Journal of Orthopaedic & Sports Physical Therapy 2OOl;3l (l):&l5

Anterior Tibial Translation During Different Isokinetic Quadriceps Torque in Anterior Cruciate Ligament Deficient and Nonimpaired Individuals
)oanna Kvist, RP7; PhD1 Christian Karlberg2 Bjorn Gerdle, MD, PhD3 )an Gillquist, MD, PhD4
Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Study Design: Factorial quasi-experimental design. Objectives: To quantify the effect of different levels of isokinetic concentric and eccentric knee extensor torques on the anterior tibial translation in subjects with anterior cruciate ligament (ACL) deficiency. Electromyogram (EMG) activity of 4 leg muscles was recorded in order to detect any co-activation of extensors and flexors. Background: The rehabilitationafter an ACL injury is of importance for the functional outcome of the patient. In order to construct a rehabilitation program after that injury, it is important to understand the in vivo relationships between muscle force and tibial translation. Methods and Measures: Twelve patients with unilateral ACL injury and 11 uninjured volunteers performed 36 repetitions of a quadriceps contraction at different isokinetic concentric and eccentric torque levels, on a KinCom machine (60"-s-I), with simultaneous recordings of tibial translation (CA-4000) and EMG activity from quadriceps and hamstrings muscles. Tibial translations and EMG levels were normalized to the maximum of each subject. Results: The individual anterior tibial translation increased with increased quadriceps torque in a similar manner in both quadriceps contraction modes in all legs tested. During concentric mode, translation was similar in all groups, but during eccentric mode, the mean translation was 38% larger in the ACL injured knees. No quadriceps-hamstringsco-activation occurred in any test or group. Conclusions: An ACL deficient knee can limit the translation within a normal space during concentric muscle activity but not during eccentric activity. That limitation depends on other mechanisms than hamstrings co-activation. ] Orthop Sports Phys Ther 2001;31:4-15.

Key Words: concentric contraction, eccentric contraction, knee kinematics, rehabilitation

Spom Medicine and Physical Therapy, Department of Neuroscience and Locomotion, Faculty of Health Sciences, Linkoping University, SE-581 85 Linkoping Sweden. Research engineer, Rehabilitation Medicine, Department of Neuroscience and Locomotion, Faculty of Health Sciences, linkoping University, SF-581 85 linkoping Sweden. Rehabilitation Medicine, Department of Neuroscience and locomotion, Faculty of Health Sciences, Linkoping Universitp SF-581 85 Linkoping Sweden. Sports Medicine, Department of Neuroscience and locomotion, Faculty of Health Sciences, linkoping University,SF-581 85 Linkoping, Sweden. Study was approved by the ethical committee at Linkoping University, Sweden and partially funded by the Swedish Foundation for Health Care Sciences and Allergy Research "Vdrdalsstiftelsen," the Swedish Medical Research Council, the Swedish Council for Research in Spom and the Faculty of Health Sciences Linkoping University. Send correspondence to: loanna Kvist, Sports Medicine and Physical Therapy, Department of Neuroscience and Locomotion, Faculty of Health Sciences, Linkijping University, SF-581 85 Linkoping Sweden. E-mail: loanna.Kvist@hul.liu.se

he long-term prognosis for sports participation after an anterio r cruciate ligament (ACL) injury is uncertain, both after nonsurgical treatment and after reconstruction of the ruptured ligament.11.27 The rehabilitation after an ACL injury is of importance for the functional outcome of the patient. Various exercises included in a rehabilitation program may produce potentially harmful forces for the knee joint and stretch out the secondary restraints o r the reconstructed ligament. Exercises that operate without joint compression produce large anterior shear force ~ ~ and . ACL ~ ~ strain14 . ~ compared ~ to weight-bearing exercises that have been strongly recommended as the best form of exercise after an ACL injUry.6.14.nv25.28.39Nanweight-bearing activities using isokinetic dynamometers, such as KinCom or Cybex, are used to quantitatively evaluate the progress of muscle rehabilitation and training.15 The recommendation to avoid non-weight-bearing exercises during the early rehabilita-

