This study evaluated the efficacy of antiviral therapy with different nucleoside analogs in patients with HBV-related cirrhosis and EV. Eligible patients with HBV-related cirrhosis who consulted two tertiary hospitals in Beijing, China, the Chinese Second Artillery General Hospital and Chinese PLA General Hospital, were enrolled in the study from January 2005 to December 2009.

The study population had HBV-related cirrhosis and EV. All patients had serum HBV DNA > 500 copies/mL as measured by polymerase chain reaction (PCR) (Fluorescence Quantitation kit; Shanghai Kehua Bio-engineering, Shanghai, China). Exclusion criteria included: history of hepatitis C or D, viral hepatitis, evidence of alcoholic cirrhosis, history of surgery for portal hypertension (splenectomy, shunt or devascularization), suspected liver cancer, hepatic encephalopathy or hepatorenal syndrome, or life-threatening diseases.

In accordance with clinical practice guidelines[11-13], endoscopic eradication of varices was performed in cases with a history of bleeding. Other treatment modalities, such as polyene phosphatidylcholine, reduced glutathione and diuretics, were administered as required. Propranolol was prescribed to all patients except for those who could not tolerate, or had contraindications to, beta-blocker use.

Because no evidence-based suggestions on antiviral therapy in patients with esophageal varices were available, the benefit and potential harm (such as drug resistance and adverse effects) of antiviral therapy were introduced to the patients. Patients received one of the following antiviral treatments immediately after enrollment: lamivudine (LAM) (100 mg/d, Glaxo Wellcome, United Kingdom), adefovir (ADV) (10 mg/d, Tianjin Pharmaceutical, China), entecavir (ETV) (0.5 mg/d, Bristol-Myers Squibb, United States), telbivudine (LdT) (600 mg/d, Novartis, Switzerland), a combination of LAM and ADV (LAM 100 mg/d + ADV 10 mg/d), or no antiviral therapy according to their wishes. As antiviral therapy may have a greater risk in patients with esophageal varices, it was not prescribed if the patients refused it for reasons such as adverse effects and resistance.

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Ethics Committee of both Hospitals. All patients provided written informed consent prior to enrollment in the study.

Patients were followed up at a 3-mo interval over 5 years, and levels of HBV DNA, glutamic-pyruvic transaminase (GPT), glutamic-oxaloacetic transaminase (GOT), total bilirubin, serum albumin, prothrombin time, platelet count, diameter of portal vein and splenic area on ultrasound scan were recorded.

Endoscopic findings were graded according to the criteria of the Japanese Association of Portal Hypertension[14] as Grade I, II or III, and given a score of 1, 2 or 3, respectively. Bleeding was scored as 4. Patients whose varices were eradicated were scored as 0. Those who showed newly visible, small vessels after endoscopic eradication, but did not reach grade I were scored as 0.5. Patients without a history of bleeding underwent endoscopic examination every 2 follow-up visits (at a 6-mo interval). Cases with a history of bleeding underwent endoscopic eradication of the varices and follow-up at month 3, and then at a 6-mo interval over the remaining study period. Variceal scores were recorded.

The end-point of this observational study was bleeding or time to study conclusion (December 2009) if the patient did not experience any bleeding. The duration from study enrollment to bleeding or to study conclusion was defined as non-bleeding duration.

The incidence of adverse events, discontinuations and deaths were documented. Complete blood counts, blood urea nitrogen, creatinine, creatine kinase, lactic acid and electrolytes were also monitored.

The software STATA 10.0 (Stata Corporation, United States) was used in the statistical analysis of data. The bleeding rates (defined as the cumulative proportion of patients who experienced an esophageal variceal bleeding episode during the study) in the antiviral and control groups were compared using the χ² test. Multivariate Cox proportional hazard regression was used to identify factors related to esophageal variceal bleeding. Non-bleeding duration of the different groups was determined using the Kaplan-Meier model. Increases in variceal score were compared between groups using the Student’s t test. The portal vein diameter and splenic area on ultrasound scan, levels of GPT, GOT, total bilirubin, serum albumin, prothrombin time and platelet counts were also compared using the Student’s t test. All tests were two-sided and a P value of < 0.05 was considered statistically significant.

