Survival of patients with pacing‐induced cardiomyopathy upgraded to CRT does not depend on defibrillation therapy

Background: Permanent right ventricular pacing (RVP) results in cardiac dyssynchrony that may lead to heart failure and may be an indication for the use of cardiac resynchronization therapy (CRT). The study aimed to evaluate predictors of outcomes in patients with pacing‐induced cardiomyopathy (PICM) if upgraded to CRT. Methods: One hundred fifteen patients, 75.0 years old (IQR 67.0–80.0), were upgraded to CRT due to the decline in left ventricle ejection fraction (LVEF) caused by the long‐term RVP. A retrospective analysis was performed using data from hospital and outpatient clinic records and survival data from the National Health System. Results: The prior percentage of RVP was 100.0% (IQR 97.0‐100.0), with a QRS duration of 180.0 ms (IQR 160.0‐200.0). LVEF at the time of the upgrade procedure was 27.0% (IQR 21.0‐32.75). The mean follow‐up was 980 ± 522 days. The primary endpoint, death from any cause, was met by 26 (22%) patients. Age > 82 years (HR 5.96; 95% CI 2.24‐15.89; P =.0004) and pre‐CRT implantation LVEF < 20% (HR 5.63; 95%CI 2.19‐14.47; P =.0003), but neither the cardioverter‐defibrillator (ICD) implantation (HR 1.00; 95%CI 0.45–2.22; P = 1.00), nor the presence of atrial fibrillation (HR 1.22; 95%CI 0.56–2.64; P =.62), were independently associated with all‐cause mortality. Conclusion: Advanced age and an extremely low LVEF, but neither the presence of atrial fibrillation nor implanting an additional high voltage lead, influence the all‐cause mortality in patients after long‐term RVP, when upgraded to CRT.

Keywords: cardiac resynchronization therapy upgrade; chronic right ventricular pacing; defibrillator; pacing‐induced cardiomyopathy

• Abbreviations
• ACE‐I angiotensin‐converting–enzyme inhibitors
• AF atrial fibrillation
• ARB angiotensin‐receptor blocker
• AV block atrioventricular block
• BiV biventricular
• CDRIE cardiac device‐related infective endocarditis
• CIED cardiac implantable electronic devices
• CRT‐D cardiac resynchronization therapy defibrillator
• CRT‐P cardiac resynchronization therapy pacemaker
• EDD end‐diastolic dimension
• ESD end‐systolic dimension
• HFrEF heart failure with reduced ejection fraction
• HV height voltage
• ICD IMPLANTABLE CARDIOVERTER DEfiBRILLATOR
• LBBB left bundle branch block
• LVEF left ventricular ejection fraction
• MI myocardial infarction
• MR mitral regurgitation
• MRA mineralocorticoid‐receptor antagonists
• NYHA The New York Heart Association Functional Classification
• OMT optimal medical therapy
• PICM pacing‐induced cardiomyopathy
• PM pacemaker
• RF ablation radiofrequency ablation
• RVIT right ventricular inflow tract dimension
• RVP right ventricular pacing burden
• SCD sudden cardiac death
• STEMI ST elevation myocardial infarction
• sVT sustained ventricular tachycardia
• TAPSE tricuspid annular plane systolic excursion
• VF ventricular fibrillation

There is evidence that right ventricular pacing (RVP) may cause heart failure with a reduced ejection fraction (HFrEF), and the risk of occurrence depends on the duration and percentage of the RVP.[1] In paced patients with a normal ejection fraction, the risk of HFrEF development ranges between 10% and 25% in 4‐6 years; however, pacing‐induced cardiomyopathy (PICM) symptoms may show up even after one month of RVP with ≥20% burden of pacing.[[2]] The inappropriate electrical pulse propagation in heart muscle by omitting the native His‐Purkinje conduction system, as well as the subsequent contractile dyssynchrony, is considered to be the cause of PICM. Therefore, cardiac resynchronization therapy (CRT) was introduced to correct the dyssynchrony.[[4]] The current European Society of Cardiology guidelines give II B recommendations for an upgrade from right ventricular to biventricular (BiV) pacing in patients who develop worsening Heart Failure (HF) due to a high percentage of RVP, despite Optimal Medical Therapy (OMT).[6] However, there are only a few data on whether the addition of implantable cardioverter defibrillator (ICD) to CRT in the PICM population affects survival.[[7], [9]]

