Master Techniques in Otolaryngology - Head and Neck Surgery: Reconstructive Surgery, 1ed.

14. Reconstruction of the Composite Oromandibular Defect: Fibular Free Flap

Richard E. Hayden

INTRODUCTION

In 1973, Daniels and Taylor described the transfer of the first free skin flap, the groin flap pedicled on the superficial circumflex iliac artery (SCIA). Panje, Baker, et al. were the first to bring this free flap to head and neck surgery in 1975 when they used the free groin flap for tongue reconstruction. As a medical student, I worked with Rollin Daniels, MD at McGill University, performing microvascular reattachment of fingers. As a resident at the University of Toronto in the late 1970s, I studied under John Fredrickson, MD. Ostrup and Fredrickson had performed the first vascularized mandible reconstruction when they successfully transferred free vascularized rib for mandibular reconstruction in dogs in 1975. I worked with Fredrickson in 1977 using free groin flaps with attached iliac crest bone to repair oromandibular defects. The groin flap with bone was not successful as the SCIA reliably supplied the skin of the groin but not the iliac bone.

The modern history of oromandibular reconstruction using free flaps dates to 1978 when Taylor introduced the composite iliac crest flap pedicled on the deep circumflex iliac artery (DCIA). The DCIA not the SCIA supplied the iliac crest bone, and musculocutaneous perforators from the DCIA also supplied the skin overlying the iliac crest. His technique provided, for the first time, a reliable osseocutaneous flap with bone stock that could closely reduplicate the contours of the original mandible up to hemimandibular and angle-to-angle defects. An additional feature was that the bone stock was sufficient to easily accept osseointegrated implants. However, the soft tissue component was usually excessively bulky for most intraoral soft tissue reconstructions. It was comprised of a cuff of internal oblique, external oblique, and transversalis muscles with overlying adipose tissue and skin. The cutaneous branches of the DCIA coursed through this triad of abdominal muscles, medial to the iliac crest, to supply the overlying adipose tissue and skin. Therefore, a cuff of these muscles was required with the bone to protect the cutaneous blood supply. This, combined with the frequently slack abdominal wall and abdominal adipose tissue distribution found in many of these patients requiring oromandibular reconstruction, frequently led to an excessively bulky soft tissue component. This bulk was often a poor match for the intraoral soft tissue defect. Also, the requirement that this bulky soft tissue had to remain aligned with the adjacent iliac bone to avoid torsion or compression of the cutaneous blood supply led to a suboptimal intraoral soft tissue reconstruction.

Ramasastry identified the ascending branch of the DCIA as a separate vascular supply to the internal oblique muscle. Urken combined the findings of Taylor and Ramasastry to develop a composite flap that used the abundant reliable iliac bone covered intraorally with the equally reliable internal oblique muscle. He then externalized the excessively thick musculocutaneous component incorporating it as an external monitoring segment into the closure of the neck wound, ultimately discarding it at a later time. This solved the problems associated with attempting to close the intraoral defect with an excessively bulky muscle–adipose tissue–skin component. There remained, however, a geographic mandate wherein the size of the attached internal oblique muscle paddle limited the size of the intraoral soft tissue defect that could be reliably closed. Other issues that plagued the iliac crest flap were the relatively short vascular pedicle and the significant morbidity associated with the donor site. Nonetheless, it remained my first choice for oromandibular reconstruction from 1978 to 1985.

The radial forearm free flap was introduced to the West in 1981 and quickly became one of the most frequently used soft tissue free flaps for reconstruction of head and neck surgical defects for the next two decades. Harvest of a portion of the radius bone with this flap was popularized by Soutar in the early 1980s, but the small amount of available bone stock (one-third of the circumference of the radius) limited its application and therefore its popularity for oromandibular reconstruction. The potential morbidity associated with a pathologic fracture of the radius deterred many surgeons.

The free scapular skin flap was introduced by Dos Santos in 1980, and the free parascapular skin flap was introduced the following year by Nassif. Both flaps were vascularized by the cutaneous branches of the circumflex scapular artery. In 1985, Swartz and Banis developed the scapular composite osseocutaneous flap based on the circumflex scapular artery that they demonstrated supplied not only the two skin flaps but also the lateral border of the scapula. This flap provided a dramatic advance for oromandibular reconstruction. The skin paddle(s) could be huge, were usually relatively thin and pliable, and could be rotated 180 degrees on the axis of the bone. The skin paddle(s) could easily be contoured to intraoral or facial defects or to through-and-through defects. The lateral border of the scapula provided less bone stock than did the iliac crest but much more than did the radius. The average adult donor scapula could easily duplicate a hemimandible. The bone stock was inadequate for reliable osseointegration in the 1980s and 1990s although the smaller implants currently available make osseointegration in this setting modestly successful. The donor morbidity was less than that produced by the iliac crest flap and certainly much less than that associated with a potential pathologic fracture of the radius, an inherent risk with the radial forearm flap. The vascular pedicle was longer and of greater caliber than that of the iliac crest flap. The angular branch of the thoracodorsal system was identified as a separate vascular supply to the tip of the scapula by Coleman in 1992. This enhanced the options for contouring the bone segment of the scapular flap. The biggest negative associated with this flap for oromandibular reconstruction was the need for patient positioning to allow for flap harvest and head and neck surgical ablation. Synchronous two-team ablation and flap harvest can be done with the patient on a beanbag, in lateral 45-degree position, head turned on a neurosurgical headholder, but exposure is somewhat compromised for each team, and a synchronous contralateral neck dissection is extremely difficult. Asynchronous ablation and reconstruction demand a time-consuming turning of the patient during the case. Nonetheless, this was my absolute favorite flap for oromandibular reconstruction between 1985 and 1990.

Taylor had also introduced the free fibular flap in 1978 based on the peroneal artery and its venae comitantes. It was used primarily for long bone replacement until 1990 when Hidalgo demonstrated its utility for oromandibular reconstruction. It has been my preferred choice for oromandibular reconstruction since 1990 and by 2000 was unquestionably the gold standard of reconstructive surgeons worldwide.

