Author + information
- Received August 29, 2016
- Revision received February 14, 2017
- Accepted February 24, 2017
- Published online October 16, 2017.
- Charles A. Henrikson, MDa,∗ (, )
- M. Rizwan Sohail, MDb,
- Helbert Acosta, MDc,
- Eric E. Johnson, MDd,
- Lawrence Rosenthal, MDe,
- Roman Pachulski, MDf,
- Dan Dan, MDg,
- Walter Paladino, MDh,
- Farhat S. Khairallah, MDi,
- Kent Gleed, MDj,
- Ibrahim Hanna, MDk,
- Alan Cheng, MDl,
- Daniel R. Lexcen, PhDl and
- Grant R. Simons, MDm
- aKnight Cardiovascular Institute, Oregon Health & Science University, Portland, Oregon
- bDivisions of Infectious Diseases and Cardiovascular Diseases, Mayo Clinic College of Medicine, Rochester, Minnesota
- cTrinity Medical Center, Rock Island, Illinois
- dStern Cardiovascular Foundation, Germantown, Tennessee
- eDivision of Cardiovascular Medicine, UMass Memorial Medical Center, Worcester, Massachusetts
- fSouth Texas Heartbeat, San Antonio, Texas
- gPiedmont Medical Center, Atlanta, Georgia
- hCatholic Health Partners, Youngstown, Ohio
- iTallahassee Memorial Hospital, Tallahassee, Florida
- jAlegent Health, Omaha, Nebraska
- kPrinceton Baptist Medical Center, Birmingham, Alabama
- lMedtronic plc, Minneapolis, Minnesota
- mDivision of Electrophysiology, Englewood Hospital and Medical Center, Englewood, New Jersey
- ↵∗Address for correspondence:
Dr. Charles A. Henrikson, Electrophysiology Service, Knight Cardiovascular Institute, Oregon Health & Science University, 3181 Southwest Sam Jackson Park Road, UHN 62, Portland, Oregon 97201.
Objectives This study sought to determine whether the nonabsorbable TYRX Antibacterial Envelope (TYRX) reduces major cardiovascular implantable electronic device (CIED) infections 12 months after implant.
Background TYRX is a monofilament polypropylene mesh impregnated with minocycline and rifampin specifically designed to hold a CIED in place and elute antimicrobials over time. There are limited data on its ability to reduce CIED infections.
Methods We prospectively enrolled patients who underwent generator replacement with an implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy device (CRT), treated with TYRX. The primary endpoints were major CIED infection and CIED mechanical complications. Given the differences in infection rates among ICD and CRT patients, 3 different control populations were used: a published benchmark rate for ICD patients, and both site-matched and comorbidity-matched controls groups for CRT patients.
Results Overall, a major CIED infection occurred in 5 of 1,129 patients treated with TYRX (0.4%; 95% confidence interval: 0.0% to 0.9%), significantly lower than the 12-month benchmark rate of 2.2% (p = 0.0023). Among the TYRX-treated CRT cohort, the major CIED infection rate was 0.7% compared with an infection rate of 1.0% and 1.3% (p = 0.38 and p = 0.02) in site-matched and comorbidity-matched control groups, respectively. Among the ICD group, the 12-month infection rate was 0.2% compared with the published benchmark of 2.2% (p = 0.0052). The most common CIED mechanical complication in study patients was pocket hematoma, which occurred in 18 of the 1,129 patients (1.6%; 95% confidence interval: 0.8 to 2.5), which is comparable with a published rate of 1.6%.
- antibacterial envelope
- antibiotic prophylaxis
- implantable cardioverter-defibrillator
Infection is a major complication of cardiovascular implantable electronic device (CIED) therapy and is associated with substantial morbidity, mortality, and expense (1–3). Although systemic antibiotics help reduce CIED infections (4,5), the incidence of CIED infections is increasing (6–9). Hence, novel adjunctive measures to prevent infection could improve CIED therapy outcomes.
