Many patients achieve satisfactory range of motion (ROM) after total knee arthroplasty. 1 If postoperative motion is limited, treatment with manipulation can facilitate participation in physical therapy exercise programs to obtain adequate flexion. 2,5
Occasionally, limited motion persists and causes significant functional impairment that requires operative treatment. Williams et al 19 reported improvement in motion in nine patients with limited flexion after posterior cruciate retaining total knee arthroplasty who were treated with arthroscopic posterior cruciate ligament release and manipulation. However, two (22%) of the patients required revision total knee arthroplasty. Nicholls and Dorr 13 reported satisfactory results in 11 of 12 patients who had revision of a total knee arthroplasty for stiffness. The cause of limited motion was malpositioned components in eight of the patients. The current authors have observed that stiffness can occur after posterior cruciate retaining and substituting total knee arthroplasties that are well aligned with appropriate sized components. This study was conducted to review the pathologic findings and results of surgical treatment for arthrofibrosis after total knee arthroplasty.
MATERIALS AND METHODS
This study was based on a review of the clinical, radiographic, and pathologic findings of six knees in five consecutive patients who required revision total knee arthroplasty because of limited motion after primary total knee arthroplasty. All of the patients had an underlying diagnosis of osteoarthritis and were treated with a prolonged course of physical therapy after the primary total knee arthroplasty. Revision total knee arthroplasty was offered to patients who presented with soft tissue contracture that impaired knee function and who were unwilling to accept the limitations associated with stiffness. The revision operations were performed by one surgeon. Minimum followup was 2 years.
Patients were included in the study if the revision surgery was performed because of complaints of stiffness and pain. Patients who had revision surgery performed because of pain but did not have limited motion were excluded from the study group. Patients who had revision surgery performed because of mechanical failure of the components, wear, or prior infection also were excluded. Included were only well-aligned knees with appropriate-sized components in which no cause of stiffness was identified other than soft tissue contracture. Patients with an underlying diagnosis of inflammatory arthritis were excluded.
All knees were exposed through a medial parapatellar approach with extended tibial tubercle osteotomy. Scar tissue was excised from the suprapatellar pouch, extensor mechanism, and medial and lateral gutters. The tibial polyethylene insert and femoral component were removed, leaving the tibial baseplate and patellar components in place. The posterior cruciate ligament, if present, and scar from the medial and lateral posterior compartments were excised. The femur was revised using a posterior stabilized implant. A revision femoral component was chosen of the same size or slightly smaller than the removed component to provide a wider flexion space and potentially improve knee flexion. A tibial posterior stabilized insert was chosen of appropriate thickness to permit approximately 5° hyperextension and full flexion of the knee without dislocation of the femoral component over the tibial post.
Postoperative rehabilitation was the same as that used for routine primary total knee arthroplasty. Constant passive motion was started on the day of surgery and increased daily within the patient’s pain tolerance. Weightbearing activity was not restricted. A knee immobilizer was used during ambulation until the patient could actively perform straight leg raise against gravity and then it was discontinued. Physical therapy was performed daily during the inpatient hospitalization and continued on an outpatient basis for approximately 3 months.
Heterotopic bone formation was quantitated using the system of Harwin et al. 8 Heterotopic bone was assessed on lateral radiographs obtained within 1 week before surgery and at the most recent followup. Knee ROM and Knee Society score 9 were recorded before surgery and at the most recent followup. Radiographs obtained before the primary arthroplasty were not available for five of the six knees so it was not possible to assess whether the initial components were excessively large or thick.
Representative soft tissue specimens were obtained from the patellofemoral and posterior compartments. Sections were stained with hematoxylin and eosin and examined using light microscopy.
To determine if the pathologic features of the periarticular soft tissue in patients who had revision surgery for stiffness were different from those for patients who had revision surgery for other reasons, histologic specimens were obtained from a control group of a consecutive series of six knees in six patients who had revision total knee arthroplasty performed for reasons other than stiffness. Representative soft tissue specimens were obtained from the patellofemoral and posterior compartments.
