Authors: Goro Yoshioka, Takanori Yamaguchi, Atsushi Tanaka, Hikari Sakai, Junji Koyamatsu, Toshiharu Umeki, Kohei Kaneta, Yoshiko Sakamoto, Atsushi Kawaguchi, Koichi Node
Categories: Original Article, echocardiography, heart failure, left atrial strain, pacemaker
Source: ESC Heart Failure
Doi: 10.1002/ehf2.14973
This study aimed to investigate the clinical impact of pre‐procedural left atrial strain (LAS) in patients undergoing permanent pacemaker implantation (PPI).
This single‐centre retrospective study enrolled 434 patients who were admitted for transvenous PPI between 2010 and 2020. After excluding patients with persistent atrial fibrillation, PPI for complete atrioventricular block, severe valvular disease, history of open‐heart surgery and those without LAS data, 172 patients were analysed. The LAS was measured using commercially available software to calculate the average strain value of the apical four‐ and two‐chamber views before PPI. The primary composite endpoint was hospitalization due to heart failure or cardiovascular death. Cox proportional hazard models were used to evaluate risk factors for the primary composite endpoint. The mean patient age was 78 ± 8 years, and 42% of the patients were men. PPI was performed for sick sinus syndrome in 64% and second‐degree atrioventricular block in 36% of the patients. The pre‐procedure left atrial reservoir strain (LASr) was 28 ± 11%. The median follow‐up period was 4.7 years, and the primary endpoint was observed in 23 (13%) patients. In multivariate Cox proportional risk analysis, LASr was independently associated with the primary composite endpoint (hazard ratio, 1.08 per 1% decrease; 95% confidence interval, 1.02–1.15; P = 0.007). The receiver operating characteristic curve of the LASr for the primary composite endpoint showed a cutoff value of 21% (area under the curve 0.657, P = 0.004). The prognostic impact of LASr was consistent with that of sick sinus syndrome and atrioventricular block.
A decreased pre‐procedure LASr was associated with long‐term adverse outcomes after PPI use.
Keywords: echocardiography, heart failure, left atrial strain, pacemaker
Permanent pacemaker implantation (PPI) is a well‐established therapy for bradycardia, ^1^ and guidelines worldwide recommend PPI for symptomatic bradycardia. ^2^ , ^3^ , ^4^ Generalization of PPI procedures and technical innovations in pacemaker devices have been achieved in recent decades, ^5^ , ^6^ and adverse events associated with PPI procedures have decreased.
However, adverse cardiac events after PPI use, especially hospitalization for heart failure (HF) and cardiovascular death, remain important issues. ^7^ Therefore, better risk stratification is required to predict adverse events in the remote phase after PPI use. Previous studies have reported that the right ventricular (RV) pacing rate or RV pacing lead location predicts adverse events such as hospitalization due to HF and cardiovascular death. ^8^ , ^9^ , ^10^ However, limited data are available regarding the pre‐procedural prediction of these adverse events.
Left atrial (LA) strain (LAS) is an echocardiographic parameter that reflects the atrial function. ^11^ LAS has been reported to predict new‐onset non‐valvular atrial fibrillation (AF), AF recurrence after catheter ablation, ^12^ , ^13^ embolic events ^14^ and adverse cardiovascular events. ^11^ LAS is also associated with LA structural remodelling and pressure. ^15^ The clinical utility of LAS in patients with HF has recently been reported and is expected to be useful in explaining the pathophysiology of HF. ^16^ , ^17^ However, the detailed clinical characteristics of LAS before PPI use and its prognostic impact on long‐term outcomes after PPI use remain poorly evaluated.
Therefore, the present study evaluated the prognostic impact of pre‐procedural LAS on HF or death in patients undergoing PPI.
This single‐centre, retrospective observational study was conducted at Saga University Hospital, Japan (Figure 1). We included 434 patients who received transvenous PPI for sick sinus syndrome (SSS) or atrioventricular block (AVB) between January 2010 and December 2020. The exclusion criteria were persistent AF at the time of PPI (n = 31), complete AVB (CAVB) (n = 127), history of open‐heart surgery (n = 8) and severe valvular disease (n = 15). Patients with persistent AF were excluded because of the difficulty in measuring each component of the LAS. Patients with CAVB were excluded because of the difficulty in accurately assessing LAS due to atrioventricular conduction defects (Figure 2). Patients with a history of open‐heart surgery or severe valvular disease were also excluded because of the potential impact of these conditions on the LAS and outcomes. Ultimately, 172 patients, including 110 with SSS and 62 with second‐degree AVB, were examined.
Figure 1 Flow diagram of the study cohort. AF, atrial fibrillation.
