NHANES indicates National Health and Nutrition Examination Survey. See Table 1 footnote for definition of “other” race/ethnicity. Error bars indicate 95% confidence intervals.
For fitness deciles, 1 = low and 10 = high. Fitness estimated by maximal oxygen consumption. BMI indicates body mass index; HbA1c, glycosylated hemoglobin; HDL-C, high-density lipoprotein cholesterol; METs, metabolic equivalent tasks; and NHANES, National Health and Nutrition Examination Survey. To convert mg/dL of total cholesterol and HDL-C to mmol/L, multiply values by 0.0259.
Carnethon MR, Gulati M, Greenland P. Prevalence and Cardiovascular Disease Correlates of Low Cardiorespiratory Fitness in Adolescents and Adults. JAMA. 2005;294(23):2981–2988. doi:10.1001/jama.294.23.2981
Author Affiliations: Department of Preventive Medicine (Drs Carnethon, Gulati, and Greenland) and Division of Cardiology, Department of Medicine (Drs Gulati and Greenland), Feinberg School of Medicine, Northwestern University, Chicago, Ill.
Context Population surveys indicate that physical activity levels are low in the United States. One consequence of inactivity, low cardiorespiratory fitness, is an established risk factor for cardiovascular disease (CVD) morbidity and mortality, but the prevalence of cardiorespiratory fitness has not been quantified in representative US population samples.
Objectives To describe the prevalence of low fitness in the US population aged 12 through 49 years and to relate low fitness to CVD risk factors in this population.
Design, Setting, and Participants Inception cohort study using data from the cross-sectional nationally representative National Health and Nutrition Examination Survey 1999-2002. Participants were adolescents (aged 12-19 years; n = 3110) and adults (aged 20-49 years; n = 2205) free from previously diagnosed CVD who underwent submaximal graded exercise treadmill testing to achieve at least 75% to 90% of their age-predicted maximum heart rate. Maximal oxygen consumption (O2max) was estimated by measuring the heart rate response to reference levels of submaximal work.
Main Outcome Measures Low fitness defined using percentile cut points of estimated O2max from existing external referent populations; anthropometric and other CVD risk factors measured according to standard methods.
Results Low fitness was identified in 33.6% of adolescents (approximately 7.5 million US adolescents) and 13.9% of adults (approximately 8.5 million US adults); the prevalence was similar in adolescent females (34.4%) and males (32.9%) (P = .40) but was higher in adult females (16.2%) than in males (11.8%) (P = .03). Non-Hispanic blacks and Mexican Americans were less fit than non-Hispanic whites. In all age-sex groups, body mass index and waist circumference were inversely associated with fitness; age- and race-adjusted odds ratios of overweight or obesity (body mass index ≥25) ranged from 2.1 to 3.7 (P<.01 for all), comparing persons with low fitness with those with moderate or high fitness. Total cholesterol levels and systolic blood pressure were higher and levels of high-density lipoprotein cholesterol were lower among participants with low vs high fitness.
Conclusion Low fitness in adolescents and adults is common in the US population and is associated with an increased prevalence of CVD risk factors.
There is strong and consistent evidence from observational studies that physical inactivity and poor cardiorespiratory fitness (ie, fitness) are associated with higher morbidity and mortality from all causes, including cardiovascular disease (CVD) and cancer.1 United States population reports describe an increasingly less physically active society, with marked downturns in reported physical activity during adolescence and young adulthood.1- 3 Physical activity has been described in the population and in relation to health outcomes; however, prior to the current National Health and Nutrition Examination Survey (NHANES),4 data were not available to quantify objectively determined cardiorespiratory fitness in the US population.
A recently published large international case-control study attributed 12.2% of myocardial infarction in the world's population to physical inactivity.5 The extent to which physical inactivity affects the risk of heart disease through its negative impact on cardiorespiratory fitness, which is associated with a high prevalence of other CVD risk factors, is not known at the population level. The objectives of this report are to describe the prevalence of low fitness in the US population of adolescents and adults younger than 50 years and to relate low fitness to CVD risk factors in this population.
