Compared with maximal exercise testing, submaximal exercise testing appears to have greater applicability to physical therapists in their role as clinical exercise specialists. This review contrasts maximal and submaximal exercise testing. Two major categories of submaximal tests (ie, predictive and performance tests) and their relative merits are described. Predictive tests are submaximal tests that are used to predict maximal aerobic capacity. Performance tests involve measuring the responses to standardized physical activities that are typically encountered in everyday life. To maximize the validity and reliability of data obtained from submaximal tests, physical therapists are cautioned to apply the tests selectively based on their indications; to adhere to methods, including the requisite number of practice sessions; and to use measurements such as heart rate, blood pressure, exertion, and pain to evaluate test performance and to safely monitor patients.

Although maximal exercise testing is considered the gold standard for assessing maximal aerobic capacity, the role of such testing is limited in people whose performance may be limited because of pain or fatigue rather than exertion and in cases where maximal exercise testing is contraindicated. Submaximal exercise testing overcomes many of the limitations of maximal exercise testing, and it is the method of choice for the majority of individuals seen by physical therapists in that these individuals are likely to be limited physically by pain and fatigue or have abnormal gait or impaired balance. This article contrasts maximal and submaximal exercise testing and describes the clinical application of submaximal testing. The strengths and limitations of both predictive and performance submaximal tests and the means of maximizing validity and reliability of data are presented. Predictive tests are submaximal tests that are used to predict maximal aerobic capacity. Typically, heart rate (HR) or oxygen consumption (V̇o2) at 2 or more workloads is measured.1,2 A predicted V̇o2 value is obtained by extrapolating the relationship between HR and V̇o2 to age-predicted maximal heart rate (HRmax). Performance tests involve measuring the responses to standardized physical activities that are typically encountered in everyday life. Finally, we discuss the use of submaximal exercise testing in clinical decision making and the implications for professional education and research.

Maximal Versus Submaximal Exercise Tests

Maximal exercise tests either measure or predict maximum oxygen consumption (V̇o2max) and have been accepted as the basis for determining fitness.37 They have served as a standard against which to compare other measures.8 Maximum oxygen consumption is dependent on the ability of the oxygen transport system to deliver blood and the ability of cells to take up and utilize oxygen in energy production.9 Theoretically, a maximal test is defined by the plateau of V̇o2 with further increases in workload.10,11 Other indexes used to assess maximal effort include obtaining HRmax within 15 beats per minute (bpm) of age-predicted HRmax (ie, 220–age) and a respiratory exchange ratio >1.10 (ratio of metabolic gas exchange calculated by carbon dioxide production divided by V̇o2).12 Maximum oxygen consumption is typically expressed relative to body weight (ie, mL·kg−1·min−1),13 which enables individuals of different body masses to be compared. When a maximal test is performed but the criteria for V̇o2max are not met, the maximal V̇o2 achieved is termed a “V̇o2peak.”14 Few individuals reach a true V̇o2max, and V̇o2peak values are often incorrectly reported as maximal values.14 The intraindividual day-to-day variation in measuring V̇o2max is between 4% to 6% in individuals with no known cardiopulmonary pathology or impairment.15,16 In people with various diagnoses, such as those with chronic obstructive pulmonary disease (COPD), this variation is between 6% and 10%.17

There are several limitations to assessing maximal performance with a V̇o2max test. Unless an individual is able to attain a V̇o2max without fatiguing first or being limited by musculoskeletal impairments or other problems, the results of the test are invalid. In addition, higher levels of motivation are required by the individual, and maximal tests require additional monitoring equipment (eg, electrocardiograph machine) and trained staff and are labor intensive.1,13,18

In comparison with maximal tests, submaximal exercise tests and their applications have been less well developed, which we find surprising given the large number of patient types and individuals who should be able to benefit from nonmaximal exercise tests. For the purpose of this review, submaximal tests are classified as either predictive tests or performance tests.

Clinical Application of Submaximal Exercise Testing


Exercise constitutes a physiologic stress that may pose a greater risk to people with various diagnoses than to people without pathology or impairment. The space for testing must be sufficient to minimize injury should the patient fall or have an arrest. All physical therapists should have current certification in cardiopulmonary resuscitation. Emergency procedures and basic equipment need to be in place to ensure that the individual has immediate care until paramedical or medical assistance arrives. There are other critical needs for exercise testing. Basic emergency supplies, including a sugar source for people with diabetes, should be on hand. A portable oxygen source and suction device should be accessible. People who are stable and who have a history of angina should have their antianginal medication, and the physical therapist should have access to this medication. Monitoring equipment should be maintained and regularly calibrated.

Indications for testing and any contraindications for testing should be determined before testing.19 In the presence of relative contraindications, the person may require additional monitoring (eg, 12-lead electrocardiography) or be cleared for such testing by an appropriate medical practitioner. A high proportion of people over the age of 65 years without known cardiac disease have a high incidence of cardiac dysrhythmias,20,21 which may necessitate greater attention to monitoring cardiac status during exercise. Individuals requiring additional monitoring or considered to be at hemodynamic risk should be tested in a setting with medical personnel present.


Maximal exercise testing has a role in the assessment of maximal aerobic capacity or functional work capacity. Because people are frequently limited by cardiopulmo-nary, musculoskeletal, and neuromuscular impairments and complaints such as exertion, dyspnea, fatigue, weakness, and pain during their activities of daily living, maximal testing is often contraindicated or of limited value. In people without cardiopulmonary or musculo-skeletal impairments, the reserve capacity of the cardiopulmonary and musculoskeletal systems is thought to be barely tapped during daily activities.22 In people with pathology, this reserve can be greatly reduced, and a greater than usual proportion of a person's maximal capacity may be needed to perform routine activities.23

Exercise constitutes a major physiological stress that can lead to untoward responses in patients as well as in individuals without known pathology. In addition, such testing is resource intensive and thus should be applied and performed judiciously. The purposes of maximal tests include determination of V̇o2max and use as diagnostic or treatment outcome tools. Submaximal exercise tests can be used to predict V̇o2max, to make diagnoses and assess functional limitations, to assess the outcome of interventions such as exercise programs, to measure the effects of pharmacological agents, and to examine the effect of recovery strategies on exercise performance.13,19,2325

