If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Patients with cirrhosis of the liver have increased hepatic and gastrointestinal lymph flow that may contribute to the formation of ascites and pleural effusions. Increased lymph flow, which is due to postsinusoidal portal hypertension, causes a high rate of flow through the thoracic duct. Because of the high flow rates, disrupted lymphatic vessels in patients with cirrhosis of the liver may fail to close, a situation that results in chylous ascites, pleural effusions, or chylous fistulas. Chylous fistulas deplete proteins, fluid, and lymphocytes and thus lead to volume depletion and coagulopathy. Herein we describe an unusual case in which a high-output traumatic thoracic duct-cutaneous fistula developed in a patient with cirrhosis and led to volume depletion and coagulopathy. Correction of the portal hypertension with placement of a transjugular intrahepatic portosystemic shunt led to closure of the fistula and normalization of accompanvina metabolic abnormalities.
Hepatic lymph flow is regulated predominantly by the pressure gradient between the capillary and interstitial compartments within the liver. This pressure gradient increases with postsinusoidal portal hypertension associated with cirrhosis of the liver.
In addition, portal hypertension results in splanchnic venous congestion and increased gastrointestinal lymph flow. Although measurement of hepatic and gastrointestinal lymphatic flow is difficult in humans, increased hepatic and gastrointestinal lymphatic flow has been documented in patients with hepatic cirrhosis and partial hypertension.
Hepatic and gastrointestinal lymph constitutes a substantial portion of thoracic duct lymph flow, and increases in hepatic or gastrointestinal lymphatic flow should manifest as increased lymphatic flow across the thoracic duct. Indeed, patients with increased portal pressure have increased thoracic duct lymph flow, and successful treatment of portal hypertension is associated with decreases in thoracic duct flow.
It has been reported to be a rare complication of central venous catheterization. Chylothorax and chylous fistulas are the major clinical manifestations of thoracic duct injury. Treatment of thoracic duct injury depends on the degree of injury, and in severe cases, surgical ligation of the thoracic duct may be necessary; however, the subsequent lymphatic congestion of the lower extremities and gastrointestinal lymphatic system can be associated with significant morbidity.
We present an unusual case of thoracic duct injury due to iatrogenic trauma in a patient with cirrhosis of the liver that resulted in a high-output chylous fistula. The use of a transjugular intrahepatic portosystemic systemic shunt (TIPS) to decrease the portal pressure dramatically affected the output of the fistula, and the physiologic mechanism for this observation is discussed.
Report of Case
A 61-year-old woman was referred to Mayo Clinic Rochester for evaluation of chronic liver disease due to hepatitis C and persistent drainage from the catheter site in a left subclavian vein. Two years before admission, the patient underwent assessment because of long-term asymptomatic increases in serum aminotransferase values. A liver biopsy showed chronic hepatitis with cirrhosis, and antibodies to hepatitis C virus were positive. No evidence of decompensated liver disease was detected, and the patient did well until 3 months before admission when she had hematemesis due to a variceal hemorrhage. Assessment at her local hospital revealed esophageal varices; she received blood transfusions, and the varices were treated with endoscopic sclerotherapy. Two months before admission, she had another variceal hemorrhage. Two weeks before admission, she sought treatment at her local hospital because of fever, increasing cough, and shortness of breath. A large right pleural effusion and ascites were clinically evident. Chest roentgenography confirmed the presence of a large right pleural effusion, and thoracentesis was performed. Cytologic study was negative for malignant cells, and biochemical parameters were consistent with portal hypertension. A left subclavian central venous catheter was placed, and intravenous vancomycin and ceftazidime therapy was initiated for presumed spontaneous bacterial peritonitis. A fluid leak around the site of the catheter was detected within 12 hours after placement. This leak continued despite replacement of the catheter over a guidewire. The catheter was subsequently removed, but drainage continued from the puncture site. Because of the persistent drainage of a large volume of fluid, the patient was transferred to our institution.
On arrival, the patient appeared ill; her blood pressure was 70/40 mm Hg, and her pulse rate was 110 beats/min. A large amount of fluid was noted in an external drainage collection bag overlying the left subclavian vein puncture site. A physical examination revealed dullness to percussion of the right lung base, with ausculatory changes consistent with pleural effusion. Additional findings were hepatosplenomegaly, stage 2 encephalopathy, and a small amount of peripheral edema. No ascites was noted.
