Prevention and Treatment of Knee Arthrofibrosis

Introduction

Loss of a normal range of knee motion following an injury or ligament reconstruction is a potentially devastating complication. Many investigations and publications have appeared since the early 1990s regarding this problem, and it continues to be one of the most frequently reported complications after surgical procedures such as autologous chondrocyte implantation (ACI), knee ligament reconstructions, and total knee arthroplasty (TKA). Although the definition and use of the term arthrofibrosis varies among authors, it implies a loss of knee flexion, extension, or both compared with the contralateral normal knee. Stephenson and associates estimated that the number of patients who undergo major knee surgery in the United States that are subsequently affected by arthrofibrosis is at least 85,000 a year. Of these, 21,000 are at risk of requiring additional surgery, and 54,000 may have an unsuccessful outcome of treatment for this complication.

Primary arthrofibrosis is caused by an exaggerated inflammatory response to an injury or surgical procedure, followed by the production of fibroblastic cells and an increase in the deposition of extracellular matrix proteins. Proliferative scar tissue or fibrous adhesions form within the joint, which can result in either localized or diffuse involvement of all of the compartments of the knee and the extra-articular soft tissues ( Fig. 38-1 ). In the most severe cases, dense scar tissue obliterates the normal peripatellar recesses, suprapatellar pouch, intercondylar notch, and articular surfaces. Formation of dense scar tissue in the infrapatellar region may result in a patella infera and permanent limitation of normal knee flexion and extension. The consequent pain and restricted knee motion resulting from arthrofibrosis may lead to disabling events, including severe quadriceps atrophy, loss of patellar mobility, patellar tendon adaptive shortening, patella infera, and articular cartilage deterioration.

The incidence of arthrofibrosis after knee ligament reconstructive surgery varies according to several factors. In our experience, two of the most commonly noted factors are knee dislocations and major concurrent operative procedures ( Table 38-1 ). Studies show wide disparity among authors regarding the treatment of this problem, especially the postoperative time that intervention is warranted and what type of treatment should be implemented for losses of knee extension and flexion and patella mobility.

Note that limitations in knee flexion and extension may occur because of reasons other than arthrofibrosis. Any acute injury causing pain and swelling may cause a short-term loss of knee motion, which usually resolves as symptoms subside. Other factors influencing the restoration of knee motion include a mechanical block from a displaced bucket-handle meniscus tear, an impinged or improperly placed anterior cruciate ligament (ACL) graft, a cyclops lesion, improper graft tensioning that constrains normal knee joint motion, and graft fixation at 30 degrees or more of extension. In addition, contracture of the posterior capsular structures may limit knee extension. If not resolved, these problems may result in the development of what is termed secondary arthrofibrosis , complicating the course of treatment.

Normal knee motion varies, but most individuals have some amount of hyperextension, which averages 5 degrees in men and 6 degrees in women. This degree of hyperextension is required so that the normal “screw-home” mechanism that allows the knee to be stabilized in extension during the stance phase may occur, with the quadriceps muscle relaxed. Soucie and colleagues provided the following ranges of knee motion, according to gender and age from a sample of 674 healthy subjects: (1) ages 9 to 19 years: 2.4 to 142.3 degrees for girls and 1.8 to 142.2 degrees for boys, (2) ages 20 to 44 years: 1.6 to 141.9 degrees for women and 1.0 to 137.7 degrees for men, and (3) ages 45 to 69 years: 1.2 to 137.8 degrees for women and 0.5 to 132.9 degrees for men.

It is generally assumed that at least 125 degrees of knee flexion is required for activities of daily living. Rowe and coworkers measured knee kinematics during functional activities in a group of 20 subjects aged 49 to 80 years and reported that 90 degrees of flexion is required for gait and slopes, 90 to 120 degrees for stairs and sitting, and 135 degrees for bathing. Patients not involved in athletics tolerate small flexion deficits. Significant flexion deficits of less than 90 degrees cause problems with squatting, stair climbing, and sitting. Athletes have a poorer tolerance for even minor losses of flexion, which affect running and jumping. Cosgarea and associates reported that deficits of 10 degrees or more of flexion were associated with decreased running speed.

