Expanding the utility of modified vascularized femoral periosteal bone-flaps: An analysis of its form and a comparison with a conventional-bone-graft

Abstract

Background

Vascularized medial femoral condyle (MFC) corticoperiosteal bone-flap is a well-accepted technique when dealing with tissue defects or infection. Its role in refractory conditions and in the possible use for options concerning modifications of this bone-flap compared to a conventional iliac bone graft (conventional-graft) are rarely discussed.

Methods

We reviewed 21 consecutive cases concerning alternatives with some modifications of original MFC bone-flap surgery used to treat refractory conditions with bone defects, necrosis, or infection in the extremities. We present our devised approaches for this boneflap, and especially modifications of the grafted bone (including strut bone, perforator to the vastus medialis muscle, and the use of one vascular pedicle for some bone flaps) as well as the combined use of artificial bone as hybrid bone transplantation. We also compared the clinical results of 21 cases that received a conventional-graft.

Results and Conclusions

Following flap placement, 100% of the nonunion sites healed in an average of 2 months, which was significantly shorter than 5.5 months for the conventional-graft. The results showed the expanding possibility for options with regard to the form and options of this bone-flap as well as the shortening the duration of treatment, especially at the site of an infected distal tibia, insertion of the Achilles tendon on the posterior aspect of calcaneal osteomyelitis, distal end of the clavicle, clavicle or forearm with a bone defect, small bones with refractory conditions, and a femur without implant failure. However, it was not efficient for treating a forearm without bone defect.

1. Introduction

A living periosteum bone graft, including the inner cambium layer, has an advantage of providing an adequate cell source and producing growth factors. A living bone flap has become a popular and powerful tool for refractory conditions (i.e. nonunions, osteonecrosis, infection, or lack of soft tissue coverage). The vascularized medial femoral condyle (MFC) corticoperiosteal bone flap was introduced by Sakai and Doi in 1991 as a means for achieving a bony union under unfavorable conditions.¹ Subsequent reports have described the possibility for alternative options (i.e. combination with a vascularized vastus medialis muscle flap, wrapping around a conventional autologous strut bone graft, and pedicled vascularized bone graft).^1–8 In this study, we describe our modification of grafted bone at the donor site as well as a bone-grafting procedure at the recipient site and the clinical results of an MFC corticoperiosteal bone graft. We also compare the clinical results of nonunions treated with a combination of rigid fixation and iliac bone graft, and discuss the most recommended non-union sites and the best situations for treatment with a bone flap.

2. Materials and methods

2.1. Surgical method

The vascularized corticoperiosteal graft was elevated from the medial femoral condyle, which is based on the articular branch of the descending genicular artery and vein and consists of periosteum.⁹ As popularized by Sakai et al,^1,2 the flap was raised with the periosteum attached to a thin layer of outer cortical bone to protect the cambium layer from injury. Using this approach, flap osteogenicity is maintained without adversely affecting its malleability.

One advantage of a vascularized periosteal bone flap is that there are several possibilities for molding the form. Fig. 1 shows a schematic diagram of the classification of our cases as follows: a pedicle bone flap, bony tailoring into various forms (including dicortical autologous strut bone from the posterior corner or perforator to the vastus medialis muscle), and one vascular pedicle for some bone flaps. In some cases, we wrapped the non-union site over the artificial bone composed of 1) beta-tricalcium phosphate (β-TCP) or 2) hybrid of collagen and low-crystalline calcium phosphate with a vascularized periosteal bone flap. The pedicle is always located close to the femur under the medial border of the vastus medialis muscle and routinely giving one or two branches to the vastus medialis. This allows for the inclusion of small muscle segments that can be useful for collapsing dead spaces or providing a tissue cover. Bony tailoring of the corticoperiosteal portion was an amenable procedure, as this thin bone flap is elastic and readily conforms to the recipient bed configuration. The deep (spongiosa) surface of the flap is carefully cut with an oscillating microbone saw to the periosteum. After an osteotomy, the bone was bent at the osteotomy cutting line. Thereafter, rotation of the bony paddles is possible. The flap would then be wrapped around the non-union site or placed next to the plate. In two cases, we elevated and separated two bone flaps each using the same vascular pedicle. In one case, two bone flaps were grafted in the scaphoid and lunate of the same wrist. In another case, two bone flaps were grafted separately at the anterior and posterior position across the plate, which fixed clavicle.