Figure 2 Left atrial (LA) strain (LAS) in the patient with sick sinus syndrome (SSS) and complete atrioventricular block (CAVB). The LAS measurement was set to begin in the QRS complex, and a global LA strain–time curve was generated. The LA reservoir strain (LASr) was measured as the height between the top and bottom of the strain curve and averaged over two‐ and four‐chamber views. LA contractile strain (LASct) was measured at end‐diastole, and LA conduit strain (LAScd) was calculated as the difference between LASr and LASct. Patients with CAVB were excluded because of the difficulty in accurately assessing LAS due to atrioventricular conduction defects.
All PPI procedures were performed in accordance with the American College of Cardiology/American Heart Association/Heart Rhythm Society 2008 Guidelines for Device‐Based Therapy of Cardiac Rhythm Abnormalities and Guidelines for Non‐Pharmacotherapy of Cardiac Arrhythmias (Japan Circulation Society 2011). ^2^ , ^3^ All patients provided informed consent for the use of PPI. Ethical approval for this study was obtained from the Institutional Review Board of Saga University Hospital (approval reference 2021‐05‐R‐01).
Experienced sonographers performed comprehensive transthoracic echocardiography using commercially available iE33 (Philips Medical Systems, Bothell, WA, USA) or Vivid E9 machines (GE Medical Systems, Milwaukee, WI, USA) prior to PPI. Standard techniques were used to obtain M‐mode, two‐dimensional (2D) and Doppler measurements. Left ventricular (LV) ejection fraction (LVEF) was estimated using the standard biplane Simpson method. 2D LAS and LV strains were measured offline using a commercially available software program (TOMTEC Imaging Systems, Munich, Germany). 2D LAS was calculated using the biplane method by automatically tracing the LA border at end‐systole in apical two‐ and four‐chamber views. The LAS measurement was set to begin in the QRS complex, and a global LA strain–time curve was generated (R–R gating method). ^11^ In this study, LA reservoir strain (LASr) was measured as the height between the top and bottom of the strain curve and averaged over two‐ and four‐chamber views. LA contractile strain (LASct) was measured at end‐diastole, and LA conduit strain (LAScd) was calculated as the difference between LASr and LASct (Figure 2). The global longitudinal strain of the left ventricle (LVGLS) was obtained via speckle tracking using two‐, three‐ and four‐chamber views. The software program automatically detected LV endocardial boundaries. The maximum negative value of systolic deformation represents the maximum contractile force for each segment of the LV wall. The average of these values for each segment was defined as LVGLS.
The following data were collected from the medical records at the time of initial hospitalization for PPI: medical history, signs and symptoms, blood test results, transthoracic echocardiography, 12‐lead electrocardiogram (ECG) (including baseline rhythms), degree of AVB and fascicular block, and clinical outcomes. Post‐procedure ECG findings were reviewed to measure QRS duration.
The primary composite endpoint was defined as the composite of hospitalization for HF or cardiovascular death. HF was diagnosed using the diagnostic criteria of the 2021 European Society of Cardiology (ESC) guidelines for all time periods from 2010 to 2020, ^18^ in which HF was diagnosed by the presence of at least one sign (ascites, rales or radiographic evidence of pulmonary congestion) and one symptom (oedema, dyspnoea or orthopnoea), regardless of the ejection fraction (EF). The secondary endpoints included the individual components of the primary composite endpoint and all‐cause mortality. Clinical follow‐up was performed during the clinic visits. The patients had scheduled clinical visits at 1, 6 and 12 months after PPI and every 12 months thereafter. The cause of death was determined through a detailed review of the medical records and death certificates. The results of the device interrogation at 12 months were reviewed to confirm the RV percentage pacing rate and arrhythmic events. AF was defined as a documented paroxysmal AF event prior to PPI or detection during interrogation within 1 year after PPI. For the patients who were followed up in the regional comprehensive network (n = 30), we confirmed their clinical outcomes by telephone calls to their attending physicians. The LAS was reassessed in patients with follow‐up echocardiographic data at 12 months (±3 months). The frequency of LVEF reduction, defined as <40%, was evaluated in patients with follow‐up echocardiography beyond 1 year after PPI.
Inter‐observer and intra‐observer variability for LASr was studied in a group of 10 randomly selected subjects by one observer, repeated twice, and by two observers who were blinded to the clinical outcomes and each other's measurements.