All patients, regardless of their activity level, have greater problems with loss of knee extension. A loss of just 5 degrees of extension may produce a flexed-knee gait, fatigue the quadriceps muscles, and cause patellofemoral pain. Limitation of more than 20 degrees may cause a functional limb-length discrepancy. Many years ago, Perry and colleagues demonstrated the effect of a loss of extension on joint contact pressures, quadriceps muscle activity, and fatigue. These investigators measured the extensor forces required to stabilize the flexed knee during simulated weight bearing in cadaver specimens. The quadriceps force required to stabilize the knee was 75% of the load on the head of the femur at 15 degrees of flexion, which increased to 210% at 30 degrees of flexion and to 410% at 60 degrees of flexion. The quadriceps force was equal to 20% of average maximum quadriceps strength at 15 degrees of flexion but increased to 50% at 30 degrees of flexion. Surface pressure in the tibiofemoral and patellofemoral joints also increased with greater amounts of knee flexion.

It has become evident that the prevention of knee arthrofibrosis is paramount and preferred over using the currently available treatment options for this complication. This chapter discusses the risk factors and preventive measures for loss of knee motion after knee injury and surgery. In addition, conservative medical and therapeutic treatment intervention strategies are discussed that are usually successful in resolving a transient limitation of knee motion if initiated early postoperatively. Surgical procedures for severe cases of restriction of knee motion are presented. Our clinical studies involving 650 ACL-reconstructed knees are included to provide support for treatment recommendations.

Terminology and Classification Systems

A variety of terms have been used to describe, define, or classify a limitation of knee flexion and extension, including arthrofibrosis, flexion contracture, ankylosis, infrapatellar contracture syndrome, and motion loss. Some of the terms indicate a specific process, but others simply imply a limitation of motion compared with the contralateral normal knee. Ankylosis is one such generic term that indicates stiffness of joints that occurs for any reason and may represent loss of knee flexion, extension, or both. The term motion loss is another general phrase used to describe deviations from the amount of flexion and extension compared with the contralateral normal knee. Conversely, arthrofibrosis describes a specific cause for a limitation of knee motion, which is the formation of diffuse scar tissue or fibrous adhesions within a joint. Flexion contracture indicates a loss of extension attributable to any cause and is accompanied by a relative shortening of the posterior soft tissues (in either the capsule or muscles). Petsche and Hutchinson state that arthrofibrosis and flexion contracture are general terms, and it is necessary to indicate a specific cause for loss of motion.

Classification systems have been proposed by various authors to describe limitations in knee motion, based either on the anatomic location of scar tissue and adhesions or on the amount of knee motion in the affected joint ( Table 38-2 ). Sprague and coworkers first introduced a system based on the arthroscopically observed location of scar tissues and adhesions in 24 patients who had a severe limitation of knee motion. Paulos and associates described three stages of knee motion restrictions caused by infrapatellar contracture syndrome. Del Pizzo and colleagues and Shelbourne and coworkers developed classification systems based on the deviation of knee motion compared with the opposite knee. Blauth and Jaeger described a system based on the total arc of motion in the affected knee.

We have developed a classification system based on the anatomic sites at which scar tissue, adhesions, and adaptive shortening of soft tissues occur ( Fig. 38-2 and Table 38-3 ). This system is advantageous because it highlights the areas that must be addressed surgically, as is discussed.

Risk Factors after Knee Ligament Reconstruction

A variety of factors appear to influence the requirement for treatment intervention for knee motion limitations ( Table 38-4 ) and the final amount of knee flexion and extension gained after ligament reconstruction. These factors are related to the severity of the injury, the timing of surgery, preoperative treatment, technical aspects of the ligament reconstruction, and the postoperative course and rehabilitation. Although it remains uncertain why some knees are more likely to form an abnormal scar response to trauma and surgery than others, knowledge of these factors may help reduce this problem.

Severity of the Injury

Patients who sustain knee dislocations are at increased risk for developing motion complications ( Fig. 38-3 ). These injuries frequently occur from high-energy accidents that produce extreme soft tissue swelling and edema, hematomas, muscle damage, and other multiple trauma that must be resolved before consideration of knee soft tissue reconstructions. Early operative intervention when possible is advised for multiple ligament injuries that involve the lateral and posterolateral structures where acute repair and augmentation procedures are performed (see Chapter 17 ). For all other dislocations, surgery is delayed with the limb immobilized in a posterior plaster splint or bivalved cast with a posterior pad to prevent posterior tibial subluxation. Even in these severe knee injuries, it is possible to start immediate range of motion and prevent scar tissue formation that may compromise the outcome of a subsequent ligament reconstruction. Too often, these serious knee joint injuries are not treated with an early motion and function program. When surgical treatment is elected, the reconstructive and repair procedures of torn ligaments, capsular structures, and menisci are performed in a manner that allows immediate knee motion to be instituted postoperatively.

