Dose Insider / Analyses / Why people stop

Why People Stop GLP-1 MedicationsHow Many Quit, Why, and How Many Come Back

Updated November 20, 2026 · 16 min read · JAMA Netw Open 2025, JMCP 2026

In US health records, 64.8% of adults without diabetes who started a GLP-1 medicine had stopped within a year. With type 2 diabetes it was 46.5%.

53.6%of adults who started a GLP-1 had stopped within a year, 72.2% within two
Stopped within a yearAdults who started a GLP-1 medicine, shareWithout diabetes64.8%With type 2 diabetes46.5%Stopped within a yearAdults who started a GLP-1 medicine, shareWithout diabetes64.8%With type 2 diabetes46.5%

Across both groups, 53.6% had stopped within a year and 72.2% within two.

On this page

Cost and insurance lead the reasons people give, ahead of side effects, and more than a third of those who stop are back on a GLP-1 within a year.

This page goes through the studies behind those numbers: how many people stop and how soon, what made them stop, who is more likely to, and what happens afterwards. One calculator places your own weight loss on the largest study's curve.

How many people stop taking a GLP-1 within a year?

The largest study of the question used the health records of 125,474 US adults with overweight or obesity who filled a first prescription for liraglutide, semaglutide or tirzepatide between 2018 and 2023. A person counted as having stopped on the first day they had gone 60 days or more with no medicine on hand.1

Of 100 people who startRodriguez 2025
Who
After
StoppedStill on treatment

Each dot is one person in a hundred. Stopped means 60 days or more without a GLP-1 on hand, counted from pharmacy fills. People who later came back are still counted as stopped here.

Across the whole group, 53.6% had stopped within one year and 72.2% within two. The gap between the two groups is wide. Within two years, 84.4% of people without diabetes had stopped, against 64.1% of people with type 2 diabetes.1

The authors point to one likely cause: insurance coverage of these drugs is limited for people without type 2 diabetes.1

A gap in pharmacy fills is all this number records. It does not say why the fills ended, and it counts people who paused and returned later. Both questions have their own sections below.

Why do studies give such different numbers?

Before this study, published one-year stopping rates ran from 37% to 81%.1 Four recent studies show where the spread comes from.

StudyWho was countedStill on treatment at 1 year
Rodriguez 20251125,474 adults, with and without diabetes, records from 30 health systems46.4%
Xie 20262126,984 insured adults without diabetes, 2019 to 202424.5%
Ali 202639,189 adults without diabetes at one academic medical center18.6%
Marshall 2026433,607 health plan members without diabetes, started 2021 to mid-202433.2% to 60.9%, by start year

Rodriguez reports 53.6% stopped; the table shows the remainder. Xie counted gaps of more than 60 days in insurance claims. Ali counted prescriptions that covered at least 336 days.

Three things move the result. The first is diabetes: every study limited to people without it finds fewer still on treatment. The second is the year people started, which the next section covers. The third is what the study can see.

The academic center study could see prescriptions, and it found that 36.7% of patients had a prescription that covered less than 32 days. The average was 3.6 days. Of those patients, 82.7% received no further prescription for a weight medicine that year.3 The study could not see why. A prescription that covers a few days looks less like a decision to quit than like a start that never happened.

Switching is another blind spot. In the insurance claims study, 20.6% of people changed from one GLP-1 to another within the year, most often from liraglutide to semaglutide. A study that follows a single drug counts them as having stopped. People who switched were more likely to be on treatment at one year than people who did not: 36.4% against 21.4%.2

Are people staying on longer than they used to?

Yes, in the one study that compared start years. It followed 33,607 health plan members without diabetes who started semaglutide or tirzepatide, and measured how many were still on treatment a year later.4

Still on it after a yearMarshall 2026
Medicine

Health plan members without diabetes. The bars are different groups of people, one for each start year, and the study did not randomise anyone to a drug.

Among people who started in 2021, one in three was still on treatment a year later. Among those who started in the first half of 2024, six in ten were. The authors' first explanation is supply: the shortages ended. They also name better handling of dose increases and side effects, and weight management programs offered alongside the drug.4

The tirzepatide bars sit above the semaglutide bars in both years they share. That is a comparison of different people at different times, so it does not show that one drug is easier to stay on. The four study authors named in the disclosure work for the company that paid for the analysis.4

How soon do people stop?

Early. In a national survey of 440 adults who had stopped a GLP-1, 54.6% said they had stopped within six months of starting and 79.7% within twelve.6

The records study gives the same picture from the pharmacy side. Among people who stopped and were followed afterwards, the median time on treatment was 176 days, about 25 weeks. Their median weight change by then was a loss of 2.9%, and a quarter of them had lost nothing.1

People who stopped, at the point they stopped
Median time on treatment176 days
Median weight change−2.9%
Stopped within 6 months (survey)54.6%

Twenty-five weeks is close to the length of the dose schedule itself. Zepbound starts at 2.5 mg and goes up by 2.5 mg no sooner than every four weeks, so the 15 mg dose arrives in week 21 at the earliest.7 A person on that schedule who stops at the median has had at most five weeks on the top dose. The Zepbound dosing schedule lays out the dates.

