Congenital pseudarthrosis of the tibia is a rare disorder that is difficult to treat successfully. Despite varied initial clinical presentations, this condition results in a pathologic fracture in nearly all cases. Frequently, there is an accompanying fracture of the fibula. Both fractures are recalcitrant to standard treatment, and patients with this condition often require multiple surgical procedures in an effort to achieve fracture union and a functional extremity. Historically, many patients eventually have had painful low-functioning limbs or even amputation1-3.
Numerous treatment methods have been used to obtain and maintain healing of congenital pseudarthrosis of the tibia. Intramedullary fixation1,4,5, free vascularized fibular graft6,7, and Ilizarov bone transport8,9 have all been used, with union rates ranging from 14% to 86%1,2,4,8,9. The results of intramedullary fixation that were reported in two recent studies were encouraging. In a retrospective review in which the Charnley-Williams procedure was used, Johnston defined important variables of the technique that lead to more consistent union rates5. Concomitant fibular osteotomies, with or without fixation, yielded superior clinical and radiographic results compared with leaving the fibula intact. In addition, Dobbs et al. reported the long-term results for twenty-one consecutive patients who were managed with the Williams rod technique4. Union occurred in 86% of the patients, but healing took an extended period of time (average, sixteen months), and there was a 57% rate of refracture after a mean duration of follow-up of 14.2 years.
Recently, two types of recombinant human bone morphogenetic protein (rhBMP) have been approved for the clinical treatment of fractures in specific situations10,11. Multiple studies have demonstrated the safety and efficacy of rhBMPs when used for the treatment of open tibial fractures in adults as well as for the treatment of tibial nonunions12-14. Thus far, there are limited data regarding the use of rhBMP for the treatment of congenital pseudarthrosis of the tibia, with recent studies specifically investigating BMP-714-16.
The purpose of the present study was twofold. The first purpose was to examine the effect of rhBMP-2 on congenital pseudarthrosis of the tibia when combined with the Williams intramedullary fixation technique. We hypothesized that the addition of rhBMP-2 would improve the union rate and would decrease the time to union. The second purpose was to determine if any detrimental effects secondary to the use of rhBMP-2 or the collagen sponge carrier became clinically apparent in this population of skeletally immature patients over a minimum duration of follow-up of four years.
We performed a retrospective review of our first seven patients in which rhBMP-2 (INFUSE; Medtronic Sofamor Danek, Memphis, Tennessee) was used in addition to the Williams intramedullary rod technique for the treatment of congenital pseudarthrosis of the tibia. Institutional review board approval was obtained for the present study. Compassionate use exemption for rhBMP-2 was obtained from the manufacturer for each procedure performed prior to United States Food and Drug Administration approval of the product for specific conditions.
The patients were managed between November 1998 and March 2004. The mean age of the patients at the time of surgery was 6.8 years (range, 1.9 to 11.7 years). The patients included five girls and two boys. Four patients had neurofibromatosis. All patients had an established pseudarthrosis of the tibia after a fracture. Two patients (Cases 1 and 5) had previous surgical treatment of the pseudarthrosis with intramedullary rod fixation of the tibia and bone-grafting (Table I). Union occurred in one of these patients (Case 5) with previous treatment, but a refracture occurred 4.5 years later, when the patient was nine years old.
Surgical Technique
A similar surgical technique was used by four different surgeons (including one of the authors [B.S.R.]) for the management of the seven patients. With use of a tourniquet to control bleeding, a longitudinal anterior approach to the tibial pseudarthrosis site was made. The soft tissue within the pseudarthrosis site was excised, and the bone edges were débrided. With use of a separate lateral incision, the fibula was osteotomized to allow complete apposition of the ends of the tibia. The fibular osteotomy was performed at the level of the tibial defect or the level of a fibular pseudarthrosis, if present. The medullary canal of the distal tibial fragment was drilled with sequentially larger drill bits to allow it to accept either a 1/8-in (0.32-cm) or 3/16-in (0.48-cm) intramedullary rod. The male end of the assembled Williams rod was then introduced into the distal tibial fragment and was advanced in an antegrade fashion, exiting through the distal part of the tibia, the talus, the calcaneus, and the sole of the foot, with the ankle and foot held in a neutral position. The female end of the rod was then advanced in a retrograde fashion into the proximal fragment, stopping distal to the proximal tibial physis. The Williams rod was uncoupled, and the intramedullary portion of the rod was left in place. The distal extent of the tibial intramedullary rod was left to the discretion of the surgeon, with the options being either transfixation of the ankle joint with the rod or leaving the rod end in the distal tibial metaphysis. In cases in which the ankle joint was not transfixed, the Williams rod had been machined preoperatively to allow distal interlocking for better stabilization of the distal tibial fragment.
