Femoral reconstruction by single, folded or double free vascularised fibular grafts

The British Association of Plastic Surgeons (2004) 57, 550 555 Femoral reconstruction by single, folded or double free vascularised fibular grafts K. Muramatsu*, K. Ihara, M. Shigetomi, S. Kawai Department of Orthopedic Surgery, Yamaguchi University School of Medicine, 1-1-1 Minami-Kogushi, Ube, 755-8505 Yamaguchi, Japan Received 3 December 2002; accepted 15 August 2003 KEYWORDS Femur; Fibula; Vascularised bone graft; Stress fracture; Hypertrophic change Summary We reviewed 17 patients for a mean of 25 months period after free vascularised fibular transfer to reconstruct massive bone defect or recalcitrant nonunion of the femur. There were 11 cases of posttraumatic nonunion and six patients had a large bony defect following resection of bone tumour. Ten patients underwent double or folded and seven patients underwent single vascularised fibula graft transfer. Mean bony defect of the femur was 6.5 cm and mean length of grafted fibula was 15 cm. Revision surgery due to postoperative vascular complications was required in five cases. Twenty-three out of 24 (96%) vascularised fibulas were transferred successfully. The resultant outcome was successful in 15 out of 16 (94%) patients with confirmed bone union. Stress fracture occurred in three inlay fibula grafts. Hypertrophic change of the fibula graft was significantly noted in inlay grafts as compared to onlay grafts. All patients could walk without brace at a mean of 11 months postoperatively. Donor-site morbidity was minimum. Vascularised fibula grafting is a reliable and safe reconstructive procedure for massive femur defects. Folded or double fibula grafts cannot prevent stress fractures and the key point is to rigidly stabilise the femur in an anatomically aligned position. Q 2004 The British Association of Plastic Surgeons. Published by Elsevier Ltd. All rights reserved. Free vascularised fibular graft (FVFG) has become an established procedure for the treatment of massive bone defects after bone tumour resection 1,2 and recalcitrant nonunion. 3,4 Fibula is probably the most suitable donor bone for reconstruction of a large defect in a long bone because of its length, geometrical shape and mechanical strength. Of all long bone chronic nonunions, those involving femur are the most disabling and reconstruction with FVFG is favorable in these cases. However, some authors pointed out that *Corresponding author. Tel.: þ81-836-22-2268; fax: þ81-836- 22-2267. E-mail address: muramatu@po.cc.yamaguchi-u.ac.jp FVFG to the femur is technically very difficult with a high rate of immediate and late complications. 4,5 The recipient bone is located deeper than other long bones and has only one major artery. Internal and external bony fixation is very difficult because of the strong muscle forces across the femur. Length discrepancy more than 2 cm of the femur causes gait disturbance. Among them, the most important problem is late stress fracture of FVFG due to low initial mechanical strength. 6,7 Late stress fracture is probably prevented by hypertrophic change of FVFG but until now, to our knowledge, little is known about stress fracture and hypertrophy of FVFG in femur reconstruction. In this study, we reviewed a consecutive series of S0007-1226/$ - see front matter Q 2004 The British Association of Plastic Surgeons. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.bjps.2003.08.021

Femoral reconstruction by single, folded or double free vascularised fibular grafts 551 17 patients after femur reconstruction by FVFG, assessing clinical results and complications. In particular, we focused on the cause of stress fracture and discussed the relationship between mechanical stimulation and hypertrophy of the FVFG. Patients and methods We retrospectively reviewed 17 patients, who underwent FVFG transfer due to a massive bone defect or recalcitrant nonunion of the femur between 1989 and 2001. All free fibular transfers were carried out in our university hospital and the follow-up of the patients was subsequently carried out at our outpatient clinic. The mean age of the patients was 39 years (5 74) and the mean followup period was 25 months (6 60). The recorded data of each patient included age, sex, diagnosis, duration of follow up, recipient site, fixation devices, length of bone defect, type of FVFG, graft length, recipient vessels, technique of vascular anastomosis, graft survival, postoperative complications, occurrence of stress fracture, hypertrophic change of FVFG 8 and functional recovery (Table 