The TYRX antibacterial envelope (TYRX), impregnated with minocycline and rifampin, was cleared by the Food and Drug Administration in 2008 for stabilization of CIED implants. The first-generation TYRX is made of nonabsorbable polypropylene mesh and elutes over time. Although an absorbable version was recently cleared for use, all patients in this study were treated with the first-generation, nonabsorbable product. Several nonrandomized retrospective studies have demonstrated that TYRX use is associated with a 60% to 100% relative risk reduction for CIED infection (10–12). However, TYRX envelope performance in a large, prospective CIED population has not been previously reported.
We report on the findings of 2 prospective registry studies conducted to evaluate the effectiveness of TYRX in reducing CIED infections in high-risk patients undergoing implantable cardioverter-defibrillator (ICD) (Citadel Study) and cardiac resynchronization therapy (CRT) (Centurion Study) implantations.
Citadel and Centurion were separate multicenter, prospective, cohort studies that enrolled patients undergoing CIED replacement or upgrade with an ICD (Citadel) or CRT (Centurion) with TYRX. After the interim analysis in 2012, the studies were combined and the analysis plan was updated to analyze patients with TYRX in a single analysis.
The CIED procedure was performed according to usual standards of care including administration of pre-procedural intravenous antibiotics.
Patient follow-up visits were scheduled for 1 to 8 weeks, 3 months, 6 months, and 12 months after implantation. At each visit, wound sites were inspected and patients were assessed for presence of CIED infection, mechanical complication, and other adverse events. The trial protocol was approved by the institutional review board at each participating center, or a central institutional review board registered with the Food and Drug Administration and the Office of Human Rights Protections (Goodwyn Institutional Review Board, Cincinnati, Ohio). All patients provided written, informed consent.
Eligible patients included those >18 years of age undergoing CIED replacement with an ICD (Citadel) or CRT (Centurion), were >18 years with a TYRX (Figure 1), and could follow-up in person. Use of an antibiotic eluting flat sheet (TYRX ST Antibacterial Soft Tissue Device, TYRX, Inc., Monmouth Junction, New Jersey) was permitted in lieu of the TYRX Envelope. The antibiotic eluting flat sheet was used in <5% of patients.
Patients were excluded from enrollment if they had a contraindication to receiving TYRX, a current CIED infection, a planned lead extraction, a clinical diagnosis of an active infection at the time of the CIED procedure, were pregnant, could not provide appropriate signed informed consent, had a life expectancy of <6 months, or were expected to undergo heart transplantation within 6 months.
Both studies were nonrandomized registries; hence, no formal control group was used. In an attempt to place the findings in proper context, previously published or site-matched and comorbidity-matched controls were used. Additionally, because complication rates differ between ICD and CRT implants, different control groups were selected to compare the relative event rates for the ICD and CRT cohorts. For the ICD cohort, a 12-month published rate of 2.2% was used (13). For the CRT cohort, 2 control groups were used: a retrospective site-matched control group and retrospective comorbidity-matched control group–derived Medicare claims data. The Online Appendix provides details on the selection of control groups.
The 2 primary endpoints of the study were major CIED infection and CIED mechanical complication (14,15). Major CIED infection was defined according to the Centers for Disease Control and Prevention’s surgical site infection criteria (16) as a device infection involving any part of the anatomy other than the incision or subcutaneous tissue, which was opened or manipulated during the CIED implantation operation, or as endocarditis defined according to the modified Duke Criteria (17) (Online Table S1).
CIED mechanical complications included generator or lead malfunction requiring pocket revision, generator or lead dislodgement or migration, lead fracture, skin erosion, wound dehiscence, and generator pocket hematoma (Online Table S1).
All primary clinical endpoints and deaths were adjudicated by an independent Clinical Events Committee consisting of 2 CIED implanting physicians and an infectious disease physician (Online Table S2). The Clinical Events Committee reviewed all adverse events and primary outcome events. Strict definitions for CIED infection were used to minimize bias. Two Clinical Events Committee members adjudicated endpoint events. In the event of disagreement, a third member provided final ruling.
