A two-year prospective multicenter study of a case series was conducted to evaluate the outcomes of operative treatment of unstable distal radial fractures with anatomically shaped angular stable plate fixation. Nine participating centers were involved, and institutional review board approval was obtained in each center. The study was conducted under the overall supervision of the AO Foundation Clinical Investigation and Documentation Center.
The inclusion criterion was a displaced fracture of the distal part of the radius in a patient between the ages of twenty-one and eighty years who provided written informed consent to participate in the study. The operative procedure was intended to be performed within ten days after the injury, even when the fracture had displaced following an alternative method of treatment such as cast immobilization or external fixation.
Exclusion criteria included existing medical disorders that have an impact on bone physiology such as a malignant tumor and hyperparathyroidism as well as vascular insufficiency of the upper limb. Polytraumatized patients with an Injury Severity Score25 of >16 points were also excluded. Patients with known alcohol or drug dependency, those who knew that they would be unable to participate, and those participating in other clinical trials of a drug or device were also excluded.
Baseline Evaluation
Patient demographics at the time of the injury, including sex, age, height, weight, limb dominance, smoking history, concomitant medical disorders, and medications, were documented and entered into an individual journal. The baseline injury characteristics included the type of accident, level of energy of the injury, associated injuries, fracture type according to the Müller-AO Comprehensive Classification26, associated soft-tissue injury at the fracture site, and delay between the injury and surgery.
The patients were interviewed regarding their upper-limb function as it was one week prior to the injury to determine a baseline Disabilities of the Arm, Shoulder and Hand (DASH) score with use of the extended three-module questionnaire. A DASH score of 0 points reflects no disability, whereas a score of 100 points signifies maximum disability27-30. In addition, each patient completed the Short Form-36 (SF-36) patient-rated health-status instrument, which is based on large cross-sectional pools of normative data with which comparisons can be made. An SF-36 score of 100 points represents "best health" while a score of 0 points reflects "poorest health."31-33 Baseline pain was also specifically assessed between the first and third postoperative days by asking the patient to rate his or her level of pain with use of a visual analogue scale, with 0 indicating no pain and 10 representing very severe pain.
Anteroposterior and lateral radiographs were made for each patient on admission, intraoperatively, immediately postoperatively, and at subsequent scheduled follow-up evaluations.
Follow-up Examinations
The study protocol required a follow-up evaluation at six weeks, six months, one year, and two years. The patient examination included measurements of wrist and forearm motion and grip strength and identification of any soft-tissue or neurologic changes. Wrist motion and forearm motion were documented both as absolute values on the injured side and as relative values as compared with the contralateral ("healthy") side. The measurements on the healthy side at the different follow-up intervals were compared by repeated-measures analysis of variance and were found to be not significantly different from each other. Therefore, the mean of all available measurements on the "healthy" side was calculated and was used as a reference value. Physician-rated evaluations included rating of their satisfaction with the procedure with use of a 0 to 10-point visual analogue scale and documentation of any unexpected occurrences or complications. Patients were asked to use the 0 to 10-point visual analogue scale to rate their pain, both at rest and with motion, as well as their overall satisfaction with the result of the treatment at each follow-up visit.
Overall physician-rated outcomes were documented with use of the scoring system of Gartland and Werley34 at the six-month, one-year, and two-year follow-up visits, and patient-rated outcomes were documented with use of the DASH and the SF-36 questionnaires. Any adverse occurrences were documented at one and two years throughout the entire two-year study period.
The radiographs from each center were digitized and were evaluated by independent radiologists. The quality of the postoperative radiographs was defined as "not usable," "bad," "poor," "acceptable," "good," or "very good," and data were obtained only from radiographs that were rated as "acceptable," "good," or "very good." The length of the implant was used to derive a correction factor for magnification. The measurements were made on high-quality paper prints of the digitized radiographs. The radial angle, radial length, ulnar variance, and articular step-off and gap were measured on the anteroposterior radiograph. The palmar tilt angle and articular step-off and gap were measured on the lateral radiograph. These measurements were recorded according to the criteria of Kreder et al.35. The arthritis status was assessed according to the system of Knirk and Jupiter36 at baseline and at the one and two-year follow-up examinations, with Grade 0 indicating no arthritis, Grade 1 indicating slight joint-space narrowing, Grade 2 indicating marked joint-space narrowing with osteophyte formation, and Grade 3 indicating complete loss of the cartilage space with osteophyte and cyst formation.