All five patients in the study group were female with a mean age of 65.7 years. Four knees were right and two were left. Four of the revised knee arthroplasties were posterior cruciate retaining and two were posterior cruciate substituting (Table 1). One patient (Patient 5) presented with right knee pain and ROM from 20° to 50° 2 years after primary cruciate retaining total knee arthroplasty. The knee was revised successfully and the patient achieved relief of pain and improvement in motion to 0° to 85°. Degenerative arthritis in the contralateral knee then became sufficiently symptomatic that primary left total knee arthroplasty was performed. To minimize any possible causes of stiffness in the left knee, a primary posterior stabilized prosthesis was used and full ROM was obtained during the operation. The patient achieved flexion from 10° to 85° by the time of her discharge from the hospital but lost motion despite the use of home constant passive motion, daily physical therapy, and adequate pain control with oral analgesics. Two months after the operation, flexion was 10° to 45°. Manipulation was performed using a continuous epidural anesthetic. The patient remained in the hospital for 4 days to receive inpatient physical therapy while the epidural catheter was used to facilitate pain control. At the time of discharge, knee motion was 0° to 90°. Despite continued outpatient physical therapy and home constant passive motion, flexion was limited from 5° to 65°, 3 months after arthroplasty. The patient was readmitted and an epidural catheter was placed to facilitate inpatient physical therapy. The patient remained in the hospital 1 week and achieved motion from 5° to 85° by the time of her discharge. Despite continued outpatient physical therapy and use of oral narcotics for pain control, 1 year after the operation, knee motion was limited from 10° to 30° flexion. Grade 1 heterotopic bone was present along the superior edge of the femoral component (Fig 1). The left knee then was revised using the technique previously described. Dense scar was found covering the patella (Fig 2) and throughout the knee. The femoral component was revised to a smaller, slightly undersized posterior stabilized prosthesis; scar tissue was excised and the tibial tubercle was recessed to lengthen the extensor mechanism. Although Grade 1 heterotopic bone reformed after the revision surgery, the patient’s pain was relieved and she has 0° to 80° flexion.
Knee ROM and Knee Society score were improved after the revision operations (Table 1). All patients thought that they had benefited to some degree from the operation. One complication, which was a fracture of the medial femoral condyle, occurred during removal of a posterior stabilized prosthesis in a patient with osteoporosis (Patient 2). This required use of a constrained condylar femoral component. The patient experienced improvement in extension of 30° after the operation but did not gain more flexion. Arthroscopic lysis of adhesions was performed 6 months after the revision arthroplasty to improve flexion, but the ROM remained unchanged.
Dense scar tissue was observed consistently in the knees of all patients (Fig 3). Cells in the dense scar tissue were ovoid and spindle-shaped with typical characteristics of fibrocartilage. Some ovoid cells were surrounded by a clear lacuna space, similar to that observed in chondrocytes. Although heterotopic bone was present in two knees preoperatively, no direct mechanical obstruction to knee motion was caused by the bone masses observed during surgery. Heterotopic ossification was observed in five of the six knees after revision arthroplasty.
Two of the knees in the control group were revised for instability, two were revised for wear, and two for prior infection. The mean patient age was 67.3 years. Four of the patients were female and two were male. The mean preoperative arc of knee motion was 82.5°. Histologic analysis of the periarticular soft tissue specimens showed fibrocartilage in five of the six knees in the control group. However, the collagen and fibrocartilage metaplasia appeared qualitatively more dense in the study group than in the control group.
Stiffness after total knee arthroplasty has been attributed to retention of the posterior cruciate ligament, collagen disease, or technical errors in operative technique. 13,19 The current authors have observed arthrofibrosis after posterior cruciate retaining and substituting total knee arthroplasty. This indicates that stiffness cannot be avoided by removing the posterior cruciate ligament.
Observations in the current study of dense scar-containing fibrocartilage suggests that stiffness can result from pathologic scar formation. Previous studies have shown that increased compressive forces on connective tissues results in fibrocartilaginous metaplasia, including increased extracellular matrix synthesis. 4,7 Therefore, mechanical compression, such as the passive ROM exercises used to maximize knee motion after total joint arthroplasty, can produce mechanical stresses on soft tissues sufficient to induce fibrocartilage metaplasia. It has been shown that in human digital A1 pulleys, 17 in human tibialis posterior tendon and ligaments, 18 and in other human tendons, 15,18 compressive forces can regulate the development of fibrocartilaginous tissue with gene expression, synthesis, and accumulation of proteoglycans similar to the extracellular matrix of fibrocartilage. However, fibrocartilage metaplasia was present consistently in the study and control groups, although it was qualitatively more dense in the study group. This suggests that both populations have the potential to have fibrocartilage metaplasia develop and the fibrocartilage metaplasia may not necessarily lead to arthrofibrosis. What triggers the proliferation of extensive scar tissue formation is not clear. Some patients may be predisposed to this condition or may have it develop as a response to the surgical trauma and postoperative rehabilitation.
Salter observed that passive ROM causes articular cartilage metaplasia. 16 Limiting motion may restrict the development of fibrocartilage metaplasia. However, limiting motion after total knee arthroplasty until wound healing occurs also can lead to stiffness. 11
Although heterotopic bone rarely causes direct mechanical obstruction to knee motion, an association is seen between heterotopic bone formation and stiffness. 2,6,8 Furia and Pellegrini 6 suggested that the heterotopic bone may be indicative of a systemic predisposition for fibrosis and joint stiffness after surgical insult. This concept is supported by the current findings of heterotopic bone in five of six knees after revision surgery for stiffness.
