Thoracic Duct Embolization for Nontraumatic Chylous Effusion. Gregory J. Nadolski, MD ; and Maxim Itkin, MD

Transcription

1 CHEST Thoracic Duct Embolization for Nontraumatic Chylous Effusion Experience in 34 Patients Gregory J. Nadolski, MD ; and Maxim Itkin, MD Original Research DISORDERS OF THE PLEURA Background: Thoracic duct embolization (TDE) is an acceptable alternative procedure for treating traumatic chylothorax. The purpose of this study is to demonstrate efficacy of TDE in treating nontraumatic chylous effusions. Methods: A retrospective review of 34 patients was conducted assessing technical and clinical success of TDE for nontraumatic chylous effusions. Results: Thirty-four patients (mean age, 59 years; 27 female patients) with nontraumatic chylous effusions underwent TDE. Presentations included 21 unilateral chylothoraces (61.8%), nine bilateral chylothoraces (26.5%), two isolated chylopericardiums (5.9%), and two pleural effusions with chylopericardium (5.9%). TDE was technically successful in 24 of 34 patients (70.6%). The thoracic duct could not be catheterized in four of 34 (11.8%). Cisterna chyli was not visualized in six of 34 patients (17.6%), and, thus, TDE was not attempted. Follow-up was available for 32 patients. Four lymphangiographic patterns were observed: (1) normal thoracic duct in 17.6% of patients (six of 34), (2) occlusion of thoracic duct in 58.8% (20 of 34), (3) failure to opacify thoracic duct in 17.6% (six of 34), and (4) extravasation of chyle in 5.9% (two of 34). Clinical success varied with the lymphangiographic pattern. The clinical success rate was 16% (one of six) in cases of normal thoracic duct, 75% (15 of 20 patients) in occlusions of the thoracic duct, 16% (one of six) in cases of failure to opacify the thoracic duct, and 50% in two cases of chyle extravasation. Lymphangiography alone cured two patients (6.5%). Conclusion: TDE was most successful in cases of thoracic duct occlusion and extravasation. Lymphangiography is important for identifying the cause of chylous effusions and selecting patients who benefit most from TDE. CHEST 2013; 143(1): Abbreviations: TDE 5 thoracic duct embolization Nontraumatic chylothorax is a relatively rare condition in which the intestinal lymph (chyle) leaks into the pleural cavity. The causes of nontraumatic chylothorax include malignancy, 1 congenital diseases, Manuscript received March 1, 2012; revision accepted June 4, Affiliations: From the Department of Radiology, Section Interventional Radiology, Hospital of the University of Pennsylvania, Philadelphia, PA. Funding/Support: The authors have reported to CHEST that no funding was received for this study. Correspondence to: Maxim Itkin, MD, Department of Radiology, Section Interventional Radiology, Hospital of the University of Pennsylvania, 3400 Spruce St, Philadelphia, PA 19104; itkinmax@gmail.com 2013 American College of Chest Physicians. Reproduction of this article is prohibited without written permission from the American College of Chest Physicians. See online for more details. DOI: /chest idiopathic diseases, systemic diseases (eg, systemic lupus erythematosus, Behçet disease), and infection (eg, TB). 2 Malignancy, specifically lymphoma, is responsible for the majority of cases of reported nontraumatic chylothorax.1,3,4 Traditionally, chylothorax initially has been treated conservatively with diet modification (eg, low fat diet, nothing by mouth). Thoracentesis, pleurodesis, and surgical management (eg, thoracic duct ligation, tube drainage) have been performed for cases that fail conservative management. Percutaneous treatment of the chylothorax was developed as a minimally invasive alternative to surgical treatment. 5,6 The treatment consists of diagnostic pedal lymphangiography followed by transabdominal catheterization of the cysterna chyli or lumbar lymphatics Original Research

2 with embolization of the thoracic duct proximal to the chyle leak. In addition to the minimally invasive nature of the procedure, which results in reduction of mortality and morbidity, the ability to identify the location of the chyle leak and variation in thoracic duct anatomy potentially improves the outcome. In the largest series of 109 patients with traumatic chylothorax, Itkin et al 6 reported an 88% success rate (attempted interventions) after thoracic duct embolization and needle interruption with minimal morbidity. The goal of this study is to summarize our experience with thoracic duct embolization of chylous effusions not associated with recent surgery or trauma. Materials and Methods Permission from the Institutional Review Board of the Hospital of University