Test-Retest reliability of isokinetic knee extension and flexion torque measurements in persons with spastic hemiparesis.

0

The purpose of this st y w to evaluate and compare t e isokinetic torque measurements in the involved and uninvolved knee musculature of 20 subjects with spastic bemiparesis An isokinetic dynamometer was used to measure maximal voluntary knee extension and flexion at 60' and 120/s. Peak torque P7) and average peak torque (APT) data were collected from five repetitions on two separate occasions. Average peak torque was defined as the mean of the PT values obtained during each of thefive repetitions. Spasticity was measured in the involved knee musculature prior to isokinetic testing using the Ashworth Scale. Pearson Product-Moment Correlation Coefficients and intraclass correlation coefficients ICCs) were high >.90) for both knees for PT and APT at both angular velocities. No clinically meaningful differences were found between the Pearson correlation coefficients and the ICCs of the involved versus the uninvolved knee for any testing conditions. We concluded that isokinetic evaluation of torque, as measured by PT and APT in subjects with spastic bemiparesis, can yield reliable results in both extremities [Tripp EJ, Harris SR. Test-retest reliability of isokinetic knee extension and flexion torque measurements in persons with spastic bemiparesis. Phys ther 199-1;71..390-396] Key Words: Hemiplegia, evaluation; Muscle performance, lower extremity,spasticity.

Muscular strength of paretic muscles in patients with central nervous system (CNS) lesions is difficult to quantify objectively. Muscular strength has been described as the maximal voluntary force produced during a movement when joint angle, limb velocity, and type of muscle contraction are defined.1 Spasticity, often a clinical manifestation of CNS lesions, should be considered when conducting strength testing with individuals who have CNS lesions. Lance defined spasticity as a

motor disorder characterized by a

velocity-dependent increase in the

tonic stretch reflexes ("muscle tone")

with exaggerated tendon jerks, resulting

from hyperexcitability of the

stretch reflex, as one component of the

upper motor neuron syndrome.2(p485) Clinical and research reports have documented that spasticity impedes volitional movement or interferes with force production.3- Spasticity levels can fluctuate in relationship to position changes, excessive voluntary effort, stress,3 room temperature, bladder fullness, and the presence of decubiti.9 The effect of spasticity on the consistency or reliability of muscle performance measurements has been questioned.3,10

Isokinetic dynamometry has been used to measure muscular performance in physical therapy clinical practice and in research laboratories since the late 1960s.11-13 Isokinetic evaluation of muscular force production requires the use of an electromechanical or hydraulic device to measure torque produced by a muscle at a fixed angular velocity of joint movement. Isokinetic testing and exercise training research has predominantly involved healthy individuals or patients with musculoskeletal injuries. The reliability of some isokinetic measurements of muscle performance has been documented for healthy subjects.1 17

Recently, several studies6,1-24 have used isokinetic dynamometry for muscle performance testing in persons with CNS lesions. Only two published studiesl8,25 have addressed test-retest reliability of isokinetic muscle performance measurements in subjects with CNS lesions. Annstrong and colleaguesl8 tested 10 patients with multiple sclerosis (MS) and 20 healthy control subjects with the Cybex II dynamometer.* High test-retest reliability r=.99) was reported for knee extensor and flexor peak torque data collected during one testing session at all velocities tested O', 70', 135', 190', 230@, and 275*/s). When tested across three sessions over an 11-week period, however, knee muscle peak torque measurements in 3 subjects with MS (who were unfamiliar with isokinetic testing) were found to be highly variable with large peak torque differences. The authors suggested that these differences may be due to a learning effect and to initial timidity in using the apparatus.

In a published abstract, Koziowski25 reported acceptable levels of test-retest reliability of isokinetic measurements in 11 subjects with chronic hemiplegia using the Cybex II dynamometer. The subjects performed maximal voluntary knee extension and flexion at 30' and 180'/s on two occasions separated by 3 days. Mean peak torque data were collected bilaterally during five repetitions at each velocity. Reliability coefficients (the specific type was not reported) ranged from .82 to .98 for the involved musculature and from .90 to .93 for the uninvolved musculature. A limitation of this study was the reduced amount of data for analysis resulting from the inability of several subjects to generate torque at 180'/s with the involved knee muscles.

