FRAMING HEALTH MATTERS

Obesity, Health at Every Size, and Policy

Andrea Bombak, MA

presented by the studies was likely overesti- is associated with chronic diseases that may negatively affect in- dividuals’ health and the sustainability of the health care system. Despite in- mated attributed to confounding of the effects creasing emphasis on obesity as a major health care issue, little progress has been of diet and , low follow-up rates, self- ’ made in its treatment or prevention. Individual approaches to obesity treatment, report of participants weight by phone or mail, largely composed of weight-loss , have not proven effective. Little direct and many participants’ use of subsequent diets evidence supports the notion of reforms to the obesogenic environment. Both following the studies in question. Regardless, these individualistic and environmental approaches to obesity have important the trials demonstrated no consistent health limitations and ethical implications. The low levels of success associated with benefits, including sustained . In- these approaches may necessitate a new non–weight-centric public health deed, one-to-two thirds of dieters regained strategy. Evidence is accumulating that a weight-neutral, nutrition- and physical more weight than initially lost. activity–based, Health at Every Size (HAES) approach may be a promising chronic Only recently have the long-term physio- disease-prevention strategy. (Am J Public Health. Published online ahead of print logical changes following weight loss begun to December 12, 2013: e1–e8. doi:10.2105/AJPH.2013.301486) be illuminated. In part, this gap in knowledge arises from low numbers of weight-loss main- 1 Obesity is defined as having a weight-loss outcomes, and environmental re- tainers. Recent evidence on the physiological (BMI; defined as weight in kilograms divided forms devoted to enhancing livability, accessi- changes that follow weight loss in obese by the square of height in meters) in excess of bility, and equity. Evidence is accumulating animal models and humans suggests organ- ’ 30. Obesity is associated with numerous that a weight-neutral, nutrition- and physical isms endocrine systems actively promote 3 chronic health conditions, including diabetes, activity---based, Health at Every Size (HAES) weight regain. Such changes involve highly hypertension, heart disease, and certain can- approach may be a promising chronic-disease integrated alterations including reduced sati- cers.1 The directionality of such associations is prevention, and overall well-being, strategy. ety, increased hunger, suppressed energy ex- largely unknown, confounding may be present, penditure, a decrease in nutrient availability, fl and causality has only definitely been assigned EXISTING APPROACHES enhanced metabolic exibility, an increase in fi to obesity with respect to osteoarthritis and energy ef ciency and storage in peripheral ovarian cancer.2 Despite these limitations, to Current public health approaches to obesity tissues, a decline in adipose energy depletion counter the health effects of obesity-associated largely consist of promoting individual weight signaling from leptin and insulin, altered neu- conditions, individuals frequently are encour- loss dieting or alterations to obesogenic envi- ral activity, and alterations in hepatic, adipose 3 aged to lose weight to improve individual and ronments. tissue, and skeletal muscle metabolism. Im- population health. However, diet-induced portantly, similar changes have been identi- weight loss stimulates somatic and psycholog- Weight Loss Dieting fied in humans for a full year after weight 4 ical ‘homeostatic pressures’ that induce weight A particularly potent argument against pub- loss. One year following weight loss, sub- regain.3 These mechanisms include hormonal lic health’s existing anti-obesity tactics is the jective hunger, ghrelin (hunger stimulating alterations, reduced satiety and energy expen- proven ineffectiveness of weight loss attempts hormone), gastric inhibitory polypeptide, and diture, and increased hunger.3,4 These adap- at an individual level. Indeed, as early as 1992, pancreatic polypeptide were all elevated tations stimulate weight regain in more than the American National Institutes of Health among and obese individuals who 90% of weight losers.5,6 In acknowledgment (NIH) released a consensus statement that lost weight in an intervention. Decreases in of the limited effectiveness of individual ap- dieting is an ineffective method to produce insulin, leptin, polypeptide YY, and cholecys- proaches to weight loss, increasing emphasis sustained weight loss.7,8 Mann et al.9 con- tokinin were also observed.4 These findings has been placed on environmental reforms. ducted a comprehensive review of the long- align with the reduced rates of nonresting and However, when weight loss is the key motiva- term consequences of weight loss dieting in resting energy expenditure and thermic effect tor of such changes, they are hindered by a long-term randomized trials, observational of feeding present in individuals who have limited evidence base and ethical difficulties. studies without a control group, and prospec- sustained a loss of at least 10% of body weight These concerns suggest public health would tive studies without randomization. The studies over a year.10 Such processes may help explain benefit from a shift in focus from weight loss to reviewed by Mann et al.9 were methodologi- the very high rates of recidivism in weight-loss disease prevention for individuals of all ages cally biased to show long-term weight-loss dieters.5,6 Given these rates of recidivism, it is and sizes, with a focus on health rather than maintenance. The effectiveness of dieting concerning that weight regain is largely

Published online ahead of print December 12, 2013 | American Journal of Public Health Bombak | Peer Reviewed | Framing Health Matters | e1 FRAMING HEALTH MATTERS

