Why is the missed-dose window five days for one weekly agent and four for another?

The missed-dose instructions in this class are close to each other but not identical, and I cannot work out whether the differences are pharmacology or arbitrary.

What I have read across the prescribing information for the weekly agents:

• one says a missed dose can be administered within five days, and after that to skip it and resume on the normal day;
• another says four days;
• a third says three days;
• and each of them separately specifies a minimum interval between two doses, which is a different number again.

These are all weekly subcutaneous agonists with multi-day half-lives. Five days, four days, three days. Is that a real reflection of different pharmacokinetics, or three regulatory committees rounding differently?

What I would find genuinely useful is the arithmetic behind the windows. If I could see what the exposure consequence of using a window to its limit actually is, I would understand what the number is protecting against. I am asking about how the labels were derived, not about what to do on any particular week.

One correction to the accepted answer's model that makes it slightly conservative in a direction worth naming: absorption is not instantaneous, and for these agents it is slow enough to blunt the overshoot.

Semaglutide reaches maximum plasma concentration roughly one to three days after a subcutaneous dose. That is a flat absorption phase, not a spike. So the "peak" in the accepted answer's model is an idealisation that assumes the whole dose lands at once; in reality each dose is spread over a couple of days.

The consequence for a stacking calculation is that two doses two days apart overlap in their absorption phases rather than producing two distinct additive peaks, and the resulting maximum is lower and later than the instantaneous model predicts. A reasonable expectation is that a computed +16 % becomes something like +10 to +13 % in a real concentration-time curve. The direction is unambiguous — slow absorption always smooths peaks — and the magnitude depends on the absorption rate constant, which is why I am giving a range and not a number.

Two things follow:

• The accepted answer's numbers are usefully wrong in the safe direction. They overstate the overshoot, which is the right way for a back-of-envelope model to fail.
• The same smoothing is why peak-related tolerability arguments in this class are weak in general. Nobody in this class experiences a sharp concentration spike from anything, which is the pharmacokinetic reason that dose stacking within a label window has modest symptomatic consequences and the reason the withdrawal-of-drug direction is the more consequential error.

Worth adding one more term the model omits: inter-individual variability. Reported exposure variability for these agents runs at a coefficient of variation in the region of 30 % or more. A 16 % computed overshoot sits well inside the spread between two people on the same dose. That does not make the overshoot irrelevant — it is a change in your own exposure, and your own baseline is what your tolerance was calibrated to — but it does mean the label windows are drawn inside the noise, which tells you they were set conservatively.

The daily agents follow a completely different rule and it is worth understanding why, because the contrast makes the weekly logic clearer.

For a daily agent the instruction is generally to skip and resume, with no catch-up dose. Two reasons, and they pull in the same direction:

1. The fractional consequence of one missed dose is small. On a daily schedule with an accumulation ratio around 1.4, one missed dose drops the level by less than a third and the next dose almost restores it. Missing one of seven weekly exposures is not comparable to missing one of one.
2. The catch-up window would be measured in hours and is not usable. With a 13-hour half-life, taking a dose 12 hours late and the next one on schedule 12 hours after that is a real overshoot, and nobody can be relied upon to compute a window that narrow. So the instruction is a categorical skip rather than a conditional window.

The oral weekly-equivalent case is the odd one. Oral semaglutide is dosed daily but has the same roughly 7-day half-life as the injectable, giving an accumulation ratio around 7.5 and a very flat concentration-time profile. One missed daily dose changes total exposure by on the order of a few per cent, so a catch-up dose would be pointless as well as unnecessary — and the absorption of an oral dose depends on tightly specified fasting and fluid conditions, so doubling up introduces an absorption-variability problem on top. Skip and resume, for once, for reasons of both pharmacokinetics and formulation.

The genuinely different case is the daily agent with an explicit re-initiation rule. The liraglutide weight-management label instructs that if more than three days have elapsed since the last dose, dosing is re-initiated at the starting dose and re-escalated. Three days is about 5.5 half-lives, which is essentially complete washout — under 3 % of steady state remaining. That is a washout-based rule, and it is the only one in the class that clearly is.

A record-keeping note, since the arithmetic above is only usable if you know when the last dose actually happened.

Every calculation in this thread takes the form "days since the previous dose". People consistently overestimate the reliability of their own memory of that. A missed dose is by definition a break in routine, and a break in routine is exactly the circumstance under which the previous date becomes uncertain — the whole reason it was missed is usually that the week was irregular.

What makes the difference in reported practice is trivial: a written date on the vial or the carton, updated at the time of each administration rather than afterwards. Not a phone reminder, which tells you the schedule rather than the history, and not a mental note. The specific failure mode is reconstructing "I think it was Tuesday" and being wrong by a day in the direction that turns a legitimate catch-up into an interval below the label minimum.

Secondary benefit: a dated log is the only way to notice a slow drift. Somebody who moves the day forward by a day whenever it is convenient will, over six months, discover their schedule has rotated most of the way round the week, and each individual shift was inside the label allowance while the cumulative pattern was not a weekly schedule at all.