new england

The
journal of medicine
established in 1812 march 9, 2006 vol. 354 no. 10

A Comparison of Entecavir and Lamivudine

for HBeAg-Positive Chronic Hepatitis B
Ting-Tsung Chang, M.D., Robert G. Gish, M.D., Robert de Man, M.D., Adrian Gadano, M.D., José Sollano, M.D.,
You-Chen Chao, M.D., Anna S. Lok, M.D., Kwang-Hyub Han, M.D., Zachary Goodman, M.D., Ph.D., Jin Zhu, Ph.D.,
Anne Cross, Ph.D., Deborah DeHertogh, M.D., Richard Wilber, M.D., Richard Colonno, Ph.D.,
and David Apelian, M.D., Ph.D., for the BEHoLD AI463022 Study Group*

A BS T R AC T

Background
Entecavir is a potent and selective guanosine analogue with significant activity From the National Cheng Kung University
against hepatitis B virus (HBV). Medical College, Tainan, Taiwan (T.-T.C.);
California Pacific Medical Center, San
Francisco (R.G.G.); University Hospital
Methods Rotterdam, Rotterdam, the Netherlands
In this phase 3, double-blind trial, we randomly assigned 715 patients with hepa- (R.M.); Hospital Italiano, Buenos Aires
(A.G.); the University of Santo Tomas,
titis B e antigen (HBeAg)–positive chronic hepatitis B who had not previously re- Manila, the Philippines ( J.S.); Tri-Service
ceived a nucleoside analogue to receive either 0.5 mg of entecavir or 100 mg of la- General Hospital, Taipei, Taiwan (Y.-C.C.);
mivudine once daily for a minimum of 52 weeks. The primary efficacy end point the University of Michigan, Ann Arbor
(A.S.L.); Severance Hospital, Yonsei Uni-
was histologic improvement (a decrease by at least two points in the Knodell necro- versity College of Medicine, Seoul, Korea
inflammatory score, without worsening of fibrosis) at week 48. Secondary end (K.-H.H.); Armed Forces Institute of Pa-
points included a reduction in the serum HBV DNA level, HBeAg loss and serocon- thology, Washington, D.C. (Z.G.); Bristol-
Myers Squibb Pharmaceutical Research
version, and normalization of the alanine aminotransferase level. Institute, Wallingford, Conn. ( J.Z., A.C.,
R.W., R.C.); the University of Connecticut,
Results Farmington (D.D.); and GlobeImmune,
Louisville, Colo. (D.A.). Address reprint
Histologic improvement after 48 weeks occurred in 226 of 314 patients in the ente- requests to Dr. Chang at the Department
cavir group (72 percent) and 195 of 314 patients in the lamivudine group (62 per- of Internal Medicine, National Cheng
cent, P = 0.009). More patients in the entecavir group than in the lamivudine group Kung University Hospital, 138 Sheng-Li
Rd., Tainan, Taiwan 704, or at ttchang@
had undetectable serum HBV DNA levels according to a polymerase-chain-reaction mail.ncku.edu.tw.
assay (67 percent vs. 36 percent, P<0.001) and normalization of alanine amino-
transferase levels (68 percent vs. 60 percent, P = 0.02). The mean reduction in serum *Other members of the Benefits of En-
tecavir for Hepatitis B Liver Disease
HBV DNA from baseline to week 48 was greater with entecavir than with lamivu- (BEHoLD) AI463022 Study Group are
dine (6.9 vs. 5.4 log [on a base-10 scale] copies per milliliter, P<0.001). HBeAg se- listed in the Appendix.
roconversion occurred in 21 percent of entecavir-treated patients and 18 percent of
N Engl J Med 2006;354:1001-10.
those treated with lamivudine (P = 0.33). No viral resistance to entecavir was de- Copyright © 2006 Massachusetts Medical Society.
tected. Safety was similar in the two groups.

Conclusions
Among patients with HBeAg-positive chronic hepatitis B, the rates of histologic, vi-
rologic, and biochemical improvement are significantly higher with entecavir than
with lamivudine. The safety profile of the two agents is similar, and there is no evi-
dence of viral resistance to entecavir. (ClinicalTrials.gov number, NCT00035633.)

