HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): V. Paul Addonizio Valluvan Jeevanandam Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Prendergast, T. W. Articles by Jeevanandam, V. Search for Related Content PubMed PubMed Citation Articles by Prendergast, T. W. Articles by Jeevanandam, V.

Ann Thorac Surg 1997;64:142-147
© 1997 The Society of Thoracic Surgeons

---

Original Articles: Cardiovascular

Management of Left Ventricular Assist Device Infection With Heart Transplantation

Section of Cardiac and Thoracic Surgery, Temple University Health Sciences Center, Philadelphia, Pennsylvania

Accepted for publication January 20, 1997.

	Abstract

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Background. Left ventricular assist devices (LVADs) are being used as bridges to heart transplantation (HT). Infection of the LVAD in this patient population represents a serious complication, as simple LVAD removal or delaying HT may result in death. To improve outcomes in this group of patients, we performed HT in the presence of LVAD infection.

Methods. Eighteen patients underwent LVAD implantation followed by HT. Ten underwent HT in the absence of LVAD infection (group 1); and 8, in the presence of LVAD infection (group 2). All patients were treated similarly except for modification of immunosuppression in group 2 patients.

Results. Infectious and noninfectious complications were equivalent between the two groups. There was no difference between groups in regard to intraoperative deaths (one versus none), long-term survival (8/10 versus 7/8), wound complications (three versus none), and mean length of hospital stay after HT (21 versus 26 days).

Conclusions. Patients with LVAD infection are too seriously ill to allow LVAD removal or delay of HT. Transplantation in the face of infection is an effective treatment option.

	Introduction

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Since the 1980s, left ventricular assist devices (LVADs) have been used to support the circulation in patients in cardiogenic shock. In the setting of postcardiotomy cardiogenic shock, LVADs allow survival in up to 45% of patients who otherwise would not survive [1–3]. Indeed, the early success with the LVAD in supporting cardiac function in patients in postcardiotomy cardiogenic shock encouraged several groups to study the use of the LVAD in supporting patients awaiting heart transplantation (HT). These results have been encouraging, and this therapy is now accepted as an appropriate bridge to transplantation [4–6].

Experience with the LVAD as a bridge to transplantation at our institution confirmed the benefits of this therapy in patients in whom complications secondary to LVAD support did not develop. However, use of the LVAD in patients awaiting transplantation is not without risks. Complications of LVAD support include bleeding, thromboembolic events, pulmonary dysfunction, right heart failure, and infection [7]. Infection, which is of particular concern in patients awaiting transplantation, has been reported to occur in as many as 53% of patients with an LVAD [8]. Included are infections of the LVAD pocket and drivelines as well as intravascular infections of the LVAD inflow conduit and valves. The persistent contamination associated with the infected prosthesis can result in sepsis and associated septic complications (eg, septic embolization) [9]. Options for treatment of infection include simple removal of the LVAD, long-term treatment with antimicrobial therapy, replacement of the LVAD, or HT. Indeed, patients with LVAD–related infections are often so gravely ill that removing the LVAD or delaying HT to allow resolution of infection is likely to result in death.

We speculated that the infected prosthesis (ie, the LVAD) represented an infected foreign body and therefore needed to be removed in total. Because many patients awaiting HT would not tolerate LVAD removal alone, we undertook removal of the LVAD with concomitant HT in this group of patients. This report retrospectively analyzes the outcomes of patients who underwent HT in the presence of documented LVAD–related infection.

	Material and Methods

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
From 1991 to 1995, 22 patients at our institution underwent LVAD implantation as a bridge to HT. In 4 of these patients, complications unrelated to the device developed and precluded transplantation. The remaining 18 patients form the study group. All patients were operated on by the same group of surgeons using similar operative technique. The first three LVADs used were Thoratec (Berkeley, CA) models and the last 15, Thermo Cardiosystems (Woburn, MA) HeartMate (TCI) LVADs. Thoratec LVADs were placed in the conventional extracorporeal position. The first two TCI LVADs were placed intraperitoneally and the subsequent TCI LVADs, between the rectus muscle and the posterior rectus fascia (ie, preperitoneally).

Patients receiving an LVAD as a bridge to HT were divided into two groups on the basis of the presence or absence of infection at the time of HT. Group 1 (n = 10) consisted of patients without infection at the time of HT. These patients manifested no significant clinical signs of infection (eg, fever or leukocytosis), had no signs of localized LVAD infection (eg, erythema or purulent drainage), and had no positive blood cultures between the time of LVAD implantation and the time of HT. Group 2 (n = 8) consisted of patients with infection at the time of transplantation. Patients were considered as harboring infection if they had fever and at least two positive blood cultures within 1 week of transplantation. Blood cultures were considered positive if they yielded bacteria or fungus or both. Driveline or LVAD pocket infections were defined by positive cultures obtained from needle aspiration or intraoperative incision and drainage in the clinical setting of a localized infection at the LVAD site.

