Patient expectations after a total knee arthroplasty (TKA) include pain relief, improved ambulatory ability, and improved knee flexion. The amount of flexion achieved postoperatively is determined by the amount of preoperative flexion, especially if the flexion was less than 75°.7,8,13,18,23 Therefore, preoperative restriction of knee motion is a challenging dilemma for surgeons.
There is limited information regarding results of primary TKAs in patients with a preoperative arc of motion of 50° or less, and in patients with spontaneous osseous ankylosis. 1-3,11,12,14-18,20,24 However, we know from the literature that the surgical approach in these patients can be challenging because of difficulty in patellar ever-sion.1,11,12,20 For example, identifying the joint line may be difficult in ankylosed knees where the patella usually is fused to the anterior femur. Technical problems include lack of adequate exposure, the need for extensile surgical approaches, the risk of patellar tendon avulsion, difficulty balancing the flexion-extension gaps, component malpositioning, extensor mechanism management, patellar mal-tracking, avulsion of the collateral ligaments, and difficulty in wound closure.1,11,12,20 Investigators have reported an inconsistent range of motion (ROM) and a high risk of complications and revision rates, especially in patients with ankylosed knees.16,19 However, others have reported that postoperative ROM in patients with stiff and ankylosed knees can be the same as ROM in patients with flexible knees, and that there were low complication rates in these replaced knees.18,26 Numerous studies had small patient cohorts,1-3,11,13,15,17,18,24 and although a higher complication rate and suboptimal outcome after TKA were reported, there may be a tendency to underestimate complications, especially in patients with ankylosed knees.19
We wondered whether the function and motion of patients with stiff knees would be better than those in patients with ankylosed knees. We asked whether there would be differences in ROM and function for knees ankylosed in extension and knees ankylosed in flexion. Finally, we asked whether the complication or revision rates would differ.
MATERIALS AND METHODS
We retrospectively reviewed the records of 1247 primary TKAs performed by the senior author (SB) from January 1994 to December 2002. One hundred six (13%) were performed in knees with a preoperative flexion arc less than 50°. Of the 106 knees, we then reviewed 52 patients (90 knees; bilateral in 38 patients and unilateral in 14 patients) with a minimum followup of 2 years. We retrieved the patients' clinical and radiographic data prospectively entered in a computerized database. The preoperative data included the patient demographics, the position of ankylosis, arc of flexion, Knee Society knee and function scores, ambulatory capacity, operative approach, need for extensile approaches, intraoperative problems, and prosthesis type. The mean followup was 6.5 years (range, 2-10 years).
The patients were classified into two groups: Group 1 included patients with stiff knees (64 knees in 35 patients; bilateral in 29 patients, and unilateral in six patients), and Group 2 included patients with ankylosed knees (26 knees in 17 patients). Twelve knees in eight patients (bilateral in four patients, and unilateral in four patients) were ankylosed in extension (Group 2a), and 14 knees in nine patients (bilateral in five patients, and unilateral in four patients) were ankylosed in flexion (Group 2b) (Fig 1A-B). The mean sagittal alignment in the spontaneous osseous ankylosed knees was 0°-15° in Group 2a and 35°-65° in Group 2b. The mean age of the patients was 53.7 ± 13.3 years (range, 21-75 years) (Group 1, 59 ± 11 years; Group 2, 42 ± 11 years). There were 32 women and 20 men.
Thirty-two patients had simultaneous bilateral TKAs under one anesthesia. Six patients had staged bilateral TKAs and 14 had a unilateral TKA. Ipsilateral total hip arthroplasties (THA) were required in 22 patients. Four patients had bony ankylosis of the hips and knees; all four joints were replaced in a staged manner. Rheumatoid arthritis (RA) was the most common pre-operative diagnosis (Table 1). Nine patients in the spontaneous osseous ankylosis group were unable to walk preoperatively. Eight patients needed a cane for ambulation. The mean duration of knee fusion in the spontaneous ankylosis group was 3.5 years (range, 2-6 years). None of the patients with spontaneous osseous ankylosis had pain preoperatively, whereas all patients with stiff knees had severe pain preoperatively. Of the patients with stiff knees, 21 patients with bilateral fixed flexion deformities were bedridden for at least 6 months preoperatively. Fourteen patients needed support for ambulation. Thirty-two patients were unable to climb stairs and 20 patients could climb stairs in some way. All patients wanted a mobile joint for better sitting and walking. Patients with ankylosed knees were informed about postoperative pain recurrence, and all patients were informed about the high risk of complications.
