Abstract
Background
Assessing maximal oxygen consumption (VO2max), the gold standard for assessing cardiorespiratory fitness, is often impractical in large-scale studies. We derived sex-specific VO2 max and graded exercise test duration (GXTd) prediction equations from 6-minute walk test (6MWT) performance.
Methods
Data were from 564 Coronary Artery Risk Development in Young Adults (CARDIA) participants (mean age: 61.5 years; 58% women; 40% Black) who completed the 6MWT and symptom-limited modified Balke treadmill graded exercise test at the Year 35 (2021-2022) follow-up exam. Sex-stratified samples were randomly split (2/3 for training, 1/3 for testing) to derive and evaluate prediction equations. Stepwise linear regression identified predictors of VO2max and GXTd from 6MWT distance. Sex-specific VO2max CARDIA equations were compared with Burr and FRIEND equations. Models’ accuracies were evaluated by comparing the predicted values to measured values using Student’s t-test and Pearson correlation coefficients (r). Bland−Altman plots used to evaluate agreement between measured and predicted values.
Results
Sex-specific VO2max CARDIA equations explained 53% and 57% of the variance in men and women, respectively, with strong correlations between measured and predicted values (r = 0.73 and 0.78). The Burr and FRIEND equations showed larger biases and weaker correlations compared with the CARDIA equations. The GXTd equations explained 59% and 62% of the variance in men and women, with strong correlations (r = 0.71 and 0.72) and no significant mean differences between observed and predicted.
Conclusions
The CARDIA prediction equations for VO2max and GXTd from 6MWT enhance accuracy and accessibility, providing a practical tool for large-scale studies and clinical assessments, particularly in aging populations.
Keywords: Physical fitness, VO2peak, Cohort studies
Introduction
Cardiorespiratory fitness (CRF), the ability of the circulatory and respiratory systems to deliver oxygen to working muscles during physical activity and exercise, provides important diagnostic and prognostic information.1 Specifically, strong scientific evidence supports the inverse association between CRF and cardiovascular disease,2 all-cause mortality,3 and other adverse health outcomes such as obesity, type 2 diabetes, hypertension, and metabolic syndrome.2,4,5 These associations highlight the utility of CRF assessments for screening and risk stratification in adults.
The criterion measure of CRF is the direct measurement of maximal oxygen consumption (VO2max) during a maximal graded exercise test (GXT). However, this method is resource-intensive, requiring specialized equipment to capture gas exchange, trained personnel, and maximal exertion from participants, making the protocol less practical to implement in large-scale, population-based studies. To address these limitations, indirect measures of CRF, such as submaximal exercise tests and field-based protocols, have been developed to estimate VO2max.6 Among these alternatives, the 6-minute walk test (6MWT), a submaximal exercise test measuring the distance walked in 6 min, is easy to administer across a wide-range of expertise and skill levels, requires minimal equipment, is well-tolerated and safe across clinical and non-clinical populations, and has demonstrated its validity in CRF classification in comparison to the criterion in adults.7 Several studies have validated the utility of the 6MWT to predict VO2max; however, its generalizability is limited by small sample sizes, homogeneity in study populations (eg, predominantly White adults), or a focus on specific clinical populations.8-11 Further, research has shown that sex differences in fitness exist, with lower fitness levels in women compared to men due to differences in body composition, hemoglobin concentrations, and maximal cardiac output.12
Several cohort studies, such as the Coronary Artery Risk Development in Young Adults (CARDIA) study and Cooper Center Longitudinal Study, have relied on the duration of GXT (graded exercise test duration (GXTd); ie, time to exhaustion) to estimate CRF.5,13 GXTd has demonstrated strong predictive validity for VO2max in smaller laboratory-based studies of active and sedentary young adults (r ≥ 0.88),14,15 as well as within middle-aged to older adults in the population-based CARDIA cohort (r = 0.83).16 Predicting GXTd from 6MWT distance could provide an additional practical tool for CRF estimation, particularly in aging populations where maximal exertion tests are less feasible.
Developing sex-specific VO2max prediction equations using the 6MWT in a racially/ethnically diverse cohort could enhance their accuracy in predicting CRF in diverse populations. In addition, equations that estimate treadmill performance from the 6MWT can support continuity in CRF assessment over time, particularly when maximal testing becomes less feasible in aging or large-scale, population-based studies. Linking past treadmill assessments with current 6MWT performance preserves the integrity of CRF tracking over time and enhances our ability to study its impact on future health. This study aimed to develop sex-specific prediction equations for VO2max based on 6MWT distance within the CARDIA cohort, a diverse and well-characterized longitudinal prospective cohort study. It also sought to examine the accuracy of the newly developed equations in comparison to existing 6MWT and non-exercise VO2 prediction equations commonly used in clinical and research settings. Comparing these equations will allow us to evaluate whether the new CARDIA-derived equations offer improved accuracy and fit for a cohort of middle-aged to older adults relative to existing approaches. By addressing gaps in the literature, these newly developed equations could provide a practical, accessible, and non-invasive method for CRF assessment, particularly in aging populations or settings where traditional testing is impractical.
