Exercises Following Anterior Cruciate Ligament Reconstructive Surgery: Biomechanical Considerations and Efficacy of Current Approaches

Abstract

To enable a safe, effective return to daily functional activities and to prevent premature knee joint osteoarthritis, adults undergoing anterior cruciate ligament (ACL) surgery require carefully designed and appropriate rehabilitation strategies. In this article we critically examine the contemporary body of literature concerning the application and outcomes of various knee exercise rehabilitation protocols described in the literature in the context of the surgically repaired ACL injured knee. These data were obtained from the MEDLINE (1991–2007), CINAHL (1982–2007), and SPORTS DISCUS (1975–2007) databases, using the terms anterior cruciate ligament, exercise, and ligament, showed that two lines of research currently dominate this topic: (1) the role of open versus closed chain exercises approaches; (2) the impact of these exercises on knee stability and function post-ACL surgery. In light of the small number of studies and their methodological limitations, however, more research clearly is recommended to establish if closed-chain kinetic exercises are conclusively more efficacious than open-chain exercise for this condition, as predicted by the present literature base.

Keywords:

• anterior cruciate ligament injury
• exercise
• knee joint
• reconstructive surgery
• rehabilitation

HISTORICAL BACKGROUND

In efforts to restore optimal lower limb function after anterior cruciate ligament (ACL) reconstructive surgery, a variety of exercise protocols have evolved. In the 1990s, most often advocated were nonweightbearing or “open kinetic chain” knee extensor resistive exercises introduced only after the operated knee joint had been immobilized for 6 weeks or more (Arms et al. 1984), a strategy followed by a slow progression to functional activities (Frndak and Berasi 1991; Lieber 1992; Stanish and Lai 1993). Designed to protect the healing knee joint, limit excessive stresses on the knee extensor mechanism (Morrissey et al. 2002), and offset potential problems due to excessive patellofemoral joint stresses (Ross, Denegar, Winzenried 2001), this careful protracted postsurgical exercise approach conflicted with an already established knowledge base concerning the detrimental influence of suboptimal muscle “use” on joints such as the knee.

Subsequent attempts to include more aggressive exercise approaches into early rehabilitation protocols often stress the surgical site unduly, however, and did not sufficiently influence knee extensor torque production as was deemed desirable, regardless of graft type (Heijne and Werner 2007). In addition, performing repetitive knee extension movements in an “open kinetic” chain mode tended to foster excessive anterior tibial displacement (Beynnon and Johnson 1996; Heijne and Werner 2007; Kirkley, Mohtadi, Ogilvie 2001), thus potentially countering the stability objective of the surgery.

Because knee joint stability is an essential prerequisite for full functional recovery, exercise strategies following ACL reconstructive surgery that are not injurious to the joint and can maximise knee extensor torques clearly are indicated (Morrissey et al. 2004). Not surprisingly, therefore, several researchers interested in advancing the science of ACL reconstructive surgery and rehabilitation processes have tried to carefully examine factors that impact the postoperative application of exercise for the ALC reconstructed knee from both a positive and negative perspective and through a variety of basic and clinical empirical studies.

What should constitute “best practices” in light of this complex injury is unclear, however, since most current review and scientific publications in this area emanate from a limited number of investigative teams, and some of this data yield conflicting or equivocal results. Moreover, most have not discriminated between lesion sites and how this impacts the broader context of the problem of the ACL injury and reconstructive approaches, in addition to exercise applications.

Because achieving consistently successful management and rehabilitation outcomes for the patient after ACL reconstruction remains a challenge, we set out carefully to review the available data concerning factors influencing the healing and functional recovery processes associated with an ACL rupture and its surgical reconstruction. Then we examined the available evidence concerning the efficacy of undertaking “closed kinetic chain” exercises, which might be difficult for some to carry out if other tissue damage or diseases prevail, even though these potentially can heighten functional outcomes, versus “open kinetic chain” knee extensor exercises for rehabilitation, which might be useful if weightbearing is not recommended for the post-ACL reconstructed knee or if multiple injuries to the knee or lower limb prevail. In doing this we hoped to offer some guidelines for what might constitute “best practices” for guiding the rehabilitation process after ACL reconstruction, plus directives for future research.

Methods

Using the keywords anterior cruciate ligament, biomechanics, exercise, and exercise therapy we located relevant basic science and intervention-related articles in the existing databases published in English over the past 25 years. We then read and organized this material so as to provide a brief overview of this topic from an anatomical and physiological standpoint. Then in this article the effects of exercise on the ACL are discussed including a model of hypothetical changes in ACL graft strength after surgery as this pertains to the application of exercise. Rehabilitation and related studies that pertain to “open” versus “closed” kinetic chain exercises approaches in the context of ACL reconstruction were reviewed specifically, along with their benefits and limitations. Finally, we provide some recommendations for rehabilitation specialists regarding the application of exercise after ACL ligament reconstruction, plus some future directions for adding to the desired knowledge base in this area.

Topic Overview

In terms of knee function, the ACL is extremely important because in addition to serving a proprioceptive role (Johannson et al. 1990), it provides 86% of the support required by an individual against anterior displacement of the tibia on the femur when weightbearing. It also facilitates lateral and medial rotation of the knee joint (Louie and Mote 1987; Palmitier, Scott, Chao 1991; Seto et al. 1989) and provides lateral, medial, and posterior stability to the knee joint. Consequently, injury to this ligament can result in considerable functional disability.

Fortunately, even though the potential for functional recovery at the knee is decreased markedly without reconstructive surgery (Stanish and Lai 1993), a surgically reconstructed ACL tends to provide for a strong, stable knee joint that has the potential to enable individuals to return to their former activities (Frndak and Berasi 1991). Thus, surgical reconstruction has become the basic approach used in the management of an ACL injury.

The literature reveals, however, that numerous factors are likely to have an impact on an optimum return to function after ACL reconstructive surgery. These include, but are not limited to, the status of the knee joint at time of injury, the extent of the injury, the status of the surrounding knee joint structures, including muscle, timing of surgery (Almekinders et al. 1995), tibial fixation methods (Vergis and Gillquist 1995), incision type (Dalldorf, Alexander, Lintner 1998), and the differential impact of external and internal stressors on the healing process over the course of the postsurgical period. Other factors are the degree of joint force and resultant line of pull that rehabilitation exercises produce (Skinner et al. 1986); the nature of the exercises in terms of intensity, mode, frequency, and duration; and their impact on knee joint proprioception and on the ACL structure itself. Others that are not well researched include gender, age, mode of injury, fitness level, body mass, health status, lifestyle, and personal goals and needs (see Table 1).

Table 1. Considerations in the Choice of Exercises to Foster Optimal Rehabilitation in People Undergoing Anterior Cruciate Ligament Reconstructive Surgery

• Duration after injury

• Duration after surgery

• Host characteristics

• Age

• Gender

• Health status

• Musculoskeletal status

• Prior injury history

• Weight

• Nature of the injury

• Nature of the surgery/surgical techniques

• Lifestyle needs

In particular, the impact of exercise technique on the postsurgical site, which is commonly constituted by allografts or autografts constructed from the Achilles, patellar, or gracilis tendons, or bone – patellar-tendon – bone autografts (Marrale, Morrissey, Haddad 2007) is likely to be an important rehabilitation-related consideration. For example, even though the surgically reconstructed ligament using the patellar tendon is initially almost 170% times stronger than a healthy intact ACL (Frndak and Berasi 1991; Stanish and Lai 1993), if the load applied across an individual joint exceeds the threshold for accommodating this, this may cause graft failure, regardless of surgery type (Kurosoka, Yoshiya, Andrish 1987; Medvecky, Zazulak, Hewett 2007; Vergis and Gillquist 1995). Moreover, it must be noted that even though these grafts initially are secured quite firmly by screw plugs placed at the ends of the replacement or reconstructed ligament deep into the tibial plateau, as time progresses, these graft fixation points are found to be weaker than other points of the graft (Seto et al. 1989) and, over time, tend to shift to the graft itself. That is, as the healing process ensues, and the fixation sites become incorporated by bone, the ACL graft becomes weaker, and its strength can decline to approximately 50% of the healthy ACL during the first postoperative year (Seto et al. 1989; Stanish and Lai 1993). Further, even though both Washerloc and tandem washers and screws provide structural fixation properties in young human tibia that should be appropriate for intensive rehabilitation, it cannot be assumed that this applies to all surgical sites and older tibial bone. As well, caution is still advised because 57% of the interference screw fixations using human tissue were found to fail at loads below 500 N (Magen, Howell, Hull 1999), graft laxity or failure is correlated with an earlier return of range of motion (Almekinders et al. 1995), and press-fit fixation using bone–patellar-tendon–bone grafts still can fail under cyclic loads comparable with those applied in accelerated rehabilitation conditions (Seil et al. 1998).

While this decline in graft strength clearly has a large surgical and biological basis (Vergis and Gillquist 1995) and strongly is related to graft revascularization and “ligamentization” processes (Amiel, Kleiner, Akeson 1986), it is still important for the clinician to prevent failed ACL reconstruction due to extraneous factors (George, Dunn, Spindler 2006). As proposed by Stanish and Lai (1993), one contributory cause for graft failure may involve trauma- or microtrauma-induced overloading of the reconstructed knee as a result of inappropriate dosage or performance of various exercises and the fact that failure strain and tissue strength are reduced significantly once the graft is implanted (Fleming, Oksendahl, Beynnon 2005). In this respect, we further propose that some instances of failure toward graft incorporation may result from graft weakness as observed by Seto, Brewster, Lombardo et al. as a result of an inadequate healing process, due to the lack of adequate stress on the autograft, which has a viable cell population at the time of implantation that can respond to mechanical strain (Fleming, Oksendahl, Beynnon 2005). That is, healing tissue, unless adequately stressed, tends to be of poorer quality than normal tissue and partially could explain why there appears to be an extraordinary yielding of tissue, especially if one considers that the autograft is initially 65%–75% stronger than an intact ACL. Motion also may influence the extent to which the autograft, initially dependent on synovial fluid nutrition (Amiel, Kleiner, Roux et al. 1986), is successful in fostering healing.

