Clinical outcomes with initial oral triple and dual therapy in echocardiography-defined connective tissue disease-associated pulmonary hypertension: a retrospective single-center cohort study

Abstract

Background

Connective tissue disease-associated pulmonary hypertension (CTD-PH) is a serious complication, but evidence on early oral treatment intensification in patients followed primarily by echocardiography is limited. We evaluated outcomes associated with initial selexipag-containing oral triple therapy vs. dual therapy in echocardiography-defined CTD-PH.

Methods

This single-center retrospective cohort included consecutive adults who initiated PH-targeted therapy between April 2020 and June 2025. Treatment was selected by treating physicians in routine care and was not randomized: dual therapy comprised macitentan plus tadalafil, whereas triple therapy additionally included selexipag. Baseline clinical, laboratory, and echocardiographic variables were obtained from standardized institutional assessments and reassessed at week 26. The primary endpoint was the change in the number (0–3) of European Society of Cardiology/European Respiratory Society (ESC/ERS) noninvasive low-risk criteria met: World Health Organization functional class (WHO-FC) I–II, 6-minute walk distance (6MWD) >440 m, and B-type natriuretic peptide (BNP) <50 ng/L or N-terminal pro-B-type natriuretic peptide (NT-proBNP) <300 ng/L. Complete cases with all three components available at both time points were analyzed without imputation. Disease-progression events were recorded through 30 weeks, and adverse events were assessed during follow-up.

Results

Of 119 patients, 59 received triple therapy and 60 received dual therapy. The groups had similar baseline functional and biomarker profiles (6MWD, 290.0 vs. 286.9 m; NT-proBNP, 1,455 vs. 1,473 pg/mL); mean age was 45.5 vs. 47.2 years in the triple and dual therapy group respectively, and 93% vs. 83% were women. At week 26, the change in low-risk criteria differed between groups (P=0.003), with 57.6% vs. 33.3% meeting all the three criteria. Mean 6MWD increased by 156 vs. 104 m (adjusted difference, 52 m; 95% confidence interval (CI): 29–78; P<0.001), and WHO-FC distributions differed at week 26 (P=0.007). Exploratory analysis yielded an hazard ratio (HR) of 0.75 (95% CI: 0.39–1.43) for the first disease-progression event. At least one adverse event occurred in 53/59 (89.8%) vs. 47/60 (78.3%) patients; diarrhea occurred in 24/59 (40.7%) vs. 19/60 (31.7%), nausea in 21/59 (35.6%) vs. 16/60 (26.7%), and jaw pain in 16/59 (27.1%) vs. 11/60 (18.3%), with no significant between-group differences.

Conclusions

In this real-world cohort of patients with echocardiography-defined CTD-PH, initial selexipag-containing oral triple therapy was associated with greater improvement in noninvasive ESC/ERS risk status and functional outcomes than dual therapy over 26 weeks, supporting further evaluation of noninvasive risk-guided oral treatment intensification in CTD-associated pulmonary vascular disease. However, given the small, single-center, nonrandomized retrospective design, these findings cannot establish superior efficacy or safety and require prospective multicenter confirmation with systematic hemodynamic assessment.

Highlight box.

Key findings

• Initial selexipag-containing oral triple therapy was associated with greater improvement in European Society of Cardiology/European Respiratory Society (ESC/ERS) noninvasive low-risk status, 6-minute walk distance, and World Health Organization functional class at 26 weeks than dual therapy. Disease progression was numerically lower with triple therapy, but the difference was not statistically significant.

What is known, and what is new?

• Risk-based treatment and combination therapy are established approaches in pulmonary arterial hypertension; however, research on oral triple therapy in echocardiography-defined connective tissue disease-associated pulmonary hypertension (CTD-PH) is limited.

• This real-world study directly compared selexipag-containing oral triple therapy with dual therapy and incorporated longitudinal noninvasive ESC/ERS risk assessment and exploratory prognostic analyses.

What is the implication, and what should change now?

• Serial noninvasive risk assessment may help monitor treatment response and guide treatment intensification in echocardiography-defined CTD-PH. Prospective multicenter studies with systematic hemodynamic confirmation are needed before the routine adoption of early oral triple therapy.

Introduction

Pulmonary hypertension (PH) is a serious cardiopulmonary complication of connective tissue disease (CTD) and is associated with impaired exercise capacity, progressive right-heart dysfunction, and adverse outcomes (1-3). Importantly, CTD-associated PH (CTD-PH) represents a heterogeneous clinical category that may include pulmonary arterial hypertension (PAH), PH related to left-heart disease or interstitial lung disease, chronic thromboembolic disease, and mixed mechanisms. CTD-associated PAH (CTD-PAH) is therefore a specific hemodynamic phenotype within the broader spectrum of CTD-PH and requires invasive hemodynamic confirmation. Transthoracic echocardiography is valuable for identifying patients with a high probability of PH and for longitudinal assessment, but it does not establish precapillary PH or replace right-heart catheterization (RHC) for definitive hemodynamic classification (4-8).

