- Email: [email protected]
Are hamstrings activated to counteract shear forces during isometric knee extension efforts in healthy subjects?
Journal of Electromyography and Kinesiology 14 (2004) 307–315 www.elsevier.com/locate/jelekin
Are hamstrings activated to counteract shear forces during isometric knee extension efforts in healthy subjects? Idsart Kingma , Sietske Aalbersberg, Jaap H. van Diee¨n Institute for Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit, Van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands Received 2 July 2003; received in revised form 15 December 2003; accepted 6 January 2004
Abstract The hamstring muscles have the potential to counteract anterior shear forces at the knee joint by co-contracting during knee extension efforts. Such a muscle recruitment pattern might protect the anterior cruciate ligament (ACL) by reducing its strain. In this study we investigated to what extent co-activation of the knee flexors during extension efforts is compatible with the hypothesis that this co-activation serves to counteract anterior tibial shear forces during isometric knee extension efforts in healthy subjects. To this aim, it is investigated whether co-activation varies with the required knee extension moment, with the knee joint angle, and with the position of the external flexing force relative to the knee joint. With unaltered moment and muscle activation, distal positioning of the flexing force on the tibia causes higher resultant (muscular plus external) forward shear forces at the knee as compared to proximal positioning. In ten subjects, knee flexor and extensor EMG was measured during a quasi-isometric positioning task for a range (5–50 degrees) of knee flexion angles. It was found that the co-activation of the knee flexors increased with the extension moment, but this increase was less than proportional (p< 0:001). The extension moment increased 2.7 to 3.4 times, whereas the activation of Biceps Femoris and Semitendinosus increased only a factor 1.3 to 2.0 (joint angle dependent). Furthermore, a strong increase in co-activation was seen near full extension of the knee joint. The position of the external extension load on the tibia did not affect the level of co-contraction. It is argued that these results do not suggest a recruitment pattern that is directed at reduction of anterior shear forces in the knee joint during sub-maximal isometric knee extension efforts in healthy subjects. # 2004 Elsevier Ltd. All rights reserved. Keywords: EMG; Knee; ACL; Hamstrings; Shear forces
1. Introduction During knee extension efforts, quadriceps activation causes the tibia to be pulled forward relative to the femur. Even in full extension, the patellar tendon has a considerable angle with the long axis of the tibia, causing knee extension efforts to be accompanied by a substantial shear force on the tibia [10]. In healthy subjects, the anterior cruciate ligament is assumed to play an important role in counteracting forward shear forces on the tibia. Full activation of the quadriceps is estimated to cause a force in the anterior cruciate ligament (ACL) of over 500 N [14]. This force depends on Corresponding author. Tel.: +31-20-4448492; fax: +31-204448529. E-mail address: [email protected] (I. Kingma).
1050-6411/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2004.01.003
v
the knee angle with a maximum being reached at 15 of knee flexion [14]. Co-contraction of the hamstring muscles can potentially assist in counteracting forward tibial shear because the hamstrings’ force has a backward shear component. However, a disadvantage of this strategy is that, in order to maintain the same extension moment, the quadriceps force also has to be increased to counteract the flexion moment of the hamstrings. In a simulation study, Liu and Maitland [12] estimated that cancelling out forward tibial shear by hamstrings co-contraction would result in about a doubling of the required quadriceps force in gait. Due to the relatively small knee extension moments required in gait, protective co-contraction may not be needed in healthy subjects. However, when large extension moments are required, for instance up to 100 Nm in the forward lunge movement, substantial co-contrac-
308
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
tion is found [2]. In addition, mechanical [18] as well as electrical [11] stimulation of the ACL has been shown to elicit hamstrings activation. Apart from fatigue [19], three main factors might affect the amount of hamstrings co-contraction. These factors are (1) the required extension moment, (2) the joint angle and (3) the position of the external flexing force with respect to the knee joint. With respect to the first factor, increasing quadriceps activation leads to an increase of the forward shear force on the tibia. If the hamstrings play a role in counteracting those shear forces in healthy subjects, increased hamstrings activation is expected when the extension effort increases. However, Grabiner et al. [6] reported an unchanged hamstrings activation with extension efforts ranging from 10 to 90% maximum voluntary contraction. Possibly, the absence of increased hamstrings co-activation in the study of Grabiner et al. [6] is related to the second factor, the influence of the knee joint angle. The v knee flexion angle in their experiment was 10–20 . Hamstrings co-contraction appears to be effective in reducing forward tibial shear forces only when the knee v v is flexed more than 15 [3] to 22 [13]. With respect to the third factor, the position of the external flexing force, Zavatsky et al. [20] estimated that, with a flexing force applied at 40 cm distance to the knee joint, the extension effort by the quadriceps would result in an ACL force of 5 times the external flexing force. With the same moment applied by an external force at a distance of 10 cm, the ACL would not be loaded because the external force is large enough to fully cancel out the forward shear force of the quadriceps [20]. A protective hamstrings co-contraction would therefore be useful when the external flexing force is applied at a distance of 40 cm, but it would be superfluous when the external flexing force is applied at a distance of 10 cm from the knee joint. The aim of this study was to find out to what extent co-activation of the knee flexors during extension efforts is compatible with the hypothesis that this coactivation serves to counteract anterior tibial shear forces during sub-maximal isometric knee extension efforts in healthy subjects. To this aim, it is investigated whether hamstrings co-activation, during a quasi-isometric positioning task, varies with (1) the required knee extension moment, (2) the joint angle and (3) the position of an external flexing force relative to the knee joint. In addition, the actual knee angle during the positioning task is measured to evaluate the difficulty of maintaining a constant target angle.
2. Methods 2.1. Subjects and procedure Ten healthy young subjects (7 males and 3 females, age 22 2 yrs, bodyweight 71:8 8:9 kg, shank length 393:6 52:5 mm) participated in the experiment after providing informed consent. In a seated position with v the hip joint in approximately 70 flexion, the knee joint of the right leg of the subjects was aligned with the axis of the loading apparatus. (External) flexing moments of 20, 50 and 80 Nm were attained using an external force that was applied perpendicular to the long axis of the shank. Since the moment is defined as force times moment arm, the ratio between the (external) posterior force and the flexion moment depends on where the force is applied. For all three moment levels, we applied the external force at three distances from the knee joint axis (95, 212 and 307 mm). Therefore, for each moment level, the three external force positions resulted in three levels of external backward shear force (Fig. 1). Note that moment levels were constant within subjects over external force positions. However, moments were not completely constant between subjects and between knee flexion angles due to the moment caused by the lower leg mass. With the external force at 95 mm (and ignoring the lower leg mass), the backward shear force caused by the external force is assumed to have about the same magnitude as compared to the forward shear force due to quadriceps contraction, so that the extension effort would not cause ACL loading [20]. At 307 mm, the backward shear force due to the external force is considerably smaller than the forward shear force due to the activation of the quadriceps, resulting in an increased forward shear force challenge to the knee joint (and thus resulting in an increased loading of the ACL) compared to the external force position close to the joint. The order of the moments and distances was randomised over subjects. From markers attached to the leg, subjects received real-time feedback regarding their knee joint angle on a computer screen in front of them. For each measurement, subjects started with their knee at approximately v 90 flexion. The subjects were requested to extend their knee to a target angle (indicated by a line on the computer screen). The external force was applied as soon v as the knee joint reached an angle in approximately 70 flexion. After holding the knee at the target angle for 5 s, the subjects were allowed to flex the knee and relax for at least 15 s before the next target angle was presented. Only the isometric part of each trial was used for further analysis. Target angles ranged from 5 to v v 50 , in steps of 5 , so that a total of 90 extension efforts were performed (10 angles 3 moment levels 3 force positions). In addition, flexion efforts were per-
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
309
Electromyographic data (EMG) from selected leg muscles were obtained, using disposable surface EMGelectrodes (Ag/AgCl; square 5 5 mm pick-up area) that had been attached at a center-to-center electrode distance of 2.5 cm after cleaning and abrasion of the skin. Surface EMG was measured on the Rectus Femoris (RF), Vastus Medialis (VM), Vastus Lateralis (VL), Biceps Femoris (BF) and Semitendinosus (ST). Electrode locations were based on Seniam guidelines [7]. During the holding task, EMG signals were sampled at 1000 Hz (Porti 17, TMS, Enschede, The Netherlands; 22 bits AD conversion after 20x amplification, input impedance > 1012X, CMMR >90 dB for the relevant range of frequencies), after band-pass filtering with 10 and 350 Hz cut-off frequencies. At the same time, marker positions were sampled at 100 Hz with the Optotrak system. EMG and marker position data were synchronised in time using a pulse generated by the Optotrak system. 2.3. Data processing
Fig. 1. A graphical representation of the experimental setup. A constant external moment was applied at the knee through the use of a wheel. Subjects were requested to hold the knee at a specific angle while they received real time feedback regarding the knee angle. The three dotted arrows indicate the locations where the flexion force was applied during the extension trials. Those locations (95, 212 and 307 mm distance from the knee joint) were obtained by displacing the attachment to the knee in proximal or distal direction. By increasing the distance from the knee joint, the magnitude of this force was reduced without affecting the applied flexion moment.
