Yes, in a specific and measurable sense. A 2024 Nature study found that fat tissue in both humans and mice retains altered cell level activity after appreciable weight loss, and that in mice those changes are written into the epigenome and come with faster rebound weight gain. Separately, appetite hormones measured a year after a diet had still not returned to where they started.
Neither finding says regain is fixed. Both say the body you are maintaining is not the body you had before you lost the weight.
What the 2024 Nature study found
Researchers at ETH Zurich and collaborators used single nucleus RNA sequencing on adipose tissue from humans and mice. In both species, fat tissue that had been through obesity and then appreciable weight loss still showed transcriptional changes, meaning the cells were running a different pattern of gene activity than tissue that had never been through it.
In the mice they went further and looked at the epigenome, the layer of chemical marks that decides which genes are available to be read. Those marks were persistently altered in adipocytes, the changes impaired how the cells responded to metabolic stimuli, and mice carrying them showed accelerated rebound weight gain when high fat feeding resumed. The authors named the phenomenon an obesogenic memory.
Where the human half of that claim stops
This is the part worth being pedantic about, because the headline version of this study lost it immediately. The human tissue showed transcriptional changes. The epigenetic marks, the impaired cell function and the accelerated rebound were demonstrated in mice.
The paper's own conclusion is careful in exactly that way. It says the findings indicate an obesogenic memory in mouse adipocytes and probably other cell types, and that these changes seem to prime cells for pathological responses. Probably and seem are the authors' words. A study can be genuinely important and still be one step short of the claim people want to make with it.
The human evidence that is not in doubt
The hormonal version of this memory was settled more than a decade earlier. A 2011 trial in the New England Journal of Medicine put 50 people through a ten week very low energy diet, during which they lost a mean of about 13.5 kg, then measured nine circulating appetite mediators at baseline, at ten weeks and again at 62 weeks.
A year after the initial weight loss, leptin, peptide YY, cholecystokinin, insulin, ghrelin, gastric inhibitory polypeptide and pancreatic polypeptide were all still significantly different from each person's own starting levels. So was subjective hunger. The signals that push toward eating more had not stood down after twelve months of the new weight being the actual weight.
The direction of each change is worth reading, because together they describe a coordinated argument rather than a random drift. Immediately after the weight loss, leptin fell, along with peptide YY, cholecystokinin, insulin and amylin, all signals associated with fullness. Ghrelin rose, along with gastric inhibitory polypeptide and pancreatic polypeptide. Subjective appetite rose with them. Every one of those moves points the same way, toward eating more, and a year later most of them had not moved back.
That is the same shape of result as the adipose work, arriving from a completely different direction. When two unrelated measurement techniques both find that the previous state left a residue, the residue is probably real.
What a memory does and does not determine
It sets a bias, not an outcome. A mouse with altered adipocytes still needs the high fat diet to regain. A person whose ghrelin is elevated still has to eat the food that appetite is arguing for. The finding explains why maintenance takes ongoing effort rather than why it fails.
Where it genuinely changes the plan is in how long you expect to need a system. If you have been treating maintenance as a recovery period that eventually ends, the evidence says the thing you would be recovering from has not gone anywhere. Whether a set point can be lowered is the closest live question to this one, and the yo-yo pattern is what the 2024 authors explicitly connect their finding to.
It also sits alongside the tissue level fact that fat cells do not disappear when you lose weight. Different mechanism, same conclusion: the structure stays and the programming changes.
Why OffRamp's radar has no finish line
Most weight apps are shaped like a course. You set a goal, you follow a plan, the plan ends. OffRamp is deliberately not shaped that way, and this research is the reason.
Its Regain Radar holds a baseline you set and compares your 7 day rolling average against it indefinitely. There is no week twelve, no graduation screen, no point at which the comparison stops. If your average sits at 2 percent above baseline for seven consecutive days it flags drift, and it will do that in month two and in year four with the same rule. When a day has no weight in it, the window carries the last average forward rather than breaking the streak, because the system has to survive real life to still be running in year four.
The honest argument for that design is the one above. If the tissue and the hormones both hold a record of the higher weight, then a monitoring system that expires is a monitoring system that stops exactly when the biology does not. The set point and adaptation picture makes the same case from the energy side.
None of this is medical advice or guidance about any medication. Decisions about treatment for obesity belong with your prescriber.
