One NHS GLP-1 prescription item dispensed in England, August 2025 to July 2026 (5,697,913 items).
NHSBSA English Prescribing Data, retrieved 15 Oct 2026Data · GLP-1
GLP-1 by the numbers.
How many people use GLP-1 medicines, and what do the trials show?
In England, the NHS dispensed one GLP-1 prescription item every 5.53 seconds between August 2025 and July 2026. In the US, Medicare Part D processed one GLP-1 claim every 1.05 seconds in 2025. In two large trials, average weight loss was 15% to 21%; in one, two-thirds returned within a year of stopping.
Updated
Every few seconds
How often a GLP-1 is dispensed.
Counts of prescriptions and claims, divided by the seconds in the period. The arithmetic is in the downloads.
The same in July 2026 alone (566,555 items), the latest month published.
NHSBSA English Prescribing Data, July 2026One Medicare Part D GLP-1 claim in the US in 2025 (30,143,652 claims).
CMS Medicare Part D Spending by Drug, 2025Medicare Part D gross GLP-1 spend per second in 2025 ($39.6bn for the year, before rebates).
CMS Medicare Part D Spending by Drug, 2025England
NHS GLP-1 prescriptions, 2014 to 2025.
Tirzepatide (Mounjaro) went from 9 NHS items in all of 2023 to 652,662 in 2024 and 2,640,219 in 2025. All GLP-1 items rose from 733,855 in 2014 to 4,824,227 in 2025. NHSBSA English Prescribing Data, 2014–2025.
By month, items rose from 60,443 in January 2014 to 566,555 in July 2026, 9.4 times as many.
- Tirzepatide
- Semaglutide
- Other GLP-1s
See the numbers
| Year | All GLP-1 items | Semaglutide | Tirzepatide |
|---|---|---|---|
| 2014 | 733,855 | 0 | 0 |
| 2015 | 779,463 | 0 | 0 |
| 2016 | 847,222 | 0 | 0 |
| 2017 | 938,313 | 0 | 0 |
| 2018 | 1,093,380 | 2 | 0 |
| 2019 | 1,332,864 | 79,105 | 0 |
| 2020 | 1,622,931 | 323,781 | 0 |
| 2021 | 2,006,629 | 632,183 | 0 |
| 2022 | 2,628,772 | 1,059,823 | 0 |
| 2023 | 2,850,037 | 1,207,666 | 9 |
| 2024 | 3,139,829 | 1,542,440 | 652,662 |
| 2025 | 4,824,227 | 1,548,832 | 2,640,219 |
What this leaves out. Private prescriptions, which were most UK weight-loss use until the NHS tirzepatide rollout in 2025, and hospital dispensing. Semaglutide includes Ozempic, Rybelsus and Wegovy.
Cost. List-price cost in January to July 2026 was £617.8m, already more than the whole of 2025 (£608.1m). That is before confidential discounts, and the step up from September 2025 reflects a tirzepatide list-price change, not only more items.
United States
Medicare Part D GLP-1 spending, 2020 to 2025.
Gross Medicare Part D spend on GLP-1 medicines rose from $7.2bn in 2020 to $39.6bn in 2025, and claims from 6,018,487 to 30,143,652. CMS Medicare Part D Spending by Drug, 2020–2025.
See the numbers
| Year | Gross spend | Claims |
|---|---|---|
| 2020 | $7,216,574,629 | 6,018,487 |
| 2021 | $10,194,232,322 | 7,885,425 |
| 2022 | $14,206,699,695 | 10,406,887 |
| 2023 | $22,606,361,233 | 16,553,308 |
| 2024 | $27,904,791,306 | 22,164,471 |
| 2025 | $39,626,929,489 | 30,143,652 |
of US adults say they have ever taken a GLP-1 (12% are taking one now), February to March 2026. It was 12% ever in April to May 2024.
KFF Health Tracking Poll, March 2026 toplineof adults in Great Britain used a GLP-1 in the past year, 2.9% to support weight loss, January to March 2025.
Jackson et al., BMC Medicine, 2026Trials
What happened after people stopped.
In the STEP 1 extension, people on semaglutide 2.4 mg had lost 17.3% of their weight on average by week 68. A year after stopping, at week 120, the net loss was 5.6%: they had regained about two-thirds of what they lost. Wilding et al., Diabetes, Obesity and Metabolism, 2022.
