Clinical, Prognostic and Biological Features of High-Risk Cardiometabolic Phenotype: the REMODEL Study

BackgroundConventional cardiometabolic risk markers (blood pressure, glucose, lipids) incompletely capture biological vulnerability in hypertension. We aimed to identify distinct risk phenotypes and define their clinical, prognostic, and proteomic features. MethodsREMODEL is a prospective observational cohort of asymptomatic adults with essential hypertension who underwent 24-hour ambulatory blood pressure monitoring and standardized cardiovascular magnetic resonance (CMR). Among 83 candidate predictors, outcome-informed feature selection identified four variables with best discrimination: NT-proBNP, Romhilt-Estes ECG score, and CMR-derived indexed interstitial and myocyte volumes. Unsupervised clustering using the KAMILA algorithm (without outcome labels) determined two clusters. The primary outcome was a composite of acute coronary syndromes, first heart failure hospitalization, stroke, and all-cause mortality. Proteomics quantified 192 cardiovascular-related proteins (Olink CVD II/III). ResultsOf 885 participants, two clusters emerged: low-risk (n=787) and high-risk (n=98). Over 60 [37,73] months, the high-risk cluster had markedly worse event-free survival (log-rank P<0.001) and remained independently associated with outcomes (adjusted HR 8.89, 95% CI 4.75-16.65). High-risk individuals were younger, more often male, had greater visceral adiposity, worse renal function, higher 24-hour blood pressures, and adverse remodeling on CMR/biomarkers. Proteomics identified 26 enriched proteins implicating myocardial stress, extracellular matrix remodeling/fibrosis, immune activation, apoptosis, and vascular dysfunction. ConclusionsMultimodal clustering reveals a high-risk hypertensive phenotype with distinct structure-biology coherence and substantially elevated long-term risk, supporting integrated imaging-biomarker-proteomic approaches for refined stratification.