IPM Take
Obesity care has turned the scale into a verdict.
Weight goes down, the intervention worked. Weight returns, the patient failed.
A decade-long MRI follow-up suggests the biology may be less punitive. Participants regained their lost weight on average, yet reductions in visceral fat were partly maintained and were associated with better long-term metabolic outcomes. Every 10% reduction in visceral fat during the original intervention was associated with a 28% lower risk of developing type 2 diabetes during follow-up.
That does not mean weight no longer matters, and it does not justify routine MRI scanning. It does mean policymakers, payers and clinicians should question whether body weight alone is an adequate measure of treatment value.
A patient can regain kilograms without returning to exactly the same metabolic state. Calling that failure may be clinically crude, economically shortsighted and unnecessarily stigmatising.
Executive Summary
Researchers followed participants from two 18-month randomised lifestyle trials, CENTRAL and DIRECT-PLUS, which tested dietary strategies including low-fat guidance and Mediterranean diet variants, together with structured physical activity. Of 381 participants who had MRI data at the beginning and end of the original interventions, 366, or 96%, were reached during follow-up conducted five or 10 years later. The cohort’s mean age at follow-up was approximately 57 years.
Body weight had returned to baseline on average. However, waist circumference and abdominal fat stores, including visceral, deep subcutaneous and superficial subcutaneous fat, remained below baseline. Reductions in liver fat were not maintained, while pancreatic fat increased above its original level.
Greater intervention-induced visceral fat loss was associated with sustained improvements in insulin resistance, overall cardiometabolic risk and metabolic-syndrome severity. Most notably, each 10% reduction in visceral adipose tissue was associated with a 28% lower risk of incident type 2 diabetes during follow-up, with a hazard ratio of 0.72 and a 95% confidence interval of 0.54 to 0.94. Visceral fat was the only measured fat depot whose reduction independently predicted lower future diabetes risk.
The findings are important but should not be interpreted as proof that reducing visceral fat by 10% directly causes a 28% reduction in diabetes. The original dietary interventions were randomised, but participants were not randomly assigned to achieve particular amounts of visceral fat loss. This long-term analysis therefore identifies an association between the biological response and later outcomes.
The study population was also highly selective. At baseline, 91.3% of participants were men, the mean BMI was 30.6 kg/m² and around 64% met criteria for metabolic syndrome. The results therefore require confirmation in women, younger populations and more ethnically, socially and geographically diverse groups.
The supplied coverage accurately highlights the central policy signal: weight regain did not necessarily erase every cardiometabolic benefit achieved during the intervention.
Why it matters
- Policymakers and public authorities: Obesity strategies should measure health improvement, not only kilograms lost. Outcomes such as waist circumference, glucose control, blood pressure, lipids, physical function and disease prevention may offer a fuller picture.
- HTA bodies and payers: Reimbursement models built around rigid percentage weight-loss thresholds may miss clinically meaningful responses that are not visible on the scale.
- Clinicians and providers: Weight regain should trigger reassessment and renewed support, not an automatic conclusion that the original intervention achieved nothing.
- Patients and advocates: The findings challenge the language of personal failure that often surrounds weight regain. Biological benefits may persist even when maintaining maximum weight loss proves difficult.
- Researchers and industry: Future obesity trials should examine body composition and long-term cardiometabolic outcomes alongside total body weight, while ensuring that expensive imaging does not become an unnecessary barrier to care.
The patient steps onto the scale.
The number has gone back up.
In many obesity pathways, that single moment defines the entire story.
The diet failed. The intervention failed. The treatment failed. Sometimes, the patient is made to feel that they failed.
A long-term MRI study published in Circulation suggests the body may be keeping a more complicated record.
Researchers followed participants from two randomised lifestyle trials for five or 10 years after the original interventions ended. During the trials, participants changed their diet, increased physical activity and lost weight and internal fat. Years later, their body weight had returned to baseline on average.
But their bodies had not completely returned to where they started.