The NHANES is a nationally representative sample of the noninstitutionalized civilian US population. Race/ethnicity was self-reported, and participants were identified using a complex, stratified, multistage probability cluster design that oversampled non-Hispanic blacks, Mexican Americans, persons aged 60 years and older, and low-income individuals so that nationally representative estimates of health could be generated in these often understudied population groups. Beginning in 1999, NHANES became a continuous biannual survey rather than the periodic survey that it had been in the past. This report includes survey years 1999-2000 and 2001-2002. A detailed description of the study design and sampling methodology for NHANES is available.6,7
The overall response rate for those who completed the household interview was 81.9% (9965/12 160), and the response rate for those examined in mobile examination units was 76.3% (9282/12 160). Participants aged 12 through 49 years without existing medical conditions, diagnosed CVD, physical limitations, or abnormal hemodynamic parameters (ie, systolic blood pressure >180 mm Hg, diastolic blood pressure >100 mm Hg, or heart rate >100/min) were eligible to participate in the cardiovascular fitness component of NHANES. Of the 8457 participants in the age range for the cardiovascular fitness examination, 5315 (3110 adolescents [12-19 years] and 2205 adults [20-49 years]) completed the examination. Further details of the exclusion criteria are available.6,8 The NHANES protocol was reviewed and approved by the National Center for Health Statistics institutional review board. All participants provided written informed consent at the time of the household interview and the mobile clinic examination; additional parental consent was required for adolescents younger than 18 years.
The cardiovascular fitness component of NHANES was implemented in 1999 to provide nationally representative data on cardiovascular fitness and its relation to health conditions.8 Because it was not feasible to conduct maximal exercise tests in a population sample this size at multiple examination sites, a submaximal treadmill exercise protocol was used. The initial goal of the submaximal test was to elicit 75% of the age-predicted maximum heart rate (220 − age). During the course of data collection the protocol was modified to allow adolescents and adults to achieve up to 90% and 85%, respectively, of their age-predicted maximum heart rate. To achieve this goal within the allotted test time of 8 minutes (2-minute warm-up and two 3-minute stages), the protocol was designed to administer test protocols of varying difficulty (ie, speed, grade)8 to participants based on age, sex, body mass index (BMI), and self-reported participation in physical activities.
All tests were supervised by trained technicians. Participants walked or ran on treadmills. Heart rate was measured at the end of warm-up and each minute during recovery using an automated monitor with 4 electrodes connected to the thorax and abdomen. At the end of each stage, heart rate and blood pressure were measured. Maximal oxygen consumption (O2max) was estimated by measuring the heart rate response to reference levels of submaximal work. Higher O2max is indicative of more favorable cardiorespiratory fitness. Details of the protocols and fitness calculation formulas are available.8
Although submaximal exercise testing is used clinically to diagnose CVD, it is less than ideal for estimating high fitness, as truly fit persons may not have been asked to perform to their highest level.9 Thus, we focused our report on “low fitness” in the population by using categorizations of low, moderate, and high fitness based on reference percentile cut points as recommended in the NHANES fitness assessment manual.8 Low (<20th percentile), moderate (20th-59th percentiles), and high (≥60th percentile) fitness was defined for adults based on published data from the Aerobics Center Longitudinal Study9,10 and for adolescents based on the FITNESSGRAM program.11,12
Data were collected at all study sites by trained personnel according to standardized procedures.6 Sociodemographic information (ie, age, sex, race/ethnicity, educational attainment, personal and family medical history) was collected during the household interview. Physical examinations and laboratory measurements were performed in a mobile examination center. Weight and height were measured using standard methods and digitally recorded. Body mass index was calculated as weight in kilograms divided by the square of height in meters. Waist circumference was measured horizontally at the uppermost border of the ilium at the end of a normal expiration. Blood pressure was measured 3 to 4 times (first measurement was excluded) with participants in the seated position using a mercury sphygmomanometer.
