METHODS

A Nomogram to Predict Exercise Capacity from a Specific Activity Questionnaire

and Clinical Data Jonathan Myers, PhD, Dat Do, BS, William Herbert, PhD,

Paul Ribisl, PhD, and Victor F. Froelicher, MD

Recent investigations suggested that clinical exercise testing can be optimized by individuaC izing the protocol, depending on the purpose of the test and the subject tested. This requires some knowledge of a patient’s exercise capacity before beghtning the test. The accuracy of a simple physical activity questionnaire and readily available dinical data in predicting subsequent treadmill performance was examined. A brief, setf administered questionnaire (VSAQ) was deveC oped for veterans who were referred to exercise testing for clinical reasons. The VSAQ was designed to determine which specific daily activities were associated with symptoms of cardiovascuiar disease (fatigue, chest pain and shortness of breath). Two hundred twelve consec- utive patients (mean age 62 -c 8 years) referred for maximal exercise testing were studied. Clinical and demographic variables were added to VSAQ responses in a stepwise regression model to determine their ability to predict treadmill per fonnance. Only metabolic equivalents by VSAQ, and age were significant predictors of treadmill performance; these 2 variabies yielded kO.62 (SEE 1.43; p ~O.OOl), and explained 67% of the variance in exercise capacity. 7he regression equation reflecthtg the relation between age, VSAQ and exercise capacity was: achieved met& boiic equivalents = 4.7 + 0.97 (VSAQ) - 0.06 (age). Using this equation, a nomogram was developed. incorporating the VSAQ with the nomogram requires only a few minutes, and yields a reasou ably accurate estimate of a patient’s exercise capacity. Although the present equation is popu iationapecific, a similar approach in different populations may be useful for individualizing pro tocois for clinical exercise testing.

(Am J Cardioi 1994;73:591-!599)

From the Cardiology Division, Palo Alto Veterans Affairs Medical Center and Stanford University, Palo Alto, California. Manuscript received June 30, 1993; revised manuscript received and accepted September I, 1993.

Address for reprints: Jonathan Myers, PhD, Cardiology Division (11 IC), Palo Alto Veterans Affairs Medical Center, 3801 Miranda Av- enue, Palo Alto, California 94304.

E xercise testing is performed frequently in patients with heart disease to evaluate cardiopulmonary function, and assess clinical status, therapeutic ef-

ficacy and patient prognosis. Maximal exercise capacity, commonly expressed in metabolic equivalents (METS), is 1 of the more important measurements used clinical- ly, owing to its impact on prognostic, vocational and financial concerns. These issues are of particular impor- tance in the veteran population, because exercise capac- ity is a major determinant of disability. However, max- imal exercise testing is not feasible in all settings owing to financial and physical limitations, time restraints and increased patient risk. Thus, methods of assessing func- tional status without exercise testing are desirable in some cases.

The New York Heart Association functional classifi- cation is 1 method used to assess the cardiovascular sta- tus of patients with heart disease.’ This system, which classifies patients according to the degree of symptoms associated with daily activities, has several limitations. First, it divides patients into 1 of 4 categories; thus, it is very general. Second, the degree to which the system correlates with direct measures of exercise capacity is unclear. Third, it has been reported that physicians and patients frequently disagree on classifications of activi- ties, which results in patients being misclassified.2-s Goldman et al2 used a specific activity scale consisting of a series of questions concerning daily activities which were designed to classify patients by functional class. Although its performance was superior to those of the New York Heart Association and Canadian Cardiovas- cular Society6 systems, 32% of patients were incorrect- ly classified, and only a modest correlation (r = 0.66) was observed with treadmill time on the Bruce proto- col. Hlatkey et al7 recently developed the Duke Activi- ty Status Index in which responses concerning physical activities were empirically weighted on the basis of difficulty. These investigators reported a correlation co- efficient between the Duke Activity Status Index and maximal oxygen uptake of r = 0.58 when using a self- administered questionnaire; however, the correlation im- proved to r = 0.80, when patients were interviewed personally by a staff member. Other studies, using ac- tivity assessments alone or in combination with clinical or demographic data, have reported associations with ex- ercise capacity ranging from r = 0.29 to O.87.2,7-‘”

