Chapter 9: Surgical Management of Portal Hypertension

In the 15 years since publication of the second edition of this book, the once preeminent role of the general surgeon in the management of portal hypertension (PHT) has virtually vanished. The ever-wider application of percutaneous catheter–based portal decompressive methodologies, more aggressive pharmacologic and endoscopic approaches, and the maturation of liver transplantation into a technically feasible and predictably successful treatment for end-stage liver disease have supplanted the role of the general surgeon in the contemporary management of patients with bleeding varices. Randomized trial evidence that percutaneous transjugular intrahepatic portacaval shunt (TIPS) construction results in survival and variceal rebleeding outcomes equivalent to those after surgical shunt underscored this progression.1 Portosystemic shunt and esophagogastric devascularization procedures, for a half century the mainstay of the management of patients with cirrhosis and PHT, are now so rarely performed that few currently active general surgeons have carried any of them out in the past quarter century. Such operations are no longer even part of the training curriculum for general surgery residents.

This chapter is, therefore, to a substantial degree a reflection on general surgeons’ historical involvement with the management of PHT and a testimonial to the inexorable advance of medical technologies. Others have offered similar valedictories.² Herein are archived general principles about the various types of portal decompressive shunts and devascularization procedures, perhaps of occasional contemporary use when the interventional radiology team is out of town and the airport is closed by bad weather or perhaps in the developing world where well-stocked angiography suites are rare but technically skilled surgeons are not.

General Considerations

The major complications associated with hepatic cirrhosis and PHT include gastrointestinal (GI) tract hemorrhage, ascites, hypersplenism, hepatic encephalopathy, and liver failure. The last two complications are commonly progressive and irreversible and, as markers for advanced liver disease, are indications for consideration of transplantation—a topic discussed elsewhere in this text.

Studies of the natural history of cirrhosis and PHT suggest that esophagogastric varices will develop in approximately 30% of patients with compensated cirrhosis and in 60% of patients with decompensated disease.³ The risk of bleeding from large varices is 20% to 30% per year. After an initial bleed, the risk of rebleeding is 75% to 80%, with the highest risk being within the first 6 months to 1 year. Acute variceal bleeding carries a mortality rate of approximately 30% (range, 15%–50%), with most deaths occurring in poor-risk Child class C patients as the consequence of progressive liver failure.

Pathophysiologic studies of portal venous pressure have yielded several important observations: (1) normal portal venous pressure is 5 to 7 mm Hg; (2) variceal bleeding does not occur until the portal venous pressure exceeds 12 mm Hg; and (3) reduction of the portosystemic pressure gradient by 50%, or to an absolute pressure of ≤12 mm Hg, generally prevents variceal bleeding. PHT is defined as a portal pressure of greater than 10 mm Hg.

A majority of patients with PHT present with upper (or lower) GI tract hemorrhage from esophageal, gastric, hemorrhoidal, or stomal varices or from portal hypertensive gastropathy. In Western countries, the leading cause of portal hypertensive bleeding remains alcoholic liver disease; in developing countries, such hemorrhage arises most commonly as a consequence of various infectious disorders (viral hepatitis, schistosomiasis).

After the source of GI tract hemorrhage has been identified as portal hypertensive in origin, proper initial management includes blood and volume restitution, pharmacologic modification of portal venous pressure, and (for upper GI bleeding) endoscopic therapy by variceal sclerotherapy or banding. Because variceal rebleeding in this setting is commonplace, one or another form of percutaneous transjugular or direct TIPS construction is warranted. Transfer to a center where such procedures are frequently performed may be necessary.

Because anesthesia, major surgery, blood loss, and diversion of hepatic portal blood flow are tolerated poorly by decompensated Child class C cirrhotics, whenever possible, patients with persistent jaundice, intractable ascites, spontaneous encephalopathy, and advanced muscle wasting are better managed by endoscopic variceal therapy or banding or TIPS (or, when appropriate, orthotopic liver transplantation). Not all such patients are irreversibly ill; temporary control of bleeding followed by vigorous nutritional and metabolic resuscitation frequently can improve Child class and prognosis in patients with cirrhosis and therefore improve their likelihood of long-term survival. Similarly, documented abstinence from alcohol for 6 to 12 months in a previously recalcitrant alcoholic with cirrhosis makes consideration of liver transplantation ethically reasonable.⁴

Another important prognostic feature is the presence of a significant coagulopathy, defined as an international normalized ratio greater than 1.8 despite correction with blood products. Such patients have an excessive risk of intra- and postoperative hemorrhage, and surgery should be withheld. Thrombocytopenia from hypersplenism is commonplace in patients with variceal bleeding, with platelet counts commonly around 60,000/mL. Successful portal decompression generally reverses hypersplenism and restores platelet counts to normal levels.⁵

Special concerns accompany the anesthetic management of the patient with cirrhosis, especially those undergoing shunts or liver transplantation.^6,7 Generalized arteriovenous shunting in these patients may involve the pulmonary circulation, resulting in significant hypoxemia. Hepatocellular dysfunction may result in slowed clearance of anesthetic agents and sedatives that are normally metabolized by the liver. Cirrhosis is associated with an upregulation of benzodiazepine receptors in the brain, thus making the use of such sedative agents relatively contraindicated. Fluid and electrolyte disorders, such as extracellular volume excess, respiratory alkalosis, and total body potassium deficiency, can be anticipated. Repletion of blood volume in an actively bleeding patient may result in numerous problems with volume shifts and dilutional coagulopathy. Endotracheal intubation of a patient who has recently bled is fraught with the risk of aspiration of gastric contents—a problem that may be exacerbated when concurrent ascites results in significantly raised intraabdominal pressure.