The BMP-2 was prepared, as per the manufacturer's directions, with bovine collagen sponges. These BMP-2 saturated collagen sponges were carefully placed around the tibial defect site. Before the sponges were wrapped completely around the tibia, the defect was filled with autogenous bone graft that had been harvested from the iliac crest (Figs. 1-A through 1-D). The subcutaneous tissues were then closed over the sponge with use of interrupted sutures. The fibular osteotomy site was fixed with an intramedullary Steinmann pin when possible.
Postoperatively, the patient was managed with a non-weight-bearing above-the-knee cast with the knee held in flexion for three months. The patient was then fit with an ankle-foot orthosis that included anterior and posterior shells and was allowed to begin protected weight-bearing. With evidence of radiographic healing, the patient was allowed increased activities. When the tibia was healed and remodeled, the patient was given the choice to discontinue the use of the ankle-foot-orthosis or, if he or she felt the need, to continue using the orthosis for protective support.
Surgical Treatment
All seven patients were managed with a Williams rod. In four patients, the rod crossed the ankle joint and ended in the talus or calcaneus. Six patients had concomitant fibular osteotomies with intramedullary fixation. The one patient who did not undergo a fibular osteotomy had such severe preoperative fibular resorption at the pseudarthrosis site that an osteotomy was unnecessary and intramedullary fibular fixation was impossible. BMP-2 dosing was left to the discretion of the surgeon, with the dose ranging from 4.2 to 12 mg. The average BMP-2 dose was 8.5 mg (Table I).
Radiographic and Clinical Assessment
Postoperatively, anteroposterior and lateral radiographs of the involved extremity were made at approximately six-week intervals. These radiographs were reviewed to determine radiographic healing, defined as evidence of bridging across the transverse tibial cortical defects. Clinical office notes were reviewed to determine functional outcome according to the criteria defined by Johnston, in which a grade-1 outcome was characterized by unequivocal union with maintenance of alignment without additional surgical treatment, a grade-2 outcome was characterized by equivocal union and/or deformity for which additional surgery was required or anticipated, and a grade-3 outcome was characterized by persistent nonunion or refracture, pain or instability of the limb, and a need for full-time orthotic support5. Knee and ankle motion, activity tolerance, foot position, residual limb-length discrepancy, and postoperative complications were also recorded for each patient.
Source of Funding
There was no external funding source.
Primary surgical intervention augmented with rhBMP-2 resulted in radiographic union of the pseudarthrosis in five of the seven patients at an average of 6.4 months (range, 3.7 to 8.1 months) postoperatively (Figs. 2-A through 2-C). The primary functional outcome was classified as grade 1 for five patients, grade 2 for one, and grade 3 for one (Table II). The patient with a grade-2 outcome (Case 6) was found to have limited radiographic signs of healing at the pseudarthrosis site with persistence of a transverse cortical defect at eleven months after surgery. This patient, in whom the intramedullary rod crossed the ankle joint, complained of heel pain at that time. A radiograph revealed that the intramedullary rod had migrated distally to a subcutaneous position in the heel. The patient underwent a rod exchange with a customized Williams rod that allowed proximal interlocking. Repeat autogenous bone-grafting with BMP-2 augmentation was performed. The pseudarthrosis showed radiographic signs of healing six weeks later. The patient with a grade-3 outcome (Case 2) had received the largest dose of BMP (12 mg) among all of the patients according to body weight. She had a persistent painful nonunion despite repeat bone-grafting after the index procedure, and the parents subsequently elected for her to have a below-the-knee amputation 2.8 years after the index procedure.
