Isokinetic, electrophysiologic, and clinical function relationships following tourniquet-aided knee arthrotomy.

Isokinetic, Electrophysiologic, and Clinical Function Relationships Following Tourniquet-aided Knee Arthrotomy 

Few data exist pertaining to the validity of isokinetic muscle function tests, particularly in describing their interrelationships with other common clinical assessments. The purpose of this study was to critically analyze the maximal voluntary knee torque, motor unit activity, range of motion, and gait sequelae in a series of 95 patients who had tourniquet-aided meniscectomy or intra-articular loose-body removal. Prior to arthrotomy, affected and contralateral knee data differed minimally. In the first days following arthrotomy, the patients invariably demonstrated severe gait and muscle mechanical impairments, with grossly abnormal quadriceps femoris muscle motor unit activity. Of 32 patients consenting to one-month postoperative diagnostic electromyographic and nerve conduction velocity testing, 17 (53%) were found to have femoral neuropathies, with 8 of the 17 having other thigh and leg neuropathies as well. Patients with neuropathy recovered more slowly and scored significantly lower on functional and electrophysiological measures than patients without neuropathy. One month postarthrotomy, knee extensor torques, motor unit activity, and gait scores averaged about half the normal values, and knee flexion torques and ROM were about 75% of normal. Isokinetic device measurements were moderately related to other clinical measures of post-arthrotomy outcome (.58 [is less than] r [is less than] .80). Knee torque measurements alone do not adequately characterize functional capacity. [Krebs DE: Isokinetic, electrophysiologic, and clinical function relationships following tourniquet-aided knee arthrotomy. Phys Ther 69:803-815, 1989]

Key Words: Kinesiology/biomechanics, lower extremity; Knee; Tests and measurements, functional; Tourniquets.

No published empirical evidence exists that isokinetic device assessments are meaningfully related to other clinical measures, despite their widespread use in physical therapy.[1] To assess the usefulness of isokinetic device measurements and to describe the interrelationships among commonly used clinical measures, I studied an empirical model of motor recovery involving patients who had tourniquet-assisted knee arthrotomy.

Pneumatic tourniquets are used in extremity surgery to provide a clear operative field, to allow visualization for ease of arthrotomy, and possibly to aid in prevention of iatrogenic infection.[2] Saunders and colleagues,[3] Krebs,[4] and Dobner and Nitz,[5] however, suggest that motor unit inhibition and frank neuropathy may be frequent sequelae of tourniquet-assisted knee arthrotomy. Knee muscle weakness immediately following arthrotomy is well known[4,6]; even after 6 to 12 months of rehabilitation, weakness of 10% to 20% on the involved side compared with the uninvolved side is present.[7] The wide range of outcomes following arthrotomy provides an excellent model for the investigation of common clinical measurements because the onset and cause of the impairments are known and the impairments will improve with time following the arthrotomy. Therapists typically assume that organ-level variables such as manual muscle test, isokinetic, and isometric measurements are correlated with one another and that they convey information about the patient's functional status.[1] For example, if isokinetic and MMT measures are highly related, then elaborate isokinetic measures may be redundant with this simpler clinical measure. If isokinetic values, however, are unrelated to accepted variables such as MMT values, gait competency, and time following surgery prior to return to work, then isokinetic data might contribute no useful information.[8] In short, isokinetic, electrophysiological, and clinical measures of organ-level and person-level postoperative function should be moderately correlated to justify their continued use by clinicians.

The purposes of the present investigation were to assess the short-term sequelae of tourniquet-aided knee arthrotomy and to describe the interrelationships among organ-level and person-level measures of lower extremity function. I hypothesized 1) that motor unit activity, knee range of motion, muscle mechanical competency, and gait would be at least moderately related (hypothesized Pearson product-moment correlation [r] = .60) and 2) that neuropathy would adversely affect organ-level and person-level performance measures.

Method

Subjects

A total of 95 serially referred patients (74 male, 21 female), ranging in age from 15 to 75 years (X[Bar] = 35.2 years), were studied (Tab. 1). The assessment sessions were conducted in two general phases: 1) Inpatient assessments were performed immediately following arthrotomy while the patients were still hospitalized (n = 94), and 2) outpatient assessments were performed 1 to 14 days (X[Bar] = 3 days) prior to arthrotomy (n = 14) and then beginning 2 to 5 weeks (X[Bar] = 4 weeks) after arthrotomy (n = 38). One patient was assessed only postoperatively as an outpatient. All patients gave informed consent prior to participation.

Sixty-three patients had medial meniscectomies, and 18 patients had lateral meniscectomies. Fourteen patients had intracapsular chondroid or osteoid loose bodies removed. Two patients developed pulmonary emboli following arthrotomy and were subsequently eliminated from the study. All patients were otherwise healthy, and no patient had a history of peripheral neuropathy or of neuropathogenic disorders such as diabetes mellitus or alcohol abuse.