A pre-specified interim analysis of events occurring during the first 3 months after implantation was completed for the first 1,000 subjects enrolled in both studies. The analysis compared the incidence of major CIED infection with a success criteria set as follows: 1,000 subjects from Citadel and Centurion have an infection rate with a statistically significant reduction of the upper bound of a 1-sided 95% confidence interval as compared with the Gould et al. (13) 3-month benchmark rate of 1.88%. The alpha was set to 0.05.
The primary efficacy endpoint, major CIED infection rate at 12 months, was estimated using a 4-sided Clopper-Pearson 95% confidence interval for each treatment group (18). A 2–sample, 1-sided Fisher exact test was used to test the primary efficacy endpoint versus the benchmark rate and to compare the rates of major CIED infection in the CRT cohort and the site-matched retrospective controls. A site-adjusted comparison of the CRT cohort and the site-matched retrospective controls was also conducted using logistic regression, with random site effects. All other statistical testing between cases and controls were performed using a Fisher exact test, chi-square test, or Student t test. Testing versus the comorbidity-matched rates for both infections and mortality used 1-sample tests. The comorbidity matched rates are based on all Medicare data and not a sample; hence, values are considered fixed and not estimates. The tests of the infection rates against the comorbidity-matched infection rates were 1-sided, exact binomial tests (i.e., Clopper-Pearson tests). The mortality tests are based on the Kaplan-Meier estimated mortality rates and their SE at the given timepoint. Two-sample, 2-sided Student t tests were used for the case versus control tests and 1-sample, 2-sided Student t tests for the comparison with the fixed comorbidity-matched rate.
Medicare database comorbidity-matched controls
To match the entire population of Medicare controls with CRT patients, an inverse-probability-weighted estimator calculation was used to estimate treatment effects as evidenced by their inclusion (19,20). A total of 15 covariates were used to balance the inverse-probability-weighted analysis and the model was estimated using a Probit specification to weight the individual variance.
ICD cases and combined cohorts
Baseline characteristics were compared using a Student t test for continuous variables or a Fisher’s exact/chi-square test for categorical variables. A Fisher’s exact test was used for sparse data (i.e., np <5 or nq <5). CIED infection and mechanical complication data among ICD and CRT subjects were compared using the chi-square or Fisher’s exact tests. A Fisher’s exact test was used for sparse data.
The Data Management Committee informed the steering committee that the 2 primary objectives of the studies had been met at the first prespecified interim analysis. At that time, there was 1 major infection out of 1,000 subjects at 3 months compared with a 3-month benchmark rate of 1.88% (p < 0.001). The Data Management Committee recommended the enrolled cohorts be combined and followed for 12 months, the results of which are presented herein.
A total of 1,262 prospective patients were enrolled in both studies from 55 U.S. academic, community, and Department of Veteran’s Affairs Medical Centers between December 30, 2009, and May 7, 2013 (Online Table S3). Among them, 1,129 were included in the analysis (Figure 2), with 459 patients in the ICD group (40.6%) and 670 in the CRT group (59.3%). Of the 670 subjects enrolled in the CRT group, 578 were analyzed because retrospective site-matched controls were not available for 92 of the cases.
Combined study (CRT and ICD)
Patient demographics and primary endpoints
Patient demographics for the combined cohorts are shown in Table 1. CRT patients had a significantly higher prevalence of patient and procedure characteristics previously reported to be associated with increased risk for CIED infections (13,21–23).
The primary endpoint for major infections occurred in 5 of 1,129 patients at 12-months follow-up (0.4%; 95% confidence interval: 0.02% to 0.95%) (Table 2) as compared with the pre-specified benchmark rate of 2.2% (p < 0.003) (Figure 3).
There were minor infections in 12 of the 1,129 patients (1.1%) (Table 2). Although these were defined as infections confined to the skin or subcutaneous tissue, the CIED was explanted in 1 case (Table 3). This explantation was a decision by the local clinical team. The remaining 11 patients with minor infections were managed with antimicrobial therapy alone and none relapsed during the follow-up period.
