A Systematic Review on Natural Medicines for

the Prevention and Treatment of Alzheimers

Disease with Meta-Analyses of Intervention
Effect of Ginkgo
Mengmeng Yang,
*
,a
Dan Dan Xu,
*
,a
Yan Zhang,
*
Xinyou Liu,
*
Robin Hoeven

and William Chi Shing Cho

*
Department of Pharmacy, Tangdu Hospital
The Fourth Military Medical University, Xian, China

Faculty of Life Sciences, University of Manchester

Manchester, United Kingdom

Department of Clinical Oncology, Queen Elizabeth Hospital

Hong Kong SAR
Abstract: We performed a systematic review to evaluate the efcacy of natural medicines for
the treatment of Alzheimers disease (AD) in randomized controlled trials (RCTs). Disease-
specic and intervention terms were searched in MEDLINE, EMBASE, the Cochrane Li-
brary and PsycINFO to identify RCTs for the AD intervention of natural medicines, and
searched for literatures in English language. The RCTs compared natural medicines and
either placebo or orthodox medication in AD patients. The quality of literature was evaluated
by Jadads score and the Cochrane assessing tool to reduce the risk of bias. Meta-analysis
and the heterogeneity of results across the trials were performed. Out of the literatures, 21
clinical reports were included in this review that satised the particular selection criteria.
Apart from Ginkgo, other treatments we came across had minimal benets and/or the
methodological quality of the available trials was poor. The meta-analyses showed that
Ginkgo had better outcomes than the placebo, with the standardized mean difference (SMD)
between Ginkgo and the placebo on cognition being 1.62 (95% CI: 2.69 to 0.56) and
on activities of daily living being 1.55 (95% CI: 2.55 to 0.55), with the existence
of signicant heterogeneity across studies. The meta-analysis for assessing the prevention
effect of Ginkgo against AD suggested that risk ratio (RR) is 1.06 (95% CI: 0.92 to 1.22)
between Gingko and the placebo, with no signicant heterogeneity across studies (test for
Correspondence to: Dr. William Chi Shing Cho, Department of Clinical Oncology, Queen Elizabeth Hospital, 30
Gascoigne Road, Kowloon, Hong Kong SAR. Tel: (852) 2958-5441, Fax: (852) 2958-5455, E-mail:
williamcscho@gmail.com or Dr. Mengmeng Yang, Clinical Pharmacy Services, Department of Pharmacy,
Tangdu Hospital of the Fourth Military Medical University, 569 Xinsi Road, Xian, China. Tel: (86) 29-8477-
7547, Fax: (86) 29-8477-7154, E-mail: myang td@163.com
a
These authors contributed equally to this work.
The American Journal of Chinese Medicine, Vol. 42, No. 3, 505521
2014 World Scientic Publishing Company
Institute for Advanced Research in Asian Science and Medicine
DOI: 10.1142/S0192415X14500335
505
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
heterogeneity, p 0:49). Our results suggest that Ginkgo may help established AD patients
with cognitive symptoms but cannot prevent the neurodegenerative progression of the
disease.
Keywords: Alzheimers Disease; Ginkgo; Meta-Analysis; Natural Medicines; Systematic
Review.
Introduction
Alzheimers disease (AD) is the most common form of dementia, and is dened as a
progressive neurodegenerative disease characterized by cognitive deterioration together
with behavioral disturbances and declining activities. It is hypothesized that AD is caused
by the reduced level of the neurotransmitter acetylcholine, the deposition of beta-amyloid,
or tau protein abnormality in the brain. Nowadays the available therapeutic medications
mainly include cholinesterase inhibitors (tacrine, rivastigmine, galantamine, or donepezil),
NMDA receptor antagonist (memantine), anti-inammatory drugs (NSAIDs), nootropic
agent (piracetam) and a variety of antioxidants (vitamins) (Selkoe, 2001). However, the
success of these treatment strategies cannot be predicted and often offer very limited
benets in clinic.
Due to the limitation of pharmacological interventions, the therapies from natural origin
provide alternative and complementary choices for patients with AD (Perry et al., 1999). In
this review, we carried out a systematic analysis of randomized controlled trials (RCTs), to
assess the efcacy of natural medicines vs. placebo or active control in the prevention or
treatment for AD. This review aims to provide a comprehensive assessment on the inter-
vention of natural medicines in clearly specied AD or dementia of AD type. The non-
Alzheimer types of dementia, including vascular dementia, dementia with Lewy body,
mixed dementia and Parkinsons disease, were excluded. Natural medicines included in
this review comprise original or extracted forms of naturally occurring medicines. The
quality of RCTs was assessed with Jadads score, followed by Cochrane Collaborations
Risk of bias tool. The summary intervention effect estimate is calculated as a weighted
average of the intervention effects estimated in the individual studies by means of random-
effects meta-analysis, which give us a deeper insight in the efcacy of natural medicines
for AD.
Methods
Search Strategy
Disease-specic terms (Alzheimer disease or dementia) combined with intervention
terms (phytotherapy or phytomedicine or natural product or natural medicine or
herb) were searched in MEDLINE from 1950 to July 2013, EMBASE from 1980 to July
2013, the Cochrane Library from 1990 to July 2013 and PsycINFO from 1990 to July 2013
to identify RCTs for AD intervention of natural medicines in the English language.
506 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Inclusion Criteria
