 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1999 The Journal of Bone and Joint Surgery, Inc. Results of Herbert-Screw Fixation with Bone-Grafting for the Treatment of Nonunion of the Scaphoid*  , ST. JOHN'S, NEWFOUNDLAND, CANADAInvestigation performed at The Health Sciences Centre and St. Clare's Mercy Hospital, St. John's
We examined twenty-one patients in whom an established nonunion of the scaphoid had been treated with Herbert-screw fixation with bone-grafting; our purpose was to determine the long-term morbidity of the procedure (at an average of forty-eight months; range, seventeen to fifty-five months). According to the modified scaphoid outcome scoring system, eighteen patients had a good or excellent result. No patient had failure of the hardware. Radial deviation was significantly decreased (an average of 17 degrees; 55 percent) compared with that of the contralateral wrist (p = 0.02). The other ranges of motion and the grip strength were also decreased, although the difference was not found to be significant with the numbers available. Three patients had a persistent nonunion, and two of them also had loosening of the hardware. Three patients had a tender hypertrophic scar. On the basis of our data, we concluded that Herbert-screw fixation with bone-grafting is effective for the treatment of nonunion of the scaphoid.
Scaphoid fractures are common injuries that occur as a result of forced dorsiflexion. These fractures may be overlooked both radiographically and clinically as their presentation after a low-energy injury may be subtle. Radford et al. suggested that a delay between the time of the fracture and the operation was the major factor predisposing to a postoperative nonunion especially when the fracture involved the proximal pole13. Robbins et al. stated that one of three acute fractures has soft-tissue interposition, which may eventually lead to a nonunion16. It is known that 5 to 10 percent of scaphoid fractures fail to unite after nonoperative treatment4,7,10. The nonunions may be due to a delay in treatment, inadequate immobilization, displacement of fragments, or instability caused by a concurrent ligamentous injury10,12,16. There is general agreement that a symptomatic nonunion of the scaphoid should be treated in order to improve function of the wrist and hand and to prevent long-term osteoarthritis. Fixation with a Herbert screw and bone-grafting has been considered quite effective for this purpose. Herbert and Fisher reported on the first thirteen nonunions that had been treated with Herbert-screw fixation alone; none united5. Those authors recommended that bone-grafting be used for the treatment of all nonunions. Warren-Smith and Barton found that Herbert-screw fixation with bone-grafting yielded a high rate of success; eighteen of twenty-two patients who had had a pseudarthrosis of the scaphoid had osseous union, and the functional results were better than those obtained with bone-grafting alone18. The purpose of the current study was to analyze the long-term results (average, forty-eight months; range, seventeen to fifty-five months) of Herbert-screw fixation with bone-grafting for the treatment of nonunion of the scaphoid. We know of no other study on this subject.
Forty patients in whom a fracture of the scaphoid had been treated with Herbert-screw fixation by the same surgeon (D. S. S.) were identified. The criteria for inclusion in the study were: (1) an established nonunion of the scaphoid, (2) an interval of at least seventeen months since the operation, (3) a medically healthy condition, and (4) an age of more than eighteen years, with a closed epiphysis. Twenty-one patients met these four criteria. The average age of the patients at the time of the operation was thirty years (range, seventeen to fifty-five years). Only one patient was female. The average time from the initial injury to the operation was fifty-one months (range, three to 192 months). Fourteen of the twenty-one patients had a fracture of the right scaphoid, and seven had a fracture of the left scaphoid. Ten of the injuries were on the dominant side. Twelve patients had fallen onto the dorsiflexed wrist, and the remaining nine were unable to recall the exact mechanism of the injury. One patient was receiving Workers' Compensation. The patients answered a questionnaire so that we could determine their perception of the outcome. The subjective results were then classified with use of a modified scaphoid outcome scoring system (Table I)16. A perfect score (10 points) was considered an excellent result; 8 or 9 points, a good result; 6 or 7 points, a fair result; and 5 points or less, a poor result.
