Balloon pulmonary angioplasty vs riociguat in patients with inoperable chronic thromboembolic pulmonary hypertension: A systematic review and meta‐analysis

Abstract Backgrounds No previous meta‐analyses have compared the efficacy and safety of BPA with riociguat therapy in inoperable CTEPH patients. Methods Relevant published studies were searched in the PubMed, Embase and ClinicalTrial.gov databases. Results Twenty‐three clinical trials including 1454 patients (631 underwent BPA; 823 underwent riociguat therapy) were analyzed. BPA was associated with a greater improvement in RAP (mean difference (MD) = −3.53 mmHg, 95% CI: [−4.85, −2.21] vs MD = −1.05 mmHg, 95% CI: [−1.82, −0.29]); mPAP (MD = −15.02 mmHg, 95% CI: [−17.32, −12.71] vs MD = −4.19 mmHg, 95% CI: [−5.58, −2.80]); PVR (standard MD = −1.32 woods, 95% CI: [−1.57, −1.08] vs standard MD = −0.65 woods, 95% CI: [−0.79, −0.50]); NYHA functional class (RR = 6.78, 95% CI: [3.14, 14.64] vs RR = 1.49, 95% CI: [1.07, 2.07]); and 6MWD (MD = 71.66 m, 95% CI: [58.34, 84.99] vs MD = 45.25 m, 95% CI: [36.51, 53.99]) than riociguat treatment. However, the increase in CO was greater with riociguat (MD = 0.78 L/min, 95% CI: [0.61, 0.96]) than with BPA (MD = 0.33 L/min, 95% CI: [0.06, 0.59]). No significant difference in cardiac index (CI) was found between BPA (MD = 0.40 L/min/m2, 95% CI: [0.21, 0.58]) and riociguat (MD = 0.40 L/min/m2, 95% CI: [0.26, 0.54]). The most common complications of BPA were pulmonary injury (0.3%‐5.6%) and pulmonary edema (0.8%‐28.6%). The most common adverse events of riociguat were headache, dizziness, hypotension and nasopharyngitis. Conclusions Our meta‐analysis indicates that BPA might be associated with greater improvements in exercise tolerance and pulmonary hemodynamics except for cardiac output and cardiac index than riociguat therapy. However, both of them were well tolerated.

output and cardiac index than riociguat therapy. However, both of them were well tolerated.

K E Y W O R D S
balloon pulmonary angioplasty, chronic thromboembolic pulmonary hypertension, efficacy, riociguat, safety

| INTRODUCTION
Chronic thromboembolic pulmonary hypertension (CTEPH) is a rare, potentially life-threatening disease of the pulmonary vasculature. 1 CTEPH has been proposed to develop when a pulmonary embolism does not resolve and transforms into fibrous tissue that occludes the pulmonary artery. [2][3][4][5] Patients with untreated CTEPH are at high risk of progressive pulmonary hypertension, right heart failure and death. 5 For patients with CTEPH, the gold standard treatment is potentially curative pulmonary endarterectomy (PEA). 1,6 For patients ineligible for PEA or those who have recurrent or persistent pulmonary hypertension after surgery, drug treatment with riociguat is beneficial and recommended by the guideline. 7 At present, riociguat is the only medicine approved for the treatment of both PAH and inoperable, persistent or recurrent CTEPH. 8,9 Balloon pulmonary angioplasty (BPA) is an emerging option that promises hemodynamic and functional benefits for inoperable patients. 10 BPA relies on the use of telescoping catheters placed in a central vein, through which wires and balloons are guided to mechanically disrupt chronic clot material and relieve pulmonary vascular obstruction. Recently, additional groups have reported that BPA improved symptoms and hemodynamic parameters in patients with peripheral-type CTEPH. 11 In addition, repeated PEA is not a feasible way due to high perioperative risk. Therefore, BPA appears to be an alternative and less invasive technique and may address some of the limitations of PEA. [12][13][14] In the past 5 years, BPA has been widely performed worldwide. BPA currently carries a class IIb recommendation for the treatment of inoperable CTEPH according to the most recent European guidelines 15 and may be considered in patients who are technically inoperable or carry an unfavorable risk during surgery.
One meta-analysis compared the efficacy of medical therapy, which included pulmonary vasodilators, against BPA in patients with inoperable CTEPH. 16 In this meta-analysis, six studies on BPA and 15 studies on medical therapy, including various pulmonary vasodilators with heterogeneity, were pooled. The conclusion showed high-quality evidence on the use of pulmonary vasodilators while only moderate-quality evidence on BPA in improving both hemodynamics and exercise capacity. However, no previous meta-analyses have evaluated and directly compared the efficacy and safety of BPA with those of riociguat therapy. Additional clinical trials on BPA have recently been published. Therefore, the aim of this new meta-analysis was to evaluate and compare the efficacy and safety of BPA with riociguat therapy, including 11 recent clinical trials on BPA, in inoperable CTEPH patients.

