Author + information
- Received December 16, 2014
- Revision received February 10, 2015
- Accepted February 14, 2015
- Published online March 1, 2015.
- Matthew M. Zipse, MD,
- Joseph L. Schuller, MD,
- David S. Steckman, MD,
- David F. Katz, MD,
- Wendy S. Tzou, MD,
- Duy T. Nguyen, MD,
- Ryan G. Aleong, MD,
- Raphael K. Sung, MD,
- Christine Tompkins, MD,
- Paul D. Varosy, MD and
- William H. Sauer, MD∗ ()
- ↵∗Reprint requests and correspondence:
Dr. William H. Sauer, Section of Cardiac Electrophysiology, University of Colorado Hospital, 12401 East 17th Avenue, B136, Aurora, Colorado 80045.
Objectives The study sought to characterize the performance of implanted leads among a cohort of patients with cardiac sarcoidosis (CS) and implantable cardiac-defibrillators (ICDs).
Background An ICD is indicated for some patients with CS for the prevention of sudden cardiac death. CS can lead to myocardial inflammation and scar that may interfere with lead performance.
Methods We performed a case-control study within the cohort of patients at the University of Colorado Hospital with CS and an ICD (n = 48) compared with randomly selected controls (n = 117) who had other indications for an ICD. We compared the measured lead parameters at the time of routine interrogation to assess the differences between groups over time. The mean duration of follow-up was 51 months. Survival analysis was performed by the method of Kaplan and Meier and by Cox proportional hazards regression.
Results There was no significant difference in measured lead impedance, capture thresholds, or sensed electrograms at implantation between the CS and control groups. There were no significant differences between the mean parameters between groups over the follow-up period. However, patients with CS have a high incidence of significant (>50%) drop in measured electrograms (16 of 46 [33%] CS patients vs. 4 of 117 [3.4%] controls; hazard ratio: 10.49, 95% confidence interval: 3.47 to 31.67). As a result of alterations in lead parameters, 2 patients (4.3%) required lead revision, and 6 patients (13%) required ICD testing to ensure adequate detection of induced ventricular fibrillation.
Conclusions Reductions over time in ICD sensing of P- and/or R-wave electrograms are common in patients with CS. Although further investigation is needed to determine the mechanism of these changes, these findings suggest that patients with CS who have an ICD should be closely monitored for clinically relevant changes in P- and R-wave amplitudes.
Sarcoidosis is a granulomatous disorder with cardiac involvement diagnosed in approximately 5% of cases, although it is seen in up to 25% of sarcoidosis patients on autopsy (1–7). An ICD is indicated in many patients with cardiac sarcoidosis (CS) due to the high risk of ventricular arrhythmias and sudden cardiac death in these patients (8). CS patients with implantable cardiac-defibrillators (ICDs) have demonstrated a high rate of both appropriate ICD therapies and electrical storm in recent observational data (9–12).
Although the high incidence of appropriate shocks observed in this population may offer supportive evidence for ICD implantation, the long-term outcomes of the ICD transvenous leads and stability of lead parameters is unknown. In addition, in a large cohort of patients with CS and ICDs, there was an 11% incidence of lead-related complications requiring revision (9).
Lead malfunction can lead to the delivery of inappropriate therapies or withholding of appropriate therapies and ultimately necessitate ICD testing and/or lead revision (13–15). The effect of the progressive nature of CS, manifesting both as inflammation and scar, on the transvenous ICD system and lead function over time is unknown. Therefore, we sought to evaluate the performance of implanted leads and measured electrograms in a cohort of patients with CS and ICDs.
Study population and design
We performed a nested case-control study of ICD patients, comparing those with CS (cases) to randomly selected nonsarcoidosis cardiomyopathy patients (controls) followed at the University of Colorado Hospital. All patients included in the CS cohort had biopsy-proven extracardiac sarcoidosis with evidence of myocardial involvement. Device interrogation and clinical and demographic data collection were performed retrospectively in a standardized fashion. Patients were excluded if device follow-up occurred at an outside institution or if only the immediate peri-implantation interrogation data were available. Patients with left ventricular (LV) assist devices were also excluded, as LV assist devices are known to cause alterations in lead parameters (16). The study protocol was reviewed and approved by the Colorado Multiple Institutional Review Board.
