Overweight and obesity are directly related to increased risk of several chronic diseases and impaired physical function, whereas physical activity and cardiorespiratory fitness are most often inversely associated with similar increased risks (25,38). In observational studies both overweight and obesity are correlated with physical activity and cardiorespiratory fitness, with sedentary and unfit persons having greater weight for height, body mass index (BMI, kg·m−2), and percent body fat (2,5–8,11,14,15,17,32). Data from controlled clinical trials have shown that increases in physical activity result in weight loss and changes in body composition and fat distribution (4,9,16,35,36,40–42). Thus, it is possible that some, perhaps much, of the overweight and obesity seen in U.S. populations is caused by a sedentary, physically inactive lifestyle. This topic is reviewed in detail elsewhere in these proceedings.
It is difficult to know how much of the higher morbidity and mortality seen in overweight or obese individuals results from the elevated weight and how much results from inactivity. Whereas most observational studies of the relationship between physical activity or cardiorespiratory fitness and morbidity or mortality include adjustment for some aspect of body composition, many of the studies of overweight or obesity and morbidity and mortality have not included data on physical activity.
Barlow et al. (2) published the first report that included analyses specifically designed to evaluate the relation of cardiorespiratory fitness to mortality in individuals classified as normal, overweight, or obese based on their BMI. Based on their results these investigators suggested the hypothesis that moderate to high levels of cardiorespiratory fitness protect against much, if not most, of the increased mortality that accompanies overweight and obesity. This hypothesis is the basis for this review of the available evidence.
The material presented in this review addresses three specific questions:
- 1. Do higher levels of physical activity or cardiorespiratory fitness attenuate the increased risk of morbidity and mortality in overweight or obese persons?
- 2. If the evidence supports the hypothesis that higher levels of physical activity and cardiorespiratory fitness attenuate the increased risk associated with overweight and obesity, do overweight or obese individuals who are physically active or fit actually have a lower risk of morbidity and mortality than normal weight individuals who are sedentary (i.e., what is the magnitude of the protective effect of activity or fitness in overweight and obese individuals)?
- 3. Which is a more important predictor of mortality in individuals, overweight and obesity or inactivity and low fitness?
METHODS AND DEFINITIONS
We evaluated several different measures of health, overweight or obesity, and sedentary habits. Outcome variables considered for this report include all-cause mortality, the presence or development of cardiovascular disease, hypertension, stroke, CHD, type 2 diabetes mellitus, cancer, and functional health status. We found no articles on stroke or functional health status that met the inclusion criteria described below, and these variables are not mentioned further. Exposure variables included body habitus (BMI or body composition as estimated by hydrostatic weighing or skinfold thickness, and fat distribution as estimated by waist girth), and evidence of a sedentary life style (low levels of self-reported occupational or leisure-time physical activity or low levels of cardiorespiratory fitness as determined by objective laboratory assessment). These exposure variables have limitations. Current opinion holds that it is the amount and distribution of body fat that are the principal causes of high rates of chronic disease seen in overweight or obese individuals, but few studies have included direct measurement of these parameters. Physical activity is difficult to quantify, and self-report of activity patterns is the most common measurement approach used in studies of physical activity and health. Although many of the activity questionnaires and procedures have established validity, the amount of variance explained in criterion measures of energy expenditure by these methods is low to moderate (27). Cardiorespiratory fitness is assessed objectively, but it is relatively expensive and logistically complicated to perform in epidemiological studies. Limitations of cardiorespiratory fitness assessments are that these measurements only reflect activity over recent months before the assessment and are at least partly determined by heredity.
Identifying Source Material
Our objective was to identify all published reports that were from prospective investigations and included one of the outcome measures listed above, in which results were presented in strata of body habitus cross-tabulated by categories of physical activity or cardiorespiratory fitness. We excluded reports in which outcomes were given by strata of body habitus with adjustment for activity or fitness or by strata of activity or fitness with adjustment for body habitus. Such reports typically did not include sufficient information that would allow us to fully characterize the independent effect of one exposure variable across the full range of the other. We also excluded review articles and cross-sectional analyses.
