OUP user menu

Prevalence of bacterial colonization of generator pockets in implantable cardioverter defibrillator patients without signs of infection undergoing generator replacement or lead revision

Thomas Kleemann, Torsten Becker, Margit Strauss, Ngoc Dyck, Udo Weisse, Werner Saggau, Ulrich Burkhardt, Karlheinz Seidl
DOI: http://dx.doi.org/10.1093/europace/eup334 58-63 First published online: 27 October 2009

Abstract

Aims This study was designed to evaluate the prevalence of bacterial colonization of generator pockets in implantable cardioverter defibrillator (ICD) patients without signs of infection and to analyse the impact of bacterial colonization on the incidence of device infection during follow-up.

Methods and results In 122 ICD patients undergoing generator replacement or surgical lead revision between January 2006 and July 2008, microbiological cultures of generator pockets and extracted leads were consecutively obtained. Patients with clinical evidence of a device infection were excluded. Positive cultures from the generator pocket and leads were found in 40 (33%) patients. The most common bacteria isolated were coagulase negative staphylococci (68%). During a median follow-up time of 203 days after the revision device infection occurred in three [7.5%, confidence interval (CI) 1.6–20.4%] patients with a positive culture vs. two (2.4%, CI 0.3–8.5%) patients with a negative culture (P = 0.33). Time from revision to infection was 108 ± 73 days in patients with positive culture vs. 60 ± 39 days in patients with negative culture (P = 0.50).

Conclusion A third of ICD patients undergoing generator replacement or lead revision have an asymptomatic bacterial colonization of generator pockets. After revision 7.5% of these patients develop a device infection with the same species of microorganism.

  • Device infection
  • ICD
  • Swab culture

Introduction

The rate of device implantations is increasing with ageing of the general population and the formulation of new indications.1,2 Although conferring obvious benefits, the use of these implantable devices is associated with serious complications, including infections. Reported incidence rates of cardiac device infections range from 0.13% to 19.9% and 0.0% to 3.2% for pacemakers and implantable cardioverter defibrillators (ICD), respectively.38 Device replacement or repeat interventions are known to be associated with an increased rate of device infection.9 Harcombe et al.10 observed a five-fold increased complication rate in patients undergoing elective generator replacement compared with patients receiving first implantation of a permanent pacemaker. The most common pathogens of pacemaker/ICD infection are skin flora, including Staphylococci and Corynebacteria.11 Although several studies have focused on the mechanism of device infection, there is a considerable uncertainty about the sources of device infection.12,13 One pathophysiological mechanism is contamination of the pocket at the time of device implantation. Repeat interventions may reduce the patient's defence to the bacteria which might become clinically relevant after revision. Therefore, the aim of the study was to evaluate the prevalence of bacterial colonization of generator pockets in ICD patients without signs of infection and to analyse the impact of bacterial colonization on the incidence of device infection during follow-up.

Methods

In 122 ICD patients without signs of infection undergoing generator replacement or surgical lead revision at the Cardiac Arrhythmia Institute of Ludwigshafen, Germany, between January 2006 and July 2008, microbiological cultures of generator pockets and extracted leads were obtained. Patients with clinical evidence of a device infection were excluded. One patient was excluded due to signs of infection at surgery. Informed consent was obtained from all patients before participation in the study. The study was performed in corporation with one surgeon who performed all surgical interventions in the study. His speciality was lead revision. The study population, therefore, represents consecutive patients undergoing ICD replacement or lead revision performed by this surgeon. Surgical interventions performed by other surgeons were not included into the study (n = 102 with generator replacement). Consequently, the number of patients with lead revision included into the study was higher than the number of patients needing only generator replacement.

