David L. Longworth

Diagnosis and Management of Infections Involving Implantable Electrophysiologic Cardiac Devices

Worldwide, there are approximately 3.25 million functioning pacemakers and 180 000 functioning implantable cardioverter defibrillators (Warkentin D. Personal communication). Infection rates for these devices range from 1% to 7% (1-4), and the optimal method for management of such infection has yet to be defined (5-16). Since 1988, the Cleveland Clinic Foundation in Ohio has been a referral center for treatment of device-related infections because of the tools and techniques used there to extract leads for transvenous pacemakers and implantable cardioverter defibrillators (17). To define optimal management of these infections, we examined our recent experience at the Foundation. Methods We reviewed information on all patients treated at the Cleveland Clinic Foundation between 1 January 1995 and 31 August 1998 who had explantation of pacemakers or implantable cardioverter defibrillators. Only patients who satisfied the case definition of infection were included. Infection was defined as the presence of local warmth, erythema, swelling, edema, pain, or discharge from the device pocket along with a positive culture from the device, device pocket, blood, or lead. Device-associated endocarditis was defined as the presence of lead or valvular vegetation on surface or transesophageal echocardiography. Relapse was defined as the recurrence of infection with a similar type of organism documented by antibiogram within 1 year of treatment of the original infection. Statistical calculations were performed by using EpiInfo, version 6 (Centers for Disease Control and Prevention, Atlanta, Georgia). Results Four hundred sixty-seven consecutive patients had explantation of their devices, leads, or both because of mechanical or infectious complications. Three hundred forty-four patients were excluded because no organisms were isolated from the device pocket or explanted hardware. One hundred twenty-three patients satisfied the criteria for device-related infection. Ninety-nine patients (81%) had had their device placed at another institution. One hundred nineteen patients (97%) had transvenously implanted leads, and 3% (n =4) had leads implanted through the epicardium. The study included 87 men and 36 women (mean age, 66 16 years [range, 14 to 93 years]). Comorbid conditions included coronary artery disease in 64% of patients (n =79), coronary bypass surgery in 32% (n =39), diabetes mellitus in 26% (n =32), anticoagulation in 19% (n =23), atrial fibrillation in 19% (n =23), malignant conditions in 6% (n =7), oral corticosteroid use in 5% (n =6), and hemodialysis in 2% (n =3). Corticosteroid use and diabetes mellitus occurred more frequently in patients with polymicrobial infections (12% and 38%, respectively). Infection occurred early (0 to 28 days after device placement) in 25% of patients (n =31), occurred late (29 to 364 days after device placement) in 33% of patients (n =41), and was delayed (at least 365 days after device placement) in 42% of patients (n =51). Patients most commonly presented with erythema and pain over the pocket (Table 1). Sixty-nine percent of patients (n =85) presented with symptoms localized to the pulse generator pocket, 20% (n =25) presented with a combination of local and systemic symptoms, and 11% (n =13) presented with systemic signs and symptoms alone. Table 1. Clinical Presentation Thirty-three percent of patients (n =40) were bacteremic, but of these, only 24 (60%) had systemic symptoms. Eighty-one percent of pulse generator pocket cultures (n =99) were positive, and of patients with positive cultures, 24% (n =24) had bacteremia. The most common pathogens were coagulase-negative staphylococci in 68% of infections (n =83), Staphylococcus aureus in 24% (n =29), and enteric gram-negative bacilli in 17% (n =21). Thirteen percent of infections (n =16) were polymicrobial. Antibiotics were given to 76% of patients (n =94) before they presented for device extraction. Twenty-three percent of patients (n =28) received intravenous antibiotics alone (median, 5 days), 33% (n= 40) received intravenous antibiotics (median, 8 days) followed by oral antibiotics (median, 20 days), and 20% (n =25) received oral antibiotics alone (median, 15 days). One patient received topical antibiotic treatment for 30 days. Sixty-three percent of patients (n =78) had been previously hospitalized for a mean duration of 11 days (range, 1 to 100 days), and 33% (n =40) had had a previous surgical intervention without full removal of all hardware. Echocardiography was performed on 64 patients (52%). Forty-five patients (37%) received transthoracic echocardiography, 8 (7%) received transesophageal echocardiography, and 11 (9%) received both types. Thirteen patients, 12 with a pacemaker and 1 with an implantable cardioverter defibrillator, had vegetations on valves, leads, or both. The device and all lead material were completely removed in 95% of patients (117 of 123). Six