Frequency and Prognostic Significance of T-wave Inversions in Athletes: A Systematic Review and Meta-analysis

Abstract

Background

T-wave inversions (TWI) are a challenging finding in athletes' electrocardiograms. This study aimed to determine the frequency and prognosis of TWI in athletes.

Methods

A systematic review and meta-analysis were conducted following PRISMA guidelines and registered in PROSPERO (CRD42023494038). Data sources were PubMed, Scopus, and Web of Science. We included original research articles reporting on the prevalence or underlying substrates of TWI and providing data on athletes participating in any type of sports, sex, or race. Risk of bias was evaluated using the Joanna Briggs Institute Critical Appraisal Tool for Cross-Sectional Studies. Random-effects models were used. Sensitivity and subgroup analyses (sex, age, ethnicity, diagnostic criteria) were predefined.

Results

Of 4,509 articles identified, 115 were selected (258,954 individuals). Substantial heterogeneity was observed among the study results. The global TWI prevalence in athletes > 16 years according to the International criteria was 4% (3%, 1%, and 0% in anterior, inferior, and lateral leads, respectively). Prevalence was higher in Black athletes. TWI was more frequent in athletes than non-athletes (prevalence ratio=1.49; 95% confidence interval: 1.13–1.97). Cardiomyopathy diagnosis among athletes with TWI showed high variability, but in most studies, prevalence was <1%. Acute cardiovascular event incidence was low in most of the studies with high heterogeneity.

Conclusions

TWI is more frequent in athletes than non-athletes but remains low overall. While it may reflect physiological exercise related ventricular remodelling, thorough evaluation is essential to exclude structural heart disease. The risk of cardiovascular events is uncertain due to low statistical power and high heterogeneity.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s40798-026-01115-9.

Key Points

• The prevalence of TWI in athletes is 4% (International-Seattle criteria), and it is more common in Black athletes. This pooled-point estimate should be interpreted with caution due to high heterogeneity between studies.

• TWI are more frequent in athletes than in the general population (prevalence ratio: 1.49; 95% confidence interval: 1.13-1.97).

• The prevalence of structural cardiac disease among athletes with TWI varies across studies.

• The incidence of clinical cardiovascular events in athletes with TWI is generally low; however, the risk in this population remains uncertain because of the limited statistical power and substantial heterogeneity between studies.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s40798-026-01115-9.

Introduction

Sudden cardiac death (SCD) is a rare but striking event in young athletes, most often associated with underlying structural heart diseases or cardiac channelopathies [1]. These events have a great impact due to their occurrence in individuals who are otherwise perceived as healthy. As a preventive strategy, pre-participation cardiovascular screening is recommended to detect potentially life-threatening cardiac abnormalities in athletes [2], although this practice remains controversial [3].

The electrocardiogram (ECG) is among the primary tests used in this initial screening [4]. It is cost-effective, widely accessible, and valuable for detecting both structural and electrical cardiac alterations, though its routine use is still debated [5–7]. Interpreting the ECG of athletes is particularly challenging because physiological cardiac adaptations may closely mimic pathological findings [5]. These adaptations include increased ventricular mass, enlarged chamber dimensions, greater wall thickness, as well as autonomic changes related to increased vagal tone or reduced sympathetic activity [7]. Several criteria have been developed to interpret the ECG of athletes, including those from the European Society of Cardiology (ESC) [8], the Seattle [9], and more recently the International criteria [5, 10].

T-wave inversion (TWI) is a relatively common repolarisation abnormality observed in athletes [11]. However, the presence of TWI is considered an abnormal ECG finding, with the exception of ST elevation followed by TWI in V1–V4 in black athletes and TWI in V1–V3 in athletes <16 years old [10]. TWI poses a diagnostic challenge [11], as it can reflect either benign ventricular remodelling associated with intense training or an early manifestation of cardiomyopathies that predispose athletes to SCD [12–14]. To aid in distinguishing physiological from pathological TWI, various ECG parameters have been proposed [5, 11, 15], such as J-point elevation, ST-segment morphology, the anatomical location of TWI, and the depth of inversion.

The aims of this systematic review were: (i) to determine the prevalence of TWI in athletes and whether it is higher than in the general population; (ii) to identify structural heart diseases associated with TWI; and (iii) to evaluate its prognostic significance in this population.

Methods

Design of the Study

A systematic review and meta-analysis were conducted in accordance with the Cochrane Handbook for Systematic Reviews of Interventions, version 5.1.0 [16], and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [17]. The protocol was prospectively registered in the International Prospective Register of Systematic Reviews (PROSPERO) under registration number CRD42023494038. A specific protocol was prepared, and no amendments were made to the information provided at registration or within the protocol.

PICOS Format of the Research Questions

Three research questions were developed using the Population, Intervention, Comparator, Outcome, and Study Design (PICOS) framework (Supplementary Table 1) and are summarised as follows:

• Question 1: a. What is the prevalence of TWI in athletes?; b. Is the prevalence of TWI higher in athletes compared to non-athletes?

• Question 2: Is TWI in athletes associated with underlying structural cardiac disease?

