Genes, Behavior, and the Social Environment: Moving Beyond the Nature/Nurture Debate (2006)

Myles S. Faith, Ph.D. and Tanja V. E. Kral, Ph.D.*

Weight and Eating Disorders Program

Obesity is one of the most pressing public health disorders in the United States and other westernized societies. Its prevalence is increasing worldwide and it is associated with concerning medical comorbidities, most notably the metabolic syndrome and type 2 diabetes [1-4]. Hence, innovative research that elucidates the causes of obesity has become an increasingly important focus for the National Institutes of Health. A challenge to this mission, however, is that fact that obesity is a “complex disorder.” For most individuals in the population, obesity results from multiple genetic and environmental factors that may interact with, or may be correlated with, each other. Genes operate additively and through gene-gene interactions to influence body weight [5].

The topic of genetic and social environmental influences on obesity, and how they interact, is a unique topic for which conceptual frameworks are scarce. Research within each domain appears to have advanced largely within independent “camps,” each of which has undergone major advances in the past decade. Research into the genetics of human obesity has become increasingly sophisticated with respect to molecular technologies, biostatistics, and efficient design strategies; however, as illustrated in this report, these studies generally did not measure specific aspects of the social environment. Research into social environmental influences on obesity has expanded its scope

of coverage from interpersonal variables to potential consequences of a broader “toxic environment;” however, these studies generally did not collect DNA or use genetically informative designs. Hence, there appears to be room for greater scientific synergy between the domains.

There are two overarching aims to the present report: (a) to review evidence for genetic and social-environmental influences on obesity, respectively, and the types of methodologies used to establish these associations, and (b) to consider opportunities for greater methodological synergy between the two domains. The report strives to foster ideas for new research that bridge genetic and social-environmental research, as they relate to obesity and obesity-promoting behaviors. Conceptual frameworks that posit potential interactions or covariation among genetic and social environmental factors are proposed.

Figure C-1 presents the conceptual framework around which the present report is organized. The model posits that genetic and social-environmental factors promote obesity through their independent influences on intermediary behavioral variables. These intermediary phenotypes may induce a positive energy balance (i.e., greater energy intake than expenditure) that, when sustained, promotes obesity. Although physiological variables are not depicted in the model, they clearly are central to energy balance regulation and the putative behavior phenotypes listed in the figure. The model is intended to reflect much of the current literature, in that correlations or interactions among the social environment and genetic factors are not explicitly posited. However, as reviewed in this report, certain studies challenge this assumption and suggest that expansions of this model may help guide future research. The final section of this report suggests additional research that would test interactions and correlations among genetic and social-environmental variables.

The following section of the report, Section 3, addresses putative social-environmental influences on obesity-promoting behaviors and obesity, corresponding to pathways b and c in Figure C-1. Section 4 addresses evidence for selected refined behavioral traits that have been associated with obesity in some studies, corresponding to the “putative behavioral phenotypes” noted in the figure. Section 5 addresses putative genetic influences on obesity-promoting behaviors and obesity, corresponding to pathways a and c in the figure. Section 6 addresses evidence for potential interactions among genetic, social, environmental, and behavioral influences on obesity. The data presented in this section challenge the premise that genetic and environmental factors do not interact or cannot influence each other. Section 7 suggests additional research questions and designs that

FIGURE C-1 Conceptual model relating genetic and social-environmental factors to obesity. In this figure, the effects of genetic and social-environmental factors, respectively, are posited to operate through putative behavioral phenotypes that promote positive energy balance. Although not depicted, genetic and social-environmental factors are posited to impact on physiological variables as well.

might test new questions concerning the interplay between genes, social environment, behavior, and obesity.

It should be noted that the term “obesity,” used throughout this report, was not necessarily measured in the same way across all the reviewed studies. Most studies defined obesity based on the body mass index (BMI; kg/m²), which is a reasonable proxy measure of total body fat, at least in population studies. Guidelines by the National Heart, Lung, and Blood Institute stipulate a BMI between 25.0 and 29.9 as “overweight,” and greater than 30.0 as “obese.” More refined body composition measures were used in some studies.

Given the range of topics covered in this report, a table of contents for the major report sections and subsections is provided for the reader (Table C-1).

For the purposes of this report, a broad definition of “social environment” is used. Specifically, as defined by Barnett and Casper [6], “Human social environments encompass the immediate physical surroundings, social relationships, and cultural milieus within which defined groups of

people function and interact. Components of the social environment include built infrastructure; industrial and occupational structure; labor markets; social and economic processes; wealth; social, human, and health services; power relations; government; race relations; social inequality; cultural practices; the arts; religious institutions and practices; and beliefs about place and community. [ … ] Social environments can be experienced at multiple scales, often simultaneously, including households, kin networks, neighborhoods, towns and cities, and regions.”

This section reviews evidence for potential social-environmental influences on obesity and obesity-promoting behaviors, corresponding to paths b and c in Figure C-1. The social-environmental variables include two “macroenvironmental” variables and two “microenvironmental” variables. Macroenvironmental factors operate across larger communities or populations, specifically, exposure to components of the “toxic environment” and socioeconomic status (SES); “microenvironmental” factors, on the other hand, refer to smaller groups of individuals or family members, specifically, the “social facilitation” of overeating that occurs in group settings and parent-child feeding dynamics. The social-environmental variables reviewed below are not necessarily independent of each other, but are presented individually for ease of presentation.

The two macroenvironmental factors reviewed below are (i) exposure to the “toxic environment” and (ii) SES. These particular factors are reviewed because there is a reasonable database providing information on these variables and because of their potential relevance for obesity prevention.

Brownell coined the term “toxic environment” [7, 8], referring to a pervasive series of social and economic changes that have occurred in the United States during in the past several decades. Brownell argues that these changes have caused the rising obesity prevalence, even though strong causal inferences cannot be easily made from these observational trends. These changes are outlined in detail elsewhere [9-12], but include the increased portion sizes and the “super-sizing” of commercially available foods, the proliferation of fast-food restaurants, the reduced cost of fast-food products, the increasing access to energy-dense foods in schools, the increased use of labor saving devices that reduce physical activity, and reduced opportunities for physical activity in schools and at safe playgrounds.

Data have been published that are consistent with the notion that some of these changes may have contributed to the rising obesity prevalence. As reviewed elsewhere [13], for example, data on national food supply and utilization from the U.S. Marketing System indicate that the overall energy availability per capita in the United States increased by 15 percent between 1970 and 1994, a period during which there was also an increase in per capita availability of dietary fat, increased consumption of added fats (commonly found in snack or confectionary foods), reduced milk intake, and increased soft-drink intake. During this period, there was an increased number of households with two or more television sets, home video recorders, and home computers.

Despite these findings, several caveats are warranted. First, although these aforementioned findings are consistent with a causal influence (i.e., pathways b and c in Figure C-1), evidence for a causal relationship per se is limited [13]. Much of the evidence comes from observational studies that could not control for potential confounding factors or did not directly test associations between participant weight status and exposure to putative environmental risk factors. Second, specific aspects of the “toxic environment” that have the greatest impact on obesity are unknown [13]. Third, findings from certain studies did not support expected predictions. For example, in a cohort of over 7,000 children who were 36 to 59 months of age and from low-income families, child obesity status was not associated

with access to playgrounds, proximity to fast-food restaurants, or neighborhood crime level [14].

Finally, it has not been tested whether exposure to the toxic environment is related to genotype. That is, individuals with obesity-predisposing genes may be particularly responsive to the effects of such a “toxic” environment. In addition, certain individuals may be more likely to seek out or expose themselves to aspects of the toxic environment. The topic of gene-environment correlations as a topic for additional research is discussed further in Section 7.

Several studies (e.g., [15-17]) have documented an inverse relationship between SES and obesity in previous years. In a recent review, Ball and colleagues [18] examined 34 articles to test the hypothesis that persons from lower SES strata are at increased risk of weight gain. Their hypothesis was supported for predominantly non-African American samples, but not for African American samples. Reviewing relevant studies, they found little support for a relationship between SES and weight gain among African Americans. In contrast, depending on the particular indicator for SES that was used (i.e., occupational status, education, and income), they found that lower SES was associated with an increased risk of weight gain in non-African American individuals. Specifically, the authors found an inverse association between occupational status and weight gain for men and women. When SES was assessed using education as the indicator, the relationship became less strong (particularly among men). Using income level as the particular indicator for SES, findings for associations between weight gain and SES were inconsistent for both men and women. Finally, the authors noted a differential rate of weight gain by SES and attributed that finding to an early onset of weight gain in a person’s life, when parental SES may still be influential.

Prospective analyses of the National Longitudinal Survey of Youth [19] found that children from lower SES families were more likely to have been overweight during the prior year than children from higher SES families. Negative associations between obesity status and household income and parental education were found even when controlling for ethnicity and other demographic variables.

Several mechanisms could underlie the link between low SES and obesity. Factors such as limited access to resources, poor knowledge of nutrition and health, increased exposure to fast-food outlets, and limited physical activity due to deprived or unsafe neighborhoods [20, 21] have been suggested to influence energy intake and energy expenditure and, consequently, body weight. For instance, in an ecological study of 267 postal districts in Melbourne, Australia, families living in the poorest SES strata

had 2.5 times the exposure to fast-food outlets and thus increased access to relatively inexpensive, calorically dense foods compared to families from the wealthiest SES strata [22].

The relationship between SES and obesity may also be influenced by differential costs of less or more nutritious foods. For instance, in a series of elegant analyses, Drewnowski documented that the cost of healthy, nutrient-dense foods such as fruits and vegetables were reliably more expensive than more energy-dense, less nutritious foods [23-25]. Possibly for this reason, the availability of fruits and vegetables in adolescents’ homes was shown to be greater among families from high compared to low SES strata [26]. These data suggest that families from lower SES strata have overall fewer monetary resources to purchase more nutrient-dense, healthy foods [23, 25, 27].

Reduced access to recreational facilities or parks in deprived neighborhoods also may contribute to diminished energy expenditure and thus increased body weight in individuals of lower SES [28].

In summary, lower SES may contribute to the onset of obesity in that it provides an environment which promotes the intake of calorically dense foods while it reduces the need or the opportunity for physical activity.

The two microenvironmental influences reviewed in this section are social facilitation of eating and parental feeding practices. These particular factors are reviewed because there is a reasonable database providing information on these variables and, in regards to feeding practices, because of its potential relevance for obesity prevention.

There is reliable evidence that total energy intake at meals is increased significantly when eating in the presence of other people, a phenomenon termed “social facilitation” [29]. This phenomenon would be represented by pathway b in Figure C-1. De Castro [30] studied 63 adults who maintained a 7-day continuous food diary and recorded the number of people present at each meal. Results indicated that energy intake during meals that were eaten alone was significantly lower compared to energy intake during meals that were consumed in the presence of others. This was observed for total energy intake (410 vs. 591 kcals), carbohydrate intake (190 vs. 241 kcals), fat intake (157 vs. 230 kcals), and protein intake (65 vs. 100 kcals). Satiety ratings were 30 percent greater following meals eaten with others compared to meals eaten alone.

Additional analyses of de Castro’s data indicated that the social facili-

tation effect was greater for meals consumed in the presence of a spouse, family member, or friend compared to less familiar or unknown companions, suggesting that enhanced social interactions and discussions were the underlying mechanisms [31]. Indeed, de Castro and de Castro [30] argued that physiological signals that relate to appetite and meal size can be overridden by social interactions. Specifically, they found that reported total energy intake at meals was positively correlated with time since prior meal consumption, but only for meals eaten alone. When others were present at meals, there was no longer a significant association, suggesting that post-prandial meal regulation may be “disrupted by the presence of other people” (p. 246).

Laboratory studies have also demonstrated this social facilitation phenomenon. Edelman et al. [32] showed that overweight and normal-weight subjects consumed more lasagna when eating in groups of 4 or 5 persons compared to when eating alone, and that there was no significant difference between the weight groups in terms of this phenomenon. Klesges et al. documented the social facilitation effect in a restaurant setting, with the effect being more pronounced for women than men. Kimm and Kissileff [33] also demonstrated the social facilitation of eating in a cafeteria setting.

