Nearly 100 GLP-1 Vision Loss Suits in NJ Hinge on Causation Evidence

For product liability and mass tort lawyers, the GLP-1 NAION filings are a textbook failure-to-warn dispute where general causation is everything. With nearly 100 cases in New Jersey since 2025 and a Philadelphia cohort emerging, the litigation will likely pivot on whether mixed epidemiological studies survive Daubert scrutiny and whether diabetes itself—not the drug—explains the vision loss.

Novo Nordisk and Eli Lilly are now defending a growing product-liability docket over claims that their blockbuster GLP-1 receptor agonists Ozempic, Wegovy, and Zepbound can cause a rare and potentially blinding optic nerve condition. The litigation centers on nonarteritic anterior ischemic optic neuropathy, or NAION, a sudden and often permanent loss of vision caused by reduced blood flow to the optic nerve. According to reports citing USA Today, nearly 100 lawsuits have been filed in New Jersey since 2025, with a separate group emerging from Philadelphia. The core allegation is not that the drugs are defective per se, but that manufacturers failed to warn patients and prescribers about a potential NAION risk.

“

Novo Nordisk and Eli Lilly are now defending a growing product-liability docket over claims that their blockbuster GLP-1 receptor agonists Ozempic, Wegovy, and Zepbound can cause a rare and potentially blinding optic nerve condition.

The scientific record, however, is far from settled. Some recent studies have raised concerns about an increased risk of NAION among users of specific GLP-1 drugs, but others have found no association. One positive study drew its data from patients at an institution that specializes in eye care, making its sample unrepresentative of the general population. That distinction matters in both medicine and law: a specialized ophthalmology center may over-select patients already prone to optic nerve problems. Because of the mixed results, doctors and researchers quoted in the coverage say there is not enough evidence to state with confidence that these drugs cause NAION.

The absolute numbers reinforce the rarity of the condition. Even under the worst-case assumption in one analysis from the Wall Street Journal—that semaglutide doubled the risk—the average would still be only two cases per 10,000 people taking the drug. NAION itself is uncommon and typically occurs in people 50 and older, and its underlying causes are not fully understood. That rarity cuts both ways: it is scientifically hard to prove a drug effect in a small patient population, but it also makes the profound injury—permanent vision loss—especially devastating for plaintiffs and persuasive to juries.

For the defense, Novo Nordisk and Eli Lilly have emphasized an epidemiological confounder that sits at the heart of the litigation. People with type 2 diabetes, the primary initial population for many GLP-1 drugs, are already at higher risk of blindness and other vascular complications. The companies deny any significant relationship between GLP-1s and NAION and have argued in court that the underlying disease, rather than the drug, may explain the outcomes. Novo Nordisk has also asserted that semaglutide has a well-established safety profile, and both companies continue to contest general causation—the legal requirement that a plaintiff prove the drug is capable of causing the injury in anyone before proving it caused the injury in them.

The litigation has classic mass-tort features. A concentration of filings in New Jersey and Philadelphia suggests early attempts at coordination or consolidation, which could eventually lead to a multidistrict litigation proceeding if the cases qualify for federal MDL treatment. Failure-to-warn claims will likely turn on what the manufacturers knew, when they knew it, and whether the FDA-approved label adequately reflects the available evidence. At this stage, the mixed scientific literature provides room for both sides: plaintiffs can point to positive studies and argue that even an uncertain risk required a warning, while defendants can argue that a label need not include every unresolved hypothesis and that the absolute risk is too low to support causation.