Importance of Correcting Isokinetic Peak Torque for the Effect of Gravity when Calculating Knee Flexor to Extensor Muscle Ratios

Transcription

1 Importance of Correcting Isokinetic Peak Torque for the Effect of Gravity when Calculating Knee Flexor to Extensor Muscle Ratios MICHAEL FILLYAW, THOMAS BEVI, and LISA FERNANDEZ The purpose of our investigation was to compare, for the hamstring and quadriceps femoris muscles, peak torque values uncorrected for gravity with the peak torque values corrected for gravity and to determine the effect of making this correction on the hamstring to quadriceps femoris muscle peak torque ratio at slow and fast isokinetic speeds. We measured peak torques isokinetically at 60 /sec (slow) and 20 /sec (fast) in 25 female university soccer players. The gravity effect torque (GET) is the torque resulting from the effect of gravity on the combined weight of the leg and dynamometer arm at the precise angle of extension and flexion peak torque. The GET was added to the measured quadriceps femoris muscle peak torque and subtracted from the hamstring muscle peak torque to yield gravity corrected values. Failure to consider GET greatly underestimated quadriceps femoris muscle torque and overestimated hamstring muscle torque and the ratio between these torques at both speeds. Whereas the uncorrected hamstring to quadriceps femoris muscle peak torque ratio increased as speeds went from 60 /sec to 20 /sec, the gravity corrected ratio significantly decreased. Clinicians must remember the importance of making the gravity correction in patients with reduced torque output where the gravitational torque is a greater percentage of the measured torque to ascertain correctly the relative strength of antagonists inversely affected by gravity. Key Words: Gravitation, Knee, Muscle contraction. Achieving a balance in the relative strength of the knee flexor and extensor muscles is a recognized goal of rehabilitation,,2 athletic conditioning, 3, and injury prevention. 5,6 Development of the Cybex II isokinetic dynamometer* has enabled researchers to study the relationship of knee flexor to extensor muscle torque dynamically at preselected angular velocities. One way to look at this relationship is to examine the hamstring to quadriceps femoris muscle peak torque ratio. The hamstring to quadriceps femoris torque ratio has been reported Mr. Fillyaw is Research Physical Therapist, Physical Therapy Department, University of Vermont, Burlington, VT 0505 (USA). Mr. Bevins is Lecturer, Physical Therapy Department, University of Vermont. Ms. Fernandez is Assistant Athletic Trainer, University of Vermont. This article was submitted June 29, 98; was with the authors for revision 26 weeks; and was accepted May 28, 985. *Cybex, Div of Lumex, Inc, 200 Smithtown Ave, Ronkonkoma, NY 779. for groups of high school 7,8 and professional 9 football players, postoperative knee cases, and nonathletic adults. 0-2 An interesting finding common to all but one of these studies is the increase in the hamstring to quadriceps femoris torque ratio from slow to fast movement speeds. Torque decreases as the speed of a muscle contraction increases.,-5 This is true of both antagonist muscles performing a reciprocal movement, such as knee extension and flexion. Therefore, for the hamstring to quadriceps femoris torque ratio to increase as speed increases, the decrease in quadriceps femoris torque must be larger than the concomitant decrease in hamstring torque. 7,0,6 Curiously, the investigators who have reported an increase in the hamstring to quadriceps femoris torque ratio at fast isokinetic speeds have not suggested any biomechanical, neurophysiological, or other mechanism to explain the increase. If, in fact, the ratio increases as movement speed increases, some morphological or functional differences between the muscles must account for the change. In the absence of a proposed mechanism or evidence of morphological differences, an alternative thesis may be advanced. Namely, that the hamstring to quadriceps femoris peak torque ratio may not increase at fast isokinetic speeds but is found to increase because of an error in methodology. This error is a failure to correct for the gravity effect torque (GET) resulting from limb weight when measuring peak torques of the hamstring and quadriceps femoris muscles. Correcting for the effect of gravity increases quadriceps femoris torque values and decreases hamstring torque values. 7,8 Richards has reported the hamstring to quadriceps femoris peak torque ratio at slow and fast isokinetic speeds using gravity corrected torque values. 2 The magnitude of the GET as a correction factor in calculating hamstring and quadriceps femoris peak torques and their ratio at varying isokinetic speeds, however, has not been investigated adequately. Clinically, to plan Volume 66 / Number, January