The mechanism underlying social facilitation of eating has been termed “time-extension” [29, 34] and has received the most empirical support. Specifically, the presence of people at a meal serves to lengthen meal time which, in turn, promotes further energy intake. The point is important to the present paper because, as presented in Section 5, there is evidence that the tendency to eat with others may be genetically influenced. Thus, the fact that some individuals are more likely to eat in the presence of others may not be a random event; rather, eating in the presence of others may be a trait that is influenced by genes that indirectly promote social facilitation of eating at meals.

An area of active research concerns parental feeding practices and parent-child feeding dynamics that might promote a positive energy balance and overweight in young children. Review of this literature reveals two specific feeding practices that are prospectively associated with increased body weight and weight gain in infants and children. These practices are, first, bottle-feeding as opposed to breast-feeding, and, second, parental use of restrictive child feeding practices. With respect to breast-feeding practices, prospective epidemiology studies have shown that childhood and adolescent obesity rates were reduced among infants who were breast-fed as opposed to never breast-fed [35] and among infants who were breast-fed for longer compared to shorter durations [36, 37]. In one seminal study, the prevalence

of overweight was studied in 8,186 girls and 7,155 boys, 9 to 14 years of age, who were participating in a national growth and development study [38]. Among children who were mostly or exclusively breast-fed during the first 6 months of life, compared to children who were mostly or exclusively formula-fed, the odds ratio for being overweight was 0.78. This held true when controlling for maternal BMI and other variables reflecting SES and lifestyle activities. It should be noted that not all studies replicated this significant association [39], and that one study found the association to be true in non-Hispanic white families but not African American families [40].

The mechanisms for the apparent protective effect of breast-feeding on overweight development were unknown, although recent data implicate parental feeding patterns as a possible factor. Specifically, mothers who breast-fed their infants were less restrictive in their feeding practices (as measured by self-report questionnaire) than mothers who bottle-fed their infants [41]. As discussed in the next section, restriction of child eating may impede a child’s ability to self-regulate food intake and instead teach a child to eat in response to external cues [42]. Whether or not this is the actual mechanism needs to be clarified in future research.

An extensive literature has examined which parental feeding practices, if any, are associated with increased child food intake during meals and increased weight status [43]. Investigators have measured feeding practices by parent-report questionnaires, direct observation, or analysis of videotapes, with the most common assessment tool being the parent-report Child Feeding Questionnaire [44]. A recent review of this literature concluded that, across the range of parental feeding domains that have been studied, only restriction of child eating was consistently associated with increased child total energy intake and weight status [43]. Parents who restrict their children’s access to foods tend to have heavier children. No other feeding domains were associated with childhood obesity, including use of food to calm infants and children, feeding on schedule, pushing child to eat more, and provision of structure during feeding, or using food as a reward [45, 46].

Several mechanisms by which parental restriction may promote increased child energy intake and body weight have been proposed. First, restrictive feeding practices may impede on a child’s ability to adhere to internal hunger and satiety cues (i.e., impaired self-regulation) and thereby teach children to eat in response to external cues (e.g., portion size, time of day). Among preschool children, the ability to self-regulate food and energy intake across meals was poorer among children whose parents reported elevated efforts to control child eating [42]. Second, restricting children’s

access to foods may have the counterproductive effect of making those “forbidden” foods more desirable [47]. Third, restriction of foods may teach children to eat in the absence of hunger, that is, to continue eating despite being full when food is available [48].

At the same time, the body of evidence suggests that parental restriction of child eating is elicited, at least in part, by a child’s increased body weight [43, 49]. Indeed, in one study, the association between restrictive feeding practices and increased child weight gain was only seen in children who were born at high risk for obesity [49]. As in other realms of child development, there appears to be a bidirectional association such that parental restriction of child eating partially is elicited by child’s weight, which in turn may exacerbate further child weight gain. This also suggests a possible gene-environment correlation such that genes and environmental conditions that promote childhood obesity are interrelated. The topic of gene-environment correlations is discussed in Section 7.

This section reviews refined behavioral traits that have been associated with obesity in cross-sectional or prospective investigations. As such, it addresses the putative behavioral phenotypes listed in Figure C-1. Obesity results from an imbalance between energy input and energy output. The daily energy surplus that is necessary to promote weight gain is small; specifically, Hill et al. [50] estimated that a sustained daily energy surplus above a person’s daily energy requirements as small as 100 kcal/day is sufficient to promote weight gain. For this reason, it is desirable to identify refined behavioral traits that are related to positive energy balance and obesity. Identifying such intermediary traits may help elucidate the pathways through which the social environment and/or genes promote obesity.

In the 1970s and early 1980s there was much interest in identifying an “obese eating style” [51-57] which differentiates lean and obese individuals’ eating behavior. It has been argued that intraindividual differences in various eating behaviors may underlie the disparity in energy intake and body weight among both groups. In light of the recent obesity epidemic, the search for distinctive patterns of food intake among individuals with differing body sizes continues to be of great importance. Following is a description of selected eating traits which may represent behavioral phenotypes of obesity.

During the late 1960s results from a series of experiments conducted by Schachter and colleagues [58-60] suggested that the eating behavior of obese individuals is greatly influenced by the immediate (food) environment. In particular, the eating behavior of obese individuals was believed to be controlled by external cues related to the perception of time, taste and sight of food, and the number of highly palatable food cues present [53, 60, 61], rather than by internal physiological cues of hunger.

Subsequent studies [62, 63] failed to replicate consistent differences between lean and obese individuals in their responsiveness to external food-related and non-food-related cues. These studies found large intraindividual variability among individuals across all weight groups in their response to external cues. However, this early research on “external eating” developed into a more promising line of research on the trait of dietary “disinhibition.”

Disinhibition refers to the loss of self-imposed cognitive control of eating behavior in response to external or emotional stimuli, and is the behavioral trait that most consistently differentiates between obese and nonobese individuals [64]. Obese subjects show greater disinhibition scores than do nonobese individuals [65, 66] and degree of disinhibition is strongly associated with energy intake [64, 67], weight status and weight gain [68, 69], weight fluctuations [65], binge eating [70], and body fat [71].

In summary, dietary disinhibition, a characteristic that associated with external eating, may represent a behavioral phenotype which is relevant to obesity and obesity-related traits.

In recent years there has been much debate over whether obesity is the result of impairment in the regulation of energy intake. One way to study food and energy intake in individuals is to examine satiation (or intrameal satiety). Satiation refers to the process leading to the termination of eating. It is assessed by measuring food and energy intake during a single meal which subjects consumed ad libitum.

To date only a limited number of studies is available that investigated the effects of dietary manipulation on satiation in both normal-weight and overweight/obese subjects. A study conducted by Bell and Rolls [72] was designed to examine the effects of energy density across three levels of dietary fat on intake in both lean and obese women. Results demonstrated that the energy density of the meals significantly affected subjects’ energy intake across all levels of dietary fat. The response to the dietary manipulation was similar between lean and obese women. All women consumed approximately 20 percent less energy in the condition of low energy density compared to high energy density.

Likewise, studies which examined the effects of varying the portion size [73, 74] or the portion size and the energy density of food [75] on subjects’ ad libitum intake found no significant difference in the eating response of lean and obese individuals. Both groups consumed significantly more energy when the portion size or the portion size and the energy density of food were increased. A longitudinal study [76] conducted in children analyzing nutritional data from nationally representative databases (i.e., CSFII 94-96; NFCS 77-78) found that portion sizes of commonly consumed foods were positively related to children’s energy intake and body weight.

As outlined above, laboratory studies for the most part failed to detect significant differences between lean and obese individuals in their response to the dietary manipulation of the energy density and/or portion size. One of the great difficulties in accurately assessing food intake in obese populations has been their altered eating behavior when being monitored. As several studies on self-reported food intakes have indicated, obese individuals underreport their intakes to a greater extent than do lean individuals [77]. The measured energy intakes of obese subjects in a controlled laboratory setting may likewise be compromised by the fact that their food intake is being monitored.

Despite these null findings, there is some evidence that when self-selecting their diets obese individuals tend to consume overall greater amounts of foods that are higher in energy density than do their lean counterparts. In a study conducted by Westerterp-Plantenga et al. [78] obese women reported consuming larger portions and an overall greater percentage of their total energy intake from foods that are higher in energy density than did lean women.

There is some evidence [79] of a difference in the pattern of cumulative intakes within a meal between lean and obese individuals. While lean individuals showed a decrease in their eating rate over the course of a meal, obese and latent obese as well as restrained subjects [80] failed to do so. The authors suggested that this difference in the pattern of cumulative intakes over the course of a meal may indicate that lean and obese individuals experience satiation differently.

Recent findings from neuroimaging studies confirmed intrameal differences between lean and obese individuals. It has been shown that the hypothalamic response following glucose ingestion was significantly delayed (~4-9 min) in obese individuals compared to their normal-weight counterparts [81]. These findings suggest that obesity may be associated with an abnormal neuronal activity in certain regions of the brain [82], some of which are believed to cause a delayed response in satiation over the course of a meal.

In summary, the finding of a potentially delayed satiation in obese individuals is of interest in that it may point to differences in the experience of hunger and fullness between lean and obese individuals. Innovative re-

search designs need to be developed to further study satiation as a possible phenotype for obesity.

Another approach to examine energy intake regulation among individuals is the study of satiety. Satiety, defined as the effects of a food or a meal after eating has ended [83], can be studied by administering a fixed amount of a given food or nutrient (preload) and, after a predetermined delay, measure its effects on subsequent intake (test meal).

Among adults, there is conflicting evidence that obese individuals experience satiety differently and compensate for energy less accurately than do lean individuals. Data generated from an experiment that was designed to compare effects of carbohydrate and fat on eating behavior in lean and obese individuals [84] suggest that obese restrained females show a relative insensitivity to the satiating power of fat in that they did not adjust their energy intake as well as did their lean counterparts after the ingestion of a high-fat preload. Outcomes from other investigations [85, 86], however, failed to detect differences in caloric compensation (i.e., satiety) among individuals with differing body sizes.

Studies have found that young children have the ability to adjust food intake at test meals in response to preloads, although compensation often is incomplete and differs between children. Johnson and Birch [42] found that children with poorer caloric compensation abilities tended to be heavier than children with better compensation abilities. On the other hand, other studies have failed to detect this same association in young children [87]. Thus, whether or not this trait reliably relates to a child’s proneness for obesity remains to be further investigated.

In summary, the degree to which an individual is able to compensate for energy may represent an eating trait that distinguishes the lean from the obese. It is possible that a predisposition for obesity moderates developmental changes in compensation ability as environmental factors start to override internal feelings of hunger and satiety.

The reinforcing value of food can be defined as the extent to which an individual will work for a given food or food group when an alternative commodity (e.g., money) is concurrently available. Typically assessed on a computer keyboard that required “bar presses” on the keyboard, the reinforcing value of food represents the highest amount of work (i.e., bar presses) an individual will emit to earn access to food. Thus, the measure represents “drive” or hedonic motivation for foods (i.e., food reward). The

paradigm is based on behavioral economics theory, which builds upon an extensive animal literature and research in the additions [88]. In a series of controlled studies, Epstein and colleagues have found that obese individuals score higher on measures of food reward than nonobese individuals [89-91]. This trait has proven to be one of the more consistent behavioral phenotypes that relates to weight status and, as described in Section 6, has even been linked to specific genes related to dopamine pathways.

In 1962, Ferster and colleagues put forward the idea that obese individuals take larger bites and eat faster than do normal-weight individuals and that the obese would eat less if they ate more slowly [92]. Subsequent experiments tested potential differences between lean and obese individuals in their eating style, including rate of eating, bite size, and the amount and rate of chewing.

