Patients
Between January 2003 and October 2006, twenty-two consecutive patients with twenty-seven malunited fractures of the upper extremities underwent a corrective osteotomy with the use of preoperative simulation and a custom-made osteotomy template. All patients were followed for more than twelve months (range, fourteen to thirty-one months; mean, twenty-two months) (see Appendix). There were fourteen male and eight female patients with a mean age of thirty-two years (range, ten to seventy-two years) at the time of surgery. Our institutional review board approved this study. After informed consent was obtained from patients for participation in the study, a preoperative simulation, corrective osteotomy with use of the custom-made template, and physical and radiographic examinations were carried out.
Inclusion criteria included an osseous deformity of the radius, ulna, or humerus and an age of at least ten years at the time of surgery. Exclusion criteria included malunions of the hand and carpus, deformity with marked shortening that would require gradual lengthening, intra-articular involvement of the adjacent joint, and an age of less than ten years. Deformity of small bones and an age of less than ten years were considered exclusion criteria because the reliability of the operative technique with use of a surgical template for smaller bones had not been established and because active remodeling is expected in younger patients18.
This study included four distal humeral malunions (a cubitus varus deformity), ten diaphyseal malunions of the forearm, and eight malunited distal radial fractures. In the ten forearm malunions, the radius was malunited in four, the ulna was malunited in one, and both the radius and the ulna were malunited in five. The average age at the time of injury was 5.6 years for the patients with a cubitus varus deformity, twenty years for those with a forearm malunion, and forty-nine years for those with a malunion of the distal end of the radius. Initial treatment had consisted of closed reduction and cast immobilization in eighteen patients and open reduction and internal fixation in four patients.
At the time of initial presentation to our hospital, all patients demonstrated 5° to 45° of angular deformity of the bone on plain radiographs. Surgery was indicated because of a functional deficit in nineteen patients and an unsightly appearance in three patients. Nine of the ten patients with a forearm malunion had difficulty turning a doorknob and receiving a coin in the open palm because of restricted forearm rotation. The other patient (Case 7), who had malunited fractures of both bones of the forearm, reported recurrent anterior dislocation of the radial head accompanied by pain when he pronated the affected forearm. Three patients with a malunited forearm fracture (Cases 6, 11, and 12) and one patient with a malunited distal radial fracture (Case 20) had distal radioulnar joint subluxation. All of the patients with malunited distal radial fractures reported restricted range of wrist motion, moderate wrist pain, and decreased grip strength, and they experienced difficulty in using the affected hand in daily activities. One patient with cubitus varus (Case 1) could not perform simple activities, such as bringing food to the mouth, with the affected hand because of restricted elbow flexion. The other three patients with cubitus varus complained that the elbow had an unsightly appearance.
Simulation Technique
The affected and the contralateral limbs of all patients were scanned with use of a computed tomography scanner (LightSpeed Ultra 16; General Electric, Waukesha, Wisconsin), at a scan time of 0.5 sec, a scan pitch of 0.562:1, a tube current of 10 to 50 mA, and a tube voltage of 120 kV. Digital data from 0.625-mm slices were sent to a workstation (Precision Workstation 650; Dell, Round Rock, Texas). The malunited bones were segmented, and three-dimensional surface models were constructed by applying three-dimensional surface generation of the cortex of the bone19 with use of the original computer program based on the Visualization Toolkit (Kitware, Clifton Park, New York). On the computer, the deformity of the affected bone was evaluated by superimposing20 it with the goal model (e.g., the mirror image of a contralateral, normal bone), which can be further determined in terms of rotation around and translation along one unique axis, i.e., the three-dimensional deformity axis, with use of the screw displacement axis technique21-23 (Fig. 1, A) (see Appendix for Videos 1 and 2). On the basis of the information obtained about the deformity axis, a corrective osteotomy was simulated24,25 (Figs. 1, B and 2).