Microbiological cultures

Collections of the specimens for culture were obtained during revision in the operating room. The incision was started after the patient was prepped and draped for the procedure under sterile conditions. The pocket location of all patients was subpectoral. After opening the pocket, it was inspected for fluid or visual evidence of infection. In case of visual evidence of infection, the patient was excluded from the present study. A swab specimen was obtained from the deep portion of the pocket after the removal of the device from the pocket. Microbiological cultures were taken from the extracted leads as well. Thus, in two-thirds of patients two or more microbiological cultures were obtained. Before closing the pocket, it was routinely soaked with a polyvidon-iod solution (Braunol®). The time between specimen acquisition and processing in the microbiology laboratory ranged from 2 to 6 h. The fibrous capsule of the old pocket was not routinely removed or modified during the procedure.

Blood sampling, antibiotic treatment, and patient follow-up

All patients underwent the following work-up before surgery: complete clinical examination, routine blood sampling including complete blood counts, creatinine and C-reactive protein assays. Imaging studies consisted of chest X-ray and transthoracic echocardiography. All patients received periprocedural antimicrobials. Cefuroxime therapy was started prior to the surgery and was continued for 3 days. Cefuroxime therapy had also been used routinely for the first ICD implantation as single-shot therapy. Patients were discharged routinely from hospital on Day 1 or 2 after surgery following a chest X-ray, inspection of the wound, and a routine ICD control. After discharge, all patients visited the defibrillator outpatient clinic every 3 months as well as in case of any adverse event. Evaluation in ICD clinic consisted of clinical examination of the patient, inspection of the wound, and interrogation of the ICD. The median follow-up was 203 days after the revision.

A device infection was defined as previously described.11 Clinical evidence of device infection included local signs of inflammation at the generator pocket, including erythema, warmth, fluctuance, wound dehiscence, erosion, tenderness, or purulent drainage. Presence of ICD-related endocarditis was clinically confirmed when valvular or lead vegetations were detected by echocardiography, or if the Duke criteria for infective endocarditis were met.14,15 Vegetation was defined as an oscillating intracardiac mass on the electrode leads, cardiac valve leaflets, or endocardial surface in the setting of valve or lead infection confirmed by imaging in more than one echocardiographic plane, and positive blood and/or lead tip cultures.14,16,17

A device infection was microbiologically confirmed based on positive cultures from the generator pocket, lead(s), or blood (in the presence of local inflammatory signs at generator pocket or absence of another source of bacteraemia and resolution of blood stream infections after device explantation). The type of organism and its antibiogram were then compared with the cultures obtained from the previous intervention. A device infection with the same organism was based on the biotype and the antimicrobial susceptibility pattern.

Statistical analysis

Absolute numbers, percentages, means and standard deviation, median and upper/lower quartile were computed as appropriate. The Wilcoxon rank-sum test was used for the comparison of continuous variables in two groups. Categorical variables were compared by using the χ2 or Fisher's exact test, as appropriate. Exact confidence intervals (CI) were calculated according to the Pearson–Clopper formula. All P-values were two-tailed. A P-value less than 0.05 was considered to be statistical significant. The tests were performed using the SAS statistical package, version 8.02 (Cary, NC, USA).

Results

Characteristics of the patients

A total of 122 ICD patients underwent generator replacement or surgical lead revision between January 2006 and July 2008 and were included into the study. The median time from last surgical intervention to the present surgical revision was ∼3 years. The low median time to revision was due to the high number of lead revisions. Two-thirds of the interventions were lead revisions. Demographic and clinical characteristics of the study subjects are summarized in Table 1. The median age was 67 years. Seventy-seven percent were male. In all patients, leads were placed transvenously and the pulse generator was placed in the subpectoral area. Half of the patients had a primary prophylactic indication for an ICD. The prevalence of renal impairment was not different between both groups (Table 1). No patient in each group was on dialysis. C-reactive protein was not different between both groups. The median C-reactive protein in both groups was <0.5 mg/dL.

