- Email: [email protected]
Postlaryngectomy Pharyngocutaneous Fistula: Incidence, Predisposing Factors, and Therapy
Otolaryngology–Head and Neck Surgery (2005) 133, 689-694
ORIGINAL RESEARCH
Postlaryngectomy Pharyngocutaneous Fistula: Incidence, Predisposing Factors, and Therapy Jacopo Galli, MD, Eugenio De Corso, MD, Mariangela Volante, MD, Giovanni Almadori, MD, and Gaetano Paludetti, MD, Rome, Italy OBJECTIVE: The pharyngocutaneous fistula (PCF) is a serious complication after total laryngectomy, and its etiology is not well understood yet. The aim of our study was to evaluate predisposing factors, incidence, and management of this complication. STUDY DESIGN AND SETTING: This was a retrospective study of 268 patients who underwent total laryngectomy in our clinic (January 1990-December 2001). A number of factors potentially predisposing to PCF formation were evaluated. RESULTS: A PCF was observed in 16% of patients. Systemic diseases, previous radiotherapy, supraglottic origin of tumor, and concurrent radical neck dissection were significantly associated with PCF. Spontaneous closure was noted in 28 patients, whereas a surgical closure was necessary in 15 patients. CONCLUSIONS: In presence of a specific risk factor, PCF can be expected; nevertheless, its prevention remains very difficult. Moreover, given the high percentage of spontaneous closure, we suggest the “wait and see” approach for 28 days before proceeding with a surgical approach. EBM Rating: C © 2005 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved.
In the present study, we reviewed the clinical data of 268 patients treated with total laryngectomy in our clinic with the aim of evaluating the incidence, predisposing factors, and treatment of PCF.
MATERIALS AND METHODS
he pharyngocutaneous fistula (PCF) is a common and troublesome complication in the early postoperative period after total laryngectomy.1-4 Many factors have been indicated as predisposing to fistula formation5-8: preoperative radiotherapy, surgical technique, concurrent neck dissection, suture material used for pharyngeal reconstruction, surgeon’s ability, tumor stage, preoperative tracheostomy, early oral feeding, preoperative and postoperative hemoglobin levels, presence of systemic diseases, and nutritional status.9-11 However, to date no conclusive evidence has been gathered about the relative importance of these factors.
Between January 1990 and December 2001, 268 consecutive total laryngectomies for laryngeal or hypopharyngeal squamous cell carcinoma were performed at the Department of Otolaryngology of Policlinico Gemelli in Rome. The neoplasms were staged following the International Union Against Cancer (UICC) TNM system.12 All the lesions were squamous cell carcinomas with various degrees of differentiation. Seventy-eight patients (29.1%) had received previous radiotherapy; the remaining 190 (70.9%) underwent laryngectomy as their primary treatment. Of the total 268 patients, standard total laryngectomy was performed in 250 patients, and surgery was extended to the pharynx (pyriform sinus and/or base of the tongue) in 18 patients. The operative technique was consistent through the study period. The pharyngoesophageal T closure was performed in 2 layers. Until 1998, the first layer was a chromic catgut suture with interrupted stitches, and the second was a continuous chromic catgut suture. After 1998, only Vicryl suture material was used. A nasoesophageal feeding tube was inserted during the operation and was usually removed on the 10th postoperative day. A percutaneous endoscopic gastrostomy (PEG) was never performed, even in the patients with fistula. Radical
From the Institute of Otolaryngology, Università Cattolica del Sacro Cuore, Rome, Italy. Reprint requests: Dr Eugenio De Corso, Institute of Otolaryngology,
Università Cattolica del Sacro Cuore, Largo Gemelli n.8, 00168 Rome, Italy. E-mail address: [email protected].
T
0194-5998/$30.00 © 2005 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved. doi:10.1016/j.otohns.2005.07.025
690
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
Table 1 Results of TNM staging of 268 patients submitted to total laryngectomy in the Department of Otolaryngology of the Policlinico Gemelli, in Rome, Italy
N0 N1 N2
T2
T3
T4
128 (47.7%) 26 (9.7%) 2 (0.7%)
43 (16.5%) 5 (1.8%) 2 (0.7%)
45 (16.7%) 4 (1.4%) 13 (4.8%)
neck dissection was associated with the primary operation in 18 patients and functional neck dissection in 73 patients. The operations were performed by 3 different surgeons with a similar degree of experience, as evaluated on the basis of learning curve, and with similar years in practice and number of laryngectomies performed. If a fistula developed, the first treatment was conservative, consisting of medical treatment with antibiotics and anti-inflammatory drugs; suspension of oral feeding
with position of nasogastric tube or parenteral nutrition; and daily local medication, including drainage of fluids from the fistulous tract, removal of necrotic tissue, local cleaning with antibiotic solutions, and a compressive dressing. If the fistula did not close spontaneously within 28 days of primary surgery, an operative closure was usually performed. In our series, patients who developed PCF (fistula group; 43 patients) were compared with those without postlaryngectomy complications (nonfistula group; 225 patients). We evaluated a series of variables, including smoking and drinking habits, preoperative radiotherapy, surgical technique and suture materials, concurrent neck dissection, tumor stage, presence of systemic diseases, and the respective incidence of the variables in the 2 groups. Statistical comparison of categorical variables was performed with Pearson’s chi-square test, and a P value less than 0.05 was considered significant. Odds ratio (OR) and 95% confidence interval (95% CI) were obtained by multivariate logistic model performed by the SSPS for Windows version 9.0 statistical software.
