Evidence Hub

Neurosensory Risk After Semaglutide or Tirzepatide Is Lower Than Matched Antidiabetic and Bariatric Surgery Cohorts and Tracks Glycemic Burden

Marwaha A, Gregg C, Venkatakrishnan A, Murugadoss K, Varma G, Shiv S, Soundararajan V

2026

16,143 participants

Abstract

Glucagon-like peptide-1 receptor agonists (GLP-1RAs) are increasingly used for type 2 diabetes and obesity, but emerging evidence from randomized trials and real-world anecdotes suggest that neurosensory adverse events merit systematic investigation. We analyzed 670,422 initiators of semaglutide or tirzepatide across a large-scale federated EHR network and ascertained 14 new-onset neurosensory phenotypes over 12 months using AI-assisted clinical-note curation harmonized with ICD-10-CM codes. Among 128,377 at-risk GLP-1RA initiators — new users with at least two prescriptions within 12 months and complete 12-month follow-up, free of any studied phenotype at baseline — 29,519 (23.0%) had at least one newly documented phenotype, most commonly hypoesthesia/numbness in 16,143 patients (12.6%) and paresthesia in 15,415 patients (12.0%). Outcome ascertainment was validated by physician-led curation, with the review of 223 AI-identified events yielding a positive predictive value of 88.8% (95% CI, 84.0–92.3). In propensity-score–matched analyses, GLP-1RA initiators were associated with a 23% lower risk than initiators of the antidiabetic medications metformin, dipeptidyl-peptidase-4 (DPP-4) inhibitors, or sodium-glucose co-transporter-2 (SGLT2) inhibitors (19.7% versus 25.6%; RR, 0.77; 95% CI, 0.75–0.79; q12%, but showed no ordered association with achieved weight loss (P=0.867), baseline body-mass index (BMI, P=0.784) or BMI reduction (P=0.740). Maximum attained…

How strong is the evidence?

Evidence Hub reads this as a randomized trial, done in people. Its evidence level is 2 of 5 (High · randomized trial). Randomized trials of a reasonable size are the strongest single-study design. Reported size: 16,143 participants or records.

How the five levels of evidence work. The level is assigned automatically from the abstract and can be wrong, so check the paper.

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free