Basic and Clinical Endocrinology 7th International student edition Edition


Obesity & Overweight

Marc K. Hellerstein MD, PhD

Elizabeth J. Parks PhD


Obesity is a disorder of body composition defined by a relative or absolute excess of body fat and characterized by several remarkable features. Its prevalence has increased dramatically over the past several decades, both in the industrialized and developing worlds. It is now no exaggeration to state that obesity is an international epidemic. Moreover, obesity is no longer a disorder of adults; prevalence in children has accelerated rapidly, so that 25% of United States children are overweight or obese. At the same time, body fatness has taken on a central—even obsessive—place in popular culture as images and pressures from nonmedical as well as medical sources are ubiquitous and disapproving. This central epidemiologic paradox sets off obesity as occupying a special place among the physical ailments in contemporary society.

Uncertainty about the etiology, pathogenesis, and treatment of obesity is a key element in this paradox. Although the health risks associated with excess body fat are increasingly well documented, therapies remain generally ineffective. Persistent therapeutic failure, accelerating prevalence, and rising morbidity of excess body weight are occurring despite enormous efforts by the medical profession, pharmaceutical industry, and weight-loss industry (in addition to popular culture). This dissonance between therapeutic effort and efficacy represents a second paradox in this area.

The objective of this chapter is to present obesity with a focus on its best-understood aspect—the physiology of energy and fat balance—and to use these physiologic insights as a framework for discussion of current therapies. Because most patients seen by a physician in the United States have probably thought or worried about “getting fat” and typically have a number of questions and misconceptions, many physicians are uneasy about discussing “fatness” with their patients and their patients' children. A central goal here is thus to help health care providers answer questions that may be asked by their patients. Toward this end, what is not known will be directly identified and noted in addition to what is known.

Within this framework, the chapter will focus on the following issues faced by health care providers: (1) Why should obesity be treated, and when? (2) What are the causal factors and pathogenic pathways contributing to overweight, and which of these can be manipulated therapeutically? (3) What specific treatments are available or under investigation, and what are their demonstrated efficacy and risks? (4) Which other medical conditions are affected by body fatness and its treatment? (5) What areas of the field remain controversial or unknown? (6) What are some of the common questions and misconceptions that patients have, and how can we best respond to them?

Definition & Diagnosis of Overweight & Obesity

Obesity is best defined as the presence of an abnormal absolute amount or relative proportion of body fat. The presence of excess body fat usually—not always—results in higher body weight. The criteria for “abnormal” amounts of body fat or body weight can be purely statistical (ie, based on population means) or, more usefully, on epidemiologic associations with adverse health


events. The classification criterion used in most large studies has been the body mass index (BMI), derived by dividing the body weight (in kilograms) by the square of the height in meters—or by dividing the weight in pounds by the square of the height in inches and multiplying by 703. Because most of the health outcome data are with BMI, this is at present the recommended basis for classifying overweight and obesity (Table 20-1). The term “morbid obesity” has also been used to emphasize the extreme health risks of body weights above 150 kg.

Although it may be most convenient to use body weight for identification of obese patients, it is important to keep the true definition clear: obesity is a disorder of body fat stores (adiposity). Maintaining this emphasis will help us when thinking about pathogenesis and treatment, wherein body weight and body composition must often be clearly separated.

One important complication regarding diagnosis relates to body fat distribution. All body fat is not created equal. Central or visceral-abdominal obesity (“apple-shaped”) is associated with substantially different metabolic profiles and cardiovascular risk factors than gluteal-femoral obesity (“pear-shaped”). Overweight women tend to have different body fat distribution than men, and metabolic and cardiovascular risks vary in parallel. Hypothesized reasons for these differences in risk for adverse health consequences are discussed below. For the clinician, the distinction is extremely important, and overweight patients should be broadly classified as having central or gluteal-femoral types. This can be done quickly and easily by measurement of the waist circumference, using a tape measure. High risk abdominal obesity is defined as waist circumference > 102 cm (> 40 inches) in men and > 88 cm (> 35 inches) in women.

Table 20-1. Classification of overweight and obesity based on body mass index (BMI).




 < 18.5







   Class I


   Class II


   Class III (extreme)

> 40.0


Prevalence of Obesity

The late 20th century has witnessed an extraordinary change in the epidemiology of human body composition. What was in 1950 or 1960 a prevalent condition in the West—particularly the United States and certain Northern European countries—that was observed primarily in older adults has expanded in all respects. The prevalence has reached “epidemic” proportions in the West, increasing even in the past decade in United States adults from 12% obese (defined as BMI > 30) in 1991 to 17.9% in 1998. The prevalence of overweight (defined as BMI > 25) has increased even more, representing more than 60% of men and more than 55% of women in 1994. In some ethnic or racial populations (eg, Mexican-American or African-American women), the prevalence of overweight is greater than 65%. Children are increasingly affected and represent the fastest-growing population group of overweight and obese people in the United States. Children in certain ethnic or racial groups are disproportionately affected. Perhaps most significantly, the rest of the world is catching up to the West. Obesity is now prevalent in regions where it was historically rare, such as South America, Central America, Southeast Asia, Australia, and Africa. This international change is linked to urbanization, with attendant changes in physical activity and a shift away from traditional diets.

The prevalence of obesity is also affected by age. Body composition and the prevalence of obesity vary during the life cycle (Table 20-2). Body weight tends to peak in United States men and women in the sixth and seventh decades, then to fall with further increasing age.

Health Consequences of Overweight & Obesity

Why identify and treat overweight or obesity? Some authors have suggested that body weight should be considered in a manner analogous to characteristics such as height: largely inherited, naturally variable within a population, and best accepted in an individual rather than opposed through interventions. Others have pointed out the considerable social and medical costs of the widespread popular concern about and therapies for perceived excess fatness. These criticisms express considerable truth—the social, psychological and financial costs of both lay and medical approaches to adiposity are indeed great.

The problem is that obesity is not a benign variant. Convincing data exist from a number of studies documenting the substantial disease burden associated with overweight and obesity. In 1991, death and disability attributable to obesity (BMI > 30) has been estimated


at 280,000–325,000 people per year in the United States. Previous studies have given similar estimates. Few factors, with the exception of cigarette smoking, exact as large a public health toll. Excess body fat must be taken seriously from a health perspective.

Table 20-2. Combined prevalence (%) of overweight (BMI > 25) and obesity (BMI > 30) in United States adults by age, gender, and race or ethnicity (19601994).


NHES I 196062 (Ages 2074)

NHANES I 197174 (Ages 2074)

NHANES II 197680 (Ages 2074)

HHANES 198284 (Ages 2074)

NHANES III 198894 (Ages 20)













































































Both sexes
















White men





White women





Black men





Black women





White, non-Hispanic men





White, non-Hispanic women



Black, non-Hispanic men



Black, non-Hispanic women



Mexican-American men




Mexican-American women



NHES I = National Health Examination Survey I.
NHANES I = National Health and Nutrition Examination Survey I.
NHANES II = National Health and Nutrition Examination Survey II.
HHANES = Hispanic Health and Nutrition Examination Survey I.
NHANES III= National Health and Nutrition Examination Survey III.
N/A = Not available

It is difficult to be certain about the causal link between diseases and body fat. These relationships are confounded by the many metabolic and hormonal alterations that either contribute to or derive from excess adiposity. For example, low cardiorespiratory fitness, though closely linked with body weight, is a strong and independent predictor of cardiovascular disease and mortality. In some studies, people in the > 35 BMI category with a good level of fitness have a lower mortality rate than people in the 25–30 BMI category with poor fitness. The specific mechanisms or agencies by which body fat promotes adverse outcomes are not always known. This will be a critical task for future research.

Some of the more serious medical conditions associated with overweight can be listed. Although a number


of differences exist among ethnic groups in the United States with regard to prevalence and complications of overweight and obesity, the graded and significant relationship between BMI and prevalence of these health comorbidities applies for all groups.