Table 2 Relationship between PCF formation and investigated risk factors Variables Sex M F Median age Smoker No Yes Alcohol intake None ⬍1 L ⬎1 L Systemic disease Diabetes mellitus Cardiopathy Chronic bronchitis Neurological Previous radiotherapy Concurrent neck dissection Radical Functional Region Supraglottis Glottis Subglottis Transglottis Pharyngeal-larynx Type of laryngectomy Standard Extended Suture material used Vicryl Chromic catgut ns, not significant.
Fistula (n ⫽ 43)
Nonfistula (n ⫽ 225)
Significance (P)
39 (90.6%) 4 (9.4%) 62
212 (94.2%) 13 (5.8%) 64
ns ns
6 (13.9%) 37 (86.1%)
21 (9.3%) 204 (90.7%)
ns ns
11 18 14 20 4 9 5 2 18
54 109 62 68 20 21 19 8 60
(25.5%) (41.86%) (32.5%) (46.5%) (9.3%) (20.9%) (11.6%) (4.6%) (41.8%)
6 (13.9%) 11 (25.5%) 12 23 2 3 3
(27.9%) (53.4%) (4.6%) (6.9%) (6.9%)
(24%) (48.4%) (27.5%) (30.2%) (8.8%) (9.3%) (8.4%) (3.5%) (26.6%)
ns ns ns ⬍0.05 ns ⬍0.05 ns ns ⬍0.05
12 (5.3%) 62 (27.5%)
⬍0.05 ns
35 156 11 8 15
(15.5%) (69.3%) (4.8%) (3.5%) (6.6%)
⬍0.05 ns ns ns ns
40 (93%) 3 (7%)
210 (93.3%) 15 (6.6%)
ns ns
21 (48.8%) 22 (52.2%)
109 (48.4%) 116 (51.6%)
ns ns
Galli et al
Postlaryngectomy Pharyngocutaneous Fistula: . . .
691
4 weeks since primary surgery (wait-and-see period), we observed only a reduction of the size of PCF; therefore a surgical closure was required. In all cases we performed a direct suture of the pharyngeal mucosa. The choice of the surgical technique was dictated by the small diameter of the fistulas. No patient presented dysphagia due to fibrous stenosis, after conservative or surgical treatment. Radiotherapy had been performed in 78 out of 268 patients, in a period ranging from 6 months to 20 years before laryngectomy, with a dosage varying from 42 to 70 Gy. Among these cases, 18 developed PCF (n ⫽ 3, 1-2 cm; n ⫽ 15, 2-3 cm). We did not find differences in dosage of radiation in these groups with respect to patients who received preoperative radiotherapy and did not develop fistula after total laryngectomy (n ⫽ 60). Nevertheless, the small number of patients doesn’t allow us to draw a meaningful inference. The relationships between PCF formation and investigated parameters are summarized in Table 2. In our series, the presence of systemic diseases, previous radiotherapy, the origin of the tumor from the supraglottic region, and the concurrent radical neck dissection were all significantly associated with the development of the fistula (46.5% vs 30.2%, 41.8% vs 26.6%, 27.9% vs 15.5%, and 13.9% vs 5.3%, respectively; Fig 1). Conversely, there were no statistically significant differences between the fistula and the nonfistula groups with regard to age, sex, smoking, drinking habits, concurrent functional neck dissection, and type of total laryngectomy (standard or extended to pharyngeal mucosa). Furthermore, differences concerning the suture material and the surgeon’s ability did not affect the outcome of the 2 groups. Finally, in our series, T stage was not significantly associated with fistula formation (Table 3). The OR and the 95% CI obtained from a multivariate logistic model including terms for cigarette smoking, alcohol intake, previous radiotherapy and neck dissection (radical and functional), and associated systemic disease (diabetes mellitus, cardiopathy both as congestive heart failure and ischemic heart disease, chronic bronchitis, neurological disease) are shown in Table 4. We observed a significant correlation between PCFs and the supraglottic tumor site (OR: 1.5; 95% CI: 0.37-6.17), preoperative radiotherapy (OR: 1.94; 95% CI: 0.91-4.14), concurrent radical neck
Figure 1 Significant differences in incidence of systemic disease (SD), previous radiotherapy (PR), supraglottic carcinoma (SGC), and radical neck dissection (RND) in the fistula and nonfistula groups.
RESULTS The patient population consisted of 17 women and 251 men, ranging in age from 22 to 88 years (median age, 64 years). The majority of the patients (241, 89.9%) were smokers, 28.3% of the patients were classified as heavy drinkers (1 liter of wine per day or more), and 47.4% as moderate drinkers (less than 1 liter per day). Eighty-eight patients (32.8%) had a chronic systemic disease such as chronic congestive heart failure or ischemic heart disease (11.2%), diabetes mellitus (8.9%), chronic bronchitis (8.9%), or neurological disease (3.8%). The primary tumor was supraglottic in 47 patients (17.5%), glottic in 179 (66.8%), subglottic in 13 (4.9%), transglottic in 11 (4.1%), and pharyngolaryngeal in 18 (6.7%). The TNM staging of the patient population is shown in Table 1. A PCF developed in 43 out of 268 patients (16%), within an average of 10 days from surgery. In 28 out of 43 patients (65.1%), fistula size ranged between 1 cm and 2 cm. Applying conservative management, we observed spontaneous closure in all patients, within an average of 20 days from the appearance of the PCF (range, 5-28 days). In 15 out of 43 patients (34.9%), all submitted to previous radiotherapy, the diameter of the PCF ranged between 2 cm and 3 cm; after
Table 3 TNM staging in fistula group (n ⴝ 43) and nonfistula group (n ⴝ 225) T2
N0 N1 N2
T3
T4
F
Non F
F
Non F
F
Non F
22 (51%) 2 (4.6%) 2 (4.6%)
106 (47%) 24 (10.6%) 0
10 (23%) 2 (4.6%) 0
33 (14%) 3 (1.3%) 2 (0.8%)
3 (6.9%) 0 2 (4.6%)
42 (18.6%) 4 (1.7%) 11 (4.8%)
F, fistula group; Non F, nonfistula group.