ICD interrogation reports and medical records were retrospectively reviewed by a trained physician. Clinical characteristics, ICD manufacturer, system parameters, and lead configuration were recorded. The presence of a lead placed on advisory or recall was also noted. Lead sensing, impedance, and capture threshold parameters were collected at implant and recollected with each routine device interrogation at follow-up. Parameters for right atrial and coronary sinus leads were also recorded when available. Of note, for the purposes of data collection for this study, sensed electrograms were recorded only in sinus rhythm (i.e., reductions in sensed P-wave electrograms were excluded if these were due to the presence of atrial fibrillation or other atrial arrhythmias at the time of device interrogation). The sensed electrogram amplitude was recorded in millivolts, lead impedance was recorded in ohms, and lead capture threshold was measured as volts with variable pulse widths. For standard comparisons to be made, thresholds were expressed in microjoules by the following formula:Occurrences of appropriate and inappropriate therapies (shocks and/or antitachycardia pacing) were documented. Subsequent lead-related interventions were also recorded.
Criteria for diagnosis of CS and ICD implantation
Patients in this study were considered to meet criteria for the clinical diagnosis of CS by the 2006 Revised Japanese Ministry of Health and Welfare criteria (7), modified to include electrophysiologic testing with right ventricular (RV) voltage mapping and also with an emphasis on delayed contrast enhancement by cardiac magnetic resonance, which was considered a major criteria. This schema overlaps with the more recently released 2014 Heart Rhythm Society expert consensus guideline for the diagnosis of CS (8). All patients met both criteria for clinical diagnosis of CS or presence of “probable” CS. Those patients diagnosed with CS underwent ICD implantation at the discretion of the consulting electrophysiologist. Not every patient diagnosed with CS underwent ICD implantation. ICD implantation was performed using standard and accepted technique, and no patients in either arm of the study were implanted with a subcutaneous ICD. If ICD testing was not performed at the time of implantation due to a contraindication, patients returned for testing at 3 months post-implant.
Maximum, minimum, and mean values of lead parameters for each patient were calculated for sensed electrogram amplitudes, thresholds, and impedances for both sarcoidosis and control patients. A 50% reduction in sensed electrograms, a 50% change in impedance, and/or an 8-fold increase in pacing energy required for myocardial capture were considered to be significant changes in lead parameters. The number of patients with significant reduction in sensing, capture thresholds, and changes in impedance were compared between the cardiac sarcoid patients and control patients. Survival analysis was performed by the method of Kaplan and Meier and by Cox proportional hazards regression. The Pearson chi-square test was used to compare clinical characteristics between patients with and without lead malfunction by univariate analysis. Statistical analyses were performed using the SPSS (version 18.0, IBM, Armonk, New York) statistical software program, and statistical significance was defined as a 2-sided p value <0.05.
Clinical and electrophysiologic characteristics of study cohort
Table 1 displays the clinical characteristics of the study cohort. We identified 48 cases with CS and an implanted ICD that were followed for a mean of 51 ± 26 months after implantation. Among these patients, 95% were implanted for primary prevention. Sixty-five percent of CS patients received a single-chamber ICD, 24% of patients received a dual-chamber ICD, and 11% of patients received a biventricular ICD. Seventeen percent of patients in the CS group had evidence of LV dysfunction by transthoracic echocardiogram or cardiac magnetic resonance imaging. RV dysfunction was noted in 46% of cases. The mean LV ejection fraction was 56 ± 12%. No CS patient who underwent ICD testing had a high defibrillation threshold or delayed detection of induced ventricular fibrillation (VF) due to ventricular electrogram undersensing.
The control group consisted of 117 patients with other indications for ICD implantation, followed for a mean of 52 ± 19 months after implantation: 77% of patients of the control cohort received an ICD for primary prevention, 74% of patients in the control arm had chronic LV systolic dysfunction (of at least mild severity), and 42% had ischemic heart disease; 50% received a single-chamber ICD, 22% of patients received a dual-chamber ICD, and 28% of patients received a biventricular ICD. The mean LV ejection fraction was 32 ± 15%.
Changes in lead parameter measurements over time
There was no significant difference in capture thresholds and impedance measurements at implant between CS patients and controls, nor were there significant changes in these mean parameters in either group over the follow-up period (Table 2). Aside from 2 patients who underwent lead revision for lead fracture of a Sprint Fidelis ICD (Medtronic, Mounds Views, Minnesota) lead in the control arm, there were no patients who required lead revision for isolated changes in lead impedance.
