Search

Portal hypertension can be classified into two main groups, presinusoidal (intrahepatic and extrahepatic) and intrahepatic. Presinusoidal extrahepatic portal hypertension is unlikely to be associated with hepatocellular failure. Increased splenic flow plays a minor role in its cause. Renal cortical perfusion may be reduced. Causes of presinusoidal intrahepatic portal hypertension include schistosomiasis, myeloproliferative diseases, sarcoidosis, congenital hepatic fibrosis, chronic arsenic poisoning, and primary biliary cirrhosis. Portosystemic shunt operations are well tolerated unless there is associated hepatocellular damage. In cirrhosis, the portal hypertension is presinusoidal, sinusoidal, and postsinusoidal. The problem of bleeding esophageal varices is largely that of hepatocellular failure. The Budd-Chiari (hepatovenous obstruction) syndrome has many causes, including “congenital” webs and trauma. These may be susceptible to a surgical approach.

A case of Wilson’s disease, in which extrapyramidal motor singns manifested only three months and Kayser-Fleischer ring eight months after splenorenal anastomosis, is reported.

The hemodynamic alterations occurring in portal hypertension are com­ plex and difficult to evaluate. Investigation and understanding of the vas­ cular pathophysiology assume clinical importance because operative treat­ ment very often alters splanchnic blood flow suddenly and dramatically. I t is generally conceded that hepatic function is a major determinant in the degree of well being of patients with portal hypertension, especially after complications of portal hypertension are treated by portosystemic shunt. The concept that hepatic function is not the sole determining factor is em­ phasized by a number of clinical studies. Callow et al. (I), Conn & Linden­ muth (2) and Jackson et al. (3) have analyzed the results of shunt therapy in controlled, randomized studies. Survival rates in the surgical groups are not superior to those in the groups treated medically. In a review of the subject, Grace et al. (4) came to the same conclusion. Although survival rates were similar, modes of death differed. In the medically treated control group, hemorrhage from esophageal varices was the major cause while in patients undergoing portacaval shunt, liver failure was the most common cause of death. Johnson et al. (S) reported the results of nonshunting procedures in the control of esophageal varices. None of the long-term survivors were troubled with hepatic encephalopathy or liver failure. In bilharzial cirrhosis hepato­ cellular function usually remains good but Hassab (6) found that 20.6 per cent of such patients undergoing portacaval shunt developed encephalopathy. After nonshunting operations on 3SS patients with the same disease only one patient developed encephalopathy in the absence of bleeding. Mikkelsen et al. (7) reported on 36 patients, in 17 of whom the portal vein was patent, liver function was excellent and there was no evidence of liver disease on gross examination. Following portacaval shunts, 10 developed signs of encephalopathy and five died of liver failure. In those patients with spon­ taneously occluded portal veins these complications were infrequent. The foregoing clinical observations suggest that the sudden complete diversion of portal blood flow, as occurs following an end-to-side or side-to­ side (8, 9) portacaval shunt, may be detrimental to the liver under certain circumstances. This detrimental effect has been observed in dogs (10), spider monkeys (11), and man (12). Warren et al. (13, 14) studied hemo

Currently, preoperative hemodynamic measurements are of limited value in the selection of patients for portosystemic shunt surgery. Wedged hepatic vein pressure measurement accurately assesses portal pressure in alcoholic liver disease and can be helpful in the diagnosis and follow-up of portal hypertension but not in the selection of patients for surgery. Though theoretically useful, hepatic blood flow measurement has not been shown to correlate with results of surgery. New methods have been developed for measurement of portal blood flow preoperatively, which use the umbilical vein approach. Assessment of hepatic arterial response to portal flow diversion is also feasible with the patient under local anesthesia. Hopefully, one of these measurements will prove useful in selecting those patients who can undergo portosystemic shunt without a high risk of postoperative encephalopathy or hepatic failure.

Of 48 patients given selective mesenteric arterial infusion of vasopressin for the control of gastrointestinal hemorrhage, 28 were actively bleeding from varices secondary to portal hypertension, 14 were infused electively at the time of portosystemic shunt surgery, and 6 were infused because of hemorrhage from an arterial or capillary source demonstrated on prior selective arteriography. The indications, technique, and results of the vasopressin infusions are described.

The surgical contributions to the therapy of lifethreatening hemorrhage from esophageal varices due to portal hypertension during the past twentyfive years clearly support a few virtually incontestable conclusions. Analysis of the data derived from both controlled and uncontrolled studies has identified that (1) successful portosystemic shunt of appropriate size will eliminate the threat of variceal hemorrhage ; (2) the risk of morbidity and mortality due to encephalopathy and liver failure is increased under these conditions [l-4]. This risk is of sufficient magnitude such that life expectancy for patients undergoing prophylactic portacaval shunt is not improved and may actually be reduced when compared to that of a comparable group of patients treated by medical means alone, even though the latter group is subject to the continuing risk of exsanguinating hemorrhage. In the light of this evidence, it is difficult to escape the conclusion that acute diversion of portal venous blood from the liver is fraught with hazard. This conclusion has been supported by extensive laboratory investigations [S] primarily in the dog, which have demonstrated biochemical and morphologic changes consistent with liver injury which frequently leads to the untimely death of the animal [6]. The inability to predict the magnitude of this risk for any individual patient within a cohort of cirrhotic patients requiring portacaval shunt therapy has suggested that individual differences in preoperative hemodynamic status may be a factor. C’ertainly it has been common clinical knowledge that s’everely restricted portal venous flow to the liver M manifested by patients with cavernous transformation of the portal vein or portal venous thrombosi,s is associated with a distinct survival rate whereas patients who suffer acute ligation of the portal vein rarely survive. Furthermore, in 1963, Warren and his colFrom the Department of Surgery, University of Miami and Affiliated Hospitals, Miami, Florida. This work was partially supported by USPHS Grant He 08411-05 and RR 261, University of Miami Clinical

Adult cadaveric dissections and simulated surgical approaches were performed to assess surgical feasibility of re-establishment of this natural portosystemic shunt. Existing anatomic relations and absence of endothelial lining prevent the use of the ductus in surgical treatment of portal hypertension although dilatation is possible.

Ninety-two splanchnic arteriograms performed to demonstrate the portal vein were analysed. Patients who had portal hypertension were grouped together and their hepatic and gastroduodenal arterial patterns were graded from 4+ to 1+ according to their filling. The arterial pattern in the patients without portal hypertension were graded similarly. Total splenic blood flow (liters per minute) was measured in all patients using intraarterial 133 Xenon. Arterial pressure tracings were obtained during catheterization. Of fifty-five patients with portal hypertension twenty-two had poor arterial filling of the hepatic and gastroduodenal arteries. In the remaining thirty-three patients with good arterial filling, twenty-four had either portal vein occlusion, a congenital abnormality of the hepatic artery, a spontaneous internal portosystemic shunt or hepatocellular disease (with jaundice) which was thought to contribute to the good arterial pattern. Nine patients had good filling for no apparent reason. The poor contrast filling was reversed following a successful portacaval shunt. All thirty patients without portal hypertension had good filling of their arterial system. A possible portal venous-hepatic arterial nervous reflex is discussed to account for these findings. A siphon of blood from the hepatic artery into the splenic artery is also considered.

Transthoracoesophageal ligation of bleeding esophageal varices was carried out in 55 patients with a survival of 71%. All good-risk patients (Child criteria) survived; 75% of moderate-risk patients and 48% of poor-risk patients survived. Nine patients underwent exploratory surgery for recurrent variceal bleeding following a portosystemic shunt, and 46 patients underwent variceal ligation as the first step of a two-stage procedure for the control of bleeding esophageal varices. Sixty-four percent of the latter group had a subsequent portosystemic shunt, usually of the splenorenal type, to avoid encephalopathy and liver failure associated with portacaval shunt. Group A and B patients had 52% cumulative survival at five years and 24% cumulative survival at ten years. Transthoracoesophageal ligation of varices, followed preferably by an end-to-side splenorenal shunt and splenectomy, remains the most satisfactory method for the control of esophageal bleeding secondary to Laennec cirrhosis.

In spite of ideal present day management, 25% of patients who bleed from esophageal varices continue to succumb to hemorrhage. Those who fail to respond to conservative measures require surgical intervention. Major surgery on a patient already taxed by exsanguinating hemorrhage and hepatic decompensation is accompanied by a high mortality. White et al,1Piccone and LeVeen,2and Christopherson et al3have recently described the use of an umbilical systemic vein shunt with low operative mortality to control bleeding from esophageal varices. Their reports indicate that flows obtained under existing portosystemic pressure differentials were sufficient to abate exsanguination. Gott et al4have described a non thrombogenic coating technique for surfaces using graphite, heparin sodium, and benzalkonium chloride. Zarins et al5have recently reported their results with splenofemoral or external jugular shunts to relieve portal hypertension in dogs utilizing the Gott graphite-benzalkonium-heparin coating. We6have developed