Five attending orthopedic surgeons in the same university-affiliated hospital performed the arthrotomies using an 8.5-cm pneumatic tourniquet that was inflated to 350 to 450 mm Hg and applied in the proximal third of the thigh. Tourniquet applications ranged from 14 to 98 minutes (X[Bar] = 60.6 minutes, s = 18.4). Immediately following arthrotomy, each patient's affected limb was snugly wrapped to the mid-thigh with elastic bandages to minimize leg and knee joint edema.

Inpatient Assessments

Ninety-four patients were studied during hospitalization, beginning the first to the seventh day (X[Bar] = 1.9 days) following arthrotomy. Electromyographic, MMT, and ambulatory weight-bearing assessments were similar to those described in a previous article.[4]

During inpatient EMG analyses, patients were positioned supine with the knee at 0 degrees of flexion. Skin resistance was reduced by abrasion and alcohol. The 1.5-cm surface electrodes were separated by 0.5 cm and placed 10 to 15 cm proximal to the lateral border of the medial femoral condyle over the quadriceps femoris muscle. Electrode locations were recorded and marked on the skin to allow precise placement replication on ensuing days. A linear-envelope EMG signal was obtained by low-pass (20 Hz) filtering and rectifying the raw EMG data with a Cyborg J33 portable EMG unit,(*) which could be taken to the patient's bedside if necessary. The maximum smoothed and rectified electromyographic (srEMG) amplitude held for 2 consecutive seconds was recorded from five 5- to 10-second isometric contractions, with a 30- to 60-second rest interval between contractions and without feedback to the patient.

A break-test MMT was conducted with the patient sitting and the knee at 0 degrees at flexion. The numerical scoring scale is shown in Table 2.

Patients then performed isometric quadriceps femoris muscle exercise for 20 minutes and were allowed to use EMG machines for feedback.[4] The electrodes remained in place throughout the 20-minute treatment session. The srEMG and MMT assessments were repeated at the end of the 20-minute exercise period.

Following the exercise periods, patients practiced straight leg raising, if the patient was able, for 10 minutes. Crutch-walking instruction ensued, allowing weight bearing to tolerance. The numerical scoring scale for weight-bearing tolerance is given in Table 2.

Outpatient Assessments

For 14 consecutive months, participating orthopedists were contacted biweekly for the names and telephone numbers of patients scheduled for knee arthrotomy. Fourteen of 20 patients contacted consented to participate in preoperative assessments. Thirty-eight of the 43 patients contacted postarthrotomy consented to postoperative assessments (Tab. 1).

All outpatient tests were performed both preoperatively and postoperatively, except the questionnaire interview, which was completed only postoperatively. If the postoperative diagnostic electromyogram or nerve conduction velocity (EMG/NCV) results were abnormal, outpatient assessments were repeated in three weeks.

The EMG/NCV assessments and the functional assessments were performed on the same day. Functional assessments were performed without knowledge of the EMG/NCV outcomes and vice versa. During the functional studies, the unaffected limb was always tested first to maximize patients' familiarity with the tests and equipment and thus to reduce error variability in the scores of the affected extremity, which were of primary interest.

Electromyogram and nerve conduction velocity. All EMG/NCV assessments were performed with a TECA TE4 electromyograph([single dagger]) by one investigator who was blind to the patients' tourniquet time. Teflon([double dagger])-coated 27-gauge monopolar needle electrodes were used in an EMG analysis of the quadriceps femoris, hamstring, triceps surae, anterior tibialis, and lumbar paraspinal muscles. The EMG results were judged abnormal in the presence of any spontaneous resting potentials (ie, fibrillations, positive sharp waves, bizarre high frequency discharges) and increased insertional activity.[9] Femoral, peroneal, and tibial motor and sural sensory NCV tests were performed using standard techniques; H-reflex and F-wave peroneal and tibial NCV tests were conducted using the method described by Johnson.[9]

To further pursue the tourniquet's role in the genesis of postarthrotomy neuropathy, a special EMG test was performed on two patients over 6 ft([section]) tall who had very long rectus femoris muscles with positive EMG findings. Needles were placed in the proximal third of the rectus femoris muscle just distal to the inguinal ligament, at the junction of the proximal and middle thirds (at the level of the tourniquet application), and in the distal third. Increased insertional activity, fibrillations, and positive sharp waves were present at and below the level of the tourniquet application, but not proximal to the area of tourniquet application during arthrotomy.

Motor unit activity. An isometric 2-second, maximal effort srEMG test of the quadriceps femoris muscle was performed identically to those of the inpatient evaluations. A 10-second integration of the root mean square (RMS)-averaged EMG signal (IEMG) during maximal quadriceps femoris muscle contraction was then performed using a Cyborg BL900 electromyograph(*) to better characterize the quadriceps femoris muscle's motor unit recruitment over prolonged contractions.

