- Email: [email protected]
New Generalized Equation for Predicting Maximal Oxygen Uptake (from the Fitness Registry and the Importance of Exercise National Database)
New Generalized Equation for Predicting Maximal Oxygen Uptake (from the Fitness Registry and the Importance of Exercise National Database) Peter Kokkinos, PhDa,b,c,d,*, Leonard A. Kaminsky, PhDe, Ross Arena, PhDf, Jiajia Zhang, PhDg, and Jonathan Myers, PhDh,i Impaired cardiorespiratory fitness (CRF) is closely linked to chronic illness and associated with adverse events. The American College of Sports Medicine (ACSM) regression equations (ACSM equations) developed to estimate oxygen uptake have known limitations leading to well-documented overestimation of CRF, especially at higher work rates. Thus, there is a need to explore alternative equations to more accurately predict CRF. We assessed maximal oxygen uptake (VO2 max) obtained directly by open-circuit spirometry in 7,983 apparently healthy subjects who participated in the Fitness Registry and the Importance of Exercise National Database (FRIEND). We randomly sampled 70% of the participants from each of the following age categories: <40, 40 to 50, 50 to 70, and ‡70 and used the remaining 30% for validation. Multivariable linear regression analysis was applied to identify the most relevant variables and construct the best prediction model for VO2 max. Treadmill speed and treadmill speed 3 grade were considered in the final model as predictors of measured VO2 max and the following equation was generated: VO2 max in ml O2/ kg/min [ speed (m/min) 3 (0.17 D fractional grade 3 0.79) D 3.5. The FRIEND equation predicted VO2 max with an overall error >4 times lower than the error associated with the traditional ACSM equations (5.1 – 18.3% vs 21.4 – 24.9%, respectively). Overestimation associated with the ACSM equation was accentuated when different protocols were considered separately. In conclusion, The FRIEND equation predicts VO2 max more precisely than the traditional ACSM equations with an overall error >4 times lower than that associated with the ACSM equations. Published by Elsevier Inc. (Am J Cardiol 2017;120:688e692) Cardiorespiratory fitness (CRF) is defined as the capacity of the cardiovascular and pulmonary systems to meet the oxygen demands of skeletal muscle during physical work. Accordingly, a subject’s highest oxygen uptake (VO2 max) during physical work represents maximal aerobic capacity. CRF is closely linked to risk for chronic disease and associated adverse events.1e4 Direct assessment of VO2 max is performed in a laboratory or clinical setting by open-circuit spirometry, requiring specialized equipment and trained personnel, limiting a widespread application. Thus, regression a Department of Cardiology, Veterans Affairs Medical Center, Washington, DC; bDepartment of Clinical Research and Leadership, George Washington University School of Medicine and Health Sciences, Washington, DC; cDepartment of Cardiology, Georgetown University School of Medicine, Washington, DC; dDepartment of Exercise Science, University of South Carolina, Arnold School of Public Health, Columbia, South Carolina; eFisher Institute of Health and Well-Being and Clinical Exercise Physiology, Ball State University, Muncie, Indiana; fDepartment of Physical Therapy and Integrative Physiology Laboratory, College of Applied Health Sciences, University of Illinois at Chicago, Chicago, Illinois; gDepartment of Epidemiology and Biostatistics, University of South Carolina, Columbia, South Carolina; hDepartment of Cardiology, Veterans Affairs Palo Alto Health Care System, Palo Alto, California; and iDepartment of Cardiology, Stanford University, Stanford, California. Manuscript received March 16, 2017; revised manuscript received and accepted May 17, 2017. See page 692 for disclosure information. *Corresponding author: Tel: (202) 745-8430; fax: (202) 745-2261. E-mail address: [email protected] (P. Kokkinos).
0002-9149/17/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.amjcard.2017.05.037
equations for walking and running speeds were developed and have been adapted over the years by the ACSM to estimate energy requirements at submaximal steady states (defined as a constant VO2) and subsequently used to estimate VO2 at maximal, nonesteady-state work rates during a progressive exercise test.5 This estimated metabolic demand for a given work rate has conventionally been expressed in metabolic equivalents (METs). One MET is considered to be equivalent to the amount of oxygen (O2) consumed at rest (w3.5 ml O2/ kg/min).6,7 However, the steady state assumed by these equations is influenced significantly by several factors including the subject’s age, health and fitness status, walking efficiency, and handrail use.7,8 In addition, the most commonly used equations for walking and running speeds were developed nearly 4 decades ago, using a specific protocol and were based on relatively few (<100), young (19 to 26 years old) participants in most studies. These factors contribute to the well-documented overestimation of VO2 max (MET levels) that occurs during a progressive exercise test.7,9 Because the primary indication for exercise testing is for clinical reasons,10 the objective of this study was to develop clinically applicable equations, derived from direct assessment of VO2 max to more accurately define peak MET levels for different exercise testing protocols. Methods Our cohort included 7,983 apparently healthy subjects (4,798 men and 3,183 women) who participated in The www.ajconline.org
Miscellaneous/Equation for Maximal Oxygen Uptake
Fitness Registry and the Importance of Exercise National Database (FRIEND), established in 2014. FRIEND is a multi-institutional initiative with the primary objective of establishing normative CRF values for the United States across the adult life span.11 Participants completed a graded exercise treadmill test in one of the 8 participating CPX laboratories (see Acknowledgment) with geographic representation from Connecticut, Indiana, Illinois, Louisiana, Maryland, North Carolina, Tennessee, and Texas. VO2 max was measured directly in all participants by open-circuit spirometry. The procedures used for acquiring and managing the FRIEND registry data have been previously reported. In brief, the CPX laboratories contributing data to FRIEND all used valid and reliable calibration and testing procedures and using experienced personnel qualified to conduct exercise tests. Although some variations in laboratory equipment, protocols, and procedures existed, the characteristics of all participating CPX laboratories are consistent with recommendations provided in recently guidelines.12,13 The indication for the exercise tests was determination of CRF before entry into an exercise program or research
689
was further investigated by the variable selection criteria including Akaike information criterion, corrected Akaike information criterion, the Sawa Bayesian Information Criterion-adjusted R-square statistic, the Predicted Residual Sum of Squares Statistic, and the average square error over the validation data (validation ASE). Results For each of the evaluation criteria, the precision of the model improved significantly when treadmill speed and the interaction of treadmill speed and grade were introduced in the model. The addition of the remaining variables did not change the model appreciably. This was confirmed by the average squared error that included 70% of the entire cohort and validation sample (the remaining 30%) for each variable entering the model (Figure 1). Consequently, treadmill speed and treadmill speed grade were the variables considered in the final model as predictors of measured VO2 max. Subsequently, the following equation was generated.
