Mid-term prognosis and risk stratification in patients with post-operative pulmonary hypertension: Insights from the Japanese Association of Congenital Heart Disease Registry (JACPHR) Authors: Naofumi F Sumitomo, Taku Ishii, Keiko Uchida, Susumu Hosokawa, Reina Ishizaki, Yoshiyuki Furutani, Hidekazu Ishida, Shinichi Takatsuki, Kei Inai, Shigetoyo Kogaki, Hiroyuki Fukushima, Yuichi Tamura, Hiroyuki Yamagishi, Shozaburo Doi, The Japanese Association of Congenital Heart Disease Registry (JACPHR) members Published: 2026 ABSTRACT AbstractBackgroundPost-operative pulmonary hypertension (PH) following biventricular repair of congenital heart disease (CHD) remains a high-risk condition, but its prognosis and risk stratification are poorly defined. This study aimed to identify prognostic factors and develop a simple clinical risk score.MethodsPatients with post-operative PH enrolled in the Japan CHD–Pulmonary Hypertension Registry (2021–2024) were analyzed. Continuous and categorical variables were assessed by log-rank testing, and a composite risk score was developed to predict cardiovascular events. Prognostic performance was evaluated by Kaplan–Meier and Cox regression analyses.ResultsAmong 84 patients (median age 11.6 years; 55% men), eight cardiovascular events occurred during a mean follow-up of 18.2 months: three all-cause deaths, three hospitalizations for heart failure, and two cases of worsening functional class. Mean pulmonary artery pressure ≥30 mmHg, pulmonary vascular resistance index ≥4.0 WU·m2, B-type natriuretic peptide ≥40 pg/mL or NT-proBNP ≥300 pg/mL, and WHO functional class ≥III were significant prognostic factors. An equal-weighted score (0–4 points) stratified patients into low- (0–1), moderate- (2–3), and high-risk (4) groups with distinct survival (hazard ratio, 3.61; p <0.001).ConclusionsThis nationwide, registry-based study provides new insights into prognostic determinants and risk stratification in patients with post-operative PH. This study identified practical prognostic factors and proposed a simple, exploratory equal-weighted risk score. Prognostic significance was evident even at thresholds below conventional criteria, highlighting the vulnerability of this population and supporting the score's potential clinical utility pending external validation.Keywords: Pulmonary hypertension, Registry, Congenital heart defect, Postoperative, Prognostic risk factors Introduction Pulmonary hypertension (PH) associated with congenital heart disease (CHD) remains a clinically significant condition, particularly in patients with residual or persistent PH following biventricular surgical repair, referred to as post-operative PH [1]. Its prevalence accounts for approximately 6.3–6.5% of all PH cases, and 17.6–30% when idiopathic and heritable forms are excluded [2], [3], [4]. Post-operative PH has also been observed in 3% of patients with simple CHD after repair and in 4–28% of all CHD patients, depending on defect type [5], [6]. Although recent PH-targeted therapies have improved short-term management, the long-term prognosis of post-operative PH remains poorly understood [2], [7]. Survival in this group is reportedly comparable to idiopathic PH and worse than Eisenmenger syndrome, highlighting their high-risk status [1], [5]. Previous studies have reported outcomes such as mortality and transplantation, but most have been descriptive rather than predictive, focusing little on factors associated with either soft (e.g., 6-minute walk distance (6MWD), WHO functional class) or hard endpoints [1], [3], [4], [6]. Consequently, prognostic determinants in post-operative PH remain largely unclear. Moreover, CHD-PH encompasses a broad spectrum of pathophysiologies, including group 1 pulmonary arterial hypertension, group 2 PH due to left heart disease, and group 5 PH associated with complex CHD, often modified by comorbidities such as lung disease or chromosomal abnormalities [5], [8]. This complexity, combined with the absence of standardized management strategies, underscores the need for multicenter data to define disease characteristics, treatment practices, and prognostic factors. However, no registry-based study to date has specifically focused on prognostic assessment in this subgroup. The Japan Congenital Heart Disease–Pulmonary Hypertension Registry (JACPHR) was established in 2021 to address the lack of nationwide data on CHD-associated PH in Japan. This multicenter registry systematically collects clinical and hemodynamic data on patients with CHD-PH from 48 specialized institutions. As of July 2025, 474 cases have been registered, representing a broad spectrum of congenital heart