Chinese Herbal
Medicine and Interferon in the Treatment of Chronic Hepatitis
B: A Meta-Analysis of Randomized, Controlled Trials
October
2002, Vol 92, No. 10 | American Journal of Public Health
1619-1628
© 2002
Michael McCulloch, LAc, MPH,
Michael Broffman, LAc, Jin Gao, MD, PhD and John M. Colford,
Jr, MD, PhD
Michael McCulloch and
Michael Broffman are with the Pine Street Clinic, San Anselmo,
Calif. Michael McCulloch and Jin Gao are with the Institute of
Biophysics, China Academy of Sciences, Beijing, China. Michael
McCulloch and John M. Colford Jr, are with the School of
Public Health, University of California at Berkeley.
Correspondence: Requests for
reprints should be sent to John M. Colford Jr, MD, PhD, UC
Berkeley School of Public Health, 140 Warren Hall MC 7360,
Berkeley, CA 94720 (e-mail: jcolford@socrates.berkeley.edu).
ABSTRACT
Objectives. This meta-analysis was conducted to examine
the effectiveness of Chinese herbal medicine
(either alone or with interferon alfa) in treating
chronic hepatitis B.
Methods. We searched the TCMLARS, AMED, CISCOM,
EMBASE, MEDLINE, and Cochrane Collaboration
databases and then hand-searched the articles’
bibliographies.
Results. Chinese herbal medicine significantly
increased seroreversion of HBsAg and was equivalent
to interferon alfa in seroreversion of HBeAg and
hepatitis B virus (HBV) DNA; Chinese herbal medicine combined
with interferon alfa significantly increased seroreversion
of HBsAg, HBeAg, and HBV DNA. The Chinese herbal
medicine active component bufotoxin combined with
interferon alfa significantly increased HBeAg and
HBV DNA seroreversion. The Chinese herbal medicine
active component kurorinone was equivalent to interferon alfa
in seroreversion of HBeAg and HBV DNA.
Conclusions. Although the quality of existing
studies was poor, these data suggest that further
trials of Chinese Herbal Medicine and interferon in
chronic hepatitis B infection are justified.
INTRODUCTION
Traditional Chinese medicine is an established segment of the
health care delivery system in China. In planning for
allocation of health care resources, an important
question for China’s health care authorities is
whether traditional Chinese medicine functions best
as a stand-alone therapy or in close integration with
allopathic medical care. However, little formal assessment
of its clinical effectiveness has been conducted. In
this study, we sought to evaluate the clinical
evidence for its effectiveness in the treatment of
chronic hepatitis B and to examine the quality of
the published data.
As one of the core techniques used within traditional
Chinese medicine, Chinese herbal medicine is
commonly used in China in the treatment of
hepatitis. In this meta-analysis of randomized, controlled
trials we examined the effectiveness of Chinese herbal medicine
in the treatment of chronic hepatitis B when used as a
stand-alone therapy and when used in combination with
interferon alfa. The control group in each case was
patients treated with interferon alfa alone.
Infection with hepatitis B virus is a significant public
health concern. Worldwide, an estimated 2 billion
people are infected with the hepatitis B virus (HBV).1
A total of 350 million people have the chronic form
of hepatitis B infection, 75% of whom live in Asia.1
Chronic infection increases the risk for primary liver
cancer.
Endemic hepatitis B infection in Asia’s large
population contributes to primary liver cancer’s position
as the fourth leading cause of cancer death worldwide (after
lung, stomach, and colorectal cancers).
Successful treatment of hepatitis B infection has long been
defined as loss of detection of hepatitis B surface
antigen (HBsAg). Meta-analysis has shown that HBsAg
clearance occurs in only 6% of patients with
chronic hepatitis B who are treated with interferon
alfa.5
Observational studies have shown that such
clearance occurs spontaneously in 4% to 29% of people with
chronic infection.
However, patients can develop HBsAgnegative chronic
infection, a clinical course with a more serious prognosis
than that of patients who are HBsAg positive.1
Therefore, some authors support the use of
hepatitis B e antigen (HBeAg) and HBV DNA as
markers of active viral replication and infectivity.
HBeAg clearance occurs in 18% to 40% of patients with
chronic hepatitis B who are treated with interferon
alfa and spontaneously in 15% to 60% of people with
chronic infection.
Furthermore, patients can also develop HBeAg-negative
chronic infection, which, as with HBsAgnegative
patients, signals a poor prognosis.