• Question 3: Does the presence of TWI in athletes confer an increased risk of cardiovascular events or SCD?

Search Strategy

A systematic literature search was conducted in three electronic databases: PubMed, Scopus, and Web of Science. Studies published from the inception of each database up to September 2023 were considered for inclusion. An updated search was performed in PubMed to capture additional studies published between September 2023 and June 2025. The search strategy was structured around two main conceptual domains, with terms mapped to relevant Medical Subject Headings (MeSH):

1. Repolarisation abnormalities: “Repolarization”, “T-wave”, “ST segment”, “Abnormal ECG”, “Abnormal EKG”, “Electrocardiographic abnormalities”, “Electrocardiographic patterns”, “Electrocardiographical abnormalities”, and “Electrocardiographical patterns” (MeSH term: Electrocardiography, Ambulatory).

2. Athlete population: “Athletes”, “Sport”, and “Athlete’s heart” (MeSH terms: Athletes, Sport, Cardiomegaly, Exercise-Induced).

Within each conceptual group, search terms were combined using the Boolean operator “OR”, and the two conceptual groups were linked using “AND” to construct the final search strategy.

Eligibility Criteria and Study Selection

Eligibility criteria included: (i) original research articles reporting on the prevalence or underlying substrates of TWI; (ii) studies providing data specifically on athletes, with or without comparisons to non-athlete populations; (iii) inclusion of male and/or female participants; and (iv) inclusion of participants of any race, with particular attention to Black and Caucasian populations; (v) participation in any type of sport, at any level; and (vi) no age restrictions. Studies were also included if they reported diagnostic procedures used to rule out structural cardiomyopathies or provided follow-up data on athletes with TWI to assess the risk of cardiovascular events.

Exclusion criteria encompassed studies focusing on arrhythmias, prolonged or shortened QT intervals, other channelopathies, atrioventricular conduction abnormalities, ventricular hypertrophy, or those solely dedicated to morphological assessments (e.g., echocardiography, cardiac magnetic resonance [CMR] imaging) of the athlete’s heart.

Study selection followed a three-step screening process: (1) title screening, (2) abstract screening, and (3) full-text review of potentially eligible articles. In those studies analysing the same set of athletes in more than one publication, the study with the largest sample size or the longest follow-up was selected. Two independent reviewers (JS and RE) assessed each study. Discrepancies were resolved through discussion and consensus with a third reviewer (JC-B).

Additionally, reference lists of all included studies were screened to identify any relevant articles not retrieved through the initial search strategy.

Variables of Interest and Data Extraction

From each included study, two authors independently extracted the following variables of interest: first author; year of publication; country; study design; sample size; participant characteristics (sex, race, age range, and details of sports participation); prevalence of TWI in athletes and, if applicable, in control groups; TWI location (anterior, inferior, or lateral); criteria used to define TWI (e.g., ESC, Seattle, International, or other); data on associated structural cardiac abnormalities; follow-up duration; and incidence of cardiovascular outcomes during follow-up. In the event of discrepancies between the two reviewers (JS and RE), a third reviewer (JC-B) was consulted to reach consensus.

TWI was defined according to one of the following accepted criteria: (i) the 2010 ESC guidelines define TWI as a negative T wave ≥2 mm in ≥2 contiguous leads, excluding leads III and aVR [8]; (ii) the Seattle and International criteria define TWI as a negative T wave ≥1 mm in 2 contiguous leads, excluding leads aVR, II and V1 [5, 9, 10]. We used the International-Seattle Criteria as the primary criteria for this analysis. The location of TWI was classified based on the affected electrocardiographic leads as follows: (i) anterior (V2–V4); (ii) inferior (II–aVF); and (iii) lateral (V5–V6 and/or I–aVL). Studies presenting results on the infero-lateral leads were classified as on the inferior leads. In athletes ≤ 16 years, the presence of TWI in V2 or V2–V3 was considered juvenile TWI pattern and was analysed independently.

Risk-of-Bias Assessment

To assess the risk of bias in the studies included in the meta-analysis, we used the Joanna Briggs Institute (JBI) Critical Appraisal Tool for Analytical Cross-Sectional Studies [18]. This tool evaluates key methodological domains, including the clarity of inclusion criteria; the definition of the study setting and participant recruitment; the validity and reliability of exposure and outcome measurements; the identification and management of potential confounding factors; and the appropriateness of the statistical analyses employed. Each domain was rated as adequate, not adequate, unclear, or not applicable (for non-analytical, descriptive studies). A global index was assigned: adequate when all the domains were adequate, unclear when at least one domain was unclear and the rest were adequate, and non-adequate when at least one domain was not adequate.

Statistical Analysis

The meta-analyses were conducted using RStudio and the R package 'meta'. In studies reporting the prevalence of TWI in both athletes and control groups, the SE of the difference in prevalence between the two groups was calculated using the following formula:

$$\sqrt{\frac{p1(1-p1)}{n1}+\frac{p2(1-p2)}{n2},}$$

where p1=prevalence in athletes, n1=sample size of athletes, p2=prevalence in controls, n2=sample size of controls.

