SGLT2 inhibitors and clinical outcomes in LVAD patients – a retrospective cohort study

Purpose

An impact of sodium glucose transporter 2 inhibitors (SGLT2-i) on mortality and hospitalizations in patients supported with left ventricular assist device (LVAD) is unknown.

Methods

We retrospectively evaluated 403 consecutive patients who underwent LVAD implantation since 1/3/2018 till 12/31/2021 in the single center and received SGLT2-i therapy following surgery. Patients with less than six months follow up and patients supported with chronic hemodialysis were excluded. All hospitalizations and mortality events were recorded. Primary end-point was defined as all-cause mortality and a secondary end-point as time to first hospitalization following LVAD implantation. Multivariable cox regression model was applied with SGLT2-I treatment as a time-dependent covariate and heart transplantation and LVAD explanation as a competing risk events. We performed propensity score matched analysis (PSMA) and matched patients with SGLT2-I in 1:2 fashion with patients without such a therapy. Age, gender, diabetes mellitus (DM), LVAD type, chronic kidney disease, body mass index and INTERMACS status prior to LVAD implantation were incorporated into the multivariable model and in the PSMA.

Results

SGLT2-i were initiated in 44 (16.5%) out of 266 patients that were included in the final analysis (32 (73%) received Dapagliflozin and 12 (27%) Empagliflozin).  Overall, 147 (55%) were implanted with Heartware (Medtronic, Minneapolis, MN, USA) and 119 (45%) with HeartMate 3 (Abbott, Chicago, IL, USA) LVAD. The median age at the time of LVAD implantation was 59 (IQR 52-64), 226 (85%) were male and 81 (30.5%) had DM. During the median follow up time of 812 days (IQR 432-1123), 65 (24.4%) patients died, 27 (10.2%) underwent heart transplantation and 9 (3.4%) underwent LVAD explanation. Treatment with SGLT2-I was associated with non-statistically significant 70% lower risk of mortality (95% CI 0.07-1.16, p value=0.098) in multivariable model. Following propensity score matching this association became statistically significant (HR= 0.22, 95% CI 0.047-0.989, p value=0.048). No differences were detected in time to first all cause hospitalization (p value=0.993), heart failure hospitalization (p value=0.264), hospitalization for bleeding (p value 0.251), driveline infection (p value 0.337) or arrhythmic event (p value 0.971) with SGLT2-I as compared with patients without such a treatment in multivariable model.

Conclusions

SGLT2-i treatment is associated with reduction in all-cause mortality rate in LVAD patients following propensity score matching.