Abstract
Thirty-nine patients were retrospectively reviewed after a wrist arthrodesis using a Synthes wrist fusion plate and iliac crest bone graft. Information was obtained from review of patient files, a questionnaire to assess pain, function and work status, and clinical assessment of grip strength, forearm rotation and fingers motion. All wrist fusions united except that the index carpometacarpal joint failed to unite in one patient. Thirty-seven patients were satisfied with the procedure, noting a reduction in wrist pain after fusion, but all reported some limitation of function. The wrist fusion plate was removed in six patients and a further four patients experienced minor symptoms over the dorsal aspect of the middle finger metacarpal.
INTRODUCTION
Arthrodesis is the preferred surgical option for end-stage wrist disease. Plate fixation has reduced the incidence of nonunion but can cause extensor tendon irritation requiring plate removal. The Synthes wrist fusion plate (Mathys Medical Bettlach, Switzerland) is designed specifically for wrist arthrodesis and does not require intra-operative contouring. It is tapered distally to provide low profile fixation on the middle finger metacarpal, thereby theoretically reducing the risk of extensor tendon irritation. It is, however, significantly more expensive than a 3.5 mm dynamic compression plate (DCP – original implant of the Association for the Study of Internal Fixation, AO/ASIF, Mathys Medical Bettlach, Switzerland).
PATIENTS AND METHODS
Thirty-nine consecutive wrist fusions, performed between August 1994 and July 2000, were identified from the practices of two surgeons. Patient notes and X-rays were reviewed, in which pre-operative grip strengths and ranges of motion were recorded. Twenty-two patients were examined clinically and ten interviewed by telephone. These 32 also answered a questionnaire covering residual symptoms, function and work status. The other seven patients could not be contacted, but sufficient follow-up information was obtained from their clinical notes with an average follow-up of 15 (range, 4–30) months. The overall duration of follow-up averaged 26 (range, 4–78) months. X-rays were taken at 6 and 12 weeks and then until union. Further X-rays were taken at review.
Operative technique
The procedure was performed under axillary block and general anaesthesia with a pneumatic tourniquet on the upper arm.
A dorsal approach was used in 38 wrists, a palmar approach in one. The third extensor compartment was opened and the EPL tendon retracted. The second and fourth compartments were reflected subperiosteally and the articular branch of the posterior interosseous nerve was resected. The middle finger metacarpal was exposed subperiosteally and the radiocarpal joint, midcarpal joints (scaphocapitate and capitolunate) and the index and middle finger carpometacarpal joints were then decorticated until cancellous bone was exposed. A tricortical bone block was harvested from the iliac crest if avascular bone, such as Kienböck’s disease or an avascular proximal pole of scaphoid, was present. Cancellous bone was also harvested from the iliac crest and inserted into the decorticated areas.
The best-fitting titanium Synthes wrist fusion plate was then selected and the wrist was fused in 10° of extension (38 cases), with neutral radio-ulnar deviation and neutral rotation. A straight plate was utilized for the wrist which was fused through a palmar approach. The metacarpal screws were inserted first, followed by those into the distal radius. A further screw was placed obliquely through the capitate to the base of the middle finger metacarpal. The periosteal/capsular flap was then closed over the plate and the extensor retinaculum was repaired under the EPL tendon. The skin was closed with interrupted 5-0 nylon sutures and a palmar plaster slab was applied for 2 weeks. A removable thermoplastic wrist splint was then applied in reliable patients though the others wore a fibreglass cast for 4 weeks. Finger exercises were commenced on the first post-operative day. The one patient who had a palmar approach had had a previous palmar plate applied to a distal radial fracture.
RESULTS
There were 31 men and eight women. Their average age at the time of fusion was 41 (range, 22–64) years. The right wrist was fused in 29 patients and the left in ten. The dominant hand was fused in 30 and the nondominant in nine. Twenty-six of the cases involved workers compensation or litigation.
The pre-operative diagnoses are shown in Table 1, the most frequent being scaphoid fracture nonunion (16 cases) and scapholunate advanced collapse (seven cases).
Patients in this series had undergone an average of 2.25 surgical procedures prior to fusion. Wrist arthroscopy was performed in 22 patients prior to fusion to determine their suitability for alternative motion preserving procedures (Table 2). All patients cited pain as the reason for the fusion. Stiffness and weakness were also reported pre-operatively. The short bend plate was utilized on 28 occasions, a standard bend plate in ten and the straight plate in one case.
All 39 wrists fused but one developed a mildly symptomatic nonunion of the index carpometacarpal joint.
Patients graded their pre-operative pain on an analogue scale from 0 to 4, with 0 as no pain and 4 as very severe pain. The average preoperative score was 3.3 (range 2.5–4.0 SD 0.47) compared with 1.0 (range 0–2 SD 0.79) post-operatively.
Thirty-seven of the 32 patients with follow-up were satisfied with the procedure and stated they would have the same procedure again. The other two patients indicated they had only gained marginal benefit.
All patients reported some loss of function. The most commonly reported restrictions were picking up small objects, fatigue when writing and changing of hands or increased difficulty with perineal care (Table 3).
Pre-operative grip strength in the involved hand averaged 21 (range, 2–43) kg. The mean post-operative grip strength was 31 (range, 6–64) kg. Comparison with the opposite side showed an improvement from 44% pre-operatively to 66% post-operatively. There was little difference between the comparative grip strengths of those who had the dominant or the nondominant side fused. Rotation of the forearm was unaffected with pronation and supination greater than 70° in all but three patients. No patient has required further procedures for the distal radio-ulnar joint.
