Author + information
- Received October 11, 2016
- Revision received November 29, 2016
- Accepted November 29, 2016
- Published online June 19, 2017.
- Sean D. Pokorney, MD, MBAa,∗ (, )
- Craig S. Parzynski, MSb,
- James P. Daubert, MDa,
- Donald D. Hegland, MDa,
- Paul D. Varosy, MDc,
- Jeptha P. Curtis, MDb and
- Sana M. Al-Khatib, MD, MHSa
- aDuke University Medical Center, Durham, North Carolina
- bCenter for Outcomes Research & Evaluation, Yale School of Medicine, New Haven, Connecticut
- cUniversity of Colorado Denver School of Medicine, Denver, Colorado
- ↵∗Address for correspondence:
Dr. Sean D. Pokorney, Electrophysiology Section, Duke University Medical Center, DUMC 3845, Durham, North Carolina 27710.
Objectives This analysis evaluated temporal trends and factors associated with the use of dual-coil implantable cardioverter-defibrillator (ICD) leads.
Background Data suggest that dual-coil ICD leads are not associated with lower mortality and can be more difficult to extract than single-coil leads.
Methods A total of 435,772 patients at 1,690 hospitals underwent ICD lead insertion in the National Cardiovascular Data Registry’s ICD Registry between April 2010 and December 2015. Hospitals were classified into 3 pre-specified groups (low, decreasing, or high use) based on the frequency of dual-coil lead use.
Results Nationally, the use of dual-coil leads has decreased over time, from 87% of ICD leads in early 2010 to 55% at the end of 2015. Hospitals in the low-use (n = 292) or decreasing-use (n = 561) group had more ICDs inserted by electrophysiologists compared to the high-use (n = 837) group (90% or 80% vs 46%; p < 0.001 for both) and more extractions performed (median 7 or 11 vs 2; p < 0.001 for both). Despite statistical differences, there were no clinically significant differences in patient characteristics across all 3 groups.
Conclusions Although the use of dual-coil ICD leads has decreased over time, it continues to represent the majority of insertions in the United States. Hospital-level factors, but not patient factors, were associated with use of dual-coil ICD leads. Whether decreasing dual-coil ICD lead use has improved patient outcomes remains unknown and should be examined in large, multicenter, contemporaneous patient groups.
In the landmark implantable cardioverter-defibrillator (ICD) trials, the decision to insert a single- versus dual-coil ICD lead was at the physician’s discretion (1,2). Studies of the potential benefits of dual- versus single-coil ICD leads have yielded conflicting results. Some studies have shown that dual-coil ICD leads do not change defibrillation threshold (DFT) relative to single-coil leads (3–6), whereas other studies have found a statistically significant decrease in DFT with a dual-coil ICD lead (7,8). Although secondary analyses from the MADIT-CRT (Multicenter Automatic Defibrillator Implantation Trial with Cardiac Resynchronization Therapy) and the SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) had conflicting findings regarding the association between dual-coil ICD leads and DFT, neither analysis showed a statistically significant difference in mortality or first shock efficacy between single- and dual-coil ICD leads (3,8). Potential disadvantages of dual-coil ICD leads include higher cost, more frequent venous obstruction, and increased adverse events during lead extraction (9,10).
The proportion of single-coil ICD leads being used has increased over time (11), but no data on the characteristics associated with single- versus dual-coil ICD use have been published. This study explored patient and hospital characteristics associated with trends in the use of single- versus dual-coil leads over time.
Data from the National Cardiovascular Data Registry’s (NCDR) ICD Registry were used. The quality of data entered into the registry was ensured by quality checks, outlier analyses, and audits (12). Version 2.1 of the ICD Registry, which included lead data, began in the second quarter of 2010 and was used for this analysis.
