Repeated procedures at the generator pocket are a determinant of implantable cardioverter‐defibrillator infection

Abstract

Background

Rates of cardiac‐device infections have increased in recent years, but the current incidence and risk factors for infection in patients with implantable cardioverter‐defibrillators (ICDs) are not well known.

Hypothesis

The increasing number of ICD infections is related to accumulated pocket manipulations over time.

Methods

This single‐center, prospective study included patients that underwent ICD implantation from 2008 to 2015. The endpoint was time to infection. Multivariate analysis was performed to identify independent risk factors related to infection.

Results

The study included a total of 570 patients, of whom 419 (73.5%) underwent a first implantation. Mean age was 59 ± 14 years, and 80% were male. During a median follow‐up of 36 months (interquartile range, 18–61 months; 1887 patient‐years), infection was identified in 26 patients (4.56%), an incidence of 14.9 × 1000 patient‐years. Median time to infection was 9.7 months (interquartile range, 1.35–23.4 months), and 38.5% were late infections (beyond 12 months of follow‐up). In patients with replacement implants, the incidence was 3‐fold higher than in first implantations (27.7 vs 9.1 × 1000 patient‐years; P = 0.002). Cox regression identified 2 independent predictors of ICD infection: cumulative number of interventions at the generator pocket (hazard ratio: 1.92, 95% confidence interval: 1.42‐2.6, P < 0.001) and pocket hematoma (hazard ratio: 7.0, 95% confidence interval: 2.7‐17.9, P < 0.0001).

Conclusions

The incidence of infection in ICD patients is greater than previously reported, largely due to late infections. Each new cumulative intervention at the same generator pocket nearly doubles the risk of infection.

1. INTRODUCTION

The use of cardiac implantable electronic devices (CIED) has increased exponentially over the last decade, and the rate of CIED‐related infection seems to be rising at an even greater rate.1, 2 Such infections are a feared complication, associated with higher morbidity and mortality despite correct treatment, and increase financial costs significantly.3, 4 Although many studies have analyzed the incidence of complications in patients with CIED, most have been retrospective, have assessed pacemaker‐related infections, and focused only on the first months after implantation.5, 6, 7, 8, 9 Furthermore, the clinical characteristics of patients with implantable cardioverter‐defibrillators (ICDs) may differ from those of pacemaker patients. For example, higher incidence of ventricular dysfunction and heart failure can predispose to infection in fragile patients.6 Several studies have also shown that the more complex the device, the higher the incidence of infection.7, 8 Finally, CIED infection can be easy to diagnose and manage in some cases, but in others a high degree of vigilance is required, and it is desirable to have a standardized definition of infection and a multidisciplinary approach that can be lacking in retrospective series.9, 10, 11 For all of these reasons, the current rate of infection in the ICD population is probably underestimated and not well understood, with rates varying between 1% and 9.5% in different series, depending on the type of ICD and the duration of follow‐up.2, 4, 5, 7, 8, 12

The aim of this prospective study in patients with ICD implant was to assess the incidence of ICD‐related infections during long‐term follow‐up and identify predictors of infection.

2. METHODS

2.1. Patients

Our center has an ongoing prospective registry of all patients who have undergone ICD implantation since May 2008, when all ICD implantations were centralized in the arrhythmia unit. This registry was specifically designed to identify ICD‐related complications during follow‐up. For the study, we selected all patients with ICD implantation, whether first‐time or replacement implantations. Patients were followed from inclusion in the cohort until death, loss to follow‐up, discontinuation of ICD treatment, new replacement due to end of life of the ICD, or until December 31, 2015. (For details about peri‐implant patient management and infection management, see Supporting Information, section 1, in the online version of this article.)

2.2. Data collection and follow‐up

Patient data were collected at the time of ICD implantation, at each follow‐up visit to the ICD clinic, and during hospital admissions. Patient‐related variables included demographics, comorbidity, analytics, electrocardiographic data, medications, and indication. Procedure‐related variables were previous procedure, number of procedures on the same pocket, number of leads implanted, number of leads abandoned, and type of ICD.

