Key points
- Pharyngo-cutaneous fistula (PCF) occurred in 31% of cohort of 242
patients
- 81% of PCFs ultimately healed
- 52% of PCFs underwent at least 1 operation
- PCF was not associated with a delay in adjuvant radiotherapy nor
increased risk of future dilatation
- PCF is associated with a significantly worse 5 year overall survival
(44% vs 24%, , OR 1.7, p=0.001)
Introduction
Total laryngectomy (TL) is used as a primary treatment for advanced
stage laryngeal or hypopharyngeal cancer, as salvage in patients with
recurrent disease after failure of organ preserving treatments or as a
functional treatment in patients with a dysfunctional larynx1-4.
Postoperative complications including pharyngocutaneous fistula (PCF)
are common1-5 with a recent systematic
review6 reporting rates of up to 58%. PCF often
requires additional surgery, delays oral feeding/voice rehabilitation,
decreases quality of life and increases hospital stay and
costs3,7,8. Furthermore, PCF may delay postoperative
(chemo)radiotherapy, thus jeopardizing optimal oncological
treatment4,9 and potentially lead to increased risk of
neo-pharyngeal stricture formation.
Many papers describe risk factors for the occurrence of PCF including
prior chemoradiotherapy, hypopharyngeal cancer, extensive pharyngeal
resection and reconstruction, neck dissections, low BMI and low skeletal
muscle mass (sarcopenia)4,10,11. Far fewer articles
deal with the management of PCF and its consequences.
In this article we present our single center experience of 242 patients
operated over a 10 year period. We evaluate our management strategy with
regards to PCF (conservative versus operative), closure success rate and
time to closure. We also evaluate whether PCF led to prolonged hospital
stay, a delay in adjuvant radiotherapy, increased the need of future
neopharyngeal dilatations, or had an adverse effect on overall survival.
A rough estimate of extra cost incurred by PCF patients is also given.
Materials and Methods
We performed a retrospective cohort study of all patients undergoing
total laryngectomy at [blinded] over the 10 year period January 2008
to December 2017. The indication for total laryngectomy was either
primary, salvage or functional.
Patients’ demographic, staging, treatment and outcome data were
collected using electronic patient records. All patients were discussed
in our multidisciplinary tumour board and underwent total laryngectomy
with or without (partial) pharyngectomy. For patients where the pharynx
could not be closed primarily, a variety of flaps were used including
pectoralis major with or without skin island, free radial forearm flap
(FRFF), anterior lateral thigh flap (ALT), jejunum interposition and
gastric pull up. Occasionally a pectoralis major muscle onlay flap (also
reported in the literature as an interposition flap) was used to
reinforce a primarily closed pharynx.
Postoperatively, all patients were fed via a feeding tube for the first
10 days before undergoing a contrast swallow. If this showed no leaks,
the patient would begin with clear fluids and quickly build up to a
normal diet. If there was a leak, the patient would be kept nil-by-mouth
and a further contrast swallow study organised for the following week
(unless a PCF manifested itself in during this period).
For this study, PCFs were defined by their clinical manifestation on the
skin and patients who had leakage on the contrast swallow but never
manifested a PCF were counted as not having a PCF.
The majority of PCFs occurred shortly after surgery, though some
patients were discharged either with a feeding tube or on oral diet only
to present with a PCF later. These patients were classified as having a
PCF. Still further patients developed a PCF many months later secondary
to an intervention such as dilatation of a stricture. These patients
were classified as no PCF as the development of a PCF was considered a
complication of the second procedure, not of the initial laryngectomy.
In those patients who developed a PCF, initial management was
conservative using a variety of wound dressings. For fairly dry PCFs,
Eusol (Edinburgh University solution of lime) wound dressings were used.
For more productive PCFs either iodine soaked gauze or an alginate
dressing. For the most productive a stoma bag had to be used.
Occasionally scopolamine patches were used to reduce saliva production.
The indication for operative management of a fistula was up to the lead
surgeon’s judgement. This depended on many factors including the size of
the fistula, its response to conservative therapy, the patient’s
fitness, and the availability of a suitable method of reconstruction
(the favoured reconstruction being a salivary bypass tube and pectoralis
major flap).