Some of the reduction in visceral fat, the fat stored deep inside the abdomen around internal organs, was still present. Waist circumference and several abdominal fat compartments also remained below baseline. Those who had lost more visceral fat during the original intervention showed better long-term metabolic profiles and a lower subsequent risk of type 2 diabetes.
The finding challenges one of the most deeply embedded assumptions in obesity policy: that weight regain erases success.
It may erase part of it.
It may not erase all of it.
The fat that matters may not be the fat the scale measures
A bathroom scale cannot distinguish visceral fat from subcutaneous fat, muscle, water or fat stored in the liver and pancreas.
It records total mass.
That simplicity makes weight cheap, practical and useful. It also makes it biologically incomplete.
Visceral fat is metabolically active. Excess accumulation is associated with insulin resistance, dyslipidaemia and other cardiometabolic abnormalities. The new study used 3.0-Tesla MRI to separate visceral fat from deep and superficial subcutaneous fat and to quantify ectopic fat in the liver and pancreas.
At the beginning of the original trials, the 381 participants with repeat MRI scans had a mean visceral fat area of 148.6 cm². Following the 18-month interventions, this fell to 109.0 cm². At long-term follow-up, it had risen to 124.8 cm², meaning some of the lost visceral fat returned, but the average remained below baseline.
Body weight, by contrast, had fully returned to baseline on average.
This divergence matters.
It suggests weight regain may not always recreate the same distribution of fat or the same cardiometabolic risk profile. A person can return to an earlier weight while retaining some changes in abdominal composition, behaviour or metabolic function.
The researchers described this as a lasting metabolic legacy. That language is provocative, but the data support caution rather than celebration.
The study did not show that temporary dieting permanently protects everyone. It did not show that repeated weight cycling is harmless. It did not establish that regaining weight should be ignored.
It showed that biological progress may be partly durable even when the scale suggests a complete reversal.
A 28% lower diabetes risk, with an important caveat
The most striking result concerned type 2 diabetes.
For each 10% reduction in visceral fat achieved during the intervention, the risk of developing type 2 diabetes during long-term follow-up was 28% lower after adjustment for factors including initial fat levels, weight change, diet group and physical activity. The estimate was statistically significant, but the confidence interval remained relatively wide.
This is not equivalent to saying that clinicians can guarantee a 28% risk reduction by reducing a patient’s visceral fat by 10%.
The original trials randomly assigned dietary interventions. They did not randomly assign the amount or type of fat that each participant would lose. People who lost more visceral fat may have differed biologically or behaviourally from those who lost less, even after statistical adjustment.
The relationship is therefore persuasive, not definitive.
It should generate new trials, not new promises.
The study also found that changes in other fat depots were associated with some long-term metabolic measurements, but only visceral fat loss consistently predicted a lower risk of future diabetes. Liver fat returned to baseline, while pancreatic fat increased beyond baseline at follow-up.
This complexity undermines the idea that there is one simple hierarchy in which every kilogram lost is beneficial and every kilogram regained is harmful in exactly the same way.
Where fat is stored may matter.
When it is lost may matter.
How long particular changes persist may matter.
The metabolic effects of an intervention may continue after its most visible effect has faded.
Obesity reimbursement is still built around the scale
Health systems need practical thresholds. They cannot perform MRI scans on every person receiving lifestyle support or obesity medication.
But practical does not have to mean simplistic.
In England, NICE guidance states that clinicians should consider stopping semaglutide when a person has lost less than 5% of their initial weight after six months. For tirzepatide, clinicians should decide whether treatment should continue when weight loss remains below 5% after six months on the highest tolerated dose, taking individual benefits and risks into account.
These rules have a rational basis. Medicines carry costs and adverse effects, and continuing treatment without meaningful benefit may be inappropriate.
The problem is how benefit is defined.
Percentage weight loss is measurable and supported by trial evidence. Yet it may not capture every clinically relevant response. A patient could lose less than 5% of body weight while improving glycaemic control, blood pressure, waist circumference, mobility, sleep apnoea or visceral fat.
Conversely, someone may lose substantial weight while continuing to face serious metabolic risk.
