Secondary Logo

Journal Logo


Epidemiologic Research in the Face of an Obesity Epidemic

Potter, John D.

Author Information
doi: 10.1097/01.ede.0000197405.28469.a3
  • Free

There is a commonsense view of obesity that says, simply, that it is the result of an imbalance between energy intake (food) and energy output (exercise). Although this view obviously has considerable merit, as well as testable implications for intervention, it is an incomplete explanation. Energy–flux imbalance is necessary but is probably not sufficient. Thus, it may be reasonable to ask: what other conditions have to be present to induce obesity? That some genetic influences are in play may seem obvious but, to date, the evidence is sparse. However, there are some rare conditions in which a genetic defect is closely associated with hyperphagia and obesity, and these may hold clues.

Second, eating disorders are one way in which energy balance can be disturbed. Eating disorders are either more common now or more frequently diagnosed. Do they cast any light on obesity? Third, what explains the paradoxical relationship between poverty and obesity in the developed world? Finally, the 20th century was characterized by an epidemic of coronary heart disease—an epidemic that began to decline for, as yet, unclear reasons in Western countries in the 1960s and 1970s. There have also been epidemics of obesity in previous centuries. Is it possible that these are epidemics in the communicable disease sense?


The prevalence of overweight in the United States has increased markedly over the last 10 to 15 years such that, by the beginning of the 21st century, the prevalence of a body mass index (BMI) of 25 kg/m2 or greater was greater than 55% in almost every state in the Union.1

This trend is not restricted to the United States nor to adults. Indeed, the increase in the prevalence of overweight children from 1974 to 1997 in Brazil was approximately 3-fold to almost 15%.2 In contrast, there was little increase in the prevalence of overweight among children in China during the 1990s, and in Russia, there was actually a decline, almost certainly reflecting the impact of the economy.2 In the United States, the prevalence of overweight children has risen from approximately 15% in 1971 to more than 25% in 1994,2 consistent with the adult data.


At its simplest, this large increase in obesity must reflect an imbalance between energy intake and output. Although this is almost certainly a necessary condition, it is clearly not sufficient. An obvious starting point in our understanding is to consider obesity as a risk factor—or as an intermediate variable between some other exposure and disease—for a variety of disorders (eg, diabetes, coronary heart disease, hypertension, and some cancers), all of which may be clustered together in various ways (and with other diseases such as gout) to form the so-called metabolic syndrome.

However, it is also possible to think of obesity as a dependent variable and to think of the factors that influence energy balance. These include not only eating patterns and exercise, but also germline genetics, the perinatal environment, imprinting and epigenetics, and reproductive history. Behaviors such as exercise and eating patterns, as well as the perinatal environment, are variously influenced by the physical environment, cultural practices, socioeconomic status, and genetics. Some of these factors are further influenced by agriculture, the food supply, and advertising (Fig. 1). Thinking about disease causation in this way—with multiple biologic intermediates as well as multiple social, behavioral, and biologic steps—does not fit easily with the way in which epidemiologists have approached causal thinking nor does it lend itself easily to our approaches to data analysis (see also McMichael3).

The chain of disease causality that includes obesity.


The central controls on appetite and energy are not fully understood, but they include both sympathetic and endocrine influences. Among the latter are leptin secreted by fat and insulin by the pancreas and a variety of peptide hormones secreted by the intestine (eg, cholecystokinin) and stomach (eg, ghrelin).4 Prader-Willi syndrome has usually been thought of as a genetic model of obesity. It is the result of the loss of expression of paternally derived alleles of maternally imprinted genes at 5q11-13. The perinatal phenotype includes fetal growth retardation and, from age 2 years onward, hyperphagia, developmental delay, and short stature.5 Children with this disorder have elevated plasma ghrelin levels and show a positive correlation between the ghrelin levels and hunger. However, unlike in normal individuals, ghrelin levels do not decline with food intake.6 Holland and colleagues5 have proposed that, rather than being a genetic model of obesity, the phenotype is the result of a starvation signal (high ghrelin) in the presence of abundant food. One question worthy of further research is whether there are attenuated versions of this syndrome, and whether they might contribute to the risk of obesity in a society such as ours in which food is abundant and cheap.

