- Email: [email protected]
Results of Revision Metacarpophalangeal Joint Surgery in Rheumatoid Patients Following Previous Silicone Arthroplasty
Results of Revision Metacarpophalangeal Joint Surgery in Rheumatoid Patients Following Previous Silicone Arthroplasty Scott D. Burgess, MD, Michiyuki Kono, MD, Peter J. Stern, MD From the University of Cincinnati, Cincinnati, OH; and Kaiser Permanente, Anaheim, CA.
Purpose: Primary silicone metacarpophalangeal (MCP) joint arthroplasties have good results that deteriorate with time. The purpose of this study was to assess indications, patient satisfaction, and clinical and radiographic results following revision surgery in rheumatoid patients who had previously undergone silicone MCP arthroplasty. Methods: Twenty hands in 18 patients (62 implants) had revision silicone MCP arthroplasties between 1986 and 2005 and had a greater than 1-year follow-up period (mean 5 y). A retrospective chart review was performed to collect preoperative and intraoperative data. Patients were then re-examined and administered a questionnaire addressing subjective outcome and satisfaction. Results: Intraoperatively, 76% of the implants were fractured. Thirteen of 17 synovial biopsies revealed giant cell foreign body reaction. Preoperatively, the average arc of motion was from 16° to 50°, and ulnar drift was 24°. Postoperatively, the average arc of motion was from 20° to 54°, and ulnar drift was 13°. X-rays of 14 hands revealed that 15 of 44 revised implants had fractured. Sixteen patients (18 hands) were available to complete questionnaires. Twelve patients (14 hands) were satisfied and 3 were dissatisfied. Five of 16 patients would not have the revision again. These patients had worse average postoperative ulnar drift (30° vs 9°) than the other 11 patients. All patients except one who had preoperative pain had at least moderate pain relief, and of the 6 patients who had revision surgery because of pain, 5 were satisfied. Conclusions: Revision silicone arthroplasty provides excellent pain relief, and the majority of patients were pleased with their results. Objective results, however, were generally poor. Soft tissue reconstruction is more difficult to achieve than the primary procedure, as evidenced by minimal improvement in ulnar drift, a high rate of implant fracture, and no change in arc of motion. (J Hand Surg 2007;32A:1506 –1512. Copyright © 2007 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Therapeutic IV. Key words: Metacarpophalangeal joint, revision arthroplasty, rheumatoid arthritis, silicone arthroplasty.
ince the development of the flexible silicone implant in the 1960s, silicone arthroplasty has become a mainstay in the treatment of rheumatoid arthritis of the metacarpophalangeal (MCP) joint. The implant was designed as a spacer to provide stability and maintain alignment while allowing motion during the period of soft tissue healing, called encapsulation. Numerous studies have confirmed the benefits of primary silicone MCP arthroplasty, including pain relief and improvements in dexterity, range of motion, strength, and appearance.1–14 Com-
S 1506
The Journal of Hand Surgery
plications, including implant fracture and resection for deep infection, have been few. Although some long-term studies have reported a recurrence of ulnar drift and loss of motion with time,3,7,9,15 patients and physicians have remained satisfied with the results, and reports of revision arthroplasty have been infrequent.10,11,13 Because stability of the joint requires integrity of the soft tissue envelope, we were concerned that the results of revisions would not be as good as those reported following primary arthroplasty. The purpose
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
of this study was to assess indications, patient satisfaction, and clinical and radiographic results following revision surgery in rheumatoid patients who had previously had silicone MCP arthroplasty.
Materials and Methods After institutional review board approval was obtained, a retrospective review was performed by identifying all patients who had implant MCP arthroplasty from 1980 to 2005 via a database search using current procedural terminology (CPT) codes. From this initial group, patients were selected for inclusion in this study who had rheumatoid arthritis and had had revision arthroplasty of previously placed silicone MCP joints. Initial chart review yielded 33 hands in 28 patients (109 arthroplasties) who had had revision silicone arthroplasty during the study period. Attempts were made to contact all patients. Seven patients had died without documented 1-year follow-up evaluation. Three patients could not be located and did not have adequate chart documentation for inclusion. Two patients who could not be contacted but had sufficient documentation in the chart, with the minimum 1-year follow-up period, were included in this study. Twenty hands in 18 patients, involving 62 joints, qualified for inclusion with minimum 1-year follow-up data. All patients were women. Two had bilateral revisions; the other 16 had revisions of a single hand. There were 16 right hands and 4 left hands. The average age at time of primary MCP arthroplasty was 54 years (range 26 –76) and at the time of revision was 61 years (range 26 –78). Average time from primary arthroplasty to revision surgery was 6 years (range 6 mo–12 y). Five implants in 3 hands were revised within 1 year of the primary procedure, 3 for dislocation and 2 for implant rotation. The remaining hands underwent revision at an average of 7 years after the primary procedure (range 2–18 years). All women were diagnosed with rheumatoid arthritis; one had the juvenile form. Sixteen patients (18 hands) were contacted and completed our questionnaire. Ten patients (9 hands) were evaluated in 2006; the remaining 6 patients (7 hands) were evaluated in 1998, in the initial phase of the study, with all x-rays and data sheets available for review. Twelve patients (14 hands) were able to return for examination and x-rays. The other 4 were too debilitated to travel but consented to examinations at home. Of the two deceased patients, 1 had 23 months of close follow-up evaluation, during which time she underwent two additional revision surgeries.
