Acute gastroesophageal variceal hemorrhage is one of the defining clinical complications of decompensated cirrhosis and carries high morbidity and mortality. There have been significant advances in the medical management over the last 20 years for variceal hemorrhage, including the integration of endoscopic band ligation, vasoactive drugs, and antibiotic prophylaxis as standard of care. However, acute gastroesophageal variceal hemorrhage still carries a 6-week mortality rate of 15-20%.
The rationale for the use of vasoactive medications, including vasopressin and its analog terlipressin, and somatostatin along with its analogs octreotide and vapreotide, is to produce splanchnic vasoconstriction and reduce portal blood inflow and portal pressure, thus decreasing portal-variceal flow and pressure. Numerous trials have examined the effectiveness of vasoactive drugs in the setting of acute variceal hemorrhage, and a recent meta-analysis of the 30 best studies demonstrates that the use of vasoactive agents is associated with a reduction in 7-day mortality, improved control of bleeding, lower transfusion requirements, and a shorter hospital stay when compared to placebo. Interestingly, the aforementioned meta-analysis does not find a significant difference in hemostasis when comparing the different vasoactive drugs, although the reported quality of evidence is low. Conversely, the analysis does reveal that octreotide (11 trials) and terlipressin (7 trials) provide a significant advantage in acute hemostasis when compared to placebo.
In this issue of Hepatology, Seo et al. provide a well-executed multicenter study, although not double-blinded, that validates prior conclusions. This large, noninferiority, randomized trial investigated the use of the three different vasoactive drugs combined with endoscopic therapy on 780 patients presenting with acute variceal hemorrhage. Patients with known cirrhosis who presented with upper gastrointestinal hemorrhage at 11 centers throughout Korea were randomly allocated to receive terlipressin, somatostatin, or octreotide, followed by endoscopic therapy, in particular endoscopic band ligation if varices were present. The primary endpoint was control of acute bleeding without the need for “rescue” therapy, development of rebleeding, or death during the first 5 days after the index bleed. Sixty-seven percent (780/1160) of all screened patients were included in the final analyses after the exclusion of those who bled from a nonvariceal origin, had advanced hepatocellular carcinoma, or were lost to follow-up. The authors state that the 5-day treatment success was equivalent among the three drug arms (terlipressin, 86.2%; somatostatin, 83.4%; octreotide, 83.8%, P = 0.636), and that 88.5% of all enrolled patients had acute hemostasis. When comparing the three drug arms, the authors did not appreciate a statistical difference in the initial control of bleeding, 5-day rebleeding, or death (P = 0.752, P = 0.739, and P = 0.929, respectively). Given the intent of a noninferiority trial, the authors conclude that terlipressin, somatostatin, and octreotide are not inferior as adjunctive therapy with endoscopic band ligation for the management of acute variceal hemorrhage. However, we do note that the absolute values of the confidence interval lower boundary for somatostatin (0.095) and octreotide (0.090) nearly reach the authors' designated cutoff margin of 0.1 for relevant inferiority when each is compared to terlipressin. This study also reveals an overall 6-week mortality rate of 12% and that there are no differences in death among the three medication groups (P = 0.891). With regard to adverse events, hyponatremia is observed more frequently among patients receiving terlipressin (11.5%, P < 0.001), but other adverse event occurrences are not statistically different.
Another important finding from this article is the confirmation of clinical predictors for treatment failure. Many researchers have attempted to establish parameters that are important in the prognostication of treatment failure. One variable, the hepatic venous pressure gradient (HVPG) ≥20 mm Hg, when measured shortly after control of an initial variceal hemorrhage, clearly predicts treatment failure as well as 1-year mortality.[4-6] However, since HVPG measurement is not readily available at most institutions, additional studies have investigated their cohorts to discover clinical characteristics that are associated with treatment failure. Indices of hepatic disease severity, such as Child-Pugh classification, consistently correlate with treatment failure.[6, 7] A recent study further illustrates this point by demonstrating the use of another disease severity index, the Model for Endstage Liver Disease (MELD) score, when included into a modified risk-stratification model accurately predicts 6-week mortality risk among those presenting with acute variceal hemorrhage. Using multivariate analysis on their large cohort, Seo et al. report that low initial systolic blood pressure, elevated serum creatinine, the presence of active bleeding during the index endoscopy, bleeding from a gastric origin, and Child-Pugh Class C disease, are all associated with 5-day treatment failure. Inspecting their Supporting Tables 1 and 2, the authors' cohort clearly exhibits a considerable drop in 5-day treatment success when gastric varices are present (88-91% treatment success for esophageal varices and gastroesophageal varices [GOV] type 1 compared to 55-87% for GOV type 2 and isolated gastric varices) and when patients have advanced liver disease (88-93% treatment success for Class A and B compared to 65-76% for Class C). Clearly, all of the above parameters need to be taken into account to tailor treatment strategies based on patients' risk. Low-risk patients would likely still do well using a “trimmed-down” approach (monotherapy or shorter courses of therapy), whereas high-risk patients would likely benefit more with additional methods of portal pressure reduction.
One suggestion is to use preemptive or “early” transjugular intrahepatic portosystemic shunt (TIPS) placement among selected high-risk patients. This statement is supported by a pivotal multicenter trial that randomized high-risk patients (defined as having Child-Pugh class B disease and active bleeding or Child-Pugh Class C disease, <13 points) presenting with acute variceal bleeding to either receive TIPS placement within 72 hours of the initial bleeding episode or standard care consisting of vasoactive therapy along with endoscopic band ligation. The study group shows that patients who receive TIPS placement have a significant advantage in control of hemorrhage and survival benefit compared to those receiving standard combined therapy. The same collection of study centers also confirms these findings in a follow-up validation study. Seo et al. note that only 3 patients out of 90 treatment failures received TIPS in their study (study enrollment was prior to the “early” TIPS publications), and we could assume that some of their treatment failures would likely have benefited from early TIPS. Then again, this study emphasizes the effectiveness of combined endoscopic and vasoconstrictive therapy since the authors are still able to achieve a 6-week survival of 88% without applying “early” or rescue TIPS, even when the vast majority of enrolled patients had either Child B (42.7%) or Child C (32.8%) disease. Further prospective studies are still needed to justify how we risk-stratify patients and properly assign cost-effective therapy prior to these approaches being implemented as standard practice.
Based on the considerable amount of literature supporting the use of vasoactive medications, these drugs are universally accepted and included in society guidelines, including the American Association for the Study of Liver Diseases (AASLD), as adjunctive therapy with endoscopic band ligation for the management of acute variceal hemorrhage. In the United States, octreotide is the only commercially available vasoactive drug but it is not approved by the Food and Drug Administration (FDA) for this indication. In fact, no medication is currently approved by the FDA for the treatment of acute or chronic clinical manifestations of portal hypertension. The protocol for drug approval by the FDA typically requires a randomized controlled trial involving an experimental drug against the standard of care in order to establish efficacy as well as safety. In this instance, there are no current accepted medications to use as controls and finding a sponsor for one of the available drugs will be difficult since most are generic. Thus, we stand at a crossroads. American researchers are faced with a conundrum on how to participate in the worldwide efforts within this important field of research. Perhaps our professional societies could meet with the FDA and develop a plan of action on how to obtain drug labeling for the management of acute variceal hemorrhage and portal hypertension. Then investigators in the U.S. will be able to successfully participate with industry on studies that could lead to advances in the treatment of this lethal complication of chronic liver diseases.
Brett E. Fortune, M.D., M.Sc., and Roberto J. Groszmann, M.D., FRCP
Yale University School of Medicine New Haven, CT