Treatment of Hepatitis B e Antigen–Positive Chronic Hepatitis with Telbivudine or Adefovir: A Randomized Trial FREE

—The Editors

In chronic hepatitis B virus (HBV) carriers, elevated serum HBV DNA levels have been associated with increased risk for end-stage complications, such as cirrhosis and hepatocellular carcinoma (1 - 4). Corresponding evidence supports the concept that antiviral therapy can ameliorate liver injury and reduce or delay the progression of cirrhosis and the incidence of hepatocellular carcinoma (5 - 9). Effectively maintained viral suppression has been correlated with these effects ((5), (10)). Recent studies further suggest that, with direct-acting anti-HBV nucleosides and nucleotides, therapeutic efficacy and the emergence of resistance are related to the degree of viral suppression achieved early in the course of treatment (11 - 18). Specifically, achieving low serum HBV DNA levels in the first 6 to 12 months—preferably levels undetectable by polymerase chain reaction (PCR) assay—has been associated with greater subsequent efficacy and less resistance, with observations now extending to treatment periods beyond 2 years. Maximizing early viral suppression is therefore emerging as an important therapeutic goal for patients with hepatitis B.

Telbivudine and adefovir are approved for use in the United States and elsewhere ((8), (19 - 22)). Telbivudine, a thymidine nucleoside analogue, had greater antiviral efficacy than lamivudine in phase IIb and III trials (22 - 23). Adefovir, an adenosine nucleotide analogue, had greater efficacy than placebo in phase III trials ((8), (21)). Telbivudine and adefovir have not been compared directly, although HBV DNA reductions after 1 year of telbivudine in phase III trials seemed greater than those reported with adefovir in similar patient populations ((8), (21 - 22)).

We compared the degree of telbivudine- and adefovir-induced viral suppression after 24 and 52 weeks of treatment to assess the relative antiviral efficacy and safety of these agents. Because trial data ((8), (22)) suggest that the clinical antiviral effects of telbivudine may be greater than those of adefovir, we also assessed antiviral efficacy and safety in patients switched to telbivudine after receiving adefovir treatment for 24 weeks. Finally, we evaluated potential relationships between the degree of viral suppression achieved after 6 months of therapy and subsequent efficacy outcomes at 1 year.

In this open-label trial, we randomly assigned hepatitis B e antigen (HBeAg)–positive patients with chronic hepatitis B to receive telbivudine or adefovir for 52 weeks (groups A and B, respectively) or adefovir for 24 weeks and then telbivudine for the remaining 28 weeks of the study (group C), evaluating the results at weeks 24 and 52. Screening began on 21 October 2004, and the last patient's last visit was on 16 August 2006. The study was conducted in compliance with the Declaration of Helsinki and in accordance with good clinical practice guidelines and applicable local regulations. All patients provided written informed consent, and all participating institutions received ethics committee approval. There was no data safety monitoring board.

We conducted this trial at 16 outpatient gastroenterology clinics at academic centers in Hong Kong, Australia, Canada, France, Korea, Singapore, Taiwan, Thailand, and the United States. Study investigators recruited participants from among their clinic patients after reviewing medical records and completing screening procedures to assess eligibility. Eligible patients were men or women with chronic hepatitis B 18 to 70 years of age, with no history or signs of hepatic decompensation, positivity for serum hepatitis B surface antigen (HBsAg), positivity for serum HBeAg, serum alanine aminotransferase (ALT) level between 1.3 and 10 times the upper limit of normal, and serum HBV DNA levels of at least 6 log₁₀ copies/mL. We excluded patients who were pregnant, breastfeeding, or co-infected with hepatitis C or D virus or HIV; had other known causes of liver disease, a history or signs of pancreatitis or hepatocellular carcinoma, or potentially confounding concomitant medical conditions; had ever been treated for hepatitis B with nucleoside or nucleotide analogues or had received interferon or other immunomodulatory agents within 12 months of screening; or had used alcohol or illicit drugs in the past 2 years. Patients with elevated serum creatinine levels, hemoglobin levels less than 110 g/L for men or less than 100 g/L for women, an absolute neutrophil count less than 2 × 10⁹ cells/L, platelet counts less than 100 × 10⁹ cells/L, α-fetoprotein levels greater than 50 µg/L, serum amylase or lipase levels at least 1.5 times the upper limit of normal, a prothrombin time prolonged by more than 3 seconds above the upper limit of normal, albumin levels less than 34 g/L, or total bilirubin levels at least 2 times the upper limit of normal were also excluded.

