Portal hypertension may be defined as a persistent increase in the portal vein pressure to above normal. I*3 Because the portal system is relatively inaccessible, pressures are not often measured in children, and an operational definition is important. The diagnosis should be considered if the signs of portal hypertension are observed. Some of the signs, such as esophageal varices, are essentially pathognomonic of portal hypertension, and when they are present the diagnosis may be established with confidence.
Several different ways have been used to measure the portal pressure resulting in some confusion about the values necessary to diagnose it. A pressure of above U mmHg by direct measurement at the time of surgery is considered abnormal, as is a splenic pulp pressure of more than 16 mmHg. J Catheterization of the hepatic vein allows for the measurement of hepatic wedge pressure, which reflects the portal venous pressure unless there is a substantial pre-sinusoidal resistance and pressure grathent. Measured by this technique, the portal pressure is normally about 4 mmHg higher than that of the inferior vena cava to overcome the resistance of the hepatic sinusoidal vascular bed.1 Pressure grathents of greater than 4 mmHg are abnormal. In studies of children, the most often reported values are those obtained by direct manometry at the time of surgery. Because of the influence of general anesthesia, these values may be somewhat lower than those obtained under other circumstances, but they are similar to pressures recorded in adults with portal hypertension.2,3
There are two basic causes for increased portal venous pressure: an increase in the resistance of the portal system to blood flow or an increase in the portal blood flow. l The latter is quite rare and will be discussed no further than to say that occurrences of arterio-portal venous shunts have been reported to increase portal blood flow and pressure. Resistance to flow can occur anywhere in the portal circulatory system. Although there continues to be some controversy regarding some disorders, a functional classification categorizing disorders by the site of increased resistance is useful (Table). Anatomic disturbance of the portal vein or its major branches may result in resistance to flow and what is termed prehepatic or extrahepatic portal hypertension. Parenchymal liver disease may result in portal hypertension. Here the categorization of some diseases is difficult and controversial. If the obstruction to flow is in the small portal tributaries before the sinusoidal bed, the result is presinusoidal portal hypertension. The hepatic wedge pressure will be normal in prehepatic and presinusoidai hypertension which serves to functionally divide diseases causing portal hypertension into two groups which differ clinically in some important ways. Parasinusoidal or sinusoidal porta) hypertension results from diseases increasing the resistance to flow through the sinusoidal bed. Central venous disease causes postsinusoidal hypertension. Finally, obstruction of the veins draining the liver or increased central venous pressure will result in hepatic congestion and suprahepatic portal hypertension.
DIAGNOSING PORTAL HYPERTENSION
Diagnosing the cause of portal hypertension involves the usual modalities for diagnosing parenchymal liver disease, especially liver biopsy, and techniques of imaging the portal venous system. A large or firm liver, jaundice, ascites, and somatic wasting suggest parenchymal liver disease and should lead to biochemical and histological evaluation. The absence of evidence of liver disease should lead to imaging of the portal vein as a first step. Traditionally, imaging has been performed by injection of contrast material while obtaining serial x-rays. The injection may be into the splenic pulp, and as the contrast material flows from the spleen into the splenic and portal veins, the anatomy of the system is defined. The risks of splenoportography include bleeding into the peritoneal cavity, and the procedure should not be performed if coagulopathy or thrombocytopenia are present. Failure to visualize the portal vein after splenic injection suggests splenic vein thrombosis. Another approach for visualizing the portal vein is umbilical vein catheterization, which is accomplished by dissecting the preperitoneal fat above the umbilicus. Current techniques of angiography allow for excellent definition of the portal system in the venous phase of a superior mesenteric arteriogram. This is the preferred technique at present. The emergence of high resolution ultrasound threatens to preclude the need for injection angiography. Ultrasound can define the anatomy of the portal system in most cases, and newer Doppler technology may allow for the measurement of portal blood flow. It is the /rot imaging technique we have used in cases of suspected portal obstruction.
