We performed a retrospective review of the results of operative excision of a proximal radioulnar synostosis, performed by one of us (J. B. J.) from 1987 to 1995, in a consecutive series of seventeen patients (eighteen limbs). The average age of the fifteen men and two women was thirty-seven years (range, nineteen to forty-eight years). All of the patients were skeletally mature, and the osseous synostosis had developed as a result of a musculoskeletal injury. Patients who had a congenital proximal radioulnar synostosis or who had an associated head injury with sufficient residual cognitive or physical limitations, or both, that would preclude them from following postoperative instructions14 were excluded. None of the patients had an associated burn injury. The dominant limb was involved in ten of the sixteen patients who had involvement of one limb. Before the injury, eleven patients had worked as manual laborers, four were employed in a white-collar occupation, and two were homemakers.
The initial injury was a posterior fracture-dislocation of the elbow in eleven limbs, a posterior Monteggia fracture-dislocation in two, a traumatic rupture of the distal biceps tendon in two, an isolated fracture of the radial head in two, and a floating elbow (a concomitant fracture of the humeral diaphysis and the proximal aspects of the radial and ulnar diaphyses) in one. All eleven limbs that had a posterior fracture-dislocation of the elbow had a fracture of the radial head, and six also had a fracture of the proximal aspect of the ulna (the coronoid process or the olecranon process, or both). In addition, one limb that had a posterior fracture-dislocation of the elbow had a concomitant fracture of the humeral diaphysis and another limb that had a posterior fracture-dislocation had a fracture of the distal aspect of the radius (Table I). Two patients had associated polytrauma, and one of them had a closed head injury.
All eighteen limbs had at least one operative procedure for treatment of the initial injury; five had two procedures and four had three procedures. The initial management of three patients was complicated by postoperative deep infection.
Eight patients (eight limbs) were referred to us less than twelve months (average, nine months; range, six to eleven months) after the initial injury and had resection of the synostosis without delay. The average time from the injury to the resection for the seventeen patients was nineteen months (range, six to forty-eight months) (Table I). In each patient, the radiographic appearance of the heterotopic bone (a well defined margin and clear trabeculation) suggested maturation of the process.
On examination, complete absence of active or passive rotation of the forearm was documented. The heterotopic ossification in seven limbs extended to the distal aspect of the humerus, resulting in a complete osseous ankylosis of the ulnohumeral articulation in an average position of 53 degrees (range, 20 to 90 degrees) of flexion of the elbow. The average flexion of the eleven limbs that had retained ulnohumeral motion was 111 degrees (range, 60 to 140 degrees) and the average extension was -26 degrees (range, -60 to 0 degrees), for an average total arc of ulnohumeral motion of 85 degrees (range, 30 to 140 degrees). None of the three patients who had had a postoperative deep infection had active drainage at the time of presentation. All eleven patients who had worked as laborers were unemployed.
The extent of heterotopic ossification was classified with a modification of the system proposed by Vince and Miller, who defined the location as type I (the distal third of the forearm), type II (the middle third), or type III (the proximal third)28. The specific locations of the proximal radioulnar synostoses in the present study were further divided into three subgroups: A indicated a synostosis at or distal to the bicipital tuberosity (Fig. 1-A); B, a synostosis of the radial head and the proximal radioulnar joint (Fig. 1-B); and C, a synostosis contiguous with heterotopic bone extending to the distal aspect of the humerus (Fig. 1-C). Four limbs had a type-IIIA synostosis; seven, type-IIIB; and seven, type-IIIC.
The history and results of the most recent physical examination in the office, as recorded by the operating surgeon, were obtained by review of the medical records. We also reviewed any radiographs made at the most recent follow-up examination.
The results of the index operation were determined with use of the classification system developed by Failla et al. in a previous study evaluating the outcome of operative treatment of post-traumatic proximal radioulnar synostosis11. This system is based on previously reported normative data for the ranges of rotation of the forearm that are needed for fifteen functional tasks22. An excellent result indicates arcs of pronation and supination of 50 degrees or more (a total of 100 degrees or more), which is the amount needed to perform all fifteen tasks without modification of behavior or restriction of activity; a good result indicates more than 30 degrees of pronation and of supination, which is the amount needed for ten of the fifteen tasks; a fair result indicates a total arc of rotation of more than 30 degrees, which is the amount of rotation needed for six tasks; and a poor result indicates a gain of less than 30 degrees of total rotation, which is enough rotation to perform only three of the tasks.
