Reliability of isokinetic measurements of hip muscle torque in young boys.

The Purpose of this study was to determine the reliability of measurements of hip muscle torque obtained with the Cybex (registered trademerk) II isokinetic dynamometer from healthy young boys. Twenty-nine healthy boys, aged 6 to 10 years, were tested two times at one- to two-week intervals. Torque of the hip flexors and extensory and the hip abductors and adductors was assessed at angular velocities of 30 degrees and 90 degrees/sec The torque generated by each muscle group at 30 degree/sec was similar to the torque generated at 90 degree/sec Intraclass correlation coefficients (ICCS) were used to deter mine test retest reliability for measurements of each of the muscle groups at both angular velocities The ICCs were higher for hip flexion and extension than they were for hip abduction and adduction. The highest ICC (ICC = .84) was found for hip extension at 90 degrees/sec The ICCs for hip abduction and adduction at both angular velocities were less than .60. Factors that possibly contributed to the relatively low reliability in hip abduction and adduction are discussed. [Burnett CN, Betts EF, King WM Reliability of isokinetic measurements of hip muscle torque in young boys. Phys Ther 70244-249, 19901 Key Words: Hip; Kinesiology/biomechanics, trunk; Muscle performance, measurement; Tests and measurements, functional In the early 1970s, Alexander and Molnar first reported the use of the Cybex registered trademark) dynamometer to measure the muscle strength of knee extensors and flexors and elbow extensors and flexors in children 7 to 15 years of age.1,2 In a subsequent study of children, Molnar et al reported intratest, intertest, and intertester variability in obtaining isokinetic torque measurements.3 The torque of the shoulder and hip flexors, extensors, and abductors; knee extensors and flexors; and elbow extensors and flexors were tested at an angular velocity of 30 degrees/sec. Differences between the measurements obtained 1) on three trials of one test, 2) on two tests conducted 7 to 10 days apart, and 3) on two tests conducted by two examiners for all muscle groups of each child were not statistically significant. The investigators concluded that isokinetic testing is a simple, easily applicable, and reliable way to determine muscular strength in children with normal intelligence, as well as in children with a mild degree of mental deficiency. Their results have been replicated, in part, by Gilliam et al.4 Most recently, Weltman et al reported the reliability of isokinetic measurements of torque obtained with the KIN-COM (registered trademark) dynamometer in prepubertal boys aged 6 to 11 years. 5 Test-retest reliability coefficients were calculated from dominant versus nondominant motions of knee, elbow, and shoulder extension and flexion at 30' and 90"/,sec angular velocities. The coefficients for knee and elbow motion and shoulder extension ranged from .72 to.91. Coefficients for shoulder flexion at the two speeds were <.70. Weltman et al concluded that the isokinetic measurements were reliable for all of the muscle groups they tested with the exception of the shoulder flexors.5 We considered it important to establish the reliability of measurements obtained with the Cybex (registered trademark) II isokinetic dynamometer* in children so that these measurements can be used for this population. Essentially no information is available on isokinetic testing as it relates to normal development. We have little idea at what age isokinetic performance is similar to that of the adult. We do not know whether the Cybex (registered trademark) II dynamometer can be used to obtain reliable measurements from healthy children. We also do not know the functional implications of isokinetic testing. Until recently, measurement of muscle strength has been primarily by manual muscle testing. Manual muscle testing, however, does not appear useful for determining strength at its upper limits. Several investigators have provided information on the use of various myometers in measuring muscle strength of healthy children and children with muscle weakness.6,7 These myometers have been used in assessing weakness at the low end of the force continuum and for measuring small muscle groups in healthy children. They have not been discriminating in early muscle loss because of the need of the therapist to provide appropriate resistance. For instance, in children with Duchenne muscular dystrophy (DMD), measurements of hip muscle strength are necessary because of the hip's important contribution to independent ambulation. Edwards et al have observed that a simple relationship does not exist between loss of muscle strength and function such as walking in children with DMD.8 They suggest that future studies examine critical changes in the stabilizing muscles of the pelvis throughout the entire course of the disease, requiring a more objective measuring tool. The purpose of this study was to determine the reliability of torque measurements obtained with the Cybex (registered trademark) II dynamometer by testing the hip muscles of healthy young boys. Method Subjects Twenty-nine healthy boys, aged 6 to 10 years, with no history of known hip pathology were tested two times at one- to two-week intervals. Eight subjects were 6 years old, 3 were 7 years old, 7 were 8 years old, 6 were 9 years old, and 5 were 10 years old. The subjects' height and weight were recorded prior to testing. One limb, selected by lower limb preference as establishing by ball kicking, was tested. The children willingly participated in