We applied the Freeman–Tukey double arcsine transformation to the observed proportions in the selected studies. This transformation is one of the recommended methods when the distribution of the observed proportions is skewed or when the observed proportions are below 0.2 or above 0.8. The pooled results were back-transformed and are presented as proportions for ease of interpretation. We also applied a continuity correction to studies with zero events, assuming a raw prevalence of 0.5% in these studies. The classic DerSimonian–Laird random-effects model was used to estimate the pooled prevalence.

Homogeneity statistics were used to assess the degree of unexplained variability in effect estimates across studies, commonly referred to as between-study heterogeneity. The I² statistic was used to quantify this heterogeneity and estimates the proportion of total variation attributable to true heterogeneity rather than sampling error. If heterogeneity was moderate or high (I² ≥ 50%), meta-regression analyses were performed to identify potential variables that might explain this heterogeneity. Initially, the following variables were considered individually: year of publication, region/country, study sample size, mean age of participants, proportion of men, proportion of Black athletes, type of screening, type of sport, and type of athlete. Subsequently, a multivariable meta-regression analysis was performed including all variables that individually explained at least some of the observed heterogeneity. Bubble plots were used to visualise the effect of the variables on between-study heterogeneity.

Finally, the estimated pooled prevalence of TWI was stratified according to variables already known to influence its prevalence, specifically age (≤ 16 years and >16 years) and ethnicity. We also performed stratified analyses according to sex and the TWI diagnostic criteria applied. For studies reporting TWI prevalence in athletes aged ≤18 years, these data were included in the ≤16 years age stratum.

Although standard methods have limited utility for assessing potential publication bias in meta-analyses of proportions, we explored this potential bias using funnel plots and Egger's regression asymmetry test, with statistical significance set at p < 0.05.

In studies evaluating the incidence of cardiovascular events during follow-up, annualised incidence rates were calculated to standardise differences in follow-up duration across studies. Specifically, the cumulative incidence was divided by the mean or median follow-up time reported in each study, to estimate annual incidence rates per 1,000 athletes.

Equity, Diversity, Inclusion Statement and Data Availability

The collection and analysis of the data were stratified by sex and ethnicity. The author team is balanced by gender, and one of the co-authors is a medical student who contributed to this work as part of his master’s thesis. Data extracted from included studies, data used for all analyses, analytic code, any other materials used in the review are available upon request to the corresponding author.

Results

Literature Search

The initial literature search identified 5963 potentially relevant articles, of which 1454 were found to be duplicates. After applying the selection criteria and adding 10 articles obtained through other sources, 107 articles were retained for analysis. Additionally, 8 articles were also included after the additional search to June 2025 was performed, resulting in a final total of 115 studies included in the analysis (Fig. 1).

Fig. 1.

Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) flow diagram. TWI T-wave inversion

Prevalence of TWI in Athletes (Question 1-a)

To address the question of the prevalence of TWI in athletes, 110 articles comprising a total of 258,954 individuals were included in the analysis [4,12-15,19–123]. The main characteristics of the studies are summarised in Supplementary Table 2. Twelve studies were not included in the meta-analyses as the criteria to define TWI were not standardized or properly defined [19, 20, 74, 81, 97, 98, 100, 106, 109, 111, 115, 119].

Prevalence of TWI in Athletes > 16 Years

Forty-two studies used the International-Seattle criteria [12, 31, 34, 44, 47, 48, 50, 52, 53, 56–58, 60, 62, 65, 66, 68, 73, 79, 83, 84, 86–89, 91–95, 104, 107, 110, 113, 114, 116–118, 120–123] and 32 studies applied the ESC criteria [4, 13, 14, 25, 26, 29–31, 35, 36, 38–42, 44, 45, 49, 66, 68, 69, 75–77, 80, 83, 101–103, 105, 108, 110, 112]. The heterogeneity between studies was very high; in most meta-analyses the I² was > 80%.

The overall prevalence of TWI in athletes >16 years according to the International-Seattle criteria was 4% (95% confidence interval -CI 3–6%). Heterogeneity was very high with prevalence ranging from 0 to 42%. Prevalence by anatomical location was 3% (95%CI 2–4%), 1% (95%CI 0–2%) and 0% (95%CI 0–1%) in the anterior, inferior and lateral leads, respectively (Figs. 2 and 3). Heterogeneity was very high with prevalence ranging from 0 to 20%, 0 to 7%, and 0 to 6% in the anterior, inferior and lateral leads, respectively.

Fig. 2.

Forest plots displaying the pooled prevalence of T-wave inversion (TWI) across all leads in athletes aged >16 years following the International-Seattle criteria for ECG interpretation. CI confidence interval

Fig. 3.

Forest plots displaying the pooled prevalence of T-wave inversion (TWI) in athletes aged >16 years following the International-Seattle criteria for ECG interpretation in: A anterior leads (V2–V4); B inferior leads (II, aVF); C lateral leads (I, aVL, V5-V6). CI confidence Interval

Funnel plots are presented in Supplementary Figure 1, and Egger's test for funnel plot asymmetry suggested the presence of publication bias or substantial heterogeneity among studies (p-value<0.05).

