Individuals with body mass index values below 20 in middle age were significantly more likely than those of normal weight to develop dementia later on, a retrospective study of nearly 2 million people indicated.
Researchers in Great Britain found that those who were underweight had a higher risk of dementia than people with a normal weight (rate ratio 1.34, 95% CI 1.29-1.38; P<0.0001), with an average follow-up time of about 9 years.
And the incidence of dementia fell for each increasing BMI category, the researchers found, with very obese people having the lowest risk (RR 0.71, 95% CI 0.64-0.78;P<0.0001), according to the report published April 10 in the The Lancet Diabetes & Endocrinology.
“If increased weight in mid-life is protective against dementia, the reasons for this inverse association are unclear at present,” wrote the authors, who were led by Nawab Qizilbash, MSc, DPhil, an epidemiology lecturer at the London School of Hygiene and Tropical Medicine. “Many different issues related to diet, exercise, frailty, genetic factors, and weight change could play a part.”
In an accompanying comment to the study, Deborah Gustafson, PhD, from SUNY Downstate Medical Center in New York, wrote that the findings by Qizilbash and colleagues are not the “final word” on the subject and should be interpreted cautiously.
“The published literature about BMI and dementia is equivocal,” she wrote. “To understand the association between BMI and late-onset dementia should sober us as to the complexity of identifying risk and protective factors for dementia.”
The data were gathered from the United Kingdom Clinical Practice Research Datalink (CPRD), which included BMI records from people age 40 or older between 1992 and 2007. The median age of the participants was 55 at baseline; data collection began in 1987 and continued to July 2013. Fifty-five percent were female.
Underweight was defined as having a BMI of <20, healthy weight as 20-24.9, overweight as 25-29.9, and obese as ≥30. Those who had a record of previous dementia were excluded, as were those with fewer than 12 months of data before being indexed.
Follow-up continued until the participant's final data record, transfer out of practice, death, or the first record of dementia. More than 45,000 people had a first diagnosis of dementia -- an overall rate of 2.4 cases per 1,000 person-years, noted the authors.
When researchers controlled for potentially confounding factors like smoking and alcohol status, diabetes, or cardiovascular disease history, the results remained largely the same.
The findings contradicted earlier, smaller studies which have shown that being overweight in mid-life is associated with an increased risk of dementia later in life. But studies have also shown that being overweight in late-life can reduce dementia risk -- a situation that some have called the "obesity paradox."
Qizilbash and colleagues found that at 80 years of age, the incidence of dementia was 9.9% for both sexes (95% CI 9.5%-10.3%) for underweight people, 6.5% (95% CI 6.4%-6.6%) for those with a normal weight, 5.2% (95% CI 5.0%-5.3%) for overweight people, and 4.9% (95% CI 4.7%-5.0%) for obese people for both sexes. Beyond age 80, the gradient for risk of dementia was steeper for women than it is for men.
Researchers also looked at whether their results were affected by the competing risk of all-cause mortality, which has a J-shape associated with BMI, with the greatest risks at the extremes. "The rate ratios of dementia for obese versus healthy weight people became somewhat attenuated, but a more than 20% lower dementia risk persisted at all degrees of obesity," they wrote.
Limitations of the study included a potential for selection bias, because nearly half of the eligible people were not included in the study either because they did not have a BMI record or they didn't have at least 1 year of previous health records. In addition, relatively few variables were available for leaving a possibility of residual confounding.
Some factors like socioeconomic status, physical activity level, and ethnic origin might have changed the relationship between BMI and dementia, the authors noted.
Gustafson noted also that there's no consistent definition of "mid-life," and what counts as mid-life versus late-life is unclear. Additionally, details about dementia sub-types and validity of dementia diagnosis in the CPRD were unavailable, and the data for the age of dementia diagnosis was unclear. (Originally printed in MedPage Today)
