Content uploaded by Paul Ernsberger
Author content
All content in this area was uploaded by Paul Ernsberger on Dec 29, 2017
Content may be subject to copyright.
Biomedical Rationale for a Wellness
Approach to Obesity: An Alternative to
a Focus on Weight Loss
Paul Ernsberger* and Richard J. Koletsky
Case Western Reserve School of Medicine
The direct medical hazards of obesity, although real, have been overstated.
Because current remedies for obesity have little long-term effectiveness, no con-
trolled clinical trial has demonstrated improved longevity after weight loss. In
contrast, advances in drug therapy for diabetes, hypertension, and high choles-
terol allow obese persons affected by these conditions to live healthier lives. Fur-
thermore, weight cycling may cause much of the cardiovascular risk associated
with obesity. Repeated loss and regain of weight increases human deaths from
heart disease, and in obese laboratory animals weight cycling increases blood
pressure, enlarges the heart, damages the kidney, increases abdominal fat depos-
its, and promotes further weight gain. Additional health risks in obesity may be
caused by hazardous treatments for obesity, as illustrated by heart disease caused
by diet pills. Obese patients often lack full access to medical services owing in part
tosocialstigmaandlowself-esteem,whichimpairself-careactivities,andthebias
of health professionals. These barriers, along with the prevalence of poverty
among the obese, may contribute to the association of obesity with poor health.
Medical beliefs about obesity are shaped by expert panels that are highly selective
in the data they consider. Experts included on government consensus panels have
been disproportionately drawn from the ranks of diet clinic directors, which
might explain the congruence between panel recommendations and the economic
interests of the diet industry. One remedy is a wellness approach focused on
healthy lifestyle, positive attitude to health and self-care, and a disregarding of
predetermined weight standards in favor of preventing further weight gain and
reducing risk factors. Medical conditions common in obese patients, including
*Correspondence concerning this article should be addressed to Paul Ernsberger, Department of
Nutrition,CaseWesternReserve University,Cleveland, OH44106-4906 [e-mail:pre@po.cwru.edu].
Journal of Social Issues, Vol. 55, No. 2, 1999, pp. 221–260
221
© 1999 The Society for the Psychological Study of Social Issues
43
hypertension, type-2 diabetes, hyperlipidemia, and sleep apnea, are dealt with
directly and aggressively rather than relying on weight loss as the primary treat-
ment. This new approach should improve the physical and mental well being of
obese patients.
In this review, we address the prevailing view of obesity as a major threat to
public health and find that this paradigm is based on incomplete consideration of
theevidence.Inparticular, life expectancy is reduced more byunderweightthanby
overweight. Although many diseases are more common in obese patients, in many
cases a direct causal link cannot be made. Weight loss alone has limited effective-
ness as a treatment for chronic medical conditions, whereas lifestyle enhancement
canimprove health independent of the loss of body fat. Repeated loss and regain of
weight may make a significant contribution to obesity-associated disease. Also,
obese persons are stigmatized and often have low socioeconomic status, which
contributes independently to disease and reduced life expectancy. We advocate a
wellness approach focused on healthy lifestyle and treating disease in the obese,
rather than treating obesity as a disease.
Body Weight, Risk Factors, and Disease
What are the health hazards of obesity? In looking to the medical literature,
there are hundreds of studies we can turn to for information. How can we decide
which studies provide the best information? Epidemiologists and statisticians rank
different types of medical studies according to their validity, reliability, and
generalizability (Table 1; Hill, 1966). The gold standard for guiding medical prac-
tice is the controlled clinical trial. In the case of obesity, a controlled trial would
require a group of obese persons to be divided into two groups: one to be cured of
their obesity by some form of treatment, and a comparison group that would not be
treated. This type of study is impossible, of course, because there are no effective
treatments for obesity. Next to controlled trials, the most trustworthy source of
information is the prospective study. A well-known example is the Framingham
study,inwhich5,000residentsoftheBostonsuburbwereweighedandgivenmed-
ical exams in 1948 and their health was followed for the next 30 years (Lissner et
al., 1991). The largest of these controlled trials involved 1.8 million Norwegians,
who were weighed and examined and then followed for 10 years (Waaler, 1984).
Prospectivestudiesinform us about the predecessors of disease,andfactorssuch as
body weight that can predict whether someone stays alive and healthy or sickens
and dies prematurely. Dozens of prospective studies have examined the long-term
health consequences of obesity, and these are summarized in Table 2. Prospective
studies are the best source of information on the relationship between body weight
and health and are the focus of this article.
222 Ernsberger and Koletsky
44
Retrospective studies compare persons having a particular disease (the cases)
with other people who are healthy (the controls). These studies attempt to recon-
structthe cause of a disease by comparing the medical histories of its sufferers com-
paredtohealthycontrols.Forexample, using a case-control design it was found that
people with pulmonary hypertension are 23 times more likely to have a history of
taking diet pills than are matched controls (Abenhaim et al., 1996). The retrospec-
tive design is most commonly used for relatively uncommon disorders, such as pul-
monary hypertension, where population-based prospective studies are impossible.
Theweakesttypeofmedical study is the cross-sectional survey. Cross-sectional
studies are also the least difficult to carry out, requiring no more than a questionnaire
or a single medical exam. Typically, participants are asked their height and weight
and what diseases their doctor has told them they have. Obese persons usually report
more medical diagnoses, particularly diabetes, hypertension, high cholesterol, heart
disease, arthritis, and gallstones. Compared to prospective studies, there are many
problemsandconfounding factors with cross-sectional studies. Oneexampleisdiag-
nostic bias. Doctors are trained to expect certain diseases in obese persons and may
diagnose them more readily. For whatever reason, cross-sectional studies generally
paint a much more unfavorable picture of the health of obese persons than controlled
prospective studies. In this report we have focused on prospective studies that do not
rely on a doctor’s diagnosis, but rather use the simplest and most verifiable measure
of health: age at the time of death.
Wellness Approach to Obesity 223
Table 1. Hierarchy of Medical Studies
Type of study Design Available for obesity?
Controlled clinical trial Obesity is reversed by an
effective treatment. Health
outcomes compared between
treated and untreated groups.
Not possible, because all
therapies lack long-term
effectiveness.
Prospective study Group is monitored for many
years, and the relationship
between initial body weight
and later health outcomes is
determined.
Yes. Nearly every prospective
epidemiological has included
body weight.
Retrospective case-control study Persons with a particular
disease are compared to
healthy controls matched for
age, gender, and background.
Yes.
Cross-sectional survey Groups of people are
surveyed, and the number of
diagnoses reported by people
of different body weights is
recorded.
Yes. This is the type of study
mainly considered by obesity
experts.
45
Table 2. Population Characteristics and Outcome of Epidemiological Studies
of Obesity and Total Mortality
Study
population Relative hazard
of obesity Likely prevalence of
weight loss practices Reference
Young nurses Exceptionally high Very high Manson et al., 1995
Holders of individual life
insurance policies Very high Very high Lew, End, & Wilber,
1979
Harvard alumni Very high High Lee, Manson,
Hennekens, &
Paffenbarger, 1993
Residents of affluent
Boston suburb
(Framingham)
High High Garrison, Feinleib,
Castelli, & McNamara,
1983
Neighbors and relatives of
American Cancer Society
volunteers
High High Lew et al., 1979
Residents of Finland Moderate Moderate Rissanen et al., 1989;
Rissanen et al., 1991
White women in
Charleston, South Carolina Moderate Moderate Stevens et al., 1992
Black women in
Charleston, South Carolina None Low Stevens et al., 1992
Civil servants in rural
eastern Finland None (Women)
Low (Men) Low Tuomilheto et al., 1987
German construction
workers Inverse Low Brenner et al., 1997
Dutch civil servants None (Women)
Low (Men) Low Tuomilehto et al., 1987
Black Kaiser Permanente
subscribers None (Women)
U-shaped* (Men) Low Wienpahl, Ragland, &
Sidney, 1990
San Francisco
longshoreman Inverse relation Low Borhani, Hechter, &
Breslow, 1963
Residents of villages in
rural Italy None Very low “Incidence and
Prediction of Coronary
Heart Disease,” 1982
Residents of villages in
rural Scotland Inverse relation Very low Garn, Hawthorne,
Pilkington, & Pesick,
1983
Elderly populations Inverse relation Very low See text discussion.
Residents of American
Samoa None Very low Crews, 1989
Residents of Micronesia Inverse relation Very low Vandenbroucke et al.,
1984
46
Even though cross-sectional studies are the least reliable (Table 1) and are sel-
domreliedupon when other information is available,obesityexpertsrely primarily
on these health surveys. For example, reviews of the health consequences of obe-
sity focus almost entirely on the fact that certain diseases are more common in
obese people than in lean ones (Akram et al., 1997; Bray, 1996; Pi-Sunyer, 1993).
Thus, we will briefly discuss each of these conditions in turn. We will first discuss
the evidence that the disease is more common among the obese, then we will
explore possible disease processes that might logically link obesity to the pathol-
ogy, and finally we will address the question of whether weight loss is an effective
treatment for the condition.
Type-2 diabetes is probably the condition most strongly linked to obesity. In
type-1 diabetes, the pancreas fails to produce enough insulin. Without frequent
injections, the patient will die. In contrast, insulin levels are normal or even high in
type-2diabetes,but blood sugar is elevated because insulinisunableto carry out its
functions. Type-2 diabetes does not lead to as severe a set of complications as
type-1, and the biggest danger of the disease is as a risk factor for coronary heart
disease. Type-1 diabetes usually begins before age 40, and decreases life expec-
tancy by an average of 8 years (Bale & Entmacher, 1977). Type-2 usually occurs
laterinlife,andreduceslife expectancy by only 2 to 4 years. Diabetics over the age
of 70 have the same life expectancy as nondiabetics of the same age.
The link between obesity and type-2 diabetes is undeniable. One oft-cited
statistic is that 80% of persons with type-2 diabetes are obese (Bray, 1996;
Pi-Sunyer, 1993). Less often mentioned is the fact that in the demographic group
with the highest incidence of type-2 diabetes, women in their 50s, the overall prev-
alence of obesity is about 60% (Flegal, Carroll, Kuczmarski, & Johnson, 1998). In
Wellness Approach to Obesity 225
Residents of Fiji Inverse relation Very low Hodge, Dowse, Collins,
& Zimmet, 1996;
Collins, Dowse,
Cabealawa, Ram, &
Zimmet, 1998
New Zealand Maori None Very low Salmond, Beaglehole,
& Prior, 1985
Native Americans of the
Pima tribe Inverse relation
(women)
Inverse up to BMI
of 40 (men)
Very Low Hanson et al., 1995
Note: Studies are ranked according to the degree of relative risk associated with obesity. The incidence
of weight loss behavior is based either on direct report or on the characteristics of populations known to
be associated with a high prevalence of dieting practices.
*Highest mortality at body mass index (BMI) < 20; lowest mortality at BMI of 28 (moderately over-
weight). Risk increase only for BMI > 40.
47
one study cited by the World Health Organization (Akram et al., 1997), diabetes
was 40 times more common in women with a body mass index (BMI; kg body
weight per meter of height squared) of 30 relative to lean women with a BMI
less than 22. Although most studies have not documented such an extreme risk,
an association does exist and has been verified in prospective studies, wherein
obese persons are found to be more likely to develop diabetes (Edelstein et al.,
1997). However, a relationship between obesity and subsequent development of
diabetes was not found in all studies. In a study of a multiethnic population in
Israel, there was no increase in risk of diabetes except when BMI was greater than
31 (Modan et al., 1986).
Does excess body fat lead directly to type-2 diabetes? Studies have consis-
tently shown that the disease is genetic in origin. If one identical twin over the age
of 50 has type-2 diabetes, there is a 91% chance that the second twin will also
develop it (Barnett, Eff, Leslie, & Pyke, 1981). Although it is true that identical
twinssharetheirenvironment and diet while growing up, by the sixthdecadeoflife
the effect of childhood environment is probably minimal. Furthermore, there was
littleeffect of BMI on the age of onset for diabetes, as the first twin to bediagnosed
had an average BMI of 23.8 while the second twin, who stayed free of diabetes up
to 15 years longer, had an average BMI of 24.9. Given that diabetes type-2 is a
genetic disease and that most of its victims are obese, it follows that the genes
causing diabetes must also facilitate weight gain. There is now excellent evidence
that this is the case.
Excess levels of insulin are the first abnormality to appear in future diabetics,
followedbyexcessive weight gain. This was showninaprospective study in which
levels of insulin were found to predict the subsequent rate of weight gain during a
17-year follow-up (Sigal et al., 1997). People with a family history of diabetes are
heavier than those lacking this genetic endowment, and this high-risk group also
shows more weight gain over time (Lapidus, Bengtsson, Lissner, & Smith, 1992).