~ ~ ~ Arthromter A computerized goniometer linkage measuring 3 tibial rotations and sagittal translation (CA4000, OSI Inc, Hayward, Calif) was used to measure tibial translation (b in Figure l ) . The system is composed of 3 parts, the femoral and tibial frames and a rotation module. Three goniometers in the rotation module measure the relative rotations between the femur and tibia. The potentiometer for sagittal motion mounted in the tibial frame (Figure 1B) registers the difference in position between a springloaded patellar pad and the fixation point on the t i b ial tuberosity during knee motion. The sagittal plane direction is perpendicular to the tibial frame. In our study, only the sagittal plane translation (mm) and the knee flexion angle (degrees) were analyzed. The r e p r o d u ~ i b i l i t yand ~ . ~validitys ~~~ of the measurement system has been shown to be good. The potentiometer registering knee extension-flexion (a in Figure 1) was aligned with an approximate knee flexion axis in the center of the lateral femur epicondyle and the alignment was checked repeatedly during the examination. The system was zeroed at the beginning of each test with the subject lying on the examination table and the knee relaxed in full extension. Data METHODS were sampled from the 4 potentiometers by a computer at a rate of 100 Hz. The extension-flexion Subjects curves form a loop (Figure 2) and the translation is calculated as the difference between the arms of the Patient group Twelve patients (4 women, 8 men; loop at the same flexion angle. In our study both the mean age 29 years, age range 14-37 years) with arthroscopically verified unilateral ACL deficiency were maximal distance between the arms of the loop (maximal translation) and the distance at 20" of included in this study (Table 1A). The patients were recruited consecutively from patients admitted to the knee flexion were calculated. The calculation of the County Hospital of Norrkiiping, Sweden for knee he- translations from the curves has been described premarthrosis since February 1995. Except for the diag- viously.1.13.20.23.35 nostic arthroscopy with partial meniscectomy when Dynamomty and Ekctromyogram An isokinetic dynamometer (KinCom 500H. Chattecx Corporation, necessary, no previous surgery had been performed Tenn) was used to measure muscle torque and knee in the injured or contralateral knee. Of 19 patients, 12 agreed to participate. The patients were tested be- joint position. The flexion angle for the KinCom and EMG were zeroed with the leg in full extension, ustween 15 and 28 months after injury (19 ? 3 ing the CA4000 zeroing. Subjects were seated commonths) (Table 1A). Activity level was evaluated with the Tegner score.32 fortably in the dynamometer chair and secured by Before the injury, 10 of the patients had participated body straps and with an angle of -110" between the alignment of the spine and the femur. in sports on a competitive level. At the time of testElectromyogram signals were recorded from the ing, 8 had decreased their level of activity. Knee symptoms at the time of investigation were evaluated vastus medialis, vastus lateralis, biceps femoris and semitendinosus muscles by surface electrodes. The with the Lysholm score.32Five patients had a normal skin was first dry shaved and then cleaned with an alscore (> 94) and only 3 had problems in daily life cohol and ether solution (4:l). Two recording silver(< 84). The individual characteristics are shown in chloride electrodes (Medicotest, Blstykke, Denmark), Table 1A. abraded with redux paste (Medicotest, Blstykke, Control group Eleven volunteers (6 women and 5
J Orthop Sports Phys Ther-Volume 31 .Number 1 *January 2001

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

tion period is not founded on sound data.I0 Specifically, Beynnon et a14 have shown that the ACL strain is almost the same in weight and non-weight-bearing exercises in normal knees. The effect of different exercises on the ACL or the secondary restraints in case of ACL deficiency, can be measured as the degree of anterior translation occurring during the exe ~ i s e . ~ In . the ~ nonimpaired ~ ~ * ~ ~ knee ~ the ~ an~ ~ terior shift of the tibia is eventually resisted by the ACL which is the main restraint to this motion5; but in an anterior cruciate deficient knee, secondary restraint provided mostly by the posterior capsule, menisci, and collateral ligaments5 may be less efficient. In previous studies we reported that, in normal knees, anterior translation of the tibia increased in proportion to the amount of isokinetic knee extensor torque.'.* A larger translation can be expected in the ACL injured knee, which might be harmful for the structures providing secondary restraint to joint translation. Quadriceps-hamstringsco-activation has also been suggested to limit anterior translatiOn.3.18253138 Therefore the purpose of our study was to investigate the effect of different levels of isokinetic concentric and eccentric knee extensor torques in the sitting position, on the anterior tibial translation, in subjects with ACL deficiency and noninjured individuals. Electromyogram (EMG) activity was recorded from quadriceps and hamstring muscles in order to detect any co-activation of extensors and flexors.

men, mean age 27 years, age range 19-34 years) with no previous knee problems, served as controls. Most of them were sports active on a physical fitness level (Tegner score < 6)32and only 2 were active in competitive sports (Table 1B).

Equipment

TABLE 1. Individual characteristics. A. Anterior cruciate ligament deficient group Activity lime level from before injury injury to test (Tegner Surgeryt (months) score) LM LM
15 18 17 21 22 18 16 28 17 21 18 19 9 9 9 7 10 9 9 9 7 6 7 4

Activity level at test time 90N Conc ner Lysholm Lachman quad peak score) score test* (mm) (Nm)
(Teg-

Case
1 2 3 4 5 6 7 8 9 10 11 12

Sex

Age (years)

Concomitant injuries* PCL part, MCL part, MM, LM LM MCL part MCLpart,MM MM, LM LM MM MCL part LM

Ecc quad peak (NW

32 M M 24 M 29 M 1 4 F 2 5 F 25 F 36 M 31 M 29 M 36 M 26 F 37

MM LM LM

5 7 7 4 3 3 7 9 7 6 5 4

71 96 80 99 85 89 100 100 99 80 91 91

511 1 617 519 10111 918 9/23 7/11 6/12 5/9 W8 10113 10116

217l79 1731162 2201177 73/85 1501142 1411153 1901162 291/290 2611189 1871150 2201167 1671180

'1
247/205 156/188 11 1/87 207/2 14 189/204 264/248 387/298 275/248 0 263/219 225/229

0. Control group
Case

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Sex

Age (years)

Activity level at test time (Tegner score)

90N Lachman test* (mm)ll

Conc quad peak (Nm)lI

Ecc quad peak (Nm)lI

PCL indicates posterior cruciate ligament; MCL, medial collateral ligament; MM, medial meniscus; LM, lateral meniscus; part, partial rupture of. t Pdrtial meniscectomy. $ Lachman test measured with CA-4000. 5 Noninjuredlinjured knee. 11 Rightlleft knee. '1 Could not do the eccentric tests because the concentric peak torque at MVC in the injured leg was less than 50'7'' of that in the noninjured leg. t No eccentric tests because of pain.

