Changes in torque and electromyographic activity of the quadriceps femoris muscles following isometric training.

Key Words: Electromyography; Exercise, general; Functional training and activities, Lower extremity, knee; Muscle performance, lower extremity.

Isometric contractions are used for strengthening exercises in the rehabilitation of orthopedic and sports-related injuries.[1-5] Following acute injuries or during immediate postoperative periods, isometric exercises are often used to increase strength.[1] Isometric exercise is also used in the rehabilitation of individuals with pathologies in which resisted movements through the full range of motion (ROM) are contraindicated because of crepitus, pain, effusion, and insufficient healing.[2-5] Because isometric exercises do not necessarily require equipment, they may be performed anywhere. We believe these exercises are appropriate for the early phase of strengthening programs for healthy individuals.[6,7]

Although support exists for the use of isometric training, the literature regarding the correct use of isometric exercise in rehabilitation is controversial. Conflicting reports exist as to whether isometric training performed at one angle will result in force gains only at the angle used for training (angular specificity)[8-10] or whether increases in force production may occur when measurements are taken at other angles.[11-13] In a review, Fleck and Schutt reported

. . . strength gained via isometric contraction

is very specific to the joint

angle at which training is performed.

Isometric training must be performed

at several points in the range of motion

of a joint if strength gains are

desired throughout the range of

motion.[1](p450)

Only a few investigations have examined whether training with isometric exercises increases force production at the joint angle at which the isometric exercises take place (angle specificity)[8-10] or whether gains occur when force is measured at other angles.[11-13]

Lindh[8] compared changes in torque of isometric knee extension by exercising 10 female subjects at two joint angles. Results indicated significant increases in isometric torque at the specific angle at which exercises took place, but no increase in torque at the other joint angle. Based on their research, the authors stated that "strength increase was mainly specific according to the angle at which the knee was exercised."[8](p33) Gardner[9] found that all subjects, except control subjects, had significant increases in isometric force when measurements were obtained at the training angle, but no improvement occurred at the other angles. Gardner concluded that "strength increases are quite specific to the position at which a limb is exercised."[9](p99) Belka[10] found that when wrist flexors were exercised with isometric contractions, improvements in isometric wrist flexion force were noted when measurements were taken at the angle at which subjects exercised, but no significant change in force occurred at other angles.

Meyers[11] reported data that refute the finding of angular specificity. Isometric elbow flexion force at angles of 90 and 170 degrees was measured for 33 subjects. Significant increases in force were found at joint angles of both 170 degrees (the joint angle of exercise) and 90 degrees (the joint angle at which no exercise took place). This increase in force at the 90-degree joint angle indicates that force increases could be found when the elbow was positioned up to 80 degrees away from the angle at which exercise took place. Knapik et al[12] studied 12 subjects and their responses to an isometric elbow extension exercise program. Results indicated significant increases in isometric torque at the angle of exercise, as well as other joint angles. The authors concluded that isometric training at one joint angle increases torque production with the elbow positioned 20 degrees from the training angle. These results indicate that training at 90 degrees of elbow flexion will result in an increase in torque production at that training angle, as well as through a ROM of 70 to 110 degrees.

More recently, Thepaut-Mathieu et al[13] compared three experimental groups that trained with isometric contractions at three different angles of elbow flexion (25 [degrees], 80 [degrees], and 120 [degrees]) with a control group. Results indicated that when the muscle was shorter (25 [degrees] of elbow flexion), the force gain was more limited to the training angle (reporting increases in force only 20 [degrees] away from the angle of exercise) and least specificity was observed in the group that trained with the muscle in the lengthened position (120 [degrees] of elbow flexion) (reporting increases in force at 75 [degrees] of elbow flexion away from the angle of exercise). The authors also reported that "an increase in maximal integrated EMG [electromyographic activity] accompanied the improvement of maximal voluntary contraction at the training angles."[13](p1500)

The increase in force production of muscles following isometric training has been reported to be the result of both muscle hypertrophy and the ability of the nervous system to recruit motor units.[14,15] Muscle hypertrophy, characterized by an increase in the size of individual muscle fibers, is a response to resistance training.[16-18] Some investigators,[19,20] however, have suggested that increases in force production may be dependent on the extent to which the muscle is activated by the nervous system. These changes in the nervous system following resistance training have been referred to in the early literature as "learning"[21,22] and more recently as "neural adaptation."[14]

Electromyographic (EMG) recordings of motor unit activity during maximal voluntary contractions before and after resistance training have been used to evaluate neural adaptation.[14] These recordings indicated that resistance-trained muscles exhibited recruitment of a greater number of motor units and a greater firing rate of each motor unit during a maximal contraction than did untrained or less trained muscles. Moritani and DeVries[19] and Hakkinen et al[20] concluded that the increased number of motor units firing at higher frequencies indicates increased activity of the muscle and, therefore, increased ability of the muscle to generate force. DeLuca et al[23] theorized that during contraction, a functional reserve of motor units appears to be present. These motor units are presumably not readily available, despite the effort at maximal voluntary contraction. Part of the increased motor unit activity following training may be related to learning to more fully activate some of those motor units that were not previously active. To date, no studies have examined the changes in neural adaptation at joint angles other than the specific joint angle at which the isometric exercises took place.

Specificity of training as related to isometric exercise and changes in force production at various joint angles is not well understood and requires further investigation. The influence of the length of the muscle at which isometric training takes place and how changes in muscle length affect force gains and neural adaptation has not been studied extensively.

