Left Ventricular Assist Device Related Infections: Does the Time of Onset Matter?

Abstracts S255 Pittsburgh Medical Center, Pittsburgh, PA; 2Cardiothoracic Surgery, University of Pittsburgh Medical Center, Pittsburgh, PA. Purpose: Driveline site infection (DLI) is a major cause of morbidity and mortality in patients with continuous flow left ventricular assist devices (CF-LVADs). Due to wide variability in patient-specific factors and infection progression, the appropriateness, timing and durability of surgical debridement is not clear. Our group sought to characterize CF-LVAD patients with DLIs that required surgical intervention of the driveline site. Methods: A retrospective review was done of all CF-LVAD DLIs at our institution from January 2010 to October 2015 and those that required surgical intervention of the driveline site were further analyzed. Surgical interventions were debridement with or without transposition to a new exit site. Those that required simultaneous LVAD pump exchange or pocket revision were excluded. Results: There were a total of 61 instances of DLI in 31 patients, from 172 primary CF-LVAD implants. Patient characteristics are further described in table 1. The mean time to first DLI from implant was 339 days (median 243). The most common drive line site culture organisms was MSSA (35%). All patients with DLI received antibiotics. Of the 31 DLI patients, 21 (67.7%) went on to have drive line (DL) site debridement and 2 others had simultaneous pump exchange. 12 (57.1%) of debridement patients went on to have another DLI, with a mean time to next DLI of 254 days (median 85), and 9 (42.9%) required a second DL site surgery. Mean time to next surgery 358 days (median 114 days).Interestingly the cultures from the debridement only matched the DL site cultures in 14 (66.6%) patients. Conclusion: Despite antibiotics, a high percentage of patients with DLIs required surgical intervention. DLIs recurred in over half of surgical patients, most within a year, with almost half requiring a second debridement. Staphylococcal species were the most common etiology, but about a third of patients had surgical cultures that were different from the superficial driveline cultures.

Table 1 Patients with Driveline Site Infection Variable

Sub-variable

Patients Mean Age at Implant (sd) Female Caucasian Mean BMI (sd) Diabetes Ischemic etiology Intention to Treat

31 54.5 (12.9) 6 (19.4) 22 (71.0) 33.7 (7.4) 13 (41.9) 13 (41.9) Bridge to Transplant 11 (35.5) Destination Therapy 15 (48.4) 24 (77.4) 7 (22.6) MSSA 11 (35.5) MSSA 4 (12.9) Coagulase negative Staph 4 (12.9) 4 (12.9) 21 (67.7) 14 (66.6)

Heartmate II HVAD DL site organism

Traumatic Etiology Debridement Matched DL site and surgical organisms Recurrent DLI after debridement Recurrent debridement

No (%)

12 (57.1) 9 (42.9)

6( 94) Surgical Management of Driveline Infections in Patients with Ventricular Assist Devices M. Pieri ,1 A. Scandroglio,1 G. Arlt,2 V. Falk,2 T. Krabatsch,2 E. Potapov.3  1San Raffaele Hospital, Milan, Italy; 2Deutsches Herzzentrum Berlin, Berlin, Germany; 3Deutsches Herzzentrum Berlin, Berlin, Italy. Purpose: Chronic ascending driveline and tunnel infections in patients with implantable left ventricular assist devices (LVAD) carry increasing risk for pump infection, thromboembolic events, decreased quality of life and increased hospitalization. The interventional strategies are extremely

heterogeneous. We report our experience of surgical management of refractory driveline infections without mediastinitis by translocation and wrapping of the driveline with omentum major. Methods: We reviewed data of VAD patients who underwent surgical treatment by translocation and wrapping with omentum for severe chronic driveline infection. In nine patients with exit site infection, the VAD cable was disconnected, placed intraabdominally, wrapped with omentum major and relocated on the other side of the abdomen creating a new exit site. In three patients with an infected fistula in the middle of the subcutaneous tunnel and clean exit site the fistula was excised and the cable relocated in the abdominal cavity and covered with omentum major, maintaining the previous exit site. Results: Twelve patients were treated between 1.1.2010 and 27.10.2015; 11 (92%) were male, and mean age was 55±15 years. Ten patients had a HeartWare HVAD (HeartWare International, Framingham, MA) as LVAD, 1 patient 2 Heartware HVAD as BIVAD, and 1 a HeartMate II device (Thoratec Corp., Pleasanton, CA). The pathogens isolated from the drive line are listed in table 1. Three episodes of postoperative bleeding requiring surgical revision were recorded: 1 perihepatic bleeding, 1 from the old driveline tract, and 1 abdominal bleeding. Nine patients were discharged from hospital, 1 died, and 2 are still in hospital at present. Out of 12 patients 2 had recurrence of infection within the first postoperative year. Conclusion: Driveline relocation with use of omentum major is an effective procedure. However, elevated risk for perioperative complications makes this surgical technique a last resort in cases of chronic severe driveline infection.