The risk of type 2 diabetes mellitus increases linearly with increasing BMI in the general population. In comparison with a normal-weight individual, BMI > 40 (obesity class 3) in a person under 55 years of age increases the risk of type 2 diabetes mellitus by 18.1-fold in men and 12.9-fold in women. Classification as overweight (BMI 25–30) increases the prevalence of type 2 diabetes mellitus threefold to fourfold in both men and women under 55 years of age. The relative risks associated with excess body weight are less strong in older people but are still present (eg, about twofold for overweight and three- to sixfold for obesity class 3).

One of the most remarkable and worrisome features of the recent change in demographics of obesity is the rapid growth of type 2 (“adult”-type) diabetes in children and teenagers. In many clinics, particularly in urban centers, the majority of new childhood diabetics are overweight, not ketosis-prone, and lacking in immunologic markers of type 1 diabetes. Childhood type 2 diabetes had until the past decade been an extremely unusual condition, often associated with a strong familial component and mendelian inheritance (eg, glucokinase deficiency in some kindreds). The cause and pathogenesis of this extraordinary change in childhood diabetes remain uncertain, but the association with the increasing prevalence of obesity, sedentary lifestyle, or specific dietary factors is difficult to ignore. If this syndrome of childhood obesity and diabetes becomes established as a common consequence of the urbanized industrial world, the long-term health burden may be enormous (eg, for premature heart disease, peripheral vascular disease, renal failure, blindness, and erectile dysfunction).


High blood pressure is the most common health condition associated with overweight or obesity in the United States. The relationship between hypertension prevalence and body weight for both men and women under 55 years of age is steeply graded, with relative risk increasing from 1.6 for overweight to 2.5–3.2 for obesity class 1 (BMI 30–35) and 3.9–5.5 for obesity classes 2 and 3 (BMI 35–40 and >40). The prevalence of hypertension is much higher with aging even in normal-weight individuals. The absolute number of persons affected by overweight or obesity is still substantial in those over age 55 despite a less steep relationship with body weight.


The presence of overweight or obesity is associated with high blood cholesterol, but the strength of the relationship is only moderate (less than a twofold relative risk at all weight categories and ages). The relationship to other blood lipid parameters such as hypertriglyceridemia, low HDL cholesterol, and altered composition of lipoprotein particles is much stronger. The difference in serum triglyceride concentrations between individuals with BMI < 21 and those with BMI > 30 is about 65 mg/dL (in women) and 62–118 mg/dL (in men, depending upon age). For HDL cholesterol, each 1-unit BMI change is associated with an HDL-cholesterol decrease of 1.1 mg/dL for young men and 0.7 mg/dL for young women (Figure 20-1).


The prevalence of coronary artery disease exhibits a significant linear relationship with BMI for both genders and all ages. The relative risk for overweight persons is not statistically significant, but obesity classes 1–3 exhibit relative risks generally between 1.5 and 3.0, with the highest values in obesity class 3. Cerebrovascular accidents are also associated with obesity.


The prevalence of gallbladder disease is strongly associated with higher BMIs. The relative risks with increasing BMI range from fourfold to 21-fold elevated in men and 2.5-fold to 5.2-fold in women under age 55.


Osteoarthritis is also more prevalent in association with higher body weights, and symptoms often improve with weight loss. Several types of cancer are more prevalent in adults who are overweight or obese. Examples include endometrial, breast, prostate, and colon cancer. Sleep apnea is also highly associated with BMI > 30. Menstrual irregularity, infertility, and polycystic ovarian syndrome are more prevalent in premenopausal women with higher BMIs.


Figure 20-1. NHANES III age-adjusted prevalence of low HDL-cholesterol according to body mass index. Low HDL-cholesterol defined as < 35 mg/dL in men and < 45 mg/dL in women.

Etiology & Pathogenesis

The cause of obesity is multifactorial; that is, a large number of “pathways”—biochemical, dietary, and behavioral—exist that can contribute to the accrual of body fat (Figure 20-2). The nonmendelian (polygenic) inheritance patterns of obesity and body weight are also consistent with the operation of multiple interacting factors.

In a general sense, however, the pathogenesis of obesity (accrual of excess body fat) is simple and clear. There must have been an imbalance between the energy intake and the energy needs of the tissues—or, more specifically, between the intake or synthesis of fat and its oxidation by tissues—at some point in the life of the individual. Moreover, the resulting storage of surplus energy and fat in the only large energy storage depot available in humans (triacylglycerols in adipocytes) has not induced an effective feedback adaptation to dissipate accrued stores.

At its core, then, obesity is a disorder in the operation of a balancing system—ie, an interactive disorder—wherein the system for matching energy intake to expenditure is disrupted. What do we know about energy and nutrient balance systems in humans? Although much is not understood, a number of basic principles have emerged about the operating rules of this system. Understanding these physiologic principles about macronutrient balances is essential for an understanding of the pathogenesis and treatment of obesity.

Some Features of the Macronutrient & Energy Balancing Systems in Humans Can Be Described as Follows


The relationship between daily intake and size of body stores for the different macronutrients (carbohydrates, fats, and proteins) are very different, with important consequences for tissue fuel selection.

The relationship between intake and body stores for each macronutrient (Figure 20-3) shows the unique status of body carbohydrate stores. In addition to the extremely limited storage pool of carbohydrates (< 1000–2000 kcal in the human body), several biochemical facts all lead to the prediction that a system for sensing and preserving body carbohydrate stores over the short term must exist: Fats cannot be converted to carbohydrates in animals; conversion of amino acids to carbohydrates draws upon essential protein stores and is therefore undesirable; and the brain depends almost exclusively on glucose as a fuel under most conditions. Failure to conserve body carbohydrate stores in times of limitation would therefore have predictable adverse phenotypic consequences for the organism (protein wasting or neuroglycopenia), on the one hand. Inability to store excess carbohydrates indefinitely (due to limited whole-body storage capacity for glycogen) in times of surplus is similarly constraining—in the other direction. In contrast, a system for short-term sensing of body fat stores is not likely to be essential, as attested by the great surplus of fat calories relative to daily needs that exists in the body (Figure 20-3).

This quantitative prediction has been strongly supported by experimental data. Dietary carbohydrate intake stimulates its own oxidation, and dietary carbohydrate deprivation reduces whole-body oxidation of carbohydrates. These adaptations in fuel selection of the body to carbohydrate intake occur within a few days. An important feature of this response is that the conversion of surplus dietary carbohydrate to fat through the de novo lipogenesis pathway cannot be a quantitatively important process within this system. If de novo lipogenesis represented a quantitatively significant disposal route, it would act as a safety valve for dietary carbohydrate and obviate the necessity to oxidize surplus carbohydrate stores. Experimental evidence in humans makes it clear that surplus carbohydrate energy is not converted to fat. Instead, when extra carbohydrate calories are added to a mixed diet that includes fats, the oxidation of dietary fat is suppressed. The respiratory quotient (RQ) in this setting increases to 0.98–1.00, reflecting nearly complete replacement of fats by carbohydrate in the whole-body fuel mixture. Stable isotope metabolic measurements have shown that this RQ does not reflect conversion of carbohydrate to fat, followed by direct oxidation of fatty acids in tissues (eg, the former in the daytime, the latter at night). Thus, carbohydrate intake can lead to accrual of body fat, but this occurs by sparing oxidation of dietary fat rather than by conversion of carbohydrate to fat.


Short-term metabolic mechanisms have been identified for matching carbohydrate intake and oxidation but not for dietary fats.

The metabolic mechanisms responsible for matching whole-body oxidation of carbohydrates to their dietary intake are shown (Figure 20-4). The major players


are liver, pancreas, and adipose tissue. Hepatic glucose production closely parallels liver glycogen content. When stores are high, glycogenolysis and glucose production are relatively insensitive to suppression by insulin; when stores are low, the opposite occurs. Expansion of liver glycogen stores therefore leads to increased input of glucose into the circulation. In response, pancreatic insulin secretion increases and serum insulin concentrations rise. As a result, adipocyte lipolysis is inhibited and tissue uptake and oxidation of glucose are stimulated, though glucose production is not suppressed (the liver is insulin-resistant). The net effect is greater oxidation of glucose and less of fat even under postabsorptive conditions (Figure 20-4). Interestingly, increased endogenous glucose production mimics exogenous carbohydrate intake with regard to stimulating whole-body carbohydrate oxidation, suppressing fat oxidation, and accelerating accrual of body fat. Indeed, the best predictor of obesity in Sprague-Dawley rats placed on ad libitum high-fat diets is a high rate of baseline hepatic glucose production and its resistance to suppression by insulin.


