The mean duration of fistula is shown according to closure technique. Fistula duration was longest for primary closure at 14.0 weeks. For pectoralis onlay, mean fistula duration was 9.0 weeks. Fistula duration was shortest for patients closed using interposed free tissue. The longer duration of fistula for patients who underwent primary closure was statistically significant (P = .004). Error bars indicate standard deviation.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Patel UA, Moore BA, Wax M, et al. Impact of Pharyngeal Closure Technique on Fistula After Salvage Laryngectomy. JAMA Otolaryngol Head Neck Surg. 2013;139(11):1156–1162. doi:10.1001/jamaoto.2013.2761
Copyright 2013 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.
No consensus exists as to the best technique, or techniques, to optimize wound healing, decrease pharyngocutaneous fistula formation, and shorten both hospital length of stay and time to initiation of oral intake after salvage laryngectomy. We sought to combine the recent experience of multiple high-volume institutions, with different reconstructive preferences, in the management of pharyngeal closure technique for post–radiation therapy salvage total laryngectomy in an effort to bring clarity to this clinical challenge.
To determine if the use of vascularized flaps in either an onlay or interposed fashion reduces the incidence or duration of pharyngocutaneous fistula after salvage laryngectomy compared with simple primary closure of the pharynx.
Multi-institutional retrospective review of all patients undergoing total laryngectomy after having received definitive radiation therapy with or without chemotherapy between January 2005 and January 2012, conducted at 7 academic medical centers.
Academic, tertiary referral centers.
The study population comprised 359 patients from 8 institutions. All patients had a history of laryngeal irradiation and underwent laryngectomy between 2005 and 2012. They were grouped as primary closure, pectoralis myofascial onlay flap, or interposed free tissue. All patients had a minimum of 4 months follow-up.
Main Outcomes and Measures
Fistula incidence, severity, and predictors of fistula.
Of the 359 patients, fistula occurred in 94 (27%). For patients with fistula, hospital stay increased from 8.9 to 12.1 days (P < .001) and oral diet initiation was delayed from 10.5 days to 29.9 days (P < .001). Patients were grouped according to closure technique: primary closure (n = 99), pectoralis onlay flap (n = 40), and interposed free tissue (n = 220). Incidence of fistula with primary closure was 34%. For the interposed free flap group, the fistula rate was lower at 25% (P = .07). Incidence of fistula was the lowest for the pectoralis onlay group at 15% (P = .02). Multivariate analysis confirmed a significantly lower fistula rate with either flap technique. For patients who developed fistula, mean duration of fistula was significantly prolonged with primary closure (14.0 weeks) compared with pectoralis flap (9.0 weeks) and free flap (6.5 weeks).
Conclusions and Relevance
Pharyngocutaneous fistula remains a significant problem following salvage laryngectomy. Use of nonirradiated, vascularized flaps reduced the incidence and duration of fistula and should be considered during salvage laryngectomy.
The management of advanced laryngeal cancer has fundamentally changed over the past 30 years. Quiz Ref IDThe Department of Veterans Affairs (VA) Laryngeal Cancer Study demonstrated that combination therapy with chemotherapy and radiation therapy allowed preservation of the larynx without a decrement in disease control compared with total laryngectomy (TL) with adjuvant radiation therapy (RT).1 This organ preservation approach was validated by the results of the Intergroup Radiation Therapy Oncology Group 91-11 (RTOG 91-11) study, which demonstrated the superiority of primary concurrent chemoradiation therapy (CRT) over induction chemotherapy–radiation therapy (IC-RT) and RT alone.2 Permutations of CRT continued to emerge, with primary TL increasingly restricted to extensive extralaryngeal tumor spread or to the salvage setting for persistent or recurrent disease. Chemoradiation therapy for organ preservation within the context of a multidisciplinary treatment evaluation received widespread support from the oncology community.3