VO2 max in ml O2 =kg=min ¼ speed ð0:17 þ fractional grade 0:79Þ þ 3:5 study. Participant screening was specific to each laboratory’s procedures to rule out contraindications for exercise testing and for risk stratification. Inclusion criteria included age 20 years (mean age 47 13 years) and free from overt cardiovascular disease (coronary artery disease, peripheral artery disease, or heart failure) or chronic obstructive pulmonary disease at the time of the test and peak respiratory exchange ratio 1.0. Any tests terminated for abnormal clinical findings or before achieving voluntary maximal effort were excluded.14 Maximal effort was determined by a peak respiratory exchange ratio 1.0. We applied multivariable linear regression analysis to identify the most relevant variables associated with VO2 max to construct the best prediction model for VO2 max estimation. We randomly sampled 70% of the participants from the following age categories: <40, 40 to 50, 50 to 70, and 70 years. The remaining 30% of participants were used for validation. The variables considered in variable selection were (1) treadmill speed (m/min), (2) treadmill grade (%), (3) interaction between treadmill speed and grade, (4) heart rate at rest, (5) blood pressure at rest, (6) history of cardiovascular disease, (7) exercise time (minutes), (8) body mass index (BMI), (9) height, (10) age (years), and (11) gender. The stepwise selection was adopted and selection stopped when the candidate for entry had a p value >0.15 and the candidate for removal had a p value <0.15. During the selection process, models are constructed using 70% of the cohort. Then, the prediction errors for the models were found using the validation data (30% of the cohort), which were used to decide when to terminate the selection procedure. The order of variables that entered into the final model was (1) treadmill speed, (2) treadmill speed grade, (3) gender, (4) exercise time, (5) age, (6) systolic blood pressure at rest, and (7) BMI. The contribution of the variables selected to predict VO2 max
where speed is in meters per minute. Model G:F D BMI: Comparisons between the predicted values based on the ACSM equations and the FRIEND equation for the entire group and for each protocol are presented in Table 1. The ACSM equations overestimate VO2 max, as previously reported since the first edition of the ACSM guidelines5 with a percent error for the entire cohort of 21.4 24.9 versus 5.1 18.3 for the FRIEND equation. For different protocols using the ACSM equations, the percent error ranged from 32.0 5.1% for the modified Balke to 39.8 26.4% for the BSU/Bruce-Ramp protocol. In contrast, the percent error using the FRIEND equation ranged from 1.7 15.4 for the Bruce protocol to 12.8 21.0 for the Bruce-Ramp. Thus, overall, the FRIEND equation consistently outperformed the existing ACSM equations for predicting VO2 max for each protocol. Tables 2 to 6 present the MET levels per stage for different protocols estimated by the ACSM and FRIEND equations. Compared with the FRIEND protocol, the MET levels achieved are consistently overestimated by the ACSM equations when using the Bruce protocol and underestimated when using the Modified Bruce (Table 2). In general, for the remaining protocols (Balke-Ware, modified Naughton, modified Balke, and Ramp), workloads defined by MET levels 5.0 were underestimated by the ACSM equations, whereas those >7.0 METs are overestimated by the ACSM equations (Tables 3 to 6). Discussion Model G:F D BMI: In the present study, we derived a new equation (FRIEND equation) that estimates directly measured VO2 max more accurately than the currently used ACSM equations. The overall error of the FRIEND equation is >4 times lower than the error associated with
690
The American Journal of Cardiology (www.ajconline.org)
Figure 1. Average squared error change with variables considered in the multiple regression models (models A to G). The graph represents 70% of the entire cohort data set (open circles), and the remaining 30% (x), used for validation. The stepwise selection was used and variables with high contribution to VO2 max in ml O2/kg/min entered to the model first. Lower average squared error values indicate greater predictive value of the model. Note that the interaction between treadmill speed and grade (model B) are the strongest predictors of measure VO2 max.
