- Email: [email protected]
Ulnar Nonunion After Osteoclasis for Rotational Deformities of the Forearm
Ulnar Nonunion After Osteoclasis for Rotational Deformities of the Forearm John F. Dalton IV, MD, Paul R. Manske, MD, J. Clint Walker, MD, Charles A. Goldfarb, MD From the Department of Orthopaedic Surgery, School of Medicine, Washington University in St. Louis, St. Louis, MO.
Purpose: Osteoclasis, a minimally invasive technique to rotate the radius and ulna, is used commonly to correct forearm rotational deformities in children. The purpose of this investigation was to evaluate objectively osteotomy healing in patients treated with osteoclasis, with specific attention given to the risk for nonunion. Methods: We identified 69 extremities in 65 children treated with osteoclasis and performed retrospective chart and radiographic reviews to evaluate the time to union of the radius and ulna and factors influencing healing. Results: The average rotational correction was 90°. Twenty-one ulnas had either delayed union or nonunion. Forty-eight of the forearms healed in less than 3 months. Factors correlated with a significantly decreased union rate included increased patient age, percutaneous technique, osteoclasis site in the proximal ulna, and primary diagnoses other than congenital radioulnar synostosis. Preoperative forearm position, magnitude of position correction, and treatment of the periosteum were not associated with changes in union rates. Conclusions: Forearm osteoclasis has a delayed union rate of 16%. Timely union of the ulna appears to be influenced by both patient-centered factors and surgical technique. (J Hand Surg 2006;31A:973–978. Copyright © 2006 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Therapeutic, Level IV. Key words: Ulna nonunion, osteoclasis, osteotomy, forearm, synostosis.
arked rotational deformity of the forearm in children and adolescents can limit upper-extremity use and surgery may be indicated to improve patient function. Although occurring most commonly in children with congenital radioulnar synostosis, rotational abnormalities also are seen in patients with brachial plexus palsy, cerebral palsy, and arthrogryposis. In patients with a fixed or severe rotational deformity of the forearm osteoclasis is a commonly used technique for correction.1,2 Although a variety of other procedures have been described— including osteotomy of only the radius,3 resection of the proximal radioulnar synostosis,4,5 osteoclasis through the radioulnar synostosis,3,6 or radius and ulna osteotomy with internal fixation7—surgeons use osteoclasis because of its surgical sim-
M
plicity and the previously reported outstanding results.1,2 For more than 40 years we have used 2-stage osteoclasis to correct marked rotational abnormalities of the forearm.2 In stage I a minimally invasive osteotomy of both the radius and ulna is performed; the extremity then is splinted in the preosteotomy position. In stage II (⬇10 days after stage I to allow for early callus formation) the forearm is rotated into the desired position and held in a long-arm (above elbow) cast until healed. No internal or external fixation is used. In 1995 Lin et al1 reported the outcome of 23 patients with forearm rotational deformity as a result of neuromuscular, developmental, or congenital problems treated with this osteoclasis technique. Good or excellent results were achieved in more than 90% of their patients with an average correction of The Journal of Hand Surgery
973
974
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
81° and 69° for pronation and supination deformities, respectively. Three complications were reported: 2 nonunions and 1 recurrence of the deformity. The functional results of the osteoclasis procedure continue to be satisfactory. The purpose of this investigation was to evaluate a larger cohort of patients treated with this osteoclasis method with specific attention given to the risk for nonunion and the factors affecting healing.
Materials and Methods This investigation was performed at the Shriners Hospital for Children. After obtaining institutional review board approval we used an operating room logbook to identify all patients who had a 2-staged forearm osteoclasis performed between 1970 and 2004. The patients identified through this search were evaluated further with thorough chart and radiographic reviews. Sixty-five patients who had 69 osteoclases by 1 of 3 attending surgeons were identified. There were 38 male and 27 female patients. Osteoclasis was performed at an average age of 10 years (range, 2–20 y). The diagnosis was obstetric brachial plexus palsy for 28 extremities, congenital radioulnar synostosis for 24 extremities, spastic diplegia for 6 extremities, arthrogryposis for 5 extremities, multiple osteochondromatosis for 2 extremities, and fracture malunion, Maffucci’s enchondromatosis, Conradi-Hünermann syndrome, and Volkmann’s ischemic contracture for 1 extremity each. All patients who had forearm osteoclasis by the general methods described by Manske et al2 for rotational deformity of the forearm were included. Any patient with functional limitations caused by the forearm position was eligible for inclusion. Most patients had functional limitations attributable to a fixed pronation or supination deformity of more than 60°. For example patients with fixed pronation deformities have difficulty receiving objects in the palm (ie, holding coins). Patients with fixed supination deformities may accommodate by abducting the shoulder but still may have difficulty using a keyboard in a comfortable position. Patients with either supination or pronation deformities have difficulty using the affected extremity as a helper hand for 2-handed activities such as carrying large objects. A few patients had bilateral deformities in which both forearms were fixed in the same position. Patients who had rotational osteotomy of only the radius and patients treated by any other method than osteoclasis were excluded. Differences in osteoclasis technique
(including osteotomy site and specific osteotomy technique) were based purely on the preference of the attending surgeon. Procedure In 17 of 69 extremities stage I of the procedure was performed by percutaneously drilling the forearm bone cortices with a K-wire and the bones were fractured by manipulation. In 52 extremities stage I was performed open, accessing both forearm bones through small (1- to 2-cm) incisions; the cortices were drilled and the fracture was completed either by manipulation or with an osteotome. After stage I all forearms were splinted in the preoperative position. Stage 2 was performed an average of 10 days (range, 6 –19 d) after stage I. The forearm was rotated through the osteoclasis site into the desired position as determined by the attending surgeon. The position was based on the concept of placing the dominant extremity in a position of slight pronation and placing the nondominant extremity in a position of slight supination. All forearms were immobilized in a longarm splint or cast; internal fixation was not used. All patients were followed up until clinical and/or radiographic union of the forearm bones. Evaluation Parameters A retrospective chart review identified the following preoperative data: diagnosis, age at the time of the initial osteoclasis procedure, and forearm position. Surgical notes were used to determine the method of exposing the forearm bones, osteotomy technique, and time between each stage of the procedure. Final forearm position, type and length of immobilization, and time to clinical union were recorded from the postoperative medical records. The osteotomy site was identified on the surgical radiographs. Adequate radiographs were available in 66 of 69 extremities. In calculating osteotomy location we measured the length of the forearm bones on the lateral radiograph and expressed the distance from the proximal end of the bone to the osteotomy site as a percentage of the total length. Of 132 osteotomies 128 were performed within the central three fifths of the bone—that is, 20% to 80% from the proximal end. Only 4 osteotomies were in the proximal or distal 20% of the forearm. Therefore we defined the osteotomy location as proximal when placed in the proximal 40% of the bone, middle when placed in the central 40% to 60% of the bone, and distal when placed in the distal 40% of the bone. With these zone definitions there was a satisfactory
Dalton et al / Ulnar Nonunion After Osteoclasis
distribution of osteotomies between the proximal, middle, and distal forearm. An analysis of postoperative radiographs was performed to evaluate for bony union. Radiographs typically were obtained at 2, 6, and 10 weeks after stage 2 of the osteoclasis procedure. If union was not documented by the treating physician at 10 weeks most patients were re-evaluated every 6 weeks until union or until a diagnosis of nonunion was made. Adequate postoperative radiographs were available for 61 extremities to determine the date of union. In the remaining 9 extremities the time of union was determined clearly from clinical records, which were made and documented by the treating physician based on a clinical examination and radiograph interpretation. Based on the time to union of the forearm bones extremities were subdivided into union, delayed union, and nonunion groups. The union group included patients with bridging callus across 3 cortices on anteroposterior and lateral projections of the forearm at 90 days or less. The delayed union group included patients with radiographic or clinical union occurring without a secondary procedure more than 90 days after stage II of the osteoclasis procedure. The remaining patients were placed in the nonunion group. These patients required bone grafting and internal fixation. (One patient required 2 bone-grafting procedures.) The average time from the initial procedure to grafting was 299 days (range, 175– 602 d). The final forearm position was documented from the medical records. All functional data were obtained from occupational therapists’ notes concerning forearm rotation. These data were produced by standardized measurement techniques in which rotation is determined clinically with a goniometer. A statistical analysis was performed to assess the impact of age, gender, diagnosis, preoperative forearm position, degree of rotational change, osteotomy site, and surgical technique on union rates. The statistical significance of differences between groups was obtained using analysis of variance, the Kruskall-Wallace test, or the Fisher exact test where appropriate with the -error level set at 0.05. When the overall analysis of variance was significant (p ⱕ .05), Tukey’s Standardized range test was performed to determine the pairwise comparisons of means that were significantly different. All data were analyzed with statistical software (SAS version 8.2, 1999; SAS Institute Inc., Cary, NC).
975
Results Time to Union Forty-eight of the ulnas healed within 90 days (union group). Twenty-one of the 69 extremities had a delayed union or nonunion, with 11 ulnas in the delayed union group and 10 in the nonunion group. Both forearm bones united in 48 extremities in fewer than 90 days; the average healing time in this group was 56 days (range, 31–90 d) for the radius and 57 days (range, 31– 84 d) for the ulna. In the delayed union group (11 of 69 extremities) the radius healed at an average of 100 days (range, 46 –207 d) and the ulna united at an average of 159 days (103– 305 d). The time to union of both the radius and ulna in the delayed union group was significantly longer than that in the union group (p ⬍ .001). In addtion, within the delayed union group the difference in healing times of the radius and ulna was significant (100 vs 154 d, respectively) (p ⫽ .03). The nonunion group (10 of 69 extremities) also had delayed healing of the radius in addition to nonunion of the ulna. The average time for radius healing was 136 days (range, 53–308 d). This too was significantly longer than the radiuses in the union group (p ⬍ .001). When the days to union of radiuses and ulnas of all 69 extremities were analyzed as continuous variables the healing of 1 bone correlated highly with the other (r ⫽ 0.81). Age The average age of all patients at the time of osteoclasis was 10 years (range, 2–20 y). The average ages per group were as follows: union group, 9 years (range, 2–17 y); delayed union group, 9 years (range, 4 –15 y); and nonunion group, 14 years (range, 6 – 20 y). The patients in the nonunion group were significantly older than the patients in the other 2 groups (p ⬍ .01). Gender The ulnas in 33 of the 38 males healed within 90 days compared with those in 15 of the 27 females. This difference was statistically significant (p ⫽ .02). An odds ratio showed that males were approximately 3.5 times more likely to attain union within 90 days than females (odds ratio, 3.57; 95% confidence interval, 1.22–10.4). Primary Diagnosis Birth brachial plexus palsy and congenital radioulnar synostosis were the primary diagnoses for 52 of the extremities in this study. Seventeen of the 28 extrem-
976
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
ities with birth brachial plexus palsy healed in fewer than 90 days whereas delayed union occurred in 6 of 28 extremities and nonunion in 5 of 28 extremities. In comparison 21 of 24 extremities with synostosis united within 90 days, 3 of 24 extremities united in a delayed fashion, and none had a nonunion. The difference between the 2 diagnoses was significant (p ⫽ .02). The small number of extremities with other primary diagnoses precluded statistical analysis of union rates; however, a comparison of the synostosis group with all other combined diagnostic groups also was performed. The other diagnosis groups had a combined union rate within 90 days of 60%, a delayed union rate of 18%, and a nonunion rate of 22%. Statistical comparison showed a significantly higher union rate in extremities with congenital radioulnar synostosis (p ⫽ .03). Forearm Rotational Position Before surgery 37 extremities had a pronation deformity, 31 extremities had a supination deformity, and 1 extremity was in the neutral position. A comparison of the forearm position between the union, delayed union, and nonunion groups was performed: a pronation deformity was found in 27, 6, and 4 extremities, respectively, and a supination deformity was found in 20, 5, and 4 extremities, respectively. Twenty of 31 extremities with a supination deformity united within 90 days. There was a delayed union in 5 of 31 forearms and nonunion in 6 of 31 extremities. Twenty-seven of 37 extremities with a pronation deformity united within 90 days. There was a delayed union in 6 extremities and nonunion in 4 of the extremities. These differences were not statistically significant (p ⫽ .76). The average change in forearm position after surgery was 90° (range, 40°–135°). The magnitude of positional change did not differ significantly between the 3 groups (p ⫽ .76). Osteotomy Location Based on the 66 extremity radiographs analyzed 12 ulnas had a proximal ulna osteotomy, 47 a middle osteotomy, and 7 a distal osteotomy. A proximal osteotomy had a significantly lower incidence of union in fewer than 90 days compared with middle and distal osteotomies (p ⬍ .05). None of the distal osteotomies developed nonunion compared with 4 of 12 proximal osteotomies and 8 of 47 middle osteotomies (p ⫽ .007).
