|Home | About | Journals | Submit | Contact Us | Français|
Many beliefs about obesity persist in the absence of supporting scientific evidence (presumptions); some persist despite contradicting evidence (myths). The promulgation of unsupported beliefs may yield poorly informed policy decisions, inaccurate clinical and public health recommendations, and an unproductive allocation of research resources and may divert attention away from useful, evidence-based information.
Using Internet searches of popular media and scientific literature, we identified, reviewed, and classified obesity-related myths and presumptions. We also examined facts that are well supported by evidence, with an emphasis on those that have practical implications for public health, policy, or clinical recommendations.
We identified seven obesity-related myths concerning the effects of small sustained increases in energy intake or expenditure, establishment of realistic goals for weight loss, rapid weight loss, weight-loss readiness, physical-education classes, breast-feeding, and energy expended during sexual activity. We also identified six presumptions about the purported effects of regularly eating breakfast, early childhood experiences, eating fruits and vegetables, weight cycling, snacking, and the built (i.e., human-made) environment. Finally, we identified nine evidence-supported facts that are relevant for the formulation of sound public health, policy, or clinical recommendations.
False and scientifically unsupported beliefs about obesity are pervasive in both scientific literature and the popular press. (Funded by the National Institutes of Health.)
Passionate interests, the human tendency to seek explanations for observed phenomena, and everyday experience appear to contribute to strong convictions about obesity, despite the absence of supporting data. When the public, mass media, government agencies, and even academic scientists espouse unsupported beliefs, the result may be ineffective policy, unhelpful or unsafe clinical and public health recommendations, and an unproductive allocation of resources. In this article, we review some common beliefs about obesity that are not supported by scientific evidence and also provide some useful evidence-based concepts. We define myths as beliefs held to be true despite substantial refuting evidence, presumptions as beliefs held to be true for which convincing evidence does not yet confirm or disprove their truth, and facts as propositions backed by sufficient evidence to consider them empirically proved for practical purposes.
When standards for evidence are considered, it is critical to distinguish between drawing conclusions from scientific evidence and making decisions about prudent actions. Stakeholders must sometimes take action in the absence of strong scientific evidence. Yet this principle of action should not be mistaken as justification for drawing conclusions. Regardless of the urgency of public health issues, scientific principles remain unchanged. We find the language of the Federal Trade Commission to be apt: its standard for making claims is “competent and reliable scientific evidence,” defined as “tests, analyses, research, studies, or other evidence … conducted and evaluated in an objective manner … using procedures generally accepted … to yield accurate and reliable results.”1
The scientific community recognizes that randomized experiments offer the strongest evidence for drawing causal inferences. Nevertheless, at least since the 1960s, when Sir Austin Bradford Hill spearheaded the scientific activities that led to the acceptance of the claim that smoking causes lung cancer and to his classic writing on association and causation,2 the scientific community has acknowledged that under some circumstances (i.e., when it is unethical or unfeasible to conduct a randomized study and when observed associations are not plausibly due to confounding), inferring causality in the absence of data from randomized, controlled trials is necessary and appropriate. However, the fact that the appropriateness of inferring causality holds only under certain circumstances is sometimes discounted by those who are eager to garner support for a proposal in the absence of strong data from randomized studies.
Notably, the circumstances that justify drawing a conclusion of causation from nonexperimental data are rarely met in clinical and public proposals regarding obesity. It is possible to conduct randomized studies of even the most sensitive and invasive obesity procedures, as exemplified by recent articles in the Journal. Moreover, observational associations germane to the causes, treatment, and prevention of obesity are subject to substantial confounding, fraught with measurement problems, and typically small and inconsistent.3 Such observational associations are often found to differ from those later obtained by more rigorously designed studies.4 Hence, in the present discussion, we generally conclude that a proposition has been shown to be true only when it has been supported by confirmatory randomized studies. References to published studies are used sparingly herein, with a more comprehensive listing provided in the Supplementary Appendix, available with the full text of this article at NEJM.org.
We review seven myths about obesity, along with the refuting evidence. Table 1 provides anecdotal support that the beliefs are widely held or stated, in addition to reasons that support conjecture.
Myth number 1: Small sustained changes in energy intake or expenditure will produce large, long-term weight changes.
Predictions suggesting that large changes in weight will accumulate indefinitely in response to small sustained lifestyle modifications rely on the half-century-old 3500-kcal rule, which equates a weight alteration of 1 lb (0.45 kg) to a 3500-kcal cumulative deficit or increment.5,6 However, applying the 3500-kcal rule to cases in which small modifications are made for long periods violates the assumptions of the original model, which were derived from short-term experiments predominantly performed in men on very-low-energy diets (<800 kcal per day).5,7 Recent studies have shown that individual variability affects changes in body composition in response to changes in energy intake and expenditure,7 with analyses predicting substantially smaller changes in weight (often by an order of magnitude across extended periods) than the 3500-kcal rule does.5,7 For example, whereas the 3500-kcal rule predicts that a person who increases daily energy expenditure by 100 kcal by walking 1 mile (1.6 km) per day will lose more than 50 lb (22.7 kg) over a period of 5 years, the true weight loss is only about 10 lb (4.5 kg),6 assuming no compensatory increase in caloric intake, because changes in mass concomitantly alter the energy requirements of the body.
