Low-Fat Diet and Chronic Disease Prevention: the Women’s Health Initiative and Its Reception

Low-Fat Diet and Chronic Disease Prevention:
the Women’s Health Initiative and Its Reception
“The Low-Fat Diet Doesn’t Prevent Chronic Disease—or Does
Alice Ottoboni, Ph.D.
It?”10 is the title of one of the many responses that attempted to
Fred Ottoboni, M.P.H., Ph.D.
explain the disappointing findings. It summarizes the conundrum
that the trial presented to the nutrition authorities.
ABSTRACT
The dietary trial was designed in accordance with the prevailing
expert opinion of international nutrition authorities, who believed
The Women’s Health Initiative Randomized Controlled Dietary
that a reduction in total fat intake would reduce the risks of breast
Modification Trial was designed to study a low-fat diet, a nutritional
and colorectal cancers. There was less agreement on how reducing
approach to prevention of chronic diseases that was considered
total fat intake would affect heart disease, even though saturated fat
promising. The negative findings from the trial were both
was a generally accepted risk factor for heart disease. Thus, the
unexpected and disappointing to nutrition authorities. The authors’
public responses to the findings articulated an unwillingness to
heart-disease component of the study was added to the protocol
believe the finding that a low-fat diet did not prevent breast or colon
because it was anticipated that a reduction in total fat would be
cancer or heart disease. The negative results should stimulate work
accompanied by a reduction in saturated fat. Fruits, vegetables, and
on alternate hypotheses, and reconsideration of the long-standing
grains were also considered to have a role in chronic disease
proscription against dietary fat.
prevention, but this hypothesis also had not been tested in a long-
term, randomized trial.
Introduction
Design of the Trial
The Women’s Health Initiative (WHI) was launched in 1991
with the purpose of addressing the most common causes of death,
The dietary modification trial5,6,7 enrolled 48,835 postmeno-
disability, and impaired quality of life in postmenopausal women:
pausal women aged 50 to 79 years, recruited between 1993 and
heart disease, breast cancer, colorectal cancer, and osteoporotic
1998 from 40 clinical centers throughout the United States. Each
fractures. Three randomized controlled clinical trials were
clinical center had its own principal investigator and staff. The
proposed to test promising approaches to prevention of these
women were randomly assigned to the dietary modification group,
chronic diseases. The three trials were hormone therapy, calcium
referred to as the intervention group (n = 19,541; 40%), or the
plus vitamin D supplementation, and dietary modification.
comparison group (n = 29,294; 60%). The women in the
The WHI program was the largest federally funded study of
intervention group were asked to lower their fat intake to 20% of
women’s health ever undertaken, with the expenditure of more than
their total calories, and to eat five or more fruit/vegetable servings
$700 million since its inception in 1991. To put the massive WHI
and six or more grain servings a day. The comparison group was
effort in more personal terms: “After 12 years, 7.5 million forms,
asked not to make any dietary changes. Both diet groups were
and 1 million clinic visits, we have reached the most exciting phase
followed closely during the 8.1 years of the study with clinic visits
of the Women’s Health Initiative—the results!”1 The results from
and periodic questionnaires. The intervention group also
the hormone therapy trial were published earlier, in 2002.2 Results
participated in an intensive behavioral modification program using
from the calcium plus vitamin D supplementation trial3,4 and the
group sessions, self-monitoring techniques, and other strategies
dietary modification trial were published in 2006.5,6,7 A WHI
aimed to motivate and support reductions in dietary fat and increase
Extension Study through 2010 has been funded to follow
consumption of vegetables, fruit, and grains.
participants through the next four years.
