THoRAc CARDIOVASC SURG
Delayed infection of the automatic implantable cardioverter-defibrillator Current recognition and management Three cases of delayed infection of automatic implantable cardioverter-defibrillator devices without systemic manifestations are reported. Computed tomographic scan of the heart revealed fluid deep to the patch in each case. Sonication of explanted automatic implantable cardioverter-defibrillator patches facilitated the recovery of adherent microorganisms in one case. Management of this previously unrecognized problem is outlined.
G. Hossein Almassi, MD, Gordon N. Olinger, MD, Paul J. Troup, MD, Peter D. Chapman, MD, and Lawrence R. Goodman, MD, Milwaukee. Wis.
h e use of the automatic implantable cardioverterdefibrillator (AICD) for the management of medically refractory malignant ventricular arrhythmias is an established therapeutic modality. These devices, like other foreign bodies, are prone to infection with a variety of microorganisms, but delayed infection of the hardware without concomitant infection elsewhere or other predisposing factors has not previously been recognized. We report our experience with three such patients in whom de novo infection of the devices developed at intervals of 7 to 31 months after implantation.
Case report CASE I. On a routine checkup, 18 months after a successful myocardial revascularization and implantation of AICD patches and a generator, a 70-year-old man was found to have induration of the skin over the lower corner of the AICD generator pocket without actual erosion. He was afebrile and had no other physical findings of infection. He had a white blood cell count of 7600/mm 3 with normal differential, and the chest roentgenogram showed no pleural effusion or cardiomegaly. A computed tomographic (CT) scan of the heart revealed fluid densities between the AICD patches and
From the Department of Cardiothoracic Surgery. Division of Cardiology, and Department of Diagnostic Radiology. The Medical College of Wisconsin, Milwaukee, Wis. Received for publication April 30, 1987. Accepted for publication June I, 1987. Address for reprints: G. Hossein Alrnassi, MD. Medical College of Wisconsin, Department of Cardiothoracic Surgery. 8700 W. Wisconsin Ave.. Milwaukee. WI 53226.
the heart. Gram stain of a specimen obtained by CT-guided transthoracic needle aspiration showed many polymorphonuclear neutrophil leukocytes and a few gram-positive bacteria. The generator pocket was opened in the operating room and contained cloudy, purulent material. The AICD patches and sensing leads were explanted through a left anterior thoracotomy. Except in the areas of infected AICD patches. the pericardial cavity was obliterated with dense adhesions. Inflow-outflow catheter irrigation with a 10% povidone-iodine solution was established and continued for 10 days. Gram stain and cultures from the pericardial and generator pocket fluid failed to show any microorganism. The patient was treated empirically with a 2-week course of cephazolin. Since he was believed still to be at risk for sudden death, he subsequently underwent successful implantation of a custom-made transvenous spring lead and a subcutaneous patch lead defibrillator system. CASE 2. A 70-year-old man came to his cardiologist with induration of the skin over the subcutaneous pocket containing the buried defibrillator leads 7 months after a successful triple-vessel coronary artery bypass grafting combined with subendocardial resection, cryoablation of left ventricular arrhythmogenic foci, and implantation of AICD sensing leads and patches. Generator implant was not required, because postoperative electrophysiologic study revealed no inducible ventricular tachycardia and the patient remained free of recurrent arrhythmias. His physical examination was notable only for the induration of skin over the subcutaneous leads without actual erosion. The total white blood cell count was normal with no leftward shift. Chest x-ray film revealed left pleural effusion. A CT scan of the chest disclosed a fluid density underneath the posterior AICD patch. At operation via a left anterior thoracotomy, a large, sterile left pleural effusion was drained and the sensing leads and infected AICD patches were removed. The infection was limited to the posterior AICD patch only. Inflow-outflow irrigating tubes with a 10% povidone-iodine solution were established and
Volume 95 Number 5 May 1988
continued for 2 weeks. Cultures of material from beneath the posterior patch electrode subsequently grew Candida albicans. Results of a thorough search for systemic candidiasis were negative. He was treated with a limited course (300 mg) of intravenous amphotericin B and was discharged in good condition. At 3 months' follow-up, he remains free of any problem. CASE 3. A 75-year-old-man underwent five-vessel coronary artery bypass and implantation of AICD patches in 1984. Appropriate electrophysiologic studies revealed no inducible sustained ventricular arrhythmia, and no pulse generator was implanted. Two and a half years later the patient began to have pain and discomfort over the subcutaneous leads in the left upper quadrant of the abdomen. He was afebrile and the only pertinent positive finding was induration of skin over the leads without erosion. The total white blood count was normal with no leftward shift, and the chest x-ray film showed no pleural effusion. A CT scan of the heart revealed fluid density beneath both AlCD patches. Needle aspiration of the subcutaneous pocket revealed purulent material containing many white blood cells but no bacteria. The infected AICD patches and leads were explanted through a left anterior thoracotomy. The infection was localized to the patches, and the rest of the pericardium was unaffected. Continuous catheter irrigation with a 10% povidone-iodine solution was established. The cultures from the initial subcutaneous pocket aspirates remained negative. However, cultures of the AICD patch after sonication grew Staphylococcus epidermidis, and the same organism was grown from the pericardiaI patches on routine microbiologic culture techniques 3 days later. The patient was treated with a 2-week course of cefazolin to which the organism was sensitive and left the hospital in good condition.