Muscle mechanical competency. Manual muscle testing of the quadriceps femoris and hamstring muscles, using a break test, was conducted and recorded as described previously. Knee flexor and extensor torque assessments were conducted according to the methods described by Mira et al.[10] A Cybex [R] Orthotron [TM] isokinetic dynamometer([~~]) was used to determine the maximum dynamic torque of the quadriceps femoris and hamstring muscles between 0 and 90 degrees of knee flexion at 60 [degrees] and 105 [degrees]/sec. Isometric knee extension and flexion peak torques were assessed by locking the dynamometer lever arm at 20, 45, and 60 degrees of knee flexion. Each test was performed three times, with the hip flexed to 80 degrees.[11] A 5- to 10-minute rest period was allowed between tests. The maximum torque resulting from each effort was recorded. An additional isokinetic test was conducted to assess the quadriceps femoris muscle's endurance by counting the number of successive repetitions performed (without rest between repetitions) at 105 [degrees]/sec, which exceeded one half the maximum quadriceps femoris muscle torque previously recorded at the same speed. The dial indicator from which torque values were read was mechanically damped (restrained) to prevent inertial overshoot relative to the actual torque values. The isokinetic device was calibrated weekly during the investigation period. At no time during the study was the device found to be out of calibration.

Range of motion. Knee flexion active range of motion (AROM) and passive range of motion (PROM) were determined with a goniometer while the patient was positioned supine and the hip flexed to 90 degrees. Knee extension ROM was determined with the patient sitting and the hip flexed 80 degrees.

Gait. Ambulation quality on smooth level surfaces, stair climbing, hopping, and squatting were scored according to the criteria listed in Table 3. The scores were summed and divided by 4 to obtain an average "Gait" score for each patient.

Data Analysis

Passive-range-of-motion, srEMG, IEMG, and isokinetic and isometric torque values were standardized for each patient by dividing the affected limb's value by the homologous value obtained from the unaffected limb, thus forming a ratio. The PROM, srEMG, and IEMG ratio thus obtained, multiplied by 100, equals percentage of normal (% PROM, % srEMG, and % IEMG) for each patient. The two isokinetic and three isometric knee extensor muscle ratios were summed, divided by 5, and multiplied by 100% to yield a composite quadriceps femoris muscle function index (QFI); the five knee flexor torque ratios were treated similarly to derive the hamstring muscle function index (HFI).[10] The alpha level was set at .05 unless otherwise noted in the text. Multivariate statistical methods were used that controlled chance variability in the dependent variables because multiple comparisons on correlated dependent variables may artificially inflate the alpha level.(12) The general linear model(12,13) was used for between-variable comparisons for both univariate and multivariate analyses. This statistical model allows the effect of either categorical or continuous variables to be independently or simultaneously assessed and permits valid comparisons of groups with unequal numbers of members. All analyses were performed on an IBM Model 4341 computer(#) using the Statistical Analysis System.[13] Because of the large number and relatively normal distribution of data points, parametric statistics could be used, thereby increasing the statistical power of the investigation.[13]

Reliability of the dependent measures was assessed by correlating the pre-operative and postoperative values from the normal limb of patients who had normal postoperative EMG/NCV results. Patients with abnormal EMG/NCV results had significant normal-limb variables' mean differences, both relative to their preoperative performance and to other subjects, possibly because their activity patterns differed from those of the cohort without neuropathy; therefore, only those patients with normal EMG/NCV results participated in the reliability assessments. A minimum sample size of 19 was needed for the specific hypothesis tests to detect a statistically significant correlation of .60 with a level of probability of .05 and a power of .80.

Results

Inpatient Assessments

A total of 222 inpatient visits were analyzed. Preexercise and postexercise results from the first four inpatient days are summarized in Table 2. Only five patients remained hospitalized longer than four days. The most striking finding of the inpatient assessments was the profound weakness of the quadriceps femoris muscle (Tab. 2). On the first inpatient assessment after arthrotomy, but prior to exercise, no patient had greater than a Fair MMT value, and all patients had severe impairments of maximal srEMG amplitude ([is less than] 30 [Micro] V). By the end of the second or third day's exercise treatment, most patients could extend their knee against gravity and could perform straigh-leg-raising exercises. Both motor unit activity and muscle force improved more during the exercise period than during the ensuing 24 hours (Tab. 2). Postexercise srEMG and MMT results were correlated significantly, but rather weakly (r [is less than] .04, p [is less than] .05), with lenght of time after surgery. Measurements of motor unit activity and muscle "strength" (srEMG and MMT results, respectively) were correlated significantly (r = .60, p [is less than] .001), both before and after exercise. Weight-bearing tolerance varied directly with length of time after surgery (r = .50, p [is less than] .000001), with preexercise and postexercise MMT results (r = .58 and .64, respectively; p [is less than] .000001), and with srEMG results (r = .39 for both preexercise and postexercise measurements, p [is less than] .0001). An analysis of variance (ANOVA) of inpatient srEMG results from patients on whom EMG/NCV tests were performed post-operatively revealed both relative (Fig. 1) and absolute (Fig. 2) motor unit activity impairment of the group with neuropathy (F [is greater than] 4.8, p [is less than] .01).

