


Is the drug that reshaped modern medicine actually as safe as the prescriptions suggest?
Approximately 11% of Americans are currently taking GLP-1 medications — a figure that would have seemed implausible five years ago. Prescriptions increased over 2000% between 2019 and 2022. Serena Williams lost 31 pounds on Zepbound and returned to professional tennis at age 44, competing at Wimbledon and the U.S. Open in 2026. The cultural momentum behind these drugs is, at this point, nearly impossible to overstate. But momentum and safety are different things. And the gap between the two is where patients are getting hurt.
The reassuring headline from a Harvard Medical School and CDC study published on April 8, 2025 in the Annals of Internal Medicine is this: fewer than four emergency department visits occurred per 1,000 semaglutide patients. For a drug taken by tens of millions of people, that number sounds manageable. Clinical advocates point to it frequently.
Look closer and the picture becomes more complicated.
Nearly 70% of those emergency visits involved gastrointestinal problems — nausea, vomiting, abdominal pain, diarrhea, constipation. Approximately 6% involved allergic reactions. Approximately 17% involved hypoglycemia, dangerously low blood sugar. The FDA updated Ozempic’s warning label in September 2023 to include ileus, a severe gastrointestinal disorder that causes the bowel to stop moving. These are not minor inconveniences. For some patients, they represent extended hospitalisation, permanent organ damage, or conditions that do not resolve when the drug is discontinued.
The question is not whether serious adverse effects are rare. The question is whether patients were told they were possible at all.
Todd Engel lost vision in both eyes to a condition called NAION — non-arteritic anterior ischemic optic neuropathy — after taking Ozempic for diabetes. His first eye went blind on New Year’s Eve 2023. His second eye went black a year later. His case is among the most documented of what researchers now recognise as a pattern, not an anomaly.
Research published in JAMA Ophthalmology identified a striking association: diabetic patients taking semaglutide faced a 400% increased risk of NAION compared to patients on other diabetic medications. For overweight patients taking the drug, that figure rose to 700%. The FDA database contains over 1,000 reports of optic neuropathy linked to GLP-1 medications since 2024, drawn from nearly 190,000 total reports filed in that period. The same database records 1,928 reports of pancreatitis since 2024.
Novo Nordisk is now facing approximately 3,800 pending lawsuits. Legal analysts, as of July 2026, project total liability potentially exceeding $2 billion. No settlements or completed trials have yet occurred. But the volume of litigation — and the specific nature of the claims — suggests this is not a matter of isolated misfortune.
Clinical trials, by design, measure what researchers think to look for. They are not built to capture what patients actually experience at scale, in their own words, across years of use.
An AI analysis of over 400,000 Reddit posts found that nearly 4% of GLP-1 users reported menstrual irregularities. Common symptoms emerging from that analysis included chills, hot flashes, and fatigue — none of which appear prominently in standard prescribing information. Researchers were careful to note these findings do not establish causation. But they identify signals that warrant formal investigation, and they raise the uncomfortable question of how many symptoms have gone unreported simply because patients assumed they were unrelated to the drug.
The dermatological findings are, perhaps, the most striking. A multinational study presented at the European Academy of Dermatology and Venereology Congress 2026 found that GLP-1 users were more than four times as likely to report nail detachment — 39% of users versus 9% of non-users. Some 66% of GLP-1 users reported at least one nail disorder. A companion study on hair loss found 50% of GLP-1 users reported hair loss compared to 34% of non-users, with 26% of users experiencing hair loss that began after treatment started versus 12% of controls.
These are not side effects listed on the standard patient information sheet. They are findings that emerged years after these drugs entered mass circulation.
Ozempic was originally approved and dosed for blood sugar control in type 2 diabetic patients. Wegovy — its weight-loss equivalent — is capped at a higher maximum dose specifically because weight loss requires more aggressive titration. The distinction matters. Millions of patients have been prescribed Ozempic off-label for weight loss, at doses calibrated for a different clinical purpose, without the benefit of long-term safety data in non-diabetic populations.
The rapid expansion of use into entirely new patient populations has created, in effect, a vast uncontrolled experiment. Long-term effects such as pancreatitis, kidney injury, and vision problems may persist for months or years. When patients discontinue the medication — which many eventually do — rapid weight regain has been widely documented, raising questions about dependency and whether manufacturers adequately communicated the long-term implications of treatment.
Stanford researchers have identified a naturally occurring molecule called BRP that produced Ozempic-like appetite suppression in animal studies without the accompanying nausea, constipation, or significant muscle loss, acting specifically through the hypothalamus. That research is early-stage. But it suggests the side effect burden of current GLP-1 medications is not an inevitable feature of appetite suppression — it may be a feature of this particular mechanism that better-targeted alternatives could eventually avoid.
The clinical consensus has not turned against these drugs. For many patients — particularly those managing type 2 diabetes or severe obesity — the benefit-to-risk calculation remains favourable. The Harvard and CDC data, whatever its limitations, does confirm that catastrophic outcomes are not common at a population level.
But the framing of that reassurance deserves scrutiny. Rare, at the scale of tens of millions of prescriptions, still means thousands of people. The patients who experience NAION, or gastroparesis, or nail detachment, or vision loss — they are not statistical noise. They are individuals who deserve to have understood what they were agreeing to.
The central question facing regulators, manufacturers, and clinicians is not whether Ozempic works. It is whether the people taking it were given enough information to make an informed choice.
I think that question, honestly, has not been answered satisfactorily. Worth considering — before the next prescription is written.