Patient Demographics and Baseline Characteristics
The study was initiated in January 2002. The plan was to recruit 150 patients, and this was accomplished by January 2004. The mean age of the cohort was fifty-one years, with a range of twenty-one to eighty years. Eighty-eight (59%) of the patients were women, with a mean age of fifty-seven years (range, twenty-one to eighty-three years). This group was significantly older than the sixty-two men, who had a mean age of forty-two years (range, eighteen to seventy-one years) (p < 0.001).
The majority of the patients (88%) were right-handed, and twenty-seven (18%) were smokers at the time of injury. Twenty-two (15%) had at least one concomitant medical disorder, with cardiovascular (7%) and metabolic (5%) diseases being the most common. The estimated baseline DASH score was 0 points for 108 patients (72%) and <5 points for an additional twenty-two patients (15%). The overall mean estimated baseline DASH score was 2 points.
Eighty-four patients (56%) were employed outside the home at the time of the injury. Of the sixty-three who were considered to have been "not working," thirty-four were retired, twenty were homemakers, five were students, and four were unemployed. The employment status of the remaining three patients was unknown.
Eighty-six (57%) of the fractures occurred from a fall from a standing height. Eighty-two (55%) of the patients injured the dominant hand. Forty-five fractures (30%) occurred in the home, thirty-eight (25%) occurred in a traffic collision, thirty-four (23%) were associated with sports or leisure activities, sixteen (11%) were work-related, fourteen (9%) were from other causes, and the cause of three (2%) was not documented. Men sustained the fracture more often in traffic collisions or work-related accidents, whereas forty-one of the forty-five injuries that occurred at home were in women. Almost half of the sports and leisure-related injuries were considered to be high-energy.
According to the Müller-AO Comprehensive Classification, 18% of the fractures were Type A, 11% were Type B, and 71% were Type C. The ulna was involved in 48% of the fractures, and the majority of the ulnar lesions were reported to be a distal radioulnar dislocation. Three of the fractures were open. Nineteen patients had a concomitant injury; seven of these injuries involved the contralateral limb, six involved the lower limb, and six involved another site in the body.
Operative Data
The delay between the injury and the surgery ranged from zero to ten days, with a mean of five days. Forty fractures (27%) were treated operatively within forty-eight hours after the injury.
The surgical approach was left to the discretion of the treating surgeon, but all utilized one or more 2.4-mm Synthes implants. The surgical approach was volar in 113 patients (75%), dorsal in twenty-three (15%), and combined dorsal and volar in fourteen (9%). Eleven (7%) of the patients had plate fixation of the radial column with either a dorsal or a volar implant. The 2.4-mm implant system comprises low-profile titanium-alloy plates and screws available in a variety of shapes to permit implant placement to match the individual fracture components. The design features include a low-profile plate with rounded edges designed to minimize the potential for tendon and soft-tissue irritation. Plates for dorsal application include L, T, and oblique distal limb designs as well as a straight implant that is precontoured to fit along the radial column. The volar plates have two separate shapes, one designed specifically for more distal articular fractures and the other designed for extra-articular dorsally displaced fractures. All of the implants feature locking-screw fixation distally and "combination" holes proximally that can accept either locking or conventional screws (Figs. 1-A and 1-B).
A median of four (range, one to seven) locking-head screws were used in the distal portion of the implant, and a median of two (range, zero to five) locking-head screws were used in the proximal portion. The number of locking-head screws used increased proportionately with the complexity of the fracture, with 15% and 11% of the screws used for Type-A and B fractures, respectively, being of the locking-head variety, in contrast to 74% of the screws used for the Type-C fractures.
Fifteen fractures (10%) were supplemented with Kirschner wire fixation, and four others were treated with temporary external fixation. In eighteen patients (12%), the fixation was supplemented with bone graft or a bone substitute during the procedure. The majority of the fractures were treated by surgeons who had utilized these implants to stabilize at least ten fractures. Only one fracture was treated by an unsupervised junior assistant.