Outpatient Assessments

Table 4 lists the major sample characteristics and outpatient data from the functional and electrophysiological assessments for the 38 patients seen postoperatively. No significant differences existed between the patients with and those without preoperative examinations regarding age, sex, and tourniquet application durations; arthrotomy type or location; and post-operative QFI, HFI, Gait, ROM, or electrophysiological assessments (p [is greater than] .15). The preoperative data, therefore, may be representative of the entire sample's data. Sixty outpatient visits were analyzed.

Electromyogram and nerve conduction velocity. Three of the 14 patients seen preoperatively refused diagnostic EMG/NCV testing. None of the 11 preoperative EMG/NCV tests showed any abnormalities. One of the 11 patients on whom preoperative EMG/NCV tests were performed developed a postoperative pulmonary embolism and was subsequently dropped from the study. All other patients seen preoperatively had post-operative EMG/NCV analyses. (One patient, a 31-year-old man who was seen 3 days as an inpatient and whose tourniquet time was 56 minutes, was found to have quadriceps femoris muscle neuropathy 25 days following arthrotomy. A family emergency developed, however, forcing his departure prior to further outpatient assessment. His EMG/NCV results are included in this section of the article only and are not included in Tab. 4.)

The 32 EMG/NCV tests performed an average of four weeks postarthrotomy revealed 17 patients (53%) had quadriceps femoris muscle abnormalities indicative of neuropathy; 15 patients had normal EMG/NCV results. No patients had paraspinal or contralateral limb abnormalities. Repeat EMG/NCV tests were performed on eight patients who had positive findings (Tab. 1). By the ninth postoperative week, 3 of the 8 EMGs remained abnormal, whereas the others no longer showed signs of neuropathy. The patients with positive findings were asked to return for another follow-up. One patient complied; his EMG/NCV results at 12 weeks were normal.

Motor unit activity. Preoperatively, both the 2-second maximum amplitude srEMG signal and the 10-second IEMG signal were approximately equal in the affected and the unaffected limbs (%srEMG = 102% [+ or -] 20%, %IEMG = 105% [+ or -] 20%). Four weeks after surgery, the %srEMG and %IEMG values were 46% [+ or -] 5% and 50% [+ or -] 5%, respectively.

Both preoperatively and postoperatively, the unstandardized srEMG and IEMG test results were well related within the affected as well as the unaffected limb (r = .80, p [is less than] .001). The standardized electrophysiological variables (%srEMG and %IEMG) were also highly related when compared both prior to and following arthrotomy (r [is greater than] .95, p [is less than] .001).

Patients with neuropathic EMG/NCV findings had significantly lower %srEMG and %IEMG (F = 15.1 and 24.3, respectively; p [is less than] .0001) than those with normal EMG/NCV values. The affected quadriceps femoris muscle unstandardized IEMG and srEMG values were also significantly lower in patients with neuropathy (F = 13.2, p [is less than] .001). Quadriceps femoris muscle motor unit activity improved at a faster rate following surgery than before surgery in patients without femoral neuropathy (Fig. 2).

The preoperative unaffected quadriceps femoris muscle srEMG and IEMG values were highly related to postoperative values in patients with normal EMG/NCV levels (r = .80 and .93, respectively; p [is less than] .01). Therefore, srEMG and IEMG tests appear to be reliable measures of quadriceps femoris muscle motor unit activity.

Muscle mechanical competency. Figure 3 depicts the relationship of the unaffected to the affected limb's knee extensor torques prior to and following arthrotomy. Preoperatively, the affected knee extensor torques (QFI) were 77% [+ or -] 11% of the contralateral limb values, whereas one month postoperatively the QFI was 46% [+ or -] 4%. The patients from the first postoperative examination who had abnormal EMG/NCV findings and who returned for repeat tests improved their extensor torques only to 59% [+ or -] 11% of normal values at 9 weeks postarthrotomy. No significant interactions of the affected and unaffected knee torques were found among the isokinetic speeds or the isometric joint angles. That is, although the postarthrotomy knee produced significantly less torque, the measured values at a given isokinetic speed or isometric knee angle tended to rank-order similarly for both the affected and unaffected limbs. Figures 4 and 5 depict the relationships among the four-week postoperative knee flexion and extension torques. Multivariate and univariate ANOVAs demonstrated significant neuropathic impairments in all torque data (p [is less than] .05). No significant difference attributable to neuropathy was found in the knee extensor endurance test (p [is greater than] .20).

The postarthrotomy limb's individual extensor torques--the components of the QFI--were correlated (r = .50--.80, p [is less than] .001) with the srEMG and IEMG test results. No significant correlations, however, were found for the normal limb.