All patients were examined daily until hospital discharge, usually 24 to 48 hours after implantation. Thereafter, they were evaluated as outpatients for the first month and then at least once every 6 months, or as needed in the case of ICD‐related events. Out‐of‐clinic events related to the device were gleaned from the electronic medical records that are unified throughout our healthcare system (primary, specialty, and emergency care, and all hospital admissions and discharge records). The database is reviewed periodically to maintain data quality and to identify possible missing data in follow‐up.

2.3. Endpoint and definitions

The primary endpoint of the study was to assess the time to ICD‐system infection and to determine related predictive factors. Local pocket infection was defined when local signs were evident (eg, localized cellulitis, swelling, discharge, wound dehiscence, erosion, or pain), the patient showed no systemic symptoms (eg, fever, rigors, night sweats, shock, or embolisms), and blood cultures were negative. Local cultures were used to confirm the diagnosis. Endovascular infection was present when, regardless of the state of the pocket, the patient had systemic signs and symptoms of infection supported by repeated positive blood cultures and/or vegetations in leads or valves, as identified by transesophageal echocardiography. Depending on the time of appearance, infections were classified as early if they appeared in the first month after device implantation; deferred, if they appeared between 2 and 12 months; and late, if they occurred >12 months after implantation. A secondary objective of the study was to analyze the prevalence of ICD‐implant complications and time to death.

In this study, first implantation refers to patients with no history of any device implanted before the inclusion event. ICD replacement was defined as any generator change due to battery depletion, upgrade, or system complication. Reintervention was defined as any surgical procedure required to manage a noninfectious complication of the ICD implant without generator change. The number of pocket interventions was defined as the cumulative number of surgical interventions involving the generator pocket (ie, first implantation, replacements, and reinterventions). Any blood collection in the pocket with swelling, pain, or functional impairment was registered as hematoma.

2.4. Statistical analysis

Descriptive statistics were reported for patients and procedure characteristics. Continuous variables are presented as mean ± SD for normally distributed variables, and medians with interquartile range (IQR) for non‐normally distributed variables. Categorical variables are presented as frequencies with percentages. To compare patients with and without infection, the Student t test or Wilcoxon rank‐sum test was used to compare continuous variables, and the Fisher exact test was used for categorical variables. Survival curves were determined by the Kaplan‐Meier method, with comparisons of cumulative events based on logarithmic transformation. Patients were censored at death, at their last follow‐up, or at the end of life of the ICD implanted during the study period, requiring a new replacement. To evaluate the independent effects of clinically significant risk factors, a multivariate Cox regression analysis was performed including those covariates that were significant in crude models. An analysis of interaction between covariates was performed. Hazard ratio (HR) and 95% confidence interval (CI) were reported. Incidence of infection was calculated as the rate of infection per 1000 patient‐years. Analysis was performed using SPSS Statistics version 22 (IBM Corp., Armonk, NY), with significance predefined as P < 0.05.

3. RESULTS

3.1. Patient characteristics

Between May 2008 and December 2015, 570 patients underwent ICD implantation at our institution. The mean age of the cohort was 59 ±14 years; 80% of the patients were male. The most common cardiac disease was ischemic cardiomyopathy (n = 249; 43.7%). A severely depressed left ventricular ejection fraction was present in 335 (58.8%) patients, and primary prevention was the most common indication (62%). A first implantation was registered in 419 patients (73.5%). A single‐chamber device was used in 360 patients (63.2%), a dual‐chamber device in 100 (17.5%), and resynchronization in 110 (19.3%). Clinical characteristics are detailed in Table 1.

Table 1.