In group 2 patients, antimicrobial therapy was initiated at the first clinical sign of localized or systemic infection. Usually this was manifested as fever, leukocytosis, localized erythema at the LVAD site, or signs of systemic sepsis. When these signs appeared, broad-spectrum antibiotic coverage was promptly initiated along with needle aspiration or incision and drainage of the localized infection. Cultures were grown whenever possible, and antimicrobial coverage was adjusted according to the culture results. The culture-directed antibiotics were continued until the time of HT. The duration of antimicrobial therapy ranged from 1 day to 1 week and was determined by organ availability. Although these aggressive local and systemic measures for treatment of infection resulted in resolution of leukocytosis and fever in 5 patients (all of whom had driveline infections), all patients in group 2 manifested positive blood cultures within the week of HT. Postoperative antibiotic coverage was continued for 4 to 6 weeks and was based on the results of preoperative and intraoperative cultures and concomitant in vitro antibiotic-sensitivity testing.

The first six transplants were implanted using an atrial anastomotic technique [10], and the remainder in the series were performed using a bicaval anastomosis [11]. At the time of transplantation, all prosthetic material (device and graft) was removed in total. The wound and explant site were copiously irrigated with culture-directed antimicrobial solution delivered by Pulsavac (Zimmer, Warsaw, IN). Group 1 patients underwent irrigation with warm, sterile saline solution containing cefazolin sodium, 1 g/L. Group 2 patients underwent irrigation with warm, sterile saline solution containing antibiotics, 1 g/L, directed at specific organisms recovered preoperatively from the LVAD site or blood cultures. The antibiotics used for irrigation in group 2 included cefazolin, vancomycin, and amphotericin.

Inotropic support was provided as necessary to maintain cardiac index at or above 2.4 L•min^-1• m^-2, mean systolic blood pressure greater than or equal to 75 mm Hg, and pulmonary capillary wedge pressure between 10 and 20 mm Hg. We defined baseline inotropic support as dopamine hydrochloride at 2.5 µg•kg^-1• min^-1 and epinephrine at 2 µg/min. Inotropic support that exceeded the usual doses was considered to be greater than baseline.

Our usual immunosuppression protocol of cyclosporine, steroids, and azathioprine was altered for this special group of patients with infection. Instead, initial immunosuppression consisted of only steroids, with 1 g of methylprednisolone given intravenously at the onset of reperfusion. This dose was tapered to 125 mg every 8 hours for 1 day, then 125 mg every 12 hours for 1 day, and finally 125 mg per day for 1 day. Prednisone was then begun at 80 to 100 mg/day and tapered gradually to maintenance doses over the ensuing months. When liver transaminase enzyme levels were normal, azathioprine administration was initiated at a dose of 2 mg • kg^-1• day^-1 and adjusted thereafter so that the white blood cell count remained greater than 4 x 10⁹/L. Also beginning on postoperative day 2 or 3, cyclosporin A was given in divided doses twice per day beginning at 5 mg • kg^-1• day^-1 to achieve blood levels of 300 ng/dL as measured by whole-blood radioimmunoassay (Abbott, Abbott Park, IL). Group 1 patients received the same steroid schedule; however, azathioprine and cyclosporin A were begun on the day of transplantation.

For all patients, the following data are reported: type of LVAD used, duration of LVAD support, maximum temperature on the day before transplantation, white blood cell count at the time of operation, presence and source of infection in the perioperative period, results of perioperative blood cultures, results of LVAD explant cultures (LVAD pocket, graft, and valve), agents used for perioperative antibiotics, incidence of infectious and noninfectious postoperative complications, and survival. Postoperative cultures were obtained only when specific clinical indications arose (ie, patients 1 and 8 in group 1 and patient 2 in group 2). Results were analyzed using Student's paired t test with significance defined by a p value of less than 0.05.