1507
The other patient died after 2 years of documented follow-up evaluaton. All patients had a minimum of a 1-year follow-up period with an average of 5 years (range 1–26 y). The authors performed all examinations, read all radiographs and administered all questionnaires for this study. We recognize that the senior author is not an objective observer and bias may be introduced. Outcomes of revision were measured objectively by patient examination and radiographs at most recent follow-up visit. The patient examination included measurement of MCP joint range of motion and ulnar deviation using a goniometer and was performed by one of the authors. Ulnar drift was measured as the angle between the longitudinal axes of the metacarpal and the proximal phalanx. Postoperative range of motion and ulnar drift were obtained for 57 of 62 joints, excluding 5 that underwent MCP joint arthrodesis during the study period. Preoperative x-rays were evaluated for the presence of grades B and C endosteal erosions according to criteria established by Creighton et al (Fig. 1).16 These erosions were also evaluated on the latest follow-up films for evidence of progression. Postrevision follow-up x-rays were evaluated for evidence of implant fracture according to criteria defined by Bass et al.17 Subjective outcomes were measured using a questionnaire created to evaluate the patient’s indication for revision, subjective outcome, and overall satisfaction of both primary and revision surgery. Satisfaction was measured in 2 ways. First, patients were asked to stratify their level of satisfaction on a 5-point scale (greatly satisfied, mildly satisfied, neither satisfied nor dissatisfied, mildly dissatisfied, or greatly dissatisfied). Second, patients were asked if they would have the surgery again. Statistical differences were evaluated using the Student’s t-test. The decision for surgery was made collaboratively between the surgeon and the patient. The surgeon’s indication for recommending revision (Table 1), as documented in the chart, was most commonly recurrent ulnar drift (n⫽11). Other indications were implant dislocation/rotation (n⫽3), pain (n⫽5), and as part of a flexor digitorum profundus tenodesis for tendon rupture (n⫽1). The patients’ primary reasons for seeking revision surgery, according to the administered questionnaire (Table 1), were most commonly pain and functional difficulties, particularly in handling small objects. The predominant procedure was revision of all 4 joints, performed in 12 hands. There were 5 single joint revisions, one double, and one triple. Two patients (5 fingers) had MCP arthrodesis
1508
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
Figure 1. X-ray showing endosteal erosions of the proximal phalanx (grade B in middle finger, grade C in index finger).
during the study period. Our indications for MCP arthrodesis included inadequate soft tissue surrounding the failed implant, inadequate bone to accept a new implant, and severe recurrent deformity. All surgeries were performed through a transverse incision, and joints were approached from the ulnar side. Where identifiable, the radial collateral ligament was preserved and reattached to the metacarpal shaft through a drill hole. The medullary canal usually contained a thin jacket of reactive tissue and silicone debris, which had surrounded the stem; this was removed. Further preparation included gentle burring or broaching of the medullary canal. All sharp bony surfaces are squared off to prevent fissuring (and possible fracture) of the implant. The largest possible implant was inserted. The extensor digitorum tendon was sewn through a drill hole to the dorsal base of the proximal phalanx and the radial sagittal bands were reefed to keep the extensor ten-
don centrally located. The skin incision was closed over a drain. The postoperative therapy protocol was the same as that used for primary arthroplasty. Patients were placed in a bulky compressive dressing, supported with a volar forearm-based splint to the fingertips, keeping the MCP joints extended. The dressing was removed on postoperative day 3–5, and a dynamic outrigger splint was fabricated by a certified hand therapist for use during the day. A static splint was worn at night. Active range of motion was initiated in the dynamic splint, which was worn for 6 weeks. Static night splinting was continued for 3 months and, in some cases, longer if there was any evidence of recurrent ulnar drift. Intraoperative abnormalities included 52 fractured implants, 3 dislocated stems, and 2 rotated implants. Eleven replaced implants had no fracture, but many were noted to be deformed or cracked. Synovial specimens (the jacket surrounding the implant) were routinely sent for pathologic analysis by the surgeon who performed the majority of revisions in this series. The other surgeon in this series did not send specimens for analysis. Specimens were not sent in cases of revision for known cause of failure, such as dislocation or rotation of the implant or tendon, rupture or if the revision was to an arthrodesis. Seventeen specimens were sent for pathologic analysis, with 13 revealing a giant cell foreign body reaction containing refractile material consistent with silicone. The other four specimens demonstrated chronic inflammation and fibrosis but no giant cell reaction or foreign body material.
Results Preoperatively there was 16° of extension (range 20° hyperextension to 80°) and flexion to 50° (range 0° to 80°) for an average arc of 34°. Postoperative range of
Table 1. Primary Reasons for Revision Surgery Per Hand Recurrent deformity Pain Implant dislocation/rotation Physician recommendation FDP rupture Dexterity loss Reason not given
Surgeon
Patient
11 5 3 0 1 0 0
0 6 0 5 0 4 5
FDP, flexor digitorum profundus. According to surgeon and patient, responses obtained from surgeon’s notes and patient’s questionnaire.
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
1509
Table 2. Preoperative and Postoperative Range of Motion and Deformity Results Preoperative Extension Flexion Arc of motion Ulnar drift
14° (-20° to 80°)* 56° (0° to 80°) 34° 24° (-20° to 60°)†
Postoperative 20° (-5° to 60°)* 55° (0° to 95°) 34° 15° (-10° to 90°)†
p Value .114 .136 .003
*Minus (-) indicates hyperextension. †Minus (-) indicates radial deviation.
motion averaged 20° of extension (range 5° hyperextension to 60°) and 54° of flexion (range 0° to 95°), for an average arc of 34°. Changes in range of motion were not statistically significant (p⬎.1). Preoperative ulnar drift was 24° (range 20° radial drift to 60° ulnar drift), which improved to 13° (range 10° of radial drift to 90°) at the most recent follow-up evaluation (p⫽.003) (Table 2). Grip and pinch strengths were not routinely recorded preoperatively. Postoperatively, grip averaged 7 kg (range 1 to 16 kg) and pinch averaged 3 kg (range 1 to 6 kg). For the 14 hands with follow-up x-rays, 15 of 44 revised implants had fractured, based on criteria defined by Bass et al.17 Radiographic evaluation also demonstrated that 38 of 53 implants with available preoperative x-rays had endosteal bony erosions present in either the metacarpal or the proximal phalanx. All except 3 implants in one hand had no further progressions of these bony changes on most recent follow-up films. Two hands (5 joints) had MCP joint arthrodesis during the study period. One patient had conversion of a single, failed, primary arthroplasty to an arthrodesis during her revision procedure because the medullary canal of the proximal phalanx was insufficient to accept a silicone implant (Figs. 2, 3). The second patient had conversion of all 4 MCP joints to arthrodeses in 2 separate procedures following revision. During the first procedure, one digit was fused and a second treated with implant removal for rotated implants. In the second procedure, the remaining 3 MCP joints were fused due to recurrent deformity and inadequate soft tissue including absent extensor tendon over the MCP joint. All 5 arthrodeses successfully fused. Of the 13 pathologic specimens that revealed giant cell foreign body reaction containing refractile material consistent with silicone particles, 8 hands had endosteal erosions on pre-revision x-rays and 5 did not show erosions. Of the other 4 specimens that demonstrated chronic inflammation and fibrosis but no giant cell reaction or foreign body material, 2
hands had endosteal erosions on pre-revision x-rays and 2 did not. The pathologic findings of silicone debris did not correlate with x-ray findings of endosteal erosions (p⫽.35). There were no infections. Twelve implants in 3 hands were revised a second time. One patient, already discussed, was converted to MCP arthrodeses of all 4 joints in 2 separate procedures. Two addi-
Figure 2. X-ray showing fractured implants with narrowed medullary canal of the middle finger proximal phalanx.
