Pigmented Villonodular Synovitis does not Influence the Outcome Following Cruciate-retaining Total Knee Arthroplasty: A Case- Control Study with Middle-term Follow-up Minimum 5-Year Follow-up

Background: Pigmented villonodular synovitis (PVNS) is a rare synovial disease with benign hyperplasia, which has been successfully treated with total knee arthroplasty (TKA). The purpose of this study was to investigate the middle-term follow-up results of cruciate-retaining (CR) total knee arthroplasty (TKA) in PVNS patients. Methods: 17 patients with PVNS were treated with CR TKA between 2012 and 2014. 68 patients with osteoarthritis (OA) who received CR TKA as control group. The two groups matched in a 4:1 ratio based on age,gender, body mass index (BMI) and follow-up time. In the two groups, the range of motion (ROM), Knee Society Score (KSS), revision rate, disease recurrence, complications and the survivorship curve of Kaplan-Meier implant were compared. Results: All patients were followed up at least 7 years. At the last follow-up, no patients with recurrence of PVNS were found. There was no difference in KSS between the two groups, and no patients experienced infection and death. In the PVNS group, 3 patients had stiffness complications but did not need revision. 1 patient underwent revision surgery due to periprosthetic fracture. At 7 years,the implant survivorship was 90.0% in the PVNS group and 96.6% in the control group (p = 0.54). Conclusions: This study demonstrates that the function of PVNS patients who received CR TKA has been significantly improved. The survival rate of implants in PVNS patients is similar to that of OA patients. Consequently, PVNS patients who treated with CR TKA may be an achievable option. However, patients with PVNS should pay more attention to the occurrence of postoperative stiffness complications. cruciate cruciate-substituting (PS); BMI: KSS: knee society score; PCL: cruciate

had stiffness complications but did not need revision. 1 patient underwent revision surgery due to periprosthetic fracture. At 7 years,the implant survivorship was 90.0% in the PVNS group and 96.6% in the control group (p = 0.54).
Conclusions: This study demonstrates that the function of PVNS patients who received CR TKA has been significantly improved. The survival rate of implants in PVNS patients is similar to that of OA patients. Consequently, PVNS patients who treated with CR TKA may be an achievable option. However, patients with PVNS should pay more attention to the occurrence of postoperative stiffness complications.
Total knee arthroplasty (TKA) has been successfully used to improve function and relieve pain in PVNS patients. According to some articles, the survival rate of prosthesis ranges from 80% to 85% in PVNS patients who have been followed up for more than 10 years [9].
In addition, the incidence of postoperative complications of TKA also provided acceptable results in a large number of patients diagnosed with PVNS [10]. However, it is still controversial to retain or substitute posterior cruciate ligament (PCL) in TKA when treated with PVNS patients. Due to limited case studies of knee PVNS, there was few data have been available on the clinical outcomes of PVNS patients who underwent cruciateretaining (CR) TKA. Moreover, there was a lack of evaluation of the results of control group and middle-term study in previous studies. Therefore, the intention of this research was to investigate the middle-term clinical outcomes of patients with PVNS who underwent CR TKA, and compared with a group of patients with knee osteoarthritis (OA) who underwent CR TKA.

Study cohort
With the approval of the Institutional Review Committee, we conducted a retrospective study, 17 patients diagnosed as knee PVNS (PVNS group), and 68 patients diagnosed as knee OA (control group), all patients were treated with CR TKA between 2012 and 2014. In the PVNS group, according to Jaffe's classification, all patients were diagnosed as diffuse PVNS by pathology [11]. The two groups matched in a 4:1 ratio based on age, gender, body mass index (BMI) and follow-up time. Before surgery and at last follow-up after surgery including range of motion (ROM), Knee Society Scores (KSS) [12], disease recurrence, revision rate, complications and the survivorship curve of Kaplan-Meier implant were compared between the two groups.

Clinical and radiographic features of patients with PVNS before CR TKA
All patients with PVNS had knee pain before receiving CR TKA. 7 patients (41%) underwent at least one synovectomy. 5 patients (29.4%) were associated with patella dislocation or subluxation, and 10 patients (59%) had limited knee mobility with an average ROM of 89°( range 78°-101°).
X-ray showed typical features of the end-stage knee joint in patients with PVNS, including narrowing of joint space and cystic destruction (Fig. 1). The joint space almost disappeared in 6 patients.