The purpose of our study was to examine the effect of 8 weeks of isometric training of the quadriceps femoris muscles, at different joint angles, on changes in torque and EMG activity. More specifically, this study examined whether isometric training resulted in changes in torque and EMG activity only at the angle trained (angle specificity) or whether the training at one angle resulted in changes at other angles.

Three hypotheses were tested: (1) Isometric exercises of the quadriceps femoris muscles at a specific joint angle for 8 weeks would not increase torque or EMG activity at the same angle of exercise, (2) isometric exercises of quadriceps femoris muscles at a specific joint angle for 8 weeks would not increase torque or EMG activity at any other angles of knee flexion at which exercise did not take place, and (3) the position of the knee joint at which the 8 weeks of isometric exercise took place would not influence changes in torque and EMG activity.

Method

Subjects

One hundred seventeen female subjects between the ages of 20 and 36 years (X[bar]=23.8, SD=4.4), with no previous history of pathology of the knee, were recruited for this study. Subjects were volunteers and signed an institutionally approved informed consent form. Subjects not involved in any exercise activity at the start of this study agreed to avoid lower-extremity exercise and activities other than those prescribed by the research protocol. Subjects involved in aerobic exercise activity at the start of this study agreed not to increase the intensity or frequency of the activity during the 8 weeks of isometric training. Subjects involved in resistance training were not allowed to participate in this study.

One hundred seven subjects, with a mean age of 23.9 years (SD=4.3), completed the study. Three subjects were eliminated from the study as a result of injuries sustained in activities unrelated to the study, and 7 subjects were dropped as a result of noncompliance with the exercise program.

Equipment

Maximal isometric torque was measured using the leg extension apparatus of the Cybea[R] II dynamometer.(*) The dynamometer was calibrated weekly according to the manufacturer's instructions.[24]

A preamplifier bipolar electrode,(+) consisting of two silver/silver chloride electrodes and a solid-state preamplifier in a plastic enclosure, was used. The EMG signal was processed by the GCS 67 Multichannel Electromyographic System.(+) Collection, management, analysis, and reporting of the EMG data were controlled by the DATASAMP customized software package,(+) which allowed the user to specify the sampling rate and the length of the sampling interval. Signals were monitored for artifacts on a Tenmar oscilloscope.[double dogger] A standard universal goniometer was used to measure the angle of knee flexion.

Procedure

Instruction session. On their first day in the study, the subjects received instructions for all procedures, signed the informed consent form, and learned to perform a maximal isometric contraction. Each subject's dominant lower extremity was determined by asking the subject to kick a ball. The limb used to kick the ball was considered the dominant lower extremity. The limb not used to kick the ball was considered the nondominant lower extremity and was chosen for data collection.

The subjects practiced performing maximal isometric contractions at an angle of 45 degrees of knee flexion. The subjects were stabilized in the chair (seat angle set at 110 [degrees] of hip flexion) with Velcro[R] straps[subsection] placed at the chest, pelvis, and midthigh. The mechanical axis of rotation of the dynamometer[25] lever arm was aligned with the lateral joint line of the knee. The resistance pad at the end of the lever arm was strapped to the anterior tibia approximately 2.5 cm superior to the medial malleolus. The subjects were instructed to grab the sides of the padded seat for added stabilization when they performed contractions. The 45-degree angle of knee flexion was established using a manual goniometer, and the dynamometer was locked at this angle by setting the speed selector to 0 [degrees]/s.

Following a warm-up consisting of three submaximal isometric contractions, each subject was instructed to push (extend the knee) as smoothly as possible against the lever arm resistance pad and to gradually build up force until a maximal effort was exerted. During the isometric contraction, the investigator (WDB) verbally encouraged the subject to perform maximally: "Are you ready? Now push! Push harder! Harder! And relax." The contraction lasted approximately 4 seconds. The amount of torque produced during each contraction was graphed at a speed of 5 mm/s on the Cybex[R] strip chart recorder, which was set at a damping of 2.[25]

Each subject practiced maximal isometric knee extension contractions in sets of three, separated by 10 seconds of rest between contractions and 1 minute of rest between sets, until the following criteria were met: (1) Two of three consecutive contractions were within 10% of each other,[26] (2) the peak torques of these two contractions were maintained for no longer than 1 second,[27] and (3) the rise (build-up) times of these two contractions were no longer than 2 seconds.[28] Once subjects were able to produce a maximal isometric contraction meeting the criteria, they were scheduled to be pretested the next day.

Pretest. The actual pretest data (the maximal isometric torque and EMG activity of the extensor muscles of the knee during maximal isometric contraction) were obtained on the day following the successful instructional session. The skin was prepared by scrubbing the area with alcohol-soaked cotton to reduce electrical resistance. The distance between the medial epicondyle of the femur and the anterior superior iliac spine of the nondominant limb was measured with a tape measure, and the bipolar electrode was placed along this line one fourth of the distance proximal to the medial epicondyle. The ground electrode was placed 8.0 cm lateral to the bipolar electrode. The bipolar electrode and the ground electrode were attached to the appropriate sites using double-sided adhesive tape and remained in the same place throughout the data-collection session. Conductive gel[~~] was used to fill the holes in the tape over each electrode.