Another problem that arises in considering the questions under review in this report is that many of the major studies on obesity and health did not include measurement of physical activity or fitness. In reports on obesity and health in which investigators indicated that they adjusted for physical activity, there is often no, or at best an incomplete, description of how the physical activity was measured. In other instances activity was assessed so crudely, perhaps with a single global question, that little confidence could be placed in the measurement. In these circumstances determination of the true independent effects of body habitus and energy expenditure on outcomes is impossible because of the great difference in the reliability of assessment of the two exposure variables. For these reasons, we believe that the most valid approach to evaluate the independent contributions of body habitus and energy expenditure to health is to use objective and valid measures of each. Measurement of body habitus and cardiorespiratory fitness in a laboratory or clinical setting meets this criterion, although we also include reports in which physical activity was measured with a standardized method that has been shown to have a strong and graded association with health outcomes.
We began with papers identified in the 1996 publication, Physical Activity and Health: A Report of the Surgeon General (38). Next, we performed computer searches with keywords related to our identified outcome measures in combination with keywords related to energy expenditure (physical fitness, cardiorespiratory fitness, exertion, exercise, or physical activity) and body habitus (BMI, overweight, obesity, or fat distribution). We also searched our personal files and reference lists in identified published articles. We initially identified more than 700 potentially relevant articles. After exclusions based on the above criteria, this process identified 24 articles that are included in this report.
Critical Analysis of Published Articles.
Both authors reviewed each article on the final list. We summarized results from the articles in Tables 1–6, with a different table for each outcome measure. Each table includes the study reference; a brief description of the study population; the method of assessing physical activity or fitness; the method of assessing body habitus; a description of outcomes, the number of events, and the length of follow-up (where applicable); confounding variables considered in the analyses; and a summary of the study results. We used the evidence-based approach for rating the quality of evidence that was described in the report on obesity treatment guidelines from the National Institutes of Health (25). There are no randomized controlled clinical trials that provide data on the questions under consideration in this report (Evidence Category A or B), and we did not include any consensus judgment reports (Evidence Category D). Therefore, all reports included in our results are classified in Evidence Category C (prospective observational studies).
;)
Table 1: All-cause mortality by cross tabulation of body habitus and cardiorespiratory fitness or physical activity.
Table 1A: continued.
Table 2: Cardiovascular disease (CVD) mortality by cross tabulation of body habitus and cardiorespiratory fitness.
Table 3: Coronary heart disease (CHD) by cross tabulation of body habitus and physical activity.
Table 3A: continued.
Table 4: Hypertension (HTN) by cross tabulation of body habitus and physical activity.
Table 5: Type 2 diabetes mellitus (DM) by cross tabulation of body habitus and physical activity or cardiorespiratory fitness.
Table 5A: continued.
Table 6: Cancer by cross tabulation of body habitus and physical activity.
Table 6A: continued.
{tabft}BMI, body mass index (kg·m−2); py, person-years; RR, relative risk; CI, confidence interval.
{tabft}BMI, body mass index (kg·m−2); py, person-years; RR, relative risk; LTPA, leisure time physical activity; CHD, coronary heart disease.
{tabft}LTPA, leisure time physical activity; BMI, body mass index (kg·m−2); py, person-years; RR, relative risks; CI, confidence intervals.
{tabft}LTPA, leisure time physical activity; BMI, body mass index (kg·m−2); py, person-years; RR, relative risks; CI, confidence intervals.
RESULTS
We present separate tables on the relationship of body habitus and physical activity or cardiorespiratory fitness to the outcomes of all-cause mortality, cardiovascular disease mortality, CHD, hypertension, type 2 diabetes mellitus, and cancer.
All-cause Mortality
A summary of the review of reports on all-cause mortality is presented in Table 1. There are four separate reports on cross-tabulations of cardiorespiratory fitness and BMI, body composition, or fat distribution. All these reports are from the Aerobics Center Longitudinal Study (ACLS). The first of these papers, published in 1989, had a short period of follow-up and the number of deaths was small, especially in women (3). Another limitation of this report is that the characterization of body habitus was limited to BMI categories of underweight, normal weight, and moderately overweight. The primary purpose of the study was to evaluate the relation of cardiorespiratory fitness to mortality, and detailed analyses are lacking to clarify the independent effects of the two exposures. Nonetheless, fit men and women tended to have lower all-cause mortality rates than unfit individuals in each BMI stratum. In this population, the highest mortality occurred in individuals with a BMI < 20. This is most likely a result of a lack of adjustment for smoking status and the presence of chronic disease at baseline.
