Bony Hypertrophy in Vascularized Fibular Grafts

Abstract

Background: Vascularized fibula graft (VFG) transfer is an established method of repairing large skeletal defects resulting from trauma, tumor resection, or infection. It obviates the process of creeping substitution that conventional bone grafts undergo and therefore exhibits better healing and improved strength. The aim of this study is to evaluate hypertrophy in VFG. Methods: We retrospectively reviewed patients undergoing VFG and studied immediate and late postoperative radiographs. Orthogonal views were measured for width of graft cortex and intramedullary canal, as well as adjacent recipient bone. Changes were measured for total cross sectional area, cortical area, intramedullary area, and graft width. Results: Thirty patients were included in the analysis, with recipient sites including 3 forearm, 4 humerus, 12 tibia, and 11 femur. Mean follow-up was 7.6 years (0.5-24.9 years). Patients’ mean age was 31 (16-59 years). Average hypertrophy was 254% in early postoperative period and 340% in the late postoperative period. There was rapid graft hypertrophy in early postoperative period that plateaued with time. The width of the graft increased over time but didn’t exceed the width of the adjacent recipient bone. In the later postoperative period, the size of graft intramedullary canal increased. Upper and lower extremity grafts showed similar hypertrophy. Conclusions: Using VFG to treat large skeletal defects is an attractive option in part due to the graft’s ability to hypertrophy. We describe an early period of periosteal hypertrophy, followed by endosteal hypertrophy. These processes have relevance to function, mechanical strength, and surgical decision-making.

Introduction

Vascularized fibular graft (VFG) transfer is a method of autologous bone grafting in which fibula is transplanted with an intact vascular supply to a skeletal defect. Unlike other methods of bone grafting, VFGs require preservation and reconnection of a vascular pedicle from the donor site to the graft site. This became a possibility only after the development of microvascular anastomotic techniques in the 1960s. The technique of VFG was first described in 1975 by Taylor et al ¹ in their case report of a tibial defect repaired using a vascularized graft from the contralateral fibula. Since then, indications for VFG have expanded to include fixation of skeletal defect from infection, ² tumor, ³ congenital pseudarthrosis, ⁴ and trauma. ⁵ VFG recipient sites can be both lower^4,6 and upper extremities,^7,8 and the technique can be used in both adult and pediatric populations. ⁹

The fibula is an ideal donor of vascularized bone grafts due largely to the organization of its vascular supply, but also due in part to its bony composition. ¹⁰ The fibula receives both endosteal and periosteal vascular supply from a singular source^1,10—the peroneal artery and veins—minimizing the required number of surgical anastomoses. In addition, anterior tibial artery provides circulation to the proximal fibula and the growth plate and can be an option if needed. The relatively large vessel diameter, in the range of 1.5 to 3 mm,^10,11 further eases the process of surgical anastomosis with the recipient site vasculature. The density of cortical bone, triangular cross-section, and tubular structure of the fibula combine to provide resistance to torsional stress.^12,13 Additionally, while the fibula is most similar in size to the radius and ulna, it can be transferred to a variety of bony defects. VFGs can be modified to fit into intramedullary canals of larger diameter bones ⁷ or into a double-barrel graft to replace defects of larger cross-sectional area (CSA). ¹⁴

With their intact vascular supply, VFGs have a biological advantage over conventional grafting methods. The latter heal by creeping substitution, ¹⁵ in which nonvascularized bone serves as a scaffold for angiogenesis and deposition of new osteoid. In VFGs, however, the preservation of microvasculature by surgical anastomosis between the recipient and the graft allows for survival of osteogenic cells. ¹⁶ Possible complications of bone grafting include fracture fatigue, graft nonunion, and graft resorption, ¹⁰ and these can result from incomplete bone formation over the necrotic trabecular scaffold in creeping substitution. In VFGs, these risks are minimized. The overall rate of union for VFG is >85%, with the exception of the diagnosis of trauma where rate of union is approximately 75%.^17,18 The reported average time to union is approximately 8 months. ¹⁷ Due to the technical demands of microvascular anastomotic techniques and the rate of complications, VFG is usually limited to reconstruction of large skeletal defects (ie, ≥10 cm) or difficult nonunions that have been unsuccessfully treated with nonvascularized bone grafting methods. ¹⁹

