Electrical stimulation versus electromyographic biofeedback in the recovery of quadriceps femoris muscle function following anterior cruciate ligament surgery.

Following an anterior cruciate ligament (ACL) reconstruction, immobilization and restricted use of the operative limb results in quadriceps femoris muscle atrophy and weakness. Significant disuse atrophy can occur as early as the first several days after surgery and certainly within the first few weeks following surgery. [1] A primary focus of ACL rehabilitation protocols, therefore, is the recovery of quadriceps femoris muscle force production and function. We believe it is important that patients begin exercising the quadriceps femoris musculature during the initial post-operative period in order to reduce muscle atrophy, recover muscle function, and recover and maintain knee range of motion (ROM).

Postoperative Strength

Recovery

In order to minimize the atrophic effects of immobilization and disuse following ACL reconstruction, quadriceps femoris muscle strengthening exercises are typically begun on the first postoperative day. These exercises commonly include quadriceps femoris muscle setting (QS) and straight leg raises (SLRs). Often, these exercises are difficult for patients to perform correctly during the initial postoperative weeks because of pain, edema, and possibly the distortion of normal joint receptor activity. [2,3] Krebs et al [3] have suggested that an anterior capsular incision, similar to that used in the ACL procedure, may cause a temporary disruption in the normal joint receptor activity, leaving the patient without accurate proprioceptive feedback from the quadriceps femoris muscle. Normal feedback patterns may not return for up to 3 weeks after surgery. Consequently, initial attempts to contract the quadriceps femoris muscle may result in inadequate levels of motor unit activation, incomplete contractions, and ultimately a less-than-optimal rate of force recovery.

In an effort to maximize a patient's effort to contract the quadriceps femoris muscle and enhance the rate of force production, clinicians may choose to augment the exercise with a training modality that facilitates higher levels of motor unit activity and more complete contractions during exercise. Because force development is a result of both neural and muscular elements, [4-6] the training method should facilitate both. One conventional choice is electrical stimulation (ES), which artificially activates the intramuscular branches of a motor nerve, causing a muscle contraction. The efficacy of ES as a facilitator of quadriceps femoris muscle strengthening has been examined by numerous researchers under various conditions. In healthy subjects, no differences in force measurements have been found when comparing electrical stimulation training with voluntary isometric exercise (VIE) training. [7-10] Among subjects undergoing postoperative knee rehabilitation (ie, ACL reconstruction), however, ES alone and ES combined with VIE have been shown to be more effective than VIE alone in retarding quadriceps femoris muscle atrophy and loss of force production following surgery. [11-13] Although this modality does involve the peripheral motor structures and pathways, training with ES may only minimally involve the higher motor pathways. [5]

Another modality that has been used to facilitate greater levels of muscle activity and more complete muscle contractions is electromyographic (EMG) biofeedback. Biofeedback allows a subject to monitor a voluntary contraction. Via surface electrodes, muscle activity from a targeted muscle is converted to a visual or audible feedback signal. The patient can use this feedback to augment diminished joint receptor feedback and better monitor the quality of the muscle contraction (ie, the level of muscle activity). We believe that biofeedback, unlike ES, requires the patient to integrate all levels of the nervous system while relearning to contract the quadriceps femoris muscle. Several studies have compared biofeedback-facilitated quadriceps femoris muscle exercise with exercise alone in healthy individuals [14-16] and in patients following knee surgery [17,18] and have demonstrated greater peak torque, [14,15] greater EMG output, [16] an increased rate of knee extension recovery, [17] and an increased rate of peak torque recovery [18] with the use of biofeedback.

Although both ES and biofeedback have been shown to be more effective than voluntary exercise alone in the recovery of quadriceps femoris muscle function, there are no studies in the literature comparing the two modalities. The purposes of this study were to compare ES and biofeedback as adjuncts to quadriceps femoris muscle strengthening exercises and to determine whether differences exist in the rate of recovery of peak torque output and knee ROM during a 6-week training period following an ACL reconstruction.

