Between January 1992 and December 1995, ninety-six adults who were less than seventy years old and had an intra-articular fracture of the distal aspect of the radius were managed with an operation performed by the senior one of us (K. D.) at the Department of Orthopedic Surgery of Ogori Daiichi General Hospital or of the Yamaguchi University School of Medicine. Patients who had an open fracture or a volar lip fracture of the distal aspect of the radius were excluded. The patients were prospectively randomized into two treatment groups: thirty-eight patients had arthroscopic reduction and external fixation, and fifty-eight patients had open reduction and internal fixation. Our protocol for treatment of a fracture of the distal aspect of the radius was based on whether a satisfactory reduction could be achieved by closed manipulation. A reduction was defined as satisfactory when there was no dorsal angulation; ulnar variance was within three millimeters of that of the contralateral, normal wrist; and the intra-articular step-off was less than one millimeter as seen on plain radiographs or computerized tomography scans. Immobilization in a sugar-tong cast was used for stable fractures, and external fixation was used for unstable fractures when satisfactory reduction was lost after release of traction. If reduction could not be achieved with closed manipulation, the patient was assigned to one of two treatment groups. Patients who were seen at the Ogori Daiichi General Hospital were managed with arthroscopically guided reduction with Kirschner-wire pinning and external fixation. Patients who were seen at the Yamaguchi University School of Medicine had conventional open reduction with pinning, with or without external fixation, or plate-and-screw fixation. The study protocol was reviewed and approved by the Institutional Review Board at each institution.
Four of the thirty-eight patients managed with the arthroscopic procedure and ten of the fifty-eight patients managed with open reduction were lost to follow-up before the minimum duration of twenty-four months, and they were excluded from the study. Thirty-four patients managed with the arthroscopic procedure and forty-eight patients managed with open reduction were available for analysis in the present study.
Group Managed with Arthroscopic Reduction
This group consisted of twenty men and fourteen women, who had an average age of fifty-two years (range, twenty to sixty-nine years). Seventeen fractures were on the dominant side, and seventeen were on the nondominant side. The mechanisms of injury included a fall on the road (ten patients), a fall from a height (fourteen), a traffic accident (nine), and a sports-related fall (one). A fall from a standing position was considered a low-energy injury, but all other mechanisms involved a high-energy injury with a high degree of metaphyseal comminution and articular incongruency. The fractures were classified with use of the system of Frykman10 and the AO/ASIF classification8,17 (Table I). Of the thirty-four patients, twenty had three-dimensional computed tomography scans, reformatted for preoperative planning of the arthroscopically guided reduction. The average time from the injury to the operative fixation was approximately five days (range, one to ten days). None of the patients had failure of the arthroscopic procedure and thus none had open reduction and internal fixation.
Operative Technique
The patient is positioned supine on the operating table, with the arm draped free on a hand-table. A roll is placed beneath the ipsilateral pelvis to aid in obtaining the bone graft from the iliac crest. A compressive elastic bandage is wrapped around the forearm to retard extravasation of fluid into the muscle compartments during arthroscopy. A tourniquet is routinely applied to the upper arm and inflated. We do not use intraoperative fluoroscopy for reduction.