Baseline characteristics and comparison of patient demographics and ICD features in patients with and without infection

	No Infection, 544 (95.43)	Infection, 26 (4.56)	P Value
Age, y	59.5 ± 14.5	57.8 ± 17.5	0.61
≤20 or ≥75 y	76 (13.97)	7 (26.92)	0.084
Male sex	434 (76.1)	23 (88.5)	0.3
Smoker	246 (45.2)	13 (50)	0.68
CPD	108 (20)	5 (19.2)	0.93
DM	182 (33.45)	11 (42.3)	0.35
HTN	286 (52.6)	14 (53.8)	0.89
Dyslipidemia	254 (46.7)	9 (34.6)	0.33
Nephropathy	83 (15.25)	2 (7.7)	0.4
Hepatopathy	24 (4.4)	2 (7.7)	0.33
Hgb, g/dL	13.4 ± 3.6	13.7 ± 1.8	0.75
Leukocytes, ×109/L	8.8 ± 2.5	7.8 ± 2.28	0.77
Cr, mg/dL	1.05 ± 0.43	1.02 ± 0.45	0.91
CrCl, mL/min/1.73 m2	87 ± 32	87.7 ± 27	0.152
ESRD	2 (0.36)	1 (3.8)	0.131
BMI >30 kg/m2	124 (22.8)	4 (15.4)	0.37
NYHA class			0.53
I	138 (25.3)	4 (15.4)
II	293 (53.9)	18 (69)
III	113 (20.8)	4 (14.4)
QRS duration >120 ms	231 (42.4)	11 (42.3)	0.98
AF/flutter	79 (14.5)	2 (7.7)	0.8
LVEF, %			0.15
50	129 (23.7)	9 (34.6)
30–50	90 (16.6)	6 (23)
<30	325 (59.7)	11 (42.4)
Cardiomyopathy			0.58
Ischemic	240 (44)	9 (34.6)
Dilated	155 (28.5)	5 (19.2)
Valvular	11 (2)	2 (7.7)
Hypertrophic	44 (8)	2 (7.7)
ARVC	13 (2.5)	1 (3.9)
Channelopathy	27 (5)	3 (11.5)
Others	54 (10)	4 (15.4)
Antithrombotic agents
None	239 (44)	10 (40)	0.91
ASA	219 (40.2)	9 (32)	0.82
Clopidogrel	9 (1.6)	1 (4)	0.94
Double antiaggregation	25 (4.6)	3 (11.5)	0.12
VKA	51 (9.4)	3 (11.5)	0.72
Heparin use	17 (3.1)	3 (11.5)	0.057
Indication
First procedure	407 (74.8)	12 (46)	0.002
Replacement
Battery change	99 (18.2)	9 (34.6)	0.07
System complications	19 (3.5)	2 (7.7)	0.24
Upgrade	19 (3.5)	3 (11.5)	0.057
Type of ICD
Single chamber	344 (63.2)	16 (61.6)	0.9
Dual chamber	95 (17.5)	5 (19.2)	0.79
Resynchronization	105 (19.3)	5 (19.2)	0.99
No. of leads (IQR)	1.0 (1.0–2.0)	1.5 (1.0–2.25)	0.38
No. of pocket interventions (IQR)	1.0 (1.0–2.0)	2.5 (1.0–4.0)	0.001
Abandoned lead	38 (6.98)	6 (23)	0.01
Mean (IQR)	1 (1–1)	1 (1–1.5)	0.043
Procedure‐related complications
DVT	5 (0.94)	1 (3.8)	0.12
Lead dislocation	15 (2.7)	1 (3.8)	0.531
Hematoma	20 (3.67)	6 (23)	0.0001
Lead problems	13 (2.38)	1 (3.8)	0.48
Pneumothorax	6 (1)	0 (0)	0.9
Need for reintervention	24 (4.41)	2 (7.69)	0.33

Abbreviations: AF, atrial fibrillation; ARVC, arrhythmogenic right ventricular cardiomyopathy; ASA, acetylsalicylic acid (aspirin); BMI, body mass index; CPD, chronic pulmonary disease; Cr, creatinine; CrCl, creatinine clearance; DM, diabetes mellitus; DVT, deep‐vein thrombosis; ESRD, end‐stage renal disease; Hgb, hemoglobin; HTN, hypertension; ICD, implantable cardioverter‐defibrillator; IQR, interquartile range; LVEF, left ventricular ejection fraction; NYHA, New York Heart Association; SD, standard deviation; VKA, vitamin K antagonist.