of Pennsylvania (protocol number ) was obtained, and a Health Insurance Portability and Accountability Act waiver was granted for this retrospective review prior to initiation of this study. Between 1998 and 2010, 34 consecutive patients with nontraumatic chylous effusions presented for pedal lymphangiogram and thoracic duct embolization ( Table 1 ). No patients were excluded from analysis. Nontraumatic chylothorax was defined as chylothorax with no evidence of recent trauma, surgery, or intervention in the thorax. All thoracic duct embolizations were performed under moderate sedation. Prophylactic IV antibiotics (cefazolin 1 g) was administered prophylactically before the procedure. Initially, a diagnostic pedal lymphangiogram of both lower extremities was performed. To opacify the pedal lymphatic vessels and facilitate lymphatic vessel dissection, isosulfan blue 1% (Lymphazurin) mixed with lidocaine 1% was injected into the soft tissues between the toes. After several minutes, a small horizontal incision was made on the dorsum of the foot, and a lymphatic duct was dissected and cannulated using a 30-gauge needle (COOK). Ethiodized oil (Ethiodol) was then injected through the needle using a dedicated lymphangiogram pump (Cordis Corporation, Johnson & Johnson Services, Inc). The amount of contrast injected depended on physician preference and ranged from 5 to 20 ml for both legs. At the completion of the injection, 20 ml of normal saline was injected into the lymphatic vessels to propel the contrast and facilitate opacification of the thoracic duct and cisterna chyli. After identification of a target lymphatic vessel (ie, cisterna chyli, the thoracic duct, or one of its tributaries), the vessel was accessed percutaneously using a 21- or 22-gauge, cm Chiba needle (COOK) via a transabdominal approach. Reasonable efforts were made to avoid access through the large intestine or aorta. A stiff in guidewire (V18 Control; Boston Scientific Corporation) was advanced through the needle into the cisterna chyli and further into the thoracic duct. A 65-cm Rapid Transit microcatheter (Cordis Corporation, Johnson & Johnson Services, Inc) was advanced into the thoracic duct over the wire. Iodinated water-soluble contrast was injected through the microcatheter to define the chyle leak. Embolization of the thoracic duct was then performed. Over the course of the study period, different embolization materials were used to occlude the thoracic duct. Embolization coils initially were used to pack the lumen of the thoracic duct and impede flow. Later, with the availability of liquid embolics, glue (Trufill; Cordis Corporation) or onyx (Covidien) alone or in combination with intravascular coils were used to occlude the thoracic duct. When used, Trufill glue was diluted 1:2.5 with ethiodol to delay polymerization; onyx was used according to the manufacturer s instructions. After the procedure, the output from pleural drainage tubes was monitored, and if decreased significantly, a food challenge test (ice cream or cream) was prescribed. If negative, the diet was advanced, and the chest drains were removed. In patients with a slow leak or without a chest tube, serial follow-up chest radiographs were used to monitor the success of the treatment. If embolization was technically successful, but failed clinically, repeat embolization was performed. Technical success was defined as catheterization and embolization of the thoracic duct and the clinical success was defined as resolution of chylous effusion with removal of chest tubes. Results A total of 34 patients (mean age, 59 years; 27 female patients) with nontraumatic chylous effusions were referred for thoracic duct embolization (TDE). Twenty-one patients presented with unilateral chylothorax (61.8%; 11 right sided, 10 left sided). Nine patients presented with bilateral chylothorax (26.5%). Two patients presented with isolated chylopericardium (5.9%) and two with pleural effusion and chylopericardium (5.9%). The etiologies of the chylous effusions are listed in Table 1. All patients had failed conservative management with dietary restriction. Two of the patients were referred following unsuccessful thoracic duct ligation and another two had failed surgical pericardial window. All patients referred for chylothoraces secondary to lymphoma had failed to respond to treatment with chemotherapy. In all patients, the pedal lymphangiogram was successful. Thoracic duct catheterization and embolization was technically successful in 24 of 34 patients (70.6%). TDE was repeated twice in four of these patients for a persistent thoracic duct leak. The thoracic duct could not be