Discussion An increasing number of patients with life-threatening ventricular arrhythmias are being managed with implantation of AICD devices, with or without concomitant myocardial revascularization and surgical ablative procedures. Results in several large series of patients with these devices have been published with a variety of complications reported." Delayed primary infection of these systems has not heretofore been recognized. In our series of 122 patients with one or more ventricular patch electrodes, we have had six infectious complications (4.9%). In three of these patients, the infection of the AICD devices was associated with infective foci elsewhere in the body (staphylococcal antecubital phlebitis, mixed flora empyema, and staphylococcal sternal infection) and happened within 2 weeks of the implant. However, all three of the patients described in this report had delayed infection (7 months to 31 months after the initial operation) without prior perioperative infection; none had systemic symptoms, and actual erosion of the skin was absent. Primary bacterial infection of the pericardial cavity is an uncommon and potentially lethal disease, most
Infection of AICD
frequently seen in children." In adults, it is mainly seen in immunocompromised patients such as those receiving long-term hemodialysis.' Patients with this disease are systemically ill and have high fever."! In sharp contrast to patients with primary purulent pericarditis, our patients had no systemic manifestations of the intra pericardial infections. The reasons for the lack of systemic response to the presence of infection with these devices is not clear. It is conceivable that operation on the heart and pericardium blocked or altered the lymphatic drainage and made it less accessible to the body's immune system. In all patients described in this report, the patch electrodes were placed within the pericardial space. None of our patients were immunosuppressed and all had positive results to a battery of skin test antigens. The most likely mechanism of infection is the seeding of patches at the time of implant with low virulence, slow growing, slime producing Staphylococcus epidermidis" These bacteria produce a surface biofilm, which is a complex structure composed of coalescing microcolonies of bacteria and a surrounding fibrous anionic matrix of exopolysaccharide, glycocalyx.'? This biofilm facilitates adhesion of microorganisms to the prosthetic surface and provides nutrient flow, while at the same time protecting them from the host defense mechanisms and antibacterial agents.?'!' Slime production by some strains of Staphylococcus epidermidis may be an important virulence factor that may cause delayed infection of prosthetic material." Such organisms were noted in case 3 and suspected in case I, in which Gram stain showed gram-positive cocci. Infection caused by microorganisms that could not be cultured with routine bacteriologic culture techniques could not be excluded in either case I or 2. All three patients had myocardial revascularization combined with AICD device implantation. In addition, in case 2, the patient had intraoperative mapping and cryoablation. These procedures tend to be lengthy and are associated with more equipment and personnel in the operating room than routine myocardial revascularization. Any or all of these factors may increase the likelihood of bacterial seeding of the operative field. Thus far, a single infection involving the generator pocket and patch electrodes has occurred in a patient 2V2 months after implantation via a left anterior thoracotomy. This patient was receiving immunosuppressive therapy attendant to a renal allograft, however. In case 3, 3 days were required for growth of Staphylococcus epidermidis by conventional culture techniques, indicative of the slow growth characteristics of this organism. In each patient, the infection was localized to the immediate capsule around the patch electrode(s). The
The Journal of Thoracic and Cardiovascular
A/massi et a/.
Fig. 1. CT scan of chest showing fluid density between heart and AICD patch leads (arrows).
pericardium remote from these areas was unaffected and was tightly adherent to the epicardial surface of the heart. This feature facilitates the operative management, because once the patch electrode is identified, it can easily be removed without need for further dissection. Radical debridement of the pericardium does not appear to be necessary. Removal of the lead system and pulse generator in its entirety, as advocated for pacemaker infection, 12. 13 is essential. The radiographic appearance of the normal AICD device has been recently described. 14 In none of our three patients did the radiograph appear abnormal. On the other hand, CT scan of the heart was strongly suggestive of the diagnosis. Fluid density between the heart and an AICD patch, particularly in the presence of a previously normal CT appearance, indicates that a local infection is likely (Fig. I). Several of our patients have had chest CT scans in the initial postoperative period for different reasons, and no fluid collection has been identified underneath the AICD patches. CT-guided aspiration of the fluid collection beneath the AICD patch lead may help establish the etiologic agent of the condition preoperatively. However, when there is evidence of purulent fluid within the generator or lead pocket, negative results of a tap should not deter surgical explantation of the pulse generator and the entire lead system. This is particularly true if induration of the skin is present over the generator or lead pocket. These appear to be late, rather than early, manifestations of infection, because all of the patients described here already had an extensive process present intrapericardially at the time skin changes were first noted.