The WHI Randomized Controlled Dietary Modification Trial
Results of the Trial
was designed to study a low-fat diet, a dietary approach to
prevention of cardiovascular disease, colorectal cancer, and breast
Diet Compliance: Dietary fat intake was significantly lower in the
cancer. Although unproven, nutrition scientists thought the
intervention group than in the comparison group. The difference
approach was promising. The findings were published
between groups in change from baseline for percentage of energy
simultaneously in three separate scientific papers, the first dealing
from fat was 10.7% at year 1, and this difference between groups
with breast cancer,5 the second with colorectal cancer,6 and the third
was mostly maintained throughout the trial, although by year 6, the
with cardiovascular disease.7
difference in mean fat intake between the two groups decreased
All three papers from the dietary modification trial presented
somewhat, to 8.2% of energy intake. The intervention group also
findings that were not only a surprise to the nutritionists, but also in
achieved statistically significant increases in vegetable, fruit, and
sharp disagreement with its long-held dietary recommendations, as
grain servings. Vegetable and fruit consumption was higher in the
diagrammed in the Food Guide Pyramid8 (now known as
intervention group by at least one serving per day, and grain
MyPyramid9).
consumption was higher by half a serving per day.5,6,7
10
Journal of American Physicians and Surgeons
Volume 12
Number 1
Spring 2007

---

Breast Cancer: The number of women who developed invasive
diseases “does not mean that anyone should abandon a proven
breast cancer (annualized incidence rate) was 665 (0.42%) in the
healthy diet.”13
intervention group and 1,072 (0.45%) in the comparison group. It
Responses of nutrition experts can be grouped into four major
was concluded that among postmenopausal women, a low-fat
categories: disappointment and/or disbelief, criticism of study
dietary pattern did not result in a statistically significant reduction
design, attribution of results to participant noncompliance, and
in invasive breast cancer risk over an 8.1-year mean follow-up
statements that findings were incomplete or immaterial.
period. However, because some of the findings, although not
statistically significant, indicated that there might be a reduced risk
Disappointment and/or Disbelief: Many of the principal
associated with a low-fat dietary pattern, the authors suggested that
investigators as well as other health scientists expressed feelings of
the planned longer nonintervention follow-up might yield a more
disappointment and/or disbelief about the lack of statistically
definitive comparison.5
significant benefits. Tim Byers, WHI principal investigator,
University of Colorado Health Sciences Center in Denver, said,
Colorectal Cancer: Despite dietary changes, there was no
“We are scratching our heads over some of these results.”12
evidence that the intervention reduced the risk of invasive
Margery Gass, principal investigator for the University of
colorectal cancer. There were 201 women with invasive colorectal
Cincinnati clinical center said, “The women [in the study] worked
cancer (0.13% per year) in the intervention group and 279 (0.12%
very hard to change their eating patterns. That is another reason the
per year) in the comparison group. These results led to the
finding are disappointing…. We really hoped the dietary
14
conclusion that a low-fat dietary intervention did not reduce the risk
modification would produce a major benefit in their health.”
of colorectal cancer in postmenopausal women during the mean 8.1
In an effort to assuage the almost universal feelings of disbelief
years (SD, 1.7) of follow-up.6
in the results of the study, Rossouw, the WHI project officer, offered
a ray of hope: “Scientists will observe the women until 2010, when
Cardiovascular Disease: Low-density lipoprotein cholesterol
we could hear a whole new message. I would not worry about the
levels and diastolic blood pressure were significantly reduced by
headlines of today as far as low fat and breast cancer are concerned.
15
3.55 mg/dL and 0.31 mm Hg, respectively, in the intervention vs.
They may be wrong.”
the comparison group. However, high-density lipoprotein
cholesterol, triglycerides, glucose, and insulin did not significantly
Flawed Study Design: Criticisms of the study cited its duration,
differ between the groups. The number who developed coronary
size, selection of participants, failure to distinguish between “good
heart disease (CHD), stroke, or cardiovascular disease (CVD)
fat” and “bad fat,” and inadequate lifestyle change (e.g. no exercise).
