Author + information
- Received December 17, 2014
- Revision received January 30, 2015
- Accepted February 16, 2015
- Published online March 1, 2015.
- Maria J. Brosnan, MBBS∗,†∗ (, )
- Guido Claessen, MBBS‡,
- Hein Heidbuchel, MBBS, PhD§,
- David L. Prior, MBBS, PhD∗,† and
- Andre La Gerche, MBBS, PhD∗,‡,‖
- ∗Department of Cardiology, St. Vincent’s Hospital, Fitzroy, Australia
- †Department of Medicine, St. Vincent’s Hospital and University of Melbourne, Fitzroy, Australia
- ‡Department of Cardiovascular Medicine, University Hospitals Leuven, Leuven, Belgium
- §Hasselt University and Heart Center, Jessa Hospital, Hasselt, Belgium
- ‖Baker IDI Heart and Diabetes Institute, Melbourne, Australia
- ↵∗Reprint requests and correspondence:
Dr. Maria Brosnan, Department of Cardiology, St. Vincent’s Hospital, Melbourne, PO Box 2900, Fitzroy, Victoria 3065, Australia.
Objectives The objective of this study was to test the hypothesis that T-wave inversion in the right precordial leads (TWIV2-3) reflects lateral displacement of the heart such that the surface electrocardiographic (ECG) leads overlie a greater proportion of the right ventricle (RV).
Background TWIV2-3 on ECG is more frequently observed among endurance athletes (EAs) than in the general population, the underlying mechanism for which is unclear.
Methods Sixty-eight EAs and 41 nonathletic control subjects underwent ECG and cardiac magnetic resonance imaging (CMRI). In addition to standard measurements of biventricular function and volume, novel measurements of cardiac displacement and orientation were analyzed from horizontal long-axis images. These included RV wall thickness in diastole (RVd), cardiac-to-hemithorax area ratio (CHTx%), percentage of circumferential displacement of the RV apex toward the axilla (%LatD), and the angle of interventricular septum with respect to the thoracic midline (∠septal).
Results All cardiac volume, RVd, CHTx%, %LatD, and ∠septal values were greater in EAs than in controls. Compared to EAs without TWIV2-3, EAs with TWIV2-3 (n = 26) did not have greater RV wall thickness or cardiac volumes (RVd = 4.9 vs. 4.8 mm, p = 0.695; LVEDV = 231 vs. 229 ml, p = 0.856; RVEDV = 257 vs. 254 mL, p = 0.746), but all measurements of cardiac displacement toward the axilla were greater (%LatD = 45.6% vs. 37.9%, respectively, p < 0.0001; ∠septal = 54.23° vs. 48.63°, respectively, p = 0.001; and CHTx% = 46.3% vs. 41.9%, respectively, p = 0.048).
Conclusions In healthy EAs, TWIV2-3 is associated with displacement of the RV toward the left axilla rather than RV dilatation or hypertrophy. TWIV2-3 may be explained by the position of the RV relative to that of the surface ECG leads.
Right precordial T-wave inversion in leads V2 to V3 (TWIV2-3) is uncommon in the general population but is observed in up to 85% of subjects with confirmed arrhythmogenic right ventricular cardiomyopathy (ARVC), a genetically determined cardiomyopathy which is associated with an increased risk of sudden cardiac death (1,2). Thus, current guidelines consider TWIV2-3 an abnormal finding, which should prompt further investigation to exclude structural heart disease when found in athletes (3,4). TWIV2-3, however, may be more prevalent among highly trained endurance athletes (EAs) (5,6) and some ethnic athletic populations (7) than among nonendurance athletes and the general population (8,9).
The mechanisms underpinning the greater prevalence of TWIV2-3 among EAs have not been elucidated. Zaidi et al. (10) investigated the logical hypothesis that TWIV2-3 might be associated with right ventricular (RV) dilation and hypertrophy but found little relationship between TWIV2-3 and RV dimensions. One possible explanation could be the modest accuracy of 2-dimensional echocardiographic measurements of RV mass and volumes (11).