Method

Subjects

Thirty patients (16 male, 14 female), ranging in age from 15 to 44 years ([Mathematical Expression Omitted], SD=7.8) participated in this study. All patients had suffered acute tears of the ACL and had undergone an arthroscopic-assisted bone-tendon-bone patellar tendon autograft surgical reconstruction. During a preoperative evaluation by a staff physical therapist and an athletic trainer (LB), patients were familiarized with the proposed postoperative training and testing protocol. All patients signed an informed consent form and were then assigned to one of two groups. During the first 6 postoperative weeks, group 1 (n=15) used ES in conjunction with quadriceps femoris muscle exercises, and group 2 (n=15) used an EMG biofeedback unit to monitor muscle activity during quadriceps femoris muscle exercises. Assignment of the patients to the two groups was performed with a randomized matching procedure, controlling for age and gender (Tab. 1). Patients whose diagnosis indicated medial collateral or posterior cruciate ligament involvement, or who were to undergo additional procedures (eg, high tibial osteotomy), were excluded from the study. In group 1, 8 right knees and 7 left knees were the operative knees; in group 2, 9 right knees and 6 left knees were the operative knees. The study design and informed consent procedure were approved by the participating physicians.

Instrumentation

The instruments used in this study were a MyoTrac[TM] EMG biofeedback unit, (*) a Myocare Plus electrical stimulation unit, (1) and a Cybex[R] II is isokinetic dynamometer with dual-channel chart recorder. (2) The Myocare Plus is a programmable electrical stimulation unit.

[TABULAR DATA OMITTED]

Program parameters for all ES units were a rectangular waveform with a maximal output of 50 mA, delivered at a rate of 35 pulses per second, with a ramp up time of 4 seconds, a hold time of 10 seconds, a ramp down time of 2 seconds, and an off time of 20 seconds. Two 5.08- X 10.16-cm (2- X 4-in) Myocare/Tenzcare[TM] (1) self-adhering silicone rubber electrodes were attached to one stimulator output channel. The active electrode was placed proximally over the femoral nerve, and the dispersive electrode was placed distally, approximately 5 to 7 cm superior to the patella and over the vastus medialis muscle. The Myocare Plus was used in conjunction with voluntary quadriceps femoris muscle contractions during QS and SLR exercises.

The MyoTrac[TM] is a portable EMG biofeedback device that offers an integrated EMG signal, records muscle activity between 0.5 and 1,000 [mu]V, and has both a visual light bar display and an audible signal that sounds when the patient exceeds a predetermined EMG threshold level. Disposable surface electrodes, (*) consisting of three silver-plated electrodes spaced triangularly in a foam-backed, 5.08-cm-(2-in-) diameter adhesive disk, were used. With the knee in an extended position, a single disk was placed 3 to 5 cm above the superior border of the patella and 2 to 3 cm medially, so as to focus on the motor unit activity of the extensor mechanism during QS and SLR exercises. The MyoTrac[TM] was used by patients during QS and SLR exercises as a monitor and facilitator of knee extensor use.

The Cybex[R] dynamometer was used to record the torque (in foot-pounds [1 ft.lb=1.356 N.m]) produced by patients during the 6-week postoperative evaluation of quadriceps femoris muscle isometric peak torque. We chose to administer this test 6 weeks following surgery for two reasons: (1) We believe that, by testing as early as possible, the torque measure might reflect a degree of the neural reorganization that we believe precedes hypertrophy, and (2) this was the earliest date that participating physicians would allow a maximal effort against resistance. The chart recorder torque damping was set at 2. Dynamometer torque calibration was performed on a regular basis, but was not verified at each testing session. One tester (LB), who was not blind to group assignment, performed all dynamometer tests. We made no attempt to assess the test-retest reliability of our torque measurements, but we believe, based on previous works, [19,20] that these measurements are reliable.

For goniometric measures of active knee extension, the same tester (LB) used a standard, transparent plastic goniometer, 17.78 cm (7 in) long with a 360-degree scale and 1-degree increments. Although intratester reliability was not determined prior to data collection, previous studies [21,22] have demonstrated high intratester reliability for these measurements.

Procedure

All patients had been familiarized with the training and testing protocol during a preoperative therapy session. We met with each patient again 1 day after surgery and assigned the patient either an ES device or a biofeedback device. These devices were used in conjunction with a home program of QS and SLR exercises for the first 4 weeks following surgery. Subjects were given log sheets to document the number and duration of exercise sessions per day in addition to the number of repetitions performed during each session and the intensity at which they were done (ie, the percentage of 50 mA output for patients using ES and the EMG threshold level for patients using biofeedback).

