am. j. epidemiol. 1985-siconolfi-382-90

AMERICAN JOURNAL OF EPIDEMIOLOGY Vol. 121, No. 3Copyright © 1985 by The Johns Hopkins University School of Hygiene and Public Health Printed in U.SA.All rights reserved

A SIMPLE, VALID STEP TEST FOR ESTIMATING MAXIMALOXYGEN UPTAKE IN EPIDEMIOLOGIC STUDIES1

STEVEN F. SICONOLFI,2 CAROL EWING GARBER, THOMAS M. LASATER ANDRICHARD A. CARLETON

Siconolfi, S. F. (The Memorial Hospital, Pawtucket, Rl 02860), C. E. Garber, T.M. Lasater and R. A. Carteton. A simple, valid step test for estimating maximaloxygen uptake in epidemiologic studies. Am J Epidemiol 1985; 121:382-90.

The authors' modification of the Astrand-Rhyming Cycle Ergometer Test is ofshort duration, has low initial and peak work rates and was in an earlier studyapplied for population fitness testing (N = 587) at a survey center after othercardiovascular risk factor measures were obtained in the home. To add fitnesstesting in the home, the authors have designed a safe, brief 10 inch (25.4 cm)high step test for estimating maximal oxygen uptake (V02mu). Measured maximaloxygen uptake for step tests has been shown to be approximately 10% higherthan that reported for cycle tests. All test Instructions and stepping rates wereincluded on a cassette tape; heart rates were monitored by a digital tachographduring the last 30 seconds of stepping. Maximal oxygen uptake was measureddirectly on a bicycle, estimated by the step test, and measured by the authors'bike test in 48 men and women aged 19-70 years who took part in a communityfitness program in Pawtucket, Rhode Island in January-February 1983. No signif-icant differences in maximal oxygen uptake were found between the bicycleprotocols. The step test estimate of maximal oxygen uptake (VOi™*) was signif-icantly higher (12%) than directly measured WO3mmx, reflecting the expecteddifference between stepping and cycling. The correlation between direct andboth estimates was 0.92. The cross-validation correlation between the estimateswas 0.98. The authors' protocol provides accurate estimates of maximal oxygenuptake and is safe and suitable for in-the-home assessment of fitness of peopleaged 19-70 years for epidemiologic studies.

exercise, physical; physical fitness

Physical fitness has been reported to be factors. Additionally, physical fitness influ-inversely related to coronary disease inci- ences other coronary disease risk factors—dence (1). Consequently, it is important to smoking prevalence (2), overweight (3),estimate physical fitness in any epidemio- high density lipoproteins (4), and bloodlogic study of coronary heart disease risk pressure (2). Maximal oxygen uptake

Received for publication January 16, 1984 and in * Current address: Department of Physical Educa-final form May 1, 1984. tion, Health and Recreation Studies, Lambert Field

Abbreviations: Kgmmin"1, kilogram meter per House, Rm. 113, Purdue University, West Lafayette,minute; 1-min"1, liter of oxygen per minute; 1 MET, IN 47907. (Send reprint requests to Dr. Siconolfi at3.4 milliliters of oxygen per kilogram of body weight this address.)per minute; ml-kg"'-min"1, milliliters of oxygen per This work was supported in part by Grant HLkilogram of body weight per minute; VO2mM, maximal 23629 from the National Heart, Lung, and Bloodoxygen uptake. Institute.

1 From the Division of Cardiology and of Health The authors thank the following people for tech-Education, The Memorial Hospital, and the Depart- nical support in this project Mary Marcos, Robertment of Medicine and Community Health, Brown Snow, Paula Beaudin, Julie Durakis and George Mis-University, Pawtucket and Providence, RI. sailidis, Jr.

382

by guest on August 7, 2012

http://aje.oxfordjournals.org/D

ownloaded from

http://aje.oxfordjournals.org/

STEP TEST ESTIMATES OF MAXIMAL OXYGEN UPTAKE 383

(VO2mM) is the most widely acceptedmeasure of fitness. This measure repre-sents the maximum rate of oxygen trans-port to exercising muscles. Maximal oxygenuptake is influenced by ventilation, cardiacoutput, vascularization, and oxygen utili-zation by muscles. In short, it is a singlemeasure of the working capacity of thecardio-respiratory-va8cular systems as aunit. Regular physical activity has beenshown to increase this measure of fitness(5).

