Long-term Outcomes of MCP Surface Replacement Arthroplasty in Patients With Rheumatoid Arthritis

Abstract

Background: Surface replacement arthroplasty (SRA) can be used in the treatment of rheumatoid arthritis (RA) affecting the metacarpophalangeal (MCP) joint. The authors of this study sought to investigate the outcomes of MCP SRA in patients with RA. Methods: Retrospective review of medical records and an institutional joint registry were used to gather data on 80 MCP SRAs performed in 27 patients with RA. Data collected included demographics, SRA revisions, reoperations, complications, pain, and MCP arc of motion. Results: The mean postoperative follow-up was 9.5 years (range, 2.1-20.5 years), with all SRAs achieving at least 2 years of follow-up. Thirteen digits (16%) underwent revision arthroplasty, and 29 (36%) required reoperation. The 5-, 10-, 15-, and 20-year rates of survival from implant revision were 95%, 85%, 80%, and 69%, respectively. The 5-, 10-, 15-, and 20-year rates of survival from overall reoperation were 80%, 65%, 55%, and 46%, respectively. Metacarpophalangeal joint arc of motion, grip strength, and pain levels significantly improved following surgery. Conclusions: Metacarpophalangeal SRA can offer benefit to patients with RA for improvement in function and pain. High overall reoperation rates remain concerning; however, most do not involve arthroplasty revision.

Introduction

Silicone metacarpophalangeal (MCP) joint arthroplasty has been used to treat symptomatic rheumatoid arthritis (RA) since the 1960s.^1-7 Silicone implants have been shown to improve function and pain at the MCP joint in some patients with RA.^8-11 Despite this, high rates of mechanical failure have been reported in silicone implants after long-term follow-up in patients with RA. ¹ Surface replacement arthroplasty (SRA) was developed as an alternative surgical option for arthritic conditions affecting the MCP joints (Figure 1).^12,13

Figure 1.

Metacarpophalangeal surface replacement arthroplasty implant.

Although outcomes of silicone MCP arthroplasty in patients with RA have been well described, information on the outcomes of SRA of the MCP joint in patients with RA is limited. ¹ The goal of this study was to assess the outcomes of SRA for MCP joints affected by RA at our institution. Specifically, we sought to investigate revision operation rates, complications, clinical outcomes, and risk factors for revision.

Materials and Methods

Data Collection

Following institutional review board approval, electronic medical records (EMRs) and an institutional joint registry were retrospectively reviewed. Data were collected from charts of patients with RA who underwent SRA of the MCP joint over a 12-year period at a single institution. Ninety-three SRAs to treat RA were performed during this time period. A minimum 2-year follow-up inclusion criterion was used, excluding 13 arthroplasties. Variables collected included patient demographics, operation and implant details, complications, and reoperations. Revision operations were defined as any occurrence of reoperation requiring replacement or alteration of the implant. Secondary operations were defined as operations occurring after the original arthroplasty that did not directly involve the implant. Clinical outcomes regarding function and pain were assessed using an institutional joint registry and EMRs. Patients reporting pain as none, mild, moderate, or severe on a standardized joint registry form or during follow-up recorded in the EMRs was recorded as such. Subjective improvement was determined by the patients stating they felt better than they were preoperatively by reporting this to their surgical team via completion of a standardized joint registry form including the options “worse,” “the same,” “somewhat better,” or “much better” or by reporting this to their team during follow-up appointments as recorded in EMRs.

Study Population and Arthroplasty Procedure

The study group included 80 MCP SRAs in 27 patients with RA. Operations were carried out between December 1993 and December 2005 at a single institution. Surface replacement arthroplasties were performed by 3 surgeons who were all beyond 5 years from fellowship completion and were classified as specialists with at least level III experience. ¹⁴ Table 1 summarizes the cohort demographics. The surgical indication for patients undergoing MCP arthroplasty with the SRA component was pain and/or deformity with functional impairment in the setting of RA (Figure 2). Cement was used in 71 (89%) cases, whereas bone graft was used in none. At the time of operation, primary medical management for RA included nonsteroidal anti-inflammatory drugs/hydroxychloroquine (n = 5, 6%), methotrexate (n = 28, 35%), or other disease-modifying antirheumatic drugs (n = 29, 36%). Eighteen (23%) joints had unknown or no documented medical therapy at the time of primary operation. Twenty-seven (34%) joints were additionally managed with prednisone. The mean follow-up time was 9.5 years (range, 2.1-20.5 years).

Table 1.

Cohort Demographics.

Demographics	Results
Joints replaced	80
Patients	27
Mean age, y (range)	52.8 (27-78)
Female, No. (%)	74 (92.5)
Mean BMI, kg/m2 (range)	24.0 (14.1-33.1)
Hand, No. (%)
Right	45 (56.3)
Left	35 (43.8)
Digit, No. (%)
Second finger	22 (27.5)
Third finger	25 (31.3)
Fourth finger	19 (23.8)
Fifth finger	14 (17.5)

Note. BMI = body mass index.

Figure 2.

Preoperative radiograph of a 66-year-old woman with advanced rheumatoid arthritis causing pain and deformity and postoperative radiograph showing surface replacement implants in the same patient.

The SRA implant is a 2-component implant with a proximal component consisting of cobalt-chromium articulation and titanium stem meant to promote bony ingrowth and an all-polyethylene distal component bearing surface. The SRA technique involves resection of the metacarpal and proximal phalanx articular surfaces. Following preparation of the intramedullary canal, the distal and proximal implant components are inserted, followed by soft-tissue balancing of the ligaments and centralization of the extensor tendon as indicated after closure of the capsule. Postoperatively, the hand is immobilized with the MCP joints extended, allowing for interphalangeal motion, for 2 to 6 weeks depending on the severity of disease, deformity, and adequacy of stabilization. Following the appropriate period of immobilization of the MCP joints, the patient is graduated to a removable splint and supervised therapy. Unless there are special considerations or concerns, such as complications, a less-than-typical postoperative course, or patient-specific barriers to returning for regular follow-up, patients are followed up at regular intervals in the first year, then 2 years, and then invited to return at 5 years and every 5 years thereafter.

Statistical Analysis

Survival curves were generated using the Kaplan-Meier model. Cox proportional hazard regression was used for analysis with this model. Statistical significance was determined for values of P < .05. JMP statistical software was used for data analysis.

Results

Revision Operation and Implant Survival

Thirteen arthroplasties (16%) required revision operation (Table 2). The 5-, 10-, 15- and 20-year rates of revision-free survival were 95%, 85%, 80%, and 69%, respectively (Figure 3). Indication for revision included subluxation (n = 8), ulnar drift (n = 4), and implant loosening (n = 1). Revision occurred in 5 index fingers, 3 long fingers, 3 ring fingers, and 2 little fingers. Of those having undergone revision, 8 implants were revised to a silicone arthroplasty, 4 had a new surface replacement (SR) component inserted, and 1 had the same proximal component originally implanted reinserted following metacarpal shortening to obtain proper reduction in the treatment of subluxation with instability. No complications, additional revisions, or secondary operations occurred in 10 of the implants following revision at a mean postrevision follow-up of 1.0 years (range, 0-3.7 years). The outcome of revision operation was unknown in 3 cases who were lost to follow-up after revision. A body mass index greater than 25 was associated with a statistically significant greater risk of revision operation (P < .01) (Supplemental Table). No other significant risk factors for revision operation were identified.

Table 2.

Revision Operations and Complication Data.

Measure	Results
Arthroplasties requiring revision	13 (16%)
Replacement of multiple components	8 (62%)
Replacement of proximal components	3 (23%)
Replacement of distal components	1 (8%)
Reinsertion of proximal component	1 (8%)
Indication for revision
Subluxation	8 (62%)
Ulnar drift	4 (31%)
Implant loosening	1 (8%)
Mean time to revision operation	90 mo (38-192 mo)
Joints requiring secondary operation	20 (25%)
Total secondary operations	38
Indications for secondary operation
Correct deformity	23 (61%)
Synovectomy	7 (18%)
Extensor mechanism repair	5 (13%)
Other	3 (8%)
Joints with at least 1 complication	51 (64%)
Total complications	101
Soft tissue contracture	38 (38%)
Subluxation/Functional: Instability	31 (31%)
Limited motion	9 (9%)
Dislocation	5 (5%)
Other	18 (18%)
Mean time to first complication	48 mo (0-180 mo)

Figure 3.

Survival from arthroplasty revision.

Reoperations

Overall, 36% (n = 29) of arthroplasties required a reoperation, including revision and secondary operations. The 5-, 10-, 15-, and 20-year rates of survival from reoperation were 80%, 65%, 55%, and 46%, respectively (Figure 4). Twenty joints underwent 38 secondary operations for complications during the follow-up period (Table 2). Most secondary operations were carried out to correct deformity and alignment. Other indications for secondary operation were removal of ectopic bone in 2 fingers and reduction of a dislocation in 1 arthroplasty. Eight joints required 1 secondary operation, 6 joints required 2 secondary operations, and 6 joints required 3 secondary operations.

Figure 4.

Overall survival from reoperation.

Functional and Clinical Outcomes

Pain ratings significantly improved following arthroplasty, with 87% of joints producing no or mild pain postoperatively, compared with 43% preoperatively (P < .0001) (Table 3). Pain was reported as “none” in 73% of arthroplasties postoperatively. The mean MCP arc of motion significantly improved from 29.4° to 49.3° postoperatively (P < .0001). The mean extension lag significantly improved from 56.7° to 23.3° postoperatively (P < .0001). Grip strength also significantly improved after the procedure (P = .04). Thirteen (48%) patients subjectively reported improvement in their condition since undergoing arthroplasty. Nine (33%) patients subjectively reported their condition being worse since arthroplasty. Among these 9 patients, 5 had persistent postoperative pain rated at greater than no/mild and 7 had a persistent extension lag deformity after SRA.

Table 3.

Clinical Outcomes Following Arthroplasty.

Outcome measure	Preoperative	Postoperative	P value
Joints with no or mild pain, %	43	87	<.0001
MCP arc of motion, deg	29.4	49.3	<.0001
MCP extension lag, deg	56.7	23.3	<.0001
Grip strength, kg	9.7	12.8	.04

Note. MCP = metacarpophalangeal.

Complications

Fifty-one arthroplasties (64%) experienced at least 1 complication (Table 2). Among these, 25 fingers had 1 complication, 12 had 2 complications, 7 had 3 complications, 6 had 4 complications, and 1 arthroplasty had 7 complications. Twenty patients (74%) experienced at least 1 complication. The most common complication was a soft-tissue contracture at 38%. Functional instability and instability from subluxation occurred in 31% of fingers. Other complications not tabulated in Table 2 were delayed wound healing or drainage (n = 4), superficial wound edge necrosis (n = 4), ligament laxity/instability (n = 3), tendon/ligament rupture (n = 3), heterotopic bone (n = 2), proximal implant loosening (n = 1), and synovitis (n = 1). There were no infections noted. Mean time to the first complication was 48 months (range, 0-180 months).

Discussion

The MCP joint is commonly affected in patients with RA.^15,16 Arthroplasty is a common surgical treatment for patients with significant pain and/or deformity.^4,5 Use of the silicone implant remains the criterion standard, but implant fracture and recurrent instability are common over medium- to long-term follow-up.^1-4,6,7,17 The SRA was designed and introduced as an alternative to silicone. The aim of this study was to examine our institution’s experience with the SRA MCP arthroplasty.

Our rates of revision operation were greater than those reported in the use of silicone implants for MCP arthroplasty in patients with RA.^1,18 Trail et al ¹⁸ reported a 5.7% revision rate in a cohort of silicone MCP arthroplasties in patients with RA. Similarly, Goldfarb and Stern ¹ reported a 7% revision rate. For comparison, 16.3% of arthroplasties in our study underwent revision. Boe et al ¹⁷ reported 5-, 10-, and 15-year rates of survival of 98%, 95%, and 95%, respectively, from silicone implant revision, whereas our study’s rates of survival from SRA revision were 95%, 85%, and 80%, respectively, for the same time points. These results suggest that patients with RA receiving an MCP SRA may be at greater risk of revision than those undergoing a silicone arthroplasty. While SRA fracture did not occur in this cohort, as opposed to those reporting silicone revision, modular implants depend largely on soft tissues to provide stability, and in the setting of RA, soft-tissue compromise may predispose patients with SRA to instability and deformity necessitating revision.

Our definition of SRA loosening was based on implant subsidence, osteolysis, or implant migration in the volar/dorsal or ulnar/radial directions on plain radiographs comparing subsequent images at least 6 months apart. The 1 patient experiencing implant loosening presented with pain at the affected joint with radiographic evidence of loosening. Further imaging studies with computed tomography demonstrated at least 3 mm of osteolysis about the proximal component. Anecdotally, the joint that underwent revision due to implant loosening did not have any cement used for the arthroplasty.

No joints experienced additional operations or complications following revision operation. This suggests that complications arising in the SR implant can be adequately managed surgically. However, the fact that no further surgery was performed does not necessarily mean that all the patients were satisfied. Most joints were revised to a silicone implant, whereas some continued to rely on SR components. One would expect similar outcomes to occur in those joints revised to a silicone implant as previously described.^1,18 As mentioned previously, one of the joints had the same implant reinserted during the revision operation. This was feasible because the implant was determined to not be worn or damaged, likely a result of volar joint subluxation.

Most patients in our study achieved significant pain relief following surgery. Goldfarb and Stern ¹ found discouraging results in silicone MCP arthroplasties, with only 27% achieving a pain-free outcome. These results may suggest patients with RA undergoing MCP SRA are more likely to have a pain-free outcome than those receiving a silicone implant. While arc of motion typically improved after surgery, an extension lag remained in most joints. Despite this, the extension lag tended to be less prominent after SRA, further signifying improvements in range of motion. Among the 9 patients reporting worsening of their condition following SRA, most continued to have both pain and an extension lag deformity. We find that dissatisfied patients likely report worsening of their condition due to an interplay of both pain and functional deficits, with functional deficit being the primary driver in our experience.

In a long-term evaluation of MCP arthroplasty, Goldfarb and Stern’s ¹ study of silicone arthroplasties demonstrated patients subjectively reported functional satisfaction in only 38% of cases. We also found that a large proportion of patients reported worsening function over time, with only 48% of patients feeling that their condition had improved at a mean 9.5-year follow-up. Our results, in conjunction with Goldfarb and Stern’s, indicate many patients with RA will experience dissatisfaction following MCP arthroplasty, regardless of the type of implant used over time.

It is important to note that MCP arthroplasty for RA presents unique challenges when compared with other joints of the hand. ³ Most arthroplasties experienced at least 1 complication. Many complications were observed years after initial operation. Most complications did not involve failure of the implant, but were complications involving the soft tissues. This may be explained in part by the inherent soft-tissue deficiencies associated with the inflammatory condition. ³ In addition, many patients required secondary operations for the management of complications at the affected site. Our results elaborate on those challenges previously described when surgically managing RA of the MCP joint.^1,18 It also highlights the deficiencies of currently available implants and the need for future implant innovation to help improve the stability of the deficient soft-tissue constraints.

Our study has limitations. The retrospective data collection and use of an institutional joint registry limit the data available for analysis. Operations were performed by multiple surgeons at our institution. In addition, we did not evaluate functional scores that may have been influenced by arthroplasty. Furthermore, we did not have patient-reported outcome measures available for these patients preoperatively or postoperatively. Pain levels and patient-reported subjective improvement may be prone to recall and social desirability biases; these, along with ROM values, are limited by the retrospective collection of data from the joint registry and EMR. Given this was a retrospective review, we did not perform a power analysis. Finally, given the limited number of end points, we were not able to perform a multivariate analysis.

Our experience with the SRA implant in the management of patients with RA demonstrated good pain relief and overall improvement of motion. However, complication rates were significant, with most being soft-tissue complications, including contracture and recurrent deformity. In addition, reoperation rates were noteworthy, and the overall revision rate at 9.5 years of average follow-up was 16%. In conjunction with other studies, the clinical results of SRA implants deteriorated with time. This study provides important considerations regarding MCP arthroplasty with SRA components while highlighting the inherent deficiencies at stabilizing an inherently soft-tissue-deficient joint in patients with RA.