Introduction

Cardiovascular adaptation to exercise can occur in individuals who undertake high volumes of aerobic sport. Such cardiac changes seen in athletes termed the ‘athlete’s heart’ (AH), may mimic mild phenotypes of cardiomyopathies. These physiological changes can often be observed on a resting electrocardiogram (ECG) and echocardiogram and can be influenced by demographics including sex, age, ethnicity and sporting discipline [1,2,3]. It is therefore important to integrate all these factors when assessing athletes and guide investigations appropriately, in order to help distinguish physiological change from a pathological process.

Historically, the left ventricle (LV) in athletes has been characterized in detail with studies demonstrating how physiological adaptation to exercise may be influenced in adolescents and adults by ethnicity [3,4,5] sporting discipline and gender [1, 6,7,8]. By contrast, there are fewer studies detailing physiological adaptation to exercise in the right ventricle (RV) [9,10,11]. Furthermore, the majority of these studies detail the impact of physiological changes in adults as opposed to characterizing right ventricular geometric changes of the adolescent athlete (aged 13–18 years). There are even fewer studies discussing the potential impact of ethnicity on right ventricular adaptation to exercise.

Arrhythmogenic right ventricular cardiomyopathy (ARVC) has an estimated prevalence of between 1 in 1000 and 1 in 5000 [12]. It is characterized by right ventricular pathology and arrhythmias where there is impaired desmosome function. When subjected to mechanical stress this causes myocyte detachment and cell death. Macroscopically, there is fibrous or fibro-fatty replacement of the myocardium affecting the RV. Phenotypic changes of ARVC can manifest on the echocardiogram as increased right ventricular size and as abnormal T wave inversion (TWI) on the electrocardiogram (ECG). These pathological findings can overlap with physiological changes seen in athletes where chamber dilatation and TWI are well recognised [13,14,15,16,17].

Exercise may exacerbate pathology in those with known ARVC or induce phenotypical change in genotype positive carriers at an earlier stage than would have occurred in a sedentary individual. Multicentre post-mortem series (n = 42) in ARVC has shown that the majority of patients with ARVC die suddenly (81%) with nearly half of these occur during exercise [18]. It is therefore critical that sports participants with potential ARVC are advised appropriately.

Given the potential association of exercise with greater adverse effects, guidelines have recommended restriction in competitive sports for those with ARVD [19].

This study aims to characterize right heart size / function and electrical changes in the adolescent heart from a cohort of male academy footballers (aged 13–18). The impact of ethnicity, indexing to body surface area (BSA) is assessed as well as the potential overlap with ARVC criteria.

Methods

Study design

A total of 1087 academy male football players within the Football Association (FA) underwent mandatory cardiac screening. All players were aged between 13 and 18 years old (mean age 16.0 ± 0.5 years). Written informed consent for screening was obtained from each player by the team doctor. Written informed consent from a parent or guardian was required for athletes younger than 16 years of age, in accordance with the FA governance department. Athletes underwent assessment with a health questionnaire, physical examination, 12-lead ECG and echocardiography.

For age group comparison of the RV, similar data was collected on footballers > 18 years (n = 114, mean age 21.4 ± 3.0 years, all male).

Echocardiographic examination

Two-dimensional echocardiography was performed with Philips (CPX50; iE33, Sonos 7500) and GE Vivid I (Tiral, Israel) machines. Right heart size and function measurements were in accordance with national guidelines [20, 21], see Table 1.

Table 1 Echocardiographic measurements acquired. Adapted from [20, 21]
Full size table

Statistical analyses

Values are expressed as mean ± standard deviation (SD) or percentages as appropriate. Group differences were analysed using ANOVA (with Tukey post hoc test) or the Kruskal–Wallis test with (with Dunn’s post hoc test) where appropriate. Upper reference values for RV dimensions were calculated as mean + 1.96SD for normally distributed data. For non-normally distributed data reference intervals were calculated (MedCalc Software, Ostend, Belgium) using the 2.5th percentile and 97.5 percentile for the lower and upper intervals respectively. The chi-square test was used to assess proportional differences between groups where appropriate. A 2-tailed probability value < 0.05 was considered to indicate significance. Statistical analysis was performed with SPSS software, version 25 (Chicago, IL).