The surgeries were performed under tourniquet control using a medial parapatellar approach. The posterior cruciate ligament was sacrificed because of severe deformity. We used an extra-medullary alignment jig for the tibia and an intramedullary alignment jig for the femur. The patella was not replaced. We performed a patelloplasty, which included soft tissue release from the lateral patella, division of the patellofemoral ligament, and patellar rim cautery to provide partial denervation and osteophyte removal. We removed 2-4 mm of the articular surface of the patella in patients with RA in the knee.
We performed an extensile exposure (rectus snip) in 10 of the 64 knees in Group 1 to enable eversion of the patella and knee flexion. Soft tissue procedures included lateral retinacular release in 14 knees and lateral gutter débridement in 34 knees. Twenty-one knees required extensive posterior capsulotomy and subperiosteal elevation of the gastrocnemius from the posterior femur. Six knees required an additional 5-mm distal femoral cut to correct the severe preoperative flexion deformity (> 30°). Fourteen patients had a postoperative residual flexion deformity of 10°. The femoral component was upsized in one of 17 knees to prevent flexion instability. Extensive medial capsular sleeve dissection, including stripping the superficial medial collateral ligament, was performed in 33 knees to correct the preoperative varus deformity (> 20°). The iliotibial band was released from the Gerdy's tubercle in 14 knees with a preoperative valgus deformity of 20°. Posterior-stabilized components were used in 57 knees (23 Insall-Burstein II, (Zimmer Inc, Warsaw, IN), (34 NexGen, Zimmer Inc), and an Insall-Burstein II constrained condylar prosthesis (Zimmer Inc) was used in seven knees (Fig 1C-D), with an asymmetric laxity greater than 1 cm for the capsular and ligamentous insufficiency. All patients in Group 1 had at least a 90° arc of flexion after wound closure.
In Group 2a (patients with spontaneous osseous ankylosis in extension), we performed a quadriceps VY-plasty5 early during exposure to facilitate patellar eversion (Fig. 2). The patella was fused to the anterior femur in all patients (Fig. 2B), and was osteotomized and reflected laterally after the arthrotomy. The tibiofemoral joint line was visible in all knees, even in patients with bony ankylosis. We performed the osteotomy along the joint line using a curved osteotome to separate the tibia and femur. We used precaution to preserve as much bone as possible, and to preserve the medial and lateral soft tissue sleeve. A subperiosteal femoral peel was needed in four knees (three patients). An Insall-Burstein II constrained condylar prosthesis (Zimmer Inc) was used in all knees and allowed for sacrifice of both cruciate ligaments while providing valgus-varus constraint; an operative step considered essential for this type of fused knee.11,12 The VY-plasty was repaired with the knee in 45° flexion to prevent extension lag. The mean arc of motion was 78° (range, 70°-90°) after closing the capsule and repairing the quadriceps.
All knees in Group 2b (patients with spontaneous osseous ankylosis in flexion) were exposed through a medial parapatellar arthrotomy. We made no attempt to lengthen or shorten the quadriceps expansion. The quadriceps was elevated from the anterior femur, and adhesions in lateral and medial gutters were excised. We performed a lateral retinacular release to facilitate exposure and patellar eversion. We performed a rectus snip in four knees (three patients), and the iliotibial band was elevated from Gerdy's tubercle in 10 knees (seven patients) with concomitant valgus and external rotation deformity. The medial periosteal sleeve was elevated, and the osteotomy was performed at the level of the joint line to separate the femur and the tibia. Posterior soft tissue release could be done only after the bone cuts because of poor observation. All patients required a posterior capsulotomy with subperiosteal elevation of the medial and lateral heads of the gastrocnemius. They also required a 5-mm distal femoral cut to correct the flexion deformity. The femoral component was upsized in 10 knees to equalize the flexion and extension gaps. An Insall-Burstein II constrained condylar pros-thesis (Zimmer Inc) was used in eight knees, a Legacy® constrained condylar prosthesis (Zimmer) was used in two knees (Fig. 2C-D), and a posterior-stabilized prosthesis (NexGen®, Zimmer Inc) was used in four knees. Ten of the 14 knees had a residual flexion deformity of 10°. One patient had a 35°-residual flexion deformity. The patient had postoperative skeletal traction applied through a lower tibial pin for 2 weeks, which corrected the flexion deformity to 10°. After closure, the mean arc of flexion was 87° (range, 80°-100°).
