News · Heart & Metabolic
The weight came back. The visceral fat loss didn't.
A Circulation follow-up scanned 366 people 5 and 10 years after two diet trials and found that despite complete weight regain, visceral fat stayed partly down, and 10% less of it cut diabetes risk 28%.
Based on peer-reviewed follow-up of two randomized trials in Circulation
- Researchers followed up the CENTRAL and DIRECT-PLUS diet and exercise trials 5 and 10 years after they ended, publishing in Circulation.
- 366 of 381 eligible participants returned, a 96% follow-up rate that is unusual at this distance.
- Everyone regained the weight they had lost. That is the premise, not a failure of the study.
- Despite that, visceral fat and both subcutaneous fat layers partly held their trial-era reductions.
- Liver fat and pancreatic fat did not. Both came back fully, and pancreatic fat overshot.
- Each 10% loss of visceral fat during the trial tracked with better insulin resistance and metabolic syndrome scores years later.
- Only visceral fat loss independently tracked with less diabetes: 10% less carried 28% lower risk (HR 0.72).
- Follow-up is observational even though the original interventions were randomized. Nobody was assigned to keep visceral fat off.
Almost everyone who loses weight on a diet trial puts it back. That is the finding that makes diet research feel futile. A study in Circulation went back to two such trials a decade later, put people in an MRI scanner, and found the scale had been hiding something.
Despite complete weight regain, waist circumference and abdominal fat depots, including visceral adipose tissue, partially preserved their intervention-induced achievements at long-term follow-up.
Why go back at all
The two source trials were serious pieces of work. During the 18-month CENTRAL and DIRECT-PLUS randomized controlled trials, participants achieved considerable reductions in abdominal and ectopic fat, testing Mediterranean diet variants and low-fat patterns alongside structured physical activity.
What nobody had until now was the far end. The team examined the long-term postintervention cardiometabolic profile associated with these changes, which required getting the same people back into a scanner years after the intervention stopped.
They largely managed it: 366 out of 381 eligible participants, a 96% return rate that most follow-up studies never approach.
What stayed down, and what did not
The depots split cleanly, and the split is the story.
Visceral fat and both subcutaneous layers held part of their gains. Postintervention reductions of intrahepatic fat and intrapancreatic fat were fully and excessively gained during follow-up - liver fat returned to baseline, and pancreatic fat ended up worse than where it started.
So the body did not simply rewind. Some compartments tracked the scale; others kept a memory of the intervention long after the weight came back.
The number that changes the target
Losing visceral fat during the trial predicted how people were doing a decade later. Each 10% intervention-induced loss of visceral adipose tissue, superficial subcutaneous fat and intrapancreatic fat was associated with long-term improvements in insulin resistance and metabolic syndrome scores.
For actual disease, only one depot mattered. Only 10% visceral adipose tissue loss was independently associated with a 28% lower risk of incident type 2 diabetes.
Not weight. Not total fat. The deep abdominal compartment specifically, and the effect survived adjustment for weight change, diet adherence and physical activity at follow-up.
What this does to the scale
The authors put it plainly: visceral fat loss rather than weight loss emerges as a key target for durable cardiometabolic health.
That reframes what “the diet failed” means. A person who regained every kilogram may still be carrying a durable metabolic benefit that a bathroom scale is structurally incapable of detecting.
What the follow-up cannot show
The original interventions were randomized. This part is not. Nobody was assigned to lose visceral fat, so people who lost more of it may differ in ways the models did not capture, and the diabetes figure runs from 0.54 to 0.94 - unlikely to be nothing, uncertain in size.
MRI at three time points across a decade is also a demanding protocol, and the people who kept turning up are not a random sample of those who started.
The authors keep the claim conditional: lifestyle interventions may yield long-term improvements in cardiometabolic measures despite weight regain. Which is a more encouraging sentence than most weight-loss research produces.
People also ask
What is visceral fat, and how is it different from the fat I can pinch?
Visceral fat sits deep in the abdomen, packed around the organs, and is invisible from outside. The fat you can pinch is subcutaneous, sitting under the skin. They behave differently: visceral fat is metabolically active and drains directly to the liver, which is the usual explanation for why it tracks with insulin resistance more strongly than total weight does. Telling them apart requires imaging, which is why this study used MRI rather than scales.
If the weight all came back, what was the point?
That is precisely the finding. Weight regain is the normal outcome of diet trials and is usually treated as failure. Here the scans showed that visceral and subcutaneous fat had partly stayed down even as the scale returned to baseline, and that the visceral portion was the part linked to lower diabetes risk years later. It suggests the scale was measuring the wrong thing all along.
Why did liver and pancreatic fat behave differently?
The study documents it without explaining it. Reductions in intrahepatic and intrapancreatic fat were fully regained, and pancreatic fat was regained excessively, meaning it ended up above where it started. Those depots appear to track current weight much more closely than visceral fat does. Why is an open question.
How solid is a 28% reduction in diabetes risk?
It comes with real caveats. The original interventions were randomized, but nobody was randomized to lose visceral fat specifically, so this part of the analysis is observational. People who lost more visceral fat may differ in ways the model did not capture. The confidence interval runs from 0.54 to 0.94, so the effect is unlikely to be zero but its size is uncertain.
Can I target visceral fat specifically?
Not directly, and anyone selling that is overselling. What these trials tested were ordinary diet patterns, including Mediterranean variants, combined with structured physical activity. Visceral fat responded well to those, more so than to weight loss as such. That is a reason to keep going with diet and exercise even when the scale stalls, not a reason to buy anything.