Also, people with diabetes in their family have higher insulin levels, independent
of their body weight (Ishikawa, Pruneda, Adams-Huet, & Raskin, 1998). This is
consistent with the idea that a diabetes-inducing “thrifty gene” promotes obesity
andweightgain(Groop& Tuomi, 1997). As suggested by the epidemiologist Peter
Bennett, “Insulin resistance, therefore, may be the key defect that independently
[italics in original] leads to obesity, hypertension, and diabetes, and accounts for
the well-known associations among obesity, serum insulin levels and glucose
intolerance. Insulin resistance, rather than obesity, may be the principal determi-
nant of diabetes” (Bennett, 1986).
Bennett’s model for the development of diabetes is illustrated in Figure 1.
Geneticfactorsinitiatethediseaseprocess,leadingtoinsulinresistanceand a com-
pensatory increase in insulin production by the pancreas. High levels of insulin
lead to weight gain, which can further exacerbate insulin resistance. After many
years, the system decompensates and blood glucose rises. Dietary factors may also
226 Ernsberger and Koletsky
48
contributetothedevelopmentofdiabetes. In one study, nutritional habits were sur-
veyed in a population followed prospectively for 20 years (Feskens et al., 1995).
Future diabetics ate more saturated fat and took in fewer vegetables and less vita-
minC.Thus although genetic factors predominate,lifelongdietary habits can mod-
ify genetically determined risk.
How effective is weight loss as medical treatment for diabetes? Obviously,
blood glucose levels fall within hours of skipping a meal, so there can be an imme-
diate apparent benefit. Simply because blood glucose falls in the short run, how-
ever, does not mean that the chronic disease of diabetes has been reversed. Many
studies have looked at the short-term benefits of weight loss programs, usually
while the participants are still on a restrictive regimen. Many authors have
reviewed the apparent risk factor improvements that take place during the initial
phase of weight loss and in early maintenance (Akram et al., 1997). Here we will
consider only those few studies that included at least a 6-month follow-up. For
type-2 diabetics, the critical factor to monitor is glycosylated hemoglobin, which
provides a picture of the prevailing glucose levels over several weeks. A review of
all the controlled trials of weight loss in type-2 diabetes showed that there were
initialimprovementsimmediatelyaftertheweightlossprogram, but follow-up at 6
to18monthsshowedadeteriorationback to starting values, even when weight loss
persisted (Ciliska, Kelly, Petrov, & Chalmers, 1995). In 21 experimental groups
where there was follow-up, persisting benefit was found only in 3, despite main-
tained weight losses of 3 to 9 kg. A 1-year follow-up study of a behavior modifica-
tion weight loss program showed that diabetics were actually worse off for having
lostweight after 1 year. Overall, it appears that existing very-low-calorie or behav-
ioral programs have no beneficial effect at 1 year of follow-up, even when weight
loss is maintained. In a prospective study of the relationship between weight
change and mortality in type-2 diabetics, those who lost weight had a higher, not
lower, risk of death (Chaturvedi, Fuller, & the WHO Multinational Study Group,
1995).Formoderatelyoverweightdiabetics (BMI < 26), successful weight loss led
Wellness Approach to Obesity 227
Fig.1. Hypothetical model,based oncurrent human geneticstudies, forthe relationship betweengenetic
defects in insulin signaling, circulating insulin, obesity, and the development of diabetes.
49
to a tripling of the death rate. For severely obese diabetics (BMI > 29), there was a
small but significant (16%) reduction in death rate associated with weight loss.
Why are diabetics who are losing weight more likely to die? One possibility is that
weight loss occurs as diabetes progresses and worsens. Thus those that are losing
weight may be those with the most severe disease. However, if diabetics with
severe disease lose weight even as they continue to deteriorate, the question
remains whether weight loss is best treatment for type-2 diabetes.
Can weight loss prevent diabetes? Most prospective studies suggest that
weight gain precedes the onset of type-2 diabetes. However, other studies do not
showthis.Inan Israeli prospective study, formerly obese personshadahigherinci-
denceofdiabetes(14%)thancurrently obese subjects who had gained weight (6%;
Modan et al., 1986). In every category of current BMI, weight losers had a higher
risk of developing diabetes than those who gained weight or stayed stable. Among
the Pima Indians, who have the highest rate of type-2 diabetes in the world, weight
gain is not related to the development of the disease (Charles et al., 1993). Despite
poor results from weight reduction trials in diabetics, weight loss is still usually
considered the cornerstone of diabetes treatment. As noted in the New England
JournalofMedicine(Wood&Bierman,1986),“Tradition,asopposed to scientific
evidence, has had a remarkable influence on the prescription of dietary therapy for
diabetes.”
Hypertension is two to three times more common in obese women and men
than in the leanest members of the population (Akram et al., 1997). Hypercholes-
terolemia,ontheother hand, is only weakly related to body weight. Thecorrelation
coefficient between BMI and cholesterol level in most studies is about 0.1 in
women and 0.2 in men (Garn, Bailey, & Block, 1979). Blood pressure correlates
better with BMI, usually around 0.4 in middle age (Ernsberger & Nelson, 1988a).
Putanotherway, if you know awoman’sBMI,you can predict her cholesterollevel
with an accuracy only 1% better than pure chance. Blood pressure can be predicted
a little better on the basis of BMI, with an accuracy 16% better than chance.
Doeshypertensionresult directly from excess body fat? Obesityandhyperten-
sion commonly accompany each other in human populations (Sims, 1982), but the
process leading from enlarged fat stores to high blood pressure is unknown. A
weakness in our current understanding is the lack of an animal model for obesity-
related hypertension. Dogs become hypertensive when their daily diet is supple-
mentedwith2 lb of lard (Rocchini, Moorehead, Wentz, & Deremer, 1987), but this
may be a direct effect of excess saturated fat intake. Rats fed an obesity-inducing
diet of sweetened lard for one year show a slight increase in blood pressure when
measured by one method, but not by a second method (Buñag, Krizsan, & Itoh,
1990).Obesitycaused by overfeeding has no effect onbloodpressure(Contreras &
King, 1989; Ernsberger & Nelson, 1988b). Zucker fatty rats have normal blood
pressures (Buñag & Barringer, 1988). Thus, obese animal models do not show ele-
vated blood pressures.
228 Ernsberger and Koletsky
50
Ananimalmodel of hypertension in obesitydevelopedatCase Western Reserve
University involves the expression of a specific obesity gene on a background of
genetic hypertension (Ernsberger, Koletsky, Baskin, & Collins, 1996; Ernsberger,
Koletsky, Baskin, & Foley, 1994; Koletsky, Boccia, & Ernsberger, 1995). In this
model, lean and obese siblings are both hypertensive, but the obese animals unex-
pectedly have somewhat lower blood pressures. Thus, the net impact of genetic obe-
sity on blood pressure is actually protective. However, the association of obesity
with reduced blood pressure is lost under certain circumstances. For example, if the
animals are fed a diet high in salt, blood pressure will rise markedly in the obese rats
so that they are more hypertensive than the lean rats. Another example is weight
cycling, wherein the loss of weight on a low-calorie diet is alternated with ad libitum
refeeding.Weight-cycledobese rats have very highbloodpressures,exceeding those
of their lean siblings (Ernsberger, 1994; Ernsberger, Koletsky, Baskin, & Collins,
1996). Therefore, under nutritional stresses such as excess salt consumption or
“yo-yo dieting,” obesity becomes a liability to the cardiovascular system.
Why do obese animals have normal blood pressures? Weight stability might
distinguish obese rats from obese humans. Obese animals maintain steady body
weights and consume a constant amount of food on a day-to-day basis. In contrast,
obese humans frequently alternate between fasting and bingeing and show large
fluctuations in body weight as they lose and then regain a major proportion of their
body fat (Blackburn et al., 1989). An animal model of this so-called “yo-yo
syndrome” has been characterized (Ernsberger & Nelson, 1988b). Animals made
obese by overfeeding developed hypertension only when subjected to periodic
supplemented fasts. Blood pressure fell during weight loss, but after then rose,
overshooting the original level. Because this form of hypertension occurs during
therefeedingperiodfollowing a fast, it is known asrefeedinghypertension.Hyper-
tension results from refeeding in dogs, pigs, mice, and rats (Ernsberger &
Koletsky,1993).Geneticallyobeseanimalsshowevengreaterrefeedinghyperten-
sion when weight cycled, and also develop enlarged hearts and significant kidney
damage (Ernsberger & Koletsky, 1993; Ernsberger et al., 1994; Ernsberger,
Koletsky, Baskin, & Collins, 1996; Koletsky et al., 1995). Thus when obese
animalsare made to alternately lose and regain weight, they develop diseases com-
parable to those experienced by obese humans.
For human patients, a fall in blood pressure has been observed immediately
following caloric restriction (Landsberg & Young, 1978), leading to guidelines
recommending weight loss in hypertensive patients. However, there is little infor-
mation on long-term changes in blood pressure, particularly during weight regain.
Inpatient studies (Jung, Shetty, & James, 1980) show rapid rebound of blood
pressure upon full restoration of caloric intake. A report on the effects of weight
regain after therapeutic weight loss showed increases in systolic blood pressure as
well as cholesterol, triglycerides, and plasma glucose (Pfohl, Luft, Blomberg, &
Schmülling, 1994).
Wellness Approach to Obesity 229
51
The cause of hypertension in obese persons might be the cycles of weight loss
and regain they undergo, rather a direct influence of fat tissue. This would explain
theweakrelationshipbetweenobesity and hypertension in cultures and societies in
which weight loss practices are uncommon (see Table 2). This “yo-yo hypothesis”
requires further testing in human populations.
Weight loss is the most common nondrug treatment recommended for hyper-
tension. However, as is the case for diabetes, the long-term results are disappoint-
ing, even when losses are maintained, including massive amounts of weight lost
after gastric surgery (Ernsberger & Nelson, 1988a). Promising results have been
obtained with trials of multifactorial interventions that have included exercise,
sodium restriction, healthful overall diets, and stress management along with
weight loss (Beckmann et al., 1995; Whelton et al., 1998). For multifactorial pro-
grams, it is impossible to say how much of the blood pressure reduction is due
to the small loss of body weight (< 5 kg) that accompanies these lifestyle
modifications.
As noted above, cholesterol levels are only weakly correlated with body
weight. Indeed, even individuals with so-called morbid obesity selected for gastric
surgerydonotshowelevatedcholesterollevels(A. M. Wolf, Beisiegel, Kortner, &
Kuhlmann, 1998). The weak relationship between fatness and cholesterol levels
can probably be accounted for by dietary factors. A diet high in fat and low in fiber
and vegetables can, through separate and independent actions, lead to weight gain
and increases in cholesterol, as indicated in the paradigm delineated in Figure 2.
Similarly, a sedentary lifestyle independently promotes weight gain and elevated
cholesterol. The relevant question is where to intervene. We would argue that
modification of dietary and exercise habits is a more direct approach and more
likely to achieve long-term results. Moreover, medications with proven effective-
ness are available to treat excessive cholesterol levels.
230 Ernsberger and Koletsky
Fig. 2. Hypothetical model for the relationship between dietary saturated fat, obesity, and elevated
cholesterol levels.
52
Type-2 diabetes, hypertension, and elevated cholesterol are primarily of
concern because they lead to coronary heart disease, the number one killer in the
Western world. What is the relationship between obesity and atherosclerosis? The
evidence from prospective studies is contradictory. The Framingham study in par-
ticular seems to indicate a higher rate of diagnosed coronary heart disease in obese
participants (Higgins, Kannel, Garrison, Pinsky, & Stokes, 1988). However, in the
Framingham study weight cycling could account for all of the excess risk associ-
ated with obesity (Lissner et al., 1991). Two comprehensive reviews have covered
the available literature and concluded that there is no consistent relationship
between body weight or body fatness and coronary heart disease (Barrett-Connor,
1985; Williams, Jones, Bell, Davies, & Bourne, 1997).
Arthritis, or more specifically the osteoarthritis associated with old age, has
been linked to obesity in cross-sectional and prospective studies (Felson, 1996). In
contrasttothe diseases mentioned so far, awell-understoodprocesslinks increased
body weight to osteoarthritis through increased wear and tear on the joints.
Osteoarthritis is the only disease that has been proven to be directly caused by high
body weight. It should be noted, however, than thin persons are not exempt from
osteoarthritis.
Gallstones are typically listed as one the health consequences of obesity
(Akram et al., 1997; Bray, 1996; Pi-Sunyer, 1993). Even though gallstones are
much more common in obese persons than in lean ones, excess body weight is
almost certainly not a direct cause of gallstones. That is because large and rapid
weight losses are a proven direct cause of gallbladder disease (Everhart, 1993;
Vezina et al., 1998). Obese persons are more at risk primarily because they are
more likely to engage in crash dieting. An additional role may be played by excess
dietary fat, similar to the interplay between lifestyle factors, obesity, and choles-
terol (Figure 2).
The Biological Response to Caloric Restriction
What physical changes happen during a weight loss diet, and how long do they
take? Timing is an important clue to understanding what is going on in the body.