During the first postoperative session, patients in both groups were reviewed and subjectively evaluated on QS and SLR performance and were then instructed in the performance of these exercises in conjunction with the device they had been assigned. The patients in group 1 were instructed to contract the quadriceps femoris muscle with a maximal effort for the 10-second hold time and to relax during the 20-second off time. These patients were asked to increase the intensity with increased tolerance. Through the course of the study, the patients increased the intensity from an average of 15 mA initially to a final average of 40 mA. Biofeedback instruction included finding a threshold EMG level that the patient could achieve only by contracting with a maximal effort. The patients in group 2 were instructed to contract the quadriceps femoris muscle to their EMG threshold level, to maintain the audible signal for 10 seconds, and to rest for 20 seconds. The patients were initially able to achieve a threshold level of 7 to 10 [mu]V and progressed to 50 to 60 [mu]V over the 6-week exercise period.

In an effort to standardize exercise instruction, the following procedure was followed for all patients:

1. Verbal exercise cues were limited to the following: (1) for the QS exercise protocol, "press the back of the knee into the table," "look for the kneecap to shift and the foot to raise," and finally "try to keep the hip muscles as relaxed as possible," and (2) for the SLR exercise protocol, "maintain a QS throughout the lifting and lowering of the leg."

2. Instruction was always given by one or both of the experimenters (VD and LB), and both experimenters were present at all first sessions.

3. Patients were given written information and instructions regarding the general exercise protocol in addition to specific operating instructions for the exercise device they had been assigned.

Patients in both groups were asked to use the exercise device three times a day for 30 minutes at each session, but were told to record on their log sheets only the actual number of sessions and repetitions completed. These records were kept for the first 4 weeks, beginning 1 day after surgery. For the 2-week interim between the 1-month home exercise program and the 6-week Cybex[R] test, ES- or biofeedback-facilitated QSs and SLRs (3 sets of 10 repetitions each) were performed three times per week at the Knoxville Orthopedic Clinic (Knoxville, Tenn) as part of a progressively aggressive rehabilitation protocol that was a modification of that described by Noyes et al. [1] (Five patients, 3 from group 1 and 2 from group 2, lived out of town and followed this protocol at more convenient clinical sites.) This protocol progressed from week 1 to week 6 to include the addition of ankle weights (week 2), progressive resistive exercises (week 4), and quadriceps femoris muscle isometric contractions at fixed points (90[degrees], 60[degrees], and 45[degrees] of flexion) against the Cybex[R] dynamometer (week 4) (Tab. 2). These workouts were supervised by one of the experimenters

[TABULAR DATA OMITTED]

(LB) and by a member of the Knoxville Orthopedic Clinic's physical therapy staff. Throughout the 6-week training period, all patients followed the same protocol and used only their assigned modality to facilitate quadriceps femoris muscle contraction during QSs and SLRs.

Data Collection

During the sixth postoperative week, quadriceps femoris muscle isometric peak torque was measured using a Cybex[R] II isokinetic dynamometer. Each patient was tested while in a seated position, with the hips at approximately 80 to 90 degrees of flexion and the trunk and tested limb stabilized with self-adhesive straps. The knee was positioned 60 degrees from full extension to avoid stressing the graft and graft site at this early postoperative stage. The test involved three maximal 3-second contractions with 10-second rest periods. [23] Patients were allowed to grasp the sides of the Cybex[R] table during the test. Verbal commands of "ready" and "go" were given to the patients, and contraction time was counted aloud by the experimenter. All patients were tested on the nonoperative limb first. For each limb, the best of the three contractions was recorded as the maximal torque produced. We did not gravity-correct these torque values and believe that, because all subjects were tested isometrically at one specified angle and no comparisons were made between knee flexors and extensors, our torque values were not compromised by this omission. A ratio of torque values in the operative to nonoperative limbs was determined for each patient. This ratio served as the measure of postoperative force recovery following the 6-week training period.

Goniometric measurements of active knee extension were taken weekly for the first 6 weeks following surgery. Although we were unable to collect all six measurements for five patients, all patients had recorded measurements for weeks 1, 2, 4, and 6 (which coincided with scheduled appointments with the patients' physician); therefore, these four measurements were used in the data analysis. For the goniometric measurements, the patients were seated on an examination table, with their legs extended the length of the table and their hands in their lap. Patients were allowed to contract the quadriceps femoris muscle two or

[TABULAR DATA OMITTED]

three times as a warm-up in preparation for one maximal contraction and one measurement of active extension. These measurements were taken prior to the exercise session for that day. Patients were encouraged to produce a maximal effort, but there was no continuous prompting throughout the measurement.

