Journal of Otology & RhinologyISSN: 2324-8785

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.
bahis siteleri bahis siteleri bahis siteleri casino siteleri

Research Article, J Otol Rhinol Vol: 6 Issue: 4

Postlaryngectomy Pharyngocutaneous Fistula:Analysis of Possible Risk Factors

Bilici S*, Gokduman AR and Yigit O

Department of Otorhinolaryngology and Head and Neck Surgery, University of Health Sciences, Istanbul Training and Research Hospital, Turkey

*Corresponding Author : Suat Bilici, MD
Department of Otorhinolaryngology & Head and Neck Surgery, University of Health Sciences, Istanbul Training and Research Hospital
Abdurrahman Nafiz Gurman, Street No: 1, Postal code: 34098, Istanbul, Turkey
Tel: 0532 233 46 00
Fax: +902126320063
E-mail: [email protected]

Received: June 13, 2017 Accepted: June 28, 2017 Published: July 05, 2017

Citation: Bilici S, Gokduman AR, Yigit O (2017) Postlaryngectomy Pharyngocutaneous Fistula:Analysis of Possible Risk Factors. J Otol Rhinol 6:4. doi: 10.4172/2324-8785.1000321


Objective: The aim of this study was to analyse factors that predispose patients to pharyngocutaneous fistula (PCF) in total laryngectomy, with a focus on intraoperative primary tracheooesophageal fistula (TEF) with voice prosthesis and anti-reflux prophylaxis.
Methods: This retrospective cohort included 77 patients who underwent total laryngectomy (TL). Potential risk factors included intraoperative primary TEF with voice prosthesis, anti-reflux prophylaxis, previous radiotherapy (RT), diabetes mellitus, concurrently neck dissection and tumour stage.
Results: The global PCF rate was 46.3%. No statistically significant difference was noted between the fistula positive and negative groups for these parameters, except for hospitalisation time.
Conclusion: Anti-reflux prophylaxis was not significantly associated with the incidence of PCF. Primary TEF and voice prosthesis did not increase the incidence of PCF.

Keywords: Pharyngocutaneous fistula; Voice prosthesis; Anti-reflux prophylaxis; Trachea-oesophageal fistula


Pharyngocutaneous fistula (PCF) is communication between the pharyngo-oesophageal and cervical skin. It mostly occurs at the level of the surgical incision or around a tracheostomy and appears as accumulation of saliva [1]. PCF is the most frequent local complication after total laryngectomy (TL) and adversely affects the healing process of patients, prolongs hospitalisation time and delays oral feeding [2]. In addition, adjuvant therapies of the postoperative period are interrupted. The incidence of PCF has been reported to range from 3% to 65% [3].

Various risk factors associated with the occurrence of PCF have been studied, such as previous comorbid diseases (congestive cardiac and renal failure, chronic bronchitis, diabetes mellitus, malnutrition, anaemia and hypoalbuminaemia), previous radiotherapy (RT), chemotherapy (CT), and prior tracheostomy, but the reports are inconsistent. Possible risk factors for PCF include synchronous neck dissection, the type of pharyngeal closure, an advanced stage of the tumour, and in the postoperative period, nutritional disorders and infections [1,4]. Inadequate surgery and haematoma of the surgical wound are factors with a known association with increased incidence [5]. The incidence of gastro-oesophageal reflux (GER) is highly increased (82%) among laryngectomies, and GER may play a role in fistula formation after laryngectomy [6,7].

Many procedures have been utilised for voice rehabilitation after TL. Currently, primary tracheo-oesophageal fistula (TEF) and voice prosthesis are generally recognised globally for speech restoration after laryngectomy. The most important disadvantages of voice prostheses are deterioration of functions due to frequent breakdown of their valvular systems, which leads to endoprosthetic incontinence. Therefore, sound communication ceases, psychosocial problems occur and financial expenditures increase [8]. At the Department of Otorhinolaryngology of Istanbul Training and Research Hospital, we have performed intraoperative primary TEF with voice prosthesis (Provox 1, 2, Vega) (Atos Medical AB, Hörby, Sweden) in consenting patients since 2009. A recent literature review provides no information about the incidence of PCF in patients with vocal prostheses, so the aim of this present study was to analyse the effects of intraoperative primary TEF with voice prosthesis on the incidence of PCF.

