We reviewed 25 patients who had undergone resection of a primary bone sarcoma which extended to within 5 cm of the knee with reconstruction by a combination of a free vascularised fibular graft and a massive allograft bone shell. The distal femur was affected in four patients and the proximal tibia in 21. Their mean age at the time of operation was 19.7 years (5 to 52) and the mean follow-up period 140 months (28 to 213). Three vascularised transfers failed. The mean time to union of the fibula was 5.6 months (3 to 10) and of the allograft 19.6 months (10 to 34). Full weight-bearing was allowed at a mean of 21.4 months (14 to 36). The mean functional score at final follow-up was 27.4 (18 to 30) using a modfied 30-point Musculoskeletal Tumour Society rating system. The overall limb-salvage rate was 88%. The results of our study suggest that the combined use of a vascularised fibular graft and allograft is of value as a limb-salvage procedure for intercalary reconstruction after resection of bone tumours around the knee, especially in skeletally immature patients.
Since first described in the mid-1970s for the reconstruction of traumatic bone defects, the vascularised fibular graft has become increasingly popular for reconstructing the defects left after the resection of bone tumours.1 It has the capacity to unite with the host bone and remodel and hypertrophy in response to mechanical stress.2 Furthermore, it has the ability to survive infection, radiotherapy and chemotherapy.3
However, problems have occurred when using an isolated free vascularised fibular graft for the reconstruction of defects in weight-bearing bones of the lower limb, especially around the knee. The small diameter of the fibula in relation to the host bone can subject it to excessive stress. Multiple fatigue fractures are common, and may require further immobilisation or fixation. It can also be difficult to achieve solid bony fixation between the fibula and the small residual epiphyseal bone segment after an intercalary juxta-articular resection around the knee.4,5 This is particularly important in the growing child if the growth plate is resected and a limb-length discrepancy is anticipated.6
In 1993 Capanna, Bufalini and Campanacci7 described the combined use of an allograft shell and a free vascularised fibular graft for intercalary reconstruction of the lower limb. This combination provided better primary stability and protection for the vascularised fibular graft from excessive mechanical stress. The use of an allograft bone shell also allowed the preservation of a small epiphyseal bone segment with its articular surface to which it could be fixed with screws.
We have reviewed the results of the combined use of a massive bone allograft and free vascularised fibular graft in the primary management of bone tumours around the knee.
Patients and Methods
Between May 1998 and June 2006, we treated 25 patients with a primary malignant bone tumour which extended to within 5 cm of the knee by resection of the tumour and reconstruction with a free vascularised fibular graft and a massive allograft. We used this technique if at least 1 cm of tumour-free juxta-articular bone was available. All the patients were followed up for at least 28 months after their surgery.
There were 16 males and nine females with a mean age at the time of surgery of 19.7 years (5 to 52). A total of 17 patients were skeletally immature. Their pre-operative details are shown in Table I⇓, and their operative details in Table II⇓.
Post-operatively, the patients mobilised partially weight-bearing on an ischial weight-bearing knee brace once union of the free vascularised fibular graft had occurred. This was maintained until radiographs showed union of the allograft and initial hypertrophy of the vascularised fibular graft, at which point, full weight-bearing was allowed. Serial CT was used to evaluate the morphological changes in the vascularised fibular graft.8 The mid-point of the transferred fibula was used as the fixed point for direct measurement of the circumference of the fibula to assess the progress of hypertrophy. Functional evaluation of the patients was undertaken at the final follow-up using the modified 30-point Musculoskeletal Tumour Society rating score for the lower limb.9
Descriptive statistics were calculated using means and proportions as appropriate to the type of data. The survival rates of the patients, limbs and the reconstruction were estimated using the Kaplan-Meier method. Student’s t-test, the chi-squared test, linear regression and correlation coefficient analysis were used for statistical measurements. A result was considered to be significant when the p-value was ≤ 0.05. All statistical analyses were done using SPSS version 10.0 (SPSS Inc., Chicago, Illinois).
The mean length of follow-up was 140 months (28 to 213). The viability of the vascularised fibular graft was assessed with a bone scan in 21 patients (84%) and from a skin paddle in four (16%). The skin paddle is a flap from the harvested fibula, monitored for colour changes, temperature, turgor, capillary refill and bleeding on pinprick. Failure, detected by scanning and confirmed by clinical follow-up, occurred in three patients (12%). In two, the reconstruction fractured, failed to unite and eventually had to be revised to salvage the limb. The third patient developed deep post-operative infection. The reconstruction failed to unite and fractured: an above-knee amputation was carried out.
The mean time to union of the fibular graft was 5.6 months (3 to 10) and for the allograft 19.6 months (10 to 34). All except two of the allografts united primarily. Of these, one was successfully treated by autogenous bone grafting and replating, the other required amputation for deep infection. Partial weight-bearing was allowed at a mean of 7.4 months (4 to 18) and full weight-bearing at a mean of 21.4 months (14 to 36).
Stress fractures occurred in three patients at a mean of 10.3 months (5 to 18) after the primary operation. All were managed by immobilisation in a cast with eventual union after a mean of two months (1 to 3).
