The
new england journal
of
medicine
n engl j med 354;10 www.nejm.org march 9, 2006
1011
original article
Entecavir versus Lamivudine for Patients
with HBeAg-Negative Chronic Hepatitis B
Ching-Lung Lai, M.D., Daniel Shouval, M.D., Anna S. Lok, M.D.,
Ting-Tsung Chang, M.D., Hugo Cheinquer, M.D., Zachary Goodman, M.D., Ph.D.,
Deborah DeHertogh, M.D., Richard Wilber, M.D., Richard C. Zink, Ph.D.,
Anne Cross, Ph.D., Richard Colonno, Ph.D., and Lori Fernandes, M.D.,
for the BEHoLD AI463027 Study Group*
From the Queen Mary Hospital, Hong
Kong, China (C.-L.L.); Hadassah-Hebrew
University Hospital, Jerusalem, Israel
(D.S.); University of Michigan, Ann Arbor
(A.S.L.); National Cheng Kung University
Medical College, Tainan, Taiwan (T.-T.C.);
Universidade Federal do Rio Grande do
Sul, Porto Alegre, Brazil (H.C.); Armed
Forces Institute of Pathology, Washing-
ton, D.C. (Z.G.); University of Connecti-
cut, Farmington (D.D.); and Bristol-Myers
Squibb Pharmaceutical Research Institute,
Wallingford, Conn. (R.W., R.C.Z., A.C.,
R.C., L.F.). Address reprint requests to
Dr. Lai at the University Department of
Medicine, Queen Mary Hospital, Hong
Kong, China, or at hrmelcl@hkucc.hku.hk.
*Other members of the Benefits of En-
tecavir for Hepatitis B Liver Disease
(BEHoLD) AI463027 Study Group are
listed in the Appendix.
N Engl J Med 2006;354:1011-20.
Copyright © 2006 Massachusetts Medical Society.
ABSTRACT
BACKGROUND
Entecavir is a potent and selective antiviral agent that has demonstrated efficacy in
phase 2 studies in patients with hepatitis B e antigen (HBeAg)–negative chronic hepa-
titis B.
METHODS
In this phase 3, double-blind t rial, we randomly assigned 648 patients with HBeAg-
negative chronic hepatitis B who had not previously been treated with a nucleoside
analogue to receive 0.5 mg of entecavir or 100 mg of lamivudine once daily for a
minimum of 52 weeks. The primary efficacy end point was histologic improvement
(a decrease by at least two points in the Knodell necroinflammatory score, without
worsening of fibrosis).
RESULTS
Histologic improvement after 48 weeks of treatment occurred in 208 of 296 patients
in the entecavir group who had adequate baseline liver-biopsy specimens that could
be evaluated (70 percent), as compared with 174 of 287 such patients in the lami-
vudine group (61 percent, P = 0.01). More patients in the entecavir group than in the
lamivudine group had undetectable serum hepatitis B virus (HBV) DNA levels ac-
cording to a polymerase-chain-reaction assay (90 percent vs. 72 percent, P<0.001)
and normalization of alanine aminotransferase levels (78 percent vs. 71 percent,
P = 0.045). The mean reduction in serum HBV DNA levels from baseline to week 48
was greater with entecavir than with lamivudine (5.0 vs. 4.5 log [on a base-10 scale]
copies per milliliter, P<0.001). There was no evidence of resistance to entecavir.
Safety and adverse-event profiles were similar in the two groups.
CONCLUSIONS
Among patients with HBeAg-negative chronic hepatitis B who had not previously
been treated with a nucleoside analogue, the rates of histologic improvement, viro-
logic response, and normalization of alanine aminotransferase levels were signifi-
cantly higher at 48 weeks with entecavir than with lamivudine. The safety profile
of the two agents was similar, and there was no evidence of viral resistance to
entecavir. (ClinicalTrials.gov number, NCT00035789.)
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at UNIVERSITY OF HONG KONG on August 1, 2007 .
The
new england journal
of
medicine
n engl j med 354;10 www.nejm.org march 9, 2006
1012
H
epatitis B virus (HBV) infection is a
serious global health concern. Approxi-
mately 350 million people worldwide are
chronically infected, and 500,000 to 1.2 million
deaths per year are attributed to HBV-associated
complications.
