FACTORS THAT AFFECT POSTOPERATIVE REHABILITATION AND OUTCOMES

Considerable advances have been made in the treatment of complete anterior cruciate ligament (ACL) ruptures and reconstruction methods since the mid 1980s. These include the appropriate selection of patient candidates and criteria that should be achieved before surgery, such as resolving limitations of knee motion, muscle atrophy, gait abnormalities, pain, and joint effusion. Appropriate graft selection, harvest, implantation, tensioning, and fixation are all paramount to achieving a reliable and desirable outcome, and considerable attention has been devoted to these principles in the orthopaedic literature.^18,19,62

Appropriate postoperative rehabilitation after ACL reconstruction is critical to achieve normal knee function and prevent complications such as arthrofibrosis and reinjury. The goals of postoperative therapy are to regain normal knee motion, gait mechanics, lower extremity muscle strength, coordination, proprioception, and neuromuscular indices using exercises and modalities that are not deleterious to the healing graft. Dye introduced the concept of restoring normal knee osseous and soft tissue homeostasis after ACL injury and reconstruction using an appropriate combination of medical and therapeutic measures.^45,46 The exercise program should not produce undue forces on the patellofemoral or tibiofemoral compartments or result in chronic joint effusions. Unfortunately, few investigations have studied the effect of specific exercises and treatment modalities commonly used after ACL reconstruction. In fact, entire protocols have appeared and been used extensively based on clinical observations and retrospective analyses instead of prospective, randomized, controlled clinical trials.^83,179

One of the difficulties in conducting rehabilitation investigations is the multitude of factors that may affect both the initial and the long-term recovery after ACL reconstruction. In addition, few ACL ruptures are truly isolated in nature, because approximately 80% of patients sustain concomitant bone bruises^52,93 and 60% suffer meniscus tears.^130,136 The factors that affect recovery include

Authors have noted no benefit^28,117 and, in some cases, deleterious outcomes^{120,176,182,205} when ACL reconstruction is performed early after injury before the resolution of limitations in knee motion, muscle atrophy, swelling, and pain. Meighan and coworkers¹¹⁷ conducted one of the few prospective, randomized investigations regarding this issue. These investigators followed 31 patients who received either an early ACL hamstring reconstruction (within 2 wk of the injury) or a delayed reconstruction (8–12 wk) to determine outcomes up to 1 year postoperatively. The early reconstruction group had significantly less knee extension and flexion at 2 weeks postoperatively, and greater deficits in quadriceps isokinetic work and power at 12 weeks postoperatively, than the delayed group. No other differences were noted in outcome throughout the study period. The authors concluded that early reconstruction did not provide any benefit to athletic individuals. Shelbourne and associates¹⁸² noted an increased rate of arthrofibrosis in patients who underwent acute ACL reconstruction (within 1 wk of the injury) compared with those in whom the reconstruction was delayed for 21 days or more. Shelbourne and colleagues^166,177 have heavily emphasized the need to restore normal knee motion when possible before surgical intervention. The exception is the presence of a mechanical block to extension, such as a bucket-handle meniscus tear or ruptured ACL that is impinged in the intercondylar notch. In these cases, early surgical intervention is warranted to repair the meniscus tear or remove the ACL, regain full knee extension and flexion, and then proceed later with ACL reconstruction.

An important issue is whether the postoperative rehabilitation process should be governed and progressed according to the amount of time that has elapsed or by well-defined and measurable criteria. Although some programs allow a return to full activities according to a certain amount of time, such as 6 months postoperative, others will not release patients until specific muscle and function objectives have been achieved regardless of the amount of time that has passed since surgery. There is no consensus among investigators regarding which factors are the most important to measure to determine whether a rehabilitation program has been successful in allowing patients to return to activities safely.

Modern studies of ACL bone–patellar tendon–bone (B-PT-B) autograft reconstruction typically report low failure rates of approximately 5% to 10%.^{10,67,129,143,181} Failure is defined as an increase in anteroposterior (AP) displacement of 6 mm or greater compared with the normal contralateral limb or a fully positive (grade II or III) pivot shift test. However, the percentage of grafts that undergo some amount of elongation, resulting in 3 to 5 mm of increased AP displacement or a mildly positive (grade I) pivot shift test, is quite variable and ranges from 5% to 50%.^{1,7,72,129,143,173,174,206} Therefore, although the overall rate of failure is low, some authors express concern that any amount of abnormal anterior tibial translation may be detrimental to the knee joint over the long term.^20,148 Whether graft elongation occurs as a result of technical aspects of the operative procedure, inconsistencies in maturation of the collagenous and bony components of the construct, the rehabilitation program, or a combination of these controllable and uncontrollable factors is unknown. Only one study has been conducted, to our knowledge, that followed knees reconstructed with ACL B-PT-B autografts with serial KT-2000 testing throughout the postoperative period (2 yr), discussed in detail later in this chapter.¹⁰

The purpose of this chapter is to review the current knowledge surrounding rehabilitation after ACL autogenous reconstruction. The scientific basis for immediate knee motion, early weight-bearing, specific exercises, and evaluation criteria for return to activity are reviewed. The authors’ ACL postoperative rehabilitation programs formulated from the scientific literature and nearly 3 decades of empirical clinical experience are detailed in Chapter 13, Rehabilitation of Primary and Revision Anterior Cruciate Ligament Reconstruction.

AUTOGENOUS ACL GRAFT MATURATION IN HUMANS

More than 100,000 ACL reconstructions are performed in the United States each year.¹⁴⁶ Few studies have been conducted on the maturation process of ligament grafts in humans; only a minuscule sampling (<1%) of B-PT-B and semitendinosus-gracilis (STG) grafts have undergone histologic analysis after implantation. The strong potential for sampling error prevents conclusions on when graft maturation is complete, indicated by when the transplanted tissue resembles a normal ACL’s histologic, structural, biomechanical, and material properties.

Rougraff and coworkers¹⁶⁴ performed biopsies on 23 B-PT-B autografts 3 weeks to 6.5 years after implantation. The authors reported that all patients had “clinically stable” knees; however, Lachman, pivot shift, and knee arthrometer data were not provided. The patients were allowed full weight-bearing immediately postoperatively and returned to sports between 3 and 6 months after surgery. The authors described four stages of ligamentization. The first stage occurred during the first 2 postoperative months and was characterized by an increased number of fibroblasts compared with time zero, preservation of mature collagen, and early neovascular ingrowth. The second stage, composed of rapid remodeling, occurred from 2 to 10 months postoperatively. A rapid increase in the number of metabolically active fibroblasts and replacement of approximately two thirds of mature collagen with immature matrix was noted. The third stage, in which the fibroblast count slowly decreased, collagen matrix matured, and vascularity decreased, occurred from 1 to 3 years postoperatively. In the final stage, the grafts were noted to be less cellular and vascular and appeared similar to those of control ACLs. The study did not include electron microscopy to evaluate the pattern of collagen fibril diameters in the grafts.

Rougraff and Shelbourne¹⁶⁵ performed biopises on nine B-PT-B autografts between 3 and 8 weeks after implantation. All patients followed an accelerated rehabilitation protocol¹⁷⁹ and agreed to the biopsy for investigational purposes. All specimens showed evidence of survival of portions of the original tendon. Vascular invasion was present at 3 weeks postoperative, with increased cell counts compared with controls found in all samples. All specimens had areas of acellularity and degeneration; however, these were small (no more than 30% of the biopsied area) in comparison with the areas of vascularity and normal-appearing tendon tissue.

Petersen and Laprell¹⁵³ examined the insertion of B-PT-B and STG autografts to bone in 14 knees that required revision ACL reconstruction. The time from the primary ACL reconstruction to revision ranged from 6 to 37 months. The hamstring specimens revealed a fibrous insertion of the graft into the periphery of the tibial bone tunnel. The patellar tendon (PT) specimens taken from the femoral tunnel showed fibrocartilage at the bone-tendon interface and resembled the chondral insertion of a normal ACL.

Johnson⁹⁴ performed biopsies on 20 STG ACL grafts between 20 days and 44 months postoperative. All patients required follow-up arthroscopy for various symptoms, at which time biopsy samples were taken from the tendon graft and surrounding fibrous tissue. The reconstructed ACL was a composite of these two distinctly different tissues, which had diverse histologic properties. The 3-month postoperative tendon samples appeared normal histologically, with no signs of inflammation and organized collagen bundles. Specimens taken beyond 3 months showed continued tissue maturation and crimping of collagen bundles similar to those of the native ACL. Specimens taken from the fibrous tissue had a disorganized cellular pattern and hypervascularity at 3 months postoperative and, although with time they developed increasing amounts of collagen, never resembled a normal tendon.

Beynnon and associates²³ examined the knees of a B-PT-B autograft recipient 8 months after implantation after the patient had died of causes unrelated to the knee joint. The patient’s rehabilitation program included immediate continuous passive motion (CPM) from 0° to 90°; however, he underwent a prolonged period of partial weight-bearing for 12 weeks for unknown reasons. At 4 months postoperative, he was examined clinically and had a stable reconstruction. The postmortem examination revealed 6 mm of increased AP displacement, a 13% reduction in ultimate failure loads, and a 53% reduction in energy absorbed at failure compared with the contralateral native ACL. Histologic analyses were not conducted. Interestingly, the ultimate failure load of the patient’s contralateral native ACL of 1015.6 N was well below the expected value of 1725 ± 269 N.¹³²

Delay and colleagues⁴⁰ inspected a retrieved whole B-PT-B autograft 18 months after implantation in a patient who died from a traumatic injury. Whereas the graft was mainly cellular and resembled a normal ACL morphologically, there were areas of acellularity deep within the tendinosis portion and also in the femoral intra-articular region where the autograft emerged from the tibial tunnel. The graft had been placed in an over-the-top position that the authors speculated could have caused increased stresses, resulting in inconsistent remodeling.

Falconiero and coworkers⁵³ obtained biopsies of 35 B-PT-B and 8 STG autografts in 48 patients from 3 to 120 months after implantation. The rehabilitation program the patients followed was not described. Significant differences were noted in vascularity, cellularity, and fiber pattern between grafts that were examined 3 to 6 months postoperative and those that were observed greater than 12 months after implantation. Grafts that were biopsied 7 to 12 months postoperatively were described as transitional in terms of vascular maturity. No difference was found between fiber patterns and vascularity in grafts that had been implanted 7 to 12 months before the biopsy and those that had been implanted more than 12 months before the biopsy. The authors concluded that whereas graft maturity generally occurs over a 12-month period, some autografts may reach full maturation even earlier.

Although it appears from the literature reviewed that ACL autografts undergo increasing maturation postoperatively, the small sample size of examined grafts and the host of both intrinsic and extrinsic variables affects this process and prevents definitive conclusions. Questions remain regarding whether the healing process produces a graft that has histologic, ultrastructural, and biomechanical characteristics equivalent to those of a native ACL. It is probable that the delicate nature and microgeometry of native ligament fibers is never regained, but that the ligament functions more as a checkrein to resist gross knee displacements rather than replicating normal kinematics.⁸⁹ Recent studies have demonstrated that even though B-PT-B and STG autograft reconstructions frequently restore normal AP displacements as measured with a KT-2000 arthrometer (<3 mm side-to-side difference), normal knee kinematics may not be restored, causing abnormal joint motions when the knee is subjected to weight-bearing activities.¹⁴⁸ The delay in maturation of ACL allografts and the potential effects on postoperative recovery are discussed in Chapter 5, Biology of ACL Graft Healing.