Figure 20-2. Biochemical and physiologic pathways to body fat accrual


Figure 20-3. Comparison of storage pool size and daily throughput for three macronutrients. A typical Western diet is represented (40% fat, 45% carbohydrate, 15% protein; 2400 kcal/d in lean, 3000 kcal/d in obese) and compared for a lean (weight, 70 kg; fat, 20%) and an obese (weight, 100 kg; fat, 45%) man. Note different scales.


Figure 20-4. Metabolic mechanisms by which surplus carbohydrate intake leads to body fat accrual by suppressing dietary fat oxidation rather than by conversion of carbohydrates into fats. (CHO, carbohydrate; DNL, de novo lipogenesis; HGP, hepatic glucose production; ox, oxidation; INS, insulin.) “?” refers to signal of surplus CHO in organism (not proven, but likely an increase in hepatic glycogen stores).





In contrast, intake of surplus dietary fat has essentially no impact on whole-body fuel selection or almost any other short-term hormonal or metabolic measures. Dietary fat is “invisible” in the short term; there are no metabolic mechanisms for sensing or adjusting expenditures in response to changes in intake of fat. As a result, dietary fat can be stored without inducing any short-term counterresponse in whole-body substrate oxidation.

There is also evidence from animal studies that tissue carbohydrate stores may exert feedback effects on appetite over the short term as part of the mechanism for maintaining whole-body carbohydrate stores. If that is true, factors that promote carbohydrate oxidation will tend to stimulate food intake until tissue glycogen stores are restored. Some investigators have proposed that appetite-modulating effects of tissue carbohydrate stores may explain the lower energy intakes typically observed on high-carbohydrate, low-fat diets (see below).

These fundamental differences in fuel metabolism and storage between carbohydrates and fats may go a long way toward explaining our susceptibility to storage of fat in the body, especially when challenged by increased dietary fat intake. Carbohydrate stores are regulated tightly, while fat stores are not.


Carbohydrates and fats compete for tissue oxidation. Fat oxidation generates products (NADH, acetyl-CoA, ATP) that inhibit glucose uptake, glycolysis, and oxidation of pyruvate in muscle and liver (the Randle effect, or glucose-fatty acid cycle). In the other direction, carbohydrate utilization generates products (malonyl-CoA, α-glycerophosphate, oxaloacetate, citrate) that inhibit fatty acyl-CoA transport into mitochondria, stimulate their reesterification to triglycerides in the cytosol, inhibit ketogenesis from acetyl-CoA, and activate de novo lipogenesis. Carbohydrate utilization thereby inhibits fatty acid oxidation. Interestingly, carbohydrate oxidation tends to win out over fat oxidation; if both fuels are available (with insulin), the carbohydrate is preferentially oxidized. These biochemical observations are congruent with the physiologic model based on macronutrient pool sizes and metabolic fluxes wherein carbohydrates are predicted to have oxidative priority compared with fats (see above).


Fats (two-carbon precursors) cannot be converted to carbohydrates (three- to six-carbon compounds) because the required enzymes of the glyoxylate shunt do not exist in animals. A functional block also appears to exist under most conditions from carbohydrates to fats (ie, de novo lipogenesis is a pathway only used as a last-resort), as discussed above.

Because metabolic interconversion between energy macronutrients is neither free-flowing nor symmetric, it is therefore misleading to discuss a common “energy” currency. Instead, independent macronutrient systems exist in the body's energy economy, and it is inaccurate to discuss “energy intake” or “energy balance” as though there were a common energy currency in the body. It is more accurate physiologically to consider carbohydrates and fats as independent—though interacting—currencies. These considerations have led to the notion of “macronutrient” balances. According to this model, we should do our accounting of fat and carbohydrate balances separately.


The energy balance equation is dynamic, not static, and this may explain why there is long-term relative stability of body weight and fat stores.

One of the key points about energy balance and weight that is not generally appreciated is that most people do not continuously and relentlessly gain or lose weight. Over the long term, there tends to be relative stability of body weight and body fat stores in most individuals, at least in comparison with the energy throughput of the system. Quantitative estimates make this point clearly. The annual flux of energy through the body (about 1,000,000 kcal/yr) is enormous compared with the typical change in energy content of the body (< 2.5 lb, or 10,000 kcal/yr). This represents an “error” of less than 1% with regard to the balancing of intake and expenditure.

Why is relative weight stability the rule? To answer this question, it is useful to consider what would happen if a person in energy balance simply added a cup of ice cream (500 kcal) every night, and did everything else exactly the same (activity, other food intake). Would the person gain 3500 kcal (almost 0.45 kg) of weight and fat every week for the rest of his life (ie, 227 kg after 10 years)?

The answer is no. The physiologic reason is that weight gain or loss results in increases or decreases, respectively, in total energy expenditure (TEE). Thus, the energy balance system is adaptive (dynamic), not fixed. Moreover, the changes in TEE can be extremely large in response to changes in body weight. It is now well documented that adaptations of TEE to altered body weight are greatly in excess of what would be quantitatively predicted from standard energy costs of tissues calculated in weight-stable humans, whether obese or nonobese. Recent studies in humans by Leibel and colleagues at Rockefeller University make it possible to calculate the consequences on TEE of gaining or losing weight.

If a person gains 10% of body weight by overeating (ie, 7–12 kg in nonobese or obese subjects), about two-thirds


of the gain will be as body fat (5–8 kg) and one-third (2–4 kg) as fat-free mass (FFM). FFM is the main determinant of resting energy expenditure (REE; see below) and also affects TEE; the relationship is generally about 45–50 kcal TEE/kg FFM per day both in obese and in nonobese individuals. A 10% weight gain would therefore be expected to increase TEE by 100–200 kcal/d. Conversely, a 10% weight loss consists of about 80% fat and 20% FFM (2–3 kg), so that the expected reduction in TEE would be 100–150 kcal/d (2–3 kg FFM × 50 kcal/kg FFM/d).

But in fact, after stabilization at a 10% weight gain, increases in TEE are much greater than these values. TEE increases by 870 kcal/d (Table 20-3). Accordingly, there is a change in the relationship between TEE and FFM; the ratio increases from 45–50 to 55–60 kcal TEE/kg FFM per d. Carrying the extra weight (or fat) apparently changes the energy costs of all tissues rather than just adding energy costs of new tissue. Change in REE appears to account for only a minority (about 150 kcal/d) of the increased TEE; most of the TEE increase (500–700 kcal/d) relates in some manner to costs of activity.

Conversely, a stable 10% weight loss reduces TEE by 450–550 kcal/d; the relationship between TEE and FFM now falls to about 40 kcal/kg FFM/d (Table 20-3). In this direction, REE accounts for about half of the change in TEE.

The consequences of these changes are profound. If one were to gain 12 kg, he could then eat about 850 kcal/d above his previous intake level and maintain a new stable weight (at the expense of higher weight and fatness). Put differently, if one chose to eat 850 kcal/d above his current intake, he would likely gain about 12 kg and stop there because a new balance between intake and TEE would be reached.

Table 20-3. Adaptive nature of energy expenditure in response to weight gain or loss.


Weight (kg)

FFM (kg)

Fat (kg)

kcal TEE / kg FFM

TEE (kcal/d)

Δ TEE (kcal/d)

Weight gain








 . . .

      10% gain














 . . .

      10% gain







Weight loss








 . . .

      10% loss














 . . .

      10% loss







FFM = Fat-free mass.
TEE = Total energy expenditure.

In the other direction, if one were to lose about 17 kg, he then would need to eat about 550 kcal/d less than current intake to maintain this new weight indefinitely. A dieting individual who chooses to reduce the intake by 550 kcal would lose about 17 kg and stop there.

Weight stabilization is therefore reestablished even in the face of permanent changes on the intake side of the balance equation. This has a resemblance to a long feedback loop. An example of this feedback loop has been reported by Ravussin and colleagues, who found that low REE is the best predictor of subsequent long-term weight gain in young Pima Indians. However, REE (and TEE) increase when weight is gained by these individuals, which results in achievement of a new state of energy equilibrium. Thus, the thermogenesis due to weight gain ultimately balances the initial energy surplus relative to TEE.

