Intra-articular fractures of the distal end of the radius that are characterized by volar displacement of part of the articular surface are uncommon and tend to be associated with high-velocity trauma5,8,16,30.
Although the findings of a number of studies have supported the effectiveness of operative treatment of such fractures6,7,9,11,17,22,25,31, few reports have described the morphology of the fracture in detail. Instead, many authors simply used the eponymic descriptions Barton or reversed Barton fracture1,2,5,6,19,20,23,28,29.
The purpose of the present retrospective study was to evaluate the long-term functional and radiographic results of open reduction and internal fixation of such fractures. A number of variables, such as the age of the patient, the fracture pattern, and the interval between the injury and the operation were evaluated statistically with regard to their influence on the outcome.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Orthopaedic Hand Service, Massachusetts General Hospital, WAC-527, Boston, Massachusetts 02114.
‡Orthopaedische Chirurgie F. M. H., Mittelstrasse 54, 3012 Bern, Switzerland.
§Oberartz, Kantonspital, Aarau, Switzerland.
Fifty-two volar marginal intra-articular fractures of the distal end of the radius were treated operatively by the senior two of us (J. B. J. and D. L. F.) at Massachusetts General Hospital in Boston and Kantonspital in Aarau, Switzerland, from 1982 to 1993. Forty-nine patients (forty-nine fractures) were available for follow-up, and they form the basis of the present retrospective study. Operative treatment was indicated for volar intra-articular fractures that were associated with at least two millimeters of articular displacement; those for which previous non-operative or operative treatment had failed; and those associated with carpal subluxation, trauma involving the ipsilateral upper extremity, or soft-tissue injury. Similar fractures that had been treated operatively at our institutions but not by the senior two of us, as well as those that were associated with a devascularizing injury or a complete amputation, were excluded from the study.
Twenty-nine patients were male and twenty were female; the average age was forty-three years (range, sixteen to seventy-six years). Thirty fractures involved the right wrist and nineteen, the left; thirty fractures involved the dominant extremity. At the time of the injury, eighteen patients were white-collar workers, eighteen were employed at manual labor, seven were homemakers, five were retired, and one was a student (Table I).
Twenty patients had sustained the fracture in a motorcycle or moped accident; twelve, in a fall from a standing height; ten, in a bicycle accident; four, in a sports-related incident; and three, in an automobile accident. The average age of the patients who had been injured in a motorcycle or moped accident was thirty-two years, compared with fifty years for those who had been injured in a bicycle accident and fifty-five years for those who had been injured in a fall (Table I).
Eight patients had polytrauma, and five others had additional trauma involving the ipsilateral upper extremity. Eight patients had an open fracture, and all of these injuries were classified as grade I according to the criteria described by Gustilo and Anderson14.
The fractures were classified according to the Comprehensive Classification of Fractures21 by the senior two of us solely on the basis of the operative findings. All of the fractures were type-B3 injuries; that is, the volar aspect of the distal articular surface of the radius was involved while the dorsal aspect remained intact. The injuries were further subdivided into two B3.1 fractures (characterized by a small volar fragment, with the sigmoid notch intact), three B3.2 fractures (characterized by a large volar fragment that included the sigmoid notch), and forty-four B3.3 fractures (characterized by comminution of the volar fragment) (Fig. 1). In twenty-five of the B3.3 fractures, the volar fragment was split into two components; in the other nineteen, the volar fragment and the metaphysis were split into multiple components.
Twenty-four patients had an associated fracture of the ulnar styloid process, and three had radiographic evidence of alteration of the normal relationship between the scaphoid and lunate bones.
Thirty-five patients were managed with primary open reduction and internal fixation (Figs. 2-A, 2-B, 2-C, 2-D, 2-E through 2-F). Three of these patients (Cases 10, 30, and 46) initially had been managed at another institution and had been referred to us because of postoperative displacement of the fracture. Nine other patients initially had been managed with closed reduction and immobilization in a plaster cast and had been referred to us because of displacement of the fracture while the limb was in the cast. The remaining five patients had been managed with provisional stabilization of the fracture in a plaster splint without definitive reduction because of associated life-threatening injuries that required emergency treatment.
