Subjects
We acquired in vivo kinematic data by means of three-dimensional computed tomography in order to investigate the pathological condition of forty-two wrists in thirty-seven patients with rheumatoid arthritis who had continuous wrist pain, instability, and/or extensor tendon rupture. We assessed the suitability of the midcarpal joint for partial arthrodesis by means of three-dimensional computed tomographic images with use of a bone surface model made with the wrist in a neutral position.
In eighteen of the forty-two wrists, the midcarpal joint was found to be well preserved. Eight of these eighteen wrists were in patients who elected not to have surgery or dropped out of the study. The remaining ten wrists were selected for partial arthrodesis; all had severe radiolunate damage, but all had a relatively well-preserved and congruent midcarpal joint. All patients were female, and the average age was fifty-eight years (range, twenty-three to seventy-one years). The average duration of rheumatoid arthritis was eleven years (range, five to seventeen years), and the average interval between surgery and postoperative image acquisition was nineteen months (range, five to forty-two months).
A radiolunate arthrodesis was used in six wrists; all had a severely damaged radiolunate joint but a normal radioscaphoid articulation. A radioscapholunate arthrodesis was chosen for four wrists in which destruction and subluxation or dislocation of the radioscaphoid joint were detected on direct intraoperative observation. All subjects consented to participate in the present study.
Surgical Technique
All of the procedures were performed by two specialists in hand surgery (H.M. and T.M.), who are the senior authors of the present study. The wrist joint was approached through a dorsal midline skin incision. The dorsal compartments were separated and elevated radially from the radius. In addition, we performed tendon transfer in three wrists, synovectomy in all ten wrists, hemiresection of the ulnar head in seven wrists, and a Darrach procedure in three wrists. The capsule of the radiocarpal and distal radioulnar joints was opened, and synovectomy, hemiresection of the ulnar head, and/or a Darrach procedure was performed as appropriate. The articular surfaces of the scaphoid, lunate, and radius were exposed, and the damaged articular cartilage and cortical bone were removed. For bone graft, we used cancellous bone harvested from the iliac crest in five of the ten wrists and the excised ulnar head in all ten wrists. For fixation, three or four Kirschner wires were used in six wrists and two Kirschner wires with a staple were used in four wrists. For radiolunate arthrodesis, the lunate was fixed in the neutral position with use of enough bone graft to restore height and to keep the distal surfaces of the scaphoid and lunate precisely aligned to produce smooth midcarpal motion with good congruity of the distal surfaces of the scaphoid and lunate. For radioscapholunate arthrodesis, the scaphoid and lunate were each positioned with a temporary Kirschner wire, which was used as a joystick. The scaphoid and the lunate were carefully aligned to the neutral position by fitting each to the proximal articular surface of the capitate. The average flexion of the lunate and scaphoid with respect to the longitudinal axis of the radius measured 51° ± 7° and 5° ± 4°, respectively, consistent with the natural posture as previously reported21. The decision to perform a radiolunate or radioscapholunate arthrodesis was made after direct observation of the cartilage status and the congruity of the radioscaphoid joint.
Postoperative immobilization involved the use of a long arm cast for approximately three weeks, followed by a removable splint for two weeks. The patients were able to perform active motion of the wrist and forearm four weeks after the operation. Complete osseous union was achieved in all patients within three months.
Image Acquisition
The technique that we used for in vivo three-dimensional kinematic evaluation has been described in detail previously3,15-20. We performed three-dimensional computed tomography on the wrists preoperatively and postoperatively with use of a clinical helical-type scanner with an image slice thickness of 0.625 mm (LightSpeed XTra; GE Healthcare, Waukesha, Wisconsin). We made images with the wrist in three different positions: neutral (in which the third metacarpal and the forearm axis were aligned), maximum wrist flexion, and maximum wrist extension. Data were saved in a standard format (Digital Imaging and Communications in Medicine) that is used commonly for transferring and storing medical images.
Segmentation and Registration
Regions of individual bones were segmented semi-automatically with use of a software program for image analysis (Virtual Place-M; AZE, Tokyo, Japan). The software generated three-dimensional surface bone models with use of the marching cubes technique22. The kinematic variables were calculated by registering the segmented bones14 from one position to another with use of the software. The accuracy of volume-based registration has been discussed previously15.
