Systemic Infection Related to Endocarditis on Pacemaker Leads
Clinical Presentation and Management
Background Endocarditis related to pacemaker (PM)-lead infection is a rare but serious complication of permanent transvenous pacing. To determine in which situations the diagnosis should be evoked and to determine optimal management, we reviewed our experience with endocarditis related to PM-lead infection.
Methods and Results Fifty-two patients were admitted for endocarditis related to PM-lead infection. The presentation was acute in 14 patients, with onset of symptoms in the first 6 weeks after the last procedure on the implant site, and chronic in 38 patients. Fever occurred in 86.5%. Clinical and/or radiological evidences of pulmonary involvement were observed in 38.4%. Pulmonary scintigraphy disclosed pulmonary infarcts in 31.2%. Local complications were found in 51.9%. Elevated C-reactive protein was found in 96.2%. A germ was isolated in 88.4% of patients and was a Staphylococcus in 93.5%. Transthoracic echocardiography demonstrated vegetations in only 23% of patients, whereas transesophageal echocardiography disclosed abnormal appearances on the PM lead in 94%. We systematically tried to ablate all the material. Two techniques were used: percutaneous ablation or surgical removal during extracorporeal circulation. All patients were treated with antibiotics after removal of the infected material. Two patients died before lead removal and 2 after surgical removal; the predischarge mortality was 7.6%, and the overall mortality was 26.9% after a follow-up of 20.1±13 months.
Conclusions The diagnosis of endocarditis related to PM-lead infection should be suspected in the presence of fever, complications, or pulmonary lesions after PM insertion. Transesophageal echocardiography should be performed to look for vegetations. Staphylococci are involved in the majority of these infections. The endocardial system must be entirely removed and appropriate antibiotic therapy pursued for 6 weeks.
Endocarditis related to PM-lead infection is a rare but serious complication of permanent transvenous pacing. The reported incidence after permanent endocardial PM implantation varies in the literature from 0.13% to 7%.1 2 3 4 5 6 7 Medical treatment alone is rarely successful, and several studies have suggested that the material should be removed quickly for optimal management.8 9 10 Such ablation can be performed by use of percutaneous or surgical techniques; the relative place of these techniques is not clear but probably depends on the size of the vegetations and on the time elapsed before diagnosis. The natural history of such infection is dismal, with a mortality rate of 33%. We reviewed our experience with endocarditis related to PM-lead infection to determine (1) the clinical situations in which the diagnosis should be suspected, (2) the relative value of biological markers, echocardiography, and pulmonary scintigraphy to support the diagnosis, and (3) its optimal management.
Between 1992 and 1995, 52 patients were admitted with suspected infection on PM leads. This study was not based on the examination of cardiac pacing files; rather, patients were referred from peripheral hospitals to our tertiary referral center after initial management by physicians from various specialties.
All patients underwent the following procedures:
1. Blood cultures on 3 consecutive days (three cultures from each patient with additional cultures if the temperature was >38.5°C or <36°C). Cultures at the site of battery implantation were performed when appropriate (wounds, local infection, or PM exteriorization). All cultures were performed before antibiotic therapy except in the case of patients referred from other institutions who were already receiving antibiotic therapy and in whom a germ had been isolated before admission.
2. Cultures of the leads and of the PM were done after ablation. In the majority of patients, the ablation was performed without interruption of the antimicrobial therapy. We did not look for rarer pathogens such as Legionella or Q fever, which have been found in endocarditis.
3. Repeated blood counts and measurements of creatinine, CRP, CICs, fibrinogen, and the erythrocyte sedimentation rate.
4. TTE and TEE performed by two different investigators. Serial echocardiograms for suspicion of an infected lead were performed by the same two investigators (except for four patients).
5. Chest roentgenogram.
6. Ventilation perfusion pulmonary scintigraphy before and after lead ablation in the majority of cases.
To exclude other sources of infection, the following were used:
1. Dental pantomogram, sinus radiographs, and specialist referral where appropriate.
2. Abdominal ultrasound.
3. Specialist referral and accessory investigations when warranted by the clinical findings.
Ablation of the Material
We routinely tried to ablate all the material (PM and all leads). Two techniques were used:
1. Percutaneous ablation. The leads were removed by simple traction. If the tip was fixed in the myocardium, a stylet wire (Cook pacemaker-lead–extraction system) was used, and if the lead stopped in the sternoclavicular angle, special sheaths or a lasso introduced via the femoral vein was used to complete the ablation.