From 2011 up to 2017 in the Department of Electrocardiology, Upper‐Silesian Medical Center, Katowice, we performed 1326 CRT procedures, including first‐device implantation and the replacement of generators. We retrospectively reviewed the charts of 181 patients who underwent the upgrade to CRT with or without defibrillation therapy because of newly diagnosed and/or remaining symptomatic HF despite OMT. At the time of hospitalization, when the upgrade was done, we performed a comprehensive clinical evaluation with all patients, including echocardiographic examination, interrogation of the device, and review of charts from the first implantation, if available. If the medical history suggested ischemic disease as being the cause of decreased left ventricular ejection fraction (LVEF), we performed coronary angiography, and if this was the case, 3‐6 months after coronary angioplasty, we reassessed LVEF. We excluded patients with native intraventricular conduction disturbances and with the presence of reversible causes of non‐ischemic cardiomyopathy, for example, tachycardiomyopathy, chronic alcohol consumption, or thyrotoxicosis. For the remaining patients, the indications to upgrade from conventional PM or ICD to cardiac resynchronization therapy were the signs of PICM defined as cases with RV pacing burden of ≥40% as well as the decline of LVEF < 50% and HF worsening. This required a change of stimulation mode to CRT in the absence of factors adversely affecting prognosis other than RVP. According to the data from clinical studies confirming the development of PICM both in patients with preserved[[3], [10]] and reduced LVEF,[[2], [4], [11]] we decided to qualify the analysis for all patients described above, regardless of the initial LV systolic function. The decision to use or not to use ICD depended on the discretion of both electrophysiologists and heart failure specialists. The data to evaluate all‐cause mortality was received from the Silesian Regional Branch of National Health Fund, the obligatory health insurer in Poland. The BiV pacing burden and arrhythmic events were analyzed retrospectively using available data from medical records, including hospitalizations, outpatients, and telemonitoring.

Absolute numbers and percentages are shown for discrete variables and were compared using the χ2 test. Continuous variables were expressed as a mean ± SD or median (IQR) and compared using the Student's T‐test and the one‐way ANOVA test or the Mann‐Whitney U‐test and the Kruskal‐Wallis H‐test, respectively. An optimal cutoff value of LVEF and age to predict all‐cause mortality were determined using receiver‐operating characteristic (ROC) curve analysis. Time‐dependent comparisons were made using Kaplan‐Meier estimates with the log‐rank test as follows: (a) For the whole population of upgraded patients to assess the effect of resynchronization, (b) for the population of patients upgraded from PM to CRT Pacemaker (CRT‐P) versus CRT defibrillator (CRT‐D) to assess the impact of the ICD implantation. To determine which factors were independently related to clinical outcome, a multivariate Cox proportional hazard regression analysis was performed. A two‐tailed P‐value of.05 was considered statistically significant. All analyses were done using MedCalc Statistical Software version 18.9.1 (MedCalc, Belgium).

A total of 115 patients, 75.0 years old (IQR, 67.0‐80.0), of whom 20 were females, without native bundle branch block, were included. Notably, 79 patients (68.7%) were upgraded from PM and 36 (31.3%) from ICD. In the PM group, 40 patients (50.6%) with no history of ventricular arrhythmia were upgraded to CRT‐P, whereas 39 patients (49.4%) were upgraded to CRT‐D, of whom four (10.3%) for secondary prevention of sudden cardiac death (SCD). All patients in the ICD group were upgraded to CRT‐D, including 17 persons (47.2%) implanted for secondary prevention of SCD. The mean follow‐up was 980 ± 522 days.

