European Journal of Sport Science, vol.

1, issue 2
©2001 by Human Kinetics Publishers and the European College of Sport Science

Activation and Torque Deficits

in ACL-Reconstructed Patients
4 Months Post-Operative
Caroline Nicol, Nicolas Gouby, Jean-Marie Coudreuse, Christian Flin,
Jean-Michel Viton, Alain Delarque, Christian Brunet, and Paavo V. Komi

This study compares knee extension and flexion torques and electromyographic (EMG)
activity of normal and anterior cruciate ligament (ACL)-reconstructed knees during
maximal unilateral isometric and isokinetic tests performed 4–5 months after ACL
reconstruction. The subjects consisted of 3 age- and activity-matched groups of 6
subjects: a healthy control group (Ctrl) and 2 groups of patients, with Kenneth-Jones
technique using autologous patellar-ligament graft (KJ group) and autologous graft from
the semitendinosus tendon (ST group). When compared to the Ctrl group values, each
patient group had significant bilateral extension torque deficits in isometric and at slow
velocity concentric conditions. In the Op leg, this deficit was associated with lower
quadriceps EMG activation. In all tests, bilateral hamstring co-activation level was lower
in the patient groups. Flexion torque deficits were observed in the Op leg of both patient
groups and in the Nop leg of the KJ group, with minor differences in either hamstring
EMG activation or quadriceps co-activation. The inter-leg difference in extension and
flexion torques were significant for both patient groups, but only in the isometric strength
test, and with no difference in activation. These data demonstrate bilateral knee extension
and flexion torque deficits a few months after ACL reconstruction. The observed deficit
in quadriceps activation emphasizes the interest to perform combined EMG and strength
testings quite early during the rehabilitation period.

Key Words: ACL, EMG, activation deficit, torque deficit, co-activation.

Key Points:
• Protective deficit in quadriceps activation partly explains the long-lasting recovery in
extension torque production in ACL-reconstructed patients.
• An isometric strength test is useful in revealing deficits in maximal activation and
torque production in such patients.
• Based on observed bilateral deficit, responses of the operated leg should be compared
not only with the non-operated leg, but also with the legs of healthy control subjects.

Introduction
The anterior cruciate ligament (ACL) plays major mechanical and kinesthetic roles at the knee
joint. Anatomic studies have clearly defined the composition, orientation, and attachment sites of
the ACL (1). Mechanical studies with cadavers have demonstrated that the ACL ligament plays a
primary role in restraining anterior tibial displacement with respect to the femur (2). In addition
to this function, the ACL also restrains internal knee rotation (3). The ACL tension varies
depending on the type of movement (knee extension, flexion, and/or rotation), angular position
(3), and intervention of active muscle forces (4, 5). Equilibrium under variable loading and
geometric conditions of the joint would require synergism of all active and passive components.
The literature reports of the existence of a complex but systematic sensory-motor synergy around
the anatomical structures of the knee. In vivo mechanical studies with strain transducers have
suggested that cruciate receptors may act as sensors of tibial position and movement (6).
Histologic investigations have identified in the ACL two types of mechanoreceptors, similar to
Ruffini corpuscules and Golgi tendon organs, that could respond to changes in ligament tension
and provide the central nervous system with information on the position, velocity, and
acceleration of the tibial motion (7). ACL rupture may thus be suggested to lead to a vicious
circle in which the combined increased laxity and sensorial deficits would induce a functional
instability, thus increasing the risk of additional lesions.

ACL injury is particularly high in contact sports that involve high acceleration and/or side-
cutting movements, and more frequent in sportive females than in males (8). In spite of its
surgical reconstruction, the ACL injury is generally followed by long-term weakness of
especially the knee extensor muscles. Deficits of 30 to 50% in maximal knee extension torque
have been reported 3 and 9 months after ACL reconstruction (9, 10) as well as long-lasting
deficits of 25% and 15%, respectively, 1 year and 2 years post surgery (11). Muscle weakness
has been mostly attributed to a loss of muscle mass caused by immobilization or limited use of
the injured limb. Disuse atrophy does not seem, however, to fully account for the prolonged
decrease in knee extension strength in acute disuse. According to Suter and Herzog (12), only
little attention has been paid to the possible contribution of inhibition to the muscle weakness
and associated long-term functional and structural changes of the knee joint after surgery. Using
the interpolated twitch technique in ACL patients several months after the ligament rupture,
these authors demonstrated knee extensor inhibition in the injured as well as in the contralateral
limb. These bilateral activation deficits were 2 to 3 times higher than the 10% inhibition of the
control group. On the other hand, Hurley et al. (13) reported that severe initial knee extensor
inhibition (30–45%) might interfere with the rehabilitation process and compromise the
functional recovery, whereas a small amount of initial inhibition seems to allow for strength
gains early in the rehabilitation process.

Observation of long-lasting torque decrement, even after ACL-reconstruction, raises the problem
of the protective neuromuscular adjustments that could delay or even limit the functional
recovery. The present paper therefore makes an attempt to evaluate after a few months of
recovery the effects of ACL injury and reconstruction on the neuromuscular function in
commonly used isometric and isokinetic strength tests.