Types of Studies
All published, double-blinded, randomized, placebo-controlled trials examining efcacy of
natural medicines in patients with AD were included.
Types of Participants
Diagnostic criteria included Diagnostic and Statistical Manual of Mental Disorders (DSM),
National Institute of Neurological and Communicative Disorders and Stroke-Alzheimers
Disease and Related Disorders Association (NINCDSADRDA).
Types of Interventions
Comparisons of natural medicines and either placebo or orthodox medication were per-
formed.
Types of Outcome Measures
(1) Cognition: Mini-Mental State Examination (MMSE), Cognitive subscale of the Alz-
heimers Disease Assessment Scale (ADAS-Cog), or Syndrom-Kurz test (SKT); (2) Ac-
tivities of daily living (ADL): The AD Activities of Daily Living International Scale (ADL-
IS), Gottfries-Brne-Steen-Activities of Daily Living (GBS-ADL) scale, Geriatric Evalu-
ation by Relatives Rating Instrument (GERRI), or Nrnberger Alters-Beobachtungsskala
(NAB); (3) Global impression: Alzheimers Disease Cooperative Study Clinical Global
Impression of Change (ADCS-CGIC), Clinical Dementia Rating Scale (CDR), Clinical
Global Impression (CGI), or Gottfries-Brne-Steen (GBS); (4) Behavior/Neuropsychology:
The Neuropsychiatric Inventory (NPI) or Hamilton Rating Scale for Depression (HRSD);
(5) Quality of life: Alzheimer Disease Related Quality of Life (ADRQL); (6) Communi-
cation/Social Interaction: Verbal Fluency Test (VFT).
Data Collection and Analysis
Two authors (MY, DDX) were independently involved in database searching, reference
screening and inclusion assessment. Any discrepancies were resolved by consensus with
the other authors (YZ, XL, RH, WCC). One author (MY) performed data extraction and
constructed a worksheet including authors and publication year, country, duration, sample
size, trial type, age of the participants, diagnostic system, outcome measures, and details of
intervention for each study. Another author (DDX) veried the extracted data. Since de-
mentia was reported as several different types, the studies were included with either
reported AD cases only or presented sub-group analyses for AD patients.
Selection of Studies with Jadads Scores
The quality of the studies was assessed using Jadads scores by two authors (MY, DDX)
(Jadad et al., 1996) and nally conrmed by another author (WCC). Scores were assigned
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 507
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
if the trial was described as randomized (1 point), double-blind (1 point) or had with-
drawals/dropouts (1 point). The additional points could be awarded (1 extra point in each
case) if the method of randomization or double blinding was appropriate. Conversely,
points could be deducted (1 point in each case) if the study was described as randomized or
double blind, but the methods were inappropriate. The trials with Jadads scores 3 were
included in this review.
Assessment of Risk of Bias in Included Studies
Trials with Jadads scores 3 were subjected to Cochrane Collaborations tool for
assessing the risk of bias. The assessment domains include selection bias (random sequence
generation; allocation concealment), performance bias (blinding of participants and per-
sonnel; blinding of outcome; incomplete outcome data), reporting bias (selective report-
ing), and other bias (other sources of bias).
Assessment of Heterogeneity and Data Synthesis
All statistics were conducted with Review Manager (RevMan) version 5.2 (Copenhagen:
The Nordic Cochrane Centre, the Cochrane Collaboration, 2012). For a specic outcome,
the heterogeneity of results across the trials, which was analyzed by the intention-to-treat
(ITT) approach, was evaluated. A xed-effects model, without weighting of the studies,
was used when there was a lack of signicant heterogeneity (p > 0:10. If signicant
heterogeneity was observed (p 0:10), a random effects model, which assigns a weight to
each study based on individual study variance as well as between study variance, was used
to pool the results together. The standardized mean difference (SMD) for the same outcome
is used as a summary statistic in meta-analysis and presented in a forest plot. The likelihood
of publication bias was assessed in the context of visual inspection of funnel plots.
Results
Description of the Included Studies
Study ow diagram in Fig. 1 illustrates the result of the search and the process of screening
and selecting studies for inclusion in the review. 1228 records were retrieved by the
electronic searches, of which 704 were duplicates and 491 were excluded according to
inclusion criteria in the Methods for general relevance. 35 full-text articles with Jadad
scores 3 were considered as potentially eligible after screening. 14 studies were further
excluded without explicit presentation for dementia of the AD type. The remaining 21
studies were included in this review, of which 8 were the interventions of Salvia ofcinalis
(Akhondzadeh et al., 2003a), Melissa ofninalis (Akhondzadeh et al., 2003b), Crocus
sativus (Akhondzadeh et al., 2010), curcumin (Baum et al., 2008), huperzine A (Rai
et al., 2011; Xu et al., 1999) and Fuzhisan (Bi et al., 2011). A Characteristics of Included
Studies table (Table 1) comprises a succinct summary of the trials information, including
508 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
T
a
b
l
e
1
.
C
h
a
r
a
t
e
r
i
s
t
i
c
s
o
f
I
n
c
l
u
d
e
d
S
t
u
d
i
e
s
(
J
a
d
a
d
S
c
o
r
e

3
)
N
a
t
u
r
a
l
M
e
d
i
c
i
n
e
F
i
r
s
t
A
u
t
h
o
r
a
n
d
Y
e
a
r
C
o
u
n
t
r
y
D
u
r
a
t
i
o
n
T
y
p
e
S
a
m
p
l
e
S
i
z
e
A
g
e
D
i
a
g
n
o
s
t
i
c
S
y
s
t
e
m
O
u
t
c
o
m
e
M
e
a
s
u
r
e
s
D
e
t
a
i
l
s
o
f
I
n
t
e
r
v
e
n
t
i
o
n
G
i
n
k
g
o
b
i
l
o
b
a
V
e
l
l
a
s
e
t
a
l
.
(
2
0
1
2
)
F
r
a
n
c
e
5
y
e
a
r
s
R
a
n
d
o
m
i
s
e
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
c
l
i
n
i
c
a
l
t
r
i
a
l
n

2
8
5
4

7
0
y
e
a
r
s
D
S
M
-
I
V
,
N
I
N
C
D
S
-
A
D
R
D
A
,
M
M
S
E
a
n
d
C
D
R
D
S
M
-
I
V
a
n
d
N
I
N
C
D
S
-
A
D
R
D
A
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
I
h
l
e
t
a
l
.
(
2
0
1
2
)
U
k
r
a
i
n
e
2
4
w
e
e
k
s
R
a
n
d
o
m
i
s
e
d
c
o
n
t
r
o
l
l
e
d
t
r
i
a
l
n

4
0
4
(
n

3
3
3
f
o
r
A
D
)

5
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
S
K
T
,
N
P
I
,
N
P
I
c
a
r
e
g
i
v
e
r
d
i
s
t
r
e
s
s
s
c
o
r
e
,
A
D
C
S
-
C
G
I
C
,
A
D
L
-
I
S
,
D
E
M
Q
O
L
-
p
r
o
x
y
a
n
d
V
F
T
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
L
e
B
a
r
s
e
t
a
l
.
(
2
0
0
2
)
G
e
r
m
a
n
y
,
U
S
A
5
2
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
-
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
s
t
u
d
y
n

