A new, long-term study finds that midlife obesity raises the risk of dementia in women. However, calorie intake and physical inactivity do not.
Sarah Floud, Ph.D., of the Nuffield Department of Population Health at the University of Oxford in the United Kingdom, is the lead author of the study.
As Floud and her colleagues explain in their paper, some previous studies have found an association between a low body mass index (BMI) and the likelihood of receiving a diagnosis of dementia within the next 5–10 years.
Other studies that lasted a decade or less have also linked poor diet and lack of exercise with the incidence of dementia.
However, all of the above may be the result of reverse causality, meaning that they may be consequences, rather than causes, of dementia. This situation could well be possible, explain the authors, because dementia typically affects cognition a decade before the person formally receives a diagnosis.
During this preclinical stage, the condition can slowly but gradually affect behavior, impair mental and physical activity, reduce the intake of food and calories, and cause weight loss.
Furthermore, explain the authors, some recent meta-analyses have pointed out that although in the short term, a low BMI may be associated with dementia as a result of reverse causality, over a longer period, obesity is positively associated with dementia.
Either way, prospective studies over longer periods are necessary to settle the matter of how BMI connects to dementia risk. Floud and her team set out to do exactly this.
Their findings appear in the journal Neurology.
The team examined 1,136,846 women in the U.K. They had an average age of 56 years and were free of dementia at the start of the study, between 1996 and 2001.
The women gave information about their height, weight, calorie intake, and physical activity, and the researchers clinically followed them until 2017 through the National Health Service records. These records also noted any hospital admissions for dementia.
For their study, the scientists considered a BMI of 20–24.9 as “desirable,” 25–29.9 as overweight, and 30 and over as obese. They classified women who exercised less than once a week as inactive and those who exercised at least once weekly as active.
Using Cox regression models, the team calculated the links between BMI and dementia incidence over the follow-up period, adjusting for age, height, education, smoking, alcohol intake, use of menopausal hormones, residential area, and area deprivation.
Over the study period, 89% of the participants had no mention of dementia in their health records. At 15 years after the start of the study, 18,695 women had received a dementia diagnosis.
Women who had obesity at the beginning of the study were 21% more likely to develop dementia than women who had a “desirable” BMI.
More specifically, 2.2% of the women with obesity went on to develop dementia in the long term, compared with 1.7% of those with a healthy BMI.
Although the findings revealed that low calorie intake and a lack of physical activity had a link with higher dementia risk in the first decade of the study, these associations gradually faded after that period, and neither calorie intake nor inactivity had a significant association with dementia risk.
Floud comments on the findings, saying, “Some previous studies have suggested poor diet or a lack of exercise may increase a person’s risk of dementia.”
“However, our study found these factors are not linked to the long-term risk of dementia. […] The short-term links between dementia, inactivity, and low calorie intake are likely to be the result of the earliest signs of the disease, before symptoms start to show,” she emphasizes.
“On the other hand, obesity in midlife was linked with dementia 15 or more years later. Obesity is a well-established risk factor for cerebrovascular disease. Cerebrovascular disease contributes to dementia later in life.”
Sarah Floud, Ph.D.
The study is limited by the fact that it only involved women, which means that the findings may not apply to men.
The authors of a linked editorial also mention “the absence of time-dependent dynamic analyses of BMI […], crude measurement of dietary habits, and residual confounding” as study limitations.