From this perspective, obesity is the “solution” adopted by the nutrient system for reestablishing balance when faced with surplus fat or energy. Again, weight gain may be seen as a kind of long feedback loop that compensates for baseline deficiencies in TEE or appetite regulation.




The three determinants of body fat stores—intake, expenditure, and nutrient partitioning—are each under complex control (Figures 20-1, 20-4and 20-5). A few interesting features are worth emphasizing here.

  1. Appetite control pathways—Some of the hypothalamic nuclei involved in control of ingestive behavior are shown in Figure 20-5. Like most complex adaptive networks, interference with one pathway tends to induce compensatory changes in alternative or redundant pathways. This redundancy and network-like structure of the neurobiology of appetite control may in part explain difficulties in maintaining efficacy of some anorexigenic agents over the long term.
  2. Nutrient partitioning—Two points that are generally not appreciated regarding the role of macronutrient partitioning deserve to be emphasized. First is the role that muscle protein anabolic capacity can play in determining body fatness. Differences in the capacity to increase the mass of muscle tissue (muscle is normally the largest organ in the mammalian body, representing 40–45% of body weight) may underlie differences in the amount of fat tissue accrued under conditions of nutrient surplus (ie, spillover into fat stores may occur if lean tissue anabolism is blocked). Also, as discussed above, muscle tissue per se represents a calorigenic factor. Genetics, hormones, exercise, and other regulators of muscle metabolism may therefore play an important role in the pathogenesis of obesity.

Figure 20-5. Schematic of appetite regulatory pathways in humans. (PVN, paraventricular nucleus; ARC, arcuate nucleus; PFA, prefornical area; LHA, lateral hypothalamic area; NTS, nucleus of tractus solitarius; SNS, sympathetic nervous system; CCK, cholecystokinin; POMC, pro-opiomelanocortin; NPY, neuropeptide Y.) (Modified, with permission, from Schwartz MW et al: Central nervous system control of food intake. Nature 2000;404:661.)

A second interesting relationship is that observed between insulin resistance and body composition. It is widely appreciated that insulin resistance syndromes generally occur in association with overweight and excess adiposity. The impact of insulin resistance on body weight and macronutrient balance is less often considered. In Pima Indians of the Southwestern United States, resistance of peripheral tissues to insulin-mediated glucose uptake predicts lower subsequent body-fat stores and less weight gain. A physiologic explanation is that peripheral insulin resistance increases lipolysis while reducing


glucose oxidation; the result favors fat oxidation. The interesting implication here is that peripheral insulin resistance, occurring in association with obesity, may represent another long-loop feedback system to allow oxidation of the surplus fat that previously could not be oxidized. Similar to the increases in REE, peripheral insulin resistance may represent a long-term physiologic “solution” to the problem of macronutrient imbalance for fats. As discussed above, both of these long-term adaptations for achieving whole-body fat balance are necessary because no short-term control system exists for dietary or body fat. Of course, the adverse health consequences of insulin resistance (diabetes, hyperlipidemia, etc) may make this adaptation a poor biologic decision—from the health perspective—in modern society.

The consequences of hepatic insulin resistance on body weight are quite different from those of peripheral tissue (muscle and fat) insulin resistance. Increased hepatic glucose production and resulting hyperinsulinemia suppress whole-body fat oxidation and predict weight gain (see above).

  1. Energy expenditure—The major determinant of REE is FFM. Thus, “the lean stay lean” on the one hand, and anything that leads to increased metabolically active (lean) tissue will allow equilibration of body weight to be reached at higher levels of energy intake. Moreover, TEE increases out of proportion to changes in FFM when body weight is gained (see above). The net effect is to dampen the effect of any changes in energy intake on body weight, as just discussed.

Adipocytes, the storage cells for surplus energy in the body, have their own internal metabolic control systems that can affect the whole-body macronutrient economy.

Body fat stores are influenced by the number and the size (lipid content) of fat cells. Early studies suggested that the number of fat cells was fixed after early childhood and that subsequent gains in adipose mass must reflect increases in fat cell size. This is no longer believed to be true, however; adult humans have preadipocytes and the capacity to form new fat cells. However, the actual extent to which adipocytes differentiate and proliferate from precursors in adults remains unknown. Adipocyte lipogenesis, lipolysis, and avidity for uptake of fatty acids from circulating triglycerides can also influence the systemic availability and balances of lipids. Activity of the intravascular enzyme lipoprotein lipase in adipose tissue relative to muscle also may affect the proportion of fats that are stored or oxidized. The specific effects of these processes on clinical obesity have not been established.

The health risks of obesity are also modulated by the distribution of body fat. Compared with peripheral obesity (“pear-shaped”), midline or truncal obesity (“apple-shaped”) is associated with greater health risks. Visceral fat stores are believed to be more lipolytically active and have greater potential to impact liver metabolism given that their venous drainage flows directly to the liver. Visceral obesity is an important predictor of abnormal plasma lipoprotein concentrations; individuals with truncal obesity have higher lipoprotein and triglyceride production rates and elevations in postprandial lipemia. Even a small amount of weight loss may have a dramatic health benefit if it occurs in the abdominal region (see Chapter 19).


A number of controversies or unresolved questions pertaining to these matters deserve comment.

Genes Versus Environment

Estimates of genetic heritability have ranged from 20% to as high as 80%. There is no doubt that genetic factors predispose individuals toward different body composition. Many authors have taken this to mean that obesity is a genetic disease and that body fat content is a quantitative genetic trait. Yet the notion that obesity is a genetic disorder—even a complex genetic disorder—is misleading in important ways. This becomes apparent from consideration of a few facts.

  1. The prevalence of obesity has increased markedly worldwide in recent years, yet genes have not changed. Obviously, environmental factors explain population changes that have occurred.
  2. The Pima Indians who live on reservations in Arizona have the highest known prevalence of obesity and diabetes in the world. But their genetic “cousins” living in the mountains of Mexico have almost no obesity.
  3. Variability exists between populations—not only among individuals—and changes occur within populations when migration occurs.

A resolution of these observations is that the phenotypic expression of genes for obesity are environment-specific—ie, obesity is a disorder of the interaction between genes and environment.

Intake Versus Expenditure

The best predictors of future risk for obesity have been on the expenditure side rather than the intake side of


the energy balance equation. For example, hours of television watched per day predicts subsequent weight gain in children, and low REE predicts weight gain in Pima Indians. The relation between weight gain and secondary increases in TEE suggests that interventions to increase TEE independent of weight gain could allow energy equilibration on higher caloric intakes at nonobese body weights. Indeed, the more effective weight control strategies have acted to increase TEE (eg, exercise, nicotine, amphetamines; see below). In contrast, most treatment strategies aimed at controlling food intake have proved difficult to sustain. The role of diet in preventing weight gain cannot be ignored, however; in prospective studies, dietary fat intake predicts weight gain in children and adolescents.

Body Fat Set-Point Versus Settling Point

A subtle but important debate relates to the explanation for the relative weight stability generally present in people. Does this stability imply an active servomechanism (a system that actively resists changes in body weight), or can it be explained on a passive basis? Do the changes in TEE and fat oxidation observed when people gain or lose weight reflect a hormonally or neurally directed response to restore body composition toward a sensed and regulated value? Or do these energetic changes simply reflect intrinsic physiologic costs and efficiencies of the body?

This philosophic debate has importance for the way we think about interventions. If the entire nutrient balance system is fighting our therapeutic attempts at weight loss and avidly defending an overweight person's current body fatness, the situation is indeed bleak. If, on the other hand, all that is needed is a few adjustments in the system parameters (TEE, etc) to achieve a new settling point, the implication is quite different. Unfortunately, the answer remains controversial.

Some authors have questioned whether the assumption of short-term weight stability is accurate in the first place. Many animals exhibit cyclic weight changes in parallel with seasonal availability of food. Under these conditions, the nutrient balance system might well be programmed to alternate between fat anabolic and fat catabolic modes rather than defending a certain body fatness consistently. If humans are like this, then relatively small signals of food availability might induce physiologic programs of weight gain or loss that are currently unrecognized. This interesting perspective deserves further consideration.