Methods
CARDIA is a population-based longitudinal investigation of the development of cardiovascular risk factors and subclinical and clinical cardiovascular disease. Beginning in 1985-1986, 5115 Black and White young adults (18-30 years of age) were enrolled for an in-person examination at one of four clinical centers across the United States (Birmingham, Alabama, United States; Chicago, Illinois, United States; Minneapolis, Minnesota, United States; and Oakland, California, United States), with 10 additional in-person examinations held approximately every 2-5 years,17 to date. This study includes a subgroup of participants who attended the Year 35 (Y35; 2021-2022) in-person follow-up examination and consented to additional measures including a symptom-limited treadmill GXT and 6MWT, as part of two separate ancillary studies (CARDIA ACT-HF and CARDIA Function, respectively). All CARDIA participants provided written consent. Study procedures for the Y35 core exam and all ancillary studies were approved by the University of Alabama at Birmingham (UAB) Institutional Review Board (IRB); all other Field Centers agreed to rely on the UAB IRB as the single IRB for CARDIA.
Participant descriptives
Standardized questionnaires and protocols were used to collect participant characteristics, such as age, sex assigned at birth (male or female), self-identified race (Black or White), education [high school (or equivalent) degree or less; or associate’s degree or more], smoking status (current, former, never), and meeting physical activity guidelines (≥300 exercise units as assessed by the self-reported CARDIA Physical Activity History questionnaire, which estimates a minimum weekly frequency and duration).18 Height was measured to the nearest 0.5 cm using a stadiometer and weight to the nearest 0.5 lb (0.2 kg) using a calibrated balance-beam scale, both taken in light clothing without shoes. Body mass index (BMI) was calculated using measured weight in kilograms and height in meters (kg/m2) and BMI categories were defined as healthy weight (≤24.9 kg/m2), overweight (25.0-29.9 kg/m2), and having obesity (≥30 kg/m2).
Graded exercise treadmill testing
Medical eligibility for the GXT was determined during a screening using the American College of Sports Medicine criteria.19 Participants were ineligible to participate in the GXT if they had a history of heart, blood vessel, or lung conditions, severe hypertension at rest, or had an abnormal electrocardiogram prior to exercise. Similar to prior assessments at the baseline, Y7, and Y20 CARDIA exams, participants underwent a symptom-limited GXT using a modified Balke protocol with physician supervision20 at the Y35 follow-up exam to measure CRF.13 The modified Balke protocol consisted of a 2 min warm-up (2.0 mph, 0% grade) followed by up to nine 2 min stages (up to 18 minutes total) of progressively increasing difficulty, beginning at a workload of 4.1 METs and ending at 19.0 METs at stage nine, and a 3 min recovery (2.0 mph, 0% grade).13 The first stage in the protocol began at a treadmill speed of 3.0 mph and 2% grade, stages 2 through 6 maintained 3.4 mph with 6% grade for the stage 2 and then increased by 4% in each stage up to stage 6, 4.2 mph at 22% and 25% grade for stages 7 and 8, respectively, followed 5.6 mph at 25% grade for stage 9.
Heart rate, blood pressure, and a 12-lead ECG were obtained on each participant at rest (5 min), and heart rate, blood pressure, 12-lead ECG, and pulse oximetry were obtained at the end of each stage, at maximum exercise, and every minute for 3 min post-exercise. Respiratory exchange ratio (RER) and ratings of perceived exertion (RPE; 6-20 scale) were recorded near the end of each stage and at maximal exercise.21 Oxygen consumption was measured to obtain VO2max using an integrated metabolic measurement system (Ultima CardiO2, MGC Diagnostics, Saint Paul, Minnesota, United States). In addition to VO2max, total treadmill GXTd was recorded in total minutes and seconds and subsequently calculated as total seconds.
Six-minute walk test
The 6MWT was conducted as part of the CARDIA Function ancillary study which included a battery of five physical performance measures and was administered by trained research staff who followed a standardized protocol and script. Participants were instructed to walk as fast as safely possible to cover as much distance during the 6 min period on an indoor hallway course approximately 30 feet in length. Participants were provided with updates on time remaining every minute (eg, 1 min down, 5 to go) and at the last 30 s. At the halfway point (ie, 3 min), participants were provided with verbal encouragement. The total distance covered during the period was recorded to the nearest meter.
