Over the last three decades, the operative treatment of distal radial fractures has become increasingly common as compared with nonoperative treatment1-7. Over the last fifteen years, there has been a trend toward internal plate-and-screw fixation for the treatment of these fractures8-16 and away from percutaneous fixation with Kirschner wires or external fixation. An argument that is commonly made in favor of internal fixation of fractures of the distal part of the radius is that, similar to the findings that have been reported in association with other periarticular fractures, early movement of the wrist would be beneficial in order to obtain better wrist motion and thus better function4,16-18. To our knowledge, there are no comparative scientific data to support this concept. In fact, data regarding external fixation that either immobilizes the wrist or allows wrist motion suggest that early wrist mobilization is not as important as the overall alignment of the bone in terms of final wrist motion19.
We performed a clinical trial comparing mobilization of the wrist joint within two weeks (early motion) or at six weeks (late motion) after volar plate fixation of a fracture of the distal part of the radius in order to test the null hypothesis (the primary study question) that there are no differences in the flexion-extension arc three months after surgery. Secondary study questions included comparisons of range of motion, grip strength, physician-based wrist ratings, pain, upper extremity-specific health status, and determinants of the arc of flexion-extension of the wrist at three and six months after a fracture of the distal part of the radius.
Location and Eligibility Criteria
The study protocol was approved by our institution's Human Research Committee. Between May 2004 and November 2005, patients who satisfied the inclusion criteria were invited to participate in the study. The inclusion criteria were (1) an age of eighteen years or more, (2) a fracture of the distal part of the radius and no other skeletal injury, and (3) operative treatment with a fixed-angle volar plate and screws alone. All patients were enrolled in a tertiary care medical center in the practice of three orthopaedic hand and upper extremity surgeons (D.R., J.J., C.M.).
Randomization: Sequence Generation, Allocation, and Implementation
After providing informed consent, sixty patients were enrolled and randomized to either early or late wrist mobilization according to a sequence determined by a computer random-number generator (Excel 2003, Windows XP; Microsoft, Redmond, Washington). Thirty patients were assigned to the early motion arm of the study, and thirty were assigned to the late motion arm. Randomization and its implementation and the allocation of patients were executed by a research coordinator.
Interventions
Patients in both arms of the study had a volar plaster splint applied as part of the postoperative dressing; it was converted to a custom thermoplastic volar splint at the first postoperative visit. Patients in both study arms were encouraged to perform active and active-assisted digital and forearm motion exercises and to use the injured arm for light functional tasks prior to and immediately following surgery. The exercises of patients in both study arms were supervised by certified hand therapists.
Patients in the late motion group were advised to wear the thermoplastic splint at all times, except when showering. Wrist motion exercises were initiated at the time of the six-week postoperative visit (average, forty-nine days postoperatively; range, forty-two to fifty-nine days postoperatively).
Patients in the early motion group were taught to remove the thermoplastic splint and to perform active and active-assisted wrist motion exercises at the time of the first postoperative visit (average, eight days postoperatively; range, seven to thirteen days postoperatively). They also were encouraged to wean out of the splint in order to exercise the wrist during activities of daily living. No attempt was made to confirm adherence to the recommended protocol.
All fractures were internally fixed with a volar plate and screws with use of a standard Henry volar-radial approach20. The majority of the fractures (twenty-eight in the early motion group and twenty-nine in the late motion group) were treated with a Hand Innovations DVR plate (Miami, Florida); two fractures (one in each group) were treated with a Synthes 2.4-mm LCP volar column distal radius plate (Synthes, Paoli, Pennsylvania); and one fracture in the early motion group was treated with an Acu-Loc Targeted Distal Radius Plate (Acumed, Hillsboro, Oregon).
Blinding
Due to the nature of the study, it was not possible to blind the care providers (the orthopaedic surgeon and hand therapist) to treatment allocation. Functional and radiographic evaluation was performed by a blinded observer who was not involved in the patients' care.
Outcomes and Evaluation
Study evaluations were performed at three months (between two and four months) and six months (between five and eight months) after surgery. At each visit, patients had bilateral measurement of motion and grip strength; completed both an ordinal pain scale (with 0 indicating no pain and 10 indicating the worst possible pain) and the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire21; and were evaluated according to the modified Gartland and Werley score22 and the Mayo wrist score23. The uninvolved wrist was used for comparative measurements. Two patients sustained a distal radial fracture on the contralateral side while participating in the study. To provide control data for the comparisons of range of motion and grip strength in these two patients, two gender, age, and occupation-matched controls were selected from the study sample.