Surgical treatment
All surgeries were performed by the same surgeon. All patients underwent standard midline skin incisions and medial parapatellar arthrotomy. For patients with PVNS, the pathological synovium was radically resected and histopathologically assessed. Firstly, the proximal tibia was resected on the coronal plane perpendicular to the tibial axis, and the sagittal plane was inclined backwards by 6-8 °, preserving PCL. The distal femur was excised with 5-7 ° valgus. Provide sufficient space between the femur and tibia for synovectomy. All patients received the same type of CR prosthesis (LINK, Germany, Gemini MK II). There were no restrictions placed on range of motion, activity postoperatively or weight-bearing status.

Statistical analysis
All statistical data were used Shapiro-Wilks test to check the normality of continuous variables. If the data were normally distributed, the two groups would be compared using student t-test; on the contrary, the non-parametric test would be performed. And the chisquared test or Fisher's exact test was performed for categorical variables. Implant survivorship was generated with 95% confidence intervals (CI) by Kaplan-Meier method.
The data were analyzed with the SPSS 19.0. p < 0.05 was considered significant.

Results
All patients were followed up at least 7 years ( Table 1). The results of CR TKA in the two groups were shown in Table 2. During the last follow-up, no clinical or radiological signs of prosthetic loosening were found in the PVNS group, indicating no recurrence of PVNS ( Figure 2). There was no notable difference in ROM, knee society clinical and functional scores between the two groups before surgery and at last follow-up after surgery. All patients were able to exercise moderately without crutches. In the PVNS group, 1 patient underwent revision because of periprosthetic fracture, and compared with control group, 3 patients had stiffness complications 1 year after operation (17.6% vs 1.5%, P = 0.005), but no revision was needed. In the control group, 2 patients underwent revision (1 unstable revisions and 1 infection) (p = 0.56). At 7 years, the implant survivorship without any revision was 90.0% in the PVNS group and 96.6% in the control group, there was no notable difference between the two groups (p = 0.54) (Figure 3).

