(Circulation. 1995;92:614-621.)
© 1995 American Heart Association, Inc.

Clinical Trials for Claudication

Assessment of Exercise Performance, Functional Status, and Clinical End Points

William R. Hiatt, MD; Alan T. Hirsch, MD; Judith G. Regensteiner, PhD; Eric P. Brass, MD, PhD; and the Vascular Clinical Trialists¹

From the Department of Medicine, Section of Vascular Medicine, University of Colorado Health Sciences Center, Denver; the Vascular Medicine Program, University of Minnesota Medical School, Minneapolis; and the Department of Medicine, Harbor-UCLA Medical Center, Los Angeles, Calif.

Correspondence to William R. Hiatt, MD, Section of Vascular Medicine, University of Colorado Health Sciences Center, 4200 E Ninth Ave, Box B-180, Denver, CO 80262.

Key Words: peripheral vascular disease • tests • clinical trials • exercise

 
Peripheral arterial disease (PAD) affects a large proportion of the general population, with an age-adjusted prevalence of approximately 12% and a prevalence of intermittent claudication of 3% to 7%.^{1 2} In symptomatic persons, the limited lower extremity arterial supply cannot meet the dynamic metabolic demand of the muscles during ambulatory activities, resulting in the symptom of claudication. Claudication is associated with a severe limitation in walking ability,³ which may adversely affect social, leisure, and occupational activities in many patients.⁴

The treatment of all patients with PAD is initially directed at cardiovascular risk factor modification, since these individuals have a high future risk of cardiovascular mortality.⁵ Severely affected patients who have ischemic rest pain or tissue loss are candidates for interventional therapy (bypass surgery or angioplasty) to maintain limb viability.^{6 7} However, since the majority of patients with claudication are not at short-term risk of limb loss, the primary therapeutic goal is to improve exercise performance and community-based functional status.

The past decade has witnessed a marked increase in the evaluation and utilization of therapies to treat patients with claudication.⁸ Percutaneous transluminal angioplasty is considered an appropriate intervention for patients with "earlier stages of symptomatic disability" due to claudication,⁹ and the American Heart Association has recently recommended that invasive interventions are appropriate for patients with incapacitating claudication.⁶ In addition, there is increased interest in medical therapies for claudication. Exercise training elicits well-established and clinically important changes in treadmill exercise performance and community-based walking ability.^{3 10 11} Recent pharmacological advances have led to a greater use of drugs to treat claudication, with new agents in clinical development. Examples include drugs that alter blood rheology¹² and drugs that improve ischemic skeletal muscle metabolism.¹³

A clinical classification to evaluate therapies for PAD has been proposed by Rutherford¹⁴ and adopted by the vascular surgery and interventional radiology communities.⁶ This classification system uses a ranking of (1) symptom severity, (2) the subject's ability to complete an exercise test of 5 minutes at 2 mph, 12% grade, and (3) the degree of arterial occlusion as estimated by the ankle systolic blood pressure response to exercise. Other published guidelines have defined clinical success of claudication therapy as an increase in walking time or distance on a treadmill protocol set at 2 mph, 12% grade.¹⁵ Importantly, these evaluation procedures have several deficiencies that do not allow for a comprehensive, reproducible assessment of the broad range of functional limitations in patients with PAD. Future standards must be developed that address the clinically relevant changes in exercise performance and functional status that occur in response to the treatment of claudication. Our goal was to critically review the utility of testing methodologies that have been applied in clinical trials for the treatment of claudication. Our recommendations regarding the future use of exercise tests and functional status measures are intended to emphasize clinically relevant outcomes and the appropriate evaluation of new therapies.