Consider the concept of ideal weight. According to the ideal weight hypothesis,
there is a small range of BMI values that we all should attain. If your BMI is higher
than this ideal, then your blood sugar, blood pressure, and blood cholesterol should
increase proportionately with each excess pound. As you lose weight, your risk fac-
tors should decline according to the proportion of excess weight you have removed.
Thus if you lose 10% of your excess weight, you should lose 10% of your excess
blood pressure. Does this happen? The answer appears to be no. These risk factors
are normalized very quickly, in many cases before a significant amount of weight is
lost. Blood sugar, of course, drops within a few hours of skipping a meal and stays
low all the time when one is on a low-calorie diet. Blood sugar stays low even when
Wellness Approach to Obesity 231
53
one goes off the diet, because one’s tissues readily take up nutrients after a period of
deprivation. Blood pressure also drops quickly on a very- low-calorie diet, usually
within a week, before there is much loss of body fat (Ernsberger & Nelson, 1988a).
On more moderate caloric restriction, there is less of a drop in blood pressure and it
takes longer to develop. Drops in cholesterol take a few weeks to register, but the
full drop in cholesterol is achieved after only a fraction of excess weight is lost. In
fact, “bad” cholesterol (LDL, or low-density lipoprotein) reaches its lowest level 8
weeks into a diet and actually rises with additional weight loss (Andersen, Wadden,
Bartlett,Vogt,&Weinstock,1995).Improvementsincholesterolprofileweremore
related to improved dietary habits such as reduced saturated fat intake than to the
amount of weight loss. Therefore, the reductions in risk factors for heart attack that
areseenwith reduced-calorie regimens are notsolelytheresult of body fat loss.This
argues against the ideal-weight hypothesis.
Why do risk factors drop so quickly compared to the slow and gradual loss of
body fat? One explanation is the fall in risk factors is a biological response to mild
starvation, rather than the result of reducing body fat stores. We tested this idea
with a meta-analysis of trials of weight loss for the treatment of blood pressure
(Ernsberger & Nelson, 1988a). Comparison of the kilograms of weight lost on the
diet to the blood pressure reduction showed no relationship. On the other hand, if
welookattherate of weight loss per week, there was a strong relationship.Gradual
weight loss had little effect, whereas there were dramatic falls in blood pressure
with rapid weight loss, especially with very-low-calorie formulas. By the same
token, diets providing very few calories (300–800 per day) led to large falls in
blood pressure, whereas more moderate diets with 1,200–1,800 calories seldom
lowered pressure. This means that the process of “going on a diet” and entering a
state of deprivation results in a lowering of blood pressure. Part of the benefit may
result from the consumption of healthier “diet” foods such as fruits and vegetables
and the avoidance of high-fat and high-sodium “junk” foods. In the long run, how-
ever, this blood-pressure-lowering effect is not sustained after the diet ends.
Mostpatients with high cholesterol are treated first with a weight loss regimen
before any drug therapies are given. However, despite short-term changes, there is
limited long-term effectiveness of weight loss as cholesterol-lowering therapy
(Andersen et al., 1995; R. N. Wolf & Grundy, 1983). In one study, obese patients
were placed on a 1,000-calorie diet until they got all the way down to their insur-
ance table weight (R. N. Wolf & Grundy, 1983). Importantly, their diet was
controlled so that they consumed the same amount and type of dietary fat before
and after losing weight. After a temporary dip during the weight loss process,
levelsofLDLcholesterolreturnedtomatchthe starting obese level. Thus provided
that dietary fat intake is not changed, weight loss does not improve cholesterol
levels. One exception to this is weight loss by surgical methods such as gastric or
biliopancreaticbypass.By interfering with the absorptionofdietary fat, these oper-
ations can significantly lower cholesterol. However, even following massive
232 Ernsberger and Koletsky
54
weight losses of 50 kg or more, a fall in cholesterol level is not always seen (A. M.
Wolf et al., 1998). Similarly, the new drug orlistat can lower cholesterol by
preventing the uptake of dietary fats from the digestive tract. The effectiveness of
these surgical and pharmaceutical interventions does not arise from loss of body
fat, but from a direct intestinal action.
Undeniably, weight loss programs can benefit health. This is especially true
when these programs emphasize permanent lifestyle changes and encourage exer-
ciseandhealthierfood choices. On the other hand, positivelifestylechangescanbe
encouraged without a primary focus on weight loss. Thus exercise programs and
low-fatdietscanyieldrealandlastingimprovementsinriskfactorswhilefailingto
correctobesity(Dengel,Katzel,& Goldberg, 1995). Importantly, improvements in
cholesterol and other risk factors stemming from improved diet are maintained as
long the dietary guidelines are followed and do not dissipate with time.
The poor effectiveness of weight loss stands in marked contrast to the increas-
ing effectiveness of medications. New and highly potent treatments for type-2 dia-
betes,hypertension,andhigh cholesterol have appeared in the last few years. Often
new and old drugs can be combined for even greater effectiveness. A drawback of
relying on weight loss as a first line of treatment is that chronically ill patients may
bedeniedtrulyeffective pharmaceutical therapy while pursuing the elusive goal of
permanent weight loss. Weight reduction should be considered at most an adjunct
to treatment rather than a primary goal.
Weightlossnotonlymaybeineffectiveas a tool for managing chronic disease
but may even cause harm in the long run. Weight loss is not usually permanent,
regardless of the intervention used. During regain of lost weight, all of the
short-term benefits of weight loss are undone, and in many cases risk factors
become worse during weight regain than they were at the starting level
(Ernsberger, Koletsky, Baskin, & Collins, 1996). Blood pressure shows large
increases during the relapse to obesity in humans and in laboratory animals.
Worsening of risk factors during regain probably accounts for the increased heart
attack deaths seen in persons who lose and regain weight.
Another potentially harmful effect of weight loss regimens is that they can
triggerbingeeating. Binging is never healthy, but inpersonswithchronicdisease it
can do serious harm. Diabetics need to maintain a steady level of food intake to
keep their blood sugar levels tightly regulated. Alternately starving and bingeing
cancompromisebloodsugar control. Similarly, chaotic eating patterns canhamper
therapy for high blood pressure and cholesterol.
The position of weight loss in medicine today can be compared to the role of
bloodletting 150 years ago. Bloodletting became popular because doctors found
that if feverish patients were bled, their fever would break and their skin would
become cool and clammy. Thus bloodletting improved the symptoms of sick
patients. Of course, we now know that blood loss creates a state of shock that
lowers body temperature but ultimately increases the risk of death. Similarly,
Wellness Approach to Obesity 233
55
weight loss produces short-term improvements in symptoms but may not be
ultimately beneficial. Before weight loss can be removed from its exalted status as
a therapy, a revolution in medicine may be required comparable to the one that
brought an end to the practice of bloodletting.
To summarize, weight loss programs can produce many short-term benefits
that have been documented in literally thousands of medical studies. However,
those few trials with follow-up beyond 6 months have failed to show lasting bene-
fit. Weight loss programs can actually do harm by diverting the patient from more
effective and reliable treatments with modern drugs or with sustainable lifestyle
changes. Furthermore, when lost weight is regained the patient may be worse off
than when she started because of the harmful consequences of weight cycling. All
oftheselimitationsofweight loss programs compel the creation of a new paradigm
that incorporates the health-promoting aspects of permanent lifestyle changes
without a focus on the correction of obesity.
Body Weight and Mortality
How hazardous is obesity? A condition is primarily considered hazardous
when it shortens life expectancy. To establish this, it must first be shown that
people having the condition consistently die sooner than appropriate controls.
Evidence for a causal relationship must then be obtained. Alternatives to the
hypothesis of simple causality must be refuted. Furthermore, reversing the condi-
tion should extend life.
The relationship between initial body weight, as measured by BMI and subse-
quent mortality, is complex and varies markedly between different prospective
epidemiological studies (Table 2). Most studies have shown a U-shaped relation-
ship between BMI and mortality, with both low and high body weights associated
withincreasedrisk of death (Andres, 1980).Othersstudieshave shown no relation-
ship (Keys, 1981, 1989). A few studies have shown a steady decline in mortality
with increasing BMI, and a few have shown a steady increase in mortality with
increasing weight up to a relative risk (RR) between 1.5 and 2 for the extremely
obese. The latter studies, listed in the top few rows of Table 2, have been empha-
sized in the setting of public health policy concerning body weight. A quantitative
meta-analysis of 23 major studies showed a U-shaped curve in the combined data
for men and women, with increased mortality when BMI was less than 23 or
greater than 28(Troiano, Frongillo,Sobal,& Levitsky,1996). Toplace this inper-
spective, a BMI of 21 corresponds to the old “ideal weight” set by Metropolitan
Life Insurance. Thus, being less than 10% “overweight” or being underweight by
any amount raises the risk of death. At the other end of the spectrum, being more
than 35% “overweight” also constitutes a hazard.
Whichismoreimportantinthe overall population, the hazards of underweight
or the hazards of overweight? In a study of Finnish women ages 25–64, the leanest
234 Ernsberger and Koletsky
56
fifthandthefattestfifthofthepopulationbothshowedaRRofabout1.5(Rissanen
et al., 1991). For women over 65, only underweight was a hazard. Therefore, the
impactof underweight and overweight are very nearly the same. Nearly equivalent
results have been obtained in many populations, whenever U-shaped distribution
of mortality versus BMI has been obtained. Based on his study of 1.8 million
Norwegians, Waaler (1984) estimated that 4.3% of deaths in persons aged 30–79
could be attributed to obesity. Especially if the elderly are included, the number of
excess deaths attributable to underweight may be higher. Therefore, a rational
approach to public health would dictate devoting equal time to treating under-
weight as overweight.
Some authors acknowledge that moderately fat people have increased life
expectancy, but nonetheless argue that obesity is dangerous (Lee & Manson,
1998). It has been claimed that because cigarette smokers weigh an average of 2 to
3 kg less than nonsmokers, the excess mortality among thin people can be
explained by the higher numbers of smokers in that group. However, in numerous
studies, mild obesity was found to be protective in nonsmokers and smokers alike.
Adjustment of mortality data for smoking has no effect on its relationship to body
weight. Importantly, only light to moderate cigarette smokers tend to be lean. In
many studies, persons who consume at least two packs a day actually weigh more
than nonsmokers. Thus, the smokers at the highest risk for disease are not
unusually lean.
Another explanation offered for the increased mortality among underweight
persons is the following: The reason that thin people are not obese is because they
have a hidden, undetectable disease that causes them to lose weight and then, after
many years of thinness, causes their death. Thus even though lean people have a
shortenedlifespan,itisstilladvantageousto be thin. However, there is no evidence
that “occult wasting disease” even exists. Patients with unexplained weight loss
tend to either deteriorate or die within a single year, or to regain their lost weight
along with their health (Reife, 1995; Wallace, Schwartz, LaCroix, Uhlmann, &
Pearlmann, 1995). Significant weight loss is a late symptom in cancer, appearing
only after the tumor is large enough to have metabolic effects or cause pain.
Because weight loss occurs only in the advanced stages of serious disease, nearly
all persons who have become lean because of fatal wasting disease would be
screened out of prospective studies during the initial physical examination. Any
persons with previously existing fatal wasting disease unintentionally included in
such studies should die within the first two years of follow-up. However, in most
studies, excluding deaths occurring in the first 5 to 10 years has no effect on the
outcome. Finally, it is unlikely that fatal wasting disease is a major cause of
thinness in the general population. Genetic predisposition and the desire to con-
form to cultural and medical standards for body weight are the two most probable
causes of leanness.
Wellness Approach to Obesity 235
57
In the Ontario Longitudinal Study of Aging (Hirdes & Forbes, 1992), mortal-
ity washighest in underweightmen (BMI < 20;RR = 1.00)and lowest inthe over-
weight group (BMI = 25–30; RR = 0.50). Mortality in the severely obese men
remained lowerthan that ofthe underweights (BMI > 30;RR =0.75). Anyrole for
“occult wasting disease” can be ruled out, because only those reporting good or
excellent heath at baseline were included. A representative population sample in
Italy showed lowest mortality in women at a BMI of 32 and in men at a BMI 29,
both in the severely obese range (Seccareccia, Lanti, Menotti, & Scanga, 1998).
Age has a powerful impact on the relationship between body weight and
mortality. Studies of people between the ages of 25 and 45, such as first-time life
insurancebuyersandyoung nurses, tend to show the greatest hazard of obesity(see
Table 2, top two lines). In contrast, among the elderly, low body weight is a strong
risk for subsequent death but there is little hazard attributable to obesity until
extreme levels (BMI > 35) are reached (Andres, Elahi, Tobin, Muller, & Brant,
1985; Stevens et al., 1998; Tayback, Kumanyika, & Chee, 1990). In nearly all of
these studies, subjects with current health problems that could lead to weight loss
were excluded. In most cases, smokers were not included in the study, and mortal-
ity occurring shortly after weight determination was excluded, so as to avoid any
possible influence of “occult wasting disease.” Thus, low body weight is a reliable
harbinger of decline and death in persons over 60 years of age.