New Guidelines (Dietary Reference Intakes)

In 2002, new recommendations were established by the Food and Nutrition Board of the National Academy of Science's Institute of Medicine. The revised guidelines, called the Dietary Reference Intakes (DRIs) for Macronutrients, were designed to promote diets that minimize chronic disease risk. The guidelines state that adults should obtain 45–65% of their daily energy intake (calories) from carbohydrates, 10–35% from protein, and 20–35% from fat. This latter recommendation for fat reflects the changing focus of national dietary advice from previous recommendations of less than 30% of energy from fat to the current recommendations that allow more flexibility in the mix of macronutrients.

The report recommends a minimum intake of carbohydrates (at least 130 g per day, an amount chosen to ensure enough glucose for brain metabolism). The report also advises that “added” sugars (those incorporated into foods and beverages during their production) should not represent more than 25% of total energy. The recommendation of an upper level of intake of sugars is based in part on evidence that people whose diets were high in added sugars had lower intakes of essential nutrients. Moreover, overconsumption of carbohydrates leads to elevations in body weight, though not through fat synthesis (as mentioned above). Rather, overconsumption of carbohydrate can lead to an accumulation of body weight by displacing stored fat that would usually be burned. The new DRIs recognize that overconsumption of carbohydrate, fat, or protein can lead to increased body weight. Diets that focus on delivering the primary source of energy from only one macronutrient therefore have limitations.

Lower-fat, higher-carbohydrate diets have been associated with elevations in blood triglycerides and reductions in HDL cholesterol, typically associated with lower LDL cholesterol levels. Although the atherogenicity of this lipid profile remains uncertain, there is suspicion that these diets will contribute to heart disease development. Furthermore, the consumption of sugars in a single meal may lead to return of hunger sooner than after consumption of a mixed meal. By contrast, the benefits of diets higher in complex carbohydrates may include increased micronutrient intake and an increased fiber intake, the latter of which is associated with several health benefits related to the gastrointestinal tract. Diets high in complex carbohydrates can also provide greater satiety. The vast majority of individuals in the National Weight Control Registry (a national survey of individuals who have lost 30 lb [13.6 kg] and have maintained that loss for at least 1 year) report that they consume low-fat, high-carbohydrate diets to help them manage their weight, in addition to participating in voluntary exercise activities of about 400 kcal/d (3000 kcal/wk; see below).

Diets higher in fat (> 30% energy) are associated with a higher incidence of obesity in many population


studies. Furthermore, unless careful attention is paid to the type of fat, higher-fat diets can contain large amounts of saturated and trans fatty acids (the latter found in partially hydrogenated vegetable oils). Because these fats raise blood cholesterol concentrations, the new DRIs continue to focus on the avoidance of saturated fat. On the other hand, some benefits of higher-fat diets have been observed when the primary fats consumed are monounsaturated. LDL cholesterol can be lowered by these diets, while HDL cholesterol is not lowered. For some individuals, higher-fat diets can also be more satisfying even when calories are restricted, thereby helping attempts at weight loss.

Higher-protein diets (> 25% of energy) can also have some undesirable effects. These diets can be virtually devoid of fiber and, with their low content of carbohydrate, are also low in water content, leading to constipation. In addition, higher-protein foods also tend to be high in saturated fat and cholesterol. Lastly, there is some concern that higher-protein diets might increase the filtration load on the kidney and promote the development of renal disease as well as increasing urinary calcium losses and contributing to osteoporosis. The benefits of diets higher in protein have only recently been investigated. It is likely that high-protein diets do result in an increased sense of satiety, thereby leading to a reduction in food intake and to weight loss, perhaps without induction of the hypertriglyceridemia associated with high-carbohydrate and low-fat diets.

The new DRIs recognize that different diet strategies work for different people and are therefore more flexible than previous recommendations regarding macronutrient intake for individuals trying to lose weight and those attempting to maintain their body weight. The new guidelines also focus on the importance of balancing diet with exercise, advocating at least an hour a day of moderate physical activity.



Caloric Intake Restriction

Dietary and behavioral therapy—in particular, the prescription of a hypocaloric diet—has been the mainstay of obesity treatment for the past 40 years. Caloric restriction alone has unfortunately been a generally ineffective approach. The use of very-low-calorie diets (800 kcal/d) in individuals with BMI > 30 can produce an initial weight loss of about 2 kg/wk, and an average total weight loss of 20 kg after 4 months. However, medical supervision is critical because the loss of electrolytes can be substantial. Intake of at least 1 g/kg of ideal body weight per day of protein of high biologic value is important in preserving lean body mass. After going off the very-low-calorie diet, maintaining the weight loss has proved to be difficult, particularly if patient education has not been incorporated into the therapy. Part of the recidivism may be attributable to the physiologic and behavioral adaptations that occur in response to energy deprivation, including reduced energy expenditure (as discussed above). The behavioral consequences of semistarvation may also play a role in recidivism—eg, fatigue, malaise, and loss of motivation. It simply may not be possible for most people to persist indefinitely in a state of perceived deprivation, particularly when stresses of life arise (as they always do). Programs which are more structured and combine diet (energy restriction to 1200–1800 kcal/d), exercise, and behavioral therapy may have greater success, though the efficacy of these programs is also debated. Programs of this type can produce weight losses in the range of 9–14 kg over 5–6 months, with attrition as low as 20–24%. Average weight losses of 9–10% have been the rule, with about 60–80% maintenance of weight loss after 1 year. Although most participants regain weight after leaving these programs, the amount regained after 1–2 years is frequently less than the original weight loss. The major problem relates to longer-term follow-up. Data are much scarcer for more than 3-year follow-up, but the data that exist suggest almost 100% relapse after 3–5 years. A recent study through the National Weight Control Registry has identified a large number of individuals who have successfully maintained weight loss. The most common features identified among these individuals is voluntary activity of about 3000 kcal per week (ie, 400–450 kcal of daily exercise) and a low-fat diet. It is hoped that the attributes of programs that help individuals lose weight and the characteristics of those who successfully keep weight off will help in the design of future programs incorporating energy restriction. It must be recognized, however, that net benefits to health or longevity have yet to be proved from weight loss diets and that the vast majority of participants regain the great majority of their weight when followed for 3–5 years.

Fat Absorption Inhibitors

Orlistat is a drug that reduces fat absorption by blocking pancreatic lipase in the intestine. In a clinical trial, patients lost 10% of initial body weight after 1 year of use on a 30% fat diet. The major side effect is “intestinal leakage” or fecal fat loss, which can be controlled in


patients who combine this drug with the consumption of a reduced fat (35% of energy) diet. The long-term efficacy and safety of fat absorption inhibitors has yet to be determined.

Altered Macronutrient Composition of Diet


Interestingly, the one dietary prescription that has most consistently resulted in long-term weight loss is to change the macronutrient composition (to a lower percentage of fat) rather than the total energy content of the diet. Individuals who switch to diets containing less than 25% fat for reasons other than weight loss (eg, to reduce dietary cholesterol and saturated fats or to reduce cancer risk) experience on average a 2–3 kg weight loss over the first 2-month period. For example, in a large study in nurses of the effects of dietary fat restriction on breast cancer risk, randomization into the low-fat arm (without directives to lose weight) resulted in an average > 4 kg weight loss. Many other high-carbohydrate, low-fat diet studies have confirmed these observations. Potential explanations for weight loss include reduced total energy intake due to the feeling of fullness related to the bulk of low-fat, high-fiber diets or due to their lack of palatability; or metabolic explanations related to satiety effects of body glycogen stores. Epidemiologic correlations between fat content of diet and adiposity in populations indirectly support—but do not prove—the importance of macronutrient composition of diet.