Zero to four years after the most recent follow-up visit at which a history was recorded and a physical examination was performed, eight patients were contacted by mail and seven, by telephone; we were unable to contact two patients. Using an unpublished assessment form developed by the American Shoulder and Elbow Surgeons for the evaluation of problems related to the elbow, the patients rated their pain, their capacity to perform routine tasks, and their satisfaction with the result of the procedure. With 0 points indicating no pain and 10 points indicating the worst pain possible, each patient was asked to rate the pain when it was most severe, at rest, when lifting an object, when performing a task that involved repeated movement of the elbow, and at night. The scores were added together, and the total was subtracted from fifty and divided by two to reflect the overall pain score for each patient. A higher score indicated less pain; the maximum possible score was 25 points. Each patient also rated his or her ability to perform five tasks: buttoning the top button of a shirt, toileting, combing hair, tying shoes, and eating with utensils. A score of 0 points indicated that the patient was unable to perform the task; 1 point, the task was extremely difficulty; 2 points, the task was somewhat difficult; and 3 points, the task was not difficult. The five scores were added together and multiplied by two, for a maximum possible score of 30 points. In addition, the patients rated their satisfaction with the result of the operative procedure on a scale of 0 (not at all satisfied) to 10 points (very satisfied).
Statistical Analysis
The average rotation of the forearm at the most recent follow-up examination was compared between patients who had had a low-energy injury and those who had had a high-energy injury, between patients who had been managed with a free fat graft and those who had not, between patients who had been managed with a hinged elbow distractor and those who had not, and between patients who had had an early resection (less than twelve months after the injury) and those who had had a late resection (twelve months or more after the injury). The average rotation was also compared among the subtypes of proximal radioulnar synostosis. Ninety-five per cent confidence intervals for the average rotation of the forearm are reported. Non-parametric tests were used to calculate the p values because the distributions of average rotation were not Gaussian and the variances were dissimilar. With use of the Number Cruncher Statistical System (J. L. Hintze, Kaysville, Utah), the Mann-Whitney U test was performed for comparison of two groups and the Kruskal-Wallis test was done for comparison of three groups. A p value of 0.05 was considered significant, and no adjustment was made for multiple comparisons.
Operative Technique
All patients had general endotracheal anesthesia and were positioned supine with the involved arm on a hand-table and a pad placed under the ipsilateral scapula. A sterile tourniquet was applied to the arm to achieve a bloodless field.
The specific operative approach was dictated in part by the location of the synostosis, whether there was a contiguous osseous ankylosis extending to the distal aspect of the humerus, whether there was a capsular contracture of the elbow, and the status of the overlying soft-tissue envelope. We used a posterior midline skin incision and elevated skin flaps as needed to gain access to the medial and lateral aspects of the forearm and elbow. Previous posterior incisions were incorporated into the wound, whereas previous lateral and anterior incisions were not.
Type-IIIA and IIIB synostoses were approached by elevation of the extensor carpi ulnaris and the supinator from the ulna, to expose both the synostosis and the proximal third of the radius. For type-IIIC lesions, the ulnar and radial nerves were identified proximal to the elbow and were dissected free of any osseous entrapment. The ulnar nerve was released for at least six centimeters proximal and distal to the cubital tunnel and was left anteriorly transposed in the subcutaneous tissues. The radial nerve was isolated between the brachioradialis and brachialis and was traced beyond its division into the superficial and posterior interosseous branches. The median nerve and the brachial artery were not routinely visualized but remained in a protected position superficial to the brachialis muscle.
The synostosis was defined at its proximal and distal extent and was resected with an oscillating saw, with the surrounding soft tissues protected by small retractors. The extent of the resection was controlled by both direct vision and intraoperative fluoroscopy. The radial head was resected in two limbs. The decision to resect or preserve the radial head was made on the basis of the presence of incongruity of the articular surface, radiocapitellar osteoarthrosis, or disruption or distortion of the proximal radioulnar joint. Bone was resected until full or nearly full passive motion of the elbow and pronation and supination of the forearm could be achieved intraoperatively.
Any bleeding osseous margins on the radius and ulna at the sites of the resection were cauterized and were covered with a thin layer of bone wax. In eight limbs, a small amount of free fat, obtained from either the buttock or the upper limb, was placed between the two bones of the forearm. The use of free fat was discontinued after treatment of the eighth patient, in whom the fat dislodged into the subcutaneous tissue without recurrence of the synostosis.
In seven limbs, excision of the heterotopic bone extended onto the humerus both anteriorly and posteriorly. A concurrent complete anterior and posterior elbow capsulectomy was performed. Three of these patients, as well as three other patients who had associated chronic instability of the ulnohumeral articulation, had application of a hinged elbow distractor (Compass Hinge Distractor; Smith and Nephew Richards, Memphis, Tennessee), which was in place for an average of four weeks (range, three to six weeks) postoperatively. One of these patients also had open reduction and internal fixation of a fracture fragment of the coronoid process, and two had a fascial arthroplasty of the ulnohumeral articulation. An additional four limbs that had a stiff ulnohumeral joint unrelated to heterotopic ossification had an anterior and posterior capsulectomy of the elbow at the time of the resection (Table II).
The pneumatic tourniquet was released, and care was taken to obtain hemostasis before closure of the wound. A suction drain was used in each limb for twenty-four hours postoperatively. No adjuvant non-steroidal anti-inflammatory medication, such as indomethacin, or low-dose radiation was used in any patient during the postoperative period. Although the risk of subsequent malignant change is probably low6,18, our radiation therapy department was reluctant to use radiation therapy prophylactically in these young adults. We decided not to use non-steroidal anti-inflammatory agents mostly because of their potential for gastrointestinal, renal, and hematological complications as well as the need for compliance on the part of the patient15,21,23.
A controlled physical therapy program was begun on the first postoperative day and consisted of active and active-assisted rotation of the forearm. Rotation of the forearm against resistance was begun by four weeks after the excision when a distractor had not been applied. For the six patients who had a hinged elbow distractor, rotation of the forearm was restricted until the device was removed.
Three patients had an early postoperative complication (Table II). In one (Case 7), one of the five-millimeter-long fixation pins that attached the hinged elbow distractor to the humerus broke, at the pin-humerus interface, between the third and fourth postoperative weeks. The frame was removed the next week without residual problems related to the retained pin. In another patient (Case 12), the proximal aspect of the ulna fractured several weeks after the reconstructive procedure; the fracture occurred during a therapy session in which passive mobilization was used against our advice. It was treated with open reduction and internal fixation with a contoured plate and screws, which permitted continuation of the rehabilitation program. The third patient (Case 9) had dislodgment of a free fat graft into the subcutaneous tissues, which was detected with palpation. It did not cause any symptoms, and it was gradually resorbed without sequelae.
The most recent clinical and radiographic follow-up evaluation was performed an average of thirty-four months (range, twenty-four to sixty months) after resection of the synostosis. Follow-up radiographs revealed a recurrence in one patient (Case 16). This patient had had a type-IIIC synostosis (no motion of the elbow) and was the only one in the series who had sustained a closed head injury. The patient, who had had a preoperative ankylosis at -20 degrees of extension, had gained 65 degrees of ulnohumeral motion. She deferred additional operative treatment.
The results of the questionnaire were available for fifteen limbs in fourteen patients; the patient who had a recurrent synostosis was not included, and we were not able to contact two patients. According to the American Shoulder and Elbow Surgeons elbow assessment form, the average postoperative pain score was 18 points (range, 6 to 25 points) and the average activity score was 23 points (range, 4 to 30 points). The average satisfaction score was 8 points (range, 0 to 10 points). Only one patient expressed dissatisfaction with the result, and issues regarding disability compensation were believed to be at least partially responsible for this dissatisfaction.
At the most recent follow-up examination, the average total arc of rotation of the forearm was 139 degrees (range, 65 to 165 degrees)—the average supination was 61 degrees (range, 25 to 85 degrees), and the average pronation was 78 degrees (range, 40 to 90 degrees)—for the seventeen limbs that did not have a recurrence (Table II).
For the eleven limbs in which the heterotopic ossification did not involve the ulnohumeral articulation (types IIIA and IIIB), the average total arc of ulnohumeral motion at the most recent follow-up examination was 106 degrees (range, 75 to 135 degrees): the average flexion was 126 degrees (range, 105 to 135 degrees), and the average extension was -15 degrees (range, -45 to 0 degrees). One patient (Case 16) in whom heterotopic ossification had caused simultaneous radioulnar and ulnohumeral synostoses (type IIIC) achieved an arc of 65 degrees of ulnohumeral motion, but the radioulnar synostosis recurred. The other six limbs with type-IIIC synostosis achieved an average total arc of ulnohumeral motion of 98 degrees (range, 60 to 110 degrees): the average flexion was 127 degrees (range, 100 to 135 degrees), and the average extension was -28 degrees (range, -40 to -20 degrees) (Table II).
According to the classification system of Failla et al.11 for the evaluation of rotation of the forearm, sixteen limbs had an excellent result (Figs. 2-A, 2-B, 2-C, 2-D, 2-E, 2-F, 2-G and 2-H), one had a fair result, and one had a poor result.
Statistical Analysis
In an analysis that excluded the patient who had recurrence of the radioulnar synostosis, we found the average total arc of rotation of the forearm to be comparable between patients who had had a low-energy traumatic injury (138 degrees; 95 per cent confidence interval, 22 degrees) and those who had had a high-energy injury or who had needed more than one operative procedure during the initial hospitalization to treat the initial injury (140 degrees; 95 per cent confidence interval, 16 degrees); between the patients who had been managed with a free fat graft (138 degrees; 95 per cent confidence interval, 21 degrees) and those who had not (141 degrees; 95 per cent confidence interval, 16 degrees); and between the patients in whom rotation of the forearm had been restricted for the first four weeks postoperatively because of a hinged elbow distractor (136 degrees; 95 per cent confidence interval, 37 degrees) and those in whom rotation had not been restricted (140 degrees; 95 per cent confidence interval, 11 degrees).
Although the range of rotation of the forearm was slightly greater in the patients who had had the resection within twelve months after the injury (144 degrees) than in those who had had a late resection (134 degrees), the difference could not be shown to be significant with the numbers available. Likewise, the patients who had had a type-IIIA synostosis had slightly less rotation of the forearm (126 degrees) than those who had had a type-IIIB or IIIC synostosis (143 degrees), but the difference could not be shown to be significant.