Between September 2006 and July 2012, 15175 young individuals (aged between 14 and 35 years) underwent cardiac evaluation in the UK, as part of a pre-participation screening program established by the charitable organization Cardiac Risk in the Young. For the purpose of this study selected athletes were defined as individuals involved in regular competition (regional, national or international) and training more than four hours per week. The final population comprised 3423 athletes. Cardiac evaluation consisted of a self-report health questionnaire, physical examination and a 12-lead ECG. On an individual basis, when indicated for clinical reasons, ECG abnormalities or research, transthoracic echocardiography was also performed. All the athletes included in this analysis were asymptomatic and normotensive at rest (<140/90 mmHg). Ethnicity, number of hours of exercise-training/week and level of competition were self-assigned.

A 12-lead ECG was performed using GE® Marquette Hellige (Milwaukee, WI, USA) or Philips® Pagewriter Trim III (Bothell, WA, USA) electrocardiographs, with a paper speed of 25 mm/s and amplification of 0.1 mV/mm, as previously described.12 Heart rate and QRS axis were calculated; intervals, durations, and voltages were measured in each lead. The ECGs were analyzed by the cardiologist responsible for each screening session and later read independently by two cardiologists, highly experienced in sports cardiology, cardiomyopathies and primary electrical cardiac diseases. All ECGs were interpreted and categorized as normal (without changes or with training-related changes) or abnormal (with changes that justified additional investigations for exclusion of cardiac pathology) in accordance with the Seattle criteria (SC) and the Refined Criteria (RC).13–16

Transthoracic echocardiography was performed by a cardiologist or a senior cardiac physiologist, using a Philips® Sonos 7500, Philips® iE33, or Philips® CPX50 (Bothell, WA, USA) and Acuson Computed Sono-graph® 128XP/10c (San Jose, CA, USA). Standard views and chamber measurements were performed as previously suggested.17 Left ventricular wall thickness was measured in two-dimensional (2D) parasternal short axis, at the level of the mitral valve and papillary muscles, and right ventricular dimensions were measured in apical 4-chamber view and right ventricular outflow in parasternal short-axis view at the aortic valve level. Color tissue Doppler and 2D continuous and pulsed Doppler were performed using standard parasternal and apical views. Assessment of diastolic function included pulsed Doppler across the mitral valve and tissue Doppler velocity imaging of the septal and lateral mitral valve annulus and free wall tricuspid valve annulus. An experienced cardiac physiologist independently repeated all cardiac measurements.

The sporting disciplines of all the athletes included in the study were categorized according to the classification proposed by Mitchel et al. in the 36th Bethesda Conference guidelines.11 This classification is based on the peak of static and dynamic components achieved during exercise, evaluated by maximal oxygen uptake and maximal voluntary contraction.

The volume of training was analyzed according to the number of hours performed per week and the competitive level was stratified according to the nature of the competition – regional, national or international.

The following categories were formulated, according to which the distribution of abnormal ECG changes was analyzed:

• •

  Intensity of sport: low/moderate vs. at least one high static or dynamic component (categories III and/or C of the Mitchell et al. classification)11

• •

  Competitive level: regional vs. national/international

• •

  Training volume: ≤20 vs. >20 hours/week

• •

  Type of sport: high static vs. high dynamic component.

Written consent was obtained from individuals aged 16 years or over and from a parent for those aged less than 16 years. The National Research Ethics Service, Essex 2 Research Ethics Committee, granted ethical approval in the UK.

Continuous variables with normal distribution were expressed as means and standard deviations. Normality was tested with the Kolmogorov-Smirnov test. Categorical variables were expressed as frequencies and percentages. Statistical comparison of baseline characteristics was performed using the chi-square test or Fisher's exact test, when appropriate, and the Student's t test or the Mann-Whitney test for continuous variables. To identify independent predictors of abnormal SC, multivariate analysis using a binary logistic regression model (enter method) was performed. Two-tailed tests of significance are reported. For all comparisons, a p value of <0.05 was considered statistically significant. When appropriate, 95% confidence intervals (CI) were calculated. The statistical analysis was performed with SPSS version 21.0 (SPSS® Inc., Chicago, IL, USA).

