Home › The Protocol Vault › Survodutide

Survodutide

BI 456906

Metabolic & Fat LossInjectable📊 Correlative data

Survodutide (BI 456906) is a metabolic & fat loss research compound. Dual GLP-1/glucagon receptor agonist engineered for appetite suppression plus increased energy expenditure and hepatic benefit.

Research & educational use only. The information below summarizes published research and mechanisms. It is not medical advice or a recommendation for human use. The protocol that uses it — dosing, sequence and what to retest — is inside Skool ($10/mo).

Survodutide quick facts

Reported research dose0.3mg-4.8mg
RouteSubq
Frequency1-2x Weekly (Split Dose)
Half-life~Weekly dosing
FormsInjectable
Evidence levelPhase 2/3 human
Coach Cam’s take

Strong MASH/liver data emerging. Same GLP-1+glucagon logic as Mazdutide.

How Survodutide works

Dual GLP-1/glucagon receptor agonist engineered for appetite suppression plus increased energy expenditure and hepatic benefit.

Proposed benefits

Researched for fat oxidation, appetite and energy regulation, insulin sensitivity and endurance capacity.

Where to get Survodutide

Buy Survodutide at Flawless Compounds →
Use code CAMERON at checkout

Bacteriostatic water is the diluent — sterile water with 0.9% benzyl alcohol, which is what lets a vial be drawn from more than once. It does not come with the vial, and unlike the compound it is bought again every time.

Need bacteriostatic water? Get it at AminoWell USA (my company) → Code CAMERON.

The evidence for Survodutide

Graded by what exists behind each claim.

✅ Clinically validated

📊 Correlative data

🧪 Theoretical / extrapolated

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What Survodutide actually does

Survodutide is a glucagon receptor and GLP-1 receptor dual agonist, and unlike the rest of this class it has a paper describing how the ratio was chosen. Nineteen dual GCGR/GLP-1R agonists were screened on cyclic AMP potency in CHO-K1 cells stably expressing the human GCGR and the human GLP-1R, then re-tested on endogenously expressed receptors — MIN6 insulinoma cells for the mouse GLP-1R, rat primary hepatocytes for the GCGR — and BI 456906 was picked for what the authors call balanced dual pharmacology Thomas 2024. That is one step further than any other multi-agonist in this catalog gets, and it still stops short: the paper prints no numeric potency ratio. ‘Balanced’ is a word.

What was measured instead, and it is unusually concrete. The two arms were each given their own in vivo readout. GLP-1R engagement was scored as improvement in an oral glucose tolerance test at 30 nmol/kg — area under the curve 54% versus vehicle for survodutide against 45% for semaglutide, so the GLP-1 arm is doing roughly what a selective GLP-1 does. GCGR engagement was scored at 100 nmol/kg as hepatic nicotinamide N-methyltransferase mRNA, up 15- to 17-fold, and plasma FGF21, up as much as sevenfold Thomas 2024. Those two are liver-specific glucagon signatures, and they are the closest thing to a measured arm-by-arm contribution published for any drug in this class.

An aside that nobody in this market has noticed. The biomarker used to prove the glucagon arm is working is induction of NNMT in the liver Thomas 2024. NNMT is the enzyme that a different compound in this same Vault, 5-Amino-1MQ, is sold to inhibit. Nobody has asked what happens when a person runs both, and the question is not idle: if hepatic NNMT induction is downstream of the receptor that produces survodutide’s extra weight loss, an NNMT inhibitor taken alongside it is being asked to block the marker of the very arm the user is paying for.

Where in the body each receptor is, which is the mechanism question that actually got answered. Single-cell work found the GCGR barely detectable in the area postrema and the arcuate nucleus of the hypothalamus, while the GLP-1R is expressed in both. Fluorophore-labeled survodutide reached the circumventricular organs and the adjacent hindbrain and hypothalamic nuclei without uniformly crossing the blood-brain barrier, and c-Fos labeling showed activation of multiple food-intake nuclei. The decisive control: a long-acting GCGR agonist on its own did not activate satiety regions and did not reduce food intake — but still reduced body weight Zimmermann 2026.