Three-Dimensional Evaluation of Joint Congruity
The three-dimensional bone models created from computed tomography slice data could be viewed and qualitatively evaluated from any direction and on any cutting plane with use of originally developed software (Orthopedics Viewer; Osaka University, Osaka, Japan)20. We qualitatively evaluated joint congruity by means of three-dimensional investigation of the bone surface models with use of this software preoperatively and postoperatively (Fig. 1).
Three-Dimensional Quantification of Range of Motion
The radius and all carpal bones except the pisiform were registered, and the relative motion of each bone was calculated. In general, the displacement of the moving body from one position to another can always be represented as a rotation around and translation along a unique axis, which is called a screw axis23. We defined this screw axis as an axis of rotation. The three-dimensional range of motion was calculated as the angle of rotation around the axis of rotation18.
Global wrist motion during wrist flexion-extension before the radiolunate or radioscapholunate arthrodesis, which was represented by capitate motion relative to the radius, was compared with postoperative global wrist motion, represented by capitate motion relative to the fixed lunate.
Midcarpal motion during wrist flexion-extension before the arthrodesis, which was represented by capitate motion relative to the scaphoid, lunate, and triquetrum, was compared with postoperative values.
We also evaluated radiocarpal and scapholunate motion preoperatively and postoperatively.
Three-Dimensional Evaluation of Direction of Motion
To quantify the three-dimensional direction of motion, we defined a grid for the radius, which was the orthogonal reference system described by Belsole et al. (Fig. 2, A)24. A consistent orthogonal reference system was established in the radius in the neutral position, as previously reported3,25.
After determination of each axis of rotation, the angle of the axis of rotation relative to the axis of wrist flexion and extension (the Z axis of the orthogonal reference system) was calculated as viewed in the axial plane of the radius, defined as the ZX plane. We defined this angle as the radially and palmarly-directed opening angle, as reported previously (Fig. 2, B)19,20. We evaluated the radially and palmarly-directed opening angle with regard to global wrist motion and midcarpal motion.
Statistical Analysis
All data were expressed as the mean and the standard deviation. Quantitative data on the wrists before and after the operation were compared with use of standard statistical formulas based on the Mann-Whitney U test. Differences were deemed to be significant at p < 0.05.
Source of Funding
This research was financially supported by the Nakatani Foundation of Electronic Measuring Technology Advancement, which did not play a role in the investigation. The financial support was used for the supplies, including two computers used for analysis.
Radiolunate Arthrodesis
Global Wrist Motion
The average global range of wrist motion during wrist flexion-extension was 59° ± 4° before radiolunate arthrodesis and 48° ± 21° after radiolunate arthrodesis. The postoperative value was 81% of the preoperative value, but the difference was not significant (p = 0.34).
The preoperative direction of global wrist motion was close to the sagittal plane, whereas the postoperative direction of global wrist motion was more oblique from radiodorsal to ulnopalmar along the so-called dart-throwing motion plane (Fig. 3). The radially and palmarly-directed opening angle was 5° ± 13° preoperatively, compared with 25° ± 10° postoperatively; this difference was significant (p = 0.01) (Table I).
Midcarpal Motion
The residual postoperative midcarpal motion and joint congruities between the scaphoid, lunate, and capitate were well preserved in all cases. The three-dimensional animation showed that all of the joints in the wrist were moving smoothly and congruently and that scapholunate motion appeared to contribute to global wrist motion after the radiolunate arthrodesis (see Appendix). The average preoperative ranges of capitate motion relative to the scaphoid, lunate, and triquetrum during wrist flexion-extension were 21° ± 9°, 44° ± 8°, and 36° ± 5°, respectively. After radiolunate arthrodesis, the average capitate motions relative to the scaphoid, lunate, and triquetrum were 19° ± 12°, 48° ± 21°, and 31° ± 16°, respectively. The postoperative capitate motion relative to the lunate was 109% of the preoperative value. The postoperative capitate motions relative to the scaphoid and triquetrum were slightly less than the preoperative values (90% and 87%, respectively). However, there was no significant difference in any range of motion in the midcarpal joint before as compared with after radiolunate arthrodesis (p = 1.00 for capitate motion relative to the lunate, p = 0.75 for capitate motion relative to the scaphoid, and p = 0.75 for capitate motion relative to the triquetrum).