The aim of this study was to enhance the prediction of exercise capacity using a qu’ stionnaire and clinical data. However, the specific goal was to use this infor-

PREDICTING EXERCISE CAPACITY 591

TABLE I Clinical Characteristics of Subjects (n = 212; mean k SD)

Age (years) Height (inches) Weight (lb) Body mass index (kg/m2) Resting heart rate (beats/mm) Resting systolic blood pressure (mm Hg) Resting drastolic blood pressure (mm Hg) Hypertension Previous acute myocardial infarction Previous coronary bypass surgery Previous coronary angioplasty Medications at time of test

p blocker Calcium antagonist Digoxin Nitrates

62 2 8 69 2 3

187 -c 37 28 k 5 74-e 15

136 rt 22 852 12 98 (47%) 63 (30%) 41 (19%) 27 (13%)

46 (22%) 84 (40%) 16 (8%) 65 (31%)

mation to establish individualized ramping protocols that derive maximal exercise responses from patients within a standard period of 8 to 12 minutes. Individualized ap- proaches to exercise testing, including targeting a test duration, were recommended recently on both empir- ic’&i6 and experimental 17-21 bases. We have relined an activity questionnaire (VSAQ) that is designed specifi- cally for veterans who were referred to exercise testing for clinical reasons. The observation that only 2 vari- ables, age and METS by the VSAQ, were significant predictors of achieved exercise capacity lent itself to the development of a nomogram.

METHODS Patients Two hundred twelve consecutive patients

(mean age 62 f 8 years) referred for exercise testing for clinical reasons between February and September 1992 were included in the study. Clinical characteristics of the 207 men and 5 women are presented in Table I. Only patients achieving a symptom-limited effort (fatigue, leg fatigue, chest discomfort and shortness of breath) were included in the study. Patients were excluded if their ex- ercise test was sign-limited or submaximal (postrny- ocardial infarction before discharge), or terminated by the supervising physician present for other reasons. The normal therapeutic medication regimen was maintained. Forty-six patients were receiving B blockers, 84 calci- um antagonists and 16 digoxin. Sixty-three patients had history of myocardial infarction, and 41 had undergone bypass surgery.

Exercise testing: All patients underwent maximal exercise testing using an individualized ramp treadmill protocol.17,‘8 This test individualizes both warm-up and peak walking speeds (based on a given patient’s height, fitness and familiarity with treadmill walking), and ramp rate (rate of change in speed and grade) to yield a test duration of approximately 10 minutes. Using software developed at our institution, a microcomputer automati- cally increased work load after an individualized walk- ing speed and predicted value for maximal exercise ca- pacity were entered. Standardized equations were used to determine the predicted oxygen uptake (in METS) based on treadmill speed and grade.22 The estimated value obtained is called “predicted” exercise capacity,

whereas the value obtained from the linal speed and grade on the treadmill is called “achieved’ exercise ca- pacity. Blood pressure was recorded in alternate minutes throughout the test, whereas a 1Zlead electrocardiogram was recorded each minute. Subjective levels of exertion were quantified using the Borg 6-to-20 scale.23 Standard clinical criteria for terminating the test were followed,i5 but no heart rate or time limit was imposed, and a max- imal effort was encouraged.

Questionnaire: Before testing, the VSAQ was given to each patient (Table II). The VSAQ consists of a list of activities presented in a progressive order according to METS. Patients were instructed to determine which activities may typically cause fatigue, shortness of breath, chest discomfort or claudication pain during daily activities. The MET values associated with each activity were derived from various sources,15,22,24 and are in general agreement with the recently published Compendium of Physical Activities.25 Current activity status of each patient was also estimated on a l-to-4 scale as: 1 = very sedentary or bed rest; 2 = sedentary; 3 = moderately active; and 4 = very active.

Statistii Simple univariate regression was per- formed, with peak METS estimated from treadmill speed and grade as the dependent variable (i.e., achieved ex- ercise capacity), and METS estimated by VSAQ as the independent variable (i.e., predicted exercise capacity). A stepwise multiple regression procedure was then per- formed with achieved exercise capacity as the dependent variable, and age, METS by VSAQ, current activity sta- tus, body mass index, smoking history (yes/no and pack- years), history of myocardial infarction, history of con- gestive heart failure, use of B blockers, and resting heart rate and blood pressure as independent variables. Statis- tical Graphics Corporation Software (Rockville, Mary- land) was used for the regression procedures and de- scriptive statistics.