IMMEDIATE KNEE MOTION

The scientific basis supporting immediate knee motion after ACL reconstruction is well established.^{18,78,87,131,134,135,158,163} Early knee joint motion decreases pain and postoperative joint effusions, aids in the prevention of scar tissue formation and capsular contractions that can limit normal knee flexion and extension, decreases muscle disuse effects (Fig. 12-1), maintains articular cartilage nutrition, and benefits the healing ACL graft.* There is a consensus in the medical community that immobilization is detrimental to the knee joint structures and may result in a permanent limitation of knee motion, prolonged muscle atrophy, patella infera, and articular cartilage deterioration.^{33,70,95,139–141,150,170}

FIGURE 12-1 The graph demonstrates the decrease in thigh circumference on the 7th and 21st postoperative days for ACL arthroscopic versus open reconstruction. The difference between open and arthroscopic reconstructions is statistically significant (P < .05).

(Redrawn from Noyes, F. R.; Mangine, R. E.; Barber, S.: Early knee motion after open and arthroscopic anterior cruciate ligament reconstruction. Am J Sports Med 15:149–160, 1987.)

In 1983, Noyes and associates¹³³ first published recommendations for immediate knee motion after ACL B-PT-B autograft reconstruction after a successful 4-year experience with this program. Advances in knee ligament reconstructive and fixation techniques allowed the controlled movement of the knee from 0° to 90° and did not disrupt the healing graft. In 1987, these authors¹³⁴ published one of the first clinical studies that assessed joint effusion, joint motion, muscle atrophy, and the integrity of ACL grafts in a series of 18 patients who underwent ACL allograft reconstruction. Patients were randomly placed into either an immediate motion group that began CPM on the 2nd postoperative day or a delayed motion group that initiated passive and active-assisted flexion and extension exercises on the 7th postoperative day. All other parameters of the postoperative rehabilitation program were identical for the two groups, including exercises, bracing, and modalities. The results showed no differences between groups for knee joint effusion, the amount of knee flexion and extension achieved by 3 months postoperative, decrease in thigh circumference loss (Fig. 12-2), and frequency and type of pain medication used while in the hospital. There was no difference between groups in AP displacements at 12 months postoperative. Patients in whom the ACL graft had been implanted through an arthrotomy had significantly greater thigh circumference loss compared with those who had an arthroscopic-assisted technique, noted by the 7th postoperative day.

FIGURE 12-2 The graphs show the circumference results after initiating intermittent passive motion on the 2nd postoperative day (motion group) versus the 7th postoperative day (delayed motion group). The change in thigh circumference from the preoperative value is shown. There was a nearly identical decrease in the joint circumference measurements in both groups. Note the marked loss of thigh circumference in both groups.

(From Noyes, F. R.; Mangine, R. E.; Barber, S.: Early knee motion after open and arthroscopic anterior cruciate ligament reconstruction. Am J Sports Med 15:149–160, 1987.)

Richmond and colleagues in 1991¹⁵⁸ compared the effectiveness of CPM when used for the first 4 postoperative days to utilization during the first 14 days in 19 patients who underwent ACL B-PT-B autograft reconstruction. CPM was used for at least 6 hours a day. All patients also performed passive motion exercises from 0° to 90° three times daily. There were no differences between the groups at 6 weeks postoperative for swelling, thigh girth measurements, knee flexion and extension, time to regain full motion, or AP displacements. Immediate motion was concluded to be safe and not deleterious to the healing grafts in the initial postoperative period. No long-term benefit of CPM was demonstrated.

Rosen and coworkers¹⁶³ followed 75 patients who had ACL B-PT-B autogenous reconstruction and were sorted into three groups based on a knee motion and rehabilitation program. Patients who began active-assisted early knee motion exercises had similar results in regard to knee motion achieved at week 1 and at monthly intervals for the first 6 postoperative months as those who used CPM an average of 7 hours a day for 4 weeks postoperatively. There were no deleterious effects of early motion on AP displacements at the final follow-up evaluation, 6 months postoperative. The authors concluded that the CPM device offered no advantage and increased the cost of physical therapy by $1,800.

Beynnon and associates²¹ conducted in vivo measurements of the elongation of B-PT-B autografts in 20 patients immediately after implantation. In 11 knees, the length of the graft increased after 14 knee flexion-extension motion cycles, and in 9 patients, the graft length decreased. The predicted maximum increase in anterior tibial translation after 20 knee motion cycles was only 1 mm.

The current authors¹³⁵ followed 207 knees that underwent arthroscopically assisted ACL allograft reconstruction and immediate knee motion. All patients began CPM for 8 to 10 hours a day immediately after surgery. In addition, passive and active-assisted motion exercises were initiated on the 2nd postoperative day. Partial weight-bearing was allowed immediately postoperative and was gradually progressed to full by 4 to 6 weeks based on resumption of normal gait mechanics. Any patient who demonstrated a limitation of knee flexion or extension from predetermined goals was placed into a specific treatment program as early as the 7th postoperative day. The program included overpressure exercises and modalities initially (Fig. 12-3), followed, if required, by a serial extension cast program (Fig. 12-4) and in rare instances, arthroscopic débridement of scar tissues (Fig. 12-5). Of the 207 knees, 189 (91%) regained at least 0° to 135° of motion without additional intervention. Eighteen knees were placed into the phased treatment program, of which 14 regained normal motion and 2 lacked 5° of full extension. Two other patients who had not complied with the rehabilitation program had a permanent significant limitation of motion. The incidence of postoperative motion problems was related to concomitant surgical procedures: medial collateral ligament (MCL), 23%; meniscus repair, 12%; iliotibial band extra-articular procedure, 10%; and isolated ACL, 4%.

FIGURE 12-3 Knee overpressure exercises for limitation of extension (A) and flexion (B–D).

FIGURE 12-4 Extension cast application requires overpressure initially (A). The cast is worn for 48 hours (B) and is then split (C) and converted into a night splint.

FIGURE 12-5 Arthroscopic débridement of scar tissues.

In a second study from the authors’ center,¹³¹ 443 knees that had an arthroscopic-assisted ACL B-PT-B autograft reconstruction were followed to determine the effectiveness of an immediate knee motion program and treatment protocol for limitations of flexion and extension. A CPM device was not used in this group of patients. A normal range of motion (ROM) was achieved in 436 knees (98%) and a mild limitation of extension of 5° was detected in 7. Treatment intervention was required in 23 knees, all of which achieved normal knee motion. The 7 patients whose knees had mild losses of extension refused additional treatment. Only 3 knees (<1%) required an arthroscopic lysis of adhesions that, combined with an in-patient epidural program, resulted in successful resolution of the motion limitations. The incidence of knee motion problems based on concomitant surgical procedures was 22% for MCL repairs, 18% for patellar realignment procedures, 8% for meniscus repairs, and 6% for isolated ACL reconstructions.

In 2002, Henriksson and coworkers⁷⁸ compared the outcome of 5 weeks of immobilization with early knee ROM exercises in 45 patients after ACL B-PT-B autograft reconstruction. Patients in the early ROM group began motion exercises from 0° to 90° from the 7th postoperative day. The early ROM group demonstrated greater extension and flexion for the first 20 weeks. Patients in the immobilization group generally required more physical therapy visits to achieve full knee motion.

EARLY WEIGHT-BEARING

Whereas many authors have recommended early partial or full weight-bearing immediately after ACL reconstruction, few studies have examined this factor and its potential effect on both short- and long-term recovery. The effect of immediate full weight-bearing on the outcome in knees with noteworthy articular cartilage damage or those in which major concomitant operative procedures were required is unknown. In addition, gait abnormalities that may ensue from immediate full weight-bearing due to pain, knee joint effusion, and muscle weakness have not been examined. It does appear from the literature and the authors’ experience that immediate partial weight-bearing is safe and not deleterious to the healing graft.¹⁰

In 1991, Ohkoshi and associates¹⁴⁴ recommended early weight-bearing exercises after ACL reconstruction based on results of a study in which shear forces on the tibia during standing were predicted from an analytical model at various trunk and knee flexion angles (Fig. 12-6). Although weight-bearing was normally prohibited in the initial postoperative period at the time of this publication, these investigators reported that the calculated shear forces were negative in all positions, with increasing posterior drawer forces found as trunk flexion angles increased (at knee flexion angles of 30° and 60°). Co- contraction of the quadriceps and hamstrings was observed at all knee and trunk flexion angle positions, with increasing hamstrings activity measured with increasing trunk flexion angles. Exercises performed in the standing position with the knees flexed and the trunk anteriorly flexed, such as half-squatting, were believed not only to be safe but also to have the potential advantages of increasing muscle strength, endurance, and proprioception and decreasing bone atrophy.

FIGURE 12-6 Shear force (kilograms) per body weight (kilograms) while standing on both legs at various flexion angles of the knee and the trunk.

(From Ohkoshi, Y.; Yasuda, K.; Kaneda, K.; et al.: Biomechanical analysis of rehabilitation in the standing position. Am J Sports Med 19:605–611, 1991.)

Tyler and colleagues in 1998²⁰³ evaluated the effect of immediate full weight-bearing after ACL B-PT-B autograft reconstruction on ROM, patellofemoral pain, and AP displacement in a randomized, controlled trial. The results of 25 patients who were allowed immediate full weight-bearing were compared with those of 20 patients in whom no weight-bearing was allowed for 2 weeks postoperative. All patients performed immediate ROM exercises and followed an identical rehabilitation protocol (which was not detailed). There was no difference between groups for AP displacement, ROM, and vastus medialis obliquus (VMO) electromyographic activity at follow-up an average of 7 months postoperatively. Patients in the delayed weight-bearing group had a higher incidence of anterior knee pain than those in the immediate weight-bearing group (35% and 8%, respectively). Extension deficits of 5° or greater were reported in 14% of patients in the delayed weight-bearing group and in 20% in the immediate weight-bearing group. There was no correlation between loss of knee extension and anterior knee pain. Patients in the immediate weight-bearing group had significantly greater VMO electromyographic activity at 2 weeks postoperative compared with those in the delayed weight-bearing group, and the authors hypothesized that this aided in lowering the incidence of anterior knee pain by preventing the reflex inhibition of the VMO. Approximately 15% in each group had greater than 5 mm of increased AP displacement at follow-up.

POSTOPERATIVE BRACING

The effects of postoperative and functional braces on pain, return of normal knee motion, complications, sports activity performance, balance, and proprioception have been assessed by many investigators (Table 12-1). Most studies report that postoperative braces do not provide significant benefit in the amount of time postoperatively that normal range of knee motion is achieved, isokinetic lower extremity muscle strength, AP displacements, lower limb symmetry assessed by single-leg hop tests, and functional knee rating and activity level scores.* Negative effects of braces have been reported regarding quadriceps strength,¹⁶⁰ running and turning times,²¹⁸ and hamstrings isokinetic peak torque.²⁶ One study reported that a functional brace improved standing balance, but not more difficult tasks such as balance after a forward hop.²⁶ In regard to proprioception, two investigations reported significant, although small, improvements in static joint sense of 1° to 2° with brace utilization (Table 12-2).^26,219 The clinical impact of these improvements remains questionable. Risberg and coworkers¹⁵⁹ reported no improvement in the threshold to detect passive motion with application of a functional brace an average of 24 months after ACL B-PT-B autograft reconstruction. These investigators also found no impairment in proprioception in the ACL-reconstructed knees compared with those of controls.

Marx and associates¹¹⁴ published a survey of data collected from 1998 to 1999 conducted by the American Academy of Orthopaedic Surgeons. Significant variation was reported among surgeons on the use of braces in the postoperative period, with half of the surgeons indicating they prescribed a brace for the first 6 postoperative weeks and the remaining half indicating no brace utilization. Surgeons who performed more ACL reconstructions reported less brace use. Approximately 60% of all surgeons responded that they recommended a brace for sports participation.

