- Email: [email protected]
Undertreatment and Overtreatment of Patients With Infected Antiarrhythmic Implantable Devices
Undertreatment and Overtreatment of Patients With Infected Antiarrhythmic Implantable Devices J. Ernesto Molina, MD, PhD Division of Cardiovascular and Thoracic Surgery, University of Minnesota, Minneapolis, Minnesota
Background. Infection of implantable defibrillators or pacemakers is a serious complication, reported with increasing frequency probably because of an increase in the total number of devices implanted due to a change in trends in the treatment of arrhythmias. This review is aimed to provide guidelines on how to deal with these infections and which method is most likely to be successful. Methods. This is a review of 38 patients with infected antiarrhythmic implantable devices under three different plans of therapy. There were 17 implantable cardioverter defibrillators and 21 pacemakers. In 27, infection occurred after primary implantation (15 pacers, 12 implantable cardioverter defibrillators), and in 11 after replacement (six pacers, five implantable cardioverter defibrillators). Three therapeutic plans were identified. Group I (n 5 12) received intravenous antibiotics without removal of the antiarrhythmic implantable device, but with relocation to a different area or plane, and with or without the use of a topical irrigating-suction system. Group II (n 5 19) had complete removal of the system, 2 weeks of intravenous antibiotics, and implantation of a new unit followed by 10 more days of
antibiotics. Group III (n 5 7) underwent complete removal, 6 weeks of antibiotics, implantation of a new unit, and another 6 or more weeks of antibiotic therapy. Results. Failure occurred in 100% of cases in group I. Groups II and III had complete clearing of infection and successful reimplantation of new systems with no recurring infections. Follow-up was 8 months to 5 years. Two deaths occurred, both in group I. Hospitalization for groups I and III was 104 days and 65 days, respectively, versus 22 days for group II. No deaths occurred in group II or III. Conclusions. With an infected antiarrhythmic implantable device, immediate removal of the entire unit is recommended, followed by 2 weeks of intravenous antibiotics, implantation of a new system, and 10 more days of postoperative antibiotics. This regimen is sufficient to cure the problem. No attempts should be made to save an infected system from removal because it endangers the patient’s life, prolongs hospitalization, increases costs, and most likely will fail. (Ann Thorac Surg 1997;63:504 –9) © 1997 by The Society of Thoracic Surgeons
T
reports, continuous antibiotic irrigation of the pulse generator pocket for several days, or relocation, can save a few of these units [15–18]. However, once an infection is established, closed irrigation has never been the standard therapy [13, 14, 19]. Infection after implantation of these devices is usually not apparent immediately after operation. Symptoms can develop days or even weeks later. To provide some guidelines on how to deal with these infections, I reviewed patients with infected pacemakers or ICDs after either first implantation or replacement.
he incidence of infection after implantation of pacemakers and defibrillators is supposed to be low [1– 6]. Recently, however, infection has been reported more frequently and addressed as a clear potential problem. To enumerate some of the reasons, we must consider changing trends in the treatment of arrhythmias. First, pacemakers are used very commonly to treat cardiac arrhythmias, particularly bradycardia. Second, automatic implantable defibrillators (also called internal cardiac defibrillators, or ICDs) are recommended more frequently to treat ventricular tachycardia or fibrillation; for other arrhythmias, many cardiologists prefer them to pharmacologic therapy as a first-line approach [7–10]. The incidence of infections, therefore, will possibly increase, not only in percentages, but also in total numbers, because more devices are implanted more often. In general, infection of pacemakers is not as serious as infection of ICDs [11–15]. The pacemaker unit is smaller, and part of its leads often can be left behind without impairing clearance of the infection (although usually the pulse generator must be removed). According to several
Accepted for publication Sep 21, 1996. Address reprint requests to Dr Molina, University of Minnesota, Box 182, 420 Delaware St SE, Minneapolis, MN 55455.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Material and Methods Patient Population There were 38 patients with infections seen between 1985 and 1996, either with implantation done at my institution or referred to my institution from other centers: 17 with ICDs and 21 with pacemakers. In 27 of them, infection occurred after the primary implantation: 15 with pacemakers and 12 with ICDs. In the remaining 11 patients, infection occurred after replacement of the pulse generator, including 6 with pacemakers and 5 with ICDs. Patient age ranged from 5 months to 79 years. The mean age was 34 years for those with pacemakers versus 62 years with ICDs. 0003-4975/97/$17.00 PII S0003-4975(96)01033-8
Ann Thorac Surg 1997;63:504 –9
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
505
Table 1. Bacteriologic Diagnosis of 38 Patients With Infected Implantable Cardioverter Defibrillators or Pacemakers Gram Smear Bacterium Staphylococcus aureus Staphylococcus epidermidis Enterobacter Streptococcus Klebsiella Unknown Total a
Culture
Blood Culture
1
2
1
2
1
2
6 5 0 0 0 0
7 13 3 1 1 2
14 17 3 1 1 0
0 0 0 0 0 2
4 1 1 ... ... ...
(1a) ... (1a) ... ... ...
11
27
36
2
6
(2a)
Deaths.
Patients with ICDs had three combinations of electrodes: (1) only epicardial patches placed directly on the heart during thoracotomy or midline sternotomy (n 5 10); (2) a transvenous nonthoracotomy system implanted in combination with a subcutaneous or chest wall patch (n 5 4); and (3) only transvenous ICD systems implanted (n 5 3). All the transvenous systems were the MedtronicTransvene model (Medtronic, Inc, Minneapolis, MN). The epicardial patches were either Medtronic or CPI (Cardiac Pacemakers, Inc, St. Paul, MN).
Infection Site At the time of diagnosis, 30 patients had symptoms and signs of infection at the generator pocket site; another 5 had erosion with or without inflammation along the leads. For 2 patients, I could not determine what part of the system was infected, but both had fever and bacteremia; externally, there were no signs of inflammation or erosion over the lead or over the pulse generator system. For 1 patient, the sternotomy site was infected, presumably as the primary focus.
Timing Fourteen patients had a pattern of early infection— within 30 days after the original implantation or replacement (mean, 18 days). In this group, I found no evidence of erosion or displacement of the pulse generator or leads. I also noted a pattern of late infection—at least 30 days after the implantation or replacement—in 24 patients (range, 30 to 600 days; mean, 150 days). The late infections occurred secondary to erosion of the pulse generator (n 5 7), erosion of the leads (n 5 3), direct trauma (n 5 1), or primary pocket infection (n 5 9). Bacteremia with no identified focus occurred in 2 patients. In 2 other patients, the infection appeared later in the form of infected hematoma, where fluid had been left undrained at the pulse generator site.
