We conducted a prospective, randomized, controlled, multicenter trial involving three hospitals: Haukeland University Hospital in Bergen, Norway; St. Olav's University Hospital in Trondheim, Norway; and Stavanger University Hospital in Stavanger, Norway. The study was approved by the Norwegian Data Inspectorate and the Regional Ethics Committee for Medical Research (REK-Vest) (approval number, 040/99-007.99.). Written informed consent was obtained from each patient.
Device
The dynamic fixator (DYNAWRIST) is a flexible distractor through which continuous dynamic traction is applied to the fracture site. The circular bars and a distal and proximal pin retainer constitute the system. Both bars contain a spring, the tension of which can be adjusted (to a maximum of 30 N). The inner bar pushes the pin retainers apart, whereas the outer bar pulls the pin retainers toward each other (Fig. 1), and the resultant force of the two bars causes traction at the fracture site. The attachment of the bars to the two pin retainers is designed with sufficient degrees of freedom to permit movement in both the coronal and sagittal planes. Thus, the distraction at the fracture site is maintained during flexion and extension as well as during radial and ulnar deviation. Unlike some previous flexible wrist fixators, in which the axis of rotation was determined by the fixator6, this device does not interfere with the natural movements of the wrist (Fig. 2).
The proximal and distal pin retainers are of different sizes. Because the half-pins have the same length (80 mm), the distance between the bars is smaller proximally than distally (i.e., the bars are not parallel) (Fig. 1). This creates an ulnar vector force that acts in the plane of the bars and pulls the fractured end of the radius to the radial side and pulls the carpus (and the distal fragment) to the ulnar side. The movement of the wrist and this resultant force provide "multiplanar ligamentotaxis," which may be capable of restoring nearly normal anatomy at the fracture site (Figs. 3-A through 3-H).
The static bridging fixators used in the present study were the Hoffmann II Compact bridging external fixator (Stryker, Mahwah, New Jersey) (used in Bergen and Stavanger) and the Pennig (static) wrist bridging external fixator (Orthofix SRL, Verona, Italy) (used in Trondheim).
Patients (Table I)
Seventy consecutive patients with unstable fractures of the distal part of the radius (AO types A3, C1, C2, or C3) were randomized either to dynamic external fixation or to closed reduction and external fixation with a static, bridging fixator (Fig. 4). The study group included fifty-one women and nineteen men ranging in age from eighteen to seventy years. A sample size calculation showed that in order to show a =5° difference in distal or volar tilt of the distal radial fragment with a 5% significance level and 80% power, thirty-five individuals would be needed in each group.
The patients were randomly allocated to one of the two groups with the aid of a random-numbers generator and with use of sealed, opaque envelopes. The envelopes were opened immediately after the doctors on call had informed the patients and had obtained their consent, just before the operation.
We included AO-type-A and C fractures, fewer than ten days old, that could not be maintained in a reduced position following closed reduction and cast immobilization or that were judged to be unstable on the basis of the initial radiographs. All open fractures as well as fractures requiring supplemental percutaneous Kirschner-wire fixation were excluded from the study. Finally, in order for a fracture to be included in the study, the incongruity of the radiocarpal and distal radioulnar joints had to be reducible by closed means.
The fractures were classified according to the AO/ASIF system14 by one senior orthopaedic surgeon at each hospital (P.H., K.R., and V.F.). In the dynamic fixator group, twenty fractures were extra-articular (AO type A) and fifteen were intra-articular (AO type C). In the static fixator group, nineteen fractures were extra-articular and sixteen were intra-articular. The energy of the injury was classified as low (a simple fall from a standing position; n = 59) or high (any other injury; n = 11)15. Thirty-eight fractures (54%) were on the nondominant side (Table I).
Surgical Technique
Twenty orthopaedic surgeons and residents treated the fractures. The majority of the operations were performed by trainee surgeons as part of their daily work.
