Five-year radiographic and clinical outcomes after arthroscopic synovectomy of the ankle in rheumatoid arthritis: A clinical trial

Abstract

Background

Rheumatoid arthritis (RA) affects ankle joints in up to one half of patients with established disease, causing inflammation and damage. Arthroscopic synovectomy removes inflamed tissue to improve joint function, but long-term outcomes and the potential role of adjunctive therapies are limited.

Materials and methods

A total of 176 patients who had a preoperative diagnosis of RA according to the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria, radiographic Larsen grade ≤ 3, and who underwent arthroscopic synovectomy from May 2013 to May 2019 were prospectively enrolled. Weight-bearing anteroposterior and lateral plain radiographs of the ankle were performed annually after initial surgery. The Larsen grade was used to evaluate the progression of ankle joint damage in RA, and patient-reported outcomes (American Orthopaedic Foot and Ankle Society [AOFAS] Ankle–Hindfoot Scale and the Foot and Ankle Outcome Score [FAOS]) were collected annually. The primary outcome measure was 5-year AOFAS score. Baseline characteristics, including age, body mass index (BMI), duration of symptoms before surgery, Larsen grade before surgery, and other potentially related factors, including number of platelet-rich plasma (PRP) injections and change in BMI from baseline, were recorded.

Results

There were 138 patients included, all with minimum 5-year follow-up data. The overall reoperation rate was 13% (95% confidence interval [CI] 6.8–18.9%; 18 of 138). According to multivariable analysis, 5-year AOFAS scores were associated with number of PRP injections (correlation coefficient = 2.09 [95% CI 1.47–2.71]; P < 0.001), duration of symptoms before surgery (correlation coefficient = 0.42 [95% CI 0.14–0.70]; P = 0.01), Larsen grade before surgery (correlation coefficient = 0.28 [95% CI 0.06–0.49]; P = 0.034), and mean BMI change from baseline (correlation coefficient = −1.23 [95% CI −1.57 to −0.89]; P < 0.001). When comparing the number of PRP injections (0, 1–2, or ≥ 3), patients who had serial PRP injections (≥ 3) had diminished functional and radiographic deterioration over time.

Conclusions

Arthroscopic synovectomy improves symptoms for ankle RA and appears to slow—but not halt—radiographic deterioration over 5 years. Serial PRP injections and reduction in BMI from baseline could be associated with better clinical outcomes and slower joint degeneration, which needs to be verified by randomized controlled trials.

Level of evidence

Level II, prospective cohort study.

Trial registration Research Registry, researchregistry10878. Registered 24 November 2024—retrospectively registered, http://researchregistry.knack.com/researchregistry10878

Supplementary Information

The online version contains supplementary material available at 10.1186/s10195-025-00878-3.

Introduction

Rheumatoid arthritis (RA) is a chronic, autoimmune disease characterized by inflammation, pain, and progressive joint destruction [1, 2]. The etiology and mechanism of RA are unclear; however, it is believed to result from complex interactions between genetic and environmental factors [3, 4], leading to a breakdown of immune tolerance and symmetric involvement of multiple joints, including the ankles [5–7]. Tibiotalar arthritis occurs in 15–52% of patients with adult-onset RA, and the occurrence of juvenile-onset RA may exceed 60% [8, 9]. Ankle involvement in RA can be particularly debilitating, causing pain, stiffness, and loss of function [1, 10–13]. Pharmacological therapies, such as disease-modifying antirheumatic drugs (DMARDs) and biologics, are the cornerstone of treatment aimed at reducing inflammation, slowing disease progression, and preserving joint function [14, 15]. However, despite these advancements, some patients with RA continue to experience persistent ankle symptoms despite medical management [16, 17].

Arthroscopic synovectomy is a minimally invasive surgical procedure that has been used to treat ankle symptoms in patients with RA [18, 19]. This procedure involves the removal of inflamed synovial tissue through small incisions using arthroscopic instruments. By reducing the inflammatory burden within the joint, arthroscopic synovectomy can improve symptoms, slow radiographic progression, and may potentially delay the need for more aggressive surgical interventions, such as ankle arthrodesis or total ankle arthroplasty [20–24]. It is hypothesized that arthroscopic synovectomy, based on the internationally established Larsen grading system, would provide lasting pain relief and improved ankle function for patients in the early stages of the disease [25, 26].