Current PAH treatment strategies are guided by risk assessment and disease severity. For many patients with low- or intermediate-risk PAH, initial combination therapy with an endothelin receptor antagonist (ERA) and a phosphodiesterase type 5 inhibitor (PDE-5i) is recommended, whereas more intensive regimens incorporating parenteral prostacyclin therapy are generally considered in selected high-risk patients (4,5). Selexipag, an oral selective prostacyclin-receptor agonist, has been shown to reduce morbidity and mortality events when added to background PAH therapy (1,9). However, evidence supporting an initial oral triple regimen containing selexipag remains limited, although recent real-world data in catheter-confirmed PAH have begun to emerge (10,11). Earlier studies of intensive or upfront triple therapy predominantly enrolled patients with catheter-confirmed PAH and often included parenteral prostacyclin analogues (12,13). Contemporary CTD-PAH literature further emphasizes early detection and timely, risk-adapted treatment initiation, while recognizing persistent uncertainty regarding long-term outcomes (11). These data, however, cannot be directly extrapolated to patients encountered in routine rheumatology practice who may be identified and monitored primarily using noninvasive assessments. In addition, treatment intensity in observational practice is influenced by physician judgment, disease severity, anticipated tolerance, comorbidities, and other measured and unmeasured factors. Consequently, nonrandomized comparisons of dual and triple therapy are susceptible to confounding by indication and should not be interpreted as estimates of causal treatment efficacy.

Nevertheless, carefully characterized real-world cohorts may provide descriptive information regarding treatment patterns, longitudinal changes in noninvasive risk status, functional outcomes, and short-term clinical events in populations underrepresented in randomized trials. We therefore evaluated observed 26-week changes in European Society of Cardiology/European Respiratory Society (ESC/ERS) noninvasive low-risk criteria and other clinical measures among patients with echocardiography-defined CTD-PH receiving physician-selected initial oral dual or selexipag-containing triple therapy. Between-group analyses and prognostic analyses were considered exploratory and hypothesis-generating rather than confirmatory estimates of comparative treatment effect. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-2748/rc).

Methods

Study design and ethics

This single-center retrospective observational cohort study was conducted at the Department of Rheumatology and Immunology, The First Affiliated Hospital of Soochow University. Consecutive patients who initiated PH-targeted therapy between April 2020 and June 2025 were identified from the institutional clinical database and screened according to prespecified eligibility criteria. Because this was a retrospective available-cohort study, no formal a priori sample-size calculation was performed; all eligible patients during the study period were included.

The study did not mandate diagnostic procedures, treatment initiation, treatment intensity, or dose escalation. All clinical decisions were made independently by the treating physicians as part of routine clinical care before retrospective study inclusion. The study was approved by the Medical Ethics Committee of The First Affiliated Hospital of Soochow University (No. 20251285), and written informed consent was obtained from all participants before study inclusion. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Study population and diagnostic classification

Eligible patients were adults aged 18–75 years with a CTD diagnosis documented by the treating rheumatologist and echocardiographic findings suggestive of PH within 6 months before initiation of PH-targeted therapy. For retrospective cohort identification, echocardiographic eligibility was operationally defined by an estimated systolic pulmonary artery pressure (sPAP) ≥40 mmHg together with supportive findings, including tricuspid regurgitation and/or pulmonary artery enlargement. This definition was used solely for cohort identification and was not considered equivalent to invasive confirmation of PH or PAH.

RHC was not required for study inclusion. Because some diagnostic evaluations had been performed at referring or outside institutions, documentation of previous catheterization and complete invasive hemodynamic measurements, including mean pulmonary arterial pressure, pulmonary arterial wedge pressure, and pulmonary vascular resistance, was not systematically available in the institutional database. Complete RHC data therefore could not be reliably reconstructed and were not used for cohort definition or statistical analyses.

Patients were excluded if the available clinical records indicated another established cause of PH, including idiopathic PAH, congenital heart disease-associated PAH, left-heart disease, severe obstructive or restrictive lung disease, pulmonary embolism, chronic thromboembolic PH, or another identifiable cause. PH subtype was therefore determined from the available clinical, imaging, and diagnostic records rather than systematic invasive hemodynamic phenotyping. Accordingly, the study population is referred to throughout as having echocardiography-defined CTD-PH, which should not be considered equivalent to catheter-confirmed CTD-PAH.

The primary analysis required complete paired baseline and week-26 assessments for all three components of the prespecified noninvasive low-risk endpoint. Patients without evaluable paired assessments were excluded before formation of the final analytic cohort. No patient in the final analytic cohort was lost to follow-up during the 26-week primary evaluation period. Disease-progression events were additionally recorded through 30 weeks. The overall study flow and follow-up schedule are summarized in Figure S1.

Treatment

Patients were classified according to the physician-selected initial PH-targeted treatment strategy documented in routine clinical practice. Patients in the dual-therapy group received oral macitentan (10 mg once daily) combined with tadalafil, initiated at 20 mg once daily and up-titrated to 40 mg once daily according to individual tolerability. Patients in the triple-therapy group additionally received selexipag, which was initiated at a low dose and gradually up-titrated according to individual tolerability to a maximum of 1,600 μg twice daily.

Treatment assignment was not randomized or protocol-mandated. The choice of dual or triple therapy and subsequent dose adjustment were made by the treating physicians according to the overall clinical assessment, including symptoms, functional status, laboratory and echocardiographic findings, comorbidities, and anticipated treatment tolerance. For this retrospective analysis, treatment exposure was classified according to the initial treatment strategy. The exact interval between initiation of individual treatment components and detailed information on temporary interruption, discontinuation, switching, or subsequent treatment modification were not uniformly available and were therefore not modeled separately.

Outcome measures and follow-up assessments

Clinical assessments were obtained at baseline and at approximately 26 weeks after treatment initiation. Functional status was assessed using World Health Organization functional class (WHO-FC), and exercise capacity was evaluated using the 6-minute walk distance (6MWD) according to standardized institutional procedures. Laboratory measurements, including N-terminal pro-B-type natriuretic peptide (NT-proBNP), hemoglobin, and hepatic and renal biochemical variables, were obtained from the hospital’s accredited central laboratory.