 
The evaluation of claudication severity requires a comprehensive approach, using both objective exercise testing and "patient-focused" clinical outcomes, such as outpatient walking ability, physical activity, or quality of life. In the PAD population, useful evaluation procedures must have adequate precision and accuracy across the full spectrum of claudication symptoms. Because of the heterogeneity of functional status in patients with claudication, optimal testing procedures for a patient with pain after walking a half block may be very different than for a patient who can walk six blocks but who is primarily impaired in work-related physical activities. Claudication testing should also maintain accuracy as patients improve their range of function in response to a therapeutic intervention. In addition, the test characteristics must be stable with respect to time and nature of the intervention. Since PAD is a chronic disease, multiple tests over a period of months are usually required to fully evaluate the therapy. The introduction of any learning effect or bias induced by repetitive testing would make assessment of the efficacy of the intervention difficult. Finally, claudication testing should be acceptable to the patient and the physician and should be performed on readily available equipment.

 
There are several noninvasive vascular tests that accurately diagnose PAD. These tests include recordings of systolic blood pressures (the ankle-brachial index, or ABI), pulse volumes, and duplex ultrasound imaging with blood flow velocity measurements in the lower extremity. Critically, none of these hemodynamic tests correlate well with the degree of exercise impairment or functional limitation in patients with claudication.^{16 17} In addition, a number of medical treatments are known to significantly improve treadmill exercise performance and functional status without a change in ABI or leg blood flow.^{10 18} Therefore, monitoring changes in peripheral hemodynamics alone in clinical trials is an inadequate measurement of treatment success in patients with claudication.

 
Treadmill testing is designed to obtain objective data on the functional exercise capacity of patients with claudication. Two contrasting approaches were used in the PAD population: constant-load (single-stage) testing and graded exercise testing.

Constant-Load Exercise Testing
The traditional constant-load treadmill test, initially applied for clinical use in the PAD population by Carter,¹⁹ is conducted at low speed (1.5 to 2.0 mph) at a defined grade (ranging from 0% to 12%).^{19 20 21} The distance at which the patient first notices the onset of claudication pain is recorded (initial claudication distance, or ICD), and the test is terminated when the patient reaches a maximal level of claudication pain (absolute claudication distance, or ACD).

Despite its long-standing clinical utility, the constant-load treadmill test demonstrates coefficients of variation of 30% to 45% for ICD and ACD.^{20 22} This large variability often necessitates repetitive testing in individual patients to establish the true baseline walking distance or time.^{12 23} Overall, international experience with this testing strategy indicates that as many as 20% of potential patients are excluded on entry to a study due to exercise tests that exceed the treadmill's range of accuracy.^{12 23 24}

Because of the heterogeneity of walking ability in the PAD population, it is difficult to select a single optimal fixed workload for testing. If the grade is too steep, many patients will experience pain very early in the testing protocol and be incapable of performing a meaningful test. In contrast, if the grade is set too low, high-functioning patients may walk almost indefinitely, terminating the test due to symptoms other than claudication, precluding assessment of claudication severity. Thus, lower and upper ranges for walking distance are necessary for the constant workload treadmill test, and walking distances that fall outside these arbitrary limits are considered invalid.²¹ The upper limit requirement is particularly problematic in therapeutic trials, because patients who demonstrate a large degree of clinical improvement may not be claudication-limited as assessed by constant-load testing on study exit, leading to an underestimation of treatment efficacy.^{24 25}

Repetitive testing over several weeks on the constant-load treadmill is associated with changes in exercise performance that complicate statistical analyses. For example, patients in a control group who have multiple tests at the same constant workload experience an increase in maximal walking distance.^{12 20} This temporal augmentation of exercise distance in patients not on active therapy limits the ability to assess treatment effect in study subjects. The mechanism underlying this effect is unknown but may represent a "learning curve" as patients repeat the test, a classic placebo effect, or it may be due to improved walking biomechanics as subjects become familiar with the treadmill apparatus. In untreated (control) patients, the magnitude of improvement in walking distance from repetitive testing is typically 36% for the ICD and 25% for the ACD¹² but may be as great as 100% for the ACD over a 2-week period.²⁶ This large treatment-independent response must be recognized in the power calculation for any trial in which constant-load testing is used.