Previous studies [18, 27] have reported favorable short-term outcomes following arthroscopic synovectomy in patients with RA. However, long-term follow-up data are limited, and the role of adjunctive therapies, such as platelet-rich plasma (PRP), in enhancing outcomes remains controversial. PRP is a concentrated suspension of autologous platelets in a small volume of plasma. It contains high concentrations of growth factors and cytokines that have been shown to promote tissue healing, reduce inflammation, and stimulate cell proliferation and differentiation [5, 28–30]. Many researchers [31–37] have investigated PRP application in RA and suggested its benefits for RA management with the improvement of local joint inflammation by inhibiting inflammatory factors in synovial fluid.

This study aims to evaluate the 5-year radiographic and clinical outcomes following arthroscopic synovectomy of the ankle in patients with early stage RA. In addition, this study will explore the potential role of PRP as an adjunctive therapy in this patient population. By doing so, this study intends to provide valuable insights into the long-term effectiveness of arthroscopic synovectomy and the potential benefits of PRP in the management of ankle symptoms in patients with RA.

Materials and methods

Study design and patients

From May 2013 to May 2019, a total of 176 patients who had a preoperative diagnosis of RA according to the 2010 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria [38] and underwent unilateral arthroscopic synovectomy of the ankle were prospectively enrolled in this study. The ACR/EULAR 2010 classification criteria for RA clarified that RA is diagnosed when a patient has at least one joint with unexplained synovitis and a combined score of ≥ 6 (maximum of 10) across four domains (number and location of affected joints, serologic category, abnormal acute phase reactions, and the 6-week duration requirement) [38]. Patients were eligible if they experienced synovitis-related ankle pain and dysfunction that had not improved with systemic treatment for at least 6 months, and radiographic alterations that were grade 3 or lower using the Larsen grading system (Table 1) [25]. Larsen grade ≤ 3 was chosen to ensure minimal joint space narrowing and no cartilage degeneration, aligning with the concept of early intervention. Patients were excluded from enrollment if they had a history of trauma, previous ankle surgery, severe radiographic destruction or gross deformity, or previous intra-articular injection of either a steroid within the last 3 months or PRP within the past 6 months. Patients with Larsen grade ≥ 4 were excluded owing to irreversible cartilage loss, where arthroscopic synovectomy offers limited benefit compared with total ankle arthroplasty or arthrodesis [15, 39]. All patients continued their existing antirheumatic therapy comprising nonsteroidal anti-inflammatory drugs (NSAIDs) and conventional DMARDs (methotrexate, sulfasalazine, hydroxychloroquine, or penicillamine). During follow-up, rheumatologists could adjust the drug dose or switch the type of biologic agent according to the condition, but the data of treatment adjustment were not systematically recorded. The protocol for this study was approved by the ethics committee of the institution, and written consent was obtained from all participants. This prospective cohort study complied with the code of ethics of the World Medical Association (Declaration of Helsinki) and was performed in line with the Strengthening the Reporting of Observational Studies in Surgery (STROCSS) criteria [40].

Table 1.

Radiographic grading system for rheumatoid arthritis

Grade	Description
0	Intact bony outlines and normal joint space
1	Erosion < 1 mm in diameter or joint space narrowing
2	One or several small erosions, diameter > 1 mm
3	Marked erosions
4	Severe erosions, where there is usually no joint space left and the original bony outlines are partly preserved
5	Mutilating changes, where the original bony outlines have been destroyed

Data from Larsen et al. [32]

Surgical techniques

A tourniquet was used to prevent intraoperative bleeding after general anesthesia was administered. Traditional anteromedial and anterolateral portals are used for diagnostic ankle arthroscopy. A standard small-joint arthroscopy setup was used in this procedure, including a 2.7–4.0-mm, 30° arthroscope for optimal visualization in the tight ankle joint space, a pump-controlled irrigation to maintain joint distension, small-joint shavers (2.0–3.5 mm), radiofrequency probes for synovial ablation, and graspers for tissue removal. The pathologically hypertrophic synovium was systematically removed from all compartments using mechanical shavers and thermal devices, prioritizing areas of maximal inflammation (e.g., medial gutter and posterior recess). More attention should be paid to careful cartilage inspection and minimal bone resection to avoid iatrogenic instability. Rough articular cartilage was shaved, and loose debris was removed. The ultimate objective was to perform a complete synovectomy of all pathologic tissue, and hemostasis was achieved before completion of the surgery. The synovium located on the posterior malleolus was also debrided using the traditional posterolateral and posteromedial portals. After surgery, walking boots were used to allow weight bearing as tolerated. It is worth noting that, because of a lack of strong clinical evidence, the adjunct therapy of intraoperative and/or postoperative injections of PRP was based on the patients’ preference after consulting with treating surgeons.

