Primary site pathology. BCCA indicates basal cell carcinoma; ORN, osteoradionecrosis; and SCCA-IV, stage IV squamous cell carcinoma.
Militsakh ON, Werle A, Mohyuddin N, Toby EB, Kriet JD, Wallace DI, Girod DA, Tsue TT. Comparison of Radial Forearm With Fibula and Scapula Osteocutaneous Free Flaps for Oromandibular Reconstruction. Arch Otolaryngol Head Neck Surg. 2005;131(7):571-575. doi:10.1001/archotol.131.7.571
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
To compare our experience with the osteocutaneous radial forearm free flap (group 1) (n = 108) with other commonly used osteocutaneous free flaps (group 2) (n = 56) such as the fibula and scapula in single-stage oromandibular reconstruction.
Retrospective case review.
Tertiary-care academic medical center.
One hundred sixty-three consecutive patients who underwent 164 mandibular reconstructions with osteocutaneous free flaps.
Main Outcome Measures
Assessment of preoperative and intraoperative variables for both groups. We compared recipient-site complication rate, intensive care unit stay, total hospital stay, and postoperative function.
The most common donor site used was the radius (n = 108 [66%]), followed by the fibula (n = 36 [22%]) and scapula (n = 20 [12%]). Mean follow-up was 29 months (range, 1-116 months). Group 2 patients had larger soft tissue and/or bony defects. Surgical and medical complication rates and major donor site morbidity in group 1 were similar or better when compared with those in group 2. The lengths of the intensive care unit (4 vs 7 days; P = .009) and hospital stays (13 vs 15 days; P = .06) were shorter in group 1. Although the microvascular success rate was similar in both groups, the local wound complication rate was significantly better for group 1. The difference for the length of intensive care unit stay was statistically significant and potentially amounts to more than $6000 of savings. Functional outcomes, including the ability to tolerate oral diet, tracheostomy presence, and dental rehabilitation, were similar between the groups.
The primary site long-term morbidity, donor site morbidity, and postoperative function of osteocutaneous radial forearm free flaps are comparable to those of other commonly used osteocutaneous free flaps such as the fibula and scapula when used in single-stage oromandibular reconstruction.
Oromandibular reconstruction continues to be a challenge to the head and neck reconstructive surgeon. Composite resection for neoplasm or trauma due to high-speed missiles frequently results in large soft tissue and bony defects. For functional and cosmetic reasons, it is important to maintain the crucial 3-dimensional anatomical relationships, including mandibular continuity. Important physiologic functions such as deglutition, mastication, articulation, and oral competence can be severely affected if these anatomical relationships are not preserved.
Since the development of microvascular free-tissue transfer, osteocutaneous free flaps (OCFFs) have been a valuable option for the oromandibular reconstruction. Other techniques to restore mandibular continuity, which include mandibular bridging with a reconstruction plate or nonvascular bone grafting, have been used with varying success for certain mandibular defects.1- 3 Several donor sites are available for OCFF, with the fibula, scapula, and iliac crest being the most commonly used. The osteocutaneous radial forearm free flap (OCRFFF) was described in 1978,4 but its widespread use was limited due to early reports of significant donor site morbidity.5,6 The use of prophylactic plate fixation of the donor radius bone has decreased donor site morbidity to that seen with the fasciocutaneous flap and subsequently widened its use.7
The fasciocutaneous radial forearm free flap (FCRFFF) offers mobile, pliable, thin, sensate soft tissue without added bulk. These soft tissue characteristics allow for facile reconstruction of a variety of external skin defects and intraoral defects. These advantages of the FCRFFF made this a popular option for the reconstruction of oral soft tissue defects.8- 10 The OCRFFF offers a 1-stage reconstruction of mandibular and soft tissue defects while maintaining all the soft tissue advantages of FCRFFF.
We performed a retrospective medical chart review of the 163 consecutive patients at the University of Kansas Medical Center, Kansas City, who underwent 164 mandibular reconstructions from August 26, 1994, through April 9, 2003. This study was approved for exempt classification by the institutional review and privacy boards of the University of Kansas Medical Center, Kansas City. The patient data were collected and stored in compliance with all regulations of the Health Insurance Portability and Accountability Act of 1996.
The patients were divided into 2 groups. Group 1 consisted of patients who underwent oromandibular reconstruction using the OCRFFF with a technique previously described.11 Group 2 consisted of patients who underwent scapula and fibula OCFF (OCSFF and OCFFF, respectively) reconstructions.
