NATIONAL INSTITUTE OF HEALTH
DEVELOPMENT CONFERENCE STATEMENT
49. HEALTH IMPLICATIONS OF OBESITY
Health Implications of Obesity
NIH Consensus Statement 1985 Feb 11-13; 5(9): 1-7
49. HEALTH IMPLICATIONS OF OBESITY National Institutes of Health Consensus Development Conference Statement
February 11-13, 1985
This statement was originally published as:
Health Implications of Obesity. NIH Consensus Statement 1985 Feb 11-13; 5(9):1-7.
For making bibliographic reference to the statement in the electronic form displayed here, it is recommended that the following format be used:
Health Implications of Obesity. NIH Consens Statement Online 1985 Feb 11-13 [cited year month day~; 5(9):1-7.
Current knowledge of human obesity has progressed beyond the simple generalizations of the past. Formerly, obesity was considered fully explained by the single adverse behavior of inappropriate eating in the setting of attractive foods. The study of animal models of obesity, biochemical alterations in man and experimental animals, and the complex interactions of psychosocial and cultural factors that create susceptibility to human obesity indicate that this disease in man is complex and deeply rooted in biologic systems. Thus, it is almost certain that obesity has multiple causes and that there are different types of obesity.
To assess the health implications of obesity, new knowledge and new epidemiologic observations have introduced a variety of complications that must be addressed. Thus, a reassessment of definitions and measurements of obesity is required. There is controversy surrounding the interpretation of data showing an association of body weight with morbidity and mortality. The interpretations of data from different studies have been complicated by the confounding effects of smoking behavior, the coexistence of diseases other than obesity, and variations in methods of data collection and followup. Because population samples in some studies have not been representative of the U.S. population, there have been uncertainties as to how far their conclusions can be generalized for recommendations for dietary advice and treatment.
There is evidence that an increasing number of children and adolescents are ovenweight. Even though all overweight children will not necessarily become ovenweight adults, the increasing prevalence of obesity in childhood is likely to be reflected in increasing obesity in adult years. The high prevalence of obesity in our adult population and the likelihood that the nation of the future will be even more obese demand a reassessment of the health implications of this condition.
For the special purpose of resolving the pressing questions relating to the health implications of obesity, the NIH Office of Medical Applications of Research, the National Institute of Arthriti
Based on indices of body fat, studies of large populations have shown that there is a continuous
relationship between RW or BMI and morbidity and mortality. Thus, it becomes important to establish ranges of these indices as guidelines for developing appropriate and effective approaches for the treatment and prevention of obesity.
Since the amount of body fat, as estimated by the above indices, is a continuous variable within the population, all quantitative definitions of obesity must be arbitrary. The panelists agree that an increase-in body weight of 20 percent or more above desirable body weight constitutes an established health hazard. Significant health risks at lower levels of obesity can present hazards, especially in the presence of diabetes, hypertension, heart disease, or their associated risk factors.
WHAT IS THE EVIDENCE THAT OBESITY
HAS ADVERSE EFFECTS ON HEALTH
Clinical observations have long suggested a connection of obesity (particularly in its extreme forms) with a variety of illnesses. Obesity creates an enormous psychological burden. In fact, in terms of suffering, this burden may be the greatest adverse effect of obesity. At the present time, the strongest evidence that obesity has an adverse effect on physical health comes from population-based prevalence (cross-sectional) and cohort (followup) studies. These data are complemented by weight reduction trials.
The most comprehensive data on prevalence of cardiovascular disease (CVD) risk factors and obesity are the National Health and Nutrition Examination Surveys (NHANES). NHANES I was conducted from 1971 through 1974 and NHANES n from 1976 through 1980. Both were based on a representative sample of residents of the United States.