Fig. 1.

Schematic diagrams of our classification for various forms of vascularized corticoperiosteal grafts from the femur. A: Bone flap with dicortical autologous strut bone from the posterior corner of the femur. B: Bone flap wrapping around the artificial bone composed of β-TCP. C: Pedicle bone flap. D: One vascular pedicle for some bone flaps. E: Bone tailoring into various forms. F: Bone flap that includes a perforator to the vastus medialis muscle.

2.2. Subjects

Our study was done at the Department of Orthopedics, Tokushima Prefectural Central Hospital, Tokushima (Japan). This study was conducted between Apr. 2011 and Oct. 2013. After approval from ethical committee total 42 patients were included in study. A retrospective chart review was performed on 21 consecutive patients who received a vascularized corticoperiosteal graft (Table 1) as well as 21 site- and age-matched consecutive patients who received an iliac bone graft as a comparison (Table 2). A comparative review of the clinical outcomes in patients treated with a bone flap and iliac bone graft was carried out by evaluating the union rate, time to union, and other characteristics. For statistics analysis, a student's t-test and chi-squared test were used with a significance level of p < 0.05 for all tests. The average age of the patient population treated by bone flap was 47 years (range: 14–71 years), which was not significantly different from the average age of patients treated with an iliac bone graft. The locations of the nonunion sites were the center of femur, distal tibia, clavicle, radius, ulna, and small carpal bones. In this study we present our experience with this type of graft in 21 patients, including 18 patients with difficult nonunions, 2 patients with infection at the transitional zone between bone and tendon, and one patient with other bony problems. Eighteen free transfers and three pedicle flaps were used for the 21 patients. Most of the patients were smokers. The duration of the nonunion prior to the use of a vascularized bone graft ranged from 3 months to 30 months (median = 10 months), which was not significantly different from the duration of the nonunion in patients receiving an iliac bone graft (range: 3–24 months; median = 6.8 months).

Table 1.

Details of 21 Cases treated by vascularized periosteal bone flap.

	Location of nonunions	Gender	Age (years)	Smoking	DM	Length of nonunion (months)	Type of the graft	Time to union (months)
1	Femur1	M	35	+	–	9	C	2
2	Femur2	M	39	+	+	30	B + C	1
3	Femur3	M	45	+	–	12	C	1.5
	Average of the group		40			17		1.5
4	Tibia infection1	M	51	–	+	18	F	3.5
5	Tibia2	F	71	–	–	6	E	3
6	Tibia3	M	40	+	–	7	E	3
	Average of the group		54			10		3.2
7	Clavicle1	M	38	–	–	3	E	3
8	Clavicle2	M	53	+	–	9	E	1
9	Clavicle3	M	41	+	–	6	D	1.5
10	Clavicle distal end4	M	41	–	–	6	E	2
	Average of the group		43			6		1.9
11	Radial neck1	M	14	–	–	6	A	1.5
12	Shaft of the radius2	M	48	+	+	8	B	1
13	Shaft of the radius3	M	32	+	–	3	E	1
14	Shaft of the radius infection4	M	64	+	+	12	F	1.5
15	Shaft of the ulna5	M	59	+	+	18	B	1.5
	Average of the group		43			9		1.3
16	Scaphoid and lunate1	M	59	+	–	5	D	2
17	Scaphoid2	M	38	+	–	12	E	2
18	Scaphoid3	M	46	+	–	16	E	2
19	Scaphoid4	M	51	+	–	9	E	3
	Average of the group		49			11		2.3
20	Calcaneal infection1	M	58	+	+	12	B + F	2
21	Calcaneal infection2	M	73	–	–	8	E	2
	Average of the group		66			10		2
	Average		47			10		2.0

Table 2.

Details of 21 cases treated by iliac bone graft.