However, as these organ preservation strategies disseminated from academic to community settings, concerns were raised about the toxic effects of primary CRT, the complications associated with salvage surgery, and even the efficacy of CRT.4 Heightened scrutiny of the VA Laryngeal Cancer Study and RTOG 91-11 data on salvage TL revealed that major wound complications occurred in as many as 59% to 64% of patients, with pharyngocutaneous fistula (PCF) after concurrent CRT developing in at least 30% of patients, and in some series, up to 75%.5,6 Although other factors such as postoperative anemia, prior tracheostomy, and concurrent neck dissection have been linked to PCF formation, subsequent reports drew further attention to the increased wound healing problems encountered with salvage TL after CRT.7,8
Institutions that participated in the seminal trials of CRT were logically in the vanguard of not only identifying the problems but also proffering recommendations to minimize complications of salvage TL.9 Although it is intuitive that the introduction of vascularized tissue into a compromised recipient bed would improve wound healing, debate persists on the benefits of transferring vascularized tissue and the best techniques to accomplish this—to either patch or bolster the pharyngeal closure suture line with either regional or free flaps.10
Quiz Ref IDSeveral studies have demonstrated that free tissue transfer can decrease the rate of major complications, PCF formation, and reoperation in post-CRT salvage TL.9,11 A free tissue patch incorporated into the suture line was shown to decrease PCF formation from 50% to 18%, with notable decreases in the rate of stricture and feeding tube dependence in these patients as well.11 However, when free flaps were used to reinforce the suture line in an on-lay technique, there was no difference in the rates of PCF formation, although the resulting wounds were more likely to heal faster and without additional surgical procedures than those not reinforced.12 Other studies describe the application of pectoralis myofascial pedicled flaps to reinforce primary neopharyngeal closures, and published data demonstrate a variable but favorable reduction on PCF formation.13-15
Despite numerous publications on this topic, no consensus exists as to the best technique, or techniques, to optimize wound healing, decrease pharyngocutaneous fistula formation, and shorten both hospital length of stay and time to initiation of oral intake after salvage laryngectomy. This has been attributed to the limitations of small cohort sizes, retrospective analyses, and institutional variation, whether from philosophical bias or technical capability.10,12 We sought to combine the recent experience of multiple high-volume institutions, with different reconstructive preferences, in the management of pharyngeal closure technique for post-RT salvage TL in an effort to bring clarity to this clinical challenge. Accordingly, the purpose of this study was to assess the possible impact of closure technique on the incidence and severity of fistula following salvage TL.
A retrospective review of all patients undergoing TL after having received definitive RT with or without chemotherapy between January 2005 and January 2012 was conducted at 7 academic medical centers. Participating centers included The Johns Hopkins Hospital, Baltimore, Maryland; Greater Baltimore Medical Center, Baltimore; University of Alabama, Birmingham; Oregon Health Sciences University, Portland; Ochsner Medical Center, New Orleans, Louisiana; University of Nebraska Medical Center, Omaha; and Northwestern University, Chicago, Illinois. Institutional review board approval was granted at each participating center. Patients with extended laryngectomies or pharyngotomy defects that could not be closed primarily or who received postoperative RT were excluded. Pharyngotomy closure type included simple primary closure, the use of pedicled pectoralis myofascial flaps as an onlay over the closed pharyngotomy, and the interposition of free fasciocutaneous tissue into the pharyngeal defect. Choice of pharyngotomy closure was at the discretion of the surgeon at the time of surgery. All patients had a minimum of 4 months follow-up.
Patient, disease, and treatment-related factors were retrospectively collected and compiled into a master database. Demographic factors including preoperative laboratory values, comorbidities, and alcohol and tobacco status were recorded. Disease factors included the initial and recurrent tumor staging and concurrent procedures performed with TL (eg, tracheoesophageal puncture, gastrostomy placement, neck dissection). Treatment factors included the interval between completion of radiation and salvage TL, type of pharyngotomy closure technique, and formation and duration of fistula. Pharyngotomy closure was categorized into the following 3 groups: primary closure (suture or stapler), pectoralis myofascial onlay flap, and interposed fasciocutaneous free flap (radial forearm, ulnar, and anterolateral thigh), with a small number of musculocutaneous flaps also included in this group.
