Population-Based Reference Standards for Cardiovascular Fitness among U.S. Adults

Physical inactivity is an important etiologic factor in many chronic conditions, including obesity, type 2 diabetes, and several manifestations of cardiovascular disease (^2,5,8,13). Although the benefits of exercise are well known, the majority of the U.S. population does not engage in activities that are consistent with the levels recommended for health (⁶).

An important outcome of physical activity is cardiovascular fitness, measured as maximal oxygen uptake (V˙O_2max), a health-related component of overall physical fitness that relates to the ability of the circulatory and respiratory systems to supply oxygen to the body during sustained physical activity (¹¹). Though genetic factors explain approximately 40-60% of the variation in V˙O_2max (^3,4), cardiovascular fitness can also be improved in most individuals who participate in appropriate physical activities (^1,11). There is compelling evidence to support the concept that increased cardiovascular fitness can decrease many chronic health risks (^9,13,14). To this end, increasing the number of U.S. adults who participate in physical activity, particularly vigorous activity, that promotes the development and maintenance of cardiovascular fitness is a national health objective (¹²).

Our group recently provided national estimates of V˙O_2max (estimated from submaximal exercise testing) and sex- and age-specific cardiovascular fitness levels among U.S. adults, based on data from the National Health and Nutrition Examination Survey (NHANES 1999-2000 and 2001-2002) (⁷). The cardiovascular fitness levels for adults in NHANES were established using V˙O_2max cut-points derived from the Aerobics Center Longitudinal Study (ACLS) (²). In our recent study, we found that fitness levels were lowest among non-Hispanic black adults, particularly among non-Hispanic black women. However, differences in the distribution of low, moderate, and high fitness among the various groups examined may have been caused by a potential for bias or misclassification because the cut-points for categorizing and reporting cardiovascular fitness levels in NHANES were based on data from the ACLS, a mostly non-Hispanic white sample. The purpose of the present study was to determine the distribution of cardiovascular fitness levels among U.S. adults based on estimated V˙O_2max cut-points derived from NHANES, a large, nationally representative sample of the U.S. population. In turn, the new cut-points established in this study might be useful reference standards to categorize cardiovascular fitness levels of U.S. adults for health risk assessment purposes.

METHODS

The NHANES survey design is a stratified, multistage probability sample of the civilian noninstitutionalized U.S. population (http://www.cdc.gov/nchs/nhanes.htm). The current NHANES includes oversampling of Mexican-American and non-Hispanic black persons to improve estimates for these groups. Approximately 9965 persons aged 2 months to 85 yr were studied in NHANES 1999-2000 and 11,039 persons in NHANES 2001-2002. The NHANES protocol was reviewed and approved by the National Center for Health Statistic's institutional review board. Written informed consent was obtained from all participants prior to any testing.

Testing procedures and details of the NHANES protocol are available elsewhere (http://www.cdc.gov/nchs/about/major/nhanes/). All individuals aged 12-49 yr were eligible to participate in the cardiovascular fitness-testing component of NHANES. Participants were screened prior to testing using questionnaires and a physical examination. Certain medical conditions (e.g., angina and self-reported lung or breathing problems) excluded individuals from testing, as well as medication usage (e.g., antiarrhythmias and beta-blockers), and physical limitations (e.g., bone or joint problems that could be made worse by walking) that would interfere with cardiovascular responses to exercise. Health technicians, all trained using a common protocol, performed the fitness examination. Participants were assigned to one of eight submaximal treadmill test protocols, based on sex, age, body mass index (BMI, kg·m⁻²), and self-reported level of physical activity. The goal of each protocol was to elicit a HR that was approximately 75% of the age-predicted maximum (220 − age) by the end of the test. Each protocol included a 2-min warm-up, two 3-min exercise stages, and a 2-min cool-down period.

The primary outcome for the fitness component was estimated V˙O_2max (mL·min⁻¹·kg⁻¹), calculated from the HR response to known levels of submaximal work. Based on sex- and age-specific criteria, the estimated V˙O_2max was also categorized and reported as a low, moderate, or high level of cardiovascular fitness. These cut-points were established from data for adults ages 20-49 yr from the ACLS (^2,10). Low cardiovascular fitness was defined as an estimated V˙O_2max below the 20th percentile of the same sex and age group; a moderate fitness level was a value between the 20th and 59th percentiles, and high fitness was at or above the 60th percentile (Table 1).

