Ozempic long term side effects: what the evidence actually covers, year by year

The question changes around month eighteen. In the first weeks on semaglutide, people ask how long the nausea will last. By the second year the nausea is long gone, the weight is mostly off, and a quieter question shows up at the follow-up visit: what is this doing to me if I stay on it for ten years? That is the real search behind Ozempic long term side effects, and it deserves a more precise answer than the one it usually gets, which is some version of "these drugs have been studied for a long time."

That sentence is true for the drug class and misleading for the drug. Exenatide, the first GLP-1 medication, reached pharmacies in 2005. Liraglutide followed in 2010, Ozempic in late 2017, and Wegovy, the 2.4 mg weight loss dose of semaglutide, in 2021. So the class has two decades behind it, but a person starting Wegovy today is taking a dose that has been in routine use for about five years, and that has been followed in a controlled trial for about four.

The original element in this piece is an evidence horizon ledger. For each long term concern patients actually raise, we list the longest controlled human evidence that exists for semaglutide, what kind of evidence it is, and where it stops. No single label or review lays it out this way, and the pattern it reveals is the useful part: the risks we can measure well are the common ones, and the rare ones sit almost entirely in observational data.

Where the long term data comes from

Two trials carry most of the weight. STEP 5 followed adults with obesity on semaglutide 2.4 mg for 104 weeks and found average weight loss of about 15 percent against under 3 percent on placebo, with a side effect profile at two years that looked much like the one at twelve weeks: mostly gastrointestinal and mostly early. The SELECT trial is the big one. It enrolled more than 17,000 adults with overweight or obesity and existing heart disease but no diabetes, followed them for an average of about 40 months, and found a 20 percent reduction in heart attack, stroke and cardiovascular death. A follow up analysis showed weight loss held for up to four years in that population.

That is genuinely good long term evidence. It is also the edge of the map. Anything you read about year six, year eight or year fifteen on semaglutide specifically is reasoning from shorter data, from liraglutide, or from insurance claims.

The evidence horizon ledger, concern by concern

Gastrointestinal effects. Longest controlled evidence: about four years, randomized. Nausea and vomiting are overwhelmingly a titration problem that fades, which our breakdown of how long semaglutide side effects last covers in days. The long term signal is different: in SELECT, about 17 percent of people on semaglutide stopped because of adverse events against roughly 8 percent on placebo, mostly for GI reasons. Long term tolerability is a real issue, just not usually a dangerous one.

Pancreatitis, bowel obstruction and gastroparesis. Longest controlled evidence: trials too small to detect them reliably. The strongest signal is observational. A 2023 JAMA analysis of insurance records found GLP-1 users taking the drugs for weight loss had higher rates of pancreatitis, bowel obstruction and gastroparesis than people on bupropion and naltrexone. The absolute rates were low, the study cannot prove cause, and it is the reason severe persistent abdominal pain on these drugs is never something to wait out.

Gallbladder disease. Longest controlled evidence: about four years, and the signal is consistent. Rapid weight loss itself raises gallstone risk, which is why gallstones on a GLP-1 are better understood as a weight loss effect than a drug toxicity.

Thyroid C cell tumors. Longest controlled human evidence: about four years, with no clear signal, set against a boxed warning built on rodent data. Our piece on the thyroid boxed warning walks through who that warning is a hard stop for. The honest gap: medullary thyroid cancer is rare enough and slow enough that four years of trial data cannot rule a small effect in or out.

Vision, specifically NAION. Longest controlled evidence: none designed to detect it. A 2024 retrospective study from Massachusetts Eye and Ear linked semaglutide to a higher rate of nonarteritic anterior ischemic optic neuropathy, a sudden painless loss of vision in one eye. In June 2025 the European Medicines Agency's safety committee concluded that NAION is a very rare side effect of semaglutide, affecting up to 1 in 10,000 people. Rare, real, and entirely discovered after approval.

Muscle and bone. Longest controlled evidence: body composition substudies of a year or two. Losing weight fast takes some lean mass with it, and whether that matters over a decade depends heavily on protein and strength training. The bone question is covered in our look at GLP-1s and bone density.

What the studies do not tell you

Three gaps matter more than any single side effect. First, nobody has randomized people to ten years of semaglutide, so a slow cumulative effect would show up in registries, years late, the way NAION did. Second, SELECT enrolled people with established heart disease, and much of the benefit side of the ledger may not transfer to a healthy 32 year old using the drug for twenty pounds. Third, the trials studied branded drug at labeled doses. Compounded versions, which the FDA has warned about repeatedly, have no long term safety record at all.

A once a year self audit for anyone past month twelve

This is the practical half of the ledger. Once a year, ideally at the same visit, walk through five checks with your prescriber. One: has any abdominal pain in the past year been severe, persistent, or felt through to the back? That is the pancreatitis and gallbladder screen, and a yes means imaging, not reassurance. Two: any sudden change in vision in one eye, even if it partly recovered? That goes to an eye doctor promptly. Three: are you still doing resistance training at least twice a week and eating near your protein target? If not, the muscle and bone risk is the one you can actually control. Four: is the dose you are on still the lowest one that holds your weight? Many people never step down once they reach goal. Five: are you getting the drug from a licensed pharmacy with a real label? If any answer worries you, that is the conversation for that visit.

The takeaway

The long term picture of semaglutide is better than the fear and thinner than the marketing. Four years of randomized data look reassuring for the common effects and even protective for the heart in people who already have heart disease. Beyond four years, and for the rare events, we are relying on registries and case reports that arrive after the fact. For people who expect to need treatment indefinitely, that uncertainty is worth weighing openly against the alternatives, including surgery, which has decades of follow up behind it. Our guide on when weight loss surgery makes sense lays out that comparison, and the NIDDK overview of obesity medications is a sound neutral reference for the full list of approved options.