Technical Factors at Surgery

Improper ACL graft placement has been frequently noted to cause loss of knee motion. On the tibia, grafts placed anterior to the native ACL insertion result in impingement on the roof of the intercondylar notch when the knee is extended. Yaru and colleagues recommended that, with passive extension, there should be a 3-mm clearance between the anterior portion of the intercondylar notch and the graft to prevent impingement. Grafts placed lateral to the insertion site impinge on the lateral wall of the notch. In addition, Romano and coworkers found that ACL grafts placed too far anteriorly in the tibial tunnel can cause knee extension deficits, and grafts placed too far medially in the tibial tunnel can cause knee flexion deficits. On the femur, an excessive anterior graft placement causes deleterious forces on the graft, leading to limitations in flexion and potential graft failure ( Fig. 38-4 ).

Overtensioning ACL grafts may lead to abnormal knee kinematics. In addition, the degree of knee extension during graft tensioning and fixation may affect postoperative motion. Austin and associates noted in a cadaver study that the amount of graft tension (44 N or 89 N) did not effect knee extension; however, tensioning the graft in knee flexion was associated with extension deficits. The authors reported that grafts tensioned and fixed at 30 degrees of flexion had more than a 12 degree increase in knee flexion after ACL reconstruction compared with those tensioned and fixed at 0 degrees of flexion. From a two-part biomechanic and clinical study, Nabors and colleagues suggested that grafts tensioned in full extension result in a low incidence of knee motion loss. In their series of 57 patients who underwent patellar tendon autograft ACL reconstruction, only one patient had a mild (5 degrees) loss of extension.

Concurrent medial collateral ligament (MCL) repair with ACL reconstruction has been associated with an increased risk of knee arthrofibrosis. Harner and associates postulated that concurrent MCL primary repair may cause a limitation of knee motion attributable to the disruption of the medial capsule because the procedure does not restore the normal tissue planes, resulting in scar formation and a heightened pain response. Our studies and the studies of Shelbourne and coworkers led to the recommendation many years ago to treat the majority of combined ACL/MCL ruptures conservatively. This allows healing of the medial-side injury followed by ACL reconstruction in appropriately indicated patients (see Chapter 19 ). There are exceptions to this rule, and any surgical procedure on the medial side of the knee joint should be carefully observed postoperatively for a limitation of motion or, in acute knee injuries, the development of heterotopic soft tissue ossificiation requiring treatment.

Pathophysiology and Experimental Prevention Studies

In 1972, Enneking and Horowitz published one of the first studies on the pathophysiology of a joint contracture after immobilization in humans. The investigators described in a series of case reports the presence of fibrofatty connective tissue in the infrapatellar fat pad, suprapatellar pouch, and posterior recesses of the knee joint with eventual obliteration of the joint cavity. Over time, this tissue was replaced with mature fibrous connective tissue.

Tissue homeostasis is maintained by the normal level of cell growth and proliferation along with the production and turnover of the extracellular matrix. Polypeptides known as cytokines or growth factors are responsible for the constant signaling that occurs both between local cells (paracrine) and within cells (autocrine). Cytokines exist as small proteins that signal cells in response to injury and infection. One cytokine, interleukin-1 (IL-1), has been found to stimulate platelets and macrophages to release a variety of growth factors, including transforming growth factor-β (TGF-β). TGF-β is thought to be one of the factors (among others ) mediating Dupuytren's contracture. Cytokines regulate all aspects of tissue remodeling and may act positively or negatively on tissue damage. This variability results in a wide range of potential responses to injury among patients.

TGF-β, released by platelets, tendon fibroblasts, and the joint capsule, has been identified as a key cytokine that both initiates and ends the process of tissue repair. Although other growth factors are involved in tissue remodeling, TGF-β is unique in its actions that enhance the deposition of extracellular matrix and regulate the actions of other cytokines. An overexpression of TGF-β ₁ results in progressive accumulation of matrix in tissues and an elevated number of myofibroblasts, leading to fibrosis. Border and Noble believe that repeated injury or trauma may also cause autoinduction of TGF-β beyond normal levels, creating a “chronic, vicious circle of TGF-β overproduction” resulting in tissue fibrosis that may also occur in organ systems such as the kidney, liver, and lung. Investigators have documented that elevated levels of myofibroblasts and TGF-β occur rapidly, within 2 weeks following injury in experimental models, and that these changes are similar to those observed in humans with chronic elbow joint contractures.