Read that last sentence slowly, because it is the whole drug. Appetite suppression is the GLP-1 arm, working in the brain. The extra weight is the glucagon arm, working somewhere else. That is a clean anatomical dissociation, published, and it explains why a dual can pass the ceiling a pure GLP-1 hits without simply making people hungrier-proof.

The half-life engineering. A C18 fatty diacid is attached to the peptide as the half-life-extending principle, which supports once-weekly human dosing Zimmermann 2022. Note the number: C18. The albumin-binding side chains in this class are not interchangeable, and chain length is one of the few structural parameters that is actually published.

Cell, rodent, human — and where it stops

Step one, in cells. cAMP potency at the human GCGR and human GLP-1R in CHO-K1 cells, confirmed on endogenous receptors in MIN6 cells and rat primary hepatocytes Thomas 2024.

Step two, in mice, two strains and two questions. Diet-induced obese mice, 30 nmol/kg once daily, body weight down 25% from baseline; diabetic db/db mice at 10 and 20 nmol/kg once daily, HbA1c down 0.4 to 0.6 percentage points Thomas 2024. In the discovery paper, pharmacological doses produced greater weight loss than maximally effective doses of semaglutide, and the mechanism was split between increased energy expenditure and reduced food intake Zimmermann 2022.

Step three, phase 1, and the tolerability signal is in the first human study rather than the last. A single-rising-dose study in 24 men with a BMI of 20 to under 30 (NCT03175211) and a multiple-rising-dose study in 125 adults with a BMI of 27 to 40 (NCT03591718), escalated over 6 weeks (Part A) or 16 weeks (Part B). Placebo-corrected weight loss reached −5.79% at week 6 and −13.8% at week 16. Plasma amino acids and glucagon both fell, which is target engagement at the GCGR measured in people. 10 of 80 subjects (12.5%) discontinued in Part A, most commonly for a cardiac or vascular adverse event (6 subjects, 7.5%); 8 of 45 (17.8%) discontinued in Part B, mainly gastrointestinal Jungnik 2023.

Step four, phase 2 in MASH, and this one met its primary endpoint. 293 participants with biopsy-confirmed MASH and fibrosis stage F1 to F3, randomized 1:1:1:1 to survodutide 2.4, 4.8 or 6.0 mg weekly or placebo, 48 weeks, structured as a 24-week rapid escalation followed by 24 weeks of maintenance (NCT04771273). Primary endpoint: histologic improvement in MASH with no worsening of fibrosis — 47%, 62% and 43% against 14% on placebo, P < 0.001 with a quadratic dose-response as the best-fitting model. Liver fat down by at least 30% in 63%, 67% and 57% against 14%. Fibrosis improved by at least one stage in 34%, 36% and 34% against 22% Sanyal 2024.

Those last two rows are not the same result and the difference is the point. Steatohepatitis improved against placebo by 33 to 48 percentage points. Fibrosis improved by 12 to 14 points against a placebo arm that was itself at 22%. Inflammation is the endpoint this drug moved; scar is the endpoint that decides outcomes. And the dose-response is non-monotonic — 4.8 mg beat 6.0 mg on every one of the three histology rows — which the trialists handled honestly by fitting a quadratic rather than pretending it was linear.

Step five, phase 3 SYNCHRONIZE-1, which also met its primary endpoints. 725 participants (241 at 3.6 mg, 242 at 6.0 mg, 242 on placebo), mean BMI 37.9, mean body weight 108.8 kg, 76 weeks (NCT06066515). Co-primary: percent body weight change and the proportion losing at least 5%. Result: −12.2% (95% CI −13.6 to −10.8) at 3.6 mg and −13.0% (−14.4 to −11.6) at 6.0 mg against −5.4% (−6.9 to −4.0) on placebo, with 72.6%, 71.9% and 46.3% losing at least 5%, P < 0.001 le Roux 2026.

