Research over the past two decades has provided an unprecedented expansion of our knowledge about the physiologic and molecular mechanisms regulating body weight and body fat [1]. One great step was the cloning of genes corresponding to five types of obesity in experimental animals that were due to single genes–so-called monogenic obesity syndromes–and the ensuing characterization of the human counterparts to these syndromes [2,3]. Subsequent research has added a number of other genes with lesser effects to the list of genes modifying obesity [4]. Extensive molecular and reverse genetic studies (mouse knockouts) have helped to identify critical pathways regulating body fat and food intake, and have validated or refuted the importance of previously identified pathways [2].

This chapter reviews this rapidly expanding literature from two perspectives. The first is an epidemiologic approach, considering environmental agents that affect the human being. The second views body weight regulation from a “set-point” or homeostatic approach by considering the way in which one part of the metabolic system communicates with another and how this system may be “overridden” by hedonic or pleasure centers.

The epidemic of overweight is occurring on a genetic background that does not change as fast as the epidemic has been exploding. It is nonetheless clear that genetic factors have an important role in its development [2,4]. One analogy for the role of genes in overweight is that “genes load the gun and a permissive or toxic environment pulls the trigger.”

Identification of genetic factors involved in the development of obesity increases yearly. From the time of the early twin and adoption studies more than 10 years ago [2], the focus has been on evaluating large groups of individuals for genetic defects related to the development of overweight [3,4]. These genetic factors can be divided into two groups: the rare genes that produce excess body fat and a group of more common genes that underlie susceptibility to becoming overweight – the so-called ‘susceptibility’ genes [2,4]. The field of genetic factors has been given a recent boost from genome-wide association studies, in which variants in large populations of tens of thousands of people are examined [4]. Using this genome-wide association strategy, 17 genes are now known that account for a small fraction of the variance in human body weight [5]. The most important of these is the FTO gene, which accounts for half of this effect. All of these genes are thought to affect the regulation of food intake.

Underlying the following discussion is the reality that genetic responses to the environment differ between individuals and affect the magnitude of the weight changes. Several genes have such potent effects that they produce overweight in almost any environment where food is available. Leptin deficiency is one of them [3]. Most other genes that affect the way body weight and body fat vary under different environmental influences have only a small effect. That these small differences exist and differ between individuals accounts for much of the variability in the response to diet that we see.

Over the past decade it has become clear that infants who are small for their age are at higher risk for metabolic diseases later in life. This idea was originally proposed by Professor David Barker and is often called the Barker or Developmental Origins of Health and Disease hypothesis [6]. Several examples illustrate its role in human obesity. The first was the Dutch winter famine of 1944, in which the calories available to the residents of the city of Amsterdam were severely reduced by the Nazi occupation [7]. During this famine, intrauterine exposure occurred during all parts of the pregnancy; caloric restriction during the first trimester increased the subsequent risk of overweight in the offspring [7].

Two other examples that fall into the category of fetal imprinting are the increased risk of obesity in offspring of mothers with diabetes [8] and in the offspring of mothers who smoked during the individual’s intrauterine period [9,10]. In a study of infants born to Pima Indian women before and after the onset of maternal diabetes, Dabelea et al. [8] noted that the infants born after diabetes developed were heavier than those born to the same mother before diabetes developed [11,12]. The risk for overweight at age 3 years was predicted by smoking at first prenatal visit with an odds ratio (OR) of 2.16 (95% confidence interval [CI] 1.05–4.47). Despite being smaller at birth, these infants more than caught up by age 3 years [11,12].

Smoking during pregnancy increases the risk of overweight at entry to school from just under 10% to over 15% if smoking continued throughout pregnancy and to nearly 15% if it was discontinued after the first trimester, indicating that most of the effect is in the early part of pregnancy [13,14].

One way to view the etiology of increased body fat is from the epidemiologic or environmental perspective. Food, medications, viruses, toxins and sedentary lifestyle can each act on the host to produce increased fatness. We need to remember, however, that for each of these agents there are genetic components.

Food is an environmental agent for obesity

We obtain all of our energy from the foods we eat and the beverages we drink. Thus, without food there could be no life, let alone excess stores of fat. The cost of this food is an important determinant of food choices. In addition to cost and total quantity, food styles of eating and specific food components may be important in determining whether or not we become fat.