lesions and PH subtypes. This comprehensive real-world dataset provides a unique platform to investigate epidemiology, treatment patterns, and prognostic factors in CHD-PH, including post-operative PH. The first JACPHR report has already described real-world pharmacological management in Japan [9]. Building on this foundation, the present study aimed to leverage JACPHR data to identify prognostic determinants of mid-term outcomes in patients with post-operative PH. Based on these findings, we further sought to construct a simple and clinically applicable risk score using readily available clinical parameters. Methods Study design of the JACPHR The JACPHR is a nationwide, multicenter registry conducted across 48 specialized institutions in Japan. Data are collected primarily on prospective basis, with retrospective entry of key information such as PH onset, underlying CHD, and medical history when available. All cases are confirmed by right heart catheterization according to standardized definitions, with detailed baseline data on cardiac anatomy, comorbidities, genetic syndromes, hemodynamics, and treatment profiles. PH in the JACPHR is defined as a mean pulmonary arterial pressure (mPAP) >20 mmHg or a pulmonary vascular resistance index (PVRi) ≥3 Wood units·m2 by catheterization. The registry enables longitudinal tracking of cardiovascular events (e.g., all-cause death, lung transplantation, clinical worsening, hospitalizations) and treatment changes. Exclusion criteria include patients scheduled for surgery or catheter intervention likely to improve PH within 6 months, those who had recently undergone such procedures, and those with unstable PH without a definitive diagnosis. Participating centers contribute patient data using a standardized protocol with uniform hemodynamic definitions and collection procedures to ensure consistency across sites. Trained investigators enter data into a secure, web-based capture system (Microsoft Azure) incorporating built-in logic and range checks to minimize errors. All entries are centrally monitored for quality and completeness, anonymized using site codes and patient identifiers so that only the registering site can re-identify patients. The central data management team reviews and verifies all records before database locking and analysis. Ethical considerations All participants/guardians provide written informed consent before enrollment. The overall JACPHR protocol, covering all participating centers, was approved by the institutional review boards of each institution, including Keio University Hospital (approval-number: 20140293), and conducted in accordance with the principles of the Declaration of Helsinki. Neither patients nor the public were involved in the design, conduct, reporting, or dissemination plans of our research. Study population This study included patients with post-operative PH enrolled in the JACPHR between August 2021 and December 2024. Eligible patients of any age had CHD with biventricular repair, regardless of subtype, and residual or persistent PH confirmed by right heart catheterization performed on or after January 1, 2005. The following were excluded: (i) patients after Glenn/Fontan procedures with transpulmonary gradient >6 mmHg or PVRi ≥3 WU·m2, and (ii) those with complex CHD involving pulmonary atresia/stenosis and multiple major aortopulmonary collateral arteries after unifocalization, with mPAP >20 mmHg in any lobar pulmonary artery. For patients followed at multiple institutions, registration was completed by the site performing regular catheter evaluations. Cases were classified as “incident” if consent was obtained within 3 months of diagnostic catheterization and “prevalent” otherwise; for prevalent cases, catheterization data at PH diagnosis were also recorded. Data collection Data were collected prospectively according to a standardized registry protocol, with retrospective entry of key variables when available. For this analysis, variables included demographics (birth date, sex, height, and weight), underlying CHD (diagnosis and surgical history), comorbidities, genetic syndromes, airway abnormalities, functional status (WHO functional class; Modified Ross Heart Failure Classification for infants and young children) [10], hemodynamic data from right heart catheterization, laboratory parameters (bilirubin, creatinine, uric acid, BNP, or NT-proBNP), and treatment profiles (PAH-targeted drugs, anticoagulants, and other medications). Follow-up data were obtained at predefined events (all-cause death, lung transplant, hospitalization for heart failure, WHO class worsening, therapy escalation, severe drug adverse