131: Percutaneous Therapy for Traumatic Chylothorax

Iatrogenic disruption of the thoracic duct is an uncommon but potentially serious complication of thoracic surgery, particularly esophagectomy.¹ The thoracic duct conveys chyle and lymph from the liver, intestines, abdominal wall, and lower extremities into the systemic venous circulation, and depending on diet and activity, the flow of chyle through the thoracic duct can reach several liters per day.² Among other components, this fluid contains essential proteins, lipids, and lymphocytes. The clinical sequelae of unremitting chylous effusions can be severe and life-threatening, including immunosuppression, respiratory compromise, dehydration, cachexia, and death.

Conservative management of chylous pleural effusions includes chest tube drainage, cessation of oral intake, and institution of total parenteral nutrition to decrease the physiologic production of chyle. If lymph production can be minimized, low-output leaks of less than 500 mL per day can sometimes heal spontaneously, although this process may take several weeks, during which time the patient may become nutritionally and immunologically depleted. As a result, some have called for earlier diagnosis and intervention to avoid metabolic compromise.^3,4 Repeat thoracotomy and direct thoracic duct ligation typically are performed early to stop high-output leaks; however, the morbidity and mortality of reoperation in this patient population is a serious consideration, with complication rates approaching 40%.⁵

Recently, percutaneous thoracic duct embolization (TDE) has been introduced as a minimally invasive technique for controlling high-output chylothorax (Fig. 131-1).⁶ In this chapter, we review the indications, preprocedural assessment, and techniques of percutaneous TDE.

The method of TDE for the control of chylothorax requires a comprehensive understanding of the anatomy of the lymphatic system, especially of the thoracic duct and cisterna chyli and their many anatomical variations. TDE is a two-stage procedure, beginning with pedal lymphangiography which is used to visualize the cisterna chyli or dominant upper lumbar lymphatics for possible cannulation. Alternatively, pelvic intranodal lymphangiography can be employed. Once a suitable retroperitoneal target has been identified, the thoracic duct is accessed percutaneously and embolized, typically from a right anterior oblique transabdominal approach to avoid the aorta, or a right posterior oblique transhepatic approach. If the thoracic duct cannot be cannulated, maceration of the cisterna chyli and upper lumbar lymphatics is undertaken to divert the flow of chyle into the retroperitoneum. One variation of TDE entails retrograde cannulation of the thoracic duct via its ostium near the left angulus venosus, using a coaxial catheter system delivered from a left brachial or basilic vein approach. The retrograde transvenous approach is less reliable, however, owing to the difficulties encountered in locating and seating a catheter in the ostium of the thoracic duct under fluoroscopy and the complexity of passing a wire and microcatheter through its competent terminal valves.

The primary indication for TDE is an incessant chylous pleural effusion arising in the setting of suspected traumatic thoracic duct disruption. TDE is generally ineffective for managing nontraumatic chylothorax secondary to infiltrative, obstructive, or malignant disease. True chylous effusions must also be distinguished from pseudochylous effusions, which can result from tuberculosis or rheumatoid disease. A true chylous effusion will contain chylomicrons and exhibit triglyceride levels greater than 110 mg/dL, whereas a pseudochylous effusion will have cholesterol as the dominant lipid component (greater than 200 mg/dL) and chylomicrons will be absent.^7,8

Preoperative imaging is critical to determining the safest approach because of the close proximity of the cisterna chyli to the right renal artery and aorta. Thin-slice, fat-suppressed, heavily T2-weighted MR imaging of the thoracolumbar region in the coronal and axial planes is used to localize the cisterna chyli and determine its relationship to adjacent structures (Fig. 131-2A). It is also important for the patient to have a normal coagulation profile because the needle will traverse many abdominal and retroperitoneal structures on the way to its periaortic target.

TDE is a 4- to 6-hour event and can be separated into two distinct procedures. A lymphangiogram is performed first to visualize the thoracic duct (see Fig. 131-2B). Once opacified, the thoracic duct is cannulated and embolized. The transit time for oil-based contrast material to travel from the dorsal foot (i.e., lymphangiogram site) to the upper abdomen is extremely variable and is the rate-limiting step for the procedure.