Comparison to existing VO2max prediction equations
Two existing prediction equations, the Burr equation22 and FRIEND equation,23 were used for comparison with the newly developed CARDIA-derived prediction equation. The Burr equation, developed in a sample of working-aged adults (aged 20-60 years), is frequently used for predicting VO2max from 6MWT distance. In addition to 6MWT distance, the equation incorporates body weight, sex, resting heart rate, and age as predictors of VO2max. The FRIEND equation, by contrast, is a non-exercise VO2max equation using age, sex, and body weight as predictor variables. The FRIEND equation was derived from the FRIEND registry, a diverse sample of healthy US men and women. These equations were selected because they are widely used in both clinical and research settings, represent different approaches to VO2max estimation (exercise-based vs non-exercise), and have been previously validated. We aimed to examine whether our newly developed equations would offer better or worse performance in predicting VO2max.
Statistical analysis
For the current analyses, inclusion criteria included achieving at least one of the following indicators of a maximal effort during the GXT: exercise to volitional exhaustion, max heart rate ≥ 85% of age-predicted maximal heart rate using the Tanaka formula,24 and/or RER ≥ 1.10. Participant characteristics are reported, overall and stratified by sex, as means and standard deviations (SD) for continuous variables and proportions for categorical variables. Student’s t tests or Chi Square (χ2) tests were used to examine differences in participant characteristics by sex. The sample was stratified by sex, and stratified samples were randomly split into training sets (2/3 within strata) for model derivation and test sets (1/3 within strata) for validation. No additional cross-validation procedures were conducted because model performance was evaluated in this independent hold-out test set. Candidate predictors were selected a priori based on prior literature, clinical relevance, and practical availability, and included 6MWT distance (m), age (years), BMI (kg/m2), physical activity status, resting heart rate (bpm), blood pressure category, education, and smoking status. Bidirectional stepwise selection was then applied within the sex-specific training datasets to identify the final multivariable linear regression models for VO2max and GXTd. All model selection steps were confined to the training sets; final model performance was evaluated in the independent test sets. The models’ accuracies were evaluated in the test sets by comparing the measured and predicted VO2max (CARDIA, Burr, and FRIEND) and GXTd values using Student’s pairwise t-test and Pearson correlation coefficients (r). Additionally, Bland−Altman plots were generated to assess the agreement between measured and predicted values. All analyses were conducted in R (Version 4.3.2). Statistical significance was determined using an alpha level of 0.05.
Results
The study sample comprised 564 participants, including 58.2% women and 40.4% Black individuals, with a mean age of 61.5 years (SD = 3.6, range 53-68) and a mean BMI of 28.6 kg/m2 (SD = 5.7). Participant characteristics of the overall sample and by sex are presented in Table 1. Men demonstrated higher height and weight, a higher proportion meeting physical activity guidelines, greater CRF (both VO2max and GXTd) and better 6MWT performance compared to women. Participants included in the analytical sample (consented to GXT, were medically eligible, and met inclusion criteria) were more likely to self-identify as White, have a lower BMI, have higher educational attainment, were more likely to meet physical activity guidelines, and have never smoked compared to the remaining Y35 CARDIA participants not included in the analyses (ie, did not complete both GXT and 6MWT) (Table S1).
Table 1.
Participant characteristics, overall, and by sex.