The hypertrophy of VFG after surgery has been frequently observed. ¹⁶ Multiple studies have shown that hypertrophy of the graft is more common in the lower limbs, and when the graft is not stabilized by internal fixation (ie, with increased mechanical loading).^16,20 It has also been shown that grafts atrophy in the absence of a mechanical load. ²¹ The length of the graft has not been shown to affect bone union or graft hypertrophy. ²² These findings suggest that VFGs heal like segmental fractures, and the grafts can respond to biomechanical stress. ¹⁶ However, even without mechanical stress, VFG hypertrophy can occur and is associated with increased vascularity of the periosteum; this would suggest a mechanism different from reactive callus formation in normal fracture healing. ²³ VFG hypertrophy has also been shown to be limited to the borders of the original bone and does not continue indefinitely, ²³ another distinction from reactive callus formation.

The purpose of this study is to quantify hypertrophy and to better understand the mechanisms of VFG hypertrophy, including periosteal and endosteal components of bone hypertrophy. In addition, we aim to compare the hypertrophy for grafts to the lower limbs relative to upper limb grafts.

Methods

We retrospectively reviewed charts for patients who underwent VFG reconstruction of long bone skeletal defects between 1980 and 2008 at Massachusetts General Hospital in Boston, United States, and Shanghai Sixth People’s Hospital in Shanghai, China. Only those patients in whom VFG was used to bridge a long bone skeletal defect were included in the study. Patients were excluded for other indications of VFG (such as femoral head osteonecrosis) or other configurations of bone graft (along side of existing bone to add mechanical support, or in conjunction with allograft). Patients were included if they had at least one anteroposterior (AP) and lateral X-ray during the early postoperative period and at least one AP and lateral X-ray after radiographic evidence of graft union (Figure 1). We identified 30 patients who had adequate imaging data available for review. This study was approved by respective institutional review boards.

Figure 1.

Radiographs of VFG tibial reconstruction from a study subject with adequate radiographs. External stabilization of the graft is present immediately post operation (a and b), but absent in later postoperative radiographs (c and d). AP (a and c) and lateral (b and d) radiographs were obtained for early postoperative follow-up and after graft union.

VFG = vascularized fibula graft.

Analysis of Graft Hypertrophy via Imaging Data

All available imaging was converted to a common digital format. Cortical width and intramedullary width on both AP and lateral views was measured at the midpoint of the VFG as well as the adjacent recipient bed (Figure 2). Measurements were made for the initial postoperative X-rays and images obtained at final follow-up after evidence of union. CSA of both recipient and graft was calculated at each time point using the average of widths measured from the AP and lateral radiographs. Total CSA of a fibular graft was assumed to approximate a triangle, whereas medullary CSA was assumed to approximate a circle. Recipient total and intramedullary areas were assumed to approximate circles. Cortical CSA was calculated by taking the difference of total CSA and intramedullary CSA. Graft hypertrophy was calculated using modification of de Boer and Wood’s ¹⁶ definitions, where instead of using graft diameter we employed total graft width and total CSA for measurements. CSA might better represent the 3 dimensional shape of fibula and recipient bone compared to graft and recipient bone diameter. Index is defined as the ratio of a parameter (eg, cortical CSA, intramedullary CSA or total CSA) measured at the graft site to the value of the parameter for the neighboring recipient bone (Equation 1).

Figure 2.

Example measurements of total bone width and intramedullary width for recipient and VFW are shown on AP (right) and lateral (left) radiographs. Dashed lines indicate total bone width and solid lines indicate intramedullary width; cortical width was taken as their difference. Total, cortical, and intramedullary surface areas were calculated using the average of bone diameters measured from AP and lateral radiographs.

VFG = vascularized fibula graft.

$$Index=\frac{CSAgraft}{CSArecipient}$$ (1)

Hypertrophy is defined as the change in Index for total CSA over time (Equation 2).

$$Hypertrophy=\frac{\begin{array}{l}IndexofTotalCSAat\\ finalfollowup-\\ IndexofTotalCSAat\\ earlyfollowup\end{array}}{\begin{array}{l}IndexofTotalCSAat\\ earlyfollowup\end{array}}$$ (2)

Similarly, cortical and intramedullary hypertrophy values were calculated for changes over time in index for cortical and intramedullary CSA, respectively. The Index of total graft width was calculated using a modified version of Equation 1 in which CSA was replaced with the average of the total width from AP and lateral radiographs.

The parameters calculated using the aforementioned equations were plotted against postoperative time, including total CSA, medullary CSA, cortical CSA, and Index of total graft width. All graphs were generated using Microsoft Excel; best-fit lines for all graphs were generated using the same program. Independent sample t-test was used to compare hypertrophy between upper and lower extremity groups.

Results

Patient Characteristics

Patient mean age at the time of surgery was 31 years (range 16-59 years). Mean follow-up was 7.6 years (range 6 months to 24.9 years). The recipient sites included 3 forearm, 4 humerus, 12 tibia, and 11 femur (Supplementary Table 1). VFG reconstruction indications included 7 acute trauma, 17 infected nonunions, 1 congenital defects, and 5 tumors (Supplementary Table 1).

Hypertrophy of VFG Occurs Rapidly in Early Postoperative Period

Figure 3 shows hypertrophy values in % for all the patients at final follow-up. As shown in the chart, VFG undergoes hypertrophy rapidly in the early postoperative period with values more than 100% even for patients with short follow-ups. For most of the patient with follow-up <1000 postoperative days, hypertrophy varied from 100% to 300%. Average hypertrophy for patients with follow-up < 1000 days (19 patients) was 254%. For patients with follow-up >1000 days, hypertrophy values mostly ranged from 300% to 400%. Average hypertrophy for patients with follow-up >1000 days (11 patients) was 340%. These data favor the theory that hypertrophy occurs rapidly in the initial postoperative period and then plateaus over time.

Figure 3.

Each data point on the graph represents an individual patient’s VFG hypertrophy value at the final follow up with % hypertrophy represented on the Y axis and follow-up in days on X axis.

VFG = vascularized fibula graft.

Hypertrophy of VFG Is Similar for Upper and Lower Limb Recipient Sites

When stratified into upper and lower limb recipient sites, VFG hypertrophy is similar in both groups (Figure 4). Overall average hypertrophy was 290% for upper extremity group and 284% for lower extremity group (P = 0.88). Hypertrophy in patients with lower limb grafts averaged 331% for patients with >1000 days follow-up and averaged 259% for patients with <1000 days of follow-up. In patients with upper limb grafts, hypertrophy averaged 366% and 233%, respectively. Patients with lower limb graft recipient sites demonstrated a higher variability in graft behavior for patients with <1000 days of follow-up relative to patients with grafts to the upper limb (Figure 4).

Figure 4.

Hypertrophy data divided into (a) upper extremity and (b) lower extremity.

VFG Graft Width Is Limited by Borders of Recipient Bone

The index of graft width approached 1 but did not exceed the recipient bone width for any patient (Figure 5). This follows the same pattern as the total CSA hypertrophy [Figure 5 and Figure 1). Index of total graft width averaged 0.6 for patients with <1000 days follow-up and 0.7 for patients with >1000 days follow-up.

Figure 5.

The index of total graft width, increases to but doesn’t exceed 1.

Cortical and Intramedullary Hypertrophy Follow Opposite Trends

Cortical hypertrophy values seem to decrease in patients with later follow-up (Figure 6a). In contrast, intramedullary hypertrophy seems to be higher in patients with later follow-up dates (Figure 6b).

Figure 6.

Changes in VFG (a) cortical area and (b) IM area. Each data point on the graph represents an individual patient’s hypertrophy value.

VFG = vascularized fibula graft; IM = intramedullary.

Discussion

VFG is an established surgical technique used in the repair of large (>10 cm in size) bony defects or for those unsuccessfully repaired with traditional grafting techniques. The advancement of surgical techniques continues to increase the indication for the use of VFG. The literature on the mechanism of VFG hypertrophy is limited to small retrospective studies. Several retrospective studies indicate a correlation between increased mechanical loading of the limb or increased graft vascularity and increased hypertrophy of the graft.^10,16,23 In a study of patients with VFG reconstruction following removal of lower limb tumors, El-Gammal et al, ²⁴ reported 2 peaks in the rate of graft hypertrophy: at 6 to 12 months post-operation and again at 18 to 24 months. Patients younger than 20 years of age were also shown to have a faster rate of hypertrophy and reached maximum hypertrophy at an earlier time point than older patients. Although studies generally agree to the existence of a point of maximum hypertrophy, there is variability in when that point is reached. Lazar et al report the time point to be 2 years post-operation, ²⁵ while Wei et al ²⁶ reported it to be 3 years.