Dynamic Fixator
With the patient under general or regional anesthesia, two threaded half-pins (diameter, 3 mm; length, 80 mm) were inserted parallel to the palm into the index finger metacarpal through small incisions, and two half-pins of the same size and length were inserted into the shaft of the radius through a limited incision (radial to the second dorsal compartment) to avoid the extensor tendons and the sensory branches of the radial nerve16. The half-pins were inserted into the radial shaft from the radial side, parallel to those in the index finger metacarpal, with the forearm in neutral rotation. The pins were inserted with use of a jig to secure the correct distance between the two pin groups and insertion at 90° to the shafts of the metacarpals and radius. The placement of the pins was checked radiographically to ensure bicortical fixation (Fig. 3-C). The fixator was then applied radial to the wrist, exerting a distraction force at the fracture site. The tension in the springs was adjusted until radial length was acceptable (i.e., the pins in the radial shaft were parallel to those in the index metacarpal) (Figs. 1 and 3-C). In all cases, the fractures were reduced sufficiently by distraction forces alone and no closed manipulation was performed.
Static Fixator
The distal half-pins were inserted in the dorsoradial aspect of the shaft of the index finger metacarpal. The proximal pins were inserted, with use of an open placement technique16, 5 to 10 cm proximal to the wrist joint dorsoradially and at 90° to the shaft of the radius. The pins were tested manually to ensure that they were solidly anchored, and fluoroscopy was used to confirm bicortical fixation. The fracture was then reduced by means of longitudinal traction and closed manipulation under fluoroscopic control. The frames were then locked, and a radiograph was made to assess the quality of the reduction.
Postoperative Treatment
Postoperatively, a therapist provided a written instructional form showing how to exercise the fingers, forearm, elbow, and shoulder. In the dynamic fixator group, the patients were allowed to start motion of the wrist in all directions on the first postoperative day. Formal physiotherapy was offered to some patients with long-lasting stiffness. The device was removed at a mean of six weeks (range, six to nine weeks), with some of the most unstable fractures in the dynamic fixator group being treated for as long as nine weeks.
Evaluation
The wrists were assessed clinically and radiographically at the time of removal of the fixator and at three, six, and twelve months postoperatively. Independent observers at each hospital performed the measurements (three senior orthopaedic surgeons [P.H., K.R., and V.F.] performed the anatomical assessments, and therapists performed the functional evaluation). Two patients died between the six-month and twelve-month follow-up evaluations. They were included in the study analyses, but their twelve-month evaluations were not available. Two other patients, one from each group, dropped out during the follow-up period, and their data are also incomplete.
Anatomical Assessment
The fracture position and the intra-articular step-off and gap between fragments were assessed before and after surgery, at the time of removal of the fixator, and at the six and twelve-month follow-up evaluations. Change in position during the period of fixation was assessed by measuring the radial length (relative to the ulna), the radial tilt (volar or dorsal) of the articular surface of the distal part of the radius, and the inclination (coronal plane angulation) of the articular surface of the distal part of the radius17. The examiners were not blinded with regard to the radiographs.
Functional Assessment
At all postoperative reviews, a functional assessment was performed by an independent physiotherapist with measurements of the mobility (range of motion) of the wrist joint and supination and pronation of the forearm. Pain was assessed with use of a visual analog scale (with 0 representing no pain and 10 representing severe pain), and grip strength was assessed with use of a Jamar dynamometer (Sammons Preston, Bolingbrook, Illinois). Appearance and complications were also assessed. The Disabilities of the Arm, Shoulder and Hand (DASH) score was assessed after one year on the basis of a self-evaluation with use of a Norwegian translation of the DASH questionnaire18,19. With the DASH score, 100 points represent total disability and 0 point represents no disability.
The complications that were assessed included redisplacement, malunion, distal radioulnar joint instability, pin-track infection, pin loosening (or other pin-site problems), premature fixator removal, osteomyelitis, wound infection, tendon injury or rupture, nerve problems (including carpal tunnel syndrome, ulnar tunnel syndrome, superficial radial nerve paresthesia, or neurapraxia), complex regional pain syndrome, severe finger stiffness or contracture, arthritis, nonunion, and refracture.