Of the five patients with radiographic evidence of union after the initial procedure, four patients had maintenance of the union at an average of 5.4 years (range, 3.9 to 7.3 years). One patient (Case 3) sustained an ipsilateral tibial fracture distal to the healed pseudarthrosis site 2.5 years after the primary intervention. This patient initially had been managed with a customized Williams rod with distal interlocking without transfixation of the ankle. She sustained the new tibial fracture, with rod breakage, through one of the distal interlock holes in the rod. She subsequently underwent rod revision with ankle transfixation, autogenous bone-grafting, and repeat BMP-2 augmentation at the new fracture site, and the new fracture healed.
After an average duration of follow-up of 6.0 years (range, 4.0 to 9.0 years), five patients were pain-free without activity limitations and two patients reported mild ankle pain. Four patients did not use orthoses, two patients wore protective orthoses to varying degrees, and one patient was active with a below-the-knee amputation prosthesis, Full knee motion was reported for all patients. Ankle motion was full in two patients who did not undergo fixation across the ankle joint, and it was mildly limited in another patient, but three patients had no ankle motion. All feet were in a plantigrade position (Table III).
After a minimum duration of follow-up of four years, no intermediate-term deleterious effects related to the use of BMP-2 or the collagen sponges have been recognized in any patient. There were no instances of disturbance of normal physeal growth, development of primary or secondary tumors in the patients (with or without neurofibromatosis), localized tissue reactions at the site of the BMP-soaked collagen sponge, allergic reactions in patients who underwent repeat applications of the BMP-soaked collagen sponge, or infections.
Congenital pseudarthrosis of the tibia remains a major treatment challenge, and there is no consensus on the optimum treatment. The ultimate goals are to achieve bone-healing that is stable enough to allow functional use of the extremity and to minimize the risk of refracture. In the present small series, our results were promising, with healing of the pseudarthrosis at an average of 6.4 months in five of seven patients.
One difficulty in evaluating any form of treatment of congenital pseudarthrosis of the tibia is the scarcity of published long-term follow-up studies. As one of the major goals is the durability of treatment, extended follow-up is paramount. Dobbs et al. reported the average 14.2-year results for patients with congenital pseudarthrosis of the tibia that had been treated with an intramedullary rod and autogenous bone-grafting4. The initial healing rate was 86% (eighteen of twenty-one patients), with a mean time to union of sixteen months. With long-term follow-up, however, the rate of refracture at the site of the initial pseudarthrosis was 57% (twelve of twenty-one patients). Although the authors attributed three of these fractures to removal of the intramedullary rod, the other nine fractures appear to have occurred despite the maintenance of internal fixation. As indicated by the functional outcome classification system of Johnston, these patients can have equivocal osseous union with minimum symptoms (a grade-2 outcome). If there is tenuous union, one might expect refracture with hardware fatigue over time as the pseudarthrosis site is stressed. Patients with a grade-2 outcome initially may be evaluated as having an osseous union, only to have a refracture with time because of the underappreciated tenuous union. With use of the surgical technique outlined in the present study, we found that, of the five patients who had initial union, four had maintenance of the union at an average of 5.4 years (range, 3.9 to 7.3 years). One patient (Case 3) sustained a refracture through a screw-interlock site distal to the original pseudarthrosis, which healed again after revision surgery, bone-grafting, and a second administration of BMP-2. We have been impressed by the healing response produced in conjunction with rhBMP-2 augmentation, as evidenced by the grade-1 outcome in five of our seven patients.
Despite the smaller number of patients and the shorter follow-up period in the present study, a comparison of our results with those reported by Dobbs et al.4 reveals a number of interesting findings. The use of BMP-2 in our patients did not lead to an increase in the initial union rate (71% compared with 86%), but the overall initial healing appears comparable. We found improvement in the average time to initial union in comparison with the rate reported by Dobbs et al. (6.4 compared with sixteen months). Finally, it appears that our patients who were managed with BMP-2 had a higher rate of maintenance of union in comparison with those in the report by Dobbs et al. (80% compared with 57%), but the smaller number of patients and shorter duration of follow-up in our study prevent us from making a definitive statement about the maintained union rate.