Blood specimens were processed locally, then stored and shipped to central laboratories for analysis. Levels of total serum cholesterol and triglycerides (measured in the morning examination session only) were measured enzymatically, levels of high-density lipoprotein cholesterol (HDL-C) were measured using precipitation, and levels of low-density lipoprotein cholesterol (LDL-C) were calculated using the Friedewald equation.13 Plasma glucose levels (morning examination session only) were processed using the reference analytic method.14,15 Glycosylated hemoglobin (HbA1c) values were standardized to the method used in the Diabetes Control and Complications Trial.16
Overweight and obesity were identified in adults and adolescents by the World Health Organization BMI criterion.17 While participants with previously diagnosed hypertension and elevated resting blood pressure at the time of examination were excluded from participating in the fitness test,8 we identified previously undetected hypertension in adults according to modified guidelines (identification in this survey was based on a single measurement) from the Report of the Seventh Joint National Committee.18 Hypertension in adolescents was identified as the 90th percentile of systolic or diastolic blood pressure based on age, sex, and height.19 Glucose level was categorized according to cut points defined by the 2004 American Diabetes Association guidelines.20 The metabolic syndrome was determined in adults by National Cholesterol Education Program Adult Treatment Panel III guidelines21 in the subset of adults with measured levels of glucose and triglycerides. The metabolic syndrome was identified in adolescents with 3 or more of the following: triglycerides level of 110 mg/dL (1.4 mmol/L) or greater, HDL-C level of 40 mg/dL (1.0 mmol/L) or less (girls and boys), waist circumference greater than or equal to the sex-specific 90th percentile, fasting glucose level greater than 100 mg/dL (5.6 mmol/L), or hypertension as defined above.19
All analyses were stratified by age category (adolescent and adult) and sex. Participant characteristics were described using means and proportions and their 95% confidence intervals (CIs). The age-adjusted estimated population prevalence (and its 95% CI) of low fitness in the sample was compared across ethnicity using χ2 tests. To evaluate the pattern of association between fitness and CVD risk factors, we categorized estimated O2max into age- and sex-specific deciles and plotted the age- and race-adjusted means of each CVD risk factor. Next, we calculated age- and race-adjusted means and 95% CIs by categories of fitness and compared the low- and moderate-fitness categories with the high-fitness category using F tests. Logistic regression modeling was used to calculate odds ratios and 95% CIs for each of the categorical risk factors (eg, hypertension), comparing participants with low fitness with those in the moderate- or high-fitness categories. SAS version 9.1 (SAS Institute Inc, Cary, NC) and SUDAAN version 9.0.1 (Research Triangle Institute, Research Triangle Park, NC) were used to conduct all analyses; SUDAAN was used to account for the complex sampling design and to apply sampling weights to produce national estimates.7 Statistical significance was determined at P<.05.
Clinical characteristics (ie, lipid levels, glycemia measures, and blood pressure) of the sample were generally within the normal reference ranges, and the prevalence of newly identified hypertension and diabetes was low (Table 1). While the percentage of adolescents reporting no vigorous or moderate physical activity in the previous 30 days was below 15%, more than a quarter of adult men and women reported no activity. Among adults, the mean BMI (26.8) indicated overweight, and nearly a fifth (19%) of adults had the metabolic syndrome.
Nineteen percent (19.2%) of the surveyed population—an estimated 16 million adolescents and adults younger than 50 years—were in the low fitness category, and 33.6% of adolescents (approximately 7.5 million US adolescents) and 13.9% of adults (approximately 8.5 million US adults) had low fitness. Among adolescents, the prevalence of low fitness was similar between females (34.4%) and males (32.9%) (χ2 = 0.73; P = .40), but among adults the prevalence of low fitness was significantly higher in females (16.2%) compared with males (11.8%) (χ2 = 5.2; P = .03). Figure 1 shows the prevalence of low fitness by racial/ethnic group. Low fitness was more prevalent across age groups and sex in blacks, Mexican Americans, and those reporting “other” ethnicities than in non-Hispanic whites.
Figure 2 displays the association between deciles of fitness estimated by O2max and selected CVD risk factors across age-sex groups. Anthropometric measures, ie, BMI and waist circumference, demonstrated the most consistent inverse associations with fitness. Total cholesterol levels demonstrated a graded inverse association with fitness among adolescents, whereas among adults a trend was less evident. There was a graded positive association between HDL-C levels and fitness in adolescent and adult males, whereas a similar pattern could not be observed in females. Systolic blood pressure was highest among adults and adolescent males in the lowest fitness category, but there was no association in adolescent females. Although participants in the lowest fitness decile had correspondingly higher HbA1c values than those in the highest decile, association across deciles was not graded.