After 12 months of follow-up, there were 50 CIED mechanical complications (4.4%) (Table 2). Generator pocket hematoma was the most common mechanical complication, occurring in 1.6% of cases.
The baseline patient characteristics of the 578 site-matched prospective and retrospective CRT subjects are outlined in Table 4. Within the CRT case cohort, significantly more patients had diabetes, and were on oral anticoagulation, taking aspirin, or on antiplatelet therapy than the patients found within the CRT site-matched controls. Additionally, there were significant differences in New York Heart Association functional class between the 2 cohorts. The greatest proportion of patients enrolled were indicated for a device replacement due to battery depletion or device upgrade (Table 5). The proportion of patients in these 2 categories differed significantly between the CRT case and site-matched cohorts. The number of leads implanted during the procedure differed significantly (p = 0.002); however, the total number of indwelling leads at the end of the procedure was not different between the cohorts (p = 0.52).
The inverse-probability-weighted balancing results for the comorbidity-matched controls demonstrated balance in all 15 covariates (Online Table S4). Consistent with this finding, a Probit estimation of the treatment on the weighted covariates was unable to predict which patients received TYRX, whereas a Probit estimation of the treatment on the unweighted covariates had a McFadden R-Square of 0.26, a measure of explanatory power that is nontrivial for patient-level data.
A total of 4 CIED infections (2 pocket, 1 endocarditis, 1 bacteremia and endocarditis) were observed in the CRT case cohort treated with TYRX, compared with 6 CIED infections in the site-matched control group (Table 6). Within the CRT case subjects, the mean time from generator replacement to infection diagnosis was 185 ± 98 days. Before 90 days, no pocket infections were observed among the CRT case group, compared with 6 pocket infections among the site-matched controls. Among both case and control groups, staphylococcal species of bacteria were the most common cause of infection (Table 6).
The primary study endpoint, major CIED infection rate at 12 months, was 0.7% among CRT case subjects implanted with the TYRX, 1.0% in the site-matched control group (p = 0.38), and 1.3% in the comorbidity-matched control group (p = 0.02) (Table 6). The site-adjusted analysis of the case and control subjects resulted in a p value of 0.53. The observed 80% relative reduction in infection rate for case subjects at 90 days was significant when compared with the comorbidity-matched controls (p = 0.02), but not significant when compared with the site-matched controls (p = 0.06). After 90 days, the infection rates between the cohorts were not significantly different. At 12 months after implantation, the all-cause mortality rate was 11.0% in the CRT case cohort, 4.3% in the site-matched control group, and 13.4% in the comorbidity-matched control group. The 12-month mortality rate in the CRT case subjects was significantly higher than the site-matched control (p < 0.001), and nonsignificantly lower than the comorbidity-matched control cohort (p = 0.089) (Table 4). The mortality rates at 90 days were not significantly different. Deaths in the CRT case cohort were consistent with the causes of death expected in an older cohort of patients receiving a CRT and none were adjudicated as being related to the TYRX envelope.
This analysis demonstrates that the use of TYRX for CIED replacement procedures was associated with a significantly low rate of major CIED infections relative to published controls without an increase of device-related mechanical complications. These findings are consistent with 2 earlier retrospective studies using case- and propensity-matched controls (24).