Following attachment of the electrodes, the subjects were positioned and stabilized on the dynamometer using the procedure previously described for the instruction session. Isometric torque and EMG data were collected from the quadriceps femoris muscles of the nondominant lower extremity at every 15-degree angle from 15 to 105 degrees of knee flexion (15 [degrees], 30 [degrees], 45 [degrees], 60 [degrees], 75 [degrees], 90 [degrees], and 105 [degrees]) for all subjects. The joint angle was measured by a manual goniometer. The order of testing the joint angles was randomized. Subjects were allowed three submaximal contractions at each joint angle prior to the first maximal voluntary isometric contraction at that angle. Three maximal isometric knee extensor contractions were then performed at the joint angle, with 10 seconds of rest allowed between contractions. One minute of rest was given between tests at each joint angle.

At the same time the subject was verbally instructed to push, a computer key was activated, which not only began the EMG data collection but also changed the color on the monitor screen, thereby providing each subject with a visual cue to contract with a maximal effort. During each contraction, the investigator provided the same verbal encouragement as was used during the instruction session. The contractions lasted for 4 seconds, during which time the torque was graphed and the EMG activity was recorded. The simultaneous muscle contraction and activation of the EMG recording device ensured synchronization of the time interval for later data derivation.

The EMG signal was recorded for a period of 4 seconds. The amplified (gain=2,000) and filtered EMG signal (time constant=2.5 milliseconds, bandwidth=40-4,000 Hz) was processed using the root-mean-square (RMS) procedure and a sampling rate of 1,500 samples per second.

The graph from the strip chart recorder was analyzed to determine the peak torque and time to peak torque. Using the computer software, the mean RMS was computed at a time interval of 0.5 second before and after the time to peak torque, providing a 1-second peak torque RMS. The peak torque at each angle and the EMG activity associated with this peak torque (1-second peak torque RMS) obtained during the maximal isometric knee extension contraction were used during data analysis.

We found no report in which normalization of EMG activity was performed in a training study. Therefore, the decision was made not to normalize the EMG signal. Given that previous research demonstrated that reliability of EMG recordings being removed and replaced could achieve an acceptable range of .80 and above without normalization,[29,30] a pilot study was performed to measure the reliability of the EMG measurement. Twelve subjects, 3 per group, exercised for 2 weeks using procedures described. Reliability testing of pretest to posttest EMG data was performed on the 3 control subjects using the most conservative intraclass correlation coefficient (ICC) available (formula 1,1)[31] and yielded an ICC of.92. We therefore proceeded with the experiment using the procedure established in the pilot study (no normalization), knowing that good reliability would have to occur in the control group for the EMG data to be relevant.

Training. Following pretesting, the subjects were randomly assigned to one of four groups. Group 1 (age range=20-30 years, X[bar]=23.9, SD=4.3) trained with isometric contractions with the knee flexed to 30 degrees, group 2 (age range=20-36 years, X[bar]=24.0, SD=5.1) trained at a knee flexion angle of 60 degrees, and group 3 (age range=20-36 years, X[bar]=23.2, SD=4.1) trained at a knee flexion angle of 90 degrees. Group 4 (age range=20-34 years, X[bar]=24.2, SD=4.1) did not train.

Subjects in groups 1 through 3 trained three times a week for 8 weeks. Exercise took place on the dynamometer with the subject stabilized with Velcro[R] straps placed at the chest, pelvis, and midthigh to secure the subject in the chair. The tested lower extremity was secured to the lever arm resistance pad with Velcro[R] straps about the subject's ankle. A manual goniometer was used to place the subject's knee at the appropriate angle for exercise (30 [degrees], 60 [degrees], or 90 [degrees] of knee flexion).

Each subject performed 20 repetitions of 6-second maximal isometric knee extension contractions at the assigned joint angle. Ten seconds of rest was allowed between contractions. Verbal encouragement was given to the subjects during the exercise session.

Performance of each exercise session by each subject was recorded on an attendance sheet to document compliance with the training program. If a subject missed a scheduled exercise session, she had to make up the exercise session on another day during the same week or during the next week (requiring an exercise frequency of four times per week during the week following the missed session). Any subject not exercising for greater than 4 days was dropped from the study.

Posttest. After the 8 weeks of training and 2 days of rest, all subjects were retested using the same procedure described for the pretest. Torque and EMG data were again collected and analyzed as described for the pretest. The difference between the pretest and posttest data for each dependent variable, torque and EMG activity, was used for data analysis.

Data Analysis

Reliability of the measurements of the two dependent variables (torque and EMG activity) was evaluated by performing separate ICCs (formula 1,1) for torque and EMG activity for the pretest-posttest data of the control group across all angles of knee flexion.[31] The mean changes (from pretest to posttest) and standard deviations in muscle torque and EMG activity during maximal voluntary isometric contraction of the quadriceps femoris muscles were computed for all four groups at all angles. A two-way (4 x 7, group x angle) multivariate analysis of variance (MANOVA) for repeated measures on one factor (angle) was initially performed to determine whether significant differences existed in the torque and EMG activity at each level of group and each level of angle. After the significant MANOVA, two follow-up two-factor analyses of variance (ANOVAs) for repeated measures on one factor (angle) were performed to make specific univariate comparisons for each dependent variable (torque, EMG activity).[32]

Comparisons of the simple main effects, involving the comparisons of the various levels of one factor (angle) at each separate level of the other factor (group), were performed by calculating several follow-up one-way ANOVAs. The Tukey test was used for post hoc analysis for each significant one-way ANOVA. Significance for all statistical analyses was accepted at the .05 level of probability.