Objective muscle performance or strength testing in persons with CNS lesions is important during rehabilitation for (1) monitoring changes in paresis, or weakness, that are likely to affect activities of daily living and functional skill performance and (2) assessing the efficacy of treatment interventions aimed at improving muscle strength or dynamic motor capacity. Some researchers have reported a relationship between isokinetic evaluation of muscle strength and certain functional skills in persons with hemiparesis secondary to stroke. Watkins et al2 suggested that isokinetic testing at the maximum speed at which a muscle can generate torque may serve as an indicator of a patient's functional capacity. The reliability of measurements obtained with the isokinetic dynamometer on patients with CNS lesions needs to be investigated before it can be considered an appropriate evaluation tool for testing this population. As with all clinical measurements, the reliability of isokinetic measurements must be examined for all types of patients for each joint and at each angular velocity tested.12

The purpose of this study was to evaluate and compare the test-retest reliability of isokinetic torque measurements in the involved spastic knee musculature and the uninvolved knee musculature of persons with CNS lesions. To promote homogeneity, the sample studied was limited to persons with spastic hemiparesis resulting from a unilateral intracranial lesion, secondary to either cerebrovascular accident (CVA) or traumatic brain injury (TBI). Because spasticity levels are known to fluctuate clinically3,9 and because spasticity is frequently cited as impeding volitional movement and interfering with force production,3the following hypothesis was developed: The test-retest reliability of isokinetic strength measurements would be lower in the involved spastic knee musculature than in the uninvolved knee musculature. Method Subjects The subjects were 20 volunteers (16 male, 4 female), ranging in age from 22 to 68 years -X=40.4, SD = 15.8). They were recruited primarily from outpatient rehabilitation clinics affiliated with the University of Wisconsin Hospital and Clinics and from a stroke support group in the Madison, Wis, area. Thirteen subjects had spastic hemiparesis secondary to CVA, and 7 subjects had spastic hemiparesis secondary to TBI. Ten subjects were right hemiparetic, and 10 subjects were left hemiparetic. Time from onset of lesion to testing ranged from 1.0 to 11.4 years X-=4.1, SD=2.8).

Inclusion criteria for the subjects were (1) a minimum of 6 months post-onset of cerebral lesion to minimize the chance of spontaneous motor recovery during the course of data collection; (2) lower-extremity involvement with evidence of motor dysfunction, spasticity in the knee extensor or flexor muscles (sitting or supine position) as measured with the Ashworth Scale (grades 1-4),26 and abnormal synergy as described by Brunnstrom; (3) at least Fair strength28 in the knee extensors and at least Poor strength, in the knee flexors; (4) active knee extension to at least 15 degrees from full O*) extension in the sitting position and active or active-assisted knee flexion to at least 90 degrees in the prone position; (5) ability to ambulate in the home environment with or without assistive devices (6) ability to communicate and follow testing instructions; and (7) no medical contraindications. Signed informed consent was obtained from each subject prior to testing. Instrumentation

The LIDO Active Isokinetic System' was used for testing. According to its manufacturer, the LIDO isokinetic dynamometer has features for automatic internal calibration, to control for overshoot and to correct for gravity. A computerized system is used for data collection, analysis, and storage. Procedure

To document the presence of spasticity in the involved extremities, knee flexor and extensor spasticity was graded by the same investigator (EJT) prior to isokinetic testing using the Ashworih Scale.26 The involved knee was passively extended and flexed in the order of slow to fast movements, two to four times, in both sitting and supine positions.

Ashworth Scale grades are as follows: O=no increase in tone; 1=slight increase in tone, giving a "catch" when the limb is moved in flexion or extension; 2=more marked increase in tone, but limb is easily flexed (or extended); 3=considerable increase in tone, and passive movement is difficult; and 4=limb is rigid in flexion or extension. Results of the spasticity measurements are presented in Table 1. To establish interrater reliability for the spasticity rating, a second rater performed the same assessment (sitting position only) on 13 of the subjects, for a total of 26 grades, during one of the two data-collection sessions. The second rater was an experienced physical therapist who was trained in the spasticity assessment and rating procedure. Neither rater was allowed to observe the other rater during testing. The results remained confidential until all spasticity data were collected.