Beynnon and colleagues¹⁸ concluded from a review of this topic that there “appears to be a consensus among investigators that, during the early phase of recovery, the use of a rehabilitation brace results in fewer problems with swelling, lower prevalence of hemarthrosis and wound drainage, and less pain compared to rehabilitation without a brace; however, at longer-term follow-up, rehabilitation bracing does not appear to have an effect on clinical outcome.” The authors recommended postoperative bracing in complex multiligament reconstructions for protection of healing grafts and unloading braces in medial and posterolateral reconstructions to allow early partial weight-bearing.

LOWER EXTREMITY MUSCLE STRENGTH ATROPHY AND RECOVERY AFTER SURGERY

Lower extremity muscle weakness represents an unresolved problem after ACL reconstruction.* Studies have reported that the magnitude of quadriceps atrophy and strength loss exceeds 20% and 30%, respectively, in the first few months postoperatively.^{54,55,68,117,160} Deficits in semitendinosus and gracilis muscle volume of 10% and 30%, respectively, have been noted after STG reconstructions.^86,211 Conflicting data have been published regarding the return of normal quadriceps and hamstrings muscle strength 6 months or longer postoperatively.

The pathophysiology of the loss of muscle size and strength after major knee surgery remains speculative. Konishi and coworkers¹⁰¹ proposed abnormal gamma loop function in the quadriceps muscles due to the lack of normal sensory function in the reconstructed ACL as a potential explanation. Other authors advocated that the loss of native ACL mechanoreceptors, which have an important role in enhancing normal activity of gamma motoneurons, was a potential cause of this problem.^{90–92,187,192,193} Young and associates²²⁴ attributed quadriceps atrophy to a reduction in muscle fiber size from histologic findings in 14 patients. Indeed, some clinical investigators have suggested that fast-twitch muscle fibers (accounting for ~60% of the rectus femoris muscle¹⁸⁷) become hypotrophic after ACL injury and reconstruction.^13,48 Residual abnormal anterior tibial displacement after reconstruction may play a role in persistent quadriceps weakness.⁸⁴ Finally, the magnitude of preoperative muscle weakness may inhibit the ability of a patient to regain normal strength and endurance after surgery.

The question of whether differential atrophy exists between type I and II muscle fibers in the quadriceps remains unclear. Haggmark and colleagues⁷¹ obtained muscle biopsies before surgery and after 5 weeks of postoperative immobilization in nine patients who underwent ACL reconstruction. A decrease in the cross-sectional area (CSA) of type 1 fibers and a reduction in the oxidative capacity of quadriceps muscle were noted, indicating a rapid fall in aerobic capacity. The authors concluded that selective atrophy of type 1 fibers occurred and that type 2 fibers were not significantly affected. In contrast, Gerber and coworkers⁶³ studied the changes in the quadriceps muscles of 41 chronic ACL-deficient knees and reported that neither histologic nor electron microscopy showed any selective loss of muscle fiber type. These authors concluded that there was no scientific rationale for selective rehabilitation of type 1 or type 2 fibers and that vigorous training of quadriceps muscles was essential after ACL injuries. Lorentzon and associates¹⁰⁹ studied the morphology of muscle types with biopsy and isokinetic testing of 18 male subjects with chronic ACL deficiency. These authors did not find any evidence of selective type 1 or type 2 fiber atrophy. In addition, there was no correlation between isokinetic test data and muscle size or morphologic changes. These authors concluded that decreases in muscle strength cannot be explained by muscle atrophy or structural change. They speculated that nonoptimal activation of muscles during voluntary contractions is the most causative mechanism of strength decrease found in patients with chronic ACL deficiency.

A comparison of high- and low-intensity electrical muscle stimulation (EMS), combined with an intensive rehabilitation program, on recovery of quadriceps strength and gait mechanics was conducted in two investigations.^189,191 A variety of ACL grafts implanted among 110 patients were included in these studies, including Achilles tendon allografts, B-PT-B allografts, STG autografts, and B-PT-B autografts. Although high-intensity EMS significantly increased rectus femoris strength recovery (compared with the opposite limb) compared with the low-intensity EMS protocol, this effect was not as pronounced for patients who underwent ACL PT autogenous reconstruction. The strength of the quadriceps correlated with flexion and extension excursions of the knee during stance. Patients who were treated with high-intensity EMS walked with more normal excursions, which the authors attributed to the gains in quadriceps strength. Low-intensity EMS, delivered via portable units for home use, was ineffective in promoting return of normal quadriceps strength.

Fitzgerald and colleagues⁵⁸ reported modest (effect size, 0.48) improvements in quadriceps isometric peak torque 16 weeks after ACL reconstruction when a “modified” EMS-training protocol was incorporated into the postoperative rehabilitation program of 21 patients. EMS was delivered with the patient positioned supine, lying passively with the knee in full extension. This protocol was modified from previously reported training methods, in which high-intensity EMS was delivered with the patient seated in an isokinetic dynamometer, the knee positioned in approximately 60° of flexion, and the patient actively contracting the quadriceps during application of the electrical current.^41,189,190 The EMS protocol was modified for patients who experienced patellofemoral pain during the training sessions. The protocol involved 10 contractions to maximum patient tolerance, conducted twice weekly for 16 weeks. At 12 weeks postoperative, the experimental group demonstrated significantly greater quadriceps isometric peak torque compared with that of a matched group of patients who did not receive EMS training (75.9 ± 16.8 and 67.0 ± 19.9, respectively; P < .05). However, no significant differences were measured at 16 weeks postoperative between these groups. A greater proportion of the EMS-trained subjects initiated agility training at 16 weeks compared with the control group (62% and 32%, respectively; P < .05). Still, the authors concluded that the original high-intensity EMS-training protocol was preferred when possible based on patient tolerance and equipment availability.

Investigators and clinicians have recently questioned whether rehabilitation programs should focus more efforts on eccentric training to reduce the early postoperative loss of muscle CSA, volume, and strength following ACL reconstruction.64–66,¹⁰⁵ In an exhaustive review of the literature on the effectiveness of concentric, eccentric, or combined concentric-eccentric training in normal, uninjured subjects, Wernbom and coworkers²⁰⁷ concluded that there was no evidence to support one mode of training over another with regard to achieving superior muscle hypertrophy. These authors assessed published training programs according to frequency, intensity, and duration of work on CSA and volume of the quadriceps and elbow flexor muscle groups. The observation was made that training protocols to induce muscle hypertrophy may differ from those designed to produce maximum strength.

Eccentric training is believed by some authors to be superior to concentric training owing to its potential to overload the muscle and produce greater increases in muscle size and strength.^64,66 There are concerns with high-force eccentric training after ACL reconstruction, including the potential for inducing damage to the muscle and healing graft. A repeated, gradual, and progressive exposure to this type of training was successfully accomplished by Gerber and associates⁶⁴ who conducted eccentric training in patients who had either ACL B-PT-B autograft (N = 20) or STG autograft (N = 20) reconstruction. The patients were randomly assigned to either a 12-week program of eccentric exercises or a standard rehabilitation program. All patients followed a similar protocol for the first 3 postoperative weeks that emphasized regaining full knee motion and basic quadriceps function. Then, patients in the eccentric exercise group initiated progressive exercises using a recumbent eccentric ergometer. The duration and intensity of the negative-work training gradually increased throughout the 12-week period. The patients underwent magnetic resonance imaging (MRI) 3 and 15 weeks postoperative.

The volume and CSA measurements of the quadriceps in the reconstructed knees improved significantly (P < .001) in both groups (Fig. 12-7) between the pre- and the post-training time periods. The increases in these measurements were significantly greater (P < .001), by more than twofold, in the eccentrically trained patients than in the standard rehabilitation group. In the eccentric group, quadriceps volume increased 23.1 ± 12.9% and peak CSA increased 24.2 ± 12.6%. In comparison, in the standard rehabilitation group, quadriceps volume increased 8.8 ± 9.3% and peak CSA increased 9.3 ± 9.4%. The increase was not related to the type of graft used for the ACL reconstruction. Significant increases in these measurements were also found in the quadriceps in the noninvolved side.

FIGURE 12-7 Changes in the volumes of the quadriceps and gluteus maximus muscles in the involved and uninvolved lower extremities during the 12-week training period after treatment with a semitendinosus-gracilis or bone–patellar tendon–bone graft. The asterisks indicate a significant difference in muscle-volume improvement between the eccentric and the standard rehabilitation groups (P < .005).

(From Gerber, J. P.; Marcus, R. L.; Dibble, L. E.; et al.: Effects of early progressive eccentric exercise on muscle structure after anterior cruciate ligament reconstruction. J Bone Joint Surg Am 89:559–570, 2007.)

There were no significant differences between the training groups in the improvement in volume and CSA of the hamstring muscles (Fig. 12-8). When analyzed according to graft type, significant improvements were noted in these measurements in the B-PT-B autograft patients (P ≤ .006), but not in the STG autograft patients. A reduction in gracilis muscle volume of nearly 20% was found 3 weeks postoperative in the STG group, which increased to a deficit of 35% by 15 weeks postoperative. The authors concluded that the gradual progressive eccentric exercise program safely and effectively improved quadriceps structure in comparison with a standard rehabilitation program. Whether the short-term results of this investigation will demonstrate longer-term benefits are unknown.

FIGURE 12-8 Changes in the volumes of the hamstrings and gracilis muscle in the involved and uninvolved lower extremities during the 12-week training period after treatment with a semitendinosus-gracilis or bone–patellar tendon–bone graft. There was no significant difference in muscle-volume changes between the eccentric and the standard rehabilitation groups.

(From Gerber, J. P.; Marcus, R. L.; Dibble, L. E.; et al.: Effects of early progressive eccentric exercise on muscle structure after anterior cruciate ligament reconstruction. J Bone Joint Surg Am 89:559–570, 2007.)

Shaw and colleagues¹⁷⁵ conducted a prospective, blinded, randomized trial to determine the effectiveness of instituting quadriceps exercises immediately after ACL reconstruction on a number of parameters. A total of 91 patients who underwent either a B-PT-B or an STG autograft reconstruction were randomized into either the quadriceps exercise–training group or the no-quadriceps exercise–training group. The quadriceps training included straight leg raises and isometric quadriceps contractions performed three times daily for the first 2 postoperative weeks. After 2 weeks, all patients were entered into a similar rehabilitation program. At 6 months postoperative, no difference was found between groups for isokinetic quadriceps strength or lower limb symmetry on functional hop testing.

It is important to note that several authors have documented quadriceps muscle inhibition from an experimentally induced knee joint effusion.^{147,194,199,200} This problem has been documented in separate studies during walking, jogging, and landing from a jump. Therefore, it is imperative that knee joint effusion be avoided and, if noted, treated immediately to lessen its deleterious impact on quadriceps function.

OPEN VERSUS CLOSED KINETIC CHAIN EXERCISES: BIOMECHANICAL, IN VIVO, AND CLINICAL STUDIES

The process of graft maturation and healing is assumed to be influenced by strains and forces applied to the ACL during weight-bearing and exercises. Whereas a general consensus exists that some strain is necessary to promote the process of ligamentization,²⁰ questions remain regarding the amount of load that is safe and the amount that may produce graft elongation. In addition, ligamentization (the return to normal native ACL characteristics) has never actually been shown to occur with graft remodeling and healing. Since the mid 1980s, investigations have attempted to measure, through either analytical or direct means, force and strain incurred on the ACL during common exercises used in rehabilitation. These exercises are typically referred to as either closed kinetic chain (CKC) or open kinetic chain (OKC). During CKC exercises, the foot is fixed to a platform or surface, the motion at the knee joint is accompanied by predictable motions at the hip and ankle joints, and the entire limb is loaded such as during a leg press or squat. In OKC exercises, the foot is mobile and not fixed to a surface, and motion at the knee joint occurs independent of motion at the hip and ankle joints. Common examples include the leg extension and hamstring curl exercises.