Although general anesthesia and major surgery in a hypovolemic, malnourished, coagulopathic, chronically ill patient with cirrhosis might seem to result in an overwhelmingly elevated morbidity and mortality, Orloff and Bell successfully performed emergency side-to-side portacaval shunts (PCSs) in more than 450 patients.8 Although the early post-shunt mortality rate was 16%, further variceal bleeding was eliminated, and the 5-year survival rate exceeded 70%. However, others have reported mortality rates of 30% to 50% after emergency PCS surgery.

Today almost all such patients would optimally be treated initially with endoscopic therapy and then by TIPS, with immediate control of bleeding in virtually all such patients. However, when TIPS is performed on an emergency basis in actively bleeding patients who are inadequately resuscitated, the mortality rate remains excessive.

Fig. 9.1 Esophageal transection using the end-to-end anastomosis stapling device. The stapler is introduced into the distal esophagus via a high gastrostomy. The distal esophagus has been carefully dissected, with careful attention to the identification and preservation of the vagus nerves. A heavy monofilament suture is then tied around the “open” central rod of the stapler, after which the stapler is fired and removed. The tissue “donut” from the stapler, which is the area of the esophageal transection and reanastomosis, is checked for completeness, and the gastrostomy is closed.

Laparotomy and transgastric staple transection of the esophagus may be the optimal approach for emergency control of esophageal variceal bleeding in which endoscopic treatment has failed and for some reason TIPS is not feasible. Such a procedure is rapid, relatively straightforward, and effective (at least in the short term), and it does not significantly interfere with consideration of a shunt or liver transplantation if the patient survives to be a candidate for either ( Fig. 9.1).

Direct operative attack on bleeding esophageal and gastric varices has been a well-established therapeutic concept for more than a century. The aim is to reduce inflow to the bleeding gastroesophageal varices and is dependent on the extent of devascularization, with better bleeding control being achieved with more extensive procedures.

Certain approaches, such as direct oversewing of variceal columns after thoracotomy and longitudinal esophagotomy, are obsolete. Other variations on the devascularization theme combine splenectomy; gastroesophageal variceal plexus ligation; and occlusion of the coronary, gastroepiploic, and short gastric veins.⁹ It is important that a sufficient length of esophagus (at least 7 cm) is mobilized to be devascularized from below the diaphragm. Moreover, gastric devascularization should totally devascularize the entire greater curvature and lesser curvature of the stomach in a manner similar in extent to a highly selective vagotomy. An extremely aggressive devascularization procedure, originally promoted by Sugiura and Futagawa, entails a staged thoracotomy and laparotomy to perform splenectomy, esophageal mucosal transection, and a painstaking ligation of esophagogastric venous collaterals.10 The incidence of encephalopathy is 10% or less after devascularization but depends on the underlying liver disease and its severity.

Staple transection of the esophagus obviously does not treat gastric varices or portal hypertensive gastropathy. It may be rendered more difficult or impossible to perform if the esophagus is edematous, inflamed, or scarred because previous endoscopic variceal therapy. Because the vagus nerves may be divided during esophageal transection, pyloroplasty may be required to prevent postoperative gastric outlet obstruction.

Devascularization procedures can be used for patients who have extensive portal venous thrombosis and who continue to have significant gastric or esophageal variceal bleeding despite pharmacologic and endoscopic therapy. Extensive devascularization for these patients can reduce the risk of major bleeding for several years and may be the only reasonable therapeutic solution in bleeding patients who have no shuntable splanchnic vessels.

Portal Decompressive Shunt Surgery

Because variceal hemorrhage results from underlying PHT, reduction of portal pressure to normal physiologic levels invariably halts such bleeding. Bypassing splanchnic venous outflow obstruction by connecting the hypertensive portal system to the systemic venous circulation, either directly or by means of various autogenous or synthetic conduits, once proved to be a highly effective means of halting variceal hemorrhage.

