Predicting maximal oxygen uptake from treadmill testing in trained and

)N

y \L

Predicting maximal oxygen uptake from treadmill testing in trained and untrained women A. Daniel Martin, PhD; Morris Notelovitz, MD, PhD: Carol Fields, MPE: and Joseph O'kroy, BSb Gainesville, Florida This study was conducted to develop an equation to predict maximal oxygen uptake from exercise time during a standardized treadmill test in women aged 29 to 75 years before and after aerobic training. Treadmill tests were administered to 181 women with measurement of maximal oxygen uptake, and an equation predicting maximal oxygen uptake was derived: maximal oxygen uptake (mi· kg-' . min-') = 10.34 + 1.29 (exercise time), r = 0.88, standard error of the estimate = 2.1, P < 0.0001. Thirty-three women were retested after 6 and 12 months of aerobic exercise training. Maximal oxygen uptake was predicted from the equation developed and compared with the measured values at 6 and 12 months. The predicted and measured maximal oxygen uptake values after training were not significantly different. These results indicate that it is possible to predict maximal oxygen uptake for clinical purposes with a single equation from exercise time in untrained and trained women across a wide range of ages. (AM J OBSTET GVNECOL 1989;161:1127-32.)

Key words: Maximal oxygen uptake, exercise testing, aerobic training

Maximal treadmill testing is frequently conducted to establish exercise prescriptions and to evaluate the results of aerobic exercise training programs. Maximal oxygen uptake (Vo 2 max) is considered the best indicator of aerobic fitness, but its measurement requires expensive equipment and trained operators. Another potential problem with the measurement of Vo 2 max is that the technique requires the subjects to wear a nose clip and breathe through a mouthpiece, which limits the ability to communicate during the test. One means of estimating aerobic fitness, without the problems associated with the measurement ofVo 2 max, uses regression equations to predict Vo 2 max from exercise time and other parameters. Several popular protocols have corresponding equations to predict Vo 2 max, but most of these equations were based on male or cardiac patient populations. \·9 Because of the known differences in the aerobic power of men and women and cardiac patients versus healthy subjects, there is a need to develop and validate clinically useful equations to predict Vo 2 max from exercise performance in healthy women across a wide range of ages. We present the results of maximal treadmill exercise testing in 181 women beFrom the Department of Physical Therapy, College of Health Related Professions, Uni,'ersity of Florida," and The Center for Climacteric Studies, Inr." Supported in part by a grant from Nautilus Sports Medicine, Inc. Received for publication February 3, 1989; revised May 15, 1989; accepted June 1, 1989. Reprint requests: Norris Notelovitz, MD, PhD, The Center for Climactrric Studies, Inc., 222 S. W. 36th Terrace, Gainesville, FL 32607. 6/1/14419

tween the ages of 29 and 75 and develop an equation to predict Vo 2 max from the results of maximal treadmill exercise testing before and after aerobic training.

Method Subjects were recruited as part of a study to investigate the relationship between exercise and bone mineral density. The study protocol called for a series of maximal exercise tests to be administered as part of the initial evaluation. The first test was conducted to screen subjects for electrocardiographic abnormalities (;;.2mm of ST segment depression, frequent ventricular arrhythmias, or conduction disturbances) and other cardiovascular symptoms. Oxygen uptake was not measured during the first test. One hundred eighty-one subjects were judged to have normal responses to maximal exercise and were administered a second maximal test, approximately 1 week later, which included measurement of Vo 2 max. Thirty-three subjects were also tested after 6 and 12 months of aerobic training to determine if the equation developed from the pre training data could accurately predict Vo,max after exercise training. None of the subjects were taking f3-adrenergic receptor-blocking medications or other medications likely to influence the results of exercise testing. Subjects fasted for 12 hours before testing, and informed consent in accordance with institutional policy was obtained before data collection. Testing was conducted on a Quinton motorized treadmill following a protocol starting at a speed of 3.0 mph (80.5 m/min) and 1% grade. to The speed remained constant throughout the test, and grade was

1127

1128

Martin et al.

November 1989 Am J Obstet Gynecol

Volume 161 Number 5

50

......-. ,.... c: II

E ,.....

•

40

•

•

I

C)

-. ~

E ........ >< ca E

C\I

•

•

30

•

20

•

••

I

0

-> 4

8

12

16

20

24

Exercise Time (min) Fig. 1. Scatter plot of exercise time and Vo 2 max for validation subjects with regression line. Vo 2 max (mi· kg-I. min-I) = 10.34 + 1.29 (exercise time), r = 0.88, standard error of the estimate = 2.1, P < 0.0001.

Table I. Physical characteristics of the subjects Cross-validation (N = 60)

Age (years) Weight (kg) Height (cm)

54.8 ± 0.7 60.8 ± 0.6 161.3 ± 0.4

55.0 ± 0.8 60.2 ± 0.7 161.8 ± 0.6

54.2 ± 1.0 62.0 ± 0.9 161.0 ± 0.8

Values are shown as mean ± SEM.

increased by I % every minute until the subjects became exhausted or oxygen uptake did not increase despite increases in the grade of the treadmill. Subjects walked throughout the entire test and were not allowed to support themselves on the handrail at any time. Treadmill speed and grade were calibrated according to the manufacturer's directions. Metabolic measures were recorded with a Beckman metabolic measurement cart and included Vo 2 max, maximal carbon dioxide production, maximal ventilation, and respiratory exchange ratio. Exercise time was recorded to the nearest tenth of a minute. The carbon dioxide and oxygen analyzers and the volume measuring turbine in the metabolic measurement cart were regularly calibrated according to the manufacturer's instructions. An electrocardiogram and heart rate were monitored every minute with a Quinton electrocardiograph.

Simple, multiple, and polynomial regression models predicting Vo 2 max were calculated with a General Linear Models procedure in the Statistical Analysis System. II In addition to exercise time, age, height, and weight were also evaluated for inclusion into the prediction equations. A validation and cross-validation regression analysis algorithm was used to generate and verify the appropriateness of the regression equation. A validation regression equation, which was based on a random sample of approximately two thirds . (N = 121) of the subjects, was computed. This equation was then used to predict the Vo 2 max of the cross-validation subjects (N = 60). The cross-validation group's measured Vo 2 max and the Vo 2 max predicted by the validation equation were then compared with correlation analysis and a paired t test. The ability of the validation equation to accurately predict Vo 2 max after exercise training was evaluated in 33 of the validation group subjects after 6 and 12 months of exercise training. The training consisted of three exercise sessions per week on either treadmills or stationary cycle ergo meters for 20 minutes at 70% to 85% of measured maximal heart rate. Maximal oxygen uptake was measured as described earlier. The exercise time achieved on the 6- and 12-month tests were used to calculate a predicted Vo 2 max with the validation equation, which was then compared to the measured

Table II.

ET, Exel respiratory

Vo 2 max a tical sign if the mean

Results

There, cal charac validation No com testing rei evidenced ratio, vent There we

November 1989

J Obstet Gynecol

Predicting maximal oxygen uptake

Volume 161 Number 5

-

1129

T ""

•

c: .-

E.

42

T""

•

C)

.

-E

35

>< ca

28

~

E

N

0 .>

21

"C

Q) ...,

(.) ."C

Q)

a-

D.

14 7 7

14

21

28

35

42

Measured V02max (ml.kg- 1.min- 1) Fig. 2. Relationship between predicted and measured V02max for cross-validation group. The difference between the two measures was not significant (P = 0.52); however, the correlation between measured and predicted V0 2max was significant, r = 0.S9, standard error of the estimate = 1.7, P < 0.0001. ~ssion

models General LinAnalysis Sysheight, and into the pre)ss-validation generate and ion equation. b. was based f two thirds l. This equa2max of the lss-validation ax predicted mpared with

to accurately 'as evaluated ter 6 and 12 consisted of treadmills or es at 70% to limal oxygen The exercise ts were used Ie validation Ie measured

Table II. Results of exercise testing Total sample (N = 181)

