Accepted answer
The honest answer is that no peer-reviewed degradation-kinetics dataset exists for these peptides reconstituted in unbuffered bacteriostatic water, and every number in circulation is an in-use limit borrowed from a formulated product. That does not make the numbers useless — it means you have to know what they are before you lean on them.
What is actually documented
Here is the real set of documented storage windows for approved incretin products, with the basis for each stated explicitly. Read the last column, because it is the whole point of the table.
| Product / presentation | Condition | Documented window | Basis of the number |
|---|
| Semaglutide, multi-dose pen | 2–8 °C, after first use | 56 days | label in-use limit — product meets specification through the period |
| Semaglutide, multi-dose pen | 15–30 °C, after first use | 56 days | label in-use limit; same window as refrigerated |
| Semaglutide, single-dose pen | 8–30 °C, before use | 28 days | label excursion allowance, then discard |
| Tirzepatide, single-dose presentation | up to 30 °C | 21 days | label excursion allowance, single use only |
| Liraglutide, multi-dose pen | 2–8 °C or 15–30 °C, after first use | 30 days | label in-use limit |
| Dulaglutide, single-dose pen | up to 30 °C | 14 days | label excursion allowance |
| Exenatide twice-daily, multi-dose pen | up to 25 °C, after first use | 30 days | label in-use limit |
| Any of the above | frozen | discard | explicit label instruction — do not use if frozen |
| Research cake in bacteriostatic water | 2–8 °C | no published figure | none — every quoted number is extrapolated from the rows above |
Three observations from that table that matter more than any single row.
First, the numbers track the presentation, not the molecule. Semaglutide is 56 days in one presentation and 28 in another. Tirzepatide is 21 days and dulaglutide 14 days, and neither difference reflects a fourfold difference in intrinsic peptide stability — it reflects single-dose versus multi-dose design, preservative presence, and how much excursion the sponsor chose to characterise. Quoting "semaglutide lasts 56 days" as a property of semaglutide is a category error.
Second, refrigerated and room-temperature in-use windows are often identical. Both semaglutide multi-dose rows are 56 days. If the limit were set by chemical degradation you would expect a large temperature dependence. It is not — it is set by preservative effectiveness and by how long the sponsor tested, which is why the two conditions collapse to the same figure.
Third, "do not freeze" is unanimous. That is the one instruction where every label agrees, and it is the one instruction people most often ignore.
Estimating what you cannot look up
If you want to reason about your own conditions, the tool is temperature-dependence arithmetic with the assumptions stated out loud. Peptide degradation in solution in this temperature range typically has a Q10 — factor change in rate per 10 °C — somewhere around 2 to 3. Take Q10 = 2.5 and use the 56-day refrigerated window at about 5 °C as the anchor:
- From 5 °C to 25 °C is ΔT = 20 °C, which is 20/10 = 2.0 Q10 steps. Rate multiplier = 2.5^2.0 = 6.25. Equivalent time = 56 / 6.25 = 9.0 days.
- From 5 °C to 30 °C is ΔT = 25 °C = 2.5 steps. Multiplier = 2.5^2.5 = 9.9. Equivalent time = 56 / 9.9 = 5.7 days.
- From 5 °C to 40 °C is ΔT = 35 °C = 3.5 steps. Multiplier = 2.5^3.5 = 24.7. Equivalent time = 56 / 24.7 = 2.3 days.
Now notice something important: the label allows 56 days at 15–30 °C, and line 2 of that calculation predicts 5.7 days. The calculation is far more conservative than reality for the formulated product. Why? Because the anchor is wrong — 56 days is not a degradation endpoint, so dividing it by a rate ratio is dividing the wrong quantity. The arithmetic is still useful, but only as a relative statement: whatever your material's real half-life is, moving it from 5 °C to 40 °C multiplies the degradation rate by roughly 25. That ratio is the transferable part. The absolute days are not.
What this means for a research vial
An unbuffered reconstitution in bacteriostatic water differs from the pen formulation in ways that mostly work against you: no buffer, so pH is uncontrolled and drifts with dissolved carbon dioxide; typically no surfactant, so the air-liquid interface is unprotected and interfacial aggregation is faster; a stopper entered many times rather than a sealed cartridge; and no stability programme behind any of it.
The commonly reported practice of a 28-day refrigerated in-use window is therefore defensible not because it is a measured degradation limit but because it is the preservative-effectiveness window from the compendial antimicrobial testing chapter, which is a real and relevant constraint. Treating 28 days as a microbiological limit is sound. Treating it as a chemical guarantee is not supported by anything.
The trial programmes that established these molecules' efficacy — the STEP series for semaglutide [1] and SURMOUNT-1 for tirzepatide [2] — used temperature-controlled clinical supply of the formulated product throughout. Nothing in that literature speaks to a reconstituted research solution, and it should not be cited as if it did.
edited 6 Aug 2025 by sian_llewellyn — added the placebo-arm figures
The observation that refrigerated and room-temperature in-use windows are identical is the single most clarifying thing in this thread. – Dr_Idris_Coulibaly 8 months ago Dividing the wrong quantity by a rate ratio — yes. Half the confident numbers online are exactly that error. – Dr_Ilse_Vandenberg 7 months ago 8The pH drift point deserves more attention. Unbuffered water is not a controlled solvent. – rukhsana_iqbal 1 months ago add a comment