Signs Elevations in body temperature (greater than 37.2°C) and in white blood cell count (greater than 7,000/mL) were considered positive signs of infection. At the time the infection was demonstrated, 13 patients had temperatures less than 37.2°C; the remaining 25 had tempera-
tures greater than 37.2°C. Similarly, 20 patients had white blood cell counts less than 7,000/mL, even though their infection was obvious. The remaining 18 patients had white blood cell counts greater than 7,000/mL. For all 38 patients studied, white blood cell counts varied from 2,500 to 28,600/mL (mean, 9,700/mL).
Bacteriologic Diagnosis The vast majority of infections were caused by Staphylococcus pathogens: S aureus (n 5 14), S epidermidis (n 5 17), Enterobacter (n 5 3), nonhemolytic Streptococcus (n 5 1), and Klebsiella (n 5 1). In 2 patients, no organisms grew in the purulent material, probably because of previous treatment with antibiotics. The three tests used to establish a bacteriologic diagnosis and the degree of severity were Gram stain of the purulent material, cultures from infected sites, and blood cultures. All 38 patients had these three tests done. Table 1 shows the number of positive and negative Gram smears. Only 11 patients had positive results; the majority (27 patients) had negative Gram smears. More reliably, cultures from the infected sites were positive in 36 patients and negative in only 2. Bacteremia was demonstrated in 6 patients: S aureus (n 5 4), S epidermidis (n 5 1), and Enterobacter (n 5 1). Despite all treatment efforts, 2 patients with bacteremia died.
Treatment Protocols Three treatment plans were implemented depending on the individual physician’s preference. In group I (n 5 12), patients were given intravenous antibiotics. An attempt was made to save the pacemaker or ICD by moving it to a different area, by cleaning the pocket, or by reclosing it after the Gram smear was reported negative. In some patients, a closed irrigating-suction catheter system was placed in the pocket in an attempt to clear the infection. In group II (n 5 19), the entire pacemaker or ICD was immediately removed, followed by 2 weeks of intravenous antibiotics before a new device was implanted, and 10 more days of antibiotics after the second implantation. In group III (n 5 7), the pacemaker or ICD was immediately removed, followed by 6 weeks of intravenous antibiotics before a new device was implanted; this was followed by 6 more weeks of antibiotics after the second implantation. The only difference between groups 2 and
506
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
3 was the length of treatment. It was important to determine whether a shorter treatment period was just as effective as the prolonged hospitalization required in protocol 3. All patients whose ICDs were removed remained hospitalized and were provided with R-2 pads on the chest wall connected to an external defibrillator until the second device was implanted. Two of these patients experienced tachyarrhythmia that required cardioversion. One patient’s ventricular tachycardia recurred twice, but was successfully cardioverted both times. The other patient had only one episode, also treated with electrical cardioversion using the same system. These patients also continued to receive antiarrhythmic drugs throughout the observation period until the second device was implanted. Of the patients with infected pacemakers, 5 were fully dependent on the system. Two were treated by exteriorizing the “infected” endocardial lead at the infected pocket site and connecting it to an external pacemaker pulse generator. They were maintained on this system until their infection cleared and a new pacemaker unit with new leads was implanted (at the same time as the old infected lead was removed). One patient, who had a unipolar system, had a subdermal electrode implanted under local anesthesia in the area of the abdomen to allow pacing with the external pacer. One child had a temporary epicardial lead placed in the ventricle after the infected transvenous lead was completely removed. This lead was removed when the new pacemaker was implanted. The fifth patient was given a temporary transvenous lead at the time that the entire infected pacemaker was removed. In the implementation of protocols 2 and 3, the new system was implanted in a different area from the original infected site. Half of the time it was in the ipsilateral site, but using a different route. All the patients with transvenous devices, except for 1, originally had these leads implanted in the subclavian vein. The other patient had the lead in the jugular vein. For an ipsilateral reimplantation of a new defibrillator, I used the internal jugular vein technique previously described [20], tunneled the lead down behind the clavicle, and implanted the ICD in the retropectoral area. For infected pacemakers, I chose to implant the new unit in the opposite side using the subclavian vein.
Results Morbidity and Mortality I looked at three factors in particular: (1) the need for reoperation after the initial maneuvers to clear the infection (for group I, this meant any operation after relocation, insertion of irrigating-suction systems, or reclosing of the pocket after local revision; for groups II and III, either complete or partial removal of the pacemaker or ICD), (2) the length of hospitalization, and (3) death due to the infection or related complications. All 12 group I patients required a reoperation to remove the entire pacemaker or ICD when the first approach failed. Their mean (6standard deviation)
Ann Thorac Surg 1997;63:504 –9
length of hospitalization was 104 6 13.9 days (range, 82 to 120 days). The two deaths in the entire series were in this group, in which there was reluctance to remove the entire device from the beginning. By the time the initial treatment failed, these 2 patients had already suffered complications from the infection, including multiorgan failure, renal shutdown, and septicemia. None of the 19 patients in group II required a reoperation. The mean length of hospitalization was 22 6 4 days (range, 17 to 31 days) for patients with ICDs and 7 days for patients with pacemakers, including the time after a new system was implanted in a different area. None of these patients died. Of the 7 group III patients, 2 required a reoperation. Both of them retained leads thought to be causing continuous infection. In 1, a leftover transvenous lead was removed during the second operation. In the other, only the pulse generator was removed during the initial operation; later the leads were removed as well. The mean length of hospitalization in this group was 65 6 4.4 days (range, 61 to 72 days). None of these patients died.
