December 20, 2019
A large prospective study of more than 1 million women in the United Kingdom showed that midlife obesity was associated with a 21% increased risk for all-cause dementia at 20-year follow-up, whereas diet and exercise were not.
"Obesity in midlife is a risk factor for dementia in late life. Physical inactivity and poor diet may not be," lead author Sarah Floud, PhD, senior epidemiologist in the Cancer Epidemiology Unit, Nuffield Department of Population Health, Oxford University, United Kingdom, told Medscape Medical News.
The study was published online December 18 in Neurology.
Cause or Consequence?
Although diet and inactivity were linked to greater dementia risk in the first 5 to 10 years of the study, there were no significant associations over the longer term.
Some previous studies suggest poor diet or a lack of exercise may increase dementia risk. However, Floud noted that this current study of 1.1 million women shows the link between these factors may reflect behavioral changes that manifest as poor dietary habits and physical inactivity in the prodromal stage of dementia.
There may be some "reverse causality" in play, she said. Individuals who develop dementia tend to experience a preclinical phase that lasts up to a decade. During this time, many become less physically active, consume fewer calories, and lose weight. This could have altered the baseline measures and affected outcomes in the first decade of the study.
In other words, reverse causality could be the reason that a significant association with obesity emerged later and why short-term associations with low body mass index (BMI), low caloric intake, and inactivity fizzled out over time.
"Because we have long follow-up and a large number of cases, we can also show that in the long term ? when most of the women developed dementia ? the association between midlife inactivity and dementia was weak and close to nonexistent," Floud said.
The aim of the study was to "determine whether midlife obesity is a cause of dementia and whether low BMI, low caloric intake (in women who had not recently changed their diet because of disease), and physical activity are causes or merely consequences of the gradual onset of dementia," the investigators write. They analyzed data from the Million Women Study.
This observational study included 1,136,846 UK women born from 1935 to 1950 and who were enrolled from 1996 to 2001. The average age was 56 years at study entry and 74 years at last follow-up.
Assessments were completed at regular intervals after enrollment, from <5 years, 5 – 9 years, 10 – 14 years, or ≥15 years. Dementia diagnosis was based on first ICD-10 code mentioning dementia in hospital records.
Obesity was defined as a BMI of ≥30 kg/m2. During a mean of 18 years' follow-up, 30,957 women developed dementia.
Result of Vascular Disease?
After 15 years or more, women who were obese at baseline had a relative dementia risk of 1.21 (95% confidence interval [CI], 1.16 – 1.26; P < .0001) vs women with a BMI of 20 – 24 kg/m2 at study entry.
In contrast, during the first 5 years of the study, dementia risk related to obesity was lower (relative risk [RR], 0.65; 95% CI, 0.49 – 0.85).
An opposite trend was observed with respect to low caloric intake. Women in this group had an elevated risk for dementia in the first 5 years (RR, 1.75; 95% CI, 1.42 – 2.14; P < .0001). However, the relative risk "decreased sharply with duration of follow-up," the researchers report. At ≥15 years, there was no significant risk (RR, 1.03; 95% CI, 0.94 – 1.13).
The investigators observed a similar pattern regarding inactivity. In the first 5 years, physically inactive women were almost 60% more likely to develop dementia (RR, 1.59; 95% CI, 1.31 – 1.92; P < .0001) compared to their counterparts who were of normal weight. Again, however, the association approached a null value after ≥15 or more years (RR, 1.05; 95% CI, 1.02 – 1.08; P = .004).
Interestingly, the investigators found that baseline obesity was strongly associated with vascular dementia (RR, 1.41; 95% CI, 1.29 – 1.56; P < .0001) at the ≥15-year follow-up. However, they found no such association regarding Alzheimer disease (RR, 1.01; 95% CI, 0.93 – 1.08; P = .87).
"The short-term associations of dementia with physical inactivity, low caloric intake, and low BMI are likely to be mainly due to preclinical dementia affecting behavior and weight loss," the researchers note.
"In this population, midlife obesity is the only factor examined that is likely to be causally related to dementia, perhaps chiefly through its effects on vascular disease," they write.
"Further work is needed on how obesity increases the risk of dementia," Floud said. Future research plans include exploring whether a measure of central obesity such as waist-to-height ratio would yield the same results.
In addition, the investigators want to expand the research population to determine whether the risk for younger women who are obese is similar to that of middle-aged women.
WHO Guidelines Challenged
In an accompanying editorial, May A. Beydoun, PhD, and Mika Kivimaki, FMedSci, note that the findings confirm the hypothesis that the short-term effects of preclinical dementia are associated with weight loss, physical inactivity, and poor eating habits and that the long-term effect of midlife obesity on dementia risk is likely "more causal and amenable to prevention."
The null findings on physical inactivity and poor dietary habits are noteworthy because they challenge the 2019 guidelines by the World Health Organization (WHO), they add.
To reduce the likelihood of cognitive decline or dementia, the WHO recommends physical activity and a Mediterranean-like diet for adults with normal cognition or mild cognitive impairment.
However, they note that the evidence supporting the WHO recommendations is "uncertain" because none of the studies to date has been large enough or was conducted over a long enough period to show that physical activity and a healthy diet prevent dementia.
The Million Women Study is the "largest observational investigation on this topic to date, and the first with sufficient statistical power to stratify the follow-up into 5-year periods to address reverse causation," they add.
Despite the study's "notable strengths," its findings should be interpreted in light of several limitations, the editorialists note. These include a lack of generalizability to men, not tracking changes in BMI over time, and relying on self-report with respect to diet.
"More long-term studies are needed to examine time-dependent changes in weight status, physical activity, diet, and their bidirectional associations with dementia and its preclinical phases, with further inclusion of neuroimaging and dementia markers," they write.
The study was supported by funding from the UK Medical Research Council and Cancer Research UK. The editorial was supported by the Intramural Research Program of the National Institues of Health, the National Institute on Aging, the UK Medical Research Council, and the Academy of Finland. Floud, Beydoun, and Kivimaki have disclosed no relevant financial relationships.