The isometric extensor torque ratio at 60 degrees of flexion correlated most strongly, in comparison with the other isometric and isokinetic components of the QFI, the EMG/NCV results, HFI, %PROM, %srEMG, %IEMG, and the Gait assessments (.60 [is less than or equal to] r [is less than or equal to] .90; p [is less than] .001). The isometric knee extensor torque at 60 degrees of flexion correlated .92 with the QFI combined score.

Patients with neuropathic quadriceps femoris muscles had significantly lower (p [is less than] .01) MMT scores than those with normal EMG/NCV results (X[bar] = 3.7 [+ or -] 0.5 and 4.2 [+ or -] 0.6, respectively). The affected quadriceps femoris muscle postoperative MMT values correlated moderately (p [is less than] .001) with the affected limb's PROM (r = .65) and motor unit recruitment (r = .60 for both srEMG and IEMG test results), as well as the isometric (.60 [is less than] r [is less than] .75) and isokinetic values (r = .69 and .58 for the 60 [degrees] and 105% sec torques, respectively).

Table 5 lists the preoperative and postoperative correlations from patients with normal EMG/NCV values for the individual components of the QFI and the HFI. Although Figure 3 shows slight mean differences between the preoperative and postoperative normal limb torques, neither the univariate nor the multivariate ANOVAs showed any significant differences, either overall or among the individual components. Apparently, QFI and HFI are reliable measures in this sample.

Range of motion. The operated knee-to-nonoperated knee PROM ratio (%PROM) was 98% [+ or -] 1% preoperatively and 77% [+ or -] 3% postoperatively; the eight patients who returned for the follow-up postoperative assessment regained 95% [+ or -] 2% of the normal knee's PROM by nine weeks postarthrotomy.

Affected limbs with normal EMG/NCV results averaged 124 [+ or -] 21 degrees of flexion, whereas the neuropathic cohort could flex the affected knee to only 102 [+ or -] 26 degrees (p [is less than] .01). The ratios of the affected limb's to the unaffected limb's PROM values for each patient are listed in Table 4. Preoperatively, the affected knee lacked an average 1.5 [+ or -] 1 degrees of full passive extension and 4 [+ or -] 7 degrees of full active extension. At the first postoperative outpatient examination, the knee came to within an average of 0.8 [+ or -] 0.5 degree of full passive extension and 1.0 [+ or -] 0.6 degree of full active extension.

The normal limb's PROM and AROM preoperative to postoperative correlations in patients with normal postoperative EMG/NCV results were very high (r [is greater than] .90); thus, PROM and AROM measurements are probably reliable in this study.

Gait. Prior to arthrotomy, the Gait assessment score (Tab. 3) averaged 3.9 [+ or -] 0.4. At four weeks postarthrotomy, Gait assessment scores averaged 2.7 [+ or -] 0.3, or 54% of the maximum score. Table 4 reveals that the four-week Gait profile was normal (ie, 5.0) in only two patients. Those patients (Patients 7 and 35) also ranked highly in motor unit activity, knee flexor and extensor mechanical performance, and ROM. Indeed, the overall postoperative Gait scores were moderately related to QFI, HFI, %PROM, and %srEMG (r = .75, .71, .74, and .50, respectively; p [is less than] .001) (Fig. 6).

Gait was significantly affected by neuropathy. Patients with positive EMG/NCV results averaged 2.0 [+ or -] 0.3, and the cohort without neuropathy averaged 3.4 [+ or -] 0.3 (F = 7.9, p [is less than] .01).

Discriminant analysis. An omnibus stepwise discriminant analysis to "predict" neuropathy, given the patient's standardized and unstandardized QFI, HFI, Gait, %PROM, and %IEMG scores was performed. The QFI discriminated most efficiently among the EMG/NCV outcomes (Wilk's Lambda = .65, p [is less than] .001, priors proportional). Data easily obtainable in a clinic (ie, MMT and PROM values) were not significant discriminators of neuropathy.

Effect of standardization. No sex differences were found in the multivariate ANOVA of standardized electrophysiological and functional data (p [is greater than] .12). On the unstandardized measures, however, the male patients generated more torque (Tab. 6) and more EMG activity than the female patients. The male patients' affected limb IEMG and srEMG values exceeded those of the female patients by over 350%; the male patients' contralateral quadriceps femoris muscle values were over 200% greater than those of the female patients (F [is greater than] 10.0, p [is less than] .001).

Discussion

Isokinetic, electrophysiologic, and clinical variables apparently measure some common factors following arthrotomy, but each is sensitive to unique phenomena. In particular, knee extensor torques were significantly, but mildly, related to gait function. Thus, measures of person-level functional capacity such as gait proficiency probably depend on a number of factors including, but not limited to, muscle force production and ROM.[14]

Isokinetic Measurement Validity

Knee torque measurements appear to be reliable, moderately useful indicators of person-level functional status. Less than two thirds of the variability in other clinical measures could be attributed to any combination of isometric and isokinetic torques (r [is less than or equal to] .80). As indicators of organ-level function (ie, muscle mechanical competency), however, the peak torque values were more useful than MMT values; peak torque more efficiently discriminated EMG/NCV outcomes. Dynamometric data are frequently reported without reference to clinical meaningfulness. The present data are especially important because they indicate that isokinetic device measurements must be supplemented with other clinical outcomes to adequately gauge functional capacity.

