Thoracic Duct-to-Azygous Vein Lymphovenous Anastomosis for Lymphatic Decompression: Initial Experience and Early Outcomes

Abstract

Background

Thoracic duct obstruction can cause devastating lymphatic complications, and definitive treatments remain lacking. This investigation evaluated our initial experience and early outcomes using a surgically created thoracic duct-to-azygous vein lymphovenous anastomosis (TDA-LVA) to provide lymphatic decompression in critically ill neonates and infants with thoracic duct obstruction.

Methods

All children who underwent TDA-LVA creation at the Children’s Hospital of Philadelphia between December 2019 and January 2024 were retrospectively reviewed. Preoperative and postoperative clinical characteristics were compared among all patients. Primary end points of analysis included survival to discharge and at last follow-up.

Results

Eight children (median weight, 4.5 kg; interquartile range [IQR], 3.5-5.0 kg) underwent successful TDA-LVA creation. All procedures were tolerated very well. One patient (1 of 8 [12.5%]) required chest washout on postoperative day (POD) 7, and 1 TDA-LVA was unsuccessfully revised on POD 21. The median hospital length of stay after TDA-LVA creation was 138 days (IQR, 64-241 days). Overall, 6 of 8 patients (75.0%) survived to hospital discharge, all of which (6 of 6 [100.0%]) remained alive at a median follow-up of 165 days (IQR, 18-521 days) after discharge. Postoperative TDA-LVA patency was confirmed with conventional lymphangiography in 5 of 8 patients (62.5%), and up to POD 70.

Conclusions

A TDA-LVA can restore lymphatic fluid drainage into the systemic venous circulation and remain patent for >2 months in carefully selected patients with thoracic duct obstruction. Further investigations are needed to determine the long-term durability and potential impact on overall quality of life and survival in children with complex lymphatic disorders.

In Short.

• ▪The surgical creation of a thoracic duct-to-azygous vein lymphovenous anastomosis can restore lymphatic drainage into the systemic venous circulation and remain patent for >2 months in critically ill children with thoracic duct obstruction.
• ▪A thoracic duct-to-azygous vein lymphovenous anastomosis may provide a unique opportunity to decompress the lymphatic circulation in select patients with complex lymphatic disorders.

The thoracic duct is the largest lymphatic vessel in the body and typically drains into the systemic venous circulation through a lymphovenous junction near the confluence of the left jugular and subclavian veins.¹ Congenital or acquired etiologies of thoracic duct obstruction can lead to pathologically elevated pressures within the lymphatic circulation that cause severe, refractory, and life-threatening complications (anasarca, respiratory failure, effusions, and electrolyte imbalances, etc).²

Definitive treatments for thoracic duct obstruction are lacking, and supportive management strategies typically center on decreasing lymphatic fluid production (low-fat diets, interrupting enteral feeds, diuresis, etc) and a limited number of interventions that aim to reduce lymphatic leaks.1, 2, 3, 4, 5 Although there is growing evidence that other novel percutaneous and surgical interventions may be able to effectively decompress the lymphatic circulation in select patients with complex lymphatic disorders, experiences remain limited, and the long-term durability of these interventions is unclear.1, 2, 3, 4, 5

This investigation reports our initial experience and early outcomes using a surgically created thoracic duct-to-azygous vein lymphovenous anastomosis (TDA-LVA) to provide durable decompression of the lymphatic circulation in critically ill neonates and infants with thoracic duct obstruction.

Patients and Methods

All children who underwent TDA-LVA creation at the Children’s Hospital of Philadelphia between December 2019 and January 2024 were retrospectively reviewed with Institutional Review Board approval. Primary end points of analysis included survival to hospital discharge and at last follow-up. Clinical characteristics were analyzed the week before and after TDA-LVA creation.

Statiscial analysis

Categorical variables are presented as frequencies with percentages, and continuous variables are presented as medians with an interquartile range (IQR). Study variables were compared using Fisher exact or Wilcoxon rank sum tests, respectively.

Surgical Technique

All patients underwent extensive preoperative lymphatic imaging.^2,4,5 After obtaining informed consent, the thoracic duct was percutaneously accessed, as described previously.⁴ Next, a right thoracotomy was performed to expose the posterior mediastinum. The thoracic duct was dissected free and ligated cephalad to the site of the intended anastomosis, and the caudal aspect was occluded with a microvascular clamp (Micrins) or bulldog clamp (V. Mueller) before transection. The free lumen of the thoracic duct was then prepared further for the anastomosis (Figure 1; Video 1).

Figure 1.

(A) The thoracic duct is dissected under a high-powered surgical microscope. (B) It is imperative to preserve surrounding adventitia to maintain its integrity when creating the thoracic duct (TD)-to-azygous (Az) vein lymphovenous anastomosis (LVA) using an (C) end-to-end or (D) end-to-side technique.