composed of fat; the weight regained does not distribution and subsequent risk.13 In certain on health.28 In summary, the efficacy and replace bone mass or lean mass lost during populations, BMI may not adequately measure safety of weight-loss dieting is questionable, as weight loss.8,9 Furthermore, weight loss may adiposity or differentiate between populations, are the health consequences of the most not be harmless and may increase stress, particularly with respect to athletes and athletic common forms of obesity.16,17 In addition to release of persistent organic pollutants, and populations.14 Perhaps most importantly, BMI being ineffective, focusing on individuals as risk of osteoporosis.11 is a population-level proxy measure of obesity, the cause and likeliest cure for obesity pro- Those individuals who do sustain substan- and an individual’s BMI may say little con- motes weight stigma, which may be particu- tial weight loss over time generally must cerning that individual’s health.15 Additionally, larly health debilitating.26 maintain high levels of dietary restraint, overweight status (BMI = 25---30) has been physical activity, and self-monitoring behav- shown to be associated with decreased mor- Environmental Antiobesity Reforms iors.11 This is evident in reports from members tality risk in US, Canadian, and international An ecological perspective on obesity causa- of the National Weight Control Registry samples.11,16 ---18 Obesity (BMI > 30) has even tion was first proposed by Egger and Swin- (NWCR). The NWCR is composed of individ- demonstrated a protective or neutral effect burn29 in 1997 and further developed by uals who have sustained a 30-pound weight among some chronic disease or older popula- Swinburn, et al.30 in 1999. Their definition loss for at least a year.5 The registry has been tions.17,19,20 Minor (£ 5%) intentional weight of the obesogenic environment included the critiqued for its nonrepresentativeness of the loss may reduce mortality risk in obese in- micro- and macrocomponents of individuals’ US population, its reliance on self-reported dividuals with related comorbidities, but physical, political, economic, and sociocultural data, and its high rates of loss to follow-up.5 weight loss heightens mortality risk among circumstances. More recently, the obesogenic Ogden et al.12 conducted a latent class cluster healthy obese individuals.21 Therefore, in ad- environment is thought to be composed of analysis of the NWCR to detect distinctive dition to being extraordinarily difficult to factors such as the reduced time-cost of food, methods, experiences, and perspectives on maintain, weight loss may not always be health the increased availability of high-calorie, weight loss and weight maintenance among enhancing, particularly for the large proportion nutrient-poor food, and the increased motori- successful weight losers. Although a majority of obese persons who may be cardiometaboli- zation and mechanization of daily life.31 These of weight losers were satisfied with their cally healthy,22 or whose ill health may be factors are thought to have all simultaneously weight loss and reported good health and attributable to numerous non---weight-related produced an energy imbalance in the daily healthy behaviors, particularly high levels of confounding factors.2 lives of individuals and consequent population- exercise, other results were more distressing. Aiming for a narrow “healthy” range of wide .31 For example, more than one quarter (26.9%) body size may induce individuals to engage in Rectifying the obesogenic environment of participants reported high rates of stress, disordered eating practices.23 These behav- would require comprehensive government re- depression, and dissatisfaction with their iors may remain hidden, and these individuals forms.31 Existing evaluations of health inter- weight loss. This cluster struggles with their may still be valorized for their “healthiness,” ventions appear to support the notion that the weight, frequently weight cycle, and compared particularly by others trying to lose weight. most effective interventions for chronic disease with other members of NWCR, report poorer This is evident among members of, and in the prevention and health promotion focus on health. Interestingly, these individuals are espousal of seemingly bulimic practices in, larger-scale interventions that make healthy trying to maintain the greatest weight loss. The commercial weight-loss groups.24,25 That is, choices more convenient and affordable, re- cluster identified as trying to maintain the weight-loss groups’ weight-loss strategies may strict unhealthy products, focus on community second largest weight loss (9.9%), reported include advice similar to bulimic practices of development, and support healthy social, eco- low levels of physical activity, frequent skip- binge eating and subsequent compensation nomic, and environmental policy.32 The most ping of meals, and the poorest levels of health through dietary restriction or compulsive ex- cost-effective approaches to obesity, as deter- apart from those who had lost more weight. ercising. Furthermore, a public health ap- mined by modeling studies, are those that The relatively poor psychological and physical proach in which obese individuals are viewed include a taxation on unhealthy foods, restric- health of the individuals in these 2 clusters, as unhealthy burdens on the health care tion of junk food advertisement to children, and their struggles to maintain weight health- system and individually responsible for their and improved nutrition labeling.33 Targeted ily, suggest that they may be striving to weight may promote weight stigma.26 This family-based programs for obese children and maintain an overly extreme weight loss. stigma may produce ill health through direct school-based campaigns to increase physical Current public health approaches assume stress-induced neuroendocrine pathways or education and reduce sweetened beverage that intentional weight loss is always positive adverse coping mechanisms.26 Weight bias is consumption and television viewing are also and that a BMI in the range of 18.5 to 24.9 highly prevalent among health care profes- cost-effective in models.33 These modeling is necessarily the healthiest range for all in- sionals, and this may pose a barrier to in- studies and reviews of previous public health dividuals. Importantly, despite its widespread dividuals receiving adequate health care.27 interventions have led proponents of obeso- adoption, BMI is acknowledged to be a Additionally, weight discrimination may di- genic environmental reforms to extrapolate crude measurement of obesity that may not minish an individual’s socioeconomic stand- that obesity prevalence could be lessened adequately account for regional body fat ing, which will have further deleterious effects through structural modifications that would

e2 | Framing Health Matters | Peer Reviewed | Bombak American Journal of Public Health | Published online ahead of print December 12, 2013 FRAMING HEALTH MATTERS