Myth number 2: Setting realistic goals for weight loss is important, because otherwise patients will become frustrated and lose less weight.
Although this is a reasonable hypothesis, empirical data indicate no consistent negative association between ambitious goals and program completion or weight loss.8 Indeed, several studies have shown that more ambitious goals are sometimes associated with better weight-loss outcomes (see the Supplementary Appendix).8 Furthermore, two studies showed that interventions designed to improve weight-loss outcomes by altering unrealistic goals resulted in more realistic weight-loss expectations but did not improve outcomes.
Myth number 3: Large, rapid weight loss is associated with poorer long-term weight-loss outcomes, as compared with slow, gradual weight loss.
Within weight-loss trials, more rapid and greater initial weight loss has been associated with lower body weight at the end of long-term follow-up.9,10 A meta-analysis of randomized, controlled trials that compared rapid weight loss (achieved with very-low-energy diets) with slower weight loss (achieved with low-energy diets — i.e., 800 to 1200 kcal per day) at the end of short-term follow-up (<1 yr) and long-term follow-up (=1 year) showed that, despite the association of very-low-energy diets with significantly greater weight loss at the end of short-term follow-up (16.1% of body weight lost, vs. 9.7% with low-energy diets), there was no significant difference between the very-low-energy diets and low-energy diets with respect to weight loss at the end of long-term follow-up.10 Although it is not clear why some obese persons have a greater initial weight loss than others do, a recommendation to lose weight more slowly might interfere with the ultimate success of weight-loss efforts.
Myth number 4: It is important to assess the stage of change or diet readiness in order to help patients who request weight-loss treatment.
Readiness does not predict the magnitude of weight loss or treatment adherence among persons who sign up for behavioral programs or who undergo obesity surgery.11 Five trials (involving 3910 participants; median study period, 9 months) specifically evaluated stages of change (not exclusively readiness) and showed an average weight loss of less than 1 kg and no conclusive evidence of sustained weight loss (see the Supplementary Appendix). The explanation may be simple — people voluntarily choosing to enter weight-loss programs are, by definition, at least minimally ready to engage in the behaviors required to lose weight.
Myth number 5: Physical-education classes, in their current form, play an important role in reducing or preventing childhood obesity.
Physical education, as typically provided, has not been shown to reduce or prevent obesity. Findings in three studies that focused on expanded time in physical education12 indicated that even though there was an increase in the number of days children attended physical-education classes, the effects on body-mass index (BMI) were inconsistent across sexes and age groups. Two meta-analyses showed that even specialized school-based programs that promoted physical activity were ineffective in reducing BMI or the incidence or prevalence of obesity.13 There is almost certainly a level of physical activity (a specific combination of frequency, intensity, and duration) that would be effective in reducing or preventing obesity. Whether that level is plausibly achievable in conventional school settings is unknown, although the dose–response relationship between physical activity and weight warrants investigation in clinical trials.
Myth number 6: Breast-feeding is protective against obesity.
A World Health Organization (WHO) report states that persons who were breast-fed as infants are less likely to be obese later in life and that the association is “not likely to be due to publication bias or confounding.”14 Yet the WHO, using Egger’s test and funnel plots, found clear evidence of publication bias in the published literature it synthesized.15 Moreover, studies with better control for confounding (e.g., studies including within-family sibling analyses) and a randomized, controlled trial involving more than 13,000 children who were followed for more than 6 years16 provided no compelling evidence of an effect of breast-feeding on obesity. On the basis of these findings, one long-term proponent of breast-feeding for the prevention of obesity wrote that breast-feeding status “no longer appears to be a major determinant” of obesity risk17; however, he speculated that breast-feeding may yet be shown to be modestly protective, current evidence to the contrary. Although existing data indicate that breast-feeding does not have important antiobesity effects in children, it has other important potential benefits for the infant and mother and should therefore be encouraged.
Myth number 7: A bout of sexual activity burns 100 to 300 kcal for each participant.
The energy expenditure of sexual intercourse can be estimated by taking the product of activity intensity in metabolic equivalents (METs),18 the body weight in kilograms, and time spent. For example, a man weighing 154 lb (70 kg) would, at 3 METs, expend approximately 3.5 kcal per minute (210 kcal per hour) during a stimulation and orgasm session. This level of expenditure is similar to that achieved by walking at a moderate pace (approximately 2.5 miles [4 km] per hour). Given that the average bout of sexual activity lasts about 6 minutes,19 a man in his early-to-mid-30s might expend approximately 21 kcal during sexual intercourse. Of course, he would have spent roughly one third that amount of energy just watching television, so the incremental benefit of one bout of sexual activity with respect to energy expended is plausibly on the order of 14 kcal.
