Arthrofibrosis of the Knee

David Magit, MD Abstract Andy Wolff, MD Better understanding of surgical timing, improved surgical Karen Sutton, MD technique, and advanced rehabilitation protocols has led to Michael J. Medvecky, MD decreased incidence of motion loss after anterior cruciate ligament injury and reconstruction. However, motion loss from high-energy, multiligament injuries continues to compromise functional outcome. Prevention, consisting of control of inflammation and early motion, remains the key element in avoiding motion loss. However, certain techniques, such as manipulation under anesthesia in conjunction with arthroscopic lysis of adhesions, are reliable treatment options. Open surgical débridement is rarely necessary and should be considered only as a salvage procedure. A greater understanding of the pathogenesis of arthrofibrosis and related inflammatory mediators may result in novel therapies for treating the patient with motion loss.

Dr. Magit is Orthopaedic Surgeon, The Greenwich Sports and Shoulder oss of knee motion is a devastat- tients with arthrofibrosis.1 Service, Orthopaedic and Neurosurgery Ling consequence of both single- Arthrofibrosis represents a wide Specialists PC, Greenwich, CT. Dr. and multiligamentous injury and spectrum of disease, ranging from lo- Wolff is Fellow, Steadman Hawkins their reconstruction. Increased recog- calized to diffuse involvement of all Clinic, Vail, CO. Dr. Sutton is Resident, nition of this problem in the past two compartments of the knee and of the Department of Orthopaedics and decades has led to better prevention extra-articular soft tissues. Preven- Rehabilitation, Yale University School of and improved management of these tion of motion loss remains essential Medicine, New Haven, CT. Dr. Medvecky injuries. Despite these advances, to successful outcome. In the patient is Assistant Professor, Department of however, motion loss remains a prob- who experiences motion loss despite Orthopaedics and Rehabilitation, Yale lematic consequence of knee liga- preventive measures, treatment op- University School of Medicine. ment injury. tions include static or dynamic brac- None of the following authors or a The incidence of motion loss var- ing, manipulation under anesthesia, member of their immediate families has ies according to the degree of injury. and arthroscopic or open débride- received anything of value from or owns Motion loss is less severe after single- ment. In recalcitrant cases, arthrode- stock in a commercial company or ligament, low-energy injury than af- sis in the older patient or total knee institution related directly or indirectly to ter high-energy, multiligament in- arthroplasty may be required. the subject of this article: Dr. Magit, Dr. jury. The etiology of motion loss is Wolff, Dr. Sutton, and Dr. Medvecky. multifactorial, involving a combina- Normal Knee Motion Reprint requests: Dr. Magit, The tion of mechanical and biologic fac- Greenwich Sports and Shoulder tors. Major risk factors include tech- Normal knee motion involves a Service, Orthopaedic and Neurosurgery nical errors during intra-articular combination of longitudinal axial, ro- Specialists PC, 6 Greenwich Office ligament reconstruction and extra- tation, varus/valgus angulation, and Park, 10 Valley Drive, Greenwich, CT articular procedures, injury severity, flexion/extension arcs. Flexion and 06831. timing of surgery, delayed postoper- extension can be categorized into ative physical rehabilitation, hetero- three sub-arcs: terminal extension, J Am Acad Orthop Surg 2007;15:682- topic ossification, prolonged immo- active function, and passive flexion. 694 bilization, infection, and complex The arc of terminal extension, also Copyright 2007 by the American regional pain syndrome. Recently, called “screw home,” begins at the Academy of Orthopaedic Surgeons. authors have begun to examine pos- limit of passive extension. This arc sible genetic differences among pa- moves from 10° of flexion to 5° of hy-

682 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al

Figure 1

Knee motion can be measured by the standard goniometric method (A) or by measuring the heel-height difference (B). (Panel A reproduced and panel B adapted from Schlegel TF, Boublik M, Hawkins RJ, Steadman JR: Reliability of heel-height measurement for documenting knee extension deficits. Am J Sports Med 2002;30:479-482.)

perextension. The arc of terminal ex- gait. Without flexion beyond 125°, Measuring Motion Loss tension is rarely used in normal gait the patient may report an inability but is thought to allow for quadriceps to squat. Small flexion deficits in Proper care for the patient with re- muscle relaxation during the stance the athlete, however, can produce cent ligamentous knee injury or phase. The arc of active function marked changes in performance. reconstruction requires accurate de- ranges from 10° to approximately Loss of flexion ≥10° can affect run- tection of motion loss. The most 120°, which covers the range needed ning speed. Severe flexion deficits common method involves placing a for most activities of daily living, in- <90° affect the ability of even the goniometer over the lateral knee cluding sitting and stair climbing. most sedentary patient to sit or joint line in the midsagittal position, The arc of passive flexion begins at climb stairs. using the greater trochanter and lat- approximately 120° and continues to eral malleolus as reference points the passive limit of an applied exter- Extension Loss (Figure 1, A). Several studies have nal force. Typically, passive flexion is Extension loss is poorly tolerated demonstrated high inter- and in- 140° in men and 143° in women. and can be more difficult to manage traobserver reliability with this However, flexion to 165° is seen in than loss of flexion. As little as 5° of method.4 A second method involves societies in which full kneeling or extension loss can produce a notice- measuring the heel-height differ- squatting is common, such as in Ja- able limp during ambulation, strain ence, which is done with the patient pan, India, and the Middle East.2 the quadriceps muscle, and contrib- in the prone position (Figure 1, B). In ute to patellofemoral pain. During general, 1 cm of heel-height differ- ence correlates to 1° of knee flexion Flexion and Extension weight bearing on a flexed knee, the contracture. This technique may be Deficits quadriceps muscle force required to stabilize the knee is 75% of the load helpful in detecting subtle degrees of Flexion Loss on the femoral head at 15° of flexion, motion loss (<10°).5 The functional effects of knee 210% at 30°, and 410% at 60°.3 With motion loss vary depending on pa- increased joint contact pressure, the Classification of Motion tient activity. In general, flexion to clinical consequences are increased Loss 125° is adequate for completing ac- quadriceps muscle activity and fa- tivities of daily living and usually tigue, and, ultimately, patellofemo- Determining the true incidence of does not adversely affect normal ral arthrosis. motion loss can be difficult, given