facilitate energy expenditure or inhibit caloric Kirk et al.35 conducted a scoping review on established that middle-class US familes were intake. obesogenic environment research. The review more likely to eat at fast-food restaurants than Evidence for the Obesogenic Environment. determined that most studies focused on the were lower or higher income individuals, Appeals to the obesogenic environment may physical microenvironment, and many studies whereas higher income individuals were more appear to be an effective and nonstigmatizing used physical activity, rather than diet or BMI, likely to eat at full-service restaurants. An and approach to obesity policy. However, a num- as outcome measures. Few studies focused on Sturm41 found that California children and ber of limitations must be considered regard- the economic or political (such as household or adolescents’ proximity to fast-food outlets and ing these strategies. Particularly problematic workplace policies) microenvironment. Also convenience stores, and distance from larger is the lack of direct empirical evidence that limited were studies conducted on the macro grocery stores, is not associated with healthy environments necessarily predispose individ- level, including urban development, health and food consumption or BMI. Lee42 determined uals to developing obesity, which calls into transportation systems, the media, and the food that elementary-school children residing in question the validity of using obesity as industry. Most importantly, the review found poorer and minority neighborhoods in the ajustification for environmental reforms. A that such studies are methodologically stymied United States had greater access to both fast- recent review, for example, highlighted that all by a general inability to measure potential food outlets and grocery stores. Food access evidence used to support the obesogenic environmental effects appropriately and was also not found to independently predict environment is observational and inconsis- a comprehensive theoretical framework that weight gain in children over time. This access tent, and the proposed relationship between would help conceptualize the intricate and also did not account for socioeconomic and environment and obesity must be viewed as multifaceted obesogenic environment. minority differences over time in weight gain, presumptive.34 Models are thus based on Reviews of existent research suggest only which disadvantaged poorer and minority stu- hypothetical contributors to an overly sim- limited support for the obesogenic environ- dents. Interestingly, higher income and White plistic energy balance model, rather than ment. A systematic review conducted by Giskes majority neighborhoods also have a greater empirical evidence.33 Furthermore, obesity et al.38 on environmental effects on fat and share of fast-food outlets, of all food establish- may not even indicate poor health, and re- energy intake established there was limited ments, compared with minority or poorer duced obesity prevalence may not improve evidence that environment influenced fat or neighborhoods.42 Junk food in schools was not population health. Kirk et al.35 assert that energy intake and that any such evidence was a determinant in the development of obesity a main issue hindering effective obesogenic produced from observational, cross-sectional by the eighth grade in the United States.43 environment research is nonconsensus over studies. A more recent review on food-related A recent study by the US Department of what aspects of the environment, a necessarily environmental factors also was unable to con- Agriculture asserted that assessments of the complex, dynamic, and multilevel concept, firm the impact of the obesogenic environ- affordability of a healthy diet are largely de- should be implicated under the obesogenic ment.36 However, some reviews have found pendent on the metric used to measure food environment rubric. Other issues persist in support for an association between the built prices. For example, if food is measured by environmental obesity research. Randomized environment and BMI or obesity.30 Greater portion, it appears less expensive than if food is control trials are difficult to conduct, and support is available for a relationship between measured by weight or calorie.44 impossible for the highest level of upstream physical activity accessibility and BMI or obe- Ethical Considerations of Environmental environmental determinants given small sity than for food environments and BMI or Reforms. Ethical implications of broad-based number of units (e.g., high-end food poli- obesity.30 Evidence suggests walkability, den- antiobesogenic environmental reforms are also cies).36 Thus the highest quality of research sity, accessibility of recreational spaces, and essential to consider. As limited interventions originates from natural experiments or quasi- attractiveness increase physical activity and have been implemented, this involves consid- experimental designs.36 However, the majority may lower risk of elevated BMI.37 Isolated ering their potential implications on target of environmental research remains observa- longitudinal studies have found some support audiences based on proposed projects and tional.35,37 Causality is difficult to assess, as for environmental factors and their effect on deployed rhetoric. Broad alterations to the multilevel studies can only suggest causality, and BMI and obesity.37 Accessibility to healthy environment may increase stigma on obese cross-sectional studies may not establish cau- food increase is associated with dietary quality, individuals who do not lose weight, despite sality.35,36 Difficulties arise in identifying but less evidence exists to support a link to BMI these environmental modifications.45 Should valid, reliable, and consistent dietary, social or obesity.37 This may be an artifact of less individuals not lose weight despite environ- and material resource, and spatial mea- research focusing on diet, because of dietary mental reforms, this might be viewed as proof sures.37 Accounting for all individual level measures-related difficulties.35 that their obesity is the result of an unwilling- factors and their potential mediating and Recent independent studies also failed to ness to practice healthy choices.46 Such stigma confounding effects is particularly challeng- confirm a basic relationship between environ- may contribute to poorer health outcomes.26 ing.37 Furthermore, there may be numerous ment and obesity. McPhail et al.39 found that Obesogenic environment approaches may unknown or unmeasured effects or multiple Canadian adolescents’ socioeconomic class or contribute to a homogenizing view of lower- interactions of social and individual factors access to fast food did not affect their likelihood income lifestyles and the perception of lower- that may affect interpretation.36 to consume fast food. Kim and Leigh40 income individuals as passive and lacking in