PRP was prepared using a commercially available product, which commonly yielded a platelet concentration factor of more than six times over whole blood and with approximately 80% platelet recovery rate. The leukocytes in this PRP are no more than four times the concentration of the whole blood. For each patient, 40 mL of whole blood was collected into the blood collection tubes with the anticoagulant. The PRP kits used two-time spins to collect 3–4 mL of PRP (first time: 800g 10 min and second time: 1100g 10 min) at room temperature by a nurse. To assess the efficiency of production, DEPA classification of PRP was adopted. The DEPA classification of PRP is based on four different components: (1) the dose of injected platelets, (2) the efficiency of the production, (3) the purity of the PRP obtained, (4) the activation process [41]. To improve transparency and reproducibility, Minimum Information for Studies Evaluating Biologics in Orthopaedics (MIBO) was also used for specific biological detail with the delivery or administration of PRP [42]. Immediately after PRP preparation, intraarticular injection of PRP was performed.

Baseline and outcome data collection

Baseline characteristics, including age, sex, education level, body mass index (BMI), and symptom duration before surgery, were self-reported by patients. Patients received standard weight-bearing anteroposterior and lateral plain radiographs of the ankle that were interpreted without knowledge of the PRP status at baseline and annually after surgery. The Larsen grading system was used to grade radiographic rheumatoid arthritis (RRA). The three observers were all radiologists with more than 5 years of experience in bone and joint imaging diagnosis. Before the evaluation, they received 2 h of training on Larsen grading criteria. After training, the Fleiss’ kappa value was calculated through the pre-experiment test consisting of 20 samples. The results of the preliminary experiment showed a kappa of 0.78 (95% confidence interval [CI] 0.72–0.84), indicating good consistency. When the interrater difference was more than one grade, a consultative mechanism was activated: two junior radiologists reviewed the images together, and if there was still disagreement, the lead radiologist (with 15 years of experience) made the final decision. This occurred in 12% of the total estimates (17 of 138 patients), and the mean difference between final adjudication and initial assessment was 0.6 grades. Three assessors were also completely unaware of the patients’ PRP treatment status (only the need to assess radiographic progression of RA). The treatment-assignment information was kept by the independent study staff, and assessors were able to identify imaging data only by number and could not correlate clinical data. Blinding was lifted after statistical analysis was completed.

For patient-reported outcomes, well-validated self-report instruments were adopted, including the American Orthopaedic Foot and Ankle Society (AOFAS) Ankle–Hindfoot Scale [43] and the Foot and Ankle Outcome Score (FAOS) [44, 45]. These outcomes were collected annually. The AOFAS the primary patient-reported outcome of this study, which is a validated, nine-item scale (range, 0–100 points; higher scores indicate less pain and better function), consisting of three subscales (pain, function, and alignment). FAOS consists of five subscales: pain, other symptoms, activities of daily living, sport and recreation function, and foot- and ankle-related quality of life. The minimal clinically important difference (MCID) is the smallest change in a clinical outcome measure that is considered to be meaningful to patients [46].

Statistical analysis

All statistical tests were two-sided and conducted at the 0.05-level of significance. All statistical analyses were conducted using SPSS, version 24.0, in this study. Normality assessment was conducted via Shapiro–Wilk tests and Q–Q plots of residuals and formal testing for heteroscedasticity using Breusch–Pagan tests (P > 0.05 for all models). Continuous variables are represented as mean ± standard deviation, and categorical variables are represented as number (percentage). For univariable analysis, the comparison of 5-year AOFAS with continuous and categorical variables was evaluated using the Kruskal–Wallis test and linear regression, respectively. When the yielded P-value was < 0.2 in the univariable analysis, that variable was included in the multivariable analysis. Multivariable linear models were then constructed for the association of these variables with 5-year AOFAS. For secondary analyses and subgroup analyses, false discovery rate (FDR) control (Benjamini–Hochberg procedure) was used to correct for multiple comparisons, with a threshold of 0.05. Propensity score matching (PSM) was used to analyze the three groups of PRP injections (0, 1–2, ≥ 3), with the number of PRP injections as a covariate (1:1 matching, matching coefficient = 0.02).

Results

Of the initial 176 patients, there were 29 losses to follow-up within 5 years, and 9 incomplete follow-up records were available. Ultimately, 138 patients completed a minimum 5-year follow-up and were included for analysis. The 29 patients who were lost to follow-up (16.5%) included 18 patients (62.1%) who relocated outside the catchment area, 10 patients (34.5%) who declined further participation owing to personal reasons, and 1 patient (3.4%) who died from unrelated causes (Fig. 1). Supplementary Table S1 presents the baseline characteristics between completers and noncompleters; all variables were well balanced (absolute standardized mean difference [SMD] < 0.1), so the findings are unlikely to be substantially biased by attrition. Comparative analysis showed no significant differences (P > 0.20) in baseline characteristics between completers and noncompleters. Sensitivity analyses of 12 revision cases showed that 5-year AOFAS improved from 68.2 ± 12.1 to 82.4 ± 9.3 (P < 0.001), while the reoperation rate decreased to 4.3% (6/138). Minimum observed scores (AOFAS = 45) were also assigned to all revision cases, and mean 5-year AOFAS remained significantly improved (76.8 ± 11.6 versus baseline, P = 0.002) through worst-case scenario imputation. Among the 138 patients, 42 patients did not receive any intra-articular injection of PRP, 60 patients received single/double injection of PRP, and 36 patients received serial (≥ 3, range 3–9) PRP injections. Comparative analysis showed no significant differences (P > 0.20) in baseline characteristics between the non-PRP group versus PRP group (Table 2). During follow-up, 32% of patients reported DMARD dose escalation, and 18% switched biologic agents (e.g., etanercept to adalimumab) owing to poor disease activity control. Because treatment adjustment data were not fully recorded, it was not possible to quantify the specific effect on AOFAS scores or joint degeneration. Subgroup analysis showed no significant difference in the change in AOFAS score between the treatment-adjusted group and the treatment-unadjusted group (P = 0.12), with caution regarding sample size limitations (n = 43 in the adjusted group). The overall reoperation rate was 13% (18 of 138). The procedures included 12 patients who received revision arthroscopy and 6 who underwent ankle arthrodesis. Notably, for patients who received revision arthroscopic synovectomy within 5 years, the outcome of the last follow-up was carried forward for the final analysis.

Fig. 1.

Flowchart of the study

Table 2.

Comparison of baseline characteristics between non-PRP group versus PRP group

	Non-PRP group (N = 42)		PRP group (N = 96)		P-value
	N	%	N	%
Sex
Male	13	31.0	29	30.2	0.120
Female	29	69.0	67	69.8
Larsen grade before surgery
0, 1	25	59.5	56	58.3	0.056
2, 3	17	40.5	40	41.7
Age at baseline, years
Mean ± SD	51.0 ± 8.2		51.9 ± 7.6		0.546
BMI at baseline
Mean ± SD	24.6 ± 3.6		25.7 ± 3.1		0.090
Duration of symptoms before surgery, months
Mean ± SD	12.94 ± 8.23		14.05 ± 7.15		0.451
Preoperative AOFAS score
Mean ± SD	63.18 ± 5.28		61.85 ± 6.24		0.201
Preoperative FAOS score
Mean ± SD	52.27 ± 12.07		51.11 ± 10.85		0.594

This table shows insignificant differences regarding sex, Larsen grade before surgery, age at baseline, body mass index (BMI) at baseline, duration of symptoms before surgery, preoperative AOFAS score, and preoperative FAOS score between the two groups

SD standard deviation

Five-year clinical and radiographic outcomes after arthroscopic synovectomy

Regarding patient-reported outcomes, scores on the AOFAS (year 1: 89.04 ± 5.79; year 2: 88.32 ± 5.64; year 3: 84.11 ± 6.36; year 4: 81.32 ± 6.78; and year 5: 78.51 ± 6.60) and FAOS (year 1: 82.27 ± 6.41; year 2: 81.1 ± 5.74; year 3: 77.22 ± 5.53 year 4: 74.26 ± 6.32; and year 5: 70.67 ± 5.46) were both generally stable in the initial 2 years after surgery, followed by a gradual decrease in scores over time. The distribution characteristics of AOFAS scores were: median 1-year follow-up, 89 (interquartile range [IQR] 85–92) and median 5-year follow-up, 78 (IQR 75–82); only 12% of patients with scores > 90; and no significant left or right bias (coefficient of skewness = −0.32). The mean AOFAS and FAOS scores at the final follow-up were still significantly better than the baseline assessments, reaching the MCID.