	Location of nonunions	Gender	Age (years)	Smoking	DM	Length of nonunion (months)	Time to union (months)	Union ± (total number of bone graft)
1	Femur1	M	63	+	–	8	2	+
2	Femur2	M	61	+	–	9	16	− (2 times)
3	Femur3	F	36	+	–	6	1.5	+
4	Femur4	M	54	–	–	6	4	+
5	Femur5	M	69	+	+	6	7	+
6	Femur6	M	26	–	–	6	9	+
7	Femur7	M	56	–	–	7	3.5	+
8	Femur8	M	21	–	–	12	2	+
	Average of the group		48			7.5	5.6	Union Rate 89%
9	Tibial infection1	M	63	–	+	3	6
10	Tibial infection2	M	58	–	–	3	9	− (3 times)
11	Tibia3	M	39	+	+	6	8
	Average of the group		53			4	7.7	Union Rate 60%
12	Clavicle distal end1	M	42	–	–	3	nonunion
13	Clavicle2	M	56	+	+	3	4.5
14	Clavicle3	M	53	–	+	6	9
15	Clavicle4	M	46	+	–	4	3
	Average of the group		49			4	5.5	Union Rate 750%
16	Shaft of the radius1	M	52	+	–	4	2
17	Shaft of the radius2	M	34	+	–	6	1.5
18	Shaft of the radius3	M	54	+	+	3	2.5
	Average of the group		47			4	2	Union Rate 100%
19	Scaphoid1	M	49	–	+	24	5
20	Scaphoid2	M	52	+	–	12	12	− (2 times)
21	Scaphoid3	M	23	–	–	6	3
	Average of the group		41			14	6.7	Union Rate 75%
	Average		48			6.8	5.5	Union Rate 80%

3. Results

3.1. Over-all results (Table 1)

Concerning the elevation of the vascularized corticoperiosteal graft, in one patient (a 14-year-old male with non-union of the radial neck) among 21 patients, the descending genicular artery was completely deficient, and therefore the superior medial genicular artery was used as a vascular pedicle.

Following flap placement, 100% of the nonunion sites healed within an average of 2.0 months, and all cases healed within 3.5 months. Complete healing of mesh-skin graft over the vascularized corticoperiosteal graft was observed in 4 patients. No relapse of osteomyelitis was observed in 3 of all 4 patients (tibia, calcaneous, and forearm). One relapsed calcaneal osteomyelitis, which was subsided after minimum debridement in a patient with decubitus after spinal cord injury (No. 21). However, acceptable healing was observed between calcaneal tuberosity and Achilles tendon in both two cases (No. 20, 21). Almost complete healing of bone flap over the articular surface was observed in carpal bones and adjacent to elbow joint, as functional restoration of the damaged joints as well as radiographic finding was acceptable (No. 11, 16).

Although limited donor site morbidity of this procedure is well known, in one case of a 53-year-old male with clavicle nonunion, a supracondylar femur fracture was encountered, which required anatomically stable internal fixation (No. 8). Thereafter, the patient developed psychological problems and could not keep the rest of his legs. Nevertheless, the rate of complications seemed to be rare, as described in previous reports.^1–6 However, great care must be taken not to cause fracture when harvesting the cortical bone from the poster-medial corner of femur.

3.2. Comparative results with iliac bone-grafting (Tables 1 and 2)

The union rate and time to complete union for the iliac bone grafted cases was 80% (p > 0.05) and 5.5 months (p < 0.01), respectively, compared to cases treated with a bone flap. In the cases treated with an iliac bone graft, all 3 cases of the radius without bone defect healed within an average of 2 months. In one case of a distal end of the clavicle, bone union was not achieved. The union rates of the femur and tibia were 89% and 60%, according to the cases with additional iliac bone graft. The average union rate and time to union for nonunions of the forearm using a bone flap and iliac bone graft were 100 and 100%, and 1.3 and 2 months, respectively. For the clavicle, the union rate and time to union were 100 and 75%, and 1.9 and 5.5 months, respectively. For the tibia, the union rate and time to union were 100 and 60%, and 3.2 and 7.7 months, respectively. Although bone union of the tibia was achieved finally, recurrent osteomyelitis with a fistula was observed in all 2 cases. For the femur, the union rate and time to union were 100 and 89%, and 1.5 and 5.6 months, respectively. However, due to the small number of cases, the data from each site were not significantly different.

