12-MINUTE STATIONARY CYCLE PERFORMANCE AS A PREDICTOR OF 1.5-MILE RUN TIME James Hodgdon, Ph.D. Linda Hervig, M.A. Lisa Griswold, M.S. Jeff Terry, B.S. Charlie Le, B.S. LCDR Kenneth Sausen, Ph.D. Paul Miller, M.D.

Naval Health Research Center Report No. XX-XX supported by the Bureau of Naval Personnel, under Work Unit No. 60601. The views expressed in this report are those of the authors and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government. Approved for public release; distribution is unlimited. This research has been conducted in compliance with all applicable federal regulations governing the protection of human subjects in research.

Naval Health Research Center P.O. Box 85122 San Diego, California 92186-5122 1

SUMMARY Issue Recently there has been an interest in developing alternative methods of testing cardiovascular-respiratory capacity for inclusion in the Navy Physical Readiness Test (PRT). Reasons for this interest include a desire to allow sailors and officers to be tested on the devices that they use for training and to provide “low-impact” tests that can be used by sailors or officers who cannot run as a result of medical problems. Objective The objective of this study was to explore the use of a stationary exercise cycle test as an alternative to the 1.5-mile run in the PRT. Approach The Life Fitness model 95Ci stationary cycle was used for this exploration. Participants in this study were 59 U.S. Navy sailors and officers (36 male, 23 female) who consented to participate. Fifty-one participants completed all aspects of the study. For this sample, men and women differed in stature, weight, and body composition, but not in age. Participants were tested twice. Visits were usually separated by 1 day. During each visit, they had their body weight measured and were administered the stationary cycle test. The test was a 12-minute bout on the stationary cycle during which, the participant pedaled at self-selected pedaling rates and work loads. At the end of the 12-minute test period, the number of calories expended was recovered. Participants also ran 1.5 miles for time as part of their command PRT or mock PRT. These runs took place within 3 weeks of stationary cycle testing. Results The multiple regression analysis yielded the following equation to predict 1.5-mile run time from cycle performance: 1.5 - mile run time = 8.02 + 2.99 ×

body weight (lb) + 1.296 × gender(0,1) , where gender is coded kcals in 12 minutes

0 for males and 1 for females. The multiple correlation coefficient for this equation is 0.732, and the standard error of estimate is 1.05 minutes (1:03).

i

Analysis of the regression residuals revealed no obvious biases related to physical characteristics. The residuals were not correlated significantly with age (r = 0.15, P = 0.31), stature (r = 0.07, P = 0.63), body mass (r = -0.03, P = 0.82), fat-free mass (r = 0.02, P = 0.91), fat mass (r = -0.11, P = 0.47), percent body fat (r = -0.12, P = 0.43) or gender (r = 0.00, P = 1.00). However, the residuals were significantly correlated with performance, F1,47 = 10.16, P = 0.003. For run times greater than ~ 12.5 minutes, the predicted run times were generally less than measured performance. For run times less than 12.5 minutes, the predicted run times were generally greater than the measured performance. The predictive accuracy of this test is similar to that of the Navy’s currently employed 12minute elliptical trainer test, and better than that of the Navy’s currently used swim test. Conclusion Use of the stationary cycle for PRT testing provides a viable mode for testing individuals who cannot run due to physical limitations on weight-bearing activities or who do not like to run and prefer to train using the stationary cycle.

ii

INTRODUCTION The U.S. Navy currently employs time to complete a 1.5-mile run or time to complete a 500yd or 450-m swim as tests of cardiovascular-respiratory capacity within its semiannual Physical Readiness Test (PRT) (Deputy Chief of Naval Operations, 2005). Recently there has been an interest in developing alternative methods of testing cardiovascular-respiratory capacity for inclusion in the PRT. Reasons for this interest include a desire to allow sailors and officers to be tested on the devices that they use for training and to provide low-impact tests that can be used by sailors or officers who cannot run as a result of medical problems. Development of an alternative test based on 12-minute performance on an elliptical trainer has already been completed (Parker, Griswold, & Vickers, 2006). In this test, the participant works for 12-minutes on one of a number of specified elliptical trainers. The number of calories accumulated for the 12-minute period is recorded. The number of calories and the participant’s weight are used in an equation to predict an equivalent 1.5-mile run time. This test has been accepted for use in the PRT and will be included in the next PRT cycle. Most Navy fitness centers include stationary exercise cycles among their exercise equipment. Like the elliptical trainer, these exercise machines have potential to serve as cardiovascularrespiratory capacity test devices. Additionally, both the U.S. Army and U.S. Air Force offer stationary cycle alternatives for determination of aerobic capacity (Secretary of the Air Force, 2006; Secretary of the Army, 1998). The objective of this study was to explore the use of a stationary exercise cycle test as an alternative to the 1.5-mile run in the PRT. The approach taken was to use a similar format to that used with the elliptical trainer, namely measurement of the number of calories indicated on the stationary cycle after a 12-minute work bout. Because running is a weight-bearing activity, and stationary cycle exercise is not, the number of calories expended needs to be adjusted by the participant’s body weight to determine an equivalent run time. METHODS The Life Fitness model 95Ci stationary cycle (Schiller Park, IL) was used for this exploration. This cycle was selected because it appears to be the one most commonly purchased recently by Navy fitness centers.

1

Subjects Participants in this study were 59 U.S. Navy sailors and officers (36 male, 23 female). Participants were briefed on the purposes, procedures, and risks associated with participation in this study. They indicated their desire to participate by giving their informed consent. Fifty-one participants completed all aspects of the study. Their physical characteristics are provided in Table 1. For this sample, men and women differed in stature, weight and body composition, but not in age. Table 1. Participant Characteristicsa Age (y) Height (in) Weight (lb) % body fat a

Male (N = 33) 30.5 (9.0) 69.3 (2.4) 193.3 (28.3) 22.2 (4.8)

Female (N = 18) 27.1 (5.9) 65.4 (3.2)b 152.4 (20.2)b 30.2 (7.5)b

Total (N = 51) 29.3 (8.1) 67.9 (3.3) 178.9 (32.3) 25.0 (7.0)

Values shown are means (1 SD). Male and female subjects differ (P < 0.01).

b

Procedures Participants reported to the testing facility twice. Visits were usually separated by 1 day. During each visit, they had their body weight measured and were administered the stationary cycle test. During the first visit, they also had stature and body girths measured (neck and abdomen for men; neck, waist, and hips for women), and had their body composition estimated using the U.S. Navy formulas (Hodgdon & Beckett, 1984a, 1984b). During the first visit, participants were also asked about their experience exercising on the stationary cycle. Participants also ran 1.5 miles for time as part of their command PRT or mock PRT. These runs took place within 3 weeks of stationary cycle testing. The PRTs were conducted in accordance with Navy Instruction 6110.1h (Deputy Chief of Naval Operations, 2005). The distribution of classifications for their run times is shown in Figure 1. Stationary Cycle Test The stationary cycle test was conducted using a Life Fitness model 95Ci. The test was conducted with the cycle set to “manual” mode and “bike” mode (in bike mode, the resistance is fixed for a given work level, and changing the pedaling frequency changes the work rate correspondingly). Participants were given a 5-minute warm-up at a light load. During the warmup period, participants were encouraged to experiment with different levels and pedaling speeds

2

to help identify the work level that

30

Number of Participants

they would use to begin the test. Following the warm-up period, the cycle was stopped and a new work

20

session was set up. The time was set for 12 minutes and the workout level 10

Gender F

0

M Outstanding

Excellent

Good

Satisfactory

was set to the value selected by the participant. The participant then pedaled for 12 minutes. Participants

Fail

were encouraged to give the same PRT Run Time Classification

level of effort that they would during Figure 1. Classification of PRT run times.

a PRT run. During the 12-minute

test, participants were allowed to vary the work level and pedaling rate. They were required to hold on to the handlebars during the test and were not allowed to stand up to pedal. At the end of 12 minutes, the CLEAR button was pressed to freeze the display, and the total calories expended was read off the display and recorded. The participant was then allowed to cool down for 5 minutes. Analysis The modeling approach undertaken was similar to that used with the elliptical trainer (Parker, Griswold, & Vickers, 2006). The weight-adjusted cost of running 1.5 miles is similar for all individuals. The energy expenditure, in kilocalories, is equal to 1.093 x body weight in pounds (McArdle, Katch, & Katch, 1991, p. 182). The number of kilocalories expended during the 12minute cycle bout can be used to estimate the equivalent 1.5-mile run time. The estimation would be proportional to the fraction: number of calories required to run 1.5 miles divided by the number of calories expended during the bike test, multiplied by the test time. That is, Est. time ∝ a 1

1.093 × body weight (lb) × 12 minutes + a 0 , where a1 and a0 are regression constants kcals in 12 minutes

needed to adjust for differences between cycling and running as exercise modes. (Running is weight bearing, cycling is not. Cycling involves a lesser proportion of the muscle mass than running).

