OUP user menu

Sinus node dysfunction in atrial fibrillation patients: the evidence of regional atrial substrate remodelling

Hung-Yu Chang, Yenn-Jiang Lin, Li-Wei Lo, Shih-Lin Chang, Yu-Feng Hu, Cheng-Hung Li, Tze-Fan Chao, Wei-Hsian Yin, Shih-Ann Chen
DOI: http://dx.doi.org/10.1093/europace/eus219 205-211 First published online: 6 July 2012

Abstract

Aims It remains unclear as to whether regional atrial substrates of certain areas of the atrium in patients with atrial fibrillation (AF) can be related to sinoatrial node dysfunction. We investigated the relationship between the biatrial substrate characteristics and sinus node function in these patients.

Methods and results The study enrolled 34 patients (aged 57 ± 11 years old; 20 males) who underwent catheter ablation for symptomatic paroxysmal AF. Sinus node dysfunction was defined as having corrected sinus node recovery time longer than 550 ms. Atrial substrate analyses of both atria and atrial conductive properties were investigated in patients with (Group 1) and without sinus node dysfunction (Group 2). The mean global bipolar voltage of both atria and the atrial refractory period were similar between the two groups. Regional analysis showed that the mean bipolar voltage for patients in Group 1 was lower than in Group 2 (1.0 ± 0.3 vs. 2.1 ± 0.7 mV, P < 0.001) only in the sinus node region, while the electrophysiological properties were similar for both groups in other anatomic regions of both atria. The right atrial total activation time was significantly longer (97 ± 9 vs. 89 ± 10 ms, P = 0.023) and the conduction velocity along the crista terminalis was significantly slower (1.0 ± 0.2 vs. 1.2 ± 0.3 m/s, P = 0.019) in Group 1 patients than in Group 2 patients.

Conclusion In patients with AF, regional atrial remodelling near the sinus node area was associated with sinus node dysfunction.

  • Atrial fibrillation
  • Catheter ablation
  • Mapping
  • Sinus node

Introduction

The association between sinoatrial node dysfunction (SND) and atrial tachyarrhythmia has been identified for a long time.1 Atrial fibrillation (AF) develops in up to 50% of patients with SND,2 and the development of AF may further exacerbate SND due to the high atrial rate.3 However, evidence that AF leads to SND, or vice versa, has been controversial.4,5 Previous study has shown that pacing-induced AF can cause SND in a canine model, and that the SND recovers within a week after AF termination.4 On the other hand, the bradycardia-related inhomogeneous refractoriness in SND enhances the AF vulnerability.5

Many previous studies have also shown that SND is associated with poor right atrial (RA) substrate properties. Sanders et al.6 observed diffuse atrial anatomic and structural abnormalities in patients with SND. In patients with both congestive heart failure and SND, substrate remodelling with structural changes were demonstrated.7 In the rabbit hearts AF model, Kirchhof and Allessie showed that a high degree of sinoatrial entrance block was present that protected the pacemaker fibres in the centre of the sinus node against the rapid fibrillatory impulses. However, the high activation rate in the sinoatrial border during AF prevented the sinus impulses from exiting the sinus node.8

It remains unclear as to whether regional atrial substrates of certain areas of the atrium in patients with AF can be related to SND. Tachycardia-induced remodelling of the sinoatrial node ion channel expression has recently been demonstrated.9 We hypothesized that the sinoatrial node regional remodelling, which can down-regulate the ion channel expression, decreasing automaticity of spontaneous sinus activity, and slowing down the conduction velocity along the sinoatrial border, can be demonstrated in the electrophysiological laboratory with the assistance of a three-dimensional electroanatomic mapping system. The purpose of this study was to investigate the relationship between regional atrial substrate remodelling and sinus node function in AF patients.

Methods

Study population

The study enrolled 34 patients (aged 57 ± 11 years old; 20 males) with frequent episodes of drug refractory, symptomatic paroxysmal AF who underwent catheter ablation in our institute. All anti-arrhythmic drugs except amiodarone were discontinued for at least five half-lives before the procedure. Four (11.8%) patients continued taking amiodarone before the procedure.