Sixteen patients were managed operatively on the day of the injury; fourteen, within two days after the injury; nine, within one week after the injury; seven, within two weeks after the injury; and three, more than two weeks after the injury. General anesthesia was used for twenty-five patients and a brachial plexus block, for the remaining twenty-four. Each procedure was performed, under pneumatic tourniquet control, through the distal limb of the Henry approach15. Internal fixation was achieved with a 3.5-millimeter T-plate for forty-seven patients, with additional screws or Kirschner wire being used to secure osseous fragments in eleven. Two of the forty-seven patients were managed with adjunctive fixation with an external device that extended from the second metacarpal to the distal part of the radial diaphysis. In both of these patients, the external fixation device was used intraoperatively to provide longitudinal traction to facilitate the reduction of a comminuted fracture and was left in place because of associated soft-tissue swelling. The fracture in one other patient (Case 12) was secured with screws alone. The remaining patient (Case 46), who had been referred to us after previous operative fixation had failed, had two separate fragments that were secured with two 2.7-millimeter straight plates. Autogenous iliac-crest bone graft was added to the site of the fracture in five patients (Cases 3, 4, 14, 16, and 37) to help to support the reduction of impacted articular fragments.
The average duration of the operation was eighty-three minutes (range, thirty-five to 175 minutes). Postoperatively, the wrist was supported in a volar splint for an average of three weeks (range, zero to six weeks). Active motion of the wrist and forearm was permitted after removal of the splint, and the patient was allowed to use the hand and wrist for activities of daily living by six weeks after the operation.
The reduction of the fracture was assessed on the immediate postoperative anteroposterior and lateral radiographs by the surgeon who had performed the operation. The radio-ulnar index (representing the distance between the distalmost aspect of the sigmoid notch of the radius and the distalmost aspect of the ulnar head) and the accuracy of the realignment of the articular surface in the frontal and sagittal planes were measured in millimeters. The angulation of the articular surface of the distal part of the radius in the frontal plane, termed ulnar angulation, as well as that in the sagittal plane, termed volar angulation, were measured in degrees.
The patients were followed at regular intervals by the senior two of us. At the time of the most recent follow-up examination, they were asked to rate the pain as absent, mild (occasional pain that did not interfere with activities), moderate (pain that was noted only during exertional activities), or severe (pain at rest or that interfered with activities of daily living). The occupational status of each patient was recorded, and the influence of pain and function of the wrist on the ability of the patient to perform occupational and leisure activities was noted.
Active pronation and supination of the forearm as well as active extension, flexion, and radial and ulnar deviation of the wrist were measured with use of a hand-held goniometer, and the measurements were compared with those for the contralateral forearm and wrist. Grip strength was measured with use of a Jamar dynamometer (Asimow Engineering, Los Angeles, California) and was recorded as a percentage of that of the contralateral, uninjured hand. Sensibility in the hand was determined by light touch as well as two-point discrimination. Finally, the presence of a hypertrophic operative scar or visible deformity about the wrist was documented.
Anteroposterior and lateral radiographs of both the injured and the contralateral wrist were made at the time of the most recent follow-up examination and were compared with those made immediately postoperatively. The radio-ulnar index, articular congruity, and ulnar angulation of the distal part of the radius in the frontal plane, and the volar angulation of the articular surface in the sagittal plane, were measured by two of us (C.-L. T. and T. F.). Radiographic changes consistent with post-traumatic osteoarthrosis were evaluated according to the system described by Knirk and Jupiter18, in which grade 0 indicates no evidence of osteoarthrosis; grade I, slight narrowing of the joint space; grade II, marked narrowing of the joint space and formation of osteophytes; and grade III, complete loss of the joint space with bone-to-bone contact as well as formation of marginal osteophytes and subchondral cysts.
The subjective and objective criteria were incorporated into an over-all assessment of the outcome with use of the wrist-scoring system of Gartland and Werley12 as well as the system of Green and O'Brien as modified by Cooney et al.4.
The association of dichotomous variables (such as the presence or absence of an associated fracture of the ulnar styloid process) with a fair or poor outcome as opposed to an excellent or good outcome was evaluated with use of the Fisher exact test (Number Cruncher Statistical System, J. L. Hintze, Kaysville, Utah). The association of continuous variables (such as age and ulnar angulation) with a fair or poor outcome as opposed to an excellent or good outcome was evaluated with use of the Mann-Whitney test (Number Cruncher Statistical System). A p value of 0.05 or less (that is, a probability of 5 per cent or less that the observed association was due to chance) was considered significant. All of the tests were two-tailed.
