Long-term effects of sustained virologic response on the development of esophageal varices in compensated cirrhosis: “Is the bang worth the buck?”

Authors

  • Richard K. Sterling M.D., M.Sc.

    Corresponding author
    1. Division of Gastroenterology, Hepatology, and Nutrition, Virginia Commonwealth University Health System, Richmond, VA
    • Virginia Commonwealth University, Hepatology Section and Liver Transplant Program, VCU Medical Center, 1200 East Broad Street, West Hospital, Room 1492, Richmond, VA 23298-0341
    Search for more papers by this author
    • fax: 804-828-4945


  • See Article on Page 2069.

  • Potential conflict of interest: Dr. Sterling is on the speakers' bureau of, and received grants from, Roche and Merck. He is on the speakers' bureau of Vertex. He advises and received grants from Bayer. He also received grants from GlaxoSmithKline and Bristol-Myers Squibb.

The natural history of compensated cirrhosis in those with chronic hepatitis C virus (HCV) has been well described. In the landmark study by Fattovich et al. of 384 compensated subjects, the 5-year risk of hepatocellular carcinoma (HCC) was 7% and the risk of hepatic decompensation was 18%.1 Of the 355 patients who remained tumor-free, 65 (18%) developed at least one episode of ascites (8.7%), jaundice (1%), hepatic encephalopathy (1.5%), or variceal bleeding (4%), and the mean time to decompensation was 37 months (range, 3-137). In a more recent study, 131 of 352 (37%) subjects with compensated HCV-induced cirrhosis who were followed for a median of 14.4 years developed decompensation.2 Of the 77 (59%) subjects who were without HCC, 66 (86%) developed ascites, 22 (28%) developed portal hypertensive bleeding, and 21 (27%) developed hepatic encephalopathy. Importantly, those with varices had twice the rate of decompensation compared to those without varices (65% versus 33%). Therefore, development of portal hypertension seems to be an important predictor of decompensation and increased mortality.3

Abbreviations:

HCC, hepatocellular carcinoma; HCV, hepatitis C virus; HVPG, hepatic venous pressure gradient; SVR, sustained virologic response.

The development of varices is one of the hallmarks of significant portal hypertension and the incidence of new varices in those with cirrhosis is <5%/year.4 In those without varices, the development of varices is related to the severity of underlying liver disease and the presence of increased hepatic venous pressure gradient (HVPG) of more than 10 mm Hg. In the study by Groszmann et al. which examined use of beta-blockers to prevent esophageal varices in patients with stable cirrhosis (62% with HCV) without esophageal varices at baseline, the rate of developing varices was similar between those subjects randomized to Timolol and placebo (42 of 108, 39% versus 41 of 105, 40%) during a mean follow-up of 55 months.5 Although the majority of varices were small, a few patients in each group developed large varices and subsequently bled. However, varices developed less frequently in those with a baseline HVPG < 10 mm Hg and in those who had less than 10% decrease in HVPG at 1 year.

The potential benefit to HCV therapy, in addition to sustained virologic response (SVR), is improvement in outcomes. Because many treated individuals who achieve SVR do not have significant fibrosis, this benefit may not be realized for several years, if not decades. However, although those with advanced fibrosis have poorer response to current therapy,6 they also have the most to gain. In support of this, studies have shown that those with advanced cirrhosis who achieve SVR have fewer clinical outcomes including liver failure, variceal bleeding, and HCC2, 7-9 (Table 1).

Table 1. Benefits of Virologic Response on Outcomes in Patients with HCV and Advanced Fibrosis
Author (reference) YearNStudy DesignStudy PopulationOutcome
  1. HALT-C, Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis; HCC, hepatocellular carcinoma; HVPG, hepatic venous pressure gradient; SVR, sustained virologic response; stage 3 or advanced fibrosis = bridging fibrosis and stage 4 = cirrhosis.

Rincon et al. (13) 200620ProspectiveStage 3 or 4 fibrosis and HVPG > 5 mm Hg studied before and after HCV therapyHVPG deceased in all but one patient (mean 28%) and was higher in those with virologic response vs. those without virologic response (26% vs. 13%, P = 0.05).
Bruno et al. (9) 2007920RetrospectiveCompensated cirrhosis, followed for a mean of 96 monthsLower incidence (per 100 person years) of liver complications, HCC, and liver-related death were 0, 0.66, and 0.19 in those with SVR compared to 1.88, 2.10, and 1.44 in those without SVR (P < 0.001).
Veldt et al. (7) 2007479RetrospectiveAdvanced fibrosis or cirrhosisSVR was associated with a reduction in liver-related death (P = 0.024) and liver failure (P = 0.001)
Roberts et al. (14) 200747ProspectiveCompensated cirrhosis, studied before and after HCV therapyThose with SVR had reductions in HVPG compared to those without (−2.1 vs. 0.6; P = 0.05). In those with baseline HVPG > 5 mm Hg, 71% of those with SVR had a decrease in HVPG of 20% compared to those without SVR (20%, P = 0.01).
Shiffman et al. (16) 2009764ProspectiveHALT-C trialThose with a ≥ 4-log decline in HCV RNA during the lead-in phase had fewer clinical outcomes (P = 0.003) regardless of whether they were randomized to maintenance therapy or not. No significant reduction in clinical outcomes was observed in the 30 patients who had persistent viral suppression during maintenance therapy.
Bruno et al. (2) 2009194ProspectiveCompensated cirrhosisThose with SVR had lower annual rates (per 100 person years) of decompensation (0.03), HCC (1.9), or liver-related mortality (1.0) compared to those without SVR (3.8, 2,8, and 2.4) or those who were untreated (3.5, 2.9, 3.0), respectively.