Materials and Methods

Subjects
A total of 18 subjects, ranging from 18 to 28 years, volunteered for the study. From these, the
control group (C) included 6 healthy subjects (5 males and 1 female) with no previous records of
any lower extremity problems. This group was matched, with respect to age and athletic level,
with two patient groups. The first patient group (KJ) included 4 males and 2 females with
reconstructed ACL by the Kenneth-Jones technique (autologous patellar-ligament graft). The
second patient group (ST) included 3 men and 3 women with reconstructed ACL from an
autologous graft consisting of four loops of the semitendinosus tendon. Table 1 summarizes data
on each group, including the mean times from accident to surgery and from surgery to testing.
None of the patients had any pain, knee effusion, or exhibited any type of knee-joint damage
besides the ACL rupture. None of the subjects had undergone isokinetic training or testing before
participating in this study.

Methods
The testing protocol included a series of isometric and concentric knee extensions and flexions
on a specific isokinetic dynamometer (Cybex Norm). The subjects sat on the dynamometer chair
in a secured position. Special care was taken to align the dynamometer rotation axis with the
estimated knee joint anatomical flexion-extension axis. In order to familiarize the patients with
the tests and to make them more confident while testing the operated leg (Op), the non-operated
leg (Nop) was systematically tested first. In addition, the tests were preceded by a warm-up and
training periods with visual feedback to teach the patients to produce a fast rise in torque and to
maintain the maximal effort over the entire range of motion.

In the isometric strength tests, two maximal trials were performed at the knee angles of 65° in
extension and 40° in flexion (0° corresponding to complete knee extension). The subjects were
verbally encouraged to produce maximal torque output as fast as possible and to maintain the
maximal effort for 5 s. The trials were separated by a recovery period of 5 min. Four warm-up
contractions were systematically performed with gradually increasing efforts before the maximal
test. The isokinetic testing series included 3 maximal tests of 4 successive knee extension-flexion
movements (at 60, 180, and 240° · s–1). Each test was performed over a 95° range (5–100°) and
was preceded by a familiarization over 3 repetitions at the predetermined velocity.

Torque and angular displacement were recorded in parallel with surface electromyographic
(EMG) activity detected from 6 selected muscles. The overall recording was performed at a
sampling frequency of 1000 Hz (ME3000P; MEGA systems, Finland), and transmitted via an
optic fiber to a PC. Bipolar technique with disposable surface electrodes (Contrôle Graphique
Médical) was used to record the EMG activity from (a) the major knee extensors: vastus medialis
(VM), vastus lateralis (VL), and rectus femoris (RF); (b) the knee flexors: biceps femoris (BF),
semitendinosus and semimembranosus (ST-SM); and (c) the lateral gastrocnemius (LG) of both
legs. Each EMG signal was band-pass filtered (20–500 Hz) and pre-amplified nearby the site of
the electrode.

Data Analysis
From the recorded maximal isometric performances, the trial presenting the highest peak torque
was kept for further analysis. For the isokinetic strength tests, the two intermediate repetitions,
which showed lesser variability than the other two, were selected for subsequent analysis. For
each of the selected trials, the digitized EMG signals were rectified and afterwards low-pass
filtered at 100 Hz in parallel with the torque signal. The subsequent part of the analysis was then
specific to each mode of testing.

The specific analysis of the isometric strength performances examined the peak torque values
during the first 2,500 ms from the start of the torque rise. The EMG activity of each muscle was
averaged for a period of 1 s (from –500 to +500 ms) around the peak torque. These averaged
EMG values were then used to calculate mean EMG activation of the quadriceps femoris (VM,
VL, RF) and hamstring (BF, ST-SM) muscle groups. The lateral gastrocnemius (LG) EMG
activity varied greatly among subjects and was not analyzed any further. For the isokinetic
performances, the torque and EMG signals were first averaged for the 2 successive repetitions of
each selected velocity (Figure 1A). Mean torque and mean EMG activity of each muscle were
then calculated for the respective mean extension and flexion phases. Mean torque was chosen
instead of peak torque to provide more insight into the total contribution of the involved muscles.
To estimate the relative co-activation level, the mean activity of each antagonist muscle was
expressed in percentage of its agonist activity during the opposite phase of movement. The
hamstring muscle group presented in some subjects a short initial burst of activity at the
beginning of the extension movement that has been previously described by Solomonov et al.
(1987) and Aagaard et al. (2000). This hamstring burst differed clearly from the rest of the
antagonist EMG activity that was steady thereafter at a lower level throughout extension. The
level of co-activation was calculated for the second half of the respective extension and flexion
movements.

Figure 1 — A. Average recording for a patient of 2 consecutive knee extension-flexion movements at 60° · s–1
including knee torque, angular displacement, mean quadriceps femoris (QF), hamstrings (H), and lateral
gastrocnemius (LG) rectified EMG activity. B. Mean value of each signal corresponding to 4 equal parts of
the angular displacement in extension (from 100 to 5°) and in flexion (from 5 to 100°). Reference 100%:
maximal Nop leg value. (Note: See color originals of this figure at <www.humankinetics.com/ejss>.)