The impact of age on the relative hazard of obesity is illustrated in Figure 3,
which shows data from a prospective study of British civil servants that were
adjusted for smoking habit (Jarrett, 1986). In the youngest age group in the study,
those aged 40–44, the mortality rate doubled from the lightest to the heaviest sub-
group. This doubling of risk agrees with a report from the Nurses’ Health Study,
which found an identical increase in women of about the same age (Manson et al.,
1995). The Nurses’ Health Study reported its findings only as relative risk, thereby
emphasizing the twofold elevation in hazard. What was left unstated is the very
low risk of dying—regardless of weight—at these young ages. Expressed differ-
ently, at age 44 you have a 98% chance of living to age 54 if you are lean, and 96%
odds if you are severely obese (see Figure 3). What is considerably more important
from a public health standpoint is the effect of weight on mortality at ages 60–64,
wheremortalityis highest for those who meet lifeinsurancecriteria.Obese persons
face a 2% greater chance than lean individuals of dying between 44 and 54, but a
6%lesserchanceofdyingbetween64and74.Asa result, the net adverse impact of
obesity on median life expectancy is minimal to nonexistent.
The fact that obese persons have a normal life expectancy presents a paradox,
since the incidence of a number of serious risk factors is increased in obesity. The
solution to this puzzle is that there are advantages as well as disadvantages to being
heavy (Ernsberger & Haskew, 1987). Obese persons are less likely to later develop
cancer,as shown in studies in France, Hawaii, Iowa, Massachusetts, Norway, Puerto
Rico, and Scotland, as well as the Hypertension Detection and Follow-Up Program
236 Ernsberger and Koletsky
58
and Seven Countries studies. The obese are also protected against infectious dis-
eases, chronic obstructive pulmonary disease, osteoporosis, mitral valve prolapse,
intermittent claudication, renovascular hypertension, eclampsia, premature birth,
anemia, type-1 diabetes, peptic ulcer, scoliosis, and suicide (Ernsberger & Haskew,
1987). These health benefits of obesity might potentially offset its hazards.
Wellness Approach to Obesity 237
Fig. 3. Mortality rate during 10 years of follow-up as a function or starting BMI in the Whitehall study
(Jarrett, 1986). For purposes of illustration, only the youngest (40–44) and oldest (60–64) groups are
included. Data were adjusted for smoking. Subjects were British civil servants. The dashed line indi-
cates“ideal weight” definedby lifeinsurance standards (BMI = 21),and dotted linemarks the threshold
for the current definition of overweight (BMI > 25).
59
Obesity is also associated with improved survival in several diseases. Heavy
persons with hypertension, type-2 diabetes, and high cholesterol have a more
favorable prognosis than thin people with these same ailments (Ernsberger &
Haskew, 1987). Obese hypertensives have been shown to outlive lean hyper-
tensivesinmorethan 15 separate controlled prospective studies (Barrett-Connor&
Khaw, 1985). In one study, 43% of nonobese hypertensives died, whereas only
26% of obese hypertensives died. In the Systolic Hypertension in the Elderly Pro-
gram, mortality fell by 35% for each 5 kg/m2increment in BMI. Lean type-2 dia-
betics tend to have a more severe form of the disease than obese diabetics (Ross et
al., 1997). Retinopathy is three times more common in lean than in obese type-2
diabetics. Among placebo-treated type-2 diabetics in the University Group Diabe-
tesProgram,those who were obese had lower mortalityduring5years of follow-up
than did the nonobese (2.8% vs. 7.2%).
Obesitymaycushion the risk of highcholesterol.Heavypersons exhibit eleva-
tions in literally every proposed risk factor for atherosclerosis. Despite this, a
comprehensivereviewofthemedical literature came to the conclusion that obesity
is not associated with plaque formation in blood vessels (Barrett-Connor, 1985).
Even among persons weighing more than 300 pounds, no increase in coronary
plaque has been found at autopsy. Because heavy people have elevated risk factors
that fail to translate into a higher incidence of disease, the danger associated with a
given risk factor level must differ between the lean and the obese. For example,
a total cholesterol of 300 is probably not as grave a sign for a fat patient as for a
thin one.
Although hypertension, diabetes, and hyperlipidemia have reduced complica-
tions and mortality in heavy persons, this does not mean these conditions are
benign in obesity, nor does it mean that diabetics and hypertensives should be
encouragedtogainweight, since this may worsen their condition. However, it does
mean that the threat to the health and longevity of fat people posed by diabetes and
hypertension has been overestimated because of the failure to take into account the
ameliorating influence of obesity on these conditions. Thus obesity is associated
with a decreased incidence of a few diseases even though the incidence of many
other ailments is increased among the obese.
Epidemic increases in the incidence of obesity in the United States and around
theworldhavebeen documented in numerous independent reports since the 1980s.
Figure 4 shows time trends in national data broken down by gender and race. Since
1960, there have been progressive increases in average BMI in all subgroups,
paralleling the rising prevalence of obesity. In almost a mirror image pattern,
mortality rates have fallen simultaneously. The causes of death contributing to the
declines in mortality are those usually linked to obesity, especially heart attack and
stroke. If obesity were truly a major contributor to premature death, as some have
claimed (Manson et al., 1995) then we would expect a rising death toll, especially
with the sharp rise in obesity since 1977.
238 Ernsberger and Koletsky
60
An example of short-term time trends in a specific population comes from
Minneapolis-St. Paul, Minnesota. Between 1981 and 1986, the average BMI in
that population rose 1.2 kg/m2in women and 0.6 kg/m2in men (Sprafka, Burke,
Folsom, Luepker, & Blackburn, 1990). Yet during the same time interval and in
the same population the incidence of coronary heart disease fell 13% in women
and 20% in men. Paralleling the decrease in coronary disease were falls in risk
factors: diastolic blood pressure (down 0.9 and 1.1 mm/Hg in women and men,
respectively), and cholesterol (down 5.8 and 5.2 mg/dl). Thus even as human
populations fatten, risk factors and diseases usually linked to obesity continue to
decline. The most likely explanation is that improvements in lifestyle, such as
reduced intakes of cholesterol, saturated fat, and sodium and increased exercise,
have improved health but not led to widespread loss of weight. Also much of the
decline in mortality rates in Western cultures has been credited to improved
detection and treatment of hypertension with improved medications. Hyperten-
sion is the primary health risk faced by the obese, and if adjustment is made for
the higher blood pressures in obese persons, their elevated rates of heart attack
and stroke are eliminated (Keys, 1981). Thus improved treatment of hyperten-
sion in recent years may have eliminated much of the cardiovascular risk associ-
ated with obesity.
Setting aside for the moment the beneficial or ameliorating effects of obesity,
there are a number of ailments, including hypertension, that are more common
amongtheobese.Doesthismean that obesity is a cause of these diseases? The only
way to securely establish this would be to conduct a randomized trial in which one
group of obese persons would undergo weight reduction while a control group
received comparable counseling on nutrition and exercise without specific advice
to lose weight. Any differences in disease incidence or mortality between the
groups could then be attributed to the effect of weight loss. Unfortunately, such a
trial is impossible, because no safe and effective treatment for obesity exists.
What of the mortality risks of underweight? Would a weight gain program
lower the risk of death in excessively lean persons? The answer appears to be
yes. In the case of anorexia nervosa, weight gain is the preferred course of treat-
ment and is life saving. The same is also true for protein-calorie malnutrition.
Thereisalsoevidencethatweightgainis beneficial in lean persons suffering from
neither anorexia nor malnutrition. Elderly patients enrolled in a weight gain
program who increased their weight by at least 5% showed reduced mortality and
extended longevity (Keller, 1995). Successful weight gain also reduced the inci-
dence of falls and recurrent infections. In patients with lung disease having a
BMI of 25 or less, a gain of 2 kg or more reduced mortality (Schols, Slangen,
Volovics, & Wouters, 1998). Thus controlled clinical trials of weight gain in
underweights show decreased mortality, but no equivalent trial exists to show
benefits of weight loss.
Wellness Approach to Obesity 239
61
Fig. 4a. Time trends for obesity and total mortality by race and gender between 1960 and 1994.
Age-adjusted total mortality data are from the National Center for Health Statistics. Mean BMI by
gender and race are from the National Health and Nutrition Examination Surveys (Kuczmarski, Flegal,
Campbell, & Johnson, 1994).
62
Fig. 4b.
63
Markeddiscrepanciesin citation rate can be foundforreviewscovering the haz-
ards of obesity. An outstanding skeptical appraisal of the health hazards of obesity
from the prestigious Annual Reviews in Medicine (Fitzgerald, 1981) has been cited
an average of less than once per year since its publication (0.7 citations a year).
A comprehensive critical review with more than 400 references (Ernsberger &
Haskew, 1987) has been cited only 2.3 times per year. In contrast, reviews offering
muchgrimmerassessments of the health of obese persons, composed by the director
of a hospital weight loss clinic, have been cited at an annual rate of 9.2 (Van Itallie,
1979) and 34.7 (Van Itallie, 1985). Thus articles describing obesity as extremely
hazardous are quoted more extensively in the scholarly literature than equivalent
articles drawing more skeptical conclusions. This trend can also be seen in media
coverage of medical findings, in which extensive attention is devoted to reports of
obesity hazards and scant attention is paid to the extensive contrary evidence.
All of the above discussion and all of the studies listed in Table 2 concern
healthypeople in the general population. In fact, persons with active disease of any
sort are automatically excluded from epidemiological studies. However, in medi-
calpracticeonetreatspatientsafflictedwithdisease. Medical advice to lose weight
is not confined to persons who are otherwise healthy. On the contrary, it is people
whoare ill who are most likely to be counseled to lose weight. We should therefore
askabouttherelationship between body weight and mortality in the unwell. Inhos-
pitalized patients, BMI upon admission was compared to the odds of dying in the
hospital, after adjustment for disease severity (Potter, Schafer, & Bohi, 1988). As
shown in Figure 5, the highest risk of death was in people who were underweight
(BMI < 21). At all ages, the risk of death declined with increasing weight. For
patients aged 50 to 79, the best chance of survival was at a BMI of 40, usually
considered“morbidlyobese.”Forpatients in their 80s, the optimum BMI is around
32, which is considered “severely obese.” The authors of this study considered
whether the excess mortality in nonobese patients was caused by malnutrition or
chronicillness.To test this, they examinedweightchange between hospitalizations
for those who were hospitalized more than once. There was no increase in risk of
death for those who were losing weight prior to hospitalization, indicating that it is
long-term, possibly lifelong, leanness that raises the risk of in-hospital death.
These findings have been independently confirmed in other studies. The Study to
Understand Prognoses and Preferences for Outcomes and Risks of Treatments
(SUPPORT)foundalackofoverweight(BMI<25)tobeanindependentpredictor
of mortality in hospitalized patients after stratified adjustment for 15 different
physiological and demographic variables (Galanos et al., 1997). (The accompany-
ing editorial was entitled “A Mean Outcome for the Lean” [Oud & Haupt, 1997].)
Thus even though the absence of obesity may be favorable for young and healthy
individuals,withtheonsetof acute illness or old age, moderate obesity can become
protective. These results are consistent with the notion that expanded fat stores
facilitate survival when the individual is challenged by illness or aging.
242 Ernsberger and Koletsky
64
Fig.5. RelationshipbetweenBMI onadmission andin-hospital mortality indifferent age groups(Potter,
Schafer,&Bohi, 1988).Mortality isadjusted fordiagnosis andother prognosticfactors. Thedashed line
indicates “ideal weight” defined by life insurance standards (BMI = 21), and the dotted line marks the
threshold for the current definition of overweight (BMI > 25).
65
Socioeconomic Status, Culture, Obesity, and Health Risks
Obesity is strongly related to socioeconomic status (SES; Gortmaker, Must,
Perrin, Sobol, & Dietz, 1993; Jeffery, French, Forster, & Spry, 1991; Lantz et
al., 1998; Morris, Cook, & Shaper, 1992; Sorensen, 1995; Wamala, Wolk, &
Orth-Gomér, 1997). Obesity can be either a cause or a consequence of poverty
(Sorensen, 1995). Discrimination and stigma can result in unemployment or
low-paying work (Gortmaker et al., 1993). Thus it seems that obesity arises first,
then low SES results through societal discrimination and stigma. On the other
hand,theloss of employment increases the risk of weightgain(Morriset al., 1992),
so it is also possible that a decline in SES can presage the development of obesity.
Poverty has been strongly linked to low-quality nutrition, which can result in
weightgainbecauseexcess calories must be consumed to maintainadequateintake
of vital nutrients. Persons of low SES have increased intakes of dietary fat and get
less exercise, both of which can promote weight gain (Jeffery et al., 1991). Both
obesity and low SES are associated with low self-esteem, high job stress and poor
health habits (Wamala et al., 1997).