Electrophysiological study and sinus node function test

The details have been described in our previous works.10,11 In brief, after providing informed consent, each patient underwent an electrophysiological study and catheter ablation while in the fasting non-sedated state. The sinus node function was assessed after the ablation procedure. The sinus node recovery time was assessed at cycle lengths of 500, 450, 400, and 300 ms after a 60 s pacing train and determined as the time from the stimulus artefact to the earliest atrial activity. The corrected sinus node recovery time (CSNRT) was calculated as the difference between the sinus node recovery time and baseline cycle length. Sinoatrial node dysfunction was defined as CSNRT longer than 550 ms in >2 cycle lengths. Patients were divided into two groups: with SND (Group 1, n = 13) and without SND (Group 2, n = 21).

Electroanatomic mapping

Biatrial electroanatomic mapping was performed by a three-dimensional mapping system (NavX, Version 7.0, St Jude Medical, Minnetonka, MN, USA). Either an irrigated-tip 3.5 mm ablation catheter or a 4 mm tip conventional ablation catheter was selected as the roving catheter. After completing an intact atrial geometry construction, the RA geometry was further divided into the following seven parts, consistent with our previous work12: high anterolateral wall, low anterolateral wall, high posterior wall, low posterior wall, high septum, low septum, and cavotricuspid isthmus (Figure 1).

Figure 1

The division of the right atrium into seven segments in anteroposterior and posteroanterior views.

The roving catheter was used to detect the local activation time (relative to the time reference signal recorded by the coronary sinus catheter) and local voltages while being swiped throughout the left and right atrium during sinus rhythm and coming into contact with the atrial wall. The bipolar electrograms were filtered between 32 and 300 Hz and recorded digitally. High-density mapping was performed along the crista terminalis and superior vena cava–RA junction. The signal from the roving catheter was used to build the sequential activation and voltage map. In the sinus rhythm activation map, the sinus node region was identified as the area within 0.5 cm radius of the earliest activation site.

Atrial substrate and voltage analysis

The left atrial (LA) and RA bipolar peak-to-peak voltages (PPV) were obtained from more than 350 points approximately equally distributed within both atria. The mean PPV of both atria, and of different atrial parts were analysed. The low-voltage zone (LVZ) was defined as the area with a bipolar PPV voltage amplitude of ≤0.5 mV. The LVZ index was calculated as the LVZ surface area divided by the total atrial surface area. An index of the heterogeneity of the bipolar voltage amplitude was obtained by calculating the coefficient of variance (defined as the standard deviation/mean value) of the voltage of all points.

Atrial conduction

The atrial total activation time was defined as the time elapsed between the earliest activation at any site in the right atrium and the latest activation at any site in the left atrium. P-wave duration and PR interval were measured simultaneously from 12 leads of the surface electrocardiography and averaged from a series of five consecutive beats. The conduction velocity along the crista terminalis was calculated by dividing the linear distance between the high and low crista terminalis by the difference of their local activation times.

Effective refractory period

Atrial effective refractory periods (ERPs) were evaluated at twice diastolic threshold at driving cycle lengths of 600 and 500 ms. A decremental technique was used, starting with an S2 coupling interval of 400 ms, which was decreased in decrements of 10 ms. The atrial ERP was defined as the longest coupling interval failing to propagate to the atrium.

Statistical analysis

The results are presented as the mean value ± standard deviation or percentiles. The Student' t-test was used to analyse the continuous data and the χ2 test with a Pearson correction or Fisher's exact test was used to compare the non-parametric data in the different groups. Statistical significance was selected at a value of P < 0.05.

Results

Patient characteristics

The study consisted of 34 patients who underwent catheter ablation for AF with a history of paroxysmal AF over 6 ± 5 years. The arrhythmias were refractory to 2 ± 1 anti-arrhythmic drugs. Twelve patients (35%) had hypertensive cardiovascular disease, 3 (9%) diabetes mellitus, 9 (27%) hyperlipidaemia, and 1 (3%) a history of stroke. The mean LA diameter was 38 ± 6 mm, and mean left ventricular ejection fraction was 59 ± 6%. Table 1 demonstrates the clinical characteristics in both groups and all clinical parameters were similar between groups, except that the incidence of permanent pacemaker implantation was higher in Group 1 than that in Group 2 (23 vs. 0%, P = 0.048), and the pre-ablation heart rate was slower in Group 1 than that in Group 2 (65 ± 6 vs. 75 ± 16 b.p.m., P = 0.037).