Denmark), were placed 2 cm apart on the skin above each muscle. The muscles were located by palpation during a submaximal isometric contraction. Electre myogram was recorded by a bipolar isolated amplifier (EMGArnp, Braintronics BV ISO-2104, Almere, Netherlands). Signals from the EMG amplifier and the dynamometer were acquired simultaneously by a data acquisition processor (DAP 2400/6 Microstar Laboratories, Inc, Bellevue, Wash), running parallel with a host personal computer (PC). This system includes its own operating system, DAPL, and data acquisition commands which performed a wide range of standard data acquisition tasks that process data in real-time. The sampled signals were buffered before being sent to the PC, which freed the PC to perform other functions momentarily (eg, displaying and saving data) (Figure 3). The bandwidth of the EMG sig-

nals from surface electrodes was less than 1 kHz. The EMG signals were, therefore, sampled with a rate of 2 kHz." The torque and flexion angle signals were sampled with a rate of 40 Hz and interpolated to the same rate as EMG signals. All signals were amplified and analogue-todigital converted with 12-bit accuracy in the signal range 5V. An analogue low-pass filter was used to eliminate aliasing of the sampled signals, EMG at 800 Hz and torque at 10 Hz. To avoid the influence of movement artifacts and low-frequency noise of the EMG signal a high-pass filter of 16 Hz was used. For details concerning the data acquisition system for registration and analysis see Karlsson et al." The signal energy of the raw EMG signal at the different angles was computed using a low-pass filter (transfer function H(z) = O.OO5/ [ l - 0.995 z-'1 ) of the enerJ Orthop Sports Phys Ther.Volume 31 .Number 1 .January 2001

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

FIGURE 1. CA-4000 system mounted on a subject (a) potentiometer registering knee extension-flexion, (b) potentiometer registering sagittal plane motion.

gy of the signal. This procedure was used with the purpose of removing the fast fluctuations of the EMG signal.'2

Procedures
A laminute warm-up on an ergometer bicycle was completed before the CA-4000 system was secured to the lower extremity with straps. An instrumented Lachman test was performed with the subject strapped to a special seat with the knee flexed to 20'. Tibial translation was recorded by the CA-4000 system when the u p per tibia was pushed/pulled by 90N measured by a force handle. The total translation at 90N was used as a reference for translations during activity3' The muscle tests were then performed using the isokinetic dynamometer at the angular velocity 60O.s-I. In the ACL deficient group, the healthy leg was always tested first. In the control group, the right or left leg was tested in random order. A maximal concentric test was first done with 3 repetitions for both quadriceps and hamstrings. Six tests with 3 r e p etitions each were then performed. The subjects were instructed to work actively in the extension phase (quadriceps concentric contraction), relax in the flexion phase (without hamstrings concentric
J Orthop Sports Phys Ther.Volume 31 .Number 1 .January 2001

contraction), and follow a predetermined line on the computer screen indicating 10, 30, 50, 60, 80, and loo%, respectively, of their individual peak torque calculated from their maximal voluntary contraction (MVC) test. The same procedure was then done for the eccentric testing. The subjects were instructed to relax in the extension phase (without hamstrings eccentric contraction) and try to resist when the lever arm of the dynamometer flexed the knee (quadriceps eccentric contraction). The reliability of achieving different individual torque levels by this method is good (mean variation 4.7% 2 0.96%, 95% confidence interval 3.6-6.1% for the concentric tests and 7.9% + 19%. 95% confidence interval 4.9-10.8% for the eccentric tests) Because of pain and discomfort risk, 1 subject with ACL deficiency was not allowed to perform the eccentric test because the concentric peak torque at MVC in the injured knee was less than 50% of that in the noninjured leg. One subject with ACL deficiency and 1 control subject also did not complete the eccentric testing due to pain.

Data Analysis
All subjects performed 3 repetitions in the 6 different tests (18 repetitions) for each leg and each isoki7

22

4~ -10

10

30 50 70 Kneejoint angle (flexion - extension)

90

110

FIGURE 2. Example of the CA-4000 data for 1 individual with anterior cruciate ligament deficiency (injured leg). The position of the tibia in relation to the femurlpatella is shown for each fifth knee joint angle during the eccentric isokinetic test (the different curves are for the different percent-torque tests). For each curve, the anterior values represent the active phase (ie, flexion in the eccentric tests) and the posterior values, the passive phase. Symbols for the different torque tests: 20% (+), 40% (+), 50% (01, 60% (A), 70%(0), 80% (-), 90% (o), 100% (1).

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

netic mode (concentric and eccentric). These 18 repetitions were then individually analyzed in 10 subgroups (10, 20, 30%. etc torque) depending on the quadriceps torque reached. The borders were 5 1 4 %

for the 10% category (1524% for the 20%, etc). For example, if a subject who was instructed to do 3 r e p etitions on 30% of quadriceps peak torque at MVC achieved 22, 31, and 36%, then the first repetition

2000.
1000.
0.1 -

Biceps

-1000

mv -2000
2000

Semitend

Flexion

deg
0
1

10

11

FIGURE 3. Representative trial of isokinetic knee extension-flexion with active extension phase (concentric quadriceps contraction) and relaxed flexion phase (without active hamstrings contraction) in the injured leg at 80% of MVC torque test. The 4 top graphs show electromyogram activity for muscles vastus medialis, vastus lateralis, biceps femoris, and semitendinosus (mV). The y-axis is adjusted to the MVC test. The 2 bottom graphs show the knee extension force (N) and knee joint angle (degrees).
8