Now do the subtraction nobody prints. The phase 1 headline is −13.8% placebo-corrected at 16 weeks Jungnik 2023. The phase 3 result is −13.0% absolute at 76 weeks against a placebo arm that lost 5.4%, so the placebo-corrected effect is about 7.6 percentage points le Roux 2026. The number that circulates for this compound comes from a 45-person dosing cohort at week 16; the number from 725 people at week 76 is roughly half of it. Nothing was hidden — the two figures are simply not the same quantity, and they are quoted as if they were.

The obstacles, one at a time. (1) No published potency ratio, in any of the six papers, so ‘balanced’ cannot be checked Thomas 2024. (2) No human head-to-head against semaglutide or tirzepatide — the only comparison ever run was in mice Zimmermann 2022. (3) The cardiac and vascular discontinuations in phase 1 Part A have never been explained in the published record Jungnik 2023. (4) No body composition data, so the energy-expenditure mechanism has never been checked against a scan in a person. (5) The circumventricular-organ dissociation was shown in mice with a fluorophore-labeled compound Zimmermann 2026, not in humans.

Survodutide pharmacokinetics — how much of it actually gets in

The card says ‘~Weekly dosing’, which is a schedule. No human half-life for survodutide appears in the abstracts of any of the six papers behind this page. What follows is what is actually published and what it bounds.

What holds it in the body. A C18 fatty diacid, attached as the half-life-extending principle explicitly to support once-weekly dosing in humans Zimmermann 2022. The mechanism is reversible binding to circulating albumin: bound drug is too large for glomerular filtration and shielded from peptidases, and the small free fraction is continuously replenished. That is the same trick as every weekly incretin, and the chain length is the part that differs between them.

What degrades it. Proteolysis of the free fraction by plasma and tissue peptidases, plus renal handling of whatever is unbound. It is a peptide, not a cytochrome substrate, so the interaction that matters is not metabolic — it is delayed gastric emptying changing the absorption rate of oral drugs taken with it.

The number that bounds the half-life, and it is a trial design rather than a measurement. The MASH trial escalated over 24 weeks before a 24-week maintenance phase Sanyal 2024, and phase 1 escalated over 6 or 16 weeks Jungnik 2023. A 24-week escalation is extreme even by the standards of this class, and it says two things: steady state on a weekly peptide arrives in about 4 to 5 half-lives, so a weekly schedule implies a half-life on the order of 5 to 7 days; and tolerability, not pharmacokinetics, is what set the pace. Two practical consequences: the first three weeks at any new dose under-represent that dose, and stopping leaves measurable exposure for three to four weeks.

The oral barrier. Absolute. An acylated peptide swallowed meets gastric acid, pancreatic proteases and brush-border peptidases, with hepatic first-pass extraction behind them. Every dose in every survodutide study was injected.

The injectable route, and a card discrepancy worth naming. Every published dose was subcutaneous, once weekly. This site’s card lists 0.3 mg to 4.8 mg. The registered phase 3 doses are 3.6 mg and 6.0 mg le Roux 2026 and the phase 2 MASH doses were 2.4, 4.8 and 6.0 mg Sanyal 2024, so the card’s ceiling sits below the dose that produced the pivotal result and its floor sits below any dose ever studied as a maintenance dose.

What would have to be true, and how you would know it was not

Four predictions. The second one says the compound will fail to move the marker that matters most in liver disease, and it is the one worth running.

1. GGT and the aminotransferases should fall, and fast. Liver fat fell by at least 30% in 57 to 67% of the phase 2 participants Sanyal 2024, and hepatic steatosis is the commonest reason a healthy person’s GGT and ALT sit high. Draw a CMP and GGT at baseline and at 12 weeks. This is the fastest visible signal this compound produces and it usually moves before the scale does.

2. The ELF score will not move by six months, and this cuts against the product. The ELF score is a serum fibrosis panel, and fibrosis is precisely the endpoint that separated least in the phase 2: at least one stage of improvement in 34 to 36% on drug against 22% on placebo over 48 weeks Sanyal 2024. A 12-to-14-point separation over a year does not become a detectable ELF change in one person in six months. Predict it flat, retest at 12 months, and treat any 6-month movement as assay noise or as weight loss rather than as reversal of scar.