There are sufficient data in our view, therefore, to suggest that even though ACL grafts can fail for a variety of mechanical reasons, therapists aiming to optimise dynamic mobility of the ACL reconstructed knee still need to evaluate carefully how to increase function of the dynamic restraints of the knee without compromising the reconstructed passive restraints (Medvecky et al. 2007). At present, however, no conclusions on the best way to do this can be drawn, because, first, none of the above authors reported on the graft's strength beyond a time period of 5 years; none examined what specific exercise frequency or dosage would be indicated/contraindicated taking the site and type of ACL repair into consideration; similarities or differences in exercise effects on various types of reconstruction approaches remain unclear; and data using well-defined endpoints and validated measurement tools are scarce. As well, even though Seto et al. (1989) identified that the ACL was composed of three distinct yet interwoven bundles, namely, (a) the anteromedial, (b) the intermediate, and (c) the posterolateral bundles, and that these respond differentially to tension, exercises applied to restore knee function in the context of ACL reconstructive surgery do not seem to have been applied in light of these facts.

Hypothetical Changes in ACL Graft Strength

To broadly conceptualize the relationship between strength of the fixation site with and without applied stresses and the ACL graft strength, over time, we elected to extrapolate the research discussed to this point to form Figure 1. The maximum tension curve hypothetically represents the largest amount of load that the graft unit is likely to endure without encountering the possibility of additional injury.

Figure 1. Conceptual model of the relationship among anterior cruciate ligament graft strength, fixation strength, and functional tension over the early rehabilitation period, depicting that the tension applied to the healing graft should be selected with the view that its strength is changing along with fixation strength over time. The specific amount of tension in light of surgery type remains to be determined, however.

If we generally accept that the majority of postoperative clients will participate in a rehabilitation protocol for approximately 6 months, Figure 1 shows that during this period the maximum tension possible on the graft is increasing and may continue to increase for an additional month or 2, lasting no more than 7 or 8 months. This is within the time period where the client is no longer attending a formal rehabilitation program and is probably returning to preinjury activities. If the client attempts to return to activities that place significant stress on the reconstructed ACL, it is possible they might find the ligament has less stability than anticipated or that it “underperforms” comparatively speaking. That is, the purpose of this figure is to try to identify at what point in time exercises should be introduced in order to facilitate return to function and how these should be introduced in order to avert potential graft failure.

This idea is not entirely new and concurs with opinions of others who suggest that in addition to surgical reconstruction of the ACL injured knee, a carefully tailored postsurgical rehabilitation protocol that is developed in light of the individual patient's unique situation and the type of surgery that has ensued is more likely to fully restore limb function than one that is not carefully applied (e.g., Henning, Lynch, Glick 1985). Indeed, even though rehabilitation plays an extremely important role in the recovery process of individuals after ACL reconstruction, the current science base indicates this may be significantly compromised if the individual is exposed to either an excessively conservative or an excessively overzealous exercise regimen. Similarly, introducing exercise prescriptions that are less than satisfactory for mediating functional outcomes or that are not appropriate for the stage of healing or nature of the injury itself is likely to result in less than satisfactory long-term outcomes.

Other Factors Influencing Exercise Applications Post-ACL Surgery

Postsurgical Joint Status

Among the many considerations that are warranted in efforts to positively impact recovery of the ACL reconstructed knee, appropriate initiation of non‐weightbearing exercises designed to mitigate knee extensor muscle atrophy postsurgery is crucial. This is because even though the primary goal of all ACL rehabilitation programs is to initiate knee joint movements as soon as possible after surgery to prevent muscle atrophy (Frndak and Berasi 1991), anatomical research has shown that when the knee extensors contract, their line of action through the patellar tendon can cause anterior tibial displacement (Wilk, Reinold, Hooks 2003). Thus, even though carrying out such exercises soon after surgery may allay muscle atrophy, such exercises, if not carefully implemented, could inadvertently impact optimal healing of the reconstructed knee joint. It is also possible that the rapid wasting of the knee extensors seen post-ACL surgery is a protective mechanism designed to minimize anterior tibial motion of the reconstructed knee joint. Alternately, the rapid muscular atrophy seen postoperatively may be associated with the injury and its concomitant effects, such as inflammation, and is thus an important protective mechanism in this respect, rather than a postsurgical immobilization response, requiring exercise to counteract this.

Joint Range of Motion

Another factor warranting consideration and potentially impacting recovery is the joint range used in carrying out knee extensor movements. For example, Stanish and Lai (1993) found that movements applied to the knee joint after ACL surgery that extended beyond 30° of knee flexion increased joint swelling, but if performed in the final degrees of extension, there was a noted decrease in quadriceps inhibition, as well as an improved healing rate. In addition to joint swelling, patello-femoral pain syndrome symptoms, which could affect optimal recovery rates, commonly are aggravated when the knee is exercised beyond 30° of knee flexion (Palmitier et al. 1991).

Even though it appears that movements within the range of 0–30° of knee extension would be beneficial to the healing process, however, the benefits of this restricted approach still would be contingent on the specific ACL bundle injured, because, as previously discussed, the intermediate and posterolateral ACL bundles become taut when the knee is extended and the anteromedial fibers become taut when the knee is flexed. Thus, even though Bynum, Barrack, and Alexander (1995) successfully applied passive movements extending from 0° to 60° to all their study subjects each day after surgery until discharge for 12 hours, in our view this exercise approach probably should be advocated only if the ACL lesion has occurred in the anteromedial fibers.

In summary, even though surgery effectively assists in improving knee stability after an ACL injury, early movement of the postsurgical knee must not jeopardize the integrity of the ACL graft fixation sites, or place the limb in joint ranges known to increase excessive tension in the ACL, or both. Furthermore, passively placing the joint at the extreme ranges of joint flexion and extension where the ACL becomes taut or at 30° of knee extension should be avoided (Frndak and Berasi 1991; Palmitier, Scott, Chao 1991).

Exercise Dosage

The quantity of exercises an individual performs in a given period after ACL reconstruction surgery may similarly impact recovery. For example, as indicated in a study performed by Grana and Muse (1988), even low workload cycling (i.e., at 60% of MVO₂) can increase the average anterior laxity of both normal and ACL reconstructed knees. Although the difference was not significant (p < 0.01), average laxity between 25 participants increased 12%. Normal knee laxity increased 21%. Of all the participants, six had reconstructive surgery of their ACL, and their combined results showed an increase in average laxity (i.e., an increase of 15%). Similar outcomes have been demonstrated on an in vitro animal model (Burroughs and Dahners 1990), by Skinner, Wyatt, Stone et al. (1986) in athletes, and a study by Kirkley, Mohtadi, and Ogilvie (2001).

As well, an inappropriate exercise dosage that results in muscle fatigue may not only foster knee joint instability as discussed by Skinner, Wyatt, Stone et al. (1986), it also may jeopardize healing of the newly constructed graft (Kirkley, Mohatadi, Ogilvie 2001). That is, even though exercises prescribed for a particular type of ACL reconstruction may benefit the recovery process, a rehabilitation protocol that is not carefully applied may prove more harmful than beneficial in both the short- and long-term postsurgical periods. In addition, the clinician may want to acknowledge and take into account the probable presence of other tissue lesions that could have occurred concomitantly at the time of the ACL injury.

Exercise Mode

In addition to the importance of identifying the “safe” time period, appropriate knee joint range, and exercise dosage for implementing knee extensor exercises post-ACL surgery, much recent debate in this area has focused on whether the individual should perform these in an “open” or a “closed” kinetic chain mode.

In some cases, several authors have shown that exercising the knee extensors in an “open” kinetic chain mode substantially increases anterior displacement of the tibia on the femur, and, according to Morrissey, Hudson, Drechsler et al. (2000), it is probable that this form of exercise is contributory to the 9% increase in joint looseness that can occur postoperatively. By contrast, carrying out knee extensor exercises in a “closed” kinetic chain mode may minimise ACL stresses (Beynnon et al. 1995; Ohkoshi et al. 1991; Palmitier et al. 1991; Yack, Collins, Whieldon 1993) or produce ACL strain responses that are equal to or similar to those produced during other rehabilitation exercises (Heijne et al. 2004). Moreover, according to Wilk, Escamilla, Fleisig, et al. (1996) a well-controlled exercise routine of static or dynamic squats performed in the immediate postoperative period is unlikely to place any excessive stress on the ACL surgical site because they result in a net mean posterior displacement force at the femorotibial joint.

Beynnon, Fleming, Johnson et al. (1995) showed, however, that exercises performed in an “open” chain mode could be applied post-ACL surgery if restricted to particular ranges. This was because the ACL was strained least when it was held statically at 60° and 90° of flexion, while it was strained most at 30° and 15° of knee extension. Other experiments performed by Beynnon, Fleming, Johnson et al. (1995) showed ACL strain increased during active extension between 50° and full extension (Beynnon and Fleming 1998) as compared with active flexion. This phenomenon was reversed when resistance was added against the direction of movement. Thus, flexion created more anterior strain than did extension. As well, ACL strain now occurred from full extension to 45° of knee flexion.