We prefer three arthroscopic portals for visualization of the reduction. These include a volar portal between the flexor carpi radialis tendon and the radial artery (Fig. 1), the conventional dorsal 3—4 portal (between the extensor pollicis longus tendon and the extensor digitorum communis tendons), and the dorsal 4—5 portal (between the extensor digitorum communis tendons and the extensor digiti minimi tendon). Before the wrist is suspended in the traction tower, the volar portal is made under direct vision through a one-centimeter longitudinal skin incision between the flexor carpi radialis tendon and the radial artery. The volar aspect of the capsule is exposed after blunt dissection, and a small (three-millimeter) incision is made parallel to the capsular fibers. A 2.7-millimeter cannula is inserted into the volar portal, the wrist is suspended in the traction tower, and five kilograms of traction is applied. Plain posteroanterior and lateral radiographs are made to confirm the reduction of the fracture (Figs. 2-A, 2-B, 2-C, 2-D). An inflow cannula is inserted through a portal located ulnar to the extensor carpi ulnaris tendon (the dorsal 6U portal), and the joint is distended. A 21-gauge needle is inserted between the extensor pollicis longus tendon and the extensor digitorum communis tendons (the dorsal 3—4 portal) and between the extensor digitorum communis tendons and the extensor digiti minimi tendon (the dorsal 4—5 portal). The needle should pass easily into the radiocarpal joint, without impinging on either the carpal bones or the distal end of the radius, to ensure that the arthroscopic cannula will not be inserted into a fracture plane or an intercarpal joint. The portals are then made, and a blunt cannula is inserted.
The arthroscope (Sapphire View; Stryker, Kalamazoo, Michigan), a 2.7-millimeter instrument with a 30-degree angled field of vision, is introduced through the volar portal, and the shaver is used to remove the remaining hematoma for a clear view of the joint surface. From the volar portal, the degree of comminution, separation, and depression of the dorsal fracture fragments is assessed. Next, the arthroscope is introduced into the dorsal 3—4 portal, and the volar surface and the lunate fossa are examined.
After the fracture fragments have been identified, they are reduced with manual compression with or without the use of tenaculum forceps; a percutaneous Kirschner wire with a diameter of 1.5 millimeters or more is used as a joystick on individual fracture fragments. Fracture fragments can best be identified by first placing a needle into the joint. The Kirschner-wire joystick is placed six to eight millimeters proximal to the needle. The needle is then removed before manipulation and reduction of the fracture fragment. A probe or a blunt trocar can also be used to reduce and disimpact the fracture fragments.
The radial styloid fragment is usually reduced first. This provides an anatomical landmark to which remaining fragments can be reduced. A small incision is made in the anatomical snuff box to avoid the radial artery and the superficial radial nerve. Usually, a two-millimeter Kirschner wire is placed through the skin incision and is inserted through the styloid fragment into the radial shaft. Volar tilt and radioulnar inclination of the articular surface must be anatomically reduced at this stage, as the reduced styloid fragment will be the control fragment for reduction of the medial so-called die-punch fragment. Too much traction can overreduce the radioulnar inclination and cause dorsal tilt of the fragment. To prevent these deformities, the wrist should be placed in a position of slight flexion with reference to the preoperative plain radiographs made with the wrist in traction. Pin-caps are placed to avoid injury to the surgeon throughout the rest of the procedure. If present, the lunate die-punch fragment then is reduced to the radial styloid fragment (Figs. 2-E and 2-F). With comminuted fractures, it is common to identify both a volar and a dorsal die-punch fragment. The volar fragment, in particular, tends to rotate dorsally during traction due to ligamentotaxis. Traction is released slightly, and the wrist is placed in slight flexion so that the central depressed fragment can be elevated with use of a small periosteal elevator placed through the small skin incision into the metaphyseal bone proximal to the metaphyseal fracture line. The periosteal elevator can be used as a joystick to elevate and reduce the fracture fragment under arthroscopic control. If a sagittal gap is present, manual lateral compression can be used to aid in reduction (Fig. 2-E). The volar die-punch fragment then is fixed to the radial styloid fragment with a 1.5-millimeter Kirschner wire placed from the dorsal 4—5 portal under the control of the arthroscope in the dorsal 3—4 portal (Fig. 2-F).
The dorsal rim fragment is difficult to view from the dorsal portals. The arthroscope is therefore placed again through the volar portal, and a 1.5-millimeter Kirschner wire is inserted from the dorsal 3—4 or 4—5 portal to be used as a joystick to reduce the remaining dorsal fragment to the previously fixed styloid and volar die-punch fragments. The dorsal fragment is then pinned.