| Total | Female | Male | p-value | |
|---|---|---|---|---|
| Characteristics | N = 564 | N = 328 | N = 236 | |
| Testing center | .91 | |||
| Birmingham | 138 (24.5%) | 79 (24.1%) | 59 (25.0%) | |
| Chicago | 89 (15.8%) | 53 (16.2%) | 36 (15.3%) | |
| Minnesota | 127 (22.5%) | 71 (21.6%) | 56 (23.7%) | |
| Oakland | 210 (37.2%) | 125 (38.1%) | 85 (36.0%) | |
| Race | .006 | |||
| Black | 228 (40.4%) | 149 (45.4%) | 79 (33.5%) | |
| White | 336 (59.6%) | 179 (54.6%) | 157 (66.5%) | |
| Age (years) | 61.5 (3.6) | 61.4 (3.7) | 61.7 (3.5) | .30 |
| Height (cm) | 169.6 (9.3) | 164.1 (6.7) | 177.3 (6.7) | <.001 |
| Weight (kg) | 82.4 (18.1) | 77.2 (17.9) | 89.7 (15.8) | <.001 |
| Body mass index (kg/m2) | 28.6 (5.7) | 28.6 (6.4) | 28.5 (4.7) | .80 |
| Healthy weight | 163 (28.9%) | 105 (32.0%) | 58 (24.6%) | |
| Overweight | 194 (34.4%) | 103 (31.4%) | 91 (38.6%) | |
| Obesity | 207 (36.7%) | 120 (36.6%) | 87 (36.9%) | |
| Education | .21 | |||
| Associate degree or more | 383 (76.9%) | 234 (79.1%) | 149 (73.8%) | |
| High school or less | 115 (23.1%) | 62 (20.9%) | 53 (26.2%) | |
| Smoking status | .43 | |||
| Current | 36 (6.5%) | 17 (5.3%) | 19 (8.1%) | |
| Former | 115 (20.8%) | 67 (21.0%) | 48 (20.4%) | |
| Never | 403 (72.7%) | 235 (73.7%) | 168 (71.5%) | |
| Meeting physical activity guidelines | .005 | |||
| Yes | 279 (50.0%) | 145 (44.8%) | 134 (57.3%) | |
| No | 279 (50.0%) | 179 (55.2%) | 100 (42.7%) | |
| Resting HR (bpm) | 65.1 (10.2) | 65.6 (9.7) | 64.4 (10.8) | .17 |
| Resting SBP (mmHg) | 121.3 (15.0) | 121.0 (16.3) | 121.8 (13.0) | .50 |
| Resting DBP (mmHg) | 72.4 (9.9) | 72.2 (10.3) | 72.6 (9.3) | .63 |
| GXT duration (s) | 329.7 (144.1) | 281.5 (125.2) | 396.4 (142.3) | <.001 |
| VO2max (mL/kg/min) | 26.0 (7.3) | 24.0 (6.7) | 28.9 (7.3) | <.001 |
| 6MWT distance (m) | 477.4 (102.8) | 467.7 (95.1) | 490.8 (111.4) | .010 |
Each column reports mean (SD) or N (%). Student’s t and χ2-tests were used to examine potential sex differences. Not meeting or meeting physical activity guidelines based on a threshold of <300 or ≥300 exercise units. BMI categories: healthy weight: <25 kg/m2, overweight: 25.0-29.9 kg/m2, obesity ≥ 30 kg/m2.Abbreviations: 6MWT, six-minute walk test; DBP, diastolic blood pressure; GXT, graded exercise test; HR, heart rate; s, seconds; SBP, systolic blood pressure; VO2, maximal oxygen consumption.
Predicting VO2max
The newly developed CARDIA sex-specific VO2max prediction equations following stepwise model selection are presented in Table 2. In both men and women, the final model included 6MWT distance, age, BMI, meeting or not meeting physical activity guidelines, resting heart rate, and smoking status as significant predictors of VO2max, resulting in adjusted R2 values of 0.53 and 0.57, respectively. Table 3 presents validation metrics for each equation, including mean predicted VO2max, mean difference from measured values with 95% confidence intervals, p-values, Pearson correlation coefficients (r), and standard error of the estimate (SEE) for men and women. Measured and predicted VO2max values based in their respective sex-specific equations are presented in Table 3. The differences between measured and predicted VO2max were close to 0 in both men and women (p = .63 and .20, respectively), with strong Pearson correlations between measured and predicted values in both men and women (r = 0.73 and 0.78, respectively).
Table 2.
Newly developed cardiorespiratory fitness equations.
| VO2max (mL/kg/min) | |
| Men | VO2 = 64.920 + (0.018 × 6MWT)−(0.344 × age)−(0.781 × BMI) + (2.934 × PA)−(0.099 × RestingHR)−(3.952 × Smoking) |
| Women | VO2 = 37.633 + (0.012 × 6MWT)−(0.044 × age)−(0.600 × BMI) + (2.209 × PA)−(0.044 × RestingHR)−(2.692 × Smoking) |
| GXT duration (s) | |
| Men | GXTd = 1110.422 + (0.393 × 6MWT)−(8.577 × age)−(12.465 × BMI) + (54.362 × PA)−(2.048 × RestingHR)−(83.698 × Smoking) |
| Women | GXTd = 667.671 + (0.373 × 6MWT)−(4.204 × age)−(9.275 × BMI) + (46.503 × PA)−(2.090 × RestingHR)−(82.922 × Smoking) |
Inputs: 6MWT (meters); age (years); BMI (kg/m2), PA (0 = not meeting guidelines, 1 = meeting guidelines), RestingHR (bpm); smoking (0 = former or never smoker; 1 = current smoker).
Abbreviations: 6MWT, 6-minute walk test; BMI, body mass index; GXT, graded exercise test; PA, physical activity; RestingHR, resting heart rate; VO2, maximal oxygen consumption.
Table 3.