Data from each survey were linked using the unique survey participant identifier (SEQN). The initial sample consisted of 2214 nonpregnant adults aged 20-49 yr who had an estimated V˙O_2max value recorded. Participants who were categorized as "other Hispanics" and "other races, including multiracial" were excluded from the analyses because of inadequate sample sizes for these groups. Of the 2214 participants in the initial sample, six participants who had extreme estimated V˙O_2max values (≥87.5 mL·min⁻¹·kg⁻¹, at or above the 99th percentile of the sample distribution), and all participants with missing or negative sampling weights (N = 230) were excluded, giving a final sample of 1978 adults who had complete data available for the analyses. Based on data from this sample and the corresponding population-based sample weights, identical sex- and age-specific criteria was used to establish new cut-points to categorize estimated V˙O_2max as a low, moderate, or high level of cardiovascular fitness (Table 2). We then reanalyzed the distribution of low, moderate, and high cardiovascular fitness using these new cut-points.

Data were analyzed using SAS (Version 9.1, SAS® Institute, Cary, NC) survey procedures. The full sample was analyzed using the 4-yr full sample medical examination weights (WTMEC4YR) to estimate means and 95% confidence intervals (CI), and the masked variance units (pseudo-primary sampling units (SDMVPSU) and pseudo-stratum (SDMVSTRA)) to estimate those SEM. Continuous variables are presented as a mean (± SE), including age, BMI, and estimated V˙O_2max. Categorical variables are presented as a frequency (N), weighted frequency (based on sampling weights), and percentage (with corresponding 95% CI), including sex, race, and cardiovascular fitness level. Percentages are based on the weighted frequencies. The association between row (race or race-sex groups) and column variables (cardiovascular fitness level) was analyzed using the Rao-Scott χ² test for proportions. Statistical significance was established at α = 0.05 a priori. All pairwise comparisons were adjusted using the Bonferroni method.

RESULTS

Characteristics of the 1978 adults included in the sample, along with the corresponding population-based estimates, are provided in Table 3. The sex distribution of the sample was similar to that of the population (Table 3A), whereas the race distribution was approximately 29% Mexican-American, 20% non-Hispanic black, and 51% non-Hispanic white. As shown in Table 3B, non-Hispanic white subjects were significantly older (P < 0.017) than the other race groups. Similarly, non-Hispanic white males and females were significantly older than their Mexican-American counterparts (P < 0.008). Estimated V˙O_2max was significantly lower (P < 0.017) among non-Hispanic black than Mexican-American and non-Hispanic white adults. Estimated V˙O_2max was also significantly lower (P < 0.008) among non-Hispanic black females than Mexican-American and non-Hispanic white females, but there were no differences among males. Finally, BMI was significantly lower (P < 0.017) among non-Hispanic white than non-Hispanic black and Mexican-American adults, and BMI was lower (P < 0.008) in non-Hispanic white and Mexican-American females than non-Hispanic black females.

For each sex and age category, the revised cut-points based on NHANES data shown in Table 2 were slightly higher than those based on the cut-points derived from ACLS data (Table 1). Figure 1A shows the distribution of cardiovascular fitness levels by race, as reported by us previously (⁷), and Figure 1B shows the distribution of cardiovascular fitness levels by race based on the revised cut-points. There was a significant association between race and cardiovascular fitness level (P < 0.0001), regardless of the cut-points used. Although the percentage of individuals in the low fitness category was higher in all race groups based on NHANES (new) compared with ACLS (old) cut-points, this difference was greatest among non-Hispanic black adults. Using the NHANES cut-points, 31.9% (95% CI: 26.0-37.8%) of non-Hispanic black adults were categorized as having a low cardiovascular fitness level compared with approximately 23% using the ACLS cut-points.

Figure 2A shows the distribution of cardiovascular fitness levels by race-sex groups, as reported previously (⁷), and Figure 2B provides the distribution of cardiovascular fitness levels by race-sex groups based on the revised cut-points. There was a significant association between race and cardiovascular fitness level for each sex (P < 0.0001), regardless of the cut-points used. The percentage of individuals in the low fitness category was also higher for all race-sex groups based on the NHANES-derived cut-points compared with the ACLS-derived cut-points. This difference was greatest among non-Hispanic black men; 29% of non-Hispanic black men had a low fitness level using NHANES compared with 15% using ACLS cut-points. However, the absolute percentage of individuals categorized as having a low cardiovascular fitness level was still highest among non-Hispanic black women; 35% using the NHANES compared with 31% using the ACLS derived cut-points.