n engl j med 354;10 www.nejm.org march 9, 2006 1001

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

C
hronic hepatitis b affects more HBeAg-positive patients and 5.9 percent of HBeAg-
than 350 million people worldwide and negative patients.22,23
each year is responsible for more than Entecavir (Baraclude, Bristol-Myers Squibb) is
1 million deaths from cirrhosis and hepatocellular a potent and selective inhibitor of HBV DNA
carcinoma.1,2 Suppression of hepatitis B virus polymerase.24 Preclinical studies in woodchucks
(HBV) replication is a principal goal of long-term demonstrated the antiviral activity of entecavir and
hepatitis B therapy.3 Studies of long-term lamivu- its potential to reduce the incidence of hepato-
dine treatment in patients with hepatitis B e anti- cellular carcinoma and improve survival.25 In a
gen (HBeAg)–positive chronic hepatitis B have randomized phase 2 trial in HBeAg-positive and
found that maintenance of virologic suppression HBeAg-negative patients who had not previously
is associated with improved histologic findings received a nucleoside analogue, entecavir at a
in the liver.4,5 Liaw and coworkers have shown dose of 0.5 mg once daily resulted in a signifi-
that lamivudine treatment delays the progression cantly greater mean reduction in HBV DNA at 22
of disease among patients with chronic hepatitis B weeks than did lamivudine at a dose of 100 mg
and advanced liver disease, probably by suppress- once daily (4.72 vs. 3.36 log copies per milliliter,
ing viral replication and hepatic necroinflamma- P<0.001).26 Entecavir had a safety and adverse-
tion.6 In addition, two recent prospective studies reaction profile similar to that of lamivudine. The
from Taiwan found that among persons infected efficacy and tolerability of entecavir were also
with HBV, higher plasma HBV DNA levels are as- demonstrated in phase 2 and phase 3 trials of
sociated with an increased risk of hepatocellular patients whose disease was refractory to lamivu-
carcinoma, reinforcing the importance of control- dine.27,28 The current study was designed to com-
ling viral replication among patients with chronic pare the efficacy and safety of 48 weeks of thera-
hepatitis B.7,8 py with entecavir with 48 weeks of therapy with
Five medications are currently available for the lamivudine (Epivir-HBV, GlaxoSmithKline) in pa-
treatment of HBeAg-positive chronic HBV infec- tients with HBeAg-positive chronic hepatitis B who
tion: interferon alfa, lamivudine, pegylated inter- had not previously received a nucleoside analogue.
feron alfa-2a, adefovir dipivoxil, and entecavir.
Interferon alfa is effective in only 30 to 40 per- ME THODS
cent of patients and is associated with a high
incidence of adverse events.9,10 Lamivudine sup- Study Design
presses viral replication and results in HBeAg This study was a double-blind, double-dummy,
seroconversion in 16 to 17 percent of patients and randomized, controlled trial. Patients were recruit-
histologic improvement in 52 to 56 percent after ed from 137 centers worldwide, including Europe
one year of therapy.11,12 However, within four (41 centers), North America (40), Asia (26), Aus-
years, the rate of emergence of lamivudine-resis- tralia (12), and South America (18), and received
tant HBV mutants approaches 70 percent,13-15 and entecavir at a dose of 0.5 mg once daily or lami-
resistance is usually associated with a rebound in vudine at a dose of 100 mg once daily for a mini-
viral load and often with exacerbation of hepati- mum of 52 weeks. Treatment assignments were
tis.16-19 Pegylated interferon alfa-2a appears to allocated centrally on the basis of permuted block
have better efficacy than lamivudine in HBeAg- sizes of four that were assigned within each
positive patients 6 months after the completion center.
of a 48-week treatment course; however, the rate The study was conducted in accordance with
of adverse events associated with pegylated inter- the ethics principles of the Declaration of Helsinki
feron alfa-2a is higher than it is with lamivudine.20 and was consistent with Good Clinical Practice
One year of treatment with adefovir dipivoxil re- guidelines and applicable local regulatory require-
sults in a median reduction in the HBV DNA level ments. Written informed consent was obtained
of 3.52 log (on a base-10 scale) copies per millili- from all randomly assigned patients.
ter, histologic improvement in 53 percent of pa- The study was designed by the sponsor (Bristol-
tients, and HBeAg seroconversion in 12 percent Myers Squibb) in collaboration with expert hepa-
of patients.21 After three years of treatment, resis- tologists who comprised the Benefits of Enteca-
tance to adefovir develops in up to 3 percent of vir for Hepatitis B Liver Disease (BEHoLD) Study

1002 n engl j med 354;10 www.nejm.org march 9, 2006

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 entecavir for hb e a g -positive chronic hepatitis b

Group. The sponsor collected the data, monitored within 24 weeks before randomization; prior la-
the conduct of the study, performed the statistical mivudine therapy lasting more than 12 weeks; an
analyses, and coordinated the writing of the alpha fetoprotein level greater than 100 ng per
manuscript with all authors. Data were unblind- milliliter; a history of ascites requiring diuretics
ed for statistical analysis after the database was or paracentesis; and previous treatment with en-
locked. The authors had access to the complete tecavir.
study reports, were actively involved in data analy-
sis and interpretation, and approved the final Efficacy End Points
manuscript. The academic authors vouch for the The primary efficacy end point was the propor-
veracity and completeness of the data and the data tion of patients with histologic improvement, de-
analyses. fined as improvement by at least two points in the
Clinical-management decisions were defined Knodell necroinflammatory score with no wors-
by protocol and made at week 52 on the basis of ening in the Knodell fibrosis score at week 48,
the results of the Quantiplex (Chiron) branched- relative to baseline.29 Biopsies were scheduled to be
chain DNA and HBeAg assays on serum samples performed at baseline (unless one had been per-
obtained at the visit at week 48. Patients who formed within one year before randomization) and
had a response (defined by an HBV DNA level at week 48. Liver-biopsy specimens were evaluated
below 0.7 megaequivalents [MEq] per milliliter by a central, independent histopathologist who
and HBeAg loss) or a nonresponse (defined by was unaware of patients’ treatment assignment,
an HBV DNA level of 0.7 MEq per milliliter or biopsy sequence, and clinical outcome. For analy-
greater) discontinued study treatment at week 52. ses of the primary efficacy end point, patients
Patients who had only a virologic response (de- who could be evaluated had adequate baseline bi-
fined by an HBV DNA level below 0.7 MEq per opsy specimens with a Knodell necroinflammatory
milliliter and no HBeAg loss) were offered con- score of at least 2.
tinued study therapy for up to 96 weeks. Secondary efficacy end points at week 48 in-
cluded the reduction in HBV DNA level from base-
Study Population line and the proportion of patients with undetect-
Eligible patients were 16 years of age or older and able HBV DNA, as measured by the Roche COBAS
had HBeAg-positive chronic hepatitis B and com- Amplicor PCR assay (version 2.0; lower limit of
pensated liver function (a total serum bilirubin quantification, 300 copies per milliliter); the
level of 2.5 mg per deciliter [42.8 μmol per liter] decrease in the Ishak fibrosis score; HBeAg loss;
or less; a prothrombin time not more than three HBeAg seroconversion (HBeAg loss and the ap-
seconds longer than normal or an international pearance of HBe antibody); and normalization of
normalized ratio not greater than 1.5; a serum serum alanine aminotransferase. Normalization
albumin level of at least 3.0 g per deciliter; and of alanine aminotransferase was defined by the
no history of variceal bleeding or hepatic enceph- protocol as an alanine aminotransferase value
alopathy). Eligible patients also had detectable less than 1.25 times the upper limit of normal;
hepatitis B surface antigen (HBsAg) for at least the data were subsequently reanalyzed according
24 weeks before screening, evidence of chronic to a more stringent definition of normalization
hepatitis on a baseline liver-biopsy specimen ob- (an alanine aminotransferase value no greater
tained within 52 weeks before randomization, than the upper limit of normal).
evidence of HBV DNA by any commercial assay at Patients who had a response at week 48 and
least 4 weeks before screening, an HBV DNA level discontinued treatment were followed for 24 weeks
of at least 3 MEq per milliliter by the branched- after treatment. In this way we investigated wheth-
chain DNA assay at screening, and a serum ala- er the virologic and serologic benefits of antiviral
nine aminotransferase level 1.3 to 10.0 times the therapy were sustained.
upper limit of normal at screening.
Exclusion criteria included coinfection with Safety Analysis
hepatitis C, hepatitis D, or the human immuno- The safety analysis included data from all 709
deficiency virus; the presence of other forms of treated patients during treatment, including the
liver disease; use of interferon alfa, thymosin α, second year of treatment for patients who contin-
or antiviral agents with activity against hepatitis B ued for more than 52 weeks. The primary safety