Low SES is a powerful predictor of death from cardiovascular disease. In a
representativenationalU.S.sample, low income was associated with aRRof3.9 in
womenand3.3 in men (Lantz et al., 1998). After adjustment for cigarette smoking,
alcoholintake,exercise level, and body weight, the RRswere3.8in women and 3.1
inmen.Thusknownriskfactors accounted for only a small portion of the mortality
risk of low SES. Evidence points to psychological stress and limited access to
health care as the primary source of the high risk of premature death in the lower
socioeconomic classes. Low SES is an especially strong predictor of death in
diabetics, presumably because of impaired delivery of necessary medical care
(Rosengren, Welin, Tsipogianni, & Wilhelmsen, 1989).
After controlling for SES in the Americans’ Changing Lives study, the RR
from underweight was 2.03 whereas the RR for overweight was 0.94 (Lantz et al.,
1998). Thus in comparing individuals with equivalent social status, there is no sig-
nificant increase in risk of death from high body mass. In the National Health
Examination Follow-Up Survey, mortality rates were higher in obese persons
(BMI > 30) in poverty compared to the nonpoor obese, whereas poverty had a
much smaller adverse effect on survival for leaner persons (Tayback et al., 1990).
Similarly, a study in Finland foundhigh risks associated with obesity(BMI > 34)
in the upper classes (RR = 1.4 to 1.7) but not in lower social classes (RR = 1.2;
Rissanen et al., 1989). Also, underweight (BMI < 19) was not a hazard in the upper
classes(RR= 1.1) but was a strongriskinlower classes (RR = 1.9). Thecontrasting
risks of overweight and underweight in high-SES and low-SES groups might
account for the discrepancies between different epidemiological studies.
Several epidemiological studies were controlled for SES because they used
defined occupational groups (Table 2). The 10-year San Francisco longshoremen
244 Ernsberger and Koletsky
66
study, for example, showed lower mortality in nonsmokers who were markedly
obese (BMI > 29; 6.0% died) than in those who were of “ideal weight” (BMI < 22;
10.2% died; Borhani, Hechter, & Breslow, 1963). An influence of “occult wasting
disease” is unlikely, because all the subjects were involved in physically demand-
ing work. Almost identical results were obtained for nonsmoking employees of
People’s Gas in Chicago: In 14 years, 20.4% of the “ideal weight” (BMI < 23.5)
workers died, whereas 16.9% of the markedly obese (BMI > 29) men died (Dyer,
Stamler, Berkson, & Lindberg, 1975). The Whitehall Study in England also
showedaprotectiveeffectofmoderateoverweightonmortalityinBritishcivilser-
vants (Jarrett, Shipley, & Rose, 1982). A comparable study of Paris civil servants
showedthehighest mortality in those who conformed to thelatestweightstandards
(BMI < 24.4), whereas mortality was 34% lower in those who were overweight
(BMI of 24.4 to 27) and 31% lower in those who were obese (BMI > 30;
Filipovsky, Ducimetière, Darné, & Richard, 1993). Adjusting the relative risks for
smoking and removing the deaths in the first 5 or 10 years of follow-up did not
remove the apparent protective effect of obesity. In each of these studies of
employment groups, those participants who registered the low body weights
associatedwithincreaseddeath rates were not emaciated by any means, but wereat
or even slightly above weights recommended by the life insurance tables. The use
of healthy, actively employed persons rules out any possibility of weight loss due
to chronic wasting disease. The inescapable conclusion is that underweight is a
significant contributor to excess mortality, and may be equally important to over-
weight in its impact on longevity in the population.
The studies of body mass and mortality listed in Table 2 are arranged roughly
in order of the degree of hazard found to be associated with obesity. A striking
relationship, which has not been commented on in any previous review, is that the
five studies showing a high hazard of obesity have subjects drawn from the high-
est SES. This includes suburban nurses, Harvard alumni, persons who can afford
individual life insurance policies, residents of an affluent suburb (Framingham,
Massachusetts),andothers.Theseare social groups in which the stigma of obesity
wouldbehighest.Nurses,inparticular,areunderstrongpressuretoloseweightas
a healthful example to their patients. In contrast, study populations showing no
risk of obesity or even an inverse relationship between body weight and mortality
tendtobeof low SES and to belong tonon-Westernculturesthatdo not value thin-
ness as strongly as affluent Americans. The incidence of obesity was very low in
the studies of Harvard alumni, nurses, and insurance policyholders. The average
BMI of all these groups was well below the U.S. average. In contrast, obesity was
prevalent in the populations where it was perceived as neutral or even advanta-
geous. The world’s highest rates of obesity are found in the South Pacific and in
Native American groups. Yet in these non-Western cultures, there is no increased
risk of cardiovascular disease even with extreme obesity (BMI of up to 40). This
raises the possibility that the stigma, discrimination, and stress faced by obese
Wellness Approach to Obesity 245
67
persons of high SES in Western cultures may be a major contributor to cardiovas-
cular mortality.
The possible relationships between obesity, SES, and health are schematized
in Figure 6. Obesity can lead to low SES through social stigma and the process
of discrimination. Obesity is more common among minority groups, including
Native Americans, Blacks, Hispanics, and Jews, further magnifying discrimina-
tion. Low SES can also contribute to weight gain and obesity by limiting exercise
opportunities and diminishing nutritional quality. These factors can promote poor
health at the same time as they contribute to obesity. Low SES is a strong indepen-
dentpredictor of poor health and premature death, probably through psychological
stress and reduced access to health care.
Weight Cycling and Obesity-Associated Illness
A further difference between the populations at the top and the bottom of
Table 2 is the likely prevalence of dieting and participation in weight loss
programs. High SES is associated with increased efforts to lose weight (Rodin,
1993). The relative hazard of obesity is higher in younger age groups, which are
also most likely to pursue weight loss. Weight loss methods may be hazardous in
themselves, as reviewed elsewhere in this issue. Nurses in particular are exposed
to opportunities for pharmaceutical, surgical, and very-low-calorie diet interven-
tions, which can be hazardous. There is also strong evidence that the regain of
weightthatalmostalways follows successful weight loss can be harmful. Thusthe
harmful effects of weight cycling may contribute to the high mortality risk of obe-
sity in populations in which weight loss is prevalent, in contrast to the low
246 Ernsberger and Koletsky
Fig. 6. Hypothetical model for the relationship between SES, obesity, and adverse health outcomes.
68
mortalityriskinpopulations in which obesity is the norm and likely tobeaccepted
or tolerated.
Fatalconsequencesofrefeedingafterstarvationarewellknowninfaminevic-
tims and anorexics (Weinsier & Krumdieck, 1981). Obese dieters may not be
immune to the adverse effects of refeeding after significant weight loss (Brownell
& Rodin, 1994). Obese humans typically show repeated loss and regain of large
amounts of weight (Blackburn et al., 1989; Rodin, Radke-Sharpe, Rebuffë-Scrive,
&Greenwood,1990). Men with large fluctuationsinweightbetween the ages of 20
and 40 have increased systolic and diastolic blood pressure and cholesterol
(Hamm,Shekelle,&Stamler,1989). These yo-yo dieters are two times more likely
to die of coronary heart disease, even after adjustment for known risk factors, than
aremenwithstableor steadily increasing weight (Hamm et al., 1989; Lissner et al.,
1991).Fluctuationsinbody weight have been shown in many other major epidemi-
ological studies to have deleterious cardiovascular effects resulting in increased
mortality (Blair, Shaten, Brownell, Collins, & Lissner, 1993; Ernsberger &
Koletsky, 1993; Lissner et al., 1991).
A few studies have not found that weight cycling increases the risk of mortal-
ity. However, the populations lacking a risk associated with weight cycling tend to
be those with low rates of dieting for weight loss, such as a study of elderly men in
Baltimore (Lissner, Andres, Muller, & Shimokata, 1990). The lack of an effect of
weight cycling in a population with a low level of intentional weight loss implies
that minor random fluctuations in weight do not pose a health hazard.
Animal studies strongly support the existence of adverse effects of weight
cycling (Ernsberger & Koletsky, 1993; Ernsberger & Nelson, 1988b; Ernsberger
etal.,1994; Ernsberger, Koletsky, Baskin, &Collins,1996; Koletsky et al., 1995).
Weight cycling induces hypertension, enlargement of the heart, thickening of the
heart wall, increased levels of the stress hormones adrenaline and noradrenaline,
progressive kidney damage, redistribution of fat deposits to the abdomen,
increased efficiency of weight gain, and exacerbation of obesity. The last effect,
progressive weight gain, was found only in genetically obese animals, but not in
naturally lean animals. This implies that a genetic susceptibility may determine
whether cycles of weight loss and regain lead to progressive accumulation of fat,
possibly in humans as well as in rats.
In the Framingham study, the increase in cardiovascular disease in obese
patients could be entirely explained by taking two facts into account (Lissner et al.,
1991). First, obese people were more like to go up and down in weight than thin
people. Second, weight cycling was associated with increased rates of death from
cardiovasculardisease.Obesepeople who maintained a high but steady weight had
only an average risk of death or cardiovascular disease. This finding raises the
possibility that much of increased risk of disease and death in obese people is the
result of repeated cycles of weight loss and regain.
Wellness Approach to Obesity 247
69
Bias, Conflicts of Interest, and Medical Beliefs About Obesity
Given the number of studies with contrary findings, why the emphasis on the
few studies that show a strong upward trend of mortality with increasing body
weight? One answer is selective citation. Other factors being equal, studies report-
ing adverse effects of obesity are quoted more frequently than those that are more
neutral. Citation analysis can document this. Let us take two articles that appeared
inthesameissue of Annals of Internal Medicine aspartoftheproceedings of a con-
ference convened by Congress. The first article reviewed more than 100 prospec-
tive epidemiological studies and concluded that there was no correlation between
obesity and heart disease (Barrett-Connor, 1985). The second article presented a
long list of diseases associated with obesity based on cross-sectional studies and
made the case for adverse health effects of obesity (Van Itallie, 1985). The
antiobesity article was cited an average of 34.7 times per year through 1996,
according to the Science Citation Index, whereas the neutral article was cited an
average of 9.6 times per year. The lighter citation of the first article cannot be
attributed to lower academic standing, because the lead author, Elizabeth
Barrett-Connor, is a well-known epidemiologist with more than 339 publications
in major medical journals, whereas the second author, Theodore Van Itallie, was
thedirectorofaweightlossclinic at Presbyterian-St. Luke’s Hospital in New York
City at the time of the conference.
Another example of selective citation can be found in two reports, both
appearingintheJournaloftheAmericanMedicalAssociationandbothreporting
results of the Framingham study. The first report, entitled simply “Body Build
and Mortality: The Framingham Study,” showed an increased risk of death in
underweight but not in overweight (Sorlie, Gordon, & Kannel, 1980) and was
accompanied by an editorial entitled “Beware the Lean and Hungry Look.” The
second article in the same journal, reporting on the very same study, reached the
opposite conclusion, stating that underweight persons had the lowest risk and
every increment in body weight increased mortality risk (Garrison, Feinleib,
Castelli, & McNamara, 1983). The second report made no attempt to explain the
discrepancy with the previous report on the same group of Framingham
residents. However, the first report covered 5,146 men and 6,829 women. The
second report covered 1,976 men and no women. No rationale was given for
excluding more than 3,000 men and all of the women from this reanalysis, but
obviously the selective inclusion of Framingham subjects could account for the
reversalofdatatrendsin the second report. Even though it included only one fifth
the number of subjects, the second Journal of the American Medical Association
report has been cited 7.9 times per year through 1996, whereas the original report
has been cited at an annual rate of 5.5 times. Citing authors show a clear prefer-
ence for articles that assign a high risk to obesity, regardless of journal stature or
data quality.
248 Ernsberger and Koletsky
70
The National Institutes of Health frequently convene panels of experts to dis-
cuss important and controversial issues in medicine and to arrive at compromise or
consensus statements that are almost universally agreed upon. This process has
worked well for many topics, as experts representing a spectrum of opinion are
brought together and reach agreement on the current state of knowledge in their
field.Amajorproblemwithconsensuspanels, however, is that many of the experts
represent special interest groups (Tong, 1991). This is clearly a problem with NIH
panels on obesity, on which the multibillion-dollar interests of the weight loss
industry have been well represented.
Beforeexaminingthe specific issues, we mustreviewthe possible conclusions
that favor the fortunes of the weight loss industry. These are summarized in Table
3. Weight loss clinics, pharmaceutical firms developing and marketing diet pills,
and diet industry concerns represent an enormous economic interest. Diet industry
sales alone, excluding medical clinics, amount to more than $60 billion a year, as
reviewed elsewhere in this issue. These economic interests are favored by procla-
mations from governmental panels identifying obesity as a major health risk,
establishing a very low limit for the definition of obesity, and minimizing the
hazards of treatment. The economic interest of pharmaceutical firms is illustrated
by a major grant program initiated by Knoll Pharmaceuticals, manufacturers of the
diet pill sibutramine. In a letter addressed to physicians across the nation, the
company offered generous grant support “to advance the understanding of obesity
as a major health problem.” Knoll and Wyeth-Ayerst have sponsored continuing
education programs across the country with the aim of promoting “awareness” of
the hazards of obesity.