We formulated an alternative hypothesis. Exercise-induced cardiac remodeling (“the athlete’s heart”) can be profound (12), and, given the anatomical constraints imposed by the thorax, the heart is displaced progressively along the left anterior chest wall toward the axilla. As a result, it might be expected that a larger proportion of the RV lies adjacent to the right precordial electrocardiographic (ECG) leads. Thus, it may be that TWIV2-3 may be better explained by the anatomical position of the RV within the thorax than the size or structural characteristics of the RV. In other words, the normal TWI observed in V1 may be observed in lower and more lateral ECG leads simply because these leads are now overlying the body of the RV rather than being in proximity to the cardiac apex. To address this hypothesis, we investigated whether greater lateral cardiac displacement was associated with TWIV2-3 in EAs and nonathletic controls using cardiac magnetic resonance imaging (CMRI).
Sixty-eight elite endurance athletes (EAs) volunteered to participate. EAs were defined as individuals participating in more than 10 h of intense exercise per week, and all were participating in long-distance events ranging from the marathon to the ultraendurance triathlon. All athletes were healthy and asymptomatic. Other than TWIV2-3, there was no clinical evidence of ARVC in any athlete after thorough investigations.
A nonathletic control cohort consisted of 20 healthy volunteers and 21 subjects who underwent CMRI for clinical indications but in whom no cardiac pathology was identified. Written informed consent was obtained from all subjects, and the protocol was approved by the St. Vincent’s Hospital Human Research Ethics Committee in accordance with the declaration of Helsinki.
Standard 12-lead ECG was performed with all subjects at rest, in the supine position, at 10 mm/mV and 25 mm/s. Measurements were made with digital calipers and included heart rate, QRS axis, and QRS duration. Sokolow-Lyon s Sokolow-Lyon scores for left ventricular (LV) hypertrophy (SV1 + RV5, mV) and RV hypertrophy (RV1 + SV5, mV) were recorded as continuous variables. The presence of incomplete right bundle branch block (RBBB) was noted, defined as an rSR′ morphology in V1 with a QRS duration of <120 ms. The R/S transition zone was defined as the precordial lead where R-wave amplitude exceeded S-wave amplitude. TWI was measured in each lead separately and considered significant if deeper than −0.1 mV. TWIV2-3 was defined as T-wave inversion in lead V2 or in V2 and V3. The presence of bifid T waves was also noted and considered to represent TWIV2-3 if the negative portion was deeper than −0.1 mV in leads V2 and V3.
Cardiac magnetic resonance imaging
CMRI was performed with a 1.5-T scanner (Signa Excite, GE Healthcare, Waukesha, Wisconsin; or Achieva, Philips Medical Systems, Best, the Netherlands), using a dedicated cardiac multiphased array coil and cardiac gating during breath-hold. Cine imaging was used to obtain a contiguous short-axis stack (8-mm slice thickness without gaps) covering the LV and RV from the apex to a level well above the atrioventricular groove. Endocardial and epicardial borders were manually traced with customized software (RightVol, Leuven, Belgium) to quantify volumes at end-diastole (EDV) and end-systole (ESV).
Measurements of RV displacement were performed, as shown in Figure 1. The end-diastolic frame was selected from a horizontal long-axis cine acquisition, and analysis was performed using open-source DICOM version 4.1.2 viewing software (OsiriX, Geneva, Switzerland). The following linear planes were defined and are illustrated in Figure 1: an anteroposterior mid thorax line passing through the mid-point of the sternum and thoracic vertebrae (Figure 1, line A), and a line orthogonal to and bisecting line A, passing through the axilla (Figure 1, line B). The angle of the interventricular septum (∠septal) was then measured with respect to line A, in a plane passing through the atrioventricular crux and the cardiac apex (Figure 1, line C). The point at which line C crossed the internal thoracic wall was labeled RV apex. The distance from the sternum (Figure 1, line A) to the RV apex was then manually traced, following the bony thorax, and defined as L1 (in cm). This line was then extended, following the thorax, to meet line B, and defined as L2 (cm). This lateral displacement (L1:L2) ratio was then calculated and expressed as a percentage (%LatD). Finally, the area of the left hemithorax was traced (HTx), as was the area of the whole heart (Figure 1, line C) within the left hemithorax. The cardiac area/hemithorax area ratio was then calculated and expressed as a percentage (CHTx%). The thickness of the RV free wall was measured in the short axis view from an end-diastolic frame at the level of the papillary muscles. All measurements were performed blinded to subject identity or ECG findings.
