Some people lose weight and still get type 2 diabetes

Does losing weight always protect you from type 2 diabetes?

No. In this 9-year German study, one group of people lost weight and kept it off, and still saw their blood sugar get worse and their type 2 diabetes risk climb. Forty-one percent of that group developed diabetes, compared with none of the lower-risk participants.

Weight loss is the single most recommended way to prevent type 2 diabetes, and for most people it works. But “most people” is not everyone. Researchers in Tübingen, Germany, had already sorted people at risk into six metabolic groups, which they call clusters. Two of those clusters stood out as high risk: cluster 3, where the insulin-making cells in the pancreas are already weak, and cluster 5, made up of older adults with a higher BMI, severe insulin resistance, and fat stored in the liver. This new analysis asked a simple question: do those two high-risk groups get the same payoff from losing weight as everybody else?

What the data show

The researchers started with 190 people who finished a full 24-month lifestyle program in the Tübingen Lifestyle Intervention Program. They then tracked them for an average of 8.7 years, give or take about 1.6 years. Sixty of those people had lost at least 3 percent of their body weight and still kept it off at the long-term check-in, with an average reduction of 8 percent. That is real, sustained weight loss, the kind most prevention programs consider a success.

Among those 60 successful weight losers, the results split sharply by cluster. The 17 people in cluster 5 saw their fasting blood sugar rise more than it did in the 33 people from the lower-risk clusters 1, 2, 4, and 6, and more than in the 10 people in cluster 3. Their 2-hour glucose, the reading taken after a sugar drink, also rose more than in cluster 3. Each of those differences had less than a 1 in 20 chance of being coincidence (P less than 0.05). Their ability to make insulin dropped more than in cluster 3, with roughly a 1 in 100 chance of being coincidence (P = 0.01), and more than in the lower-risk clusters, at about a 1 in 20 chance (P = 0.05).

Then comes the number that matters most. Over the follow-up, 41 percent of cluster 5 participants developed type 2 diabetes, which works out to 7 of the 17 people in that group. In the low-risk clusters, 0 percent did, none of the 33. In cluster 3, 10 percent did, or 1 of 10. That is an absolute gap of 41 percentage points between cluster 5 and the low-risk clusters, and 31 percentage points between cluster 5 and cluster 3. All of these people had lost a comparable amount of weight and kept it off for years.

Dr. Kumar’s Take

The wrong takeaway here is “weight loss does not work.” It works. It worked so well in the low-risk clusters here that not a single person progressed to diabetes. The real message is narrower and more useful: a small subset of people has a metabolic problem that weight loss alone does not fix.

What I think is happening in cluster 5 is that the damage has already moved past the point where lifestyle change alone can reverse it. These are older people with a lot of insulin resistance and fat in the liver, and the beta cells that produce insulin appear to be failing on their own timeline. Losing 8 percent of your body weight lightens the load, but it does not rebuild cells that are wearing out. That is why their insulin secretion kept dropping even as the scale went down.

The clinical implication is that we should stop treating “prevent diabetes” as one strategy applied to one undifferentiated group of at-risk people. If someone does everything right for two years, keeps the weight off for nine, and their fasting glucose is still creeping up, that is not a willpower problem. That is a signal to escalate, not to repeat the same advice louder.

How strong is the evidence?

This is a long follow-up on people who went through a real, structured, 2-year program, and that is a meaningful strength. Nine years of tracking is far more informative than the 6-month or 12-month windows most lifestyle studies use. The clusters were also defined before this analysis, not invented afterward to fit the results.

The main limitation is size. Only 60 people made up the successful weight-loss group, and cluster 5 contained just 17 of them. When 7 people out of 17 develop a condition, a couple of cases in either direction would shift the percentage a lot. The authors say plainly that the finding needs to be repeated in a prospective study designed for this question before anyone treats it as settled. This was also a single program in one region of Germany, so it is not clear how well the cluster patterns transfer to other populations.

What this means for you

If you are working on preventing diabetes, none of this changes what you should do first. Lose weight if you carry extra, and keep it off. For most people that is enough, and the low-risk clusters in this study are the proof.

What changes is the follow-up. Weight is a useful marker, but it is not the outcome you actually care about. The outcome is your blood sugar and your ability to make insulin, and this study shows the two can move in opposite directions. Someone whose glucose keeps rising despite years of maintained weight loss deserves a different conversation, possibly including medication, rather than another round of diet coaching.

Practical Takeaways

Ask your doctor to track your fasting glucose and your A1c over time, not just your weight, so you can see whether your blood sugar is actually responding to the changes you are making.
If you have kept weight off for a year or more and your glucose numbers are still drifting upward, treat that as a reason to escalate care rather than a personal failure.
Pay attention to a fatty liver diagnosis if you have one, since severe insulin resistance with liver fat was a defining feature of the group that did poorly here.
Keep the lifestyle work going regardless, because sustained weight loss still carried this group’s lower-risk peers to a 0 percent diabetes rate over nine years.

FAQs

Can I find out which diabetes risk cluster I belong to?

Not through a standard checkup, at least not yet. The Tübingen clusters were built from detailed research measurements, including tests of how well the body responds to insulin and how much insulin the pancreas releases, plus imaging of liver fat. Most of that goes well beyond a routine fasting glucose or A1c. That said, the general shape of cluster 5 is recognizable in a clinic: older age, higher BMI, strong signs of insulin resistance, and fat in the liver. If that description fits you, it is worth telling your doctor that you may need closer monitoring than the usual once-a-year glucose check.

Why would insulin production keep falling if someone is losing weight?

Weight loss mainly helps by reducing insulin resistance, which means your body needs less insulin to do the same job. It does not directly repair the beta cells in the pancreas that manufacture insulin. In people whose beta cells are already declining, the demand side improves while the supply side keeps shrinking, and eventually supply loses the race. That is the pattern this study picked up in cluster 5, where insulin secretion fell further than in the comparison groups even though the weight loss was comparable.

Does this mean lifestyle programs are a waste of time for high-risk people?

No, and it would be a mistake to read it that way. Cluster 3 participants, who also start with weak insulin production, had a 10 percent diabetes rate over nine years, which is much closer to the low-risk groups than to cluster 5. Weight loss also brings benefits this study was not measuring, including blood pressure, cholesterol, joint pain, and liver health. The argument here is for adding something on top for one specific group, not for taking the lifestyle work away.

Bottom Line

Sustained weight loss prevented type 2 diabetes almost completely in the lower-risk participants of this 9-year study, but it failed one group. People in cluster 5, defined by older age, higher BMI, severe insulin resistance, and liver fat, saw their blood sugar rise and their insulin production fall despite an average 8 percent weight loss held for years, and 41 percent of them developed diabetes. The sample is small and needs confirmation, but it points at something clinicians should take seriously: for a minority of high-risk people, diet and exercise alone are not going to be enough, and waiting to find that out costs them years.

Read the full study