Septicemia On admission, 6 patients had positive blood cultures (3 in group I, 2 in group II, and 1 in group III). The 6 patients with bacteremia underwent a two-dimensional echocardiographic examination to rule out endocarditis. The examination was negative for vegetation in the cardiac valves or thrombus attached to the transvenous leads. In group I, 1 patient with an ICD and 1 with a pacemaker had S aureus cultured up to 24 and 48 hours after admission. Another patient with a defibrillator had Enterobacter cultured up to 24 hours after admission. Both of these ICD patients died, even though their blood culture had eventually become negative (one at 72 hours, the other at 48 hours). In 1 of these patients, the ICD in the subcutaneous space was removed and placed deeper under the muscular layer. Because the infection progressed, the entire ICD was later removed. Pulmonary infection developed, requiring reintubation and respiratory support. The infection never cleared, and multiorgan failure led to his death. In the other ICD patient who died, the ICD was not removed initially because the infection was thought to be in the sternal incision of a coronary bypass operation. Later, however, the ICD had to be removed because of a progressive infection, which eventually led to the patient’s death. In group II, 1 ICD patient had S aureus cultured up to 24 hours after admission, and a pacemaker patient had S epidermidis cultured up to 48 hours. After the cultures became negative, both patients eventually recovered. The first patient also had aortic and mitral valve prostheses implanted (before the ICD); because of the fear of endocarditis in these valves, a transesophageal echocardiogram was obtained several times, as well as a transthoracic examination. Results of both, however, were negative for endocarditis. This patient has remained clear of infection up to the present (3 years). The other patient with a pacemaker implant also had a negative twodimensional echocardiographic examination.
Ann Thorac Surg 1997;63:504 –9
In group III, 1 pacemaker patient had S aureus cultured on admission; the culture became negative at 24 hours. This patient recovered. Results of transthoracic twodimensional echocardiographic examination were negative.
Planned Treatment The difference between group I and the other two groups was striking: When an attempt was made to save the ICD or pacemaker, the failure rate was 100%. Eventually, all 12 had to have their devices removed, and only 10 achieved healing and total recovery; 2 died. In contrast, when the devices were totally removed from the beginning (groups II and III), the success rate was 100%. All 26 of the group II and group III patients had a new device implanted, and all recovered.
Antibiotics Two questions were especially important when treating these patients. First, after removal of the ICD or pacemaker, how long should antibiotics be given before it is safe to implant a new device? Second, after implantation of the new device, how long should antibiotics be given to prevent reinfection? For group II, I considered 2 weeks of antibiotics to be adequate before implanting a new device. After the reimplantation, I gave antibiotics for 10 days, then discontinued them totally. With this regimen, there were no cases of reinfection. Healing was total, with no further interventions required for up to 6 years (the follow-up period). For group III, antibiotics were given for 6 weeks before implanting a new device. After the reimplantation, antibiotics were continued for 36 days. Most of these patients continued to receive antibiotics for longer periods. There were no cases of reinfection. The follow-up period has ranged from 8 months to 10 years.
Follow-up Follow-up ranged from 8 months to 11 years in the entire group of 38 patients. Because the treatment applied to all patients was implemented contemporaneously, the range of follow-up is the same for all groups. However, the patients in group I who had failed treatment were eventually treated using therapeutic plan II or III. Therefore, their follow-up, observing strictly the results of treatment I, had a range of 3 months to 2 years. All of the patients who had treatment plan II or III are currently receiving no antibiotic therapy. The mean follow-up period for groups II and III is 6 6 2 years. The follow-up of group I also reflects the fact that during the past 2 years, no patient has been subjected to type I treatment because of its observed ineffectiveness.
White Blood Cell Count A count greater than 7,000/mL was considered abnormal. When antibiotics were given, the white blood cell count in all patients started to decrease within 24 hours. In groups II and III, in whom the entire ICD or pacemaker was removed, the white blood cell count decreased
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
507
slowly to normal over 48 hours. After that time, it did not rise again. In group I, however, in whom the ICD or pacemaker was left in, the white blood cell count remained above 7,000/mL, even after the initial 48 hours. It remained abnormal for up to 2 weeks until the decision was made to remove the device. Thus, the white blood cell count, when elevated, was a valid index of persistent infection, despite what was thought clinically to be adequate treatment.
Temperature The patients who had elevated temperatures on admission had a steady decrease over 48 hours while receiving antibiotics. Usually the temperature reached 36.7°C by 72 hours. Group I patients remained somewhat febrile at that time, but the difference was not statistically significant. Thus, temperature alone was not a reliable sign to assess the effectiveness of treatment.
Comment The incidence of infection after implantation of pacemakers has been considered low, according to the literature. It turns out to be quite substantial, however, when accurate and meticulous records are kept [6]. The infection risk is equal when the unit is first implanted versus when it is replaced. Treating pacemaker patients who have an infection is difficult and sometimes complicated. For ICD patients, however, infections are even more serious; if not treated properly and promptly, they can be fatal. Some surgeons and cardiologists rely heavily on the Gram smear results from the infected site to decide whether to remove the ICD. Unfortunately, the Gram smear is not a reliable test, as shown by this report. Because ICDs are much larger than pacemakers, they have larger pockets that are more difficult to clear. Furthermore, if epicardial patches are used, infection extends rather easily into the chest cavity and into the heart. Interventions to remove ICDs can be formidable. The salvage techniques described by Gupta and associates [11] and Lee and colleagues [16] apply only when the infection at the pulse generator site has not extended along the leads. Most of the ICD patients in my study had infections of the leads as well. Even when an infected ICD may look salvageable with a muscle flap [17] or with relocation [18], such approaches are less practical than removing the entire device and implanting a new one later [14], particularly now that transvenous systems are available. Anecdotal reports have described salvage of pacemakers and ICDs with a closed irrigation-suction system [15, 16]. However, this is a prolonged treatment and carries no guarantee of working; in fact, it fails more often than it succeeds. Hospitalization is lengthy and costs are considerable. Infection seems to be only dormant, and the patient is never truly cured. In general, the intravenous systems can be removed relatively easily. However, patients who have undergone patch implantation on the heart need to be evaluated more carefully; removal of these patches is a much more serious intervention.
508
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
Fig 1. Computed tomographic scan of the chest in a 73-year-old man from group II showing CPI epicardial patches (arrows), one patch anterior and one posterolateral. The patient had an infection of the pulse generator site, but there was no evidence of any fluid accumulation or infection at the site of the patches. The patches were not removed in this case, but part of the leads and the pulse generator itself were. The patient is now cured of any infection and is not taking antibiotics.