When fitness was categorized into 3 levels, participants across age-sex groups with low fitness had overall significantly higher mean BMIs and waist circumferences than those with high fitness (Table 2). Mean systolic blood pressures and total cholesterol levels showed a similar pattern by fitness categories, but some trends did not achieve statistical significance in adult females (systolic blood pressure) and adult males (total cholesterol). Diastolic blood pressures were significantly higher, while HDL-C levels were significantly lower, among adults with low vs high fitness. Comparable patterns were evident for HDL-C levels in adolescent males, but there was no association in adolescent females. The inverse associations between fitness category and mean values of triglycerides, glucose, and HbA1C did not universally achieve statistical significance.
Adolescents and adults with low fitness were 2 to 4 times more likely to be overweight or obese than were participants in the moderate or high fitness categories (Table 3). Adolescents who were less fit were more likely to have hypercholesterolemia and the metabolic syndrome, though the association with the metabolic syndrome achieved statistical significance only in adolescent males. Newly identified hypertension, low HDL-C levels, and hypercholesterolemia (among women only) were more prevalent among adults with low vs moderate or high fitness. Nonsignificant increases in prevalence of the metabolic syndrome were observed among the least-fit participants.
Findings from this report indicate that low cardiorespiratory fitness affects approximately 1 out of 5 persons aged 12 through 49 years in the US population—with a disproportionate impact on adolescents, adult females, and nonwhite minorities. The most striking indication of the health burden of low fitness in the US population is the strong association among low fitness, obesity, and CVD risk factors that is already present in adolescents and young adults.
There is a high prevalence of low fitness in adolescents and a strong correlation between fitness and the presence of CVD risk factors. The relatively higher prevalence of low fitness in adolescents compared with adults in this report may be due in part to study exclusion criteria and the method for defining low fitness. Adults with existing hypertension or other clinical CVDs, which are rare conditions in adolescents, did not participate in the treadmill component of the examination. This potentially biased the sample of adults toward being more fit. Alternatively, because aerobic capacity declines with age,22,23 the standards for defining adequate fitness in adolescents are higher than those for adults. The categories of fitness used in this study were derived from an external reference standard for adolescents that was based on a smaller select sample.11,12 Our data suggest that few adolescents in an ethnically diverse representative US sample meet these fitness expectations.
Although adolescents are not generally considered at risk for having clinical CVD events in the short term, the development of risk factors during adolescence and young adulthood sets the stage for heart disease in the middle and older ages. Much of this evidence has come from longitudinal studies such as the Coronary Artery Risk Development in Young Adulthood (CARDIA) study,24 the Bogalusa Heart Study,25 the Amsterdam Growth and Health Longitudinal Study (AGHLS),26 and the Cardiovascular Risk in Young Finns Study.27 With the exception of CARDIA and the AGHLS, these studies have not investigated the relationship between objectively determined fitness and development of CVD risk factors, and only the AGHLS measured fitness during adolescence. While we are unable to relate adolescent fitness to the subsequent development of risk factors in this cross-sectional survey, we can extrapolate from previous research to suggest that a large segment of US youth who are unfit and overweight are at risk for negative cardiovascular consequences as they age.
Low fitness is more prevalent among females compared with males, and among females, disparities by race/ethnicity were apparent. Absolute levels of cardiorespiratory fitness differ between females and males,22 but fitness categorizations in this study were made based on a sex-specific referent population. Because physical activity, the health behavior most strongly associated with fitness, is consistently lower among women,2 our findings of lower fitness among women are not surprising. Racial/ethnic differences in fitness may also be explained by comparatively lower levels of activity in nonwhites vs whites. Sex and racial/ethnic differences in low fitness nearly parallel the prevalence of obesity in the population28 and may account for the higher prevalence of overweight and obesity in females compared with males.29 The relatively higher prevalence of diabetes and hypertension in racial/ethnic minorities may be attributable, in part, to poorer fitness and physical inactivity in these subgroups. The notable exception that may not be due to health behaviors is the generally higher levels of HDL-C reported among females and blacks.
