Fasciocutaneous and jejunal free flaps for circumferential hypopharyngeal defect reconstruction: A 22‐year multicenter cohort study

It remains unclear whether a tubed fasciocutaneous or jejunal free flap (FCFF and JFF) is preferable for reconstruction of circumferential pharyngolaryngoesophageal defects.


| INTRODUCTION
Free flap reconstruction of circumferential pharyngolaryngoesophageal (CPLO) defects after resection of locally advanced tumors is a complex procedure, which requires experienced multidisciplinary surgical teams.In the first part of the 20th century, reconstructions were performed using regional flaps and gastric pull-up procedures. 1,2The advent of reconstructive microsurgery in the latter half of the century has led to the implementation of free tissue transfer to improve surgical and functional outcomes.][3][4] Nowadays, both FCFF and jejunal free flap (JFF) techniques are commonly performed for reconstruction of CPLO defects throughout the world.FCFF reconstructions can be divided into two main reconstructive procedures: a radial forearm free flap (RFFF) or anterolateral thigh free flap (ALTFF).Each technique has procedurespecific advantages and drawbacks.Recent reviews have indicated that JFF reconstructions have lower fistula and stricture rates, while having similar functional outcomes compared to FCFF reconstructions. 5,6However, reported disadvantages of JFF reconstructions include higher donor-site morbidity and higher flap failure rates, leading to additional interventions and prolonged hospital stay. 1 Although patient-related factors play an important role, the reconstructive modality of choice also depends on institutional preferences and the surgeons' expertise. 7,86][7][8][9] The pooling of heterogeneous data from studies with a single reconstruction group in recent systematic reviews and meta-analyses introduces substantial bias, which affects the validity of the conclusions.Moreover, most studies have only reported short-term outcomes without sufficient data on long-term complications.Considering these factors, up to now, no international consensus on which procedure offers superior outcomes has been reached.
The present retrospective study aimed to compare both recipient and donor-site complications between CPLO defect reconstructions using FCFFs (ALTFFs and RFFFs) and JFFs.In addition, the literature was reviewed for other comparative studies on CPLO defect reconstruction.

| Study design
A multicenter retrospective cohort study was performed, including all consecutive patients who underwent a CPLO reconstruction between January 1st 2000 and December 31st 2022 at two large tertiary referral centers (the Erasmus MC Cancer Institute, University Medical Center Rotterdam, The Netherlands, and the Leiden University Medical Center, The Netherlands).The Erasmus MC Cancer Institute accommodates the largest head and neck oncology center in The Netherlands, which performs approximately 40 total laryngectomies each year, of which on average eight patients with circumferential defects.
Patients were eligible for inclusion if there was a circumferential defect after pharyngolaryngoesophageal resection and the reconstruction was performed using a free jejunal or tubed fasciocutaneous flap.Reconstructions of partial defects, gastric pull-up reconstructions, and reconstructions using regional pedicled flaps were excluded.
The indication for free flap reconstruction and flap choice was set at a weekly multidisciplinary head and neck cancer meeting.A factor that affected flap choice was the availability of a gastrointestinal or plastic surgeon for the planned surgery date, often independent of patient-related factors.Previous abdominal surgery was not an absolute contra-indication for the use of a JFF.The electronic medical records of all patients were reviewed.
Multiple reconstructive surgeons performed the reconstructions, most of which used a trapezoid-like spatulated design to tube and inset the flap.A lens-type flap was occasionally used.In order to minimize flap ischemia time, some surgeons chose to tube the flap at the donor-site prior to transection of the pedicle.If available, fascia was used to reinforce the suture line by incorporating a secondary closure layer.For JFF reconstructions, performed by gastrointestinal surgeons, the flap was harvested following laparotomy by means of a stapler device.Subsequently, the vascular anastomoses were performed, after which the flap was inset.
The medical ethics committee of the Erasmus MC approved the study prior to the start of the data collection (MEC-2021-0467).