Cadaver Studies

Grood and coworkers⁶⁹ biomechanically examined the knee extension exercise in cadavers to determine the conditions in which the ACL was loaded, the quadriceps force that developed during knee flexion-extension, and the forces that were incurred on the extensor mechanism as the knee was extended. Both ACL-intact and ACL-sectioned conditions were created, and the OKC knee extension exercise was simulated both without resistance and with 31 N (7 lb) of resistance applied with an ankle weight at the foot. Although the quadriceps force required to extend the knee remained at a constant value of 177 N between 50° and 15°, it rose rapidly to 350 N at 0° (Fig. 12-9). The addition of resistance doubled the quadriceps force that was required to extend the knee, resulting in large muscle forces on the patellofemoral joint. There was no change in the quadriceps force required to extend the knee when the ACL was removed. However, loss of the ACL resulted in an increase in anterior tibial displacement from 0° to 30° of extension without resistance. The addition of 31 N of resistance resulted in further increases in anterior tibial displacement throughout the range of extension, with a mean increase of 3.8 mm at 15° (Fig. 12-10). The authors concluded that the range of 0° to 30° of flexion should be avoided during rehabilitation of ACL-deficient or ACL-reconstructed knees and in patients with patellofemoral symptoms.

FIGURE 12-9 The ratio of quadriceps force to leg weight is shown for five specimens as a function of flexion angle. The lower group of curves shows the force measured when extending the leg against its own weight. The upper group of curves shows the forces measured when 31 N was placed at the foot.

(From Grood, E. S.; Suntay, W. J.; Noyes, F. R.; Butler, D. L.: Biomechanics of the knee-extension exercise. Effect of cutting the anterior cruciate ligament. J Bone Joint Surg Am 66:725–734, 1984.)

FIGURE 12-10 The increased anterior tibial displacement, relative to the intact knee, that resulted when the anterior cruciate ligament was removed and 31 N was placed at the foot is shown for five specimens. There is an increased displacement in the range of 30° to full extension compared with the intact knee with weights at the foot.

(From Grood, E. S.; Suntay, W. J.; Noyes, F. R.; Butler, D. L.: Biomechanics of the knee-extension exercise. Effect of cutting the anterior cruciate ligament. J Bone Joint Surg Am 66:725–734, 1984.)

Critical Points OPEN VERSUS CLOSED KINETIC CHAIN EXERCISES

Cadaver Studies

• Quadriceps force required to extend the knee remains at a constant value of 177 N between 50° and 15°, then rises rapidly to 350 N at 0°.

• Addition of 7-pound weight at the ankle doubles the quadriceps force required to extend the knee.

• Avoid 0° to 30° of flexion open kinetic chain (OKC) during rehabilitation of ACL-deficient or ACL-reconstructed knees or in patients with patellofemoral symptoms.

• Isolated hamstring contractions decrease ACL strain relative to normal passive strain from 0° to 120°.

• Isolated quadriceps isometric and isotonic contractions increase ACL strain from 0° to 45°.

• Isometric co-contractions of quadriceps and hamstrings increase ACL strain from 0° to 30°.

• During the squat exercise, the addition of a hamstrings load causes significant decrease in ACL graft load, most evident between 15° and 45°.

Effect of Open and Closed Kinetic Chain Exercises on Anterior Tibial Displacement in Intact and ACL-deficient Knees

• Overall, significantly greater anterior tibial displacements found in ACL-deficient knees during OKC activities than during closed kinetic chain (CKC) exercises.

• CKC squat exercises produce less anterior tibial translation than knee extension in ACL-deficient knees.

• Anterior tibial translation is significantly less during weight-bearing conditions than that measured during Lachman test and non–weight-bearing conditions.

• CKC exercises cause increased joint compression forces and co-activation of the gastrocnemius and quadriceps muscles and result in decreased anterior tibial displacements in ACL-deficient knees.

Calculated ACL Forces, Tibiofemoral Compressive Forces, and Muscle Forces in Human Subjects

• Three-dimensional models used to predict internal muscle forces, tibiofemoral compressive forces, tension in the ACL and posterior cruciate ligament (PCL), and patellofemoral compressive forces during OKC and CKC exercises.

• Greater tibiofemoral compressive forces occur during CKC exercises than during OKC knee extension and are greatest when the knee is fully flexed.

• CKC exercises produce greater co-contraction between quadriceps and hamstrings compared with knee extension; magnitude is dependent on trunk position and knee flexion angle.

• CKC squat produces double the amount of hamstring activity as the leg press and knee extension exercises.

Measurement of In Vivo ACL Strain in ACL-deficient and Intact Knees during Common Open and Closed Kinetic Chain Exercises

• Beynnon et al.^{17,22,59,61,76}: series of studies, arthroscopically implanted a Hall effect transducer into the anteromedial fibers of the normal ACL

• Magnitude of ACL strain lowest in isometric contractions of hamstring muscles; contractions quadriceps and hamstrings 30°, 60° and 90°; isometric quadriceps contractions 60° and 90°; passive knee flexion and extension.

• Highest ACL strain in isometric quadriceps contractions 15° flexion with 30 Nm of extension torque, squatting without and with resistance, active knee flexion and extension with 45 N weight boot.

Muscle Recruitment Patterns during Common Open and Closed Kinetic Chain Exercises

• Onset activity for all quadriceps muscles simultaneous during CKC leg press.

• During OKC knee extension, onset of vastus lateralis, vastus medialis longus, and rectus femoris occurs before vastus medialis obliquus.

• Peak levels quadriceps activation 201% and 207% of a maximum voluntary isometric contraction (MVIC) for single-leg squats and step-ups, respectively.

• Hamstring activity (biceps femoris muscle only) approximately 20% to 40% MVIC, could have been influenced by trunk flexion angle.

Comparative Clinical Studies

Bynum et al.³¹: Prospective, Randomized Study: OKC or CKC Program after ACL B-PT-B Autograft

• Significantly lower values mean anterior tibial translation and patellofemoral pain severe enough to restrict activities in CKC group.

• Significantly higher values patient rating of the end result of the operation in the CKC group.

• Authors recommended the exclusive use of CKC protocols after ACL reconstruction.

Mikkelsen et al.¹¹⁹: Prospective, Randomized Study: CKC or Combined CKC-OKC Program (Began OKC Exercises Week 6) after ACL B-PT-B Autograft

• Higher percentage of patients in combined OKC-CKC group returned to pre-injury sports levels than those in the CKC group.

• Authors recommended combined protocol under carefully supervised conditions.

Three investigations same laboratory: No benefit from OKC short-term postoperative period.

Beynnon et al.²⁴: Prospective, Randomized Double-Blind Study of Two Rehabilitation Programs (“Accelerated” Weight-Bearing, OKC, and Traditional) after ACL B-PT-B Autograft

• No differences between groups for anteroposterior (AP) displacement, International Knee Documentation Committee grading, activity levels, subjective function and patient satisfaction, or single-leg hop tests.

• Authors concluded that rehabilitation programs that allow early weight-bearing and early use of significant quadriceps contractions have the same effect as programs that delay weight-bearing and use of the quadriceps muscles.

Renstrom and associates¹⁵⁷ measured the strain of cadaver ACL specimens under simulated isometric contractions of various muscle groups. The authors reported that isolated hamstring contractions decreased ACL strain relative to normal passive strain at all knee flexion angles tested (0°–120°). Isolated quadriceps isometric and isotonic contractions increased ACL strain from 0° to 45° of flexion (Fig. 12-11); the greatest magnitude (5% above the normal passive strain) was measured at full knee extension. Isometric co-contractions of the quadriceps and hamstrings also increased ACL strain from 0° to 30°, with the maximum increase of 5% measured at 0° and 15° of flexion. The authors concluded that isolated quadriceps isometric and isotonic exercises, and isometric co-contraction exercises, could produce potentially harmful forces on healing ACL grafts.

FIGURE 12-11 Mean passive normal, simulated quadriceps isometric, and simulated quadriceps isotonic strain patterns.

(From Renstrom, P.; Arms, S. W.; Stanwyck, T. S.; et al.: Strain within the anterior cruciate ligament during hamstring and quadriceps activity. Am J Sports Med 14:83–87, 1986.)

More and colleagues¹²² used a cadaver model that incorporated quadriceps and hamstrings muscle loads to examine knee kinematics and ACL loads during the squat exercise. In the intact knees, the addition of a hamstrings load resulted in a significant reduction of anterior tibial translation and internal tibial rotation during flexion. After the ACL was sectioned, the amount of anterior tibial translation increased significantly between 15° and 45° during the squat compared with the intact knee. After ACL reconstruction, maximal graft tension was measured at full extension. During the squat exercise, the addition of a hamstrings load caused a significant decrease in graft load that was most evident between 15° and 45°. The authors concluded that the squat exercise may be safe in the early postoperative period after ACL reconstruction.

Calculated ACL Forces, Tibiofemoral Compressive Forces, and Muscle Forces in Human Subjects

Several investigators used two-dimensional mathematical models from in-vivo experimental measures to calculate muscle forces, tibiofemoral shear, and compressive forces during isometric and isokinetic exercises.* The analytical models did not allow for the authors to determine the magnitude of the ligament forces incurred during various exercises.

Yasuda and Sasaki²²² calculated anterior and posterior drawer forces exerted on the tibia in 20 healthy adult males during isometric contractions of the quadriceps and hamstrings at knee flexion angles ranging from 5° to 90° (Fig. 12-12). During the quadriceps isometric contractions, maximum anterior shear forces were measured at 5° of knee flexion. As the angle of knee flexion increased, the calculated anterior shear force decreased. The mean knee flexion angle at which the anterior drawer force changed to a posterior drawer force during this exercise was 45.3° ± 12.5°. During isometric hamstring contractions, posterior shear forces were measured at all knee flexion angles. The authors concluded that early postoperative rehabilitation should include quadriceps isometrics exercises with knee flexion greater than 70° and hamstrings isometric exercises at all flexion angles. The question was raised of the efficacy of quadriceps isometric training at high knee flexion angles for adequate muscle training.

FIGURE 12-12 Shear force in isometric contraction of the quadriceps (A) or hamstrings (B) at various flexion angles of the knee. The abscissa shows the flexion angle of the knee, and the ordinate shows the coefficient of a linear function that gives the shear force in proportion to tension of the quadriceps (A) or the hamstrings (B). Plus and minus values of the coefficient mean the anterior and posterior drawer force to the tibia. Each polygonal line shows a different subject. (A) The anterior drawer force is given to the tibia in the area of small flexion angle of the knee, and the posterior drawer force is given in the area of large flexion angle. (B) The posterior drawer force is given to the tibia at every flexion angle of the knee.

(A and B, From Yasuda, K.; Sasaki, T.: Exercise after anterior cruciate ligament reconstruction. The forces exerted on the tibia by the separate isometric contractions of the quadriceps or the hamstrings. Clin Orthop Relat Res 220:275–283, 1987.)

Lutz and colleagues¹¹⁰ analyzed forces at the tibiofemoral joint during OKC and CKC exercises in five healthy subjects. A two-dimensional model was used to calculate tibiofemoral shear and compression forces during maximal isometric contractions at 30°, 60°, and 90° of knee flexion. The OKC knee extension contraction produced the greatest amount of anterior shear force (Fig. 12-13; 285 ± 120 N at 30° of flexion). CKC exercises produced significantly less anterior shear force at all flexion angles. In addition, CKC exercises produced significantly greater compressive forces and muscular co-contraction at the same knee flexion angles at which the OKC exercises produced maximum shear forces and minimum muscular co-contraction. The authors recommended CKC exercises after ACL injury or reconstruction.

FIGURE 12-13 Graph of the average tibiofemoral shear forces observed during the closed kinetic chain (CKC) leg press, the open kinetic chain extension exercise (OKC_ex), and the open kinetic chain flexion exercise (OKC_fl).

(From Lutz, G. E.; Palmitier, R. A.; An, K. N.; Chao, E. Y.: Comparison of tibiofemoral joint forces during open-kinetic-chain and closed-kinetic-chain exercises. J Bone Joint Surg Am 75:732–739, 1993.)

Ohkoshi and associates¹⁴⁴ used a two-dimensional model derived from radiographs and electromyographic analyses to predict shear forces on the tibia during standing at various trunk and knee flexion angles in 21 normal male subjects. Co-contraction of the quadriceps and hamstrings was observed at all knee and trunk flexion angle positions. Hamstrings activity increased with increasing trunk flexion angles. The calculated shear forces were negative in all positions, with increasing posterior drawer forces found as trunk flexion angles increased (at knee flexion angles of 30° and 60°). The authors concluded that exercises done in the standing position with the knees flexed and the trunk anteriorly flexed (such as half-squatting) could be performed safely in the early stages after ACL reconstruction.

Other investigators used three-dimensional models to predict internal muscle forces, tibiofemoral compressive forces, tension in the ACL and posterior cruciate ligament, and patellofemoral compressive forces during OKC and CKC exercises.^{51,99,201,210} Kaufman and coworkers⁹⁹ reported that an isokinetic OKC knee extension exercise produced mean tibiofemoral compressive forces of 4.0 ± 0.7 times BW at 60°/sec and 3.8 ± 0.9 BW at 180°/sec; however, these were calculated at 55° of knee flexion. Anterior shear forces existed between 40° and full extension, potentially loading the ACL.

Escamilla and coworkers⁵¹ and Wilk and associates²¹⁰ reported that greater tibiofemoral compressive forces occurred during CKC exercises than during OKC knee extension (Fig. 12-14) and that these compressive forces were greatest when the knee was fully flexed. The calculated joint compression forces were dependent on the position of the trunk relative to the knee and ankle joints, with greater forces generated when the body was positioned directly over the knee. Peak ACL tensile forces occurred only during OKC exercises near full extension and were calculated as 0.20 × BW. CKC exercises produced greater co-contraction between the quadriceps and the hamstrings compared with knee extension; however, the magnitude was dependent on the trunk position and knee flexion angle. In addition, the squat produced approximately double the amount of hamstring activity as the leg press and knee extension exercises. Knee extension produced greater quadriceps activity, leading the investigators to conclude that the ACL could be loaded from 0° to 60° during this exercise.

FIGURE 12-14 Mean and standard deviation of forces during squat (triangle), leg press (square), and knee extension (circle). Tibiofemoral compressive force, posterior cruciate ligament (PCL; +), anterior cruciate ligament (ACL; –) tensile force, and patellofemoral compressive force are shown.

(From Escamilla, R. F.; Fleisig, G. S.; Zheng, N.; et al.: Biomechanics of the knee during closed kinetic chain and open kinetic chain exercises. Med Sci Sports Exerc 30:556–569, 1998.)

Toutoungi and colleagues²⁰¹ determined cruciate ligament forces from analytical modeling in normal knees during OKC isokinetic and isometric exercises and during CKC double-leg and single-leg squatting. During isokinetic and isometric extension, peak ACL forces occurred from 35° to 40° of nearly 400 N, or 0.55 × BW. However, during isokinetic extension, ACL forces decreased significantly with increasing dynamometer speed, from 349 N at 60°/sec to 254 N at 180°/sec. Small forces were incurred on the ACL during squats at knee flexion angles less than 50°. The authors concluded that isokinetic flexion and squats were safe to perform in the early postoperative period after ACL reconstruction, but that isokinetic knee extension should be avoided until graft healing is well advanced.

Shields and coworkers¹⁸³ examined the single-leg squat in normal subjects to determine the effect of resistance to both flexion and extension and knee flexion angle on lower extremity muscle activity. Resistance to flexion and extension was set as a percentage of BW, being either 0%, 4%, or 8%. The results revealed that co-contraction of the quadriceps and hamstrings occurred throughout the squatting exercise (0°–40°). Although the quadriceps had greater activity than the hamstrings at all levels of resistance, the quadriceps-to-hamstrings ratio decreased with higher levels of resistance. Biceps femoris activity increased during knee flexion with resistance from approximately 12% of that of a maximum voluntary isometric contraction (MVIC) during low resistance (0% BW) to 27% of that of a MVIC during high resistance (8% BW). The authors suggested that this CKC exercise done under controlled conditions with resistance to both flexion and extension was effective in increasing the dynamic control of the knee joint.

Muscle Recruitment Patterns during Common OKC and CKC Exercises

Stensdotter and coworkers¹⁹⁷ studied onset time and amplitude of the quadriceps muscles during a knee extension and simulated leg press isometric contraction in 10 healthy subjects. The onset of activity for all of the quadriceps muscles was simultaneous during the CKC activity (Fig. 12-15). During the OKC task, the onset of vastus lateralis, vastus medialis longus, and rectus femoris occurred before that of VMO. The mean amplitude for rectus femoris was significantly greater during the OKC exercise, whereas the mean amplitude for VMO was significantly larger during the CKC exercise.

FIGURE 12-15 Group mean values (standard error of the mean [SEM]) for onsets of activity relative to onset of force increase. There was no difference in electromyographic onset time between muscle portions relative to force in CKC. In OKC, the latency between onset of activity and onset of force was shorter for the vastus medialis obliquus than for all other muscle portions, whereas the latency between onset of activity in rectus femoris relative to onset of force was longer than all others.

(From Stensdotter, A. K.; Hodges, P. W.; Mellor, R.; et al.: Quadriceps activation in closed and in open kinetic chain exercise. Med Sci Sports Exerc 35:2043–2047, 2003.)

Beutler and associates¹⁶ measured quadriceps and hamstring activation in healthy subjects during two CKC activities. The peak levels of quadriceps activation were 201% and 207% of a MVIC for single-leg squats and step-ups, respectively. Hamstring activity (biceps femoris muscle only) was approximately 20% to 40% MVIC, which could have been influenced by trunk flexion angle, which was not controlled for in this study. The authors concluded that both of these CKC exercises were effective for males and females for achieving maximal quadriceps contraction for strength training or rehabilitation.

Salem and colleagues¹⁶⁸ evaluated the kinematics and kinetics of the ankle, knee, and hip in eight ACL-reconstructed and contralateral knees during a two-legged squat. The patients were tested a mean of 30 ± 12 weeks postoperative; all had implemented the two-legged squat into their rehabilitation at least 6 weeks prior to testing. Two distinctly different strategies were noted for generating the joint torques required to perform the exercise. In the noninvolved limb, equal distribution of hip and knee muscular effort was noted. However, in the reconstructed limb, patients used a more hip-dominant strategy, thereby reducing the knee extension peak torque effort. The authors cautioned that patients may use substitution methods during bilateral CKC activities by either shifting the effort from the reconstructed limb to the contralateral limb or adopting a hip-dominant strategy, limiting the potential effectiveness of these exercises.

PATELLOFEMORAL JOINT CONSIDERATIONS

As previously discussed, the quadriceps force that develops during knee flexion-extension and the forces incurred on the extensor mechanism as the knee extends were experimentally measured in the authors’ laboratory.⁶⁹ The investigation demonstrated that the quadriceps force required to extend the knee remained at a constant value of 177 N between 50° and 15° of knee flexion, and then rose rapidly to 350 N at 0°. The addition of a small amount of resistance (31 N) doubled the quadriceps force that was required to extend the knee, resulting in large forces on the patellofemoral joint. This was the first investigation to recommend avoidance of quadriceps exercises in the range of 0° to 30° of flexion during rehabilitation of the ACL-deficient or ACL-reconstructed knee and in knees with patellofemoral symptoms.

Doucette and Child⁴³ used computed tomography to measure the patellar congruence angle in patients with symptomatic lateral patellar compression syndrome under three conditions: with the lower extremity relaxed, holding an OKC (knee extension) position, and holding a CKC exercise position. Measurements were made in 10° increments from 0° to 40° of flexion. There were significant differences at 0°, 10°, and 20° of flexion between the OKC position and both the CKC and the relaxed conditions. At each flexion angle, significantly greater lateral patellar tracking occurred during the OKC exercise. During all three conditions, patellar congruence progressively improved from 0° to 40° of flexion. The authors concluded that a quadriceps contraction has less influence on patellar tracking at 30° of flexion than at 0° of flexion owing to the increased stability of the patella as it moves into the intercondylar groove as the knee is flexed. In low, functional knee ROMs, CKC exercises were recommended owing to the improved patellar positioning and decreased joint irritation in symptomatic patellofemoral patients.

Steinkamp and associates¹⁹⁶ calculated knee moments, patellofemoral joint reaction forces, and patellofemoral joint stresses in 20 normal subjects at four knee flexion angles (0°, 30°, 60°, and 90°) during the leg press and knee extension exercises. All three parameters were significantly greater at 0° and 30° in the leg extension exercises than during the leg press exercise (Fig. 12-16). The opposite was true at the high knee flexion angles, for which all parameters were significantly greater in the leg press exercise. The authors concluded that the leg press placed minimal stress on the patellofemoral joint in the functional ROM (low knee flexion angles) and noted empirically that patients with patellofemoral disorders frequently complained of pain with the knee extension exercise. During the leg press exercise, compressive forces were higher but distributed over a larger contact area, whereas during the leg extension exercise, compressive forces were lower but concentrated over a smaller contact area.

FIGURE 12-16 Mean ± standard deviation of patellofemoral joint reaction force at four flexion angles.

(From Steinkamp, L. A.; Dillingham, M. F.; Markel, M. D.; et al.: Biomechanical considerations in patellofemoral joint rehabilitation. Am J Sports Med 21:438–444, 1993.)

Witvrouw and colleagues^212,213 conducted investigations on 51 patients with isolated patellofemoral pain who were randomly assigned to either an OKC or a CKC 5-week exercise program. The OKC exercises consisted of maximal static quadriceps contractions with the knee in full extension, straight leg raises in the supine position, leg adduction exercises in the lateral decubitus position, and short arc movements from 10° of flexion to terminal extension. The CKC exercises included seated leg press, double or single one third knee bend, stationary biking, rowing machine, step-up and step-down exercises, and jumping on a minitrampoline. The patients were evaluated upon completion of the program and then 5 years later. There were no significant differences between the two treatment groups for the majority of parameters evaluated. At the 5-year evaluation, 92% of the OKC group was participating in sports compared with 60% of those in the CKC group. The OKC group demonstrated, on visual analog scales, less swelling, pain on descending stairs, and pain at night compared with the CKC group. Owing to the overall lack of differences between the groups in the majority of parameters studied, the authors recommended a combination of both OKC and CKC exercises in patients with patellofemoral symptoms.

Dye⁴⁵ reviewed the role of loading in patellofemoral pain, including the amount sustained by a single blow (such as during a fall onto the pavement) and the amount sustained with repeated smaller loads that disrupt normal tissue homeostasis (such as climbing up or down stairs, kneeling, squatting, or jumping). The stress on the patellofemoral joint depends not only on the load applied but also on the surface areas of the patella and femur that may be in contact at any given point in time. Estimated loads on the patellofemoral joint with weight-bearing activities ranged from 3.3 times BW with stair climbing to 7.6 times BW with squatting to up to 20 times BW with jumping.^156,186 The ability of the entire knee joint to absorb and distribute these forces is dependent on what Dye⁴⁵ termed the envelope of function, or “that range of loading applied across the joint that is compatible with and probably inductive of maintenance of tissue homeostasis.” The envelope includes three zones. The first is the zone of subphysiologic underload, or diminished loads (e.g., prolonged bedrest), that cause muscle atrophy or calcium loss. The second zone, that of homeostatic loading, represents the range of acceptable loading in which tissue homeostasis is maintained. This zone is highly variable between individuals, because knee joints accept loads that range from less than 1 to nearly 8 times BW.⁴⁵ The third zone, supraphysiologic overload, represents loads that exceed the knee joint’s ability to accept and distribute forces in a manner that maintains homeostasis. The author expressed that patellofemoral pain could frequently be diminished by simply lowering the forces to the patient’s asymptomatic envelope of function.

The authors recommend avoidance of exercises that place large forces on the patellofemoral joint after ACL reconstruction, especially the leg extension exercise in the range of 30° to 0° of knee flexion. Patients who develop patellofemoral symptoms postoperatively must be carefully followed and the rehabilitation program altered as required to avoid activities, such as deep squatting and jumping, that place high forces on the patella.

ALTERATIONS IN GAIT, NEUROMUSCULAR FUNCTION, AND PROPRIOCEPTION AFTER ACL RECONSTRUCTION

Chronic ACL deficiency produces marked alterations in gait during a variety of activities.^{4,5,15,137,149,208} During level walking, ACL-deficient subjects demonstrate significantly decreased external knee flexion moments and increased external knee extension moments compared with healthy control subjects.²⁰⁸ The resultant quadriceps avoidance gait pattern has been identified in these knees in several investigations: in 16 of 32 knees that also had varus malalignment in the authors’ laboratory,¹³⁷ in 7 of 8 subjects who were greater than 7 years postinjury by Wexler and coworkers,²⁰⁸ and in 12 of 16 (75%) subjects by Berchuck and associates.¹⁵ Wexler and coworkers²⁰⁸ found that changes in sagittal plane knee moments were more pronounced as the amount of time after the injury increased. Berchuck and associates¹⁵ reported that gait adaptations were present in both the injured and the contralateral limbs owing to the symmetrical function required for weight-bearing activities. Patel and colleagues¹⁴⁹ found that patients with ACL deficiency had a significantly reduced peak external flexion moment during jogging and stair climbing that correlated with significantly reduced quadriceps strength. Andriacchi and coworkers⁵ used a finite-element model from three-dimensional cartilage volumes created from MRI to predict progression of osteoarthritis (OA) in normal and ACL-deficient knees. The model predicted a more rapid rate of cartilage thinning in the ACL-deficient knees, especially in the medial tibiofemoral compartment. The investigators concluded that this was due to a shift in the normal load-bearing regions of the knee joint during weight-bearing activities and stressed the importance of restoring proper gait mechanics after ACL reconstruction.

Several investigations have documented altered gait biomechanics, neuromuscular function, and proprioception many months or years after ACL reconstruction.^{27,30,42,50,106,214} Unfortunately, the majority of these investigations did not provide detailed information regarding the postoperative rehabilitation program. Timoney and associates¹⁹⁸ were among the first investigators to document altered gait kinematics in a study conducted on 10 male patients an average of 10 months (range, 9–12 mo) after ACL B-PT-B autograft reconstruction. Compared with a control group, the patients had a significantly lower mean external knee flexion moment at midstance (3.74% and 2.02%, respectively) and a significantly lower mean heel-strike transient value (66.6 and 44.3 BW/sec, respectively). The patients’ reconstructed limb had a significantly lower mean midstance external knee flexion moment than the uninvolved limb (2.02% and 3.10%, respectively). The patients did not demonstrate a quadriceps-avoidance gait, because a net external flexion moment was present throughout most of the stance phase.

Devita and colleagues⁴² conducted gait analysis testing in eight patients who had a B-PT-B autograft ACL reconstruction 3 weeks and 6 months after surgery and an “accelerated” rehabilitation program. The data were compared with those collected from 22 healthy subjects. Although the ACL-reconstructed patients walked with normal kinematic patterns 6 months postoperatively, they demonstrated altered joint torque and power patterns at the hip and knee. The hip extensors provided more vertical support and forward progression during the first half of stance compared with the contribution measured in the normal subjects. There was also a decrease in the magnitude of extensor torque at the knee in early stance compared with that in the healthy subjects.

Kowalk and coworkers¹⁰² conducted gait analyses on 7 patients who had a B-PT-B autograft reconstruction (mean, 6 mo postoperative; range, 3.2–11.3 mo) and on 10 healthy subjects. The analysis consisted of ascending three steps that were attached to two force plates. The authors reported statistically significant reductions for peak moment, power, and work in the reconstructed knees along with significant increases in excursion, moment, and power at the contralateral ankle joint. The patients compensated during the stair ascent task by generating increased power at the contralateral ankle.

Ernst and associates⁵⁰ measured lower extremity kinematics in 20 patients who underwent a B-PT-B autograft reconstruction (mean, 9.8 mo postoperative; range, 8–15 mo) and in 20 matched normal subjects. The subjects performed a single-leg vertical jump and a lateral step-up. The knee extension moment of the ACL-reconstructed subjects was less than that of their contralateral lower extremity and those of the lower extremities of the controls during take-off and landing on the vertical jump and the lateral step-up exercise. The authors suggested that this finding was related to either weakness of the quadriceps femoris muscle or some alteration of the neuromuscular system in which hip or ankle extensors were recruited. Patients may compensate during these activities and not adequately recruit the quadriceps musculature, thereby reducing the potential effectiveness of the exercises.

Bush-Joseph and colleagues³⁰ measured knee kinematics and kinetics in 22 patients who had ACL B-PT-B autograft reconstruction 22 ± 12 months postoperative. The data were compared with a control group of 22 subjects. All patients were satisfied with the results of the reconstruction and all had negative Lachman and pivot shift tests. Quadriceps and hamstrings isokinetic peak torques were similar to those of the control group and the contralateral limbs. Peak external moments were similar between the ACL and the control subjects for light activities such as walking and stair climbing. However, a decrease was noted in the peak external flexion moment (net quadriceps moment) in the ACL-reconstructed knees compared with that of controls during jogging and a jog-and-cut maneuver (13.3 ± 3.9% BW × height [Ht] and 16.1 ± 4.2% BW × Ht, respectively; P = .024). The decrease in the peak external flexion moment during jogging significantly correlated with quadriceps muscle strength at 60°/sec (R² = 0.464, P < .001) and at 180°/sec and 240°/sec (P < .02). This correlation between quadriceps muscle strength and decreased external flexion moment was not found during the jog-and-cut task. Subjects in the ACL-reconstructed group with the weakest quadriceps muscles had the greatest reductions in the peak flexion moment during jogging. The investigators concluded that functional adaptations during more demanding athletic activities were present in patients who had well-functioning ACL reconstructions and minimal decreases in quadriceps strength compared with controls. The authors speculated that further improvements in lower extremity strength may improve these adaptations and reinforced the value of ensuring comprehensive quadriceps training after ACL reconstruction.

Wojtys and Huston²¹⁴ measured lower extremity muscle strength, endurance, reaction time, and time to reach peak torque in a group of 25 patients who received a B-PT-B autograft reconstruction at 6, 12, and 18 months postoperative. A group of 40 healthy subjects served as controls. At 18 months postoperative, 88% of the reconstructed knees were within 3 mm of increased AP displacement compared with the opposite limb and 80% of the patients believed they had regained their preinjury sports activity level. However, quadriceps peak torque of the reconstructed limbs was equal to that of the opposite limbs in only 72% (Fig. 12-17). Quadriceps endurance failed to reach the level of the opposite limb throughout the study period. Time to reach peak torque of both the quadriceps (Fig. 12-18) and the hamstrings (Fig. 12-19) was significantly slower than the uninvolved limb at all test sessions.

FIGURE 12-17 Comparison of quadriceps peak torque over time.

(Redrawn from Wojtys, E. M.; Huston, L. J.: Longitudinal effects of anterior cruciate ligament injury and patellar tendon autograft reconstruction on neuromuscular performance. Am J Sports Med 28:336–344, 2000.)

FIGURE 12-18 Comparison of the time to peak torque of the quadriceps over time.

(Redrawn from Wojtys, E. M.; Huston, L. J.: Longitudinal effects of anterior cruciate ligament injury and patellar tendon autograft reconstruction on neuromuscular performance. Am J Sports Med 28:336–344, 2000.)

FIGURE 12-19 Comparison of the time to peak torque of the hamstring muscles over time.

(Redrawn from Wojtys, E. M.; Huston, L. J.: Longitudinal effects of anterior cruciate ligament injury and patellar tendon autograft reconstruction on neuromuscular performance. Am J Sports Med 28:336–344, 2000.)

Although deficits in proprioception after ACL reconstruction have been documented by several investigators,^{12,27,106,112} others have shown that no significant difference exists after surgery.^{32,82,88,155,159} The differences in findings are based on variability in test procedures, time from injury to ACL reconstruction, the presence of associated injuries to the menisci and articular cartilage, the rehabilitation program, and the use of internal versus external control limbs.¹⁵⁵ It is also difficult to understand whether statistically significant differences reported by authors represent clinically relevant findings. The two most common tests for proprioception are threshold for detection of passive motion (TDPM) and joint position sense (JPS). Differences in the magnitude of error in the patients’ ability to reproduce a specific joint position (flexion angle) postoperatively during these tests are frequently less than 1° to 2° compared with either internal or external control data.

Lephart and coworkers¹⁰⁶ reported significantly decreased kinesthetic awareness in 12 ACL-reconstructed knees compared with the uninvolved limb near the terminal end of knee motion. The ACL-reconstructed knees demonstrated a longer TDPM at 15° of knee flexion (mean difference, ~1.5°, moving from 15° of flexion to full extension). However, no significant difference was found between limbs in kinesthetic awareness at higher knee flexion angles (≥45°). MacDonald and associates¹¹² reported a significant difference in the mean TDPM in ACL-reconstructed (0.17°–0.22°) as well as ACL-deficient (0.14°) knees when the mean values were compared with those of internal controls (uninvolved knees). However, this study found no difference in mean TDPM values when ACL-deficient and ACL-reconstructed knee data were compared with external control knee values. Neither of these studies conducted JPS testing.

Risberg and coworkers¹⁵⁹ found no significant difference in mean TDPM between ACL-reconstructed knees and internal or external controls. The patients were tested a mean of 24 months (range, 11–32 mo) after B-PT-B autogenous ACL reconstruction and a supervised rehabilitation program. No effects were found in TDPM values when a functional knee brace was applied to the reconstructed knees. Fischer-Rasmussen and Jensen⁵⁷ also found no difference in mean TDPM between ACL-reconstructed and ACL-deficient knees compared with external controls. These authors did find a significant difference in JPS at 60° of flexion between ACL-reconstructed and external controls and between ACL-deficient and external controls. No difference was found between the three groups in JPS at full extension.

Co and coworkers³² measured TDPM and JPS in 10 patients a mean of 31.6 months after ACL B-PT-B autogenous reconstruction. All of the knees had less than 3 mm of increased AP displacement and a significantly decreased quadriceps isokinetic peak torque of the reconstructed limb compared with the opposite limb (84 ± 24 Nm and 98 ± 24 Nm, respectively; P = .005). An external control group of 10 subjects was included in the investigation. No significant differences were found between the ACL-reconstructed limbs and both the internal control (contralateral) and the external control limbs for JPS. However, a significant difference was found for TDPM between the ACL-reconstructed limbs and both the internal control (contralateral) and the external control limbs. The ACL-reconstructed limbs had a more accurate response than the external controls. The authors concluded that, in successfully ACL reconstructed knees (demonstrating <3 mm of increased AP displacement), proper rehabilitation and training may help overcome loss of proprioception usually demonstrated in chronic ACL-deficient knees. The authors did not measure proprioception in ACL-reconstructed knees with less optimal results for AP displacement or poorer rehabilitation training methods.