Despite these limitations, constant-load treadmill testing has been well accepted by both patients and physicians and provides an important historical database from prior exercise training studies,^{27 28} trials of drug therapy,¹² and surgical intervention studies.^{11 29} The treadmill equipment is widely available, and testing requires minimal preparation. Thus, the constant-load treadmill has been an important (albeit imperfect) cornerstone for clinical research in PAD for over 25 years.

Graded Treadmill Testing
In the 1960s, Bruce and colleagues³⁰ developed graded treadmill protocols to perform functional assessments of patients with cardiac disease. These protocols were characterized by an initially low work demand that could be sustained by even the most impaired patient. The workload was then increased until each patient reached a definable, reproducible peak workload during a test of moderate duration. The success of these protocols was due in part to their ability to reproducibly define peak exercise performance across the full spectrum of cardiac impairment. Graded treadmill testing in patients with coronary artery disease has proven its utility in studies of exercise training, drug therapy, coronary angioplasty, and coronary artery bypass surgery.^{31 32 33 34}

Recently, the graded testing concepts developed for patients with cardiac disease have been extended to patients with PAD. Table 1 shows the design of two graded treadmill protocols that have been validated in the PAD population as compared with the "standard" constant-load protocol. Using graded treadmill protocols, the within-subject coefficient of variation for the ICD is 15% to 25% and for the ACD is approximately 12% to 13%.^{16 22} In addition, the between-subjects variability is minimized with the graded protocol (Table 2). The graded exercise test has been shown to accommodate PAD patients with varied disease severity without modification of the rate of increase in workload. Previous experience demonstrates that the graded treadmill test is well accepted by nearly all PAD patients, including patients over the age of 70 years.³⁵ Therefore, patients need not be excluded from study on the basis of age or initial ABI.

Sequential testing of untreated patients with the graded treadmill test is not associated with temporal improvement in the initial or absolute claudication distances or times.^{16 22} This stability of measurement facilitates clinical trials, as fewer subjects are required to define relatively small but clinically important changes in exercise performance. Since the graded test may be providing a true physiological assessment of the claudication-limited peak exercise performance (analogous to exercise limitations in cardiac output in patients with ischemic heart disease), the absence of a repetitive testing effect is not surprising.

Changes in exercise performance using the graded treadmill test have been shown to correlate with improved outpatient ambulatory function in patients with PAD.^{10 36} This association is important because it allows even small-percentage improvements in treadmill performance to be interpreted as clinically relevant. As the workload of the test sequentially increases, the meaning of a 1-minute change early in the test versus at higher workloads is quite different. Thus, percentage improvements in performance time or distance should not be considered comparable between the fixed and graded treadmill designs. The performance of graded exercise testing protocols requires treadmill equipment that is widely available at most clinics and hospitals that perform cardiac stress testing and extensive physician, technician, and patient experience with these systems.

Measurement of Oxygen Consumption With Exercise Testing
The measurement of oxygen consumption during exercise testing provides objective physiological information regarding peak exercise performance, the ventilatory responses to exercise, and the metabolic cost of submaximal workloads.^{25 37} The oxygen consumption measurement is reproducible¹⁶ and is modified with both medical and surgical treatments of claudication.^{10 36} However, measurement of oxygen consumption requires expensive equipment and trained personnel, and thus availability of this technique is primarily limited to research sites.

 
Treadmill testing is the primary means of assessing changes in the physiological status of the claudicant at baseline and in response to therapy. However, the treadmill test alone does not directly define the degree of functional disability of the patient in the community setting. Therefore, questionnaires that describe disease-specific as well as more global aspects of function are essential measures of therapeutic outcome in PAD clinical trials. The battery of questionnaires and activity monitors that are now available to investigators can provide a critical assessment of outcomes in the treatment of claudication. These instruments not only assess changes in walking ability and physical activity levels but also describe the impact of claudication and its treatment on a patient's global health status.

