GLP-1 receptor agonists, dual and triple agonists and amylin analogues in one database, with a primary source behind every number. The maps are built from the database and rebuilt on every update; the feed below records what changed.
Every database update is a short entry: which data were added and which maps moved. This is the «blog» of the atlas — the maps live, the entries stay.
One e-mail per new entry in the feed — never more often than the database changes. Confirmation by link; one-click unsubscribe.
New results found by the atlas monitor in September–October 2026 were added after two independent extractions with verbatim quotes. In type 2 diabetes, CagriSema 1.0/1.0 mg reduced body weight by 12.4 % versus 9.1 % with tirzepatide 5 mg at week 60 (REIMAGINE 5), and the eloralintide + tirzepatide combination (EloraTZP) by up to 23.3 % versus 14.8 % with tirzepatide 15 mg and 3.0 % with placebo at week 48.
Company releases are marked as such; peer-reviewed values replace them when the papers appear.
The cardiovascular outcome trial of orforglipron versus insulin glargine in 2,749 adults with type 2 diabetes at increased cardiovascular risk (median follow-up 104 weeks) is now on the forest plot. The primary endpoint, 4-point MACE, met non-inferiority: hazard ratio 0.84 (95 % CI 0.59–1.20). Cardiovascular death: 0.47 (0.25–0.90).
Two independent extractions agreed; every hazard ratio was checked against the text of the publication.
Three new peer-reviewed publications and one company announcement were picked up by the atlas monitor and added after two independent extractions with verbatim quotes. In people with obesity and type 2 diabetes, retatrutide 12 mg reduced body weight by 20.8 % (efficacy estimand; placebo 4.0 %) at 80 weeks in TRIUMPH-2, and survodutide 6.0 mg by 13.1 % (placebo 3.1 %) at 76 weeks in SYNCHRONIZE-2. The amylin analogue petrelintide is now shown per dose from the ZUPREME-1 paper.
Discontinuation percentages computed from the participant counts reported in the trial flow diagrams are accepted when both extractions agree.
The outcome layer now goes beyond the primary composites. From the 13 cardiovascular and kidney outcome trials we added 34 hazard ratios for secondary and safety endpoints, and the primary results of outcome trials in heart failure, sleep apnoea, MASH, chronic kidney disease and Alzheimer's disease are shown as effects versus placebo. The registry layer was rebuilt: trials registered under new international non-proprietary names or company codes are now matched, and six molecules were added.
Every added value was extracted twice independently and accepted only when both extractions agreed and the quote was found verbatim in the source.
We added the Chinese molecules that were missing and went through the companies' own press releases and stock-exchange filings (HKEX, Shanghai and Shenzhen), plus publications and EASD 2026 abstracts. Every number carries a verbatim quote from its source and a link to it; the quotes were checked automatically against the saved source texts. These values are marked «from company release or filing» and have not yet been through the two-model check.
109 arms of Chinese-origin molecules in the database. Where sources disagree (e.g. GLORY-2: journal vs press release) the card says so.
Every cell of the matrix was checked against the current FDA labels (openFDA, DailyMed), the EMA medicines register, NMPA announcements, ClinicalTrials.gov (1,598 interventional studies of the 20 molecules) and published or press-released phase 3 results. Each cell now opens its evidence. The matrix grew from 12 to 18 molecules.
Classification codes: approved / positive phase 3 / phase 3 ongoing or results pending / phase 2 / negative.
The cardiovascular and kidney outcome trials are now broken down by endpoint. For all 13 trials in the forest plot, hazard ratios with 95 % confidence intervals and event counts were extracted from the primary publications by two independent language models; all 85 values agree between the models and each carries a verbatim quote. The primary-endpoint hazard ratios of the first release were confirmed without change.
The molecule × indication matrix is still a first-release classification and will be reconciled with labels and registries next.
Every arm in the database was re-extracted from its primary source — the full paper, the congress abstract or the company press release — by two independent language models, each value with a verbatim quote and its location in the source. Values on which both models agree are now marked «verified-2» and replace the first-release seed figures; the supporting quote is shown when you click a point. Values on which the models disagree stay unchanged until a human reviewer decides.
Cardiovascular outcome hazard ratios (forest plot) and the indication matrix are still first-release values and will be verified next.
First release. The base covers the key phase 2–3 trials in obesity and type 2 diabetes (STEP, SURMOUNT, REDEFINE, retatrutide phase 2, survodutide, mazdutide, MariTide, amycretin, orforglipron) plus 42 key results from the EASD 2026 congress (Milan, 29 Sep – 3 Oct) drawn from our archive of 1,309 abstracts and 19 session transcripts.