A right internal jugular central venous catheter was placed, and the central venous pressure was measured at 1 mm Hg. Chest roentgenography confirmed the presence of a moderate-sized right pleural effusion. Laboratory tests yielded the following results: hemoglobin, 10.1 g/dL; leukocyte count, 8.4 × 109/L (81% neutrophils, 14% monocytes, and 5% lymphocytes); platelet count, 68,000 × 109/L; so dium, 136 mEq/L; potassium, 5.1 mEq/L; chloride, 100 mEq/L; bicarbonate, 18 mEq/L; and creatinine, 1.8 mg/dL. Liver tests showed total bilirubin of 1.0 mg/dL (direct, 0.3), aspartate aminotransferase of 31 U/L (normal, 12 to 31), alkaline phosphatase of 160 U/L (normal, 108 to 282), pseudocholinesterase of 1 U/mL (normal, 11 to 25), albumin of 1.6 g/dL, total protein of 6.1 g/dL., and prothrombin time (PT) of 21.2 seconds (international normalized ratio [INR], 2.1). The partial thromboplastin time (PTT) was 47.7 seconds, D-dimer was greater than 500 ng/mL, fibrinogen was 198 mg/dL, and venous ammonia was 39 µg N/dL (normal, less than 50). Findings on an abdominal ultrasound study were consistent with cirrhosis and showed no focal hepatobiliary abnormalities or ascites.
Analysis of the fluid obtained from the drainage collection bag yielded a total protein level of 2.9 g/dL., glucose level of 181 mg/dL, leukocyte count of 2,400/mL (91% lymphocytes), and triglyceride level of 41 mg/dL with a cholesterol value of 17 mg/dL. Lipids were present in the form of plasma lipoproteins and chylomicrons. Gram's stain and culture of the fluid were negative. The patient was resuscitated with albumin infusions and crystalloids, resulting in an increase in the central venous pressure to 5 to 8 mm Hg. The increase in central venous pressure was associated with a substantial increase in the fluid in the drainage collection bag. Within 48 hours after initial resuscitation of the patient, the output of fluid was approximately 10 L/day.
The patient's condition stabilized sufficiently to allow lymphangiography to be performed on day 5 of hospitalization, findings of which demonstrated a large leak from the thoracic duct in the area near its junction with the left subclavian vein (Fig. 1). Surgical repair was planned, but because of loss of coagulation factors due to high output from the fistula, a severe coagulopathy developed. Correction of the coagulopathy was initiated, but on day 6 of hospitalization, the patient had upper gastrointestinal bleeding, resulting in a decrease in the hemoglobin concentration from 9.6 to 5.3 g/dL. At that time, the PT was 48.8 seconds (INR, 4.9), PTT was 75.8 seconds, and platelet count was 28,000 × 109/L. Esophagogastroduodenoscopy showed fresh red blood in the stomach, obliterated esophageal varices, and evidence of portal hypertensive gastropathy with mucosal oozing. The patient's coagulopathy was corrected with fresh frozen plasma, platelets, and cryoprecipitate, and she received a transfusion of 4 U of packed erythrocytes. Twelve hours after her initial bleeding had been stabilized, she had a recurrent upper gastrointestinal hemorrhage, and her hemoglobin concentration decreased from 10.9 to 5.1 g/dL. At that time, the PT was 15.9 seconds (INR, 1.6), PTT was 33.8 seconds, fibrinogen was 153 mg/dL, and platelet count was 32,000 × 109/L; findings on thromboelastography were within normal limits.
Because of continued portal hypertension-related gastrointestinal bleeding in the setting of a corrected coagulopathy, the patient underwent insertion of a TIPS on hospital day 7. Before placement of the TIPS, the free hepatic vein pressure was 0 mm Hg, and direct portal puncture revealed a portal pressure of 22 mm Hg for a portosystemic pressure gradient of 22 mm Hg. A 6.8-cm Wallstent (Schneider Inc., Minneapolis, Minnesota) was placed, bridging the hepatic and portal vein, and dilated to 10 mm in diameter. The portosystemic pressure gradient at the conclusion of the procedure was 12 mm Hg. After the procedure, no complications occurred, and no gastrointestinal bleeding was evident. On the basis of hourly measurements of drainage volume, the output from the thoracic duct fistula dramatically decreased immediately after TIPS placement (Fig. 2); 72 hours after TIPS placement, drainage from the fistula site had discontinued completely. The patient was dismissed a few days later; no fluid had reaccumulated in the area of the left subclavian vein puncture site. At that time, chest roentgenography showed near-resolution of the right pleural effusion. A follow-up assessment 6 weeks after TIPS placement demonstrated no evidence of recurrent fistula, pleural effusion, or ascites.