The immediate postoperative radiographs were evaluated by the surgeon who had performed the operation. The radio-ulnar index, as measured in millimeters, was 0 for thirty-two patients, +1 for three, +2 for two, -1 for six, -2 for three, and -3 for three. (A positive value indicated that the ulna was longer than the radius, and a negative value, that the radius was longer than the ulna). Articular incongruity was measured as the maximum incongruity in either the sagittal or the frontal plane. In thirty patients, the articular reconstruction was considered to be anatomical (that is, there was no incongruity in either plane) as seen on the immediate postoperative radiographs. Of the remaining nineteen patients, twelve had a one-millimeter gap, six had a 1.5 or two-millimeter gap, and one had a three-millimeter gap.
The average ulnar angulation was 22 degrees (range, 0 to 32 degrees), and the average volar angulation was 7 degrees (range, -9 to 16 degrees). The normal volar angulation of the articular surface was lost in three patients, in whom the volar angle ranged from -4 to -9 degrees.
Six complications occurred in the early postoperative period, before the fracture had healed. One patient (Case 32) had deep venous thrombosis of the lower extremity and was managed with anticoagulation therapy, three patients (Cases 14, 30, and 31) had signs and symptoms of sympathetic maintained pain and were managed with sympathetic blockade and physical therapy for the hand, one patient (Case 21) had a loss of reduction and was managed with a shortening osteotomy of the ulna, and one patient (Case 30) had scapholunate dissociation and was managed with closed reduction and percutaneous pinning. Fourteen late complications, all of which necessitated an additional operative procedure, developed in twelve patients. Five patients (Cases 1,2,8,11, and 27) had pain over the operative incision or the implant and were managed with removal of the plate, three patients (Cases 16, 18, and 19) had median-nerve compression in the carpal canal and were managed with release of that nerve, two patients had tenosynovitis involving the flexor carpi radialis tendon (Case 27) or tendons in the first dorsal extensor compartment (Case 7) and were managed with tenosynovectomy, two patients had subluxation of the distal radio-ulnar joint and were managed with the procedure described by Sauvé and Kapandji24 (Case 31) or a hemiresection arthroplasty (Case 25), one patient (Case 30) had severe post-traumatic osteoarthrosis involving the radiolunate articulation and was managed with an arthrodesis of that joint, and one patient (Case 16) had spontaneous rupture of the extensor pollicis longus tendon and was managed with transfer of the extensor indicis proprius tendon.
After an average duration of follow-up of fifty-one months (range, twenty-four to 117 months), twenty-eight patients stated that they had no pain in the involved wrist, eleven had mild pain, seven had moderate pain, and three (Cases 3, 29, and 30) had severe pain (Table II). Thirty-one patients had returned to their previous occupation without restrictions; ten had retired either before the injury or for reasons that were unrelated to the injury; four had changed jobs for reasons that were unrelated to the injury; three had changed jobs because of problems related to other injuries that had occurred simultaneously with the radial fracture; and only one (Case 28), who had sustained a multi-fragmented fracture, had changed from full-time to part-time employment because of sequelae of the radial fracture. Thirty-nine patients were able to perform leisure activities without any restrictions, six had occasional discomfort with such activities, and four (Cases 14, 28, 32, and 33) had substantial discomfort with such activities.
Active pronation of the involved forearm averaged 72 degrees (range, 45 to 90 degrees) and active supination averaged 78 degrees (range, 30 to 110 degrees), compared with 71 and 82 degrees, respectively, for the contralateral forearm.
Active extension of the involved wrist averaged 58 degrees (range, 20 to 90 degrees) and active flexion averaged 66 degrees (range, 20 to 90 degrees), compared with 64 and 75 degrees, respectively, for the contralateral wrist.
Active radial deviation of the involved wrist averaged 22 degrees (range, 5 to 40 degrees) and active ulnar deviation averaged 42 degrees (range, 25 to 60 degrees), compared with 23 and 46 degrees, respectively, for the contralateral wrist.
Thirty-five patients had grip strength that was equal to that on the contralateral side; twelve, within 10 per cent of that on the contralateral side; and two, within 25 per cent of that on the contralateral side.
Forty-two patients had an operative scar that was unremarkable in appearance, and seven had a hypertrophic scar. Four patients had residual sensitivity to touch over the scar.
Follow-up radiographs revealed that five patients had reversal of the normal volar angulation of the radial articular surface; the volar angle in these five patients ranged from -1 to -6 degrees. The other forty-four patients had an average volar angulation of 9 degrees (range, 0 to 16 degrees) (Table III).
The radio-ulnar index, as measured in millimeters, averaged 0.04 (range, +3 to -3). In ten patients the ulna was longer than the radius, and in nine the distal end of the ulna was proximal to the sigmoid notch of the radius. Ulnar angulation averaged 22 degrees (range, 0 to 32 degrees).
