Who were enrolled in the EMPA-Kidney trial?

• A. Patients with CKD and T2DM
• B. Patients with a broad range of CKD with and without T2DM
• C. Patients with macro-albuminuria (uACR >300 mg/g)
• D. Patients with low eGFR (<45 mL/min/1.73m2)

Answer B is correct

Hi. I'm Will Herrington and it's a pleasure to present the results of the EMPA-KIDNEY Trial on behalf of its collaborative group. EMPA-KIDNEY was initiated by the University of Oxford who led its design, analysis and reporting with a steering committee of international experts. It was initiated in 2016 following reviewing the results of the renal analyses of the EMPA-REG OUTCOME trial. Its population-based design makes it applicable despite the fact there have been 11 other SGLT2 Inhibitor trials reported since. It was designed to assess the effects of an SGLT2 Inhibitor in a broad range of patients with chronic kidney disease at risk of progression, and we wanted to include at least a third with diabetes and at least a third without.

We identified them with simple criteria based on eGFR and albuminuria. To be eligible, you're required to have an eGFR between 20 and 45 at screening and historically or an eGFR between 45 and 90 with some evidence of albuminuria putting you at risk. We only excluded patients with polycystic kidney disease or kidney transplant, all other types of kidney disease were eligible.

Participants were required to be on an investigator-judged, clinically appropriate dose of a renin angiotensin system inhibitor when it was indicated and tolerated. We had about 15% who weren't on a RAS inhibitor and they were then randomized one-to-one to empagliflozin 10 milligrams once daily versus a matching placebo and it was event-driven. It was event-driven until, we had a set number of primary outcomes, which were cardiovascular death or kidney disease progression in a composite. Kidney disease progression itself was divided up into a clinical component of the need to initiate maintenance dialysis or receive a kidney transplant or death from kidney failure. Then a biochemical component which was based on eGFR changes, which were required to be sustained. That was either a 40% decline or more from the randomization value or an eGFR less than 10.

Here is the result of the primary composite outcome. Among those allocated to placebo, there were 558 primary outcomes, 432 amongst those allocated to empagliflozin representing a highly statistically significant reduction in the risk of the primary outcome by 28% and a 95% competence interval between 18% and 36%. Of these 990 outcomes, the vast majority, 888, were kidney disease progression outcomes, and you can see the results of kidneys disease progression alone on the left-hand side of this slide. Consequently, the numbers of cardiovascular deaths were actually very low within the composite, lower than we are expecting, and there wasn't a clear result because of the low power for this outcome.

The point estimate of a 16% risk reduction is consistent with the totality of the evidence. When it's combined with the need to start a maintenance dialysis or receiver kidney transplant, there was a very clear 27% risk reduction on these hard clinical components of the composite.

We predefined only three key subgroups for emphasis and analysis because they were scientifically most relevant at the time we were designing the trial, and the first was by diabetes status. You can see the overall result of the trial, the 28% risk reduction in the diamond, and you can see plotted above it, those with and without diabetes. You can see that there are over 450 outcomes in people without diabetes in this trial, a really important new amount of evidence and there was no evidence of effect modification, a heterogeneity P value of 0.06. The overall result is a best estimate in people with and without diabetes.

Here is the second subgroup by kidney function. You can see that there's a large amount of data, there are over 500 events in those with an eGFR less than 30 and no evidence at all of any difference of effect by baseline kidney function. In fact, here's a post-doc analysis. When we looked at the 254 patients with an eGFR between 15 and 20 by the time they were randomized, and you can see that there isn't at all any evidence of a threshold below which you don't get these benefits on the primary outcome. At least within the range we've studied and we've studied almost down to the point that people have kidney failure, so they reduce risk and it's irrespective of level of eGFR.

This is a key subgroup, which has attracted a lot of attention, which is the final key subgroup, and it's by baseline level of albuminuria. What we can see here is that we have evidence for a trend, a P value of 0.02 suggesting differences across albuminuria and perhaps a smaller effect in people with the lowest levels. But because those with the lowest levels are progressing more slowly, we only have 84 primary outcomes, a very wide composite interval, which includes 28%. So as a steering committee, we elected to bring forward a series of exploratory analyses based on change of eGFR slope at the time.

Here are some of those results. What you can see here is the annual rate of change of eGFR in the placebo group, which was 2.75 ml per minute per year. We saw the expected drop in eGFR, that reversible dip on a commencing empagliflozin, followed by nearly a halving of the rate of decline in those allocated to empagliflozin, down to 1.37 ml per minute per year on average. That was the difference as well, 1.37 ml per minute per year. That result here is now plotted in a diamond, 1.37 ml per minute. Absolute difference. What you can see in the box is the rate of decline amongst those allocated to placebo group, ranging from just under 1 ml per minute per year in those with A1 levels, up to over 4 in those with A3 levels of albuminuria. Here you can see the results of the empagliflozin arm, and the difference between the two. You can see there was a clear slowing of the rate of decline irrespective of baseline albuminuria. Those that had A1 levels, slowed by 0.78 ml per minute per year, and those with A3 levels, at 1.76. What we can see here is that empagliflozin appears to be slowing the rate of progression irrespective of the level of kidney function, and this much more sensitive outcome is identifying that for us.

EMPA-KIDNEY concentrated on kidney disease progression, but we did pre-specify key secondary outcomes based on cardiovascular events and hospitalization. I'll draw your attention to the all-cause hospitalization result here. We did identify 14% risk reduction, and this was true across all of the major terms which defined hospitalizations, including procedures, vascular events, infections. It was a general effect, no specific outcome that high-drove this result.

Then our safety outcomes, empagliflozin in CKD was shown to be similarly safe as it is in other types of patients treated with this class of drugs. We did see six versus one ketoacidosis events. We know that these drug class increases the risk of glycemic ketoacidosis and ketoacidosis, but we didn't see large numbers, and the risk was low in CKD. For the other outcomes, there were broadly similar results, and you can see that hyperkalemia was non-significantly reduced by 17%, and acute kidney injury by 22%. Those remarkably line up, although non-significant in this trial, with the estimates you see from the totality of the other trials, which we have meta analyzed and published separately in The Lancet.

Final conclusions from the EMPA-KIDNEY trial, is that we randomized 6,609 participants, they had CKD with a broad range of causes, and we had large numbers with the types of patients understudied in previous trials, some with low levels of kidney function, some with low levels of albuminuria. Overall, we found that empagliflozin safely reduced the composite primary outcome of kidney disease progression or cardiovascular death by 28%. Those relative benefits were consistent, broadly speaking, in patients with and without diabetes and across the full range of eGFR studied, at least down to 20, with no evidence of any attenuation of effect by eGFR. We did see some effect modification by level of albuminuria, but we brought forward exploratory analyses of eGFR slope and the chronic eGFR decline analysis suggests there would still be benefits in all albuminuria subgroups, even those that are progressing more slowly, provided that they were treated for long enough.

Full details are available in our publication in New England Journal of Medicine. Please do visit our website where there are copies of these slides and a series of other interesting information for your patients and for yourselves. Thank you to the collaboration.

Share this page with your colleagues and friends: