Union Rates and Complications After Thumb Metacarpophalangeal Fusion

Abstract

Background: Fusion of the thumb metacarpophalangeal joint (MPJ) can be performed using tension band wiring (TBW) or plate and screw (PS) fixation. This study evaluated results and complications using these techniques. Methods: A retrospective review of patients who underwent thumb MPJ fusion at our institution from 2010 to 2016 was performed. Patients with >1 year follow-up were included. Demographic information, indication for fusion, time to fusion, and complications were collected. Final radiographs were examined and alignment measured. Results: There were 56 thumbs in 53 patients (42 women and 11 men) including 12 TBW and 44 PS. The mean age was 60.9 years, and follow-up was 32.4 months. Twenty-eight of 44 plates were nonlocking, and 16 were locking. Of the locking plates, 7 of 26 used all locking screws, and 9 of 26 had a combination of locked and nonlocked screws. The mean flexion angle for TBW was 16.5° and PS was 12.8°. The mean coronal angle for TBW patients was 4.0° ulnar and PS was 2.5° ulnar. The overall union rate was 95%. There were 12 complications, 9 in the PS group. The TBW complications were painful hardware requiring removal. Eight complications in the PS group occurred in patients with locked plates. Five of the delayed or nonunions occurred in patients with locked plates and 4 of these were in plates with all screws locked. Conclusion: Complications using PS or TBW are not infrequent. Alignment with both techniques is similar, but use of locked plates specifically increases the rate of delayed or nonunions. We do not recommend routine use of locked plates for fusion of the thumb MPJ.

Introduction

Arthrodesis of the thumb metacarpophalangeal joint (MPJ) is indicated for a variety of posttraumatic and degenerative disorders.^8,9,13 Fusion can be utilized to restore stability in cases of chronic collateral ligament insufficiency,⁷ to relieve pain in the arthritic joint,^8,11 or in conjunction with basal joint arthroplasty in the setting of MPJ hyperextension.¹

Multiple surgical techniques have been described to fuse the thumb MPJ including tension band wiring (TBW), intramedullary devices, and plate and screw (PS) fixation.^2-4,8-13 Generally, these reports have yielded high rates of fusion (ranging from 85-100%) and relatively low rates of complications regardless of technique (though complication rates as high as 30% have been reported).⁶ Each has advantages and disadvantages, and at our institution, we typically use either TBW or PS. TBW is a straightforward, cost-effective means of achieving fusion, but the hardware may be palpable and require removal. It may be easier to fuse the thumb in slight flexion, but more difficult to position the thumb in neutral alignment in the coronal plane. In contrast, PS produce stable fixation, and the option of a locked construct can be useful in osteoporotic bone. Use of a straight plate makes varus/valgus alignment easier, but bending the plate is required to achieve a flexed position.

The purpose of the present study was to review the results of thumb MPJ arthrodesis. Our hypothesis is that MPJ arthrodesis is a reliable procedure, yielding high rates of fusion. We further hypothesize that TBW would result in higher rates of hardware removal.

Methods

A retrospective review of patients who underwent thumb MPJ fusion at our institution from 2010 to 2016 was performed. Our surgical database was searched for the Current Procedural Terminology (CPT) code 26850, corresponding to MPJ fusion. Institutional review board approval was obtained and informed consent was waived per our institutional protocol. All patients who underwent thumb MPJ fusion were included. Patients who underwent MPJ fusion of any other digit were excluded. The indications for fusion included hyperextension instability during basilar joint arthroplasty measuring more than 30°, posttraumatic arthritis, and posttraumatic instability, all refractory to conservative treatment.

Data were collected including demographic information, indication for fusion, time to fusion, and complications. Delayed union (Figure 1) was defined as healing which took >6 months but which ultimately healed. Nonunion was defined as an arthrodesis which failed to heal >12 months postoperatively or hardware breakage. Final radiographs were examined on our Picture Archiving and Communication System (PACS, Sectra AB, Linköping, Sweden), and measurement of flexion angle and varus/valgus alignment were made using the PACS software. Patients were excluded if they had less than 1 year of follow-up, inadequate radiographs, or fusion using a technique other than TBW or PS. Statistical analysis was performed using Fisher exact test at a 0.05 level of significance.