Remarkably, each of the precepts in Table 3 has been adhered to precisely by
each NIH panel, as detailed below. The congruence between the interests of the
diet industry and the content of NIH statements does not establish undue influence
of industry over governmental deliberations, but it does suggest the need for
reform, particularly of the manner in which expert panels are assembled.
Wellness Approach to Obesity 249
Table 3. Medical Precepts Favoring the Growth and Profitability of the Weight-Loss Industry
Precept Outcome
Exaggeration of the ill effects of obesity Facilitation of third-party payments; increased
motivation to seek weight control services
Setting body weight standards as
low as possible Expansion of client base
Overstatement of the long-term
benefits of weight loss Increased utilization and reimbursement for
services
Minimization of the ill effects of obesity
treatments and weight cycling Repeat utilization of services
71
In 1979, a panel of weight loss clinic directors and surgeons, chaired by
Theodore Van Itallie, was selected by the NIH to evaluate surgical methods of
weight loss. The panel gave a strong endorsement to intestinal bypass surgery for
weight loss (Van Itallie & Burton, 1979). However, by 1979, severe and even fatal
complications of intestinal bypass were known (Halverson, Scheff, Gentry, &
Alpers, 1980). Intestinal bypass was virtually abandoned shortly after the panel
endorsement was published.
Describing themselves as “an impartial panel of 15 senior level professionals”
(Burton, Foster, Hirsch, & Van Itallie, 1985), an NIH panel was convened in 1985
to consider the question of whether obesity constituted a health risk. Nineteen
experts on obesity testified to the panel. A remarkable aspect of the consensus
statement is that it almost completely contradicted the conclusions of the experts
who testified to the panel. A list of these contradictions is provided in Table 4. In
each case, the conclusions of the panel favored the weight loss industry (compare
Tables 3 and 4). The deliberations might have been affected by the fact that both
thepanelchairand the chair of the planning committee directed weight lossclinics.
When confronted with evidence of systematic bias, the panel chair would only
state that the allegation “exceeds the bounds of scientific collegiality” (Hirsch,
1987).
250 Ernsberger and Koletsky
Table 4. Contrasts Between Evidence Presented and Conclusions Reached: 1985 NIH Panel
Evidence presented
(Ernsberger, 1987) Panel conclusion
(Burton, Foster, Hirsch, & Van Itallie, 1985
Excess lean body mass is more hazardous
than excess fat, because nonobese
overweights have the most risk factors
(Van Itallie, 1985; Figure 5 therein)
Refers to body weight and body fat synonymously
Insurance height-weight tables derived
improperly from faulty data (Harrison, 1985) Endorsed insurance tables and called for those
with common ailments to reduce even further
Poor diet and lack of exercise may be primary
causes of disease, with obesity only a
bystander (Stallones, 1985)
Identified obesity as a direct cause of disease and
a disease unto itself
Genetic factors lead to disease and to obesity,
with obesity only a bystander (Iverius &
Brunzell, 1985; Stallones, 1985)
Genetic factors not acknowledged
Epidemiologic risk of obesity in children is
highly uncertain (Johnston, 1985) Children must “bring their weight within
normal limits”
Obesity is not related to atherosclerosis
(Barrett-Connor, 1985) Obesity causes atherosclerosis
Obesity is actually protective in the Seven
Countries Study (Kluthe & Schubert, 1985) Findings dismissed because study groups are
“not representative of the U.S. population”
Weight standards should be adjusted for age
(Andres, Elahi, Tobin, Muller, & Brant, 1985) Weight standards based on insurance customers at
age 25. No adjustment for age.
72
The NIH constituted a panel of obesity experts to serve as a standing expert
panel on obesity, known as the National Task Force for the Prevention and Treat-
ment of Obesity. The composition of this governmental panel is shown in Table 5.
TheTaskForce rendered a series of official statements, each published in the Jour-
nal of the American Medical Association. The first statement covered very-
low-caloriediets(VLCDs),referring to liquid diets such as Optifast (provided only
by doctors) and Slim-Fast (over-the-counter) that provide 300–500 calories a day
(NIH National Task Force, 1993). The panel considered these to be both safe and
effective, but only under medical supervision. This position, of course, would
strengthen the position of hospital-based weight loss programs, which offer liquid
diets under medical supervision. However, the endorsement by the Task Force of
liquid diets as a medical intervention came as the Optifast fad, fostered by Oprah
Winfrey’s temporary weight loss, was fading and most hospitals had terminated
their liquid diet programs, including Mt. Sinai Hospital of Cleveland, where
Optifast was first formulated.
The second statement concerned weight cycling (NIH National Task Force,
1994).Theconclusion,widely repeated in the media, was that weight cycling isnot
harmful and the increased mortality found in persons who repeatedly lose and
regain weight should not deter anyone from joining a weight loss program. The
overwhelming evidence that weight cycling promotes cardiovascular disease and
death is reviewed in the previous section. Recognition of the hazards of repeated
loss and regain of weight is a serious threat to the weight loss industry, however, as
they depend on repeat business for their survival. Given the high prevalence of
dieting, if only persons who had never dieted before were encouraged to join
weight loss programs, the industry would collapse overnight.
The most recent statement from the Task Force concerned long-term treat-
ment of obesity with drugs (NIH National Task Force, 1996). The panel endorsed
the use of diet drugs, such as fenfluramine and phentermine, for periods of over 1
year or even for life. The report minimized the concerns over serious and fatal side
effects from these drugs, because these risks “must be viewed in the context of the
major excess in morbidity and mortality attributable to obesity.” A few months
after the publication of the panel’s conclusions, fenfluramine was withdrawn from
themarketbecause of incidences of severeheartdamage and a number offatalities.
In each of its three reports, the Task Force issued opinions that were favorable
to weight loss clinics and to pharmaceutical firms marketing diet pills. Could the
conflicts of interest listed in Table 5 have played a role? As we pointed out in a
letter to Journal of the American Medical Association, seven of the nine members
of Task Force were directors of weight loss clinics (Ernsberger & Koletsky, 1995).
A more complete picture of the conflicts of interest of Task Force members was
disclosed as part of their subsequent report (NIH National Task Force, 1996).
None of the arrangements or relationships listed in Table 5 are illegal in any
way. However, the remarkably consistent position taken by the Task Force on
Wellness Approach to Obesity 251
73
Table 5. Conflicts of Interest Among Members of the National Task Force
on the Prevention and Treatment of Obesity
Scientific
member Affiliation Employment
conflict Other conflict
R. L. Atkinson,
MD University of
Wisconsin,
Madison
Weight loss clinic;
performed surgery
for weight loss
Knoll Pharmaceuticals
W. H. Dietz,
MD, PhD Tufts University
School of Medicine Weight loss clinic Knoll Pharmaceuticals;
Hoffman-LaRoche
J. P. Foreyt, PhD Baylor College of
Medicine Nutrition
Research Clinic
Behavior
modification
weight loss
program
Advisor to Calorie Control
Council (industry group);
author of popular weight
loss book (Warner Press)
N. J. Goodwin, MD HEALTH WATCH
Promotion Service Unclear Unknown
J. O. Hill, PhD University of
Colorado, Denver Behavior
modification
weight loss
program; consultant
to Duke Diet and
Fitness Center
Advisor to Calorie Control
Council (industry group);
Proctor & Gamble; Amgen;
Knoll Pharmaceuticals;
Hoffman-LaRoche;
International Life Sciences
Institute (industry group)
J. Hirsch, MD
(chaired NIH
consensus panel)
Rockefeller
University Research weight
loss clinic Hoffman-LaRoche;*
Nutrasweet Co;* Diet Center;
Weight Watchers;
Millennium Co.
F. X. Pi-Sunyer,
MD Columbia
University Weight loss clinic Scientific Advisory Boards of
Weight Watchers International
Inc., American Home Products,
Wyeth-Ayerst and Knoll
Pharmaceuticals; Executive
Director, Weight Watchers
Foundation; Consultant to
Hoffman-LaRoche, Genentech,
Eli Lilly; Neurogen; Parke-Davis
R. L. Weinsier,
MD, DrPH University of
Alabama,
Birmingham
Weight loss clinic Scientific Advisory Board,
Weight Watchers International
Inc.; Sandoz Nutrition
(makers of Optifast)
R. Wing, PhD University of
Pittsburgh Behavior
modification
weight loss
program
Scientific Advisory Board,
Weight Watchers International
Inc.; Eli Lilly; Ross Laboratories;
International Life Sciences
Institute (industry group)
*Conflict disclosed in NIH National Task Force, 1994. All other conflicts undisclosed prior to 1996.
74
matters that affect the weight loss industry raises questions about the NIH’s selec-
tion of panelists. Furthermore, the Task Force endorsed liquid diet products and
diet pills when members had accepted payments from the manufacturers of such
products. The Task Force is a true governmental body and forms a part of the NIH.
Anarmof the NIH, the WeightInformationNetwork,distributes copies of the Task
Force reports and other writings of the panel members. It also distributes pam-
phlets, at taxpayer expense, for the lay public, with strongly worded warnings
about the dangers of exceeding body weight guidelines. Other pamphlets dismiss
or minimize concerns dieters might have about weight cycling, development of
gallbladder disease as a result of dieting, or side effects of medication. The under-
lying message from the Task Force and the NIH itself is that the $60 billion a year
investedinthepursuitofthinnessintheUnitedStatesistoolittle,andconsiderably
more money and attention needs to be devoted to weight control. We ask whether
the money and effort expended on the generally unsuccessful pursuit of thinness
might be better spent on directly promoting lifestyle change.
References
Abenhaim, L., Moride, Y., Brenot, F., Rich, S., Benichou, J., Kurz, X., Higenbottam, T., Oakley, C.,
Wouters, E., Aubier, M., Simonneau, G., & Begaud, B. (1996). Appetite-suppressant drugs and
therisk ofprimary pulmonaryhypertension. NewEngland Journalof Medicine,335, 609–616.
Akram, D.-S., Astrup, A. V., Atinmo, T., Boisson, J.-L., Bray, G. A., Carroll, K. K., Chunming, C.,
Chitson, P., Dietz, W. H., Hill, J. O., Jequier, E., Komodiki, C., Matsuzawa, Y., Mollentze, W. F.,
Moosa, K., Noor, M. I., Reddy, K. S., Seidell, J., Tanphaichitr, V., Uauy, R., & Zimmet, P.
(1997). Obesity: Preventing and managing the global epidemic. Report of a WHO consultation
on obesity. Geneva, Switzerland: World Health Organization.
Andersen, R. E., Wadden, T. A., Bartlett, S. J., Vogt, R. A., & Weinstock, R. S. (1995). Relation of
weight loss to changes in serum lipids and lipoproteins in obese women. American Journal of
Clinical Nutrition, 62, 350–357.
Andres,R. (1980).Effect ofobesity ontotal mortality.International Journalof Obesity,4(4), 381–386.
Andres,R., Elahi, D., Tobin, J.D., Muller,D. C., & Brant, L.(1985). Impactof age onweight goals.An-
nals of Internal Medicine, 103(6, part 2), 1030–1033.
Bale, G. S., & Entmacher, P. S. (1977). Estimated life expectancy of diabetics. Diabetes, 26(5),
434–438.
Barnett, A. H., Eff, C., Leslie, R. D., & Pyke, D. A. (1981). Diabetes in identical twins: A study of 200
pairs. Diabetologia, 20(2), 87–93.
Barrett-Connor, E. L. (1985). Obesity, atherosclerosis, and coronary artery disease. Annals of Internal
Medicine, 103(6, part 2), 1010–1019.
Barrett-Connor, E. L., & Khaw, K. T. (1985). Is hypertension more benign when associated with
obesity? Circulation, 72(1), 53–60.
Beckmann, S. L., Os, I., Kjeldsen, S. E., Eide, I. K., Westheim, A. S., & Hjermann, I. (1995). Effect of
dietary counselling on blood pressure and arterial plasma catecholamines in primary hyperten-
sion. American Journal of Hypertension, 8(7), 704–711.
Bennett, P. H. (1986). More about obesity and diabetes [Letter to the editor]. Diabetologia, 29(10),
753–754.
Blackburn, G. L., Wilson, G. T., Kanders, B. S., Stein, L. J., Lavin, P. T., Adler, J., & Brownell, K. D.
(1989). Weight cycling: The experience of human dieters. American Journal of Clinical
Nutrition, 49(Suppl.), 1105–1109.
Wellness Approach to Obesity 253
75
Blair, S. N., Shaten, J., Brownell, K., Collins, G., & Lissner, L. (1993). Body weight change, all-cause
mortality, and cause-specific mortality in the multiple risk factor intervention trial. Annals of
Internal Medicine, 119, 749–757.