In an earlier study of 624 CIED implantations with TYRX performed at 10 U.S. centers, Bloom et al. (10) demonstrated a low CIED infection rate (0.48%) after an average follow-up of 1.9 ± 2.4 months. A subsequent investigation by Kolek et al. (1) compared CIED infection rates in 260 patients who had ≥2 pre-specified risk factors for CIED infection and were implanted with TYRX to 639 case-matched retrospective control patients with the same number of pre-specified risk factors who were implanted without TYRX. The CIED infection rate was significantly lower in the TYRX group (0.4% vs. 3.0%; odds ratio [OR]: 0.13; 95% confidence interval: 0.02 to 0.95; p = 0.04). Similarly, Mittal et al. (12) compared CIED infection rates in 275 propensity-matched pairs who underwent CIED implantation, with or without TYRX. The CIED infection rate was significantly lower in the TYRX group (1.1% vs. 3.7%; p < 0.05). Finally, a retrospective analysis by Shariff et al. (25) reported no infections in a cohort of CIED patients with TYRX implanted relative to a non-TYRX cohort infection rate of 1.7% after 6 months of follow-up. The variability in the infection rates across these studies is likely due to different cohorts being studied and variable definitions being used for what constitutes and infection. Nevertheless, all of them demonstrate a significant reduction in CIED infections with TYRX use when compared with usual standard of care. Our study results are consistent with these observations.
In the present study, 4.4% of patients experienced a post-implantation mechanical complication. The most common was generator pocket hematoma, occurring in 1.6% of patients, followed by lead dislodgement and migration (1.2%). These findings are similar to prior publications (26,27) and do not suggest and increased risk of these events associated with the TYRX Envelope. There were a total of 96 deaths which is similar to prior reports (28); none of the deaths were TYRX related.
Among the CRT cohort, we observed a low rate (0.7%) of major infection in those who received the TYRX envelope that was statistically significantly lower than a comorbidity-matched control group, but not different than a site-matched control group. This latter discrepancy may be due to an unexpectedly lower rate of infection among the site-matched control group who also appeared to have fewer comorbidities and risk factors, including diabetes, heart failure, and use of anticoagulant or antiplatelet agents. Additionally, all reported infections in the control group occurred within 3 months and no late infections were reported. This is highly unusual and strongly suggests under-reporting of late events or limitations in collecting events during the chart review process (type II error). Nevertheless, the infection rate compares favorably with prospective and retrospective published CIED infection rates in the absence of the TYRX Envelope, which vary from 1.88% to 2.30% over 2.7 to 12.0 months of follow-up (Table 7) (13,21,22,26,27,29).
A higher 12-month event rate of 1.3% was observed among a large Medicare claims-based database of >21,000 patients who underwent CRT replacement procedures, which we consider to be a more reliable control dataset due its size and because it was matched by comorbidity. It is conceivable that a claims-based methodology also underestimated the actual major infection rate in this population, due to coding errors, failure to bill, failure to treat invasively, or patient death before system extraction, but the magnitude of these errors was probably small. Because this comorbidity-matched control event rate is derived from such a large database, and the matching criteria were more detailed, we consider it a robust result for comparison with the CRT case population, which was observed to have a statistically significant 46% relative reduction in the rate of major infection at 12 months. Of note, this claims-based control event rate compares favorably with prior published studies, and it may represent an important modern benchmark for future analyses of infection rates among CRT patients who undergo CIED replacement procedures.
The CRT case population was noted to have a significantly higher 12-month mortality than the CRT site-matched control population (49.5% relative difference; p < 0.001), but a nonsignificantly lower mortality rate than the Medicare control group (34% relative reduction). None of the deaths were related to the TYRX envelope and there were no differences in mortality within the first 90 days after implantation. Importantly, the CRT case mortality rate (11.0%) and Medicare comorbidity-matched control mortality rate (13.4%) were similar to a previously reported rate (9.9%) from a cohort of ICD/CRT replacements in the National Cardiovascular Data Registry (30). Similar to the infection results, the differences in comorbidities between the cases and site-matched controls may be responsible for the late difference in mortality and, thus, the Medicare control population may represent a more robust comparison.
The Citadel study (ICD Cohort) found a 12-month infection rate of only 0.2% (1 in 459), as compared with a 12-month published control infection rate of 2.2% (13). This is a highly significant difference; there was no difference in the rates of mechanical complication between the groups.
