Click here to watch a video of double free-flap reconstruction

The use of microvascular free flaps for the reconstruction of defects following the resection of head and neck cancer is a complex but routine procedure. However, single flaps may not be sufficient for some defects that are either too large or warrant composite tissues. In particular, resection of advanced tumours of the oral cavity results in complex oromandibular defects that often involve bone, oral lining, external skin, and soft tissue. The free fibular osteocutaneous (FO) flap is well established as a workhorse flap for mandible reconstruction,1 which provides 25 to 30 cm of straight bone of good quality that can be contoured, as well as a skin paddle for soft tissue coverage when needed. The pedicle has an acceptable length and its vessels have a good diameter. It is therefore our preferred option for restoring mandibular defects and for lining the oral cavity.

However, the size of the skin paddle is limited1 and may not be supplied by the same vessel as the bone.2 Thus, with larger composite defects, a single fibula flap cannot provide sufficient soft tissue coverage and a second skin flap may be necessary. Some surgeons nevertheless elect to avoid a second free flap by choosing either a pedicled flap or alloplastic material. We therefore set out to demonstrate the feasibility of resorting to double free flap surgery in head and neck reconstruction.

Our choice for additional soft tissue is the anterolateral thigh (ALT) flap that provides up to 630 cm² of skin.3 On occasions when the vascularity of the fibula flap skin paddle is deemed borderline, the ALT can be harvested with multiple skin islands so as to cover both the inner lining and the external skin. Harvest of the FO and ALT flaps can proceed at the same time as tumour excision, without the need for patient re-positioning, which is an important logistical advantage. Like most surgeons, whenever possible we prefer using separate anastomoses for double flaps rather than sequential linking or ‘flow through’,4 5 6 as some studies5 6 suggest that the latter has more complications (possibly due to increased thrombogenicity or a ‘steal’ phenomena).

We conducted a retrospective case review of patients in our institution with head and neck cancer who had reconstruction with planned simultaneous double free flaps over a 2.5-year period (from November 2010 to August 2013). For all cases we deployed two surgical teams; reconstructions were performed (one surgeon) at the same time as tumour excision (other surgeons). Preoperatively, handheld Doppler probes were used to locate the skin perforators for both flaps. The peroneal artery was sacrificed in the harvest of fibula flaps and adequacy of the remaining vessels was screened by palpation of the dorsalis pedis and posterior tibial pulses. An angiogram was used in only one patient with a history of peripheral vascular disease.

The FO flap was harvested first using a lateral approach; a sterile tourniquet was placed on the upper thigh but not inflated. A skin island was harvested in nine out of 10 fibula flaps. In one patient, the skin island was not perfused by the peroneal artery and thus not harvested. In another, the vascularity of the skin island was deemed suboptimal and therefore not used. The fibula flap was kept in situ after isolation of its vascular pedicle while the ALT was harvested. Intramuscular perforators to the thigh skin island were skeletonised in all cases so as to completely visualise the vessels. Once the surgical margins were deemed clear by frozen sections, the final dimensions of the ALT flaps were determined when the final defect was defined.

Whenever possible, intermaxillary fixation was used to hold the mandible and maxilla in an optimal position, and ‘by eye’ the fibula was osteotomised to fit (average 1-2 osteotomies). Two sets of mini-plates were used per osteotomy site so as to maximise rotational stability. The use of 2.5 x or 3.5 x loupes by the reconstructive surgeon allowed micro-anastomoses of the vessels, whilst insetting of the flap was completed.

A 58-year-old man was referred to our centre with a second recurrence of a squamous cell carcinoma of his tongue. Three years earlier, he had had a partial right glossectomy with a selective neck dissection for a pT2N0 lesion. One year later he underwent a complete neck dissection for a right nodal recurrence, and another year later he had had a reconstruction with a pectoralis major myocutaneous flap (PMMF) after total glossectomy for local tumour recurrence. After the tumour was resected, he had a bony defect from one angle of the mandible to the other, and a soft tissue defect that involved the entire inferior oral cavity down to the chin and anterior neck skin, which left a 3-cm rim of lower lip (Fig 1).

