Change in the age-adjusted prevalence of overweight and obesity in men (top) and women (bottom) aged 20 to 74 years from 1960-1994. BMI indicates body mass index. Adapted from National Institutes of Health.13
Weighted prevalence of type 2 diabetes mellitus according to body mass index (BMI) in adults aged 20 years or older. From the Third National Health and Nutrition Examination Survey, 1988-1994. (Data from Maureen Harris, PhD, written communication, March 1, 1999.)
Age-adjusted prevalence of hypertension (mean systolic blood pressure >140 mm Hg, mean diastolic blood pressure >90 mm Hg, or currently using antihypertensive medication) by body mass index (BMI) and sex. From the Third National Health and Nutrition Examination Survey, 1988-1994. Adapted from National Institutes of Health.13
National Task Force on the Prevention and Treatment of Obesity. Overweight, Obesity, and Health Risk. Arch Intern Med. 2000;160(7):898-904. doi:10.1001/archinte.160.7.898
More than half of adult Americans are overweight or obese, and public health recommendations call for weight loss in those who are overweight with associated medical conditions or who are obese. However, some controversy exists in the lay press and in the medical literature about the health risks of obesity. We review briefly the large body of evidence indicating that higher levels of body weight and body fat are associated with an increased risk for the development of numerous adverse health consequences. Efforts to prevent further weight gain in adults at risk for overweight and obesity are essential. For those whose present or future health is at risk because of their obesity and who are motivated to make lifestyle changes, a recommendation for weight loss is appropriate.
The health consequences of overweight and obesity have been the subject of controversy in the lay press1 and in the professional literature.2,3 For example, in January 1998, an editorial in the New England Journal of Medicine stated, "Unfortunately, the data linking overweight and death, as well as the data showing the beneficial effects of weight loss, are limited, fragmentary, and often ambiguous."4 Other authors have suggested that the health risks of overweight are overstated, and that, for most individuals, the risks of weight loss outweigh the benefits.5,6 In addition, there have been some widely cited data published that call into question the belief that obesity per se is harmful, and hypothesize that obesity is just a marker for unhealthful behaviors such as a sedentary lifestyle.7,8 Some of the points raised by critics have merit. Obesity is associated with sedentary behaviors and with less healthful diets, which could exert independent effects on health. There is certainly need for additional research directly pertaining to the effects of obesity on health and to the health effects of weight loss in those already obese. However, the data linking overweight and obesity to adverse health outcomes are well established and incontrovertible. Our purpose was to review briefly the evidence linking overweight and obesity with morbidity and mortality.
One area of controversy involves the definitions of overweight and obesity. Whereas obesity refers to an excess of body fat,9 most data regarding the effects of obesity on health rely on measurement of body weight. Historically, weight-for-height tables, such as the Metropolitan Life Insurance Company tables,10 were used to define the normal weight range. However, such tables have major limitations, including use of an unvalidated estimate of frame size, reliance on primarily white populations, and derivation of the tables from mortality data, which may not accurately reflect obesity-related comorbidities. More recently, relative body weight has been defined by the body mass index (BMI), which is calculated by dividing the weight (in kilograms) by the square of the height (in meters).
Body mass index is believed to be a better measure of adiposity than weight-for-height tables, although neither directly measures body fat level.11 The correlation between BMI and body fatness is age and sex dependent but appears to be independent of ethnicity.12 A limitation of BMI is that it does not incorporate body fat distribution, which is an independent predictor of health risk.13 Therefore, use of BMI and body fat distribution are useful to evaluate health risk due to overweight and obesity. In addition, because body fat is not directly measured, some muscular persons (such as athletes) may be misclassified as obese using BMI alone. The third edition of the Dietary Guidelines for Americans14 published in 1990, used an age-adjusted BMI cutoff for overweight, with a lower limit of 27 in adults aged 35 years or older. This adjustment was based on life insurance data showing that the BMI associated with minimum mortality rate increased with increasing age.15 Subsequent data showed that for men, a BMI of 24 is associated with minimum mortality, regardless of age, although this relationship is not observed for women, and in both sexes, weight gain during adult life is associated with increased risk of heart disease and death.16 This risk is modified by race and ethnicity, as well as sex and age. Based on these and other data, the fourth edition of the Dietary Guidelines for Americans, issued in 1995,17 deleted the age adjustment, recommending that adult men and women maintain a healthy weight corresponding to a BMI from 19 to 25. The National Institutes of Health (NIH) Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults,13 published in 1998, operationally define overweight as a BMI of 25 to 29.9 and obesity as a BMI of at least 30. Defining overweight as a BMI of at least 25 is consistent with recommendations of the World Health Organization18 and most other countries. Using the definition of obesity as a BMI of at least 30 places approximately 25% of US adult women and 20% of US adult men in this category.19 The prevalence of obesity is even higher in some minority populations. For example, 37% of non-Hispanic black women and 34% of Mexican American women are obese.19 As shown in Figure 1, the percentage of individuals in the overweight but not obese category has remained relatively stable during the past 30 years. However, the prevalence of obesity has increased more than 50% within the past 10 to 15 years (from 14.5% to 22.5%).19 This most likely reflects a shift toward higher weight in all categories, with a consequent reduction in the percentage of persons meeting criteria for normal weight.