Table 1 Measured and estimated oxygen consumption based on the American College of Sports Medicine (ACSM)* and the Fitness Registry and the Importance of Exercise National Database (FRIEND) equations (mean SD) Protocol BALKE BRUCE BRUCE-RAMP RAMP MODIFIED BALKE MODIFIED BRUCE MOD-NAUGHTON MANNUAL-I MANNUAL-II ENTIRE COHORT
N
Age
Measured Maximal Oxygen Uptake
FRIEND Maximal Oxygen Uptake
% Error FRIEND
ACSM Equations
% Error ACSM
353 936 2,224 230 108 38 407 3017 670 7,983
5414 4013 4813 5416 569 5314 579 4311 5615 4714
24.88.5 40.211.1 31.39.6 26.311.7 16.22.8 28.39.7 23.65.3 37.910.7 20.48.5 32.911.8
25.05.6 38.47.7 33.86.4 26.09.7 16.02.3 29.38.6 23.72.5 37.59.0 21.67.8 33.39.7
5.317.8 -1.715.4 12.821.0 4.324.4 -0.84.6 6.7 20.0 3.817.1 1.214.2 9.1 18.6 5.118.3
29.99.3 45.99.0 42.08.3 26.714.0 11.13.3 35.110.8 26.55.6 42.911.5 23.510.2 39.012.6
23.019.6 17.620.3 39.826.4 1.832.3 -32.05.1 27.324.9 14.016.1 14.417.2 15.820.1 21.424.9
* Walking and running speed equations were used.
the values obtained from the ACSM equations (5.1 18.3% vs 21.4 24.9%, respectively). The overestimation of the ACSM equations is accentuated when protocols are considered separately. In general, a common theme that was apparent for all protocols studied was that workloads associated with MET levels 5.0 were underestimated by the ACSM equations and those >7.0 METs were overestimated. The main factor that distinguishes the FRIEND equation and renders it superior to the ACSM equations in predicting VO2 max is that the FRIEND equation was constructed to predict a known value of directly measured VO2 max. In contrast, the ACSM equations predict VO2 max using an extrapolation of VO2 measured at a submaximal workload, assuming that a steady state
was achieved. However, steady state is influenced by a number of factors, including age. Because most subjects in these studies were relatively young (19 to 26 years old), VO2 max will be overestimated by the ACSM equations when applied to older populations. Moreover, steady state is rarely achievable during a progressive exercise test. Additional attributes of the FRIEND equation include the relatively large database of 7,983 participants, both men (n ¼ 4,798) and women (n ¼ 3,183) and black (n ¼ 395) and white (n ¼ 7,588), with a wide age range (20 to 91 years), derived from different regions across the United States. This allowed us to examine whether gender and race had a significant impact on the predictive power of the model. Finally, because the FRIEND equation is based on a
Miscellaneous/Equation for Maximal Oxygen Uptake
691
Table 2 Metabolic equivalent (MET) level estimates for the Bruce and modified Bruce protocols according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Table 4 Metabolic equivalent (MET) level estimates for the modified Naughton protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Stages
Stages
1 2 3 4 5 6 7
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM Equations FRIEND Equation 1.7 2.5 3.4 4.2 5.0 5.5 6.0
10 12 14 16 18 20 22
4.6 7.0 10.2 13.5 14.9 17.0 19.3
4.2 6.0 8.3 10.5 13.0 14.8 16.8
Modified Bruce Stages 1 2
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM equations FRIEND equation 1.7 1.7
0.0 0.5
2.3 3.5
3.2 3.7
1 2 3 4 5 6 7 8 9 10 11 12 13
Speed (miles/hour)
Grade (%)
Estimated MET Level ACSM equations
Estimated MET Level FRIEND equation
1.0 1.5 2.0 2.0 2.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
0 0 3.5 7.0 10.5 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
1.8 2.1 3.5 4.5 5.4 6.4 7.4 8.5 9.5 10.5 11.6 12.6 13.6
2.3 3.0 4.0 4.5 4.9 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4
Table 3 Metabolic equivalent (MET) level estimates for the Balke-Ware protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Table 5 Metabolic equivalent (MET) level estimates for the modified Blake protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Stages
Stages
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Speed (miles/hour)
Grade (%)
Estimated MET Level ACSM equations
Estimated MET Level FRIEND equation
3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 29 20 21.00 22.00 23.00 24.00 25.00 26.00
4.0 4.4 4.9 5.4 5.8 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4 9.9 10.4 10.8 11.3 11.7 12.2 12.6 13.1 13.5 14.0 14.5 14.9 15.4
5.5 5.7 5.9 6.1 6.3 6.5 6.7 6.9 7.1 7.3 7.5 7.7 7.9 8.1 8.3 8.5 8.7 8.9 9.1 9.3 9.5 9.7 9.9 10.1 10.3 10.5
1 2 3 4 5 6 7 8 9 10 11 12
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM equations FRIEND equation 2.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
.00 .00 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
2.5 3.3 4.3 5.4 6.4 7.4 8.5 9.5 10.5 11.6 12.6 13.6
3.6 4.9 5.4 5.8 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4