the study, and a parent signed a consent form approved by The Ohio State University Biomedical Sciences Human Subject Review Committee. Procedure The Cybex (registered trademark) II isokinetic dynamometer was used to measure peak torque of the hip flexors, extensors, abductors, and adductors at angular velocities of 30 degrees and 90 degrees/sec. Calibration of the unit was performed at the beginning of each test day according to the guidelines printed in the Cybex (registered trademark) II manual. Throughout the testing, the recorder was run at a 5-mm/sec paper speed and the damping was maintained at a setting of 2. Positioning. The children were positioned supine to test the hip flexors and extensors. The axis of the dynamometer was aligned with the greater trochanter, and the short input adaptor was used so that the adaptor cuff could be positioned approximately midway between the hip and knee. The distance from the anterior superior iliac spine to the adaptor cuff was constant for each child's test-retest measurements. The pelvis and upper trunk were stabilized with straps, and the child grasped the corners of the table above his head to prevent him from sliding on the table. Each child was asked to submaximally flex and extend the hip two or three times through a 90-degree arc of movement to familiarize himself with the 30 degrees/sec test speed. After a brief rest, he was then asked to repeat the procedure giving his best effort through four consecutive repetitions. The testing procedure at 90 degrees/sec was repeated with one procedural difference. The child was asked to continue the 90 degrees/sec test past the four repetitions to the point of fatigue or until the torque dropped to 50% of the original effort. Testing of the slow velocity preceded testing of the fast velocity in each child. The sequencing of the angular velocity was not randomized because we followed the 30 degrees/sec test with the endurance test at 90 degrees/sec. Inconsistencies in the end point for the endurance test precluded our reporting this information. The children were positioned side lying to test the hip abductors and adductors. The axis of the dynamometer was aligned with the greater trochanter of the hip, and placement of the adaptor cuff midway between the hip and knee was maintained at the same distance from the greater trochanter for each of the two trials. The pelvis and lower leg were stabilized, and the child grasped the table extension pad with one hand. Four repetitions were completed through a 45degree arc of movement at both the 30 degrees and 90 degrees/sec velocities. The child was maintained in a side-lying position with the hip in a neutral extended position to minimize substitution during hip abduction by the hip flexor muscles. Angular velocity selection. The 30 degrees and 90 degrees/sec angular velocities were chosen based on some initial clinical data obtained from repeated measures of four children with DMD (C N Burnett, E F Betts; unpublished data; 1987). These children's hip flexion and extension were tested at 30 degrees, 60 degrees, 90 degrees, and 120 degrees/sec angular velocities. The children with DMD were unable to generate consistent measurable torque at 120 degrees/sec. We therefore elected to exclude the 120 degrees/sec angular velocity in testing the healthy children in our current study. Because the differential in torque at the remaining angular velocities (30 degrees, 60 degrees, and 90 degrees/sec) in the earlier testing of children with DMD was small (4 ft.lb t) and because we believe, based on our observations in our laboratory, that peak torques generated at 30 degrees and 60 degrees/sec angular velocities are similar, we selected 30 degrees and 90 degrees/sec angular velocities as the test velocities for the healthy children in this study. Damping selection. The Cybex Corporation suggests a damping of 3 when testing the hip so that the damping circuit in the recorder suppresses overshoot artifact. We chose to use a damping of 2 to obtain some effect of filtering artifact, yet minimize the damping effect on the suppression of muscular torque. Minimizing the damping effect on the suppression of muscular torque was important in our study because the torque generated by the children was small compared with that generated by adults.9 Data Analysis A mean peak torque value (in foot-pounds) was determined for each subject by averaging the peak torque at each of the four repetitions. The values were read directly from the strip chart recorder. Means and standard deviations for each muscle group at each angular velocity were calculated. A Student's paired t test was used to determine whether the torques at each velocity for the test and retest were significantly different. To determine reliability of the measurements, intraclass correlation coefficients (CCS) were used to test for agreement between the paired scores of each subject under every test condition.10 A single-factor, repeatedmeasures analysis of variance was performed to generate the variance components necessary to compute the ICCs (formula 1,1).11" Results Mean peak torque (MPT), standard deviations, and t-test values for each muscle group at each angular velocity are presented in Tables 1 and 2. No significant difference was found between the group means of test 1 and test 2 (p > .05). Intraclass correlation coefficients for each muscle group at each angular velocity varied between .49 for hip adduction at 90 degrees/sec and .84 for hip extension at 90 degrees/,sec (Tab. 3). The highest ICCS, and therefore the highest reliability, were demonstrated at 90 degrees/sec in