Cardiac Structure and Function in Elite Female and Male Soccer Players

This cross-sectional study describes electrocardiographic and echocardiographic findings in healthy elite US soccer players.

Key Points

Question

What are normal electrocardiographic and echocardiographic findings in elite female and male US soccer players?

Findings

In this cross-sectional study of 238 elite US soccer players, we found that male athletes frequently displayed common training-related electrocardiography changes, whereas female athletes had significantly more abnormal electrocardiograms per the International Recommendations for Electrocardiographic Interpretation in Athletes. Both female and male athletes frequently exceeded American Society of Echocardiography normative value standards for basic echocardiographic parameters.

Meaning

Elite US soccer players frequently present with training-related electrocardiography patterns and echocardiographic parameters that are above guideline-defined normal ranges.

Abstract

Importance

Population-specific normative data are essential for the evaluation of competitive athletes. At present, there are limited data defining normal electrocardiographic (ECG) and echocardiographic values among elite US soccer players.

Objective

To describe ECG and echocardiographic findings in healthy elite US soccer players.

Design, Setting, and Participants

This cross-sectional study analyzed Fédération Internationale de Football Association–mandated screening sessions performed at US Soccer National Team training locations from January 2015 to December 2019. US women’s and men’s national team soccer players undergoing mandated cardiovascular screening were included.

Main Outcomes and Measures

Normal training-related and abnormal ECG findings were reported using the International Recommendations for Electrocardiographic Interpretation in Athletes. Echocardiographic measurements of structural and functional parameters relevant to cardiovascular remodeling were assessed relative to American Society of Echocardiography guideline–defined normal ranges.

Results

A total of 238 athletes (122 [51%] female; mean [SD] age, 20 [4] years; age range, 15-40 years) were included. Male athletes demonstrated a higher prevalence of normal training-related ECG findings, while female athletes were more likely to have abnormal ECG patterns (14 [11%] vs 0 in male cohort), largely accounted for by abnormal T-wave inversions. Echocardiography revealed no pathologic findings meeting criteria for sport restriction, but athletes frequently exceeded normal ranges for structural cardiac parameters responsive to exercise-induced remodeling including body surface area–indexed left ventricular (LV) mass (58 of 113 female athletes [51%] and 67 of 114 male athletes [59%]), indexed LV volume (89 of 115 female athletes [77%] and 76 of 111 male athletes [68%]), and LV wall thickness (37 of 122 female athletes [30%] and 47 of 116 male athletes [41%]). Age-stratified analysis revealed age-dependent increases in LV wall thickness, mass, and volumes among female athletes and LV wall thickness and mass among male athletes.

Conclusions and Relevance

These data represent the first set of comprehensive normative values for elite US soccer players and one of the largest sport-specific echocardiographic remodeling studies in female athletes. Abnormal ECG findings were more common in female athletes, while both female and male athletes frequently exceeded clinical normality cut points for remodeling-associated echocardiographic parameters.

Introduction

Soccer, internationally referred to as football, is the world’s most popular sport, and soccer players are increasingly encountered in clinical cardiovascular practice. Clinicians are commonly asked to assess soccer players with symptoms suggestive of cardiovascular disease and to oversee preparticipation screening of asymptomatic athletes. In response to numerous highly publicized cardiac arrest cases occurring during soccer matches,¹ the Fédération Internationale de Football Association (FIFA) mandates preparticipation screening with electrocardiography (ECG) and echocardiography prior to international competition.^2,3 It is anticipated that this practice will become more widely adopted during return to competitive soccer in the wake of the coronavirus disease 2019 (COVID-19) pandemic given concerns about cardiac complications of infection.^4,5,6

Exercise-induced cardiac remodeling is the process by which the cardiovascular system adapts to the hemodynamic stress imposed by athletic training.⁷ Determinants of the structural, functional, and electrophysiologic manifestations of exercise-induced cardiac remodeling include age,⁸ duration of sport exposure,⁹ sex,¹⁰ ethnicity,¹¹ and sport type.^12,13,14 Accurate differentiation of benign exercise-induced cardiovascular adaptations from high-risk pathology, a process that requires sport- and sex-specific normative data, underlies the successful clinical evaluation of symptomatic and asymptomatic athletes. To date, normative data integrating ECG and echocardiography among elite US soccer players, particularly women, have been lacking.