RESULTS Exercise responses: Exercise test results are listed

in Table III. The mean maximal perceived exertion was 18.3 f 2.2, suggesting a maximal effort was achieved by most patients. The mean maximal heart rate of 130 f 23 beats/min was lower than that expected for age, reflect- ing the fact that many patients were limited by symp- toms due to cardiovascular disease and by B-blocker ef- fects. In questioning patients immediately after the test, 154 (73%) were limited mainly by fatigue or leg fatigue, 30 (14%) by chest pain, 13 (6%) by shortness of breath and 11 (5%) by leg pain that appeared to be due to pe- ripheral vascular disease, and 4 (2%) stopped owing to dizziness.

Achieved versus predicted exercise capacity: The mean maximal MET value achieved for the group was 7.1 f 3.0, whereas that predicted from the VSAQ was 6.3 + 2.3. The correlation coefficient between the 2 values was 0.79 (SEE 1.42; p <O.OOl). Using stepwise multiple regression to predict exercise capacity, the first variable to enter the equation was peak METS by VSAQ (R2 = 0.63; p <O.OOl). Age was the only other variable that added significantly to predicting exercise capacity (adding 4% to the explanation of variance in peak

592 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 73 MARCH 15, 1994

r TABLE II The Veterans Specific Activity Questionnaire

Draw One Line Below the Activities You Are Able to Do Routinely witmmal or No Symptoms, Such as

METS Shortness of Breath, Chest Discomfort, Fatigue

1 - Eating, getting dressed, working at a desk. 2 - Taking a shower.

- Walking down eight steps. 3 - Walking slowly on a flat surface for one or two blocks.

- A moderate amount of work around the house, like vacuuming, sweeping the floors or carrying groceries.

4 - Light yard work, i.e., raking leaves, weeding or pushing a power mower. - Painting or light carpentry.

5 - Walking briskly, i.e., four miles in one hour. - Social dancing, washing the car.

6 - Play nine holes of golf carrying your own clubs. Heavy carpentry, mow lawn with push mower.

7 - Perform heavy outdoor work, i.e., diggrng, spading soil, etc. - Play tennis (singles), carry 60 pounds.

8 - Move heavy furniture. - Jog slowly, climb stairs quickly, carry 20 pounds upstairs.

9 - Bicycling at a moderate pace, sawing wood, jumping rope (slowly). 10 - Brisk swimming, bicycle up a hill, walking briskly uphill, jog six miles per hour. 11 - Cross country ski.

- Play basketball (full court). 12 - Running briskly, continuously (level ground, eight minutes per mile). 13 - Any competitive activity, including those which involve intermittent sprinting.

- Running competitively, rowing, backpacking.

METS = metabolic equivalents.

achieved METS), yielding R = 0.82, and R2 = 0.67 (SEE 1.43, p <O.OOl). Current activity status, body mass index, history of myocardial infarction, history of congestive heart failure, use of B blockers, history or pack-years of smoking, resting heart rate and resting systolic or dia- stolic blood pressure did not add significantly to the pre- diction of exercise capacity.

The final regression equation reflecting the relation between age, VSAQ and exercise capacity was: METS = 4.7 + 0.97 (VSAQ) - 0.06 (age). Figure 1 shows the relation between METS predicted from VSAQ and age versus actual METS achieved. A nomogram to predict exercise capacity using VSAQ and age is shown in Figure 2. The data are presented also in tabular form by age group in Figure 3.