Results

Subjects

Of the total group (n = 1087), mean age was 16.4 ± 0.5y. Eight hundred and twenty six (76%) were white footballers (WFs), 166 (15%) were black footballers (BFs) and 95 (9%) were mixed race footballers (mixed black/white—MFs). BFs had significantly higher weight and BSA than either WFs or MFs (Table 2).

Table 2 Background demographics of academy footballers < 18 years
Full size table

The adult footballers cohort (n = 114) had a mean age of 21.4 ± 3.0y (Table 3). Their mean height, weight and BSA were significantly greater than the academy footballer cohort. Training time (hours per week) was similar between both academy and adult groups (13.3 ± 5.3 vs. 13.9 ± 5.1 h). Academy footballers on average had been training for 8 years. No academy footballer reported any cardiac symptoms suggestive of underlying cardiac pathology or any family history of cardiomyopathy or sudden cardiac death. Self-reported medical conditions included hay fever (n = 14); acne (n = 14); asthma (n = 18); diabetes (n = 1) and juvenile idiopathic arthritis and inflammatory bowel disease (n = 1). One of the cohort was prescribed oral steroids for an underlying medical condition.

Table 3 Background demographics comparing academy footballers (age < 18 years) with footballers > 18 years
Full size table

RV dimensions in academy footballers—normal ranges

Normal ranges for the various right heart echocardiographic dimensions measured in this study are demonstrated in Table 4 and compared with published normal ranges in adults [22,23,24,25] and athletes [10, 23]. For absolute RV dimensions, parameters in the academy footballer cohort exceeded those published in joint American and European guidance [22] by between 5% (for RVD1) to 59% (for RVOT2). Of the 4 adult RV published reference ranges described in Table 4, the ranges for RV dimensions obtained in the current study were most similar to the adult RV normal ranges published by Kou et al. [23] and Addetia et al. [25]. < 10% of any of the measurements were greater than the ranges published by Kou et al. [23] and < 5% were greater than those published by Addetia et al. [25]. Reference ranges from Popple et al. [10] who studied a smaller group of academy footballers (n = 100) were mainly larger than those seen in the current study. The upper limits of reference ranges in the current study (aside to RVOT2) were similar to those published by D’Ascenzi et al. [11] who studied adult competitive athletes from different sporting disciplines.

Table 4 Right heart normal ranges from current study compared with published ranges in non athlete adults and adolescent/adult athletes
Full size table

Electrocardiographic characteristics

TWI in precordial leads was observed in leads V1-V2, V1-V3 or V1-V4 in 4.3%, 3.3% and 1.3% of the cohort, respectively. These patterns of TWI were more frequently seen in BFs than MFs or WFs. TWI in leads V1-V2 was seen in 13% of BF compared with 7.3% of MFs and 2.2% of WFs (P = 0.005). TWI in leads V1-V3 was seen in 13% of BFs compared with 6.3% of MFs and 1% of WFs (P = 0.005). TWI in leads V1-V4 was seen in 4.2% of BFs compared with 2.1% of MFs and 0.6% of WFs (P = 0.001). TWI inferiorly or laterally was a rare occurrence, being seen in 0.1% and 0.2% of the cohort respectively. ECG characteristics of the whole academy footballer group and comparisons between ethnicity is shown in Table 5.

