Improved Ventricular Assist Device Outcomes Through Standardized Processes

Background:   Ventricular assist devices (VADs) are being used more frequently in patients with end-stage heart failure.  Infections remain a common cause of major morbidity and mortality.  From 2000-2006 (P1), our infection rate (IR) was 5.01 /1000 VAD days (d).  Process improvements implemented 2006-2008 (P2) resulted in a decreased IR to 1.2 /1000 VAD d, however, consistency of compliance was not completely achieved.

Objective: Determine if the VAD IR could be reduced in 2008-2010 (P3) through improved compliance with standardized pre, intra and post-operative processes.  

Methods:   Following IRB approval, infections in all patients (pt) with a VAD implanted in P2 (N=96) were compared to those implanted in P3 (N=116).  In P3, pt were followed for the duration of their device to determine infection incidence and risk factors.  Standardized pre-operative processes were: nasal screening for Staphylococcus aureus (SA) with mupirocin administration if culture was pending or SA +; 4% (CHG) bathing X 2 pre-op, pre-op wash with a 2% CHG disposable cloth.  Intra-operative processes included: antibiotic prophylaxis with vancomycin + aztreonam, +/- fluconazole (if hospitalized > 5 d) with timing, and specified dose/re-dose, skin prep with ChloraPrep®, insulin drip to keep glucose < 150.  Post-operative processes were: pt education, sterile dressing changes, and compliance with an abdominal binder to secure the driveline. Infections of the driveline, pocket, surgical site or primary VAD associated blood stream infection (VAD-BSI) were included.

Results: Infections developed in 32 pts (P3) vs. 21 pts (P2); median onset was 117 d, (range 5-647 d) vs. 89 d in P2 (p<0.05). IR decreased to 1.0 /1000 VAD d (P3=30,952 actual VAD d) vs. 1.2 / 1000VAD d (P2=17,535 actual VAD d). Compliance with P3 processes for 32 pts: nasal screenings (75%), results available pre-operatively (50%), mupirocin if + or pending: 8/11(73%), timely vanco (84%) and aztreonam (84%). Of 6 pts who developed a VAD infx within 30 d of implant (day 5, 10, 14, 18, 20, 26), there was no statistical significance in compliance with preoperative processes vs. the other 26 infected pts. There were 19 Gram positive infx: (17 staphylococcus, 2 enterococcus) and 12 Gram negative (7 P. aeruginosa, 2 Enterobacter sp.,1 Neisseria sp., 2 Klebsiella sp.) and 1 no culture.

Conclusions: Reductions of VAD infections and delay in the time to infection can be achieved through multi-disciplinary standardization of all operative processes. Peri-operative practices continue to represent the greatest opportunity for infection prevention; however, prevention in the now prolonged post discharge period is also essential and the most challenging.