Pathologic bone formation typically occurs in association with trauma, head injury, total joint arthroplasty, and some bone tumors. 13 Fibrocartilage may represent a precursor of heterotopic bone and cause stiffness after total knee arthroplasty as a result of operative trauma. All of the patients in the study group were female and had an underlying diagnosis of osteoarthritis, which has been associated with heterotopic bone formation. 8 However, no specific characteristics were identified that distinguish this group of patients from the general population of patients with osteoarthritis who undergo total knee arthroplasty.
Recurrence of the pathologic soft tissue stiffening would be expected with additional operative trauma. However, revision of the femoral component and excision of scar resulted in improved knee function. This may be related to the intentional creation of relatively loose extension and flexion spaces, which would not be desired during routine primary total knee arthroplasty.
Arthrofibrosis after total knee arthroplasty also can be treated arthroscopically. 10,19 However, arthroscopy appears most beneficial for treatment of patellar clunk syndrome and to explore knees with satisfactory ROM. 3 The average preoperative arc of motion of knees in the study group was only 36°. Although arthroscopy after a total knee arthroplasty is technically feasible to perform, insertion of the arthroscope into a markedly stiff knee with an arthrofibrotic patellofemoral compartment can create scratches in the femoral component and visualization is limited.
Postoperative Knee Society scores, which were improved significantly in the current study, were not attributable to increased ROM alone. Patients also experienced variable relief of pain. Patients who undergo revision total knee arthroplasty for arthrofibrosis can achieve some improvement in motion and pain, but the degree of improvement is difficult to predict.
This population of patients is distinctly different from those who present with unexplained pain, but without stiffness after primary total knee arthroplasty. Knees that are explored for unexplained pain typically remain painful after revision arthroplasty, whereas those having limited motion can be improved with operative treatment. 12
1. Anoachi YS, McShane M, Kelly F, et al: Range of motion in total knee replacement. Clin Orthop 331:87–92, 1996.
2. Daluga D, Lombardi AV, Mallory TH, et al: Knee manipulation following total knee arthroplasty. Analysis of prognostic variables. J Arthroplasty 6:119–128, 1991.
3. Diduch DR, Scuderi GR, Scott WN, et al: The efficacy of arthroscopy following total knee replacement. Arthroscopy 13:166–171, 1997.
4. Evanko SP, Vogel KG: Proteoglycan synthesis in fetal tendon is differentially regulated by cyclic compression in vitro. Arch Biochem Biophys 307: 153–164, 1993.
5. Fox JL, Poss R: The role of manipulation following total knee replacement. J Bone Joint Surg 63A: 357–362, 1981.
6. Furia JP, Pellegrini VD: Heterotopic ossification following primary total knee arthroplasty. J Arthroplasty 10:413–419, 1995.
7. Gillard GC, Reilly HC, Bell-Booth PC, et al: The influence of mechanical forces on the glucosaminoglycan content of the rabbit flexor digitorum profundus tendon. Connect Tissue Res 7:37–46, 1979.
8. Harwin SF, Stein AJ, Stern RE, et al: Heterotopic ossification following primary total knee arthroplasty. J Arthroplasty 8:113–116, 1993.
9. Insall JN, Dorr LD, Scott RD, et al: Rationale of the Knee Society clinical rating system. Clin Orthop 248:13–14, 1989.
10. Markel DC, Luessenhop CP, Windsor RE, et al: Arthroscopic treatment of peripatellar fibrosis after total knee arthroplasty. J Arthroplasty 11:293–297, 1996.
11. Mattingly PC, Bentley G, Cohen ML, et al: Preliminary experience with the Geomedic total knee replacement. Rheumatol Rehabil 16:241–247, 1977.
12. Mont MA, Senna FK, Krackow KA, et al: Exploration of radiographically normal total knee replacements for unexplained pain. Clin Orthop 331: 216–220, 1996.
13. Nicholls DW, Dorr LD: Revision surgery for stiff total knee arthroplasty. J Arthroplasty 5:573–577, 1990.
14. Pazas JE, Miller MD, Rosier RN: Pathologic bone formation. Clin Orthop 245:269–281, 1989.
15. Robbins JR, Evanko SP, Vogel KG: Mechanical loading and TGF-beta regulate proteoglycan synthesis in tendon. Arch Biochem Biophys 342:203–211, 1997.
16. Salter RB: The biologic concept of continuous passive motion of synovial joints. The first 18 years of basic research and its clinical application. Clin Orthop 242:12–25, 1989.
17. Sampson SP, Badalamente MA, Hurst LC, et al: Pathology of the human A1 pulley in trigger finger. J Hand Surg 16A:714–721, 1991.
18. Vogel KG, Ordog A, Pogany G, et al: Proteoglycans in the compressed region of human tibialis posterior tendon and in ligaments. J Orthop Res 11:68–77, 1993.
19. Williams RJ, Westrich GH, Siegel J, et al: Arthroscopic release of the posterior cruciate ligament for stiff total knee arthroplasty. Clin Orthop 331: 185–191, 1996.