The later reports from the ACLS focus specifically on the questions addressed in this review (2,18,19). These reports are from extended follow-up in the ACLS cohort, but because the number of events remains low in women, the reports include data for men only. Different selection criteria were used for the three studies, with the principal difference being whether individuals with chronic disease were excluded at baseline. The studies also differ in the classification and in the specific measure of body habitus used in analyses. Results, nevertheless, are consistent across the three studies. Men with higher levels of BMI, percent body fat, fat mass, fat-free mass, and waist girth who also were fit had lower death rates than unfit men with similar levels of body habitus measurements. In fact, in all analyses fit men with higher body habitus measurements, indicating obesity and/or overweight, had death rates not significantly different from fit men with low body habitus measurements. In these data men with high body habitus measurements had elevated mortality only if they also were unfit.
We also found two studies that used all-cause mortality as the outcome and physical activity and BMI as the exposure variables (28,32). The results were similar to those seen for cardiorespiratory fitness in the ACLS cohort. Active men in each BMI stratum had lower death rates than inactive men in the same stratum, and active men in the high BMI group had lower death rates than sedentary men in the low BMI group. A limitation of these reports is that the upper BMI categories included individuals with only mild to moderate overweight.
Cardiovascular Disease Mortality
We only found three reports on cardiovascular disease mortality (defined as ICD-9 codes 390 to 449.9) that met our review criteria (see Table 2). These include two studies from extended follow-up in the ACLS cohort discussed in the section on all-cause mortality (18,19) and one additional report from the same study (10). Results of these analyses were similar to that for all-cause mortality. Unfit men had substantially higher cardiovascular disease death rates than fit men in the same body habitus group (BMI, percent body fat, fat mass, and fat-free mass). Although there was an inverse trend for cardiovascular disease deaths across fitness categories in individuals with a BMI ≥ 27.0 (rates 17.9, 10.6, and 9.5/10,000 man-years), the trend was not significant (P = 0.24) (10). Fit men in the highest body habitus groups had lower death rates than unfit men in the lowest body habitus groups. However, there was a tendency, even in the fit men, for cardiovascular disease death rates to be higher in higher body habitus groups than in the lower body habitus groups, which suggests that higher fitness does not eliminate excess cardiovascular disease mortality in individuals in the higher body habitus groups. This is in contrast to results from the studies on all-cause mortality where no excess risk was seen in the fit men who were overweight and/or obese (Table 1).
CHD
There were seven studies that included CHD as an outcome (Table 3), sometimes combined as fatal and nonfatal events and sometimes as fatal events , and BMI and physical activity as exposure variables. BMI was determined by self-report in five of the studies and measured in the other two studies. Physical activity was estimated from self-reports of leisure-time physical activity in all studies, and one study also included data on self-reported occupational activity (33). All reports included data for men, but only one study included data for women (33). The report on women includes only 12 events, and the results are inconsistent, both within exposure groups for the women and in relation to the results for men in this and other papers.
Results of the cross-tabulation of body habitus and physical activity in relation to CHD were similar to those described earlier for all-cause and cardiovascular disease mortality. Physically active men had lower rates of CHD than inactive men in all studies, and in all BMI strata. In general, active men in the high BMI group had lower CHD rates than inactive men in the low BMI group. In some studies the difference between these divergent groups was striking. For example, Morris et al. (23) reported that men with a BMI ≥ 27.0 who regularly participated in vigorous exercise had a heart attack rate of 1.3/1000 man-years, and men with a BMI < 24.0 who were inactive had a rate of 5.5/1000 man-years. Salonen et al. (33) reported that men with a BMI ≥ 27.0 in an active occupation and with high leisure time physical activity had a CHD mortality rate of 7.4/1000. In contrast, men with a BMI < 27.0 who were sedentary at work and in leisure time had a mortality rate of 25.9/1000. Regular physical activity appeared to provide substantial protection against CHD, especially in overweight men.
Hypertension
We found only two papers with hypertension as the outcome that met our inclusion criteria (Table 4). Paffenbarger et al. (26,30) followed two cohorts of college alumni for up to 15 yr for the development of incident physician-diagnosed hypertension. Validity of self-report of both exposure and outcome assessments was high in these well-educated men. In general, active men in both studies had a lower risk of developing hypertension than sedentary men in different BMI categories, although this difference was significant only in the two highest BMI categories in the Harvard men. Among the inactive University of Pennsylvania alumni, the incidence of hypertension was greater than among active men in all BMI categories. None of these differences were statistically significant.
