Thursday May 8 2008
Underweight and obese people were at higher risk
“Obesity doubles the risk of Alzheimer’s” is the headline in the Daily Express. It and other news sources report new research which found that people who are obese have an increased risk of all types of dementia. Being underweight “increased the risk of any kind of dementia by 36 per cent, while being obese increased it by 42 per cent. For Alzheimer’s disease, being obese increased the risk by 80 per cent“, the newspaper says.
The study that the newspaper story is based on has some limitations, as it combines the results of 10 studies that were of variable quality and characteristics. The increases in risk of dementia from obesity were not statistically significant, and the 80% increase in risk of Alzheimer’s was only of borderline significance, which means that it may still be a chance finding.
Alzheimer's disease is a form of dementia with characteristic clinical features and changes to the brain that can only be seen at autopsy and, currently, the causes are unknown. While age and heredity are the most established risk factors, those of overweight and obesity remain uncertain. It seems plausible that being obese may be associated with an increased risk of dementia, in particular vascular dementia, as obesity frequently occurs in conjunction with raised blood pressure, high cholesterol and possibly smoking, all of which increase the risk of damage to the blood vessels in the body. However, further research will be needed to clearly establish any link.
Where did the story come from?
Dr May A Beydoun and colleagues at Johns Hopkins University and the University of Iowa carried out this research. No sources of funding for this research were reported. It was published in the peer-reviewed publication by The International Association for the Study of Obesity: Obesity Reviews.
What kind of scientific study was this?
This was a systematic review with meta-analysis in which the authors aimed to clarify obesity’s influence upon dementia by combining the findings from different studies that have looked at the association between body mass index (BMI) and other measures of body fat and different forms of dementia in later life.
The researchers conducted a search of a medical database (PubMed) for all English language articles published between January 1995 and June 2007 which included the keywords “dementia” and “obesity”. All studies had to:
- be prospective cohort studies (following up a group of people over a period of time);
- have an initial cohort size of 100 or more people;
- include only participants aged at least 40 at the start of the study;
- have some measure of BMI or obesity/overweight;
- follow up participants for at least two years;
- look at any dementia outcomes (Alzheimer’s disease or vascular dementia); and
- include a statistical measure of risk that the researchers could use in their analyses.
The researchers then pooled results from the studies to see how obesity, overweight, BMI or any weight change affected the risk of the different forms of dementia. They considered factors that could have differed between the studies and therefore would have an effect when combining results, such as sex, age group, length of follow up, and other lifestyle factors and co-morbidities that the participants may have had. They also considered the effect of publication bias, which means that they looked for evidence that studies showing non-significant results may have not been published.
What were the results of the study?
researchers identified 10 studies (four from the US, others from Sweden, Finland, Japan and France) and were able to combine the results of seven in a meta-analysis. The studies had variable methods and inclusion criteria: two included men only; four included participants aged in their 40s at the start; while in six, participants were over 65. The studies had different methods of screening for and diagnosing dementia and used different main exposures of interest (BMI, weight change, underweight, overweight or obesity). They differed in length of follow up from between 30 months and 36 years; and in sample size from 382 to 10,136. Overall, combining all the studies, they had 1,007,911 person-years of follow up.
The researchers carried out meta-analyses of association between any type of dementia and being of normal weight compared with being underweight, overweight or obese (for men and women combined and after adjustment for confounding factors of lifestyle, co-morbidity, genetic influence and socioeconomic features). The increase in risk of dementia from being obese (the 42% quoted by the newspapers) and being overweight were both not statistically significant. Only the 36% increase in risk of dementia from being underweight was found to be significant.
The researchers then combined the results from the four studies looking specifically at the link between obesity and Alzheimer’s, and the three studies looking at the link between obesity and vascular dementia (again, for men and women combined and following adjustment for confounding factors). The risk of Alzheimer’s from obesity was increased by 80% but this was only just significant (95% confidence interval [CI] 1.00 to 3.29; if the smaller figure here had been any less than 1.00 the results would have been non-significant). The 73% increase in risk of vascular dementia was non-significant (95% CI 0.47 to 6.31).
However, the risks from obesity upon dementia, Alzheimer’s and vascular dementia all became significantly increased when the researchers carried out analyses looking only at studies that had followed people up for 10 years or more (two studies for dementia, two for Alzheimer’s and one for vascular dementia) and those that had only included people aged less than 60 at the start of the study (one study for each of the dementia outcomes).
The researchers describe the overall relationship between BMI and dementia risk as being U-shaped, with higher risks for those underweight and overweight-obese compared with those in the middle with normal weight.
What interpretations did the researchers draw from these results?
The researchers conclude that their meta-analysis shows a “moderate association between obesity and risks for dementia and Alzheimer’s disease”. They say that further studies are needed to understand the possible biological mechanisms for this and to determine what would be considered an optimal weight.
What does the NHS Knowledge Service make of this study?
This carefully conducted meta-analysis has examined a number of studies in order to evaluate the links between weight and the risk of dementia. However, the results should be interpreted with care and their significance has been slightly overemphasised by the newspapers.
- There was only a significantly increased risk of dementia in people who were underweight (the 36% increased risk reported by the papers); the 42% increase in risk associated with obesity was a non-significant result. The increase in risk of Alzheimer’s in obese people was 80% as reported by the newspapers, but this was only of borderline significance and there remains the possibility that it may be a chance finding.
- The literature search included studies from only one medical database, only those published over a 12-year period and only those which used the keywords “dementia” and “obesity”. Other relevant studies may have been missed by the search.
- The individual studies had very variable methods, inclusion criteria, outcomes that they were looking at and the possible confounding factors that were considered. It is difficult to tell at what time intervals measures of weight were taken and whether it was established that all of the participants were free from dementia at the start of the included studies. The studies had also used different statistical measures of risk of dementia. These things all add some error when evaluating individual studies and combining results in a review.
Alzheimer's disease is a form of dementia with characteristic clinical features and changes to the brain that can only be seen at autopsy and, currently, the causes are unknown. While age and heredity are the most established risk factors, those of overweight and obesity remain uncertain. It seems plausible that being obese may be associated with an increased risk of dementia, in particular vascular dementia, as obesity frequently occurs in conjunction with raised blood pressure, high cholesterol and possibly smoking, all of which increase the risk of damage to the blood vessels in the body. However further research will be needed to clearly establish any links.
Sir Muir Gray adds...
Another reason to remember to walk those extra 3,000 steps a day.