A nationwide sample of patients, aged 18 years or over, with type 2 diabetes drawn from a large anonymised US commercial healthcare database was studied.

The three arms of the trial compared cohorts of patients who were started on canagliflozin with patients started on other diabetes therapies.

• The first cohort compared 21,431 patients on canagliflozin vs 77,463 on a DPP-4i

• The second compared 25,806 patients on canagliflozin vs 32,676 on a GLP-IRA

• The third compared 18,924 patients on canagliflozin vs 115,435 on a sulphonylurea.

The main outcome measures were hospital admission for heart failure, and a composite cardiovascular endpoint of hospital admission for acute myocardial infarction, or ischaemic or haemorrhagic stroke.

The canagliflozin arms showed a reduction in heart failure admissions by 30-49%, and no difference in outcomes for the composite cardiovascular endpoint.

The hazard ratios (HR) for heart failure admission in the 30-month follow-up period were:

• 0.70 (95% CI: 0.54-0.92) for canagliflozin vs a DPP4i

• 0.61 (95% CI: 0.47-0.78) for canagliflozin vs a GLP-1RA

• 0.51 (95% CI: 0.38-0.67) for canagliflozin vs a sulphonylurea

The HR values for the composite cardiovascular endpoint were:

• 0.89 (95% CI: 0.68-1.17) for canagliflozin vs a DPP4i

• 1.03 (95% CI: 0.79-1.35) for canagliflozin vs a GLP-1RA

• 0.86 (95% CI: 0.65-1.13) for canagliflozin vs a sulphonylurea

The reduction in admissions for heart failure duplicates the results of two recent large placebo controlled randomised trials CANVAS (canagliflozin) and EMPA-REG OUTCOME (empagliflozin).^1,2

Current UK guidance still recommends low cost hypoglycaemic agents such as sulphonylureas despite adverse cardiovascular outcome data.

These findings will help inform new diabetes guidance, and in the meantime may help GPs who are selecting hypoglycaemic agents for their patients.

Dr Matthew Lockyer

Patorno E, Goldfine AB, Schneeweiss S et al. Cardiovascular outcomes associated with canagliflozin versus other non-gliflozin antidiabetic drugs: population based cohort study. BMJ 2018;360:k119