Validation of newly developed cardiorespiratory fitness prediction equations.
| Equation | Mean ± SD | Mean difference (95% CI) | p-value | r | SEE |
|---|---|---|---|---|---|
| VO2max (mL/kg/min) | |||||
| Men | |||||
| Measured | 29.25 ± 6.99 | ||||
| CARDIA Equation | 29.52 ± 6.05 | 0.27 (−0.83, 1.37) | .627 | 0.73 | 4.81 |
| Burr equation | 32.38 ± 6.95 | 3.13 (1.81, 4.45) | <.001 | 0.65 | 6.56 |
| FRIEND Equation | 30.52 ± 4.80 | 1.27 (−0.03, 2.58) | .056 | 0.58 | 5.85 |
| Women | |||||
| Measured | 24.21 ± 7.22 | ||||
| CARDIA Equation | 23.64 ± 5.47 | −0.57 (−1.44, 0.30) | .196 | 0.78 | 4.52 |
| Burr equation | 28.04 ± 6.80 | 3.83 (2.84, 4.82) | <.001 | 0.68 | 6.98 |
| FRIEND equation | 20.29 ± 5.18 | −3.92 (−4.92, −2.92) | <.001 | 0.64 | 6.58 |
| GXT duration (s) | |||||
| Men | |||||
| Measured | 408.17 ± 148.96 | ||||
| CARDIA equation | 407.22 ± 113.44 | −0.95 (−24.86, 22.96) | .937 | 0.71 | 104.67 |
| Women | |||||
| Measured | 291.75 ± 135.66 | ||||
| CARDIA equation | 280.14 ± 103.19 | −11.61 (−29.76, 6.54) | .208 | 0.72 | 94.07 |
Abbreviations: Burr, Burr equation; CARDIA, CARDIA equation; FRIEND, FRIEND Registry Equation; GXT, graded exercise test; r, Pearson correlation; SEE, standard error of the estimate; VO2, maximal oxygen consumption.
The Burr equation, derived from a younger and less diverse cohort, produced VO2max estimates that overestimated measured values in this sample of middle-aged to older adult CARDIA participants by 3.13 mL/kg/min in men and 3.83 mL/kg/min in women, on average. In men and women, the Burr equation showed correlations with measured VO2max of r = 0.65 and 0.68, respectively, and SEE of 6.56 and 6.98 mL/kg/min. The FRIEND equation yielded VO2max predictions that were closer to measured values than the Burr equation but still demonstrated mean differences of 1.27 mL/kg/min for men and −3.92 mL/kg/min for women. Correlations with measured VO2max were r = 0.58 in men and r = 0.64 in women, and SEE values were 5.85 and 6.58 mL/kg/min, respectively. Bland−Altman plots comparing measured and predicted VO2max values for the newly developed CARDIA equation, Burr equation, and FRIEND equation are presented in Figure 1.
Figure 1.
Bland−Altman Plots for predicted versus measured VO2max for men and women based on the CARDIA, Burr, and FRIEND equations. Solid center line represents mean bias. Dashed outer lines represent upper and lower limits of agreement (±1.96SD). Plots A, B, and C represent men; Plots D, E, and F represent women. BURR, Burr equation; CARDIA, CARDIA equation; FRIEND, FRIEND Registry Equation; VO2, maximal oxygen consumption.
Predicting GXTd
6MWT distance, age, BMI, physical activity status, resting HR, and smoking status were significant predictors of GXTd in both men and women (Table 2). Presented in Table 2, the newly developed CARDIA prediction equations resulted in adjusted R2 values of 0.59 and 0.62 in men and women, respectively. The differences between measured and predicted GXTd were close to 0 in both men and women (p = .94 and .21, respectively), with strong Pearson correlations between measured and predicted values in both men and women (r = 0.71 and 0.72, respectively) (Table 3). In contrast to VO2max, GXTd was underestimated in both men and women (mean differences: −0.95 s and −11.61 s, respectively). Bias and limits of agreement for GXTd measured and predicted values are presented in Bland Altman plots in Figure 2.
Figure 2.
Bland−Altman Plots for predicted versus measured GXTd for men and women based on the CARDIA equation. Solid center line represents mean bias. Dashed outer lines represent upper and lower limits of agreement (±1.96 SD). CARDIA, CARDIA equation; GXTd, graded exercise test equation.