1
2
2

4
5
y
e
a
r
s
I
C
D
-
1
0
a
n
d
D
S
M
-
I
I
I
-
R
A
D
A
S
-
C
o
g
a
n
d
G
E
R
R
I
1
2
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
L
e
B
a
r
s
e
t
a
l
.
(
2
0
0
0
)
G
e
r
m
a
n
y
,
U
S
A
2
6
w
e
e
k
s
D
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,

x
e
d
d
o
s
e
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
s
t
u
d
y
n

2
4
4

4
5
y
e
a
r
s
I
C
D
-
1
0
a
n
d
D
S
M
-
I
I
I
-
R
A
D
A
S
-
C
o
g
,
G
E
R
R
I
a
n
d
C
G
I
C
1
2
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
L
e
B
a
r
s
e
t
a
l
.
(
1
9
9
7
)
U
S
A
5
2
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
-
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
s
t
u
d
y
n

3
2
7

4
5
y
e
a
r
s
I
C
D
-
1
0
a
n
d
D
S
M
-
I
I
I
-
R
A
D
A
S
-
C
o
g
,
G
E
R
R
I
a
n
d
C
G
I
C
1
2
0
m
g
/
d
a
y
e
x
t
r
a
c
t
o
f
G
b
i
l
o
b
a
:
P
l
a
c
e
b
o
S
c
h
n
e
i
d
e
r
e
t
a
l
.
(
2
0
0
5
)
U
S
A
2
6
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
a
r
a
l
-
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
t
r
i
a
l
n

5
1
3

6
0
y
e
a
r
s
M
M
S
E
A
D
A
S
-
C
o
g
a
n
d
A
D
C
S
-
C
G
I
C
1
2
0
m
g
o
r
2
4
0
m
g
/
d
a
y
e
x
t
r
a
c
t
o
f
G
i
n
k
g
o
:
P
l
a
c
e
b
o
N
a
p
r
y
e
y
e
n
k
o
e
t
a
l
.
(
2
0
0
9
)
U
k
r
a
i
n
e
2
2
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
a
r
a
l
-
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
t
r
i
a
l
n

3
9
5
(
n

2
1
4
f
o
r
A
D
)

5
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
a
n
d
M
M
S
E
S
K
T
,
V
F
T
,
C
D
T
,
N
P
I
,
H
R
S
D
,
a
n
d
G
B
S
-
A
D
L
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
D
e
K
o
s
k
y
e
t
a
l
.
(
2
0
0
8
)
U
S
A
6
.
1
y
e
a
r
s
R
a
n
d
o
m
i
z
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
-
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
m
u
l
t
i
c
e
n
t
e
r
s
t
u
d
y
n

5
2
3

7
5
y
e
a
r
s
D
S
M
-
I
V
a
n
d
C
D
R
D
S
M
-
I
V
2
4
0
m
g
/
d
a
y
e
x
t
r
a
c
t
o
f
G
i
n
k
g
o
:
P
l
a
c
e
b
o
M
a
z
z
a
e
t
a
l
.
(
2
0
0
6
)
I
t
a
l
y
2
4
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
s
t
u
d
y
n

7
6
5
0

8
0
y
e
a
r
s
D
S
M
-
I
V
a
n
d
S
K
T
S
K
T
,
M
M
S
E
,
a
n
d
C
G
I
1
6
0
m
g
/
d
a
y
e
x
t
r
a
c
t
o
f
G
i
n
k
g
o
:
P
l
a
c
e
b
o
:
d
o
n
e
p
e
z
i
l
(
5
m
g
/
d
a
y
)
K
a
n
o
w
s
k
i
a
n
d
H
o
e
r
r
(
2
0
0
3
)
G
e
r
m
a
n
y
2
4
w
e
e
k
s
M
u
l
t
i
-
c
e
n
t
e
r
,
d
o
u
b
l
e
-
b
l
i
n
d
,
d
o
n
e
p
e
z
i
l
-
c
o
n
t
r
o
l
l
e
d
,
r
a
n
-
d
o
m
i
z
e
d
t
r
i
a
l
n

2
2
2
>
5
4
y
e
a
r
s
D
S
M
-
I
I
I
-
R
,
S
K
T
,
M
M
S
E
,
a
n
d
M
A
D
R
S
A
D
A
S
-
C
o
g
,
S
K
T
,
C
G
I
,
a
n
d
N
A
B
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
K
a
n
o
w
s
k
i
e
t
a
l
.
(
1
9
9
7
)
G
e
r
m
a
n
y
2
4
w
e
e
k
s
P
r
o
s
p
e
c
t
i
v
e
,
r
a
n
d
o
m
i
z
e
d
,
d
o
u
-
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
,
m
u
l
t
i
-
c
e
n
t
e
r
s
t
u
d
y
n