Materials and Methods

This retrospective cohort study included the analysis of the charts of patients diagnosed with laryngeal squamous cell carcinoma (SCC) who underwent TL in the tertiary care hospital. This study was approved by the Research Ethics Committee of Istanbul Training and Research Hospital and was carried out from 2009 to 2017, involving 89 patients who underwent TL. Pre-treatment staging was classified according to the 2002 laryngeal cancer classification of the Union for International Cancer Control/American Joint Cancer Committee (UICC/AJCC). Final pathology reports were reviewed, and all patients with positive or close (2 mm or less) margins were excluded from the study (5 patients). A total of 7 patients with stage T 4b were also excluded from the study. In total, 77 patients (97.4% male) with a mean age of 61.4 ± 8.9 years were included the study. These patients were studied retrospectively for the occurrence of PCF. We also included salvage procedures after previous functional laryngeal surgeries. Table 1 shows the distribution of these procedures. The following were considered potential risks of PCF formation: the age and gender of patients, hospitalisation time, previous RT, concomitant neck dissection, tumour stage, the presence of diabetes mellitus, antireflux treatment and intraoperative TEF with voice prosthesis.

Primary radiotherapy n
Primary laryngeal cancer-  RT failure 5
Surgery without radiotherapy
LF Cordectomy 4
Total 9
Radiotherapy combination with surgery
LF Cordectomy+RT 1
Primary tongue cancer+RT 1
Total 4

Table 1: Contribution of RT and previous surgeries.

The surgical procedure was standardised among all surgeons. Primary TL included removal of the hyoid bone and infrahyoid muscles for oncological safety. Pharyngeal closure was always performed using 4/0 absorbable Vicryl (polyglactin 910) suture material in all patients, using a “T” shaped technique. A nasogastric feeding tube (18F) was placed intraoperatively in all the patients, and no patient was fed orally postoperatively. As a routine, a primary TEF was created during surgery to place the voice prosthesis for the patients who gave consent. The TEF was opened approximately 0.5 cm inferior to and along the midline fistula line of the tracheostomy. In particular, a short cricopharyngeal myotomy was performed and a tracheostomy was constructed by suturing the skin flap as far back as possible to the posterolateral tracheal cartilage.

Patients who could not provide consent were admitted for voice prostheses after a minimum of 6 months of operative time. All the patients underwent speech exercises. Two suction drains were used and generally removed when producing less than 10 cc in 24 hours. Postoperative infection was prevented with antibiotic therapy with cefazolin and ornidazole, which was generally started intraoperatively and continued for the following 10 days. Skin sutures were routinely removed on postoperative day 8. Enteral feeding through the NGT was started on the second postoperative day, and oral feeding was started on postoperative day 10 if no signs of fistula were evident. A PCF was diagnosed by observation of saliva secretions on the tracheostomy and/or suture lines. Patients were given water to facilitate fistula visualisation. PCF was also evaluated as positive in patients who developed fistulas early after oral feeding. No anti-reflux prophylaxis protocol was administered to the patients.

In total, 36 patients were complicated by the onset of PCF. This group was matched and compared with a control group of 41 patients without PCF. We investigated and compared the risk factors for PCF between these groups, and we statistically evaluated the intraoperative TEF with voice prosthesis in terms of the occurrence of PCF.

Statistical Analysis

The descriptive statistics included the mean, standard deviation, median, minimum, maximum frequency and ratio values of the data. The distribution of the variables was measured by the Kolmogorov- Smirnov test. Independent sample t tests and Mann-Whitney U-tests were used to analyse the quantitative data. The Chi-square test was used for analysis of qualitative data. A value of p<0.05 was accepted as significant. Statistical analysis was conducted using the SPSS 22.0 for Windows (SPSS/IBM, Chicago, IL) software package.


Overall, 77 laryngeal SCC patients who underwent TL were analysed. The global incidence of PCF was 46.8% (36 patients). Mean hospital stay was 25.0 days.

The outcomes concerning the analysed parameters are shown in Table 2. No statistically significant difference was noted between the fistula positive and negative groups for any parameters, except for hospitalisation time. The mean age of the patients with and without PCF was 62.3 and 60.5 years, respectively. The fistula positive group had a mean length of hospitalisation of 36 days versus 15.4 days for the non-PCF group (p=0.000). Concomitant neck dissection, diabetes mellitus, previous RT and previous laryngeal surgeries did not statistically increase the risk of PCF. The majority of patients were at stage III and IV a (65 patients). Patients diagnosed with advanced stage tumours were not at increased risk of PCF (p=0.925). PCF occurrence was not statistically different among patients receiving anti-reflux prophylaxis, a proton pump inhibitor (PPI), ranitidine, PPI+Alginate and no prophylaxis.