Figure 1.

Thumb with delayed union at 8 months postoperatively. This arthrodesis healed at final follow-up. (a) Posteroanterior view. (b) Lateral view.

Surgeries were performed through a dorsal longitudinal incision overlying the MPJ. The extensor mechanism was split in the midline, a capsulotomy was performed, and the collateral ligaments released. The joint surfaces were prepared using hand tools to healthy appearing cancellous bone on each side of the joint. For the PS construct, the implant was positioned on the dorsal aspect of the metacarpal and proximal phalanx and secured with screws. Compression was applied either through the plate when combination holes that allowed for compression were available on the implant (Figure 2) or through the fusion site prior to fixation when they were not. Nonlocking plates, locked plates with all locked screws, and locked plates with a hybrid (locked and nonlocked screws) were used.

Figure 2.

Thumb fused with plates and screws. (a) Posteroanterior view. (b) Lateral view.

When using the TBW technique, the joint was secured with two 0.045-inch Kirschner wires (K-wires) placed in parallel from the dorsal metacarpal into the shaft of the proximal phalanx. A TBW was placed transversely in the dorsal shaft of the proximal phalanx. The wire was passed around the K-wires in figure-of-8 fashion, tightened, cut short, and buried deeply. The K-wires were bent and cut short and the free ends rotated and impacted with the cut ends deep (Figure 3).

Figure 3.

Thumb fused with tension band wire. (a) Posteroanterior view. (b) Lateral view.

Results

There were 56 thumbs in 53 patients (42 women and 11 men) who met the inclusion criteria. There were 29 right and 27 left thumbs included. Twelve patients underwent fusion by TBW and 44 by PS. The mean age was 60.9 years (range: 18-77 years). The average age in the TBW group was 62 years and in the PS group was 60 years. The mean time to follow-up was 32.4 months (range: 12-86 months).

Thirty-two thumbs (57%) in 29 patients were treated for hyperextension deformity of the MPJ in the setting of carpometacarpal (CMC) joint arthroplasty, while 24 (43%) for instability with degenerative changes at the MPJ only. Of the 29 patients with concomitant CMC arthroplasty, there were 24 women and 5 men with an average age of 62.4 years (range: 48-77 years). In this group, 24 thumbs had the MPJ addressed with PS and 8 with TBW. Of the 24 patients treated for isolated MPJ pathology, there were 18 women and 6 men with an average age of 59 years (range: 18-76 years). In this group, 20 had the joints addressed with PS and 4 with TBW. There were 4 nonunions or delayed unions in the CMC group and 2 in the MP only group (P = .69).

Of the 44 plates that were used, 28 were nonlocking, and 16 were locking. Of the locking plates, there were 7 with locking-only screws, and 9 had a combination of locked and nonlocked screws.

The mean sagittal alignment (flexion angle) for patients treated with TBW was 16.5° and for PS was 12.8°. The mean coronal alignment for patients treated with TBW was 4.0° ulnar compared with PS which was 2.5° ulnar.

The overall complication rate was 21% (n = 12). There were 9 complications in the PS group: nonunion (n = 3), delayed union (n = 3), wound infection (n = 1), complex regional pain syndrome (n = 1), and intractable pain (n = 1). The 3 complications in the TBW group were all symptomatic, painful hardware requiring removal. All patients in the TBW group healed. The difference in complication rate between TBW and PS was not significant (P = .99). There was no statistically significant difference in union rate between PS and TBW (P = .33).

Among the PS group, 8 of 9 complications occurred in patients in whom locked plates were used. The difference in complication rate between locked and nonlocked plates was significant (P = .0005). Five of the 6 complications related to union occurred in locking plates compared with 1 with nonlocked plates, also a significant difference (P = .018). Of these 5, 80% (n = 4) occurred in plates where all the screws were locked compared with 1 in which a hybrid construct was used (P = .013).