The fibular flap provides up to 25 cm of vascularized bone, enough to reconstruct almost any defect of the mandible. It is important to maintain 6 cm of fibula and interosseous membrane proximally and distally to maintain the integrity of the knee and ankle joints. The bone is densely cortical, and I began placing osseoin-tegrated implants at the primary surgery in 1990. Implant retention over decades has been excellent. The skin of the proximal lateral calf is often supplied by musculocutaneous perforators through the soleus muscles. The skin of the distal lateral calf is supplied by septocutaneous branches of the peroneal artery, which traverse the posterior crural septum between the peroneal musculature and the soleus muscle. These branches are more common in the lower half of the septum. Preoperative assessment of the cutaneous branches makes the harvest quicker and more reliable (see Preoperative Planning). Almost the entire surface of the lateral lower leg can be reliably harvested with rectangular skin paddles up to 20 × 10 cm. This is more than enough skin for almost all intraoral or through-and-through defects encountered. Multiple skin paddles are available depending on the number of cutaneous branches. In the patient with only a single cutaneous branch, large skin paddles are still possible and segmental deepithelialization of the paddle can produce multiple skin paddles for synchronous intraoral and extraoral coverage. The skin of the calf is usually thin and pliable but can be thick in obese patients; but this situation also applies to iliac crest and scapular flaps in these patients.

The quick and easy harvest is performed remote from the head and neck with the patient in the supine position. The donor morbidity is far less than either the iliac crest or scapula donor site. The vascular pedicle has large caliber vessels. The length of the vascular pedicle is somewhat relative. The absolute length of the peroneal pedicle from the proximal fibula is relatively short, but if distal fibula is to be used, which is most common, the proximal bone is discarded, lengthening the pedicle considerably. Conversely, when reconstructing an extensive mandibular defect, the bone is not shortened and the vascular pedicle goes directly to vessels in the ipsilateral neck while the proximal fibula replaces the ipsilateral mandible and the distal fibula replaces the anterior and contralateral mandible.

Contouring the fibula with its thick cortical bone requires ostectomies as “green stick” osteotomies are impossible. Preplanning (see Preoperative Planning) optimizes vascular pedicle length and its geometry with neck vessels. Elliptical skin paddles up to 6 cm wide and 20 cm long can usually be closed primarily leaving a vertical scar in the posterior lateral lower leg. Larger skin paddles require a skin graft.

The ease of harvest, reliability, huge skin and bone units available, adequate bone stock for predictable osseointegration, and low donor morbidity combine to make this the most useful and popular reconstructive option for oromandibular defects in the modern era.

When considering a fibula flap for reconstruction of an oromandibular defect, I take the usual thorough history, including coverage of issues that affect postoperative care and wound healing such as weight loss, diabetes, hypothyroidism, and substance abuse. It is also important to ask specifically about a history of previous fractures to the lower limbs, total knee replacement, or a history of deep venous thrombosis since this aspect of the medical history is frequently overlooked or forgotten by the patient. A history of diabetes or peripheral vascular disease is also important but I have found that the surprises that I have encountered are almost always associated with recanalized peroneal veins following deep vein thrombosis (known or unknown) or fracture. Any history of gait disturbance, unilateral ankle swelling, pain, or discomfort needs to be acknowledged, considered, and recorded.

In addition to the usual complete physical examination, I pay close attention to the lower extremities. In particular, evidence of venous stasis or vascular insufficiency is noted. Significant varicosities can signal significant deep venous problems. Pedal temperature and pulses must be examined. Stigmata of previous surgery, trauma, or burns are significant. The symmetry of the calves and ankles, the thickness, and hair-bearing qualities of the lower leg skin are assessed. Most important, perhaps is an assessment of the vascular supply to each lower leg. There are rare anatomic vascular anomalies in which the ever-present peroneal artery may be dominant to the foot (peronea magna). This is the artery that will be harvested. Simple assessment of pedal pulses will not demonstrate these cases. I assess and record the skin supply and three-vessel runoff of each lower leg with objective, recorded Doppler and CT angiography. This is one of the deciding factors on choosing which fibula to harvest (see Preoperative Planning). This identifies any anatomic anomalies including arterial stenoses associated with peripheral vascular disease while also illustrating the cutaneous branches and their anatomic correlates.

With up to 200 cm2 of skin and 25 cm of bone available, the fibula flap is indicated for almost all oroman-dibular reconstructions. The bone is sufficient in length to reconstruct almost total mandible defects and can be “double barreled” to provide mass sufficient to reduplicate the dentulous mandible for anterior and lateral defects. Before I started to harvest very large rectangular skin paddles, I treated through-and-through oral–facial–mandibular defects differently. For small intraoral combined with large extraoral defects, I would use the vascularized flexor hallucis longus (FHL) muscle attached to the fibula to fill the intraoral defect leaving the skin paddle to close the large extraoral facial defect. Intraoral mucosa, both buccal–labial and lingual, was sewn to the FHL muscle much as was done with the intraoral internal oblique muscle when it covered iliac crest bone replacing the mandible. These muscles quickly atrophy, leaving a tight mucosal surface over the intraoral neomandible. For defects that had large synchronous intraoral and facial soft tissue defects, I would use the scapula flap. Since I started harvesting very large skin paddles with the fibula, I find less use for the scapula flap and hardly ever use muscle only intraorally. The large skin paddles can be divided as separate perforator flaps when multiple cutaneous branches exist or a large skin paddle can simply be divided into separate skin surfaces by deepithelializing intervening segments. This allows for closure of large intraoral and facial defects with bone replacement. Very large intraoral defects such as subtotal or total glossectomy defects in concert with mandibular and facial skin defects are still repaired with scapular flaps incorporating very large scapular and parascapular skin paddles.