Costs of food

Economic factors may have an etiologic role in explaining the basis for the intake of a small number of “excess calories” over time that leads to overweight [1]. What we consume is influenced by the price we have to pay for it. In the recent past, particularly since the beginning of the 1970s, the prices of foods that are high in energy density (fat and sugar-rich) have fallen relative to other items. The Consumer Price Index rose by 3.8% per year from 1980 to 2000 [15] compared with the rise in food prices, which rose by 3.4% per year. In the period 1960–1980, when there was only a small increase in the prevalence of overweight, food prices rose at a rate of 5.5% per year – slightly faster than the Consumer Price Index, which grew at a rate of 5.3% per year. The relative prices of foods high in sugar and fat have decreased since the early 1980s compared with those of fruits and vegetables. By comparison, Finkelstein et al. [15] note that between 1985 and 2000, the prices of fresh fruits and vegetables rose 118%, fish 77% and dairy 56%, compared with sugar and sweets, which rose only 46%, fats and oils 35% and carbonated beverages only 20%. Is it any wonder that people with limited income eat more sugar and fat-containing foods?

Quantity of food eaten

Eating more food energy over time than we need for our daily energy requirements produces extra fat. In the current epidemic the increase in body weight is on average 0.5–1 kg/year. The amount of net energy storage required by an adult to produce 1 kg of added body weight, 75% of which is fat, can be calculated by using a few assumptions. One kilogram of adipose tissue contains about 7000 kcal (29.4 mJ) of energy. If the efficiency of energy storage were 50%, with the other 50% being used by the synthetic and storage processes, we would need to ingest 14 000 kcal (58.8 mJ) of food energy. As there are 365 days in the year this would be an extra 20 kcal/day (40 kcal/day × 365 day/year = 14600 kcal) [16]. For simplicity we can round this to 50 kcal/day or the equivalent of 10 teaspoons of sugar.

Has the intake of energy increased? The energy intake (kcal/day) was relatively stable during the first 80 years of the 20th century. During the last 20 years, however, there was a clear rise from about 2300 kcal/day to about 2600 kcal/day, or an increase of 300kcal/day. This is more than enough to account for the 50 kcal/day net (100 kcal gross) required to produce the 1 kg weight gain each year [17].

Portion size

Portion sizes have dramatically increased in the past 40 years [18] and now need reduction. One consequence of the larger portion sizes is more food and more calories [17]. The US Department of Agriculture (USDA) estimates that between 1984 and 1994 daily calorie intake increased by 340 kcal/day or 14.7%. Refined grains provided 6.2% of this increase, fats and oils 3.4%, but fruits and vegetables only 1.4% and meats and dairy products only 0.3%. Calorically sweetened beverages that contain 10% high-fructose corn syrup (HFCS) are made from these grain products. These beverages are available in containers of 12, 20 or 32 oz, which provide 150, 250 or 400 kcal if all is consumed. Many foods list the calories per serving, but the package often contains more than one serving. In 1954, the burger served by Burger King weighed 2.8 oz and had 202 kcal. By 2004, the size had grown to 4.3 oz and 310 kcal. In 1955, McDonald’s served French fries weighing 2.4 oz and having 210 kcal. By 2004, this had increased to 7 oz and 610 kcal. Popcorn served at movie theaters has grown from 3 cups containing 174 kcal in 1950 to 21 cups with 1700 kcal in 2004 [19]. Nielsen & Popkin [18] have examined the portion sizes consumed by Americans and have shown the increased energy intake associated with the larger portions of essentially all items examined. Guidance for intake of beverages suggests intake of more water, tea, coffee and low-fat dairy products with lesser consumption of beverages that contain primarily water and caloric sweeteners [20]. The importance of drinking water as an alternative to consuming calories is suggested in a recent study. There was an inverse relationship between water intake expressed per unit of food and beverage intake and total energy intake. When the water intake was less than 20 g/gram of food + beverages, energy intake was 2485 kcal/day. At the highest quartile, when water intake was at or above 90 g/gram of food + beverage, energy intake had fallen to 1791 kcal/day. Thus, drinking water may be one strategy for lowering overall energy intake [21].

Energy density

Energy density interacts with portion size to affect how much is eaten. Energy density refers to the amount of energy in a given weight of food (kcal/g). Energy density of foods is increased by dehydrating them or by adding fat. Conversely, lower energy density is produced by adding water or removing fat. When energy density of meals was varied and all meals were provided for 2 days, the participants ate the same amount of food, but as a result obtained more energy when the foods were higher in energy density. In this experiment, they obtained about 30% less energy when the meals had low rather than high energy density [22,23]. When energy density and portion size were varied, Kral et al. [24] showed that both factors influence the amount that is eaten. The meals with low energy density and small portion sizes provided the fewest calories (398 kcal vs 620 kcal) [24].

Styles of eating

Breastfeeding is a case in which the style of eating can be associated with later weight gain. In infants, breast milk is their first food, and for many infants their sole food for several months. There are now a number of studies showing that breastfeeding for more than 3 months significantly reduces the risk of being overweight at entry into school and in adolescence when compared with infants who are breastfed for less than 3 months [25]. This may be an example of “infant imprinting” [26,27].