In summary, regular physical activity appears to reduce the risk of developing hypertension in men with elevated BMI, but this association is less marked than was observed for the mortality outcomes reviewed earlier. Reduction in risk for active men was greatest in men in the high BMI categories.
Type 2 Diabetes Mellitus
Five studies with type 2 diabetes mellitus as the outcome variable are summarized in Table 5. One study was in women and four studies included only men. Four of the studies relied on self-reported physical activity, height and weight, and presence of type 2 diabetes mellitus from mail-back questionnaires. Active individuals in these studies had a lower incidence of type 2 diabetes than sedentary participants, and this association tended to be stronger among those in the highest BMI categories. For example, there was a 39% lower risk (P = 0.005) in the active men with a BMI > 26.4 when compared with sedentary men in the Physicians’ Health Study, and this association was absent or much weaker in the lighter men (21). Data from the Honolulu Heart study were similar to those reported in the Physicians’ Health Study, with the additional observation that the trend across activity categories in men with a BMI ≥ 28.2 was not significant for all men combined or men with a glucose < 225 mg·dL−1 (6). Results for men and women were similar in these studies.
Studies relating physical activity and type 2 diabetes mellitus that rely on self-reported physical activity and presence of diagnosed diabetes may underestimate the true risk of sedentary habits in relation to this disease. This is because of the limitations of self-reported exercise habits and also because approximately 50% of the prevalent cases of type 2 diabetes are undiagnosed, which leads to substantial misclassification on both the exposure and outcome variables. Wei et al. (39) recently reported a greater than 3-fold increased risk for developing type 2 diabetes in men with low cardiorespiratory fitness when compared with high fit men. The odds ratio was 3.7 when adjusted for age, history of parental diabetes, and length of follow-up, and this was only reduced to 3.2 after additional adjustment for BMI, high blood pressure, high density lipoprotein (HDL)-cholesterol, total cholesterol, triglycerides, cigarette smoking status, alcohol intake, and change in fitness. In addition to an objective measurement of the activity exposure, diabetes in this study also was determined objectively by measuring fasting plasma glucose concentrations at baseline and at follow-up. These authors reported a steep inverse gradient of risk across fitness categories among individuals with BMI of < 27.0 (P < 0.001) and ≥ 27.0 (P < 0.006) (see Fig. 1). Fit men with a BMI of ≥ 27.0 had a slightly lower incidence of type 2 diabetes than unfit men in the BMI < 27.0 group. Wei et al. (39) also reported a substantially increased risk of developing impaired fasting glucose (fasting plasma glucose ≥ 110 to < 126 mg·dL−1). The odds ratio for impaired fasting glucose in men with low fitness when compared with men with high fitness was 2.1 in the fully adjusted model. The results were similar among men in high and low strata of BMI.
Figure 1: Incidence of type 2 diabetes (per 1000 man-years) among cardiorespiratory fitness levels according to BMI (kg·m\-2). White bars indicate the low fitness group, diagonal lines the moderate fitness group, and black bars the high fitness group. Adapted by permission from 39. Wei, M., L. W. Gibbons, T. L. Mitchell, J. B. Kampert, C. D. Lee, and S. N. Blair. The association between cardiorespiratory fitness and impaired fasting glucose and type 2 diabetes mellitus in men. Ann. Intern. Med. 130:89–96, 1999.
Being active and fit appears to reduce the risk of type 2 diabetes, at least in heavy individuals. This association was strongest in the study with objective measures of both exposures and outcome.
Cancer
We found four studies in which cancer was the outcome (Table 6). Physical activity was assessed in all studies by self-report. Self-reported height and weight were used to calculate BMI in one study. Two studies were in women only, and the other two studies included both women and men. In general, we consider the results for cancer inconclusive. In several of the analyses cancer risk was higher in sedentary individuals, but the results in the highest BMI groups were not consistent. For example, sedentary women with a BMI ≥ 25.0 had nearly five times higher endometrial cancer risk than highly active women in the same BMI category (20). However, there was a nonsignificant trend of higher risk of breast cancer across activity categories in women with a BMI > 25.7 (37). In the same study, a steep inverse gradient of breast cancer risk across activity groups was found in women with a BMI < 22.8. It is difficult to generalize the results presented in Table 6 because of the inconsistent results associated with different cancer outcomes. In some studies the outcome was a single type of cancer, and in one study cancer at all sites was the outcome of interest.