1,2
A common variant of HBV in-
fection occurs in patients who test negative for
hepatitis B e antigen (HBeAg) and positive for an-
tibodies against HBeAg (anti-HBe) and in whom
serum HBV DNA and alanine aminotransferase
levels remain persistently or intermittently elevat-
ed.
3-5
The median worldwide prevalence of HBeAg-
negative disease in hepatitis B surface antigen
(HBsAg)–positive carriers was reported to be 33
percent in 2002 and is increasing.
5
HBeAg-negative HBV develops spontaneously
through mutations in the precore or core promoter
regions of the viral genome such that HBeAg is
no longer expressed or is down-regulated, and it
has been suggested that this gives the mutant an
immunologic advantage over wild-type HBV.
3,4
However, HBeAg-negative chronic hepatitis B is
a heterogeneous condition, and wild-type HBV
may also be responsible for disease activity in
some patients.
6,7
The clinical profile of HBeAg-
negative chronic hepatitis B differs from that of
HBeAg-positive disease in that patients are typi-
cally older,
4
serum HBV DNA levels are usually
lower,
8,9
and liver disease tends to fluctuate.
10-12
Patients with HBeAg-negative chronic hepatitis B
have more advanced liver disease, and the likeli-
hood of spontaneous remission is very low.
4,11
The end point of treatment for HBeAg-negative
chronic hepatitis B is unknown. HBeAg loss or
seroconversion cannot be used to assess response,
and treatment usually focuses on suppression of
HBV DNA and normalization of alanine amino-
transferase levels.
13
Effective suppression of HBV
DNA without development of resistance among
HBeAg-negative patients has been associated with
improved histologic findings in the liver and long-
term clinical benefit.
14-16
Treatment guidelines
support the use of interferon, lamivudine, or ad-
efovir for HBeAg-negative chronic hepatitis B in
patients with viremia and elevated alanine ami-
notransferase levels.
17-20
Entecavir (Baraclude, Bristol-Myers Squibb) is
a potent and highly selective inhibitor of HBV
DNA polymerase.
21
In a double-blind, randomized
phase 3 study of HBeAg-positive patients who
had not previously received a nucleoside analogue,
entecavir resulted in significantly higher rates of
histologic, virologic, and biochemical improve-
ment than did lamivudine, with a similar safety
profile.
22
The current study was designed to
compare the efficacy and safety of entecavir with
that of lamivudine (Epivir-HBV, GlaxoSmithKline)
after 48 weeks of treatment in patients with
HBeAg-negative chronic hepatitis B who had
not previously received a nucleoside analogue.
Methods
Study Design
The design of this randomized, double-blind, con-
trolled study was nearly identical to that of the
st ud y of HBe Ag-posit ive pat ient s repor t ed by Chang
et al. elsewhere in this issue of the Journal.
22
Pa-
tients were recruited from 146 centers worldwide,
including Europe and the Middle East (68 centers),
Asia (25), Australia (11), North America (30), and
South America (12), and received 0.5 mg of ente-
cavir or 100 mg of lamivudine once a day for a
minimum of 52 weeks. Clinical management de-
cisions were made at week 52, on the basis of the
results of branched-chain DNA assays (Quanti-
plex, Chiron) and alanine aminotransferase assays
on serum samples obtained at week 48.
Patients who had a response (defined by an
HBV DNA level below 0.7 megaequivalents [MEq]
per milliliter according to branched-chain DNA
assay and an alanine aminotransferase level be-
low 1.25 times the upper limit of the normal
range) or a nonresponse (defined by an HBV DNA
level of 0.7 MEq per milliliter or greater) were to
discontinue study treatment.
Patients who had a response at week 48 and
discontinued treatment were followed for 24 weeks
after the cessation of treatment. In this way, we
investigated whether the virologic and biochem-
ical benefits of antiviral therapy were sustained
after the discontinuation of treatment. Patients
who had only a virologic response (defined by an
HBV DNA level below 0.7 MEq per milliliter and
an alanine aminotransferase level of at least 1.25
times the upper limit of normal) were offered con-
tinued therapy for up to 96 weeks.