We used a fibula flap with its overlying skin island along with a large ALT flap (Fig 2). After anastomosis of the two sets of vessels, bleeding from the edge of the fibula flap skin island appeared rather sluggish. So the ALT was used for both intraoral lining and external skin cover. A strip of the ALT flap was de-epithelialised for suturing to the lower lip remnant (Fig 3). There were no major complications and the patient was discharged on the 14th postoperative day. There was a good contour at follow-up (Fig 4); the patient used a percutaneous endoscopic gastrostomy (PEG) for feeding preoperatively but regrettably could not resume oral feeding after this surgery and therefore remained reliant on the PEG.

All tumours were stage T4a, with nodal status ranging from N0-N3 (Table). During the study period, there were six male and six female patients who had double free flap surgery. Their ages ranged from 31 to 88 (mean, 55) years. In 10 of them, a free fibula flap was combined with an ALT flap harvested from the same limb; in eight of them a skin island was harvested with the bone. One patient had bilateral ALT flaps for reconstruction of an extensive tumour of the tongue and floor of the mouth without bone involvement. Another patient had a free fibula flap combined with an anteromedial thigh flap, due to absence of suitable perforators upon dissecting the ALT flap.

The mean total operating time was 660 minutes, which included the time for frozen section results. Postoperative hospital stays ranged from 11 to 82 days; nine patients were discharged home within 3 weeks. Patient 10 stayed 80 days. She declined further surgery for an intraoral dehiscence, which was therefore treated conservatively. Patient 7 stayed 82 days, as his recovery was complicated by a carotid blowout on the 11th postoperative day for which he had a surgery; subsequently a pseudomonas wound infection was treated with antibiotics. After surgery, seven patients were able to resume oral feeding sufficient to maintain their body weight; the remainder relied on tube feeding. Five patients received adjuvant treatment (4 had chemoradiation and 1 only had radiotherapy).

Minor postoperative complications (fluid collections, fistulae) occurred in 67% of these patients and usually resolved with conservative management. More serious complications occurred in 33% of the patients (carotid blowout, wound dehiscence/infection, and fluid collections treated surgically). In one patient, a haematoma was treated by debridement of the soft tissue portion of the free fibula flap that had been de-epithelialised and ‘buried’. There were no instances of total flap loss; two patients were taken back to theatre for exploration and their flaps were salvaged. One of them (patient 10) had venous congestion of the fibula skin flap (used for intraoral lining), which was salvaged but remained swollen and indurated. In view of a concomitant intraoral wound dehiscence, the swollen skin island was debrided and a pedicled ipsilateral pectoralis major flap was harvested to close the intraoral wound. Regrettably, although the pedicled flap survived, the intraoral wound dehisced again, and the patient declined to have further surgery so her wound was managed with daily dressings (see above).

Two (17%) out of the 12 patients had tumour recurrence during the follow-up period, and a further two (17%) had distant metastases. Survival from the time of surgery ranged from 60 to 303 days. The patient survival rate at 6 months was 91%, and at 1 year was 64%. At the time of writing this paper, only seven of the 12 patients had been followed up for at least 2 years, three (43%) of whom were still alive.

Following resection of advanced oral cancers, it is our standard practice to use double free flaps when needed for reconstruction of complex oromandibular defects, particularly those involving large defects of both bone and soft tissue. In most cases, the indication for double free flaps was the requirement for bone and soft tissue/skin not provided by the skin island of a FO flap. This practice is by no means universal; some surgeons are reluctant to contemplate a second free flap due to the perceived increase in technical complexity, operating time, and risk of complications. Alternative strategies include substitution of the fibular flap with a metal reconstruction plate, combined with a soft tissue flap for resurfacing7; combining a fibular free flap with pedicled regional flaps, such as the deltopectoral flap, PMMF,8 or latissimus dorsi myocutaneous flap. Some centres regard such cases as ‘inoperable’ and offer palliative treatment only.