Discussion
The results of this study underscore the utility of the 6MWT as a practical and accurate measure to include in models estimating CRF, both VO2max and GXTd, in large-scale, population-based studies. Our findings build upon previous research demonstrating the validity of 6MWT distance as a predictor of VO2max, with further development of sex-specific prediction equations for both VO2max and GXTd in a population-based sample of middle-aged to older adults. The strong correlation between predicted and measured values for both VO2max and GXTd, along with the lack of significant differences, indicates that the prediction models are robust. The inclusion of age, BMI, meeting or not meeting physical activity guidelines, resting heart rate, and smoking status as predictors in addition to 6MWT distance in our models provide additional, simple to assess, participant characteristic information aiding in the accuracy of CRF prediction.
The new VO2max CARDIA prediction equations for men and women explained 53% and 57% of the variance in VO2max, respectively, with strong correlations between measured and predicted VO2max values (r = 0.73 and 0.78). Although small biases were observed, with VO2max being slightly overestimated in men and underestimated in women, the mean differences were not statistically significant, indicating that the equations performed well across sex-specific groups. For context, the Burr equation,22 selected due to its widespread use in estimating VO2max from 6MWT distance, overestimated VO2max in this sample in both men and women. Correlations between predicted and measured VO2max were r = 0.65 for men and r = 0.68 for women. This overestimation is likely attributable to differences in the samples used for equation development; the Burr equation was derived from a sample of healthy, working-aged individuals with higher CRF levels than those in our study. The FRIEND equation,23 developed using a diverse sample of healthy US men and women, also showed mean differences from measured VO2max values, tending to overestimate VO2max in men and underestimate it in women, with correlations of r = 0.58 and r = 0.64. Together, these observations highlight how VO2max prediction equations derived in distinct populations yield different levels of agreement when applied to a middle-aged to older adult cohort.
The newly developed GXTd prediction equations also demonstrated robust performance with R2 values of 0.59 and 0.62 in men and women, respectively. Measured and predicted GXTd values were strongly correlated (r = 0.71 and r = 0.72, respectively) and showed no significant differences (p = .94 and p = .21, respectively). Highlighting the utility of 6MWT distance for estimating GXTd in large-scale studies where time to exhaustion is used as a proxy for CRF. By accurately estimating GXTd, our equations expand the application of the 6MWT, offering an additional, practical metric for CRF assessment that aligns with a historical treadmill protocol in CARDIA.13 Moreover, these equations provide a valuable tool for longitudinal cohort studies, such as CARDIA, where treadmill-based tests were routinely performed in earlier study phases but may no longer be feasible as cohort participants age. The ability to harmonize historical treadmill data with current 6MWT performance ensures continuity in CRF assessment and supports the exploration of CRF trends and their associations with long-term health outcomes.
Despite the strengths of this study, several limitations should be acknowledged. First, while the newly developed VO2max and GXTd prediction equations demonstrated robust predictive performance, the explained variance (R2) remains moderate, indicating that additional predictors could improve accuracy of the estimates. Including variables such as objectively measured physical activity or body composition measures might improve model performance but would increase complexity and resource demands, potentially reducing the equations’ practical utility. Second, these equations were developed using data from the CARDIA cohort including the CARDIA-modified Balke treadmill protocol and 6MWT. As such, their performance may not generalize to other populations or to studies employing different exercise testing protocols, particularly for GXTd predictions. Furthermore, participants had to enroll in the GXT ancillary study and meet ancillary specific inclusion criteria which introduces a potential selection bias, as individuals willing to participate in a GXT may possess a certain level of fitness and physical function, potentially influencing the study outcomes.25 Additionally, The CARDIA Y35 in-person exam occurred during COVID-19, possibly limiting further limiting participation. Finally, our results may not be generalizable to other race and ethnic groups (eg, Asian adults) or age groups, as they were not represented in CARDIA. Future research should validate these equations across diverse cohorts and protocols to ensure broader applicability.
Conclusions
The development of sex-specific prediction equations for VO2max and GXTd from 6MWT performance represents a significant advancement in the field of CRF assessment. These equations are particularly valuable for large-scale epidemiological studies where traditional GXT may not be feasible.
Supplementary Material
Contributor Information
Bjoern Hornikel, Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama, United States.
Erin E Dooley, Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama, United States.
Christopher Barrett Bowling, Durham Veterans Affairs Medical Center (VAMC), Durham Veterans Affairs Geriatric Research Education and Clinical Center, Durham, North Carolina, United States; Department of Medicine, Duke University, Durham, North Carolina, United States.
Baojiang Chen, Department of Biostatistics and Data Science, UTHealth Houston School of Public Health in Austin, Austin, Texas, United States.
Pablo Martinez-Amezcua, Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States.
David R Jacobs, Jr, Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, Minneapolis, Minnesota, United States.
Mercedes Carnethon, Department of Preventive Medicine, Northwestern University, Chicago, Illinois, United States.
Barbara Sternfeld, Division of Research, Kaiser Permanente Northern California, Pleasanton, California, United States.