Data from NHANES II were analyzed by comparing several parameters for the subjects at or above, or below, the 85th percentile of the reference population.* At or above this cutoffpoint, males have a BS 'v7.8 and females have a BS >97.3. This analysis showed a strong association between the prevalence of obesity and CVD risk factors. Based on these criteria, the prevalence of hypertension (blood pressure greater than 160/95) is 2.9 times higher for the overweight than for the nonoverweight. The prevalence is 5.6 times higher for the young (20 through 44 years old) overweight than for the nonoverweight subjects in this age group. The prevalence is twice as high for the obese older (45 through 74 years old) group as it is for the nonoverweight subjects of the same age. The prevalence of hypercholesterolemia (blood cholesterol over 250 mg/dl) in the young overweight age group is 2.1 times that ofthe nonoverweight group; overweight and nonoverweight subjects show similar prevalences for hypercholesterolernia after age 45.
* Noninstitutionalized, nonpregnant U.S. residents, ages 20 to 29, 197S1980.
Levels of blood pressure and serum cholesterol vary with levels of obesity in a continuous manner. This relationship holds for the so-called normal as well as the elevated range in observational studies. Intervention studies confirm that levels of blood pressure and serum cholesterol can be reduced by weight reduction.
The prevalence of reported diabetes is 2.9 times higher in overweight than nonoverweight persons in otherwise not representative of the U.S. population. The greater the degree of overweight, the higher the mortality ratio or excess death rate. Both mortality ratio and excess deaths per 1,000 per year increase with length of followup. Two small groups of insurance policyholders who reduced weight to acceptable levels for standard insurance had a decline in mortality to normal. In the insurance studies, the increased mortality with overweight was observed in normotensive men and women, without other major impairment, who would have been eligible for standard insurance rates except for being overweight: Smokers were not differentiated from nonsmokers. In the Framingham and ACS studies, the increase in excess mortality with increasing degrees of overweight was present in both smokers and nonsmokers.
The pattern of excess mortality variation with relative weight is illustrated in men ages 15 to 39 years at entry from data in the 1979 Build Study
Relative to Mortality
Average Weight Ratio
For those with relative weight of 125 to 135 percent at entry, the aggregate mortality ratio was 134 percent, as shown above. When mortality was analyzed by duration, the mortality ratio increased from 110 percent at the 0 to 5-year interval to 169 percent at the 15 to 22-year interval. The weight class for lowest mortality shown above is below the average weight category. There is higher mortality m the lowest relative weight class, 65 to 75 percent of average. In extreme obesity ("morbid'2 obesity), the mortality ratio has been reported in a small series as being of the order of 1 200 percent. A recent analysis has shown that the body mass index of minimum mortality, derived from the data in the 1979 Build Study, increases with age in a straight line relationship, the lines for male and female being virtually identical. The 1959 and 1983 Metropolitan Life Insurance Company tables of ranges of weight with minimal mortality do not provide for any age variation.
The increase in mortality versus relative weight is steeper in men and women under age 50 than in older persons, and the increase with duration is also steeper. These findings suggest that particular attention should be paid to efforts to reduce weight in younger patients.
Recent studies suggest that the distribution of fat deposits may be a better predictor of mortality than BMI or RW. If confirmed, it may be important in the future to measure fat distribution in addition to using height-weight tables.
WHAT ARE THE APPROPRIATE USES AND LIMITATIONS OF EXISTING HEIGHT-WEIGHT TABLES?
The mortality and morbidity related risks of obesity are influenced by concurrent risk factors such as smoking. Tables do not provide information on body fat distribution or degree of obesity. Frame size as used for estimation of lean (fat-free) body mass is subjectively determined in the 1959 tables. The use of elbow width to judge frame size, as suggested in the 1983 tables, may or may not eliminate the problem. . Age is not taken into account.
Body Mass Index
Body weight in kilograms
BMI = (Height in meters)2
is a simple measurement highly correlated with other estimates of fatness. It minimizes the effect of height and is useful for descriptive or evaluative purposes. It has the advantage of permitting comparison of populations. The major limitation of the BMI is that it is difficult to interpret this mathematical index to patients and to relate it to weight that must be lost.