The order of isokinetic testing (involved versus uninvolved knee) was randomly assigned. The isokinetic testing was performed at two angular velocities (ie, 60* and 120*/s), which were also randomized to reduce order effects.

Following the spasticity measurement, each subject was seated on the LIDO bench with the back support set at a 90-degree sitting angle. The mechanical axis of the input shaft was visually aligned with the axis of rotation of the subject's knee when resting at 90 degrees of flexion. The thigh stabilization pad was applied distally and secured firmly. The padded ankle cuff was applied just above the malleoli and secured by a self-adhesive strap. To avoid extraneous body movement, large self-adhesive straps were applied horizontally across the pelvis and diagonally across the trunk. For upper-extremity stabilization, the subjects grasped handgrips positioned lateral to their mid-thighs. Subjects who were unable to maintain an adequate grip placed their involved hand comfortably on their lap during both test and retest sessions.

Individual knee extension and flexion range-of-motion (ROM) limits were set, based on previously attained goniometric measurements and manual muscle testing of the involved knee for each subject. To maintain consistency, these same ROM limits were used for the second day of testing for each subject. Upper and lower ROM limits varied from 5 to 15 degrees from full O') extension and from 85 to 90 degrees of flexion. The gravity correction feature was eliminated to prevent cancellation of torque attributable to spastic restraint, which would likely occur as a result of this feature on the LIDO" device. The gravity-correction feature was eliminated by removing the extremity from the input shaft and allowing the shaft to swing freely through the ROM.

Following set-up, each subject performed approximately five submaximal repetitions of knee extension and flexion for each of the conditions prior to actual testing. These "warmup" trials also allowed for familiarization with the equipment and the testing procedure. After a rest period of 1.5 to 2.0 minutes, the actual testing was initiated. The length of the rest period was based on the subjects' readiness to proceed and their denial of fatigue.

Each subject performed five continuous maximal-effort repetitions of knee extension and flexion through the preselected ROM at both angular velocities, followed by identical repeat tests with the contralateral extremity. Subjects were instructed to "work as hard and as fast as you can" for both knee extension and flexion. During testing, verbal reinforcement was provided for each repetition (ie, kick up hard" for extension and "pull down hard" for flexion). The use of the visual feedback from the monitor was not encouraged because of the likelihood of visual impairment in some of our subjects. Rest periods of approximately 5 minutes separated the tests between angular velocities and between extremities. These periods were set arbitrarily, based on the subjects' denial of fatigue and readiness to proceed to the next part of the test. A retest of the entire procedure was performed with each subject 2 to 4 days later at approximately the same time of the day to minimize diurnal variation. Data Analysis Percentage of agreement and the weighted Kappa coefficient29 were used for determining interrater reliability of the spasticity ratings using the Ashworth Scale. The isokinetic data were collected using an IBM personal computer and the LIDOACT version 2.2) software program. Data included bilateral knee extensor and flexor peak torque (PT) and average peak torque (APT) from the five maximal voluntary repetitions for the two angular velocities tested for both trials. Peak torque was defined as the highest torque generated during any one of the five repetitions. Average peak torque was defined as the mean of the PT values obtained during each of the five repetitions. To determine test-retest reliability of the PT and APT measurements, Pearson ProductMoment Correlation Coefficients (r) and intraclass correlation coefficients (ICC) were used as reliability coefficients. The ICC formula 2,1) of Shrout and Fleiss was chosen to measure the degree of agreement between trials. The trials were considered random effects; therefore, the findings may be generalized to other trials for a particular subject. In addition, the means and standard deviations for the isokinetic data were computed. Results

For the interrater reliability of the spasticity measurements, there was perfect agreement on 13 (50%) of the 26 measurements using the Ashworth Scale.27 Agreement within one point was found for 86.6% of the measurements. There was no disagreement greater than two grades per muscle. The weighted Kappa value was .40.

For the involved knee isokinetic testing (extensors and flexors) at both angular velocities, the Pearson correlation coefficients ranged from .95 to .97 for PT and from .94 to .98 for APT. The corresponding ICCs were .91 to .97 for PT and .90 to .98 for APT (Tab. 2). For the 60'/s testing condition, the Pearson correlation coefficients ranged from .95 to .98. Similarly, for the 120'/s testing condition, Pearson correlation coefficients ranged from .94 to .97. The corresponding ICC values were .90 to .98 for 60*/s testing and .92 to .97 for 120*/s testing (Tab. 2).

