Complications of Cirrhosis: Clinical Perspectives and Implications CME Disclosures Helena Cortez-Pinto, MD, PhD Introduction The final stage of liver disease is punctuated by several major complications: ascites, hepatorenal syndrome (HRS), spontaneous bacterial peritonitis (SBP; a potentially lethal infection of the ascitic fluid that occurs in the absence of a local source of infection), variceal hemorrhage, and hepatic encephalopathy. Although liver transplantation is indicated whenever these complications are present, there are still many patients who have to be managed through them, either while waiting for a liver or because they have contraindications to liver transplantation. The appropriate management of these patients is crucial, not only because these complications can be life-threatening, but also because it is now clear that influencing the pretransplant status of patients results in better transplant outcomes. During this year's meeting of the European Association for the Study of the Liver (EASL), this broad topic was not primarily addressed in formal lectures, but rather most of the new data were presented in poster sessions. This clinical overview addresses some of the more important new data related to complications of cirrhosis -- ascites/ SBP, varices, and hepatic encephalopathy, as presented during these meeting proceedings. Ascites/SBP/HRS Background Ascites is one of the more frequent complications of liver cirrhosis and will develop in approximately 50% of patients with compensated cirrhosis who are followed for 10 years. Development of this complication is an ominous sign in that approximately 50% of patients with ascites die within 2 years if not transplanted. It is usually associated with a marked deterioration in nutritional status and in quality of life. The management of this clinical problem is challenging for the patient and physician alike because it requires very careful monitoring of the pharmacologic treatment, with special attention given to fluid and weight balance. Treatment of ascites has been based on a sodium-restriction diet and the progressive administration of diuretics, starting with spironolactone (100 mg/day) and furosemide (40 mg/day), and increasing progressively every 3-5 days, if weight loss is inadequate, until a maximum dosage of 400 mg/day of spironolactone and 160 mg/day of furosemide is reached. Fluid restriction is not necessary in treating most patients with cirrhosis and ascites. The chronic hyponatremia usually seen in patients with cirrhotic ascites is seldom morbid. In this setting, attempts to rapidly correct hyponatremia with hypertonic saline can lead to more complications than the hyponatremia itself. Preliminary data suggest that aquaretic drugs may have some promise in correcting hyponatremia, and may also be useful in treating HRS. Therapeutic paracentesis may be used in patients with refractory ascites; its use in patients with diuretic-sensitive ascites is more controversial. Administration of albumin at a dose of 5- 10 g/L of ascitic fluid removed has been strongly advised to prevent paracentesis-induced circulatory dysfunction in large-volume paracentesis (> 5 L). Treatment and Complications Ascites. Alessandria and colleagues, from San Giovanni Battista Hospital in Turin, Italy, compared low-dosage infusion of albumin (4g/L of ascitic fluid removed) with the standard dose of 8g/L of ascitic fluid removed in 25 patients undergoing large-volume paracentesis. In both groups, the amount of ascitic fluid removed was similar, as was plasma renin activity. The study authors found a similar incidence of paracentesis- induced circulatory dysfunction, suggesting that low doses of albumin are as effective as standard doses in the prevention of this complication -- a finding that may support a significant cost reduction in the management of patients with cirrhotic ascites. Renal failure. Another significant complication of ascites in patients with cirrhosis is the development of renal failure. Ballesté and colleagues, from Badalona and Barcelona, Spain, assessed the incidence and prognosis of different types of nonorganic renal failure (ie, functional, reversible renal failure) in 263 consecutive cirrhotic patients followed for 41 ± 3 months after their first ascites. They considered 3 types of such renal failure: prerenal failure when associated with a depletion of intravascular volume; renal failure associated with infection not leading to HRS; and HRS as defined by preestablished