Outcomes of interest included the number of PCFs, duration of PCF,
management of PCF (conservative vs operative) and method of
reconstruction of the PCF. Some patients died from complications of the
fistula and this was also noted. In the longer term, we looked at
whether PCF led to a delay in adjuvant radiotherapy (delay was defined
as starting radiotherapy >6 weeks after surgery) and
whether the PCF correlated need for dilatation of a neopharyngeal
stenosis during follow up. Lastly, we looked at PCF impact on overall
survival and make a rough estimation of the extra costs involved in
patients with PCF.
Statistical analyses were performed using SPSS® Statistics 20.0 (IBM,
Armonk, NY). Overall survival and hospital length of stay was calculated
using the Kaplan–Meier method and Cox-mantel log-rank test for
comparison. The chi-squared test or binary logistic regression analysis
was used as appropriate for univariate analysis.
Results
A total of 242 patients (199 males, 43 females) underwent total
laryngectomy (TL) with or without (partial) pharyngectomy (TLPP) in the
study period. Treatment indication was primary (n=115), salvage (n=102)
and functional (n=25). Details of the cohort general characteristics
stratified by indication is shown in table 1.
Postoperative course /LOS
Mean average length of hospital stay was 21 days (16 days for patients
without PCF and 30 days for patients with a PCF, OR 2.38,
p=<0.001), Figure 1. Out of 242 patients, 51 (21%) returned
to theatre: 39/51 (76%) for the treatment of PCF, 9/242 (4%) for
hemostasis, 1/242 (0.4%) for a chyle leak and 2/234 (0.8%) for
exploration of a free flap (one flap successfully salvaged, the other
unfortunately lost (jejunum interposition) giving a free flap failure
rate of 1/17 (4.5%)).
Incidence of PCFs
A total of 75/242 (31%) patients developed a PCF and these were managed
as reported in table 2. In total, 61/75 (81%) fistulas were
successfully closed, 2/75 (2.7%) fistulas were still open at last
follow-up (respectively 3 and 29 months following laryngectomy) and 12
patients died with an open PCF. These patients are discussed in more
detail below.
Conservative Management of PCF
Initial PCF management was conservative with a feeding tube and a
variety of wound dressings including Eusol, alginate dressings and/or
iodine soaked gauze. Thirty-six out of 75 (48%) PCFs were treated this
way with a success rate of 30/36 (83%). The remaining 6 patients died
with open PCFs. Median time to closure was 41 days (see figure 2).
Operative management of PCF
The decision to return to theatre was taken by the lead surgeon and
depended on a variety of factors including failure of conservative
management and fitness of the patient. For most patients, this was
during the same hospital stay as the initial laryngectomy, but for a few
patients, the surgical management of the PCF was performed months later.
In 39/75 (52%) PCFs, the patient underwent at least one operation for
their fistula. Twenty-three of 29 patients had one operation, 6 patients
had 2 operations, 3 patients had 3 operations, 5 patients had 4
operations, 1 patient had 6 operations and 1 patient had 11 operations
for his fistula. Twenty-two of 39 (56%) patients had a regional flap
reconstruction (either pectoralis major or delto-pectoral). Of the
operatively managed patients, the PCF healed in 31/39 (79%). Of the
patients managed with regional flap reconstruction for their PCF 14/20
(70%) healed. Of the patients managed with local operative techniques
17/19 (89%) of PCFs healed.
Median time to closure in the operatively managed patients was 61 days.
Kaplan-Meier analysis of time with fistula vs management (conservative
vs operative) showed no significant differences (p= 0,119, see Figure
2).
Salivary bypass tubes
In patients undergoing operative management of their PCF, a salivary
bypass tube was used in 32/39 (82%) of operated patients. Of the 7
patients managed without a salivary bypass tube 4 had a regional flap
reconstruction and 3 a local closure.
Does PCF lead to delayed adjuvant radiotherapy for primary laryngectomy
patients?
Post-operative radiotherapy was indicated in 100/115 primary patients.
Postoperative radiotherapy was started >7 weeks following
operation in 28/100 primary patients. Twelve of 28 patients had a PCF
versus 16/28 who had a delay despite no PCF. Binary logistic regression
showed no correlation between PCF and delay to radiotherapy (p=0.06).