2 Fig.. Gravity effect torque (GET) (PAS SIVE) and combined GET and active muscular torque (ACTIVE). GETθ = FgCosθ, (r) is the maximal torque resulting from the force of gravity during the passive fall when the limb and input arm are θ degrees from the horizontal. FgCos θ 2 (r) + (contractile force) (r) is the torque recorded when the limb and input arm are θ 2 degrees from the horizontal. GETθ 2 = FgCos θ 2 (r) is the component of the recorded torque resulting from the gravity effect. correct preventive conditioning or postinjury rehabilitation programs, therapists must understand the true nature of the relationship between the hamstring and quadriceps femoris muscles and how this relationship is affected by changing isokinetic speeds. The purpose of our investigation was to compare, for the hamstring and quadriceps femoris muscles, peak torque values uncorrected for gravity with the peak torque values corrected for gravity and to determine the effect of gravity correction on the hamstring to quadriceps femoris peak torque ratio at 60 / sec and at 20%sec. We expected that when the GET was accounted for, the hamstring to quadriceps femoris peak torque ratio would not change at increasing isokinetic speeds. METHOD Subjects We tested 27 women on the university varsity soccer team as part of a preseason medical examination. Prior to testing, each subject gave informed consent. Two subjects who had a history of reconstructive knee surgery exhibited serious asymmetries in either their ipsilateral antagonist or contralateral homologous muscles and were excluded from the data analysis for this study. The remaining 25 subjects had no history of knee surgery and were not injured at the time of testing. The mean age of the women was 9 ±.22 years; the mean weight, ±.38 kg; and the mean height, ± 5.2 cm. Experimental Procedure The preseason medical examination isokinetic testing protocol required that both legs of each subject be tested isometrically (0 /sec) at 90 degrees of knee flexion and dynamically in a knee extension-flexion pattern at 60%sec and 20 / sec. Three to five warm-up trials were allowed at each test speed. Five trials, with at least 0 seconds of rest between each trial, were recorded at both test speeds. The order of testing was the same for each subject (isometric, 60 / sec, 20 /sec). We will report only the isokinetic data from the right leg. Paired / tests revealed no significant difference between left and right legs for any variable (p >.05) Subjects sat on the testing table with the knee at 90 degrees offlexionand the axis of rotation of the dynamometer aligned with the lateral femoral condyle. TABLE Gravity Corrected and Uncorrected Mean Peak Torque and Ratio at Two Isokinetic Speeds a Isokinetic Speed 60 /sec Uncorrected data Gravity corrected data 20 /sec Uncorrected data Gravity corrected data Hamstring Muscle 55.2 (± 9.25) 6.66 (± 8.97) 33.7 (±0.07) (± 9.) Quadriceps Femoris Muscle 82.9 (±6.73) 87.0 (±6.3) 2.5 (± 8.65) 9.57 (± 8.52) a Mean peak torque in ft.lb (standard deviations in parentheses). Ratio (Hamstring to Quadriceps).67 (±.3).5 (±.0).79 (±.23).5 (±.9) Standard stabilization straps were used across the thigh and pelvis to minimize extraneous hip motions. Subjects grasped the side of the seat and kept the trunk in contact with the backrest during testing. Subjects were instructed to kick and bend the leg as hard and as fast as possible through a complete range of motion. Verbal encouragement was given during every trial. Data Reduction and Analysis Torque in foot-pounds was recorded on the graph of the dual channel recorder using the 80 ft. lb scale at a damping setting of 2. The torque at the peak point in each extensor and flexor muscle torque curve (peak torque) and the corresponding angle (angle of peak torque) were measured manually using a Cybex II data chart card. These torque values taken directly from the graph paper were uncorrected for the effect of gravity. Gravity corrected hamstring and quadriceps femoris peak torque values were determined as recommended by Nelson and Duncan. 7 (Appendix, Fig. ). We chose this accurate and clinically feasible method in favor of the alternate method suggested by Winter et al 8 or the formula published by Lumex, Inc, 9 each of which has disadvantages. The technique of Winter et al would be difficult to apply in the clinic because it requires an accelerometer, a piece of equipment not common to a physical therapy department. According to J. Seederger (June 98), Cybex recognizes that the formula for gravity correction published in the Cybex II handbook is incorrect. Use of this formula does not result in a zero GET as it should when the force arm is vertical. The GET is the torque produced by the combined weight of the limb and dynamometer arm at the precise angle at which peak torque occurred during each knee extension and flexion movement. Because the angles of peak torque for hamstring and quadriceps femoris muscles varied slightly from trial to trial, we calculated the GET for each trial. Hence, gravity corrected hamstring and quadriceps femoris peak torque values were determined for allfivetrials at each isokinetic speed and used in the data analysis. ft. lb=.356 N. m. 2 PHYSICAL THERAPY