Initial LVEF at the time of the first implantation in the PM group was 42.3 ± 11.6%, and in the ICD group, 28.9 ± 9.1%. The mean decrease in LVEF during the observation period in the PM group and in the ICD group was Δ15.9 ± 11.9%, P < .0001, and Δ3.95 ± 8.5%, respectively; P = .01. Thus, LVEF at the time of the upgrade procedure in the PM group was no different than in the ICD group: 27.0% (IQR 20.0‐33.0) versus 26.5% (IQR 22.0‐30.0), P = .25. The median of pacing burden was 100.0% in both groups (IQR 97.0‐100.0). What was different was the pacing period before the upgrade procedure for the PM and the ICD groups, which was 7.9 years (IQR 4.4‐9.4) and 4.7 years (IQR 2.5‐5.8), respectively; P = .0008. The median width of the paced QRS complex was 180.0 ms (IQR 160.0‐200.0) in both groups. QRS narrowing after upgrade to CRT was observed in 96 (85.7%) of patients. The mean shortening of QRS duration time was 20% compared to initial right‐ventricular‐paced QRS (ΔQRS 30 ms; 95% CI 30–40 ms; P < .0001). In Table 1, we have presented the characteristics of patients at the time of upgrade procedure, divided into three subgroups: upgraded from PM to CRT‐P (the PM/CRT‐P group, left ventricular lead implantation), upgraded from PM to CRT‐D (the PM/CRT‐D group, left ventricular and high voltage lead implantation) and upgraded from ICD to CRT‐D (the ICD/CRT‐D group, left ventricular lead implantation).

1 TABLEClinical and echocardiographic characteristics for the groups of patients upgraded from: (A) PM to CRT‐P, (B) PM to CRT‐D, (C) ICD to CRT‐D (at the time of up‐grade)

1 Abbreviations: PM, pacemaker; CRT‐P, cardiac resynchronization therapy pacemaker, CRT‐D, cardiac resynchronization therapy defibrillator, STEMI, ST‐elevation myocardial infarction; AV Block, atrioventricular block; AF, atrial fibrillation; NYHA, The New York Heart Association Functional Classification; MR, mitral regurgitation; EDD, end‐diastolic dimension; ESD, end‐systolic dimension; LVEF, left ventricular ejection fraction; RVIT, right ventricular inflow tract dimension; TAPSE, tricuspid annular plane systolic excursion; LBBB, left bundle branch block; RVP, right ventricular pacing burden; ACE‐I, angiotensin‐converting–enzyme inhibitors; ARB, angiotensin‐receptor blocker; MRA, mineralocorticoid‐receptor antagonists.

During follow up, the median percentage of BiV stimulation did not differ between the subgroups: 98.5% (IQR 97.0‐100.0) for the PM/CRT‐P group, 98.0% (94.6‐99.7) for the PM/CRT‐D group and 98.4% (IQR 95.0‐100.0) for the ICD/CRT‐D group respectively; P = .74. The BiV pacing burden in the atrial fibrillation (AF) group was as high as in the sinus rhythm group: 98.0% (IQR 94.6‐99.4) versus 98.6% (IQR 96.45‐100.0), respectively; P = .095.