Transthoracic echocardiography was performed by experienced sonographers according to institutional protocols. Recorded variables included right ventricular diameter (RVD), right atrial diameter (RAD), aortic root diameter (AAO), left ventricular end-diastolic diameter (LVEDd), estimated sPAP, tricuspid regurgitation pressure gradient, and tricuspid annular plane systolic excursion (TAPSE). Echocardiographic variables were assessed before treatment initiation and again at week 26 whenever an evaluable examination was available.

The primary endpoint was the change from baseline to week 26 in the number of ESC/ERS guideline-defined noninvasive low-risk criteria met. The count ranged from 0 to 3 and comprised WHO-FC I–II, 6MWD >440 m, and BNP <50 ng/L or NT-proBNP <300 ng/L. All the three components were required at both baseline and week 26 for inclusion in the primary-endpoint analysis; patients without complete paired data were excluded before analysis, and missing values were not imputed.

Secondary outcomes included changes in WHO-FC distribution, 6MWD, NT-proBNP, echocardiographic variables, and routine laboratory measures. First disease-progression events through 30 weeks were analyzed separately as exploratory time-to-event outcomes.

Adverse events were retrospectively abstracted from available clinical records during follow-up. Because the study was not designed as prospective pharmacovigilance, adverse-event analyses were considered descriptive, and information on severity grading, causality, dose modification, and treatment interruption was not uniformly available.

Statistical analysis

Descriptive statistics

Continuous variables were summarized as mean ± standard deviation (SD) or median with interquartile range (IQR), according to their distribution. Categorical variables were presented as frequencies and percentages. Normality was assessed using the Shapiro-Wilk test. Baseline characteristics were compared using Student’s t-test or the Mann-Whitney U test for continuous variables and Pearson’s Chi-squared test or Fisher’s exact test for categorical variables, as appropriate. Because treatment allocation was nonrandomized, the absence of statistically significant baseline differences was not interpreted as evidence of treatment-group balance or exchangeability.

Treatment-response analyses

The primary outcome was analyzed as the change from baseline to week 26 in the number of ESC/ERS guideline-defined noninvasive low-risk criteria met. Because this ordinal outcome ranged from 0 to 3 and showed substantial ties, within-group changes were evaluated using the Wilcoxon signed-rank test, and exploratory between-group comparisons of change were performed using the Wilcoxon rank-sum test. Hodges-Lehmann estimates with corresponding 95% confidence intervals (CIs) were reported as descriptive contrasts where appropriate. The proportion of patients meeting all the three low-risk criteria at week 26 was compared using the Chi-squared test or Fisher’s exact test.

Secondary continuous outcomes were analyzed using paired t-tests or Wilcoxon signed-rank tests for within-group comparisons, depending on the distribution of paired differences. Exploratory between-group comparisons at week 26 were performed using analysis of covariance (ANCOVA), with the corresponding baseline value included as a covariate. Adjusted mean differences with 95% CIs were reported as adjusted between-group contrasts. WHO-FC distributions were compared using the Chi-squared test or Fisher’s exact test, whereas paired categorical comparisons within groups were evaluated using the McNemar test when applicable.

Because treatment assignment was determined by treating physicians in routine clinical practice, confounding by indication could not be excluded. Propensity-score matching or inverse probability weighting was not performed because the modest sample size and incomplete capture of potentially important determinants of treatment selection, including invasive hemodynamics and other clinical factors, would have limited the reliability of propensity-based adjustment. Accordingly, all between-group comparisons were interpreted as exploratory associations rather than causal estimates of comparative treatment effect.

Exploratory prognostic analyses

Exploratory prognostic analyses were performed to investigate factors associated with short-term disease progression. Time-to-event methods were applied only to the separate 30-week disease-progression endpoint. Kaplan-Meier curves were compared using the log-rank test, and Cox proportional hazards regression models were used to estimate hazard ratios (HRs) with 95% CIs. To reduce the risk of model overfitting, the multivariable Cox model was restricted to a limited number of clinically relevant variables identified from the univariable analyses, whereas additional echocardiographic variables were evaluated only in exploratory extended models. The proportional hazards assumption was assessed using Schoenfeld residuals and a global proportional hazards test. No evidence of violation of the proportional hazards assumption was observed for the treatment-only model (P=0.63) or the multivariable model (global P=0.48); all covariate-specific tests were also nonsignificant.

Missing data and statistical considerations

Patients without complete paired baseline and week-26 assessments required for the primary and secondary longitudinal analyses were excluded before formation of the final analytic cohort. Consequently, the 119 patients included in the final cohort had complete data for the variables used in the reported primary, secondary, and Cox regression analyses, and no imputation was performed. No patient in the final analytic cohort was lost to follow-up during the 26-week primary evaluation period. This complete-case cohort definition may have introduced selection bias by excluding patients without paired assessments, potentially limiting the representativeness and generalizability of the analytic cohort.

All statistical tests were two-sided, with P<0.05 considered statistically significant. No adjustment for multiple comparisons was applied to secondary, prognostic, or subgroup analyses; therefore, these findings were interpreted as exploratory. All analyses were performed using R software (version 4.4.1).

Results

Patient characteristics

A total of 119 patients with echocardiography-defined CTD-PH were included in the study, of whom 59 received initial oral triple therapy and 60 received initial oral dual therapy. The mean age was 46.4±15.5 years, and 88% of patients were female. The most common underlying CTDs were systemic lupus erythematosus (39%) and mixed CTD (31%), followed by systemic sclerosis (15%) and Sjögren’s syndrome (SS, 14%).

Age, sex distribution, and underlying CTD subtype did not differ significantly between the initial triple-therapy and dual-therapy groups (Table 1). Baseline demographic, clinical, biomarker, and echocardiographic characteristics are summarized in Table 1. Because treatment allocation was nonrandomized, baseline P values were not interpreted as evidence of treatment-group balance or exchangeability.