2. Surgical removal during extracorporeal circulation.
The choice of the technique of ablation (surgical versus percutaneous) was based on the size of the vegetations on TEE, the presence of an alteration of the tricuspid valve, and the general condition of the patient. After the seventh patient, we adopted a standardized policy for lead removal (Fig 1⇓): (1) if TEE demonstrated vegetation >10 mm, the material was surgically removed; (2) if TEE did not demonstrate vegetations or demonstrated vegetation ≤10 mm, the material was percutaneously removed.
All patients received intravenous antibiotics before lead ablation. This therapy was adapted to the germ found at the blood culture. If blood cultures were negative, empirical antistaphylococcal therapy was given (usually vancomycin and amikacin, except for patients referred from other institutions with alternative treatment combinations that were clinically effective). Two antibiotics were usually used, but 3 were administered in case of pulmonary infection. Antibiotic therapy was continued intravenously for 2 weeks after lead ablation and subsequently by mouth for 4 weeks; the antibiotic was adapted to the germ found at lead culture. Antibiotic therapy was discontinued after outpatient consultation to verify the absence of recurrent symptoms and normalization of the systemic parameters of inflammation.
Pacing After Ablation
An abdominal PM with permanent epicardial stimulation was inserted after surgical or percutaneous removal in patients requiring a permanent ventricular stimulation. In the other cases, the indication for permanent pacing was reevaluated. If necessary, a new transvenous permanent PM was inserted 1 to 2 months after the percutaneous removal of the material.
An outpatient clinic was used for follow-up 4 weeks after discharge. Further follow-up was obtained by phone contact with the patients and their physicians.
Data were analyzed by paired t tests or in two-by-two tables using the χ2 test and Fisher’s exact test. A value of P<.05 was accepted as significant.
Fifty-two patients (21 women and 31 men aged 64.7±17.7 years) were referred to our institution for suspected infection on PM leads.
The presentation was acute in 14 patients (6 women and 8 men aged 72.6±12.7 years), with onset of symptoms in the first 6 weeks after the last procedure on the implant site, and chronic in 38 patients (15 women and 23 men aged 61.9±18.6 years), with >6 weeks from the last procedure on the implant site to the onset of symptoms. Only 3 patients in the chronic group had the first symptom of infection 6 to 12 weeks after the last procedure.
In the acute group, 9 patients (64.3%) had a definite infective endocarditis on the basis of the Duke criteria, and 5 (35.7%) had a possible infective endocarditis with one major clinical criteria but fewer than three minor clinical criteria.
In the chronic group, 36 patients (94.7%) had a definite infective endocarditis on the basis of the Duke criteria, and 2 (5.3%) had a possible infective endocarditis with one major clinical criteria but fewer than three minor clinical criteria.
Thirteen patients (25%; 3 of 14 in the acute group and 10 of 38 in the chronic group) had impaired immunity due to diabetes mellitus in 3 patients, a neoplasm in 4, corticosteroid therapy in 2, alcohol abuse in 2, hemodialysis for chronic renal failure in 1, and an eosinophilic myocardiopathy in 1.
In the acute group, the last intervention at the PM site was a first implantation in 11 patients (2 of whom required a new intervention due to lead dysfunction) (Fig 2⇓). In the remaining 3 patients, it was a lead change or repositioning several months after the first PM implantation.
Four patients had temporary endocardial pacing before implantation. In two patients, a hematoma occurred after the PM insertion that required evacuation in one case.
In the chronic group, the last intervention was a first implantation in 14 patients, elective PM replacement in 3, lead replacement in 8, and PM exteriorization or local infection in the remaining 13 (Fig 2⇑). The mean number of local interventions at the PM site was 2.5±1.7 (range, 1 to 8) with a mean lead number of 2.1 (range, 1 to 4). One patient had temporary endocardial pacing after right femoral vein catheterization, and 2 patients had an electrophysiological study with a right femoral vein catheterization just before or just after the PM insertion. Three patients developed a hematoma after implantation.
Symptoms (Fig 3⇓)
In the acute group, the time elapsed between the last intervention and the first symptom was 4 days (range, 1 to 12 days). Fever was present in 13 patients (92.8%). Local pain at the PM site was found in 6 patients (42.8%), followed by a purulent discharge in 3 of them. Pneumonia was observed in 2 patients (14.3%) and aseptic pulmonary embolism in another patient (7.1%). Only 1 patient had isolated local symptoms. Fever was isolated without pulmonary or local symptoms in 6 patients. One patient died from pulmonary embolism.
