Author response.

We would like to thank Dr Houck for his comments on our article. First, we would like to clarify a misconception regarding the intent of the perturbation training program. Nonoperative management of active individuals with anterior cruciate ligament (ACL) injury remains controversial. Conventional wisdom suggests that most active young individuals with complete ACL rupture who do not undergo reconstruction continue to experience knee instability. A return to preinjury activities after surgery and postoperative rehabilitation, however, may not be possible for many months. Our perturbation training program has been directed at those individuals who want to or must delay reconstruction. The challenge becomes identifying and training individuals who have the potential to return to sporting activities for a defined period of time without extending their original knee injuries. Patients identified by a screening examination as "potential copers" who go on to participate in perturbation training have a significantly improved chance of returning to complete a competitive sports season without experiencing knee instability. (1) The long-term effect of participating in perturbation training, returning to sports, and not undergoing surgical reconstruction is currently being assessed by our collaborators in Norway, where patient management includes reconstruction only when nonoperative management fails.

The purpose of our study was to try to elucidate mechanisms behind the proven effects of perturbation training first demonstrated by Fitzgerald et al. (1) In his commentary, Dr Houck equated lower co-contraction with proof of success of the training. Perturbation training has previously been shown to successfully reduce the risk of extending the original knee injury based on functional measures and number of giving-way incidents. (1) The present study attempted to identify mechanisms underlying the improved knee stability. Thus, we do not see a lower co-contraction index by itself as an indication that subjects benefited from the training; rather, we observed that those who benefited from training showed reduced co-contraction and then speculated as to the long-term outcome of such a strategy.

Dr Houck's question about our use of an alpha level of P<.10 was addressed several times in the review process. Electromyographic (EMG) data are inherently variable even in subjects without injuries during routine tasks such as walking. (2) For that reason, we have historically used a P value of .10 for EMG analysis. We believe differences within this range represent real differences, and we are much more concerned about making a type II error than a type I error in these situations.

We agree with Dr Houck that people with ACL deficiency may activate muscles such as the gluteus maximus or soleus in an attempt to stabilize the knee. Indeed, our previous work has shown that "copers" do not use only one muscle stabilization strategy; rather, they use many different strategies, all characterized by rapid muscle responses. (3) Perturbation training was designed to elicit rapid responses in muscles that are activated to successfully complete the tasks that comprise the training; each individual responds uniquely to the demands placed on the knee. Which muscles are involved in an individual's strategy likely depends on a complex interaction of anatomical, biomechanical, neurological, and innate factors. Our primary focus on muscles that affect the knee in weight bearing emerged from the understanding that the shear forces associated with episodes of giving way are particularly detrimental to cartilage. (4) Other characteristics found in people with ACL deficiency such as reduced knee flexion (that would reduce shock absorption) and high knee muscle co-contraction (that would increase joint compression) also could exacerbate joint destruction. People with ACL deficiency use highly variable muscle activation patterns to stabilize their knees, sometimes involving muscles that do not cross the knee joint. The training may be so successful at allowing appropriately identified people with ACL deficiency to return to their sports for a limited time because we have accepted that there is no single pattern that suits all patients in all tasks.

Our selection of muscle co-contraction as a primary outcome measure was based on initial findings that patients with ACL deficiency and knee instability ("noncopers") tended to have high muscle co-contraction (3) and recent findings that potential copers also have higher muscle co-contraction during a destabilizing standing task compared with subjects without injuries. (5) Our calculation of muscle co-contraction is not simply the ratio of muscle activity in 2 muscles. Rather, our calculation is a point-by-point calculation (at the sampling rate of 980 Hz) using the integrated activity of [(less active muscle/more active muscle) x (sum of the integrated activity of both muscles)] and then integrating over the resulting curve. In this way, our calculation is sensitive to the timing of muscle activity, and unlike a simple ratio, it is normalized to the overall magnitude of both muscles combined. For this reason, changes in muscle co-contraction cannot be interpreted as resulting from a change in the magnitude of activity of one muscle.

With regard to how the co-contraction index would change with altered knee flexor or extensor activity, we will refer to the Figure for ease of illustration. The Figure shows the individual linear envelopes of the vastus lateralis (VL) and lateral hamstring (LH) muscles along with the trace that results from calculating co-contraction with our equation. We have generated graphs of EMG data to represent the conditions in Dr Houck's example of potential copers using lower VL activation along with no change in LH activation. As shown in the Figure, the co-contraction index is nearly identical in both cases because, although the ratio of VL/LH is higher in the bottom graph, the sum of their magnitudes is lower, thus resulting in little change in the co-contraction index. Although there is no question that dynamic stability in an ACL-deficient knee requires muscle co-contraction, functional outcomes suggest that a high magnitude of muscle co-contraction is not advantageous. We continue to study the co-contraction phenomena in those with ACL injury, osteoarthritis, and other disorders.

A central hypothesis of our research is that most patients use a joint-stiffening strategy to stabilize their knee after ACL injury, and this is based on concepts from the field of motor control. We describe an increase in the degrees of freedom as indicating motor learning. In contrast, Dr Houck points out that there are conditions where reducing the degrees of freedom signifies refinement of motor strategy, and he cites a study that used a slower application of force over a much longer distance than is used in our testing and training protocols. Consequently, an initial decrease in degrees of freedom seems a more appropriate theoretical underpinning for our work. Based on our clinical experience, patients show little joint movement during the initial 1 to 3 treatments of perturbation training, even though the speed of application is slower than what will occur in later treatments. We try to facilitate increased degrees of freedom (decreased muscle co-contraction) from the very beginning of training with verbal cues such as "respond only when you feel the board move" and so on. Findings of higher muscle co-contraction and a reduced knee flexion angle before training lend support to our joint-stiffening hypothesis, although, as Dr Houck pointed out, we have not actually measured muscle force in vivo. Although the magnitude of knee angle change after training is very small, it does indicate change toward increasing the degrees of freedom and, in the context of improved functional outcomes, has clinical significance. Most individuals complete training within a month, so it is unknown how knee kinematics change with further opportunity for motor learning.

We would again like to thank Dr Houck for his insight and questions on our study. The opportunity to engage in a discussion on this important topic will help advance the design of future studies and ultimately will lead to improved management of the patient with ACL deficiency.

Terese L Chmielewski, PT, PhD

Wendy J Hurd, PT, MS

Katherine S Rudolph, PT, PhD

Michael J Axe, MD

Lynn Snyder-Mackler, PT, ScD, SCS, ATC 

References

(1) Fitzgerald GK, Axe MJ, Snyder-Mackler L. The efficacy of perturbation training in nonoperative anterior cruciate ligament rehabilitation programs for physical active individuals. Phys Ther. 2000;80:128-140.

(2) Ivanenko YP, Poppele RE, Lacquaniti F. Five basic muscle activation patterns account for muscle activity during human locomotion. J Physiol. 2004;556:267-282.

(3) Rudolph KS, Axe MJ, Buchanan TS, et al. Dynamic stability in the anterior cruciate ligament deficient knee. Knee Surg Sports Traumatol Arthrosc. 2001;9:62-71.

(4) Setton LA, Mow VC, Howell DS. Mechanical behavior of articular cartilage in shear is altered by transection of the anterior cruciate ligament. J Orthop Res. 1995;13:473-482.

(5) Chmielewski TL, Hurd WJ, Snyder-Mackler L. Elucidation of a potentially destabilizing control strategy in ACL deficient non-copers. J Electromyogr Kinesiol. 2005;15:83-92.