Table 1. Baseline characteristics of study patients.

Variable	Initial triple therapy (N=59)	Initial dual therapy (N=60)	Overall (N=119)	P value
Gender (female)	55 [93]	50 [83]	105 [88]	0.09
Age, years	45.5±15.5	47.2±15.6	46.4±15.5	0.60
Etiology				0.70
SLE	21 [36]	26 [43]	47 [39]
SSc	8 [14]	10 [17]	18 [15]
SS	9 [15]	8 [13]	17 [14]
MCTD	21 [36]	16 [27]	37 [31]
6-MWD, m	290.0±118.8	286.9±117.6	288.4±117.7	>0.90
WHO-FC				0.60
I or II	6 [10]	10 [17]	16 [13]
III	40 [68]	37 [62]	77 [65]
IV	13 [22]	13 [22]	26 [22]
NT-proBNP, pg/mL	1,455.0±1,814.3	1,473.1±1,834.0	1,464.1±1,816.5	>0.90
RVD, mm	42.9±8.2	41.1±6.0	42.0±7.2	0.20
RAD, mm	45.5±8.5	43.8±6.8	44.6±7.7	0.30
AAO, mm	31.2±3.0	31.5±3.9	31.3±3.5	0.13
LVEDd, mm	42.3±5.6	43.5±7.9	42.9±6.8	>0.90
sPAP, mmHg	79.1±22.7	75.3±21.4	77.2±22.0	0.30
TRPG, mmHg	73.4±22.0	70.4±21.1	71.9±21.5	0.40
TAPSE, mm	17.8±3.3	17.6±3.1	17.7±3.2	0.60
LVEF, %	63.3±6.1	62.5±5.8	62.9±5.9	0.40
ALT, U/L	17.5±12.5	19.0±10.5	18.2±11.5	0.20
AST, U/L	20.8±13.0	21.6±8.5	21.2±10.9	0.07
SCr, μmol/L	63.3±18.8	60.7±14.5	62.0±16.8	0.60
Hb, g/L	130.0±22.4	124.8±19.7	127.4±21.2	0.20

Values are n [%] or mean ± SD. Complete-case analysis was used for all comparative analyses. 6MWD, 6-minute walk distance; AAO, aortic root diameter; ALT, alanine aminotransferase; AST, aspartate aminotransferase; Hb, hemoglobin; LVEDd, left ventricular end-diastolic diameter; LVEF, left ventricular ejection fraction; MCTD, mixed connective tissue disease; NT-proBNP, N-terminal pro-B-type natriuretic peptide; RAD, right atrial diameter; RVD, right ventricular diameter; SCr, serum creatinine; SD, standard deviation; SLE, systemic lupus erythematosus; sPAP, systolic pulmonary artery pressure; SS, Sjogren’s syndrome; SSc, systemic sclerosis; TAPSE, tricuspid annular plane systolic excursion; TRPG, tricuspid regurgitation pressure gradient; WHO-FC, World Health Organization functional class.

Noninvasive low-risk criteria

In the echocardiography-defined CTD-PH cohort, the proportion of patients meeting the predefined noninvasive low-risk criteria increased significantly from baseline to week 26 in both treatment groups (both P<0.001). The proportion increased from 42.4% (25/59) to 84.7% (50/59) in the triple-therapy group and from 40.0% (24/60) to 78.3% (47/60) in the dual-therapy group (Figure 1A, Table S1). The proportions of patients showing improvement in noninvasive low-risk status by week 26 were also compared between the two treatment groups (Table S2). The distribution of the number of criteria met also shifted after treatment (Figure 1B). At week 26, 57.6% of patients in the triple-therapy group met all the three noninvasive low-risk criteria, compared with 33.3% in the dual-therapy group.

Figure 1.

Changes in ESC/ERS noninvasive risk status and their association with disease progression. (A) Achievement of ESC/ERS noninvasive low-risk criteria at baseline and week 26. (B) Distribution of patients meeting 0–3 noninvasive low-risk criteria. (C) Comparison of changes in the number of ESC/ERS noninvasive low-risk criteria met. (D) Disease progression-free survival according to the number of invasive criteria. ESC/ERS, European Society of Cardiology/European Respiratory Society.

Both treatment strategies significantly increased the number of noninvasive low-risk criteria met (P<0.001 for both groups; Figure 1C, Table S3). Within-group analyses showed a greater median increase in the triple-therapy group than in the dual-therapy group [Hodges-Lehmann estimate: 2.00 (95% CI: 2.00–2.50) vs. 1.50 (95% CI: 1.50–2.00)]. The between-group comparison showed a significantly greater improvement with initial triple therapy [W=1238, P=0.003; Hodges-Lehmann difference approximately 1.00 (95% CI: approximately 0.00–1.00)]. The detailed distributions of patients meeting 0–3 noninvasive low-risk criteria at baseline and at week 26 are provided in Table S4.

Kaplan-Meier curves were constructed according to the number of noninvasive low-risk criteria met at baseline (Figure 1D). Over a maximum follow-up of 30 weeks, no disease-progression events occurred among patients who met all the three criteria at baseline. Patients who met none of the criteria had the highest event frequency. The overall difference among the four groups did not reach statistical significance (log-rank P=0.18).

Secondary endpoints

Echocardiography-estimated pulmonary pressure

Echocardiography-estimated sPAP decreased significantly in both treatment groups over 26 weeks. In the triple-therapy group, estimated sPAP decreased from 79.1±22.7 mmHg at baseline to 57.4±21.3 mmHg at week 26, corresponding to a mean change of −21.8 mmHg (95% CI: −25.3 to −18.2; P<0.001). In the dual-therapy group, estimated sPAP decreased from 75.3±21.4 to 54.8±22.2 mmHg, corresponding to a mean change of −20.6 mmHg (95% CI: −24.7 to −16.4; P<0.001). After adjustment for baseline sPAP, the between-group difference in sPAP change was not statistically significant (Table 2).

