The milligram is not the instruction
Peptide pages talk in milligrams as if the unit carried its own meaning. It does not. A milligram is a mass. What it is — a labeled maintenance dose, a trial arm, a rodent experiment, or a number someone repeated — lives in the sentence around it. Drop the sentence and the milligram starts doing other jobs: a start dose pretends to be a target, a mouse paper pretends to be a protocol, a compounded vial pretends to be a pen.
This reference reports published regimens as research literacy, so you can check the evidence yourself. It does not pick a dose for a person. The four-kind split below is the whole method. Once you can name which kind you are looking at, most of the internet’s dosing arguments collapse into a category error.
Cite the source beside the number, and keep the compound, formulation, population, units, route, frequency, and duration that the source actually reported. Missing details are gaps. They are not blanks to fill.
Four kinds of milligram
Approved products have a label. Human studies have a protocol. Animal and cell work have an experimental dose. Forums have a convention. Those four objects share a unit and almost nothing else. A label is specific to one named product and indication. A protocol is specific to one trial. Neither licenses the other, and neither licenses a research-chemical vial.
| Kind | Answers | Blind to |
|---|---|---|
| Label Wegovy 2.4 mg weekly; Forteo 20 mcg daily; Vyleesi 1.75 mg as needed. | What this named product, indication, and jurisdiction tells a prescriber to give. | Any other product of the same molecule. Ozempic is not Wegovy. Egrifta WR is not Egrifta SV. |
| Trial protocol Cagrilintide 0.3–4.5 mg weekly, up to 6 weeks' escalation (Lau 2021). | What that study administered, to whom, for how long, including the ramp. | Approval, interchangeability with a label, and whether a grey-market vial matches the trial material. |
| Preclinical A milligram-per-kilogram rodent dose that does not convert by body weight to a human. | What a mouse, rat, or cell experiment used. | Human pharmacokinetics, human tolerability, and almost every community 'human equivalent.' |
| Community IGF-1 LR3 100–500 mcg, labeled anecdotal in this reference. | What a forum, vendor page, or knowledge-transfer document currently repeats. | Whether anyone measured it in people. Provenance is the grade, not confidence. |
The documented-range box used on insight pages is the planner’s number with the provenance badge attached, so the two surfaces cannot disagree. For clinical-tier molecules it is the published maintenance or target span, not the titration start and not a dose to pick. Semaglutide’s box is Wegovy’s 1.7–2.4 mg once weekly, not the 0.25 mg first month (Wegovy prescribing information).
Same molecule, three products
Semaglutide is the cleanest demonstration that the milligram is not portable. Injectable Wegovy for chronic weight management titrates from 0.25 mg once weekly and usually maintains at 2.4 mg, with 1.7 mg as an alternative. Injectable Ozempic for type 2 diabetes maintains at 0.5, 1, or 2 mg. Oral Rybelsus is a daily tablet at a different milligram entirely, because almost none of a swallowed peptide survives the gut — the oral GLP-1 story. One sequence. Three labeled objects.
STEP 1, the trial that turned 2.4 mg into a cultural event, judged once-weekly 2.4 mg over 68 weeks in adults with overweight or obesity without diabetes (Wilding et al., NEJM 2021). It did not judge the 0.25 mg start, and it did not judge a compounded multi-dose vial. Tirzepatide repeats the pattern at a different scale: SURMOUNT-1 assigned 5, 10, or 15 mg once weekly after a 20-week escalation (Jastreboff et al., NEJM 2022); the Zepbound label’s recommended maintenance dosages are those same three strengths, and 2.5 mg is the start, not the target (Zepbound prescribing information).
Quoting a titration start as if it were the dose the trial measured is how a 0.25 mg first month and a 2.4 mg result end up in the same sentence as if they were one regimen.
A protocol is a ramp, then a target
Investigational peptides do not have a label, so the honest object is the protocol. That still is not a start-at-target instruction. Cagrilintide’s 26-week phase 2 assigned 0.3, 0.6, 1.2, 2.4, or 4.5 mg once weekly, with up to six weeks of escalation inside the 26 (Lau et al., Lancet 2021). The 10.8% mean loss everyone quotes is the 4.5 mg arm at week 26, after the climb. The week is not flat is how to read the clock; this page is how to read the milligram.
Survodutide makes the ramp even larger. Phase 2 in obesity assigned 0.6, 2.4, 3.6, or 4.8 mg once weekly for 46 weeks, of which 20 were escalation (le Roux et al., Lancet Diabetes Endocrinol 2024). The primary, planned-treatment analysis at 4.8 mg was a mean 14.9% loss against 2.8% on placebo. The ~19% figure that travels with the molecule is the actual-treatment reading of the same trial, not a second study. Two estimands, one protocol — the same split two numbers, one trial walks for retatrutide. Start-at-4.8 is exactly the mistake the twenty-week ramp exists to prevent.