CONCLUSIONS
The review performed for this report addressed three questions:
1.
Do higher levels of physical activity or cardiorespiratory fitness attenuate the increased risk of morbidity and mortality in overweight or obese persons?
Evidence statement.
Overweight and obese individuals who are active and fit have lower rates of disease and death than overweight and obese individuals who are inactive and unfit. This inverse gradient of risk across activity or fitness categories is present in various strata of body habitus and frequently is steeper in the higher categories of body habitus variables.
Rationale.
Active and fit individuals in nearly all the studies summarized in Tables 1–6 had lower rates of morbidity and mortality than sedentary and unfit individuals in all strata of body habitus. These findings remained after adjustment for age and other potential confounding variables and were seen for the various specific outcomes included here. An active and fit way of life apparently provides protection against several chronic diseases, and this association typically is present in all strata of body habitus. The inverse gradient of risk across levels of activity or fitness often was steeper in the higher strata of body habitus than among leaner individuals. For example, Morris et al. (23) report heart attack rates in men with a BMI ≥ 27.0 of 7.3/1000 man-years in sedentary men and only 1.3/1000 man-years in regular vigorous exercisers. Corresponding rates in men with a BMI < 24.0 were 5.5/1000 man-years in sedentary men and 1.9/1000 man-years in regular vigorous exercisers.
2. Do overweight or obese individuals who are physically active or fit actually have a lower risk of morbidity and mortality than normal weight individuals who are sedentary (i.e., what is magnitude of the protective effect of activity or fitness in overweight and obese individuals)?
Evidence Statement:
Overweight or obese individuals who are active and fit are less likely to develop obesity-related chronic diseases and have early death than normal weight persons who lead sedentary lives.
Rationale.
Overweight or obese individuals who were active and fit had morbidity and mortality rates that were at least as low, and in many instances much lower, than normal weight individuals who were sedentary. Perhaps the strongest evidence supporting this conclusion comes from the reports by Lee et al. (18,19). These investigators followed a large cohort of men for an average of 8.5 yr. The strength of this study is that objective measurements were available for cardiorespiratory fitness and several measures of body habitus, including BMI, percent body fat, fat mass, fat-free mass, and waist circumference. Furthermore, the study was large enough to allow for analyses in smokers and nonsmokers, with exclusion for early mortality and with adjustment for several important potential confounders. In this report unfit men in the normal range of body habitus (BMI 19–25, percent body fat < 16.7%, or waist circumference < 87 cm) had more than a two-fold higher mortality risk than fit men in the highest category of body habitus (BMI ≥ 27.8, percent body fat ≥ 25%, or waist circumference ≥ 99 cm). This apparent protection of moderate to high cardiorespiratory fitness extends to men who clearly are obese (BMI > 30) (2).
3. Which is a more important predictor of mortality in individuals, overweight or obesity or inactivity and low fitness?
Evidence statement.
Inactivity and low cardiorespiratory fitness are as important as overweight or obesity as predictors of mortality, at least in men.
Rationale.
The question considered here relates to risk for individuals and not to the importance of either of these exposures as public health problems. It is clear that overweight and obesity are associated with increased risk of morbidity and mortality (25) and that an active way of life protects against morbidity and mortality (38). Because many studies on obesity and health have not adequately accounted for the possible confounding effects of physical activity, the independent effects of these two exposures are unclear. It is tempting, based on the reports reviewed here from the ACLS (2,18,19), to conclude that low fitness is a more important predictor of mortality in individuals than overweight or obesity, because excess mortality risk is largely eliminated in overweight or obese men who are also fit. In our judgment, such a conclusion would be premature. Other reports reviewed here did not have the specific purpose of addressing this question, and there are limitations on some of the analyses that did not focus specifically on this issue. In addition, data from the ACLS are available only for men, and this cohort is from middle to upper socioeconomic strata, with few members of minority groups. This homogeneity is an advantage in relation to internal validity of the study but presents limitations for generalization. Replication of the studies reported from the ACLS is needed.
Issues and limitations.
All studies reviewed here are from Evidence Category C. It is unlikely that there will ever be randomized clinical trials to address specifically the questions posed in this report because of the required size, complexity, and cost. Thus, the best evidence we are ever likely to have on this issue will come from prospective observational studies.