Measurement issues have made the in-clusion of physical fitness difficult in epi-demiologic studies. Direct measurement ofmaximal oxygen uptake is expensive, timeconsuming, and aversive to respondents. Inaddition, it requires a fixed facility, becausethe equipment is not very portable. In 1980,we developed a submaximal bicycle test touse in a population survey (6). The test wasa modification of the Astrand-RhymingCycle Ergometer Test (5) and was not asaversive or as time consuming as directmeasurement. However, the fitness teststill required a separate appointment andvisit to our Survey Center, whereas the restof the survey was conducted^in the homesof our randomly selected sample. The needfor travel and other inconveniences for the587 respondents resulted in a reduction ofthe in-home response rate of 75-55 percent. The obvious answer to the problemwas to develop a simple, safe, and valid testthat could be administered in the home.The result was the step test described inthe present paper. In addition to the vali-dation of the new step test against directlymeasured maximal oxygen uptake, it wasalso important for us to cross-validate thestep test against the cycle ergometer test aswell. These data are also included in thepresent report.

MATERIALS AND METHODS

Subjects

The subjects were 48 men and womenaged 19-70 years, who participated in acommunity fitness program in Pawtucket,

Rhode Island in January-February 1983and who volunteered to participate. Uponarrival at the laboratory, subjects com-pleted a brief medical questionnaire de-signed to screen for cardiovascular contra-indications to exercise. Volunteers on med-ications known to affect heart rate wereexcluded, as were subjects with symptomsconsistent with angina pectoris, or with ahistory of prior myocardial infarction, orwith a resting blood pressure over 180/100mmHg.

Design

The usual way to evaluate a submaximalexercise test for validity of estimates ofmaximal oxygen consumption is to com-pare the estimates to directly measured ox-ygen uptake values obtained during maxi-mal exercise with the same exercise modal-ity. In the present study, we chose a bicycletest protocol to ascertain maximal oxygenconsumption rather than a step test pro-tocol. The use of a bicycle protocol in theolder and inactive adult increases safety byallowing smaller gradations in work and ahigher quality electrocardiogram. In addi-tion, the logistics of oxygen consumptionare far more difficult with a moving (e.g.,step test) compared with a stable (e.g.,treadmill, bicycle test) head. Moreover, itwas important for us to relate to our earlierwork (6) and to determine the cross-vali-dation between the step test and the sub-maximal cycle ergometer test. Others (7-10) have consistently shown that peak ox-ygen uptake measured during bicycle exer-cise bears a predictable relationship to thatmeasured using treadmill protocols. Valuesduring pedalling exercise average 11 percent (range, 8-15 per cent) lower than withtreadmill exercise (7-10). Kasch et al. (11)have shown that maximal oxygen consump-tion during stepping and treadmill exerciseagreed within an average difference of 0.5per cent (mean 11 per cent; range 8-15 percent lower than treadmill running) (7-10).These data and our own extensive experi-ence with maximum bicycle exercise tests



ownloaded from


384 SICONOLFl ET AL.

led us to use the bicycle test for validatingthe step test.

Subjects performed a sub-maximal pedalergometer test according to our previousprotocol (6) followed by a ten minute restperiod. They then performed a volitionalmaximum exercise test on the same pedalergometer. After at least 24 hours and notgreater than seven days, all subjects re-turned to the laboratory for the administra-tion of the step test protocol. The step testwas not given the same day to reduce anyeffects of fatigue. Informed consent wasobtained from all participants.