3

The outcome measure for the cycle test, then, was the participant’s body weight divided by 12-minute calorie expenditure values. The first test was treated as a familiarization bout, and the value derived from the second test was used in multiple regression analysis to predict the run time recorded during the participant’s PRT. A gender-coded (0 = male, 1 = female) dummy variable was also included in the regression analysis, in the event that there was bias in the run time prediction associated with gender as had been found with the elliptical trainer test (Parker SB, Griswold L & Vickers RR, 2006). Statistical analyses were carried out using SPSS for Windows, version 15 (SPSS, Inc., Chicago, IL). RESULTS Table 2 shows the results for the performance measures in this study. Men and women differed in the number of kilocalories expended and in their 1.5-mile run times, but not in the ratio of body weight to kilocalories expended. Table 2. Performance Resultsa 1.5-mile run time (min) Kilocalories in 12 minutes Body weight (lb)/kcals a

Male (N = 33) 12.4 (1.4) 140.5 (26.6) 1.412 (0.277)

Female (N = 18) 14.0 (2.0)b 105.3 (16.1)b 1.479 (0.301)

Total (N = 51) 12.9 (1.8) 128.1 (28.8) 1.435 (0.285)

Values shown are means (1 SD). Male and female subjects differ (P < 0.01).

b

Figure 2 shows the relationship between the ratio of body weight to kilocalories expended in 12 minutes and 1.5-mile run time. The correlation coefficient for this relationship is 0.599, and the standard error of estimate is 1.46 minutes (1:27.6). The figure displays the regression line predicting run time from the body weight–calorie ratio as well as the 95% confidence limits for the regression. One can see that two data points lie outside the 95% confidence intervals. These two cases were found to lie consistently outside the 95% confidence limits for every model that involved run time predicted from calories. Because of this, these two cases were dropped from the analyses. The multiple regression analysis yielded the following equation to predict 1.5-mile run time from cycle performance:

4

Gender

20.0

18.0

Gender F M

F M

3

18.0

16.0

1.5-mile Run Time (min)

1.5-mile Run Time (min)

25

16.0

14.0

12.0

14.0

12.0

10.0

10.0

R Sq Linear = 0.359 8.0

8.0 0.80

1.00

1.20

1.40

1.60

1.80

2.00

2.20

10.0

Body Weight (lb) / Kcals Expended, Trial 2

12.0

13.0

14.0

15.0

16.0

Run Time Predicted From Equation 1

Figure 2. Relationship between measured 1.5-mile run time and the ratio of body weight to calories expended in 12 minutes. Lines indicate the regression and 95% confidence intervals. Points flagged with case numbers represent outliers subsequently dropped from the analysis.

1.5 - mile run time = 8.02 + 2.99 ×

11.0

Figure 3. Comparison of measured 1.5-mile run time with that predicted from Equation 1. Dashed line is the line of identity. Solid lines indicate the 95% confidence intervals for the prediction

body weight (lb) + 1.296 × gender(0,1) kcals in 12 minutes

(Equation 1),

where gender is coded 0 for males and 1 for females. The multiple correlation coefficient for this equation is 0.732, and the standard error of estimate is 1.05 minutes (1:03). This result is displayed graphically in Figure 3. DISCUSSION Prediction Bias Analysis of the regression residuals revealed no obvious biases related to physical characteristics. The residuals were not correlated significantly with age (r = 0.15, P = 0.31), stature (r = 0.07, P = 0.63), body mass (r = -0.03, P = 0.82), fat-free mass (r = 0.02, P = 0.91), fat mass (r = -0.11, P = 0.47), percent body fat (r = -0.12, P = 0.43) or gender (r = 0.00, P = 1.00). However, the residuals were significantly correlated with performance (F1,47 = 10.16, P = 0.003). Figure 4 is a Bland-Altman plot showing the regression residuals (difference between predicted and measured run times plotted as a function of the mean of the predicted and measured run times. Indicated on the plot are horizontal reference lines showing the mean difference between measured and predicted run times (solid line at y-value 0) and the 95%

5

Figure 4. Bland-Altman plot showing residuals as a function of the mean of the predicted and measured run times. Horizontal lines indicate zero difference (solid) and 95% confidence intervals (dashed) for the residuals. Slanted line indicates the regression of the residuals on the mean of the run times, also shown with 95% confidence intervals.