The study was conducted in accordance with
the ethics principles of the Declaration of Helsinki
and was consistent with Good Clinical Practice
guidelines and applicable local regulatory require-
ments. Written informed consent was obtained
from all randomly assigned patients.
The study was designed by the sponsor (Bristol-
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at UNIVERSITY OF HONG KONG on August 1, 2007 .
Entecavir for HB
e
A
g
-Negative Chronic Hepatitis B
n engl j med 354;10 www.nejm.org march 9, 2006
1013
Myers Squibb) in collaboration with expert hepa-
tologists who comprised the Benefits of Enteca-
vir for Hepatitis B Liver Disease (BEHoLD) Study
Group. The sponsor collected the data, monitored
the conduct of the study, performed the statisti-
cal analyses, and coordinated the writing of the
manuscript with all authors. Data were unblind-
ed for statistical analysis after the database was
locked. The authors had access to the complete
study reports, were actively involved in data analy-
sis and interpretation, and approved the final
manuscript. The academic authors vouch for the
veracity and completeness of the data and the data
analyses.
Study Population
Eligible patients were 16 years of age or older and
had HBeAg-negative chronic hepatitis B and com-
pensated liver function (a total serum bilirubin
level of 2.5 mg per deciliter [42.8 μmol per liter]
or less, a prothrombin time not more than three
seconds longer than normal or an international
normalized ratio not greater than 1.5, a serum
albumin level of at least 3.0 g per deciliter, and
no history of variceal bleeding or hepatic enceph-
alopathy). Eligible patients also had detectable
HBsAg for at least 24 weeks before screening,
evidence of chronic hepatitis on a baseline liver-
biopsy specimen obtained within 52 weeks be-
fore randomizat ion, evidence of HBV DNA by any
commercial assay at least 2 weeks before screen-
ing, undetectable HBeAg, detectable anti-HBe, a
serum HBV DNA level of at least 0.7 MEq per mil-
liliter according to the branched-chain DNA assay
at screening, and a serum alanine aminotrans-
ferase level 1.3 to 10.0 times the upper limit of
normal at screening.
Exclusion criteria included coinfection with
hepatitis C, hepatitis D, or the human immuno-
deficiency virus; the presence of other forms of
liver disease; use of interferon alfa, thymosin α,
or antiviral agents with activity against hepatitis
B within 24 weeks before randomization; previ-
ous lamivudine therapy lasting more than 12
weeks; an alpha fetoprotein level greater than
100 ng per milliliter; a history of ascites requir-
ing diuretics or paracentesis; and previous treat-
ment with entecavir.
Efficacy End Points
The primary and secondary efficacy end points
were the same as those in the study of HBeAg-
positive patients,
22
except that HBeAg loss and
seroconversion were not secondary end points in
the present study. The primary efficacy end point
was the proportion of patients with histologic
improvement, defined as improvement by at least
two points in the Knodell necroinflammatory
score, with no worsening in the Knodell fibrosis
score at week 48, relative to baseline.
23
Secondary efficacy end points at week 48 in-
cluded the reduction in the HBV DNA level from
baseline and the proportion of patients with un-
detectable HBV DNA, as measured by the Roche
COBAS Amplicor polymerase-chain-reaction (PCR)
assay (version 2.0; lower limit of quantification,
300 copies per milliliter); the decrease in the Ishak
fibrosis score; and normalization of serum alanine
aminotransferase (less than 1.0 times the upper
limit of normal).
Safety Analysis
The safety analysis included data from all 638
treated patients during treatment, including the
second year of treatment for patients who contin-
ued for more than 52 weeks. The primary safety
end point was the proportion of patients who dis-
continued the study medication because of clinical
or laboratory-determined adverse events. Other
safety evaluations included analyses of adverse
events, serious adverse events, and deaths. Flares
of hepatitis during treatment were defined as el-
evations in the alanine aminotransferase level to
more than twice the baseline level and to more
than 10 times the upper limit of normal. Post-
treatment flares were defined as elevations in
alanine aminotransferase to more than twice the
reference level and to more than 10 times the up-
per limit of normal, where the reference level was
the lesser of the baseline value and the end-of-
treatment value.