Herbal medicine is in common use in many parts of the
world. A 1997 survey estimated that 34% of the
American public use alternative medicine; among the
survey respondents, 12% reported the use of herbal
medicine within the prior 12 months.In China, Chinese herbal medicine is used as a
treatment adjunct or alternative to interferon alfa
and accounts for 30% to 50% of total medicine
consumption,
with low cost and low toxicity. Interferon alfa, by
contrast has a very high cost and significant toxicities.
English-language journals have published few randomized
trials of Chinese herbal medicine for the treatment
of hepatitis.
A far larger body of literature exists in
Chineselanguage journals. For centuries, textbooks
have discussed treatment strategies handed down in
the oral and literary tradition of Chinese herbal therapy.
Over the past 50 years, modern Chinese-language medical journals
have more formally assessed the effectiveness of these treatment
strategies. From case reports came observational studies and,
over the past decade, randomized, controlled trials. Although
these medical journals report only studies from the past
50 years, these data represent a distillation of
the accumulated historical experience of the body
of traditional Chinese medicine. The field has
grown substantially, from 1 published randomized, controlled
trial of Chinese herbal medicine for the treatment of
chronic hepatitis B in 1991 to 221 in 1999. Until recently,
however, time-consuming searching by hand was the only
means of accessing Chinese-language data sources.
Online availability of the Chinese-language TCMLARS
database now allows rapid searching of journal abstracts to
quickly locate clinical trials data published in
China after 1984. TCMLARS contains more than 330
000 references and abstracts to literature on traditional
Chinese medicine, drawn from more than 600 Chinese
biomedical journals and 100 specialty journals.
Searching is straightforward, and scanned articles
can be ordered via e-mail. Approximately 10% of the
abstract database has been translated into English.
Using TCMLARS and searches of Western medical literature,
we examined 2 hypotheses: (1) that treatment with
Chinese herbal medicine could serve as an
alternative therapy when interferon alfa is not
available or acceptable, and (2) that Chinese herbal medicine
used in combination with interferon alfa could enhance the
effectiveness of interferon alfa. We were interested in assessing
the effectiveness of Chinese herbal medicine when used
either as a stand-alone therapy or in combination with interferon
alfa and also in examining the quality of the published data.
METHODS
Study Selection
We searched for articles in TCMLARS (1984–2000), MEDLINE
(1966–2000), the Cochrane Database of Systematic
Reviews (Cochrane Collaboration, 1992–2000),
CISCOM (Centralised Information Service for
Complementary Medicine), EMBASE (Excerpta Medica,
1974–2000), and AMED (Allied & Complementary Medicine
Resources, 1985–2000), with articles in all languages included
for consideration. We used the following keywords and medical
subject headings: hepatitis B; hepatitis B, chronic; drugs,
Chinese herbal; medicine, Chinese traditional; medicine, oriental
traditional; interferon; and interferons. These resources
were supplemented by the hand-searching of articles’
bibliographies, nonindexed medical and professional
journals, and the Chinese-language and
English-language libraries and files of the authors. Two authors
(M. M. and M. B.) translated the Chinese-language articles.
We searched for additional data, both published and
unpublished, through communications with a senior
investigator and collaborator at the China Academy
of Sciences (J. G.). To define a standardized control
regimen, we included only studies in which the control group
used interferon alfa at a dosage of at least 1 million units
administered 3 times weekly; we excluded studies in which
the control group used very low doses of interferon alfa,
different comparison treatments such as gamma
interferon, other drugs, or other herbal
treatments.
In the first stage of our systematic review, we identified
studies describing the use of Chinese herbal
medicine and interferon alfa in the title or
abstract (n = 587). We retained for further review
studies in which interferon alfa was administered to the
control group (n = 49). For the meta-analysis, we retained
only those 27 studies (1) that were randomized,
controlled trials of Chinese herbal medicine alone
(vs interferon alfa) or Chinese herbal medicine
combined with interferon alfa (vs interferon alfa)
for the treatment of hepatitis B (Table 1 )
and (2) that provided data on the number of
responders and nonresponders for any of the 3
endpoints: HBsAg (n = 18 studies), HBeAg (n = 27),
and HBV DNA (n = 20). We defined Chinese herbal medicine as
the 311 botanical and animal-product medicines that are
commonly used in clinical practice by practitioners
of traditional Chinese medicine and enumerated in a
current herbal medical textbook used at the
Shanghai University of Traditional Chinese Medicine.
When we found multiple reports of the same patient data,
we selected for review only the most recently
published data (n = 1).