Laboratory methods

Heart rates were monitored with an elec-trocardiograph and recorded every minuteusing a CM6 lead. The cycle ergometer(Monark, Quinton Instruments, Seattle,WA) was calibrated prior to testing with a1 kg weight. The cycle ergometer was ad-justed so that the leg had a slight bend atthe knee when the pedal was in the downposition. The pedal rate was kept constantfor both the cycle tests at 50 rpm with theaid of a metronome. Oxygen uptake (VO2)was measured only during the maximumexercise test. The VO2 was measured witha pneumotachograph (Hewlett-Packard,Vertek Series, Lexington, MA), and S-3AOxygen Analyzer (Applied Electrochemis-try, Inc., Sunnyvale, CA), and a BeckmanLB-2 Medical Gas Analyzer (Beckman,Schiller Park, IL). Analyzers were cali-brated before each test with a gas thatpreviously had concentrations of oxygenand carbon dioxide chemically analyzed bythe Scholander method (12).

In addition to heart rate monitoring withan electrocardiograph, heart rates duringthe stepping protocol were also measuredwith a digital tachometer (Pulseminder,Carolina Products, Inc., Hempstead, NY)attached by a simple clip to the ear lobe.

Submaximal exercise test protocols

The submaximal cycle ergometer test hasbeen previously described (6). Briefly, thistest consists of a modification of the As-

trand-Rhyming Test (13). Work rates areprogressively increased until subjects reach70 per cent of their maximum heart rate.Upon reaching this target heart rate, sub-jects are then allowed to obtain steadystate. Steady state was concluded to bepresent when consecutive (one minuteapart) heart rates differed by <5 beats-min"1. The mean steady state heart rateused was the average of the last two heartrate readings. The VO2mai (liter-min"1:liters of oxygen consumed per minute) wasestimated from the Astrand-RhymingNomogram (13) using the mean steadystate heart rate and the final exercise rate.The age correction factor developed by I.Astrand (5) was not used. Separate regres-sion equations for male and female subjectsthat we had previously developed in pilotwork (n = 50) were used to correct theVO2mM. These equations were:

for males: Y = 0.348 (Xi)- 0.035 (X2)+ 3.011

for females: Y = 0.302 (X,)

- 0.19(X2)

(1)

(2)+ 1.593

where Y is the VO2 (liters • min"1), Xi is theVO2 (liters-min"1), from the Astrand-Rhyming Nomogram (13) (not correctedfor age), and X2 is the age in years (6).

The step test protocol consisted of step-ping on a portable 10 inch (25.4 cm) highbraced box (12 in (30.5 cm) wide, 18 in (45.7cm) long and 10 in (25.4 cm) high) for threeminutes per stage for a maximum of threestages. (Detailed plans for step construc-tion are available from the authors on re-quest.) The stepping rates for the stageswere 17, 26, and 34 steps per minute, re-spectively. These stepping rates approxi-mate the MET (1 MET = 3.5 ml of oxygenper kg of body weight per minute) levelsrequired for the first three stages (4, 7, 10METs) of the Bruce treadmill exercise testprotocol. Table 1 summarizes the step testprotocol. During the last 30 seconds of thethird minute of each stage of stepping,



ownloaded from



TABLE 1

Step test protocol for estimating maximal oxygenuptake in 48 men and women aged 19-70 years who

took part in a community fitness program inPawtucket, Rhode Island in January-February 1983

IRest

nRestin

Supping*rate/min

17_26-

34

Metro nometrate/mill

68_

104_

136

EstimatedMET»

4-7

-10

Time(min)

31313

* Stepping is defined as a full body lift (10 in (25.4cm)) with extended leg and returning to the originalposition.

t The metronome rate is set so that each beatrepresents a foot movement (e.g., "up, up, down, down"represents one complete stepping action). By settingthe metronome rate at 68, the respondent would bemaking 17 full ascensions/descensions per minute.

heart rates were recorded at 2:30, 2:45 and3:00 minutes. If the average of these threeheart rates did not equal or exceed 65 percent of the age predicted maximum heartrate (estimated as 220 — age in years), thenthe subject was instructed to complete stage2. Similarly, if the heart rates did not reachthe 65 per cent target level at the comple-tion of stage 2, subjects then completedstage 3. Each stage was separated by atleast one minute of sitting rest. All testinstructions and stepping rates (as metro-nome rates) were included on a cassettetape that was played for the subject afterheart rate monitoring equipment was se-cured to the patient. Heart rates were mon-itored by a sensor attached to the subjectby an ear clip. Estimates of oxygen con-sumption for the last stage completed weremade using the following equations:

Stage 1: VO2 (1/min)= 16.287 X Wt (kg)/l,000



These equations were developed fromthose for stepping given in the AmericanCollege of Sports Medicine Guidelines for

Exercise Testing and Exercise Prescription(14). The VO2 estimated from this submax-imal stepping test and the mean exerciseheart rate were then used to estimateVO2mai from the Astrand-Rhyming Nomo-gram (13). Age-adjusted maximal oxygenconsumption was then computed based onthe regression equations developed earlier(6) for the cycle ergometer test (equations1 and 2).