3 RESEARCH We used an analysis of variance to study the repeated measures design involving the three factors (method, speed, trial). In this design, the error term for each significance test is the interaction of that effect with the subject effect. 20 We performed follow-up testing using Duncan's multiple range test to detect differences between individual means. RESULTS Table presents the means and standard deviations of the uncorrected and corrected peak torque and ratio data at each isokinetic speed. The difference between the uncorrected and corrected data is the GET. Figures 2 and 3 plot the data for each variable from Table. Table 2 summarizes the analysis of variance for each variable. Correcting for gravity yielded increases in mean quadriceps femoris peak torque of.5 ft.lb at 60%sec and of 7 ft.lb at 20 / sec over the respective uncorrected torque values. Gravity correction yielded about an 8 ft.lb difference in hamstring peak torque at each speed. This is reflected in the highly significant method main effect for each variable. The effect of the gravity correction on the hamstring to quadriceps femoris peak torque ratio as isokinetic speed increased is revealed by the highly significant interaction of method and speed for the ratio variable (Tab. 2, Fig. 3). When calculated with uncorrected peak torque data, the ratio increased from.67 at 60 /sec to.79 at 20 /sec, but decreased from.5 at the slow speed to.5 at the fast speed when calculated with gravity corrected torque values (Tab. ). A Duncan's multiple range test performed on the ratio means in Table showed that all differences between means were significant (p <.0). The highly significant speed main effect (Tab. 2) for both hamstring and quadriceps femoris peak torques reflected the large torque decreases that occurred in each muscle group at the fast isokinetic speed regardless of the method used to determine peak torque. The effect of measurement over five trials was analyzed as a separate source of variation. The trial main effect was not significant for any variable. The interaction between speed and trial for hamstring peak torque was significant. BMDP8V, Department of Biomathematics, University of California, Los Angeles, CA Fig. 2. Effect of gravity correction on peak torque. Gravity corrected and uncorrected peak torques at each isokinetic speed. Hamstring muscle peak torque decreased an equal amount from 60 /sec to 20 /sec if corrected for gravity or not. Uncorrected quadriceps femoris muscle peak torque decreased more than the gravity corrected torque at the fast speed. Key: x = corrected quadriceps femoris, o = uncorrected quadriceps femoris, = corrected hamstring, A = uncorrected hamstring muscles. The linear effect of trials was examined for each speed using orthogonal contrast coefficients. At 60%sec, peak torque increased significantly from Trial ( = 9.72 ft.lb) to Trial 5 ( = ft.lb) (F=9.; df=,; p<.0). At 20 / sec, peak torque decreased significantly from Trial ( =30.9 ft.lb) to Trial 5( = ft.lb)(f= 5.09;df=,; p<.0). DISCUSSION When the knee extension-flexion movement is performed on the Cybex dynamometer, knee flexion is aided by gravity and knee extension is opposed by it. Even in the absence of an active hamstring muscle contraction, a flexor muscle torque curve is produced by the combined weight of the lower leg and dynamometer arm passively falling through a 90 arc from full extension to flexion. Clearly, this torque is a component of any flexor torque curve produced by an active hamstring contraction during isokinetic testing. Conversely, a quadriceps femoris contraction may raise the limb and dynamometer arm through a full range of knee extension but register no extensor torque curve if the preset angular velocity is not reached. Obviously, torque is produced by the quadriceps femoris muscle, but in this case is just not measurable by the Cybex. Indeed, the extensor torque necessary to raise the Volume 66 / Number, January