PREHEPATIC OR EXTRAHEPATIC PORTAL HYPERTENSION
Most large series of portal hypertension in childhood report a predominance of extrahepatic portal hypertension.4 However, the relative frequency of the causes of portal hypertension in children is actually unknown because no objective analysis of the frequency and magnitude of portal hypertension in the common parenchymal liver diseases has been undertaken. In extrahepatic disease, the signs and symptoms of portal hypertension usually dominate the clinical picture, whereas in parenchymal liver disease, they may be trivial compared to the other signs of liver disease. This tends to accentuate the reported incidence of extrahepatic portal hypertension. Put another way, the patient who presents with the signs of portal hypertension will probably be found to have extrahepatic disease, but the patient with cirrhosis can almost always be found to have portal hypertension.
The obstruction is almost always in the portal vein. Obstruction of the splenic vein occurs in 15% to 20% of cases and the superior and/or inferior mesenteric veins are involved in a few. Frequently, cavernous transformation of the portal vein, a probable attempt at rechanneling the obstructed segment, accompanies the obstruction. Rarely, portal vein obstruction occurs with congenital anomalies, but the most common cause is thought to result from thrombosis. Stime authors report a significant correlation between the occurrence of neonatal omphalitis and umbilical catheterization and the development of extrahepatic portal vein thrombosis, but others feel the evidence for such an association is weak. 2,4 No prospective study of the risk of umbilical catheterization for developing portal vein thrombosis has been published. Certainly, portal vein thrombosis can occur as a complication of pancreatitis, abdominal infections, and some systemic disorders, but these are rare. Also, portal vein obstruction may complicate the course of parenchymal liver disease with portal hypertension and reduced portal blood flow. The majority of patients have no antecedent history to suggest an etiology.
INTRAHEPATIC PORTAL HYPERTENSION
The common theme among the various intrahepatic diseases that cause portal hypertension is increased resistance to blood flow. ' Although portal blood flow through the liver is at a standstill in many cases, this is a function of the relative resistance of the liver and the collateral circulation. In some cases of severe cirrhosis, the portal flow is thought to be reversed, but whether this ever actually happens is controversial. The most common pathologic finding to account for increased resistance is fibrosis, which causes reduced compliance and constriction of the vascular bed. Fibrosis may affect the presinusoidal portal radicles (presinusoidal intrahepatic portal hypertension), the sinusoidal bed (parasinusoidal hypertension), or the central veins (postsinusoidal). Rarely, liver diseases without fibrosis produce portal hypertension, and the mechanism is usually unclear.
CLASSIFICATION OF PORTAL HYPERTENSION
Worldwide, one of the most common causes of portal hypertension and the most common cause of intrahepatic presinusoidal hypertension in children is schistosomiasis. ' Ova of the parasites laid in the submucosa of the colon embolize into the portal vein radicles where they produce a foreign body reaction and scarring. Other diseases which produce presinusoidal hypertension include congenital hepatic fibrosis and idiopathic portal fibrosis, an unusual disease in which there is moderate fibrosis of the portal areas with sclerosis of the portal vein radicles.
Portal hypertension generally accompanies cirrhosis.1,6 The site of increased resistance in cirrhosis is variable, and no pattern of cirrhosis holds true in this regard. This variability may contribute to the different patterns of symptoms seen in patients with any type of cirrhosis. Biliary cirrhosis usually produces presinusoidal hypertension, whereas postnecrotic cirrhosis most often causes parasinusoidal hypertension. Diseases with more lobular involvement, such as chronic active liver disease with cirrhosis, produce parasinusoidal hypertension.
Although portal hypertension is not often an important clinical problem, parasinusoidal hypertension results from all active lobular diseases such as hepatitis. l The magnitude of the hypertension in hepatitis in children is not known, but it probably contributes significantly to the splenomegaly observed in children with diseases such as giant cell hepatitis.