criteria. It is interesting to note that these investigators found that 49% of cirrhotic patients with ascites developed some type of nonorganic renal failure, prerenal failure being the most frequent (27.4% of all patients) during follow-up. This complication was associated with a significantly worse prognosis, suggesting that efforts to prevent prerenal failure should be implemented. HRS, a common and serious problem in patients with advanced liver disease, is a functional form of renal failure that can occur in end-stage liver disease. Its pathogenesis involves vasodilation of the splanchnic vessels. The vasoconstrictor terlipressin* may be useful in patients with HRS. Gluud and colleagues from the Cochrane Hepato-Biliary group in Copenhagen, Denmark, presented a systematic review of randomized controlled trials assessing the effect of terlipressin in the treatment of HRS. They found 3 published and 2 ongoing trials eligible for inclusion. All patients had cirrhosis and HRS type 1 or type 2. The most commonly evaluated dose of terlipressin was 1 mg given every 12 hours for 5-15 days. Survival rates were 62% for patients randomized to terlipressin and 20% for those in the control group, with terlipressin demonstrating a significant beneficial effect on creatinine clearance. The study authors concluded that terlipressin may be considered for patients with HRS in spite of its potential adverse effects, namely ischemia of myocardium, gut, or digits, although evidence is still needed for definitive treatment recommendations, especially regarding dose and duration of therapy. SBP. Nousbaum and colleagues, from France, presented data regarding the rapid diagnosis and prevalence of SBP in cirrhotic ascites. The diagnosis of SBP in patients with cirrhosis requires an ascitic fluid polymorphonuclear cell count > 250/mm3. Using leukocyte esterase reagent strips (the type used for urinalysis), these investigators collected data on a total of 2123 ascites samples in 1069 patients in this multicenter prospective study. The samples were tested with commercially available reagent strips, with each strip read twice by different members of the team to avoid investigator bias. One hundred and seventeen SBP episodes (5.5% prevalence) were identified among the samples. Among these, 56 were associated with a positive ascitic fluid culture. Using the threshold of 2+ for positivity of the reagent strip (ie, a positive reagent strip result of 2 or more), sensitivity was 43.6% and specificity was 99.2% for the diagnosis of SBP, with a positive predictive value of 76.1% and negative predictive value of 96.8%. With regard to the sensitivity of the test, these results are not as good as those previously reported by other groups -- although specificity was excellent. Thus, the study authors concluded that a negative result on reagent strip cannot exclude the diagnosis of SBP. Varices Background The management of portal hypertension encompasses therapy to prevent hemorrhage in patients who have never bled (primary prophylaxis), treatment for those who are actively bleeding, and therapy for those who have previously bled (secondary prophylaxis). In primary prophylaxis, after a screening endoscopy, if medium- or large-sized varices are present, nonselective beta-blockers (ie, propranolol or nadolol) remain the treatment of choice. The drug dose is increased until heart rate decreases by 25%, with a frequency between 50 and 60 beats/minute. Specific hemostatic treatments for variceal bleeding include vasoactive drugs that decrease portal pressure, such as terlipressin* or somatostatin, endoscopic band ligation, and surgical portosystemic shunts or TIPS (transjugular intrahepatic portosystemic shunt). As secondary prophylaxis, both beta- blockade and endoscopic band ligation can be used. When possible, hepatic venous pressure gradient (HVPG) should be monitored, with the goal of reduction to < 12 mmHg; otherwise, the goal of reducing heart rate by 25% is acceptable. HVPG provides valuable prognostic information in the pharmacologic management of portal hypertension in helping to predict the risk of variceal hemorrhage. Primary Prophylaxis With regard to prophylaxis of the first variceal bleed, Drastich and colleagues, from the Czech Republic, conducted a prospective, multicenter, randomized trial to compare the efficacy and safety of propranolol therapy vs endoscopic variceal ligation in the prophylaxis of the first variceal bleed. Of 73 cirrhotic patients with high-risk