In 1877, Eck performed the first PCS in dogs. Pavlov et al performed the first substantive investigations of the metabolic effects of portosystemic shunts in 1893. Vidal, in 1903, performed the first PCS in a patient with ascites. The patient died 4 months later, probably from sepsis, liver failure, and encephalopathy. A resurgence of interest in total portosystemic shunts paralleled the beginning of arterial reconstructive surgery in the 1940s and 1950s, and such procedures, usually end-to-end or side-to-side PCS, became commonplace therapy for patients with variceal hemorrhage in the two decades after the Second World War.¹¹

Several complicating factors became evident. A characteristic neuropsychiatric disorder of memory loss; altered levels of consciousness; behavioral changes; and (in its advanced stages) stupor, coma, and death was noted to be a frequent and unpredictable result in patients who had undergone portosystemic shunt.¹² This syndrome, hepatic or portosystemic encephalopathy (“hepatic coma”), remains incompletely understood and, in varying degrees, has continued to plague all forms of portal decompression. Pavlov had shown that dogs undergoing PCS become listless and anorectic, suffer premature death, and have hepatic atrophy at autopsy. Extensive investigations in the 1950s and 1960s, summarized most elegantly by Starzl et al, suggested that an equivalent phenomenon in humans undergoing PCS, accelerated hepatic atrophy and progressive loss of hepatocellular function occur because diversion of portal flow deprives the liver of a splanchnic venous trophic factor (insulin seemed a likely candidate) necessary for normal hepatocellular function and regeneration.¹³

General acceptance of the “hepatotrophic theory,” combined with the demonstration that PCS improved survival only minimally in prospective randomized trials, led to a significant decline in the performance of PCS after the 1980s. Recognizing, on the one hand, the ability of PCS to control bleeding but, on the other hand, the evident advantage of maintaining portal perfusion following devascularization procedures, Warren et al introduced the concept of selective shunts for variceal decompression in the 1960s.¹⁴ Even later, in the 1980s, the concept of partial PCSs, designed to lower portal pressures to nonbleeding but not “subnormal” levels, was advanced.

Three types of surgical decompression have been used to treat gastroesophageal varices. All aim to reduce variceal pressure and provide durably effective control of variceal bleeding. Total shunts divert all portal venous blood flow from the liver, selective shunts divert only the gastric and splenic components of portal flow, and partial shunts reduce the portal pressure but may maintain some portal flow. The more central the shunt, the greater the continuing patency rate—because of the high flow—but the more certain is loss of first-pass portal perfusion of the liver. Use of prosthetic materials increases the risk of thrombosis. Factors that determine the choice of shunt are technical feasibility based on preoperative evaluation of the splanchnic vessels, the surgeon’s preference and experience, and the subsequent possibility of liver transplantation.

Operative approaches to each of these procedures will not be described in great detail but will be diagrammed; specifics can be found in reference texts. The key steps are adequate operative exposure of the vessels to be connected and careful operative technique in fashioning the veno-venous anastomoses. The overall perioperative management of the condition of patients with underlying cirrhosis and impaired liver function is the other key factor in a successful outcome.

Total Portosystemic Shunts

The goal of a total portosystemic shunt is complete portal decompression. Theoretically, such shunts have the highest likelihood of protecting against further variceal rebleeding. Total shunts include PCSs, mesocaval shunts, and proximal or central splenorenal shunts. The shunts are usually 15 to 25 mm in diameter ( Fig. 9.1; Fig. 9.2; Fig. 9.3; Fig. 9.4; Fig. 9.5).

In experienced hands, the patency rate of PCSs is more than 90%, with excellent control of variceal bleeding. In some 1700 PCSs performed by Orloff, fewer than 10 (<0.1%) have been associated with proven variceal rebleeding (Orloff MJ, personal communication). However, others have reported higher rebleeding rates, ranging from 3% to 17%. Shunt failure, when it occurs, almost always results from an attempt to anastomose a partially or completely thrombosed portal vein to the inferior vena cava (IVC) rather than performing an alternative portal decompressive procedure.

Unfortunately, total PCS is associated with a substantial risk of encephalopathy (30%–50%) and acceleration of liver failure. Notwithstanding the fact that a number of such patients with postoperative neuropsychological deterioration can be treated medically (protein-restricted diet, oral antibiotics, and/or lactulose) with mitigation of symptoms, most authorities discouraged and discarded this particular shunt construction.

Drapanas et al theorized that a conduit from a tributary of the portal vein (such as the superior mesenteric vein [SMV]) to the systemic venous circulation might preserve prograde portal flow (and continued hepatic portal perfusion) and still afford adequate portal decompression and protect against further variceal hemorrhage.¹⁵ Development of the prosthetic interposition mesocaval shunt quickly followed. Although mesocaval shunts continue to be performed, their long-term efficacy is questionable. Initially constructed using large-caliber Dacron grafts, mesocaval shunts resulted in slow flow through a thrombogenic synthetic conduit, with predictable shunt thrombosis and variceal rebleeding. Interposition of autologous tissue (most commonly internal jugular vein) as mesocaval graft was shown to produce the highest likelihood of long-term patency.¹⁶ Importantly, the underlying premise that hepatic portal venous perfusion can be preserved by a mesocaval shunt has not been borne out; such patients have a likelihood of postoperative complications equivalent to that of patients undergoing standard PCS.

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Clinical Outcomes of Balloon-Occluded Retrograde Transvenous Obliteration and Balloon-Occluded Antegrade Transvenous Obliteration
2. Paracentesis and the LeVeen and Denver Shunts
3. Management of Duodenal Varices
4. Endoscopic Classification and Management of Varices

Stay updated, free articles. Join our Telegram channel

Join