| Outcomes
The primary outcomes of interest were fistula and stricture formation.A fistula was defined as any abnormal connection between skin or trachea and the reconstructed CPLO segment requiring surgical revision.A stricture was defined as any abnormal narrowing of the free flap anastomosis requiring (outpatient) dilation or surgical revision.
Secondary outcomes were total flap failure, other recipient-site complications, donor-site complications, systemic complications, length of hospital stay, the start of oral intake and speech.Total flap failure was defined as failure of the flap due to microvascular or intraoperative failure.To account for the success of the reconstruction, our study also assessed reconstructive failure, defined as the delayed need to salvage the entire reconstruction due to persistent (wound healing) complications or afunctionality of the neopharynx during follow-up.
Outcomes following flap failure were analyzed separately.If an additional pectoralis major flap was used to close the neck defect, its associated complications were not registered as donor-site complications related to the free flap harvest.Anastomotic suture line leakage was defined as any abnormal connection between the skin or trachea and the reconstructed hypopharyngeal segment that spontaneously resolved following prolonged tube feeding, antibiotic therapy, and/or scopolamine therapy.The severity of all complications was graded using the international Clavien-Dindo classification (Figure 1). 10 A literature review of similar studies was performed by including solely comparative studies identified from a pool of four independent systematic reviews on CPLO reconstructions, covering the period 1994-2022. 1,5,6,13

| Statistical analysis
Numeric data were analyzed using Student's t tests or Mann-Whitney U tests.Categorical data were analyzed using chi-square and Fisher's exact tests for binary variables.Binary logistic regression analyses were performed for the primary outcome variables to identify potential risk factors, based on the following predetermined variables: age at time of surgery, active smoking at time of surgery, preoperative radiotherapy, perioperative albumin level, T-stage, and presence of comorbidity.All statistical analyses were performed using SPSS version 28 (IBM Corp., Armonk, NY).Two-sided p-values <0.05 were considered statistically significant.

| RESULTS
Between January 1st 2000 and December 1st 2022, 112 patients underwent a total laryngectomy, pharyngectomy, or laryngopharyngectomy leading to a CPLO-defect at our institutions (Erasmus MC Cancer Institute n = 101, Leiden University Medical Center n = 11).All patients received a single-stage CPLO-segment reconstruction using free tissue transfer, consisting of 35 FCFFs and 77 JFFs.The FCFF group comprised the following tubed flaps: ALTFFs (n = 25, including tensor fascia latae myocutaneous free flaps [n = 2] due to the absence of available perforators) and RFFFs (n = 10).The baseline characteristics of all cases are shown in Table 1.Median follow-up time after surgery was 22.1 months (IQR, 10.5-57.1 months), which did not differ significantly between the FCFF and JFF groups.There were no significant differences in primary endpoints between the institutions in the outcomes of FCFF reconstructions.Cricopharyngotomy (n = 1), laryngofissure (n = 1), marginal mandibular resection (n = 1), carotid endarterectomy (n = 1), hemithyroidectomy (n = 1), and cervical spine surgery (n = 4, e.g., anterior cervical discectomy and fusion, or unknown approach of the cervical spine surgery).

| Primary outcomes
of RFFFs (5 of 10, 50%) differed significantly from JFF reconstructions ( p = 0.008), whereas ALTFF reconstructions (7 of 25, 28%) did not.Median time to fistula closure did not differ significantly between FCFF and JFF reconstructions (160 days, IQR 78-179 days).Of all 21 fistulas which required additional surgery, only 11 (52%) could be successfully closed permanently (FCFF 7 of 12, 58%; JFF 4 of 9, 44%).Anastomotic suture-line leakage, which spontaneously resolved following tube feeding and/or antibiotic therapy occurred in 23% of all patients and did not differ significantly between the groups.The relative risk of fistula formation across all patients was not significantly lower if a pectoralis major flap was used to cover the neck defect (RR 0.8, 95% CI 0.3-2.1).
Strictures occurred in 39 of 112 patients (35%) including 17 of 35 patients (49%) in the FCFF group and 22 of 77 patients (29%) in the JFF group ( p = 0.04).Strictures occurred at a median time of 150 days (IQR, 99-310 days) after surgery.The rates of adjuvant radiotherapy (FCFF 29%, JFF 31%) and chemotherapy (FCFF 9%, JFF 1%) were not significantly different between both groups.Post hoc analysis of stricture cases did not demonstrate any significant associations between patients who had received adjuvant chemotherapy or radiotherapy and those who had not.For both primary outcomes, there was no observed effect modification after correcting for significant baseline differences between the groups (Tables 1 and S1, Supporting Information).