Accordingly, we conducted a prospective 5-year study of US national team soccer players participating in FIFA-mandated preparticipation screening with several discrete objectives. First, we sought to develop normative data characterizing ECG and echocardiographic findings among asymptomatic elite soccer players. Next, we aimed to determine the prevalence of abnormal ECG findings and their association with underlying pathology as determined by concomitant echocardiography. It is anticipated that the results of this study will provide valuable reference standards during future clinical evaluations among the large global population of competitive female and male soccer players.

Methods

This prospective cross-sectional study included female and male members of US Soccer national teams who underwent FIFA-mandated preparticipation screening prior to international World Cup competitions during a 5-year study period (January 2015-December 2019). Repeat evaluation was performed in a subset of individuals participating in multiple competitions. Screening was performed at US Soccer training sites by experienced clinicians affiliated with the Massachusetts General Hospital Cardiovascular Performance Program. The Mass General Brigham institutional review board approved this study and waived the need for informed consent based on use of deidentified data obtained during routine clinical care.

Data Acquisition

Demographic data including race was self-reported by participants. ECGs were performed by experienced cardiologists (T.W.C. and A.L.B.) using standard 12-lead placement and commercially available ECG software (Cardea 20/20 ECG; Cardiac Insight) after 5 to 10 minutes of quiet rest. Two-dimensional transthoracic echocardiography was performed using a commercially available system (Vivid-Q; GE Healthcare). Two-dimensional, pulsed-Doppler, and color tissue Doppler imaging from parasternal, apical, and subcostal transducer positions was performed with 3-beat acquisition.¹⁵ Two-dimensional frame rate of 50 to 75 seconds and tissue Doppler frame rate of more than 120 seconds were maintained for all images.

ECG Data Analysis

ECGs were reviewed in a digital printable format, which included athlete age, sex, and ethnicity by 1 of 2 readers (T.W.C. and B.J.P.) with experience in athlete ECG interpretation. Readers were blinded to all other athlete characteristics. ECGs were coded to define the presence/absence of all components of the International Recommendations for Electrocardiographic Interpretation in Athletes (henceforth referred to as the international criteria).¹⁶ ECGs were defined as abnormal if they contained 1 or more abnormal findings or 2 or more borderline findings.¹⁶ Quantitative measurements including heart rate, PR, QT, corrected QT intervals (calculated by the Bazett formula¹⁷), QRS duration, and P and R wave axis were calculated automatically and confirmed by visual inspection. Corrected QT intervals of 450 milliseconds or more were manually confirmed using the tangent method on the lead with the longest absolute QT interval.¹⁸

ECG definitions were as defined in the international criteria,¹⁶ with additional criteria as follows in cases where exact definitions were not present. Voltage criteria for left ventricular (LV) hypertrophy was quantified as a dichotomous variable based on either Romhilt Estes¹⁹ score more than 4 or fulfilment of Sokolow Lyon criteria²⁰ (S in V1 + R in V5 or V6 ≥ 35 mm). Right ventricular hypertrophy was defined based on the Sokolow Lyon criteria (R wave in V1 + S wave in V5 or V6 > 10.5 mm).²¹ Early repolarization was defined as a J point elevation of more than 1 mm in 2 leads in either lateral (V4-V6, I, aVL) and/or inferior territories (II, III, aVF).²² Sinus arrhythmia was defined as PP interval variation more than 120 milliseconds. T-wave inversions (TWI) were defined as 1 mm or more in depth in 2 or more continuous leads excluding III, aVR, and V1.