At least one other rare genetic model of hyperphagia exists among carriers of mutations in the melanocortin-4 receptor gene, for whom the concurrent risk of binge eating and obesity is very high.7 Polymorphisms in at least one previously suspected gene (GAD2), however, do not appear to be related to obesity.8

Insulin and insulin-like growth factors (IGFs) at abnormal doses or timing in animals are also associated with obesity. Administering exogenous insulin to a rat at week 3 of pregnancy or to the pups themselves in the first 2 postnatal weeks results in obese offspring.9 Both insulin and IGFs affect various aspects of brain and nervous system maturation, and the administration of maternal insulin results in obese offspring that have structural and functional changes in the hypothalamus that may explain the phenotype.9

Other aspects of early-life exposures alter growth and maturation rates. For instance, high fat exposure in utero causes the early onset of puberty in mice.10 Furthermore, maternal waist-to-hip ratio, particularly in the presence of obesity, is an important predictor of larger offspring in humans, Overall, each 0.1-unit increase in waist-to-hip ratio is associated with 120 g greater birth weight, 5 mm greater birth length, and 3 mm greater head circumference; these increases are twice as great among the offspring of women with a BMI of 30 kg/m2 or more.11

Thus, inherited mutations and early-life exposures induce permanent changes that influence eating behavior, growth rates, obesity, and disease risk. At least some of these exposures probably induce epigenetic set-point changes rather than genetic damage. This growing body of evidence that very early-life exposures have consequences for adult disease poses special challenges for the way in which epidemiologists design studies as well as for the size of those studies. How much of the population obesity burden these influences may explain is quite unclear.


Fat, salt, and sugar are all rare in nature, with the exception of salt for coastal dwellers. Thus, there has been no evolutionary pressure to develop feedback loops to prevent overconsumption. Furthermore, although being obese would have been a serious disadvantage to our hominid ancestors, the tendency to obesity (the ability to store fat in the seasons of abundance to survive lean times) would have been a specific advantage, and this tendency is likely to have been selected for, particularly in the northern hemisphere up to the last Ice Age approximately 12,000 years ago.

Accordingly, some eating disorders, particularly binge eating, may have some genetic underpinning over and above the specific severe genetic disorders noted previously. That we live in a culture where food is abundant and cheap, and where competitive eating is now a “sport,” compounds this tendency. Other influences may turn binge eating into a way of life, leading to obesity, as well as inducing even more psychologically disordered behavior such as bulimia nervosa. It is worth remembering, however, that the ancient Romans had massive feasts and that the essential vomitorium was found in all the better homes.


Indeed, concerns about eating habits, obesity, and exercise have been with us, at least in Western Europe, for over 2000 years. In 50 BC, Lucretius wrote: “In primitive times, lack of food gave languishing bodies to death; now, on the other hand, it is abundance that buries them.” Socrates is recorded as saying: “...moderate exercise reduces to order, according to their affinities, the particles and affections that are wandering about the body...”

Morgagni,12 the 18th century anatomist, described intraabdominal fat accumulation in android obesity and also described the association between such visceral obesity and hypertension, hyperuricemia, atherosclerosis, and obstructive sleep apnea.

A study by Als et al13 of a collection of portraits from a defined geographic region (Berne, Switzerland) over a considerable span of time—the 14th to 20th century—shows greater obesity among men than women and among those over 40 years than those younger. It also shows that, in the 18th century, almost 80% of men over 40 and more than 50% of women over 40 were overweight and that this proportion among men subsequently declined through to the 20th century. This 18th century obesity epidemic is also, more anecdotally, represented in the works of Hogarth.14 The 18th century, too, is notable for the prevalence of gout: Samuel Johnson and Benjamin Franklin are well-documented sufferers. Gout is related both to weight and weight gain.15 That this is self-induced and the result of dietary behavior can be attested to, in part by Boswell, Samuel Johnson's biographer, who noted “The goose is a silly bird: too much for one, not enough for two.”

Thus, ours is not the first prosperous society to suffer from the consequences of an overabundance of food, associated with an apparent ability to override satiation. This may be a complete explanation of human epidemics of obesity. However, it does raise a question as to whether these epidemics could be epidemics in the truly infectious sense. As we learn more about the role of infectious agents in cancer, cardiovascular heart disease, and other chronic diseases, could we entertain the possibility that infectious agents could also cause obesity? How might we explore this hypothesis with epidemiologic tools?


