Efmedaglutide Alfa (TG103)_Dosage, benefits & legal status
Also known as TG103, TJ103, SYSA1803, efmedaglutide
Two regulatory filings, and nothing published to read.
Common vial size is what vendors typically sell, not a recommendation. Enter your own vial in the calculator.
What is Efmedaglutide Alfa (TG103)?
Plain language first. The technical label stays, but it never stands alone.Overview
This is a weekly GLP-1 injection in late-stage Chinese development, and structurally it is the same kind of molecule as dulaglutide: a GLP-1 welded to an antibody fragment rather than a small fatty peptide. The reason it earns a page is the gap between its paperwork and its evidence. Two applications have been filed with China's regulator and four mid and late-stage trials have finished, including a 675-patient phase 3 in obesity. Not one of those trials has published a result or posted one to a registry. So every impressive figure you will read about it, including the claim that it outperforms dulaglutide, traces back to a company announcement rather than to data anyone outside the company has seen.
Efmedaglutide alfa is a long-acting GLP-1 receptor agonist fusion protein. The FDA substance registry records it as a fusion protein of human sequence origin, two identical subunits of 286 residues, with two N-linked and ten O-linked mammalian glycosylation sites, and its systematic name identifies it as a dimeric GLP-1(7-37) variant fused to a hybrid human IgD and IgG4 Fc, expressed in CHO DG44 cells. The IgD element is corroborated by the peer-reviewed phase 1b report. Two marketing applications have been accepted in China, for obesity in October 2025 and for type 2 diabetes in June 2026. Acceptance is not approval, and the compound is not approved anywhere.
What it's used for
Investigational for obesity and for type 2 diabetes. Marketing applications have been accepted in China for both, but it is not approved in any country.
How does Efmedaglutide Alfa (TG103) work?
What it does in the body, and where.The pharmacology is the ordinary GLP-1 story. What is distinctive is the fusion partner, and what is missing is any published evidence of what the molecule does in the trials that matter.
- 01An unusual Fc choiceMost GLP-1 fusion drugs hang the peptide off an IgG Fc or off albumin. The substance registry's systematic name for this one identifies a hybrid Fc built from human IgD and IgG4, and the peer-reviewed phase 1b report independently describes the linkage as via IgD. That is an uncommon design, and the registry records two identical 286-residue chains with two N-linked and ten O-linked glycosylation sites, the heavy O-linked count being characteristic of an IgD hinge.
- 02The GLP-1 end is conventionalThe active portion is a variant of GLP-1(7-37), the natural active fragment of the hormone. Once it reaches the receptor the effects are the class standard: glucose-dependent insulin release, glucagon suppression, slowed stomach emptying and reduced appetite. Nothing in the design suggests a novel mechanism, so the interesting questions about this compound are about evidence rather than biology.
- 03What the one published study showedA 48-patient phase 1b reported week 12 weight loss of 7.24% at 15 mg weekly against 1.76% on placebo. Its own authors describe the finding as preliminary. For a phase 1b that is a reasonable signal, and it is the entire published efficacy record for this molecule.
- 04The gap that defines this compoundFour completed phase 2 and phase 3 trials have produced no publication and no posted registry results, including the 675-patient phase 3 in obesity that supports the October 2025 filing. A head-to-head against dulaglutide is registered with 632 patients, but it is open-label and has released nothing. Every circulating claim, from more than one percentage point of HbA1c by week 8 to over two percentage points, to outperforming dulaglutide, exists only in press releases. A filing accepted on unpublished data is normal regulatory practice; quoting those numbers as though they were published results is not.
A schematic of the compound's mechanism, not a diagnostic image.
What else is it used for?
Each group carries its own evidence level.Approved means regulators have cleared it for that purpose. Investigational means it is in trials. Early research means it is being studied and is not established.
The evidence gap is the story
Investigational- No dataFour completed phase 2 and phase 3 trials with zero publications and zero posted registry results, including a 675-patient phase 3 in obesity.
- human-earlyThe only published efficacy data is a 48-patient phase 1b: week 12 weight loss 7.24% at 15 mg against 1.76% placebo, called preliminary by its authors.
- Human trialThe registered head-to-head against dulaglutide has 632 patients, is open-label, and has released no data.
Who actually develops it
none- RegulatoryCSPC Baike (Shandong) sponsors every registered trial and filed both marketing applications, despite the compound being widely attributed to TJ Biopharma or I-Mab.
- RegulatoryIt originated at I-Mab, was licensed to CSPC in 2018, and TJ Biopharma became licensor only after an April 2024 restructuring.
- No dataThree codes circulate for one molecule: TG103, TJ103 and SYSA1803. CSPC's filing uses three product names and never mentions TG103.
Filed, not approved
Investigational- RegulatoryBoth Chinese events are accepted applications, not approvals: obesity in October 2025 and type 2 diabetes in June 2026.
