GLP-1 Trial Dropout Rates Explained: What the Numbers Actually Mean

Which “dropout” number are you actually looking at?

If a news article says “more than half quit,” that's the bottom row. If a trial paper says “7% discontinued,” that's the top row. They're not the same thing — and confusing them is the single most common mistake in coverage of these drugs.

GLP-1 trial dropout rates explained: what “dropout” actually means

A GLP-1 — short for glucagon-like peptide-1 receptor agonist — is a class of injected or oral drugs that lower appetite and slow stomach emptying. The class includes Wegovy and Ozempic (semaglutide), Zepbound and Mounjaro (tirzepatide), Saxenda (liraglutide), and the newly approved oral Foundayo (orforglipron). Trials of these drugs all report some version of “dropout” — but they don't all mean the same thing.

1. Trial completion

The percentage of participants who finished follow-up — meaning they showed up for the final study visit. Trial completion measures whether the study itself was reliable. In modern GLP-1 obesity trials, completion is usually 75% to 90%. People can complete a trial without taking the drug for the full time, and they can stop the drug while still finishing follow-up. These are tracked separately.

2. Treatment discontinuation (any reason)

The percentage who stopped the study drug at any point, for any reason. It includes side effects, but also pregnancy, moving, loss of interest, “withdrew consent,” and being unreachable for follow-up. In SURMOUNT-1, overall treatment discontinuation was 14.3% to 16.4% on tirzepatide — and 26.4% on placebo. More people quit the placebo arm than the active arm. So “treatment discontinuation” tells you almost nothing about tolerability on its own.

3. Discontinuation due to adverse events

The cleanest tolerability number. It's the percentage who stopped because they had a side effect — what trials call an adverse event (any unwanted medical occurrence during the study, whether or not it's clearly drug-related). This is the number you want if you're asking “did people stop because the drug made them sick?” In Phase 3 GLP-1 obesity trials it runs roughly 4% to 11%. In SELECT, the longest semaglutide trial we have at a mean of about 3.3 years, it ran 16.6%.

4. GI-specific discontinuation

A narrower version of #3. The percentage who stopped specifically because of stomach side effects — nausea, vomiting, diarrhea, constipation, or stomach pain. This is the one to look at if your main worry is the GI symptoms everyone hears about. It usually runs 3% to 6%, even though experiencing GI symptoms is much more common (more on that below).

5. Real-world discontinuation

The totally different beast. Instead of being measured inside a controlled trial, real-world discontinuation comes from insurance claims or pharmacy refill data — researchers look at whether people kept filling their prescription. It rolls cost, prior authorization denials, supply shortages, life disruptions, achieving goals, and side effects into one number. This is where you get the famous “85% quit in 2 years” and “more than half quit in 1 year” figures. It's a real measurement, but it answers a different question than trial dropout.

The four-question test

Before you let any GLP-1 dropout number worry you, ask:

1. Who was studied? Adults with obesity? Adults with type 2 diabetes? Both?
2. What drug and dose? Higher doses tend to have more discontinuation. Different drugs are not interchangeable.
3. What kind of discontinuation? One of the five above.
4. Compared with what? Placebo dropout matters. A 7% AE-driven dropout looks fine next to 3% placebo, and alarming next to 0.5% placebo.

Run any headline through those four questions and you'll find that most of them collapse.

GLP-1 trial dropout rates by drug: the verified 2026 table

We pulled every row below from the original peer-reviewed publication, manufacturer official trial release, or current FDA label. Sources are cited at the bottom of the page.

The master trial dropout table

Four patterns in the data

What this table is not

It's not a ranking. You cannot read it as “Saxenda has the highest dropout, so Saxenda is the worst.” The trials had different populations, different durations, different titration schedules, different comparator arms, and different definitions of “GI event.” Newer drugs have newer titration protocols designed to reduce GI side effects. The honest comparison is drug vs placebo in the same trial, not drug vs drug across different trials. That's why we showed “excess over placebo” — that number is fair within a row.

For a full head-to-head on weight loss outcomes, see our semaglutide vs tirzepatide dosage comparison.

How many people stop because of stomach side effects?

This is one of the biggest misreadings of trial data online. Trial reports show two GI numbers: how many people had nausea, vomiting, diarrhea, or constipation at any point, and how many stopped the drug because of it. The first number is much bigger than the second. Most people who get nauseated keep going. Some adjust their dose down and stay on. Most GI events happen during the first few weeks of dose escalation, then fade.

Here's the gap in plain numbers:

Reading that one row from STEP 1 carefully: nearly three out of four people on Wegovy 2.4 mg had some kind of GI event during the trial. About one out of 22 people stopped the drug because of one. That ratio — roughly 16 to 1 — is the actual signal hiding inside the scary headline.