1510
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
(31%) stated they would not undergo the revision if they had it to do over again. These 5 patients included 1 patient who categorized herself as mildly satisfied, 1 neither satisfied nor dissatisfied, 1 who was mildly dissatisfied, and 2 who were greatly dissatisfied. In order to contrast groups, the objective data were compared between those who would have the surgery done again (pleased) against those who would not (displeased). The pleased patients improved to a postoperative range of motion of 17° extension to 54° flexion and ulnar drift of 9°, while the displeased patients improved to a range of motion of 30° extension to 54° flexion and ulnar drift of 30° (Table 4). This difference was statistically different for residual extensor lag (p⫽.007) and for ulnar drift (p⬍.001), but not for flexion (p⫽.49). The pleased patients, however, had slightly less initial deformity than the displeased group. The pleased patients started from a range of 13° extension to 49° flexion and drift of 24°, while the displeased patients started from a range of 25° extension to 57° flexion and drift of 27°. Figure 3. X-ray of revised MCP silicone arthroplasties and arthrodesis of the middle finger MCP joint.
tional patients had second revision procedures of all 4 joints for recurrent deformity and implant failure at 10 and 12 years. There were 32 additional procedures performed during the study period on those hands being studied, for an average of 2 per hand (range 0 –5). These included interphalangeal arthrodesis, boutonniere reconstruction, nodule excision, carpal tunnel release, and wrist procedures. Questionnaire Patients were asked to indicate their primary reason for undergoing revision (Table 1). The most common reason was pain (n⫽6). The second most common reason was that the physician recommended revision (n⫽5); of these patients, of those who listed a secondary reason 2 were because of pain, 1 because of weakness and 1 because of tendon rupture. The remaining 4 patients stated they sought revision for difficulty in handling small objects. Satisfaction was measured in 2 ways (Table 3). First, patients were asked to stratify their level of satisfaction on a 5-point scale. Twelve of the 16 patients were either greatly or mildly satisfied. Two patients were greatly dissatisfied, with 1 patient each in the other categories. Second, patients were asked if they would have the surgery again. Five of the 16
Discussion Swanson’s 1972 report emphasized that the silicone implant acted as a spacer following resection arthroplasty, providing stability and allowing early motion while the soft-tissue envelope healed.1 Ultimate stability is provided by the developing capsule, which in turn protects the implant from fracture. He termed this process “encapsulation.” Implant fracture can also be minimized by allowing the stems to piston within the medullary canals, thus dissipating the forces of flexion and allowing the implant to find the best position with respect to the axis of motion.1 Because stability is eventually provided by the soft tissues, fractured implants do not imply loss of function. In fact, Swanson noted that most patients with radiographic evidence of implant fractures were doing well and did not notice any functional change.1 This concept is supported in the literature, as very few fractured implants have been revised. Long-term studies with a greater than 5-year follow-up period
Table 3. Patient Satisfaction Number of Patients Greatly satisfied Mildly satisfied Neither satisfied nor dissatisfied Mildly dissatisfied Greatly dissatisfied
8 4 1 1 2
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
1511
Table 4. Preoperative and Postoperative Range of Motion for Pleased and Displeased Patients Pleased
Extension Flexion Ulnar drift
Displeased
Preoperative
Postoperative
Preoperative
Postoperative
Postoperative p Value
13° 49° 24°
17° 54° 9°
25° 57° 27°
30° 54° 30°
.007 .49 ⬍.001
have reported fracture rates of 0% to 63% with few revisions.2,3,7,10,11,15 Trail et al have reported the largest number of revisions, revising 76 of 1336 joints, 39 with fractured stems. With a revision rate of 3%, they concluded that radiographic implant failure does not require revision surgery.13 Gellman also revised 54 implants, all for recurrent deformity due to implant fracture.11 He notes, however, that there were 126 fractured implants in his series, and the remainder had no recurrence of deformity or loss of function. It appears that the soft tissues are more critical to long-term stability and function than the implant. However, the soft tissue support seems to weaken with time and progressive disease. Goldfarb and Stern reported at an average follow-up period of 14 years, an implant fracture rate of 63%.15 Clinical outcomes worsened. Only 38% of patients were satisfied with hand function and only 27% were pain free. In cases where revision is necessary, there is presumably greater insult to the soft tissues, not only from the second surgical procedure but from the stretching that has occurred with recurrence of deformity and the underlying destruction caused by the rheumatoid arthritis. As such, soft tissue reconstruction may be more difficult to achieve in the revision arthroplasty. Our revision series showed generally poor results from an objective standpoint. As a group, there was no significant change in the flexion range (preoperative 16° to 50°, postoperative 20° to 54°) and a small improvement in ulnar drift (preoperative 24°, postoperative 13°) at just over a 5-year follow-up period. In addition, there was a high implant fracture rate (34%) in the revisions, suggesting that the soft tissues were unable to support the forces at the joint, leading to excessive demand on the implant. We believe that these poor results are most likely due to incompetent soft tissues, secondary to multiple surgeries, active rheumatoid disease, foreign body reaction caused by particulate debris, or a combination of these. Although we found no correlation between silicone debris and endosteal erosion on x-ray, the effect of silicone particulate matter on soft tissues
and implant longevity is not currently known, and we did not alter our treatment based on these findings. Despite these poor objective results, the majority of patients were pleased with the results and would have the surgery done again. These pleased patients had similar postoperative MCP flexion but less extensor lag and ulnar drift than the displeased patients. The pleased patients had slightly less preoperative deformity to begin with, mainly affecting extension. The poor average correction for the entire group could suggest that greater preoperative deformity is associated with weaker soft tissues and subsequently worse correction. Patient displeasure also seems to be related to the ability to correct this deformity. Presumably, if some soft tissue stability can be maintained, good results can be obtained from both an objective and subjective standpoint. As such, careful patient selection is critical. An annual follow-up evaluation may be worthwhile. If there is evidence of progressive deformity or the patient complains of pain or loss of dexterity, a discussion can be held regarding revision MCP arthroplasty. It is very important that the patient understands that goals are limited and that malalignment, pain, and decreased dexterity may recur over time. There was a disparity between the physician’s documented reason for revision and the patient’s reason for undergoing revision. Physicians most commonly cited deformity for the reason to operate, whereas patients most commonly cited pain. This difference may be due to documentation styles, patient recall bias, or a true disconnect between the goals of the surgeon and the patient. Chung et al studied patient reasons for seeking MCP implant arthroplasty using the Michigan Hand Outcomes Questionnaire in 33 surgical and 28 nonsurgical patients.18 They found that worse hand function and increased pain correlated with seeking surgical treatment. Hand appearance was not significantly correlated with seeking arthroplasty. They did not, however, evaluate differences between patient and physician reasons for surgery. Future studies investigating this potential difference in motivation could shed some light on this issue.