Reproducibility of CMRI measures
Measurements of LV and RV volumes were performed by 2 blinded interpreters (A.L.G. and G.C.). Measurements of RV displacement were performed by a single blinded interpreter (M.B.) and repeated by a second blinded observer (A.L.G.), using a subset of 25 randomly selected subjects. Interobserver reproducibility of measurements of RV volume and displacement were assessed using intraclass correlation coefficient analysis and are reported as coefficients (95% confidence interval [CI]).
Values are means ± SD or percentages, as appropriate. Group differences between EA and controls as well as those between EA with TWIV2-3 and those EA without TWIV2-3 were analyzed using independent sample t tests with Levene’s test for equality of variances for continuous variables, and chi-square or Fisher’s exact test for categorical variables. Univariate binary logistic regression analysis was performed to test for characteristics associated with presence or absence of TWIV2-3. From this analysis, covariates identified as having significant association with TWIV2-3 were entered into a multivariate model (with forward stepwise selection) to determine predictors of TWIV2-3. Linear regression analysis was used to test for an association between QRS axis and cardiac volumes. A 2-tailed p value of <0.05 was considered significant throughout. All statistical analyses were performed using SPSS version 21 software (IBM, Armonk, New York).
Demographic, morphometric, electrocardiographic, and CMRI characteristics of EAs and controls are shown in Table 1. Compared to controls, EAs were younger and taller, and a smaller proportion of the group was female.
Compared to controls, EAs had slower heart rates, a more rightward QRS axis, larger scores for LVH and RVH, and a greater prevalence of incomplete RBBB (Table 1). In EAs, the R/S transition zone was later than that of controls, and right precordial TWIs and bifid T waves were far more prevalent. No control subject displayed TWI beyond lead V2.
CMRI findings in endurance athletes versus nonathletic controls
Right- and left-sided cardiac volumes were larger, and the RV free wall was thicker in EAs than in controls (Table 1). Compared to controls, EAs had a larger left hemithorax area, and the heart occupied a relatively larger percentage of this area (expressed as CHTx%). Furthermore, in EAs, the RV and cardiac apex were displaced farther toward the left axilla, represented by greater L1, %LatD, and ∠septal than those of controls.
CMRI findings and associated electrocardiographic features in endurance athletes with right precordial T-wave inversion
Electrocardiographic and CMRI characteristics of EAs with and without right precordial TWI are presented in Table 2. QRS axis, RVH score and R/S transition zone were similar in EAs with and without TWIV2-3, as were right and left sided cardiac volumes and RVd wall thickness. On the other hand, as compared with EAs without TWIV2-3, EAs with TWIV2-3 had a larger CHTx%, %LatD and ∠septal representing greater displacement of the RV toward the axilla. Examples of 2 representative athletes are presented in Figures 2A and 2B.
CMRI findings in athletes with bifid T-waves in the right precordial leads
CMRI and other ECG findings in the 10 athletes with bifid right precordial TWI were similar to those with deep TWIV2-3. Compared to athletes with no TWI, all cardiac volumes were similar, but ∠septal (56° vs. 49°, respectively, p = 0.001) and %LatD (47% vs. 38%, respectively, p < 0.0001) were larger.
CMRI findings in non-athletic controls with right precordial T-wave inversion
Only 1 nonathletic control subject demonstrated right precordial TWI, which was isolated to leads V2. Compared to the remainder of the nonathletic control group, this subject demonstrated larger cardiac volumes (LVEDV of 178 vs. 159 ml, respectively; RVEDV of 212 vs. 167 mL, respectively), and larger measurements of CHTx% (44% vs. 39%, respectively), ∠septal (47° vs. 40°, respectively), and %LatD (45% vs. 35%, respectively). As expected with only 1 subject, none of these differences reached statistical significance.
