (−27, 51)−7 (−43, 36)−10 (−24, 21)0.675Variables are expressed as median (first, third quartile) or numbers (%). Absolute and percentage change in haemodynamic and physiological parameters measured at diagnosis and at first BPA Variables are expressed as median (first, third quartile) or numbers (%). Patients taking PDE5i and ERA dual therapy had poorer haemodynamic and physiological parameters at baseline versus either riociguat or PDE5i monotherapy. Thus, percentage change in parameters was analysed to account for the greater potential for absolute change with higher baseline values. PDE5i and ERA dual therapy was associated with a greater percentage decrease in PVR (p=0.008), RAP (p=0.005) and NT-proBNP (p=0.004) and a greater percentage increase in CO (p<0.001) and CI (p<0.001) (figure 2, table 2). Percentage change in haemodynamic parameters from diagnostic right heart catheterisation to immediately before first BPA. (A) mPAP, (B) CO and (C) PVR. BPA, balloon pulmonary angioplasty; CO, cardiac output; ERA, endothelin receptor antagonist; mPAP, mean pulmonary artery pressure; PDE5i, phosphodiesterase type-5 inhibitor; PVR, pulmonary vascular resistance; RIO, riociguat. After multivariable adjustment for baseline haemodynamic severity, PDE5i and ERA dual therapy remained independently associated with greater reduction in PVR compared with riociguat monotherapy (adjusted difference 97 dynes·s·cm⁻⁵, p=0.028) and PDE5i monotherapy (adjusted difference 85 dynes·s·cm⁻⁵, p=0.042). Dual therapy was also associated with greater increase in CO versus riociguat monotherapy (0.53 L/min, p=0.007) and PDE5i monotherapy (0.51 L/min, p=0.008). No significant differences were observed for mPAP following adjustment (figure 3). Baseline-adjusted pairwise comparison of change in haemodynamic parameters from diagnostic right heart catheterisation to immediately before first BPA. Covariates included baseline mPAP, RAP, CO, CI and PVR. Horizontal bars represent unadjusted 95% CIs. (A) mPAP, (B) CO and (C) PVR. BPA, balloon pulmonary angioplasty; CI, cardiac index; CO, cardiac output; ERA, endothelin receptor antagonist; mPAP, mean pulmonary artery pressure; PDE5i, phosphodiesterase type-5 inhibitor; PVR, pulmonary vascular resistance; RAP, right atrial pressure; RIO, riociguat. Agreement between second RHC and first BPA To determine whether haemodynamics measured just prior to first BPA were comparable to those obtained at second RHC under identical conditions to the first RHC, an agreement analysis was performed. No time-dependency effects of any drug group on absolute change in parameters were observed (online supplemental table 2). 37 patients had a second RHC between diagnosis and first BPA. For these patients, comparing parameters measured at second RHC to first BPA yielded no significant differences in mPAP (p=0.659), CO (p=0.166) and PVR (p=0.331). Changes from diagnosis to second RHC were significantly correlated with changes from diagnosis to first BPA for mPAP (r=0.460, p=0.004), CO (r=0.651, p<0.001) and PVR (r=0.748, p<0.001). Assessing agreement between parameters measured at second RHC versus first BPA yielded ICCs of 0.679 (range: 0.457–0.821) for mPAP (p<0.001), 0.671 (range: 0.451–0.815) for CO (p<0.001) and 0.829 (range: 0.693–0.909) for PVR (p<0.001), and Bland-Altman analysis demonstrated no significant bias between parameters measured at second RHC and first BPA (online supplemental figure 1). Periprocedural complications When stratified according to PVR and mPAP at first BPA, overall complication rates did not differ significantly between haemodynamic strata, which were selected post-hoc as clinically plausible thresholds (table 3). However, lung injury occurred more frequently in patients with higher PVR (>540 dynes·s·cm⁻⁵ vs ≤540 dynes·s·cm⁻⁵: 14.1% vs 4.6%, p=0.049) and higher mPAP (>40 mm Hg vs ≤40 mm Hg: 15.7% vs 1.3%, p=0.001). The direction of this association was consistent across a range of PVR thresholds (520–560 dynes·s·cm⁻⁵) and mPAP thresholds (35–45 mm Hg), although statistical significance varied with threshold selection, likely reflecting the small number of lung injury events (online supplemental table 3). Other procedural complications including pulmonary artery perforation, dissection, haemoptysis and access site complications were not associated with haemodynamic severity at first BPA. These findings should be interpreted as exploratory. Table 3Periprocedural complications of first BPA stratified by PVR and mPAPPVR ≤540 dynes·s·cm-5PVR >540dynes·s·cm-5P valuemPAP ≤40mm HgmPAP >40 mm HgP valuePatients, n (%)87 (55.1)71 (44.9) 75 (47.5)83 (52.5) Complicated procedures, n (% of procedures)17 (19.5)20 (28.2)0.25814 (18.7)23 (27.7)0.194Thoracic complications, n (% of procedures)15 (17.2)19 (26.8)0.17511 (14.7)23 (27.7)0.054Lung injury, n (% of procedures)4 (4.6)10 (14.1)0.049*1 (1.3)13 (15.7)0.001*Haemoptysis, n (% of procedures)9 (10.3)12 (16.9)0.2478 (10.7)13 (15.7)0.482Pulmonary artery perforation, n (% of procedures)2 (2.3)4 (5.6)0.4102 (2.7)4 (4.8)0.684Pulmonary artery dissection, n (% of procedures)3 (3.4)4 (5.6)0.7022 (2.7)5 (6.0)0.447Access site injury, n (% of procedures)4 (4.6)5 (7.0)0.7324 (5.3)5 (6.0)1BPA, balloon pulmonary angioplasty; mPAP, mean pulmonary artery pressure; PVR, pulmonary vascular resistance. Periprocedural complications of first BPA stratified by PVR and mPAP BPA, balloon pulmonary angioplasty; mPAP, mean pulmonary artery pressure; PVR, pulmonary vascular resistance. When analysed according to bridging therapy approaches, overall complication rates were similar between groups (23.8% PDE5i, 23.9% dual therapy, 22.4% riociguat, p=1.00, table 4). Rates of lung injury were numerically higher in the dual therapy bridged group (13.4%), although this difference was not statistically significant. Table 4Periprocedural complications of first BPA stratified by bridging therapy approachBridging therapy prior to BPAP valuePDE5iPDE5i+ERARiociguatPatients, n (%)42 (27)67 (42)49 (31) Complicated procedures, n (% of procedures)10 (23.8)16 (23.9)11 (22.4)1Thoracic complications, n (% of procedures)9 (21.4)16 (23.9)9 (18.4)0.754Lung injury, n (% of procedures)3 (7.1)9 (13.4)2 (4.1)0.235Haemoptysis, n (% of procedures)6 (14.3)10 (14.9)5 (10.2)0.798Pulmonary artery perforation, n (% of procedures)3 (7.1)1 (1.5)2 (4.1)0.377Pulmonary artery dissection, n (% of procedures)2 (4.8)2 (3.0)3 (6.1)0.710Access site injury, n (% of procedures)3 (7.1)4 (6.0)2 (4.1)0.909BPA, balloon pulmonary angioplasty; PDE5i, phosphodiesterase type-5 inhibitor; ERA, endothelin receptor antagonist. Periprocedural complications of first BPA stratified by bridging therapy approach BPA, balloon pulmonary angioplasty; PDE5i, phosphodiesterase type-5 inhibitor; ERA, endothelin receptor antagonist. Discussion Here, we directly compared the effect of different classes of PH medical therapies as a bridge to first BPA on haemodynamic and physiological parameters. Our analysis of this large cohort of consecutive patients with CTEPH found that all bridging therapies were associated with improvements in haemodynamic parameters. PDE5i and ERA dual therapy was associated with greater improvements in haemodynamic parameters between diagnosis and first BPA compared with either PDE5i or riociguat monotherapy. These differences persisted after adjustment for differences in baseline severity, suggesting an additive effect of combination therapies and supporting a potential haemodynamic contribution of ERA in inoperable CTEPH. Bridging therapy to first BPA Most (97%) of the originally included 195 patients received PH medical therapies as a bridge to BPA, reflecting the current international consensus.5 21 Although riociguat is recommended as first-line therapy prior to BPA due to its more robust clinical evidence compared with off-label PH-targeted drugs, PDE5i and ERA combination therapy was more commonly used in patients with greater baseline haemodynamic severity.5 In our study, riociguat monotherapy was associated with improvement in haemodynamic data including mPAP, CO and PVR, consistent with existing evidence.16 22 However, it was not associated with significantly greater improvements in haemodynamic parameters than PDE5i monotherapy. This is consistent with prior findings demonstrating haemodynamic benefit with PDE5i, supporting their continued use as an alternative to riociguat.23 PDE5i and ERA dual therapy treated patients experienced greater improvements in haemodynamic parameters than either monotherapy group, suggesting an additive effect of ERAs. This reflects the haemodynamic findings of the MERIT-1 trial investigating the use of macitentan in patients with inoperable CTEPH, as well as the BENEFiT trial of bosentan.24 25 Studies, including an ancillary follow-up of the RACE trial, have reported additional haemodynamic improvement with the sequential use of medical therapy and BPA, suggesting that the strategy of targeting microvasculopathy with pharmacological agents and subsegmental disease with BPA might provide additive effects in the treatment of patients with inoperable CTEPH.16 17 Although all bridging strategies were associated with reduced mPAP and PVR from baseline to first BPA, lung injury remained associated with greater haemodynamic burden at the time of intervention. Patients with PVR >540 dynes·s·cm⁻⁵ and mPAP >40 mm Hg experienced higher rates of lung injury, consistent with prior studies linking elevated mPAP and PVR to increased thoracic complications of BPA, although with slightly different thresholds of 6.6 Wood’s units (528 dynes·s·cm⁻⁵) for PVR and 45 mm Hg for mPAP.26,28 While registry data emphasise operator experience as a major determinant of BPA complication rate, our findings indicate that haemodynamic burden also contributes to procedural risk.29 30 The higher rate of lung injury in the dual therapy bridged group may reflect their more severe baseline haemodynamics rather than treatment effect and is clinically important, although not statistically significant. In this cohort, greater haemodynamic improvement was observed with PDE5i and ERA dual therapy compared with monotherapy strategies, while haemodynamic responses with riociguat monotherapy appeared similar to those with PDE5i monotherapy. Haemodynamic parameters are used as markers of CTEPH severity and are widely used in clinical trials as markers to assess treatment effect.8 