Clinical trials lose people early too, though far fewer. In the Zepbound withdrawal trial, 783 people started the drug with regular visits, and 14.4% had stopped by week 36. Adverse events were the most common reason, at 6.8%.7

Why do people stop?

Pharmacy data show a gap, not a reason. Two studies went after the reasons in different ways. One team read the medical charts of 288 people who had stopped semaglutide or tirzepatide for obesity within a year, at a health system in Ohio and Florida, and recorded the main reason for each.5 The other was the national survey, where people could name several reasons.6

Reasons for stopping2025 and 2026
Source

Charts: adults without diabetes who stopped within a year, one main reason each. Survey: adults with or without diabetes, any number of reasons.

Cost and coverage

Cost or insurance was the main reason in 47.6% of the charts. The category covers an insurance denial, a manufacturer discount coupon that expired, and an out-of-pocket price the person could not afford. It was more common among people who stopped after the third month (54.0%) than among those who stopped before it (40.6%), and more common with Medicaid (76.5%) or Medicare (61.0%) than with private insurance (43.0%).5

The survey explains how people end up there. Of those who had stopped, 31.3% had filled their first prescription without insurance, 35.4% had used a copayment coupon and 22.0% a free sample.6 A coupon or a sample pays for the start of a treatment and not for the rest of it.

Income shows up in the records study as well, but only for people with diabetes. Those with incomes above $80,000 stopped at a rate 28% lower than those with incomes under $30,000.1

Side effects

Side effects were the main reason in 14.6% of the charts, and nausea, abdominal pain, vomiting and diarrhea were the ones written down most. Among people who stopped in the first three months the share was 21.0%, and after that 8.7%.5 In the survey, where several answers were allowed, 33.3% named side effects, and people with diabetes named them much more often than people without: 50.1% against 28.1%.6

In the trials the share was smaller. In the two main Zepbound trials, 4.8%, 6.3% and 6.7% of people on 5, 10 and 15 mg stopped for good because of adverse reactions, against 3.4% on placebo. Most of them stopped in the first few months, and stomach and gut reactions were the usual cause.7 In the Wegovy trials it was 6.8% on 2.4 mg against 3.2% on placebo, with nausea (1.8%), vomiting (1.2%) and diarrhea (0.7%) leading.8

Both labels give the prescriber room here. The Wegovy label says that if a dose is not tolerated during the dose increases, the next increase can be delayed by four weeks. The Zepbound label says to consider a lower maintenance dose if the current one is not tolerated; 5 mg, 10 mg and 15 mg are all maintenance doses.7,8 The side effects timeline shows which weeks are the hardest.

Shortages

For 11.8% of the people in the chart review, the reason was that the pharmacy could not fill the prescription. They had started in 2022 or 2023.5 The larger records study names shortages as a limit on its own results: it had too little data to measure their effect.1

Reaching the goal

In the survey, 26.1% of people who had stopped said they had reached their weight goal.6 In the chart review this reason was rare: 4 of 288 records said treatment was paused or stopped after reaching the desired weight.5 The two methods reach different people. A survey asks the person directly, while a chart holds only what was said at a visit, and a person who feels finished may not come back to say so.

Too little weight loss, and where the drug came from

Unsatisfactory weight loss was the main reason in 1.7% of the charts, and another 2.4% had left for a compounded version.5 The survey adds a detail about source. About one in five people who had stopped had bought from a compounding pharmacy (9.8%) or a website (9.1%), and 14.4% had been given the drug by a friend or relative. People who used a compounding pharmacy or a website were more likely to stop because they felt the drug was not working: 17.5%, against 6.6% of those who used a retail or mail-order pharmacy.6

Do people who lose more weight stay on longer?

Yes. In the records study, each 1% of body weight lost went with a 3.3% lower rate of stopping for people without diabetes and a 3.1% lower rate for people with it.1 The calculator compounds that figure over your own weight loss.

My weight loss on the curveRodriguez 2025
Unit
Group in the study
Weight change—
Against no change—

Log your weight each week and the app works out the percentage for you.

Open the app

Along the bottom: percent of starting weight lost. Up the side: the rate of stopping, where 100% is a person whose weight has not changed. This is a pattern across 125,474 records, not a forecast for one person.

Weight lostRate of stopping, no diabetesRate of stopping, type 2 diabetes
5%15% lower15% lower
10%29% lower27% lower
15%40% lower38% lower
20%49% lower47% lower

Our arithmetic: the study's 3.3% and 3.1% per point of weight lost, compounded.