A major predictor of health risk associated with obesity is body fat distribution. Body fat may be preferentially located in the abdomen (android obesity pattern) or surrounding the hips and thighs (gynoid obesity pattern). Numerous studies have demonstrated that the android pattern often reflects an accumulation of fat surrounding the abdominal visceral organs and is associated with a variety of metabolic derangements, including dyslipidemia, hypertension, and glucose intolerance.20 Thus, even at the same level of overweight, the individual with a greater amount of visceral fat is more likely to have, or to have develop, many of the serious health conditions associated with obesity. Sex21 and ethnicity22,23 have an impact on body fat distribution and adipose tissue metabolism. Although methods accurately assessing visceral fat (such as computerized tomography and magnetic resonance imaging) are expensive and not suitable for routine clinical use, an easily obtained surrogate measure is waist circumference. Waist circumference of at least 88 cm (35 inches) in women or 102 cm (40 inches) in men has been associated with increased health risk.13 It should be stressed, however, that although these cutoffs provide a reasonable estimate of risk within populations, there is significant interindividual variability in the amount of visceral fat at a given waist circumference.
For most of the following conditions, obesity and overweight are used interchangeably, but are further defined by BMI or percentage overweight when such information is provided. In general, weight, rather than body fat, has been measured. The differing definitions used among studies may account, in part, for apparent discrepancies in assessing degree of health risk or in prevalence figures.
An estimated 15.6 million adults in the United States, 8% of men and women aged 20 years or older, have diabetes.24 Type 2 diabetes mellitus, previously known as non–insulin-dependent diabetes mellitus or adult-onset diabetes, accounts for about 90% to 95% of all diagnosed cases of diabetes.25 Data from the Third National Health and Nutrition Examination Survey (NHANES III) indicate that among those in whom type 2 diabetes mellitus has been diagnosed, 67% have a BMI of at least 27, and 46% have a BMI of at least 30 (Maureen Harris, PhD, written communication, March 1, 1999). Risk for development of diabetes increases dramatically as degree of overweight increases (Figure 2), and several studies note increasing risk at relatively low levels of BMI, as well as with even modest amounts of weight gain after 18 years of age.26- 28 Visceral adiposity increases the risk for hyperinsulinemia and glucose intolerance at a given BMI.29,30 The dramatic increase in obesity during the past decade has been accompanied by a 25% increase in the prevalence of type 2 diabetes mellitus.24 Thus, overweight and obesity, especially in those with a more central body fat distribution, have a clear and easily demonstrable relationship to the development of type 2 diabetes mellitus.
Obesity has an association with coronary heart disease (CHD), presumably through its impact on risk factors, including hypertension, dyslipidemia, impaired glucose tolerance, and type 2 diabetes mellitus.31 As can be seen in Figure 3, the increase in prevalence of hypertension begins at relatively low levels of overweight.
