The influence of training status on right ventricular morphology and segmental strain in elite pre-adolescent soccer players

Twenty-two highly trained, male youth SP and 22 age- and sex-matched recreationally active individuals (CON) were recruited to participate in the study. The SP were recruited from two Category one Premier League youth soccer academies. Category one academies represent the highest achievable grade for professional youth soccer academies in the English Premier League. The CON participants were recruited from a local high school close to one of the Premier League Academy training grounds.

The SP training profiles were as follows: 4.5 ± 1.5 years training, 11 ± 1 months per year training, 4 ± 1 training sessions per week and 9.4 ± 2.4 h per week of training. This volume of exercise training had been consistent for the entirety of their active training years. SP played one competitive match per week and had been engaged in competitive soccer matches for 4 ± 2 years. For one club, 14 boys from the Under-12 (U12) squad and their parents were approached, of which three were not enrolled because of either personal circumstance (n = 2) or a soccer related injury (n = 1). At the second club, researchers provided information to 15 U12 players and their parents, of which 2 were recovering from injury, one was released from the club after signing up from the study, and one signed up and simply did not attend the testing. This resulted in 11 participants from both clubs, with a total of 22 SP volunteering for the study. CON participants took part in compulsory physical education of 2 h per week (the same as SP), were all recreationally active and without engagement in systematic training. The CON self-reported 1.53 ± 1.77 h per week of physical activity (Beaumont et al. 2019).

All participants underwent a physical examination and completed a medical history questionnaire. Exclusion criteria included the use of any medications that would influence cardiovascular function and any personal or early family history of cardiovascular disease. Written informed parental and participant consent was obtained prior to participation. All procedures performed in the study were in accordance with the ethical standards of the Declaration of Helsinki and the study was reviewed and approved by a local University ethics committee.

The study employed a prospective, cross-sectional, cohort assessment of cardiac structure and function in highly trained pre-adolescent SP and CON. All testing took place at the training grounds of the two soccer clubs and at a local school for the CON participants. Participants were instructed to refrain from exercise on the day preceding the test. Furthermore, all participants were also informed to refrain from consuming any drinks containing sugar or caffeine, as well as the consumption of any food in the two hours preceding the testing session.

Physical activity and training questionnaires (Rowland et al. 2009) were completed prior to the testing. Following this, stature, sitting height and body mass were measured. Maturity status was quantified using maturity offset (Sherar et al. 2005; Unnithan et al. 2018). Resting arterial blood pressure was recorded in the left arm by an automated blood pressure cuff (Boso, Medicus, Jungingen, Germany) and heart rate was assessed by a 12-lead electrocardiogram (ECG) (CardioExpress SL6, Spacelabs Healthcare, Washington US). Resting echocardiographic measurements were taken in the left lateral decubitus position. All SP (n = 22) and a sub-sample (n = 15) of the CON participants underwent an assessment of peak oxygen uptake (\(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\)) to reflect maximal aerobic capacity (Cortex MetaMax 3B, Cortex Biophysik GmbH, Leipzig, Germany) on a cycle ergometer (Lode, Corival, Groningen, Netherlands) (Unnithan et al. 2018).

Normality of data was assessed using the Shapiro–Wilk test. For normally distributed data, a Student’s independent t test was used to compare RV structure, function, ε and SR in SP and CON. TDI-A’ was not normally distributed and consequently a Mann Whitney U test was performed to evaluate inter-group differences. The relationship between RV morphology and function (ε and SR) and their relationship with \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) was determined with Pearson correlation coefficients. A sample size of 22 SP provided a (1–β) of 80% at an alpha level of 0.05. Cohen’s D effect sizes were calculated for all inter-group comparisons. All statistical data was analysed using SPSS version 23.0, Chicago IL), with statistical significance granted at p < 0.05.

Larger RVD₁, RVD₂ and RVD₃ were noted in the SP compared to the CON in both absolute terms and after scaling (Table 1). The RV:LV ratio was greater (p = 0.001) in the SP compared to the CON (Table 1). RVFAC was greater (p = 0.02) in the SP compared to the CON (Table 1). There were no significant differences with regard to peak tissue velocities adjusted for RV length [TDI-S’adj: (SP) 0.17 ± 0.02 vs. (CON) 0.17 ± 0.04 cm.s⁻¹.mm⁻¹], [TDI-E’adj: (SP) 0.21 ± 0.04 vs. (CON) 0.19 ± 0.04 cm.s⁻¹.mm⁻¹] and [TDI-A’adj: (SP) 0.11 ± 0.03 vs (CON) 0.11 ± 0.01 cm.s⁻¹.mm⁻¹].