Early work by Dodd et al. [52] found that obese individuals ate more, ate at a faster rate, and took in larger bites than nonobese individuals. Some investigators [93] confirmed that obese individuals ate faster than lean individuals, however, others [57, 94] did not replicate this finding. The conflicting outcomes may have been due to methodological issues related to how the rate of eating during a meal was manipulated, as well as the failure to control for meal size in early studies.

An interesting finding has been the difference in the rate of sucking in infants who were born at high or low risk for obesity based on maternal pre-pregnancy BMI. That is, at 3 months of age, infants born at high risk for obesity displayed greater nutritive sucking rates on an artificial nipple than did infants born at low risk for obesity [95]. Moreover, among all infants, increased sucking rate was predictive of increased weight gain during the first two years of life [95, 96].

In adults, the rate of eating appears to be related to food consumption in both obese and nonobese individuals. Spiegel et al. [97] tested the effects of bite size on ingestion rate, satiation and meal size, and found that decreasing the bite size of test foods was associated with a lower ingestion rate for the whole meal. Interestingly, this decrease in the rate of eating was offset by an increase in meal duration such that overall meal sizes did not differ across conditions. This result was found true for both lean and obese individuals.

Spiegel [98] gave lean and obese men access to a buffet-style meal during which they could choose between different flavors, kinds of foods, and make their own sandwiches. Results showed that lean and obese subjects did not differ significantly in their average bite size of different foods, local ingestion rate (g ingested/min), chew efficiency (g ingested/chew), and

chew frequency (chews/s). However, obese men consumed more energy per minute than lean men, a difference that was due to the higher energy density of the foods consumed by obese subjects, in particular the greater energy density of the sandwiches. Thus, there may be an interrelation between a greater rate of eating and the tendency to eat more energy-dense foods among obese individuals. Lean and obese individuals may respond similarly to the physical properties of foods, while they may be differing in the food preferences and food choices they make, which can promote a positive energy balance.

While there has been much interest in identifying distinct taste qualities that are more or less preferred by lean versus obese populations, these studies have yielded mixed results. For instance, studies conducted to examine differences in taste preferences between lean and obese individuals [99, 100] suggested that obesity is associated with an overall heightened preference for high-fat stimuli. Others, however, could not confirm these findings [101, 102] in that they failed to find differences between lean and obese individuals for overall pleasantness scores or liking for foods with different predominant taste qualities.

Another difficulty is to find evidence for the conception that sensory preferences influence food choices in both lean and obese individuals. Epidemiologic data have documented a positive relationship between weight status and dietary fat intake [103, 104]. Based on a comprehensive review of animal, epidemiological, and clinical studies exploring the relation between fat intake and obesity, Bray and Popkin [105] concluded that dietary fat is an important contributor to obesity in certain individuals. Likewise, data from a cross-sectional study [106] using food frequency questionnaires indicated a positive association between the consumption of red meats, fish, oil, poultry, eggs, fats, oils, and condiments and BMI and a negative association between the consumption of legumes, soy, tofu, fruit juice, cold cereals, and vegetables and BMI.

In summary, the consumption of certain foods and/or macronutrients (i.e., dietary fat) has been associated with increased weight status, however, this relationship may not hold true for all individuals. In general, it has been difficult to establish clear associations between weight status and intake of single foods or food groups [107].

Fast-food restaurants and the promotion of predominantly calorically dense, relatively inexpensive foods, are considered by some to be the

cornerstones of the “toxic” environment. That is, increased consumption of fast foods has been associated with increased weight status. Among 891 adults enrolled in the “Pound of Prevention Study,” greater frequency of eating fast foods was significantly associated with higher total energy intake, higher percent fat intake, more frequent consumption of hamburgers, french fries, and soft drinks, and less frequent consumption of fiber and fruit. Over 3 years, each additional fast-food meal/week was associated with an excess weight gain of 0.72 kg beyond the average weight gain observed during that period. In a prospective study of over 3,000 young adults enrolled in the Coronary Artery Risk Development in Young Adults (CARDIA) study, frequency of fast-food restaurant visits at baseline (visits/week) predicted excess 15-year weight gain and worsening of insulin resistance in Caucasian and African American respondents [108].

Visiting fast-food restaurants may promote obesity by promoting increased consumption of energy-dense foods. Prentice and Jebb [109] reviewed the nutritional content of the foods sold at three popular fast-food outlets. The average energy densities for the three menus were 1.7-fold greater than the average British diet. Other potential mechanisms by which fast-food restaurants may promote a positive energy balance include the increased portion sizes of foods (e.g., “super-sizing”).

In conclusion, fast-food establishments may put certain individuals at an increased risk for the overconsumption of calories. As noted in Section 7, additional research is needed to test whether certain obesity-promoting genotypes moderate this association (i.e., a gene-environment interaction) or whether individuals with certain genotypes may be more likely to seek out such restaurants (i.e., a gene-environment correlation).

Living in the modern-day environment has decreased the need for individuals to be physically active. Decreased physical activity, and thus decreased energy expenditure, has been negatively associated with BMI [110-112] and maintenance of weight loss [68, 113]. The following two sections will highlight two activity-related activities in particular. One is television viewing and its association with weight status, the other is nonexercise activity thermogenesis (NEAT).

Increased television viewing (TVV) has been associated with increased energy intake and body weight [114]. The mechanisms for this association

are multifold. For one, increased TVV increases sedentary behavior which in turn is likely to displace time spent in physical activity.

Second, TVV also provides a setting during which food, especially energy-dense snack foods, can be consumed. A study conducted by Francis et al. [114] showed that TVV viewing was associated with increases snack food consumption in girls who were 5, 7, and 9 years old which in turn predicted girls’ increase in BMI from age 5 to 9. Thus, TVV has been shown to be a risk factor for excessive snack consumption and in turn increased weight status, especially for those individuals who are predisposed for obesity. Another study [115] also demonstrated that physical activity (negatively associated) and TVV (positively associated) were the only significant predictors, beyond baseline BMI, of BMI in children between the ages of 3 and 4 years during a 3-year study phase.

A third mechanism by which increased TVV may lead to increased energy intake may involve the increased exposure to food advertising. A study conducted by Henderson and Kelly [116] was designed to analyze the content of food advertising appearing on either general market or African American TV programming. The results of the study showed that African American TV programs included more food advertisements, more advertisements for unhealthy foods such as fast food, candy, soda, or meat, and made more weight-related claims and those related to the fat content of foods compared to advertisements that appeared in general market television. Thus, food advertisements seem to be targeted at and tailored to specific populations to increase product sales.

In summary, television viewing has been associated with increased energy intake and weight status among individuals. It remains to be further investigated whether TVV is a behavior that, through its association with sedentarianism, may be fostered through an individual’s biology, as the following section on NEAT alludes to.

NEAT has been defined as energy expenditure that is associated with daily activities such as sitting, standing, walking, and talking and as such is different from purposeful, planned physical activity [117]. NEAT can further be divided into thermogenesis that is associated with posture (standing, sitting, and lying) and that associated with movement (ambulation). Research conducted by Levine and colleagues [117] has shown that obese individuals, on average, were seated longer per day and spent less time in an upright position, compared to lean individuals. Overall this difference accounted for an additional energy expenditure of 352 calories per day, on average, for lean individuals. Interestingly, the difference in NEAT was not due to the differential body weights of the study participants per se, but

seemed to be inherent in an individual’s biological/genetic makeup. That is, even after the study team had obese subjects lose a considerable amount of weight (8 kg) and had lean individuals gain weight (4 kg), the two subject groups did not change their original posture allocation.

These data suggest that interindividual differences in posture allocation (i.e., NEAT) may be genetically determined. The authors of the respective research [117] speculate that “( … ) obese and lean individuals respond differently to the environmental cues that promote sedentary behavior” (p. 586). This type of research, again, provides a rich ground to further integrate more genetic-based research with studies on the social environment.

This section reviews evidence for genetic influences on obesity and obesity-promoting behaviors, corresponding to paths b and c in Figure C-1. The section is divided into two subsections. Subsection 5a examines evidence for genetic influences on BMI and body fat measure, while subsection 5b examines evidence for genetic influences on obesity-promoting behaviors related to food intake (i.e., pathway b in Figure C-1). Within each subsection, data are presented for studies that estimate heritability of the phenotype, followed by studies that tested the influence of specific genes or genomic regions.

“Heritability” refers to the extent to which variability in a trait is influenced by genetic variations within a population, and can be subdivided into “narrow-sense” or “broad-sense” heritabilities [118]. The former refers solely to additive genetic influences on the trait, whereas the latter refers to nonadditive interactions among genes. Beyond heritability, the remaining variance in weight status is due to environmental influences which can be partitioned into “shared environment” or the “nonshared environment” influences. Shared environment refers to aspects of the home environment that are perfectly shared by siblings from the same home (e.g., food in the home cupboards, the number of television sets at home). The nonshared environment refers to those aspects of the environment that are uncorrelated among siblings (e.g., differential interactions with parents or peers, or differential life experiences). Apropos to this report, specific as-

pects of the environment are rarely directly measured in behavioral genetics studies of obesity. Consequently, investigators generally could not test the influence of putative social-environmental factors when modeling genetic influences on fat mass.

Heritability of BMI has been estimated from a variety of designs, but most commonly from twin and adoption studies that compared phenotypic correlations for body fat between groups of individuals varying in genetic relatedness. To the extent that fat mass is genetically influenced, phenotypic correlations will be greater among individuals who are more genetically similar (e.g., monozygotic, MZ, twins) than less genetically similar (e.g., dizyogotic, DZ, twins). Using biometric statistical models, heritability is estimated along with the magnitude of shared and nonshared environmental factors.

Maes et al. [118] put forward a most comprehensive review of this literature, the results of which provide indisputable support for a heritable component to BMI and fat mass. Heritability estimates fall in the range of 20 to 80 percent when estimated from family studies that compared parent-child and sibling correlations, 20 to 60 percent when estimated from adoption studies, and 50 to 90 percent when estimated from twin studies. Although the heritability estimates vary sizably, the most accurate estimates arguably come from twin studies, which have methodological strengths over other designs [118]. Also, studies of twins reared apart (i.e., twins separated during childhood and therefore not exposed to the same home environment) generally yield some of the higher heritability estimates. Stunkard et al. [119] conducted the first study of twins reared apart, the results of which estimated heritability at ~65 to 75 percent for BMI. Allison et al. [120] obtained comparable heritability estimates when pooling an international sample of twins reared apart from seven countries. Finally, the inconsistent results that have been reported across studies were likely due to small sample sizes.

Results of longitudinal behavior genetic studies suggest that there are age-specific genetic effects on BMI, such that different obesity-promoting genes may become active at different ages across the lifespan. This has been documented throughout the lifespan [121, 122]. Thus, although some genes exert a consistent influence over time and are partially responsible for the “tracking” of BMI, other genetic influences may appear at different stages of a child’s development. Apropos to the theme of this paper, considering developmental milestones may be especially important for future studies testing the interplay of genetic and social-environmental influences on obesity.

Beyond heritability, unmeasured genotype studies have provided clues into the nature of environmental influences on obesity. First, most studies find no evidence for shared home environmental influences on fat mass variability in adulthood. Rather, most environmental influences on fat

mass appear to be of the nonshared variety. In regards to this report, efforts to identify social-environmental influences on weight status might focus on unique life experiences that are unshared among family members. Second, among the few studies finding evidence for shared environmental influences on weight status, most examined pediatric samples. For example, in an analysis of over 3,500 twin pairs who were 4 years old, shared environmental factors accounted for 24 percent of variance in weight adjusted for height in boys and 25 percent of the variance in girls [123]. Jacobson and Rowe [124] reported shared environmental influences on BMI in white adolescent females, but not in African American adolescent females or adolescent males who were white or African American. Specific aspects of the shared environment were not tested in these studies.