Numerous studies in adult men and women10,30- 36 report an association between low fitness and increased mortality from cardiovascular and other causes, including cancer.1,37,38 The relationship between low fitness and cardiovascular mortality in particular is proposed to be mediated by the development of CVD risk factors including hypertension, diabetes, dyslipidemia, and the metabolic syndrome.24,39 Further, physical activity training in efforts to improve fitness have been shown to lower the risk of developing risk factors, independent of changes in weight.40
Obesity and overweight could be described as the seminal public health problem today.41 Body mass index and waist circumference, which are estimates of overall and central adiposity, respectively, demonstrated the most consistent association with fitness in this study, as evidenced by the highest mean values among the least-fit persons that decreased in a nearly graded fashion with increasing fitness. While the correlation between low fitness and other CVD risk factors was not as consistent or dose-dependent, patterns indicated a generally worse cardiovascular profile among the least-fit participants. We attempted to relate low fitness to the 10-year CVD risk using the Framingham Risk Score among adults, but scores were generally too low in this relatively young population sample selected to be at low risk. Despite this, evidence of higher BMI, total cholesterol level, and blood pressure with lower fitness status suggests that less fit persons are at risk for developing CVD if these deleterious patterns persist.
To our knowledge, this is the first attempt to characterize fitness, determined by objective treadmill testing, in adolescents and adults at the population level. However, these findings are not without limitations. In this sample of US adults, fitness was estimated from a submaximal treadmill test protocol.42 In contrast to symptom-limited maximal protocols with or without direct measurement of oxygen consumption, submaximal testing is inferior for determining fitness because of its reliance on prediction formulas.9,42 These prediction formulas, which are based on estimates of predicted maximum heart rate that are often too low for physically fit persons43 and may not apply to women,22 further compound the risk of error. We were less confident in the ability of the submaximal protocol to accurately estimate “high” fitness. Consequently, we chose not to focus on the prevalence of this protective factor in the population and were unable to quantify the level of fitness that correlated with favorable CVD risk factor levels in the population. Thus, recommendations for achieving specific levels of fitness cannot be made based on this report.
Although treadmill testing is generally considered a safe diagnostic test when conducted in the presence of trained medical personnel, there remains a small risk of death during exercise. To balance the need for public health information about fitness with risks to the individual survey participants, we surveyed a population at very low risk for complications from exercise. Besides the obvious exclusion of older adults who are at higher risk for CVD, individuals with previously diagnosed hypertension were potentially the largest additional group of high-risk individuals who were not tested. An important consequence of this restriction is that we are unable to make generalizations to the population at greatest risk for CVD and the complications of low fitness—ie, older adults and individuals with existing risk factors for CVD. As a result, these data likely represent an underestimate of the true prevalence of low fitness in the population. There remains a paucity of population-level research on the distribution of fitness in older persons and persons with comorbid CVD.
Despite these limitations, this report indicates that low fitness is a prevalent and important public health problem in the US population. The consequences of declines in physical activity over time are now evident by the large proportion of society with low levels of fitness. The correlations we report between low fitness and CVD risk factors suggest a potential trend of increasing morbidity and mortality from chronic diseases—the first sign of which is the burgeoning obesity epidemic. Historical evidence from the campaign to educate about the dangers of cigarette smoking indicates that education efforts, particularly among youth, can retard and reverse these negative health behaviors. Thus, it is plausible that a similar education campaign about the health benefits of physical activity to improve cardiorespiratory fitness, in combination with changes in health care policy to make environments more conducive to physical activity, could begin to reverse this serious public health problem.
Corresponding Author: Mercedes R. Carnethon, PhD, Department of Preventive Medicine, Feinberg School of Medicine, Northwestern University, 680 N Lake Shore Dr, Suite 1102, Chicago, IL 60611 (firstname.lastname@example.org).
Author Contributions: Dr Carnethon had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Carnethon, Gulati.
Analysis and interpretation of data: Carnethon, Gulati, Greenland.
Drafting of the manuscript: Carnethon.
Critical revision of the manuscript for important intellectual content: Gulati, Greenland.
Statistical analysis: Carnethon, Gulati.
Obtained Funding: Carnethon, Greenland.
Financial Disclosures: None reported.
Funding/Support: Financial support for data collection was provided by the National Center for Health Statistics (NCHS), a part of the Centers for Disease Control and Prevention, Department of Health and Human Services. Dr Carnethon was supported in part by a career development award from the National Heart, Lung, and Blood Institute, National Institutes of Health (5 K01 HL73249-02).
Role of the Sponsor: All data used in this study were collected by the NCHS; the NCHS had no role in the design and conduct of the study; the analysis and interpretation of the data; or the preparation, review, or approval of the manuscript.