Published online ahead of print December 12, 2013 | American Journal of Public Health Bombak | Peer Reviewed | Framing Health Matters | e3 FRAMING HEALTH MATTERS

agency.47 Furthermore, this view may serve as stroke, an increase in the number of years change over this time period.57 Prevalence of a mask for gender-, race-, or socioeconomic- living with said diseases, and fewer number of prediabetes or diabetes, however, increased based discrimination and essentialist argu- years lived free of CVD. Critically, this study substantially by 14% during these years. Given ments, and the moralization of a health is- did not control for cardiorespiratory fitness the stabilization in obesity over this time, this sue.47---53 The obesogenic environment ratio- (CRF), physical activity, diet, social support, again suggests a severance from the seemingly nale has also been criticized as being overly income, occupation, weight history, or weight unassailable link between increasing obesity simplistic and dismissive of the role of neo- loss drug use. The study’s only controls related and elevation in diabetes risk. Importantly, liberalism, its systematic overproduction and to smoking status, number of cigarettes smoked although 61% of obese adolescents and 49% overconsumption, and consequent societal per day, and the exclusion of individuals with of overweight adolescents had at least 1 CVD anxieties and values.48,52 Importantly, broad- diabetes, hypertension, and hypercholesterol- risk factor, 37% of adolescents of normal based reforms may become increasingly in- emia at start of follow-up.54 weight also had at least 1 CVD risk factor.57 vasive and truly limit the freedom of choice of To better understand underlying cardiome- individuals.46 A compelling argument can be tabolic risk in US adults of various weight Independent Effects of Fitness and made for the type of environmental restruc- classes, Wildman et al.22 examined data from Diet on Health turing proposed by obesogenic environment the National Health and Nutrition Examination Increasing evidence suggests fitness and diet arguments, without relying on mobilizing fear Survey (NHANES) from 1999 to 2004. They may affect health independent of weight status, of fat to support their institution.46 By focusing determined the prevalence of overweight and and that obesity and fitness are nonmutually on other measures of health and well-being, or obese adult individuals who are metabolically exclusive. Physical inactivity has a biologically sustainability, livability, and accessibility, valu- fit (characterized by 0 or 1 of the following plausible, temporally consistent, and dose- able and legitimately effective reforms could be abnormalities: low high density lipoprotein dependent relation to risk factors, chronic undertaken. In summary, limited evidence cholesterol level or elevated blood pressure, morbidity, and mortality.58,59 Low CRF has an suggests that environment affects obesity C-reactive protein, homeostasis model assess- even stronger effect on these outcomes than prevalence in a direct, simplistic manner, and ment of insulin resistance value, levels of tri- does physical inactivity.58 In fact, fitness may reforms based on restructuring particular en- glycerides, and fasting plasma glucose) and be as relevant a predictor of mortality as vironments to remedy obesity rates have im- normal-weight adult individuals presenting diabetes mellitus and other CVD risk factors, portant ethical implications. Furthermore, as with cardiometabolic risk clustering. The au- and it may be a stronger predictor than BMI, the empirical evidence suggests, none of these thors found that more than one half of the obesity, or abdominal obesity.59 actions may necessarily result in thinner pop- overweight participants (51.3%) and approxi- Obese individuals who engage in moderate ulations. mately one third of the obese participants intensity physical activity for 150 minutes per (31.7%) were metabolically well. By contrast, week have half the death rates and lower rates CARDIOMETABOLIC HEALTH nearly one quarter of normal-weight individ- of CVD than their unfit, normal-weight coun- AND OBESITY uals (23.5%) had cardiometabolic risk cluster- terparts.58 Furthermore, physically active ing. A predictor of metabolic health was overweight or obese individuals may have Obesity is associated with a number of physical activity.22 greater cardiovascular fitness than inactive in- conditions including diabetes, hypertension, Similarly, data from the 2009---2010 NHANES dividuals, regardless of weight status.60 Both heart disease, and certain cancers1; however, suggest that rates of high cholesterol are low and CRF and obesity were found to affect self-rated causality has yet to be determined in most declining as of 1999 in the United States.55 health (SRH) among adolescent Portuguese instances.2 Furthermore, it is essential not to Additionally, in the American Bogalusa Heart girls. However, the association between obesity overlook potential ill health in normal-weight Study, the children participants experienced an and SRH was eliminated in multivariate statis- individuals. Critically assessing the studies that increase in obesity during the years 1974 to tics, suggesting that CRF may mediate this shaped biomedical views of the risks conferred 1993; however, their rates of hypertension de- relationship.61 In California, children involved by obesity is also imperative for better un- creased.56 This indicates that despite fears of in a school prevention program, fitness im- derstanding the relationship between obesity rising disease risk resulting from increased obesity proved, although obesity rates did not.62 This and health. In this respect, longitudinal studies levels, these risks are less prevalent than assumed. again suggests that obesity and fitness are not utilizing disease outcomes such as cardiovas- This may be the result of better diets, more mutually exclusive. Furthermore, if potentially cular disease (CVD), diabetes, or stroke may be physical activity, and medication use. beneficial programs are only assessed based on especially important to consider. For example, May et al. also present a complex view of weight loss outcomes, they may be substan- Pardo Silva et al.54 reported, based on data cardiometabolic health in US adolescents from tially undervalued. from the highly esteemed Framingham Heart the NHANES from 1999 to 2008.57 Obesity In US adults of various weight classes, Study, that obesity before middle-age in both stabilized during this time at approximately healthy lifestyle habits (e.g., moderate drinking, men and women was associated with numer- 20%. Prevalence of prehypertension or hy- not smoking, regular exercise, and fruit-and- ous negative health outcomes. These included pertension and borderline-high or high low- vegetable consumption) significantly decreased lower life expectancy, , density lipoprotein cholesterol also did not the risk of mortality for all individuals,