There is also evidence from long-term observational studies that overweight is a predictor of cardiovascular atherosclerosis independent of its effects on traditional risk factors.32,33 This increase in relative risk occurs at levels of overweight frequently considered clinically insignificant by some (eg, 72% increased risk for fatal or nonfatal CHD in middle-aged men with a BMI of 25-29 compared with men having a BMI of <23).34 The relationship between degree of overweight and the development of CHD may be modified by age, sex, body fat distribution, degree of fitness, and ethnicity.31,35,36 Obesity also contributes to the development of congestive heart failure through its relationship to systemic hypertension, and, in normotensive and hypertensive obese patients, through increases in stroke volume and cardiac output along with diastolic dysfunction.31 In patients with severe obesity, dilated cardiomyopathy may lead to sudden death through predisposition to arrhythmias.31 Right heart dysfunction can also occur, secondary to obstructive sleep apnea or obesity hypoventilation syndrome with secondary pulmonary hypertension or as a consequence of left heart failure.37 Citing the well-established contribution of obesity to the development of CHD, the American Heart Association recently added obesity to its list of major risk factors for CHD.38
Sleep apnea is defined as a cessation of airflow during sleep for at least 10 seconds, which can be central (no spontaneous respiratory effort), obstructive (continued respiratory effort), or mixed (central and obstructive components present). Sleep apnea can lead to severe arterial hypoxemia, recurrent arousals from sleep, increased sympathetic tone, pulmonary and systemic hypertension, and cardiac arrhythmias.13 Obesity is a major correlate of sleep apnea in men and women, with those having a BMI of at least 30 at greatest risk.13,39,40 Obesity is believed to change upper-airway geometry through loading of the wall of the pharynx or through increased deposition of periluminal fat.39 There are fairly well-established data indicating that symptoms of sleep apnea improve with weight loss.13,39
In addition to predisposing to obstructive sleep apnea, obesity can lead to pulmonary compromise in a number of ways, including decreases in respiratory compliance due to mechanical factors, such as increased weight on the thoracic cage and abdomen, as well as changes in lung compliance.39 Severe obesity can lead to the obesity-hypoventilation syndrome, an alteration in respiratory control, with decreased responsiveness to carbon dioxide and hypoxemia, accompanied by alterations in respiratory muscle function.39 Central distribution of body fat has been identified as a predisposing factor for obesity-hypoventilation syndrome and sleep apnea.41,42
Overweight may increase the risk for ischemic, but not hemorrhagic, stroke. In the Nurses' Health Study,43 ischemic stroke risk was 75% higher in women with a BMI of greater than 27 and 137% higher in those with a BMI of greater than 32, compared with women who had a BMI of less than 21.
The risk for gallstones and cholecystectomy increases with increasing body weight.44 Data from the NHANES III indicate that the prevalence of gallbladder disease increased from 9.4% in the lowest quarter of the BMI distribution to 25.5% in the highest quarter among women, and from 4.6% in the lowest quarter of the BMI distribution to 10.6% in the highest quarter among men.45
Nonalcoholic steatohepatitis (NASH) is a pathologic condition with all of the features of fatty infiltration, inflammation, and fibrosis seen in alcoholic liver disease, but occurring in the absence of excessive alcohol intake.46 It is more common in persons with obesity and with type 2 diabetes mellitus.46 Most series report that 69% to 100% of affected patients are overweight, with most patients being 10% to 40% overweight.46 The disease has also been described in severely obese patients who have undergone rapid weight loss after jejunoileal bypass. The natural history is uncertain, but in a few patients, NASH progresses to liver failure and need for transplantation. Although the pathogenesis of NASH is unknown, increases in visceral adiposity, free fatty acids, and hyperinsulinemia that occur with obesity are postulated to play a role.46 Resolution of NASH has been reported with sustained weight loss.47
Increasing body weight increases the risk for osteoarthritis, presumably through increased stresses on weight-bearing joints,48,49 although dietary and metabolic factors have also been implicated.49,50 Age, female sex, and BMI are independent predictors of disabling knee osteoarthritis,51 and persons with a BMI of at least 30 have a markedly increased risk for knee osteoarthritis compared with those with more modest overweight (BMI, 25-29.9).52 Increased body weight has also been shown in cross-sectional and longitudinal studies to be a risk factor for hyperuricemia and gout.50 However, increased body weight appears to offer protection against the development of osteoporosis.53
In women, increased body weight is associated with menstrual irregularities, presumably through insulin-mediated effects on the ovarian stroma.54 Infertility occurs at both ends of the weight spectrum, and overweight has been found to be a risk factor for primary ovulatory infertility.55 Maternal obesity is a significant risk factor for the development of gestational diabetes mellitus, and maternal obesity has also been found to be associated with an increased risk of neural tube defects.56,57