Among the 36 subjects from the ICD/CRT‐D group, sustained ventricular tachycardia or ventricular fibrillation (sVT/VF) episodes were found in 16 patients (44.4%), including seven patients (19.4%) implanted for primary prevention of SCD. In this group, only one subject suffered multiple inadequate interventions due to RV lead damage. In the PM/CRT‐D group of 39 patients, VT/VF episodes were found in three subjects (7.69%) implanted for primary prevention of SCD and in one subject (2.56%) implanted for secondary prevention of SCD. The high voltage (HV) therapies appeared in three of 4 patients, and all of them were classified as adequate. Among the 40 subjects in the PM/CRT‐P group, sVT episodes were found in four patients (10%), leading to subsequent upgrade to CRT‐D in three cases: (a) Male, 60 years old, with coronary artery disease and stable angina – for whom sustained and stable VT occurred during the HF decompensation episode 3 years after upgrade procedure. (b) Female, 57 years old, with stable NYHA II dilated cardiomyopathy – for whom sustained VT 190 bpm with syncope occurred 15 months after upgrade to CRT‐P. (c) Male, 70 years old, with NYHA III non‐ischemic chronic HF – unstable sVT appeared after an infectious episode during hospitalization. Re‐upgrade to CRT‐D was performed after the exclusion of cardiac device‐related infective endocarditis (CDRIE). The procedure was challenging and required subclavian vein angioplasty. In the last case, a 90‐year‐old male, with sVT episodes during infection with fever 2 years post upgrade to CRT‐P procedure, was qualified to conservative treatment with amiodarone. The patient died due to decompensated HF 1 month later in the district hospital with no documented episodes of arrhythmia.

During the follow‐up after upgrade to CRT in subjects with no previous VT history, the number of sVT episodes was four times lower in patients upgraded from PM than in patients upgraded from ICD: 9.3% versus 36.8%, respectively; P = .003. Additional analyses show that the only factor that influences the probability of VT occurrence in patients upgraded from PM to CRT‐P or CRT‐D for primary prevention of SCD was LVEF of ≤25%, AUC 0.74; sensitivity 87.50; specificity 64.79; P = .009.

A total of 26 patients died during the observation period, of whom 11 (27.5%) from the PM/CRT‐P group, seven (17.9%) from the PM/CRT‐D group, and eight (22.2%) from the ICD/CRT‐D group; P = .60. Up to 13 out of 26 deaths (50.0%), including elderly patients, occurred at home, and the cause of death was pronounced by a general practitioner without a postmortem examination or the possibility of reading and analyzing data from the implanted device.

Patients who upgraded to CRT‐P without ICD therapy option were on average, 5 years older; however, they presented slightly better LV systolic function, lower LV dimensions, and less frequent regional LV akinesis compared to the PM/CRT‐D and the ICD/CRT‐D groups (Table 1). In addition, the PM/CRT‐P patients required less frequent use of loop diuretics and mineralocorticoid‐receptor antagonists (MRA) to control the symptoms of HF. However, survival prognosis between groups was not different for comparison between the CRT‐P and the CRT‐D group, log‐rank P = 1.00, as well as for comparison between the PM/CRT‐P and the PM/CRT‐D group, log‐rank P = .98 (Figures 1 and 2).

Differences in all‐cause mortality were not visible between the groups of patients with CRT‐D implanted in primary and secondary prevention of SCD, log‐rank P = .83.

Neither paced QRS width, log‐rank P = .72, nor QRS morphology concordant with the Strauss criteria for LBBB, log‐rank P = .63, had prognostic value for all‐cause mortality in the entire study group. Similarly, the presence of AF, probably due to the high rate of BiV pacing, did not affect mortality, log‐rank P = .62. The number of deaths in the group of patients with previous myocardial infarction with ST‐segment elevation (STEMI) and/or scars visible at echocardiography (primarily ischemic cardiomyopathy) was higher than in the rest of the population (20; 29.4% vs 6; 12.8%, P = 0.04). However, this did not affect survival after the upgrade, log‐rank P = .14. Patients with the New York Heart Association (NYHA) Functional Class IV and NYHA Class II/III, 4 (57.1%), and 21 (20.0%), respectively, died during the observation, P = .02. Kaplan‐Meier survival curves separated early, in the first year of observation, log‐rank P < .0002, Figure 3. In patients with severe mitral regurgitation (MR), there was increased mortality (8 [36.4%] patients died) compared to patients with mild to moderate MR, 16 [18.4%] patients died), p = 0.07. Severe MR affects survival, log‐rank P = .02, Figure 4. Extremely low LVEF at the time of the upgrade was found to be a significant negative predictor of survival, with cut‐off value for LVEF < 20%, AUC 0.68, sensitivity 30.77%, and specificity of 92.13%, and in survival analysis, log‐rank P < .0001, Figure 5. In the subgroup of patients with primarily normal LV systolic function, with a mean decrease in LVEF of Δ 23.9 ± 7.8%, P < .0001, extremely low LVEF at the time of the upgrade also seemed to be a negative predictor of survival (trend, log‐rank P = .09), with cutoff value for LVEF < 20%, AUC 0.775, sensitivity 40.0%, and specificity of 100.0%. The outcome of this subgroup did not differ from the outcome of patients with baseline LVEF < 50% (log‐rank P = .81).