of 123 patients did not have removal of all hardware. One hundred nineteen patients received antibiotics after hardware was removed. Fifty-eight percent of patients (n =71) received intravenous antibiotics alone (median, 28 days), 35% (n =43) received intravenous antibiotics (median, 7 days) followed by oral antibiotics (median, 16 days), and 4% (n =5) received oral antibiotics alone (median, 24 days). The mean interval from the date of explantation of the infected device to reimplantation of a new device was 7 days (median, 5 days [range, 0 to 68 days]) (Table 2). Of 13 patients with vegetations, 7 underwent reimplantation a median of 7 days (range, 5 to 25 days) after extraction, 5 did not require further device therapy, and 1 had a new device implanted at another institution. Thirty percent of patients (n =37) did not have their devices reimplanted during the index hospitalization. Eighteen percent (n =22) did not require further device therapy, and 6% (n =7) had their devices reimplanted at other institutions. Two patients were treated with antibiotics alone, 2 had lead removal only, and 1 left the treatment center against medical advice. One patient was treated with ablation, 1 died of congestive heart failure, and 1 declined to undergo reimplantation for the long QT syndrome. Table 2. Analysis of Results according to Subgroup The median hospital stay was 8 days (range, 1 to 65 days). One patient died of congestive heart failure before discharge. Follow-up information was available for 115 patients (94%) at a mean interval of 56 weeks after discharge (range, 1 to 194 weeks). Eighty-six percent of patients (n =106) were still alive at the last follow-up. The relapse rate was 3% (n =4). Relapse occurred in 1 of 117 patients who had all hardware removed and in 3 of 6 patients who did not have complete hardware removal. The patient who experienced relapse despite complete hardware removal was the only patient to have a new device reimplanted in the old pocket. All other patients had new devices reimplanted at a new site at a later date. The crude mortality rate was 8% (n =10), and causes of death were not related to surgery or infection (respiratory failure [n =3], congestive heart failure [n =3], multiorgan failure [n =1], pneumonia [n =1], myocardial infarction [n =1], and renal failure [n =1]). Discussion Although our study of our clinical experience in the treatment of device-related infection was not randomized, it provides substantial insight into an effective approach in a large group of patients. Our study also describes the management delivered by the community at large and documents numerous unsuccessful attempts to treat infections without removing all hardware. The length of hospital stay for these unsuccessful attempts at device salvage was similar to if not longer than that for complete extraction of all hardware. Some authors (5-11) have advocated conservative therapythat is, antibiotic therapy with hardware in placefor device-related infections. However, conservative therapy has a limited role in the treatment of such infections and was effective for only three patients in our study group. Forms of conservative treatment, such as partial device removal, pocket debridement, and repositioning of the device, are best used as a bridge to full explantation. Complete explantation, as recommended by other studies (12-16) and supported by our study, is an effective form of management for device-related infections. Conservative therapy may be tried for a short period but should generally be reserved for patients who cannot tolerate any surgical procedure (13). Table 2 summarizes the various clinical conditions, duration of antibiotic treatment, date of reimplantation, and treatment outcome for our patients. Bacteremia or endocarditis on presentation was not a contraindication for ultimate reimplantation. Successful reimplantation was accomplished when patients were afebrile and cleared of bacteremia after extraction. Some authors have recommended reimplantation as early as 36 hours after explantation in patients with only local symptoms of device-related infection (18). In our study, 19% of bacteremic patients presented with only local symptoms. Therefore, because most blood cultures turn positive within 48 hours, it is prudent to wait for results before considering reimplantation (19). Also, 18% of our patients no longer required device therapy or had reasonable alternatives after their devices were removed. Therefore, the need for reimplantation in patients with infected devices should be reassessed. Our study has important limitations. Eighty-one percent of our patients initially presented to other institutions, and 33% had previous failed treatment attempts. This study therefore has a potentially significant referral bias because treatment of referred patients was difficult, complicated, and high risk. Optimal treatment of infected pacemaker and implantable defibrillator devices involves complete explantation of all hardware, followed by antibiotic the

