Chylothorax is a rare but serious complication of esophagectomy, with an incidence ranging from 0.2–10.5%, and generally due to an undetected injury of the thoracic duct[2, 3]. The thoracic duct originates from the upper portion of the cisterna and vertically enters the chest through the aortic orifice of the diaphragm, passing between the azygos vein and the aorta, and then terminates at the junction between the left subclavian vein and the jugular vein, called left venous angle. In total, the thoracic duct measures 36–45 cm in length and 2–3 mm in diameter. However, many classification systems have been described to characterize the types of variation, such as complete left-sided course, complete right-sided course, proximal and distal duplications, plexiform variation, and absence of the cisterna chyli[4–6]. One of the significant anatomical variants about proximal and distal partial duplication of the thoracic duct was previously detected, showing that the thoracic duct may be duplicated anywhere along its course, both proximally and distally, before entering the left venous angle[5]. Although thoracic duct may vary along any aspect of its course and terminate in the venous system either as single or multiple channels, lesions are mostly reported in the distal portion of it. Most of the lesions occur near the arch of the aorta and azygos, where the esophagus and the thoracic duct are the closest. In the present case, we supposed that it happened to be a distal partial duplication of the thoracic duct; a hypothesis that was confirmed by the second operation. Because the first esophagectomy and intrathoracic esophagogastric anastomosis was performed via left thoracotomy, we opted to perform the second operation through the left original incision. During the second operation, our approach was impeded by adhesions between intrathoracic gastric wall and lung. After a large amount of chylous fluid in the mediastinum was drained away, an injury on the side wall of the thoracic duct was revealed at 4 cm below the carina, with chylous fluid leaking. Even if we performed mass ligation of the thoracic duct at the supradiaphragmatic area again, and further ligated both ends of the injury site, it was still gradually leaking at the same place of the damaged thoracic duct, which was confusing for us.
Fortunately, in the further exploration, we found the hidden collateral branch at the back of the thoracic duct and supposed that this was the cause of distal partial duplication of the thoracic duct. In this variation, the thoracic duct is partially duplicated distally from a shared trunk and then joining to form a single vessel that terminates at the left jugulovenous angle. Because of exposure of the thoracic duct via left thoracotomy, we speculated that the hidden collateral branch arose from the unrevealed right-sided duplicated duct and flowing into the damaged left-sided thoracic duct. As the point of entry into the thoracic duct was opposite the injury of the side wall, it was explained that ligation of the thoracic duct was ineffective whether at the supradiaphragmatic area or at both ends of the injury. First, care must be taken to adequately ligate both trunks because of the shared proximal trunk in case of distal duplication. Second, even after the trunk of the thoracic duct had been ligated twice at different levels, and both ends of the injury had been ligated, the chylous fluid from the undetected and unligated right-sided trunk still drained into the damaged left-sided trunk via the collateral branch opposite the injury and leaked out through the injury. Finally, we ligated the collateral branch with confirmation of chylous leakage cessation.
Chylous leakage should be suspected postoperatively when the drainage is > 500 mL/d with milky features. Usually, the effusion is evident when the patient resumes oral intake. Treatment of chylous leakage can be either conservative or surgical. In this case, conservative treatment including adequate drainage, complete fasting and total parenteral nutrition was not effective, even though the injection of octreotide did not reduce chylous production. Timing of surgical treatment is still debated. Some surgeons have suggested that loss of > 1 L/d of chyle for 1 wk is evidence of failure of conservative treatment, while others wait for > 2 wk and intervene only if there are severe metabolic and nutritional imbalances[7]. In general, we prefer the former view of early surgery. However, as the patient was admitted to the ICU due to respiratory failure, we could only resort to surgery on postoperative day 34 when he was better.
The presence of anatomical variations, together with the presence of multiple small lymphatic vessel tributaries, is probably responsible for the incidence of postoperative chylothorax, irrespective of the surgeon’s meticulous attention and skill. To reduce the incidence of postoperative chylothorax, prophylactic ligation of thoracic duct has been considered an effective method, and in some cases, it would be performed regardless of the thoracic duct injury[8, 9]. However, a variety of reasons such as anatomical variation, incomplete ligation and thoracic duct wall lesion during the isolation process might be responsible for the ligation failure[10]. Thus, during surgery for esophageal cancer that involves the thoracic duct, we recommend preoperative administration of oral cream and possible ligation of any thick tributaries from the thoracic duct. However, during surgery, it can be difficult to recognize such thick tributaries in practice. In this case, perhaps lymphangiography data were useful, which provided us with the exact course of the thoracic duct.