Statistical Analysis
Mean values with standard deviations were calculated separately for each time point and each measurement. To account for the repeated measures in the study, we used a mixed model for repeated measures with an autoregressive correlation structure to estimate the differences between the two products for all time points. The standard deviations and p values for comparisons of the mean values were calculated with use of dummy variables for each of the time points in the mixed models. The level of significance was set at p = 0.05.
Source of Funding
The design and development of the new fixator was supported by grants from the Norwegian Industrial and Regional Development Fund (SND).
Anatomical Results
Before surgery, ten patients had an intra-articular step-off or gap of 2 to 6 mm between fragments. After surgery, all but one patient in each group (both with a 2-mm step-off or gap) had congruent articular surfaces.
There were no significant differences between the two groups in terms of radial length either preoperatively or immediately after application of the device. However, at the time of removal of the fixator (at six weeks) and at the time of the latest follow-up (at one year), there was no shortening in the dynamic fixator group but there was a mean shortening of 2 mm (range, 1 to 4 mm) in the static fixator group (p = 0.05). There were no significant differences between the two groups in terms of tilt or inclination of the articular surface of the distal part of the radius at any time (Table II).
Functional Results
At the time of removal of the device, the mean wrist extension (and standard deviation) was 24° ± 18° in the dynamic fixator group, compared with 13° ± 12° in the static fixator group (p < 0.01). Wrist extension in the dynamic fixator group was also significantly better at three, six, and twelve months. The mean amounts of wrist flexion, radial deviation, and pronation-supination at the time of removal and at three months were significantly better in the dynamic fixator group than in the static fixator group (p < 0.05). At the time of the one-year follow-up, there was no significant difference between the two groups in terms of wrist or forearm range of motion (Table III).
The mean grip strength for the dynamic fixator group was 46% of the value on the contralateral side at three months, 70% at six months, and 90% at one year, compared with 34%, 63%, and 83%, respectively, in the static group. These differences were not significant.
There was no significant difference between the two groups in terms of the mean visual analog scale pain score at any time interval (data not shown). Likewise, the mean DASH score at the time of the latest follow-up was 8 ± 8 in the dynamic fixator group and 7 ± 9 in the static fixator group.
Complications
In the static fixator group, one fracture (an AO type-A3 fracture) redisplaced with considerable radial shortening during the treatment period. The fracture was re-reduced, and the fixator was changed to a nonbridging type. One patient from the dynamic fixator group who had an AO type-C3 fracture underwent open reduction and internal fixation because of persistent articular incongruity. Two other patients with an AO type-C3 fracture (one in each group) had a 2-mm step-off or gap and, according to the treatment protocol, should have undergone open reduction; however, neither patient had additional surgery.
Fifteen patients (43%) in the dynamic fixator group had a superficial pin-track infection, compared with four patients (11%) in the static fixator group (p < 0.01). All infections resolved after local wound cleaning or treatment with antibiotics. There were no chronic infections. One patient in the dynamic fixator group and two patients in the static fixator group had loosening of one or two pins at the end of the fixation period, but no pins were removed prematurely.
Four patients (11%) in the static fixator group and two patients (6%) in the dynamic fixator group had transient symptoms of dystrophy with swelling and finger stiffness. Two patients in the static group had thickening of the palmar aponeurosis at the time of the latest follow-up but had no sign of complex regional pain syndrome. There were no neurologic complications. There were no nonunions.