The new study does not invalidate current stopping rules. It does suggest that treatment assessment should be broader than a binary scale-based pass or fail.
This is especially important as obesity medicines become integrated into cardiovascular and diabetes prevention. A therapy is not valuable because it makes a number smaller. It is valuable if it reduces disease, disability and premature death.
HTA bodies and payers should increasingly ask whether evaluation frameworks reflect those outcomes.
MRI cannot become the new gatekeeper
There is an obvious danger in responding to an inadequate measurement by replacing it with an inaccessible one.
MRI can provide detailed information on visceral and ectopic fat. It is also expensive, capacity-intensive and unnecessary for most routine obesity care.
Making MRI-defined visceral fat loss a condition of treatment would create a new layer of inequality. Well-funded health systems and private clinics could provide sophisticated metabolic phenotyping, while patients in lower-resource settings would continue to be assessed by scales and tape measures, or receive no structured care at all.
The goal should not be universal internal-fat imaging.
It should be better outcome selection.
Waist circumference, blood pressure, HbA1c, fasting glucose, lipids, liver markers, physical function and medication requirements can provide additional information at far lower cost. No single measure is perfect, but together they can prevent the scale from monopolising the definition of progress.
Future research may also identify validated blood biomarkers or imaging proxies that estimate visceral-fat change without full MRI assessment. Those tools will still need to demonstrate that they improve decisions, not simply produce more data.
Weight regain is often treated as a moral event
The study’s most immediate impact may be psychological rather than technological.
Weight regain is common after lifestyle intervention. Obesity is a chronic, biologically complex condition influenced by genetics, appetite regulation, medications, stress, sleep, food environments, income and access to continuing treatment.
Yet health systems often offer time-limited interventions and expect permanent results.
When weight returns, responsibility moves rapidly from the programme to the patient.
The programme ended.
The support disappeared.
The environment remained unchanged.
The patient is then blamed for responding predictably to biology and policy.
The new findings do not make weight regain desirable. They make the language of total failure harder to justify.
A person who improved their diet, became more active, reduced visceral fat and lowered metabolic risk did not achieve nothing simply because some or all of the weight later returned.
The intervention may have been incomplete.
The benefits may have diminished.
But some biological protection may remain.
That changes how clinicians should speak to patients. It also changes what health systems should do next.
Instead of discharging people after a short programme and waiting for relapse, care pathways should offer long-term maintenance, periodic reassessment and rapid re-entry when risk begins to rise. Obesity should be managed as a chronic condition, not as a temporary project with a final weigh-in.
The study itself is not representative enough
The enthusiasm around these findings must also confront a major limitation.
More than 91% of participants were men. The trials recruited people with abdominal obesity or dyslipidaemia, and the cohort had a mean BMI of approximately 30.6 kg/m² at baseline.
That limits generalisability.
Fat distribution differs by sex, age, hormonal status, ancestry and disease profile. Visceral-fat thresholds and relationships with diabetes may not be identical across populations.
The next studies must include substantially more women and participants from diverse ethnic, geographic and socioeconomic backgrounds. They should examine whether the relationship holds after different lifestyle interventions, obesity medicines and metabolic or bariatric surgery.
Researchers should also determine whether actively targeting visceral-fat reduction produces better outcomes than targeting total weight loss, rather than only observing that patients who lose more visceral fat later do better.
Until then, visceral fat is a promising therapeutic target, not a new universal treatment endpoint.
The scale is useful. It should not be sovereign.
Weight remains clinically relevant.
Large and sustained weight reduction can improve multiple obesity-related conditions. Tracking body weight is affordable, understandable and feasible almost everywhere.
But the scale should be treated as one measurement, not a judgement.
The FIT follow-up suggests the body may preserve part of the benefit of an earlier intervention even after weight regain. It also suggests that visceral fat may be more closely connected to durable metabolic health than total weight alone.
That should change the political conversation around obesity.
Success should not be reduced to appearance.
Failure should not be assigned at the first sign of regain.
Treatment value should not be defined by kilograms alone.
The weight came back.
The evidence suggests not everything else did.