Echocardiographic Data Analysis

Echocardiographic measurements were performed by a single experienced echocardiographer (A.L.B.) on commercially available software (EchoPac, version 6.5; GE Healthcare). All measurements were made according to American Society of Echocardiography guidelines.²³ LV relative wall thickness (RWT) was calculated as 2 × (posterior wall thickness) / LV end-diastolic dimension. LV volumes and LV ejection fraction (LVEF) were measured using the modified Simpson biplane method of disks. LV mass was calculated using the area-length technique. Chamber dimensions and LV mass were indexed to body surface area, which was calculated using the Mosteller formula.²⁴ Tissue velocities were measured offline from 2-dimensional pulsed-wave Doppler images. Tissue and spectral Doppler measurements were performed in triplicate. As most individuals lacked sufficient tricuspid regurgitation for Doppler evaluation, pulmonary artery systolic pressure was estimated from pulmonary artery acceleration time.²⁵

Intraobserver variability analysis was performed through blinded reassessment of 2 key metrics (LV mass and LV end-diastolic volume [LVEDV]) for 10 randomly selected individuals from the female and male cohorts (n = 20). Correlation for each measurement was assessed using linear regression with the following results: LV mass (R² = 0.97) and LVEDV (R² = 0.97).

Statistical Analysis

Continuous variables are presented as mean (SD), with the range (minimum − maximum) included for ECG and echocardiographic measurements. Categorical variables are presented as number (%). Significance of differences across groups was assessed using t test, χ² testing, or 1-way analysis of variance. Distributions of important echocardiographic parameters were compared with sex-specific, general population reference values.²³ Two-sided P values had a significance threshold of .05.

Results

Study Population

A total of 238 athletes (122 female athletes [51.3%] and 116 male athletes [48.7%]) from the US National Soccer teams were included in this study (Table 1). The mean (SD) age was 20 (4) years (range, 15-40 years). Female athletes were predominantly White (87 [71%]) with fewer Black (23 [19%]) and Hispanic (8 [7%]) players, whereas male players were evenly divided between Black (40 [34%]), Hispanic (38 [33%]), and White individuals (37 [32%]). Compared with male soccer players, the female soccer players in this sample were older and had smaller body size (height, weight, body mass index, and body surface area).

Table 1. Clinical and Electrocardiographic Characteristics by Sex.

Clinical characteristic	Mean (SD)
	Female athletes (n = 122)	Male athletes (n = 116)
Age, y	21 (5.4)	18 (1.2)a
Height, in	66 (2.6)	71 (2.9)a
Weight, lb	136 (16.2)	160 (16.2)a
BMI	22 (1.8)	22 (1.6)a
Body surface area, m2	1.7 (0.1)	1.9 (0.1)a
Race/ethnicity, No. (%)
Hispanic	8 (7)	38 (33)a
Black	23 (19)	40 (34)a
White	87 (71)	37 (32)a
Otherb/unknown	4 (3)	1 (1)
ECG data
Heart rate, beats/min	56 (9)	58 (9)
Intervals, mean (SD) [range], ms
PR	154 (26) [104 to 244]	166 (24) [114 to 252]a
QRS	87 (8) [66 to 108]	94 (9) [72 to 122]a
QTc	415 (23) [361 to 481]	403 (21) [349 to 460]a
QRS axis, mean (SD) [range], °	73 (23) [−69 to 147]	75 (21) [−40 to 119]
International ECG criteria, No. (%)
Normal ECG findings
Increased QRS voltage for LVH	13 (11)	74 (64)a
Increased QRS voltage for RVH	1 (1)	17 (15)a
Incomplete right bundle branch block	5 (4)	15 (13)a
Early repolarization	35 (29)	97 (84)a
Black athlete repolarization pattern	3 (3)	7 (6)
Sinus bradycardia	85 (70)	74 (64)
Sinus arrhythmia	57 (47)	50 (43)
Ectopic atrial rhythm	5 (4)	13 (11)
Junctional rhythm	2 (2)	0
First degree AV block	8 (7)	12 (10)
Mobitz type 1 2° AV block	0	0
Borderline ECG findings
Left axis deviation	1 (1)	1 (1)
Left atrial enlargement	2 (2)	4 (3)
Right axis deviation	1 (1)	0
Right atrial enlargement	2 (2)	2 (2)
Complete right bundle branch block	0	1 (1)
Abnormal ECG findings
T-wave inversion	9 (7)	0a
ST-segment depression	1 (1)	0
Pathologic Q waves	2 (2)	0
Complete left bundle branch block	0	0
QRS ≥140 ms duration	0	0
Epsilon wave	0	0
Ventricular pre-excitation	0	0
Prolonged QT interval	1 (1)	0
Brugada type 1 pattern	0	0
Profound sinus bradycardia (<30 beats per min)	0	0
PR interval ≥400 ms	0	0
Mobitz type II 2° AV block	0	0
3° AV block	0	0
≥2 Premature ventricular contractions	0	0
Tachyarrhythmia
Atrial	0	0
Ventricular	0	0