Age Fistula (-) Fistula (+) p
Mean. ± S.D./n-% Median Mean. ± S.D./n-% Median
60.5   ± 7.8 60.0 62.3   ± 10.0 62.0 0.387 t
Hospitalisation time 15.4   ± 8.2 12.0 36.0   ± 25.2 29.0 0.000 m
Previous laryngeal surgery (-) 31   75.6%   33   91.7%   0.061
(+) 10   24.4%   3   8.3%  
Diabetes mellitus (-) 38   92.7%   32   88.9%   0.563
(+) 3   7.3%   4   11.1%  
Previous RT (-) 37   90.2%   31   86.1%   0.573
(+) 4   9.8%   5   13.9%  
Anti-reflux prophylaxis Lansoprozol IV 15   36.6%   17   47.2%   0.821
Ranitidine IV 7   17.1%   5   13.9%  
Lansoprozol IV+Alginat 9   22.0%   7   19.4%  
No prophylaxis 10   24.4%   7   19.4%  
Neck dissection (-) 6   14.6%   4   11.1%   0.646
(+) 35   85.4%   32   88.9%  
Tumour stage (T) II 7   17.1%   5   13.9%   0.925  
III 15   36.6%   14   38.9%  
IV 19   46.3%   17   47.2%  
Voice rehabilitation Oesophageal speech 17   41.5%   10   27.8%   0.209
Primary TEF with VP 24   58.5%   26   72.2%  
tt test / mMann-whitney u test / X² Ki-kare test

Table 2: Correlation between variables and incidence of pharyngocutaneous fistula.

Fifty patients had undergone TEF with intraoperative voice prosthesis. No statistically significant difference was noted between the group with voice prosthesis and the group without prosthesis in terms of the risk of developing PCF (p=0.209)

In the group with PCF, 16 subjects were quickly treated conservatively with neck compression dressing, antibiotics and local hygiene in 5 days, on average. Conversely, 20 patients needed further surgery. Of these, 8 cases required only surgery revision and primary closure, while 5 patients required secondary surgical reconstruction using a pectoralis major myocutaneous flap transposition. The remaining 7 patients underwent surgical repair using a sternocleidomastoid muscle flap transposition.


The recent evolution of non-surgical organ preservation protocols and functional laryngeal surgeries for treatment of laryngeal squamous cell carcinoma (SCC) has led to the frequent performance of TL as a salvage procedure. Timely restoration of laryngeal functions, such as speech and swallowing, is essential, so PCF is an important complication. It usually occurs shortly after total laryngectomy and can carry potentially devastating sequelae, including retardation of the swallowing function, in addition to a delay in the onset of adjuvant RT when indicated. Conversely, a loss of voice is the most disturbing consequence of laryngectomy for most patients. Therefore, the decision to perform a laryngectomy is always directly related to the likelihood of voice rehabilitation.

In the present study, risk factors were examined that could play a role in the occurrence of PCF after total laryngectomy. In addition, intraoperative primary TEF with voice prosthesis was analysed for its effects on the incidence of PCF.

Previous RT has a known effect on the occurrence of PCF, but no consensus exists in the literature. Some authors have reported no statistically significant difference [9-13], whereas others found relevant relations between PCF formation and previous RT [14,15]. Virtaniemi et al. [10] reported that PCF occurred earlier in previously irradiated patients. A recent meta-analysis by Paydarfar et al. [16] reported a greater severity and duration of fistula in patients who underwent preoperative RT than in patients who did not. Another study indicated that RT dosages greater than 5000 cGy increased the PCF incidence rates [17], although this variable showed no statistically significant differences in our samples (p=0.573).

Some authors found that simultaneous laryngectomy and neck dissection increased the risk of fistula formation [10,18]. By contrast, we found no significant efficacy in the occurrence of fistula following neck dissection, which we performed concurrently with TL (p=0.646). Similarly, previous laryngeal cancer surgeries had no statistically significant effect on the occurrence of PCF, but the rate of fistula-negative operations was approximately triple that of fistulapositive ones.

Pharyngeal closure can be performed using different methods. We used 4/0 Vicryl stitches and a T-shaped manual closure for all patients to close the pharynx. Manual stitches take longer and may result in more irritation and necrosis in the pharyngeal mucosa, as well as contamination of the surgery site with saliva. In addition, the junction regions in T-shaped pharyngeal closures remain as weak spots [5]. Therefore, the PCF incidence may be high. In the last 25–30 years, mechanical closure with a pharyngeal stapler has increased in popularity, and these weak spots do not arise when a stapler is used [19]. The meta-analysis by Shah et al. [20] reported that stapling reduced the rate of PCF.