Technique An analysis of the differences in stage I osteotomy technique showed that open procedures had a significantly higher rate of union within 90 days (40 of 52 extremities versus 8 of 17 extremities) and a significantly lower nonunion rate (4 of 52 extremities versus 6 of 17 extremities) compared with percutaneous procedures (p ⬍ .01). There was no difference in the rate of delayed union (8 of 52 extremities versus 3 of 17 extremities). Postoperative Care The method (long-arm cast, short-arm cast, splint) and duration of immobilization did not have a significant influence on the union rate.
Discussion In adults the nonunion rate of forearm fractures is low; reported rates vary between 2% and 3% with current fixation techniques.8 Fracture healing in children is more reliable, in part because of the thickened periosteum that contributes to the improved healing potential. Aside from stabilizing the fracture the periosteum of developing bones facilitates fracture healing through its high vascularity and osteogenic potential.9,10 Therefore nonunion of forearm fractures in children is rare, with reported nonunion rates as low as 0.03%.11 Nonunion or delayed union of the ulna is a rare complication after osteotomy as well. Ulnar-shortening osteotomy is performed most commonly in adults, with union rates afterward of 74% to 100%.12–14 In this article we report union rates after ulna osteotomy in children. The role of age in bone healing has been evaluated previously. In a review of 30 mixed-location diaphyseal nonunions in pediatric patients, Lewallen and Peterson15 concluded that nonunion was extremely rare in children younger than 10 years of age, with 27 of 30 nonunions occurring in patients between the ages of 12 and 16 years. Only 4 of the 30 nonunions, however, involved the ulna. In a review of 1,968 pediatric diaphyseal forearm fractures Adamczyk and Riley11 noted that there were only 6 nonunions and all occurred in patients between the ages of 13 and 16 years. In our investigation of pediatric forearm osteotomies the average age of patients attaining union, in either a timely or delayed fashion, was 9 years whereas in those failing to unite it was 13 years. Our patients with birth brachial plexus palsy had significantly lower union rates than those with congenital radioulnar synostosis. Although the etiology
Dalton et al / Ulnar Nonunion After Osteoclasis
of superior healing for synostosis versus brachial plexus palsy patients (88% vs 61%) is not entirely clear the role of peripheral nerve dysfunction may be important. Peripheral nerve injury has been shown to play a part in fracture healing. In a basic scientific investigation of the effects of peripheral nerve lesions on fracture healing, Aro16 showed that the union of rat tibia fractures was affected negatively by cutting the sciatic nerve. Adequate blood supply is another critical factor for bony healing. In the adult ulna the proximal diaphysis receives its blood supply from a major nutrient artery that enters the anterior surface of the bone 7.5 cm distal to the tip of the olecranon. The artery courses proximally, providing the primary blood supply to the proximal portion of the ulna. Distal to the entry point of this artery, several small perforating vessels from the anterior interosseous artery provide the primary blood supply.17 Therefore there is a relative watershed zone at the junction of the proximal and middle thirds of the ulna just distal to the nutrient artery.11,17 Fractures in this watershed zone of the ulna have been associated with poor healing. Szabo and Skinner18 reported nonunions in 7 of 28 closed, isolated ulna fractures. They determined that fractures in the proximal third of the ulna were prognostic of increased nonunion. Delayed healing of other long-bone fractures including the tibia19 and humerus20 has been noted when fracture occurs just distal to the entry of the nutrient artery. In our patients ulnar union rates decreased significantly when the osteoclasis was performed in the proximal third of the ulna. The blood supply to the proximal diaphysis of the ulna may contribute to the lower union rates of our osteotomies performed at this location. The healing time of the radius and ulna correlated highly. In the delayed union and nonunion groups the radii took significantly longer to heal when compared with those in the union group. The relationship between the healing of one bone in the forearm to the other suggests that factors affecting healing of the ulna also affect the radius in the same extremity; however, only 1 radius experienced nonunion. The reason for preferential healing of the radiuses over the ulnas in our patients is not clear. Forearm osteoclasis as described originally by Manske et al2 and reported here does not use any internal or external fixation. As designed originally the procedure avoided internal fixation to avoid complications such as pin track infections and the need to remove hardware. The high union rate reported initially2 suggested that the hardware was not neces-
977
sary. Other reported forearm osteoclasis techniques7 use internal fixation in the form of a single K-wire across the osteotomy site with a single-stage correction. A report7 of 4 extremities treated with this technique showed healing of all osteotomies within 8 weeks. The small number of extremities in this series does not allow a conclusion to be drawn concerning the use of limited fixation. The current study has several weaknesses. First, although most radiographs were available for review a few were not. The assessments of the radiographic healing status at the time of treatment as performed by both orthopedic surgeons and radiologists, however, were available in the medical records and there was agreement in all cases between the radiologist and the orthopedic surgeon. In addition, despite general adherence to the originally reported surgical technique there were slight variances including the method of cortical perforation (K-wire or drill) and the method of osteoclasis (manual or with an osteotome). Separation of the extremities based on these differences in methods produced small subgroups, which precluded meaningful statistical analysis. It might be expected that because percutaneous procedures are less invasive they would have better healing; however, this was not the case as we found a significantly higher union rate after open procedures. Finally, because this was a retrospective study based on medical record and radiograph reviews functional data were not obtained. Functional limitations are the primary indication for osteoclasis but based on the data acquired for this study we cannot comment on the functional changes provided by the surgery. Although the purpose of this study did not include an evaluation of functional results we believe that osteoclasis remains an effective technique for addressing forearm rotational deformities. A midshaft or distal osteotomy should lead to more reliable healing. The risk of nonunion in older patients and patients with birth brachial plexus palsy must be considered. No other complications were noted in this group of 66 patients. The neurovascular injuries reported with other techniques including radial and ulnar nerve injury, ischemic contracture, and circulatory embarrassment3,6,21,22 were not seen in this patient population. Received for publication August 8, 2005; accepted in revised form March 15, 2006. 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.
