Mental health

Memory gone pear shaped?

“Pear-shaped women face increased risks of mental problems and memory loss in later life,” reported The Sun.

The story is based on a study in 8,745 postmenopausal women aimed at determining if body mass index (BMI) and hip-waist ratio (HWR), were associated with cognitive function. It found that in women with a higher BMI, those who were 'pear-shaped' (low HWR) tended to have slightly lower scores of cognitive function than their 'apple shaped' (high HWR) counterparts.

These findings do not mean that pear-shaped women are at risk of cognitive problems in later life. This type of study cannot show cause and effect as both body measurements and cognitive function were assessed at the same time. It is also important to note that all the women in this study had good cognitive function and none had dementia or cognitive impairment.

Where did the story come from?

The study was carried out by researchers from Northwestern, Wake Forest and Rush Universities, the Universities of Pittsburgh and of Iowa, the Medical College of Wisconsin and the Fred Hutchinson Cancer Research Center, all in the US. The original trial on which the study is based was funded by the US National Heart, Lung and Blood Institute.

The study was published in the peer-reviewed Journal of the American Geriatrics Society.

Most papers including The Sun , incorrectly gave the impression that being pear shaped was a risk factor for cognition problems, when the study only found an association between poorer cognitive function and pear-shaped women with a higher BMI. All of the news stories failed to mention that this cross-sectional study cannot tell us if the waist-hip ratio had any effect on cognitive function, since both were measured at the same time.

Headlines of ‘impairment’ of memory and memory ‘problems’ are also very misleading considering that all of these women were in good cognitive health, i.e. they showed no signs of dementia or memory problems.

What kind of research was this?

This was a cross-sectional analysis of data from the Women’s Health Initiative (WHI), a large cohort study investigating common causes of disease and mortality in postmenopausal women aged 50 to 79. The WHI also included a randomised controlled trial within the cohort examining the possible effects of hormone replacement therapy (HRT) on health.

The researchers point out that dementia, particularly Alzheimer’s disease, is a major public health concern, with some studies suggesting associations between Alzheimer’s disease and vascular disorders such as coronary heart disease and high blood pressure. Obesity and being overweight increase the risk of vascular disorders and so possibly the risk of dementia.

In this study, the researchers set out to explore the relationship between waist-hip ratio, BMI category and cognition score. To do this they looked at data from the women who were participating in the HRT trial. Before the trial began, the women had several measurements taken including body measurements, cognitive function and various other health and lifestyle factors. Analysing cross-sectional associations between these factors cannot demonstrate cause and effect.

What did the research involve?

The researchers used data on 8,745 healthy women aged 65 to 79 who had been enrolled in the initial WHI hormone trials cohort. The women in this study had completed questionnaires on a number of factors that might influence their health, such as previous hormone use, history or cardiovascular disease, stroke and diabetes, various lifestyle factors and factors such as age, income and education.

The women completed a 10-point validated test for cognitive function called the Modified Mini-Mental State Examination (3MSE). This measures cognitive abilities such as temporal and spatial orientation, immediate and delayed recall, verbal fluency and abstract reasoning. The tests were conducted in private by trained technicians.

The women also had their blood pressure taken, weight and height measured, BMI calculated and waist and hip measurements taken.

The researchers used standard statistical methods to determine any association between the scores from the 3MSE test and BMI and waist circumference. They also looked for any correlations between the cognitive score and other possible confounders such as age, education, blood pressure and smoking history.

What were the basic results?

The researchers found that a large proportion of the women (over 70%) were classified as overweight or obese. Overall, cognitive scores decreased slightly as BMI increased. Women who reported having a stroke, heart disease, diabetes or high blood pressure also had lower cognitive scores. The researchers adjusted their findings to take account of these factors, as well as age and education.

They found that for every one-unit increase in BMI, the cognitive score decreased by 0.988 points. BMI had the most pronounced association with poorer cognitive functioning in women with smaller waist measurements and waist-hip ratio. In the quartile of women with the highest waist-hip ratio (the top 25%), cognitive scores increased as BMI increased.

How did the researchers interpret the results?

The researchers say the relationship between cognition and BMI is complex and modified by the women’s waist-hip ratio. The findings suggest that excess fat distributed around the middle, as estimated by a higher waist-hip ratio, is associated with higher cognitive function scores in older women, while excess fat around the hips is associated with poorer cognitive function. They suggest that oestrogen levels may play a role in this complex relationship.


This is a cross-sectional analysis of data from a large group of postmenopausal women who were about to participate in the Women’s Health Initiative trials of hormone replacement therapy.

This study’s strengths include its size (8,745 women), and its accurate collection of a large amount of medical and health information from the women. It also considered the numerous confounders that could affect the observed relationship between BMI and waist-hip measurements and cognitive scores.

The analysis demonstrated a complex relationship between obesity, body measurements and cognitive function, with cognitive scores appearing to decrease as BMI increased, and a higher waist-hip ratio (indicating higher central fat mass) appearing to have a ‘protective’ effect. However, it cannot prove that body shape can affect cognitive ability in this way due to several limitations:

  • Body measurements, BMI and cognitive function were all assessed at the same time. It is unclear whether one caused the other, or whether some unmeasured factor is affecting both.
  • All of the women in this study, whatever their BMI and waist-hip ratio, were found to be at a high level of cognitive functioning and had no cognitive impairment. Therefore, any associations in the news between body shape and memory problems or memory impairment, are misleading. Also, the results should not be extrapolated to mean that women of a certain body shape have a higher or lower risk of dementia or Alzheimer’s disease.
  • The cognitive testing was limited to only one measure. The test used was a measure of global cognitive functioning and did not extensively assess specific areas of cognition. In addition, it is unclear whether the small 0.988 point change with every one unit increase in BMI would have any actual implications for women in their normal day-to-day functioning.
  • The results should not be interpreted to mean that obesity, and in particular central obesity (the apple shape), is protective to health. Obesity is a risk factor for many health problems, including high blood pressure and cholesterol, heart disease and diabetes, and central obesity has been associated with greater risk of these diseases.
  • The analysis only included postmenopausal white women, so it may not be possible to generalise the results to other ethnic groups.

Further research into the relationship between obesity, fat distribution and cognitive function is needed.

NHS Attribution