Introduction Sudden cardiac death in young athletes is a devastating event. The screening and detection of potentially life-threatening cardiac pathology by ECG is difficult due to high numbers of false-positive results, especially in the very young. The Seattle ECG criteria (2013) were introduced to decrease false-positive results. We compared the Seattle ECG criteria with the European Society of Cardiology (ESC) ECG criteria of 2005 and 2010 for cardiac screening in high-level junior soccer players.
Methods During the 2012–2013 season, all data from cardiovascular screenings performed on the youth division of two professional soccer clubs were collected. The total study population consisted of 193 male adolescent professional soccer players, aged 10–19 years. Five players dropped out of this study.
Results Applying the ESC criteria of 2005 and 2010 to our population resulted in a total of 89 (47%) and 62 (33%) abnormal ECGs. When the Seattle ECG criteria were applied, the number of abnormal ECGs was 6 (3%). The reduction was mainly due to a reclassification of the long QT cut-off value and the exclusion of right atrial enlargement criteria. All ECG abnormalities using the Seattle criteria related to T-wave inversion criteria.
Conclusion The Seattle ECG criteria seem very promising for decreasing false-positive screening results for high-level junior soccer players.
- Cardiology prevention
Statistics from Altmetric.com
On 19 March 2012, a talented Dutch junior soccer player suffered a cardiac arrest during practice. He was rushed to the hospital, where he died the next day at the age of 13. The loss of a child is a devastating event that has a high impact on the local community. When a child dies during sport activities, which should promote health, the impact is even bigger.
Aiming to prevent cardiac events during sports, the European Society of Cardiology (ESC) proposed a screening protocol in 2005.1 This protocol was based on 25 years of experience gained by Corrado et al,2 and consisted of a questionnaire, a physical examination and an ECG. When abnormalities are found, further examination, including an echocardiogram, is warranted.
These recommendations triggered a scientific debate about the most appropriate screening strategy. Maron et al and the American Heart Association3 stated that the addition of an ECG to screening would be impractical due to expected high false-positive test results that would lead to unnecessary further testing, anxiety about outcomes for the athlete and, possibly, unmerited disqualification from sports. In practice, some sporting programmes include ECG screening, others do not.4
There has been much research to optimise the ECG criteria and minimise false-positive screening results. This caused the ESC ECG criteria to be revised by Corrado et al5 in 2010. Recently, after an international summit in 2012, a second revision was made by Drezner et al6 in 2013, known as the ‘Seattle criteria’.
We aimed to compare the ECG criteria outcome of the ESC recommendations of 2005 and 2010 and the Seattle ECG criteria of 2013 in talented junior soccer players aged 10–19 years.
During the 2012–2013 season, all data from cardiovascular screenings performed on the youth division of two professional soccer clubs were collected. All youth teams of both clubs played at the highest national level. The total study population consisted of 193 male adolescent professional soccer players, aged 10–19 years. Written informed consent was obtained from all participants and their parents/legal guardians, and an Independent Review Board statement was provided.
A standard 12-lead resting ECG was collected by a sports physician (Welch Allyn CardioPerfect software, V.1.6.4). All ECGs were scored by the principal investigator, using the criteria provided by Corrado et al1 in 2005, the revised criteria provided by Corrado et al5 in 2010 and the criteria provided by Drezner et al6 in 2013.
It should be noted that the ESC criteria of 2005 and 2010 were developed for athletes aged 12–35 and the Seattle criteria were developed for athletes aged 14–35.
Data on age (at the screening), ethnicity, height, weight and blood pressure were collected. Ethnicity was self-reported. The options of choice were White-Caucasian, Black-African (Morocco/Turkey/Other), Black-Caribbean, Asian, Mixed and Other. Body surface area was calculated using the Mosteller formula (√(L(cm)×M(kg))/3600).7 Training intensity was calculated by averaging 5 weekly programmes. It was established whether the player was international.
Positive screening result
When there was an abnormal finding on the questionnaire, the physical examination and/or the ECG, the athlete was referred to the (paediatric) cardiologist for further primary examination. These primary examinations included an echo cor, exercise testing and 24 h ECG monitoring. When these tests were not able to clear the athlete, additional secondary testing was performed including, among others, a cardiac MRI. When a cardiovascular disease was diagnosed, we excluded the player for this study. For the ECG screening criteria, we used the Seattle criteria of 2012.6 Reference ranges for the echo cor were based on the data provided by Prior and La Gerche.8
Youth players with a known cardiovascular disease or diagnosed with a cardiovascular disease by a (paediatric) cardiologist during the season 2012–2013 were excluded.