Just as it is important to recognize that some widely held beliefs are myths so that we may move beyond them, it is important to recognize presumptions, which are widely accepted beliefs that have neither been proved nor disproved, so that we may move forward to collect solid data to support or refute them. Instead of attempting to comprehensively describe all the data peripherally related to each of the six presumptions shown in Table 2, we describe the best evidence.
Presumption number 1: Regularly eating (versus skipping) breakfast is protective against obesity.
Two randomized, controlled trials that studied the outcome of eating versus skipping breakfast showed no effect on weight in the total sample.20 However, the findings in one study suggested that the effect on weight loss of being assigned to eat or skip breakfast was dependent on baseline breakfast habits.20
Presumption number 2: Early childhood is the period in which we learn exercise and eating habits that influence our weight throughout life.
Although a person’s BMI typically tracks over time (i.e., tends to be in a similar percentile range as the person ages), longitudinal genetic studies suggest that such tracking may be primarily a function of genotype rather than a persistent effect of early learning.21 No randomized, controlled clinical trials provide evidence to the contrary.
Presumption number 3: Eating more fruits and vegetables will result in weight loss or less weight gain, regardless of whether any other changes to one’s behavior or environment are made.
It is true that the consumption of fruits and vegetables has health benefits. However, when no other behavioral changes accompany increased consumption of fruits and vegetables, weight gain may occur or there may be no change in weight.22
Presumption number 4: Weight cycling (i.e., yo-yo dieting) is associated with increased mortality.
Although observational epidemiologic studies show that weight instability or cycling is associated with increased mortality, such findings are probably due to confounding by health status. Studies of animal models do not support this epidemiologic association.23
Presumption number 5: Snacking contributes to weight gain and obesity.
Randomized, controlled trials do not support this presumption.24 Even observational studies have not shown a consistent association between snacking and obesity or increased BMI.
Presumption number 6: The built environment, in terms of sidewalk and park availability, influences the incidence or prevalence of obesity.
According to a systematic review, virtually all studies showing associations between the risk of obesity and components of the built environment (e.g., parks, roads, and architecture) have been observational.25 Furthermore, these observational studies have not shown consistent associations, so no conclusions can be drawn.
Our proposal that myths and presumptions be seen for what they are should not be mistaken as a call for nihilism. There are things we do know with reasonable confidence. Table 3 lists nine such facts and their practical implications for public health, policy, or clinical recommendations. The first two facts help establish a framework in which intervention and preventive techniques may work. The next four facts are more prescriptive, offering tools that can be conveyed to the public as well established. The last three facts are suited to clinical settings.
Myths and presumptions about obesity are common. Several presumptions appear to be testable, and some of them (e.g., effects of eating breakfast daily, eating more fruits and vegetables, and snacking) can be tested with standard study designs. Despite enormous efforts promoting these ideas, research often seems mired in the accrual of observational data. Many of the trials that have been completed or are in progress do not isolate the effect of the presumed influence and the findings are therefore not definitive.
Many of the myths and presumptions about obesity reflect a failure to consider the diverse aspects of energy balance,35 especially physiological compensation for changes in intake or expenditure.36 Some myths and presumptions involve an implicit assumption that there is no physiological compensation whatsoever (i.e., the 3500-kcal rule) or only minimal compensation (e.g., a reduction in snacking as a means of reducing weight). In other cases, there is an implicit assumption of overcompensation (e.g., eating breakfast daily or increasing the intake of fruits and vegetables as a means of reducing weight). Proponents of other unsupported ideas fail to consider that people burn some amount of energy even without engaging in the activity in question (e.g., increased sexual activity). In addition, interested parties do not regularly request the results from randomized, long-term studies that measure weight or adiposity as an outcome. Therefore, the presented data are rife with circumstantial evidence, and people are not informed that the existing evidence is not compelling (e.g., breakfast consumption). Furthermore, some suggested treatment or prevention strategies may work well (e.g., increasing the consumption of fruits and vegetables) but only as part of a multifaceted program for weight reduction. Yet such a strategy is often presented as though it will have effects in isolation and even among persons not participating in weight-loss programs. We must recognize that evidence that a technique is beneficial for the treatment of obesity is not necessarily evidence that it will be helpful in population-based approaches to the prevention of obesity, and vice versa.