e4 | Framing Health Matters | Peer Reviewed | Bombak American Journal of Public Health | Published online ahead of print December 12, 2013 FRAMING HEALTH MATTERS

irrespective of initial BMI. Obese individuals developments that inhibit hedonic urges for hyperlipidemia, hypertension, and coronary benefitted the most from the adoption of overconsumption. These developments may heart disease.71 healthy lifestyle habits.63 In fact, obese indi- also improve capacity for goal-oriented behav- Thus, obese individuals, particularly abdomi- viduals who adopted all 4 healthy lifestyle iors, inhibitory control, and executive function- nally obese individuals, may suffer health risks habits, had the lowest risk of mortality com- ing.69 Studies independently linking diet to associated with the endocrine actions of adipose pared with every other weight strata and health may also help explain why lower fruit tissue. However, independent pathways related lifestyle combination.63 Over 6 years, US adults and vegetable consumption is not always an to healthy dietary consumption and CRF may who experienced changes in fatness and fitness indicator of obesity.70 Obese individuals may provide obese individuals protection from displayed alterations in their risk for incidence still consume a diet high in nutritious quality, adverse health outcomes. This may aid in of , hypertension, and hy- without experiencing weight loss. explaining obesity paradoxes; some obese indi- percholesterolemia.64 This reduction in risk viduals may not suffer from increased mortality persisted even when the changes in fatness Adipose Tissue Effects on Health risk and may be metabolically healthy,20 whereas were controlled for changes in fitness and The mentioned studies indicate that over- some thinner individuals may be metabolically when changes in fitness were controlled for weight and obese individuals may be physi- compromised despite their “healthy” BMI. changes in fatness. However, these adjustments cally fit, consume a nutritious diet, and expe- for changes in fitness or fatness did attenuate rience high cardiometabolic health. This HEALTH-AT-EVERY-SIZE AS the reductions seen in health risk. In Katzmar- supports the hypothesis that 2 separate A NEW PARADIGM zyk and Lear’s65 systematic review on the mechanisms are at work in obesity-related effect of physical activity on chronic disease conditions. A specific pathway involves excess The public appears to be increasingly sup- risk factors in obese individuals, only modest adipose tissue, particularly visceral adipose portive for less weight-centric approaches to benefits for obese individuals with respect to tissue, affecting health. A second pathway public health.72,73 HAES advocates healthy chronic disease risk factors were evident. influences health through lifestyle factors, such diets and health-sustaining physical activity, However, even in intervention groups without as physical activity and diet, independent of fat but is weight neutral. Its proponents do not a dietary component, significant improvements and size. Of consequence in this regard is view weight as an adequate indicator of health were produced in some studies in blood pres- recent research into adipose tissue and its or weight loss as an appropriate end goal. sure, insulin, glucose, triglycerides, C-reactive functioning as an endocrine organ to affect HAES promotes consuming healthful foods, protein, and cholesterol measures. both health and behavior. honoring internal cues of hunger and satiety, All these studies exemplify what are referred Adipose tissue develops in numerous parts engaging in enjoyable physical activity, and to as obesity paradoxes. These are studies that of the body, mainly between muscle and skin, advocating against fat stigma and in support of provide findings that seemingly contradict although it is also present surrounding internal social justice, rectifying health disparities, and obesity epidemiology orthodoxy.20 The first organs.71 These different depots have dissimi- for ethical treatment of people of all sizes.11,74 recognized is that obese in- lar metabolic profiles. Adipose tissue functions The dietary component of HAES is based on dividuals with CVD are more likely to survive in providing mechanical support, insulation, an intuitive eating model.11, 7 4 Intuitive eating than their normal weight counterparts. Other and as a storage site for excess fuel.71 Such fuel operates on the assumption that is it external obesity paradoxes identified by McAuley and is stored as triglycerides and released as fatty cues, such as being coaxed to eat to excess in Blair20 include that being overweight (BMI = acids in response to hormonal and sympathetic familial or social settings or engaging in weight- 25.0---29.9) may be protective in relation to signaling. Additionally, adipose tissue releases loss dieting, that leads to chronic eating prob- mortality compared with being of normal its own endocrine signals, such as leptin. These lems and contributes to diet-related ill health. weight; that a large component of overweight signals include adipokines and other cell types Intuitive eating gains support from early studies and obese populations are metabolically that help regulate food intake and energy conducted on toddlers by Clara Davis, which healthy; and that being physically fit, even expenditure.71 These cells have effects on were subsequently replicated in more recent when obese, may eliminate the association numerous body systems and functions includ- studies.75,76 These studies examined the diets of between obesity and increased mortality. ing energy homeostasis, blood pressure, the small children who were allowed to eat from an Evidence relating to diet and obesity may immune system, and the complex, highly reg- assortment of nutritious foods. No exhortations eventually be synthesized into an additional ulated, process of adipose tissue expansion. were given to eat more or less in general or of obesity paradox. Diet may have an effect on Leakage of nutrients into other organs (e.g., any particular food. Researchers found that health and mortality, independent of weight liver, muscle) during periods of nutritional young children, although exhibiting substantial measures. Healthy diet was related to a signifi- overflow, or excessive or altered adipokines, variety in consumption at meals and occasional cantly reduced risk of all-cause or CVD mortality may result in abnormal somatic functioning.71 food jags, ate relatively the same amount of among US, German, and British samples.66---68 The dysregulated activities of adipokine- calories and macronutrients over a 24-hour Joseph et al.69 hypothesized that physical related processes form the physiological period over 6 days. The children always had activity may also improve eating behavior. underpinnings of numerous obesity-related sufficient energy for activity. Therefore, children Physical activity may induce neurocognitive morbidities including type II diabetes mellitus, will self-compensate and can meet their own