| Secondary outcomes
All recipient-site complications are shown in Table 2. Total flap failure did not differ significantly between FCFF (2 of 35 patients; 6%) and JFF (5 of 77 patients; 7%) reconstructions.The two failed FCFF reconstructions were salvaged by tubed RFFFs (n = 2).Four failed JFF failures were managed using a gastric pull-up (n = 4).In one case after flap removal, no further attempt at reconstructing the conduit was performed, with the defect being closed caudally and cranially including pectoralis major flap coverage (n = 1).d Total number of patients is lower than the total number of complications, as some patients had multiple complications.
In the long-term there was a higher proportion of reconstructive failures in the FCFF (4 of 35 patients; 11%) compared to JFF group (2 of 77 patients; 3%).However, this difference was not statistically significant.All four FCFF reconstructive failure cases were ALTFFs, which were ultimately unsuccessful due to persistent fistulas (n = 2) or severe neopharyngeal stenosis (n = 2).The flaps were salvaged using a pectoralis major flap (n = 1), JFF reconstruction (n = 1), and gastric pull-up (n = 2).The two reconstructive failures in the JFF group were due to persistent fistulas (n = 2), which were managed using a tubed ALTFF (n = 2).
In total, six donor-site complications occurred in six patients in the FCFF group and 35 in 19 patients in the JFF group.There was a significantly higher rate of severe donor-site complications in the JFF (18%) versus FCFF (0%; p = group (Table 3).Multivariable analysis identified low perioperative albumin levels (p = 0.018) and prior cerebrovascular disease (7% vs. 31%, p = 0.03), as significant risk factors for severe adverse donor-site events in the JFF group.Median serum albumin levels were comparable between groups (FCFF 28 g/L, IQR 25-32; JFF 27 g/L, IQR 23-31).No other significant risk factors for fistulas, strictures, flap failures, or severe adverse donor-site events were identified.
Severe cardiopulmonary complications requiring ICU admittance occurred in 8.6% (3 of 35) of FCFF and 6.6% (5 of 77) of JFF patients, respectively, which did not differ significantly.One patient in the JFF group had a postoperative CVA (1%).There was a higher mortality rate in the JFF group (6.5%) than the FCFF group (2.9%), but this difference was not statistically significant.
Median length of stay was significantly longer for the FCFF group (26 days; IQR, 17-40 days) than for the JFF group (18 days; IQR, 15-26 days; p = 0.03).Patients resumed oral intake significantly later after FCFF reconstruction than following JFF reconstruction with a median of 26 days (IQR, 14-57 days) versus 14 days (IQR, 10-19 days), respectively ( p = 0.003).The proportion of patients without flap or reconstructive failure that could not resume oral intake after surgery was higher following FCFF (4 of 29 patients, 14%) than JFF reconstructions (3 of 70 patients, 4%), although this difference was not statistically significant.Speech data was only available for 86 patients (19 FCFF and 67 JFF patients) who had a successful primary reconstruction.There were no significant differences in speech outcomes between the reconstructive groups (Table 4).

| Previous comparative studies
1,12 Unfortunately, these studies do not provide unequivocal data regarding which flap yields better postoperative surgical and functional outcomes with the least donor-site complications.Additional outcome data regarding comorbidity and perioperative characteristics can be found in Data S1.

