Why weight returns: the physiology behind the withdrawal trials

The withdrawal trials tell you what happens. This page is about why, from the studies that measured the mechanism directly. None of them used a GLP-1; that is the point. The adaptation they describe is a response to weight loss by any route.

Appetite hormones move, and stay moved

In 2011 Sumithran and colleagues put 50 adults with overweight or obesity on a very-low-energy diet for 10 weeks. They lost about 13.5 kg. Before, after, and one year later, the team measured circulating hormones and asked about hunger. After the loss, leptin, peptide YY, cholecystokinin, insulin and amylin were lower and ghrelin was higher, which together push towards eating more. At one year, with a mean regain of about 5.5 kg, the hormone changes were still present and hunger ratings were still higher than at baseline (PubMed 22029981).

The direction of every one of those changes is towards regaining. The body reads a lower weight as a deficit to be corrected and keeps signalling until it is. A GLP-1 receptor agonist acts on the appetite side of this: it slows gastric emptying and acts on central appetite circuits, so the signals to eat are quieter while the drug is on board. Take the drug away and the signals are still there, which is what the STEP 1 extension recorded: two-thirds of the loss back in a year (PubMed 35441470).

Energy expenditure falls further than the weight alone predicts

A lighter body costs less to run. That part is arithmetic. The additional part, often called metabolic adaptation or adaptive thermogenesis, is a drop in resting expenditure beyond what the change in body composition predicts. The most-cited measurement is Fothergill's follow-up of 14 Biggest Loser contestants: six years after the competition, resting metabolic rate was about 500 kcal per day below what their body composition predicted, despite most of them having regained much of the weight (PubMed 27136388). It is a small, extreme sample; the size of the effect in ordinary weight loss is smaller and debated, but its existence is not.

Hall's 2011 model puts the two together. Because expenditure falls as weight falls, a fixed reduction in intake does not produce a fixed rate of loss. It produces a curve that decelerates and approaches a new equilibrium over years, with about half the total change in the first year (PubMed 21872751). The curve you see in STEP 1 and SURMOUNT-1 has that shape for that reason. When the drug stops and intake rises, the same dynamics run in reverse.

What the lean-mass question adds

Lean tissue is the largest determinant of resting expenditure. If a share of the weight lost is muscle, resting expenditure falls by more than if it were all fat, and regain tends to arrive as fat first. The DXA substudies put the lean share of GLP-1 weight loss at roughly a quarter to two-fifths in the pivotal trials. That is the mechanistic reason protein and resistance training sit at the centre of this site: they are the levers with evidence for changing that share.

What this means in practice

• Regain after stopping is the expected physiological response, not a personal failure. Plan for it before it happens.
• The drug does not reset the set point. It holds appetite down while it is taken. The withdrawal trials show what happens when that hold is released.
• The parts you can move without the drug are lean mass, activity, sleep and monitoring. None reproduces the drug effect; together they are what the evidence supports for holding more of a loss.
• The decision to continue, change or stop belongs with your prescriber. The maintenance planner produces the questions to take to that conversation.