Less Invasive Ventricular Assist Device Implantation Lowers Rethoracotomy Rate for Bleeding and Reduces Postoperative Hospital Stay: Multicenter Experience

Objectives: Avoiding full sternotomy may reduce the invasiveness of left ventricular assist device (LVAD) implantation. Studies comparing sternotomy to less-invasive (LIS) approach is limited to anecdotal case reports or only few patients. The aim of this study was to compare the outcome between these two surgical approaches.

Methods: Data of two high volume VAD centers were collected and analyzed. Inclusion criteria were patients supported with LVAD between January 2014 and December 2018 using less invasive (LIS) or full sternotomy approach. The LIS and Sternotomy approaches were compared for preoperative characteristics/hemodynamic profile and postoperative hemodynamic profile, ICU/hospital stay, adverse event rates, and overall survival.

Results: Out of 342 implanted VADs during this period, LIS approach was used in 101 patients (30%). The LIS approach was feasible in all patients with no need for conversion. The preoperative characteristics from both groups were identical except for more patients on venoarterial extra corporeal membrane oxygenation (24 vs. 10%, p = 0.003), lower INTERMACS profile (p < 0.001) and more redo surgeries (p = 0.06) in sternotomy patients. Propensity score matching was performed to match for these parameters and resulted in 100 patients remaining in each group and were compared. Mild and moderate RV failure rates were comparable. However, higher rate of severe RVF in sternotomy group was documented (27 vs. 11% p = 0.02). Meanwhile, the acute severe RVF requiring right ventricular assist device was comparable (15% in sternotomy vs. 8% in LIS group, p = 0.2). Stroke, pump thrombosis, GI bleeding, and driveline infection rates were comparable between both groups (p = 0.3, 0.9, 0.4, and 0.5, respectively). The duration of intensive care unit was comparable (p = 0.2). The duration of hospital stay was significantly longer in sternotomy group (34 days [25–57] vs. 27 days [22–42] in LIS group, p = 0.009). The short and long-term survival remains comparable between the groups (p = 0.93).

Conclusion: In this matched group of patients, LIS approach seems to result in lower postoperative morbidities (lower reexploration rate for bleeding and severe RV failure) and shorter hospital stay. Nevertheless, the survival remains similar between the groups.

No conflict of interest has been declared by the author(s).