692
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
Table 4 Estimates of fistula risk according to selected individual characteristic Variable
Regression coefficient
Standard error
OR
95% CI
0.26
0.39
1.30
0.59-2.84
0.05 1.20 0.08 0.73 0.06 0.09 0.66
0.42 0.98 1.50 0.81 1.51 1.52 0.38
1.05 3.34 0.55 1.92 0.51 0.38 1.94
0.45-2.43 0.18-6.11 0.21-7.77 0.62-18.52 0.25-9.7 0.14-10.3 0.91-4.14
1.78 0.61
0.63 0.43
5.96 1.48
1.70-15.87 0.78-4.34
0.4 0.4 0.5 0.6 0.5
0.7 0.6 0.7 0.9 0.6
1.5 0.6 0.3 1.4 0.8
0.37-6.17 0.1-2.4 0.1-1.5 0.3-0.5 0.3-0.9
Smoker Yes Alcohol intake Yes Systemic disease Diabetes mellitus Cardiopathy Chronic bronchitis Neurological Previous radiotherapy Concurrent neck dissection Radical Functional Region of neoplasm Supraglottic Glottic Sottolgottic Transglottic Pharyngeal-larynx OR, odds ratio; CI, confidence interval.
dissection (OR: 5. 96; 95% CI: 1.70-15.87), and systemic diseases (OR: 3.34; 95% CI: 0.18-6.11) in particular, and a high correlation with cardiopathy (OR: 1.92; 95% CI: 0.6218.52).
DISCUSSION After total laryngectomy, the appearance of a PCF results in a delayed outcome and rarely in carotid artery rupture.13 Generally PCF develops just above the tracheostoma, at the weakest point of the suture line of the pharyngeal mucosa, and it is always associated with surrounding tissue necrosis. Other localization of the salivary fistulas can occur higher in the neck, at the level of the junction between the pharyngeal mucosa and the base of the tongue. The presence of saliva results in surrounding tissue infection and microvenous thrombosis with a progressive tissue loss. The corresponding neck skin becomes tender and dark red in color. In previous studies a variable incidence of PCFs after total laryngectomy has been reported, ranging from 3% to 65%, with an average time of appearance of 10 days after surgery.1-4 Our observations of 43 fistulas in our group of 268 patients (16%) and a mean time of 10 (standard deviation ⫾4) days are therefore consistent with the mean values reported in the literature. Multivariate statistical analyses show a significant correlation between PCF and supraglottic tumor site, preoperative radiotherapy, and concurrent radical neck dissection. Several authors11,13 reported similar data, and particularly in agreement with our results, Virtaniemi et al1 have recently shown that supraglottic tumors have significantly
more fistulas than glottic ones. Cavalot et al,14 however, have considered preoperative radiotherapy as the crucial factor in fistula formation, especially when delivered at a full dose. In our series, confirming several previous studies,14,15 other factors, such as T stage, concurrent functional neck dissection, and extended total laryngectomy were not significantly associated with PCF. Nevertheless, in this study there are too few patients with ‘extended’ resections to pharyngeal mucosa, pyriform sinus, or base of the tongue to allow us to reach a meaningful conclusion. Indeed, in accordance with several authors,5,11 we believe that a closure under tension, caused by a resection of large amounts of pharyngeal mucosa, could increase the incidence of fistula formation. Furthermore, multivariate statistical analyses show a significant correlation between fistula formation and systemic diseases, particularly cardiopathy. Even though the exact mechanism of this correlation is not known at the moment, a reduced immune response and an impaired mechanism of wound healing, due to microangiosclerosis, could be the most likely factors involved.15 Similarly, Redaelli de Zinis et al16 have described ischemic cardiopathy as a specific risk factor for medical morbidity after total laryngectomy. Moreover, Hier et al9 and Lavelle and Maw2 have underlined the significance of postoperative hemoglobin level (⬍12.5 g/dL) and intraoperative transfusions because of their indication of patients’ general conditions. In the study reported by Horgan and Dedo,17 the surgeon’s ability played an important role in the formation of fistulas. In general, we agree with such observations. How-
Galli et al
Postlaryngectomy Pharyngocutaneous Fistula: . . .
ever, in our series we could rule out this variable because in our clinic, during the observation period, total laryngectomy treatment was performed by 3 surgeons with comparable surgical experience. Recently, Lavelle and Maw2 underlined the importance of the suture material in fistula development. A higher incidence of PCF was observed after the use of continuous Nylon suture. Verma et al18 reported a sharp drop in the rate of fistula formation after switching from chromic catgut to Vicryl. In our case records, the suture material did not seem to be a predisposing factor, because the use of different materials did not correlate with the incidence of PCF in the 2 groups. PCF diagnosis is exclusively clinical and is made during the postoperative course. Recently some authors have suggested that fever in the first postoperative day seems to play a prognostic role in PCF development.19 They found a high correlation between fever (⬎101.5°F) that developed in the first 48 hours and eventual fistula formation. Therefore, they proposed that most fistulas are the consequence of an inadequate closure, not of a wound breakdown, as generally believed. Furthermore, these same authors demonstrated that early intervention (before third postoperative day) decreased morbidity and duration of hospitalization. Concerning the treatment, some authors20 refrain from waiting for a spontaneous closure because this may take a long time (often more than 1 month), prolonging the patient’s stay in the hospital and resulting in swallowing difficulty due to fibrosis in the majority of cases. Our management of PCF is generally conservative because we have observed that chemical cauterization with a silver nitrate stick and medical treatment can sometimes contribute to a spontaneous closure of the fistula, as we observed, with a mean time of 20 days. In our series, small fistulas (1-2 cm), without previous exposure to radiotherapy, responded well to conservative management whereas large PCFs (2-3 cm), mainly related to previous radiotherapy, do not close spontaneously with a conservative approach. We suggest that in these cases it is not wise to wait any longer than a month to plan surgery. Nevertheless, with this approach we obtained a reduction of PCF size, so the small diameter reached did not require reconstruction with local flaps (deltopectoral and pectoralis major myocutaneous flaps) or free flaps.