Cora E Lewis, Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama, United States.
Stephen Sidney, Division of Research, Kaiser Permanente Northern California, Pleasanton, California, United States.
Priya Palta, Department of Neurology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, United States.
Kelley P Gabriel, Department of Epidemiology, University of Alabama at Birmingham, Birmingham, Alabama, United States.
Supplementary material
Supplementary material is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.
Funding
The Coronary Artery Risk Development in Young Adults Study (CARDIA) is supported by contracts 75N92023D00002, 75N92023D00003, 75N92023D00004, 75N92023D00005, and 75N92023D00006 from the National Heart, Lung, and Blood Institute (NHLBI). Support also provided by R01HL149796 (to K.P.G.) and R01AG062502 (to C.B.B.). This research was supported by the Paffenbarger-Blair Fund for Epidemiological Research on Physical Activity from the American College of Sports Medicine Foundation.
Conflicts of interest
None declared.
Data availability
CARDIA data are available upon reasonable request from the CARDIA Coordinating Center. CARDIA investigators are eager to collaborate with investigators interested in using CARDIA data. Please see the CARDIA website (https://sites.uab.edu/cardia/for-researchers/publications/) for publication policies and for a list of CARDIA Representatives. CARDIA data are also publicly available on the NIH-supported BioLINCC and dbGaP platforms.
Author contributions
Bjoern Hornikel conceived the study idea, analyzed the data, interpreted the findings, and drafted this manuscript. Kelley P. Gabriel and Erin E. Dooley helped with study conceptualization, interpretation, manuscript drafting, correction, and finalization. Baojiang Chen and David R. Jacobs, Jr. provided analytic support, helped with interpretation of findings, and manuscript drafting, correction, and finalization. C. Barrett Bowling, Pablo Martinez-Amezcua, Mercedes Carnethon, Barbara Sternfeld, Cora E. Lewis, Stephen Sidney, and Priya Palta provided critical review of the study conceptualization and aided in the manuscript correction and finalization. All authors have contributed to the manuscript and approved the submitted version.
References
- 1. Arena R, Myers J, Williams MA, et al. Assessment of functional capacity in clinical and research settings. Circulation. 2007;116:329-343. 10.1161/CIRCULATIONAHA.106.184461 [DOI] [PubMed] [Google Scholar]
- 2. Carnethon MR, Gidding SS, Nehgme R, Sidney S, Jacobs DR Jr., Liu K. Cardiorespiratory fitness in young adulthood and the development of cardiovascular disease risk factors. JAMA. 2003;290:3092-3100. 10.1001/jama.290.23.3092 [DOI] [PubMed] [Google Scholar]
- 3. Ross R, Blair SN, Arena R, et al. ; Stroke Council. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation. 2016;134:e653-e699. 10.1161/CIR.0000000000000461 [DOI] [PubMed] [Google Scholar]
- 4. Blair SN, Kampert JB, Kohl HW III, et al. Influences of cardiorespiratory fitness and other precursors on cardiovascular disease and all-cause mortality in men and women. JAMA. 1996;276:205-210. 10.1001/jama.1996.03540030039029 [DOI] [PubMed] [Google Scholar]
- 5. Blair SN, Kohl HW, Paffenbarger III, et al. Physical fitness and all-cause mortality: a prospective study of healthy men and women. JAMA. 1989;262:2395-2401. 10.1001/jama.1989.03430170057028 [DOI] [PubMed] [Google Scholar]
- 6. Noonan V, Dean E. Submaximal exercise testing: clinical application and interpretation. Phys Ther. 2000;80:782-807. 10.1093/ptj/80.8.782 [DOI] [PubMed] [Google Scholar]
- 7. Dourado VZ, Nishiaka RK, Simões MSMP, et al. Classification of cardiorespiratory fitness using the six-minute walk test in adults: comparison with cardiopulmonary exercise testing. Pulmonology. 2021;27:500-508. 10.1016/j.pulmoe.2021.03.006 [DOI] [PubMed] [Google Scholar]
- 8. Cahalin LP, Mathier MA, Semigran MJ, Dec GW, DiSalvo TG. The six-minute walk test predicts peak oxygen uptake and survival in patients with advanced heart failure. Chest.1996;110:325-332. 10.1378/chest.110.2.325 [DOI] [PubMed] [Google Scholar]