- No dataThe June 22, 2026 date in circulation is TJ Biopharma's press release date; CSPC announced the same acceptance on June 18. Neither gives a day-level regulatory date.
- RegulatoryAbsent from the 2025 Chinese reimbursement list, consistent with not being approved. No US label, verified against a dulaglutide control that returned seven records.
Identity and the look-alike name
none- lab-analysisTwo identical 286-residue chains, a GLP-1(7-37) variant on a hybrid human IgD and IgG4 Fc, made in CHO DG44 cells. Ten O-linked glycosylation sites fit an IgD hinge.
- No dataIt is not efsubaglutide alfa, a different molecule from a different developer that is already approved and reimbursed in China.
- No dataListings exist only as protein research reagents at microgram to milligram analytical scale, priced as laboratory material and in one case sold out. Not injectable product.
The first error is calling it approved. Both Chinese regulatory events are accepted applications, and the compound is approved in no country. The second is the date: June 22, 2026 is TJ Biopharma's press release date for the diabetes filing, while CSPC announced the same acceptance on June 18, and neither gives a day-level regulatory date, so June 2026 is as precise as the record allows. The third is attribution. The developer is effectively CSPC Baike (Shandong), the sponsor of every registered trial and the filer of both applications, not TJ Biopharma or I-Mab as usually stated, and the same molecule travels under three codes, TG103, TJ103 and SYSA1803. The fourth, and the most consequential for a reader, is quoting its efficacy. Nothing beyond a 48-patient phase 1b has been published or posted, so the HbA1c figures above one and above two percentage points and the claim of outperforming dulaglutide are press-release material; the dulaglutide comparison does have a registered 632-patient trial behind it, but that trial is open-label and has released nothing. Finally, it is routinely confused with efsubaglutide alfa, a different fusion protein from a different company that is already approved and reimbursed in China.
The published record
Papers about Efmedaglutide Alfa (TG103)
Searched: efmedaglutide, efmedaglutide alfa, TG103, TJ103, SYSA1803.
Left out: efsubaglutide (a different GLP-1-Fc fusion from a different developer that IS approved and reimbursed in China, with its own slug. The two are routinely swapped); efpeglenatide (a different molecule again, three-chain and aglycosylated, with its own slug); dulaglutide (the comparator in the registered head-to-head and an approved drug with a very large literature of its own); GLP-1 (the receptor class term, tens of thousands of records); CSPC (a company name spanning a large portfolio); I-Mab (the originating company, whose other programs are unrelated).
14 papers tagged human · 2 not reviews or lab · 2 clinical trial publications
11 registered trials · none with posted results
Furthest stage reached: Phase 3, completed, no results posted
Counted from Europe PMC indexing, not read by a person; a human tag can land on reviews, commentary or lab papers; clinical trial publications is the narrower count. A preprint and its journal version count once.
At a glance
Four things to know about Efmedaglutide Alfa (TG103)
Evidence
Investigational
Category
Weight loss and metabolic
Common vial
Varies
Half-life
Long enough for weekly dosing; no published figure located
How much, and do you cycle it?
The range, and whether you take breaks.15 mg weekly was the top dose in the published phase 1b, once weekly. Continuous weekly dosing in trials Trial dose
Trialed as a chronic weekly treatment for obesity and diabetes, titrated rather than cycled.
Long enough for weekly dosing; no published figure locatedNo figure
Dosed once weekly in all trials, so the half-life must be of the order of days, but no numerical value appears in a source we could reach. The published phase 1b reports pharmacokinetics in its title, yet no value from it could be extracted from the accessible record.
You can compare this against the whole library to see where it sits.
How is it given?
Dosing context and source notes.Investigational and not sold, so no vial draw can be shown and the calculator does not apply. The highest dose in the only published study was 15 mg once weekly in a 48-patient phase 1b. Doses used in the four completed phase 2 and phase 3 trials have not been published.
How do you store it?
Before mixing, and after.Nothing to store: it is not sold anywhere.
Efmedaglutide Alfa (TG103) has never been marketed, so there is no vial, pen or bottle to keep at home. Supplies for a trial are made, stored and dispensed by the sponsor and the trial pharmacy under conditions set in the protocol. Anything offered for sale under this name did not come from that supply chain.
What are the side effects and cautions?
Known cautions and warnings for this compound.Who should avoid it
- Treating either Chinese filing as an approval: both are accepted applications and the compound is approved nowhere
- Quoting any efficacy figure beyond the 48-patient phase 1b, since nothing else has been published or posted
- Reading the claim that it beats dulaglutide as a trial finding, when the registered head-to-head is open-label and has released no data
- Confusing it with efsubaglutide alfa, a different molecule from a different developer that is already approved in China
Stop and get help
- Severe or persistent abdominal pain, a possible sign of pancreatitis in this drug class
- Signs of an allergic reaction, including rash, swelling or difficulty breathing
- Persistent vomiting with an inability to keep fluids down
Common effects
- Nothing established beyond a 48-patient phase 1b study.