A centralized computer-generated process assigned eligible patients to each of the 3 treatment groups by using block randomization with block sizes of 3, implemented with an automated voice-response system. Eligible patients were randomly assigned in a 1:1:1 ratio to receive 600 mg (3 tablets) of telbivudine per day for 52 weeks (group A), 10 mg (1 tablet) of adefovir dipivoxil per day for 52 weeks (group B), or 10 mg of adefovir per day for 24 weeks followed by 600 mg of telbivudine per day for the remaining 28 weeks (group C). This design compared responses to telbivudine and adefovir at week 24, consistent with the primary study goal, and provided 3 treatment comparisons at week 52 (group A vs. group B vs. group C), as appropriate to secondary goals. Patients completing this study were offered enrollment in a follow-up study of continuing treatment with telbivudine if indicated.

Telbivudine was supplied by Idenix Pharmaceuticals, Cambridge, Massachusetts, and adefovir (Hepsera, Gilead Sciences, Foster City, California) was purchased commercially. The brittle nature of adefovir tablets precluded treatment blinding by routine overencapsulation. Investigators were blinded to HBV serologic data from baseline until week 52.

Routine laboratory values, including HBV DNA levels, were obtained along with clinical status at screening; baseline; and weeks 2, 4, 8, 12, 16, 24, 32, 40, 48, and 52. Other HBV serologic markers were measured beginning at week 12. A central reference laboratory (Quintiles Transnational, Research Triangle Park, North Carolina) conducted all laboratory-based assessments (HBV markers, serum chemistries, hematology, and urinalyses). Staff at Quintiles Transnational collected clinical data and entered them into a database in accordance with standardized data management and quality assurance procedures. Adherence was evaluated at each study visit; evaluation methods included pill counts and discussions with study participants.

Serum HBV DNA levels were assessed at each study visit by using the COBAS Amplicor PCR assay (Roche Molecular Systems, Branchburg, New Jersey), which has a lower limit of detection of 300 copies/mL. The primary treatment comparison was HBV DNA reduction from baseline values at week 24, with a secondary comparison at week 52. Other secondary efficacy measures included comparisons of mean residual HBV DNA levels, proportions of patients with HBV DNA who were PCR-negative or had HBV DNA values less than 5, 4, or 3 log₁₀ copies/mL; serum ALT normalization; HBeAg loss and seroconversion; HBsAg loss and seroconversion; and primary treatment failure. Primary treatment failure was defined as completion of at least 24 weeks of treatment without achieving 2 consecutive serum HBV DNA values less than 5 log₁₀ copies/mL, the minimal response threshold recommended by treatment guidelines when the study was designed (24 - 25). After study completion, suboptimal response was defined for exploratory analyses as a serum HBV DNA level of 3 log₁₀ copies/mL or greater at week 24, consistent with reports of greater resistance and poorer efficacy outcomes for patients with this level of residual viremia (11,17 - 18).

Viral breakthrough was defined as an on-treatment increase in HBV DNA over 2 consecutive determinations of greater than 1 log₁₀ from nadir, or a single determination at week 52, consistent with recent recommendations (26 - 27). Resistance (viral breakthrough with documented resistance mutations) was identified by PCR amplification of HBV DNA from sera of patients with viral breakthrough at week 52, followed by automated DNA sequencing of the entire 344-codon reverse transcriptase domain of the HBV polymerase gene at an independent reference laboratory (Delft Diagnostic Laboratory, Delft, the Netherlands).

At all study visits, we evaluated clinical adverse events, regardless of drug attributability, and discontinuation of study treatment for any reason; confirmed serum creatinine level elevations of 44.2 µmol/L or greater (≥0.5 mg/dL) from baseline; and conducted graded assessments of laboratory abnormalities. Adverse event evaluations included spontaneous patient reports, open-ended questioning by the investigator, and physical examinations. They were documented on case report forms that included investigator assessments of possible cause and drug attributability. We used standard, recommended, open-ended questions to elicit adverse event information rather than checklists or targeted questionnaires. Staff at Quintiles Transnational monitored each study site at 6- to 8-week intervals to ensure complete data entry into case report forms, including adverse events, and to verify source data.