This latter finding was not observed by Heijne, Fleming, Renstrom et al. (2004), who recorded equivalent ACL strain values between flexion and extension cycles of open and closed kinetic chain exercises. However, Beynnon, Fleming, Johnson et al. (1995) and Heijne, Fleming, and Renstrom et al. (2004) did not study individuals with reconstructed ACLs and how they would respond biomechanically to the same weighted knee extension exercises. Likewise, they did not examine whether this response might differ depending on lesion site. An additional problem in the more recent study by Perry, Morrissey, King et al. (2005), who found equality of knee laxity outcomes 14 weeks after ACL reconstructive surgery, was that the final quadriceps loads were more than 10 times lower in the “open” chain exercise group than that of the “closed” kinetic chain exercise training group, and the hamstrings load in the “open” chain kinetically trained group was 33% higher than that used to train the “closed” chain exercises group. The time from injury to surgery, which was longer in the “open” chain trained group, and body mass, which was higher in the “closed” chain group, although nonsignificant, could have impacted the outcome, as could the activities undertaken outside of physiotherapy, which included many weightbearing activities.

In addition, because the strain characteristics on the ACL vary tremendously depending on the specific form of exercise and the range within which the exercise is performed, the strain pattern placed on the reconstructed ACL may be radically different when isometric, rather than isokinetic, exercises are used. Palmitier, An, Scott et al. (1991) showed there was significant stress applied to the ACL when the quadriceps were contracted in the range of both 0° to 30°, which would support the results of Beynnon, Fleming, Johnson et al. (1995); Heijne, Fleming, Renstrom et al. (2004); and Sato, Higuchi, Terauchi et al. (2005). The increased ACL strain however, also was found with isometric quadriceps exercises beyond 60° of flexion. In the same article, Palmitier, An, Scott et al. reviewed a study that found that isometric quadriceps exercises at 0° and 22° of knee flexion resulted in 5–17 times as much strain as normal gait on a flat surface. Although the experiment consisted of only two subjects and provided no reference to angles that were measured in the gait cycle, it again emphasized the necessity for caution in applying exercises for maximizing rehabilitation outcomes after ACL surgery.

A reasonable body of data does shows however, that in the normal lower extremity, the quadriceps to hamstring strength ratio is approximately 3 to 2. Therefore, those study results indicating that there is an increase in the strain placed on the ACL within the ranges of full extension to midrange knee flexion (e.g., Arms et al. 1984) also may reflect the fact that the quadriceps muscles are capable of overpowering the hamstrings within this range and thus can cause anterior displacement of the tibia, and subsequent ACL strain. Although the postoperative knee may perform quite differently than that of the healthy knee, Fleming, Ohlen, Renstrom et al. (2003) found that it may be possible to increase the activity of the quadriceps muscles without increasing the strain on the ACL by applying a compressive load to the joint. While there may be less concern today with this issue because of advanced graft fixation techniques (Fitzgerald 1997), Fleming, Ohlen, Renstrom et al.'s results supported the use of closed kinetic chain exercises as an alternative approach to open kinetic exercises for maximizing the surgically reconstructed ACL recovery process, and for lowering the possibility of the individual inadvertently experiencing undue ACL strain as a result of some forms of exercise.

Additionally, irrespective of this potential postsurgical benefit, what is certain is that all exercises that are performed in an “open” kinetic chain mode are likely to lack the specificity and sensory feedback that “closed” kinetic chain exercises promote. In addition, as identified in several studies, to be done safely, “open” kinetic chain exercises need to be executed within a range that is potentially difficult for an individual who has undergone reconstructive knee joint surgery to attain. As well, intra-articular swelling and patello-femoral pain symptoms could both be aggravated by exercises within this range.

Moreover, despite the limitations revealed in the present database, the use of “closed” kinetic chain or weightbearing exercises seems promising if we consider they can achieve the following:

1. Promote normal muscle activation and coactivation.

2. Help maintain or promote muscle strength and endurance.

3. Provide sensory feedback, i.e., tactile and proprioceptive inputs that can help to minimize pain via the pain gating mechanisms.

4. Increase quadriceps muscle activation without increasing strain on the ACL reconstructed knee (Fleming et al. 2003).

5. Provide the benefit of functional specificity of training principle to the limb (Palmitier et al. 1991), and activate the gastrocnemius muscle, which may be protective because it is an ACL antagonist (Fleming et al. 2001).

6. Induce stronger contractions within the hamstrings muscles, which serve as an important knee joint protective mechanism.