Timing matters: GI side effects are heavily front-loaded. Across the STEP and SURMOUNT programs, most GI events happened during the first 20 weeks of dose escalation, then dropped off. If you're in your first month and feel terrible, you are at the moment in treatment when most other people felt terrible too — and most of them did not quit. This is not a reason to push through severe symptoms in silence. It's a reason to call your prescriber about dose adjustment instead of stopping entirely.

Why real-world GLP-1 dropout rates look so much higher than trial rates

In a 2021–2023 cohort of 125,474 U.S. adults studied in JAMA Network Open, 64.8% of people without type 2 diabetes and 46.5% of people with type 2 diabetes stopped their GLP-1 within one year. [17] A separate Prime Therapeutics study of commercially insured weight-loss starters in 2021 found that only 32.3% were still on the drug at one year and only 15% at two years — the source of the “1 in 7” headline. [18] Both findings are real. Neither is about whether the drug works.

Trials remove barriers that real life doesn't

Here's what a clinical trial gives a participant that real life almost never gives a patient:

• The drug is free. Trial participants don't pay the $1,000+ list price.
• Supply is guaranteed. Trial supply isn't subject to shortages.
• Titration is supervised. Trial protocols slowly increase the dose with built-in pauses.
• Follow-up is intensive. Trial visits happen every 4 to 8 weeks with a study coordinator on call.
• Side-effect management is structured. Anti-nausea support, dose holds, and dose adjustments happen by protocol.
• Selection is filtered. Trial participants are pre-screened — people unlikely to tolerate the class often don't make it in.

Now remove all of that. Charge $1,000 a month. Make the patient call their insurance for prior authorization. Make the pharmacy out of stock for three weeks. Schedule the next prescriber visit for three months out. Hand them a self-injection pen and a one-page side-effect handout. That's real life. The drop in persistence is not a mystery.

The trial-vs-real-world gap, broken down

Editorial synthesis combining Prime Therapeutics, Blue Health Intelligence, Evernorth, JAMA Network Open, and Cleveland Clinic data. Ranges reflect how different studies measure different things. Not a measured share from a single cohort.

Honest takeaway: Non-tolerability factors account for most of the estimated real-world burden in this synthesis. If cost and coverage barriers were reduced, real-world discontinuation would probably move closer to trial discontinuation — though it wouldn't become identical, because real life still includes lower-touch follow-up, switching, intentional pauses, and personal preference. The drug isn't the only thing under stress when someone stops a GLP-1. The system around the drug is too.

The five real reasons people stop a GLP-1

1. Cost

The list price of Wegovy is around $1,350 a month. Zepbound is roughly the same. Even with commercial insurance and manufacturer copay cards, many patients face out-of-pocket exposure of $300 to $600 a month. A Danish national cohort study published in 2025 tracked 77,310 semaglutide users and found that people in lower-income areas were 14% more likely to discontinue within a year, and people aged 18 to 29 were 48% more likely. Both findings line up with the cost driver — younger and lower-income patients face the steepest affordability problem.

2. Coverage changes and prior authorization

The other half of the cost story. Even when a patient is on coverage in January, formularies update, employers re-bid, and a drug that was tier 2 in 2024 may need prior authorization in 2025. An AJMC analysis of ACA marketplace plans found that 86% required prior authorization and 96% included quantity limits for obesity GLP-1 coverage. [26] When a renewal denial lands, the patient often stops — not by choice but by force.

3. Side effects

This is the part that gets the most airtime in popular coverage. In claims-based discontinuation studies it sits behind cost, but in the Evernorth 2025 consumer report, 43.7% of consumers who had used and stopped GLP-1s cited side effects or safety concerns as a reason, with 30.9% citing financial or insurance barriers. [21] The methodologically honest read: both are major drivers, and which one comes first depends on whether you're counting refill gaps or asking people why they stopped.

4. Supply shortages

This was a major driver in 2022 and 2023, when Wegovy and Zepbound were both on the FDA shortage list and patients faced months-long pharmacy gaps. The FDA resolved the tirzepatide injection shortage on October 2, 2024 (reaffirmed December 19, 2024 after a legal challenge), and the semaglutide injection shortage on February 21, 2025. [25] As of mid-2026, pharmacy-level supply hiccups still happen, especially for newer doses, but supply as a quit driver has been falling for two years.

5. Goal achievement and intentional pause

This one almost never makes the headlines. A meaningful share of people stop because they hit a weight or A1c goal and decide to pause. Some plan to restart if they regain. Prime Therapeutics found that 26% of patients had switched GLP-1 drugs during therapy by year two, indicating this often isn't a hard “stop” — it's more like a “shift.” The pharmacy-claims dataset codes all of it as “discontinued.”

If a coverage loss is forcing you off, see how to appeal a Zepbound prior authorization denial.