Roberts and associates¹⁶¹ reported that bilateral proprioception deficits were found in 20 patients who had a B-PT-B autogenous ACL reconstruction for chronic ruptures compared with an external control group. The patients were tested an average of 2 years postoperatively; however, no data were given regarding AP displacement or their overall functional status. The patients had a significantly higher TDPM in the reconstructed limb compared with external controls from starting positions of 20° and 40° for both extension (1.0° and 0.75°, respectively) and flexion. The differences between the reconstructed and the external control limbs for flexion were 1.0° at the starting position of 20° and 0.5° at 40°. Patients also had a higher TDPM in their contralateral limb compared with external controls from starting positions of 20° and 40° for both extension and flexion. However, there was high variability in the data, which could influence the significance of these findings. For instance, the range of values for threshold toward extension at the 20° flexion starting position in the patients’ reconstructed limbs was 1.0° to 6.0°, with a median of 1.0°. The range of values for this test in the external control group was 0.5° to 2.25°, with a median of 0.75°. Further, the authors indicated that although some patients in their series demonstrated a marked “decrease in proprioception ability,” others had small deviations from the group median. The number of patients with the marked proprioception deficits was not provided.

Bonfim and colleagues²⁷ conducted an investigation to detect sensory and motor deficits in 10 controls and 10 patients who had B-PT-B autogenous ACL reconstruction. The authors measured JPS, TDPM, latency of hamstring muscles, and maintenance of an upright stance position. The patients were tested an average of 18 months (range, 12–30 mo) postoperatively. The ACL-reconstructed knees demonstrated significantly decreased JPS (at 0°, 15°, 30°, 45°, and 60° of flexion) and significantly higher TDPM for both flexion and extension compared with the noninvolved side. The patients also showed longer latency of the hamstring muscles, and increased body sway during single-leg stance, in the reconstructed limb compared with the noninvolved side. The authors concluded that these findings were due to the disrupted ACL mechanoreceptors that were not restored by the reconstruction.

Hopper and coworkers⁸² reported no significant differences in JPS measured during full weight-bearing in nine patients who underwent STG ACL reconstruction. No external control group was incorporated into this investigation. Reider and associates¹⁵⁵ found no significant differences in JPS or TDPM between ACL-reconstructed knees and external controls at 3 weeks, 6 weeks, and 3 months after surgery. At 6 months postoperative, the ACL-reconstructed knees had significantly better JPS mean values compared with the external control group (5.67° and 7.53°, respectively; mean difference, –1.86°). There was no difference between ACL-reconstructed knees and external controls in TDPM mean values 6 months postoperatively.

The differences in test methodology and populations studied do not allow definitive conclusions to be reached regarding the effect of ACL reconstruction on knee joint proprioception. Whether the finding of less than 1° to 2° of difference in JPS or TDPM between ACL-reconstructed knees and external controls is clinically relevant remains to be determined. Studies have not assessed whether specific rehabilitation exercises (such as OKC vs. CKC) influence knee joint proprioception postoperatively.

CLINICAL STUDIES AND OUTCOME OF ACL REHABILITATION PROGRAMS

Effect of Postoperative Exercises on AP Knee Displacements

The authors conducted a study to determine the effect of rehabilitation exercises and time elapsed postoperatively on AP knee displacements after ACL B-PT-B autograft reconstruction performed by a single surgeon.¹⁰ A total of 142 patients were followed a minimum of 2 years postoperatively; 90 had the operation for chronic ACL ruptures and 52, for acute ruptures (reconstruction performed within 12 wk of the injury). One experienced examiner conducted KT-2000 testing throughout the study period (134 N) at 8, 12, 16, 20, 24, 52, and 128 weeks after surgery. A total of 938 arthrometer measurements were collected, for an average of 7 per patient.

The rehabilitation program was divided into four phases. The assisted ambulatory phase represented the length of time patients spend using crutch or cane support, which lasted until approximately the 4th to 8th week postoperative. During this phase, ROM exercises, straight leg raises (in all four planes), quadriceps muscle isometrics, EMS, and CKC exercises (minisquats, toe raises) were performed. The early strength-training phase lasted from between approximately the 4th to 8th week to the 12th to 16th week and added balance, proprioceptive, and gait-training exercises. The intensive strength-training phase focused on progressive resistive exercises, swimming, bicycling, ski machines, stair climbing machines, and running programs. This phase varied between patients, but typically lasted from approximately the 12th to 16th week to the 24th and 52nd postoperative week. The final phase, return to sports, was entered when patients successfully completed the intensive strength-training program. Return to running and light sports was allowed at 6 months postoperative, and full competitive sports, at approximately the 8th postoperative month in this group of patients.

Critical Points CLINICAL STUDIES AND OUTCOME OF ACL REHABILITATION PROGRAMS

Effect of Postoperative Exercises on Anteroposterior Knee Displacements

Barber-Westin et al.¹⁰: Studied effect of rehabilitation exercises and time elapsed postoperatively on AP knee displacements after ACL B-PT-B autograft reconstruction performed by a single surgeon.

• 142 patients followed minimum 2 years postoperatively.

• One experienced examiner conducted KT-2000 testing: total 938 arthrometer measurements were collected, average of 7 per patient.

• Most recent follow-up: 121 (85%) had normal displacements (<3 mm of increased AP displacement compared with the contralateral side) and 21 (15%) had abnormal displacements.

• 14 patients had 3 to 5.5 mm of increased displacement, 7 patients had 6 mm or more.

• No association between the initial onset of the abnormal displacements and the amount of time that had elapsed since surgery or the phase of rehabilitation in which they were detected.

Home- versus Clinically Based Supervised Physical Therapy Programs

• Four randomized investigations compared home-based rehabilitation with supervised clinically based physical therapy programs.

• All concluded similar outcomes between these programs. Did not provide data on the sports activities patients returned to (asymptomatically vs. symptomatically), details on end-stage functional progression of programs (e.g., plyometric training), long-term outcome.

• Essential components for home-based program: preoperative patient education sessions, a comprehensive written description of the postoperative program, periodic monitoring of patient progress by a therapist.

At the most recent follow-up evaluation, 121 (85%) of the patients had normal displacement values (<3 mm of increased AP displacement compared with the contralateral side) and 21 (15%) had abnormal displacements. Fourteen patients had 3 to 5.5 mm of increased displacement and 7 patients had 6 mm or more, indicating graft failure. There was no significant difference in these measurements between the knees operated on for acute ACL ruptures and those with chronic ruptures.

There was no association between the initial onset of the abnormal displacements and the amount of time that had elapsed since surgery. The abnormal displacements were first detected a mean of 46 ± 51 weeks and a median of 24 weeks (range, 6–208 wk) postoperatively. In 7 patients, the abnormal displacements were detected less than 20 weeks postoperatively; in 9 patients, these were detected between 20 and 52 weeks; and in 5 patients, these were detected longer than 1 year postoperatively.

There was also no association between the initial onset of abnormal displacements and the phase of rehabilitation in which they were detected. The percentage of knees with less than 3 mm, 3 to 5 mm, and greater than 5 mm during each phase of rehabilitation is shown in Figure 12-20. In 8 patients, the abnormal displacements were first detected during the early strength-training phase; in 6 patients, during the intensive strength-training phase; and in 7 patients, after return to sports. There was no relationship between the type of sports activity patients had returned to and the presence of abnormal AP displacements.

FIGURE 12-20 The percentage of knees distributed in the three arthrometer categories according to the phases of rehabilitation.

(Redrawn from Barber-Westin, S. D.; Noyes, F. R.; Heckmann, T. P.; Shaffer, B. L.: The effect of exercise and rehabilitation on anterior-posterior knee displacements after anterior cruciate ligament autograft reconstruction. Am J Sports Med 27:84–93, 1999.)

The study demonstrated that the operation and the rehabilitation program were effective and resulted in an acceptable failure rate of 5%. It is important to note that the patients in this investigation were widely varied with regard to the goals of the operation, age, injury chronicity, prior treatment of the injury, and condition of the menisci and articular cartilage. The rehabilitation program was individualized as required and was based on patient symptoms and functional testing.

ACCELERATED REHABILITATION: INDICATIONS, CONTRAINDICATIONS, AND OUTCOME

The phrase accelerated rehabilitation was introduced by Shelbourne and Nitz in 1990¹⁷⁹ to describe a program that incorporated rapid allotments for strength and functional training and return to full activities after B-PT-B autogenous ACL reconstruction. Light sports were permitted by 2 months and full activity by 4 to 6 months postoperative. Since then, many authors have used this term to describe other facets of rehabilitation and the authors agree with Beynnon and colleagues’ assessment¹⁸ that “there is little consensus in the literature about what composes an accelerated versus a more conservative rehabilitation program.” Shelbourne and Nitz’s report¹⁷⁹ provided outcome data on 73 of 237 (30%) patients who participated in the accelerated program. The results of these patients were compared retrospectively with those of others who had participated in a more conservative program. The authors reported that the accelerated program was more effective in restoring knee motion and preventing complications related to arthrofibrosis. There was no difference between programs in quadriceps strength at 1 year postoperative, patient perception of the results of surgery, or AP displacements on KT-1000 testing.

Subsequent reports by Shelbourne and coworkers^176,177 emphasized the need for preoperative rehabilitation to restore full ROM, decrease swelling, and resume normal gait and leg control. The longest postoperative follow-up of this program reported by these investigators to date (mean, 4 yr; range, 2–9.1) showed a low failure rate of 3%, a low arthrofibrosis rate of 1%, and an average time to return to sports of 6.2 ± 2.3 months. The authors did not report what types of athletic activities patients participated in before their injury or at the most recent follow-up evaluation. As well, the authors noted that some patients had problems with the accelerated program in terms of pain and chronic joint effusion, especially those whose quadriceps strength was less than 65% of that of the opposite limb. The long-term effects of the return to strenuous training in the early postoperative period (5 wk on average) on articular cartilage and future joint arthrosis are unclear.

Majima and associates¹¹³ reported that accelerated rehabilitation (as described by Shelbourne¹⁷⁹) produced a significant increase in the incidence of joint effusions in 32 patients who underwent STG reconstruction compared with 18 patients who were treated with a conservative protocol. Within the first 8 postoperative weeks, a total of 15 aspirations (in 3 patients) had been done in the conservative group compared with a total of 60 aspirations (in 13 patients) in the accelerated group (P < .01). No difference was found between the two protocols in regard to return of muscle strength (after 9 mo postoperative), IKDC scores, or knee function. Patients in the accelerated group had a trend toward greater AP displacement, as 20% had greater than 3 mm of increased displacement compared with 13% in the conservative group.

Roi and colleagues¹⁶² published a detailed accelerated rehabilitation program that was used to return a professional soccer player to his sport 90 days after an ACL STG reconstruction. The athlete did not sustain concomitant ligament or meniscus injuries and had minimal pain and limitations with daily activities after the injury, all of which allowed surgery to be performed just 4 days later. Although the treatment protocol was successful in this case, the authors qualified their aggressive program as one that “may only be possible with an individual who has the resources (time and money) to invest in unlimited access to rehabilitation facilities and personnel.”¹⁶² The authors used objective criteria to advance the program including ROM, KT-2000, functional, isokinetic, aerobic, and anaerobic testing.

Wilk²⁰⁹ questioned the applicability and indications of accelerated rehabilitation. Citing the ultimate goal of ACL reconstruction and postoperative rehabilitation as the return of normal homeostasis to the knee joint⁴⁷ over the long term, he cautioned that aggressive rehabilitation may have more risks than benefits, especially in regard to its impact on preexisting articular cartilage damage and bone bruising.

Yu and Paessler²²⁵ investigated the influence of an aggressive rehabilitation protocol on tunnel widening after four-strand STG reconstruction. At 6 months postoperative, patients who followed the aggressive protocol (immediate full weight-bearing, OKC exercises begun at the 6th week, running allowed from the 8th to the 10th week) had significantly greater tibial tunnel widening on posteroanterior (PA) and lateral radiographs compared with a group of patients who followed a more conservative protocol. There was no difference between groups in mean KT-1000 arthrometer values.