Results

The mean values for the pooled pretest measurements and for the pooled posttest measurements across all knee flexion angles of the control group for torque were 112.71 N.m (SD= 29.30) and 112.81 N.m (SD=27.93). The ICC value calculated for torque pretest-posttest data of the control group across all angles of knee flexion was .97. The mean values for pooled pretest and posttest measurements of the control group for EMG activity were 4.85 (SD=2.16) and 5.02 (SD=2.27) mV/s. The ICC value for EMG pretest-posttest data of the control group was .89.

The mean changes in torque and EMG activity during maximal voluntary isometric contraction of the quadriceps femoris muscle determined for each level of the independent variable of group (30 [degrees], 60 [degrees], 90 [degrees] control) and angle (15 [degrees], 30 [degrees], 45 [degrees], 60 [degrees], 75 [degrees], 90 [degrees], and 105 [degrees]) are presented in Tables 1 and 2. Table 1 presents data for the 107 subjects according to their mean change in muscle torque, and Table 2 presents data for changes found in EMG activity.

The two-way (4 x 7, group x angle) MANOVA for repeated measures on one factor (angle) for torque and EMG activity at each level of group and each level of angle was found to be significant for group (F= 17.90; df=6,204; P < .05), trial (F=4.34; df=12,92; P < .05), and interaction (F=5.43; df=36,372; P < .05) using the Wilks' Lambda criterion. Two separate follow-up two-factor ANOVAs for repeated measures on one factor revealed a significant interaction effect (F=13.84; df= 18,618; P < .05) for the dependent variable of torque and a significant interaction effect (F=5.40; df= 18,618; P < .05) for the dependent variable of EMG activity. Following these significant interactions, analysis of the simple main effects was performed by calculating a one-way ANOVA for each level of angle across each group for the dependent variable torque (Tab. 3) and a one-way ANOVA for each level of angle across each group for the dependent variable EMG activity (Tab. 4).

The one-way ANOVAs across groups (and Tukey post hoc test) indicated that torque and EMG activity were significantly greater for all groups exercised than for the control group, not only at the specific angle at which exercise took place, but also at other angles. In addition to significant changes in torque and EMG activity found at the 30-degree angle of measurement, subjects training at the 30-degree angle of knee flexion (group 1) showed a significantly greater increase in torque and EMG activity at the 15-, 45-, and 60-degree angles of measurement than did the group that did not exercise (Figs. 1, 2). The fact that subjects exercising at a 60-degree angle of knee flexion (group 2) increased torque significantly more than did the control group at the 15-, 30-, 45-, and 75-degree angles of measurement (Fig. 3) and increased EMG activity significantly more than did the control group at the 30-, 45-, 60-, 75-, and 90-degree angles of measurement (Fig. 4) indicates that the increase in torque and electrical activity at angles at which exercise did not take place occurred to a similar degree.

Subjects training at the 90-degree angle of knee extension (group 3) showed a greater increase in torque than did the control group at all seven angles of measurement, suggesting increases in torque production at positions at least 75 degrees apart (Fig. 5). Similarly, group 3 demonstrated greater EMG changes than did the control group at six of the seven angles measured (30 [degrees], 45 [degrees], 60 [degrees], 75 [degrees], 90 [degrees], and 105 [degrees]) (Fig. 6). In comparison with groups 1 and 2, group 3 appeared to show the greatest amount of increase in force production at angles away from the angle of exercise and was the only group to show increased torque at the 90 and 105 degrees angles of measurement.

Discussion

Specificity of Isometric Exercise

Isometric exercises of the quadriceps femoris muscles of the knee at a joint angle for 8 weeks will increase torque and EMG activity of those muscles when the knee is at that same position. At each angle of knee flexion at which isometric exercise was performed (30 [degrees], 60 [degrees], and 90 [degrees] of knee flexion), torque and EMG activity increased significantly more than the data collected at the same angle for the control group. As early as 1953, Hettinger and Muller[33] reported that isometric exercise produced increased muscular strength. Since this early study, force gains following isometric training have frequently been described in the literature.[8-13,19-22]

The results of our study are also supported by previous research[13,19,20] that showed increased motor unit activity (as evidenced by increased EMG activity) following resistance training. Moritani and DeVries[19] reported that EMG activity increased following 8 weeks of training. The increased EMG activity following isometric training produced in our investigation adds support to the theory of DeLuca et al,[23] who suggested that a functional reserve of motor units exists that is not readily available for use during maximal contractions and that resistance training may allow the subject to learn to fully activate all motor units available.

Increases at Angles Away From Angle of Exercise

Further evaluation of our results indicates that isometric exercises of the knee at a specific joint angle for 8 weeks will increase torque or EMG activity at angles other than the angle of exercise. At each angle of knee flexion at which exercise took place (30 [degrees], 60 [degrees], and 90 [degrees] of knee flexion), torque and EMG activity increases were significantly greater than in the control group to at least 30 degrees beyond the angle at which exercise occurred.

Increases of EMG activity at angles away from the actual exercise angle following isometric training has not been previously reported. Our findings of increased torque at angles at which exercise did not occur following isometric exercise (30 [degrees] for group 1, 45 [degrees] of transfer to lower angles of knee flexion and 15 [degrees] of transfer to greater angles of knee flexion for group 2, and at least 75 [degrees] of transfer for group 3) are consistent with research by Meyers,[11] Knapik et al,[12] and Thepaut-Mathieu et al,[13] who reported increased torque at other angles. The increase of torque found in our study is in contrast to the findings of studies by Lindh,[8] Gardner,[9] and Belka,[10] who reported that isometric training effects were specific to the joint angle at which training is performed.