The severe deformities and osteopenia made surgery technically demanding for all three groups. Small components often were needed in patients with bony ankylosis in flexion because the anterior femoral condyles were destroyed in some of the knees. The tibial and femoral components were cemented. Pulsatile lavage was used before cementing the implants.
Patients received antibiotic prophylaxis with intravenous cefazolin (2 g 30 minutes before tourniquet followed by 1 g every 8 hours for 3 days) and antithrombotic prophylaxis with subcutaneous enoxaparin (40 mg from the night before surgery until 10 days postoperatively).
The postoperative regimen included: epidural analgesia, knee immobilization for 2 days, gravity-assisted regaining of flexion from the third postoperative day, and walking with support from the fourth postoperative day. Progressive resistance exercises were started 2 weeks postoperatively to strengthen the quadriceps and were continued for 1 year postoperatively. All patients used support while walking for 3 months postoperatively. In patients with a VY-plasty, knee flexion beyond 45° was delayed for 1 week and a knee brace was used for ambulation for 6 weeks.
We evaluated the patients preoperatively and postoperatively at intervals of 6 weeks, 3 months, 6 months, 1 year, and yearly thereafter until the final followup between October and November 2004. Preoperative and postoperative clinical evaluations were performed according to The Knee Society recommendations.10 Radiographic evaluations of alignment were performed on preoperative and followup postoperative radiographs, which included anteroposterior (AP) standing, supine lateral, and skyline patellar views according to The Knee Society guidelines.6 Patients were evaluated by an independent orthopaedic surgeon (RM) unless the patient was lost to followup or could not attend followup. We measured the postoperative arc of flexion and postoperative Knee Society knee and function scores at the final followup. Data regarding the intraoperative and immediate postoperative complications were retrieved from the operating notes of the chief surgeon (SB). Complications were classified as major if they necessitated additional or repeat surgery and had an effect on the long-term outcome; other complications were deemed minor. Revision for any reason was documented.
Knee Society scores and knee flexion were compared using the Independent samples t test. The demographic data, lateral release rates, complication rates, and survivorship rates were compared using the chi square test. Similarly, the results of TKAs in patients with ankylosis in extension were compared with the results in patients with ankylosis in flexion. A p value of 0.05 was considered significant. The survival rate was calculated using a Kaplan-Meier survivorship analysis. Statistical analysis was performed using SPSS software (SPSS Inc, Chicago, IL).
The postoperative Knee Society knee and function scores in patients with stiff knees were higher (p < 0.001) than scores of patients with ankylosed knees (Table 2). Patients with stiff knees had a greater (p < 0.001) postoperative arc of motion than patients with ankylosed knees (whether in extension or flexion). However, patients with ankylosed knees had more improvement (p < 0.001) in flexion (Table 3). The mean preoperative flexion deformity was 31° (range, 0°-65°). Postoperatively, flexion deformity persisted in 10 knees in Group 1 (mean, 9°; range, 5°-15°) and in seven knees in Group 2b (mean, 10.7°; range, 10°-15°).
We observed no differences between preoperative and postoperative arcs of motion, Knee Society knee scores, lateral release rates, and complication rates between knees ankylosed in extension and knees ankylosed in flexion (Table 2). At final followup, the mean tibiofemoral alignment was 4° valgus (range, 0°-9° valgus) in both groups (Table 4). Lateral release rates were higher (p < 0.001) in patients with ankylosed knees (26 of 26 knees in 17 patients) than in patients with stiff knees (14 of 64 knees in 35 patients).