This analysis included primary and secondary prevention, initial ICD insertions in the ICD Registry between April 2010 and December 2015. A total of 481,517 patients underwent initial ICD insertion during the study period. There were 44,765 (9.3%) patients with missing or insufficient data to classify the inserted ICD lead as single or dual coil; these patients were excluded from the analysis. An additional 980 patients (0.2%) were excluded because their ICD insertion was outside the study date range. After these exclusions, the study population consisted of 435,772 patients at 1,690 hospitals.
The national and hospital-level percentages of inserted single- versus dual-coil ICD leads were calculated for each quarter between the second quarter of 2010 and the fourth quarter of 2015. Assuming a censored normal distribution, we used latent class growth curve statistical methodology to model the frequency of dual-coil ICD lead use over time by grouping the hospitals into 3 separate phenotypic groups: low-, decreasing- (high use at the beginning and low use at the end), or high-use hospitals (13). The group trajectories were identified from the data by allowing hospitals with similar trends to be grouped using the latent growth curve analysis. A 3-group (phenotype) model was pre-specified based on clinical significance, model fit statistics, and the ability to discriminate between groups by the latent class growth curve model. The model calculated the probability of individual hospitals belonging to each discrete, mutually exclusive group, and the hospital was assigned to the single group with the highest probability of membership. Based on the modeling, there was a projected or predicted trend in dual-coil use for each group over time, and this closely matched the actual or observed trend in dual-coil use for each group over time.
Using hospital characteristics and aggregated patient-level information, a multinomial logistic model was simultaneously modeled to evaluate the association between these characteristics and the identified groupings of dual-coil ICD lead use. Based on the available data in the ICD Registry, factors were selected a priori and tested for an association with a change in trend of dual-coil ICD lead use. Factors evaluated for their potential impact on dual-coil ICD use included ejection fraction, baseline use of amiodarone, baseline use of sotalol, baseline use of dofetilide, DFT or upper limit of vulnerability testing at insertion, chronic obstructive pulmonary disease, ischemic heart disease, New York Heart Association (NYHA) classification, renal disease, primary versus secondary prevention device, device insertion by an electrophysiologist, academic versus nonacademic inserting hospital, region of the inserting hospital, and ratio of ICD insertions to lead extractions. Each characteristic was first evaluated individually. Variables significant at the p < 0.20 level were entered into a multivariable model. Only characteristics significant at a p < 0.05 level were retained in the final model. For each significant characteristic we determined the probability of belonging to each of the 3 dual-coil use groups.
To account for hospitals with missing quarterly data, all models were weighted by the hospital’s ICD volume proportional to the size of the entire population. To ensure the consistency of the results, a sensitivity analysis was performed by limiting the hospitals included in the analysis to those that had ICD data for all 23 quarters of the analysis (n = 569).
For all analyses, SAS version 9.3 (SAS Institute, Cary, North Carolina) was used. For modeling the hospital groupings, the open source Proc Traj program (CITE) was used (14). The study complied with the Declaration of Helsinki, and the project was approved by the Yale Institutional Review Board.
Among the 435,772 ICDs inserted during the study period, 72% (n = 313,333) had a dual-coil ICD lead. Within the overall study group, patients with an ICD had a mean age of 64.5 years, 71% were men, 57% had ischemic heart disease, and 78% had a primary prevention indication for the device (Table 1). A lower percentage of patients in the high-use group (66%) had their device inserted by an electrophysiologist, relative to the low-use (78%; p < 0.001) and decreasing-use (73%; p < 0.001) groups. There were statistically significant differences in the characteristics between patients in the low-, decreasing-, and high-use groups because of the large number of patients in the analysis; however, the numeric differences between the characteristics of the 3 cohorts were not clinically significant (Table 1).