Guidelines for Test Selection

There are numerous submaximal tests from which to choose. These tests have been developed to meet the needs of people with various functional limitations and disabilities and the needs of older adults. In our opinion, however, inappropriate selection may lead to either understressing or overstressing the individual. Such understressing or overstressing of the person, in our view, can lead to invalid conclusions because of ceiling or floor effects, and the testing may be hazardous. The goal of testing should be to produce a sufficient level of exercise stress without physiologic or biomechanical strain. Factors that we believe should be considered in selecting the appropriate test include the person's primary and secondary pathologies and how these pathologies physically affect the person's daily life. Other factors include cognitive status, age, weight, nutritional status, mobility, use of walking aids or orthotic or prosthetic devices, independence, work situation, home situation, and the person's needs and wants. People who may be medically unstable and at risk for an arrest may need to be tested in the presence of a cardiologist or pulmonary specialist or by a physical therapist in a specialized setting where emergency services are on hand. The population for which a given test was developed, the degree of validity and reliability of measure-ments obtained with each test, and test sensitivity also should be considered (Appendix). Reports in the literature on the common submaximal tests described in this article vary with respect to the adequacy of establishing validity, reliability, and sensitivity; thus, test interpretation may be limited. These limitations should be considered in the selection of each test. Physical therapists should determine what information will be added by performing an exercise test and how that information will alter clinical decision making.

Pretest Workup

A detailed medical and surgical history is needed to identify the indications for an exercise test and to alert the physical therapist about any underlying conditions (eg, cardiovascular, pulmonary, musculoskeletal, or neurological dysfunction or the presence of diabetes, hyper-tension or heart block requiring a pacemaker, anemia, thyroid dysfunction, obesity, deformity, vertigo, or impaired cognitive function). The therapist should be aware of medications (indications, response, and side effects) that can influence the test procedures and the response to the exercise. Laboratory tests and investigations that may be relevant include electrocardiograms, echocardiograms, pulmonary function tests, investigations of peripheral vascular function, blood chemistry tests, bone density measures, radiographs, scans, thyroid function tests, glucose tolerance tests, autonomic nervous system function tests, sleep studies, nutritional assessment, and tests for level of hydration.

The effect of each medication on exercise response and the medication's side effects should be known to the person administering the test. Beta blockers, for example, attenuate normal HR and blood pressure (BP) responses to exercise and contribute to fatigue in some people. The purpose of the test must be clear so that the person can be appropriately premedicated (eg, with antidysrhythmic drugs, inotropic drugs, anticoagulants, antithrombolytics, bronchodilators, vasodilators, diuretics, and analgesics). For example, medications such as bronchodilators and analgesics have peak effect times; thus, it is important to ensure that these medications are at peak effect during the test and that this effect is replicated on subsequent tests.

People with a history of angina should be screened carefully. The objective of submaximal testing is to test the individual below the work rate that induces angina. The person's anginal history will divulge the range of activities and the activities that are not associated with symptoms. Labile angina, angina at rest, and frequent premature ventricular contractions (PVCs) at rest are, in our opinion, absolute contraindications to exercise testing in the absence of a cardiologist unless in a specialized setting where physical therapists are qualified to perform such testing. Premature ventricular contractions can be detected reliably only with electrocardiograms and not by palpation or verbal report. A detailed anginal history, including what triggers episodes of angina and the frequency of self-medication with anti-anginal medication and its effect, should be recorded. Any history of chest discomfort or pain from any cause should be noted by the tester. We also believe that any medication should be checked for its expiration date and should be available in the event it is needed during or after the test. A person with a history of angina and for whom antianginal medication is prescribed, in our opinion, should be considered at risk even if the medication has not been required for a prolonged period. We advise that a risk assessment should be conducted for every individual, regardless of whether a maximal exercise test or a submaximal exercise test is being performed. This assessment will help to determine which test is appropriate, predict an adverse response to testing, identify the level of monitoring needed, and whether there are any contraindications to submaximal exercise testing.

Standardization of Procedures

A primary concern about submaximal exercise testing is the lack of standardization of the procedures. We believe that general procedures should include informing the person about the type and purpose of the test and instructing the person to avoid any strenuous activity for 24 hours prior to testing and to avoid a heavy meal, caffeine, or nicotine within 2 to 3 hours of testing.19 Medications taken prior to testing should be noted by the examiner, and, if appropriate, their use should be consistent from one test to the next. The individual should become familiar with the equipment and test procedures to minimize anxiety. Many tests require one or more practice sessions. If time and resources do not permit these practice sessions, we argue that the test should not be performed because the results, in our view, will not be valid. Appropriate rest periods, in our opinion, need to be scheduled between practices and between the last practice and the actual test. We have previously shown that performance of a submaximal treadmill walking test requires at least one practice session, even in young subjects without functional impairments, in order for the measurements to be valid.26 For some individuals, more practice sessions are justified. The number of practice sessions required to make the results valid, in our opinion, is dependent on the test and on the experience and functional capacity of the person being tested. Verbal encouragement in sub-maximal testing should be standardized to ensure that this does not affect the person's performance.27 Failure to calibrate both exercise devices and monitoring equipment can lead to erroneous results.


Basic measures of exercise responses include HR, BP, respiratory rate (RR), rating of perceived exertion (RPE), and breathlessness. Depending on the person's history and other variables, the examiner may find other measures to be useful (eg, a 3-lead electrocardiogram, arterial saturation assessed using a pulse oximeter, cadence, ratings of fatigue and discomfort or pain). Because tests are performed over a wide area or circuit, monitoring equipment should be portable. Repeated measurements of each variable of interest, in our opinion, should be taken prior to the exercise test to ensure a stable baseline, at various points during the test (depending on the type of test), and during the cool-down period, if applicable, and these measurements should be repeated during recovery to ensure that the measures have returned to baseline levels. As a precaution, we recommend that the person should not leave the testing area until all measures have returned to within 10% of resting values.19 Based on the history of the person being tested, additional monitoring may be indicated to maximize the safety of the test.

Because the measurement of BP is an important part of exercise testing, the validity of these measurements should be maximized with an appropriately sized cuff, its position on the midshaft of the humerus, its tightness, the cuff deflation rate, and the position of the stethoscope over the brachial artery as it courses over the antecubital fossa.28 Skill in recording BP is essential, given that many people have undiagnosed or poorly controlled hypertension.