TABLE 1 —Chinese Herbal Medicine (CHM) for Chronic
Hepatitis B: Study Diagnoses and Herbal Medicines Used
|
Author
|
Quality
Scorea
|
Diagnosis
|
CHM
or CHM + IFN-
Group, Average Age ± Range
|
IFN-
Group, Average Age ± Range
|
Herbal
Treatment
|
|
|
|
Cai24
(1997)
|
0
|
CAH,
CPH, CAH + LC
|
31.5
± 10.3
|
35.4
± 9.9
|
Kurorinone
|
|
Chen19
(2000)
|
1
|
CHB
|
NS
|
NS
|
Kurorinone
|
|
Dai25
(1998)
|
1
|
CAH,
CPH
|
NS
|
NS
|
Artemesia
capillaris,
Astragalus membranaceus, Peonia rubra,
Polygonum multiflorum, Poria cocos, Pseudostellaria
heterophylla
|
|
Fu26
(1997)
|
0
|
CHB
|
NS
|
NS
|
Agrimonia,
Astragalus membranaceus, Atractylodes
alba, Carthamus tinctorum, Ligusticum
wallichium, Codonopsis pilosula, Gardenia
jasminoidis, Gentiana scabra, Glycyrrhiza
uralensis, Imperata cylindrica, Peonia
alba, Peonia rubra, Prunus persica,
Poria cocos, Pueraria lobata, Rheum
officinale, Salvia multiorrhiza, Sparganium
longifolium, Schisandra chinensis, Curcuma
longazedoaria, Zingiberis officinalis, Zizyphus
jujuba
|
|
Hao27
(1996)
|
1
|
CHB,
CPH
|
NS
|
NS
|
Ganpi
jiaonang combination (ingredients not specified)
|
|
Huang28
(1999)
|
1
|
CHB
|
32.5
± 6.7
|
30.8
± 5.6
|
Phyllanthus,
pseudoginseng
|
|
Huang29
(2000)
|
1
|
CHB
|
35.7
± 11.3
|
37.2
± 11.7
|
Artemesia
capillaris,
Atractylodes alba, bupleurum, Glycyrrhiza
uralensis, Hypericum japonicum, Magnolia
officinalis, Polygonum cuspidatum, Polyporus
umbellatus, Poria cocos, Rheum
officinale, Salvia multiorrhiza
|
|
Jing30
(2000)
|
1
|
CHB
|
28.7
(NS)
|
27.6
(NS)
|
Cuscuta
chinensis,
Ganoderma lucidum, Juglans regia, Sophora
subprostata
|
|
Li31
(1998)
|
1
|
CHB
|
30.8
± 5.7
|
32.8
± 6.9
|
Phyllanthus,
pseudoginseng
|
|
Li32
(1999)
|
0
|
CHB
|
NS
|
NS
|
Phyllanthus,
Polygonum cuspidatum, Schisandra
chinensis
|
|
Li33
(1997)
|
1
|
CAH
|
NS
|
NS
|
Agrimonia
pilosa,
Isatis indigotica, Scutellaria barbata,
Scutellaria baicalensis, Nidus vespae,
Oldenlandia diffusa, Polygonum
cuspidatum, Smilax glabra
|
|
Li34
(2000)
|
1
|
CHB
|
33.7
± 7.8
|
31.0
± 7.8
|
Alpinia,
Atractylodes alba, bupleurum, Coix
lachryma-jobii, Curcuma longa, Dryopteris
crassirhizo, Eclipta prostrate, Oldenlandia
diffusa, Isatis indigotica, Loranthus
parasiticus, Magnolia officinalis, Patrinia
villosa, Pinellia ternata, baijiangcao,
Scutellaria baicalensis
|
|
Liu35
(1999)
|
1
|
CHB
|
CHM
group: 32.6 ± 14.6; CHM + IFN-
group: 34.6 ± 17.8
|
35.7
± 20.5
|
Agrimonia
pilosae,
Bruca javanica, litchi, Dryopteris
crassirhizo, Punica granatum, Prunus
mume, Siegesbeckia orientalis, Stemonia
japonica
|
|
Lu36
(1992)
|
1
|
CAH,
CPH
|
NS
|
NS
|
Achyranthis
bidentata,
aloe, Amyda sinensis, Artemesia
capillaris, Astragalus membranaceus, Atractylodes
alba, Citrus medica, Curcuma longa,
Eclipta prostrate, Gallus gallus
domesticus, Gardenia jasminoidis, Gentiana
macrophylla, Imperata cylindrica, Isatis
indigotica, Lithospermum arnebia, Loranthus
parasiticus, mouton, Oldenlandia diffusa,
Peonia rubra, Polygonum cuspidatum, Salvia
multiorrhiza
|
|
Qian37
(1999)
|
1
|
CAH
|
37.4
(NS)
|
36.4
(NS)
|
Carthamus
tinctorum,
Ligusticum wallichium, Lithospermum
arnebia, Polygonum cuspidatum,
pseudoginseng, Salvia multiorrhiza, Scutellaria
baicalensis
|
|
Shen38
(2000)
|
1
|
CHB
|
31.2
(NS)
|
32.3
(NS)
|
Bufotoxin
|
|
Song39
