1184 Training & Testing
A Maximal Cycle Test with Good Validity and High Repeatability in Adults of All Ages
Affiliations
Key words
▶ cardiorespiratory fitness ● ▶ reproducibility ● ▶ exercise test ● ▶ physical fitness ● ▶ cycle ergometer ● ▶ measurement ●
L. Eriksen1, J. S. Tolstrup1, S. Larsen2, M. Grønbæk1, J. W. Helge2 1 2
National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark Centre of Healthy Aging, Department of Biomedical Sciences, University of Copenhagen, Copenhagen N, Denmark
Abstract
▼
In 11 680 individuals (18–85 years) maximal oxygen consumption (VO2max) was estimated indirectly in a maximal cycle test using a prediction model developed in a young population (15–28 years). A subsample of 182 individuals (23–77 years) underwent 2 maximal cycle tests with VO2max estimated indirectly in both tests and measured directly in one test. Agreement between the direct measurement and the indirect estimate of VO2max and repeatability of the indirect estimates of VO2max were examined by Bland-Altman plots, limits of agreement (LOA) and coefficient of repeatability (CR). The indi-
Introduction
▼ accepted after revision April 22, 2014 Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1376973 Published online: September 26, 2014 Int J Sports Med 2014; 35: 1184–1189 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0172-4622 Correspondence Louise Eriksen National Institute of Public Health University of Southern Denmark Øster Farimagsgade 5A 1353 Copenhagen Denmark Tel.: + 45/655/07 728 Fax: + 45/392/08 010 [email protected]
Maximal oxygen consumption (VO2max) is the criterion measure of cardiorespiratory fitness, which is an independent predictor of morbidity and mortality [16]. Direct measurement of VO2max implies measurement of oxygen and carbon dioxide in expired air and is considered the gold standard physiological test [1]. Direct measurement is time-consuming and requires expensive equipment and specific expertise. Thus, several indirect and simpler procedures based on heart rate during exercise [5], covered distance for a given time [10], maximal speed in shuttle run [19, 20] and maximal time and/or power on treadmill or ergometer cycle [3, 4, 9, 14, 29] have been developed to predict VO2max. Indirect tests may be maximal (performed to voluntary exhaustion) or sub-maximal, the latter considered less precise. In the Danish Health Examination Survey 2007–2008 (DANHES) [11] 11 680 participants ages 18–85 years performed a maximal cycle exercise test. VO2max was estimated indirectly based on maximal power output (MPO) obtained by the participants in the cycle test. The prediction model of
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
rect method (mean VO2max = 3 132 ml · min − 1) underestimated VO2max as compared to the direct method (mean VO2max = 3 190 ml · min − 1) in men (bias: 58 ml · min − 1 (95 % LOA − 450 and 565)) and overestimated VO2max in women (mean VO2max = 2 328 vs. 2 258 ml · min − 1, bias: − 70 ml · min − 1 (95 % LOA − 468 and 328)). The mean difference between the 2 indirect estimates was non-significant (men: − 11.9 ml · min − 1, women: 18.3 ml · min − 1) with a CR of 279 ml · min − 1 (8.9 %) in men and 274 ml · min − 1 (11.7 %) in women. The validity of the indirect method was good despite minor sex-specific bias. Owing to this bias we suggest a new prediction model of VO2max. The maximal cycle test was highly repeatable.
VO2max was originally developed by Andersen [4] from directly measured VO2max among 535 men and women ages 15–28 years. Nonetheless, this model might not be suitable for our population with a wider age range, since age has been found to contribute significantly to the prediction of VO2max independent of MPO with lower values of VO2max for older ages [4, 25]. In large population studies like DANHES, it is inevitable that physiological tests are performed under the instruction and guidance of several test leaders due to the huge number of participants. This will most likely induce a higher variation in the results, and it is therefore important to investigate whether such tests have high repeatability, even when being carried out under the guidance of different test leaders. The objective of the present study was to assess the agreement between VO2max estimated indirectly (applying a prediction model originally developed in adolescents and young adults) and measured directly (gold standard) among an adult population spanning a wide age range. A second objective was to test the repeatability of the maximal cycle test under the guidance of different test leaders.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Authors
Methods
▼
Study population A random subsample of adult citizens 18 years and older (n = 180 103) from 13 Danish municipalities was invited to participate in the Danish Health Examination Survey in 2007–2008 [11]. A total of 18 065 individuals (10 %) accepted the invitation, and based on a screening interview on health status and results of blood pressure measurements participants were eligible for participation in a maximal exercise test (n = 11 680), a sub-maximal test (n = 4 403) or no test (n = 1 982) for indirect estimation of maximal oxygen consumption (VO2max). Out of those 11 680 participants eligible for the maximal test 10 973 (94 %) completed the test with a valid test result. In 3 municipalities a sexand age-stratified sample of participants (n = 182) who underwent the maximal test with a valid test result were invited to come back after a week to perform a retest with indirect estimation as well as direct measurement of VO2max. Informed consent was obtained from all participants, and the protocol was reviewed by the Scientific Ethical Committee B for the Capital Region of Denmark (H-B-2007-050). The study was conducted in accordance with the ethical standards of the International Journal of Sports Medicine [13].
Measurement of VO2max In order to ensure safe participation and achieve optimal risk minimisation in the maximal cycle, test participants were subjected to a screening by means of an interview and measurement of blood pressure. In brief, one or more of the following conditions contraindicated participation in the cycle test: any heart-related disease; chest pain or pressure; moderate hypertension; consumption of antihypertensives; cardiac or pulmonary medication; pregnancy; muscle, joint or skeletal problems. On the first test day (test day 1) VO2max was estimated indirectly using an ergometer cycle (Ergomedic 839E, Monark Exercise AB, Vansbro, Sweden). Heart rate was monitored with chest strap and watch (Polar RS100, Polar Electro Oy, Kempele, Finland). The test started with a 5-min warm-up at a load of 75 watts for women and 100 watts for men. After the warm-up the work load was increased by 35 watts every 2 min until voluntary exhaustion. The size of the initial load and increments were chosen given the notion that younger as well as elderly participants should be able to complete the test within a reasonable time frame [2, 21]. Pedal frequency was freely chosen by participants within the limits 60–80 rpm, while trained participants were allowed to exceed 80 rpm. The participants were verbally encouraged to continue for as long as possible, and the perceived exertion of the participants was evaluated using Borg’s scale [8]. VO2max was estimated from MPO, that is, the highest achieved load adjusted for the proportion of the last stage completed: VO2max (l · min − 1) = 0.16 + (0.0117 · MPO (watt)) [4]. The maximal cycle tests were managed by trained scientific personnel. On the second test day (test day 2) the following week a similar test procedure was followed, yet on a different exercise ergometer cycle of the same model as on the first test day (Ergomedic 839E, Monark Exercise AB, Vansbro, Sweden). During this cycle test direct measurement of pulmonary VO2 as well as indirect estimation of VO2max was performed, enabling assessment of repeatability of the cycle test on 2 different test days as well as
agreement between the indirect and direct measurement conducted in one and the same test. The measurements were performed continuously using an automated metabolic cart (Quark b2, Cosmed Srl., Rome, Italy), calibrated before each test according to the manufacturer’s specifications. Before each test a volume calibration and a calibration of the gas analysers were performed using gases of known composition. The respiratory variables were averaged every 20 s. The greatest 20-s averaged VO2 (l · min − 1) value during the test was taken as the VO2max. Achievement of VO2max was accepted when a respiratory exchange ratio of 1.15 and/or maximal heart rate (220 − age) were present, or a levelling off or decline in VO2 was reached [15, 22]. All tests on test day 2 were managed by one experienced and formally qualified exercise physiologist. No attempt was made to control the participants’ food and beverage intake and physical activity prior to the tests.