Zeichen and coworkers compared tissue samples of 13 patients with arthrofibrosis obtained a mean of 16 months after the inciting knee ligament surgery with those of 8 patients undergoing primary ACL surgery. Histologic examination from the patients with arthrofibrosis showed marked synovial hyperplasia, infiltration of inflammatory cells, and vascular proliferation with an increased level of collagen types VI and III. None of the control subjects demonstrated these findings. These authors noted that this confirmed the observations of Murakami and associates regarding the association between an inflammatory process and proliferation of fibroblasts. The suggestion was made that collagen type VI could play a contributory role in the deposition of extracellular matrix that leads to arthrofibrosis.

There is evidence that during the exaggerated fibrotic response that occurs in knee arthrofibrosis, a certain amount of tissue contraction or shrinkage occurs. Specific α-smooth muscle actin (ASMA)-expressing myofibroblasts generate tissue contraction during wound healing and are responsible for collagen overproduction during fibrotic diseases. TGF-β is capable of up regulating ASMA and collagen in fibroblasts.

Unterhauser and associates measured the amount of ASMA-containing fibroblastic cells in arthrofibrotic tissue to determine whether the number of these cells was increased over that found in normal tissue. Tissue biopsies were obtained from 9 patients who underwent surgery for arthrofibrosis that had developed after ACL reconstruction. The patients all had primary arthrofibrosis and underwent the debridement procedure between 4 and 12 months after the ACL reconstruction. Tissue samples were also taken from 5 patients who underwent ACL reconstruction for chronic ligament deficiency and from 8 patients who had follow-up arthroscopy after ACL reconstruction for meniscus pathology; all 13 of these control patients showed no signs of arthrofibrosis. Significantly higher expressions of ASMA-positive fibroblastic cells were found in the collagen bundles and remaining fatty tissue in the study group compared with both control groups (23.4% of the total cell count vs 2.3% and 10.8%, respectively; P < .001). These authors concluded that this cell type is involved in scar formation and scar tissue contraction during the course of primary arthrofibrosis. They speculated that the expression of ASMA is most likely variable according to the time from the onset of the disease to tissue biopsy and probably decreases as the disease progresses.

Unterhauser and associates also reported that the fatty tissues in the arthrofibrotic knees were replaced by a dense collageneous network with parallel fiber orientation. A significantly higher total cell count and lower vessel density were measured in the arthrofibrotic knees compared with the control group that underwent ACL reconstruction. Similar findings were noted by Mariani and colleagues who biopsied 17 knees with arthrofibrosis that occurred after a variety of surgical procedures. The tissue samples were obtained less than 6 months from the onset of stiffness in 9 patients, from 6 to 12 months in 5 patients, and more than 12 months in 4 patients. The histological evaluations demonstrated collagen-producing fibroblastic tissue with differing amounts of cellularity and vascularization. This study supports the hypothesis that arthrofibrotic tissue undergoes progressive remodeling and acquires the characteristics of mature collagen, all within the first 6 months after the onset of joint stiffness. There was no association between the amount of knee motion and time from surgery or the degree of tissue maturation. The severity of loss of motion was associated only with the location and amount of adhesive tissue in the knee joint.

Many investigations have suggested that antifibrotic agents may be of potential use in preventing or combating arthrofibrosis because they inhibit the TGF-β signaling pathway. However, the effectiveness of various agents such as decorin, hyaluronic acid, chitosan, mitomycin C, and various IL-1 receptor antagonists in high-level human studies for early or later-stage arthrofibrosis has not been investigated to date.

The use of powerful antiinflammatory medications, typically used for rheumatoid arthritis, may be useful in the future in patients who do not respond to high doses of oral corticosteroids. Magnussen and coworkers reported a series of four patients who were treated with an intraarticular injection of Kineret (anakinra), an IL-1 receptor antagonist, to treat persistent postoperative inflammation and scarring after ACL reconstruction. The injections were performed on postoperative days 36, 38, 41, and 97. The patients regained normal or nearly normal knee motion; three did not undergo other procedures but one did require debridement of scar tissue from the anterior compartment. Brown and associates also used intraarticular anakinra injections in eight patients, four of whom had chronic refractory arthrofibrosis and four of whom had “limited” arthrofibrosis with varying results. There was no control group or validated rating system used, and range of motion and joint effusion were based on “clinician estimate.” With so few data published to date, no recommendation can be made regarding the use of Kineret for knee arthrofibrosis at present.