In the chronic group, the time elapsed between the last intervention and the first symptom was 25±28 months (range, 1 to >120 months), and the time elapsed between symptoms and the diagnosis was 8±12 months (range, 0 to 48 months).
Fever and chills were present in 32 patients (84.2%) with asthenia, and wasting was present in 10. Twenty-two (68.7%) of these patients described episodes of fever, with an episode of septic shock in 2 cases. Eleven patients had repeat episodes requiring several trials of antimicrobial therapy.
Arthralgia was present in five patients (13.2%). A history of spondylitis (8 to 12 months before the diagnosis) was found in two patients (5.3%); in one, the occurrence of repeated episodes of spondylitis led to the diagnosis.
Local symptoms occurred in 21 patients (55.3%): inflammation or discomfort in 3 patients, infection in 13, and exteriorization of the material in 5. Local infection or material exteriorization was isolated in 7 patients. Local symptoms were recurrent in 13 patients and were responsible for the great number of local interventions at the PM site before the diagnosis of PM-lead infection (3.86±2 local interventions versus 1.37±0.49 in the population without local infection or exteriorization; P<.0003 by t test). The time elapsed between the last intervention and the first symptom was 26.7±28.8 months when patients had local symptoms versus 23.5±29.0 months in patients without local symptoms (P=NS).
Pulmonary lesions were found in 17 patients (44.7%). A pleural effusion was present in 4 patients (pleural fluid cultures performed in 3 cases were negative). Pneumonia with pulmonary consolidation occurred in 6 patients and was recurrent in 2 cases; a germ was isolated in 4 of these 6 cases. One case of pneumonia evolved into a pulmonary abscess, and one was associated with septic shock. Interstitial pneumonia was found in 1 patient. Previous recurrent clinical pulmonary embolism, documented by pulmonary scintigraphy, was found in 3 patients. Five patients had a history of recurrent bronchitis with fever, cough, and expectoration. Eight of these 17 patients had undergone extensive investigations and had received antimicrobial therapy for these pulmonary symptoms; however, the diagnosis of infection related to the PM had not been specifically considered.
Blood Results (Fig 4⇓)
In the acute group, erythrocyte sedimentation rate and CRP were increased in 13 patients (92.8%). CICs were found in 10 patients (71.4%). Blood cell counts showed leukocytosis in 7 patients (50%), in all cases due to an increase in polymorphonuclear cells.
In the chronic group, erythrocyte sedimentation rate and CRP were increased in 37 (97.4%) and 36 (94.7%) patients, respectively. CICs were observed in 75.9% of patients. Blood cell counts showed leukocytosis in 24 patients (63.2%), in all cases due to an increase in polymorphonuclear cells.
Echocardiography (TTE and TEE) (Fig 5A and 5B⇓)
In the acute group, TTE did not disclose abnormalities related to the PM lead except in 1 patient in whom a large vegetation was visible in the right ventricle. No further examination was performed in this patient, who died suddenly with symptoms of pulmonary embolism. TEE, performed in 12 patients, found abnormal echoes on the PM lead in 91.7% of patients. Vegetations were seen in 8 patients: ≥10 mm in 1 patient (11 and 14 mm); ≥5 but <10 mm in 5 patients (4 in the atrium and 1 in the ventricle); and <5 mm in 2 patients (1 in the atrium associated with a vegetation on the tricuspid valve and 1 in the ventricle associated with a sleevelike prolongation onto the lead). This sleevelike appearance was found in 4 patients (isolated in 2 and associated with a vegetation in 2).
In the chronic group, interpretable TTE was possible in 33 patients. TTE disclosed vegetation on the PM lead in 10 (30.3%) of these patients. TEE performed in 38 patients was normal in 2 and abnormal in 36 (94.7%); a vegetation was observed in 34 patients (≥20 mm in 7 patients; ≥10 but <20 mm in 7 patients; ≥5 but <10 mm in 17 patients; and <5 mm in 3 patients), and a sleevelike appearance was visible in 10 patients (26.3%, associated with vegetations on all but one occasion). A vegetation was detected on the tricuspid valve in 6 patients (15.8%), and an additional perforation was evidenced in 1 case. One patient, with a PM lead in the left ventricle through a patent foramen ovale, had vegetations on the mitral valve.