One difference between the current obesity epidemic and those of earlier times is the fact that, in the United States, it is particularly among the poor and the less educated that the highest prevalence of obesity is seen.16 Education, furthermore, is also positively associated with greater physical activity.16 Nonetheless, the trend toward increasing prevalence of obesity is seen in every educational and income stratum.


There has been a steady increase in the availability and consumption of empty calories from sugar. Enough carbonated and noncarbonated soft drinks are produced each year to provide the average American with 68 gallons and 85,000 kcal.17 Soda pop now provides more energy than any other single food in the U.S. diet.17

This trend is seen more widely, with the worldwide consumption of caloric sweeteners (sucrose, glucose, fructose, syrups, honey, and so on) having risen from around 225 kcal in 1962 to around 300 kcal in 2000.2 In the United States at the present time, there is a general increase in the intake of energy-dense, nutrient-poor foods and snacks, with a resulting increase in overall daily energy intake; restaurant portion sizes are increasing; there is a growing rate of consumption of salty snacks, fast foods, and added sugars; intake of vegetables and fruits remains low; and the average child is awake and sedentary (TV, games, movies) for 6 hours.2

The United States is not alone in its increasingly sedentary behavior and the relation between this behavior and obesity. For example, in China, where physical activity bears the same inverse relationship, for instance to colon cancer risk,18 as has been reported almost universally, the relative risk of obesity among individuals with a motor vehicle versus those without is 1.8.19 Fourteen percent of households in China acquired a motor vehicle between 1989 and 1997,2 and obesity is on the rise. Indeed, internationally, there is a strong correlation between per-capita car ownership and breast, intestine, prostate, kidney, and colon cancers but not with esophageal, lung, or thyroid cancers (Potter, unpublished data). Although these can legitimately be ignored as just ecologic data, the association between motor vehicle use and obesity on the one hand and the risk of a variety of cancers associated with obesity on the other makes this a plausible causal association and obesity the plausible intervening variable. Again, this raises issues, as suggested previously, about the way to undertake studies that account for complex causal chains, not just risk factors.


One proper test of the association between physical activity and obesity is the exercise intervention study. From our own work, we now have data that confirms the obvious—exercise in previously sedentary individuals reduces body mass and improves a number of other important aspects of health.20–24 Intervention strategies should be considered part of the epidemiologist's tool kit.


From the evolutionary perspective, there is probably selection pressure in favor of overeating and in favor of fat storage. In the developed world, social and environmental support for overeating is widespread, and cheap food is abundant. There are no immediate biologic consequences of overconsumption of fat and sugar. Similarly, selection pressure in favor of energy conservation has probably played a role as a survival mechanism. In the developed world (and, increasingly, everywhere), there is social and environmental support for inertia as well as an abundance of cheap transport. Finally, humans experience no immediate biologic feedback on the consequences of inertia and obesity.

Other factors may contribute to a constellation of causes for obesity that includes, but is not limited to, disruption of energy balance. What are the immediate causes of reduced physical activity and increased energy intake that seem to be occurring worldwide? In response to this question, we must turn, minimally, to aspects of the way in which we produce and distribute food, the influence of advertising, and the way in which we organize transportation. These are policy, not biology questions, but they are essential to understanding, and certainly to remediating, the obesity epidemic.

For epidemiologists, the challenge is to devise studies that help us understand the complexities of the causal nets that lead to obesity—and to identify strategies that disrupt the weakest links in the nets. For public health professionals, the challenge is to develop strategies that will break the links.


John D. Potter is Director of the Public Health Sciences Division of the Fred Hutchinson Cancer Research Center, Seattle. He has contributed extensively to our understanding of the causes and prevention of colorectal neoplasia. His current research focus is on genetic and environmental causes of colorectal cancer as well as on understanding obesity as both a risk factor and an outcome.