Similarly, the age at the time of the upgrade we revealed as a negative predictor of survival, AUC 0.66, sensitivity 30.77%, and a specificity of 93.26% for the cutoff value > 82 years old; in survival estimation, log‐rank P < .0002, Figure 6.

In multivariate analysis, ages over 82 years and an LVEF of < 20% were found to be the only independent predictors of all‐cause mortality (Table 2, Figure 7).

2 TABLEUnivariate and multivariate Cox regression analyses to identify parameters associated with all‐cause mortality after the upgrade to CRT

2 Abrreviations: CRT‐P, cardiac resynchronization therapy pacemaker; CRT‐D, cardiac resynchronization therapy defibrillator; STEMI, ST elevation myocardial infarction; AF, atrial fibrillation; LVEF, left ventricular ejection fraction; MR, mitral regurgitation.

About 20‐30% of CRT implantations are currently upgrading procedures from either PM or ICD.[13] Although the CRT implant success rate has reached 94‐98% in experienced centers,[11] and LVEF improvement is observed by up to 80% of patients,[[3], [14]] the risk of complications in upgrading procedures is one of the highest among cardiac implantable electronic devices (CIED) procedures.[15] In the population of patients with implanted PM qualified to upgrade to CRT, it is necessary to resolve up to two clinical problems: (a) whether changing the stimulation mode to resynchronization therapy by LV lead implantation or His bundle pacing will improve the prognosis; and (b) whether implantation of an additional HV lead or replacing the pacing lead with an HV one is reasonable, taking into account higher perioperative risk and potential adverse lead‐tricuspid valve interactions. The balance of clinical benefit and periprocedural complication risk in patients upgraded to CRT is uncertain. Standard prognostic factors, such as sex, HF etiology, sinus rhythm, QRS width, LV volume, and tissue Doppler assessment of dyssynchrony, are not reliable in this population.[[16], [18]] There are many possible reasons for this.

First, when compared with subjects who are qualified for CRT de novo, the upgrade cohort is more advanced in age and has additional comorbidities that affect survival.[[5], [19], [21]] On average, the frailty syndrome is diagnosed in 68% of patients who qualify for an upgrade procedure,[22] which has an impact on their physical fitness and exercise capacity, as described by the NYHA class. In a population of elderly patients with low LVEF, the major cause of death is non‐arrhythmic and linked to HF progression, severe comorbidities, and a generally poor condition, so it reduces the impact of HV therapy on the overall prognosis.[23] Similarly, in the presented population, advanced age over 82 years and NYHA class IV affected survival, but ICD therapy did not.

Second, the patterns of electromechanical dyssynchrony induced by RVP and LBBB are different. The paced QRS width is often larger than in the native LBBB and similar to the findings of the presented study, it significantly exceeds 150 ms.[[13], [19]] This demonstrates more advanced electrical dyssynchrony, which leads to PICM development.[[2], [16], [20]] In addition, RVP‐induced mechanical dyssynchrony, which is assessed by two‐dimensional, speckle‐tracking echocardiography, is characterized by more frequent early apical septal segment contraction, compared to more frequent basal or mid‐septal segments activation in LBBB, worsened LV torsion and delayed untwisting.[[24]] The narrowing of QRS complexes, as well as the resolution of electromechanical dyssynchrony, for example, apical rocking and septal flash in echocardiography after CRT, is a well‐described predictor of reverse LV remodeling, LVEF improvement, and a reduction in the number of ventricular arrhythmia episodes and mortality.[[11], [16], [21], [24]] Correction of severe dyssynchrony seems to be crucial in the PICM population in the absence of a history of ventricular arrhythmias influencing survival despite the use of HV therapy.[[2], [7], [9]] Our findings, by directly comparing the PM/CRT‐P versus the PM/CRT‐D group, seems to confirm these observations in the patients at high risk of PICM, without severe, irreversible LV dysfunction.