The consensus panel recommends that physicians adopt this measure as an additional factor in evaluating patients and that nomograms be used to facilitate calculations of BS.
FOR WHAT MEDICAL CONDITIONS CAN
WEIGHT REDUCTION BE RECOMMENDED?
Weight reduction may be lifesaving for patients with extreme obesity, arbitrarily defined as weight twice the desirable weight or 45 kg (100 pounds) over desirable weight. When obesity is accompanied by severe cardiopulmonary manifestations, as in the Pickwickian syndrome, weight reduction should be part of the treatment for this medical emergency.
In view of the excess mortality and morbidity associated with obesity (as discussed above), weight reduction should be recommended to persons with excess body weight of 20 percent or more above desirable weights in the Metropolitan Life insurance Company tables (using the rnidpoint of the range for a medium-build person). In the 1983 tables, 20 percent over desirable weight is a higher weight than would be obtained by the use of the 1959 tables. The maximum increase is found in those of short stature and does not exceed 17 percent for men or 13 percent for women. Although not a specific recommendation of the panel, use of the lower weights as goals would be advisable in the presence of any of the complications or risk factors summarized below. The body mass index values, which correspond to 20 percent above desirable weight, are 27.2 and 26.9 for men and women, respectively, using the 1983 tables and 26.4 and 25.8 for men and women, respectively, using the 1959 tables. These values are not substantially different from the BMI values for men and women identified with the lower cutoffpoint for overweight as determined by the National Center for Health Statistics27.8 and 27.3 for men and women, respectively (NHANES II population, bare feet, no)
1. In infancy and childhood, we must search for biological (genetic, metabolic, or anthropometric) markers as predictors of adult obesity. Having such predictors would permit the study of the development of the disease, would provide a powerful epidemiological tool, and would allow treatment to begin very early in life.
2. The factors that regulate the regional distribution of fat and methods to assess the distribution must be developed. We need to define the mechanism by which body fat distribution is associated with adverse effects of obesity.
3. Regulation of energy balance is complex, but many aspects have begun to yield to investigation.
Promising leads are:
1. effects of the central and autonomic nervous systems and the endocrine system.
2. adipose tissue cellularity (in tissue culture) and metabolism.
3. the role of various components of thermogenesis in the overall control of energy balance.
4. control of food intake (e.g., endogenous opioids).
5. satiety factors (e.g., gut hormones).
6. Studies utilizing cultural and physical measurements in several cultures, including minority, low socioeconomic, and rapidly changing cultures, should be conducted.
7. The data from large CAHD cohort studies initiated 20 to 30 years ago should be identified and archived. Archiving should be encouraged for data obtained from ongoing and future studies.
8. Relative risk tables that incorporate both fat distribution and height-weight data should be
Great advances of modern biological science as applied to obesity can generate new information that can now be tested at the bedside. Clinical investigation utilizing the biological advances is timely. The best of public health sciences, including the anthropological and sociological, should be brought into the study of the prevention of obesity.
The evidence is now overwhelming that obesity, defined as excessive storage of energy in the form of fat, has adverse effects on health and longevity. Obesity is clearly associated with hypertension, hypercholesterolemia, NIDDM, and excess of certain cancers and other medical problems. Height and weight tables based on mortality data or the body mass index are helpful measures to determine the presence of obesity and the need for treatment. Thirty-four million adult Americans have a body mass index greater than 27.8 (men) or 27.3 (women). At this level of obesity, which is very close to a weight increase of 20 percent above desirable, treatment is strongly advised. When diabetes, hypertension, or a family history for these diseases is present, treatment will lead to benefits even when lesser degrees of obesity are present.
Obesity research efforts should be directed toward elucidation of biologic markers, factors regulating the regional distribution of fat, studies of erlergy regulation, and studies utilizing the techniques of anthropology, psychiatry, and the social sciences.