Bone-grafting is performed when the fracture is a severely comminuted, high-energy injury with bone defects in the metaphyseal region of the distal part of the radius after arthroscopically assisted reduction. We prefer to use cancellous bone grafts obtained from the iliac crest. The bone chips are packed through the small skin incision that was previously used for the insertion of the periosteal elevator, proximal to the metaphyseal fracture line between the extensor pollicis longus tendon and the extensor digitorum communis tendons.
After reduction of the fracture of the distal aspect of the radius, associated cartilage and soft-tissue injuries should be evaluated and treated. Carpal chondral fractures can be debrided and loose fragments, removed. Degenerative tears of the triangular fibrocartilage complex are left without débridement. There were no traumatic peripheral tears of the triangular fibrocartilage complex in this series. Radial tears with a small bone fragment of the ulnar rim of the distal part of the radius are fixed to the reduced die-punch fragment with use of a 0.8-millimeter Kirschner wire inserted from the 6R portal.
Partial tears of the scapholunate or the lunotriquetral ligament are debrided, and a complete separation between the involved carpal bones (a grade-III or IV tear, according to the classification system of Geissler12) is reduced with use of Kirschner-wire joysticks. In addition, the intercarpal joint is transfixed with 1.2-millimeter Kirschner wires, without direct operative repair of the tears, for ten to fifteen weeks.
We prefer to use an external fixator (Ace Colles Fixator; Ace Medical, Los Angeles, California) to maintain longitudinal length and alignment after arthroscopically guided reduction and percutaneous pinning (Figs. 2-G and 2-H). An associated fracture near the base of the ulnar styloid process (type 2B of the AO/ ASIF classification system17) can cause instability of the triangular fibrocartilage complex and, consequently, the distal radioulnar joint. If present, it is stabilized with a tension-band wiring technique, with use of one or two smooth 0.8-millimeter Kirschner wires.
Postoperative Care
Patients who have had arthroscopically guided reduction, percutaneous pinning, and external fixation, with or without a bone graft from the iliac crest, begin range-of-motion exercises for the fingers and rotation of the forearm immediately after the operation. The external fixation device is removed six to seven weeks after the operation, depending on the surgeon's evaluation of the postoperative radiographs. After the device has been removed, active exercises for motion of the wrist are performed for the first two weeks, after which passive exercises are performed. Dynamic and static progressive splints are used after two weeks of passive exercise if the patient has not achieved the goals of a minimum of 60 degrees of pronation-supination of the forearm and 100 degrees of flexion-extension of the wrist.
Group Managed with Open Reduction
This group consisted of fifteen male and thirty-three female patients, who had an average age of fifty-five years (range, seventeen to sixty-nine years). Twenty-three fractures were on the dominant side, and twenty-five were on the nondominant side. The mechanisms of injury included a fall on the road (twenty-one patients), a fall from a height (four), a traffic accident (twenty-two), and a sports-related fall (one). The fractures were classified according to the Frykman system10 and the AO/ASIF system17 (Table I).
All forty-eight patients had open reduction through a volar or dorsal approach. A dorsally displaced fracture is approached through a dorsal longitudinal incision. The extensor retinaculum between the third and fourth extensor compartments is reflected, and the wrist capsule is divided in line with the skin incision. A volarly displaced fracture is approached through a volar longitudinal incision. The incision is made through a proximally extended carpal tunnel incision, with reflection of the pronator quadratus from the radius.
Twenty-three patients had fixation with an AO plate and screws; eight of them had fixation with a dorsal plate and fifteen, with a volar plate. Fourteen patients had external fixation and pinning with Kirschner wires. Eleven patients had pinning, which was performed with Kirschner wires in six and with conehead wedging screws (Alphatec, Palm Springs, California) in five. Twelve patients had insertion of bone graft from the iliac crest. In four patients, hydroxyapatite graft was added to the bone graft.