varices (> 5 mm in diameter), 33 received propranolol (at a dose sufficient to decrease baseline heart rate by 25%) and 40 underwent endoscopic variceal ligation until eradication. They found no significant difference in the actuarial risk for first variceal bleed between the 2 groups (9% for endoscopic variceal ligation and 20% for the propranolol group). As to associated complications, there was 1 bleed related immediately to endoscopic variceal ligation and no complications in the propranolol group. The study authors concluded that both propranolol and endoscopic variceal ligation seem to be safe and effective in this setting. In fact, previous randomized controlled trials have shown that variceal banding is as effective as beta-blockers for preventing a first variceal bleed. However, it is still recommended to initiate management with beta-blockers if there are no contraindications because these agents are inexpensive, easy to use, and relatively safe, considering prophylactic band ligation of varices in those patients who cannot tolerate the appropriate doses of beta-blockers. Another interesting study concerning portal hypertension-related bleeding was presented by Beauchant and colleagues from France. This study aimed to examine French management practices in this setting. This cross-sectional survey examined 127 bleeding events and involved 103 men and 23 women; mean age was 58.5 years. Cirrhosis was linked to alcohol in 85% of patients. The bleeding event was the first episode in 79 (62%) patients, 40 (51%) of whom had known cirrhosis and 25 had prior endoscopy (with stage II esophageal varices found in 18 of 25, and prophylaxis with beta-blockers given in 14 of 18 cases). The main causes of bleeding were esophageal varices in 73% of cases, other varices in 10% of cases, and portal hypertensive gastritis in 5%. Vasoactive treatment was given in 89% of cases, endoscopic ligature in 42%, sclerosis in 21%, and/or tamponade in 6% of cases. The study authors concluded that although management was usually in line with general consensus, one third of known cirrhotics did not have endoscopy or prophylaxis before the first bleeding event. Secondary Prophylaxis As mentioned previously, HVPG provides valuable prognostic information in the pharmacologic management of portal hypertension. After a variceal bleed, a spontaneous decrease in HVPG has been observed in some patients treated with endoscopic procedures, but the impact of a spontaneous hemodynamic response on clinical outcome in these patients has not been extensively investigated. In this setting, Aracil and colleagues conducted a study to assess the influence of such a spontaneous response in the prevention of variceal rebleeding. This study included 99 patients treated with an endoscopic procedure (ligation or sclerotherapy) to prevent rebleeding after an acute bleeding episode. HVPG measurements were performed at baseline (once the acute bleeding was controlled) and again 1-3 months later (once varices had been eradicated). A spontaneous decrease in HVPG by > 10% was observed in 22% of patients (responders). Variceal eradication occurred more frequently in responders than in nonresponders (94% vs 58%; P < .001), and variceal rebleeding was less frequent in responders (20% vs 42% in nonresponders at 2 years; P = .04). On the basis of these findings, the study authors suggested that efficacy of endoscopic therapy may be related to spontaneous hemodynamic response. TIPS have long been used to lower portal pressures in the setting of variceal hemorrhage and refractory ascites, their major associated complications being the development of hepatic encephalopathy and shunt stenosis, which is seen in up to 50% of patients within 6 months. The polytetrafluorothethylene (PTFE)-covered stents* appear to have the potential to improve TIPS patency. During this year's EASL meeting, Angeloni and colleagues prospectively compared TIPS safety and 1-year patency in 56 patients using PTFE-covered stents vs in 87 patients treated with conventional stents. The 1-year probability of remaining free of shunt dysfunction was significantly higher in the covered stent group compared with the conventional stent group: 82% vs 57.5%; P = .003. The probability of rebleeding was significantly lower among those patients in the covered stent group. Although the number of hepatic encephalopathy episodes was similar between the 2 groups in the first 12 months after TIPS, 6 of the 56 patients treated with covered stents had