978
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
Corresponding author: Charles A. Goldfarb, MD, Department of Orthopaedic Surgery, School of Medicine, Washington University in St. Louis, One Barnes-Jewish Hospital Plaza, Ste 11300 West Pavilion, St. Louis, MO 63110; e-mail: [email protected]. Copyright © 2006 by the American Society for Surgery of the Hand 0363-5023/06/31A06-0016$32.00/0 doi:10.1016/j.jhsa.2006.03.006
References 1. Lin HH, Strecker WB, Manske PR, Schoenecker PL, Seyer DM. A surgical technique of radioulnar osteoclasis to correct severe forearm rotation deformities. J Pediatr Orthop 1995; 15:53–58. 2. Manske PR, McCarroll HR Jr, Hale R. Biceps tendon rerouting and percutaneous osteoclasis in the treatment of supination deformity in obstetrical palsy. J Hand Surg 1980; 5:153–159. 3. Simmons BP, Southmayd WW, Riseborough EJ. Congenital radioulnar synostosis. J Hand Surg 1983;8:829 – 838. 4. Kanaya F, Ibaraki K. Mobilization of a congenital proximal radioulnar synostosis with use of a free vascularized fasciofat graft. J Bone Joint Surg 1998;80A:1186 –1192. 5. Sugimoto M, Masada K, Ohno H, Hosoya T. Treatment of traumatic radioulnar synostosis by excision, with interposition of a posterior interosseous island forearm flap. J Hand Surg 1996;21B:393–395. 6. Green WT, Mital MA. Congenital radio-ulnar synostosis: surgical treatment. J Bone Joint Surg 1979;61A:738 –743. 7. Murase T, Tada K, Yoshida T, Moritomo H. Derotational osteotomy at the shafts of the radius and ulna for congenital radioulnar synostosis. J Hand Surg 2003;28A:133–137. 8. Anderson LD, Meyer FN. Nonunion of the diaphysis of the radius and ulna. Instr Course Lect 1988;37:157–159. 9. Lindaman LM. Bone healing in children. Clin Podiatr Med Surg 2001;18:97–108. 10. Rodriguez-Merchan EC. Pediatric skeletal trauma: a review and historical perspective. Clin Orthop 2005;432:8 –13.
11. Adamczyk MJ, Riley PM. Delayed union and nonunion following closed treatment of diaphyseal pediatric forearm fractures. J Pediatr Orthop 2005;25:51–55. 12. Darrow JC Jr, Linscheid RL, Dobyns JH, Mann JM III, Wood MB, Beckenbaugh RD. Distal ulnar recession for disorders of the distal radioulnar joint. J Hand Surg 1985; 10A:482– 491. 13. Minami A, Ogino T, Minami M. Treatment of distal radioulnar disorders. J Hand Surg 1987;12A:189 –196. 14. Rayhack JM, Gasser SI, Latta LL, Ouellette EA, Milne EL. Precision oblique osteotomy for shortening of the ulna. J Hand Surg 1993;18A:908 –918. 15. Lewallen RP, Peterson HA. Nonunion of long bone fractures in children: a review of 30 cases. J Pediatr Orthop 1985;5: 135–142. 16. Aro H. Effect of nerve injury on fracture healing. Callus formation studied in the rat. Acta Orthop Scand 1985;56: 233–237. 17. Wright TW, Glowczewskie F. Vascular anatomy of the ulna. J Hand Surg 1998;23A:800 – 804. 18. Szabo RM, Skinner M. Isolated ulnar shaft fractures. Retrospective study of 46 cases. Acta Orthop Scand 1990;61:350 – 352. 19. Nelson GE Jr, Kelly PJ, Peterson LF, Janes JM. Blood supply of the human tibia. Am J Orthop 1960;42A:625– 636. 20. Esterhai JL Jr, Brighton CT, Heppenstall RB, Thrower A. Nonunion of the humerus. Clinical, roentgenographic, scintigraphic, and response characteristics to treatment with constant direct current stimulation of osteogenesis. Clin Orthop 1986;211:228 –234. 21. Hankin FM, Smith PA, Kling TF Jr, Louis DS. Ulnar nerve palsy following rotational osteotomy of congenital radioulnar synostosis. J Pediatr Orthop 1987;7:103–106. 22. Ogino T, Hikino K. Congenital radio-ulnar synostosis: compensatory rotation around the wrist and rotation osteotomy. J Hand Surg 1987;12B:173–178.