formed at the same 10 target angles, but with only one moment level (60 Nm) and with only one force position (307 mm). Subjects were allowed to rest as long as they needed when they felt fatigued. 2.2. Measurements An opto-electronic movement registration system (Optotrak, Northern Digital Inc., Canada) was used to provide a target angle and real-time feedback of the knee joint angle to the subjects. LED markers were attached to the medial side of the leg at the medial malleolus, the medial tibial condyle, the medial condyle of the femur, and proximally at the medial side of the upper leg. Markers were aligned in such a way that the real-time feedback showed an angle of zero degrees between the two markers on the upper leg and the two markers on the lower leg when the subject placed the leg on a flat table and relaxed the muscles as much as possible.
EMG signals were rectified and averaged over 3 s during the isometric part of each trial (starting 1 s after the target angle was reached). As noted before, the net moment at the knee joint was not constant over target angles, due to the weight of the lower leg, the foot and the arm of the loading device, which caused an additional moment that varied with the knee joint angle. This additional moment was calculated for each positioning task using the recorded marker positions, the mass and centre of mass location of the foot and lower leg according to Plagenhoef [15], and the mass and centre of mass location of the loading device. From those data, the additional net moment due to the foot, lower leg and loading device was calculated to v range from 5:1 0:6 Nm at 50 of knee flexion to v 16:1 3:1 Nm at 5 of knee flexion. The success in maintaining a constant target angle over the 3 s period was quantified using the skin marker positions to calculate the standard deviation of the knee angle in that period. 2.4. Statistics For all muscles, repeated measures ANOVA’s were applied to the rectified and averaged EMG data of the extension efforts. The independent variables were joint angle (10 levels), moment level (3 levels), and external force position (3 levels). The same ANOVA’s were applied to the ratio of BF and ST activation to extension moment level and to the standard deviation of the joint angle during the recorded period. In cases where significant interactions with joint angles were found, separate ANOVA’s with moment level and external
310
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
force position as independent variables were applied for each joint angle. 3. Results An example of the EMG (rectified and averaged over 3 s) of two selected muscles in one external load position shows that there was, as can be expected, substantial variation among subjects in the level of the raw EMG (Fig. 2). However, the pattern of change with joint angle was comparable over subjects. During extension efforts all muscles showed a significant (p< 0:001) and evidently non-linear change in activ vation with joint angle (Fig. 3). Especially below 20 of flexion, there was a strong increase in co-contraction of the hamstring muscles with the quadriceps muscles with decreasing knee flexion in all muscles. This increase in co-contraction could only be explained to a minor extent by an increase in actual net moment due to the mass of the lower leg, foot, and arm of the loading device. For instance, for the flexors (and even more v for the extensors) the activity at a target angle of 5 was roughly twice as high as compared to a target v angle of 20 . In contrast, the increase in actual moment amounted only on average 3.2 Nm (3.4% of the total moment at the highest moment level up to 9.7% of the total moment at the lowest moment level). The high level of co-contraction near full knee extension v becomes apparent when, at a target angle of 5 , the activation of the hamstring muscles is compared between flexion and extension efforts. At the strongest extension effort of 96:1 3:1 Nm (80 Nm moment due
to the external force + 16:1 3:1 Nm additional moment due to lower leg, foot and loading device) the BF activation was about the same as the activation at a flexion effort of 43:9 3:1 Nm (60 Nm moment due to the external force + 16:1 3:1 Nm additional moment due to lower leg, foot and loading device). Not surprisingly, the activation of all muscles increased with the applied moment level (p< 0:002). However, hamstring activation did not increase proportionally with the moment level, since the ratio of hamstring activation to (actual) extension moment was significantly affected by moment level (p< 0:001 for both the BF and ST). In fact, the hamstrings coactivation increased less than proportional with the moment level, since the highest level of extension moment was (dependent on the joint angle) 2.7–3.4 times the lowest level of extension moment, whereas the BF and ST activation at the highest extension moment level ranged (dependent on the joint angle) from 1.3 to 2.0 times the activation at the lowest extension moment level. A significant moment level with target angle interaction was found for all muscles (p< 0:002). Fig. 3 shows that the effect of moment level was larger at smaller knee flexion angles. Separate ANOVA’s per joint angle showed that the effect of moment level was significant for all three quadriceps muscles in all joint angles (p< 0:05). For the BF the effect of moment level was v significant at all joint angles except at 50 of flexion, and for the ST the effect of moment level was only sigv nificant for knee flexion angles less than 30 . There was no significant main effect of external force position
Fig. 2. Average EMG against flexion angle during isometric knee extension and flexion efforts in a joint angle positioning task. This figure shows the variation in EMG level over subjects. EMG data (VM and BF) for one external load position (307 mm), averaged over 3 moment levels, is shown for individual subjects. Each line represents the EMG of one subject. a.u. = arbitrary units.
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
311
Fig. 3. Average EMG against flexion angle during isometric knee extension and flexion efforts in a joint angle positioning task. In each graph, one line represents the flexion efforts (60 Nm moment applied by an external force at a distance of 307 mm from the knee joint). The other three lines are extension efforts, averaged over three distances (95, 212 and 307 mm) relative to the knee joint at which the external force was applied. Those three lines represent the three moment levels (20, 50 and 80 Nm). The five graphs display results for 5 muscles around the knee (RF = Rectus Femoris, VM = Vastus Medialis, VL = Vastus Lateralis, BF = Biceps Femoris, ST = Semitendinosus). Note the scale differences between knee extensors and flexors. Error bars indicate one standard deviation over 10 subjects (for clarity of the graph error bars are not shown for the 50 Nm moment). a.u. = arbitrary units.