Further worth noting is that 5 patients started adjuvant radiotherapy
with an open PCF and in all these 5 patients the PCF closed during the
radiotherapy.
Does PCF lead to increased risk of stricture formation necessitating
dilatation ?
Inflammation, multiple surgeries and prolonged tube feeding associated
with PCF may plausibly lead to increased long-term risk of neopharyngeal
stricture formation, potentially necessitating dilatation. In our
cohort, a total of 44/242 (18%) of patients ultimately underwent
dilatation during follow-up, 15/44 (34%) had had a PCF. Once again,
binary logistic regression showed no statistically significant
correlation between PCF and need for dilatation (p=0.623).
Does PCF adversely affect perioperative or long-term survival?
Laryngectomy is a physiologically demanding procedure, often performed
in patients with little physiological reserve. In our cohort, there were
3/242 (1.2%) perioperative deaths (defined as deaths within 30 days of
laryngectomy). One 76 year old patient had an injury to his common
carotid artery intraoperatively which had to be ligated unfortunately
leading to an ultimately fatal ischaemic stroke. A 67 year old patient
had a fatal myocardial infarction on the first postoperative day.
Finally a 76 year old male had necrosis of his pectoralis flap leading
to carotid blow-out 1 month after laryngectomy.
In total 12/75 (16%) patients died with an open PCF which had never
healed since the original laryngectomy. In these patients we considered
the PCF a complication of the laryngectomy. (Other patients ultimately
died of blow-outs associated with pharyngo-cutaneous fistulas due to
either local recurrence or complications of dilatation. As the PCF in
these cases was not considered a complication of laryngectomy, they are
not described below.)
In 1/12 patients with a persistent fistula, the cause of death was
unknown. In 3/12 patients there was a separate cause of death (lung
embolus, peritonitis and small cell lung cancer). One patient died from
distant metastasis. Three of 12 had complications due to their fistula
(the blow-out mentioned above, 2 had spondylitis) but no sign of cancer.
The remaining 4 patients had biopsy confirmed local residual disease
(which retrospectively was probably the reason their PCF had not
healed).
We can therefore say that 3/75 (4%) patients with PCF died as a direct
result of it without sign of cancer.
When we analyzed long term overall survival comparing patients who
developed a post-operative PCF and those who did not, we found a
significant difference (5 year overall survival 24% versus 44%
respectively, OR 1.7, p=0.001, Figure 3).
Costs associated with PCF formation
A complete analysis of the costs associated with PCF formation is beyond
the scope of this article. A proper analysis would have to include not
only inpatient costs, but also costs for outpatient care and costs
associated with inability to work.
Given the differences in LOS and that a hospital bed in our institution
very roughly costs 633 Euros per day we can estimate that a PCF patient
costs an extra 8862 euros in bed costs alone. This substantial sum does
not even take into account that many PCF patients underwent multiple
admissions for the management of their PCF, with extra imaging and
surgery (PCF patients underwent an average of 1.1 (81/75) operations
compared to patients without PCF). Outside of the hospital, PCF patients
also required more intensive nursing and tube feeding. Several patients
with particularly persistent PCFs even underwent hyperbaric oxygen
therapy. Along with delays in returning to work, all these factors
vastly increase the costs to the healthcare system in case of PCF.
Discussion
In line with the literature6,12, roughly a third of
our developed a PCF with a roughly 50/50 split between conservative and
operative management. PCFs successfully closed in 61/75 (81%) cases
with both conservative and operative approaches having a similar success
rate. We can assume that smaller fistulas were treated conservatively
and larger fistulas surgically and this high closure rate reflects the
correct strategy for these PCFs. Three of 75 (4%) patients died of
complications from their PCF without sign of cancer. Furthermore, PCF
patients had a poorer overall survival (5 year overall survival PCF 24%
versus 44% in the no-PCF group, OR 1.7, p=0.001). PCF does not however
seem to delay the start of adjuvant radiotherapy (p=0.06). Indeed, in 5
patients the PCF was deemed small enough to all adjuvant radiotherapy to
begin despite an open PCF. PCF also does not increase the risk of future
dilatation.