Among the 1,690 hospitals included in the analysis, the majority (50%) were in the high-use group, whereas only 17% of hospitals were in the low-use group (Table 2). Hospitals in the high-use group had lower ICD insertion volumes (median 89) compared to hospitals in the low-use (median 147; p < 0.001) and decreasing-use (median 181; p < 0.001) groups. The high-use hospitals also had lower quarterly extraction volumes (median 2) relative to hospitals in the low-use (median 7; p < 0.001) or decreasing-use (median 11, p < 0.001) groups. Electrophysiologists inserted a higher proportion of ICDs in hospitals in the decreasing-use (median 80%; p < 0.001) and low-use (median 90%; p < 0.001) groups relative to the high-use (median 46%) group. There was geographic variation in the use of dual-coil ICD leads (Table 2). Academic hospitals and nonprofit hospitals were more likely to be in the low- or decreasing-use groups.
Temporal trends in use of dual-coil ICD leads
The use of dual-coil ICD leads decreased from a high of 87% in early 2010 to 55% of ICD leads by the end of 2015 (Figure 1). Similarly, the use of polytetrafluoroethylene (PTFE) dual-coil leads dropped from a high of 16% in 2010 to 3% and 0.7% by the end of 2012 and 2015, respectively (Figure 1). The use of dual-coil ICD leads has been consistently decreasing in a linear relationship over time. Over the study period, the high-use group continued to use dual-coil ICD leads in nearly 90% of their insertions with little change, whereas the decreasing-use group dropped from >90% dual-coil ICD lead at the beginning of the study to <50% dual-coil ICD leads 23 quarters later (Figure 2). Less than 30% of patients in the low-use group received a dual-coil ICD lead as of the end of 2015.
Characteristics associated with use of dual-coil ICD leads
A multivariable model identified the proportion of ICD insertions at a hospital by an electrophysiologist and the ratio of insertions to extractions as the factors associated with being in a specific group of dual-coil ICD lead use. The probability of a hospital being in the low- or decreasing-use groups increased as the proportion of ICDs performed by an electrophysiologist increased, whereas the probability of being in the high-use group decreased from 56% to 37% with an increase in the percentage of ICDs inserted by electrophysiologists from 0% to 100% (Figure 3A). As the ratio of new ICD insertions to extractions decreased (meaning the number of extractions increased), the probability of being in the low- or decreasing-use groups increased (Figure 3B). There was a nearly 50% probability that a site performing no extractions would be a high-use site, but the probability of being a high-use site decreased to approximately 10% among sites that performed 4 extractions for every 5 ICDs inserted.
There were 569 hospitals that reported an ICD insertion in each of the 23 quarters within the study period. Among those hospitals, 297,220 patients had initial ICD insertions during the study period. These patients and hospitals had similar characteristics to the overall patient group (Online Tables 1 and 2). In this smaller group of patients and hospitals, the decreasing-, low-, and high-use groups had similar temporal trends and proportions of dual-coil ICD leads as those observed in the overall population. Restricting the multivariable model to hospitals with data for all 23 quarters did not change the association between dual-coil ICD lead use and the proportion of leads inserted by electrophysiologists or the ratio of insertions to extractions.
Although data on the effect of dual-coil ICD leads on DFTs are mixed, no data have shown improvement in first shock efficacy or adjusted mortality. Despite the lack of proven benefit from dual-coil ICD leads, more than 7 in 10 ICD leads inserted over the course of this study were dual-coil leads. Although the proportion of dual-coil ICD leads has decreased over time, they still accounted for approximately 55% of inserted leads as of the end of 2015. Higher extraction volumes and a larger proportion of ICDs inserted by electrophysiologists were both associated with low or decreasing use of dual-coil ICD leads.
Dual-coil leads may result in a modest reduction in DFT (7,8). In the largest published analysis (n = 1,652), there was a statistically significant lower DFT of 1.8 J for dual-coil ICD leads compared with single-coil ICD leads (17.6 J vs 19.4 J; p < 0.001) (8), whereas other studies showed no difference in DFT between single- and dual-coil leads (3–6). Relative to older devices, the current generation of ICDs has higher maximum energy of 35 to 41 J and has lower DFTs because of biphasic waveforms, active cans, and pectoral insertion sites (15). These advances in defibrillation technology raise questions about the incremental benefit of a modest decrease in DFT from dual-coil ICD leads (16–18).