Design and Manufacturing of the Custom-Made Osteotomy Template
To reproduce the preoperative simulation during the actual surgery, we developed an operative method using a custom-made osteotomy template that was designed on the basis of a preoperative three-dimensional computer simulation with use of commercially available software (Magics RP; Materialise, Leuven, Belgium) and was embodied as a plastic model through rapid prototyping technology (Eden250; Objet Geometries, Rehovot, Israel, or Viper si2; 3D Systems, Rock Hill, South Carolina) with medical grade resin (Figs. 3 and 4). The custom-made osteotomy template has a shape that closely fits the bone surface and an osteotomy slit or slits and drill-holes that guide the insertion of the Kirschner wires. The slit guides the precise osteotomy cut; and the two sets of Kirschner wires, inserted through the drill-holes at an angle of the deformity, indicate that the reduction is completed when they become parallel to each other. A reduction guide to maintain the parallel position of the Kirschner wires is prepared preoperatively in the same manner as is the custom-made osteotomy template (see Appendix for Video 3).
Surgical Technique
During the procedure, the template was placed on the bone surface, the bone was osteotomized through a slit in the template, and the deformity was corrected as simulated preoperatively (Figs. 5 and 6). We were able to perform all osteotomies as preoperatively simulated. This was followed by internal fixation (plates and screws were used in twenty-four bones; Kirschner wires, including tension band wiring, were placed in two bones; and both methods of fixation were used in one bone). The template was fitted to the bone surface with reference to the characteristic configuration of the malunion and anatomical landmarks (e.g., the Lister tubercle, humeral condyles, and olecranon fossa). In patients with a malunited forearm fracture, the distance between the osteotomy site and the radial or ulnar styloid was calculated preoperatively on the computer and also served as a reference. A rotational osteotomy was performed on ten bones, a closing wedge osteotomy was done on five bones, and an opening wedge osteotomy with bone graft was performed on twelve bones. The amount of correction, which was the rotation around the three-dimensional deformity axis, ranged from 13° to 67°, with an average of 32°. In two rotational osteotomies, 3-mm and 2-mm shortenings along the deformity axis were required. In the other three rotational osteotomies, 2, 10, and 4-mm lengthenings with an interposition bone graft were required. These shortenings and lengthenings were performed to correct radioulnar discrepancies. The osteotomy template and the reduction guide were designed with consideration of these longitudinal adjustments. Open reduction of distal radioulnar joint subluxation in three patients with malunited forearm fractures (Cases 6, 11, and 12) and osteosynthesis of an ulnar styloid nonunion in one patient with a malunited distal radial fracture (Case 21) were combined with the corrective osteotomy.
The average time between the initial injury and the corrective osteotomy was thirteen years (range, two to twenty-seven years) for cubitus varus deformity, thirty months (range, six to 100 months) for malunited forearm fractures, and twelve months (range, five to twenty-three months) for malunited distal radial fractures.
Radiographic and Clinical Evaluation
Radiographic and clinical evaluations were conducted for all patients before surgery and at the most recent follow-up evaluation. Union was considered complete when the osteotomy line had disappeared and osseous trabecular continuity was confirmed. For cubitus varus deformities, the humerus-elbow-wrist angle26 (defined by the longitudinal humeral axis and a line passing through the proximal and distal midpoints of the radius and ulna) and the tilting angle27,28 (the anterior tilt of the articular condyles with respect to the long axis of the humerus) were examined with anteroposterior and lateral radiographs of the upper extremity made with the forearm in a supinated position. For malunited forearm fractures, anteroposterior and lateral radiographs made with the forearm in neutral position or in full supination were compared with those of the contralateral, normal side. The angular deformities of the radius and/or ulna were measured in reference to the contralateral forearm with use of the radiographs made with the forearm in the same position29. The greater angle of deformity between the anteroposterior and lateral radiographs was defined as the radiographic deformity angle. For malunited distal radial fractures, volar tilt, radial inclination, and ulnar variance were evaluated from wrist radiographs30. The values used in our study were the average of those of two independent reviewers (K.O. and H.M.). For clinical evaluation, ranges of motion of the adjacent joint were measured. Forearm rotation was measured with use of a goniometer with the humerus in a vertical position and the elbow in 90° of flexion31. Wrist flexion-extension was measured with the goniometer placed along the axis of rotation of the respective joints and with the forearm in a neutral position9,32. Grip strength was measured with use of a Jamar dynamometer (Matsumiya Medical Instruments, Tokyo, Japan) and was recorded as a percentage of that of the contralateral, normal side. Pain at the adjacent joint was graded as none (no pain), mild (occasional pain with excessive use of the hand), moderate (persistent, but endurable, pain), or severe pain necessitating analgesic control. The level of satisfaction was graded by the patient as very satisfied, satisfied, neither satisfied nor dissatisfied, dissatisfied, or very dissatisfied33.
Statistical Methods
The differences between the radiographic values of the extremity that had been operated upon and the values from the normal side, and the differences between the preoperative and postoperative range of motion and grip strength for each deformity group, were determined by paired t test. Preoperative and postoperative scores for pain in each deformity group were compared with use of the Wilcoxon signed-rank test. Significance was established at p < 0.05.
Cubitus Varus Deformity (Table I)
All osteotomy sites had united by an average of 3.3 months (range, nine to twenty-two weeks) after surgery. The average humerus-elbow-wrist angle and tilting angle were 20° (varus alignment) and 19°, respectively, before surgery and 2° (valgus alignment) and 28° after surgery. In one cubitus varus deformity (Case 4), although the deformity was completely corrected and fixed with Kirschner wires and suture wires, reduction was lost in the early postoperative course. The patient did not want additional surgery, and the osteotomy site united with moderate displacement. The humerus-elbow-wrist angle in this patient had been 21° (varus alignment) before surgery, 7° (valgus alignment) just after surgery, and 5° (varus alignment) at the most recent follow-up evaluation. Except for this patient, the average humerus-elbow-wrist angle improved from 19° (varus alignment) before surgery to 5° (valgus alignment) after surgery. One patient with a cubitus varus deformity (Case 1) who had elbow hyperextension and restricted flexion attained normal range of elbow motion after surgery (Figs. 7 through 11). The range of elbow joint motion of the other patients with cubitus varus did not change significantly. Three patients were very satisfied with the operation, and one patient (Case 4) was neither satisfied nor dissatisfied with the operation. One patient (Case 2) complained of mild discomfort around the hardware, which was subsequently removed.
Malunited Forearm Fracture (Table II)
The osteotomy sites united an average of sixteen weeks (range, eight to twenty-six weeks) after surgery. The average angle of deformity before surgery was 16° (range, 5° to 33°) compared with the normal side. It was well corrected to 1° (range, 0° to 3°) after surgery. No distal radioulnar joint discrepancy was observed on the radiographs of the involved forearm compared with the normal side. The average range of forearm pronation and supination significantly improved from 60° and 19°, respectively, before the operation to 82° and 73° after the operation (p < 0.01 for both). Restricted forearm supination persisted in one patient (Case 5) with malunited fractures of both bones of the forearm, although the malunions were well corrected. In this patient, the initial injury had occurred when the patient was seven years old, and corrective surgery was performed when she was sixteen years old. It was believed that changes in joint configurations and soft-tissue contractures during this long period of time were the cause of the residual restricted forearm supination. Five patients reported pain in the adjacent joint before surgery; four of them (Cases 6, 11, 12, and 14) experienced pain in the distal radioulnar joint and one (Case 7) had pain in the proximal radioulnar joint. The preoperative pain experienced by all of these patients disappeared or decreased substantially after surgery. Painful recurrent dislocation of the radial head in one patient (Case 7), with malunited fractures of both bones of the forearm, disappeared after corrective osteotomy without soft-tissue reconstruction, and the patient was able to resume sports activity. Six patients were very satisfied, and four patients were satisfied with the surgery.
Malunited Distal Radial Fracture (Table III)
All osteotomy sites had united an average of ten weeks (range, eight to thirteen weeks) after surgery. The mean volar tilt, radial inclination, and ulnar variance were -17°, 14°, and 3.4 mm, respectively, before the operation and 8°, 23°, and 0.6 mm at the final follow-up evaluation. The average wrist flexion and extension improved from 33° and 54°, respectively, before surgery to 62° and 66° after surgery (p < 0.01 for both). The average ranges of forearm pronation and supination improved from 58° and 69°, respectively, before surgery to 79° and 78° after surgery (p = 0.042). Restricted forearm supination persisted in one patient (Case 20), although the malunion was well corrected. In this patient, we believe that preexisting distal radioulnar subluxation that was not addressed well during surgery was the cause of the residual forearm rotation loss. The average grip strength improved from 42% to 86% of that of the normal side. All eight patients experienced wrist pain before surgery, which disappeared or decreased after surgery (p = 0.013). Three patients (Cases 16, 21, and 22) complained of discomfort around the hardware, which was subsequently removed. Four patients were very satisfied, three were satisfied, and the remaining patient (Case 20) was neither satisfied nor dissatisfied with the operation.
Complications
Postoperative partial loss of correction occurred in one patient with cubitus varus deformity. One patient who underwent surgery for cubitus varus deformity and three patients who underwent surgery for a malunited distal radial fracture complained of hardware-related pain or discomfort that necessitated hardware removal. Distal radioulnar subluxation in one patient with a malunited distal radial fracture persisted after surgery. No other major complications, including nonunion, neurovascular compromise, or infection, were observed.
Malunion of the forearm bones, cubitus varus deformity, and a malunited distal radial fracture are typical posttraumatic deformities of the upper extremity. Symptoms and functional impairments related to these deformities may cause serious disabilities26,30,31,34-36. Although corrective osteotomies have been performed to improve the function and appearance of the extremity, it is not easy to correct three-dimensionally complex osseous deformities accurately2-4,7,8,16,31,37. Previous studies have suggested the usefulness of frontal and sagittal radiographs in the preoperative planning of a corrective osteotomy, although estimation of three-dimensional deformities with two-dimensional images has limitations5,6,9,12,38,39. Failure to make an accurate correction may lead to inferior clinical results, especially in the upper extremity, where anatomical bone configuration is of considerable importance to function3-5,8,9,31,35,38,39. To solve these problems, we developed a novel computer-assisted system to guide corrective osteotomy15-17. Our program can indicate the optimum pattern and plane of corrective osteotomy by calculating the axis and amount of three-dimensional deformity. The osteotomy template navigates the surgical procedure to realize the preoperative simulation.
Restricted forearm rotation is the key problem associated with malunions of the forearm bones4,35,36,39. Correct rotational alignment, restoration of normal length, and achievement of axial alignment of both bones are necessary to obtain a good range of forearm rotation40. Cadaver and clinical studies have suggested that angular deformity of the radius and/or ulna of >10° causes limitations of forearm rotation3,4,31,39, although the relation between the severity of the deformity and the decrease in forearm rotation has not been clearly established. Contracture of the interosseous membrane, the distal and proximal radioulnar joint capsules, and other surrounding soft tissues is an additional cause of restricted forearm rotation. In long-standing forearm deformity, a combination of these factors is probably responsible for the substantial pathology of restricted range of forearm rotation36. In corrective surgery, the challenge is to reduce two parallel rotating long bones while maintaining the congruity of the adjacent joints35. Trousdale and Linscheid36 reported the clinical results of corrective osteotomy of forearm malunion. The average arc of forearm rotation improved to only 102° in chronic forearm malunions treated more than twelve months after the initial injury, whereas it improved to 156° in those treated within twelve months of the initial injury. They concluded that corrective osteotomy for a posttraumatic malunion was best performed within twelve months of the initial fracture. They did not refer to any radiographic parameters. In our series, the average forearm rotation in the patients managed late and in those managed early improved to 152° and 159°, respectively, and angular deformities were well corrected in both groups. Although restricted forearm supination remained in one patient (Case 5), who had undergone surgery nine years after the initial injury, the range of forearm rotation had improved by 40° at the time of the final follow-up. These results indicate that an accurate three-dimensional correction of a deformity of the forearm bones can yield reasonable improvement in forearm motion, even when performed long after the initial injury.
Cubitus varus deformity is a malunion of the distal end of the humerus that generally includes varus, internal rotation, and hyperextension deformities8,9,28,41. In the past, it was considered a cosmetic problem, and correction of varus deformity alone has been an accepted practice26,42,43. Recently, because joint laxity34,44 and tardy ulnar nerve palsy45,46 have been reported as late complications, several investigators have advocated that correction of frontal plate angular deformity is not enough and that rotational deformity should also be corrected8,9,41. In fact, internal rotation of the distal end of the humerus of >25° exists in 22% of patients with a cubitus varus deformity47. In our opinion, this should not be overlooked, considering its effect on shoulder motion and the appearance of the extremity48. Previously reported attempts at three-dimensional correction, however, were based on preoperative planning with use of data from plain radiographs and changes in the range of shoulder motion9,28,43. This procedure was also criticized for its technical difficulty and the poor osseous contact achieved at the osteotomy site26. In contrast, our system provides a simple and accurate correction based on three-dimensional data. The contact area at the osteotomy site can be visualized easily with use of three-dimensional images, which allow practical planning. The results of the present study, in which the postoperative humerus-elbow-wrist angle was an average of 5° and was within 3° of that of the contralateral, normal side, were the same as or better than those of previous studies8,41,42. Yamamoto et al.28 performed a three-dimensional corrective osteotomy in seven patients, and a hyperextension deformity ranging from 5° to 20° persisted after surgery in three patients. In our series, the tilting angle improved to 28°, on the average, and was almost the same as that of the contralateral, normal side in all patients, including one patient (Case 1) who had 35° of hyperextension deformity before surgery.
Malunion of the distal end of the radius is one of the most common deformities of the upper extremity. Dorsal tilt, radial shortening, and a decrease in radial inclination have been cited by several investigators when attempting to plan three-dimensional correction with use of plain radiographs1,2,49-53. Fernandez49,50 suggested preoperative planning with use of frontal and sagittal plane radiographs and reported good clinical results, with postoperative ranges of 50° of flexion and 57° of extension at the wrist at the time of the final follow-up. However, dorsal tilt of the distal end of the radius remained in eight of the thirty-five patients, and radial shortening of >2 mm persisted in four of the thirty patients, in whom ulnar head resection was not performed. Athwal et al.38 introduced a computed tomography-based computer-assisted three-dimensional surgical planner that calculates the corrected position of the distal end of the radius, including an evaluation of rotational deformity with use of the contralateral, normal wrist as the template. They applied an intraoperative guidance system, which linked the preoperative plan to an optical tracking device. In six patients, the radiographic parameters of radial inclination, volar tilt, and ulnar variance relative to the contralateral, normal wrist were 2°, 1°, and 0.4 mm, respectively. At the time of the final follow-up, the average ranges in wrist flexion and extension were 47° and 42°, respectively. However, an optical tracking system requires bulky equipment and computers, monitors, and a system operator to be present during the surgery. In our series, the error between the preoperative simulation and postoperative result was <1° for both radial inclination and volar tilt and 0.3 mm for ulnar variance. As for wrist range of motion, the average ranges of wrist flexion and extension at the time of the final follow-up were 62° and 66°, respectively. The radiographic and clinical results of correction for malunited distal radial fractures in the present study were comparable or superior to those of previous studies, and the small custom-made template was quite practical. We could exactly adjust the postoperative ulnar variance in computer simulation and could avoid ulnar head resection, which has often been performed with conventional osteotomies49,50,54.
The preliminary results in our twenty-two patients indicate that this simulation technique is a clinically reliable method. Three-dimensionally complex deformities can be accurately corrected with a simple one or two-plane osteotomy. The shortcomings of our technique include radiation exposure during computed tomography scanning, the time and effort required for computer simulation, the cost of the custom-made template, and the use of software that is not available to the public. A previous experimental study showed that radiation exposure with this system can be reasonably reduced compared with typical doses of radiation from conventional diagnostic computed tomography scanning55. Computer simulation takes two to three hours for a trained operator. The time and cost of manufacturing the template are three to eight hours and about US$50, respectively. The software is currently used only in our institute for this specific clinical study; however, we plan to distribute it in the near future.
Because this study is based on a preliminary series, it has several limitations such as a relatively small number of patients in each of the corrective osteotomy groups and the absence of a control group. An additional criticism is that a minor deformity after a distal radial fracture could be corrected reasonably well without use of this technique. Therefore, further investigation is needed to determine its clinical value. However, the results of this series are encouraging. We hope that the three-dimensionally accurate correction realized by our technology will contribute to the field of deformity correction of the upper extremities.
Note: The authors thank Takeshi Yoshida, MD, Kakuro Denno, MD, Mitsuru Horiki, MD, and Koichi Tada, MD, from the Department of Orthopaedic Surgery, Kansai Rosai Hospital; Kozo Shimada, MD, from the Department of Orthopaedic Surgery, Osaka Koseinenkin Hospital; Shunichi Henmi, MD, from the Department of Orthopaedic Surgery, Ikeda Municipal Hospital; Yoshiharu Nakamura, MD, from the Department of Orthopaedic Surgery, Sakai Municipal Hospital; and Ryoji Nakao, computer programmer, Department of Orthopaedic Surgery, Osaka University Graduate School of Medicine, for their contributions to this study.