4. Case presentation

4.1. Case 1 (Fig. 2)

Fig. 2.

Case 1: A 14-year-old male with nonunion of the proximal radius. A: Pre-operative 1) radiograph and 2) CT scan. The arrow indicates nonunion of the proximal radius with progressive bone absorption. The proximal radial head was united to the ulna, as shown by the white arrow of the CT scan. B: Surgical findings. 1) The arrow shows a large cavity between the proximal radial head and shaft. 2) The dicortical strut bone with vascularized periosteal bone flap elevated by the superior medial genicular artery. 3) The vascular pedicle of the bone flap was elongated by a vein graft. C: Serial radiological findings. 1) Immediately after, and 2–4) 3 months after the surgery. At 3 months after the surgery, the bone union was observed at the posterior region (shown by the white arrow), which was covered by the bone flap, but not the anterior region (shown by the pointed arrow) at this point. D: 1–2) Radiograph and 3–4) Physical findings at 3 years after surgery.

A 14-year-old male had received surgery of the proximal radius and ulna 6 months before our initial consultation. Bone union of the proximal radius was not achieved and the proximal radial head was united to the ulna, as shown in the computed tomography (CT) scan. Alignment of elbow showed cubitus valgus. Because the bone defect was evident, a dicortical strut bone was created from the medial posterior corner of the femur with a vascularized periosteal bone flap. Since the descending genicular artery was completely deficient, the superior medial genicular artery was used as a vascular pedicle. The vascular pedicle was small and short, and therefore a vein graft was necessary to complete this case. The vascularized dicortical strut bone was inserted into both the proximal and distal cavities of the radius, and the periosteal bone flap was placed on the posterior region of the nonunion site. Three months after the surgery, a bony union was observed at the posterior region that was covered by the bone flap, but not the anterior region. Twelve months after the surgery, the plate was removed and additional fixation with another headless compression screw was performed to prevent impending fracture. Three years after the surgery, the skeletal growth was complete, the alignment of the elbow was normalized, and active movement of the elbow and forearm were maintained within an almost normal range without pain. The Mayo Elbow Performance score was 100.

4.2. Case 2 (Fig. 3)

Fig. 3.

Case 2: A 59-year-old male with nonunion of the shaft of the ulna. A: Pre-operative radiograph. B: Surgical findings. The ulna was stabilized with a new locking plate and the scar tissue around the nonunion was removed. 1) After removal of the fibro-osseous scar tissue at the nonunion site. 2) An artificial bone was grafted into the cavity. 3) A vascularized periosteal bone flap was fixed over the artificial bone. C: A CT scan at 1.5 months after surgery. D: A schematic diagram and serial radiological findings of the repaired site. Serial radiograph (a, b, c) at 1, 4, and 8 months after the surgery in order from the left side. a: Site of the vascularized periosteal bone flap over artificial bone. b: Opposite site of the locking plate that is adjacent to sites a and c. c: Conventional-bone-grafted site, which is opposite the site covered by the bone flap across the locking plate. The bone union was serially confirmed in order from site a, b, and c.

A 59-year-old male had received osteosynthesis of the ulna 18 months before our first consultation. The radiograph showed a hypertrophic nonunion and loosening of screws. Because the patient was a heavy manual worker with heavy smoking habits and the nonunion had been untreated for a long time, we applied a vascularized bone flap in addition to the more rigid fixation. An artificial bone was grafted under the vascularized periosteal bone flap. A periosteum flap was fixed by a bone anchor, which served as a marker for the site of the bone flap. Bone union was achieved with the periosteal bone flap area after 1.5 months. However, the opposite site that had received an autologous cancelous bone graft required a much longer time to achieve bone union (6–8 months). Serial progression of the bone union and absorption of the artificial bone were recognized by radiography.

4.3. Case 3 (Fig. 4)

Fig. 4.