Pedal Lymphangiography

The patient is positioned supine on a fluoroscopy table, and both feet are sterilely prepped. Moderate procedural sedation is employed for patient comfort. Prophylactic intravenous antibiotics are administered.

Because the cisterna chyli is more commonly located in the right upper abdomen, right pedal lymphangiography generally provides for more efficient and definitive opacification of this structure, but either or both feet may be used. Standard pedal lymphangiography is performed by injecting 0.25 to 0.5 mL of methylene blue dye between the web spaces of the toes. The dye is taken up by the lymphatics, which can be visualized as blue streaks under the skin. After 1% lidocaine local anesthesia, a small incision is made on the dorsum of the foot, and a lymphatic vessel is isolated and cannulated with a 30-gauge needle lymphography catheter. The lymphatic is secured to the needle with silk ties, and a gentle test injection is performed with a 3-mL saline syringe to assess for leaks. The lymphography catheter is then connected to an injector, and iodized oil (Lipiodol, Laboratoire André Guerbet, Aulnay-sous-Bois, France) is infused at a rate of 8 to 12 mL per hour to a maximal administered volume of 20 mL. During infusion, serial spot radiographs are obtained of the lower extremity, pelvis, and abdomen to track the cephalad opacification of the lymphatic system (Fig. 131-3). Transit times from the foot to the cisterna chyli are typically 1 to 3 hours. A saline bolus can be administered behind the column of iodized oil, if necessary, to speed its transit through the lymphatics. Once the thoracic duct is visualized, the infusion can be stopped and attention turned to the cisterna chyli and/or dominant lumbar lymphatics for needle cannulation or disruption. The foot incision is closed with 3-0 Prolene vertical mattress sutures and covered with a sterile dressing.

Thoracic Duct Cannulation

Once the cisterna chyli is visualized, the patient's abdomen is prepped from the inferior costal margin to just below the umbilicus. Planning a percutaneous approach requires careful examination of the preoperative MRI. The position of the aorta and right renal artery are referenced to bony landmarks on the scan. A slightly inferior right anterior oblique approach is used most often to avoid these two vascular structures. The approximate starting position is 3 to 5 cm to the right of midline just below the costal margin. After local anesthesia, a small dermatotomy is made, and a 21- or 22-gauge, 15- to 20-cm Chiba needle (Cook Medical, Inc., Bloomington, IN) is directed toward the cisterna chyli using a “gun-site” technique (Fig. 131-4). This technique helps to minimize drifting of the needle as it travels through visceral structures. Just before the needle reaches the cisterna chyli, the C-arm is placed into an anteroposterior position. The needle can be seen “tenting” the lymphatic channels. Entry is made with a brisk and deliberate motion. A 0.018-in, 150-cm stiff guidewire (V-18 Control wire, Boston Scientific, Natick, MA) is advanced into the duct (Fig. 131-5). The needle is exchanged for a microcatheter. Predilation with a 4-French inner dilator and stiffening cannula from a nonvascular access kit (MAK-NV Introducer System, Merit Medical Systems, Inc., South Jordan, UT) is employed if the microcatheter fails to track appropriately. Given the likelihood of transintestinal passage, it is recommended that the access system be kept as small as possible, ideally 3 to 4 French. Direct hand injection of iodinated contrast medium is then performed to locate the site of thoracic duct injury. The point of extravasation should be carefully documented with imaging; in the event TDE is unsuccessful, precise localization of the injury will help guide subsequent surgical en bloc ligation.

Embolization Procedure (a.k.a. Type 1 TDE)

The microcatheter is positioned up to, or ideally across, the site of thoracic duct injury. Embolization is performed by depositing 4- to 6-mm diameter platinum-fibered microcoils (Cook Medical, Inc., Bloomington, IN) along the entire length of the thoracic duct to within a few centimeters of the entry site. As the final step, a 2:1 mixture of ethiodized oil and N-butyl cyanoacrylate with 0.5- to 1-g of tantalum powder (components included in Trufill n-BCA Liquid Embolic System, Codman & Shurtleff, Inc., Raynham, MA) is injected to seal off the lower thoracic duct and lymphatic entry site (Fig. 131-6).