DISCUSSION Since the development of the New York Heart As-

sociation and Canadian Cardiovascular Society classifi- cation systems for patients with heart disease, several attempts have been made to categorize patients by func- tional class more precisely without exercise testing. These methods have been useful, because it is not al- ways feasible for patients to perform maximal exercise. These estimations have been shown to serially appraise clinical and symptomatic status in a reproducible, al- though general manner. 2,26 In addition to questionnaires on current or past activity status, or both, certain clini- cal variables including symptoms, risk factors and de- mographic data have been used also to predict exercise capacity. However, all these approaches have had limi- tations. Limitations in precision and consistency in the widely used New York Heart Association and Canadian Cardiovascular Society classification systems are well

TABLE Ill Responses to Exercise Testing (mean 2 SD)

Maximal heart rate (beatslmin) 130 + 23 Maximal perceived exertion 18.3 2 2.2 Exercise capacity achieved (METS) 7.1 f 3.0 Exercise capacity predicted by VSAQ 6.3 -t 2.3 Exercise capacity predicted by VSAQ and age 7.05 f 2.4 Maximal treadmill time (min) 10 It 4

METS = metabolic equwalents: VSAQ = veterans specific activity questionnaire.

documented.2-ss26 In more specific and detailed efforts since the development of these scales, correlations with exercise capacity have varied widely (r = 0.29 to 0.87).2,7-13

In this study, with the exception of age, no clinical or demographic variable added significantly to the pre- diction of exercise capacity. In most previous studies, age, body composition and gender were important vari- ables in predicting exercise capacity. Age alone has cor- related with exercise capacity in the range of r = -0.3 to 4X5.9~27-29 Correlations between exercise capacity and body composition, expressed as percent body fat, body mass index or weight, have ranged from r =-0.37 to -0.66.9~10~30 The multiple R of 0.82 derived from age and VSAQ in the present study is nearly identical to the widely cited effort to predict exercise capacity reported by Bruce et a128 (multiple R = 0.81), using a combina- tion of 6 variables (sex, age, activity status, weight, smoking history and height). Other studies, using ques- tionnaires alone or in combination with the aforemen- tioned demographic and clinical variables, have observed widely ranging associations with exercise capacity and have reported SEES generally ranging in the order of 1 to 2 mTs.7,9,‘0,30

PREDICTING EXERCISE CAPACITY 593

An “acceptable” regression equation and SE depend for an “optimal” protocol that immediately followed. Al- on the purpose of the equation and on the precision though several advantages to individualized tests were needed for its specific use. The present objectives dif- described recently (i.e., individualizing work rates, tar- fered from those in previous studies in that the goat was geting the test duration, and ramp protocols),‘421 1 im- to estimate exercise capacity to individualize work rates pediment to the use of this approach is that a reason-

12-

6-

PREDICTED METS = A74 + 0.97lVSAQl- O.WAGE’) 0

t = OAt, SEE o 1.4% p<0.007 Q&..‘- 0 ./-

Ol I I I I I I I 0 2 4 6 6 10 12 14 1

ACHIEVED METS FIGURE 1. Rektiae between metabok equivalents (METS) prdkted by multiple regresdon and those achieved. Wter lines lmpment9o%~li~v5AQ=spscmcactivlty~-refer~

AGE (wars)

METS BY QUESTIONNAIRE

13 -

:: -- 10 -. 9 --

78 -- 6 --

: -- 3 -- 2 -- l-

15 +

14 --

13 --

12 --

11 . .

10 -*

9 --

6 --

7 --

6 . .

5 --

4 --

3 --

2 -- II

PREDICTED EXERCISE CAPACITY

W-W

FlRuRE2.Nomogmntopmdiiexerebe cz3pety. using alight edge, exercke u6mcltykPmdietedbasedonageaml -~spedficactivityc- mire. MEls = -k aquivden&

594 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 73 MARCH 15,1994

able approximation of exercise capacity before the test is needed.

The results shown in Figure 1 reflect those using the VSAQ in consecutive patients referred for exercise test- ing over an 8-month period, excluding those in whom questions pertaining to limiting symptoms would not apply (i.e., those whose tests were terminated due to ar- rhythmias, an abnormal blood pressure response, and so forth). The association between VSAQ, age and exercise capacity in this study [multiple R = 0.82; SEE 1.43) sug- gests that for a targeted test duration of 10 minutes, 90% of tests would last between approximately 8 and 12 min- utes. This is within the recommended test duration range outlined in the recent American Heart Association Ex- ercise Standards’” and that suggested by experimental data.t7*t9 We developed a nomogram to facilitate appli- cation of these data (Figure 2). Incorporating the VSAQ with the nomogram requires only a few minutes, yields a reasonably accurate prediction of achieved METS and obviates the need for more cumbersome calculations.