CONCLUSIONS Our results confirm that PCF remains a troublesome complication of the early postoperative period after total laryngectomy.1,15,16 In the presence of specific risk factors, PCF can be expected; nevertheless, its prevention remains very difficult. Our data show that the presence of systemic diseases (with particular regard to cardiopathy), previous radiotherapy, radical neck dissection, and the origin of the
693
tumor from the supraglottic region can all be important predisposing factors, or at least underlying causes, in fistulas’ arising. On the other hand, we do not find significant differences between the 2 groups concerning suture material, surgical technique, and surgeon ability (evaluated on the basis of a learning curve), despite the fact that several authors2,18 have considered these as predisposing factors to PCF. It must be underlined that in the present study, surgery was performed by different but almost equally well-experienced surgeons. Therefore our results indirectly indicate that surgeon ability represents a significant variable impacting on the eventual development of PCF. Concerning the treatment, we believe that an observation period of 28 days after the primary surgical procedure is the best approach for PCF. Indeed, in small PCFs without previous radiotherapy exposure, a conservative approach permits spontaneous closure without stenosis or dysphagia. But in larger PCFs, radiotherapy related, it determines reduction of its size, reducing the need of reconstruction surgery with local or free flaps.
REFERENCES 1. Virtaniemi JA, Kumpulainen EJ, Hirvikoski PP, et al. The incidence and etiology of postlaryngectomy pharyngocutaneous fistulae. Head Neck 2001;23(1):29 –33. 2. Lavelle RJ, Maw AR. The aetiology of post-laryngectomy pharyngocutaneous fistulae. J Laryngol Otol 1972;86(8):785–93. 3. Thawley SE. Complications of combined radiation therapy and surgery for carcinoma of the larynx and inferior hypopharinx and inferior hypopharynx.. Laringoscope 1981;91:677–700. 4. Bressan K, Rasmussen H, Rasmussen PA. Pharyngocutaneous fistulae in totally laryngectomized patients. J Laryngol Otol 1974;88:835– 42. 5. Ikiz AO, Uca M, Guneri EA, et al. Pharyngocutaneous fistula and total laryngectomy: possible predisposing factors, with emphasis on pharyngeal myotomy. J Laryngol Otol 2000;114(10):768 –71. 6. Fradis M, Podoshin L, Ben David J. Post-laryngectomy pharyngocutaneous fistula—a still unresolved problem. J Laryngol Otol 1995; 109(3):221– 4. 7. Markou KD, Vlachtsis KC, Nikolaou AC, et al. Incidence and predisposing factors of pharyngocutaneous fistula formation after total laryngectomy: is there a relationship with tumor recurrence? Eur Arch Otorhinolaryngol 2004;261:61–7. 8. Parikh SR, Irish JC, Curran AJ, et al. Pharyngocutaneous fistulae in laryngectomy patients: the Toronto Hospital experience. J Otolaryngol 1998;27(3):136 – 40. 9. Hier M, Black MJ, Lafond G. Pharyngo-cutaneous fistulas after total laryngectomy: incidence, etiology and outcome analysis. J Otolaryngol 1993;22(3):164 – 6. 10. Johansen LV, Overgaard J, Elbrond O. Pharyngo-cutaneous fistulae after laryngectomy: influence of previous radiotherapy and prophylactic metronidazole. Cancer 1988;61(4):673– 8. 11. Natvig K, Boysen M, Tausjo J. Fistulae following laryngectomy in patients treated with irradiation. J Laryngol Otol 1993;107(12): 1136 –9. 12. Sobin LH, Wittekind C. UICC International Union Against Cancer: TNM classification of malignant tumours. 5th ed. New York: WileyLiss; 1997. p. 33-7. 13. McCombe AW, Jones AS. Radiotherapy and complications of laryngectomy. J Laryngol Otol 1993;107(2):130 –2.
694
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
14. Cavalot AL, Gervasio CF, Nazionale G, et al. Pharyngocutaneous fistula as a complication of total laryngectomy: review of the literature and analysis of case records. Otolaryngol Head Neck Surg 2000; 123(5):587–92. 15. Papazoglou G, Doundoulakis G, Terzakis G, et al. Pharyngocutaneous fistula after total laryngectomy: incidence, cause, and treatment. Ann Otol Rhinol Laryngol 1994;103(10):801–5. 16. Redaelli de Zinis LO, Ferrari L, Tomenzoli D, et al. Postlaryngectomy pharyngocutaneous fistula: incidence, predisposing factors, and therapy. Head Neck 1999;21(2):131– 8.
17. Horgan EC, Dedo HH. Prevention of major and minor fistula after laryngectomy. Laryngoscope 1979;89:250 – 60. 18. Verma A, Panda NK, Mehata S, et al. Post-laryngectomy complications and their mode of management: an analysis of 203 cases. Indian J Cancer 1989;26:247–54. 19. Friedman M, Venkatesan TK, Yakovlevmd A, et al. Early detection and treatment of post-operative pharyngocutaneous fistula. Otolaryngol Head and Neck Surg 1999;121:378 – 80. 20. Soylu L, Kiroglu M, Aydogan B, et al. Pharyngocutaneous fistula following laryngectomy. Head Neck 1998;20(1):22–5.
ORIGINAL RESEARCH
Postlaryngectomy Pharyngocutaneous Fistula: Incidence, Predisposing Factors, and Therapy Jacopo Galli, MD, Eugenio De Corso, MD, Mariangela Volante, MD, Giovanni Almadori, MD, and Gaetano Paludetti, MD, Rome, Italy OBJECTIVE: The pharyngocutaneous fistula (PCF) is a serious complication after total laryngectomy, and its etiology is not well understood yet. The aim of our study was to evaluate predisposing factors, incidence, and management of this complication. STUDY DESIGN AND SETTING: This was a retrospective study of 268 patients who underwent total laryngectomy in our clinic (January 1990-December 2001). A number of factors potentially predisposing to PCF formation were evaluated. RESULTS: A PCF was observed in 16% of patients. Systemic diseases, previous radiotherapy, supraglottic origin of tumor, and concurrent radical neck dissection were significantly associated with PCF. Spontaneous closure was noted in 28 patients, whereas a surgical closure was necessary in 15 patients. CONCLUSIONS: In presence of a specific risk factor, PCF can be expected; nevertheless, its prevention remains very difficult. Moreover, given the high percentage of spontaneous closure, we suggest the “wait and see” approach for 28 days before proceeding with a surgical approach. EBM Rating: C © 2005 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved.
In the present study, we reviewed the clinical data of 268 patients treated with total laryngectomy in our clinic with the aim of evaluating the incidence, predisposing factors, and treatment of PCF.
MATERIALS AND METHODS
he pharyngocutaneous fistula (PCF) is a common and troublesome complication in the early postoperative period after total laryngectomy.1-4 Many factors have been indicated as predisposing to fistula formation5-8: preoperative radiotherapy, surgical technique, concurrent neck dissection, suture material used for pharyngeal reconstruction, surgeon’s ability, tumor stage, preoperative tracheostomy, early oral feeding, preoperative and postoperative hemoglobin levels, presence of systemic diseases, and nutritional status.9-11 However, to date no conclusive evidence has been gathered about the relative importance of these factors.
Between January 1990 and December 2001, 268 consecutive total laryngectomies for laryngeal or hypopharyngeal squamous cell carcinoma were performed at the Department of Otolaryngology of Policlinico Gemelli in Rome. The neoplasms were staged following the International Union Against Cancer (UICC) TNM system.12 All the lesions were squamous cell carcinomas with various degrees of differentiation. Seventy-eight patients (29.1%) had received previous radiotherapy; the remaining 190 (70.9%) underwent laryngectomy as their primary treatment. Of the total 268 patients, standard total laryngectomy was performed in 250 patients, and surgery was extended to the pharynx (pyriform sinus and/or base of the tongue) in 18 patients. The operative technique was consistent through the study period. The pharyngoesophageal T closure was performed in 2 layers. Until 1998, the first layer was a chromic catgut suture with interrupted stitches, and the second was a continuous chromic catgut suture. After 1998, only Vicryl suture material was used. A nasoesophageal feeding tube was inserted during the operation and was usually removed on the 10th postoperative day. A percutaneous endoscopic gastrostomy (PEG) was never performed, even in the patients with fistula. Radical
From the Institute of Otolaryngology, Università Cattolica del Sacro Cuore, Rome, Italy. Reprint requests: Dr Eugenio De Corso, Institute of Otolaryngology,
Università Cattolica del Sacro Cuore, Largo Gemelli n.8, 00168 Rome, Italy. E-mail address: [email protected].
T
0194-5998/$30.00 © 2005 American Academy of Otolaryngology–Head and Neck Surgery Foundation, Inc. All rights reserved. doi:10.1016/j.otohns.2005.07.025
690
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
Table 1 Results of TNM staging of 268 patients submitted to total laryngectomy in the Department of Otolaryngology of the Policlinico Gemelli, in Rome, Italy
N0 N1 N2
T2
T3
T4
128 (47.7%) 26 (9.7%) 2 (0.7%)
43 (16.5%) 5 (1.8%) 2 (0.7%)
45 (16.7%) 4 (1.4%) 13 (4.8%)
neck dissection was associated with the primary operation in 18 patients and functional neck dissection in 73 patients. The operations were performed by 3 different surgeons with a similar degree of experience, as evaluated on the basis of learning curve, and with similar years in practice and number of laryngectomies performed. If a fistula developed, the first treatment was conservative, consisting of medical treatment with antibiotics and anti-inflammatory drugs; suspension of oral feeding
with position of nasogastric tube or parenteral nutrition; and daily local medication, including drainage of fluids from the fistulous tract, removal of necrotic tissue, local cleaning with antibiotic solutions, and a compressive dressing. If the fistula did not close spontaneously within 28 days of primary surgery, an operative closure was usually performed. In our series, patients who developed PCF (fistula group; 43 patients) were compared with those without postlaryngectomy complications (nonfistula group; 225 patients). We evaluated a series of variables, including smoking and drinking habits, preoperative radiotherapy, surgical technique and suture materials, concurrent neck dissection, tumor stage, presence of systemic diseases, and the respective incidence of the variables in the 2 groups. Statistical comparison of categorical variables was performed with Pearson’s chi-square test, and a P value less than 0.05 was considered significant. Odds ratio (OR) and 95% confidence interval (95% CI) were obtained by multivariate logistic model performed by the SSPS for Windows version 9.0 statistical software.
Table 2 Relationship between PCF formation and investigated risk factors Variables Sex M F Median age Smoker No Yes Alcohol intake None ⬍1 L ⬎1 L Systemic disease Diabetes mellitus Cardiopathy Chronic bronchitis Neurological Previous radiotherapy Concurrent neck dissection Radical Functional Region Supraglottis Glottis Subglottis Transglottis Pharyngeal-larynx Type of laryngectomy Standard Extended Suture material used Vicryl Chromic catgut ns, not significant.