Purpose: Osteoclasis, a minimally invasive technique to rotate the radius and ulna, is used commonly to correct forearm rotational deformities in children. The purpose of this investigation was to evaluate objectively osteotomy healing in patients treated with osteoclasis, with specific attention given to the risk for nonunion. Methods: We identified 69 extremities in 65 children treated with osteoclasis and performed retrospective chart and radiographic reviews to evaluate the time to union of the radius and ulna and factors influencing healing. Results: The average rotational correction was 90°. Twenty-one ulnas had either delayed union or nonunion. Forty-eight of the forearms healed in less than 3 months. Factors correlated with a significantly decreased union rate included increased patient age, percutaneous technique, osteoclasis site in the proximal ulna, and primary diagnoses other than congenital radioulnar synostosis. Preoperative forearm position, magnitude of position correction, and treatment of the periosteum were not associated with changes in union rates. Conclusions: Forearm osteoclasis has a delayed union rate of 16%. Timely union of the ulna appears to be influenced by both patient-centered factors and surgical technique. (J Hand Surg 2006;31A:973–978. Copyright © 2006 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Therapeutic, Level IV. Key words: Ulna nonunion, osteoclasis, osteotomy, forearm, synostosis.
arked rotational deformity of the forearm in children and adolescents can limit upper-extremity use and surgery may be indicated to improve patient function. Although occurring most commonly in children with congenital radioulnar synostosis, rotational abnormalities also are seen in patients with brachial plexus palsy, cerebral palsy, and arthrogryposis. In patients with a fixed or severe rotational deformity of the forearm osteoclasis is a commonly used technique for correction.1,2 Although a variety of other procedures have been described— including osteotomy of only the radius,3 resection of the proximal radioulnar synostosis,4,5 osteoclasis through the radioulnar synostosis,3,6 or radius and ulna osteotomy with internal fixation7—surgeons use osteoclasis because of its surgical sim-
M
plicity and the previously reported outstanding results.1,2 For more than 40 years we have used 2-stage osteoclasis to correct marked rotational abnormalities of the forearm.2 In stage I a minimally invasive osteotomy of both the radius and ulna is performed; the extremity then is splinted in the preosteotomy position. In stage II (⬇10 days after stage I to allow for early callus formation) the forearm is rotated into the desired position and held in a long-arm (above elbow) cast until healed. No internal or external fixation is used. In 1995 Lin et al1 reported the outcome of 23 patients with forearm rotational deformity as a result of neuromuscular, developmental, or congenital problems treated with this osteoclasis technique. Good or excellent results were achieved in more than 90% of their patients with an average correction of The Journal of Hand Surgery
973
974
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
81° and 69° for pronation and supination deformities, respectively. Three complications were reported: 2 nonunions and 1 recurrence of the deformity. The functional results of the osteoclasis procedure continue to be satisfactory. The purpose of this investigation was to evaluate a larger cohort of patients treated with this osteoclasis method with specific attention given to the risk for nonunion and the factors affecting healing.
Materials and Methods This investigation was performed at the Shriners Hospital for Children. After obtaining institutional review board approval we used an operating room logbook to identify all patients who had a 2-staged forearm osteoclasis performed between 1970 and 2004. The patients identified through this search were evaluated further with thorough chart and radiographic reviews. Sixty-five patients who had 69 osteoclases by 1 of 3 attending surgeons were identified. There were 38 male and 27 female patients. Osteoclasis was performed at an average age of 10 years (range, 2–20 y). The diagnosis was obstetric brachial plexus palsy for 28 extremities, congenital radioulnar synostosis for 24 extremities, spastic diplegia for 6 extremities, arthrogryposis for 5 extremities, multiple osteochondromatosis for 2 extremities, and fracture malunion, Maffucci’s enchondromatosis, Conradi-Hünermann syndrome, and Volkmann’s ischemic contracture for 1 extremity each. All patients who had forearm osteoclasis by the general methods described by Manske et al2 for rotational deformity of the forearm were included. Any patient with functional limitations caused by the forearm position was eligible for inclusion. Most patients had functional limitations attributable to a fixed pronation or supination deformity of more than 60°. For example patients with fixed pronation deformities have difficulty receiving objects in the palm (ie, holding coins). Patients with fixed supination deformities may accommodate by abducting the shoulder but still may have difficulty using a keyboard in a comfortable position. Patients with either supination or pronation deformities have difficulty using the affected extremity as a helper hand for 2-handed activities such as carrying large objects. A few patients had bilateral deformities in which both forearms were fixed in the same position. Patients who had rotational osteotomy of only the radius and patients treated by any other method than osteoclasis were excluded. Differences in osteoclasis technique
(including osteotomy site and specific osteotomy technique) were based purely on the preference of the attending surgeon. Procedure In 17 of 69 extremities stage I of the procedure was performed by percutaneously drilling the forearm bone cortices with a K-wire and the bones were fractured by manipulation. In 52 extremities stage I was performed open, accessing both forearm bones through small (1- to 2-cm) incisions; the cortices were drilled and the fracture was completed either by manipulation or with an osteotome. After stage I all forearms were splinted in the preoperative position. Stage 2 was performed an average of 10 days (range, 6 –19 d) after stage I. The forearm was rotated through the osteoclasis site into the desired position as determined by the attending surgeon. The position was based on the concept of placing the dominant extremity in a position of slight pronation and placing the nondominant extremity in a position of slight supination. All forearms were immobilized in a longarm splint or cast; internal fixation was not used. All patients were followed up until clinical and/or radiographic union of the forearm bones. Evaluation Parameters A retrospective chart review identified the following preoperative data: diagnosis, age at the time of the initial osteoclasis procedure, and forearm position. Surgical notes were used to determine the method of exposing the forearm bones, osteotomy technique, and time between each stage of the procedure. Final forearm position, type and length of immobilization, and time to clinical union were recorded from the postoperative medical records. The osteotomy site was identified on the surgical radiographs. Adequate radiographs were available in 66 of 69 extremities. In calculating osteotomy location we measured the length of the forearm bones on the lateral radiograph and expressed the distance from the proximal end of the bone to the osteotomy site as a percentage of the total length. Of 132 osteotomies 128 were performed within the central three fifths of the bone—that is, 20% to 80% from the proximal end. Only 4 osteotomies were in the proximal or distal 20% of the forearm. Therefore we defined the osteotomy location as proximal when placed in the proximal 40% of the bone, middle when placed in the central 40% to 60% of the bone, and distal when placed in the distal 40% of the bone. With these zone definitions there was a satisfactory