Abbreviations: AV, atrioventricular; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); ECG, electrocardiography; LVH, left ventricular hypertrophy; RVH, right ventricular hypertrophy.

^aP < .05.

^bThe other category includes people who self-identified as Asian/Pacific Islander, >1 race/ethnicity, and/or indicated that the available options failed to adequately describe them.

ECG Findings

ECG findings including basic normative data and the prevalence of specific ECG findings as defined by the international criteria are presented in Table 1. Compared with male athlethes, female athletes had shorter PR (mean [95% CI], 154 [149-159] vs 166 [162-171] milliseconds; P < .001) and QRS intervals (mean [95% CI], 87 [85-88] vs 94 [93-96] milliseconds; P < .001) but longer corrected QT intervals (mean [95% CI], 415 [411-419] vs 403 [399-407] milliseconds; P < .001). Male athletes demonstrated a higher prevalence of normal training-related ECG findings including voltage criteria for LV hypertrophy (74 [64%] vs 13 [11%]; P < .001) and right ventricular hypertrophy (17 [15%] vs 1 [1%]; P < .001), incomplete right bundle branch block (15 [13%] vs 5 [4%]; P = .03), and early repolarization pattern (97 [84%] vs 35 [29%]; P < .001). The prevalence of abnormal ECGs was higher in female athletes than among male athletes (14 of 122 [11%] vs 0 of 116; P < .001). The most common abnormal ECG pattern was pathologic TWI (9 of 14 [69%]), which were most commonly isolated to the anterior leads (7 of 9) and less often involved anterolateral (V3-V6) and inferior (II, III, aVF) leads (2 of 9). Additional abnormal ECG findings included septal Q waves (2 of 14 [14.3%]), inferior ST-segment depressions (1 of 14 [7.1%]), and a mildly prolonged corrected QT interval (481 milliseconds; 1 of 14 [7.1%]).

Echocardiographic Findings

Echocardiographic data are presented in Table 2. Compared with female athletes, male athletes had greater maximum wall thickness (mean [95% CI], 10.0 [9.8-10.3] vs 8.9 [8.7-9.1] mm; P < .001), RWT (mean [95% CI], 0.38 [0.37-0.39] vs 0.36 [0.35-0.37]; P = .04), indexed LVEDV (mean [95% CI], 81 [79-84] vs 70 [68-72] mL/m²; P < .001), and indexed LV mass (mean [95% CI], 108 [105-111] vs 91 [88-93] g/m²; P < .001). Only 1 athlete (a 28-year old woman) had an LV end-diastolic dimension more than 60 mm, measuring 61 mm. As shown in Figure 1A, 37 of 122 female athletes (30%) and 47 of 116 male athletes (41%) exceeded American Society of Echocardiography upper limits of normal for LV wall thickness. LV wall thickness of 12 mm or more was seen in 14 male athletes (12%) but only in 1 female athlete (1%), and maximum LV wall thickness among all individuals was 13 mm (present in only 2 male athletes).

Table 2. Echocardiographic Findings by Sex.

