The relationship between body weight and mortality is
the subject of intense debate. In this study, the association between
body mass index (BMI) and overall mortality, and mortality from specific
causes was examined in women. Body weight was directly related to
all-cause mortality in middle-aged women, after smoking and pre-existing
illness were taken into account. The study found that the leanest women
had the lowest mortality, and even women of average weight had higher
There was a 'J'-shaped relationship between BMI and
overall mortality when women who smoked were included.
Women with the lowest mortality had a BMI between
19.0 and 26.9 (measured in kg/m2).
Of the deaths of obese women (BMI>29.0), 53% could
be attributed to their obesity.
When former or current smokers were excluded from the
analysis, the relationship between mortality and BMI became more direct.
Women with the lowest mortality had a BMI of < 22.
The most direct association between BMI and mortality
occurred when both cigarette smoking and disease-related weight loss
were accounted for.
These optimal analyses included only women who had
never smoked and who had stable weight in the previous 4 years (i.e. no
weight changes of 4 kg or more).
The leanest women (those with a BMI < 19) had the
lowest mortality. Mortality among obese women (those with a BMI of >
29) was over twice that for the leanest women.
The apparent excess risks associated with leanness
were found to be artefacts in this study, and were eliminated after
cigarette smoking and subclinical disease were accounted for.
Obese women were four times more likely to die due to
cardiovascular disease and five times more likely to die due to coronary
heart disease (CHD) than the leanest women.
Reported hypertension, diabetes and raised serum
cholesterol levels were two to six times more prevalent among women in
the heavier categories.
Self-reported dietary fat and cholesterol intake
varied very little in relation to BMI category.
Mortality due to cancer in obese women was twice that
in the leanest group of women.
This increased prevalence of death due to cancer was
predominantly due to colon, breast and endometrial cancers.
There was a trend towards higher mortality due to
coronary heart disease, other cardiovascular diseases and cancer, even
among women of average weight and those who were mildly overweight.
Never-smoking women with a BMI of 27.0-28.9 had a
relative risk of death due to cardiovascular disease up to 1.8 times
greater than the leanest group of women (BMI < 19) .
Even women with a BMI of 22.0-24.9 had a higher
relative risk of death due to cardiovascular disease than the leanest
group of women.
The relative risk of death due to cancer was higher
in the group whose BMI was 19.0-21.9 than in the leanest group of women.
A weight gain of 10 kg or more since 18 years of age
was associated with an increased mortality in middle adulthood.
Women who had gained 10-19 kg since 18 years of age
had a relative risk of overall mortality 1.2 times that of women whose
weight had remained stable (defined as no weight fluctuations of 4 kg or
Women who had gained 20 kg or more since 18 years of
age had a relative risk of overall mortality 1.6 times that of women
whose weight had remained stable. The waist-to-hip ratio was a strong
predictor of death due to CHD.
BMI was a stronger predictor of overall mortality
than the self-reported waist-to- hip ratio, however, women whose
waist-to-hip ratio was in the highest fifth of the cohort had a relative
risk of death due to CHD over eight times higher than that of the lowest
This 16-year study started in 1976 and followed
115195 female registered nurses aged 30-55 years. Data were collected by
questionnaires, and were controlled to eliminate bias due to
pre-existing diseases. Participants completed questionnaires every two
years, supplying age, current weight and height, current and past
cigarette smoking, other risk factors, medical history and newly
diagnosed major illnesses. In 1980, data on food frequency, physical
activity, and participants' weight at 18 years of age were also
Overall mortality was the primary end-point and,
where possible, the cause of death was determined. The mortality rate
for a specific category of BMI was calculated by dividing the number of
deaths by the cumulated number of person — years of follow-up.
Participants were grouped into seven categories of BMI as follows: <
19.0,19.0-219,220-249, 25.0-26.9,27.0-28.9,29.0-31.9 and > 32. The
relative risk for a specific BMI category was used as a measure of the
strength of association. This was calculated by dividing the mortality
rate by the rate for the leanest BMI category ( < 19) . The
proportion of deaths attributable to adiposity was calculated by
dividing the difference between the mortality rate of a specific
category and the rate for the leanest category by the rate for the