Median length of hospital stay in our cohort was 14 days. This is
somewhat higher than in other reports13 and may be due
to our protocol of keeping patients until their contrast swallow study
at 10 days which then allows supervised initiation of feeding and voice
rehabilitation. In any case, length of hospital stay was significantly
longer in the PCF group than in the no-PCF group (figure 1) which
together with the extra nursing care required, extra imaging studies,
extra operations and delayed return to work vastly increase the costs
associated with PCF.
Given the mortality and morbidity associated with PCF, many surgeons
have suggested strategies to reduce their incidence. Pre-operatively
there is unfortunately often little time for “pre-habilitation”.
Intra-operatively, modifications of technique such as the use of barbed
sutures14, stapler closure15,16,
salivary bypass tubes17, sealants18or even transoral robotic total laryngectomy19,20 have
all been suggested. Reinforcing primary pharyngeal closure with a
pectoralis major muscle-only overlay has also been suggested21. This latter strategy is clearly more invasive and
is probably only appropriate in high risk patients, though the
number-needed-to treat and the extra morbidities and costs associated
with an interposition flap are all areas for further investigation. The
extra hour of operative time and morbidity of such an interposition flap
takes well outweigh the morbidity associated with dealing with a PCF at
a later date. Even if patients do suffer a PCF after interposition,
there is also the hope that the PCF will be less severe and could be
conservatively managed.
Post-operative techniques to reduce the risk of PCF such as reducing
saliva (scopolamine patches, botuline injections, parotid duct
ligation)22, early oral fluids23 and
vacuum drains24/endoscopic negative pressure
therapy25 all have their proponents though remain
controversial.
Unfortunately, when there is such a plethora of varying advice in the
literature, it is often because no single management option is
particularly more effective than any other. Therefore, it remains
unclear what the optimal management strategy for prevention and
treatment of PCF is.
Conclusions
PCF remains a troublesome complication after total laryngectomy.
Patients with PCF often have prolonged or multiple hospital admissions
with associated higher morbidity and costs. Their overall survival is
also significantly worse than patients who do not develop a PCF. The
risk factors for PCF are numerous and described elsewhere in the
literature, as are potential ways to reduce this risk. PCF however does
not seem to lead to a delay to adjuvant radiotherapy, nor to an
increased risk of stricture formation.
References
1. Van der Putten L, De Bree R, Kuik D, et al. Salvage laryngectomy:
oncological and functional outcome. Oral oncology.2011;47(4):296-301.
2. Theunissen EA, Timmermans AJ, Zuur CL, et al. Total laryngectomy for
a dysfunctional larynx after (chemo) radiotherapy. Archives of
Otolaryngology–Head & Neck Surgery. 2012;138(6):548-555.
3. Paydarfar JA, Birkmeyer NJ. Complications in head and neck surgery: a
meta-analysis of postlaryngectomy pharyngocutaneous fistula.Archives of Otolaryngology–Head & Neck Surgery.2006;132(1):67-72.
4. Lansaat L, Van Der Noort V, Bernard SE, et al. Predictive factors for
pharyngocutaneous fistulization after total laryngectomy: a Dutch Head
and Neck Society audit. European Archives of
Oto-Rhino-Laryngology. 2018;275(3):783-794.
5. Goepfert RP, Hutcheson KA, Lewin JS, et al. Complications, hospital
length of stay, and readmission after total laryngectomy. Cancer.2017;123(10):1760-1767.
6. Liang J-W, Li Z-D, Li S-C, Fang F-Q, Zhao Y-J, Li Y-G.
Pharyngocutaneous fistula after total laryngectomy: a systematic review
and meta-analysis of risk factors. Auris Nasus Larynx.2015;42(5):353-359.
7. Hasan Z, Dwivedi R, Gunaratne D, Virk S, Palme C, Riffat F.
Systematic review and meta-analysis of the complications of salvage
total laryngectomy. European Journal of Surgical Oncology (EJSO).2017;43(1):42-51.