The prevalence of TWI was higher among Black athletes than in Caucasian athletes, both overall (13% vs 3%) (Fig. 4), and across all leads (Supplementary Figs. 2–3). In the analysis stratified by sex, women showed a higher overall prevalence compared to men (7% vs. 5%) (Fig. 5). This higher prevalence was observed in the anterior and inferior leads but not in the lateral leads (Supplementary Figs. 4–5).

Fig. 4.

Forest plots showing the prevalence of T-wave inversion (TWI) in Caucasian and Black athletes aged >16 years, using the International-Seattle criteria for ECG interpretation: A Caucasian athletes; B Black athletes. CI confidence Interval

Fig. 5.

Forest plots showing the prevalence of T-wave inversion (TWI) in male and female athletes aged >16 years, using the International-Seattle criteria for ECG interpretation. A Male athletes. B Female athletes. CI confidence interval

In the meta-regression analyses the only variable partially explaining the heterogeneity between studies was age: the higher the age of the athletes included in the studies, the higher the prevalence of TWI (Supplementary Fig. 6).

When the ESC criteria were applied, the overall prevalence of TWI was similar, at 4% (Supplementary Fig. 7). Prevalence by anatomical location was 3%, 1% and 0% in the anterior, inferior and lateral leads, respectively.

Prevalence of TWI in Athletes ≤ 16 Years

To analyse the prevalence of TWI in young athletes (≤ 16 years/18 years), 29 studies were included [15, 23, 27, 28, 32, 33, 37, 43, 45, 46, 51, 52, 54–56, 61, 63, 64, 66–72, 85, 90, 96, 99]. Of these, 19 studies reported the prevalence of juvenile TWI (V1–V3), which was 6%. Additionally, 17 studies provided data on TWI in other specific locations [23, 28, 32, 33, 37, 43, 45, 46, 54, 61, 64, 67, 70–72, 85, 90], showing prevalences of 2% in the extended anterior leads, 1% in the inferior leads, and 0% in the lateral leads (Supplementary Fig. 8).

Prevalence of TWI in Athletes vs Non-athletes (Question 1-b)

For this analysis, the diagnostic criteria for TWI used in each study (International-Seattle, ESC, or other) were considered. A total of 14 studies were included in this analysis [13, 21–23, 27, 28, 31, 34, 37, 42, 53, 54, 70, 102].

The presence of TWI in any lead was more frequent in athletes compared to non-athletes, with a prevalence ratio (PR) of 1.49 (95% CI 1.13–1.97) (Fig. 6-A). Specifically, TWI in the inferior leads was more prevalent among athletes (PR=1.70; 95% CI 1.23–2.35). The differences in TWI prevalence in the anterior and lateral leads were not statistically significant but the direction of the association was consistent and meaningful (Fig. 6). A sensitivity analysis excluding studies in which the prevalence of TWI was 0% in either athletes or non-athletes yielded consistent results (data not shown).

Fig. 6.

Forest plots showing the prevalence ratio of T-wave inversion in athletes versus non-athletes, aged >16 years, based on lead distribution and the criteria for ECG interpretation used in each study: A all leads combined; B anterior leads (V2–V4); C inferior leads (II, aVF); D lateral leads (I, aVL, V5–V6). PR prevalence ratio; CI confidence interval

Risk-of-Bias Assessment Question 1

The qualitative assessment of risk of bias is presented in Supplementary Table 3. Most concerns were identified in the first two domains, which relate to the definition of the sample selection criteria and the study setting. Among the included studies, 53 were judged adequate, 31 showed some uncertainty, and 26 were considered inadequate.

Relationship Between TWI with Structural Cardiac Diseases in Athletes (Question 2)

Thirty-four studies assessed the presence of structural cardiomyopathies and included at least 30 athletes with TWI (Table 1). In 19 of these studies [13, 28, 31, 34, 37, 42, 44, 46, 50, 53, 57, 61, 66, 67, 75, 78, 83, 87, 103], the prevalence of structural cardiac disease among athletes with TWI was <1%. Nine studies [4, 15, 32, 41, 64, 101, 121, 122, 124] reported a prevalence between 1 and 10%, while six studies [14, 26, 72, 125–127] reported a prevalence exceeding 10%. The most frequently identified structural cardiomyopathies were hypertrophic cardiomyopathy, arrhythmogenic ventricular cardiomyopathy, and dilated cardiomyopathy. Notably, TWI in the lateral leads was most frequently associated with a confirmed diagnosis of cardiomyopathy.

Table 1.