Patients who had an external fixation device were managed with the same postoperative program as those who had an arthroscopically guided operation. Patients who had a stable fracture after internal fixation and thus were not managed with an external fixator wore a splint for four to six weeks. They began a program of active range-of-motion exercises, with removal of the splint, during the first two to four weeks, assuming that no displacement occurred. During the first two weeks, the patients also started a program of active and passive exercises for motion of the digits and rotation of the forearm. They progressed to a passive range-of-motion program, including exercises for the wrist for another two weeks, followed by resistive exercises for strengthening.
Follow-up Evaluations
A functional evaluation was performed at an average of thirty-one months (range, twenty-four to forty-seven months). Pain, swelling, grip strength, sensibility, and range of motion of the wrist and forearm were evaluated. The range of motion of the injured limb was expressed as degrees. The grip strength of the hand on the side of the injury, as measured with a Jamar adjustable dynamometer (Asimow Engineering, Los Angeles, California), was compared with that of the contralateral hand.
Posteroanterior and lateral radiographs of the injured wrist were available for all patients and were used for various measurements (Figs. 2-I and 2-J). The radiographs made at the time of the latest follow-up were evaluated for joint congruity and were supplemented with a sagittal computerized tomography scan of the wrist for all of the patients who had had the arthroscopic procedure and for thirty-five patients who had had open reduction3,16. The articular incongruity on the computerized tomography scan was measured with the arc method originally described by Cole et al.3. Radial inclination, volar tilt, and ulnar variance were calculated preoperatively and postoperatively, and the groups were compared with respect to these findings. The scapholunate angle, capitolunate angle, and scapholunate gap were measured to detect any carpal instability. The scapholunate angle, determined by the axes of the scaphoid and lunate, was measured on the lateral radiograph made with the wrist in neutral flexion-extension and neutral deviation and the forearm in neutral pronation-supination.
Osteoarthritis of the radiocarpal joint and the distal radioulnar joint was graded, according to the criteria of Knirk and Jupiter13, at the latest follow-up evaluation (Table III). Grade 0 indicates no osteoarthritis; grade I, slight narrowing of the joint space; grade II, marked narrowing of the joint space and the formation of osteophytes; and grade III, full-thickness loss of articular cartilage and the formation of osteophytes and cysts.
The data were quantified with two scoring systems: the system of Green and O'Brien as modified by Cooney et al.4,8, in which subjective and objective clinical data are used to determine a score, and the system of Gartland and Werley8,11, in which clinical and radiographic data are used. The result for each fracture was graded as excellent, good, fair, or poor with use of both scoring systems.
Statistical testing for significance was performed with the paired t test for paired continuous variables. Spearman's rank correlations were calculated for comparisons of two-scaled variables. The level of significance was set at p = 0.05.
There was no significant difference between the groups with respect to the age or gender of the patients, the side (dominant or nondominant) of involvement, the number of days between the injury and the operative treatment, the severity of the fracture according to AO/ASIF classification17, or the number of patients who had insertion of a bone graft from the iliac crest.
The duration of follow-up averaged thirty months, with a range of twenty-four to forty-seven months, for the patients managed with the arthroscopic procedure and thirty-two months, with a range of twenty-four to forty-six months, for those managed with open reduction (Table II).
According to the criteria of Gartland and Werley11, fifteen (44 percent) of the wrists treated with the arthroscopic procedure had an excellent result, thirteen (38 percent) had a good result, six (18 percent) had a fair result, and none had a poor result. In the group treated with open reduction, twelve (25 percent) had an excellent result, fifteen (31 percent) had a good result, nineteen (40 percent) had a fair result, and two (4 percent) had a poor result (Table II). According to the modified system of Green and O'Brien4, fifteen (44 percent) of the wrists treated with the arthroscopic procedure had an excellent result, thirteen (38 percent) had a good result, six (18 percent) had a fair result, and none had a poor result. In the group treated with open reduction, eleven (23 percent) had an excellent result; eight (17 percent), a good result; nineteen (40 percent), a fair result; and ten (21 percent), a poor result. Twenty-eight (82 percent) of the wrists treated with the arthroscopic procedure and eighteen (38 percent) of those treated with open reduction had an excellent or a good result according to both systems. This difference between the groups was significant (p < 0.001). The scores, assessed with use of the two systems, for the patients who were less than fifty years old and for those who were fifty years old or more were compared, and the difference was not found to be significant, with the numbers available, in either treatment group.