We believe this last point is particularly important because most functions of daily living and activity, especially sporting activities, require the lower extremity muscles to be used interactively in a closed kinetic chain mode. By contrast, many functional movements and types of contractions cannot be replicated in an open kinetic chain mode. Consequently, as pointed out by Palmitier, An, Scott et al. (1991), if exercises designed to rehabilitate the ACL reconstructed knee joint are carried out in isolation, their application might limit the potential for full functional recovery, even if graft “safety” is not a factor.

Application of Closed Kinetic Chain Exercises After Anterior Cruciate Ligament Surgery

The small number of controlled studies that have examined the use of closed kinetic chain exercises for improving function, plus some of their limitations, are shown in Table 2. As outlined by Fitzgerald (1997) and Bynum et al. (1995), who conducted a randomised study of open and closed chain exercises after ACL reconstruction, closed kinetic chain exercises are safe and effective and offer several functional advantages over open kinetic chain exercises after ACL reconstructive surgery. These include improvements in knee stability, pain, satisfaction, and the ability to rise from sitting to standing and to return to normal activities and sports (see Table 2).

Table 2. Summary of Randomized Controlled Studies of Closed Versus Open Chain Exercise Applications Following Anterior Cruciate Ligament Reconstructive Surgery

Authors	Sample	Study Design	Finding	Limitations
Bynum, Barrack, and Alexander (1995)	Patients with ACL reconstruction	2 grp comparison with tests at a min 1 yr	At 19 months, CKC grp had better stability, less pain, more satisfaction with outcome	Ex. dosage uncertain/unequal Baseline laxity status unknown No data on functional outcomes
Hooper, Hill, Drechsler, et al. (2002)	32 patients	2 grp comparison 2 and 6 wks postsurgery 2 grp comparison CKC vs CKC plus OKC	OKC and CKC resulted in similar isokinetic torque values	No other measures reportedUnclear exercise dosage
Mikkelson, Werner, and Eriksson (2000)	44 patients		CKC plus OKC yielded higher numbers of patients returning to prior sports, and returned 2 months earlier than those who carried out CKC only with no compromise in stability at 6 m	Not certain therapist/assessors were blinded
				Exercise dosage unclear/unequal
				No laxity measures in directions other than anterior laxity assessed
				Demographic data at baseline not clearly shown to demonstrate equivalency grps
Morrisey, Hudson, Drechsler, et al. (2000)	36 patients	Pts received OKC or CKC exs for 4 wks and were tested for knee laxity at 2 and 4 wks	OKC grp had 9% greater laxity, but this felt to reflect greater laxity in OKC grp	Short-term study Exercise dosage unclear
Morrissey, Drechsler, Morissey, et al. (2002)	43 patients	CKC and OKC exs were compared for pain effects	At 2 and 6 wks there were no differences in pain scores	No other measures reported
				All pts received PT
Perry, Morrissey, King, et al. (2005)	49 patients recovering from ACLR surgery received formal PT plus either OKC or CKC exercises 3x per wk	RCT with tests at 8 and 14 wks	No diff grps in knee laxity or function	–All had PT
				–Follow-up time limited
				–Bicycling included
				–Dosage exs not equitable
				–Proprioception exs included
Abbreviations: ACL=anterior cruciate ligament; ACLR: anterior cruciate ligament replacement; CKC=closed kinetic chain exercises; OKC=open kinetic chain exercises; PT=physical therapy; Pts=patients; RCT=randomized controlled trial; diff=differences; grps=groups; wks=weeks.

Although this finding was not consistent with that of Hooper, Hill, Drechsler, et al. (2002), Palmitier, An, Scott, et al. (1991) felt that it was both the movements involved and the resultant muscle effects that constituted a sound rationale for this rehabilitation approach. Moreover, the authors further suggested that because each weightbearing exercise combines every type of contraction (i.e., isometric, concentric, eccentric) within a movement, the neural adaptations following closed kinetic chain exercises are likely to be quite different from those achieved as a result of exercises performed in isolation. Thus, although nonweightbearing exercises may provide strength and endurance gains to the exercised muscles (Hooper, Hill, Drechsler et al. 2002), their mode of activation is unlikely to replicate that used in functional activities or that desired for effectively stabilizing the postsurgical knee joint when the individual attempts to ambulate for the first time, even though Hooper et al. (2001) found insignificant effects under controlled level walking and stair walking conditions. As well, performing nonweightbearing knee extension exercises with the knee held in a slightly flexed position would train the knee extensors to contract when the knee is in the frontal plane alone and without the hamstrings coactivation that occurs in weightbearing and that facilitates knee stability (Palmitier et al. 1991). Palmitier, An, Scott et al. (1991) further indicated that studies that have followed individuals who performed joint isolation exercises have identified these individuals as having increased knee varus and valgus laxity.