Peak torques do apparently provide a useful indication of muscle (organ-level) mechanical impairment. Motor unit activity (%srEMG) and knee extensor torque (QFI) were moderately related (r = .72) and equally depressed (46% of the normal limb values) one month following tourniquet-aided arthrotomy, although they were nearly normal (ie, roughly equivalent bilaterally) prior to surgical intervention. These results support the commonly held belief that muscle torque depends on motor unit integrity.[3-5]

In this study, isometric torque data provided as much information, or more, about person-level clinical measures as isokinetic torque. In particular, isometric torque at 60 degrees of knee flexion accounted for 85% of the variance (information) contained in QFI. Lankhorst et al studied patients with knee arthritis using isokinetic and isometric measures similar to those of the present study and drew several similar conclusions.[15] If clinicians are unable to measure any other postoperative outcome, therefore, most information is gleaned from measuring isometric knee extensor torque at 60 degrees of flexion.[10,11,16,17] Only concentric, relatively slow angular velocities, however, were tested. Low-velocity torque productivity may be more closely related to isometric output than high angular velocity[18] or eccentric force production. Furthermore, gravity correction,[19] measures of power, and shape of the torque output curve were not analyzed in this study, so it is possible that more sophisticated isokinetic measurements would reveal different relationships. The present literature, however, indicates that these measurements yield no more information than simply examining peak torques.[20,21]

No significant neuropathy effect was found in knee extensor muscle endurance, perhaps because of the muscle fiber type most affected by neuropathy. Using electron microscopy and electrophysiological techniques on rat muscle, Pachter et al concluded, "Type II [fast] fibers were most affected by the denervation process."[22] A predominance of type II fibers in the quadriceps femoris muscle is highly related to differentially better high-speed isokinetic performance than to endurance and slow-speed characteristics.[23] The present findings, however, may also be explained more parsimoniously by the "fatigue" test's insensitivity.[24] Although no fiber-typing studies were performed on these patients, selective loss of type II fibers would explain some of the functional impairments (eg, Gait) demonstrated in these patients. Many functional activities, including ambulation, require maximum knee movement faster than 105 [degrees]/sec.

Overall, knee flexor torques were relatively less affected than knee extensor torques. The affected limb's flexors could exert an average of 58% of the normal side's isokinetic torque, whereas the comparable extensor torque values were about 30% of the contralateral homologue. The pattern of isometric values was symmetrical. The affected limb's flexor torques were in 70% range; the extensor isometric values were about 15% lower, in the 55% range. Data from the study of Campbell and Glenn indicate that this differential flexor and extensor impairment persists, to a lesser degree, up to one year postoperatively, even when patients engage in 6 to 20 weeks of isokinetic exercise.[7] This reported postarthrotomy flexor and extensor differential impairment is consistent with the present data, demonstrating that most of the patients undergoing tourniquet-aided arthrotomy have knee extensor, but not knee flexor, neuropathies.

Electrophysiological Measurements

The relationship of %srEMG and %IEMG to other clinical variables, including time after surgery and presence or absence of femoral neuropathy (Fig. 1), indicates that clinicians should probably evaluate postarthrotomy patients with the simple srEMG unit to help predict those patients who may require follow-up EMG/NCV diagnosis and more extensive therapy. If these results can be replicated in other clinical settings, a patient whose quadriceps femoris muscle electrical activity improves little with exercise and whose absolute electrical activity is low should probably be monitored carefully for other neuropathic signs. This within-subject srEMG estimate of change in motor unit activity, at least in patients with severely paretic quadriceps femoris muscles, appears to reflect real differences in quadriceps femoris muscle function. Because the relationship of electrical activity to other neuromuscular factors is in part a function of the recording apparatus, electrode positions, and skin preparation,[25] these results only imply that some relationship exists between srEMG, as measured in this study, and muscle mechanical competency. Given the large number of subjects in this study and the observed moderate relationship between srEMG and commonly accepted clinical variables (MMT, days postsurgery, weight bearing), it appears that motor unit activity measured by srEMG is a reasonably valid clinical indicator of muscle function during the immediate postoperative period and, if interpreted cautiously, can contribute meaningfully to clinical decisions.

The results of this study, however, do not indicate that srEMG values are valid functional indicators for healthy subjects. The error bars in Figure 2 reveal that srEMG variability increased as amplitude approached normal values. Healthy individuals share this wide intersubject electrical variability.[26]

Electromyographic and Nerve Conduction Velocity Assessments

Overall, it appears that although tourniquet-aided arthrotomy may lead to neuropathy in some patients, nearly all patients have motor unit recruitment abnormalities; thus, reduced muscle function is not entirely attributable to neuropathy. The EMG/NCV tests for neuropathy apparently are not sensitive to the full range of motor unit abnormalities present in patients postarthrotomy. A nonneuropathic mechanism (hence, sparing of EMG/NCV abnormalities), perhaps within the affected muscle fibrils,[27-31] may be responsible for the suppression of srEMG and IEMG amplitude. For example, in Patient 9 (Tab. 4), the 87% depression of motor unit activity in the presence of normal EMG/NCV levels is difficult to explain unless the inability to recruit motor units results from some non-neuropathic muscle disorder.

