Background: Clinical electrocardiographic (ECG) guidelines for athlete's heart are based upon cross-sectional data. We aimed to longitudinally evaluate the influence of endurance training on the ECG and compare the prevalence of ECG abnormalities defined by contemporary criteria. Methods: A group of 66 training-naïve individuals completed a six-month training programme with resting ECGs and cardiopulmonary exercise tests performed at baseline, two and six months. Data were analysed using repeated measures analysis of variance and the prevalence of ECG abnormalities compared between proposed criteria. Results : Maximal oxygen consumption increased from 45.4 ± 7.1 to 50.3 ± 7.1 ml·kg ^-1 ·min ^-1 (p < 0.05) pre-to-post training. ECG changes included, bradycardia (60 ± 12 vs. 53 ± 8 beats·min ^-1 ; p < 0.05), shorter P wave duration (106 ± 10 vs. 103 ± 11 ms; p < 0.05), reduced QTc (413 ± 27 vs. 405 ± 22 ms; p < 0.05), and increased left ventricular Sokolow-Lyon index (2.45 ± 0.66 vs. 2.62 ± 0.78 mV; p < 0.05). 85% of individuals showed ≥1 'training-related' ECG finding at six months vs. 68% at baseline. Using the 2013 Seattle Criteria, 4 ECGs were 'abnormal' at baseline and 3 at month six vs. 2 at baseline and 1 at month six, using the 2017 International Consensus. Prevalence of 'borderline' findings did not increase with training (11% at baseline and six months). Conclusion: Six-months endurance training leads to a greater prevalence of 'training-related' but not 'borderline' or 'training-unrelated' ECGs. 'Borderline findings' may not necessarily represent training-related cardiac remodelling in novice athletes following a six-month training intervention. KEY MESSAGES This study aimed to assess the longitudinal ECG changes following six months of endurance training, in training-naïve individuals, and whether these ECG changes support the revisions made to the recent 2017 international consensus criteria. The prevalence of 'training-related' findings were increased with six months of endurance training, however the prevalence of the revised 'borderline' criteria, according to the 2017 international consensus, did not increase and the associated quantitative ECG data (e.g. P-wave amplitude, QRS axis and QRS duration) remained unchanged. Further clinical consideration may be warranted for individuals within the early phase of exercise engagement presenting with 'borderline' ECG abnormalities, defined by the International criteria.