Fistula (n ⫽ 43)
Nonfistula (n ⫽ 225)
Significance (P)
39 (90.6%) 4 (9.4%) 62
212 (94.2%) 13 (5.8%) 64
ns ns
6 (13.9%) 37 (86.1%)
21 (9.3%) 204 (90.7%)
ns ns
11 18 14 20 4 9 5 2 18
54 109 62 68 20 21 19 8 60
(25.5%) (41.86%) (32.5%) (46.5%) (9.3%) (20.9%) (11.6%) (4.6%) (41.8%)
6 (13.9%) 11 (25.5%) 12 23 2 3 3
(27.9%) (53.4%) (4.6%) (6.9%) (6.9%)
(24%) (48.4%) (27.5%) (30.2%) (8.8%) (9.3%) (8.4%) (3.5%) (26.6%)
ns ns ns ⬍0.05 ns ⬍0.05 ns ns ⬍0.05
12 (5.3%) 62 (27.5%)
⬍0.05 ns
35 156 11 8 15
(15.5%) (69.3%) (4.8%) (3.5%) (6.6%)
⬍0.05 ns ns ns ns
40 (93%) 3 (7%)
210 (93.3%) 15 (6.6%)
ns ns
21 (48.8%) 22 (52.2%)
109 (48.4%) 116 (51.6%)
ns ns
Galli et al
Postlaryngectomy Pharyngocutaneous Fistula: . . .
691
4 weeks since primary surgery (wait-and-see period), we observed only a reduction of the size of PCF; therefore a surgical closure was required. In all cases we performed a direct suture of the pharyngeal mucosa. The choice of the surgical technique was dictated by the small diameter of the fistulas. No patient presented dysphagia due to fibrous stenosis, after conservative or surgical treatment. Radiotherapy had been performed in 78 out of 268 patients, in a period ranging from 6 months to 20 years before laryngectomy, with a dosage varying from 42 to 70 Gy. Among these cases, 18 developed PCF (n ⫽ 3, 1-2 cm; n ⫽ 15, 2-3 cm). We did not find differences in dosage of radiation in these groups with respect to patients who received preoperative radiotherapy and did not develop fistula after total laryngectomy (n ⫽ 60). Nevertheless, the small number of patients doesn’t allow us to draw a meaningful inference. The relationships between PCF formation and investigated parameters are summarized in Table 2. In our series, the presence of systemic diseases, previous radiotherapy, the origin of the tumor from the supraglottic region, and the concurrent radical neck dissection were all significantly associated with the development of the fistula (46.5% vs 30.2%, 41.8% vs 26.6%, 27.9% vs 15.5%, and 13.9% vs 5.3%, respectively; Fig 1). Conversely, there were no statistically significant differences between the fistula and the nonfistula groups with regard to age, sex, smoking, drinking habits, concurrent functional neck dissection, and type of total laryngectomy (standard or extended to pharyngeal mucosa). Furthermore, differences concerning the suture material and the surgeon’s ability did not affect the outcome of the 2 groups. Finally, in our series, T stage was not significantly associated with fistula formation (Table 3). The OR and the 95% CI obtained from a multivariate logistic model including terms for cigarette smoking, alcohol intake, previous radiotherapy and neck dissection (radical and functional), and associated systemic disease (diabetes mellitus, cardiopathy both as congestive heart failure and ischemic heart disease, chronic bronchitis, neurological disease) are shown in Table 4. We observed a significant correlation between PCFs and the supraglottic tumor site (OR: 1.5; 95% CI: 0.37-6.17), preoperative radiotherapy (OR: 1.94; 95% CI: 0.91-4.14), concurrent radical neck
Figure 1 Significant differences in incidence of systemic disease (SD), previous radiotherapy (PR), supraglottic carcinoma (SGC), and radical neck dissection (RND) in the fistula and nonfistula groups.
RESULTS The patient population consisted of 17 women and 251 men, ranging in age from 22 to 88 years (median age, 64 years). The majority of the patients (241, 89.9%) were smokers, 28.3% of the patients were classified as heavy drinkers (1 liter of wine per day or more), and 47.4% as moderate drinkers (less than 1 liter per day). Eighty-eight patients (32.8%) had a chronic systemic disease such as chronic congestive heart failure or ischemic heart disease (11.2%), diabetes mellitus (8.9%), chronic bronchitis (8.9%), or neurological disease (3.8%). The primary tumor was supraglottic in 47 patients (17.5%), glottic in 179 (66.8%), subglottic in 13 (4.9%), transglottic in 11 (4.1%), and pharyngolaryngeal in 18 (6.7%). The TNM staging of the patient population is shown in Table 1. A PCF developed in 43 out of 268 patients (16%), within an average of 10 days from surgery. In 28 out of 43 patients (65.1%), fistula size ranged between 1 cm and 2 cm. Applying conservative management, we observed spontaneous closure in all patients, within an average of 20 days from the appearance of the PCF (range, 5-28 days). In 15 out of 43 patients (34.9%), all submitted to previous radiotherapy, the diameter of the PCF ranged between 2 cm and 3 cm; after
Table 3 TNM staging in fistula group (n ⴝ 43) and nonfistula group (n ⴝ 225) T2
N0 N1 N2
T3
T4
F
Non F
F
Non F
F
Non F
22 (51%) 2 (4.6%) 2 (4.6%)
106 (47%) 24 (10.6%) 0
10 (23%) 2 (4.6%) 0
33 (14%) 3 (1.3%) 2 (0.8%)
3 (6.9%) 0 2 (4.6%)
42 (18.6%) 4 (1.7%) 11 (4.8%)
F, fistula group; Non F, nonfistula group.
