Blackwell KE. Unsurpassed Reliability of Free Flaps for Head and Neck Reconstruction. Arch Otolaryngol Head Neck Surg. 1999;125(3):295–299. doi:10.1001/archotol.125.3.295
To review the outcome and incidence of perioperative complications in patients undergoing microvascular free flaps for reconstruction of the head and neck region.
A prospective case series.
An academic tertiary care otolaryngology–head and neck surgery program.
One hundred fifteen patients who underwent 119 consecutive free flaps performed by 1 surgeon during a 32-month period.
Reconstruction primarily by means of radial forearm, fibula, and rectus abdominis flaps (95% of the flaps selected for reconstruction).
Main Outcome Measure
The incidence of perioperative reconstructive and medical complications.
There was 1 perioperative death (0.8%). Among the surviving patients, there was 1 case of complete flap failure, resulting in an overall flap survival of 99.2%. There were 2 additional cases (1.8%) of partial flap necrosis. Perioperative reconstructive complications occurred during 10.1% of the hospitalizations, half of which required additional surgical intervention. Notable perioperative medical complications occurred in 17.1% of the patients.
Despite their reliance on small-vessel anastomoses for survival, free flaps are extremely reliable with regard to the incidence of flap necrosis, which contributes to a low incidence of perioperative complications. Selection of flaps that have proven dependability contributes to a successful outcome. While this technique frequently requires lengthy surgery in an elderly patient population, the perioperative mortality and morbidity are acceptable. Because of their unsurpassed reliability, free flaps have become the preferred method of reconstruction for most patients with major defects in the head and neck region.
DURING the decade after its description by Ariyan1 and Baek et al2 in 1979, the pectoralis major myocutaneous flap was widely considered to be the standard flap for reconstruction of major defects in the head and neck regions. Although the first microscope-assisted transfer of a free flap was reported in 1973,3 until recently there has been limited enthusiasm in the United States for applying free flaps for reconstruction of head and neck defects.4 This reluctance arose from several perceived potential shortcomings of microvascular tissue transfer. Such concerns included questions regarding the reliability of a technique that was dependent on small-vessel vascular anastomoses for a successful outcome and the potential for an adverse impact on the costs and complications of therapy. The current study was undertaken to document the reliability and safety of free flaps for reconstruction of defects in the head and neck.
The current series describes the career experience of a single surgeon performing microvascular free flaps for reconstruction of the head and neck region. I performed 119 microvascular free flaps in 118 surgeries in 115 patients during 117 periods of hospitalization occurring between August 1, 1995, and April 30, 1998. One patient underwent simultaneous transfer of 2 free flaps, while 3 patients underwent sequential transfer of 2 free flaps to manage either a complication of therapy or cancer recurrence. The study group consisted of 82 men and 33 women who had an age range of 19 to 88 years at the time of surgery. Preoperative comorbidity was scored according to the American Society of Anesthesiology scale. Twenty patients (17.4%) were in class 1 (normal); 58 (54.4%) in class 2 (mild systemic illness); 23 (20.0%) in class 3 (severe systemic illness that limits activities); and 14 (12.2%) in class 4 (severe systemic disease that causes immobilization and may be life-threatening).
One hundred eleven patients had defects arising from the treatment of head and neck cancer. Three patients had severe craniofacial trauma, and 1 patient was treated for a benign neoplasm of the mandible. Of the classifiable tumors, 74% of the flaps were used to repair defects that arose after the treatment of stage T3 or T4 primary cancers. Forty-one (34.4%) of the 119 free flaps were used in patients who had previously undergone radiation therapy. Ninety free flaps were carried out at UCLA Medical Center, Los Angeles, Calif, 16 flaps were performed at the West Los Angeles Veterans Affairs Medical Center, and 13 flaps were performed at UCLA-affiliated Los Angeles County hospitals.