(Fig. 4) on the activation of any of the knee extensors (p> 0:172) or flexors (p> 0:667). In addition, external force position did not significantly interact with moment level for any of the knee extensors (p> 0:298) or flexors (p> 0:199). The interaction between external force position and target angle was not significant for
all muscles (p> 0:231) except for the ST (p¼ 0:024). However, in separate ANOVA’s per joint angle, the effect of external force position did not reach significance in the ST (p> 0:079) in any of the joint angles. The success in maintaining a constant target knee angle, as quantified by the standard deviation of the
312
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
Fig. 4. Average EMG against flexion angle during isometric knee extension efforts in a joint angle positioning task. In each graph, one line represents the flexion efforts (60 Nm external load applied at a distance of 307 mm from the knee joint). The other three lines are extension efforts, averaged over three moment conditions (20, 50 and 80 Nm). Those three lines represent the three distances (95, 212 and 307 mm) relative to the knee joint at which the external force was applied. The five graphs are organized in the same way as in Fig. 2. For clarity of the graph error bars are not shown for the 212 mm external force position. a.u. = arbitrary units.
joint angle during the positioning task recorded over 3 s, was significantly affected by joint angle (p< 0:038) and external force position (p < 0:043). With respect to the target angle, this standard deviation showed a v v gradual decrease from 0.30 at a target angle of 5 of v v knee flexion to 0.18 at a target angle of 20 of knee v flexion. Beyond 20 of knee flexion the standard deviation of the knee angle during the positioning task
remained rather stable. With respect to external force position, an average standard deviation of 0.22, 0.21 v and 0.18 was found for the short, middle and long distance, respectively. Moment level did not significantly affect the knee angle standard deviation (p¼ 0:832), nor were there any interaction effects on the knee angle standard deviation. Like the joint angle, the EMG of the quadriceps also showed more fluctuation with small
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
flexion angles. Those fluctuations were not accompanied by concomittant fluctuations of the hamstrings, but rather by a constant increase of hamstrings EMG (Fig. 5).
313
4. Discussion 4.1. Co-contraction and extension moment level Quadriceps activation leads to a forward shear force on the tibia. If the hamstrings are playing a role in counteracting those shear forces in healthy subjects, increased hamstrings activation is expected when the extension effort increases without a change in joint angle. The results do indeed show an increase of the hamstrings EMG with an increase in extension moment. However, the increase in activation of the muscles was less than proportional. This is in agreement with the results found by Grabiner et al. [6] for a v flexion angle of 10 to 20 . The current study extends v this finding to a range of knee joint angles of 5 to 50 . Together, these results show that hamstrings cocontraction does not increase proportionally with forward shear forces induced by the quadriceps. This argues against a potential role of the hamstrings in the protection of the ACL during isometric extension efforts in healthy subjects within the force ranges used in this study. However, it should be noted that this argument depends on linearity of the relation between muscle force and EMG at a constant joint angle. 4.2. Co-contraction and joint angle
Fig. 5. Typical example of rectified averaged EMG (using a low pass 2nd order butterworth filter with 3 Hz cut-off frequency) at 3 different target flexion angles. This example is subject 2, performing an extension trial with the closest external force position and the highest moment level. With small flexion angles, the level as well as the variability of the quadriceps EMG (highest 3 lines in each graph) increases. The EMG of the hamstring muscles (lower 2 lines) also increases, without showing a clear correlations with the fluctuation in quadriceps EMG (upper 3 lines).
With smaller knee flexion angles, the shear force component of the hamstrings decreases with the flexion angle. The same effort of the hamstrings will thus result in smaller backward shear forces in smaller knee flexion angles. Within the range of 20–50 degrees of flexion, the increase in co-contraction as found in this study, could therefore be interpreted as supporting the idea of the hamstrings protecting the ACL. However, hamstrings co-contraction appears to be effective in reducing forward tibial shear forces only v v when the knee is flexed more than 15–22 . In the 0–15 flexion range, hamstrings co-contraction is likely to be counter-productive with regard to forward shear force reduction. The reason is that, in order to neutralize the flexion moment produced by the hamstrings, the quadriceps activation also has to be increased. Modelling work suggests that, close to full extension, the resulting increase in forward shear force by increasing the quadriceps activation outweighs the backward shear force that is produced by the hamstrings [8,13]. In vivo measurement of ACL strain supports this idea by v showing that, at 15 of knee flexion, simultaneous activation of the hamstrings and quadriceps increases rather than decreases ACL strain [3]. So, hamstrings co-contraction can only compensate the forward directed shear produced by the quadriceps at larger flexion angles. The strong increase in co-
314
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
contraction near full extension is therefore inconsistent with the idea of the hamstrings protecting the ACL. The results of the current study show a strong increase in EMG of almost all muscles near full extension. The EMG increase of the quadriceps (VL þ RF þ VM) near full extension can in part be explained by the decreasing muscle length, which requires increased activation to maintain a constant force level [16]. In addition, the mass of the lower leg and the arm of the loading device cause slightly higher moments in the extended knee. Increases in hamstrings (BF þ ST) EMG activity, with decreasing flexion angle cannot be explained by an effect of muscle length, since flexion trials show a decreasing hamstrings activity with decreasing flexion angle. The results of this study suggest that near full extension the activation of the BF at on average 96 Nm total extension moment rises up to a level that is comparable to 44 Nm total flexion moment. As is evident from Fig. 3, the ratio of hamstrings activation in extension efforts to flexion efforts quickly drops for larger flexion angles. Although the high level of co-contraction near full extension may enhance knee stability, it is, as stated earlier, ineffective in terms of reducing forward shear forces at the tibia. Therefore, the increase in hamstrings EMG near full extension also argues against a specific ACL-protective role during sub-maximal isometric contractions in healthy subjects. This raises the question what the function might be of the increase in hamstrings co-contraction near full extension. In part, this increased cocontraction may be related to precision demands in the positioning task in the current study. The subjects reported that holding the knee at the target angle was more difficult near full extension. This could be related to a reduced force-producing capacity of the quadriceps near full extension. In addition, the actual moment also increases because the weight of the leg causes a larger moment near full extension, but this moment increase was small relative to the total moment, especially in the condition with an (external) moment level of 80 Nm. An increase in muscular effort leads to an increase in the difficulty of a positioning task [17] and this may therefore explain the increase in co-contraction near full extension. Despite the increase in co-contraction, the current data revealed an increased standard deviation of the knee angle near full extension, showing that subjects were less successful in holding the knee joint angle near full extension compared to a more flexed position. However, precision demands may not be fully responsible for the increase in hamstrings activation v near full extension. Aagaard et al. [1] used slow (30 /s) isokinetic maximal contractions, which do not depend on precision, and found an increase in hamstrings EMG to extension moment ratio near full extension.