Resistance Analysis
Two sampling schemes were used to identify
emerging HBV polymerase substitutions that may
be associated with reduced susceptibility to ente-
cavir. Paired samples from 211 randomly selected
patients in the entecavir group were genotypically
analyzed. HBV DNA was extracted and amplified
with the use of PCR, and amino acids 1 through
344 of the reverse transcriptase were sequenced as
described elsewhere.
24
Substitutions that emerged
during therapy were inserted into recombinant
clones and analyzed in cell-culture phenotypic
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at UNIVERSITY OF HONG KONG on August 1, 2007 .
The
new england journal
of
medicine
n engl j med 354;10 www.nejm.org march 9, 2006
1014
assays for susceptibility to entecavir.
24
The second
sampling scheme involved genotypic and pheno-
typic analyses of all paired samples from all pa-
tients meeting the criterion for virologic rebound
(defined as a confirmed increase in the HBV DNA
level by at least 1 log [on a base-10 scale] copy per
milliliter from the nadir value, according to PCR
assay, while the patient was receiving the study
medication).
Statistic al Analysis
A two-stage evaluation was planned. First, non-
inferiority to lamivudine was tested, and if non-
inferiority was established, a second test for su-
periority was conducted. The planned sample size,
315 per group, had 90 percent power to demon-
strate noninferiority with respect to the primary
eff icacy end point, assuming response rates of 60
percent for lamivudine and 64 percent for enteca-
vir, a 25 percent rate of missing biopsy specimens
obtained at week 48, and a −10 percent boundary
for the 95 percent lower confidence limit for the
difference in proportions. The study had a single
primary end point (histologic improvement).
Patients with missing or inadequate biopsy
specimens obtained at week 48 were considered
not to have had a histologic response. In propor-
tion analyses of HBV DNA levels and alanine
aminotransferase levels, treated patients with a
missing value for an end point were considered
not to have had a response for that end point. To
compare the means of continuous variables, we
used t-tests based on linear regression models,
adjusted for baseline measurements. There were
no interim analyses of efficacy. All reported P val-
ues are two-sided and were not adjusted for mul-
tiple testing.
Results
Study Population
Of 1468 patients who were enrolled and screened,
648 were randomly assigned to treatment (331 to
the entecavir group and 317 to the lamivudine
group), and 638 (325 in the entecavir group and
313 in the lamivudine group) received, in a blinded
fashion, at least one dose of study drug. Of these
638 patients, 583 had adequate baseline liver-biopsy
specimens with a Knodell necroinflammatory
score of 2 or greater; 31 of 296 patients receiving
entecavir (10 percent) and 37 of 287 patients re-
ceiving lamivudine (13 percent) who had baseline
liver-biopsy specimens had missing specimens at
week 48. The two treatment groups were well
balanced at baseline (
Table 1
). Of the 820 pa-
tients not randomly assigned to treatment, 774
did not meet one or more of the inclusion criteria.
Three hundred eleven patients assigned to
receive entecavir (96 percent) and 296 patients
assigned to receive lamivudine (95 percent) com-
pleted 52 weeks of treatment. No patient discon-
tinued treatment because of treatment failure or
lack of efficacy during the 52-week, blinded treat-
ment period.
Histologic and Biochemical Response
After criteria for noninferiority were met, we con-
ducted tests for superiority. Histologic improve-
ment occurred in significantly more patients in
the entecavir group than in the lamivudine group
at week 48 (70 percent vs. 61 percent, P = 0.01)
(
Table 2
). There was no histologic improvement
in 26 percent of patients in the lamivudine group
and 19 percent of patients in the entecavir group.
Treatment with entecavir or lamivudine resulted
in improved Ishak fibrosis scores in 36 percent and
38 percent of the patients, respectively (P = 0.65).
The alanine aminotransferase level was normal-
ized in significantly more patients treated with
entecavir than those treated with lamivudine at
week 48 (78 percent vs. 71 percent, P = 0.045)
(
Table 3
).
Virologic and Serologic Response
The mean reduction from baseline in serum HBV
DNA levels according to PCR assay at week 48
was significantly greater in the entecavir group
than in the lamivudine group (5.0 log copies vs.