Of the 119 flaps, 97 (81.5%) were performed for reconstruction of defects in the oral cavity or oropharynx, 11 (9.2%) were for skull base or midface defects, and 11 (9.2%) were for pharyngoesophageal defects. Free flap selection was based on specific patient and defect characteristics, but an effort was made to select donor sites that offered long vascular pedicles that contained large-diameter blood vessels whenever feasible. As a result of this approach, radial forearm, fibula, and rectus abdominis flaps together accounted for 95% of the donor sites selected, being used in 69, 27, and 17 cases, respectively. Latissimus dorsi free flaps were used in 3 cases, iliac crest in 2, and parascapular in 1. Radial forearm flaps were used for reconstruction of all pharyngoesophageal defects and were the most common flaps selected for reconstruction on the oral cavity and oropharynx. All of the fibula flaps were used for reconstruction of segmental mandibular defects. Rectus abdominis free flaps were the most common flap used for skull base reconstruction.
A 2-team approach for simultaneous tumor resection and free flap reconstruction was used whenever possible in patients undergoing immediate reconstruction of cancer-related defects. All microvascular anastomoses were hand sutured with the use of an operating microscope with either 9-0 or 10-0 nylon. Recipient artery selection in order of frequency included the facial artery, the superior thyroid artery, the lingual artery, the transverse cervical artery, and the occipital artery. End-to-side anastomosis to the internal jugular vein and end-to-end anastomosis to an internal jugular vein side branch were the most common methods of venous revascularization, followed by end-to-end anastomosis to the external jugular vein. Anticoagulation was used in all patients by means of intraoperative topical heparin solution irrigation of the donor and recipient vessels and postoperative daily administration of low-dose aspirin for the duration of the hospitalization.
The incidence of perioperative reconstructive and medical complications was recorded prospectively during the study period by means of a personal computer spreadsheet database. A complication was classified as perioperative if it occurred within 30 days of surgery.
There was 1 perioperative death (0.8%) in this series of 118 operations. This patient had an intraoperative cardiopulmonary arrest secondary to a tension pneumothorax while undergoing an iliac crest free flap to reconstruct an anterior segmental mandibulectomy defect. Postoperatively, he developed adult respiratory distress syndrome and progressive multiorgan system. The free flap remained well perfused until the time of his death on postoperative day 10.
Reconstructive complications that occurred during the perioperative period are listed in Table 1. Among the patients who survived the perioperative period, the rate of successful free flap transfer was 99.2% (117 of 118 free flaps). A fibula free flap was lost secondary to anastomotic venous thrombosis of 2 peroneal venae comitantes on postoperative day 2. At the time of unsuccessful attempted flap salvage, no evidence of anastomotic error, kinking or extrinsic compression of the vascular pedicle, or local infection was noted. Further workup disclosed a protein C deficiency as the most likely cause of the venous thrombosis. After correction of this hypercoagulable state, successful secondary reconstruction was carried out by means of a fasciocutaneous radial forearm free flap in conjunction with a mandibular reconstruction plate, which was later replaced by a fibula flap harvested from the contralateral extremity.
There were 2 cases (1.8%) of partial free flap necrosis. One case was caused by an error in flap insetting. A radial forearm flap used for total lower lip reconstruction was folded over a palmaris longus tendon static sling with excessive tension, resulting in ischemic necrosis of the ulnar margin of the flap. In the second case of partial flap necrosis, the entire skin paddle of an osteocutaneous fibula free flap was lost because of an ischemia-reperfusion injury sustained after failure of the arterial anastomosis on postoperative day 4. The bone graft of this flap was successfully salvaged by revision of the arterial anastomosis, but the skin paddle was replaced with a pedicled latissimus dorsi flap 1 week later.
The most common reconstructive complication was partial skin graft failure of the donor site over the flexor carpi radialis tendon, which occurred in 4 (6%) of 69 patients who underwent radial forearm free flaps. All such cases were successfully treated by forearm splinting and local wound care until reepithelialization occurred. Long-term donor site morbidity was seen in 2 patients. One patient developed laxity of the abdominal wall after harvest of a rectus abdominis flap, and another patient complained of chronic dysesthesia in the distribution of the superficial branch of the radial nerve after radial forearm flap harvest. Overall, 6 of the 12 reconstructive complications required further surgical intervention. Thus, the incidence of reoperation to treat perioperative reconstructive complications was 5.0% for the entire series.
