139-LB: The Redox Balance Predicts Both Cardiovascular (CV) and Renal Outcomes in CANVAS

Abstract

We tested whether circulating metabolites reflecting cytosolic (lactate/pyruvate) or mitochondrial (β-hydroxybutyrate/acetoacetate [AcAc]) redox balance are associated with major clinical outcomes in patients with type 2 diabetes.

The CANVAS study randomized participants to placebo (Plb) or canagliflozin (100mg or 300mg) (Tx) and followed them for up to 6 years. Metabolic markers were measured at baseline (2,373 in Tx [dose combined] and 1,185 in Plb), at year 1 (2,022 in Tx and 949 in Plb) and year 2 (1,793 in Tx and 834 in Plb). By linear mixed model analysis, plasma free fatty acids, AcAc, and β-hydroxybutyrate increased with Tx vs. Plb, whereas plasma pyruvate decreased with Tx while glycerol and lactate did not change significantly. Of note, these changes were independent of the background antihyperglycemic medication (metformin, sulfonylureas, insulin, or combinations thereof). The incidence of hospitalized heart failure or cardiovascular death (HHF/CVD) was 10.2%, the incidence of the renal composite endpoint (CombR; i.e., 40% reduction in eGFR or renal replacement/death) was 3.9%. In Cox models adjusting for sex, age, BMI, HbA_1c, baseline systolic blood pressure, diabetes duration, smoking, baseline use of diuretics or statins, history of heart failure, prior CV disease, and Tx, baseline AcAc was a negative predictor of HHF/CVD (HR=0.83 [95%CI 0.75-0.94]), while baseline lactate was a positive predictor (HR=1.42 [95%CI 1.20-1.68]). Adding baseline ACR and eGFR to the same multivariate model for CombR as the outcome, AcAc was still a negative (HR=0.84 [95%CI 0.75-0.94]) and lactate still a positive predictor (HR=1.42 [95%CI 1.20-1.67]).

For CV and renal outcomes, higher baseline plasma lactate is an independent positive risk marker potentially reflecting a reducing cytosolic redox shift, but higher baseline plasma AcAc is a protective factor, which may be related to efficient mitochondrial supply/usage of reducing equivalents. Further analysis is needed to test these hypotheses.