"Microdosing tirzepatide" is searched thousands of times a month, and the approved products were not built for it. The smallest amount on the label is 2.5 mg, and the label itself says that amount is only for starting treatment. This page reports what the label states, the one randomised trial arm found below 2.5 mg, what the published literature on microdosing consists of, and what the smaller amounts are as syringe units in a powder vial. It does not describe a practice to follow.
GLP1Ledger's microdosing page covers who promotes the practice and why; this page stays with the figures. The full label schedule is on the tirzepatide dosage page, and Medibact's Tirzepatide guide covers the compound itself.
What the label states
The ZEPBOUND label (Eli Lilly, effective 2026-08-28) sets a starting dosage of 2.5 mg subcutaneously once weekly for 4 weeks, then increases of 2.5 mg after at least 4 weeks on the current dose, to maintenance dosages of 5, 10 or 15 mg once weekly, with 15 mg as the maximum. One sentence in section 2.1 settles the question of where the approved range begins: "The 2.5 mg dosage is for treatment initiation and is not approved as a maintenance dosage." On the label, every amount below 5 mg is a step on the way somewhere else, and nothing below 2.5 mg exists.
The approved products are solutions. Single-dose pens and vials hold 2.5 mg in 0.5 mL (5 mg/mL); multi-dose vials and KwikPens hold four 2.5 mg doses of 0.6 mL at 4.17 mg/mL. Neither is designed to deliver a fraction of its dose.
The one randomised arm below 2.5 mg
Before the phase 3 programme, Eli Lilly's phase 2 trial (NCT03131687; Frias et al., Lancet 2018, PMID 30293770) randomised adults with type 2 diabetes to tirzepatide 1, 5, 10 or 15 mg once weekly, dulaglutide 1.5 mg or placebo, for 26 weeks. The 1 mg arm is the only randomised exposure below the label's starting dose found for this page. The results posted on ClinicalTrials.gov (read 2026-09-23):
| Arm | Weight change, kg | ≥5% weight loss | ≥10% weight loss |
|---|---|---|---|
| Placebo | -0.4 | 0% | 0% |
| Tirzepatide 1 mg | -0.9 | 13.5% | 5.8% |
| Tirzepatide 5 mg | -4.8 | 47.3% | 16.4% |
| Tirzepatide 10 mg | -8.7 | 70.6% | 39.2% |
| Tirzepatide 15 mg | -11.3 | 62.3% | 37.7% |
| Dulaglutide 1.5 mg | -2.7 | 22.2% | 9.3% |
The abstract adds that HbA1c fell by 1.06 percentage points on 1 mg against 0.06 on placebo, so the 1 mg arm did move blood sugar, and that gastrointestinal adverse events occurred in 23.1% of the 1 mg arm, against 9.8% on placebo and 66.0% on 15 mg. Three limits apply to reading anything into it for microdosing: the population had type 2 diabetes, the trial lasted 26 weeks, and 1 mg was a fixed arm rather than a reduced dose after a higher one.
What the published record on microdosing consists of
A PubMed search for microdosing AND (semaglutide OR tirzepatide OR GLP-1) on 2026-09-23 returned ten records. Six are relevant: two letters in Diabetes Care on smaller amounts from multi-dose semaglutide pens (PMIDs 39808463 and 40540673), and a letter and a perspective in Obesity on the same with multi-dose tirzepatide pens (PMIDs 42571951 and 42244165, the second by pharmacists at the University of North Carolina Medical Center), a nurse-practitioner report on the practice's drivers and dosing errors (PMID 42201545), and a narrative review in Cureus (PMID 42668762). The other four are unrelated, among them a mouse psilocybin study and a laboratory assay paper. None of the ten is a trial. The nurse-practitioner report describes the practice as "subtherapeutic" and names gastrointestinal tolerability and "vanity weight" as its drivers, with affordability behind the shift to unsupervised strategies.
The arithmetic for amounts below 2.5 mg
In a powder vial, the concentration is the milligrams in the vial divided by the water added. Units on a U-100 syringe = mg ÷ mg/mL × 100.
| Amount | 5 mg/mL | 10 mg/mL | 20 mg/mL |
|---|---|---|---|
| 0.5 mg | 10 units | 5 units | 2.5 units |
| 1 mg | 20 units | 10 units | 5 units |
| 1.25 mg | 25 units | 12.5 units | 6.25 units |
| 1.5 mg | 30 units | 15 units | 7.5 units |
| 2 mg | 40 units | 20 units | 10 units |
Where those concentrations come from: 5 mg/mL is a 10 mg vial with 2 mL; 10 mg/mL is a 10 mg vial with 1 mL or a 30 mg vial with 3 mL; 20 mg/mL is a 20 mg vial with 1 mL. At 20 mg/mL the smallest rows fall between the 2-unit marks of a 1 mL barrel, which is the arithmetic reason sub-milligram amounts are read at lower concentrations or on a 0.3 mL barrel. The mg to units page sets out the formula and the tenfold misreadings; the tirzepatide reconstitution page works each vial size, and the calculator preset for tirzepatide vials does it for any figure entered.
Why the half-life changes what a small weekly amount means
The ZEPBOUND label gives an elimination half-life of approximately 5 to 6 days. With once-weekly injection, about 38 to 45% of each injection is still present when the next one is given. Summed over repeated weeks, the peak level at steady state is about 1.6 times a single injection's (5-day half-life) to 1.8 times (6-day half-life), and steady state is approached after four to five half-lives, 20 to 30 days. That is arithmetic from the label figure, not a measured result for sub-2.5 mg amounts; no pharmacokinetic study of them was found. It does mean a small weekly amount is not a small level after a month.
What is not established
No trial has tested weekly amounts below 2.5 mg for weight management. No trial has tested taking reduced portions from multi-dose pens. The one sub-2.5 mg arm on record was in type 2 diabetes, fixed from the start, and lasted 26 weeks. The label describes no use below 2.5 mg. Any figure for a "microdose" in circulation is therefore practice, not evidence, and this page reports none as a plan. Decisions about amounts belong with a prescriber.
Sources and dates
Opened 2026-09-23: the ZEPBOUND label via openFDA (set id 487cd7e7-434c-4925-99fa-aa80b1cc776b, effective 2026-08-28), sections 2, 3 and 12.3; ClinicalTrials.gov v2 record NCT03131687 with its posted outcome measures; Frias JP et al., Lancet 2018;392:2180-2193 (PMID 30293770), abstract; the PubMed search described above and the records it returned (PMIDs 42668762, 42571951, 42244165, 42201545, 40540673, 39808463 and four unrelated), via NCBI E-utilities. Accumulation arithmetic by this site. Corrections go to the contact page.