However, these simpler alternatives have their drawbacks. The problems associated with an alloplastic plate with a soft tissue flap for composite mandible reconstruction are well documented,9 10 11 there being high rates of delayed plate exposure and recourse to salvage procedures.12 In the long term, use of vascularised bone (particularly in the FO flap) is more successful for mandible reconstruction,2 and was our first choice in all cases, with the possible exception of patients with a short life expectancy (<6 months). Recourse to a regional pedicled soft tissue flap instead of a free flap is based on its perceived advantage in being technically easier to harvest and involving shorter operating times.9 13 There is also a perceived lower risk of complications through avoiding a second set of microanastomoses. The PMMF is the most commonly used regional flap,14 but the vascularity of its skin paddle (like that of other regional flaps used in head and neck reconstruction) tends to be suboptimal; if the muscle is too short, more of the skin paddle results in a ‘random-pattern’. Crucially, the skin islands tend to be positioned at the most distal portions and thus have the poorest vascularity in the most critical parts.15 Chen et al16 recommends avoiding PMMFs to line the oral cavity due to a high rate of bone exposure from dehiscence.

On the contrary, surgeons such as Bianchi et al17 have actually demonstrated better outcomes with double free flaps compared to a combination of one free flap with one pedicled flap. The bulk of the muscle pedicle in regional flaps can interfere with the inset and vascularity of a concomitant free flap,13 and the tendency for muscle atrophy and gravitational effects can adversely affect the final results of reconstruction. Chen et al16 demonstrated a lower failure rate with two free flaps (2.8%) compared with the combination of one free and one pedicled flap (9%). They speculated that the bulky PMMF pedicle may actually compress the free flap pedicle, citing the 14% to 33% frequency of internal jugular vein thrombosis after radical neck dissection covered with pedicled flaps.18 19 The skin island of a regional flap also tends to be thicker, less pliable, and thus may interfere with intraoral function. Regional flaps may be limited in other ways (eg lack of necessary tissue components or specific tissue volume), which compromise the final aesthetic and functional outcomes.20

Although on average, a single free flap can take 1.5 hours longer than a PMMF to harvest, Tsue et al21 found that the operating time for double flaps can be 3 hours shorter than for a one free and one pedicled combination. They explained this by citing possible bias by surgeons choosing to use a second pedicled flap, when the resection time was longer, and surgeons working faster whenever two free flaps were anticipated. Guillemaud et al22 found no significant difference in the duration of surgery and complication rate when comparing double free and one free and one pedicled surgeries. In the end, the duration of surgery should not be a factor in determining the type of reconstruction.23

Proposed indications for the use of double free flaps are listed in the Box.20 The reconstruction of defects resulting from tumour resection in the head and neck region is a challenge, particularly when a composite of tissues is required or the defect is too large to cover by a single flap. Recourse to two free flaps allows more versatility and flexibility when reconstructing such complex defects. The best osseous and soft tissue elements may be independently selected, yielding appropriate tissue characteristics for ideal defect reconstruction. Using two separate thin pliable free flaps rather than bulky pedicled flaps may allow easier insetting and better restoration of the 3-dimensional anatomical boundaries,24 and thus both the functional and aesthetic outcomes can be addressed. With free flaps, there is also the potential for including other components such as nerves for sensate flaps.24

Good-quality soft tissue coverage is needed to reduce the risk of plate exposure12; even when the skin component of the FO flap can provide adequate surface cover, there is usually an overall shortage of soft tissue. Soft tissue reconstruction is as important as bone reconstruction25 in determining a satisfactory outcome, as deficiency of the latter tissues is poorly tolerated in the head and neck,26 and may lead to inadequate obliteration of dead spaces (eg from resection of masticators, buccal fat pad, and parotid). This causes accumulation of fluid which may become secondarily infected,16 and threaten micro-anastomoses and lead to contractures, and poor cosmetic outcomes or functionality that can lead to trismus, as well as contraction of the floor of the mouth with tethering of the tongue with difficulties in swallowing and speech.27 Therefore, even in the absence of bone loss, a double free flap reconstruction can be advantageous especially if soft tissue loss is substantial or beyond the reach of pedicled alternatives.