The osseocutaneous fibular flap as I now use it can easily replace hemitongue, symphysis-to-condyle mandibular defects even if they include the buccal mucosa. Equally easy to replace is the body-to-body mandibular defect with loss of the anterior half of the tongue even if the defect includes the chin. A lateral defect involving through-and-through cheek, facial defect, lateral mandible, floor of the mouth, and buccal lining is easy to replace, but as this defect extends more medial to include substantial amounts of mobile tongue, this technique loses its effectiveness since tongue tethering becomes a bigger factor. Intraoral defects can also include the oropharynx. The fibula composite flap can easily provide replacement of the tongue base, tonsil, or palate region along with the lateral and posterior mandible. Importantly, these defects rarely include an external facial defect. Also important is the need to preplan these reconstructions so that the distal fibula is at the condylar location, thereby allowing much more freedom with the orientation of the vascular pedicle (see Preoperative Planning). This technique also places the skin with the best blood supply (that of the distal leg) in the region of the soft tissue defect, thereby minimizing the risk of fistula formation. This is doubly important when reconstructing through-and-through defects. The most reliable skin paddle should always be used for the intraoral closure, thereby reducing the risk of fistula formation that could threaten the vascular anastomoses and thereby the entire reconstruction. Less reliable skin used for the external facial defect, if compromised, can be more easily replaced while preserving the microvascular anastomoses in the neck, the vascularized neomandible, and intraoral flap.

Absolute:	Insufficient vascularity to the lower leg and foot
Relative:	Only one lower leg or foot
	Severe atherosclerotic changes on angiography
	History of deep venous thrombosis: noninvasive venography recommended
	History of trauma, fractures, or surgery on lower extremity: may choose contralateral leg, evaluate carefully

Preoperative planning is more important with this flap than with most. The vascular supply to the potential flap is paramount, so if one leg is deficient and the other is not, the leg with the adequate blood supply is obviously chosen. In such cases, the issues reviewed below are still important although ideal design may have to be compromised.

In those patients who have adequate vascularity to both potential fibula flaps, there are important preoperative considerations required to maximize the utility of the flap. A fibula flap from either leg can be used to successfully reconstruct oromandibular defects. However, one side is always superior for each defect. There are constants and variables that need to be systematically assessed preoperatively in each case to determine the best choice of a flap.

The constants are as follows:

• The lateral aspect of the fibula is best for plate and screw application and should represent the external aspect of the neomandible.

• The peroneal vascular pedicle runs on the posteromedial aspect of the fibula so that this medial surface of the fibula should be avoided or protected from excessive manipulation or instrumentation and should never be the recipient of plates or screws.

• The cutaneous branches leave the peroneal vascular pedicle on the medial aspect of the fibula and course posterior to the bone toward the skin. In the proximal lower leg, they may perforate the soleus muscle on their way to the skin. In the more distal lower leg, most of these cutaneous branches run directly from the peroneal vessels into the posterior crural septum and on to the skin (septocutaneous). Some cutaneous branches can run into the intermuscular septum and then posteriorly through the lateral aspect of the soleus muscle to the skin (septomusculocutaneous).

• The anterior aspect of the fibula is the best location for placing osseointegrated implants especially at the time of primary reconstructive surgery. This approach minimizes risk to the vascular pedicle medial to the bone and to the cutaneous branches posteriorly. This means that the anterior aspect of the fibula is the ideal gingival aspect of the neomandible.

• The natural location of the skin paddle is posterolateral to the bone (Fig. 14.1).

• The distance between the peroneal pedicle on the medial aspect of the fibula and the skin, that is, the length of the cutaneous branches is finite. Therefore, rotation of the skin paddle anteriorly over the anterior aspect of the bone is also limited if one is to avoid excessive traction or compression of these cutaneous branches and subsequent vascular compromise of the skin paddle. There is usually sufficient arc of rotation available for the skin paddle to reconstruct intraoral defects from buccal sulcus to floor of the mouth, but lingual defects often put too much strain on the skin paddle brought from the lateral side (Fig. 14.2). Rotating the skin paddle posteriorly over the posterior aspect of the bone puts much less strain on the cutaneous branches since they are not draped over the lateral and anterior aspects of the fibula and is therefore useful for lingual and more medial defects. Therefore, for large intraoral soft tissue defects, the bone can be turned upside down so that the posterior aspect of the fibula now represents the gingival aspect of the neomandible, and this orientation gives much more latitude to the placement of the skin paddle especially for closing large intraoral defects involving the tongue. The posterior aspect of the fibula is less favorable for placing implants at the time of primary surgery but is equally good for delayed implant placement (Fig. 14.3).

• Skin paddles can be rotated 90 degrees on the axis of the fibula around their perforator(s) allowing for greater leeway in closing complex soft tissue defects. This is especially true for complex body-body mandibular defects with large anterior intraoral and extraoral soft tissue defects. It is critical, in such cases, to design the osteotomies so that the cutaneous supply, if there is only one, is geographically associated with the appropriate piece of fibula (usually the central segment) (Fig. 14.4).

• The best cutaneous branches usually exit the intermuscular septum in the distal leg. Occasionally, the only cutaneous perforator is found about 6 cm proximal to the lateral malleolus, in other words, at the point where fibula bone harvest ends. It does not mean that the skin paddle harvest must end 6 cm above the lateral malleolus. In these cases, design the skin paddle to extend distally to incorporate this single perforator and be very careful when making the distal bone cut. This single septocutaneous branch can carry a very large skin paddle extending all the way to the proximal calf.

• The vascular pedicle is short before it joins the proximal bone.

• The nutrient vessel is less important for mandible reconstruction than it may be for long bone reconstruction since ostectomies are common for mandible reconstruction. The blood supply to the osteotomized bone is primarily periosteal and segmentally provided by the peroneal artery. This allows for sections of bone to be removed without compromising more distal bone. Though not imperative, when parts of the fibular bone are removed, I recommend preserving the intervening periosteum between bone segments, especially that along the medial aspect of the fibula near the vascular pedicle (Fig. 14.5A). Try to avoid creating bone segments less than 2 cm in length as avascular segments within a viable flap have been reported. There would be some concern regarding the viability of the fibula proximal to the attachment of the peroneal vascular pedicle if a segment of periosteum were to be circumferentially removed. The proximal bone relies on the integrity of the periosteum associated with the peroneal vessels to survive..

