Statistical analysis:
Normally distributed continuous variables are reported as mean ± standard deviation, other continuous variables are reported as median (25th-75th percentile). Categorical variables are reported using frequencies and percentages.
Reproducibility analyses were conducted offline on 15 randomly selected anonymized recordings using a two-way mixed model with absolute agreement between measures. The main investigator (ES) performed intra-observer analysis with a two-week interval between measurements. Rater two (MGS) conducted inter-rater analysis, separated in time and place from rater one, on the same set of recordings from each participant. The raters could select different pictures and loops randomly for analysis and re-analysis.
To explore the interplay between AF, athletic status, and echocardiographic values, we conducted a two-way between-groups analysis of variance with athletic status and AF status as independent variables. In addition, a one-way analysis of variance with polynomial contrasts investigating for a linear trend between SD-TPS and years of regular exercise was performed in all participants without a history of AF to explore the associations between prolonged endurance exercise and LA MD.
To assess the association between AF and SD-TPS, we performed a logistic regression analysis with AF as the dependent variable, adjusting for weight, height, CHA2DS2-VASc score, GLS, QRS width, and indexed maximal LA volume (LAVImax) in all participants and in athletes. We also performed a hierarchical logistic regression adding LASr to the model to evaluate the independent incremental value of SD-TPS in identifying athletes with AF.
Tests with P -values <0.05 were considered statistically significant. All analyses were performed using SPSS Statistics version 26 (IBM, Armonk, NY).