3. HbA1c should fall while fasting glucose is watched for the opposite. Draw HbA1c and fasting insulin at baseline, 12 and 24 weeks. Glucagon receptor agonism raises hepatic glucose output; GLP-1 receptor agonism lowers it. The published amino-acid and glucagon suppression Jungnik 2023 says the GCGR arm is genuinely engaged, so the informative pattern is weight falling while fasting glucose creeps up — the specific signature of a glucagon arm outrunning its counterweight, and the reason this class is studied carefully in people already on insulin.

4. The mechanistic read-out nobody orders. Survodutide reduced plasma amino acids in humans, and that was reported as GCGR target engagement rather than as a side note Jungnik 2023. A plasma amino acid panel is orderable, is not on any standard wellness profile, and is the only marker on this page that reports on the glucagon arm specifically. If it does not fall, the vial is not doing the thing that distinguishes this compound from semaglutide.

What nobody has tested yet

Five experiments. The second one has never been considered by anyone and is available to this site’s own readership.

1. The potency ratio is still unpublished. Nineteen dual agonists were screened on cAMP potency at both receptors and the selection logic was published — without the numbers Thomas 2024. One table would let survodutide, mazdutide, pemvidutide and retatrutide be ranked on the single axis that separates them, and it exists in a file at the sponsor.

2. Nobody has asked what an NNMT inhibitor does to this drug. Hepatic NNMT mRNA induction, 15- to 17-fold, is the assay used to prove the glucagon arm is engaged Thomas 2024. 5-Amino-1MQ — sold in this same catalog for fat loss — inhibits that enzyme. Whether the two oppose each other, do nothing to each other, or combine is unknown, unstudied, and directly relevant to people who are already stacking them. A rodent study with four arms and a hepatic mRNA readout would answer it in a month.

3. Nobody has explained the non-monotonic dose-response. In the MASH trial, 4.8 mg beat 6.0 mg on MASH resolution, on liver fat and on fibrosis Sanyal 2024. Either 6.0 mg is genuinely worse for the liver — which a glucagon arm large enough to drive hepatic substrate flux could plausibly cause — or it is dropout at the high dose. The trial data separate those two and the analysis has not been published.

4. Nobody has published body composition. The entire argument for a glucagon arm is that it raises energy expenditure rather than only suppressing intake Zimmermann 2022. If that is true, the fat-to-lean ratio of the weight lost should be better than a pure GLP-1’s. No DXA has been reported at any dose in any survodutide trial.

5. Nobody has looked at the phase 1 cardiac and vascular discontinuations again. Six of eighty subjects left Part A for a cardiac or vascular adverse event Jungnik 2023 and no later publication in this set returns to it. Six events in eighty people is either an early-escalation artefact or the most important safety sentence this compound has, and only the sponsor’s dataset can say which.

Survodutide — its own safety story, not its class's

The class block on this page is written for GLP-1 receptor agonists. Five things below belong to this molecule specifically.

1. The escalation is the safety story. The phase 2 spent 24 of its 48 weeks escalating Sanyal 2024 and phase 1 escalated over 6 or 16 weeks Jungnik 2023. A drug that needs half a year to reach its dose in a monitored trial is not a drug to titrate quickly outside one, and essentially every reported intolerance in this program happened during escalation rather than at steady state.

2. The gastrointestinal numbers, at MASH doses. Nausea 66% against 23% on placebo, diarrhea 49% against 23%, vomiting 41% against 4%; serious adverse events 8% against 7% Sanyal 2024. Two in five participants vomited. That is the cost side of the dual-agonist trade and it is larger than the single-receptor drugs in this catalog.

3. The cardiac and vascular discontinuations, which are specific to this compound’s record. In phase 1 Part A, 6 of 80 subjects (7.5%) discontinued for a cardiac or vascular adverse event Jungnik 2023. Glucagon receptor agonism raises heart rate and cardiac output; that is expected pharmacology rather than a surprise. What is missing is any published size for the heart-rate change in this program, so a person taking it has no reference range to compare their own pulse against.