Data Analysis

For measurements of peak torque, a ratio of isometric peak torque between the operative and the nonoperative limbs at the sixth postoperative week was calculated for each patient. The mean percentage of recovery of the nonoperative limb's peak torque was computed for each group and compared using an independent-samples t test.

The ROM data for weeks 1, 2, 4, and 6 were analyzed using a two-way analysis of variance (ANOVA) (groups X weeks) for repeated measures on the second factor. A Newman-Keuls multiple-comparison post hoc test for significance was performed following the determination of a significant F value.

Results

The results of the peak torque evaluation, shown in Table 3, revealed that the patients in group 2 recovered a significantly greater percentage of their nonoperative limb's peak torque [Mathematical Expression Omitted] than did the patients in group 1 [Mathematical Expression Omitted] P=.044).

The rate at which patients regained ROM during the first six postoperative weeks was represented by measurements of active knee extension at weeks 1, 2, 4, and 6. Results of the ANOVA, presented in Table 4, revealed no significant effects for groups (F=0.12; df=1,27; P=.738). A significant effect for weeks, however, did exist (F=6.84; df=3,81; P<.001), demonstrating the expected recovery of joint ROM for all patients from week 1 through week 6. Post hoc comparisons of the ROM data pooled across groups showed no significant difference between weeks 1, 2, and 4, but did reveal a significant difference (P<.01) between the ROM at week 6

Table 4. Results of Analysis of Variance of Differences Between Groups in of Range of Motion.

Source             df   SS        MS     F      P
Between subjects
  Groups (A)        1      4.50    4.50  0.11    .738
  Error            27   1065.04   39.45
Within subjects
  Weeks (B)        3     220.58   73.53  6.84   <.001
  A X B            3       9.17    3.06  0.28   .837
  Error            81    871.22    10.76

versus the ROMs at weeks 1, 2, and 4. This difference (3.4 [degrees]), however, was less than intratester measurement errors reported by Rothstein et al [19] (3.5 [degrees]-5 [degrees]) and Boone et al [20] (4 [degrees]).

Discussion

This study was designed to compare ES and biofeedback as adjuncts to quadriceps femoris muscle strengthening exercises following ACL reconstruction. The primary objectives of postoperative joint rehabilitation are the recovery of muscle force and joint ROM. The results indicate that there was greater recovery of isometric peak torque by use of biofeedback than by use of ES and that there was no difference in the recovery of active knee extension when each of these modalities was used.

The recovery of quadriceps femoris muscle force was represented in this study as the percentage of the nonoperative limb's peak torque generated by the operative limb during a 6-week postoperative evaluation. The patients in this study who used biofeedback to augment quadriceps femoris muscle exercises demonstrated a greater peak torque recovery [Mathematical Expression Omitted] than patients who used ES [Mathematical Expression Omitted]. Delorme and Watkins [4] suggest that strength gains are based on two components: (1) the neural changes associated with heightened motor unit activation and more organized patterns of activation, collectively referred to as "motor learning," and (2) the actual morphological changes that result in hypertrophy. The magnitude of these strength gains is a function of the degree and duration of the exercise effort. [24] In this study, the observed differences in torque recovery between groups 1 and 2 may have been a result of the work effort during exercise sessions. We believe biofeedback-assisted exercise requires the patient to formulate a motor strategy, initiate the muscle contraction, and voluntarily maintain the contraction during the hold time. In contrast, ES provides an artificial stimulus to contract and will maintain the contraction independent of patient effort.

Although the patients in this study who used ES followed the artificial stimulus with a voluntary contraction, these patients did not have to initiate the contraction or exert voluntary effort to maintain it, which might account for the differences in torque recovery observed between the two groups. In our experience, patients using ES can become passive and let the "machine" do the work. If this passivity occurs while the patient is being supervised by a therapist, the exercise can be corrected. Home workouts, however, which account for a majority of the exercises performed during the first month following ACL reconstruction, may be compromised. This is an important phase of the rehabilitation process, and the provision of a home exercise device that does not allow passivity during exercise may encourage a more effective exercise program. Although the daily exercise logs indicated that all patients performed approximately the same number of home exercise repetitions per day, it is possible that the work effort, and consequently the training intensity, was consistently greater for the patients who used biofeedback compared with those who used ES and may have accounted for their increased recovery of quadriceps femoris muscle peak torque.