Case 3: A 59-year-old male with Kienböck's Disease and delayed-union of the scaphoid. A: Pre-operative MRI. B. Intraoperative view^1,2 during harvesting of a medial condyle corticoperiosteal flap showing the two bone flaps.^3,4 The two bone flaps were placed in the spaces, which were created after curettage of the sclerotic lunate and the scar tissue at the fracture site of the scaphoid. The one pedicle for two bone flaps was anastomosed at the snuffbox. C: Radiograph at 6 months after surgery.

A 59-year-old male with Kienböck's Disease and delayed-union of the scaphoid was treated successfully with one pedicle for two bone flaps. Bone union and revascularization of the scaphoid and lunate were confirmed by CT and magnetic resonance imaging (MRI) as well as by radiographic findings. A decrease in wrist pain and minimum improvement of range of motion (ROM) were clinically confirmed at one year follow-up. The Mayo Wrist score remained at 65, indicating satisfactory (60–80) improvement due to the restriction of the ROM.

4.4. Case 4 (Fig. 5)

Fig. 5.

Case 4: A 39-year-old male with nonunion of the femur. A: Pre-operative radiograph. B: C. Surgical findings. A pedicled vascularized bone graft from the medial supracondylar region of the femur was fixed over artificial bone, which was packed in the spaces created after curettage of the nonunion tissue. D: Radiograph and CT at 1 month after surgery showing the bone union at the bone flap area.

A 39-year-old male with nonunion of the femur had received closed intramedullary nailing 2.5 years before our initial consultation. Local exploration confirmed that the pedicled bone flaps induced bone union within 1 month of treatment. An increase in the transverse diameter and bone union area was also observed due to the thickening of the bone flap.

4.5. Case 5 (Fig. 6)

Fig. 6.

Case 5: A 51-year-old man with chronic post-traumatic osteomyelitis and infected nonunion of the tibia. A: Pre-operative photograph and radiograph. The combined muscular (vastus medialis) and corticoperiosteal flap was used to solve a compound bone-soft tissue loss. B: Postoperative radiograph. This radiograph at 6 weeks after free tissue transplantation showing Ilizarov fixator to correct deformity and contracture. C: Postoperative photograph, radiograph, and CT. At 3 years after the reconstructive surgery, the extremity was capable of full load-bearing with complete bone union and active movement of his ankle.

A 51-year-old male with chronic post-traumatic osteomyelitis and infected nonunion of the tibia had received plate fixation after open fracture 18 months prior to our initial consultation. Because of this presentation, four surgical procedures were performed during 18 months, with two procedures performed to change the implant and two procedures for debridement. The plate was removed and external fixation was performed by means of a Hoffman fixator. To control infection, cover a fistula with an evidence of Pseudomonas aeruginosa as well as bone union, we preferred a vascularized corticoperiosteal graft with the vastus medialis muscle covered by a mesh-skin graft. After the pseudarthrosis gap was cleared out, a conventional autologous bone graft was packed into the cavity. The vascularized bone flap was grafted over the free autologous bone graft. After stabilization of the free bone and muscle flap (6 weeks after surgery), a pes equinus and claw toe deformity was corrected by a gradual correction method using an Ilizarov fixator. The extremity was capable of full load-bearing. Three years after the reconstructive surgery, the patient was satisfied with the result and the active movement of his ankle was preserved with minor pain. The clinical rating system for the ankle and hindfoot was 82, and the Foot and Ankle Disability Index (FADI) total score was 80.8. The main cause of the deduction was due to the limited ROM.

4.6. Case 6 (Fig. 7)

Fig. 7.

Case 6: A 41-year-old male with nonunion of the distal end of the clavicle. A: Pre-operative radiograph. B: Postoperative radiograph at 2 months after the vascularized corticoperiosteal graft. C: Postoperative CT at 2 months after the reconstructive surgery. The bone union was observed from the anterior portion covered by the bone flap, but not yet at the posterior portion.