1512
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
Finally, it should be noted that despite the overall inability to improve motion and ulnar drift, revision silicone arthroplasty was very successful in alleviating pain. Of the 6 patients who underwent revision surgery because of pain, 5 were satisfied and would have the procedure performed again.
8.
9. Received for publication September 27, 2006; accepted in revised form July 30, 2007. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Corresponding author: Scott D. Burgess, MD, Michigan Hand Center, 1111 Leffingwell NE, Ste 200, Grand Rapids, MI 49525; e-mail: [email protected] Copyright © 2007 by the American Society for Surgery of the Hand 0363-5023/07/32A10-0001$32.00/0 doi:10.1016/j.jhsa.2007.07.026
References 1. Swanson AB. Flexible implant arthroplasty for arthritic finger joints: rationale, technique, and results of treatment. J Bone Joint Surg 1972;54A:435– 455. 2. Wilson YG, Sykes PJ, Niranjan NS. Long-term follow-up of Swanson’s silastic arthroplasty of the metacarpophalangeal joints in rheumatoid arthritis. J Hand Surg 1993;18B:81–91. 3. Bieber EJ, Weiland AJ, Volenec-Dowling S. Silicone-rubber implant arthroplasty of the metacarpophalangeal joints for rheumatoid arthritis. J Bone Joint Surg 1986;68A:206 –209. 4. Beckenbaugh RD, Dobyns JH, Linsheid RL, Bryan RS. Review and analysis of silicone-rubber metacarpophalangeal implants. J Bone Joint Surg 1976;58A:483– 487. 5. Blair WF, Shurr DG, Buckwalter JA. Metacarpophalangeal joint implant arthroplasty with a silastic spacer. J Bone Joint Surg 1984;66A:365–370. 6. Mannerfelt L, Andersson K. Silastic arthroplasty of the metacarpophalangeal joints in rheumatoid arthritis. J Bone Joint Surg 1975;57A:484 – 489. 7. Kirschenbaum D, Schneider LH, Adams DC, Cody RP. Arthroplasty of the metacarpophalangeal joints with use of
10.
11.
12.
13.
14.
15.
16.
17.
18.
silicone-rubber implants in patients who have rheumatoid arthritis. Long-term results. J Bone Joint Surg 1993;75A: 3–12. Vahvanen V, Viljakka T. Silicone rubber implant arthroplasty of the metacarpophalangeal joint in rheumatoid arthritis. A follow-up study of 32 patients. J Hand Surg 1986; 11A:333–339. Olsen I, Gebuhr P, Sonne-Holm S. Silastic arthroplasty in rheumatoid MCP-joints. 60 joints followed for 7 years. Acta Orthop Scand 1994;65:430 – 431. Hansraj KK, Ashworth CR, Ebramzadeh E, Todd AO, Griffin MD, Ashley EM, Cardilli AM. Swanson metacarpophalangeal joint arthroplasty in patients with rheumatoid arthritis. Clin Orthop 1997;342:11–15. Gellman H, Stetson W, Brumfield RHJ, Costigan W, Kuscher SH. Silastic metacarpophalangeal joint arthroplasty in patients with rheumatoid arthritis. Clin Orthop 1997;342: 16 –21. Rothwell AG, Cragg KJ, O’Neill LB. Hand function following silastic arthroplasty of the metacarpophalangeal joints in the rheumatoid hand. J Hand Surg 1997;22B:90 –93. Trail IA, Martin JA, Nuttall D, Stanley JK. Seventeen-year survivorship of Silastic metacarpophalangeal joint replacement. J Bone Joint Surg 2004;86B:1002–1006. Chung KC, Kostis SV, Kim HM. A prospective outcomes study of Swanson metacarpophalangeal joint arthroplasty for the rheumatoid hand. J Hand Surg 2004;29A:646 – 653. Goldfarb CA, Stern PJ. Metacarpophalangeal joint arthroplasty in rheumatoid arthritis. A long-term assessment. J Bone Joint Surg 2003;85A:1869 –1878. Creighton JJ, Steichen JB, Strickland JW. Long-term evaluation of silastic trapezial arthroplasty in patients with osteoarthritis. J Hand Surg 1991;16A:510 –519. Bass RL, Stern PJ, Nairus JG. High implant fracture incidence with Sutter silicone metacarpophalangeal joint arthroplasty. J Hand Surg 1996;21A:813– 818. Chung KC, Kostis SV, Kim HM, Burke FD, Wilgis EFS. Reasons why rheumatoid arthritis patients seek surgical treatment for hand deformities. J Hand Surg 2006;31A:289 – 294.