Maximal exercise test protocol

For the direct measurement of maximaloxygen uptake, subjects pedaled for oneminute with no resistance. Thereafter, theexercise rate was increased by 24.5 Watt(150 Kgm • min"1 (kilogram meters per min-ute)) each minute until the subject couldnot maintain a pedal rate of 50 rpm. Heartrate and oxygen consumption were meas-ured during the last 30 seconds of eachminute. Oxygen uptake was considered tobe maximum when the difference betweenthe final two VO2 measurements was lessthan 0.25 liters-min"1 and the respiratoryquotient for the last measurement wasequal to or greater than 1.0.

Pearson product-moment correlations(15) were computed between the directlymeasured peak oxygen uptake and the es-timates from the submaximal step and thebicycle tests. In addition, correlations werealso computed between the two submaxi-mal estimates of peak oxygen consumption.Differences among the means were ana-,lyzed with a subject by treatment ANOVAdesign (15).

RESULTS

All subjects tested were able to completeboth submaximal exercise tests. The di-rectly measured mean values for maximaloxygen uptake for different age groupsshown in table 2 are very similar to normsprovided by the American College of SportsMedicine (14), suggesting that our sampleis representative of the general population.

Correlations and standard errors of esti-mates of oxygen consumption for the threetests are presented in table 3. The overall



ownloaded from


386 SICONOLFI ET AL,

TABLE 2

Means (± standard deviations) of age, weight, directly measured maximal oxygen uptake (VOjmmJ, and peakheart rate in all subjects, and by age and sex, for 48 men and women aged 19-70 years who took part in a

community fitness program in Pawtucket, Rhode Island in January-February 1983

All subjects

Age (years)19-3435-50

51+

SexMalesFemales

n

48

241410

1929

Age (yean)

37 ±13

27 ±441 ±556±7

41 ±1435 ±11

Weight (kg)

68 ±14

66 ±1569 ±1573 ± 8

82±860±9

Directly measuredVO,_(mlkg-1min-')'

29.1 ± 7.6

31.8 ± 6.929.8 ± 7.521.7 ± 4.1

33.2 ± 8.326.4 ± 5.8

Peek heart rate(beats per minute)

172 ± 14

179 ± 11173 ± 10156 ±12

172 ± 18172 ± 12

' Directly measured VOtn»» by bicycle ergometer.

TABLE 3

Pearson product-moment correlations (r) andstandard error of estimates (SEE) between all values

of VOtmml (liters • min'1) in all subjects, and by age andsex, for 48 men and women aged 19-70 years who

took part in a community fitness program inPawtucket, Rhode Island in January-February 1983

Direct Directmeasure- measure-ment vs. raent vs.«t«p test* bike test*

Biketert*vs. step

test* '

SEE SEE SEE

Allsubjecta 48 0.92 0.30 0.92 0.29 0.98 0.17

Age (years)19-34 25 0.95 0.24 0.93 0.28 0.98 0.1535-50 13 0.95 0.26 0.95 0.27 0.97 0.16

51+ 10 0.89 0.16 0.89 0.16 0.94 0.20

SexMaleFemale

19 0.79 0.43 0.83 0.39 0.94 0.2029 0.72 0.17 0.61 0.19 0.82 0.14

* Estimates of VOSM1.