Dalton et al / Ulnar Nonunion After Osteoclasis
distribution of osteotomies between the proximal, middle, and distal forearm. An analysis of postoperative radiographs was performed to evaluate for bony union. Radiographs typically were obtained at 2, 6, and 10 weeks after stage 2 of the osteoclasis procedure. If union was not documented by the treating physician at 10 weeks most patients were re-evaluated every 6 weeks until union or until a diagnosis of nonunion was made. Adequate postoperative radiographs were available for 61 extremities to determine the date of union. In the remaining 9 extremities the time of union was determined clearly from clinical records, which were made and documented by the treating physician based on a clinical examination and radiograph interpretation. Based on the time to union of the forearm bones extremities were subdivided into union, delayed union, and nonunion groups. The union group included patients with bridging callus across 3 cortices on anteroposterior and lateral projections of the forearm at 90 days or less. The delayed union group included patients with radiographic or clinical union occurring without a secondary procedure more than 90 days after stage II of the osteoclasis procedure. The remaining patients were placed in the nonunion group. These patients required bone grafting and internal fixation. (One patient required 2 bone-grafting procedures.) The average time from the initial procedure to grafting was 299 days (range, 175– 602 d). The final forearm position was documented from the medical records. All functional data were obtained from occupational therapists’ notes concerning forearm rotation. These data were produced by standardized measurement techniques in which rotation is determined clinically with a goniometer. A statistical analysis was performed to assess the impact of age, gender, diagnosis, preoperative forearm position, degree of rotational change, osteotomy site, and surgical technique on union rates. The statistical significance of differences between groups was obtained using analysis of variance, the Kruskall-Wallace test, or the Fisher exact test where appropriate with the -error level set at 0.05. When the overall analysis of variance was significant (p ⱕ .05), Tukey’s Standardized range test was performed to determine the pairwise comparisons of means that were significantly different. All data were analyzed with statistical software (SAS version 8.2, 1999; SAS Institute Inc., Cary, NC).
975
Results Time to Union Forty-eight of the ulnas healed within 90 days (union group). Twenty-one of the 69 extremities had a delayed union or nonunion, with 11 ulnas in the delayed union group and 10 in the nonunion group. Both forearm bones united in 48 extremities in fewer than 90 days; the average healing time in this group was 56 days (range, 31–90 d) for the radius and 57 days (range, 31– 84 d) for the ulna. In the delayed union group (11 of 69 extremities) the radius healed at an average of 100 days (range, 46 –207 d) and the ulna united at an average of 159 days (103– 305 d). The time to union of both the radius and ulna in the delayed union group was significantly longer than that in the union group (p ⬍ .001). In addtion, within the delayed union group the difference in healing times of the radius and ulna was significant (100 vs 154 d, respectively) (p ⫽ .03). The nonunion group (10 of 69 extremities) also had delayed healing of the radius in addition to nonunion of the ulna. The average time for radius healing was 136 days (range, 53–308 d). This too was significantly longer than the radiuses in the union group (p ⬍ .001). When the days to union of radiuses and ulnas of all 69 extremities were analyzed as continuous variables the healing of 1 bone correlated highly with the other (r ⫽ 0.81). Age The average age of all patients at the time of osteoclasis was 10 years (range, 2–20 y). The average ages per group were as follows: union group, 9 years (range, 2–17 y); delayed union group, 9 years (range, 4 –15 y); and nonunion group, 14 years (range, 6 – 20 y). The patients in the nonunion group were significantly older than the patients in the other 2 groups (p ⬍ .01). Gender The ulnas in 33 of the 38 males healed within 90 days compared with those in 15 of the 27 females. This difference was statistically significant (p ⫽ .02). An odds ratio showed that males were approximately 3.5 times more likely to attain union within 90 days than females (odds ratio, 3.57; 95% confidence interval, 1.22–10.4). Primary Diagnosis Birth brachial plexus palsy and congenital radioulnar synostosis were the primary diagnoses for 52 of the extremities in this study. Seventeen of the 28 extrem-
976
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
ities with birth brachial plexus palsy healed in fewer than 90 days whereas delayed union occurred in 6 of 28 extremities and nonunion in 5 of 28 extremities. In comparison 21 of 24 extremities with synostosis united within 90 days, 3 of 24 extremities united in a delayed fashion, and none had a nonunion. The difference between the 2 diagnoses was significant (p ⫽ .02). The small number of extremities with other primary diagnoses precluded statistical analysis of union rates; however, a comparison of the synostosis group with all other combined diagnostic groups also was performed. The other diagnosis groups had a combined union rate within 90 days of 60%, a delayed union rate of 18%, and a nonunion rate of 22%. Statistical comparison showed a significantly higher union rate in extremities with congenital radioulnar synostosis (p ⫽ .03). Forearm Rotational Position Before surgery 37 extremities had a pronation deformity, 31 extremities had a supination deformity, and 1 extremity was in the neutral position. A comparison of the forearm position between the union, delayed union, and nonunion groups was performed: a pronation deformity was found in 27, 6, and 4 extremities, respectively, and a supination deformity was found in 20, 5, and 4 extremities, respectively. Twenty of 31 extremities with a supination deformity united within 90 days. There was a delayed union in 5 of 31 forearms and nonunion in 6 of 31 extremities. Twenty-seven of 37 extremities with a pronation deformity united within 90 days. There was a delayed union in 6 extremities and nonunion in 4 of the extremities. These differences were not statistically significant (p ⫽ .76). The average change in forearm position after surgery was 90° (range, 40°–135°). The magnitude of positional change did not differ significantly between the 3 groups (p ⫽ .76). Osteotomy Location Based on the 66 extremity radiographs analyzed 12 ulnas had a proximal ulna osteotomy, 47 a middle osteotomy, and 7 a distal osteotomy. A proximal osteotomy had a significantly lower incidence of union in fewer than 90 days compared with middle and distal osteotomies (p ⬍ .05). None of the distal osteotomies developed nonunion compared with 4 of 12 proximal osteotomies and 8 of 47 middle osteotomies (p ⫽ .007).