(1994)
|
0
|
CHB
|
15.6
(NS)
|
14.8
(NS)
|
Aconite
carmichaeli,
Agastache rugosa, Amomum cardamom, Astragalus
membranaceus, Atractylodes alba, Citrus
reticulata, Epimedium, Glycyrrhiza
uralensis, Panax ginseng, Poria
cocos, Rehmannia glutinosa
|
|
Wang40
(1997)
|
1
|
CAH,
CPH
|
NS
|
NS
|
Astragalus
membranaceus, Cassia tora, chouteng, dibo, guicao, huangpi, longye, Salvia
multiorrhiza
|
|
Wang41
(2000)
|
1
|
CHB
|
38.5
(NS)
|
36.4
(NS)
|
Eupolyphaga,
hirudo, qichan, tabanus
|
|
Wang42
(2000)
|
1
|
CHB
|
33.4
(NS)
|
35.0
(NS)
|
Astragalus
membranaceus
|
|
Wu43
(1997)
|
1
|
CHB
|
38.2
± 6.5
|
36.4
± 7.9
|
Salvia
multiorrhiza
|
|
Wu44
(1998)
|
1
|
CHB
|
36.5
(NS)
|
36.0
(NS)
|
Atractylodes
alba,
Amyda sinensis, bupleurum, amomum, Citrus
reticulata, Dioscorea opposita, Glycyrrhiza
uralensis, Lycium chinensis, Panax
ginseng, Pinellia ternata, Peonia
alba, Poria cocos, Rheum officinale,
Scutellaria barbata, Trionyx sinensis
|
|
Zhang45
(1999)
|
1
|
CHB
|
28.4
(NS)
|
27.6
(NS)
|
Bufotoxin
|
|
Zhang46
(1997)
|
1
|
CHB
|
36
(NS)
|
34.5
(NS)
|
Long
dan xie gan tang (ingredients not specified)
|
|
Zhang47
(1999)
|
1
|
CHB
|
20–50
(NS) for both CHM & CHM+ IFN-
groups
|
20–50
(NS)
|
Bupleurum,
erhoutao, Hypericum japonicum, Imperata
cylindrica, Panax ginseng, jixueteng,
tianwangye, wuahuaxueteng
|
|
Zhao48
(1996)
|
1
|
CAH;
CPH
|
36.0
± 11.1
|
34.2
± 13.4
|
Astragalus
membranaceus, Prunus persica, Curcuma longa zedoaria
|
|
Zhou49
(1999)
|
1
|
CHB
|
36.2
± 10.3
|
35.6
± 11.0
|
Atractylodes
alba,
Astragalus membranaceus, Crinis
carbonisatus, Prunus persica, Phyllanthus,
Polygonum multiflorum, Poria cocos, Rehmannia
glutinosa, Salvia multiorrhiza, Schisandra
chinensis, Taraxacum mongolicum
|
|
Note.
CAH = chronic active hepatitis B; CHB = chronic
hepatitis B; CPH = chronic persistent hepatitis B;
IFN-
= interferon alfa; LC = liver cancer; NS = not
specified.
|
|
aPoints
awarded for modified Jadad scale criteria (how
studies randomized patients and handled dropouts
or withdrawals): low score = 0 or 1; high score =
2 or 3; maximum possible total score = 3.
|
|
We retained studies that reported the use of different forms
of interferon alfa (interferon alfa, n = 20 studies;
interferon alfa-1b, n = 2; interferon alfa-2a, n =
3; interferon alfa-2b, n = 2) in the treatment or
control groups. Previous research has documented
similarities in the effectiveness of the different forms
of interferon alfa in the treatment of hepatitis B.
Data Abstraction
Two reviewers (M. M. and M. B.), who were blinded to author,
affiliation, and journal title, reviewed the 27 studies.
The following data were abstracted through
standardized forms: publication year; diagnosis;
average patient age; definition of diagnosis used;
Chinese herbal medicine treatment used; type of interferon
alfa used; interferon alfa doses; whether the treatment
arm involved Chinese herbal medicine alone or
Chinese herbal medicine combined with interferon
alfa; the total number of subjects in each
treatment arm; and the number of treatment responders in
each treatment arm for any of the endpoints HBsAg, HBeAg,
and HBV DNA. Any disparities in data abstraction were
resolved through a consensus process in which a
third investigator served as arbitrator (J. M. C.).