- 9. Lipkin DP, Scriven AJ, Crake T, Poole-Wilson PA. Six minute walking test for assessing exercise capacity in chronic heart failure. Br Med J (Clin Res Ed). 1986;292:653-655. 10.1136/bmj.292.6521.653 [DOI] [Google Scholar]
- 10. Miyamoto S, Nagaya N, Satoh T, et al. Clinical correlates and prognostic significance of six-minute walk test in patients with primary pulmonary hypertension. Am J Respir Crit Care Med. 2000;161:487-492. 10.1164/ajrccm.161.2.9906015 [DOI] [PubMed] [Google Scholar]
- 11. Zugck C, Krüger C, Dürr S, et al. Is the 6-minute walk test a reliable substitute for peak oxygen uptake in patients with dilated cardiomyopathy? Eur Heart J. 2000;21:540-549. 10.1053/euhj.1999.1861 [DOI] [PubMed] [Google Scholar]
- 12. Wilmore JH, Costill DL, Kenney WL, Physiology of sport and exercise. Vol 20. Human Kinetics Champaign, IL; 2004. [Google Scholar]
- 13. Sidney S, Haskell WL, Crow R, et al. Symptom-limited graded treadmill exercise testing in young adults in the CARDIA study. Med Sci Sports Exerc. 1992;24:177-183. [PubMed] [Google Scholar]
- 14. Pollock ML, Foster C, Schmidt D, Hellman C, Linnerud AC, Ward A. Comparative analysis of physiologic responses to three different maximal graded exercise test protocols in healthy women. Am Heart J.. 1982;103:363-373. 10.1016/0002-8703(82)90275-7 [DOI] [PubMed] [Google Scholar]
- 15. Pollock ML, Bohannon RL, Cooper KH, et al. A comparative analysis of four protocols for maximal treadmill stress testing. Am Heart J. 1976;92:39-46. [DOI] [PubMed] [Google Scholar]
- 16. Hornikel B, Dooley EE, Sternfeld B, et al. Validation of a time-to-exhaustion-derived V̇O2max prediction equation and secondary criteria agreement: the CARDIA study. Exercise Sport Movement. 2025;3(3):esm.0000000000000047. 10.1249/esm.0000000000000047 [DOI] [Google Scholar]
- 17. Lloyd-Jones DM, Lewis CE, Schreiner PJ, Shikany JM, Sidney S, Reis JP. The Coronary Artery Risk Development in Young Adults (CARDIA) study: JACC Focus Seminar 8/8. J Am Coll Cardiol. 2021;78:260-277. 10.1016/j.jacc.2021.05.022 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Pettee Gabriel K, Sidney S, Jacobs DR Jr., et al. Convergent validity of a brief self-reported physical activity questionnaire. Med Sci Sports Exerc. 2014;46:1570-1577. 10.1249/mss.0000000000000278 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Liguori G, Feito Y, Fountaine C, Roy BA; American College of Sports M. ACSM’s Guidelines for Exercise Testing and Prescription. Eleventh ed. Wolters Kluwer; 2022. [Google Scholar]
- 20. Myers J, Forman DE, Balady GJ, et al. Supervision of exercise testing by nonphysicians. Circulation. 2014;130:1014-1027. 10.1161/CIR.0000000000000101 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14:377-381. https://pubmed.ncbi.nlm.nih.gov/7154893/ [PubMed] [Google Scholar]
- 22. Burr JF, Bredin SS, Faktor MD, Warburton DE. The 6-minute walk test as a predictor of objectively measured aerobic fitness in healthy working-aged adults. Phys Sportsmed. 2011;39:133-139. 10.3810/psm.2011.05.1904 [DOI] [PubMed] [Google Scholar]
- 23. Myers J, Kaminsky LA, Lima R, Christle JW, Ashley E, Arena R. A reference equation for normal standards for VO(2) max: analysis from the fitness registry and the importance of exercise national database (FRIEND Registry). Prog Cardiovasc Dis. 2017;60:21-29. 10.1016/j.pcad.2017.03.002 [DOI] [PubMed] [Google Scholar]
- 24. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited. J Am Coll Cardiol. 2001;37:153-156. 10.1016/s0735-1097(00)01054-8 [DOI] [PubMed] [Google Scholar]
- 25. Pettee Gabriel K, Whitaker KM, Duprez D, et al. Clinical importance of non-participation in a maximal graded exercise test on risk of non-fatal and fatal cardiovascular events and all-cause mortality: CARDIA study. Prev Med. 2018;106:137-144. 10.1016/j.ypmed.2017.10.025 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
CARDIA data are available upon reasonable request from the CARDIA Coordinating Center. CARDIA investigators are eager to collaborate with investigators interested in using CARDIA data. Please see the CARDIA website (https://sites.uab.edu/cardia/for-researchers/publications/) for publication policies and for a list of CARDIA Representatives. CARDIA data are also publicly available on the NIH-supported BioLINCC and dbGaP platforms.