Of the 3423 athletes included in the final analysis, mean age was 20.1±5.0 years, 2468 (72.1%) were male and the majority were Caucasian (90.1%). Athletes were engaged in 43 sporting disciplines, with rugby (35.0%), football (9.7%), swimming (6.3%) and rowing/canoeing (4.6%) being the most popular sports. Almost half of the overall population (46.3%) were engaged in sports with a high static and/or high dynamic component (category III and/or C). Regarding the competitive level, 795 (23.2%) athletes were involved in regional, 1277 (37.3%) in national and 1352 (39.5%) in international competitions. The mean number of training hours/week was 16.2±7.7 (5-50 hours), with 714 (20.9%) of athletes training for >20 hours/week.

Based on the SC, approximately 80% of the overall population (n=2731) had alterations in the 12-lead ECG: normal/training-related in 2482 (72.5%) and abnormal/training-unrelated in 225 (6.6%) athletes. Among the abnormal ECG changes, the most prevalent were left atrial enlargement in 76 athletes (2.1%), left axis deviation in 47 (1.4%), T-wave inversions in 42 (1.2%) and Wolff-Parkinson White pattern in 28 (0.01) (Table 1). These abnormal ECG findings were more frequent in athletes involved in high-intensity sports (8.0% vs. 5.4%; p=0.002) (Figure 1A). Athletes with abnormal SC were more likely to be of Black/Caribbean descendent (8.0% vs. 5.0%; p=0.047) and to compete at national/international level (82.7% vs. 76.3%; p=0.028) in sports with a high static or dynamic component (56.0% vs. 45.6%; p=0.002). Athletes involved in disciplines with predominantly high dynamic intensity (classes C–I/II) had a higher rate of abnormal ECG changes (9.1% vs. 5.5%; p<0.001), but there was no significant difference in sports with high static intensity (III-A/B) (4.5% vs. 6.7%; p=0.153). Comparing only athletes involved in disciplines with isolated high dynamic intensity (classes C-I/II) with those in high static intensity (classes III-A/B), the former had a higher rate of abnormal ECG changes (9.1% vs. 4.5%; p=0.014). Exercising >20 hours/week was not significantly associated with the presence of abnormal ECG changes (Table 2).

Figure 1.

Distribution of abnormal (training-unrelated) ECG findings (A: Seattle criteria; B: Refined Criteria) according to intensity of sport – low-moderate vs. high (categories III and/or C). Max O2: maximal oxygen uptake; MVC: maximal voluntary contraction. Adapted from Mitchell et al.11

(0.82MB).

By multivariate analysis, national/international level (odds ratio [OR] 1.50, 95% confidence interval [CI] 1.04-2.14; p=0.027) and high-intensity sport (OR 1.55, 1.18-2.03; p=0.002) were independent predictors of abnormal ECG changes, with a trend for Black/Caribbean ethnicity (OR 1.58, 95% CI 0.95-2.62; p=0.078) (Table 3). The combination of the two independent variables revealed a significantly higher frequency of abnormal ECG findings in athletes competing in high-intensity sports at national/international level (8.7% vs. 5.5%; p<0.001) (Figure 2A).

Figure 2.

Distribution of abnormal ECG changes (A: Seattle criteria; B: Refined Criteria) according to the combination of type/intensity of sport and competitive level. Int.: international; Nat.: national.

(0.19MB).

Interpretation of ECGs with the RC led to a decrease of more than half in the number of abnormal/training-unrelated ECG changes, to 103 (3.0%) athletes, including among those involved in high-intensity sports (4.2% vs. 2.0%; p<0.001) (Figure 1B). These athletes with abnormal RC were more likely to be female (47.6% vs. 27.3%; <0.001) and to be involved in disciplines with predominantly high dynamic intensity (classes C-I/II) (4.7% vs. 2.3%; p<0.001), but there was no significant difference in disciplines with predominantly high static intensity (III-A/B) (2.6% vs. 3.0%; p=0.700). Comparing only athletes involved in disciplines with isolated high dynamic intensity (classes C-I/II) with those in high static intensity (classes III-A/B), there was no significant difference in prevalence of abnormal ECG changes (4.7% vs. 2.6%; p=0.127). Competing at national/international level (82.5% vs. 76.6%; p=0.161) and training >20 hours/week (19.4% vs. 20.9%; p=0.715) were not significantly associated with abnormal RC. However, the combination of intensity of sport and competitive level revealed significant abnormal ECG changes in athletes involved in high-intensity sports at national/international level (4.7% vs. 2.1%; p<0.001) (Figure 2B).