Published online ahead of print December 12, 2013 | American Journal of Public Health Bombak | Peer Reviewed | Framing Health Matters | e5 FRAMING HEALTH MATTERS

nutritional needs in the absence of external to continue in behaviors that have an inde- Columbia’s antiobesity campaigns, high levels cues.75,76 In practice, intuitive eating essentially pendent benefit on health, rather than poten- of support were found for adopting weight- advocates eating what one wants when hungry tially growing discouraged and abandoning neutral public health language.83 Adopting and stopping when one is full.76 As early as healthy behavioral efforts. antiweight bias training in medical school was 1991, this approach was proposed as a method deemed feasible. Also supported was the for preventing obesity. This suggests that in- SYNTHESIS OF BEST APPROACHES implementation of obesity research guidelines tuitive eating has long been viewed as a method that ensures research includes measures of of achieving a healthful body size, even among In summary, an exclusive focus on either socioeconomic status, diet, physical activity, those who problematize obesity. individualistic or obesogenic accounts of obesity and weight cycling to better delineate their Critics of HAES fear size acceptance may may be flawed in producing effective health independent effects on health, and reports lead to excessive consumption and weight promotion policy. Individualistic weight-loss fo- potential adverse outcomes of lifestyle changes gain.11 However, in the 6 completed random cuses are largely ineffective and may be stig- or weight loss. Unfortunately, less support was control trials of HAES interventions found no matizing. Furthermore, weight loss may not be found for government-funded HAES studies, negative effects, including weight gain, and necessary for health improvement in obese hindering the development of a more substan- there was maintenance or improvement in persons and may be harmful. It is undeniable tial HAES evidence base. Proceeding with behavioral, psychological, clinical, and physio- that a person’s sociocultural, economic, and initial small changes in terms of language and logical outcomes.11,77 Intriguingly, when an physical circumstances affect their well-being, education may facilitate future willingness to HAES intervention among premenopausal but the current obesogenic environment rheto- fund such research, however. For example, women employed a control group undergoing ric presents too simplistic a conceptualization of well-evaluated HAES teacher-resources that a weight-loss intervention that incorporated such a complex and dynamic process. As a re- encourage healthy behaviors, without trigger- a social support component, no significant sult, it tends to oversimplify important socio- ing body-image issues or eating disorders, have differences in eating behaviors between the economic, cultural, and ethnic considerations of been produced and may be used in place in groups was evident over time. However, de- well-being, and homogenizes diverse life ways. more weight-centric school health programs.84 creases in hunger and overall energy intake in At worst, it may mask other forms of discrimi- An analogous approach to obesogenic envi- some HAES participants were evident.78,79 nation in its focus on “target populations.”81 ronment proponents’ focus on broader struc- Also encouraging is that even with weight The most effective method of addressing tural issues is Health in Every Respect regain, maintained behavioral practices, such chronic disease may be to focus on health (HIER).85 HIER expands the sphere of HAES as those produced in HAES interventions, have rather than weight. By encouraging physical to include a more holistic perspective on health sustained health benefits.80 Among Israeli activity and healthy nutrition among all in- that incorporates socioeconomic and psycho- adult participants in a long-term weight-loss dividuals, everyone, regardless of weight status, social dimensions. It investigates more intricate intervention, partial weight regain occurred may benefit with respect to health and well- health-effecting pathways, rather than overly over time. This partial weight regain stimulated being. Individuals’ mental and physical health relying on nutrition and physical activity, and deterioration in a number of biomarkers may benefit from environments more suited to offers a critique of evidence-based medicine. (insulin, leptin, triglycerides, monocyte chemo- active transportation and increased access to It defies the healthism and individualism attractant protein 1, chemerin, and retinol- produce, whether significant aggregate reduc- that may permeate existent public health binding protein 4). However, sustained dietary tions in obesity are also produced. Further- approaches. HIER advocates political action that alterations produced benefits in other bio- more, weight stigma, which may produce more directly implicates the socioeconomic markers (fetuin-A, hsCRP, adiponectin, HDL-C, harmful weight and appearance preoccupation factors that are far more relevant to health than progranulin, and vaspin levels), in spite of in all individuals, may be ameliorated by lifestyle factors.85 This approach may extend partial weight regain.78 Thus, it may be that as focusing on creating salubrious environments the individual benefits of a HAES-approach to excess adipose tissue exerts independent effects and encouraging healthy behaviors for all more environmental reforms, and underscores on cardiometabolic health, so too do sustained people. This is likely more effective than the commitment of HAES to social justice and dietary behaviors. This finding is very impor- focusing exclusively on obese individuals and ethical practice.11,74,85 Most importantly, this tant given the high rate of weight regain among on a potentially unattainable goal, such as approach aligns with compelling international weight loss losers.5,6 Adopting a HAES ap- weight loss.46 Obesity rates are declining.82 evidence that the most effective approach to proach appears to produce longer-lasting Therefore, preventing and treating chronic improving population health is to redress so- behavioral changes than do weight-loss inter- disease, rather than promoting weight loss, may cioeconomic disparities.86 Previous weight- ventions.11,77 These maintained behavioral be a greater public health priority, regardless of centric approaches to public health have changes may induce healthier somatic func- weight-loss proponents’ views on the benefits largely been ineffective.9 However, evidence tioning, regardless of weight status. By not of weight loss for population health. suggests that weight-neutral physical activity using weight changes as a marker for health, Initiating such changes within the current and nutrition-based approaches may be individuals may be less discouraged by weight public health environment may be difficult to a promising new direction for encouraging stabilization or gain. They may be more likely implement. In a policy analysis of British lasting wellness in all individuals. j