Effect of Muscle Length

Differences in joint angles at which isometric training took place during our study caused the quadriceps femoris muscles to be in a shortened position (group 1, training at a knee flexion angle of 30 [degrees]), a midrange position (group 2, training at a knee flexion angle of 60 [degrees]), or a lengthened position (group 3, training at a knee flexion angle of 90 [degrees]). With the quadriceps femoris muscles in a shortened position, less torque was produced than when muscles were in a lengthened position. In addition, carryover of training effects to other angles was greater for muscles exercised in a lengthened position. We believe the most effective use of isometric training appears to be accomplished when muscles are exercised in a lengthened position.

Clinical Implications

The results of our study suggest that, if the goal of resistance training in rehabilitation is to increase force production at a specific angle of knee flexion because a knee pathology exists causing a force deficit at that particular joint angle, isometric exercise should be performed at the specific angle of the force deficit. If the goal is to provide general increases in force production, however, then our results suggest that an efficient way to increase force production throughout the entire ROM is to exercise the quadriceps femoris muscles in the lengthened position. In addition, if exercise in the extended position of the knee is contraindicated for some reason, our results indicate that isometric exercise of the quadriceps femoris muscles with the knee flexed (lengthened position) will have an effect even at the angles at which patients cannot safely perform exercise.

Limitations and Suggestions for Future Research

One limitation of our study is the relatively homogeneous sample of subjects, who were young, female, and healthy. Although subjects were recruited from the age range of 20 to 36 years, most volunteers were college students and all subjects were free of knee pathology. Future research is needed to determine whether similar results would occur across a wider age range and whether male subjects should be included.

The results are limited because of the use of healthy subjects. The notion that angular specificity of isometric training occurs is commonly assumed in secondary sources (textbooks),[6,7,34-38] despite a limited amount of data from primary sources. We find this astonishing. Every primary research report on the specificity of isometric exercise (though few in number) has described the use of healthy subjects.[8-13] Most of the studies used a small sample size and a questionable design. Because past research has used healthy subjects, a comparison using healthy subjects is useful and relevant. In addition, our study (using healthy subjects) is valuable because of the large ample size, a good design, and the attempt to provide clinical implications for the use of isometric exercises based on the results of this study. Future research, however, is needed to examine the effects of isometric training in subjects with weakness as a result of pathology or disuse.

Additional limitations include the fact that data were collected on only one muscle group, the quadriceps femoris muscles, and the fact that muscle hypertrophy was not measured. Whether muscle groups surrounding other joints would react the same is not known. In addition, the influence of hypertrophy of muscles following training on torque production was not investigated in our study, but may be of interest in future research.

Conclusions

The purpose of this study was to investigate the changes occurring in torque and EMG activity of the quadriceps femoris muscles following 8 weeks of resistance training using isometric exercise. Based on the results of the statistical analysis and within the limitations of the study, the following conclusions were reached: 1. Isometric training of the quadriceps 

femoris muscles at a specific

joint angle significantly increased

torque and EMG activity at that

specific joint angle of exercise. 2. Isometric training of the quadriceps

femoris muscles at a specific

joint angle resulted in at least a

30-degree increase in torque and

EMG activity at angles at which

exercise did not occur, regardless

of the joint angle at which exercise

occurred. These results refute

some earlier research reporting

that isometric training increases

torque only at the specific angle of

exercise. 3. Exercises performed with the muscles

in a lengthened position (knee

flexion angle of 90 [degrees]) produced the

greatest increase of torque (with at

least 75 [degrees] of arc in which force

measurements increased) and EMG

activity (with at least 60 [degrees] of arc in

which force increased) at angles at

which exercise did not take place

following isometric training. The

results suggest that the most effective

use of isometric training appears

to be accomplished when

exercise is performed with muscles

in a lengthened position.

Enhanced understanding of isometric exercise as a result of the findings of this study will hopefully enable the clinician to provide more effective and scientifically based rehabilitation when using isometric training programs.

Acknowledgment

We thank Jimmy Ishee, PhD, for preparation of this manuscript. (*) Cybex, Div of Lumex Inc, 2100 Smithtown Ave, Ronkonkoma, NY 11779. (+) Therapeutics Unlimited Inc, 2835 Friendship St, Iowa City, IA 52240. [double dogger] Tenmar, 858 E Congress Park Dr, Centerville, OH 45459. [subsection] Velcro USA Inc, 406 Brown Ave, Manchester, NH 03103.