The subsequent part of the analysis (Figure 1B) consisted of calculating the mean value of each
signal corresponding to 4 equal parts (~24°) of the angular displacement in extension as well as
in flexion. These angular displacements (100–76, 76–52, 52–28, 28–5°) are indicated on Figure
1B by the corresponding mean knee joint positions of 88, 64, 40, and 16°. The maximal values
obtained for each signal in the Nop leg were then used as 100% reference to allow comparison of
the relative EMG and torque patterns between the Nop and the Op legs. The obtained individual
values were then averaged for each group.

To calculate the relative differences in torque and activation between the two legs, the maximal
torque and EMG values of the Op leg were then expressed as a percentage of the corresponding
values measured in the Nop leg. In the control group, the left leg values were taken as 100%
reference.

Statistical Analysis
Differences in absolute torque value and agonist EMG activation as well as in relative level of
co-activation were examined by a three-way (Group x Leg x Velocity) analysis of variance
(ANOVA) with repeated measures on the last two factors. When this analysis revealed
significant differences, a HSD Tukey post hoc test was used to compare the experimental
conditions. The level of statistical significance was set at p = .05, and the minimum effect
intensity (EI) was set at 2%.

Results
Independently of the tested velocity, group effects, and leg effects, the results differed clearly in
extension and flexion. These conditions therefore are presented separately. In the case of
significant main effects and interactions, the latter are shown with statistical significance. When
the absolute changes in torque and EMG activity were significant, they are often illustrated as
corresponding relative changes for reasons of convenience.

Maximal Knee Extension Tests

Figure 2 presents the averaged values of extension torque, quadriceps EMG activity, and
hamstring co-activation for each of the 4 testing conditions (0, 60, 180, and 240° · s–1) and for
the operated (Op) and non operated (Nop) legs of each patient group. As the Ctrl group presented
only minor and non-significant inter-leg variations (0 ± 7% in torque, 2 ± 11% in quadriceps
EMG and 0 ± 5% in hamstring co-activation), this group is represented in Figure 2 by the mean
values only. However, for statistical inter-group comparisons, the Nop and Op legs of the patient
groups were compared with the respective left and right legs of the Ctrl group.
Figure 2 — Patient group mean (±SD) values corresponding to the operated (Op) and non operated (Nop)
extension leg torque, quadriceps AEMG activity, and relative hamstring co-activation level in each of the 4
tests (0, 60, 180, and 240° · s–1). The control (Ctrl) group is represented by mean left and right leg values.
Note: *p < .05 and ***p < .001 in the Group x Velocity comparison; ###p < .001 in the Group x Leg x Velocity
comparison; £p < .05 and ££p < .01 in the Group x Leg comparison. (Note: See color originals of this figure at
<www.humankinetics.com/ejss>.)

For the extension torque, the analysis revealed significant main effects of Velocity (p < .001),
Leg (p < .001), and Group (p < .01), accompanied by significant two-way interactions between
Group and Velocity (F6,45 = 12.48; p < .001; EI = 3.5%), Group and Leg (F2,15 = 6.90; p < .01; EI
= 1.4%), Velocity and Leg (F3,45 = 109.39; p < .001; EI = 15%) as well as between Group and
Leg and Velocity (F6,45 = 6.96; p < .001; EI = 1.9%).

For the Group x Velocity effect, post hoc analysis revealed that the patient groups differed from
the Ctrl group during isometric testing (KJ, ST: p < .001) and at low (60° · s–1) velocity (KJ: p <
.001; ST: p < .05), while no differences were observed at the higher velocities. When expressed
in relative values, the torque deficits of the patient groups averaged 34 ± 2% in isometric and 24
± 1% at 60° · s–1. The Group x Leg effect resulted from the bilateral similarity of the Ctrl group,
while each patient presented around 30% torque deficit in the Op leg (KJ, ST: p < .001) and 20%
in the Nop leg (KJ, ST: p < .001) as compared to the Ctrl leg values. The Velocity x Leg effect
reflected a significant 30 ± 10% inter-leg difference in the isometric testing condition (p < .001)
but none in the other tests. More specifically, the interaction between the three factors revealed
in the isometric condition and for the KJ patient group only, a significant torque deficit in the Op
leg as compared to the Nop leg (p < .001). When compared to the unilateral Ctrl leg values, each
patient group presented at least 30% lower torque values in the Op leg and 25% in the Nop leg.
Post hoc analysis revealed for the Op leg significant torque deficits in isometric (KJ, ST: p <
.001) and at 60° · s–1 concentric velocity (KJ: p < .001; ST: p < .05). In the Nop leg, both patient
groups had a significant torque deficit in isometric strength test (p < .001); the ST group
presenting also deficits in the concentric tests at 60° · s–1 (p < .05) and at 240° · s–1 (p < .05).

