The prevalence of heart failure has become substantial in the United States with over 5.5 million Americans affected and approximately 700,000 new cases diagnosed annually.1,2 Implantable left ventricular assist devices (LVADs) are increasingly being used in patients with end-stage heart failure as a bridge to transplantation and as a destination therapy in those not deemed candidates for heart transplant.3 Newer generation continuous-flow LVAD pumps are associated with improved survival and decreased complications, including infection, primarily attributed to smaller size and the simplified pump mechanism compared with pulsatile-flow LVAD pumps.4 Infection rates have historically ranged from 25% to 80% with 1-year mortality being 5–6-folds higher in those that develop an LVAD infection (LVADI).5,6
Despite the high incidence of LVADI, an optimal antimicrobial surgical infection prophylaxis (SIP) regimen is not well established for either pulsatile-flow or newer generation continuous-flow LVADs. Recently published practice guidelines for antimicrobial prophylaxis in surgery acknowledged the limited body of literature regarding optimal SIP for LVAD implantations.7 These guidelines advocate that clinicians should modify their SIP protocols for LVAD implantation based on individual institution’s pathogen prevalence and antimicrobial resistance patterns. A recent national survey of 21 centers performing LVAD implantations found significant interinstitution variability in SIP regimens.8 At the three Mayo Clinic campuses, SIP regimens vary between surgeons. The regimens generally consist of either a single-drug regimen (vancomycin, cefazolin, or both) or a multidrug (3–4 drug) regimen (vancomycin, ciprofloxacin or cefepime, fluconazole, ±rifampin). Using our institutional database, we sought to compare LVADI risk between the different SIP regimens used among patients undergoing continuous-flow LVAD implantation.
This was a single-center, retrospective cohort study of 290 patients who underwent continuous-flow LVAD implantation between February 2007 and March 2015 at Mayo Clinic, Rochester. The primary objective was to compare the risk of LVADI between the single-drug and multidrug SIP regimens. Secondary objectives were to compare the 90-day and 1-year all-cause mortality between SIP regimen groups, as well as compare complication rates that could be associated with antibiotic use, such as occurrence of Clostridium difficile diarrhea and new-onset end-stage renal disease (ESRD). Study protocol was reviewed and approved by the Mayo Clinic Institutional Review Board (IRB).
Definitions used in this study are consistent with those published by Nienaber et al.9 Cases of LVADIs were classified by a team of Infectious Diseases physicians using the International Society of Heart Lung Transplantation (ISHLT) criteria.10 Only LVAD-specific and LVAD-related infections were included in this analysis. These are infections that are specific to LVAD recipients (do not occur in non-LVAD patients) and are related to the device hardware including pump and cannula infections, pocket infections, and percutaneous driveline infections. We also included patients who had bloodstream infection (BSI) related to LVADI, as well as endocarditis and mediastinitis (LVAD-related infections). Antimicrobial SIP was defined as antimicrobials that were started anytime within 24 hours of LVAD implantation procedure and continued for variable duration at the discretion of the treating clinician. Single-drug SIP was defined as receipt of either vancomycin, cefazolin, or both. Multidrug SIP was defined as single-drug SIP plus any other antimicrobial agent. Time to onset of infection was defined as the interval between LVAD implantation and LVADI diagnosis or the onset of symptoms, if known, whichever came first. Delayed sternal closure was defined as after LVAD implantation, when the patient’s skin was approximated with sterile occlusive dressing applied to nonapproximated sternum with planned return to the operating room for chest exploration and presumptive sternal closure, typically within 48 hours.
Patients were identified from a database of all LVAD implantations performed at Mayo Clinic campus, Rochester, from February 2007 to March 2015. Patients included were ≥18 years of age and had received continuous-flow LVADs, regardless of indication. Those patients who were reimplanted with continuous-flow LVADs due to previous device malfunction of pulsatile-flow LVAD were also included. Patients who received older generation LVADs that employ pulsatile-flow and centrifugal pump mechanism were excluded, as these devices are no longer used in contemporary practice. Patients who did not have antimicrobial prophylaxis information available were excluded as well. Patients who were receiving antimicrobial therapy at the time of LVAD implantation for empiric or definitive treatment of a diagnosed infection were also excluded. Data were collected via chart review and from the preexisting database that included antimicrobial SIP regimen received. We also abstracted patient demographics, baseline characteristics, infection risk factors, surgical complications, and microbiological data.