Quality Scoring
Five of these trials compared an injected active ingredient
extracted from a Chinese herbal medicine with injected
interferon alfa
and thus could have included double-blinding within
the study design. However, in the remaining 22 studies, blinding
was obviously not possible because those studies compared
an orally administered Chinese herbal medicine with
injected interferon alfa. Thus, we created a
modified scale based on the method of Jadad,
limiting our assessment of study quality to how
studies randomized patients and handled dropouts or
withdrawals (low score = 0 or 1; high score = 2 or
3; maximum possible total score = 3).
Statistical Analysis
We used the Stata statistical software package (version 6.0;
Stata Corp, College Station, Tex) for data management
and analysis. We calculated relative risk of cure
from the data in the original studies for use in
the meta-analysis. These relative risks were calculated
as the probability of seroreversion in the treated group
divided by the probability of seroreversion in the control
group. Thus, relative risk values greater than 1.0 are
consistent with a beneficial effect of Chinese
herbal medicine used alone (vs interferon alfa) or
Chinese herbal medicine in combination with
interferon alfa (vs interferon alfa). In 4 of the studies,
we encountered individual contingency table cells with
no patients.
In calculating relative risk for these studies, the
value 0.5 was added to all 4 cells of the
contingency table.
Confidence intervals for the relative risks were
estimated by the Woolf method.
Studies with missing data were excluded from analysis (n
= 4). We used the Egger et al. regression asymmetry test
to examine our meta-analysis data for publication bias
We constructed our groupings for meta-analysis as follows:
(1) to assess the effectiveness of Chinese herbal
medicine as a stand-alone therapy, all studies of
Chinese herbal medicine alone (vs interferon alfa)
were analyzed together (Figure 1;
Table 2);
(2) to assess the effectiveness of Chinese herbal medicine
as an adjunct to interferon alfa, all studies of Chinese herbal
medicine combined with interferon alfa (vs interferon alfa)
were analyzed together (Figure 1;
Table 3);
(3) to examine the effectiveness of specific active
components extracted from Chinese herbal medicines,
subanalyses of those active components were
conducted when 2 or more studies reporting use of the same
active component were available. Within each of these
groupings, the outcome we studied was seroreversion
of 3 dichotomous endpoints: HBsAg, HBeAg, and HBV
DNA. Using these endpoints, we calculated the
treatment effect of Chinese herbal medicine alone (vs
interferon alfa) and Chinese herbal medicine
combined with interferon alfa (vs interferon alfa);
we report these results as relative risk of cure,
with 95% confidence intervals. A relative risk of cure >
1 indicates effectiveness of the treatment evaluated.
FIGURE 1 —Meta-analysis forest plots for
seroreversion among patients with chronic hepatitis B
infection treated with Chinese herbal medicine: pooled
relative risks and 95% confidence intervals, by trial.
TABLE 2 —Seroreversion in Chronic Hepatitis B:
Chinese Herbal Medicine (CHM) Alone vs Interferon Alfa (IFN-)
|
|
Seroconversion,
No. of Responders/Total No. of Patients in Group
|
|
|
Study
|
IFN-
|
CHM
|
RR
(95% CI)
|
|
|
|
Hepatitis
B surface antigena
|
|
Dai25
(1998)
|
2/15
|
2/15
|
1.00
(0.16, 6.20)
|
|
Li31
(1998)
|
2/25
|
2/30
|
0.83
(0.13, 5.50)
|
|
Li32
(1999)
|
1/32
|
3/38
|
2.53
(0.28, 23.1)
|
|
Li33
(1997)
|
0/40
|
1/40
|
3.00
(0.13, 71.5)
|
|
Lu36
(1992)
|
15/79
|
33/97
|
1.79
(1.05, 3.05)
|
|
Song39
(1994)
|
5/30
|
20/50
|
2.40
(1.01, 5.72)
|
|
Wang40
(1997)
|
2/11
|
33/64
|
2.84
(0.79, 10.1)
|
|
Zhao48
(1996)
|
0/30
|
10/30
|
21.0
(1.29, 342)
|
|
Overall
|
27/262
|
104/364
|
2.00
(1.35, 2.97)
|
|
Hepatitis
B e antigenb
|
|
Cai24
(1997)
|
23/50
|
28/63
|
0.97
(0.64, 1.45)
|
|
Chen19
(2000)
|
17/29
|
15/29
|
0.88
(0.55, 1.41)
|
|
Dai25
(1998)
|
7/15
|
5/15
|
| | |