Measures of exertion, breathlessness, fatigue, discomfort or pain, and well being in response to physical activity or exercise are important exercise responses reported by the person being examined. Many people, particularly older people, more readily and reliably monitor and act on their complaints, rather than using measures such as HR to guide their activities or exercise intensity.


The interpretation of the submaximal exercise test results is based primarily on the type of test conducted, its indications (eg, assessment, diagnostic, exercise prescription), specified outcomes, and, in some instances, norms (Appendix). Submaximal exercise tests can be used to predict aerobic capacity or to assess the ability to perform a standardized exercise or task. In addition, measurements taken before, during (where applicable), and after the test can yield valuable information regarding the person's exercise response. These values can be compared across subsequent tests. They can alert the physical therapist to undue pretest arousal (a measure of the adequacy of the pretest standardization), exaggerated exercise responses, and delayed recovery, which are consistent with deconditioning or pathology, or both. Comparison of the responses with pretest and posttest measurements is particularly useful for assessing the effect of an intervention such as an exercise program. In this case, a reduction in submaximal exercise responses such as HR, RR, and BP can be consistent with improved aerobic conditioning or movement economy, or both. Movement economy refers to the efficient use of energy during movement (ie, not excessive V̇o2 for a given activity or work rate).

Predictive Submaximal Exercise Tests

Modified Bruce Treadmill Test


The Bruce Treadmill Test5 is a maximal test that was designed to diagnose coronary heart disease. Some preliminary stages have been added to the original test, which has given rise to the use of the Modified Bruce Treadmill Test in people with other conditions.29,30 Compared with the original test, which starts at 1.7 mph at a grade of 10%, the modified test has a zero stage (1.7 mph at 0% grade) and a one-half-stage (1.7 mph at 5% grade) (Tab. 1). Predictive equations for estimating V̇o2max have been developed and can be used with the original and modified tests. Bruce et al5 developed the first predictive equations, which are population-specific for active and sedentary adults with and without cardiac conditions. Individuals must be correctly classified to determine which equation is appropriate. Foster et al30 later developed a regression equation applicable to all men based on a sample of 230 men of various ages with a variety of clinical conditions (symptomatic angina, n=14; postmyocardial revascularization surgery, n=36; outpatient cardiac rehabilitation surgery, n=63; preventative medicine program, n=90, and athletes, n=27) and activity levels. The details of the Modified Bruce Treadmill Test are provided in the Appendix.

Table 1.

Modified Bruce Treadmill Test: Protocola

Reliability and validity.

Bruce et al5 reported Pearson product moment correlation coefficients (r) between predicted V̇o2max and measured V̇o2max of .94 for without cardiac conditions (n=292), .93 for women without cardiac conditions (n=509), and .87 for men with cardiac disease (n=153). Foster et al30 compared predicted V̇o2max and measured V̇o2max for the general equation and the population-specific equations introduced by Bruce et al.5 The average predicted error was −0.6 mL·kg−1·min−1 for the general equation versus −2.0 mL·kg−1·min−1 for the population-specific equations. The correlation between measured V̇o2max and predicted V̇o2max for the general equation was high (r=.96), with a multiple correlation coefficient (R) of .98 and a standard error of the estimate (SEE) of 3.5 mL·kg−1·min−1.30

Strengths and weaknesses.

The Bruce Treadmill Test and the Modified Bruce Treadmill Test are widely used, especially for the diagnosis of coronary heart disease, and, as a result, normative data are available. Compared with the original protocol, which starts with a large workload, the modified protocol has a more gradual initial workload. Thus, we contend that the modified protocol is more applicable for individuals with low functional capacity. The large increases in workload in the original protocol, however, allowed the test to be completed within 6 to 9 minutes.23

Single-Stage Submaximal Treadmill Walking Test


Ebbeling et al12 developed the Single-Stage Submaximal Treadmill Walking Test (SSTWT), which can be used by individuals of various ages and fitness levels. The test was developed on a sample of 139 volunteers with no health problems (67 men and 72 women) aged 20 to 59 years. The subjects were randomly assigned to either an estimation group (n=117) or a cross-validation group (n=22). Subjects walked on a treadmill at a constant speed, ranging from 2.0 to 4.5 mph at grades of 0%, 5%, and 10%, with each stage lasting 4 minutes. A maximal test was then performed. The regression equation used to estimate V̇o2max was based on data obtained from the estimation group from the 4-minute stage at a grade of 5%. The details of the SSTWT are given in the Appendix.

Reliability and validity.

The SSTWT was validated by correlating the estimated V̇o2max and the measured V̇o2max in the cross-validation group. A correlation (r) of .96 was obtained, with a multiple correlation (R) of .86 (SEE=4.85 mL·kg−1·min−1).

Strengths and weaknesses.

This test is suitable for testing people with various diagnoses in clinical and research settings. It consists of only a warm-up session and a single stage on the treadmill. This test, in our view, is useful for assessing people who are prone to fatigue.

Further research is needed to validate this test in people with various diagnoses, in individuals over 60 years of age, and in both unfit and highly trained individuals. Further studies are needed to establish its sensitivity to detect change. Finally, because this test is based on HR, factors that affect HR must be controlled; otherwise, the test results will be invalidated.

Astrand and Ryhming Cycle Ergometer Test


The Astrand and Ryhming (A-R) Cycle Ergometer Test, which is used to predict V̇o2max by use of a cycle ergometer, is based on the linear relationship between V̇o2 and HR.31 Astrand and Ryhming31 noted that, in subjects aged 18 to 30 years, the men had an average HR of 128 bpm at 50% of V̇o2max and an average HR of 154 bpm at 70% of V̇o2max, and the women had an average HR of 138 bpm at 50% of V̇o2max and an average HR of 164 bpm at 70% of V̇o2max. A nomogram was developed by Astrand and Ryhming31 to estimate V̇o2max (Fig. 1), and later an age-correction factor was incorporated to account for the decrease in HRmax with age (Tab. 2).32 Modification of the A-R nomogram were proposed by Legge and Banister33 and by Hartung and colleagues9,34 to improve the accuracy of the equation. A revision to the A-R nomogram was also proposed by Siconolfi et al.35 The details of the A-R Cycle Ergometer Test are presented in the Appendix.