Predictors of right precordial T-wave inversion in endurance athletes
Univariate predictors of TWIV2-3 are shown in Table 3. On multivariate analysis, %LatD was the only variable significantly associated with the presence of TWIV2-3, with an odds ratio (OR) of 11.52 per 10% increment in this ratio (95% CI: 10.59 to 12.53, p = 0.001). Area under the receiver operating curve (ROC) was 0.774 (95% CI: 0.661 to 0.887, p < 0.0001), with %LatD >42.8% predicting TWIV2-3 with 74% sensitivity and 78% specificity (Figure 3). When separately considering those athletes with bifid T-waves in V2-V3, the findings were similar, with %LatD the only variable on multivariate analysis associated with this finding (OR: 12.10 per 10% increment, 95% CI: 10.54 to 13.89, p = 0.007). Area under the ROC curve was 0.856 (95% CI: 0.742 to 0.970, p = 0.001) with %LatD >43.2% predicting the presence of bifid T-waves in the right precordial leads with 90% sensitivity and 80.5% specificity.
Association between QRS axis and right ventricular volumes
QRS axes did not differ between those subjects with and those without TWIV2-3; however, linear regression analysis demonstrated a significant association between QRS axis and RV volume, with each 10° increment in axis associated with a 4.4-ml increase in RVEDV in athletes (p = 0.036) and a 5.9-ml increase in nonathletes (p = 0.005).
Reproducibility of CMRI measurements between interpreters
There was excellent agreement between interpreters for measurements of cardiac displacement and cardiac volumes, with intraclass correlation coefficients as follows: ∠septal was 0.938 (95% CI: 0.53 to 0.973), L1 was 0.958 (95% CI: 0.889 to 0.982), L2 was 0.933 (95% CI: 0.849 to 0.971), %LatD was 0.932 (95% CI: 0.837 to 0.971), LVEDV was 0.985 (95% CI: 0.971 to 0.992), and RVEDV was 0.982 (95% CI: 0.965 to 0.991).
Lateral displacement of the cardiac apex was first described in endurance athletes more than a century ago by the Swedish physician Henschen (13), who used the simple technique of cardiac percussion and auscultation. Our study is the first to consider this basic examination finding as a potential explanation for right precordial TWI, which is observed on the ECGs of approximately 1 in 7 healthy endurance athletes in the absence of pathological electrical or structural cardiac abnormalities (5). We extended the cardiac enlargement and displacement observed in athletes by Henschen (13) by using modern CMRI techniques and related these changes to the standard position of the ECG leads. The fact that the RV apex is displaced inferiorly and laterally relative to the sternum means that a greater proportion of the RV is positioned under ECG leads that have been placed more inferiorly and laterally. Thus, the TWI normally observed in V1 is frequently observed also in V2 and occasionally in V3 in athletes. We demonstrated a strong association between the extent of leftward cardiac displacement and the extent of TWI in the precordial leads. Thus, it seems that TWIV2-3, which is usually associated with pathological changes in the RV in nonathletes, is explained by very simple geospatial relationships in EAs.
Previous studies focusing on cardiac volumes and dimensions as assessed by ECG have found no relationship between right precordial TWI and RV dimensions in athletes (10). Our observations provide a robust validation of these findings, given the improved accuracy of CMRI measurements of RV dimensions relative to that of ECG (14). We observed that CMRI-derived cardiac volumes, although significantly larger in EAs than in controls, did not correlate directly with the presence of TWIV2-3. However, we observed that, as a secondary result of cardiac enlargement within the constraints of the thorax, lateral cardiac displacement was prominent in EAs. In EAs with TWIV2-3 it was demonstrated that the RV was displaced relative to the thorax such that it was placed against the anterior chest wall with the apex extending closer to the left axilla. As such, this means that a greater portion of the RV is placed behind the ECG leads situated on the anterior chest wall and to the left of the sternum, namely leads V2 and V3. This is well illustrated by the examples in Figures 2A and 2B. Although the ventricular volumes are similar in the 2 athletes (Figure 2A), in athlete A, the right atrium and more than one-half of the RV extend into the right hemithorax. In athlete B, the heart is displaced leftward such that the entire RV sits under the sternum or in the left hemithorax (Figure 2B). As a result, in athlete B, much more of the RV lies in direct proximity to the precordial ECG markers, resulting in TWIV2-3 (Figure 2B). This cardiac displacement is not appreciable on ECG because the surface anatomical “window” is adjusted to align with cardiac landmarks. In comparison, CMRI provides accurate assessment of RV structure and cardiac displacement can be assessed relative to the thorax.