The link runs in both directions, and the study cannot separate them. A drug that is working gives a reason to keep paying for it. A person who stays on longer also has more time to lose weight. The study handled this by updating each person's weight over time, which reduces the second effect without removing it.1

Three other things went with stopping in the same analysis.1

FactorNo diabetesType 2 diabetes
Age 65 or older18% higher rate of stopping28% higher
A moderate or severe stomach or gut side effect19% higher38% higher
Income above $80,000, against under $30,000No clear link28% lower

How many people come back after stopping?

More than a third within a year. The records study followed 41,792 people after they stopped and counted the first new fill of any GLP-1, including a different one from before.1

Of 100 people who stoppedRodriguez 2025
Who
Time since stopping
Weight regained since stopping
Started againDid not

Each dot is one person in a hundred who had stopped. The regain buttons change the sentence, not the dots: the study reports regain as a rate, 2.8% higher per point regained without diabetes and 2.3% with it.

Weight was the clearest signal for coming back, as it was for staying. Each 1% of weight regained after stopping went with a 2.8% higher rate of restarting for people without diabetes.1 Compounded, a person who had regained 10% restarted at a rate about a third higher than a person whose weight had held.

Two groups came back less often. People aged 65 or older without diabetes restarted at a rate 27% lower than younger people. People who had a moderate or severe stomach or gut side effect the first time restarted at a rate 18% lower.1

The academic center study looked for the same thing in its own prescriptions and found restarts and switches to be rare within twelve months.3 It could only see prescriptions written inside one institution, while the records study also picked up fills written elsewhere.

What happens to weight after stopping?

It comes back, at a pace that has been measured. In the Zepbound withdrawal trial, everyone took the drug for 36 weeks and lost 20.9% of their weight on average. Half were then switched to placebo. Over the next year their weight rose 14.0% from where it stood, while the half who stayed on lost a further 5.5%.7

A 2026 review in the BMJ pooled 37 studies of weight medicines with 9,341 participants. After the medicine was stopped, weight returned at an average of 0.4 kg a month, and the gains in heart and metabolic measures were projected to be gone within 1.4 years. Regain was faster than after diet and exercise programs, by 0.3 kg a month.9

The page on what happens when you stop Zepbound puts the trial's two paths into your own weight.

What does the evidence say about my reason?

The reasons above each have a different answer in the data. Pick the one closest to yours.

My reasonStudies and labels

The app keeps your doses, weight and symptoms in one place to show a prescriber.

Open the app

What do the labels say about a gap in treatment?

The two labels treat a gap differently, and neither sets out a full plan for coming back after months away.

SituationZepbound labelWegovy injection label
One dose missedTake it within 4 days (96 hours). After that, skip it and take the next dose on the usual day.Take it if the next dose is more than 2 days away. If less, skip it.
Two or more doses missed in a rowNo instructionRestart the dose increases at a lower dose, to reduce stomach and gut side effects

Zepbound prescribing information, section 2.3; Wegovy prescribing information, section 2.4.7,8

The Wegovy patient leaflet adds that after two or more missed weeks, the person should contact their healthcare provider about how to restart.8 The Zepbound label is silent on longer gaps, so the dose to resume on is the prescriber's decision. The missed dose page works out the dates for a single late dose.

What is worth writing down?

Three records line up with what the studies measured.

Weight, once a week. It was the strongest signal in the records study for staying on treatment, and regain after stopping was the strongest signal for coming back.1

Side effects, with the dose and the day after the injection. Most people who quit a trial over side effects did so in the first few months, during the dose increases,7 and both labels let the prescriber slow down or step back. A dated list makes that conversation short.

Fill dates and what each fill cost. Cost and coverage ended more treatments than anything else, sometimes at a date that could be seen coming, such as the day a discount coupon expires.5

What can these studies not tell you?

Pharmacy and prescription records see fills, not injections. The records study notes that people who stretched their supply by skipping doses were not captured, and that some people may have been misclassified by the 60-day rule.1

The chart review covers one health system, and 11.1% of its records gave no reason at all.5 The survey relies on what 440 people remembered and chose to report.6 Neither can say how often a stated reason was the whole story.

The start-year comparison was paid for by the company that employs its authors.4 One author of the chart review reports advisory fees from the makers of both drugs, and the survey was funded by the philanthropy Arnold Ventures.5,6

All of these studies describe the United States, where the price a person pays depends on their insurance. None of them tested a way to help people stay on treatment.

Short answers to common questions

What percentage of people stop GLP-1 medications?

In the largest US records study, 53.6% had stopped within one year and 72.2% within two. Without diabetes the figures were 64.8% and 84.4%. A newer analysis found that six in ten people who started in the first half of 2024 were still on treatment a year later, so the rate has been improving.

What is the number one reason people stop?