Owing to the lack of randomized, prospective studies on the return to high-impact loading activities in the early postoperative period after ACL reconstruction, there remain unanswered questions regarding appropriate patient candidates, indications, and contraindications for this program. In the authors’ opinion, contraindications include (1) noteworthy bone bruising on MRI or articular cartilage damage visualized on arthroscopy, (2) concomitant complex meniscus repairs, reconstructions of other knee ligaments, patellar realignment procedures, or articular cartilage restorative procedures, (3) postoperative joint effusion, pain, and (4) inability of the patient to regain normal gait, neuromuscular control, and muscle strength. Indications include use of a B-PT-B autograft with secure internal fixation, serial evaluation of AP displacement postoperatively, return of normal gait and muscle strength, a compliant patient, a carefully supervised environment, and completion of a neuromuscular-retraining program in all athletes prior to return to full competition.

CRITERIA FOR PATIENT RELEASE AND RETURN TO SPORTS ACTIVITIES

Different authors have described objective and subjective criteria that patients must achieve before clearance is given to return to strenuous sports activities. These include assessment of lower limb symmetry on single-leg functional tests, lower extremity muscle strength on an isokinetic dynamometer, AP displacement on knee arthrometer testing, trials-of-function of activities of increasing demand, and ROM.

Kvist¹⁰³ detailed multiple criteria to be fulfilled before release to full sports activities that included rehabilitation, surgical, and other factors. The assessment included objective muscle strength and performance, pain, effusion, ROM, functional knee stability, static knee stability, associated injuries, and psychological and social aspects. The patient had to demonstrate less than 15% deficit on isokinetic and single-leg hop tests, no pain or effusion, full ROM, functional knee stability, and static knee stability on KT-1000 testing. The influence of associated meniscal, cartilage, or other ligamentous injuries; psychological issues such as motivation or fear of reinjury; and social factors such as lost time owing to occupational demands could not be predicted on the ability to return to sports successfully after ACL reconstruction.

Discharge criteria at the authors’ center after ACL reconstruction based on patient goals for athletics and occupations, the rating of symptoms, KT-1000 testing, muscle strength testing, and function testing have been previously described (Table 12-4).⁷⁵ First, patients complete the Cincinnati Sports Activity Scale¹¹ and the Occupational Rating Scale¹¹ in order to document sports and occupational levels that are desired after surgery (see Chapter 44, The Cincinnati Knee Rating System). Upon completion of the rehabilitation program, pain, swelling, and giving-way are rated on the Cincinnati Symptom Rating Scale.¹¹ The patient must not experience symptoms at the level of activity that she or he wishes to participate in prior to discharge. KT-2000 testing is performed at 134 N of total ‘AP force and must be within normal limits (<3 mm difference between limbs) prior to allowance of return to strenuous activities. Muscle strength testing is performed with an isokinetic dynamometer to ensure that adequate strength exists prior to the initiation of the running and cutting programs. At least two single-leg hop function tests are completed and limb symmetry calculated as described previously.^9,128 Patients desiring to return to high-risk activities complete a neuromuscular-retraining program before release to full competition. Recommendations for the types of activities that patients should consider resuming are made based on the results of these criteria and the condition of the articular cartilage and menisci. Patients in whom articular cartilage damage was visualized during the ACL reconstruction (especially those with grade 2B or 3; see Chapter 47¹³⁸), or those in whom the majority of function of one or both menisci has been compromised are encouraged to return to low-impact activities to preserve the knee joint for as long as possible.

RISKS FOR REINJURY AND FUTURE JOINT ARTHROSIS

An important component of the surgery and rehabilitation decision making process is an understanding of the risks of reinjury to either the reconstructed or the contralateral knee and the chance of developing future joint arthrosis. A few studies have reported high rates of reinjury upon return to sports.^{145,154,169,204} Salmon and coworkers¹⁶⁹ followed 67 patients for 7 to 13 years postoperatively after ACL B-PT-B autogenous reconstruction and a conservative rehabilitation program. Patients were kept non–weight-bearing for the first 4 postoperative weeks, with ROM permitted only from 30° to 90°. Competitive sports were not allowed until 9 months after surgery. A total of 23 patients (34%) sustained an ACL injury after return to sports; 9 (13%) patients reinjured the reconstructed ACL and 15 (22%) ruptured the contralateral ACL. The authors related that their patients had not participated in a neuromuscular-retraining program, which may have been at least partially responsible for the high reinjury rate. At 13 years postoperative, 37% of the patients reported swelling after moderate activity and 44% had a limitation in knee extension.

Orchard and associates¹⁴⁵ followed Australian football players over eight seasons to determine the occurrence of ACL injuries. The overall ACL injury incidence was 0.82/1000 athlete exposures. The incidence for noncontact ACL injuries was 0.62/1000 athlete exposures. All ACL injuries in this study were treated with reconstruction. The players who sustained an ACL injury had approximately a 10 times greater incidence of sustaining further ACL injury within the first 12 months of surgery compared with uninjured players. After 12 months, these players still had a greater than 4 times increased risk of sustaining a noncontact ACL injury, which was evenly divided between the reconstructed and the contralateral knee.

Pinczewski and colleagues¹⁵⁴ presented high reinjury rates after return to activity in a cohort of patients who received either a B-PT-B (90 patients) or an STG autogenous (90 patients) ACL reconstruction and who were observed for 10 years. The rehabilitation program included no brace, immediate partial weight-bearing, jogging after 5 weeks, and return to sports 6 months postoperatively if knee stability was confirmed. The overall ACL reinjury rate, including both the reconstructed and the contralateral sides, was 27%; 22% for STG grafts and 30% for B-PT-B grafts. A total of 19 patients (11%) ruptured the ACL graft: 8% (7/90) in the B-PT-B group and 13% (12/90) in the STG group. A total of 29 patients (16%) sustained contralateral ACL ruptures: 22% (20/90) in the B-PT-B group and 10% (9/90) in the STG group.

Wright and coworkers²¹⁶ reported early outcomes of the Multicenter Orthopaedic Outcome Network (MOON) investigation that documented reinjury rates in both ACL-reconstructed and contralateral knees in 235 patients followed for 2 years after autograft or allograft reconstruction. The reinjury rates included only patients who underwent ACL revision or primary reconstruction in the contralateral knee. The authors acknowledged the limitations of the study of no on-site physician examination of the patients who had not undergone further surgery and a short-term follow-up. It is feasible that other patients sustained reinjuries that were not detected. There were seven tears of the intact ACL in the contralateral knee and seven ruptures of ACL grafts, for an overall incidence rate of 3% for each knee. No information was provided regarding the type of postoperative rehabilitation programs the patients participated in or functional outcome.

Published rates of radiographically documented OA after ACL reconstruction vary tremendously (Table 12-5).* It is important to note that some studies report an increased incidence of mild OA (IKDC grade nearly normal, Ahlback grade 1) changes in the knee and not moderate or severe degeneration. Whether mild OA changes on standing radiographs represent actual articular cartilage deterioration is questionable. Two investigations concluded that standing radiographs have low sensitivity rates in detecting mild articular cartilage damage. Wright and associates²¹⁵ determined the sensitivity and specificity of standing AP and weight-bearing PA radiographs in detecting Outerbridge grade 2 (fragmentation and fissuring, <½ inch in diameter) chondral lesions in a group of 349 patients. Although both radiographic techniques had specificity rates greater than 90%, both had poor sensitivity rates for detecting grade 2 changes in the medial (3% AP, 6% PA) and lateral (16% AP, 6% PA) compartments. Lysholm and colleagues¹¹¹ reported a similar poor correlation between mild OA (Outerbridge grade 2) and the Ahlback classification² rating of standing radiographs.

Determining the risk of developing future OA after ACL reconstruction is difficult for many reasons. Many factors may influence the development of articular cartilage deterioration, including the effects of the initial impact injury on the subchondral bone (bone bruising^34,52,80,195), loss of meniscus tissue (either before, during, or after the ACL reconstruction), patient age, time from injury to reconstruction, number of reinjuries prior to reconstruction, failure of the reconstruction to restore normal AP displacement and knee kinematics, the rehabilitation program, complications (such as infection or arthrofibrosis), final knee motion achieved, and the activity level the patient resumes after surgery. In many studies, these variables are not controlled for, which when combined with the high incidence of the detection of only mild radiographic OA changes, makes reaching conclusions on this topic difficult. In addition, most long-term studies have an unfortunately high rate of transfer bias (patients lost to follow-up) and do not compare the radiographic findings of the reconstructed knee with the contralateral normal knee.

One conclusive association with future joint OA is loss of meniscus tissue. Nearly every long-term study reports a statistically significant correlation between prior meniscectomy and radiographic OA. Hart and coworkers⁷⁴ conducted a 9- to 13-year follow-up study on 31 patients who underwent ACL B-PT-B reconstruction. Both weight-bearing PA radiographs and single-photon emission computed tomography (SPECT) were used to detect OA. When the population was sorted according to meniscectomy (15 meniscus intact, 16 meniscectomy), the SPECT scans detected OA in 31% of the patients compared with just 7% in whom meniscus tissue had been preserved.

Salmon and coworkers¹⁶⁹ reported data on 43 patients who underwent radiographic assessment 13 years postoperatively in which the IKDC radiographic scores were 37% normal; 42% nearly normal; 19% abnormal; and 2% severely abnormal. In comparison, at 7 years postoperative, 44 patients underwent radiographic assessment and 59% were rated as normal; 34% nearly normal; and 7% abnormal. The difference in the proportion of patients with an abnormal radiographic score between evaluations was significant (P = .01) and correlated with meniscectomy and greater amount of anterior tibial displacement on Lachman testing. The majority of moderate to severe degeneration occurred in the medial tibiofemoral compartment. No associations were found between radiographic and IKDC symptom and function scores.

Pinczewski and colleagues¹⁵⁴ reported 10-year radiographic postoperative data on 59 B-PT-B patients in whom the IKDC scores were normal in 61%, nearly normal in 36%, and abnormal in 3%. Sixty-nine of the STG patients in this series were radiographically assessed and the IKDC scores were normal in 81%, nearly normal in 17%, and abnormal in 1%. The IKDC scores correlated with the single-leg hop test and graft type.

Shelbourne and Gray¹⁷⁸ followed 482 patients who received a B-PT-B autogenous ACL reconstruction between 5 and 15.8 years postoperatively. The patients were sorted according to prior meniscectomy and articular cartilage damage noted at the time of the ACL reconstruction. A clear association was found between meniscectomy and radiographic OA. A normal or nearly normal radiographic IKDC score was reported in 97% of patients who had intact menisci compared with just 64% of patients who had one or both menisci removed and articular cartilage damage. The authors concluded that ACL reconstruction offered a protective effect in knees that had normal menisci and no articular cartilage damage. Most authors agree that early knee joint stabilization procedures should be considered prior to reinjuries and subsequent meniscus and articular cartilage damage.

In a review of literature on the long-term consequence of ACL injuries, Lohmander and associates¹⁰⁸ proposed an overall long-term mean value of greater than 50% of patients who sustained this injury will develop OA. The authors combined data from ACL-injured and ACL-reconstructed studies and were not able to account for the effect of meniscectomy or reinjuries. Many problems in study design and loss of follow-up (transfer bias) were cited; even so, the authors concluded that there is a lack of evidence to support that ACL reconstruction reliably plays a protective role against the development of OA.

As additional well-designed long-term studies on the outcome of ACL reconstruction become available, the issue of the operation’s ability to provide chondroprotective effects should become clearer. Studies should document the chronicity of the ACL rupture and whether reinjuries occurred before the reconstruction, the status of the menisci during the entire observation period, the condition of the articular cartilage noted at the reconstruction (rated by a validated system, Chapter 47), athletic activity levels patients returned to, and radiographic findings of both knees.