The summary of the data collected for each patient is presented in Table 1. Major medical comorbidity was documented as the number of preoperative medical problems significantly affecting 1 or more physiologic systems (eg, cardiovascular, pulmonary, gastrointestinal tract, metabolic, and immune) for each patient, and the average was recorded for each group. Medical complications were recorded as any major medical complication (eg, pneumonia, myocardial infarction, major cardiac arrhythmia, or deep venous thrombosis) that occurred in the perioperative period. The tourniquet time was used as measurement of the harvesting time because it was consistently recorded for most of the OCRFFF and OCFFF. Surgical complications were recorded for each individual category such as wound infection or breakdown, fistulas, hematoma, hardware related, microvascular related, and donor site morbidity. We used unpaired t test, χ2 test, and regression analysis to analyze continuous and nominal data.
A total of 164 oromandibular reconstructions with OCFF were performed during the study period. One hundred eight patients underwent 108 OCRFFF reconstructions (66%) (group 1). Fifty-six patients (34%) underwent reconstruction with OCFFF (n = 36) or OCSFF (n = 20) (group 2). The mean age for groups 1 and 2 was equal at 62 years, and the ranges were similar (16-93 vs 13-93 years). The sex distribution was also similar between groups (male-female ratio, 68:40 [63% male] vs 36:20 [64% male]).
Medical history factors such as tobacco and alcohol use and comorbidities were also comparable between the 2 groups (Table 2).
Fourteen (9%) of the 164 patients underwent a composite resection because of significant mandibular osteoradionecrosis; 2 patients had secondary reconstructions of mandibular defects from previous composite resections at outside facilities; and 1 patient had reconstruction after a self-inflicted gunshot wound to the face. The remainder of patients underwent oncologic extirpative procedures followed by immediate reconstruction. Oral cavity neoplasms constituted 61%, oropharyngeal constituted 22%, and the remainder consisted of skin tumors, salivary gland neoplasms, primary bone tumors, and soft tissue tumors (Figure). The most common malignancy was squamous cell carcinoma, with most of these (97%) being stage IV disease.
Most of our patients (87%) underwent radiation therapy (XRT) as part of their treatment. Sixty-eight (42%) had preoperative XRT, and, in 14 patients, it resulted in mandibular osteoradionecrosis requiring composite resection of the nonviable tissue. Group 1 had a larger proportion of patients undergoing postoperative XRT after the primary surgical oncologic extirpative procedure (52% vs 38%); however, this was not statistically significant (P = .08). The average radiation exposure as measured in total dose of radiation to the primary site was the same in both groups (Table 3).
Posterior and lateral mandibular defects (ramus-body and body subunits12 only) were more often reconstructed with OCRFFF (84% vs 16%), whereas for more anterior defects (involving the hemisymphysis and/or symphysis subunit), the OCRFFF was used with the same frequency as other OCFFs (49% vs 51%). Mean bony defect in the group 1 was 6.6 cm (range, 3-12 cm) compared with 11 cm (range, 6-23 cm) in group 2. Soft tissue harvested when measured by the skin paddle area was the highest for the OCSFF (mean, 107 cm2), followed by the OCFFF (mean, 96 cm2) and then the OCRFFF reconstructions (mean, 75 cm2). Sixty-three radial bone grafts (58%) and 51 fibular and scapular grafts (91%) were intraoperatively osteotomized to contour the reconstructed part of the mandible. Neural anastomosis of recipient nerve to one of the antebrachial cutaneous nerves was performed in 61 (56%) of the OCRFFF. In only 2 flaps (4%) in group 2 (both were OCFFF with anastomosis to the lateral sural cutaneous nerve) was an attempt made to reinnervate the free flap.
Several plating systems were used to fixate the bone graft. The 2.4-mm locking reconstruction plate (LRP) was used in 97 patients, the 2.0-mm LRP was used in 44 patients, and earlier in the study period, the titanium hollow-screw reconstruction plate system was used in 18 patients. Three other reconstructions were performed with miniplates and 1 with a nonlocking compression plate. The most commonly used plating systems (titanium hollow-screw reconstruction plate system, 2.4-mm LRP, and 2.0-mm LRP) were used in both groups at similar frequencies (11:64:29 and 7:33:15). Fifteen patients in the OCRFFF group underwent a reconstruction using a double-barrel technique in which vascularized bony graft is osteotomized and folded on itself to offer twice the circumference of the ordinary radial bone graft.
