Taking the question as asked, rather than the general version of it. One third of STEP 4 participants held most of their loss without the drug, and nobody has convincingly characterised who they are. That subgroup is the most interesting unanswered question in the field and it is routinely flattened into the two-thirds headline.

DebRD_ATL said:

One third of STEP 4 participants held most of their loss without the drug, and nobody has convincingly characterised who they are.

Agreed, though "tolerable" needs defining. A dose you tolerate by eating almost nothing is not tolerated, it is being paid for somewhere else — usually in lean mass, sometimes in adherence three months later.

Dr.Martinez said:

I am eight weeks into deliberately reducing rather than stopping, and the appetite change came back faster than the weight did.

Can confirm the pattern Dr.Martinez describes. Extending the interval and reducing the dose are pharmacologically different. Reducing the dose lowers the whole exposure curve evenly; extending the interval keeps the peak and drops the trough. Since the appetite effect tracks the trough, interval extension tends to give you good days and bad days rather than a uniformly smaller effect, which most people find harder to live with.

Adding the clinical framing, because it changes how the question reads.

Epigenetic implications of GLP-1 therapy and maintenance dosing: emerging evidence suggests that sustained weight loss and metabolic improvement may produce epigenetic changes (DNA methylation, histone modification) that persist beyond drug discontinuation[1].

This is speculative but fascinating: could long-term GLP-1 agonist treatment "reprogram" metabolic gene expression? If so, it would explain why some patients maintain weight loss better than others after discontinuation.

More research needed, but the concept of pharmacologically-induced epigenetic remodeling is intellectually exciting.