Table 2. Changes in clinical, echocardiographic, biomarker, and laboratory parameters from baseline to week 26.

Variable	Initial triple therapy (N=59)				Initial dual therapy (N=60)				Treatment effect/difference
	Baseline	Week 26	Mean change from baseline to week 26	P value	Baseline	Week 26	Mean change from baseline to week 26	P value	Adjusted between-group difference	P value
sPAP, mmHg	79.1±22.7	57.4±21.3	–21.8 (–25.3, –18.2)	<0.001	75.3±21.4	54.8±22.2	–20.6 (–24.7, –16.4)	<0.001	–0.10 (–5.39, 4.85)	0.92
6MWD, m	290.0±118.8	446.1±53.8	156 (130, 182)	<0.001	286.9±117.6	390.8±106.0	104 (81, 127)	<0.001	52 (29.68, 78.09)	<0.001
NT-proBNP, pg/mL	1,455.0±1,814.3	433.4±712.5	–1,022 (–1,414, –629)	<0.001	1,473.1±1,834.0	433.0±638.9	–1,040 (–1,415, –665)	<0.001	4.7 (–183.3, 192.69)	0.96
RVD, mm	42.9±8.2	40.1±7.5	–2.8 (–4.3, –1.2)	0.001	41.1±6.0	38.2±5.9	–2.9 (–4.5, –1.3)	0.003	0.92 (–1.06, 2.89)	0.36
RAD, mm	45.5±8.5	42.7±7.7	–2.8 (–4.4, –1.1)	0.001	43.8±6.8	40.0±6.1	–3.8 (–5.5, –2.2)	<0.001	1.77 (–0.22, 3.81)	0.08
AAO, mm	31.2±3.0	31.3±3.4	0.1 (–0.7, 0.9)	0.79	31.5±3.9	32.2±3.1	0.6 (0.0, 1.3)	0.053	–1.50 (–1.54, 0.21)	0.14
LVEDd, mm	42.3±5.6	44.1±4.9	1.8 (0.6, 3.0)	0.003	43.5±7.9	45.5±6.1	2.1 (0.8, 3.4)	<0.001	–1.04 (–2.14, 0.66)	0.30
TRPG, mmHg	73.4±22.0	55.2±21.0	–18.2 (–22.0, –14.4)	<0.001	70.4±21.1	50.7±21.7	–19.7 (–23.7, –15.6)	<0.001	0.87 (–2.88, 7.39)	0.39
TAPSE, mm	17.8±3.3	18.8±3.0	1 (0.3, 1.6)	0.004	17.6±3.1	19.3±3.1	1.7 (1.0, 2.4)	<0.001	–1.47 (–1.5, 0.22)	0.14
LVEF, %	63.3±6.1	65.4±4.5	2.2 (0.6, 3.7)	0.009	62.5±5.8	64.8±4.5	2.4 (1.0, 3.8)	0.003	0.46 (–1.14, 1.83)	0.65
ALT, U/L	17.5±12.5	17.7±17.4	0.3 (–2.0, 2.5)	0.59	19.0±10.5	21.0±12.7	2.1 (–0.4, 4.5)	0.20	–1.03 (–5, 1.59)	0.31
AST, U/L	20.8±13.0	19.2±10.6	–1.6 (–3.5, 0.2)	0.14	21.6±8.5	22.1±8.7	0.4 (–1.4, 2.3)	0.47	–2.05 (–4.59, –0.08)	0.04
SCr, μmol/L	63.3±18.8	62.2±18.9	–1.2 (–3.8, 1.5)	0.25	60.7±14.5	63.2±17.4	2.4 (–0.7, 5.5)	0.08	–1.61 (–7.19, 0.75)	0.11

Values are n (%), mean ± SD or least-squares mean change (95% CI). Treatment effects were estimated using ANCOVA with treatment group and baseline value as covariates. sPAP and TRPG were estimated by transthoracic echocardiography and should be interpreted as noninvasive estimates rather than invasive hemodynamic measurements. 6MWD, 6-minute walk distance; AAO, aortic root diameter; ALT, alanine aminotransferase; ANCOVA, analysis of covariance; AST, aspartate aminotransferase; CI, confidence interval; LVEDd, left ventricular end-diastolic diameter; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro-B-type natriuretic peptide; RAD, right atrial diameter; RVD, right ventricular diameter; SCr, serum creatinine; SD, standard deviation; sPAP, systolic pulmonary arterial pressure; TAPSE, tricuspid annular plane systolic excursion; TRPG, tricuspid regurgitation pressure gradient.

Exercise capacity and biomarker response

Exercise capacity, as assessed by 6MWD, improved in both groups, with a significantly greater improvement in the triple-therapy group. The mean improvement in 6MWD was 156 m in the triple-therapy group and 104 m in the dual-therapy group, yielding an adjusted between-group difference of +52 m (95% CI: 29–78; P<0.001). Serum NT-proBNP levels decreased substantially in both groups, with a mean reduction of approximately 1,000 ng/L by week 26. No significant between-group difference was observed (P=0.96; Table 2).

Echocardiographic parameters

Echocardiographic indices, including RVD, RAD, and TAPSE, changed from baseline in both groups. Between-group comparisons showed no statistically significant differences for most parameters. A small between-group difference was observed for TAPSE (between-group difference, −1.47 mm), but this did not reach statistical significance (P=0.14). Other echocardiographic measures, including TRPG, left ventricular ejection fraction, LVEDd, and AAO, did not differ significantly between groups.