• When employing the double-barrel technique, sufficient bone needs to be removed between the two “barrels” of fibula to allow the peroneal vascular pedicle to turn without torsion, kinking, or compression. Otherwise, the distal “barrel” might be devascularized (Fig. 14.5B).

• When using a double-barrel technique, be sure that the segment of fibula to be used for the gingival aspect of the neomandible is the “barrel” associated with the cutaneous branch. A segment (barrel) of fibula not associated with the cutaneous supply can be used for the basilar aspect of the neomandible.

• The left and right legs provide flaps with different geometry.

FIGURE 14.1 Leg flexed at the hip and knee illustrating flap harvest positioning. Note that the posterior crural septum starts near the junction of the proximal and middle thirds of the fibula and diverges posteriorly from the fibula as it runs to the distal leg. The most reliable cutaneous perforators traverse this septum, and any skin paddle should be oriented over the septum and not the fibula. Six centimeters of fibula should be preserved proximally and distally to ensure knee and ankle stability. The location of the peroneal nerve is indicated.

FIGURE 14.2 A right fibula osseocutaneous flap in its natural orientation illustrates the peroneal vascular pedicle from the proximal lower leg. The flat lateral border of the fibula is best suited for plate and screw fixation and therefore the best surface of the lateral aspect of the neomandible. The anterior aspect of the fibula is ideal for intraoperative placement of osseointe-grated implants. The peroneal vascular pedicle can be seen coursing along the posteromedial aspect of the fibula between tibialis posterior and FHL muscle remnants. Cutaneous perforators can be seen traversing the intermuscular septum on the posterior aspect of the fibula. The arrow illustrates the limited arc of rotation of the skin paddle dictated by the length of the cutaneous perforators when the skin paddle is rotated anterior to the bone. This is not of concern when the soft tissue defect is extraoral or intraoral only to the floor of the mouth. More medial intraoral defects demand a better flap orientation.

FIGURE 14.3 A right fibula osteocutaneous flap rotated 180 degrees along its long axis, still illustrating the peroneal vascular pedicle from the proximal lower leg. The lateral aspect of the right fibula (the lateral aspect of the neomandible) is now oriented toward the left, and the posterior aspect of the fibula is uppermost. Osseointegrated implants can be placed in the posterior aspect of the bone, but placement should be delayed. Intraoperative implant placement in the posterior bone increases the risk of vascular compromise to the flap. The arrowillustrates the much more generous arc of rotation of the skin paddle available when the skin paddle is rotated posterior to the bone. This rotation is better suited for reconstruction of more lingual (medial) defects. Figures 14.2 and 14.3 together illustrate only two of the four variables available within each leg during preoperative selection of which leg to harvest. These two variables, anterior surface of the fibula placed superior or inferior in the neomandible, are doubled when one turns the entire flap around end-to-end. Remember, depending on available recipient vessels, the vascular pedicle can exit the neomandible posterior or anterior in the neck and if anterior can be anastomosed to ipsilateral or contralateral neck vessels. These variables (options) are then doubled again if both legs have potential for a fibula flap.

FIGURE 14.4 A right fibula osseocutaneous flap used to reconstruct an angle-to-angle mandibular defect, requiring significant soft tissue replacement of both intraoral and facial defects. The flap has been designed around a major cutaneous perforator that is associated with that portion of the fibula used to replace the anterior mandible. The intermuscular septum proximal and distal to that perforator has been sectioned allowing the skin paddle to rotate 90 degrees on the bone axis illustrated by the arrows. For large mandibular defects, requiring most of the fibula, the vascular pedicle always exits the neomandible posteriorly.

FIGURE 14.5 A. Illustration of a right fibula following ostectomy. Note the preserved periosteum and associated soft tissues with peroneal vessels supplying both the proximal and distal segments of bone. Note the flat lateral aspect of the fibula will be used as the lateral aspect of the neomandible. The arrow illustrates the intention to double barrel the fibula relocating the anterior aspect of each bone segment (shaded) into opposition. Ideally, the anterior aspect of each bone segment would have the periosteum incised longitudinally and reflected medially and laterally to expose the bone to be opposed. B.Illustration of the intended double-barrel procedure. Arrow illustrates the intended opposition of the denuded anterior surfaces of each fibula segment. Note that sufficient bone has been removed to allow repositioning of the segments without undue traction, torsion, or compression of the peroneal vascular pedicle connecting the two. Sufficient bone removal allows this soft tissue envelope containing the peroneal vascular pedicle to be positioned medial to the cut ends of both bone segments at the turn, avoiding potential compression at the interface between the double-barreled fibula and the native mandible.

The Variables

• The bone defect: Is it a lateral, anterior, posterior, or combination of defects? How long is the defect? Will there be sufficient fibula for the double-barrel technique? Is there a defect that involves the condyle?

• The soft tissue defect: How large is the defect? Is it intraoral, facial, or through and through? Does the anterior soft tissue defect extend to both mandibular body segments? Is the oropharynx involved in the defect?

• Vessel availability in the neck: Ipsilateral, contralateral, or bilateral availability? What is the recipient vessel proximity to the neomandible? (Important in the vessel-depleted neck) Preoperative CT neck with contrast can be helpful.

• Cutaneous branch(es) of the flap. How many? Septocutaneous or musculocutaneous? Location (proximal, mid, or distal)? Is there more than one skin paddle available? Doppler or CT angiography of the legs can be helpful.

• Geometry of the flap: The peroneal vascular pedicle can be brought off the neomandible posteriorly, anteriorly, or go to the contralateral neck. This is the variable over which the surgeon has the greatest control and therefore will receive the most attention in preoperative planning. To best visualize this 3D plan, the beginning surgeon should make the templates as illustrated in (Figs. 14.2 and 14.3) (cut them out if helpful).

• Assessing the bone and soft tissue defects and keeping the constants always in mind, the variables can then be systematically assessed to determine the best leg from which to harvest the fibula flap.