Data are reported as n (%) or mean ± SD, unless otherwise noted.

3.2. Follow‐up and noninfectious complications

The median follow‐up was 36 months (IQR, 18–61 months), with a total of 1887 patient‐years. Fourteen patients (2.3%) were lost during follow‐up, mainly due to change of residence. As the median of follow‐up in these patients was 32 months (IQR, 16.6–50.0 months) months, all of them were included in the analysis. Death occurred in 108 (18.9%) patients, mostly due to exacerbation of heart failure (n = 53; 49%), and 11 (1.9%) patients received cardiac transplantation. Noninfectious device‐related events were registered in 62 patients (Table 1). Pocket hematoma was the most common problem (26 patients; 4.56%) and was related to the type of antiaggregation therapy, with a maximum incidence in patients with heparin bridge (18% vs 5.3%, P = 0.005; odds ratio: 7.4, 95% CI: 2.2‐24.6; see Supporting Information, Figure 1, in the online version of this article).

Figure 1.

(A) Time to infection after first ICD implantation (n = 419) and after replacement (n = 151). (B) Time to death, by presence or absence of device‐related infection. Abbreviations: ICD, implantable cardioverter‐defibrillator.

Surgical reintervention for system revision was reported in 26 patients (4.56%). There were 27 events: 15 lead dislodgements, 3 large hematomas, and 9 lead‐related failures.

3.3. Infections

Infection was identified during follow‐up in 26 (4.56%) patients, resulting in an incidence of 14.9 (95% CI: 9.4‐20.4) per 1000 patient‐years (see Supporting Information, Table, in the online version of this article). Clinical presentation was a local infection in 20 patients (pocket abscess in 12 and erosion in 8) and a systemic infection in 6 patients. The median time to infection was 9.7 months (IQR, 1.35–23.4 months); 23% were early infections, 38.5% deferred, and 38.5% late, of which 15.4% occurred >2 years after implantation.

In patients with a first implantation, the incidence of infection was 9.11 (95% CI: 4.1‐14.21) per 1000 patient‐years vs 27.76 (95% CI: 13.23‐42.3) per 1000 patient‐years in patients with replacements (P = 0.002; Figure 1A). Patients with infection were more likely to have had a replacement procedure (n = 14 [54%]; P = 0.002), abandoned leads (n = 6 [23%]; P = 0.009), or a history of pocket hematoma (n = 5 [19.2%]; P = 0.0001). No other clinical characteristics differed between patients with and without infection (Table 1).

The microbiological spectrum was varied, with coagulase‐negative staphylococci being the most frequent agent (n = 17; 65.3%). Staphylococcus aureus was present in 1 patient with acute infection and in another with very late infection (77 months). In 1 patient, the causative microorganism was not identified. In all patients with these infections, the whole system was removed. Following extraction, there was 1 death due to septicemia and another due to massive ictus. In the remaining cases, 3 patients received no new device and 21 patients received a new system, implanted at a different location. Two patients had a new infection within the first year after reimplantation. Although there was a trend toward increased mortality within the first year after implantation in patients with infection (P = 0.08), the long‐term mortality risk was similar for all patients (P = 0.75; Figure 1B).