J Orthop Sports Phys Ther.Volume 31 .Number 1 .January 2001

was categorized as 20% torque, the second as 30% torque and the third as 40% torque. If more than 1 repetition was categorized in the same category, the mean value was used. In the ACL deficient group, the values for the maximal translations were normalized to the translation at the maximal voluntary contraction test (100% test) for the noninjured leg, and in the control group it was normalized to each legs translation at MVC. In both groups, the translations at 20" of flexion during active motion for each leg was normalized to the static translation for the same leg in 20" of knee flexion with 90N load (Lachman t e ~ t ) . ~ In ' the ACL deficient group the translation in 20" of flexion in the injured knee was also normalized to the translation of the uninjured knee at the 90N Lachman test. From EMG, root mean square was calculated and used in the analysis. The root mean square was normalized to the root mean square at the MVC test for the same leg. The mean value of the root mean square between 20-40" and 40-60" of flexion at each test was normalized to the root mean square at the same flexion angle intervals as the maximal torque test, as described p r e v i o ~ s l y . ~ ~ ~ ~ ~ ~ Due to technical problems, some EMG files were not used in the final analysis. The technical p r o b lems were mainly related to pressure on the surface electrodes of the hamstrings due to the fact that the subject was sitting. In order to avoid this, the hamstrings electrodes were placed in a bolster of foam rubber. Other problems that occurred more seldom were motion artifacts and problems with wiring. Each raw EMG was inspected visually in order to detect recordings with technical problems. In the concentric mode, from a total of 460 files for each muscle (sum of 1840 files), 26 files for the biceps femoris muscle (11 in the noninjured leg, 13 in the injured leg, and 2 in the control group) and 45 for the semitendinosus muscle (25 in the noninjured leg, 0 in the injured leg, and 20 in the control group) were excluded. In the eccentric mode, from a total of 400 files for each muscle (sum of 1600 files), 4 files for the vastus lateralis muscle (in the control group), 36 files for the biceps femoris muscle (20 in the noninjured leg, 9 in the injured leg, and 7 in the control group), and 72 for the semitendinosus muscle (16 in the noninjured leg, 6 in the injured leg, and 50 in the control group) were excluded. Statistics All statistics were calculated using Statistics, version 4.3 (Statsoft, Inc, Tusla, Okla) and mean values 2 1 standard deviation are given in the text, tables, and figures. With 80% power, the design of this study allowed detection of a translation difference of 1.8 mm and EMG differences of 1% of MVC as significant on a 5% level. An analysis of variance (ANOVA design: 2-way, between 10 torque categories within 2 legs) for repeated measures and post hoc Scheffe tests was used to
J Orthop Sports Phys Ther-Volume 31 .Number 1 .January 2001

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

test the effect of muscle torque categories across limbs on the dependent variables of tibial translation, flexion angle (where the maximal tibial translation occurred), and EMG activity (% of MVC). Separate ANOVA's were used for control and ACL deficient groups. The "limb" factor involved a comparison between the right and left leg in the control group and between the injured and noninjured leg in the ACL deficient group. Students' t tests for dependent variables were used to compare the dynamic and static translation, the concentric and eccentric peak muscle torque at MVC between the right and left leg in the control group, and the injured and noninjured leg in the ACL deficient group. The students' t tests for independent variables were also used as comparisons of the dynamic and static translation and the concentric and eccentric peak muscle torque at MVC between the injured leg in the ACL deficient group and the control group. A multiple analysis of variance (MANOVA) was used to directly compare the injured leg in the ACL deficient group with the control group. This analysis evaluated the muscle torque categories on the dependent variables of tibial translation, flexion angle where the maximal tibial translation occurred, and EMG activity between the injured leg in the ACL deficient group and the control group. The Pearson product moment correlation was used to measure relationships between quadriceps torque and the logarithm for tibial translations (the group correlation was calculated as mean of the individual r values). Raw translation and torque data were used for comparisons of the translation and peak muscle torque at the maximal voluntary contraction test. Otherwise, individually normalized data was used (see data analysis section). A significance level of P < .05 was used.

RESULTS
In the control group, there were no significant differences between the right and left leg in translation at the different torque levels. Therefore, the mean value of translation in both legs was used. In the ACL deficient group, the concentric quadriceps peak torque at maximal voluntary contraction (MVC) was 16% lower (a mean difference of 30 Nm 2 44, t = 2.33, df = 11, P = .04) (Table 2) in the injured leg compared to the noninjured, but there was no difference in the eccentric quadriceps peak torque at MVC. Both peak torques normalized to the individual body mass index, were significantly smaller in the ACL injured legs compared to the control group ( t = 2.52, d f = 21, P = .02, and t = 2.79, df = 18, P = .01, respectively) (Table 2). The static tibial translation in the 90N Lachman test was 35% larger (a mean difference of 4.1 mm 2 4.1) in the injured leg compared to the noninjured ( t = 3.42, df = 11, P = .006) (Table 2). The maxi-

TABLE 2. Peak muscle torque during maximal voluntary contraction test and active and static translation (mean and standard deviation). Anterior cruciate ligament deficient injured leg
161 2 53 224 2 76 6.6 2 1.9 9.0 2 2.7 10.722.8 11.722.5 11.424.5 4.1 2 4.1

n Conc quad (Nm)* Ecc quad (Nm)* Conc quad (Nm)/BMI (kglm2)* Ecc quad (Nm)/BMI (kglm2)* Max-aAll conc (mm)* Max-aATT ecc (mm)' STT (mm)* STTdiff (mm)*
12 10 12 10 12 10 12

n
12 10 12 10 12 10 12

P Anterior cruciate Anterior ligament deficient cruciate noninjured leg ligamentt