events, or new surgical/catheter interventions) or at least annually. Heart failure hospitalization was defined as any admission deemed necessary by the attending physician for the management of heart failure. Worsening of WHO functional class was defined as any increase by at least one class. A complete list of variables is provided in Supplementary Table S1. Statistical analysis Event-free survival was estimated using the Kaplan–Meier analysis, with group comparisons by the log-rank test. Associations between clinical variables and outcomes were evaluated using Cox proportional hazards regression. Hazard ratios were estimated using univariate Cox proportional hazards models, while Kaplan–Meier curves with log-rank tests were used to illustrate differences in event-free survival between groups defined by the selected thresholds. Candidate predictors for the risk score were selected based on their ability to discriminate event-free survival in Kaplan–Meier analyses together with clinical relevance, while univariate Cox regression was used to estimate effect sizes. Cut-off values for continuous variables were first derived using time-dependent receiver operating characteristic (ROC) analysis, then refined by exploring adjacent thresholds to ensure adequate group sizes, discrimination, and clinical interpretability. When either BNP or NT-proBNP was missing, the missing value was complemented by the available counterpart, prioritizing the biomarker with fewer missing cases, in accordance with guideline-based risk strata (BNP ≥50 pg/mL or NT-proBNP ≥300 ng/L) [5]. Because BNP and NT-proBNP reflect similar pathophysiological processes and are commonly used interchangeably in guideline-based risk stratification for pulmonary hypertension, this substitution was considered clinically acceptable. Sensitivity analyses confirmed the validity of this substitution. Final thresholds for survival analyses were therefore determined by balancing statistical robustness and practical applicability. Potential confounding and collinearity among predictors were evaluated using pairwise correlations (Pearson and Spearman) and by calculating variance inflation factors (VIFs) and tolerances. Multivariable Cox regression was planned if ≥20 events were observed; otherwise, analyses were limited to univariate models, and a composite risk score was developed. Weighting strategies for candidate predictors were considered, but equal weighting was applied if event numbers were insufficient for stable coefficient estimation, especially when individual hazard ratios were similar. This approach minimized overfitting and preserved model stability and clinical utility. Continuous variables were tested for normality using histograms and the Shapiro–Wilk test. Normally distributed data were expressed as mean ± SD, and non-normal data as median with interquartile range (IQR). A two-sided p-value <0.05 was considered statistically significant. Analyses were performed using R (v4.5.0; R Foundation, Vienna) and IBM SPSS Statistics (v30.0; IBM Corp., Armonk, NY, USA). Results Patient characteristics Of the 144 patients with post-operative CHD-PH registered in the JACPHR, 85 had follow-up data. After excluding one patient with substantial missing data, 84 patients were included in the final analysis. Approximately half were men, and the median ages at repair, PH diagnosis, and JACPHR enrollment were 1.0, 3.7, and 11.6 years, respectively. The median interval between the last surgery and PH diagnosis was 1.6 years. Regarding CHD classification, pre- and post-tricuspid shunts accounted for 75% of cases. Genetic or chromosomal abnormalities were identified in approximately 60%. A detailed list of underlying congenital heart disease diagnoses is provided in Supplementary Table S2. At enrollment, most patients (90.5%) were in WHO functional class I or II. The median BNP level was 29 pg/mL. Hemodynamic median values were: right atrial pressure, 7.5 mmHg; mPAP 42 mmHg; PVRi 5.2 Wood units·m2; and pulmonary artery wedge pressure (PAWP) 11.3 mmHg. Nine patients (13.6%) had PAWP ≥16 mmHg. Over 30% were not receiving PAH-targeted drugs, while single-, dual-, and triple-agent therapy were used in 26.2%, 20.2%, and 17.9%, respectively (Table 1). Table 1Patient characteristics, data at registration, and follow-up data.N = 84Patient's backgroundMale46 (54.8)Age at biventricular repair1.0 (0.6–4.2)Age at postoperative PH diagnosis3.7 (1.6–11.0)Years from last surgery to PH diagnosis1.6 (0.8–5.6)Age at registration11.6 (5.7–22.4)Body surface area at registration (m2)0.93 (0.67–1.40)Cardiac anomalyPre-tricuspid shunt26 (31.0)Post-tricuspid shunt37 (44.0)Right-heart obstruction10 (11.9)Left-heart obstruction2 (2.4)Malposition of great artery6 (7.1)Other3 (3.6)Genetic anomalyn = 47 (60.0%)21 trisomy29 (34.5)22q11.2DS7 (8.3)Heterotaxy4 (4.8)Airway disease15 (17.9) Data at registrationWHO functional classI53 (63.1)II23 (27.4)III8 (9.5)IV0Laboratory dataBNP (pg/ml) (N = 68)29.0 (13.5–50.5)NT-pro BNP (pg/ml) (N = 19)304 (132–501)Total bilirubin (mg/dl) (N = 78)0.6 (0.4–0.7)Uric acid (mg/dl) (N = 67)4.9 (4.1–5.9)Creatinine (mg/dl) (N = 84)0.46 (0.38–0.57)Catheter dataRight atrium pressure (mmHg)7.5 ± 3.2Mean pulmonary artery pressure (mmHg)42.0 (36.0–49.0)Pulmonary vascular resistance index (W·U × m2)5.2 (3.6–7.8)Pulmonary artery wedge pressure (mmHg)11.3 ± 3.8Number of PAH-targeted drugs030 (35.7)122 (26.2)217 (20.2)315 (17.9) Follow-up dataFollow-up duration (month)18.2 ± 7.6EventTotal8 (9.5)Death3 (3.6)Hospitalization3 (3.6)Worsening of WHO functional class2 (2.4)Data are expressed as mean ± SD, median (IQR1–3), or n (%). For laboratory data, the number of patients with available results is indicated (due to missing values).BNP, B-type natriuretic peptide; PAH, pulmonary arterial hypertension; PH, pulmonary hypertension. Patient characteristics, data at registration, and follow-up data. Data are expressed as mean ± SD, median (IQR1–3), or n (%). For laboratory data, the number of patients with available results is indicated (due to missing values). BNP, B-type natriuretic peptide; PAH, pulmonary arterial hypertension; PH, pulmonary hypertension. Follow-up outcomes The mean follow-up duration from registration was 18.2 months. During this period, eight events occurred: three all-cause deaths, three hospitalizations for heart failure, and two cases of worsening WHO functional class (Table 1). Kaplan–Meier analysis showed a cumulative survival rate of 96.0% at 3.3 years, and a cumulative cardiovascular event–free survival rate of 82.1% at 3.3 years (Fig. 1). Kaplan–Meier curves for overall survival and cardiovascular event–free survival in the study cohort. Identification of prognostic thresholds and predictive markers In time-dependent ROC analyses at 12, 24, and 36 months, mPAP, PVRi, and BNP consistently demonstrated strong prognostic performance as continuous predictors. Cut-off values ranged from 29 to 35 mmHg for mPAP, from 4.9 to 5.2 WU·m2 for PVRi, and from 40 to 61 pg/mL for BNP, with high discrimination at 12 and 24 months and moderate performance at 36 months (Supplementary Table S3). Thresholds of 30–35 mmHg (mPAP), 4–5 WU·m2 (PVRi), and 40–60 pg/mL (BNP) all stratified survival curves. Because these adjacent thresholds showed similar discriminatory performance, the lower values were selected to facilitate earlier risk identification and improve clinical interpretability. The predefined thresholds (30 mmHg for mPAP, 4.0 WU·m2 for PVRi, and 40 pg/mL for BNP) were therefore applied in subsequent analyses. Using predefined thresholds, univariate log-rank analyses identified three continuous variables significantly associated with lower event-free survival: PVRi ≥4.0 WU·m2 (p <0.001), mPAP ≥30 mmHg (p = 0.001), and BNP ≥40 pg/mL (p = 0.03). BNP was available in 68 patients. When 15 patients with NT-proBNP ≥300 ng/L were added using the prespecified surrogate threshold, log-rank analyses yielded consistent results with BNP alone, showing no change in the prognostic trend (p = 0.03). Among categorical variables, only WHO functional class ≥III (p <0.001) and use of triple PAH-targeted therapy (p = 0.02) were significantly associated with outcomes (Fig. 2). Kaplan–Meier curves for event-free survival stratified by prognostic factors. The curve for triple PAH-targeted therapy is shown for reference only, as this factor was excluded from risk score development due to confounding by disease severity. BNP, B-type natriuretic peptide; FC; functional class; MPAP, mean pulmonary artery pressure; PAH, pulmonary arterial hypertension; PVRi, pulmonary vascular resistance index. Sensitivity analyses of triple therapy Among the five candidate factors, the association between triple PAH-targeted therapy and adverse events was likely confounded by disease severity. At baseline, patients on triple therapy had higher mPAP, PVRi, and more frequent WHO class ≥III than those on fewer agents (Supplementary Table S4). As sensitivity analyses restricted to milder hemodynamics, subgroup analyses were performed in patients with PVRi <4.0 WU·m2 (n = 52) and those with both PVRi <4.0 and mPAP <30 mmHg (n = 30). However, no events occurred in either group, precluding evaluation of event-free survival. To further address potential reverse causation, landmark-style