CREDENCE Takes SGLT2 Inhibitors 'to the Next Level'

Harpreet S. Bajaj, MD, MPH


April 25, 2019

The enormity of the need for better treatments for diabetic kidney disease is underscored by the fact that it has been almost 18 years since the last advancement in this area, with the advent of renin-angiotensin-aldosterone system (RAAS) blockers.

It's not surprising, then, that considerable excitement had built up in anticipation of results from the CREDENCE trial, which were presented recently in Melbourne, Australia, at the International Society of Nephrology congress and published simultaneously in the New England Journal of Medicine.[1] This trial shows an unequivocal benefit of canagliflozin in diabetic kidney disease, similar to the straightforward secondary prevention and cardiovascular risk reduction revealed for empagliflozin in the EMPA-REG OUTCOME trial more than 3 years ago.[2]

Although the composite renal endpoint reduction (a post hoc analysis at that time) was a pleasant surprise in EMPA-REG, nephroprotection was subsequently corroborated in the CANVAS program[3] with canagliflozin, and more recently in the DECLARE-TIMI 58 trial[4] with dapagliflozin, albeit as a secondary outcome.

Unequivocal Renal Benefits

Notably, CREDENCE randomly assigned participants to receive canagliflozin or placebo on top of RAAS blocker therapy (as a mandatory study inclusion criterion), and still observed a staggering 30% relative risk reduction in the primary composite outcome of end-stage kidney disease, doubling of serum creatinine, or renal or cardiovascular death.

In addition to the statistically significant primary outcome benefit, several clinically important renal outcomes, as well as the composite major acute cardiovascular events (nonfatal myocardial infarction, nonfatal stroke, and cardiovascular death), were reduced significantly.

In a subgroup analysis, the primary composite endpoint reduction appears to be at least as robust, if not more, in the 60% of trial participants who had an estimated glomerular filtration rate (eGFR) between 30 and 60 mL/min/1.73 m2.

Nephroprotection with sodium-glucose cotransporter-2 (SGLT2) inhibitors, especially at a low eGFR, is thought to be glucose independent. This is quite apparent in CREDENCE, where a minimal end-of-study A1c difference (0.1%) was observed between the trial arms in favor of canagliflozin. This concept of glucose-independent nephroprotection is further set to be tested in the ongoing Dapa-CKD trial and the upcoming EMPA-KIDNEY trial, both of which include a proportion of participants without diabetes.

Broadly, CREDENCE is yet another positive outcome trial—with renal endpoints reported as the primary objective for the first time for an SGLT2 inhibitor—that adds weight to the ongoing paradigm shift of clinical practice guidelines toward emphasizing reduction of complications as the primary goal of diabetes management rather than focusing solely on a glucose target approach.

Limitations to Consider

Let me point out two limitations that should be considered while interpreting the results from CREDENCE:

  1. Further review and analyses may be necessary to provide clarity on the 27% of participants who discontinued the study drug within the median 2.6 years of follow-up. This percentage appears to be comparatively worse in CREDENCE than in most recently completed outcome trials in diabetes, which have an approximately 5% drug discontinuation rate per year.

  2. As acknowledged in the publication, one of the statistical limitations of CREDENCE is that the trial was stopped early at a planned interim analysis, which may increase the risk of overestimating effect sizes. Notwithstanding, the extremely tight (ie, small magnitude and spread) number needed to treat of 22 (95% confidence interval, 15-38) for the primary outcome over 2.5 years appears to be staggeringly clinically significant.

What About the Amputation Risk?

Anticipation (and anxieties) were even greater for this randomized trial to confirm (or not) the amputation and fracture signals observed in the CANVAS program. A collective sigh of relief was felt in the Melbourne presentation hall when no such increase in adverse outcomes was reported.

The next pertinent question, then, is: How do we integrate the dissimilar amputation results in the cardiovascular versus kidney disease outcome trial programs of canagliflozin?

Here are a few speculative guesses:

Could differing populations have resulted in the observed amputation risk differential? Notably, unlike CANVAS, CREDENCE used only the lower dose of canagliflozin. Could this difference in study protocol, together with the less potent glucose lowering with canagliflozin at lower eGFR (and possibly less diuresis as a result), explain the discrepancy?

Could practice and trial protocol adaptations after the CANVAS amputation revelations have influenced the subsequent amputation risk differential among the trial arms in CREDENCE?

In addition to a more careful and comprehensive prospective assessment of foot complications (including full foot exams) mandated by a protocol amendment, investigators were recommended to interrupt study medication for any significant foot-related condition associated with amputations (including lower-extremity infection, skin ulcer, osteomyelitis, gangrene, and critical limb ischemia) until the condition resolves.

However, if indeed the above practice and protocol changes were effective, should we have expected a lower rate of amputations overall? Such a speculative explanation appears to fall flat if you consider the relatively high amputation rates observed in both trial arms (no statistical difference between an event rate of 12.3 in the canagliflozin group and 11.2 per 1000 person-years in the placebo group).

Could the amputation signal in CANVAS be a statistical anomaly (ie, a type 1 error)? That interpretation, though possible, should only be entertained cautiously if none of the above or any other plausible explanations provide clues to the singular adverse findings in CANVAS.

Speculations aside, we must remember that data are data, especially when the evidence is coming from a high-quality randomized trial such as CREDENCE. Hence, the correct interpretation of the nonobservance of an amputation signal in CREDENCE may be that in real-life clinical situations, when physicians choose patients similar to those enrolled in the trial and practice regular foot care and counseling, the risk (if any) for amputations may be mitigated.

Overall, CREDENCE takes the positive outcome evidence accruing for SGLT2 inhibitors in type 2 diabetes (since EMPA-REG OUTCOME in 2015) to the next level, and indeed may be considered a game-changing landmark trial for diabetic kidney disease, much like the pivotal EMPA-REG OUTCOME trial for secondary cardiovascular prevention.

Follow Medscape on Facebook, Twitter, Instagram, and YouTube


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.