Characteristic	Mean (SD) [range]		P value
	Female athletes (n = 122)	Male athletes (n = 116)
Structural
Interventricular septum, mm	8.3 (1.2) [6-12]	9.3 (1.3) [6-13]a	<.001
Posterior wall, mm	8.6 (1.2) [6-11]	9.8 (1.4) [7-13]a	<.001
Relative wall thickness	0.36 (0.06) [0.23-0.54]	0.38 (0.06) [0.26-0.56]a	.04
Maximum wall thickness, mm	8.9 (1.2) [6-12]	10 (1.3) [7-13]a	<.001
LV end-diastolic dimension, mm	48 (3.6) [39-61]	52 (3.7) [39-60]a	<.001
LV end-systolic dimension, mm	31 (3.0) [24-40]	34 (3.7) [25-45]a	<.001
LV outflow tract, mm	19 (1.7) [15-22]	21 (1.9) [16-28]a	<.001
Aortic root: sinuses of valsalva, mm	26 (2.5) [20-33]	29 (2.7) [23-36]a	<.001
Left atrium antero-posterior dimension, mm	33 (4.0) [22-44]	37 (3.6) [28-47]a	<.001
Right ventricular basal diameter, end-diastole, mm	38 (4.2) [27-51]	42 (4.4) [32-54]a	<.001
LV length, mm	83 (6.6) [48-101]	89 (5.7) [73-101]a	<.001
LV mass, g	155 (27) [105-225]	205 (34) [130-309]a	<.001
LV mass/BSA, g/cm2	91 (13) [64-122]	108 (16) [74-156]a	<.001
Volumes and ventricular function
LV end-diastolic volume, mL	120 (22) [74-180]	155 (26) [105-228]a	<.001
LV end-diastolic volume/BSA, mL/cm2	70 (11) [46-96]	81 (14) [58-127]a	<.001
LV end-systolic volume, mL	43 (10) [23-87]	58 (12) [34-98]a	<.001
LV ejection fraction, %	64 (5.5) [47-75]	63 (5.3) [49.6-77]a	.03
Spectral and tissue doppler
Mitral inflow velocity
Early (E wave), cm/s	95 (18) [57-133]	92 (17) [59-143]	.25
Late (A wave), cm/s	40 (10) [22-73]	42 (11) [20-68]	.09
E/A ratio	2.5 (0.72) [1.3-5.6]	2.3 (0.68) [1.3-6.0]a	.04
Lateral e′, cm/s	18 (2) [14-20]	21 (3) [17-27]a	<.001
Pulmonary artery acceleration time, ms	173 (20) [122-214]	166 (19) [121-210]a	.02
Estimated pulmonary artery systolic pressure, mm Hgb	26.1 (4.8) [18-41]	27.7 (4.8) [18-41]a	.02

Abbreviations: BSA, body surface area; LV, left ventricular.

^aP < .05.

^bEstimated pulmonary artery systolic pressure was calculated using pulmonary artery acceleration time according to published formula.²⁵

Figure 1. Left Ventricular (LV) Wall Thickness and Remodeling Patterns Among Female and Male Elite Soccer Players.

A, The distribution of maximal LV wall thickness is shown for female and male soccer players. Overall, 31% of female athletes and 41% of male athletes had wall thickness above the American Society of Echocardiography–defined normal range.²³ The maximal wall thickness observed was 1.3 cm. B, Patterns of LV remodeling in the study population are shown, stratified based on relative wall thickness and sex-specific LV mass index, according to guideline framework.²³ ULN indicates upper limit of normal.