For continuous variables, normally distributed data were reported as mean ± standard deviation, and nonparametric data were reported as median and inter‐quartile range (IQR). For categorical variables, data are presented as counts and percentages. Comparisons of continuous variables between groups were performed using Student's t test or Kruskal–Wallis test, as appropriate. Categorical variables were compared using the chi‐squared test or Fisher's exact test, as appropriate. The cumulative incidence of each endpoint was calculated using the Kaplan–Meier method. The effects of LASr, LASct and LAScd on primary composite endpoints were determined using a multivariate Cox proportional hazard regression model adjusted for confounding factors as Model 1 (age and sex), Model 2 [age, sex and body surface area (BSA)], Model 3 (age, sex and LVGLS) and Model 4 [age, sex and LA volume index (LAVI)]. To evaluate the relationship between the primary composite endpoint and LASr, patients were divided into two groups according to the LASr cutoff value, and the free survival time of the primary composite endpoint was determined by Kaplan–Meier estimation and compared between the groups using the log‐rank test. The cutoff value was determined using the Youden index based on the primary composite endpoint throughout the follow‐up period. Subgroup analyses were performed using Cox proportional hazard regression models based on the type of bradycardia (SSS or AVB), presence of chronic kidney disease (CKD) [cutoff: estimated glomerular filtration rate (eGFR) ≤60], LAVI (cutoff: 34 mL/m^2^), location of RV lead, RV percentage pacing rate at 1 year (cutoff: 50%) and detection of AF within 1 year after PPI, after excluding subgroups of <50 patients. No adjustment for confounding factors was made in the subgroup analysis due to the small number of events in each subgroup. The cutoff value of the LAVI and RV percentage pacing rate was based on a previous report. ^8^ , ^19^ In addition, the effects of pre‐procedural LASr on LVEF reduction and its timing in the chronic phase after PPI were determined using a multivariate Cox proportional hazard regression model adjusted for the confounding factors of age, sex or LVGLS.
All statistical analyses were performed using JMP PRO® 16 software (SAS Institute Inc., Cary, NC, USA). Statistical significance was defined as a two‐sided P value <0.05.
The baseline patient characteristics, procedural information and medications at discharge are shown in Table 1. The baseline transthoracic echocardiographic parameters are presented in Table 2. The mean age of the patients was 78 ± 8 years, and 42% were male. A total of 64% had SSS, and 36% had second‐degree AVB. Echocardiography was performed for a median of 7 (IQR, 3–18) days prior to PPI.
The median duration of follow‐up after PPI was 4.7 (IQR, 2.3–7.9) years. Overall, the primary composite endpoint was reached in 23 of the 172 (13%) patients. Individual components of the primary composite endpoint were reached in 15 of 172 (9%) patients hospitalized for HF and in 8 of 172 (5%) patients hospitalized for cardiovascular death. The patient characteristics and echocardiographic parameters in the primary and non‐primary endpoint groups are shown in Tables 1 and 2, respectively. LASr, LASct and LAScd were significantly lower in the primary endpoint group than in the non‐primary endpoint group. In multivariate Cox proportional hazards analyses, LASr was independently related to the primary composite endpoint even after covariate adjustments, including LAVI (Table 3). Similarly, LASct and LAScd were independently associated with the primary composite endpoint in different multivariate models (Table 3). Based on these results, we selected LASr from the three LAS parameters for subsequent analyses.
The receiver operating characteristic (ROC) curve for the primary composite endpoint showed a cutoff value of 21% (area under the curve 0.657, P = 0.004). Figure 3 shows Kaplan–Meier curves of clinical events after PPI in patients with LASr ≥ 21% and those with LASr < 21%. Higher event rates were observed in patients with LASr < 21% than in those with LASr ≥ 21% (log‐rank, P = 0.002). The results were consistent between the individual components of the primary composite endpoint (hospitalization for HF and cardiovascular death) and all‐cause death.
Figure 3 Clinical events after permanent pacemaker implantation (PPI). (A) Primary composite endpoint (hospitalization for heart failure or cardiovascular death). (B) Hospitalization for heart failure. (C) Cardiovascular death. (D) All‐cause death. LASr, left atrial reservoir strain.
We performed a subgroup analysis stratified by the type of bradycardia, presence of CKD, LAVI, location of the RV lead and RV percentage pacing rate at 1 year. The baseline clinical characteristics and endpoints for each bradycardia type are shown in Table S1. Cox proportional hazards models demonstrated that LASr was related to the primary composite endpoint in both SSS and second‐degree AVB groups (Table 4). The results of other subgroup analyses, including LAVI, also showed that LASr was related to the primary endpoint.