In this study, all patients with epicardial patches underwent a computed tomographic scan. This accurately detected any fluid collection or thickening around the patches (Fig 1). Usually, patients with infected patches were much sicker clinically than those with only erosions and contaminated pocket sites. Of the 14 patients with patch systems, 2 had no evidence in their patches of either inflammation or thickening according to their computed tomographic scan. Therefore, in these 2 patients, an incision was made first at the costal margin, before any other manipulations. The leads were divided at that level; then, exerting traction on the cardiac end of the lead, this was divided as high as possible. This incision and the tunnel tract were closed, and the incision was irrigated with antibiotics. The pocket site was opened and the generator with the lead remnants was removed. These 2 patients have remained clear of infection up to the present (5 and 1.5 years, respectively, after a new system was implanted). In the other 12 patients with patch systems, the patches either were crinkled with thickening or showed fluid collection pockets. All of these patches were removed primarily. One helpful confirmatory test is the gallium scan because when negative, it indicates the absence of infection; however, is not always necessary. Figure 1 shows the computed tomographic image for a patient who did not have the patches removed and also had a negative gallium scan. If the gallium scan is obtained, it should be done before any other local intervention, because areas of recent operation will appear falsely positive up to 10 days postoperatively. Based on the present review, I recommend that if any antiarrhythmic implantable device becomes infected, all of it should be removed and antibiotic therapy should be started. With previous epicardial patch implants, we
Ann Thorac Surg 1997;63:504 –9
recommend a computed tomographic scan and possibly a gallium scan to rule out infection in those areas before operative treatment is undertaken. The present study shows that after incomplete removal of a pacemaker system, part of the transvenous lead in continuity with the pacemaker system can be left in the patient’s body without causing the infection to continue. In other words, a piece of transvenous pacemaker lead is probably easier to sterilize with antibiotics once the pulse generator is removed. In fact, in 8 of my patients, the leads were only partially removed, but the infection was completely cleared nonetheless. In 3 of my ICD patients, the leads were only partially removed; 2 patients had a patch system and 1 had only transvenous leads. Very rarely, if at all, is cardiopulmonary bypass needed to remove transvenous pacemaker leads adhering in the right ventricle. The use of lead extractors has been very successful [21, 22]. I therefore recommend immediate removal of the pulse generator, with the leads if possible, followed by intravenous antibiotics; after 2 weeks, a new generator and leads can be implanted safely. I do not recommend attempting to salvage infected ICDs or pacemakers. Total immediate removal is safest. After the second implantation, 10 days of antibiotics is sufficient to prevent reinfection; longer treatment does not seem to be justified.
References 1. Winkle RA, Mead RH, Ruder MA, et al. Long-term outcome with the automatic implantable cardioverter-defibrillator. J Am Coll Cardiol 1993;13:1353– 61. 2. Wunderly D, Maloney J, Edel T, et al. Infections in implantable cardioverter defibrillator patients. PACE 1990;13(Part 1):1360– 4. 3. Bluhm G, Nordlander R, Ransjo U. Antibiotic prophylaxis in pacemaker surgery: a prospective double blind trial with systemic administration of antibiotics versus placebo at implantation of cardiac pacemakers. PACE 1986;9:720– 6. 4. Shahian DM, Williamson WA, Martin D, et al. Infection of implantable cardioverter defibrillator systems: a preventable complication? PACE 1993;16:1956– 60. 5. Parsonnet V, Bernstein AD, Neglia D. Non-thoracotomy ICD implantation: lessons to be learned from permanent pacemaker implantation. PACE 1995;18(Part 1):1597– 600. 6. Grimm W, Flores BF, Marchlinski FE. Complications of implantable cardioverter defibrillator therapy: follow-up of 241 patients. PACE 1993;16(Part 2):218–22. 7. Sperry RE, Stambler BS, Wood MA, et al. Reduction of antiarrhythmic drug use in patients receiving implantable cardioverter defibrillators. PACE 1996;19:61– 4. 8. Williamson BD, Man KC, Niebauer M, et al. The economic impact of transvenous defibrillation lead systems. PACE 1994;17:2297–303. 9. Zipes DP, Roberts D. Result of the international study of the implantable pacemaker cardioverter-defibrillator. A comparison of epicardial and endocardial lead systems. Circulation 1995;92:59– 65. 10. Kelly PA, Cannom DS, Garan H, et al. The automatic implantable cardioverter-defibrillator: efficacy, complications and survival in patients with malignant ventricular arrhythmias. J Am Coll Cardiol 1988;11:1278– 86. 11. Gupta S, Prevel CD, Shaheen K, et al. Wound complications and treatment of the infected implantable cardioverter defibrillator generator. J Cardiac Surg 1993;8:671–7. 12. Perez I, Nunain SO, Osswald S, et al. Management of patients with infected implantable cardioverter defibrillator systems. PACE 1994;17(Part 2):780.
Ann Thorac Surg 1997;63:504 –9
13. Gro¨gler FM, Frank G, Greven G, et al. Complications of permanent transvenous cardiac pacing. J Thorac Cardiovasc Surg 1975;69:895–904. 14. Choo MH, Holmes DR, Gersh BJ, et al. Permanent pacemaker infections: characterization and management. Am J Cardiol 1981;48:259– 64. 15. Hurst LN, Evans HB, Windle B, et al. The salvage of infected cardiac pacemaker pockets using a closed irrigation system. PACE 1986;9:785–92. 16. Lee JH, Geha S, Rattehalli NM, et al. Salvage of infected ICDs: management without removal. PACE 1996;19:437– 42. 17. Taylor RL, Cohen DJ, Widman LE, et al. Infection of an implantable cardioverter defibrillator: management without removal of the device in selected cases. PACE 1990;13(Part 1):1352–5.
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
509
18. Garcia-Rinaldi R, Revuelta JM, Bonington JM, et al. The exposed cardiac pacemaker. Treatment by subfascial pocket relocation. J Thorac Cardiovasc Surg 1985;81:136– 41. 19. Olinger GN, Chapman PD, Troup PJ, et al. Stratified application of the automatic implantable cardioverter defibrillator. J Thorac Cardiovasc Surg 1988;96:141–9. 20. Molina JE, Dunnigan AC, Crosson JE. Implantation of transvenous pacemakers in infants and small children. Ann Thorac Surg 1995;59:689–94. 21. Colavita PG, Zimmern SH, Gallagher JJ, et al. Intravascular extraction of chronic pacemaker leads: efficacy and followup. PACE 1993;16:2333– 6. 22. Espinosa RE, Hayes DL, Vlietstra RE, et al. The dotter retriever and pigtail catheter: efficacy in extraction of chronic transvenous pacemaker leads. PACE 1993;16:2337– 48.
Bound volumes available to subscribers Bound volumes of the 1996 issues of The Annals of Thoracic Surgery are available only to subscribers from the Publisher. The cost is $105.00 (outside US add $27.00 for postage) for volumes 61 and 62. Each bound volume contains a subject and author index, and all advertising is removed. The binding is durable buckram with the name of the journal, volume number, and year stamped on the spine. Payment must accompany all orders. Contact Elsevier Science Inc, 655 Avenue of the Americas, New York, NY 10010; or telephone (212) 633-3950 (facsimile: (212) 633-3990).