This nonrandomized study used prospectively collected data from patients treated with TYRX and compared with historical controls, which may introduce recall and selection bias. Additionally, the use of periprocedural antibiotics was not protocolized. We attempted to reduce these biases for the comparison with the Citadel/Centurion patient cohort by selecting a historical published control cohort, site-matched control cohorts, and comorbidity-matched controls that received similar device types with the same infection risk profiles. Nevertheless, the retrospective methodology of the site-matched control group seems to have been fraught with difficulty. We suspect that, at each site, the control patient selection was biased toward patients who were followed up at the study center, and presented to the device clinic reliably, thereby rendering the site-matched group a poor comparator. The Medicare claims database seems to be a more appropriate comparator; however, this has the billing and coding limitations of any administrative database. It is likely that the identification of infections within control groups underestimated the true infection rates in these populations; however, as stated, the claims-based database was probably less susceptible to this flaw. The WRAP-IT (World-wide Randomized Antibiotic Envelope Infection Prevention Trial; NCT02277990), a prospective, randomized, controlled study is currently underway to evaluate infection reduction and mortality associated with the use of TYRX to definitively address these issues.
The use of TYRX was associated with a low rate of major CIED infections in high-risk patients undergoing ICD/CRT replacement procedures and resulted in a significant reduction in major CIED infections compared with a published cohort without TYRX at 12 months.
COMPETENCY IN MEDICAL KNOWLEDGE: CIED infection is a feared complication of device implantation. Proper skin preparation, meticulous sterile technique, and pre-procedural intravenous antibiotics have been shown to be effective, but additional methods are needed to further decrease the incidence of device infection.
TRANSLATIONAL OUTLOOK: These studies provide evidence that the use of an antibiotic envelope reduces the incidence of device infection. Further study in a randomized trial evaluating infection reduction with a TYRX absorbable antibiotic envelope is underway to confirm these findings.
Dr. Henrikson was the local principal investigator for contracted research for the Citadel/Centurion Studies from TYRX, Inc. Dr. Sohail has received funding from TYRX, Inc. for prior research (Bloom et al. PACE 2011 34, 133-42); and honoraria/consulting fees from Medtronic Inc., Spectranetics, and Boston Scientific. Drs. Pachulski, Dan, Paladino, Gleed, Hanna, and Simon were the local principal investigators for contracted research for the Citadel/Centurion Studies; and they received support from TYRX, Inc. Dr. Cheng is employed by Medtronic, Inc.; and holds equity ownership in the company. Dr. Lexcen is employed by Medtronic, Inc.
The Citadel/Centurion Studies were sponsored by TYRX, Inc. Employees of the sponsor participated in the study design and study process, but were blinded to the adjudicated primary endpoint data from the Clinical Events Committee and statistical analyses until the results were reported by the Interim Data Monitoring Committee. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
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
- cardiovascular implantable electronic device
- cardiac resynchronization therapy
- cardiac resynchronization therapy with DEFIBRILLATOR
- cardiac resynchronization therapy with pacing function only
- implantable cardioverter-defibrillator
- TYRX antibacterial envelope
- Received August 29, 2016.
- Revision received February 14, 2017.
- Accepted February 24, 2017.
- 2017 American College of Cardiology Foundation
- Da Costa A.,
- Kirkorian G.,
- Cucherat M.,
- et al.
- de Oliveira J.C.,
- Martinelli M.,
- Nishioka S.A.,
- et al.
- Voigt A.,
- Shalaby A.,
- Saba S.
- Greenspon A.J.,
- Patel J.D.,
- Lau E.,
- et al.
- Clopper C.,
- Pearson E.S.
- Guo S.,
- Fraser M.W.
- Wooldridge J.M.
- Lekkerkerker J.C.,
- van Nieuwkoop C.,
- Trines S.A.,
- et al.
- Shariff N.,
- Eby E.,
- Adelstein E.,
- et al.
- Krahn A.D.,
- Lee D.S.,
- Birnie D.,
- et al.
- Kramer D.B.,
- Kennedy K.F.,
- Noseworthy P.A.,
- et al.
- Prutkin J.M.,
- Reynolds M.R.,
- Bao H.,
- et al.