Changes in electrogram sensing over time
There was no significant difference in the mean R-wave amplitude at time of ICD implantation between patients with CS (11.8 ± 5.2 mV) and other cardiomyopathies (10.9 ± 4.9 mV). There also was no significant difference in the mean minimum R-wave amplitude over the course of follow-up (CS 9.4 ± 4.5 mV vs. other cardiomyopathies 9.6 ± 4.5 mV). Similarly, there was no significant difference in P-wave sensing at implant or follow-up between groups (Table 2).
There was, however, a significant difference in the number of CS patients compared with controls who had a >50% drop in R-wave amplitude over follow-up (10 of 48 [21%] vs. 2 of 117 [1.7%]; hazard ratio [HR]: 12.27, 95% confidence interval [CI]: 2.67 to 56.47). Similarly, a greater proportion of CS patients experienced a 50% drop in P-wave sensing over follow-up (6 of 20 [30%] vs. 2 of 59 [3.4%]; HR: 7.94, 95% CI: 1.60 to 39.41). Combined, the number of CS patients (16 of 48, 33%) compared with controls (4 of 117, 3.4%) who experienced drops in sensed electrograms (R or P wave) was also significant (16 of 48 [33%] vs. 4 of 117 [3.4%]; HR: 10.49, 95% CI: 3.47 to 31.67). The incidence of a 50% reduction in sensing is shown in a time-to-event Kaplan-Meier format in Figure 1, which demonstrates time-dependent reduction in intracardiac electrogram sensing remote from the implant date. Seven of 10 patients who experienced significant reductions in sensed electrograms did so 3 or more years after ICD implantation. Figure 2 depicts individual and averaged sensing trends from implant to the last recorded interrogation in CS cases and controls.
Clinical predictors of reductions in electrogram sensing
Univariate analysis revealed that longer-duration follow-up (>5 years) was a significant predictor of reductions in R-wave sensing (odds ratio [OR]: 9.82, 95% CI: 1.79 to 53.78) (Table 3). A history of atrial arrhythmias was also a significant predictor of reductions in sinus rhythm P-wave sensing (OR: 18.00, 95% CI: 1.19 to 271.48), even though sensing values were only recorded in sinus rhythm. When P- and R-wave reduction events were combined, the presence of LV dysfunction (OR: 4.33, 95% CI: 1.15 to 16.26), RV dysfunction (OR: 4.84, 95% CI: 1.33 to 17.67), and follow-up beyond 5 years (OR: 6.60, 95% CI: 1.75 to 24.85) were significant predictors of reductions in sensing. Although not achieving statistical significance, integrated bipolar leads appeared to provide a protective effect against reductions in sensing compared with dedicated bipolar defibrillator leads (OR: 0.31, 95% CI: 0.06 to 1.65; p = 0.23).
Clinical events related to lead malfunction or undersensing in patients with CS
Some cases of lead undersensing in the CS cohort resulted in significant clinical events. Defibrillation threshold (DFT) testing was performed at the discretion of the patient’s primary electrophysiologist when reduced ventricular sensing was recognized. Six patients underwent DFT testing (all with R waves of <3). In 2 cases, the result of DFT testing led to lead revision. In 1 of these cases, the patient had received multiple inappropriate shocks secondary to T-wave oversensing and had undersensing of VF at the time of DFT testing (Figure 3), necessitating a command internal shock. Three cases of P-wave reduction in CS patients were of no significant clinical consequence, although all cases of low sensed electrograms necessitated continued clinical follow-up.
Association of changes in clinical status in subjects with reductions in electrogram sensing
Immunosuppression was managed in conjunction with each patient’s primary pulmonologist; 54% of CS patients in our cohort received 1 or a combination of immunomodulating medications, most commonly including prednisone, methotrexate, and mycophenolate mofetil. One CS patient with significant reduction in measured R waves was preceded by the patient’s own elective discontinuation of all medications, including prednisone. Subsequent reinitiation had no significant effect on the recovery of sensing, although the reduction in R-wave amplitude in this case was of no clinical consequence. No other cases with reductions in sensing were associated with identified changes in immunosuppression regimen or dosing.