2
1
6
>
5
4
y
e
a
r
s
D
S
M
-
I
I
I
-
R
,
S
K
T
,
M
M
S
E
,
a
n
d
M
A
D
R
S
C
G
I
I
t
e
m
2
,
S
K
T
,
a
n
d
N
A
B
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 509
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
T
a
b
l
e
1
.
(
C
o
n
t
i
n
u
e
d
)
N
a
t
u
r
a
l
M
e
d
i
c
i
n
e
F
i
r
s
t
A
u
t
h
o
r
a
n
d
Y
e
a
r
C
o
u
n
t
r
y
D
u
r
a
t
i
o
n
T
y
p
e
S
a
m
p
l
e
S
i
z
e
A
g
e
D
i
a
g
n
o
s
t
i
c
S
y
s
t
e
m
O
u
t
c
o
m
e
M
e
a
s
u
r
e
s
D
e
t
a
i
l
s
o
f
I
n
t
e
r
v
e
n
t
i
o
n
Y
a
n
c
h
e
v
a
e
t
a
l
.
(
2
0
0
9
)
B
u
l
g
a
r
i
a
2
2
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
a
c
t
i
v
e
-
c
o
n
t
r
o
l
l
e
d
,
m
u
l
t
i
-
c
e
n
-
t
e
r
t
r
i
a
l
n

9
6

5
0
y
e
a
r
s
N
I
N
C
D
S
-
A
D
R
D
A
,
C
D
T
,
a
n
d
S
K
T
S
K
T
,
N
P
I
,
A
D
L
s
u
b
-
s
c
o
r
e
o
f
G
B
S
,
H
R
S
D
,
C
D
T
a
n
d
V
F
T
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
d
o
n
e
p
e
z
i
l
(
i
n
i
t
i
a
l
l
y
5
m
g
,
a
f
t
e
r
4
w
e
e
k
s
1
0
m
g
/
d
a
y
)
:
E
G
b
7
6
1
&
d
o
n
e
p
e
z
i
l
c
o
m
-
b
i
n
e
d
(
s
a
m
e
d
o
s
e
s
)
M
a
u
r
e
r
e
t
a
l
.
(
1
9
9
7
)
G
e
r
m
a
n
y
3
m
o
n
t
h
s
D
o
u
b
l
e
-
b
l
i
n
d
,
r
a
n
d
o
m
i
z
e
d
,
p
l
a
-
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
p
a
r
a
l
l
e
l
-
g
r
o
u
p
d
e
s
i
g
n
n

2
0
5
0

8
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
a
n
d
D
S
M
-
I
I
I
-
R
S
K
T
,
t
r
a
i
l
m
a
k
i
n
g
t
e
s
t
,
A
D
A
S
,
a
n
d
C
G
I
2
4
0
m
g
/
d
a
y
E
G
b
7
6
1
:
P
l
a
c
e
b
o
S
.
o
f

c
i
n
a
l
i
s
A
k
h
o
n
d
z
a
d
e
h
e
t
a
l
.
(
2
0
0
3
)
(
S
O
)
I
r
a
n
4
m
o
n
t
h
s
P
a
r
a
l
l
e
l
g
r
o
u
p
,
p
l
a
c
e
b
o
c
o
n
-
t
r
o
l
l
e
d
,
m
u
l
t
i
c
e
n
t
e
r
t
r
i
a
l
n

4
2
b
e
t
w
e
e
n
6
5
a
n
d
8
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
,
A
D
A
S
-
c
o
g
a
n
d
C
D
R
A
D
A
S
-
c
o
g
a
n
d
C
D
R
-
S
B
S
a
l
v
i
a
o
f

c
i
n
a
l
i
s
e
x
t
r
a
c
t
,
6
0
d
r
o
p
s
/
d
a
y
:
P
l
a
c
e
-
b
o
M
.
o
f

n
i
n
a
l
i
s
B
u
r
n
s
e
t
a
l
.
(
2
0
1
1
)
U
K
1
2
w
e
e
k
s
D
o
u
b
l
e
-
b
l
i
n
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
,
p
l
a
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
,
r
a
n
d
o
m
-
i
z
e
d
t
r
i
a
l
n