DISCUSSION

The cardiovascular fitness component was implemented in NHANES in 1999 to provide nationally representative data on cardiovascular fitness among the noninstitutionalized U.S. population. One purpose in doing so was to establish prevalence estimates for specific groups that may be at increased health risk due to poor cardiovascular fitness. We recently reported that non-Hispanic black adults, and non-Hispanic black women in particular, had lower cardiovascular fitness than other groups in the United States (⁷). We found no evidence to support our earlier suggestion that differences in the distribution of low, moderate, and high cardiovascular fitness among the various groups might have been caused by a potential for bias or misclassification because the cut-points used to categorize and report fitness levels in NHANES were derived from the ACLS, which was composed of a mostly non-Hispanic white sample. Despite using revised cut-points to establish cardiovascular fitness levels that were derived from NHANES, which is composed of a more representative sample of the U.S. population, we still found that non-Hispanic black adults, and non-Hispanic black women in particular, had the highest percentage of low cardiovascular fitness, compared with their Mexican-American and non-Hispanic white counterparts. In fact, the percentage of non-Hispanic black adults and non-Hispanic black women who had a low cardiovascular fitness level was actually higher in the present study compared with our earlier report (⁷).

As depicted in the figures, the distribution of low, moderate, and high cardiovascular fitness differed substantially among the various race and race-sex groups. This difference was most apparent among females, with roughly 35% of non-Hispanic black women categorized has having a low cardiovascular fitness level, compared with 22% of Mexican-American and 18% of non-Hispanic white women. Thus, on a population basis, over one third of all non-Hispanic black women likely have a low cardiovascular fitness level. The present report supports our earlier findings and provides further evidence that non-Hispanic black women in particular may be at increased health risk due to poor cardiovascular fitness. This suggests a need for health promotion specialists to target efforts toward increasing cardiovascular fitness among non-Hispanic black women to help decrease the prevalence of low fitness in this group. This suggestion is supported by recent efforts of public health professionals in Richmond, VA, to engage relatively high-risk residents (who were disproportionately African American and female, had higher BMI levels, and were more likely to report a family history of chronic disease compared with the general community) into a multipronged intervention focused on dietary change and increased physical activity levels (¹⁵).

There are a number of important limitations in the present study. First, the estimates for the distribution of cardiovascular fitness levels in the population were derived from estimated V˙O_2max values based on submaximal exercise testing. There is error involved with estimated V˙O_2max from this type of testing because the results are based on extrapolation to HR_max established from a theoretical formula. In addition, errors in this estimate may be compounded by any factor that influences the HR response to exercise (e.g., temperature and stress) or the measurement of HR itself (e.g., equipment used to measure HR). Second, the cross-sectional nature of the NHANES survey does not allow us to make causal inferences as to the underlying reason(s) for the higher percentage of non-Hispanic black adults who have a low cardiovascular fitness level compared with other major race groups in the United States. Potential explanatory factors might include differences in BMI, income, physical activity level, and body composition, all of which will be explored in our future studies. Third, it should be noted that these findings might not be generalizable to all U.S. adults because of the select age and physical status of the sample examined. Finally, by excluding participants categorized as "other Hispanics" and "other races, including multiracial," we are unable to provide a precise estimate of estimated V˙O_2max (and fitness level distribution) for "all races" combined.

The goal of this article was to present the distribution of cardiovascular fitness levels among various race groups in the United States using appropriate cut-points that were derived from national estimates of V˙O_2max (estimated from submaximal exercise). In turn, the established cut-points might be used to categorize and report cardiovascular fitness levels of adults in the U.S. population for health risk assessment purposes. Despite the limitations noted, the reference standards presented in this article should provide a more representative estimate of cardiovascular fitness levels among U.S. adults, which in turn might be useful as an evaluation tool by health care providers and other professionals.

Glen Duncan was supported by K01 DK61999; funds from this grant were used to defray publication costs.