Purpose: Primary silicone metacarpophalangeal (MCP) joint arthroplasties have good results that deteriorate with time. The purpose of this study was to assess indications, patient satisfaction, and clinical and radiographic results following revision surgery in rheumatoid patients who had previously undergone silicone MCP arthroplasty. Methods: Twenty hands in 18 patients (62 implants) had revision silicone MCP arthroplasties between 1986 and 2005 and had a greater than 1-year follow-up period (mean 5 y). A retrospective chart review was performed to collect preoperative and intraoperative data. Patients were then re-examined and administered a questionnaire addressing subjective outcome and satisfaction. Results: Intraoperatively, 76% of the implants were fractured. Thirteen of 17 synovial biopsies revealed giant cell foreign body reaction. Preoperatively, the average arc of motion was from 16° to 50°, and ulnar drift was 24°. Postoperatively, the average arc of motion was from 20° to 54°, and ulnar drift was 13°. X-rays of 14 hands revealed that 15 of 44 revised implants had fractured. Sixteen patients (18 hands) were available to complete questionnaires. Twelve patients (14 hands) were satisfied and 3 were dissatisfied. Five of 16 patients would not have the revision again. These patients had worse average postoperative ulnar drift (30° vs 9°) than the other 11 patients. All patients except one who had preoperative pain had at least moderate pain relief, and of the 6 patients who had revision surgery because of pain, 5 were satisfied. Conclusions: Revision silicone arthroplasty provides excellent pain relief, and the majority of patients were pleased with their results. Objective results, however, were generally poor. Soft tissue reconstruction is more difficult to achieve than the primary procedure, as evidenced by minimal improvement in ulnar drift, a high rate of implant fracture, and no change in arc of motion. (J Hand Surg 2007;32A:1506 –1512. Copyright © 2007 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Therapeutic IV. Key words: Metacarpophalangeal joint, revision arthroplasty, rheumatoid arthritis, silicone arthroplasty.
ince the development of the flexible silicone implant in the 1960s, silicone arthroplasty has become a mainstay in the treatment of rheumatoid arthritis of the metacarpophalangeal (MCP) joint. The implant was designed as a spacer to provide stability and maintain alignment while allowing motion during the period of soft tissue healing, called encapsulation. Numerous studies have confirmed the benefits of primary silicone MCP arthroplasty, including pain relief and improvements in dexterity, range of motion, strength, and appearance.1–14 Com-
S 1506
The Journal of Hand Surgery
plications, including implant fracture and resection for deep infection, have been few. Although some long-term studies have reported a recurrence of ulnar drift and loss of motion with time,3,7,9,15 patients and physicians have remained satisfied with the results, and reports of revision arthroplasty have been infrequent.10,11,13 Because stability of the joint requires integrity of the soft tissue envelope, we were concerned that the results of revisions would not be as good as those reported following primary arthroplasty. The purpose
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
of this study was to assess indications, patient satisfaction, and clinical and radiographic results following revision surgery in rheumatoid patients who had previously had silicone MCP arthroplasty.
Materials and Methods After institutional review board approval was obtained, a retrospective review was performed by identifying all patients who had implant MCP arthroplasty from 1980 to 2005 via a database search using current procedural terminology (CPT) codes. From this initial group, patients were selected for inclusion in this study who had rheumatoid arthritis and had had revision arthroplasty of previously placed silicone MCP joints. Initial chart review yielded 33 hands in 28 patients (109 arthroplasties) who had had revision silicone arthroplasty during the study period. Attempts were made to contact all patients. Seven patients had died without documented 1-year follow-up evaluation. Three patients could not be located and did not have adequate chart documentation for inclusion. Two patients who could not be contacted but had sufficient documentation in the chart, with the minimum 1-year follow-up period, were included in this study. Twenty hands in 18 patients, involving 62 joints, qualified for inclusion with minimum 1-year follow-up data. All patients were women. Two had bilateral revisions; the other 16 had revisions of a single hand. There were 16 right hands and 4 left hands. The average age at time of primary MCP arthroplasty was 54 years (range 26 –76) and at the time of revision was 61 years (range 26 –78). Average time from primary arthroplasty to revision surgery was 6 years (range 6 mo–12 y). Five implants in 3 hands were revised within 1 year of the primary procedure, 3 for dislocation and 2 for implant rotation. The remaining hands underwent revision at an average of 7 years after the primary procedure (range 2–18 years). All women were diagnosed with rheumatoid arthritis; one had the juvenile form. Sixteen patients (18 hands) were contacted and completed our questionnaire. Ten patients (9 hands) were evaluated in 2006; the remaining 6 patients (7 hands) were evaluated in 1998, in the initial phase of the study, with all x-rays and data sheets available for review. Twelve patients (14 hands) were able to return for examination and x-rays. The other 4 were too debilitated to travel but consented to examinations at home. Of the two deceased patients, 1 had 23 months of close follow-up evaluation, during which time she underwent two additional revision surgeries.