Restaurants and fast-food establishments

Eating outside the home has increased significantly over the past 30 years. There are now more fast-food restaurants (277 208) than churches in the USA [28]. The number of fast-food restaurants has risen since 1980 from 1 per 2000 people to 1 per 1000 Americans. Of the 206 meals per capita eaten out in 2002, fast-food restaurants served 74% of them. Other important figures are that Americans spent $100 billion on fast food in 2001, compared to $6 billion in 1970. An average of three orders of French-fried potatoes are ordered per person per week, and French-fried potatoes have become the most widely consumed vegetable. More than 100 000 new food products were introduced between 1990 and 1998. Eating outside the home has become easier over the last four decades as the number of restaurants has increased, and the percent of meals eaten outside the home reflects this. In 1962, less than 10% of meals were eaten outside the home. By 1992 this had risen to nearly 35%, where it has remained. In a telephone survey of body mass index (BMI) in relation to proximity to fast-food restaurants in Minnesota, however, Jeffrey et al. [29] found that eating at a fast-food restaurant was associated with having children, with eating a high-fat diet and having a high BMI, but not with proximity to the restaurant.

Eating in a fast-food restaurant also changes the foods consumed [30,31]. Paeratakul et al. [30] compared a day in which individuals ate at a fast-food restaurant with a day when they did not. On the day when food was eaten in the fast-food restaurant, less cereal, milk and vegetables were consumed, but more soft drinks and French-fried potatoes were eaten. Similar findings were reported by Bowman et al. [31]. who reported in addition that on any given day, over 30% of the total sample group consumed fast food. In this national survey, several other features were also associated with eating at fast-food restaurants, including being male, having a higher household income and residing in the US South. Children who ate at fast-food restaurants consumed more energy, more fat and added sugars, and more sweetened beverages than children who did not eat at fast-food restaurants.

Night-eating syndrome

The original description of the night-eating syndrome was published in a classic paper by Stunkard in 1955 and updated recently [32]. Recent studies have refined this syndrome. It consists of individuals who eat more than 50% of their daily energy intake during the nighttime [1].

Frequency of food intake

Frequency of eating may increase the risk of obesity. Crawley & Summerbell [33] showed that among males, but not females, the number of meal-eating events per day was inversely related to BMI. Males with a BMI of 20–25 ate just over 6 times per day compared to less than 6 times for those with a BMI >25kg/m².

Eating breakfast is associated with eating more frequently, and there are data showing that eating breakfast is associated with lower body weight. Eating breakfast cereal has been related to decreased BMI in adolescent girls. Using longitudinal data on adolescent girls, Barton et al. [34] showed that as cereal intake per week increased from 0 to 3 times per week, there was a small, but significant, decrease in BMI.

Calorically sweetened soft drinks

One of the consequences of the lower farm prices in the 1970s was a drop in the price of corn, which made inexpensive the production of corn starch which is converted to HFCS. With the development of the isomerase technology in the late 1960s which could convert starch into the highly sweet molecule, fructose, manufacture of soft drinks entered a new era [35]. From the early 1970s through the mid-1990s, HFCS gradually replaced sugar in many manufactured products, and almost entirely replaced sugar in soft drinks manufactured in the USA. In addition to being cheap, HFCS is very sweet. We have argued that this “sweetness” in liquid form is one factor driving the consumption of increased calories which are needed to fuel the current epidemic of obesity.

The relationship of soft-drink consumption to calorie intake, to body weight and to the intake of other dietary components has been examined in both cross-sectional and longitudinal studies [36]. Of the 11 cross-sectional studies examining the relation of caloric intake and soft-drink consumption, nine found a moderately positive association. Among the four longitudinal studies, the strength of the association was slightly stronger. The authors conclude that when humans consume soft drinks there is little caloric compensation. That is, the soft drinks are “added” calories and do not lower the intake of energy in other forms. The strengths of these relationships were stronger in women and in adults. Not surprisingly, they found that studies funded by the food industry had weaker associations than those funded by independent sources.

Several studies on the consumption of calorically sweetened beverages in relation to the epidemic of overweight have received significant attention [35]. Ludwig et al. [37] reported that the intake of soft drinks was a predictor of initial BMI in children in the Planet Health Study. They went on to show that higher soft-drink consumption also predicted the increase in BMI during nearly 2 years of follow-up. Those with the highest soft-drink consumption at baseline had the highest increase in BMI. In one of the few randomized well-controlled intervention studies, Danish investigators [38] showed that individuals consuming calorically sweetened beverages during 10 weeks gained weight, whereas subjects drinking the same amount of artificially sweetened beverages lost weight. Equally important, drinking sugar-sweetened beverages was associated with a small, but significant, increase in blood pressure. Women in the Nurses’ Health Study [39] also showed that changes in the consumption of soft drinks predicted changes in body weight over several years of follow-up. In children, a study focusing on reducing intake of “fizzy” drinks and replacing them with water showed slower weight gain than for those not advised to reduce the intake of fizzy drinks [40].