TDA-LVAs were created under a high-powered surgical microscope (Mitaka) with multiple 9-0 or 10-0 nylon sutures using an end-to-end or end-to-side technique, depending on the anatomy (Figure 1; Videos 2, 3). The completed TDA-LVA was inspected, and if needed, manual pressure was gently applied until adequate hemostasis was achieved to avoid placing additional stitches that could potentially obstruct the microvascular anastomosis. Before chest closure, TDA-LVA patency was confirmed with a pinch test (Video 4) and/or injection of blue dye (Video 5).

Results

In total, 8 neonates and infants underwent TDA-LVA creation, most of which had complex congenital lymphatic disorders (5 of 8 [62.5%]) and/or congenital heart disease (3 of 8 [37.5%]) (Supplementary Table). The thoracic duct obstruction occurred in 1 patient (1 of 8 [12.5%]) after decannulation from extracorporeal membrane oxygenation, and 2 of 8 patients (25.0%) had recently undergone cardiac surgery. None of the patients had a surgical history of thoracic duct ligation.

The median age and weight at time of TDA-LVA creation was 123 days (IQR, 97-179 days) and 4.6 kg (IQR, 3.5-5.0 kg), respectively. The median hospital length of stay before TDA-LVA creation was 42 days (IQR, 24-73 days). All patients had anasarca and were supported with mechanical ventilation during the week before TDA-LVA creation.

All TDA-LVAs (8 of 8 [100.0%]) were successfully created with a median operative duration of 318 minutes (IQR, 280-352 minutes). All procedures were tolerated very well, and clinical characteristics did not change significantly during the first postoperative week compared with the week before TDA-LVA creation (Table 1). Chest tube output during the first postoperative week was often minimal (15 mL/d [IQR, 9-37 mL/d]; 3 mL/kg/d [IQR, 3-6 mL/kg/d]); however, 1 patient (1 of 8 [12.5%]) required chest washout on postoperative day (POD) 7 due to a progressive hemothorax.

Table 1.

Patient Characteristics One Week Before and After Thoracic Duct-to-Azygous Vein Lymphovenous Anastomosis Creation

Variable	1 Week Before TDA-LVA Creation	1 Week After TDA-LVA Creation	P Value
Weight, kg	4.4 (3.4-5.3)	4.8 (3.5-5.8)	.782
Urine output, mL/kg/d	60.7 (48.3-78.9)	77.0 (53.4-92.3)	.405
Diuretic dose, mg/kg/d	0.73 (0.64-1.32)	0.72 (0.49-1.81)	.632
Fraction of inspired oxygen, %	27 (24-32)	34 (30-36)	.086
Respiratory rate, min−1	40 (36-44)	43 (39-46)	.705
Positive end-expiratory pressure, cm H2O	8 (6-8)	9 (5-9)	.715
Peak inspiratory pressure, cm H2O	28 (22-29)	31 (27-36)	.272
Mean airway pressure, cm H2O	13 (10-13)	15 (11-17)	.303
Dopamine dose, μg/kg/min	0.00 (0.00-0.75)	0.00 (0.00-2.25)	.873

Data are presented as median (interquartile range).

TDA-LVA, thoracic duct-to-azygous vein lymphovenous anastomosis.

The median follow-up after TDA-LVA creation among all patients analyzed was 281 days (IQR, 115-561 days), and the median hospital length of stay was 138 days (IQR, 64-241 days) (Table 2). Overall, 6 of 8 patients (75.0%) survived to hospital discharge. The 6 who survived to discharge remained alive at last follow-up (165 days [IQR, 18-521 days] after discharge, and 391 days [IQR, 174-748 days] after TDA-LVA creation; Table 2).

Table 2.

Hospital Course and Survival Outcomes

Variable	Data Value
Hospital length of stay after TDA-LVA creation, d, N = 8	138 (64-241)
Follow-up after TDA-LVA creation, d, N = 8	281 (115-561)
Survival to hospital discharge, n/8 (%)	6 (75.0)
Died before discharge, n/8 (%)	2 (25.0)
Follow-up after hospital discharge, d, N = 6	165 (18-521)
Remain alive at last follow-up, n/6 (%)	6 (100.0)

Data are presented as median (interquartile range) or as n/N (%). TDA-LVA, thoracic duct-to-azygous vein lymphovenous anastomosis.

Postoperative TDA-LVA patency was confirmed using conventional lymphangiography in 5 of 8 patients (62.5%) at a median 40 days (IQR, 20-64 days) after surgery (Figure 2; Video 6). One additional patient (1 of 8, 12.5%) with recurrent postoperative pleural effusions and anasarca after initial improvement was found to have mildly elevated gradients across the anastomosis (4 mm Hg) on POD 19, which was unsuccessfully revised on POD 21. Meanwhile, one TDA-LVA (1 of 8 [12.5%]) was found to be occluded on POD 28, and the patient ultimately died in the hospital on POD 249 due to progression of their underlying disease. One additional patient (1 of 8 [12.5%]) died in the hospital on POD 26 due to complications unrelated to the TDA-LVA procedure.

Figure 2.