In the formal evaluation, the ankle radiographs of 138 patients were graded independently by three reviewers, with Fleiss’ kappa of 0.82 (95% CI 0.78–0.86), indicating excellent interreader reliability. The intraclass correlation coefficients (ICC) of grade 0–1, 2–3, and 4–5 were 0.88, 0.79, and 0.91, respectively. The blind design effectively reduced observer bias. Blinding was maintained until data analysis was completed to ensure objective interpretation of the results. On preoperative plain radiographs, 24 patients (17.4%) had RA Larsen grade 0, 57 patients (41.3%) had Larsen grade 1, 33 patients (23.9%) had Larsen grade 2, and 24 patients (17.4%) had Larsen grade 3. At the final follow-up, Larsen grade 1 was present in 75 patients (54.3%), Larsen grade 2 in 30 (21.7%), Larsen grade 3 in 12 (8.7%), and Larsen grade 4 in 21 (15.2%). A total of 60 patients (43.5%) showed a progression in degenerative changes in their ankles when compared with the preoperative radiographs, while 72 patients (56.5%) showed no change at all (Fig. 2).

Fig. 2.

Arthroscopic synovectomy of the ankle in rheumatoid arthritis and radiological images collected preoperatively and postoperatively. A One patient with RA who underwent arthroscopic synovectomy of the ankle with serial PRP injection was registered with Larsen grade 0 preoperatively and Larsen grade 1 at 5-year follow-up; B one patient with RA underwent arthroscopic synovectomy of the ankle with no PRP injection, with preoperative Larsen grade 1 and 5-year follow-up Larsen grade 3

Univariable analysis of prognostic factors

Table 3 presents demographic and clinical characteristics and evaluation using univariable analyses. The univariable analysis yielded P-values < 0.2 for the following variables: age at baseline, BMI at baseline, mean BMI change from baseline, duration of symptoms, Larsen grading, and number of PRP injections. The multivariable analysis was then conducted with these variables included.

Table 3.

Univariable correlation between demographic/clinical characteristics and 5-year AOFAS

Categorical variables	AOFAS score (mean ± SD)	P-value*
Sex		0.211
Male (n = 42)	78.7 ± 6.5
Female (n = 96)	77.7 ± 6.4
Number of PRP injections#		0.0013
0 (n = 42)	76.7 ± 4.9
1–2 (n = 60)	76.3 ± 4.8
≥ 3 (n = 36)	85.0 ± 3.4
Intraoperative PRP injection		0.293
No (n = 60)	78.7 ± 6.0
Yes (n = 78)	78.6 ± 6.6
Larsen grade before surgery		0.013
0, 1 (n = 81)	81.1 ± 5.9
2, 3 (n = 57)	75.1 ± 5.4

Continuous variables	Mean ± SD	Correlation coefficient	P-value†
Age at baseline, years	51.6 ± 7.9	−0.225	0.017
BMI at baseline	25.3 ± 3.3	−0.235	0.079
Mean BMI change from baseline**	0.74 ± 4.6	−1.469	< 0.001
Duration of symptoms before surgery, months	13.61 ± 7.77	0.913	0.021
Preoperative AOFAS score	62.74 ± 5.63	0.017	0.814
Preoperative FAOS score	51.83 ± 11.39	0.011	0.613

Boldface P-values indicate variables included in the multivariable analysis (P < 0.2)

^*Kruskal–Wallis test

^†Univariable linear regression

^#Number of injections during the study period (intraoperatively and postoperatively)

^**Mean BMI change from baseline = (mean BMI from years 1 through 5) − BMI at baseline

Multivariable analysis of prognostic factors

The main conclusion is based on multivariate linear models that have passed tests of heteroscedasticity (Breusch–Pagan P > 0.05) and collinearity diagnostics (variance inflation factor [VIF] < 5), and the robustness of the results is not directly affected by multiple comparisons. Sensitivity analysis showed that the difference between the unadjusted and adjusted effect estimates was < 10%, suggesting that the main conclusion was reliable. In the multivariable analysis (Table 4), the 5-year AOFAS was associated with number of PRP injections (correlation coefficient = 2.09 [95% CI 1.47–2.41]; P < 0.001), duration of symptom before surgery (correlation coefficient = 0.42 [95% CI 0.14–0.70]; P = 0.01), Larsen grade before surgery (correlation coefficient = 0.28 [95% CI 0.06–0.49]; P = 0.034), and mean BMI change from baseline (correlation coefficient = −1.23 [95% CI, −1.57 to −0.89]; P < 0.001).

Table 4.