The use of two simultaneous free flaps undoubtedly poses technical difficulties, by increasing potential patient morbidity and is time-consuming. Although it is not our intention to promote double free flap reconstruction as a ‘routine’ reconstruction procedure, we wish to highlight it as an option, at least for tumours that are often deemed ‘inoperable’. Balasubramanian et al28 demonstrated that advanced ‘inoperable’ tumours such as T4b (in 7 of 21 cases) can be safely operated on; having double free flap reconstruction in the armamentarium allows surgeons to be more aggressive with extirpation. With careful patient selection, the duration of surgery, hospital stays, and complications need not be prohibitive compared to single free flap operations.25 Wei et al20 suggest that double free flaps should be restricted to patients with primary cancers, avoiding their use in those with recurrent cancers or second primaries. Nevertheless, in our series three patients presented with recurrent cancer. Individual patients should be assessed on a case-by-case basis—a PMMF could be considered to cover the skin of the neck, whilst reconstruction plates may be used to reconstruct short posterior or lateral mandible defects, particularly in those with a short life expectancy.

Our study shows that double free flap reconstruction can be worthwhile in patients with T4 tumours with a flap survival rate of 100% and a patient survival rate of 64% at the time of going to press. Just over half of our patients were able to resume oral feeding, which is somewhat lower than that in some other studies,28 29 and may be related to the locally advanced extent of their tumours, particularly with regard to tongue involvement.

1. Hidalgo DA, Rekow A. A review of 60 consecutive fibula free flap mandible reconstructions. Plast Reconstr Surg 1995;96:585-96. CrossRef

2. Tan BK, Wong CH. An anomalous septocutaneous perforator to the skin paddle of the fibula osteocutaneous flap originating from the posterior tibial artery. J Plast Reconstr Aesthet Surg 2009;62:690-2. CrossRef

3. Chiu T, Wong EW, Burd A, Vlantis A. Perforator transfer in the antero-lateral thigh flap. J Plast Reconstr Aesthet Surg 2013;66:1012-3. CrossRef

4. Lin PY, Kuo YR, Chien CY, Jeng SF. Reconstruction of head and neck cancer with double flaps: comparison of single and double recipient vessels. J Reconstr Microsurg 2009;25:191-5. CrossRef

5. Wei FC, Demirkan F, Chen HC, Chen IH. Double free flaps in reconstruction of extensive composite mandibular defects in head and neck cancer. Plast Reconstr Surg 1999;103:39-47. CrossRef

6. Wei FC, Celik N, Chen HC, Cheng MH, Huang WC. Combined anterolateral thigh flap and vascularized fibula osteoseptocutaneous flap in reconstruction of extensive composite mandibular defects. Plast Reconstr Surg 2002;109:45-52. CrossRef

7. Boyd JB, Mulholland RS, Davidson J, et al. The free flap and plate in oromandibular reconstruction: long-term review and indications. Plast Reconstr Surg 1995;95:1018-28. CrossRef

8. Ariyan S. The pectoralis major myocutaneous flap. A versatile flap for reconstruction in the head and neck. Plast Reconstr Surg 1979;63:73-81. CrossRef

9. Blackwell KE, Buchbinder D, Urken ML. Lateral mandibular reconstruction using soft-tissue free flap and plates. Arch Otolaryngol Head Neck Surg 1996;122:672-8. CrossRef

10. Cohen M, Schultz RC. Mandibular reconstruction. Clin Plast Surg 1985;12:411-22.

11. Shpitzer T, Gullane PJ, Neligan PC, et al. The free vascularized flap and the flap plate option: comparative results of reconstruction of lateral mandibular defects. Laryngoscope 2000;110:2056-60. CrossRef

12. Wei FC, Celik N, Yang WG, Chen IH, Chang YM, Chen HC. Complications after reconstruction plate and soft-tissue free flap in composite mandibular defects and secondary salvage reconstruction with osteocutaneous flap. Plast Reconstr Surg 2003;112:37-42. CrossRef