correlation coefficients for each estimateversus the directly measured peak VO2lMI

were both 0.92. The standard error in pre-dicting the directly measured VO2max was0.30 liters min"1 and 0.29 liters-min"1 forthe step and bicycle protocol, respectively.As discussed subsequently, it is known thatbicycle testing has consistently lower (10per cent) VO2mjU[ when compared to steptesting. The relationship between the di-

rectly measured V02max and that estimatedfrom the stepping protocol is shown in fig-ure 1. In figure 1, the line of "identity" hasbeen offset to reflect this difference. Therelationship for the estimate from the sub-maximal bicycle test to the measured valueis shown in figure 2. The overall correlationcoefficient for the cross-validation betweenthe submaximal exercise tests was 0.98.The standard error in predicting the bicycleestimate from the step estimate was 0.17liters • min"1. The relationship between thetwo estimates is shown in figure 3. In figure3, the line of "identity" has been offset toreflect the difference in exercise modalities.

An analysis of variance was computed forthe three values of peak oxygen consump-tion. This analysis showed that the esti-mate from the step test was significantly(p < 0.01) higher than the directly meas-ured or submaximal bike test estimatedpeak oxygen uptake. As expected, the steptest estimates were approximately 12 percent higher than the directly measured val-ues from the bicycle test for peak oxygenuptake. There was no significant differencebetween the submaximal bicycle estimateand the directly measured peak oxygen up-take. The means and standard deviationsof results from the three protocols areshown in table 4.



ownloaded from



c

'e

X<

OJO

• >

QLJ

AS

UR

UJ

_l1—uUJa:Q

3

3

2

2

1

1

0

. 5 -

. 0 -

. 5 -

. 0 -

. 5 -

. 0 -

. 5 -

- /

' /

s/ /

/-K 1

/ /

'• *Y* '''

/ £f /

/ J\ // * '

y/xy r »0.92 SEE=0.30 / -mln"'

yO.9l x-0.01n = 48

1 1 1 1 10.5 I.O 1.5 2.0 2.5 3.0 3.5 4.0

STEP TEST EST. OF ^ 0 2 MAX U-mirf')

FIGURE I. Directly measured (bicycle protocol) maximal oxygen uptake (VOIM.) versus estimated VOj,,.,from the step test protocol in 48 men and women aged 19-70 years who took part in a community fitnessprogram in Pawtucket, Rhode Island in January-February, 1983. The solid line ( ) is the line of identitywith a 10 per cent adjustment for the expected difference between exercise modalities. The thick broken line(-— - —) is the regression line. The thin broken line ( ) represents ±1 standard error of estimate.

DISCUSSION

Oxygen uptake measured at maximal ex-ercise will vary according to the exercisemodality (16). Kasch et al. (11) showed lessthan 0.5 per cent mean difference betweenpeak oxygen values measured during tread-mill exercises as compared to a step testprocedure. Shephard et al. (17) also showedlittle variation between step testing proce-dures and treadmill procedures. However,bicycle exercise compared to treadmill orstep testing uniformly shows a discrepancy.Verstappen et al. (18) report that the oxy-gen uptake was 14 per cent higher on thetreadmill when compared to bicycle exer-cise. Others who have compared bicycle totreadmill or step testing procedures haveshown differences of 8-15 per cent (7-11).These studies suggest that peak oxygenconsumption measured during bicycle ex-ercise should be between 8-15 per centlower than that measured during step ortreadmill testing protocols. In the present

study, we found a comparable differencebetween the step test estimate of maximaloxygen uptake and the directly measuredpeak oxygen uptake from the bicycle pro-tocol. This difference was approximately 12per cent. However, the difference betweenthe estimated peak oxygen uptake from thesubmaximal bicycle protocol to the directlymeasured peak oxygen uptake (during bi-cycle exercise) should only reflect measure-ment variations. The peak oxygen con-sumption estimated from the submaximalbicycle protocol was within 0.08 liters-min"1 of the directly measured peak oxygenuptake. These differences are similar tothose we have reported earlier (6).