Follow‐up echocardiography performed at 12 months (±3 months) after PPI was available for 134 of the 172 patients. Ninety‐four patients had a pre‐procedural LASr ≥ 21%, among whom 73 patients maintained an LASr ≥ 21% (Group A) and 21 had an LASr < 21% at follow‐up echocardiography (Group B). In contrast, 40 patients had a pre‐procedural LASr < 21%, with the LASr recovering to ≥21% in 6 patients (Group C) while 34 still had an LASr < 21% (Group D) (Figure S1A). Group B also showed a similar event rate to Group D during a median follow‐up of 3 (IQR, 1.5–5.5) years after follow‐up echocardiography (primary composite 24% vs. 27%, P = 0.825). In contrast, the event rates in Groups A and C were equally low (primary composite 8% vs. 0%, P = 0.320) (Figure S1B).
LVEF data beyond 12 months after PPI were available for 135 patients, with a median follow‐up echocardiographic time of 5.3 (IQR, 3.1–8.1) years. A reduced LVEF, defined as <40%, was observed in 11 of 135 (11%) patients. Even after adjusting for age, sex or LVGLS using multivariate Cox proportional hazard models, LASr was independently associated with reduced LVEF after PPI (Table S2).
The intra‐observer variability of the intraclass correlation coefficient was 0.979, and the inter‐observer variability of the intraclass correlation coefficient was 0.936.
The main findings of the present study were as (1) LASr was independently related to the primary composite endpoint, which was defined as the composite of hospitalization for HF or cardiovascular death; (2) the ROC curve of the LASr for the primary composite endpoint showed a cutoff value of 21% (area under the curve 0.657, P = 0.004); and (3) regardless of the type of bradycardia or LAVI, LASr was independently related to the primary composite endpoint. To the best of our knowledge, this is the first report to demonstrate the prognostic impact of pre‐procedural LASr on the long‐term outcomes after PPI. Patients with a reduced pre‐procedure LASr should be closely followed up and considered for optimal medical therapy for HF at appropriate time. In addition, as LASr is also associated with decline of LVEF after PPI, LAS would be useful information in determining indication for cardiac physiological pacing including left bundle branch pacing. ^20^
Hospitalization for HF and cardiovascular death after PPI use remain important clinical issues. ^7^ , ^21^ High RV pacing burden has been reported as a risk factor for post‐PPI events. RV pacing induces abnormal myocardial contractions, similar to left bundle branch block, ^21^ and RV percentage pacing rate >50% has been reported to decrease LVEF and increase rates of hospitalization due to HF. ^8^ Apex pacing has been associated with abnormal myocardial contractions and HF development. ^9^ In addition, apex pacing has been reported to be associated with an abnormal pattern of particle image velocity measurement of intracardiac flow patterns by echocardiography. ^22^ However, reports on the risk stratification of atrial function and cardiovascular events using speckle tracking methods prior to PPI use are limited.
LAS is a relatively novel parameter that reflects LA function, and LA deformation was assessed using LA 2D speckle tracking imaging. ^23^ This measurement has advantages over conventional echocardiographic parameters (E/A, tissue Doppler and pulmonary vein flow), as it is free from angle alignment and is less influenced by loading status. ^24^ LAS has been reported to be a prognostic factor in various cardiovascular diseases, including HF. In patients with chronic severe mitral regurgitation who underwent successful mitral valve repair surgery, LAS was reported to be an independent predictor of long‐term outcomes. ^25^ Several studies of patients with HF have established the clinical implications of LAS. Regardless of the HF phenotype [HF with reduced EF (HFrEF) or HF with preserved EF (HFpEF)], LAS has been shown to be a predictor of clinical outcomes. ^17^ , ^26^ , ^27^ The present study demonstrated an association between reduced LASr and adverse events, including hospitalization for HF, after PPI.
LA actively cooperates with the LV throughout the cardiac cycle. Therefore, LAS partially reflects LV function and is associated with LV end‐diastolic volume and LV diastolic function. ^15^ As such, it may be a prognostic factor for ventricular and atrial cardiomyopathy. ^28^ , ^29^ The present study also showed that LASr was a predictive factor for reduced LVEF after PPI therapy, suggesting that reduced atrial function precedes and/or causes LV dysfunction in this cohort. Further studies are needed to clarify the causal relationship between LA and LV dysfunctions.
LAS has been reported to be lower in patients with AF than in non‐AF patients ^11^ and predicts new‐onset AF. ^11^ , ^30^ A recent study reported that LAS is a useful predictor of LA fibrosis, which predicts AF recurrence after thoracoscopic ablation. ^31^ These reports suggest that a decrease in LAS is associated with the primary endpoint of 5 year AF recurrence, partly due to the development of AF during the follow‐up period in patients with decreased LAS. However, patients with persistent AF before PPI were excluded from this study, and LASr was consistently associated with the primary composite endpoint, even in the subgroup of patients without AF detection during the follow‐up period. This suggests that LAS is associated with outcomes independent of AF development.