Borhani, N. O., Hechter, H. H., & Breslow, L. (1963). Report of a 10-year follow-up study of the San
Francisco longshoremen. Journal of Chronic Diseases, 16, 1251–1266.
Bray, G. A. (1996). Health hazards of obesity. Endocrinology and Metabolism Clinics of North
America, 25(4), 907–919.
Brenner, H., Arndt, V., Rothenbacher, D., Schuberth, S., Fraisse, E., & Fliedner, T. M. (1997). Body
weight, pre-existing disease, and all-cause mortality in a cohort of male employees in the
German construction industry. Journal of Clinical Epidemiology, 50(10), 1099–1106.
Brownell, K. D., & Rodin, J. (1994). Medical, metabolic, and psychological effects of weight cycling.
Archives of Internal Medicine, 154, 1325–1330.
Buñag, R. D., & Barringer, D. L. (1988). Obese Zucker rats, though still normotensive, already have
impaired chronotropic baroreflexes. Clinical and Experimental Hypertension, 10(Suppl. 1),
257–262.
Buñag, R. D., Krizsan, D., & Itoh, H. (1990). Diminished cardiovascular responsiveness to vagal
stimulation in obese rats. American Journal of Physiology, 259, R842–R848.
Burton, B. T., Foster, W. R., Hirsch, J., & Van Itallie, T. B. (1985). Health implications of obesity: An
NIH Consensus Development Conference. International Journal of Obesity, 9, 155–170.
Charles, M. A., Pettitt, D. J., Saad, M. F., Nelson, R. G., Bennett, P. H., & Knowler, W. C. (1993).
Development of impaired glucose tolerance with or without weight gain. Diabetes Care, 16,
593–596.
Chaturvedi, N., Fuller, J. H., & the WHO Multinational Study Group. (1995). Mortality risk by body
weight and weight change in people with NIDDM: The WHO Multinational Study of Vascular
Disease in Diabetes. Diabetes Care, 18, 766–774.
Ciliska,D., Kelly,C., Petrov, N., & Chalmers, J. (1995).A reviewof weightloss interventionsfor obese
people with non-insulin-dependent diabetes mellitus. Canadian Journal of Diabetes Care, 19,
10–15.
Collins, V. R., Dowse, G. K., Cabealawa, S., Ram, P., & Zimmet, P. Z. (1998). High mortality from
cardiovascular disease and analysis of risk factors in Indian and melanisian Fijians. Inter-
national Journal of Epidemiology, 25, 59–69.
Contreras,R. J., &King, S.(1989). High fat/sucrosefeeding attenuatesthe hypertension ofspontaneously
hypertensive rats. Physiology and Behavior, 46, 285–291.
Crews, D. E. (1989). Multivariate prediction of total and cardiovascular mortality in an obese Polyne-
sian population. American Journal of Public Health, 79(8), 982–986.
Dengel, J. L., Katzel, L. I., & Goldberg, A. P. (1995). Effect of an American Heart Association diet,
with or without weight loss, on lipids in obese middle-aged and older men. American Journal of
Clinical Nutrition, 62, 715–721.
Dyer, A. R., Stamler, J., Berkson, D. M., & Lindberg, H. A. (1975). Relationship of relative weight and
body mass index to 14-year mortality in the Chicago People’s Gas Company study. Journal of
Chronic Diseases, 28(2), 109–123.
Edelstein, S. L., Knowler, W. C., Bain, R. P., Andres, R., Barrett-Connor, E. L., Dowse, G. K., Haffner,
S.M., Pettitt, D.J., Sorkin, J.D., Muller, D.C., Collins, V. R., & Hamman, R. F.(1997). Predic-
tors of progression from impaired glucose tolerance to NIDDM: An analysis of six prospective
studies. Diabetes, 46(4), 701–710.
Ernsberger, P. (1987). NIH Consensus Conference on Obesity: By whom and for what? Journal of
Nutrition, 117, 1164–1165.
Ernsberger, P., & Haskew, P. (1987). Health implications of obesity: An alternative view. Journal of
Obesity and Weight Regulation, 6, 55–137.
Ernsberger, P., & Koletsky, R. J. (1993). Weight cycling and mortality: Support from animal studies.
Journal of the American Medical Association, 269, 1116.
Ernsberger,P., &Koletsky, R. J. (1995). Weightcycling. Journalof theAmerican MedicalAssociation,
273, 998–999.
Ernsberger, P., Koletsky, R. J., Baskin, J. Z., & Collins, L. A. (1996). Consequences of weight cycling
in obese spontaneously hypertensive rats. American Journal of Physiology: Regulatory, Inte-
grative and Comparative Physiology, 270, R864–R872.
254 Ernsberger and Koletsky
76
Ernsberger, P., Koletsky, R. J., Baskin, J. Z., & Foley, M. (1994). Refeeding hypertension in obese
spontaneously hypertensive rats. Hypertension, 24, 699–705.
Ernsberger, P., Koletsky, R. J., Collins, L. A., & Bedol, D. (1996). Sympathetic nervous system in
salt-sensitive and obese hypertension: Amelioration of multiple abnormalities by a central
sympatholytic agent. Cardiovascular Drugs and Therapeutics, 10(Suppl. 1), 275–282.
Ernsberger, P., & Nelson, D. O. (1988a). Effects of fasting and refeeding on blood pressure are deter-
mined by nutritional state, not by body weight change. American Journal of Hypertension, 1,
153S–157S.
Ernsberger, P., & Nelson, D. O. (1988b). Refeeding hypertension in dietary obesity. American Journal
of Physiology, 254, R47–R55.
Everhart, J. E. (1993). Contributions of obesity and weight loss to gallstone disease. Annals of Internal
Medicine, 119, 1029–1035.
Felson, D. T. (1996). Weight and osteoarthritis. American Journal of Clinical Nutrition, 63(Suppl.),
430S–432S.
Feskens, E. J., Virtanen, S. M., Rasanen, L., Tuomilehto, J., Stengard, J., Pekkanen, J., Nissinen, A., &
Kromhout, D. (1995). Dietary factors determining diabetes and impaired glucose tolerance: A
20-year follow-up of the Finnish and Dutch cohorts of the Seven Countries Study. Diabetes
Care, 18(8), 1104–1112.
Filipovsky, J., Ducimetière, P., Darné, B., & Richard, J. L. (1993). Abdominal body mass distribution
and elevated blood pressure are associated with increased risk of death from cardiovascular
diseases and cancer in middle-aged men: The results of a 15- to 20-year follow-up in the Paris
prospective study I. International Journal of Obesity, 17, 197–203.
Fitzgerald, F. T. (1981). The problem of obesity. Annual Review of Medicine, 32, 221–231.
Flegal,K. M.,Carroll, M. D., Kuczmarski, R. J., &Johnson, C.L. (1998).Overweight andobesity inthe
United States: Prevalence and trends, 1960–1994. International Journal of Obesity, 22(1),
39–47.
Galanos, A. N., Pieper, C. F., Kussin, P. S., Winchell, M. T., Fulkerson, W. J., Harrell, F. E. Jr., Teno, J.
M., Layde, P., Connors, A. F. Jr., Phillips, R. S., & Wenger, N. S. (1997). Relationship of body
mass index to subsequent mortality among seriously ill hospitalized patients. Critical Care
Medicine, 25(12), 1962–1968.
Garn, S. M., Bailey, S. M., & Block, W. D. (1979). Relationships between fatness and lipid level in
adults. American Journal of Clinical Nutrition, 32(4), 733–735.
Garn, S. M., Hawthorne, V. M., Pilkington, J. J., & Pesick, S. D. (1983). Fatness and mortality in the
West of Scotland. American Journal of Clinical Nutrition, 38(2), 313–319.
Garrison, R. J., Feinleib, M., Castelli, W. P., & McNamara, P. M. (1983). Cigarette smoking as a
confounder of the relationship between relative weight and long-term mortality: The
FraminghamHeart Study. Journalof theAmerican MedicalAssociation, 249(16),2199–2203.
Gortmaker, S. L., Must, A., Perrin, J. M., Sobol, A. M., & Dietz, W. H. (1993). Social and economic
consequences of overweight in adolescence and young adulthood. New England Journal of
Medicine, 329, 1008–1012.
Groop, L. C., & Tuomi, T. (1997). Non-insulin-dependent diabetes mellitus—A collision between
thrifty genes and an affluent society. Annals of Medicine, 29(1), 37–53.
Halverson, J. D., Scheff, R. J., Gentry, K., & Alpers, D. H. (1980). Jejunoileal bypass: Late metabolic
sequelae and weight gain. American Journal of Surgery, 140(3), 347–350.
Hamm, P., Shekelle, R. B., & Stamler, J. (1989). Large fluctuations in body weight during young adult-
hood and twenty-five-year risk of coronary death in men. American Journal of Epidemiology,
129, 312–318.
Hanson, R. L., McCance, D. R., Jacobsson, L. T. H., Narayan, K. M. V., Nelson, R. G., Pettitt, D. J.,
Bennett,P. H., & Knowler, W.C. (1995).The U-shaped association between bodymass index
and mortality: Relationship with weight gain in a Native American population. Journal of
Clinical Epidemiology, 48, 903–916.
Harrison, G. G. (1985). Height-weight tables. Annals of Internal Medicine, 103(6, part 2), 989–994.
Higgins, M., Kannel, W., Garrison, R., Pinsky, J., & Stokes, J. (1988). Hazards of obesity—The
Framingham experience. Acta Medica Scandinavica, 723(Suppl.), 23–36.
Hill, A. B. (1966). Principles of Medical Statistics (8th ed.). Oxford, UK: Oxford University Press.
Wellness Approach to Obesity 255
77
Hirdes, J. P., & Forbes, W. F. (1992). The importance of social relationships, socioeconomic status and
health practices with respect to mortality among healthy Ontario males. Journal of Clinical
Epidemiology, 45, 175–182.
Hirsch, J. (1987). Reply to letter of Dr. Ernsberger. Journal of Nutrition, 117, 1166.
Hodge, A. M., Dowse, G. K., Collins, V. R., & Zimmet, P. Z. (1996). Mortality in Micronesian
Nauruansand Melanesianand IndianFijians isnot associatedwith obesity.American Journal of
Epidemiology, 143, 442–455.
Incidence and prediction of coronary heart disease in two Italian rural population samples followed-up
for 20 years. (1982). Acta Cardiologia, 37(2), 129–145.
Iverius, P. H., & Brunzell, J. D. (1985). Obesity and common genetic metabolic disorders. Annals of
Internal Medicine, 103(6, part 2), 1050–1051.
Jarrett, R. J. (1986). Is there an ideal body weight? British Medical Journal, 293(6545), 493–495.
Jarrett, R. J., Shipley, M. J., & Rose, G. (1982). Weight and mortality in the Whitehall Study. British
Medical Journal, 285(6341), 535–537.
Jeffery, R. W., French, S. A., Forster, J. L., & Spry, V. M. (1991). Socioeconomic status differences in
health behaviors related to obesity: The Healthy Worker Project. International Journal of Obe-
sity, 15, 689–696.
Johnston, F. E. (1985). Health implications of childhood obesity. Annals of Internal Medicine, 103(6,
part 2), 1068–1072.
Jung, R. T., Shetty, P. S., & James, W. P. (1980). Nutritional effects on thyroid and catecholamine
metabolism. Clinical Science, 58(3), 183–191.
Keller, H. H. (1995). Weight gain impacts morbidity and mortality in institutionalized older persons.
Journal of the American Geriatrics Society, 43, 165–169.
Keys, A. (1981). Overweight, obesity, coronary heart disease, and mortality: The W. O. Atwater
Memorial Lecture, 1980. Progress in Clinical and Biological Research, 67, 31–46.
Keys, A. (1989). Longevity of man: Relative weight and fatness in middle age. Annals of Medicine, 21,
163–168.
Kluthe, R., & Schubert, A. (1985). Obesity in Europe. Annals of Internal Medicine, 103(6, part 2),
1037–1042.
Koletsky, R. J., Boccia, J., & Ernsberger, P. (1995). Acceleration of renal disease in obese SHR by
exacerbation of hypertension. Clinical and Experimental Pharmacology and Physiology, 22,
S254–S256
Kuczmarski, R. J., Flegal, K. M., Campbell, S. M., & Johnson, C. L. (1994). Increasing prevalence of
overweight among US adults: The National Health and Nutrition Examination Surveys, 1960 to
1991. Journal of the American Medical Association, 272, 205–211.
Landsberg, L., & Young, J. B. (1978). Fasting, feeding and regulation of the sympathetic nervous
system. New England Journal of Medicine, 298, 1295–1301.
Lantz, P. M., House, J. S., Lepkowski, J. M., Williams, D. R., Mero, R. P., & Chen, J. (1998). Socio-
economic factors, health behaviors, and mortality: Results from a nationally representative
prospective study of US adults. Journal of the American Medical Association, 279(21),
1703–1708.