(annualized incidence rates) were 1,000 (0.63%), 434 (0.28%), and
Jacques Rossouw, WHI project officer, suggested that “some
1,357 (0.86%) in the intervention group, and 1,549 (0.65%), 642
of the hypotheses used to design the project may have been
flawed, or become outdated while the project was underway” and
(0.27%), and 2,088 (0.88%), respectively, in the comparison group.
it “may have been too short, or studied women who were too old or
It was concluded that over a mean of 8.1 years, dietary intervention
just too healthy.”13
(reduced total fat intake and increased intake of vegetables, fruits,
Ruth Kava, Director of Nutrition at the American Council on
or grains) did not significantly reduce the risk of CHD, stroke, or
Science and Health, asked: “The activity levels weren’t
CVD in postmenopausal women.7
described—if they had increased their exercise, would that have
increased their weight loss and thus decreased at least their risk of
Responses to the Trial from Nutritionists
heart disease? Would the results have been different with a longer
follow-up period?” She added that perhaps a greater degree of
As might be expected when strongly held beliefs are called into
supervision and/or education would have made a difference.10
question, nutrition experts responded to the results of the WHI diet
Kelly Brownell, director of Yale’s Rudd Center for Food
study quickly and vigorously. Because so many of the responses
Policy and Obesity, said that members of the WHI committee who
were negative, officials at the National Institutes of Health, under
designed the trial had serious questions: “Cancer and heart disease
whose aegis the study was conducted, rejected criticism that they
can take decades to develop. Would an eight-year trial be long
mishandled the study with the following comment:
enough? Would the women in the test group fully report their eating
… [P]eople are upset only because it took controversial
habits? Self-reports of dietary intake are notoriously inaccurate.”15
topics and upset accepted notions. “The strength of the
reaction has been commensurate with the strength of the
Participant
:
Noncompliance The women in the intervention
dogma it overturned,” says Jacques Rossouw, WHI project
group achieved only a modest though statistically significant
officer for the NIH.11
reduction in their percentage of energy from fat, from 38% to 29%.
“Many of the principal investigators emphasized that in many
A statement issued by the Harvard School of Public Health (HSPH)
ways, the project was very well conceived, designed, and executed,
Department of Nutrition cited the view of some nutrition experts
and has produced valuable information,” according to a report in
that benefits from the low-fat approach may have become more
the
.12
Washington Post
However, the general agreement that the
apparent had the women reached the target of 20%.16
study was properly conducted did not mute dissatisfaction with the
outcome. Nor did it persuade nutritionists to change the advice they
Findings Incomplete or Immaterial: Elizabeth Mayer-Davis,
will give to clients. As Jean Wactawski-Wende, one of the authors
director of the Center for Research in Nutrition and Health
of the breast and colorectal cancer papers, remarked, there is no
Disparities, University of South Carolina, wrote: “My view is that
question that a diet low in fats and high in fruits, vegetables, and
for several reasons the WHI results are not entirely negative or
grains is a very healthy diet. Lack of proved effect on chronic
positive; they are simply incomplete.”17
Journal of American Physicians and Surgeons
Volume 12
Number 1
Spring 2007
11

---

Marcia Stefanick, chair of the WHI steering committee, said
had been aware of problems with the low-fat diet for many years.
that additional studies that distinguish between “good fats” and
The HSPH Department of Nutrition wrote:
“bad fats” need to be done.18 In another interview, Stefanick noted
The dietary fat reduction arm of the WHI was
that “the women weren’t asked to differentiate between ‘good fats’
controversial from the beginning. Members of the HSPH
and ‘bad fats,’ which is emphasized in current guidelines for heart
Department of Nutrition argued that the hypothesis that a
disease reduction.”19
reduction in total fat would have major health benefits was
not supported by existing data…. The findings from the
The Lipid Hypothesis
Women’s Health Initiative Dietary Modification Trial came
as a surprise to many Americans who have been hearing for
The lipid hypothesis, the label for the diet-heart connection that
years that reducing fat is important for long-term health. Yet
attributes heart disease to consumption of animal fats, has a long
long-term follow-up studies such as the Nurses Health
history. It had its origins in the 1950s when nutrition pioneer Ancel
Study have consistently found little relation between the
Keys and his wife Margaret discovered the cuisine of southern Italy
percentage of calories from fats and risks of breast cancer,
and Greece, which they named the Mediterranean diet. The Keyses
colon cancer, or coronary heart disease. Such studies are one
assumed wrongly that their Mediterranean diet was low in animal
reason why major reviews of diet and health in the last five
fats because it was largely pastas, olive oil, vegetables, fruits, and
years, including those conducted by the U.S. Institute of
wine.20 Convinced that the Mediterranean diet was responsible for
Medicine and the U.S. Dietary Guidelines Committee, have
low rates of coronary heart diseases (CHD) because it was low in
moved away from advocating low fat intake to emphasis on
animal fat, they set out to obtain statistics on fat consumption and
the type of fat.15
CHD in other countries, including those that bordered the
In order to retain the lipid hypothesis as a foundation for
Mediterranean, for which there were such statistics.