hip flexion and extension followed by 30 degrees/sec in the same motions. Lower ICCs were computed for hip abduction and adduction than for hip flexion and extension at both angular velocities. Discussion Mean Peak Torque The mean range of torque values for hip flexors at 30 degrees and 90 degrees/sec angular velocities was 15 to 18 ft.lb. The mean range of torque for hip extensors at the two angular velocities was 25 to 28 ft.lb. The values for hip extensors are similar to the values Sockolov et al obtained during isokinetic testing of hip extension in 13 boys aged 6 to 11 years (approximately 20-30 ft.lb),12 The torque values we obtained for hip flexion at 30 degrees and 90 degrees/sec angular velocities were higher than the values reported by Sockolov et al (6-11 ft.lb). The angular velocity, positioning, and damping used by Sockolov et al were not reported; however, all of these variables play a role in the magnitude of torque values and may explain differences observed between the subjects in our study and the subjects Sockolov et al studied. Molnar and Alexander reported that hip musculature could be measured reliably by the Cybex (registered trademark) II dynamometer.2,3 However, they included no information on the positioning or techniques used in testing the hip muscles or the torque values obtained. The effect of gravity was a consideration in our study. Correcting for the effect of gravity in the entire limb is more difficult than correcting for the effect of gravity in the shank segment. In correcting for the effect of gravity in the lower limb, the individual is placed in a sitting position and the lower limb is extended and then allowed to flex passively. In correcting for the effect of gravity in the entire limb, not only the lower portion of the leg but also the thigh must passively drop to the table. Because the leg must be relaxed, knee motion may occur and interfere with the free fall of the lower extremity. We positioned the children supine with their hips as close to the end of the table as possible and still adequately supported their pelvis. The examiner (CNB or EFB) lifted the limb to 90 degrees of hip flexion and allowed the knee to assume a flexed position, then placed her hand at the anterior knee joint and allowed the leg to fall. The heel often softly touched the edge of the table. It was difficult for the children to completely relax the limb, and the gravitational torque (passive) obtained for most children in each age group was not reproducible. Gravity correction, however, was not essential to the purpose of our study, which was to determine the reliability of the maximum test-retest torque measurements obtained for each boy. The weight of the leg was assumed to be the same within the two-week test-retest period, and the test positioning was consistent. In studies where absolute torque values are important or where change in the size of the lower extremity is a factor, we recognize that a need exists to determine the role of gravity. The high standard deviations of the group means of test 1 and test 2 are, in part, related to the marked age spread of the subjects tested. Other factors, identified in the discussion section under clinical implications, also influenced the standard deviations. Reliability In our study, the ICCs were highest for hip extension, next highest for hip flexion, and lowest for hip abduction and adduction. These findings were true at both the 30 degrees and 90 degrees/sec angular velocities. Only the correlation coefficient for hip extension (.84) approached the standard for good reliability.13 Molnar et al reported little intratest, intertest, and intertester variability in testing several muscles, including hip abductors, extensors, and flexors.3 They did not separate these hip muscles from the other muscles studied, however, in the results or discussion sections of their article. The two highest ICCs indicated that 75% of the variance of hip flexion measurements at the 90'/sec angular velocity and 84% of the variance of hip extension measurements at 90'/ sec were due to true measurement variability between subjects. The remaining 25% and 16% of the variability of the hip flexors and hip extensors, respectively, is unexplained. The unexplained variability and therefore the reliability of measurement in hip abduction and adduction is lower than in hip flexion and extension. Even though the sample size in this study was small, we were interested in determining whether evidence exists that the muscle performance of younger children is less consistent than that of older children. When the ICCs for the 11 younger children (age 6-7 yr) were compared with those of the 18 older children (age 8-10 yr), we found that the ICCs of the younger children were substantially higher than those of the older children in hip flexion at 30 degrees/sec and in abduction and adduction at both angular velocities. in the younger children, the ICCs for hip abduction and adduction were also higher than those for hip flexion and extension. The older boys may have performed less consistently during hip abduction and adduction than the younger boys because of the greater mass of the lower extremity and, with the knee extended during abduction and adduction, the need to control a long lever. A larger sample should be tested to determine whether these findings can be substantiated. The performance of one 9-year-old subject (Subject 23) was markedly different from test I to test 2. He had been up most of the night during a campout the night before test 2, and he complained of fatigue during testing. When this subject's data were