The functional results in the present series were quite good (Table III). Two patients (Cases 3 and 5) required treatment for residual ankle valgus, as has been noted in other series3,15. We found relatively well-preserved ankle function, even in cases in which intramedullary fixation initially crossed the ankle joint but migrated proximally into the tibia during the growth of the patient. Although the patients in the present series had some residual problems, such as valgus ankle deformity, limb-length discrepancy, and complete ankle stiffness, obtaining early and lasting union of the tibial pseudarthrosis leads to the best clinical results.
To our knowledge, we are the first to report on the use of BMP-2 for the treatment of congenital pseudarthrosis of the tibia, and no deleterious effects due to BMP-2 were seen. The use of BMP-7 to treat congenital pseudarthrosis of the tibia has been previously reported, with mixed success. Lee et al. reported on five patients with congenital pseudarthrosis of the tibia who were managed with bone-grafting, fixation, and rhBMP-717. Union was achieved in only one of five patients after twelve months. Variables in the surgical technique involving the use of BMP-7 may have contributed to poorer outcomes as the authors used allograft and unilateral static dynamic external fixation for the tibia in that study. Three other case reports on the use of rhBMP-7 demonstrated healing of congenital pseudarthrosis of the tibia18-20. The results of those studies suggest that rhBMP-7 should be combined with autogenous graft and that optimum fixation of the pseudarthrosis is extremely important. Those reports suggest that additional investigation with use of randomized controlled studies is needed to elucidate the optimum use of bone morphogenetic proteins for the treatment of congenital pseudarthrosis of the tibia.
The safety profile of BMPs in clinical practice has been excellent to date. Despite concerns, we are not aware of any published reports of systemic toxicities in skeletally immature patients directly related to the use of BMP21,22. Because of concerns about the effect of BMPs on the immature physis and the potential for tumorigenesis, both commercially available types of BMP (INFUSE; Medtronic Sofamor Danek, and OP-1; Stryker Biotech, Hopkinton, Massachusetts) have listed contraindications in their product inserts regarding their use in skeletally immature patients. Although there may be a theoretical risk of adverse events in association with the use of BMP in skeletally immature patients, we did not identify any complications related to its use in the present series. Still, a word of caution regarding its use is warranted. We are not aware of any clinical trials of BMP in pediatric patients or in patients with potentially premalignant conditions, such as neurofibromatosis. Despite the reasonable safety profile with this product in the adult population, the lack of data regarding pediatric patients should be noted and discussed with patients and parents or guardians prior to use. The United States Food and Drug Administration recognizes BMP-2 as a device. Its use in patients with congenital pseudarthrosis of the tibia is off-label, meaning that the physician must fully inform the patient and/or parents of the potential risks, and everyone must be willing to take the risk of possible unknown long-term effects, particularly because the use of BMP does not require an investigational new drug application, investigational device exemption, or review by an institutional review board.
Despite the reputation of congenital pseudarthrosis of the tibia as a disorder that is recalcitrant to healing, the treatment method that we used appears to have improved the outcomes, with reliable initial healing of the lesion, a relatively rapid healing time (6.4 months), and durable results in four of five patients at an average of 5.4 years. Although the doses used in the present study were determined empirically, we currently use doses in the range of 4.2 to 8.4 mg. Avoiding higher doses should decrease the potential for osteoclastic effects that have been noted when excessive BMP is utilized10. In addition, through a complete review of the literature and review of our case series, we have found no adverse events in association with the use of rhBMP-2 in a population of skeletally immature patients. We conclude that the use of rhBMP-2 can lead to earlier union in patients with congenital pseudarthrosis of the tibia and may increase the chance of long-term success in the management of these patients.
Note: The authors thank Robert Welch, DVM, PhD, and Wade Shrader, MD, for their contributions to this research.