Volume 15, Number 11, November 2007 683 Arthrofibrosis of the Knee

16 Table 1 motion loss. Sisto and Warren reported motion problems in 6 of 20 Sprague Pathoanatomic Classification of Motion Loss patients (30%). Shapiro and Freed- 17 Group Pathoanatomy man reported a 57% incidence of postdislocation motion loss. In this 1 Discreet bands or a single sheet of adhesions traversing the study, patients required an additional suprapatellar pouch procedure to restore motion. 2 Complete obliteration of the suprapatellar pouch and peripatellar gutters with masses of adhesions 3 Multiple bands of adhesions or complete obliteration of the Risk Factors for Motion suprapatellar pouch with extracapsular involvement with Loss bands of tissue from proximal patella to anterior femur Technical Errors Reproduced with permission from Sprague NF III, O’Connor RL, Fox JM: Proper graft placement in ACL re- Arthroscopic treatment of postoperative knee fibroarthrosis. Clin Orthop Relat Res 1982;166:165-172. construction is essential to reducing motion loss.18 Graft placement ante- rior to the native ACL insertion on Table 2 the tibia results in impingement on Classification of Motion Loss of the Knee Based on Deviation From Full the roof of the intercondylar notch in Flexion and Extension extension. Lateral placement on the tibia produces impingement on the Group Extension Flexion Severity lateral wall of the intercondylar 1 <5° >110° Mild notch. Placement too far anterome- 2 5°-10° 90°-110° Moderate dially has been shown to limit flex- 19 3 >10° <90° Severe ion. On the femoral side, the most common error is graft placement too Reproduced with permission from Del Pizzo W, Fox JM, Friedman ML, et al: far anterior, which causes excessive Operative arthroscopy for the treatment of arthrofibrosis of the knee. Contemp Orthop 1985;10:67-72. strain on the graft, leading to limited flexion and potential graft failure.20,21 Other sources of graft impinge- the number of available classification extension loss >10° with patella in- ment include the intercondylar schemes. Sprague et al6 defined mo- fera. notch and adjacent posterior cruci- tion loss based on a pathoanatomic ate ligament (PCL). Whereas moder- distribution (Table 1). Later classifi- Incidence of Motion ate notch impingement can cause cation schemes focused on knee Loss pain, effusion, and extension loss, range of motion (ROM). Del Pizzo et severe impingement can cause abra- al7 graded motion loss by evaluating Early studies reported that as many sion and graft failure.22 Contraction deviation from full flexion and exten- as 35% of patients who underwent of the quadriceps during knee exten- sion, with severe motion loss consid- acute anterior cruciate ligament sion results in increased impinge- ered to be >10° from full extension (ACL) reconstruction or repair devel- ment; thus, a notchplasty should be and <90° of flexion (Table 2). Blauth oped loss of knee motion.10 However, performed to provide 3 mm of clear- and Jaeger8 graded motion loss based with advances in surgical technique ance between the graft and the inter- on full arc of motion as grade I (mild, and accelerated rehabilitation proto- condylar roof. Impingement of the ROM >120°), grade II (moderate, cols, the incidence has markedly de- ACL graft on the PCL also may lim- ROM 80° to 120°), grade III (severe, creased, to as low as 0% to 4%.11-14 it flexion when the angle of the tib- ROM 40° to 80°), and grade IV (ex- Motion loss is more common with ial tunnel is too steep (80°).23 treme, ROM < 40°). The most recent multiligamentous high-energy injury classification system was introduced than with single-ligament low-energy Graft Tension by Shelbourne et al,9 who compared injury. Noyes et al15 reported a 23% The relationship between motion motion loss on the affected side with incidence of motion loss in a group of loss and graft tension remains con- the normal contralateral limb. The patients with concomitant ACL re- troversial. Some have suggested that authors identified four types: 1, nor- construction and medial collateral lig- increased graft tension results in ex- mal flexion and extension loss <10°; ament (MCL) repair. Traumatic knee cess constraint on the joint.24 How- 2, normal flexion and extension loss dislocation, which produces variable ever, in their biomechanical study, >10°; 3, flexion loss >25° and exten- patterns of ligamentous instability, Markolf et al25 showed that although sion >10°; and 4, flexion loss >30° and results in the highest incidence of a high degree of graft pretension may

