Mediators between canagliflozin and renoprotection vary depending on patient characteristics: Insights from the CREDENCE trial

Aim: To identify the mediators between canagliflozin and renoprotection in patients with type 2 diabetes at a high risk of end-stage kidney disease (ESKD).Methods: In this post hoc analysis of the CREDENCE trial, the effect of canagliflozin on potential mediators (42 biomarkers) at 52 weeks and the association between changes in mediators and renal outcomes were evaluated using mixed-effects and Cox models, respectively. The renal outcome was a composite of ESKD, serum creatinine doubling or renal death. The percentage of the mediating effect of each significant mediator was calculated based on changes in the hazard ratios of canagliflozin after additional adjustment of the mediator.Results: Changes in haematocrit, haemoglobin, red blood cell (RBC) count and urinary albumin-to-creatinine ratio (UACR) at 52 weeks significantly mediated 47%, 41%, 40% and 29% risk reduction with canagliflozin, respectively. Further, 85% mediation was attributed to the combined effect of haematocrit and UACR. A large variation in mediating effects by haematocrit change existed among the subgroups, ranging from 17% in those patients with a UACR of more than 3000 mg/g to 63% in patients with a UACR of 3000 mg/g or less. In the subgroups with a UACR of more than 3000 mg/g, UACR change was the highest mediating factor (37%), driven by the strong association between UACR decline and renal risk reduction.Conclusions: The renoprotective effects of canagliflozin in patients at a high risk of ESKD can be significantly explained by changes in RBC variables and UACR. The complementary mediating effects of RBC variables and UACR may support the renoprotective effect of canagliflozin in different patient groups.