Thoracic Duct Embolization for Delayed Chyle Leak After Lewis-Tanner Esophagectomy

Background

Esophageal cancer is the seventh most common neoplasia worldwide and is responsible for approximately 450 000 deaths per year [1]. Radical esophagectomy is a potentially curative treatment, requiring two/three-field lymphadenectomy and carrying a high risk of postoperative complications [2]. Chyle leak (CL) has a reported incidence of 1–9% and it has been associated with a higher rate of postsurgical morbidity and mortality [3]. Chylothorax usually occurs early in the postoperative period and treatment should be timely and effective to prevent patient deterioration. Delayed CL is less common and is thought to be due to an occult leak or late diagnosis [4]. Treatment options range from fully conservative approaches to surgical ones. Interventional radiology has been described as an effective treatment in the management of main thoracic duct (TD) fistulas.

Case Report

Our patient was a 54-year-old man referred to our institution with an ulcerated, bleeding mass localized 1 cm from the gastroesophageal junction, occupying 2/3 of the esophageal lumen and extending cranially for 8 cm. Biopsies confirmed a poorly-differentiated, high-grade, esophageal adenocarcinoma. PET-CT scan revealed cT3 N0 M0. Multidisciplinary discussion indicated neoadjuvant chemotherapy with FLOT: 4 cycles Docetaxel 50 mg/m2+ Oxaliplatin 1.85 mg/m2+ Leucovorin 200 mg/m2+ 5-Florouracil 600 mg/m2 i.v. The patient had a good response to the neoadjuvant treatment and underwent radical subtotal Lewis-Tanner esophagectomy and gastric pull-up reconstruction. During en bloc lymphadenectomy, the TD was identified at the base of the right chest, dissected, clipped, and divided. The chest was drained with 2 closed water-seal drains; 1 abdominal drain was placed at the level of the hiatus.

Postoperatively, total parenteral nutrition with Olimel N9E and vitamins was administered and then discontinued 6 days after surgery. The abdominal drain was removed on postoperative day 3, when drainage was less than 80 mL. The apical chest tube was removed after a negative chest X-ray (Figure 1), also on postoperative day 3. Finally, the basal chest drain was removed on day 6, with draining less than 100 mL in the last 24 h. No quality test of drain fluids was performed. The postoperative course was uneventful, and the patient was discharged in good clinical condition on postoperative day 8.

The patient was clinically assessed in the outpatient department 1 week after discharge. He had no concerns and the clinical examination was negative. Skin staples were removed and his diet increased from soft to frequent light meals. One month after hospital discharge, his general practitioner referred him to the emergency department because of severe abdominal, scrotal, and lower-limb edema. The patient had noticed scrotal edema 2 weeks earlier, with subsequent insurgence to the lower limbs and abdominal edema. A blood test revealed a mild increase of WBC (11.94), hypoalbuminemia (2.99 g/dL), hyponatremia (133 mEq/L), and BUN elevation (83 mg/dL). The chest X-ray (Figure 2) showed unilateral right pleural ef-fusion with atelectasis. A chest-abdomen CT scan confirmed the X-ray findings, revealing compression and left shift of the mediastinal structures with ground-glass appearance in the left hilum and inferior left lung lobe. Abdominal sequences showed presence of fluid in the upper quadrants in the subphrenic spaces. Oral contrast with soluble solution was also given, showing good passage through the neo-esophagus and pylorus, and no fistulas.

The patient was admitted to the surgical ward and a 24 French chest drain was inserted: 1400 mL of milky fluid was drained, thus confirming CL. Triglyceride dosage on the pleural fluid sample was 399 mg/dL and a microbiological culture was negative. Treatment was initially conservative, with fasting and fat-free parenteral nutrition with Clinimix N12G20E (1400 kcal/day) and Octreotide 0.1 mcg 3 times a day. Hyponatremia and hypoalbuminemia were managed with albumin and NaCl infusions. X-rays on day 5 showed right lung re-expansion and residual pleural effusion (Figure 3).

For the following 10 days, drain output remained stable at 2800–3000 mL per day.

On day 11, TDE was carried out by ultrasound-assisted percutaneous injection of iodinated contrast agent into the right inguinal lymph node [5]. A slow diffusion of contrast was observed in the abdominal and pelvic lymphatic network up to the cisterna chyli (CC) and TD. After identification of the target ducts, cannulation was performed with a right transabdominal approach using a 21 G sterile needle and fluoroscopic guidance. A 3-French microcatheter was advanced into the TD over a 0.014” guidewire. Two sources for leaking were identified just at the level of the clips on the TD. A 4-mm metallic coil was placed at the superior aspect of the thoracic duct to provide a matrix for glue polymerization. TD was embolized with a mixture of glue proximally injected in the duct. The microcatheter was removed immediately after glue injection (Figure 4).

No immediate complication was observed, and the following day an X-ray showed further reduction of pleural effusion and complete lung expansion. A progressive reduction of the drainage was observed during subsequent days (Figure 5). Parenteral nutrition was gradually replaced by a Medium-Chain-Fat dietary scheme and the patient was discharged for home-care aid with 90 mL of output per day. Fat food was progressively reintroduced in the diet, and the chest drain was finally removed 11 days after thoracic ducts embolization (Figure 6).