Table 5 Comparison of ECG data in academy footballers according to ethnicity
Full size table

Comparison with ARVC diagnostic criteria

There was significant overlap with certain ARVC parameters for academy football players (Table 6). 12% of the academy football cohort would fulfil major absolute dimension ARVC criteria for RVOTp which fell to 3% following indexing for BSA (Fig. 1). Similarly, 6.1% of footballers would fulfil major criteria for absolute RVOT1 dimension and 23% would fulfil minor criteria. When indexed, these values fell to 2% and 19%, respectively (Fig. 2). 0.4% of the cohort fulfilled ARVC major criteria for RV FAC and 12.1% reached thresholds for minor criteria. Overlap between ECG changes observed in the cohort of academy footballers and criteria for ARVC was also seen. ARVC major criteria for TWI was seen in 3.3% of the cohort. This was more prevalent in BFs (12%) when compared to MFs (6.3%) or WFs (3.3%), P = 0.005. Overall, with the physiological changes in RV dimensions and ECGs seen, 0.2% of the cohort would fulfil diagnosis for ‘definite’ ARVC (for dimension and ECG changes). 2.2% of the cohort would fulfil diagnosis for ‘borderline’ ARVC for dimension and ECG changes and this was seen more frequently in BAs (9.9%) compared with MAs (3.9%) or WAs (0.6%), P = 0.0005. Following further evaluation with cardiac MRI, exercise testing, signal averaged ECG, holter monitoring and familial evaluation, no footballer was diagnosed with ARVC.

Table 6 Comparison of ECG and echocardiographic data from academy footballers against criteria for ARVC
Full size table
Fig. 1
figure 1

Distribution of values for parasternal long axis right ventricular outflow tract dimensions in adolescent footballers. Absolute dimensions are shown in upper chart and indexed values in lower chart

Full size image
Fig. 2
figure 2

Distribution of values for parasternal short axis right ventricular outflow tract dimensions (RVOT1) in adolescent footballers. Absolute dimensions are shown in upper chart and indexed values in lower chart

Full size image

RV functional and geometric parameters in academy and adult footballers

Right heart size (absolute and indexed dimensions) and functional echocardiographic results for academy and adult footballers are reported in Table 7. TAPSE was significantly higher in adults compared with academy players (24.5 ± 4.6 mm vs. 24.0 ± 3.7 mm). For both groups, mean values for functional parameters (RV FAC; S’ RV and TAPSE) were within normal published limits and there was no significant difference between age groups for S’ RV and RV FAC. Right atrial dimensions were similar between the groups.

Table 7 Echocardiographic RV data for academy and adult footballers
Full size table

Analysis of absolute right ventricular dimensions demonstrated that only the RVOT-1 dimension was significantly higher in the adult footballers when compared to the younger academy players. When indexed to BSA this difference was not apparent. There was no significant difference in absolute measurements between the groups for the remainder of the right ventricular outflow or the right ventricular dimensions. When indexed to BSA the academy players (with significantly smaller BSA when compared to adult footballers) demonstrated larger RVD1 and RVD3 values.

Impact of ethnicity

Functional and geometric echocardiographic parameters were compared between ethnicities in the academy cohort (Table 8). All mean values for functional parameters (TAPSE; RV FAC and S’ RV) were normal and there were no significant differences between ethnicities. There were no significant differences in either absolute or indexed measurements between ethnicities for any of the right heart dimension measures undertaken.

Table 8 Impact of ethnicity on echocardiographic RV data for academy footballers
Full size table

Discussion

This study is the largest study to date to provide normal RV dimension data reference ranges for academy footballers of different ethnicities. We highlight that physiological adaptation due to sporting activity can cause increased RV size in adolescents. The RV dimensions of this cohort of academy footballers is larger than joint reference ranges for adults published by the American and European society of echocardiography [22]. In the current study, up to 59% of certain measurements exceeded those of normal adult ranges published by Lang et al. [22]. The adult reference ranges produced Kou et al. [23] and Addetia et al. [25] are most comparable to the RV dimension ranges for academy footballers obtained in this study. Absolute RV dimensions in academy footballers were similar to adult footballers.

The finding of increased RV dimensions in this study of academy footballers has not been previously appreciated in this age group of footballers with this sample size. RV dimensions have been shown by Zaidi et al. to be greater in athletes (mean age 22 years) when compared to controls [26]. Similar findings were demonstrated by Baggish et al. [27] in elite rowers (mean age 25 years) when compared with controls.