	Results

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Of the 18 patients undergoing HT after LVAD support as a bridge to transplantation at our institution, only 10 patients had transplantation in the absence of LVAD infection (group 1) (Table 1). Eight patients who were considered too ill to tolerate either persistence of an infected indwelling LVAD prosthesis or simple LVAD removal underwent LVAD removal with concomitant HT in the presence of LVAD infection (group 2) (Table 2). The etiology of cardiomyopathy was similar in both groups and included ischemia, valvular disease, viral illness, and idiopathic causes. The type of LVAD used was similar in each group; a few patients seen early (2 in group 1 and 1 in group 2) received Thoratec LVADs, and later patients (8 in group 1 and 7 in group 2) were given TCI LVADs. Group 1 patients were supported by the LVAD for an average of 37 days and group 2 patients 38.5 days (p = not significant).

As expected, group 2 patients were more likely to demonstrate signs of systemic infection at the time of HT than their group 1 counterparts. Group 2 patients had higher average temperatures at the time of transplantation than the group 1 patients (37.8°C versus 37.4°C). Similarly, the average white blood cell count at the time of operation in group 2 was 17.1 x 10⁹/L versus 8.4 x 10⁹/L in group 1 (p = 0.02). Whereas all group 2 patients had positive blood cultures within the week preceding operation, blood cultures from group 1 patients showed no growth during this period. Various organisms were grown from the blood of group 2 patients including gram-positive organisms (Staphylococcus aureus, Staphylococcus epidermidis, and group D enterococcus), and fungus (Candida albicans). In no patient was the bacteremia or fungemia cleared by the time of HT. In all of these patients, the last documented culture before HT was positive.

The source of the infection in group 2 patients was variable. Left ventricular assist device valve endocarditis was present in 2 patients. Infection of the LVAD inflow conduit occurred in 1 patient. The remaining 5 patients had infection of the driveline. All patients whose blood culture was positive for fungus were ultimately found to have fungal vegetations in the LVAD valve or inflow conduit; in retrospect, this would not be expected to resolve with antimicrobial treatment alone. Indeed, the presence of such fungal organisms in blood and subsequently in the LVAD explant culture would seem to indicate that prosthesis removal is urgent.

Despite the different sources of LVAD infection, it is noteworthy that organisms grown from LVAD explant cultures correlated positively with organisms grown from blood cultures in all group 2 patients for whom both types of cultures were obtained. Our future investigations will try to elucidate the reasons for the high rate of infections among LVAD recipients and perhaps suggest structural changes in LVADs that may reduce the risk of infection (eg, valve or graft materials that may be less prone to endocarditis).

Not surprisingly, group 2 patients were more likely to receive higher-dose and broader-spectrum antibiotics for longer durations of treatment than patients in group 1. Whereas no patient in group 1 was on a regimen of antibiotics prior to HT, all 8 patients in group 2 were receiving such treatment prior to transplantation. Postoperatively, group 1 patients generally underwent only 3 days of antibiotic therapy with either a second-generation cephalosporin or a combination of vancomycin and gentamicin sulfate in patients allergic to penicillin. Only 1 patient (patient 8) in group 1 required additional antibiotic therapy. This patient had a fever postoperatively and was treated empirically with ampicillin for 10 days.

Group 2 patients, on the other hand, required prolonged postoperative antibiotic administration with a variety of agents (see Table 2). Generally, the patients were on a regimen of intravenous antibiotics for a total of 4 to 6 weeks (including preoperative antibiotics). Four weeks' total antibiotic duration was used for bacterial driveline infections and six weeks' duration for intravascular or fungal infections. No patients were kept on long-term oral antibiotic suppression for LVAD–related infection.

Patients who underwent HT in the presence of infection required substantially greater inotropic support than these without infection. Seven of the 8 patients with infection required inotropic support exceeding our baseline, and only 3 of 10 without infection patients required additional inotropic support (p = 0.04).

Despite differences in signs of infection, use of antibiotics, and need of inotropic support between groups 1 and 2, measures of outcome after HT were remarkably similar for the two groups. Specifically, the incidence of infection after transplantation was equal between groups. In group 1, pneumonia in 1 patient and fevers in another were treated empirically with antibiotics and resolved. In group 2, 1 patient had a localized infection of the driveline site and underwent incision and drainage of an abscess at this site. Sternal dehiscence, a potentially serious complication after HT, did not occur in our series. Three patients in group 1 sustained wound problems requiring minimal intervention (2 patients with fascial dehiscence and 1 with a sternal click); no wound problems were seen in group 2.

The incidence of serious noninfectious complications was not significantly different between groups 1 and 2. The only cerebrovascular accident after transplantation occurred in the first patient in group 2. The only intraoperative death in the series involved a group 1 patient with severe coagulopathy. The single late death in group 1 occurred 9 months after transplantation and was due to noncompliance. The sole death in group 2 also occurred 9 months after transplantation and was due to graft arteriopathy.

