Phenotypic Clustering of Idiopathic Pulmonary Arterial Hypertension: Insights Into Pulmonary Vascular and Cardiometabolic Co-Morbidity Trajectories


Kaymaz C., Kultursay B., Tokgoz H. C., Tanyeri S., Bulus C., KESKİN B., ...Daha Fazla

Pulmonary Circulation, cilt.16, sa.3, 2026 (SCI-Expanded, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 16 Sayı: 3
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1002/pul2.70362
  • Dergi Adı: Pulmonary Circulation
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, BIOSIS, Chemical Abstracts Core, EMBASE, Directory of Open Access Journals, Academic Search Ultimate (EBSCO), Health Research Premium Collection (ProQuest)
  • Anahtar Kelimeler: clustering, comorbidity, EUPHRATES, phenotype, pulmonary arterial hypertension
  • Çukurova Üniversitesi Adresli: Evet

Özet

Idiopathic pulmonary arterial hypertension (IPAH) exhibits significant clinical heterogeneity, necessitating a precision medicine approach. This study aimed to identify distinct IPAH phenotypes using machine learning-based clustering and to evaluate their longitudinal therapeutic responses and long-term survival. We analyzed 297 IPAH patients using hierarchical agglomerative clustering based on baseline demographics, hemodynamics, and comorbidities. The cohort was characterized by two distinctive yet overlapping phenotypic patterns: Cluster 1 (n = 198, “Pulmonary Vascular Phenotype”), comprising younger patients (median age: 40) with low comorbidity burden but severe hemodynamic impairment, and Cluster 2 (n = 99, “Cardiometabolic Phenotype”), consisting of older, predominantly female patients (median age: 65) with high prevalence of obesity (78%), diabetes (73%), and hypertension (57%). Despite comparable baseline multiparametric risk scores, Cluster 1 exhibited more impaired echocardiographic right ventricular-arterial coupling, lower left ventricular filling pressure and higher pulmonary vascular resistance (p < 0.001). Triple combination therapy was significantly more frequent in Cluster 1 (41.6% vs. 25.5%, p = 0.048). Linear mixed-effects modeling over 12 months demonstrated that Cluster 1 achieved superior functional recovery (6-min walk distance increase: 75 m vs. 42 m, p < 0.001) and more profound neurohormonal stabilization (Final NT-proBNP: 55 vs. 145 pg/mL, p = 0.017). However, during a median follow-up of 763 days, 5-year survival was similar between groups (Log-rank p = 0.452), and cluster membership was not associated with mortality in adjusted Cox regression (HR: 0.71; 95% CI: 0.47–1.07; p = 0.099). These findings suggest that while vascular phenotypes demonstrate more robust responses to therapy, mortality is driven by distinct factors—hemodynamic severity in younger patients versus metabolic frailty in the comorbid population.