Background: Glucose variability has been associated with cardiovascular (CV) outcomes in type 2 diabetes, however, the interplay between glucose variability, empagliflozin and CV death has not been explored. In the EMPA-REG OUTCOME trial, empagliflozin reduced the risk of CV death by 38%. We explore post-hoc the association between HbA1c variability and CV death, and the potential mediating effects of HbA1c variability on empagliflozin’s CV death reductions.Methods: In total, 7,020 patients with type 2 diabetes and established CV disease received placebo, empagliflozin 10 mg or 25 mg. We defined within-patient HbA1c variability as standard deviation, coefficient of variation and range of HbA1c measurements (%) post-baseline. First, we compared HbA1c variability until week 28 and 52 by Wilcoxon tests. We explored the association between CV death and HbA1c variability in placebo and pooled empagliflozin arms separately with landmark analyses at week 28 and 52, and additionally with HbA1c variability as a time-dependent co-variate. We used Cox regression models adjusted for baseline risk factors including changes in HbA1c from baseline to week 12, and the interaction term HbA1c variability* treatment.Results: HbA1c variability was lower with empagliflozin compared to placebo. In all Cox analyses, high HbA1c variability increased the risk for CV death in both treatment arms with no interaction with treatment.Conclusions: HbA1c variability was reduced by empagliflozin and high values of HbA1c variability were associated with an increased risk of CV death. Empagliflozin’s reduction in CV death did not appear to be mediated by reductions in HbA1c variability.ClinicalTrials.gov number, NCT01131676. Registered May 27 2010. https://clinicaltrials.gov/ct2/show/NCT01131676