Fructose consumption, either in beverages or food, may have an additional detrimental effect. It has been linked to the development of cardiometabolic risk factors and the metabolic syndrome in participants in the Framingham Study [41]. Cross-sectionally, individuals consuming ≥1 soft drink per day had a higher prevalence of the metabolic syndrome (OR 1.48; 95% CI 1.30–1.69) and an increased risk of developing the metabolic syndrome over 4 years of follow-up. It may also increase the risk of gout [42] and diabetes [43].

Dietary fat

Dietary fat is another component of the diet that may be important in the current obesity epidemic [44–46]. In epidemiologic studies, dietary fat intake is related to the fraction of the population that is overweight [45]. In an 8-year follow-up of the Nurses’ Health Study, Field et al. [47] found a weak overall association of percent fat and a stronger effect of animal fat, saturated fat and trans fat on fatness. In experimental animals, high-fat diets generally produce fat storage. In humans, the relationship of dietary fat to the development of overweight is controversial. It is certainly clear that ingesting too many calories is essential for the increase in body fat. Because the storage capacity for carbohydrate is very limited, it must be oxidized first. Thus, when people overeat, they oxidize carbohydrate and store fat. When fat is a large component of a diet, the foods tend to be “energy dense” and thus overconsumption is easy to achieve.

Low levels of physical activity

Epidemiologic data show that low levels of physical activity and watching more television predict higher body weight [48]. Recent studies suggest that individuals in US cities where they had to walk more than people in other cities tended to weigh less [49]. Low levels of physical activity also increase the risk of early mortality [50]. Using normal weight, physically active women as the comparison group, Hu et al. [51] found that the relative risk of mortality increased from 1.00 to 1.55 (55%) in inactive lean women compared with active lean ones, to 1.92 in active overweight women, and to 2.42 in women who are overweight and physically inactive. It is thus better to be thin than fat and to be physically active rather than inactive.

Television has been one culprit blamed for the reduced levels of physical activity, particularly in children. The first suggestion that TV viewing was associated with overweight was published by Gortmaker and Dietz. Using data from the National Health Examination Survey [52] and the National Longitudinal Study of Youth [53], they found a linear gradient from 11–12% overweight in children watching 0–2 hours/day to over 20–30% when watching more than 5 hours/day. Since that time a number of studies have shown that in both children and adults, those who watch TV more are more overweight. By one estimate about 100 kcal of extra food energy is ingested for each hour of TV viewing. In studies focusing on reducing sedentary activity, which largely means decreasing TV viewing, there was a significant decrease in energy intake with increased activity [54]. In the Early Childhood Longitudinal Study, investigators found that between kindergarten and third grade, children watching more TV (OR 1.02) and eating fewer family meals together (OR 1.08) predicted a modest increase in weight [55].

Effect of sleep time and environmental light

Sleep time declines from an average of 14.2 ± 1.9 (mean ± SD) hours/day in infancy (11.0 ± 1.1 hours/day by 1 year of age) to 8.1 ± 0.8 hours/day at 16 years of age [56]. Sleep time declined across the cohorts from 1974 to 1993 due largely to later bedtime, but similar arising time. Nine epidemiologic studies have been published that relate shortness of sleep time with overweight. Six of these studies are cross-sectional in design, and three are longitudinal. The earliest of these studies was only published in 1992, but most were published after 2002. They include both children and adults. In a small case-control study involving 327 short-sleepers compared with 704 controls, Locard et al. [57] found that short-sleepers were heavier than the controls.

In two large cross-sectional studies in children, Sekine et al. [58] and von Kries et al. [59] found that there was a dose-dependent relationship between the amount of sleep and the weight of children when they entered school. Von Kries et al. [59] studied 6862 children aged 5–6 years whose sleeping time was reported in 1999–2000 by the parent, and followed-up in 2001–2002. Overweight in this study was defined as a weight for height greater than the 97th percentile. Children with reported sleeping time of less than 10 hours had a prevalence of overweight of 5.4% (95% CI 4.1–7.0), those who slept 10.5–11.0 hours per night had a prevalence of 2.8% (95% CI 2.3–3.3) and those who slept more than 11.5 hours had a prevalence of overweight of 2.1% (95% CI 1.5–2.9). Among the 8274 children from the Toyama Birth Cohort in Japan [58], there was a graded increase in the risk of overweight, defined as a BMI above 25 kg/m²