The Pearson correlation coefficients for the uninvolved knee testing (extensors and flexors) at both angular velocities ranged from .94 to .98 for both PT and APT. The corresponding ICC values were .93 to .97 for PT and .92 to .97 for APT (Tab. 2). For the 60*/s testing condition, Pearson correlation coefficients ranged from .94 to .98. Similarly, for the 120*/s testing condition, Pearson correlation coefficients ranged from .95 to .98. The corresponding ICC values were .92 to .97 for 60*/s testing and .95 to .97 for 120*/s testing (Tab. 2). Pearson correlation coefficients and ICC values for isokinetic testing of individual extensor and flexor muscles for both knees are also shown in Table 2.

A visual analysis of the correlation coefficients (both Pearson correlation coefficients and ICCs) for the involved and uninvolved knee testing revealed that they were quite similar, with no clinically meaningful differences. Descriptive statistics for the isokinetic data are presented in Table 3. Discussion

The results of this study investigating the test-retest reliability of isokinetic torque measurements in persons with spastic hemiparesis secondary to CVA or TBI do not support the original hypothesis. On the contrary, high correlation coefficients (.90 and above) were found for both Pearson correlation coefficients and ICCs for both knees under all conditions of testing. No clinically meaningful differences were found between the Pearson correlation coefficients and the ICCs for the involved and uninvolved knees under any conditions of testing.

The correlation coefficients in this study are similar to those involving test-retest reliability on healthy subjects using the LIDO" Active Isokinetic System.17 Armstrong et al18 reported slightly higher coefficients (r=.99) in their isokinetic intrasession test-retest reliability study involving persons with MS and healthy subjects using the Cybex II dynamometer.

Our findings are also similar to those of Kozlowski, who tested subjects with chronic hemiplegia. A direct comparison between our findings and those of Kozlowski, however, would not be justified for the following reasons: (1) the small sample size (N=ll) in Kozlowski's study, (2) differences in the angular velocities used (30' and 180'/s in Kozlowski's study versus 60' and 120'/s in our study), (3) differences in the isokinetic dynamometers used Cybex II in Kozlowski's study versus LIDO in our study), and (4) possible differenccs in reliability coefficients used. Several of the subjects in Kozlowski's study were unable to generate torque at the higher velocity, which further limited the data for analysis. In our study, all subjects were able to generate torque at both angular velocities tested. There are several possible reasons for the high correlation coefficients obtained in our study. We used specific predetermined inclusion criteria. We used a standardized testing setup and protocol and controlled for order effects, fatigue, learning effects, and diurnal variation. Four of the subjects had previous experience with isokinetic testing and exercise. (The raw data for these subjects were highly consistent across trials.) In addition, the subjects were highly motivated and were genuinely interested in participating in the research. The majority of the subjects in this study had relatively low spasticity, especially in the sitting position, which was the test position. The supine position measurements were also comparatively low (Tab. 1). McLellan reported that in mild spasticity, stretch reflexes appear to be suppressed by voluntary effort. Therefore, low levels of spasticity may have a minimal effect on the consistency of torque production over time.

Another explanation for the high reliability coefficients may be that spasticity, despite its fluctuating nature, does not have a significant effect on isokinetic test-retest reliability. As Kozlowski argued,