measured VO2 max value, the need for 2 equations to account for different treadmill speeds is eliminated. Collectively, these attributes make the FRIEND equation more applicable to different populations. In contrast, a potential limitation of the FRIEND equation is the exclusion of diseased populations in the FRIEND registry. In conclusion, the accuracy of the FRIEND equation for predicting VO2 max is superior to that associated with the ACSM equations (overall error >4 times lower). Because the majority of treadmill tests are performed for clinical reasons and the importance of CRF in the clinical setting, this overwhelming superiority warrants that manufacturers of exercise assessment apparatus (treadmills) and exercise
692
The American Journal of Cardiology (www.ajconline.org)
Table 6 Metabolic equivalent (MET) level estimates for the Ramp protocol According to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations Stages 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Speed (mph)
Grade (%)
Estimated MET Level ACSM equation
Estimated MET Level FRIEND equation
.50 1.0 1.5 2.0 2.5 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
.00 .00 .00 .00 .00 .00 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12 13 14 15 16 17 18 19 20 21 22 23 24 25
1.4 1.8 2.1 2.5 2.9 3.3 3.7 4.1 4.5 5.0 5.4 5.8 6.2 6.6 7.0 7.4 7.8 8.3 8.7 9.1 9.5 9.9 10.3 10.7 11.2 11.6 12.0 12.4 12.8 13.2 13.6
1.7 2.3 3.0 3.6 4.3 4.9 5.1 5.3 5.4 5.6 5.8 6.0 6.2 6.4 6.5 6.7 6.9 7.1 7.3 7.4 7.6 7.8 8.0 8.2 8.4 8.5 8.7 8.9 9.1 9.3 9.4
laboratories consider adapting the FRIEND equation for the estimation of VO2 max during exercise testing. Acknowledgment: Cardiorespiratory Fitness Registry Board Members: Co-Chairs: Lenny Kaminsky and Ross Arena, Paige Briggs, Peter Brubaker, Daniel Forman, Carol Garber, Carl J Lavie, Jonathan Myers, and Mahesh Patel. FRIEND Consortium contributors: Ball State University (Leonard Kaminsky), Brooke Army Medical Center (Kenneth Leclerc), Cone Health (Paul Chase), Johns Hopkins University (Kerry Stewart), Pennington Biomedical Research Center (Timothy Church), Southern Connecticut State University (Robert Axtell), Taylor University (Erik Hayes), University of Illinois-Chicago (Jacob Haus), University of Tennessee, Knoxville (David Bassett).
Disclosures The authors have no conflicts of interest to disclose.
1. Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE. Exercise capacity and mortality among men referred for exercise testing. N Engl J Med 2002;346:793e801. 2. Kokkinos PF, Faselis C, Myers J, Narayan P, Sui X, Zhang J, Lavie CJ, Moore H, Karasik P, Fletcher R. Cardiorespiratory fitness and incidence of major adverse cardiovascular events in US Veterans: a cohort study. Mayo Clin Proc 2017;92:39e48. 3. Kokkinos P, Myers J, Faselis C, Panagiotakos DB, Doumas M, Pittaras A, Manolis A, Kokkinos JP, Karasik P, Greenberg M, Papademetriou V, Fletcher R. Exercise capacity and mortality in older men: a 20-year follow-up study. Circulation 2010;122:790e797. 4. Kokkinos P, Myers J. Exercise and physical activity: clinical outcomes and applications. Circulation 2010;122:1637e1648. 5. ACSM Guidelines for Exercise Testing and Prescription 1st Edition. Philadelphia, PA: Lea and Fabiger, 1975. 6. Glass S, Dwyer GD, eds. ACSM Metabolic Calculations Handbook. Baltimore, MD: Lippincott Williams & Wilkins, 2007. 7. ACSM’s Guidelines for Exercise Testing and Prescription 10th Edition. Baltimore, MD: Wolters Kluwer, 2017. 8. Myers J, Kaminsky L, Lima R, Christle J, Ashley E, Arena R. A reference equation for for normal standards for VO2 max: analysis from the fitness registry and the importance of exercise database (FRIEND registry). Prog Cardiovasc Dis 2017. http://dx.doi.org/10. 1016/j.pcad.2017.03.002. 9. Arena R, Myers J, Williams MA, Gulati M, Kligfield P, Balady GJ, Collins E, Fletcher G; American Heart Association Committee on Exercise, Rehabilitation, and Prevention of the Council on Clinical Cardiology; American Heart Association Council on Cardiovascular Nursing. Assessment of functional capacity in clinical and research settings: a scientific statement from the American Heart Association Committee on Exercise, Rehabilitation, and Prevention of the Council on Clinical Cardiology and the Council on Cardiovascular Nursing. Circulation 2007;116:329e343. 10. Ross R, Blair SN, Arena R, Church TS, Despres JP, Franklin BA, Haskell WL, Kaminsky LA, Levine BD, Lavie CJ, Myers J, Niebauer J, Sallis R, Sawada SS, Sui X, Wisloff U; American Heart Association Physical Activity Committee of the Council on Lifestyle and Cardiometabolic Health; Council on Clinical Cardiology; Council on Epidemiology and Prevention; Council on Cardiovascular and Stroke Nursing; Council on Functional Genomics and Translational Biology; 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:e653ee699. 11. Kaminsky LA, Arena R, Beckie TM, Brubaker PH, Church TS, Forman DE, Franklin BA, Gulati M, Lavie CJ, Myers J, Patel MJ, Pina IL, Weintraub WS, Williams MA; American Heart Association Advocacy Coordinating Committee, Council on Clinical Cardiology, and Council on Nutrition, Physical Activity and Metabolism. The importance of cardiorespiratory fitness in the United States: the need for a national registry: a policy statement from the American Heart Association. Circulation 2013;127:652e662. 