8. Busoni M, Deganello A, Gallo O. Pharyngocutaneous fistula following
total laryngectomy: analysis of risk factors, prognosis and treatment
modalities. Acta Otorhinolaryngologica Italica. 2015;35(6):400.
9. Casasayas M, Sansa A, García-Lorenzo J, et al. Pharyngocutaneous
fistula after total laryngectomy: multivariate analysis of risk factors
and a severity-based classification proposal. European Archives of
Oto-Rhino-Laryngology. 2019;276(1):143-151.
10. Ganly I, Patel S, Matsuo J, et al. Postoperative complications of
salvage total laryngectomy. Cancer: Interdisciplinary
International Journal of the American Cancer Society.2005;103(10):2073-2081.
11. Bril SI, Pezier TF, Tijink BM, Janssen LM, Braunius WW, de Bree R.
Preoperative low skeletal muscle mass as a risk factor for
pharyngocutaneous fistula and decreased overall survival in patients
undergoing total laryngectomy. Head & neck.2019;41(6):1745-1755.
12. Virtaniemi JA, Kumpulainen EJ, Hirvikoski PP, Johansson RT, Kosma
VM. The incidence and etiology of postlaryngectomy pharyngocutaneous
fistulae. Head & Neck: Journal for the Sciences and Specialties
of the Head and Neck. 2001;23(1):29-33.
13. Shenson JA, Craig JN, Rohde SL. Effect of preoperative counseling on
hospital length of stay and readmissions after total laryngectomy.Otolaryngology–Head and Neck Surgery. 2017;156(2):289-298.
14. Elmallah RK, Khlopas A, Faour M, et al. Economic evaluation of
different suture closure methods: barbed versus traditional interrupted
sutures. Annals of translational medicine. 2017;5(Suppl 3).
15. Calli C, Pinar E, Oncel S. Pharyngocutaneous fistula after total
laryngectomy: less common with mechanical stapler closure. Annals
of Otology, Rhinology & Laryngology. 2011;120(5):339-344.
16. Aires FT, Dedivitis RA, Castro MAF, Bernardo WM, Cernea CR, Brandao
LG. Efficacy of stapler pharyngeal closure after total laryngectomy: a
systematic review. Head & neck. 2014;36(5):739-742.
17. Montgomery WW. Salivary bypass tube. Annals of Otology,
Rhinology & Laryngology. 1978;87(2):159-162.
18. Stephenson KA, Pandey S, Lubbe DE, Fagan JJ. Use of surgical sealant
in the prevention of pharyngocutaneous fistula after total laryngectomy.Head & neck. 2018;40(12):2606-2611.
19. Lawson G, Mendelsohn A, Fakhoury R, et al. Transoral Robotic Surgery
Total Laryngectomy. ORL. 2018;80(3-4):171-177.
20. Krishnan G, Krishnan S. Transoral robotic surgery total
laryngectomy: evaluation of functional and survival outcomes in a
retrospective case series at a single institution. ORL.2017;79(4):191-201.
21. Guimarães AV, Aires FT, Dedivitis RA, et al. Efficacy of pectoralis
major muscle flap for pharyngocutaneous fistula prevention in salvage
total laryngectomy: a systematic review. Head & neck.2016;38(S1):E2317-E2321.
22. Guntinas‐Lichius O, Eckel HE. Temporary reduction of salivation in
laryngectomy patients with pharyngocutaneous fistulas by botulinum toxin
A injection. The Laryngoscope. 2002;112(1):187-189.
23. Timmermans AJ, Lansaat L, Kroon GV, Hamming-Vrieze O, Hilgers FJ,
van den Brekel MW. Early oral intake after total laryngectomy does not
increase pharyngocutaneous fistulization. European Archives of
Oto-Rhino-Laryngology. 2014;271(2):353-358.
24. Roberts ST, Hobson M, Eisenberg R. A simple drainage technique to
manage pharyngocutaneous fistula following laryngectomy. Clinical
Otolaryngology. 2019;44(6):1211-1213.
25. Loeck J, von Lücken H-J, Kehrl W, Loske G. Endoscopic negative
pressure therapy (ENPT) of a post-laryngectomy pharyngocutaneous
fistula: first report of a new treatment method. HNO.2019;67(2):77-79.