4 Fig. 3. Effect of gravity correction on hamstring to quadriceps femoris muscle ratio. The hamstring to quadriceps femoris muscle torque ratio increased when calculated with uncorrected torque values but decreased with the gravity corrected torque values. Key: x = uncorrected, o = corrected. combined weight of the leg and dynamometer arm is never a component of the knee extensor torque curve produced during isokinetic testing. Simply stated, theflexortorque curve is inflated and the extensor torque curve is deflated by an amount equal to the GET. This results in observed torque values that underestimate actual quadriceps femoris torque and overestimate actual hamstring torque. The magnitude of this effect increases as the speed of contraction increases because the joint torque a muscle can generate by means of muscle contraction decreases as speed of contraction increases while the limb weight remains constant.,-5 Therefore, at fast isokinetic speeds, the GET becomes a larger percentage of observed torque. Uncorrected for GET at fast isokinetic speeds, quadriceps femoris torque will decrease more than hamstring torque. Several investigators have reported this, 7,0,6 and the results of our study confirm thisfinding.uncorrected quadriceps femoris peak torque decreased by 8% and uncorrected hamstring peak torque decreased by only 0% at the fast isokinetic speed. This resulted in the significant increase from.67 to.79 in the hamstring to quadriceps femoris ratio (Tab. ). Predictably, if peak torque measures are not corrected for the gravitational torque resulting from limb weight, increasing isokinetic speed will result in larger hamstring to quadriceps femoris ratios. Correcting for gravitational torque in measuring hamstring and quadriceps femoris peak torque yielded different results. Increasing isokinetic speed produced a 3% decrease in corrected quadriceps femoris peak torque, but a 6% decrease in corrected hamstring peak torque. This resulted in a significant decrease from.5 to.5 in the hamstring to quadriceps femoris ratio. Because limb weight is a constant, unaffected by changes in movement speed, only a change in the angle of peak torque could account for different gravity effect torques at the two speeds. Indeed, a significant decrease in the angle of quadriceps femoris peak torque occurred at the fast isokinetic speed. At 20 /sec, the angle of quadriceps femoris peak torque was 7 degrees compared with 6 degrees at 60 /sec (paired t= 0.5,df= 2,p<.00).TheGET is larger at the smaller angle (faster speed) because of the longer moment arm as the limb approaches horizontal. Thus, although quadriceps femoris peak torque decreased at the fast speed when calculated by both gravity corrected and uncorrected methods, the decrease was larger when using the uncorrected method. The uncorrected quadriceps femoris peak torque decreased by 0 ft. lb from 60 /sec to 20%ec (Tab. ). The decrease was only 37.5 ft.lb in the gravity corrected torque values. The resulting 2.5 ft.lb difference was of sufficient magnitude to produce the highly significant method by speed interaction for the quadriceps femoris muscle (Tab. 2, Fig. 2). Although the angle of hamstring peak torque increased from 33 degrees at 60 /sec to 37 degrees at 20 / sec, evidently this difference was not large enough to affect significantly the moment arm. Similar decreases (2.95 to 2.33 ft.lb) occurred in hamstring peak torque from the slow to fast speed regardless of which method, gravity uncorrected or corrected, was used to determine peak torque resulting in the insignificant method by speed interaction for the hamstring peak torque (Tab. 2). Table demonstrates that, although both hamstring and quadriceps femoris corrected peak torque values decreased as speed increased, the magnitude of the decrease was greater for the hamstring muscle. Thus, the ratio between gravity corrected hamstring peak torque and the gravity corrected quadriceps femoris peak torque decreased as speed increased. In contrast to our results, Schlinkman 2 and Richards 2 corrected for the effect of gravity when measuring the hamstring to quadriceps femoris peak torque ratio at varying isokinetic speeds but found an increase in the ratio at increased speed. Schlinkman tested 32 high school football players for knee muscle imbalance on the Cybex II. 2 A Cybex data reduction computer was used to determine gravity corrected hamstring to quadriceps femoris peak torque ratios at three isokinetic speeds. Although statistical significance is not reported, the ratio increased from.5 at 60 /sec to.66 at 200 /sec and to.67 at 300 /sec. Richards corrected for the effect of gravity when measuring the hamstring to quadriceps femoris ratio at two isokinetic speeds in women. 2 She re- 26 PHYSICAL THERAPY