The length of hospital stay averaged 2.5 (range, 1–5) days and time off work varied from 2 days to 16 months. Compensable cases had a greater length of time off work (4.8 versus 2.2 months) and were less likely to return to a similar level of work than noncompensable cases (Table 5). Heavy manual labourers were able to return to pre-injury duties or to a similar level of functioning or work in eight of 17 cases which was less than for light manual workers (eight of 14) (Table 4).
The plate was removed in six patients between 13 and 16 months post-operatively because of residual extensor tendon irritation. Four other patients reported minor symptoms due to extensor irritation but did not want the plate removed. Complications are divided into early and late (Table 6), with ten complications in 39 patients, and four cases returning to theatre. One patient required drainage of an infected iliac crest haematoma and another required surgical removal of an iliac crest drain that had been inadvertently sutured. Two patients with no pre-operative carpal tunnel symptoms developed acute carpal tunnel syndrome post-operatively requiring open carpal tunnel release at 15 and 21 days, respectively. A carpal tunnel release was performed in five patients prior to fusion and three had an open carpal tunnel release at the time of fusion, including the patient with the palmar approach.
There were two late complications. One patient presented with a painful nonunion of the index carpometacarpal joint at 22 months, but his symptoms were not sufficient to require further surgery. Another patient sustained an undisplaced fracture of the middle finger metacarpal after a fall 32 months after fusion. This was treated successfully in a cast.
DISCUSSION
Plate fixation is now the most commonly reported technique of wrist arthrodesis. The 3.5 mm dynamic compression plate requires intra-operative contouring, which increases operative time, and is relatively prominent, particularly distally over the metacarpals where soft-tissue coverage is thin. In contrast the Synthes wrist fusion plate is precontoured and probably saves operative time. It is also tapered distally both in thickness and width which reduces its prominence and, theoretically, soft-tissue irritation. The wrist fusion plate is considerably more expensive than the standard 3.5 mm dynamic compression plate, a disadvantage that can arguably be offset by decreased operative time and a reduced incidence of removal.
Field et al. (1996) reported a plate removal rate of 65% with the 3.5 mm dynamic compression plate or T-plate and 34% of 3.5 mm dynamic compression plates were removed in O’Bierne et al.’s (1995) series. The 15% rate of removal in our series is considerably lower than these, strongly suggesting that use of the lower profile wrist fusion plate reduces the incidence of plate removal, although patients may still report some minor symptoms due to tendon irritation. However, there are some situations in which the wrist fusion plate cannot be used, such as when there is also a distal radial fracture with metaphyseal/diaphyseal extension: its proximal length is then insufficient.
Grip strength improved considerably in our patients after wrist fusion, although not to normal values. O’Driscoll et al. (1992) measured maximal grip strength with the wrist in 33–40° of extension and 2–7° of ulnar deviation depending upon hand dominance. Improvement in strength compared to the other side has been previously reported at between 65% and 75% (Field et al., 1996; O’Bierne et al., 1995). O’Driscoll et al. (1992) reasoned that the inability of the patients to achieve greater grip strength is due to the position of fusion: arthrodesis in 10° extension may cause weaker grip although it is considered more cosmetic and functional than one fused in 30° extension.
The use of iliac crest bone graft is controversial. Sorial et al. (1994) used only a local bone sliding radial graft and Weiss and Hastings (1995) used local radial metaphyseal bone graft harvested from below the plate. Hartigan et al. (2001) used proximal carpal bone graft, performing a proximal row carpectomy at the time of fusion. In our opinion patients with avascular bone due to fracture nonunions of the scaphoid or Kienböck’s disease need this excised, and carpal height restored with a block of iliac crest graft. Bruno et al. (2001) demonstrated that the volume of packed bone graft available from the distal radius averaged 2.7 ml compared with 5.3 ml from the iliac crest. Furthermore, iliac crest bone graft is mechanically and biologically superior to radial bone graft. These differences become more apparent with increasing age and with immobilization or disuse of a limb (Biddulph, 1999). Harvest of the iliac crest bone does however contribute to increased morbidity as shown in our study, and it contributed to the complication rate and the average inpatient stay of 2.5 days. This is higher than the length of stay reported by Weiss and Hastings (2001), who performed 24 of 28 operations on an outpatient basis.
Our overall complication rate was higher than expected, but less than the 45% reported by Field et al. (1996). Two patients developed acute carpal tunnel syndrome requiring surgical release post-operatively. A previous report indicated a post-operative incidence of carpal tunnel syndrome of 25% in patients undergoing a wrist fusion (Field et al., 1996).
Inclusion of the carpometacarpal joints within the fusion mass is also controversial. There is only a minor degree of movement of these joints and the fusion plate only crosses the middle finger carpometacarpal joint. However if this plate remains in situ, a small degree of movement at this joint may potentially lead to a fatigue fracture of the plate. This fatigue should occur in the part of the plate that overlies the distal carpus and carpometacarpal joint as this is the most tapered region and thus the weakest point as measured in a straight plate (Richards et al., 1993). Bolano and Green (1993) recommended that fusion of the wrist should include the index and middle finger carpometacarpal joints in patients with heavy labour jobs. They had to revise two of 18 radio-capitohamate fusions due to carpometacarpal joint pain. However problems associated with nonunion of the index finger carpometacarpal joint can be troublesome, as shown in our series.
One fracture of the middle finger metacarpal occurred in our series. Despite the use of smaller screws for fixation of the plate, the distal screw holes and end of the plate are stress risers which can result in a fracture of this bone after a fall. Hartigan et al. (2001) also reported this complication in two of 17 wrist fusions.