View this table:
Table 1

Clinical characteristics of implantable cardioverter defibrillator patients undergoing generator replacement or surgical lead revision: patients with positive vs. negative cultures

Positive culture (n = 40)Negative culture (n = 82)P-value
Clinical characteristics
 Age (years)66 (59–71)*67 (58–75)*0.92
 Male (%)83740.23
 Coronary artery disease (%)53550.80
 Dilated cardiomyopathy (%)25280.30
 Ejection fraction <40%60670.44
 Diabetes (%)15180.65
 Renal impairment (%)980.98
Chambers
 Single (%)28240.71
 Dual (%)37490.24
 Biventricular (%)35270.35
No prior revisions (%)68620.57
Time from last intervention (months)40 (12–57)38 (10–60)0.68
Type of present intervention
 Generator replacement only (%)38300.44
 Lead revision only (%)37470.68
 Generator replacement and lead revision (%)25280.72
  • *Median values are given with interquartile ranges (the 25th and 75th percentiles).

  • Serum creatinine ≥124 µmol/L.

Microbiological cultures

Positive cultures from the generator pocket and leads were found in 40 (33%) patients. The most common bacteria isolated were coagulase negative staphylococci (68%, Table 2). Gram-positive organisms were found in nearly all patients with positive cultures (95%), whereas gram-negative bacteria were cultivated in only one patient. In detail, coagulase negative staphylococci were found in 27 patients, Staphylococcus aureus in one patient, other gram-positive cocci in four patients, gram-positive bacilli in one patient, Propionibacterium acnes in seven patients, Corynebacterium in one patient, Pseudomonas spp. in one patient, and Trichosporon beigelii in one patient. Polymicrobial cultures were present in three patients (Table 2). Seventy-one percent of bacteria were cefuroxime sensitive. In 82 patients, two or more microbiological cultures were obtained. In 59 of these patients none of the samples became positive. In 22 patients, either the pocket culture or the lead culture became positive, whereas only one patient had positive cultures in both the generator pocket and extracted lead but with different bacteria specimen.

View this table:
Table 2

Bacterial isolates

Isolated n (%)
Gram-positive organism
Staphylococcus epidermidis19 (48)
 Other coagulase negative staphylococci8 (20)
Staphylococcus aureus1 (2.5)
 Other gram-positive cocci4 (10)
Bacillus species1 (2.5)
Propionibacterium acnes7 (17.5)
Corynebacterium species1 (2.5)
Gram-negative organism
Pseudomonas luteola1 (2.5)
Other organism
Trichosporon beigelii1 (2.5)

Clinical events

Follow-up data were available for all patients. No patient died during the follow-up. During a median follow-up time of 203 days after revision device infection occurred in three (7.5%, CI 1.6–20.4%) patients of the positive culture group vs. two (2.4%, CI 0.3–8.5%) patients of the negative culture group (P = 0.33). Time from revision to infection was 108 ± 73 days in patients with positive culture vs. 60 ± 39 days in patients with negative culture (P = 0.50). In all three patients of the positive culture group, the organism responsible for the device infection was identical to the one obtained during the prior intervention. In one patient no microorganism was identified. Under the suspicion of pulmonary infection, this patient was treated with antibiotics for several weeks before the diagnosis of device infection was made. One patient presented with pocket infection, three patients suffered from lead endocarditis and one patient had lead and tricuspid valve endocarditis. Clinical characteristics and microbiology of these five cases are summarized in Table 3. All patients were subsequently cured with complete removal of the device. None of the remaining patients demonstrated signs of an ongoing or new post-operative infection during the follow-up period.