official dietary policy, something that will excuse the failure of
The studies of Keys and his followers became widely popular,
studies to confirm the concept, without lessening its validity, is
and the nutrition and medical communities enthusiastically
required. This requirement could be met by new data showing that it
embraced the diet-heart connection of the lipid hypothesis. Later,
is not total dietary fat that is the culprit, but rather only the saturated
laboratory reports showing an association between blood
fat and trans fat components of total fat that are to blame for heart
cholesterol levels and CHD, plus reports that cholesterol was found
and other chronic diseases. Walter Willett, in Chapter Four of a
in coronary artery occlusions in people who died of heart disease,
Harvard Medical School diet guide, summarizes traditional
strengthened the credibility of the lipid hypothesis, because
nutrition’s rationale for why dietary fat restriction should apply
cholesterol is a lipid found only in association with animal fats.
only to saturated and trans fats, and not to total fat.22
Thus, cholesterol became accepted as a major risk factor for the
It seems incomprehensible that authorities in a scientific
lipid hypothesis.
discipline would be unaware of the wealth of data in the scientific
Many large, long-term prospective studies that examined the
literature that contradict the basis for its official position on dietary
question of a link between dietary fat and chronic disease, beginning
fat intake. Thus, the responses suggest an unwillingness to look
with the Framingham Heart Study of the post-World War II era, have
objectively at all relevant data, plus a lack of motivation to
been conducted in the United States and abroad in the last five
reexamine long-held nutrition dogma.
decades. Despite this tremendous expenditure of time and money, a
The lesson to be drawn from the WHI study is the importance
statistically significant association between dietary fat intake and
of a fundamental principle of scientific inquiry: Believe the data. If
chronic disease still eludes proponents of the lipid hypothesis.21
a well-designed study yields results that negate the hypothesis
The epidemiology on which the low-fat diet was based had a
being tested, the data must not be rejected until a valid reason for
number of early critics, but scant attention was paid to them until
the late 1980s, when a Scandinavian cardiologist began to question
doing so is found. If no errors of omission or commission can be
the purported relationship between cholesterol and cardiovascular
found, the possibility exists that some unsuspected variable may
disease. The result was a thoughtful, factual, and thoroughly
have been operating. Hence, negative results could give important
referenced book describing cholesterol myths, in which Uffe
direction to future research.
Ravnskov explains the flaws in the studies that led to the lipid
Instead of simply reaffirming support for a low-fat diet,
hypothesis and the fallacies that have supported it.20
nutritionists might ask whether the role of some other
There is ample scientific evidence that the lipid hypothesis is
macronutrient and/or micronutrient might be worth exploring.
not a valid theory, but rather a creature of misused and/or
incompetent epidemiology. A new book by Colpo21 explains the
Implications for Chronic Disease Prevention
misuses of epidemiology in the genesis of the lipid hypothesis, and
describes how and why politics and the food and pharmaceutical
Academic nutrition experts like to believe that they have been
industries actively promote the very profitable lipid hypothesis.
