BMI May Have Threshold Effect on Diabetes Risk

Liam Davenport

September 08, 2020

Body mass index (BMI) appears to have a threshold effect on diabetes risk that could explain why aggressive weight loss strategies can reverse diabetes, suggests a large UK data analysis.

Prof Brian Ference

"This means that to prevent diabetes, both BMI and blood sugar should be assessed regularly," said lead researcher Professor Brian Ference, from the University of Cambridge, in a news release.

"Efforts to lose weight are critical when a person starts to develop blood sugar problems."

The research was presented at the European Society of Cardiology Congress 2020 on August 31, which was held digitally due to the coronavirus pandemic.

Polygenic Diabetes Risk Score

Prof Ference and colleagues examined data from the UK Biobank on more than 445,000 individuals, of whom more than 30,000 developed type 2 diabetes during follow-up for an average of 8 years.

They found that a polygenic diabetes risk score derived from millions of gene variants was associated with a 2.5-fold increased risk of developing the disease, a figure that was dwarfed by the 11-fold risk increase seen with BMI.

The team nevertheless found that "BMI and a polygenic predisposition appear to have independent and additive effects on the risk of developing diabetes", Prof Ference said in his presentation.

While "BMI appears to have a much more powerful impact on the risk of diabetes than polygenic predisposition", he said, "a polygenic score for diabetes can help to refine the estimates of diabetes risk at all levels of BMI".

Prof Ference said that the most important finding was that lifelong exposure to increased BMI appeared to have the same effect on the risk of diabetes as short-term exposure to raised BMI, implying that "BMI has a threshold effect on the risk of diabetes, rather than a cumulative effect over time".

This, he said, suggests that "most cases of diabetes can be either prevented by keeping BMI below the threshold at which dysglycemia occurs, or reversed as the diabetes develops by lowering BMI below the threshold of dysglycaemia".

This calls for "aggressive therapeutic weight loss to reduce BMI below each person’s threshold of dysglycaemia, as an attempt to reverse diabetes before irreversible beta cell injury occurs". At the same time, glucose control is vital "to prevent symptoms as well as prevent micro and macro-vascular complications".

Diabetes Reversal

Speaking to Medscape News UK, Prof Ference said that the finding of a threshold rather than cumulative effect of BMI on diabetes risk supports the findings of aggressive weight loss studies that have achieved diabetes reversal.

These include the landmark DiRECT trial, in which a calorie-restricted liquid diet followed by gradual food reintroduction and a weight loss maintenance programme allowed 36% of patients to attain remission of their diabetes and sustain it for 24 months.

More recently, an analysis published in The New England Journal of Medicine, indicated that the metabolic benefits of gastric bypass surgery and diet were similar and were related to weight loss itself rather than anything specific to the surgery.

"In the past, our goal has been to treat somebody with their glucose lowering medicine and simply recommend weight loss, but some of the therapies we gave, like insulin…actually cause an increase in weight, which would prevent them from going below their threshold and not cure their diabetes," said Prof Ference. What's more, this strategy could lead to them "requiring more and more medicines".

Prof Ference added that the goal is now "not simply controlling the glucose, which is extremely important, but also there’s a dual goal to aggressively lower weight in order to reduce somebody below their threshold in an explicit attempt to reverse diabetes".

Clues

Professor Partha Kar, associate national clinical director for diabetes for NHS England and a consultant in diabetes medicine at Portsmouth Hospitals NHS Trust, stressed that the study is focused on type 2 and not type 1 diabetes, which cannot be reversed.

He told Medscape News UK that the results "could offer clues into the future as regards thresholds, which indeed may help in further obesity management and prevention strategies around type 2 diabetes".

"It would be of course desirable to see further details of this work in a peer reviewed publication to take the findings into further account," he added.

Anna Morris, assistant director of research strategy and partnerships at Diabetes UK, commented: "The study shows that for everyone, even those with a higher genetic risk, having a lower weight is associated with reduced risk of developing type 2 diabetes.

"It adds to a vast body of evidence showing the critical role of weight management in the prevention of type 2 diabetes and the importance of getting support to lose any extra weight," she told Medscape News UK.

However, Roy Taylor, professor of medicine and metabolism at Newcastle University, Newcastle upon Tyne, who led the DiRECT study, was sceptical of the clinical utility of the results, suggesting that 'big data' analyses such as this cannot determine an individual’s risk of developing diabetes.

Unknowns

Prof Ference began his presentation by saying that both observational data and randomised trials have shown that individuals with diabetes "appear to have a two-fold higher risk of experiencing cardiovascular events; therefore, preventing diabetes would be an effective strategy to reduce cardiovascular disease".

He went on to note that "although obesity is a powerful risk factor for diabetes, recently polygenic scores consisting of millions of genetic variants have also emerged as a strong risk factor".

It could therefore be possible to use a polygenic risk score to focus efforts on preventing diabetes in people with the highest genetic predisposition, but "whether this will be an effective strategy to prevent diabetes is unknown".

To determine the relative contribution of BMI and polygenic risk score to diabetes risk, and how they could be integrated, the team looked at data on 445,765 individuals enrolled in the UK Biobank, of whom 54% were female and the mean age at entry was 57.2 years.

They constructed a polygenic risk score from 6.9 million genetic variants associated with the development of diabetes, and divided the patients into quintiles of increasing risk.

Prof Ference explained: "BMI was evaluated in both observational analyses using each participants measured BMI at the time of enrolment, and in Mendelian randomisation analyses using a BMI genetic score composed of 255 variants associated with BMI at genome wide level of significance."