References

[1] Fleck SJ, Schutt RC. Types of strength training. Orthop Clin North Am. 1983;14:449-458. [2] Blackburn TA. Rehabilitation of anterior cruciate ligament injuries. Orthop Clin North Am. 1985;16:241-269. [3] Montgomery JB, Steadman JR. Rehabilitation of the injured knee. Clin Sports Med. 1985;4: 333-343. [4] Nitz A. Physical therapy management of the shoulder, Phys Ther. 1986;66:1912-1919. [5] Torg JS, Balduini FC, Bonci C, et al. A modified Bristow-Helfat-May procedure for recurrent dislocation and subluxation of the shoulder. J Bone Joint Surg [Am]. 1987;69:904-913. [6] Mangine R, Hickman TP, Eldridge VL. Improving strength, endurance, and power. In: Scully RM, Baines MR, eds. Physical Therapy. Philadelphia, Pa: JB Lippincott Co; 1989: chap 40. [7] Sanders MT. Weight training and conditioning. In: Sanders B, ed. Sports Physical Therapy. East Norwalk, Conn: Appleton & Lange; 1990: chap 14. [8] Lindh M. Increase of muscle strength from isometric quadriceps exercises at different knee angles. Scand J Rehabil Med. 1979;11:33-36. [9] Gardner G. Specificity of strength changes of the exercised and nonexercised limb following isometric training. Research Quarterly. 1963;34:98-101. [10] Belka D. Comparison of dynamic, static and combination training on dominant wrist flexor muscles. Research Quarterly. 1968;39: 241-250. [11] Meyers CR. Effects of two isometric routines on strength size and endurance in exercised and nonexercised arms. Research Quarterly. 1967;38:430-440. [12] Knapik JA, Mawdsley RH, Ramos MU. Angular specificity and test mode specificity of isometric and isokinetic strength training. Journal of Orthopaedic and Sports Physical Therapy. 1983;5:58-65. [13] Thepaut-Mathieu C, Van Hoecke J, Maton B. Myoelectrical and mechanical changes linked to length specificity during isometric training. J Appl Physiol. 1988;64:1500-1505. [14] Sale DG. Neural adaptation to resistance training. Med Sci Sports Exerc. 1988;20: S135-S145. [15] Bandy WD, Lovelace-Chandler V, McKitrick-Bandy B. Adaptation of skeletal muscle to resistance training. Journal of Orthopaedic and Sports Physical Therapy. 1990;12:248-255. [16] Yarasheski KE, Lemon PWR, Gilloteaux J. Effect of heavy resistance training on muscle fiber composition in young rats. J Appl Physiol. 1990;69:434-437. [17] MacDougall JD, Sale DG, Alway SE, et al. Muscle fiber number in biceps brachii in body builders and control subjects. J Appl Physiol. 1984;57:1399-1405. [18] Gollnick PD, Timson BF, Moore RL, et al. Muscular enlargement and number of fibers in skeletal muscle of rats. J Appl Physiol. 1981;50: 936-943. [19] Moritani T, DeVries HA. Neural factors versus hypertrophy in the time course of muscle strength gain. Am J Phys Med. 1979;58:115-130. [20] Hakkinen K, Komi PV, Alen M. Effect of explosive type strength training on isometric force- and relaxation-time, electromyographic and muscle fibre characteristics of leg extensor muscles. Acta Physiol Scand. 1985;125: 587-600. [21] Ward J, Fisk G. The difference in response of the quadriceps and the biceps brachii muscles to isometric and isotonic exercise. Arch Phys Med Rehabil. 1964;45:614-619. 22 Smith LE. Strength increments following massed and distributed practice relative to motor learning. Med Sci Sports Exerc. 1974;6: 154-157. [23] DeLuca CJ, LeFevers RS, McCue MP, et al. Behavior of human motor units in different muscles during linearly varying contractions. J Physiol (Lond). 1982;329:113-128. [24] Isolated Joint Testing and Exercise: A Handbook for Using the Cybex II and UBXT. Ronkonkoma, NY: Cybex, Div of Lumex Inc; 1984. [25] Sapega AA, Nicholas JA, Sokolow D, Saraniti A. The nature of torque "overshoot" in Cybex isokinetic dynamometry. Med Sci Sports Exerc. 1982;14:368-373. [26] Caldwell LS, Chaffin DB, Dukes FN, et al. A proposed standard procedure for static muscle strength testing. Am Ind Hygiene Assoc J. 1974; 12:201-206. [27] Hislop HJ. Quantitative changes in human muscular strength during isometric exercise. J Amer Phys Ther Assoc. 1963;43:21-38. [28] Kroemer KHE, Marras WS. Towards an objective assessment of the "maximal voluntary contraction" component in routine muscle strength measurements. Eur J Appl Physiol. 1980;45:1-9. [29] Soderberg GL, Cook TM. Electromyography in biomechanics. Phys Ther. 1984;64: 1813-1820. [30] Yang JF, Winter DA. Electromyography reliability in maximal and submaximal isometric contractions. Arch Phys Med Rehabil. 1983;64: 417-420. [31] Shrout PE, Fleiss JL. Intraclass correlations: uses in assessing rater reliability. Psychol Bull. 1979;86:420-428. [32] Huck SW, Cormier WH, Bounds WG. Reading Statistics and Research. New York, NY: Harper & Row, Publishers Inc; 1974:87-90. [33] Hettinger T, Muller EA. Cited by: Hislop HJ. Quantitative changes in human muscular strength during isometric exercise. J Amer Phys Ther Assoc. 1963;43:21-38. [34] Kisner C, Colby LA. Therapeutic Exercise: Foundations and Techniques. Philadelphia, Pa: FA Davis Co; 1985:112. [35] Nicholas JA, Hershman EB. The Lower Extremity and Spine in Sports Medicine. St Louis, Mo: CV Mosby Co; 1986:571. [36] Athletic Training and Sports Medicine, Park Ridge, Ill: American Academy of Orthopaedic Surgeons; 1991:783. [37] Prentice W. Rehabilitation Techniques in Sports Medicine. St Louis, Mo: Mosby-Year Book Inc; 1990:39. [38] McArdle WD, Katch FI, Katch VL. Exercise Physiology: Energy, Nutrition, and Human Performance. Philadelphia, Pa: Lea & Febiger; 1986:379.

Commentary

The purpose of this study was to investigate neural and strength changes occurring after exercise training in various joint configurations. The study's main attribute is that it represents an important area of study because increasing strength is a major goal of physical therapists. The authors claim that, by exercising a muscle in a lengthened position, strength "carries over" to other lengths more effectively.