Descriptive variables were summarized by mean (SD) values for continuous variables and number (percentage) for categorical variables. A t test was used to compare differences in means between groups. One-way analysis of variance was used to analyze relationships between categorical factors and continuous responses. A contingency analysis was used to analyze relationships between categorical factors and responses. Univariate analyses were performed using the Pearson bivariate correlation coefficient. Multivariate analysis was performed to analyze the combined effects of multiple independent variables (age, closure type, T classification, primary site, hemoglobin, and albumin) on fistula formation. P < .05 was considered statistically significant. Statistical analysis was performed using Jmp 10 software (SAS Institute Inc).
A total of 359 patients who underwent salvage laryngectomy were included in the study, with a mean age of 64 years and a male to female ratio of 4:1. Bivariate analysis examining the association between patient characteristics and the development of pharyngocutaneous fistula is summarized in Table 1. Quiz Ref IDOf the 359 patients, 94 (26%) developed postoperative pharyngocutaneous fistula. Sex, tobacco use, and incidence of comorbidity were similar between fistula and nonfistula groups and did not have apparent association with fistula. Similarly, there was no significant difference in preoperative hemoglobin or albumin levels between the fistula and nonfistula groups. Use of prior chemotherapy with radiation was also not significantly associated with fistula. Mean age was noted to be marginally lower for patients with fistula (62 years) compared with patients without fistula (65 years).
Intraoperative and postoperative factors and their association with fistula are given in Table 2. Concomitant neck dissection or placement of a tracheo-esophageal prosthesis did not affect fistula development. The rate of fistula varied considerably depending on closure technique. In 99 patients who underwent primary closure of the pharynx with no flap support, a fistula developed in 34 (34%). However, the fistula rate was significantly lower (6 of 40 [15%]) when an onlay pectoralis myofascial flap was used (P = .02). For the 220 patients undergoing interposed free flap closure of the pharynx, fistula developed in 54 (24%) (P = .07). Of these 220 patients with interposed flap reconstruction, most were radial forearm (n = 170) and anterolateral thigh (n = 25). The remaining patients had ulnar flaps, and there was a small collection of musculocutaneous flaps. There was no statistical difference in fistula rates among subgroups of these 220 patients. Of the 260 patients who underwent any type of flap closure, there was a single case of flap failure, yielding a flap success rate of 99.6%. Irrespective of closure method, patients with fistula experienced a significantly longer mean hospital stay (12.1 vs 8.9 days; P < .001). Mean duration until oral diet initiation was similarly prolonged for patients with fistula (29.9 vs 10.5 days; P < .001).
Of the 94 patients with fistula, 83 ultimately closed, with 11 patients experiencing prolonged fistula that remained open through the available follow-up period in this study. There was no significant association between the 11 persistent fistulas and the technique of pharynx closure. Fifty-two patients did return to the operating room 1 or more times to attempt fistula closure. There was no significant difference in the use of such operative intervention between groups based on closure technique. Overall duration of fistula averaged 9.1 weeks. When stratified by closure technique, mean duration of fistula for patients who initially underwent primary closure was 14.0 weeks. This duration was significantly shorter for patients with pectoralis onlay flaps (9.0 weeks) as well as interposed free flap (6.5 weeks) (Figure). This difference in fistula duration between primary closure and either flap technique was statistically significant (P = .004). Findings from multivariate analysis demonstrated that patient age, closure technique, T classification, and primary site were independent predictors of fistula formation (P = .01; R2 = 0.47).
Pharyngocutaneous fistula remains a common complication following salvage TL, with profound deleterious effects on wound healing, nutrition, hospitalization duration, initiation of oral diet, use of medical resources, cost, and quality of life. Meta-analysis of risk factors contributing to fistula following laryngectomy include anemia, prior tracheotomy, preoperative RT, and preoperative RT with concurrent neck dissection.7 Especially notable is that the incidence, severity, and duration of fistula in patients who received preoperative RT are greater than those in patients who did not. Our study confirms the previously published finding5,6,8 and the almost universally held belief that pharyngocutaneous fistula after salvage laryngectomy remains common. With simple primary closure, the fistula rate in our study was 34%, which remains considerably higher than what is generally seen in the nonsalvage setting. In addition, when fistula did occur in our studied population, hospital stay was significantly prolonged, as was delay to oral diet initiation. Accordingly, we believe that this particular clinical problem is worthy of further study and also feel justified in considering surgical strategies to reduce fistula incidence in this population.