Figure 1.

The Astrand and Ryhming nomogram. Estimated maximum oxygen consumption (V̇o2max) can be determined by reading horizontally from the body weight scale (step test) or workload scale (cycle test) to the oxygen uptake (V̇o2) scale. The predicted V̇o2max value is obtained by connecting the point on the V̇o2 scale (V̇o2, liters per minute) with the corresponding point on the pulse rate scale (in beats per minute). The horizontal line extends from the workload and V̇o2 scales to the pulse rate scale. Where the line intersects the max V̇o2 scale (liters per minute) is the estimate of the individual's V̇o2max. Reprinted with permission from Astrand I. Aerobic capacity in men and women with special reference to age. Acta Physiol Scand. 1960;49(suppl 169):2–92.

Table 2.

Astrand and Ryhming Cycle Ergometer Test: Correction Factor for Age-Predicted Maximal Heart Ratea

Reliability and validity.

Astrand32 reported a correlation (r) of .71 between the measured V̇o2max and the estimated V̇o2max in the original A-R Cycle Ergometer Test and a correlation (r) of .78 between the measured V̇o2max and the A-R Cycle Ergometer Test using the age-correction factor. Teraslinna et al36 reported a correlation (r) of .69 between the original A-R Cycle Ergometer Test and the measured V̇o2max and a correlation (r)of .92 using the age-correction factor in a sample of 31 sedentary men. Kasch37 reported that the A-R Cycle Ergometer Test predicted a V̇o2max that was too low (by 21%) in 83 men aged 30 to 66 years. Other researchers38,39 have reported similar findings. Hartung et al,34 in a study of women aged 19 to 70 years, found that the A-R method overestimated V̇o2max by 3% to 21%. In addition, an overestimation of V̇o2max by the A-R method has been documented in women who were pregnant.40

Legge and Banister33 reported a correlation (r) between their revised nomogram and the measured V̇o2max of .98. Hartung and colleagues9,34 reported a correlation (r) of .95 between the measured V̇o2max and the estimated V̇o2max using their revised nomogram. However, the revised nomogram still predicted a V̇o2max that was too low (by 8.1 mL·kg−1·min−1) in a sample of sedentary and trained men.9 In a sample of women (n=38) aged 19 to 47 years, the revised nomogram overestimated V̇o2max by 18.5%.34

Strength and weaknesses.

The A-R Cycle Ergometer Test is one of the most frequently used submaximal cycle ergometer tests.41,42 This test has been a standard used by fitness facilities as part of fitness evaluations and to develop a training plan and evaluate the results.43 The protocol uses HR, which is easy to measure. Limitations of the test include the margin of error in the predicted V̇o2max values. The protocol can elicit lower-extremity discomfort in some people, which may invalidate the results.

Canadian Aerobic Fitness Test


The Canadian Aerobic Fitness Test (CAFT), formerly known as the Canadian Home Fitness Test, is unique in that it was designed to promote fitness testing at home. The CAFT was developed on a sample of 1,544 individuals (699 men and 845 women) aged 15 to 69 years.44 The CAFT is a measure of fitness and is based on the duration of the step test and a 10-second recovery of HR (Tabs. 3 and 4). Norms for the recovery HR in men and women have been reported,44 and a “Physical Fitness Evaluation Chart” for various age groups is available (Tab. 5). In addition, Jette et al45 developed a regression equation for the CAFT to predict V̇o2max. A sample of 59 individuals, aged 15 to 74 years, completed the CAFT and then underwent a progressive treadmill test to evaluate V̇o2max.

Table 3.

Canadian Aerobic Fitness Test: Starting Tempo of the Stepping Exercise Based on Age and Sexa

Table 4.

Canadian Aerobic Fitness Test: Ceiling Postexercise Heart Ratesa

Table 5.

Canadian Aerobic Fitness Test: Physical Fitness Evaluation Charta

The CAFT was modified (mCAFT)46,47 following reports that it predicted a V̇o2max that was too low in women aged 20 to 30 years and in heavy, older, and well-trained individuals.4850 Use of too few stages can produce a ceiling effect, and, if the target HR is not attained, the V̇o2max prediction may be too low.46 The modification allows an individual to complete the number of stages necessary to reach a target HR within 85% of the age-predicted maximum. Weller et al46 developed 2 additional stages for the original CAFT for individuals who exceed stage 6. A new regression equation was also developed.51 The details of the CAFT52 are shown in the Appendix.

Reliability and validity.

The reliability of measurements of recovery time for HR for the CAFT was determined using a sample of 102 individuals (r=.79).44 In terms of validity, the regression equation developed by Jette et al45 had a multiple correlation (R) of .905 (SEmeas=4.08 mL·kg−1·min−1). The regression equation for the mCAFT demonstrated the same strength between the predicted V̇o2max and the measured V̇o2max as the original equation for the CAFT (mCAFT, r=.88; CAFT, r=.99), but there was a lower mean square error (mCAFT=37.0 and CAFT=63.3).

Strengths and weaknesses.

The CAFT is a step test and, therefore, is inexpensive to administer and requires no electricity or calibration. A person's power output can be calculated within 6% to 7% if the individual steps in time with the beat, stands erect on the top step, and places both feet flat on the ground at the end of each stepping cycle.11

This test may not be suitable for people whose ability to balance is diminished because no handrail is used. It is also difficult to monitor individuals while they are stepping. Because we believe there is a ceiling effect, we contend that the original protocol appears to be more suited for assessing individuals who are unfit. Further research is needed to validate both the CAFT and the mCAFT with people with various diagnoses.

12-Minute Run Test


The 12-Minute Run Test (12-MRT) was developed by Cooper53 in 1968. This test is based on the work of Balke,54 which indicated that various run-walk tests could relate V̇o2 to either the distance covered in a given period of time or the time taken to cover a given distance. A sample of 115 men with no health problems aged 17 to 52 years completed two 12-MRTs and a V̇o2max test on a treadmill, and a regression equation was developed. The details of the 12-MRT are shown in the Appendix.

Reliability and validity.

Test-retest reliability (r) of measurements obtained with the 12-MRT was reported by Cooper53 to be .90. In terms of validity, Cooper53 reported a correlation (r) of .90 between the 12-MRT distance and V̇o2max. Jessup et al55 reported a lower correlation (r) of only .13 between the 12-MRT and V̇o2max in a sample of male subjects with no health problems aged 18 to 23 years. Safrit et al56 reported findings similar to those of Jessup et al.55

Strengths and weaknesses.

