Flap Coverage of Infected Ventricular Assist Devices Influences Patient Outcomes

The use of left ventricular assist devices (LVADs) for patients with end-stage cardiac failure awaiting heart transplantation has become increasingly common. However, ventricular assist device-related infections remain a major problem complicating their long-term use. Poor data exist to determine how to manage these infections after operative debridement.

Patients who underwent insertion of a ventricular assist device and had a subsequent readmission for LVAD infection at the University of Rochester Medical Center from 2012 to 2022 were identified through accessing the medical records archives of the hospital. Patients were followed retrospectively for an average of 3.2 years. Patient demographics, preoperative diagnosis/disease state, type of ventricular assist device inserted, postoperative day of ventricular assist device infection onset, infectious organism identified at initial washout, infectious organism identified at time of definitive device coverage, timing of coverage procedure after the initial washout for infection, type of flap used for coverage, 90-day complications after definitive coverage, and lifetime return to operating room for infection were reviewed. Comparison analysis with a χ 2 test was used to analyze outcomes.

Of 568 patients admitted with an LVAD-related infection 117 underwent operative debridement. Of these, 34 underwent primary closure, 31 underwent closure with secondary intention (negative pressure wound therapy with split thickness skin grafting), and 52 were closed with a flap (pectoralis, omental, latissimus, or vertical rectus abdominus musculocutaneous flap). There was a statistically significant higher incidence of return to the operating room (RTOR) for infection over a lifetime with primary closure compared with secondary intention and flap reconstruction ( P = 0.01, 0.02), but no difference in 90-day complications ( P = 0.76, P = 0.58). Eighty-three patients had a positive culture upon definitive coverage with 24 having a postsurgical complication, 15 of which required lifetime RTOR for infection. Thirty four were closed with negative cultures with 9 having a complication and 4 requiring RTOR for infection. This was not statistically significant for complications or RTOR ( P = 0.79, 0.40). Culture data were further substratified into bacterial cultures (n = 73) versus fungal cultures (n = 10), and there was no statistically significant difference between these compared with complications or RTOR ( P = 0.40, 0.39).

Coverage of infected LVADs with locoregional flaps or allowing to granulate using wound vac therapy has a decreased lifetime RTOR for future infections for these patients without increase in 90-day complications. Timing of RTOR should not be impacted by positive cultures provided there is healthy granulation tissue in the wound.