4. The glucose direction is not the class’s. A pure GLP-1 lowers hepatic glucose output. This one contains an arm that raises it, and human target engagement at that arm has been demonstrated by falling plasma amino acids and glucagon Jungnik 2023. In anyone on insulin or a sulfonylurea — populations these trials excluded — it is the background agent that needs attention, and the risk window is escalation.

5. The honest limit of the record. Roughly 1,150 people across the phase 1, phase 2 and phase 3 reports here, the longest exposure 76 weeks le Roux 2026, and the pivotal population had a mean BMI of 37.9. There is no published cardiovascular outcome result, no body composition, and no data in people with fibrosis stage F4. Approval status is not the same thing as outcome data, and this compound has neither yet.

Sources read for this page

Survodutide — safety, predicted from mechanism

Predicted from mechanism, not from a human safety trial. How that reasoning works →

What the mechanism predicts

Derived from the molecule, not a trial.

What has actually been reported

How to reduce the risk

Same mechanism as the prediction.

What it does to your bloodwork

A fact about the assay.

Don't run this if

The honest unknown

Not medical advice. If you take prescription medication or have a diagnosed condition, check this with a pharmacist or doctor.

Survodutide — interference & stacking

Predicted from mechanism, not from an interaction study. How mechanism-predicted claims are made →

What Survodutide moves on your bloodwork

Expected direction, not a measured one.

The pattern with this class is that the frightening-looking numbers (lipase) are usually fine and the boring ones (ferritin, B12) are where the real damage happens.

🔒
The dose is the easy part. Making Survodutide actually work is what's behind Skool:
Running it
  • How to work up to it, and when not to
  • When to take it, and why that window
  • Cycle length
  • Time off between cycles
  • Fasted or fed, and when in the day
  • Needle gauge and injection site
  • Coach Cam's personal notes
Stacking it
  • Which compounds push the same lever, and why the dose adds up faster than people count
  • What blunts it — the stacks that waste your money
  • What compounds the risk, so a side effect arrives sooner than any one of them suggests
  • Coach Cam's read on running it alongside the rest of your protocol

Everything above is free and stays free. Skool is where it becomes a plan — Survodutide in an order, with the rest of what you're running.

Unlock in Skool — $10/mo →

Bloodwork to run alongside Survodutide

Baseline first, then again at 8–12 weeks.

MarkerWhat it’s watching for
HbA1c (Hemoglobin A1c)Baseline
Fasting InsulinThe earliest marker of metabolic change
LipasePancreatitis monitoring
Comprehensive Metabolic Panel (CMP)Liver especially — glucagon agonism acts on the liver directly

The “On a GLP-1 (Semaglutide / Tirzepatide)” panel covers these in one order — 11 markers, $148.05 with the discount applied.

Check results you already have → · All 103 markers A–Z

Survodutide — frequently asked questions

What is Survodutide?

Survodutide (BI 456906) is a metabolic & fat loss research compound. Dual GLP-1/glucagon receptor agonist engineered for appetite suppression plus increased energy expenditure and hepatic benefit.

Is the full Survodutide protocol on this page?

The reported research dose is on this page, along with how Survodutide works and the evidence behind it. The protocol — how to work up to it, frequency, cycle length, time off, what not to stack it with and Coach Cam's notes — is inside Skool.

What is the half-life of Survodutide?

Survodutide has an approximate half-life of ~Weekly dosing, which is part of what determines how often it's dosed.

What's the evidence behind Survodutide?

Current evidence level: Phase 2/3 human. Survodutide is offered for research purposes only and is not an approved medicine.

What Survodutide is used for

Survodutide appears under 2 goals in the goal router.

🔥 Lose fatAppetite & satiety signaling📉 Metabolic health & insulin sensitivityIncretin & satiety signaling📉 Metabolic health & insulin sensitivityHepatic fat & fatty liver

Where this goes next

Go deeper$10/mo

The pages here are the frameworks. The protocols — the dosing, the order to correct things in, the week-by-week schedule and what to retest — are inside Skool.

← Explore the full Protocol Vault

↑ Back to on this page