Figure 5. Change in aerobic classification as a function of initial run-time classification.

confidence intervals for the differences (dashed lines). The data show a tendency for the difference between predicted and measured run time to decrease as the run times become greater. For run times greater than 12.5 minutes, the predicted run times were generally less than measured performance. For run times less than 12.5 minutes, the predicted run times were generally greater than the measured performance. The correlation between the prediction equation residuals and run-time performance implies the aerobic performance classification will vary depending upon the test chosen. Figure 5 depicts the change in PRT aerobic classification using the results of the stationary cycle test as a function of the classification based on the 1.5-mile run. In this figure, the Navy PRT classifications have been assigned numeric values as indicated in the legend. The scale values increase with better run performance. The results depicted in this figure reflect findings shown in the Bland-Altman plot. Those who performed at “good medium” or better on the run are more likely to do less well using the cycle, and those perform at less than “good medium” on the run are more likely to improve their score by using the cycle.

6

This should not be an unexpected finding. Cycling and running are very different activities, as noted above. Individuals who are proficient at running or who like to run might be expected to perform better at running than at cycling. Conversely, those who are not proficient at running or do not like to run might well be expected to find it easier to exercise on the cycle. Alternative Models The model presented is one of several that were developed. It has the advantage of being based on theoretical considerations, and is similar in form to the predictive equation used in the elliptical trainer test (Parker, Griswold, & Vickers, 2006). The best fit model developed was: 1.5 − mile run time = 16.773 − 6.160 ×

kcals in 12 minutes + 1.311 × gender(0,1) body weight (lb)

(Equation 2)

where gender is coded 0 for males and 1 for females, as above. The multiple correlation coefficient for this model was 0.756 and the standard error of estimate is 1.01 (1:00.6). While this model offers a better fit (but not significantly so) to the data than Equation 1, this equation has two disadvantages when compared with Equation 1. First, it is not consistent with theory and thus may simply represent capitalization on chance for its better fit. Second, when the equation is solved for zero calories, the result is a time (16:46 for men, 18:05 for women) that represents a passing score for the PRT. Thus, an individual could perform no work at all and pass the PRT. If Equation 1 is solved for zero calories expended, the result is an infinite run time. However, it should be noted that if Equation 1 is solved for an infinite number of calories expended (the term, body weight (lb) kcals in 12 minutes ⇒ 0 ), the result is a time of 8:01 for men and 9:20 for women. These values represent the best theoretical performance that can be estimated using Equation 1. From our study sample, excellent performances on the cycle were those for which the ratio, body weight (lb) kcals in 12 minutes ≈ 1 . Based on this ratio, the best performances are 11:01 for men, 12:18 for women. These findings reinforce those indicated in Figures 4 and 5. Individuals who can run 1.5-miles in times less than these certainly should not use the cycle as part of their PRT. Theoretically, predicted performance on the cycle should match that on the 1.5-mile run across the full range of possible performances. Equation 1 clearly does not do this. To explore

7

such a model, the regression of body weight (lb) kcals in 12 minutes and gender was calculated with an intercept equal to zero. The resulting equation was: 1.5 - mile run time = 8.376 ×

body weight (lb) + 1.438 × gender(0,1) kcals in 12 minutes

(Equation 3).

The multiple regression coefficient for this equation was 0.695, and the standard error of estimate was 1.899, almost double that of Equation 1. It was decided to recommend Equation 1 rather than this one (Equation 3) because Equation 1 offered more accurate prediction over the range of performances exhibited by our study participants, and by extension, the bulk of Navy personnel. Comparison With Other Tests Based on these results, prediction of 1.5-mile run time from energy expenditure during a 12minute bout on the stationary cycle compares well with that for the elliptical trainer test. The standard error of estimate for the elliptical trainer prediction of run time is 0:57 and that for prediction from the stationary cycle is 1:03. The stationary cycle test appears to have greater validity that the other Navy PRT alternative, the 500-yd swim. Hodgdon reanalyzed the data used to develop the 500-yd swim performance standards (Buono, 1987), and found the standard error of estimate for the prediction of run time from swim time to be 1:55. This test also should be a more valid indicator of cardiovascular-respiratory capacity than the test previously employed by the Air Force. This 12-minute test is a maximal test, just like the run. The Air Force test is a submaximal test, based on the Åstrand-Rhyming formula, and relies on heart rate after 6 minutes of steady-state work to predict maximal aerobic capacity ( V O2max ). Measured in our laboratory, correlations between V O2max predicted from Åstrand-Rhyming and measured by open-circuit spirometry have values of approximately 0.5 (Hodgdon, unpublished data). Air Force researchers have found the Air Force Cycle Ergometry Test to be difficult to administer (too many unusable tests due to failure of the heart rate to stabilize or achieve sufficient magnitude) and not accurate enough (De Wolf, Chao, Richardson, Constable, & Flatten, 1997; Flatten, De Wolfe, Constable, O’Dowd, & Richardson, 1997) for Air Force purposes. The cycle ergometry test is now an alternative for Air Force personnel. Performance