A total of 117 patients who fulfilled the study criteria were enrolled from January 2005 to December 2009. Of these, 79 patients received antiviral therapy, including ETV (n = 29), ADV (n = 28), LAM (n = 15), LdT (n = 4) and combination treatment with LAM and ADV (n = 3). Patients not receiving antiviral therapy (n = 38) were followed up as a control group, and received the same treatment as the antiviral group minus antiviral therapy. Of the total study population, 63 patients (42 in the antiviral group and 21 in the control group) who had a history of bleeding underwent endoscopic eradication of varices, while the other 54 cases (37 in the antiviral group and 17 in the control group) without a history of esophageal variceal bleeding did not undergo endoscopic eradication treatment. Primary prevention was not performed in this study population. In addition, a total of 95 patients (65 patients in the antiviral group and 30 patients in the control group) received propranolol. Both propranolol and endoscopic eradication patterns were not significantly different between the two groups (P = 0.801 and P = 0.846, respectively). The antiviral treatment and control groups were well matched with regard to gender, age, hepatic function, grade of EV, HBV DNA level, HBeAg status, portal vein diameter and splenic section area on ultrasound (Table 1).

By week 12 of treatment, HBV DNA decreased to undetectable levels (defined as < 500 copies/mL, by PCR) in 58 (73.4%) of the 79 patients in antiviral treatment group. This was observed in 23 (79.3%) patients in the ETV group, 19 (67.8%) patients in the ADV group, 12 (80.0%) patients in the LAM group, 2 patients in the Ldt group (n = 4) and 2 patients in the LAM/ADV treatment group (n = 3). The number of patients achieving undetectable HBV DNA levels increased to 67 at week 24. At week 48, a total of 71 patients in the antiviral group achieved undetectable HBV DNA levels. None of the patients in the control group (n = 38) achieved undetectable HBV DNA during the study.

The portal vein diameter and area of splenic section tended to increase at the conclusion of the study compared to baseline. However, there was a smaller mean increase in the diameter of the portal vein and area of splenic section on ultrasound in the antiviral group compared to the control group (Table 2). Analysis of biochemical profiles (alanine aminotransferase, aspartate aminotransferase, serum albumin, total bilirubin and prothrombin time) showed progressive improvements in liver function in the group receiving antiviral treatment (Table 2). Median Child-Pugh score increased from 8 to 10 in control patients (P = 0.018), and decreased from 8 to 7 in the antiviral group (P = 0.686).

By the end of the study, the bleeding rate was significantly decreased in the antiviral group (n = 79) compared to the control group (n = 38) (29.1% vs 65.8%, P < 0.001). The bleeding rate was also statistically different when all the patients were stratified into the endoscopic eradication group and non-endoscopic intervention group (Table 3). The mean ± SD increase in variceal score was reduced in the antiviral treatment group compared with the control group (Table 3).

Multivariate Cox regression analysis was performed to identify the factors associated with increased or decreased esophageal bleeding. The results showed that antiviral therapy (OR = 11.3, 95%CI: 3.1-38.5; P < 0.001), endoscopic eradication of varices (OR = 15.8, 95%CI: 4.1-51.1; P < 0.001), baseline portal vein diameter (OR = 39.1, 95%CI: 1.6-842.6; P = 0.025) and baseline serum HBV DNA level (OR = 0.8, 95%CI: 0.6-1.1; P = 0.042) were independent factors related to esophageal bleeding.

Four cases showed undetectable HBV DNA initially, however, viral load rebound was observed during follow-up, which was defined as virological breakthrough. All 4 cases bled in contrast to other patients who benefited from antiviral therapy.

In the LAM group (n = 15), three cases were found to have virological breakthrough during the study. All 3 cases had no history of bleeding and therefore were not treated with endoscopy. In the first case, virological breakthrough was detected at week 48 (1.8 × 10⁵ copies/mL) and ADV was administered immediately, however, the patient exhibited re-bleeding at week 53 and the HBV DNA level reached 3.3 × 10⁵ copies/mL. In the second case, the patient bled at week 192 and eventually died due to bleeding. Examination revealed HBV DNA at 5.1 × 10⁵ copies/mL. In the third case, the patient bled at week 90 and HBV DNA breakthrough was detected on examination (3.1 × 10⁸ copies/mL). The patient was treated with vasoactive drugs, endoscopic eradication of varices and switched to ETV after cessation of bleeding. HBV DNA levels were subsequently undetectable at 12 wk after treatment, and the patient was bleed-free for 48 wk by the end of the study.

One patient in the ADV group (without a history of bleeding and endoscopic intervention) experienced bleeding at week 130. HBV DNA was 9.8 × 10³ copies/mL at examination. The patient was switched to ETV after cessation of bleeding and was bleed-free for 36 wk by the end of the study.

As more cases suffered from virological breakthrough and bleeding, the bleeding rate in the LAM group was not statistically different from that in the control group (8/15 vs 25/38, P = 0.531), while the ETV (5/29) and ADV (8/28) groups showed a lower bleeding rate (P < 0.001 and P = 0.006, respectively) compared to the control group. ETV was statistically better than LAM (P = 0.019), and no significant differences were found between the other agents.