Multivariable linear regression analysis of 5-year AOFAS

	Correlation coefficient (95% CI)	P-value
Age at baseline	−0.06 (−0.36 to 0.25)	0.79
BMI at baseline	−0.14 (−0.44 to 0.17)	0.36
Mean BMI change from baseline	−1.23 (−1.57 to −0.89)	< 0.001
Duration of symptoms before surgery, months	0.42 (0.14–0.70)	0.01
Larsen grade before surgery	0.28 (0.06–0.49)	0.034
Number of PRP injections	2.09 (1.47–2.71)	< 0.001

Boldface P-values indicate significant association with 5-year AOFAS score (P < 0.05)

Comparison of functional and radiographic outcomes by number of PRP injections

This study further explored the role of intra-articular PRP injection by comparing functional and radiographic outcomes by the number of PRP injections. In this study, the PRP kit produced a mean 1.143 billion platelets per mL with a mean (5.9×) platelet enrichment ratio and 65% platelet recovery rate, and a mean 26.2 million leukocytes per mL with a mean (4.1×) leukocyte enrichment ratio and 50% leukocyte recovery rate. The final DEPA score is BCC, which means high dose of injected platelets, low efficiency of the process, and heterogeneous PRP. After propensity score matching, the baseline characteristics were balanced (SMD < 0.1), and the results showed that serial PRP injections (≥ 3 times) were still significantly associated with improved 5-year AOFAS scores (Table 4). In this study, clear and consistent trends were observed that the functional deterioration could be diminished by intra-articular injection of PRP (Fig. 3). The incidence of RRA progression (defined as at least a one-grade increase in Larsen grade) at 5 years was significantly lower for the serial PRP group (no injections: 27/42, 64.3%; 1–2 injections: 24/60, 40%; ≥ 3 injections: 9/36, 25%, P < 0.001).

Fig. 3.

Patient-reported outcomes AOFAS Ankle–Hindfoot Scale (A) and FAOS (B) were collected preoperatively and postoperatively. Considerable improvements in patient-reported outcomes were maintained over the 5-year follow-up period, although there was a gradual decline as time progressed. AOFAS American Orthopaedic Foot and Ankle Society, FAOS Foot and Ankle Outcome Score

Discussion

The current study is the largest prospective study with at least a 5-year follow-up regarding the arthroscopic synovectomy in the treatment of RA [18, 47–55]. In addition, an important strength of the current study was that the radiographic evaluation was performed on an annual basis. This study has evaluated the 5-year radiographic and clinical outcomes after arthroscopic synovectomy of the ankle in patients with early stage RA, with a particular focus on the role of PRP as an adjunct therapy. The results provide valuable insights into the long-term efficacy of this surgical procedure and the potential benefits of PRP in enhancing outcomes.

The findings of this study are consistent with previous studies that have reported significant improvements in pain, swelling, and function following arthroscopic synovectomy in patients with RA [18–20]. Specifically, a considerable enhancement in patient-reported outcomes was observed, as measured by the AOFAS Ankle–Hindfoot Scale and the FAOS. These improvements were maintained over the 5-year follow-up period, although there was a gradual decline in scores as time progressed. The gradual decline may be attributed to the progressive nature of RA, which can lead to ongoing joint damage despite successful synovectomy. On the basis of the previous study reporting the foot and ankle surgery, the MCID was set at 8.9 for the AOFAS score and 3.3–8.9 for the FAOS score [56, 57]. The results suggest that both AOFAS and FAOS scores were significantly improved even at the 5-year follow-up. Furthermore, the improvements of both AOFAS and FAOS scores at the final follow-up exceeded the upper limit of the range used for the MCID on both scales, suggesting its clinical significance. Although there is a theoretical range of AOFAS scores (0–100 points), the actual data distribution in this study did not show a significant ceiling effect (only 12% of patients scored > 90 points). The residual normality test (Shapiro–Wilk P = 0.12) and heteroscedasticity test (Breusch–Pagan P = 0.08) all met the premise of linear regression. Therefore, it was reasonable to use multivariate linear regression to analyze the association between variables and AOFAS scores. While direct comparative analysis was limited by study design, previous literature [18, 35, 47, 52, 53, 55, 58–62] shows that 13% reoperation rate compares favorably with those in open synovectomy studies (Table 5). This suggests that arthroscopic synovectomy may occupy a valuable niche in early stage RA ankle management.

Table 5.

Key outcomes regarding the management of rheumatoid arthritis previously reported in literature

Literature	Type of study	Interventions	Sample size	Outcomes	Re-intervention rate	Use of PRP
Blajovan et al. [45]	Comparative observational study	Arthroscopic synovectomy versus open synovectomy of joints	30 versus 23	HAQ 1.4 ± 0.4 versus 1.6 ± 0.5 at 6-month follow-up	3.3% (1/30) versus 8.7% (2/23) at 1-year follow-up	None
Liu et al. [60]	Cohort study	Open arthrolysis with hinged external fixation, total synovectomy, and ulnar nerve release and anteriorization	48	MEPS 87.40 at 10-year follow-up	2.1% (1/48) at 10-year follow-up	None
Badsha et al. [35]	Case series	PRP injection for knee and wrist joints	4	VAS and DAS 28 scores improved at 4 and 8 weeks after injection and sustained for up to 1 year of follow-up	0 at 5-year follow-up	Yes
Triolo et al. [52]	Cohort study	Arthroscopic synovectomy of the knee joint	57	Laurin score significantly improved from 3.91 ± 1.3 to 8.2 ± 2.0 after surgery	27.3% (18/57) at 4-year follow-up	None
Kanbe et al. [61]	Cohort study	Arthroscopic synovectomy plus capsular release of the shoulder joint	54	DAS 28 significantly improved from 4.29 ± 0.87 to 3.11 ± 0.92; the JOA score significantly improved from 36.65 ± 7.66 to 84.61 ± 12.74 at 5-year follow-up	66.7% (36/54) at 5-year follow-up	None
Lee et al. [53]	Cohort study	Arthroscopic synovectomy of the wrist joint	56	VAS decreased from 6.3 to 1.7, and the mean Mayo Wrist Score improved from 48 (5–70) to 76 (55–100)	25.0% (14/56) at 8-year follow-up	None
Choi et al. [18]	Case series	Arthroscopic synovectomy of the ankle joint	18	77.8% (14/18) clinical success at 5-year follow-up	0 at 5-year follow-up	None
Goetz et al. [58]	Case series	Combined arthroscopic and radiation synovectomy of the knee joint	32	Stable improvement of knee function for a minimum of 5 years, but surgical re-interventions were frequently observed at the 14-year assessment	15.6% (5/32) at 5-year follow-up	None
Tanaka et al. [55]	Comparative observational study	Arthroscopic synovectomy versus open synovectomy of the elbow joint	23 versus 23	For elbows with a preoperative arc of flexion of  < 90°, arthroscopic synovectomy provided significantly (P < 0.05) better function than open surgery at 13-year follow-up evaluation	Recurrent synovitis in six elbows versus in three elbows	None
Mäenpää et al. [59]	Cohort study	Open synovectomy of the elbow joint	103	No significant improvement in range of motion was detected, but pain relief and patient satisfaction were favorable	23.3% (24/103) at 5-year follow-up	None
Akagi et al. [62]	Case series	Open synovectomy for ankle joint	20	Only two ankles showed recurrence of synovitis at 15-year follow-up, and no patient complained of severe ankle pain disturbing the activities of daily life	0 at 15-year follow-up	None

PRP platelet-rich plasma, HAQ Health Assessment Questionnaire, MEPS Mayo Elbow Performance Score, VAS visual analog scale, DAS 28 Disease Activity Score using 28 joints, JOA Japanese Orthopaedic Association

Notably, this study also investigated the impact of PRP as an adjunct therapy during and after arthroscopic synovectomy. While the use of PRP in RA has been controversial, with some studies reporting beneficial effects [5, 31–33, 63] and others showing no significant improvement [64], the results of this study suggest that PRP use was associated with slower functional deterioration in observational data. Specifically, after controlling for baseline differences, patients who received serial PRP injections tended to correlate with better clinical outcomes and showed a slower rate of radiographic and functional deterioration compared with those who did not receive PRP. However, it should be noted that this was an observational cohort study not designed to definitively establish the efficacy of PRP in this context, and PRP assignment was based on patient preference rather than randomization. Therefore, causality cannot be established, and unmeasured residual confounding (e.g., patient adherence and socioeconomic status) may still have influenced the results. Further randomized controlled trials are needed to confirm these hypothesis-generating findings.