8

on a 1.5-mile run is now the principal method of assessing aerobic capacity for Air Force personnel. The Army cycle test consists of measurement of the time required to pedal 6.2 mi (10 km) on a Monark cycle ergometer (Monark Exercise AB, Vansboro, Sweden) with the resistance set at 2 kp for all individuals, or to pedal 6.2 mi on a bike with only one gear engaged (Secretary of the Army, 1998). The time measured is not adjusted for body weight, and the resistance is the same for all individuals. These tests are biased in favor of larger individuals who perform better because they can apply a greater muscle mass to accomplish this fixed task. CONCLUSIONS In this study, we developed an equation to predict 1.5-mile run time from the indicated energy expenditure during a 12-minute exercise bout on a stationary exercise cycle, specifically, the Life Fitness model 95Ci, which is most commonly used in Navy fitness centers. The predictive accuracy of this test was similar to that of the Navy’s currently employed 12-minute elliptical trainer test, and better than that of the Navy’s currently used swim test. Predictions using the equation were not biased by height, weight, fat mass, fat-free mass, or body fat content. However, we found that predicted run times were slightly overestimated for those whose run times were less than 12.5 minutes, and slightly underestimated for individuals whose run times were greater than 12.5 minutes. Use of the stationary cycle for PRT testing provides a viable mode for testing individuals who cannot run due to physical limitations on weight-bearing activities or who do not like to run and prefer to train using the stationary cycle.

9

REFERENCES

Buono, M. J. (1987). Validity of the 500 yard swim and 5 kilometer stationary cycle ride as indicators of aerobic fitness (NHRC Report No. 87-27). San Diego, CA: Naval Health Research Center. , San Diego, CA. Deputy Chief of Naval Operations: Manpower, Personnel, Training and Education. (2005). Physical Readiness Program (Chief of Naval Operations Instruction 6110.1h). Washington, DC: Department of the Navy. De Wolfe, G., Chao, S., Richardson, D., Constable, S., & Flatten, P., (1997). Invalid cycle ergometry assessment outcomes at five Air Force bases (Report AL/PS-TR-19970056). Brooks AFB, TX: U.S. Air Force Armstrong Laboratory. Flatten, P., De Wolfe, G., Constable, S., O’Dowd, R., & Richardson, D. (1997). Modifications to the Air Force Cycle Ergometry protocol: Impact on pass, fail, and invalid outcomes (Report AL/PS-TR-1997-0165). Brooks AFB, TX: U.S. Air Force Armstrong Laboratory. Hodgdon, J. A., & Beckett, M. B. (1984b). Prediction of percent body fat for U.S. Navy men from body circumferences and height (NHRC Report No. 84-11). San Diego, CA: Naval Health Research Center. Hodgdon, J. A., & Beckett, M. B. (1984a). Prediction of percent body fat for U.S. Navy women from body circumferences and height (NHRC Report No. 84-29). San Diego, CA: Naval Health Research Center. McArdle, W. D., Katch, F. L., & Katch, V. L. (1991). Exercise physiology: Energy, nutrition, and human performance 3rd ed. Philadelphia, PA: Lea & Febiger. Parker, S. B., Griswold, L., & Vickers, R. R., Jr. (2006). Development of an elliptical training physical fitness test (NHRC Report No. 06-06). San Diego, CA: Naval Health Research Center. Secretary of the Air Force. (2006). Fitness program (Air Force Instruction 10-248). Washington, DC: Department of the Air Force. Secretary of the Army. (1998). Physical fitness training. Field manual no. 21-20. Washington, DC: Department of the Army.

10