Background. Infection of implantable defibrillators or pacemakers is a serious complication, reported with increasing frequency probably because of an increase in the total number of devices implanted due to a change in trends in the treatment of arrhythmias. This review is aimed to provide guidelines on how to deal with these infections and which method is most likely to be successful. Methods. This is a review of 38 patients with infected antiarrhythmic implantable devices under three different plans of therapy. There were 17 implantable cardioverter defibrillators and 21 pacemakers. In 27, infection occurred after primary implantation (15 pacers, 12 implantable cardioverter defibrillators), and in 11 after replacement (six pacers, five implantable cardioverter defibrillators). Three therapeutic plans were identified. Group I (n 5 12) received intravenous antibiotics without removal of the antiarrhythmic implantable device, but with relocation to a different area or plane, and with or without the use of a topical irrigating-suction system. Group II (n 5 19) had complete removal of the system, 2 weeks of intravenous antibiotics, and implantation of a new unit followed by 10 more days of
antibiotics. Group III (n 5 7) underwent complete removal, 6 weeks of antibiotics, implantation of a new unit, and another 6 or more weeks of antibiotic therapy. Results. Failure occurred in 100% of cases in group I. Groups II and III had complete clearing of infection and successful reimplantation of new systems with no recurring infections. Follow-up was 8 months to 5 years. Two deaths occurred, both in group I. Hospitalization for groups I and III was 104 days and 65 days, respectively, versus 22 days for group II. No deaths occurred in group II or III. Conclusions. With an infected antiarrhythmic implantable device, immediate removal of the entire unit is recommended, followed by 2 weeks of intravenous antibiotics, implantation of a new system, and 10 more days of postoperative antibiotics. This regimen is sufficient to cure the problem. No attempts should be made to save an infected system from removal because it endangers the patient’s life, prolongs hospitalization, increases costs, and most likely will fail. (Ann Thorac Surg 1997;63:504 –9) © 1997 by The Society of Thoracic Surgeons
T
reports, continuous antibiotic irrigation of the pulse generator pocket for several days, or relocation, can save a few of these units [15–18]. However, once an infection is established, closed irrigation has never been the standard therapy [13, 14, 19]. Infection after implantation of these devices is usually not apparent immediately after operation. Symptoms can develop days or even weeks later. To provide some guidelines on how to deal with these infections, I reviewed patients with infected pacemakers or ICDs after either first implantation or replacement.
he incidence of infection after implantation of pacemakers and defibrillators is supposed to be low [1– 6]. Recently, however, infection has been reported more frequently and addressed as a clear potential problem. To enumerate some of the reasons, we must consider changing trends in the treatment of arrhythmias. First, pacemakers are used very commonly to treat cardiac arrhythmias, particularly bradycardia. Second, automatic implantable defibrillators (also called internal cardiac defibrillators, or ICDs) are recommended more frequently to treat ventricular tachycardia or fibrillation; for other arrhythmias, many cardiologists prefer them to pharmacologic therapy as a first-line approach [7–10]. The incidence of infections, therefore, will possibly increase, not only in percentages, but also in total numbers, because more devices are implanted more often. In general, infection of pacemakers is not as serious as infection of ICDs [11–15]. The pacemaker unit is smaller, and part of its leads often can be left behind without impairing clearance of the infection (although usually the pulse generator must be removed). According to several
Accepted for publication Sep 21, 1996. Address reprint requests to Dr Molina, University of Minnesota, Box 182, 420 Delaware St SE, Minneapolis, MN 55455.
© 1997 by The Society of Thoracic Surgeons Published by Elsevier Science Inc
Material and Methods Patient Population There were 38 patients with infections seen between 1985 and 1996, either with implantation done at my institution or referred to my institution from other centers: 17 with ICDs and 21 with pacemakers. In 27 of them, infection occurred after the primary implantation: 15 with pacemakers and 12 with ICDs. In the remaining 11 patients, infection occurred after replacement of the pulse generator, including 6 with pacemakers and 5 with ICDs. Patient age ranged from 5 months to 79 years. The mean age was 34 years for those with pacemakers versus 62 years with ICDs. 0003-4975/97/$17.00 PII S0003-4975(96)01033-8
Ann Thorac Surg 1997;63:504 –9
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
505
Table 1. Bacteriologic Diagnosis of 38 Patients With Infected Implantable Cardioverter Defibrillators or Pacemakers Gram Smear Bacterium Staphylococcus aureus Staphylococcus epidermidis Enterobacter Streptococcus Klebsiella Unknown Total a
Culture
Blood Culture
1
2
1
2
1
2
6 5 0 0 0 0
7 13 3 1 1 2
14 17 3 1 1 0
0 0 0 0 0 2
4 1 1 ... ... ...
(1a) ... (1a) ... ... ...
11
27
36
2
6
(2a)
Deaths.
Patients with ICDs had three combinations of electrodes: (1) only epicardial patches placed directly on the heart during thoracotomy or midline sternotomy (n 5 10); (2) a transvenous nonthoracotomy system implanted in combination with a subcutaneous or chest wall patch (n 5 4); and (3) only transvenous ICD systems implanted (n 5 3). All the transvenous systems were the MedtronicTransvene model (Medtronic, Inc, Minneapolis, MN). The epicardial patches were either Medtronic or CPI (Cardiac Pacemakers, Inc, St. Paul, MN).
Infection Site At the time of diagnosis, 30 patients had symptoms and signs of infection at the generator pocket site; another 5 had erosion with or without inflammation along the leads. For 2 patients, I could not determine what part of the system was infected, but both had fever and bacteremia; externally, there were no signs of inflammation or erosion over the lead or over the pulse generator system. For 1 patient, the sternotomy site was infected, presumably as the primary focus.
Timing Fourteen patients had a pattern of early infection— within 30 days after the original implantation or replacement (mean, 18 days). In this group, I found no evidence of erosion or displacement of the pulse generator or leads. I also noted a pattern of late infection—at least 30 days after the implantation or replacement—in 24 patients (range, 30 to 600 days; mean, 150 days). The late infections occurred secondary to erosion of the pulse generator (n 5 7), erosion of the leads (n 5 3), direct trauma (n 5 1), or primary pocket infection (n 5 9). Bacteremia with no identified focus occurred in 2 patients. In 2 other patients, the infection appeared later in the form of infected hematoma, where fluid had been left undrained at the pulse generator site.