The “Human Obesity Gene Map” is the most comprehensive annually updated compendium of specific genes that have been associated with obesity and obesity-related phenotypes (e.g., physiological and metabolic measures, peptides and hormones, and behavioral traits) [125]. Initially published in 1994, the Human Obesity Gene Map summarizes evidence from the following classes of human studies: (a) obesity due to single-gene or digenic mutations, (b) obesity associated with Mendelian disorders, such as Prader-Willi syndrome or Bardet-Biedl syndrome, (c) “association studies” that test whether candidate genes are associated with obesity phenotypes among samples of unrelated participants, and (d) “linkage studies” that test for causal associations between genomic regions and obesity phenotypes in cohorts of families. Within each of these categories, studies have been methodologically heterogeneous with respect to participant characteristics, sample sizes, phenotype measurements, and data analytic strategies.

From this voluminous literature, several broad conclusions about specific gene effects on weight status can be made. First, the number of genes shown to be statistically associated with fat mass and obesity-related traits increased dramatically over the past decade. This is true across each of the aforementioned study categories, as summarized in Table C-2. Thus, the platform of specific genes that might contribute to obesity is very large and involves loci throughout the genome. Similarly, the number of genes that might interact with the social environment could also be large.

Second, the failure to replicate positive associations has been a common occurrence. For example, although 204 genomic regions for obesity-related phenotypes were identified from 50 genome scans in the 2004 report, replication of positive findings was found for only 38 genomic regions. Similarly, although there were 358 significant associations with 113 candidate genes from association studies, only 18 positive associations were

replicated across five studies. Probable reasons for nonreplication including population stratification (i.e., the combination of individuals from different racial/ethnic backgrounds), publication bias, Type 1 errors, and insufficient statistical power [126].

Third, among the phenotypes investigated that were not body composition measures, the vast majority were metabolic or physiological measures rather than measures of food intake, appetite, or food preferences. Thus, as described below, behavioral measures have been largely unrepresented in genotype studies and this may represent an opportunity for future research.

Fourth, single-gene mutations likely account for a small percent of the cases of human obesity in the general population. For most obese individuals, obesity likely results from the influence of multiple genes on different chromosomes that work additively and through gene-gene interactions [5]. Among the cases of monogenic obesity reported in the literature, most have been related to mutations in the melanocortin 4 receptor (MC4R) gene [125].

Finally, for most individuals in the population, the specific physiological mechanisms by which genes influence obesity probably involve both energy intake and expenditure pathways. A detailed discussion on this topic is beyond the scope of this report, which is more geared towards behavioral phenotypes, but is provided elsewhere [127-129]. Several genes have been implicated in the regulation of energy expenditure, including those related to mitochondrial uncoupling proteins (i.e., UCP1, UPC2, and UCP3 genes), the adrenergic systems (i.e., β₂-AR and β₃-AR genes), and the growth and development of the adipocyte (PPARγ gene) [125, 127]. Although these genes are involved in energy expenditure pathways, Loos and Bouchard [127] point out that most genetics studies examined these genes in relation to obesity-related phenotypes and not energy expenditure phenotypes per se.

The physiological pathways related to appetite are complicated and, as reviewed by Badman and Flier [130], involve the integration of short-term satiety signals from the gut to the brain along with longer-term homeostatic systems. Figure C-2 depicts an overview of these physiological pathways, which involve the integrated signaling of POMC, AGRP, MC4R, and NPY systems. As Loos and Bouchard note [127], there have been relatively inconsistent findings linking obesity phenotypes to genes for these proteins. Perhaps the most encouraging findings in the literature involve the MC4R gene, which, as discussed in the next section, has been associated with human food intake in a few preliminary studies.

A relatively small number of studies have tested genetic influences on eating phenotypes, independent from body fat. They provide data pertinent

to pathway a in Figure C-1. These studies are summarized below, divided into those that estimated heritability and those that tested for specific genes or genomic regions.

A seminal investigation by de Castro studied adult twin pairs who recorded food and beverage intake continuously in the free-living environment for 7 days [131-134]. The initial cohort consisted of 109 identical and 86 fraternal twin pairs, who used diaries to record food and beverage intake, time of food consumption, amount of food consumed, food preparation methods, the number of other people present when eating, hunger and thirst levels, depression, anxiety, and perceived food attractiveness and palatability. These data lead to a series of publications on the broader genetic-environment architecture of adult food intake. In an initial report, significant heritability estimates were documented for reported total energy intake, weight of food intake, fat intake, carbohydrate intake, protein intake, and water intake, with the remaining variance due to nonshared environmental factors [132].

In a multivariate analysis that tested genetic and environmental influences on meal-specific energy intake, cumulative daily energy intake, and weight status, results supported the presence of independent genetic and nonshared environmental influences on meal-specific energy intake [133]. Genes accounted for 46 percent of the variance in the frequency with which meals were eaten and 56 percent of the variance in meal size, respectively. Thus, there may be genes and environmental influences on food intake at specific eating episodes that are different from those factors related to habitual dietary intake. Specific social-environmental influences on meal intake were not reported initially.

Gender differences in the heritability of food intake were reported in other twin cohorts (see below) and, apropos to this report, may be an important issue for studying the interplay of social-environmental and genetic influences on obesity.

There were genetic influences on the full range of appetite and eating phenotypes studied by de Castro [131-139] including premeal hunger levels, time of meals, premeal stomach content, energy intake of high-palatability foods, energy intake of low-palatability foods, and overconsumption of foods rated as being high versus low in palatability. Apropos to this report, the environmental conditions that promote social facilitation of eating may be genetically influenced, suggesting a gene-environment correlation. These data challenge the conceptual framework in Figure C-1, suggesting that covariation between genes and the social-environmental should be modeled. As de Castro concluded, “genes appear to affect the

physiologic, psychologic, and social context in which eating occurs. In the past, heredity and environment were seen to operate separately on behavior. The present results indicated that there may not be such a clear separation […] heredity ends up having a strong influence on the nature of the environment in which individuals immerse themselves” (p. 554). This issue is revisited in the final section of this report.

Building upon de Castro’s work, other twin studies tested the heritability of dietary intake patterns. In a study of 4,640 adult twins who completed the National Cancer Institute food-frequency questionnaire, the heritability of consuming foods high in fat, sugar, and salt was 15 percent for women and 30 percent for men [140]. The heritability of consuming “healthy foods,” including fruits and vegetables, was 15 percent for women and 30 percent for men. Thus, heritability estimates were larger for men than women and, for women only, there were significant shared environmental influences on both food categories.

A recent study of 5,250 male and female twin pairs, 16 years of age, examined the heritability of frequency of breakfast consumption [141]. Results indicated significant gender differences, in that heritability estimates were higher for boys while shared environmental influences were higher for girls. The authors concluded that “Breakfast eating is moderated differently in adolescent boys and girls. Unlike boys, girls are much influenced by the family and pair-specific environment. In girls, environmental influences may override genetically driven factors” (p. 512).

With one exception, all studies testing the heritability of eating traits used self-report measures of dietary intake. A recognized drawback to this method is the fact that respondents tend to underreport food intake, a finding that is more common among obese individuals [142]. To bypass this problem, Faith et al. [143] studied 36 MZ and 18 DZ twins who consumed a buffet lunch in a controlled feeding laboratory. The meal provided servings of 27 foods and beverages, including chicken nuggets, hot dog sandwiches, apples, grapes, carrots, chocolate cookies, and donuts. Results indicated that shared environmental factors had the biggest influence on total energy intake, accounting for 48 percent of the variance. Additive genetic factors accounted for 33 percent of the variance, and nonshared environmental factors accounted for 19 percent of the variance. Thus, the results suggest that both genes and the shared environment can influence total energy intake, although specific aspects of the environment were not measured. The study also suggests the potential value of laboratory-based protocols for study eating phenotypes, a point that is reviewed in the final section of this report for future research directions.

There is a dearth of information on specific genes that influence dietary patterns, at least for most individuals in the population [144]. There is a subset of individuals whose obesity resulted from single gene mutations and who were markedly hyperphagic [125]; however, these individuals are relatively uncommon in the population. Studies that used association or linkage designs to identify specific genes or genomic regions in larger cohorts have been relatively uncommon and represent an area for future research.

A series of reports documented associations between the serotonin (5-hydroxy-tryptamine, 5-HT) receptor gene and reported energy intake. Because serotonin has been shown to reduce food intake and may play a role in the etiology of eating disorders, it was deemed an appropriate candidate gene by some investigators. Aubert et al. [145] studied 276 unrelated overweight and obese adults who were genotyped for the −1438 GA polymorphism of the 5-HT_2A receptor gene. Results indicted that reported daily energy intake from 3-day food records was significantly associated with genotype. Specifically, individuals carrying the A allele of the gene (i.e., genotypes G/A or A/A) consumed less total energy per day than individuals not carrying the A allele (i.e., genotype G/G). Comparable findings were made by the same investigators in an analysis of 370 children and adolescents, 10 to 20 years old, who were participants in the Stanislas Family Study. Youth carrying the A allele consumed less daily energy intake than youth not carrying the allele, even when controlling for age, sex, weight, and height [146].

Associations between MC4R polymorphisms and eating traits have been reported, further implicating the role of this gene in obesity onset. In a study of 500 children with severe obesity, 5.8 percent of the sample was found to have mutations in the MC4R gene that were not found in nonobese controls [147]. The investigators then compared the ad libitum test meal intake of children who had fully inactive or partially inactive MC4R polymorphisms. When served a standardized breakfast, children with fully inactive mutations consumed more food than children with partially inactive mutations. In another study, MC4R genotype was associated with Binge Eating Disorder status in a cohort of severely obese Caucasians adults and nondieting controls [148].

Several linkage studies investigated genomic regions associated with reported Dietary Restraint, Eating Disinhition, and Hunger levels, as measured by the Eating Inventory (EI) [149]. Steinle et al. [150] studied 624 related individuals, from 28 Amish families, who completed the EI. Results of a genome-wide linkage analysis revealed five chromosomal regions that contained genes for EI subscales. Specifically, markers for Restraint were detected on chromosomes 3 (LOD score = 2.5) and 6 (LOD score = 2.3);

markers for Disinhibition were detected on chromosomes 7 (LOD score = 1.6) and 16 (LOD score = 1.4); and a marker for Hunger was detected on chromosome 3 (LOD score = 1.4). Specific genes on these regions were not identified. Thus, overall, it was dietary disinhibition which was identified as the behavior that had the highest heritability estimate (0.40 ± 0.10) and showed the strongest association with obesity phenotypes.

In a subsequent linkage analysis of the EI conducted in 660 adults from the Quebec Family study, fine-mapping strategies identified the Neuromedian (NMB) gene on chromosome 15 as a possible gene contributing to eating behaviors and obesity [151]. A specific polymorphism of NMB gene (i.e., the p.P73T polymorphism) was associated with scores on the Disinhibition and Hunger subscales. Specifically, individuals with the T/T genotype had higher Disinhibition and Hunger scores than individuals with the P/T or P/P genotypes. Moreover, 6-year weight gain was significantly greater among individuals with the T/T genotype compared to the other genotypes. This study implicates a specific gene that appears to promote increased food intake and weight gain. Specific social-environmental factors were not identified in this report but remain an avenue for future research.

The conceptual framework for this paper (Figure C-1) does not explicitly posit interactions among social-environmental and genetic factors due to the paucity of research examining gene-environment interactions in the human obesity literature. This section reviews the handful of studies that have addressed this issue, which provides a potential framework for future investigations (see Section 7). The section is divided into subsections examining (a) the potential moderating effects of the social environment on the relationship between genetics and obesity, and (b) the potential moderating effects of genetic factors on the relationship between the social environment and obesity. The term “moderator” is used as is conventional for multiple regression analysis [152]. Specifically, a variable X is considered a “moderator” when the relationship between two other variables, Y and Z, depends on the level of X. Figures C-2 and C-3 depict a modified version of Figure C-1, allowing for interactions among the social environment and genetic factors. The social environment serves as the moderator in Figure C-2, while genetic factors serve as the moderator for Figure C-3.