In quadriceps EMG activation, the analysis revealed a significant Group and Leg interaction
(F2,15 = 3.71; p < .05; EI = 16.6%) that reflected 20 ± 3 % lower quadriceps activation in the Op
leg of the patient groups (KJ: p < .05; ST: p < .01) as compared to the reference leg of the Ctrl
group. The inter-leg difference did not reach the level of statistical significance.

For the hamstring co-activation, the analysis revealed a significant main effect of Group (F2,15 =
4.91; p < .05; EI = 79.1%). The relative co-activation level that averaged 27 ± 3% for both legs
in the Ctrl group was significantly lower (p < .05) in the KJ and ST patient groups, with 16 ± 4%
and 14 ± 3%, respectively.

Maximal Knee Flexion Tests

Similar to the results in extension, the Ctrl group presented only minor and non-significant
differences between the left and the right leg values: 3 ± 5% in flexion torque, 6 ± 20% in
hamstring AEMG activity and 1 ± 0.5% in quadriceps co-activation. For this reason, the Ctrl
group is represented in Figure 3 by the mean leg values together with the respective Op and Nop
leg values of each patient group. For the statistical comparisons, the Nop and Op legs of the
patient groups are compared to the respective left and right legs of the Ctrl group.
Figure 3 — Patient group mean (±SD) values for the operated (Op) and non operated (Nop) flexion leg
torque, averaged hamstring EMG (AEMG) activity, and relative quadriceps co-activation level in each of the
4 testing conditions (0, 60, 180, 240° · s–1). The control (Ctrl) group is represented by its mean left and right
leg values. (Note: See color originals of this figure at <www.humankinetics.com/ejss>.)

For the flexion torque, the statistical analysis revealed significant main effects of Leg (p < .01)
and Velocity (p < .001) accompanied by significant interactions, although of limited influence (<
2%) between Group and Leg (F2,15 = 9.57; p < .01; EI = 1.7%) and Leg and Velocity (F3,45 =
4.33; p < .01; EI = 0.6%). As compared to the Ctrl leg values, each patient group presented a
significant flexion torque deficit of at least 20% in the Op leg (KJ, ST: p < .001), with a slight
7% but significant deficit (p < .05) in the Nop leg of the KJ group. The Leg x Velocity effect
reflected significant inter-leg difference, but only in the isometric testing condition (p < .001).

For the hamstring AEMG activity, the analysis did not reveal any significant main or interaction
effects. As shown in Figure 3, patient and Ctrl groups presented very similar and constant
AEMG values among the 4 testing conditions.
For the quadriceps co-activation during the flexion movement, the analysis indicated significant
main effects of Leg (F1,15 = 5.16; p < .05; EI = 8.8%) and Velocity (F3,45 = 5.87; p < .01; EI =
27.5%) but no interactions. With regard to the leg effect, its influence may be considered as
negligible, as the inter-leg difference in co-activation corresponded to only 1%. The velocity
effect revealed that, independent of the subject group, quadriceps co-activation was slightly
lower (7.5%) in the isometric as compared to dynamic tests. In the latter tests the co-activation
level averaged 10% at 60° · s–1 (p < .01) and 9.5% at 180° · s–1 (p < .05).

Discussion
The major findings of the present study were as follows: despite the limited number of subjects,
both ACL-reconstructed patient groups presented significant extension and flexion torque
deficits in the Op leg as compared to Ctrl subjects with comparable anthropometric
characteristics and athletic background. Independently of the type of surgery, the torque deficits
were more prominent in extension than in flexion (32 vs. 22%), and were larger in isometric and
at slow isokinetic concentric velocity. The torque deficits were associated in the patient groups,
with 20% lower EMG of the quadriceps muscle group. Interestingly, the parallel analysis of Nop
leg values revealed a mean 21% extension torque deficit in the patient groups and 7% flexion
torque deficit in the KJ patient group, with no significant differences in the agonist EMG. On the
other hand, the hamstring co-activation level was bilaterally and independent of the testing
condition significantly lower in the KJ and ST patient groups than in the Ctrl group (16 and 14%
vs. 27%). The inter-leg comparison revealed for both patient groups significant extension and
flexion torque deficits in the Op leg, particularly in the isometric testing condition.