Signs Elevations in body temperature (greater than 37.2°C) and in white blood cell count (greater than 7,000/mL) were considered positive signs of infection. At the time the infection was demonstrated, 13 patients had temperatures less than 37.2°C; the remaining 25 had tempera-
tures greater than 37.2°C. Similarly, 20 patients had white blood cell counts less than 7,000/mL, even though their infection was obvious. The remaining 18 patients had white blood cell counts greater than 7,000/mL. For all 38 patients studied, white blood cell counts varied from 2,500 to 28,600/mL (mean, 9,700/mL).
Bacteriologic Diagnosis The vast majority of infections were caused by Staphylococcus pathogens: S aureus (n 5 14), S epidermidis (n 5 17), Enterobacter (n 5 3), nonhemolytic Streptococcus (n 5 1), and Klebsiella (n 5 1). In 2 patients, no organisms grew in the purulent material, probably because of previous treatment with antibiotics. The three tests used to establish a bacteriologic diagnosis and the degree of severity were Gram stain of the purulent material, cultures from infected sites, and blood cultures. All 38 patients had these three tests done. Table 1 shows the number of positive and negative Gram smears. Only 11 patients had positive results; the majority (27 patients) had negative Gram smears. More reliably, cultures from the infected sites were positive in 36 patients and negative in only 2. Bacteremia was demonstrated in 6 patients: S aureus (n 5 4), S epidermidis (n 5 1), and Enterobacter (n 5 1). Despite all treatment efforts, 2 patients with bacteremia died.
Treatment Protocols Three treatment plans were implemented depending on the individual physician’s preference. In group I (n 5 12), patients were given intravenous antibiotics. An attempt was made to save the pacemaker or ICD by moving it to a different area, by cleaning the pocket, or by reclosing it after the Gram smear was reported negative. In some patients, a closed irrigating-suction catheter system was placed in the pocket in an attempt to clear the infection. In group II (n 5 19), the entire pacemaker or ICD was immediately removed, followed by 2 weeks of intravenous antibiotics before a new device was implanted, and 10 more days of antibiotics after the second implantation. In group III (n 5 7), the pacemaker or ICD was immediately removed, followed by 6 weeks of intravenous antibiotics before a new device was implanted; this was followed by 6 more weeks of antibiotics after the second implantation. The only difference between groups 2 and
506
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
3 was the length of treatment. It was important to determine whether a shorter treatment period was just as effective as the prolonged hospitalization required in protocol 3. All patients whose ICDs were removed remained hospitalized and were provided with R-2 pads on the chest wall connected to an external defibrillator until the second device was implanted. Two of these patients experienced tachyarrhythmia that required cardioversion. One patient’s ventricular tachycardia recurred twice, but was successfully cardioverted both times. The other patient had only one episode, also treated with electrical cardioversion using the same system. These patients also continued to receive antiarrhythmic drugs throughout the observation period until the second device was implanted. Of the patients with infected pacemakers, 5 were fully dependent on the system. Two were treated by exteriorizing the “infected” endocardial lead at the infected pocket site and connecting it to an external pacemaker pulse generator. They were maintained on this system until their infection cleared and a new pacemaker unit with new leads was implanted (at the same time as the old infected lead was removed). One patient, who had a unipolar system, had a subdermal electrode implanted under local anesthesia in the area of the abdomen to allow pacing with the external pacer. One child had a temporary epicardial lead placed in the ventricle after the infected transvenous lead was completely removed. This lead was removed when the new pacemaker was implanted. The fifth patient was given a temporary transvenous lead at the time that the entire infected pacemaker was removed. In the implementation of protocols 2 and 3, the new system was implanted in a different area from the original infected site. Half of the time it was in the ipsilateral site, but using a different route. All the patients with transvenous devices, except for 1, originally had these leads implanted in the subclavian vein. The other patient had the lead in the jugular vein. For an ipsilateral reimplantation of a new defibrillator, I used the internal jugular vein technique previously described [20], tunneled the lead down behind the clavicle, and implanted the ICD in the retropectoral area. For infected pacemakers, I chose to implant the new unit in the opposite side using the subclavian vein.
Results Morbidity and Mortality I looked at three factors in particular: (1) the need for reoperation after the initial maneuvers to clear the infection (for group I, this meant any operation after relocation, insertion of irrigating-suction systems, or reclosing of the pocket after local revision; for groups II and III, either complete or partial removal of the pacemaker or ICD), (2) the length of hospitalization, and (3) death due to the infection or related complications. All 12 group I patients required a reoperation to remove the entire pacemaker or ICD when the first approach failed. Their mean (6standard deviation)
Ann Thorac Surg 1997;63:504 –9
length of hospitalization was 104 6 13.9 days (range, 82 to 120 days). The two deaths in the entire series were in this group, in which there was reluctance to remove the entire device from the beginning. By the time the initial treatment failed, these 2 patients had already suffered complications from the infection, including multiorgan failure, renal shutdown, and septicemia. None of the 19 patients in group II required a reoperation. The mean length of hospitalization was 22 6 4 days (range, 17 to 31 days) for patients with ICDs and 7 days for patients with pacemakers, including the time after a new system was implanted in a different area. None of these patients died. Of the 7 group III patients, 2 required a reoperation. Both of them retained leads thought to be causing continuous infection. In 1, a leftover transvenous lead was removed during the second operation. In the other, only the pulse generator was removed during the initial operation; later the leads were removed as well. The mean length of hospitalization in this group was 65 6 4.4 days (range, 61 to 72 days). None of these patients died.
Septicemia On admission, 6 patients had positive blood cultures (3 in group I, 2 in group II, and 1 in group III). The 6 patients with bacteremia underwent a two-dimensional echocardiographic examination to rule out endocarditis. The examination was negative for vegetation in the cardiac valves or thrombus attached to the transvenous leads. In group I, 1 patient with an ICD and 1 with a pacemaker had S aureus cultured up to 24 and 48 hours after admission. Another patient with a defibrillator had Enterobacter cultured up to 24 hours after admission. Both of these ICD patients died, even though their blood culture had eventually become negative (one at 72 hours, the other at 48 hours). In 1 of these patients, the ICD in the subcutaneous space was removed and placed deeper under the muscular layer. Because the infection progressed, the entire ICD was later removed. Pulmonary infection developed, requiring reintubation and respiratory support. The infection never cleared, and multiorgan failure led to his death. In the other ICD patient who died, the ICD was not removed initially because the infection was thought to be in the sternal incision of a coronary bypass operation. Later, however, the ICD had to be removed because of a progressive infection, which eventually led to the patient’s death. In group II, 1 ICD patient had S aureus cultured up to 24 hours after admission, and a pacemaker patient had S epidermidis cultured up to 48 hours. After the cultures became negative, both patients eventually recovered. The first patient also had aortic and mitral valve prostheses implanted (before the ICD); because of the fear of endocarditis in these valves, a transesophageal echocardiogram was obtained several times, as well as a transthoracic examination. Results of both, however, were negative for endocarditis. This patient has remained clear of infection up to the present (3 years). The other patient with a pacemaker implant also had a negative twodimensional echocardiographic examination.