Ventilation Perfusion Pulmonary Scintigraphy (Fig 6A⇓)
In the acute group, scintigraphy was not performed in two patients (one with a bilateral pneumonia and one who died suddenly). Pulmonary embolism occurred in three patients (21.4%) before PM-lead ablation. Scintigraphy disclosed embolism in two patients; the other patient died suddenly of pulmonary embolism. One of the surviving patients had three vegetations visible during TEE, one of them 6 mm long; the other had two large vegetations (14 and 11 mm) associated with a sleevelike appearance on the PM lead.
In the chronic group, pulmonary embolism was found before PM-lead ablation in 13 patients (34.2%). Five of these patients had no clinical history suggestive of pulmonary embolism, 3 described clinical pulmonary embolism, 4 had pneumonia (recurrent in 2 cases), and 1 had recurrent bronchitis. There was no significant statistical difference in vegetation size between patients with and without pulmonary embolism.
Bacteriology (Fig 7⇓)
In the acute group, a germ was isolated in 10 patients (71.4%) in blood culture in 7 (50%) of 14 patients, local cultures in 3 (21.4%) of 14 patients, and lead culture in 9 (69.2%) of 13 patients. The germ was an S. aureus in 5 patients (50%) (3 MetiS, 2 MetiR), an S. epidermidis in 3 patients (30%) (2 MetiS, 1 MetiR), an S. species MetiR in 1 patient (10%), and a Pseudomonas aeruginosa G in 1 patient (10%). Only 1 patient had another potential source of infection (leg ulceration). All patients received antibiotic therapy before lead ablation. The duration of antibiotic therapy before lead ablation was 10.4±8.2 days in patients with a positive lead culture versus 16.2±11.3 days in patients with a negative lead culture (P=NS). No germ was detected in 4 patients, but 2 of these had recurrent fever when intravenous antibiotics were discontinued. TEE had disclosed vegetations in the 5 patients with negative lead cultures.
In the chronic group, a germ was detected in 36 patients (94.7%) in blood cultures in 24 (63.2%) of 38 patients, local cultures in 12 (33.3%) of 36, and lead cultures in 29 (80.6%) of 36. The germ was an S. aureus MetiS in 5 patients (13.9%), an S. epidermidis in 27 (75%; 20 MetiS, 3 MetiR, and 4 MetiR+MetiS), another S. in 4 (2 capiris, 1 specie, 1 cohnii) (11.1%; 3 MetiS, 1 MetiR), and a Gram-negative bacillus (1 Serratia, 1 Escherichia coli, and 1 Enterobacter cloacae) in 3 patients (8.3%). Two different germs were found in 7 patients, 4 with two different S. epidermidis MetiR and MetiS, 1 with S. aureus and S. epidermidis found in several blood cultures, and 2 with an S. associated with a Gram-negative bacillus. All but 1 patient received an antibiotic therapy before lead removal. Two antibiotics were used in 18 of 28 and 3 of 7 patients with a positive or a negative lead culture, respectively, and three antibiotics were used in 10 of 28 and 4 of 7 patients with a positive or a negative lead culture, respectively (P=NS by Fisher’s exact test). The duration of antibiotic therapy before lead ablation was 9.7±6.1 days in patients with a positive lead culture versus 15.3±6.2 days in patients with a negative lead culture (P=.04).
The search for a possible source of infection, other than local contamination, was always negative, regardless of the germ found.
In our population, all but one of the cases of S. aureus in the acute group were associated with the presence of local symptoms. The time between the last intervention and the first symptom was 6.4±5.2 months when the responsible germ was an S. aureus versus 27.6±26.8 months when the germ was an S. epidermidis (P=.0010 by Welsh test).
Eight patients in the acute group and 27 in the chronic group met the Duke pathological criteria (positive lead culture). The clinical criteria (Table 2⇑) enabled us to classify these patients in the definite endocarditis group before lead cultures were done for 2 (25%) and 16 (59.3%) of the patients in the acute and chronic groups, respectively. If we had considered local symptoms as a major clinical criteria, 5 (62.5%) and 21 (77.8%) of the patients in the acute and chronic groups, respectively, would have been classified as having definite endocarditis by the clinical criteria alone. If we had considered local symptoms as a minor clinical criteria, only 1 more patient in the chronic group would have been classified as having definite endocarditis by the clinical criteria alone. If we had considered pulmonary symptoms as a major clinical criterion, 3 (37.5%) and 19 (70.4%) of the patients in the acute and chronic groups, respectively, would have been classified as having definite endocarditis by the clinical criteria alone. If we had considered pulmonary symptoms as a minor clinical criterion, no more patients would have been classified as having definite endocarditis. If we had considered local and pulmonary symptoms as major clinical criteria, 7 (87.5%) and 23 (85.2%) of the patients in the acute and chronic groups, respectively, would have been classified as having definite endocarditis by the clinical criteria alone.