5 ported a significant increase in the gravity corrected hamstring to quadriceps femoris peak torque ratio from.7 at 30 /sec to.60 at 80 /sec. The disparity between the results of Richards and Schlinkman and the results of our study may be because of differences in test subjects or isokinetic speeds. Richards's female subjects were older (mean age, 38. years), more heterogeneous with respect to age, and had a broad cross section of levels of physical activity. Our subjects were younger, trained athletes. Testing at different isokinetic speeds results in different angles of peak torque. Indeed, we found the angle of quadriceps femoris peak torque significantly decreased from 6 degrees to 7 degrees (0 = full extension) and the angle of hamstring peak torque significantly increased from 33 degrees to 37 degrees from 60 /sec to 20%ec. RESEARCH Richards reported angles of peak torque for the quadriceps femoris muscle to be 72 degrees at 30%sec and 5 degrees at 80 /sec and the angle of the hamstring peak torque to be 8 degrees at both speeds. Because calculation of the GET is a function of the angle at which peak torque occurs during knee motion, comparisons of hamstring to quadriceps femoris ratios between studies that report different angles of peak torque are TABLE 2 Analysis of Variance Summary Tables a Variable Source Quadriceps Femoris Muscle Peak Torque (ft.lb) Subjects (P) Method (M) Error (MxP) Speed (S) Error (SxP) Trial (T) Error (TxP) Method by Speed Error (MxSxP) Method by Trial Error (MxTxP) Speed by Trial Error (SxTxP) Method by Speed by Trial Error (MxSxTxP) Hamstring Muscle Peak Torque (ft.lb) Subjects (P) Method (M) Error (MxP) Speed (S) Error (SxP) Trial (T) Error (TxP) Method by Speed Error (MxSxP) Method by Trial Error (MxTxP) Speed by Trial Error (SxTxP) Method by Speed by Trial Error (MxSxTxP) Ratio: HamstringiQuadriceps Subjects (P) Method (M) Error (MxP) Speed (S) Error (SxP) Trial (T) Error (TxP) Method by Speed Error (MxSxP) Method by Trial Error (MxTxP) Speed by Trial Error (SxTxP) Method by Speed by Trial Error (MxSxTxP) SS df MS F NT b NT NT P a Each source of variation is tested by its interaction with subject. b NT = No valid test of significance. Volume 66 / Number, January

6 ineffectual. Schlinkman did not report the angles of peak torque for either muscle group but did test at 60%ec as we did. 2 Interestingly, identical ratios of.5 resulted in his football players and the soccer players when peak torques were measured at the same isokinetic speed. Future isokinetic research of the effect of speed on the ratio of hamstring to quadriceps femoris torques should be conducted using the same speeds so that valid comparisons can be made. Differences in fiber type of the quadriceps femoris muscle of the different subject samples also may account in part for the disparate findings between our study and those of Richards and Schlinkman. 2,2 Thorstensson et al reported a significant positive correlation between fast twitch muscle fiber in the vastus lateralis muscle and peak torque of knee extensors at 80 /sec. They concluded that a high percentage of fast twitch fibers is one prerequisite for performing fast contractions with appreciable torque outputs. Gregor et al reported significant positive correlations between percentage of total fast twitch fiber area in the vastus lateralis muscle of female athletes and maximal torque at isokinetic speeds of /sec and faster. 22 A high percentage of fast twitch fibers in the quadriceps femoris muscle would put it at a physiologic advantage and lead to relatively smaller decreases in extensor torque than in flexor torque at fast isokinetic speeds. This would be reflected as a decrease in the hamstring to quadriceps femoris ratio at the fast speed as reported. Even in the absence of morphological differences because of muscle fiber type, we postulate that athletes involved in a kicking sport might have an advantage in the performance of knee extensionflexion on the Cybex at fast speeds. Athletes are accustomed to the maximum effort demanded by isokinetic testing. In particular, soccer players who are practiced in developing maximum torque and limb velocity to execute a soccer kick may have a training bias favoring their ability to develop large extensor torques at fast movement speeds. Because each subject was tested at 600 /secfirstfollowed by 20 /sec, we cannot rule out the possibility that test order may have introduced a systematic error and influenced our results. Thorstensson et al reported no significant differences in knee extensor torques between a group performing first isometric and then isokinetic contractions from slow to fast speeds and a group performing the same number of contractions in a randomized order. Costain and Williams used different testing sequences to reduce the possibility of error resulting from fatigue or learning but did not report if a statistical analysis was done to test for the effect order of isokinetic speed had on torque output. 