The 12-MRT requires no specialized equipment and allows more than one individual to be tested at a time. We suggest that this test is appropriate for assessing the cardiopulmonary fitness of individuals with high levels of function. The 12-MRT has been modified as a 12-Minute Walk Test (12-MWT), which we believe is more appropriate for the rehabilitation setting.

This test was developed using a male population. No cross-validation group was used to validate the equation. The 12-MRT requires a constant level of motivation, and the individual must pace herself or himself. The wording of the instructions makes this a potentially maximal exercise test, so well-defined testing criteria are needed to ensure that it is a submaximal exercise test. Finally, this test fails to account for age or body weight, which can influence exercise responses.18

20-Meter Shuttle Test


The 20-Meter Shuttle Test (20-MST)57,58 assesses maximal aerobic power. This test was designed for children, adults attending fitness classes, and athletes participating in sports requiring constant stopping and starting. The test requires subjects to run between 2 lines spaced 20 m apart at a pace set by signals on a pre-recorded cassette tape (Fig. 2). Starting speed is 8.5 km·h−1, and the frequency of the signals is increased 0.5 km·h−1 each minute. When the subject can no longer maintain the set pace, the last completed speed (ie, stage) is used to predict V̇o2max. Leger and Lambert59 found that maximal speed, subsequently termed “maximal aerobic speed” (MAS), for 2-minute stages in the 20-MST, could predict V̇o2max, with a correlation (r) of .84 (SEE=10.5%). A regression equation was developed on a sample of 188 boys and girls aged 8 to 19 years.58 Another regression equation was developed for adults based on a sample of 77 adults (53 men and 24 women) aged 18 to 50 years, in which age is held constant at 18 years.58 Norms have been established for children aged 6 to 17 years.57 Berthoin et al60 modified the 20-MST by incorporating 1-minute stages rather than 2-minute stages because they reported that faster speeds could be achieved when the work stages were shorter. The details of the 20-MST are presented in the Appendix.

Figure 2.

The setup and protocol for the 20-Meter Shuttle Test. Reprinted with permission from Leger LA, Mercier D, Gadoury C, Lambert J. The multistage 20-meter shuttle run test for aerobic fitness. J Sports Sci. 1988;6:93–101.

Reliability and validity.

In terms of reliability, the test-retest correlation (r) for the 20-MST was reported to be .89 for children (n=139) aged 8 to 19 years and .95 for adults (n=81) aged 20 to 45 years.58 Leger et al58 reported a correlation (r) of .71 (SEE=5.9 mL·kg−1·min−1) between the 20-MST and measured V̇o2max in children and a correlation of .90 in adults. Paliczka et al61 confirmed the validity of measurements obtained with the 20-MST by demonstrating a high correlation between 20-MST and V̇o2max (r=.93), as well as with a 10-km race time (r=−.93). The test has been further validated on active women.62 The 20-MST was reported to yield valid and sufficiently sensitive measurements such that the intensity of exercise could be modified for children with asthma.63

Strength and weaknesses.

The 20-MST is based on an individual's MAS. This test has multiple stages, enabling a wide range of fitness levels to be tested. It requires little equipment, and more than one individual can be tested at a time. The 20-MST is unique because it paces the individual with the use of sound signals on a pre-recorded cassette tape.

Due to the frequent stopping and starting of this test, we believe that it is important to screen the individual prior to testing to ensure that she or he is suitable. The test may not be suitable for some individuals due to the progressive increments of speed each minute and the requirement to pivot when they run between 2 lines. For example, this test may not be suitable for elderly people or those with musculoskeletal impairments. Some individuals may find it difficult to pace themselves with the signals. Finally, testing criteria are needed to ensure that the test is submaximal.

1-Mile Track Walk Test (Rockport Fitness Test)


The 1-Mile Track Walk Test (1-MTW), also known as the Rockport Fitness Test, estimates V̇o2max across a range of age groups and fitness levels. The prediction equations were developed based on a sample of 390 volunteers with no health problems (183 men and 207 women, aged 30 to 69 years).64 Each individual performed a minimum of two 1-MWTs on separate days. The walk times in the 2 tests had to be within 30 seconds. All individuals also performed a V̇o2max test on a treadmill. This test has also been validated on individuals with mental retardation.65,66 Variations in the distance used with this test have been reported (eg, 1-mile run/walk,67,68 1.5-mile run,68 2-mile run69). The details of the 1-MWT are given in the Appendix.

Reliability and validity.

Kline et al64 reported the reliability (r) of measurements obtained for the last quarter-mile HRs to be .93 (SEE=7.6 bpm) and the reliability (r) of measurements obtained for the total time for the 2 trials of the track walk to be .93 (SEE=0.26 minute). The validity of the regression equation was established by having a validation group (n=174) and a cross-validation group (n=169). The sample of 390 volunteers were assigned to the validation and cross-validation groups based on alternate case assignment (ie, odd-even case selection). Descriptive statistics revealed no difference between the 2 groups. The correlation (r) between the predicted V̇o2max and the actual V̇o2max was .93 (SEE=0.325 L·min−1) for the validation group and .92 (SEE=0.355 L·min−1) for the cross-validation group.

Strengths and weaknesses.

This test, in our view, is applicable to a wide range of individuals. It requires little specialized equipment and uses the familiar activity of fast walking. Thus, we contend that it is suitable for use in the rehabilitation setting. The test was cross-validated, which confirms the accuracy of prediction.

The test and regression equations need to be validated in patient groups. Finally, research is needed to determine whether a practice test improves the prediction.

Performance Submaximal Tests

Self-Paced Walking Test


The Self-Paced Walking Test (SPWT)70 is an exercise test developed for elderly and frail individuals. It consists of free walking at 3 speeds down an indoor corridor (ie, 250 m). Various exercise responses can be assessed such as speed, time, stride frequency, stride length, HR, and predicted V̇o2max. The test was developed on 24 individuals aged 64 to 66 years. Ten active students aged 19 to 21 years served as a comparison group. Each subject performed the SPWT and a progressive cycle ergometer test. Only 17 elderly individuals could complete the cycle ergometery test, whereas all of them completed the SPWT.