The average values (and standard deviations) for the range of motion, as measured on the involved side, for the patients managed with the arthroscopic procedure and for those managed with open reduction were, respectively, 57 ± 7.5 and 53 ± 10.8 degrees of extension (p = 0.05), 51 ± 10.0 and 45 ± 11.6 degrees of flexion (p = 0.01), 16 ± 5.7 and 12 ± 3.7 degrees of radial deviation (p = 0.001), 24 ± 6.2 and 19 ± 8.2 degrees of ulnar deviation (p = 0.001), 79 ± 11.9 and 79 ± 12.1 degrees of supination, and 76 ± 12.7 and 78 ± 13.0 degrees of pronation (Table II). With the numbers available, we did not detect an association between the maximum step and gap displacement as assessed with computerized tomography scans at the time of follow-up and the range of motion. The average grip strength of the involved hand was 86 ± 14.2 percent of that of the contralateral, normal hand for the patients managed with the arthroscopic procedure and 76 ± 20.1 percent for the patients managed with open reduction (p = 0.01).
On the average, radial inclination improved from 16 ± 6.8 degrees to 21 ± 4.3 degrees in the patients managed with the arthroscopic procedure, but it remained unchanged in those managed with open reduction (Table III). Volar tilt improved from -16 ± 12 degrees to 0 ± 8.7 degrees in the group managed with the arthroscopic procedure and from -3 ± 15 degrees to 4 ± 12 degrees in the group managed with open reduction (p = 0.05). Ulnar variance averaged 4 ± 2.8 millimeters before the reduction and 1 ± 2.1 millimeters at the latest follow-up evaluation in the group managed with the arthroscopic procedure, compared with 2.4 ± 3.0 and 2.5 ± 2.5 millimeters, respectively, in the group managed with open reduction (p = 0.005).
Nine (26 percent) of the thirty-four patients who had the arthroscopic procedure and twelve (25 percent) of the forty-eight who had open reduction had insertion of bone graft from the iliac crest. However, no significant difference was detected between the patients who had and those who had not had bone-grafting in either group with respect to the average ulnar variance measured at the time of the latest follow-up (p = 0.145).
Articular incongruity (step and gap displacement) was measured on the plain radiographs made before the reduction and at the latest follow-up evaluation as well as on the computerized tomography scans made at the most recent evaluation. The average maximum step displacement on the plain radiographs improved from 2.6 ± 1.2 millimeters preoperatively to 0.9 ± 0.8 millimeter postoperatively in the group managed with the arthroscopic procedure and from 2.4 ± 1.2 millimeters to 0.8 ± 0.9 millimeter in the group managed with open reduction. The average maximum gap displacement on the computerized tomography scans made at the follow-up evaluation was 1.0 ± 1.0 millimeter for the group managed with arthroscopy and 0.7 ± 1.0 millimeter for the group managed with open reduction. The average maximum gap displacement on the most recent radiographs was 0.3 ± 0.5 millimeter for the group managed with arthroscopy and 0.8 ± 0.8 millimeter for the group managed with open reduction; this difference was significant (p = 0.005).