The preoperative direction of capitate motion relative to the scaphoid during wrist flexion-extension was oblique from radiodorsal to ulnopalmar and the radially and palmarly-directed opening angle was 43° ± 25°, whereas the preoperative directions of capitate motion relative to the lunate and triquetrum were closer to the sagittal plane than to that of the scaphoid and the radially and palmarly-directed opening angles were 8° ± 13° and 7° ± 15°, respectively. After radiolunate arthrodesis, the direction of capitate motion relative to the scaphoid was oblique from radiodorsal to ulnopalmar, the radially and palmarly-directed opening angle was 41° ± 21°, and the directions of capitate motion relative to the lunate and triquetrum were 25° ± 10° and 24° ± 13°, respectively. The postoperative direction of capitate motion relative to the lunate was significantly more oblique than the preoperative value (p = 0.03). The directions of capitate motion relative to the scaphoid and triquetrum did not differ significantly before as compared with after radiolunate arthrodesis (p = 0.63 and 0.10, respectively) (Table I).
Radiocarpal Motion
The preoperative range of lunate motion relative to the radius was 17° ± 9°. The preoperative range of scaphoid motion relative to the radius during wrist flexion-extension was 44° ± 11°, whereas the postoperative value was 30° ± 15°. The range of scaphoid motion relative to the radius did not differ significantly before as compared with after radiolunate arthrodesis (p = 0.11) (Table I).
Scapholunate Motion
The preoperative range of scaphoid motion relative to the lunate during wrist flexion-extension was 30° ± 10°, whereas the postoperative value was 30° ± 15° (Table I). The difference was not significant.
Radioscapholunate Arthrodesis
Global Wrist Motion
The average preoperative and postoperative ranges of global wrist motion were 83° ± 24° and 47° ± 14°, respectively. The postoperative value was 57% of the preoperative value.
The preoperative direction of global wrist motion was close to the sagittal plane, whereas the postoperative direction of global wrist motion was more oblique from radiodorsal to ulnopalmar along the dart-throwing motion plane as compared with the preoperative direction of global wrist motion (Fig. 4). The radially and palmarly-directed opening angle was -2° ± 5° before surgery and 19° ± 11° after surgery; this difference was significant (p = 0.02) (Table I).
Midcarpal Motion
The residual postoperative midcarpal motion and joint congruities between the scaphoid, lunate, and capitate were well preserved in all cases (Fig. 1) (see Appendix). The average preoperative capitate motions relative to the scaphoid, lunate, and triquetrum during wrist flexion-extension were 18° ± 7°, 53° ± 21°, and 37° ± 14°, respectively. After radioscapholunate arthrodesis, the average capitate motion relative to the lunate (with the scaphoid) and triquetrum measured 47° ± 14° and 28° ± 13°, respectively. The postoperative capitate motions relative to the lunate and triquetrum were slightly less than the preoperative motion (88% and 76% of the preoperative values, respectively). However, there was no significant difference in midcarpal motion before radioscapholunate arthrodesis as compared with after radioscapholunate arthrodesis.
The preoperative direction of capitate motion relative to the scaphoid, lunate, and triquetrum during wrist flexion-extension was almost along the sagittal plane, and the radially and palmarly-directed opening angles were 3° ± 15°, 2° ± 6°, and 2° ± 3°, respectively. After radioscapholunate arthrodesis, the direction of capitate motion relative to the lunate (with the scaphoid) and triquetrum was oblique from radiodorsal to ulnopalmar, and the radially and palmarly-directed opening angles were 19° ± 11° and 22° ± 11°, respectively. The postoperative direction of capitate motion relative to the lunate and triquetrum became significantly more oblique than the preoperative value, from radiodorsal to ulnopalmar (p = 0.02 and 0.02, respectively) (Table I).
Radiocarpal Motion
Before radioscapholunate arthrodesis, the radiocarpal joint was incongruent between the scaphoid, lunate, and radius during wrist flexion-extension in all cases. The preoperative ranges of scaphoid and lunate motion relative to the radius were 66° ± 21° and 31° ± 15°, respectively (Table I).
Note: The authors thank Akira Goto, MD, PhD, Kunihiro Oka, MD, PhD, and Ryoji Nakao, computer programmer, Department of Orthopaedic Surgery, Osaka University, for assistance during parts of the experimental procedure.