Endotrophin is a risk marker of complications in CANagliflozin cardioVascular Assessment Study (CANVAS): a randomized controlled trial

Abstract

Abstract Background Enhanced de-novo collagen type VI (COL VI) formation has been associated with kidney and cardiovascular fibrosis. We hypothesized that endotrophin (ETP), a product specifically generated during collagen type VI formation, may be prognostic for heart failure (HF), cardiovascular death (CVD), kidney endpoints, and all-cause mortality in patients with type 2 diabetes. Methods We measured ETP in plasma (P-ETP) and urine (U-ETP) samples collected at baseline and follow-up (year 3) from the randomized controlled trial, CANagliflozin cardioVascular Assessment Study (CANVAS), by use of the PRO-C6 ELISA measuring COL VI formation and ETP. At baseline, plasma and urine samples were available for 3531 and 3423 patients, respectively. At year 3, plasma and urine samples were available for 2178 (61.7%) and 2070 (60.5%) patients, respectively Patients were followed for a median of 6.1 years, and endpoints included: incident HF, CVD, three kidney composite endpoints, and all-cause mortality. Backward selection was used to identify variables to be included in the analyses. Robustness of the association with outcome was assessed by bootstrap analyses. Results In univariable analysis, P-ETP predicted all investigated outcomes (all p < 0.0001), remained independently associated with all outcomes after adjustment for conventional risk factors (all p < 0.004), and increased C-statistics of the models for the outcomes HF, CVD, HFCVD, all-cause mortality, and kidney composite 2 (ΔC ≥ 0.002). In bootstrap analysis, P-ETP was retained with a frequency ranging from 41.0 to 98.4% for all outcomes. Levels of U-ETP were associated with outcomes in univariable analysis, but associations with most outcomes were lost after adjustment for conventional risk factors. The increase in P-ETP over time was greater with increasing albuminuria stage (p < 0.0001) and was independently associated with the kidney endpoints (p < 0.03). In the placebo arm, the increase in P-ETP was prognostic for all-cause mortality (HR [95% CI]; 1.14 [1.05–1.23], p = 0.003). Whereas levels of P-ETP were not impacted by treatment, levels of U-ETP significantly increased with canagliflozin treatment. Conclusions P-ETP generated during COL VI formation predicts cardiovascular, kidney and mortality outcomes in patients with type 2 diabetes. As ETP identifies patients at increased risk of experiencing relevant outcomes, it may be used for patient enrichment in future clinical trials. Trial Registry Number (ClinicalTrials.gov Identifier): NCT01032629