Statistics To obtain a visual assessment of the relationship between the different measurements of VO2max, scatterplots and Pearson’s correlation coefficient were used. Bland-Altman plots [7] were used to illustrate the extent of agreement between the direct and indirect measurement of VO2max in the maximal cycle test on test day 2 by plotting the difference between the direct and indirect measurements for each individual against their direct measurement (the gold stand) instead of against the mean of the direct and indirect measurements, since the gold standard is expected to be closer to the “true value” [17]. The mean difference (bias) represents the degree of systematic error between the 2 measurements [7]. The hypothesis of zero bias was tested using a paired t-test. Stepwise multiple regression analysis with the direct measurement as the dependent variable was performed to derive a new equation for indirect estimation of VO2max. The repeatability of the cycle test was evaluated by the coefficient of repeatability (CR) (CR = 1.96 · standard deviation (SD) of the mean difference between the 2 indirect estimates of VO2max) and Bland-Altman plots with the difference between estimates by the 2 indirect methods for each individual against their mean value of the 2 indirect estimates. The significance level was set at P < 0.05. Statistics were calculated using STATA/IC 12.
Results
▼
Participants A total of 182 participants (91 women and 91 men) were enrolled in the present study. The results from 4 participants were excluded from the analyses due to unreliable direct test measurements resulting from technical problems with the equipment. Thus, data from 88 men and 90 women who all achieved one of the 3 cut-off criteria, were analysed. The study population is characterised by means and SD or ranges in ▶ Table 1. As expected, for the indirect estimates as well as the ● direct measurements of VO2max, mean VO2max was higher for men compared to women, e. g. 3 120 ml · min − 1 and 2 347 ml · min − 1, respectively, from the indirect estimate on test day 1. Indirectly and directly determined VO2max in men and women on the 2 different test days across age groups are illustrated ▶ Fig. 1. Overall, a lower VO max was observed with increasin ● 2 ing age in men and women.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Training & Testing 1185
1186 Training & Testing
Test day 1 age (years) weight (kg) height (cm) BMI (kg · m − 2) indirect VO2max (ml · min − 1) indirect VO2max (ml · min − 1 · kg − 1) MPO (Watt) maximal heart rate (bpm) direct VO2max (ml · min − 1) direct VO2max (ml · min − 1 · kg − 1)
Test day 2
48.8 (23–77) 80.3 (9.7) 180 (6) 24.7 (2.8) 3 120 (625) 39.2 (8.3) 253 (56) 174 (16)
3 132 (655) 39.4 (8.9) 254 (56) 175 (14) 3 190 (637) 40.1 (8.3)
Test day 1
Test day 2
44.4 (22–72) 66.9 (9.7) 168 (6) 23.8 (3.0) 2 347 (451) 35.5 (7.9) 185 (41) 174 (14)
2 328 (481) 35.2 (7.4) 185 (41) 176 (13) 2 258 (459) 34.1 (6.5)
Men
Men
3 800
Table 1 Characteristics of the population. Means and standard deviations (SD) or range.
Women (n = 90)
Women
5 000
3 600 Indirect VO2max (ml min –1)
3 400
3 000 2 800 2 600 2 400
4 000
3 000
Pearson’s r =0.91
2 200
22–29
30–39
40–49 Age (years)
n=4
n=24
n=21
Indirect VO2max day 1
50–59
60–77
n=16
n=23
Direct VO2max day 2
Indirect VO2max day 2 Women
3 800 3 600 3 400 3 200 3 000 2 800 2 600 2 400 2 200 2 000 1 800 1 600 1 400
00 50
00 40
00 30
00 20
00
00
10
50
00 40
00
Direct VO2max (ml min –1)
1 600 1 400
30
00 10
1 800
00
1 000
2 000
VO2max (ml min –1)
Pearson’s r = 0.92
2 000
20
VO2max (ml min –1)
3 200
22–29
30–39
n=13
n=24
40–49 Age (years)
Indirect VO2max day 1
n=22
50–59
60–77
n=19
n=12
Direct VO2max day 2
Indirect VO2max day 2
Fig. 1 VO2max (ml · min − 1) determined indirectly and directly in men and women on separate test days across age groups. Data are mean ± SEM.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
Fig. 2 Scatter plot indirect vs. direct measurement of VO2max (ml · min − 1) for men (n = 88) and women (n = 90).
Validity A scatter plot of the indirect estimate vs. the direct measurement of VO2max from test day 2 showed that data for both men and women were near the line of equality, and Pearson’s correla▶ Fig. 2). tion coefficient was 0.92 for men and 0.91 for women (● ▶ Fig. 3. Bias was positive in Bland-Altman plots are shown in ● men (58 ml · min − 1) and negative in women ( − 70 ml · min − 1). The agreement between the 2 methods for an individual was considered by the degree of random error indicated by the 95 % limits of agreement (LOA) in the Bland-Altman plot. Thus, since differences between the 2 measurements were normally distributed, 95 % of the differences fell within the range of − 450 and 565 ml · min − 1 for men and − 468 and 328 ml · min − 1 for women. The results of a paired t-test on the indirect estimate and direct measurement showed that the mean difference was significantly different from zero for men as well as women (58 ml · min − 1 (95 % confidence interval (CI): 3–113) and − 70 ml · min − 1 (95 % CI: − 112–( − 27)). Due to this systematic bias we made a new prediction model derived from a multiple linear regression with direct VO2max as the dependent variable. The model initially included MPO (watt), sex, age, body weight (kg), and the interaction sex · MPO. Age and the interaction sex · MPO were found to be non-significant. From the results of the final regression analy▶ Table 2) we achieved a new prediction model: VO max sis (● 2 (ml · min − 1) = (MPO in watt · 10.1) + (sex (male = 0, female = 1) · (− 145)) + (body weight in kg · 6.6) + 84.5 (adjusted R2 = 0.91, root mean squared error (RMSE) = 215 ml · min − 1).
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Men (n = 88)
Training & Testing 1187
200 0 –200 –400
400 200 0 –200 –400 –600 1 000
2 000
3 000 Direct VO2max (ml ·min
5 000
4 000 )
600
Women 600 400 200 0 –200 –400
3 000 Direct VO2max (ml ·min
4 000
4 000
5000 –1
)
400 200 0 –200 –400 –600 1 000
2 000
3 000 Women
–600 1 000
2 000
Mean indirect VO2max day 1 and day 2 (ml ·min
–1
Direct VO2max – indirect VO2max (ml ·min –1)
1 000
2 000
3000
4000
Mean indirect VO2max day 1 and day 2 (ml · min
5 000
5000 –1
)
–1
)
Fig. 3 Bland-Altman plot of difference of direct and indirect measurements of VO2max (ml · min − 1) against the gold standard (direct measurements of VO2max) in men (n = 88) and women (n = 90).
Fig. 4 Bland-Altman plot of difference of the indirect measurements of VO2max (ml · min − 1) on test day 1 and test day 2 against the average measurements in men (n = 88) and women (n = 90).