Why do we think or claim we know things that we actually do not know? Numerous cognitive biases lead to an unintentional retention of erroneous beliefs.37,38 When media coverage about obesity is extensive, many people appear to believe some myths (e.g., rapid weight loss facilitates weight regain) simply because of repeated exposure to the claims.39 Cognitive dissonance may prevent us from abandoning ideas that are important to us, despite contradictory evidence (e.g., the idea that breast-feeding prevents obesity in children). Similarly, confirmation bias may prevent us from seeking data that might refute propositions we have already intuitively accepted as true because they seem obvious (e.g., the value of realistic weight-loss goals). Moreover, we may be swayed by persuasive yet fallacious arguments (Whately provides a classic catalogue40) unless we are prepared to identify them as spurious.
Fortunately, the scientific method and logical thinking offer ways to detect erroneous statements, acknowledge our uncertainty, and increase our knowledge. When presented with an alleged truth, we can pause to ask simple questions, such as, “How could someone actually know that?” Such a simple question allows one to easily recognize some beliefs as spurious (e.g., 300 kcal is burned during sexual intercourse). Moreover, we often settle for data generated with the use of inadequate methods in situations in which inferentially stronger study designs, including quasi-experiments and true randomized experiments, are possible, as recently illustrated (see the Supplementary Appendix). In addition, eliminating the distortions of scientific information that sometimes occur with public health advocacy would reduce the propagation of misinformation.
The myths and presumptions about obesity that we have discussed are just a sampling of the numerous unsupported beliefs held by many people, including academics, regulators, and journalists, as well as the general public. Yet there are facts about obesity of which we may be reasonably certain — facts that are useful today. While we work to generate additional useful knowledge, we may in some cases justifiably move forward with hypothesized, but not proved, strategies. However, as a scientific community, we must always be open and honest with the public about the state of our knowledge and should rigorously evaluate unproved strategies.
Supported in part by a grant (P30DK056336) from the National Institutes of Health.
Dr. Astrup reports receiving payment for board membership from the Global Dairy Platform, Kraft Foods, Knowledge Institute for Beer, McDonald’s Global Advisory Council, Arena Pharmaceuticals, Basic Research, Novo Nordisk, Pathway Genomics, Jenny Craig, and Vivus; receiving lecture fees from the Global Dairy Platform, Novo Nordisk, Danish Brewers Association, GlaxoSmithKline, Danish Dairy Association, International Dairy Foundation, European Dairy Foundation, and AstraZeneca; owning stock in Mobile Fitness; holding patents regarding the use of flaxseed mucilage or its active component for suppression of hunger and reduction of prospective consumption (patents EP1744772, WO2009033483-A1, EP2190303-A1, US2010261661-A1, and priority applications DK001319, DK001320, S971798P, and US971827P); holding patents regarding the use of an alginate for the preparation of an aqueous dietary product for the treatment or prevention of overweight and obesity (patent WO2011063809-A1 and priority application DK070227); and holding a patent regarding a method for regulating energy balance for body-weight management (patent WO2007062663-A1 and priority application DK001710). Drs. Brown and Bohan Brown report receiving grant support from the Coca-Cola Foundation through their institution. Dr. Mehta reports receiving grant support from Kraft Foods. Dr. Newby reports receiving grant support from General Mills Bell Institute of Health and Nutrition. Dr. Pate reports receiving consulting fees from Kraft Foods. Dr. Rolls reports having a licensing agreement for the Volumetrics trademark with Jenny Craig. Dr. Thomas reports receiving consulting fees from Jenny Craig. Dr. Allison reports serving as an unpaid board member for the International Life Sciences Institute of North America; receiving payment for board membership from Kraft Foods; receiving consulting fees from Vivus, Ulmer and Berne, Paul, Weiss, Rifkind, Wharton, Garrison, Chandler Chicco, Arena Pharmaceuticals, Pfizer, National Cattlemen’s Association, Mead Johnson Nutrition, Frontiers Foundation, Orexigen Therapeutics, and Jason Pharmaceuticals; receiving lecture fees from Porter Novelli and the Almond Board of California; receiving payment for manuscript preparation from Vivus; receiving travel reimbursement from International Life Sciences Institute of North America; receiving other support from the United Soybean Board and the Northarvest Bean Growers Association; receiving grant support through his institution from Wrigley, Kraft Foods, Coca-Cola, Vivus, Jason Pharmaceuticals, Aetna Foundation, and McNeil Nutritionals; and receiving other funding through his institution from the Coca-Cola Foundation, Coca-Cola, PepsiCo, Red Bull, World Sugar Research Organisation, Archer Daniels Midland, Mars, Eli Lilly and Company, and Merck.
We thank Drs. Kyle Grimes and S. Louis Bridges for their suggestions on an earlier version of the manuscript.
The views expressed in this article are those of the authors and do not necessarily represent the official views of the National Institutes of Health.
No other potential conflict of interest relevant to this article was reported.
Disclosure forms provided by the authors are available with the full text of this article at NEJM.org.