4.3. Co-contraction and external force position Within each moment level, actual moments (and thus quadriceps activation) varied over joint angles and over subjects due to the moment caused by the lower leg. However, within subjects and joint angles, actual moments as well as applied moments were constant over the different external force positions. Zavatsky et al. [20] calculated the effect of the position of the external force on the ACL strain. Ignoring the leg mass, they predicted that at a distance of 400 mm the ACL strain would be five times the external force, but at 100 mm distance the external force would be large enough to fully cancel out the forward shear force by the quadriceps, and consequently the ACL would not be loaded. Thus, a protective hamstrings cocontraction will only be useful when the external force position is at a distance of more than 100 mm from the knee joint. However, no main effect of external force position was observed for any of the flexors or extensors. An interaction of external force position with target angle was found for the ST, but Fig. 4 shows that this effect was very small and therefore does not provide a convincing argument in favour of ACL protection by the hamstrings in healthy subjects during submaximal isometric contractions. Finally, it is noteworthy that Butler et al. [4] estimated the failure load in the ACL at 173066 N in young donors. This is much larger than the estimated force on the ACL produced by the quadriceps muscle during maximal isometric extension efforts alone, which is about 520 N [14]. Moreover, the contractions in our study were at a submaximal level. Therefore, the ACL should be able to counteract shear forces in the isometric contractions used in this study and there may thus have been no need for co-contraction from the perspective of ACL protection. This may not be the case in highly dynamic and powerful activities, for instance during elite sports performance. The moments in the current study were, however, at least as high as the moments that are reached in gait. In conclusion, within the range of quasi-static loading conditions tested in this study, changes in cocontraction with joint angle, moment level and position of the external force in healthy subjects did not appear to be primarily directed at reducing forward tibial shear forces. Therefore, the results of this study do not support the hypothesis that the central nervous system generates recruitment patterns that aim to protect the ACL by actively counteracting shear forces in the knee during submaximal isometric contractions in healthy subjects. However, this does not rule out the possibility that such a mechanism could be operative when shear forces are of larger magnitude or when the loading of the ACL is highly dynamic. Previous work showed a reflex arc from the ACL to the hamstrings to be operative [9,18]. In a goat
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
model, Fujita et al. found this reflex arc to be operative under dynamic subfailure and failure loading, but not under static subfailure loading [5]. References [1] P. Aagaard, E.B. Simonsen, J.L. Andersen, S.P. Magnusson, F. Bojsen-Moller, P. Dyhre-Poulsen, Antagonist muscle coactivation during isokinetic knee extension, Scandinavian Journal of Medicine and Science in Sports 10(2) (2000) 58–67. [2] T. Alkjaer, E.B. Simonsen, S.P. Peter Magnusson, H. Aagaard, P. Dyhre-Poulsen, Differences in the movement pattern of a forward lunge in two types of anterior cruciate ligament deficient patients: copers and non-copers, Clinical Biomechanics 17(8) (2002) 586–593. [3] B.D. Beynnon, B.C. Fleming, R.J. Johnson, C.E. Nichols, P.A. Pope, M.H. Pope, Anterior cruciate ligament strain behavior during rehabilitation exercises in vivo, American Journal of Sports Medicine 23(1) (1995) 24–34. [4] D.L. Butler, E.S. Grood, F.R. Noyes, A.N. Sodd, On the interpretation of our anterior cruciate ligament data, Clinical Orthopaedics 196 (1985) 26–34. [5] I. Fujita, T. Nishikawa, H.E. Kambic, J.T. Andrish, M.D. Grabiner, Characterization of hamstring reflexes during anterior cruciate ligament disruption: in vivo results from a goat model, Journal of Orthopaedic Research 18(2) (2000) 183–189. [6] M.D. Grabiner, K.R. Campbell, D.L. Hawthorne, D.A. Hawkins, Electromyographic study of the anterior cruciate ligamenthamstrings synergy during isometric knee extension, Journal of Orthopaedic Research 7(1) (1989) 152–155. [7] H.J. Hermens, B.F. Freriks, R. Merletti, D. Stegeman, J. Blok, G. Rau, C. Disselhorst-Klug, G. Ha¨gg, Seniam-European Recommendations for Surface Electromyography, Roessingh Research and Development, Enschede, 1999. [8] A. Imran, J.J. O’Connor, Control of knee stability after ACL injury and repair: interaction between hamstrings contraction and tibial translation, Clinical Biomechanics 13(3) (1998) 153–162. [9] H. Johansson, P. Sjolander, P. Sojka, A sensory role for the cruciate ligaments, Clinical Orthopaedics & Related Research 268 (1991) 161–178. [10] E. Kellis, V. Baltzopoulos, The effects of the antagonist muscle force on intersegmental loading during isokinetic efforts of the knee extensors, Journal of Biomechanics 32(1) (1999) 19–25. [11] M.R. Krogsgaard, P. Dyhre-Poulsen, T. Fischer-Rasmussen, Cruciate ligament reflexes, Journal of Electromyography and Kinesiology 12(3) (2002) 177–182. [12] W. Liu, M.E. Maitland, The effect of hamstring muscle compensation for anterior laxity in the ACL-deficient knee during gait, Journal of Biomechanics 33(7) (2000) 871–879. [13] J.J. O’Connor, Can muscle co-contraction protect knee ligaments after injury or repair? Journal of Bone and Joint Surgery Br 75(1) (1993) 41–48. [14] M.G. Pandy, K.B. Shelburne, Dependence of cruciate-ligament loading on muscle forces and external load, Journal of Biomechanics 30(10) (1997) 1015–1024. [15] S. Plagenhoef, F.G. Evans, T. Abdelnour, Anatomical data for analyzing human motion, Research Quarterly for Exercise and Sport 54(2) (1983) 169–178. [16] K.B. Shelburne, M.G. Pandy, A musculoskeletal model of the knee for evaluating ligament forces during isometric contractions, Journal of Biomechanics 30(2) (1997) 163–176. [17] A.B. Slifkin, K.M. Newell, Variability and noise in continuous force production, Journal of Motor Behaviour 32(2) (2000) 141–150.
315
[18] M. Solomonow, R. Baratta, B.H. Zhou, H. Shoji, W. Bose, C. Beck, R. D’Ambrosia, The synergistic action of the anterior cruciate ligament and thigh muscles in maintaining joint stability, American Journal of Sports Medicine 15(3) (1987) 207–213. [19] M.E. van Lent, M.R. Drost, F.A. v.d. Wildenberg, EMG profiles of ACL-deficient patients during walking: the influence of mild fatigue, International Journal of Sports Medicine 15(8) (1994) 508–514. [20] A.B. Zavatsky, D.J. Beard, J.J. O’Connor, Cruciate ligament loading during isometric muscle contractions. A theoretical basis for rehabilitation, American Journal of Sports Medicine 22(3) (1994) 418–423.
Idsart Kingma is an assistant professor at the Faculty of Human Movement Sciences of the ‘Vrije Universiteit Amsterdam’ where he is teaching courses on ergonomics and biomechancis. He has an MSc in human movement sciences and he finished his PhD on the biomechanics of lifting in 1999. Currently, his main research interests are mechanical aspects of low back loading and neuromuscular control of joint stability. Idsart Kingma was the first author of 17 and a coauthor of 23 papers in international scientific journals. Sietske Aalbersberg received her M.Sc. in Human Movement Sciences and is a medical doctor since 2002. Currently, she is a PhD student in Human Movement Sciences at the Vrije Universiteit in Amsterdam, The Netherlands. The focus of her research project is the neuromuscular control of the human knee with a specific interest in control strategies after ACL rupture.
Jaap van Diee¨n worked as a researcher at the Institute for Agricultural Engineering (IMAG-DLO) in Wageningen, the Netherlands from 1986 to 1996. He obtained a PhD in Human Movement Sciences from the Faculty of Human Movement Sciences at the ‘Vrije Universiteit Amsterdam’ the Netherlands in 1993. Since 1996 he has been affiliated to this faculty, since 2002 as professor of biomechanics. He is chairing the ergonomics program at this faculty. In addition, he is the head of a research group focusing on mechanical and neural aspects of musculoskeletal injuries. His main research interest is on control of muscles of the trunk and upper extremity, where especially the interaction of muscle coordination, fatigue, joint load and stability is an important research topic. Jaap van Diee¨n was the first author of over 40 and a co-author of a similar number of papers in international scientific journals. In addition he has (co-) authored numerous abstracts and book chapters in the international literature and technical reports and publications in Dutch. He serves on the editorial boards of the Journal of Electromyography and Kinesiology and Human Movement Sciences and is a regular reviewer for several other journals and funding agencies.