12. Myers J, Forman DE, Balady GJ, Franklin BA, Nelson-Worel J, Martin BJ, Herbert WG, Guazzi M, Arena R; American Heart Association Subcommittee on Exercise, Cardiac Rehabilitation, and Prevention of the Council on Clinical Cardiology, Council on Lifestyle and Cardiometabolic Health, Council on Epidemiology and Prevention, and Council on Cardiovascular and Stroke Nursing. Supervision of exercise testing by nonphysicians: a scientific statement from the American Heart Association. Circulation 2014;130:1014e1027. 13. Myers J, Arena R, Franklin B, Pina I, Kraus WE, McInnis K, Balady GJ; American Heart Association Committee on Exercise, Cardiac Rehabilitation, and Prevention of the Council on Clinical Cardiology, the Council on Nutrition, Physical Activity, and Metabolism, and the Council on Cardiovascular Nursing. Recommendations for clinical exercise laboratories: a scientific statement from the American Heart Association. Circulation 2009;119:3144e3161. 14. Kaminsky LA, Arena R, Myers J. Reference standards for cardiorespiratory fitness measured with cardiopulmonary exercise testing: data from the Fitness Registry and the Importance of Exercise National Database. Mayo Clin Proc 2015;90:1515e1523.
0002-9149/17/$ - see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.amjcard.2017.05.037
equations for walking and running speeds were developed and have been adapted over the years by the ACSM to estimate energy requirements at submaximal steady states (defined as a constant VO2) and subsequently used to estimate VO2 at maximal, nonesteady-state work rates during a progressive exercise test.5 This estimated metabolic demand for a given work rate has conventionally been expressed in metabolic equivalents (METs). One MET is considered to be equivalent to the amount of oxygen (O2) consumed at rest (w3.5 ml O2/ kg/min).6,7 However, the steady state assumed by these equations is influenced significantly by several factors including the subject’s age, health and fitness status, walking efficiency, and handrail use.7,8 In addition, the most commonly used equations for walking and running speeds were developed nearly 4 decades ago, using a specific protocol and were based on relatively few (<100), young (19 to 26 years old) participants in most studies. These factors contribute to the well-documented overestimation of VO2 max (MET levels) that occurs during a progressive exercise test.7,9 Because the primary indication for exercise testing is for clinical reasons,10 the objective of this study was to develop clinically applicable equations, derived from direct assessment of VO2 max to more accurately define peak MET levels for different exercise testing protocols. Methods Our cohort included 7,983 apparently healthy subjects (4,798 men and 3,183 women) who participated in The www.ajconline.org
Miscellaneous/Equation for Maximal Oxygen Uptake
Fitness Registry and the Importance of Exercise National Database (FRIEND), established in 2014. FRIEND is a multi-institutional initiative with the primary objective of establishing normative CRF values for the United States across the adult life span.11 Participants completed a graded exercise treadmill test in one of the 8 participating CPX laboratories (see Acknowledgment) with geographic representation from Connecticut, Indiana, Illinois, Louisiana, Maryland, North Carolina, Tennessee, and Texas. VO2 max was measured directly in all participants by open-circuit spirometry. The procedures used for acquiring and managing the FRIEND registry data have been previously reported. In brief, the CPX laboratories contributing data to FRIEND all used valid and reliable calibration and testing procedures and using experienced personnel qualified to conduct exercise tests. Although some variations in laboratory equipment, protocols, and procedures existed, the characteristics of all participating CPX laboratories are consistent with recommendations provided in recently guidelines.12,13 The indication for the exercise tests was determination of CRF before entry into an exercise program or research
689
was further investigated by the variable selection criteria including Akaike information criterion, corrected Akaike information criterion, the Sawa Bayesian Information Criterion-adjusted R-square statistic, the Predicted Residual Sum of Squares Statistic, and the average square error over the validation data (validation ASE). Results For each of the evaluation criteria, the precision of the model improved significantly when treadmill speed and the interaction of treadmill speed and grade were introduced in the model. The addition of the remaining variables did not change the model appreciably. This was confirmed by the average squared error that included 70% of the entire cohort and validation sample (the remaining 30%) for each variable entering the model (Figure 1). Consequently, treadmill speed and treadmill speed grade were the variables considered in the final model as predictors of measured VO2 max. Subsequently, the following equation was generated.