Ann Thorac Surg 1997;63:504 –9
In group III, 1 pacemaker patient had S aureus cultured on admission; the culture became negative at 24 hours. This patient recovered. Results of transthoracic twodimensional echocardiographic examination were negative.
Planned Treatment The difference between group I and the other two groups was striking: When an attempt was made to save the ICD or pacemaker, the failure rate was 100%. Eventually, all 12 had to have their devices removed, and only 10 achieved healing and total recovery; 2 died. In contrast, when the devices were totally removed from the beginning (groups II and III), the success rate was 100%. All 26 of the group II and group III patients had a new device implanted, and all recovered.
Antibiotics Two questions were especially important when treating these patients. First, after removal of the ICD or pacemaker, how long should antibiotics be given before it is safe to implant a new device? Second, after implantation of the new device, how long should antibiotics be given to prevent reinfection? For group II, I considered 2 weeks of antibiotics to be adequate before implanting a new device. After the reimplantation, I gave antibiotics for 10 days, then discontinued them totally. With this regimen, there were no cases of reinfection. Healing was total, with no further interventions required for up to 6 years (the follow-up period). For group III, antibiotics were given for 6 weeks before implanting a new device. After the reimplantation, antibiotics were continued for 36 days. Most of these patients continued to receive antibiotics for longer periods. There were no cases of reinfection. The follow-up period has ranged from 8 months to 10 years.
Follow-up Follow-up ranged from 8 months to 11 years in the entire group of 38 patients. Because the treatment applied to all patients was implemented contemporaneously, the range of follow-up is the same for all groups. However, the patients in group I who had failed treatment were eventually treated using therapeutic plan II or III. Therefore, their follow-up, observing strictly the results of treatment I, had a range of 3 months to 2 years. All of the patients who had treatment plan II or III are currently receiving no antibiotic therapy. The mean follow-up period for groups II and III is 6 6 2 years. The follow-up of group I also reflects the fact that during the past 2 years, no patient has been subjected to type I treatment because of its observed ineffectiveness.
White Blood Cell Count A count greater than 7,000/mL was considered abnormal. When antibiotics were given, the white blood cell count in all patients started to decrease within 24 hours. In groups II and III, in whom the entire ICD or pacemaker was removed, the white blood cell count decreased
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
507
slowly to normal over 48 hours. After that time, it did not rise again. In group I, however, in whom the ICD or pacemaker was left in, the white blood cell count remained above 7,000/mL, even after the initial 48 hours. It remained abnormal for up to 2 weeks until the decision was made to remove the device. Thus, the white blood cell count, when elevated, was a valid index of persistent infection, despite what was thought clinically to be adequate treatment.
Temperature The patients who had elevated temperatures on admission had a steady decrease over 48 hours while receiving antibiotics. Usually the temperature reached 36.7°C by 72 hours. Group I patients remained somewhat febrile at that time, but the difference was not statistically significant. Thus, temperature alone was not a reliable sign to assess the effectiveness of treatment.
Comment The incidence of infection after implantation of pacemakers has been considered low, according to the literature. It turns out to be quite substantial, however, when accurate and meticulous records are kept [6]. The infection risk is equal when the unit is first implanted versus when it is replaced. Treating pacemaker patients who have an infection is difficult and sometimes complicated. For ICD patients, however, infections are even more serious; if not treated properly and promptly, they can be fatal. Some surgeons and cardiologists rely heavily on the Gram smear results from the infected site to decide whether to remove the ICD. Unfortunately, the Gram smear is not a reliable test, as shown by this report. Because ICDs are much larger than pacemakers, they have larger pockets that are more difficult to clear. Furthermore, if epicardial patches are used, infection extends rather easily into the chest cavity and into the heart. Interventions to remove ICDs can be formidable. The salvage techniques described by Gupta and associates [11] and Lee and colleagues [16] apply only when the infection at the pulse generator site has not extended along the leads. Most of the ICD patients in my study had infections of the leads as well. Even when an infected ICD may look salvageable with a muscle flap [17] or with relocation [18], such approaches are less practical than removing the entire device and implanting a new one later [14], particularly now that transvenous systems are available. Anecdotal reports have described salvage of pacemakers and ICDs with a closed irrigation-suction system [15, 16]. However, this is a prolonged treatment and carries no guarantee of working; in fact, it fails more often than it succeeds. Hospitalization is lengthy and costs are considerable. Infection seems to be only dormant, and the patient is never truly cured. In general, the intravenous systems can be removed relatively easily. However, patients who have undergone patch implantation on the heart need to be evaluated more carefully; removal of these patches is a much more serious intervention.
508
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
Fig 1. Computed tomographic scan of the chest in a 73-year-old man from group II showing CPI epicardial patches (arrows), one patch anterior and one posterolateral. The patient had an infection of the pulse generator site, but there was no evidence of any fluid accumulation or infection at the site of the patches. The patches were not removed in this case, but part of the leads and the pulse generator itself were. The patient is now cured of any infection and is not taking antibiotics.