692
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
Table 4 Estimates of fistula risk according to selected individual characteristic Variable
Regression coefficient
Standard error
OR
95% CI
0.26
0.39
1.30
0.59-2.84
0.05 1.20 0.08 0.73 0.06 0.09 0.66
0.42 0.98 1.50 0.81 1.51 1.52 0.38
1.05 3.34 0.55 1.92 0.51 0.38 1.94
0.45-2.43 0.18-6.11 0.21-7.77 0.62-18.52 0.25-9.7 0.14-10.3 0.91-4.14
1.78 0.61
0.63 0.43
5.96 1.48
1.70-15.87 0.78-4.34
0.4 0.4 0.5 0.6 0.5
0.7 0.6 0.7 0.9 0.6
1.5 0.6 0.3 1.4 0.8
0.37-6.17 0.1-2.4 0.1-1.5 0.3-0.5 0.3-0.9
Smoker Yes Alcohol intake Yes Systemic disease Diabetes mellitus Cardiopathy Chronic bronchitis Neurological Previous radiotherapy Concurrent neck dissection Radical Functional Region of neoplasm Supraglottic Glottic Sottolgottic Transglottic Pharyngeal-larynx OR, odds ratio; CI, confidence interval.
dissection (OR: 5. 96; 95% CI: 1.70-15.87), and systemic diseases (OR: 3.34; 95% CI: 0.18-6.11) in particular, and a high correlation with cardiopathy (OR: 1.92; 95% CI: 0.6218.52).
DISCUSSION After total laryngectomy, the appearance of a PCF results in a delayed outcome and rarely in carotid artery rupture.13 Generally PCF develops just above the tracheostoma, at the weakest point of the suture line of the pharyngeal mucosa, and it is always associated with surrounding tissue necrosis. Other localization of the salivary fistulas can occur higher in the neck, at the level of the junction between the pharyngeal mucosa and the base of the tongue. The presence of saliva results in surrounding tissue infection and microvenous thrombosis with a progressive tissue loss. The corresponding neck skin becomes tender and dark red in color. In previous studies a variable incidence of PCFs after total laryngectomy has been reported, ranging from 3% to 65%, with an average time of appearance of 10 days after surgery.1-4 Our observations of 43 fistulas in our group of 268 patients (16%) and a mean time of 10 (standard deviation ⫾4) days are therefore consistent with the mean values reported in the literature. Multivariate statistical analyses show a significant correlation between PCF and supraglottic tumor site, preoperative radiotherapy, and concurrent radical neck dissection. Several authors11,13 reported similar data, and particularly in agreement with our results, Virtaniemi et al1 have recently shown that supraglottic tumors have significantly
more fistulas than glottic ones. Cavalot et al,14 however, have considered preoperative radiotherapy as the crucial factor in fistula formation, especially when delivered at a full dose. In our series, confirming several previous studies,14,15 other factors, such as T stage, concurrent functional neck dissection, and extended total laryngectomy were not significantly associated with PCF. Nevertheless, in this study there are too few patients with ‘extended’ resections to pharyngeal mucosa, pyriform sinus, or base of the tongue to allow us to reach a meaningful conclusion. Indeed, in accordance with several authors,5,11 we believe that a closure under tension, caused by a resection of large amounts of pharyngeal mucosa, could increase the incidence of fistula formation. Furthermore, multivariate statistical analyses show a significant correlation between fistula formation and systemic diseases, particularly cardiopathy. Even though the exact mechanism of this correlation is not known at the moment, a reduced immune response and an impaired mechanism of wound healing, due to microangiosclerosis, could be the most likely factors involved.15 Similarly, Redaelli de Zinis et al16 have described ischemic cardiopathy as a specific risk factor for medical morbidity after total laryngectomy. Moreover, Hier et al9 and Lavelle and Maw2 have underlined the significance of postoperative hemoglobin level (⬍12.5 g/dL) and intraoperative transfusions because of their indication of patients’ general conditions. In the study reported by Horgan and Dedo,17 the surgeon’s ability played an important role in the formation of fistulas. In general, we agree with such observations. How-
Galli et al
Postlaryngectomy Pharyngocutaneous Fistula: . . .
ever, in our series we could rule out this variable because in our clinic, during the observation period, total laryngectomy treatment was performed by 3 surgeons with comparable surgical experience. Recently, Lavelle and Maw2 underlined the importance of the suture material in fistula development. A higher incidence of PCF was observed after the use of continuous Nylon suture. Verma et al18 reported a sharp drop in the rate of fistula formation after switching from chromic catgut to Vicryl. In our case records, the suture material did not seem to be a predisposing factor, because the use of different materials did not correlate with the incidence of PCF in the 2 groups. PCF diagnosis is exclusively clinical and is made during the postoperative course. Recently some authors have suggested that fever in the first postoperative day seems to play a prognostic role in PCF development.19 They found a high correlation between fever (⬎101.5°F) that developed in the first 48 hours and eventual fistula formation. Therefore, they proposed that most fistulas are the consequence of an inadequate closure, not of a wound breakdown, as generally believed. Furthermore, these same authors demonstrated that early intervention (before third postoperative day) decreased morbidity and duration of hospitalization. Concerning the treatment, some authors20 refrain from waiting for a spontaneous closure because this may take a long time (often more than 1 month), prolonging the patient’s stay in the hospital and resulting in swallowing difficulty due to fibrosis in the majority of cases. Our management of PCF is generally conservative because we have observed that chemical cauterization with a silver nitrate stick and medical treatment can sometimes contribute to a spontaneous closure of the fistula, as we observed, with a mean time of 20 days. In our series, small fistulas (1-2 cm), without previous exposure to radiotherapy, responded well to conservative management whereas large PCFs (2-3 cm), mainly related to previous radiotherapy, do not close spontaneously with a conservative approach. We suggest that in these cases it is not wise to wait any longer than a month to plan surgery. Nevertheless, with this approach we obtained a reduction of PCF size, so the small diameter reached did not require reconstruction with local flaps (deltopectoral and pectoralis major myocutaneous flaps) or free flaps.