n engl j med 354;10 www.nejm.org march 9, 2006 1003

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

end point was the proportion of patients who dis- Patients with missing or inadequate biopsy
continued the study medication because of clini- specimens obtained at week 48 were considered
cal or laboratory-determined adverse events. Other not to have had a histologic response. In propor-
safety evaluations included analyses of adverse tion analyses of HBV DNA values, HBV serologic
events, serious adverse events, and deaths. Hepati- data, and alanine aminotransferase levels, treated
tis flares during treatment were defined as eleva- patients with a missing value for an end point
tions in the alanine aminotransferase level to more were considered not to have had a response for
than twice the baseline level and to more than 10 that end point. To compare the means of con-
times the upper limit of normal. Post-treatment tinuous variables, we used t-tests based on linear
flares were defined as elevations in alanine ami- regression models, adjusted for baseline measure-
notransferase to more than twice the reference ments. There were no interim analyses of efficacy.
level and to more than 10 times the upper limit of All reported P values are two-sided and were not
normal, where the reference level was the lesser adjusted for multiple testing.
of the baseline value and end-of-treatment value.
R E SULT S
Resistance Analysis
An extensive resistance analysis was undertaken Study Population
to identify emerging HBV polymerase substitutions Of 1056 patients who were enrolled and screened,
that may be associated with reduced susceptibil- 715 were randomly assigned (357 to the entecavir
ity to entecavir. All 339 available paired samples group and 358 to the lamivudine group), and 709
from patients in the entecavir group were sub- (354 in the entecavir group and 355 in the lami-
mitted for genotypic analysis. HBV DNA was ex- vudine group) received, in a blinded fashion, at
tracted and amplified by polymerase chain reac- least one dose of study drug. Three hundred four-
tion (PCR), and amino acids 1 through 344 of the teen patients in each treatment group had adequate
reverse transcriptase were sequenced as described baseline liver-biopsy specimens with a Knodell
elsewhere.30 Substitutions that emerged during necroinflammatory score of 2 or greater. The two
therapy were inserted into recombinant clones and treatment groups were well balanced at baseline
analyzed in cell-culture phenotypic assays for any (Table 1).
effect on susceptibility to entecavir.30 Within the Among those who started receiving the study
lamivudine group, genotypic and phenotypic analy- drug, 340 patients assigned to the entecavir group
ses were performed only on samples from patients (95 percent) and 321 patients assigned to the la-
meeting the criterion for virologic rebound (de- mivudine group (90 percent) completed 52 weeks
fined as a confirmed increase in HBV DNA by at of treatment. The difference in these proportions
least 1 log copy per milliliter from the nadir val- was attributable mainly to the greater number of
ue, according to a PCR assay, during the admin- discontinuations due to adverse events in the la-
istration of study medication). mivudine group (nine, vs. one in the entecavir
group) and the greater number of losses to fol-
Statistical Analysis low-up in the lamivudine group (eight vs. three,
A two-stage evaluation of efficacy was planned.31 respectively).
First, noninferiority to lamivudine was tested, and
if noninferiority was established, a second test for Histologic and Biochemical End Points
superiority was conducted. The planned sample More patients in the entecavir group than in the
size, 315 per group, had 90 percent power to dem- lamivudine group underwent a biopsy at week 48.
onstrate noninferiority with respect to the primary Twenty-two and 45 patients, respectively (P = 0.004),
efficacy end point, assuming response rates of with adequate baseline biopsy specimens had spec-
60 percent for lamivudine and 64 percent for en- imens obtained at week 48 that could not be
tecavir, a 25 percent rate of missing biopsy spec- evaluated (Table 2). The primary analysis met the
imens obtained at week 48, and a −10 percent initial test for noninferiority, and we proceeded
boundary for the 95 percent lower confidence to test for superiority. Entecavir treatment result-
limit for the difference in proportions. The study ed in a significantly higher rate of histologic im-
had a single primary end point (histologic im- provement than lamivudine treatment at week
provement). 48, with 72 percent and 62 percent of the patients,

1004 n engl j med 354;10 www.nejm.org march 9, 2006

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 entecavir for hb e a g -positive chronic hepatitis b