Vo 2 max (Umin) Vo2 max (mi· kg-I. min-I) ET(min) HRmax (beats/min) Ve max (Umin) Ve ma jVo2max RER (Vco 2/Vo 2max)

1.575 26.1 12.2 175.S 64.S 41.4 l.lS

± 0.020 ± 0.3

± 0.2 ± 1.0

± 0.9 ± 0.45 ± 0.01

Validation (N = 121)

1.554 26.0 12.1 175.9 64.0 41.5 l.l9

0.024 0.4 0.3 1.2 ± l.l ± 0.6 ± 0.01

± ± ± ±

Cross-validation (N = 60)

1.617 26.2 12.4 175.6 66.4 41.2 l.l7

± 0.034 ± 0.5 ± 0.4

± 1.5 ± 1.6 ± 0.7 ± 0.02

ET, Exercise time; HR_, maximal heart rate; Ve_, maximal ventilation; Ve_/Vo 2 max, ventilatory equipment for oxygen; RER, respiratory exchange ratio; Veo2 , maximal carbon dioxide production. Values are shown as mean ± SEM.

Vo 2max achieved on the 6- and 12-month tests. Statistical significance was set at p < 0.05; data are shown as the mean ± SEM. Results

There were no significant differences in the physical characteristics between the validation and crossvalidation groups (Table I). No complications resulted from exercise testing. The testing represented maximal effort for the subjects as evidenced by the high values for respiratory exchange ratio, ventilatory equivalent for oxygen, and heart rate. There were no significant differences in the results

of exercise testing between the validation and crossvalidation groups (Table II). The age, height, and weight of the 33 subjects who trained for 12 months were as follows: 58.9 ± 1.3 years, 151.0 ± 4.9 cm, and 67.9 ± 5.1 kg. The subjects in the training group significantly increased (p < 0.05) their Vo 2 max values from the pretraining period (26.5 mi· kg-I. min-I) to 6 months (29.2 ml . kg-I. min-I); however, the 6- and 12-months (29.0 ml . kg-I. min-I) values were not significantly different (p > 0.05) from each other. The improvement in Vo 2 max was similar (p > 0.05) in cycle and treadmill groups. The validation and cross-validation scheme of regres-

1130

Martin et al.

November 1989 Am

J Obstet Gynecol

42

,.. I

C

.-

35

E ,..

.

I

C) ~

28

E

•

>< as

E

('II

0

tJ

Pretraining 6 months 12 months

21

.> 14 4

8

12

20

16

24

Exercise Time (min) Fig. 3. Relationship between exercise time and Vo 2 max before training and after 6 and 12 months of training. Pretraining line represents Vo 2 max predicted by the validation equation as a function of time, and the symbols show the relationship between measured Vo 2 max and exercise time for 33 subjects after 6 and 12 months of aerobic exercise training. There were no significant differences in the measured and predicted Vo 2 max when the 6-month (r = 0.90) and 12-month (r = 0.88) exercise times were used in the validation equation, indicating that a single equation can be used to predict Vo 2 max from exercise time for both trained and untrained women.

sion analysis was conducted to verify the validation regression formula with data collected from a similar subject sample. The Vo 2 max of the validation group was predicted with a simple regression equation with exercise time as the independent variable; age, height, and weight did not significantly contribute to the equation, and the polynomial equations were not statistically significant. The relationship between exercise time and Vo 2 max for the validation group is shown in Fig. 1. The equation to predict Vo 2 max, which is based on the 121 validation subjects, is: Vo 2 max (mi· kg-I. min-I) = 10.34 + 1.29 (exercise time), r = 0.88, standard error of the estimate = 2.1, P < 0.0001. When the validation equation was used to predict Vo 2max for the cross-validation group, the correlation between the cross-validation group's measured Vo 2 max and the Vo 2 max predicted by the validation equation was r = 0.89, standard error of the estimate = 1.7 mi· kg-I. A paired t test between these two variables revealed a mean difference of 0.15 mi· kg-I. min-I, which was not significant (p = 0.52). A scatter plot of the cross-validation group's pretraining measured and predicted Vo 2 max is shown in Fig. 2. Fig. 3 shows the regression line predicting Vo 2 max from exercise time based on the validation equation (pretraining), and

the relationship between exercise time and measured Vo 2max following after 6 and 12 months of exercise training in 33 subjects. The validation equation was used to predict Vo 2 max, using the exercise time achieved at 6 and 12 months; when compared with the measured Vo 2 max, the correlations were r = 0.90 and r = 0.88 at 6 and 12 months, respectively. The differences between the Vo 2 max predicted by the validation equation (using the exercise-time achieved at 6 and 12 months) and the Vo 2 max measured after 6 months (0.12 mi· kg-I. min-I, p = 0.65) and 12 months (-0.29 mi· kg-I. min-I,p = 0.31) of training were not statistically significant. Comment

American women of all ages are becoming more familiar with the benefits of regular aerobic exercise: increased aerobic power and endurance, 12 decreased body weight and body fat,13 lower levels of blood cholesterol, triglycerides, and increased high-density lipoprotein cholesterol. 14 Many women are seeking medical advice for specific recommendations with regard to exercise programs and objective means of assessing their progress. A maximal exercise test performed before an exercise program is begun can be used not only

Volume 161 Number 5

to detenT segment ( determinl appropri, Repeat eJ creasing. testing pi their trail will em ph effortless predicted achieved women a( fromexel of assessi difficultie fractions. The pi propriate The moc creases \I duration training (' with a mt after 6 a and 14.6 ercised t{ the high change r. (Table II reported ages. 2 . 3 The st. which is mi· kg-I regresslO dard err< band of Vo 2 max. ror of th favorabl) based on protocols exercise 1 ror of th( et al.I pr Bruce pI above av tween ex dard en Froelicht found a { of 0.87 mi· kg-I test resul

vember 1989 bstet Gynecol

neasured exercise :tion was 'ise time with the 0.90 and Ie differe validaved at 6 ed after and 12 training

nore faTise: in~creased

lod chonsity ling medregard ssessing med benot only

Volume 161 Number 5

to determine the presence or absence of ischemic ST segment depession and serious arrhythmias but also to determine the maximal heart rate for prescribing the appropriate exercise intensity and estimate Vo"max. Repeat exercise tests can be an effective means of increasing adherence to exercise programs. Subsequent testing provides positive feedback to the subjects that their training efforts have resulted in improvement or will emphasize that gains in aerobic fitness will not come effortlessly. Our results indicate that Vo"max can be predicted with good accuracy from exercise time achieved with this protocol in untrained and trained women across a wide range of ages. Predicting Vo,max from exercise time can provide a clinically useful means of assessing aerobic fitness without the expense and difficulties of measuring ventilation and expired gas fractions. The protocol used in this study appears to be appropriate for testing sedentary and trained women. The moderate constant speed and gradual grade increases were well-tolerated by the subjects, and the duration of the tests was not unduly long. The pretraining exercise time ranged from 3.6 to 21.1 minutes, with a mean of 12.2 minutes. The mean exercise times after 6 and 12 months of training were 14.5 ± 0.50 and 14.6 ± 0.55 minutes, respectively. Our subjects exercised to maximal levels of exertion as evidenced by the high values for maximal heart rate, respiratory exchange ratio, and the ventiltory equivalent for oxygen (Table II). The Vo,max values were similar to those reported by other workers with women of similar ages. 2 ,3 The standard error of the estimate of the equation, which is based on the validation subjects, was 2.1 mi· kg-I. min-I. The 95% confidence limit around a regression equation can be approximated by ± 2 standard error of the estimate, yielding a 95% confidence band of 4.2 mi· kg-I. min" 1 around the predicted Vo 2 max. The correlation coefficient and standard error of the estimate found with our subjects compare favorably with equations used to predict oxygen uptake based on exercise time achieved during various testing protocols. Bruce et aI.' found a correlation between exercise time and Vo"max of 0.91, with a standard error of the estimate of 3.5 ml . kg- 1 • min- I.' Alexander et al. I predicted Vo,max from exercise time with the Bruce protocol in a sample of 25 young subjects with above average fitness and found the correlations between exercise time and Vo"max to be 0.95 with a standard error of the estimate of 3,0 mi· kg-I. min I. Froelicher et al." tested 77 military volunteers and found a correlation between exercise time and Vo,max of 0.87 with a standard error of the estimate 4.7 mi· kg-I. min with the Bruce protocol, and the Balke test resulted in r = 0.80 and standard error of the es-

Predicting maximal oxygen uptake

1131

timate of 4.0 ml . kg-I. min -I." Foster et al." found a correlation of 0.97 with a standard error of the estimate of 2.6 ml . kg- 1 • min -I between exercise time and Vo 2 max with a group of subjects that included cardiac patients and healthy and trained subjects. The equation developed from the validation subjects was validated by predicting Vo,max for the crossvalidation group by means of the equation developed with the validation group data and comparing this predicted value with the cross-validation group's measured Vo,max. The measured and estimated measures of Vo,max were not statistically or practically different at baseline or after 6 or 12 months of training. Bruce et aI.' found different equations to be necessary when predicting Vo 2 max in healthy sedentary and trained men, but our data indicate that our equation is appropriate for women regardless of training status. Although our equation has good predictive capability, there are some restrictions to its use. The equation should not be used in tests that last less than approximately 7 minutes or longer than approximately 20 minutes. The number of data points outside the 7- to 20-minute range is small, and it is possible that our equation will not fit the data obtained from tests not ending within the 7- to 20-minute range. For the Vo,max predicted by this equation to be accurate, it is vital that subjects not support themselves on the handrails or otherwise reduce the weight supported by the legs. Allowing subjects to support some of their body weight with the arms during the test can cause Vo"max to be significantly overestimated." 16 The use of 13adrenergic blocking medications may also alter the relationship between exercise test and Vo"max. Hughson and Smyth '7 have shown that the use of f3-adrenergicblocking medications slows the kinetics of oxygen uptake; use of this formula to predict the Vo"max of subjects taking f3-blockers may result in Vo,max being overestimated. The applicability of our equation for use with female cardiac patients or with male subjects is not addressed in this study. In conclusion, we have found and validated a strong predictive relationship between treadmill exercise time and Vo,max in trained and untrained women ranging in age from 29 to 75 years. Provided the assumptions under which the data were collected are followed, it is possible to predict Vo,max with sufficient accuracy for clinical purposes with only a treadmill and electrocardiograph. REFERENCES 1. Alexander JF, Liang MT, Stull GA, Serfas Re, Wolfe DR, EwingJ L A comparison of the Bruce and Liang equations for predicting Vo,max in young male adults. Res Q 1984;55:383-7. 2, Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake and nomographic assessment of functional aerobic im-

Martin et al.

3. 4.

5.

6.

7.

8.

9.

10.

pairment in cardiovascular disease. Am Heart J 1973; 85:546-62. Drinkwater BL, Horvath SM, CL Wells. Aerobic power of females, ages 10 to 68. J Geriatr 1975;30:385-94. Froelicher VF, Thompson AJ, Noguera I, Davis G, Stewart AJ, Triebwasser JH. Prediction of maximal oxygen consumption: comparison of the Bruce and Balke treadmill protocols. Chest 1975;68:331-6. Naughton J. The National Exercise and Heart Disease Project: development, recruitment, and implementation. In: Brest AN, ed. Cardiovascular clinics. Philadelphia: FA Davis, 1978 vol 9:205-22. Haskell WL, Savin W, Oldridge N, DeBusk R. Factors influencing estimated oxygen uptake during exercise testing soon after myocardial infarction. Am J Cardiol 1982;50:299-304. Pollock ML, Bohannon RL, Cooper KH, et al. A comparative analysis of four protocols for maximal treadmill stress testing. Am Heart J 1976;92:39-46. Rod JL, Squires RW, Pollock ML, Foster C, Schmidt DH. Symptom-limited graded exercise testing soon after myocardial revascularization surgery. .I Cardiac Rehab 1982; 2: 199-205. Sullivan M, McKirnan MD. Errors in predicting functional capacity for post-myocardial infarction patients using a modified Bruce protocol. Am HeartJ 1984; 107:48692. Notelovitz M, Fields C, Caramaelli K, Dougherty M,

November 1989 Am J Obstet Gynecol

11. 12.