684 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al lead to fraying of the graft over the not at increased risk for developing Figure 2 femoral tunnel, it does not lead to a arthrofibrosis. loss of full knee extension. Con- versely, undertensioning the graft Soft-tissue Injury may produce anteroposterior laxity The magnitude of soft-tissue and with subsequent instability, poor bony injury associated with ligament graft healing, and failure. injury is directly related to the patho- genesis of motion loss. Typical mech- Graft Choice anisms include sports-related trauma The relationship between exten- and knee dislocation secondary to sion loss and graft choice has been high-energy motor vehicle accidents. questioned. In a prospective, nonran- The global nature of multiple trau- domized study comparing hamstring matic injuries often results in de- and patellar tendon ACL reconstruc- layed surgical timing, compromised tions, Pinczewski et al26 reported that wound healing, alterations in surgi- 31% of the patellar tendon group and cal approach, and delayed mobiliza- 19% of the hamstring group had ex- tion. Each factor plays a major role in tension deficits at 5 years postoper- the pathogenesis of motion loss. atively. These results were not sta- tistically significant. Sajovic et al27 Timing of Surgery found no significant difference in Timing of surgery remains con- ROM in their recent prospective, ran- troversial. Many authors differenti- Lateral radiograph demonstrating domized trial comparing hamstring ate acute from delayed reconstruc- posterior heterotopic ossification after and patellar tendon ACL reconstruc- tion at 3 weeks. Delayed surgery multiligamentous reconstruction. tions at 5-year follow-up. The au- allows time for soft-tissue healing, thors concluded that graft choice is resumption of full ROM, and im- not related to the development of proved strength. tions, they did find a higher overall motion loss after ACL reconstruc- Shelbourne et al30 reviewed 169 reoperation rate in the acute cases. tion. young athletes after acute ACL re- They recommend elective subacute construction. Patients in whom reconstruction, provided that mo- Extra-articular Procedures ACL ligament reconstruction was tion and effusion goals have been at- Increased trauma around the knee done within 1 week of injury had a tained. during open arthrotomy has been statistically significant increased in- Sterett et al32 found no association cited as a risk factor for motion loss cidence of arthrofibrosis (P < 0.05) between surgical timing of ligament after ACL reconstruction. Harner et compared with patients in whom reconstruction and incidence of mo- al11 theorized that concomitant ACL reconstruction was delayed tion loss. The authors cited minimal MCL repair during ACL reconstruc- >3 weeks. preoperative active ROM (0°-120°), tion produced motion loss via dis- Harner et al11 found a 37% rate of active quadriceps control, and the ruption of the medial capsule and re- postoperative motion loss after ability to perform a straight-leg raise sultant interference with normal acute ACL reconstruction (16 of 43 as essential determinants of a suc- knee kinematics. Alternatively, ad- patients) versus 5% after chronic re- cessful outcome. However, 8% of pa- ditional trauma to the capsule may construction. Wasilewski et al31 tients in the acute group required a cause an elevated fibrotic response evaluated 87 patients who under- second surgery for symptomatic scar as well as increased pain and swell- went reconstruction for ACL injury tissue (6 of 80 patients). ing, resulting in quadriceps inhibi- at the acute (<1 month), subacute (1 Given the multiple techniques tion. With extensive extra-articular to 6 months), and chronic (>6 used for reconstruction, variable def- soft-tissue dissection or trauma, the months) stage. Arthrofibrosis was initions of timing and classification, development of soft-tissue calcifica- found in 22% of knees managed and lack of prospective studies, firm tion or myositis ossificans also may with ACL reconstruction in the conclusions regarding this ongoing result in secondary loss of motion28 acute stage, in 0% managed in the debate remain elusive. The key (Figure 2). Conversely, Cosgarea et subacute stage, and in 12.5% man- factor remains understanding the al29 found that patients who under- aged in the chronic stage. Although mechanism and severity of injury went intra-articular procedures (eg, Bach et al24 did not find differences as they relate to the preoperative meniscal repair, partial meniscecto- in postoperative motion between level of inflammation, strength, and my) after ACL reconstruction were acute and chronic ACL reconstruc- ROM.

Volume 15, Number 11, November 2007 685 Arthrofibrosis of the Knee

Immobilization ing. Blood samples were taken from Clinical Findings Immobilization after knee liga- each patient, and DNA was evalu- ment injury and reconstruction re- ated for loci human leukocyte anti- The term arthrofibrosis describes a mains a well-established risk factor gen (HLA)–A, –B, –C, and –DQB1. wide spectrum of conditions, all of for motion loss. Hooper and Wal- Compared with a control group, pa- which ultimately result in motion ton33 reported that 46% of patients tients with arthrofibrosis were less loss. The spectrum of involvement who began motion 2 weeks after likely to have allelic group HLA- ranges from purely localized to vastly surgery experienced loss of motion. Cw*07 and more likely to have HLA- diffuse, with varying degrees of intra- 34 In comparison, Zarins and Rowe Cw*08. It is unknown whether these and extra-articular extension. found nearly complete return of alleles represent potential increased knee extension in their patients, Localized Intra-articular risk for arthrofibrosis by either in- with return of motion after 1 week Extension creased susceptibility or decreased of cast immobilization following lig- Fullerton and Andrews36 first defense mechanisms. ament reconstruction. Complica- reported hypertrophy of the ACL tions stemming from prolonged cast- graft and protrusion of bone at the ing and the benefits of early motion Pathophysiology tibial attachment site. Jackson and have led to a shift toward a shorter Schaefer37 further characterized this period of immobilization. Increased understanding of the mo- finding, now known as “cyclops syn- lecular pathways responsible for or- drome,” as extension loss after ACL Infection thopaedic disease has provided new reconstruction caused by a mechan- Infection may cause motion loss insight into the pathogenesis of mo- ical block from hypertrophic fibrous after knee ligament surgery. Intra- tion loss. Tissue organization and tissue attached to the tibial insertion articular inflammatory mediators re- homeostasis depend on constant sig- point of the ACL graft (Figure 4). sponding to infection cause joint naling of cytokines and locally act- Various explanations of the cy- synovitis and toxic degeneration to ing growth factors. These cytokines clops syndrome have been described the articular surface. Motion loss is coordinate cell growth, differentia- in the literature. Anterior graft caused directly by fibrous scar for- tion, and programmed cell death via placement can cause repetitive trau- mation produced by local cytokine constant signaling between local ma of graft against bone, leading to activation, and indirectly by pain cells (paracrine) as well as among impingement on the top of the notch produced by joint swelling and irrita- themselves (autocrine). and graft hypertrophy with fibrous tion. Transforming growth factor-β tissue proliferation. Fibrous tissue (TGF-β), which is released by plate- originating from the drilling debris Complex Regional Pain lets, plays a critical role in the pro- of the tibial tunnel is another possi- Syndrome cess of tissue repair.35 At the site of ble cause. Fibroproliferative tissue Complex regional pain syndrome injury, TGF-β and platelet-derived may develop in the patient who does inhibits effective patient participa- growth factor initiate a cascade of not achieve full extension in the ear- tion in proper postoperative rehabil- ly postoperative period because of events resulting in the production of itation. Resultant quadriceps shut- pain or immobilization. extracellular matrix proteins and down, muscle atrophy, and avoidance protease inhibitors as well as inhibi- of mobility establish a vicious cycle, Localized Intra-articular tion of proteolytic enzyme produc- leading to the development of intra- With Extra-articular tion. Formation of extracellular ma- articular adhesions and arthrofibro- Extension sis. Patella infera is a secondary con- trix occurs at the site of injury, Paulos et al38 were among the first cern in these patients. consisting of an aggregation of col- to describe infrapatellar contraction lagen, fibronectin, and proteogly- syndrome (IPCS), which manifests Genetic Risk Factors cans. With an increase in local as significant reduction of both flex- Some patients develop motion concentration, the autoregulatory ion (>25°) and extension (>10°) with loss despite efforts to reduce or elim- mechanism of TGF-β results in feed- an associated decrease in patellar inate the aforementioned risk factors. back inhibition. Overexpression of mobility, characterized as patellar This has led researchers to consider TGF-β can result in progressive dep- entrapment. The cause of IPCS was whether there may be a genetic pre- osition of matrix and tissue fibrosis. attributed to either exaggerated disposition to arthrofibrosis after in- TGF-β overexpression leads to fibro- pathologic fibrous hyperplasia of the jury. Skutek et al1 evaluated 17 pa- sis and exists in the other organ sys- anterior soft tissues beyond normal tients with arthrofibrosis after ACL tems, such as the kidney, liver, and healing or certain risk factors asso- reconstruction with autologous graft- lung (Figure 3). ciated with knee surgery (eg, poor