A first explanation for these findings may be that disuse and knee pain are more likely to affect
primarily the extensor muscle chain, leading to quadriceps amyotrophy and associated
performance reduction. It should be mentioned that disuse amyotrophy alone does not account
for the prolonged decrease in knee extension strength observed in acute disuse (16) and chronic
ACL-tear (17). As emphasized by recent studies on ACL patients (12, 18), more attention should
be paid to the additional contribution of inhibition to the measured muscle weakness and
associated long-term functional and structural changes of the knee joint after surgery. With
regard to the potential role of deficit in activation of the knee extensors of the operated leg, the
present EMG analysis revealed a significantly lower quadriceps activation in both patient groups
as compared to the Ctrl group. As illustrated in Figure 1 for one KJ patient at 60° · s–1, this
activation deficit was not position-specific but covered the entire range of motion, resulting in a
slow rate and reduced maximal torque development. Similarly, incomplete voluntary activation
of knee extensor muscles has been observed in patients after ACL rupture (12, 13, 18) and ACL
reconstruction (12) as well as after menisectomy (19), anterior knee pain (20), experimental knee
effusion (21), and extensive knee injury (22, 23). In the latter study of Hurley et al. (23), it was
shown that severe initial knee extensor inhibition (30–45%) may interfere with the rehabilitation
process and compromise the functional recovery, whereas a small amount of initial inhibition
seems to allow for strength gains early in the rehabilitation process. This emphasizes the interest
to perform quite early in the recovering period combined EMG and torque tests in order to detect
potential inhibition that could delay the functional recovery. On the other hand, the underlying
neurophysiological mechanisms of the observed activation deficits are not yet fully understood.
Intervention of several protective reflexes induced primarily by ACL strain (14, 28), but also by
stretch of joint capsule and associated muscle-tendon complex (4, 14, 28, 38), may reduce the
anterior pulling force on the tibia and knee joint, via the extensor muscle inhibition and
hamstring activation. In case of ACL-deficiency, Beard et al. (29) reported an increased latency
of the protective reflexes that persisted after ACL-reconstruction. Interestingly, evidence of
reinnervation of free patellar autograft used for ACL-reconstruction has been recently reported in
dogs by Barrack et al. (30). These authors suggested that a successful reconstruction includes a
tendon graft reinnervation by mechanoreceptors, allowing an afferent feedback to stabilize the
knee. It is noticeable that such an effective graft reinnervation, observed in 50% of the animals,
took place at least 6 months post-surgery.

The variability in the relative content of mechanoreceptors and free nerve endings are large
among different anatomical structures. It is therefore important to consider the delays inherent to
the different afferent pathways. As shown in cat studies (31), nociceptors are innervated by type
IV (C) afferent fibers with a slow conduction velocity (<1 m · s–1), whereas mechanoceptors are
innervated by type III (A delta) afferent fibers with faster conduction velocity (2.5 to 20 m · s–1).
Nociceptor-induced reflex activation of the hamstring muscle group does not appear to be fast
enough to counteract the anterior draw movement of the tibia during rapid leg extension. In
addition, strain-induced reflex inhibition does not explain the present observation of EMG
deficits that lasted in most subjects from the initiation till the end of contraction, both in
isometric and isokinetic conditions. On the other hand, studies based on healthy subjects have
revealed reflex inhibition of agonistic muscle activation in situations of high force production
(24, 25), with a trend to be overcome in trained subjects (26). In the present study, however, the
observed deficit in quadriceps activation was not more accentuated in the isometric strength test,
which allows for the production of larger torques, than in the other testing conditions (Figure 2).
This may be explained partly by the fact that the isometric extension test was performed at 65° of
knee flexion, a position in which quadriceps contraction has been reported to produce no anterior
or posterior tibial translation (27).

Among other stimuli that could sensitize free nerve endings, chemicals released during acute
inflammation have been reported in cats to lead to hyper-excitability of spinal neurons for the
afferent input from the inflamed knee as well as for the input from regions of the contralateral
leg (31, 32). This effect is counteracted by progressive enhancement of the tonic descending
inhibition (33, 34). In humans, both voluntary and reflex EMG inhibitions have been reported to
last over a few days after eccentric exercise induced ultrastructural muscle damage (35). It is thus
suggested that polymodal group III and IV muscle afferent fibers might also be involved and
lead to a pre-synaptic inhibition of alpha motoneurons in case of knee effusion. This was not,
however, the case in the present study as none of the patients experienced any knee pain and/or
effusion. Finally, considering the protective antagonistic hamstring effect against the anterior
tibial load (4, 14, 36), much emphasis has been put in rehabilitation programs on the hamstring
strengthening after ACL rupture (37–39). It should be mentioned, however, that Bencke et al.
(40) were unable to demonstrate a significant effect of a prophylactic training program on
healthy athletes. In the present study, the hamstring co-activation of both patient groups was
about 10% lower than in the Ctrl group in all tested conditions but remained in the 15–35%
range reported in the literature on isokinetic strength tests (15). It is suggested, however, that due
to the automatic braking of the leg lever arm by the dynamometer, isokinetic strength tests may
not be considered as the proper way to evaluate the exact co-activation level of natural lower
limb movements.