Performance of the SPWT is correlated with V̇o2max and is independent of age.71 Following an exercise program, the speed of walking was reported to increase, whereas HR remained unchanged.72 A predicted V̇o2max can be obtained from estimating V̇o2 from an aerobic demand curve and then extrapolating a predicted V̇o2max from V̇o2 and HR. To date, this test has been used primarily with older individuals.7173 The details of the SPWT are presented in the Appendix.

Reliability and validity.

The test-retest reliability for measurements obtained with the SPWT when it was repeated a few days later for the older group was ±5.2%, ±4.7%, and ±11% for the fast, normal, and slow paces, respectively.70 The younger group varied by ±7%, but no difference was found between the 2 tests on separate days. In terms of validity, the assessments (ie, standardized HR from the SPWT and a progressive cycle test) were correlated (r=.79).70

Strengths and weaknesses.

The SPWT assesses cardiopulmonary fitness as well as walking efficiency, both of which are beneficial in daily activities.70 This test is suitable for individuals requiring mobility devices or when a treadmill or cycle ergometer is not indicated. This test may also be suitable for monitoring an older person's mobility status over time, including the effects of aging and the effect of using mobility aids and devices.70 The information obtained from this test can provide safety guidelines (eg, for crossing an intersection safely requires a speed of 3.5 ft/s).24 Individuals who are at risk for injury while crossing an intersection may be identified. Individuals who are not able to walk at this speed should be identified as being not safe, and alternative means of mobility or mobility aids need to be recommended.

This test is limited because it does not provide a measure of endurance and may not be sufficiently sensitive to test individuals with higher levels of function. For some individuals with diminished function, it may be too difficult to complete the 3 selected walks with only 5 minutes of rest.

Modified Shuttle Walking Test


The Modified Shuttle Walking Test (MSWT) was modified from the 20-MST to provide a standardized progressive test for obtaining a symptom-limited maximum performance in individuals with chronic airway obstruction (CAO).74,75 The individual walks up and down a 10-m course at incremental speeds of 0.17 m/s each minute dictated by a prerecorded audio signal on a cassette deck74 (Tab. 6, whereas the original 20-MST required the individual to run a 20-m distance at a starting speed of 8.5 km/h with increments of 0.5 km/h each minute.58

Table 6.

Modified Shuttle Walking Test: Protocola

A sample of 35 individuals with CAO aged 45 to 74 years was used to develop the test.74 This test has been further validated on individuals with pacemakers.76 Singh and colleagues74,75 have recommended the MSWT for use as an assessment tool for individuals with a wide range of cardiac and respiratory disabilities. The details of the MSWT are shown in the Appendix.

Reliability and validity.

The measurements obtained with this test were replicable in a sample of 10 individuals after one practice trial.74 The mean difference between trials 2 and 3 was −2.0 m (95% confidence interval of −21.9 to 17.9 m). The validity of measurements obtained with the test, which was established by comparing the distance completed during the MSWT with the distance completed during the 6-Minute Walk Test (6-MWT), was moderate (rho=.68).74 The HRs, however, were higher on the MSWT, indicating a greater cardiovascular response. A strong relationship (r=.81 and r=.88) was observed on comparing the V̇o2max recorded in 2 tests using treadmill walking with the V̇o2max recorded during the MSWT.75

Strengths and weaknesses.

The MSWT requires little equipment and is easy to administer. The audio signal standardizes the increments in walking speed and motivates the individual. We believe that the initial speed is sufficiently slow to be used with most types of patients. No individual in the studies attained the highest level (ie, level 12).74,75 This test can be used to prescribe an appropriate walking speed for an exercise program by evaluating the individual's HR and RPE responses at the various stages.74

This test, however, requires a near-maximal effort by having the speeds continue to increase. We believe, therefore, that it is essential to monitor the individual during the test to ensure that she or he is responding appropriately. Familiarizing the individual with the pacing required for the test may require some time.

Bag and Carry Test


The Bag and Carry Test (BCT)77 is used to assess a task that evaluates both endurance and muscle force. The BCT involves walking a circuit carrying a 0.9-kg package for 7.5 m, up and down a 4-step flight of stairs, and back 7.5 m. On the completion of each circuit, 0.9 kg is added to the package until the individual can no longer complete the circuit. It requires 10 minutes to complete. A sample of 61 women aged 48 to 93 years was recruited from the community and a residential home. Fifty-six subjects completed the test. The maximal weight they could carry up and down the stairs ranged between 3 and 26 kg. The test developers concluded that this test was easy to administer and suitable for testing individuals with higher levels of function. The details of the BCT are presented in the Appendix.

Reliability and validity.

The test-retest reliability of measurements obtained with the BCT was established by administering the BCT 3 days later (r=.89). Maximal HR was 90%±10% of the HR achieved during the V̇o2peak test on a cycle ergometer. The BCT correlated with force of the quadriceps femoris muscle (r=.43), hamstring muscle (r=.54), gastrocnemius muscle (r=.52), and soleus muscle (r=.62).77

Strength and weaknesses.

The BCT is designed to integrate endurance, muscle force, and balance capability and is based on an everyday activity. This test, in our view, is easy to administer and can be used in research and clinical settings. However, it may be difficult to replicate the test with 4 steps. A platform or landing should be at the top of the stairs to allow the individual to turn around safely.

The guidelines for administering this test are not well described in the literature. There are no specifications regarding the height of the steps, whether the individual is allowed to use a handrail for support, or whether a practice trial is required. In the absence of criteria for administering the test, this test could become a maximal test if the individual is not properly monitored during the test. We argue that this test has the potential to be a very useful submaximal exercise test if the individual is timed as opposed to being scored by only the weight he or she carried. The number of circuits completed in a specified time could be measured, or the time to complete the circuit while carrying a specified weight and walking at a safe and comfortable pace could be scored.