There were few complications in this series. In the early postoperative period, five patients managed with the arthroscopic procedure and six managed with open reduction had severe stiffness of the ipsilateral finger joints. Reflex sympathetic dystrophy subsequently developed in two of those five and three of those six patients. The stiffness and the symptoms of reflex sympathetic dystrophy had resolved in all patients at the time of the latest follow-up. Pin-track infections developed in four patients managed with the arthroscopic procedure and in one patient managed with open reduction, although the infections were mild and healed after removal of the pins. Carpal tunnel syndrome developed between three and six months postoperatively in one patient managed with the arthroscopic procedure and in two patients managed with open reduction. These patients were managed unsuccessfully with nonoperative treatment for three months, and a carpal tunnel release was performed. The symptoms had resolved at the most recent follow-up examination.
The Kirschner wires that were used in the arthroscopic procedures were removed at an average of twelve weeks (range, eight to fifteen weeks) postoperatively. The plates and screws used for internal fixation in the patients who had open reduction were removed at an average of eight months (range, six to fifteen months) postoperatively.
All of the radial fractures were solidly united at an average of 9.2 ± 2.5 weeks after the arthroscopic procedures and 9.4 ± 2.4 weeks after the open reductions. With the numbers available, we found no significant difference between the two groups with respect to the average time to union (p = 0.282).
Posttraumatic osteoarthritis of the radiocarpal joint was assessed on the plain radiographs and the computerized tomography scans with use of a modification of the four-grade system of Knirk and Jupiter13. According to this system, eighteen of the wrists treated with the arthroscopic procedure had grade-0 (no) osteoarthritis; thirteen, grade-I osteoarthritis; and three, grade-II osteoarthritis. In the group treated with open reduction, twenty had grade-0; twelve, grade-I; twelve, grade-II; and four, grade-III. Therefore, some osteoarthritis of the radiocarpal joint developed in 47 percent of the patients managed with the arthroscopic procedure and in 58 percent of those managed with open reduction. The difference between the groups with respect to the development of osteoarthritis was found to be significant (p = 0.014).
At the most recent follow-up evaluation, a strong association was found between the maximum step and gap displacement and evidence of osteoarthritis of the radiocarpal joint on computerized tomography scans (p < 0.001). However, we found no association between the scores for osteoarthritis, according to the modified system of Knirk and Jupiter13, and the scores for overall outcome, as assessed with the two scoring systems, in either group (p = 0.376).
Of the twenty-five wrists treated with arthroscopy and for which the operative record fully described the findings with regard to the ligaments, eight had a grade-II or grade-III lesion of the scapholunate interosseous ligament, according to the system of Geissler12, and four had a grade-II or grade-III lesion of the lunotriquetral interosseous ligament. Seven injuries of the scapholunate interosseous ligament were found in the group managed with open reduction. All of the involved joints were immobilized by transfixation with Kirschner wires, without direct operative repair, for an average of twelve weeks. In the group managed with the arthroscopic procedure, four wrists had a traumatic tear (class 1D) and eight had a degenerative tear (class 2C, 2D, or 2E) of the triangular fibrocartilage complex, according to the criteria of Palmer18. Four traumatic tears of the triangular fibrocartilage complex that were associated with an avulsion fracture of the ulnar rim of the distal aspect of the radius were fixed with Kirschner wires under arthroscopic guidance. Injuries to the triangular fibrocartilage complex were not recorded for the group managed with open reduction.
None of the wrists had demonstrable carpal instability, including scapholunate dissociation or dorsal or volar intercalated segmental instability, except for one, which had dorsal intercalated segmental instability after open reduction and pinning. None of the wrists that had an injury of the scapholunate or lunotriquetral interosseous ligament was seen to have any pattern of carpal instability on the radiographs made at the most recent follow-up evaluation.
The average scapholunate angle was 43 ± 8.3 degrees in the group managed with the arthroscopic procedure, compared with 44 ± 7.0 degrees in the group managed with open reduction. The average capitolunate angle was 17 ± 29 degrees and 18 ± 5.2 degrees, respectively.