23 This is an issue that has received little attention in the isokinetic literature and yet may have important ramifications for both research and clinical testing. Because peak torque and angle of peak torque vary from trial to trial, we chose to use all five trials so that each ratio was calculated between truly antagonistic flexor and extensor torques. If we had chosen the largest torque and its respective angle from among all trials, as has been reported, 23 ' 2 we might unwittingly have affected the hamstring to quadriceps femoris peak torque ratio. By including all five trials in the analysis, we prevented mismatching quadriceps femoris peak torque with a hamstring peak torque from a different trial. In the analysis, therefore, trial is shown as a separate source of variation. Differences among trials were not significant for any variable, which indicates no significant improvement or decrement over trials (Tab. 2). The only significant effect related to trials was the speed by trial interaction for hamstring peak torque. Examination of individual trial means at each speed revealed that, at the slow speed, hamstring peak torque increased from Trial to Trial 5 and, at the fast speed, the largest torque occurred on Trial and decreased thereafter to Trial 5. Although each trend was significant, we cannot account for this finding. Even though the method for calculating the GET published by Lumex, Inc, is incorrect, the magnitude of error is not great. For our data, their method overestimates the GET for the quadriceps femoris muscle by about.5 ft.lb, underestimates the GET for hamstring muscles by about 0.8 ft.b, and underestimates the ratio by about.025. Nevertheless, we prefer the Nelson and Duncan method for its accuracy and simplicity. 7 Our results suggest that the GET is of sufficient magnitude to change significantly the peak torque measurements of both the hamstring and quadriceps femoris muscles. Failure to consider the GET seriously underestimates actual quadriceps femoris torque and overestimates hamstring torque at both slow and fast isokinetic speeds. This leads to incorrect assessment of the muscles' true ability to develop torque during an isokinetic movement. Herein lies the major motivation for making the gravity correction. When uncorrected torque values are used to calculate the ratio of peak torques from antagonist muscles inversely affected by gravity, specious ratios result. The problem is magnified at faster isokinetic speeds. Clinically, the influence of the gravitational torque in patients with reduced torque output would be even greater because the gravitational torque is a larger percentage of the measured torque. We believe that failure to correct for the GET resulting from limb weight when measuring hamstring and quadriceps femoris torque has led to erroneous conclusions about the relative strength of these muscles at all isokinetic speeds. CONCLUSION The results of our study demonstrate the need to make the correction for the effect of gravity when measuring hamstring and quadriceps femoris torque isokinetically in an antigravity position. This is necessary not only to obtain valid measurements of muscle torque but, even more important, to ascertain the relative strength of antagonists inversely affected by gravity. The error in not correcting for the effect of gravity becomes magnified at faster isokinetic speeds where absolute muscle torque values are small and the proportion of GET to the measured muscle torque increases. Whereas the hamstring to quadriceps femoris ratio has been reported to increase at fast isokinetic speeds, we found a significant decrease in the ratio after we corrected for the effect of gravity. Additional research is needed to replicate the findings of our study and to continue to investigate the relationship of hamstring to quadriceps femoris muscles in other groups of subjects, including men, the elderly, and untrained individuals. 28 PHYSICAL THERAPY

7 RESEARCH REFERENCES. Campbell DE, Glenn W: Rehabilitation of knee flexor and knee extensor muscle strength in patients with meniscectomies, ligamentous repairs, and chondromalacia. Phys Ther 62:0-5, Goslin BR, Charteris J: Isokinetic dynamometry: Normative data for clinical use in lower extremity (knee) cases. Scand J Rehabil Med :05-09, Laird D: Comparison of quadriceps to hamstring strength ratios of an intercollegiate soccer team. Athletic Training 6:66-67,98. Slagle G: The importance of pre-testing the knee joint. Athletic Training : , Hunter S, Cain T: Preseason isokinetic knee evaluation of professional football athletes. Athletic Training : , Burkett LN: Causative factors in hamstring strain. Med Sci Sports Exerc 2:39-2, Gilliam TB, Sady SP, Freedson PS, et al: Isokinetic torque levels for high school football players. Arch Phys Med Rehabil 60:0-, Parker MG, Holt D, Bauman E, et al: Descriptive analysis of bilateral quadriceps and hamstring muscle torque in high school football players. Abstract. Med Sci Sports Exerc :52, Davies GJ, Kirkendall DT, Leigh DH, et al: Isokinetic characteristics of professional football players:. Normative relationships between quadriceps and hamstring muscle groups and relative to body weight. Abstract. Med Sci Sports Exerc 3:76-77,98 0. Wyatt MP, Edwards