One of the major concerns about dual-coil ICD leads is the increased difficulty with lead extraction. One study of 2,176 ICD leads (82% dual coil) showed that extractions of dual-coil ICD leads were associated with longer procedure times and a higher risk of major complications, such as pericardial tamponade due to superior vena cava tear or right atrial avulsion that required emergent sternotomy (9). Another extraction study of 1,385 ICD leads (67% dual coil) found that 30-day mortality associated with dual-coil ICD leads was 2.7 times higher than that associated with single-coil ICD leads (95% confidence interval: 1.6 to 4.5; p < 0.001) (19). Dual-coil leads have also been associated with higher rates of venous occlusion relative to single-coil leads (10). The risks with dual-coil leads combined with a lack of data supporting clinical benefit highlight the need for a more definitive investigation on the outcomes associated with dual-coil ICD leads in large, multicenter studies.
Despite the decreasing use of dual-coil leads over time, dual-coil leads accounted for 55% of all ICD leads by the end of 2015. Despite data suggesting lower rates of fibrosis with PTFE leads, the rates of PTFE dual-coil ICD leads also decreased over time and represented a minimal portion of the dual-coil ICD leads (20). Recent data from LATITUDE, Boston Scientific's remote monitoring system, showed lower rates of dual-coil lead use (45% in LATITUDE vs. 64% in the ICD Registry in 2014), with a consistent yearly decrease of approximately 10% in dual-coil use from 2010 through 2014 (11). Our analysis may differ because it is a larger sample size of nearly 9 times (435,772 vs. 49,958 leads) as many leads since 2010 and nearly 19 times (77,828 vs. 4,213 leads) as many leads in the year 2014. Our analysis also includes patients from all ICD vendors, as well as patients not followed by remote monitoring. The study from LATITUDE had limited patient-level data and no site-level data on the inserting sites. In our analysis, the variation in use of dual- versus single-coil ICD leads appeared to be largely due to hospital factors.
Hospitals in the low- or decreasing-use groups were more likely higher volume, academic hospitals performing more extractions and with a larger proportion of insertions by electrophysiologists. Electrophysiologists are more likely to be involved in lead extraction than general cardiologists who insert devices, which may partially explain the association between greater electrophysiology insert volume and lower use of dual-coil leads. Similarly, hospitals that perform extractions are likely higher volume inserting hospitals with more outpatient insertions. The association between sites and dual-coil lead use was further demonstrated by the meaningful regional variation between low-, decreasing-, and high-use hospitals. Hospital-level variation in dual-coil use was also observed in MADIT-CRT (8). In our analysis, 50% of the hospitals (837 of 1,690) were in the high-use group, and these hospitals were still using dual-coil leads nearly 90% of cases as of the end of 2015.
In our study, there were no clinically significant differences in the characteristics of patients in the low-, decreasing-, or high-use groups. Patient characteristics such as black race, end-stage renal disease on dialysis, amiodarone use, and NYHA functional class IV heart failure have been reported to be associated with high DFTs (21), and it is these patients at risk for inadequate defibrillation safety margins who stood to benefit from the potentially lower DFT with a dual-coil lead. In our analysis, black race, amiodarone use, NYHA functional class IV heart failure, and end-stage renal disease were not associated with higher dual-coil lead use. Based on these findings, providers do not appear to be selecting dual-coil leads for patients with a higher likelihood of elevated DFTs. The study from LATITUDE showed an association between dual-coil lead use and higher mortality (11), but the effect of single-coil lead use on mortality needs to be examined in additional large, contemporaneous patient groups.