An anterior thoracotomy incision through an appropriate intercostal space (usually the fifth) provides excellent exposure for removal of the infected hardware. The incision can be extended across the sternum if further exposure of the right side of the heart is required. Continuous local irrigation of the pericardia I cavity with a 10% povidone-iodine solution combined with systemic antibiotic coverage is used to sterilize the spaces. We have adopted the policy that, when the effluent is clear, saline irrigation is substituted for povidone-iodine and is continued for 24 hours. If cultures from the saline effluent are negative, irrigation is discontinued and the catheters are removed. In the single patient who was dependent on AICD for the management of arrhythmia, a transvenous spring lead and a thoracic subcutaneous patch lead configuration provided a satisfactory alternative to repeat intrathoracic implantation. IS We believe that patients with AICD hardware or a generator, whether intravascular, extravascular, or both, should be regarded in a manner similar to those with prosthetic valves or other prosthetic devices in terms of antibiotic prophylaxis during any procedure that may result in bacteremia. Antibiotic impregnated patches, if and when available, may provide another adjunctive measure for prevention of delayed infection. Alteration of the dynamics of pericardiaI lymphatic drainage and the possible role of patch electrodes in modifying pericardiaI response to infection are subjects in need of further investigation. REFERENCES I. Echt OS, Armstrong K. Schmidt P. et al. Clinical experience, complications, and survival in 70 patients with the automatic implantable cardioverterjdefibrillator. Circulation 1985;71 (pt 2):289-96. 2. Marchlinski FE, Flores BT, Buxton AE, et al. The automatic implantable cardioverter-defibrillator: efficacy. complications, and device failures. Ann Intern Med 1986; 104:481-8. 3. Mirowski M, Reid PRo Winkle RA. et al. Mortality in patients with implanted automatic defibrillators. Ann Intern Med 1983;98:585-8. 4. Mirowski M, Reid PR, Mower M. et al. Clinical performance of the implantable cardioverter defibrillator. PACE 1984;7:1345-50. 5. Platia EV, Griffith LSC, Watkins L. et al. Treatment of malignant ventricular arrhythmias with endocardial resection and implantation of the automatic cardioverterdefibrillator. N Engl J Med 1986;314:213-6. 6. Rane HS, Lahari KR, Desai AG. Pyo-pericarditis in children. Indian J Pediatr 1984;51:305-8. 7. Mako J. Jansen J. Bognar B. Farago A. Purulent pericarditis caused by Staphylococcus aureus in two
Volume 95 Number 5 May 1988
patients undergoing haemodialysis. Int Urol Nephrol 1985;17:79-83. Fyfe DA, Hagler DJ, Puga FJ, Driscoll DJ. Clinical and therapeutic aspects of Haemophilus influenzae pericarditis in pediatric patients. Mayo Clin Proc 1984;59:41522. Tollefson DF, Bandyk DF, Kaebnick HW, et al. Surface biofilm disruption. Arch Surg 1987;122:38-42. Costerton JW, Irvin RT, Cheng KJ. The bacterial glycocalyx in nature and disease. Annu Rev Microbiol 1981;35:299-324. Costerton JW, Irvin RT, Cheng KJ. The role of the bacterial surface structures in pathogenesis. CRC Crit Rev Microbiol 1981;8:303-38.
Infection of AICD 9 1 1
12. Chao MH, Gersh BJ, Pluth JR. Infected epicardial pacemaker systems. J THORAC CARDIOVASC SURG 1981; 82:794-6. 13. Ruiter JH, Degener JE, Van Mechelen R, et al. Late purulent pacemaker pocket infection caused by Staphylococcus epidermidis. PACE 1985;8:903-7. 14. Goodman LR, Troup PJ, Thorsen MK, Youker JE. Automatic implantable cardioverter-defibrillator: radiographic appearance. Radiology 1985;155:571-3. 15. Wetherbee IN, Chapman PD, Klopfenstein HS, et al. Nonthoracotomy internal defibrillation. J Am Coll Cardiol 1987;10:406-11.
Bound volumes available to subscribers Bound volumes of THE JOURNAL OF THORACIC AND CARDIOVASCLiLAR SURGERY are available to subscribers (only) for the 1988 issues from the Publisher. at a costof $51.00 ($68.00 international) for Vol. 95 (January-June) and Vol. 96 (July-December). Shipping charges are included. Each bound volume contains a subject and author index and all advertising is removed. Copies are shipped within 60 days after publication of the last issue of the volume. The binding is durable buckram with the JOURNAL name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. ContactThe C. V. Mosby Company, Circulation Department, 11830 Westline Industrial Drive, St. Louis, Missouri 63146, USA; phone (800) 325-4177, ext. 351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.