There was a significant correlation between the time of union of the free vascularised fibular graft and that of the allograft (p = 0.029). We believe that this reflects the biological effect of the vascularised fibular graft on the allograft. The earlier the fibula united, the earlier did the allograft.
There was a significant correlation between the time of union of the diaphyseal end of the allograft and the stability of the fixation method (p = 0.013). The allograft healed earlier after a stable plating than with minimal fixation using screws and wires.
Hypertrophy of the fibular graft of more than 20% on serial CT was considered to be significant.2,8 Significant hypertrophy was present in 19 of the 21 patients (90.5%) after a mean of 101.7 months (24 to 213). The mean hypertrophy was 55.7% (12% to 150%).
There was a significant correlation between the extent of hypertrophy of the fibular graft and the time to union of the fibula (p = 0.05). When the fibula healed early partial weight-bearing started earlier with early loading of the fibula (p = 0.05).
There was a positive significant correlation between the extent of hypertrophy of the fibular graft and the time to union of the allograft (p = 0.025). When the allograft took longer to unite there was more loading of the fibula. Hypertrophy was greater with minimal fixation using wires and screws than with stable plating (p = 0.05). Stress fractures of the fibula significantly increased the rate and magnitude hypertrophy of the graft (p = 0.000).
There were 17 of these in 12 of the 25 patients (48%) (Table III⇓). Management was successful in nine, but required removal of the graft in two and amputation in one.
The mean time at which fracture of the allograft occurred was 29.9 months (11 to 75). Four of these fractures healed after immobilisation in a cast in a mean time of 2.75 months (2 to 4). Two healed after plating and cancellous bone autografting in three and four months, respectively. Three failed to unite, all in patients in whom vascularisation of the fibular graft had failed. Limb salvage was achieved in two by further reconstruction with allograft and bone cement in one and by another free vascularised fibular graft and allograft in the other. One patient had an above-knee amputation because of deep resistant infection and was the only case of deep infection in the series (Table III⇑).
Six donor-site complications occurred in five patients. One required lengthening of the tendon for flexion deformity of the big toe (Table IV⇓).
At the last follow-up, 15 patients had a mean limb-length discrepancy of 2.4 cm (1 to 6). One had a bone-lengthening procedure for shortening of 6 cm. The remaining 14 are either using a shoe raise to compensate for shortening or waiting until they reach skeletal maturity to consider a lengthening procedure. Growth of the residual epiphysis was observed in skeletally immature patients (Fig. 1⇓), but the extent and rate of growth of the remaining epiphysis were beyond the scope of this study.
A total of 14 secondary surgical procedures were required in nine patients to achieve union or to treat complications. In one patient four secondary operations were necessary (Table V⇓).
After exclusion of five patients (two local recurrences and three failed vascularised transfers) the mean modified 30-point musculoskeletal tumour society functional score at the final follow-up after a mean of 139.8 months (28 to 213) was 27.4 (18 to 30). All patients had a good functional range of movement and a stable knee. Of the five excluded patients, three had undergone amputation (two for local recurrence and one for deep infection) and two revision of the reconstruction (Table VI⇓).
Local recurrence occurred in two patients at five and 24 months post-operatively. Both were managed by amputation. One showed no evidence of disease at final follow-up after 89 months. The other developed a pulmonary metastasis 42 months after the primary surgery and died 20 months later. Local recurrence was attributed to inadequate surgical margins in both patients and confirmed by postoperative pathological examination of the resected specimens. Distant metastases to the lung occurred in three patients (12%) at a mean 44 months (21 to 69). One died 62 months after primary surgery. The other two were treated by pulmonary lobectomy and showed no evidence of disease at final follow-up 160 and 165 months, respectively, after surgery.
Kaplan-Meier survivorship analysis showed that the two-year limb survival from amputation for local recurrence, infection and failure of the vascularised transfer was 88% (sd 13) (95% confidence interval (CI) 75 to 99), with no amputation expected to occur after the second year. The success rate of reconstruction at three years was 88% (sd 13) (95% CI 75 to 99) and at five years 79% (sd 16) (95% CI 63 to 95).
Restoration of good function to a limb and maintaining it in the long-term after resection of a bone tumour remains a challenge. The ideal reconstruction should be biologically similar, resist infection and be both strong and durable.10,11
The knee is a common site for primary bone tumours. Most patients can be treated with a limb-salvage procedure. In most cases, when the tumour extends into the meta-epiphysis, an intra-articular resection which includes the articular surface is required. Under these circumstances, the reconstructive options include an endoprosthesis,12 an osteo-articular allograft13 or a combination of the two.14
Endoprosthetic replacement has many advantages, such as early stability, mobilisation and weight-bearing, a high level of emotional acceptance by the patient and rapid restoration of function with a good functional outcome. However, problems such as infection, mechanical failure and aseptic loosening may limit the long-term survival of the prosthesis especially in skeletally immature patients.15,16 Despite recent improvement in the long-term results of endoprosthetic replacement as a result of improvements in design and materials,17 the risk of revision increases with time. In 2007, Myers et al18,19 noted that the risk of revision surgery at 20 years was more than 50% for the distal femur regardless of the type of prosthesis, up to 75% for a fixed-hinge prosthesis and 30% for a rotating-hinge prosthesis in the proximal tibia.
The use of an extendible endoprosthesis in the management of bone tumours in skeletally immature patients can help to maintain equality of leg length. However, the high rate of secondary operations and the complications associated with the use of this type of prosthesis have limited its popularity.20 The use of a non-invasive extendible endoprosthesis has limited the need for secondary operations and has been associated with a lower rate of complications.21 However, longer follow-up is needed to evaluate this technique fully.
A massive allograft can be used as an osteo-articular or intercalary supplement and can be combined with an endoprosthesis. The allograft may take up to 24 months to unite: a rate of nonunion of up to 20%22,23 and a rate of infection as high as 20% have been reported.22–24 The incidence of fracture is between 15% and 45% depending on which definition of fracture is used.25 A further problem with osteoarticular allografts is degeneration of the articular surface. The high rate of complications and revision surgery associated with the use of an osteo-articular allograft has led some authors to believe it should only be considered a temporary solution in the management of malignant bone tumours.13
The improved accuracy of imaging techniques has encouraged preservation of the epiphyseal plate in both the distal femur and proximal tibia. In these sites, intercalary reconstruction may be achieved with massive bone allografts alone26 or in association with a free vascularised autologous fibular graft,7 or by bone transport using the Ilizarov technique.10
Musculo et al26 described a series of 13 patients with a high-grade metaphyseal osteosarcoma around the knee who were treated by transepiphyseal resection and reconstruction with an intercalary allograft. They encountered deep infection in one patient, nonunion in two and fracture in three. Removal of the allograft was required in four of these patients. The mean functional score using the musculoskeletal tumour society system was 27 points at final follow-up.
Deijkers et al27 reported 35 patients who had intercalary allograft reconstruction of the lower limb after resection of a tumour. The reconstruction was epidiaphyseal in 14, metaphyseal in nine, and diaphyseal in 12. The allograft reconstruction was augmented with a free vascularised fibular graft in ten patients. They reported a survival rate of 79% at ten years for epidiaphyseal reconstructions and 89% for metadiaphyseal reconstructions. There were three cases of infection, 12 of fracture and nine of nonunion requiring treatment from the 35 grafts.
The results of our study indicate that the use of a free vascularised fibular graft and shell allograft in combination is a reliable method of reconstructing an intercalary defect around the knee after resection of a bone tumour. The primary rate of union of the allograft in our series was 92%. The mean time to union of the allograft was 19.6 months (10 to 34) which could be explained by the fact that 18 patients (72%) received pre- and post-operative chemotherapy.28 The study also confirmed the importance of stable fixation of the allograft in achieving early union.29
The correlation between the time to union of the allograft, that of the fibula and the extent of fibular hypertrophy may reflect the fact that the fibula, in addition to its biological role in allograft union, shares in weight-bearing until there is solid union of the allograft. The biological role of the fibula was seen on long-term follow-up CT in which fusion between the fibula and the allograft and hypertrophy of the fibula were observed (Fig. 2⇓).
The Kaplan-Meier survivorship analysis showed a patient survival rate of 88% at two years and of the reconstruction of 79% at five years, with no change expected after these periods of time. These figures support the favourable long-term results of this method of reconstruction when compared with other procedures.
Although resection was intraepiphyseal in 13 patients (52%) only two had a local recurrence (8%) and both had an intercalary resection. This figure is similar to previously published rates of local recurrence (5% to 10%).30
Despite the high rate of local complications in our series (48%) treatment was successful in 75% of patients. We believe that the fact that allograft fractures unite in 67% of patients reflects the biological capability of the free vascularised fibular graft. In the three patients who failed to achieve union, vascularisation of the fibular graft also failed. Only one patient had a deep infection which compares favourably with a rate of 10% to 20% in other published series of allograft reconstruction.24,31
In skeletally immature patients, the choice of reconstruction is more difficult. The most common options in this age group are rotationplasty or the use of an extendible prosthesis. Rotationplasty provides good long-term function, but its psychological impact cannot be overlooked. The use of an extendible prosthesis is associated with a high rate of complications requiring secondary operations.21,32 In the present study, 17 patients were skeletally immature at the time of surgery and 15 had a mean limb-length discrepancy of 2.4 cm (1 to 6) at final follow-up. Only one patient has undergone bone lengthening. The others compensate for the limb-length discrepancy by using a shoe raise or are awaiting skeletal maturity before considering a lengthening operation.
Since it was first reported in 1993,7 the combined use of a free vascularised fibular graft and allograft has gained popularity and several studies have reported good results with an excellent functional outcome, a high rate of union, and low rates of infection and fracture.33–38 The reliability of this technique has encouraged surgeons to expand its indications and to use a free vascularised fibular graft to successfully treat nonunion of the allograft.39
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.
- Received December 19, 2008.
- Accepted June 3, 2009.
- © 2009 British Editorial Society of Bone and Joint Surgery