intractable episodes of hepatic encephalopathy, and none of the patients in the conventional stent group suffered from recurrent intractable hepatic encephalopathy. The study authors concluded that although PTFE-covered stent-grafts have a higher patency and a lower rebleeding rate, a higher incidence of intractable hepatic encephalopathy may occur. Kalambokis and colleagues, from Greece, compared the hemodynamic effects of terlipressin, somatostatin, and terlipressin plus somatostatin in cirrhotic patients with portal hypertension; they also examined the effect of each treatment on sodium renal excretion. Terlipressin monotherapy significantly reduced portal vein velocity and portal flow volume, and increased mean arterial pressure (MAP) and systemic vascular resistance along with a decrease in cardiac output. In patients without ascites, it significantly increased fractional excretion of sodium, although it did not change in patients with ascites. Somatostatin monotherapy did not alter portal hemodynamics, whereas it significantly reduced MAP, heart rate, and -- in patients with ascites -- fractional excretion of sodium. The addition of terlipressin to the regimen induced similar changes in hemodynamic parameters and sodium excretion to those observed after receiving terlipressin alone, although the increase in MAP was significantly lower than in those patients receiving terlipressin alone. The study authors concluded that the use of terlipressin as monotherapy appears to be advantageous, whereas a combination of somatostatin and terlipressin does not seem to exert an additive portal hypotensive effect in cirrhotic patients. Prevention of Infection Another important issue concerning the management of cirrhotic patients with upper gastrointestinal bleeding is the risk of infection. During General Session 1: Clinical Hepatology, Fernandez and colleagues from Barcelona, Spain, presented the results of a randomized, multicenter, controlled trial, comparing oral norfloxacin (400 mg twice daily for 7 days; n = 55) vs intravenous ceftriaxone (1 g/day for 7 days; n = 53) in the prevention of bacterial infections in 108 cirrhotics with severe liver failure and gastrointestinal bleeding within 10 days after inclusion. They found that the probability of developing bacterial infections during the study period was significantly lower in patients receiving ceftriaxone than in those treated with norfloxacin (11% vs 27%; P = .02), with all types of infections except pneumonia occurring less frequently in the ceftriaxone group (urinary tract infection, 5.7% vs 14.5%; spontaneous bacteremia, 0% vs 1.8%; pneumonia, 3.8% vs 1.8%). The study authors concluded that intravenous ceftriaxone was more effective than oral norfloxacin in the prevention of bacterial infections in cirrhotic patients with severe liver failure and upper gastrointestinal bleeding. Hepatic Encephalopathy Background Hepatic encephalopathy is a complex neuropsychiatric syndrome due to hepatic failure. In most cases, the diagnosis is clinical, and relies on the exclusion of other causes of disordered mental status. In difficult cases, the use of brain imaging (computed tomography or magnetic resonance) may be useful. Unfortunately, there have been no major advances in this area during the last decade. Therapy is primarily based on supportive care and the treatment of common precipitating factors, such as gastrointestinal bleeding, infections, renal and electrolyte disturbances, psychoactive medications, volume depletion, constipation, excessive dietary protein intake, and the presence of shunts (surgical or TIPS). Oral lactulose is still the first-line pharmacologic treatment for hepatic encephalopathy, the recommended dose being 15-45 mL given orally every 6-8 hours, adjusted to achieve 3 bowel movements. Several oral antibiotics can be used in this setting, such as neomycin (dose = 3-6 g/day), metronidazole (dose = 250 mg/day), and rifamixin (dose = 1200 mg/day). More controversial therapies include L-ornithine L-aspartate, flumazenil,* and bromocriptine, whose efficacy have not been demonstrated. Risk Factors and Treatment Sánchez-Muñoz and colleagues, from Seville, Spain, presented the results of a study comparing the impact of hepatic encephalopathy on survival in 1037 patients with hepatitis C virus (HCV)-related cirrhosis and 180 patients with HCV-related cirrhosis who were coinfected with HIV. Hepatic encephalopathy was more often seen in coinfected patients (14% vs 7.3%; P < .001), and