Data analysis was performed using Excel (2003).
Of the 193 players eligible for the study, four dropped out because they no longer played at a high level at the date of the screening. One player was excluded due to an already diagnosed atrioventricular nodal tachycardia (AVNT). A total of 188 players were included in this study. All players with a positive screening result could be cleared by the (paediatric) cardiologist using only the primary additional testing (echo cor, exercise testing and 24 h ECG monitoring).
The population was evenly distributed over the different age categories with an age range of 10–19 years. The average blood pressure was 115/69 mm Hg. A total of 29% of the players were of non-Caucasian ethnicity, with a total of 15% being of black ethnicity. A relative high number of non-Caucasian players were from Turkey, Morocco or the Caribbean (22/54, 41%). There were 9% international players. For the population demographics, see table 1.
The average heart rate was 69 bpm. No player had a heart rate below 40 bpm. PQ time ranged from 108 to 222 ms. The QRS time range was 68–117 ms. Corrected QT time ranged from 353 to 468 ms. An overview of the ECG characteristics is shown in table 2.
Training-related (group 1) ECG changes
Almost three-quarters (72%) of our population had one group 1 change and 64 persons (34%) had more than one group 1 change. Most group 1 changes were seen in the category of sinus arrhythmia (29%), sinus bradycardia (28%) and incomplete right bundle branch block (RBBB) pattern (26%). For an overview of training-related (group 1) ECG changes, see table 3.
The ESC and Seattle criteria
ESC ECG criteria of 2005
Applying the 2005 ESC ECG criteria to our population resulted in 89 (47%) abnormal ECGs. The highest numbers of abnormal ECGs are found with the voltage criteria (27%), long QT criteria (15%), right atrial enlargement criteria (10%) and right ventricular hypertrophy (RVH) criteria (7%).
ESC ECG criteria of 2010
Applying the 2010 ESC criteria resulted in a total of 62 (32%) abnormal ECGs. This is a reduction of 32% compared with the criteria of 2005. The biggest reduction was caused by the exclusion of the isolated voltage criteria. The highest numbers of abnormal ECGs were found with the long QT criteria (15%), right atrial enlargement criteria (10%), RVH criteria (5%) and T-wave inversion criteria (3%).
Seattle ECG criteria of 2013
When applying the 2013 Seattle ECG criteria, the number of abnormal ECGs was 6 (3%). This is a reduction of 90% compared with the 2010 criteria. The biggest reduction was caused by the adjustment of the long QT cut-off value and the exclusion of the right atrial enlargement criteria. All abnormal ECGs were found with the T-wave inversion criteria.
Data for players with ECGs with T-wave inversions are displayed in table 5. Seven of the nine players with T-wave inversion showed a convex (domed) ST segment elevation followed by an (end portion) negative T-wave in V1–V3/V4. An example is shown in figure 1. Two of the nine players showed T-wave inversion in the lateral leads.
All these ECGs are considered abnormal by the ESC criteria of 2005. The 2010 ESC and the Seattle criteria both recognise a common early repolarisation variant in black athletes of Afro-Caribbean decent characterised by domed/convex ST segment elevation followed by T-wave inversion confined to leads V1–V4 (as shown in figure 1). Therefore, athletes #2, #3 and #7 would be classified as normal by these criteria. This early repolarisation pattern does not apply to Caucasians, and thus athletes #1, #4 and #9 are classified as abnormal. Likewise, T-wave inversion beyond V4 (ie, into V5 or V6) is abnormal (athletes #5 and #6). It should also be noted that persistent juvenile T-wave inversion in the anterior precordial leads occurs in up to 8% of prepuberty athletes (ie, 14 or younger).9 ,10 Thus, this may be contributing to abnormal findings in athletes #1 and #4 (although as neither the ESC criteria nor the Seattle criteria define juvenile T-wave inversion, these ECGs are classified as abnormal if following the criteria strictly).