Main characteristics of the studies assessing the prevalence of cardiomyopathies in adult athletes with T-wave inversion and principal findings

Study	Year	Country	N TWI	N diagnosed CM	Prevalence (%)	Dx test	Diagnosis
Studies using echocardiogram
Magalski et al. [75]	2008	USA	35	0	0	Eco
Papadakis et al. [28]	2009	UK	67	0	0	Eco
Chandra et al. [42]	2014	UK	429	0	0	Eco
Berge et al. [44]	2015	Norway	70	0	0	Eco
Calò et al. [46]	2015	Italy	136	0	0	Eco
Ghani et al. [50]	2016	UK	42	0	0	Eco
Waase et al. [87]	2018	USA	32	0	0	Eco
Pambo et al. [66]	2021	Ghana	63	0	0	Eco
Halasz et al. [67]	2021	Italy	202	0	0	Eco
Calò et al. [124]	2019	Italy	136	2	1.47	Eco	2 HCM
Migliore et al. [32]	2012	Italy	158	4	2.53	Eco	3 ACM + 1 HCM
Gati et al. [41]	2013	UK	100	3	3.00	Eco	3 HCM
Wilson et al. [101]	2012	Qatar	83	3	3.61	Eco	1 ACM + 2 HCM
Pelliccia et al. [26]	2008	Italy	123	39	31.71	Eco	17 HCM + 6 MP + 4 ACM + 4 HTN + 3 DCM + 3 AoE + 1 ICH + 1 Myocarditis

Studies using echocardiogram and magnetic nuclear resonance as diagnostic tests
Le et al. [78]	2010	USA	44	0	0	Eco + MNR
Papadakis et al. [31]	2011	France	272	0	0	Eco + MNR
Brosnan et al. [103]	2014	Australia	97	0	0	Eco + MNR
Malhotra et al. [53]	2017	UK	103	0	0	Eco + MNR
Finocchiaro et al. [57]	2017	UK	108	0	0	Eco + MNR
Zaidi et al. [34]	2013	UK	66	0	0	Eco + MNR
Sheikh et al. [37]	2013	UK	116	0	0	Eco + MNR
D'Ascenzi et al. [61]	2019	Italy	354	1	0.28	Eco + MNR	1 NCCM
Ferrari et al. [122]	2024	Brazil	430	5	1.16	Eco + MNR	3 HCM + 2 NCCM
Sheikh et al. [4]	2014	UK	386	5	1.30	Eco + MNR	5 HCM
Ho et al. [121]	2023	Singapore	562	12	2.14	Eco + MNR	6 HCM + 1 ACM + 1 DCM + 1 NCCM + 3 hypertensive CM
McClean et al. [15]	2019	Qatar	143	10	6.99	Eco + MNR	?
Tranchita et al. [72]	2024	Italy	39	8	20.51	Eco + MNR	5 HCM + 3 DCM
Sheikh et al. [128]	2018		100	21	21.00	Eco + MNR	19 HCM + 1 NCCM + 1 Fabry
Maestrini et al. [125]	2024	Italy	143	59	41.26	Eco + MNR	42 HCM + 10 DCM + 3 ACM + 4 NCCM
Schnell et al. [14]	2015	France	155	69	44.52	Eco + MNR	56 HCM +?
De Lazzari et al. [126]	2021	Italy	112	59	52.68	Eco + MNR	39 HCM + 2 DCM + 2 NCCM + 2 IHD + 11 fibrosis

Studies using other diagnostic methods/non-specified methods
Kiss et al. [13]	2015	Hungary	34	0	0	?
Wasfy et al. [83]	2015	USA	7	0	0	?
Malhotra et al. [64]	2020	UK	337	8	2.37	Eco + Other	5 HCM + 2 ACM + 1 DCM

N TWI number of athletes with T-wave inversion, N Diagnosed CM number of athletes with diagnosed cardiomyopathy, Dx Test diagnosis test, UK United Kingdom, Eco echocardiography, MNR magnetic nuclear resonance, HCM hypertrophic cardiomyopathy, ACM arrhythmic ventricular cardiomyopathy, DCM dilated cardiomyopathy, NCCM non-compaction cardiomyopathy

Relationship Between TWI with a Higher Risk of Cardiovascular Events or Sudden Cardiac Death (Question 3)

Two distinct outcomes were identified: (i) the delayed diagnosis of a structural heart disease in athletes who initially presented no detectable pathology, and (ii) the occurrence of major adverse cardiovascular events, including SCD, syncope, or ventricular arrhythmias.

Thirteen studies evaluated the incidence of newly diagnosed structural heart disease [14, 26, 31, 50, 59, 61, 67, 70, 121, 123, 126, 128, 128]. The estimated annual incidence rate ranged from 0 to 54.9 cases per 1000 athletes (Table 2). The variability in follow-up duration was also very high. Consistent with previous findings, the most commonly diagnosed conditions were hypertrophic cardiomyopathy, arrhythmogenic ventricular cardiomyopathy, and dilated cardiomyopathy. These new diagnoses were most frequently related to the presence of TWI in the lateral leads.

Table 2.