VO2 max in ml O2 =kg=min ¼ speed ð0:17 þ fractional grade 0:79Þ þ 3:5 study. Participant screening was specific to each laboratory’s procedures to rule out contraindications for exercise testing and for risk stratification. Inclusion criteria included age 20 years (mean age 47 13 years) and free from overt cardiovascular disease (coronary artery disease, peripheral artery disease, or heart failure) or chronic obstructive pulmonary disease at the time of the test and peak respiratory exchange ratio 1.0. Any tests terminated for abnormal clinical findings or before achieving voluntary maximal effort were excluded.14 Maximal effort was determined by a peak respiratory exchange ratio 1.0. We applied multivariable linear regression analysis to identify the most relevant variables associated with VO2 max to construct the best prediction model for VO2 max estimation. We randomly sampled 70% of the participants from the following age categories: <40, 40 to 50, 50 to 70, and 70 years. The remaining 30% of participants were used for validation. The variables considered in variable selection were (1) treadmill speed (m/min), (2) treadmill grade (%), (3) interaction between treadmill speed and grade, (4) heart rate at rest, (5) blood pressure at rest, (6) history of cardiovascular disease, (7) exercise time (minutes), (8) body mass index (BMI), (9) height, (10) age (years), and (11) gender. The stepwise selection was adopted and selection stopped when the candidate for entry had a p value >0.15 and the candidate for removal had a p value <0.15. During the selection process, models are constructed using 70% of the cohort. Then, the prediction errors for the models were found using the validation data (30% of the cohort), which were used to decide when to terminate the selection procedure. The order of variables that entered into the final model was (1) treadmill speed, (2) treadmill speed grade, (3) gender, (4) exercise time, (5) age, (6) systolic blood pressure at rest, and (7) BMI. The contribution of the variables selected to predict VO2 max
where speed is in meters per minute. Model G:F D BMI: Comparisons between the predicted values based on the ACSM equations and the FRIEND equation for the entire group and for each protocol are presented in Table 1. The ACSM equations overestimate VO2 max, as previously reported since the first edition of the ACSM guidelines5 with a percent error for the entire cohort of 21.4 24.9 versus 5.1 18.3 for the FRIEND equation. For different protocols using the ACSM equations, the percent error ranged from 32.0 5.1% for the modified Balke to 39.8 26.4% for the BSU/Bruce-Ramp protocol. In contrast, the percent error using the FRIEND equation ranged from 1.7 15.4 for the Bruce protocol to 12.8 21.0 for the Bruce-Ramp. Thus, overall, the FRIEND equation consistently outperformed the existing ACSM equations for predicting VO2 max for each protocol. Tables 2 to 6 present the MET levels per stage for different protocols estimated by the ACSM and FRIEND equations. Compared with the FRIEND protocol, the MET levels achieved are consistently overestimated by the ACSM equations when using the Bruce protocol and underestimated when using the Modified Bruce (Table 2). In general, for the remaining protocols (Balke-Ware, modified Naughton, modified Balke, and Ramp), workloads defined by MET levels 5.0 were underestimated by the ACSM equations, whereas those >7.0 METs are overestimated by the ACSM equations (Tables 3 to 6). Discussion Model G:F D BMI: In the present study, we derived a new equation (FRIEND equation) that estimates directly measured VO2 max more accurately than the currently used ACSM equations. The overall error of the FRIEND equation is >4 times lower than the error associated with
690
The American Journal of Cardiology (www.ajconline.org)
Figure 1. Average squared error change with variables considered in the multiple regression models (models A to G). The graph represents 70% of the entire cohort data set (open circles), and the remaining 30% (x), used for validation. The stepwise selection was used and variables with high contribution to VO2 max in ml O2/kg/min entered to the model first. Lower average squared error values indicate greater predictive value of the model. Note that the interaction between treadmill speed and grade (model B) are the strongest predictors of measure VO2 max.
Table 1 Measured and estimated oxygen consumption based on the American College of Sports Medicine (ACSM)* and the Fitness Registry and the Importance of Exercise National Database (FRIEND) equations (mean SD) Protocol BALKE BRUCE BRUCE-RAMP RAMP MODIFIED BALKE MODIFIED BRUCE MOD-NAUGHTON MANNUAL-I MANNUAL-II ENTIRE COHORT
N
Age
Measured Maximal Oxygen Uptake
FRIEND Maximal Oxygen Uptake
% Error FRIEND
ACSM Equations
% Error ACSM
353 936 2,224 230 108 38 407 3017 670 7,983
5414 4013 4813 5416 569 5314 579 4311 5615 4714
24.88.5 40.211.1 31.39.6 26.311.7 16.22.8 28.39.7 23.65.3 37.910.7 20.48.5 32.911.8
25.05.6 38.47.7 33.86.4 26.09.7 16.02.3 29.38.6 23.72.5 37.59.0 21.67.8 33.39.7
5.317.8 -1.715.4 12.821.0 4.324.4 -0.84.6 6.7 20.0 3.817.1 1.214.2 9.1 18.6 5.118.3
29.99.3 45.99.0 42.08.3 26.714.0 11.13.3 35.110.8 26.55.6 42.911.5 23.510.2 39.012.6
23.019.6 17.620.3 39.826.4 1.832.3 -32.05.1 27.324.9 14.016.1 14.417.2 15.820.1 21.424.9
* Walking and running speed equations were used.