Table 1. Demographic and Baseline Characteristics of the Patients.*

Characteristic Entecavir (N = 354) Lamivudine (N = 355) P Value

Age — yr 35±13 35±13 1.00
Male sex — no. (%) 274 (77) 261 (74) 0.26
Race or ethnic group — no. (%)†
Asian 204 (58) 202 (57) 0.91
White 140 (40) 141 (40)
Black 8 (2) 8 (2)
Other 2 (<1) 4 (1)
Region — no. (%)
Asia or Australia 172 (49) 167 (47) 0.76
Europe 84 (24) 88 (25)
North America 47 (13) 55 (15)
South America 51 (14) 45 (13)
Knodell necroinflammatory score‡ 7.8±2.98 7.7±2.99 0.67
Ishak fibrosis score‡ 2.3±1.27 2.3±1.29 1.00
Score ≥3 (bridging fibrosis) — % 34 32 0.68
Score ≥5 (cirrhosis) — % 8 8 1.00
Mean HBV DNA level
By branched-chain DNA assay — MEq/ml 2.56±1.05 2.61±1.03 0.52
By PCR assay — log copies/ml 9.62±2.01 9.69±1.99 0.64
HBeAg-positive — no. (%) 348 (98) 351 (99) 0.55
HBeAg antibody–negative — no. (%) 342 (97) 346 (97) 0.52
Viral genotype — no. (%)
A 94 (27) 100 (28) 0.24
B 68 (19) 77 (22)
C 111 (31) 90 (25)
D 37 (10) 49 (14)
F 20 (6) 12 (3)
Other, indeterminate, or missing 24 (7) 27 (8)
Alanine aminotransferase — IU/liter 140.5±114.3 146.3±132.3 0.53
Prior anti-HBV therapy — no. (%)
Interferon 46 (13) 46 (13) 1.00
Lamivudine 10 (3) 10 (3) 1.00

* Plus–minus values are means ±SD. The Knodell necroinflammatory score can range from 0 to 18, with higher scores
indicating more severe hepatitis. The Ishak fibrosis score can range from 0 to 6, with higher scores indicating more ex-
tensive fibrosis and scores of 5 or higher indicating cirrhosis. HBV denotes hepatitis B virus, PCR polymerase chain re-
action, and HBeAg hepatitis B e antigen.
† Race or ethnic group was determined by the investigator.
‡ Adequate baseline biopsy specimens were available for 329 patients in the entecavir group and 330 patients in the lamivu-
dine group.

respectively, reaching this primary end point (dif- percent of the patients, respectively (P = 0.41). In
ference estimate, 9.9; 95 percent confidence in- 8 percent of the entecavir-treated patients and 10
terval, 2.6 to 17.2; P = 0.009) (Table 2). Treatment percent of the lamivudine-treated patients, the
with entecavir and lamivudine resulted in im- Ishak fibrosis score worsened. The alanine ami-
proved Ishak fibrosis scores in 39 percent and 35 notransferase level was normalized in a higher

n engl j med 354;10 www.nejm.org march 9, 2006 1005

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. Histologic Improvement at Week 48 (Primary Study End Point).*

Entecavir Lamivudine Difference Estimate P

End Point (N = 354) (N = 355) (95% CI)† Value
Adequate baseline biopsy specimen with Knodell 314 314
necroinflammatory score ≥2 — no.
Histologic improvement — no. (%)‡ 226 (72) 195 (62) 9.9 (2.6–17.2) 0.009
No improvement — no. (%) 66 (21) 74 (24)
Mean Knodell necroinflammatory score§
Baseline 8.2 8.1
Week 48 4.4 4.6

* Higher scores on the Knodell necroinflammatory scale indicate more severe hepatitis. CI denotes confidence interval.
† The difference estimate was calculated for the entecavir group as compared with the lamivudine group.
‡ Histologic improvement was defined as an improvement by at least two points in the necroinflammatory component
of the Knodell score, with no worsening in the fibrosis component of the score.
§ There were 292 treated patients in the entecavir group and 269 treated patients in the lamivudine group with adequate
pairs of biopsy specimens.

proportion of entecavir-treated patients than la- cent) had a protocol-defined response (HBV DNA
mivudine-treated patients at week 48 (68 percent level of less than 0.7 MEq per milliliter and HBeAg
vs. 60 percent, P = 0.02) (Table 3). loss). Two hundred forty-seven patients in the en-
tecavir group (70 percent) and 165 in the lamivu-
Virologic and Serologic End Points dine group (46 percent) had a virologic response
HBV DNA levels in the entecavir group fell through- (HBV DNA level of less than 0.7 MEq per milli-
out treatment; HBV DNA was undetectable by PCR liter, without HBeAg loss). There was a nonre-
assay in 67 percent of the patients in this group sponse (HBV DNA level, ≥0.7 MEq per milliliter)
at week 48 (Table 3 and Fig. 1A). In contrast, viral in 19 patients in the entecavir group (5 percent)
loads in the lamivudine group remained distributed and 94 in the lamivudine group (26 percent).
over a wide range of values, and by week 48, HBV Among the patients with a protocol-defined
DNA was undetectable by PCR assay in 36 per- response, 82 percent of those in the entecavir
cent of the patients in that group (P<0.001 for the group (61 of 74) and 73 percent of those in the
comparison with the entecavir group). The mean lamivudine group (49 of 67) had a sustained re-
reduction from baseline in the serum HBV DNA sponse 24 weeks after the discontinuation of treat-
level by PCR assay at week 48 was also greater in ment. Of the patients with a response at week 48
the entecavir group (6.9 log copies per milliliter) who discontinued therapy, 41 percent (25 of 61)
than in the lamivudine group (5.4 log copies per of those in the entecavir group and 41 percent
milliliter, P<0.001) (Table 3 and Fig. 1B). (20 of 49) of those in the lamivudine group had
HBeAg loss and seroconversion occurred in undetectable HBV DNA by PCR, and 84 percent
22 percent and 21 percent, respectively, of the (51 of 61) and 82 percent (40 of 49), respectively,
patients in the entecavir group (Table 3). These had normalization of alanine aminotransferase.
rates were similar to those in the lamivudine
group (20 percent and 18 percent, respectively; Resistance
P = 0.45 for the comparison of HBeAg loss be- There was no evidence of emerging resistant vari-
tween the groups and P = 0.33 for the compari- ants to entecavir by week 48 among the 339 eval-
son of HBeAg seroconversion between the groups). uated patients assigned to receive entecavir. During
HBsAg loss occurred in six patients in the en- entecavir therapy, substitutions were noted at sev-
tecavir group (2 percent) and four in the lami- eral residues within the viral reverse transcriptase,
vudine group (1 percent). but none appeared in more than three patients,
and none were associated with reduced suscepti-
Responses at Week 48 and After Treatment bility to entecavir when examined in phenotypic
At week 48, 74 patients in the entecavir group (21 assays. Six patients in the entecavir group (2 per-
percent) and 67 in the lamivudine group (19 per- cent) and 63 patients in the lamivudine group

