What did TRIUMPH-3 do with participants who could not tolerate a step?

This came up because two published schedules for the same agent differ.

I have read the primary source rather than the summary, which has left me with more questions.

I understand the headline. I do not understand the footnotes, and the footnotes look important.

How should I read this, and where are the traps?

The steady-state arithmetic is worth doing once, because it explains most of what people find confusing about weekly dosing.

Escalating in response to a plateau is a specific and common error of reasoning. A plateau after four to six months is the expected trajectory in every trial in the class — the curve flattens because energy expenditure falls with mass, not because the receptor stopped working. A dose step may still be reasonable; "the loss stopped" is not by itself the reason.

The relevant detail is that holding at a step for longer than four weeks before escalating does appear to reduce cumulative gastrointestinal burden, which is unsurprising given that adverse events cluster in the one to two weeks after each increase. What it does not do is change where you end up, provided you get there.

The caveat that matters: dose decisions on a licensed medicine belong with a prescriber, and dose decisions on research-use-only material belong to a category where nobody has any obligation to you at all.

Write down in advance what would make you hold a step, because deciding that in the middle of a bad week is not when you are at your most analytical.

edited 7 Aug 2026 by tandem_gradient — clarified the distinction between purity and content

For a compound with a seven-day half-life, dose timing is much less consequential than people expect and dose rate is much more consequential.

The maintenance question turns on what you are maintaining. If the objective is weight maintenance, the withdrawal-extension data suggests that a fraction of the therapeutic dose retains a substantial fraction of the effect. If the objective is the cardiovascular or renal endpoint, the trials that demonstrated those endpoints used the full dose, and extrapolating downward is not supported by anything.

The underlying point is that where the label ladders differ between agents, the differences track the potency ratio and the tolerability profile rather than anything deeper. It is worth reading the ladders side by side once, because the pattern — small starting dose, four-week steps, a defined maintenance range and a defined maximum — is identical in structure across the class.

SURMOUNT-4 is the withdrawal trial to read on the maintenance question: after an open-label lead-in, randomised withdrawal produced substantial regain in the placebo arm while continued treatment produced continued loss. It is the cleanest available answer to "what happens if I stop".

The short version: four-week steps because that is steady state, and the ladder is about the side effects, not the result.