A flat line that deserves a second look
If you skim the results table, the kidney finding looks like a non-event. Estimated glomerular filtration rate, eGFR, the standard measure of how well your kidneys are filtering, changed by an average of 1.04 at six months. The confidence interval ran from 4.56 lower to 2.49 higher. p = 0.56. Not significant.
A lot of readers stop there and file it under "no effect on kidneys".
That would be a mistake, and understanding why is one of the most useful things in this whole paper.
Why "no change" can be good news
Type 2 diabetes is one of the leading causes of kidney failure. Left alone, kidney function in many people with diabetes drifts downward year after year. So when you look at a group of people with diabetes over six months and their eGFR has not dropped, that is not nothing. Holding the line is the goal.
There is also a quirk specific to SGLT2 inhibitors worth knowing. When you start one, eGFR often dips a little in the first few weeks. That early dip reflects the drug reducing pressure inside the kidney's filtering units, which is thought to be protective over the long run. Doctors expect it. If you ever see your eGFR drop slightly after starting one of these drugs, ask about this before assuming something has gone wrong.
Six months is not enough time to see the real effect
The authors say directly that the follow-up in these real-world studies was not long enough to observe the kind of kidney benefit seen in large SGLT2 trials. They point to DECLARE-TIMI 58, with a median follow-up of 4.2 years, and CANVAS, with a mean of 3.6 years.
Kidney protection is a slow story. You measure it in years, not months. A six month window mostly tells you "nothing bad happened", which is a fair and useful thing to know, but not the full picture.
The albumin numbers tell the more interesting story
eGFR measures filtration. Albumin in the urine measures leakage, and it tends to move earlier.
Two of the included studies reported albumin results.
In Carretero-Gómez and colleagues, the total urine albumin to creatinine ratio (UACR) dropped by 15.14 mg/g, and in people with macroalbuminuria it dropped by 63.18 mg/g at 26 weeks. Both results had p values under 0.0001.
In Díaz-Trastoy and colleagues, albuminuria data was available for 127 of 212 patients. Ten progressed to a worse category. Thirteen regressed to a better one.
Microalbuminuria is an independent risk factor for progressive chronic kidney disease and cardiovascular events in type 2 diabetes. Reducing it is a meaningful win even when eGFR has not had time to move.
Why GLP-1 drugs matter more as kidneys weaken
Part 1 touched on this and it deserves a fuller explanation.
SGLT2 inhibitors are recommended for people with type 2 diabetes and chronic kidney disease when eGFR is 30 or above and UACR is over 30 mg/g, because of their protective effect on the kidney through the renin-angiotensin-aldosterone system.
But below an eGFR of 30, the paper notes, the HbA1c lowering from an SGLT2 inhibitor becomes negligible. That makes sense when you remember how they work. They lower sugar by making the kidney dump it in urine. A kidney that is not filtering much cannot dump much.
GLP-1 receptor agonists do not care about kidney function for their glucose effect. They work through the pancreas, the stomach and the brain. So as kidney function falls, the GLP-1 partner keeps doing the glucose work the SGLT2 partner can no longer manage, while the SGLT2 drug may still offer kidney protection in its own right.
That is one of the strongest arguments for the pairing in people with diabetic kidney disease.
The trial that stopped early for good reasons
The paper also mentions FLOW, a trial of semaglutide in 3,534 people with type 2 diabetes and chronic kidney disease. It was stopped early because it met pre-specified efficacy criteria, meaning the benefit on kidney disease progression was clear enough that continuing would not have been fair to the placebo group.
And in a post hoc analysis of the STEP 2 trial (1,210 people with type 2 diabetes and overweight or obesity), semaglutide cut the proportion of people with microalbuminuria to 11.5% compared with 22.4% on placebo at week 68.
So the GLP-1 side of this combination has its own kidney evidence building up, separate from the SGLT2 side.
A small confusing detail worth knowing
In the paper's GRADE summary table, the comment next to the eGFR result reads that the combination "may reduce eGFR". If you read that quickly, it sounds alarming, because a falling eGFR means worsening kidneys.
It is not what the data shows. The change was not statistically significant, and the confidence interval included both small drops and small rises. The wording is standard GRADE language for the direction of the point estimate, not a warning. But it is a nice example of how careless phrasing in a summary table can scare someone who never reads the underlying numbers.
The takeaway
The kidney story here is not "no effect". It is "no harm over six months, and early signs of less leakage". Combined with the longer-term trials of each drug class, that is an encouraging place to be.
If you have diabetes, the two tests to know by name are eGFR and UACR. Ask for both. Ask what they were last time and what direction they are heading. Those two numbers will tell you more about your kidneys than almost anything else.
This series discusses a peer-reviewed meta-analysis of observational studies. Nothing here is medical advice. Do not start, stop or combine prescription medications without your doctor.
