| Medication | Trial | Week | Mean change | Placebo |
|---|---|---|---|---|
| Semaglutide 2.4 mg (Wegovy) | STEP 1 — Wilding et al., NEJM 2021 | 68 | −14.9% | −2.4% |
| Tirzepatide 5 mg (Zepbound) | SURMOUNT-1 — Jastreboff et al., NEJM 2022 | 72 | −15.0% | −3.1% |
| Tirzepatide 10 mg | SURMOUNT-1 | 72 | −19.5% | −3.1% |
| Tirzepatide 15 mg | SURMOUNT-1 | 72 | −20.9% | −3.1% |
| Oral semaglutide 25 mg (Wegovy pill) | OASIS 4 — Wharton et al., NEJM 2025 | 64 | −13.6% | −2.2% |
| Semaglutide 7.2 mg (Wegovy HD) | STEP UP — Wegovy prescribing information §14 (Study 8); Wharton et al., Lancet Diabetes Endocrinol 2025 | 72 | −18.8% | −3.9% (paper) |
| Liraglutide 3.0 mg (Saxenda) | SCALE — Saxenda prescribing information §14 (Study 1); Pi-Sunyer et al., NEJM 2015 | 56 | −7.4% | −3.0% |
| Tirzepatide vs semaglutide | SURMOUNT-5 — Aronne et al., NEJM 2025 | 72 | −20.2% vs −13.7% | — |
Trials report two estimands: treatment-policy (everyone randomized, whether or not they stayed on the drug) and trial-product (only people who stayed on treatment, which runs higher; STEP UP paper: −18.7% vs −20.7%). This site uses treatment-policy figures for cross-trial comparison: STEP 1 −14.9%, SURMOUNT-1 −20.9%, OASIS 4 −13.6%, STEP UP −18.8%. Where the prescribing information and the published paper differ slightly for the same group, the site uses the prescribing information (STEP UP: −18.8% in the label, −18.7% in the paper; SCALE: −7.4% in the label, −8.0% in the paper; ATTAIN-1: −11.1% for the marketed 17.2 mg dose in the label, −11.2% for the investigational 36 mg formulation in the paper).
Intermediate weeks are read from the published mean-change figures in each paper at the visits reported (typically every 4 weeks to week 20, then every 8) and linearly interpolated between them. Titration schedules in the trials reached the maintenance dose by week 16 (semaglutide) or week 20 (tirzepatide 15 mg); the early part of each curve reflects that ramp.
The shaded band runs from 0.55× to 1.35× the trial mean. It is an illustrative band, not a clinical reference interval: it is not a confidence interval or a percentile range, and it is not calculated from the trial data. It is there as a reminder that people in the trials were spread widely around the mean. For example, in SURMOUNT-1 (tirzepatide 15 mg, everyone randomized, week 72), 90.9% lost at least 5% of their body weight and 56.7% lost at least 20% (Zepbound prescribing information, Table 2). Where your numbers sit against the band or the mean does not say whether a result is normal, whether a treatment is working or whether anything should change, so every tool on this site shows your result as plain numbers next to the trial mean, with no grade. Questions about your dose or progress belong with your prescriber.
Body-composition studies suggest roughly 25–35% of weight lost can be lean mass, varying by drug and study. A 2026 meta-analysis estimated about 25% for tirzepatide and about 35% for semaglutide. Adequate protein and resistance training may help preserve it. 1.2–1.6 g per kg per day is the range most commonly cited in obesity-medicine guidance for preserving lean mass during pharmacological weight loss. With a lot to lose, it is applied to goal weight, not current weight: current weight gives a number that overshoots what muscle needs and is hard to eat on a suppressed appetite. The protein calculator uses the goal weight you enter (current weight as an upper bound); calculators that ask for height use a healthy-range reference weight, the weight at BMI 25 for that height, whenever you are above it. Already near goal: current weight.
Some provider links are affiliate links: if you sign up through one, WeighWeekly may earn a commission at no extra cost to you. Today that means compounded-GLP-1 telehealth providers in the Katalys network whose applications have been approved; those links are marked "Paid affiliate link" or carry rel="sponsored". Providers still under review, manufacturer programs (LillyDirect, NovoCare), savings cards and every other provider are listed without an affiliate link, and a provider with an FDA warning letter from 2025 on never gets one.
Commission never sets the order. Every price table, price card and ranking has the same two billing views. Annual prepay, the default, shows and sorts by a one-year plan's monthly equivalent, always with the amount billed upfront, the term and the refund terms, and the month-to-month price after it; a provider without a one-year plan keeps its month-to-month price. Monthly billing shows and sorts by the month-to-month price, cheapest first, with any first-month price beside it; a price that applies only to the first weeks is labelled that way. Membership, visit and shipping fees are listed apart from the drug where the provider's site states them. Route tables are sorted by the first year from the first dose in the active view: brand-name rows count every fill (a fill lasts 28 days, so a 52-week year is 13 fills, titration fills included), compounded rows count each charge or the one-year bill. Cost per pound is shown only for brand-name routes: it divides the first year by the trial loss at week 52 for that same dose, and no compounded product was tested in those trials, so compounded rows are compared on price only. The cheapest cards and lists follow the state you pick: providers that exclude it move to a separate group, and providers that publish no state list are marked "coverage to be confirmed" and left out of the cheapest cards. Each price shows the date its provider's page last confirmed it; a price that could not be re-confirmed is shown as the last known price, with that date.
The dataset at /trials.json is CC BY 4.0. Cite this site as the source when reusing the interpolated curves; cite the original papers for endpoint values.
One short email a week on GLP-1 treatment: what the trials and prescribing labels describe at each stage, and changes in cost and access. General information, not tied to a particular medication or dose. For the trial average for your own medication, use the curve calculator.
No spam. Unsubscribe any time. Educational information only, not medical advice.