Glucagon-like peptide-1, or GLP-1, is a hormone your gut releases after a meal. It signals the pancreas to release insulin, tells the brain you are full, and slows how fast food leaves the stomach. Over the past decade, researchers have synthesized peptides that mimic or amplify this signal, creating what scientists call GLP-1 receptor agonists. Most published research on these compounds has focused on blood sugar control and weight, but a quieter body of work has been asking a different question: do they also protect the kidneys?
A 2026 systematic review and meta-analysis published in The Journal of International Medical Research brought together ten large randomized controlled trials, covering more than 26,000 participants, to answer that question as rigorously as possible. The researchers looked at two primary outcomes: how much body weight changed and how much a kidney stress marker called the urinary albumin-to-creatinine ratio changed. They also looked at a secondary marker of kidney filtration capacity. Crucially, they split results by whether participants had type 2 diabetes, because most earlier studies focused heavily on diabetic populations.
The findings are drawing attention because they suggest the kidney signal may not be limited to people with diabetes, though the evidence for that non-diabetic group is less certain. Here is what the trial record showed and why researchers say the details matter.
What the urinary albumin-to-creatinine ratio measures
Healthy kidneys act as filters. They keep large proteins like albumin inside the bloodstream while clearing waste into urine. When the filtration barrier is under stress, small amounts of albumin begin leaking into urine. Scientists measure this by comparing albumin to creatinine, a waste product, in a urine sample. The resulting ratio, often abbreviated UACR, rises when kidneys are under strain and falls when the strain eases.
Even mild UACR elevation is associated with increased cardiovascular and kidney risk in both diabetic and non-diabetic populations. Reducing it is therefore considered a meaningful target in research settings, not just a number on a lab report. The meta-analysis used percentage change in UACR as one of its two primary outcomes precisely because it captures early kidney stress in a way that is clinically interpretable.
Body weight findings across 10 trials
Across all ten included trials, participants receiving a GLP-1 receptor agonist lost a mean of 5.85 kilograms more than those receiving a placebo or standard care. The 95 percent confidence interval ran from roughly 3.9 to 7.8 kilograms, meaning the researchers are confident the true effect falls somewhere in that range.
The meta-analysis rated the certainty of evidence for this weight outcome as high, the top rating on the scale researchers use to express confidence. Despite some statistical heterogeneity across studies, which simply means individual trial results varied around the average, every single included study pointed in the same direction: participants on the active compound lost more weight. That consistency across diverse trial designs strengthened the overall confidence rating.
Kidney marker changes and the diabetes distinction
On the UACR outcome, the pooled result showed a mean reduction of about 28 percent compared with placebo or standard care. The confidence interval stretched from roughly 18 to 38 percent, and the researchers rated certainty of evidence as moderate, one step below high.
When the team split results by glycemic status, an interesting pattern emerged. In the subgroup of trials enrolling people with type 2 diabetes, the reduction in UACR was approximately 25.7 percent, and there was essentially no statistical heterogeneity between studies. That is an unusually clean result for a meta-analysis of this size, suggesting the effect in diabetic populations is consistent and reliable.
In the subgroup without type 2 diabetes, the average reduction was actually numerically larger at about 30.9 percent, but the heterogeneity statistic was very high at 96 percent. That number signals that individual trial results were scattered widely around the average, which makes the pooled figure much harder to interpret with confidence. The meta-analysis authors were careful to flag this: the directional signal in non-diabetic populations is encouraging, but the imprecision and variability mean firm conclusions cannot be drawn yet.
What drove the variation in kidney results
To understand why individual trials produced such different UACR results, the researchers ran a statistical technique called meta-regression. This lets analysts test whether specific factors, such as how long a trial lasted or how high participants' baseline UACR was, can explain the scatter in outcomes.
Two factors emerged as meaningful. First, the UACR benefit appeared to attenuate with longer treatment duration. Put differently, the reduction looked larger in shorter trials and somewhat smaller in longer ones. Second, participants who started with higher baseline albuminuria showed greater benefit. This suggests the peptide class may have the most measurable kidney impact when kidney stress is already elevated at the start of treatment.
Notably, the specific GLP-1 receptor agonist used and the glycemic status of participants were not consistent moderators of the UACR outcome. The researchers interpreted this as evidence that the effect may be a class-level property rather than something unique to a single compound or to people with diabetes.
The filtration rate secondary outcome
The estimated glomerular filtration rate, or eGFR, measures how efficiently the kidneys filter blood. A higher eGFR is generally better. The meta-analysis found a modest difference between groups: participants on GLP-1 receptor agonists had an eGFR about 0.82 milliliters per minute per 1.73 square meters lower than the comparison group. There was no heterogeneity across studies on this metric.
The authors noted this small eGFR difference is consistent with a well-recognized pattern in kidney research: treatments that lower pressure inside the kidney's filtering units often produce a temporary, modest eGFR dip that does not reflect actual kidney damage. Whether this explanation applies fully here is a question the trial record did not definitively resolve, and the authors urged careful interpretation of this secondary metric.
Limitations the researchers identified
The team applied rigorous quality checks throughout. They assessed risk of bias in each included trial, examined funnel plots for signs of publication bias, ran trim-and-fill analyses to account for potential small-study effects, and performed leave-one-out sensitivity analyses, where each study is temporarily removed to test whether a single trial is driving the overall result. The main conclusions held up across these tests.
Still, several limitations deserve mention. The non-diabetic subgroup analysis was limited by the high heterogeneity and by the inability of the existing trial data to separate participants with normal glucose from those with prediabetes or intermediate glycemic phenotypes. Those distinctions might matter biologically and could explain some of the scatter. The attenuation of effect over longer durations is also not fully explained by the current data, and future trials designed specifically for non-diabetic chronic kidney disease populations would be needed to draw firmer conclusions. The researchers also noted that different trials measured UACR at different time points and used different baseline populations, adding another source of variability.