Strengths of the Report

This is an important area of study, it has been suggested since the 1950s that neural training can result in significant increases in muscle strength without muscular hypertrophy. Neural effects have been shown to cause "strength" increases early in exercise training, whereas muscular changes appear much later.[2]

The most powerful conclusion presented in this report is that, by exercising the quadriceps femoris muscle group in a lengthened position, the carryover of strength to other joint angles is greater. If this result is true, it will have a significant impact on the practice of physical therapy.

The authors investigated neural factors affecting strength. Using the surface electromyogram (EMG) as a "window" into the nervous system, the authors attempted to quantify the neural drive to the quadriceps femoris muscles. This is a more enlightened approach to understanding the physiological basis of strength gain. Unfortunately, too many studies currently ascribe any strength increase to muscle hypertrophy. We now know that this is a naive point of view.

Weaknesses of the Report

The study attempted no measure of muscle hypertrophy. By measuring EMG activity, the authors can make a statement about neural factors, and the reader is left to believe that the "remaining" effect is due to the muscle. This is an incomplete experimental design. The authors could have simply estimated hypertrophy by measurement of thigh girth, or they could even have used magnetic resonance imaging (MRI) as a more ambitious approach.[3] It is thus extremely difficult to interpret their data without some estimate of the magnitude of muscle hypertrophy that occurred during the time course of the training period.

Electromyographic measurements were made from a single muscle, but torque was measured from the entire quadriceps femoris muscle group. In several cases, there is a poor correlation between EMG change and torque change. In Figure 1, for example, there is almost no torque increase at joint angles above 60 degrees and yet large EMG increases are presented for these angles in Figure 2. How could EMG activity increase with no torque increase? Did muscles atrophy.? This type of uninterpretable data could simply result from the fact that the EMG data are not representative of the torque data. if this is true, it casts a grave shadow of doubt on the study.

From the report, it appears that EMG electrodes were placed primarily over the rectus femoris muscle. Because the rectus femoris muscle strength comprises only about 20% of the quadriceps femoris muscle strength,[4] the EMG measurements would not necessarily reflect torque changes that arise from the contributions of all four quadriceps femoris muscles.

There is no presentation of the time course of neural and torque changes. if we are to believe that neural changes are contributing differentially to the strength gain at various joint angles, it would be important for us to see the time course of such changes. We would expect the early changes to be neural and the later changes to be due to muscular hypertrophy. The key question, however, is, What proportion of the strength gain at each time period is due to muscular versus neural factors? Differentiating between neural and muscular mechanisms of strength change is extremely important in that it will then permit application of this experiment to clinical practice. If the effect the authors have shown were primarily neural in nature, we would then focus our efforts on neural mechanisms of training (eg, biofeedback), whereas if the effect were muscular, our efforts would be focused on mechanical inducement of hypertrophy (eg, high-resistance training).

The authors present the data as if they are simply training one "big" muscle group with nearly identical properties. It has been demonstrated that there is significant architectural and fiber-type heterogeneity between the various heads of the quadriceps femoris muscles. in addition, recent experimental and theoretical studies suggest that the various quadriceps femoris muscles may have their optimal length at different joint angles (see Figure 7 of Herzog et al[5]). Thus, some of the observed effect may simply be due to the fact that in one joint configuration, a muscle was trained more closely to its optimal length, whereas in another situation, a different muscle was closer to its optimal length. This would provide a simple mechanical basis for the differential effect observed. This effect was clearly neglected by the authors. Again, muscle cross-sectional area estimates could address this question.

The statistical analysis for the experimental design was inappropriate. The two-way analysis of variance (ANOVA) that tested for the main effects of experimental group and angle was inappropriately reanalyzed as seven individual one-way ANOVAs in order to test for angle-specific effects. This application of multiple analyses to the same data set increases the probability of Type I error (false positive) and therefore suggests that some of the authors' "significant" effects were merely due to incorrect statistical analysis.

The control group was not brought into the clinic for "sham" testing. It would have been helpful to bring in the control group (group 4) three times per week in order to account for the "coaching" factor. It has been demonstrated that "human contact" is a powerful motivator, and it could be that some of the neural effects observed were due to the dramatic amount of attention that the subjects in the experimental groups received. This is a minor point, however, because although this effect may change the magnitude of torque or EMG increase, it is not likely that it could account for the differential effect of training at different joint angles.

Richard L Lieber, PhD Department of Orthopaedics Biomedical Sciences Graduate Group University of California, San Diego and Veterans Administration Hospital 3350 La Jolla Village Dr San Diego, CA 92161

David J Pierotti, PhD Postdoctoral Fellow Muscle Physiology Laboratory Department of Orthopaedics Biomedical Sciences Graduate Group University of California, San Diego and Veterans Administration Hospital 3350 La Jolla Village Dr San Diego, CA 92161

References

[1] Delorme TL, West FE, Shriber WJ. Influence of progressive resistance exercise on knee function following femoral fractures. J Bone Joint Surg [Am]. 1950;32:910-924. [2] Moritani T, DeVries HA. Neural factors versus hypertrophy in the time course of muscle strength gain. Am J Phys Med. 1979;58:115-130. [3] Fukunaga T, Roy RR, Shollock FG, et al. Physiological cross-sectional area of human leg muscles based on magnetic resonance imaging. J Orthop Res. 1992;10:926-934. [4] Wickiewicz TL, Roy RR, Powell PJ, Edgerton VR. Muscle architecture of the human lower limb. Clin Orthop. 1983;179:317-325. [5] Herzog W, Abrahamse SK, ter Keurs HEDJ. Theoretical determination of force-length relations of intact human skeletal muscles using the cross-bridge model. Pflugers Arch. 1990; 416:113-119.