The cross-validation between the sub-maximal bicycle test and the newly devel-oped step test was high with correlationcoefficients ranging from 0.89 to 0.98 forthe age groups. The correlations for malesand females were lower (0.79 and 0.72, re-spectively), but this reflects the truncated



ownloaded from


388 SICONOLFI ET AL.

4.0T

0 . 50.5 1.0 1.5 2.0 2.5 3.5 4 .0

BICYCLE TEST EST. OF V02 MAX (f-mirf')

FIGURE 2. Directly measured (bicycle protocol) maximal oxygen uptake (VO,_.) versus estimatedfrom submaximal bicycle test in 48 men and women aged 19-70 years who took part in a community fitnessprogram in Pawtucket, Rhode Island in January-February, 1983. The solid line ( ) is the line of identity.The thick broken line (-—-—). is the regression line. The thin broken line ( ) represents ±1 standarderror of estimate.

range of maximal oxygen uptake for eachsex. The standard error of estimate rangedfrom 0.14 to 0.21 liters min"1 for the sexand age groups examined in this study. Thehigh correlation between the step and bi-cycle estimates of oxygen uptake and thesmall standard error of estimate indicatethat the step test protocol is valid for theevaluation of peak oxygen uptake. Theportability of the step test platform and thenecessary heart rate equipment makes thetest well suited for testing by mobile sur-veyors visitingly randomly selected homesand respondents for epidemiologic studiesof fitness.

The step test provides estimates of max-imal oxygen uptake in adults of low fitnesslevels. For example, 54 per cent of oursubjects completed the teat at Stage I,which is equivalent to 3-4 METs, while 40per cent of our subjects completed the testat Stage II, equivalent to approximately 7METs. Only 6 per cent of the subjectsneeded to complete Stage III of the protocol

(approximately 10 METs). By using athreshold heart rate of 65 per cent of theage predicted maximum, the step test pro-tocol can be completed by the majority ofthe normal population at relatively low lev-els of exercise. This low threshold heartrate permits estimates of maximal oxygenconsumption for subjects over age 36 yearsat heart rates less than the heart rate usedin other submaximal exercise tests (13) andshould, therefore, be safer to use.

The subjects of this investigation werechosen to represent the general population;the results will probably be comparable,therefore, in similar subject populations.The validity of this protocol is unknownfor patients, children or well-trained ath-letes. Therefore, studies are in the planningstage and will be reported when completed.

The simplicity of this exercise test hasbeen demonstrated by the training of layvolunteers to administer the test at publicservice risk factor screening events (19).Approximately 1,000 step tests were admin-



ownloaded from


STEP TEST ESTIMATES OF MAXIMAL OXYGEN UPTAKE

4 . 0 - r

389

CO 0.50.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0

STEP TEST EST. OF V0 2 MAX (i-mirT1)

FIGURE 3. Estimated maximal oxygen uptake (V0 lmn) from the submaximal bicycle test versus estimatedVet,™, from the step test protocol in 48 men and women aged 19-70 years who took part in a communityfitness program in Pawtucket, Rhode Island in January-February, 1983. The solid line ( ) is the line ofidentity with a 10 per cent adjustment for the expected difference between exercise modalities. The thickbroken line (—— -—) is the regression line. The thin broken -line ( ) represents ±1 standard error ofestimate.

TABLE 4

Means (± standard deviations) of directly measured and step test and bicycle test estimates of maximal oxygenin all subjects, and by age and sex, for 48 men and women aged 19-70 years who took part in a

community fitness program in Pawtucket, Rhode Island in January-February 1983

Directly measured VO,_«>(liters mill"1)

Step test VO,_»(liters • mill"1)

Bike tert V O , _ '(liters • minM)

All subjects

Age (years)19-3435-50

51+

Sex' Male

Female

48

251310

1929

2.00 ± 0.74

2.14 ± 0.762.08 ± 0.851.57 ± 0.35

2.69 ± 0.691.64 ± 0.24

2.24 ± 0.75t

2.44 ± 0.79t2.16 ± 0.68f1.87 ± 0.59t

2.99 ± 0.60t1.74 ± 0.26t

2.06 ± 0.71

2.23 ± 0.782.05 ± 0.681.69 ± 0.68

2.77 ± 0.591.59 ± 0.24

•Estimates of VO,™.t Significantly (p < 0.01) different than directly measured V0,o

istered at these events with no untowardoccurrences, demonstrating the field ap-plicability and safety of this test.

