Unipolar atrial electrogram morphology is affected by age: evidence from high-resolution epicardial mapping

Abstract Background It is unknown which features of unipolar atrial electrogram (U-AEGM) morphology are affected by ageing and whether age-related changes in U-AEGM morphology are equally distributed throughout the right and left atria. Patients and methods Epicardial high-resolution mapping was performed in patients undergoing coronary artery bypass grafting surgery during sinus rhythm (SR). Mapping areas include the right atrium (RA), left atrium (LA), pulmonary vein area (PVA) and Bachmann’s bundle (BB). Patients were categorized into a young (age < 60) and aged (age ≥ 60) group. U-AEGM were classified as single potentials (SPs, one deflection), short double potentials (SDPs, deflection interval ≤ 15ms), long double potentials (LDPs, deflection interval > 15ms) and fractionated potentials (FPs, ≥3 deflections). Results A total of 213 patients (age: 67 (59–73) years; young group N = 58, aged group N = 155) were included. Only at BB, the proportion of SPs (p = 0.007) was significantly higher in the young group, while the proportion of SDPs (p = 0.051), LDPs (p = 0.004) and FPs (p = 0.006) was higher in the aged group. After adjusting for potential confounders, older age was associated with a reduction in SPs [regression coefficient (β): −6.33, 95% confident interval (CI): −10.37 to −2.30] at the expense of an increased proportion of SDPs (β: 2.49, 95% CI: 0.09 to 4.89), LDPs (β: 1.94, 95% CI: 0.21 to 3.68) and FPs (β: 1.90, 95% CI: 0.62 to 3.18). Conclusions Age-related remodeling particularly affects BB as indicated by the decreased amount of non-SP at this location in the elderly. Key Messages Ageing preferentially affects the morphology of unipolar atrial electrograms recorded at Bachmann’s bundle. At Bachmann’s bundle, the proportion of short double-, long double- and fractionated potentials increase during ageing at the expense of a decrease in the proportion of single potentials, reflecting aggravation of abnormalities in conduction. The increase in abnormal unipolar atrial electrograms at Bachmann’s bundle during ageing supports the concept that Bachmann’s bundle may play an important role in development of age-related arrhythmias such as atrial fibrillation.


Introduction
Ageing is one of the principal risk factors for development of atrial fibrillation (AF) [1][2][3]. The reported prevalence of AF ranges between 0.12% and 0.16% in subjects younger than 49 years and increases up to 17% above 80 years or older [4]. Although ageing is currently one of the most important factors resulting in worldwide health issues such as AF [5], data on age-related changes in atrial electrophysiology is scarce.
In canine models, Anyukhovsky et al. [1] demonstrated that the average action potential duration was significantly longer in the right and left atria of elderly dogs compared to adult dogs. Also, action potential duration heterogeneity was more pronounced in atrial tissues of old dogs. In humans, endocardial mapping studies revealed that ageing is associated with slowing of conduction at the right atrium (RA) and the presence of diffusely spread low-voltage areas [6]. Recently, Van der Does et al. [7] revealed that ageing was Unipolar atrial electrogram; ageing; coronary artery bypass grafting; sinus rhythm; epicardial mapping associated with more conduction disorders, especially at Bachmann's bundle (BB) and the RA. However, current data on age-related changes in electrogram (EGM) morphology is limited to the RA. Kistler et al. [6] demonstrated in 41 patients that ageing was related to a larger number of double and fractionated potentials along the crista terminalis. However, it is unknown which features of EGM morphology are affected by ageing and whether the age-related changes in EGM morphology are equally distributed throughout the right and left atria (LA). The purpose of this study is therefore to investigate the influence of age on unipolar atrial EGM (U-AEGM) morphology recorded from the RA and LA including BB in a large cohort of patients undergoing cardiac surgery.

Study population
The study population included adult participants, who underwent coronary artery bypass grafting (CABG) surgery at Erasmus Medical Center in Rotterdam between March 2012 to August 2020 (8,9). This project has been approved by the institutional medical ethics committee (MEC2010-054/MEC2014-393) [8,9], and written informed consent was obtained from all participants before enrolling. Baseline demographic and clinical profiles (e.g. age, gender, body mass index and underlying heart diseases) were attained from the hospital's electronic medical system. None of the patients included in this study had a history of AF. All included participants were grouped into a young (age < 60) and aged (age ≥ 60) group, which was based on previous studies [10,11].