It should be emphasized that echocardiography remains a very important element of prognostic assessment. As was recently presented, the severity of RVP‐induced mechanical dyssynchrony and PICM development is mostly dependent on the initial lowered LVEF.[[2], [4], [12]] Unfortunately, in a large proportion of our patients, LVEF was at least mildly impaired at the time of the first implantation, which could worsen the prognosis. But significant decreases in LVEF were seen even in groups with initially normal LV contractility, as was described in other publications.[[3], [10]] In the presented study, severe HFrEF with LVEF < 20% before the upgrade was a predictor of mortality. Additionally, the LVEF value with a cutoff of ≤25% was a predictor of sVT occurrence in patients upgraded from PM with no history of previous ventricular arrhytmia. However, the limited size of the cohort prevents establishing a definite correlation. Recently, Tayal et al[24] presented the importance of the initial low LVEF value with a cutoff of <24% as a negative predictor for a response to CRT. In turn, Chang et al[18] documented a cutoff of LVEF ≥43.5%, indicating 100% specificity for clinical improvement. Importantly, Gage et al[19] documented that ischemic HF in patients with high RVP burden does not exclude the beneficial effects of resynchronization therapy. Similarly, in the present study, a history of STEMI or areas of myocardial akinesia visible in echocardiography did not affect mortality. Also, severe functional MR as a predictor of overall mortality proved to be significant in univariate analysis but not in multivariate analysis.

There was no relationship between RVP time and mortality in the studied cohort and some other studies.[[14], [16]] However, there is data showing that shorter RVP time is in favor of predicted resynchronization success.[26] These observations may be essential for the decision on early intervention in each case of suspected PICM, especially when HF progression is fast.

Last but not least, surprisingly, chronic AF had no impact on mortality in the PICM population. Preserved atrioventricular (AV) conduction causes a low biventricular pacing rate in patients with high rate AF and native LBBB. This phenomenon is responsible for the lack of effect of CRT on LVEF and survival.[27] In patients with impaired AV conduction and high RVP burden before the upgrade procedure, this problem is negligible.[[16]] However, our patients were predominantly paced from RV before the upgrade with a median of RVP burden equal to 100% (the PICM population), which suggests the lack of AV conduction and predicts a good response to CRT in the majority of them. The reason for the high stimulation rate was second‐ and third‐degree atrioventricular block in 60.9% of patients and pharmacotherapy for HF and associated ventricular arrhythmias with beta‐blockers, digoxin, and amiodarone. The BiV pacing burden, even in the AF subgroup, was very high, with a median of 98.0% (IQR 94.6‐99.4), and was sufficient to achieve reverse LV remodeling and the clinical response. Therefore, it proves that, in the upgraded population with the diagnosis of PICM, widespread atrioventricular disturbances ensure the effect of resynchronization even in patients with AF.

Advanced age and extremely low LVEF, but not the presence of atrial fibrillation, influence all‐cause mortality in patients after long‐term right ventricular pacing and when upgraded to CRT. In patients with PICM upgraded from PM, implanting an additional high‐voltage lead or replacing pacing lead with a high‐voltage one, brings no additional clinical benefit.

Danuta Loboda was associated with study concept, data collection, statistics, data analysis, and manuscript preparation. Michal Gibinski was associated with data analysis and manuscript preparation. Karolina Zietek and Sylwia Gladysz‐Wanha were associated with data collection. Jacek Wilczek and Rafal Gardas were associated with data analysis and manuscript review. Krzysztof S. Golba was associated with statistics, data analysis, and critical revision of the manuscript.