We are looking at the RV, but why are the T-waves negative?
The genesis of negative, biphasic, or notched T-waves in the right precordial leads in healthy and diseased states is still not completely understood. It has been demonstrated in animal and human models that epicardial activation and repolarization of an enlarged RV occurs after that of the LV, and ST-T polarity is positive at the earliest and negative at the latest site of epicardial repolarization, resulting in negative or biphasic T waves in the ECG leads which look at the RV (15). In subjects with ARVC, a strong correlation between RV volumes, as assessed by ventriculography, and the extent of right precordial TWI has been demonstrated. Although Nava et al. (16) did not have the benefit of CMRI to assess cardiac displacement, they theorized that these findings were probably due to RV dilatation displacing the LV backward (16). Similarly, Awa et al. (17) theorized that, in children with congenital heart disease, the presence of bifid or biphasic TWIV2-3 might be related to the proximity of the RV to the anterior chest wall as a result of cardiac enlargement, with the chest leads recording a more localized epicardial deflection than would be expected with normal cardiac orientation. Our observation that it is the location of the RV with respect to the anterior chest wall rather than cardiac dimensions per se, which result in deep or bifid TWIV2-3, supports the theories of these authors. Our findings also offer a possible explanation for TWIV2-3 in nonathletes who have conditions resulting in RV displacement, such as pectus excavatum (18–20).
Associated ECG findings
Compared to nonathletic controls, EAs were found to have a more rightward QRS axis, and the QRS transition occurred later. However, neither of these findings correlated with the presence of TWIV2-3. Although this was perhaps surprising, a similar lack of association between QRS transition and TWI has been reported previously (16). Furthermore, we found that a more rightward QRS axis was associated with bigger RV volumes in both athletes and controls, supporting the notion that a larger RV mass will result in the ECG finding of a more rightward QRS axis but that the presence of TWIV2-3 relies on other anatomical factors such as cardiac displacement.
Higher-than-expected prevalence of right precordial T-wave inversion in the endurance athlete cohort
We have previously reported that TWIV2-3 is approximately 3 times more prevalent in endurance athletes than in nonendurance athletes (5). The prevalence of right precordial TWIs in the current cohort of 68 endurance athletes was higher than that which we reported in the 251 endurance athletes in the aforementioned study but similar to that recently reported in a cohort of highly trained endurance athletes (6). All subjects were healthy study volunteers, not included on the basis of ECG abnormalities, and therefore selection bias would not have been expected to explain the high prevalence of TWIV2-3. Differences in cohort demographics may be a more likely explanation. Compared with our previous study (5), the current cohort was older (35 ± 8 vs. 22 ± 5 years of age, respectively), and a large proportion were ultraendurance athletes participating in events such as the Ironman triathlon (21). Given this difference, it is perhaps reasonable to hypothesize that those who had performed the greatest volume of training (ultraendurance athletes) over the longest periods of time (an average of 10 ± 9 years in the current cohort) (21) would have greater cardiac remodeling and displacement, which would explain the higher prevalence of TWI. To address this speculative hypothesis, a larger cohort with a broader age range and better-defined training duration would need to be studied. Nevertheless, the high prevalence of TWIV2-3 in the current cohort was an advantage in that it increased the statistical power for the comparison between athletes with and without TWIV2-3.
