Some people don't
know when to quit. You would think that after the debacle over the
grossly inflated estimates of so-called obesity-related deaths from
the US Center for Disease Control that the fat police would have the
decency to just shut up. But the scary junk science stories about
killer fat just keep coming. The latest is an alarmist study in the
New England Journal of Medicine titled "Children's Life Expectancy
Being Cut Short by Obesity."
The study, by a team led by Jay Olshansky
of the University of
Illinois, makes the astonishing and
quite unsupported claim that for the first time since at least 1900
children born in the United States today will
live shorter lives than their parents due to obesity-caused
mortality. The study's only problems are:
1)
It flies in
the face of the standards of contemporary evidence-based medical
research since the "study" is not based on any scientific evidence
but simply on a "collective judgment" about disease and mortality
trends.
2)
It is
contradicted by a mass of real scientific evidence from the last
fifty years about the relationship between obesity and
death.
To begin with, it is hard to describe the
Olshansky study as science at all since there is no empirical
evidence, other than "trends" cited in support of the author's
claims. As the authors note, "from our analysis of the effect of
obesity on longevity, we conclude that the steady rise in life
expectancy during the past two centuries may soon come to an end."
According to Olshansky and his co-researchers, Americans would live
longer if everyone "who is currently obese were to lose enough
weight to obtain an 'optimal' BMI, which we defined as a BMI of
24."
The problem with this claim is that it is
contradicted by a significant number of empirical studies of the
connection between weight and premature mortality, including one by
David Allison, one of Olshansky's co-authors. According to Allison's
1999 study in JAMA "Annual Deaths Attributable to Obesity in the
United
States", which claims that obesity
causes about 300,000 premature deaths a year, individuals with BMI's
of 20 ("normal") have the same risk of early death as those with
BMI's of 30 ("obese"). Both of these groups have higher risks of
dying than those who are overweight with a BMI of 25. Again, the
1995 Nurses Study found that those nurses with BMI's of 19-24.0 had
almost identical mortality rates as nurses with BMI's of 25-31.9. So
how can an "ideal" BMI of 24 possibly lead to living longer?
Second, the US National Center for Health
Statistics, in its most recent US Life Tables (2000), reports that
overall life expectancy at birth was 76.9 compared with 47 in 1900.
Overall life expectancy increased by more than six years from 1970
to 2000, despite the claim that for the last ten years there has
been an unprecedented obesity epidemic in which weights have
increased by up to 75%.
Again, the author's worry about the
connection of childhood obesity with such diseases as Type 2
diabetes is not supported by the scientific evidence. With the change in the
definition of diabetes (from fasting blood sugar of 140-126)
millions of new diabetics were minted overnight, just as millions of
new overweight and obese were created with the "revised" BMI
categories.
But even with this change there is little
evidence of a substantial increase in blood sugar levels across the
population. Despite the claims of an obesity-fueled diabetes
epidemic, CDC figures show that during the 1990's the disease
increased only from 8.2 to 8.6%, a statistically nonsignificant
increase. Finally, a number of recent studies have shown that
dieters have a higher risk of the disease than others. Studies have
also reported that the most effective strategy for avoiding Type 2
diabetes is not through weight loss by through changes in physical
activity and diet.
But what really contradicts the claim
that today's children will live shorter lives than their parents due
to obesity is the ever-growing number of serious scientific studies
that over the last 50 years that have looked at the link between
obesity and premature death. These studies refute Olshansky's claim
of a reduction in mortality if everyone who is currently obese were
to reduce to a BMI of 24.
Starting in 1950's in the midst of
another alleged obesity epidemic that had the medical community
declaring obesity to be the number one threat to the nation's
health, there has been a steady stream of studies that have shown
that being overweight is not a cause of increased mortality. The
Framingham study of some 5,000
people over 30 years found, for instance, that the worst life
expectancies were for the thinnest men. According to Reubin Andres
of the US National Institute on Aging, none of the scientific
studies on weight and mortality supports the claim that excessive
weights lead to increased mortality. As he noted "The major studies
of obesity and mortality fail to show that overall obesity leads to
greater risk." This conclusion is echoed in the Seven Country Study
by Ancel Keys. In a recent re-examination of Keys data, researchers
found that being overweight -- a BMI of 25-30 -- had no effect on
mortality and even the obese still had lower death rates than thin
men.
Another longitudinal study involving
almost two million Norwegians followed for ten years found those
with BMI's from 26-28 ("overweight") had the highest life
expectancy. As the
journal Obesity Research noted in 2002, "Compared with normal
weight, overweight and obesity did not significantly increase
all-cause mortality risk."
All of this suggests that the Journal's
latest study is more likely the stuff of science fiction rather than
science. Despite the alarmist cries of declining life expectancies,
the scientific evidence of the last half century does not support
the claim that obesity will mean shorter lives for our children.
As the Journal put it
so well in 1998, "The data linking overweight and death, as well as
the data showing the beneficial effects of weight loss, are limited,
fragmentary and often ambiguous. Most of the evidence is either
indirect or derived from observational epidemiologic studies, many
of which have serious methodologic flaws." Indeed. Perhaps the
Journal's editors could paste that comment up somewhere in their
office so that it might be consulted every time another "scientific"
obesity study arrived.
John Luik is a health policy analyst
currently writing a book about obesity in America.