Discussion −1
Table 2 Linear regression analysis with direct VO2max (ml · min ) as dependent variable. Coefficient MPO (watt) gender body weight (kg) constant
10.1 − 145.8 6.6 84.5
p-value
95 % CI
< 0.01 < 0.01 < 0.01
9.5–10.8 − 231.6–( − 60.1) 3.3–10.0
Gender coefficient is 0 for males and 1 for females
Repeatability We found a high degree of linear correlation between the indirect estimates of VO2max from test day 1 and test day 2. Pearson’s correlation coefficient was 0.98 for men and 0.96 for women. Repeatability of the indirect estimates carried out by different test leaders on test day 1 and test day 2 was assessed by a Bland-Altman plot along with CR for men and women ▶ Fig. 4). The mean difference between the 2 indirect estimates (● was − 12 ml · min − 1 (95 % LOA − 290 and 267 ml · min − 1) for men and 18 ml · min − 1 (95 % LOA − 256 and 292 ml · min − 1) for women. These minor biases for men as well as women were not significantly different from zero. The CR was 279 ml · min − 1 and 274 ml · min − 1 corresponding to a percentage of the mean indirect VO2max of 8.9 % and 11.7 % for men and women, respectively.
▼
With this study, we found good agreement between indirectly estimated VO2max and directly measured VO2max (gold standard) among an adult population ages 23–77 years, though with minor but significant positive and negative bias for men and women, respectively. Owing to this bias we suggested an adjusted prediction model for VO2max valid for healthy adults of all ages. Furthermore, we found high repeatability of the maximal cycle test instructed by different test leaders on different test days.
Validity Comparing the results of the indirect estimate and direct measurements of VO2max we found high correlation coefficients for men as well as women. Nonetheless, the Bland-Altman plots revealed that the indirect estimation on average resulted in statistically significant lower VO2max for men and higher VO2max for women compared to VO2max measured directly, i. e., we found systematic bias. According to the limits of agreement, 95 % of men and women would be predicted to have a VO2max deviating from the direct measurement at a maximum of + 450 and − 565 ml · min − 1, and + 468 and − 328 ml · min − 1, respectively. In terms of percentages these errors correspond to predicted values within + 14 % and − 18 % of the average direct measurement for men and within + 21 % and − 15 % for women.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
400
–600
Direct VO2max – indirect VO2max (ml ·min –1)
Men
600 Direct VO2max – indirect VO2max (ml ·min –1)
Direct VO2max – indirect VO2max (ml ·min –1)
Men 600
The sizes of these errors might not be considered small enough to make use of the indirect measurement method for comparison between individuals. However, for the purpose of obtaining valid mean estimates of VO2max on a population level as was the case in the Danish Health Examination Survey, we consider the accuracy of the indirect test to be good. The prediction model of Andersen [4], which we applied for estimation of VO2max in DANHES, was initially derived from regression analysis based solely on MPO yielding a coefficient of determination (R2) of 0.79. The prediction model did not include weight, sex and age, since it only explained an additional 3 %, 2 % and 1 % of the model, respectively. If we only include MPO in the prediction model of VO2max we obtain an R2 of 0.89 and a regression equation almost identical with the regression derived by Andersen: VO2max (L · min-1) = (MPO in watt · 0.0114) + 0.21 vs. (MPO in watt · 0.0117) + 0.16. This is in line with a prediction model of Hawley and Noakes [14] based on MPO (Wpeak) from a maximal cycle test with direct measurement of VO2max among 100 trained cyclists. The slope of their model is identical with ours, whereas the constant in their model is higher, implying that a given maximal load yields a higher predicted VO2max in trained cyclists compared to our participants. The model of Andersen [4] yielded statistically significant bias in different directions for men (58 ml · min − 1) and women ( − 70 ml · min − 1) when applied to our population. Therefore, despite sex explaining only 2 % of our model, it was included in our new prediction model given the fact that the regression coefficient was of considerable size ( − 145.8). There was no interaction between sex and MPO, and thus no reason to derive separate regression models for men and women. Furthermore, we found body weight to be an explanatory factor with a regression coefficient of 6.6, although it only explained 1 % of the model. In a prediction model of Storer et al. [25] age, body weight and sex were included in addition to MPO, though only explaining a total 3 % of the variation in VO2max. Unlike Andersen [4] and Storer et al. [25] we did not find age to contribute significantly in the prediction of VO2max – most likely due to the fact that age is correlated with MPO. Overall we confirm our hypothesis and find the prediction model of Andersen suitable for estimating VO2max in our population when not divided by sex. However, due to the sex-specific bias, we recommend a new regression model for prediction of VO2max from MPO taking sex and body weight into account achieving an R2 of 0.91 and a RMSE of 213 ml · min − 1. In comparison, using Andersen’s model to our data yields an R2 of 0.89 and a RMSE of 240 ml · min − 1. We consider the new regression model highly suitable for predicting VO2max in the DANHES population as well as in future large population studies of healthy adults of all ages.
our 2 indirect estimates was close to zero and statistically insignificant for men ( − 12 ml · min − 1 (95 % CI: − 42.0–18.2)) as well as women (18 ml · min − 1 (95 % CI: − 11.0–47.6)) with a CR of 279 ml · min − 1 (8.9 %) and 274 ml · min − 1 (11.7 %), respectively. The absence of systematic bias suggested that there was no general improvement from test to retest due to a learning effect and that the use of different test leaders did not result in any over- or underestimation of VO2max. This is in line with the conclusion of Bingisser et al. [6] who compared repeatability of directly measured VO2max in trained vs. untrained adults and found no systematic bias in either of the groups from test to retest. However, they found a lower CR in MPO for the group of trained (4 %) compared to the untrained (11 %) subjects. It is suggested that the routine and experience of the trained subjects enabled them to perform more consistently during the maximal test and vice versa for the untrained subject. In 2 studies testing test-retest repeatability of directly measured VO2max in maximal cycle tests the CRs were higher compared to our study [27, 28] Thus, the repeatability of our indirect measurements of VO2max is better compared to the repeatability found in comparable studies using direct measurement. Perhaps the variation in the direct measurements is partly caused by technical conditions that are not present in the more simple indirect measurements. A study of Kuipers et al. [18] specifically tested the variation in VO2max and MPO among 10 physically active men ages 24–38 years on cycle ergometer. They found that the variation on the individual level in VO2max exceeded that in MPO. It was not possible to conclude whether the discrepancy between variation in VO2max and MPO reflected a biological phenomenon or resulted from errors in the assessment of VO2max. We have not identified other comparable studies testing repeatability of indirect estimates of VO2max from maximal cycle tests by means of other than correlation coefficients. The correlation coefficient is not an ideal measurement of repeatability since a high correlation coefficient does not necessarily mean that the methods agree [7]. Thus, comparison to other studies is difficult due to differences in protocols, populations and applied statistics. Irrespective of the applied measurement method of VO2max, be it direct or indirect, one must be aware that test-retest variation exists. However, the variation does not seem to be higher in our indirect method compared to the variations found in several direct measurement methods. Although no universal criteria exist for test reproducibility, VO2 has been considered reproducible if values are within ± 10 % on separate days [23]. Taking into consideration that part of the variability found in our study could be due to uncontrolled pre-test food and fluid intake as well as physical activity and the fact that the test was not necessarily repeated at the same time of the day, we consider the repeatability of the cycle test to be high.