AM: Comparison of quadriceps and hamstring torque values during isokinetic exercise. Journal of Orthopaedic and Sports Physical Therapy 3:8-56,98. Scudder G: Torque curves produced at the knee during isometric and isokinetic exercise. Arch Phys Med Rehabil 6:68-72, Richards D: Dynamic strength characteristics during isokinetic knee movements in healthy women. Physiotherapy Canada 33:-9, Moffroid M, Whipple R, Hofkosh J, et al: A study of isokinetic exercise. Phys Ther 9:735-77, 9. Thorstensson A, Grimby G, Karlsson J: Forcevelocity relations and fiber composition in human knee extensor muscles. J Appl Physiol 0:2-6, Osternig LR, Bates BT, James SL: Isokinetic and isometric torque force relationships. Arch Phys Med Rehabil 58:25-257, Rankin JM, Thompson CB, Armstrong W: Isokinetic evaluation of men and women intercollegiate athletes. Abstract. Med Sci Sports Exerc :6, Nelson SG, Duncan PW: Correction of isokinetic and isometric torque recordings for the effects of gravity. Phys Ther 63:67-676, Winter DA, Wills RP, Orr GW: Errors in the use of isokinetic dynamometers. Eur J Applied Physiol 6:397-08, Isolated Joint Testing: A Handbook for Using Cybex II and the UBXT. Bay Shore, NY, Lumex, Inc, Howell DC: Statistical Methods for Psychology. Boston, MA, PWS Publishers, Schlinkman B: Norms for high school football players derived from Cybex data reduction computer. Journal of Orthopaedic and Sports Physical Therapy 5:23-25, Gregor RJ, Edgerton VR, Perrine JJ, et al: Torque velocity relationships and muscle fiber composition in elite female athletes. J Appl Physiol 7: , Costain R, Williams AK: Isokinetic quadriceps and hamstring torque levels of adolescent, female soccer players. Journal of Orthopaedic and Sports Physical Therapy 5:-200,98 2. Holmes JR, Alderink GJ: Isokinetic strength characteristics of the quadriceps femoris and hamstring muscles in high school students. Phys Ther 6:9-98,98 APPENDIX Method to Calculate Gravity Corrected Peak Torque. The lower limb is positioned exactly as for testing with all recorder and goniometer adjustments made as usual. 2. The speed is set at 30 /sec, the recorder to 30 ft lb scale and 25 mm/sec chart speed. 3. The limb and input accessories are "weighed" by allowing the limb to fall passively from full knee extension (0 ) through an arc of 90 degrees knee flexion.. Peak torque (GETθ ) and the angle of peak torque (θ ) are measured from the resultant torque curve and position angle channel, respectively. 5. The gravity effect torque (GET) can be calculated now at any angle of knee flexion (θ 2 ) using the formula adapted from Nelson and Duncan. 7 The angle chosen (θ 2 ) will be the angle at which peak torque occurs for the muscle being tested. Because peak torque occurs at different angles for the hamstring and quadriceps femoris muscles, a separate GET is calculated for each. GETθ (cosθ2) 2 ) GETθ 2 = COS0 COSθ 6. Each GETθ GET0 2quads is added to the respective uncorrected quadriceps femoris muscle peak torque, and the GETθ 2 hams is subtracted from its uncorrected hamstring muscle peak torque value to yield the gravity corrected hamstring and quadriceps femoris peak torque values. Commentary Fillyaw et al, in this article, and Nelson and Duncan in a recent article, should be commended for enlightening the clinician-researcher on this basic measurement error in isokinetic dynamometer testing in antigravity positions. In determining the strength of hamstring and quadriceps femoris muscle groups, the authors selected the most common isokinetic assessment area of the body. Fillyaw et al have identified the effects of gravity in the recorded torque values at slow (60 /sec) and fast (20 /sec) velocities for the two muscle groups. This gravitational influence was termed gravity effect torque (GET), which was calculated by considering the combined weight of the dynamometer arm and the leg and foot. The protocol Fillyaw et al used for determining the gravity correction factor was based on the method proposed by Nelson and Duncan. This procedure requires the subject to relax completely the knee extensor mechanism while the leg accelerates to a preset terminal velocity during a 90 degree range of motion. To ascertain complete relaxation of the extensor muscle group, I would recommend electromyograph monitoring, whenever possible, of the muscle group to ensure total relaxation. The clinically treated patient with a knee problem, for example, may have difficulty in willingly relaxing the muscles during this procedure. Thus, a number of trials may be necessary before a valid, reliable measure has been determined. Fillyaw et al found a significant difference between peak torque values uncorrected and corrected for gravity at each velocity using the GET calculation. The quadriceps femoris muscle peak torque value increased, whereas the Volume 66 / Number January