Newly diagnosed cases of structural cardiomyopathy and incidence of cardiovascular events in studies with follow-up. The studies are ordered by the estimated incidence rate per 1,000 persons·year⁻¹

Study	Year	Country	Athletes with TWI	Events	Cumulative incidence (%)	Follow-up years	Estimated follow-up persons-years	Estimated incidence rate per 1000 persons·year	Type of events
Diagnosis of cardiomyopathy in the follow-up
Arroja et al. [59]	2018	Switzerland	8	0	0	4	32	0	–
De Lazzari et al. [126]	2021	Italy	53	0	0	3.6	190.8	0	–
Halasz et al. [67]	2021	Italy	202	0	0	3.2	646.4	0	–
Abela et al. [70]	2023	Malta	47	0	0	3.4	159.8	0	–
Ho et al. [121]	2023	Singapore	562	0	0	2.2	1236.4	0	–
Orchard et al. [123]	2023	New Zealand	43	0	0	6.1	262.3	0	–
D'Ascenzi et al. [61]	2019	Italy	354	1	0.003	4	1416	0.71	1 Non-compacted CM
Papadakis et al. [31]	2011	France	272	3	0.011	5.8	1577.6	1.90	1 HCM
Salido et al. [128]	2025	Spain	28	3	0.107	25	700	4.29	3 HCM
Ghani et al. [50]	2016	UK	40	1	0.025	4.3	172	5.81	1?
Pelliccia et al. [26]	2008	Italy	81	5	0.062	9	729	6.86	3 HCM, 1 ACM,1 DCM
Nishimura et al. [129]	1981	Japan	6	1	0.167	5	30	33.33	1 HCM
Schnell et al. [14]	2015	France	91	5	0.055	1	91	54.95	2 HCM, 1 DCM, 2 arrhythmic events
Clinical cardiovascular events
Nishimura et al. [129]	1981	Japan	6	0	0	5	30	0	–
Pelliccia et al. [24]	2000	Italy	27	0	0	3.4	91.8	0	–
Papadakis et al. [31]	2011	France	272	0	0	5.8	1577.6	0	–
Sheikh et al. [37]	2013	UK	116	0	0	5	580	0	–
Chevalier et al. [38]	2013	France	2	0	0	4.2	8.4	0	–
Wasfy et al. [83]	2015	USA	94	0	0	3	282	0	–
Bakken et al. [107]	2016	Qatar	13	0	0	2	26	0	–
Ghani et al. [50]	2016	UK	40	0	0	4.3	172	0	–
Arroja et al. [59]	2018	Switzerland	8	0	0	4	32	0	–
D'Ascenzi et al. [61]	2019	Italy	354	0	0	4	1416	0	–
De Lazzari et al. [126]	2021	Italy	53	0	0	3.6	190.8	0	–
Halasz et al. [67]	2021	Italy	202	0	0	3.2	646.4	0	–
Abela et al. [70]	2023	Malta	47	0	0	3.4	159.8	0	–
Ho et al. [121]	2023	Singapore	562	0	0	2.2	1236.4	0	–
Orchard et al. [123]	2023	New Zealand	43	0	0	6.1	262.3	0	–
Salido et al. [128]	2025	Spain	28	0	0	25	700	0	–
Pelliccia et al. [26]	2008	Italy	81	1	0.012	9	729	1.37	1 SCD
Malhotra et al. [64]	2020	UK	337	8	0.024	10.6	3572.2	2.24	8 SCD
Schnell et al. [14]	2015	France	91	3	0.033	1	91	32.97	1 SCD, 1 syncope, 1 HF
Kervio et al. [35]	2013	France	6	1	0.167	2	12	83.33	1 SCD

CM cardiomyopathy, HCM hypertrophic cardiomyopathy, DCM dilated cardiomyopathy, ACM arrhythmic ventricular cardiomyopathy, SCD sudden cardiac death, HF heart failure

Twenty studies analysed the incidence of major cardiovascular events among athletes with TWI [14, 24, 26, 31, 35, 37, 38, 50, 59, 61, 64, 67, 70, 83, 107, 121, 123, 126, 129, 128]. The estimated annual incidence rate ranged from 0 to 83.33 events per 1,000 athletes per year (Table 2). The variability in follow-up duration was also very high.

Discussion

In this systematic review and meta-analysis, the 95% confidence interval of the estimated pooled-point prevalence of TWI in adult athletes ranged from 3 to 6%. However, heterogeneity among studies was very high, and this estimate should be interpreted with caution. TWI most frequently involved the anterior leads and was more common in Black athletes. Its prevalence was almost 50% higher among athletes than among non-athletes. A minority of athletes with TWI were diagnosed with an underlying structural cardiomyopathy, with the likelihood of such a diagnosis increasing when TWI affected the lateral leads. In most studies, the prognosis of TWI in athletes was favourable; however, the evidence is limited and highly heterogeneous, making it difficult to determine the actual risk among these athletes.

Prevalence of TWI in Athletes

In this study, we report that the prevalence of TWI is 49% higher in athletes than in non-athletes. However, TWI is considered an abnormal ECG finding, with certain exceptions, such as in Black athletes and in athletes younger than 16 years, when TWI affects the anterior leads [10]. The results of our meta-analysis indicate that the prevalence of TWI among athletes, when assessed using the International Criteria, ranges from 3 to 6% and predominantly affects the anterior leads. There is considerable heterogeneity among studies, with some reporting a prevalence of 0%, while others report prevalences exceeding 20%. The meta-regression indicated that athletes’ age was the only variable explaining the heterogeneity between studies. This variable probably reflects the cumulative lifetime training volume of these athletes. Several studies have reported a positive correlation between the intensity of training, the duration of participation in high-performance programmes, or the level of physical fitness, and the prevalence of TWI [47, 56, 60, 105, 112].