Assessment of LV hypertrophy/remodeling geometry according to clinical guideline definitions²³ (Figure 1B) revealed that a majority of athletes had either eccentric hypertrophy (RWT ≤0.42 and LV mass index above sex-specific normal ranges; 48 of 113 female athletes [42%] and 46 of 114 male athletes [41%]) or normal LV geometry (RWT ≤0.42 and normal LV mass index; 50 of 113 female athletes [44%] and 40 of 114 male athletes [35%]). Concentric remodeling patterns (either concentric remodeling or concentric hypertrophy) were more common in male athletes (28 of 114 [25%]) than in female athletes (15 of 113 [13%]) (P < .05). Diastolic function was normal in all athletes, but female athletes had lower lateral LV early tissue Doppler velocity (E′) (mean [95% CI], 18 [17-19] vs 21 [20-22] cm/s; P < .001) than their male counterparts. None of the athletes with abnormal TWI (all female) had evidence of pathology by echocardiography, but compared with female athletes without TWI, this group had higher indexed LV mass (mean [95% CI], 102.0 [92.8-111.3] vs 90.1 [87.7-92.6] g/m²; P = .02) and indexed LVEDV (mean [95% CI], 77.5 [68.8-86.2] vs 69.2 [67.2-71.2] mL/m²; P = .03). Four players (2 male athletes and 2 female athletes) had bicuspid aortic valves, but none had clinically relevant valvular dysfunction or aortic enlargement. There were no cases identified of significant valvular regurgitation (≥moderate in severity) or valvular stenosis.

The distribution of echocardiographic parameters compared with clinical normative cut points is shown in Figure 2. A total of 58 of 113 female athletes (51%) and 97 of 114 male athletes (59%) demonstrated elevated body surface area–indexed LV mass, and an even higher proportion of athletes were found to have increased indexed LVEDV (89 of 115 female athletes [77%] and 76 of 111 male athletes [68%]). Increased RV basal diameter was more common in male than female athletes (60 of 112 [54%] vs 16 of 118 [14%]; P < .001). While the mean LVEF was in the normal range and slightly higher for female athletes (mean [95% CI], 64 [63-65] vs 63 [62-64]; P = .03), 17 of 231 athletes with measured LVEF (7.4%; 8 female and 9 male) had an LVEF less than 55%, with 9 below the guideline-defined normal ranges of less than 54% for female athletes (n = 7) and less than 52% for male athletes (n = 2). Selected LV structural parameters stratified by age groups are shown in Figure 3. Among female athletes, there was a significant increase in mean LV wall thickness, indexed LVEDV, and LV mass index with increasing age. In contrast, male athletes showed age-dependent increases in LV wall thickness and LV mass index but not indexed LVEDV.

Figure 2. Structural and Functional Parameters Among Female and Male Elite Soccer Players.

Sex-specific distributions of 6 key structural and functional echocardiographic parameters are shown for members of the US Women’s and Men’s national soccer teams, along with American Society of Echocardiography–defined normal ranges and the percentage falling within and outside the normal range. BSA indicates body surface area; LA A/P, left atrial antero-posterior dimension; LV, left ventricular; LVEDD, left ventricular end-diastolic dimension; LVM, left ventricular mass; RV, right ventricular; RVEDD, right ventricular end-diastolic dimension.

Figure 3. Differences in Key Left Ventricular (LV) Structural Parameters Across Different Age Groups.

Three key structural echocardiographic parameters tightly linked to the process of exercise-induced cardiac remodeling are shown stratified by sex and age group in this sample of elite US soccer players. Values are presented as mean (SD) and percentage of individuals meeting criteria for abnormal measurements (% abnormal). Senior Men’s national team did not participate in screening program during study period. LA indicates left axis; RA, right axis; RV, right ventricular.

A total of 34 individuals (24 female, 8 male) underwent repeat evaluation after a mean (SD) of 2.4 (0.9) years. Within individuals, LV wall thickness and LV mass both increased by a mean (SD) of 0.08 (0.13) cm (95% CI, 0.03-0.12; P = .001) and 9.6 (14) g (95% CI, 5.0-14.2; P < .001), respectively, while LV volumes remained stable. Of this group, 4 of 34 (all female) athletes (7.1%) had ECG abnormalities on initial evaluation (3 with abnormal TWI and 1 meeting criteria for 2 borderline findings), all of whom had normal ECG findings on subsequent evaluation. Echocardiograms performed at the time of the initial abnormal ECG findings were only notable for increased LV mass above ASE normality thresholds in 3 of 4 and increased indexed LVEDV in 2 of 4.