engaged in prevention of chronic diseases since the mid-1950s with
This exhaustive treatise, with more than 1,400 references to major
the publication of their government-sponsored dietary guidelines.
scientific journals, builds on the work of Ravnskov in exposing the
These guidelines, based on the lipid hypothesis, ushered in an era of
fiction of the lipid hypothesis.
fat-phobia that is still with us today. Now, five decades after
adoption of these guidelines as official nutrition policy, we have a
Reflections on Reactions to the WHI Study
generation of obese Americans, beset with chronic diseases. Not
only have the guidelines failed to keep their promise of preventing
The response to the WHI diet study results make it clear that
cardiovascular diseases, but they may well have played a role in
traditional nutritionists are still wedded to the concept of the lipid
causing an exceptionally large increase in the incidence of lipid
hypothesis, despite the fact that arbiters of official nutrition policy
abnormalities, type II diabetes, and metabolic syndrome.
12
Journal of American Physicians and Surgeons
Volume 12
Number 1
Spring 2007

---

As Dr. Sylvan Weinberg, former president of the American
Fred Ottoboni, M.P.H., Ph.D., and Alice Ottoboni, Ph.D., are retired public
College of Cardiology, observed:
health scientists. E-mail: ottoboni@775.net.
This [low-fat] diet can no longer be defended by appeal to
the authority of prestigious medical organizations or by
REFERENCES
1
WHI—the results are in! WHI Matters: A Publication of the Women’s
rejecting clinical experience and a growing medical literature
Health Initiative 2006;11:1.
suggesting that a much-maligned low-carbohydrate, high-
2
Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of
protein diet may have a salutary effect….23
estrogen plus progesterone in healthy postmenopausal women.
Yet it appears that the only change in official recommendations
JAMA 2002;288:321-333.
will be an inconsequential redefinition of dietary fat.
3
Jackson RD, LaCroix AZ, Gass M, et al. Calcium plus vitamin D
A basic flaw that led nutritional science to the detour down the
supplementation and the risk of fractures. N Engl J Med
lipid-hypothesis path was neglect of hard science to confirm
2006;354:669-683.
associations. Even the most competent of epidemiologic studies
4
Wactawski-Wende J, Kotchen JM, Anderson GL, et al. Calcium plus
can only determine association. Other scientific disciplines are
vitamin D supplementation and the risk of colorectal cancers. N Engl J
required to prove cause-effect relationships. Validation of
Med 2006;354:684-696.
5
epidemiologic associations must involve laboratory investigations
Prentice RL, Caan B, Chlebowski RT, et al. Low-fat dietary pattern and
of biochemical and physiologic mechanisms, and clinical studies.
risk of invasive breast cancer: the Women’s Health Initiative
Randomized Controlled Dietary Modification Trial. JAMA
Unfortunately, nutritionists seem to have little interest in how
2006;295:629-642.
biochemistry, physiology, and other medical sciences might
6
Beresford SA, Johnson KC, Ritenbaugh C, et al. Low-fat dietary
provide nutritional studies with a solid scientific foundation.
pattern and risk of colorectal cancer: the Women’s Health Initiative
Prevention of nutritionally based chronic diseases is a long-term
Randomized Controlled Dietary Modification Trial. JAMA
strategy. It cannot replace the current diagnose-and-treat scenario in
2006;295:643-654.
the near term for palliation of today’s nutritional diseases. However,
7
Howard BV, Van Horn L, Hsia J, et al. Low-fat dietary pattern and risk
as a start, an objective, scientifically based nutrition program that
of cardiovascular disease: the Women’s Health Initiative Randomized
educates the public about the true relationship among
Controlled Dietary Modification Trial. JAMA 2006;295:655-666.
macro/micronutrients, health, and disease may eventually eliminate
8
Ottoboni A, Ottoboni F. The Food Guide Pyramid: will the defects be
or at least delay the onset of the modern nutritional diseases that are
corrected? J Am Phys Surg 2004;9:109-113.
