Driveline infections (DLIs) remain a major source of morbidity for patients requiring long-term ventricular assist device (VAD) support. We aimed to assess whether VAD driveline exit site (DLES) (abdomen versus chest wall) is associated with DLI. All adult patients who underwent insertion of a HeartWare HVAD or HeartMate II (HMII) between 2009 and 2016 were included. Driveline infection was defined as clinical evidence of DLI accompanied by a positive bacterial swab and need for antibiotics. Competing risks analysis was used to assess the association between patient characteristics and DLI. Ninety-two devices (59 HMII) were implanted in 85 patients (72 men; median age 57.4 years) for bridge to transplant or destination therapy. VAD DLES was chest in 28 (30.4%) devices. Median time on VAD support was 347.5 days (IQR 145.5, 757.5), with 28 transplants and 29 deaths (27 on device). DLI occurred in 24 patients (25 devices) at a median of 140 days (IQR 67, 314) from implant. Staphylococcus aureus accounted for 15 infections (60%). Freedom from infection was 72.8% (95% confidence interval [CI] 53.1–78.0%) at 1 year and 41.9% (95% CI 21.1–61.5%) at 3 years. In competing risks regression, abdominal DLES was not predictive of DLI (hazard ratio, HR 1.65 [95% CI 0.63, 4.29]), but body mass index (BMI) >30 kg/m2 was (HR 2.72 [95% CI 1.25, 5.92]). In conclusion, risk of DLI is high among patients on long-term VAD support, and a nonabdominal DLES does not reduce this risk. The only predictor of DLI in this series was an elevated BMI.
From the *Division of Cardiothoracic Surgery, Stanford University, Stanford, California
†Division of Cardiac Surgery, University of Alberta, Edmonton, Canada
‡Faculty of Medicine and Dentistry, University of Alberta, Edmonton, Canada
§Division of Critical Care, University of Alberta, Edmonton, Canada
¶Department of Pediatrics, University of Alberta, Edmonton, Canada.
Submitted for consideration June 2017; accepted for publication in revised form August 2017.
Disclosure: The authors have no conflicts of interest to report.
Correspondence: Jennifer Conway, MD, Department of Pediatrics, University of Alberta 8440 112 Street NW, Edmonton AB T5G-2B7, Canada. Email: email@example.com.