View this table:
Table 1

The clinical characteristics of atrial fibrillation patients with and without sinus node dysfunction

Group 1 (n = 13)Group 2 (n = 21)P value
Age (years)61 ± 855 ± 120.11
Male6 (46%)14 (67%)0.24
Body mass index (kg/m2)23 ± 324 ± 20.52
Duration of atrial fibrillation (years)7 ± 65 ± 30.26
Anti-arrhythmic drug (n)2 ± 12 ± 10.99
Use of amiodarone before the procedure2 (15%)2 (10%)0.63
Hypertension6 (46%)6 (29%)0.46
Diabetes2 (15%)1 (5%)0.54
Hyperlipidaemia4 (31%)5 (24%)0.70
Cerebrovascular attack1 (8%)0 (0%)0.38
Coronary artery disease1 (8%)1 (5%)1.00
CHADS2 score1 ± 10 ± 10.19
Permanent pacemaker implantation3 (23%)0 (0%)0.048
Systolic blood pressure before ablation (mmHg)130 ± 19128 ± 160.78
Diastolic blood pressure before ablation (mmHg)75 ± 1077 ± 100.66
Heart rate before ablation (b.p.m.)65 ± 675 ± 160.037
Left atrial diameter (mm)39 ± 837 ± 50.28
Right atrial enlargement2 (15%)1 (5%)0.54
Left ventricular end diastolic diameter (mm)47 ± 447 ± 50.87
Left ventricular ejection fraction (%)59 ± 659 ± 60.72
Right ventricular systolic pressure (mmHg)30 ± 728 ± 60.43

Sinus node function

Table 2 demonstrates the electrophysiological characteristics in both patient groups. The sinus node recovery time was significantly longer in Group 1 patients than that in Group 2 (2115 ± 398 vs. 1190 ± 283 ms, P < 0.001). The baseline RR interval was similar between both groups, and the CSNRT was significantly prolonged in Group 1 patients compared with Group 2 patients (1042 ± 390 vs. 348 ± 125 ms, P < 0.001). The CSNRT was significantly longer in patients with pacemaker implantation than that in other Group 1 patients (1393 ± 61 vs. 942 ± 392 ms, P = 0.023). There were no differences in the sinus node recovery time and the CSNRT, regardless of amiodarone intake.

View this table:
Table 2

The electrophysiological characteristics of atrial fibrillation patients with and without sinus node dysfunction

Group 1 (n = 13)Group 2 (n = 21)P value
Sinus node recovery time (ms)2115 ± 3981190 ± 283<0.001
Baseline R–R interval (ms)943 ± 112837 ± 2070.17
Corrected sinus node recovery time (ms)1042 ± 390348 ± 125<0.001
Arrhythmogenic veins and triggers
 Pulmonary veins13 (100%)19 (91%)0.51
 Extra-pulmonary vein ectopies2 (15%)4 (19%)1.00
Electroanatomic mapping
 LA mapping points269 ± 40288 ± 550.26
 RA mapping points195 ± 44228 ± 720.15
 Sinus node area mapping points8 ± 210 ± 20.11
 LA mean peak-to-peak voltage (mV)1.3 ± 0.41.7 ± 0.60.09
 RA mean peak-to-peak voltage (mV)1.3 ± 0.51.5 ± 0.50.22
 RA regional voltage mapping
  Sinus node area mean voltage (mV)1.0 ± 0.32.1 ± 0.7<0.001
  High anterolateral wall mean voltage (mV)1.2 ± 0.41.3 ± 0.60.76
  Low anterolateral wall mean voltage (mV)1.9 ± 0.92.2 ± 1.20.71
  High posterior wall mean voltage (mV)1.9 ± 0.72.1 ± 0.60.59
  Low posterior wall mean voltage (mV)1.3 ± 0.91.3 ± 0.50.92
  High septum mean voltage (mV)1.7 ± 0.31.7 ± 0.60.94
  Low septum mean voltage (mV)1.4 ± 0.61.7 ± 0.50.30
  Cavotricuspid isthmus mean voltage (mV)1.6 ± 0.91.4 ± 0.40.34
 Coefficient of variance of LA voltage0.98 ± 0.110.96 ± 0.250.80
 LA low voltage zone index (%)34 ± 1330 ± 190.59
 Coefficient of variance of RA voltage1.00 ± 0.160.96 ± 0.160.48
 RA low voltage zone index (%)33 ± 1428 ± 100.24
Superior vena cava-RA junction to sinus node distance (mm)9 ± 44 ± 70.026
Atrial conduction
 Atrial total activation time (ms)123 ± 14113 ± 200.13
 LA total activation time (ms)92 ± 1089 ± 160.62
 RA total activation time (ms)97 ± 989 ± 100.023
 Conduction velocity along the crista terminalis (m/s)1.0 ± 0.21.2 ± 0.30.019
 P-wave duration (ms)102 ± 1497 ± 140.32
 PR interval (ms)172 ± 15161 ± 190.09
Atrial effective refractory period (ms)230 ± 13231 ± 360.93
  • LA, left atrial; RA, right atrial.