A second objective of postoperative joint rehabilitation is the recovery of joint ROM. Active knee extension is an essential component of a normal and functional gait. The results of this study indicate that no significant difference exists between ES and biofeedback in influencing the rate at which patients recover full extension. This finding was somewhat surprising because group 2 demonstrated greater quadriceps femoris muscle torque recovery compared with group 1 during the same 6-week period. Because the degree of active knee extension is a function of quadriceps femoris muscle contraction, one might expect the trend in quadriceps femoris muscle torque to be reflected in the measurements of active extension. This was not the case, however, and a review of the data (Tab. 2) and of the postoperative protocol suggests that measures of ROM may not have been a valid measure of progressive changes in quadriceps femoris muscle function. Neither group demonstrated a significant loss of extension immediately following surgery and therefore had very little motion to recover. This result is a direct consequence of recent changes in surgical procedures and immobilization time frames. [25]

Prior to these advances, active knee extension was often difficult to recover and therefore a primary focus of the rehabilitation protocol. Joint motion was restricted for at least 6 weeks postoperatively, which allowed a greater potential for the development of conditions such as hamstring muscle contracture and adhesion formation that can lead to joint contracture. Depending on the tendencies of the individual to develop these conditions, knee extension could be significantly limited. Our current rehabilitation protocol does not call for restricted motion, and full extension is encouraged on the first postoperative day. The protocol itself thus reduces the chances of contracture by preventing the conditions mentioned above. Consequently, the individual's predisposition to restrictive conditions is no longer as much of a limiting factor.

Our current protocol may result in less loss of ROM and less variation in ROM among individuals. The patient's ROM, therefore, may not be as contingent on the progressive changes in quadriceps femoris muscle torque following ACL reconstruction. At the initiation of this study, this protocol had only recently been adopted, and, although we expected a decreased incidence of significantly limited ROM, we were not expecting the minimal limitations we observed. We believe that we obtained these measurements in a reliable manner, using the same instrument and the same tester for all measurements. [21,22] We must question the validity of this protocol, however, for the assessment of ES and biofeedback as postoperative training modalities.

Clinical Implications

Although the clinician may be more concerned with the neuromuscular aspects and functional outcomes of postoperative exercise, the patient may have more fundamental concerns related to comfort and control during rehabilitation. Some patients find the percutaneous stimulation of muscle to be uncomfortable, regardless of the intensity used. In addition, some patients do not like the concept of an artificial muscle contraction that they are not initiating and seemingly have no control over. Augmenting voluntary muscle contractions with biofeedback instead of ES, therefore, offers an optional training modality that may be more clinically integrative as well as more comfortable and appealing to the patient.

Perhaps the most favorable approach to postoperative muscle strengthening is an approach that uses a combination of these modalities. Protocols of this nature have been described in the literature, [26] and several training devices have been developed that offer a biofeedback-triggered ES mode. Augmenting voluntary exercise with a modality such as this would still require the patient to formulate motor strategies to initiate the contraction. Electrical stimulation could then be used, however, to enhance voluntary effort and provide a contraction that may be of even greater intensity than that achieved voluntarily at a given stage of rehabilitation.

Conclusions

Although ES has been a conventional choice as an adjunct to quadriceps femoris muscle strengthening exercises, the results of this study demonstrate that biofeedback can be used to facilitate a more rapid recovery of quadriceps femoris muscle peak torque than that achieved with ES alone and a comparable recovery of full knee extension. It is suggested that biofeedback may better facilitate the neural changes that accompany strength gains as compared with ES.

(*) Thought Technology Ltd, RR#1, Rte 9N, PO Box 380, West Chazy, NY 12992.

(1) Health Care Specialties Division/3M, 3M Center, St Paul, MN 55144.

(2) Cybex, Div of Lumex Inc, 2100 Smithtown Ave, Ronkonkoma, NY 11779.

V Draper, PhD, is Research Associate, Knoxville Orthopedic Clinic, 1128 Weisgarber Rd, Knoxville, TN 37909 (USA). Address correspondence to Dr Draper.

L Ballard, MS, ATC, is Staff Athletic Trainer, Orthopedic Rehabilitation Center, Knoxville Orthopedic Clinic.

This article was submitted July 3, 1989, and was accepted January 14, 1991.

References

[1] Noyes FR, Mangine RE, Barber S. Early knee motion after open and arthroscopic anterior cruciate ligament reconstruction. Am J Sports Med. 1987;15:149-159.