CONCLUSIONS Our results confirm that PCF remains a troublesome complication of the early postoperative period after total laryngectomy.1,15,16 In the presence of specific risk factors, PCF can be expected; nevertheless, its prevention remains very difficult. Our data show that the presence of systemic diseases (with particular regard to cardiopathy), previous radiotherapy, radical neck dissection, and the origin of the
693
tumor from the supraglottic region can all be important predisposing factors, or at least underlying causes, in fistulas’ arising. On the other hand, we do not find significant differences between the 2 groups concerning suture material, surgical technique, and surgeon ability (evaluated on the basis of a learning curve), despite the fact that several authors2,18 have considered these as predisposing factors to PCF. It must be underlined that in the present study, surgery was performed by different but almost equally well-experienced surgeons. Therefore our results indirectly indicate that surgeon ability represents a significant variable impacting on the eventual development of PCF. Concerning the treatment, we believe that an observation period of 28 days after the primary surgical procedure is the best approach for PCF. Indeed, in small PCFs without previous radiotherapy exposure, a conservative approach permits spontaneous closure without stenosis or dysphagia. But in larger PCFs, radiotherapy related, it determines reduction of its size, reducing the need of reconstruction surgery with local or free flaps.
REFERENCES 1. Virtaniemi JA, Kumpulainen EJ, Hirvikoski PP, et al. The incidence and etiology of postlaryngectomy pharyngocutaneous fistulae. Head Neck 2001;23(1):29 –33. 2. Lavelle RJ, Maw AR. The aetiology of post-laryngectomy pharyngocutaneous fistulae. J Laryngol Otol 1972;86(8):785–93. 3. Thawley SE. Complications of combined radiation therapy and surgery for carcinoma of the larynx and inferior hypopharinx and inferior hypopharynx.. Laringoscope 1981;91:677–700. 4. Bressan K, Rasmussen H, Rasmussen PA. Pharyngocutaneous fistulae in totally laryngectomized patients. J Laryngol Otol 1974;88:835– 42. 5. Ikiz AO, Uca M, Guneri EA, et al. Pharyngocutaneous fistula and total laryngectomy: possible predisposing factors, with emphasis on pharyngeal myotomy. J Laryngol Otol 2000;114(10):768 –71. 6. Fradis M, Podoshin L, Ben David J. Post-laryngectomy pharyngocutaneous fistula—a still unresolved problem. J Laryngol Otol 1995; 109(3):221– 4. 7. Markou KD, Vlachtsis KC, Nikolaou AC, et al. Incidence and predisposing factors of pharyngocutaneous fistula formation after total laryngectomy: is there a relationship with tumor recurrence? Eur Arch Otorhinolaryngol 2004;261:61–7. 8. Parikh SR, Irish JC, Curran AJ, et al. Pharyngocutaneous fistulae in laryngectomy patients: the Toronto Hospital experience. J Otolaryngol 1998;27(3):136 – 40. 9. Hier M, Black MJ, Lafond G. Pharyngo-cutaneous fistulas after total laryngectomy: incidence, etiology and outcome analysis. J Otolaryngol 1993;22(3):164 – 6. 10. Johansen LV, Overgaard J, Elbrond O. Pharyngo-cutaneous fistulae after laryngectomy: influence of previous radiotherapy and prophylactic metronidazole. Cancer 1988;61(4):673– 8. 11. Natvig K, Boysen M, Tausjo J. Fistulae following laryngectomy in patients treated with irradiation. J Laryngol Otol 1993;107(12): 1136 –9. 12. Sobin LH, Wittekind C. UICC International Union Against Cancer: TNM classification of malignant tumours. 5th ed. New York: WileyLiss; 1997. p. 33-7. 13. McCombe AW, Jones AS. Radiotherapy and complications of laryngectomy. J Laryngol Otol 1993;107(2):130 –2.
694
Otolaryngology–Head and Neck Surgery, Vol 133, No 5, November 2005
14. Cavalot AL, Gervasio CF, Nazionale G, et al. Pharyngocutaneous fistula as a complication of total laryngectomy: review of the literature and analysis of case records. Otolaryngol Head Neck Surg 2000; 123(5):587–92. 15. Papazoglou G, Doundoulakis G, Terzakis G, et al. Pharyngocutaneous fistula after total laryngectomy: incidence, cause, and treatment. Ann Otol Rhinol Laryngol 1994;103(10):801–5. 16. Redaelli de Zinis LO, Ferrari L, Tomenzoli D, et al. Postlaryngectomy pharyngocutaneous fistula: incidence, predisposing factors, and therapy. Head Neck 1999;21(2):131– 8.
17. Horgan EC, Dedo HH. Prevention of major and minor fistula after laryngectomy. Laryngoscope 1979;89:250 – 60. 18. Verma A, Panda NK, Mehata S, et al. Post-laryngectomy complications and their mode of management: an analysis of 203 cases. Indian J Cancer 1989;26:247–54. 19. Friedman M, Venkatesan TK, Yakovlevmd A, et al. Early detection and treatment of post-operative pharyngocutaneous fistula. Otolaryngol Head and Neck Surg 1999;121:378 – 80. 20. Soylu L, Kiroglu M, Aydogan B, et al. Pharyngocutaneous fistula following laryngectomy. Head Neck 1998;20(1):22–5.