Are hamstrings activated to counteract shear forces during isometric knee extension efforts in healthy subjects? Idsart Kingma , Sietske Aalbersberg, Jaap H. van Diee¨n Institute for Fundamental and Clinical Human Movement Sciences, Faculty of Human Movement Sciences, Vrije Universiteit, Van der Boechorststraat 9, 1081 BT Amsterdam, The Netherlands Received 2 July 2003; received in revised form 15 December 2003; accepted 6 January 2004
Abstract The hamstring muscles have the potential to counteract anterior shear forces at the knee joint by co-contracting during knee extension efforts. Such a muscle recruitment pattern might protect the anterior cruciate ligament (ACL) by reducing its strain. In this study we investigated to what extent co-activation of the knee flexors during extension efforts is compatible with the hypothesis that this co-activation serves to counteract anterior tibial shear forces during isometric knee extension efforts in healthy subjects. To this aim, it is investigated whether co-activation varies with the required knee extension moment, with the knee joint angle, and with the position of the external flexing force relative to the knee joint. With unaltered moment and muscle activation, distal positioning of the flexing force on the tibia causes higher resultant (muscular plus external) forward shear forces at the knee as compared to proximal positioning. In ten subjects, knee flexor and extensor EMG was measured during a quasi-isometric positioning task for a range (5–50 degrees) of knee flexion angles. It was found that the co-activation of the knee flexors increased with the extension moment, but this increase was less than proportional (p< 0:001). The extension moment increased 2.7 to 3.4 times, whereas the activation of Biceps Femoris and Semitendinosus increased only a factor 1.3 to 2.0 (joint angle dependent). Furthermore, a strong increase in co-activation was seen near full extension of the knee joint. The position of the external extension load on the tibia did not affect the level of co-contraction. It is argued that these results do not suggest a recruitment pattern that is directed at reduction of anterior shear forces in the knee joint during sub-maximal isometric knee extension efforts in healthy subjects. # 2004 Elsevier Ltd. All rights reserved. Keywords: EMG; Knee; ACL; Hamstrings; Shear forces
1. Introduction During knee extension efforts, quadriceps activation causes the tibia to be pulled forward relative to the femur. Even in full extension, the patellar tendon has a considerable angle with the long axis of the tibia, causing knee extension efforts to be accompanied by a substantial shear force on the tibia [10]. In healthy subjects, the anterior cruciate ligament is assumed to play an important role in counteracting forward shear forces on the tibia. Full activation of the quadriceps is estimated to cause a force in the anterior cruciate ligament (ACL) of over 500 N [14]. This force depends on Corresponding author. Tel.: +31-20-4448492; fax: +31-204448529. E-mail address: [email protected] (I. Kingma).
1050-6411/$ - see front matter # 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.jelekin.2004.01.003
v
the knee angle with a maximum being reached at 15 of knee flexion [14]. Co-contraction of the hamstring muscles can potentially assist in counteracting forward tibial shear because the hamstrings’ force has a backward shear component. However, a disadvantage of this strategy is that, in order to maintain the same extension moment, the quadriceps force also has to be increased to counteract the flexion moment of the hamstrings. In a simulation study, Liu and Maitland [12] estimated that cancelling out forward tibial shear by hamstrings co-contraction would result in about a doubling of the required quadriceps force in gait. Due to the relatively small knee extension moments required in gait, protective co-contraction may not be needed in healthy subjects. However, when large extension moments are required, for instance up to 100 Nm in the forward lunge movement, substantial co-contrac-
308
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
tion is found [2]. In addition, mechanical [18] as well as electrical [11] stimulation of the ACL has been shown to elicit hamstrings activation. Apart from fatigue [19], three main factors might affect the amount of hamstrings co-contraction. These factors are (1) the required extension moment, (2) the joint angle and (3) the position of the external flexing force with respect to the knee joint. With respect to the first factor, increasing quadriceps activation leads to an increase of the forward shear force on the tibia. If the hamstrings play a role in counteracting those shear forces in healthy subjects, increased hamstrings activation is expected when the extension effort increases. However, Grabiner et al. [6] reported an unchanged hamstrings activation with extension efforts ranging from 10 to 90% maximum voluntary contraction. Possibly, the absence of increased hamstrings co-activation in the study of Grabiner et al. [6] is related to the second factor, the influence of the knee joint angle. The v knee flexion angle in their experiment was 10–20 . Hamstrings co-contraction appears to be effective in reducing forward tibial shear forces only when the knee v v is flexed more than 15 [3] to 22 [13]. With respect to the third factor, the position of the external flexing force, Zavatsky et al. [20] estimated that, with a flexing force applied at 40 cm distance to the knee joint, the extension effort by the quadriceps would result in an ACL force of 5 times the external flexing force. With the same moment applied by an external force at a distance of 10 cm, the ACL would not be loaded because the external force is large enough to fully cancel out the forward shear force of the quadriceps [20]. A protective hamstrings co-contraction would therefore be useful when the external flexing force is applied at a distance of 40 cm, but it would be superfluous when the external flexing force is applied at a distance of 10 cm from the knee joint. The aim of this study was to find out to what extent co-activation of the knee flexors during extension efforts is compatible with the hypothesis that this coactivation serves to counteract anterior tibial shear forces during sub-maximal isometric knee extension efforts in healthy subjects. To this aim, it is investigated whether hamstrings co-activation, during a quasi-isometric positioning task, varies with (1) the required knee extension moment, (2) the joint angle and (3) the position of an external flexing force relative to the knee joint. In addition, the actual knee angle during the positioning task is measured to evaluate the difficulty of maintaining a constant target angle.
2. Methods 2.1. Subjects and procedure Ten healthy young subjects (7 males and 3 females, age 22 2 yrs, bodyweight 71:8 8:9 kg, shank length 393:6 52:5 mm) participated in the experiment after providing informed consent. In a seated position with v the hip joint in approximately 70 flexion, the knee joint of the right leg of the subjects was aligned with the axis of the loading apparatus. (External) flexing moments of 20, 50 and 80 Nm were attained using an external force that was applied perpendicular to the long axis of the shank. Since the moment is defined as force times moment arm, the ratio between the (external) posterior force and the flexion moment depends on where the force is applied. For all three moment levels, we applied the external force at three distances from the knee joint axis (95, 212 and 307 mm). Therefore, for each moment level, the three external force positions resulted in three levels of external backward shear force (Fig. 1). Note that moment levels were constant within subjects over external force positions. However, moments were not completely constant between subjects and between knee flexion angles due to the moment caused by the lower leg mass. With the external force at 95 mm (and ignoring the lower leg mass), the backward shear force caused by the external force is assumed to have about the same magnitude as compared to the forward shear force due to quadriceps contraction, so that the extension effort would not cause ACL loading [20]. At 307 mm, the backward shear force due to the external force is considerably smaller than the forward shear force due to the activation of the quadriceps, resulting in an increased forward shear force challenge to the knee joint (and thus resulting in an increased loading of the ACL) compared to the external force position close to the joint. The order of the moments and distances was randomised over subjects. From markers attached to the leg, subjects received real-time feedback regarding their knee joint angle on a computer screen in front of them. For each measurement, subjects started with their knee at approximately v 90 flexion. The subjects were requested to extend their knee to a target angle (indicated by a line on the computer screen). The external force was applied as soon v as the knee joint reached an angle in approximately 70 flexion. After holding the knee at the target angle for 5 s, the subjects were allowed to flex the knee and relax for at least 15 s before the next target angle was presented. Only the isometric part of each trial was used for further analysis. Target angles ranged from 5 to v v 50 , in steps of 5 , so that a total of 90 extension efforts were performed (10 angles 3 moment levels 3 force positions). In addition, flexion efforts were per-
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
309
Electromyographic data (EMG) from selected leg muscles were obtained, using disposable surface EMGelectrodes (Ag/AgCl; square 5 5 mm pick-up area) that had been attached at a center-to-center electrode distance of 2.5 cm after cleaning and abrasion of the skin. Surface EMG was measured on the Rectus Femoris (RF), Vastus Medialis (VM), Vastus Lateralis (VL), Biceps Femoris (BF) and Semitendinosus (ST). Electrode locations were based on Seniam guidelines [7]. During the holding task, EMG signals were sampled at 1000 Hz (Porti 17, TMS, Enschede, The Netherlands; 22 bits AD conversion after 20x amplification, input impedance > 1012X, CMMR >90 dB for the relevant range of frequencies), after band-pass filtering with 10 and 350 Hz cut-off frequencies. At the same time, marker positions were sampled at 100 Hz with the Optotrak system. EMG and marker position data were synchronised in time using a pulse generated by the Optotrak system. 2.3. Data processing
Fig. 1. A graphical representation of the experimental setup. A constant external moment was applied at the knee through the use of a wheel. Subjects were requested to hold the knee at a specific angle while they received real time feedback regarding the knee angle. The three dotted arrows indicate the locations where the flexion force was applied during the extension trials. Those locations (95, 212 and 307 mm distance from the knee joint) were obtained by displacing the attachment to the knee in proximal or distal direction. By increasing the distance from the knee joint, the magnitude of this force was reduced without affecting the applied flexion moment.