Although the results reviewed here are remarkably consistent for various endpoints, because different measures of habitual physical activity or fitness, different strata of body habitus, and different populations were used in the various studies, conclusions must remain tentative. Few studies included women, and for some that did, the number of events was small. There are virtually no data available from minority groups. Health status or change in health status often was not considered or was inadequately addressed, although results are similar for studies that excluded unhealthy individuals at baseline and those that did not. Only reports from the ACLS had the specific purpose of addressing the questions considered in this review. While other reports had some data related to the issue, the lack of focus on the relevant question often led to incomplete analyses insofar as the present review is concerned.
Crude measurements of self-reported physical activity are an important limitation of many of the studies. In some studies, physical activity appears not to have been a very important variable under consideration during the planning phase of the study. When only one or two very simple questions on habitual physical activity are included as part of the study measurements, it seems reasonable to conclude that this was not a high priority item for the investigators. Under these conditions it is not realistic to expect that adequate and appropriate analyses can be performed with physical activity as an exposure. As an extreme example, suppose that an investigator wanted to address question 3 that is posed in this report. Further suppose that in that study the investigator measured physical activity by the question “Are you physically active or not?” and assessed body habitus by computed tomography scanning. One could hardly expect to find that inactivity is as important as body habitus as a predictor of mortality in such a study, even if inactivity actually were more important. Therefore, one of the limitations to keep in mind when evaluating results from the studies reviewed here is the relative validity of the exposure measures used.
Another of the major limitations of several of the current studies is that the upper body habitus stratum often began in the mild overweight range and may not have included many truly obese individuals. The conclusions presented here should not be extended to individuals with Class II or Class III obesity (BMI ≥ 35.0).
Research recommendations.
Additional research is needed to address the questions posed for this review. We recommend that the following types of studies be conducted:
- 1. Replicate prospective observational studies of the predictive validity of physical activity or cardiorespiratory fitness and body habitus for morbidity and mortality in diverse population groups. It is especially important that studies be conducted in women, older persons, and minority groups.
- 2. Prospective observational studies with other outcome measures need to be conducted. Overweight and obesity increase risk for functional limitations associated with aging and for knee osteoarthritis. It is not clear that regular physical activity might reduce these risks in overweight or obese persons.
- 3. Improve the assessment of physical activity in prospective observational studies of populations so that less misclassification occurs for this exposure. Ideally, objective measurements of activity by electronic monitoring should be considered.
- 4. Develop creative designs for randomized clinical trials to determine the independent and interactive effects of physical activity and body habitus on subclinical markers of chronic disease. Random assignment to activity interventions with blocking on strata of body habitus is one possible approach. Because controlled trials with morbidity and mortality as outcome measures are unlikely to be carried out in relation to the issues addressed in this review, investigators should incorporate more direct measures of disease processes in smaller, logistically-feasible clinical trials. Examples of such measures are ultrasound measures of carotid wall thickness, coronary artery calcification by electron beam computed tomography, near-infrared spectroscopy assessment of lipid content of atherosclerotic plaque, endothelial function of the coronary circulation by phase contrast, magnetic resonance imaging (MRI) with infusion of adenosine, endothelial function of the peripheral circulation by reactive hyperemia in response to ischemic exercise, left ventricular function by MRI, and mechanical function of vasculature by pulse wave velocity measurements.
We thank Milton Z. Nichaman, M.D., Sc.D., for helpful comments on an earlier draft of the manuscript, Melba Morrow, M.A., for editorial assistance, and Stephanie Parker for secretarial support.
Our work is supported in part by grants from the National Institutes of Health AG06945, HL48597, and HL58608; Polar Electro Oy; and Kellogg Company.
Adress for correspondence: Steven N. Blair, The Cooper Institute, 12330 Preston Road, Dallas TX 75230. E-mail: [email protected].
REFERENCES
1. Albanes, D., A. Blair, and P. R. Taylor. Physical activity and risk of cancer in the NHANES I population. Am. J. Public Health 79: 744–750, 1989.
2. Barlow, C. E., H. W. Kohl, III, L. W. Gibbons, and S. N. Blair. Physical fitness, mortality and obesity. Int. J. Obes. Relat. Metab. Disord. 19: S41-S44, 1995.
3. Blair, S. N., H. W. Kohl, III, R. S. Paffenbarger, Jr., D. G. Clark, K. H. Cooper, and L. W. Gibbons. Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA 262: 2395–2401, 1989.