9
devastating today’s health and the medical system. A failure of
U.S. Department of Agriculture. MyPyramid.com. Available at:
prevention is a serious blow to any chance of alleviating the
www.mypyramid.gov. Accessed Dec 20, 2006.
10
growing economic burden of medical care.
Kava R. Low-fat diet doesn’t prevent chronic disease—or does it?
ACSH in Action, Spring 2006, p 2.
11 Parker-Pope T. In study of women’s health, design flaws raise
Implications for Physicians
questions. Wall Street Journal, Feb 28, 2006, p A1.
12 Stein R. New data on health: studies in confusion. Washington Post,
Medical education covers such a tremendously wide range and
Feb 19, 2006, p A1.
volume of information necessary to the diagnosis and treatment of
13 Baker L. Diet study finds little effect on chronic disease in women. UB
disease that ancillary subjects such as nutrition are given only
Reporter 2006;37(20):1. Available at: www.buffalo.edu/reporter.
superficial treatment. Physicians rightfully have assumed that
Accessed Jan 31, 2007.
nutritional science can provide whatever information they require.
14 Davis J. Low-fat, high-fruit diet may not stop disease. HealthNEWS
Medical organizations must serve their members by demanding
March 2006. Available at:www.healthnews.uc.edu/publications/
current, unbiased, accurate information from nutrition scientists.
findings/?/1552/1554. Accessed Dec 20, 2006.
15
As patients look to their physicians for all manner of advice on
Kantrowitz B, Kalb C. Food news blues. Newsweek, Mar 13, 2006, pp 44-50.
16
health, including nutrition, physicians will ultimately be held
Harvard School of Public Health Nutrition Source. Low-fat diet not a
cure-all. Available at: http://www.hsph.harvard.edu/nutritionsource/
responsible for errors. Physicians must also recognize that there is
low_fat.html. Accessed Dec 20, 2006.
no drug that will cure a nutritional disease.
17 Mayer-Davis E. Yes, diet matters. Surprising findings of WHI are not
cause to abandon nutrition. DOC News 2006;3(Apr):3. Available at:
Conclusion
http://docnews.diabetesjournals.org. Accessed Jan 31, 2007.
18 The real scoop on the women’s health studies. Stanford Magazine
An op-ed article in The Wall Street
,
Journal from a physician
2006;34(May/June):31. Available at: www.stanfordalumni.org/news/
critical of alternative therapies, presents some thoughts that
magazine. Accessed Jan 31, 2007.
nutritional scientists who continue to promote the low-fat diet,
19 Low-fat diet’s benefits for women less than expected, Stanford
despite negative results from scientific studies, should bear in
researcher says. Stanford press release, Feb 7, 2006.
20
mind. This prophecy may well predict the fate of the lipid
Ravnskov U. The Cholesterol Myths: Exposing the Fallacy that
hypothesis:
Saturated Fats and Cholesterol Cause Heart Disease. Washington,
There are the conflicting tides of belief and fact, and each
D.C.: New Trends Publishing; 2000.
21 Colpo A. The Great Cholesterol Con. Lulu Press; 2006. Available at:
has its own chronology. Things don’t change quickly, but over
www.lulu.com. Accessed Dec 20, 2006.
time a cumulative body of evidence becomes compelling. I
22 Willett WC. Chapter four. Eat Drink, and Be Healthy: the Harvard
reflected on this when I read that one major vendor of saw
Medical School Guide to Healthy
.
Eating New York, N.Y.: Free Press,
palmetto asserted that he would continue to promote the herb
Simon & Schuster; 2003.
despite the new data [that it was ineffective]. As science
23 Weinberg SL. The diet-heart hypothesis: a critique. J Am Coll Cardiol
spreads in his world, doubt will chip away at blind faith, and
2004;43:731-723.
he will find a shrinking group of believers.24
24 Groopman J. No “alternative.” Wall Street Journal, Aug 8, 2006, p A12.
Journal of American Physicians and Surgeons
Volume 12
Number 1
Spring 2007
13

---