Compressed and extended ladders — what actually changes if a step is two weeks or eight?

Accepted answer

Compression changes what your tolerance judgement means, extension mostly costs time, and there is exactly one published head-to-head comparison of escalation speeds that I am aware of. Taking those in order.

Compression: judging a dose before it has finished arriving

Using the same one-half-life-per-week accumulation model that applies to a weekly agent with a 7-day half-life:

1. After 2 doses you are at 1 − 0.5^2 = 75 % of that dose's eventual steady-state exposure.
2. After 4 doses, 1 − 0.5^4 = 93.8 %.

So a two-week rung asks you to certify tolerance of a dose while a quarter of its exposure is still in transit. Now compound that across a compressed ladder, because this is the part people miss. On a two-week ladder you step up while still 25 % short, so the exposure from rung n is still climbing when rung n+1 starts adding to it. The total exposure at any moment during a compressed escalation is lower than the nominal dose implies — you are always behind — but the rate of change of exposure is roughly double throughout.

That matters because gastrointestinal symptoms on this class track the rate of change of exposure more closely than its absolute level. Somebody who has been at a stable dose for months tolerates an exposure that would have floored them during escalation. Compression is therefore not "the same ladder, faster" — it is a ladder with roughly twice the exposure gradient at every point, which is the specific thing the tolerability literature says is provocative.

The mirror-image error also follows: someone who compresses and feels fine at the end of a two-week rung may be reading a genuine but premature signal. Their exposure has another 25 % to add and they have already committed to going up.

Extension: costs time, and one other thing

Extension is the safer direction and its main cost is calendar. Going from four to eight weeks per rung turns a 16-week semaglutide escalation into 32 weeks, and a 20-week tirzepatide escalation into 40. Everything else being equal you have spent an extra four or five months at sub-maintenance exposure.

The hidden cost is not pharmacological, it is that time at a low rung is time spent under-treated by the standard of the trials, and in the trials the weight-loss curves were still steep during escalation, so this is not lost effect so much as delayed effect. There is no signal that a slower ladder produces a worse final result. There is also no signal it produces a better one.

The genuine argument for extension is that the label itself contains the mechanism — escalation delay by an additional cycle is written into the weight-management label as an option when a dose is not tolerated. An extended ladder is a systematic application of a manoeuvre the label already licenses case by case.

The head-to-head evidence

The dual glucagon and GLP-1 receptor agonist survodutide phase 2 dose-finding trial is the one I know of that randomised escalation speed as well as target dose: it included both rapid and slower escalation to the higher target doses, and reported weight reductions of roughly −12.5 %, −15.0 %, −16.2 % and −18.7 % at 2.4, 3.6, 4.8 and 6.0 mg respectively against about −2.0 % on placebo at 46 weeks [1]. The relevant finding for this question is that the slower escalation arms were better tolerated with broadly comparable efficacy at the same target dose — which is precisely what the exposure-gradient model predicts, and it is worth noting that this is a dose-finding trial in a different molecule and generalises only as an argument, not as a number.

For the approved agents there is no randomised comparison of ladder speed. The four-week interval was carried through phase 2 into phase 3 and never tested against an alternative. That absence should temper confidence in both directions: the label ladder is not proven optimal, and departures from it are not proven safe.

Intermediate rungs and the measurement objection

There is a real reason to be careful about inventing a dose between two labelled ones, and it has nothing to do with pharmacology. It is granulation.

Work the semaglutide 1.7-to-2.4 step. Suppose a 10 mg vial reconstituted to 2.00 mL, so 5.0 mg/mL, ignoring content-assay corrections:

1. 1.7 mg is 1.7 / 5.0 = 0.34 mL = 340 µL = 34 units on a U-100 scale.
2. 2.4 mg is 480 µL = 48 units.
3. A midpoint of 2.05 mg is 410 µL = 41 units.

All three are comfortably readable, so at that concentration intermediate rungs are measurable. Now do the same at the bottom of the ladder with a more concentrated preparation — 10 mg in 1.00 mL, so 10 mg/mL:

1. 0.25 mg is 25 µL = 2.5 units.
2. A midpoint between 0.25 and 0.5 mg is 0.375 mg = 37.5 µL = 3.75 units.
3. On a syringe with half-unit graduations, 3.75 units cannot be read at all. You are choosing between 3.5 and 4.0 units, which is 0.35 and 0.40 mg — a ±7 % band around your intended dose.

So the constraint on fine-grained ladders is not whether the pharmacology supports a 0.375 mg rung; it is whether your measurement system can resolve one. The general rule is that usable dose granularity is set by the concentration you chose at reconstitution, and anyone planning intermediate rungs should pick the reconstitution volume to make the rungs land on readable graduations rather than discovering afterwards that they do not. As ever, none of this is advice about what dose anyone should be on.

edited 16 Oct 2025 by tess_amankwah — reworded for clarity after a comment