Osteoarthritis of the distal radioulnar joint was graded according to the criteria of Knirk and Jupiter13. In the group managed with the arthroscopic procedure, seventeen (50 percent) of the wrists had grade-0 (no) osteoarthritis, eleven (32 percent) had grade-I, four (12 percent) had grade-II, and two (6 percent) had grade-III. One patient had a Sauvé-Kapandji procedure one year after the injury because of chronic pain. Thirty (63 percent) of the wrists treated with open reduction had grade-0 (no) osteoarthritis, ten (21 percent) had grade-I, six (13 percent) had grade-II, and two (4 percent) had grade-III.
A comparison of the wrists with grade-0 or grade-I osteoarthritis of the distal radioulnar joint and those with grade-II or grade-III osteoarthritis did not reveal a significant difference, with the numbers available, with respect to the total scores according to either the system of Gartland and Werley11 or the system of Green and O'Brien as modified by Cooney et al.4 (p = 0.101). However, a significant association was found with respect to the range of supination and pronation of the wrists in both treatment groups (p = 0.038).
Twenty-one (62 percent) of the wrists that had the arthroscopic procedure and thirty-six (75 percent) of the wrists that had open reduction originally had a fracture of the ulnar styloid process. Seven fractures of the ulnar styloid process in the former group and five in the latter needed open reduction and internal fixation with a tension-band wiring technique. Seven fractures of the ulnar styloid process in the group managed with the arthroscopic procedure and twenty-four in the group managed with open reduction did not unite. However, no significant difference was detected, with the numbers available, between the wrists that had evidence of nonunion of the ulnar styloid process on plain radiographs and those without evidence of nonunion with regard to the total scores according to either the system of Gartland and Werley11 or the modified system of Green and O'Brien4 (p = 0.076).
There is a well established association between the development of posttraumatic osteoarthritis of the radiocarpal joint and residual displacement of articular fragments; the age, activity level, and occupation of the patient; and the follow-up interval after intra-articular fracture of the distal aspect of the radius1,7,13,19. The studies with the longest durations of follow-up in the literature have indicated that osteoarthritis of the radiocarpal joint develops over several years after an injury, especially in young adults2,7,9,13,20. The follow-up period in our study was shorter than that in other long-term reports concerning osteoarthritis after fractures of the distal aspect of the radius. However, 47 percent of the patients who had the arthroscopic procedure and 58 percent of the patients who had open reduction already exhibited early radiographic findings of osteoarthritis of the radiocarpal joint. Our series included patients who were older than those in other studies in the literature, but most of them were still active in the daily use of the involved wrist after treatment and they had the same results regardless of age.
A significant association was detected between the maximum step and gap displacement and evidence of osteoarthritis of the radiocarpal joint (p < 0.001). However, with the numbers available, we were unable to detect, in either group, significant associations between the score for osteoarthritis, according to the system of Knirk and Jupiter13, and the scores for overall outcome, according to the system of Gartland and Werley11 and that of Green and O'Brien as modified by Cooney et al.4. A gap or step displacement in the articular surface of more than one millimeter, as determined on plain radiographs after closed reduction, was considered a definite indication for the arthroscopic procedure or open reduction.
On the basis of our prospective comparative study, we found that the arthroscopically guided procedure was superior to the conventional open procedure with regard to several parameters. Specifically, the scores for outcome as assessed with the system of Gartland and Werley11 and the modified system of Green and O'Brien4, the range of flexion-extension and that of radial-ulnar deviation of the wrist, and the grip strength were better in the group managed with the arthroscopically guided procedure. These improvements may have resulted from the minimum scarring of the joint capsule and the adjacent soft tissues following the minimally invasive arthroscopic technique of reduction. Specifically, minimum scarring improved the postoperative range of motion of the wrist, except for supination and pronation, and this contributed to better scores in the overall evaluation. Supination and pronation are influenced predominately by incongruity and osteoarthritis of the distal radioulnar joint.