There was one case of bleeding in both the LdT (n = 4) and combination treatment (LAM/ADV) groups (n = 3), and due to the small number of cases, they were not included in the statistical analysis.

Kaplan-Meier analyses demonstrated that bleeding was postponed in the antiviral treatment group compared to the control group (Figure 1A). However, the curve of virological breakthrough cases was close to that of the control group (Figure 1A). The Kaplan-Meier curves of ETV, ADV and LAM crossed each other within the first 2 years, but deviated beyond the 2-year mark (Figure 1B).

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Kaplan-Meier analysis of non-bleeding duration. A: Non-bleeding duration in antiviral and control cases was compared using the Kaplan-Meier survival model. Bleeding was defined as a failed event. The occurrence of bleeding was reduced and delayed in the antiviral treatment group compared to the control group. The curve of virological breakthrough cases (dashed line) was close to that of the control group, demonstrating reduced efficacy when virological breakthrough occurred; B: Non-bleeding durations for entecavir (ETV), adefovir (ADV) and lamivudine (LAM) therapy were compared using the Kaplan-Meier survival model. Bleeding was defined as a failed event. The non-bleeding durations for ETV, ADV and LAM were similar in the first 2 years, however, differences became clear following long-term treatment, which may have been due to cumulative resistance and bleeding.

One patient died of esophageal variceal bleeding as stated above; no other severe adverse events were reported in the study. Most reported side effects were mild such as fatigue (n = 1), headache (n = 1), dyspepsia (n = 2), nausea (n = 1), dizziness (n = 1) and insomnia (n = 1). Discontinuation of treatment due to adverse events was not observed.

In this study, follow-up ended if the patient bled. Additional follow-up was included to determine the effect of antiviral therapy on patient survival. The cumulative 1-year survival rate was 97.5% (77/79) and 89.5% (34/38) for the antiviral group and control group, respectively (P = 0.086). The cumulative 2-year survival rate was 93.7% (74/79) and 86.3% (29/38) for the antiviral group and control group, respectively (P = 0.012). However, 8 patients in the control group switched to antiviral therapy. When these 8 patients were excluded, the cumulative 2-year survival rate in the control group was 70.0% (21/30), which was also lower than in the antiviral group (P = 0.002). Most of the patients in the control group switched to antiviral therapy during the additional follow-up period, and a comparison of the 5-year survival rate of the control group and antiviral group was not available. In addition to the patient who died of drug resistance and bleeding, 17 patients died (9 due to bleeding, 3 due to hepatic encephalopathy, 2 due to chronic liver failure and 3 due to liver cancer) and 17 were lost during the additional follow-up in the antiviral group. The overall 5-year survival rate in the antiviral group was 71.0%, which was higher than in the general cirrhotic patients[1].

Antiviral therapy with nucleoside analogs improves the clinical outcome of hepatitis B virus (HBV) infection. However, the emergence of resistance results in liver injury. The consequences may be worse in patients with esophageal varices (EV), in which bleeding and death often occur.

Recently, an on-line published study demonstrated that entecavir therapy reduced the risks of hepatic events in hepatitis B cirrhosis patients within 5 years. Koga et al reported an improvement in esophageal varices in patients following lamivudine (LAM) treatment in a retrospective study. However, the study only consisted of 12 patients treated with LAM and 6 controls, and there were no data on bleeding rate and the effect of virological breakthrough. As patients usually experience bleeding when virological breakthrough takes place, it is necessary to separate the patients and evaluate the harm of drug resistance and the benefit of HBV suppression in these patients to provide evidence-based treatment suggestions.

The present study evaluated the efficacy of antiviral treatment over 5 years in patients with HBV-related cirrhosis and EV, and found that antiviral therapy decreased the risk of bleeding; and agents with a high rate of virological breakthrough were ineffective in preventing bleeding. These findings for the first time provide evidence-based treatment suggestions for this special group of patients.

Antiviral therapy with nucleoside analogs may delay progression of varices, decrease the risk of bleeding and improve liver function in patients with HBV-related cirrhosis and EV. Agents with a high rate of virological breakthrough may not be effective in the long-term prevention of esophageal variceal bleeding.

Nucleoside analogs, are molecules that can inhibit HBV polymerase, and therefore decrease HBV load. They have been widely used in the treatment of hepatitis B. Virological breakthrough indicates that the viral load decreased to an undetectable level initially, and rebounded during follow-up. Virological breakthrough often leads to progressive liver injury and even severe liver failure.

The article provides statistical analysis of data from HBV patients who went through antiviral therapy to determine the role of antiviral therapy in esophageal varices and bleeding in HBV-related cirrhosis. This article provides information important to clinicians and HBV patients in general.