Post-pharyngo-laryngectomy fistula is a frequent surgical complication, resulting in delayed patient management and reduced quality of life. A recent study has shown that the appearance of a fistula can be detected early by measuring cytokines (particularly IL10) in postoperative drainage fluids. Resumption of feeding following this surgery varies between postoperative day 5 and postoperative day 15, depending on the team's habits. Early refeeding reduces the length of hospital stay and improves patients' quality of life. The decision to refeed is currently made without any clinical or biological marker of the quality of pharyngeal suture healing. The idea of this study is that good initial healing (evidenced by low levels of inflammatory cytokine in drainage fluids) allows early refeeding without putting the patient at additional risk.

Main hypothesis and research objectives Main hypothesis: the determination of cytokines in postoperative drainage fluids (FODP) could be a tool for screening patients at no risk of developing a fistula, and for whom early refeeding (as early as postoperative day 3) could be proposed without exposing the patient to an additional risk of developing a FODP.

Study objectives Primary objective: to compare the rate of post-(pharyngo)-total laryngectomy fistula between an early refeeding strategy (3rd or 4th postoperative day) and a late refeeding strategy (current standard of care: 7th postoperative day) in a low-risk fistula group defined according to the level of inflammatory cytokines in drainage fluids on the second postoperative day.

Secondary objectives

To compare an early versus a late refeeding strategy for patients at low risk of fistula defined according to the level of inflammatory cytokines in drainage fluids on the second postoperative day for:

• length of hospital stay,
• evolution of nutritional status,
• time to postoperative radio-chemotherapy, if indicated postoperatively,
• improvement in quality of life,
• post-operative complications other than pharyngeal fistula. Evaluation of different strategies left to the investigator's choice (continuation of antibiotic therapy, increased delay before resumption of feeding, early resumption of surgery) in the group of patients at high risk of fistula, based on the determination of postoperative cytokines in drainage fluids.

Primary endpoint: Pharyngo-cutaneous fistula occurring within 30 days post-operatively (yes/no).

Number of subjects: 250 Inclusion criteria Major cancer patient justifying scheduled total laryngectomy or pharyngo-laryngectomy after multidisciplinary consultation.

Holder of a social security plan. Non-inclusion criteria Pregnant and breast-feeding women Persons under curatorship, guardianship, safeguard of justice or deprived of liberty.

Any medical condition deemed incompatible with the study by the investigator. Refusal to participate. Body mass index less than 18.5 kg/m2

Course of the study Patients will be recruited from the ENT departments of the Clermont Ferrand, Saint-Etienne, Lyon and Grenoble university hospitals, the Lyon (Centre Léon Bérard) and Paris (Institut Gustave Roussy) cancer centers, and the Le Puy En Velay and Valence hospitals, all of which are accredited for cervico-facial carcinology.

A multidisciplinary consultation meeting will be required to validate the surgical procedure.

Patients will be informed of the study by the investigating physician. After a period of reflection and a chance to answer any questions, the patient will be included in the study after consent has been obtained from an investigator.

The main risk factors for the development of a fistula will be recorded pre- and intraoperatively.

Surgical criteria will be common to all investigating teams. The most important of these will be the positioning of the LDPO collection drain opposite the pharyngeal closure site. The drain may or may not be aspirative, depending on local habits. The concomitant insertion of a phonatory prosthesis or a salivary bypass are not grounds for exclusion.

Post-operative care must meet precise specifications. Drainage fluids will be collected from the drainage bottle on D2 post-op. To do this, the entire drainage bottle will be recovered, and a new bottle will be placed over the drain if the drain is retained.

The liquid contained in the drainage bottle is then sterile-separated into different samples directly in the patient's hospital ward:

• Immunological sampling. A BD Falcon™ Conical Tubes 50 mL will be sent to the immunology laboratory associated with the investigating center (Clermont-Ferrand University Hospital Immunology Laboratory for the Clermont-Fd and Le Puy en Velay sites, Lyon University Hospital Immunology Laboratory for the Grenoble, Lyon, Valence and Lyon anti-cancer center, the immunology laboratory of the Assistance Publique Hopitaux de Paris for the Gustave Roussy Institute and the immunology laboratory for the Saint-Etienne center). The sample is then centrifuged at 3,000 rpm for 10 minutes at 4°C.a "high-risk" group for CPE.
• Bacteriological sampling. This sample will be sent rapidly to the bacteriology laboratory of the investigating center for standard cyto-bacteriological analysis.

For an IL 10 level on D2 post-op of less than 72pg/mL (so-called low-risk fistula group defined from the DEFILAC pilot study), two groups of patients will be constituted:

Low-risk fistula group: OPT performed on D3 or D4 (before randomization and after IL-10 results). If TPO classified according to van la Parra ≤1, patients will be randomized into 2 groups:

• "EARLY" group: Resumption of feeding on postoperative day 3 or 4 with a liquid or mixed diet.
• "TARDIF" group: return to a mixed diet from day 7 post-op. Randomization will be centralized, computerized with random block sizes, stratified by center and patient age.

If the TOP shows images classified according to van la Parra>1, patients will join the so-called high-risk CPE exploratory arm.

If IL10 levels on D2 post-op exceed 72 pg/mL (high-risk group for post-op fistula), these patients will form a third, exploratory group, whose management will be left to the investigator's choice, with 3 possibilities (no change in investigators' habits): repeat surgery to clean the surgical site and repair the pharyngeal sutures, broad-spectrum antibiotic therapy while awaiting the bacteriological results of the cytobacteriological examination of the drainage fluid on postoperative day 2, without repeat surgery, monitoring and no change in the investigator's routine.

Postoperative complications will then be noted and classified according to the Clavien-Dindo classification over a period of 30 postoperative days.

The appearance of a salivary fistula at the tracheostoma will be objectified by the appearance of saliva in the tracheostoma. The appearance of a fistula at the cervical level will be sought either by imaging methods in the event of a cervical complication (infection, hemorrhage) if no surgical revision is decided and surgically by loss of sealing confirmed surgically if a revision is necessary.

The appearance of a subsequent fistula will be noted. The patient will have a final consultation 30 days after surgery if he or she is no longer hospitalized.

The quality of life of the patients will be assessed throughout the study by completing an anxiety questionnaire and a depression questionnaire. This assessment will be done before surgery, on D5, D15 and D30 of surgery. Before surgery and at D5, these questionnaires will be completed during hospitalization. At D15 and D30, the questionnaires will be completed during hospitalization if the patient is still hospitalized or by mail if the patient has been discharged from the surgical department having performed the pharyngolaryngectomy procedure. The date of the start of chemoradiotherapy will then be noted and the time between surgery and the start of this treatment will be noted.

Inclusion Criteria:

• Adult patient with cancer requiring a total laryngectomy or pharyngolaryngectomy scheduled after multidisciplinary consultation.
• Holder of a social security scheme

Exclusion Criteria:

Pregnant and breastfeeding women Persons under curatorship, guardianship, legal protection or deprived of liberty Any medical condition deemed by the investigator to be incompatible with the study. Refusal to participate. Body mass index less than 18.5 kg/ m2

-