Arthroscopically guided reduction is a feasible procedure, but it requires meticulous technique. Unlike previous authors6,14,15,21,22, we did not use intraoperative fluoroscopy to assist the arthroscopic reduction. Preoperative three-dimensional computerized tomography scans and plain radiographs made just after closed reduction were sufficient for visualization of the various fracture lines and planning of the fracture immobilization. Adjustment of the positions of the pins was easily performed with arthroscopic guidance, without intraoperative fluoroscopy.
The extra-articular alignment of the distal part of the radius is an important factor in the achievement of a satisfactory result. The improvement in the volar tilt that was seen in a comparison of the initial and follow-up radiographs was more prominent after the arthroscopic procedures than after the open reductions. This finding was attributed to the type and severity of the fracture, as the group that had the arthroscopic procedure included more fractures with dorsally displaced fragments (Colles-type fractures) than did the group that had the open reduction. Both arthroscopic and open reductions can reduce volar tilt to within an acceptable range; however, the average volar tilt measured at the latest follow-up evaluations after the arthroscopic reductions was significantly less than that measured after the open reductions (p < 0.05). The arthroscopic procedure tends to result in a slightly more dorsally angulated reduction of the articular surface of the distal aspect of the radius in the sagittal plane because that reduction is performed with the wrist in a neutral position under straight traction. This tendency for dorsal angulation can be prevented by placing the wrist in a slightly flexed position during reduction and fixation of the radial styloid fragment with use of manual traction applied by the assistant or early application of the external fixation during the arthroscopic portion of the procedure. We prefer to use a traction tower rather than the external fixator because the external fixator obstructs the intraoperative positioning of the arthroscopic instruments and pins.
We noted a substantial correction in the radial inclination in the group that had the arthroscopic procedure; however, the group managed with open reduction did not show such improvement. This difference may have resulted from the differences in the severity of the initial fracture displacement, and it was not important from a clinical perspective.
The improvement in ulnar variance was greater after the arthroscopic procedures than after the open reductions. This difference may have been due to the fact that, with the arthroscopic procedures, the intraoperative traction during reduction and subsequent external fixation with direct reduction and fixation of the fragments can better maintain the length of the radius relative to that of the ulna. Open reduction and internal fixation was performed without a traction apparatus. Unlike previous authors1, we did not find that the use of a bone graft from the iliac crest resulted in a significant difference with respect to the long-term improvement of ulnar variance. However, this may have been due to the limited number of wrists that had bone-grafting.
There were no data in this series to support routine repair of any associated carpal ligament injuries, as late carpal instability was not an important problem in our patients at the time of the most recent follow-up.
The greater the number of arthroscopic portals used, the easier it is to reduce the intra-articular fragments. We used three portals for the arthroscopic view and fracture reduction. Conventional arthroscopic approaches are through only the dorsal portals, but these portals did not allow visualization of all intra-articular fractures and the dorsal fragments were usually out of arthroscopic view. The volar portal provided indispensable views to enable the reduction of the dorsal intra-articular fracture fragments as most fractures of the distal part of the radius had dorsal die-punch or rim fragments. The use of the arthroscopic technique has no contraindications except for open fracture, as it is difficult to make the closed optical cavity that is necessary for arthroscopic visualization, and fractures with multiple intra-articular fragments, as it is difficult to identify the original location of the fragments.
On the basis of these results, we concluded that arthroscopically guided reduction, with use of an additional volar portal to clarify all intra-articular fracture lines, and external fixation provides more accurate reduction of displaced intra-articular fractures of the distal aspect of the radius than does open reduction and internal fixation. The arthroscopic technique minimizes capsular and soft-tissue scarring, which may lead to limitation of the range of motion of the wrist and subsequent unsatisfactory clinical results.
NOTE: The authors thank Dr. Lamont J. Cardon for his contribution to the manuscript.