The rationale behind both onlay and interposed flap reconstruction is that incorporation of nonirradiated tissue in the form of a vascularized flap should aid wound healing and prevent fistula formation. Quiz Ref IDFor patients with a pectoralis myofascial flap reinforcing the pharyngeal closure in an onlay fashion, the fistula rate was lowest at 15%. This finding was consistent with previously published studies examining pectoralis reinforcement of pharynx repair.13-15 There are several advantages to using the pectoralis flap in this setting. The muscle flap is generally harvested while the pharynx is being closed and thus adds minimally to the length of the operative case. Being a regional rotation flap, the pectoralis flap does not require a separate reconstructive team or microvascular experience. Finally, no postoperative flap monitoring is needed for the pectoralis flap that might otherwise be indicated for free flaps. This must all be weighed against the morbidity of the pectoralis flap, such as (1) shoulder dysfunction following flap harvest, (2) impaired postlaryngectomy speech due to presence of muscle over the neopharynx, and (3) excessive bulk of muscle, which is cosmetically unappealing and may result in the inability to primarily close the neck incisions.12 While neck and shoulder dysfunction is detectable after pectoralis flap use, it remains unclear to what degree this affects a patient’s quality of life.16 The present study did not specifically address the incidence of shoulder dysfunction in patients undergoing pectoralis flap reinforcement, nor was the impact of such flaps on speech measured. Finally, for cases in which the bulk of the pectoralis prevents tension-free closure of the skin flap, a portion of the pectoralis muscle can be left exposed and will heal well by secondary intention or skin graft.
Previous studies have examined the role of free tissue for salvage laryngectomy closure in both an onlay and interposed fashion. Fung et al12 have used free fascial flaps in an onlay fashion but did not demonstrate a significantly reduced rate of fistula though they did note a reduced duration of fistula with free tissue transfer. Withrow et al11 examined free tissue fasciocutaneous flaps that were inset into the pharyngeal closure and found a significant reduction in fistula rate. It is primarily this finding that has led several members of our study group to pursue this surgical strategy when dealing with salvage laryngectomy wounds. Generally, the radial forearm or anterolateral thigh flap is used and can accordingly be raised simultaneously with the resection. Inset of the flap is straightforward and can be performed in a time similar to primary suture closure. The microvascular anastomosis adds time, however, to the overall procedures and requires identification and preparation of donor vessels. With both forearm and anterolateral thigh flaps, there is minimal functional morbidity, and patient complaints are primary cosmetic.17,18 In the present study, we did not specifically address morbidity associated with free tissue transfer, which may include flap failure, hematoma, microvascular revision, and donor site morbidity. For institutions that routinely perform free tissue transfer to the head and neck, the additional length of operative time is generally modest and on the order of 1 to 2 hours, and the morbidity associated with free tissue transfer is low. However, free tissue transfer does require microvascular expertise, equipment, and flap monitoring, which add significantly to health care expenditures.
Our results demonstrated a lower fistula rate with pectoralis major onlay flaps (15%) compared with interposed free flap repairs (25%), although the difference between these 2 groups did not reach statistical significance. Quiz Ref IDHowever, there are several factors that may favor a muscle onlay technique over an insetting technique. First and foremost is that insetting a flap into the pharyngotomy effectively doubles the length of closure, increasing the potential length for fistula formation. Second, a muscle onlay reinforces the standard pharyngotomy closure with well-vascularized fascia and muscle. Third, the vascularity of the pectoralis muscle may be more robust and reliable than the peripheral edges of a free flap and thereby more capable of sealing off the pharyngotomy. Furthermore, the deep fascia surrounding the pectoralis muscle is rich in hyaluronan thought to play an important role during the earliest stages of wound healing.19,20
Interestingly, our study did not identify other factors that are commonly associated with fistula.7 Patients who developed fistulas did not demonstrate higher rates of anemia or hypoalbuminemia, as has been found in previous studies.7,21 Similarly, concomitant neck dissection and use of chemotherapy in conjunction with prior RT were not associated with fistula. One possible explanation for this is that use of vascularized tissue may mitigate the impact of these other factors on fistula development. Perhaps the beneficial wound healing effect of flap reconstruction is sufficient to overcome the presumed detrimental effects to wound healing of these other factors. With more than 70% of study patients undergoing some type of flap reconstruction, it is possible that other such variables no longer remain independent predictors of fistula. For the 33 patients with fistula who underwent primary closure, we could also not demonstrate an association with the usual predictors of fistula, though this may be attributed to the small sample size.