The challenge of limiting false-positive screening results applies particularly in the case of very young athletes. Our study shows that the ECGs of young professional soccer players have a wide range of normal values as well as a high degree of training-related ECG changes, and that use of the Seattle ECG criteria described by Drezner et al6 results in a much lower rate of false-positive screening results compared with the ESC ECG criteria of 2005 and 2010.
This study shows the ECG characteristics of a highly trained population of professional adolescent soccer players. Only a few articles describe normal ECG characteristics in an adolescent population and even fewer describe an adolescent athletic population such as the one in our study.
Rijnbeek et al11 described the ECG characteristics of 200 healthy non-athletic males aged 12–16 years. Our athletic population demonstrated a much wider range (for example, looking at the measured PR time, Rijnbeek et al measured an upper value of 178 ms, while the upper values we measured reached 222 ms). Similarly, Mason et al12 described the ECGs of a population of 79 743 participants participating in a drug trial; 1345 of them were aged 10–19 years and 776 were male. In the article, Mason describes the median value and the 2nd and 98th centiles for the PR interval (median 141 ms, 2nd–98th 104–186 ms), QRS interval (median 89 ms, 2nd–98th 65–108 ms) and axis (median 60, 2nd–98th 0–102) and QTc time (median 403 ms, 2nd–98th 352–448 ms). All median values found by Mason are comparable to our median values, yet if we compare the 2nd and 98th centiles of the PR interval and the QTc time with our population we see that the athletic population has more extreme values than the non-athletic population.
When comparing our results with the adolescent athletic population we see similar results. Sharma et al,13 for example, described the ECGs of 1000 junior athletes (mean age 15.7, range 14–18). When we compare table 2 of this article with the data displayed in tables 2 and 3 of the article by Sharma et al, we see much the same results. When our results are compared with the results found by Somauroo et al,14 however, we see less profound extreme values. These differences with our results may be explained by the fact that the population of Somauroo consists of soccer players older (mean age 16.7, range 15.3–18.3) than the population we describe.
Training-related ECG changes
This study shows a high degree (72%) of training-related ECG changes, as defined by Corrado et al5 and Drezner et al6 as group 1 changes. These high numbers of changes are also found by Sharma et al,13 Bohm et al15 and Brosnan et al16 (respectively ≥80%, 66% and 87%). All these studies, however, show much higher amounts of sinus bradycardia than ours (respectively 28% vs. 80%, 56% and 54%). An explanation for this difference could be a more nervous state of our relative younger population (respectively 14.9 years, range 11–19 vs.15.7 years, range 14–18; 21 years, range 16–38 and 20 years, range 16–35). Table 3 shows that the amount of sinus bradycardia increases with age, and in the >16 age group it is 56%. This amount correlates much better with the percentages found by Sharma et al,13 Bohm et al15 and Brosnan et al.16
ESC vs. Seattle criteria
One of the major problems with using the ECG as a screening tool is the number of false-positive screening results.
ESC ECG criteria of 2005 and 2010
Results on false-positive screening outcome using the ESC ECG criteria of 2005 ranged 10–40%2 ,17–22 (REF). With increasing knowledge on distinguishing ECG abnormalities resulting from intensive physical training from those potentially associated with an increased cardiovascular risk, a consensus statement on interpretation of ECG in athletes was published by the ESC in 2010.5 This statement divided ECG abnormalities into a group 1 (common and training-related) and a group 2 (uncommon and training-unrelated) category. Using this statement led to a decrease in false-positive screening outcomes of 40–80%.23 ,24 The false-positive screening rate using this statement ranged from 2.8% to 20%.16 ,23–29
Seattle ECG criteria of 2013
In 2012, an international group of experts convened in Seattle to update these ECG criteria into what is now known as the ‘Seattle criteria’. One of the goals of these criteria was to decrease false-positive screening results.
To our knowledge, only Brosnan et al16 have evaluated screening results using the Seattle ECG criteria and compared them to the 2010 ESC ECG criteria. Their positive screening outcome was reduced from 17% using the 2010 criteria to 4.2% using the 2013 Seattle criteria (reduction of 75%). The reduction was mainly due to a reclassification of the QTc intervals, of the T-wave inversion isolated to V1–V2 and of the ECG with either isolated right axis deviation or right ventricular hypertrophy on voltage criteria.