Figures of the first release are tagged «seed» and are being verified against the primary sources.
Horizontal axis — the price of the effect (patients discontinuing for adverse events, any discontinuation, nausea, vomiting or serious adverse events); vertical axis — the effect (weight, HbA1c or systolic blood pressure). Lower and further left is better. Shape = mechanism, colour = molecule, dashed outline = trial run in China. This is a cross-trial comparison: it shows direction, not a head-to-head result.
Data cut 14 Oct 2026Hover a point for the conditions and the source; click to highlight it in the table. Arms with no data on the chosen axes are listed in the table only.
All trial arms under the current filters, sorted by effect, molecule or duration. These filters are the same as on Map 1 — change them in either place. Label — arm and duration; colour — molecule. In T2D the weight effect is consistently smaller than in obesity without diabetes: compare within a population.
Data cut 14 Oct 2026What the class has proven on hard endpoints, and which molecules have an outcome trial at all. The forest plot shows hazard ratios with 95 % confidence intervals from the published outcome trials; the matrix shows how far each molecule has reached in each indication; the timeline of all registered trials is Map 4. Older GLP-1 receptor agonists are included as the reference history of the class.
Data cut 14 Oct 2026Hazard ratio with 95 % CI; square size ∝ trial size; left of 1.0 favours the drug. Tap a row for details and sources.
Primary endpoint versus placebo; each block has its own scale (difference, or ratio on a log scale). Molecule, mechanism and company filters above apply. Tap a row for the quote and source.
| Molecule | Obesity | T2D | CV outcomes | CKD | HFpEF | MASH | Sleep apnoea | Knee OA | Alzheimer's |
|---|---|---|---|---|---|---|---|---|---|
| Semaglutide GLP-1 | approved | approved | approved | approved | phase 3 + | approved | — | phase 3 + | negative |
| Tirzepatide GLP-1/GIP | approved | approved | approved | phase 2 | phase 3 + | phase 3 | approved | phase 3 | — |
| Liraglutide GLP-1 | approved | approved | approved | phase 2 | phase 2 | phase 2 | phase 3 + | negative | phase 2 |
| Orforglipron GLP-1 | approved | phase 3 + | phase 3 + | phase 2 | — | — | phase 3 | phase 3 | — |
| CagriSema GLP-1 (± GIP) + amylin | phase 3 + | phase 3 + | phase 3 | phase 2 | — | — | — | — | — |
| Retatrutide GLP-1/GIP/glucagon | phase 3 + | phase 3 + | phase 3 | phase 3 | — | phase 3 | phase 3 | phase 3 + | — |
| Survodutide GLP-1/glucagon | phase 3 + | phase 3 + | phase 3 | — | — | phase 3 | — | — | — |
| Mazdutide GLP-1/glucagon | approved | approved | phase 3 | phase 3 | phase 2 | phase 3 | phase 3 | — | phase 3 |
| MariTide GLP-1 + anti-GIPR | phase 3 | phase 3 | phase 3 | — | phase 3 | — | phase 3 | — | — |
| Amycretin GLP-1 (± GIP) + amylin | phase 3 | phase 3 | phase 3 | — | — | — | — | — | — |
| VK2735 GLP-1/GIP | phase 3 | phase 3 | — | — | — | — | — | — | — |
| Eloralintide amylin | phase 3 | phase 3 | — | — | — | — | phase 3 | phase 3 | — |
| Petrelintide amylin | phase 3 | phase 3 | phase 3 | — | — | — | — | — | — |
| Ribupatide (HRS9531) GLP-1/GIP | phase 3 + | phase 3 + | phase 3 | — | phase 2 | phase 2 | phase 3 | phase 3 | — |
| Ecnoglutide GLP-1 | approved | approved | — | — | — | — | phase 3 | phase 3 | — |
| Berobenatide (MET-097i) GLP-1 | phase 3 | — | — | — | — | — | phase 3 | — | — |
| Bofanglutide (GZR18) GLP-1 | phase 3 + | phase 3 | — | — | — | — | phase 3 | — | — |
| Pemvidutide GLP-1/glucagon | phase 2 | — | — | — | — | phase 3 | — | — | — |
| Efinopegdutide GLP-1/glucagon | — | — | — | — | — | phase 2 | — | — | — |
| UBT251 GLP-1/GIP/glucagon | phase 3 | phase 3 | — | phase 2 | — | phase 2 | phase 3 | — | — |
| HRS-7535 GLP-1 | phase 3 + | phase 3 + | — | phase 3 | phase 2 | — | — | — | — |