686 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al graftisometry,immobilization,mus- Figure 3 cle weakness). Paulos et al38 stratified patients with IPCS into prodromal, active, and residual stages. A patient in the prodromal stage demonstrates peri- articular inflammation and edema, quadriceps weakness and lag, poor knee extension, painful ROM, and tenderness over the patellar tendon with a decrease in patellar excur- sion. A patient in the active stage shows a dramatic decrease in patel- lar mobility, marked quadriceps at- rophy, worsening knee motion, and fat pad induration with a rigid patel- lar tendon. A patient in the residual stage presents with peripatellar and retinacular tissues that are more supple than in the active stage, with resulting marked quadriceps atrophy and loss of knee flexion and exten- sion. The critical component of the residual stage is the progression of patellofemoral arthrosis and patella infera (Figure 5). Comparison of the regulatory response of transforming growth factor-β (TGF-β)in In a series of patients with IPCS, a normal (A) and a pathologic (B) response to tissue repair. (Reproduced with 38 Paulos et al initially used conser- permission from Border WA, Noble NA: Mechanisms of disease: Transforming vative treatment, including inflam- growth factor [beta] in tissue fibrosis. N Engl J Med 1994;331:1286-1292.) mation control, followed by gentle

Figure 4 Figure 5

Cyclops lesion (arrow) located anterior to the intercondylar notch. (Repro- duced with permission from Ahn JH, Yoo JC, Yang KS, Kim JH, Wang JH: Second-look arthroscopic findings of 208 patients after ACL reconstruction. Lateral radiographs demonstrating patella infera after patellar tendon repair. The Knee Surg Sports Traumatol Arthrosc Blackburne-Peel index is shown on presentation (A/B = 0.62) (A) and after gait 2007;15:242-248.) retraining and quadriceps reactivation (A/B = 1.18) (B).

Volume 15, Number 11, November 2007 687 Arthrofibrosis of the Knee physical therapy, arthroscopic dé- tive and passive knee motion in the tension deficits ≥2° (54%), whereas bridement, and manipulation. Pa- immediate postoperative period (413 only 2 of the 22 patients braced in tients for whom these treatments of 443 patients). Of the remaining 30 hyperextension had extension defi- were unsuccessful underwent open patients, 23 were placed in an early cits (9%). The authors concluded arthrotomy with débridement; pa- postoperative treatment program, that use of a knee brace in hyperex- tients with patella infera >8 mm un- which included hyperflexion and hy- tension for at least 3 weeks after derwent tibial tubercle osteotomy. perextension exercises and serial ex- ACL reconstruction is an effective tension casting (Figure 7). In eight means of preventing postoperative Diffuse Intra-articular/ patients for whom cryotherapy, anti- extension loss. Extra-articular Extension: inflammatory medications, eleva- A Molecular Basis tion, and compression were ineffec- Manipulation Under An exaggerated synovial and in- tive, the authors advocated the use Anesthesia flammatory response may contribute of oral steroids. Intensive inpatient Dodds et al12 reported positive re- to arthrofibrosis and result in activa- physical therapy, manipulation un- sults with knee manipulation in 42 tion and proliferation of fibroblastic der anesthesia, and arthroscopic dé- knees with significant flexion or ex- cells, which produce elevated levels bridement were performed in refrac- tension deficits after ACL recon- of type VI collagen and extracellular tory cases. Overall, 98% regained struction with a patellar tendon matrix proteins.39 With heightened full knee motion, 2% had minor graft. Outcome after manipulation fibroblastic response, tissue shrink- limitations in extension, <1% re- was related to the severity of exten- age (ie, scar contraction) occurs both quired arthroscopic release of adhe- sion loss before manipulation; pa- within and outside the joint. A re- sions, and no patient developed per- tients with greater extension deficits cent study indicates that specific manent arthrofibrosis. achieved less overall final extension. alpha-smooth muscle actin–contain- The authors recommended manipu- ing fibroblastic cells play a critical Postoperative Bracing lation within 12 weeks of ACL re- role in tissue contraction associated Although controversial, the use of construction. More recent reports with wound healing.40 In arthrofi- a rehabilitation brace after ACL re- recommend manipulation within 4 brotic tissue, the authors found sig- construction is common. In a recent to 12 weeks of reconstruction when nificantly higher numbers of syno- survey of members of the American <90° of flexion is achieved despite vial cells (P ≤ 0.05), significantly less Orthopaedic Society for Sports Med- the use of physical therapy.41 vessel density (P < 0.001), and tenfold icine, 85% of respondents prescribed Overaggressive or significantly higher expression of alpha-smooth a brace for an average of 3.8 weeks delayed manipulation should be muscle actin–positive fibroblastic after ACL reconstruction.42 Whether avoided to prevent complications cells (P ≤ 0.001) compared with con- bracing can prevent loss of extension such as chondral damage, distal fe- trol tissue. remains a question. Feller et al43 mur or patella fracture, patellar ten- found no advantage in restoring knee don rupture, stimulation of myositis Treatment extension with the use of an exten- ossificans of the quadriceps, and os- sion brace. However, in a recent pro- sification of the MCL. Immediate Postoperative spective study comparing bracing Motion from 0° to 90° to bracing locked in Manipulation and Prevention is the most effective full extension during the first week Arthroscopy means of avoiding the devastating after ACL reconstruction, Melegati With delayed presentation (>6 consequences of motion loss after et al44 found that patients braced in weeks), manipulation under anes- knee ligament injury and surgery. full extension for the first week post- thesia is most effective when Improved postoperative rehabilita- operatively had lower heel-height performed in conjunction with ar- tion protocols, including the use of differences at 4- and 8-week follow- throscopic lysis of adhesions. Arthro- immediate passive and progressive up than did patients whose brace scopic management allows access to active ROM, have addressed the was unlocked twice daily for physio- focal lesions (eg, ACL nodules, loose harmful consequences of prolonged therapy. bodies) and is helpful in addressing immobilization (Figure 6). In another randomized, prospec- cases of severe diffuse arthrofibrosis Noyes et al41 advocated immedi- tive study, Mikkelsen et al45 com- refractory to closed methods as well ate knee motion and early interven- pared patients braced in full exten- as in avoiding potential catastrophic tion to prevent arthrofibrosis after sion (0°) with patients braced in complications associated with ma- ACL reconstruction. In a prospective hyperextension (−5°) after ACL re- nipulation alone. study, 93% of patients regained full construction. At 3 months, 12 of the Prior to arthroscopy, accurate ROM (0º to 135º) with the use of ac- 22 patients in full extension had ex- clinical assessment of motion loss