Gait Assessment

The contribution of organ-level improvements to patient-level improvements was apparently modest in this study. The correlations indicate that 25% to 50% of the variance in postoperative Gait performance may be attributable to motor unit behavior, muscle mechanical competency, and knee mobility. Thus, efforts to improve muscular force or endurance alone may not necessarily improve patient-level function.[1]

Standardization

Electrophysiological and isokinetic measurements were most meaningful when standardized to the individual's sound limb in this study. Other standardization methods, such as comparison with sex, weight, and age-matched normative data, may be preferable, particularly given that the contralateral limb may also have been impaired. Such normative data currently are not available.

Utility of Arthrotomy Model

Tourniquet-aided knee arthrotomy apparently provides an excellent paradigm of knee pathology that improves steadily with time. The adverse post-operative sequelae were independent of sex, age, and arthrotomy type or location. Pain alone is also an unlikely explanation for the relationships found. Objective tests not requiring patient effort (ie, EMG/NCV and PROM) were affected substantially similarly as were those tests requiring patient effort. Furthermore, little drop-out in motor unit activity was found during the 10-second contraction compared with the 2-second contraction, arguing against pain causing decreased effort. Pain should increase the variability in motor unit activity during prolonged contractions. Intra-articular edema is also an unlikely source of postarthrotomy paresis. Bryan et al stated, "Data ... revealed no benefit from suction drainage of the joint space of the knee after meniscectomy, as measured by recovery of strength and flexion."[6] The present data are consistent with prior studies indicating that pneumatic tourniquets cause muscle and nerve impairments.[3,5] The effects of neuropathy on motor unit activity and muscle mechanical function are supported by previous physiological investigations.[17,32,33] Further study, however, is needed using patients with other neuropathic disorders, such as diabetic neuropathy, Guillain-Barre syndrome, and multiple sclerosis, to determine whether the observed interrelationships among clinical measures such as motor unit activity, muscle mechanical competence, and gait performance are consistently present in other populations.

Conclusions

The moderate relationship of isokinetic variables to other clinical measures in this study indicates that isokinetic data alone do not adequately evince functional recovery. For example, because normal gait function requires rapid and forceful muscle contraction, normal motor unit activity, and adequate knee ROM and motor control skills, which are not adequately characterized by dynamometry alone, isokinetic data should be supplemented with other clinically meaningful measures. Manual muscle testing appears to be a less sensitive measure of muscle mechanical competency than isometric and isokinetic dynamometry. During the immediate postoperative period, srEMG appeared to be quite sensitive to functional changes, but as motor unit recruitment approached normalcy, specificity decreased. Of the organ-level measures assessed in the "cross-sectional," limited-time-period study, isometric and isokinetic dynamometry were the best predictors of person-level (eg, gait) function. Future longitudinal studies should compare functional gait and activities-of-daily-living torque requirements to isokinetic and isometric torque measurements. [Tabular Data 1 to 6 Omitted] [Figures 1 to 6 Omitted]

(*)Cyborg Corp, 55 Chapel St, Newton, MA 02158. ([single dagger])TECA Instruments Corp, 3 Campus Dr, Pleasantville, NY 10570. ([double dagger])E I du Pont de Nemours & Co Inc, 1007 Market St, Wilmington, DE 19898. ([section])1 ft = 0.3048 m. (~~)Cybex, Div of Lumex, Inc, 2100 Smithtown Ave, Ronkonkoma, NY 11779. (#)International Business Machines Corp, 1000 NW 51st St, Boca Raton, FL 33432.