Where this page says nothing is established, that means nobody has studied it, not that a compound is safe.
Is Efmedaglutide Alfa (TG103) approved?
And where the numbers on this page come from.No. Efmedaglutide Alfa (TG103) is not approved anywhere and is currently in company-run clinical trials. There is no legal route to obtain it outside a trial.
- Approval statusInvestigational
- RouteUnder the skin (subcutaneous injection, once weekly)
- Checked2 November 2026
- DoseFrom published clinical trials. Not approved by a regulator for this dose.
- Vial sizeFrom what vendors typically list; not a recommendation.
- Sources
What has happened recently
Two corrections matter here. The first is who develops this. It is widely attributed to TJ Biopharma or I-Mab, but CSPC Baike (Shandong) sponsors every registered trial and filed both marketing applications. The molecule originated at I-Mab, was licensed to CSPC in 2018, and TJ Biopharma (Hangzhou) became the licensor only after an April 2024 restructuring. Three codes circulate for the same compound: TG103, TJ103 and SYSA1803, the last being CSPC's internal code and visible in registry organization study identifiers. CSPC's own filing uses three different product names and never mentions TG103 at all, so the filings cannot be matched to the trial record without the TJ Biopharma announcement. The second correction is the date. A June 22, 2026 date circulates for the diabetes filing; that is TJ Biopharma's press release date, and CSPC announced the same acceptance on June 18, 2026. Neither company gives a day-level regulatory date, so June 2026 is as precise as the record supports. The obesity filing follows the same pattern, announced by CSPC on October 13, 2025 and by TJ Biopharma the following day. CSPC calls the submission a marketing authorization application while TJ Biopharma calls it a BLA: one filing, two labels. The compound is absent from the 2025 Chinese reimbursement list, consistent with it not being approved, and there is no US label: an openFDA query returned nothing while the same query returned seven records for dulaglutide as a control.
You can compare this against the whole library, or read how to check a seller.
What else is like Efmedaglutide Alfa (TG103)?
Others in GLP-1 agonists, side by side.If you're weighing Efmedaglutide Alfa (TG103), it's worth reading Efpeglenatide, Efsubaglutide Alfa, and Exenatide (Byetta / Bydureon) in the same class.
Two regulatory filings, and nothing published to read.
- Dose
- 15 mg weekly was the top dose in the published phase 1b
- How often
- Once weekly
A drug Sanofi handed back, revived as a cheaper weekly option in one country.
- Dose
- 2 to 10 mg
- How often
- Once weekly
A marketed GLP-1 almost nobody outside China has heard of.
- Dose
- 1 to 3 mg weekly for type 2 diabetes; 5 to 80 mg weekly in obesity trials
- How often
- Once weekly
The first GLP-1 drug, found in lizard venom, now mostly retired.
- Dose
- 2 mg per week (the weekly form)
- How often
- Weekly
Guides for this compound
- How to Reconstitute a Peptide
Mix peptide powder with bacteriostatic water, step by step.
- How to Read an Insulin Syringe
Read the units, convert to millilitres, and draw the exact dose.
- mcg vs mg: The Peptide Dosing Mistake to Avoid
Your vial says mg and your protocol says mcg. Here is the conversion and how to check it.
- How to Read a Peptide COA
What the purity number does and does not tell you, and how to spot an edited report.
- How to Vet a Peptide Vendor
Independent COAs, HPLC purity, and the red flags that mean walk away.
Sources
- BMC Medicine, safety, tolerability, pharmacokinetics and pharmacodynamics of TG103 injection(2024)Volume 22. The only peer-reviewed efficacy data for this compound: a 48-patient phase 1b reporting week 12 weight loss of 7.24% at 15 mg weekly against 1.76% on placebo, with the authors describing the result as preliminary. Note that the paper uses the development code TG103 throughout, while the Chinese marketing filings use the name efmedaglutide alfa and never mention that code.
- FDA substance registration record for efmedaglutide alfa(2026)UNII P7M6NNA87N. Substance class protein, protein type fusion protein, sequence origin human. Two identical subunits of 286 residues with two N-linked and ten O-linked mammalian glycosylation sites. The record's systematic name identifies a dimeric GLP-1(7-37) variant fused to a hybrid human IgD and IgG4 Fc, expressed in CHO DG44 cells. This is the source for the structural comparison with dulaglutide at two 275-residue chains.
- ClinicalTrials.gov record NCT06235086, the head-to-head against dulaglutide(2026)A registered head-to-head against dulaglutide in 632 patients. Two things matter about it: the study is open-label, limiting what a comparison can show for subjective and behavioral endpoints, and it has released no public data. The widely repeated claim that this compound outperforms dulaglutide therefore has a real registered trial behind it and no results from that trial.
Beyond these, 15 records are indexed. Browse the records