Lactobacillus Bacteremia and Endocarditis: Review of 45 Cases

Lactobacilli are part of normal gastrointestinal and genitourinary flora but are an uncommon cause of bacteremia. We reviewed the cases of 45 patients with clinically significant lactobacillus bacteremia occurring over 15 years. Underlying conditions were common, including cancer (40%), recent surgery (38%), and diabetes mellitus (27%). Twenty-two patients were in the intensive care unit at the time of onset of lactobacillus bacteremia. Eleven of the 45 patients were receiving immunosuppressive therapy, 11 were receiving total parenteral nutrition, and 23 had received antibiotics without activity against Lactobacillus prior to the occurrence of bacteremia. Bacteremia was polymicrobial in 27 patients and developed during hospitalization in 39. Thirty-one patients died, but only one death was attributable to lactobacillus bacteremia. Lactobacilli are relatively avirulent pathogens that produce bacteremia in patients with serious underlying illnesses, many of whom have received prior antibiotic therapy that may select out for the organism. While rarely fatal in itself, lactobacillus bacteremia identifies patients with serious and rapidly fatal illness.

The incidence of invasive aspergillosis among solid organ transplant recipients and implications for prophylaxis in lung transplants

Abstract: Background. Invasive aspergillosis (IA) is associated with significant morbidity and mortality in solid organ transplant recipients but data on the incidence rates stratified by type of solid organ are limited. Objective. To describe the attack rates and incidence of IA in solid organ transplant recipients, and the impact of universal Aspergillus prophylaxis (aerosolized amphotericin B or oral itraconazole) in lung transplant recipients. Patients. The 2046 patients who received solid organ transplants at the Cleveland Clinic Foundation from January 1990 through 1999 were studied. Methods. Cases were ascertained through computerized records of microbiology, cytology, and pathology reports. Definite IA was defined as a positive culture and pathology showing septate hyphae. Probable IA was clinical disease and either a positive culture or histopathology. Disseminated IA was defined as involvement of two or more noncontiguous anatomic sites. Results. We identified 33 cases of IA (28% disseminated) in 2046 patients (attack rate = 1.6%) for an incidence of 4.8 cases per 1000 patient‐years (33 cases/6813 pt‐years). Both the attack and the incidence rates were significantly higher for lung transplant recipients vs. other transplant recipients: lung 12.8% (24 cases/188 patients) or 40.5 cases/1000‐pt year vs. heart 0.4% (3/686) or 1.4 per 1000‐pt year vs. liver 0.7% (3/439) or 2.1 per 1000‐pt year vs. renal 0.4% (3/733) or 1.2 per 1000‐pt year (P < 0.01). The incidence of IA was highest during the first year after transplantation for all categories, but cases occurred after the first year of transplantation only in lung transplant recipients. The attack rate of IA in lung transplant recipients was significantly lower after institution of routine Aspergillus prophylaxis (4.9% vs. 18.2%, P < 0.05). Conclusions. The highest incidence and attack rate of invasive aspergillosis among solid organ transplant recipients occurs in lung transplant recipients and supports the routine use of Aspergillus prophylaxis for at least one year after transplantation in this group.