Technique An analysis of the differences in stage I osteotomy technique showed that open procedures had a significantly higher rate of union within 90 days (40 of 52 extremities versus 8 of 17 extremities) and a significantly lower nonunion rate (4 of 52 extremities versus 6 of 17 extremities) compared with percutaneous procedures (p ⬍ .01). There was no difference in the rate of delayed union (8 of 52 extremities versus 3 of 17 extremities). Postoperative Care The method (long-arm cast, short-arm cast, splint) and duration of immobilization did not have a significant influence on the union rate.
Discussion In adults the nonunion rate of forearm fractures is low; reported rates vary between 2% and 3% with current fixation techniques.8 Fracture healing in children is more reliable, in part because of the thickened periosteum that contributes to the improved healing potential. Aside from stabilizing the fracture the periosteum of developing bones facilitates fracture healing through its high vascularity and osteogenic potential.9,10 Therefore nonunion of forearm fractures in children is rare, with reported nonunion rates as low as 0.03%.11 Nonunion or delayed union of the ulna is a rare complication after osteotomy as well. Ulnar-shortening osteotomy is performed most commonly in adults, with union rates afterward of 74% to 100%.12–14 In this article we report union rates after ulna osteotomy in children. The role of age in bone healing has been evaluated previously. In a review of 30 mixed-location diaphyseal nonunions in pediatric patients, Lewallen and Peterson15 concluded that nonunion was extremely rare in children younger than 10 years of age, with 27 of 30 nonunions occurring in patients between the ages of 12 and 16 years. Only 4 of the 30 nonunions, however, involved the ulna. In a review of 1,968 pediatric diaphyseal forearm fractures Adamczyk and Riley11 noted that there were only 6 nonunions and all occurred in patients between the ages of 13 and 16 years. In our investigation of pediatric forearm osteotomies the average age of patients attaining union, in either a timely or delayed fashion, was 9 years whereas in those failing to unite it was 13 years. Our patients with birth brachial plexus palsy had significantly lower union rates than those with congenital radioulnar synostosis. Although the etiology
Dalton et al / Ulnar Nonunion After Osteoclasis
of superior healing for synostosis versus brachial plexus palsy patients (88% vs 61%) is not entirely clear the role of peripheral nerve dysfunction may be important. Peripheral nerve injury has been shown to play a part in fracture healing. In a basic scientific investigation of the effects of peripheral nerve lesions on fracture healing, Aro16 showed that the union of rat tibia fractures was affected negatively by cutting the sciatic nerve. Adequate blood supply is another critical factor for bony healing. In the adult ulna the proximal diaphysis receives its blood supply from a major nutrient artery that enters the anterior surface of the bone 7.5 cm distal to the tip of the olecranon. The artery courses proximally, providing the primary blood supply to the proximal portion of the ulna. Distal to the entry point of this artery, several small perforating vessels from the anterior interosseous artery provide the primary blood supply.17 Therefore there is a relative watershed zone at the junction of the proximal and middle thirds of the ulna just distal to the nutrient artery.11,17 Fractures in this watershed zone of the ulna have been associated with poor healing. Szabo and Skinner18 reported nonunions in 7 of 28 closed, isolated ulna fractures. They determined that fractures in the proximal third of the ulna were prognostic of increased nonunion. Delayed healing of other long-bone fractures including the tibia19 and humerus20 has been noted when fracture occurs just distal to the entry of the nutrient artery. In our patients ulnar union rates decreased significantly when the osteoclasis was performed in the proximal third of the ulna. The blood supply to the proximal diaphysis of the ulna may contribute to the lower union rates of our osteotomies performed at this location. The healing time of the radius and ulna correlated highly. In the delayed union and nonunion groups the radii took significantly longer to heal when compared with those in the union group. The relationship between the healing of one bone in the forearm to the other suggests that factors affecting healing of the ulna also affect the radius in the same extremity; however, only 1 radius experienced nonunion. The reason for preferential healing of the radiuses over the ulnas in our patients is not clear. Forearm osteoclasis as described originally by Manske et al2 and reported here does not use any internal or external fixation. As designed originally the procedure avoided internal fixation to avoid complications such as pin track infections and the need to remove hardware. The high union rate reported initially2 suggested that the hardware was not neces-
977
sary. Other reported forearm osteoclasis techniques7 use internal fixation in the form of a single K-wire across the osteotomy site with a single-stage correction. A report7 of 4 extremities treated with this technique showed healing of all osteotomies within 8 weeks. The small number of extremities in this series does not allow a conclusion to be drawn concerning the use of limited fixation. The current study has several weaknesses. First, although most radiographs were available for review a few were not. The assessments of the radiographic healing status at the time of treatment as performed by both orthopedic surgeons and radiologists, however, were available in the medical records and there was agreement in all cases between the radiologist and the orthopedic surgeon. In addition, despite general adherence to the originally reported surgical technique there were slight variances including the method of cortical perforation (K-wire or drill) and the method of osteoclasis (manual or with an osteotome). Separation of the extremities based on these differences in methods produced small subgroups, which precluded meaningful statistical analysis. It might be expected that because percutaneous procedures are less invasive they would have better healing; however, this was not the case as we found a significantly higher union rate after open procedures. Finally, because this was a retrospective study based on medical record and radiograph reviews functional data were not obtained. Functional limitations are the primary indication for osteoclasis but based on the data acquired for this study we cannot comment on the functional changes provided by the surgery. Although the purpose of this study did not include an evaluation of functional results we believe that osteoclasis remains an effective technique for addressing forearm rotational deformities. A midshaft or distal osteotomy should lead to more reliable healing. The risk of nonunion in older patients and patients with birth brachial plexus palsy must be considered. No other complications were noted in this group of 66 patients. The neurovascular injuries reported with other techniques including radial and ulnar nerve injury, ischemic contracture, and circulatory embarrassment3,6,21,22 were not seen in this patient population. Received for publication August 8, 2005; accepted in revised form March 15, 2006. 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.