688 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al

Figure 6

POD 0-7 • CPM (0°-90°) • Quadriceps isometric strengthening • WBAT in ACL brace (-5°) • Patellar mobilization • Cryotherapy

Week 1-3 Consider • Treat infection or CRPS • Infection • Continue progressive ROM, quadriceps

Preventive phase • CRPS strengthening, patellar mobilization • Error in technique • Check radiographic screw placement

Focused flexion and extension therapy

Flexion exercise program Week 4 Flexion loss • Rolling stool exercises • Towel pulls • Gentle manipulation under anesthesia

Extension exercise program • Towel rolls Extension loss • Dynamic extension splinting • Extension casting

Week 6 Flexion loss Arthroscopic management • Suprapatellar pouch • Arthroscopic débridement • Patellofemoral joint • Arthroscopic medial/lateral retinacular • Anterior femoral-side release graft placement • Graft débridement/excision

Extension loss • Cyclops lesion Arthroscopic management • Medial/lateral gutter • Débridement of cyclops • Gutter débridement Intervention phase • Anterior tibial-side graft placement • Graft débridement/excision

Flexion and extension loss Arthroscopic management • Intercondylar notch impingement • Notchplasty • Anterior interval • Débridement/excision of graft • Anterior interval release

Open intra-articular management • Anterior fat pad excision Open intra-articular management >8 weeks Failed arthroscopic management • Reestablish anterior interval • Excision of heterotopic ossification • Assessment of graft isometry • Limited quadricepsplasty and • Posterior capsular release arthroscopic débridement

Salvage procedures Failed open management • Osteotomy • Arthrodesis

Treatment algorithm for arthrofibrosis highlighting prevention (top) and intervention (bottom) postoperatively. ACL = anterior cruciate ligament, CPM = continuous passive motion, CRPS = complex regional pain syndrome, POD = postoperative day, ROM = range of motion, WBAT = weight bearing as tolerated

aids in identifying the location of ar- ment of the intercondylar notch can of the posterior capsule. throfibrosis. Loss of knee flexion of- affect both flexion and extension. Kim et al46 described a systemat- ten indicates involvement of the Extension loss can result from intra- ic approach when performing arthro- suprapatellar pouch, patellofemoral articular nodules adjacent to the tib- scopic débridement of an arthrofi- joint, or anterior interval. Involve- ial graft insertion and arthrofibrosis brotic knee. The use of regional

Volume 15, Number 11, November 2007 689 Arthrofibrosis of the Knee

Figure 7

When conventional postoperative physical therapy modalities are unsuccessful, the patient begins hyperextension exercises using pillows or towel rolls placed under the ankle (A). Alternately, the patient is managed with serial cylinder casting (B) followed by hyperflexion exercises, such as rolling chair motions (C). (Reproduced with permission from Shelbourne KD, Patel DV: Treatment of limited motion after anterior cruciate ligament reconstruction. Knee Surg Sports Traumatol Arthrosc 1999;7:85-92.)

anesthesia can effectively manage is essential. Depending on the sever- sions. Only 2% of patients required perioperative pain and facilitate ity of the scarring, the graft may open surgical management. postoperative rehabilitation.47 Prior need further débridement, release, or Surgery often demands extensile to portal placement, capsular disten- excision. Once complete, the knee incisions and large parapatellar ar- tion is achieved by saline injection should be ranged and motion reas- throtomies. Similar to arthroscopic into the suprapatellar pouch. sessed. débridement, a systematic approach The suprapatellar pouch is rees- Persistent loss of extension usu- should be used. Débridement in- tablished first, followed by the medi- ally indicates posterior capsular cludes anterior fat pad excision and al and lateral gutters. The anterior involvement. When necessary, a lim- reestablishment of the pretibial re- interval is identified by releasing the ited open posterior release is per- cess, as well as medial and lateral infrapatellar fat pad from the anteri- formed from a posteromedial and retinacular release. Evaluation of or tibia, allowing for reestablish- posterolateral incision.48 the ACL graft isometry follows, and ment of the pretibial recess. Medial débridement or excision are often and/or lateral retinacular release Open Surgical Techniques necessary. Finally, release of poste- may be required in the patient with Open surgical management of rior capsular soft tissues from both reduced patellar mobility or a tight arthrofibrosis represents a salvage the femur and tibia is performed patellofemoral joint. option for the rare knee that is re- by careful subperiosteal dissection. Once in the intercondylar notch, fractory to closed and arthroscopic Failure to recognize and remove any the surgeon must evaluate for graft procedures. Of the 207 knees treated extra-articular causes of motion impingement caused by notch ste- for motion loss after knee ligament loss, such as myositis ossificans, nosis. If present, a notchplasty is per- surgery in a study by Noyes et al,15 heterotopic ossification, or soft-tis- formed. Scar tissue, ACL or bony 202 responded to early rehabili- sue calcifications, can lead to pro- nodules, and loose bodies are re- tation, manipulation under anes- longed motion loss after débride- moved. Evaluation of graft integrity thesia, or arthroscopic lysis of adhe- ment.