Reassessment of LV and RV function was not uniformly triggered by changes in lead function in this study; however, 1 patient did have a significant change, which corresponded to reduction in sensing (with the development of severe biventricular dysfunction in follow-up, down from mild at the time of device implantation). No other cases with reductions in sensing were associated with deterioration of LV or RV size or systolic function.
This study details our clinical experience with CS and ICDs during long-term follow-up. When observed as a group, lead performance and sensing measurements are comparable to other nonsarcoidosis cardiomyopathies; however, there are a significant number of patients with CS who experience a late reduction in electrogram sensing. Some of these patients required ICD testing with the induction of VF to ensure adequate ICD function and lead revision. It is worth noting that as a result of DFT testing, lead revision for poor R-wave sensing at baseline was avoided in 67% (4 of 6) CS patients who demonstrated adequate sensing during VF induction.
Lead fracture and dislodgment in patients with CS
In our study, we found that ICD patients with CS had a similar rate of lead malfunction due to fracture or dislodgment compared with those with other cardiomyopathies. This is in contrast to a previous analysis of 235 CS patients, in which a 10.6% (25 of 235) rate of lead dislodgment or fracture was observed (9). In our study, there were no cases of lead dislodgment and 2 cases of lead fracture (4.3%). Both cases of fracture occurred with Medtronic Sprint Fidelis ICD leads, which carry a known risk of lead conductor fracture (17).
Potential mechanisms for late sensing reduction
Our results reveal a sizable proportion of patients with CS (33%) had a >50% reduction in a measured bipolar P- or R-wave recorded at some point in their routine device follow-up. It is interesting to speculate on the mechanisms for sensing reduction observed. Patients with CS can demonstrate a waxing and waning course in the level of inflammation and eventual scarring that may involve myocardial tissue near the ICD or pacing lead tip. If there is local scarring within the small bipolar sensing vector between the tip and ring of the sensing lead, a reduced electrogram measurement will be observed. This mechanism may occur locally at the lead tip-tissue interface, independent of clinically evident disease progression, as only 1 patient with reduced sensing had an associated global decline in ventricular function.
Another potential mechanism for the reduced electrogram sensing observed could be changes in myocardial activation due to progressive conduction disease. An altered local vector due to the development of bundle branch or fascicular block could result in reduced sensing despite unchanged impedance or pacing capture threshold. None of the patients with reductions in sensing had developed a new bundle-branch block on electrocardiogram during follow-up, although conceivably, bundle branch block could have been transient only at the time of device interrogation and recorded sensing parameters. Nonetheless, although this can offer an explanation for reductions in ventricular sensing, it cannot explain altered atrial sensing.
Finally, poor sensing could be due to unrecognized lead dislodgement, although this is an unlikely explanation for our results, as other lead parameters remained unchanged. Additionally, most of the R-wave reductions occurred well after the initial implant when lead dislodgement would have been uncommon.
Clinical implications of altered sensing in patients with CS
Our findings have clinical implications for cardiac electrophysiologists managing patients with CS and an ICD. First, consideration of the use of an ICD system that has programmable vectors for R-wave or P-wave sensing should be considered given the high rate of late reduction in sensed electrograms. Second, the standard 5.0-mV criteria that is often used for acceptable R-wave measurements at the time of implant may not be adequate for long-term management in patients with CS. Based on our data, there is a 21% chance that the 5.0-mV R-wave at implant will become <2.5 mV at 4 years, possibly necessitating lead revision.
Last, the use of a subcutaneous implantable cardiac-defibrillator (S-ICD) might be considered for certain patients who do not have an indication for pacing. The wide sensing bipole utilized by the S-ICD may protect against reductions in sensing which otherwise might be manifest as a result of local inflammation or RV scarring at the interface of endocardium and lead tip. Unfortunately, however, many patients with CS have an indication for pacing or will develop a pacing indication over the course of follow-up due to the high incidence of acquired atrioventricular block with CS. This alone may make the S-ICD a less desirable option, even if this device were to offer more reliable long-term sensing.