View this table:
Table 3

Characteristics of patients who developed device infection after revision

Patient 1Patient 2Patient 3Patient 4Patient 5
Culture result from revisionS. epidermidisS. epidermidisS. aureusNegativeNegative
Culture result from infected deviceS. epidermidisS. epidermidisS. aureusNo microorganism identifiedS. epidermidis
Cefuroxime sensitiveNoYesYesn.a.Yes
Clinical presentationPocket infectionLead endocarditisLead endocarditisLead and valve endocarditisLead endocarditis
Age (years)6867645565
SexMaleMaleMaleFemaleFemale
Number of chambers of ICD23223
Number of prior revisions21323
Time from last intervention (days)98202259821
Type of last interventionGenerator replacementGenerator replacementGenerator replacementGenerator replacement and lead revisionLead revision only

Discussion

Major findings

A third of ICD patients without signs of infection undergoing generator replacement or lead revision have positive swab cultures from the generator pockets or extracted leads. After revision, 7.5% of those patients with a positive culture develop a device infection compared with 2.4% of patients with a negative culture (P = 0.19). In all patients of the positive culture group, the organism responsible for the infection is identical to the one obtained from the prior intervention.

Prevalence and relevance of positive swab cultures in asymptomatic patients

Common criticisms of swab techniques include frequent contamination, drying, inadequate specimen quality, and the superiority of other isolation techniques.1820 Despite its limitations, numerous studies on swab culture have been done on chronic wounds,2122 diabetic foot infection,23 burns,24,25 and post-operative wound infections.26 The studies have shown good correlation between the isolation of microorganism with swabs and tissue biopsy culture in chronic wounds and in the initial evaluation of limb-threatening infections.2123 In the present study, a third of ICD patients undergoing generator replacement or lead revision had positive swab cultures from the generator pockets or extracted leads. The high rate of positive swab cultures is in accordance with previous studies which found a similar prevalence of positive cultures in asymptomatic patients.27,28 However, caution must be exercised when interpreting culture results in the absence of clinical signs and symptoms of infection. The cultivation of skin flora like Propionibacterium acnes or Trichosporon beigelii from the pocket swab or extracted leads indicates that contamination should be taken into account when making adjudgement about the swab culture results. Swab specimens were made after removal of the device from the pocket. This order might have increased the proportion of contamination. On the other hand, the cultivation of Staphylococcus epidermidis or S. aureus might be of clinical relevance as S. epidermidis and S. aureus are known to cause device infections which can even become life-threatening.11 This is underlined by the observation that 2 of 19 patients with S. epidermidis and one of one patient with S. aureus in the swab culture developed device infection. In all three patients, the bacteria responsible for the device infection had the identical biotype and same antimicrobial susceptibility pattern as the bacteria which were previously obtained from the patient during the previous revision procedure. Device infection and endocarditis are a severe complication of device therapy associated with a high mortality rate which ranges from 31 to 66% if the infected device is not removed and 18% or less with complete device removal and antimicrobial therapy.14,29,30 Early and adequate antibiotic therapy is essential for the treatment of device infection. Still the cultivation of microorganisms from infected devices and the preparation of an antibiogram need several days. Generally, calculated antibiotic therapy is started and has to be adapted after a few days when the results of the antibiogram are available. Therefore, especially at the beginning of the antibiotic therapy, swab cultures of previous interventions could be useful for starting the right antibiotic therapy.

Increased risk of device infection after elective generator replacement or lead revision

Secondary procedures such as pulse generator replacement are a risk factor for infection.9 The rate of infectious complications in patients who undergo multiple implantations of devices in their lifetime is markedly increased.31,32 In a retrospective analysis, the infection rate in young patients who underwent a median of two pacemaker implantations was 5.5% and consequently five-fold increased compared with de novo implantation.31 Early reinterventions for haematoma or lead dislodgement were the leading risk factor of infection, associated with an adjusted OR of 15.04.31 Harcombe et al.10 showed that the complication rate after elective pacemaker replacement was 6.5% vs. 1.4% after first implantation of a permanent pacemaker and the major complications were caused by erosion and infection. These findings are in accordance with the present study where four of five patients developing clinical infection had two or more prior revisions indicating the importance of re-operations in the risk of infections. The device infection rate was 4% after revision compared with 0.8% after de novo implantation. In contrast to previous reports on device infection,9,10 the number of lead infections was higher than the number of cases with only pocket infection. This result might be due to the high rate of lead revisions—two-thirds of patients underwent lead revision—and the subpectoral implantation site of the device.