FIGURE C-3 Gene-environment interaction model, with social environment as moderator. In this model, the effects of genetic risk/specific genes on obesity and obesity-promoting behaviors depend on the level of exposure to a given social-environmental factor.

Ravussin et al. [153] compared the weight status and diabetes-related comorbidities of Pima Indians living in remote rural regions of Mexico compared to those living in Arizona. The Pima Indians of Arizona have been extensively studied given their markedly high prevalences of obesity and type 2 diabetes. They are considered to be at genetically increased risk for these disorders. Compared to Pima Indians living in rural Mexico, those living in Arizona weighed significantly more (64.2 vs. 90.2 kg), had higher BMIs (24.9 vs. 33.4 kg/m²), and had higher total cholesterol levels (146 vs. 174mg/dl). Among the Pima Indians from Mexico, 11 percent of the women and 6 percent of the men had type 2 diabetes; by contrast, among the Pima Indians living in Arizona, 37 percent of the women and 54 percent of the men had the diagnoses.

Bhatnagar et al. [154] compared 247 London residents who had migrated from the Indian subcontinent of Punjabi against 117 of their siblings who still lived in India. Compared to the siblings in India, those living in London had significantly higher BMI values, systolic blood pressure, serum cholesterol, apolipoprotein B, and fasting blood glucose.

These results collectively suggest that genetic influences on the development of obesity can be mitigated by environmental conditions. However, these data do not necessarily provide support for the presence of gene-by-environment interactions within the U.S. population at a single time period. As noted in Section 7, this represents an avenue for additional research, especially when considering the potential moderating effects of SES and environmental factors associated with lower income.

Bouchard and colleagues conducted a seminal “overfeeding” study in which 12 male MZ twin pairs were fed an additional 1,000 kcal/day beyond their baseline intake levels, for 6 days per week over 100 days [155]. The investigators tested whether changes in body composition in response to overfeeding differed as a function of twinship. Outcome measures included changes in body composition and metabolic parameters. Results provided clear evidence that response to overfeeding was related to twinship. Twins were significantly correlated with respect to changes in body weight, percent fat mass, fat mass, and estimated subcutaneous fat, and visceral adiposity. Table C-3 presents changes in study outcome measures associated with experimental overfeeding and the intraclass correlation coefficients representing the within-twin pair association for change scores.

In recent years, a series of candidate gene analyses evaluated whether specific genes were associated with response to overfeeding in this cohort. As reviewed by Ukkola and Bouchard [156], a number of candidate genes showed associations. For example, a polymorphism in the adipsin gene was associated with greater increases in body weight, total fat mass, and subcutaneous fat in response to overfeeding; the Gln27Glu polymorphism of the beta 2 adrenergic receptor gene was associated with greater gains in body weight and subcutaneous fat. Few associations were found for changes in visceral adiposity. Despite the limited sample size, these analyses have been critical to the field for demonstrating how specific genes might moderate the effects of a specific environmental manipulation that promotes weight gain (i.e., overfeeding).

Finally, Epstein et al. [157] recently reported that the association between the “reinforcing value of food” phenotype (see Section 5) and ad libitum energy intake in the laboratory was moderated by the dopamine transporter gene (SLC6A3) and the dopamine 2 receptor gene (DRD2). Participants were 88 smokers of European American ancestry who were evaluated before beginning a smoking cessation treatment. With respect to the SLC6A3 gene, subjects who scored high on the reinforcing value of food and who lacked the SLC6A3*9 allele consumed more total energy than participants with other SLC6A3 genotypes. With respect to the DRD2

gene, subjects who scored high on reinforcing-value-of-food and who had the A1 allele consumed more total calories compared to participants with any other DRD2 genotype. This study is unique in that it focused on the genetics of food reward as they relate to dopamine pathways, an area of research that has been understudied and may be promising for future research.

In sum, there have been very few studies of gene-environment interac-

tion as they relate to obesity and obesity-related behaviors. As noted in the next and final section, this represents an avenue for additional research.

Building upon literature reviewed in this report, this final section reviews opportunities for additional research that would bridge two active, but so far separate, research areas: specifically, genetic and social-environmental influences on obesity. This list is not exhaustive and the ideas are not necessarily presented in order of importance. The overarching recommendation is that current knowledge on the causes of obesity may benefit from future research that explicitly tests the interactions between, or covariations among, genetic and social-environmental factors that promote obesity. This would require greater collaborations among social and behavioral scientists, physiologists, and molecular geneticists, with each discipline bringing its unique perspectives and methodological tools to a joint research effort. The potential benefits of integrative research need to be weighed against the potential drawbacks, including greater recruitment challenges, increased costs, and issues concerning adequate statistical power.

New insights into the joint influences of genetics and the social-environmental on obesity and obesity-promoting behaviors may be generated by:

• Additional prospective studies that test genetic and environmental influences on obesity development during putative “critical growth periods.” Most genetic studies reviewed in Section 5 used cross-sectional designs to test for genetic effects at a single time point, whether it was a general heritability estimate or tests of a specific gene or genomic region. However, the onset of obesity is a developmental process that may be influenced by different genetic or environmental influences at different ages. Thus, research that uses prospective designs to identify genetic and environmental influences on the developmental trajectories of body fat stores would be informative. This would be especially useful for studying putative “critical growth periods” for obesity: growth in the intra-uterine environment, “adiposity rebound” in early childhood, and adolescence [158]. The extent to which body composition changes during these periods is influenced by life experiences specific to those periods, or age-specific genetic influences, warrants additional research. If there is evidence for genetic influences at specific ages, the role of individual genes needs to be elucidated. Very few studies have used twin designs to test critical growth periods for obesity

onset. Indeed, the results of one such twin study suggested that the intra-uterine environment is a critical period for the development of adult height, but not for adult BMI [159].

Likewise, new research, so far only in the animal model, is emerging that suggests that maternal obesity during pre- and postnatal periods can have profound, genotype-specific effects on the development of obesity in offspring that is genetically predisposed to obesity (B. Levin, 2005: Oral presentation at the Society for the Study of Ingestive Behavior).

The first year of life may be an especially interesting period to study with respect to longer term obesity. Rapid weight gain during the first 4 months of life is a risk factor for obesity in childhood and adulthood. In one study of 300 full-term African American infants, rapid weight gain was defined as an increase in weight-for-age ≥ 1 SD between birth and 4 months [160]. After adjusting for confounding factors, infants who had experienced rapid weight gain by 4 months of age were 5.22 times more likely to be obese at 20 years of age compared to infants who did not experience rapid weight gain. In a separate analysis of 19,397 infants, results indicated that both birth weight and rate of weight gain were associated with an increased probability of childhood overweight at 7 years of age [161]; within each strata of birth weight, increased rate of weight gain was associated with increased childhood overweight prevalence. Potential genetic and home environmental influences on early life rate of weight gain are poorly understood and may be an important area for future research.

• Additional studies that evaluate the heritability of, or specific genes associated with, refined behavioral phenotypes related to obesity. Very little is known about the heritability of behavioral traits that are associated with obesity, particularly those reviewed in Section 4. Studies that clarify the genetic-environmental architecture of these traits would elucidate the extent to which those behavioral traits are genetically influenced, as well as the nature of environmental influences that influence those behaviors (i.e., shared vs. nonshared environmental effects). Such designs could also address important multivariate questions, including the extent to which the correlations between behaviors and body fat is influenced by the same genes (i.e., “genetic correlations”) or the same environmental factors (i.e., “environmental correlations”). Especially interesting would be heritability studies of laboratory-based behavioral traits, such as the reinforcing value of food [90, 157], delayed satiation [78], disinhibition [64], or eating in the absence of hunger [48, 162], which have been linked to obesity status.

One of the difficulties in identifying obese phenotypes and associated eating behaviors lies in the existence of several subpopulations of over-

weight and obese populations. It is likely that individuals who are gaining or losing weight (i.e., reduced obese) exhibit different eating patterns and intake behaviors than do those individuals who are obese but weight-stable.

Behavioral measures evaluated during infancy would be uniquely informative because sucking rate at 3 months of age predicts subsequent weight gain during the first 2 years of life [95, 96]. It is possible that sucking behavior is genetically influenced, because infants born at high risk for obesity have been shown to suck at greater rates when studied in the laboratory compared to infants born at low risk for obesity [95]. The heritability of infant sucking rate is unknown. On the other hand, there is considerable evidence that infants learn flavor preferences during the first year of life through environmental exposure to specific foods [163-168], as well as data that restrictive feeding patterns during infancy are associated with excess infant weight gain [169]. Thus, the roles of learning and genetics, early life sucking, appetite, and food intake needs greater attention. In addition, the identification of genes for NEAT and other refined physical activity traits would advance the field.

In summary, the obese phenotype is likely to be characterized by a conglomerate of significant behaviors related to eating and physical activity which likely work in conjunction to affect energy balance. One of the goals could be to develop (a set of) tools that capture “obese” eating behaviors and physical activity behaviors in an unobtrusive way, if possible at an early age, to make predictions of an individual’s weight development.

• Additional research that incorporates specific measures of the environment into genetics studies. Most genetic studies have not measured specific aspects of the environment. This includes aspects of the home environment, as well as components of the broader “macroenvironment” discussed in Section 3. Genetics studies provide clear evidence that obesity is influenced by the environment, with most studies suggesting that the non-shared environment is more influential. However, the identity of specific environmental influences has remained elusive, especially during child development. Adding specific measures of the environment might help address these issues.

It is noteworthy that valid measures of the home environment exist and, in principle, could be incorporated into genetic studies. One of the most extensively used instruments in the child development literature is the “Home Observation for Measurement of the Environment” (HOME) system [170]. Different versions of the HOME have been developed for different ages, specifically, infancy and toddlerhood, preschool and early childhood, school-age and middle childhood, and adolescence. The HOME has been used in at least one study of childhood obesity, finding that reduced

levels of “cognitive stimulation” at home prospectively predicted increased obesity incidence [171].

In addition to measures of the home environment, measures of the broader environment would be informative for genetics research. In principle, genetic influences on food intake or physical activity may depend on the access to parks, playground, grocery shops, fast-food restaurants, or other environmental variables associated with SES strata. Recent studies have used Geographic Information Systems to “geocode” the physical distance between individual homes and these other components to the community [14]; however, there appear to be no studies to date that have used this technology in the context of genetics of obesity. A handful of studies in the child development literature used this approach to understand the interaction between genes and the broader social environment [172, 173]; these provide useful examples for obesity researchers. For example, in a study of 1,081 MZ twin pairs and 1,061 DZ twin pairs, Caspi et al. [172] found that 20 percent of the variability in 2-year-old children’s “behaviors problems” were influenced by shared environmental factors. When a specific measure of “environmental deprivation” was added to the biometric model, however, it was found to account for 5 percent of the variance in the shared environment. Thus, geocoding and related tools that permit better measurement of the macroenvironment, or exposure to the “toxic environment,” may advance the field of genetics research.

• Additional observational and experimental research that evaluates gene-environment interactions. As noted in Section 6, there are very few studies of gene-environment interaction in the literature. This could be an important area for research with respect to macroenvironmental variables, such as SES, ethnicity, and exposure to the “toxic environment.” Thus, the effects of certain obesity-promoting genes may depend on the broader social environment in which a population lives; this is an avenue for additional research and is depicted in Figure C-4.

In addition, experimental studies that test for gene-environment interactions, in similar ways to the Quebec Overfeeding Study [155, 156], would be most informative. In principle, aspects of the “toxic environment” can be experimentally manipulated in a controlled feeding laboratory or metabolic ward, in a manner that cannot be done in the free-living environment. Examples include experimental manipulations of food portion size [74, 174], energy density [75], food deprivation status [175], and food variety [176]. These rigorous laboratory protocols, if used with genetics designs, could yield novel information regarding gene-environment interactions. Potential designs include: co-twin control designs, in which MZ twins are randomly assigned to different experimental conditions; classic twins de-

FIGURE C-4 Gene-environment interaction model, with genotype as moderator. In this model, the effects of social-environmental influences on obesity and obesity-promoting behaviors depend on genotype.

signs, in which MZ and DZ twins are used to estimate the heritability of response to an experimental manipulation; or candidate gene designs, in which participants are selected based on specific genotypes. In all cases, pertinent outcome variables could be behavioral and/or physiological measures, as well as changes in body weight if the manipulation is sustained over time.