4.5 log copies per milliliter, P<0.001) (
Table 3
and
Fig. 1B). HBV DNA levels in the entecavir group
fell continuously during treatment, and 90 percent
of the patients had undetectable levels at week 48
according to PCR assay. In contrast, HBV DNA
levels in the lamivudine group remained distrib-
uted over a wide range of values, and 72 percent
of the patients had undetectable levels at week 48
according to PCR assay (P<0.001) (
Table 3
and
Fig. 1A). HBsAg loss had occurred in one patient
in each treatment group at week 48.
Responses at Week 48 and after Treatment
At week 48, 275 patients in the entecavir group
(85 percent) and 245 patients in the lamivudine
group (78 percent) had a response (P = 0.04), as
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at UNIVERSITY OF HONG KONG on August 1, 2007 .
Entecavir for HB
e
A
g
-Negative Chronic Hepatitis B
n engl j med 354;10 www.nejm.org march 9, 2006
1015
defined by the protocol (HBV DNA level, <0.7 MEq
per milliliter, according to branched-chain DNA
assay; alanine aminotransferase level, <1.25 times
the upper limit of normal). Thirty-four patients
in the entecavir group (10 percent) and 34 in the
lamivudine group (11 percent) had a virologic re-
sponse (HBV DNA level, <0.7 MEq per milliliter;
alanine aminotransferase level, ≥1.25 times the
upper limit of normal). Three patients in the en-
tecavir group (<1 percent) and 18 in the lamivu-
dine group (6 percent) had a nonresponse (HBV
DNA level, ≥0.7 MEq per milliliter).
Table 1. Demographic and Baseline Characteristics of the Patients.*
Characteristic Entecavir (N = 325) Lamivudine (N = 313) P Value
Age — yr 44±11 44±11 1.00
Male sex — no. (%) 248 (76) 236 (75) 0.85
Race or ethnic group — no. (%)†
White 193 (59) 176 (56) 0.74
Asian 122 (38) 129 (41)
Black 8 (2) 7 (2)
Other 2 (<1) 1 (<1)
Region — no. (%)
Europe and the Middle East 156 (48) 148 (47) 1.00
Asia and Australia 106 (33) 104 (33)
South America 35 (11) 34 (11)
North America 28 (9) 27 (9)
Knodell necroinflammatory score‡ 8.0±2.7 7.7±2.8 0.18
Ishak fibrosis score‡ 2.4±1.2 2.5±1.3 0.31
≥3 (bridging fibrosis) — % 43 41 0.68
≥5 (cirrhosis) — % 5 10 0.06
Mean HBV DNA level
By branched-chain DNA assay — MEq/ml 1.2±1.0 1.2±1.0 1.00
By PCR assay — log copies/ml 7.6±1.8 7.6±1.7 1.00
HBeAg-negative — no. (%) 322 (99) 309 (99) 0.72
Anti-HBe-positive — no. (%) 323 (99) 312 (100) 1.00
Viral genotype — no. (%)
A 33 (10) 33 (11) 0.45
B 46 (14) 60 (19)
C 57 (18) 51 (16)
D 157 (48) 135 (43)
Other, indeterminate, or missing 32 (10) 34 (11)
Alanine aminotransferase — IU/liter 141±114.7 143±119.4 0.83
Prior anti-HBV therapy — no. (%) 49 (15) 45 (14)
Interferon 42 (13) 39 (12) 0.91
Lamivudine 9 (3) 12 (4) 0.51
* Plus–minus values are means ±SD. Percentages may not sum to 100 because of rounding. The Knodell inflammatory
score can range from 0 to 18, with higher scores indicating more severe chronic hepatitis. The Ishak fibrosis score is
based on a scale of 0 to 6, where 0 indicates no fibrosis and 5 or higher indicates cirrhosis. HBV denotes hepatitis B vi-
rus, PCR polymerase chain reaction, and HBeAg hepatitis B e antigen.
† Race or ethnic group was determined by the investigator.
‡ Adequate baseline biopsy specimens were available for 303 patients in the entecavir group and 293 patients in the la-
mivudine group.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
Downloaded from www.nejm.org at UNIVERSITY OF HONG KONG on August 1, 2007 .