Decorin is a human proteoglycan that has been shown in many experimental models to inactivate the effects of TGF-β and have a beneficial effect in arthrofibrotic tissues. Several years ago, Fukushima and coworkers reported that injection of decorin improved both the muscle structure and function of lacerated muscle to near-complete recovery in mice. More recently, Zhang and associates found that decorin reduced the effect of TGF-β in regard to intrinsic cell expansion in cultured synovial fluid explants from osteoarthritic knees. Abdel and colleagues assessed the effects of four intraarticular injections of decorin that were performed 8 weeks after an operation that was done to induce a knee joint contracture in rabbits. There was no significant effect on improving the flexion contractures. However, genetic analysis showed a significant alteration in several fibrotic genes that were likely associated in the arthrofibrosis process. These authors suggested that earlier injection might produce better results in terms of preventing joint contractures. Several other experimental models have demonstrated the effectiveness of decorin in inhibiting adhesion formation in intraarticular adhesions, skeletal muscle injury, kidneys, and lungs. There are no current human trials on the use of decorin for the prevention or treatment of knee arthrofibrosis.

Emami and coworkers investigated the effects of bevacizumab (Avastin), an agent that inhibits vascular endothelial growth factor, on preventing arthrofibrosis after surgery in a rabbit model. Given in either one or two injections, these authors reported reduced joint fibrosis, with better clinical and histological findings in the two-injection group. This was especially evident in the statistically significant reduced number of fibroblasts ( P = .002), inflammatory cells ( P = .01), vascularity ( P < .05), and collagenous matrix deposition ( P < .05).

Liang and colleagues applied topical hydroxycamptothecin, a chemotherapeutic drug, to rabbit knees in an attempt to prevent intraarticular adhesions after surgery in which cortical bone was removed from the femoral condyles. After 4 weeks, there was a significant reduction of adhesions in knees that received 1.0 mg/mL or 2.0 mg/mL compared with lower doses and controls ( P <.05). Another well-known chemotherapeutic agent, mitomycin C, was investigated by Yan and coworkers in the rabbit model. The agent was applied following removal of cortical bone from the femoral condyle and the effects evaluated 4 weeks later. There was a dose-dependent significant effect on suppression of the formation of adhesions, with 0.1 mg/mL required to reduce the adhesion score, the hydroxyproline content, and the fibroblast number compared with controls.

Wang and associates studied the effects of weekly intraarticular injections of 0.3 mL hyaluronic acid in immobilized rabbit knees performed for 8 weeks. Compared with the control group, the experimental group had significantly greater range of motion ( P = .002), a lower mean adhesion score ( P = .01), and lower mean collagen content ( P = .001).

In a small animal model, Bedair and associates showed that an angiotensin II systemic receptor blocker decreased fibrosis and enhanced muscle regeneration in acutely injured skeletal muscle and hypothesized that the blocker modulated TGF-β ₁ . Hagiwara and colleagues reported that the joint capsule has the potential to produce TGF-β ₁ and connective tissue growth factor, both of which play a role in causing fibrosis.

Two small case series used a single dose of low-level irradiation (700 to 800 rads) before repeat revision TKA in patients with long-standing arthrofibrosis. The idea was derived from the established use of low-dose irradiation for the prevention of heterotopic ossification before total hip arthroplasty in high-risk patients, and the potential effects of irradiation in preventing the differentiation of mesenchymal stem cells into osteoblastic precursors. Farid and associates employed this treatment method in 14 patients with severe idiopathic arthrofibrosis in an effort to suppress fibroblastic and fibroosseous proliferation postoperatively. Major improvements in knee motion and function were experienced in all but one patient, and there were no complications related to the irradiation therapy. Baier and colleagues described three patients who had similar treatment and marked improvements in knee flexion. However, these authors cautioned that the influence of radioactivity on soft tissues has not been investigated, and bony nonunion may occur. Both series were retrospective, and no control group or randomization procedures were performed.

Bosch and colleagues advanced the hypothesis that primary arthrofibrosis is the result of an immune response. Tissue samples obtained from 18 patients between 4 and 48 months from the triggering event to surgical debridement demonstrated synovial hyperplasia with fibrotic enlargement of the subintima and infiltration of inflammatory cells. These authors believed their findings supported an immune response as the cause of capsulitis leading to formation of diffuse scar tissue. The clinical implication was to avoid further surgery or aggressive manipulation while patients have acute capsulitis and to wait 3 to 6 months for the inflammatory response to resolve before debridement is considered. As is discussed, too often, there is a delay in treatment that results in a permanent limitation of knee motion. Unfortunately, even though a delay in treatment may allow tissues to “quiet down,” the resulting fibrotic tissue that forms becomes well organized and dense and treatment is thus more difficult. It is anticipated that in the future, chemical agents will become available to treat fibrotic response to injury by controlling elevated levels of myofibroblast upregulators and fibrogenic growth factors.

Prevention