In addition to the fact that the introduction of open chain extensor exercises after ACL surgery might need to be delayed to reduce or prevent the risk of any potential graft failure, according to Barber-Westin, Noyes, Heckmann et al. (1999), a rehabilitation program after ACL reconstruction that includes immediate knee motion and early weightbearing is not injurious and results in an acceptable failure rate of 5% of 142 surgeries. In contrast to this view, in the study by Mikkelson, Werner, and Eriksson (2000) that compared closed chain plus open chain exercises, there was a significantly greater improvement in quadriceps torque without reducing joint stability at 6 months in the latter group, plus an earlier return to previous activity. It was unclear, however, whether this was due to the nature of the exercises, or the increased exercise dosage, or other factors. Similarly, while a prospective, blinded, randomised trial of isometric quadriceps exercises conferred faster recovery of knee joint range of motion and stability than no exercises throughout the first two postoperative weeks, there was no comparison group to ascertain the relative impact of closed kinetic chain exercises for quadriceps strengthening in the first two postoperative weeks (Shaw, Williams, Chipase 2005).

As well, an increasing body of data has shown that proprioception in the reconstructed knee does not match that in the normal knee (Bonfirm, Jansen Paccola, Barela 2003; Friemert et al. 2006; Gomez-Barrena et al. 1999) and that proprioceptive feedback is needed to reinstate the necessary feedback that influences mechanical stability of the knee. There are also other neural adaptations that need to occur concomitantly to ensure the muscles surrounding the knee are sufficiently functional to contribute optimally to knee stability. Thus, in our view, these are all stages warranting careful consideration in planning for optimally adequate training and muscle reeducation and agree with the requirements for a successful return to premorbid activities using the six-phase paradigm suggested by Curl, Markey, and Mitchell (1983) more than 20 years ago: presurgery, to prepare the patient for postoperative rehabilitation; postoperative (or postinjury), to allow healing and to prevent thrombosis and muscle atrophy; early healing, to maintain muscle tone and joint motion in a protective device; late healing (water stage), to begin proprioceptive and agility training while regaining joint motion; healed (land stage), to gain greater agility and confidence in controlled situations; and competition, to demonstrate if the rehabilitation program has been successful. Although squatting, which produces substantial compressive joint forces, does not necessarily protect the ACL more than active flexion-extension movements of the leg, unlike the impact of increasing the level of resistance during active flexion-extension exercises, increasing the resistance level during the squat exercise does not produce significant increases in ACL strain values (Beynnon et al. 1997), and may stimulate proprioceptors in deeper knee joint tissues quite favourably. According to Beutler, Cooper, Kirkendall et al. (2002), one-legged squats and step-ups would be effective not only in muscle rehabilitation, but as functional, closed chain activities they also might be protective of the reconstructed ACL.

CONCLUSION

The application of carefully planned post-ACL reconstruction rehabilitation programs, which can help regulate the strain environment of the graft, while preventing muscle atrophy (Fleming et al. 2005) are important to successful long-term outcomes of indviduals who have undergone ACL reconstruction, even though intraoperative and postoperative factors and biological causes of ACL may be paramount (Vergis and Gillquist 1995). For example, notwithstanding surgically related and biological causes of tissue failure, ample evidence shows that due to the increased stress placed on the graft, leading to a creep response in the healing tissue (Jenkins et al. 1997), excessive anterior tibial displacement during the healing process that can occur with some forms of knee extension exercises may contribute toward excessive graft strain, excessive ligamentous laxity, or postopertative reconstruction failure. In addition, muscles surrounding the ACL reconstruction that are not properly rehabilitated also may produce joint dysfunction secondarily (Shrier 2006), as may exercise induced excessive patellofemoral joint stresses (Ross et al. 2001).

This article focused on those exercises commonly used or advocated for promoting efficient rehabilitation of the individual who has undergone ACL reconstruction. It also discussed the variations among ACL injury sites and approaches and identified that there are no currently available tools for the clinician to assess and diagnose particular lesions of the ACL. Thus, results of several exercise experiments presently reported may have been confounded by failure to identify the particular area of the ACL that is damaged. Moreover, this situation could result in exercises being advocated within a specific joint range or in a limb position that actually may be injurious to the healing graft. The influence of surgery approach and when exercise should be implemented and how this should ensue was not discussed either to any degree, even though Friemert et al. 2006 have shown that ACL surgical outcomes can be impacted by all of these factors.