690 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al

Figure 8

Salvage procedures, including mini-incision quadricepsplasty (A), the DeLee osteotomy for patella infera (B), and arthrodesis (C) for severe diffuse arthrofibrosis. X = new functional patellar tendon length, Y = patellar tendon length, Z = length of proximal translation of osteotomized tubercle. (Panel A adapted with permission from Wang JH, Zhao JZ, He YH: A new strategy for severe arthrofibrosis of the knee: A review of twenty-two cases. J Bone Joint Surg Am 2006;88:1245-1250. Panel B reproduced with permission from Richmond JC: Arthrofibrosis following knee surgery [arthroscopic/open], in Malek MM, Farelli G, Johnson D, Johnson D [eds]: Knee Surgery: Complications, Pitfalls and Salvage. New York, NY: Springer- Verlag, 2001, p 456.)

Salvage Procedures procedures. Arthrodesis is a final sal- and Lysholm II scores improved an Recently, Wang et al49 reported vage procedure in the rare patient for average of 35 points. However, in their series of mini-incision quadri- whom no other management option five of eight patients, degenerative cepsplasty with arthroscopy for pa- is successful (Figure 8, C). changes were seen on radiographs at tients who presented with severely a mean follow-up of 57 months. All arthrofibrotic knees with major Postoperative Outcomes patients were able to return to sports, intra- and extra-articular involve- Postoperative outcomes are diffi- but only one was able to achieve her ment. The authors used a short, lon- cult to interpret, given the different preinjury level of function. gitudinal incision adjacent to the su- classification systems, variable in- perolateral corner of the patella to jury patterns, combined surgical Summary perform a staged release, beginning treatment patterns, and different with the lateral retinaculum and ex- outcome measures. A summary of The incidence of motion loss after tending to the suprapatellar pouch outcomes after surgical treatment ACL injury and reconstruction has and across to the medial patellar ret- of localized intra-articular lesions, decreased because of greater under- inaculum. The last two stages con- more diffuse intra-articular arthrofi- standing of surgical timing, im- sisted of transection of the vastus in- brosis, and arthrofibrosis extending proved surgical technique, and ad- termedius and lengthening of the into the extra-articular soft tissues is vanced rehabilitation protocols. In quadriceps tendon (Figure 8, A). Fol- given in Table 3. high-energy injuries and multiliga- lowing release, arthroscopy was There are few reports on the re- ment injuries, however, the inci- performed to complete the intra- sults of severely arthrofibrotic knees dence of motion loss continues to articular portion of the procedure. that have undergone multiple surgi- cause problems for the treating sur- Flexion improved from 5° to 45° pre- cal treatments. Millett et al47 re- geon. Prevention through early mo- operatively to 120° to 150° postoper- ported on eight patients with an tion remains the key element in atively. average preoperative ROM of 62.5° avoiding motion loss. However, Proximal sliding tibial tubercle os- who underwent open débridement. techniques such as manipulation teotomy (DeLee osteotomy, Figure 8, After radical open débridement under anesthesia used in conjunc- B) is possible in the patient with pa- (mean follow-up, 57 months), pa- tion with arthroscopic lysis of adhe- tella infera that does not respond to tients gained an average of 62° of mo- sions are reliable treatment options. closed, arthroscopic, or open surgical tion. Patient satisfaction was high, Open surgical débridement is rarely

Volume 15, Number 11, November 2007 691 Arthrofibrosis of the Knee

Table 3 Outcomes After Localized Arthrofibrosis, Diffuse Arthrofibrosis, and Arthrofibrosis With Intra- and Extra-articular Extension

Index Average Average No. of Extent of Surgery/Injury Preoperative Postoperative Study Patients Arthrofibrosis (No. of Patients) ROM Surgery ROM

Jackson 13/230 Cyclops lesion ACLR with BPTB 16° to 103° Arthroscopic 4° to 138° and (13), ITB débridement Schaefer37 tenodesis (4), of nodule, MCL repair (1) MUA

Fisher and 35/959 Cyclops lesion ACLR with BPTB 6° to 119° Arthroscopic 2° to 135° Shelbourne14 (42) débridement of nodule

Klein et 46 Diffuse intra- ACLR (34), 10.4° Arthroscopic 1.7° extension al50 articular meniscal lesion extension débridement loss, 7.9° (4), ACL/PCL loss, 31.6° flexion loss (1), patel- flexion loss lofemoral (6), HTO (1)

Shelbourne 9 Diffuse intra- ACLR with BPTB 23° to 113° Arthroscopic 2° to 130° and articular (8) or HS (1) débridement, Johnson51 MUA

Hasan et 17/342 Diffuse intra- ACLR (17) 10° to 123° Arthroscopic 3° to 131° al52 articular débridement, MUA

Harner et 27/244 Diffuse intra-/ ACLR (27), 13° to 124° Arthroscopic 3° to 126° al11 extra-articular extra-articular débridement procedure (11), (14), open ITB tenodesis (8), débridement MCL repair or (6), MUA POL reefing (12) alone (1)

Cosgarea 37 Diffuse intra-/ ACLR (23), ACL 14° to 120° Arthroscopic 3° to 142° et al53 extra-articular repair (12), PCL débridement, repair (1), MCL percutaneous repair (1) LOA, lateral release, notchplasty (30), open débridement (7)

Millett et 8 Diffuse intra-/ Meniscectomy (1), 19° to 81° Arthroscopic 1° to 125° al47 extra-articular ACLR (4), débridement ORIF with (6), MUA PCL/MCL/LCLR alone (2), (1), PCL/MCLR open (1), ACL/ débridement PCLR (1) and soft-tissue release

ACLR = anterior cruciate ligament reconstruction, BPTB = bone-patellar tendon-bone, HS = hamstring, HTO = high tibial osteotomy, ITB = iliotibial band, LCLR = lateral collateral ligament reconstruction, LOA = lysis of adhesions, MCL = medial collateral ligament, MCLR = medial collateral ligament reconstruction, MUA = manipulation under anesthesia, ORIF = open reduction and internal fixation, PCL = posterior collateral ligament, PCLR = posterior cruciate ligament reconstruction, POL = posterior oblique ligament, ROM = range of motion