Several limitations of this analysis should be noted. This analysis was based on observational registry data, so the findings are subject to reporting bias. However, the ICD Registry has a rigorous quality assurance process that has performed well during data quality audits (12). Because the multivariable model was weighted, the highest-volume centers drive the results, which may make the results less generalizable to hospitals with missing quarterly data or to lower-volume hospitals, but the higher-volume hospitals have more experience with the data collection form for the ICD Registry and may be more likely to provide higher-quality data. The results may also not be generalizable to patients in geographic regions outside of the United States. The available data fields in the ICD Registry may have limited the analysis, but it is important to note that all clinically relevant factors that have been shown to be associated with DFT in a previous ICD Registry analysis were included in the current analysis (21). Furthermore, our analysis shows that hospital-level variables are more likely to affect lead selection than patient-level characteristics, and the ICD Registry has detailed hospital information.
Over the course of this study, 7 of 10 ICD leads inserted in the United States were dual-coil leads. Although the use of dual-coil ICD leads has decreased over time, dual-coil leads continue to represent the majority (55%) of ICD leads inserted in the United States. Hospital factors are associated with the use of dual-coil ICD leads, as evidenced by the fact that nearly 9 of 10 ICD leads at high-use hospitals remained dual coil by the end of the study period. Future studies should examine whether the decreasing trend in use of dual-coil ICD lead is associated with improved patient outcomes.
COMPETENCY IN MEDICAL KNOWLEDGE: Substantial variation in the use of dual-coil ICD leads exists among hospitals, whereas site-level factors and not patient-level factors were associated with rates of dual-coil ICD lead use. Despite a decline in the use of dual-coil ICD leads over time, the majority of inserted ICD leads in the United States remain dual-coil leads.
TRANSLATIONAL OUTLOOK: Prior studies have demonstrated increased risk with extraction of dual-coil ICD leads compared with single-coil leads, whereas studies have not demonstrated improved outcome measures with dual-coil ICD leads. Efforts should be made to determine whether the change in use of dual-coil ICD leads is associated with patient outcomes.
Dr. Pokorney has served as a consultant/advisory board for Medtronic and Boston Scientific; and has received research grants form Gilead and Boston Scientific. Dr. Parzynski has received salary from American College of Cardiology NCDR. Dr. Daubert has received research grants from Medtronic, Boston Scientific, and Gilead; has received honoraria from Boston Scientific, Medtronic, Gilead, and Biosense Webster; and has served as a consultant/advisory board for Boston Scientific, Medtronic, and Biosense Webster. Dr. Curtis reports ownership in Medtronic; and has received salary from American College of Cardiology NCDR. All other authors have reported they have no relationships relative to the contents of this paper to disclose. The NCDR ICD Registry is an initiative of the American College of Cardiology Foundation with partnering support from the Heart Rhythm Society. The views expressed represent those of the author(s) and do not necessarily represent the official views of the NCDR or its associated professional societies (www.ncdr.com). Angelo Auricchio, MD, PhD, served as Guest Editor for this paper.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- defibrillation threshold
- implantable cardioverter-defibrillator
- National Cardiovascular Data Registry
- New York Heart Association
- Received October 11, 2016.
- Revision received November 29, 2016.
- Accepted November 29, 2016.
- 2017 American College of Cardiology Foundation
- Bardy G.H.,
- Dolack G.L.,
- Kudenchuk P.J.,
- Poole J.E.,
- Mehra R.,
- Johnson G.
- Schulte B.,
- Sperzel J.,
- Carlsson J.,
- et al.
- Gold M.R.,
- Olsovsky M.R.,
- Pelini M.A.,
- Peters R.W.,
- Shorofsky S.R.
- Epstein L.M.,
- Love C.J.,
- Wilkoff B.L.,
- et al.
- Lickfett L.,
- Bitzen A.,
- Arepally A.,
- et al.
- Messenger J.C.,
- Ho K.K.,
- Young C.H.,
- et al.
- Epstein A.E.,
- Ellenbogen K.A.,
- Kirk K.A.,
- Kay G.N.,
- Dailey S.M.,
- Plumb V.J.
- Blatt J.A.,
- Poole J.E.,
- Johnson G.W.,
- et al.
- Hsu J.C.,
- Marcus G.M.,
- Al-Khatib S.M.,
- et al.