1507
The other patient died after 2 years of documented follow-up evaluaton. All patients had a minimum of a 1-year follow-up period with an average of 5 years (range 1–26 y). The authors performed all examinations, read all radiographs and administered all questionnaires for this study. We recognize that the senior author is not an objective observer and bias may be introduced. Outcomes of revision were measured objectively by patient examination and radiographs at most recent follow-up visit. The patient examination included measurement of MCP joint range of motion and ulnar deviation using a goniometer and was performed by one of the authors. Ulnar drift was measured as the angle between the longitudinal axes of the metacarpal and the proximal phalanx. Postoperative range of motion and ulnar drift were obtained for 57 of 62 joints, excluding 5 that underwent MCP joint arthrodesis during the study period. Preoperative x-rays were evaluated for the presence of grades B and C endosteal erosions according to criteria established by Creighton et al (Fig. 1).16 These erosions were also evaluated on the latest follow-up films for evidence of progression. Postrevision follow-up x-rays were evaluated for evidence of implant fracture according to criteria defined by Bass et al.17 Subjective outcomes were measured using a questionnaire created to evaluate the patient’s indication for revision, subjective outcome, and overall satisfaction of both primary and revision surgery. Satisfaction was measured in 2 ways. First, patients were asked to stratify their level of satisfaction on a 5-point scale (greatly satisfied, mildly satisfied, neither satisfied nor dissatisfied, mildly dissatisfied, or greatly dissatisfied). Second, patients were asked if they would have the surgery again. Statistical differences were evaluated using the Student’s t-test. The decision for surgery was made collaboratively between the surgeon and the patient. The surgeon’s indication for recommending revision (Table 1), as documented in the chart, was most commonly recurrent ulnar drift (n⫽11). Other indications were implant dislocation/rotation (n⫽3), pain (n⫽5), and as part of a flexor digitorum profundus tenodesis for tendon rupture (n⫽1). The patients’ primary reasons for seeking revision surgery, according to the administered questionnaire (Table 1), were most commonly pain and functional difficulties, particularly in handling small objects. The predominant procedure was revision of all 4 joints, performed in 12 hands. There were 5 single joint revisions, one double, and one triple. Two patients (5 fingers) had MCP arthrodesis
1508
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
Figure 1. X-ray showing endosteal erosions of the proximal phalanx (grade B in middle finger, grade C in index finger).
during the study period. Our indications for MCP arthrodesis included inadequate soft tissue surrounding the failed implant, inadequate bone to accept a new implant, and severe recurrent deformity. All surgeries were performed through a transverse incision, and joints were approached from the ulnar side. Where identifiable, the radial collateral ligament was preserved and reattached to the metacarpal shaft through a drill hole. The medullary canal usually contained a thin jacket of reactive tissue and silicone debris, which had surrounded the stem; this was removed. Further preparation included gentle burring or broaching of the medullary canal. All sharp bony surfaces are squared off to prevent fissuring (and possible fracture) of the implant. The largest possible implant was inserted. The extensor digitorum tendon was sewn through a drill hole to the dorsal base of the proximal phalanx and the radial sagittal bands were reefed to keep the extensor ten-
don centrally located. The skin incision was closed over a drain. The postoperative therapy protocol was the same as that used for primary arthroplasty. Patients were placed in a bulky compressive dressing, supported with a volar forearm-based splint to the fingertips, keeping the MCP joints extended. The dressing was removed on postoperative day 3–5, and a dynamic outrigger splint was fabricated by a certified hand therapist for use during the day. A static splint was worn at night. Active range of motion was initiated in the dynamic splint, which was worn for 6 weeks. Static night splinting was continued for 3 months and, in some cases, longer if there was any evidence of recurrent ulnar drift. Intraoperative abnormalities included 52 fractured implants, 3 dislocated stems, and 2 rotated implants. Eleven replaced implants had no fracture, but many were noted to be deformed or cracked. Synovial specimens (the jacket surrounding the implant) were routinely sent for pathologic analysis by the surgeon who performed the majority of revisions in this series. The other surgeon in this series did not send specimens for analysis. Specimens were not sent in cases of revision for known cause of failure, such as dislocation or rotation of the implant or tendon, rupture or if the revision was to an arthrodesis. Seventeen specimens were sent for pathologic analysis, with 13 revealing a giant cell foreign body reaction containing refractile material consistent with silicone. The other four specimens demonstrated chronic inflammation and fibrosis but no giant cell reaction or foreign body material.
Results Preoperatively there was 16° of extension (range 20° hyperextension to 80°) and flexion to 50° (range 0° to 80°) for an average arc of 34°. Postoperative range of
Table 1. Primary Reasons for Revision Surgery Per Hand Recurrent deformity Pain Implant dislocation/rotation Physician recommendation FDP rupture Dexterity loss Reason not given
Surgeon
Patient
11 5 3 0 1 0 0
0 6 0 5 0 4 5
FDP, flexor digitorum profundus. According to surgeon and patient, responses obtained from surgeon’s notes and patient’s questionnaire.
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
1509
Table 2. Preoperative and Postoperative Range of Motion and Deformity Results Preoperative Extension Flexion Arc of motion Ulnar drift
14° (-20° to 80°)* 56° (0° to 80°) 34° 24° (-20° to 60°)†
Postoperative 20° (-5° to 60°)* 55° (0° to 95°) 34° 15° (-10° to 90°)†
p Value .114 .136 .003
*Minus (-) indicates hyperextension. †Minus (-) indicates radial deviation.
motion averaged 20° of extension (range 5° hyperextension to 60°) and 54° of flexion (range 0° to 95°), for an average arc of 34°. Changes in range of motion were not statistically significant (p⬎.1). Preoperative ulnar drift was 24° (range 20° radial drift to 60° ulnar drift), which improved to 13° (range 10° of radial drift to 90°) at the most recent follow-up evaluation (p⫽.003) (Table 2). Grip and pinch strengths were not routinely recorded preoperatively. Postoperatively, grip averaged 7 kg (range 1 to 16 kg) and pinch averaged 3 kg (range 1 to 6 kg). For the 14 hands with follow-up x-rays, 15 of 44 revised implants had fractured, based on criteria defined by Bass et al.17 Radiographic evaluation also demonstrated that 38 of 53 implants with available preoperative x-rays had endosteal bony erosions present in either the metacarpal or the proximal phalanx. All except 3 implants in one hand had no further progressions of these bony changes on most recent follow-up films. Two hands (5 joints) had MCP joint arthrodesis during the study period. One patient had conversion of a single, failed, primary arthroplasty to an arthrodesis during her revision procedure because the medullary canal of the proximal phalanx was insufficient to accept a silicone implant (Figs. 2, 3). The second patient had conversion of all 4 MCP joints to arthrodeses in 2 separate procedures following revision. During the first procedure, one digit was fused and a second treated with implant removal for rotated implants. In the second procedure, the remaining 3 MCP joints were fused due to recurrent deformity and inadequate soft tissue including absent extensor tendon over the MCP joint. All 5 arthrodeses successfully fused. Of the 13 pathologic specimens that revealed giant cell foreign body reaction containing refractile material consistent with silicone particles, 8 hands had endosteal erosions on pre-revision x-rays and 5 did not show erosions. Of the other 4 specimens that demonstrated chronic inflammation and fibrosis but no giant cell reaction or foreign body material, 2
hands had endosteal erosions on pre-revision x-rays and 2 did not. The pathologic findings of silicone debris did not correlate with x-ray findings of endosteal erosions (p⫽.35). There were no infections. Twelve implants in 3 hands were revised a second time. One patient, already discussed, was converted to MCP arthrodeses of all 4 joints in 2 separate procedures. Two addi-
Figure 2. X-ray showing fractured implants with narrowed medullary canal of the middle finger proximal phalanx.