Repeatability Our analysis of repeatability was an expression of inter-rater reliability (different test leaders conducted the exercise tests). We found a high correlation between VO2max results estimated by different scientific personnel on test day 1 and by one experienced and formally qualified exercise physiologist on test day 2 in men (r = 0.98) as well as women (r = 0.96). Since the 2 different methods in question measured the same condition, it was not surprising that correlations were high. By comparison Andersen [4] found test re-test correlation coefficients of 0.95 and 0.96 in a group of younger men and women, respectively. These tests and retests were performed by physical education teachers and an experienced physiologist. The average difference between Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
Strengths and limitations Directly measured VO2max is considered the gold standard when it comes to assessing aerobic fitness. However, such a comprehensive measurement method is rarely applied in huge population or cohort studies. Irrespective of the method, be it direct or indirect, determination of VO2max in a maximal test is difficult since it depends on the performance and motivation of the individual participants. To make sure the participants reached maximal exhaustion in our study, the test leaders verbally encouraged the participants to continue for as long as possible. The participants were requested to maintain a cadence between 60 and 80 rpm, unless trained, in which case they could exceed
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
1188 Training & Testing
80 rpm. Pedalling rate is believed to have a negative effect on efficiency (the ratio of the work generated to the total metabolic energy cost) and thus on maximal oxygen consumption attainable. However, no firm conclusions have been drawn about the energetically optimal cadence [12], which seem to depend on age, work load, and level of physical cardiorespiratory fitness [24, 26]. It has been found that factors other than work rate, including cadence, account for less than 10 % of the variation in efficiency [12]. The inter-individual variations caused by the selected cadence in our study are therefore likely to be of minor importance. When direct measurement of VO2max is not feasible, several indirect methods for estimation of VO2max are available. Among the indirect methods, maximal testing should be preferred over submaximal testing, as it is considered to be more accurate and superior in the prediction of disease and mortality [1, 16]. However, maximal testing brings a higher risk of adverse events in high-risk individuals and thus should be used with caution. The participants in our study were all healthy based on the inclusion criteria for participation in the maximal cycle test, and our prediction model is therefore well suited for a healthy adult population.
Conflicts of interest: The authors declare no conflict of interest. References 1 American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. Ninth edition. Wolters Kluwer Health: Lippincott Williams & Wilkins, 2013; 72–75 2 American College of Sports Medicine. ACSM’s health-related physical fitness assessment manual. Third edition. Wolters Kluwer Health: Lippincott Williams & Wilkins, 2010; 145 3 American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. Seventh edition. Lippincott Williams & Wilkins, 2006; 99–100 4 Andersen LB. A maximal cycle exercise protocol to predict maximal oxygen uptake. Scand J Med Sci Sports 1995; 5: 143–146 5 Åstrand I. Aerobic work capacity in men and women with special reference to age. Acta Physiol Scand 1960; 49: 83–84 6 Bingisser R, Kaplan V, Scherer T, Russi EW, Bloch KE. Effect of training on repeatability of cardiopulmonary exercise performance in normal men and women. Med Sci Sports Exerc 1997; 29: 1499–1504 7 Bland JM, Altman DG. Statistical Methods for assessing agreement between two methods of clinical measurement. The Lancet 1986; 1: 307–310 8 Borg G. Borg’s Perceived Exertion and Pain Scales. Champaign, IL: Human Kinetics, 1998; 29–39 9 Bruce RA, Blackmon JR, Jones JW, Strait G. Exercise testing in adult normal subjects and cardiac patients. Pediatrics 1963; 32: 742–756
10 Cooper KH. A means of assessing maximum oxygen uptake. JAMA 1968; 203: 201–204 11 Eriksen L, Grønbæk M, Helge JW, Tolstrup JS, Curtis T. The Danish Health Examination Survey 2007–2008 (DANHES 2007–2008). Scand J Pub Health 2011; 39: 203–211 12 Ettema G, Lorås HW. Efficiency in cycling: a review. Eur J Appl Physiol 2009; 106: 1–14 13 Harriss DJ, Atkinson G. Ethical standards in sport and exercise science research: 2014 update. Int J Sports Med 2013; 34: 1025–1028 14 Hawley JA, Noakes TD. Peak power output predicts maximal oxygen uptake and performance time in trained cyclists. Eur J Appl Physiol 1992; 65: 79–83 15 Hermansen L, Saltin B. Oxygen uptake during maximal treadmill and bicycle exercise. J Appl Physiol 1969; 26: 31–37 16 Kodama S, Saito K, Tanaka S, Maki M, Yachi Y, Asumi M, Sugawara A, Totsuka K, Shimano H, Ohashi Y, Yamada N, Sone H. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: a meta-analysis. JAMA 2009; 301: 2024–2035 17 Krouwer JS. Why Bland-Altman plots should use X, not (Y + X)/2 when X is the reference method. Stat Med 2008; 27: 778–780 18 Kuipers H, Verstappen FT, Keizer HA, Geurten P, van Kranenburg G. Variability of aerobic performance in the laboratory and its physiologic correlates. Int J Sports Med 1985; 6: 197–201 19 Léger L, Gadoury C. Validity of the 20 m shuttle run test with 1 min stages to predict VO2max in adults. Can J Sport Sci 1989; 14: 21–26 20 Léger LA, Lambert J. A maximal multistage 20-m shuttle run test to predict VO2max. Eur J Appl Physiol 1982; 49: 1–12 21 Midgley AW, Bentley D, Luttikholt H, McNaughton LR, Millet GP. Challenging a dogma of exercise physiology: does an incremental exercise test for valid VO2max determination really need to last between 8 and 12 minutes? Sports Med 2008; 38: 441–447 22 Midgley AW, Carroll S, Marchant D, McNaughton LR, Siegler J. Evaluation of true maximal oxygen uptake based on a novel set of standardized criteria. Appl Physiol Nutr Metab 2009; 34: 115–123 23 Myers J, Froelicher VF. Optimizing the exercise test for pharmacological investigations. Circulation 1990; 82: 1839–1846 24 Sacchetti M, Lenti M, Di Palumbo AS, De Vito G. Different effect of cadence on cycling efficiency between young and older cyclists. Med Sci Sports Exerc 2010; 42: 2128–2133 25 Storer TW, Davis JA, Caiozzo VJ. Accurate prediction of VO2max in cycle ergometry. Med Sci Sports Exerc 1990; 22: 704–712 26 Vercruyssen F, Brisswalter J. Which factors determine the freely chosen cadence during submaximal cycling. J Sci Med Sport 2010; 13: 225–231 27 Wergel-Kolmert U, Agehall A, Rosenberg N, Wohlfart B. Day-to-day variation in oxygen consumption at submaximal loads during ergometer cycling by adolescents. Clin Physiol 2001; 21: 135–140 28 Wergel-Kolmert U, Wisen A, Wohlfart B. Repeatability of measurements of oxygen consumption, heart rate and Borg’s scale in men during ergometer cycling. Clin Physiol Funct Imaging 2002; 22: 261–265 29 Wisén A, Wohlfart B. Exercise testing using a cycle or treadmill: a review of various protocols. Phys Ther Rev 1999; 4: 7–20
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Training & Testing 1189
Copyright of International Journal of Sports Medicine is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
A Maximal Cycle Test with Good Validity and High Repeatability in Adults of All Ages
Affiliations
Key words
▶ cardiorespiratory fitness ● ▶ reproducibility ● ▶ exercise test ● ▶ physical fitness ● ▶ cycle ergometer ● ▶ measurement ●
L. Eriksen1, J. S. Tolstrup1, S. Larsen2, M. Grønbæk1, J. W. Helge2 1 2
National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark Centre of Healthy Aging, Department of Biomedical Sciences, University of Copenhagen, Copenhagen N, Denmark
Abstract
▼