Discussion

We report what appears to be a true case of delayed CL. Chylothorax after esophagectomy is not common but is indeed a serious complication, with a reported incidence of 1% to 9% [6,7]. CL is associated with considerable morbidity, mainly related to pneumonia with respiratory failure. Clinical management requires a prompt assessment of patients’ general condition, since they often become severely malnourished and immunocompromised due to massive loss of immunoglobulins, lymphocytes, proteins, and fat [8]. Recent studies reported mortality rates as high as 20% [9]. CL is generally considered a milky chest drain output that commonly occurs within the first week after surgery, although cases have been identified up to 13 days postoperatively [7]. Diagnosis is confirmed by triglyceride levels higher than 110 mg/dL after starting enteral feeding. Postoperative CL is mainly due to TD or damage to its tributaries or ligate failure during esophagectomy in the abdominal or thoracic phase. In addition, CL can result from damage to the cisterna chyli during nodal dissection around the celiac trunk. To prevent chylothorax, several approaches for thoracic duct identification and ligation during esophagectomy have been identified. Although some esophageal surgeons advocate for routine supradiaphragmatic thoracic duct ligation [10], others recommend excluding thoracic duct dissection as a standard procedure during esophagectomy [11]. The open transthoracic Lewis-Tanner operation seems to be associated with higher risk of thoracic duct injury in comparison to the trans-hiatal operation or to minimally invasive procedures [12]. In the present patient, identification, closure with clips, and division of the TD was carried just above the right diaphragm. Clipping the duct seems to be a perfectly reasonable thing to do and CL could be attributed to damage to tributaries of the main duct rather than injury to the thoracic duct per se or slippage of the clips [13]. An increased output (>13.5 mL/kg/day) seems to be linked to an increased incidence of CL from main thoracic duct [14], but in our patient there was no high postoperative chest drain output, thus suggesting a true delayed CL.

The Esophagectomy Complications Consensus Group categorized CL into 3 types, depending on treatments (I, enteric dietary modifications; II, total parenteral nutrition; and III, interventional or surgical therapy) and 2 types based on severity (A, <1000 mL output per day; and B, >1000 mL output per day) [15].

CL management is highly variable and there is no consensus on a predefined algorithm. Traditionally, conservative management based on dietary modification (fat-free diet) and drainage is the first-line treatment for low output (<1000 mL/day) of chylothorax. Some studies suggest the adjuvant use of somatostatin analogs, such as octreotide and etilefrine, with or without pleurodesis [16,17]. Wemyss et al [4] based the therapeutic approach on daily drain output. They reported a success rate of 65.6% with a 7-day low-fat diet in CL <500 mL/day. Drain output of >1000 mL is an indication for surgical intervention at any time point [18–20].

When conservative management fails, radical options are necessary. Most studies did not specify objective criteria for treatment failure or indications for treatment escalation, while others advocated a weight-based quantification of chyle output (eg, mL/kg), or applied the same threshold to all patients (eg, >1000 mL).

Pleurodesis is a relatively common procedure. Among the most commonly used agents that have shown effectiveness in pleurodesis are tetracycline, minocycline, bleomycin, OK-432, povidone iodine, and talc [21]. One study trialed platelet-rich fibrin glue pleurodesis in 26 patients following conservative management, with a success rate of 100% after 2 procedures [22].

Overall, surgical therapy is successful in 67–100% of cases [23]. Thoracoscopic or thoracotomy direct thoracic duct ligation are popular and successful means to achieve postoperative CL resolution. Lai et al [24] described thoracic duct ligation below the inferior pulmonary vein at the diaphragmatic hiatus encircling all the lympho-fatty tissue located along the pleura, aorta, and spine, including the azygos vein. However, multiple leaks can occur concomitantly at different sites, primarily from one of the tributaries. Therefore, ligation proximal to the aortic hiatus is preferred when a leak site is not well visualized. Pre- or intraoperatively butter, cream, or methylene blue administration can facilitate direct visualization of the leak site. Clip ligation is the most common procedure performed even if fibrin glue application to the leak site appears effective when the exact leak site cannot be identified [25]. If surgery fails and chylothorax becomes intractable, a pleuroperitoneal or pleuro-venous shunt could be considered as a last resort. Two different types of shunts are available: the active Denver pleuroperitoneal shunt and the passive Le Veen shunt, which is most commonly used [26]. The shunt is implanted in the chest and the catheter is used to transport chyle into the peritoneal cavity, where its components can be absorbed. Pleuroperitoneal shunts should not be overused, as complications are significant, including shunt occlusion, infections, bacterial translocation skin erosions, shunt relocation, and pneumoperitoneum [27,28].

Thoracic duct embolization (TDE) has become a treatment alternative for CL with high-output chylothorax, given its high success rate and minimal complications. In 1998, Cope et al [29] described TDE for first time. Lymphangiography before TDE detects the lymphatic leakage and is helpful to visualize the anatomy of lymphatic vessels before embolization. After identification of the cause of CL (extravasation or obstruction), embolization of the TD is performed proximally. N-Butyl cyanoacrylate (n-BCA) (Truefill1 Cordis, Johnson & Johnson, Warren, NJ) diluted 1: 1 in Ethiodol is used for embolization. The use of fibrin glue and cyanoacrylate has been successfully employed in the treatment of various sinuses and fistulas [30]; the TD is filled with the glue mixture proximal to the leak or occlusion. Clinical success rates of this technique range from 70% to 90%, and complications are rare, including the transgression of the needle through adjacent organs, chronic lower-limb swelling, chronic diarrhea, and ascites [31,32].

Conclusions

TDE is already described in the literature as a good alternative to surgery and the first-line treatment of high-output (>1000 mL/dL) CL. More research is needed on delayed CL, which is rare and less described. TDE can be an effective first-line approach to delayed CL, obtaining good results with a minimally invasive technique.