select search filters
roundups & rapid reactions
factsheets & briefing notes
before the headlines
Fiona fox's blog

expert reaction to dementia and body mass index

A study published in The Lancet Diabetes & Endocrinology journal has investigated a link between BMI and risk of dementia in a group of people in the UK aged over 40. The authors report an inverse correlation between BMI in middle age and dementia risk, contrary to previous suggestions.


Dr Liz Coulthard, Consultant Senior Lecturer in Dementia Neurology, University of Bristol, said:

“This is a well-designed study that contradicts previous smaller studies and demonstrates that the relationship between weight and dementia risk is not straightforward. Many dietary, environmental or genetic factors could link weight and dementia and further work should explore these so that lifestyle advice can be given to people who want to limit their risk of dementia.

“This study does not tell us that gaining weight will help to reduce the risk of dementia. Further work is required to establish any potential benefit of weight gain and, on the basis of this work, one might predict that underweight people only would benefit.

“We do know that obesity carries many other risks including high blood pressure, heart disease, diabetes and increased rates of some types of cancer. So maintaining a healthy weight is recommended.”


Dr Simon Ridley, Head of Research, Alzheimer’s Research UK, said:

“These new findings are interesting as they appear to contradict previous studies linking obesity to dementia risk. The results raise questions about the links between weight and dementia risk, clearly further research is needed to understand this fully. There is currently no sure-fire way to prevent dementia, and investment in research is vital if we are to understand how to stave off the condition. In the meantime, evidence suggests that we can reduce the risk by eating a healthy, balanced diet, exercising regularly, not smoking, and keeping blood pressure in check.

“This study doesn’t tell us that being underweight causes dementia, or that being overweight will prevent the condition. Many other studies have shown an association between obesity and an increased risk of dementia. These findings demonstrate the complexity of research into risk factors for dementia and it is important to note that BMI is a crude measure – not necessarily an indicator of health. It’s also not clear whether other factors could have affected these results.”


Prof. Tom Dening, Professor of Dementia Research, University of Nottingham, said:

“Is it time to slump on the sofa, pile into the burgers and slurp the lager? Probably not just yet!

“This is an interesting and provocative paper that will take some time to absorb (a bit like the calories, I suppose). Like many studies that show associations between possible risks and medical outcomes like dementia, it is likely to pose more questions than it answers. Are there any problems with the sample or the methods that may have led to a spurious or incorrect association? The authors mention that, of a potential sample of over 6 million people, many did not have BMI measurements or adequate clinical data, so the sample analysed was only 1/3 of what it might been. The people for whom there was not adequate data may have been at higher risk of dementia for various reasons, e.g. infrequent GP attenders or people with higher social or medical risks.

“The authors have adjusted for medication and various medical conditions, but in population based studies of this kind there are often questions about whether investigators have over or under-adjusted for possible confounders.

“They have also adjusted for death using a particular method, which they do describe. And mortality is likely to be higher in people at both ends of the BMI distribution. However, how can we know whether the amount of adjustment for death is ‘fair’?

“Furthermore, the ascertainment of dementia is based on a GP diagnosis of dementia. We know from government campaigns that the dementia diagnosis rate has been quite low and was probably <50% in the years that data were being collected. So, if there were factors biasing your chances of getting a dementia diagnosis that were affected by your body mass, then this would affect the results of this study. It is plausible that faced with a grossly obese patient, the average GP may have concentrated on the obvious medical risks and paid less attention to the cognitive issues than they might have done with a comparable patient of normal weight. Or maybe the GP didn’t refer them to a memory clinic but sent them to see a gastroenterologist or diabetologist instead.

“In other words, the study may be telling us something about the diagnostic practices of GPs, as much about the natural history of metabolic disease and dementia.”


Dr Doug Brown, Director of Research and Development at Alzheimer’s Society, said:

‘We don’t yet know enough about the link between body weight and dementia. Previous research has suggested that being overweight in midlife increases risk of developing the condition and yet this study suggests that it may actually be protective. This study of almost 2 million people also reports that being underweight in later years could increase risk of developing the condition by a third.

“This mixed picture highlights the difficulty of conducting studies into the complex lifestyle risk factors for dementia and reinforces the need for further research so we can identify the most important risk factors. While the evidence on body weight and dementia is unclear, we know that people can make positive lifestyle choices to keep their brains healthy by taking regular exercise, not smoking and following a healthy balanced diet.’


‘BMI and risk of dementia in two million people over two decades: a retrospective cohort study’ by Qizilbash et al. published in The Lancet Diabetes & Endocrinology on Friday 10th April. 


Declared interests

Dr Liz Coulthard: None declared

Dr Simon Ridley: None declared

Prof. Tom Dening: I have no conflicts of interest to declare

Dr Doug Brown: None declared

in this section

filter RoundUps by year

search by tag