GLP-1 Statistics 2026: The Numbers, Each With Its Source

The most-cited GLP-1 statistic in 2026 comes from KFF, the health policy research organization, whose Health Tracking Poll asks a nationally representative sample of US adults whether they have ever taken a GLP-1 drug. The wave fielded 24 February to 2 March 2026, among 1,343 adults, found that nearly one in five adults (18%) have ever taken a GLP-1 medication, including 12% who report currently taking one (KFF, Public Views on Prescription Drug Costs, Regulation, Affordability and TrumpRx, published 13 March 2026).

Those two numbers were identical in the previous wave. KFF's poll fielded 27 October to 2 November 2025 among 1,350 adults reported 18% ever and 12% currently, and described the 12% as an increase of 6 percentage points from May 2024 (KFF Health Tracking Poll, published 14 November 2025). KFF's May 2024 wave, published 10 May 2024, had reported 12% ever and 6% currently. Read together, current use doubled between May 2024 and November 2025 and then held flat through early 2026.

The November 2025 wave also breaks use down by condition and by source. Among adults with diabetes, 57% had ever taken a GLP-1 and 45% currently did; among adults with heart disease, 40% ever and 29% currently; among adults describing themselves as overweight or obese, 34% ever and 23% currently. Three in ten users (30%) said they took the drug primarily to lose weight. Most users (76%) got the prescription from a primary care doctor or specialist, and about one in six (17%) from an online provider or website. Just over half of users (56%) said the drugs were difficult to afford, including one in four who said very difficult (KFF, 14 November 2025).

One caution belongs in the same paragraph as the headline: these are self-reported survey answers, not pharmacy records, and the poll asks about GLP-1 drugs as a class. It does not separate semaglutide from tirzepatide, injection from tablet, or branded from compounded product.

For prescription and spending trends, the primary source is the IQVIA Institute, whose annual report on US medicine use was published on 28 April 2026 under the title U.S. Medicine Use Trends 2026. Its GLP-1 figure is a spending figure: GIP/GLP-1 agonists across diabetes and obesity contributed $14 billion in growth, with $9.6 billion concentrated in products approved for obesity and related comorbidities. That growth sits inside a larger movement: US net medicine spending rose from $548 billion in 2024 to $606 billion in 2025, a $58 billion increase, and IQVIA notes that most of that growth was driven by protected brands outside GIP/GLP-1 agonists and COVID-19 vaccines.

The report's forward view is a range, not a point: US medicine spending is projected to grow between 6% and 9% a year on a list-price basis and 4.5% to 7.5% after discounts, rebates and other price concessions, over the next five years (IQVIA Institute, 28 April 2026).

What the public report does not state is as important as what it does. It gives no count of GLP-1 patients, no prescription volume and no figure for compounded semaglutide or tirzepatide. Pages that quote a precise number of Americans on a GLP-1 and attribute it to IQVIA are usually citing a paid dataset, a press interview or a secondary article, not the published report. This page prints only the sentences that appear in the report itself.

A second, indirect signal on volume comes from the same KFF wave already cited: if 12% of roughly 260 million US adults currently take a GLP-1, that implies on the order of 30 million people. That arithmetic is ours, not KFF's or IQVIA's, and it inherits every limitation of a self-reported poll, so treat it as a rough scale and not as a statistic. For what the medicines cost an individual and how coverage works, see GLP-1 cost and savings.

The best-documented discontinuation figure is from Rodriguez and colleagues, Discontinuation and Reinitiation of Dual-Labeled GLP-1 Receptor Agonists Among US Adults With Overweight or Obesity, published in JAMA Network Open on 31 January 2025. The authors used Truveta Data, electronic health records from 30 US health systems linked to dispensing records, and followed 125,474 adults with overweight or obesity who started a GLP-1 that is labeled for both diabetes and weight management between 1 January 2018 and 31 December 2023.

At one year, 64.8% of adults without type 2 diabetes (95% CI, 64.4% to 65.2%) and 46.5% of adults with type 2 diabetes (95% CI, 46.2% to 46.9%) had discontinued. Among those who discontinued, 36.3% of the no-diabetes group and 47.3% of the diabetes group had reinitiated within a year of stopping. The 95% confidence intervals are narrow because the cohort is large; the width of the interval says nothing about why people stopped.

That why is the part this page cannot resolve. The study period runs through 2023, so it captures the years of supply shortages, coverage changes and compounded alternatives, and a discontinuation in the dispensing record can mean intolerance, cost, a coverage denial, a pharmacy switch or a deliberate pause agreed with a prescriber. The trial discontinuation rates in the next section are much lower precisely because trials remove most of those reasons. If you are weighing whether to stop or pause, the stopping or pausing a GLP-1 page describes what a record should hold for that conversation with a prescriber; the decision itself is not one a statistics page can inform.

Two pivotal trials supply almost every weight-loss percentage quoted about this class, and both are worth reading at the level of the abstract.

STEP 1, published as Once-Weekly Semaglutide in Adults with Overweight or Obesity by Wilding and colleagues in the New England Journal of Medicine on 18 March 2021, randomized 1,961 adults without diabetes to once-weekly semaglutide (the molecule in Wegovy, from Novo Nordisk) or placebo, alongside lifestyle intervention, for 68 weeks. The mean change in body weight was -14.9% with semaglutide versus -2.4% with placebo, a treatment difference of -12.4 percentage points (95% CI, -13.4 to -11.5). 86.4% of semaglutide participants lost at least 5% of body weight versus 31.5% on placebo. More semaglutide participants discontinued because of gastrointestinal events, 4.5% versus 0.8%.

SURMOUNT-1, published as Tirzepatide Once Weekly for the Treatment of Obesity by Jastreboff and colleagues in the same journal on 21 July 2022, randomized 2,539 adults without diabetes to one of three tirzepatide arms (the molecule in Zepbound, from Eli Lilly) or placebo for 72 weeks. The mean percentage change in weight at week 72 was -15.0%, -19.5% and -20.9% across the lowest, middle and highest arms versus -3.1% with placebo, and 85%, 89% and 91% of participants in those arms lost at least 5% versus 35% on placebo. Discontinuation because of adverse events was 4.3%, 7.1% and 6.2% across the arms and 2.6% on placebo.

Three things about these figures are easy to lose. They are means, so half of each group did better and half did worse. They come from trials that excluded people with diabetes and ran on a fixed escalation protocol with structured lifestyle support. And they are not head-to-head: the two trials differ in duration, population and design, so the gap between 14.9% and 20.9% is not a measured difference between the molecules. For how the two molecules differ on the label, see semaglutide vs tirzepatide; for why an individual curve rarely matches a trial mean, see GLP-1 weight loss plateau.

Side-effect percentages belong to a specific label, not to the class, because each label's table pools different trials over different durations. The FDA-approved Wegovy prescribing information (Novo Nordisk, revised 06/2026, via DailyMed, accessed 14 September 2026) lists in its Table 3, for adults with obesity or overweight, nausea in 44% of Wegovy-treated adults versus 16% on placebo, diarrhea 30% versus 16%, vomiting 24% versus 6%, constipation 24% versus 11% and abdominal pain 20% versus 10%. The same label states, for the same Table 3 trial pool, that 6.8% of adults on Wegovy permanently discontinued treatment because of adverse reactions versus 3.2% on placebo, and that the most common reasons were nausea (1.8% versus 0.2%), vomiting (1.2% versus 0%) and diarrhea (0.7% versus 0.1%).

The FDA-approved Zepbound prescribing information (Eli Lilly, revised 08/2026, via DailyMed, accessed 14 September 2026) lists in its Table 1, pooled from its two weight-reduction trials, nausea in 25% to 29% of adults across the three Zepbound arms versus 8% on placebo, diarrhea 19% to 23% versus 8%, vomiting 8% to 13% versus 2% and constipation 11% to 17% versus 5%. Across both trials, 4.8% to 6.7% of Zepbound-treated adults permanently discontinued because of adverse reactions versus 3.4% on placebo, and gastrointestinal adverse reactions occurred in 56% of each Zepbound arm versus 30% on placebo. The full label facts are on Wegovy (semaglutide) and Zepbound (tirzepatide); the plain-language overview is GLP-1 side effects.

Prices are the least stable numbers on this page, so this section prints only what a manufacturer publishes on its own site, with the date it was read. Novo Nordisk's wegovy.com savings page, read 14 September 2026, lists self-pay pricing through NovoCare Pharmacy of $349 per month for the Wegovy injection strengths and $399 per month for Wegovy HD, with a limited-time introductory offer of $199 per month for the first two monthly fills of the starting strengths through 31 December 2026, and $149 per month for the starting strength of the Wegovy pill. Eli Lilly's zepbound.lilly.com coverage and savings page, read 14 September 2026, states a regular self-pay price through LillyDirect of $299 to $699 per one-month supply of Zepbound KwikPen by strength, and a reduced price with the Zepbound self-pay savings card that brings the higher strengths down to a ceiling of $449 per month, with the card expiring 31 December 2026. These are cash prices offered by the manufacturers; what an insured patient pays depends on the plan, and the pharmacist is the person who can quote it.

Every statistic on this page answers a different question, and the mistakes people make come from swapping one for another.

A poll percentage (KFF's 18% and 12%) describes a population at a moment, from self-report, with a margin of error the poll publishes and a definition of GLP-1 that the respondent supplies. A spending figure (IQVIA's $14 billion) describes dollars at net prices after rebates, so it does not track prescriptions one-for-one. A cohort discontinuation rate (64.8%) describes what happened in dispensing records under real-world coverage, supply and cost, and it cannot tell you why. A trial mean (-14.9%, -20.9%) describes the average of a selected group on a fixed protocol with lifestyle support, and the placebo arm (-2.4%, -3.1%) is part of the result, not a footnote. A label frequency (nausea 44%) is a count of any report during a trial, regardless of severity or duration, and is only comparable within the same table. A self-pay price ($349, $299) is a manufacturer's cash offer on the day it was read.

None of these is a prediction about one person. A trial mean does not say what your scale will show; a discontinuation rate does not say whether you will stop; a nausea frequency does not say whether you will feel sick in week three. What does answer those questions is your own record, kept consistently enough to be compared against the numbers above. GLP-1 Tracker, a free iPhone app, is built for that layer only: it logs each dose with its date, your weight over time, your meals with protein and fiber estimates, and a dated note when something changes, all stored on your device. It does not compute where you sit relative to STEP 1, and it should not.

This page is dated 14 September 2026. KFF polls in waves, IQVIA publishes each spring, labels are revised and manufacturer prices change with a footnote. If any figure here matters to a decision, open the named source on the day you need it, and take the question to the prescriber and pharmacist who can see the record a statistics page cannot.