Potential study limitations
The limitations of observational research must be considered in this study. One such limitation involves the certainty of the CS diagnosis in our cohort. We used a clinical diagnostic criteria scheme that has not been correlated with autopsy studies proving a histologic diagnosis. In addition, there remains the possibility of false positive radiological studies leading to an inaccurate diagnosis. However, we believe that corroborated abnormal cardiac findings in patients with extracardiac sarcoidosis strongly suggest cardiac involvement. If there were a misclassification bias, then this bias would be toward no association (i.e., toward the null hypothesis) as patients without CS would not be expected to have any alterations in electrogram sensing, and therefore this potential bias is unlikely to explain our results.
Another limitation of the study involves the relatively small number of events, reducing our ability to evaluate potential confounders in the multivariable analysis. Our comparison to our control group is not matched by age, gender, or clinical features; however, we believe that an unmatched comparison is useful to clinicians who manage these patients. Significant differences that were observed between groups also would otherwise bias results toward the null, as there is no plausible mechanistic explanation for a protective effect from the presence of diabetes, hypertension, hyperlipidemia, coronary artery disease, or systolic dysfunction (these comorbidities were more common in patients with nonsarcoidosis cardiomyopathies).
Our analysis was also limited by lead parameter data, which was entered manually by review of device clinic records and not from automated lead integrity alerts from remote monitoring. Accordingly, significant changes in lead parameters that were transient may not have been captured; however, this limitation would have similarly affected data gathered for both groups of patients and should not affect the principle findings in this study. If anything, this study might have missed additional cases of poor sensing in CS patients if transient and the result of local inflammation resolved with immunosuppression therapy.
None of the ICD leads implanted in this study were of the passive fixation variety, so a comparison of long-term lead parameters between active and passive fixation mechanisms in CS could not be performed. Although it is possible that passive leads may be less traumatic in CS, they also limit options for lead positioning and may also pose a higher dislodgment risk. In regard to sensing bipole, it is possible that there may be an advantage conferred by use of integrated bipolar leads in CS patients in contemporary practice; this study was simply not sufficiently powered to detect a significant difference. Ultimately, despite the sum of these limitations, our study provides insight into the natural history of these challenging patients.
Patients with CS and an ICD have stable measured lead impedances and pacing thresholds observed over time. However, reductions in P-wave and/or R-wave amplitude are common in patients with CS who have an ICD and are rarely observed in other nonsarcoidosis cardiomyopathies used as a comparison group. Although we have speculated that progressive myocardial inflammation and scarring explains this observation, further investigation is needed to determine the mechanism of these changes. These findings suggest that in patients with CS referred for an ICD, operators should aim for a higher R-wave measured at implantation and should be closely monitored for clinically relevant changes in P-wave and R-wave amplitudes. DFT testing may be indicated if reductions in R waves are observed and can help avoid lead revision if adequate sensing with VF induction is observed. The use of programmable sensing vectors should also be considered at implantation in these patients, as reductions in sensed electrograms are unpredictable.
Patients with CS are at increased risk for sudden death and often receive an ICD. There are case reports and case series indicating a potential association between CS and abnormal lead function after implantation. In addition, the disease has a progressive inflammatory course that may affect cardiac tissue interacting with an implanted lead. In this observational clinical study, we found that patients with CS had a higher incidence of reduced electrogram sensing that sometimes necessitated system revision. This phenomenon became apparent over an extended period of time and thus, we recommend careful follow-up with frequent ICD interrogations in these patients. We also recommend selecting sites with higher measured electrograms at the time of lead implantation to avoid additional testing and/or lead revisions.
Dr. Sauer has received educational and research grant support from manufacturers of defibrillators. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- cardiac sarcoidosis
- defibrillation threshold
- hazard ratio
- implantable cardiac-defibrillator
- left ventricular
- odds ratio
- right ventricular
- subcutaneous implantable cardiac-defibrillator
- ventricular fibrillation
- Received December 16, 2014.
- Revision received February 10, 2015.
- Accepted February 14, 2015.
- American College of Cardiology Foundation
- Silverman K.J.,
- Hutchins G.M.,
- Bulkley B.H.
- Uusimaa P.,
- Ylitalo K.,
- Anttonen O.,
- et al.
- Birnie D.H.,
- Sauer W.,
- Bogun F.,
- et al.
- Kron J.,
- Sauer W.,
- Schuller J.,
- et al.
- Eckstein J.,
- Koller M.T.,
- Zabel M.,
- et al.
- Maisel W.H.,
- Kramer D.B.
- Kleemann T.,
- Becker T.,
- Doenges K.,
- et al.