The operating time was significantly lower in group 1 (mean, 12.9 vs 14.7 hours; P<.001). The tourniquet time was also significantly lower in group 1 (P = .01). Mean times for intensive care unit (ICU) and hospital stays were shorter for group 1 as well; however, statistically the difference was significant only for the length of ICU stay (P = .009 and P = .06, respectively). Noting that smaller defects were reconstructed in the OCRFFF group, a regression analysis was applied to determine whether the size of the soft tissue and bony defects correlated with increased hospital and ICU stays in the OCSFF and OCFFF groups. We found no correlation between the length of the hospital and ICU stays and the size of the bony defect (P = .27 and P = .54, respectively) or the size of the soft tissue defect (P = .49 and P = .43, respectively).
Primary site surgical postoperative complications were also the same or less for group 1 (Table 4). The problems with wound infection, breakdown, and postoperative fistulas were significantly higher in group 2 (P = .001). Medical postoperative complications (38 [35%] and 20 [36%] in groups 1 and 2, respectively), the rate of perioperative death (1 [1%] and 2 [4%] in groups 1 and 2, respectively), and hardware-related complications were not significantly different between the groups (Table 4). The hardware complication data were also analyzed with respect to different hardware plating systems used in both groups. Overall, the newer system (2.0-mm LRP) had a significantly lower complication rate then the older, most commonly used system (2.4-mm LRP) (8 [7%] vs 10 [18%]; P = .03, χ2 analysis).
The rate of revision of vascular anastomosis with or without hematoma evacuation because of free-flap pedicle compromise was similar in both groups (13 [12%] vs 7 [12%]). Total flap failure rate, defined as loss of bony and soft tissue, was the same in both groups (2% [2 in group 1 and 1 in group 2]). Partial flap failure, defined as partial soft tissue loss without bony compromise, was smaller in group 1 (3 [3%] vs 3 [5%]); however, this difference was not statistically significant (P = .4).
There were fewer major donor site complications in group 1 (Table 5), although this difference also was not statistically significant (P = .19). All the patients in the radial forearm group had prophylactic plating of their donor radii, and no clinically significant donor radius fractures have occurred in these patients. One patient required removal of the radius fixation plate due to hardware loosening at 12 months after the original surgery. With regard to minor complications, the radial forearm group had a 30% rate of minor flexor tendon exposure after a split-thickness skin graft of the donor site. None of these cases required surgical intervention, and all resolved without sequelae and with local wound care.
The mean follow-up length was similar for both groups (29.0 vs 29.2 months), with an overall mean follow-up of 29.0 months and median follow-up of 21 months. Most of the patients in our study presented with stage IV disease (97%) and, as expected, the average and median numbers for follow-up data were significantly affected by the large number of patients who died within first year after the surgery (n = 52 [32%]). With these patients excluded, the overall mean and median follow-up increased to 40.0 and 33.9 months, respectively. At follow-up, group 1 had a smaller portion of patients who were 100% dependent on gastrostomy tubes (32 [30%] vs 25 [45%]) (P = .12). In addition, the proportion of people with diet intake entirely by mouth was the same in both groups (48% [52 in group 1 and 27 in group 2]). A similar number of patients were able to undergo decanulation (88 [81%] vs 44 [79%]). A slightly higher proportion of patients underwent dental rehabilitation in group 1 (34 [31%] vs 13 [23%]), although this difference was not statistically significant (P = .27). Most of these patients were rehabilitated with tissue-born dentures. Three patients (5%) in group 2 undewent rehabilitation with implant dentures.
First introduced in 1978, the radial forearm free flap has been gaining popularity in head and neck reconstruction.4 Its superior soft tissue characteristics, including thin pliable skin that maintains mobility and is easy to contour, and its lack of bulk made the RFFF a workhorse flap for head and neck reconstruction. These characteristics allow a variety of soft tissue defects such as palate, tongue, floor of mouth, and external defects to be easily reconstructed. The RFFF was limited in its use for bony reconstruction in the head and neck because of the morbidity of harvesting the radius and the limited bone stock. Its use often necessitated an additional procedure, often a second free flap.9,13 Early reports of significant donor site morbidity, such as donor radius fractures, have deterred the use of OCRFFF.5,6 With the recent introduction of new techniques for the prophylactic plating of the radius bone, the donor site morbidity has drastically decreased to levels similar to harvest of the FCRFFF.7
In this study, we evaluated our single-institution experience with OCRFFF and compared this group of patients with those who underwent reconstruction with other osteocutaneous free flaps in our institution during the same time period and with the same group of surgeons.