A 41-year-old male had received conservative treatment for a fracture at the distal end of the clavicle six months prior to our initial consultation. As a bone union was not achieved, a vascularized periosteal bone flap was placed on the anterior region of the nonunion site in addition to the conventional autologous cancelous bone graft. Two months after surgery, bone union was observed on the anterior portion that was covered by the bone flap, although absorption of autologous cancelous bone was observed at the posterior region. We were able to safely remove the hook plate five months after surgery.

4.7. Case 7 (Fig. 8)

Fig. 8.

Case 7: A 57-year-old male with chronic post-traumatic left calcaneal osteomyelitis with a fistula. A: Postoperative 1, photograph, 2, radiograph after the first surgery. The HA block was packed as delivery systems for the sustained release of antibiotics into the dead space after the curettage. B: Postoperative 1, photograph, 2, radiograph after the second surgery. The vascularized corticoperiosteal bone flap (white arrow head) with the vastus medialis muscle was grafted between the Achilles tendon and calcaneal tuberosity. Mesh-skin graft was done on the periosteal bone flap. The non-absorbable HA block (Fig. 8. A-2 black arrow) was replaced by the absorbable hybrid of collagen and low-crystalline calcium phosphate (Fig. 8. B-2 black arrow) to fill the dead space.

A 57-year-old male had received Kirschner wire fixation for chronic post-traumatic left calcaneal osteomyelitis with a fistula due to open fracture twelve months prior to our initial consultation. Because of this presentation, three surgical procedures for debridement were performed during 12 months to fix the detached Achilles tendon by a bone anchor. To control infection, we planned the staged surgery. As a first step, we used porous blocks of calcium hydroxyapatite ceramic (HA block) as delivery systems for the sustained release of antibiotics into the dead space created by the aggressive debridement. At three weeks after the first surgery, the vascularized corticoperiosteal bone flap was grafted between the Achilles tendon and calcaneal tuberosity to treat the avulsion of Achilles tendon and prevent the impending pathologic avulsion fracture of a calcaneal tuberosity. We also used the vastus medialis muscle and artificial hybrid bone to fill the dead space. We preferred a mesh-skin graft directly over the vascularized bone flap to cover the soft tissue defect. One year after the reconstructive surgery, the patient was satisfied with the result, and the active movement of his ankle was preserved with minor pain. The FADI total score was 95.2.

5. Discussion

Previous reports have shown excellent clinical results of the MFC bone flap in persistent nonunions of the upper and lower extremities, including small bones.^{1–8,10–15} Small bone defect within 3 × 3 cm or poorly vascularized nonunions in which bony structural loss is minimal are good indications of the MFC bone flap. The MFC flap is usually elevated one of the following patterns, periosteal flap with a corticocancellous block bone within 3 × 3 cm, as an osteocartilaginous graft for management of articular defects or a thin corticocancellous sheet of bone that can be wrapped around tubular bones. Some studies have indicated the expanding possibilities and evolving concepts of this vascularized periosteal flap. Several techniques have been previously reported, including a pedicled bone flap, bony tailoring into various forms (including bi/tricortical grafts for structural support surrounded by a periosteal flap), and perforator to the vastus medialis muscle.^2–8 A study using cadavers showed that the distal half of the femur can be covered by a pedicled bone flap.⁸ The clinical results of our cases showed the expanding possibility for many options concerning its form as well as options for a vascularized periosteal bone flap. In the present study, we reported the technique of using one vascular pedicle for certain bone flaps as well as a hybrid bone graft composed of artificial bone and a vascularized periosteal bone flap. To the best of our knowledge, these procedures have not been previously reported. In four patients, we found early bone union at the site of the artificial bone composed of 1) β-TCP or 2) hybrid of collagen and low-crystalline calcium phosphate, when a vascularized periosteal bone flap was used compared to a conventional autologous bone graft. A surrounding periosteal flap seemed to play an important role in revascularization the artificial bone as well as the structural graft. This hybrid bone transplantation may be a minimally invasive alternative for a small bone defect.