Several sources for infection of the pacemaker/ICD pocket and electrode have been postulated. One possible source is contamination of the pocket at the time of device implantation. Bacteria may remain dormant, possibly in the state of equilibrium with the host defence mechanisms until something changes this equilibrium and infection becomes overt.33 The present study shows clearly that bacterial colonization is present in up to a third of ICD patients. Furthermore, 7.5% of these patients develop a device infection with the same species of microorganism which was cultivated from the swab cultures taken during the previous intervention. This observation underlines the hypothesis that primarily asymptomatic bacterial colonization can cause device infection after revision and may explain the increased infection rates after device replacement or lead revision compared with those after de novo implantation. Bacterial contamination is becoming a growing problem of device therapy, as there are more patients with heart failure who get ICD's, especially ICD's with cardiac resynchronization therapy.2 The risk of bacterial contamination may be correlated with the implantation duration and the number of leads, and the risk of infection may be increased in immunocompromized patients with severe heart failure.1,11 Device infection is underestimated in the calculation of device complications that are based on the trials with a follow-up duration shorter than longevity of ICD's. This fact may have an important impact on analysis about ICD cost-effectiveness over patients' lifetime.34 Further, the adverse events of the increasing number of device recalls may be underestimated.32

Strategies to prevent device infection after revision

All patients received periprocedural antibiotic therapy for 3 days with cefuroxime. Guidelines do not differentiate between de novo implantation and revision with regard to the antibiotic therapy, even if it is known that the infection rate is significantly higher in patients undergoing revision.3,31 The prolonged antibiotic therapy of 3 days with cefuroxime did not prevent device infections; even three of four patients had device infection with cefuroxime sensitive microorganisms. These findings suggest that antibiotic therapy may not sufficiently penetrate into generator pockets. Considering the high prevalence of asymptomatic bacterial colonization, further studies are warranted if a routine excision of the old generator pocket or a prolonged and calculated antibiotic therapy in patients who have contamination with S. aureus or S. epidermidis might reduce the infection rate after revision.

Limitations of the study

This study is a clinical, observational study from a single centre and therefore site-specific influences could bias the results. This includes site-specific operating facilities, antibiotic prophylaxis or pocket localization. As all revisions were performed in the operating room, these results may not be extended to revisions that are implanted in the cardiac catheterization laboratory.35

It is unknown how many of the positive cultures represent contamination during surgery. Swab specimens were made after removal of the device from the pocket. This order may increase the proportion of contamination.

Because of the speciality of the surgeon on lead revisions, the number of patients with lead revision included into the study was higher than the number of patients needing only generator replacement. Removing ICD leads sometimes is hard work and time-consuming bearing a considerable risk of contamination and may have increased the incidence of positive swab cultures. However, the incidence of device infection in the 102 patients with generator replacement not included into the study was not different compared with the incidence in the study population (4% in both groups).

Conclusions

A third of ICD patients without signs of infection undergoing generator replacement or lead revision have positive swab cultures from the generator pockets or extracted leads, indicating that asymptomatic bacterial colonization is present in up to a third of patients carrying a device. In 7.5% of these patients bacterial colonization becomes clinically relevant after revision causing device infection. The finding of S. epidermidis or S. aureus in generator pockets of asymptomatic patients is clinically relevant as bacterial colonization with S. epidermidis led in 10% (2 of 19) of patients to device infection and bacterial colonization with S. aureus in 100% (one of one). Prolonged peri-operative antibiotic therapy with cefuroxime does not prevent device infections due to cefuroxime sensitive microorganisms. Therefore, further studies to develop strategies in order to prevent device infections are warranted.

Conflict of interest: none declared.

References

View Abstract