There was no inter-group difference in peak global longitudinal ε. There was a longer (p = 0.04) time to peak for global longitudinal ε in the SP compared to CON (Table 2). Lower global SRS (p = 0.00007) and SRA (p = 0.005) respectively were also identified in the SP compared to CON (Table 2). Lower (p = 0.007) MW peak longitudinal ε was identified in the SP compared to the CON. Lower peak MW SRS (p = 0.0014) and MW SRA (p = 0.0039) were noted in SP compared to CON. Apical time to peak ε was longer (p = 0.02) in the SP compared to CON. Apical SRE was (p = 0.04) greater in SP compared to CON (Table 2). Temporal analyses across the cardiac cycle (Fig. 3) demonstrated lower (p < 0.05) mid-wall ε (see shaded areas in MID- Fig. 3) from Systole 50–95% and Diastole 10–60% in the SP compared to CON. Greater ε was noted at the apex in the SP (Fig. 3. Apical) and these approached significance at 35% diastole (p = 0.0628) and 40% diastole (p = 0.0656).

The morphological enlargements in RV cavity diameter noted in the SP in the present study are likely driven by repeated exposure to elevations in RV end-systolic wall stress from relative increases in pulmonary artery pressures (La Gerche et al. 2011). Additionally, persistently increased RV pre-load may also have contributed to the larger RV cavity dimensions observed in SP than CON. Previous evidence does exist with respect to morphological adaptations in pre-adolescent athletes, but this was noted in pre-adolescent swimmers (D’Ascenzi et al. 2017), where the horizontal body position of the sport would accentuate an increased pre-load and drive RV adaptation. In contrast to the present study, after allometric adjustments were made for body surface area, the observed RV enlargements in swimmers compared with controls were less obvious. The presence of morphological adaptations in the present study occurred in individuals exposed to soccer training (upright body position) and this would suggest that the intensity and volume of the training rather than any orthostatic stimulus could be the factor driving RV adaptation in these players (Qasem et al. 2018).

In the present study, the lack of difference in outflow tract dimensions coupled with a bigger basal inflow (RVD₁) dimensions in SP compared to CON mimics the pattern seen in adult endurance athletes (Oxborough et al. 2012b). This concurs with the type of adaptation anticipated from the training modes commonly used in repeat sprint type sports where there is a greater dynamic than static component in training and match-play (Utomi et al. 2015). Longitudinal data measured over the course of an 8-month competitive season in adult volleyball and basketball players demonstrated differential RV adaptation during the training season (D’Ascenzi et al. 2016) whereRVD₁ and RVD₂ dimensions increased, but RVOT did not. Taken together, these longitudinal data support cross-sectional data generated in the present study that RV inflow and mid cavity are more sensitive to a high-intensity training stimulus than RVOT. In line with the present investigation, unremarkable differences in RVOT consequent to a chronic pre-adolescent training stimulus in swimmers were noted with matched CON (D’Ascenzi et al. 2017).

Unequal remodelling (RV:LV) has previously been seen in adult endurance athletes (La Gerche et al. 2011; Oxborough et al. 2012b) (Arbab-Zadeh et al. 2014). In the present study, RV enlargement was disproportionate to that of the LV (Unnithan et al. 2018) in the SP compared to CON, based on the greater RV:LV ratio seen in SP. The increased RV: LV ratio in SP in conjunction with the recently reported unremarkable differences in LV chamber diameter between SP and CON (Beaumont et al. 2019) in the same cohort of SP and CON suggests that the major adaptation to soccer training would appear to occur in the RV rather than the LV. The bases for these changes may be a product of an enhanced pre-load and disproportionate wall stress-induced adaptations in the RV.

The present study is one of the first to explore functional differences with respect to training status in this population (D’Ascenzi et al. 2017). Conventional indices of global RV function such as RVFAC were larger in SP compared to CON. These findings oppose evidence in pre-adolescent swimmers that demonstrated no training-induced increases in RVFAC in response to 5-months of intensive swimming training, and lower RVFAC in the swimmers at baseline compared to age-matched controls (D’Ascenzi et al. 2017). It is possible to speculate that in the present study, the combination of a low HR at rest and/or RV enlargement in the SP required an increase in RV contractility to maintain an adequate resting cardiac output (Frank-Starling Law).

The lack of difference in peak global longitudinal strain between SP and CON is in accordance with previous findings in competitive-level pre-adolescent swimmers that demonstrated no differences vs age-matched control participants prior to the onset of training (D’Ascenzi et al. 2017). Similarly, evidence exists in adults of either reduced (La Gerche et al. 2012; King et al. 2013) or no difference in RV strain (Oxborough et al. 2012b; Utomi et al. 2015) when comparing athletic and non-athletic individuals. In concert, the lack of inter-group differences in ε and the superior RVFAC seen in the SP suggest that a direct measure of global longitudinal function (peak global longitudinal ε) is unaffected by physiological adaptation and further suggests a possible greater impact on circumferential shortening. Furthermore, RVFAC (similar to LV ejection fraction) is more dependent on ventricular load and chamber size and therefore our findings may also, in part, reflect this.

The present study extends existing work (D’Ascenzi et al. 2017) by incorporating segmental and temporal analysis of the RV FW in an adolescent athletic population. We observed that SP demonstrated lower MW shortening ε than CON, yet ε in basal and apical segments were unremarkable between groups. In comparison with existing work that assessed segmental RV strain in healthy adolescents (13 years) (Pieles et al. 2019) our values are lower at basal, MW and apex. These differences could be due to the vendor-related differences in 2D strain- derived indices (Nagata et al. 2015).