A transitional zone between bone and tendon is well known to composed of the 4 tissue zones (Zone 1 – ligament; zone 2 – fibrocartilage; zone 3 – mineralized fibrocartilage; zone 4 – bone). Healing of the border between bone and tendon is somewhat difficult, and this 4-zonal structure is also difficult to be recreated. We treated the 2 cases with avulsion of Achilles tendon from calcaneal tuberosity by the vascularized corticoperiosteal bone flap. The function of ankle of these 2 patients recovered well. The living bone flap seemed to accelerate the healing of the border between bone and tendon. Further pathological study is necessary to see the recreation the 4-zonal structure.

Relapse of osteomyelitis was observed in 1 of 4 patients (tibia, calcaneous, and forearm). This relapsed calcaneal osteomyelitis with decubitus after spinal cord injury was treated by a single surgery. Other 3 cases were treated by staged surgery composed of bone flap a few weeks after radical debridement. We speculated that the recurrence was due to a single-staged surgery, in which bone flap and debridement were performed at the same setting. Therefore, we recommend staged surgery for osteomyelitis with fistula.

The combination of rigid fixation and an iliac bone graft can effectively resolve most nonunion cases. Comparative results of our cases showed the effectiveness in shortening the duration of treatment, especially at the site of 1) an infected distal tibia, distal end of clavicle and calcaneus, 2) forearm or clavicle with bone defect, and 3) femur without implant failure. These findings underscore several site-specific advantages. In the cases of infected distal tibia treated with free bone graft, recurrent discharge from the fistula can occur up to several years after the bone union, and therefore we recommended a living tissue graft. In the case of the clavicle, rigid fixation of the distal end is difficult. Even with a hook plate, the bone union is rushed, as the implant should be removed earlier (usually within 5 months). Therefore, we recommend a living bone flap for early union at the distal end of the clavicle. Although an iliac bone graft with rigid fixation can induce bone union for most small bone defects, a comparative radiograph taken at long term follow-up showed differences in bone atrophy between the bone flap and iliac bone graft procedures (Fig. 9). Even with atrophic nonunion of the bone, a forearm without a bone defect seemed to be resolved by rigid fixation and iliac bone graft. For the treatment of a nonunion femur, revision of the femoral nail seemed to induce bone union. Therefore, we recommend a pedicled vascularized bone graft for the cases without implant failure because of the benefits of the minimally invasive surgery and the early bone union.

Fig. 9.

A comparative radiograph at 12 months after surgery in patients with a nonunion of the clavicle and implant failure treated by (A) iliac bone graft and (B) bone flap. The arrow shows bone atrophy in the case treated by (A-2) iliac bone graft, but not (B-2) bone flap.

Concerning the bone union, a vascularized periosteal seemed to be the most important element from several previous studies. Berggren et al demonstrated experimentally that the preservation of the periosteal blood supply alone, but not the medullary blood supply, can result in complete bone graft survival, even when the graft is placed in a poorly vascularized tissue bed.¹⁶ In a clinical prospective study, Vegas et al showed the same effectiveness of both periosteal-only and corticoperiosteal microvascular transfers from the medial femoral condyle.¹⁷ The periosteum is a condensed and highly vascular fibrous tissue outer layer, which binds periosteum to bone though Sharpey's fibers. The inner cambium layer contains osteoprogenitor cells that differentiate into osteoblasts during growth or repair. The upregulation of bone morphogenetic proteins (BMPs), which induce differentiation into osteoblasts or chondrocytes at the cambium layer, is also well known. Because of these osteogenic capabilities, periosteal grafts have a significant potential in the reconstruction of bone defects, and it has been shown experimentally that the association of a periosteal flap and cancelous bone is a better means by which to produce compact bone than either a vascularized periosteal flap alone or an isolated cancelous bone graft. Furthermore, the active healing potential over articular surface damage may be due to the capability of chondrogenic differentiation of the periosteum-derived cells.

In conclusion, 100% of the nonunion sites healed following flap placement. The possibility for many options of molding the form of the vascularized periosteal bone flap, including one vascular pedicle for some bone flaps, is a major advantage of this approach. This hybrid bone transplantation may be a minimally invasive alternative for a small bone defect. Our study shows the advantages of this bone flap surgery by comparing it to a conventional method.

Conflicts of interest

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. No funds were received in support of this study. All named authors hereby declare that we have no conflicts of interests of interest to disclose.