Seventy-two fractures were associated with a distal ulnar fracture, but only fifteen (21%) of these ulnar fractures were treated with internal fixation, primarily with Kirschner wires. Nine patients (6%) underwent carpal tunnel release at the time of the fracture fixation.
The mean duration of the operative procedure was eighty minutes (range, twenty to 180 minutes), with a mean fluoroscopy time of two minutes (range, zero to fifty minutes). As rated with the 0 to 10-point visual analogue scale, with 0 indicating dissatisfied and ten indicating highly satisfied, the overall surgeon satisfaction with the initial treatment, including the quality of the fracture reduction, the ease of implant application, the stability of fracture fixation achieved, and the postreduction radiographic appearance, ranged from 2 to 10 points, with a mean of 9 points.
Data Management
A set of do-files was created to label variables, implement variable transformations, calculate scores such as the SF-36, and provide consistency checks. Data inconsistencies were verified in the clinical journals by checking for erroneous data entry or directly with respective clinics, and any necessary corrections were made in the database.
All baseline and follow-up parameters were described with use of standard descriptive statistics. Continuous variables were described with use of means, standard deviations, and ranges, whereas categorical variables were tabulated as absolute and relative frequencies. Time-to-event (e.g., return-to-work) data were assessed with Kaplan-Meier analysis.
Changes in continuous outcomes between follow-up investigations were evaluated with use of repeated-measures analysis of variance, followed by a series of paired t tests when appropriate. Changes in categorical outcomes were analyzed with use of the symmetry test. Differences in continuous outcomes at each follow-up time according to categorical factors were examined with use of the Kruskal-Wallis test. P values of =0.05 were considered significant.
Source of Funding
Support was provided by the AO Foundation.
Of the 150 patients recruited for the study, 125 (83%) were available for follow-up at six months; 121 (81%), at one year; and 117 (78%), at two years. Two patients had died of unrelated causes (one at six months and one at one year after the surgery), and thirty-one could not be contacted or were unwilling to return for final follow-up. An auxiliary comparison of patients who returned for final follow-up with those who were lost at two years showed no significant difference between the populations in terms of their baseline demographic or fracture characteristics.
Pain at rest as rated by the patients on the 0 to 10-point visual analogue scale showed a significant decrease between baseline (i.e., the period including the first, second, and third postoperative days) and six weeks as well as between those two time points and the follow-up examinations at six months, one year, and two years (repeated-measures analysis of variance, p < 0.0001). It is of note that there was also a significant decrease in patient-perceived pain between one and two years postoperatively (p = 0.004) (Fig. 2).
There was no difference between the patients' ratings of their pain during activities between six months and one year; however, the mean pain rating at the two-year follow-up evaluation was significantly lower than the ratings at the six-month and one-year evaluations (p = 0.001 and 0.02, respectively) (Fig. 2).
The mean DASH score changed from a preinjury baseline of 2 points to 8 points at one year and 7 points at two years (p < 0.0001) (Fig. 3).
The combined SF-36 physical and mental health composite scores obtained at the one-year follow-up evaluations were not significantly different from those obtained at two years (p = 0.21).
The mean arc of motion (combined extension and flexion) of the contralateral wrist was 137°, whereas the mean arc of motion of the injured wrist improved from 116° at six months to 127° at one year with no further improvement at the two-year follow-up evaluation (Fig. 4). The mean arc of forearm rotation on the contralateral side was 174°, whereas the mean arc of forearm rotation on the injured side improved from 163° at six months to 170° at one year without further improvement over time (Fig. 5). Overall, consistently significant improvements of both wrist and forearm motion were observed between six months and one year (p = 0.001 for the wrist and p = 0.024 for the forearm) but not between one and two years (p = 0.41 for the wrist and p = 0.16 for the forearm).
There was also a significant improvement in the mean grip strength between six months and one year (from 25.4 kg to 29.5 kg; p < 0.0001) but not between the one and two-year follow-up periods (29.5 and 29.9 kg, respectively; p = 0.95).
Eighty-four patients were working prior to the injury. Seventy-six patients returned to work postoperatively, in a median of fifty-four days (range, one to 348 days). With the numbers studied, there was no significant difference in the rate of return to work among fracture types A, B, and C (p = 0.053).