Excluding superficial wound necrosis, there were no differences between the complication rates of stiff (11%) and ankylosed (24%) knees. The overall complication rate was 33%, with an intraoperative complication rate of 10%. Four of 26 ankylosed knees (one fracture, one hematoma needing reoperation, one peroneal nerve palsy, and one persistent stiffness) and two of 64 stiff knees (one deep infection and one revision) had major complications. Partial avulsion of the patellar tendon occurred in six knees in Group 1 and in two knees in Group 2 during patellar eversion. An intraoperative medial femoral condyle fracture occurred in one knee in Group 2a, which was treated with screw fixation and an intramedullary stem extension (Fig. 2C). One patient (one knee) had transient peroneal nerve palsy with preoperative flexion valgus deformity develop (Group 2b). Superficial wound necrosis occurred in four of 64 (7%) stiff knees, and in 13 of 26 (50%) ankylosed knees; all healed spontaneously. The duration for wound healing ranged from 14-25 days. One knee had hematoma formation that necessitated drainage (Group 2). A deep methicillin-sensitive Staphylococcus aureus infection developed in one knee, which subsided with débridement and intravenous antibiotics for 6 weeks. One patient with spontaneous osseous ankylosis in extension had no improvement in flexion postoperatively. This patient had a repeat VY-plasty 1 year after the index procedure. The patient had regained 75° flexion at the latest followup. One patient in Group 1 with RA had aseptic loosening of the femoral component 8.5 years after the index surgery, and had a subsequent revision TKA. Kaplan-Meier survivor-ship analysis using revision for any reason revealed 99% survivorship at 6.5 years followup in patients with stiff knees and 100% survivorship at 6.5 years in patients with ankylosed knees.
There is little published information regarding results of TKAs in knees with a preoperative arc of motion less than 50° and in knees with spontaneous osseous ankylosis.1-3,11,12,14-18,20,24 Several studies include only a small number of patients,1-3,11,13,15,17,18,24 and were not able to accurately compare stiff and ankylosed knees or identify all complications. Aglietti and Buzzi (20 stiff knees, six ankylosed knees) reported ankylosed knees achieved less motion than stiff knees.1 However, Montgomery et al17 reported no difference in the results after TKA when comparing ankylosed knees and knees with relatively normal motion. However, because there were few ankylosed knees (three of 84 knees), their study likely lacked statistical power to detect a difference.17 Our series (90 knees: 64 stiff versus 26 ankylosed knees) is the largest series of TKAs in patients with stiff and ankylosed knees. No previous study has compared the results of TKAs in patients with stiff knees and ankylosed knees. Our data suggest that patients with stiff knees had a better clinical outcome after TKA than patients with ankylosed knees.
Our study has several limitations. This retrospective study lacked adequate power to compare the results of TKAs in patients with knees ankylosed in extension with the results in patients with knees ankylosed in flexion. Several prostheses were used during the study, and comparison of results based on the implant used also would be prone to inadequate power. We did not use any standardized general outcome measures such as the Short Form-36 or the Western Ontario and McMaster University Osteo-arthritis Index (WOMAC). However, the data are appropriate to interpret ROM. We also note the possibility of intraobserver variation in interpreting the radiographic results.
A goal of TKA includes obtaining a painless ROM between 0°-105°.22,23 However, the postoperative ROM for stiff and ankylosed knees may not be comparable with ROM attained by a patient without preoperative stiffness.22 Some have reported lack of consistent adequate motion and a high risk of complications and revision rates in patients with stiff and ankylosed knees.16,19 Mullen reported that ROM after TKA in these patients may be indistinguishable from that achieved by a comparable group of patients with flexible knees.18 However, a stiff knee was defined as one with less than 90° arc of flexion, and the small number of patients (13 patients) limited any definite findings.
Postoperative arc of flexion ranging from 64°-103° and complications ranging from 6% to 66% have been reported after TKAs in stiff and ankylosed knees in series ranging from three to 84 knees (Table 5). Our results were comparable to those of previous studies (Table 5).