The incidence of pharyngocutaneous fistulas was 2.7% in 113 free flaps used to repair defects of the upper aerodigestive mucosa. Three cases of salivary fistulas occurred in patients who underwent radial forearm free flaps for pharyngoesophageal reconstruction, all of whom had received preoperative radiation therapy. The incidence of fistulas in patients undergoing this method of pharyngoesophageal reconstruction was 27% (3 of 11 cases), compared with no fistulas seen in 97 free flaps used for oral cavity, oropharyngeal, or paranasal sinus reconstruction. Overall, the incidence of salivary fistula was 7% (3 of 41 cases) in patients who received preoperative radiation therapy.
Major medical complications occurred during 20 (17.1%) of the 117 hospitalizations. A total of 29 medical complications were recorded. Thirteen patients experienced 1 medical complication during their hospitalization, while 6 patients experienced 2 medical complications and 1 patient experienced 4 medical complications. The lungs (n=10), heart (n=8), and gastrointestinal tract (n=6) were the most commonly affected organ systems. Specific complications were as follows: pulmonary edema, 7 cases; myocardial infarction, 5; supraventricular tachycardia, 3; pneumonia, 3; upper gastrointestinal tract hemorrhage, 3; delirium tremens, 2; and colon perforation, stroke, pseudomembranous colitis, exacerbation of myasthenia gravis, liver failure, and perioperative death, 1 each.
As outlined in Table 2, the current report and several previously published large clinical series have firmly established the reliability of microvascular techniques for reconstruction of major defects in the head and neck region.5- 9 The current series achieved the highest success rate of free flap transfer to the head and neck region reported in the English-language literature from a single surgeon or institution, although Khouri10 reported a remarkable 100% success rate in a multi-institutional report that included 149 head and neck free flaps performed by 9 surgeons during a 12-month period.
The unsurpassed reliability of free flaps reported in this series arises from a combination of factors. A successful outcome in microvascular surgery depends on avoidance of technical errors during flap harvest and microvascular transfer. Such errors include trauma to the flap's vascular pedicle or cervical recipient vessels, creation of an imperfect vascular anastomosis, and selection of poor geometry between the donor and recipient vessels, which puts the flap at an increased risk for thrombosis secondary to kinking or extrinsic compression of the pedicle. These errors are most commonly cited as the causes of free flap failure, and contemporary postgraduate training in otolaryngology–head and neck surgery appears to be effective in preventing most such errors.11
Local infection that leads to thrombophlebitis of the vascular pedicle accounts for a substantial proportion of free flap failures among experienced microvascular surgeons.8 The majority of such infections arise secondary to salivary fistulas in cases where the defect includes a segment of upper aerodigestive mucosa. In this regard, the inherent superior vascularity of the soft tissue components of free flaps likely contributes to the low incidence of salivary fistula formation, contributing to the reliability of the technique. In most regional flaps used for head and neck reconstruction, the skin paddle usually lies at least 1 angiosome distant from the primary vascular territory of the nutrient vascular pedicle. For example, the caudal skin paddle of the pectoralis major myocutaneous flap that is commonly used for head and neck reconstruction usually lies within the vascular territory of the internal mammary artery, while the flap is based on the pectoral branch of the thoracoacromial artery.12 When regional flaps are designed with a skin paddle that extends more than 1 angiosome distant from the vascular territory of the nutrient vascular pedicle, the incidence of partial flap necrosis increases significantly, heightening the risk of salivary fistula formation.
On the contrary, the soft tissue component of a free flap is usually centered within the vascular territory of cutaneous perforators from the flap's pedicle. This provides the skin paddle of a free flap with a blood supply superior to that of most regional flaps, resulting in a decreased incidence of partial flap necrosis and fistula formation. Shah et al13 reported a 29% incidence of partial flap necrosis in a series of 211 pectoralis major myocutaneous flaps used for head and neck reconstruction. In the current series, the incidence of partial flap necrosis was 1.8%, and the incidence of salivary fistula formation was 2.7%.