Author Response

The intent of this research was to investigate the effects of an isometric training program. Since the early stages of this research, we have been amazed at how little research existed on the topic of specificity of isometric training. The idea that isometrically exercising a muscle increased "strength" only at the specific angle of exercise was one of those "rules" that everyone was taught in school and that has remained unchallenged for many years. Recently, leading clinicians have advocated the use of multiple-angle isometric exercises, but (as happens all too frequently in many procedures used in physical therapy) the rationale for use of this exercise protocol was based anecdotal remarks. Therefore, a study investigating isometric training was needed.

We appreciate the time Dr Lieber and Dr Pierotti took in providing commentary for our manuscript, as well as the opportunity to respond. The main concerns of the commentators appear to relate to three primary areas: (1) No attempt was made to measure hypertrophy, (2) data were collected from one "big" muscle, and (3) our statistical analyses were inappropriate. 1. The commentators make suggestions

for a more grandiose study

incorporating the measurement of

hypertrophy as well as the investigation

of the time course for neural

and torque changes. As reported

in the article, a limitation of

the study was the inability to measure

the change in muscle size. We

agree with the commentators that

an accurate measurement of

change in muscle size would have

provided even more information

concerning the physiologic change

related to increased force production

reported in this study. Because

changes in cross-sectional area are

obscured by the changes in subcutaneous

fat and bone that could

also occur with the training program,

however, change in muscle

size as a result of training cannot

be accurately inferred from circumferential

measurements.[1-4] The

accurate measurement of change in

muscle size would require, as the

commentators point out, more

sophisticated equipment (eg, computed

tomography, ultrasonography,

magnetic resonance imaging).

The cost of such measurements

was more expensive than could

realistically be afforded in this

research.

The evaluation of the extent of

change in torque and electromyographic

(EMG) activity across time

was considered early in the planning

of this research. To measure

change across time, however,

would have added a third independent

variable to an already complicated

design expanding the current

two-way multivariate analysis

of variance [MANOVA] to a three-way

MANOVA). In addition, previous

research has already effectively

documented that early changes are

the result of neurologic changes,

followed later in the training program

by changes in muscle size.[5,6]

Therefore, the measurement across

time was not pursued in this investigation. 2. The commentators express concern

that no differentiation was

attempted between the four quadriceps

femoris muscles, more specifically

that "EMG electrodes were

placed primarily over the rectus 

femoris muscle" and that "various

heads of the quadriceps femoris 

muscles" have different architectural

and fiber-type characteristics.

Although such a differentiation

would be interesting, investigating

the response of each specific quadriceps

femoris muscle to isometric

exercise would require much

more complexity in the research

design and would provide little

additional clinical information to

the specific problem being investigated.

Analysis of the role of each

specific muscle of the quadriceps

femoris muscle group was not a

purpose of this investigation. 3. We believe that our statistical analyses

are appropriate. We followed a

significant MANOVA with analyses

of variance on each of the variates.

This method is often referred to as

the Least-Significant Difference

(LSD) test or the protected F approach

whereby the overall multivariate

test provides protection

from an inflated alpha level on the

univariate tests.[7] Two other

follow-up procedures are appropriate:

simultaneous confidence intervals

and discriminate function

analysis. Combinations of procedures

have also been suggested

such as combining LSD and Bonferroni

procedures. Although advantages

and disadvantages can be

argued for each procedure, little

consensus exists as to which

method is the best.[7]

Addressing all the questions raised by the commentators would be difficult in any one study. Because previous research was lacking in the area of isometric training, we attempted to answer the most clinically relevant questions first--What is the effect of isometric training on force at the angles at which exercise did take place, as well as at angles at which exercise did not occur (irrespective of a complete physiologic explanation)? Despite some stated limitations in this investigation, we believe the report makes a meaningful contribution to the body of knowledge needed for clinicians integrating isometric exercise into their rehabilitation protocols. Based on our research, clinicians are encouraged to use isometric exercises performed in the lengthened position of the muscle in their patients who require general strengthening of the knee extensor muscles. We agree with the commentators that the results of this study "will have a significant impact on the practice of physical therapy." William D Bandy, PhD, PT, SCS, ATC William P Hanten, EdD, PT

References

[1] Haggmark T, Jansson E, Svane B. Cross-sectional area of the thigh muscle in man measured by computed tomography. Scand J Clin Lab Invest. 1978;38:355-360. [2] Young A, Hughes I, Russell P, et al. Measurement of quadriceps muscle wasting by ultrasonography. Rheumatol Rehabil. 1980;19:141-148. [3] Weiss LW. Upper extremity assessment of muscularity in young children. Journal of Orthopaedic and Sports Physical Therapy. 1987;8: 397-401. (4) Lopresti C, Kirkendall DT, Street GM, et al. Quadriceps insufficiency following repair of the anterior cruciate ligament. Journal of Orthopaedic and Sports Physical Therapy. 1988;9: 245-249. [5] Moritani T, DeVries HA. Neural factors versus hypertrophy in the time course of muscle strength gain. Am J Phys Med. 1979;58:115-130. [6] Hakkinen K, Komi PV. Electromyographic changes during strength training and detraining. Med Sci Sports Exerc. 1983;15:455-460. [7] Bray JH, Maxwell SE. Analyzing and interpreting significant MANOVAs. Review of Educational Research. 1982;52:340-367.