The average postoperative Knee Society knee and function scores for stiff and ankylosed knees were lower than scores for mobile knees.4 This may be because our patients had severely degenerated knees and RA with polyarticular involvement.
The Knee Society scores and ROM after TKA were better in patients with stiff knees than in patients with ankylosed knees, whether ankylosed in flexion or extension.
Statistical comparison of knees ankylosed in flexion and extension was impossible given the relatively small number of patients with ankylosed knees. Although the arc of flexion obtained in patients with spontaneous ankylosis was disappointing, all but one knee had significant improvement in overall ROM, which resulted in better walking ability and markedly improved functional capabilities.
Small bone sizes, severe osteopenia in a high proportion of patients with RA, and severe soft tissue contractures made the operation technically demanding. The smaller patellar size, osteopenia, and high lateral retinacular releases because of severe deformities prompted us to retain the native patella during TKA.
During exposure, stiff knees may need a rectus snip for exposure. Knees with osseous ankylosis in extension almost always need VY quadricepsplasty before the patella can be everted to minimize the risk of patellar tendon avulsion.1
Aglietti and Buzzi recommended early quadricepsplasty to aid in exposure and patellar eversion without compromising the integrity of the patellar tendon.1 We performed early quadricepsplasty in 12 knees with osseous ankylosis in extension. This was the group of patients in whom early quadricepsplasty was essential because the quadriceps tendon was shortened from being fused in the extended position. Quadriceps lengthening was needed to aid exposure and prevent patellar tendon avulsion, and to improve postoperative range of flexion.
Although posterior-stabilized arthroplasties were performed successfully in four of 26 knees with bony ankylosis, the majority (22 of 26 knees) required a condylar constrained prosthesis (Zimmer Inc) to obtain mediolateral and AP stability. A constrained total knee prosthesis has been recommended for converting a fused knee to a TKA to substitute for deficient or absent collateral ligaments.9,24 In our experience, constrained condylar prostheses are essential in patients with knees ankylosed in extension, and also in some patients with knees ankylosed in flexion.
McAuley et al evaluated 27 TKAs in patients with a preoperative range of flexion less than 50°.16 They reported an overall complication rate of 41% with a revision rate of 18.5%.16 Similar high complication and revision rates were reported by Naranja et al in patients who had TKAs for ankylosed knees.19 The overall complication rate in patients with ankylosed knees in our series was similar to rates in previous studies.16,19 Intraoperative fractures may occur, especially in knees with osseous ankylosis in extension while trying to osteotomize the patella or the tibiofemoral joint and during forced flexion.25 Osteopenia also increases the risk of intraoperative fractures.25
Postoperative wound necrosis occurred in 13 of 26 (50%) knees (11 patients) with osseous ankylosis and in four of 64 (7%) stiff knees (3 patients). Similar high rates have been reported.11,12 The need to upsize the femoral component, deficient anterior femoral condyles (especially in knees ankylosed in flexion), correcting preoperative flexion deformity, and excessive retraction all contributed to the increased wound tension that resulted in high rates of wound necrosis. Preoperative use of soft tissue expanders may be beneficial to avoid this complication.15 Patelloplasty may aid in wound closure in selected patients with anterior overstuffing.12
The results of TKAs in the patients with preoperative stiffness and ankylosis were satisfactory, and patients were pleased to have a mobile instead of a fused knee. The ability to walk and sit in a normal fashion is important to a patient's sense of well being. Patients with stiff knees had better results than patients with ankylosed knees.
Our results were inferior to results of a standard primary TKA and we had a higher complication rate. The surgery is technically demanding and should be performed only by a surgeon with considerable experience and expertise with severe deformities and flexion-extension gap balancing. Patients need to be counseled preoperatively regarding the possibility of a suboptimal outcome compared with that of a standard TKA performed in a mobile knee, the need for prolonged physiotherapy, and the high complication rate.