3.4. Predictors of device‐related infections

Clinical parameters and implant data were assessed by univariate and multivariate analysis to determine their predictive value for device‐related infections, as reported in Table 2. In the univariate analysis, first implantation vs replacement (74.8% vs 46%; P = 0.002), number of pocket interventions (2.5 [IQR, 1.0–4.0] vs 1.0 [IQR, 1.0–2.0]; P < 0.001), abandoned leads (24% vs 7%; P = 0.009), and pocket hematoma (24% vs 3.6%; P < 0.0001) were associated with infection. In the multivariate model, the number of pocket interventions was included as a variable. As most patients with ≥1 intervention received replacements, the variable “first implantation vs replacement” was excluded from the multivariate analysis. Consequently, 2 independent factors in device‐related infection remained significant: number of pocket interventions (HR: 1.92, 95% CI: 1.42‐2.6, P < 0.0001) and pocket hematoma (HR 7.0, 95% CI: 2.7‐17.9, P < 0.0001; Figure 2). To determine the existence of interactions, we analyzed all the covariates obtained from the multivariate analysis with all possible interactions. No significant interactions were observed between abandoned lead and pocket apertures (P = 0.093), abandoned lead and hematoma (P = 0.555), or pocket apertures and hematoma (P = 0.933).

Table 2.

Multivariate analysis of factors predicting ICD infection

	Univariate Analysis, HR (95% CI)	P Value	Multivariate Analysis, HR (95% CI)	P Value
Abandoned lead	4.2 (1.6‐11.1)	0.009	1.69 (0.56‐5.08)	0.35
Pocket apertures	2.1 (1.63‐2.71)	<0.0001	1.92 (1.42‐2.6)	<0.0001
Hematoma	8.3 (2.98‐23)	0.0001	7.0 (2.7‐17.9)	<0.0001

Abbreviations: CI, confidence interval; HR, hazard ratio; ICD, implantable cardioverter‐defibrillator.

Figure 2.

Time to infection, by independent factors associated with device‐related infection. (A) number of cumulative pocket apertures, (B) pocket hematoma.

4. DISCUSSION

The present study had 3 main findings: (1) the incidence of ICD‐related infections was 14.9 cases per 1000 patient‐years of follow‐up, with a notable percentage of late infections; (2) pocket hematoma is the main complication after ICD implantation, its appearance depends on the anticoagulation regimen used, and it is directly related to ICD infection; and (3) at the time of a new intervention, the cumulative number of pocket interventions involving the same pocket is the main determinant of future ICD‐related infections, nearly doubling the risk of an event.

The expansion of primary prevention of sudden cardiac death and cardiac resynchronization therapy has significantly increased the rate of complex device implantation in an aging population, with higher comorbidity and cumulative procedures that increase the potential for infection.2 The incidence of 14.9 infections × 1000 patient‐years (4.56% in a median of 36 months of follow‐up) in the present study is consistent with recent prospective reports including only ICD patients.5, 8, 13 Previously reported rates of 1% to 1.3% of global CIED infection or incidences of 8.9 per 1000 device‐years in the ICD population may be underestimated for many reasons.5, 6, 14 Most studies focus on pacemakers, with ICD/cardiac resynchronization therapy populations underrepresented in patient cohorts implanted before the recent expansion of indications, and many are retrospective studies or have a short‐term follow‐up.14, 15

In the largest ICD‐population study published to date, Prutkin et al reported an incidence of 1.7% after a 6‐month follow‐up period.5 Romeyer‐Bouchard et al13 reported a 4.3% prevalence of infection at 2.5 years in 303 consecutive patients with resynchronization devices, with a 1.7% per annum incidence. Recently, Unsworth et al8 reported a prevalence of 1.2% in the first year and 4.8% beyond 1 year of follow‐up in a de novo ICD implantation population. The differences in related infection rates in these more homogeneous ICD populations may also be related to length of follow‐up and the inclusion of patients with or without previous procedures. Late infections (38.5% in our population) significantly influence the rate of documented cases, revealing the importance of a long‐term follow‐up when assessing infection rates in patients with ICD.4, 8, 9 The increased incidence of infection in patients with previous procedures (3× that of de novo implantations) merits in‐depth discussion, as it reflects the real world of device implantation (eg, patients with increased survival and a history of many interventions over the same generator pocket).