In 11 680 individuals (18–85 years) maximal oxygen consumption (VO2max) was estimated indirectly in a maximal cycle test using a prediction model developed in a young population (15–28 years). A subsample of 182 individuals (23–77 years) underwent 2 maximal cycle tests with VO2max estimated indirectly in both tests and measured directly in one test. Agreement between the direct measurement and the indirect estimate of VO2max and repeatability of the indirect estimates of VO2max were examined by Bland-Altman plots, limits of agreement (LOA) and coefficient of repeatability (CR). The indi-
Introduction
▼ accepted after revision April 22, 2014 Bibliography DOI http://dx.doi.org/ 10.1055/s-0034-1376973 Published online: September 26, 2014 Int J Sports Med 2014; 35: 1184–1189 © Georg Thieme Verlag KG Stuttgart · New York ISSN 0172-4622 Correspondence Louise Eriksen National Institute of Public Health University of Southern Denmark Øster Farimagsgade 5A 1353 Copenhagen Denmark Tel.: + 45/655/07 728 Fax: + 45/392/08 010 [email protected]
Maximal oxygen consumption (VO2max) is the criterion measure of cardiorespiratory fitness, which is an independent predictor of morbidity and mortality [16]. Direct measurement of VO2max implies measurement of oxygen and carbon dioxide in expired air and is considered the gold standard physiological test [1]. Direct measurement is time-consuming and requires expensive equipment and specific expertise. Thus, several indirect and simpler procedures based on heart rate during exercise [5], covered distance for a given time [10], maximal speed in shuttle run [19, 20] and maximal time and/or power on treadmill or ergometer cycle [3, 4, 9, 14, 29] have been developed to predict VO2max. Indirect tests may be maximal (performed to voluntary exhaustion) or sub-maximal, the latter considered less precise. In the Danish Health Examination Survey 2007–2008 (DANHES) [11] 11 680 participants ages 18–85 years performed a maximal cycle exercise test. VO2max was estimated indirectly based on maximal power output (MPO) obtained by the participants in the cycle test. The prediction model of
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
rect method (mean VO2max = 3 132 ml · min − 1) underestimated VO2max as compared to the direct method (mean VO2max = 3 190 ml · min − 1) in men (bias: 58 ml · min − 1 (95 % LOA − 450 and 565)) and overestimated VO2max in women (mean VO2max = 2 328 vs. 2 258 ml · min − 1, bias: − 70 ml · min − 1 (95 % LOA − 468 and 328)). The mean difference between the 2 indirect estimates was non-significant (men: − 11.9 ml · min − 1, women: 18.3 ml · min − 1) with a CR of 279 ml · min − 1 (8.9 %) in men and 274 ml · min − 1 (11.7 %) in women. The validity of the indirect method was good despite minor sex-specific bias. Owing to this bias we suggest a new prediction model of VO2max. The maximal cycle test was highly repeatable.
VO2max was originally developed by Andersen [4] from directly measured VO2max among 535 men and women ages 15–28 years. Nonetheless, this model might not be suitable for our population with a wider age range, since age has been found to contribute significantly to the prediction of VO2max independent of MPO with lower values of VO2max for older ages [4, 25]. In large population studies like DANHES, it is inevitable that physiological tests are performed under the instruction and guidance of several test leaders due to the huge number of participants. This will most likely induce a higher variation in the results, and it is therefore important to investigate whether such tests have high repeatability, even when being carried out under the guidance of different test leaders. The objective of the present study was to assess the agreement between VO2max estimated indirectly (applying a prediction model originally developed in adolescents and young adults) and measured directly (gold standard) among an adult population spanning a wide age range. A second objective was to test the repeatability of the maximal cycle test under the guidance of different test leaders.
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Authors
Methods
▼
Study population A random subsample of adult citizens 18 years and older (n = 180 103) from 13 Danish municipalities was invited to participate in the Danish Health Examination Survey in 2007–2008 [11]. A total of 18 065 individuals (10 %) accepted the invitation, and based on a screening interview on health status and results of blood pressure measurements participants were eligible for participation in a maximal exercise test (n = 11 680), a sub-maximal test (n = 4 403) or no test (n = 1 982) for indirect estimation of maximal oxygen consumption (VO2max). Out of those 11 680 participants eligible for the maximal test 10 973 (94 %) completed the test with a valid test result. In 3 municipalities a sexand age-stratified sample of participants (n = 182) who underwent the maximal test with a valid test result were invited to come back after a week to perform a retest with indirect estimation as well as direct measurement of VO2max. Informed consent was obtained from all participants, and the protocol was reviewed by the Scientific Ethical Committee B for the Capital Region of Denmark (H-B-2007-050). The study was conducted in accordance with the ethical standards of the International Journal of Sports Medicine [13].
Measurement of VO2max In order to ensure safe participation and achieve optimal risk minimisation in the maximal cycle, test participants were subjected to a screening by means of an interview and measurement of blood pressure. In brief, one or more of the following conditions contraindicated participation in the cycle test: any heart-related disease; chest pain or pressure; moderate hypertension; consumption of antihypertensives; cardiac or pulmonary medication; pregnancy; muscle, joint or skeletal problems. On the first test day (test day 1) VO2max was estimated indirectly using an ergometer cycle (Ergomedic 839E, Monark Exercise AB, Vansbro, Sweden). Heart rate was monitored with chest strap and watch (Polar RS100, Polar Electro Oy, Kempele, Finland). The test started with a 5-min warm-up at a load of 75 watts for women and 100 watts for men. After the warm-up the work load was increased by 35 watts every 2 min until voluntary exhaustion. The size of the initial load and increments were chosen given the notion that younger as well as elderly participants should be able to complete the test within a reasonable time frame [2, 21]. Pedal frequency was freely chosen by participants within the limits 60–80 rpm, while trained participants were allowed to exceed 80 rpm. The participants were verbally encouraged to continue for as long as possible, and the perceived exertion of the participants was evaluated using Borg’s scale [8]. VO2max was estimated from MPO, that is, the highest achieved load adjusted for the proportion of the last stage completed: VO2max (l · min − 1) = 0.16 + (0.0117 · MPO (watt)) [4]. The maximal cycle tests were managed by trained scientific personnel. On the second test day (test day 2) the following week a similar test procedure was followed, yet on a different exercise ergometer cycle of the same model as on the first test day (Ergomedic 839E, Monark Exercise AB, Vansbro, Sweden). During this cycle test direct measurement of pulmonary VO2 as well as indirect estimation of VO2max was performed, enabling assessment of repeatability of the cycle test on 2 different test days as well as
agreement between the indirect and direct measurement conducted in one and the same test. The measurements were performed continuously using an automated metabolic cart (Quark b2, Cosmed Srl., Rome, Italy), calibrated before each test according to the manufacturer’s specifications. Before each test a volume calibration and a calibration of the gas analysers were performed using gases of known composition. The respiratory variables were averaged every 20 s. The greatest 20-s averaged VO2 (l · min − 1) value during the test was taken as the VO2max. Achievement of VO2max was accepted when a respiratory exchange ratio of 1.15 and/or maximal heart rate (220 − age) were present, or a levelling off or decline in VO2 was reached [15, 22]. All tests on test day 2 were managed by one experienced and formally qualified exercise physiologist. No attempt was made to control the participants’ food and beverage intake and physical activity prior to the tests.