9 16 17 22 25 Results from the New International CTEPH Registry suggest that patients with CTEPH receiving riociguat have superior survival compared with patients taking other PH medical therapies or no medical therapy.31 This includes those undergoing BPA, with patients who took riociguat at any point having superior 3-year survival rates compared with those who never received riociguat. Our findings, which relate to short-term haemodynamic responses, do not preclude the survival benefits reported with riociguat in other studies. Endothelin-1 is a potent vasoconstrictor, thus, antagonising its activity alongside sustaining the nitric oxide signalling pathway might underlie the greater haemodynamic improvements observed in patients bridged with dual therapy. Though PDE5i and riociguat have a shared action on the nitric oxide signalling pathway via increasing cGMP within vascular smooth muscle cells, riociguat additionally sensitises soluble guanylyl cyclase to endogenous nitric oxide.32 This, or a currently uncharacterised mechanism of action of riociguat, might underlie improved survival in riociguat-treated patients independent of its haemodynamic effects, though more work is required to investigate the underlying cause of these differences. Limitations This secondary analysis of a prospective observational study has several limitations. While the national CTEPH multidisciplinary team in the UK considered patients suitable for BPA to have purely CTEPH, patients with a phenotype of group 1 PH may have been included. Patients were not randomised against matched comparison groups and those with worse haemodynamic data at baseline preferentially received PDE5i and ERA dual therapy, with greater potential for improvement. In mitigation, investigation of percentage changes and multivariable regression adjusting for baseline haemodynamic severity were performed, and our study represents a large national cohort reflecting real-world practices. Data on diuretic use and renal function were not collected, which may have confounded NT-proBNP and haemodynamic measurements. Adverse effects and tolerability of PH medical therapies were not collected in the registry, precluding comparison of side effect profiles. Specific PDE5i and ERA regimens were not collected. Some data were unavailable due to the nature of the registry database. The conditions under which RHC was performed may also be susceptible to confounding biases—while performed at specialist PH centres across the UK, different operators performed diagnostic RHC with different access sites being used. Haemodynamic data obtained at second RHC, used for agreement analysis, were available for a limited number of patients (n=37) in whom repeat RHC was clinically indicated, potentially biasing these interval RHC data towards worse haemodynamic values. To minimise this potential selection bias, haemodynamic parameters at first BPA were used for the main analysis, which did not differ from second RHC measurements in patients with both. The haemodynamic thresholds used to stratify complication rates were selected post-hoc. However, sensitivity analysis demonstrated that the association between haemodynamic severity and lung injury was consistent across a range of clinically plausible thresholds. This analysis did not compare bridging approaches to PEA as this approach is not uniformly prescribed; however, long-term outcomes of bridging therapy in PEA-treated patients have been described separately by our group.18 Conclusions In this real-world observational study, bridging therapy prior to BPA was associated with improvements in haemodynamic and physiological parameters, with greater changes observed in patients receiving PDE5i and ERA dual therapy. While haemodynamic severity at first BPA was associated with rates of lung injury, there were no differences in periprocedural complication rates between bridging strategies. Further prospective, randomised studies are required to validate these observations. Supplementary material Acknowledgements The authors thank the members of the UK Pulmonary Hypertension Network for their contributions to patient care and data collection. Participating centres included Royal Papworth Hospital NHS Trust, Royal Hallamshire Hospital, Royal Free Hospital, Royal Brompton Hospital, Hammersmith Hospital, Freeman Hospital and Golden Jubilee National Hospital. Footnotes Funding: The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors. Provenance and peer review: Not commissioned; externally peer reviewed. Patient consent for publication: Not applicable. Ethics approval: This study involves human participants. This was a secondary analysis of a registry and constituted a service evaluation, and thus did not require ethical approval from an ethics committee and active consent from patients. We have used the National Data Opt-outs service through MESH and have excluded patients who have opted out for their data to be used. Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information. References Associated Data Supplementary Materials Data Availability Statement All data relevant to the study are included in the article or uploaded as supplementary information.