Mapping procedure
As described in our previous studies, epicardial high-resolution mapping was carried out before extracorporeal circulation [8,9]. In short, a temporary bipolar pacemaker lead was sutured on the free wall of the RA and served as a reference electrode, and a steel wire was fixed on subcutaneous tissue of the thoracic cavity as a neutral electrode. A 128-or 192electrode array (electrode diameter: 0.65 or 0.45 mm, interelectrode distances: 2.0 mm) was used for epicardial mapping of the atria (upper left panel of Figure  1). According to a predefined mapping scheme (upper left panel of Figure 1), the electrode array was moved along an imaginary line with a fixed anatomical orientation for mapping, covering the entire epicardial atrial surface [including RA, BB, pulmonary vein area (PVA), and LA]. Mapping of the RA was performed from the inferior to superior caval vein perpendicular to the terminal crest. The PVA was accessed through the oblique sinus along the boundary of the left and right PV. Mapping of the LA was performed from the lower border of the left inferior PV towards the LA appendage. Additionally, BB was mapped from the tip of the LA appendage across the top of the LA and RA, behind the aorta towards the superior cavo-atrial junction. Five seconds of sinus rhythm (SR) was recorded at each mapping location, with a surface electrocardiogram (ECG) lead, a bipolar reference EGM, and all unipolar epicardial EGMs. Data were processed by amplification (gain 1,000), filtering (bandwidth 0.5-400 Hz), sampling (1 kHz) and analog-to-digital conversion (16 bits), and then saved on a hard disk.

Data analysis
Customized software was used to semi-automatically analyze U-AEGM morphology. U-AEGMs with injury potentials or recording sites with ≥25% missing U-AEGMs were excluded. In addition, U-AEGMs recorded during premature atrial beats or aberrant beats were eliminated. Under the premise that the deflection amplitude was at least twice the signal-tonoise ratio, the steepest negative slope of U-AEGMs was marked as the local activation time (LAT). All U-AEGM annotations were manually inspected by two researchers with consensus, and color-coded activation maps were reconstructed using the LATs on each electrode (upper right panel of Figure 1). Consistent with Konings et al. [12], U-AEGMs were categorized as single potentials (SPs), short double potentials (SDPs), long double potentials (LDPs) and fractionated potentials (FPs). In brief, SPs only consist of a single negative deflection; SDPs and LDPs contain two negative deflections with time interval between deflections of respectively ≤15ms and >15 ms; FPs were defined as ≥3 deflections. Examples of the different U-AEGM morphologies are shown in the lower left panel of Figure 1.

Statistical analysis
A Shapiro-Wilk test was applied to inspect the distribution of continuous variables. Continuous variables conforming to a normal distribution were described as mean ± standard deviation (SD), and the differences were compared using an independent t-test. Skewed distributed variables were described as median and interquartile range (IQR), and a Mann-Whitney U test was performed to compare differences between groups. Categorical variables were described as the number and percentage, and the differences were assessed by χ 2 test. A P-value (two-sides) less than 0.05 indicates significant difference.
Univariable linear regression analysis was performed to investigate which variables were associated with different U-AEGM morphologies (including SP, SDP, LDP, and FP). We further explored the independent relationship between age (independent variable) and different U-AEGMs morphologies (dependent variables) using a multivariable linear regression model. In this process of multivariable linear regression analysis, two models were performed: model I, covariates related to different U-AEGM morphologies in univariable linear regression analysis were adjusted; model II, all covariates, including body mass index, gender, hypertension, dyslipidemia, diabetes mellitus, myocardial infarction, left ventricular function, left atrial dilatation, and medication, were adjusted. Results were described as regression coefficient (β) and 95% confidence interval (CI).
Furthermore, a generalized linear model (GLM) was performed to visualize the independent relationship between age (considering as continuous variable) and different U-AEGM morphologies (considering as continuous variables), with the same adjustment for covariates in model I and model II. R software (version 4.1.3) and IBM SPSS Statistics (version 28) were used to analyze the data.