the values obtained from the ACSM equations (5.1 18.3% vs 21.4 24.9%, respectively). The overestimation of the ACSM equations is accentuated when protocols are considered separately. In general, a common theme that was apparent for all protocols studied was that workloads associated with MET levels 5.0 were underestimated by the ACSM equations and those >7.0 METs were overestimated. The main factor that distinguishes the FRIEND equation and renders it superior to the ACSM equations in predicting VO2 max is that the FRIEND equation was constructed to predict a known value of directly measured VO2 max. In contrast, the ACSM equations predict VO2 max using an extrapolation of VO2 measured at a submaximal workload, assuming that a steady state
was achieved. However, steady state is influenced by a number of factors, including age. Because most subjects in these studies were relatively young (19 to 26 years old), VO2 max will be overestimated by the ACSM equations when applied to older populations. Moreover, steady state is rarely achievable during a progressive exercise test. Additional attributes of the FRIEND equation include the relatively large database of 7,983 participants, both men (n ¼ 4,798) and women (n ¼ 3,183) and black (n ¼ 395) and white (n ¼ 7,588), with a wide age range (20 to 91 years), derived from different regions across the United States. This allowed us to examine whether gender and race had a significant impact on the predictive power of the model. Finally, because the FRIEND equation is based on a
Miscellaneous/Equation for Maximal Oxygen Uptake
691
Table 2 Metabolic equivalent (MET) level estimates for the Bruce and modified Bruce protocols according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Table 4 Metabolic equivalent (MET) level estimates for the modified Naughton protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Stages
Stages
1 2 3 4 5 6 7
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM Equations FRIEND Equation 1.7 2.5 3.4 4.2 5.0 5.5 6.0
10 12 14 16 18 20 22
4.6 7.0 10.2 13.5 14.9 17.0 19.3
4.2 6.0 8.3 10.5 13.0 14.8 16.8
Modified Bruce Stages 1 2
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM equations FRIEND equation 1.7 1.7
0.0 0.5
2.3 3.5
3.2 3.7
1 2 3 4 5 6 7 8 9 10 11 12 13
Speed (miles/hour)
Grade (%)
Estimated MET Level ACSM equations
Estimated MET Level FRIEND equation
1.0 1.5 2.0 2.0 2.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
0 0 3.5 7.0 10.5 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
1.8 2.1 3.5 4.5 5.4 6.4 7.4 8.5 9.5 10.5 11.6 12.6 13.6
2.3 3.0 4.0 4.5 4.9 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4
Table 3 Metabolic equivalent (MET) level estimates for the Balke-Ware protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Table 5 Metabolic equivalent (MET) level estimates for the modified Blake protocol according to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations
Stages
Stages
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
Speed (miles/hour)
Grade (%)
Estimated MET Level ACSM equations
Estimated MET Level FRIEND equation
3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3 3.3
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 29 20 21.00 22.00 23.00 24.00 25.00 26.00
4.0 4.4 4.9 5.4 5.8 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4 9.9 10.4 10.8 11.3 11.7 12.2 12.6 13.1 13.5 14.0 14.5 14.9 15.4
5.5 5.7 5.9 6.1 6.3 6.5 6.7 6.9 7.1 7.3 7.5 7.7 7.9 8.1 8.3 8.5 8.7 8.9 9.1 9.3 9.5 9.7 9.9 10.1 10.3 10.5
1 2 3 4 5 6 7 8 9 10 11 12
Speed Grade Estimated MET Level Estimated MET Level (miles/hour) (%) ACSM equations FRIEND equation 2.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
.00 .00 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0
2.5 3.3 4.3 5.4 6.4 7.4 8.5 9.5 10.5 11.6 12.6 13.6
3.6 4.9 5.4 5.8 6.3 6.7 7.2 7.6 8.1 8.5 9.0 9.4
measured VO2 max value, the need for 2 equations to account for different treadmill speeds is eliminated. Collectively, these attributes make the FRIEND equation more applicable to different populations. In contrast, a potential limitation of the FRIEND equation is the exclusion of diseased populations in the FRIEND registry. In conclusion, the accuracy of the FRIEND equation for predicting VO2 max is superior to that associated with the ACSM equations (overall error >4 times lower). Because the majority of treadmill tests are performed for clinical reasons and the importance of CRF in the clinical setting, this overwhelming superiority warrants that manufacturers of exercise assessment apparatus (treadmills) and exercise
692
The American Journal of Cardiology (www.ajconline.org)
Table 6 Metabolic equivalent (MET) level estimates for the Ramp protocol According to the American College of Sports Medicine (ACSM) and Fitness Registry and the Importance of Exercise National Database (FRIEND) equations Stages 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
Speed (mph)
Grade (%)
Estimated MET Level ACSM equation
Estimated MET Level FRIEND equation
.50 1.0 1.5 2.0 2.5 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0 3.0
.00 .00 .00 .00 .00 .00 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12 13 14 15 16 17 18 19 20 21 22 23 24 25
1.4 1.8 2.1 2.5 2.9 3.3 3.7 4.1 4.5 5.0 5.4 5.8 6.2 6.6 7.0 7.4 7.8 8.3 8.7 9.1 9.5 9.9 10.3 10.7 11.2 11.6 12.0 12.4 12.8 13.2 13.6
1.7 2.3 3.0 3.6 4.3 4.9 5.1 5.3 5.4 5.6 5.8 6.0 6.2 6.4 6.5 6.7 6.9 7.1 7.3 7.4 7.6 7.8 8.0 8.2 8.4 8.5 8.7 8.9 9.1 9.3 9.4
laboratories consider adapting the FRIEND equation for the estimation of VO2 max during exercise testing. Acknowledgment: Cardiorespiratory Fitness Registry Board Members: Co-Chairs: Lenny Kaminsky and Ross Arena, Paige Briggs, Peter Brubaker, Daniel Forman, Carol Garber, Carl J Lavie, Jonathan Myers, and Mahesh Patel. FRIEND Consortium contributors: Ball State University (Leonard Kaminsky), Brooke Army Medical Center (Kenneth Leclerc), Cone Health (Paul Chase), Johns Hopkins University (Kerry Stewart), Pennington Biomedical Research Center (Timothy Church), Southern Connecticut State University (Robert Axtell), Taylor University (Erik Hayes), University of Illinois-Chicago (Jacob Haus), University of Tennessee, Knoxville (David Bassett).