1. American Cancer Society Cancer Statistics Presentation 2004,Slide 34. Available at: Accessed November 1, 2005.
2. Popkin BM, Gordon-Larsen P. The nutrition transition: worldwide obesity dynamics and their determinants. Int J Obes Relat Metab Disord. 2004;28(suppl 3):S2–9.
3. McMichael AJ. Prisoners of the proximate: loosening the constraints on epidemiology in an age of change. Am J Epidemiol. 1999;149:887–897.
4. Evans RM, Barish GD, Wang YX. PPARs and the complex journey to obesity. Nat Med. 2004;10:355–361.
5. Holland A, Whittington J, Hinton E. The paradox of Prader-Willi syndrome: a genetic model of starvation. Lancet. 2003;362:989–991.
6. DelParigi A, Tschop M, Heiman ML, et al. High circulating ghrelin: a potential cause for hyperphagia and obesity in Prader-Willi syndrome. J Clin Endocrinol Metab. 2002;87:5461–5464.
7. Branson R, Potoczna N, Kral JG, et al. Binge eating as a major phenotype of melanocortin 4 receptor gene mutations. N Engl J Med. 2003;348:1096–1103.
8. Swarbrick MM, Waldenmaier B, Pennacchio LA, et al. Lack of support for the association between GAD2 polymorphisms and severe human obesity. PLoS Biol. 2005;3:e315.
9. Levin BE. The obesity epidemic: metabolic imprinting on genetically susceptible neural circuits. Obes Res. 2000;8:342–347.
10. Hilakivi-Clarke L, Stoica A, Raygada M, et al. Consumption of a high-fat diet alters estrogen receptor content, protein kinase C activity, and mammary gland morphology in virgin and pregnant mice and female offspring. Cancer Res. 1998;58:654–660.
11. Brown JE, Potter JD, Jacobs DR Jr, et al. Maternal waist-to-hip ratio as a predictor of newborn size: results of the Diana Project. Epidemiology. 1996;7:62–66.
12. Enzi G, Busetto L, Inelmen EM, et al. Historical perspective: visceral obesity and related comorbidity in Joannes Baptista Morgagni's ‘De sedibus et causis morborum per anatomen indagata’. Int J Obes Relat Metab Disord. 2003;27:534–535.
13. Als C, Stussi Y, Boschung U, et al. Visible signs of illness from the 14th to the 20th century: systematic review of portraits. BMJ. 2002;325:1499.
14. University of Wales, Lampeter. The Hogarth Archive. Available at: Accessed October 25, 2005.
15. Choi HK, Atkinson K, Karlson EW, et al. Obesity, weight change, hypertension, diuretic use, and risk of gout in men: the health professionals follow-up study. Arch Intern Med. 2005;165:742–748.
16. Drewnowski A, Specter SE. Poverty and obesity: the role of energy density and energy costs. Am J Clin Nutr. 2004;79:6–16.
17. Jacobson MF. Liquid candy: how soft drinks are harming Americans’ health. In: Interest CfSitP. Washington, DC: Center for Science in the Public Interest; 2005:1–35.
18. Hou L, Ji BT, Blair A, et al. Commuting physical activity and risk of colon cancer in Shanghai, China. Am J Epidemiol. 2004;160:860–867.
19. Bell AC, Ge K, Popkin BM. The road to obesity or the path to prevention: motorized transportation and obesity in China. Obes Res. 2002;10:277–283.
20. McTiernan A, Tworoger SS, Rajan KB, et al. Effect of exercise on serum androgens in postmenopausal women: a 12-month randomized clinical trial. Cancer Epidemiol Biomarkers Prev. 2004;13:1099–1105.
21. Irwin ML, Yasui Y, Ulrich CM, et al. Effect of exercise on total and intra-abdominal body fat in postmenopausal women: a randomized controlled trial. JAMA. 2003;289:323–330.
22. McTiernan A, Tworoger SS, Ulrich CM, et al. Effect of exercise on serum estrogens in postmenopausal women: a 12-month randomized clinical trial. Cancer Res. 2004;64:2923–2928.
23. Atkinson C, Lampe JW, Tworoger SS, et al. Effects of a moderate intensity exercise intervention on estrogen metabolism in postmenopausal women. Cancer Epidemiol Biomarkers Prev. 2004;13:868–874.
24. Frank LL, Sorensen BE, Yasui Y, et al. Effects of exercise on metabolic risk variables in overweight postmenopausal women: a randomized clinical trial. Obes Res. 2005;13:615–625.
© 2006 Lippincott Williams & Wilkins, Inc.