Discussion
This study showed that PVNS patients who received CR TKA had similar survival rates and functional outcomes as those patients diagnosed with knee OA who received CR TKA. In the average follow-up of 7.2 years, no infection, osteolysis, and knee instability were found in PVNS patients. In addition, there was no evidence of PVNS recurrence. However, 6 the risk of postoperative stiff complications increased in patients with PVNS compared with the OA patients.
PVNS is a rare benign proliferative knee joint disease. Although the pathogenesis of PVNS remains unclear, some researchers believe that traumatic bleeding may be one of the causative factors [13,14]. In previously case report, localized pigmented villonodular synovitis presenting as recurrent dislocation of the patella [15]. This is consistent with our study, in our study, 5 patients (29.4%) with PVNS had a history of patella dislocation or subluxation. Therefore, we believe that patella dislocation or subluxation may be a susceptible factor for knee PVNS. In the future clinical work, we should give sufficient attention to such patients.
Patients with PVNS who have symptomatic knee are usually treated. Because it is widely present in the anterior chamber of the knee joint, it can be treated by arthroscopic debridement for focal PVNS. However, diffuse PVNS require combined surgery, either through arthroscopy or open surgery [3,16]. If PVNS extends beyond the joint, an open surgery is required [16]. Although open surgery can reduce the local recurrence rate of diffuse PVNS, these procedures may lead to a high incidence of knee stiffness after surgery [17]. The ROM after TKA is associated with preoperative knee ROM [18], and in our study, we found that PVNS patients who underwent open synovectomy, the knee ROM was poor after CR TKA. In addition, we also found that the incidence of postoperative knee stiffness in patients with PVNS is higher than that of ordinary OA patients, even if they have not received other surgery before TKA, so we should pay more attention to the postoperative rehabilitation of patients with PVNS.
TKA is the most effective treatment for end-stage PVNS in patients who have severe OA due to PVNS progression, and the recurrence and revision rate of PVNS are lower than that with simple synovectomy [9]. The long-term results of TKA treatment of PVNS have been 7 well recorded, and some studies have reported excellent long-term survival of TKA in patients with PVNS [9,10]. However, it is still controversial to retain or substitute PCL in TKA when treated with PVNS patients. Due to limited case studies of knee PVNS, there is currently little data on the results of CR TKA in these patients.
During primary TKA, two principal designs are used: cruciate-retaining (CR)TKA and cruciate-substituting (PS) TKA. Compared with PS-TKA, the CR-TKA has been widely used because it improves the knee's ability to exercise, preserves the knee's proprioception, and increases the knee ROM and stability during knee extension and flexion [19,20].
Although PVNS and rheumatoid arthritis have different types of inflammation and mechanisms of joint destruction, they all produce chronic inflammation environment in joints, so the two have some comparability to some extent [21]. Scott and his colleagues [22] pointed out that 95% of RA patients had complete PCL during TKA surgery and believed that PCL should be preserved during surgery to maximize femoral rollback. In addition, it has been informed that satisfactory clinical and radiological outcomes have been obtained in RA patients who were followed up for an average of 10.5 years with CR TKA [23]. Miller [24] evaluated long-term outcomes of patients with RA who were followed up for 20 years after CR TKA. For any reason, the 20-year implant survival rate was 69%.
They believe that PCL dysfunction is rarely the cause of surgical failure [24]. This is consistent with our current research results. In our study, patients with PVNS who received CR TKA achieved excellent mid-term follow-up outcome.
PVNS most frequently affects the knee, although there are long-term follow-up studies and short-term complications in the previous literature on the evaluation of PVNS in arthroplasty, there is no clear middle-term follow-up control study. Previous TKA treatment in patients with PVNS was a minor cohort study, primarily to assess implant survival and function, and did not quantify the risk of postoperative complications, which 8 may be due to fewer patients [9,25]. Although Houdek et al [9] did not compare the incidence of complications to the control group, their most common complication in this study was the loss of knee ROM, which was consistent with the increased risk of stiffness found in our research. According to previous reports, the risk of revision of PVNS patients is as high as 21%, which is significantly higher than the incidence of published primary TKA for OA [9,26]. In our research, only 1 patient underwent revision because of periprosthetic fracture. In addition, the implant survivalship without any correction for 7 years after CR TKA was 90.0% and we did not find any local recurrence, similar to the OA patients who received CR TKA. However, the revision rate may become inconsistent with additional long-term follow-up.
Radiotherapy and chemotherapy may be a viable option when surgery fails to eradicate PVNS or recurrence. Medium-dose external irradiation (30-35 Gy) combined with surgical resection can reduce the recurrence rate in patients with extensive or invasive diseases [27,28]. In recent years, significant advances have been made in the treatment of diffuse PVNS [29,30]. Since PVNS often overexpress colony-stimulating factor 1 (csf1), receptortargeted chemotherapeutic drugs (csf1r) may be an effective treatment [29,30]. Although these drugs were not used in patients in this series of studies, it is believed that young patients may consider using them to alleviate symptoms and delay TKA for as long as possible.
Our research has several limitations. First, this is a small sample retrospective study, which has its potential bias and weaknesses. A prospective study should be established to objectify these findings. Second, because the PVNS patients in our research were treated with CR TKA, we were unable to compare the efficacy of different prostheses, such as PS TKA, semi-constrained or rotating hinge prostheses.

Conclusions 9
In conclusion, our study provided encouraging results for the middle-term survival of patients with PVNS who received CR TKA. Compared to OA patients, end-stage PVNS of knee treated with CR TKA had similar restoration of function outcome and implant survivorship. Our results suggested that PVNS patients who treated with CR TKA may be a achievable option. However, patients with PVNS should pay more attention to the occurrence of postoperative stiffness complications.

Ethics approval
This study was approved by the Third Hospital of Hebei Medical University and followed the Declaration of Helsinki. Informed consent was received from all patients.

Consent for publication
Not applicable.

Availability of data and materials
The detailed data and materials of this study were available from the corresponding author through emails on reasonable request.