Cytomegalovirus Infection Is a Risk Factor for Invasive Aspergillosis in Lung Transplant Recipients

Invasive aspergillosis (IA) remains a major cause of morbidity and mortality following solid organ transplantation. To assess the incidence of IA following lung transplantation and to identify risk factors for its occurrence, we performed a case-control study involving 101 patients undergoing lung transplantation at our institution from 1990 to 1995 and reviewed the findings. Fourteen patients (14%) developed IA. The mean time from transplantation to diagnosis was 15 months. Nine patients died; the mean time to death from diagnosis was 13 days. Risk factors associated with developing IA included concomitant cytomegalovirus (CMV) pneumonia or viremia and culture isolation of Aspergillus species from a respiratory tract specimen after lung transplantation. Optimal strategies to prevent IA in lung transplant recipients remain to be determined, but prevention of aspergillus airway colonization and CMV viremia and disease after transplantation may be important targets for prophylactic interventions.

Should prophylaxis for Pneumocystis carinii pneumonia in solid organ transplant recipients ever be discontinued

Solid organ transplant recipients are at risk for Pneumocystis carinii pneumonia (PCP), but the risk of PCP beyond 1 year is poorly defined. We identified 25 cases of PCP in 1,299 patients undergoing solid organ transplantation between 1987 and 1996 at The Cleveland Clinic Foundation (4.8 cases per 1,000 person transplant-years [PTY]). Ten (36%) of 28 PCP cases (transplantation was performed before 1987 in three cases) occurred > or = 1 year after transplantation, and no patient developed PCP while receiving prophylaxis for PCP. The incidence of PCP during the first year following transplantation was eight times higher than that during subsequent years. The highest rate occurred among lung transplant recipients (22 cases per 1,000 PTY), for whom the incidence did not decline beyond the first year of transplantation. We conclude that the incidence of PCP is highest during the first year after transplantation and differs by type of solid organ transplant. Extending the duration of PCP prophylaxis beyond 1 year may be warranted for lung transplant recipients.

Venous Thrombosis Associated with Peripherally Inserted Central Catheters: A Retrospective Analysis of the Cleveland Clinic Experience

Peripherally inserted central catheters (PICCs) have become popular for long courses of intravenously administration of antibiotics. Although these devices are generally regarded as safe, thrombotic complications have been associated with their use. In a retrospective review, 51 (2.47%) of 2063 patients who had a PICC placed during 1994-1996 were found to have developed a total of 52 PICC-associated venous thromboses (VTs). Two patients received the diagnosis of pulmonary embolism that was a complication of VT. Risk factors for VT identified by multiple logistic regression analysis were younger age, history of VT, discharge to a skilled-nursing facility, and therapy with amphotericin B. VT is a significant complication of PICC placement. It may occur more frequently than previously recognized and may be complicated by pulmonary embolism. Clinicians should maintain a high index of suspicion, especially for high-risk patients.

Early onset prosthetic valve endocarditis: the Cleveland Clinic experience 1992-1997.

BACKGROUND We reviewed all cases of early onset prosthetic valve endocarditis (EO-PVE) occurring less than 12 months after valve operation among 7,043 patients undergoing heart valve replacements or repairs at The Cleveland Clinic between 1992 and 1997. METHODS Cases were defined by the Duke criteria and identified through prospective surveillance. RESULTS Seventy-seven cases of EO-PVE were identified (1 per 100 procedures), and during the study period the incidence of EO-PVE decreased from 1.5% (1992 to 1994) to 0.7% (1995 to 1997) (p < 0.01). The incidence of EO-PVE for rings (0.2%; 4 of 1,992) was significantly lower than for mechanical (1.6%; 28 of 1,731) and bioprosthetic valves (1.1%; 41 of 3,320) (p < 0.001). The incidence of EO-PVE was also significantly lower for mitral valve versus aortic valve surgeries (0.6% versus 1.4%, p < 0.001). The most common pathogens causing EO-PVE were coagulase-negative staphylococci (52%), fungi (13%), Staphylococcus aureus (10%), and enterococci (8%). Patients undergoing combined surgical and medical treatment of EO-PVE had a significantly higher 30-day, 2-year, and 3-year survival than medically treated patients, although patients judged to be too ill to survive surgery accounted for two-thirds of the patients treated medically. CONCLUSIONS There is a 1% incidence rate of EO-PVE among patients undergoing valve operations at our institution, usually caused by coagulase-negative staphylococci, and combined surgical and medical treatment is associated with improved survival compared with medical treatment alone.