A second hypothesis, based on the so-called “quadriceps avoidance gait pattern,” would be the
expected neuromuscular disuse effect that would be learnt before the surgery when the knee was
unstable (41). As an acute reconstruction is rarely performed, many patients walk without an
intact ACL for some weeks or months before the surgery. This is expected to lead also to a
progressive deficiency in locomotor type actions, thus favoring the occurrence of bilateral
functional deficits. In support of this hypothesis, our results revealed extension and flexion
torque deficits in the Op leg as well as in the Nop leg of both patient groups as compared to an
age- and activity-matched Ctrl group. Similarly, Urbach et al. (18) reported moderate (~15%) but
significant bilateral activation deficit in isometric strength test in unilateral ACL deficient
patients as compared to a matched control group. These authors suggested that stabilization of
the knee joint by ACL-reconstruction might eliminate voluntary activation deficits. This
hypothesis is not supported, however, by the present activation deficits observed in both
isometric and isokinetic strength tests of the Op leg 4 months post ACL-reconstruction (Figure
2). Similarly, Hurley et al. (23) reported close to 20% bilateral inhibition of the knee extensors
22 months post ACL-surgery. In both of these studies as well as in the present results (Figures 2
& 3), the inter-leg functional differences were found to be only minor. Based on the observed
voluntary-activation deficit in the uninjured side, Urbach et al. (18) and Hurley et al. (23)
suggested that functional tests might underestimate the exact deficit when the uninjured limb
serves as reference. Although the present study did not reveal any significant activation deficit in
the Nop leg of the patient groups, the observed large bilateral torque deficits would suggest also
that functional tests might not be valid when the uninjured limb provides reference values.
Supporting the interest of a functional comparison to normal subjects, De Vita et al. (42)
reported normal walking kinematic patterns, although altered joint torque and power patterns, 6
months post ACL-reconstruction and accelerated rehabilitation. In activities such as walking,
ramp descending, stair ascending and descending, running, and cross-cutting, Ciccotti et al. (43)
reported greater knee extensor and flexor EMG activity in ACL-deficient patients as compared to
those of uninjured subjects, whereas ACL-reconstructed subjects (2 to 3 years post-surgery)
generally produced EMG profiles that were statistically similar to the normal subjects. Referring
to Hurley et al. (13), it is suggested that the specific activation deficit in the knee extension
muscles of the Op leg might compromise for a longer time period the functional recovery of the
Op leg as compared to the other. Inter-leg differences in agonistic muscle inhibition is thus
expected to interfere in the respective functional recovery of each leg. The present findings
emphasize the fact that EMG analysis may reveal persisting neuromuscular deficits that could
not be detected by kinetic or kinematic parameters alone.

Conclusion
The present study revealed bilateral extension and flexion torque deficits in ACL-reconstructed
patient groups. This would suggest potential underestimation of the actual torque deficits in the
Op leg when the analysis is based on inter-leg torque comparison only. On the other hand, the
EMG analysis suggests that protective activation mechanisms might occur and would result in
slower functional recovery, especially in the extensor muscle chain. In order to detect and
counteract such a vicious circle, it is recommended that the strength and EMG of agonist and
antagonist muscles be measured, especially during forceful actions, around 6 months post-
surgery.

