Predictors of ventricular and atrial feature tracking-derived global longitudinal strain in healthy athletes and volunteers

Abstract

Listen

Translate article

Abstract Background Feature tracking (FT) strain analysis provides insights into the mechanics of the heart and carries essential prognostic and diagnostic information without prolonging image acquisition during cardiac magnetic resonance (CMR) scans. However, data regarding CMR-derived FT strain in athletes is scarce, especially in the atria. We explored the determinants of FT global longitudinal strain (GLS) values concerning all four cardiac chambers in healthy competitive athletes and less active volunteers. Methods Long- and short-axis cine sequences were analyzed with Medis Suite v4.0 (Leiden, The Netherlands) to assess FT strain. After excluding individuals with preexisting pathologies and images with quality issues, subjects were classified by sport types following the 2020 ESC Sports Cardiology Guidelines. Multiple linear regression models were constructed to assess the impact of age, sex (male=0, female=1), weekly training hours, and sports category (power coded as 0, mixed as 1, endurance as 2) on GLS values. Atrial strain values are positive, and ventricular ones are negative. Intercepts/constants are denoted as "baseline" for easier comprehension. Results Native CMR scans acquired at a tertiary cardiac centre of 596 healthy Caucasian subjects (411 athletes: 145 females, mean age 21±6, range 14-36 years; and 185 volunteers: 86 females, mean age 24±5, range 13-35 years) were included. Athletes had a high exercise load versus the controls who performed no or maximum recreational physical activity (19±7 vs. 3±2 hours/week, p<.001). When assessing athletes (weekly training≥6h), RA GLS (baseline: 34%, p<.001) is increased in females (β=2.0) but is expected to decrease with every one-hour increase in training hours (β=-0.2). LA GLS is notably affected by sex alone (baseline: 37%, β=3.6, p<.001), suggesting markedly higher values for females. RV GLS (baseline: -30%, p<.001) is associated with both age and training hours (β=0.1 for each), making it less negative. Conversely, LV GLS (baseline: -25%, p<.0001) is significantly modulated by all studied factors, with sports category (β=0.5), age, and training hours (β=0.1 each) increasing values and sex decreasing them (β=-0.9). In less active volunteers (weekly training<6h), sex (β=4.7) is the only significant factor for RA GLS (baseline: 33%, p=.04), and sports category (β=2.7, the group included 98 recreational athletes with low training loads) for LA GLS (baseline: 38%, p=.04). No variables were retained in the model for RV GLS, but sex (β=-1.4) remained a predictor in the LV GLS model (baseline: -21%, p=.002). Conclusion GLS shows significant variations influenced by age, sex, sports category, and training hours across all four heart chambers in athletes. These effects appear mostly inconclusive among less active individuals. This data calls for further advanced analyses of athlete's hearts to refine differentiation methods between healthy adaptation and cardiac pathologies.Left atrial FT analysisModel outputs for GLS predictions