Prognostic significance and pathophysiological insights of functional phenotypes clustering

Abstract Background Functional capacity (FC) is a marker of cardiorespiratory fitness reflecting the integrity of cardiac, respiratory and peripheral metabolic systems. Cardiopulmonary exercise test (CPET) is the method of reference to quantify the functional capacity through a multiparametric approach. Functional phenotypes can be identified according with the patterns of CPET variables. We sought to explore the prognostic role of CPET-derived functional phenotypes, identified by an automatic clustering process, and to understand the underlying cardiac pathophysiology as assessed by exercise-echocardiography (Exe-Echo). Methodology 1379 unselected subjects who underwent to CPET combined with Exe-Echo were included in the study. Cluster analysis based on CPET variables (% of predicted peakVO2, anaerobic threshold, VE/VCO2, peak respiratory ratio, chronotropic index, O2 pulse, peak PetCO2, VO2/WR slope and heart rate recovery) identified 4 clusters (C1= 207, C2= 546, C3= 475, C4= 151 subjects), differing primarily for a progressive worsening of FC (% of PeakVO2: 87±18, 75±13, 60±14, 48±13%; VE/VCO2 25.3±2.7, 26.5±3.0, 30.8±4.3, 39.4±7.8; in C1, C2, C3, C4 respectively, all p<.001). Rest and exercise parameters of Exe-Echo were used to define cardiac characteristics and presented in a standardized radar clustering (fig 1). Patients were tracked for the combined end-point of death + heart failure hospitalization. Results C1 presented normal cardiac morphology and function with expected increase of stroke volume (SV) and cardiac output (CO) during exercise, without significant valvular disease; C2 and C3 showed a mild-to-medarately abnormal rest and exercise echo parameters, characterized by LV progressive dilatation, mild systolic dysfunction, mitral regurgitation (MR) and increased systolic pulmonary pressure (SPAP); C4 corresponded to severe cardiac remodelling with LV and left atrial enlargement, severe MR, rest pulmonary hypertension and severely impaired contractile reserve (reduced exercise SV and CO). The 4 clusters were strongly associated with prognosis, as shown by KM of survival free from combined event (p<.0001; fig 2). Conclusions automatic clustering of CPET parameters identified 4 clusters strongly associated with hard end-points (mortality and HF hospitalization). Exe-Echo provided cluster-related pathophysiological insights explaining the mechanisms responsible for prognosis decline. This combined CPET-ExeEcho approach may help in identifying high risk patients and in defining new therapeutic targets.Radar chart of echo variablesKM of combined end-point (death+HF)