Formulation is part of the number
Even an approved peptide will not carry its milligram across vials. Tesamorelin cut visceral fat about 15% at 2 mg subcutaneously once daily in the trial that earned the label (Falutz et al., NEJM 2007). Egrifta WR, a later formulation, is labeled at 1.28 mg once daily (Egrifta WR prescribing information). Same active, different product, different milligram. Reconstitution is where that difference becomes a concentration, which is why there is no universal bacteriostatic-water recipe.
Other labeled peptides are honest in a different direction: the number is small, or it is not a weekly habit at all. Teriparatide is 20 mcg subcutaneously once daily on the Forteo label (Forteo prescribing information) — a brief pulse, which is the whole mechanism. Bremelanotide (Vyleesi) is 1.75 mg as needed, at least 45 minutes before anticipated activity (Vyleesi prescribing information), not a weekly schedule. Putting either number into a “×/wk” table without that clause is a mistranslation.
Community numbers, and the error that hospitalizes
IGF-1 LR3 is the type specimen for a community milligram. This reference documents 100–500 mcg per administration and grades it anecdotal, because there is no human trial and no approved dose. The closest labeled window onto IGF-1 receptor agonism is mecasermin (Increlex): start 0.04–0.08 mg/kg twice daily, maximum 0.12 mg/kg twice daily, for severe primary IGF-1 deficiency (Increlex prescribing information). That is a different analog, a weight-based pediatric endocrine regimen, and not a conversion factor for LR3.
The harm-reduction case for reporting numbers, rather than refusing them, is not theoretical. In July 2024 FDA described dosing errors with compounded injectable semaglutide in which patients and clinicians mixed milligrams, milliliters, and insulin-syringe units on products that do not share the approved pens’ concentrations. People administered five to twenty times the intended dose; some were hospitalized (FDA compounding dosing-error alert). A 5-unit instruction drawn as 50 units is a literacy failure at the syringe. Hiding the labeled milligram does not prevent it. Naming the milligram, the unit, and the product does.
Which product or material? Which population? Which units, route, and schedule? Is this a start, a target, or a one-off? If the source does not say, the number is not ready to use as evidence.
- Label. Named product, indication, jurisdiction. Wegovy 2.4 mg is not Ozempic 1 mg and not a compounded vial.
- Trial. Arm, estimand, ramp, and who was enrolled. The quote is usually the target after the climb.
- Preclinical. Species and mg/kg. Do not scale a mouse into a person by body weight and call it a protocol.
- Community. Say so. Grade it anecdotal. Do not launder it into a clinical range by omitting the badge.
Common questions
- What dose of semaglutide is used for weight loss?
- On the Wegovy label, titration starts at 0.25 mg once weekly for four weeks and the usual recommended maintenance dose is 2.4 mg once weekly, with 1.7 mg as an alternative. That is not the Ozempic diabetes label (0.5, 1, or 2 mg) and not the oral tablet. STEP 1 judged 2.4 mg, not the 0.25 mg start. The documented-range box on this site for semaglutide is Wegovy's 1.7–2.4 mg maintenance span.
- Is a trial protocol the same as a safe dose?
- No. A registered or published protocol is what that study administered to that population, on that schedule, with that stopping rule. It is evidence of what was studied, not a finding that the regimen is established, approved, or appropriate for a reader. Missing details — units, route, titration, who was excluded — are gaps, not invitations to fill them in.
- Why do documented ranges skip the starting dose?
- Because the number people quote from a trial is almost always the maintenance or assigned target, and the number that actually starts in a pen is almost always lower. Wegovy starts at 0.25 mg; Zepbound at 2.5 mg; survodutide's phase 2 spent 20 of 46 weeks escalating; cagrilintide's phase 2 spent up to six of 26. The box reports the published target span. The note carries the start.
- What went wrong in the compounded semaglutide dosing errors?
- FDA's July 2024 compounding alert describes patients and clinicians mixing milligrams, milliliters, and insulin-syringe 'units' on products that do not share the approved pens' fixed concentrations. People drew 50 units instead of 5, or a prescription of 25 units went out when 5 were intended — five- to twenty-fold overdoses, some hospitalized. That is a unit-conversion failure, not a pharmacology failure.
- Does IGF-1 LR3 have an approved human dose?
- No. There are no controlled human trials of LR3. The range this reference documents is community convention, labeled anecdotal. The closest labeled product is mecasermin (Increlex), recombinant IGF-1, starting 0.04–0.08 mg/kg twice daily for severe primary IGF-1 deficiency — a different analog, a different population, and not a conversion factor.
Keep going
- How to reconstitute peptides — the vial is not a recipe
- How to read a peptide COA — what the certificate can and can't prove
- Two numbers, one trial — why every obesity study reports both
- The week is not flat — reading cagrilintide's clock
- Cycle planner — the same ranges, with provenance
- Ask the research agent for a labeled or trial regimen
Educational reference on how published regimens are reported, summarized from public labels, trial publications, and regulatory alerts and simplified in places. Not individualized medical advice or a prescription. Where a dose or regimen appears it is reported from its source as research literacy — so you can read the evidence yourself — not an instruction to dose. Follow the approved product label and instructions from a licensed prescriber and dispensing pharmacist. Specific products are named to explain why milligrams do not transfer; verify any claim against the linked primary sources.