Discussion

This study provides the first normative data set integrating ECG and echocardiography derived from a sex- and race-balanced group of elite United States soccer players. Importantly, our data represent one of the largest sport-specific studies of female athletes to date. Key findings can be summarized as follows. First, normal training-related ECG changes are highly prevalent in US male soccer players but are less common among female athletes. In contrast, female soccer players are more likely to demonstrate potentially pathologic ECG findings, which are most commonly accounted for by pathologic anterior precordial TWIs. Independent of ECG findings, female and male soccer players demonstrate a high prevalence of age-dependent structural LV parameters (ie, LV wall thickness, chamber volume, and mass) that exceed clinical cut points for normality. In aggregate, these data define normative ECG and echocardiographic parameters among healthy elite adolescent and young adult US soccer players that can be applied in clinical practice.

Normative cardiac parameter reference ranges in athletes are essential for differentiating training-related cardiovascular adaptations from potentially pathologic phenotypes. This process, the basis for determining athletic eligibility and for providing guideline-directed medical therapy for those with confirmed pathology, represents a common and often complex challenge for cardiovascular and sports medicine practitioners. Consensus documents delineating the use and interpretation of ECG¹⁶ and multimodality imaging among athletes²⁶ acknowledge both the critical importance and the current dearth of sport- and sex-specific normative data for this purpose. With important exceptions derived from men playing in the National Football League^27,28 and among women and men participating in the National Basketball Association,^12,13,29 prior series defining normal reference ranges in competitive athletes have comprised mixed sport cohorts and have focused almost exclusively on men. Results from the current study add to the existing literature by providing sex-specific ECG and echocardiographic data derived from elite athletes participating in the world’s most popular competitive sport.

Several insights regarding our ECG data are noteworthy. Elite soccer players of both sexes demonstrated a high prevalence of training-related ECG patterns. This finding underscores the value of this classification strategy as a way to identify adaptive ECG findings while minimizing the burden of false-positive testing. The overall rate of overtly abnormal ECG findings in this current study (5.9%), as defined by the most recent international criteria, was similar to previous cohorts of trained athletes from other sports (1.5%-7.5%).^{30,31,32,33,34} However, we found a notably higher prevalence of female athletes with abnormal ECG findings (11.5%), which was largely attributable to pathologic TWI confined to the anterior precordial lead distribution (V1-V3). Notably, the prevalence of anterior TWI in female athletes was similar in our cohort to a recent study of 895 female athletes from multiple sporting disciplines³⁵ (5.7% vs 6.5%, respectively), and follow-up testing for female athletes in both cohorts did not reveal explanatory cardiac pathology. Mechanisms underlying the high prevalence of anterior TWI among women remain uncertain, but prior hypotheses have included variation in sympathetic innervation and anatomical differences including increased breast tissue in female athletes.³⁵ At present, the international criteria define TWI confined to anterior precordial leads as a normal juvenile ECG pattern in adolescent White athletes younger than 16 years.^16,36 While anterior TWI have previously been shown to be more common in young female than male athletes,³⁵ our data suggest that they may represent a benign finding across a broader age range of female soccer players. Confirmation of this finding among similar cohorts of elite female athletes represents an important area of future work. Finally, a unique aspect of this study was the opportunity to examine serial cardiac data among athletes who attended multiple training camps. In this small but instructive subgroup, we identified 4 female athletes with abnormal ECG findings on initial evaluation that normalized on repeat testing 2 to 4 years later; notably 2 of these had lateral precordial (V4-V6) TWI, a finding previously shown to have variable but significant associations with underlying myopathic diagnoses.^37,38,39,40 Each of these athletes had normal corollary echocardiograms at both study points that only were noteworthy for elevated LV mass and volumes. In aggregate, this observation highlights the underappreciated plasticity of the athlete’s ECG despite the apparent absence of underlying pathology.