Standard radiographic measurements of distal radial deformity were made on the initial post-injury (pre-reduction) radiographs and on radiographs that were made three and six months after surgery, but not on the immediate postoperative radiographs. On the posteroanterior radiograph we measured the ulnar inclination of the articular surface of the radius (in degrees) and the ulnar variance (in millimeters), whereas on the lateral radiograph we measured the angulation of the radial articular surface (in degrees). We used the techniques of measurement described by Friberg and Lundström24,25.
Sample Size Calculation and Statistical Methods
We estimated a need to enroll twenty-three patients per group to achieve 90% power with an alpha of 0.05 to detect a difference of >10% in wrist flexion-extension arc, assuming a standard deviation of 10° for an effect size of 1.0 (nQuery; Saugus, Massachusetts). The target enrollment was sixty patients to account for an anticipated rate of patient loss during the trial of approximately 15%.
To confirm comparability of the two study arms after randomization, univariate analysis was performed with use of independent sample t tests for numerical variables (age, the interval between the injury and the operation) and with use of chi-square analysis for nominal variables (gender, occupation, hand dominance). Paired t tests were used to evaluate changes between three and six months after surgery.
We planned to analyze the data on an intention-to-treat basis, but there were no crossovers. A two-way repeated-measures mixed model analysis of variance was used to determine differences in the arc of wrist flexion-extension between groups at the three and six-month follow-up evaluations as well as to determine changes in flexion-extension within each group over time. A compound symmetry covariance structure was fitted to account for the within-subject variation; the Akaike information criterion (AIC) revealed good model fit to the longitudinal data. In addition, we analyzed the secondary variables with use of the same two-way repeated-measures mixed model, adjusting for changes within time and within subjects for each of the analyzed outcomes (motion, function, radiographic alignment, and pain). In addition, we sought predictors of the flexion-extension arc of the wrist at three and six months after surgery from among all of the demographic characteristics, injury characteristics, and radiographic parameters by performing univariate analysis (t tests for dichotomous independent variables and Pearson correlations for continuous independent variables) and then developing backward stepwise multiple linear regression models and entering all variables that were significant or near significant correlation (p < 0.08) in the univariate analysis.
Baseline Patient Characteristics
The study group included sixty patients (thirty-nine women and twenty-one men) with an average age of fifty-three years (range, twenty-five to eighty-three years). The early motion group comprised eleven men and nineteen women with an average age of fifty-five years. The late motion group comprised ten men and twenty women with an average age of fifty-one years. The groups were comparable in terms of gender and age distribution.
In the early motion group, sixteen patients performed desk-based work, four performed light-duty manual labor, one performed heavy-duty manual labor, eight were retired, and one was unemployed. In the late motion group, seventeen patients performed desk-based work, five performed light-duty manual labor, two performed heavy-duty manual labor, four were retired, one was disabled, and one was unemployed.
The greatest number of fractures were classified as intra-articular type-C fractures according to the AO system. In the early motion group, twelve fractures were classified as type A, one was classified as type B, and seventeen were classified as type C. In the late motion group, eleven fractures were classified as type A, seven were classified as type B, and twelve were classified as type C.
There were no differences in demographic characteristics (p > 0.05 for all), the interval between the injury and surgery (eleven days in the early motion group and eleven days in the late motion group; p = 0.81), or the initial/enrollment ordinal pain score (5.2 points for the early motion group, compared with 3.9 points for the late motion group; p = 0.49).
In the early motion group, assessment of the initial post-injury (pre-reduction) radiographs demonstrated an average volar tilt of 20° in the ten patients with a volarly displaced fracture and an average dorsal tilt of 28° in the twenty patients with a dorsally displaced fracture. In the late motion group, radiographic assessment demonstrated an average volar tilt of 20° in the eleven patients with a volarly displaced fracture and an average dorsal tilt of 23° in the nineteen patients with a dorsally displaced fracture. The average ulnar inclination of the radius was 5° in the early motion group and 8° in the late motion group. The early motion group had an average of 5 mm of positive ulnar variance, and the late motion group had an average of 8 mm of positive ulnar variance.