Coagulase-Negative Staphylococcal Sternal Wound Infections After Open Heart Operations

BACKGROUND Coagulase-negative staphylococci are commonly isolated from wounds of patients after median sternotomy; however, the epidemiology of these infections is poorly described and the morbidity, mortality, and cost of care remain undefined. METHODS Retrospectively, we studied all patients with sternal wound infections attributable to coagulase-negative staphylococci after 22,180 open heart procedures performed at the Cleveland Clinic between January 1, 1988, and December 31, 1994 (84 months). In an assessment of potential risk factors for sternal wound infections caused by coagulase-negative staphylococci, 17 patients with coagulase-negative staphylococcal sternal wound infections were compared with 29 patients who underwent open heart operations without subsequent sternal wound infections, as well as with another 22 patients in whom sternal wound infections attributable to other pathogens developed. RESULTS A total of 436 sternal wound infections were identified (19 per 1,000 procedures), of which 100 (23%) were attributable to coagulase-negative staphylococci (4.5 per 1,000). Fifty-six percent of coagulase-negative staphylococcal sternal wound infections were superficial, 27% were deep, and 17% represented mediastinitis; 14% of patients had a concomitant secondary bloodstream infection. Ninety-two percent of coagulase-negative staphylococcal isolates were methicillin resistant. The mean interval from operation to onset of infection was 24 days (range, 4 to 388 days), and most patients had purulent discharge from the chest wound, fever, and leukocytosis. Adverse outcomes included reexploration (39%), flap operation (12%), and sternectomy (5%); 89% required parenteral antibiotics for a mean of 22 days. This resulted in 2,600 additional hospital days, with an average additional direct cost per case of

Use of leflunomide in an allogeneic bone marrow transplant recipient with refractory cytomegalovirus infection

20,000. In both case-control studies, insulin-dependent diabetes mellitus was the only risk factor significantly associated with sternal wound infections attributable to coagulase-negative staphylococci (p value = 0.02 by two-tailed Fishers exact test). CONCLUSIONS Sternal wound infections attributable to coagulase-negative staphylococci had a substantial impact on cardiothoracic surgery-related morbidity.

Infections of intracardiac devices

Summary:Ganciclovir-resistant cytomegalovirus (CMV) infection is an emerging problem in transplant recipients. Foscarnet resistance and cidofovir resistance have also been described, but no previous reports have suggested treatment regimens for patients with CMV refractory to all three of these drugs. Leflunomide, an immunosuppressive drug used in rheumatoid arthritis and in rejection in solid-organ transplantation, has been reported to have novel anti-CMV activity. However, its clinical utility in CMV treatment has not been described previously. We report an allogeneic bone marrow transplant recipient who developed CMV infection refractory to sequential therapy with ganciclovir, foscarnet, and cidofovir. The patient was ultimately treated with a combination of leflunomide and foscarnet. Both phenotypic and genotypic virologic analysis was performed on sequential CMV isolates. The patients high CMV-DNA viral load became undetectable on leflunomide and foscarnet, but the patient, who had severe graft-versus-host disease (GVHD) of the liver, expired with progressive liver failure and other complications. We concluded that leflunomide is a new immunosuppressive agent with anti-CMV activity, which may be useful in the treatment of multiresistant CMV. However, the toxicity profile of leflunomide in patients with underlying GVHD remains to be defined.