[2] de Andrade JR, Grant C, Dixon AStJ. Joint distension and reflex muscle inhibition in the knees. J Bone Joint Surg [Am]. 1965;47:313-322.

[3] Krebs DE, Staples W, Cuttita D, Zickle R. Knee joint angle: its relationship to quadriceps femoris muscle activity in normal and postarthrotomy limbs. Arch Phys Med Rehabil. 1983; 64:441-447.

[4] Delorme TL, Watkins AL. Progressive Resistance Exercise. New York, NY: Appleton-Century-Crofts; 1951.

[5] Moritani T, deVries HA. Neural factors versus hypertrophy in the time course of muscle strength gain. Am J Phys Med. 1979;58:115-129.

[6] Enoka RM. Muscle strength and its development: new perspectives. Sports Med. 1988; 6:146-168.

[7] Currier DP, Lehman J, Lightfoot P. Electrical stimulation in exercise of the quadriceps femoris muscle. Phys Ther. 1979;59:1508-1512.

[8] Laughman RK, Youdas JW, Garrett TR, Chao EYS. Strength changes in the normal quadriceps femoris muscle as a result of electrical stimulation. Phys Ther. 1983;63:494-499.

[9] Currier DP, Mann R. Muscular strength development by electrical stimulation in healthy individuals. Phys Ther. 1983;63:915-921.

[10] Selkowitz DM. Improvement in isometric strength of the quadriceps femoris muscle after training with electrical stimulation. Phys Ther. 1985;65:186-196.

[11] Eriksson E, Haggmark T. Comparison of isometric muscle training and electrical stimulation supplementing isometric muscle training in the recovery after major knee ligament surgery. Am J Sports Med. 1979;7:169-171.

[12] Delitto A, Rose SJ, McKowen JM, et al. Electrical stimulation versus voluntary exercise in strengthening thigh musculature after anterior cruciate ligament surgery. Phys Ther. 1988; 68:660-663.

[13] Wigerstad-6ossing I, Grimby G, Jonsson T, et al. Effects of electrical muscle stimulation combined with voluntary contractions after knee ligament surgery. Med Sci Sports Exerc. 1988;20:93-98.

[14] Lucca JA, Recchiuti SJ. Effect of electromyographic biofeedback on an isometric strengthening program. Phys Ther. 1983;63:200-203.

[15] Hald RD, Bottjen EJ. Effect of visual feedback on maximal and submaximal isokinetic test measurements of normal quadriceps and hamstrings. Journal of Orthopaedic and Sports Physical Therapy. 1987;9:86-93.

[16] Krebs DE. Clinical electromyographic feedback following meniscectomy: a multiple regression experimental analysis. Phys Ther. 1981;61:1017-1021.

[17] Sprenger CK, Carlson K, Wessman HC. Application of electromyographic biofeedback following medial meniscectomy: clinical report. Phys Ther. 1979;59:167-169.

[18] Draper V. Electromyographic biofeedback and the recovery of quadriceps femoris muscle function following anterior cruciate ligament reconstruction. Phys Ther. 1990;70:11-17.

[19] Murray DA, Harrison E, Wood GA. Cybex II reliability and validity: an appraisal. Med Sci Sports Exerc. 1982;14:153. Abstract.

[20] Reitz C, Rowinski M, Davies G. Comparison of Cybex II and Kin-Com reliability on the measures of peak torque, work and power at three speeds. Phys Ther. 1988;68:782. Abstract.

[21] Rothstein JM, Miller PJ, Roettger RF. Goniometric reliability in a clinical setting: elbow and knee measurements. Phys Ther. 1983; 63:1611-1615.

[22] Boone DC, Azen SP, Lin CM, et al. Reliability of goniometric measurements. Phys Ther. 1978;58:1355-1360.

[23] Sisk TD, Stralka SW, Deering MB, Griffin JW. Effect of electrical stimulation on quadriceps strength after reconstructive surgery of the anterior cruciate ligament. Am J Sports Med. 1987;15:215-220.

[24] Hettinger T. Physiology of Strength. Philadelphia, Pa: Charles C Thomas, Publisher; 1961:19-20.

[25] Shelbourne KD, Nitz P. Accelerated rehabilitation after anterior cruciate ligament reconstruction. Am J Sports Med. 1990;18:292-299.

[26] Fields RW. Electromyographically triggered electrical muscle stimulation for chronic hemiplegia. Arch Phys Med Rehabil. 1987;68:407-414.