No fistulas occurred in the current series of 97 free flaps used for oral cavity, oropharyngeal, or paranasal sinus reconstruction, in part because of their superior vascularity. All salivary fistulas occurred in patients who underwent a radial forearm free flap for reconstruction of total or subtotal laryngopharyngectomy defects after previous radiation therapy. The incidence of fistula formation with this method was 27% in the current series, which corresponds well to the 29% fistula rate reported in 45 cases described in the literature with the use of this technique of pharyngoesophageal reconstruction.14- 16 While this high rate of fistula formation contributed to delayed oral intake, there were no cases of flap loss from infection, and 2 of the 3 fistulas were treated successfully with closed suction drains that were placed at the time of free flap transfer. While the rate of fistula formation is lower with the more commonly used jejunal free flap, the radial forearm flap remains my flap of choice for reconstruction of most pharyngoesophageal defects where there is insufficient remaining tissue to achieve primary closure and the cervical esophagus has been preserved. A comprehensive review of the literature indicated that 82% of patients are able to maintain their nutrition by means of an oral diet after jejunal flap reconstruction,17 but other studies indicate that most patients are unable to achieve useful esophageal or tracheoesophageal speech.18,19 The rate of oral alimentation after radial forearm flap reconstruction of the pharyngoesophagus is 90%,14- 16,20 and the quality of tracheoesophageal speech is good.21
RADIAL FOREARM, fibula, and rectus abdominis free flaps accounted for 95% of the flaps used in the current series of 119 cases. These free flaps have certain characteristics that have resulted in their evolution into the standard flaps for contemporary head and neck reconstruction. Radial forearm flaps offer a thin, pliable skin paddle that is suitable for restoration of the complex, sulcular anatomy of the oral cavity and oropharynx, and it can be easily tubed to reconstruct the pharynx. Several previous series have documented that the radial forearm flap is the most reliable and the most frequently selected flap for head and neck microvascular reconstruction.5,9,22
Fibula free flaps offer up to 25 cm of vascularized bone, which is an adequate length to span any segmental defect of the mandible. Use of preoperative color flow Doppler ultrasonography23 or intraoperative Doppler stethoscope examination24 allows for accurate detection of cutaneous perforators from the peroneal artery, enhancing the reliability of the skin paddle component of the compound osteocutaneous fibula free flap. This information also allows for planned division of the posterolateral intermuscular septum of the leg in areas not containing perforators during flap design, greatly increasing the mobility of the skin component of the flap relative to the bone graft. With this technique, large skin paddles can be reliably harvested to reconstruct extensive or complex soft tissue oromandibular defects. In the current series, the largest skin paddle successfully transferred as part of a compound fibula free flap had a surface area of 550 cm2, comprising approximately two thirds of the circumference of the leg skin. The cutaneous portions of the radial forearm and fibula flap are amenable to sensory reinnervation through neural anastomosis of the lateral antebrachial cutaneous nerve or lateral sural cutaneous nerve to appropriate recipient nerves in the head and neck. Preliminary evidence suggests that sensory reinnervation might provide a functional advantage to patients who undergo major head and neck reconstruction.25,26
Rectus abdominis flaps were usually selected for reconstruction of skull base defects or in patients who underwent subtotal or total glossectomy. In both of these indications, the bulk provided by the rectus abdominis flap is advantageous, to provide separation of the intracranial space from aerodigestive secretions in the case of skull base reconstruction, or to allow for passive obturation of the oral cavity during deglutition after major glossectomy.
Other favorable characteristics shared by these 3 free flaps that were confirmed by the current series include limited morbidity of the donor site after flap harvest27- 29 and the ability to carry out simultaneous cancer resection and flap harvest with the use of 2 surgical teams to decrease the overall length of the surgical procedure.
In many respects, patient characteristics in the current series that might affect the incidence of reconstructive complications were similar to those of previously published series.6,8,9 Most patients had defects that arose after treatment of stage T3 or T4 upper aerodigestive tract cancers, with oral cavity or oropharyngeal tumors being the most frequent site of involvement. The prevalence of preoperative radiation therapy in the current series was similar to6,8 or higher than9 those in previously published series. Technical factors used in the current series, including anastomotic technique, recipient vessel selection, and prophylactic perioperative anticoagulation, were also similar to those in previously published series. It is unlikely that these factors impacted the improved free flap survival seen in the current series of microvascular surgery.
