Adults with type 2 diabetes who had albuminuria, meaning protein in the urine, had a lower five year risk of losing kidney function if the drug they added after metformin was a GLP-1 receptor agonist or an SGLT2 inhibitor rather than an older DPP-4 pill, according to a US study of 75,455 people published by The BMJ on September 16, 2026. In the 82% without protein in the urine, the newer drugs showed no kidney edge.
The practical piece is the urine test itself. Albuminuria is measured with a cheap urine sample, and this study says that one result separates the people who stand to protect their kidneys from the people for whom the drug choice is close to a wash. If you have type 2 diabetes and are about to add a second drug, that number is what makes the conversation with your doctor real.
Key takeaways
- With protein in the urine, five year risk of kidney decline was 3.2% on a GLP-1 or SGLT2 drug, 4.6% on a sulfonylurea and 5.4% on a DPP-4 inhibitor.
- Without it, the drugs were close, but sulfonylureas carried about 31% higher risk than DPP-4 inhibitors.
- This is an observational records study, not a randomized trial, with follow up averaging 32 months.
What the study found
Researchers used health records and insurance claims for people with type 2 diabetes at moderate cardiovascular risk who started a GLP-1 receptor agonist or SGLT2 inhibitor, a sulfonylurea, or a DPP-4 inhibitor after metformin between 2014 and 2022, the BMJ Group announcement said. Average age was 60, 48% were women, and kidney function was tracked with blood tests over an average of 32 months.
Among those with albuminuria, the estimated five year risk of kidney decline was 3.2% on a GLP-1 or SGLT2 drug, 4.6% on a sulfonylurea and 5.4% on a DPP-4 inhibitor. That is 2.2 fewer cases per 100 people over five years, or about 22 fewer per 1,000, which is roughly 40% lower in relative terms.
Among the 61,583 people without albuminuria, the options were close together: 2.4% on a GLP-1 or SGLT2 drug, 2.1% on a DPP-4 inhibitor and 2.8% on a sulfonylurea. The sulfonylurea gap is 0.7 more cases per 100 people over five years, or about 7 more per 1,000, which works out to about 31% higher relative risk.
The authors concluded: “Compared with DPP-4 inhibitors, GLP-1 receptor agonists and SGLT-2 inhibitors reduce the risk of deterioration of renal function in people with albuminuria. Little evidence of renal benefit among people without albuminuria was observed over the study follow-up period.”
Dr. Kumar’s take
This is a target trial emulation, meaning the researchers arranged existing records to imitate the structure of a randomized trial. The authors call that “an approach that minimises common sources of bias in observational analyses,” which beats a plain database comparison, but they also write that “owing to the absence of randomisation, residual confounding cannot be excluded.” Doctors choose which patient gets which drug, and those reasons travel with the data. Thirty-two months of average follow up is also short for a kidney story, so a five year projection drawn from it is not a finish line.
The effect size deserves the same honesty. For people without protein in the urine, the gap between the best and worst number was under one percentage point over five years, and in that group the older sulfonylureas were the worst performer rather than the newer drugs being the best. The narrower finding is the useful one: the test that tells you which group you are in costs almost nothing, and the benefit sits almost entirely in the smaller group who test positive. For a drug that slowed kidney decline in people without diabetes, see my write up on finerenone.
What it means for you
If you have type 2 diabetes, find out whether you have albuminuria. It is a cheap urine test that plenty of people with diabetes go years without. That number, not a headline about Ozempic, is what this study says should shape the decision.
If you do have protein in the urine and are adding a drug after metformin, ask your doctor whether a GLP-1 drug or an SGLT2 inhibitor fits your situation better than a sulfonylurea or a DPP-4 inhibitor. Nothing here licenses switching a drug on your own, and the study says nothing about people already stable on a regimen that is working.
If your urine test is clean, the kidney argument for the newer drugs is weak in this data. Other reasons may still apply, including weight and other GLP-1 benefits, but kidney protection should not be the deciding one.