How to Read the SGLT2 Inhibitor Kidney Trials

The SGLT2 inhibitor kidney trials are best read as a lesson in endpoint design, not as a slogan about a drug class. EMPA-KIDNEY and the chronic kidney disease trials that preceded it reported a single composite outcome that bundled several kidney events together with cardiovascular death, and they modeled the loss of kidney function as a slope rather than a single before-and-after number. Reading them well means asking what each component of that composite contributed, how the eGFR curve was split into an early dip and a later trajectory, and why the effect held in people who never had diabetes. Those questions, not the headline percentage, are where the evidence actually sits.

What the trials counted

EMPA-KIDNEY, published in the New England Journal of Medicine, randomly assigned 6,609 people with chronic kidney disease to empagliflozin or placebo and followed them for a median of about two years. Its primary outcome was a composite: kidney disease progression, defined as end-stage kidney disease, a sustained eGFR below 10, a sustained decline in eGFR of at least 40 percent, or death from kidney causes, together with death from cardiovascular causes. Empagliflozin lowered the risk of that composite by 28 percent, with a hazard ratio of 0.72 and a 95 percent confidence interval of 0.64 to 0.82.

A composite endpoint is a design choice with a trade-off. Pooling related events raises the number of outcomes, which buys statistical power and lets a trial finish in a reasonable time, and it captures a disease that can exit through several doors. The cost is interpretability. A composite is only as meaningful as its components, so the first move in reading one is to check which event drove the result and whether that event is a hard, patient-relevant outcome or a softer surrogate. In EMPA-KIDNEY the result was carried mainly by kidney disease progression, which is the outcome the trial existed to test.

The eGFR slope and the acute dip

The second thing these trials measured was the rate of kidney function change over time, the eGFR slope. A prespecified analysis from EMPA-KIDNEY split that slope into two parts: an acute slope from randomization to about two months, and a chronic slope from two months onward. The split matters because SGLT2 inhibitors cause an expected early drop in eGFR. That initial dip is hemodynamic, reflecting reduced pressure inside the glomerulus, and it reverses when the drug is stopped. It is not kidney injury.

Reading the slope this way resolves an apparent paradox. A drug can produce a small early decline in measured filtration and still preserve function over the long run, because the chronic slope, the part after the dip, is flatter on treatment. A reader who looked only at the first few weeks would misjudge the drug. This is also why slope-based and event-based endpoints can look different in the same trial, and why an early eGFR drop on this class should not be read as harm.

Why the benefit reached people without diabetes

The feature that made EMPA-KIDNEY notable is that 54 percent of its participants had no history of diabetes. Dapagliflozin's DAPA-CKD trial, reported in 2020, had already enrolled people with and without type 2 diabetes. The reason the benefit is not tied to diabetes is mechanistic: the protective effect appears to run through kidney blood-flow regulation rather than through glucose lowering, so it does not require a diabetic kidney to work.

The pooled evidence made this explicit. A collaborative meta-analysis of the major placebo-controlled trials, published in The Lancet in 2022, found that SGLT2 inhibitors reduced chronic kidney disease progression by about 37 percent, with similar relative risks in people with and without diabetes and an overall relative risk near 0.63. The executive summary of the KDIGO 2024 clinical practice guideline for chronic kidney disease, published in Kidney International, sets out a strong recommendation to use an SGLT2 inhibitor in adults with CKD and an eGFR of at least 20 who have sufficient albuminuria, including those without diabetes. That guideline language describes what the evidence supports at a population level; it is not a decision for any individual, and this article is educational and not medical advice.

Reading the long-term follow-up

The most recent chapter is the EMPA-KIDNEY post-trial follow-up, published in the New England Journal of Medicine in early 2025. After the study drug was stopped, the difference between the groups narrowed. Kidney disease progression occurred in about 23.5 percent of the original empagliflozin group versus 27.1 percent of the placebo group, and some benefit persisted for up to roughly a year after discontinuation, but it was smaller than during treatment and appeared temporary. The lesson for a reader is that an on-treatment result and an off-treatment durability result answer different questions. A function-preserving therapy that fades after withdrawal is telling you it needs to be continued, not that its earlier effect was illusory.

Putting the appraisal together

Read as a set, these trials reward a few habits. Look at which component moved the composite. Weigh the width of the confidence interval alongside the point estimate. Convert relative reductions into absolute ones for the population you care about. Note the enrollment thresholds, because an eGFR floor of 20 and an albuminuria requirement define who the result actually applies to. And separate the on-treatment slope from what happens after the drug stops. Done that way, the SGLT2 inhibitor kidney trials become a clean case study in how modern nephrology trials are built, and in how much of their meaning is held inside the endpoint rather than the abstract.