4. Bouchard, C., J.-P. Despres, and A. Tremblay. Exercise and obesity. Obes. Res. 1: 133–147, 1993.
5. Brown, W. J., A. J. Dobson, and G. Mishra. What is a healthy weight for middle aged women? Int. J. Obes. Relat. Metab. Disord. 22: 520–528, 1998.
6. Burchfiel, C. M., D. S. Sharp, J. D. Curb, et al. Physical activity and incidence of diabetes: The Honolulu Heart Program. Am. J. Epidemiol. 141: 360–368, 1995.
7. Coakley, E. H., I. Kawachi, J. E. Manson, F. E. Speizer, W. C. Willet, and G. A. Colditz. Lower levels of physical functioning are associated with higher body weight among middle-aged and older women. Int. J. Obes. Relat. Metab. Disord. 22: 958–965, 1998.
8. Cooper, K. H., M. L. Pollock, R. P. Martin, S. R. White, A. C. Linnerud, and A. Jackson. Physical fitness levels vs selected coronary risk factors: a cross-sectional study. JAMA 236: 166–169, 1976.
9. Despres, J. P., M. C. Pouliot, S. Moorjani, et al. Loss of abdominal fat and metabolic response to exercise training in obese women. Am. J. Physiol. 261: E159–E167, 1991.
10. Farrell, S. W., J. B. Kampert, H. W. Kohl, et al. Influences of cardiorespiratory fitness levels and other predictors on cardiovascular disease mortality in men. Med. Sci. Sports Exerc. 30: 899–905, 1998.
11. Gibbons, L. W., S. N. Blair, K. H. Cooper, and M. Smith. Association between CHD risk factors and physical fitness in healthy adult women. Circulation 67: 977–983, 1983.
12. Haapanen, N., S. Miilunpalo, I. Vuori, P. Oja, and M. Pasanen. Association of leisure time physical activity with the risk of CHD, hypertension, and diabetes in middle-aged men and women. Int. J. Epidemiol. 26: 739–747, 1997.
13. Helmrich, S. P., D. R. Ragland, R. W. Leung, and R. S. Paffenbarger, Jr. Physical activity and reduced occurrence of non-insulin-dependent diabetes mellitus. N. Engl. J. Med. 325: 147–152, 1991.
14. Hsieh, S. D., H. Yoshinaga, T. Muto, and Y. Sakurai. Regular physical activity and coronary risk factors in Japanese men. Circulation 97: 661–665, 1998.
15. Huang, Y., C. A. Macera, S. N. Blair, P. A. Brill, H. W. Kohl, III, and J. J. Kronenfeld. Physical fitness, physical activity, and functional limitation in adults aged 40 and older. Med. Sci. Sports Exerc. 30: 1430–1435, 1998.
16. Kohrt, W. M., K. A. Obert, and J. O. Holloszy. Exercise training improves fat distribution patterns in 60- to 70-year-old men and women. J. Gerontol. 47: M99-M105, 1992.
17. Kushi, L. H., R. M. Fee, A. R. Folsom, P. J. Mink, K. E. Anderson, and T. A. Sellers. Physical activity and mortality in postmenopausal women. JAMA 277: 1287–1292, 1997.
18. Lee, C. D., S. N. Blair, and A. S. Jackson. Cardiorespiratory fitness, body composition, and all-cause and cardiovascular disease mortality in men. Am. J. Clin. Nutr. 69: 373–380, 1999.
19. Lee, C. D., A. S. Jackson, and S. N. Blair. US weight guidelines: is it also important to consider cardiorespiratory fitness? Int. J. Obes. Relat. Metab. Disord. 22: S2-S7, 1998.
20. Levi, F., C. La Vecchia, E. Negri, and S. Franceschi. Selected physical activities and the risk of endometrial cancer. Br. J. Cancer 67: 846–851, 1993.
21. Manson, J. E., D. M. Nathan, A. S. Krolewski, M. J. Stampfer, W. C. Willett, and C. H. Hennekens. A prospective study of exercise and incidence of diabetes among U.S. male physicians. JAMA 268: 63–67, 1992.
22. Manson, J. E., E. B. Rimm, M. J. Stampfer, et al. Physical activity and incidence of non-insulin-dependent diabetes mellitus in women. Lancet 338: 774–778, 1991.