Conversely, diets that restrict carbohydrate have enjoyed cyclic popularity according to a variety of theories promoted in the popular weight loss literature. The “Zone diet” and “carbohydrate busters” are examples. The long-term efficacy of these dietary approaches has never been demonstrated. As discussed in detail above, the underlying principle of these diets runs against most of what we know about the physiology and regulation of energy and macronutrient balances. More importantly, numerous studies have shown the opposite effects—namely, that high-carbohydrate, low-fat diets result in weight loss, not weight gain, for most participants. Despite the lack of compelling physiologic or empirical evidence in favor of low-carbohydrate diets, new examples of this class of therapy are continually arising. Some people may achieve at least a temporary weight loss from this (or any) diet due to the motivational effects of participating in a program that is based on a “theory.” Also, restriction of carbohydrate and bulk may lessen postprandial lethargy and improve subjective sense of well-being and alertness in some individuals—or it may simply restrict the universe of food choices available to a person, which may promote compliance with a program of caloric restriction.


The surgical management of morbid obesity has continued as a treatment option since jejunoileal bypass was developed in the 1950s. A wide variety of techniques are used, but all impose a risk of serious complications. Intraoperative complications, risks of anesthesia, and postoperative problems (sepsis, bacterial contamination of a blind loop, development of fatty liver and cirrhosis) are all concerns. Both early and long-term weight reductions have generally been satisfactory. The risks associated with these procedures must be balanced in morbid obesity against the limited number of other treatments available and the frequently life-threatening conditions (described above) that may develop in individuals who have suffered with the condition for many years.


Appetite Suppressant Drugs

The most widely prescribed drugs for treatment of obesity have been appetite-suppressant agents. Several classes have been used, including serotonin agonists, sympathomimetics, and, recently, leptin. Studies using these agents have also revealed high recidivism rates after discontinuation of medications. Sibutramine, a serotonin uptake inhibitor, is currently the only drug in this class approved for long-term use. In a recent study, patients who had lost at least 6 kg over a 4-week period on a very-low-calorie diet were then randomized to sibutramine or placebo for 1 year. With continued dietary counseling in both groups, patients who took sibutramine lost significantly more weight (4.9 kg versus 0.45 kg, in the drug-treated and control groups, respectively). At month 12, about 75% of subjects in the sibutramine group maintained at least the same degree of the weight loss achieved with a very-low-calorie diet, compared with 42% in the placebo group.

Because most clinicians are wary of indefinite use of drug therapy in this presumably lifelong disorder, use of anorexiants fell out of favor many years ago in the medical mainstream. A change in philosophy seemed to be evolving in the mid 1990s with the long-term administration of a combination anorexigenic formulation (fenfluramine and phentermine, or “Fen-Phen”). A study of 81 patients treated with this regimen for several years showed long-term efficacy with maintenance of weight loss. Subsequently, the well-publicized reports


of a serious adverse effect (heart valve thickening) rapidly led to the withdrawal of fenfluramine from the United States market. Skepticism about use of anorexiants has since resurfaced.

Thermogenic Agents

Thermogenic agents represent a theoretically attractive approach, particularly if low REE predicts subsequent development of obesity and if weight gain itself serves as a way of increasing TEE and fat oxidation to match intake. Several classes of thermogenic agents have been tested.


Thyroid hormone is calorigenic. Indeed, thyroid status used to be assessed (before the development of specific blood assays) longitudinally by following REE. In addition, short-term energy restriction results in striking reductions in serum triiodothyronine (T3) and thyroid stimulating hormone (TSH)—often called the “euthyroid sick syndrome”—accompanied by reduction in REE of up to 15–20%. Because weight loss tends to slow after the initial days or weeks of energy restriction, preventing this reduced REE by administration of T3 represented an obvious strategy. Many clinicians and investigators have given T3 or T4 as an adjunct to hypocaloric diet therapy. The results are unequivocal and, unfortunately, negative: restoration of serum T3 to normal or supraphysiologic levels during hypocaloric diets results in no extra loss of fat but greater losses of lean body mass and a more negative nitrogen balance. Thus, the response of the hypothalamic-pituitary-thyroid-tissue axis to energy deficiency is an essential part of the nitrogen preserving response to starvation. Extra weight loss may impress clinicians or patients during coadministration of thyroid hormone, but this in fact is exactly contrary to the goals of obesity therapy. Thyroid replacement therapy is contraindicated in combination with energy-restricted diets.


Adipose tissue expresses an atypical β-adrenergic receptor (ie, in addition to the usual β1 and β2 receptors), which has been called the β3-adrenergic receptor. The tissue distribution of the β3-adrenergic receptor in humans remains uncertain, but most expression appears to be in adipocytes. Expression is particularly prominent in brown adipose tissue in animals such as rodents that maintain large stores of brown fat, but β3 receptors are also expressed in white adipose tissue. Interest in the β3 receptor has focused on its potential role in adrenergic stimulation of lipolysis (in white adipose tissue) and thermogenesis (in brown adipose tissue). Because β3-selective adrenergic receptor agonists exhibit minimal cross-reactivity with β1 and β2 receptors, the β3 pathway has received considerable attention as a possible therapeutic candidate. Several pharmaceutical companies have developed selective β3-receptor agonists toward this end. This approach has been further motivated by epidemiologic associations between mutations in the β3 receptor gene and the development of obesity and diabetes mellitus (eg, in Pima Indians and in populations in Finland and France).

Despite these promising features, there are currently no approved β3-adrenergic receptor agonists on the market for the treatment of obesity. Clinical trials performed to date have been disappointing. Side effects have been observed in some studies, including tremor and changes in heart rate or blood pressure. There are few reports of clinical trials with β3 agonists over the past 5–6 years, and no large-scale randomized studies have been conducted.


The sympathetic nervous system is involved in all aspects of the macronutrient economy (see above): food intake (suppressive effect), energy expenditure (increase), and nutrient partitioning (favoring of lean body mass [LBM] over fat stores). Although adrenergic agonists such as amphetamines remain in use for weight loss both over the counter and by prescription, these agents are properly looked at with extreme caution by responsible clinicians. The psychoactive side-effects are significant and not dissociable from peripheral effects; the addictive potential and possibility of abuse cannot be ignored; weight recidivism occurs after cessation of therapy; and adverse medical effects include worsening of hypertension, diabetes, and coronary insufficiency.

It is nevertheless of interest that several of the most common addictive drugs of abuse in the Western world are sympathomimetic agents and, indeed, are associated with lower body weight, lower body fat, and increased REE. By far the most important of these in public health terms is nicotine, the active agent in cigarette smoke. Cigarette smoking is associated with lower body weight (3–5 kg) despite similar or higher food intake and higher REE compared with matched nonsmokers. Cessation of smoking leads to an average 3 kg weight gain, and approximately 10–13% of smokers who quit gain over 13 kg. Metabolic ward studies have confirmed the thermogenic effect of cigarette smoking (about 10–15% increased REE), quantitatively consistent with the relationship described between energy expenditure and body inTable 20-3. Pharmacologic studies have shown that this thermogenic effect operates via nicotine, can be reproduced by administration of nicotine, and depends upon release of adrenal catecholamines (epinephrine) into the circulation. The evidence


that nicotine increases REE and effectively lowers body weight in cigarette smokers is convincing.

Though of great interest from the perspective of the macronutrient economy, these actions of nicotine and cigarette smoking have devastating public health consequences. Many smokers maintain the habit because of fear of weight gain after cessation. This is particularly so for young women, nowadays the fastest-growing group of cigarette smokers. The marked increases in lung cancer already reported in women over the past generation will undoubtedly accelerate, along with other correlates of cigarette smoking (eg, coronary artery disease), as an unintended consequence of this widely used over-the-counter weight-managing strategy. Indeed, the strongest health argument against the contemporary obsession about slimness, particularly in women, may well be the high rates of cigarette smoking that it promotes.

Another addictive drug with sympathomimetic actions and profound weight-reducing actions is cocaine. Inhaled crack cocaine has unmistakable weight reducing effects.


Administration of recombinant leptin (the ob gene product) increases TEE and activates the sympathetic nervous system in ob/ob mice—in addition to reducing food intake. The net effect of leptin administration is reduction of adiposity in this leptin-deficient animal model. Clinical trials with recombinant leptin in humans have been equivocal. Administration of extremely high doses of leptin subcutaneously in combination with dietary energy restriction resulted in 7.3 kg lost after 6 months of treatment, compared with 1.4 kg weight loss in the placebo plus energy restriction group. Weight loss was greater than in the placebo group only for the highest dose of leptin (0.3 mg/kg/d), however, which resulted in serum leptin concentrations of 500–600 ng/mL—compared with baseline values in the range of 15–20 ng/mL. Weight loss in the group treated with 0.1 mg leptin/kg/d was not significantly different from placebo despite serum leptin levels in the 200–250 ng/mL range. The long-term metabolic and general health consequences of activation of the sympathetic nervous system by massively supraphysiologic leptin levels remains unknown. Experience with other sympathetic nervous system activators (discussed above) provides ample reason for caution.