978
The Journal of Hand Surgery / Vol. 31A No. 6 July–August 2006
Corresponding author: Charles A. Goldfarb, MD, Department of Orthopaedic Surgery, School of Medicine, Washington University in St. Louis, One Barnes-Jewish Hospital Plaza, Ste 11300 West Pavilion, St. Louis, MO 63110; e-mail: [email protected]. Copyright © 2006 by the American Society for Surgery of the Hand 0363-5023/06/31A06-0016$32.00/0 doi:10.1016/j.jhsa.2006.03.006
References 1. Lin HH, Strecker WB, Manske PR, Schoenecker PL, Seyer DM. A surgical technique of radioulnar osteoclasis to correct severe forearm rotation deformities. J Pediatr Orthop 1995; 15:53–58. 2. Manske PR, McCarroll HR Jr, Hale R. Biceps tendon rerouting and percutaneous osteoclasis in the treatment of supination deformity in obstetrical palsy. J Hand Surg 1980; 5:153–159. 3. Simmons BP, Southmayd WW, Riseborough EJ. Congenital radioulnar synostosis. J Hand Surg 1983;8:829 – 838. 4. Kanaya F, Ibaraki K. Mobilization of a congenital proximal radioulnar synostosis with use of a free vascularized fasciofat graft. J Bone Joint Surg 1998;80A:1186 –1192. 5. Sugimoto M, Masada K, Ohno H, Hosoya T. Treatment of traumatic radioulnar synostosis by excision, with interposition of a posterior interosseous island forearm flap. J Hand Surg 1996;21B:393–395. 6. Green WT, Mital MA. Congenital radio-ulnar synostosis: surgical treatment. J Bone Joint Surg 1979;61A:738 –743. 7. Murase T, Tada K, Yoshida T, Moritomo H. Derotational osteotomy at the shafts of the radius and ulna for congenital radioulnar synostosis. J Hand Surg 2003;28A:133–137. 8. Anderson LD, Meyer FN. Nonunion of the diaphysis of the radius and ulna. Instr Course Lect 1988;37:157–159. 9. Lindaman LM. Bone healing in children. Clin Podiatr Med Surg 2001;18:97–108. 10. Rodriguez-Merchan EC. Pediatric skeletal trauma: a review and historical perspective. Clin Orthop 2005;432:8 –13.
11. Adamczyk MJ, Riley PM. Delayed union and nonunion following closed treatment of diaphyseal pediatric forearm fractures. J Pediatr Orthop 2005;25:51–55. 12. Darrow JC Jr, Linscheid RL, Dobyns JH, Mann JM III, Wood MB, Beckenbaugh RD. Distal ulnar recession for disorders of the distal radioulnar joint. J Hand Surg 1985; 10A:482– 491. 13. Minami A, Ogino T, Minami M. Treatment of distal radioulnar disorders. J Hand Surg 1987;12A:189 –196. 14. Rayhack JM, Gasser SI, Latta LL, Ouellette EA, Milne EL. Precision oblique osteotomy for shortening of the ulna. J Hand Surg 1993;18A:908 –918. 15. Lewallen RP, Peterson HA. Nonunion of long bone fractures in children: a review of 30 cases. J Pediatr Orthop 1985;5: 135–142. 16. Aro H. Effect of nerve injury on fracture healing. Callus formation studied in the rat. Acta Orthop Scand 1985;56: 233–237. 17. Wright TW, Glowczewskie F. Vascular anatomy of the ulna. J Hand Surg 1998;23A:800 – 804. 18. Szabo RM, Skinner M. Isolated ulnar shaft fractures. Retrospective study of 46 cases. Acta Orthop Scand 1990;61:350 – 352. 19. Nelson GE Jr, Kelly PJ, Peterson LF, Janes JM. Blood supply of the human tibia. Am J Orthop 1960;42A:625– 636. 20. Esterhai JL Jr, Brighton CT, Heppenstall RB, Thrower A. Nonunion of the humerus. Clinical, roentgenographic, scintigraphic, and response characteristics to treatment with constant direct current stimulation of osteogenesis. Clin Orthop 1986;211:228 –234. 21. Hankin FM, Smith PA, Kling TF Jr, Louis DS. Ulnar nerve palsy following rotational osteotomy of congenital radioulnar synostosis. J Pediatr Orthop 1987;7:103–106. 22. Ogino T, Hikino K. Congenital radio-ulnar synostosis: compensatory rotation around the wrist and rotation osteotomy. J Hand Surg 1987;12B:173–178.