Histological factors associated with AF have recently been identified using atrial biopsy. ^32^ Increased intercellular spacing, myofibrillar loss, decreased myocardial nuclear density and amyloid deposition are also associated with structural remodelling of the atria in patients with AF in addition to fibrosis, ^32^ , ^33^ which has been reported to contribute to the development and persistence of AF. ^34^ Interestingly, a decrease in nuclear density, a surrogate for cardiomyocyte density, is associated with ageing and cardiomyocyte hypertrophy. ^33^ These histological changes were presumably related to atrial dysfunction. Collagen and amyloid accumulation, as well as cardiomyocyte hypertrophy, may decrease the distensibility of the atria. We speculate that these histological changes also occur in patients receiving PPI therapy, which may be associated with poor outcomes. The histological verification of atrial function, particularly in LAS, is warranted in the future. Once these details have been clarified, LAS may be clinically applicable as a surrogate for atrial histology in patients with atrial cardiomyopathy. In addition, histological analysis of atrial dysfunction will help to elucidate the pathogenesis of atrial cardiomyopathy and dysfunction and to identify better patient stratification and new therapeutic targets.
As a noninvasive assessment of the LA function and structural remodelling, a cardiac magnetic resonance (CMR) feature tracking‐guided LA wall motion analysis revealed that LA emptying fractions and LA booster pump are closely associated with the presence and extent of LA low‐voltage area (LVA). ^35^ Late gadolinium‐enhanced (LGE)–CMR imaging has also been developed to identify atrial fibrosis. However, LGE–CMR for the detection of fibrosis has several potential challenges, including a lack of standardized acquisition protocols and processing techniques and low image resolution relative to the thin LA wall. ^36^ , ^37^ A recent study comparing the most common LGE–CMR methods revealed considerable discrepancies in global and regional LGE extent ^38^ and between LVA and LGE. ^39^ In contrast, noninvasive LAS assessment by echocardiography has been reported to be associated with LVA ^33^ and can easily assess serial changes. The present study demonstrates not only the significance of evaluating LAS with follow‐up echocardiography but also the robustness of LAS in predicting post‐procedural outcomes. Furthermore, echocardiographic LAS assessment has the advantage of being more economical and accessible than CMR and may be an important surrogate measure of histological changes.
Several limitations of the present study warrant mention. First, this was a retrospective observational study conducted at a single centre in Japan, which limits the generalizability of our findings. It should also be noted that the enrolment duration was relatively long. However, all diagnoses and procedures were performed according to the guidelines; therefore, clinical practice was consistent for 11 years. Second, the study cohort included only patients with available LAS data prior to PPI, which is a potential source of selection bias. The methodology of LAS analysis needed to exclude persistent AF, PPI for CAVB, severe valvular disease or prior open‐heart surgery, because LAS could not be accurately evaluated. Therefore, the results of this study may only be applicable to a limited population. Third, His bundle pacing and left bundle branch pacing, which were reported to be associated with the recovery of LA function, ^20^ were not included. This partly limits the generalizability of our findings to contemporary practices. Fourth, hospitalization for HF was based on physician judgement. Fifth, 25% of our study cohort lacked data on the RV percentage pacing rate after 1 year, which was reported to have a predictive impact after PPI; therefore, we were unable to include the RV percentage pacing rate in the subgroup analysis. Sixth, because of the small number of cohorts and lack of internal and external validation, further studies are necessary to confirm the predictive values and cutoff values of LAS. Finally, the study cohort included patients with both SSS and AVB. Although subgroup analyses for each diagnosis confirmed the significance of the LAS, an investigation targeting each diagnosis is desirable.
Our findings suggest that reduced pre‐procedure LASr is significantly associated with long‐term adverse cardiovascular events after PPI use. Patients with reduced pre‐procedure LASr should be followed up with particular care. Our findings underscore the clinical importance of LAS in patients undergoing PPI treatment.
The authors have no conflicts of interest to declare.
Yoshioka, G. , Yamaguchi, T. , Tanaka, A. , Sakai, H. , Koyamatsu, J. , Umeki, T. , Kaneta, K. , Sakamoto, Y. , Kawaguchi, A. , and Node, K. (2024) Impact of left atrial strain on clinical outcomes in patients with permanent pacemaker implantation. ESC Heart Failure, 3982–3992. 10.1002/ehf2.14973.
Data are available upon reasonable request to the corresponding author.
Data are available upon reasonable request to the corresponding author.