Needle Disruption of Lymphatics (a.k.a. Type 2 TDE)

If cannulation of the thoracic duct is not possible, serial needle passes are made to intentionally disrupt the cisterna chyli and adjacent lymphatics. Completion is indicated by the extravasation of contrast in the retroperitoneal space. The needle disruption technique results in the diversion of chyle into the retroperitoneum and decompression of the thoracic duct, allowing the intrathoracic injury to heal.

Careful monitoring of the chest tube output is important for determining the success of the procedure. Chest tube output may drop rapidly, but typically, there is a gradual decline of approximately 50% each day. The chest tubes are removed when output is less than 200 mL per day and the patient passes an oral dietary challenge. Routine wound care and dressing changes are required for the pedal cutdown site until the sutures are removed 10 days postprocedure.

Although solid organs and bowel are traversed with the access system en route to the cisterna chyli and thoracic duct, to date, there have been no reported complications related to this approach. Standard complications of pedal lymphangiography include permanent tattooing of the skin with methylene blue dye, wound infection, and allergic reactions. Iodized oil-related pulmonary complications include hypoxia and chemical pneumonitis. Asymptomatic pulmonary embolism from liquid embolic adhesive injected into the thoracic duct has been reported. Overall, early complication rates are very low (3%).⁹

Delayed complications related to TDE are certainly conceivable, given the intentional redistribution of lymphatic flow induced by the procedure. A recent retrospective survey of 169 patients who underwent TDE uncovered several instances of chronic leg swelling, abdominal swelling without evidence of anasarca or ascites, and chronic diarrhea that were classified as “probably related” to the procedure (14.3% overall).¹⁰ This idea is supported by reports of chronic chylous ascites and peripheral lymphedema as possible long-term complications of surgical thoracic duct ligation,^11–13 and protein losing enteropathy after thoracic duct ligation in animal models.¹⁴ A prospective randomized investigation would be required to assess the strength of these apparent relationships.

Since traumatic chylothorax is a relatively uncommon postsurgical complication, the published data on outcomes after TDE are limited, but highly supportive of the technique. The largest case series to date reviewed the results of 109 patients who underwent TDE and reported an overall clinical success rate of 71%,⁹ comparing favorably to the success rate of 73.5% reported in the original case series of 42 patients.⁶ Importantly, the success rate for type 1 TDE in the larger series was significantly higher than for type 2 needle disruption (90% vs. 72%), and the use of both coils and liquid embolic adhesive for type 1 TDE was observed to be more effective than the use of coils alone (91% vs. 84%). Overall, TDE is a well-tolerated, minimally invasive, and effective alternative to the reoperation for patients with unremitting traumatic chylothorax.

A 50-year-old man developed a high-output (2 L per day) chyle leak after esophagectomy. The patient did not respond to conservative management and was referred for TDE.

A pre-embolization MRI shows the cisterna chyli (Fig. 131-7). After a right pedal lymphangiogram, the cisterna was opacified with an oil-based contrast agent. The puncture site was located just to the right of midline, and the image intensifier was positioned so that the needle overlay the target lymphatic structure. The “gun-site” technique is illustrated in Figure 131-8. The lymphatic system was cannulated (Fig. 131-9), and iodinated contrast was injected to define the site of thoracic duct injury (Fig. 131-10). The leak then was embolized and sealed using platinum-fibered microcoils and liquid embolic adhesive (Fig. 131-11).

Postoperatively, the effusions decreased by 50% each day, and the chest tubes were removed 3 days later. The patient was discharged home soon thereafter.

An important principle in management of lymph leaks is the notion of relative resistance. The lymph system is generally a “low-pressure” system that responds to subtle pressure differences. If a lymph drain provides a low-resistance option to lymph flow, the output is likely to continue. Alternatively, graded increases in resistance are likely to encourage lymph flow through alternative pathways. One of the functional consequences of thoracic duct “embolization” may be that it changes the dynamics of lymph flow.