692 Journal of the American Academy of Orthopaedic Surgeons David Magit, MD, et al necessary and should be considered cruciateligamentreconstruction. Am 23. Simmons R, Howell SM, Hull ML: Ef- only as a salvage procedure. Primary J Sports Med 1996;24:857-862. fect of the angle of the femoral and tib- motion loss occurs in the patient in 10. Strum GM, Friedman MJ, Fox JM, et ial tunnels in the coronal plane and in- al: Acute anterior cruciate ligament cremental excision of the posterior whom arthrofibrosis develops de- reconstruction: Analysis of complica- cruciate ligament on tension of an an- spite consideration of all known risk tions. Clin Orthop Relat Res 1990; terior cruciate ligament graft: An in factors. A greater understanding of 253:184-189. vitro study. J Bone Joint Surg Am the pathogenesis of arthrofibrosis 11. Harner CD, Irrgang JJ, Paul J, Dearwa- 2003;85:1018-1029. and of inflammatory cytokines, such ter S, Fu FH: Loss of motion after 24. Bach BR Jr, Jones GT, Sweet FA, Hager anterior cruciate ligament reconstruc- CA: Arthroscopy-assisted anterior as TGF-β, may lead to novel thera- tion. Am J Sports Med 1992;20:499- cruciate ligament reconstruction us- pies for managing this difficult prob- 506. ing patellar tendon substitution: Two- lem. 12. Dodds JA, Keene JS, Graf BK, Lange to four-year follow-up results. Am J RH: Results of knee manipulations af- Sports Med 1994;22:758-767. References ter anterior cruciate ligament recon- 25. Markolf KL, Burchfield DM, Shapiro structions. Am J Sports Med 1991;19: MM, Davis BR, Finerman GA, Slaut- Evidence-based Medicine: Level I/II 283-287. erbeck JL: Biomechanical conse- prospective, randomized studies are 13. Shelbourne KD, Nitz P: Accelerated quences of replacement of the anteri- references 41, 44, and 45. The re- rehabilitation after anterior cruciate or cruciate ligament with a patellar maining references are case-control ligament reconstruction. Am J Sports ligament allograft: I. Insertion of the Med 1990;18:292-299. cohort studies, basic research stud- graft and anterior-posterior testing. 14. Fisher SE, Shelbourne KD: Arthro- J Bone Joint Surg Am 1996;78:1720- ies, or expert opinion. scopic treatment of symptomatic ex- 1727. tension block complicating anterior Citation numbers printed in bold 26. Pinczewski LA, Deehan DJ, Salmon cruciateligamentreconstruction. Am LJ, Russell VJ, Clingeleffer A: A five- type indicate references published J Sports Med 1993;21:558-564. year comparison of patellar tendon within the past 5 years. 15. Noyes FR, Mangine RE, Barber SD: The versus four-strand hamstring tendon early treatment of motion complica- autograft for arthroscopic reconstruc- 1. Skutek M, Elsner HA, Slateva K, et al: tions after reconstruction of the ante- tion of the anterior cruciate ligament. Screening for arthrofibrosis after ante- rior cruciate ligament. Clin Orthop Am J Sports Med 2002;30:523-536. rior cruciate ligament reconstruction: Relat Res 1992;277:217-228. 27. Sajovic M, Vengust V, Komadina R, Analysis of association with human 16. Sisto DJ, Warren RF: Complete knee leukocyte antigen.Arthroscopy 2004; dislocation: A follow-up study of op- Tavcar R, Skaza K: A prospective, ran- 20:469-473. erative treatment. Clin Orthop Relat domized comparison of semitendino- 2. Freeman MA, Pinskerova V: The move- Res 1985;198:94-101. sus and gracilis tendon versus patellar ment of the normal tibio-femoral joint. 17. Shapiro MS, Freedman EL: Allograft tendon autografts for anterior cruciate J Biomech 2005;38:197-208. reconstruction of the anterior and ligament reconstruction: Five-year 3. Perry J, Antonelli D, Ford W: Analysis posterior cruciate ligaments after follow-up. Am J Sports Med 2006;34: of knee-joint forces during flexed- traumatic knee dislocation. Am J 1933-1940. knee stance. J Bone Joint Surg Am Sports Med 1995;23:580-587. 28. Patton WC, Tew WM: Periarticular 1975;57:961-967. 18. Yaru NC, Daniel DM, Penner D: The heterotopic ossification after multi- 4. Gogia PP, Braatz JH, Rose SJ, Norton effect of tibial attachment site on graft ple knee ligament reconstructions: A BJ: Reliability and validity of gonio- impingement in an anterior cruciate report of three cases. Am J Sports metric measurements at the knee. ligament reconstruction. Am J Sports Med 2000;28:398-401. Phys Ther 1987;67:192-195. Med 1992;20:217-220. 29. Cosgarea AJ, Sebastianelli WJ, De- 5. Schlegel TF, Boublik M, Hawkins RJ, 19. Romano VM, Graf BK, Keene JS, Haven KE: Prevention of arthrofibro- Steadman JR: Reliability of heel- Lange RH: Anterior cruciate ligament sis after anterior cruciate ligament re- height measurement for documenting reconstruction: The effect of tibial construction using the central third knee extension deficits. Am J Sports tunnel placement on range of motion. patellar tendon autograft. Am J Med 2002;30:479-482. Am J Sports Med 1993;21:415-418. Sports Med 1995;23:87-92. 6. Sprague NF III, O’Connor RL, Fox JM: 20. Shelbourne KD, Patel DV: Treatment 30. Shelbourne KD, Wilckens JH, Molla- Arthroscopic treatment of postopera- of limited motion after anterior cruci- bashy A, DeCarlo M: Arthrofibrosis in tive knee fibroarthrosis. Clin Orthop ate ligament reconstruction. Knee acute anterior cruciate ligament recon- Relat Res 1982;166:165-172. Surg Sports Traumatol Arthrosc 1999; struction: The effect of timing of recon- 7. Del Pizzo W, Fox JM, Friedman ML, 7:85-92. struction and rehabilitation. Am et al: Operative arthroscopy for the 21. Markolf KL, Hame S, Hunter DM, et J Sports Med 1991;19:332-336. treatment of arthrofibrosis of the al: Effects of femoral tunnel place- 31. Wasilewski SA, Covall DJ, Cohen S: Ef- knee. Contemp Orthop 1985;10:67- ment on knee laxity and forces in an fect of surgical timing on recovery and 72. anterior cruciate ligament graft. J associated injuries after anterior cru- 8. Blauth W, Jaeger T: Arthrolysis of Orthop Res 2002;20:1016-1024. ciate ligament reconstruction. Am the knee joint [German]. Orthopade 22. Howell SM, Taylor MA: Failure of re- J Sports Med 1993;21:338-342. 1990;19:388-399. construction of the anterior cruciate 32. Sterett WI, Hutton KS, Briggs KK, 9. Shelbourne KD, Patel DV, Martini DJ: ligament due to impingement by the Steadman JR: Decreased range of mo- Classification and management of ar- intercondylar roof. J Bone Joint Surg tion following acute versus chronic throfibrosis of the knee after anterior Am 1993;75:1044-1055. anterior cruciate ligament recon-

