Eight weeks is the first point anybody has measured. A post hoc analysis of 2,384 tirzepatide-treated trial participants split them at week 8 by whether they had lost 5% of their body weight [1]. Those who had not still reached a clinically meaningful result by week 72. Theirs was significantly smaller than the early responders’.
Two months in, several hundred dollars spent, and the scale has barely moved. It is the moment the most expensive decision in this market gets made, usually alone and usually without any information. It arrives long before the six-month total anybody planned for. What a slow start does and does not predict is the whole of this page. What counts as working at all is a separate question underneath it.
What was analyzed
Researchers took the tirzepatide-treated participants from the two SURMOUNT trials and sorted them by their week-8 result. That is 1,775 from the trial in people without type 2 diabetes and 609 from the trial in people with it, 2,384 in total. [1] Those who had lost at least 5% of their body weight by then were classed as early responders; those who had not were not. Both groups were then followed to week 72.
Early responders differed from the rest at baseline: they were more likely to be female, more likely to be White, and had lower starting HbA1c. That matters for how to read everything that follows, because nobody was assigned to a group. People sorted themselves by how their own bodies behaved.
What happened by week 72
Both groups achieved clinically meaningful weight reduction and improvements in cardiometabolic risk parameters including HbA1c. The early responders did significantly better on both.
The abstract does not publish the week-72 figures for each group, so this page does not quote any. What it establishes is the direction: a slow start predicted a smaller result, and it did not predict a null one.
What a year looks like outside a trial
A trial figure is the ceiling, not the expectation. A cohort of 14,046 US adults with a BMI of 30 or more measured mean twelve-month weight change in routine care [3]. It came to -13.0% on tirzepatide, -5.9% on semaglutide, and roughly -4.0% each on exenatide, dulaglutide and liraglutide. The trial ordering survived. The scale did not.
That is the number to hold against a monthly price, and it is less than half the trial figure for semaglutide. Set it beside what this market charges a month before deciding what week 8 was supposed to buy.
How wide the spread is
An analysis of 135,349 treated people sorted them by result [4]. Moderate response, between 5% and 15% lost, ran at 40% to 42% regardless of brand. Above 15% the figures were 34% on Zepbound, 26% on Wegovy, 24% on Mounjaro and 10% on Ozempic.
Those four names are two molecules. The gap inside each pair is the approved dose range and therefore the patient, which is why the brand spread is not a product ranking. What it does establish is that about four people in ten land in the middle whatever they take, so an average predicts an individual badly.
The tolerability finding is the useful one
Gastrointestinal side effects were similar between the groups in pattern, severity and time course. That undercuts an intuition a lot of people hold: that feeling sick means the drug is working, and feeling fine means it is not.
On this evidence, how much nausea somebody had told you nothing about whether they were losing weight quickly. People who lost little in eight weeks were not spared the side effects, and people who lost a lot were not punished with more of them. It is also a reason to be careful with what a seller charges for a gentler experience, since tolerability and result appear to move independently.
What this cannot tell you
Whether becoming an early responder would help. Nobody was made into one. The comparison is between two groups of people who differed before the drug started and continued to differ after, so it describes prognosis rather than cause.
There is also no placebo arm inside this analysis. Only tirzepatide-treated participants were included. So “clinically meaningful” here is measured against where people started, not against an untreated comparison. Hold that distinction whenever a result is reported as change from baseline.
The money version of the question
A buyer at week 8 is deciding whether to commit to roughly ten more months. If the published figures for each group existed, that decision could be priced. They do not, so it cannot be — and anybody presenting a precise expectation from this analysis is supplying a number the paper did not.
What can be said is narrower. Stopping at week 8 forecloses a result this analysis says was still available. Continuing is a real cost against a smaller expected return. Most people do not complete a year anyway — persistence in this market is poor and not only because of price. That is a conversation for a prescriber, not an intake form.