The mean Gartland and Werley score improved significantly from 4 points at six months to 2 points at two years (p = 0.003). The Gartland and Werley score was categorized as good or excellent for 88%, 90%, and 96% of the patients at six months, one year, and two years, respectively. Overall, a significant improvement was observed only between the one and two-year follow-up examinations (p = 0.014) (Fig. 6).
With the numbers studied, neither an ulnar lesion nor the number of plates used appeared to have an impact on the overall outcome.
Radiographic Results
Radiographs were evaluated for 128, 118, 107, and ninety-six of the 150 patients at six weeks, six months, one year, and two years, respectively. The quality of the radiographs ranged from very good to bad. At the two-year follow-up evaluation, 84% of the anteroposterior radiographs but only 61% of the lateral radiographs were considered to be acceptable.
Thirty-seven fractures (25%) showed a mean gap of 1.2 mm on the initial postoperative radiograph. This number decreased to three (3%) at one year and to none at two years.
Of the 102 intra-articular (AO Type-C) fractures evaluated postoperatively, seventy-five were recorded as having no articular step-off, twenty had a step-off of =2 mm, and the remaining seven had a step-off of >2 mm. Of the seventy-one intra-articular fractures seen at two years, sixty-one had no step-off, five had a step-off of =2 mm, and five had a step-off of >2 mm (Fig. 7).
There was no significant change in radial length, radial angle, or palmar tilt from the immediate postoperative evaluation to the final evaluation at two years. The mean radial length remained at 10 mm; the mean radial angle, at 23°; and the mean palmar tilt, at 88°.
In the series as a whole, there was a minimal change in ulnar variance from a mean of 1.2 mm at six weeks to a mean of 1.5 mm at two years. Of note, however, was a change in the ulnar variance from a mean of 1.4 mm at six weeks to a mean of 2.3 mm at two years (p < 0.0001) in the group with an extra-articular fracture (AO Type A). Only nineteen patients with an extra-articular fracture were seen at two years, and the ulnar variance measurements in those patients ranged from -0.9 to 6.7 mm.
By two years, 27% (twenty-nine) of 108 patients had an increase of at least one grade in radiographic signs of arthritis after having shown no radiographic signs of arthritis initially (Fig. 8-A). Fractures classified as Type B3, C2, or C3 had a greater tendency to be associated with an increase of one grade than did all other fracture types (Fig. 8-B).
Complications
Twenty-eight complications were reported within the two-year follow-up period (Table I). There were eight major complications that required an unexpected return to the operating room, were due to infection, or required secondary reconstructive surgery. There were twenty "minor" complications, including those that were either anticipated or did not interfere with postoperative care or the later outcome.
There were two deaths, six and twelve months after treatment, and they were related to patient factors and not to the implant or surgery. Two tendon ruptures occurred; one involved the extensor pollicis longus and was the result of a prominent screw tip from a volar implant, and one involved the flexor pollicis longus and was due to attrition from the radial edge of the distal limb of the plate. All were successfully treated with a tendon transfer. Four complications involved internal fixation with a volar 2.4-mm implant. One elderly woman had a loss of reduction with recurrence of dorsal angulation due to excessive force exerted on the operatively treated wrist within two weeks after the surgery. A corrective osteotomy was subsequently required. In another patient, the fracture healed with a 30° dorsally directed malunion. Two patients had implant loosening; a screw loosened in one patient, requiring secondary surgery for removal, and the entire plate and screw construct loosened in the other, requiring an additional surgical procedure.
Of the twenty complications that were considered to be minor, nine involved tendon inflammation. All were associated with a volarly placed implant, and implant removal was required in five patients. Carpal tunnel syndrome was noted in three patients, two patients had other nerve-related problems, and two others had findings suggestive of complex regional pain syndrome. Two additional patients had irritation secondary to a prominent Kirschner wire, and a dorsal ganglion developed in one patient.
The estimated risk for experiencing at least one local complication was 16% (twenty-four of 150; 95% confidence interval, 10.5% to 22.9%); two patients had two local complications each. The complication was most likely related to the internal fixation with plate and screws in fifteen of the twenty-six patients who experienced at least one complication.