With a target sample size of 120 patients and a 1:2 ratio of assignment to telbivudine or adefovir during the first 24 weeks, we estimated that our study had 98% power to detect a difference of 1.5 log₁₀ copies/mL in HBV DNA reduction at week 24, the primary efficacy comparison. This anticipated difference was based on previous clinical studies of these agents (8,11). The analyses presented are based on data available at each time point, consistent with the low rate of patient discontinuation; no missing data were imputed. All end points and comparisons were predefined except for exploratory evaluations of suboptimal responders and relationships between week-24 viral load and subsequent efficacy responses. For continuous variables, such as HBV DNA reduction, we used analyses of variance to compare treatment results. For categorical variables, we used logistic regression; we present the odds ratios. Our exploratory analyses of relationships between viral load at week 24 and subsequent efficacy outcomes were based on the Fisher exact test. All analyses were tested at a 2-sided α level of 0.05. We analyzed the results at weeks 24 and 52 by using unadjusted models and models adjusted for baseline covariates (HBV DNA level, age, body mass index, sex, and study site). We analyzed data by using SAS for Windows, release 8.02 (SAS Institute, Cary, North Carolina).

Idenix Pharmaceuticals and Novartis Pharmaceuticals provided funding for this study. The study protocol and statistical analysis plan were designed by the sponsors in conjunction with the academic investigators. The protocol was submitted to the U.S. Food and Drug Administration, and the study was conducted under an Investigational New Drug authorization from the U.S. Food and Drug Administration. Quintiles Transnational conducted data management and site monitoring and provided the study data to the study statistician, an employee of Idenix Pharmaceuticals, who completed preprogrammed data analyses.

Of the 208 patients screened, 72 were excluded and 136 were enrolled (Figure 1). One patient assigned to adefovir withdrew before the start of treatment. Of the remaining 135 patients, 45 received telbivudine (group A), 44 received adefovir (group B), and 46 received adefovir for 24 weeks and then telbivudine for the remaining 28 weeks (group C). Adherence to study medication was considered good (≥80%) for at least 95% of patients in each treatment group at each study visit, and no unplanned crossovers occurred. Treatment groups were well matched at baseline for the primary analysis at week 24 (group A and pooled groups B and C) and the final analysis at week 52 (groups A, B, and C) (Table 1). More than 90% of the enrolled patients were ethnically Asian, and baseline HBV DNA and ALT levels were similar across treatment groups.

We observed a consistent separation of mean serum HBV DNA levels according to treatment beginning at week 2 (Figure 2, top). At week 24, the reduction in mean serum HBV DNA level from baseline in group A differed from that in pooled groups B and C (−6.30 vs. −4.97 log₁₀ copies/mL; difference, −1.33 log₁₀ copies/mL [95% CI, −1.99 to −0.66 log₁₀ copies/mL]; P < 0.001), as did the proportion of patients whose serum HBV DNA levels were undetectable by PCR (39% vs. 12%; odds ratio, 4.46 [CI, 1.86 to 10.72]; P = 0.001) (Table 2). Serum HBV DNA levels remained at or above 5 log₁₀ copies/mL in more adefovir recipients than telbivudine recipients (42% vs. 5%; odds ratio, 0.07 [CI, 0.02 to 0.29]; P < 0.001). Similarly, group A and pooled groups B and C differed in the proportions of patients with HBV DNA levels that remained at or above 3 log₁₀ copies/mL (50% vs. 78%; P = 0.003) and 4 log₁₀ copies/mL (32% vs. 61%; P = 0.003). Rates of ALT normalization were similar (Table 2). Comparisons of all week-24 results, adjusted for baseline covariates, generally concur with the unadjusted results.