| Search strategy
We performed a review of the literature and a meta-analysis of studies that compared the efficacy and safety of BPA against those of riociguat therapy in inoperable CTEPH patients; hemodynamic parameters of right heart catheterization (RHC), 6-minute walking distance (6MWD) and New York Heart Association (NYHA) functional class were evaluated. Relevant studies were identified by searching the PubMed and Embase databases and ClinicalTrial.gov using the following search terms: ("chronic thromboembolic pulmonary hypertension" OR "chronic pulmonary embolism") AND ("percutaneous transluminal pulmonary angioplasty" OR "BPA" OR "balloon pulmonary angioplasty" OR "riociguat"). In addition, the references of all retrieved literatures were reviewed for further identification of potentially relevant studies. This meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement. 17 All studies included in this meta-analysis were published from January 2001 to September 2017.

| Selection criteria
Studies were identified for inclusion by screening "titles/abstracts and full texts" if they met all of the following criteria: (i) the subjects were diagnosed with inoperable CTEPH by demonstration of organized pulmonary thromboembolism using contrast-enhanced lung computed tomography, perfusion lung scintigraphy, and pulmonary angiography, excluding collagen vascular disease, pulmonary disease, left heart abnormality, and other systemic diseases by blood tests, pulmonary function tests, and echocardiography. Patients with residual or recurrent pulmonary hypertension after pulmonary endarterectomy were enrolled. (ii) Group 1 included patients who underwent BPA, and group 2 included patients who were administered riociguat as the first prescribed treatment or add-on medication; (iii) all included studies were retrospective or prospective clinical studies; (iv) the primary exposure investigated had to include hemodynamic parameters of RHC, NYHA functional assessments, 6MWD and brain natriuretic peptide (BNP), and complications after BPA were also assessed; Pulmonary edema was defined as X-ray opacity in the lung segment treated with BPA on the day or next day of the procedure. 18 Pulmonary vascular injury was commonly caused by wire perfusion, resulting in parenchymal bleeding with or without hemoptysis. 19 and (v) all patients were diagnosed as inoperable by experienced surgeons due to the location of thrombi and surgical accessibility, age, and comorbidities.

| Data extraction
The characteristics of the included clinical trials were independently extracted by two authors (W.W. and W.L.).

| Assessment of study quality
The Newcastle-Ottawa Scale (NOS) 20 for assessing the quality of nonrandomized studies in meta-analyses was used to assess the risk of bias, which consisted of the following three aspects: selection, comparability, and outcome. All included studies were prospective or retrospective single-arm studies. Quality assessment was independently conducted by two authors (W.W and W.L.); their results were compared, and if a consensus could not be reached, a third person (S.Z.) intervened.

| Statistical analyses
Pooled treatment effects, including NYHA functional class, 6MWD, BNP and hemodynamic parameters of RHC, were estimated using STATA software (Version 12). A P-value less than .05 for any statistical test was regarded as statistically significant. For continuous data, the inverse variance statistical method was used to measure the effect of the mean difference in each outcome. Categorical variables were compared using Chi-squared or Fisher's exact tests. We used Cochran's χ 2 -based Q test and the I-squared test to assess interstudy heterogeneity. 21 If there was no significant heterogeneity (defined as P > .10 or I 2 < 50%), the pooled outcomes were determined with the fixed effects model (Mantel-Haenszel). Conversely, the random effects model (DerSimonian and Laird) was used when significant heterogeneity was found. 22 In addition, a sensitivity analysis was performed to determine the effects of individual trials on the overall pooled results.
Furthermore, potential publication bias was considered using Begg's rank correlation test 23 and Egger's linear regression test. 24 Funnel plots were employed to assess potential publication bias.