Beyond the establishment of a relationship between the aforementioned factors and VFG hypertrophy, there is a need to quantify and classify the hypertrophy. While graft hypertrophy has been proposed by de Boer and Wood ¹⁶ to be either endosteal, periosteal, or a composite of the 2, few studies have examined whether 1 or more of the above mechanisms more accurately describe the process of VFG hypertrophy. In this study, measurement of both endosteal and periosteal hypertrophy was carried out in order to suggest a possible mechanism.

Our data suggest that there is rapid initial hypertrophy that stabilizes over time (Figure 3). Two mechanisms may explain this. First, several studies of VFG hypertrophy have demonstrated an increase in graft hypertrophy with mechanical loading and vascularization of the graft site. Both of these stimuli are present at a relatively high frequency in the immediate postoperative period and may explain the high initial rate of hypertrophy. Second, the stable values of hypertrophy with increasing postoperative time suggests a mechanism where VFG width does not go beyond the borders of the adjacent recipient bone, a result shown by previous studies and the current study.

Upper and lower extremities showed similar hypertrophy, matching the results from other studies. It has been postulated that mechanical loading plays a role in hypertrophy and hence should be different between upper and lower extremities; however, other factors such as recipient periosteum vascularity (better in upper extremity) can enhance hypertrophy. ²³ As shown in this series, the donor fibula usually doesn’t hypertrophy beyond the borders of the recipient bone; this suggests a combination of recipient bone vascularity and mechanical loading affects hypertrophy.

Analysis of periosteal and endosteal hypertrophy, as previously defined by de Boer and Wood, ¹⁶ the current study shows a higher initial cortical hypertrophy that decreases with time and a lower initial intramedullary hypertrophy that increases with time (Figure 6). The model shows periosteal hypertrophy contributing more in the early postoperative period. In the late postoperative period, our model suggests greater contribution from endosteal hypertrophy and growth of the intramedullary canal.

There exist numerous limitations in the data and the conclusions that can be drawn. We are primarily limited by the quality of imaging; all data concerning graft hypertrophy were calculated from radiographs. Computed tomography was not available from each patient. The small sample (n = 30) can limit the conclusions that are reached, particularly regarding subgroup analysis, such as the comparison between upper and lower extremity graft recipient site. Given the aforementioned technical demands, however, a sample size of 30 is among the larger cohorts found for studies of VFG in the literature. Conclusions on hypertrophy were drawn based on short, intermediate, and long-term data on hypertrophy for all the patients. It would be more interesting to follow hypertrophy over time for each individual patient; however, obtaining such long term follow-up was not feasible with the current study design. This study looked at graft hypertrophy for VFG in treating long bone defects, and didn’t evaluate hypertrophy in other indications of VFG, such as in osteonecrosis of femoral head. Double Barrel ¹⁴ and Capanna ²⁷ techniques were not evaluated, and we investigated hypertrophy only when VFG was placed in intramedullary canal to bridge long bone defect. Lastly, there is not a sufficient sample to examine the effect of fixation method on degree of hypertrophy.

Conclusion

Our study evaluated the postoperative progression of VFG by monitoring cortical and intramedullary hypertrophy, as well as total graft width, using a large cohort and with a novel calculation of hypertrophy and width. Graft hypertrophy stabilized over time as total graft width approached the width of the recipient bone; graft width did not exceed the width of recipient bone. Intramedullary area increased with postoperative time; both cortical and total graft area plateaued in the late postoperative period (Figure 7). Our findings suggest a mechanism for VFG hypertrophy that differs from the mechanism of reactive callus formation, most notably with respect to the continuous growth of intramedullary width and the limit on graft width exerted by the width of the recipient bone. These findings have the potential to impact long-term postoperative care and lifestyle management of patients who have undergone a VFG procedure.

Figure 7.

Proposed model for VFG hypertrophy: On the basis of hypertrophy data, our model shows continuous hypertrophy of the periosteum beginning in the immediate postoperative period, but delayed hypertrophy of the endosteum and growth of the intramedullary canal. Periosteal hypertrophy occurs to a lesser degree in the late postoperative period.

VFG = vascularized fibula graft.