Questionnaires
The Walking Impairment Questionnaire (WIQ) was developed and validated specifically for patients with claudication to assess treatment effects on claudication-limited walking ability. The questionnaire quantifies the patient's walking capability in terms of defined distances and speeds and rates the severity of claudication pain during usual walking activities.³ Changes in graded treadmill exercise performance correlate with changes in questionnaire scores,³ and the questionnaire responses are stable when repeated over time in control patients. The WIQ has been used previously in the PAD population to evaluate changes in community-based walking ability in response to exercise training¹⁰ and surgical interventions.³⁶ The current version of the WIQ is presented in "Appendix 1."

The PAD Physical Activity Recall (PAD-PAR) questionnaire provides a global measure of habitual physical activity levels by estimating the total energy expenditure (in MET hours per week) of the patient at work, in the home, and during leisure time.³⁸ The PAD-PAR has been modified from the original version to be more appropriate for patients with claudication who can perform only low levels of physical activity. The PAD-PAR described changes in activity levels after treatment with an exercise rehabilitation program.³⁹ The current version of the PAD-PAR is presented in "Appendix 1."

The Medical Outcomes Study (MOS SF-36) questionnaire evaluates physical function and general health perceptions as well as limitations due to mental health, social function, and vitality.^{40 41} Thus, the MOS instrument assesses multiple aspects of normal life function. Measurements of the effects of PAD on physical, social, and role functioning provide a comprehensive insight into the degree of disability experienced by the patient as a result of the disease. The physical function aspects examined in an earlier version of this questionnaire improved with exercise training therapy of claudication.³⁹

Physical Activity Monitors
Self-assessment of physical activity in patients with claudication is subject to bias, and both under- or overreporting of symptoms is possible. Physical activity monitors have the potential to reduce this bias by continuously recording physical activity during the monitoring period.³⁷ Motion sensor estimates of physical activity have been validated against indirect calorimetry in the laboratory setting.³⁷ However, the use of activity monitors requires special equipment and places additional demands on the patient, who must remain compliant with the monitoring procedures. In selected small-scale trials, activity monitors serve to further validate the patient's response to treatment.³⁹

 
Treadmill Testing
The primary exercise outcome assessment in previous clinical trials in claudication has been the change in walking distance or time on a constant-load treadmill protocol. However, the use of a constant workload has several limitations, including a large within-subject and between-subjects variability and the placebo or learning response in patients not on active therapy. Also, a single, predetermined workload is not appropriate for a heterogeneous population of patients with different walking abilities. These limitations of constant-load treadmill testing have raised important questions as to the usefulness of the technique and led to a reevaluation of trials previously conducted using this methodology.⁴² Therefore, the constant-load test is not recommended as a primary means of assessing treatment outcomes in clinical trials of claudication.

In comparison, the graded treadmill test has a large dynamic range that allows all patients to be evaluated at a quantifiable, maximal claudication end point. Thus, a wide range of patients can be included in therapeutic trials, since there are no exclusions secondary to functional disease severity. Use of a graded protocol (which is highly reproducible) will reduce the number of treadmill tests on entry, and statistically significant results can be achieved with a smaller sample size, resulting in time and cost savings. The results of clinical trials that use graded treadmill testing are more generalizable, since all patients can be analyzed after treatment regardless of the magnitude of improvement. Finally, exercise performance on a graded protocol correlates with community-based walking ability determined by questionnaire.³ Thus, it is our recommendation that a graded treadmill protocol be used in all future clinical trials to assess the benefits of surgery, angioplasty, drugs, or exercise training in the treatment of claudication.