References
1. Kennedy JC, Weinberg HW, Wilson AS. 1974. The anatomy and function of the anterior
cruciate ligament as determined by clinical and morphological studies. J Bone Joint Surg
56-A:223-35.
2. Butler DL, Noyes FR, Grood ES. 1980. Ligamentous restrains to anterior-posterior drawer in
the human knee. A biomechanical study. J Bone Surg 62-A:259-70.
3. Markolf KL, Gorek JF, Kabo JM, Shapiro MS. 1990. Direct measurement of resultant forces
in the anterior cruciate ligament. J Bone Joint Surg 72-A:557-67.
4. Renström P, Arms SW, Stanwyck TS, Johnson RJ, Pope MH. 1986. Strain within the anterior
cruciate ligament during hamstring and quadriceps activity. Am J Sports Med 14:83-87.
5. Grabiner MD. 1992. Ligamentous mechanoreceptors and knee joint function: the neurosensory
hypothesis. In: Grabiner MD, editor. Current issues in biomechanics. Champaign, IL:
Human Kinetics. pp. 237-53.
6. Paulos L, Noyes FR, Grood E, Butler DL. 1981. Knee rehabilitaion after anterior cruciate
ligament reconstruction and repair. Am J Sports Med 9:140-49.
7. Kennedy JC, Alexander IJ, Hayes KC. 1982. Nerve supply of the human knee and its
functional importance. Am J Sports Med 10:329-35.
8. De Loës M, Dahlsteedt LJ, Thomée R. 2000. A 7-year study on risks and costs of knee injuries
in male and female youth participants in 12 sports. Scand J Med Sci Sports 10(2):90-97.
9. Croisier JL, Delcour JP, Dupont G, Crielaard JM. 1993. Evaluation isocinétique des genoux
ligamentaires en pré et postopératoire. La Lettre du Médecin Rééducateur. October 10-
11.
10. Rochcongar P, Robineau S, Legall F, Jan J. 1999. Le suivi après plastie ligamentaire. In:
Rodineau J, Saillant G, editors. Les lésions isolées du ligament croisé antérieur.
Masson\bb\. pp. 248-54.
11. Fossier E, Christel P, Dijan P, Darman Z, Witvoët J. 1993. Principes et intérêt de l’évaluation
isocinétique dans les ruptures du ligament croisé antérieur. Revue Chir Orthop 79:615-
24.
12. Suter E, Herzog W. 2000. Does muscle inhibition after knee injury increase the risk of
osteoarthritis? Exercise Sports Sci Rev 28:15-18.
13. Hurley MV, Jones DW, Wilson D, Newham DJ. 1992. Rehabilitation of quadriceps inhibited
due to isolated rupture of the anterior cruciate ligament. J Orthop Rheum. 5:145-54.
14. Solomonov M, Baratta R, Zhou BH, Shoji H, Bose W, Beck C, D’Ambrosia R. 1987. The
synergistic action of the anterior cruciate ligament and thigh muscles in maintaining joint
stability. Am J Sports Med 15:207-13.
15. Aagaard P, Simonsen EB, Andersen JL, Magnusson SP, Bojsen-Moller F, Dyhre-Poulsen P.
2000. Antagonist muscle coactivation during isokinetic extension. Scand J Med Sci
Sports 10(2):58-67.
16. Perhonen M, Komi PV, Häkkinen K, Von Bondsdorff H, Partio E. 1992. Strength training
and neuromuscular function in elderly people with a total knee endoprostesis. Scand J
Med Sci Sports 2:234-43.
17. Lorentzon R, Elmqvist LG, Sjöström M, Fagerlund M, Fugl-Meyer AR. 1989. Thigh
musculature in relation to chronic anterior cruciate ligament tear: muscle size,
morphology, and mechanical output before reconstruction. Am J Sports Med. 17:423-29.
18. Urbach D, Nebelung W, Weiler HT, Friedemann A. 1999. Bilateral deficit of voluntary
quadriceps muscle activation after unilateral ACL tear. Med Sci Sports Exer 31:1691-96.
19. Shakespeare DT, Stokes M, Sherman KP, Young A. 1985. Reflex inhibition of the
quadriceps after menisectomy: lack of association with pain. Clin Physiol 5:137-44.
20. Suter E, Herzog W, Desoua K, Bray RC. 1998. Inhibition of the quadriceps muscles in
patients with anterior knee pain. J Appl Biomech 14:360-73.
21. Wood L, Ferrell WR. 1988. Pressures in normal and acutely distended human knee joints and
effects on quadriceps maximal voluntary contractions. Q J Exp Physiol 73:305-14.
22. Newham DJ, Hurley MV, Jones DW. 1989. Ligamentous knee injuries and muscle
inhibition. J Orthop Rheum 2:163-73.
23. Hurley MV, Jones DW, Newham DJ. 1994. Arthrogenic quadriceps inhibition and
rehabilitation of patients with extensive traumatic knee injuries. Clin Sci 86:305-10.
24. Westing SH, Cresswell AG, thorstesson A. 1991. Muscle activation during maximal
voluntary eccentric and concentric knee extension. Eur J Appl Physiol 62:104-8.
25. Seger JY, Thorstensson A. 1994. Muscle strength and myoelectric activity prepubertal and
adult males and females. Eur J Applied Physiol 69:81-87.
26. Aagaard P, Simonsen EB, Trolle M, Bangsbo J, Klausen K. 1996. Specificity of training
velocity and training load on gains in isokinetic knee joint strength. Acta Physiol Scand
156:123-29.
27. Sawhney R, Dearwater S, Irrgang JJ, Fu FH. 1990. Quadriceps exercise following anterior
cruciate ligament reconstruction without anterior tibial displacement. Presented at the
American Conference of the American Physical Therapy Association, Anaheim, CA.
28. Baratta RV, Solomonow M, Zhou BH, Letson D, Chuinard R, D’Ambrosia R.1988.
Muscular coactivation. The role of the antagonist musculature in maintaining knee
stability. Am J Sports Med 14:83-87.
29. Beard DJ, Kyberd PJ, Fergusson CM, Dodd CA. 1993. Proprioception after rupture of the
anterior cruciate ligament. J Bone Joint Surg 75-B:311-15.
30. Barrack RL, Lund PJ, Munn BG, Wink C, Happel L.1997. Evidence of reinnervation of free
patellar tendon autograft used for anterior cruciate ligament reconstruction. Am J Sports
Med 25:196-202.
31. Grigg P, Schaible H-G, Schmidt RF. 1986. Mechanical sensitivity of group III and IV
afferents from posterior articular nerve in normal and inflamed cat knee. J neurophysiol
55:635-43.
32. Neugebauer V, Schaible HG. 1990. Evidence for a central component in the sensitization of
spinal neurons with joint input during development of acute arthritis in cats’ knee. J
Neurophysiol 64:299-311.
33. Schaible HV, Neugebauer V, Cervero F, Schmidt RF. 1991. Changes in tonic descending
inhibition of spinal neurons with articular input during the development of acute arthritis
in the cat. J Neurophysiol 66:1021-32.
34. Cervero F, Schaible HG, Schmidt RF. 1991. Tonic descending inhibition of spinal cord
neurones driven by joint afferents in normal cats and in cats with an inflamed knee joint.
Exp Brain Res 83:675-78.
35. Komi PV, Nicol C. 2000. Stretch-shortening cycle fatigue. In: Nigg BM, MacIntosh BR,
Mester J, editors. Biomechanics and biology of movement. Champaign, IL: Human
Kinetics. pp. 385-408.
36. Yasuda K, Sasaki T. 1987. Exercise after anterior cruciate ligament reconstruction. The
forces exerted on the tibia by the separate isometric contractions of the quadriceps or the
hamstrings. Clin Orthop 220:275-83.
37. Tegner T, Lysholm J, Lysholm M, Gillquist J. 1986. Strengthening exercises for old cruciate
ligament tears. Acta Physiol Scand 57:130-34.
38. Solomonov M, Baratta R, D’Ambrosia R. 1989. The role of the hamstrings in the
rehabilitation of the anterior cruciate ligament-deficient knee in athletes. Sports Med
7:42-48.
39. Beynnon BD, Flemming BC, Johnson RJ, Nichols CE, Rentröm PA, Pope MH. 1995.
Anterior cruciate ligament strain behavior during rehabilitation exercises in vivo. Am J
Sports Med 23:24-34.
40. Bencke J, Naesborg, H., Simonsen EB, Klausen K. 2000. Motor pattern of the knee joint
muscles during side-step cutting in European team handball. Scand J Med Sci Sports
10:68-77.
41. Berchuck M, Andriacci TP, Bach BR, Reider B. 1990. Gait adaptations by patients who have
a deficient anterior cruciate ligament. J Bone Joint Surg 72-A:871-77.
42. De Vita P, Hortobagyi T, Barrier J. 1998. Gait biomechanics are not normal after anterior
cruciate ligament reconstruction and accelerated rehabilitation. Med Sci Sports Exerc
30:1481-88.
43. Cicotti MG, Kerlan RK, Perry J, Pink M. 1994. An electromyographic analysis of the knee
during functional activities. II. The anterior cruciate ligament-deficient and –
reconstructed profiles. Am J Sports Med 22:651-58.