The length of hospital stay among patients who survived the initial operation was slightly less in group 1 than in group 2 (21 ± 3 days versus 26 ± 5 days), but this difference was not significant. The long-term survival after HT did not differ between groups. Eight of 10 patients in group 1 were alive as of February 1996 with a mean to-date survival of 17.3 months. In group 2, 7 of 8 patients were alive with a mean to-date survival of 15.2 months.

	Comment

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Several groups have demonstrated that LVAD support can successfully bridge patients to HT, but the presence of LVAD infection is often considered a contraindication to HT. There is concern that active infection increases risk after transplantation because postoperative immunosuppression may promote propagation of the infection and thereby decrease patient survival. However, development of an LVAD infection in a patient awaiting HT leads to the complications of harboring such an infected prosthesis; delay in LVAD removal or HT while waiting for resolution of the infection would likely result in substantial morbidity and mortality. This morbidity and mortality includes the systemic complications of sepsis and septic embolization to the brain and other organs [9]. Under such circumstances, the benefits of HT outweigh the risks of immunosuppression in the face of active infection and allow salvage of these critically ill patients.

Our initial concern was that the presence of active infection in patients immunosuppressed after HT would lead to problems with overwhelming infection in the postoperative period. As a result, our immunosuppression regimen was empirically altered to exclude induction therapy and hold off cyclosporin A administration for 48 to 72 hours. By withholding cyclosporin A, we hoped to avoid the complication of renal failure in patients already in a compromised hemodynamic state. In addition, we withheld azathioprine in an effort to improve leukocyte availability and with it, the ability to fight infection. These alterations may allow immune defense mechanisms to function maximally in the early postoperative period. Also, the later administration of cyclosporine had major renal benefits, as evidenced by the fact that no patient required postoperative dialysis. Further, LVAD infection did not seem to increase the risk of postoperative infection in this small group of patients. This reduction to single-agent immunosuppression in the first 2 to 3 days after HT did not result in any apparent increase in the number of rejection episodes. The number of rejection episodes in group 1 and group 2 patients in the first year after HT was not significantly different (1.9 ± 0.9 and 2.2 ± 1.0, respectively).

The incidence of postoperative infectious complications was equal between the two groups. Ironically, patients in group 1 were more likely to have sternotomy wound complications than patients in group 2 in this small series. The notable absence of infectious complications in group 2 patients suggests that despite postoperative immunosuppression, preoperative LVAD infection can be overcome with avoidance of induction therapy, aggressive surgical debridement, and judicious use of high-dose, culture-directed, intravenous antibiotics. Our aggressive approach to the treatment of patients in whom LVAD infection develops while they await HT has resulted in an excellent salvage rate in this small cohort of patients.

There are noteworthy alterations in the management of patients undergoing HT in the presence of active LVAD infection. Basic principles include expedient and complete removal of the entire infected prosthesis, particularly before sepsis and septic embolization occur. These patients generally require long-term use (4 to 6 weeks) of antibiotics specific for in vitro cultures and sensitivities. Alteration in immunosuppression strategies to avoid induction therapy has seemed to help get these patients through HT safely. Generally the degree of inotropic support needs to be intensified in these patients, perhaps because the myocardial-depressant factor of sepsis is more amplified in the transplanted heart. Of all these alterations, complete removal of LVAD material is probably the most important to assure good HT outcomes.

With these alterations in HT protocols, the long-term outcomes of our LVAD–supported patients undergoing HT in the presence and absence of LVAD infection are strikingly similar. There was no significant difference between groups in regard to length of hospital stay or survival (see Tables 1 and 2). From these data, it appears that preoperative LVAD infection, including fungal and vascular infections, does not necessarily influence long-term survival in HT patients.

Our investigation does have several methodologic problems. The study encompasses only 18 patients, is retrospective, and employs two different types of LVADs implanted by two different methods. However, severe LVAD infections are infrequent enough that only a few patients are available for analysis. In our opinion, prospective randomization would risk consequences as profound as stroke and death to patients denied transplantation. Despite these methodologic limitations, our data suggest that the rate of infectious complications was acceptable in patients undergoing transplantation in the face of active LVAD infection. Long-term outcomes after HT were similar in patients having the operation in the presence and absence of LVAD infection. For patients with active LVAD infection who harbor an infected prosthesis, transplantation in the face of infection is the best salvage treatment option.

	Footnotes

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 
Presented at the Sixteenth Annual Meeting of the International Society for Heart and Lung Transplantation, New York, NY, Mar 15-18, 1996.