1006 n engl j med 354;10 www.nejm.org march 9, 2006

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 entecavir for hb e a g -positive chronic hepatitis b

Table 3. Virologic, Biochemical, and Serologic End Points at Week 48.*

Entecavir Lamivudine Difference Estimate

End Point (N = 354) (N = 355) (95% CI)† P Value
Virologic
HBV DNA <300 copies/ml by PCR assay — no. (%) 236 (67) 129 (36) 30.3 (23.3 to 37.3) <0.001
HBV DNA <0.7 MEq/ml by branched-chain DNA assay 322 (91) 232 (65) 25.6 (19.8 to 31.4) <0.001
— no. (%)
Mean change in HBV DNA from baseline by PCR assay −6.9±2.0 −5.4±2.6 −1.52 (−1.78 to −1.27) <0.001
— log copies/ml‡
Biochemical
ALT normalization (≤1× ULN) — no. (%) 242 (68) 213 (60) 8.4 (1.3 to 15.4) 0.02
Serologic
Loss of HBeAg — no. (%) 78 (22) 70 (20) 2.3 (−3.7 to 8.3) 0.45
HBeAg seroconversion — no. (%) 74 (21) 64 (18) 2.9 (−2.9 to 8.7) 0.33
HBsAg loss — no. (%) 6 (2) 4 (1) 0.6 (−1.2 to 2.3) 0.52

* Plus–minus values are means ±SD. CI denotes confidence interval, HBV hepatitis B virus, PCR polymerase chain reaction, ALT alanine
aminotransferase, ULN upper limit of normal, HBeAg hepatitis B e antigen, and HBsAg hepatitis B surface antigen.
† The difference estimate was calculated for the entecavir group as compared with the lamivudine group.
‡ There were 340 patients in the entecavir group and 324 patients in the lamivudine group with paired baseline and week-48 HBV DNA
measurements. Samples in which HBV DNA was undetectable were assigned a value of 299 copies per milliliter.

(18 percent) had virologic rebound during the first continued treatment because of an increase in
year of drug administration. Genotypic analysis alanine aminotransferase.
of isolates obtained at week 48 from the six ente- Elevations in alanine aminotransferase were
cavir-treated patients revealed no emerging sub- observed less frequently in the entecavir group
stitutions when baseline samples were compared than in the lamivudine group (Table 4). Alanine
with those obtained at week 48. Samples obtained aminotransferase flares during treatment were
at week 48 retained full phenotypic susceptibility observed in 12 entecavir-treated patients (3 per-
to entecavir. Genotypic analysis of isolates obtained cent) and 23 lamivudine-treated patients (6 per-
at week 48 from the patients in the lamivudine cent). In the entecavir group, all the alanine
group with virologic rebound revealed that 45 of aminotransferase flares during treatment were
63 (71 percent) had mutations in the YMDD mo- associated with a reduction in HBV DNA by at
tif of the HBV polymerase gene. least 2 log copies per milliliter, and all but one
were self-limiting with continued treatment. None
Safety and Adverse Events of the entecavir-treated patients had hepatic de-
Mean exposure to study therapy at the time the compensation.
database was locked was 75 weeks for entecavir In the lamivudine group, approximately half
and 65 weeks for lamivudine. The frequency of the alanine aminotransferase flares during treat-
adverse events during treatment was similar in the ment (12 of 23) were associated with increasing
two groups (Table 4). The most frequent adverse HBV DNA levels. The majority of the flares per-
events were headache, upper respiratory tract in- sisted until the time of treatment discontinua-
fection, nasopharyngitis, cough, pyrexia, upper tion. One of the lamivudine-treated patients had
abdominal pain, fatigue, and diarrhea, most of hepatic decompensation associated with a flare
which were of mild-to-moderate severity. The fre- and died during follow-up. At the time the data-
quencies of serious adverse events were also sim- base was locked, alanine aminotransferase flares
ilar in the two treatment groups. There were fewer had occurred in 2 of 134 patients (1 percent) in the
discontinuations due to adverse events in the en- entecavir group and 9 of 129 patients (7 percent)
tecavir group (one) than in the lamivudine group in the lamivudine group who underwent post-
(nine); four of the nine patients in the lamivudine treatment follow-up.
group and the patient in the entecavir group dis- Two deaths occurred during the on-treatment

n engl j med 354;10 www.nejm.org march 9, 2006 1007

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

period, both in the lamivudine group. Neither was

A
Entecavir Lamivudine
judged to be related to study therapy. One patient
died from sudden dyspnea, and one death was of
352 331 329 341 354 322 311 324 undetermined cause.
≥11
Dis cus sion
10

9 This study was a large, active-control trial of two

HBV DNA (log copies/ml)