Because this was a multi-institutional retrospective study, patients were not randomized and the choice of closure was at the discretion of the surgeon. Some institutions contributing patients used a variety of closure techniques, which may result in bias because different closure options may have been chosen for particular patient factors. Other institutions used exclusively 1 closure type for all patients, thus reducing patient selection bias. That being said, most surgeons in the group demonstrated a general preference for a particular closure technique. As such, this study is a collection a different surgeons’ experience with each of their respective preferred closure techniques. We acknowledge that this results in a number of areas of potential bias. Of note is that all patients included in this study had sufficient pharyngeal mucosa allowing for primary pharyngeal closure at the end of salvage TL. Therefore, reconstructive technique was not related to a particular defect but was rather used solely to prevent postoperative fistula. Future studies that may randomize patients to various closure techniques would reduce patient selection bias and perhaps better answer the questions regarding closure technique.
There are a number of limitations to this study making it premature to definitively conclude an optimal closure technique. As previously mentioned, this is a retrospective study combining several different surgical practice patterns, where choice of pharynx reconstruction was left to the discretion of the surgeon. The risk of bias is considerable compared with possible future studies in which patients are randomized to various closure techniques. In addition, the present study focused on the particular morbidity of fistula incidence and duration. We did not address morbidity specifically attributed to the more complex reconstruction. While we suspect the morbidity of both pedicled and free flaps is low in experienced hands, this would need to be clearly measured and demonstrated in future studies. Finally, we did not perform any financial analysis to determine the cost-effectiveness of any particular surgical strategy. Despite the increased costs associated with regional or free flap reconstruction, we expect that the benefit of lower fistula rates may lead to substantial overall heath care cost savings.
In conclusion, our study confirms that pharyngocutaneous fistula remains a troublesome complication following salvage laryngectomy. Both length of hospital stay and time until oral diet initiation are markedly prolonged for these patients. The incidence of fistula was lowest for patients who underwent pectoralis onlay flap. Use of either a pectoralis myofascial onlay flap or a fasciocutaneous free flap interposed into the pharyngeal defect resulted in decreased incidence of fistula formation. Duration of fistula was also significantly reduced with either flap technique compared with primary closure alone. The present study supports these surgical strategies for pharynx reconstruction during salvage laryngectomy.
Corresponding Author: Jeremy D. Richmon, MD, Department of Otolaryngology–Head and Neck Surgery, Johns Hopkins University, 601 N Caroline St, Sixth Floor, Baltimore, MD 21287-0910 (firstname.lastname@example.org).
Submitted for Publication: February 16, 2013; final revision received March 22, 2013; accepted March 22, 2013.
Published Online: April 10, 2013. doi:10.1001/jamaoto.2013.2761.
Author Contributions: Drs Patel and Richmon had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Patel, Wax, Rosenthal, Sweeny, Militsakh, Califano, Richmon.
Acquisition of data: Moore, Wax, Sweeny, Militsakh, Lin, Hasney, Butcher, Flohr, Arnaoutakis, Huddle, Richmon.
Analysis and interpretation of data: Patel, Wax, Califano, Richmon.
Drafting of the manuscript: Patel, Moore, Wax, Sweeny, Militsakh, Califano, Hasney, Flohr, Huddle, Richmon.
Critical revision of the manuscript for important intellectual content: Patel, Moore, Wax, Rosenthal, Militsakh, Califano, Lin, Butcher, Arnaoutakis, Huddle, Richmon.
Statistical analysis: Sweeny.
Administrative, technical, and material support: Moore, Wax, Califano, Hasney, Butcher, Arnaoutakis, Richmon.
Study supervision: Patel, Wax, Rosenthal, Militsakh, Califano, Richmon.
Conflict of Interest Disclosures: None reported.
Previous Presentation: This study was presented orally at the Combined Otolaryngology Spring Meetings, American Head & Neck Society; April 10, 2013; Orlando, Florida.