191 57 232 2 75 7.8 2 2.2 9.8 t 2.7 10.622.3 8.8 2.7 7.4 ? 2.6

n
11 10 11 10 11 10 11

P Anterior cruciate Control group ligament-CS 185 2 44 270 2 69 8.6 2 1.9 12.6 2 3.2 9.9 2 2.1 11.722.0 8.4 2 2.5 0.6 2 1.9

.04 .18 .04 .14 .81 .02

.006

.02 .01 .47 .98 .06 .02

Conc quad indicates concentric quadriceps peak torque at MVC; Ecc quad, eccentric quadriceps peak torque at MVC; BMI, body mass index (body weightllengthz); max-aAll conc, maximal active anterior tibial translation in the concentric test MVC test; max-aATT ecc, maximal active anterior tibial translation in the eccentric MVC test; STT, static tibial translation in the 90N Lachman test; Sll-diff, difference in static tibial translation between injured and noninjured leg in the anterior cruciate ligament group and left-right leg in the control group. t P value in the comparison between injured and noninjured leg in the anterior cruciate ligament deficient group. P value in the comparison between injured leg in the anterior cruciate ligament deficient group and the control group. 5 P value in the comparison of STT-difference between the anterior cruciate ligament deficient group and the control group.

*
Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

erally it remained within 105% of the individual translation at MVC (Figure 5A). During eccentric muscle activity, the mean normalized translation at all the percent torque levels was 38% larger (a mean difference of 2.8 mm 2 2.6) in the ACL injured knees, compared to the noninjured (F,,, = 50, n = 72, P < .000) (Figure 5B). Comparison of active tibial translation between conm t r i c and ecmtric tests In the ACL deficient group, the mean normalized active tibial translation was 43% larger (a mean difference of 2.3 mm 2 2.6) Active Translation in the eccentric tests compared to the concentric (F,,, = 42, n = 67, P < .000) in the injured knee, During one isokinetic cycle (extension-flexion), the position of the tibia in relation to the patella was ana- but there was no difference in the noninjured knee (Figure 5A and B). Neither was there any difference lyzed at every fifth knee joint angle. Figure 2 demonin concentric and eccentric anterior tibial translation strates an example of these data on 1 individual with for the control group. ACL deficiency (injured leg) in the eccentric tests Comparison between active and static translation at with different torque values. In all tests, the tibia followed an anterior pathway during the phase with m u s 20" offexion The Lachman test in 20' of knee flexion represents the available 90N translation envecle activity (concentric = extension; eccentric = flexlope. In concentric muscle activity, control knees ion) and a posterior pathway during the relaxation used the entire normal space (118%) and the noninphase. jured knees a similar space (131%). Utilization was The mean maximal translation occurred between similar in eccentric muscle activity. The injured knees 20 and 40" of flexion, without any significant differused only 80% of the available space in the concenences between tests o r groups (Figure 4A and B). tric mode. In the eccentric mode, already 10% quadThere was a similar increase in mean normalized translation with increasing torque in all knees during riceps torque consumed a translation space that was equal to the 90N space of the normal knee and concentric and eccentric muscle contractions. In the 155% of the normal knee's space at MVC. This control group, the mean normalized translation inmeans that the injured knees utilized the entire creased from 54 to 100% of the maximal individual pathologic space available (101%) in eccentric mode. translation (Figure 5A and B). The correlation coefficients between muscle torque and translation for the concentric tests were r = 0.84, r = 0.83 and r = Eledromyogram 0.90 for the injured, noninjured, and control group, In the control group, during the eccentric mode, respectively and for the eccentric tests r = 0.67, r = the EMG activity for vastus medialis in both flexion 0.78 and r = 0.89, respectively ( P < .05). angle intervals (2040" and 40-60") and vastus laterDuring concentric muscle activity the mean normalized translation was similar in all groups and gen- alis at 20-40" of flexion was significantly larger in the ma1 tibial translation in the eccentric MVC test was 25% larger (a mean difference of 2.9 mm 2 3.2) in the injured leg than in the noninjured ( t = 2.88, df = 9, P = .02) (Table 2), but similar in the concentric test and similar to the control group. The difference in static translation between the 2 legs was significantly greater in the ACL group than in the control group (4.1 mm 2 4.1 versus 0.6 mm ? 1.9, t = 2.53, df = 22, P = .02) (Table 2).
J Orthop Sports Phys Ther.Volume 31 .Number 1 .Janualy 2001

Concentric muscle toque (% of maximum)

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

10%

20%

30%

40%

S O %

80%

70%

80%

90%

100%

Eccentric muscle toque (% of maximum) FIGURE 4. Flexion angles (t SD) where the maximum active tibia1 translation occurred at the different concentric (A) and eccentric (B)quadriceps torques in the noninjured (0) and injured (I) leg (n = 12 concentric and n = 10 eccentric) in the ACL deficient group, and the mean of the 2 legs in the control group (t) (n = 11 concentric and n = 10 eccentric). Muscle torque is expressed as a percentage of individual maximum.

