Abstract

Background and Aims: A number of significant positive and negative signals emerged from the CANVAS Program and CREDENCE trial with the use of canagliflozin. These signals are confusing. A Likelihood of being Helped of Harmed (LHH) analysis was conducted to determine the risk, benefit ratio associated with canagliflozin use and address the signals as a continuum.

Materials and Methods: LHH was calculated from the number needed to treat (NNT) and number needed to harm (NNH) available from the absolute risk reductions reported with the outcomes of interest, in these two trials.

Results: In the CANVAS Program, LHH for major adverse cardiovascular events (MACE) points at a significant benefit with canagliflozin use in comparison to amputation (1.65), fractures (1.65) and euglycaemic diabetic ketoacidosis (euDKA) (16.67) risks. Only genital fungal infections were significant more in both sexes (0.21-M and 0.1-F) when LHH was matched against the positive outcomes. In contrast, the hHF benefits were outweighed by amputation (0.95) and fracture risks (0.95).

In CREDENCE trial, the LHH for Primary composite, Renal composite and MACE, all supported the benefits in comparison to any adverse events encountered in the trial.

The LHH from pooled data (CANVAS Program and CREDENCE trial) was in favour of all the benefits (hHF and renal composites) except for MACE matched against amputation (0.66).

Conclusion: The outcome benefits were in favour of canagliflozin in comparison to all reported adverse events, when hHF and renal composite were under consideration, in both the individual and pooled LHH analysis. However, the MACE benefits were overwhelmed by amputation risk in the pooled analysis.

Keywords: Type 2 diabetes, canagliflozin, NNT, NNH, LHH, MACE, amputation.