In the absence of gas exchange techniques, all esti- mates of exercise capacity should be viewed with cau- tion. Errors in predicting maximal oxygen uptake from treadmill time or work load are well document- ed.16%17,20,21 Although the individualized ramp approach used in this study appears to improve considerably the accuracy with which oxygen uptake is estimated from treadmill work rate, a significant degree of variability is nevertheless observed.t7 In addition to the shortcomings associated with treadmill time alone as a measure of ex- ercise capacity, the accuracy of a questionnaire such as the VSAQ greatly depends on individual differences in

perceptions of, familiarity with and tolerance to similar activities. In completing the VSAQ, patients generally underestimated their achieved exercise capacity by ap- proximately 10 and 30% in the oldest and youngest age groups, respectively (Figure 3). The nomogram devel- oped from the present data was not intended to be viewed as a substitute for exercise testing or functional classification of patients per se; its purpose was to indi- vidualize ramp rates for exercise testing.

Any effort to predict exercise capacity, by assessing physical activity or using clinical or demographic data, or both, should be considered population-specific. The present population differed from those of previous stud- ies in that only patients referred to exercise testing for clinical reasons were included. Thus, the VSAQ was tar- geted toward a patient population, many of whom were limited by symptoms, medications and other factors re- lated to cardiovascular disease. In addition to population factors, differences in these results and those obtained in previous studies may be due to differences in definitions of activity, exercise protocols, the method for determi- nation of exercise capacity (measured vs estimated max- imal oxygen uptake), and the variables included (i.e., some studies did not consider women, diseased subjects or medications). Finally, as with any clinical tool, the VSAQ requires prospective validation.

1. Criteria Committee of the New York Hean Association (Kossman CE, Chair- man). Diseases of the Heart and Blood Vessels: Nomenclature and Criteria for Di- agnosis. 6th ed. Boston: Little, Brown, 1964:112. 2. Goldman L, Hashimoto B, Cook EF, Loscalzo A. Comparative reproducibility and validity of systems for assessing cardiovascular functional class: advantages of a new specific activity scale. Cir-cula~ion 198 I;64: 1227-1234.

METS BY QUESTIONNAIRE

FIWRE 3. Pradii metabolic equivalents (METS) based on age and questionnaire by muRiple regwtssh. Shahd values relkct mean achieved metabolii equivalents for each age group.

PREDICTING EXERCISE CAPACITY 595

3. Goldman L, Cook EF, Mitchell N, Flatley M, Sherman H, Cohn PF. Pitfalls in the serial assessment of cardiac functional status. J Chronic Dis 1982;35:763-771. 4. Seizer A, Cohn K. Functional classification of cardiac disease: a critique. Am .I Cardiol 1972:30:30&308. 5. Dunselman PHJM, Kuntze EE, van B~ggen A, Beekhuis H, Piers B, Scaf AHJ, Wesseling H, Lie KI. Value of New York Heart Association classification, ra- dionuclide venniculogsaphy, and cardiopulmowy exercise tests for selection of pa- tients for congestive heart failure studies. Am Heart J 1988: 116: 1475-148 1. 6. Campeau L. Grading of angina pectoris (letter). Circulation 1976;54:522-523. 7. Hlatkey MJA, Boineau RE, Higginbatham MB, Lee KL, Mark DB, Califf RM, Cobb FR, F’ryor DB. A brief self-administered questionnaire to determine functional capacity (the Duke Activity Status Index). Am J Cardiol 1989:64:65 l-654. 8. Lee TH, Shammash JB, Ribeiro JP, Hartley LH, Sherwood J, Goldman L. Es- timation of maximum oxygen uptake from clinical data: performance of the Spe- cific Activity Scale. Am Heat? J 1988: 115:203-204. 9. Milesis CA. Prediction of treadmill performance from clinical characteristics in healthy persons. J Cardiopulmonary Rehahil 198X7:365-373. 10. Jackson AS, Blair SN, Mahar MT, Wier LT. Ross RM. Stuteville JE. Predic- tion of functional aerobic capacity without exercise testing. Med Sci Sporrs &xerc 1990;22:863-870. ll. Taylor HL, Jacobs DR Jr, Schucker B, Knudsen J, Leon AS, Debacker G. A questionnaire for the assessment of leisure time physical activities. J Chronic Dis 1978;31:741-755. 12. Siconolfi SF, Lasater TM, Snow RCK. Self-reported physical activity com- pared with maximal oxygen uptake. Am J Epidemiol 1985;122:101-105. 13. Kohl HW, Blair SN, Paffenbarger RS Jr, Macera CA, Kronenfeld JJ. A mail survey of physical activity habits as related to measured physical fitness. Am J Epi- demiol 1988;127:1228-1239. 14. Webster MWI, Sharpe DH. Exercise testing in angina pectoris: the importance of protocol design in clinical trials. Am Heart J 1989;l l7:505-508. 15. Fletcher GF, Froelicher VF, Hartley LH, Haskell W, Pollock M. Exercise stan dards: a statement for health professionals from the American Heart Association. Circulation 1990;82:228&2322. 16. Myers I. Froelicher VF. Optimizing the exercise test for pharmacologic in- vestigations. Cvcularion 1990;82:1839-1846. 17. Myers J, Buchanan N, Walsh D, Kraemer M, McAuley P, Hamilton-Wessler