The microvascular success rate for OCRFFF was greater than 98%. This was similar to the success rate in the other group in our study (98%). It was also comparable to the success rate for OCFF at other institutions.10,14 Recipient site surgical complication rate in our study was significantly better in the OCRFFF group compared with the other OCFF groups (scapula and fibula). Our data were also similar to the literature-reported complication data for FCRFFF.15 Few smaller studies have reported their experience with OCRFFF. Thoma et al6 performed a retrospective review of 60 patients, and, although the total recipient site complication rate was as high as 40%, an overall microsurgical success rate was similar at 98%. Villaret and Futran16 reviewed mandibular reconstruction with OCRFFF of posterior mandibular defects in 21 patients. They found a radial graft a suitable option for the oromandibular reconstruction.
The possibility of the concurrent harvest without the need for repositioning of the patient and the straightforward forearm anatomy allows the reconstructive surgeon to significantly decrease operating time. The longer pedicle of the forearm flap and larger-caliber vessels also technically simplify the procedure. As previously shown and confirmed by our findings, the radial bone graft can be safely osteotomized to contour the reconstructed mandible.6,17 Another advantage of the radial forearm flap is that it is a reliable means of restoring sensation of the free flap.8,9
In our review, free-flap harvest time and total operating time were significantly less in the OCRFFF group. As a consequence, operative time in combination with decreased ICU stay offers a potentially significant financial savings for the patient. At present, at our institution, with ICU bed charges of more than $2400 a day, the shorter ICU stay alone offered more than $6000 in cost savings.
Total hardware-related complication rate was similar between both groups in our study. This was also comparable to the literature-reported complication rate.14,18 Although the distribution of the type of plating system used to fixate the free-flap bone graft remained the same in both groups, over the years newer plating systems were becoming available and being used. The implementation of these newer plating systems (such as the 2.0-mm LRP) resulted in significant decrease of hardware complication rate (5% vs 18%; P = .03, χ2 analysis).19
Good functional results were achieved as most of the patients in OCRFFF group were able to resume an oral diet and undergo decanulation. The same proportion of patients in both groups underwent dental rehabilitation.
We previously reported7 and now confirm with this study that OCRFFF donor site morbidity can be significantly minimized, although a few may point out that cosmesis of the donor site is less than desirable in some patients.
The main limitation of OCRFFF is the size and the caliber of the bone graft that can be safely harvested without added donor site morbidity. The largest radial bone graft harvested in our study was 12 cm. For the less common reconstructions, such as for subtotal mandibular defects that require longer grafts, the fibula graft can offer up to 25 cm of bone. The smaller caliber of radial bone graft did not prove to have more short- or long-term complications in our set of patients, even at long follow-up. In addition, although the thin skin of the radial forearm flap can be an advantage, if a very large soft tissue defect is present and soft tissue bulk is desirable, other options such as the scapular free flap may need to be explored.
This study is a retrospective review, and there is inevitable preoperative patient selection bias for the choice of the flap. This bias was reflected in the size of the bony and soft tissue defects, as the OCFFF was chosen for larger bony defects, and OCSFF was chosen for larger soft tissue defects. We understand that a shortcoming of comparing the OCRFFF directly with other flaps at our institution is this patient selection bias; however, it is encouraging that the rate of surgical complication and postoperative course was better for OCRFFF. Moreover, when the data for all of the reconstruction procedures in our study (N = 164) were stratified by the size of the mandibular defect, the patients with larger defects did not have a direct correlation with increased hospital and ICU stays (linear regression analysis, P = .54 and P = .27, respectively). This finding is especially significant in lieu of the finding of shorter ICU and hospital stays for the OCRFFF group, indicating a shorter inpatient rehabilitation time for these patients.
The OCRFFF can be successfully used for 1-stage reconstruction of oromandibular defects. It provides enough bone for the reconstruction of the most common composite defects with the primary and donor site complication rates comparable with or better than other osteocutaneous free flaps.
Correspondence: Terance T. Tsue, MD, Department of Otolaryngology–Head and Neck Surgery, University of Kansas School of Medicine, 3901 Rainbow Blvd, Mail Stop 3010, Kansas City, KS 66160.
Submitted for Publication: May 20, 2004; final revision received November 19, 2004; accepted December 16, 2004.
Previous Presentation: This study was presented at the 6th International Conference on Head and Neck Cancer; August 8, 2004; Washington, DC; and was awarded the Best Resident Clinical Paper Award.
Acknowledgment: We thank Arlene Rodriguez for help with data collection and Debra Park, PhD, for her help with statistical analysis.