The cardiorenal benefits of SGLT2 inhibition are established, but clinicians treating older patients reasonably ask whether the trial results apply to them, and what the absolute return looks like at the age when competing mortality starts to matter. This individual participant data meta-analysis addressed that using harmonised subject-level datasets from CANVAS and CREDENCE, obtained through the Yale University Open Data Access project, restricted to participants aged 50 or over — 3,955 from CANVAS and 4,035 from CREDENCE.
For end-stage kidney disease, across 175 events, the pooled hazard ratio was 0.535 (95% CI 0.392 to 0.729) — a 46.5% relative reduction. The trial-level results differed considerably in precision: CANVAS gave 0.853 with an interval of 0.250 to 2.916, essentially uninformative, while CREDENCE gave 0.518 (0.376 to 0.714), which is where the pooled estimate comes from. The number needed to treat over three years in CREDENCE was 27.5.
Cardiovascular mortality, across 309 events, gave a pooled hazard ratio of 0.841 (0.669 to 1.058) — favourable in direction, not statistically significant, and the authors attribute that to limited power rather than absence of effect.
The practical value here is the absolute number. Twenty-seven and a half patients treated for three years to prevent one case of end-stage kidney disease, in adults over 50, is a figure worth carrying into a consultation and into a formulary discussion. In Indian practice, where dialysis access is limited and self-funded for most patients, preventing one case of kidney failure per 27 patients treated is an argument that lands differently than it does where dialysis is universally provided.
Read it for what it is. This is a re-analysis of two completed trials by an independent investigator using open data, not new evidence — it confirms and quantifies rather than discovers, and the end-stage kidney disease finding is essentially the CREDENCE result restated for an older subgroup. That is still useful: CREDENCE enrolled patients with albuminuric diabetic kidney disease, and knowing the effect holds in the over-50s within it supports treating the patients you actually see.
- Quote the absolute figure: about 27 patients treated for 3 years to prevent one case of end-stage kidney disease.
- The effect holds in adults over 50, so age alone is not a reason to withhold.
- The pooled estimate is driven almost entirely by CREDENCE; the CANVAS contribution is uninformative.
- Cardiovascular mortality pointed the right way but was underpowered — do not claim a mortality benefit from this.
- This is a re-analysis of existing trials confirming known effects, not new evidence.
The statistics, in plain English
The two trials contributed very unequally, and it is worth seeing why. CANVAS produced a hazard ratio of 0.853 with an interval running from 0.250 to 2.916 — a range so wide it is compatible with a large benefit or a large harm, because CANVAS enrolled patients at lower kidney risk and accumulated very few end-stage events. CREDENCE, designed as a kidney outcome trial in albuminuric disease, supplied nearly all the information. So the pooled figure of 0.535 is really the CREDENCE result, and it applies best to patients resembling CREDENCE participants. The number needed to treat of 27.5 is quoted from CREDENCE alone for the same reason.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for nephrology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free