SINCE LYMPH normally contains practically no red cells, the finding of gross hematochylia in patients with cirrhosis is of great interest.1 The obvious explanation for such contamination of the thoracic duct lymph is leakage of blood from distended liver sinusoids into the spaces of Disse. However, certain of our observations in both man2 and animals3 with portal hypertension are inconsistent with this hypothesis. Our previous experience4,5 with the experimental development of hematochylia in animals with aortocaval fistulae had prompted us to study this phenomenon in patients with cirrhosis as originally reported by Dumont and Mulholland.1 Additionally, we were interested in their suggestion regarding the therapeutic value of thoracic duct drainage in the management of bleeding varices and refractory ascites.
The thoracic duct was cannulated in 19 patients with a diagnosis of cirrhosis of the liver. Of these 13 had been admitted with upper