Clinical implications and future directions
CMRI is increasingly used in the assessment of asymptomatic athletes with TWIV2-3 (commonly in the setting of preparticipation screening), in whom underlying cardiomyopathy cannot adequately be excluded after ECG. In the absence of any clinical or radiological features of structural heart disease such as ARVC, increased lateral displacement of the RV apex toward the axilla may provide an explanation and provide reassurance regarding an expectedly benign prognosis. Similarly, in nonathletic subjects referred for CMRI on the basis of incidental ECG findings of TWIV2-3, lateral displacement of the RV apex could be considered as part of the assessment.
Further studies in patients with proven ARVC would be useful to determine whether cardiac displacement may contribute to the degree of observed TWI, or whether the ECG changes in this cohort represent a unique electrical substrate.
The number of nonathletic control subjects with TWIV2-3 was low, reflecting the low prevalence of TWIV2-3 in healthy, nonathletic individuals. Although a trend for the same alterations in cardiac orientation was seen in the single control subject with TWIV2-3, more subjects with TWIV2-3 would be required to determine whether the same observations hold. It is possible that other anatomical factors which are rare in EAs, such as chest wall adiposity, may need to be taken into consideration when correlating CMRI with ECG features in nonathletic subjects.
All subjects were Caucasian, thus it was beyond the scope of this study to assess whether cardiac displacement could explain the high prevalence of right precordial T-wave changes reported in subjects of black African and Afro-Caribbean ethnicity (7).
CMRI provides a unique opportunity to assess RV structure, orientation, and displacement within the thorax. We demonstrated that lateral displacement of the RV toward the left axilla is associated with progressive TWI in the right precordial ECG leads. This may explain why TWIV2-3 is more common in EAs and may also be helpful in differentiating athlete’s heart from ARVC in the absence of any other clinical or radiological features to suggest this condition.
COMPETENCIES IN MEDICAL KNOWLEDGE: In healthy endurance athletes in the absence of any clinical or radiological features of ARVC, TWIV2-3 on 12-lead ECG may be explained by lateral displacement of the right ventricle (RV) rather than RV dilatation or hypertrophy.
COMPETENCIES IN PATIENT CARE: In subjects referred for cardiac magnetic resonance imaging on the basis of ECG findings of right precordial TWI, the extent of lateral displacement of the RV apex toward the axilla is an important component of the assessment. Although ECG changes may represent structural cardiac changes, they may also reflect cardiac position within the thorax.
TRANSLATIONAL OUTLOOK: Further studies in patients with ARVC would be useful to determine whether cardiac displacement may contribute to the degree of observed TWI, or whether the ECG changes in this cohort are due to the underlying electrical substrate.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- percentage of circumferential displacement of the right ventricle apex toward the left axilla
- angle of interventricular septum with respect to the thoracic midline
- cardiac magnetic resonance imaging
- cardiac to left hemithorax area
- endurance athlete
- right precordial T-wave inversion
- Received December 17, 2014.
- Revision received January 30, 2015.
- Accepted February 16, 2015.
- American College of Cardiology Foundation
- Nasir K.,
- Bomma C.,
- Tandri H.,
- et al.
- Marcus F.I.,
- McKenna W.J.,
- Sherrill D.,
- et al.
- Corrado D.A.P.,
- Heidbuchel H.,
- Sharma S.,
- et al.
- Drezner J.A.,
- Ackerman M.J.,
- Anderson J.,
- et al.
- Wasfy M.M.,
- DeLuca J.,
- Wang F.,
- et al.
- Papadakis M.C.F.,
- Kervio G.,
- Rawlins J.,
- et al.
- Aro A.L.,
- Anttonen O.,
- Tikkanen J.T.,
- et al.
- Papadakis M.,
- Basavarajaiah S.,
- Rawlins J.,
- et al.
- Zaidi A.,
- Ghani S.,
- Sharma R.,
- et al.
- Prior D.L.,
- La Gerche A.
- ↵Henschen S. Skidlauf und skidwettlauf: eine medizinische sportsudie (A Study in Sports Medicine, Skiing and Competitive Skiing). Mitt Med Klin Upsala 1899;2.
- Chen P.S.,
- Moser K.M.,
- Dembitsky W.P.,
- et al.
- La Gerche A.,
- Burns A.T.,
- Mooney D.J.,
- et al.