Questionnaires
The major treatment goal for claudication is to improve functional status of the patient and to relieve disability. Therefore, the use of disease-specific and more general functional status measures should become a critical end point in the comprehensive assessment of the patient with claudication. For example, an intervention that improves claudication would be expected to increase both the speed and distance walked as assessed by the WIQ. A lessening of claudication and increased activity would also result in an increase in PAD-PAR scores. The physical functioning scores of the MOS questionnaire should also increase, but mental health and general health perceptions may not change unless the claudication treatment affects several dimensions of health. For example, prolonged hospitalization, surgical wounds, and bed rest are known to diminish the patient's sense of well-being.

 
Claudication is a common condition that is associated with a limitation in function. Because patients are chronically disabled, the treatment of claudication has become a major clinical focus for vascular internists, vascular surgeons, interventional radiologists, cardiologists, and the pharmaceutical industry. Thus, it is critical that future clinical trials are well designed and use appropriate testing methodologies. The integrated use of graded treadmill testing protocols and appropriate questionnaires will permit clinically meaningful changes to be accurately assessed in patients with PAD.

 
{app}The authors wish to thank Robert B. Rutherford, MD (Vascular Surgery, University of Colorado), for his critical evaluation of the manuscript. The Walking Impairment Questionnaire is reprinted with permission from Blackwell Scientific Publications Inc. The current version (see "Appendix 1") was modified by Drs Regensteiner and Hiatt to include the stair-climbing questions and an expanded range of responses to each question. The Physical Activity Recall (PAR) was originally developed by Drs William L. Haskell and Steven N. Blair at Stanford University, who have given permission to publish the questionnaire. The current version of the PAR (the PAD-PAR) was modified by Drs Regensteiner and Haskell to reflect the low level of activities in patients with claudication. The MOS questionnaire has been previously published⁴⁰ and is available from the Medical Outcomes Trust, 20 Park Plaza, Suite 1014, Boston, MA 02116-4313.

 
¹ The Vascular Clinical Trialists are a group of investigators who are members of the Society for Vascular Medicine and Biology. The positions expressed in this article reflect those of the trialists and are not an official policy statement of the society. The names and institutional affiliations of the trialists are listed in "Appendix 2."

 
Peripheral Arterial Disease Questionnaires
Walking Impairment Questionnaire (WIQ)
This questionnaire is designed to assess the degree of impairment experienced by the patient with claudication during daily activities. Question 1 is divided into two parts. Section A is specific for calf or buttocks claudication and is used to create a summary score for analysis. Section B is used to evaluate other symptoms that may limit walking ability. If the patient ranks a symptom from Section B as more severe than claudication pain (from Section A), then claudication is not the limiting symptom, and responses from Questions 2 through 4 may not be valid for assessing claudication severity. A summary score is not created for Section B.

Questions 1A, 2, 3, and 4 are expressed on a scale of 0% (unable to perform because of severe claudication) to 100% (no impairment). For Questions 2 through 4, each individual response is multiplied by its respective weight to create an individual score. All individual scores for a question are added and then divided by the maximal possible score to create the % score used for analysis.

Peripheral Arterial Disease Physical Activity Recall (PAD-PAR)
Instructions for administration: Determine for each major category (sleep, work, house or yard, recreation or leisure) the estimated number of hours per week spent within that category during the preceding week. Then, using the cards as prompts, ask about specific activities within each intensity of activity (heavy to very light). It is not expected that every hour of the week can be accounted for. However, asking the subject to estimate their total sleep hours, and the total expected hours within each major category of activity, as compared to the breakdown of activities within each major category of activity, helps the subject more reliably remember their activities. Instructions for Question 2 pertain to all three major categories of activity.

