Among 500 people with type 2 diabetes, the ones taking a GLP-1 drug had about a third the odds of painful nerve damage that the rest did. That is a big-looking gap. It is also a snapshot, and a snapshot cannot tell you which way the arrow runs.
The numbers come from a cross-sectional study published online August 20 in the journal Pain ↗, the flagship journal of the International Association for the Study of Pain. A group led by Pietro Falco and Andrea Truini at Sapienza University in Rome ran it, working with Vincenza Spallone at the University of Rome Tor Vergata and the pain physiologist Lars Arendt-Nielsen at Aalborg University in Denmark. They enrolled 500 consecutive patients with type 2 diabetes and sorted them by what they were taking: standard treatment (208 people), an SGLT2 inhibitor (152), or a GLP-1 receptor agonist (140), the class that includes semaglutide, dulaglutide, and liraglutide.
What they measured
Diabetic peripheral neuropathy is the slow loss of the sensory nerves in the feet and legs that decades of high blood sugar can cause. It shows up as numbness, tingling, and, in a hard-to-treat subset, constant burning or shooting pain. About one diabetic in two eventually develops it, and it is the leading reason people with diabetes lose a foot.
What makes this study harder to wave away than the usual database readout is how the nerve damage was graded. The team did not rely on a questionnaire. They ran neurological exams, nerve conduction studies (which measure how fast an electrical signal travels down a nerve), and skin biopsies when needed to count the tiny fibers in the skin. Pain was graded with three separate scored instruments. Objective tests like these are harder to fudge than a patient checking a box.
Across the whole group, 58.6 percent had neuropathy and 26.2 percent had the painful form. The GLP-1 users stood apart. Neuropathy of any kind showed up in 46.4 percent of them versus 60.6 percent on standard treatment. The painful form was the sharper contrast: 10.1 percent on GLP-1 drugs, against 35.1 percent on standard care and 28.3 percent on an SGLT2 inhibitor. After the authors adjusted for the obvious confounders (age, diabetes duration, blood-sugar control, and other clinical factors), GLP-1 use was tied to lower odds of neuropathy (odds ratio 0.60, meaning roughly 40 percent lower odds) and much lower odds of painful neuropathy (odds ratio 0.32, roughly a two-thirds cut). Severity and pain intensity ran lower too.
The direction problem
There is a mechanistic story that makes the finding plausible. GLP-1 receptors sit on peripheral neurons, not just in the pancreas and brain, and in animal work GLP-1 signaling has looked neuroprotective. If the drugs genuinely shield nerves, or quiet the pain pathway, that would be a real bonus on top of weight loss and glucose control.
But the study cannot show that, and the authors say so plainly. A cross-sectional design photographs everyone once. It records who was on which drug and who had nerve damage at the same moment. It cannot watch a nerve get better. So the arrow could run the other way. People with milder, shorter, better-controlled diabetes are exactly the people a clinician is most likely to start, and keep, on a GLP-1 drug, and they are also the people least likely to have advanced neuropathy in the first place. That is confounding by indication, and no amount of statistical adjustment fully removes it from a single snapshot. The 0.32 odds ratio for painful neuropathy is striking precisely because it is the kind of number that a selection effect can inflate.
This is the same shape peptidemodel has flagged before. When one analysis found tirzepatide users developed less carpal tunnel syndrome ↗, the likeliest explanation was weight loss taking pressure off the nerve, not the drug treating the nerve. And the traffic is not all one way: a separate case series gave tirzepatide its first report of cutaneous allodynia ↗, pain from ordinary touch, appearing as the dose climbed. GLP-1 drugs like semaglutide ↗, dulaglutide ↗, and liraglutide ↗ all hit the GLP-1 receptor ↗, and the sensory nervous system is clearly within reach of it. Whether the net effect on nerves is protective, harmful, or mostly a reflection of who takes the drugs is the open question this study sharpens without settling.
The authors are asking for the right next step: a longitudinal study that follows the same nerves over time. Until one runs, the honest reading is that GLP-1 users in this Rome clinic had less nerve pain, and nobody yet knows why.