The most useful setting for leptin therapy may not prove to be conditions of idiopathic obesity but states of frank leptin deficiency due either to inadequate fat stores or genetic defects in leptin production. Lipoatrophic states such as occur genetically or in HIV/AIDS may benefit from leptin administration by improving insulin sensitivity and lipoprotein concentrations. Extremely rare individuals with congenital leptin deficiency (who present clinically with massive obesity and hyperphagia in early childhood) also benefit from administration of recombinant leptin.


Though strictly speaking not a pharmacologic intervention, exercise may be the most effective prescription for weight loss. Meta-analysis has evaluated the effects of exercise training on FFM preservation during diet-induced weight loss. In individuals whose weight loss regimen included diet only, 28% of the weight lost was made up of FFM, while for those who exercised in addition to dieting, only 13% of the weight lost was made up of FFM. When combined with reduced energy intake, exercise increases the amount of weight loss above that achieved by diet alone—in addition to preserving muscle mass. Most importantly, exercise increases the likelihood of sustaining the weight loss over a longer period of time. For these reasons, programs aimed at reducing body weight must include an exercise component.

An interesting but relatively poorly understood area of energy expenditure relates to involuntary activity, or “fidgeting.”Differences in involuntary motor activity clearly exist within populations and could contribute to differences in energy expenditure. Some studies attempting to add up all the recognized components of total energy expenditure (eg, REE, thermic effect of food, measurable activity, thermic effect of exercise) have reported a consistent shortfall compared with measured 24-hour total energy expenditure. This shortfall also differs between lean and obese groups. This unmeasured energy expenditure has been attributed to “fidgeting” by some investigators. It has been difficult to quantify the energy costs of involuntary activity directly, however. The factors potentially controlling involuntary activity (eg, behavioral, hormonal, neuromuscular) are also uncertain. Nevertheless, this area represents a potentially important and generally overlooked aspect of whole-body energetics.

Nutrient Partitioning Agents

Perhaps the least-appreciated treatment strategy for obesity is to change body composition (increase lean tissue) without explicitly attempting to induce negative energy or fat balance. Though counterintuitive on the surface, this strategy has a solid rational basis and may prove an effective general approach.

The rationale is that the major determinant of REE is lean body mass (see above). Accordingly, anything that increases lean tissue will increase REE—by at least 50 kcal/kg/d. Muscle anabolic agents therefore act indirectly as whole body thermogenic agents. Consistent


with this model are reports of extremely high TEE in bodybuilders (eg, > 4000 kcal/d) out of proportion to the calculated energy expended during their weight lifting activities per se.

Several agents are known to be anabolic for lean tissue.


Androgen administration to hypogonadal adult men, prepubertal boys, or women results in nitrogen retention and gains in lean tissue, including muscle. Energy expenditure increases in proportion to lean tissue accrual. Effects on body fat stores are inconsistent; either reductions or increases (with android or abdominal distribution) may be observed depending upon energy intake.


Growth hormone administration at pharmacologic doses to adults increases LBM, stimulates lipolysis and fat oxidation, and reduces body fat stores. These effects have been reported for recombinant growth hormone in a variety of clinical settings, including wasting disorders such as AIDS and in the elderly. REE increases in proportion to FFM. Some investigators in Europe have reported that recombinant-growth hormone reduces abdominal obesity when given for 6–12 months and improves certain metabolic correlates of visceral adiposity. However, the extremely high cost of recombinant-growth hormone, along with concerns that it might directly worsen insulin resistance and hypertriglyceridemia, have to date held back widespread use of this agent for the treatment of obesity.


Aerobic, energy-utilizing exercise is usually recommended for therapeutic weight loss. As discussed above, resistance exercise programs (weight lifting) aimed at accrual of muscle (LBM) may also be useful for the indirect stimulation of thermogenesis. Prescription of supervised progressive resistance exercise in AIDS patients, for example, results in considerable gains in LBM with proportionate increases in REE and some loss (2 kg) of body fat. Combination of resistance exercise with modestly supraphysiologic androgen therapy further increases LBM gains and REE but worsens HDL cholesterol and may have other adverse effects.

Alternative Therapies

Some therapies reported in the lay press have achieved widespread use and therefore deserve comment.


Chromium is the second most widely used mineral supplement in the United States (after calcium). The claims for chromium have included “fat burning,” “muscle building,” and antihyperglycemic and hypolipidemic actions. Although some evidence exists suggesting a beneficial effect of chromium picolinate on hyperglycemia in patients with type 2 diabetes mellitus, properly controlled trials of chromium administration have failed to show any reductions in body fat or body weight or any gains in LBM. Nor does there exist any solid biochemical or endocrinologic rationale why chromium might promote fat loss or fat oxidation. Use of chromium supplements to alter body composition currently has no evidentiary basis.


The use of a Chinese “herbal fen/phen” has come into practice. These products, which contain ephedra, are marketed as“natural” tonics for weight loss, muscle building, and energy enhancement. Ephedrine alkaloids are powerful stimulants of the heart and nervous system and may be even more dangerous because the herbals can be impure. Medical complications of herbal ephedrine use are well documented.



Diet therapy, consisting primarily of weight loss prescription (hypocaloric diet) plus counting of “carbohydrate exchanges,” is widely considered to be first-line therapy for type 2 diabetes mellitus. Remarkably, most clinical and experimental data suggest that neither weight loss itself nor changes in body composition are the agents by which energy-restricted diets lower blood glucose in this disease. Short-term severe caloric restriction (to 600–800 kcal/d for 2–5 days) effectively reduces hyperglycemia (as well as hyperlipidemia) in obese type 2 diabetic patients, with only minor changes in body composition. Most of the response to weight loss diets in type 2 diabetes occurs within the first 2.3 kg; indeed, if glycemia has not improved within 2.3–4.6 kg, there is unlikely to be a response to further weight loss (a useful fact when managing diabetic patients). The short-term response to severe caloric restriction is initially mediated by reduced hepatic glucose production, which appears to be due entirely to a reduced glycogenolytic contribution. Thus, most of the beneficial effects on glycemia of hypocaloric diets appear to be attributable to a starvation effect (ie, liver glycogen depletion) rather than to a reduction in body




fat stores. The clinical implications of these remarkable physiologic observations have not been fully explored. Some clinicians have considered alternative diet strategies (eg, fasting 1 day per week), but published studies showing efficacy of such an approach are lacking.

Table 20-4. Pathways to fat accrual as guide to therapeutic interventions.




Altered energy expenditure


   Low TEE


      Sedentary lifestyle

↑ Voluntary activity (exercise, daily life activities)

Best correlate of weight maintenance; only way to “eat again”

      Low FFM (metabolizing tissue)

↑ Lean tissue anabolism (resistance exercise, hormones)

Not systematically tested; improved fit-ness as an independent health benefit?

      Low sympathetic nervous system  activity

Sympathomimetic drugs (β3-adrenergic receptor agonists, leptin, nicotine?)

Basis of most therapeutic agents (prescription and OTC, including nicotine, amphetamines)

      Thyroid or leptin deficiency

Replacement therapy

Rare (but effective if truly replacement therapy)

      Low “fidgeting” factor


Unexplored etiologic factor

   Impaired fat oxidation


      Dietary carbohydrate surplus

Carbohydrate restriction (especially energy-dense sweets)

Two-edged sword (if increased fat intake takes its place)

      Primary high respiratory quotient  (impaired fat oxidization)

Prolongation of overnight fast with or without exercise to burn fat

Strong predictor, but an unexplored therapeutic option

      Endogenous carbohydrate  availability (↑HGP)

?Inhibitors of HGP (metformin)

Effective in certain conditions (type 2 diabetes mellitus, HIV-related lipo-dystrophy)

Altered nutrient partitioning


   Impaired muscle anabolism


      Lack of muscle contraction (exercise)

Resistance exercise

Effective (↑ FFM, ↑ REE, ↓ fat)

      Androgen deficiency

Replacement therapy

Effective (↑ FFM, ↑ REE, ↓ fat)

      Glucocorticoid excess (endogenous,   exogenous)


Effective (improves FFM:fat)

      Altered adipocyte metabolism  Impaired fat mobilization (lipolysis)

?Lipolytic agents (rGH, β3, nicotine?)