Flap Harvest

Minor variations in technique for the fibula flap harvest exist but represent inconsequential personal preferences in my opinion. The common features must include an anterior and posterior dissection, fibula section, and distraction and pedicle vessel dissection. I prefer the leg to be flexed at the hip and knee with a pressure pad protecting the heel. The dissection is done under tourniquet at 350 mm Hg.

The fibula is outlined with indicators for preservation of 6 cm of proximal and distal fibula. The posterior crural septum is marked for resident education, and the location of cutaneous perforators is identified by Doppler (even if location confirmed earlier by imaging and already marked on the skin). The flap is designed to include these cutaneous branches in a size and design to fulfill the needs of the reconstruction. Whether elliptical or rectangular, the long axis of the skin paddle is centered over the posterior crural septum, not over the fibula.

The anterior incision is made through skin and subcutaneous tissue. Even if the skin paddle is designed in the distal leg, the skin incision is extended proximally to the point where the proximal osteotomy cut is to be made (preserving 6 cm of proximal fibula). For wide flaps, the skin paddle is initially elevated from anterior to posterior in a suprafascial plane until over the peroneus longus muscle. This protects the superficial peroneal nerve in the lower leg. At that point, the fascia over the peroneal musculature is incised and the skin paddle is reflected posteriorly in a subfascial plane to demonstrate the anterior aspect of the posterior crural septum. For narrow flaps, the anterior skin incision is already over the peroneal muscles, so it can be extended through the skin, subcutaneous tissue, and fascia in one step. In both wide and narrow flaps, once the posterior crural septum is exposed, the septocutaneous branch(es) are evident. The delicate paratenon over the tendon of peroneus longus muscle should not be disturbed and should not be allowed to dessicate, as both insults will compromise potential skin grafting over the muscle. I like to reinforce the skin markings with the perforators in clear view since later in the case vasoconstriction can make them less obvious. The peroneal muscles are then dissected off the lateral fibula from posterior to anterior. Care must be taken near the septocutaneous perforators at the posterior margin of the bone as these can send branches anteriorly into the peroneal muscles. Control of these branches is best accomplished a few millimeters away from the septocutaneous vessels to avoid vasospasm. The periosteum of the fibula should be left undisturbed. Anterior reflection of the peroneal musculature for the entire length of the bone to be harvested reveals the anterior crural septum. This is incised close to the fibula over the entire length of the bone to be harvested. Hugging the bone, the muscles of the anterior compartment are released from the fibula, exposing the interosseous membrane on its medial aspect. This membrane is similarly incised close to the fibula over the entire length of the bone to be removed. Careful attention must be paid to any anterior retraction of the peroneal muscles in the proximal wound as the peroneal nerve can be damaged. The anterior dissection is now complete. Some surgeons like to cut and distract the fibula at this point, but I prefer to do the posterior dissection first.

The posterior skin incision is made and carried through the subcutaneous tissue and fascia with attention to preserve the sural nerve and saphenous vein intact if possible. The subfascial dissection then reflects the skin paddle anteriorly over the gastrosoleus muscle but not all the way to the posterior crural system. I prefer to now bluntly dissect into the avascular plane between the FHL and the soleus muscles in the distal wound. I then place a finger in this space directed proximally and separate the muscles carefully from distal to proximal looking for septomusculocutaneous perforators that might pierce the soleus muscle. If encountered, I reflect the skin paddle posteriorly and anteriorly, assessing the anterior and posterior aspects of the posterior crural system to get a clear picture of the septocutaneous branches. This allows me to determine if preservation of a small cuff of soleus muscle would provide more cutaneous blood supply. If so, a small cuff of soleus is excised around such perforators and left with the skin paddle. If no such perforators are found, the soleus is separated completely from the FHL and, more proximally, from the posterolateral aspect of the proximal fibula. Also in the proximal lower leg, if I am harvesting a skin paddle based on musculocutaneous perforators, these are dissected and preserved at this time. In either case, separation of the soleus from the proximal fibula exposes the posterior aspect of the proximal fibula. The peroneal vascular pedicle can be now seen as it courses distally disappearing under the most proximal fibular attachment of the FHL. I then incise the fascia on the posterior aspect of the FHL. This maneuver releases the posterior tibial neurovascular pedicle from its close association with the medial aspect of the FHL, reducing risk of inadvertent injury. The posterior dissection is now complete.

I now section the fibula proximally and distally leaving 6 cm of fibula and interosseous membrane undisturbed at each end. It is important, especially at the proximal osteotomy, to protect the peroneal vascular pedicle during osteotomy since it runs just deep to the bone. The fibula bone flap is now distracted laterally. It is being held in the leg by the tibialis posterior muscle on its medial surface and the FHL muscle on its posterior surface and by the peroneal vascular pedicle.

Blunt dissection through the anterior aspect of the tibialis posterior muscle at the distal bone cut reveals the peroneal artery and its venae comitantes. The vascular pedicle is ligated and sectioned in the distal leg. The tibialis posterior muscle is a pennate muscle and is now dissected from distal to proximal. Beginners should follow the middle of the chevron of the muscle to stay well medial to the vascular pedicle running along the posterior medial aspect of the fibula. More experienced surgeons may prefer to expose the vascular pedicle from distal to proximal. In either case, there are muscular branches of the peroneal vessels running to the tibialis posterior and the FHL muscles, which need to be controlled and severed. Once the tibialis posterior muscle has been incised proximally over a few centimeters, the FHL is exposed on its anterior aspect. (It had already been exposed on its posterior surface during the posterior dissection.) The desired amount of FHL to be harvested with the bone is now determined, and the incision of both muscles can be undertaken working from distal to proximal in the lower leg producing ever-increasing lateral distraction of flap. I prefer to use the harmonic scalpel for speed and reliable hemostasis. As the proximal origins of the FHL are approached, the peroneal vascular pedicle can be seen diverging medially from the fibula in the proximal wound. Incision of the deep fascia now exposes the posterior tibial neurovascular pedicle converging with the peroneal vascular pedicle in the proximal lower leg. Incision of the FHL muscle is completed at its proximal origin, but the posterior tibialis muscle needs to be incised all the way to the proximal bone cut. Dissection of the vascular pedicle can now begin. Often one or two muscular branches of the peroneal vessels to the soleus muscle need to be sacrificed to obtain maximal pedicle length. No vascular insufficiency to the soleus has been demonstrated secondary to this maneuver. The peroneal artery can be ligated and sectioned at the bifurcation with the posterior tibial artery, but the veins can be dissected even more proximally in the leg to attain greater length for the venous pedicle. The lateral venae comitantes of the posterior tibial artery can often be divided where it branches over to the medial venae comitantes of the peroneal artery. This produces a very large and long peroneal venous pedicle.