Microrganisms isolated from the driveline Microrganism

Patients, n

Methicillin sensitive Staphylococcus aureus Methicillin resistant Staphylococcus aureus Non-aureus Staphylococcus spp Enterococcus faecalis Escherichia coli Enterobacter cloachae Pseudomonas aeruginosa Achromobacter xylosoxidans Corynebacterium spp

4 1 1 1 4 3 2 1 2

6( 95) Left Ventricular Assist Device Related Infections: Does the Time of Onset Matter? A.L. Lehnert , A. Hart, S. Branam, S. Brouse, M. Guglin.  University of Kentucky HealthCare, Lexington, KY. Purpose: A main obstacle for long survival with left ventricular assist devices (LVAD) is the frequency of LVAD-associated infections (LVADIs). Contamination may occur during initial surgery/admission or at a later time. Our study looks to characterize the clinical manifestations and outcomes of LVADIs at our institution, depending on the time of the onset. Methods: Patients ≥ 18 years of age implanted with LVADs between August 2009 and December 2014 were screened and included if the initial LVADI was documented and treated at our institution. Patients were stratified into 2 groups based on whether the initial infection occurred either early (<  180 days) or late (≥  180 days) after LVAD implantation. Results: Out of 37 overall LVADI episodes, 16 (43%) and 21 (57%) of the initial LVADIs occurred early or late after device implantation, respectively. Median time to first LVADI was 88 ± 35 vs. 456 ± 187 days between groups. While superficial driveline-related infection was the most common LVADI type for both groups (56% vs. 71%, p =  0.489), driveline drainage as clinical presentation was more prevalent in the late group (24% vs. 69%; p =  0.009). Early LVADIs involved more gram-positive flora, mostly Staphylococcus aureus (69% vs. 33%), whereas late LVADIs involved more gram-negative pathogens, mostly Pseudomonas aeruginosa (25% vs. 57%; p =  0.045). High rates of treatment failure were consistent between groups (88% vs. 71%, p =  0.384), as was median antibiotic duration (28 ± 63 vs. 28 ± 17 days, p =  0.939). An all-cause mortality of 19% vs. 29% and cumulative incidence

S256

The Journal of Heart and Lung Transplantation, Vol 35, No 4S, April 2016

of LVAD explantation, heart transplant, and all-cause mortality of 75% vs. 71% was noted for early and late initial LVADI. Compared with superficial LVADIs, deeper infections were associated with increased mortality (13% vs. 46%, p =  0.046). Conclusion: Clinical manifestations of initial LVADI may vary with time of onset. Early onset with likely in-hospital contamination involved more systemic symptoms and gram positive flora, and infection depth correlated with poorer outcome. Late LVADIs involved more gram negative flora along with driveline drainage. Regardless of timing, success of antibacterial treatment was dismal, with high rates of relapse. Surgical treatment with pump exchange or listing for heart transplant should be considered early after LVADI onset. 6( 96) Increased Driveline Infection with Betadine Antisepsis: Considerations for Chlorhexidine Intolerant Patients A.Y. Son ,1 L.H. Stein,1 A. Reyentovich,2 S.D. Katz,2 D.E. Smith,1 A. DeAnda Jr.,1 L.B. Balsam.1  1Cardiothoracic Surgery, NYU Langone Medical Center, New York, NY; 2Medicine, Division of Cardiology, NYU Langone Medical Center, New York, NY. Purpose: Driveline infections (DLI) are morbid complications following LVAD implantation. Chlorhexidine-gluconate (CHG) is commonly used for driveline antisepsis, but select patients have contraindications to CHG and require an alternative antiseptic. We compared DLI rates in patients using CHG versus betadine for exit site antisepsis. Methods: Between October 2011 and August 2015, 38 patients underwent primary LVAD implantation (37 HeartMate II, 1 HVAD). Driveline care was performed daily until exit site healing, and then reduced to three times weekly. Betadine was used in CHG-intolerant patients. DLI was defined according to INTERMACS criteria. Follow-up was censored by DLI occurrence, transplantation, or death. Comparisons were performed with Fisher’s exact and t-test where appropriate and time-related incidence of DLI was analyzed with the Kaplan-Meier method and log-rank test. Results: Of the 38 patients, 32 had CHG and 6 had betadine antisepsis (betadine indications: anaphylaxis, n= 1; contact dermatitis, n= 5). Four developed DLI (CHG, n= 1; betadine, n= 3) at a median of 217.5 days post-implant (IQR= 74.6-360.4). Total DLI event rate per patient year was 0.07. Causative organisms were S. aureus for betadine cases and Acenitobacter for CHG. CHG patients had significantly fewer DLI compared to betadine patients (3.1% vs. 50%, p= 0.009) and crosstab analyses showed no significant impact on DLI by race, age, gender, ischemic etiology, presence of caregiver, or INTERMACS profile. Kaplan-Meier curves demonstrated increased infection-free days in the CHG group (p= 0.001) (Figure 1). Conclusion: Betadine can be used for driveline antisepsis in VAD patients with CHG intolerance. However, given the higher risk of DLI with betadine, clinicians should be wary when choosing it as an alternative. These findings suggest the importance of larger scale studies comparing the risks and benefits of antiseptics for driveline exit site care, particularly for patients with CHG contraindications.