Continuous data were presented as median with interquartile range (IQR) for significant skewness and analyzed using Wilcoxon rank-sum test. Shapiro–Wilk test was used to assess the normality of data distribution. Categorical data were presented as proportions and analyzed using χ2 test. Kaplan–Meier method was used for survival analysis, and survival between the two groups was compared using log-rank test. A p value <0.05 was considered statistically significant. R software (R Foundation for Statistical Computing, Vienna, Austria), version 3.1.0., was used for data analysis and visualization.
Between February 2007 and March 2015, 290 patients underwent LVAD implantation at Mayo Clinic, Rochester. Of these, 16 patients who did not have HeartMate II or HeartWare device were excluded, additional 14 patients were excluded because they had missing perioperative antimicrobial data, and additional 21 patients were excluded because they were being treated for an active infection at the time of LVAD implant. Patient characteristics were similar between groups and are summarized in Table 1. A description of the SIP regimens used is listed in Table 2.
There were 199 patients in the single-drug SIP cohort and 40 in the multidrug SIP cohort (Table 2). At 90 days, LVADI occurred in five of 239 (2.1%) patients, with gram-positive organisms as the causative agent in over half of all LVADIs. In regard to LVADI risk between SIP regimens, LVADI occurred in three patients (1.5%) in the single-drug group versus two patients (5%) in the multidrug group at 90 days (p = 0.4) (Table 3). The median time to infection was 14.1 months in the overall cohort (5.3–19.1). In the multidrug group (N = 2), the median time to infection was 15.9 (10.8–30.3) vs. 12.5 (5.3–18.5) months in the single-drug group (N = 18), p = 0.5. Description of LVADI cases is outlined in Table 4.
The overall 90-day and 1-year survivals were 89.9% and 78.3% in the single-drug group and 90% and 87.1% in the multidrug group, respectively (Figure 1). Infection-free survival at 90 days and 1 year was 88.4% and 73.2% in the single-drug group and 85.0% and 82.1% in the multidrug group, respectively (Figure 2). Overall, there was no statistically significant difference observed between two groups in either overall survival or infection-free survival.
Our investigation is the largest to date that systematically evaluates and compares SIP regimens for LVAD implantation. Moreover, it is the first such study to employ LVADI case definitions proposed by ISHLT.10 We compared the efficacy of different SIP regimens in 239 patients who underwent continuous-flow LVAD implantation. Overall, we found no statistically significant differences in LVADI risk between single-drug and multidrug SIP regimen groups.
Previously published literature on this subject is limited to multi-institutional surveys and expert opinions.8,11 Surgical infection prophylaxis regimens used for LVAD implantation procedures in published studies are variable. The commonality between these studies is the utilization of multidrug regimens, which are derived largely from the SIP protocol used in the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial.3 This trial was not designed to evaluate the efficacy or superiority of a particular SIP regimen. However, since the publication of the REMATCH trial, the multidrug SIP has become standard practice at many institutions. These multidrug SIP regimens include various combinations of vancomycin, rifampin, fluoroquinolones, extended spectrum β-lactam agents, and fluconazole.12–14 However, none of these multidrug SIP regimens are based on published evidence or practice guideline recommendations. Investigators in these studies simply reported the SIP regimen used for LVAD implantation at their institution and did not include a systematic analysis of the impact of SIP regimen on postimplantation LVADIs.
Current guidelines for antimicrobial prophylaxis in surgical procedures recommend use of a β-lactam agent (such as cefazolin) for LVAD implantation.7 Vancomycin and clindamycin are recommended as acceptable alternatives in patients with history of allergic reactions to β-lactam agents such as penicillins or cephalosporins. Routine use of vancomycin as first-line agent is not recommended. For patients with known methicillin-resistant Staphylococcus aureus (MRSA) colonization, guidelines suggest that it is reasonable to add a single dose of vancomycin to the first-line recommended agent. However, if MRSA colonization status is unknown, vancomycin may be added for high-risk patients (e.g., recent hospitalization, nursing home residents, hemodialysis patients).
The primary reason for the guideline’s recommendation to use cefazolin for SIP, instead of simply switching to vancomycin, is the concern that vancomycin may be less effective for preventing surgical site infection caused by methicillin-susceptible S. aureus (MSSA).11,15 In one clinical trial11 of 885 adults undergoing cardiac surgery at a tertiary care center with high MRSA prevalence, patients were randomized to receive either vancomycin or cefazolin for 24 hours. Patients were followed up for at least 30 days (1 year for those receiving a cardiac implant). Although the overall surgical site infection rates were similar in the two groups (43 cases in the vancomycin group [9.5%] versus 39 cases in the cefazolin group [9.0%]; p = 0.8), surgical site infections caused by MSSA were significantly more common in the vancomycin group (17 cases [3.7%] versus six cases [1.3%]; p = 0.04). In another study,15 investigators reviewed 22,549 surgical procedures to assess the impact of vancomycin SIP on the development of MSSA surgical site infections. In their multivariable analysis, use of vancomycin as standard SIP was an independent predictor of higher rate of MSSA infections with adjusted odds ratio (OR) for a surgical site infection caused by MSSA being 2.79 when vancomycin prophylaxis was administered (p < 0.001).