The mechanism associated with improved outcomes is presumed to be mainly from reduction in hepatic fibrosis. Poynard and colleagues pooled data on 3010 HCV treatment-naïve patients from four large clinical trials with pretreatment and posttreatment biopsies. They observed significant improvements in both inflammation and fibrosis in those with SVR (25% improved while 7% worsened) compared to those with nonresponse (17% improved while 21% worsened).10 Importantly, they found reversal of cirrhosis in 75 of 153 (49%) patients. Similar improvements in histology with SVR have been reported by others as well.11, 12

In addition to improvements in fibrosis, antiviral therapy may also directly affect HVPG. Rincon et al. studied 20 compensated patients with advanced fibrosis, by using liver biopsy and hepatic pressure measurements before and immediately after therapy with pegylated interferon and ribavirin.13 They found that all but one patient had a significant decrease in HVPG from baseline following antiviral therapy and that those with SVR had a greater reduction than those with nonresponse. The benefits of reductions in HVPG with SVR were confirmed by Roberts et al. in 47 patients with cirrhosis.14 Although the Hepatitis C Antiviral Long-Term Treatment Against Cirrhosis (HALT-C) trial did not show overall benefit of maintenance interferon,15 improved clinical outcomes were observed in those with significant viral suppression without SVR.16 Taken collectively, these data suggest that those with chronic HCV and advanced fibrosis who achieve SVR have reduced clinical outcomes, including variceal bleeding.9 However, the impact of SVR on the de novo development of varices was not specifically assessed in these analyses and remains unknown.

In this issue of HEPATOLOGY, Bruno et al. addressed the impact of SVR on the development of esophageal varices in a subgroup analysis of a large prospective database of subjects with compensated HCV-induced cirrhosis.17 In this study, consecutive HCV-positive subjects seen between January 1989 and December 1992 with compensated, Child A cirrhosis were screened for varices. Those with hepatitis B, human immunodeficiency virus, prior history of decompensation, or HCC within 6 months were excluded. Among the 352 patients screened, 218 who were free of varices at baseline and agreed to have follow-up endoscopy were included in the analysis. All 218 subjects had regular follow-up with surveillance ultrasound for HCC every 6 months and endoscopy every 3 years to identify de novo varices. Patients received HCV therapy as determined by current practice at that time, and SVR was defined as negative HCV RNA at 6 months after stopping therapy. The primary endpoints were development of de novo varices or HCC. Of the 218 patients, 149 (68%) received HCV therapy and 23% had SVR. During the median follow-up of 11.4 years, de novo varices developed in 67 patients and was similar in untreated (22 of 69, 32%) and treated (45 of 115, 32%) patients. The distribution of varices were small (F1, 76%) while 12% each had moderate (F2) or large (F3) varices. The median time between enrollment and detection of F3 varices (5 of 8 that bled) was 8 years (range, 3-17). Importantly, none of the 34 who achieved SVR developed varices; however, 66 patients developed HCC, including seven who had SVR. Multivariate analysis identified HCV genotype 1b, baseline model for end-stage liver disease (MELD) score, and HCC as independent predictors of the development of varices in nonresponders and those who did not undergo HCV therapy. There was no association observed with platelet count, albumin, international normalized ratio, or bilirubin with de novo varices.

Bruno and colleagues concluded that SVR prevents the development of varices and that endoscopic surveillance can be delayed or avoided in these patients. Finally, they suggest that a more tailored approach based on HCV genotype, MELD score, and HCC would help indentify those patients without SVR who are at higher risk for developing varices and who would benefit from surveillance endoscopy.

As with any long-term study, there are several caveats. Although they included patients from three centers, their results may not be generalizable to all patients with HCV-induced cirrhosis. Second, because not all patients screened were included and follow-up was not complete in all patients, there may have been a type 1 error. Also, we were not told of concurrent medications that might affect portal pressures and the development of varices. Nevertheless, this is the largest study with the longest follow-up to date that addresses the impact of SVR on the development of esophageal varices.

If these results are confirmed, there are several important implications for future management of cirrhosis in those who achieve SVR. First, this study highlights that those with SVR can still develop HCC and that all subjects with cirrhosis should continue periodic surveillance for HCC according to accepted guidelines.18 Second, because those with SVR do not develop varices, it may not be necessary to expose these patients to the expense and risks of repeated endoscopies. Third, because beta-blockers used to reduce HVPG do not seem to affect the rate of development of varices,5 the current study is the first to demonstrate a pharmacologic treatment to reduce (or in this case, eliminate) the development of varices. However, before we get too excited, we must remember that current treatment to achieve SVR in those with cirrhosis is difficult and there are often increased side effects, more cytopenias, and lower response rates than those without cirrhosis.6, 19, 20 Therefore, given the cost, both in dollars and resources, the increased side effects, and decreased response rates of HCV therapy, it remains to be determined if the “bang is worth the buck” in this select group of patients.

Ancillary