Distribution of WHO-FC

The baseline WHO-FC distributions are presented in Table 1 (P=0.61). After 26 weeks of treatment, the WHO-FC distribution differed significantly between the initial triple-therapy and dual-therapy groups (Chi-squared =9.8, P=0.007). At week 26, a higher proportion of patients in the triple-therapy group was classified as WHO-FC I–II, whereas fewer patients remained in WHO-FC III–IV compared with the dual-therapy group (Tables S5,S6).

Disease-progression events

During the 30-week observation period, first disease-progression events occurred in 19 of 59 patients (32.2%) in the triple-therapy group and 29 of 60 patients (48.3%) in the dual-therapy group (Table S7). Kaplan-Meier curves showed numerically higher 30-week event-free survival in the initial triple-therapy group than in the dual-therapy group (Figure 2A). In the corresponding univariable Cox analysis, the HR for initial triple vs. dual therapy was 0.75 (95% CI: 0.39–1.43; P=0.38), and the difference was not statistically significant. Hospitalization for worsening PH was the most frequent first event in both groups, followed by clinical worsening and treatment escalation. Clinical worsening, defined as a >15% decline in 6MWD from the post-baseline peak together with WHO-FC III or IV, occurred in 3 patients (5.1%) and 10 patients (16.7%), respectively. Treatment escalation for worsening PH occurred in 2 patients (3.4%) and 6 patients (10.0%), respectively (Figure 2B).

Figure 2.

Disease progression and exploratory prognostic analyses according to treatment strategy. (A) Disease progression-free survival according to treatment group. (B) Comparison of disease-progression events rates between initial triple and dual oral therapy. (C) Time-dependent ROC curves comparing two prognostic models. AUC, area under the curve; CI, confidence interval; HR, hazard ratio; ROC, receiver operating characteristic.

Cox regression analyses

Univariable Cox regression models were fitted to identify factors associated with time to first disease-progression event in patients with echocardiography-defined CTD-PH. Variables with P<0.05 in the univariable analyses were entered into the multivariable model (Table 3). WHO-FC, 6MWD, and NT-proBNP were significantly associated with progression risk in the univariable analyses (P<0.05). After mutual adjustment, only 6MWD remained significantly associated with progression risk (HR per meter increase, 0.994; 95% CI: 0.989–0.999; P=0.01). WHO-FC and NT-proBNP were no longer significantly associated with progression risk in the multivariable model (P=0.28 and P=0.41, respectively). RAD and AAO approached statistical significance in the univariable analyses (P<0.1) and were further evaluated. Their associations were attenuated after multivariable adjustment.

Table 3. Univariable and multivariable Cox regression analysis for predictors of first disease-progression event in patients with echocardiography-defined connective tissue disease-associated pulmonary hypertension.

Variable	Univariate		Multivariate
	HR (95% CI)	P value	HR (95% CI)	P value
Treatment
Initial dual therapy	Ref
Initial triple therapy	0.73 (0.38–1.41)	0.35
Gender
Male	Ref
Female	1.64 (0.50–5.33)	0.41
Etiology	1.14 (0.89–1.47)	0.30
WHO-FC	1.79 (1.02–3.12)	0.041	0.60 (0.24–1.50)	0.28
6MWD	0.995 (0.993–0.998)	<0.001	0.994 (0.989–0.999)	0.01
NT-proBNP	1.000 (1.000–1.000)	0.007	1.000 (0.999–1.000)	0.41
RVD	1.03 (0.99–1.08)	0.15
RAD	1.04 (1.00–1.08)	0.07
AAO	1.08 (0.99–1.17)	0.09
LVEDd	1.03 (0.98–1.07)	0.25
sPAP	1.01 (1.00–1.03)	0.10
TRPG	1.01 (1.00–1.03)	0.21
TAPSE	0.96 (0.86–1.06)	0.38
LVEF	1.02 (0.97–1.08)	0.44
Hb	0.99 (0.98–1.01)	0.34

All candidate variables examined in univariable Cox regression are shown. Variables with P<0.05 in the univariable analyses were included in the multivariable model. 6MWD, 6-minute walk distance; AAO, aortic root diameter; CI, confidence interval; Hb, hemoglobin; HR, hazard ratio; LVEDd, left ventricular end-diastolic diameter; LVEF, left ventricular ejection fraction; NT-proBNP, N-terminal pro-B-type natriuretic peptide; RAD, right atrial diameter; RVD, right ventricular diameter; sPAP, systolic pulmonary artery pressure; TAPSE, tricuspid annular plane systolic excursion; TRPG, tricuspid regurgitation pressure gradient; WHO-FC, World Health Organization functional class.

Exploratory prognostic analyses

During the 30-week follow-up, two exploratory Cox-based prognostic models were evaluated: a baseline clinical model incorporating 6MWD, WHO-FC, and NT-proBNP, and an extended model additionally including RAD and AAO (Figure 2C). Discriminative performance was evaluated using time-dependent ROC curves. The baseline clinical model yielded an AUC of 0.725, whereas the extended model yielded an AUC of 0.741. The difference between the two AUC values reached statistical significance. The addition of echocardiographic parameters produced only a small increase in AUC. Therefore, although echocardiographic variables may provide incremental information, the absolute improvement in discrimination was modest and should be interpreted cautiously because no external validation cohort was available. Across these exploratory models, 6MWD remained the most consistent predictor of 30-week disease-progression events. Calibration plots based on 30 risk groups and 300 bootstrap resamples showed close agreement between predicted and observed Kaplan-Meier estimates, with a mean absolute error of <3% (Figures S2,S3). Across all models, 6MWD remained independently associated with 30-week events. Model performance was evaluated in terms of discrimination and calibration only and should be interpreted as exploratory because no external validation cohort was available.