The patients then were evaluated objectively with use of a physical examination and radiographs. Dorsiflexion, palmar flexion, radial deviation, and ulnar deviation of both the affected and the unaffected wrist were measured with use of a goniometer, and the average of three measurements for each range of motion was recorded. Grip strength of both the affected and the unaffected hand was measured with use of a Jamar dynamometer (Sammons Preston, Bolingbrook, Illinois) with the elbow flexed 90 degrees and the forearm in neutral rotation; the highest of three measurements was recorded for each hand. The presence or absence of tenderness in the anatomical snuffbox and the scaphoid tubercle also was noted. In addition, the initial treatment was classified as adequate or inadequate according to the criteria of Radford et al.13. Also, clinical healing was assessed according to the criteria of Rettig et al.15. Three standard radiographs (anteroposterior, lateral, and scaphoid with the affected wrist in full ulnar deviation and the forearm in 30 degrees of supination) were analyzed for degenerative changes, carpal collapse, the size of the intrascaphoid angle, failure of the hardware, and union. Our criterion for osseous union was trabeculation across the site of the fracture on all three radiographs. This criterion is quite strict compared with that of Daly et al., who required that at least two of four radiographs show trabeculation across the fracture site2. One of us (B. M. R.) performed and analyzed the results of the subjective and objective assessments. We used the two-sample independent group t test, with a confidence interval of 95 percent, for statistical evaluation of the range-of-motion and grip-strength data.
Operative Technique Postoperatively, all patients wore a cast for six weeks.
According to the criteria of Radford et al.13, fifteen patients were considered to have had adequate treatment, which had included a cast. The remaining six patients were deemed to have been managed inadequately: three, because radiographs had not been made; two, because of neglect on the part of the patient; and one, because the diagnosis of a scaphoid fracture had been overlooked owing to the presence of other injuries. According to the modified scaphoid outcome scoring system, one of the twenty-one patients had an excellent result, seventeen had a good result, three had a fair result, and none had a poor result. Of the three patients who had a fair result, one was receiving Workers' Compensation and had a symptomatic nonunion and the other two had moderate degenerative changes that had originated before the operation. Nineteen patients were satisfied with the long-term result of the operation. Sixteen patients reported good function of the wrist. Ten patients had stiffness of the wrist. Nine patients had pain with motion of the wrist, seven had pain and stiffness with motion of the wrist during damp weather, and one had pain while at rest; this last patient had a persistent nonunion with failure of the graft. Another patient said that he had had to change his job, from that of a truck driver to that of a salesperson, after the reconstruction of the scaphoid; this patient had evidence of osteoarthritis. One patient had migration of the screw into the wrist joint, necessitating removal of the screw. Two patients needed repeat bone-grafting, at twenty-eight and five months after the index procedure, because of a persistent nonunion; both subsequently had a successful clinical and radiographic result. A third patient (the one who was receiving Workers' Compensation) had a symptomatic nonunion. This patient had pain on palpation of the scaphoid tubercle. At the time of writing, he did not want another operation but he was still being followed. Two patients had loosening of the hardware, and both of them had a nonunion as well. Other complications included long-standing pain over the iliac crest in one patient, cellulitis at the site of the wound in two patients, and formation of a tender hypertrophic scar over the wrist incision in three patients. Filan and Herbert reported the formation of a tender hypertrophic scar at the site of the wrist incision in fifty-six (13 percent) of 431 patients3. No patient in our study had failure of the hardware. The arc of ulnar-radial deviation was decreased 40 percent compared with that of the unaffected wrist; however, with the numbers available, this difference was not significant (p = 0.08). Radial deviation was decreased an average of 17 degrees (55 percent) compared with that of the contralateral wrist; this difference was significant (p = 0.02) (Table II). Dorsiflexion and palmar flexion each were decreased an average of 20 degrees (29 percent) compared with the values for the unaffected side; however, with the numbers available, these differences were not found to be significant (p = 0.14).