Although data regarding optimal SIP regimen specifically for LVAD implantation are lacking, literature exists regarding SIP use for other cardiothoracic procedures involving implantable prosthesis. Single-dose cefazolin before pacemaker implantation was highly effective in preventing major infectious complications in a large, single-center, prospective cohort.16 Another study by de Oliveira et al.12 confirmed the benefit of single-dose cefazolin SIP prophylaxis for pacemaker and cardioverter–defibrillator implantations in a randomized, double-blinded, placebo-controlled, prospective clinical trial. The trial was prematurely stopped due to the observed benefit of cefazolin SIP. Multivariable analysis identified nonuse of perioperative cefazolin as an independent predictor of infection.12 Furthermore, recent guidelines on the management of cardiovascular implantable electronic device infections support utilization of a SIP regimen with anti-Staphylococcal coverage, preferably cefazolin for those patients without a penicillin allergy.13
Perioperative infection risk among patients undergoing LVAD implantation may differ from that of other cardiac devices. Patients undergoing LVAD implantation are at substantial risk of infection due to their critical state of illness, external driveline, presence of multiple indwelling lines and other devices, delays in sternal closure, and prolonged pre- and postimplantation intensive care unit stay.8,17,18 Gram-positive cocci, especially staphylococci, account for majority of LVADI. A recent study by Nienaber et al.9 examined the microbiologic causes of LVADIs and found the following breakdown: 45% were gram-positive bacteria, 25–30% were gram-negative bacteria, 4–7.5% fungi, and anaerobes comprised the remaining LVADIs. Mixed infections occurred in 10–15% of patients. Therefore, although single-drug SIP regimen may be appropriate for most routine LVAD implantations, a more broad-spectrum multidrug SIP regimen may be reasonable for patients with known prior infection or colonization with MRSA, drug-resistant gram-negative bacteria, and candida species.
Our investigation provides novel insights in an area that lacks substantial prior research, but the limitations of such retrospective analyses merit thorough discussion. The clinical complexity of LVAD recipients makes it difficult to attribute infections solely to the inadequacy of the SIP regimen. Surgical infection prophylaxis regimens used in our cohort were based on physician discretion rather than some predefined criteria. Although not ideal, we believe that this reflects real-world practice as there are no published studies regarding efficacy or superiority of a particular SIP regimen. Because of the low event rate in our cohort, we lack adequate power to perform a detailed multivariable analysis and adjust for all relevant covariates. However, the only measured covariates that were somewhat imbalanced between the SIP regimens are baseline creatinine, duration of surgery, and time on bypass (Table 1). We chose to report all-cause 90-day and 1-year mortality, instead of infection-related mortality, because the cause of death is often multifactorial in this critically ill population.
We have tried to minimize the case ascertainment bias in our study by using standardized and reproducible case definitions proposed by ISHLT. Moreover, difficult-to-classify cases were reviewed by a team of Infectious Diseases physicians at our institution. As a major confounding factor, patients being treated for an active infection at the time of LVAD implantation were excluded; thus, our conclusion do not apply to those patients being treated for an active infection at the time of device implantation. Another limitation is the optimal time to evaluate outcomes in this patient population. Occasionally, device infections due to indolent pathogens, such as Staphylococcus epidermidis, acquired at the time of LVAD implantation can manifest several months later. For this reason, we analyzed patients up to 12 months from the date of implantation, realizing that late-onset infections might not be related to SIP that was used during surgery and might develop infections because of other reasons (BSIs, etc.).
Optimal SIP duration is another unknown and important variable to consider. Because of variability and inconsistency in SIP duration in our study cohort, we were unable to dichotomize SIP duration in our analysis. Change in clinical practice over time may be additional factors that influence the risk of LVADI.