In patients switched from adefovir to telbivudine at week 24 (group C), mean HBV DNA levels rapidly decreased by approximately 1.4 log₁₀ copies/mL after week 24; within 8 weeks, they were nearly identical to levels in patients in group A (Figure 2, top). A corresponding increase in HBeAg seroconversion was evident in group C (Figure 2, bottom), although the differences were not statistically significant. At week 52, mean residual HBV DNA levels in groups A and C differed from those in group B (3.01 log₁₀ copies/mL and 3.02 log₁₀ copies/mL, respectively, vs. 4.00 log₁₀ copies/mL; difference, −0.99 log₁₀ copies/mL [CI, −1.67 to −0.32 log₁₀ copies/mL] and −0.98 log₁₀ copies/mL [CI, −1.64 to −0.32 log₁₀ copies/mL]; P = 0.004) (Table 3). Reductions of mean serum HBV DNA levels were greater in groups A and C (−6.56 and −6.44 log₁₀ copies/mL, respectively) than in group B (−5.99 log₁₀ copies/mL; differences, −0.57 log₁₀ copies/mL [CI, −1.41 to 0.27 log₁₀ copies/mL] [P = 0.18] and −0.45 log₁₀ copies/mL [CI, −1.28 to 0.38 log₁₀ copies/mL] [P = 0.28]). These differences were statistically significant after adjustment for baseline covariates (Table 3). More patients in groups A and C than in group B were PCR-negative at week 52, although these differences did not reach statistical significance (60% and 54% vs. 40%; P = 0.07 and 0.20, respectively). The rate of primary treatment failure (HBV DNA levels remaining >5 log₁₀ copies/mL through week 52) in group B (29%) also differed from that in group A (2%; odds ratio, 0.06 [CI, 0.01 to 0.48]; P = 0.008) and in group C (11%; odds ratio, 0.30 [CI, 0.10 to 0.96]; P = 0.042).

Loss of HBeAg was more common in group A than in pooled groups B and C at week 24, and was more common in groups A and C at week 52 (30% and 26%, respectively) than in group B (21%), although intergroup differences were not statistically significant (Tables 2 and 3). Hepatitis B e antigen seroconversion followed the same pattern. No patient experienced HBsAg loss or seroconversion. At week 52, ALT normalization occurred in 79% of patients in group A and 85% of patients in group C, compared with 85% of those in group B (P = 0.45 and 0.98, respectively).

Exploratory analyses indicated that for the 78% of adefovir recipients with suboptimal response at week 24 (HBV DNA level ≥3 log₁₀ copies/mL), those in group C displayed an additional reduction of 2.1 log₁₀ copies/mL in mean HBV DNA between weeks 24 and 52, compared with 0.8 log₁₀ copies/mL for patients in group B. Patients with a suboptimal response to adefovir who switched to telbivudine also had greater HBeAg seroconversion at week 52 than those who continued receiving adefovir (19% vs. 9%; P = 0.20), and a higher proportion was HBV DNA PCR-negative (42% vs. 24%; P = 0.11), although these differences were not statistically significant. Viral load was suppressed to less than 6 log₁₀ copies/mL by week 52 in all patients in group C, whereas levels remained above this threshold in 27% of those in group B, remaining as high as 8.9 log₁₀ copies/mL.

At week 24, serum HBV DNA levels were less than 3 log₁₀ copies/mL in more patients in group A than in pooled groups B and C (50% vs. 22%; P < 0.001). Among patients who received 52 weeks of telbivudine or adefovir (groups A and B), those with HBV DNA level <3 log₁₀ copies/mL at week 24 showed the highest rates of efficacy responses at week 52 (Table 4). In groups A and B combined, HBV DNA was undetectable at week 52 in 94% of patients with an HBV DNA level less than 3 log₁₀ copies/mL at week 24, versus 25% of patients with a viral load of 3 log₁₀ copies/mL or greater at week 24 (P < 0.001). Similarly, HBeAg seroconversion and ALT normalization at week 52 were observed in 44% and 94%, respectively, of patients with viral load <3 log₁₀ copies/mL at week 24, compared with 11% and 75% of patients with HBV DNA levels of 3 log₁₀ copies/mL or greater at week 24 (P = 0.001 and 0.04, respectively).

Viral breakthrough, defined as a confirmed increase in serum HBV DNA levels of more than 1 log above the nadir value, occurred in 4 adefovir recipients and 3 telbivudine recipients. No breakthroughs occurred in the adefovir-to-telbivudine group. All breakthroughs occurred after week 24 in patients with serum HBV DNA levels that remained at 3 log₁₀ copies/mL or greater at week 24. No codon A181V/T or N236T signature resistance mutations were detected in adefovir recipients with viral breakthrough (27). The signature M204I telbivudine resistance mutation was detected at week 52 in the 3 telbivudine recipients with breakthrough; 1 of these was accompanied by an L80V secondary change (28).