analyses were conducted (Supplementary Fig. S1). A significant difference was observed only in the 3-month landmark analysis (48.1% vs. 93.9%, p = 0.024), though event-free survival consistently favored non-triple therapy at 3, 6, and 12 months (3 mo: 48.1% vs. 93.9%; 6 mo: 55.6% vs. 96.8%, p = 0.138; 12 mo: 55.6% vs. 98.2%, p = 0.057). Based on these results, triple therapy was not included as a candidate variable in subsequent prognostic analyses or risk score development. Risk score development and stratification Univariate Cox regression analyses for the four candidate factors (mPAP ≥30 mmHg, PVRi ≥4.0 WU·m2, BNP ≥40 pg/mL or NT-pro BNP ≥300 pg/mL, and WHO functional class ≥III) showed that, although WHO functional class reached statistical significance, hazard ratio estimates varied widely with broad confidence intervals, indicating substantial instability (Supplementary Table S5). To address these limitations and to combine prognostic information, we constructed composite risk scores and evaluated two approaches: an equal-weighted score and a hazard ratio–weighted score. Notably, a moderate correlation was observed between mPAP and PVRi, with collinearity diagnostics showing a VIF of 3.9 and a tolerance of 0.25. Although this suggested some degree of overlap, the values remained within an acceptable statistical range, and both variables were retained as components, and the composite score showed preserved prognostic performance when including both mPAP and PVRi. The equal-weighted score (0–4 points, 1 point per factor) was strongly associated with adverse outcomes (HR 3.61 per 1-point increase, 95% CI: 1.75–7.44, p <0.001, Table 2). The score also demonstrated excellent discrimination with an AUC of 0.905 (Supplementary Table S6). Therefore, the equal-weighted score was adopted as the primary model because it provided comparable discrimination with slightly better model fit while avoiding unstable weighting caused by the limited number of events. Patients were stratified into risk categories based on the total score: low-risk (0–1 points), moderate-risk (2–3 points), and high-risk (4 points). Kaplan–Meier curves (Fig. 3) showed a clear separation across these groups. The overall log-rank test was highly significant (p <0.001), and pairwise comparisons also demonstrated substantial differences between the low- and moderate-risk groups (p = 0.006) and between the moderate- and high-risk groups (p = 0.016). In determining the weighted score, mPAP and PVRi were each assigned 1 point because they were highly correlated with each other, and their dichotomized Cox models showed unstable estimates due to quasi-complete separation. In contrast, BNP and WHO functional class were assigned 2 points each, as they represented more statistically stable and independent prognostic factors. Consequently, the weighted score (0–6 points) also showed significant associations with adverse outcomes (HR 2.04 per 1-point increase, 95% CI: 1.37–3.04, p <0.001) and high discrimination (AUC 0.880) (Supplementary Table S6). Still, model fit was slightly inferior (−2 log likelihood 46.069 vs. 48.500). Table 2Prognostic risk score.Score componentAssigned pointsHR (95% CI)p-ValuemPAP ≥30 mmHg1––PVRi ≥4.0 WU·m21––BNP ≥40 pg/mL or NT-pro BNP ≥300 pg/mL1––WHO functional class ≥III1––Total score0–4 points3.61 (1.75–7.44)<0.001BNP, B-type natriuretic peptide; CI, confidence interval; HR, hazard ratio; mPAP, mean pulmonary artery pressure; PVRi, pulmonary vascular resistance index. Prognostic risk score. BNP, B-type natriuretic peptide; CI, confidence interval; HR, hazard ratio; mPAP, mean pulmonary artery pressure; PVRi, pulmonary vascular resistance index. Kaplan–Meier curves for event-free survival stratified by the equal-weighted risk score. As a sensitivity analysis, the composite endpoint was redefined to include only hard outcomes (all-cause death and heart failure hospitalization), excluding worsening of WHO functional class. Kaplan–Meier analyses demonstrated that the equal-weighted risk score remained significantly associated with hard-endpoint-free survival across the three risk groups (p <0.001, Supplementary Fig. S2). Likewise, baseline WHO functional class ≥III remained significantly associated with worse hard-endpoint-free survival (p <0.001, Supplementary Fig. S2). Discussion Despite its clinical significance, studies specifically evaluating prognostic factors in post-operative PH remain limited. Most previous reports have examined broader pediatric PH or CHD-PH cohorts, leaving the post-operative subgroup underexplored. Although several registries and single-center studies have offered insights into