This study also highlights the importance of early synovectomy in achieving optimal clinical outcomes. According to previous literature, the radiographic progression rate compares favorably with progression in patients with similar-stage RA receiving DMARDs alone [65, 66]. Consistent with previous research [18], this study found that patients with shorter disease duration and minimal radiographic changes achieved the best results following arthroscopic synovectomy. This suggests that early intervention may be crucial in preventing or delaying the progression of joint damage and preserving ankle function in patients with RA [67]. It needs to be acknowledged that continued use of DMARDs/biologics may indirectly enhance the local efficacy of arthroscopic surgery by controlling the systemic inflammatory response. For example, patients in this study who received biologics (n = 58) had a smaller 5-year reduction in AOFAS scores (9.2 points) than those who received DMARDs alone (n = 80, 11.8 points), suggesting that biologics may protect joints. However, because treatment adjustments were not systematically recorded, this observation needs to be interpreted with caution. Future studies should use electronic medical record linkage to accurately capture changes in drug dose and duration of treatment.

In addition to the clinical outcomes, the radiographic changes over the 5-year follow-up period were also evaluated. The results showed that, while some patients experienced no change or even deterioration in their radiographic findings, a significant proportion of patients demonstrated improved manifestations in joint space narrowing, bone erosion, and synovitis. These improvements were more pronounced in patients who received serial PRP injections, suggesting to some extent that serial PRP injections might be associated with better-preserved joint structures.

It is worth mentioning that this study had a prospective design, which allowed us to systematically record adjunct therapies and accurately assess long-term outcomes. However, there are several limitations to this investigation. First, the sample size was relatively small, which may have affected the statistical power of analysis. Second, this study was not a randomized controlled trial, and therefore, the observed improvements cannot be definitively attributed to the use of PRP. Third, the decision on whether to undergo PRP injections was based on patient preference and recommendations from surgeons, and patients who had a good response to the first PRP injection might be more inclined to accept subsequent injections. It is conceivable that the PRP dose–response analysis and observed PRP–radiographic associations cannot establish causality owing to nonrandom allocation. A future confirmative randomized trial is needed to clarify this issue. Moreover, the analysis was limited by incomplete characterization of DMARD/biologic regimens because no attempt was made to standardize the patients’ medication regimens before or after the arthroscopic synovectomy. The current study did not incorporate a comparative analysis with the natural history of medically treated RA as well, and follow-up studies should include exploring the potential synergistic effects of combined surgical and pharmacological therapies. Finally, the long-term effects of PRP on joint structure and function and the optimal number of injections need to be further investigated in larger and more rigorous studies.

Conclusions

This study provides evidence to support the use of arthroscopic synovectomy as an effective treatment for early stage RA in the ankle. Findings support symptomatic benefit of ankle synovectomy and raise the hypothesis that adjunctive PRP may attenuate radiographic progression. However, these hypothesis-generating associations between serial PRP injections and improved outcomes need to be validated in randomized controlled trials, particularly to determine the optimal timing, dosage, and frequency of PRP injections and the long-term safety profile. In addition, larger studies with longer follow-up periods are required to fully understand the long-term effects of arthroscopic synovectomy and PRP on joint function and radiographic outcomes in patients with RA.

Abbreviations

RA

Rheumatoid arthritis

AOFAS

American Orthopaedic Foot and Ankle Society

FAOS

Foot and Ankle Outcome Score

BMI

Body mass index

PRP

Platelet-rich plasma

CI

Confidence interval

DMARDs

Disease-modifying antirheumatic drugs

ACR/EULAR

American College of Rheumatology/European League Against Rheumatism

NSAIDs

Nonsteroidal anti-inflammatory drugs

MIBO

Minimum information for studies evaluating biologics in orthopaedics

RRA

Grade radiographic rheumatoid arthritis

MCID

Minimal clinically important difference

FDR

False discovery rate

PSM

Propensity score matching

SMD

Standardized mean difference

ICC

Intraclass correlation coefficients

Author contributions

Z.S. had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. J.Z. and X.M. conceptualized and designed the study. C.Wu and F.Y. collected patients’ information and reviewed the literature. J.Z., C.Wa, and J.W. drafted the manuscript. Z.S. and X.M. revised the manuscript. All authors read and approved the final version.

Funding

This work was supported by Medical Engineering Cross Research Fund of Shanghai Jiao Tong University (grant no. YG2022ZD018); the application demonstration project of innovative medical devices in Shanghai in 2023 (grant no. 23SHS03600); and Shanghai Science and Technology Commission (grant no. 23015820500).

Data availability

No additional data are available.

Declarations

Ethics approval and consent to participate

The study was approved by the Institutional Ethics Committee of Shanghai Sixth People’s Hospital (approval no. 2010-136). Moreover, this study complied with the Code of Ethics of the World Medical Association (Declaration of Helsinki) for research involving humans. Written consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.