8 nucleoside therapies in patients with HBeAg-posi-

7 tive chronic hepatitis. When the trial began, la-
6
mivudine was the only nucleoside approved to
treat chronic hepatitis B. Not only did it serve as
5
an active comparator; it also enabled all the pa-
4
tients to receive potential benefit from antiviral
3 therapy.
2 Lamivudine treatment has been associated with
Undetect- histologic improvement and with reductions in
able
both Child–Pugh score and the development of
0 24 36 48 0 24 36 48 hepatocellular carcinoma.4,5,32 This reduction in
Weeks of Treatment
disease progression probably arises from the ef-
fective suppression of HBV replication, the un-
B derlying cause of liver necroinflammation and
10
fibrosis. However, the frequent development of
Mean HBV DNA lamivudine resistance can result in a rebound in
change at wk 48 viral load and a loss of histologic benefit.4
Mean HBV DNA (log copies/ml)

8 ¡5.4 log copies/ml

¡6.9 log copies/ml In the current study, entecavir was associated
with significantly higher rates of histologic, vi-
6 rologic, and biochemical improvement than was
lamivudine. The efficacy of entecavir appears to
Lamivudine (N= 355)
4 Entecavir (N= 354) result from its potent suppression of HBV repli-
cation. Entecavir suppressed HBV DNA by a mean
of nearly 7 log copies per milliliter, and 67 per-
2 300 copies/ml
cent of the patients in the entecavir group had
undetectable levels of HBV DNA by PCR assay
0 within 48 weeks after the start of treatment.
0 4 8 12 16 20 24 28 32 36 40 44 48
Histologic examination of the liver remains
Weeks of Treatment the definitive method of assessing disease pro-
gression, and arresting or reversing liver dam-
Figure 1. Distribution of Patients According to HBV DNA Level (Panel A) age is a principal goal of long-term hepatitis B
and Mean HBV DNA Levels through Week 48 (Panel B).
therapy. The improved histologic benefit of ente-
Panel A shows HBV DNA levels in entecavir-treated patients and lamivudine-
cavir as compared with lamivudine and its great-
treated patients. The numbers above the columns of circles are the numbers
of patients. The diameters of the circles are proportional to the percentag- er effect on viral suppression suggest that with
es of patients with specified DNA levels; the percentages represented by long-term treatment, entecavir might also reduce
the circles in each column total 100. The log HBV DNA levels indicated on the risk of end-stage liver disease and hepatocel-
the y axis reflect all observations within that exponent; an HBV DNA level lular carcinoma.
of less than 300 copies per milliliter was below the level of detection by
A direct consequence of the potent viral sup-
polymerase-chain-reaction (PCR) analysis. Panel B shows mean HBV DNA
levels by PCR through week 48 among the treated patients. The horizontal pression associated with entecavir is the absence
dashed line represents the lower limit of detection of the PCR assay. of emergence of drug resistance by week 48. In
vitro experiments and clinical studies have dem-

1008 n engl j med 354;10 www.nejm.org march 9, 2006

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 entecavir for hb e a g -positive chronic hepatitis b

Table 4. Summary of Cumulative Safety Data.*

Timing and Event Entecavir (N = 354) Lamivudine (N = 355) P Value

no. of patients (%)
During treatment
Any adverse event 306 (86) 297 (84) 0.34
Serious adverse event 27 (8) 30 (8) 0.78
Discontinuation due to adverse event 1 (<1) 9 (3) 0.02
ALT >2× baseline and >10× ULN† 12 (3) 23 (6) 0.08
ALT >2× baseline and >5× ULN‡ 37 (10) 59 (17) 0.02
Death 0 2 (<1) 0.50
Post-treatment follow-up§
ALT >2× reference value and >10× ULN¶ 2 (1) 9 (7) 0.03
ALT > 2× reference value and >5× ULN‡ 3 (2) 16 (12) 0.002

* ALT denotes alanine aminotransferase, and ULN upper limit of normal.

† According to the protocol, these findings constituted the definition of an alanine aminotransferase flare during treatment.
‡ The analysis was conducted post hoc.
§ There were 134 patients in the entecavir group and 129 in the lamivudine group who had entered post-treatment follow-
up at the time the database was locked.
¶ According to the protocol, these findings constituted the definition of a post-treatment alanine aminotransferase flare.
The reference level was the lesser of the baseline and end-of-treatment alanine aminotransferase values.