X-ray swallowing studies after gastrografin and methylene blue are commonly used to identify fistulas [5,7,21]. However, some authors have reported that scintigraphy analysis is a more objective and non-invasive method, not only for identification of the presence of PCF, but also for close monitoring of the fistula localisation and spontaneous closure [22]. We preferred water and methylene blue to facilitate fistula visualisation in our series.

Four different methods of voice rehabilitation have been used: external electrical devices (electro larynx), oesophageal speech, the creation of a TEF with insertion of voice prosthesis and the artificial larynx [23,24]. TEF and voice prostheses have been preferred during the past three decades because of the quality and clarity of the rendered voice, the ease of use and the high patient compliance. Therefore, this method is generally recognised as a routine procedure for speech restoration after TL. The use of the prosthesis may result in many minor complications, as saliva leakage, granulation tissue and fungal colonisation can frequently develop around the prosthesis [25].

TEF can involve primary (intraoperative) or secondary (postoperative) TEPs, which have been shown to be equally successful in voice rehabilitation (success rates of 65–85% and 69–83%, respectively). Aires et al. [5] implanted a TEF and voice prosthesis three months after TL. A secondary TEF was not preferred because of its additional risk and morbidity for the patient. However, in patients who did not give consent, voice prosthesis was placed at the end of the additional treatment (RT, CT) six months after the surgery. To the best of our knowledge, this is the first study to investigate whether intraoperative primary TEF with voice prosthesis is a risk factor that may affect the incidence of PCF. In our series, primary voice prosthesis was put in place in about 2/3 of the patients. No statistically significant difference was noted in the rate of PCF between the patients who had the voice prosthesis placed or not placed (Table 2). We believe that these results indicate that placement of the voice prosthesis during the operation is an appropriate choice.

Cocuzza et al. [26] showed that laryngectomies have a high incidence of pathologic gastric reflux and that the presence of pathologic gastric reflux is correlated with a partial or total voice prosthesis failure. In another study, Seikaly et al. [7] reported that GER in the postoperative period may predispose the patient to PCF formation. For these reasons, prevention of GER is important. Many anti-reflux medications, such as PPIs and H2 receptor antagonists, prevent GER by inhibiting gastric acid secretion [7,27], whereas metoclopramide increases the lower oesophageal sphincter pressure. Alginic acid produces a viscous foam barrier to prevent reflux of stomach contents into the oesophagus, and may also provide a protective lining for the oesophagus when this reflux occurs [28]. Seikaly et al. [7] reported that administration of ranitidine IV and metoclopramide IV for 7 days decreased the postoperative PCF incidence. In another study, Stephenson et al. [29] reported that prophylactic use of omeprazole intravenously for 14 days caused a statistically significant decrease in the incidence of PCF. In our study, the groups randomly received PPI, ranitidine, PPI + alginate or no prophylaxis. However, we did not determine any significant relationship between PCF incidence and anti-reflux prophylaxis.

A disagreement exists in the literature regarding the relationship between the tumour stage and PCF incidence. Aires et al. [5] reported that the T-stage was a statistically significant factor for the occurrence of PCF in their series of 94 laryngeal SCC patients. On the contrary, Busoni et al. [21] found no significant relationship between their fistula group and control group in terms of stage T2–3 and T4. This variable also showed no statistically significant difference in our study (p=0.925).

The reasons why our fistula rate was somewhat high (46.8%) may reflect the poor general health and inadequate nutritional status of our patients, which are factors known to have a negative effect on postoperative tissue healing. The social and economic status in our country means that the general health status of patients is usually poor and nutritional status insufficient. Therefore, tissue healing in the postoperative period might be adversely affected by these risk factors, thereby increasing the PCF incidence.

This study had some limitations that mainly reflected its retrospective setting. The number of patients with PCF was relatively small (36 cases), but it was representative of a single institution’s experience over the last 8 years. Larger sample sizes involving a greater numbers of patients and TEFs with voice prostheses are required to confirm our study results.


Anti-reflux prophylaxis did not significantly change the incidence of PCF. Primary TEF and voice prosthesis did not increase the incidence of PCF. When assessing which primary voice prosthesis to use for successful treatment, clinicians should be aware that the choice of an intraoperative TEF with voice prosthesis might not increase the formation of PCF.


Track Your Manuscript

Media Partners