Cost and insurance. It was the main reason in 47.6% of medical charts and the most named reason in a national survey, at 36.1%, with a further 28.2% naming a lack of coverage. Side effects came second in both.

How common is stopping because of side effects?

In the trials, between 4.8% and 6.8% of people on Zepbound or Wegovy stopped for good because of adverse reactions, against 3.2% to 3.4% on placebo. Outside trials, side effects were the main reason for 14.6% of people who stopped, and a third of survey respondents named them.

Do people on Zepbound stay on longer than people on Wegovy?

In one health plan analysis, 64.8% of people who started tirzepatide in the first half of 2024 were still on treatment a year later, against 58.6% of those who started semaglutide. The groups were not randomised, so the difference may come from who was prescribed which drug.

Can I restart after stopping?

Many people do: 36.3% of those without diabetes and 47.3% of those with it had filled a GLP-1 again within a year. The Wegovy label says to restart the dose increases at a lower dose after two or more missed weekly doses. The Zepbound label gives no instruction for a longer gap, so the starting dose is for the prescriber to set.

Will the weight come back if I stop?

On average, yes. In the Zepbound withdrawal trial, weight rose 14.0% in the year after the switch to placebo. A review of 37 studies put the average regain at 0.4 kg a month after stopping a weight medicine.

Does it matter where the drug comes from?

In the survey, people who bought from a compounding pharmacy or a website were more likely to stop because they felt the drug was not working: 17.5%, against 6.6% of people who used a retail or mail-order pharmacy.

How the calculator works

The weight calculator turns your two weights into a percentage lost and raises the study's figure to that power. For people without diabetes the study reports a 3.3% lower rate of stopping for each 1% of weight lost, so the calculator multiplies 0.967 by itself once for each point. At 10% lost that gives 0.71, a rate 29% lower. For type 2 diabetes it uses 0.969. The chart stops at 30% lost, and a weight at or above the starting weight shows no reduction.

The regain buttons work the same way with the restart figures: 1.028 per point regained without diabetes and 1.023 with it. The dot charts round the published percentages to whole people. The bars show the published percentages as they are.

The rates are hazard ratios from a model that also took in age, sex, race and ethnicity, income, health conditions and the year of the first fill. They compare groups of people in health records. They are not the chance that any one person will stop.

Sources

  1. Rodriguez PJ, Zhang V, Gratzl S, et al. Discontinuation and reinitiation of dual-labeled GLP-1 receptor agonists among US adults with overweight or obesity. JAMA Netw Open 2025;8:e2457349. doi:10.1001/jamanetworkopen.2024.57349
  2. Xie L, Anazco D, Chancay AH, et al. Glucagon-like peptide-1 receptor agonist switching and treatment persistence in adults without diabetes. JAMA Netw Open 2026;9:e261272. doi:10.1001/jamanetworkopen.2026.1272
  3. Ali D, Kudaya D, Ngige A, Chao AM. Patterns of prescription discontinuation, reinitiation, and switching of subcutaneous semaglutide and tirzepatide in adults with obesity. Diabetes Obes Cardiometab CARE 2026;1:71-75. doi:10.2337/doc25-0026
  4. Marshall LZ, Gleason PP, Friedlander N, Farley J, Urick BY. Trends in 1-year persistence and adherence among initiators of high-potency, weight loss-indicated glucagon-like peptide 1 receptor agonists. J Manag Care Spec Pharm 2026;32:281-291. doi:10.18553/jmcp.2026.32.3.281
  5. Gasoyan H, Butsch WS, Casacchia NJ, et al. Reasons for discontinuation of obesity pharmacotherapy with semaglutide or tirzepatide in clinical practice. Obesity (Silver Spring) 2025;33:2296-2303. doi:10.1002/oby.70058
  6. DiStefano MJ, Zon J, Olson EG, Levy JF, Anderson GF, Alexander GC. Discontinuation of glucagon-like peptide-1 receptor agonists among US adults: a national survey. J Manag Care Spec Pharm 2026;32:903-910. doi:10.18553/jmcp.2026.32.8.903
  7. ZEPBOUND (tirzepatide) prescribing information, NDA 217806/S-041, revised August 2026. U.S. Food and Drug Administration; sections 2.1 to 2.3, 6.1 and 14.1 (Study 4). accessdata.fda.gov
  8. WEGOVY (semaglutide) prescribing information, NDA 215256/S-025, revised June 2026. U.S. Food and Drug Administration; sections 2.2, 2.4, 6.1 and the patient information. accessdata.fda.gov
  9. West S, Scragg J, Aveyard P, et al. Weight regain after cessation of medication for weight management: systematic review and meta-analysis. BMJ 2026;392:e085304. doi:10.1136/bmj-2025-085304

Related analyses

Log each dose, your weight and what each fill cost, so a gap never comes as a surprise.

Open Dose Insider