What Is the Role of Allogeneic Cortical Strut Grafts in the Treatment of Fibrous Dysplasia of the Proximal Femur?

Despite progress in understanding the pathogenesis of fibrous dysplasia, its treatment has been subject to controversy ever since the first reports of Lichtenstein on management of the disease, emphasizing its notorious recurrent character, the wide variations in phenotype, and the lack of a successful treatment strategy [15, 16]. Over the past few decades many surgical approaches have been proposed and discarded in the management of fibrous dysplasia ranging from particular forms of grafting to a variety of types of implants or a combination of both [6]. Although curettage and bone grafting using cancellous bone have been historically the treatments of choice, the perspective of the value of this treatment in fibrous dysplasia lesions of the proximal femur has altered over the last decades as a result of increasing reports on the marginal outcome of the procedure (Table 3) [10‐12, 18, 19, 23]. We theorized that allogeneic cortical strut grafting might improve the outcome of fibrous dysplasia lesions of the proximal femur, arguing that cortical allografts would be less prone to be affected by pathological fibrous dysplasia bone and therefore less prone to resorb and fail. We found that allogeneic cortical strut grafting has a survivorship of 54% after long-term followup and that patients who presented with fracture (as opposed to impending fracture) and patients whose grafts lacked sufficient proximal fixation were at increased risk of undergoing reoperation.

This study had a number of limitations. First, the small number of patients included in our study reflects the low prevalence of symptomatic fibrous dysplasia, although our series of patients was larger than any reported thus far of which we are aware. Because of a small study size, it is possible that we were unable to detect the less common complications of allogeneic cortical bone grafting. Furthermore, we have to take into account the possibility that some of the risk factors that were not associated with treatment outcome in our study might show an association in larger studies. Second, the long span of time over which this retrospective study’s procedures were performed saw many changes in patient care. Although the indications were generally consistent over time at the study site, it is impossible to know with certainty that they were applied with precision over the nearly 35-year timeframe. Also, many of these procedures were performed before patient-reported outcomes tools came into wide use, and so we could not report patient-reported outcomes here. It is our impression based on chart review and patient surveys done after surgery that pain improved in most of these patients. Furthermore, we appreciate that the variability in treatment options for fibrous dysplasia of the proximal femur, which is no doubt the result of the heterogeneity of the disease, makes it difficult to compare different patients. Although the use of multiple trajectories, additional cancellous bone grafting, and additional curettage were taken into account in our analysis, these different approaches within the treatment with cortical allografts might affect the outcome in larger studies. Finally, although the treatment of fibrous dysplasia is centralized in The Netherlands, it is common in long-term studies that some patients were lost to followup.

Although a large proportion of patients in this study (nearly half of them) underwent revision at some point during the followup period, we consider the fact that more than half of the patients did not undergo further surgery actually a reasonably good result. The reason for this is that the condition recurs so commonly, and other described approaches actually reported even more frequent failures than we observed here [10, 23]. Most previous studies were restricted to limited followup and are therefore likely to overestimate the therapeutic effect of bone grafting, a fact that is further emphasized by studies with long-term followup generally, suggesting a poorer outcome (Table 3). Because of its mean followup of 13 years, our study gives a fair representation of the long-term effects of allogeneic cortical strut graft treatment in fibrous dysplasia lesions of the proximal femur. However, Kaplan-Meier survivorship analysis (Fig. 3A) showed that most reoperations were performed within 5 years after the primary surgery, indicating that after 5 years, failure of allogenous cortical strut grafts leading to reoperation is less likely to be expected. In addition, besides being biologically preferable to prevent graft resorption, the use of allogeneic bone has the advantage of no additional surgery being required to retrieve autogenous bone.

Slightly more than half of our patients experienced radiographic evidence of graft resorption. Again, however, we consider this a reasonably good result considering the problems reported using other techniques when dealing with proximal femoral lesions in patients with fibrous dysplasia, especially in studies addressing cancellous bone grafting in which recurrence is reported in nearly all patients [10, 23]. Although the indication for reoperation was graft resorption in the majority of the cases (61%), graft resorption was not an indication for reoperation per se, because we would only perform a reoperation in case graft resorption led to an impending or an actual fracture.

We identified several risk factors for failure of allogeneic cortical strut graft surgery in this study. In patients undergoing cortical strut allografting, we found a minimal proximal anchoring ratio in vital bone of 5% is required to enable the graft to be incorporated in the proximal femur. Therefore, proper evaluation of proximal placement preoperatively and intraoperatively should be mandatory, whereas this is not so for distal anchoring in which case anchoring in the cortex will generally suffice. Although it has previously been reported that insufficient proximal docking in healthy bone might have played a role in failure of allogeneic cortical strut grafting in two patients in the Enneking/Gearen study [7], our study is the first to clearly identify insufficient proximal anchoring of the graft and preoperative fractures as risk factors for failure of allogeneic cortical strut graft surgery.

Our data also demonstrate a higher risk of revision in patients with a preoperative fracture. Patients who sustained a fracture of the proximal femur at some point before surgery have thus a high risk of failure of allogeneic cortical strut grafting by either resorption or a consecutive fracture. We suspect that a pathological fracture is only the endpoint of a sliding scale. Guille et al. already identified involvement of the calcar femorale as a risk factor for failure [10]. If the calcar is involved, the proximal femur will considerably lose stability by loss of redistribution of stress forces [26]. Subsequently the proximal femur will be more prone to fractures. A cortical strut graft will very likely not be able to address the extensive forces applied on the proximal femur without this form of stability. Based on the results with internal fixation in other studies [1, 4, 5, 8, 10, 12‐14, 17, 20, 23‐25] and the findings from our present study, we recommend primary internal fixation in patients with extensive lesions that threaten to fracture or have already induced a fracture, because this approach has shown promising results in several studies. More importantly, although only the combination of internal fixation and autografts has so far been studied, there might be an important role for implantations in combination with cortical allografts in patients with preoperative risk factors. Although the polyostotic form of fibrous dysplasia is generally associated with a worse outcome compared with monostotic disease [6, 10], we were not able to identify polyostotic fibrous dysplasia as a risk factor for graft failure. This may be explained by the fact that polyostotic fibrous dysplasia can be profoundly variable in its course and in the extent of lesions. Patients with extensive lesions, both monostotic and polyostotic fibrous dysplasia, were primarily treated with internal fixation, because a strut graft would not properly bridge the lesion and therefore local expansion and not the type of fibrous dysplasia would be a risk factor. Furthermore, we expected that young age at the time of surgery would be a risk factor based on the hypothesis that fibrous dysplasia tends to be more active and aggressive during childhood and because patients with extensive disease are generally diagnosed at a younger age [22]. However, we were unable to demonstrate this in the present study, perhaps because of the statistical limitations imposed on our statistical analysis by the sample size.

Although this procedure can be technically challenging, we were gratified by the relative rarity of complications in this series. This compares favorably to bone grafting with autogenous bone, which can be accompanied by complications at the donor site [2, 21].

Our findings from this study suggest that cortical strut allografting may be a viable option for treatment of fibrous dysplasia of the proximal femur who have not already experienced a fracture. Surgeons should pay particular attention to the proximal fixation point of the allograft to decrease the risk of failure. Patients with a fracture have an increased risk of failure and reoperation and so should be treated with osteosynthesis.