Kroll et al30 demonstrated that free flap selection can impact the likelihood of flap survival, as some donor sites appear to have a higher inherent risk of failure than others. In reviewing their experience with 400 consecutive free flap reconstructions, Disa et al31 concluded that the use of fewer but more reliable donor sites contributed to a successful outcome in free tissue transfer. It is likely that flap selection, with a heavy reliance on radial forearm, fibula, and rectus abdominis flaps, played an important role in the 99.2% incidence of successful free flap transfer achieved in the current series. These flaps all contain long vascular pedicles that frequently can reach recipient vessels in the contralateral part of the neck without requiring vein grafting. In the current series, no patients required vein grafts to lengthen the vascular pedicle, despite the fact that recipient vessel selection was limited in 24 cases (20.3%) where the patient had undergone a previous neck dissection. Two previous large series of microvascular head and neck reconstruction have correlated the use of vein grafts with an increased risk of free flap failure.6,9 When the diameter of the blood vessels in the vascular pedicle is less than 1 mm, the risk of free flap failure is markedly increased.10 Radial forearm, fibula, and rectus abdominis flaps have vascular pedicles containing vessels that commonly have an external diameter in excess of 2 mm. The inherent dependability of these flaps is demonstrated by the fact that the only complete flap loss in this series occurred in a patient with a documented hypercoagulable state.
The incidence of perioperative morbidity and mortality was acceptable in the current series, despite the fact that microvascular reconstruction frequently requires prolonged surgery in an elderly patient population. In general, the potential for postoperative medical and reconstructive complications is more closely related to the severity of preoperative comorbidities than to patient age or the length of the surgical procedure.22,30,32- 34 In this regard, it is likely that the low incidence of medical complications and the low perioperative mortality documented in the current series when compared to previously published series6 is a reflection of the relative scarcity of comorbidities in the current series. In the current series, 72% of the patients had no notable systemic illness or a mild degree of preexisting systemic illnesses (class 1 or 2), compared with a 74% prevalence of severe systemic disease (class 3, 4, or 5) reported by Schusterman et al.9
Concerns regarding the potential adverse impact of microvascular surgery on the costs of treatments appear to be unfounded. On the contrary, evidence indicates that the low rate of complications seen in patients who undergo microvascular reconstruction may contribute to an overall cost savings in the treatment of patients with head and neck cancer. When controlling for tumor site and stage as well as preoperative comorbidities, Brown et al35 demonstrated that patients undergoing free flap reconstruction spent fewer days in the intensive care unit and hospital than patients who underwent pedicled flap reconstruction. This difference was attributed to the rarity of postoperative complications in the group with free flap reconstruction. Kroll et al36 demonstrated than the mean cost of treatment was $28,460 in a group of 145 patients who underwent microvascular reconstruction of the oropharynx, compared with a mean cost of $40,992 in 33 patients who underwent a pectoralis major myocutaneous flap for oropharyngeal reconstruction.
In summary, microvascular free flaps have been demonstrated to be an extremely reliable form of head and neck reconstruction. With careful attention to avoid technical errors and heavy reliance on a few tissue donor sites that offer long vascular pedicles containing large-diameter blood vessels, the rate of successful free flap transfer can approach 100%, which contributes to a low incidence of reconstructive complications. While the surgery is frequently lengthy, perioperative medical complications are unusual in patients who undergo a thorough preoperative evaluation to identify comorbidities. The low incidence of complications may result in an overall treatment cost savings in this patient population in comparison with previous methods of reconstruction that relied primarily on regional flaps. Because of their reliability, free flaps have evolved to become the preferred method of reconstruction for most patients with major defects in the head and neck region.
Accepted for publication October 27, 1998.
Ninety-six percent of the cases described in this series underwent simultaneous surgery by 2 teams of surgeons. The high rate of successful free flap transfer in this series is a reflection of the surgical skills of Gerald S. Berke, MD, Thomas C. Calcaterra, MD, Elliot Abemayor, MD, PhD, Joel A. Sercarz, MD, Rinaldo F. Canalis, MD, Dan J. Castro, MD, James C. Andrews, MD, and Marilene B. Wang, MD, who as the cancer ablative surgeons were responsible for the dissection and preparation of cervical recipient vessels in these cases.
Reprints: Keith E. Blackwell, MD, Box 951624 UCLA Medical Center, Los Angeles, CA 90095-1624 (e-mail: firstname.lastname@example.org).