catheterized in four of 34 (11.8%). Due to nonvisualization of the cisterna chyli, thoracic duct catheterization was not attempted in six of 34 patients (17.6%). Clinical follow-up was available for all but two patients. Overall, the intention-to-treat clinical success rate was 53% (18 of 34). In the group in which the TDE was technically successful (n 5 24), the clinical success rate was 67.7% (n 5 16). Lymphangiography alone cured two patients (5.9%). The overall complication rate was 5.9% (n 5 2). Both complications were major: One was symptomatic pulmonary artery embolization with Trufill glue, and the other was an infection of the pedal lymphangiogram incision. Four lymphangiographic patterns were identified: 1. Normal thoracic duct (17.6%; six of 34 patients): The thoracic duct was observed on lymphangiogram intact with normal transition to the subclavian vein and with free passage of contrast into the subclavian vein. In this group, the clinical success rate was 16% (one of six patients). journal.publications.chestnet.org CHEST / 143 / 1 / JANUARY

3 Table 1 Summary of Patient Information Patient Sex Age, y Indication for Procedure Surgical TD Ligation Location Visualization of Leak or TD Occlusion Catheterization Success Clinical Success 1 Female 61 LAM No Right Normal TD Yes No 2 Female 85 Lymphoma No Left Normal TD Yes No 3 Female 75 Lymphoma No Left Failure to opacify TD No attempt No 4 Female 62 Idiopathic No Right/Pericardial Occlusion Yes, twice Yes 5 Female 77 Lymphoma No Bilateral Occlusion No No 6 Male 69 Waldenstrom macroglobulinemia No Left Occlusion Yes Yes 7 Male 45 Idiopathic No Left Occlusion Yes Yes 8 Female 76 Lymphoma No Left Failure to opacify TD No attempt No 9 Female 64 Idiopathic No Right Occlusion Yes Yes 10 Female 50 Behçet disease No Left Occlusion Yes Yes 11 Female 74 Idiopathic No Right Occlusion Yes Yes 12 Female 65 Idiopathic No Left Occlusion Yes Yes 13 Female 49 Lymphoma No Right Occlusion Yes Yes 14 Female 61 Idiopathic No Bilateral Occlusion No Yes 15 Female 39 Subclavian vein occlusion No Left Occlusion Yes Yes 16 Male 54 Idiopathic No Bilateral Failure to opacify TD No attempt No 17 Female 80 Idiopathic No Right Occlusion Yes No follow up 18 Female 69 Lymphoma No Left Leak Yes No follow up 19 Female 63 Idiopathic Yes Right Normal TD Yes Yes 20 Female 71 Subclavian vein occlusion No Left Occlusion No No 21 Female 74 Idiopathic No Right Normal TD Yes No 22 Female 66 Lymphoma No Bilateral Normal TD Yes No 23 Female 21 Idiopathic chylopericardium No Pericardium Occlusion No No 24 Female 13 Gorham disease Yes Right Leak Yes Yes 25 Male 68 Waldenstrom macroglobulinemia No Bilateral Normal TD Yes No 26 Male 53 Lymphoma No Bilateral Failure to opacify TD No attempt No 27 Male 43 Idiopathic No Bilateral Occlusion Yes, twice No 28 Male 52 Idiopathic No Bilateral Failure to opacify TD No attempt Yes 29 Female 33 LAM No Bilateral Occlusion Yes, twice Yes 30 Female 62 Idiopathic No Right Occlusion Yes, twice Yes 31 Female 84 lymphoma No Right Failure to opacify TD No attempt No 32 Female 72 Idiopathic No Right Occlusion Yes Yes 33 Female 9 Idiopathic chylopericardium Pericardial window Pericardium Occlusion Yes Yes 34 Female 57 Idiopathic chylopericardium Pericardial window Pericardium Occlusion Yes Yes LAM 5 lymphangiomyomatosis; TD 5 thoracic duct. 160 Original Research

4 2. Occlusion of the thoracic duct (58.8%; 20 of 34 patients): In this pattern, no contrast passed from the thoracic duct into the subclavian vein and multiple mediastinal lymphatic collaterals were visualized ( Fig 1 ). We observed the highest clinical success rate in this group of patients: 75% (15 of 20). If TDE was technically successful, the clinical success rate was 88% (14 of 16). In one of the failed cases, embolization was performed with onyx. Later, we discovered that onyx was not effective for TDE due to the nonocclusive plug this material creates in the lymphatic fluid. In the second case, the TDE was technically successful, but the patient was lost to follow-up; therefore, the clinical outcome was uncertain. 3. Failure to opacify the thoracic duct (17.6%; six of 34 patients): This pattern was observed when lymphatic flow was obstructed at the groin level or at the level of the retroperitoneal lymphatics ( Fig 2 ). In these cases, no attempt to access or to embolize the thoracic duct could be made. However, in this group, one patient was cured following lymphangiogram alone (16%; one of six patients). 