Applying the 2005 ESC ECG criteria in our population would result in a positive screening outcome in 47% of cases. This number is unacceptable in the use of a screening tool and leads to many false positives and unnecessary additional testing. This result is higher than the previously reported 10–40%. The reason for this could be because our population consists of a high number of young (<16 years) high-level soccer players. This young age and sports level could be the reason for the higher number of ECG abnormalities.
When applying the 2010 ESC ECG criteria, the positive results drop to 32% (a reduction of 32%). This number is much higher than the previously reported 2.8–20%. Reasons for this discrepancy could be that in the 2010 criteria, the cut-off values were not clearly defined and therefore could be interpreted differently by investigators. For example, Brosnan et al16 used 120° as a cut-off value for intraventricular conduction delay (IVCD) whereas the consensus statement of 20105 recommends a cut-off value of 110°. Another example is the description of the cut-off value for short QTc syndrome. In the consensus statement of 2010, different values were mentioned (380, 360 and 330 ms) and they recommend 380 ms. Brosnan et al16 used 360 ms for the 2010 ESC ECG criteria cut-off value, whereas Drezner and colleagues23 recommend 340 ms. Furthermore, most other studies do not describe the cut-off values used and only refer to the 2010 consensus statement, which makes comparison difficult. These possible different interpretations could partly explain why we found a much higher positive screening result compared with other researchers. Another possible explanation for the higher amount of positive screening outcomes could be the aforementioned high numbers of young (<16 years) high-level soccer players in our population.
Using the Seattle ECG criteria, the positive screening outcome drops to 3%. This is a reduction of 90% compared with the 2010 criteria. The reduction was mainly due to a reclassification of the long QT cut-off value and the exclusion of the right atrial enlargement criteria.
All the positive screening results were found with the T-wave inversion/flattening category. Of the nine ECGs with T-wave flattening/inversion, seven show T-wave inversion/flattening only in the anterior leads (V1–V3/V4). Of these seven ECGs, three were of athletes with Black-Afro-Caribbean/Black-African ethnicity and are therefore classified as normal. Of the remaining four ECGs, two were of athletes aged 14 years and younger. As described by Papadakis et al9, Migliore et al10 and Drezner et al,6 the anterior lead T-wave inversion/flattening may be a juvenile pattern in athletes aged ≤14 years and could therefore be a normal finding among these athletes. If we considered these ECGs as normal, the positive screening results with the Seattle ECG criteria would decrease even further, to only 4c (2%) ECGs.
Ideally, we would have a higher population size when describing ECG characteristics and screening outcomes. Our population size is small, therefore these results should be looked at with caution. However, it is difficult to find large numbers of high-level athletes, especially below the age of 16. In sports medicine and cardiological literature, only Sharma et al13 and Migliore10 have described larger populations of similar age and sports level.
A second limitation is the absence of echocardiographic data of the players. The lack of follow-up is another limitation. As described by Pelliccia et al30 and Migliore et al,10 in a few cases the presence of repolarisation abnormalities in young athletes may represent the initial expression of underlying cardiomyopathies. Undetected cardiac abnormalities may, therefore, be present in players in this population, which may lead to exclusion from this study. However, this chance is very small.
ECG characteristics of high-level junior soccer players show more extreme range values than the non-athletic population. The Seattle ECG criteria appear to have a much higher screening specificity than the 2005 and 2010 ESC ECG criteria and seem very promising for decreasing false-positive ECG screening results for high-level junior soccer players. Future studies should extend these analyses to other age groups and examine the costs and benefits31 of screening with the Seattle criteria.
What are the new findings?
This article provides one of the first evaluations of the Seattle ECG criteria.
This article provides one of the first comparisons between the Seattle and the European Society of Cardiology (ESC) ECG criteria of 2005 and 2010.
We detail the ECG of very young professional soccer players.
How might it impact clinical practice in the near future?
Seattle ECG criteria should replace the European Society of Cardiology (ESC) ECG criteria of 2010.
ECG descriptions provided for (very) young soccer players can be used as reference.
Introduction of the Seattle ECG criteria will decrease false-positive screening results.
Contributors BB, MCdB and WN planned the study. BB and MCdB conducted the data gathering and BB evaluated the data. BB is the main author of the article and MCdB and WN coauthored the article.
Competing interests None.
Patient consent Obtained.
Ethics approval Medical Ethics Review Board of the University Medical Centre in Groningen, the Netherlands.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.