Transarticular external fixation is based on "ligamentotaxis," i.e., traction on fragments by ligaments20; however, some fragments without apparent soft-tissue attachment also may have an improved position after the application of traction21. Thus, external fixation is a versatile and useful tool for the treatment of complex distal radial fractures9. However, ligaments exhibit viscoelastic behavior, and there is a gradual loss of the final distraction force applied to the fracture site through stress relaxation. The immediate improvement in radial length can decrease considerably by the time of fixator removal. In several series of static external fixation, shortening of the radius of as much as 7 to 10 mm has been reported1,11. Even minor axial shortening of the radius has been shown to affect the functional outcome22. Thus, with a static bridging fixator, readjustment of the traction during the treatment period might sometimes be necessary to compensate for shortening of the radius. With the dynamic device, the springs maintain distraction and compensate for radial shortening. Furthermore, should it be necessary, the device easily allows for the application of additional traction force. Most patients in the present series demonstrated no radial shortening after the period of dynamic traction. In our pilot study13, twenty-four of thirty patients with dynamic traction had no radial shortening and four had only 1 to 3 mm of radial shortening. A possible explanation for this finding may be that continuous traction may stimulate stronger and more efficient callus formation, reducing the risk of shortening of the radius after removal of the dynamic fixator23.
The volar tilt of the distal radial articular surface is often restored inadequately with longitudinal traction alone. Longitudinal traction can be combined with radioulnar or dorsopalmar translation to provide "multiplanar ligamentotaxis" that is capable of restoring the normal anatomy of the distal part of the radius9,20. To achieve this objective, we have developed this dynamic external fixator system with a nonparallel distraction mechanism incorporated into the frame to allow the application of supplemental translational forces after distraction has been applied. Whether this had any influence on the final radiographic outcome in the present series is unknown. The mean volar tilt of the distal radial articular surface was better, but the improvement did not reach significance between the two groups (Table II).
In the previous series involving this fixator13, we noted fluoroscopically that, during the first five to eight days of treatment, range of motion of the wrist led to motion at the fracture site. As manipulative reduction of the fracture was not done during surgery, this motion at the fracture site induced the translational forces that had a positive effect on angular reduction. Therefore, the volar tilt of the distal radial articular surface in several fractures in the present series and the previous series13 improved during the period of dynamic traction. After about seven to ten days, however, movement occurs almost exclusively at the wrist joint. Movement at the fracture site combined with the use of an external fixator with less rigidity is also believed to stimulate callus formation9.
In the present series, we found no differences in the functional results between the groups at the latest (one-year) follow-up. However, most functional indices were significantly better at an early stage in the dynamic fixator group. These findings are similar to those seen with nonbridging fixators7. Shortening the period of recovery was a highly desirable outcome for our patients managed with the dynamic fixator.
Superficial pin-track infections were more common in the dynamic fixator group, presumably secondary to wrist motion resulting in skin irritation around the pins7. These minor pin-track infections were easily treated with cleaning and oral antibiotic therapy and did not compromise the final outcome in any case.
In the present series, we did not use supplemental Kirschner-wire fixation. Therefore, two of the most comminuted intra-articular fractures (AO type-C3 fractures) had a residual step-off after reduction. Such severely impacted fragments may not reduce with traction alone and may require percutaneous manipulation with use of supplementary Kirschner wires. In many cases, a single Kirschner wire can add substantially to the stability of fixation9.
A strength of the present study is that all patients were operatively managed with use of standardized techniques without supplementary Kirschner wires. Thus, the only variables in the series were the devices. To demonstrate that the surgical technique is simple, a large number of surgeons participated in the present study. A possible weakness is that we included both AO type-A and AO type-C fractures. It might be better if type-A and type-C fractures were evaluated in separate studies. Another weakness is that the patients had a wide range of ages. We also should have included DASH-score assessments at all follow-up evaluations. We found no way to blind those who evaluated the radiographs. However, as the examiners were not the operating surgeons, there was no reason to believe that these assessments might be biased.
We conclude that continuous dynamic traction with this new dynamic wrist fixator is an effective alternative to conventional static external fixators for the treatment of unstable extra-articular (AO type-A3) or intra-articular (AO type-C1, C2, and C3) fractures of the distal part of the radius.
Note: The authors thank medical statistician Stein Atle Lie, PhD, for help with the statistical analyses, and chief engineer Bjoernar Vassenden for help with the design of the new fixator.