One patient with infection on PM leads located in the left ventricle with vegetations on the mitral leaflets underwent surgical ablation as primary therapy.
Two patients died before lead ablation. No ablation was attempted in an 81-year-old patient with a lead segment in the superior vena cava whose general condition precluded anesthesia and surgery.
Percutaneous removal was performed as primary therapy in 38 patients with 78 leads (31 atrial and 47 ventricular leads); the mean number of leads removed was 2.06 per patient, with a maximum of 4 leads in a single patient. The vegetation size was ≤10 mm in all but 5 patients: the 3 initial patients in the series with vegetations >10 mm and 2 elderly patients with vegetation sizes of 14 and 16 mm who were in poor general condition. Ablation failed in 2 patients (5.2% of patients) in whom disruption occurred in the proximal part of the leads (1 atrial and 1 ventricular; 2.6% of leads). Disruption of the distal part occurred in 9 ventricular leads (11.5% of all leads) with recovery of all of the silicone; the wire-rope tip, however, was not recovered. The location of these tips was extravascular in 2 patients, at the right ventricular apex in 5, in the right atrium in 1, and in the pulmonary circulation (periprocedural embolization) in 1. Two cases of tricuspid regurgitation and 1 chordal disruption without regurgitation appeared after ablation of 4 leads in 3 patients without previous clinical or echocardiographic evidence of tricuspid endocarditis. Pericardial effusions were observed after 6 lead-extraction procedures in 3 patients, 1 of whom had right atrial injury noted during surgical ablation. Pulmonary scintigraphy was performed in 33 patients before and after lead ablation, leading to the demonstration of vegetation migration during ablation in 10 patients (30.3% of the patients; 10 of 66 leads, or 15.1% of the leads). TEE demonstrated lead abnormalities in 7 of these 10 patients (Fig 6B⇑).
Two (40%) of the 5 patients with vegetation size >10 mm had scintigraphic evidence of pulmonary embolism after lead removal, and pulmonary embolism was diagnosed by scintigraphy before lead removal in another patient. There was no significant statistical difference in vegetation size between patients with and without pulmonary migration. Only one patient (with a vegetation size of 40 mm) developed clinical evidence of septic pulmonary embolism with hemodynamic consequences. No deaths were observed after percutaneous ablation.
Surgical removal was performed as primary therapy in 10 patients with 21 leads (8 atrial and 13 ventricular) and after failure of percutaneous removal in 2 patients with 6 leads (3 atrial and 3 ventricular). The mean number of leads removed was 2.1 per patient, with a maximum of 4 leads in a single patient. The vegetation size was >10 mm except in 1 patient with a vegetation 8 mm in diameter who also had 2 retained lead segments in the superior vena cava. Two patients had retained lead segments in the superior vena cava; in 1 case, the lead was trapped within a thrombus that occluded the vena cava. Tricuspid valve resection was performed in 5 patients during the surgical procedure; TEE had shown abnormalities on the tricuspid valve in 3 of these 5 patients. Perioperative inspection confirmed the TEE in all but 2 cases, in which the TEE underestimated the size of the vegetations on the wire and did not detect small vegetations on the tricuspid valve. All the vegetations seen on the valves during surgery were at sites of contact with the leads. An intramyocardial infection was seen on a papillary muscle in 1 patient and at the site of insertion of the wire in another. Two patients developed pneumonia (1 patient with lung abscess after septic embolism) with collapse and died after surgery.
Comparison Between Acute and Chronic Groups
Comparisons of clinical, biological, and echocardiographic data and results of percutaneous lead removal in both groups are summarized in Table 3⇓. Symptoms, blood results, and echocardiographic data were not statistically different between the two groups. The proportions of germs isolated and of definite endocarditis according to the Duke criteria were higher in the chronic than in the acute group. All the mechanical complications of percutaneous lead removal occurred in the chronic group.
The mean follow-up was 20.1±13 months (19.1±12.7 months in the acute group, 20.5±13.2 months in the chronic group). No patients were lost to follow-up.