Echocardiographic data obtained in this study are similarly instructive and clinically relevant. Among a relatively young population of individuals (63% younger than 19 years), we observed a high prevalence of cardiac structural indices that exceeded clinical normality cut points. This finding is consistent with prior data derived from soccer athletes from primarily male cohorts^{41,42,43,44,45,46,47,48} and underscores the limitations of applying general population-based normative data to elite athlete populations. When compared with other sport-specific data, elite female and male soccer players had smaller unadjusted average chamber dimensions than professional basketball^12,13 and American-style football athletes,²⁷ an unsurprising finding given associations of LV chamber diameter with body size. Wall thickness was also slightly lower in the soccer athletes, making clear again the importance of sport-specific normative data. When compared against 2 larger, multisport samples of female athletes (n = 600⁴⁹ and n = 439¹⁰), elite female soccer players had similar LV dimensions (mean [SD], 49 [4] vs 48 [3.6] and 49 [4] mm in the 2 larger studies, respectively) and wall thickness (mean [SD], 8.2 [0.9] vs 8.9 [1.2] and 8.7 [1.2] mm), although it is difficult to draw firm conclusions here given the diversity of the larger samples. Finally, although cross-sectional, our data also suggest that optimal normative data should account for athlete age. The finding of incremental, age-dependent increases in LV mass, wall thickness, and chamber volumes in this adolescent and young adult population suggest that exercise-induced remodeling is a progressive process that continues throughout the duration of an athletic career. Future prospective, repeated-measures studies will be required to clarify the temporal nature and clinical implications of long-term exposure to elite athletics.

The normative reference data derived from this study are of particular clinical relevance in the setting of the COVID-19 pandemic. Emerging data suggest that individuals infected with COVID-19 are at risk for cardiovascular involvement including myocarditis.^50,51 Accordingly, multiple expert consensus return-to-sport testing algorithms that rely heavily on the use of ECG and echocardiography have been proposed.^4,5,6 In this context, we anticipate that preparticipation screening incorporating these modalities will occur with increasing frequency. The availability of comprehensive normative data derived from healthy athletes in the pre–COVID-19 era will serve as valuable frame of reference to differentiate athletes with clinically relevant COVID-19 complications from those with benign exercise-induced adaptation.

Limitations

We acknowledge several important limitations of this study. First, this study contains the largest database of elite US soccer players and is among the largest cohorts of single-sport female athletes, a group previously underrepresented in the literature.⁵² However, our sample size, deliberately confined to US National Team players preparing for World Cup play, may be underpowered to permit conclusive subgroup analyses. We further acknowledge the relative dearth of older men in this cohort, which was attributable to the US men’s senior national team’s failure to qualify for World Cup play during the study period. Third, individuals with ECG abnormalities but normal echocardiograms did not undergo further noninvasive imaging (ie, cardiac magnetic resonance imaging) as echocardiographic imaging was deemed to be of sufficient quality to exclude pathology in all cases. Similarly, the small group of athletes (7%) with mildly reduced LVEF did not undergo subsequent testing to assess the LV response to exercise because they were completely asymptomatic and demonstrated robust LV diastolic function. In both cases, we nonetheless acknowledge the small possibility of undetected occult pathology. Fourth, we note that normative data derived from elite athletes may be incompletely generalizable to athletes at subelite competition levels. However, cardiac parameters characteristic of healthy elite competitors likely represent the upper limits of adaptive remodeling among same-sport athletes at lower levels of competition. Finally, serial cardiac data were only available in a limited subset of athletes, rendering conclusions about the evolution and plasticity of ECG and echocardiographic data in this population preliminary.

Conclusions

In conclusion, this study provides the first comprehensive normative data set of ECG and echocardiographic data derived from elite female and male US soccer players. Our data demonstrate a high prevalence of benign training-related ECG findings and identify several specific abnormal ECG patterns with suboptimal specificity for true underlying pathology. Similar to other cohorts of elite athletes, our data confirm an elevated prevalence of adaptative cardiac structural remodeling findings that exceed general population-based clinical imaging cut points. We anticipate that the data presented in this article will serve as a valuable reference during future clinical assessments of elite female and male soccer players.