| Literature search
A total of 655 citations were identified, and 124 duplicates were removed, leaving 531 studies for screening. After reviewing the titles and abstracts, 440 publications were excluded because they did not report on BPA performed in humans or were not clinical trials about riociguat therapy. The 34 studies on inoperable CTEPH underwent full-text review; seven studies were excluded because the articles were conference abstracts or editorials, and two articles were excluded because they were related to 2D-perfusion angiography and a new index. In addition, two case reports were excluded. After reviewing the remaining studies, 17 14,18,19,[25][26][27][28][29][30][31][32][33][34][35][36][37][38] studies of BPA and 6 7,39-43 studies of riociguat therapy met the inclusion criteria and were included in the pooled analysis ( Figure 1).

| Methodological quality assessment
The quality of each included study was assessed using NOS. Thirteen studies received eight stars, nine studies 7,30,35,37,39-43 received nine stars, and one study 31 received seven stars.

| Hemodynamic parameters
The random effects model was utilized for the analysis. Regarding

| Functional capacity
BPA treatment significantly improved the NYHA class in the inoperable CTEPH patients (RR = 6.8, 95% CI: [3.14, 14.64], P = .000) ( Figure 3A). The random effects model was used in the analysis of NYHA across the studies because it was statistically heterogeneous (I 2 = 68.1% in BPA group and I 2 = 88.1% in riociguat group

| Complications
Complication rates were reported for the 17 studies. After BPA, the most common symptom among the CTEPH patients was hemoptysis, which is usually caused by wire perforation. 10 Moreover, the most common complications were pulmonary edema and pulmonary injury. Among the included studies that reported these complications, the reperfusion pulmonary edema rate ranged from 0.8% to 28.6%, and the pulmonary injury rate ranged from 0.3% to 5.6%. Only one study reported that one participant had died due to pulmonary artery wiring perforation after the procedure. In terms of riociguat treatment among the inoperable CTEPH patients, the most common adverse events observed within the six included studies were dyspepsia, headache, dizziness, hypotension and nasopharyngitis, with an incidence rate of less than 30%.
Overall, the BPA and riociguat treatments were both well tolerated.

| Sensitivity analysis and publication bias
We performed sensitivity analyses to identify the potential heteroge- Begg's rank correlation test and Egger's linear regression test were performed to assess whether there was publication bias. As the results showed, publication biases of the included studies were found

| DISCUSSION
A previous meta-analysis 16 30 The NYHA functional class also showed severe heterogeneity, which was largely attributed to one study 27 in the sensitivity analysis. The random effects model was chosen, assuming that the underlying true effects differed between studies. Formal statistical tests suggested that there was evidence of publication bias with asymmetric funnel plots and Begg's and Egger's tests.
There were some limitations that should be noted. First, all included studies were nonrandomized observational studies. Although we aimed to avoid bias through various means, due to the limitations of the meta-analysis itself, some bias still existed. However, the extent of bias was within the acceptable range. Second, in some of the included studies the medical pretreatment might influence the outcome of BPA. In order to stabilize the condition, pretreated with pulmonary vasodilators before BPA may be unavoidable in some patients. Thus, we indicated those patients' clinical characters in Table 1. Third, there should be steep learning curve to perform complete BPA unlike in prescribing riociguat. Treatment goal of BPA would also be changed depending on the operators' experience.
Therefore, selecting only initial experience of BPA or only latest experience of BPA in each institute might influence the outcome of this study. Fourth, the definition of inoperable CTEPH remains subjective and is highly dependent on the assessment of the local multidisciplinary CTEPH team based on their surgical experience. This issue is of relevance because patients enrolled in the current studies may have been considered to have potentially operable indications if evaluated by another more experienced CTEPH team. Thus, the present systematic review and meta-analysis is limited by the potential bias introduced by the lack of a standard definition of inoperable CTEPH.
Therefore, our findings should be considered carefully and confirmed with further multicenter RCTs and long-term follow-up studies.

| CONCLUSION
Our meta-analysis indicates that both BPA and riociguat improve pulmonary hemodynamic parameters and exercise tolerance. BPA might be associated with greater improvements in exercise tolerance (6MWD, NYHA functional class) and pulmonary hemodynamics (mPAP, PVR and RAP) but not CO and cardiac index compared to riociguat therapy. The most common complications of BPA were pulmonary edema and pulmonary injury. For riociguat, the most common adverse events were dyspepsia, headache, dizziness, hypotension and nasopharyngitis. Overall, both BPA and riociguat were well tolerated.
However, our findings need to be confirmed with further multicenter randomized control trials (RCTs) and prospective observational studies.