1). FVFGs were transferred in six cases of posttraumatic nonunion or bone defect without infection, five cases of infected nonunion and six cases of bone defect after tumour resection. Recipient site was the femoral neck in two patients, mid-shaft in six patients and distal femur in nine patients. Skeletal fixation of the femur was performed with a plate and screws in five patients, external fixator in six patients, fibula graft itself in three patients and intramedullary rod in remaining two patients. Single FVFG was transferred in seven patients, double FVFG transfer in seven patients and twin-barreled (folded) FVFG transfer was performed in three patients based on a single vascular pedicle. In cases with double fibula grafting, one was basically placed as inlay (intramedullary) graft and another was used as onlay graft. The knee joints were fused by double FVFGs in two cases. The mean length of bone defect was 6.5 cm (0 15) and the mean length of FVFG was 15 cm (5 25). Supplementary bone grafting with iliac cancellous bone graft was performed in all 17 patients at the junction of femur and fibula graft. The donor peroneal artery was anastomosed according to the level of the recipient site. In eight of the 10 single and folded FVFGs, the peroneal artery was anastomosed to the branch of the femoral artery by endto-end technique and in two cases by end-to-side anastomosis to the femoral artery. In seven cases of femur reconstruction with double FVFGs, seven FVFGs were performed by end-to-side anastomosis to the femoral artery and seven were performed by end-to-end anastomosis to the branch. In most cases, the peroneal veins were anastomosed to the vanae comitantes of the recipient artery and the saphenous vein. The Mann Whitney method was used to compare the variables of two different groups and a p value less than 0.05 was considered to be significant statistically. Results Graft survival A peroneal flap was simultaneously transferred to monitor the survival of the fibula graft in 19 FVFGs and bone scintigram was used in five FVFGs. Five FVFGs (26%) had vascular complications on flap monitoring, necessitating re-exploration within 5 days following the first operation. In three of these cases, the cause was venous thrombosis and in one, arterial thrombosis. After thrombectomy, three flaps survived. One flap had vascular complication due to over-stretching of the perforating vessels and it was removed at the time of re-exploration. In total, 23 out of 24 FVFGs were successfully transferred (96%). Vascular complication had no correlation with selection of the recipient vessels and anastomosis fashion. Rate of union Except in one case (Case 12) with short-term follow-up, 15 of the 16 reconstructed femurs (93%) united primarily following the vascularised fibular transfer. Although exact time of bony union could not be diagnosed on plain radiograph, five femurs united by 6 months, 5 by 1 year and 6 by 2 years. One patient had a residual nonunion of the distal bony junction but weight bearing was possible in this case by using long leg brace. Graft hypertrophy Hypertrophy of the fibular graft was assessed by the modified method of DeBoer (1989) at three levels: proximal third, middle and distal third. An index value of more than 20% was confirmed as graft hypertrophy. In five cases with bony defect less than 1 cm, no hypertrophic change occurred in the transferred onlay graft. Excluding one case (failed vascularised grafting; Case 17), all intercalary

552 Table 1 Patients and methods Case Age and sex Diagnosis Follow-up (mos) Site of involvement Fixation Bone defect (cm) FVFG number Length Recipient vessels Anastomosis Result Immediate complication Stress fracture Hypertrophy (%) Prox/Mid/Dis FWB (mos) 1 35M Infected nonunion 24 Mid-shaft IMR 9 Single 15 DFA ETE Survive Infection No 14/16/14 24 2 37F GCT 48 Distal Plate 11 Double 14 In,14 On FA ETS Survive Infection No 54/17/27, 20/0/8 17 3 39F PTN 12 Distal ESF 4 Double 21, 22 FA ETS Survive None No 30/10/8, 8/0/8 14 4 60M PTN 36 Distal Plate 5 Single 10 FA ETE Survive None No 7/9/8 6.5 5 22M PTN 24 Mid-shaft ESF 7 Folded 13, 13 DFA ETS Survive None No 42/37/32, 17/17/7 15 6 61M PTN 18 Mid-shaft ESF 1 Single 16 FA ETS Survive None No 10/8/8 12 7 46M PTN 18 Distal ESF 7 Double 25, 17 FA ETE Survive None No 32/0/0, 0/0/0 13 8 14M Infected nonunion 48 Neck Fibula 3 Single 10 DFA ETE Survive Infection No 50/50/50 15 9 74F Infected nonunion 24 Neck Fibula 1 Single 5 IGA ETE Survive None No 0/0/0 6 10 23M MFH 24 Mid-shaft IMR 12 Folded 11, 15 DFA ETE Survive None No 50/42/31, 8/0/31 