The average grip strength of the affected hands was forty-two newtons (range, twenty-four to fifty-six newtons) compared with fifty-two newtons (range, thirty-two to sixty-six newtons) for the unaffected hands. This was a 19 percent difference (ten newtons); however, with the numbers available, it was not found to be significant (p = 0.24). Nineteen of the twenty-one intrascaphoid angles, as measured on the anteroposterior and lateral radiographs, were within the normal, accepted range (25 to 45 degrees). The intrascaphoid angle is the angle of bisection between lines drawn along the volar and dorsal aspects of the scaphoid on the lateral radiograph or lines drawn along the radial and ulnar sides on the anteroposterior radiograph. The two patients who had an increased intrascaphoid angle both had union but moderate degenerative changes. Eight patients met all three criteria for clinical healing, as described by Rettig et al.15. Twenty patients met criterion A (an absence of pain on palpation); ten, criterion B (a range of motion within 10 percent of that of the uninjured wrist); and thirteen, criterion C (grip strength within 10 percent of that of the uninjured hand).
Established nonunions of the scaphoid have been treated with a variety of operative procedures. Currently, the so-called gold standard for treatment is considered to be the Russe method of bone-grafting7,17. This technique led to union in sixty-nine (90 percent) of seventy-seven patients7,17. Sixty patients (78 percent) who had had a painful wrist before the operation became symptom-free postoperatively, and sixty-eight patients (88 percent) were satisfied with the result of the operation. However, four to six months was required for healing, inevitably leading to stiffness of the wrist. In addition, the Russe technique alone cannot fully correct an abnormal intrascaphoid angle9. The use of the Herbert screw was initially reported by Herbert and Fisher, who noted that none of thirteen nonunions healed without bone-grafting5. Numerous subsequent reports have documented the early results of Herbert-screw fixation with bone-grafting, with emphasis on rates of union and range of motion4,6,11,12,16. Jiranek et al. reported the results at an average of eleven years after twenty-five patients had been managed with Russe bone-grafting alone because of a scaphoid nonunion9. Their patients had a decreased flexion-extension arc (average, 82 percent [range, 54 to 100 percent] of that on the unaffected side) and decreased grip strength (average, 84 percent [range, 52 to 100 percent] of that on the unaffected side). Bunker et al. reported on fifty scaphoid fractures, thirty-three of which were nonunions that were treated with Herbert-screw fixation or bone-grafting, or both1. At the time of the latest follow-up evaluation, an average of 14.3 months postoperatively, palmar flexion averaged 58 degrees; dorsiflexion, 61 degrees; radial deviation, 20 degrees; and ulnar deviation, 35 degrees. Grip strength was equal to that of the unaffected hand. Warren-Smith and Barton reported the results in a subgroup of twenty-two patients, from a series of fifty patients who had had a nonunion of the scaphoid, who had been managed with Herbert-screw fixation and bone-grafting18. At an average of fourteen months postoperatively, grip strength was nearly normal and the range of motion averaged 65 degrees of flexion and 61 degrees of extension. Filan and Herbert reported the results of open reduction and internal fixation with a Herbert screw that had been performed, over a thirteen-year period, in 431 patients for the treatment of a scaphoid fracture or nonunion3. Although that study was not limited to scaphoid nonunions that were treated with Herbert-screw fixation and bone-grafting, Filan and Herbert thought that this method of treatment provided better functional results than did the standard technique of Russe bone-grafting. They concluded that the Herbert screw can provide enough fixation to allow healing without use of a splint. They also noted that healing of the nonunion provided greater functional recovery and reduced the progression of osteoarthritis. Daly et al. reported the results of Herbert-screw fixation and bone-grafting that had been performed in twenty-six patients, over a fourteen-year period, for the treatment of established nonunion of the scaphoid2. The rate of union was 96 percent (twenty-five patients). The main limitation of that study, other than the relatively short duration of follow-up, was that the definition of union was trabeculation over the fracture line on two of four radiographs. Filan and Herbert noted that a number of authors have not strictly defined osseous union as seen on radiographs3. The finding by Warren-Smith and Barton that radiographic union was not necessarily associated with relief of symptoms or with good function18 may support the use of stricter criteria for radiographic union. The results with regard to range of motion in our series are similar to those reported by Jiranek et al., who used the Russe method of bone-grafting for the treatment of scaphoid nonunions9. Our patients had a relatively greater loss of radial deviation, which probably was due to early degenerative changes on the radial side of the wrist or perhaps was due to scarring associated with the