13. Blackwell KE, Buchbinder D, Biller HF, Urken ML. Reconstruction of massive defects in the head and neck: the role of simultaneous distant and regional flaps. Head Neck 1997;19:620-8. CrossRef

14. Lerrick AJ, Zak MJ. Oral cavity reconstruction with simultaneous free and pedicled composite flaps. Operat Tech Otolaryngol Head Neck Surg 2000;11:76-89. CrossRef

15. Shah JP, Haribhakti V, Loree TR, Sutaria P. Complications of the pectoralis major myocutaneous flap in head and neck reconstruction. Am J Surg 1990;160:352-5. CrossRef

16. Chen HC, Demirkan F, Wei FC, Cheng SL, Cheng MH, Chen IH. Free fibula osteoseptocutaneous-pedicled pectoralis major myocutaneous flap combination in reconstruction of extensive composite mandibular defects. Plast Reconstr Surg 1999;103:835-45. CrossRef

17. Bianchi B, Ferri A, Ferrari S, et al. Reconstruction of lateral through and through oro-mandibular defects following oncological resections. Microsurgery 2010;30:517-25. CrossRef

18. Fisher CB, Mattox DE, Zinreich JS. Patency of the internal jugular vein after functional neck dissection. Laryngoscope 1988;98:923-7. CrossRef

19. Brown DH, Mulholland S, Yoo JH, et al. Internal jugular vein thrombosis following modified neck dissection: implications for head and neck flap reconstruction. Head Neck 1998;20:169-74. CrossRef

20. Wei FC, Yazar S, Lin CH, Cheng MH, Tsao CK, Chiang YC. Double free flaps in head and neck reconstruction. Clin Plastic Surg 2005;32:303-8. CrossRef

21. Tsue TT, Desyatnikova SS, Deleyiannis FW, et al. Comparison of cost and function in reconstruction of the posterior oral cavity and oropharynx. Free vs pedicled soft tissue transfer. Arch Otolaryngol Head Neck Surg 1997;123:731-7. CrossRef

22. Guillemaud JP, Seikaly H, Cote DW, et al. Double free-flap reconstruction: indications, challenges, and prospective functional outcomes. Arch Otolaryngol Head Neck Surg 2009;135:406-10. CrossRef

23. Schusterman MA, Horndeski G. Analysis of the morbidity associated with immediate microvascular reconstruction in head and neck cancer patients. Head Neck 1991;13:51-5. CrossRef

24. Urken ML, Weinberg H, Vickery C, et al. The combined sensate radial forearm and iliac crest free flaps for reconstruction of significant glossectomy-mandibulectomy defects. Laryngoscope 1992;102:543-8. CrossRef

25. Urken ML, Weinberg H, Vickery C, Buchbinder D, Lawson W, Biller HF. Oromandibular reconstruction using microvascular composite free flaps. Reports of 71 cases and a new classification scheme for bony, soft-tissue, and neurologic defects. Arch Otolaryngol Head Neck Surg 1991;117:733-44. CrossRef

26. Andrades P, Bohannon IA, Baranano CF, Wax MK, Rosenthal E. Indications and outcomes of double free flaps in head and neck reconstruction. Microsurgery 2009;29:171-7. CrossRef

27. Urken ML, Buchbinder D, Weinberg H, et al. Functional evaluation following microvascular oromandibular reconstruction of the oral cancer patient: a comparative study of reconstructed and non-reconstructed patients. Laryngoscope 1991;101:935-50. CrossRef

28. Balasubramanian D, Thankappan K, Kuriakose MA, et al. Reconstructive indications of simultaneous double free flaps in the head and neck: a case series and literature review. Microsurgery 2012;32:423-30. CrossRef

29. Hanasono MM, Weinstock YE, Yu P. Reconstruction of extensive head and neck defects with multiple simultaneous free flaps. Plast Reconstr Surg 2008;122:1739-46. CrossRef