Follow-up analysis of cardiac indices in elite athletes using cardiomagnetic resonance imaging

Background Exercise-induced cardiac remodelling (EICR) may increase biatrial and biventricular volumes and left ventricular mass (LVM). These cardiac adaptations can resemble cardiac pathology. Current understanding of the athlete’s heart is mainly based on cross-sectional studies, with limited longitudinal follow-up in elite athletes. Purpose To explore remodelling patterns of EICR over time using cardiac indices in follow-up cardiac magnetic resonance imaging (CMR). Methods We conducted a follow-up analysis of elite athletes from the ELITE cohort (>16 years old, >10 training hours, and competing at national, international, or Olympic levels). Athletes were included if they remained active and had multiple separate CMR entries. The most recent follow-up entry was used to establish a baseline-endpoint comparison. The follow-up parameters were compared with the baseline parameters. The primary outcomes were changes in left ventricular and right ventricular end-diastolic volumes (dLVEDV and dRVEDV, respectively) and in LVM (dLVM), using smallest detectable change derived from previous literature (1-2). Secondary outcomes included changes in left- and right atrial volumes (dLAV and dRAV, respectively). LV/RV ratio and ventricular remodelling patterns (dRVEDV, dLVEDV, and dLVM) were evaluated. Results 56 elite athletes were included (41% Women; 46% Endurance, 46% Mixed) with a mean follow-up of 1.67 [0.94; 3.34] years. Compared to baseline, men had a decrease in LVM after follow-up (66.9 [60.6; 75.0] g/m2 vs 59.5 [48.8; 68.2] g/m2, p=.036), while women showed no significant dLVM (50.9 [46.9; 62.5] g/m2 vs 50.5 [45.0; 57.3] g/m2). Regression of LAV and RAV was seen in men (dLAV: -20.39, p=.000; dRAV -33.46, p=.000), while women only demonstrated RAV regression (dLAV: -6.28, p=.526; DRAV -18.59, p=.022). The LV/RV ratio transition did not differ between sexes (p = .48), with 78.6% of men and 95.5% of women maintaining a stable LV/RV ratio. The three most prevalent ventricular remodelling patterns were 1) stable dLVEDV and dRVEDV with LVM regression (n=5, 60% female), 2) a decrease in dLVM with increased dLVEDV and stable dRVEDV (n=5, 20% women), and 3) stable dLVEDV and dRVEDV, and decreased dLVM (n=5, 40% women). Sport category (p=.090) and sex distributions (p=.800) were not significantly different. Conclusion Over time, elite athletes do not demonstrate a uniform progression in EICR, emphasising inter-individual variability. Men showed significant regression in LVM over follow-up, whereas women remained stable, which could be due to differences in training intensity. Although atrial volume regression was observed, this finding should be interpreted with caution, as it may partly reflect inter-observer variability. Larger longitudinal cohort studies are needed to confirm these patterns and further elucidate cardiac adaptation in elite athletes.