. BMI and risk of dementia in two million people over two decades: a retrospective cohort study. Lancet Diabetes Endocrinol. 2015 Jun;3(6):431-6. Epub 2015 Apr 9 PubMed.


Please login to recommend the paper.


  1. The increased dementia risk associated with underweight in midlife fits well with our own unpublished results, which have been puzzled about. Our intention has been to publish a separate paper on this, where we include BMI in early midlife, late midlife, and old age, and its association with dementia. There are few such papers and it would be interesting to see if the results are replicated.

  2. It is becoming more appreciated that damage to microvessels and the blood-brain barrier may kick-start Alzheimer’s disease (AD) progression. In fact, ailments affecting the neurovasculature, such as cardiovascular disease, hypertension, stroke, and diabetes, are established risk factors for AD and other neurodegenerative diseases. These vascular risk factors are more common in obese individuals. Qizilbash et al. studied almost two million individuals to determine if body mass index (BMI) alters the chance of a person developing dementia. They found that underweight individuals have a higher risk and obese people have a lower risk of developing dementia.

    So what does this mean? BMI is not the ideal anthropometric measure to use in these studies because it gives no indication of metabolic health. Unlike BMI, waist-to-hip ratio is a better indicator of the amount of visceral fat an individual has. Visceral fat around organs is negatively associated with adverse metabolic health and can go undetected by BMI. There are many obese individuals who are metabolically healthy and underweight individuals who are not. With that said, it is important to note that subcutaneous body fat can function as an energy reserve. With aging, there is a loss of appetite due to a reduction in sensory receptors and less absorption of nutrients in the gut. Senescence leads to a loss of macronutrients (energy and protein) and micronutrients (vitamins B6, B12, and folic acid). These multiple nutrient deficiencies can be the result of physiological, social, and economic factors which Qizilbash et al. did not assess. Furthermore, dietary fat determines the lipid composition of the brain. Therefore, it is possible that the data here could be parsing out individuals based on their nutrition and metabolism that could be a consequence of many factors. Future large-scale studies that determine links between metabolic health (by measuring visceral fat, ideally by DEXA scan) and dementia risk are warranted and of great importance.

    The data in the manuscript were adjusted using sophisticated statistical modeling that is difficult to follow. No raw data is shown for BMI and dementia risk. Instead, all data used to assess BMI impact on dementia risk was adjusted for age and gender. It is well established that age is the most common risk factor for AD and women are more likely to develop the disease. The authors chose to adjust their data to eliminate these effects, presumptively to have a clear picture of body weight on cognitive impairment without known confounds. However, the way in which they adjust their data for age is unusual. They account for “age at risk” but it is ambiguous what this means aside from it being a “more meaningful time point.” The authors should define what made this time point meaningful. It is possible their data would look different had they not made this adjustment.

    They next go on to modify their data for age, sex, smoking status, alcohol status, diabetes, previous myocardial infarction, statin use, and anti-hypertensive use and find the same dementia risk for BMI. When accounting for all of the above and mortality rate, Qizilbash et al. “did a sensitivity analysis to account for the hypothesis that those who died would have had double the risk of dementia if they had survived.” The authors should scientifically justify this hypothesis. Furthermore, the definition of dementia was quite broad in scope, encompassing the diagnosis of all dementia including Alzheimer’s, Lewy body disease, or Pick’s disease, and any listed on death certificates.

    Overall, these findings may give individuals the false illusion that an unhealthy lifestyle is beneficial in preventing dementia. However, this is not the case. Unfortunately, important measurements like physical activity, blood pressure, and blood lipid values were not accounted for in this study. Good nutrition and exercise are critical for proper heart and brain health. Thus, more comprehensive studies carefully examining cognition/dementia, nutrition, metabolism, physical activity and lifestyle are needed. 

  3. In the past decade, many population-based studies have evaluated the relationship between body-mass index (BMI) and dementia. Two studies from our group have shown that obesity and even being overweight at midlife are associated with increased risk of dementia later in life (Kivipelto et al., 2005; Xu et al., 2011). In both studies, BMI was assessed in people age 40 to 60. A meta-analysis on this topic confirmed these findings (Anstey et al., 2011). However, several studies, including ours, report an association between lower BMI and prevalent dementia and a protective effect of high BMI within five years after dementia onset (Atti et al., 2008; Luchsinger and Gustafson, 2009), and low BMI and weight loss have been suggested as preclinical markers of dementia (Gustafson, 2008). Therefore, we need to pay special attention to the time windows of BMI assessment and age of dementia onset when examining their association.

    Unfortunately this issue, i.e., age at exposure, has not been properly taken into account by Qizilbash and colleagues. In their recent report they claimed that their “results contradict the hypothesis that obesity in middle age could increase the risk of dementia in old age.” BMI in Qizilbash's study was assessed in people age 40 or older (range 40 through 80-plus). Thus, age at BMI assessment included both midlife (40 to 60) and later life (60 through 80-plus). As a result, the BMI-dementia association reported in this paper reflects a mixture of midlife and later-life effects of BMI on dementia risk. In addition, other methodological limitations, such as the way the authors identified the dementia-free population and the questionable dementia ascertainment and dementia ascertainment, may affect the results.

    To better understand this paper’s contribution to the current knowledge concerning obesity and cognitive aging, it would be relevant to perform an analysis stratified by age at exposure to obesity and to estimate the extent to which cohort definition and selection bias affect the results. 


    . Obesity and vascular risk factors at midlife and the risk of dementia and Alzheimer disease. Arch Neurol. 2005 Oct;62(10):1556-60. PubMed.

    . Midlife overweight and obesity increase late-life dementia risk: a population-based twin study. Neurology. 2011 May 3;76(18):1568-74. PubMed.

    . Body mass index in midlife and late-life as a risk factor for dementia: a meta-analysis of prospective studies. Obes Rev. 2011 May;12(5):e426-37. PubMed.

    . Late-life body mass index and dementia incidence: nine-year follow-up data from the Kungsholmen Project. J Am Geriatr Soc. 2008 Jan 1;56(1):111-6. PubMed.

    . Adiposity, type 2 diabetes, and Alzheimer's disease. J Alzheimers Dis. 2009;16(4):693-704. PubMed.

    . A life course of adiposity and dementia. Eur J Pharmacol. 2008 May 6;585(1):163-75. Epub 2008 Mar 4 PubMed.

  4. You seem to have overlooked a paper published by myself and a colleague (Wotton and Goldacre, 2014) on obesity and subsequent risk of dementia. 

    This was also a very large study of people in England, although our data are from hospital episode statistics (HES) data.  We found people admitted to hospital with obesity whilst in their 30s conveyed the greatest risk of future dementia, and that this risk gradually declined with increasing age.  In people admitted to hospital with obesity in their 70s, there was no association with dementia, and for those older, the risk became reduced.


    . Age at obesity and association with subsequent dementia: record linkage study. Postgrad Med J. 2014 Oct;90(1068):547-51. Epub 2014 Aug 20 PubMed.

    View all comments by Clare Wotton

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Pound-for-Pound Protection: Does Obesity Guard Against Dementia?