With the evolution of organ sparing therapy for treatment of laryngeal and hypopharyngeal squamous cell carcinoma, chemoradiotherapy has increased over time. Thus, with laryngectomy becoming a salvage procedure, wound complications and pharyngocutaneous fistula have become an increasing concern. Wound healing complications can have a multifactorial origin including diabetes, hypothyroidism, radiation dosage, and chemotherapy [6, 8, 11]. Previous studies have confirmed the added benefit of vascularized tissue at the time of pharyngeal closure to reduce fistula and promote the return to oral diet [5, 8, 9]. In this study we review our reconstructive experience with patients who suffered pharyngocutaneous fistula after primary closure of a salvage laryngectomy wound. In our series of 20 patients we have demonstrated fistula closure to be a complex process in which only half of patients have definitive fistula closure with a single procedure. The current study supports the use of preplanned vascularized tissue use in the wound of the neopharynx after salvage laryngectomy.
It was elected to reconstruct most patients referred for fistula repair with radial forearm free flaps (RFFF), however the techniques of fistula closure varied with time and increasing experience. Each fistula required analysis of the amount of remaining mucosa, pliability and quantity of neck skin, and the presence of stricture to formulate a reconstructive plan. Patients with stricture-based fistula had the stricture incised and the free flap tissue inserted into the band of circumferential stricture, to allow passage of a 12 or 14 mm salivary bypass tube. When larger fistula (the classic three-hole defect) were closed using a single folded flap to reconstitute both the pharyngeal lining and the external skin, the vascular pedicle of the RFFF folded on itself and obstructed venous flow and as result led to refistulization. We found in the present study that patients who underwent a "patch" graft using the RFFF to reconstruct the pharyngeal lining, and then had good skin quantity and laxity to close the skin primarily overlying the free flap, had fewer leaks and required fewer additional procedures. Thus, in the patient population with minimal mucosa and skin available, we deduced over time that the combination of free flap patch closure with pedicled pectoralis muscle created the ideal reconstruction. The technique of folding the radial forearm skin paddle on itself to obtain closure of the internal and external lining did not routinely work well because of kinking the pedicle at the fold or re-fistulizing at the de-epithelialized portion of the skin paddle. An example of staged pectoralis muscle flap followed by successful folded RFFF is shown in Figure 1.
One flap failed in the series (4.5% failure rate). In this RFFF, the recipient vessels were the superior thyroid artery and external jugular vein. This flap was folded to allow the distal portion of the forearm to seal the pharynx and proximal paddle to close the neck. This patient required take-back on POD 2 and was salvaged only to require additional take-back on POD 3, complete venous clot was encountered, which was thought to be secondary the tight folding of the flap.
Rather we have favored the free flap to reconstitute the pharyngeal lining, while the pectoralis flap provides well-vascularized tissue as a substrate for skin grafting or granulating external wound closure. In patients who have already had attempts at closure via pectoralis flap, we have repositioned the pectoralis muscle to protect neck vasculature and relied on staged double free flaps to seal the neopharynx. Others have reported the benefit of using a single anterior lateral thigh flap with two skin paddles to perform internal and external epithelial lining [12]. It was noted in this study that earlier repair was less prone to complications [12]. In our patient series, the data did not support this finding. Most of our fistula patients presented for repair within the first six months (65%), and two of those 13 patients never achieved fistula closure. However, we did successfully close pharyngocutaneous fistula of more than 12 months duration in 3 patients without any complications.
As the closure technique has evolved over time, so has the selection of recipient vessels. In this patient series donor neck vessels were used 54.5%, most commonly the facial vessels, and internal mammary vessel were used in 45.5% of patients. Because neck exploration and vessel anastomosis was used in the two patients who developed carotid blowout, the use of cervical vessels in these patients was reconsidered. Both carotid blowout patients had previous neck dissection procedures, and chemoradiotherapy. The repetitive dissection of the chemoradiated neck with potential exposure to saliva for free flap vessel anastomosis was felt to significantly increase the risk for carotid blowout. To this end, the choice of recipient internal mammary vessels, as seen in Figure 2, allowed for free flap closure of the fistula without repeated dissection of neck vasculature in the field of radiation and previous surgical scarring. The combination of RFFF and internal mammary vessel is ideal because of the length of the vascular pedicle, the thin and generous quantity of skin, and the unradiated, donor vasculature. When the fistula is located suprastomally compared to in the base of tongue region there is generous length of the vascular pedicle. The use of transverse cervical vessels is also appropriate in this setting as described by others [12].
Interestingly, when the demographic factors were analyzed, the duration of the fistula prior to reconstruction did not impact the success of reconstruction. The length of time of the fistula being present probably does not increase the inflammatory and scarring factors in the wound bed, but more likely that the event of a fistula precipitates the inflammatory and scarring cascade to occur. Also, it was noted that untreated hypothyroidism was a significant comorbidity (p = 0.01) that predisposed to failure to close the fistula. This finding is also supported by previous studies [10, 13]. Screening TSH levels should be a routine part of head and neck cancer surveillance. In patients who develop fistula after primary closure of a salvage larygectomy, the need for screening TSH is heightened. In our series of patients, when elevated TSH levels were noted, institution of thyroid hormone replacement and delay of free flap closure help to improve wound healing. For example, patient 11 (Table 1), presented with wound breakdown fistula after salvage laryngectomy, unsuccessful pectoralis pedicled flap closure, and a TSH level of 97.9. The patient underwent repositioning of the pectoralis muscle to protect the neck vasculature, thyroid hormone replacement therapy, and wound care. The first free flap stage was delayed several weeks to facilitate improved healing. A tubed rectus free flap was successfully performed, as all pharyngeal mucosa was lost, to intentionally divert saliva away from the stoma and contralateral neck vessels. Finally, four weeks after the first free flap, a RFFF was used to patch the superiorly based (base of tongue) fistula. Figure 2 demonstrates a patient with similar flap staging.
Recently published literature supports the use of vascularized tissue for decreased fistula in salvage laryngectomy wounds. Driven et al. found that patients undergoing salvage surgery within one year of completing concurrent chemoradiotherapy, or high dose radiotherapy (>64 Gray) were at increased risk for pharyngocutaneous fistula, 34.2% vs. 15.7% [5]. In 2007, Withrow et al. noted that use of vascularized free flap for laryngectomy closure had 18% fistula rate compared to 50% for primary closure [9]. Others have noted improvement in rate of pharyngocutaneous fistula after using pectoralis pedicled flaps for reinforcement of a primary laryngectomy closure after chemoradiation failure [8]. In our series of 20 patients we have demonstrated fistula closure to be a complex process in which only half of patients have definitive fistula closure with a single procedure. The current study supports the use of preplanned vascularized tissue use in the wound of the neopharynx after salvage laryngectomy. Although this may add time to an oncologic procedure, simultaneous planned use of pectoralis pedicled muscle or free flap tissue does decrease fistula and its associated complications [9].