| DISCUSSION
The current study compared the outcomes of circumferential pharyngolaryngoesophageal defect reconstructions using FCFFs and JFFs.To our knowledge, this is the first study that evaluated the occurrence and severity all short-and long-term postoperative complications using the Clavien-Dindo classification.Our study illustrates that JFF reconstructions have significantly lower fistula and stricture rates than FCFF reconstructions, yet carry a substantial risk of severe donor-site morbidity, potentially leading to mortality.Furthermore, median length of hospital stay was significantly longer for FCFF patients and oral intake was resumed significantly later compared with patients in the JFF group.Other studies which directly compared outcomes of JFF and FCFF reconstructions are limited and provided inconclusive outcomes. 1,5- 9Most other studies on CPLO defect reconstructions have only reported the outcomes of a single surgical technique with highly variable complication rates, which may be explained due to the considerable heterogeneity in surgical techniques, patient selection, and different definitions of outcomes reported. 1,5he fistula rates found in our study are in line with two recent independent meta-analyses by Koh et al. and Costantino et al., which both showed a statistically significant lower fistula rate for JFF compared with FCFF reconstructions. 5,6Another meta-analysis by Bouhadana et al. reported a significant difference between JFFs and RFFFs in favor of JFFs, but not between JFFs and ALTFFs. 13Murray et al.only showed minor differences between FCFF and JFF reconstructions, with comparable fistula (13% vs. 9%, respectively) and stricture rates (16% vs. 11%, respectively). 1 Nonetheless, a major flaw of these reviews is the comparison of heterogeneous and biased data from studies with only a single reconstruction group, which substantially decreases the validity of these pooled outcomes.
The relatively higher rate of fistulas and strictures in FCFF reconstructions has been hypothesized to be caused by various factors, including the tissue mismatch of skin-to-mucosa, the use of longitudinal suture lines, and the requirement of T-junctions. 5,8,14Additionally, it has been reported that fistula formation predisposes towards stricture formation. 9In order to reduce fistula and stricture rates in FCFFs, various flap modifications have been proposed in the literature, which include the use of a Z-plasty or spatulated anastomosis at both the cranial and caudal flap anastomoses, posterior positioning of the suture line, second layer closure using fascia, chimeric flaps (e.g., addition of vastus lateralis muscle), and the use of salivary bypass tubes. 1,5After implementation of these modifications, significant improvements in fistula and stricture rates have been reported in the literature. 1 In contrast to potentially lower rates of recipient-site complications, the risk of severe donor-site complications in JFF reconstructions following median laparotomy is a major drawback.Donor-site complications in JFF reconstructions have been reported at a rate of 4.4% (16 of 368 cases) in the largest case series of JFF reconstructions. 15Shangold et al. found a rate of 5.8% (33/555 cases) in their review. 16Donor-site morbidity following JFF harvest in our series was significantly higher with 18% of patients who required surgical intervention or ICU admission.This may be related to a better registry of complications using the Clavien-Dindo classification, including sequelae of JFF harvest such as incisional hernias in the long-term, which frequently were not included in other studies, or to the comorbidities and frailty of the included patients.If a JJF is chosen for a specific patient, donor-site morbidity may be limited by laparoscopic harvest and careful patient selection. 179][20] This indicator for malnutrition and frailty can be used as a proxy for patients at risk of adverse events and should be taken into account during preoperative planning for pharyngeal reconstructions.
Donor-site complications following FCFFs are predominantly aesthetic or related to paresthesia/dysesthesia (ALTFF: 24%, RFFF: 27%), whereas wound healing problems are observed infrequently (<5%). 5,21,22While FCFF donor-site complications do not commonly necessitate surgical intervention-as was also observed in our study-it is important to take the relative donor-site morbidity profile of each flap into account.4][25][26] Reported advantages of ALTFFs include possibility of primary wound closure, availability of a large skin paddle, a long vascular pedicle and the presence of robust fascia to reinforce the suture layer. 2,8owever, in obese patients, a bulky ALTFF may hinder the ability to properly tube the flap.Recent advances such as a suprafascial ALTFF harvest may make this reconstruction available to more patients. 27igher flap failure rates have previously been reported for JFF (5%-6%) than for FCFF reconstructions (2.0%-2.8%),although our study did not observe this. 1,13he inherent vulnerability of the enteric mucosa to ischemia may predispose to a higher likelihood of flap failure in JFFs compared to more tolerant fasciocutaneous tissues. 2,28However, there was a higher proportion of late reconstructive failures in the FCFF (4 of 35 patients; 11%) compared to the JFF group (3 of 77 patients; 3%), due to a higher rate of recipient-site complications following FCFF reconstructions.Moreover, perioperative mortality has been reported to be lower after FCFF reconstructions (0%) compared to JFFs (3.3%). 13In our study we observed a higher mortality rate for the JFF group (6.5%) than the FCFF group (2.9%).However, this difference was not statistically significant in our study.
Functional outcomes (speech and oral intake) have historically been reported to be worse following JFF reconstructions compared to FCFF reconstructions. 1 JFF reconstructions have frequently been reported to produce a low-pitched wet voice caused by mucus secretions from the jejunal mucosa.Moreover, asynchronous peristaltic contractions and sensory or motoric disturbances may cause swallowing difficulties following JFF reconstructions. 1 In contrast, the advantages of FCFFs include their firmness and inherent nature to provide a stationary conduit without peristalsis.Nonetheless, our study did not identify a significant difference between patients with an FCFF or JFF reconstruction in their ability to resume oral intake, nor regarding their respective speech outcomes.In line with our study, more favorable functional outcomes after JFF reconstructions comparable to FCFF reconstructions have been reported recently, which might be the result of new insights and modifications. 5 patient-tailored approach is essential for these complex microvascular CPLO-defect reconstructions in an effort to limit morbidity.Creating a comprehensive treatment algorithm for this heterogeneous group of patients remains challenging because of the multifaceted nature of postoperative outcomes.Important considerations for flap choice in CPLO-defect reconstruction are shown in a general treatment algorithm in Figure 2.
There are limitations to this study.First, flap choice without a randomized protocol is inherent to a certain degree of confounding by indication.However, since choice of reconstruction technique was also dictated by logistic factors and to a lesser extent by patient specifics, confounding by indication may have occurred less in this study.Second, because of the infrequent nature of these complex procedures, the study size was limited.Third, no data regarding patient-reported outcomes on quality of life, swallowing or speech function were available.Finally, the heterogeneity of studies in the literature complicated the comparison of outcomes between the current study and previous studies.Future comparative (prospective) multicenter studies are required to determine which technique achieves the highest quality of life and functionality with minimal morbidity, using a core outcome set.