onstrated that resistance to entecavir develops which provided response rates that were signifi-
only, and infrequently, when HBV contains pre- cantly higher than those of an existing oral an-
existing lamivudine-associated resistance substi- tiviral agent in a large population of patients who
tutions (rtL180M and rtM204V) and when an ad- had not previously received a nucleoside analogue,
ditional substitution (rtT184G, rtS202I, or rtM250V) has demonstrated benefit as a primary therapy for
is selected.30 The potential for resistance after HBeAg-positive chronic hepatitis B.
longer exposure to entecavir treatment is currently Presented in part at the 55th annual meeting of the American
being monitored. Association for the Study of Liver Diseases, October 31, 2004,
and the 14th biennial meeting of the Asian Pacific Association
On the basis of its extensive use in both chron- for the Study of the Liver, December 14, 2004.
ic hepatitis B and HIV infection, lamivudine is Drs. Gadano, Gish, Han, and Lok report having received con-
established as a well-tolerated antiviral therapy. sulting fees from Bristol-Myers Squibb; Drs. Gish and Lok, con-
sulting fees and lecture fees from Roche and grant support from
The similar tolerability profiles of entecavir and Schering-Plough and Gilead; Drs. Gish and Goodman, consult-
lamivudine in this study are evidence that ente- ing fees from Schering-Plough; Dr. Gish, consulting fees and
cavir is well tolerated by patients with chronic lecture fees from InterMune, lecture fees from Schering-Plough,
and grant support from the National Cancer Institute; Drs.
hepatitis B; however, continued surveillance is de Man, Goodman, and Lok, consulting fees from Gilead; Dr.
necessary to confirm its long-term safety. Fewer Lok, consulting fees from GlaxoSmithKline, Innogenetics, XTL,
patients in the entecavir group than in the lami- Idun, Idenix, Nabi, PowderMed, and Anadys and grant support
from Valeant and the National Institutes of Health; Drs. Chang,
vudine group had alanine aminotransferase flares Han, and Gish, lecture fees from Bristol-Myers Squibb; Dr. de Man,
during treatment, and the flares that did occur lecture fees from GlaxoSmithKline and Gilead and grant sup-
during entecavir therapy were temporally associ- port from Biotest; Dr. Sollano, lecture fees from Novartis and
Hi-Esai; Drs. Gish, Goodman, Han, and Lok, grant support
ated with reductions in HBV DNA. Flares after from Bristol-Myers Squibb; Drs. de Man and Lok, grant support
treatment, during six months of follow-up, were from GlaxoSmithKline; Drs. de Man, Gish, and Lok, grant sup-
also uncommon among the patients assigned to port from Roche; and Drs. Goodman and Lok, grant support
from Idenix. Dr. Apelian and Dr. DeHertogh were members of
receive entecavir. the Bristol-Myers Squibb Pharmaceutical Research Institute
The advent of more potent antiviral agents for (Wallingford, Conn.) at the time this research was conducted.
the treatment of chronic hepatitis B offers the No other potential conflict of interest relevant to this article was
reported.
potential to control HBV replication and to arrest We are indebted to Bruce Kreter, Pharm.D. (Bristol-Myers
or halt the progression of liver disease. Entecavir, Squibb), for his contribution to the manuscript.

n engl j med 354;10 www.nejm.org march 9, 2006 1009

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.
 entecavir for hb e a g -positive chronic hepatitis b

appendix
Bristol-Myers Squibb scientists were D. Tenney, R. Rose, and S. Levine.
In addition to the authors, the BEHoLD AI463022 Study Group included the following investigators: North America — F. Anderson, T.
Boyer, R. Brown, D. Chua, D. Dieterich, L. Cisneros Garza, S. Gordon, S. Han, R. Herring, S. Joshi, R. Koff, D. LaBrecque, S. Lee, W.
Lee, P. Martin, S. Mehta, A. Min, G. Minuk, K. O’Riordan, P. Pockros, K.R. Reddy, R. Reindollar, R. Rubin, V. Rustgi, J. Sattler, E.
Schiff, A.O. Shakil, M. Sherman, M. Shiffman, C. Smith, N. Tsai, B. Tung, L. Tyrrell, and G. Wu. South America — A. Barone, F. Bessone,
F. Carrilho, H. Cheinquer, H. Coelho, H. Fainboim, J. Ferrandiz, M. Ferraz, R. Filho, R. Focaccia, F. Goncales, L. Lyra, H. Sette, M.
Silva, R. Terg, and E. Zumaeta. Asia, Pacific Islands, and Australia — S. Abdurachman, N. Akbar, M. Arguillas, S.W. Cho, G. Cooksley, P.
Desmond, J. George, S. Huang, E.K. Ooi, I. Kronborg, C.L. Lai, S.T. Lai, J. Lao-Tan, A. Lee, S.D. Lee, H.H. Lin, C.C. Lim, G.H. Lo, Y.W.
Luk, G. Marinos, J. Menon, I. Merican, J. Olynyk, S. Roberts, M. Rosmawati, J. Sasadeusz, W. Sievert, S. Strasser, C.K. Tan, S.S. Wu,
S.K. Yoon, and F. Zano. Europe and Middle East — Y. Baruch, M. Bassendine, A. Boron-Kaczmarska, M.R. Brunetto, G. Carosi, P. Chalu-
pa, A. Gladysz, B. Gocman, Z. Gonciarz, V. Gorbakov, W. Halota, M. Heim, Y. Horsmans, V. Isakov, J. Juszczyk, W. Kryczka, G.A.
Kullak-Ublick, J. Kuydowicz, Y. Lurie, T. Mach, F. Mazzotta, P. Mills, R. Modrzewska, D. Mutimer, F. Nevens, A. Nimer, M. Oltman,
C. Pedersen, V. Pokrovsky, V. Rafalsky, M. Rizzetto, D. Shouval, M. Sikora, G. Storozhakov, H. Thomas, R. Tur-Kaspa, H. Van Vlier-
berghe, A. Yakovlev, and K. Zhdanov.