4. Extravasation of chyle (5.9%; two of 34 patients): In this pattern, contrast clearly was identified leaking from the thoracic duct or its tributaries ( Fig 3 ). In cases where the leak was identified, one patient s chylous effusion resolved (50% clinical success rate); the other was lost to follow-up. Figure 1. Fluoroscopic image demonstrating injection of iodinated contrast agent into the thoracic duct via microcatheter (black arrowhead) with complete occlusion of the thoracic duct and development of multiple collaterals (black arrow). Figure 2. Fluoroscopic image demonstrating complete occlusion of the caudal lymphatic flow at the level of the midabdomen (black arrow). This patient had a history of previously treated lymphoma. Discussion Nontraumatic chylothorax and chylopericardium are rare conditions caused by the leakage of the intestinal lymph (chyle) into the chest and pericardial cavities. The leak can originate from the chest or abdominal cavity. 2 The causes of the leak can be idiopathic, malignant (eg, lymphoma), lymphatic vessel disease (eg, Gorham disease, lymphangiomyomatosis), systemic disease (eg, sarcoidodis, Behçet disease), or congenital. In our series, the most common causes of chylothorax were idiopathic (42%; n 5 14) and lymphoma (27%; n 5 9). Traditionally, chylothorax is initially managed conservatively by treating the underlining disease, 7 repeated thoracentesis,8 and dietary modification. 9 If this fails, surgical management with thoracic duct ligation, 10 pleurodesis, 8 pleuroperitoneal or pleurovenous shunts/ pumps, 11 or lung transplantation may be attempted. TDE is a percutaneous treatment of chylothorax, which was developed as a minimally invasive alternative to surgical interventions. 5,6 TDE consists of a diagnostic pedal lymphangiography followed by transabdominal catheterization of the cisterna chyli/thoracic duct and embolization of the thoracic duct proximal to the chyle leak. The minimally invasive nature of the procedure results in reduced mortality and morbidity compared with surgical interventions. Additionally, the ability to identify the chyle leak and variations in thoracic duct anatomy potentially improves the journal.publications.chestnet.org CHEST / 143 / 1 / JANUARY

5 Figure 3. Fluoroscopic image demonstrating extravasation of contrast (white arrow) from the cisterna chyli (white arrowhead) in patient after unsuccessful thoracic duct ligation. The white arrowhead is the distended cisterna chyli. outcome. In one series of 42 patients with traumatic and nontraumatic chyle leaks, Cope and Kaiser 5 reported a 73.8% partial response or cure rate after TDE and needle interruption with no morbidity or mortality. However, due to significant differences in pathophysiology, diagnosis, treatment, and outcome between nontraumatic and traumatic chylothorax, we believe these distinct entities should be reported separately. Previously we reported our experience with traumatic chylothorax 6 ; in the present study, we report our expe rience with nontraumatic chyle leaks. The main challenge in treating nontraumatic chylothorax is identifying the cause of the chyle leak (ie, the point of extravasation or occlusion of the thoracic duct with formation of multiple collaterals). Using pedal lymphangiography, we were able to identify the cause of the leak in only 65% (22 of 34) of patients. Knowing pedal lymphangiography fails to detect lymphatic leaks below the diaphragm, we assumed the sources of the leaks in patients with normal thoracic ducts were intestinal lymphatics, which communicated directly with the pleural cavity. In six patients, pedal lymphangiography failed to opacify the cysterna chyli or thoracic duct. Four of these patients presented with lymphoma and two had a remote history of retroperitoneal surgery. We believe that in these cases, lymphadenopathy, surgery, and/or radiation resulted in obstruction of the lymphatic flow. TDE could not be performed in these patients because neither the cysterna chyli nor thoracic duct was opacified. However, in this group, even without embolization, one patient was cured after lymphangiography alone. Overall, 5.9% of patients (n 5 2) had clinical resolution of the chylous effusion after pedal lymphangiography only. The therapeutic effects of pedal lymphangiography to cure chylous effusions have been well described in the literature. 