Figure 1. This operative portogram in an infant with congenital hepatic fibrosis demonstrates the major collaterals which develop with portal hypertension. The portal vein (P) and its major tributaries, the superior mesenteric vein (SM) and splenic vein |S) ate seen. In addition, tributaries with connections to the systemic venous circulation are also seen. The coronary (left gastric) vein (C) communicates through the gastroesophageal (variceal) plexus (V) with the azygous vein (A), a short gastric vein (G) also enters the variceal system, and the inferior mesenteric vein |IM) courses toward the hemorrhoidal veins.
Finally, postsinusoidal intrahepatic portal hypertension is produced by a few diseases which cause sclerosis of the central veins. Veno-occlusive disease is encountered most frequently in developing nations and is often the result of ingesting seeds of plants of the genera Senecio, Crotalaria, and Heliotropium in folk teas and medicines.3
SUPRAHEPATIC PORTAL HYPERTENSION
Numerically the most common cause of suprahepatic portal hypertension is heart failure, but this complication is not often important in children. Budd-Chiari syndrome produces impressive symptoms and signs, but is rare in childhood. In contrast to adults, it is most often idiopathic in children.3
CONSEQUENCES OF PORTAL HYPERTENSION IN CHILDREN
The important consequences of portal hypertension are the development of collateral circulation, hypersplenism, ascites, and portal-systemic shunting.1
Collaterals form wherever the portal and systemic venous circulations are in close approximation (Figure 1). The most common sites are in the submucosa of the esophagus and rectum, the anterior abdominal wall, and the parietal peritoneum. They result from the need to decompress the portal circuit, and represent paths of least resistance. They do not effectively decompress, and no matter how longstanding or large the collaterals may be, portal hypertension always remains. The collateral channels are abnormal vessels with attenuated walls and sometimes achieve very impressive size (Figure 2). When they have little supporting tissue, such as in the wall of the esophagus, they may rupture and bleed. Bleeding esophageal varices are the complication of portal hypertension most life-threatening to the patient.
Splenomegaly is an almost constant sign of portal hypertension, which probably results purely from mechanical forces. Hematologic disease consistent with hypersplenism is seen in far fewer patients, and probably results from stasis of blood in the splenic pulp. Why it occurs in only a fraction of patients with large spleens is not known. Thrombocytopenia, anemia, and leukopenia, especially granulocytopenia, are seen, but rarely do they impair the patient significantly. The severity of gastrointestinal bleeding may be affected by thrombocytopenia.
The forces that produce ascites are incompletely defined, but hydrostatic forces in portal hypertension play an important role. Increased lymph formation by the liver is also important, which explains the differences in the rates of ascites formation in patients with presinusoidal vs. para- and postsinusoidal portal hypertension. Patients with extrahepatic portal hypertension or intrahepatic presinusoidal hypertension, such as congenital hepatic fibrosis, usually do not have chronic ascites.
The shunting of blood through collaterals and away from the liver may result in encephalopathy. Often this is minimal, especially in extrahepatic portal vein thrombosis, but may become acute and severe in the event of gastrointestinal hemorrhage with the resultant increased protein and ammonia load.
The work-up of portal hypertension in children must be thorough because knowing the cause dictates the therapy, in large measure. Importantly, when the evaluation is complete, the site of obstruction and the existence of parenchymal liver disease should be known. The measurement of the portal pressure is recommended, but it has little clinical importance except in planning for surgical shunting. Regular complete blood counts should be recorded in all patients.3,4
Esophageal varices and to a lesser extent varices of the gastric cardia are the most important collaterals because of their propensity for bleeding. The diagnosis of esophageal varices should be by endoscopy. ' Barium studies are fraught with difficulty. Non-variceal irregularities are frequently misinterpreted and non-distented varices are missed. With endoscopy, varices can be unequivocally identified and carefully quantitated.