Timed Up & Go Test


The Timed Up & Go Test (TUGT)78 was modified from the Get-up & Go Test.79 Both tests are based on a functional task of rising from a standard armchair, walking 3 m, turning, and returning to the chair. Podsiadlo and Richardson,78 however, changed the scoring system from an observer rating of 1 to 5 to a timed version. The test was modified using a sample of 60 frail, community-dwelling, elderly individuals (23 men and 37 women, aged 60 to 90 years) and 10 volunteers with no health problems (6 men and 4 women, aged 70 to 84 years). Medical diagnoses of the study population included cerebrovascular accident (n=23), Parkinson disease (n=10), rheumatoid arthritis or osteoarthritis (n=9), and miscellaneous conditions (eg, postsurgical hip fractures, general deconditioning) (n=8).78 The TUGT has been used as a test of mobility to assess change following an exercise program for elderly individuals aged 79 to 86 years80 and aged 75 to 96 years.81 No improvement in mobility based on this test following an exercise programs was reported.80,81 The details of the TUGT are given in the Appendix.

Reliability and validity.

The interrater reliability for times obtained on the same day and the intrarater reliability tested 3 days to 5 weeks apart were good (intraclass correlation coefficient=.99 for both).78 Validity was assessed by correlating the time (in seconds) on the TUGT with the log-transformed scores on the Berg Balance Scale (r=−.72), gait speed (r=−.55), and Barthel Index of Activities of Daily Living (r=−.51). The correlations were negative, indicating that those individuals who took longer with the TUGT had lower scores on the Berg Balance Scale, with gait speed, and on the Barthel Index.

Strengths and weaknesses.

The TUGT is easy to administer, and no training is required. This test is easy to perform in research and clinical settings. The results from this test provide information related to mobility. Based on the time taken to complete the test, the level of assistance required in mobility tasks can be determined.78

A limitation of this test is that it may not detect a change following an exercise program because of the lack of sensitivity of the measure.81 Further studies are warranted to examine its sensitivity, using a larger sample, and to investigate its predictive capacity. Sensitivity could possibly be improved by increasing the distance walked or having subjects sit down and get up again at each end of the 3-m walkway, but research is needed to determine whether this is true.

12- and 6-Minute Walk Tests


The 12-MWT was introduced by McGavin and colleagues82,83 to assess the distance covered in 12 minutes in individuals with chronic bronchitis. The total distance covered in 12 minutes is recorded, and the individual is allowed to stop and rest. This test was modified from the 12-MRT described by Cooper53 for individuals without health problems. The 12-MWT has been used primarily for people with COPD,8291 but it has also been used with college-aged students.92

Butland et al93 reported that similar results could be obtained in 6 minutes. Guyatt et al94 applied the 6-MWT in individuals with heart failure. The 6-MWT has been used with individuals with end-stage lung disease,95 people with chronic heart failure,96,97 people with COPD,98100 children who are severely ill,101 people with chronic renal failure,102 and older adults between the ages of 65 and 89 years.103 Two practice tests appear to be required to obtain reproducible results,93,94 the walking circuit needs to be identical,92 and encouragement needs to be standardized.27 Walk tests with durations of 4 minutes104 and 2 minutes93 have also been reported. The details of the 6-MWT and the 12-MWT are presented in the Appendix.

Reliability and validity.

Reliability has been assessed for measurements obtained with the 12-MWT. Mungall and Hainsworth89 reported a coefficient of variation of ±8.2% over 6 tests. This statistic, however, is not a probabilistic measure, which is normally used to assess reliability. If the results of the first 2 tests were eliminated, however, the coefficient of variation was reduced to ±4.2%. Guyatt et al94 also reported that 2 practice tests are required. Other researchers95,96 have reported intraclass correlation coefficients of .96 to .99 between the second and third administrations of the 6-MWT, suggesting that only one practice test is required.

The concurrent validity of measurements obtained with the 6-MWT and the 12-MWT based on measurements of V̇o2max or V̇o2peak is not clear. Some investigators have reported a correlation between the distance covered in the 6-MWT and V̇o2peak (r=.6496 and r=.70101) as well as between the distance covered in the 12-MWT and V̇o2max (r=.4985 and r=.5282). Other researchers have reported no correlation between V̇o2max and either the distance covered in the 6-MWT94 or the distance covered in the 12-MWT.82 The physiologic demand of the walk test appears to be distinct from that of cycle ergometer tests and, therefore, may be a better indicator of function in normal daily activities.102,105 The correlation between lung function and the distance covered in the 6-MWT and the 12-MWT has also shown conflicting results.85,99,100

Strengths and weaknesses.

The 6-MWT and the 12-MWT are simple tests that are inexpensive to administer. Walking for a given time seems, in our opinion, to correspond to functional activities used in daily activities. These tests, therefore, can be administered to individuals without health problems and to patients with a variety of diagnoses. The use of a standard time rather than a predetermined distance provides a better test of endurance.82 The tests allow the individual to set her or his own pace and stop if necessary. The 12-MWT can be used to detect a change following an exercise program.86

In the literature in which these tests are described, the number of practice trials varies. Often only one practice test was given,99102 and there was considerable variation in the rest periods between tests. Some investigators95,96,100,102 reported having the subjects perform the test on the same day as the practice, whereas other investigators94,99,105 reported having the test and practice on separate days. In addition, different versions of the instructions have been used for both the 6-MWT and the 12-MWT. Some investigators101 stated in their instructions that the individual was allowed to stop if necessary, whereas other investigators85 instructed the individual to pace herself or himself so that she or he would not have to stop. The scoring of the test has also varied. Most investigators83,94,105 used the final distance (ie, that of the last test trial), whereas some investigators99 reported the longest distance of all test trials. We believe that other limitations of the timed walk tests include lack of monitoring of physiological variables while the individual is completing the test and the lack of specific performance criteria to ensure that a maximal effort is not performed.