1. Aglietti P, Buzzi R. Arthroplasty for the stiff and ankylosed knee. J Arthroplasty
2. Bae DK, Yoon KH. Kim,Song SJ. Total knee arthroplasty in stiff knees after previous infection. J Bone Joint Surg Br
. 2005;87: 333-336.
3. Bradley GW, Freeman MA, Albrektsson BE. Total prosthetic replacement of ankylosed knees. J Arthroplasty
4. Brinker MR, Lund PJ, Barrack RL. Demographic biases of scoring instruments for the results of total knee arthroplasty. J Bone Joint Surg Am
5. Coonse K, Adams JD. A new operative approach to the knee joint. Surg Gynecol Obstet
6. Ewald FC. The Knee Society total knee arthroplasty roentgeno-graphic evaluation and scoring system. Clin Orthop Relat Res
7. Gatha NM, Clarke HD,Fuchs R. Scuderi GR, Insall JN. Factors affecting postoperative range of motion after total knee arthroplasty. J Knee Surg
8. Harvey IA, Barry K,Kirby SP. Johnson R, Elloy MA. Factors affecting the range of motion of total knee arthroplasty. J Bone Joint Surg Br
9. Holden DL, Jackson DW. Consideration in total knee arthroplasty following previous knee fusion. Clin Orthop Relat Res
. 1988;227: 223-228.
10. Insall JN, Dorr LD, Scott RD, Scott WN. Rationale of the Knee Society clinical rating system. Clin Orthop Relat Res
. 1989;248: 13-14.
11. Kim YH, Cho SH, Kim JS. Total knee arthroplasty in bony ankylosis in gross flexion. J Bone Joint Surg Br
12. Kim YH, Kim JS, Cho SH. Total knee replacement after spontaneous osseous ankylosis and take down of formal knee fusion. J Arthroplasty
13. Lizaur A, Marco L, Cebrian R. Preoperative factors influencing the range of movement after total knee arthroplasty for severe osteoarthritis. J Bone Joint Surg Br
14. Lu H, Mow CS, Lin J. Total knee arthroplasty in the presence of severe flexion contracture: a report of 37 cases. J Arthroplasty
15. Mahomed N, McKee N,Solomon P. Lahoda L, Gross AE. Soft tissue expansion before total knee arthroplasty in arthrodesed joints. J Bone Joint Surg Br
16. McAuley JP, Harrer MF, Ammeen D, Engh GA. Outcome of knee arthroplasty in patients with poor preoperative range of motion. Clin Orthop Relat Res
17. Montgomery WH, Insall JN, Haas SB, Becker MS, Windsor RE. Primary total knee arthroplasty in stiff and ankylosed knees. Am J Knee Surg
18. Mullen JO. Range of motion following total knee arthroplasty in ankylosed joints. Clin Orthop Relat Res
19. Naranja RJ, Lotke PA, Pagnano MW, Hanssen AD. Total knee arthroplasty in a previously ankylosed or arthrodesed knee. Clin Orthop Relat Res
20. Rajgopal A, Ahuja N, Dolai B. Total knee arthroplasty in stiff and ankylosed knees. J Arthroplasty
21. Ritter MA, Campbell ED. Effect of range of motion in the success of total knee arthroplasty. J Arthroplasty
22. Ritter MA, Stringer EA. Predictive range of motion after total knee replacement. Clin Orthop Relat Res
23. Ryu J, Saito S, Yamamoto K, Sano S. Factors influencing the postoperative range of motion in total knee arthroplasty. Bull Hosp Jt Dis
24. Schurman JR, Wilde AH. Total knee replacement after spontaneous osseous ankylosis: a report of 3 cases. J Bone Joint Surg Am
25. Sculco TP. Management of the stiff knee. In: Callaghan JJ, Rosenberg AG, Rubash HE, Simonian PT, Wickiewicz TL, eds. The Adult Knee
; Philadelphia, PA: Lippincott Williams & Wilkins; 2003;1333-1340.
26. Spicer DD, Curry JI,Pomeroy D. Badenhausen WE Jr, Schaper LA, Suthers KE, Smith MW. Range of motion after arthroplasty for the stiff osteoarthritic knee. J South Orthop Assoc
. 2002;11: 227-230.