23. Morris, J. N., D. G. Clayton, M. G. Everitt, A. M. Semmence, and E. H. Burgess. Exercise in leisure time: coronary attack and death rates. Br. Heart J. 63: 325–334, 1990.
24. Morris, J. N., R. Pollard, M. G. Everitt, and S. P. W. Chave. Vigorous exercise in leisure-time: protection against CHD. Lancet II: 1207–1210, 1980.
25. National Institutes of Health and National Heart, Lung and Blood Institute. Clinical Guidelines on the Identification, Evaluation, and Treatment of Overweight and Obesity in Adults: The Evidence Report.
Obes. Res. 6(Suppl. 2)S51-S209, 1998.
26. Paffenbarger, R. S. J., A. L. Wing, R. T. Hyde, and D. L. Jung. Physical activity and incidence of hypertension in college alumni. Am. J. Epidemiol. 117: 245–257, 1983.
27. Paffenbarger, R. S., Jr., S. N. Blair, I.-M. Lee, and R. T. Hyde. Measurement of physical activity to assess health effects in free-living populations. Med. Sci. Sports Exerc. 25: 60–70, 1993.
28. Paffenbarger, R. S., Jr., R. T. Hyde, A. L. Wing, and C. C. Hsieh. Physical activity, all-cause mortality, and longevity of college alumni. N. Engl. J. Med. 314: 605–613, 1986.
29. Paffenbarger, R. S., Jr., R. T. Hyde, A. L. Wing, and C. H. Steinmetz. A natural history of athleticism and cardiovascular health. JAMA 252: 491–495, 1984.
30. Paffenbarger, R. S., Jr., D. L. Jung, R. W. Leung, and R. T. Hyde. Physical activity and hypertension: an epidemiological view. Ann. Med. 23: 319–327, 1991.
31. Paffenbarger, R. S., Jr., A. L. Wing, and R. T. Hyde. Physical activity as an index of heart attack risk in college alumni. Am. J. Epidemiol. 108: 161–175, 1978.
32. Rosengren, A. and L. Wilhelmsen. Physical activity protects against coronary death and deaths from all causes in middle-aged men: evidence from a 20-year follow-up of the Primary Prevention Study in Goteborg. Ann. Epidemiol. 7: 69–75, 1997.
33. Salonen, J. T., J. S. Slater, J. Tuomilehto, and R. Rauramaa. Leisure time and occupational physical activity: Risk of death from ischemic heart disease. Am. J. Epidemiol. 127: 87–94, 1988.
34. Slattery, M. L., J. Potter, B. Caan, et al. Energy balance and colon cancer–beyond physical activity. Cancer Res. 57: 75–80, 1997.
35. Stefanick, M. L. Exercise and weight control. Exerc. Sport Sci. Rev. 21: 363–396, 1993.
36. Stefanick, M. L., S. Mackey, M. Sheehan, N. Ellsworth, W. L. Haskell, and P. D. Wood. Effects of diet and exercise in men and postmenopausal women with low levels of HDL cholesterol and high levels of low density lipoprotein (LDL) cholesterol. N. Engl. J. Med. 339: 12–20, 1998.
37. Thune, I., T. Brenn, E. Lund, and M. Gaard. Physical activity and the risk of breast cancer. N. Engl. J. Med. 336: 1269–1275, 1997.
38. U.S.Department of Health and Human Services. Physical Activity and Health: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996, pp. 81–172.
39. Wei, M., L. W. Gibbons, T. L. Mitchell, J. B. Kampert, C. D. Lee, and S. N. Blair. The association between cardiorespiratory fitness and impaired fasting glucose and type 2 diabetes mellitus in men. Ann. Intern. Med. 130: 89–96, 1999.
40. Wood, P. D., W. L. Haskell, S. N. Blair, et al. Increased exercise level and plasma lipoprotein concentrations: a one-year, randomized, controlled study in sedentary, middle-aged men. Metabolism 32: 31–39, 1983.
41. Wood, P. D., M. L. Stefanick, D. M. Dreon, et al. Changes in plasma lipids and lipoproteins in overweight men during weight loss through dieting as compared with exercise. N. Engl. J. Med. 319: 1173–1179, 1988.
42. Wood, P. D., M. L. Stefanick, P. T. Williams, and W. L. Haskell. The effects on plasma lipoproteins of a prudent weight-reducing diet with or without exercise in overweight men and women. N. Engl. J. Med. 325: 461–466, 1991.