An acute, severe episode of bleeding from esophageal varices is a life-threatening event. Management includes fluid and blood resuscitation, the prevention of encephalopathy, and an effort to arrest bleeding.7 Lavage with iced saline has the dual effect of removing blood and perhaps causing mucosal vasospasm. Vasopressin, infused at a rate of 0.2 to 0.4 units/ 1.73 mp 2/min intravenously, often dramatically reduces bleeding within 30 to 60 min.2, 3, 7 Intraluminal compression tubes, such as the SengstakenBlakemore tube, are sometimes effective, but are difficult to use and are dangerous.8 For the past several years, we have not used a compression tube, but rather have used emergency endoscopic sclerotherapy to stop bleeding with great success.9
Efforts to prevent bleeding are important. Aspirin containing drugs are forbidden. Serial endoscopic sclerotherapy, although time-consuming and costly, has very sharply reduced the incidence of major variceal bleeding in our patients. It is repeated until the varices are obliterated, and again if they recur. Alternatively, a surgical porto-systemic shunt can be created to decompress the system, if bleeding is severe and recurrent. Older patients and patients with extrahepatic disease generally experience better outcomes after shunting. ,0 The type of shunt employed depends on the anatomy of the portal system. The initial shunt should probably be spleno-renal in most cases. Our surgeon prefers the side-to-side variant. Later, central shunts may be performed if necessary.
The outcome of portal hypertension depends almost entirely on its cause. Patients with extrahepatic disease are expected to survive indefinitely with little morbidity. 5 Prevention and treatment of bleeding from esophageal varices are the major foci of treatment. The type of parenchymal liver disease causing portal hypertension has much to do with the prognosis. Stable post-necrotic cirrhosis complicated by variceal bleeding has a much better prognosis than does cirrhosis with chronic active liver disease and bleeding. Postsinusoidal hypertension generally results in extreme morbidity and mortality.
Figure 2. A. Splenoportogram in a teenage boy with extrahepatic portal vein thrombosis demonstrates huge variceal collaterals (V) after injection in the spleen (SPJ. but since most of the flow was through collaterals, the portal system was not adequately visualized. B. Ultrasound in the transverse plane demonstrates the immense varices (V). C. In the coronal plane, ultrasound delineates the portal system. The splenic vein (S) and part of the portal vein (P) are visualized. The portal vein is interrupted by a cavernoma (C). often seen in extrahepatic vein thrombosis. (LI - liver. SP- spleen).
1. Reynolds TB: fonal hypertension, in Schiff L. Schiff ER (eds): Diseases of the Liver, ed 5. Philadelphia. J. B. Lipplncott. 1982. pp 393-432.
2. AIaKiIIe D. Odievre M: Liner and Biliar. Tract Disease m Children. New York. Wiley-Flammarion, 1979. pp 262-295.
3. Walia BNS, Mitra SK, Chandra RK: Portal hypertension, in Chandra RK (ed): The Liurr and Biliary System m In/ants and Children. New York. Churchill Livingstone, 1979. pp 276-295.
4. Alvarez F. Bernard O, Brunelle F. et al: Portal obstruction in children. I. Clinical investigation and hemorrhage risk. J Pediatr 1983; 103:696-702.
5. Alvarez F, Bernard O. Brunelle F. et al: Portal obstruction in children. 11. Results of surgical portosystemic shunts. J Pediatr 1983; 103:703-707.
6. Kasai M, Okamoto A. Ohi R. et al: Changes of portal vein pressure and intrahepatic blood vessels after surgery for biliary atresia, J Pediatr Surg 1981: 16:152-159.
7. Funkalsrud E: Surgical management of portal hypertension in children. Long temi results. Arch Surg 1980: 115:1042-1045.
8. Hanna SS, Warren WD, Calambos JT. et al: Bleeding varices: I . Emergency management. Can Med Assoc) 1981; 124:29-47.
9. Ayres SJ, Goff JS, Warren GH: Endoscopic sclerotherapy for bleeding esophageal varices: Effects and complications. Ann Intern Med 1983; 98:900-903.
10. Chojkier M. Conn HO: Esophageal tamponade in the treatment of bleeding varices. A decadal progress report. Dig Du Sci 1980; 25:276-272.
CLASSIFICATION OF PORTAL HYPERTENSION