Among the sixty patients who enrolled in the study and completed the initial assessment (thirty patients per study arm), fifty-six patients (93.3%; twenty-nine patients in the early motion group and twenty-seven in the late motion group) completed the three-month follow-up and fifty-four patients (90%; twenty-eight patients in the early motion group and twenty-six in the late motion group) completed the six-month follow-up. There was no crossover between the arms of the study (Fig. 1). All of the fractures healed without implant problems or major loss of alignment (>5° or 2 mm) between the initial postoperative and six-month postoperative radiographic measurements.
Additional Operations and Complications
In the early motion group, two patients had an acute median nerve neuropathy that was treated with operative release at the time of the initial operation. In the late motion group, two patients had prophylactic release of the carpal tunnel at the time of the initial operation. None of the patients had finger stiffness at the three or six-month evaluations.
In the early motion group, one patient had a carpal tunnel release five months after surgery, one patient had mild crepitation of the flexor digitorum profundus tendons over the plate with pressure over this area during active finger motion, one patient had a superficial wound infection that resolved with oral antibiotics, and one patient had slight postoperative volar subluxation of the radiocarpal articulation. In the late motion group, three patients had mild crepitation of the flexor digitorum profundus tendons over the plate with pressure over this area during active finger motion, one patient underwent implant removal at eight months because of irritation of the first dorsal compartment tendons caused by a screw that protruded on the dorsal-radial aspect of the wrist, and one patient had slight postoperative volar subluxation of the radiocarpal articulation.
Three Months Postoperatively
Three months postoperatively, there were no significant differences between the groups with regard to wrist or forearm motion on the injured side, grip strength, the Gartland and Werley or Mayo wrist scores, the ordinal pain score, the DASH score, or radiographic measurements (Tables I and II). There were no univariate predictors of wrist flexion-extension arc three months after surgery among demographic characteristics, injury characteristics, and the radiographic results.
Six Months Postoperatively
Six months postoperatively, there were no significant differences between the groups in terms of wrist or forearm motion, grip strength, the Gartland and Werley score, the Mayo wrist score, the ordinal pain score, the DASH score, or radiographic measurements (Tables III and IV).
Univariate analysis of predictors of greater flexion-extension arc of the wrist six months after surgery identified a number of factors as being significantly or nearly significantly (p < 0.08) associated: female gender (t = 1.85; p < 0.08), late initiation of wrist motion (t = -2.6; p < 0.05), and anatomical restoration of volar angulation (r = -0.28; p < 0.05) and ulnar inclination (r = -0.25; p < 0.08). The best multiple linear regression model included gender, volar angulation, and late initiation of wrist motion but only accounted for 17% of the variability in the flexion-extension arc of the wrist (adjusted R2 = 0.17, F = 4.52, p < 0.01). A model including only late initiation of wrist motion accounted for only 10% of the variability in the flexion-extension arc of the wrist (adjusted R2 = 0.10, F = 6.82, p < 0.05).
Improvement Between Three and Six Months Within Groups
Between the three and six-month assessments, there was significant improvement in the average flexion-extension arc on the injured side in both the early motion group and the late motion group, with increases of 20° (from 104° to 124°) and 19° (from 107° to 126°), respectively. The average improvement in the arc of radioulnar motion was 10° in both groups (from 45° to 55° and from 50° to 60°, respectively). The average arc of forearm rotation increased by 10° (from 168° to 178°) in the early motion group and by 7° (from 171° to 178°) in the late motion group. The average grip strength significantly increased by 4.6 kg in the early motion group and 5 kg in the late rehabilitation group (p < 0.05 for all comparisons).
Wrist rating scores, pain scores, and DASH scores also improved significantly in each group between the three and six-month evaluations (p < 0.01 for all). In the early motion group, the Mayo score improved by an average of 10 points, the modified Gartland and Werley score improved (decreased) by an average of 1.1 points, the ordinal pain score decreased by an average of 0.9 point, and the DASH score decreased by an average of 10.5 points. In the late motion group, the Mayo score improved by an average of 12 points, the modified Gartland and Werley score improved by an average of 1.8 points, the ordinal pain score improved by an average of 0.5 point, and the DASH score improved by an average of 8.9 points (p < 0.01 for all). There were no significant changes in the radiographic measurements within groups between three and six months after surgery.