formed at the same 10 target angles, but with only one moment level (60 Nm) and with only one force position (307 mm). Subjects were allowed to rest as long as they needed when they felt fatigued. 2.2. Measurements An opto-electronic movement registration system (Optotrak, Northern Digital Inc., Canada) was used to provide a target angle and real-time feedback of the knee joint angle to the subjects. LED markers were attached to the medial side of the leg at the medial malleolus, the medial tibial condyle, the medial condyle of the femur, and proximally at the medial side of the upper leg. Markers were aligned in such a way that the real-time feedback showed an angle of zero degrees between the two markers on the upper leg and the two markers on the lower leg when the subject placed the leg on a flat table and relaxed the muscles as much as possible.
EMG signals were rectified and averaged over 3 s during the isometric part of each trial (starting 1 s after the target angle was reached). As noted before, the net moment at the knee joint was not constant over target angles, due to the weight of the lower leg, the foot and the arm of the loading device, which caused an additional moment that varied with the knee joint angle. This additional moment was calculated for each positioning task using the recorded marker positions, the mass and centre of mass location of the foot and lower leg according to Plagenhoef [15], and the mass and centre of mass location of the loading device. From those data, the additional net moment due to the foot, lower leg and loading device was calculated to v range from 5:1 0:6 Nm at 50 of knee flexion to v 16:1 3:1 Nm at 5 of knee flexion. The success in maintaining a constant target angle over the 3 s period was quantified using the skin marker positions to calculate the standard deviation of the knee angle in that period. 2.4. Statistics For all muscles, repeated measures ANOVA’s were applied to the rectified and averaged EMG data of the extension efforts. The independent variables were joint angle (10 levels), moment level (3 levels), and external force position (3 levels). The same ANOVA’s were applied to the ratio of BF and ST activation to extension moment level and to the standard deviation of the joint angle during the recorded period. In cases where significant interactions with joint angles were found, separate ANOVA’s with moment level and external
310
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
force position as independent variables were applied for each joint angle. 3. Results An example of the EMG (rectified and averaged over 3 s) of two selected muscles in one external load position shows that there was, as can be expected, substantial variation among subjects in the level of the raw EMG (Fig. 2). However, the pattern of change with joint angle was comparable over subjects. During extension efforts all muscles showed a significant (p< 0:001) and evidently non-linear change in activ vation with joint angle (Fig. 3). Especially below 20 of flexion, there was a strong increase in co-contraction of the hamstring muscles with the quadriceps muscles with decreasing knee flexion in all muscles. This increase in co-contraction could only be explained to a minor extent by an increase in actual net moment due to the mass of the lower leg, foot, and arm of the loading device. For instance, for the flexors (and even more v for the extensors) the activity at a target angle of 5 was roughly twice as high as compared to a target v angle of 20 . In contrast, the increase in actual moment amounted only on average 3.2 Nm (3.4% of the total moment at the highest moment level up to 9.7% of the total moment at the lowest moment level). The high level of co-contraction near full knee extension v becomes apparent when, at a target angle of 5 , the activation of the hamstring muscles is compared between flexion and extension efforts. At the strongest extension effort of 96:1 3:1 Nm (80 Nm moment due
to the external force + 16:1 3:1 Nm additional moment due to lower leg, foot and loading device) the BF activation was about the same as the activation at a flexion effort of 43:9 3:1 Nm (60 Nm moment due to the external force + 16:1 3:1 Nm additional moment due to lower leg, foot and loading device). Not surprisingly, the activation of all muscles increased with the applied moment level (p< 0:002). However, hamstring activation did not increase proportionally with the moment level, since the ratio of hamstring activation to (actual) extension moment was significantly affected by moment level (p< 0:001 for both the BF and ST). In fact, the hamstrings coactivation increased less than proportional with the moment level, since the highest level of extension moment was (dependent on the joint angle) 2.7–3.4 times the lowest level of extension moment, whereas the BF and ST activation at the highest extension moment level ranged (dependent on the joint angle) from 1.3 to 2.0 times the activation at the lowest extension moment level. A significant moment level with target angle interaction was found for all muscles (p< 0:002). Fig. 3 shows that the effect of moment level was larger at smaller knee flexion angles. Separate ANOVA’s per joint angle showed that the effect of moment level was significant for all three quadriceps muscles in all joint angles (p< 0:05). For the BF the effect of moment level was v significant at all joint angles except at 50 of flexion, and for the ST the effect of moment level was only sigv nificant for knee flexion angles less than 30 . There was no significant main effect of external force position
Fig. 2. Average EMG against flexion angle during isometric knee extension and flexion efforts in a joint angle positioning task. This figure shows the variation in EMG level over subjects. EMG data (VM and BF) for one external load position (307 mm), averaged over 3 moment levels, is shown for individual subjects. Each line represents the EMG of one subject. a.u. = arbitrary units.
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
311
Fig. 3. Average EMG against flexion angle during isometric knee extension and flexion efforts in a joint angle positioning task. In each graph, one line represents the flexion efforts (60 Nm moment applied by an external force at a distance of 307 mm from the knee joint). The other three lines are extension efforts, averaged over three distances (95, 212 and 307 mm) relative to the knee joint at which the external force was applied. Those three lines represent the three moment levels (20, 50 and 80 Nm). The five graphs display results for 5 muscles around the knee (RF = Rectus Femoris, VM = Vastus Medialis, VL = Vastus Lateralis, BF = Biceps Femoris, ST = Semitendinosus). Note the scale differences between knee extensors and flexors. Error bars indicate one standard deviation over 10 subjects (for clarity of the graph error bars are not shown for the 50 Nm moment). a.u. = arbitrary units.