Discussion

Fusion has generally been felt to be a reliable procedure in stabilizing the thumb MPJ and relieving pain. While several techniques have been described,^2-4,8-13 we have favored performing MPJ fusion by using either TBW or PS. The results of our study demonstrate that the rate of complications using these techniques is relatively high. Complications seem to vary, however, based on the specific technique that is utilized.

In particular, nonunion or delayed union occurred in 11% of all our patients and 14% of those treated with PS. This rate is higher than that reported in the literature, though the results of fusion with locked plates specifically has not, to our knowledge, been analyzed. The rate of complications with locked plates, particularly plates with all locking screws, was significantly higher than with nonlocked plates. We speculate that locked plates have a tendency to make the construct overly stiff, thereby resulting in a longer time to union. Our complication rate is also high but the union rate of 95% is comparable to previously reported results in similar studies.^5,10 Furthermore, only 1 nonunion occurred in the 28 patients treated with nonlocked plates, a rate which is also comparable to previous studies.

Hardware-related complaints were common in patients treated with TBW, who underwent removal 25% of the time. While relatively high, our removal rate is comparable to the 23% found by Renfree and Lara ⁶ in their series of 30 patients undergoing TBW. Aside from taking special care to minimize the prominence of the ends of the K-wires and tension wires, the higher removal rate with TBW compared with PS seems unavoidable.^4,5 Unlike rigidly fixed plates, K-wires have a tendency to back out of bone, which could potentially explain this discrepancy.

Screw fixation has been suggested as a means to decrease complications in thumb MPJ fusion. Messer et al² reported on MPJ fusion using a 3.0-mm cannulated screw. While they found a high rate of fusion (100%) in 18 patients, 4 (or 22% of their patients) had complications, including 2 who required hardware removal. This rate is very similar to what we found. Using a similar technique with a cannulated screw and threaded washer, Schmidt¹⁰ had 1 nonunion in 26 patients, but no patients had hardware pain requiring removal. More recently, high rates of fusion and low complication rates have also been found with an intramedullary interlocking device.¹³ However, the data are limited to 17 patients with an average follow-up less than 5 months.

Irrespective of technique, several authors suggest fusing the MPJ in slight flexion, neutral varus/valgus alignment, and slight pronation.^2,11 We hypothesized that the thumb alignment of patients would differ based on the method of fusion. Technically, placement of the K-wires in a dorsal-proximal to distal-volar/intramedullary position while using the TBW is facilitated by having the joint in flexion. When using PS technique, on the other hand, the plate can be placed straight on the dorsal aspect of the MPJ and must be manually bent to place the joint flexion. Also, because the plate is straight, placement along the central axis of both the proximal phalanx and the metacarpal results in neutral alignment of the thumb. However, with TBW, the thumb must be maintained in a neutral position while the K-wires are placed and the TBW is tightened, which would seem to allow for deviation from neutral more easily. Despite these considerations, we did not find a marked difference in the alignment of thumbs treated with either of these techniques.

There are several limitations to our study. First, due to the retrospective nature of the study (and the timing of follow-up), we were unable to determine the exact length of time to fusion. We cannot assess whether there is a difference in the rate of healing between these techniques. Second, we were not able to account for potential variability in postoperative protocols for each patient—The length of immobilization may have been different between patients or techniques. Third, we did not assess patient outcomes based on validated functional or satisfaction scores and cannot conclude whether patients were more satisfied or functional with one or the other technique.

In conclusion, based on the results of our study, thumb MPJ fusion with either TBW or PS can be expected to result in restoration of appropriate alignment of the thumb. Complications can occur with either technique, with hardware-related issues seeming to be more frequent in patients treated with TBW and nonunion or delayed union more frequent with PS. The use of locked plates appears to substantially increase the risk for complication. Based on our results, we do not recommend the routine use of locked plates for thumb MPJ fusion. If bone quality or other factors necessitate locked screws, we recommend a hybrid configuration to minimize this risk. Newer techniques and implants, which minimize the risk for hardware prominence while maximizing compression to improve union rates may be of benefit and require additional study. For now, we continue to use both TBW and PS techniques using standard, nonlocked plates in performing thumb MPJ fusion.