Abstract Background Pulmonary hypertension (PH) is classified into five clinical diagnostic groups, including group 1 (idiopathic pulmonary arterial hypertension (IPAH) and connective tissue disease-associated PAH (CTD-aPAH)) and group 4 (chronic thromboembolic pulmonary hypertension (CTEPH)). PH is a progressive, life-threatening, incurable disease. The pathological mechanisms underlying PH remain elusive; recent evidence has revealed that abnormal metabolic activities in the endothelium may play a crucial role. Purpose Introducing a novel approach for studying PH endothelial function, building on genome-scale metabolic (GEM) reconstruction of the endothelial cell (EC) to investigate intracellular metabolism. Methods This prospective observational study includes 35 PH patients with either PAH or CTEPH. We have previously developed a platform to study the endothelium cell (EC) by reconstructing the most comprehensive endothelial genome-scale metabolic model (EC-GEM). Since the EC is in direct contact with the blood flow, we can adjust the EC-GEM by measuring extracellular metabolites from the blood and calculating the intracellular metabolic flux in the EC. We conducted individual EC-GEM models, i.e., in total 35 different models. Additionally, each model was widely analyzed; in total, 190 different intracellular activities were investigated. This enables us to develop a heatmap of the intracellular activities at individual patient levels. Patients with similar intracellular endothelial activities were combined. Results Four distinct intracellular phenotypes were identified; all were independent of the clinical classification of PH. We observed clinically relevant differences in the NT-pro BNP levels and the WHO functional classification between these phenotypes. Also, different important cellular activities were identified separating the four phenotypes. Two phenotypes were characterized by catabolic crisis and extreme citric acid cycle turnover for energy production. However, one of these also suffered from oxidative stress. Conclusion In PH patients metabolic systems analysis identified four metabolic phenotypes that correlated with NT-pro BNP levels and the WHO functional classification but were independent of the clinical diagnosis. This novel finding may enable us to identify metabolic therapeutic targets specific to the respective phenotypes and thus promote personalized medicine in patients with PH. Funding Acknowledgement Type of funding sources: Foundation. Main funding source(s): This study was funded by Novo Nordisk Foundation research grant NNF20OC0064556 and the Heart Center Research Council, Copenhagen University Hospital, Rigshospitalet, provided to Jørn Carlsen