Statistics To obtain a visual assessment of the relationship between the different measurements of VO2max, scatterplots and Pearson’s correlation coefficient were used. Bland-Altman plots [7] were used to illustrate the extent of agreement between the direct and indirect measurement of VO2max in the maximal cycle test on test day 2 by plotting the difference between the direct and indirect measurements for each individual against their direct measurement (the gold stand) instead of against the mean of the direct and indirect measurements, since the gold standard is expected to be closer to the “true value” [17]. The mean difference (bias) represents the degree of systematic error between the 2 measurements [7]. The hypothesis of zero bias was tested using a paired t-test. Stepwise multiple regression analysis with the direct measurement as the dependent variable was performed to derive a new equation for indirect estimation of VO2max. The repeatability of the cycle test was evaluated by the coefficient of repeatability (CR) (CR = 1.96 · standard deviation (SD) of the mean difference between the 2 indirect estimates of VO2max) and Bland-Altman plots with the difference between estimates by the 2 indirect methods for each individual against their mean value of the 2 indirect estimates. The significance level was set at P < 0.05. Statistics were calculated using STATA/IC 12.
Results
▼
Participants A total of 182 participants (91 women and 91 men) were enrolled in the present study. The results from 4 participants were excluded from the analyses due to unreliable direct test measurements resulting from technical problems with the equipment. Thus, data from 88 men and 90 women who all achieved one of the 3 cut-off criteria, were analysed. The study population is characterised by means and SD or ranges in ▶ Table 1. As expected, for the indirect estimates as well as the ● direct measurements of VO2max, mean VO2max was higher for men compared to women, e. g. 3 120 ml · min − 1 and 2 347 ml · min − 1, respectively, from the indirect estimate on test day 1. Indirectly and directly determined VO2max in men and women on the 2 different test days across age groups are illustrated ▶ Fig. 1. Overall, a lower VO max was observed with increasin ● 2 ing age in men and women.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Training & Testing 1185
1186 Training & Testing
Test day 1 age (years) weight (kg) height (cm) BMI (kg · m − 2) indirect VO2max (ml · min − 1) indirect VO2max (ml · min − 1 · kg − 1) MPO (Watt) maximal heart rate (bpm) direct VO2max (ml · min − 1) direct VO2max (ml · min − 1 · kg − 1)
Test day 2
48.8 (23–77) 80.3 (9.7) 180 (6) 24.7 (2.8) 3 120 (625) 39.2 (8.3) 253 (56) 174 (16)
3 132 (655) 39.4 (8.9) 254 (56) 175 (14) 3 190 (637) 40.1 (8.3)
Test day 1
Test day 2
44.4 (22–72) 66.9 (9.7) 168 (6) 23.8 (3.0) 2 347 (451) 35.5 (7.9) 185 (41) 174 (14)
2 328 (481) 35.2 (7.4) 185 (41) 176 (13) 2 258 (459) 34.1 (6.5)
Men
Men
3 800
Table 1 Characteristics of the population. Means and standard deviations (SD) or range.
Women (n = 90)
Women
5 000
3 600 Indirect VO2max (ml min –1)
3 400
3 000 2 800 2 600 2 400
4 000
3 000
Pearson’s r =0.91
2 200
22–29
30–39
40–49 Age (years)
n=4
n=24
n=21
Indirect VO2max day 1
50–59
60–77
n=16
n=23
Direct VO2max day 2
Indirect VO2max day 2 Women
3 800 3 600 3 400 3 200 3 000 2 800 2 600 2 400 2 200 2 000 1 800 1 600 1 400
00 50
00 40
00 30
00 20
00
00
10
50
00 40
00
Direct VO2max (ml min –1)
1 600 1 400
30
00 10
1 800
00
1 000
2 000
VO2max (ml min –1)
Pearson’s r = 0.92
2 000
20
VO2max (ml min –1)
3 200
22–29
30–39
n=13
n=24
40–49 Age (years)
Indirect VO2max day 1
n=22
50–59
60–77
n=19
n=12
Direct VO2max day 2
Indirect VO2max day 2
Fig. 1 VO2max (ml · min − 1) determined indirectly and directly in men and women on separate test days across age groups. Data are mean ± SEM.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
Fig. 2 Scatter plot indirect vs. direct measurement of VO2max (ml · min − 1) for men (n = 88) and women (n = 90).
Validity A scatter plot of the indirect estimate vs. the direct measurement of VO2max from test day 2 showed that data for both men and women were near the line of equality, and Pearson’s correla▶ Fig. 2). tion coefficient was 0.92 for men and 0.91 for women (● ▶ Fig. 3. Bias was positive in Bland-Altman plots are shown in ● men (58 ml · min − 1) and negative in women ( − 70 ml · min − 1). The agreement between the 2 methods for an individual was considered by the degree of random error indicated by the 95 % limits of agreement (LOA) in the Bland-Altman plot. Thus, since differences between the 2 measurements were normally distributed, 95 % of the differences fell within the range of − 450 and 565 ml · min − 1 for men and − 468 and 328 ml · min − 1 for women. The results of a paired t-test on the indirect estimate and direct measurement showed that the mean difference was significantly different from zero for men as well as women (58 ml · min − 1 (95 % confidence interval (CI): 3–113) and − 70 ml · min − 1 (95 % CI: − 112–( − 27)). Due to this systematic bias we made a new prediction model derived from a multiple linear regression with direct VO2max as the dependent variable. The model initially included MPO (watt), sex, age, body weight (kg), and the interaction sex · MPO. Age and the interaction sex · MPO were found to be non-significant. From the results of the final regression analy▶ Table 2) we achieved a new prediction model: VO max sis (● 2 (ml · min − 1) = (MPO in watt · 10.1) + (sex (male = 0, female = 1) · (− 145)) + (body weight in kg · 6.6) + 84.5 (adjusted R2 = 0.91, root mean squared error (RMSE) = 215 ml · min − 1).
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Men (n = 88)
Training & Testing 1187
200 0 –200 –400
400 200 0 –200 –400 –600 1 000
2 000
3 000 Direct VO2max (ml ·min
5 000
4 000 )
600
Women 600 400 200 0 –200 –400
3 000 Direct VO2max (ml ·min
4 000
4 000
5000 –1
)
400 200 0 –200 –400 –600 1 000
2 000
3 000 Women
–600 1 000
2 000
Mean indirect VO2max day 1 and day 2 (ml ·min
–1
Direct VO2max – indirect VO2max (ml ·min –1)
1 000
2 000
3000
4000
Mean indirect VO2max day 1 and day 2 (ml · min
5 000
5000 –1
)
–1
)
Fig. 3 Bland-Altman plot of difference of direct and indirect measurements of VO2max (ml · min − 1) against the gold standard (direct measurements of VO2max) in men (n = 88) and women (n = 90).
Fig. 4 Bland-Altman plot of difference of the indirect measurements of VO2max (ml · min − 1) on test day 1 and test day 2 against the average measurements in men (n = 88) and women (n = 90).