operative approach. Filan and Herbert believed that postoperative stiffness may be due to the operative approach or to synovial adhesions that persist because of nonunion3. Postoperative stiffness may be potentiated by use of a splint and poor physiotherapy. Interestingly, those authors concluded that a Herbert screw alone provides sufficient fixation to allow healing without use of a splint. It is conceivable that the omission of a splint and the implementation of early intensive rehabilitation may improve the range of motion of the wrist. We believe that use of a volar approach may lead to more capsular adhesions than use of a dorsal approach, probably because the former results in more intra-articular disruption from the jig, which may decrease the range of motion. The grip strength of our patients decreased an average of ten newtons (19 percent) compared with that of the unaffected hand. However, approximately half of our patients had the injury on the side of the nondominant hand, which is expected to be weaker than the dominant hand. Inoue and Sakuma demonstrated a directly proportional relationship between the duration of nonunion and decreases in grip strength and range of motion8. Seven of our patients had radiographic evidence of degenerative changes (joint-space narrowing, osteophytes, and subchondral sclerosis) around or involving the scaphoid. All seven had osteoarthritic symptoms such as pain or stiffness, or both, with damp weather or movement. Four of these seven patients had had degenerative changes and osteoarthritic symptoms before the operation; two of the four had had moderate osteoarthritis, and the remaining two had had mild radiographic changes. The other three patients, who had new degenerative changes, had had osteoarthritic symptoms before the operation, which suggests that the degenerative changes were caused by the nonunion and not by the operation. Inoue and Sakuma thought that there was a direct relationship between the duration of nonunion and the presence of osteoarthritic symptoms8. Filan and Herbert agreed with this statement and also believed that the degree of displacement was directly related to osteoarthritic symptoms and changes3. There was no radiographic evidence of carpal collapse in any of our patients. Eighteen (86 percent) of our twenty-one patients had union; this rate increased to twenty (95 percent) of twenty-one after repeat bone-grafting in two of the three patients who initially had had failure of the procedure. At the time of writing, both of these patients had a successful clinical and radiographic result. The rate of union in the current series compares favorably with those reported by Herbert et al.6, by Radford et al.13, and by Warren-Smith and Barton18. We assessed osseous union according to strict guidelines: trabeculation had to be visible across the site of the fracture on all three radiographic projections. Herbert and Fisher noted that factors that often are implicated in persistent postoperative nonunion include failure of the bone graft and incorrect placement of the screw5. Two of our three patients who had a persistent nonunion had loosening of the hardware, probably as a result of increased forces due to a lack of load-sharing. Two of the three patients had inadequate compression at the site of the nonunion, and the third had improper placement of the screw. All three nonunions were in the proximal pole of the scaphoid. The presence of these factors probably meant that these patients were at higher risk for nonunion. Herbert et al. found that fracture-healing was strongly associated with the technical success of the procedure6. Radford et al. encountered major technical problems with use of the Herbert screw in fourteen of fifty consecutive patients who had a scaphoid nonunion13. Strict adherence to technique, including fluoroscopic guidance for placement of the screw, is necessary to obtain a good result. Another advantage of Herbert-screw fixation with bone-grafting is that it promotes early active motion in contrast to Russe's method of fixation, which requires that the patient wear a cast for four to six months. The dorsal approach is preferred when the proximal-pole fragments are less than 25 percent of the size of the scaphoid7. Disadvantages of the dorsal approach are that the compression jig cannot be used to attain rigid compression before insertion of the screw and the vital dorsal blood supply to the scaphoid is in danger. None of our three patients in whom the procedure failed had the indication for a dorsal approach. Robbins et al. stated that there is no major difference, with regard to union, between the dorsal and volar approaches16. In summary, our long-term results strongly support the use of Herbert-screw fixation with bone-grafting to treat established nonunions of the scaphoid. Despite some loss of grip strength and motion (especially radial deviation), almost all of our patients were satisfied with the outcome and had regained normal function. It remains to be seen if the degenerative changes that were noted at this stage will progress after a longer duration of follow-up.
*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.
  CiteULike  Connotea  Del.icio.us  Facebook  Technorati  Twitter What's this?
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |