Corrective Osteotomies in Malunions of the Distal Radius

Do We Get What We Planned?

Abstract

Fifteen patients with symptomatic malunions of the distal radius were treated with osteotomies, corticocancellous bone grafting, and plate and screw fixation. We investigated the ability of precise preoperative planning of the size and shape of the corticocancellous bone graft to restore alignment of the radius to within 5° angular deformity and 2 mm ulnar variance as compared with the opposite uninjured wrist. Only six of 15 patients (40%) satisfied these criteria. Inter-rater reliability of radiographic assessment was greater than 0.85. Five patients had residual radial inclination or sagittal tilt greater than 10° with respect to the uninvolved wrist. Four patients had a residual ulnar variance greater than 2 mm with respect to the uninvolved wrist. Residual shortening (three of four patients), but not residual angulation, was associated with unsatisfactory pain and stiffness an average of 19.5 months after osteotomy (range, 11-32 months). We conclude that a distal radius osteotomy using a precisely planned and measured interpositional corticocancellous graft does not restore distal radius alignment in most patients, and that failure to restore length is associated with continued pain and stiffness.

Level of Evidence: Therapeutic study, Level IV (case series). See the Guidelines for Authors for a complete description of levels of evidence.

The incidence of distal radius fractures is approximately 0.27%.¹⁷ They are twice as frequent in women, and 73% are caused by falls. Two age peaks of fracture incidence occur in the second decade and sixth through eighth decades.²⁰ Among the different fracture patterns, Colles' fractures are the most frequent, representing approximately 60% of distal radius fractures.² Complications are neither infrequent nor negligible. An incidence of as much as 7.9% has been reported for compression neuropathy, 6.5% has been reported for posttraumatic arthritis, and 5.3% has been reported for malunions occurring in as much as 76% patients after nonoperative treatment.^6,7,19

Distal radius corrective osteotomies are considered for patients with painful malunions and have been recommended for younger or more active individuals with angulations greater than 25°.²³ Surgery generally is considered contraindicated in when there are dystrophic changes, inadequate bone quality, osteoarthritic changes, or fixed carpal malalignment.⁸ Corrective osteotomy of the distal radius allows reestablishment of normal anatomic relationships in the distal radioulnar joint (DRUJ) and radiocarpal joint.

Osteotomy of the distal radius can improve motion, function, and comfort level in patients with symptomatic malunited fractures of the distal radius. It is common to use a structural corticocancellous bone graft interposed in the osteotomy site to help restore alignment. This requires thorough preoperative planning and precise shaping of the bone graft to ensure insertion of the bone graft into the osteotomy produces the planned correction.^12,16,28

We tested the hypothesis that this technique can restore alignment of the distal radius to within 5° angular alignment and 2 mm ulnar variance as compared with the opposite uninjured wrist. We addressed the following specific questions: (1) Is radiographic assessment reliable?; (2) Does the osteotomy restore alignment and length of the radius?; and (3) Does residual deformity affect functional outcome?

MATERIALS AND METHODS

We retrospectively reviewed 15 patients with symptomatic established distal radius malunions. These 15 patients included all patients treated with corrective radius osteotomies with interposition of structural bone grafts and osteosynthesis between September 1997 and December 2003. The indication for the osteotomy was based on the occurrence of pain in activities of daily living and a side difference of 10° or greater in sagittal tilt and radial inclination. Exclusion criteria included advanced osteoarthritic (OA) changes in the radiocarpal joint or advanced osteoporosis. Patients presenting with excessive ulnar variance but no major anomalies in sagittal tilt and radial inclination were treated with ulnar shortening osteotomies. The mean age of the patients at the time of surgery was 38.8 years (range, 13-76 years). There were seven female patients and eight male patients. Eight right and seven left extremities were treated. The dominant arm was affected in seven patients, and the nondominant arm was affected in eight patients. Twelve patients had strictly extraarticular fractures of the distal radius. In three patients this was combined with an intraarticular nondisplaced component. Before corrective osteotomy, 11 patients had been treated nonoperatively, three had been operated on, and one had not consulted a doctor. Among the patients who were treated nonoperatively, eight were treated with short-arm casting and three were treated with long- arm casting. The mean delay from fracture to corrective osteotomy was 7 years (range, 2-19 years). Twelve patients had malunions with dorsal angulations and three patients had excessive volar angulations. No patients had an intraarticular mal- union. The average followup was 19.5 months (range, 11-32 months).

Preoperative planning defined a structural bone graft of precise shape used to restore articular surface orientation and distal radius length. Three planning techniques were used, each patient having been assessed with only one method according to surgeon preference and availability of computer-assisted software. The first method, called the sum method, consisted of simulating the osteotomy by cutting and reorienting a 1:1 scaled paper print of digitalized radiographs. The height of the four bone graft edges was calculated by adding the height of the bone gap measured on the posteroanterior (PA) and lateral radiographs after radius articular surface reorientation and radius length correction (Fig 1). The second method consisted of correcting the three-dimensional deformity through bone graft insertion into the plane of true deformity. This bone graft had a triangular-shaped section. Radial inclination and sagittal tilt were compared with the opposite side. Both angles served to obtain the angle of the triangular-shaped bone graft section (δ) and the graft insertion angle with respect to the frontal plane (β), using established tables (Fig 2).²³ A computer-assisted bone graft design based on a three-dimensional (3-D) model of correction was used as the third planning technique.^4,5 This technique used commercially available software (Pace Systems, Freiburg-Umkirch, Germany). The coordinates of 14 reference points were assessed on the affected and unaffected distal radii using a transparent grid. The coordinates were computed to generate the shape and size of the bone graft.

Fig 1A: E. (A) The level of the osteotomy is simulated on (A) PA and (B) lateral radiographs of the involved wrist. After the correct distal radius length and orientation have been obtained using the uninvolved wrist as the gold-standard, (C) the resulting radial and ulnar bone (D) defects and palmar and dorsal bone defects are added to obtain (E) the 3-D design of the bone graft 3-D. This method does not consider how changes in volar/dorsal inclination cause a change in radial/ulnar inclination.⁵

Fig 2A: C. (A) The deformity angle of the distal radius is assessed on PA and lateral radiographs by comparing the involved and uninvolved orientations of the distal radius. Angular side differences are termed δx in the frontal plane and δy in the sagittal plan. These angles are used to define the true angle of correction (δ) and the orientation of the plane of the true deformity (β), as found in established tables.²³ (B) The true angle of correction δ and the plane of the true deformity β define the angle of the wedged bone graft and (C) the direction in the transverse plane in which the bone graft should be incorporated to correct the 3-D deformity. Beta equals zero if the malunion is restricted to the sagittal plan.

The contralateral, uninvolved side was used as the reference for planning in all but one patient. This patient had bilateral involvement. A sagittal tilt of 10°, a radial inclination of 25°, and a null ulnar variance were defined arbitrarily as the reference values for this patient. If ulnar variance was greater than 2 mm on the uninvolved wrist, the osteotomy was planned to produce a variance of 2 mm or less in the surgically treated wrist. The osteotomy was always parallel with the distal radius articular surface in the sagittal plane, and perpendicular to the distal radius axis in the frontal plane. The level of the osteotomy was adjacent to the proximal edge of the radius sigmoid notch. In all patients, a tricortical bone block of the iliac crest was harvested and shaped with a saw to obtain bone edges with a precision of 1 mm or less with respect to the abacus. All operations were performed by the two senior authors (LN, CED).

The wrist was immobilized after surgery in eight patients, whereas seven patients started range of motion (ROM) exercises shortly after surgery.

The patients' records and radiographs were reviewed by two independent observers (AvC, DA) who were blinded to the results. The clinical followup included ROM assessment of the wrist in flexion-extension and pronation-supination measured with a hand-held goniometer. The grip strengths of both hands were measured with a Jamar hand dynamometer in setting II (Jamar Hand Dynamometer, Sammons Preston Inc, Bolling-brook, IL) using the best value of three attempts.¹¹ Percentages of strength between sides were calculated. Clinical results were assessed as raw scores based on the criteria described by Fernandez and the DASH functional questionnaire.^9,14 Pain was evaluated on a five-point backward scale ranging from zero, meaning severe pain, to four, meaning no pain. Four points were assigned to each of the following: flexion-extension of 130° or greater, pronation-supination of 160° or greater, and strength of 80% or greater of the uninvolved side. Three points were assigned to each of following: flexion-extension of 100° to 130°, pronation-supination of 140° to 160°, and strength of 65 to 80%. Two points were assigned to each of following: flexion- extension of 80° to 100°, pronation-supination of 120° to 140°, and strength of 40% to 65%. One point was assigned to each of following: flexion-extension of 80° or less, pronation-supination of 120° or less, and strength of 40% or less. The first part of the DASH questionnaire was used to assess impairment in activities of daily living. This section has 23 questions with a scale ranging from 1 (no difficulties with the task) to 5 (task impossible to perform). A negative coefficient was affected to the score for motion, strength, and pain because of the opposite scaling direction with respect to the DASH score. All scores were aggregated to obtain an observer-based aggregate scoring system.³³ The overall scaling direction indicated the higher the score, the worse the functional outcome.

Posteroanterior and lateral wrist radiographs from preoperative and last followup visits were analyzed and compared with the preoperative views of the unaffected side. Ulnar variance, radial inclination (angle between distal articular surface and a line perpendicular to the long axis of the radius on the PA view), and sagittal tilt (angle between distal articular surface and a line perpendicular to the long axis of the radius on the lateral view) were measured according to the criteria described by Mann et al.¹⁸ A negative sagittal tilt value indicated a dorsal tilt. Osteoarthritic changes were assessed according to the classification described by Knirk and Jupiter.¹⁵

The measurements were recorded in a Microsoft® Office Excel® 2003 data sheet (Microsoft Corporation, Walisellen, Switzerland). Interclass correlation coefficients were calculated using repeated-measures analysis of variance, with StatView 5.01 (SAS Institute, Cary, NC). The reliability was rated as acceptable if the result was 0.80 or greater. A paired t test was used to evaluate side difference in radiologic assessment of the distal radius. Groups of patients were designed according to radiologic criteria. Results were expressed as mean ± SEM. The Mann- Whitney-Wilcoxon rank-sum test was used to evaluate significance in functional outcome among groups of patients. Significance was set at p < 0.05.

RESULTS

Some patients did experience complications in this series. Five patients had pain over the implant (two volar plates and three dorsal plates), and required hardware removal before the last followup (Table 1). One patient (Patient 10) presented at the 3-month followup with Type 1 complex regional pain syndrome (CRPS-1). Grade 1 OA in the radiocarpal joint was seen in one patient before osteotomy, with no change at the last followup (Patient 9). Patient 7 had Grade 1 OA of the radiocarpal joint at the last followup. This patient also had preoperative Grade 2 OA of the DRUJ and had a Sauvé-Kapandji procedure during the osteotomy. The patient patient required tenodesis of the ulnar stump because of painful instability after the Sauvé-Kapandji operation.

TABLE 1: Clinical Data

Interobserver reliability was adequate for all tested criteria used to assess distal radius anatomy. The interclass correlation coefficient was 0.88 for the sagittal tilt, 0.87 for the radial inclination, and 0.91 for the ulnar variance.

The osteotomy did not restore alignment and length in all patients. We observed a mean side-to-side difference (p < 0.01) for the radial inclination but not for the sagittal tilt (p = 0.19) or ulnar variance (p = 0.28). Six patients (40%) had restoration within 5° of the radial surface orientation in frontal and sagittal directions and within 2 mm of the ulnar variance with respect to the uninvolved wrist (Table 2). Conversely, radial inclination was not restored within the defined criteria in nine patients, sagittal tilt not restored in five patients, and four patients had ulnar variance greater than 2 mm with respect to the uninvolved wrist (Table 3).

TABLE 2: Radiographic Assessments

TABLE 3: Functional Scores and Radiologically Assessed Differences at Last Followup

Among the four patients with an ulnar variance greater than 2 mm, three had residual pain, limitation in ROM, and weakness. In contrast, among the patients with correct anatomic restoration, only one patient (the CRPS-1 diagnosis) had major functional impairment. The mean aggregate score for six patients with restoration within 5° of the radial surface orientation in frontal and sagittal directions and within 2 mm of the ulnar variance with respect to the uninvolved wrist (Group 1) was 35.3 ± 15.7 (Table 3). The mean aggregate score was 25.6 ± 8.7 for the eight patients with restoration outside these limits (Group 2). The mean aggregate score for patients with ulnar variance of 2 mm or less (Group 3) and patients with ulnar variance greater than 2 mm (Group 4) with respect to the uninvolved wrist were 25.7 ± 8.9 and 39.7 ± 18.0, respectively. The aggregate score between patients in Groups 1 and 2 and between patients in Groups 3 and 4 were similar.

DISCUSSION

For patients with combined shortening and angulation of the distal radius, an opening wedge osteotomy with an interpositional graft may be considered for reorienting the joint surfaces and normalizing the ulnar variance.²³ The most widely used method consists of interposing a structural bone graft. This compression-resistant bone block provides stability against axial load and usually is combined with a nonlocking plate osteosynthesis.⁸ The dimensions of the structural bone graft must be adapted to the size of the bone defect to avoid overcorrection or undercorrection.

Although we used thorough preoperative planning to ascertain the size of the corticocancellous bone graft, our technique was successful in restoring distal radius alignment to within 5° angular and 2 mm axial deformity in only six of 15 patients. The preoperative planning methods were equally efficient in correcting the radius deformity in the sagittal plan, but residual excessive radial inclinations were common. One patient had a residual deformity greater than 20° in the frontal plane. Conversely, all three planning methods were unable to completely prevent insufficient correction of the ulnar variance. The small number of osteotomies performed with each method prevented additional analysis among them.

These residual deformities did not correlate with poorer outcome. This finding contrasted with the results of Prommersberger et al, who suggested gross residual deformities in the sagittal plane substantially impair functional outcome.²⁸ One patient with painful instability of the ulnar stump after a Sauvé-Kapandji operation was treated successfully with a tenodesis. Another patient who had a CRPS I develop had a poor outcome. This result had a bad influence on the overall clinical outcome in our series of patients and probably impaired data analysis because this patient had excellent anatomic restoration. Based on an aggregated raw score comparison, our results were slightly poorer than to those of Fernandez and Geissler¹⁰ but comparable with those of Shea et al.³⁰ Restoration of sagittal tilt and radial inclination correlated poorly with the functional outcome, but an additional complication such as CRPS considerably worsens the outcome.

Of the four patients with an ulnar variance greater than 2 mm with respect to the uninvolved side, three had residual pain, stiffness, and weakness of the wrist. Our results were similar to those of Oskam et al, who reported on corrective osteotomy for malunion of the distal radius.²⁵ Aro and Koivunen also reported poor outcomes in distal radius malunions after radial shortening greater than 5 mm.¹ The biomechanical effect of radial shortening has been investigated experimentally and clinically. Shortening of 6 mm or greater causes the ulna to impinge on the triquetrum and the lunate,^21,26,27 and limits forearm rotation.³ A 2.5-mm increase in ulnar length increased the ulnar load 42%.²⁶ Chronic ulnar head impaction against the triangular fibrocartilage complex and the ulnar carpal bones results in progressive triangular fibrocartilage complex deterioration, lunate and ulnar head chondromalacia, and lunotriquetral ligament attrition, all resulting in pain and weakness.¹³

For surgeons who prefer to use a structural bone graft, small errors in bone-graft sizing can be accounted for by changing the depth to which the triangular-shaped bone graft is inserted to keep bone-to-bone contact while ensuring adequate restoration of radial length. Nevertheless, thorough planning does not replace careful intraoperative fluoroscopic assessment, especially to confirm restoration of the radius length. Alternatively, morselized bone grafts have been used in combination with locking plates. In this technique, preoperative planning is not mandatory and adequate placement of the volar plate and lengthening of the radius restore alignment. The axial force is transmitted to the locking plate until completion of bone stability occurs.^24,29 Although, this technique seems reliable in primary osteosynthesis and corrective osteotomy of the distal radius,^22,29 a biomechanical study showed the unfilled hole at the site of osteotomy creates a weakness in the plate.³² Additional studies of this technique are required before it can be considered to be superior to the structural bone graft we used.

Using thorough preoperative planning neither reproducibly restored alignment nor improved the functional outcome in corrective distal radius osteotomies. Our data suggest the length of the radius must be restored to improve the functional outcome in patients with symptomatic distal radius malunions. Therefore, more efforts must be made to improve reliability in planning and performing distal radius osteotomies.

Acknowledgment

We thank V. Zdravkovic for assistance with the computer assisted design.