When pedicle vessel dissection is complete, bone wax is applied to the proximal and distal fibula stumps remaining in the leg and the tourniquet is released. Hemostasis is achieved and the flap replaced in the leg, which is wrapped in warm wet sponges and covered with an impermeable Mayo stand cover to maintain warmth.

The flap is not harvested until the ablative surgery and pathologic margin assessment are completed and recipient vessels have been prepared in the neck.

When the flap is harvested and the ischemia clock started, the pedicle vessels are sectioned and the leg wound is closed over a suction drain. If a skin graft is applied, a vacudrain or bolster dressing is placed and the lower leg placed in a posterior splint.

Flap Inset

Optimal flap orientation to facilitate inset (covered above) is a preoperative not an intraoperative concept. In fact, this concept would have been the primary preoperative consideration when choosing from which leg to harvest the fibula flap. One side is always superior, but either side can be made to work.

When the composite free fibula flap is delivered to the neck, the first thing that I do is to place the flap in its intended orientation over the mandibular defect. This allows me to measure the length of vascular pedicle required to comfortably reach the recipient vessels. This immediately dictates how much proximal fibula, if any, will need to be removed to produce optimal pedicle vessel geometry, and a quick check of the remaining fibula will demonstrate how much bone would remain to fill the rest of the mandibular defect. This is one area where preoperative 3D CT fibula ostectomy planning can be less reliable. This is especially true in those cases where much of the fibula is required for the mandibular defect and the recipient vessels are in question. It is never a problem when only distal fibula is required for the smaller defect and the neck vessels have not been disturbed by previous surgery.

The second important calculation to be made is the location of the cutaneous branch(es). The association of these branches with the bone is critical in choosing where to remove bone. Although careful removal of the periosteum from distal portions of the fibula does preserve the integrity of these cutaneous branches, they can be compromised by kinking, torsion, or traction. For this reason, I try to avoid sacrificing the bone immediately adjacent to septocutaneous perforators whenever possible. This bone–perforator relationship becomes more important in complex single skin paddle through-and-through soft tissue defects with large mandibular defects. It is also an important consideration if more than one skin paddle is employed irrespective if the size of the mandibular defect.

If proximal fibula needs to be sacrificed, I incise the periosteum along the lateral border of the proximal fibula and, using a large periosteal elevator, I elevate the periosteum off the entire proximal fibula to be removed.

Much is written about template construction for accurate fibula contouring. I employ a very inexpensive yet accurate method using 4-mm-thick Silastic sheeting. For anterior mandibular reconstruction, I make a 1-cm wide, 2D template of the basilar and of the gingival aspects of the mandible segment removed from the patient. If a single-barrel fibula replacement is planned, I use the gingival template to ensure a proper occlusal plane. If a double-barrel technique is planned, I use both templates. These templates have no curves but rather are angular to duplicate the defect with multiple straight segments comparable to the final contoured fibula. When I determine which portion of the fibula to contour, I measure the distance of the lateral aspect of each segment of the template. These measurements are accurately transferred to the lateral aspect of the fibula (Fig. 14.6). The angle required to duplicate the template is easily derived by cutting the 1-cm, 2D template from medial to lateral at each angle leaving only 1 mm of Silastic on the outer aspect of the template. The cut template is then opened and laid upon the anterior aspect of the fibula, where the triangle created by opening the template is marked on the bone with the apex on the lateral aspect of the fibula. With periosteal elevation and protection, a reciprocating saw is used to excise this wedge of fibula (Fig. 14.7) and the ostectomy “closed” on the medial aspect. Two miniplates are bent and secured with self-tapping screws to fix the new contour of the fibula before moving on to the next segment (Fig. 14.8). The same principle applies when reconstructing the angle and ascending ramus except that the Silastic template is cut from superior to inferior at the template angle (Fig. 14.9). When opened, this template is applied to the lateral aspect of the fibula and the apex of the wedge of bone to be removed is on the inferior aspect of the fibula (neomandible). That ostectomy is therefore “closed” along the superior aspect of the fibula (neomandible) (Fig. 14.10).

FIGURE 14.6 A. Illustration of a mandible from above with Silastic template (4-mm thick, 12-mm wide) overlaid on the gingival aspect of the anterior portion of the mandible that has been removed. Note that the central segment of the neomandible lies over the tooth roots, not the more anterior mentum. In this way, osseointegrated implants will be appropriately loaded in the occlusal plane. If the double-barreled technique is employed, a second template of the basilar aspect of the bone is made. Any double-barreled segments will therefore reside slightly more lateral and anterior in the neomandible. B. The Silastic template of the anterior mandible defect is cut from internal to external at the angles, leaving 1 mm of Silastic attached. This will allow outward rotation of the lateral segments illustrated by the arrows. C. When the lateral segments of the template are lined up with the central segment, the angles created accurately represent the triangular segments of fibula that need to be removed.

FIGURE 14.7 The straightened template is laid on the anterior surface of the fibula. The periosteum over the intended ostectomies is incised and reflected. A small malleable ribbon retractor is insinuated around the fibula to protect the peroneal vessels and the ostectomies marked and removed with an oscillating saw. The fibula segments are then closed at each ostectomy site and secured with two miniplates using monocortical screw at each ostectomy site.

FIGURE 14.8 The recontoured neomandible is illustrated fixed to the native mandible with mandibular plates and screws. Due to the paucity of forces on the neomandible and the thickness of fibula cortex, monocortical screws are sufficient for fixation. If intraoperative placement of osseointegrated implants is intended, this should be carried out after the neomandible is firmly fixed in its new location. This will allow maximized orientation of the implants in the occlusal plane, thereby minimizing potential cantilever forces on the implants with attendant failure rates.

FIGURE 14.9 A. Illustration of a right lateral mandible defect of body and ramus with the Silastic template replicating the intended neomandible. Note the importance of reconstructing the angle of the mandible. The anterior aspect of the reconstruction need not be at the basilar aspect of the mandible. B.The template illustrating the intended cut from superior to inferior at the angle preserving 1 mm of Silastic. C. The template opened to illustrate the exact dimensions of the intended ostectomy of the fibula required to duplicate the neomandible. This time, the template is opened, straightened, and laid on the lateral surface of the fibula, such that the apex of the triangular ostectomy is at the posterior border of the fibula. As in Figure 14.7, periosteum at the intended ostectomy site is incised and reflected. A small malleable ribbon retractor protects the peroneal vessels during the ostectomy with an oscillating saw. The fibula segments are closed at the ostectomy sites and plated.

FIGURE 14.10 Illustrates neomandible reconstruction of a right mandible defect with fibula bone.

I have employed much more sophisticated (and expensive) techniques including preoperative modeling to include bone-cutting jigs. This can save time but at some considerable expense (often not covered by insurance), and the final result, in my hands, is no different.

For simple straight line defects of the body of the mandible in the dentulous patient, no template is required if a reconstructive bone plate has been preapplied. Simple measurement of the bone defect (both gingival and basilar if a double barrel is to be used) will suffice. Always use a saw to fine-tune the cut ends to maximize bone-to-bone contact.

When multiple ostectomies are needed for the mandible reconstruction, I measure, cut and contour the fibula from proximal to distal. I plate and screw fixate each contour of the neomandible before cutting the next. The reason for this is to insure maximal bone contact for rapid, reliable bone healing.

This brings up the perennial discussion regarding miniplate versus reconstructive plate fixation. My prejudice for over 20 years has been for dual miniplate fixation at each fibula-fibula ostectomy site with heavier reconstructive plates used to secure the fibula to the native mandible. Increasing literature seems to support this notion that the stress shielding provided by one-piece reconstructive plates over vascularized bone may result in more fibula atrophy over time. For large edentulous mandibular defects (e.g., angle-to-angle), I match the fibula neomandible to the upper arch with a template and direct approximation. The anterior ostectomy sites will be fixed with miniplates and the fibula attached to the mandibular ramus remnants with separate reconstructive plates. For smaller defects in the dentulous mandible, I contour a single reconstructive plate from the mandible prior to resection or from a preoperative 3D mold. This is used to maintain occlusion and, for lateral mandibular defects, will be left in place. For defects that include the anterior mandible, I use it to maintain occlusion while I contour the neomandible. After the neomandible is fixed in place, I cut out the symphyseal aspect of the reconstructive plate leaving its proximal and distal portions to fix fibula to mandibular remnants. This is done for two reasons. First, the single reconstructive plate contoured to the mental protuberance fails to represent the occlusal plane of the neomandible but rather is too far forward. If the fibula was contoured to this plate, the patient would experience a pronounced underbite following dental rehabilitation. Second, if the fibula neomandible is contoured in its proper occlusal alignment, a significant space exists between anterior neomandible and the reconstructive plate. This can cause significant pressure on the deep aspect of the mental soft tissues with no apparent benefit. For these reasons, I remove the anterior aspect of the reconstructive plate after the neomandible is in place. An alternative for those dentulous patients with only the anterior mandible sacrificed is maxillomandibular fixation to maintain occlusion. The neomandible is then contoured to match the occlusal plane, fibula–fibula segments secured with miniplates, and larger plates used to secure fibula–mandible segments.

Attention to this anterior–posterior differential in location of the mandibular gingival versus the mental protuberance is important for optimal functional dental rehabilitation of the neomandible. Also important to remember is the medial–lateral differential associated with the mandibular body where the occlusal surface (gingival) is medial to the basilar aspect of the mandible. This is an equally important consideration often overlooked when surgeons attach a fibular segment to a reconstructive plate oriented along the inferior aspect of the mandibular body. Osseointegrated implants placed vertically in a fibula segment at this location will fail to duplicate the occlusal relationships with the maxillary dentition seen in the original mandible. For optimal dental rehabilitation, the fibula segments, both anterior and lateral, need to respect the occlusal plane. Anteriorly, this is accomplished by establishing the neomandible posterior to the mental prominence. For isolated anterior mandible defects, I have used double-barrel fibula to duplicate the vertical height of the mandible and even triple-barrel fibula to restore both mandibular height and mental protrusion.

For lateral (body) defects, I prefer the double-barrel technique but remember the superior “barrel” of the fibula (replacing the gingiva) must be located slightly medial to the inferior “barrel,” which will be replacing the basilar bone.

When very large mandibular defects are reconstructed, there is inadequate fibula length for double-barrel techniques. Other techniques have worked well. For example, placement of the fibula in a more superior position within the mandible (not along the inferior border) allows for more successful dental rehabilitation. At the same time, in these patients, a large deepithelialized skin paddle placed along the inferior mandibular aspect (with or without a bone plate for support) provides facial contouring approximating the original mandible. The height of the fibula can also be augmented secondarily with bone grafts or by osseodistraction. Of course, for the longtime edentulous patient with an atrophic mandible and no interest in or plans for dental rehabilitation, all these considerations become theoretical.

As mentioned above, when the double-barrel technique is employed, it is important to orient the superior (gingival) barrel and the intraoral skin paddle first. Any compromise of this portion of the reconstruction could lead to salivary leakage and even flap loss. The inferior (basilar) barrel should be positioned after the more important superior barrel and intraoral skin are secure and uncompromised, since the role of the inferior barrel is more aesthetic than functional. For this reason, I always choose the portion of the fibula to be used for the gingival barrel that has the best association with a cutaneous perforator. It is important to remember that successful bone union between fibula and mandible or fibular segments can be compromised by inadvertent inclusion of any soft tissues between the bony surfaces.