6( 97) Effect of Tunneling Method on Driveline Infection: Looking Beyond the Silicone-Skin Interface (SSI) Registry A.Y. Son ,1 L.H. Stein,1 A. DeAnda Jr.,1 D.E. Smith,1 S.D. Katz,2 A. Reyentovich,2 L.B. Balsam.1  1Cardiothoracic Surgery, NYU Langone Medical Center, New York, NY; 2Medicine, Division of Cardiology, NYU Langone Medical Center, New York, NY. Purpose: Newer driveline (DL) tunneling techniques have reduced driveline infection (DLI) rates by burying the velour-covered DL to create a siliconeto-skin interface. The velour portion may be buried in several ways, but it is unknown which method is superior. We compared DLI rates with three such tunneling methods following HeartMate II implantation. Methods: Between October 2011 and August 2015, 37 patients underwent primary LVAD implantation with HMII. DLs were buried using internal looping below the pump and right-sided exit (IL), curvilinear tunneling through the abdominal wall and left-sided exit (cTUN), or combination of IL and cTUN (IL-cTUN). Clinical factors (abdominal hernia or planned abdominal surgery) and patient preference (side of exit) influenced the burying method. DLI was defined according to INTERMACS criteria. Follow-up was censored by DLI occurrence, device exchange, transplantation, or death. Comparisons were performed with chi-square test and t-test or Mann-Whitney test where appropriate. Kaplan-Meier (KM) curves and log-rank test were used to evaluate time to DLI. Results: Of 37 patients, 8 had IL-cTUN, 20 had cTUN, and 9 had IL. Four developed DLI (IL-cTUN, n= 3; cTUN, n= 1) at a median of 217.5 days post-implant (IQR= 74.6-360.4). Total DLI event rate per year was 0.075. IL-cTUN had more DLI compared to cTUN and IL (37.5% vs. 5.0% vs. 0%, respectively, Χ 2= 7.701, p= 0.021) and crosstab analyses showed no significant impact on DLI or tunneling method by demographics, ischemic etiology, or INTERMACS profile. BMI and date of implant also showed no impact. KM curves demonstrated decreased infection-free days in the IL-cTUN group (p= 0.036) (Figure 1). Conclusion: Even when DL velour is buried, DLI may occur, particularly when the IL-cTUN technique is used. Our study extends beyond the findings of the SSI registry, which only compared IL and cTUN techniques. Larger scale studies are needed to understand the relationship between various methods of burying DL velour and DLI.

6( 98) Driveline Angle Is Crucial to Prevent Exit Site Infection in Patients with HeartMate II Y. Matsumoto ,1 T. Fujita,1 H. Hata,1 Y. Shimahara,1 S. Nakajima,2 T. Sato,2 O. Seguchi,2 M. Yanase,2 N. Fukushima,2 T. Nakatani,2 J. Kobayashi.1  1Cardiovascular Surgery, National Cerebral and Cardiovascular Center, Osaka, Suita, Japan; 2Transplantation, National Cerebral and Cardiovascular Center, Osaka, Suita, Japan. Purpose: Driveline infection (DLI) is a major complication after left ventricular assist device (LVAD) implantation which causes repeat hospitaliza-