References
1. World Health Organization. Hepatitis A one-year trial of lamivudine for chronic tive chronic hepatitis B. N Engl J Med
B fact sheet WHO/204. Geneva: World hepatitis B. N Engl J Med 1998;339:61-8. 2005;352:2673-81.
Health Organization, October 2000. (Ac- 13. Lau DT, Khokhar MF, Doo E, et al. 24. Innaimo SF, Seifer M, Bisacchi GS,
cessed February 10, 2006, at http://www. Long-term therapy of chronic hepatitis B Standring DN, Zahler R, Colonno RJ.
who.int/mediacentre/factsheets/fs204/en.) with lamivudine. Hepatology 2000;32:828- Identification of BMS-200475 as a potent
2. Lee WM. Hepatitis B virus infection. 34. and selective inhibitor of hepatitis B virus.
N Engl J Med 1997;337:1733-45. 14. Chang TT, Lai CL, Chien RN, et al. Antimicrob Agents Chemother 1997;41:
3. Liaw Y-F, Leung N, Guan R, Lau GK, Four years of lamivudine treatment in 1444-8.
Merican I. Asian-Pacific consensus state- Chinese patients with chronic hepatitis B. 25. Colonno RJ, Genovesi EV, Medina I, et
ment on the management of chronic hep- J Gastroenterol Hepatol 2004;19:1276-82. al. Long-term entecavir treatment results
atitis B: an update. J Gastroenterol Hepa- 15. Liaw YF, Leung NW, Chang TT, et al. in sustained antiviral efficacy and pro-
tol 2003;18:239-45. Effects of extended lamivudine therapy in longed life span in the woodchuck model
4. Dienstag JL, Goldin RD, Heathcote EJ, Asian patients with chronic hepatitis B. of chronic hepatitis infection. J Infect Dis
et al. Histological outcome during long- Gastroenterology 2000;119:172-80. 2001;184:1236-45.
term lamivudine therapy. Gastroenterol- 16. Liaw YF. Impact of YMDD mutations 26. Lai CL, Rosmawati M, Lao J, et al. En-
ogy 2003;124:105-17. during lamivudine therapy in patients with tecavir is superior to lamivudine in reduc-
5. Leung NW, Lai CL, Chang TT, et al. chronic hepatitis B. Antivir Chem Che- ing hepatitis B virus DNA in patients with
Extended lamivudine treatment in pa- mother 2001;12:Suppl 1:67-71. chronic hepatitis B infection. Gastroen-
tients with chronic hepatitis B enhances 17. Mutimer D. Hepatitis B virus infec- terology 2002;123:1831-8.
hepatitis B e antigen seroconversion rates: tion: resistance to antiviral agents. J Clin 27. Chang TT, Gish RG, Hadziyannis SJ,
results after 3 years of therapy. Hepatolo- Virol 2001;21:239-42. et al. A dose-ranging study of the efficacy
gy 2001;33:1527-32. 18. Liaw YF, Chien RN, Yeh CT, Tsai SL, and tolerability of entecavir in lamivudine-
6. Liaw YF, Sung JJY, Chow WC, et al. Chu CM. Acute exacerbation and hepatitis refractory chronic hepatitis B patients.
Lamivudine for patients with chronic hep- B virus clearance after emergence of YMDD Gastroenterology 2005;129:1198-209.
atitis B and advanced liver disease. N Engl motif mutation during lamivudine thera- 28. Sherman M, Yurdaydin C, Sollano J, et
J Med 2004;351:1521-31. py. Hepatology 1999;30:567-72. al. Entecavir is superior to continued la-
7. Yu MW, Yeh SH, Chen PJ, et al. Hepa- 19. Ayres A, Bartholomeusz A, Lau G, mivudine for the treatment of lamivudine-
titis B virus genotype and DNA level and Lam KC, Lee JY, Locarnini S. Lamivudine refractory HBeAg(+) chronic hepatitis B:
hepatocellular carcinoma: a prospective and famciclovir resistant hepatitis B virus results in Phase III study ETV-026. Hepa-
study in men. J Natl Cancer Inst 2005;97: associated with fatal hepatic failure. J Clin tology 2004;40:Suppl 1:664A. abstract.
265-72. Virol 2003;27:111-6. 29. Knodell RG, Ishak KG, Black WC, et
8. Chen CJ, Yang HI, Su J, et al. Risk of 20. Lau GK, Piratvisuth T, Luo KX, et al. al. Formulation and application of a nu-
hepatocellular carcinoma across a bio- Peginterferon alfa-2a, lamivudine, and the merical scoring system for assessing his-
logical gradient of serum hepatitis B virus combination for HBeAg-positive chronic tological activity in asymptomatic chronic
DNA level. JAMA 2006;295:65-73. hepatitis B. N Engl J Med 2005;352:2682- active hepatitis. Hepatology 1981;1:431-5.
9. Krogsgaard K, Bindslev N, Christen- 95. 30. Tenney DJ, Levine SM, Rose RE, et al.
sen E, et al. The treatment effect of alpha 21. Marcellin P, Chang TT, Lim SG, et al. Clinical emergence of entecavir-resistant
interferon in chronic hepatitis B is inde- Adefovir dipivoxil for the treatment of hepatitis B virus requires additional sub-
pendent of pre-treatment variables: results hepatitis B e antigen–positive chronic hep- stitutions in virus already resistant to la-
based on individual patient data from 10 atitis B. N Engl J Med 2003;348:808-16. mivudine. Antimicrob Agents Chemother
clinical controlled trials. J Hepatol 1994; 22. Marcellin P, Chang TT, Lim SG, et al. 2004;48:3498-507.
21:646-55. Long-term efficacy and safety of adefovir 31. Dunnett CW, Gent M. An alternative to
10. Intron A (interferon alfa-2b recombi- dipivoxil (ADV) in HBeAg+ chronic hep- the use of two-sided tests in clinical trials.
nant). Kenilworth, N.J.: Schering Corpo- atitis B patients: increasing serologic, Stat Med 1996;15:1729-38.
ration, 2002 (package insert). virologic, and biochemical response over 32. Goodman Z, Dhillon AP, Wu PC, et al.
11. Dienstag JL, Schiff ER, Wright TL, et time. Hepatology 2004;40:Suppl 1:665A. Lamivudine treatment reduces progres-
al. Lamivudine as initial treatment for abstract. sion to cirrhosis in patients with chronic
chronic hepatitis B in the United States. 23. Hadziyannis SJ, Tassopoulos NC, hepatitis B. J Hepatol 1999;30:Suppl 1:59.
N Engl J Med 1999;341:1256-63. Heathcote EJ, et al. Long-term therapy abstract.
12. Lai CL, Chien RN, Leung NW, et al. with adefovir dipivoxil for HBeAg-nega- Copyright © 2006 Massachusetts Medical Society.

1010 n engl j med 354;10 www.nejm.org march 9, 2006

The New England Journal of Medicine

Downloaded from nejm.org on November 11, 2016. For personal use only. No other uses without permission.
Copyright © 2006 Massachusetts Medical Society. All rights reserved.