Study population
A total of 213 patients with an age ranging from 37 to 84 years [median age: 67 (59-73), 85.45% male) were enrolled. This study population was categorized into a young (age < 60 years, N = 58, 82.76% male) and aged (

U-AEGM database
A total of 2,007,403 U-AEGMs (9,424 ± 2,818 per patient) with a mean SR cycle length of 886 ± 166 ms were recorded. The number of U-AEGMs obtained from the RA, BB, PVA and LA area was respectively 938,093 (46.73%), 227,465 (11.33%), 445,334 (22.19%) and 396,511 (19.75%). Characteristics of U-AEGMs recorded from all patients for every atrial region separately are shown in Supplemental Table S1. At each atrial region, the proportion of SPs was the largest (RA: 82.49%, BB: 79.76%, PVA: 83.58% and LA: 81.79%), whereas the proportion of FPs was the smallest (RA: 2.33%, BB: 2.60%, PVA: 1.52% and LA: 1.97%). Figure 2 shows a color-coded distribution of SPs, SDPs, LDPs and FPs measured at all mapping areas in four typical patients of various ages (42-, 62-, 70-and 76-year-old). These maps show that at all areas, but particularly at BB, there are age-related changes in U-AEGM morphology consisting of an increase in non-SP.

Impact of ageing on U-AEGM morphology
The impact of ageing on U-AEGM morphology in the different areas in the entire study population is summarized in Table 2. In the aged group, only at BB the proportion of SPs was significantly lower compared to the young group (p = 0.007) whereas the proportions of SDPs (border significant, p = 0.051), LDPs (p = 0.004) and FPs (p = 0.006) were higher. However, an increased age did not have significant impact on U-AEGM morphology recorded from the RA, PVA and LA.

Age-related changes in BB U-AEGM morphology
Univariable predictors of age-related changes in U-AEGMs at BB with their respective β (95% CI) are summarized in Table 3. In univariable analyses, ageing was associated     with a decrease in the proportion of SPs (β = −5.43, p = 0.006), and an increase in the proportions of LDPs (β = 1.90, p = 0.024) and FPs (β = 1.71, p = 0.005). Furthermore, the use of ACEI/ARB/AT2 antagonists was associated with an increase in the proportion of SPs (β = 4.14, p = 0.032) and a reduction in the proportion of LDPs (β = −1.88, p = 0.020); male (β = −3.46, p = 0.022) was associated with a decrease in the proportion of SDPs, while LA dilatation (β = 4.18, p = 0.024) was correlated with an increased SDP proportion. Additionally, we also found that digoxin usage was associated with a higher proportion of LDPs (β = 7.89, p = 0.040).
Variables introduced in the multivariable linear regression model of age-related changes U-AEGM morphology at BB are shown in Table 4. In model I, variables associated with different U-AEGMs morphologies (p < 0.05) in univariable linear regression were adjusted. Age was associated with a reduction in the proportion of SPs (β = −5.50, p = 0.005), and an increase in the proportions of LDPs (β = 1.78, p = 0.031) and FPs (β = 1.71, p = 0.005). No significant correlation between age and SDPs was found in model I (p = 0.128). The relation between age and U-AEGM morphology at BB is illustrated in the upper panel of Figure 3. After adjusting for all potential covariates in model II, age was associated with a decrease in the proportion of SPs (β = −6.33, p = 0.002), an increase in proportion of SDPs (β = 2.49, p = 0.042), LDPs (β = 1.94, p = 0.028) and FPs (β = 1.90, p = 0.004). The independent relationship between age and BB U-AEGM morphology is shown in the lower panel of Figure 3.

Key findings
This is the first study exploring the relation between age and morphology of U-AEGMs obtained from both atria including BB. For this purpose, we used a large sample of 2,007,403 U-AEGMs. Ageing resulted in a substantial alteration in U-AEGM morphology at BB, manifested by a reduction in the proportion of SPs and an increase in proportion of SDPs, LDPs and FPs. However, ageing had no significant effect on U-AEGM morphology recorded from the RA, PVA, and LA.