These current results demonstrate thatthe step testing protocol provides valid es-timates of cardio-respiratory fitness over a

wide age range. Inexpensive digital tach-ometers combined with the low level ofexercise needed to complete the step testmake this protocol an attractive method ofassessing cardio-respiratory fitness in thehome and other environments. We believe



ownloaded from


390 SICONOLFI ET AL.

that this test is well suited for epidemio-logic studies of apparently healthy adultsand can also be effectively used in com-munity fitness programs conducted outsideof formal laboratory settings.

REFERENCES

1. Paffenbarger RS Jr, Wing AL, Hyde RT. Physicalactivity as an index of heart attack risk in collegealumni. Am J Epidemiol 1978; 108:161-75.

2. Paffenbarger RS Jr, Hale WE, Brand RJ, et al.Work-energy level, personal characteristics, andfatal heart attack: a birth-cohort effect. Am JEpidemiol 1977; 105:200-13.

3. Thompson JK, Jarvie GJ, Lahey BB, et al. Exer-cise and obesity: etiology, physiology, and inter-vention. Psychol Bull 1982;91:55-79.

4. Wood PD, Haskell WL. The effect of exercise onplasma high density lipoproteins. Lipids 1979;14:417-27.

5. Astrand I. Aerobic work capacity in men andwomen with special reference of age. Acta PhysiolScand [Supplj I960; 169:2-92.

6. Siconolfi SF, Cullinane EM, Carleton RA, et al.Assessing VOi^u in epidemiologic studies: modi-fication of the Astrand-Rhyming test. Med SciSports Exerc 1982; 14:336-8.

7. Hermansen L, Ekblom B, Saltin B. Cardiac outputduring submaximal and maximal treadmill andbicycle exercise. J Appl Physiol 1970;29:82-6.

8. Hermansen L, Saltin B. Oxygen uptake duringmaximal treadmill and bicycle exercise. J ApplPhysiol 1969;26:31-7.

9. Kamon E, Pandolf KB. Maximal aerobic powerduring ladder climbing, uphill running and cy-cling. J Appl Physiol 1972;32:467-73.

10. McKay GA, Bannister EW. A comparison of max-

imum oxygen uptake determination by bicycleergometry at various pedaling frequencies and bytreadmill running at various speeds. Eur J ApplPhysiol 1976;35:191-200.

11. Kasch FW, Phillips WH, Ross WD, et al. A com-parison of maximal oxygen uptake by treadmilland step test procedures. J Appl Physiol1966:21:1387-8.

12. Scholander PS. Analyzer for accurate estimationof respiratory gases in one-half cubic centimetersamples. J Biol Chem 1947; 167:235-50.

13. Astrand PO, Rhyming I. A nomogram for calcu-lation of aerobic capacity (physical fitness) frompulse rate during submaximal work. J Appl Phys-iol 1954;7:218-21.

14. American College of Sports Medicine. Guidelinesfor graded exercise testing and exercise prescrip-tion. 2nd ed. Philadelphia: Lea & Febiger,1980:137-51.

15. Wiener BJ. Statistical principles in experimentaldesign. 2nd ed. New York: McGraw-Hill BookCompany, 1971:chaptere 2 and 3.

16. Astrand PO, Rodahl K. Textbook of work physi-ology. 2nd ed. New York: McGraw-Hill BookCompany, 1977:335-7.

17. Shephard R, Allan C, Benade AJ. Standardizationof submaximal exercise tests. Bull WHO1968;38:766-75.

18. Verstappen FTJ, Huppertz RM, Snoeckx LHEH.Effects of training specificity on maximal tread-mill and bicycle ergometer exercise. Int J SportsMed 1982;3:33-46.

19. Elder JP, Abrams DB, Beaudin P, et'al. Behav-ioral heart health promotion: approaches of thePawtucket Heart Health Program for changes incommunity nutrition and smoking patterns. Sym-posium Presented at the World Congress on Be-havior Therapy, 17th Annual American Associa-tion on Behavior Therapy Convention, Washing-ton, DC, December 8-11, 1983.



ownloaded from


am. j. epidemiol. 1985-siconolfi-382-90

Health & Medicine