Endocarditis recurred in a single patient after surgical removal. This patient, in whom two lead segments were left in the superior vena cava during the initial procedure, developed endocarditis on the tricuspid valve with the same germ (Serratia) 4 months after the surgical procedure. All the material was subsequently removed and a partial tricuspidectomy performed. Five months later, a contralateral PM was implanted, and 11 months later, septicemia recurred with the same germ. Extensive investigation to find a cause for the infection revealed a small intramyocardial segment of the previous PM lead. The material was removed surgically. The subsequent course was uneventful, with a follow-up of 14 months.
One patient in the acute group who underwent percutaneous removal without further subsequent permanent pacing presented 2 months later with endocarditis on the tricuspid valve that required tricuspidectomy.
Two patients died before lead ablation and 2 died after the surgical procedure. The overall predischarge mortality rate was 7.6%.
Ten patients died after hospitalization after a mean follow-up of 9.8±5.7 months, five in the acute group and five in the chronic group. Only three patients in the chronic group had symptoms of infection syndrome before death; interestingly, they were the patient without lead ablation and the patients with a wire rope not recovered in the right ventricle. The overall total mortality rate was 26.9%.
This report of a consecutive series of patients referred to a specialized tertiary referral center confirms and extends previous observations, generally case reports or small series of patients, that have documented the varying presentation of PM-lead–related endocarditis and its serious clinical implications.
In the acute form, the short time elapsed between PM implantation and the occurrence of infection facilitated the diagnosis. The vast majority of patients had systemic symptoms as well as local signs of infection. However, in one patient, isolated local symptoms occurred without fever; biological markers of systemic inflammation led to further investigations that revealed vegetations on the atrial lead on TEE. In a second patient, pneumonia occurred several days after PM implantation, in association with evidence of infection at the implant site. However, delay in the diagnosis led to death from pulmonary embolism. It therefore appears advisable to further investigate or to closely observe patients with apparently isolated local infection.
In the chronic form, the delay between the onset of symptoms and the diagnosis illustrates the difficulty in diagnosing PM-lead infection, as has been recognized previously.4 5 6 In our population, the delays in diagnosis were often related to the fact that PM-lead infection was not considered in the differential diagnosis. In other cases, possible clues to the diagnosis were ignored; for example, blood cultures positive for S. epidermidis were erroneously considered to be related to contamination of the specimens.12 13 Finally, the diagnosis was sometimes considered but wrongly excluded after the performance of inappropriate or insufficiently exhaustive tests; for example, negative results on TTE were sometimes considered by the treating physician as excluding the diagnosis of PM infection. The most common presentation was with fever or chills (84% of our population), often recurrent (68% of patients) and sometimes with associated symptoms suggestive of bronchitis. These episodes were generally not investigated further and were empirically treated by outpatient antibiotic therapy.
A history of local symptoms or signs related to the implant site was frequent (55.3% of our population) and often recurrent, and symptoms were erroneously considered to be related to local infection. The time between the last intervention and the first symptom was sometimes great (25±28 months; range, 1 to >120 months), and PM-lead infection was not considered in the differential diagnosis.
The diagnosis of systemic infection related to PM-lead infection must be systematically considered in the presence of chronic fever,6 13 14 recurrent bronchitis, or pulmonary infection15 or in case of recurrent or persistent evidence of infection at the implant site. In our population, there were no patients with clinical tricuspid regurgitation, but endocarditis of the right heart should be specifically excluded16 17 18 in cases with symptoms suggestive of pulmonary disease. Spondylitis was observed in only two (3.8% of all patients) of our patients and only in the chronic group; in one of these cases, recurrent spondylitis led to the diagnosis. Only a few cases of spondylitis as the presenting feature of PM-lead infection have been reported in the literature by Bonal et al19 and El Kohen et al.20
No patient in our series had symptoms of Osler’s endocarditis. The criteria suggested by Von Reyn et al21 are not adaptable to the systemic infection related to PM-lead infection. We have used the Duke criteria to define systemic infection related to PM-lead infection. These criteria are adapted to this pathology, but some other clinical criteria could probably be added, such as local symptoms and pulmonary infections.