Why real-world GLP-1 dropout is actually getting better fast

Prime Therapeutics published an updated analysis of 33,607 insured weight-loss patients in March 2026. [19] The 1-year persistence rate for new starters jumped from 33.2% in 2021 to 60.9% in the first half of 2024. For tirzepatide specifically, persistence was already 64% when it was introduced in 2023 and climbed to 64.8% by the first half of 2024. For semaglutide, persistence climbed from 33.2% to 58.6% over the same window.

In plain language: a person who started a GLP-1 in 2024 is roughly twice as likely to still be on it at the one-year mark as someone who started in 2021. Same drugs. Same general population. Different ecosystem around them.

What changed

• Shortages ended. Tirzepatide injection cleared the FDA shortage list on October 2, 2024; semaglutide injection cleared it on February 21, 2025.
• Coverage improved unevenly. Some commercial plans broadened GLP-1 obesity coverage; others tightened it. In a 2026 Business Group on Health survey, 67% of large employers said they covered GLP-1s for weight management. [27] CMS announced a Medicare GLP-1 Bridge program beginning July 1, 2026. Coverage is moving, but not in one direction everywhere.
• Prescriber experience grew. Endocrinologists and obesity-medicine doctors now manage these drugs daily. Better titration support means fewer early quits.
• Patient knowledge improved. Patients starting in 2024 are more informed about the dose-escalation window than patients starting in 2021 were.

If you read a headline based on 2021 data, you're reading about an environment that no longer exists. The 2024 number is closer to where the real-world rate is heading: around 40% dropping out in the first year, with most of those drops still driven by cost.

For help finding a provider with strong coverage support, see best GLP-1 providers that accept insurance.

Who's most likely to drop out — and how to lower your own risk

A persistence-barriers checklist (not a clinical tool)

An editorial checklist based on the literature, not a validated clinical score. Each “+1” reflects a common dropout risk factor from the real-world cohorts reviewed above.

Use the Find My Path quiz to explore coverage and provider options for your specific situation.

What happens to your body if you stop a GLP-1

In the STEP 1 extension trial, participants who had been on semaglutide 2.4 mg for 68 weeks were taken off the drug. Within one year, they regained about 67% of the weight they had lost. People who had been on placebo regained 95% of theirs. By the end of 120 weeks, the semaglutide group was still 5.6 percentage points below their starting weight, but most of the early loss was gone.

SURMOUNT-4 tested what happens specifically when you stop tirzepatide. After 36 weeks of open-label tirzepatide (mean weight loss 20.9%), participants were randomized to keep going or switch to placebo. From week 36 to week 88, the placebo-switch group regained 14% of body weight. The group that stayed on tirzepatide lost another 5.5%. [5] By week 88, the gap between the two groups was about 25 percentage points of body weight.

There's a newer, more hopeful wrinkle for real-world stopping. In a 2025 Cleveland Clinic analysis of patients in Ohio and Florida who stopped semaglutide or tirzepatide, regain was less rapid than the trial models predicted. [24] The likely reasons: trial discontinuation is abrupt, while real-life stopping is often gradual, with patients switching to older obesity medications, using lifestyle changes, or restarting GLP-1 therapy later when access improved.

There are also new step-down options. The ATTAIN-MAINTAIN trial, published May 12, 2026 in Nature Medicine, tested switching from injectable Wegovy or Zepbound to oral orforglipron for 52 weeks. In the tirzepatide cohort, people who switched to orforglipron maintained 74.7% of their prior weight reduction; people switched to placebo maintained 49.2%. In the semaglutide cohort, orforglipron maintained 79.3% of prior weight reduction vs 37.6% for placebo. [9] Whether this becomes standard practice will depend on payer coverage and prescriber adoption, but it's the first major signal that “what to do when you want to stop the injectable” might soon have a structured medical answer.

The bigger point: stopping a GLP-1 reverses most of the weight loss for most people, in line with how obesity behaves as a chronic condition. The drug treats the condition while you're on it. It doesn't cure it. That's worth knowing before starting, and worth talking to your prescriber about before stopping.

Can you actually compare GLP-1 dropout rates between drugs?

Here's why cross-trial ranking misleads:

• Trial protocols differ. SURMOUNT-1 ran 72 weeks; STEP 1 ran 68 weeks; SCALE Obesity ran 56 weeks. A 16-week difference is a meaningful gap in tolerability exposure.
• Populations differ. SELECT enrolled adults with established cardiovascular disease. SURMOUNT-1 excluded them. STEP 2 was specifically people with type 2 diabetes. You cannot read these three studies as if they were the same people taking different drugs.
• Comparator arms differ. Some trials use placebo. Some use older active drugs. The SURMOUNT-5 comparison between tirzepatide and semaglutide is the only modern apples-to-apples drug-vs-drug obesity trial.
• Definitions differ. What counts as a “GI adverse event” varies subtly trial by trial. Even what counts as “treatment discontinuation” can vary if titration interruptions are counted differently.
• Titration schedules differ. The 2015 SCALE Obesity protocol escalated liraglutide to 3.0 mg over 5 weeks. The 2022 SURMOUNT-1 protocol escalated tirzepatide over 20 weeks. Slower titration produces lower GI dropout — that's part of why newer trials read more favorably than older ones.