7 11 47F Infected nonunion 12 Mid-shaft Fibula 0 Single 16 On LCA ETE Survive Flap necrosis No 0/0/0 4 12 29M PTN 8 Distal ESF 11 Folded 14, 10 LCA ETE Survive Infection No 30/33/44, 37/22/11 8 13 48F OS 60 Distal Fibula 0 Double 12, 8 M Br ETE Survive Venous thrombus No 0/0/7, 0/0/0 10 14 59M Infected nonunion 31 Distal ESF 0 Double 20, 12 FA ETS Survive Venous thrombus No 0/0/7, 0/7/0 6 15 17M OS 9 Distal Plate 13 Double 16, 16 DGA, M Br ETE Survive Venous thrombus No 0/0/0, 0/0/0 6.5 16 5M OS 17 Distal Plate 15 Single 18 DGA ETE Survive Nonunion No 113/89/89 9 17 55M GCT 12 Distal Plate 11 Double 22, 22 FA, M Br ETS, ETE Survive necrosis Arterial thrombus No 0/0/0, 0/0/0 9 PTN, posttraumatic nonunion; GCT, giant cell tumour; MFH, malignant fibrous histiocytoma; OS, osteosarcoma; ESF, external skeletal fixator; IMR, intramedullary rod; DFA: deep femoral artery; IGA, inferior gluteal artery; LCA, lateral circumflex artery; M Br, muscle brance; ETE, end-to-end anastomoses; ETS, end-to-side anastomoses; FWB, full weight bearing. K. Muramatsu et al.

Femoral reconstruction by single, folded or double free vascularised fibular grafts 553 grafts showed hypertrophy at the final follow-up. There was no statistically significant difference between single grafts (excluding one child case; Case 16, mean ¼ 24.2%) and double or folded grafts (mean ¼ 19.8%) and also between the three levels of the grafted fibula. In six cases with double or folded grafts, significant hypertrophic change occurred in inlay grafts (mean ¼ 28.8%, p ¼ 0:0016) as compared to onlay grafts (mean ¼ 10.7%). Stress fracture We diagnosed a stress fracture seen as a clear fracture line on plain radiographs accompanied with symptoms of localised pain and tenderness. There were stress fractures of the fibula graft in three patients (Case 4, 7, 12) at 10, 10, 12 months postoperatively. All of them healed uneventfully with conservative treatment. All three cases had undergone intercalary double or folded fibula grafting and the grafted fibulas showed hypertrophic change at occurrence of stress fracture. These femur and fibula grafts had angular deformity of 15, 18 and 30 degrees. There was no occurrence of stress fracture in onlay fibula grafts and in cases with angular deformity less than 15 degrees. Time of full weight bearing The time until the patient could walk without braces ranged from four to 24 months (mean ¼ 11 months). There was no significant difference between single fibula transfer cases (mean ¼ 11 months) and double or folded fibula transfer cases (mean ¼ 11 months). On an average, the time of full weight bearing was delayed to 16 months postoperatively in four cases with residual osteomyelitis. In five cases with minimum bone defect and onlay fibula grafting, weight bearing of the affected limb was faster (mean ¼ 8 months) compared to that of intercalary graft cases (mean ¼ 12 months). Complications Four out of five cases with infectious nonunion of the femur had residual localized infection, but all of them healed slowly after debridement and curettage of recurring osteomyelitis and wound care. Two patients had iliac bone grafting for a second time at the junction of the femur and fibula graft. There was no local recurrence of malignant bone tumour. There were no donor-site wound complications and none of the patients perceived serious functional impairment of the donor limb. The clinical record did not suggest any objective findings of an abnormal nature related to the donor limb. Case report A 23-year-old man with a malignant fibrous histiocytoma of the left femur shaft was treated by extensive excision and the resultant 12 cm femur defect was reconstructed by a 26 cm long folded fibula graft (Fig. 1(A) (C)). After the residual femur was anatomically fixed by intramedullar interlocking rod, an 11 cm long fibula graft was placed intramedullary (inlay graft) and 15 cm fibula graft was used as onlay graft. The peroneal vessels were anastomosed to deep femoral vessels in an end-toend fashion. The monitoring peroneal flap survived without any vascular complications. Inlay fibula graft showed remarkable hypertrophy at all levels but onlay graft showed slight hypertrophy. Bony union at both junctions was achieved within 6 months postoperatively. Full weight bearing was permitted 7 months after operation and no stress fracture occurred thereafter. At the follow-up of 2 years, the patient returned to his original employment. Discussion Large femur defects resulting from bone tumours or infectious nonunion present a reconstructive challenge for today s orthopaedic surgeon. Available reconstructive methods have their limitations. The role of free bone tissue transfer for reconstruction of difficult long bone defects has been well recognized. 9,10 Among several options of donor source, fibula graft is probably the most suitable for femur reconstruction because of its structure and minimum donor-site morbidity. In our series, none of the patients perceived functional impairment of the donor limb. Our results of overall union rate of 93% are favorable as compared to other reports. Wood (1990) 4 reported femoral reconstruction by FVFG and 69% of patients healed primarily and 83% of patients achieved bony union following secondary surgery. Hsu et al. (1997) 1 reported FVFG reconstruction after tumour resection and bony union was achieved in 90% of cases at an average of 7.6 months postoperatively. Yajima et al. (1993) 5 also demonstrated high union rate of 95% at an average of 6.4 months after surgery in their cases. Our results also support that FVFG is reliable and stable procedure for the femur reconstruction.

554 K. Muramatsu et al. Figure 1 (Case 10) A 23-year-old man with a malignant fibrous histiocytoma of the left femur shaft. (A) One month postoperatively. After the residual femur was anatomically fixed by intramedullar interlocking rod, 12 cm femur defect was reconstructed by a 26 cm long folded fibula graft. (B) One year postoperatively. Bony union at both junctions was achieved within 6 months postoperatively. Inlay fibula graft showed remarkable hypertrophy at all levels but onlay graft showed slight hypertrophy. (C) Two years postoperatively. Intramedullar interlocking rod was removed without any complication. One of the advantageous phenomena of a vascularised fibula transfer is its ability to hypertrophy. 11 Although the cause of satisfactory hypertrophy is not completely understood, our higher incidence of graft hypertrophy was probably related to the mechanical stimulation due to weight bearing. 12,13 In this series, the intercalary inlay graft showed remarkable hypertrophy as compared to the onlay non-weight-bearing fibula. Interestingly, in the cases with double or folded fibula transfer, inlay graft showed significant hypertrophy compared to onlay graft. These results clearly demonstrated that hypertrophy of fibula graft is associated with mechanical stimulation caused by weight bearing. Femur reconstruction by vascularised fibula graft also has some problems. One of the problems is that this procedure is technically demanding. 5 The femur is located deeper than other long bones and has few main vessels suitable for microsurgical anastomosis. Monitoring of graft vascularity is troublesome because of the inability of the monitoring peroneal flap to reach the skin surface in some cases. In our results, the survival rate of the fibula graft was favorable in 96% of cases (23/24 fibulas). In the literature, there are few papers, which clearly confirm graft survival. Yajima et al. (1993) 5 reported that grafting was successful in 19 out of 20 patients (95%). Vascular complication occurred more frequently at the venous anastomosis site than at the arterial site. We suppose that an important factor for graft survival is not the selection of the artery and anastomosis fashion but selection of recipient vein. Careful postoperative observation of the monitoring flap and immediate re-exploration in complicated cases are also important factors. A second problem is the low mechanical strength of the fibula graft. Some authors presented a late stress fracture of grafted fibula and its reported rate is around 10% (7 16%). 5,7 Stress fracture causes delayed weight bearing of the affected limb. To resolve this problem, we have attempted double or folded vascularised fibula graft in cases with extensive femur defect. However, our results showed that stress fracture could not be prevented by using two fibulas. Interestingly, hypertrophic change was not significantly different between single and double grafts. In the case with double fibula grafting, one is fixed as inlay graft, that mainly supports weight bearing and another as onlay graft that works only as additional support. We believe that the key factor to prevent stress fracture is holding the femur in anatomical alignment using a proper fixation device. Our three cases with stress facture had undergone double fibula graft transfer and the femur had angular deformity more than 15 degrees. If the femur was stabilised in proper alignment, there would have been no cases of stress fracture even with a single fibula graft.

Femoral reconstruction by single, folded or double free vascularised fibular grafts 555 A third problem is prolonged time to full weight bearing. All patients could walk without long leg braces at final follow-up but it took a period of 11 months an average after fibula graft transfer. Yajima (1993) 5 recommended a telescoping external fixator to gradually stress the fibula graft and stimulate hypertrophy. Although we have no experience of this external fixator, rigid internal fixation using intermedullary interlocking nail or supracondylar plate may be helpful to prevent stress fracture and allow early weight bearing on the affected limb. Although there was a high rate of complications involving postoperative venous embolism or late stress fracture in our series, most of them resolved without greatly influencing the final outcome. We cannot conclude that free vascularised fibula graft is the best option for femur reconstruction, as we have no experience with alternative methods of treatment, such as bone transport method. However, in our series, FVFG transfer proved to be a reliable and safe procedure for femur reconstruction. References 1. Hsu RW, Wood MB, Sim FH, Chao EY. Free vascularised fibular grafting for reconstruction after tumour resection. J Bone Joint Surg 1997;79B:36 42. 2. Ihara K, Doi K, Yamamoto M, Kawai S. Free vascularized fibular grafts for large bone defects in the extremities after tumor excision. J Reconstr Microsurg 1998;14:371 6. 3. Duffy GP, Wood MB, Rock MG, Sim FH. Vascularized free fibular transfer combined with autografting for the management of fracture nonunions associated with radiation therapy. J Bone Joint Surg 2000;82A:544 54. 4. Wood MB. Femoral reconstruction by vascularized bone transfer. Microsurgery 1990;11:74 9. 5. Yajima H, Tamai S, Mizumoto S, Ono H. Vascularised fibular grafts for reconstruction of the femur. J Bone Joint Surg 1993;75B:123 8. 6. O Brien BM, Gumley GJ, Dooley BJ, Pribaz JJ. Folded free vascularized fibula transfer. Plast Reconstr Surg 1988;82: 311 8. 7. Hou SM, Liu TK. Reconstruction of skeletal defects in the femur with two-strut free vascularized fibular grafts. J Trauma 1992;33:840 5. 8. de Boer HH, Wood MB. Bone changes in the vascularised fibular graft. J Bone Joint Surg 1989;71B:374 8. 9. Jupiter JB, Bour CJ, May Jr JW. The reconstruction of defects in the femoral shaft with vascularized transfers of fibular bone. J Bone Joint Surg 1987;69A:365 74. 10. Weiland AJ, Daniel RK. Microvascular anastomoses for bone grafts in the treatment of massive defects in bone. J Bone Joint Surg 1979;61A:98 104. 11. Tu YK, Yen CY, Yeh WL, Wang IC, Wang KC, Ueng WN. Reconstruction of posttraumatic long bone defect with free vascularized bone graft: good outcome in 48 patients with 6 years follow-up. Acta Orthop Scand 2001;72:359 64. 12. Mizumoto S, Tamai S, Goshima J, Yajima H, Yoshii T, Fukui A, Masuhara K. Experimental study of vascularized tibiofibula graft in inbred rats: a preliminary report. J Reconstr Microsurg 1986;3:1 11. 13. Muramatsu K, Bishop AT. Cell repopulation in vascularized bone grafts. J Orthop Res 2002;20:772 8.