Wrist mobilization within two weeks after fixation of a distal radial fracture with a volar locking plate appears to be safe (with no loss of fixation or alignment), but it does not improve the arc of wrist flexion-extension or any other measure of function or health status as compared with wrist mobilization six weeks after surgery. On the basis of the data from this clinical trial, we suggest that claims that earlier wrist motion after internal fixation of fractures of the distal part of the radius can lead to improved wrist motion and function should be regarded with skepticism.
Several factors should be considered when interpreting these data. First, it would have been better to exclude patients who had a bilateral fracture because such patients are uncommon and because comparison with the contralateral wrist cannot be performed. Future studies should make this exclusion; however, we adhered to our original protocol and analysis because deviations from these predetermined protocols allow for the introduction of bias. Second, the present study was a study of effectiveness (i.e., what happens in the real world when one prescribes a given treatment) rather than efficacy (i.e., what happens under ideal conditions when patients do exactly as they are instructed to do). In respect of patient autonomy, and as required by the Human Research Committee at our hospital, no attempt was made to enforce compliance with the prescribed treatment (for instance, solid casts were not used in the late motion group). Furthermore, we did not monitor compliance with the prescribed treatment. Consequently, the present study informs us about what happens when we prescribe early or late wrist motion, but it does not tell us what would happen if patients adhered perfectly to the prescribed rehabilitation protocol. Finally, it is possible that our study would produce different results with specific variations: for instance, mobilizing the wrist within a few days after surgery, evaluating the outcomes at a very early time-point such as two or three weeks, or using hand swelling, the occurrence of hypersensitivity, or even lower overall costs as primary outcome measures. Our bias is that the majority of patients have too much pain a few days after surgery to effectively move their wrist and that such pain may hinder rehabilitation of the hand and forearm, but such suppositions require further study.
Our findings are consistent with those encountered in studies of external fixation that allowed wrist motion (so-called nonbridging external fixation). Investigators have documented improved alignment when the external fixation directly engages the distal fracture fragments (compared with external fixation that bridges the wrist and does not involve the use of ancillary Kirschner wires, bone grafts, or bone-graft substitutes to support the fracture), but the overall motion (i.e., the flexion-extension arc of the wrist) is not improved by allowing free wrist motion during the first six postoperative weeks1,26-28. In addition, studies in which external fixation that immobilized the wrist (so-called bridging external fixation) was compared with open reduction and plate-and-screw fixation also did not demonstrate significant differences in wrist motion, regardless of whether the plates were applied to the dorsal or the volar surface of the distal part of the radius1,6,29-34.
The low rate of complications in the present series is consistent with the findings of previous studies14,35-38. On the other hand, the short study period does not allow us to comment on later complications such as tendon irritation or injury and implant loosening, the two most common complications encountered after a minimum duration of follow-up of twelve months in a recent prospective cohort study of 141 consecutive patients who had been managed with a volar plate and screws for the treatment of a dorsally displaced fracture of the distal part of the radius36.
Several other observations were made. First, all patients maintained full digital motion and, on the average, obtained near full forearm rotation and DASH scores that approached population norms within six months after the injury39. Second, patients recovered pronation earlier than supination. Third, the flexion-extension arc of the wrist was more substantially and variably impaired. While our analysis demonstrated that the restoration of volar angulation of the articular surface on the lateral radiograph and the late initiation of wrist motion were significant predictors of a greater flexion-extension arc of the wrist six months after surgery, the correlations were weak and the multivariate model only accounted for 17% of the variation in flexion-extension arc of the wrist. Consequently, we can conclude that the determinants of the flexion-extension arc of the wrist—the most variable objective aspect of recovery following a fracture of the distal part of the radius—remain poorly understood. We cannot comment on later functional outcomes, but we believe that only small improvements should be expected after six months; in fact, the improvements between three and six months after surgery, although significant, were small.
Our data suggest that six weeks of postoperative wrist immobilization does not compromise wrist function and arm-specific health status (DASH scores) at three and six months after surgery when compared with mobilization of the wrist within two weeks. By the same token, we found little to condemn early wrist mobilization, and there may be benefits with respect to costs and return to function fewer than three months after the injury that were not measured in our investigation. Perhaps these factors merit additional study. 