Edward Gillen, a bald man with a gray goatee, in an interview on the left; on the right, he walks with a white cane beside a woman in a gray dress
Edward Gillen, who is suing the maker of Ozempic after alleging the drug caused him to go permanently blind.
The Law · Litigation

The FDA Eye Warning Helps on Preemption. It Doesn’t Settle Causation.

The FDA warning helps plaintiffs on preemption. It does not settle causation, and it may not come into evidence.

KC
By Karen Campbell · Nov. 9, 2026 · 2 min read

The FDA’s Nov. 9, 2026 warning helps plaintiffs on preemption, the issue the defendants most want decided early. It does not settle general causation, and it may not come into evidence at all.

Preemption

In Merck v. Albrecht (May 20, 2019), the Supreme Court said preemption of a failure-to-warn claim is a question for the judge. The manufacturer has to show “clear evidence” that it fully informed the FDA of the justification for the warning and that the FDA told it the agency would not approve the change.

The FDA has now ordered the warning. That makes the argument that the FDA would have rejected a NAION warning hard to make on this record. That is my inference, not a court’s holding.

Judge Karen S. Marston’s July 2, 2026 order in MDL 3163 grants early discovery and motion practice on general causation and on preemption and warning adequacy. It quotes the two questions preemption turns on: whether the label already contains the warnings the plaintiffs want, and whether the FDA would have rejected additional ones. If the early motions do not resolve the cases, bellwether selection follows.

General causation

The FDA wrote “potential increased risk.” It did not conclude that the drugs cause NAION. The defendants will quote that.

The warning rule also sets a low bar. Under 21 CFR 201.80(e), labeling must be revised once there is reasonable evidence of an association with a serious hazard, and “a causal relationship need not have been proved.” So the label change does not itself show causation.

The FDA’s data stand on their own as science. They also have weak points. The Sentinel result against SGLT-2 inhibitors was not significant, and a Cleveland Clinic study of about 220,000 GLP-1 users with type 2 diabetes found no statistically significant increase, according to Drug Discovery Trends. Expect all of it in the Rule 702 briefs.

Whether the label change comes in

Federal Rule of Evidence 407 bars evidence of a later measure to prove “a need for a warning or instruction.” The defendants will argue that covers the new label.

The rule allows other purposes, such as proving the feasibility of a precaution when that is disputed.

The period before the warning

These claims reach back before Nov. 9, 2026. In June 2025, the European Medicines Agency’s safety committee concluded that NAION is a “very rare” side effect of semaglutide, and European labeling was updated to say so.

What is not known

The FDA’s page describes the new label language but does not reproduce it.

Sources

FDA Drug Safety Communication, Nov. 9, 2026 · 21 CFR 201.80 · Merck v. Albrecht · Federal Rule of Evidence 407 · E.D. Pa. CMO 12, July 2, 2026 · Drug Discovery Trends

Nothing on this page is sponsored.

Next story The Law

FDA Would Have Taken 55 Years to Release One Vaccine File, Plaintiffs Say

Back to The Law

Get Plaintiff’s Brief by email