Increased hepatic lymph flow is a well-recognized consequence of cirrhosis of the liver with portal hypertension.
As the liver sinusoidal endothelium is fenestrated, plasma proteins equilibrate between the interstitium and vascular space. The concentrations of major proteins such as albumin in the hepatic interstitium approximate those in serum.
Therefore, plasma oncotic pressure has a minimal role in the accumulation of hepatic interstitial fluid. As such, hepatic lymph flow is regulated predominantly by the transcapillary hydrostatic pressure gradient (THPG). Thus, hepaticlymph formation is extremely sensitive to increases in postsinusoidal venous pressure. Increased postsinusoidal pressure induced by hepatic cirrhosis leads to an increased THPG in the liver. Concurrently, increased portal venous pressure results in increased THPG in the splanchnic capillary bed. Therefore, increased hepatic and gastrointestinal lymphatic flow is an expected consequence of postsinusoidal portal hypertension and has been documented in animal models of portal hypertension associated with hepatic cirrhosis.
Increased hepatic and gastrointestinal lymph production results in increased lymphatic flow through the thoracic duct. This relationship has been confirmed in studies involving cannulation of the thoracic duct in patients with cirrhosis.
The normal flow through the thoracic duct is approximately 1 L/day but may be as high as 20 L/day in patients with cirrhosis. In such patients, thoracic duct flow rates seem to correlate with portal pressure measurements.
In our patient, traumatic injury to the thoracic duct probably occurred during placement of the central venous catheter. Because of the increased thoracic duct flow resulting from portal hypertension, the injury produced a high-output chylous fistula. Loss of protein-rich lymph resulted in depletion of fluids and coagulation factors, leading to volume contraction and coagulopathy. Repletion of central volume led to an increase in postsinusoidal pressure and a subsequent increase in output of the fistula.
A TIPS is used in the management of complications of portal hypertension.
Successful TIPS placement decreases the portal venous pressure and the portosystemic pressure gradient. A TIPS is effective in the treatment of gastrointestinal bleeding due to portal hypertension and other complications of portal hypertension including refractory ascites and cirrhotic pleural effusions.
By decreasing postsinusoidal and portal pressure, a TIPS should decrease hepatic and gastrointestinal lymph flow. Because hepatic and gastrointestinallymph flow represents most of the thoracic duct flow in patients with cirrhosis, a TIPS should also decrease thoracic duct lymph flow. Studies have confirmed that surgical portosystemic shunts decrease thoracic duct lymph flow in both animals and humans.
but this phenomenon has not been studied in patients undergoing a TIPS procedure. Theoretically, placement of a TIPS would be expected to have a similar effect on thoracic duct flow but without the associated morbidity and mortality of operative portosystemic shunts.
The current case provided an unusual clinical opportunity for observing the effects of cirrhosis of the liver and portal hypertension on thoracic duct lymph flow. With thoracic duct injury and formation of a chylous fistula, loss of the protein-rich lymph fluid led to decreased intravascular volume, depletion of proteins, and subsequent coagulopathy. The development of gastrointestinal bleeding from portal hypertensive gastropathy in the setting of a corrected coagulopathy necessitated intervention to correct the portal hypertension. In this setting, we chose to use a TIPS, hypothesizing that this would not only control the gastrointestinal bleeding but also decrease lymphatic output of the fistula. The theoretic benefit of a TIPS to control drainage from the thoracic duct fistula was realized. In fact, the fistula output started to decrease immediately after TIPS placement and discontinued completely within 72 hours. The patient's coagulopathy resolved, as did her requirements for large volumes of albumin to maintain hemodynamic stability. She was able to avoid open surgical repair or ligation of the thoracic duct with the associated long-term consequences.
This case indirectly confirms that thoracic duct flow is substantially increased in patients with cirrhosis of the liver and portal hypertension. TIPS placement decreases portal pressure and is associated with decreased thoracic duct lymphatic flow. A TIPS should be considered in the treatment of portal hypertension-related increases of hepatic lymph flow manifesting as chylous ascites, chylous fistulas, or chylous pleural effusions in patients with cirrhosis of the liver.
Ascites, renal failure, and electrolyte disorders in cirrhosis: pathogenesis, diagnosis, and treatment.
in: McIntyre N Benhamou J-P Bircher J Rizzetto M Rodes J Oxford Textbook of Clinical Hepatology. I. Oxford University Press,
Oxford (England)1991: 429-470