Increasing body weight is associated with increased risk for certain forms of cancer, including colon cancer,58,59 endometrial cancer,60 and postmenopausal (but not premenopausal) breast cancer.61 For colon and breast cancer, some controversy remains regarding competing etiologic roles of high-fat diets vs obesity.62 An estimated 34% to 56% of cases of endometrial cancer are attributable to increased body weight (BMI, >29).63 Almost half of breast cancer cases among postmenopausal women occur in those with a BMI of greater than 29.63 In the Nurses' Health Study, women gaining more than 9 kg (20 pounds) from 18 years of age to midlife doubled their risk for breast cancer, compared with women who maintained stable weight.61
Overweight and obesity are associated with numerous additional health conditions, including carpal tunnel syndrome,64 venous insufficiency and deep vein thrombosis,65,66 and poor wound healing.67
Most observational studies have shown a U- or J-shaped relationship between BMI and mortality, with individuals at very low and very high weights at increased risk. Such a relationship is found even after attempts are made to adjust for confounding factors, such as smoking or preexisting illness.68 However, no such increased risk for death at the lowest body weights was found among nonsmoking Seventh Day Adventist men, in whom leanness was likely attributed to diet and activity rather than illness,69 although questions have been raised about this analysis.70 In addition, 2 recent reports found that the J-shaped relationship between low BMI and increased risk for death may result from the detrimental effects of lower lean body mass, rather than low body weight per se,71 and that after controlling for differences in fat loss, weight loss was associated with an increased mortality rate, whereas after controlling for differences in weight loss, fat loss was associated with a decreased mortality rate.72
Most studies clearly show an increase in mortality rate associated with BMI of at least 30.13 Individuals with a BMI of at least 30 have a 50% to 100% increased risk for death due to all causes, compared with individuals at a BMI of 20 to 25,13,68,73 with most of the increase due to cardiovascular causes. Although the relative increase in mortality rate attributable to obesity declines with age, an increased risk for death with higher BMI is seen even among individuals aged 65 to 74 years.74 Moreover, because mortality rate is already higher among older persons, even a small relative increase can translate into a very large absolute increase.
Fitness level may also have an impact on the BMI-mortality relationship. A large, observational study has found that unfit men of normal weight had a higher all-cause mortality than men who were fit but overweight (BMI, >27.8).35 That study found that men who were unfit and overweight had the highest cerebrovascular disease (CVD)–related mortality, whereas fit men of normal weight had the lowest CVD-related mortality, and that unfit men had a higher CVD-related mortality than fit men in each BMI group. Although not a controlled study, these data suggest that increased fitness may ameliorate at least some of the negative health consequences of overweight.
The BMI-mortality relationship differs somewhat between white populations and various ethnic minority populations. For example, the BMI associated with minimum mortality rate is higher (by 1-3 kg/m2) in African American than in white patients, and no relationship between BMI and mortality could be determined among a sample of Pima Indian women.13 Less information is available regarding the relationship between obesity and mortality in Asian Americans,75 Hispanic Americans, or Pacific Islanders.13,31,76 However, minority groups are often disproportionately affected by obesity and its related diseases. In addition, Asian Americans may be at increased risk for obesity-related conditions at a lower BMI because of a preferential deposition of visceral adipose tissue.77 Although populations may differ in their health risk at a given BMI, such population differences are not necessarily relevant to individuals in clinical settings.13 The NIH Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults have concluded that, based on currently available data, the use of a BMI of at least 30 as a cutoff for defining obesity appears justified regardless of race or ethnicity.13
Although beyond the scope of our report, the effects of intentional weight loss on health are of obvious importance in determining public health recommendations for weight loss in obese individuals. Numerous studies have shown that, over the short term (weeks or months), intentional weight loss in obese individuals reduces risk factors for and improves symptoms of obesity-related conditions, including heart disease, type 2 diabetes mellitus, osteoarthritis, and others.13,78- 80 Although most of these studies have measured risk-factor reduction in the short term, most81- 83 but not all84 studies measuring impact of weight loss for a year or more tend to show continuing risk factor reduction. Follow-up studies of severely obese individuals who have undergone weight loss through gastric surgery, one of which observed patients for up to 14 years after surgery,85 have shown durable improvement in type 2 diabetes mellitus85 and health-related quality of life.86 However, the applicability of these findings to those with more modest levels of obesity is not clear. At this time, there are no conclusive data proving that long-term intentional weight loss diminishes mortality rate or reduces the incidence of obesity-related disease in those who are moderately obese.87- 89 The National Institute of Diabetes and Digestive and Kidney Diseases and other cosponsors have recently announced their intention to perform a study evaluating the long-term health impact of intentional weight loss in obese individuals with type 2 diabetes mellitus.90 Although the ultimate effects of long-term weight loss on morbidity and mortality remain under study, the shorter-term decrements in risk factors and clinical improvement in symptoms strongly suggest that sustained weight reduction is of benefit for those who are obese.