Unclear role or effect on fuel selection

      Increased de novo lipogenesis

Hydroxycitrate (inhibitor of lipogenesis)

Doubtful in theory (insignificant role of de novo lipogenesis) and practice (toxicities)

Excess energy intake


   Disordered appetite system


      Primary alteration in satiety or hunger

Diet therapy and training?; change in macronutrient selection

Generally ineffective (except if it reduces dietary fat content)

      Depression or anxiety

Psychologic or pharmacologic therapy

Some antidepressant drugs reduce weight, some increase weight

      Sensitivity to food cues

Diet training; less exposure (eg, to tele-vision)

Efficacy unproven

   Dietary factors


      High-fat, low-satiety diets

Reduce fat, increase bulk, etc

Effective in population studies (reduces weight)

      High sweet, energy-dense diets

Reduce simple sugars

?Efficacy (often replaced with fat)

TEE = Total energy expenditure.
HGP = Hepatic glucose production.
β3 = β3-adrenergic receptor agonists.
rGH = Recombinant growth hormone.
FFM = Fat-free mass.
REE = Resting energy expenditure.
OTC = Over-the-counter.

Similarly, glycemic improvements in response to exercise programs have been reported without weight loss in type 2 diabetic patients. The implication here, as well, is that body weight and body fat are not the true outcome measures or treatment goals of diet therapy for obese type 2 diabetic patients.


Low-fat diets are a cornerstone of treatment for most forms of hyperlipidemia. A tension may sometimes arise, however, between optimal prescriptions for reduction of body weight and serum lipids. Low-fat, high-carbohydrate diets tend to promote weight loss (see above), but they also exacerbate hypertriglyceridemia and low-HDL syndromes. Substitution of monounsaturated fats (eg, olive oil) for carbohydrate has been recommended as a way of preventing hypertriglyceridemia on low saturated fat diets but may interfere with the weight-reducing effects of these diets.


Obesity and insulin resistance are typically present in the chronic estrus or polycystic ovarian syndrome. One hypothesis concerning the connection is that hyperinsulinemia stimulates thecal cell proliferation in the ovary, which results in constitutive steroid hormone synthesis and a loss of menstrual periodicity. Aromatization of adrenal androgens in adipose tissue also represents a postulated mechanism of altered hormone production. Treatments that improve insulin sensitivity—particularly metformin, which acts primarily on the liver to reduce endogenous glucose production and thereby lessens hepatic insulin resistance; or thiazolidinediones, which act on adipose tissue and muscle to stimulate peripheral glucose uptake—have now been shown to improve signs and symptoms of polycystic ovarian syndrome (see Chapter 13).

Table 20-5. Some unanswered questions regarding body composition and health.

1. Reasons for increasing prevalence of overweight and obesity?
      Mostly lower TEE? Role of diet composition or total energy intake?

2. Metabolic mechanisms of adverse health consequences of obesity?
      Fatness versus fitness? Direct consequence of adipose tissue or its products?

3. Failure to maintain weight loss long-term?
      True “set-point”? Role of leptin? Efficacy of exercise?

4. Role of long-term drug therapy for weight control?
      Risks and toxicities vs benefits? Long-term efficacy?

5. Mediators of beneficial response to diet in diabetes, hypertension, hyperlipidemia, etc?
      Reduced adipose mass versus negative energy balance? Other factors, such as leptin?

6. Role of adipocyte biology in systemic macronutrient balances?
      Role of proliferation, differentiation, and death of adipocytes in adults? Fat cell size versus number? Intrinsic lipolysis and  lipogenesis?

7. Health risks associated with societal obsession with thinness?
      Anorexia nervosa, impaired growth in children, and other syndromes of underweight? Reliance on cigarette smoking and  other unhealthy behaviors? Failure to emphasize fitness? Chronic symptoms of semi-starvation? Adverse psychologic and  self-image effects?

TEE = Total energy expenditure.


A close relationship exists between obesity and hypertension, but the mechanisms are unknown. Prescription of a weight loss diet is as effective as prescription of salt-restriction for the long-term management of essential hypertension. More specific guidelines for dietary control of hypertension are needed.


Weight loss (or weight gain) can alter the cholesterol saturation and lithogenicity of bile and thereby provoke gallstone formation. Gallstones are common side effects of weight loss programs. Clinicians should be alert to the possibility of biliary tract disease in patients undergoing therapeutic weight loss.


There are several conditions for which body weight has strictly mechanical—as opposed to metabolic—consequences


(osteoarthritis, low back pain, pulmonary insufficiency, etc). In these states, it is the excess weight itself that must be the treatment goal; metabolic correlates cannot substitute.


Systematic consideration of the number of pathways leading to obesity (ie, factors predisposing to body fat accrual; Figure 20-2) represents a useful exercise. Table 20-4 provides a summary of interventions tried—and their efficacy—in response to these different causes. Under the category of reduced TEE, for example, we see that increased voluntary activity has proved effective for weight loss, as have sympathomimetic agents, whereas “fidgeting” remains an unexplored treatment approach. Perusal of the other major categories similarly reveals areas of success, failure, and unexplored approaches.

It is important to recognize areas of uncertainty or ignorance regarding the control of body composition and the treatment of obesity. A number of fundamental questions remain unanswered (Table 20-5.


Bouchard C, Perusse L: Genetics of obesity. Annu Rev Nutr 1993;13:337.

Brownell KD, Kramer FM: Behavioral management of obesity. Med Clin North Am 1989;73:185.

Christiansen MP et al: Effect of dietary energy restriction on glucose production and substrate utilization in type 2 diabetes. Diabetes 2000;49:1691.

Flatt JP: Dietary fat, carbohydrate balance, and weight maintenance: effects of exercise. Am J Clin Nutr 1987;45(1 Suppl): 296.

Hellerstein MK: De novo lipogenesis in humans: metabolic and regulatory aspects. Eur J Clin Nutr 1999;53(Suppl 1):S53.

Klem ML et al: A descriptive study of individuals successful at long-term maintenance of substantial weight loss. Am J Clin Nutr 1997;66:239.

Leibel RLM et al: Changes in energy expenditure resulting from altered body weight. N Engl J Med 1995;332:621.

Miller WC: How effective are traditional dietary and exercise interventions for weight loss? Med Sci Sports Exerc 1999;31: 1129.

Mokdad AH et al: The spread of the obesity epidemic in the United States, 1991–1998. JAMA 1999;282:1519.

Must AJ et al: The disease burden associated with overweight and obesity. JAMA 1999;282:1523.

Parks EJ, Hellerstein MK: Carbohydrate-induced hypertriacylglycerolemia: historical perspective and review of biological mechanisms. Am J Clin Nutr 2000;71:412.

Pi-Sunyer X et al: Executive summary of the clinical guidelines on the identification, evaluation, and treatment of overweight and obesity in adults. Arch Intern Med 1998;71;412.

Ravussin E, Gautier JF: Metabolic predictors of weight gain. Int J Obes Relat Metab Disord 1999;23(Suppl 1):37.

Schwarz J-M et al: Short-term alterations in carbohydrate energy intake in humans. Striking effects on hepatic glucose production, de novo lipogenesis, lipolysis and whole-body fuel selection. J Clin Invest 1995;96:2735.

Wei M et al: Relationship between low cardiorespiratory fitness and mortality in normal-weight, overweight, and obese men. JAMA 1999;282:1547.

Weintraub M et al: A double-blind clinical trial in weight control. Use of fenfluramine and phentermine alone and in combination. Arch Intern Med 1984;144:1143.

Wing RR et al: Caloric restriction per se is a significant factor in improvements in glycemic control and insulin sensitivity during weight loss in obese NIDDM patients. Diabetes Care 1994;17:30.