The distal articular surface of the radius was congruent in thirty-five patients, whereas anteroposterior radiographs revealed a one-millimeter articular step-off in ten patients and a two-millimeter gap in four. Forty patients had no radiographic evidence of radiocarpal osteoarthrosis, eight had grade-I osteoarthrosis, and one had grade-II osteoarthrosis.
There were thirty-one excellent, ten good, and eight fair results according to the system described by Gartland and Werley12, and there were thirty-two excellent, nine good, five fair, and three poor results according to the modified system4 of Green and O'Brien (Table II).
When the ratings of both scoring systems were combined, nine patients (Cases 3, 8, 21, 25, 29, 30, 35, 37, and 39) had either a fair or a poor over-all result. All but one of these patients had had a B3.3 fracture, and four (Cases 3, 25, 30, and 39) had had a multifragmented fracture. In addition, five of these patients (Cases 3, 8, 21, 29, and 30) had radiographic evidence of osteoarthrosis, three (Cases 3, 8, and 35) had loss of normal volar angulation of the articular surface of the radius, and three (Cases 3, 29, and 37) had residual articular incongruity.
There was no significant difference, between the patients who had an excellent or good result and those who had a fair or poor result, with regard to age (p = 0.588), the interval from the injury to the operation (p = 0.846), a concommitant injury of the ipsilateral upper extremity (p = 1.00), the presence of an associated fracture of the ulnar styloid process (p = 0.289), articular incongruity at the time of the most recent follow-up examination (p = 0.701), the radio-ulnar index (p = 0.969), or ulnar angulation (p = 0.327). Neither the B3.3 fractures in which the volar fragment was split into two pieces (p = 1.00) nor those with multiple fragments (p = 0.701) were significantly more likely to be associated with a fair or poor outcome.
Two factors were associated with a fair or poor outcome: evidence of osteoarohrosis on the most recent follow-up radiographs (p = 0.006) and reversal of the normal volar tilt of the distal end of the radius. Three of the five patients who had a reversal of the normal volar tilt had a fair or poor outcome (p = 0.037, Fisher exact test). The average volar angle among patients who had an excellent or good result was 8.4 degrees, compared with 3.4 degrees among those who had a fair or poor result (p = 0.041, Mann-Whitney test).
The persistent use of eponyms (such as reversed Barton, Barton2, Smith25, or Goyrand13 fracture) and imprecise systems of classification (such as that described by Thomas28) has inhibited a more accurate understanding of the morphology of volar marginal intra-articular fractures of the distal end of the radius. This, in turn, has made accurate interpretation of the results of treatment of such fractures exceedingly difficult1,3,5,6,10,11,19,20,22,26-32.
Appreciation of the likelihood that a displaced marginal intra-articular fracture comprises at least two fragments will help the surgeon to plan both the operative exposure and the placement of internal fixation devices. Although we used the distal limb of the Henry approach15 routinely in the present series, it became evident that, in some patients, the operative exposure of the fracture would have been easier with use of a more ulna-based approach8. It is likely that, in some patients, excessive retraction of the median nerve led to the development of a sympathetic maintained pain syndrome or to residual hyperesthesia in the distribution of the median nerve.
The multifragmented nature of the articular surface necessitated the use of additional implants to supplement the volar T plate in eleven patients. The surgeon should be prepared to use an additional Kirschner wire or screw to secure a volarly displaced fragment consisting of the styloid process or the lunate facet.
Although a number of authors have reported excellent results after the operative treatment of these fractures1,5-7,11,30, our experience as well as that of Keating et al.17 suggest that operative intervention may not be suitable for all patients. We identified six early complications (two of which necessitated operative intervention) as well as fourteen late complications (all of which necessitated operative intervention), and we noted residual articular incongruity on the follow-up radiographs of fourteen patients (29 per cent). It is noteworthy that a substantial prevalence of malpositioning of the fracture (as seen on follow-up radiographs) also was described by Keating et al. The fact that five of the fractures in the present study healed with reversal of the normal volar angulation of the articular surface of the radius suggests that these fractures may have been classified incorrectly and should have been classified as type-C injuries with involvement of the metaphyseal and dorsal osseous column. Although this is possible, there was no evidence of such involvement either on preoperative radiographs or intraoperatively.
The present study was not designed to validate the accuracy of the Comprehensive Classification of Fractures21 with regard to the identification of the morphology of fractures on the basis of preoperative radiographs. Rather, the fracture patterns that were identified by direct intraoperative observations suggest that this classification system may be useful for the accurate documentation of the variable fracture patterns that may necessitate an operation.