e6 | Framing Health Matters | Peer Reviewed | Bombak American Journal of Public Health | Published online ahead of print December 12, 2013 FRAMING HEALTH MATTERS

About the Author 12. Ogden LG, Stroebele N, Wyatt HR, et al. Cluster environmental interventions for obesity. Prev Med. Andrea Bombak is with the University of Manitoba, analysis of the national weight control registry to identify 1999;29(6 I):563---570. Winnipeg. distinct subgroups maintaining successful weight loss. 31. Swinburn BA, Sacks G, Hall KD, et al. The global Correspondence should be sent to Andrea Bombak, Dept Obesity (Silver Spring). 2012;20(100):2039---2047. obesity pandemic: shaped by global drivers and local Community Health Sciences, S113-750 Bannatyne Ave., 13. Obesity: Preventing and Managing the Global Epi- environments. Lancet. 2011;378(9793):804---814. Winnipeg, MB Canada, R3E 0W3 (e-mail: umbombak@ demic. Technical Report Series 894. Geneva, Switzer- 32. Kriendler SA. Lifting the burden of chronic disease: cc.umanitoba.ca). Reprints can be ordered at http://www. land: World Health Organization; 2000. ’ ’ ajph.org by clicking the “Reprints” link. what has worked? What hasn t? What s next? Healthc Q. 14. Nevill AM, Stewart AD, Olds T, Holder R. Relation- This article was accepted May 28, 2012. 2009;12(2):30---40. ship between adiposity and body size reveals limitations of 33. Gortmaker SL, Swinburn BA, Levy D, et al. Chang- BMI. Am J Phys Anthropol. 2006;129(1):151---156. Acknowledgments ing the future of obesity: science, policy, and action. 15. Jutel A. The emergence of overweight as a disease Lancet. 2011;378(9793):838---847. A. Bombak has received funding from the Manitoba entity: measuring up normality. Soc Sci Med. 2006;63 Graduate Scholarship (2008-2010), Social Sciences and 34. Casazza K, Fontaine KR, Astrup A, et al. Myths, (9):2268---2276. Humanities Research Council (2009-2010), Manitoba presumptions, and facts about obesity. N Engl J Med. Health Research Council (2010-2012), Western Re- 16. Flegal KM, Graubard BI, Williamson DF, Gail MH. 2013;368(5):446---454. Excess deaths associated with , overweight, gional Training Centre for Health Services Research 35. Kirk SFL, Penney TL, McHugh TF. Characterizing and obesity. JAMA. 2005;293(15):1861---1867. (2010-2011), and the Canadian Institutes of Health the obesogenic environment: the state of the evidence Research Doctoral Research Award (2012-2014). 17. Flegal KM, Graubard BI, Williamson DF, Gail MH. with directions for future research. Obes Rev. 2010;11 The author is grateful for helpful commentary from fi Cause-speci c excess deaths associated with under- (2):109---117. Sharon Bruce. weight, overweight, and obesity. JAMA. 2007;298 36. Elinder LS, Jansson M. Obesogenic environments - (17):2028---2037. aspects on measurement and indicators. Public Health Human Participant Protection 18. Orpana HM, Berthelot J, Kaplan MS, Feeny DH, Nutr. 2009;12(3):307---315. No institutional review board approval was required for McFarland B, Ross NA. BMI and mortality: results from 37. Diez Roux AV, Mair C, eds. Neighborhoods and this literature review and commentary because it did not a national longitudinal study of Canadian adults. Obesity involve human participants. (Silver Spring). 2010;18(1):214---218. health. Ann N Y Acad Sci. 2010;1186:125---145. 19. Cohen-Mansfield J, Perach R. Is there a reversal in 38. Giskes K, Kamphuis CBM, Van Lenthe FJ, Kremers S, References the effect of obesity on mortality in old age? J Aging Res Droomers M, Brug J. A systematic review of associations between environmental factors, energy and fat intakes 1. Bray GA. A Guide to Obesity and the Metabolic 2011;2011;Epub ahead of print. among adults: is there evidence for environments that Syndrome. Boca Raton, FL: CRC Press; 2011:370. 20. McAuley PA, Blair SN. Obesity paradoxes. J Sports encourage obesogenic dietary intakes? Public Health Nutr. Sci. 2011;29(8):773---782. 2. Campos P. Does fat kill? A review of the epidemio- 2007;10(10):1005---1017. logical evidence. In: Rich E, Monaghan LF, Aphramor L, 21. Bosomworth NJ. The downside of weight loss: 39. McPhail D, Chapman GE, Beagan BL. “Too much of eds. Debating Obesity. London, UK: Palgrave MacMillan; realistic intervention in body-weight trajectory. Can Fam that stuff can’t be good”: Canadian teens, morality, and 2011:36---59. Physician. 2012;58(5):517---523. fast food consumption. Soc Sci Med. 2011;73(2):301--- 3. Maclean PS, Bergouignan A, Cornier M, Jackman 22. Wildman RP, Muntner P, Reynolds K, et al. The 307. MR. Biology’s response to dieting: the impetus for weight obese without cardiometabolic risk factor clustering and 40. Kim D, Leigh JP. Are meals at full-service and regain. Am J Physiol Regul Integr Comp Physiol. 2011;301 the normal weight with cardiometabolic risk factor fast-food restaurants “normal” or “inferior”? Popul Health (3):R581---R600. clustering: prevalence and correlates of 2 phenotypes Manag. 2011;14(6):307---315. 4. Sumithran P, Prendergast LA, Delbridge E, et al. among the US population (NHANES 1999---2004). Arch 41. An R, Sturm R. School and residential neighborhood Long-term persistence of hormonal adaptations to weight Intern Med. 2008;168(15):1617---1624. food environment and diet among California youth. Am J loss. N Engl J Med. 2011;365(17):1597---1604. 23. Aphramor L. Is a weight-centred health framework Prev Med. 2012;42(2):129---135. 5. Ikeda J, Amy NK, Ernsberger P, et al. The National salutogenic? Some thoughts on unhinging certain dietary Weight Control Registry: a critique. J Nutr Educ Behav. ideologies. Soc Theory Health. 2005;3(4):315---340. 42. Lee H. The role of local food availability in 2005;37(4):203---205. 24. Burns M, Gavey N. “Healthy weight” at what cost? explaining obesity risk among young school-aged chil- “ ” dren. Soc Sci Med 2012;74(8):1193---1203. 6. Gaesser G. Is “permanent weight loss” an oxymoron? Bulimia and a discourse of weight control. J Health The statistics on weight loss and the national weight Psychol. 2004;9(4):549---565. 43. Van Hook J, Altman CE. Competitive food sales in control registry. In: Rothblum E, Solovay S, eds. Bio- 25. Green GC, Buckroyd J. Disordered eating cognitions schools and : a longitudinal study. politics and the “Obesity Epidemic.” New York, NY: New and behaviours among slimming organization competi- Sociol Educ. 2012;85(1):23---39. York University Press; 2009:37---40. tion winners. J Hum Nutr Diet. 2008;21(1):31---38. 44. Carlson A, Frazão E. Are healthy foods really more 7. Methods for Voluntary Weight Loss and Control. 26. Puhl RM, Heuer CA. Obesity stigma: important expensive? It depends on how you measure the price. EIB- Bethesda, MD: National Institutes of Health, Office of considerations for public health. Am J Public Health. 96, US Department of Agriculture, Economic Research Medical Applications of Research; 1992. 2010;100(6):1019---1028. Service. 2012. Available at: http://www.ers.usda.gov/ publications/eib-economic-information-bulletin/eib96. 8. Goodrick GK, Poston WSC II, Foreyt JP. Methods for 27. Schwartz MB, Chambliss HO, Brownell KD, Blair SN, aspx. Accessed November 26, 2014. voluntary weight loss and control: update 1996. Nutri- Billington C. Weight bias among health professionals tion. 1996;12(10):672---676. specializing in obesity. Obes Res. 2003;11(9):1033--- 45. LeBesco K. Neoliberalism, public health, and the moral perils of fatness. Crit Public Health. 2011;21 9. Mann T, Tomiyama AJ, Westling E, Lew A, Samuels 1039. (2):153---164. B, Chatman J. Medicare’s search for effective obesity 28. Ernsberger P. Does social class explain the connec- treatments: diets are not the answer. Am Psychol. tion between weight and health? In: Rothblum E, Solovay 46. Kirkland A. The environmental account of obesity: 2007;62(3):220---233. S, eds. The Fat Studies Reader. New York, NY: New York a case for feminist skepticism. Signs (Chic). 2011;36 10. Rosenbaum M, Hirsch J, Gallagher DA, Leibel RL. University Press; 2009:25---36. (2):463---486. Long-term persistence of adaptive thermogenesis in 29. Egger G, Swinburn B. An “ecological” approach to 47. Colls R, Evans B. Critical geographies of fat/bigness/ subjects who have maintained a reduced body weight. the obesity pandemic. BMJ. 1997;315(7106):477---480. corpulence. Antipode. 2009;41(5):1011---1020. Am J Clin Nutr. 2008;88(4):906---912. 30. Swinburn B, Egger G, Raza F. Dissecting obesogenic 48. Guthman J, DuPuis M. Embodying neoliberalism: 11. Bacon L, Aphramor L. Weight science: evaluating environments: the development and application of economy, culture, and the politics of fat. Environ Plan D. the evidence for a paradigm shift. Nutr J. 2011;10:9 . a framework for identifying and prioritizing 2006;24(3):427---448.