CHD-PH, none have systematically addressed prognostic stratification in post-operative cases. To our knowledge, studies specifically addressing prognostic determinants in post-operative PH remain limited, and this study helps to fill a critical gap in the literature. Overall contribution This study represents the first registry-based analysis to clarify prognostic factors for mid-term outcomes in patients with post-operative PH and to propose a simple risk score. In contrast to previous small-scale reports, the nationwide design of the JACPHR enabled a multicenter evaluation of prognostic determinants in post-operative PH. The use of standardized hemodynamic definitions enabled harmonized case inclusion. The identification of practical clinical parameters, such as mPAP, PVRi, BNP/NT-pro BNP, and WHO functional class, further enhances the applicability of the score in routine practice. Strengths and implications for clinical practice Existing PAH risk scores and risk-stratification strategies, such as the REVEAL score, have primarily been derived from idiopathic or mixed PAH cohorts and have not been specifically validated in post-operative CHD-PH populations [5], [11]. Based on the current results, stratifying patients into low-, moderate-, and high-risk groups, the risk score provides immediate clinical utility. Low-risk patients may be monitored without unnecessary treatment escalation, while moderate- and high-risk patients may benefit from early combination therapy or timely transplant evaluation. Importantly, triple PAH-targeted therapy was not an independent prognostic factor but rather a marker for disease severity. The observed association between triple therapy and worse outcomes most likely reflects confounding by indication, as patients receiving triple therapy exhibited more advanced hemodynamics and functional impairment at baseline. Therefore, this finding should not be interpreted as evidence of harm from therapy but rather an indication that more aggressive treatment tends to be used more in severe cases. Clinical judgment should thus prioritize underlying disease severity over the mere number of drugs, reinforcing the need for individualized therapeutic strategies. Lower prognostic thresholds in post-operative PH Previous studies in other PH populations have proposed various thresholds for prognostic discrimination, but no universally accepted cut-off values have been established. For mPAP, elevated risk has been observed at values in the 20s in general PH cohorts, whereas classic severe PH cohorts reported mean baseline values around 55–60 mmHg [12], [13]. For BNP, guideline-based risk stratification defines <50 pg/mL as low risk and >300 pg/mL as high risk [5]. In pediatric PH, PVRi ≥20 WU·m2 has been linked to poor outcomes [14]. In addition, PVRi ≥6–8 WU·m2 is often regarded as a contraindication to surgical repair [15], and although derived from different disease populations, thresholds around 3–4 WU have been more consistently linked to reduced survival [16], [17]. Overall, prognostic significance was evident even at levels lower than those traditionally considered markers of severe disease, suggesting particular vulnerability in this population. Furthermore, when these factors co-occurred, their combined effect identified a subgroup with markedly worse outcomes, supporting the utility of a composite risk score rather than relying on single parameters. Validity of the composite endpoint In the JACPHR, several types of clinical worsening events are recorded during follow-up; however, in the present cohort, only three event types— all-cause mortality, hospitalization for heart failure, and worsening of WHO functional class—were actually observed during the analysis period. Therefore, these three components constituted the composite endpoint used for prognostic evaluation in this study. Although they differ in their degree of objectivity, each represents an established marker of disease progression in pulmonary hypertension. All-cause mortality is the definitive hard endpoint, heart failure hospitalization reflects clinically significant decompensation requiring intervention, and an increase in WHO functional class indicates functional deterioration that is consistently associated with adverse long-term outcomes. Taken together, the observed components provide a coherent and clinically meaningful measure of post-operative PH worsening, supporting the validity of the endpoint definition applied in this exploratory analysis. Importantly, sensitivity analyses restricted to hard outcomes yielded similar results, supporting the robustness of the main findings. Future directions