right leg compared to the left with 18% 2 45 of MVC (F,,,, = 6, n = 53, P = .018), 11% 2 28 of MVC (F,.,, = 5, n = 53, P = .026) and 16% 2 25 of MVC (F,,,, = 20, n = 49, P = .000) greater activation, respectively. In the other muscles and flexion angles, no significant differences existed between the 2 legs. The mean value for the 2 legs was used for further analysis. The normalized EMG signals of vastus medialis and vastus lateralis remained at low levels (< 12% of MVC) up to 20% torque, but then increased with increasing muscle torque, similar in both concentric and eccentric work mode with no difference between the injured and noninjured leg (Figure 6A and B). There were also similarities between the different flexion angles (20-40" and 40-60), with the excep tion of the noninjured leg in the ACL group in the eccentric mode where 20-40" had greater activation compared to 40-60" with a mean (2SD) of 16% 2 27 of MVC (F,,,, = 22, n = 82, P = .000) and 20% 5 40 of MVC (F1,,, = 15, n = 82, P = .000) for vastus medialis and vastus lateralis respectively. The
J Orthop Sports Phys Ther*Volume 31 .Number 1 *January 2001

EMG levels of the hamstring muscles were low (< 1.5% of MVC), but increased in a similar manner towards higher muscle torques. The variation in activation appeared to be larger in injured than in noninjured knees (Figure 7A and B).

DISCUSSION
Similar to the results of our previous studies,'.20 translation increased with increasing quadriceps torque in nonimpaired knees. The same was true for patients with ACL deficiency, but they had lower concentric quadriceps MVC and no change in eccentric MVC in line with findings from other authors.9J6J9~21.29 In general all knees used the entire 90N translation envelope at maximal muscle torque, but the ACL deficient knees exceeded the normal space only during eccentric muscle activity. The patients with ACL deficiency were thus able to limit translation during concentric muscle activity, but not during eccentric activity. The EMG analysis showed more or less linear increases in activation of quadri-

175

.0
C

E -

I
m

1% 125

0 .V)

loo
75 S O 25

i
10% 20% 30% 40% S O % 80% 70%

80%

90%

100%

Concentric musde toque (% of maximum)

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Eccentric musde toque (% of maximum)

FIGURE 5. Active anterior tibia1 translation in relation to concentric (A) and eccentric (B)muscle torque in the noninjured (0) and injured ( I leg ) (n = 12 concentric and n = 10 eccentric), in the anterior cruciate ligament deficient group, and the mean of the 2 legs in the control group (+) (n = 11 concentric and n = 10 eccentric) (mean with standard deviation). Translation and muscle torque are expressed as a percentage of individual maximum.

a F presented significantly more quadriceps-hamceps and very low activation of hamstrings in all strings co-activation in the noninjured leg (mean groups, and with no difference between injured and 31% of hamstrings MVC) of 5 patients (4 with ACL noninjured legs. reconstruction and 1 without surgical intervention), In different studies the quadriceps-hamstringscobut no differences in quadriceps-hamstringsco-activaactivation levels have varied widely, which tends to make the results uncertain. Draganich et a1,8 defined tion compared to the control group (mean 14-18% of hamstrings MVC). Our results could possibly vary co-activation as any normalized quadriceps or hamdue to a different set up in the isokinetic machine, strings signals that concurrently exceeded a threshold of 3% of MVC, but Solomonow et a13' and Barat- primarily designed to analyze the effect of quadriceps torque. Therefore, our patients were instructed ta et a13 estimated ceactivation between 1.5-7% and to let the return motion be passive. This could have 3-7% of MVC, respectively, as significant activation. In our study, the mean co-activation of hamstrings influenced the degree of antagonist activation also during isokinetic quadriceps work was 1.5% of MVC during the active motion. for the biceps muscle between 20-40" in the eccenIt is uncertain whether the low hamstrings activity levels that we recorded represent activation or crosstric mode and did not exceed 1% of MVC in the other tests. We were unable to register similar levels talk from the active quadriceps muscles. In fact, De of hamstrings co-activation during quadriceps work, Luca and Merletti7 reported that cross talk per se can as described by other authors, on normal individuals be associated with considerable EMG activity. In othIn er words, according to the present study, hamstrings in spite of using a similar technique.3~8~18~24.m~s1~38 the only report on ACL injured patients, Ostering et co-activation during quadriceps work could hardly be
J Orthop Sports Phys ThereVolume 31 *Number 1 *January 2001

Vastus medialis. 2040' flexion Concentric muscle contraction

Concentric muscle toque (% of maximum)

Vastus medialis. 2 0 4 0 ' flexion Eccentric muscle contraction

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Eccentric muscle toque (% of maximum)

FIGURE 6. Electromyogram activity (root mean square normalized to individual maximum) of vastus medialis during concentric (A) and eccentric (B) muscle contraction, between 20 and 40" of flexion, in the noninjured (0) and injured ( I leg ) in the anterior cruciate ligament group (n = 12 concentric and n = 10 eccentric), and the mean of the 2 legs in the control group (*) (n = 11 concentric and n = 10 eccentric) (mean with standard deviation). Muscle torque is expressed as a percentage of individual maximum.

an important way to limit anterior tibial translation. Moreover, no prominent changes in electrical activation occurred due to the ACL injury. A practical finding in this study, supported by previous in vitro is that eccentric quadriceps activity, already with low torques, reaches the limits of the normal translation envelope. We have previously demonstrated this for nonimpaired knees.* Arms et a12 demonstrated in vitro an increased strain in the ACL during eccentric compared to isometric quadriceps load. In the ACL deficient knee, the posterior capsule, the menisci, and the collateral ligaments become the primary restraints of anterior tibial translation. Therefore, the tibia will assume an a b normal anterior position during load, especially eccentric quadriceps work. In line with our in vivo findings, Renstr6m et aP6 showed in vitro that a 250N hamstrings load significantly diminished ACL strain, but had no effect if an equally large quadriceps load was applied. Similarly, Torzilli et alS3 showed in an in vitro study that not even a lOON posterior force could normalize the anterior shift of the tibia in relation to the femur that occurred after
J Onhop Sports Phys TheroVolume 31 *Number I *January 2001

sectioning of the ACL. It seems that simultaneous activation of hamstrings during knee extension in a non-weight-bearing exercise is not a viable way to avoid excessive loads. Another finding of practical importance is that concentric isokinetic muscle testing does not seem to involve a similar risk as eccentric testing. In this mode, the patients with ACL deficiency could limit the amount of translation so that it remained within the normal envelope. In the future, it would be interesting to examine tibial translation and EMG activity of the hamstrings muscles during weight-bearing exercises with different loads to provide more data about safe rehabilitation exercises after an ACL injury.