| BGM0504 GLP-1/GIP | phase 3 + | phase 3 | — | — | — | — | — | — | — |
| HDM1702 GLP-1 | phase 3 + | — | — | — | — | — | — | — | — |
| HDM1005 GLP-1/GIP | phase 3 | phase 3 | — | — | — | — | phase 3 | — | — |
| RAY1225 GLP-1/GIP | phase 3 + | phase 3 | — | — | — | — | — | — | — |
| TG103 GLP-1 | phase 3 | phase 3 + | — | — | — | — | — | — | — |
| HS-20094 GLP-1/GIP | phase 3 + | phase 3 | — | — | — | — | — | — | — |
| ASC30 GLP-1 | phase 3 | phase 3 | — | — | — | — | — | — | — |
| HDM1002 GLP-1 | phase 3 + | phase 3 | — | — | — | — | — | — | — |
| AZD5004 (ECC5004) GLP-1 | phase 3 | phase 3 | — | — | — | — | — | — | — |
| DR10624 GLP-1 + FGF21 (dual/triple) | — | — | — | phase 2 | — | phase 2 | — | — | — |
| Efsubaglutide alfa (supaglutide) GLP-1 | phase 3 | approved | — | — | — | — | — | — | — |
| PB-119 GLP-1 | phase 2 | approved | — | — | — | — | — | — | — |
| PEG-loxenatide GLP-1 | phase 3 | approved | — | — | — | — | — | — | — |
| Beinaglutide GLP-1 | approved | approved | — | — | — | — | — | — | — |
| HEC88473 GLP-1 + FGF21 (dual/triple) | — | phase 2 | — | — | — | — | — | — | — |
| Zovaglutide GLP-1 | phase 3 | — | — | — | — | — | — | — | — |
| Noiiglutide GLP-1 | phase 2 | phase 3 | — | — | — | — | — | — | — |
| Lenbelintide amylin | phase 3 | — | — | — | — | — | — | — | — |
| Efocipegtrutide GLP-1/GIP/glucagon | — | — | — | — | — | phase 2 | — | — | — |
| Macupatide GIP | phase 2 | — | — | — | — | — | — | — | — |
| Cotadutide GLP-1/glucagon | — | — | — | phase 2 | — | phase 2 | — | — | — |
approved / in label · positive phase 3 · phase 3 ongoing or results pending · phase 2 / signal only · no data · negative. Reviewed 13 Oct 2026 against FDA labels, the EMA register, NMPA announcements, ClinicalTrials.gov and phase 3 publications or press releases. Tap a cell for the evidence.
Every phase 2–3 trial of the atlas molecules in outcome and specialty indications, from first patient to primary completion, as registered on ClinicalTrials.gov and refreshed weekly. Weight and glucose trials can be switched on. Markers show the result once it is public.
Registry refresh 14 Oct 2026Bars run from study start to primary completion (ClinicalTrials.gov). Tap a row for the result and sources. Refreshed weekly from the registry; last refresh 14 Oct 2026.
Molecules of the class from approved to phase 2 with the next milestone. Tap a mechanism to filter; the table also filters by route and company.
Data cut 14 Oct 2026Highest development stage reached in any indication. Tap a molecule for status, next milestone and sources. Badges: oral formulation; developed in China; best weight change reported in this atlas (obesity, any estimand).
Statuses are tagged «verify» until reconciled with regulatory sources.
From the 1,309 congress abstracts (Milan, 29 Sep – 3 Oct 2026) we selected 42 results in 5 blocks; the abstract number links to our archive. Next layers: ObesityWeek (November) and ADA 2027.
Search the text; filter by block and molecule.
BlockWhat we are still missing and how you can help. Below the list of open requests: how complete each field is, and the gaps arm by arm.
Data cut 14 Oct 2026Specific gaps we know about. If you have the paper, poster, slides or a link — send it; subscribers can attach a PDF (seen by the editors only). Accepted contributions are marked «used in the atlas».
Filter by molecule or availability; sort by the number of missing fields.
Vancouver style; each entry links to the primary source — the article on PubMed, the registry record on ClinicalTrials.gov, or the congress abstract. Numbers match the [n] markers in the tables and tooltips.
The atlas is a database of records, not a text. Each record: trial, arm, population, duration, estimand, value, primary source, verification status. The whole page is generated from the database.