• Additional research that evaluates gene-environment correlations. Genetic studies of obesity most commonly used BMI or body fat as the primary phenotype, followed by metabolic and physiological measures, and, least commonly, behavioral measures. However, in principle, obesity-promoting genes may operate by influencing the environments into which individuals place themselves. Such a scenario is depicted in Figure C-5. That is, social-environmental measures might be conceptualized as the phenotype in a genetics study, especially if genes influence whether certain individuals will seek out “obesity-promoting” environments (e.g., fast-food restaurants). As noted in Section 3, there is evidence that obese individuals may be more likely to attend restaurants than nonobese individuals on the days that buffets are served, which would be suggestive of a gene-environment correlation. Plomin et al. [177] provide a more detailed dis-

cussion of such “active” gene-environment correlations, in which genes influence people’s tendencies to create their own environments.

The issue of gene-environment correlations is also relevant to the domain of child development and, in recent years, there has been increasing interest in the “genetics of parenting” [178-180]. Data suggest that certain parenting behaviors towards children are, in fact, elicited by child attributes and behavioral patterns that are probably genetically influenced. This may be a useful framework for studying parent-child feeding dynamics as they relate to obesity onset. As noted Section 3, there is evidence that parental restriction of child eating is elicited by child weight characteristics [43] and this in turn may exacerbate further weight gain by the child. Additional genetics studies could evaluate whether parental feeding restriction, or other parenting domains, are associated with specific candidate genes for obesity.

• Additional research that builds upon existing conceptual models for “organism-environment interactions.” Conceptual models that explicitly address the integration of genetic and social-environmental influences on behavioral traits may help guide future studies. The field of developmental behavioral genetics has addressed this issue, although not in regards to obesity per se. Several pertinent books have been published [181-186]. In addition, several longitudinal behavioral genetics studies measured specific

FIGURE C-5 Gene-environment correlation model. In this model, there is a correlation among genes and social-environmental factors that influence obesity and obesity-promoting behaviors.

aspects of the social environment and genetic factors and may provide useful models for obesity research. De Castro [187-189] has one of the few proposed models that integrates genetic and environmental influences on food intake.

• Additional institutional and/or funding mechanisms to support integrative research projects or interdisciplinary training for scientists. Interdisciplinary research of the sort reviewed in this report would likely require new collaborative relationships that bring together investigators from different “camps.” Institutional and/or funding initiatives that encourage such collaborations may help advance such efforts, given the economic and logistical challenges of such research. Initial collaboration of this sort could be exemplars for other institutions and investigators.

This report set out to highlight two distinct areas of research that share the common goal of identifying factors that contribute to weight gain and obesity in the population. The areas reviewed in this report included research on (social-) environmental factors, as well as the genetic factors, that may be associated with obesity or the onset thereof. Despite their unique focuses, the literature reviewed in this report shows that the two areas have the potential to complement each other and to stimulate future collaborations among investigators. The pathways that lead to obesity are complex and multivariate for most individuals in the population. Additional research that addresses how the genetics of obesity impacts on environmental choices made by certain individuals, and how certain environments moderate the expression of obesity-promoting genes, may advance the current state of knowledge and provide new insights for the prevention and the treatment of obesity.

1. Pi-Sunyer, F.X., Health implications of obesity. Am J Clin Nutr, 1991. 53(6 Suppl): p. 1595S-1603S.

2. Zhu, S., et al., Race-ethnicity-specific waist circumference cutoffs for identifying cardiovascular disease risk factors. Am J Clin Nutr, 2005. 81(2): p. 409-15.

3. Zhu, S., et al., Combination of BMI and waist circumference for identifying cardiovascular risk factors in whites. Obes Res, 2004. 12(4): p. 633-45.

4. Park, Y.W., et al., The metabolic syndrome: prevalence and associated risk factor findings in the US population from the Third National Health and Nutrition Examination Survey, 1988-1994. Arch Intern Med, 2003. 163(4): p. 427-36.

5. Dong, C., et al., Interacting genetic loci on chromosomes 20 and 10 influence extreme human obesity. Am J Hum Genet, 2003. 72(1): p. 115-24.

6. Barnett, E. and M. Casper, A definition of “social environment”. Am J Public Health, 2001. 91(3): p. 465.

7. Battle, E.K. and K.D. Brownell, Confronting a rising tide of eating disorders and obesity: treatment vs. prevention and policy. Addict Behav, 1996. 21(6): p. 755-65.

8. Jacobson, M.F. and K.D. Brownell, Small taxes on soft drinks and snack foods to promote health. Am J Public Health, 2000. 90(6): p. 854-7.

9. French, S.A., et al., Fast food restaurant use among adolescents: associations with nutrient intake, food choices and behavioral and psychosocial variables. Int J Obes Relat Metab Disord, 2001. 25(12): p. 1823-33.

10. French, S.A., M. Story, and R.W. Jeffery, Environmental influences on eating and physical activity. Annu Rev Public Health, 2001. 22: p. 309-35.

11. French, S.A., et al., Pricing and promotion effects on low-fat vending snack purchases: the CHIPS Study. Am J Public Health, 2001. 91(1): p. 112-7.

12. Crawford, D.A., R.W. Jeffery, and S.A. French, Television viewing, physical inactivity and obesity. Int J Obes Relat Metab Disord, 1999. 23(4): p. 437-40.

13. Jeffery, R.W. and J. Utter, The changing environment and population obesity in the United States. Obes Res, 2003. 11 Suppl: p. 12S-22S.

14. Burdette, H.L. and R.C. Whitaker, Neighborhood playgrounds, fast food restaurants, and crime: relationships to overweight in low-income preschool children. Prev Med, 2004. 38(1): p. 57-63.

15. Jeffery, R.W., et al., The relationship between social status and body mass index in the Minnesota Heart Health Program. Int J Obes, 1989. 13(1): p. 59-67.

16. Sundquist, J., M. Malmstrom, and S.E. Johansson, Cardiovascular risk factors and the neighbourhood environment: a multilevel analysis. Int J Epidemiol, 1999. 28(5): p. 841-5.

17. McMurray, R.G., et al., The influence of physical activity, socioeconomic status, and ethnicity on the weight status of adolescents. Obes Res, 2000. 8(2): p. 130-9.

18. Ball, K. and D. Crawford, Socioeconomic status and weight change in adults: a review. Soc Sci Med, 2005. 60(9): p. 1987-2010.

19. Goodman, E., The role of socioeconomic status gradients in explaining differences in U.S. adolescents’ health. Am J Public Health, 1999. 89(10): p. 1522-8.

20. Ellaway, A., A. Anderson, and S. Macintyre, Does area of residence affect body size and shape? Int J Obes Relat Metab Disord, 1997. 21(4): p. 304-8.

21. Boslaugh, S.E., et al., Perceptions of neighborhood environment for physical activity: is it “who you are” or “where you live”? J Urban Health, 2004. 81(4): p. 671-81.

22. Reidpath, D.D., et al., An ecological study of the relationship between social and environmental determinants of obesity. Health Place, 2002. 8(2): p. 141-5.

23. Drewnowski, A. and N. Darmon, Food choices and diet costs: an economic analysis. J Nutr, 2005. 135(4): p. 900-4.

24. Drewnowski, A. and S.E. Specter, Poverty and obesity: the role of energy density and energy costs. Am J Clin Nutr, 2004. 79(1): p. 6-16.

25. Drewnowski, A., Obesity and the food environment: dietary energy density and diet costs. Am J Prev Med, 2004. 27(3 Suppl): p. 154-62.

26. Neumark-Sztainer, D., et al., Correlates of fruit and vegetable intake among adolescents. Findings from Project EAT. Prev Med, 2003. 37(3): p. 198-208.

27. Darmon, N., A. Briend, and A. Drewnowski, Energy-dense diets are associated with lower diet costs: a community study of French adults. Public Health Nutr, 2004. 7(1): p. 21-7.

28. Burdette, H.L. and R.C. Whitaker, A national study of neighborhood safety, outdoor play, television viewing, and obesity in preschool children. Pediatrics, 2005. 116(3): p. 657-62.

29. Herman, C.P., D.A. Roth, and J. Polivy, Effects of the presence of others on food intake: a normative interpretation. Psychol Bull, 2003. 129(6): p. 873-86.

30. de Castro, J.M. and E.S. de Castro, Spontaneous meal patterns of humans: influence of the presence of other people. Am J Clin Nutr, 1989. 50(2): p. 237-47.

31. de Castro, J.M., Family and friends produce greater social facilitation of food intake than other companions. Physiol Behav, 1994. 56(3): p. 445-5.

32. Edelman, B., et al., Environmental effects on the intake of overweight and normal-weight men. Appetite, 1986. 7(1): p. 71-83.

33. Kim, J.Y. and H.R. Kissileff, The effect of social setting on response to a preloading manipulation in non-obese women and men. Appetite, 1996. 27(1): p. 25-40.

34. de Castro, J.M., et al., Social facilitation of the spontaneous meal size of humans occurs regardless of time, place, alcohol or snacks. Appetite, 1990. 15(2): p. 89-101.

35. Kramer, M.S. and B. Moroz, Do breast-feeding and delayed introduction of solid foods protect against subsequent atopic eczema? J Pediatr, 1981. 98(4): p. 546-50.

36. Armstrong, J. and J.J. Reilly, Breastfeeding and lowering the risk of childhood obesity. Lancet, 2002. 359(9322): p. 2003-4.

37. Liese, A.D., et al., Inverse association of overweight and breast feeding in 9 to 10-y-old children in Germany. Int J Obes Relat Metab Disord, 2001. 25(11): p. 1644-50.

38. Gillman, M.W., et al., Risk of overweight among adolescents who were breastfed as infants. JAMA, 2001. 285(19): p. 2461-7.

39. Wadsworth, M., et al., Breast feeding and obesity. Relation may be accounted for by social factors. BMJ, 1999. 319(7224): p. 1576.

40. Grummer-Strawn, L.M. and Z. Mei, Does breastfeeding protect against pediatric over-weight? Analysis of longitudinal data from the Centers for Disease Control and Prevention Pediatric Nutrition Surveillance System. Pediatrics, 2004. 113(2): p. e81-6.

41. Taveras, E.M., et al., Association of breastfeeding with maternal control of infant feeding at age 1 year. Pediatrics, 2004. 114(5): p. e577-83.

42. Johnson, S.L. and L.L. Birch, Parents’ and children’s adiposity and eating style. Pediatrics, 1994. 94(5): p. 653-61.

43. Faith, M.S., et al., Parent-child feeding strategies and their relationships to child eating and weight status. Obes Res, 2004. 12(11): p. 1711-22.

44. Birch, L.L., et al., Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of parental attitudes, beliefs and practices about child feeding and obesity proneness. Appetite, 2001. 36(3): p. 201-10.

45. Wardle, J., et al., Parental feeding style and the inter-generational transmission of obesity risk. Obes Res, 2002. 10(6): p. 453-62.

46. Baughcum, A.E., et al., Maternal feeding practices and beliefs and their relationships to overweight in early childhood. J Dev Behav Pediatr, 2001. 22(6): p. 391-408.

47. Birch, L.L. and J.A. Fisher, Appetite and eating behavior in children. Pediatr Clin North Am, 1995. 42(4): p. 931-53.

48. Fisher, J.O. and L.L. Birch, Eating in the absence of hunger and overweight in girls from 5 to 7 y of age. Am J Clin Nutr, 2002. 76(1): p. 226-31.