On a related topic, because the strain characteristics of the ACL graft have not been identified, and it is not known if a graft behaves similarly to that of an intact ACL, although it is assumed that the biomechanical behaviour of the graft never returns to normal (Beynnon and Johnson 1996), it seems prudent to conduct serial displacement measures and to avoid knee joint stresses that are likely to prove harmful rather than helpful to the graft (Barber-Westin et al. 1999). As well, the extent to which surgeons currently are replacing the graft in a similar location to an intact ACL is unclear and since this might effect tibial rotation, flexion, extension, and the associated accessory movements of the tibia on the femur, as well as mechanoreceptors within the ligament, differentially, this situation needs to be studied. Different graft types and their interaction with various exercise approaches can impact anterior tibial translation differentially after ACL surgery (Heijne and Werner 2007; Sato et al. 2005), and by knowing something about the surgical technique and the graft's response characteristics, clinicians might better understand which postsurgical strains on the ACL are likely to be safe and which are likely to be potentially injurious to this healing tissue and can try to avoid this situation more successfully. They could also gauge what type of impact the ACL surgery is likely to have on motion detection, speed, and acceleration detection as well as joint position sense and muscle coordination, and tailor their instructions accordingly.

In particular, if curves similar to that depicted in Figure 1 could be developed for the various ACL grafts, clinicians would be in a better position to advise clients about which exercises are likely to exceed the critical failure point and which are likely to do more harm than good. A better understanding of the stresses and strains imposed on the healing tissue by different forms of exercise and how these can be identified to optimally match the recovering ACL characteristics and desired outcomes would be helpful as well in this respect. Thereafter, exercise protocols could be developed on a more scientific basis and the clinicians could feel more confident that their clients will achieve optimal functional status notwithstanding the limitations of the injury and surgery. Many questions are as yet unanswered, however, and, as such, this lack of information may promote the inadvertent use of unsafe exercises or movements. At the same time, clinicians are faced daily with individuals recovering from ACL trauma, an injury that frequently afflicts and disables those enjoying a demanding active lifestyle (Gotlin and Huie 2000). What should they as professionals be administering as an efficacious ACL protocol? As discussed by Blackburn (1985), it is likely that the success of any rehabilitation effort will depend on the quality of the communication among surgeon, therapist, and trainer, and the nature of the problem in its entirety. That is, each must understand what the other is doing and must follow the biomechanical and healing restraints of the surgery and understand that the soft tissue healing is ongoing.

Yet, as was observed 25 years ago or more (Mikkelson, Werner, Eriksson 2000), we have found no consensus in the literature concerning what might constitute the best exercise program for rehabilitating the person with a reconstructed ACL. In addition to that, there were several limitations in all studies reported in Table 1. Moreover, even though most programs attempt to normalize joint range of motion, muscular strength and endurance, and joint stability, and to decrease pain and swelling, and to allow for a return to function within the individual's daily activities, few consistently examined graft healing, pain, swelling, and laxity in any rigorous prospective way. Muscle strength, a predictor of safe return to sports activity (Mikkelson, Werner, Erikson 2000), was not regularly assessed either.

As regards training using open kinetic chain knee extensor exercises in the postoperative recovery period after ACL reconstruction, it appears prudent that these exercises be restricted to those ranges where the knee extensors can be trained without harming the joint. Finally, it seems likely that weightbearing exercises that provide sensory feedback and result in total limb use are potentially more suitable for promoting functional recovery than are nonweightbearing exercises.

Indeed, a large body of data imply closed kinetic chain exercises have the potential to minimise strain on the reconstructed ACL, to decrease anterior tibial displacement, and, at the same time, to normalize knee joint physiology. Because the available evidence is not clear cut, however, we concur with Beynnon and Johnson (1996) that additional studies to document the comparative efficacy of closed kinetic chain and open chain exercises or others are indicated. In addition, while Perry, Morrissey, King et al. (2005) found no benefits of closed kinetic chain versus open kinetic chain exercises on knee laxity or leg function in the middle period of rehabilitation, and open-kinetic and closed-kinetic chain exercises may not impact differentially on the healing response of the ACL reconstructed knee (Fleming, Oksendahl, Beynnon 2005), additional studies that strive to determine the actual loads transmitted across the knee, along with the ACL graft strain characteristics during these exercises and how to relate these to the healing response of the graft, may be revealing. The potential importance of proprioception and balance training in the early rehabilitation phase, which recently was examined and yielded improvements in strength and proprioception in the proprioceptive group, with no detrimental effects on ACL laxity, should be explored further (Cooper, Taylor, Feller 2005), as should the benefits of early continuous active motion (Friemert et al. 2006).

As well, based on observations of Sato, Higuchi, Terauchi et al. (2005) and as stressed by Wilk, Reinold, and Hooks (2003), because it is likely to be important to implement a rehabilitation program in consideration of graft type, along with any other related procedures, this idea deserves study. In the absence of such data, it seems incumbent upon rehabilitation specialists to try to clinically ascertain the site of the ACL lesion and to carefully weigh how this can impact exercise choice and which movements should be emphasized or avoided. As discussed by Beynnon, Uh, Johnson et al. (2001), not only is restoration of anterior–posterior laxity values to within normal limits important, but also the biomechanical behavior of the graft produced by flexion-extension of the knee should be appreciated at all times.