Such differences in mid-wall ε between SP and CON are evident not just at peak but throughout the latter half of systole. These data highlight an important and previously undocumented advancement in our current understanding of RV ε in the athletic heart, by evidencing that sole reporting of peak global mechanics may conceal region specific adaption. The physiological significance of these differences and similarities between trained and untrained hearts of pre-adolescents requires further investigation, yet could represent a contractile reserve important during exercise, and mimics a pattern seen within basal and mid-wall segments in adult endurance athletes (Teske et al. 2009). Peak global longitudinal ε was inversely associated with resting HR. This relationship was influenced by the significant inverse relationship noted between MW strain and HR, thereby creating more supportive evidence for a possible contractile reserve for the RV during exercise (La Gerche et al. 2011).

Similarly, interrogation of the SR segmental analyses identified the potential mechanism for globally reduced SRS in SP than CON. It would appear that the slower peak RV SRS was a product of a slower MW peak SRS in the SP, without differences in other RV FW segments between groups. Similar segmental SR findings were noted in the mid-wall segment in adult endurance athletes (Teske et al. 2009). It is possible to speculate that as the larger RVD₂ of the SP has a greater number of myofibrils, any given level of deformation will take a longer time period; hence a slower SR will result (Maciver and Townsend 2008). Consequently, wall tension and intraventricular pressure during systole will be generated or released at a slower speed. During diastole, however, segmental analyses allowed the further interrogation into the source of the slower peak RV SRA in the SP, which was principally determined by slower mid-wall SRA. This apparently greater late diastolic function in SP may be the consequence of superior SRE during the early phases of ventricular relaxation, specific to the apex. Indeed, temporal analysis of strain revealed that SP showed a trend (p = 0.06) towards greater lengthening at 35% diastole. This may be coupled with the enhanced SRE in SP than CON, driving the early diastolic pressure gradient between RA and RV.

The evidence of a significant relationship between \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) and global RV ε observed in the present study has also been demonstrated in the LV in healthy adolescents (Pieles et al. 2015) and elite soccer referees (Gianturco et al. 2017). There is only one study to date that demonstrated weak, but significant relationships between global RV function and \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) in adult team sport players (Lazic et al. 2019). There is no concomitant data in the RV for elite pre-adolescent athletes. Furthermore, significant positive relationships in the present study were also identified between \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) and MW and apical SRS, respectively at rest. MW and apical SRS at rest were also both lower in the SP compared to CON. The SRS data generated in the present study were obtained at rest. There is, however, evidence that RV SRS increases progressively in proportion to the increase in pulmonary artery pressures during exercise (La Gerche et al. 2012). The combined evidence from the cross-sectional analyses between the SP and CON ε (no inter-group differences or lower in SP) and SRS global and segmental findings (lower in SP compared to CON), their subsequent relationships with \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) (moderate positive relationships) and evidence from the extant literature (La Gerche et al. 2012), suggests that a contractile reserve may exist in the SP. This functional reserve may allow the SP to generate the intrinsic RV contractility during exercise required for the high aerobic capacity needed in elite youth soccer.

As with all cross-sectional studies, the morphological and functional differences identified in this study ascribed potentially to high-intensity soccer training, could also be a product of genetic self-selection. Consequently, there is a need for a longitudinal evaluation of the effect of high-intensity soccer training on changes in RV morphology and function in males and females, as well as athletes of different ethnicities. Furthermore, the use of the cycle ergometer to obtain the \(\dot{\mathrm{V}}{\mathrm{O}}_{2\mathrm{peak}}\) data in the current study may have slightly under-estimated the aerobic capacity and associated cardio-dynamic changes in the SP. This modality was selected as it allowed us to capture in-exercise left ventricular ε and Tissue-Doppler imaging without excessive upper body motion in the same cohort (Unnithan et al. 2018). The present study identified lower global SRS in the SP compared to CON, interrogation of ECG data will allow future studies to provide possible mechanistic insight into these differences. The lack of inter-group differences in global peak ε suggests that a direct measure of global longitudinal function is unaffected by physiological adaptation. In conjunction with the superior RVFAC seen in the SP, this may suggest that high-intensity training has a possible greater impact on circumferential rather than longitudinal shortening. Two dimensional echocardiography used in the present study is unable to determine circumferential strain. Consequently, future studies interrogating the circumferential strain response to high-intensity training in this population are warranted. The present study presents initial data that suggests this approach (global and segmental analyses) may facilitate a greater understanding of the normative values of recreationally and highly fit pre-adolescents. Our findings highlighted an increased RV size and altered function in trained male pre-adolescent soccer players. Essentially these data are important when considering normal variance in this population. Echocardiography is used as a front-line investigation in the assessment of cardiomyopathy. Therefore, when considering RV structure and function to exclude conditions such as arrhythmogenic right ventricular cardiomyopathy (ARVC), it is pertinent to take into account athletic status. Our findings of increased RV size may make the structural differentiation challenging, but significantly, RV function was not depressed as is defined in the functional criteria for ARVC.