4.1. Predictors of ICD‐related infection

Although in some cases infection of the CIED system may occur by hematogenous spread, current evidence ascribes it mainly to bacterial contamination at the time of intervention.9 This contamination usually occurs as a silent colonization, but disequilibrium between the patient's immune system and the virulence of the pathogen may trigger clinical manifestations of device infection at any time in the post‐implantation follow‐up. Several studies have identified host‐related factors associated with an increased risk of CIED infection.6 Most of these confer a relative state of immunosuppression that increases vulnerability. Because most of these factors are unavoidable, it is crucial to use a meticulous technique during the implantation and to provide a careful follow‐up, as many common injuries to the pocket can increase the percentage of infections. Many studies have shown that the presence of hematoma after surgery is an independent factor associated with infection and depends on the type of anticoagulant used; most recently, Essebag et al demonstrated a >7‐fold increased risk of infections.16 Heparin bridge is clearly associated with pocket hematoma in our series (18.75% vs 5.34% in patients without this factor; P = 0.0001), as previously reported by others (such as the Bridge or Continue Coumadin for Device Surgery Randomized Controlled Trial [BRUISECONTROL], 16% vs 3.5%, P < 0.001).17 Antiaggregation therapy also has been associated with pocket hematoma,6 but this was not the case in our work, mainly due to the small number of patients included and because it was withdrawn when possible.

The presence of previous procedures has been linked with infection.6, 11 This has been evaluated in different ways depending on the study and the procedures included (eg, replacements for battery depletion, reinterventions, and upgrades). In our study, the primary determinant of infection was the cumulative number of pocket manipulations. We included patients who had a procedure to change the generator for battery depletion and also patients with reinterventions for complications commonly seen in implants (eg, lead dislodgment, hematoma evacuation, upgrades). In a historical registry of patients with pacemakers, Johansen et al reported an association between number of prior procedures and CIED infection similar to that observed in the present research.18 Our study confirms and expands these data in a current ICD‐only population, reinforcing the concept in this population at greater risk. Cumulative surgeries and other pocket stressors such as hematoma increase the risk of infection, probably due to both the likelihood of new bacterial colonization by each intervention and a reactivation of a previous latent focus.19, 20 This excess risk should be taken into account when approaching interventions in complex patients with fragility, problems with previous systems, local complications, upgrades, and other issues that are increasingly observed in routine practice. These interventions are not trivial and should be reduced as much as possible, such as by increasing ICD battery life as much as possible through proper programming of the device by the physician and technological improvements by the industry, and also by individualizing the risks/benefits of each intervention during the follow‐up.21, 22, 23

Long‐term mortality in large populations of patients with CIED infection seems to be high even after successful treatment of infection.3 In our cohort, mortality was similar in patients with and without infection. This finding could be explained by (1) the low incidence of infection caused by aggressive bacteria such as S aureus and (2) a standardized protocol with an early removal of the complete system.

4.2. Study limitations

This was a single‐center study. Multicenter studies are required for external validation of the results obtained. Patients in our cohort came from a single tertiary‐care center and therefore may not be representative of patients presenting to other centers. The population of patients with device infections is relatively small due to the prospective design of the study and the small percentage of infections. A lack of statistical power may have influenced the analysis of factors previously validated in other series. Finally, local and systemic infections were considered together for the analysis. It remains unknown whether these results would be reproduced in larger studies performing differential analysis of local and systemic device‐related infections.

5. CONCLUSION

The incidence of infection in patients with an ICD was greater than previously reported, mainly due to late infections. Accumulated stress in the generator pocket caused by local complications such as hematoma or by repeated interventions is a determinant of future infection, with each pocket manipulation doubling the HR. These data should be considered when evaluating reinterventions during ICD patient follow‐up.

Conflicts of interest

The authors declare no potential conflicts of interest.

Supporting information

Appendix S1.

Arana‐Rueda E, Pedrote A, Frutos‐López M, et al. Repeated procedures at the generator pocket are a determinant of implantable cardioverter‐defibrillator infection. Clin Cardiol. 2017;40:892–898. 10.1002/clc.22743