(Fig. 4) on the activation of any of the knee extensors (p> 0:172) or flexors (p> 0:667). In addition, external force position did not significantly interact with moment level for any of the knee extensors (p> 0:298) or flexors (p> 0:199). The interaction between external force position and target angle was not significant for
all muscles (p> 0:231) except for the ST (p¼ 0:024). However, in separate ANOVA’s per joint angle, the effect of external force position did not reach significance in the ST (p> 0:079) in any of the joint angles. The success in maintaining a constant target knee angle, as quantified by the standard deviation of the
312
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
Fig. 4. Average EMG against flexion angle during isometric knee extension efforts in a joint angle positioning task. In each graph, one line represents the flexion efforts (60 Nm external load applied at a distance of 307 mm from the knee joint). The other three lines are extension efforts, averaged over three moment conditions (20, 50 and 80 Nm). Those three lines represent the three distances (95, 212 and 307 mm) relative to the knee joint at which the external force was applied. The five graphs are organized in the same way as in Fig. 2. For clarity of the graph error bars are not shown for the 212 mm external force position. a.u. = arbitrary units.
joint angle during the positioning task recorded over 3 s, was significantly affected by joint angle (p< 0:038) and external force position (p < 0:043). With respect to the target angle, this standard deviation showed a v v gradual decrease from 0.30 at a target angle of 5 of v v knee flexion to 0.18 at a target angle of 20 of knee v flexion. Beyond 20 of knee flexion the standard deviation of the knee angle during the positioning task
remained rather stable. With respect to external force position, an average standard deviation of 0.22, 0.21 v and 0.18 was found for the short, middle and long distance, respectively. Moment level did not significantly affect the knee angle standard deviation (p¼ 0:832), nor were there any interaction effects on the knee angle standard deviation. Like the joint angle, the EMG of the quadriceps also showed more fluctuation with small
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
flexion angles. Those fluctuations were not accompanied by concomittant fluctuations of the hamstrings, but rather by a constant increase of hamstrings EMG (Fig. 5).
313
4. Discussion 4.1. Co-contraction and extension moment level Quadriceps activation leads to a forward shear force on the tibia. If the hamstrings are playing a role in counteracting those shear forces in healthy subjects, increased hamstrings activation is expected when the extension effort increases without a change in joint angle. The results do indeed show an increase of the hamstrings EMG with an increase in extension moment. However, the increase in activation of the muscles was less than proportional. This is in agreement with the results found by Grabiner et al. [6] for a v flexion angle of 10 to 20 . The current study extends v this finding to a range of knee joint angles of 5 to 50 . Together, these results show that hamstrings cocontraction does not increase proportionally with forward shear forces induced by the quadriceps. This argues against a potential role of the hamstrings in the protection of the ACL during isometric extension efforts in healthy subjects within the force ranges used in this study. However, it should be noted that this argument depends on linearity of the relation between muscle force and EMG at a constant joint angle. 4.2. Co-contraction and joint angle
Fig. 5. Typical example of rectified averaged EMG (using a low pass 2nd order butterworth filter with 3 Hz cut-off frequency) at 3 different target flexion angles. This example is subject 2, performing an extension trial with the closest external force position and the highest moment level. With small flexion angles, the level as well as the variability of the quadriceps EMG (highest 3 lines in each graph) increases. The EMG of the hamstring muscles (lower 2 lines) also increases, without showing a clear correlations with the fluctuation in quadriceps EMG (upper 3 lines).
With smaller knee flexion angles, the shear force component of the hamstrings decreases with the flexion angle. The same effort of the hamstrings will thus result in smaller backward shear forces in smaller knee flexion angles. Within the range of 20–50 degrees of flexion, the increase in co-contraction as found in this study, could therefore be interpreted as supporting the idea of the hamstrings protecting the ACL. However, hamstrings co-contraction appears to be effective in reducing forward tibial shear forces only v v when the knee is flexed more than 15–22 . In the 0–15 flexion range, hamstrings co-contraction is likely to be counter-productive with regard to forward shear force reduction. The reason is that, in order to neutralize the flexion moment produced by the hamstrings, the quadriceps activation also has to be increased. Modelling work suggests that, close to full extension, the resulting increase in forward shear force by increasing the quadriceps activation outweighs the backward shear force that is produced by the hamstrings [8,13]. In vivo measurement of ACL strain supports this idea by v showing that, at 15 of knee flexion, simultaneous activation of the hamstrings and quadriceps increases rather than decreases ACL strain [3]. So, hamstrings co-contraction can only compensate the forward directed shear produced by the quadriceps at larger flexion angles. The strong increase in co-
314
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
contraction near full extension is therefore inconsistent with the idea of the hamstrings protecting the ACL. The results of the current study show a strong increase in EMG of almost all muscles near full extension. The EMG increase of the quadriceps (VL þ RF þ VM) near full extension can in part be explained by the decreasing muscle length, which requires increased activation to maintain a constant force level [16]. In addition, the mass of the lower leg and the arm of the loading device cause slightly higher moments in the extended knee. Increases in hamstrings (BF þ ST) EMG activity, with decreasing flexion angle cannot be explained by an effect of muscle length, since flexion trials show a decreasing hamstrings activity with decreasing flexion angle. The results of this study suggest that near full extension the activation of the BF at on average 96 Nm total extension moment rises up to a level that is comparable to 44 Nm total flexion moment. As is evident from Fig. 3, the ratio of hamstrings activation in extension efforts to flexion efforts quickly drops for larger flexion angles. Although the high level of co-contraction near full extension may enhance knee stability, it is, as stated earlier, ineffective in terms of reducing forward shear forces at the tibia. Therefore, the increase in hamstrings EMG near full extension also argues against a specific ACL-protective role during sub-maximal isometric contractions in healthy subjects. This raises the question what the function might be of the increase in hamstrings co-contraction near full extension. In part, this increased cocontraction may be related to precision demands in the positioning task in the current study. The subjects reported that holding the knee at the target angle was more difficult near full extension. This could be related to a reduced force-producing capacity of the quadriceps near full extension. In addition, the actual moment also increases because the weight of the leg causes a larger moment near full extension, but this moment increase was small relative to the total moment, especially in the condition with an (external) moment level of 80 Nm. An increase in muscular effort leads to an increase in the difficulty of a positioning task [17] and this may therefore explain the increase in co-contraction near full extension. Despite the increase in co-contraction, the current data revealed an increased standard deviation of the knee angle near full extension, showing that subjects were less successful in holding the knee joint angle near full extension compared to a more flexed position. However, precision demands may not be fully responsible for the increase in hamstrings activation v near full extension. Aagaard et al. [1] used slow (30 /s) isokinetic maximal contractions, which do not depend on precision, and found an increase in hamstrings EMG to extension moment ratio near full extension.