The honest comparison is excess-over-placebo within a trial, not raw dropout rate across trials. Looking at the master table, you can fairly say “tirzepatide 5 mg in SURMOUNT-1 had +1.7 pp excess discontinuation vs placebo.” You can't fairly say “tirzepatide is more tolerable than semaglutide” unless you're citing SURMOUNT-5 — and even then the answer was only “modestly so” (6.1% vs 8.0%).

Is a high GLP-1 dropout rate a safety warning?

When SELECT reported 16.6% adverse-event-driven discontinuation on semaglutide over 3.3 years, the dropouts were overwhelmingly driven by GI symptoms — nausea, vomiting, constipation. They weren't driven by pancreatitis or cancer or anything that put the drug's safety in question. In fact, serious adverse events were lower in the semaglutide arm of SELECT than in the placebo arm (33.4% vs 36.4%), driven by the cardiovascular benefit.

For specific severe outcomes, the right reference is each drug's FDA label. The Zepbound USPI reports adjudication-confirmed acute pancreatitis at 0.2% in the active arm and 0.2% on placebo in weight-reduction trials — no signal beyond placebo. [12] Every approved GLP-1 carries a boxed warning about a possible thyroid C-cell tumor signal seen in rodent studies, and labels list contraindications for people with a personal or family history of medullary thyroid cancer or MEN 2.

A high dropout rate tells you: “this drug is hard for some people to live with.” It doesn't tell you: “this drug is dangerous.” Those are different questions with different evidence. If you're worried about safety specifically, look at:

• The boxed warnings and contraindications on the FDA label for the specific drug
• The adjudicated serious adverse event rates (not just any-AE rates)
• The cardiovascular outcomes data (semaglutide has the strongest, tirzepatide is catching up)
• The FDA Adverse Event Reporting System for post-marketing signals

How GLP-1 dropout compares to other chronic medications

A Danish national cohort study of 113,207 SGLT2 inhibitor users found a one-year discontinuation rate of 27.9% and a three-year rate of 45.9%. Statins — the most widely prescribed chronic medication in the U.S. — have non-persistence rates that often cross 50% within five years. Hypertension medications drop 30% to 40% in the first year in most large studies.

GLP-1s, sitting at roughly 40% one-year discontinuation in the most current commercial cohorts, are roughly in line with that broader pattern. The thing that makes GLP-1 dropout feel worse is the gap between what trials promised (15% to 20% weight loss) and what real-world cohorts often deliver (5% to 10%, mostly because lower doses and incomplete adherence dilute the average). That gap isn't unique to GLP-1s — it shows up in almost every chronic medication class — but it's stark here because the trial promise is so visually dramatic.

How to read any GLP-1 dropout headline in 30 seconds

Pull up the headline. Then ask:

1. Is this a clinical trial number or a real-world refill number? If it's “85% quit” or “more than half quit in one year,” it's almost certainly real-world. If it's “5% to 17%,” it's clinical trial.
2. What population? Adults with diabetes? Adults without? Older with heart disease (SELECT)? Younger and healthier (SURMOUNT-1)? Each population behaves differently.
3. What time window? Three months? One year? Two years? Four years? Dropout climbs steadily over time on every drug.
4. Compared with what? If there's no placebo or comparator rate, you can't tell signal from baseline. If 7% stopped the drug due to side effects but 3% stopped placebo for the same reason, the drug-attributable excess is 4 percentage points — not 7.

Most viral GLP-1 stories pass none of these four questions cleanly. They're not lying. They're just citing a number out of context.

Three questions to ask your prescriber before stopping a GLP-1

If your prescriber doesn't have answers, ask whether you can be referred to an obesity-medicine specialist. Per Blue Health Intelligence, GLP-1 patients prescribed by endocrinologists or obesity specialists complete the first 12 weeks at higher rates than patients prescribed by general practitioners.

For coverage continuity options, see does insurance cover Zepbound for weight loss and cheapest Zepbound without insurance.

Methodology: what we actually verified

Frequently asked questions

Sources

Related guides

• Semaglutide vs tirzepatide: dosage and comparison
• GLP-1 real world evidence vs clinical trial results: 2026 data
• Does insurance cover Zepbound for weight loss?
• How to appeal a Zepbound prior authorization denial
• Best GLP-1 providers that accept insurance
• Best GLP-1 providers that help with prior authorization
• Cheapest Zepbound without insurance (2026)
• Zepbound Savings Card: eligibility and limits