This score enabled assessment of lifetime exposure to BMI, on the risk of diabetes later in life, in contrast to simply looking at the effect of BMI via the snapshot measurement at study entry, Prof Ference told Medscape News UK. "Somebody could have just recently gained that weight, because people typically gain weight in middle age. And so they’ve only been exposed to that increment of higher BMI for a shorter period of time."

Study Details

Over an average follow-up of around 8 years, at which point the individuals had a mean age of 65.2 years, 31,298 individuals developed type 2 diabetes.

As expected, increasing quintiles of polygenic risk score were associated with an increased risk of developing diabetes, rising to a hazard ratio of 2.90 for people in the fifth versus the first quintile.

Overall, each standard deviation increase in polygenic risk score was associated with a hazard ratio for developing diabetes of 1.47.

When the researchers compared the impact of each kg/m2 increase in BMI on the risk of developing diabetes, they found that the results of the lifelong exposure and single-point observation analyses were very similar.

On the Mendelian randomisation analysis of BMI genetic score, the odds ratio of developing diabetes was 1.26 per kg/m2 increase in BMI, while that for the observational analyses of BMI on study entry was 1.22 per kg/m2 increase.

"This finding implies that BMI has a threshold effect on the risk of diabetes rather than a cumulative effect over time," Prof Ference said.

He told Medscape News UK: "If BMI had a cumulative effect, that is to say a little bit of increased BMI caused your beta cells in the pancreas to wear out over time, then we would expect that…the effect of lifelong exposure to one unit increased BMI in Mendelian randomisation would have a much larger effect than the effect of a one unit increase in BMI that occurs later in life.

"But we didn’t see that, so they appear to be approximately the same…And the reason why I think we can be confident in that conclusion is because when we compare that to the effects of LDL [low-density lipoprotein] cholesterol or blood pressure, it’s exactly the opposite."

He said that, for example, previous Mendelian randomisation studies have indicated that a 1 mmol/l increase in lifetime exposure to LDL doubles the risk of cardiovascular disease, while a 1 mmol/l increase in LDL acquired later in life increases the risk by only about 20%.

Finally, the researchers looked at the risk of diabetes with increasing polygenic risk score quintiles, stratified by BMI quintiles.

This showed that participants in the lowest polygenic score quintile but who were in the highest BMI quintile had a five-fold greater risk of diabetes than participants in the highest polygenic score quintile but with a low BMI.

Indeed, the risk of diabetes increased 2.5-fold when compared across polygenic risk score quintiles, but rose 11-fold when compared across BMI quintiles.

"These data suggest that BMI is a much more powerful risk factor for diabetes than polygenic predisposition," Prof Ference said.

"Therefore, these data suggest that perhaps the best role for a polygenic score for diabetes is to help us find the estimated risk of diabetes caused by BMI, rather than to serve as the primary method of estimating risk itself," he added.

Family History

Prof Taylor told Medscape News UK that the results show that the polygenic risk score is far less important than an individual’s family history, "and of course this is what you ask an individual sitting in front of you".

"If a person has a first degree family history of type 2 diabetes," he continued, "they’ve got a 30% chance of being at risk of getting it, so right away you’ve got a major factor that’s rather lost in this mish-mash of a polygenic score."

Explaining the practical application of family history, Prof Taylor added: "For instance, if there was just one person in the family with diabetes and that was a grandparent then it’s not very definitive.

"On the other hand, if mum and all her sisters got type 2 diabetes and you’re looking at the woman in front of you, then you’re looking at a woman who’s going to develop diabetes."

As for the claim of a threshold effect of BMI, which echoes Prof Taylor's 2015 paper on the personal fat threshold, he suggested that the curves for diabetes risk in the current study actually suggested the opposite, that BMI has a cumulative effect.

This, he said, is backed up by data from the Nurses' Health Study showing that each unit increase in BMI is linked to a stepwise increase in diabetes risk.

"That’s carefully collected information on individuals…whereas this is guessing at intangibles or, shall I say, calculating intangibles in order to try and derive information that might be clinically useful."

Prof Taylor said this underscores the limitations of analyses of large databases on hundreds of thousands of patients and millions of genetic variants.

He said: "When we have big data trying to determine information that might be useful for individuals, there is a nonsense afoot, because individuals have their own individual risk, especially family history and especially weight or BMI.

"We have to get back to brass tacks," he added. "We’re practical scientists, hopefully, in medicine, and we need definitive information on individuals, and I don’t see that this big data exploration is actually giving us information of sufficient precision to be able to allow decisions to be made."

The study was funded by the National Institute for Health Research and the University of Cambridge Biomedical Research Centre.

Prof Ference declares Research Grants: Merck, Novartis, Amgen, Esperion Therapeutics, Ionis Pharmaceuticals; Consulting Fees, Advisory Boards, Honoraria: Merck, Amgen, Regeneron, Sanofi, Novartis, Pfizer, Eli Lilly, Novo Nordisk, The Medicines Co, Mylan, Daiichi Sankyo, Silence Therapeutics, Ionis Pharmaceuticals, dalCOR, CiViPharma, KrKaPhamaceuticals, Medtronic, Celera, American College of Cardiology, European Atherosclerosis Society, European Society of Cardiology.

ESC 2020: Abstract Effect of BMI and Polygenic Scores on Lifetime Risk of Diabetes. Presented 31 August.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....