| CONCLUSIONS
This comparative multicenter study is the first to objectively report short and long-term outcomes after CPLOdefect reconstruction using FCFFs and JFFs, using the Clavien-Dindo severity classification system.In general, JFF reconstructions yield significantly lower rates of fistulas and strictures compared to FCFF reconstructions, although the former carry a significantly higher risk for severe, potentially lethal, adverse donor-site events.Given the group heterogeneity and complexity of associated risk factors, creating a comprehensive general treatment algorithm for this fragile population is currently not feasible.Flap choice should be made on an individual patient basis in order to limit potential morbidity and to offer the best quality of life.
Baseline characteristics according to reconstructive technique.
T A B L E 1 Abbreviations: AJCC, American Joint Committee on Cancer; LND, lymph node dissection; PMF, pectoralis major flap.a Statistically significant difference.b Hemithyroidectomy (n = 1).c T A B L E 2 Postoperative recipient-site complications per reconstructive treatment group by complication severity.
a CD grade I/II: conservative management, CD grade III: surgical management, CD grade IV: life-threatening complication requiring ICU management, and CD grade V: death.b Statistically significant difference.
T A B L E 3 Postoperative donor-site complications per reconstructive treatment group by complication severity.
a CD grade I/II: conservative management, CD grade III: surgical management, CD grade IV: life-threatening complication requiring ICU management, and CD grade V: death.No grade V complications were observed.b Total number of patients is lower than the total number of complications, as some patients had multiple complications.c Statistically significant difference.
T A B L E 4 Speech outcomes per reconstructive treatment group.Speech data was only available for 86 patients who had a successful primary reconstruction.Overview of previous comparative studies on CPLO defect free flap reconstruction.