4.3. Co-contraction and external force position Within each moment level, actual moments (and thus quadriceps activation) varied over joint angles and over subjects due to the moment caused by the lower leg. However, within subjects and joint angles, actual moments as well as applied moments were constant over the different external force positions. Zavatsky et al. [20] calculated the effect of the position of the external force on the ACL strain. Ignoring the leg mass, they predicted that at a distance of 400 mm the ACL strain would be five times the external force, but at 100 mm distance the external force would be large enough to fully cancel out the forward shear force by the quadriceps, and consequently the ACL would not be loaded. Thus, a protective hamstrings cocontraction will only be useful when the external force position is at a distance of more than 100 mm from the knee joint. However, no main effect of external force position was observed for any of the flexors or extensors. An interaction of external force position with target angle was found for the ST, but Fig. 4 shows that this effect was very small and therefore does not provide a convincing argument in favour of ACL protection by the hamstrings in healthy subjects during submaximal isometric contractions. Finally, it is noteworthy that Butler et al. [4] estimated the failure load in the ACL at 173066 N in young donors. This is much larger than the estimated force on the ACL produced by the quadriceps muscle during maximal isometric extension efforts alone, which is about 520 N [14]. Moreover, the contractions in our study were at a submaximal level. Therefore, the ACL should be able to counteract shear forces in the isometric contractions used in this study and there may thus have been no need for co-contraction from the perspective of ACL protection. This may not be the case in highly dynamic and powerful activities, for instance during elite sports performance. The moments in the current study were, however, at least as high as the moments that are reached in gait. In conclusion, within the range of quasi-static loading conditions tested in this study, changes in cocontraction with joint angle, moment level and position of the external force in healthy subjects did not appear to be primarily directed at reducing forward tibial shear forces. Therefore, the results of this study do not support the hypothesis that the central nervous system generates recruitment patterns that aim to protect the ACL by actively counteracting shear forces in the knee during submaximal isometric contractions in healthy subjects. However, this does not rule out the possibility that such a mechanism could be operative when shear forces are of larger magnitude or when the loading of the ACL is highly dynamic. Previous work showed a reflex arc from the ACL to the hamstrings to be operative [9,18]. In a goat
I. Kingma et al. / Journal of Electromyography and Kinesiology 14 (2004) 307–315
model, Fujita et al. found this reflex arc to be operative under dynamic subfailure and failure loading, but not under static subfailure loading [5]. References [1] P. Aagaard, E.B. Simonsen, J.L. Andersen, S.P. Magnusson, F. Bojsen-Moller, P. Dyhre-Poulsen, Antagonist muscle coactivation during isokinetic knee extension, Scandinavian Journal of Medicine and Science in Sports 10(2) (2000) 58–67. [2] T. Alkjaer, E.B. Simonsen, S.P. Peter Magnusson, H. Aagaard, P. Dyhre-Poulsen, Differences in the movement pattern of a forward lunge in two types of anterior cruciate ligament deficient patients: copers and non-copers, Clinical Biomechanics 17(8) (2002) 586–593. [3] B.D. Beynnon, B.C. Fleming, R.J. Johnson, C.E. Nichols, P.A. Pope, M.H. Pope, Anterior cruciate ligament strain behavior during rehabilitation exercises in vivo, American Journal of Sports Medicine 23(1) (1995) 24–34. [4] D.L. Butler, E.S. Grood, F.R. Noyes, A.N. Sodd, On the interpretation of our anterior cruciate ligament data, Clinical Orthopaedics 196 (1985) 26–34. [5] I. Fujita, T. Nishikawa, H.E. Kambic, J.T. Andrish, M.D. Grabiner, Characterization of hamstring reflexes during anterior cruciate ligament disruption: in vivo results from a goat model, Journal of Orthopaedic Research 18(2) (2000) 183–189. [6] M.D. Grabiner, K.R. Campbell, D.L. Hawthorne, D.A. Hawkins, Electromyographic study of the anterior cruciate ligamenthamstrings synergy during isometric knee extension, Journal of Orthopaedic Research 7(1) (1989) 152–155. [7] H.J. Hermens, B.F. Freriks, R. Merletti, D. Stegeman, J. Blok, G. Rau, C. Disselhorst-Klug, G. Ha¨gg, Seniam-European Recommendations for Surface Electromyography, Roessingh Research and Development, Enschede, 1999. [8] A. Imran, J.J. O’Connor, Control of knee stability after ACL injury and repair: interaction between hamstrings contraction and tibial translation, Clinical Biomechanics 13(3) (1998) 153–162. [9] H. Johansson, P. Sjolander, P. Sojka, A sensory role for the cruciate ligaments, Clinical Orthopaedics & Related Research 268 (1991) 161–178. [10] E. Kellis, V. Baltzopoulos, The effects of the antagonist muscle force on intersegmental loading during isokinetic efforts of the knee extensors, Journal of Biomechanics 32(1) (1999) 19–25. [11] M.R. Krogsgaard, P. Dyhre-Poulsen, T. Fischer-Rasmussen, Cruciate ligament reflexes, Journal of Electromyography and Kinesiology 12(3) (2002) 177–182. [12] W. Liu, M.E. Maitland, The effect of hamstring muscle compensation for anterior laxity in the ACL-deficient knee during gait, Journal of Biomechanics 33(7) (2000) 871–879. [13] J.J. O’Connor, Can muscle co-contraction protect knee ligaments after injury or repair? Journal of Bone and Joint Surgery Br 75(1) (1993) 41–48. [14] M.G. Pandy, K.B. Shelburne, Dependence of cruciate-ligament loading on muscle forces and external load, Journal of Biomechanics 30(10) (1997) 1015–1024. [15] S. Plagenhoef, F.G. Evans, T. Abdelnour, Anatomical data for analyzing human motion, Research Quarterly for Exercise and Sport 54(2) (1983) 169–178. [16] K.B. Shelburne, M.G. Pandy, A musculoskeletal model of the knee for evaluating ligament forces during isometric contractions, Journal of Biomechanics 30(2) (1997) 163–176. [17] A.B. Slifkin, K.M. Newell, Variability and noise in continuous force production, Journal of Motor Behaviour 32(2) (2000) 141–150.
315
[18] M. Solomonow, R. Baratta, B.H. Zhou, H. Shoji, W. Bose, C. Beck, R. D’Ambrosia, The synergistic action of the anterior cruciate ligament and thigh muscles in maintaining joint stability, American Journal of Sports Medicine 15(3) (1987) 207–213. [19] M.E. van Lent, M.R. Drost, F.A. v.d. Wildenberg, EMG profiles of ACL-deficient patients during walking: the influence of mild fatigue, International Journal of Sports Medicine 15(8) (1994) 508–514. [20] A.B. Zavatsky, D.J. Beard, J.J. O’Connor, Cruciate ligament loading during isometric muscle contractions. A theoretical basis for rehabilitation, American Journal of Sports Medicine 22(3) (1994) 418–423.
Idsart Kingma is an assistant professor at the Faculty of Human Movement Sciences of the ‘Vrije Universiteit Amsterdam’ where he is teaching courses on ergonomics and biomechancis. He has an MSc in human movement sciences and he finished his PhD on the biomechanics of lifting in 1999. Currently, his main research interests are mechanical aspects of low back loading and neuromuscular control of joint stability. Idsart Kingma was the first author of 17 and a coauthor of 23 papers in international scientific journals. Sietske Aalbersberg received her M.Sc. in Human Movement Sciences and is a medical doctor since 2002. Currently, she is a PhD student in Human Movement Sciences at the Vrije Universiteit in Amsterdam, The Netherlands. The focus of her research project is the neuromuscular control of the human knee with a specific interest in control strategies after ACL rupture.
Jaap van Diee¨n worked as a researcher at the Institute for Agricultural Engineering (IMAG-DLO) in Wageningen, the Netherlands from 1986 to 1996. He obtained a PhD in Human Movement Sciences from the Faculty of Human Movement Sciences at the ‘Vrije Universiteit Amsterdam’ the Netherlands in 1993. Since 1996 he has been affiliated to this faculty, since 2002 as professor of biomechanics. He is chairing the ergonomics program at this faculty. In addition, he is the head of a research group focusing on mechanical and neural aspects of musculoskeletal injuries. His main research interest is on control of muscles of the trunk and upper extremity, where especially the interaction of muscle coordination, fatigue, joint load and stability is an important research topic. Jaap van Diee¨n was the first author of over 40 and a co-author of a similar number of papers in international scientific journals. In addition he has (co-) authored numerous abstracts and book chapters in the international literature and technical reports and publications in Dutch. He serves on the editorial boards of the Journal of Electromyography and Kinesiology and Human Movement Sciences and is a regular reviewer for several other journals and funding agencies.