1510
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
(31%) stated they would not undergo the revision if they had it to do over again. These 5 patients included 1 patient who categorized herself as mildly satisfied, 1 neither satisfied nor dissatisfied, 1 who was mildly dissatisfied, and 2 who were greatly dissatisfied. In order to contrast groups, the objective data were compared between those who would have the surgery done again (pleased) against those who would not (displeased). The pleased patients improved to a postoperative range of motion of 17° extension to 54° flexion and ulnar drift of 9°, while the displeased patients improved to a range of motion of 30° extension to 54° flexion and ulnar drift of 30° (Table 4). This difference was statistically different for residual extensor lag (p⫽.007) and for ulnar drift (p⬍.001), but not for flexion (p⫽.49). The pleased patients, however, had slightly less initial deformity than the displeased group. The pleased patients started from a range of 13° extension to 49° flexion and drift of 24°, while the displeased patients started from a range of 25° extension to 57° flexion and drift of 27°. Figure 3. X-ray of revised MCP silicone arthroplasties and arthrodesis of the middle finger MCP joint.
tional patients had second revision procedures of all 4 joints for recurrent deformity and implant failure at 10 and 12 years. There were 32 additional procedures performed during the study period on those hands being studied, for an average of 2 per hand (range 0 –5). These included interphalangeal arthrodesis, boutonniere reconstruction, nodule excision, carpal tunnel release, and wrist procedures. Questionnaire Patients were asked to indicate their primary reason for undergoing revision (Table 1). The most common reason was pain (n⫽6). The second most common reason was that the physician recommended revision (n⫽5); of these patients, of those who listed a secondary reason 2 were because of pain, 1 because of weakness and 1 because of tendon rupture. The remaining 4 patients stated they sought revision for difficulty in handling small objects. Satisfaction was measured in 2 ways (Table 3). First, patients were asked to stratify their level of satisfaction on a 5-point scale. Twelve of the 16 patients were either greatly or mildly satisfied. Two patients were greatly dissatisfied, with 1 patient each in the other categories. Second, patients were asked if they would have the surgery again. Five of the 16
Discussion Swanson’s 1972 report emphasized that the silicone implant acted as a spacer following resection arthroplasty, providing stability and allowing early motion while the soft-tissue envelope healed.1 Ultimate stability is provided by the developing capsule, which in turn protects the implant from fracture. He termed this process “encapsulation.” Implant fracture can also be minimized by allowing the stems to piston within the medullary canals, thus dissipating the forces of flexion and allowing the implant to find the best position with respect to the axis of motion.1 Because stability is eventually provided by the soft tissues, fractured implants do not imply loss of function. In fact, Swanson noted that most patients with radiographic evidence of implant fractures were doing well and did not notice any functional change.1 This concept is supported in the literature, as very few fractured implants have been revised. Long-term studies with a greater than 5-year follow-up period
Table 3. Patient Satisfaction Number of Patients Greatly satisfied Mildly satisfied Neither satisfied nor dissatisfied Mildly dissatisfied Greatly dissatisfied
8 4 1 1 2
Burgess, Kono, and Stern / Results of Revision Silicone MCP Arthroplasty
1511
Table 4. Preoperative and Postoperative Range of Motion for Pleased and Displeased Patients Pleased
Extension Flexion Ulnar drift
Displeased
Preoperative
Postoperative
Preoperative
Postoperative
Postoperative p Value
13° 49° 24°
17° 54° 9°
25° 57° 27°
30° 54° 30°
.007 .49 ⬍.001
have reported fracture rates of 0% to 63% with few revisions.2,3,7,10,11,15 Trail et al have reported the largest number of revisions, revising 76 of 1336 joints, 39 with fractured stems. With a revision rate of 3%, they concluded that radiographic implant failure does not require revision surgery.13 Gellman also revised 54 implants, all for recurrent deformity due to implant fracture.11 He notes, however, that there were 126 fractured implants in his series, and the remainder had no recurrence of deformity or loss of function. It appears that the soft tissues are more critical to long-term stability and function than the implant. However, the soft tissue support seems to weaken with time and progressive disease. Goldfarb and Stern reported at an average follow-up period of 14 years, an implant fracture rate of 63%.15 Clinical outcomes worsened. Only 38% of patients were satisfied with hand function and only 27% were pain free. In cases where revision is necessary, there is presumably greater insult to the soft tissues, not only from the second surgical procedure but from the stretching that has occurred with recurrence of deformity and the underlying destruction caused by the rheumatoid arthritis. As such, soft tissue reconstruction may be more difficult to achieve in the revision arthroplasty. Our revision series showed generally poor results from an objective standpoint. As a group, there was no significant change in the flexion range (preoperative 16° to 50°, postoperative 20° to 54°) and a small improvement in ulnar drift (preoperative 24°, postoperative 13°) at just over a 5-year follow-up period. In addition, there was a high implant fracture rate (34%) in the revisions, suggesting that the soft tissues were unable to support the forces at the joint, leading to excessive demand on the implant. We believe that these poor results are most likely due to incompetent soft tissues, secondary to multiple surgeries, active rheumatoid disease, foreign body reaction caused by particulate debris, or a combination of these. Although we found no correlation between silicone debris and endosteal erosion on x-ray, the effect of silicone particulate matter on soft tissues
and implant longevity is not currently known, and we did not alter our treatment based on these findings. Despite these poor objective results, the majority of patients were pleased with the results and would have the surgery done again. These pleased patients had similar postoperative MCP flexion but less extensor lag and ulnar drift than the displeased patients. The pleased patients had slightly less preoperative deformity to begin with, mainly affecting extension. The poor average correction for the entire group could suggest that greater preoperative deformity is associated with weaker soft tissues and subsequently worse correction. Patient displeasure also seems to be related to the ability to correct this deformity. Presumably, if some soft tissue stability can be maintained, good results can be obtained from both an objective and subjective standpoint. As such, careful patient selection is critical. An annual follow-up evaluation may be worthwhile. If there is evidence of progressive deformity or the patient complains of pain or loss of dexterity, a discussion can be held regarding revision MCP arthroplasty. It is very important that the patient understands that goals are limited and that malalignment, pain, and decreased dexterity may recur over time. There was a disparity between the physician’s documented reason for revision and the patient’s reason for undergoing revision. Physicians most commonly cited deformity for the reason to operate, whereas patients most commonly cited pain. This difference may be due to documentation styles, patient recall bias, or a true disconnect between the goals of the surgeon and the patient. Chung et al studied patient reasons for seeking MCP implant arthroplasty using the Michigan Hand Outcomes Questionnaire in 33 surgical and 28 nonsurgical patients.18 They found that worse hand function and increased pain correlated with seeking surgical treatment. Hand appearance was not significantly correlated with seeking arthroplasty. They did not, however, evaluate differences between patient and physician reasons for surgery. Future studies investigating this potential difference in motivation could shed some light on this issue.
