Metformin users with type 2 diabetes need fewer nerve decompression surgeries, large cohort finds

Type 2 diabetes is an established risk factor for compressive neuropathies such as carpal tunnel syndrome and ulnar neuropathy at the elbow. It also raises the odds of surgical complications and repeat operations. Metformin, the most widely used first-line diabetes drug and a leading candidate geroprotector, shows neuroprotective effects in preclinical work, including AMPK-driven autophagy, reduced neuroinflammation, and axonal regeneration. No study had tested whether it changes the real-world burden of nerve compression or surgical failure, which matters for keeping hand function as people age.

The authors used the TriNetX federated electronic health record network to identify adults with type 2 diabetes diagnosed between 2007 and 2020, with up to five years of follow-up. Metformin exposure meant at least 10 prescriptions over five years; controls had none. Patients with advanced kidney disease were excluded. One-to-one propensity matching balanced age, sex, race, BMI, HbA1c, diabetic neuropathy, kidney disease, smoking, and concurrent diabetes drugs. The primary outcomes were new compressive neuropathy diagnoses and nerve decompression surgery. A separate unilateral cohort assessed ipsilateral revision 30 days to 5 years after carpal or cubital tunnel release.

The primary analysis included 260,823 matched pairs. Metformin users had lower rates of nerve decompression surgery (0.60% vs 0.72%; OR 0.831, 95% CI 0.776–0.889, P<.001) and compressive neuropathy diagnoses (2.78% vs 3.05%; OR 0.909, 0.879–0.941). Results were directionally consistent for carpal tunnel release (OR 0.844) and cubital tunnel release (OR 0.817). Carpal tunnel syndrome and cubital tunnel syndrome diagnoses were each lower as well (ORs 0.903 and 0.919).

Four prespecified sensitivity analyses examined concurrent drugs. Without insulin (153,342 pairs), the surgery association strengthened (OR 0.727, 0.662–0.799). With insulin in both groups (55,160 pairs), it faded to nonsignificance (OR 0.948, P=.410), though the diagnosis-level link persisted (OR 0.924). A similar pattern appeared for SGLT2 inhibitors and GLP-1 agonists: stronger without them (OR 0.803) and attenuated with them (OR 0.874, P=.023). The authors suggest overlapping autophagy effects, or more advanced disease in insulin-treated patients, may explain this.

Among 2,746 matched pairs with unilateral carpal or cubital tunnel syndrome who had surgery, ipsilateral revision was less frequent in metformin users (3.0% vs 4.2%; pooled OR 0.711, 0.533–0.950, P=.021, I²=0%). The effect was driven by carpal tunnel release (2.8% vs 4.2%; OR 0.653, P=.017); cubital tunnel release was not significant (OR 0.848, P=.521). Metformin users also had lower vitamin B12 (mean 646.4 vs 707.4 pg/mL, P<.001), consistent with sustained exposure and arguing that any true benefit may be underestimated.

The design cannot establish causation. Indication bias, immortal-time bias from the 10-prescription requirement, unmeasured factors such as surgical technique and symptom severity, and missing laterality data all limit interpretation. No multiple-comparison correction was applied, so the drug-stratified results are hypothesis-generating. The authors state these data do not support prescribing metformin solely to avoid nerve surgery, but that continuing therapy may offer neuroprotection beyond glycemic control. Prospective studies with nerve conduction data and verified adherence are needed.