Discussion −1
Table 2 Linear regression analysis with direct VO2max (ml · min ) as dependent variable. Coefficient MPO (watt) gender body weight (kg) constant
10.1 − 145.8 6.6 84.5
p-value
95 % CI
< 0.01 < 0.01 < 0.01
9.5–10.8 − 231.6–( − 60.1) 3.3–10.0
Gender coefficient is 0 for males and 1 for females
Repeatability We found a high degree of linear correlation between the indirect estimates of VO2max from test day 1 and test day 2. Pearson’s correlation coefficient was 0.98 for men and 0.96 for women. Repeatability of the indirect estimates carried out by different test leaders on test day 1 and test day 2 was assessed by a Bland-Altman plot along with CR for men and women ▶ Fig. 4). The mean difference between the 2 indirect estimates (● was − 12 ml · min − 1 (95 % LOA − 290 and 267 ml · min − 1) for men and 18 ml · min − 1 (95 % LOA − 256 and 292 ml · min − 1) for women. These minor biases for men as well as women were not significantly different from zero. The CR was 279 ml · min − 1 and 274 ml · min − 1 corresponding to a percentage of the mean indirect VO2max of 8.9 % and 11.7 % for men and women, respectively.
▼
With this study, we found good agreement between indirectly estimated VO2max and directly measured VO2max (gold standard) among an adult population ages 23–77 years, though with minor but significant positive and negative bias for men and women, respectively. Owing to this bias we suggested an adjusted prediction model for VO2max valid for healthy adults of all ages. Furthermore, we found high repeatability of the maximal cycle test instructed by different test leaders on different test days.
Validity Comparing the results of the indirect estimate and direct measurements of VO2max we found high correlation coefficients for men as well as women. Nonetheless, the Bland-Altman plots revealed that the indirect estimation on average resulted in statistically significant lower VO2max for men and higher VO2max for women compared to VO2max measured directly, i. e., we found systematic bias. According to the limits of agreement, 95 % of men and women would be predicted to have a VO2max deviating from the direct measurement at a maximum of + 450 and − 565 ml · min − 1, and + 468 and − 328 ml · min − 1, respectively. In terms of percentages these errors correspond to predicted values within + 14 % and − 18 % of the average direct measurement for men and within + 21 % and − 15 % for women.
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
400
–600
Direct VO2max – indirect VO2max (ml ·min –1)
Men
600 Direct VO2max – indirect VO2max (ml ·min –1)
Direct VO2max – indirect VO2max (ml ·min –1)
Men 600
The sizes of these errors might not be considered small enough to make use of the indirect measurement method for comparison between individuals. However, for the purpose of obtaining valid mean estimates of VO2max on a population level as was the case in the Danish Health Examination Survey, we consider the accuracy of the indirect test to be good. The prediction model of Andersen [4], which we applied for estimation of VO2max in DANHES, was initially derived from regression analysis based solely on MPO yielding a coefficient of determination (R2) of 0.79. The prediction model did not include weight, sex and age, since it only explained an additional 3 %, 2 % and 1 % of the model, respectively. If we only include MPO in the prediction model of VO2max we obtain an R2 of 0.89 and a regression equation almost identical with the regression derived by Andersen: VO2max (L · min-1) = (MPO in watt · 0.0114) + 0.21 vs. (MPO in watt · 0.0117) + 0.16. This is in line with a prediction model of Hawley and Noakes [14] based on MPO (Wpeak) from a maximal cycle test with direct measurement of VO2max among 100 trained cyclists. The slope of their model is identical with ours, whereas the constant in their model is higher, implying that a given maximal load yields a higher predicted VO2max in trained cyclists compared to our participants. The model of Andersen [4] yielded statistically significant bias in different directions for men (58 ml · min − 1) and women ( − 70 ml · min − 1) when applied to our population. Therefore, despite sex explaining only 2 % of our model, it was included in our new prediction model given the fact that the regression coefficient was of considerable size ( − 145.8). There was no interaction between sex and MPO, and thus no reason to derive separate regression models for men and women. Furthermore, we found body weight to be an explanatory factor with a regression coefficient of 6.6, although it only explained 1 % of the model. In a prediction model of Storer et al. [25] age, body weight and sex were included in addition to MPO, though only explaining a total 3 % of the variation in VO2max. Unlike Andersen [4] and Storer et al. [25] we did not find age to contribute significantly in the prediction of VO2max – most likely due to the fact that age is correlated with MPO. Overall we confirm our hypothesis and find the prediction model of Andersen suitable for estimating VO2max in our population when not divided by sex. However, due to the sex-specific bias, we recommend a new regression model for prediction of VO2max from MPO taking sex and body weight into account achieving an R2 of 0.91 and a RMSE of 213 ml · min − 1. In comparison, using Andersen’s model to our data yields an R2 of 0.89 and a RMSE of 240 ml · min − 1. We consider the new regression model highly suitable for predicting VO2max in the DANHES population as well as in future large population studies of healthy adults of all ages.
our 2 indirect estimates was close to zero and statistically insignificant for men ( − 12 ml · min − 1 (95 % CI: − 42.0–18.2)) as well as women (18 ml · min − 1 (95 % CI: − 11.0–47.6)) with a CR of 279 ml · min − 1 (8.9 %) and 274 ml · min − 1 (11.7 %), respectively. The absence of systematic bias suggested that there was no general improvement from test to retest due to a learning effect and that the use of different test leaders did not result in any over- or underestimation of VO2max. This is in line with the conclusion of Bingisser et al. [6] who compared repeatability of directly measured VO2max in trained vs. untrained adults and found no systematic bias in either of the groups from test to retest. However, they found a lower CR in MPO for the group of trained (4 %) compared to the untrained (11 %) subjects. It is suggested that the routine and experience of the trained subjects enabled them to perform more consistently during the maximal test and vice versa for the untrained subject. In 2 studies testing test-retest repeatability of directly measured VO2max in maximal cycle tests the CRs were higher compared to our study [27, 28] Thus, the repeatability of our indirect measurements of VO2max is better compared to the repeatability found in comparable studies using direct measurement. Perhaps the variation in the direct measurements is partly caused by technical conditions that are not present in the more simple indirect measurements. A study of Kuipers et al. [18] specifically tested the variation in VO2max and MPO among 10 physically active men ages 24–38 years on cycle ergometer. They found that the variation on the individual level in VO2max exceeded that in MPO. It was not possible to conclude whether the discrepancy between variation in VO2max and MPO reflected a biological phenomenon or resulted from errors in the assessment of VO2max. We have not identified other comparable studies testing repeatability of indirect estimates of VO2max from maximal cycle tests by means of other than correlation coefficients. The correlation coefficient is not an ideal measurement of repeatability since a high correlation coefficient does not necessarily mean that the methods agree [7]. Thus, comparison to other studies is difficult due to differences in protocols, populations and applied statistics. Irrespective of the applied measurement method of VO2max, be it direct or indirect, one must be aware that test-retest variation exists. However, the variation does not seem to be higher in our indirect method compared to the variations found in several direct measurement methods. Although no universal criteria exist for test reproducibility, VO2 has been considered reproducible if values are within ± 10 % on separate days [23]. Taking into consideration that part of the variability found in our study could be due to uncontrolled pre-test food and fluid intake as well as physical activity and the fact that the test was not necessarily repeated at the same time of the day, we consider the repeatability of the cycle test to be high.