Sinus node location

The position of the earliest sinus activation site is variable. Overall, the mean distance from the superior vena cava-RA junction to the earliest sinus activation site was 6 ± 6 mm (−10 to 18 mm). The distance was significantly longer in Group 1 than in Group 2 patients (9 ± 4 vs. 4 ± 7 mm, P = 0.026).

Atrial conduction

The P-wave duration and PR interval were similar in the two groups. The atrial total activation time and the LA total activation time did not differ between the two groups. The RA total activation time was significantly longer (97 ± 9 vs. 89 ± 10 ms, P = 0.023) and the conduction velocity along the crista terminalis was significantly slower (1.0 ± 0.2 vs. 1.2 ± 0.3 m/s, P = 0.019) in Group 1 than in Group 2 patients. Figure 2B and D are the RA activation maps of two patients with and without SND, which demonstrate a slower conduction velocity along the crista terminalis in the patient with SND.

Figure 2

Right atrial voltage and activation map of a patient with sinus node dysfunction (A,B) and a patient without sinus node dysfunction (C,D). Grey-coloured area represents the scar region with voltage amplitude of the bipolar peak-to-peak voltage of ≤0.5 mV. Note that the scar is located mainly at the sinus node area in the patient with sinus node dysfunction. An abnormal atrial electrogram with multiple rapid deflections, fractionated signals, and low amplitude is shown in (A).

Atrial substrate characteristics

Table 2 demonstrates the bi-atrial voltage collection during the sinus rhythm, and that the mean bipolar voltage of the entire right atrium and of the entire left atrium did not differ between the two groups. The coefficient of variance of the RA and of the LA voltage was similar for the two groups. The RA and LA LVZ indices were similar for the two groups. The atrial ERP were similar for the two groups.

Regional analysis showed that patients in Group 1 had lower mean bipolar voltage of the sinus node region than those in Group 2 (1.0 ± 0.3 vs. 2.1 ± 0.7 mV, P < 0.001). The electrophysiological properties were similar in the other anatomic regions of the right atrium for the two groups. Figure 2A and C are the RA voltage maps of two patients with and without SND, which show lower bipolar voltage near the sinus node area in the patient with SND.

The CSNRT was correlated with the RA total activation time (r = 0.46, P = 0.006) and was inversely correlated with the sinus node area voltage (r = −0.42, P = 0.014). The CSNRT was not correlated with LA total activation, atrial total activation time, or mean global RA voltage.

Discussion

Major findings

Our study demonstrates that in patients with paroxysmal AF, regional atrial remodelling near the sinus node area of the right atrium was associated with SND. The regional atrial remodelling was characterized by slow conduction during sinus rhythm and lower bipolar electrogram voltage of the sinus node region. Other anatomic regions were not associated with SND.

Electrophysiological and electroanatomic remodelling

We found that in patients with AF, the electrophysiological properties, such as atrial ERP, P-wave duration and PR interval, and atrial total activation time were similar in patients with or without SND. The electroanatomic studies showed that the overall biatrial substrates, presenting by the mean global biatrial PPVs, the LVZ indices, and the heterogeneities were also similar in the two groups. When analysing the regional RA substrate, the sinus node region voltage in patients with SND was significantly lower than that in patients without SND, while the substrate properties were similar in the other anatomic regions for the two groups.

Sanders et al.6 demonstrated that in the patients with unexplained sinus bradycardia or sinus pauses of >3 s, there was widespread RA low-voltage amplitude and spontaneous scarring, particularly along the crista terminalis, when compared with normal controls. However, regional analysis was not specified to the sinus node region and LA substrate was not analysed in that study. Moreover, patients with AF were excluded from that study, a significant difference from our study cohort.

Our study demonstrated similar global RA substrate characteristics in the AF patients with or without SND. Therefore, there might be two different aspects of SND. In patients with pure SND without coexisting AF, the RA substrate degeneration could be more diffuse. In patients with AF and SND, the RA substrate changes are regional, limited to the region near the sinus node. Moreover, these regional changes might be transient and reversible after the treatment of AF, as Hocini et al.13 have shown in AF patients with prolonged sinus pauses receiving catheter ablation.