Acknowledgment
The authors are grateful to Dr Reinoud Bootsma (University of the Mediterranean, Marseilles,
France) for his expert consultation to the statistical analysis.

About the Authors

C. Nicol <nicol@laps.univ-mrs.fr>, N. Gouby, , C. Flin, J.M. Viton, and A. Delarque are with
the Department of Biology of Physical Activity (UPRES-EA 3285), Marey Institute, University
of the Mediterranean, 163 avenue de Luminy, BP 910, 13288, Marseilles, France. J.M.
Coudreuse and C. Brunet are with Service of Sports Medicine, Salvator Hospital, Marseilles,
France. P.V. Komi is with the Neuromuscular Research Center, University of Jyväskylä,
Jyväskylä, Finland.\bb\

Caroline Nicol received her PhD from St Etienne University, France, after a 4-year stay in
Professor Komi’s laboratory in Jyväskylä, Finland. She is assistant professor of the Department
of Biology of Physical Activity at the University of the Mediterranean in Marseilles. She is a
member of the Scientific Board of ECSS. Her main research field deals with the neuromuscular
adjustments to fatigue, especially during the delayed recovery that characterizes stretch-
shortening cycle and eccentric-type exercises.
Nicolas Gouby is an undergraduate student of the Department of Biology of Physical Activity at
the University of the Mediterranean, Marseilles, France. His main topic of research deals with
the neuromuscular adjustments to muscle and joint afferent disturbances such as those associated
with ACL rupture and reconstruction.
Jean-Marie Coudreuse, MD, is a specialist in sports medicine and traumatology, and a graduate
student in Sport Sciences at the University of the Mediterranean. His graduate work dealt with
ACL-reconstructed patients. His research interests include the optimization of isokinetic testing
and rehabilitation protocols of different types of sport-induced pathologies.
Christian Flin, MD, is specialist in Physical and Rehabilitation Medicine in Marseille’s army
hospital. Both in practice and research, his major work deals with rehabilitation of the musculo-
skeletal system.
Jean-Michel Viton and Alain Delarque, MD, PhD, are both specialists in Physical and
Rehabilitation Medicine and professors at the Marseilles Medical School. Their main field of
interest deals with posture and movement analysis in patients with pathologies of the musculo-
skeletal system (i.e., knee pathologies, lower limb amputation). They organize the annual
Eramus intensive program on posture and movement.
Christian Brunet, MD, PhD, is a specialist in anatomy, biomechanics, and general surgery. His
main research work deals with anatomo-clinical approach of car crash–induced injuries of
cephalic, cervical, and lower limb extremities; mathematical modeling in structural and
functional anatomy; management of pancreatic and liver injuries; and perihepatic prosthesis.
Paavo V. Komi received his PhD from Pennsylvania State University. He is a professor in
biomechanics, Head of the Department of Biology of Physical Activity, and the Director of the
Neuromuscular Research Center at the University of Jyväskylä, Finland. He has served as
president of three international scientific organizations (ISB, ICSSPE, and ECSS). He is on the
editorial board of several international scientific journals. His research over the years and
presently deals with many aspects of neuromuscular performance, with a special focus on in vivo
mechanics and reflex-induced stiffness regulation of human skeletal muscle.