Echocardiography, TTE, and TEE
TTE showed vegetations in only 7% and 30% of patients in the acute and chronic groups, respectively. This rate of detection is markedly lower than previously reported rates18 22 23 24 that suggested that TTE has an 80% sensitivity in the detection of vegetations in the right heart. The explanation for this discrepancy may relate to the fact that most of our patients had PM-lead–related endocarditis with intact tricuspid leaflets. TEE showed abnormalities on the PM leads in 91.6% and 94.7% of patients in the acute and chronic groups, respectively. The superiority of TEE over TTE has been reported previously.25 The significance of the often-seen sleevelike appearance of the PM lead is uncertain but rarely occurred in isolation (3 of 14 patients). The major advantage of TEE related to its ability to demonstrate tricuspid involvement, particularly in patients whose poor echogenicity rendered TEE unreliable.
Since 1979, six series have described patients with PM-related sepsis; S. aureus and coagulase-negative staphylococcus accounted for 80% of bacteremias. 6 26 27 This rate is in accordance with the present study. The 30% and 19.4% rates of negative lead cultures in the acute and chronic groups, respectively, are surprising. However, it may be that some vegetations (particularly small atrial vegetations) did not have an infectious origin.28 In a rabbit model, it has been shown that sterile vegetations may form on an intracardiac catheter.29 Preoperative antibiotic therapy could account for these negative lead cultures equally well. Finally, the amorphous material (slime) that is universally found in staphylococcal colonies30 31 32 could also be responsible for these negative cultures by isolating the lead and the germ from the culture material. In all these patients with negative cultures, the clinical evidence of PM-lead infection was compelling. The majority of these patients had a clinical history of infection, sometimes resistant to several trials of antibiotics or with recurrent fever when antibiotics were stopped. Furthermore, in all cases, ablation of the PM leads led to resolution of the symptoms.
It is commonly accepted that the most common portal is the subcutaneous site of insertion of the pacing system. Extension along the lead into the vascular system is the usual explanation for the localization of the infection to the lead. Bacterial colonization of the lead during the course of bacteremia whose origin is not related to the pacing system has been less well documented. This mode of contamination represented 14% in the series reported by Kugener et al33 and Bryan et al.34 The presence of an important fibrous coating that isolates the lead from venous blood explains the rarity of hematogenous contamination. In our population, local contamination at the PM site was the most common portal.
In our experience, as in the study by Camus et al,27 the high rate of uncontrolled infection or relapse among patients with septicemia in relation to PM-material infection confirms the need for immediate removal of the entire pacing system. This is obvious for systemic infection related to PM-lead contamination but is also important in obvious infection of the PM pocket or the subcutaneous part of the lead. Because of the adherence of staphylococci to the surfaces of intravenous leads and the poor bactericidal activity of antibiotics on adherent bacteria, failure of antimicrobial therapy is not unexpected; thus, immediate removal of the entire pacing system should probably be performed in all cases. The incidence of staphylococcal infection confirms the need for antistaphylococcal therapy in all cases. Our protocol for antimicrobial therapy seems to be reliable after removal of material with only one relapse. We think, however, that the PM material should be removed immediately rather than attempting prolonged antibiotic therapy alone.
The choice of the technique of ablation (surgical versus percutaneous) was based on the size of the vegetations assessed on TEE, the presence of morphological changes of the tricuspid valve, and the general condition of the patient. This retrospective study did not enable us to compare both methods. Septic pulmonary embolism was a major concern after percutaneous lead removal. At the beginning of our experience, on the basis of the initial cases, we decided to remove the material percutaneously only if TEE did not disclose vegetations >10 mm. This policy was based on the observations by Mugge et al,35 who found that embolism was more frequent when vegetation size was >10 mm in endocarditis related to valve infection, and by Robbins et al,36 who found more complications when vegetation size was >10 mm in a series of 21 patients with tricuspid endocarditis. Percutaneous ablation of infected leads was found to be a reliable technique in patients with vegetations ≤10 mm. The two major complications were vegetation migration without vital involvement and tricuspid injury. Retrospectively, we did not find any significant relation between pulmonary migration and vegetation size, but the number of leads removed in patients with large vegetations was small. It must be emphasized that pulmonary embolism with vital involvement occurred in the patient with the largest vegetation size (40 mm). Additionally, 2 patients (with vegetation sizes of 14 and 20 mm, respectively) died after surgical removal. Conversely, no deaths occurred after percutaneous ablation. Perhaps we should expand the indications for percutaneous removal to include patients with larger vegetation sizes. We did not perform a surgical procedure when a wire rope was not completely recovered after percutaneous ablation, but 2 of 9 such patients (2 of 7 patients with a wire rope in the circulation) died from infection syndrome; these patients probably needed more aggressive management.
