Electrocardiographic modifications induced by breast implants

Background Echocardiography realization can be challenging in the presence of breasts implants (BI). It is less known if electrocardiograms (ECG) may be modified in the presence of BI. Methods ECG from women with BI (and without any known cardiac structural disease) were sent and analyzed by two experienced electrophysiologists (EP1 and EP2) who were blinded and completely unaware of the context of the patients (Group 1). ECG from a control matched‐group of female women without BI (Group 2) were also blindly sent for analysis. Results ECG were collected from 28 women with BI (42 ± 8 years) without any acute medical condition. A proportion of 42% of the ECG were considered abnormal by EP1 and 46% by EP2. The abnormalities were for EP1: negative T waves (5), ST depression in inferolateral leads (2), absence of R wave progression from V1 to V4 (4), left ventricular (LV) hypertrophy (1), long QT(1), early repolarization (1), short PR (1); For EP2: negative T waves (6), ST depression in inferolateral leads (2), absence of R wave progression from V1 to V4 (4), LV hypertrophy(3), long QT (1), early repolarization (1). ECG from group 2 were considered abnormal in only 1 patient (5%) for EP1, and normal in all for EP2 (P = 0.0002 between the groups). Conclusions ECG from women with BI were considered abnormal in 42% to 46% of the cases by expert readers. ECG interpretation can thus be misleading in these women.

Doing an echocardiography may be challenging in the presence of breasts implants (BI), as the ultrasound transmissions are impaired by the protheses structure. 3 It is less known if an ECG may be modified by the presence of BI.

| Patient selection
Twelve-lead ECGs obtained from women with BI were sent and analyzed by two experienced electrophysiologists (EP1 and EP2) who were blinded and completely unaware of the context (Group 1). None of the women had BI because of reconstructive surgery (breast cancer). The women gave their consent for their ECG to be collected for the purpose of the study. ECGs from a control group of women without BI (Group 2, n = 20) were also randomly and blindly sent for analysis. The control group included women from our nurse and paramedical staff. Exclusion criteria were: age > 55, any cardiovascular sign/disease (hypertension, stroke or congestive heart failure, diabetes, and dyslipidemia as defined by low-density lipoprotein ABBREVIATIONS: BI, breasts implants; ECG, electrocardiogram/electrocardiographic; EP, electrophysiologist; SHD, structural heart disease. [LDL]-cholesterol ≥160 mg/dL if age comprised between 20 and 39 years; or LDL-cholesterol ≥70 mg/dL if age above 40 years). All the ECGs were exclusively performed by the nursing staff of our department and special care was taken to place the electrodes in a correct and reproducible position, despite the presence of BI (Group 1): fourth intercostal space on the right (V1) and left (V2) border of the sternum, V4 on the fifth intercostal space on the midclavicular line, V3 midway between V2 and V4, V5 on the anterior axillary line on the same horizontal level as V4, and V6 on the mid-axillary line on the same horizontal level as V4 and V5.
All the women from both groups had an echocardiography to check for any structural heart disease (SHD) that could likely explain their ECG modification.
The electrophysiologists were asked to report the abnormalities conference, that is, QRS duration <120 ms, with Jp 0.1 mV in two or more contiguous leads of the 12-lead ECG excluding V1 to V3, and presence of an end-QRS notch or slur on the prominent R-wave. 4

| Statistical analysis
The statistical analysis was completed using GraphPad Prism 5 (San Diego, California). Numerical variables are expressed as mean ± SD. A Cohens Kappa test was used for inter-observator agreement correlation.

| RESULTS
ECGs were collected from 28 women with BI (mean age 42 ± 8 years; all of Caucasian origin except for one woman, who was from African origin). The mean time between the BI insertion and the ECG recording was 3.1 ± 2.4 years. Only one woman had an ECG before and after the insertion of her BI. None of the women had a personal history of SHD or known cardiovascular risk factors in the BI group, neither in the control group. There were no differences concerning the body mass index between the two groups (20.2 ± 5.8 in group 1 vs 22.9 ± 3.0 in group 2; P = 0.42). A proportion of 42% (12/28) of the ECGs was considered abnormal by EP1 and 46% (13/28) by EP2. The abnormalities (Table 1)  The inter-observator agreement was calculated at 92.3%.
ECGs from group 2 (38 ± 7 years, all of Caucasian origin) were considered abnormal only in one woman of group 2 (5%) for EP1 (absence of R wave progression from V1 to V4), and all normal for    an open capsulotomy on this patient. 10 One possible hypothesis could be electrical vector deviations emanating from the heart, because the different wave fronts encounter an unexcitable region (silicone) before reaching the surface of the skin ( Figure 2). This hypothesis is speculative, and needs be confirmed by experimental studies, but is an extension of the phenomenon observed with ultrasounds propagation in the presence of BI when performing an echocardiography.
It may be a reasonable advice to suggest to women who are planning to have BI insertion to have an ECG before and keep it in their file to serve as a comparison for the future, in case of the appearance of any cardiovascular symptoms, in their medical history.

| Limitations
This is a monocentric study. The number of women with ECGs before the insertion of BI is limited (n = 1). This is due to the fact that ECGs are not systematically recommended as part of the preoperative (anesthesiologist) visits in young women without SHD nor cardiovascular risk factors. A prospective study is needed, aiming to compare ECGs before BI insertion with post-operative ECG.
The data on the size of the BI in our population was not available to assess a possible correlation between the size of the BI and ECG modifications.

| CONCLUSION
ECGs obtained from women with breasts implants were considered abnormal in 42% to 46% of the cases in comparison with a control group of women without breast implants (P = 0.0002). ECG interpretation can be misleading in the context of chest pain/acute coronary syndrome occurring in these patients.

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of this article.