Lapidus, L., Bengtsson, C., Lissner, L., & Smith, U. (1992). Family history of diabetes in relation to
different types of obesity and change of obesity during 12-yr period: Results from prospective
population study of women in Göteborg, Sweden. Diabetes Care, 15, 1455–1458.
Lee,I. M.,& Manson,J. E.(1998). Bodyweight andmortality: Whatis theshape ofthe curve? Epidemi-
ology, 9(3), 227–228.
Lee, I. M., Manson, J. E., Hennekens, C. H., & Paffenbarger, R. S. (1993). Body weight and mortality:
A 27-year follow-up of middle-aged men. Journal of the American Medical Association,
270(23), 2823–2828.
Lew, E. A., End, J. A., & Wilber, J. A. (1979). The new build and blood pressure study. Transactions of
the Association of Life Insurance Medical Directors of America, 62, 154–174.
Lissner, L., Andres, R., Muller, D. C., & Shimokata, H. (1990). Body weight variability in men: Meta-
bolic rate, health and longevity. International Journal of Obesity, 14, 373–383.
Lissner, L., Odell, P. M., D’Agostino, R. B., Stokes, J. III, Kreger, B. E., Belanger, A. J., & Brownell,
K. D. (1991). Variability of body weight and health outcomes in the Framingham population.
New England Journal of Medicine, 324, 1839–1844.
256 Ernsberger and Koletsky
78
Manson, J. E., Willett, W. C., Stampfer, M. J., Colditz, G. A., Hunter, D. J., Hankinson, S. E.,
Hennekens, C. H., & Speizer, F. E. (1995). Body weight and mortality among women. New
England Journal of Medicine, 333, 677–685.
Modan,M., Karasik,A., Halkin,H., Fuchs,Z., Lusky,A., Shitrit, A., & Modan, B. (1986). Effect of past
and concurrent body mass index on prevalence of glucose intolerance and type 2
(non-insulin-dependent) diabetes and on insulin response: The Israel study of glucose intoler-
ance, obesity and hypertension. Diabetologia, 29(2), 82–89.
Morris, J. K., Cook, D. G., & Shaper, A. G. (1992). Non-employment and changes in smoking, drink-
ing, and body weight. British Medical Journal, 304, 536–541.
National Institutes of Health National Task Force on the Prevention and Treatment of Obesity. (1993).
Very low-calorie diets. Journal of the American Medical Association, 270(8), 967–974.
National Institutes of Health National Task Force on the Prevention and Treatment of Obesity. (1994).
Weight cycling. Journal of the American Medical Association, 272(15), 1196–1202.
National Institutes of Health National Task Force on the Prevention and Treatment of Obesity. (1996).
Long-term pharmacotherapy in the management of obesity. Journal of the American Medical
Association, 276(23), 1907–1915.
Oud, L., & Haupt, M. T. (1997). Body mass index and the hospitalized patient: A “mean” outcome for
the “lean.” Critical Care Medicine, 25(12), 1938–1940.
Pfohl, M., Luft, D., Blomberg, I., & Schmülling, R.-M. (1994). Long-term changes of body weight
and cardiovascular risk factors after weight reduction with group therapy and dexfen-
fluramine. International Journal of Obesity, 18, 391–395.
Pi-Sunyer, F. X. (1993). Medical hazards of obesity. Annals of Internal Medicine, 119, 655–660.
Potter,J. F.,Schafer, D. F., & Bohi,R. L.(1988). In-hospitalmortality as a function ofbody massindex:
An age-dependent variable. Journal of Gerontology, 43(3), 59–63.
Reife, C. M. (1995). Involuntary weight loss. Medical Clinics of North America, 79, 299–313.
Rissanen, A., Heliovaara, M., Knekt, P., Aromaa, A., Reunanen, A., & Maatela, J. (1989). Weight and
mortality in Finnish men. Journal of Clinical Epidemiology, 42(8), 781–789.
Rissanen, A., Knekt, P., Heliovaara, M., Aromaa, A., Reunanen, A., & Maatela, J. (1991). Weight and
mortality in Finnish women. Journal of Clinical Epidemiology, 44(8), 787–795.
Rocchini, A. P., Moorehead, C., Wentz, E., & Deremer, S. (1987). Obesity-induced hypertension in the
dog. Hypertension, 9, III64–III68.
Rodin, J. (1993). Cultural and psychosocial determinants of weight concerns. Annals of Internal Medi-
cine, 119, 643–645.
Rodin, J., Radke-Sharpe, N., Rebuffë-Scrive, M., & Greenwood, M. R. (1990). Weight cycling and fat
distribution. International Journal of Obesity, 14, 303–310.
Rosengren, A., Welin, L., Tsipogianni, A., & Wilhelmsen, L. (1989). Impact of cardiovascular risk
factors on coronary heart disease and mortality among middle aged diabetic men: A general
population study. British Medical Journal, 299(6708), 1127–1131.
Ross, C., Langer, R. D., & Barrett-Connor, E. (1997). Given diabetes, is fat better than thin? Diabetes
Care, 20(4), 650–652.
Salmond,C. E., Beaglehole,R., & Prior, I. A.(1985). Are low cholesterol valuesassociated with excess
mortality? British Medical Journal, 290(6466), 422–424.
Schols, A. M., Slangen, J., Volovics, L., & Wouters, E. F. (1998). Weight loss is a reversible factor in
the prognosis of chronic obstructive pulmonary disease. American Journal of Respiratory and
Critical Care Medicine, 157(6, part 1), 1791–1797.
Seccareccia, F., Lanti, M., Menotti, A., & Scanga, M. (1998). Role of body mass index in the prediction
of all cause mortality in over 62,000 men and women. The Italian RIFLE Pooling Project: Risk
Factor and Life Expectancy. Journal of Epidemiology and Community Health, 52(1), 20–26.
Sigal, R. J., El-Hashimy, M., Martin, B. C., Soeldner, J. S., Krolewski, A. S., & Warram, J. H. (1997).
Acute postchallenge hyperinsulinemia predicts weight gain: A prospective study. Diabetes,
46(6), 1025–1029.
Sims, E. A. (1982). Mechanisms of hypertension in the overweight. Hypertension, 4, III43–III49.
Sorensen,T. I. A.(1995). Socioeconomicaspects of obesity:Causes oreffects. International Journalof
Obesity, 19(Suppl. 6), S6–S8.
Sorlie, P., Gordon, T., & Kannel, W. B. (1980). Body build and mortality: The Framingham study.
Journal of the American Medical Association, 243(18), 1828–1831.
Wellness Approach to Obesity 257
79
Sprafka,J. M., Burke,G. L., Folsom,A. R., Luepker,R. V., &Blackburn, H. (1990).Continued decline
in cardiovascular disease risk factors: Results of the Minnesota Heart Survey, 1980–1982 and
1985–1987. American Journal of Epidemiology, 132(3), 489–500.
Stallones, R. A. (1985). Epidemiologic studies of obesity. Annals of Internal Medicine, 103(6, part 2),
1003–1005.
Stevens,J., Cai, J. W., Pamuk,E. R.,Williamson, D. F., Thun, M.J., &Wood, J. L. (1998). The effect of
age on the association between body-mass index and mortality. New England Journal of Medi-
cine, 338(1), 1–7.
Stevens, J., Keil, J. E., Rust, P. F., Tyroler, H. A., Davis, C. E., & Gazes, P. C. (1992). Body mass index
and body girths as predictors of mortality in Black and White women. Archives of Internal
Medicine, 152, 1257–1262.
Tayback, M., Kumanyika, S., & Chee, E. (1990). Body weight as arisk factor in the elderly. Archivesof
Internal Medicine, 150(5), 1065–1072.
Tong, R. (1991). The epistemology and ethics of consensus: Uses and misuses of “ethical” expertise.
Journal of Medical Philosophy, 16(4), 409–426.
Troiano, R. P., Frongillo, E. A. Jr., Sobal, J., & Levitsky, D. A. (1996). The relationship between body
weight and mortality: A quantitative analysis of combined information from existing studies.
International Journal of Obesity, 20, 63–75.
Tuomilehto, J., Salonen, J. T., Marti, B., Jalkanen, L., Puska, P., Nissinen, A., & Wolf, E. (1987). Body
weight and risk of myocardial infarction and death in the adult population of eastern Finland.
British Medical Journal, 295(6599), 623–627.
Vandenbroucke, J. P., Mauritz, B. J., de Bruin, A., Verheesen, J. H., van der Heide-Wessel, C., & van
der Heide, R. M. (1984). Weight, smoking, and mortality. Journal of the American Medical
Association, 252(20), 2859–2860.
Van Itallie, T. B. (1979). Obesity: Adverse effects on health and longevity. American Journal of
Clinical Nutrition, 32(12 Suppl.), 2723–2733.
Van Itallie, T. B. (1985). Health implications of overweight and obesity in the United States. Annals of
Internal Medicine, 103(6, part 2), 983–988.
Van Itallie, T. B., & Burton, B. T. (1979). National Institutes of Health Consensus Development
Conference on Surgical Treatment of Morbid Obesity. Annals of Surgery, 189(4), 455–457.
Vezina, W. C., Grace, D. M., Hutton, L. C., Alfieri, M. H., Colby, P. R., Downey, D. B., Vanderwerf,
R.J., White,N. F.,& Ward,R. P.(1998). Similarity in gallstone formation from 900 kcal/day di-
ets containing 16 g vs 30 g of daily fat: Evidence that fat restriction is not the main culprit of
cholelithiasis during rapid weight reduction. Digestive Disease Science, 43(3), 554–561.
Waaler, H. T. (1984). Height, weight and mortality. The Norwegian experience. Acta Medica
Scandinavica, 679(Suppl.), 1–56.
Wallace, J. I., Schwartz, R. S., LaCroix, A. Z., Uhlmann, R. F., & Pearlman, R. A. (1995). Involuntary
weight loss in older outpatients: Incidence and clinical significance. Journal of the American
Geriatrics Society, 43, 329–337.
Wamala, S. P., Wolk, A., & Orth-Gomér, K. (1997). Determinants of obesity in relation to socio-
economic status among middle-aged Swedish women. Preventive Medicine, 26(5, part 1),
734–744.
Weinsier,R. L.,& Krumdieck,C. L.(1981). Deathresulting fromoverzealous totalparenteral nutrition:
The refeeding syndrome revisited. American Journal of Clinical Nutrition, 34, 393–399.
Whelton, P. K., Appel, L. J., Espeland, M. A., Applegate, W. B., Ettinger, W. H., Kostis, J. B.,
Kumanyika, S., Lacy, C. R., Johnson, K. C., Folmar, S., & Cutler, J. A. (1998). Sodium reduc-
tion and weight loss in the treatment of hypertension in older persons: A randomized controlled
trial of nonpharmacologic interventions in the elderly (TONE). Journal of the American Medi-
cal Association, 279(11), 839–846.
Wienpahl,J., Ragland, D. R., & Sidney, S.(1990). Bodymass index and 15-year mortalityin acohort of
Black men and women. Journal of Clinical Epidemiology, 43(9), 949–960.
Williams, S. R., Jones, E., Bell, W., Davies, B., & Bourne, M. W. (1997). Body habitus and coronary
heartdisease inmen: A review with referenceto methodsof bodyhabitus assessment. European
Heart Journal, 18(3), 376–393.
Wolf,A. M., Beisiegel,U., Kortner, B.,& Kuhlmann, H.W. (1998). Doesgastric restriction surgeryre-
duce the risks of metabolic diseases? Obesity Surgery, 8(1), 9–13.
258 Ernsberger and Koletsky
80
Wolf,R. N., &Grundy, S. M.(1983). Influence ofweight reduction on plasma lipoproteins in obese pa-
tients. Arteriosclerosis, 3(2), 160–169.
Wood,F. C., &Bierman, E. L. (1986). Isdiet the cornerstonein management of diabetes? NewEngland
Journal of Medicine, 315(19), 1224–1227.
PAUL ERNSBERGER received his Ph.D. in Neuroscience and Pharmacology
from Northwestern University Medical School. After postdoctoral training at
CornellUniversityMedicalCollege, he joined the faculty at Case Western Reserve
University School of Medicine in 1989, where he is currently Associate Professor
of Nutrition, Medicine (Hypertension), Pharmacology and Neuroscience. He has
authored more than 90 peer-reviewed publications in the biomedical sciences and
has lectured nationally and internationally on the relationship between obesity and
hypertension.
RICHARDJ.KOLETSKYreceivedhisM.D.fromCaseWesternReserveUniver-
sity. After residency training at Northwestern University and the Cleveland Clinic,
he served as a research fellow at Harvard Medical School and Peter Bent Brigham
Hospital. He returned to Cleveland as the Chief of Endocrinology at St.-Luke’s
Medical Center, and he is currently Assistant Clinical Professor of Medicine and
Director of the Cleveland Weight Wellness Center. He has published and lectured
extensively on obesity and its relationship to endocrine and metabolic diseases,
combining expertise in clinical and laboratory investigations.
Wellness Approach to Obesity 259
81