and limitations The continued growth of the JACPHR will enable further validation and refinement of the proposed risk score, including subgroup analyses by CHD type, evaluation of specific drug classes, and incorporation of hard endpoints such as mortality and transplantation. This study has several limitations. The relatively small number of events reduced statistical power, precluded robust multivariable modeling, and increased the risk of overfitting. The relatively short follow-up period may have limited the ability to capture late clinical events in this population. Consequently, the proposed risk score should be considered exploratory and hypothesis-generating, pending external validation. Genetic anomalies were common in this cohort, particularly trisomy 21, which may reflect a distinct clinical phenotype of post-operative PH and may further limit the generalizability of the findings. In addition, nine patients had PAWP ≥16 mmHg at enrollment, suggesting that post-capillary components may have been present in a subset of the cohort. This potential hemodynamic heterogeneity should be considered when interpreting the study findings. Some baseline data were retrospectively collected, the analysis was limited to patients with follow-up data, and the composite endpoint included both hard and soft outcomes, all of which may introduce bias. Survival analyses were calculated from the time of registry enrollment rather than the time of PH diagnosis, which may have resulted in underestimation of early post-diagnostic events. Although JACPHR collects treatment data during follow-up, therapy intensity was analyzed only at baseline, consistent with the study's focus on prognosis from the time of enrollment. Survival analyses were based on registry enrollment rather than diagnosis, potentially missing early post-diagnostic events. Additionally, although adult cases were included, most participating centers are pediatric, leading to a predominance of pediatric and young adult cases, with a relatively young mean age of the cohort. Finally, as this study was conducted using a Japanese registry, generalizability to other populations may be limited; validation in international cohorts will therefore be essential. Despite these limitations, this study represents an important step toward standardized prognostic assessment and more personalized treatment strategies for patients with post-operative PH. Conclusion In summary, this study identified key prognostic factors associated with mid-term outcomes in patients with post-operative PH and proposed a preliminary, exploratory risk score based on readily available clinical variables. Given the limited number of events, these findings should be considered hypothesis-generating rather than definitive. The observation that even mild hemodynamic abnormalities were associated with adverse outcomes highlights the potential vulnerability of this population. As the JACPHR registry continues to expand, further validation and refinement of the proposed score in larger and independent cohorts will be essential before its adoption in routine clinical practice. CRediT authorship contribution statement Naofumi F. Sumitomo: Writing – original draft, Supervision, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Taku Ishii: Writing – review & editing, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Keiko Uchida: Writing – review & editing, Resources, Methodology, Investigation, Formal analysis, Conceptualization. Susumu Hosokawa: Writing – review & editing, Resources, Methodology, Conceptualization. Reina Ishizaki: Writing – review & editing, Resources, Methodology, Conceptualization. Yoshiyuki Furutani: Writing – review & editing, Resources, Methodology, Conceptualization. Hidekazu Ishida: Writing – review & editing, Resources, Methodology, Conceptualization. Shinichi Takatsuki: Writing – review & editing, Resources, Methodology, Conceptualization. Kei Inai: Writing – review & editing, Resources, Methodology, Conceptualization. Shigetoyo Kogaki: Writing – review & editing, Resources, Methodology, Conceptualization. Hiroyuki Fukushima: Writing – review & editing, Resources, Methodology, Conceptualization. Yuichi Tamura: Writing – review & editing, Resources, Methodology, Conceptualization. Hiroyuki Yamagishi: Writing – review & editing, Supervision, Resources, Methodology, Conceptualization. Shozaburo Doi: Writing – review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Declaration of Generative AI and AI-assisted technologies in the writing process During the preparation of this manuscript, the authors used ChatGPT (OpenAI) to refine English expressions and improve readability, as English is not their native language. The authors reviewed and edited all text and take full responsibility for the content of this publication. Sources of funding This research was supported by the Japan Agency for Medical Research and Development (AMED) under Grant Number JP21ek0109546 and MHLW FC Program Grant Number JPMH25FC1009. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Acknowledgements The JACPHR study included the following principal investigators (JACPHR members): Hiromichi Nakajima, Koji Higashi (Chiba Children's Hospital, Chiba, Japan), Shigeru Tateno, Ryota Ebata (Chiba Kaihin Municipal Hospital, Chiba, Japan), Yuichi Ishikawa, Sayo Suzuki (Fukuoka Children's Hospital, Fukuoka, Japan), Naoki Kuwabara (Gifu Prefectural General Medical Center, Gifu, Japan), Kentaro Ikeda, Kimiko Nakajima (Gunma Children's Medical Center, Gunma, Japan), Naomi Nakagawa (Hiroshima City Hiroshima Citizens Hospital, Hiroshima, Japan), Kotaro Urayama (Tsuchiya General Hospital, Hiroshima, Japan), Atsuhito Takeda, Ayako Nagai (Hokkaido University, Hokkaido, Japan), Hisanori Sakazaki (Hyogo Prefectural Amagasaki General Medical Center, Hyogo, Japan), Toshikatsu Tanaka (Hyogo Prefectural Kobe Children's Hospital, Hyogo, Japan), Junko Shiono, Lisheng Lin (Ibaraki Children's Hospital, Ibaraki, Japan), Tomohiro Watanabe (Tsuchiura Kyodo General Hospital, Ibaraki, Japan), Hitoshi Horigome, Takashi Murakami (University of Tsukuba, Ibaraki, Japan), Hirofumi Saiki, Satoshi Nakano (Iwate Medical University, Iwate, Japan), Tatsuya Onishi, Kazuya Terada (NHO Shikoku Medical Center for Children and Adults, Kagawa, Japan), Kentaro Ueno, Yoshihiro Takahashi (Kagoshima University, Kagoshima, Japan), Hideaki Ueda, Takuya Wakamiya (Kanagawa Children's Medical Center, Kanagawa, Japan), Yoichiro Hirata, Toshimi Koitabashi (Kitasato University, Kanagawa, Japan), Yusuke Nakano, Shigeo Watanabe (Yokohama City University, Kanagawa, Japan), Jun Muneuchi (JCHO Kyushu, Kyushu, Japan), Yoshihide Mitani, Hirofumi Sawada (Mie University, Mie, Japan), Toshinobu Ifuku (Miyazaki Prefectural Miyazaki Hospital, Miyazaki, Japan), Kiyohiro Takigiku, Kohta Takei (Nagano Children's Hospital, Nagano, Japan), Hirohiko Motoki, Koji Yoshie (Shinshu University, Nagano, Japan), Kenji Baba (Okayama University, Okayama, Japan), Eiji Ehara, Mitsuhiro Fujino (Osaka City General Hospital, Osaka, Japan), Kanta Kishi (Osaka Medical and Pharmaceutical University, Osaka, Japan), Futoshi Kayatani, Toshiaki Aoki (Osaka Women's and Children's Hospital, Osaka, Japan), Kenji Hoshino, Ryo Mafune (Saitama Prefectural Chilidren's Medical Center, Saitama, Japan), Satoru Iwashima (Chutoen General Medical Center, Shizuoka, Japan), Norie Mitsushita (Shizuoka Children's Hospital, Shizuoka, Japan), Mitsuru Seki, Kensuke Oka (Jichi Medical University, Tochigi, Japan), Yasunobu Hayabuchi, Yukako Honma (Tokushima University, Tokushima, Japan), Tatsuhiko Anzai (Institute of Science Tokyo, Tokyo, Japan), Hiroshi Ono, Susumu Urata (National Center for Child Health and Development, Tokyo, Japan), Tadahiro Yoshikawa (Sakakibara Heart Institute, Tokyo, Japan), Toshio Nakanishi (Sakakibara Heart Institute Clinic, Tokyo, Japan), Takanari Fujii, Toshiro Shinke (Showa University, Tokyo, Japan), Ryo Inuzuka, Atsushi Yao (The University of Tokyo, Tokyo, Japan), Jun Maeda (Tokyo Metropolitan Children's Medical Center, Tokyo, Japan), Gen Harada (Tokyo Women's Medical University, Tokyo, Japan), Yoichi Mino (Tottori University, Tottori, Japan), Keiichi Hirono (University of Toyama, Toyama, Japan), Yohei Hasebe, Yosuke Kono (University of Yamanashi, Yamanashi, Japan). We would like to thank Editage (www.editage.jp) for English language editing. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.ahjo.2026.100784. Supplementary data Supplementary Fig. S1 Landmark Kaplan–Meier Analyses of Event-Free Survival at 3, 6, and 12 months. Supplementary Fig. S2 Kaplan–Meier curves for hard-endpoint (death and heart failure hospitalization) survival stratified by prognostic factors and WHO functional class. Supplementary tables Data availability The dataset is available from the corresponding author upon a reasonable request. References Associated Data Supplementary Materials Supplementary Fig. S1 Landmark Kaplan–Meier Analyses of Event-Free Survival at 3, 6, and 12 months. Supplementary Fig. S2 Kaplan–Meier curves for hard-endpoint (death and heart failure hospitalization) survival stratified by prognostic factors and WHO functional class. Supplementary tables Data Availability Statement The dataset is available from the corresponding author upon a reasonable request.