In the study protocol, viral breakthrough was defined as an increase of serum HBV DNA to 5 log₁₀ copies/mL or greater after a decrease to below that level. During the study, this definition was superseded by the more widely accepted definition that pertains to the data reported here (27). Breakthrough per protocol occurred in 2, 1, and 0 patients in groups A, B, and C, respectively. All 3 patients were also included in the primary resistance analysis reported here.

No drug-attributed serious adverse events were reported, no patients discontinued the study because of adverse events (1 patient withdrew after becoming pregnant), and no deaths occurred. Serious adverse events not attributed to study medications included a metacarpal fracture, transient elevation of ALT level, and hospitalizations for tonsillitis and a thyroid nodule. Clinical adverse events were similar in the 3 treatment groups (Table 5). Single cases of mild (grade 1) myopathy and persistent myalgia with creatine kinase elevations were reported in telbivudine recipients after 52 and 41 weeks, respectively; treatment was continued without dose modification. The serum creatinine level was elevated (>114.9 µmol/L [>1.3 mg/dL]) in 1 adefovir recipient from weeks 24 to 52; the maximum elevation was 53 µmol/L (0.6 mg/dL) above the baseline value at week 32. Creatinine levels returned to within the normal range after week 52, concomitant with switching to telbivudine in the follow-up study. Grades 3 and 4 neutropenia were observed in 1 patient in group A and 1 patient in group C. Retesting 6 days later indicated that both conditions had resolved without dose reduction or treatment interruption.

In this study, telbivudine showed greater antiviral effects than adefovir dipivoxil after 24 weeks of treatment in HBeAg-positive patients with chronic hepatitis B. At week 24, 3 times as many telbivudine-treated patients (group A) as adefovir-treated patients (pooled groups B and C) were PCR-negative. Conversely, HBV DNA levels remained above 5 log₁₀ copies/mL in 42% of group B versus 5% of group A, suggesting more consistent antiviral suppression with telbivudine. Adefovir recipients randomly assigned to switch to telbivudine at week 24 (group C) showed a rapid, approximately 1-log incremental reduction of serum HBV DNA levels within 2 months. At 1 year, the difference in reduction of serum HBV DNA level, as assessed by residual HBV DNA levels, differed significantly when patients in groups A and C were compared with those in group B. Other measures at 1 year also consistently favored telbivudine over adefovir, although differences were not statistically significant. The incidence of viral breakthrough was similar for telbivudine and adefovir at 1 year, although resistance mutations were detected only with continuous telbivudine therapy (group A). Adverse event frequencies were similar in the 3 treatment groups; however, the occurrence of myopathy in a telbivudine recipient suggests that any persistent, unexplained muscle-related symptoms should be evaluated promptly.

Direct comparative studies are the most effective way to define the role of available agents for hepatitis B in clinical practice. A search of MEDLINE through March 2007 and hepatology conference abstracts from 2006 and 2007 identified several such studies. Phase III trials for both telbivudine and entecavir demonstrated greater antiviral efficacy at 1 year with less resistance (no resistance with entecavir) compared with lamivudine (9,22 - 23,29 - 30). A recent trial demonstrated greater antiviral effects with entecavir than with adefovir after 1 year (31).

Although all available agents have demonstrated greater antiviral efficacy than either placebo or lamivudine in phase III studies, a substantial proportion of patients receiving these agents did not achieve optimal viral suppression (7 - 9,21,29,32). In a placebo-controlled, phase III study, the serum HBV DNA level was reduced by less than 2.2 log₁₀ copies/mL after 48 weeks in 25% of adefovir recipients (33). However, few studies have explored the effects of switching therapies as the basis for a rational strategy to improve efficacy outcomes. A small trial in HBeAg-negative patients with suboptimal response to adefovir showed enhanced antiviral effects after switching to a combination of tenofovir and emtricitabine (34). In lamivudine-resistant patients, several studies have shown that viral suppression can be restored by adding or switching to adefovir or tenofovir, nucleotide analogues with complementary resistance profiles (35 - 37). Related studies suggest that in lamivudine-resistant patients, tenofovir or a higher dosage of adefovir (20 mg/d) may provide greater viral suppression than standard adefovir treatment, with potential benefits for patients with suboptimal responses to adefovir salvage therapy (38 - 39). However, adefovir dosages greater than 10 mg/d have been associated with a higher incidence of nephrotoxicity (8,40).