Other factors that may also contribute include inconsistent definitions of an “athlete”, affecting training volume and intensity; differences in participant selection criteria across studies; the type of screening (systematic versus opportunistic); the type of sport practised; and variability in the level of expertise in ECG interpretation in athletes. We could not adequately assess most of these factors in our meta-regression because the information reported in the studies was incomplete.

Ethnicity was not related to the heterogeneity of the results between studies, but we report that the prevalence of TWI is higher among athletes of Black ethnicity. This higher prevalence has been attributed to a combination of genetic, endocrine, and haemodynamic factors [66]. Furthermore, Black individuals tend to have increased myocardial wall thickness and greater left ventricular mass compared to White individuals [31, 37], and this increased ventricular mass has been associated with a higher prevalence of TWI [28, 31, 110, 129]. Interestingly, the geographical origin of athletes has also been recently related to the frequency of repolarisation abnormalities, which are significantly more common among West (6.4%) and Middle African (8.5%) athletes than among those from East (1.5%) and North Africa (< 1%) [110].

In addition, inter-observer variability and differences in expertise in the interpretation of athletes’ ECG may introduce bias in the assessment of TWI [130]. To address this issue, various initiatives, such as in-person and online training courses [131], have been implemented to enhance medical education and proficiency in ECG interpretation in the context of athletic populations.

The mechanisms underlying the presence of TWI in athletes are diverse and include: (i) ventricular remodelling. For instance, the papillary muscle/left ventricular mass ratio >=3.5%, the apical displacement of the papillary muscle, or both have been related to the presence of lateral TWI [126]. Moreover, lateral displacement of the right ventricle or the cardiac apex toward the left axilla has been also associated with progressive TWI in the anterior leads [82, 132]; (ii) structural variants, some authors have associated anterior TWI with the presence of false tendons in the left ventricle [49]; (iii) neuroanatomical changes, such as increased vagal tone [7], which activates the acetylcholine-controlled potassium channel (I_K-ACh), reducing the action potential gradient between the epicardium and endocardium. This change may affect the vectorcardiogram of repolarisation and contribute to the appearance of TWI [133]; and (iv) myocardial ischaemia. Cardiac repolarisation initiates in the epicardium and terminates in the endocardium, reflected in a positive T wave due to the shorter duration of the action potential in the epicardium relative to the endocardium [12]. In ischaemic conditions, reduced endocardial repolarisation time can diminish this gradient, potentially inverting the repolarisation vector and resulting in TWI [12].

The presence of anterior TWI is more frequently observed in athletes under the age of 16 [15, 61]. This finding has been associated with the right ventricular dominance during early adolescence [15, 28, 32, 55, 61, 70]. After the age of 16, anterior TWI generally tends to regress.

TWI and Structural Cardiomyopathy at the Diagnosis or the Follow-up

The presence of TWI requires ruling out several cardiac pathologies associated with SCD, including hypertrophic cardiomyopathy, arrhythmogenic ventricular cardiomyopathy, dilated cardiomyopathy, myocarditis, atherosclerotic coronary artery disease, long QT syndrome, and Brugada syndrome [11, 102]. Additionally, other conditions include electrolyte imbalances (such as potassium or magnesium deficiencies), certain drugs, and congenital cardiac defects [1].

In our review, we selected 35 studies comprising ≥30 athletes with TWI in which the presence of structural cardiomyopathies was specifically evaluated. The results were highly heterogeneous, but the higher prevalence observed in some studies may be attributable to deeper TWI [14, 26], predominance in the lateral leads [26, 72, 126] and the use of different imaging test for the diagnosis. The limited sample sizes and the inclusion of studies conducted in referral centres may also have contributed to the heterogeneity between studies.

When analysing by location, TWI in the lateral leads is most suggestive of underlying structural heart disease [61], most commonly hypertrophic cardiomyopathy, particularly when the inversion is ≥ 2 mm and accompanied by ST-segment depression [31, 35, 37, 46, 56, 85, 102, 121, 134]. Furthermore, normalisation of TWI during exercise testing has been reported as a favourable prognostic indicator [26, 135]. However, Zorzi et al. [136] have shown that patients with overt arrhythmogenic ventricular cardiomyopathy may also exhibit normalisation of T-waves during exercise, thereby questioning the prognostic value of this finding.

Differences in the use of diagnostic imaging tests also explain heterogeneity between studies and introduce detection bias, thereby limiting the comparability between studies. CMR has greater sensitivity than echocardiography for detecting structural cardiac abnormalities [14, 134]. Indeed, recent guidelines and expert opinion statement recommend the use of CMR in the diagnosis process of these pathologies [11, 137].