49. Faith, M.S., et al., Parental feeding attitudes and styles and child body mass index: prospective analysis of a gene-environment interaction. Pediatrics, 2004. 114(4): p. e429-36.

50. Hill, J.O., et al., Obesity and the environment: where do we go from here? Science, 2003. 299(5608): p. 853-5.

51. Adams, N., et al., The eating behavior of obese and nonobese women. Behav Res Ther, 1978. 16(4): p. 225-32.

52. Dodd, D.K., H.J. Birky, and R.B. Stalling, Eating behavior of obese and normal-weight females in a natural setting. Addict Behav, 1976. 1(4): p. 321-5.

53. Hill, S.W. and N.B. McCutcheon, Eating responses of obese and nonobese humans during dinner meals. Psychosom Med, 1975. 37(5): p. 395-401.

54. Kaplan, D.L., Eating style of obese and nonobese males. Psychosom Med, 1980. 42(6): p. 529-38.

55. Mahoney, M.J., The obese eating style: bites, beliefs, and behavior modification. Addict Behav, 1975. 1(1): p. 47-53.

56. Marston, A.R., et al., In vivo observation of the eating habits of obese and nonobese subjects, in Recent Advances in Obesity Research, A. Howard, Editor. 1975, Newman: London. p. 207-210.

57. Stunkard, A., et al., Obesity and eating style. Arch Gen Psychiatry, 1980. 37(10): p. 1127-9.

58. Schachter, S., Some extraordinary facts about obese humans and rats. Am Psychol, 1971. 26(2): p. 129-44.

59. Schachter, S., R. Goldman, and A. Gordon, Effects of fear, food deprivation, and obesity on eating. J Pers Soc Psychol, 1968. 10(2): p. 91-7.

60. Schachter, S. and L.P. Gross, Manipulated time and eating behavior. J Pers Soc Psychol, 1968. 10(2): p. 98-106.

61. Nisbett, R.E., Taste, deprivation, and weight determinants of eating behavior. J Pers Soc Psychol, 1968. 10(2): p. 107-16.

62. Nisbett, R.E. and M.D. Storms, Cognitive, social, psychological determinants of food intake, in Cognitive modification of emotional behavior, H. London and R.E. Nisbett, Editors. 1975, Aldine: Chicago.

63. Nisbett, R.E. and L. Temoshok, Is there an external cognitive style? J Pers Soc Psychol, 1976. 33: p. 36-47.

64. Lindroos, A.K., et al., Dietary intake in relation to restrained eating, disinhibition, and hunger in obese and nonobese Swedish women. Obes Res, 1997. 5(3): p. 175-82.

65. Carmody, T.P., R.L. Brunner, and S.T. St Jeor, Dietary helplessness and disinhibition in weight cyclers and maintainers. Int J Eat Disord, 1995. 18(3): p. 247-56.

66. Westenhoefer, J., A.J. Stunkard, and V. Pudel, Validation of the flexible and rigid control dimensions of dietary restraint. Int J Eat Disord, 1999. 26(1): p. 53-64.

67. Lawson, O.J., et al., The association of body weight, dietary intake, and energy expenditure with dietary restraint and disinhibition. Obes Res, 1995. 3(2): p. 153-61.

68. McGuire, M.T., et al., What predicts weight regain in a group of successful weight losers? J Consult Clin Psychol, 1999. 67(2): p. 177-85.

69. Williamson, D.A., et al., Association of body mass with dietary restraint and disinhibition. Appetite, 1995. 25(1): p. 31-41.

70. Howard, C.E. and L.K. Porzelius, The role of dieting in binge eating disorder: etiology and treatment implications. Clin Psychol Rev, 1999. 19(1): p. 25-44.

71. Provencher, V., et al., Eating behaviors and indexes of body composition in men and women from the Quebec family study. Obes Res, 2003. 11(6): p. 783-92.

72. Bell, E.A. and B.J. Rolls, Energy density of foods affects energy intake across multiple levels of fat content in lean and obese women. Am J Clin Nutr, 2001. 73(6): p. 1010-8.

73. Rolls, B.J., E.L. Morris, and L.S. Roe, Portion size of food affects energy intake in normal-weight and overweight men and women. Am J Clin Nutr, 2002. 76(6): p. 1207-13.

74. Diliberti, N., et al., Increased portion size leads to increased energy intake in a restaurant meal. Obes Res, 2004. 12(3): p. 562-8.

75. Kral, T.V., L.S. Roe, and B.J. Rolls, Combined effects of energy density and portion size on energy intake in women. Am J Clin Nutr, 2004. 79(6): p. 962-8.

76. McConahy, K.L., et al., Food portions are positively related to energy intake and body weight in early childhood. J Pediatr, 2002. 140(3): p. 340-7.

77. Heitmann, B.L. and L. Lissner, Dietary underreporting by obese individuals—is it specific or non-specific? BMJ, 1995. 311(7011): p. 986-9.

78. Westerterp-Plantenga, M.S., et al., Energy intake adaptation of food intake to extreme energy densities of food by obese and non-obese women. Eur J Clin Nutr, 1996. 50(6): p. 401-7.

79. Meyer, J.E. and V. Pudel, Experimental studies on food-intake in obese and normal weight subjects. J Psychosom Res, 1972. 16(4): p. 305-8.

80. Westerterp-Plantenga, M.S., L. Wouters, and F. ten Hoor, Restrained eating, obesity, and cumulative food intake curves during four-course meals. Appetite, 1991. 16(2): p. 149-58.

81. Matsuda, M., et al., Altered hypothalamic function in response to glucose ingestion in obese humans. Diabetes, 1999. 48(9): p. 1801-6.

82. Del Parigi, A., et al., Neuroimaging and obesity: mapping the brain responses to hunger and satiation in humans using positron emission tomography. Ann N Y Acad Sci, 2002. 967: p. 389-97.

83. Kissileff, H.R., Satiating efficiency and a strategy for conducting food loading experiments. Neurosci Biobehav Rev, 1984. 8(1): p. 129-35.

84. Rolls, B.J., et al., Satiety after preloads with different amounts of fat and carbohydrate: implications for obesity. Am J Clin Nutr, 1994. 60(4): p. 476-87.

85. Fricker, J., et al., Effect of a covert fat dilution on the spontaneous food intake by lean and obese subjects. Appetite, 1995. 24(2): p. 121-37.

86. Roe, L.S., et al., A meta-analysis of factors predicting energy compensation in preloading studies. FASEB Journal, 1999. 13: p. A871.

87. Faith, M.S., et al., Familial aggregation of energy intake in children. Am J Clin Nutr, 2004. 79(5): p. 844-50.

88. Epstein, L.H. and B.E. Saelens, Behavioral economics of obesity: food intake and energy expenditure, in Reframing Health Behavior Change with Behavioral Economics, W.K. Bickel and R.E. Vuchinich, Editors. 2000, Lawrence Erlbaum: Mahwah, NJ. p. 293-311.

89. Epstein, L.H.S., B.E, Behavioral economics of obesity: Food intake and energy expenditure, in Reframing health behavior change with behavioral economics, W.K.V. Bickel, R.E, Editor. 2000, Lawrence Erlbaum: Mahwah, NJ. p. 293-311.

90. Saelens, B.E. and L.H. Epstein, Reinforcing value of food in obese and non-obese women. Appetite, 1996. 27(1): p. 41-50.

91. Lappalainen, R. and L.H. Epstein, A behavioral economics analysis of food choice in humans. Appetite, 1990. 14(2): p. 81-93.

92. Ferster, C.B., J.I. Nurnberger, and E.B. Levitt, The control of eating. J Mathematics, 1962. 1: p. 87-109.

93. Spiegel, T.A., E.E. Shrager, and E. Stellar, Responses of lean and obese subjects to preloads, deprivation, and palatability. Appetite, 1989. 13(1): p. 45-69.

94. Kissileff, H.R. and J. Thornton, Facilitation and inhibition in the cumulative food intake curve in man, in Changing Concepts of the Nervous System, A.J. Morrison and P. Stick, Editors. 1982, Academic Press: New York. p. 585-605.

95. Stunkard, A.J., et al., Energy intake, not energy output, is a determinant of body size in infants. Am J Clin Nutr, 1999. 69(3): p. 524-30.

96. Stunkard, A.J., et al., Predictors of body size in the first 2 y of life: a high-risk study of human obesity. Int J Obes Relat Metab Disord, 2004. 28(4): p. 503-13.

97. Spiegel, T.A., et al., Bite size, ingestion rate, and meal size in lean and obese women. Appetite, 1993. 21(2): p. 131-45.

98. Spiegel, T.A., Rate of intake, bites, and chews—the interpretation of lean-obese differences. Neurosci Biobehav Rev, 2000. 24(2): p. 229-37.

99. Drewnowski, A., Food perceptions and preferences of obese adults: a multidimensional approach. Int J Obes, 1985. 9(3): p. 201-12.

100. Mela, D.J. and D.A. Sacchetti, Sensory preferences for fats: relationships with diet and body composition. Am J Clin Nutr, 1991. 53(4): p. 908-15.

101. Cox, D.N., et al., Sensory and hedonic judgments of common foods by lean consumers and consumers with obesity. Obes Res, 1998. 6(6): p. 438-47.

102. Cox, D.N., et al., Sensory and hedonic associations with macronutrient and energy intakes of lean and obese consumers. Int J Obes Relat Metab Disord, 1999. 23(4): p. 403-10.

103. Lissner, L., et al., Dietary fat and the regulation of energy intake in human subjects. Am J Clin Nutr, 1987. 46(6): p. 886-92.

104. Tucker, L.A. and M.J. Kano, Dietary fat and body fat: a multivariate study of 205 adult females. Am J Clin Nutr, 1992. 56(4): p. 616-22.

105. Bray, G.A., S. Paeratakul, and B.M. Popkin, Dietary fat and obesity: a review of animal, clinical and epidemiological studies. Physiol Behav, 2004. 83(4): p. 549-55.

106. Maskarinec, G., R. Novotny, and K. Tasaki, Dietary patterns are associated with body mass index in multiethnic women. J Nutr, 2000. 130(12): p. 3068-72.

107. Mela, D.J., Determinants of food choice: relationships with obesity and weight control. Obes Res, 2001. 9 Suppl 4: p. 249S-255S.

108. Pereira, M.A., et al., Fast-food habits, weight gain, and insulin resistance (the CARDIA study): 15-year prospective analysis. Lancet, 2005. 365(9453): p. 36-42.

109. Prentice, A.M. and S.A. Jebb, Fast foods, energy density and obesity: a possible mechanistic link. Obes Rev, 2003. 4(4): p. 187-94.

110. Parsons, T.J., C. Power, and O. Manor, Physical activity, television viewing and body mass index: a cross-sectional analysis from childhood to adulthood in the 1958 British cohort. Int J Obes (Lond), 2005. 29(10): p. 1212-21.

111. Jakicic, J.M. and A.D. Otto, Physical activity considerations for the treatment and prevention of obesity. Am J Clin Nutr, 2005. 82(1 Suppl): p. 226S-229S.

112. Wareham, N.J., E.M. van Sluijs, and U. Ekelund, Physical activity and obesity prevention: a review of the current evidence. Proc Nutr Soc, 2005. 64(2): p. 229-47.

113. Wing, R.R. and S. Phelan, Long-term weight loss maintenance. Am J Clin Nutr, 2005. 82(1 Suppl): p. 222S-225S.

114. Francis, L.A., Y. Lee, and L.L. Birch, Parental weight status and girls’ television viewing, snacking, and body mass indexes. Obes Res, 2003. 11(1): p. 143-51.

115. Jago, R., et al., BMI from 3-6 y of age is predicted by TV viewing and physical activity, not diet. Int J Obes Relat Metab Disord, 2005. 29(6): p. 557-64.

116. Henderson, V.R. and B. Kelly, Food advertising in the age of obesity: content analysis of food advertising on general market and African American television. J Nutr Educ Behav, 2005. 37(4): p. 191-6.