Most available evidence indicates that higher levels of body weight and body fat are associated with an increased risk for the development of numerous adverse health consequences. Moderating factors, including age, sex, family history, body fat distribution, diet, and physical activity clearly can have an impact on the risks of overweight for a given individual. Thus, the benefits and risks of obesity treatment must be assessed on an individual basis, as recommended in the NIH Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults.13 There may be debate regarding the impact of overweight on mortality at a BMI of 25 to 30, but it is in the group with a BMI of at least 30, now representing more than 20% of US adults, that the greatest increase in prevalence over the past decade has occurred. Therefore, efforts to prevent further weight gain in adults at risk for overweight and obesity are essential. The advice to eat a healthful diet, increase physical activity, and avoid further weight gain is appropriate for almost all individuals at or above a healthy weight. For those whose current or future health is at risk because of their obesity and who are motivated to make lifestyle changes, a recommendation for weight loss is appropriate.
Additional research is needed to determine the association between level of obesity and health risk in diverse populations; the role of fat distribution and other metabolic variables in predicting health risk in overweight individuals; the concept of metabolic fitness and the degree to which overweight individuals can reduce their disease risk through the adoption of healthful behaviors independent of weight loss; demonstration of the extent to which sustained reduction in body weight by obese persons can bring about long-term improvements in morbidity and reduced mortality rate; and safe and effective means for preventing the development of overweight and obesity. The funding of such research has been identified as a priority of the National Institute of Diabetes and Digestive and Kidney Diseases.91
Accepted for publication August 3, 1999.
The following are members of the National Task Force on the Prevention and Treatment of Obesity: Charles J. Billington, MD (Veterans Affairs Medical Center, Minneapolis, Minn); Leonard H. Epstein, PhD (State University of New York at Buffalo); Norma J. Goodwin, MD (Health Watch Information and Promotion Service, New York, NY); James O. Hill, PhD (University of Colorado Health Sciences Center, Denver); F. Xavier Pi-Sunyer, MD (St Luke's–Roosevelt Hospital Center, Columbia University, New York, NY); Barbara J. Rolls, PhD (Pennsylvania State University, State College, University Park); Judith Stern, ScD (University of California at Davis); Thomas A. Wadden, PhD (University of Pennsylvania, Philadelphia); Roland L. Weinsier, MD, DrPH (University of Alabama at Birmingham); G. Terence Wilson, PhD (Rutgers–The State University, New Brunswick, NJ); Rena R. Wing, PhD (Brown University, Providence, RI); and Susan Z. Yanovski, MD, Van S. Hubbard, MD, PhD, Jay H. Hoofnagle, MD, James Everhart, MD, and Barbara Harrison, MS (Division of Digestive Diseases and Nutrition, National Institute of Diabetes and Digestive and Kidney Disorders, National Institutes of Health, Bethesda, Md).
We thank Maureen Harris, PhD, David Allison, PhD, and David Williamson, PhD, for their helpful comments.
Corresponding author: Susan Z. Yanovski, MD, Bldg 45, Room 6AN-18, National Institutes of Health, Bethesda, MD 20892-6600 (e-mail: email@example.com). Reprints: Weight-Control Information Network, 1 WIN WAY, Bethesda, MD 20892-3665.