The diagnosis of endocarditis related to PM-lead infection should be systematically considered in patients with fever, a history of local complications, or pulmonary pathology after PM insertion. We describe two different clinical presentations: the acute form, with an early presentation with sepsis, often in conjunction with local signs of infection, and a chronic form beginning several months later. The presentation may be atypical and the presenting symptoms may occur late after the last intervention at the PM site. CRP, CIC, and pulmonary scintigraphy are of diagnostic value. TEE must be performed in search of vegetations. Staphylococci are responsible for the vast majority of these infections, especially S. epidermidis in the chronic group and S. aureus in the acute group. Immediate removal of the entire pacing system is necessary, in addition to prolonged antimicrobial therapy including antistaphylococcal agents. The technique of removal (surgical versus percutaneous) depends on the size of the vegetations; percutaneous removal seems to be safer when vegetation size is ≤10 mm.
Selected Abbreviations and Acronyms
|CIC||=||circulating immune complex|
The authors would like to acknowledge Eugéne Mac Fadden, MD, for his critical review of the manuscript and helpful suggestions; Bruno Lequeuche, MD, for his help with the data collection; and Jean Paul Berregi, MD, and the radiology department for their expert assistance in performing intravascular material removal.
- Received July 31, 1996.
- Revision received November 4, 1996.
- Accepted November 20, 1996.
- Copyright © 1997 by American Heart Association
Block K, Russi E. Right heart endocarditis. Schweiz Med Wochenschr. 1989;47:1664-1672.
Glock Y, Sabatier J, Salvador-Mazencq M, Puel P. Les endocardites sur électrodes endocavitaires de stimulateurs cardiaques: a propos de 7 cas. Arch Mal Coeur. 1986;79:483-488.
Rubio-Alvarez J, Duran-Munoz D, Sierra-Quiroga J, Garcia-Bengochea JB. Right heart endocarditis and endocardial pacemakers. Ann Thorac Surg. 1989;48:147-151.
Hartstein AI, Jackson J, Gilbert DN. Prophylactic antibiotics and the insertion of permanent transvenous cardiac pacemakers. J Thorac Cardiovasc Surg. 1978;75:216-226.
Southall E, Pallister D. Meningitis and recurrent septicemia secondary to unsuspected pacemaker infection. Br Med J. 1982;284:1530-1537.
Bonal J, Bouchiat C, De Jaureguiberry JP, Duval JL, Carli P. Spondylodiscites revelatrices d’endocardites sur électrodes endocavitaires de stimulateurs cardiaques. Arch Mal Coeur. 1994;87:1735-1738.
El Kohen M, Millaire A, Leroy O, Beuscart C, Vanjak D, Sivery B, Senneville E, Ducloux G. Endocardites infectieuses tricuspidiennes sur sonde de stimulation permanente endocavitaire. Arch Mal Coeur. 1990;83:1855-1857.
Von Reyn CF, Levy BS, Arbeit RD, Friedland G, Crumpacker CS. Infective endocarditis: an analysis based on strict case definitions. Ann Intern Med. 1981;94:505-517.
Melvin ET, Berger M, Lutzker LG, Goldberg E, Mildvan D. Noninvasive methods for detection of valve vegetations in infective endocarditis. Am J Cardiol. 1981;66:523-525.
Daelemans R, Kersschot I, Van Den Branden F, Nagler J, D’Heer H, Parizel G, Brutsaert DL. Pacemaker endocarditis: contribution of two-dimensional echocardiography. Acta Cardiol. 1984;34:293-299.
Camus C, Leport C, Raffi F, Michelet C, Cartier F, Vilde JL. Sustained bacteremia in 26 patients with a permanent endocardial pacemaker: assessment of wire removal. Clin Infect Dis. 1993;17:46-55.
Lauck G. Adherent structures on implanted leads of cardiac pacemaker. Circulation. 1994;4(suppl I):I-70. Abstract.
Peters G, Locci R, Pulverer G. Adherence and growth of coagulase negative staphylococci on surfaces of intravenous catheters. J Infect Dis. 1982;146:4789-4793.
Peters G, Pulverer G. Microbiology of staphylococcal pacemaker lead infection. PACE Pacing Clin Electrophysiol. 1986;4:205-215.
Peters G, Saborowski F, Locci R. Investigations on staphylococcal infection on transvenous endocardial pacemaker electrodes. Am Heart J. 1984;108:358-365.