13.

14.

15.

16.

17.

Schwartz AL. Cardiorespiratory fitness evaluation in climacteric women: an evaluation of two methods. AM J OBSTET GYNECOI. 1986;154:1009-13. SAS Institute. SAS user's guide. Version 5: statistics edition. Cary, North Carolina: SAS Institute, 1985. Pedersen PK, Jorgensen K. Maximal oxygen uptake in young women with training, inactivity and retraining. Med Sci Sports 1978;10:233-7. Lewis S, Haskell WI, Wood PD, Manoogian, Bailey .IE, Perira M. Effects of physical activity on weigh reduction in obese middle-aged women. AmJ Clin Nutr 1976; 29:151-6. Moore CE, Hartung GH, Mitchell RE, Kappus CM, Hinderlitter J. The relationship of exercise and diet on highdensity lipoprotein cholesterol levels in women. Metabolism 1983;32: 189-96. Foster C, Pollock ML, Rod JL, Dymond D, Wible G, Schmidt DH. Evaluation of functional capacity during exercise radionuclide angiography. Cardiology 1983;70:8593. Zeimetz GA, Mcneill JF, HallJR, Moss RF. Quantifiable changes in oxygen uptake, heart rate, and time to target heart rate when hand support is allowed during treadmill exercise. J Cardiopulmon Rehab 1985;5:525-30. . Hughson RL, Smyth GA. Slower adaptation of Va, to steady state of sub maximal exercise with beta-adenergic blockade. Eur .I Appl PhysioI1983;52:107-10.

Candida albicans: Cellular immune system interactions during different stages of the menstrual cycle Aliza Kalo-Klein, PhD, and Steven S. Witkin, PhD New York, New York Candidal vaginitis most often recurs during pregnancy and in the late luteal phase just before menstruation. We examined the influence of the stage of the menstrual cycle on the cellular immune response to Candida albicans, the efficiency of C. albicans germination in sera, and the ability of products from activated lymphoid cells to inhibit germination. C. albicans germination was maximal in sera obtained during the luteal phase. During this period the cellular immune response to Candida was reduced as was the inhibition of Candida germination by products of activated peripheral blood mononuclear cells. Variations in immune response to Candida were of much lesser magnitude in women who took oral contraceptives, which suggests that it was the marked fluctuation in progesterone or estradiol levels during the menstrual cycle that influenced the changes in the immune response to C. albicans. Thus the hormonal status of women may influence the pathogenicity of C. albicans by modulation of immune system activity. These results explain the clinical observation that candidal vaginitis infections most frequently reappear before menstruation. (AM J OSSTET GVNECOL 1989;161 :1132-6.)

Key words: Candida albicans, vaginitis, menstrual cycle, immune response

From the Immunology Division, Department of Obstetrics and Gynecology, Cornell University Medical College. Received for publication March 8, 1989; accepted l'vIay 16, 1989. Reprint requests: Dr. Steven S. Witkin, Cornell University Medical College, Department of Obstetrics and Gynecology, 515 E. 71st St., New York, NY 10021.

6/1 114000

1132

Candidal vaginitis is a common disorder caused by the opportunistic yeast-like fungi Candida albicans. I. ~ It poses a serious problem because clinical infections tend to be recurrent, are not always eradicated by standard treatments, and are often unexplained." I Hence, there is a need for a greater understanding of the involved

Volume IE Number 5

Table I. Sui

Table II contrae( SUI

pathoge to its ini The, nant wc addition contrae( didiasis contrac( vaginiti~

These 0 and prc infectiol thelial c tion, is last wee infiuenc fections to und( transien Wen both th the abil Matel Subje a histor day me contrae< triphasi biphasi( was obt of estra dioimm Lyml Periphe