Address reprint requests to Dr Prendergast, Division of Cardiothoracic Surgery, Kansas University Medical Center, Kansas City, KS 66160.

	References

Top Footnotes Abstract Introduction Material and Methods Results Comment References

 

1. Pennington DG, Samuels LD, Williams G, et al. Experience with the Pierce-Donachy ventricular assist device in postcardiotomy patients with cardiogenic shock. World J Surg1985;9:37–46.[Medline]
2. Pennock JL, Pierce WS, Wiseman CB, et al. Survival and complications following ventricular assist pumping for cardiogenic shock. Ann Surg1983;198:469–78.[Medline]
3. Pae WE, Gaines WE, Pierce WS, Waldhausen JA. Mechanical circulatory assistance for postoperative cardiogenic shock. Surg Rounds1985;7:49.
4. Phillips WS, Burton NA, Macmanus Q, Lefrak EA. Surgical complications in bridging to transplantation: the Thermo Cardiosystems LVAD. Ann Thorac Surg1992;53:482–6.[Abstract]
5. Sinisa B, Branislav R, Clay MB, et al. Heart transplantation after mechanical circulatory support: four years' experience. J Heart Lung Transplant1992;11:240–5.[Medline]
6. Burton NA, Lefrak EA, Macmanus Q, et al. A reliable bridge to cardiac transplantation: the TCI left ventricular assist device. Ann Thorac Surg1993;55:1425–31.[Abstract]
7. Frazier OH, Rose EA, Macmanus Q, et al. Multicenter clinical evaluation of the HeartMate 1000 IP left ventricular assist device. Ann Thorac Surg1992;53:1080–90.[Abstract]
8. Argenziano M, Moazami N, Catanese KA, et al. Diagnosis and management of LVAD endocarditis. J Heart Lung Transplant1996;15:S73.
9. Rose EA, Levin HR, Oz MC, et al. Artificial circulatory support with textured interior surfaces. A counterintuitive approach to minimizing thromboembolism. Circulation1994;90:1187–91.
10. Lower RR, Stofer RC, Shumway NE. Homovital transplantation of the heart. J Thorac Cardiovasc Surg1961;41:196–202.
11. Sievers HH, Weyland M, Kraatz EG, Bernhard A. An alternative technique for orthotopic cardiac transplantation with preservation of the normal anatomy of the right atrium. Thorac Cardiovasc Surg1991;39:70–2.04.[Medline]

This article has been cited by other articles:

				A. Zierer, S. J. Melby, R. K. Voeller, T. J. Guthrie, G. A. Ewald, K. Shelton, M. K. Pasque, M. R. Moon, R. J. Damiano Jr, and N. Moazami Late-Onset Driveline Infections: The Achilles' Heel of Prolonged Left Ventricular Assist Device Support Ann. Thorac. Surg., August 1, 2007; 84(2): 515 - 520. [Abstract] [Full Text] [PDF]

				W. L. Holman, B. K. Rayburn, D. C. McGiffin, B. A. Foley, R. L. Benza, R. C. Bourge, L. J. Pinderski, and J. K. Kirklin Infection in ventricular assist devices: prevention and treatment Ann. Thorac. Surg., June 1, 2003; 75(90060): S48 - 57. [Abstract] [Full Text] [PDF]

				S. M. Gordon, S. K. Schmitt, M. Jacobs, N. M. Smedira, M. Goormastic, M. K. Banbury, M. Yeager, J. Serkey, K. Hoercher, and P. M. McCarthy Nosocomial bloodstream infections in patients with implantable left ventricular assist devices Ann. Thorac. Surg., September 1, 2001; 72(3): 725 - 730. [Abstract] [Full Text] [PDF]

				R. A. Vilchez, M. C. McEllistrem, L. H. Harrison, K. R. McCurry, R. L. Kormos, and S. Kusne Relapsing bacteremia in patients with ventricular assist device: an emergent complication of extended circulatory support Ann. Thorac. Surg., July 1, 2001; 72(1): 96 - 101. [Abstract] [Full Text] [PDF]

				D. J. Goldstein, M. C. Oz, and E. A. Rose Implantable Left Ventricular Assist Devices N. Engl. J. Med., November 19, 1998; 339(21): 1522 - 1533. [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Add to Personal Folders Download to citation manager Author home page(s): V. Paul Addonizio Valluvan Jeevanandam Permission Requests Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Prendergast, T. W. Articles by Jeevanandam, V. Search for Related Content PubMed PubMed Citation Articles by Prendergast, T. W. Articles by Jeevanandam, V.