Scoring: For each activity (heavy to very light), calculate the number of hours per week spent in that activity (days per week times hours per day). Sum hours per week in each category to determine total hours per week. The amount of energy expenditure for each activity is expressed as metabolic equivalents (METs). One MET equals 3.5 mL/kg per minute of oxygen consumption. Activities are classified according to the following scale: very light (0.9 to 2.0 METs), light (2.1 to 3.0 METs), moderate (3.1 to 5.0 METs), and heavy (5.1 to 7.0 METs). Data are reported in MET hours per week (hours per week times the MET value of the activity). 1. How many hours do you sleep a night, on average? hoursx7 Sleep hours per week= 2. Explain to subject that you are going to ask about typical work activities performed during the past week (includes work for pay or regular volunteer activities). If subject not employed, go to Question 3. How many total hours did you work per week on average?Work hours per week=

Here is a listing of typical work activities (show participant Card A). Activities are classified as heavy, moderate, light, and very light, depending on their average energy demands. With your job, time may be spent in more than one category of activity. Let's start with heavy activities and then go on to moderate, light, and then very light activities. (a) Please tell me the average number of days during the last week you performed heavy activities at work. (b) Please tell me the average length of time you performed heavy activities in a day. Then, repeat above directions for all intensities of activity.

 
Members of the Vascular Clinical Trialists of the Society for Vascular Medicine and Biology include William R. Hiatt, MD, and Judith G. Regensteiner, PhD (Vascular Medicine, University of Colorado), Mark A. Creager, MD (Vascular Medicine and Cardiology, Harvard University), John P. Cooke, MD, PhD (Vascular Medicine and Cardiology, Stanford University), Alan T. Hirsch, MD, and George C. Haidet, MD (Vascular Medicine and Cardiology, University of Minnesota Medical School), Jeffrey W. Olin, DO (Interventional Vascular Medicine, Cleveland Clinic), Thom Rooke, MD (Vascular Medicine and Cardiology, Mayo Clinic), William Pearce, MD (Vascular Surgery, Northwestern University), and Jeffrey M. Isner, MD (Interventional Cardiology, St Elizabeth Hospital). Drs Hiatt, Creager, Cooke, Pearce, and Isner are recipients of NIH Academic Awards in Vascular Disease.

1. Criqui MH, Fronek A, Barrett-Connor E, Klauber MR, Gabriel S, Goodman D. The prevalence of peripheral arterial disease in a defined population. Circulation. 1985;71:510-515. [Abstract]
2. Kannel WB, McGee DL. Update on some epidemiologic features of intermittent claudication: the Framingham Study. J Am Geriatr Soc. 1985;33:13-18. [Medline]
3. Regensteiner JG, Steiner JF, Panzer RJ, Hiatt WR. Evaluation of walking impairment by questionnaire in patients with peripheral arterial disease. J Vasc Med Biol. 1990;2:142-152.
4. Olsen PS, Gustafsen J, Rasmussen L, Lorentzen JE. Long-term results after arterial surgery for arteriosclerosis of the lower limbs in young adults. Eur J Vasc Surg. 1988;2:15-18. [Medline]
5. Criqui MH, Langer RD, Fronek A, Feigelson HS, Klauber MR, McCann TJ, Browner D. Mortality over a period of 10 years in patients with peripheral arterial disease. N Engl J Med. 1992;326:381-386. [Medline]
6. Pentecost MJ, Criqui MH, Dorros G, Goldstone J, Johnston KW, Martin EC, Ring EJ, Spies JB. Guidelines for peripheral percutaneous transluminal angioplasty of the abdominal aorta and lower extremity vessels. Circulation. 1994;89:511-531. [Medline]
7. Rutherford RB. Evaluation and selection of patients for vascular surgery. In: Rutherford RB, ed. Vascular Surgery. Philadelphia, Pa: WB Saunders Co; 1984:11-18.
8. Tunis SR, Bass EB, Steinberg EP. The use of angioplasty, bypass surgery, and amputation in the management of peripheral vascular disease. N Engl J Med. 1991;325:556-562.
9. Isner JM, Rosenfield K. Redefining the treatment of peripheral artery disease: role of percutaneous revascularization. Circulation. 1993;88:1534-1557.