An additional important consideration is that the absence of detectable structural abnormalities at the time of the exploration does not exclude the possibility of future development of cardiomyopathy. TWI may represent an early phenotypic expression manifestation of an underlying condition that precedes structural abnormalities detectable via imaging by several years [14, 31, 35, 48, 102]. In this review, while half of the studies with follow-up data did not report any new cardiac diagnosis, the remainder reported incidence rates of newly diagnosed cardiomyopathies ranging from 1 to 7 per 1000 athletes per year. One study reported 5 new cases in 91 athletes followed up for 1 year. This study employed a standardised follow-up protocol, including routine CMR, underscoring the relevance of comprehensive evaluation for ruling out structural cardiomyopathies.

TWI and Clinical Cardiovascular Events

The incidence of clinical cardiovascular events in athletes with TWI was highly heterogeneous and difficult to interpret. This heterogeneity was likely related primarily to the limited statistical power of the studies, largely attributable to small sample sizes and short follow-up periods. Moreover, the results of some studies may have been affected by different forms of selection bias, including attrition bias, referral bias, and differences in screening strategies.

Athletes with TWI should undergo further evaluation to exclude underlying cardiomyopathies, and regular follow-up is recommended [11]. The clinical relevance of TWI should also be interpreted in the context of the athlete’s family history, including a history of cardiomyopathy or SCD, and personal history, particularly syncope.

Strengths and Limitations

This systematic review and meta-analysis the most comprehensive systematic review to date aimed at determining the prevalence of TWI in athletes and its prognostic significance. However, several limitations affecting comparability and the very high heterogeneity across studies must be acknowledged: (i) the criteria used to define the presence of TWI vary among studies; (ii) participant inclusion criteria differ widely, particularly in terms of age, ethnicity, sex, training intensity, and type of sport; (iii) the type of screening programme was not always systematic, with some series based on opportunistic screening or recruitment from referral centres; (iv) there is considerable variability in the diagnostic modalities employed to identify underlying cardiomyopathies; and (v) the sample sizes and follow-up durations of the studies assessing the risk of clinical events were limited, resulting in insufficient statistical power. These limitations reduce the certainty of the estimates from this meta-analysis and hamper their clinical interpretation.

Clinical Interpretation, Conclusions and Future Research

A summary of the main findings and the clinical interpretation of this systematic review and meta-analysis is shown in Fig. 7. TWI is more common in athletes than in non-athletes, and it is observed more frequently in individuals of Black ethnicity. While TWI may represent an adaptive response to sport-induced ventricular remodelling, it can also be indicative of underlying structural heart disease, especially when present in the lateral or infero-lateral leads. Accurate identification and thorough evaluation, particularly through advanced imaging modalities such as magnetic resonance imaging, are essential for appropriate clinical management. Current prospective studies assessing the clinical prognosis of athletes with TWI have insufficient statistical power and methodological limitations, hampering the ability to draw robust conclusions. International, collaborative, and standardised studies with adequate sample sizes and follow-up periods are needed to adequately address the questions of interest identified in this review:

• Prevalence: standardised systematic screening programmes, with an appropriate definition of the target population, standardised criteria for defining TWI, and trained professionals for ECG interpretation are needed. Analyses should be stratified by athlete level (recreational, competitive, and professional/elite), type of sport (endurance, strength/power, and mixed), years and intensity of training, ethnicity, and sex.

• Associated cardiomyopathies: studies should include all cases identified through systematic screening and use a standardised diagnostic protocol incorporating echocardiography, exercise testing, 24-h Holter monitoring, and CMR. Standardised longitudinal follow-up should be performed to identify athletes who subsequently develop cardiomyopathy.

• Risk of clinical events: studies should include all cases identified through systematic screening and implement standardised follow-up to capture all relevant clinical events, including SCD, syncope, and symptomatic events such as palpitations and dyspnoea. Strategies should be implemented to minimise attrition and other forms of selection bias, with time to event as the primary outcome of interest.

Fig. 7.

Summary of the main results of the meta-analyses and potential clinical implications. Q1–3: Question 1–3; TWI T-wave Inversion; CMR cardiac magnetic resonance

Abbreviations

CI

Confidence interval

ECG

Electrocardiogram

ESC

European Society of Cardiology

JBI

Joanna Briggs Institute

MeSH

Medical Subject Headings

PICOS

Population, Intervention, Comparator, Outcome, and Study Design

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

SCD

Sudden cardiac death

SE

Standard error

TWI

T-wave inversion

Author Contributions

Substantial contributions to the conception or design of the work: JC-B, RE. Acquisition: JS, JC-B, RE. Analysis: JS, RE. Interpretation of data for the work: JS, JC-B, HV, LC, SM, AB, RE. Drafting the work: JS, RE. Reviewing it critically for important intellectual content: JC-B, HV, LC, SM, AB. Final approval of the version to be published: JS, JC-B, HV, LC, SM, AB, RE. All authors read and approved the final version of the article: JS, JC-B, HV, LC, SM, AB, RE. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: JS, JC-B, HV, LC, SM, AB, RE.

Funding

This research was partially funded by CIBERCV (Instituto de Salud Carlos III).

Data Availability

Data extracted from included studies, data used for all analyses, analytic code, any other materials used in the review are available upon request to the corresponding author.

Declarations

Ethics Approval and Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Competing Interests

The authors declare no competing interests.