Volume 15, Number 11, November 2007 693 Arthrofibrosis of the Knee

struction. Orthopedics 2003;26:151- tissue. Winner of the AGA-DonJoy Arthrosc 2003;11:318-321. 154. Award 2003. Arch Orthop Trauma 46. Kim DH, Gill TJ, Millett PJ: Arthro- 33. Hooper GJ, Walton DI: Reconstruc- Surg 2004;124:585-591. scopic treatment of the arthrofibrotic tion of the anterior cruciate ligament 41. Noyes FR, Berrios-Torres S, Barber- knee. Arthroscopy 2004;20(suppl 2): using the bone-block iliotibial-tract Westin SD, Heckmann TP: Preven- 187-194. transfer. J Bone Joint Surg Am 1987; tion of permanent arthrofibrosis after 47. Millett PJ, Williams RJ III, Wick- 69:1150-1154. anterior cruciate ligament recon- iewicz TL: Open debridement and soft 34. Zarins B, Rowe CR: Combined anteri- struction alone or combined with as- tissue release as a salvage procedure or cruciate-ligament reconstruction sociated procedures: A prospective for the severely arthrofibrotic knee. using semitendinosus tendon and study in 443 knees. Knee Surg Sports Am J Sports Med 1999;27:552-561. iliotibial tract. J Bone Joint Surg Am Traumatol Arthrosc 2000;8:196-206. 48. Steadman JR, Burns TP, Peloza J: Sur- 1986;68:160-177. 42. Delay BS, Smolinski RJ, Wind WM, gical treatment of arthrofibrosis of the 35. Border WA, Noble NA: Transforming Bowman DS: Current practices and knee. J Orthop Tech 1993;1:119-127. growth factor beta in tissue fibrosis. opinions in ACL reconstruction and NEnglJMed1994;331:1286-1292. rehabilitation: Results of a survey of 49. Wang JH, Zhao JZ, He YH: A new 36. Fullerton LR Jr, Andrews JR: Mechan- the American Orthopaedic Society for treatment strategy for severe arthrofi- ical block to extension following aug- Sports Medicine. Am J Knee Surg brosis of the knee: A review of twenty- mentation of the anterior cruciate lig- 2001;14:85-91. two cases. J Bone Joint Surg Am ament: A case report. Am J Sports 43. Feller J, Bartlett J, Chapman S, 2006;88:1245-1250. Med 1984;12:166-168. Delahunt M: Use of an extension- 50. Klein W, Shah N, Gassen A: Arthro- 37. Jackson DW, Schaefer RK: Cyclops syn- assisting brace following anterior cru- scopic management of postoperative drome: Loss of extension following ciate ligament reconstruction. Knee arthrofibrosis of the knee joint: intra-articular anterior cruciate liga- Surg Sports Traumatol Arthrosc Indication, technique, and results. ment reconstruction. Arthroscopy 1997;5:6-9. Arthroscopy 1994;10:591-597. 1990;6:171-178. 44. Melegati G, Tornese D, Bandi M, Vol- 51. Shelbourne KD, Johnson GE: Outpa- 38. Paulos LE, Wnorowski DC, Green- pi P, Schonhuber H, Denti M: The role tient surgical management of ar- wald AE: Infrapatellar contracture of the rehabilitation brace in restoring throfibrosis after anterior cruciate lig- syndrome: Diagnosis, treatment, and knee extension after anterior cruciate ament surgery. Am J Sports Med long term followup. Am J Sports Med ligament reconstruction: A prospec- 1994;22:192-197. 1994;22:440-449. tive controlled study. Knee Surg 52. Hasan SS, Saleem A, Bach BR Jr, Bush- 39. Zeichen J, van Griensven M, Albers I, Sports Traumatol Arthrosc 2003;11: Joseph CA, Bojchuk J: Results of arthro- Lobenhoffer P, Bosch U: Immunohis- 322-326. scopic treatment of symptomatic loss tochemical localization of collagen 45. Mikkelsen C, Cerulli G, Lorenzini M, of extension following anterior cruci- VI in arthrofibrosis. Arch Orthop Bergstrand G, Werner S: Can a post- ate ligament reconstruction. Am J Trauma Surg 1999;119:315-318. operative brace in slight hyperexten- Knee Surg 2000:13:201-209. 40. Unterhauser FN, Bosch U, Zeichen J, sion prevent extension deficit after 53. Cosgarea AJ, DeHaven KE, Lovelock Weiler A: Alpha-smooth muscle actin anterior cruciate ligament recon- JE: The surgical treatment of arthrofi- containing contractile fibroblastic struction? A prospective randomised brosis of the knee. Am J Sports Med cells in human knee arthrofibrosis study. Knee Surg Sports Traumatol 1994;22:184-191.

694 Journal of the American Academy of Orthopaedic Surgeons