LAUSR

Introduction Serum sodium is an important prognostic marker in heart failure patients, with lower values associated with increased risk of mortality. However, prognostic value of serum sodium has not been assessed in left ventricular assist device (LVAD) patients. Hypothesis We hypothesized lower serum sodium is associated with increased risk of mortality in LVAD patients. Methods We retrospectively identified 253 consecutive patients that had LVAD implantation at a single center between 2008-2016. To minimize confounding by peri-operative factors, we evaluated serum sodium at 3 months after implantation and time to all-cause mortality with a univariate Cox proportional hazards analysis. A secondary outcome was time to hospital readmission. Groups were defined as having lower and higher sodium levels relative to the median value. Censoring occurred at death, transplant, or pump exchange. A multivariate Cox proportional hazards analysis included estimated GFR and bicarbonate levels to account for effects of renal dysfunction and acid-base disturbance on sodium levels. Results There were 195 eligible LVAD patients with a sodium value at 3 months. Average age was 53 ± 14 yrs, 77% were male, 26% black. Half (50%) received a Heartmate II device and 50% received a Heartware HVAD. Median serum sodium was 137 mmol/L. Median follow-up was 346 days with 26 deaths (13%). In a univariate analysis, increasing serum sodium at 3 months was associated with a lower risk of all-cause mortality (HR 0.826, p=0.006). Pre-operative sodium levels were not associated with mortality (HR 1.004, p=0.941). Serum sodium at 3 months remained predictive of mortality in a multivariate model with estimated GFR and HCO3 (Table). When stratified by the median, the lower sodium group had a significantly shorter time to all-cause mortality (Figure). Sodium levels at 3 months were not associated with time to readmission. Conclusions Hyponatremia in LVAD patients (sodium