(A) Conventional lymphangiography and (B) the magnified inset performed on postoperative day 70 confirming patency of a surgically created thoracic duct-to-azygous vein lymphovenous anastomosis. Brisk, free-flowing contrast is observed passing through the thoracic duct and anastomosis (arrow) into the azygous vein (∗). (C) Preoperative magnetic resonance lymphangiography in the same patient demonstrated pronounced pulmonary infiltrates (>), which (D) markedly improved on subsequent imaging acquired on postoperative day 70.

Comment

This investigation describes our initial experience and early outcomes using a TDA-LVA to surgically decompress the thoracic duct and restore lymphatic fluid flow into the systemic venous circulation in 8 neonates and infants with congenital or acquired etiologies of thoracic duct obstruction. Importantly, 75% (6 of 8) of patients who underwent TDA-LVA creation survived to hospital discharge, all of whom remained alive at a median follow-up of >1 year after surgery (Table 2). Additionally, this study provides definitive, lymphangiographic evidence that a surgically created LVA between the thoracic duct and systemic venous circulation can restore and maintain unidirectional lymphatic fluid flow into the systemic venous circulation for >2 months (Figure 2; Video 6). At the same time, TDA-LVA patency was accompanied by objective interval improvement in lymphatic perfusion abnormalities (Figure 2). Together, these findings provide important evidence that a surgically created LVA between the thoracic duct and central venous circulation may be able to provide durable lymphatic decompression in select patients with complex lymphatic disorders, including congenital and acquired etiologies of thoracic duct obstruction.1, 2, 3, 4, 5

Importantly, the main objective for creating a TDA-LVA was to provide surgical palliation for critically ill children facing improbable survival. Although the current study demonstrates that a TDA-LVA may indeed offer a safe, effective, and durable therapeutic option for carefully selected patients with complex lymphatic disorders that fail to respond to conventional treatments, it must be noted that the surgical creation of a TDA-LVA is a technically challenging and complex operation that may not be anatomically or physiologically appropriate for every patient. Specifically, individuals in whom lymphatic overload and congestion has developed due to chronically elevated central venous pressures (eg, single-ventricle congenital heart defects) or superior vena cava obstruction are unlikely to benefit from a TDA-LVA because rerouting lymphatic fluid into the azygous vein alone would not adequately offload their lymphatic circulation.^1,2 However, other novel interventions could potentially decompress the lymphatic circulation in these individuals, including a thoracic duct-to-pulmonary vein LVA, which warrant further investigation.^1,2,6

We recently reported that select individuals with complex lymphatic disorders may benefit from surgically created lymphocutaneous fistulas, which often drain more than ∼0.5 to 1.5 L of lymphatic fluid a day (∼100-300 mL/kg/d) and can lead to significant weight loss, improved edema, and decreased ventilatory support requirements.⁷ However, lymphocutaneous fistulas do not provide a cure.⁷ To some degree, this is because the precipitous loss of such large volumes of lymphatic fluid is not sustainable and may contribute to a variety of other complications that can be particularly challenging to manage in neonates and infants (eg, electrolyte imbalances, nutrient deficiencies, and immune compromise, etc).

In contrast, patients who underwent TDA-LVA creation had minimal chest tube output after surgery, as expected, and their weights and ventilatory support requirements did not significantly decrease in the early postoperative period (Table 1). Nevertheless, most patients improved overtime, which is underscored by the finding that 75.0% (6 of 8) survived to hospital discharge and remained alive at a median follow-up of >5 months after discharge (Table 2). Together, these findings further underscore that a TDA-LVA may present a unique opportunity to decompress the lymphatic circulation in select individuals with congenital or acquired etiologies of thoracic duct obstruction (Figure 2; Video 6).

In summary, the current study, in conjunction with other recent reports, provides compelling evidence that the surgical creation of a LVA between the thoracic duct and central venous circulation may benefit select individuals with complex lymphatic disorders.1, 2, 3, 4, 5 However, experiences remain limited, and additional studies are warranted to determine the long-term durability and potential impact on overall quality of life and/or survival. Moreover, several obstacles must be overcome before TDA-LVAs can be routinely applied in clinical practice, including further optimization of postoperative management strategies.^1,2,6 In this respect, our protocol for postoperative anticoagulation and antiplatelet therapy has evolved over the years. We now routinely initiate a heparin infusion within 24 hours after TDA-LVA creation and continue therapeutic anticoagulation with enoxaparin for at least 6 months. We also administer an intravenous antiplatelet (ie, cangrelor) throughout the early postoperative period. This is noteworthy because antiplatelet agents may have particular importance after the surgical creation of an LVA, as it is well-established that lymphatic endothelial cells activate platelet receptors and drive hemostasis to maintain separation between lymphatic vessels and veins during development.^1,6 Although it remains unclear whether similar mechanisms may increase the risk of thrombosis in a surgically created LVA, these considerations underscore how clinically relevant translational research models may help advance the treatment of complex lymphatic disorders.^1,6 Despite these limitations, we believe that a TDA-LVA can be performed safely, efficiently, and provide an effective treatment for select individuals with thoracic duct outlet obstruction.