With the advent of implants designed specifically for the anatomic features of the distal part of the radius and featuring so-called locking-head screws, operative fixation of unstable distal radial fractures has become increasingly widespread. In this study of patients treated with open reduction and stable plate-and-screw fixation with Synthes 2.4-mm implants, both the physician and patient-rated outcome measures continued to improve for one year after the surgery, with little or no change at the two-year follow-up evaluation. In a prospective study of eighty-seven fractures treated with volar locking implants, Chung et al.37 also observed improvement in most parameters over a twelve-month period of follow-up. Although the mean DASH score at the two-year follow-up evaluation in our study did not return to the original estimated baseline score, it must be emphasized that the mean final score of 7 points at two years reflects a high degree of patient satisfaction with minimal symptoms, and this small residual difference may be unimportant from a clinical standpoint. In a prospective study of forty-one patients treated with volar locking implants, Rozental and Blazar8 found a somewhat higher mean DASH score (14 points) at an average of seventeen months postoperatively, with an average DASH score of 12 points for patients without complications and 22 points for those with complications.
The development of the Synthes 2.4-mm implant system was based on the concept of articular fracture patterns involving specific anatomic regions of the distal part of the radius and being amenable to internal fixation with smaller, anatomically precontoured implants. While the vast majority of the fractures treated in this study were stabilized with a single volar implant, the small 2.4-mm dorsal and radial column plates proved to be extremely useful for treatment of articular fractures that required additional stabilization. In a biomechanical study, Blythe et al. compared 2.4-mm Synthes volar locking plates with smaller dorsal locking plates applied with a fragment-specific method38. While the different constructs each provided adequate stability under physiologic loads, the dorsal fixed-angle constructs proved to be stiffer and stronger. The clinical utility of fragment-specific fixation has been similarly observed in several prior studies of implants and techniques that were designed specifically for particular anatomic features of the fracture pattern11,18. In a study of eighty-one patients with a total of eighty-five intra-articular distal radial fractures, Benson et al. reported sixty-one excellent and twenty-four good results at a mean of thirty-two months after treatment with a fragment-specific approach39.
We believe that the complication rate in our study, although not unimportant, is representative of what might be anticipated with the operative treatment of unstable distal radial fractures. Tendon inflammation was the most common complication, found in nine patients, all of whom had isolated inflammation of the flexor tendons over the distal edge of a volar plate. Tendon rupture was seen in two patients. Both tendon ruptures were the result of attrition, caused by either a prominent screw tip or the protruding edge of a volar implant.
All of the implants used in the study were made of polished titanium alloy. There has been some concern that titanium may incite tendon irritation, although other authors have contested this40,41. The small plate dimensions combined with the polished titanium metallurgy, smooth rounded edges, and screw heads that are recessed into the plates seem to have minimized the risk of tendon inflammation with the dorsally applied implants.
The tendon irritation seen in association with some volar implants in this study may have been the result of these implants resting too distally on the volar cortical ridge of the distal part of the radius and presenting a prominent edge over which the flexor tendons moved. Newer volar plate designs that reside more proximally and limit this source of complications have been added to the 2.4-mm system.
The original 2.4-mm volar plate that was used in the study directed the distal locking-head screws in a proximal direction and not just below the subchondral bone, resulting in some early loss of radial height and a residual positive ulnar variance. In a series of fifty fractures treated with a volar locking plate, Drobetz and Kutscha-Lissberg observed a loss of up to 3 mm in radial length in seventeen cases and a loss of 4 mm in four cases at an average of twenty-six months42. They suggested that these losses were the result of the distal locking-head screws being placed too proximal in relation to the subchondral bone.
One limitation of this study is the approximate 20% loss to follow-up. A second limitation is that the operations were performed by several surgeons who used a variety of implant shapes and locations and the patients were evaluated by a variety of individuals throughout the course of the study. Care was taken to evaluate carefully all data points, and those at extreme variance were all reassessed. Despite these limitations, we believe that the information generated from this study provides a reasonable view of the expected outcomes of operative treatment of distal radial fractures with the 2.4-mm Synthes locking plate. 