- Semaglutide 2.4 mg, stopped at week 68
- Placebo
See the numbers
| Week | Semaglutide (digitised) | Placebo (digitised) |
|---|---|---|
| 0 | 0% | 0% |
| 4 | -2.37% | -1.04% |
| 8 | -4.35% | -1.58% |
| 12 | -6.23% | -2.13% |
| 16 | -8.11% | -2.27% |
| 20 | -10.04% | -2.52% |
| 28 | -12.41% | -2.67% |
| 36 | -14.09% | -2.72% |
| 44 | -15.53% | -2.72% |
| 52 | -16.62% | -2.67% |
| 60 | -17.01% | -2.32% |
| 68 | -17.31% | -2.03% |
| 75 | -13.85% | -1.68% |
| 80 | -11.27% | -0.99% |
| 104 | -6.68% | -0.4% |
| 120 | -5.59% | -0.05% |
Weight change from week 36 to 88 for people switched to placebo in SURMOUNT-4. Those who stayed on tirzepatide lost a further 5.5%.
Aronne et al., JAMA, 2024Weight change from week 20 to 68 for people switched to placebo in STEP 4. Those who stayed on semaglutide lost a further 7.9%.
Rubino et al., JAMA, 2021For week-by-week trial averages while on treatment, use the GLP-1 results by week tool.
Body composition
Fat and lean in the weight lost.
In the SURMOUNT-1 DXA substudy, about 74% of the weight lost on tirzepatide was fat and 26% was lean mass, almost the same split as placebo (75% and 25%). Fat mass fell 33.9% and lean mass 10.9%. Look et al., Diabetes, Obesity and Metabolism, 2025, week 0 to 72.
STEP 1 (semaglutide 2.4 mg). Among the 83 people on semaglutide with DXA results (39 on placebo), fat mass fell 9.3 kg and lean mass 5.8 kg on average by week 68. ClinicalTrials.gov, NCT03548935 results. Lean was 38% of the fat-plus-lean loss; that share is our arithmetic, not a published figure.
For comparison. A widely cited rule of thumb for diets is that about a quarter of weight lost is fat-free mass. The review that describes it also questions it. Heymsfield et al., Obesity Reviews, 2014.
Downloads
Take the data.
Every CSV behind this page. Each one lists its source, retrieval date and caveats in its columns. Retrieved 15 October 2026.
- england_glp1_items_yearly.csv
NHS GLP-1 prescription items and list-price cost in England by year, 2014 to 2026 (2026 is January to July).
Open Government Licence v3.0 - england_glp1_items_monthly_wide.csv
The same, by month and by medicine, January 2014 to July 2026.
Open Government Licence v3.0 - every_x_england.csv
The “one item every 5.53 seconds” arithmetic, row by row.
Open Government Licence v3.0 - us_cms_glp1_spending_totals.csv
Medicare Part D and Medicaid GLP-1 gross spend and claims by year.
US government public data - every_x_us_medicare.csv
The “one claim every 1.05 seconds” arithmetic.
US government public data - trial_weight_endpoints.csv
Published trial results (STEP 1, STEP 4, SURMOUNT-1, SURMOUNT-4, SURMOUNT-5) with the type of estimate.
Facts quoted from abstracts and trial registries - trial_weight_curves_digitised.csv
Week-by-week average weight change read from published figures. Every point is digitised, not published. SURMOUNT-1 was re-read marker by marker on 26 October 2026; a point that can’t be read separately is marked NOT_READABLE and left blank.
Digitised values; figures not reused - trial_weight_curves_validation.csv
Digitised values checked against the numbers printed on each figure (within 0.25 percentage points).
Facts - body_composition.csv
DXA fat and lean changes from the STEP 1 and SURMOUNT-1 substudies.
Facts quoted; some rows cited second-hand (marked)
Sources
- NHSBSA, English Prescribing Data (EPD), January 2014 to July 2026Open Government Licence v3.0. Retrieved 15 October 2026.
- CMS, Medicare Part D and Medicaid Spending by Drug, 2020 to 2025US government public data. Gross spend before rebates.
- KFF Health Tracking Poll, March 2026 topline (Q9 trend table)1,343 US adults, fieldwork 24 February to 2 March 2026.
- Jackson et al., BMC Medicine, 2026: GLP-1 use in Great BritainHousehold survey, January to March 2025.
- Wilding et al., Diabetes, Obesity and Metabolism 2022;24:1553–1564 (STEP 1 extension)Curve points digitised from Figure 1A.
- Rubino et al., JAMA 2021;325:1414–1425 (STEP 4)
- Aronne et al., JAMA 2024;331:38–48 (SURMOUNT-4)
- Look et al., Diabetes, Obesity and Metabolism, 2025 (SURMOUNT-1 DXA substudy)
- ClinicalTrials.gov NCT03548935 results (STEP 1 DXA substudy)
- Heymsfield et al., Obesity Reviews, 2014 (the quarter rule)
These are population counts and trial averages. They don’t predict what any one person will see, and they aren’t advice about any medicine. Talk to your prescriber or pharmacist about your own treatment.