The use of mechanical circulatory support in terminal heart failure patients has increased in recent years due to general shortage of donor organs, particularly in Europe. Conventional sternotomy is considered as standard approach for left ventricular assist device (LVAD) implantation. Several studies have shown the potential advantage of less invasive surgery (LIS) for LVAD implantation. This study aimed to determine the impact of LIS approach on stroke events.

Methods

Between January 2015 and March 2021, 335 consecutive patients underwent LVAD implantation using either conventional median sternotomy (CS) or LIS technique. The LIS implantation was performed using a upper partial  J-shaped sternotomy and left anterior mini-thoracotomy. Primary outcome was freedom from stroke. Secondary outcome was survival to discharge and during follow-up. Stroke was defined according to INTERMACS Protocol. Patients operated using CS approach were compared with patients undergoing LIS approach. Because these 2 groups differed in baseline parameters a propensity score analysis was computed.

Results

Study population and preoperative characteristic

A total of 242 patients (F=32; 13.0%) underwent LVAD implantation with CS and 93 patients (F=8; 8.6%) using LIS approach. The median age of LIS group was 63 years (IQR 22-66) vs. 61 years (IQR 53-66) in CS group (p=0.4). There was also no significant difference in body mass index, sex, history of diabetes, peripheral arterial disease, chronic kidney disease, history of hypertension and type of cardiomyopathy (Table 1). Median follow-up (until death, censoring for transplant/LVAD removal, or end of follow up) did not differ between the groups (612 [IQR, 160-1172] days in CS group vs 463 [IQR, 204-943] days in LIS group, p=0.4).

Overall cohort

Overall stroke rate, regardless of implanted device, was 13.7% (46/335). When comparing CS and LIS patients, in the unmatched cohort, stroke was more frequent in the LIS group (19% vs 12%, p=0.06) (Table 2). The reexploration for bleeding was necessary in 45 (19%) patients in the CS group compared with 7 (7.5%) in the LIS group (p=0.01). Postoperative RVAD use was also significantly higher in the CS group (23% in CS group vs 9.7% in LIS group, p<0.01). The duration of ICU stay for the LIS group was significantly lower than in the CS group (2 [IQR, 2-4] days vs 4 [IQR, 2-11] days, respectively, p< 0.001). All-cause hospital mortality was significantly higher in the CS group than in the LIS group (17% vs 6.5%, p=0.01). Mortality cumulative at 1 year was also significantly higher in the CS group (27.3%).

Propensity-matched cohort

Propensity matching generated 2 groups, including 98 patients in CS group and 67 in LIS group. Intensive care unit stay for the LIS group was significantly shorter than in the CS group (2 [IQR, 2-5] days vs 4 [IQR, 2-12] days, p<0.01). There were no significant differences in the incidence of stroke events (14% in CS vs 16% in LIS group; p=0.6) between the groups. Hospital mortality in the matched cohort was significantly lower in the LIS group (7.5% vs 19%; p=0.03). However, 1-year mortality showed no significant difference between both groups (24.5% in CS and 17.9% in LIS group; p=0.35) (Table 2).

Conclusions

LIS approach for LVAD implantation is a safe procedure with potential advantage in the early postoperative period. However, postoperative stroke and outcome remains comparable to sternotomy approach.