It may be erroneous to assume that

measures of strength in hemiplegic 

subjects are not reliable secondary to

the influence of spasticity or abnormal

postural reflexes.25

The angular velocities of 60' and 120'/s used in our study were in the lower end of the available range of velocities for the LIDO isokinetic dynamometer. All subjects were able to produce measurable torque at these velocities. Several studies6,18,24,25 have reported an inability of subjects with CNS lesions to produce torque at high velocities. Spasticity interferes with movement to a greater extent at high velocities than at lower velocities.5,6 Perhaps the nature of isokinetic testing (constant velocity) at our lower test velocities minimized the influence of spasticity in significantly altering the consistency of torque production across sessions. It is important to consider that spasticity has been described as a velocity-dependent increase in tonic stretch reflexes. Despite the attainment of high correlation coefficients, visual analysis of the raw data revealed a few rather dramatic inconsistencies in the results across the 2 days of testing. Seven of the 160 retests involved torque diferences of 2:27 N-m (>20 ft-lb). For example, two subjects had -40.7 N*m (>30 ft-lb) of torque differences when tested at 60'/s, one for involved knee PT flexion (subject no. 7) and the other for uninvolved knee PT and APT extension (subject no. 10). Such differences could be considered clinically significant and would suggest that caution is needed when using isokinetic evaluation for persons with CNS lesions. Conversely, several subjects were impressively consistent across the 2 days of testing, with remarkably similar torque curves. Some of these subjects had no previous experience with isokinetic dynamometry.

A limitation of our study was that the interrater reliability of the spasticity measurement had only 50% perfect agreement. The weighted Kappa value of .40 represents fair agreement.32 These results may be explained by the fluctuating nature of spasticity, the alteration of tone from the passive movement testing, and possible errors in rater recording.

A second limitation was that only two angular velocities were used. Thus, further research is needed to investigate the test-retest reliability of other angular velocities that might be considered clinically important. Clinical implications

In light of controversies surrounding the use of isokinetic measurements for clinical evaluation and treatment planning,13 findings from our study suggest that isokinetic evaluation of muscle performance in patients with spastic hemiparesis can indeed yield reliable results. Reliable measurement of muscle performance in patients with CNS lesions is a prerequisite for studying the effectiveness of therapeutic interventions as well as for analyzing relationships between muscle performance and functional outcome criteria. The results of our study suggest that the isokinetic dynamometer provides reliable measurements of muscle performance in both the involved and uninvolved lower extremities of individuals with spastic hemiparesis. We recommend caution, however, when using isokinetic testing for patients with CNS lesions, based on the high variability found across time in the raw data for a few of our subjects. Replication studies are needed to verify our findings. Future research should include subjects with other types of CNS lesions and should involve a greater range of isokinetic velocities and other types of isokinetic equipment.

Rothstein and colleagues have recently criticized the paucity of credible scientific research" on the use of isokinetic measurements by physical therapists. We hope that the results of our reliability study will contribute to the clinical usefulness of isokinetic measurements for patients with spastic hemiparesis. Conclusions

Isokinetic measurement of knee torque, as measured by PT and APT in subjects with spastic hemiparesis secondary to CVA or TBI, can yield equally reliable results in both the involved and uninvolved extremities when using a IADO dynamometer. Considerations in attaining acceptable levels of reliability may include the adherence to proper setup and to a standardized testing protocol as well as the use of a lower range of angular velocities. The high level of motivation of our subjects and their relatively low levels of spasticity may also have contributed to our attainment of high reliability coefficients. Of clinical significance in our isokinetic torque data was the wide variability between testing sessions for a few of our subjects. Acknowledgments

We thank James Agre, PhD, MD, Betty Hasselkus, PhD, OTR, and Barbara Luedke, PT, for their assistance in this project. We also thank The University of Wisconsin's Department of Rehabilitation Medicine for allowing the use of the research laboratory for data collection. References

1 Knuttgen HG. Neuromuscular Mechanisms for therapeutic and Conditioning Exercise. Baltimore, Md: University Park Press; 1976:xi.

2 Lance JW. Symposium synopsis. In: Feldman RG, Young RR, Koella \VP, eds. Spasticity: Disordered Motor Control. Miami, Fla: Symposia Specialists Inc; 1979:485-494.

3 Bobath B. Adult Hemiplegia: Evaluation and Treatment. 2nd rev ed. London, England: William Heinemann Medical Books Ltd; 1978.

4 Carr JH, Shepherd RB. Pbysiotherapy in Disorders of the Brain: A Clinical Guide. Rockville, Md: Aspen Publishers Inc; 1980.

5 Sahrmann SA, Norton BJ. The relationship of voluntary movement to spasticity in the upper motor neuron syndrome. Ann Neurol. 1977;2:460-465.

6 Knutsson E, Martensson A. Dynamic motor capacity in spastic paresis and its relation to prime mover dysfunction, spastic reflexes and antagonist co-activation. Scand J Rehabil Med. 1980; 12:93-106.