Limitations of this study include an open-label design and reduced statistical power after the primary analysis at week 24, due to reduced group sizes associated with division of the study population into 3 comparator groups. In addition, the study was not of sufficient duration to evaluate long-term clinical outcomes. Because the patients were all HBeAg-positive and nucleoside- and nucleotide-naive and were primarily Asian, the extent to which conclusions may be generalized to other patient populations with hepatitis B is uncertain. The open-ended approach that was used to elicit adverse event information may have resulted in variable or incomplete reporting of events. In analysis of variance models, observations of nondetectable HBV DNA were handled by single-value imputation (150 copies/mL); consequently, any variability of actual values for these data may not be accurately reflected in variability estimates, confidence limits, and P values.

In this trial, we randomly assigned some adefovir recipients to switch to telbivudine at 24 weeks. In clinical practice, however, treatment modification is generally considered only for patients with suboptimal antiviral responses (41). Adefovir recipients with suboptimal antiviral responses at week 24 achieved a substantial incremental decrease in viral load after switching to telbivudine, suggesting that treatment modification can be considered as early as week 24. Recent studies (42 - 43) suggest that adding adefovir to ongoing lamivudine therapy in patients with resistance may reduce subsequent adefovir resistance, compared with switching to adefovir monotherapy, and may improve viral suppression in patients with high viral load. The same advantages may also apply to patients without a resistant virus, although this has not been prospectively evaluated (44). The M204I and M204V mutations both confer resistance to lamivudine, whereas M204I is the primary basis of telbivudine resistance (28 - 29). These mutations reduce susceptibility to other l-nucleoside analogues, but not to the acyclic phosphonates adefovir and tenofovir. These findings suggest that there may be benefits of combining telbivudine with these agents, analogous to those reported for lamivudine and adefovir (28). Studies are needed to refine the definition of suboptimal early response as a basis for therapeutic decisions, to confirm that the benefits of treatment modification persist in the long term, and to compare the effects of adding telbivudine versus switching to telbivudine monotherapy in patients with a suboptimal response to adefovir.

Antiviral resistance is a serious concern with longer-term nucleoside and nucleotide therapy for hepatitis B and a relevant consideration in choice of therapy (45). The cumulative incidence of lamivudine resistance approaches 70% after 4 years; adefovir resistance was observed in 29% of HBeAg-negative patients after 5 years, and in 16 of 38 HBeAg-positive patients with virologic failure after 110 to 279 weeks (40,46). On the basis of an analysis of 120 of the 663 nucleoside-naive patients from phase III trials (47), cumulative resistance to entecavir was 0.8% at 4 years. Preliminary data suggest that the cumulative incidence of telbivudine resistance in HBeAg-positive patients approaches 20% as therapy is extended to 2 years (23). In this study, the association of HBV DNA levels at 24 weeks with efficacy outcomes and viral breakthrough at 1 year for both telbivudine and adefovir are consistent with previous analyses of telbivudine, adefovir, and lamivudine (11 - 18). These findings suggest that greater early (24-week) HBV DNA reduction may be associated with reduced risk for subsequent telbivudine resistance (18,23). Such considerations highlight the importance of regular viral load monitoring to support early intervention in patients with suboptimal virologic responses, with the goal of avoiding resistance and maintaining viral suppression (12,14,17 - 18).

The multiple agents now available for treating chronic hepatitis B provide the basis for improved control of HBV replication. Data from this study support the concept that maximizing viral suppression early in the course of therapy is linked to improved efficacy responses and less resistance, suggesting that agents providing the greatest viral suppression may be preferable as initial therapy. Data from this and previous studies indicate that telbivudine may provide greater viral suppression than adefovir and lamivudine. With regular monitoring to ensure that responses are maintained, telbivudine may have an important role in treatment regimens for patients with chronic hepatitis B (22 - 23).

In addition to the authors, investigators in this trial included Dr. Darrell Crawford, Princess Alexandra Hospital, Woolloongabba, Queensland, Australia; Dr. Seng-Gee Lim, National University Hospital, Singapore; Dr. Anuchit Chutaputti, Phramongkutklao Hospital, Bangkok, Thailand; and Dr. Thierry Poynard, Groupe Hospitalier Pitié-Salpêtrière, Paris, France.