In conclusion, we did not observe a statistically significant difference in LVADI risk and overall mortality between single-drug versus multidrug SIP for LVAD implantation. Considering the potential adverse consequences (such as emergence of resistance, toxicity, cost) related to use of broad-spectrum agents, we have modified our clinical practice in accordance with published guidelines and are now using cefazolin alone SIP for routine LVAD implantation. We believe that it is reasonable to consider a more broad-spectrum multidrug SIP for patients with known colonization or prior infection with drug-resistant organisms (such as MRSA or multidrug resistant gram-negative rods) or candida species. Institutional SIP protocols at other LVAD implant centers may be modified to reflect local pathogen prevalence and resistance profiles. Large, prospective, multicenter studies are needed to address this subject further.
1. Lloyd-Jones D, Adams RJ, Brown TM, et al. Executive summary: Heart disease and stroke statistics-2010 update: A report from the American Heart Association. Circulation 2010.121: e46e215.
2. Nienaber J, Wilhelm MP, Sohail MR. Current concepts in the diagnosis and management of left ventricular assist device
infections. Expert Rev Anti Infect Ther 2013.11: 201210.
3. Rose EA, Gelijns AC, Moskowitz AJ, et al. Long-term use of a left ventricular assist device
for end-stage heart failure. N Engl J Med 2001.345: 14351443.
4. Topkara VK, Kondareddy S, Malik F, et al. Infectious complications in patients with left ventricular assist device
: Etiology and outcomes in the continuous-flow era. Ann Thorac Surg 2010.90: 12701277.
5. Baddour LM, Bettmann MA, Bolger AF, et al. Nonvalvular cardiovascular device–related infections. Most 2003.38: 20152031.
6. Gordon RJ, Weinberg AD, Pagani FD, et al. Prospective, multicenter study of ventricular assist device infections. Circulation 2013.127: 691702.
7. Bratzler DW, Dellinger EP, Olsen KM, et al. Clinical practice guidelines for antimicrobial
prophylaxis in surgery. Idsa 2013.70: 195283.
8. Walker PC, DePestel DD, Miles NA, Malani PN. Surgical infection prophylaxis
for left ventricular assist device
implantation. J Card Surg 2011.26: 440443.
9. Nienaber JJC, Kusne S, Riaz T, et al. Clinical manifestations and management of left ventricular assist device
-associated infections. Clin Infect Dis 2013.57: 14381448.
10. Hannan MM, Husain S, Mattner F, et al. Working formulation for the standardization of definitions of infections in patients using ventricular assist devices. J Hear Lung Transplant 2011.30: 375384.
11. Acharya MN, Som R, Tsui S. What is the optimum antibiotic prophylaxis in patients undergoing implantation of a left ventricular assist device
? Interact Cardiovasc Thorac Surg 2012.14: 209214.
12. Simon D, Fischer S, Grossman A, et al. Left ventricular assist device
: Treatment and outcome. Clin Infect Dis 2005.40: 11081115.
13. Holman WL, Park SJ, Long JW, et al. Infection
in permanent circulatory support: Experience from the REMATCH trial. J Hear Lung Transplant 2004.23: 13591365.
14. Schaffer JM, Allen JG, Weiss ES, et al. Infectious complications after pulsatile-flow and continuous-flow left ventricular assist device
implantation. J Heart Lung Transplant 2011.30: 164174.
15. Finkelstein R, Rabino G, Mashiah T, et al. Vancomycin versus cefazolin prophylaxis for cardiac surgery in the setting of a high prevalence of methicillin-resistant staphylococcal infections. J Thorac Cardiovasc Surg 2002.123: 326332.
16. Bull AL, Worth LJ, Richards MJ. Impact of vancomycin surgical antibiotic prophylaxis on the development of methicillin-sensitive Staphylococcus aureus
surgical site infections. Ann Surg 2012.256: 10891092.
17. Bertaglia E, Zerbo F, Zardo S, Barzan D, Zoppo F, Pascotto P. Antibiotic prophylaxis with a single dose of cefazolin during pacemaker implantation: Incidence of long-term infective complications. Pacing Clin Electrophysiol 2006.29: 2933.
18. de Oliveira JC, Martinelli M, Nishioka SA, et al. Efficacy of antibiotic prophylaxis before the implantation of pacemakers and cardioverter-defibrillators: Results of a large, prospective, randomized, double-blinded, placebo-controlled trial. Circ Arrhythmia Electrophysiol 2009.2: 2934.
19. Baddour L, Epstein A, Erickson C, Knight B. Update on cardiovascular implantable electronic device infections and their management. Am Hear Assoc. 2010.121: 458477.
20. Califano S, Pagani F, Malani P. Left ventricular assist device
–associated infections. Infect Dis Clin North Am 2012.26: 7787.
21. Gordon RJ, Quagliarello B, Lowy FD. Ventricular assist device-related infections. Lancet Infect Dis 2006.6: 426437.