In this study, all patients with epicardial patches underwent a computed tomographic scan. This accurately detected any fluid collection or thickening around the patches (Fig 1). Usually, patients with infected patches were much sicker clinically than those with only erosions and contaminated pocket sites. Of the 14 patients with patch systems, 2 had no evidence in their patches of either inflammation or thickening according to their computed tomographic scan. Therefore, in these 2 patients, an incision was made first at the costal margin, before any other manipulations. The leads were divided at that level; then, exerting traction on the cardiac end of the lead, this was divided as high as possible. This incision and the tunnel tract were closed, and the incision was irrigated with antibiotics. The pocket site was opened and the generator with the lead remnants was removed. These 2 patients have remained clear of infection up to the present (5 and 1.5 years, respectively, after a new system was implanted). In the other 12 patients with patch systems, the patches either were crinkled with thickening or showed fluid collection pockets. All of these patches were removed primarily. One helpful confirmatory test is the gallium scan because when negative, it indicates the absence of infection; however, is not always necessary. Figure 1 shows the computed tomographic image for a patient who did not have the patches removed and also had a negative gallium scan. If the gallium scan is obtained, it should be done before any other local intervention, because areas of recent operation will appear falsely positive up to 10 days postoperatively. Based on the present review, I recommend that if any antiarrhythmic implantable device becomes infected, all of it should be removed and antibiotic therapy should be started. With previous epicardial patch implants, we
Ann Thorac Surg 1997;63:504 –9
recommend a computed tomographic scan and possibly a gallium scan to rule out infection in those areas before operative treatment is undertaken. The present study shows that after incomplete removal of a pacemaker system, part of the transvenous lead in continuity with the pacemaker system can be left in the patient’s body without causing the infection to continue. In other words, a piece of transvenous pacemaker lead is probably easier to sterilize with antibiotics once the pulse generator is removed. In fact, in 8 of my patients, the leads were only partially removed, but the infection was completely cleared nonetheless. In 3 of my ICD patients, the leads were only partially removed; 2 patients had a patch system and 1 had only transvenous leads. Very rarely, if at all, is cardiopulmonary bypass needed to remove transvenous pacemaker leads adhering in the right ventricle. The use of lead extractors has been very successful [21, 22]. I therefore recommend immediate removal of the pulse generator, with the leads if possible, followed by intravenous antibiotics; after 2 weeks, a new generator and leads can be implanted safely. I do not recommend attempting to salvage infected ICDs or pacemakers. Total immediate removal is safest. After the second implantation, 10 days of antibiotics is sufficient to prevent reinfection; longer treatment does not seem to be justified.
References 1. Winkle RA, Mead RH, Ruder MA, et al. Long-term outcome with the automatic implantable cardioverter-defibrillator. J Am Coll Cardiol 1993;13:1353– 61. 2. Wunderly D, Maloney J, Edel T, et al. Infections in implantable cardioverter defibrillator patients. PACE 1990;13(Part 1):1360– 4. 3. Bluhm G, Nordlander R, Ransjo U. Antibiotic prophylaxis in pacemaker surgery: a prospective double blind trial with systemic administration of antibiotics versus placebo at implantation of cardiac pacemakers. PACE 1986;9:720– 6. 4. Shahian DM, Williamson WA, Martin D, et al. Infection of implantable cardioverter defibrillator systems: a preventable complication? PACE 1993;16:1956– 60. 5. Parsonnet V, Bernstein AD, Neglia D. Non-thoracotomy ICD implantation: lessons to be learned from permanent pacemaker implantation. PACE 1995;18(Part 1):1597– 600. 6. Grimm W, Flores BF, Marchlinski FE. Complications of implantable cardioverter defibrillator therapy: follow-up of 241 patients. PACE 1993;16(Part 2):218–22. 7. Sperry RE, Stambler BS, Wood MA, et al. Reduction of antiarrhythmic drug use in patients receiving implantable cardioverter defibrillators. PACE 1996;19:61– 4. 8. Williamson BD, Man KC, Niebauer M, et al. The economic impact of transvenous defibrillation lead systems. PACE 1994;17:2297–303. 9. Zipes DP, Roberts D. Result of the international study of the implantable pacemaker cardioverter-defibrillator. A comparison of epicardial and endocardial lead systems. Circulation 1995;92:59– 65. 10. Kelly PA, Cannom DS, Garan H, et al. The automatic implantable cardioverter-defibrillator: efficacy, complications and survival in patients with malignant ventricular arrhythmias. J Am Coll Cardiol 1988;11:1278– 86. 11. Gupta S, Prevel CD, Shaheen K, et al. Wound complications and treatment of the infected implantable cardioverter defibrillator generator. J Cardiac Surg 1993;8:671–7. 12. Perez I, Nunain SO, Osswald S, et al. Management of patients with infected implantable cardioverter defibrillator systems. PACE 1994;17(Part 2):780.
Ann Thorac Surg 1997;63:504 –9
13. Gro¨gler FM, Frank G, Greven G, et al. Complications of permanent transvenous cardiac pacing. J Thorac Cardiovasc Surg 1975;69:895–904. 14. Choo MH, Holmes DR, Gersh BJ, et al. Permanent pacemaker infections: characterization and management. Am J Cardiol 1981;48:259– 64. 15. Hurst LN, Evans HB, Windle B, et al. The salvage of infected cardiac pacemaker pockets using a closed irrigation system. PACE 1986;9:785–92. 16. Lee JH, Geha S, Rattehalli NM, et al. Salvage of infected ICDs: management without removal. PACE 1996;19:437– 42. 17. Taylor RL, Cohen DJ, Widman LE, et al. Infection of an implantable cardioverter defibrillator: management without removal of the device in selected cases. PACE 1990;13(Part 1):1352–5.
MOLINA INFECTED ANTIARRHYTHMIC DEVICES
509
18. Garcia-Rinaldi R, Revuelta JM, Bonington JM, et al. The exposed cardiac pacemaker. Treatment by subfascial pocket relocation. J Thorac Cardiovasc Surg 1985;81:136– 41. 19. Olinger GN, Chapman PD, Troup PJ, et al. Stratified application of the automatic implantable cardioverter defibrillator. J Thorac Cardiovasc Surg 1988;96:141–9. 20. Molina JE, Dunnigan AC, Crosson JE. Implantation of transvenous pacemakers in infants and small children. Ann Thorac Surg 1995;59:689–94. 21. Colavita PG, Zimmern SH, Gallagher JJ, et al. Intravascular extraction of chronic pacemaker leads: efficacy and followup. PACE 1993;16:2333– 6. 22. Espinosa RE, Hayes DL, Vlietstra RE, et al. The dotter retriever and pigtail catheter: efficacy in extraction of chronic transvenous pacemaker leads. PACE 1993;16:2337– 48.
Bound volumes available to subscribers Bound volumes of the 1996 issues of The Annals of Thoracic Surgery are available only to subscribers from the Publisher. The cost is $105.00 (outside US add $27.00 for postage) for volumes 61 and 62. Each bound volume contains a subject and author index, and all advertising is removed. The binding is durable buckram with the name of the journal, volume number, and year stamped on the spine. Payment must accompany all orders. Contact Elsevier Science Inc, 655 Avenue of the Americas, New York, NY 10010; or telephone (212) 633-3950 (facsimile: (212) 633-3990).