Other Performance Tests

Performance tests are frequently incorporated as a measure of mobility in global physical assessments used for elderly people. The most common performance test is a measure of walking speed, which is similar to the 3 walking speeds (ie, slow, normal, and fast) used with the SPWT. Typically, a 10-ft walk106108 is used for assessing individuals who are confined indoors and a 50-ft walk107,109111 is used for all others. A 30-m walk has also been used, as this is the usual distance for pedestrian crossings.112 The instructions are for the individual to walk from a standing start at his or her regular pace and to use any mobility devices that he or she normally uses.106 The individual is timed, and the walking speed (in meters per second or feet per second) is calculated. Reports of average walking speed range from 0.74±0.29 m·s−1 for individuals aged 60 to 99 years113 to 1.1 and 1.2 m·s−1 for 70-year-old women and men with no health problems, respectively.112

The assessment of walking speed is very important for assessing independent mobility in the community. Pedestrian intersection crossing times are calculated based on a walking speed of 1.22 m·s−1.114 A walking speed of 11.5 m·min−1 is a threshold value for predicting nursing home status,107 with a normal walking speed being 70 m·min−1.107 Two factors, quadriceps femoris muscle weakness and joint impairment, are thought to be critical variables in determining walking speed, which, in turn, determines some aspect of dependency in elderly people. Variations in walking speed are due to a change in stride length rather than an alteration in frequency or cadence.113 The onset of pathology shortens the stride length and influences speed of walking.113 Chronologic age is not thought to be a primary factor in determining gait speed.113

Researchers have assessed a maximum walking speed for a given distance (eg, 30 m). In a sample of 70-year-old subjects (n=602), the maximum walking speed was the most reliable predictor of dependence in activities of daily living.112 The critical levels for the threshold of being dependent in activities of daily living was found to be a maximum walking speed of 1.7 m·s−1 in men and 1.5 m·s−1 in women. It is not entirely clear whether a decline in cardiopulmonary fitness affects walking speeds for short distances; it is more likely to be a contributing rather than a primary factor.112

Other performance tests cited in the literature include a step test. This test requires 3 boxes combined to form steps of 10, 20, 30, 40, and 50 cm in height and a handrail on the wall.112 The highest possible step height that the individual is able to climb up and down with either leg and without a rail is recorded. There are also variations of this step test.115,116 Correlations have been reported between the maximum step height up and down with a comfortable walking speed in 70-year-old men (r=.39) and women (r=.37).112

An obstacle course described by Imms and Edholm113 is used in a test that is similar to the BCT. In this test, the individual rises from a chair, walks across the room, climbs 3 stairs (rails on either side), turns around, descends the stairs, and returns to the chair. The individual is allowed to go at her or his own speed and to use a mobility aid. Two practice trials are given, and the time (in seconds) taken to complete the course is recorded. In a sample of 71 subjects (28 men and 43 women) aged 60 to 99 years, the time to complete the course was not correlated with age but was correlated with walking speed (r=−.80).113

Examples of Test Selection

  1. The patient is a 65-year-old man with severe chronic airflow limitation and right atrial enlargement. He has no history of angina but does have hypertension, which is controlled with medication. He is 18.1 kg (40 lb) overweight and is unaccustomed to physical activity. His activity is normally terminated by shortness of breath.

    Indications: to establish an exercise profile to ensure that he is safe to undertake an exercise program and to define the parameters for such a program.

    Test: 6-MWT or SPWT.

    Clinical Decision-Making Process:The 6-MWT and the SPWT are both suited for older patients with chronic lung disease. This patient is deconditioned, overweight, and hypertensive. These tests enable him to perform an activity (ie, walking) that is useful to him on a daily basis. In addition, with portable equipment, including HR monitor, BP measurement apparatus, and pulse oximeter, he can be readily monitored. Furthermore, the scale of breathlessness can be used to assess his symptoms. The physical therapist can correlate the rating of breathlessness and physiologic parameters to prescribe the parameters of an exercise program, including type of exercise, intensity, frequency, duration, continuous versus discontinuous program, and its course. These tests can be repeated at various intervals to evaluate the outcome of the training program.

  2. The patient is a 52-year-old man who had bypass surgery 10 years ago. He had one recurrence of angina. He has intermittent claudication in the left calf at a moderate walking speed.

    Indications: to establish safe exercise intensity (no anginal symptoms) and a training program for his peripheral vascular disease as well as heart disease.

    Test: Modified Bruce Treadmill Test, SPWT, or 6-MWT.

    Clinical Decision-Making Process:This patient is showing signs and symptoms of reocclusion of his coronary arteries and stenosis of a lower-extremity artery, which results in claudication. He could be a candidate for maximal exercise testing; however, if he stops because of leg pain, the test results will be limited. Alternatively, he could undergo submaximal exercise testing (eg, Modified Bruce Treadmill Test, SPWT, 6-MWT). Because of his cardiac history, precautions must be taken. Having a cardiologist present is recommended, and the treadmill test is preferable for monitoring electrocardiographic activity. If the electrocardiogram is normal, the SPWT or the 6-MWT can be performed, and one of these tests can be used to assess training response, if preferred. The parameters of the training program are set to keep the patient below his anginal threshold and his leg pain tolerable.

Summary and Conclusions

Physical therapists are clinical exercise specialists who apply exercise as an assessment and diagnostic tool and in treatment. We believe that they should have a thorough knowledge of exercise testing, including submaximal exercise testing. Physical therapists, in our opinion, need to assume a role in refining existing exercise tests and measures and to assume a leadership role in developing new tests and measures. Refinement of submaximal exercise tests is needed to increase their sensitivity as assessment, diagnostic, and treatment outcome tools and to provide valid indexes of a person's capacity to assume a given type of employment, homemaking activities, and activities of daily living. They also need to serve as a basis for exercise prescription.

We believe that there is a need for standardized submaximal ergometer tests for people with musculoskeletal limitations, people who have impaired balance, people who are overweight, people who are unable to walk on a treadmill for other reasons, and for people who require close monitoring during exercise. There is also, in our opinion, a need for the development of upper-extremity submaximal exercise tests for people with lower-extremity paresis or severe deformity.

We contend that stringent monitoring of exercise responses is essential both for test validity and for safety. When testing people with a wide range of conditions, including cardiovascular and cardiopulmonary conditions that can be life threatening, even people without known health problems can exhibit unexpected responses. People without known health problems, for example, can have cardiac dysrhythmias; and this incidence increases with advancing age.20 Safety and minimizing undue strain, in our view, are essential in planning and implementing submaximal exercise testing.

Research is also needed for the development and refinement of scales used to assess exercise response (eg, exertion, breathlessness, fatigue, discomfort or pain, and even well being associated with physical activity). Given that people are limited by their symptoms that correlate to physiologic measures, assessment of their symptoms can provide critical information about their exercise responses as well as a basis for setting the intensity of tolerable physical activity or an exercise program.


  • Both authors provided concept/idea, writing, literature collection and analysis, project management, and consultation (including review of manuscript before submission). Dr Dean also provided clerical support.


    View Abstract