Disclosures The authors have no conflicts of interest to disclose.
1. Myers J, Prakash M, Froelicher V, Do D, Partington S, Atwood JE. Exercise capacity and mortality among men referred for exercise testing. N Engl J Med 2002;346:793e801. 2. Kokkinos PF, Faselis C, Myers J, Narayan P, Sui X, Zhang J, Lavie CJ, Moore H, Karasik P, Fletcher R. Cardiorespiratory fitness and incidence of major adverse cardiovascular events in US Veterans: a cohort study. Mayo Clin Proc 2017;92:39e48. 3. Kokkinos P, Myers J, Faselis C, Panagiotakos DB, Doumas M, Pittaras A, Manolis A, Kokkinos JP, Karasik P, Greenberg M, Papademetriou V, Fletcher R. Exercise capacity and mortality in older men: a 20-year follow-up study. Circulation 2010;122:790e797. 4. Kokkinos P, Myers J. Exercise and physical activity: clinical outcomes and applications. Circulation 2010;122:1637e1648. 5. ACSM Guidelines for Exercise Testing and Prescription 1st Edition. Philadelphia, PA: Lea and Fabiger, 1975. 6. Glass S, Dwyer GD, eds. ACSM Metabolic Calculations Handbook. Baltimore, MD: Lippincott Williams & Wilkins, 2007. 7. ACSM’s Guidelines for Exercise Testing and Prescription 10th Edition. Baltimore, MD: Wolters Kluwer, 2017. 8. Myers J, Kaminsky L, Lima R, Christle J, Ashley E, Arena R. A reference equation for for normal standards for VO2 max: analysis from the fitness registry and the importance of exercise database (FRIEND registry). Prog Cardiovasc Dis 2017. http://dx.doi.org/10. 1016/j.pcad.2017.03.002. 9. Arena R, Myers J, Williams MA, Gulati M, Kligfield P, Balady GJ, Collins E, Fletcher G; American Heart Association Committee on Exercise, Rehabilitation, and Prevention of the Council on Clinical Cardiology; American Heart Association Council on Cardiovascular Nursing. Assessment of functional capacity in clinical and research settings: a scientific statement from the American Heart Association Committee on Exercise, Rehabilitation, and Prevention of the Council on Clinical Cardiology and the Council on Cardiovascular Nursing. Circulation 2007;116:329e343. 10. Ross R, Blair SN, Arena R, Church TS, Despres JP, Franklin BA, Haskell WL, Kaminsky LA, Levine BD, Lavie CJ, Myers J, Niebauer J, Sallis R, Sawada SS, Sui X, Wisloff U; American Heart Association Physical Activity Committee of the Council on Lifestyle and Cardiometabolic Health; Council on Clinical Cardiology; Council on Epidemiology and Prevention; Council on Cardiovascular and Stroke Nursing; Council on Functional Genomics and Translational Biology; 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:e653ee699. 11. Kaminsky LA, Arena R, Beckie TM, Brubaker PH, Church TS, Forman DE, Franklin BA, Gulati M, Lavie CJ, Myers J, Patel MJ, Pina IL, Weintraub WS, Williams MA; American Heart Association Advocacy Coordinating Committee, Council on Clinical Cardiology, and Council on Nutrition, Physical Activity and Metabolism. The importance of cardiorespiratory fitness in the United States: the need for a national registry: a policy statement from the American Heart Association. Circulation 2013;127:652e662. 12. Myers J, Forman DE, Balady GJ, Franklin BA, Nelson-Worel J, Martin BJ, Herbert WG, Guazzi M, Arena R; American Heart Association Subcommittee on Exercise, Cardiac Rehabilitation, and Prevention of the Council on Clinical Cardiology, Council on Lifestyle and Cardiometabolic Health, Council on Epidemiology and Prevention, and Council on Cardiovascular and Stroke Nursing. Supervision of exercise testing by nonphysicians: a scientific statement from the American Heart Association. Circulation 2014;130:1014e1027. 13. Myers J, Arena R, Franklin B, Pina I, Kraus WE, McInnis K, Balady GJ; American Heart Association Committee on Exercise, Cardiac Rehabilitation, and Prevention of the Council on Clinical Cardiology, the Council on Nutrition, Physical Activity, and Metabolism, and the Council on Cardiovascular Nursing. Recommendations for clinical exercise laboratories: a scientific statement from the American Heart Association. Circulation 2009;119:3144e3161. 14. Kaminsky LA, Arena R, Myers J. Reference standards for cardiorespiratory fitness measured with cardiopulmonary exercise testing: data from the Fitness Registry and the Importance of Exercise National Database. Mayo Clin Proc 2015;90:1515e1523.