Common surgical sense must prevail for determining the order in which to inset skin paddle(s) and fibula bone into the oromandibular defect. In any case, I like to have the bone fixated before starting the microvascular surgery. For wide open defects with excellent exposure, that means that I revascularize the flap before insetting the skin paddle. This allows for easy evaluation and control of any previously controlled subcutaneous bleeding sites before the wound is closed. There are few things more frustrating than encountering an inaccessible bleeder after bone fixation. This is another reason for establishing excellent hemostasis on the flap while it is still vascularized in the leg.

For many through-and-through defects, intraoral exposure may be severely limited after bone fixation; therefore, the intraoral soft tissue defect needs to be reconstructed first, the bone fixated second, and the flap revascularized third before external defects are closed. This still allows for more limited assessment and hemostasis after the flap is revascularized and before the face and neck are closed.

The leg is kept in a posterior splint for 5 days during which the patient is ambulated with assistance and without weight bearing. The splint is removed after 5 days and the patient encouraged to ambulate with slowly graduated weight bearing.

If a skin graft has been placed, the bolster dressing or vacudrain is removed after 8 days.

Suction drain of the lower leg is removed when drainage meets criteria.

A nasogastric tube is placed and the patient kept NPO for at least 10 days if no radiation has preceded the surgery and for 14 days if radiation has occurred.

Peridex mouth rinses.

Hourly flap check for the first 72 hours.

• Pressure trauma to the heel. Avoid by placing the patient’s foot in a foam pad and limiting the time of leg elevation.

• Instability of the knee or ankle may occur if insufficient proximal or distal fibula (interosseous membrane) is preserved.

• Nerve damage. Exercise prudent avoidance of the posterior tibial neurovascular pedicle and be careful with placement of the retractors on the proximal peroneal muscles.

• Hematoma: Avoid by exercising judicious hemostasis after flap harvest and tourniquet release and by applying bone wax to the stumps of the fibula.

• Poor skin graft take. Can be minimized by careful preservation of the paratenon over the peroneal tendon to include avoidance of desiccation of this paratenon during harvest.

• Vascular compromise of bone segments. Can be avoided by minimizing periosteal elevation and by maintaining a 2-cm minimal length on bone segments.

• Vascular compromise of the skin paddle through excessive torsion, traction, or compression of the cutaneous perforator(s) during flap inset can be avoided with proper preplanning and careful flap manipulation during inset.

In 378 consecutive fibula osseocutaneous free flaps which I have performed over 22 years, 3 flaps were lost to pedicle vessel compromise for a success rate of 99.2% for the osseous flap. In addition, six patients, who had well-vascularized bone flaps, suffered skin flap necrosis (two total and four partial but sufficient to require additional surgery). Each case could be attributed to excessive torsion or traction of the cutaneous perforators. This leaves a success rate of 97.6% for the cutaneous flap.

Three cases in which the fibula was used to reconstruct osseoradionecrosis defects suffered delayed recipient artery rupture 7 to 10 days postoperatively. In no case was the fibula bone or skin paddle lost. Each case was attributed to the poor vascularity of the wall of the recipient artery from the excessive radiation resulting in a delayed pseudoaneurysm and rupture.

Nonunion at the mandibular reconstruction site occurred in five cases and was corrected surgically by open reduction, freshening of the bone edges, and internal fixation. Each case was felt to be related to unintentional inadvertent inclusion of soft tissue between bone surfaces at the time of reconstruction. One patient suffered what appeared to be avascular necrosis of a 1.5-cm central segment of fibula in a multiosteotomized, 3D reconstruction of a very large oromandibular defect including ramus, body, symphysis, and contralateral mandibular body. All of the fibula, proximal and distal to this segment, and the overlying skin paddle remained viable over 9 years, and no further surgery was required. It was felt that the bone contouring had been overly aggressive, leaving inadequate periosteal blood supply to this small fibula segment.

No patient suffered significant functional loss in the leg or foot. All patients experienced some swelling in the operated lower leg lasting for weeks to months.

One patient required surgical evacuation of a significant hematoma in the lower leg.

For those patients who required skin grafts to the donor site, approximately 10% experienced delayed wound healing of at least some portion of the grafted site most frequently over the peroneus longus tendon. All of these wounds ultimately epithelialized with conservative management.

• Preoperative evaluation of the vascularity of the lower leg including identification of the cutaneous branches is important. Use CT angiography but do not forget the venous system. If there is a history of deep vein thrombosis, may need venogram as well.

• Preoperative evaluation of vascularity in the neck is especially important in the vessel-depleted neck.

• Using the defect requirements, the recipient vessel availability, and the constants associated with each potential fibula flap, determine preoperatively the leg that provides the superior fibula flap.

• Maximize bone-to-bone contact.

• Maximize pedicle vessel geometry.

• Optimize the relationships of the skin paddle(s) and their perforators with the bone segment(s) to be used.

• Remember to align the neomandible with the upper arch. The fibula segment does not have to be placed at the inferior rim of the mandible. Double-barrel techniques are available.

• Smooth the angles created in the anterior neomandible.

• Secure and maintain the chin pad in natural position on the neosymphysis after body-to-body mandibular reconstruction or the lip–chin complex may slip inferiorly during healing. Be wary of pressure dressings over the neosymphysis to avoid pressure necrosis of the chin pad.

• Do not tolerate postoperative seromas or hematomas in the vicinity of the vessel anastomoses.

• Head and neck surgery tray

• Reciprocating saw

SUGGESTED READING

Virgin FW, Iseli TA, Iseli CE, et al. Functional outcomes of fibula and osteocutaneous forearm free flap reconstruction for segmental mandibular defects. Laryngoscope 2010;120(4):663–667.

Baumann DP, Yu P, Hanasono MM, et al. Free flap reconstruction of osteoradionecrosis of the mandible: a 10-year review and defect classification. Head Neck 2011;33(6):800–807.

Cannon TY, Strub GM, Yawn RJ, et al. Oromandibular reconstruction. Clin Anat 2012;25(1):108–119. doi: 10.1002/ca.22019.