Of the overall population, 1345 (39.3%) athletes underwent transthoracic echocardiography. Structural findings were interpreted as abnormal in 26 (0.8%) athletes. The echocardiographic findings classified as abnormal were aortic bicuspid valve (n=7), isolated right ventricular changes (e.g. hypertrabeculation) (n=5), mitral valve prolapse (n=4), anomalous coronary origin (n=2), dilated aortic root (n=2), left ventricular hypertrophy with interventricular septum ≥15 mm (n=1), large patent foramen ovale (n=1), tricuspid valve prolapse (n=1), pulmonary valve stenosis (n=1), atrial septal defect (n=1) and cor triatriatum (n=1). These structural alterations were more frequent in athletes with both abnormal SC (2.7% vs. 0.6%; p=0.010) and abnormal RC (3.9% vs. 0.7%; p<0.001). A higher frequency of echocardiographic changes classified as abnormal was also evident in athletes engaged in high-intensity sports and competing at national/international level (1.3% vs. 0.5%; p=0.012).

Few studies have reported the association of type of sport with ECG findings in athletes, and these were mainly performed before the publication of guidelines for ECG interpretation in athletes, in which the sporting disciplines were analyzed individually and not according to the intensity of the dynamic and static components of exercise.20,21 A higher prevalence of ECG changes in athletes has been mainly reported in endurance disciplines. Pelliccia et al.,9 in 1005 athletes performing at national or international level in 38 different sporting disciplines, showed that abnormal ECG changes were more common in endurance sports such as cycling, canoeing/rowing and cross-country skiing. Athletes engaged in endurance sports exhibit more pronounced physiological structural remodeling, which may also manifest with marked ECG alterations. The present study revealed that this relationship between intensity of sport and ECG changes is also evident for non-physiological changes. However, these results should be interpreted with caution. Despite the differences in cardiac loading conditions induced by different types of exercise, some ECG abnormalities are found in athletes engaged in predominantly dynamic as well as in those in predominantly static sports. Although the association is more evident in sports with high dynamic levels, sporting disciplines in this study are typically characterized by a combination of both types of exercise (at least one type with moderate intensity), as in rugby, swimming and basketball.

The impact of competitive level on electrical remodeling is also not well established and has not been considered in previous classifications of sports. The majority of previous studies were performed in athletes with similar levels of competition, mostly at national or international level. Nonetheless, given the large number of individuals involved in regular sports training at low levels of competition (club or regional) or even not competing, it is important to analyze cardiovascular adaptations in a broad spectrum of competitive levels as well as the other factors that play a part.22 Although this idea is controversial, it has been proposed that the emotional stress of competition can induce sympathetic activation, which is a potential trigger for arrhythmias or myocardial ischemia.10 It might be assumed that sports played at higher levels of competition (national or international), independently of the intensity of physical exercise, would be associated with greater psychological stress. Nevertheless, no relation between emotional stress induced by highly competitive sport and ECG abnormalities has been reported.

Although there is an empirical assumption that the cause-effect relationship between the volume of training and electrical remodeling is obvious, there is little evidence of abnormal ECG findings and the idea is controversial. In a study by Papadakis et al.23 in adolescent Caucasian athletes, the duration of training was not an independent predictor for T-wave inversion in the precordial leads, one of the most frequent abnormalities seen in the ECGs of competitive athletes. Comparing athletes training for fewer hours a week with those training more, the latter more frequently have ECG changes, as shown by Gati et al.19 with respect to left axis deviation and left atrial enlargement.

Our results offer new insights regarding the association of intensity and level of competition with the presence of abnormal ECG findings, and highlight characteristics that should be considered together when interpreting an athlete's ECG.

The authors declare that no experiments were performed on humans or animals for this study.

The authors declare that no patient data appear in this article.

The authors declare that no patient data appear in this article.