was also the cause of death in 46% of HCV/HIV-coinfected vs 22% of HCV-infected patients. The 1- and 3-year estimated survival rates after development of hepatic encephalopathy were 38% and 23% in coinfected patients, and 52% and 31% in those with HCV-related cirrhosis; P < .05. The study authors concluded that hepatic encephalopathy was more frequent and shows a more severe course in HCV/HIV-coinfected patients, which may in part be explained by the enhanced systemic inflammatory response seen in these coinfected patients. Hyponatremia may occur in patients with cirrhosis and is associated with considerable changes in brain cells. Torre and colleagues, from several centers in Spain, hypothesized that hyponatremia-induced changes in the brain may predispose to the development of hepatic encephalopathy. They analyzed the incidence and predictive factors of hepatic encephalopathy in 70 patients with cirrhosis and refractory ascites who had no hepatic encephalopathy at entry. During a median follow-up of 10 months, 71% of patients developed at least 1 episode of hepatic encephalopathy. In multivariate analysis of several variables, including demographic and clinical data as well as liver and renal function obtained at randomization and in the absence of treatment for 5 days, only age and severe hyponatremia (medium sodium < 125 mEq/L) were associated with an independent predictive value for development of hepatic encephalopathy. Severe hyponatremia increased the risk of hepatic encephalopathy 3.2 times. This finding is interesting in that it suggests new possibilities for investigating intracerebral factors leading to the development of hepatic encephalopathy in patients with cirrhosis and refractory ascites. Abid and colleagues, from Pakistan, conducted a randomized study to examine the efficacy of intravenous infusion of L-ornithine L-aspartate* in patients with hepatic encephalopathy due to liver cirrhosis. They randomized 120 patients to receive L- ornithine L-aspartate 20 g/day (n = 60) or placebo (n = 60) in 100 mL 5% dextrose infusion over 4 hours for 4 consecutive days. The 2 groups were comparable at baseline. Hepatic encephalopathy improved by at least 1 stage on day 4 in 53 patients given L- ornithine L-aspartate vs in 44 patients on placebo (P = .016). Duration of hospital stay was shorter in the L-ornithine L-aspartate group (3.9 vs 5.6 days; P = .04). Mortality was similar in the 2 groups and there were no side effects associated with L-ornithine L- aspartate. Thus, the study authors concluded that infusions of L-ornithine L-aspartate appear to be safe and effective in the treatment of hepatic encephalopathy, resulting in improvement in hepatic encephalopathy and reduction in hospital length of stay. Concluding Remarks On the basis of data reported during this year's EASL meeting, it seems that in cirrhotic patients with diuretic-resistant ascites, large-volume paracentesis can be safely performed using lower doses of albumin, thereby reducing the associated costs. In cirrhotic ascites, attention must be paid to the prevention of volume depletion because the associated deterioration in renal function can worsen the prognosis. Additionally, the use of reagent strips for a rapid diagnosis of SBP was not confirmed to be reliable, as reported in previous studies. Additionally, as addressed in this overview, the utility of terlipressin is increasingly being recognized in the management of HRS as well as in the management of variceal bleeding, showing more favorable hemodynamic effects than somatostatin. As addressed further during these meeting proceedings, the prevention of infection during esophageal bleeding with intravenous ceftriaxone was shown to be more effective than oral norfloxacin. Also, although TIPS using PTFE-covered stent-grafts show a higher patency and lower rebleeding rate, the fact that this approach is associated with a higher incidence of intractable hepatic encephalopathy warrants consideration. As a final point of commentary, the primary prophylaxis of variceal bleeding is mandatory, and beta- blockers remain the first option in management. It is evident that the management of cirrhosis and its related complications conitnue to pose a significant challenge to the practicing gastroenterologist-hepatologist, and will remain a major focus of many research efforts in the field. Supported by an independent educational grant from Gilead. *The US Food and Drug Administration has not approved this medication for this use.