Subgroup analysis

Figure 3 presents exploratory subgroup analyses comparing initial triple therapy with dual therapy across baseline characteristics. In the overall cohort, the HR for the first disease-progression event was 0.75 (95% CI: 0.39–1.43; P=0.38), with no significant treatment-by-subgroup interactions overall.

Figure 3.

Subgroup analyses comparing initial triple with dual oral therapy across baseline characteristics. The forest plot displays HRs and 95% CIs for each subgroup, reflecting the relative risk of the outcome based on subgroup characteristics. The solid line represents the overall HR. Interaction effects between subgroups characteristics and treatment were assessed using interaction tests, with bold P values indicating statistically significant interactions. 6MWD, 6-minute walk distance; AAO, aortic root diameter; ALT, alanine aminotransferase; AST, aspartate aminotransferase; CI, confidence interval; Hb, hemoglobin; HR, hazard ratio; LVEDd, left ventricular end-diastolic diameter; LVEF, left ventricular ejection fraction; MCTD, mixed connective tissue disease; NT-proBNP, N-terminal pro-B-type natriuretic peptide; pSS, primary Sjögren’s syndrome; RAD, right atrial diameter; RVD, right ventricular diameter; SCr, serum creatinine; SLE, systemic lupus erythematosus; sPAP, systolic pulmonary artery pressure; SSc, systemic sclerosis; TAPSE, tricuspid annular plane systolic excursion; TRPG, tricuspid regurgitation pressure gradient; WHO-FC, World Health Organization functional class; y, year.

In patients with NT-proBNP >789.4 pg/mL, 6 events occurred in the triple-therapy group and 16 occurred in the dual-therapy group (HR, 0.29; 95% CI: 0.11–0.74; P=0.009; interaction P=0.003). In patients with 6MWD ≤300 m, the corresponding HR was 0.39 (P=0.02; interaction P=0.01). No significant interactions were detected for other variables, including underlying CTD subtype, WHO-FC, or imaging-derived cardiac diameters.

Adverse events

During the follow-up period, 53 of 59 patients (89.8%) in the triple-therapy group and 47 of 60 patients (78.3%) in the dual-therapy group experienced at least one adverse event, with no significant between-group difference (P=0.40). Gastrointestinal symptoms, including diarrhea (40.7% vs. 31.7%; P=0.28), nausea (35.6% vs. 26.7%; P=0.28), and vomiting (23.7% vs. 13.3%; P=0.12), were numerically more frequent in the triple-therapy group than in the dual-therapy group. Pain-related adverse events, including jaw pain (27.1% vs. 18.3%; P=0.22) and extremity pain (30.5% vs. 18.3%; P=0.10), were also numerically more frequent in the triple-therapy group. None of these between-group differences reached statistical significance (Table S8).

Discussion

In this single-center retrospective study of patients with echocardiography-defined CTD-PH, initial selexipag-containing oral triple therapy was associated with a greater increase in the number of noninvasive low-risk criteria met at week 26 than dual oral therapy. The most clinically meaningful between-group differences were observed in 6MWD and WHO-FC. Conversely, echocardiography-estimated pulmonary pressure, NT-proBNP, and most right-heart structural parameters changed in a broadly similar manner between groups. These findings suggest that treatment-associated differences were more evident in functional and composite risk-status measures than in most echocardiographic or biomarker endpoints.

These findings are consistent with the rationale for targeting the endothelin, nitric oxide, and prostacyclin pathways in pulmonary vascular disease. Previous studies of upfront triple therapy have largely involved patients with catheter-confirmed PAH, often using parenteral prostacyclin analogues, whereas selexipag has primarily been evaluated in established PAH populations receiving background therapy (9,12,13). By contrast, the present cohort consisted of patients with echocardiography-defined CTD-PH managed in routine rheumatology practice. Differences in diagnostic criteria, baseline risk profile, follow-up strategy, and treatment selection therefore limit direct comparison with those studies. Accordingly, the greater improvement observed with triple therapy in our cohort should not be interpreted as evidence that this strategy provides the same magnitude of benefit as intensive treatment regimens evaluated in catheter-confirmed PAH. Nevertheless, the observed functional improvement provides a clinical rationale for further evaluation of oral prostacyclin-pathway intensification in appropriately characterized patients. More broadly, the rapidly evolving therapeutic landscape of PAH, with increasing emphasis on individualized risk-guided treatment escalation and emerging disease-modifying therapies, further highlights the importance of optimizing treatment strategies according to patient phenotype and risk profile (14,15).

The results also reinforce the practical value of noninvasive risk assessment during follow-up. In PAH cohorts, WHO-FC, 6MWD, and BNP or NT-proBNP were used as simplified low-risk criteria with prognostic relevance (16,17). In the present echocardiography-defined CTD-PH cohort, the number of noninvasive low-risk criteria met increased in both treatment groups, with a greater improvement in the triple-therapy group. Patients who met more low-risk criteria at baseline also had fewer disease-progression events during follow-up, although the log-rank comparison was not statistically significant. Recent CTD-PAH data further support the clinical relevance of serial noninvasive low-risk criteria during follow-up (17). These findings support the use of noninvasive criteria as practical longitudinal monitoring variables. However, their interpretation requires particular caution when the underlying pulmonary vascular phenotype has not been confirmed invasively. In such patients, escalation of PAH-targeted therapy solely on the basis of noninvasive risk status may risk overtreatment or inappropriate treatment if the observed PH reflects alternative or mixed mechanisms. Noninvasive risk assessment should therefore complement, rather than replace, diagnostic confirmation or hemodynamic reassessment when clinically indicated (6,18).