-4

CONCLUSION
The individual anterior tibial translation increased with increased quadriceps torque in a similar manner in both concentric and eccentric quadriceps contraction in all legs tested, but the normalized translation was a mean 38% larger in the ACL injured knees during the eccentric contractions. Subjects with ACL

Semnendinosus. 20-40' flexion Concentric muscle contraction

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Concentric muscle contraction (% of maximum)

Semitendinosus.20-40' flexion Eccentric muscle toque

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Eccentric muscle toque (% of maximum)

FIGURE 7. Electromyogram activity (root mean square normalized to individual maximum) of Semitendinosus during concentric (A) and eccentric (6) muscle contraction, between 20 and 40" of flexion, in the noninjured (0) and injured (I leg )in the anterior cruciate ligament group (n = 12 concentric and n = 10 eccentric), and the mean of the 2 legs in the control group (*) (n = 11 concentric and n = 10 eccentric) (mean with standard deviation). Muscle torque is expressed as a percentage of individual maximum.
deficiency could l i m i t the tibial translation within a normal translation space during concentric muscle activity, but n o t during eccentric activity. Quadricepshamstring co-activation was n o t present during isokinetic quadriceps muscle testing, and, therefore, i t was n o t a factor in limiting anterior tibial translation. D'Ambrosia R. Muscular coactivation. The role of the antagonistic musculature in maintaining knee stability. Am ]Sports Med. 1988;16:113-122. Beynnon BD, Johnson RJ, Fleming BC, Stankewich CJ, Renstrom PA, Nichols CE. The strain behavior of the anterior cruciate ligament during squatting and active flexion-extension. A comparison of an open and closed kinetic chain exercise. Am J Sports Med. 1997;25:823-829. Butler DL, Noyes FR, Grood E. Ligamentous restraint to anterior-posterior drawer in the human knee. ] Bone Joint Surg Am. 1980;62A:259-270. Bynum EB, Barrack RL, Alexander AH. Open versus closed chain kinetic exercises after anterior cruciate ligament reconstruction. A prospective randomized study. Am J Sports Med. 1995;23:401406. De Luca C, Merletti R. Surface myoelectric signal crosstalk among muscles of the leg. Electroencephalogr Clin Neurophysiol. 1988;69:56&575. Draganich L, Jaeger R, Kralj A. Coactivation of the hamstrings and quadriceps during extension of the knee. ] Bone Joint Surg. 1989;71 A:1075-1081. Elmqvist LG, Lorentzon R, Johansson C, Fugl-Meyer A. Does a torn anterior cruciate ligament lead to change in the central nervous drive of the knee extensors?Eur J Appl Physiol. 1988;58:203-207. Fitzgerald G. Open versus closed kinetic chain exercise: issues in rehabilitation after anterior cruciate ligament reconstructive surgery. Phys Ther. 1997;77:1747-1754. Folksam AB. Idrottsskador. Stockholm, Folksam forlagsservice; 1994.

4.

5.

ACKNOWLEDGMENTS
The authors especially thank D r Peter Rockborn, Norrkcping County Hospital, Sweden for supporting us with A C L deficient patients and Peter Syvertsson, Rehabilitation Medicine, Department o f Neuroscience and Locomotion, Linkcping University, Sweden for assistance when testing the subjects. 6.

7. 8.

REFERENCES
1. Antonopoulos J, Gillquist J. Anterior tibia translation related to isokinetic concentric quadriceps torques. Isokinetics and Exercise Science. 1996;6:145-15 1. , 2. Arms SW, Pope MH, Johnson RJ, Fischer RA, Arvidsson I Eriksson E. The biomechanics of anterior cruciate ligament rehabilitation and reconstruction. Am J Sports Med. 1984;12:8-18. 3. Baratta R , Solomonow M, Zhou B , Letson D, Chuinard R ,
14

9.

10. 11.

J Orthop Sports Phys Ther.Volume 31 .Number 1 .January 2 0 1

12. Gerdle B , Karlsson S, Day S, Djupsjobacka M. Acquisition

and analysis of the surface electromyogram. In: WindModern Techniques in Neuhorst U, Johansson H, e d ~ . roscience Research. Springer, Berlin; 1999:705-755. Goertzen D, Lysholm M, Messner K, Gillquist J. Sagittal translation of the tibia during stair walking in normal volunteers. Reproducibility of an electrogoniometric method. lsokinetics and Exercise Science. 1 995;5:19-23. Henning C, Lynch M, Click K. An in vivo strain gage study of elongation of the anterior cruciate ligament. Am ) Sports Med. l985;13:22-26. Hubley-Kozey C, Stanish WD, Curwin SL. Rehabilitation principles and techniques. In: DeLee JC DD, ed. Orthopaedic Sports Medicine. Philadelphia, Pa: WB Saunders; 1 994:184-202. Hurley MV, Jones DW, Wilson D, Newham DJ. Rehabilitation of quadriceps inhibited due to isolated rupture of the anterior cruciate ligament. ) Orthop Rheum. 1992;5:

25. Palmitier RA, An KN, Scott SG, Chao EYS. Kinetic chain exercise in knee rehabilitation. Sports Med. 1991;ll: 402-41 3. , Arms SW, Stanwyck TS, Johnson RJ, Pope 26. Renstrom P
MH. Strain within the anterior cruciate ligament during hamstring and quadriceps activity. Am ) Sports Med.

