"You can initially lose 5 to 10 percent of your weight on any number
of diets, but then the weight comes back," said Traci Mann, UCLA
associate professor of psychology and lead author of the study. "We
found that the majority of people regained all the weight, plus more.
Sustained weight loss was found only in a small minority of
participants, while complete weight regain was found in the majority.
Diets do not lead to sustained weight loss or health benefits for the
majority of people."
Mann and her co-authors conducted the most comprehensive and rigorous analysis of diet studies, analyzing 31 long-term studies.
"What happens to people on diets in the long run?" Mann asked.
"Would they have been better off to not go on a diet at all? We decided
to dig up and analyze every study that followed people on diets for two
to five years. We concluded most of them would have been better off not
going on the diet at all. Their weight would be pretty much the same,
and their bodies would not suffer the wear and tear from losing weight
and gaining it all back."
People on diets typically lose 5 to 10 percent of their starting
weight in the first six months, the researchers found. However, at
least one-third to two-thirds of people on diets regain more weight
than they lost within four or five years, and the true number may well
be significantly higher, they said.
"Although the findings reported give a bleak picture of the
effectiveness of diets, there are reasons why the actual effectiveness
of diets is even worse," Mann said.
Mann said that certain factors biased the diet studies to make them
appear more effective than they really were. For one, many participants
self-reported their weight by phone or mail rather than having their
weight measured on a scale by an impartial source. Also, the studies
have very low follow-up rates -- eight of the studies had follow-up
rates lower than 50 percent, and those who responded may not have been
representative of the entire group, since people who gain back large
amounts of weight are generally unlikely to show up for follow-up
tests, Mann said.
"Several studies indicate that dieting is actually a consistent
predictor of future weight gain," said Janet Tomiyama, a UCLA graduate
student of psychology and co-author of the study. One study found that
both men and women who participated in formal weight-loss programs
gained significantly more weight over a two-year period than those who
had not participated in a weight-loss program, she said.
Another study, which examined a variety of lifestyle factors and
their relationship to changes in weight in more than 19,000 healthy
older men over a four-year period, found that "one of the best
predictors of weight gain over the four years was having lost weight on
a diet at some point during the years before the study started,"
Tomiyama said. In several studies, people in control groups who did not
diet were not that much worse off -- and in many cases were better off
-- than those who did diet, she said.
If dieting doesn't work, what does?
"Eating in moderation is a good idea for everybody, and so is
regular exercise," Mann said. "That is not what we looked at in this
study. Exercise may well be the key factor leading to sustained weight
loss. Studies consistently find that people who reported the most
exercise also had the most weight loss."
Diet studies of less than two years are too short to show whether dieters have regained the weight they lost, Mann said.
"Even when you follow dieters four years, they're still regaining weight," she said.
One study of dieting obese patients followed them for varying
lengths of time. Among those who were followed for fewer than two
years, 23 percent gained back more weight than they had lost, while of
those who were followed for at least two years, 83 percent gained back
more weight than they had lost, Mann said. One study found that 50
percent of dieters weighed more than 11 pounds over their starting
weight five years after the diet, she said.
Evidence suggests that repeatedly losing and gaining weight is
linked to cardiovascular disease, stroke, diabetes and altered immune
function. Mann and Tomiyama recommend that more research be conducted
on the health effects of losing and gaining weight, noting that
scientists do not fully understand how such weight cycling leads to
adverse health effects.
Mann notes that her mother has tried different diets, and has not
succeeded in keeping the weight off. "My mother has been on diets and
says what we are saying is obvious," she said.
While the researchers analyzed 31 dieting studies, they have not evaluated specific diets.
Medicare raised the issue of whether obesity is an illness, deleting
the words "Obesity is not considered an illness" from its coverage
regulations in 2004. The move may open the door for Medicare to
consider funding treatments for obesity, Mann noted.
"Diets are not effective in treating obesity," said Mann. "We are
recommending that Medicare should not fund weight-loss programs as a
treatment for obesity. The benefits of dieting are too small and the
potential harm is too large for dieting to be recommended as a safe,
effective treatment for obesity."
From 1980 to 2000, the percentage of Americans who were obese more
than doubled, from 15 percent to 31 percent of the population, Mann
noted.
A social psychologist, Mann, taught a UCLA graduate seminar on the
psychology of eating four years ago. She and her students continued the
research when the course ended. Mann's co-authors are Erika Westling,
Ann-Marie Lew, Barbra Samuels and Jason Chatman.
"We asked what evidence is there that dieting works in the long
term, and found that the evidence shows the opposite" Tomiyama said.
The research was partially supported by the National Institute of Mental Health.
In future research, Mann is interested in studying whether a
combination of diet and exercise is more effective than exercise alone.
University of California - Los Angeles. "Dieting Does Not Work, Researchers Report." ScienceDaily 5 April 2007. 3 June 2009 <http://www.sciencedaily.com/releases/2007/04/070404162428.htm>.