1512
The Journal of Hand Surgery / Vol. 32A No. 10 December 2007
Finally, it should be noted that despite the overall inability to improve motion and ulnar drift, revision silicone arthroplasty was very successful in alleviating pain. Of the 6 patients who underwent revision surgery because of pain, 5 were satisfied and would have the procedure performed again.
8.
9. Received for publication September 27, 2006; accepted in revised form July 30, 2007. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Corresponding author: Scott D. Burgess, MD, Michigan Hand Center, 1111 Leffingwell NE, Ste 200, Grand Rapids, MI 49525; e-mail: [email protected] Copyright © 2007 by the American Society for Surgery of the Hand 0363-5023/07/32A10-0001$32.00/0 doi:10.1016/j.jhsa.2007.07.026
References 1. Swanson AB. Flexible implant arthroplasty for arthritic finger joints: rationale, technique, and results of treatment. J Bone Joint Surg 1972;54A:435– 455. 2. Wilson YG, Sykes PJ, Niranjan NS. Long-term follow-up of Swanson’s silastic arthroplasty of the metacarpophalangeal joints in rheumatoid arthritis. J Hand Surg 1993;18B:81–91. 3. Bieber EJ, Weiland AJ, Volenec-Dowling S. Silicone-rubber implant arthroplasty of the metacarpophalangeal joints for rheumatoid arthritis. J Bone Joint Surg 1986;68A:206 –209. 4. Beckenbaugh RD, Dobyns JH, Linsheid RL, Bryan RS. Review and analysis of silicone-rubber metacarpophalangeal implants. J Bone Joint Surg 1976;58A:483– 487. 5. Blair WF, Shurr DG, Buckwalter JA. Metacarpophalangeal joint implant arthroplasty with a silastic spacer. J Bone Joint Surg 1984;66A:365–370. 6. Mannerfelt L, Andersson K. Silastic arthroplasty of the metacarpophalangeal joints in rheumatoid arthritis. J Bone Joint Surg 1975;57A:484 – 489. 7. Kirschenbaum D, Schneider LH, Adams DC, Cody RP. Arthroplasty of the metacarpophalangeal joints with use of
10.
11.
12.
13.
14.
15.
16.
17.
18.
silicone-rubber implants in patients who have rheumatoid arthritis. Long-term results. J Bone Joint Surg 1993;75A: 3–12. Vahvanen V, Viljakka T. Silicone rubber implant arthroplasty of the metacarpophalangeal joint in rheumatoid arthritis. A follow-up study of 32 patients. J Hand Surg 1986; 11A:333–339. Olsen I, Gebuhr P, Sonne-Holm S. Silastic arthroplasty in rheumatoid MCP-joints. 60 joints followed for 7 years. Acta Orthop Scand 1994;65:430 – 431. Hansraj KK, Ashworth CR, Ebramzadeh E, Todd AO, Griffin MD, Ashley EM, Cardilli AM. Swanson metacarpophalangeal joint arthroplasty in patients with rheumatoid arthritis. Clin Orthop 1997;342:11–15. Gellman H, Stetson W, Brumfield RHJ, Costigan W, Kuscher SH. Silastic metacarpophalangeal joint arthroplasty in patients with rheumatoid arthritis. Clin Orthop 1997;342: 16 –21. Rothwell AG, Cragg KJ, O’Neill LB. Hand function following silastic arthroplasty of the metacarpophalangeal joints in the rheumatoid hand. J Hand Surg 1997;22B:90 –93. Trail IA, Martin JA, Nuttall D, Stanley JK. Seventeen-year survivorship of Silastic metacarpophalangeal joint replacement. J Bone Joint Surg 2004;86B:1002–1006. Chung KC, Kostis SV, Kim HM. A prospective outcomes study of Swanson metacarpophalangeal joint arthroplasty for the rheumatoid hand. J Hand Surg 2004;29A:646 – 653. Goldfarb CA, Stern PJ. Metacarpophalangeal joint arthroplasty in rheumatoid arthritis. A long-term assessment. J Bone Joint Surg 2003;85A:1869 –1878. Creighton JJ, Steichen JB, Strickland JW. Long-term evaluation of silastic trapezial arthroplasty in patients with osteoarthritis. J Hand Surg 1991;16A:510 –519. Bass RL, Stern PJ, Nairus JG. High implant fracture incidence with Sutter silicone metacarpophalangeal joint arthroplasty. J Hand Surg 1996;21A:813– 818. Chung KC, Kostis SV, Kim HM, Burke FD, Wilgis EFS. Reasons why rheumatoid arthritis patients seek surgical treatment for hand deformities. J Hand Surg 2006;31A:289 – 294.