From a clinical perspective, our findings suggest that a selexipag-containing oral triple-therapy strategy warrants further investigation in selected patients with echocardiography-defined CTD-PH, particularly when serial noninvasive assessment indicates persistent functional impairment or an unfavorable risk profile. At the same time, treatment allocation in this retrospective cohort was determined by treating physicians rather than by a prespecified protocol. The apparent advantage associated with triple therapy may therefore have been influenced by baseline disease severity, physician treatment preference, monitoring intensity, anticipated tolerability, or other unmeasured clinical factors, rather than by treatment itself. For this reason, the present data should not be used to support routine initiation or escalation to triple therapy solely on the basis of noninvasive findings. Prospective multicenter studies with standardized hemodynamic confirmation and prespecified treatment allocation are required to determine which patients are most likely to benefit from early oral treatment intensification and how noninvasive risk assessment should be integrated into therapeutic decision-making.

The exploratory prognostic analyses should be interpreted with particular caution. WHO-FC, 6MWD, and NT-proBNP were associated with disease progression in univariable analyses, whereas only 6MWD remained significant after multivariable adjustment. This is consistent with previous studies linking exercise capacity to prognosis and risk stratification in PAH (19-22). The persistence of 6MWD as the most consistent prognostic marker in our analyses further supports the clinical relevance of functional assessment during follow-up, although these findings require confirmation in larger cohorts.

The addition of echocardiographic variables, particularly RAD and AAO, resulted in a modest improvement in model discrimination, suggesting that cardiac remodeling may provide incremental prognostic information beyond conventional clinical variables. The updated 2025 American Society of Echocardiography recommendations also emphasize comprehensive right-heart echocardiography for PH screening, longitudinal monitoring, risk stratification, and prognostication (18). However, the magnitude of improvement was limited and calibration was only evaluated using internal bootstrap resampling without external validation. These findings suggest that readily available echocardiographic parameters may provide incremental prognostic information and enhance noninvasive risk assessment in CTD-PH. However, the findings are exploratory and require confirmation in adequately powered prospective multicenter cohorts with independent external validation before supporting development of a clinically applicable prediction model.

Subgroup analyses suggested that patients with higher NT-proBNP levels or lower 6MWD may derive a greater relative reduction in progression events with triple therapy. However, the number of events within individual subgroups was limited, and multiple subgroup comparisons were conducted without adjustment for multiplicity. These subgroup findings should therefore be viewed as hypothesis-generating for future prospective studies rather than as sufficient evidence to guide treatment selection.

The observed adverse-event profile was generally consistent with the established safety profile of selexipag reported in previous PAH trials and real-world studies (9,23). Gastrointestinal symptoms and pain-related events were numerically more frequent in the triple-therapy group, but none of the between-group differences were statistically significant. The retrospective design limited detailed assessment of dose adjustment, adherence, adverse-event severity, and treatment interruption. Accordingly, the present study was not designed for a direct comparison of the safety profiles of the two treatment strategies.

Several strengths of the present study deserve emphasis. It represents one of the few real-world comparative studies evaluating initial selexipag-containing oral triple therapy in patients with echocardiography-defined CTD-PH and incorporates longitudinal noninvasive ESC/ERS risk assessment together with exploratory prognostic analyses using routinely available clinical and echocardiographic variables. These features provide information that may be relevant to clinical practice in rheumatology settings where longitudinal noninvasive monitoring is frequently used, while also highlighting areas that require validation in hemodynamically characterized populations.

Despite these strengths, several limitations should be acknowledged. First, the retrospective single-center design introduces potential selection bias and residual confounding, particularly because treatment allocation was not randomized. Second, PH was identified primarily using echocardiographic criteria rather than systematic right-heart catheterization; therefore, the findings should be interpreted in the context of echocardiography-defined CTD-PH rather than catheter-confirmed PAH. Third, the sample size and number of disease-progression events were limited, reducing statistical power for survival, subgroup, and exploratory prognostic analyses. Fourth, treatment escalation and follow-up assessments reflected routine clinical practice, and unmeasured clinical factors together with treatment-selection bias inherent to real-world care may have influenced treatment allocation and clinical outcomes. Finally, the exploratory prognostic analyses were limited by the relatively small number of outcome events and the absence of independent external validation. These considerations underscore the need for prospective multicenter studies with standardized hemodynamic phenotyping and prespecified treatment strategies.

Conclusions

In this retrospective real-world cohort of patients with echocardiography-defined CTD-PH, initial selexipag-containing oral triple therapy was associated with greater improvement in noninvasive ESC/ERS risk status and functional capacity than dual oral therapy, without significant differences in echocardiographic estimates of pulmonary pressure. Exploratory prognostic analyses further suggest that routinely available clinical and echocardiographic variables may contribute to longitudinal noninvasive risk assessment. These findings support further evaluation of early oral treatment intensification and integrated noninvasive monitoring in CTD-associated pulmonary vascular disease, but are insufficient to support routine triple therapy based on noninvasive assessment alone. Prospective multicenter studies with standardized hemodynamic confirmation and prespecified treatment allocation are warranted.

Supplementary

The article’s supplementary files as

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was approved by the Medical Ethics Committee of The First Affiliated Hospital of Soochow University (No. 20251285). All participants provided informed consent prior to their inclusion in the study. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Data Sharing Statement

Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2026-2748/dss

jtd-18-10-1144-dss.pdf ^{(90.4KB, pdf)}

DOI: 10.21037/jtd-2026-2748