GLP-1 Vision Injury MDL Explained: The NAION Evidence and the New Federal Docket

Why vision claims received a separate MDL

On December 15, 2025, JPML centralized 21 actions in the Eastern District of Pennsylvania and noted nine additional related actions. The complaints alleged that Ozempic, Wegovy, Saxenda, or Trulicity caused non-arteritic anterior ischemic optic neuropathy, commonly shortened to NAION. The Panel selected Judge Karen S. Marston, who already presided over GLP-1 gastrointestinal MDL 3094, but declined simply to fold the eye cases into that proceeding.

That separation protects an important boundary. MDL 3094 coordinates allegations such as gastroparesis, ileus, and intestinal obstruction. MDL 3163 concerns optic-nerve injury. Some defendants and regulatory evidence overlap, but medical mechanisms, treating specialists, expert literature, and claimed damages differ. On July 1, 2026, JPML listed 146 pending actions in the NAION MDL. Those should not be silently added to a gastrointestinal-only headline.

What NAION is—and why background risk matters

NAION is an acute ischemic optic neuropathy, usually presenting as sudden, painless vision loss in one eye, often noticed on waking. It is not the same condition as diabetic retinopathy, the retinal complication already discussed in semaglutide labeling. Clinicians consider optic-disc appearance and vascular or systemic factors. Age over 50, diabetes, hypertension, dyslipidemia, obstructive sleep apnea, smoking, and a small cup-to-disc ratio are among commonly discussed risk factors.

Those factors complicate observational research because GLP-1 medicines are prescribed to people with diabetes or obesity—conditions that can travel with NAION risks. Statistical adjustment and active comparators help, but electronic-record data may not fully capture disease severity, adherence, rapid metabolic change, anatomy, or specialist-confirmed diagnosis. Conversely, rarity makes randomized trials unlikely to accumulate many events. No single design removes every uncertainty.

The studies point in more than one direction

A Danish cohort of 424,152 people with type 2 diabetes reported NAION incidence of 0.228 per 1,000 person-years among once-weekly semaglutide users and 0.093 among nonusers, with an adjusted hazard ratio of 2.19. Other database studies have reported increased associations. These findings are important because they use large real-world populations, but they remain nonrandomized and can be affected by treatment selection, coding, surveillance, and residual confounding.

Other analyses have not replicated the same signal. A multinational database study concluded that avoidance based solely on NAION concern was not warranted. A Military Health System study reported lower risk among semaglutide-treated patients with diabetes. A pooled analysis of 96,829 randomized-trial participants identified three confirmed NAION cases in GLP-1-treated participants and five among placebo recipients, finding no increased incidence. Different populations, comparisons, event validation, and follow-up can produce materially different estimates.

What the federal court is doing now

The court's early orders built infrastructure rather than deciding the science. It created a streamlined docket, allowed direct filing under specified conditions, appointed plaintiffs' leadership, entered preservation and privilege procedures, and scheduled a science day. Science day allows the parties to educate the judge about medical literature; presentations are not sworn trial testimony and the event is not a Rule 702 ruling.

The transfer order frames common questions about development, testing, regulatory history, promotion, labeling, and capacity to cause NAION. Plaintiffs assert failure-to-warn, design-defect, and warranty theories; defendants dispute those claims. Later proceedings may test general causation and product-specific issues. A study association does not automatically satisfy evidentiary rules, and an admissibility ruling would not by itself determine whether a medicine caused one plaintiff's vision loss.

How an individual record differs from an MDL count

A useful medical chronology identifies the exact medicine, indication, dose changes, start and stop dates, onset of visual symptoms, ophthalmology findings, imaging, visual fields, optic-disc anatomy, and competing diagnoses. Diabetes control, blood pressure, sleep apnea, smoking, vascular events, phosphodiesterase-5 inhibitors, and other medications may be relevant. Sudden visual loss is a medical emergency; readers should seek urgent care rather than use a litigation page to self-diagnose or change treatment.

The 146-case count is therefore a scale marker, not a list of medically confirmed drug injuries. As of July 10, 2026, the cited public orders do not announce a universal settlement, value grid, or public eligibility standard. The accurate status is that coordinated discovery and scientific development are underway while the evidence remains contested. Patients and prescribers should make treatment decisions from current labeling and individualized medical advice, not promised lawsuit outcomes.

Reader questions

Frequently asked questions