12,13 Our intention-to-treat success rate was 52% (n 5 18). This compares favorably with a previous study by Maldonado et al, 14 who reported an overall success rate of 27% using a combined approach of conservative and surgical management for nontraumatic chylothorax. In the present study, when embolization was technically successful, the clinical success rate was 67% (16 of 24 patients). Identifying the cause of the leak, or, more importantly, excluding the presence of a normal thoracic duct, with pedal lymphangiography is crucial to the outcome of the procedure. In cases of normal thoracic duct (n 5 6), the success rate of embolization was only 16% compared with 73% in patients in whom distal occlusion or leakage from the thoracic duct was identified (n 5 22). Magnetic resonance thoracic ductography has been described by Okuda et al. 15 This technique may potentially identify patients with normal thoracic duct despite a chylous effusion, obviating the need for pedal lymphangiography. Further studies are necessary to confirm this hypothesis. We observed two major complications of TDE in this study: a symptomatic pulmonary artery embolization with Trufill glue and an infection of the pedal incision. In the former, glue escaped through the patent communication between the thoracic duct and the left subclavian vein. The total amount of glue used for a TDE was between 1 and 4 ml, and escape of part of this volume into the pulmonary arterial circulation is not supposed to cause symptoms in patients with sufficient lung reserve. However, in patients with depleted lung reserve, even a small amount of glue can result in significant symptoms. A similar complication was described by Itkin et al 6 in a healthy patient with chylothorax secondary to penetrating trauma. However, in that case, the pulmonary embolism was asymptomatic. To further prevent this complication from occurring, we recommend first embolizing the thoracic duct with coils immediately below the leak to create a scaffold on which the glue can polymerize. Conclusions Pedal lymphangiography is important for identifying the pathophysiology of nontraumatic chylous effusions. In nontraumatic chylothoraces with thoracic duct occlusion, TDE was clinically successful in 73% of patients. In contrast, in cases of nontraumatic 162 Original Research

6 chylothorax with a normal thoracic duct, TDE was unsuccessful in the majority of cases. In cases where the thoracic duct was not opacified due to occlusion of the pelvic or abdominal lymphatics, other diagnostic modalities, such as MRI ductography, could be considered as alternatives to pedal lymphangiography. Acknowledgments Author contributions: Dr Itkin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Dr Nadolski : contributed to data collection, drafting and approval of the final the manuscript, and served as principal author. Dr Itkin : contributed to design of the study, data collection, and drafting and approval of the final the manuscript. Financial/nonfinancial disclosures : The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article. References 1. Strausser JL, Flye MW. Management of nontraumatic chylothorax. Ann Thorac Surg ;31(6): Romero S. Nontraumatic chylothorax. Curr Opin Pulm Med ;6(4): Valentine VG, Raffin TA. The management of chylothorax. Chest ;102(2): Doerr CH, Allen MS, Nichols FC III, Ryu JH. Etiology of chylothorax in 203 patients. Mayo Clin Proc ; 80 ( 7 ): Cope C, Kaiser LR. Management of unremitting chylothorax by percutaneous embolization and blockage of retroperitoneal lymphatic vessels in 42 patients. J Vasc Interv Radiol ; 13 (11 ): Itkin M, Kucharczuk JC, Kwak A, Trerotola SO, Kaiser LR. Nonoperative thoracic duct embolization for traumatic thoracic duct leak: experience in 109 patients. J Thorac Cardiovasc Surg ;139 (3 ): O Callaghan AM, Mead GM. Chylothorax in lymphoma: mechanisms and management. Ann Oncol ;6 (6 ): Robinson CL. The management of chylothorax. Ann Thorac Surg ;39 (1 ): al-khayat M, Kenyon GS, Fawcett HV, Powell-Tuck J. Nutritional support in patients with low volume chylous fistula following radical neck dissection. J Laryngol Otol ; 105 (12 ): Sieczka EM, Harvey JC. Early thoracic duct ligation for postoperative chylothorax. J Surg Oncol ;61 (1 ): Azizkhan RG, Canfield J, Alford BA, Rodgers BM. Pleuroperitoneal shunts in the management of neonatal chylothorax. J Pediatr Surg ;18 (6 ): Matsumoto T, Yamagami T, Kato T, et al. The effectiveness of lymphangiography as a treatment method for various chyle leakages. Br J Radiol ;82 (976 ): Yamagami T, Masunami T, Kato T, et al. Spontaneous healing of chyle leakage after lymphangiography. Br J Radiol ;78 (933 ): Maldonado F, Cartin-Ceba R, Hawkins FJ, Ryu JH. Medical and surgical management of chylothorax and associated outcomes. Am J Med Sci ;339 (4 ): Okuda I, Udagawa H, Takahashi J, Yamase H, Kohno T, Nakajima Y. Magnetic resonance-thoracic ductography: imaging aid for thoracic surgery and thoracic duct depiction based on embryological considerations. Gen Thorac Cardiovasc Surg ; 57 ( 12 ): journal.publications.chestnet.org CHEST / 143 / 1 / JANUARY