In this study, the atrial conduction test showed that the LA total activation times were similar in patients with or without SND. In contrast, the RA total activation time was significantly longer and the conduction velocity along the crista terminalis was significantly slower in Group 1 than in Group 2 patients. The P-wave durations, and the atrial total activation times did not differ between the two groups. The prolongation of atrial conduction, the slow conduction velocity, and the decreased regional atrial voltage indicated the importance of the substrate remodelling near the sinus node area in AF patients with SND.

Although the sinus node lies typically at the subepicardial position of the junction of the superior vena cava and the right atrium next to the crista terminalis,14 the position of the leading pacemaker site can be highly variable.15,16 Using a multi-electrode array, the remodelled atria due to chronic atrial flutter were associated with more caudal localization of the sinus node complex.17 In current study, a point-by-point activation map was used, and we also demonstrated a more caudal localization of the earliest sinus activation site in patients with SND. Regional scarring at this earliest activation site was correlated with impaired sinus node function test. This finding might reflect the loss of pacemaker automatic tissue and the delay of the electrical impulse exiting from the subepicardial sinus node.

Atrial fibrillation, sinus node dysfunction, and structural abnormalities

Several mechanisms may explain the low-amplitude electrograms in patients with atrial arrhythmias. Atrial fibrillation can lead to altered patterns of connexins, changes in myocyte cellular structure, intersitital fibrosis, and apoptotic atrial myocyte death.18 Previous cellular electrophysiological study had demonstrated that diseased atrial tissue was relatively unexcitable with decrease of resting membrane potential.19 Less-depolarized myocardial cells and more interstitial fibrosis in the diseased myocardium would decrease the recorded voltage.20 Moreover, previous study had demonstrated that the voltage distribution was functionally dependent on cycle lengths and activation patterns in patients with atrial tachyarrhythmia,21 and the excised disease atrial tissue from patients with atrial arrhythmias showed slow response action potential during short atrial pacing cycle lengths.22 It could be caused by non-uniform arrangement of atrial fibres in the diseased atrium that may enhance anisotrophy and affect the conduction velocity; therefore, atrial fibre will be activated asynchronously, and this would explain the lower-amplitude electrograms.23

Diffuse atrial structural remodelling has also been shown to develop in congestive cardiac failure2 and in the presence of chronic RA stretch associated with an atrial septal defect,24 which could result in reduction in the sinus node reserve. However, the mechanism of structural changes in patients with SND alone has remained unclear. Ageing of the atrium might partially explain the impaired sinus node function. Diffuse atrial remodelling, particularly marked around the region of the crista terminalis, has been shown to be associated with ageing in human patients without structural heart disease or atrial arrhythmias, who do not have clinical presentations of SND.25 In our study, the age was similar in both groups, and the mean age was relatively younger, suggesting that the findings cannot be explained by the ageing process alone. The duration of AF is theoretically also an important factor for structural remodelling. However, there was no significant difference in AF duration between our study groups.

As yet, it is still not understood as to why diffuse atriopathy in AF patients would cause SND in some patients, but remain unapparent in others. The current study demonstrates that regional atrial remodelling near the sinus node area was associated with SND. Further studies are warranted to clarify the mechanism.

Study limitation

In the current study, it is difficult to evaluate the serial sinus function changes throughout the AF period, since it is impractical to perform sinus function test and atrial substrate mapping before catheter ablation, then allow patients to have their AF for some length of time, and re-evaluate at the time of the ablation procedure. Secondly, some patients were taking amiodarone before the ablation procedure. This might have affected some of our results, although no differences in sinus node function dependent on amiodarone intake were noted in our study. Thirdly, we did not measure the sino-atrial conduction time, and this might be one of the limitations of our study.

Conclusion

In the patients with AF, advanced regional atrial substrate change near the sinus node area was associated with a prolonged sinus node recovery time, a longer RA activation time, a lower voltage of the sinoatrial node area, and a slower conduction velocity along the crista terminalis. This study indicates the important role of the atrial substrate near the sinus node area in the maintenance of sinus nodal function in AF patients.

Acknowledgements

This study was supported by National Science Council (NSC), Grant No. 99-2628-B-075-007-MY3, 99-2627-B-008-003, and 100-2221-E-008-008-MY2; Joint foundation of VGH and NCU, grant No. VGHUST100-G 1-4-3; NSC support from the Center for Dynamical Biomarkers and Translational Medicine, National Central University, Taiwan, grant No. NSC 99-2911-I-008-100 and NSC100-2911-I-008-001.

Conflict of interest: none declared.

References

View Abstract