13.

1986;14:83-87. , Dhalberg L, Lohmander S. Osteo27. Roos H, Adalberth T

arthritis of the knee after injury to the anterior cruciate ligament of meniscus: the influence of time and age. Osteoarthritis Cartilage. 1 995;3 :261-267. . Accelerated rehabilitation after an28. Shelbourne K, Nitz P terior cruciate ligament reconstruction. A m ) Sports Med.

14. 15.

1990;18:292-299. , Kato K, Udagawa E. Characteristics of the 29. Shirakura K

isokinetic performance of patients with injured cruciate ligaments. Am ) Sports Med 1992;20:754-760. 30. Snow C, Cooper J, Quanbury A, Anderson J.Antagonist cocontraction of knee flexors during constant velocity muscle shortening and lengthening. ) Electromyogr Kinesiol. 1 993;3:78-86. 31. Solomonow M, Baratta R, Zhou BH, et al. The synergistic action of the anterior cruciate ligament and thigh muscles in maintaining joint stability. Am ) Sports Med. l987;14:

16.

145-1 54. 17. Karlsson S, Erlandson B, Gerdle B. A personal computerbased system for real-time analysis of surface EMG signals during static and dynamic conrtactions. ) Electromyogr Kinesiol. 1994;4:170-180. 18. Kellis E, BaltzopoulosV. The effects of antagonist moment on the resultant knee joint moment during isokinetic testing of the knee extensors. Eur ) Appl Physiol. 1997;76:

Journal of Orthopaedic & Sports Physical Therapy Downloaded from www.jospt.org at on April 4, 2014. For personal use only. No other uses without permission. Copyright 2001 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

253-259. , Webster-Bogaert S. Knee 19. Kramer J, Nusca D, Fowler P

flexor and extensor strength during concentric and eccentric muscle actions after anterior cruciate ligament reconstruction using the semitendinosus tendon and ligament augmentation device. Am ) Sports Med. 1993;2 1 :

207-21 3. , Lysholm J. Rating systems in the evaluation of 32. Tegner Y knee ligament injuries. Clin Orthop. 1985;198:43-49. , Deng X, Warren R . The effect of joint-compres33. Torzilli P
sive load and quadriceps muscle force on knee motion in the intact and anterior cruciate ligament-sectioned knee. Am j Sports Med. 1994;22:105-112. 34. Vergis A, Gillquist J. Fluoroscopic validation of electroniometrically measured femoral roll-glide motion in the knees of healthy and ACL deficient subjects. Tran Orthop Res Soc. 2000;486. 35. Vergis A, Gillquist 1. Sagittal plane translation of the knee during stair walking. Comparison of healthy and anterior cruciate ligament-deficient subjects. Am ) Sports Med.

285-291. 20. Kvist J, Gillquist J. Anterior tibia1 translation during eccentric, isokinetic quadriceps work in healthy subjects. Scand j Med Sci Sports. 1999;9:189-194. 21. Lorentzon R, Elmqvist LG, Sjostrom M, Fagerlund M, Fuglmeyer AR. Thigh musculature in relation to chronic anterior cruciate ligament tear: muscle size, morphology, and mechanical output before reconstruction. A m ) Sports Med. 1989;17:423-429. 22. Lutz GE, Palmitier RA, An KN, Chao EYS. Comparison of tibiofemural joint forces during open-kinetic-chain and closed-kinetic-chain exercises. ) Bone joint Surg Am.

l998;26:841-846. 36. Vergis A, Gillquist J.Sagittal plane translation of the knee

during stair walking in healthy volunteers measured by an electrogoniometer chain. Scand Med Sci Sports. 1995;

5:353-357. 37. Vergis A, Hindriks M, Gillquist J. Sagittal plane translations of the knee on anterior cruciate deficient subjects and controls. Med Sci Sports Exerc. 1997;29:1561-1566. F , Fleisig GS, Barrantine SW, An38. Wilk KE, Escamilla R drews JR, Boyd ML. A comparison of tibiofemoral joint forces and electromyographic activity during open and closed kinetic chain exercises. Am ) Sports Med. 1996;

1993;75:732-739. 23. Lysholm M, Messner K. Sagittal plane translation of the

tibia in anterior cruciate ligament deficient knees during commonly used rehabilitation exercises. Scand) Med Sci Sports. 1995;5:49-56. 24. Ostering LR, Caster BL, James CR. Contralateral hamstrings (biceps femoris) coactivation patterns and anterior cruciate ligament dysfunction. Med Sci Sports Exerc.

24:51&527. . Comparison of closed and 39. Yack H, Collins C, Whieldon T

open kinetic chain exercise in the anterior cruciate ligament-deficient knee. Am ) Sports Med. 1993;2 1 :49-54.

1995;27:805-808.

J Orthop Sports Phys Ther.Volnme 31 .Number 1 .January 2001