Regional differences in age-related EGM morphology
Double potentials (including SDPs and LDPs) and FPs are frequently associated with areas of conduction delay and/ or block [12][13][14] which play a fundamental role in the pathophysiology of atrial tachyarrhythmias such as AF. Significant alterations in EGM morphology with increasing age have been demonstrated in previous studies, although these only focused on the RA. Roberts-Thomson et al. [15] categorized 21 patients without a history of AF into three groups (age < 30 years [N = 7]; 31 < age < 59 years [N = 6]; age > 60 years [N = 8]), and compared the proportion of bipolar complex fractionated atrial EGMs (CFAE) during SR in the RA. Their results demonstrated that the proportion of CFAE in the oldest patients was significantly higher compared to the youngest patients (14.6 ± 7.7% vs 2.7 ± 2.1%, p = 0.001), but no difference was observed compared to the middle group (8.5 ± 3.5%, p = 0.14).
Unlike the study of Roberts-Thomson et al. [15], our study did not show an influence of ageing on the U-AEGM morphology at the RA. However, as the youngest group in the study of Roberts-Thomson et al. [15] was much younger than our young group, only the middle and oldest groups can be compared to our results. Then, indeed, no difference could be found in the proportion of CFAE with age, which is consistent with the results of our present study. It should still be noted that Roberts-Thomson et al. [15] used bipolar endocardial EGMs compared to our unipolar epicardial EGMs. As Van der Does et al. [16] demonstrated that there are no differences between endo-and epicardial U-AEGM morphology recorded at the RA, our findings can still be extrapolated to endocardial U-AEGMs.
In another study of 106 patients without a history of AF, Centurion et al. [17] indicated that the number of abnormal atrial EGMs (defined as fractionation duration ≥100 ms and/or ≥8 negative deflections) in the RA during SR was considerably higher in patients over 60 years compared to younger (13 to 60 years) patients (0.61 ± 1.43 vs 0.14 ± 0.44, p < 0.02). The age of the patients in the young group in the study of Centurion et al. [17] was also considerably younger compared to our young group (age range from 37 to 60 years). It could therefore be that the largest differences in EGM morphology at the RA can only be found at an earlier age than included in our present study. This implies that the largest part of age-related remodeling occurs at a younger age (before approximately 50 years) after which it gradually continues, which may explain why in the present study there were no significant alterations in U-AEGM morphologies at RA (and also LA and PVA).
However, we investigated for the first time BB and found that ageing was associated with considerable changes in U-AEGM morphology at this site. This observation suggests that BB is extra vulnerable for age related remodeling. As prior studies already suggested that BB plays an important role in the pathophysiology of AF [18,19], our observations may partly explain why elderly are more prone to develop atrial tachyarrhythmia such as AF.

Age related structural remodeling
Spach et al. [20] demonstrated that extensive collagenous septa caused by ageing resulted in electrical uncoupling of the side-to-side connections at RA, which in turn promotes variability in wave propagation direction and the complexity of EGMs. As a consequence, age-related electrophysiological alterations resulted in significant reduction in conduction velocity of transverse propagation, which makes reentry more likely to occur. Xu et al. [21] investigated the influence of ageing on atrial cellular properties in an in-vivo rat model, and demonstrated that the ratio of interstitial fibrotic areas to atrial surface area (2.1 ± 0.6% vs 1.0 ± 0.3%, p < 0.05) and cellular diameter (5.3 ± 1.1 μm vs 4.1 ± 0.8 μm, p < 0.05) at the LA were significantly higher in middle-aged rats (9 months) compared with those measured from young rats (3 months), but not at the RA. In a canine model, Anyukhovsky et al. [22] found that dogs older than 8 years had more connective tissue (8.4 ± 1.0% vs 4.8 ± 1.1%, p < 0.05) in the RA compared with dogs of 1-5 years. In addition, they found that large strands of connective tissue separated muscle bundles of elderly atrial tissue into smaller components. In our study, abnormal atrial EGMs related to ageing occurred especially at BB, indicating that age-related remodeling particularly affects BB. This observation could indicate that the parallel oriented muscle bundles contained within BB are more easily disrupted by structural remodeling. However, this hypothesis still needs further verification by subsequent histopathological studies.

Study limitations
Morphology of U-AEGMs was not repeatedly measured at different ages of each individual patient. Therefore, the influence of individual heterogeneity cannot be excluded. Due to the invasive characteristics of intraoperative mapping, it was not possible to include patients of all ages. In addition, it cannot be completely excluded that various occlusion sites affect the atria differently. However, it is expected that the influence of coronary stenosis is comparable between the young and aged group.

Conclusions
Ageing affects particularly BB as indicated by the increased amount of abnormal U-AEGM (SDPs, LDPs and FPs) recorded from this location in the elderly. This observation further supports the concept that BB may play an important role in development of age-related arrhythmias such as AF.

Author contributions
ZY, MS and NG designed the study. ZY and MS analyzed the data and draft the manuscript. AH, LN, FR and YT contributed to data acquisition and critically revising the manuscript. NG contributed to conceptual thinking, interpretation and critically revised the manuscript. All authors approved the final version to be published and agreed to be accountable for all aspects of the work.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Data availability statement
The data underlying this article will be shared on reasonable request to the corresponding author.