7 Rosenfalck A, Andreassen S. Impaired regulation of force and firing pattern of single motor units in patients with spasticity. J Neurol Neurosurg Psychiatry. 1980;43:907-916.

8 Bohannon RW, Larkin PA, Smith MB, Horton MG. Relationship between static muscle strength deficits and spasticity in stroke patients with hemiparesis. Phys Ther. 1987; 67:1068-1071.

9 Roasenda JP, Ellwood PM. A review of the physiology, measurement and management of spasticity. Arch Phys Med Rehabil. 1961;42: 167-174.

10 Davies PM. Steps to Follow: A Guide to the Treatment of adult Hemiplegia. New York, NY: Springer-Verlag New York Inc; 1985:48-49.

11 Davies GJ. A Compendium of isokinetics in Clinical Usage and Rehabilitation Techniques, 2nd ed. La Crosse, Wis: S & S Publishing, Inc; 1984.

12 Mayhew TP, Rothstein JM. Measurement of muscle performance with instruments. In: Rothstein JM, ed. Measurement in physical threapy. New York, NY: Churchill Livingstone Inc; 1985;7:57-102.

13 Rothstein JM, Lamb RL, Mayhew TP. Clinical uses of isokinetic measurements: critical issues. Phys Ther. 1987;67:1840-1844.

14 Johnson J, Siegal D. Reliability of an isokinetic movement of the knee extensors. Research Quarterly. 1978;49:88-90.

15 Mawdsley RH, Knapik JJ. Comparison of isokinetic measurements with test repetitions. Phys Ther. 1982;62:169-172.

16 Tredinnick TJ, Duncan PW. Reliability of measurements of concentric and eccentric isokinetic loading. Phys ther. 1988;68:656-659.

17 McCrory MA, Aitkens SG, Avery CM, Bernauer EM. Reliability of concentric and eccentric measurements on the Lido Active Isokinetic Rehabilitation System. Med Sci Sports Exerc [Suppl]. 1989;21:52. Abstract.

18 Armstrong LE, Winant DM, Swasey PR, et al. Using isokinetic dynamometry to test ambulatory patients with multiple sclerosis. Phys ther. 1983;63:1274-1279.

19 Bohannon RW. Knee extension torque during repeated knee extension-flexion reversals and separated knee extension-flexion dyads. Phys Ther. 1985;65:1052-1054.

20 Chen W-Y, Pierson FM, Burnett CN. Force-time measurements of knee muscle functions of subjects with multiple sclerosis. Phys Ther. 1987;67:934-940.

21 Hamrin E, Eklund G, Hillgren AK, et al. Muscle strength and balance in post-stroke patients. Ups J Med Sci. 1982;87:11-26.

22 Nakamura R, Hosokawa S, Tsuji I. Relationship of muscle strength for knee extension to walking capacity in patients with spastic hemiparesis. Toboku j Exp Med. 1985; 145:335-340.

23 Nakamura R, Watanabe S, Handa T, Morohashi I. The relationship between walking speed and muscle strength for knee extension in hemiparetic stroke patients: a follow-up study. Tohoku j Exp Med. 1988; 1 54:111-113.

24 Watkins MP, Harris BA, Kozlowski BA. Isokinetic testing in patients with hemiparesis: a pilot study. Phys Ther. 1984;64:184-189

25 Kozlowski BA. Reliability of isokinetic torque generation in chronic hemiplegic subjects. Phys Ther. 1984;64:714. Abstract.

26 Ashworth B. Preliminary trial of carisoprodol in multiple sclerosis. Practitioner. 1964; 192:540-542.

27 Brunnstrom S. Movement threapy in Hemiplegia: A Neurophysiological Approach. New York, NY: Harper & Row, Publishers Inc; 1970.

28 Kendall FP, McCreary EK. Muscles.. Testing and Function. 3rd ed. Baltimore, Md: Williams & Wilkins; 1983.

29 Cohen J. Weighted kappa: nominal scale agreement with provision for scaled disagreement or partial credit. Psychol Bull. 1968; 70:213 220.

30 Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428.

31 McLellan DL. Co-contraction and stretch reflexes in spasticity during treatment with baclofen. J Neurol Neurosurg Psychiatry. 1977;40:30-38.

32 Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977;33:159-174.