Published online ahead of print December 12, 2013 | American Journal of Public Health Bombak | Peer Reviewed | Framing Health Matters | e7 FRAMING HEALTH MATTERS

49. Warin M, Turner K, Moore V, Davies M. Bodies, case-control study. Eur J Clin Nutr. 2005;59(10):1201--- 83. O’Reilly C, Sixsmith J. From theory to policy: mothers, and identities: rethinking obesity and the BMI. 1207. reducing harms associated with the weight-centred health Sociol Health Illn. 2008;30:97---111. 67. Kant AK, Leitzmann MF, Park Y, Hollenbeck A, paradigm. Fat Studies. 2012;1(1):97---113. 50. McPhail D. What to do with the “tubby hubby”? Schatzkin A. Patterns of recommended dietary behaviors 84. McVey G, Gusella J, Tweed S, Ferrari M. A “Obesity,” the crisis of masculinity, and the nuclear family predict subsequent risk of mortality in a large cohort of controlled evaluation of web-based training for teachers in early cold war Canada. Antipode. 2009;41(5):1021--- men and women in the united states. J Nutr. 2009;139 and public health practitioners on the prevention of 1050. (7):1374---1380. eating disorders. Eat Disord. 2009;17(1):1---26. 51. McPhail D. “This is the face of obesity”: gender and 68. Akbaraly TN, Ferrie JE, Berr C, et al. Alternative 85. Aphramor L, Gingras J. Helping people change: the production of emotional obesity in 1950s and 1960s healthy eating index and mortality over 18 y of follow- promoting politicised practice in the health care pro- Canada. Radical Psychology. 2010;8(1). up: results from the Whitehall II cohort. Am J Clin Nutr. fessions. In: Rich E, Monaghan LF, Aphramor L, eds. 2011;94(1):247---253. Debating Obesity. London, UK: Palgrave MacMillan; 52. Guthman J. Weighing In: Obesity, Food Justice, and the 2011:192---218. Limits of Capitalism. Berkeley, CA: University of Califor- 69. Joseph RJ, Alonso-Alonso M, Bond DS, Pascual- nia Press; 2011:227. Leone A, Blackburn GL. The neurocognitive connection 86. Wilkinson R, Pickett K. The Spirit Level: Why Greater between physical activity and eating behaviour. Obes Rev. Equality Makes Societies Stronger. New York, NY: ’ 53. Warin M. Foucault s progeny: Jamie Oliver and the art 2011;12(10):800---812. Bloomsbury; 2010:352. of governing obesity. Soc Theory Health. 2011;9:24---40. 70. Fransoo R, Martens P, Prior H, et al. Adult Obesity in 54. Pardo Silva MC, De Laet C, Nusselder WJ, Mamun Manitoba: Prevalence, Associations, and Outcomes. Win- AA, Peeters A. Adult obesity and number of years lived nipeg, Manitoba: Manitoba Centre for Health Policy; with and without . Obesity (Silver October 2011. Spring). 2006;14(7):1264---1273. 71. Sethi JK, Vidal-Puig AJ. Thematic review series: 55. Carroll MD, Kit BK, Lacher DA. Total and high- adipocyte biology. Adipose tissue function and plasticity density lipoprotein cholesterol in adults: National Health orchestrate nutritional adaptation. J Lipid Res. 2007;48 and Nutrition Examination Survey, 2009---2010. Na- (6):1253---1262. tional Center for Health Statistics. 2012. Available at: 72. Lewis S, Thomas SL, Hyde J, Castle D, Blood RW, http://www.cdc.gov/nchs/data/databriefs/db92.htm. Komesaroff PA. “I don’t eat a hamburger and large chips Accessed November 26, 2014. every day!” A qualitative study of the impact of public 56. Freedman DS, Goodman A, Contreras OA, health messages about obesity on obese adults. BMC DasMahapatra P, Srinivasan SR, Berenson GS. Secular Public Health. 2010;10:309. trends in BMI and blood pressure among children and 73. Julie in Ontario. Our health, our future: national adolescents: the Bogalusa Heart Study. Pediatrics. dialogue on healthy weights—shift the focus away from 2012;130(1):e159---e166. health. 2011 Available at: http://ourhealthourfuture.gc. 57. May AL, Kuklina EV, Yoon PW. Prevalence of ca/2011/04/06/shift-the-focus-away-from-weight. cardiovascular disease risk factors among US adolescents, Accessed October 8, 2011. 1999-2008. Pediatrics. 2012;129(6)1035---1041. 74. Burgard D. What is “Health At Every Size”?In: 58. Blair SN, Church TS. The fitness, obesity, and health Rothblum E, Solovay S, eds. The Fat Studies Reader.New equation: is physical activity the common denominator? York, NY: New York University Press; 2009:41---53. JAMA. 2004;292(10):1232---1234. 75. Scheindlin B. “Take one more bite for me”: Clara 59. Kokkinos P, Sheriff H, Kheirbek R. Physical in- Davis and the feeding of young children. Gastronomica. activity and mortality risk. Cardiol Res Pract. 2011; 2005;5(1):65---69. 2011;Epub ahead of print. 76. Birch LL, Johnson SL, Andresen G, Peters JC, Schulte ’ 60. Lee CD, Blair SN, Jackson AS. Cardiorespiratory MC. The variability of young children s energy intake. N Engl J Med. 1991;324(4):232---235. fitness, body composition, and all-cause and cardiovas- cular disease mortality in men. Am J Clin Nutr. 1999;69 77. Bacon L, Stern JS, Van Loan MD, Keim NL. Size (3):373---380. acceptance and intuitive eating improve health for obese, female chronic dieters. J Am Diet Assoc. 2005;105(6): 61. Mota J, Santos RM, Silva P, Aires L, Martins C, Vale S. 929---936. Associations between self-rated health with cardiorespi- ratory fitness and obesity status among adolescent girls. 78. Provencher V, Bégin C, Tremblay A, et al. Health-At- J Phys Act Health. 2012;9(3):378---381. Every-Size and eating behaviors: 1-year follow-up results of a size acceptance intervention. J Am Diet Assoc. 62. Aryana M, Li Z, Bommer WJ. Obesity and physical 2009;109(11):1854---1861. fitness in California school children. Am Heart J. 2012; 163(2):302---312. 79. Leblanc V, Provencher V, Begin C, Corneau L, Tremblay A, Lemieux S. Impact of a Health-At-Every- 63. Matheson EM, King DE, Everett CJ. Healthy lifestyle Size intervention on changes in dietary intakes and eating habits and mortality in overweight and obese individuals. patterns in premenopausal overweight women: results of J Am Board Fam Med. 2012;25(1):9---15. a randomized trial. Clin Nutr. 2012;31(4):481---488. 64. Lee DC, Sui X, Church TS, Lavie CJ, Jackson AS, 80. Blüher M, Rudich A, Klöting N, et al. Two patterns of fi Blair SN. Changes in tness and fatness on the develop- adipokine and other biomarker dynmaincs in a long-term ment of cardiovascular disease risk factors. J Am Coll weight loss intervention. Diabetes Care. 2012;35(2): Cardiol. 2012;59(7):665---672. 342---349. 65. Katzmarzyk PT, Lear SA. Physical activity for obese 81. Cooper C. Fat lib: How fat activism expands the individuals: a systematic review of effects on chronic obesity debate. In: Rich E, Monaghan LF, Aphramor L, disease risk factors. Obes Rev. 2012;13(2):95---105. eds. Debating Obesity: Critical Perspectives. London, UK: 66. Zyriax BC, Boeing H, Windler E. Nutrition is Palgrave MacMillan; 2011:164---191. a powerful independent risk factor for coronary heart 82. Gard M. The End of the Obesity Epidemic. New York, disease in women—the CORA study: a population-based NY: Routledge; 2011:193.

e8 | Framing Health Matters | Peer Reviewed | Bombak American Journal of Public Health | Published online ahead of print December 12, 2013