References

[1]Mayhew TP, Rothstein JM: Measurement of muscle performance with instruments. In Rothstein JM (ed): Measurement in Physical Therapy: Clinics in Physical Therapy. New York, NY, Churchill Livingstone Inc, 1985, vol 7, pp 57-102 [2]Crenshaw AH: Surgical approaches. In Edmerson AS, Crenshaw AH (eds): Campbell's Operative Orthopedics. St Louis, MO, C V Mosby Co, 1988, pp 42-46 [3]Saunders KC, Louis DL, Weingarden SI, et al: Effect of tourniquet time on postoperative quadriceps function. Clin Orthop 143:194-199, 1979 [4]Krebs DE: Clinical electromyographic feedback following meniscectomy: A multiple regression experimental analysis. Phys Ther 61:1017-1021, 1981 [5]Dobner JJ, Nitz AJ: Postmeniscectomy tourniquet palsy and functional sequelae. Am J Sports Med 10:211-214, 1982 [6]Bryan RS, Dickingon JH, Taylor WF: Recovery of the knee following meniscectomy. J Bone Joint Surg [Am] 51:973-978, 1969 [7]Campbell DE, Glenn W: Rehabilitation of knee flexor and knee extensor muscle strength in patients with meniscectomies, ligamentous repairs, and chondromalacia. Phys Ther 62:10-15, 1982 [8]Krebs DE, Malgady RG: Understanding correlation coefficients and regression. Phys Ther 66:110-120, 1986 [9]Johnson EW: Practical Electromyography. Baltimore, MD, Williams & Wilkins, 1980, pp 46-52 [10]Mira AJ, Markley K, Green RB: A critical analysis of quadriceps function after femoral shaft fracture in adults. J Bone Joint Surg [Am] 62:61-67, 1980 [11]Currier DP: Positioning for knee strengthening exercises. Phys Ther 57:148-152, 1977 [12]Harris RJ: A Primer of Multivariate Statistics. New York, NY, Academic Press Inc, 1975, pp 13, 156-194 [13]Statistical Analysis System. Cary, NC, SAS Institute, Inc, 1979 [14]Griffin JW, McClure MH, Bertorini TE: Sequential isokinetic and manual muscle testing in patients with neuromuscular disease. Phys Ther 66:32-35, 1986 [15]Lankhorst G, Von de Stadt R, Vanderkorst JK: The relationships of functional capacity, pain, and isometric and isokinetic torque in osteoarthritis of the knee. Scand J Rehabil Med 17:167-172, 1985 [16]Goslin BR, Charteris J: Isokinetic dynamometry: Normative data for clinical use in lower extremity (knee) cases. Scand J Rehabil Med 11:105-109, 1979 [17]Smidt GL: Biomechanical analysis of knee flexion and extension. J Biomech 6:79-92, 1973 [18]Knapik JJ, Ramos MU: Isokinetic and isometric torque relationships in the human body. Arch Phys Med Rehabil 61:64-67, 1980 [19]Herzog W: The relation between the resultant moments at a joint and the moments measured by an isokinetic dynamometer. J Biomech 21:5-12, 1988 [20]Hoke B, Howell D, Stack M: The relationship between isokinetic testing and dynamic patellofemoral compression. Journal of Orthopaedic and Sports Physical Therapy 4:150-153, 1983 [21]Rothstein JM, Delito A, Sinacore DR, et al: Electromyographic, peak torque, and power relationships during isokinetic movements. Phys Ther 63:926-933, 1983 [22]Pachter BR, Eberstein A, Goodgold J: Electrical stimulation effect on denervated skeletal myofibers in rats. A light and electron microscopic study. Arch Phys Med Rehabil 63:427-430, 1982 [23]Thorstensson A: Muscle strength, fiber types and enzyme activities in man. Acta Physiol Scand [Suppl] 443:7-43, 1976 [24]Rothstein JM, Lamb RL, Mayhew TP: Clinical uses of isokinetic measurements: Critical issues. Phys Ther 67:1840-1844, 1987 [25]Basmajian JV, DeLuca CJ: Muscles Alive: Their Functions Revealed by Electromyography, ed 5. Baltimore, MD, Williams & Wilkins, 1985, pp 187-200 [26]Krebs DE, Staples WH, Cuttita DM, et al: Knee joint angle: Its relationship to quadriceps femoris activity in normal and post-arthrotomy limbs. Arch Phys Med Rehabil 64:441-447, 1983 [27]Ochoa J, Fowler TJ, Gilliatt RW: Anatomical changes in peripheral nerves compressed by a pneumatic tourniquet. J Anat 113:433-455, 1972 [28]Dahlback LO: Effects of temporary tourniquet ischemia on striated muscle fibers and motor end-plates. Scand J Plast Reconstr Surg Suppl 7, 1970 [29]Nitz AJ, Matulionis DH: Ultrastructural changes in rat peripheral nerve following pneumatic tourniquet compression. J Neurosurg 57:660-666, 1982 [30]Patterson S, Klenerman L: The effect of pneumatic tourniquets on the ultrastructure of skeletal muscle. J Bone Joint Surg [Br] 61:178-183, 1979 [31]Solonen KA, Hjelt L: Morphological changes in striated muscle during ischemia. Acta Orthop Scand 39:13-19, 1968 [32]Baranski S, Kwarecki K, Szmigielski S, et al: Histochemistry of skeletal muscle fibers in rats undergoing long-term experimental hypokinesia. Folia Histochem Cytochem 9:381-386, 1971 [33]Edstrom L: Selective atrophy of red muscle fibers in the quadriceps in long-standing knee joint dysfunction. J Neurol Sci 11:551-558, 1970

D Krebs, PhD, PT, is Associate Professor, MGH Institute of Health Professions, 15 River St, Boston, MA 02108-3402. The empirical data reported were collected at St Lukes Hospital, New York, NY.