Comparisons of RV physiological adaptation to exercise has also been made in endurance trained athletes (ET) and resistance trained athletes (RT). Utomi et al. [28] described ET RV adaptation (increased absolute diastolic area) in ET athletes (mean age 34 years) but there was limited structural changes noted in RT athletes (mean age 21 years). The findings of enlarged RV measurements in ET athletes was also identified by D’Andrea et al. [9] who analysed RV structural changes in 650 elite athletes (395 ET, mean age 29 years) and 255 (RT). Here, it was seen that RV diameters as well as right ventricular outflow tract (RVOT) diameters were significantly greater in ET athletes than either RT athletes or controls.

Differentiating RV physiological change from potential ARVC is important in those undertaking exercise due to the influence of sports participation on the progression of ARVC. Given that desmosomes play a pivotal role in intercellular integrity, endurance athletes with the genetic predisposition for ARVC are hypothesized to be most at risk for phenotypic expression. Ruwald et al. assessed the impact of exercise on probands diagnosed with ARVC [29]. Here those who participated in competitive sport had an increased risk of ventricular arrhythmia or death (hazard ratio 2.05) when compared with those who were inactive or undertook recreational sport. Furthermore symptoms developed at an earlier stage in the competitive sporting group (30 years vs. 38–41 years). In support of this, it has been reported that competitive sports activity can increase the risk of sudden cardiac death in young adults by five times [30].

Saberniak et al. assessed the impact of exercise on ARVC patients (n = 65) and their genotype positive family members (n = 45). In both groups, vigorous exercise (defined as ≥ 1440 metabolic equivalents / week over 6 years) was associated with biventricular dysfunction when compared with non athletes [31]. This notion of exercise being a trigger to deleterious sequelae of genetic mutations is supported by animal models [32] and more recently in human studies by James et al. They found increased risk of VT, heart failure and ARVC in those who are desmosomal mutation carriers who undertake endurance exercise and frequent exercise [33].

Zaidi et al. previous demonstrated overlap between physiological adaptation with Task force criteria for ARVC in adult athletes (mean absolute age 22 years) [34]. Here, ‘possible’ or ‘borderline’ criteria were seen in 51.1% and 44.5% of the cohort respectively. Our current study is the first to demonstrate significant overlap between right heart measures and ARVC task force criteria in a large cohort of academy footballers. 0.2% of the whole cohort fulfilled ‘definite’ criteria for ARVC based on RV dimension and ECG changes (compared with published ranges of ARVC incidence of 0.001–0.0002%). ‘Borderline’ criteria for ARVC was present in 2.2% of the cohort with a significant higher incidence in black (9.9%) over mixed race (3.9%) or white footballers (0.6%). These differences are largely driven by ethnic variation of T wave inversion seen on ECG. The task force criteria for the diagnosis of ARVC [35] includes the presence of TWI (V1–V3 or beyond) as potential criteria for the presence of ARVC.

In this study, standard RV function parameters were normal. However, advanced RV function techniques such as RV speckle tracking echocardiography (STE) may be of use. Dorobantu et al. [36] found that in athletes with RVOT dilatation, STE evaluation of the RV can demonstrate normal function and differentiate physiological remodelling from pathological changes in arrhythmogenic cardiomyopathy, potentially improving screening in grey-area cases.

Study limitations

The cohort size varied between the ethnic groups but was a reflection of a true world sample of academy footballers. This is an all-male cohort and results should not be extrapolated to female athletes. The data obtained were from one time point rather than a longitudinal observational study which would have allowed direct comparison of individual cardiac geometric and ECG changes over time. This study has not presented reproducibility data for echocardiographic measurements although all echocardiograms were performed by British Society of Echocardiography accredited echocardiographers.

Conclusion

This study has characterized the male academy footballer’s right heart. Right heart sizes in excess of some standard adult ranges occur frequently in academy footballers and are similar to those seen in adult footballers. It is not unusual to observe values that would overlap with criteria for ARVC. There was no inter-ethnic variability for RV dimensions identified. This work will be of value to those undertaking RV assessment in similar sporting participants (Fig. 3).

Fig. 3
figure 3

Normal RV dimensions in adolescent male footballers and factors that can help to distinguish physiology from pathology

Full size image