Repeatability Our analysis of repeatability was an expression of inter-rater reliability (different test leaders conducted the exercise tests). We found a high correlation between VO2max results estimated by different scientific personnel on test day 1 and by one experienced and formally qualified exercise physiologist on test day 2 in men (r = 0.98) as well as women (r = 0.96). Since the 2 different methods in question measured the same condition, it was not surprising that correlations were high. By comparison Andersen [4] found test re-test correlation coefficients of 0.95 and 0.96 in a group of younger men and women, respectively. These tests and retests were performed by physical education teachers and an experienced physiologist. The average difference between Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
Strengths and limitations Directly measured VO2max is considered the gold standard when it comes to assessing aerobic fitness. However, such a comprehensive measurement method is rarely applied in huge population or cohort studies. Irrespective of the method, be it direct or indirect, determination of VO2max in a maximal test is difficult since it depends on the performance and motivation of the individual participants. To make sure the participants reached maximal exhaustion in our study, the test leaders verbally encouraged the participants to continue for as long as possible. The participants were requested to maintain a cadence between 60 and 80 rpm, unless trained, in which case they could exceed
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
1188 Training & Testing
80 rpm. Pedalling rate is believed to have a negative effect on efficiency (the ratio of the work generated to the total metabolic energy cost) and thus on maximal oxygen consumption attainable. However, no firm conclusions have been drawn about the energetically optimal cadence [12], which seem to depend on age, work load, and level of physical cardiorespiratory fitness [24, 26]. It has been found that factors other than work rate, including cadence, account for less than 10 % of the variation in efficiency [12]. The inter-individual variations caused by the selected cadence in our study are therefore likely to be of minor importance. When direct measurement of VO2max is not feasible, several indirect methods for estimation of VO2max are available. Among the indirect methods, maximal testing should be preferred over submaximal testing, as it is considered to be more accurate and superior in the prediction of disease and mortality [1, 16]. However, maximal testing brings a higher risk of adverse events in high-risk individuals and thus should be used with caution. The participants in our study were all healthy based on the inclusion criteria for participation in the maximal cycle test, and our prediction model is therefore well suited for a healthy adult population.
Conflicts of interest: The authors declare no conflict of interest. References 1 American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. Ninth edition. Wolters Kluwer Health: Lippincott Williams & Wilkins, 2013; 72–75 2 American College of Sports Medicine. ACSM’s health-related physical fitness assessment manual. Third edition. Wolters Kluwer Health: Lippincott Williams & Wilkins, 2010; 145 3 American College of Sports Medicine. ACSM’s guidelines for exercise testing and prescription. Seventh edition. Lippincott Williams & Wilkins, 2006; 99–100 4 Andersen LB. A maximal cycle exercise protocol to predict maximal oxygen uptake. Scand J Med Sci Sports 1995; 5: 143–146 5 Åstrand I. Aerobic work capacity in men and women with special reference to age. Acta Physiol Scand 1960; 49: 83–84 6 Bingisser R, Kaplan V, Scherer T, Russi EW, Bloch KE. Effect of training on repeatability of cardiopulmonary exercise performance in normal men and women. Med Sci Sports Exerc 1997; 29: 1499–1504 7 Bland JM, Altman DG. Statistical Methods for assessing agreement between two methods of clinical measurement. The Lancet 1986; 1: 307–310 8 Borg G. Borg’s Perceived Exertion and Pain Scales. Champaign, IL: Human Kinetics, 1998; 29–39 9 Bruce RA, Blackmon JR, Jones JW, Strait G. Exercise testing in adult normal subjects and cardiac patients. Pediatrics 1963; 32: 742–756
10 Cooper KH. A means of assessing maximum oxygen uptake. JAMA 1968; 203: 201–204 11 Eriksen L, Grønbæk M, Helge JW, Tolstrup JS, Curtis T. The Danish Health Examination Survey 2007–2008 (DANHES 2007–2008). Scand J Pub Health 2011; 39: 203–211 12 Ettema G, Lorås HW. Efficiency in cycling: a review. Eur J Appl Physiol 2009; 106: 1–14 13 Harriss DJ, Atkinson G. Ethical standards in sport and exercise science research: 2014 update. Int J Sports Med 2013; 34: 1025–1028 14 Hawley JA, Noakes TD. Peak power output predicts maximal oxygen uptake and performance time in trained cyclists. Eur J Appl Physiol 1992; 65: 79–83 15 Hermansen L, Saltin B. Oxygen uptake during maximal treadmill and bicycle exercise. J Appl Physiol 1969; 26: 31–37 16 Kodama S, Saito K, Tanaka S, Maki M, Yachi Y, Asumi M, Sugawara A, Totsuka K, Shimano H, Ohashi Y, Yamada N, Sone H. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: a meta-analysis. JAMA 2009; 301: 2024–2035 17 Krouwer JS. Why Bland-Altman plots should use X, not (Y + X)/2 when X is the reference method. Stat Med 2008; 27: 778–780 18 Kuipers H, Verstappen FT, Keizer HA, Geurten P, van Kranenburg G. Variability of aerobic performance in the laboratory and its physiologic correlates. Int J Sports Med 1985; 6: 197–201 19 Léger L, Gadoury C. Validity of the 20 m shuttle run test with 1 min stages to predict VO2max in adults. Can J Sport Sci 1989; 14: 21–26 20 Léger LA, Lambert J. A maximal multistage 20-m shuttle run test to predict VO2max. Eur J Appl Physiol 1982; 49: 1–12 21 Midgley AW, Bentley D, Luttikholt H, McNaughton LR, Millet GP. Challenging a dogma of exercise physiology: does an incremental exercise test for valid VO2max determination really need to last between 8 and 12 minutes? Sports Med 2008; 38: 441–447 22 Midgley AW, Carroll S, Marchant D, McNaughton LR, Siegler J. Evaluation of true maximal oxygen uptake based on a novel set of standardized criteria. Appl Physiol Nutr Metab 2009; 34: 115–123 23 Myers J, Froelicher VF. Optimizing the exercise test for pharmacological investigations. Circulation 1990; 82: 1839–1846 24 Sacchetti M, Lenti M, Di Palumbo AS, De Vito G. Different effect of cadence on cycling efficiency between young and older cyclists. Med Sci Sports Exerc 2010; 42: 2128–2133 25 Storer TW, Davis JA, Caiozzo VJ. Accurate prediction of VO2max in cycle ergometry. Med Sci Sports Exerc 1990; 22: 704–712 26 Vercruyssen F, Brisswalter J. Which factors determine the freely chosen cadence during submaximal cycling. J Sci Med Sport 2010; 13: 225–231 27 Wergel-Kolmert U, Agehall A, Rosenberg N, Wohlfart B. Day-to-day variation in oxygen consumption at submaximal loads during ergometer cycling by adolescents. Clin Physiol 2001; 21: 135–140 28 Wergel-Kolmert U, Wisen A, Wohlfart B. Repeatability of measurements of oxygen consumption, heart rate and Borg’s scale in men during ergometer cycling. Clin Physiol Funct Imaging 2002; 22: 261–265 29 Wisén A, Wohlfart B. Exercise testing using a cycle or treadmill: a review of various protocols. Phys Ther Rev 1999; 4: 7–20
Eriksen L et al. A Maximal Cycle Test … Int J Sports Med 2014; 35: 1184–1189
This document was downloaded for personal use only. Unauthorized distribution is strictly prohibited.
Training & Testing 1189
Copyright of International Journal of Sports Medicine is the property of Georg Thieme Verlag Stuttgart and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