M, Froelicher VF. Comparison of the ramp versus standard exercise protocols. J Am Co!/ Cardio/ 1991;17:133~1342. 18. Myers 3, Buchanan N, Smith D, Neutel J, Bowes E, Walsh D, Fmelicher VF. Individualized ramp treadmill: observations on a new protocol. Chesr 1992;lOl: 236S241S. 19. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K, Whipp BJ. Optimizing the exercise test for cardiopulmonary assessment. J Appl Physiol 1983; 55:1558-1564. 20. Sullivan M, McKiian MD. Errors in predicting functional capacity for post- myocardial infarction patients using a modified Bruce protocol. Am Heart J 1984, 107:48&%91. 21. Haskell W, Savin W, Oldridge N, DeBusk R. Factors influencing estimated oxygen uptake during exercise testing soon after myocardial infarction. Am .I Car- did 1982;50:299-304. 22. American College of Spolts Medicine. Guidelines for Exercise Testing and Pre- scription. Philadelphia: Lea & Febiger, 1991. 23. Borg GAV. Psychophysical bases of perceived exertion. Med Sci Sporrs Exert 1982;14:377-381. 24. Pallock M, Wilmore J. Exercise in Health and Disease: Evaluation and Pre- scription for Prevention and Rehabilitation. Philadelphia: W.B. Saunders, 1990. 25. Aimworth BE, Haskell WL, Leon AS, Jacobs DR, Montoye HJ, Sallis JF, Paf- fenbarger RS. Compendium of Physical Activities: classification of energy costs of human physical activities. Med Sci Sports Exert 1993:25:71-80. 26. Cox J, Nay101 D. The Canadian Cardiovascular Society grading scale for angi- na pectoris: is it time for refinements? Ann Intern Med 1992;117:677-683. 27. Morris CK, Myers J, Fmelicher VF, Kawaguchi T, Ueshima K, Hi&g A. Nomogram for aerobic exercise capacity in men using METS and age. J Am Co/l Cardiol 1993;22: 175-182. 28. Bruce RA, Kusumi F, Hosmer D. Maximal oxygen intake and nomographic assessment of functional aerobic impairment in cardiovascular disease. Am Heart J 1973;85:54&562. 29. Froelicher VF, Allen M, Lancaster MC. Maximal treadmill testing of normal USAF aircrewmen. Aerospace Med 1974;45:310-315. 30. Kline GM, Porcari JP, Hintermeister R, Freedson PS, Ward A, McCannon RF, Ross J, Rippe J. Estimation of VOz max from a one-mile track walk, gender, age, and body weight. Med M Sporfs Exsc 198X19:253-259.

596 THE AMERICAN JOURNAL OF CARDIOLOGY VOLUME 73 MARCH 151994