117. Levine, J.A., et al., Interindividual variation in posture allocation: possible role in human obesity. Science, 2005. 307(5709): p. 584-6.

118. Maes, H.H., M.C. Neale, and L.J. Eaves, Genetic and environmental factors in relative body weight and human adiposity. Behav Genet, 1997. 27(4): p. 325-51.

119. Stunkard, A.J., et al., The body-mass index of twins who have been reared apart. N Engl J Med, 1990. 322(21): p. 1483-7.

120. Allison, D.B., et al., The heritability of body mass index among an international sample of monozygotic twins reared apart. Int J Obes Relat Metab Disord, 1996. 20(6): p. 501-6.

121. Hewitt, J.K., The genetics of obesity: what have genetic studies told us about the environment. Behav Genet, 1997. 27(4): p. 353-8.

122. Faith, M.S., S.L. Johnson, and D.B. Allison, Putting the behavior into the behavior genetics of obesity. Behav Genet, 1997. 27(4): p. 423-39.

123. Koeppen-Schomerus, G., J. Wardle, and R. Plomin, A genetic analysis of weight and overweight in 4-year-old twin pairs. Int J Obes Relat Metab Disord, 2001. 25(6): p. 838-44.

124. Jacobson, K.C. and D.C. Rowe, Genetic and shared environmental influences on adolescent BMI: interactions with race and sex. Behav Genet, 1998. 28(4): p. 265-78.

125. Perusse, L., et al., The human obesity gene map: the 2004 update. Obes Res, 2005. 13(3): p. 381-490.

126. Redden, D.T. and D.B. Allison, Nonreplication in genetic association studies of obesity and diabetes research. J Nutr, 2003. 133(11): p. 3323-6.

127. Loos, R.J. and C. Bouchard, Obesity—is it a genetic disorder? J Intern Med, 2003. 254(5): p. 401-25.

128. Korner, J. and R.L. Leibel, To eat or not to eat—how the gut talks to the brain. N Engl J Med, 2003. 349(10): p. 926-8.

129. Rosenbaum, M. and R.L. Leibel, The role of leptin in human physiology. N Engl J Med, 1999. 341(12): p. 913-5.

130. Badman, M.K. and J.S. Flier, The gut and energy balance: visceral allies in the obesity wars. Science, 2005. 307(5717): p. 1909-14.

131. De Castro, J.M., The effects of the spontaneous ingestion of particular foods or beverages on the meal pattern and overall nutrient intake of humans. Physiol Behav, 1993. 53(6): p. 1133-44.

132. de Castro, J.M., Genetic influences on daily intake and meal patterns of humans. Physiol Behav, 1993. 53(4): p. 777-82.

133. de Castro, J.M., Independence of genetic influences on body size, daily intake, and meal patterns of humans. Physiol Behav, 1993. 54(4): p. 633-9.

134. de Castro, J.M., A twin study of genetic and environmental influences on the intake of fluids and beverages. Physiol Behav, 1993. 54(4): p. 677-87.

135. de Castro, J.M., Stomach filling may mediate the influence of dietary energy density on the food intake of free-living humans. Physiol Behav, 2005.

136. Pearcey, S.M. and J.M. de Castro, Food intake and meal patterns of weight-stable and weight-gaining persons. Am J Clin Nutr, 2002. 76(1): p. 107-12.

137. de Castro, J.M., Genes and environment have gender-independent influences on the eating and drinking of free-living humans. Physiol Behav, 1998. 63(3): p. 385-95.

138. De Castro, J.M., Age-related changes in spontaneous food intake and hunger in humans. Appetite, 1993. 21(3): p. 255-72.

139. De Castro, J.M., Social facilitation of duration and size but not rate of the spontaneous meal intake of humans. Physiol Behav, 1990. 47(6): p. 1129-35.

140. van den Bree, M.B., L.J. Eaves, and J.T. Dwyer, Genetic and environmental influences on eating patterns of twins aged >/=50 y. Am J Clin Nutr, 1999. 70(4): p. 456-65.

141. Keski-Rahkonen, A.V., Genetic and environmental factors in breakfast eating patterns. Behav Genet, 2004. 34: p. 503-514.

142. Lichtman, S.W., et al., Discrepancy between self-reported and actual caloric intake and exercise in obese subjects. N Engl J Med, 1992. 327(27): p. 1893-8.

143. Faith, M.S., et al., Evidence for genetic influences on human energy intake: results from a twin study using measured observations. Behav Genet, 1999. 29(3): p. 145-54.

144. Faith, M.S. and K.L. Keller, Genetic architecture of ingestive behavior in humans. Nutrition, 2004. 20(1): p. 127-33.

145. Aubert, R., et al., 5-HT2A receptor gene polymorphism is associated with food and alcohol intake in obese people. Int J Obes Relat Metab Disord, 2000. 24(7): p. 920-4.

146. Herbeth, B., et al., Polymorphism of the 5-HT2A receptor gene and food intakes in children and adolescents: the Stanislas Family Study. Am J Clin Nutr, 2005. 82(2): p. 467-70.

147. Farooqi, I.S., G.S. Yeo, and S. O’Rahilly, Binge eating as a phenotype of melanocortin 4 receptor gene mutations. N Engl J Med, 2003. 349(6): p. 606-9; author reply 606-9.

148. Branson, R., et al., Binge eating as a major phenotype of melanocortin 4 receptor gene mutations. N Engl J Med, 2003. 348(12): p. 1096-103.

149. Stunkard, A.J. and S. Messick, The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. J Psychosom Res, 1985. 29(1): p. 71-83.

150. Steinle, N.I., et al., Eating behavior in the Old Order Amish: heritability analysis and a genome-wide linkage analysis. Am J Clin Nutr, 2002. 75(6): p. 1098-106.

151. Bouchard, L., et al., Neuromedin beta: a strong candidate gene linking eating behaviors and susceptibility to obesity. Am J Clin Nutr, 2004. 80(6): p. 1478-86.

152. Kraemer, H.C., et al., How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. Am J Psychiatry, 2001. 158(6): p. 848-56.

153. Ravussin, E., et al., Effects of a traditional lifestyle on obesity in Pima Indians. Diabetes Care, 1994. 17(9): p. 1067-74.

154. Bhatnagar, D., et al., Coronary risk factors in people from the Indian subcontinent living in west London and their siblings in India. Lancet, 1995. 345(8947): p. 405-9.

155. Bouchard, C., et al., The response to long-term overfeeding in identical twins. N Engl J Med, 1990. 322(21): p. 1477-82.

156. Ukkola, O. and C. Bouchard, Role of candidate genes in the responses to long-term overfeeding: review of findings. Obes Rev, 2004. 5(1): p. 3-12.

157. Epstein, L.H., et al., Relation between food reinforcement and dopamine genotypes and its effect on food intake in smokers. Am J Clin Nutr, 2004. 80(1): p. 82-8.

158. Dietz, W.H., Critical periods in childhood for the development of obesity. Am J Clin Nutr, 1994. 59(5): p. 955-9.

159. Allison, D.B., et al., Is the intra-uterine period really a critical period for the development of adiposity? Int J Obes Relat Metab Disord, 1995. 19(6): p. 397-402.

160. Stettler, N., et al., Rapid weight gain during infancy and obesity in young adulthood in a cohort of African Americans. Am J Clin Nutr, 2003. 77(6): p. 1374-8.

161. Stettler, N., et al., Infant weight gain and childhood overweight status in a multicenter, cohort study. Pediatrics, 2002. 109(2): p. 194-9.

162. Birch, L.L., J.O. Fisher, and K.K. Davison, Learning to overeat: maternal use of restrictive feeding practices promotes girls’ eating in the absence of hunger. Am J Clin Nutr, 2003. 78(2): p. 215-20.

163. Mennella, J.A., M.Y. Pepino, and D.R. Reed, Genetic and environmental determinants of bitter perception and sweet preferences. Pediatrics, 2005. 115(2): p. e216-22.

164. Mennella, J.A., C.E. Griffin, and G.K. Beauchamp, Flavor programming during infancy. Pediatrics, 2004. 113(4): p. 840-5.

165. Mennella, J.A., M.Y. Pepino, and G.K. Beauchamp, Modification of bitter taste in children. Dev Psychobiol, 2003. 43(2): p. 120-7.

166. Gerrish, C.J. and J.A. Mennella, Flavor variety enhances food acceptance in formula-fed infants. Am J Clin Nutr, 2001. 73(6): p. 1080-5.

167. Mennella, J.A. and G.K. Beauchamp, Experience with a flavor in mother’s milk modifies the infant’s acceptance of flavored cereal. Dev Psychobiol, 1999. 35(3): p. 197-203.

168. Mennella, J.A. and G.K. Beauchamp, Early flavor experiences: research update. Nutr Rev, 1998. 56(7): p. 205-11.

169. Fisher, J.O., et al., Breast-feeding through the first year predicts maternal control in feeding and subsequent toddler energy intakes. J Am Diet Assoc, 2000. 100(6): p. 641-6.

170. Caldwell, B.M. and R.H. Bradley, Home observation for measurement of the environment. 1984, Little Rock: University of Arkansas.

171. Strauss, R.S. and J. Knight, Influence of the home environment on the development of obesity in children. Pediatrics, 1999. 103(6): p. e85.

172. Caspi, A., et al., Neighborhood deprivation affects children’s mental health: environmental risks identified in a genetic design. Psychol Sci, 2000. 11(4): p. 338-42.

173. Kim-Cohen, J., et al., Genetic and environmental processes in young children’s resilience and vulnerability to socioeconomic deprivation. Child Dev, 2004. 75(3): p. 651-68.

174. Rolls, B.J., et al., Increasing the portion size of a sandwich increases energy intake. J Am Diet Assoc, 2004. 104(3): p. 367-72.

175. Raynor, H.A. and L.H. Epstein, The relative-reinforcing value of food under differing levels of food deprivation and restriction. Appetite, 2003. 40(1): p. 15-24.

176. Raynor, H.A. and L.H. Epstein, Dietary variety, energy regulation, and obesity. Psychol Bull, 2001. 127(3): p. 325-41.

177. Plomin, R., J.C. DeFries, and J.C. Loehlin, Genotype-environment interaction and correlation in the analysis of human behavior. Psychol Bull, 1977. 84(2): p. 309-22.

178. Loehlin, J.C., Behavior genetics and parenting theory. Am Psychol, 2001. 56(2): p. 169-70.

179. Collins, W.A., et al., Toward nature with nurture. Am Psychol, 2001. 56(2): p. 171-3.

180. Collins, W.A., et al., Contemporary research on parenting. The case for nature and nurture. Am Psychol, 2000. 55(2): p. 218-32.

181. Wachs, T.D. and R. Plomin, Conceptualization and Measurement of Organism-Environment Interaction. 1991, Washington, DC: American Psychological Association.

182. Hetherington, E.M., D. Reiss, and R. Plomin, Separate Social Worlds of Siblings. 1994, Hillsdale, NJ: Lawrence Erlbaum Associates Publishers.

183. Lynch, M. and B. Walsh, Genetics and Analysis of Quantitative Traits. 1998, Sunderland, MA: Sinauer Associates, Inc.

184. Lytton, H., Parent-Child Interaction. The Socialization Process Observed in Twin and Singleton Families. 1980, New York: Plenum Press.

185. Emde, R.N. and J.K. Hewitt, Infancy to Early Childhood. Genetic and Environmental Influences on Developmental Change. 2001, New York: Oxford University Press.

186. Reiss, D., et al., The Relationship Code. Deciphering Genetic and Social Influences on Adolescent Development. 2000, Cambridge, MA. London: Harvard University Press.

187. de Castro, J.M. and S.S. Plunkett, How genes control real world intake: palatability-intake relationships. Nutrition, 2001. 17(3): p. 266-8.

188. de Castro, J.M., Eating behavior: lessons from the real world of humans. Nutrition, 2000. 16(10): p. 800-13.