1
1
4
>
6
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
P
A
S
,
N
P
I
a
n
d
B
l
a
u
Q
O
L
s
c
a
l
e
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
A
k
h
o
n
d
z
a
d
e
h
e
t
a
l
.
(
2
0
0
3
)
(
M
O
)
I
r
a
n
4
m
o
n
t
h
s
P
a
r
a
l
l
e
l
g
r
o
u
p
,
p
l
a
c
e
b
o
c
o
n
-
t
r
o
l
l
e
d
,
m
u
l
t
i
c
e
n
t
e
r
t
r
i
a
l
n

4
2
b
e
t
w
e
e
n
6
5
a
n
d
8
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
,
A
D
A
S
-
c
o
g
,
a
n
d
C
D
R
A
D
A
S
-
c
o
g
a
n
d
C
D
R
-
S
B
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
H
u
p
e
r
z
i
n
e
A
X
u
e
t
a
l
.
(
1
9
9
9
)
C
h
i
n
a
6
0
d
a
y
s
M
u
l
t
i
c
e
n
t
e
r
,
p
r
o
s
p
e
c
t
i
v
e
,
d
o
u
-
b
l
e
-
b
l
i
n
d
,
d
o
u
b
l
e
-
m
i
m
i
c
,
p
a
r
a
l
l
e
l
,
p
o
s
i
t
i
v
e
c
o
n
t
r
o
l
l
e
d
a
n
d
r
a
n
d
o
m
i
z
e
d
t
r
i
a
l
n

6
0
N
/
A
N
I
N
C
D
S
-
A
R
D
R
A
H
D
S
-
R
,
I
A
D
L
,
G
B
S
-
S
D
S
,
T
E
S
S
,
C
G
I
,
C
C
G
I
,
e
t
c
.
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
R
a

i
e
t
a
l
.
(
2
0
1
1
)
U
S
A
1
6
w
e
e
k
s
M
u
l
t
i
c
e
n
t
e
r
,
3
-
a
r
m
,
r
a
n
d
o
m
i
z
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
d
o
s
e
-
e
s
c
a
l
a
t
i
o
n
t
r
i
a
l
n

2
1
0

5
0
y
e
a
r
s
N
I
N
C
D
S
-
A
R
D
R
A
a
n
d
M
M
S
E
A
D
A
S
-
c
o
g
,
N
P
I
,
A
D
L
a
n
d
M
M
S
E
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
C
.
s
a
t
i
v
u
s
(
S
a
f
f
r
o
n
)
A
k
h
o
n
d
z
a
d
e
h
e
t
a
l
.
(
2
0
1
0
)
I
r
a
n
1
6
w
e
e
k
s
D
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
p
a
r
a
l
l
e
l
-
g
r
o
u
p
t
r
i
a
l
n

4
6
>
5
5
y
e
a
r
s
D
S
M
-
I
V
,
N
I
N
C
D
S
-
A
D
R
D
A
,
a
n
d
M
M
S
E
A
D
A
S
-
C
o
g
a
n
d
C
D
R
-
S
B
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
C
u
r
c
u
m
i
n
B
a
u
m
e
t
a
l
.
(
2
0
0
8
)
H
o
n
g
K
o
n
g
6
m
o
n
t
h
s
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
c
e
b
o
-
c
o
n
-
t
r
o
l
l
e
d
,
r
a
n
d
o
m
i
z
e
d
t
r
i
a
l
n

3
4

5
0
y
e
a
r
s
N
I
N
C
D
S
-
A
D
R
D
A
M
M
S
E
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
F
u
z
h
i
s
a
n
(
f
o
r
m
u
l
a
)
B
i
e
t
a
l
.
(
2
0
1
1
)
C
h
i
n
a
1
2
w
e
e
k
s
R
a
n
d
o
m
i
z
e
d
,
d
o
u
b
l
e
-
b
l
i
n
d
,
p
l
a
-
c
e
b
o
-
c
o
n
t
r
o
l
l
e
d
p
i
l
o
t
s
t
u
d
y
n

2
2
N
/
A
N
I
N
C
D
S
-
A
D
R
D
A
,
D
S
M
-
I
V
,
C
D
R
,
a
n
d
M
M
S
E
A
D
A
S
-
C
o
g
a
n
d
N
P
I
F
u
z
h
i
s
a
n
e
x
t
r
a
c
t
(
1
0
m
g
/
d
a
y
)
:
P
l
a
c
e
b
o
510 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
the study participants, location, design, diagnoses, interventions and outcome measures in
the included studies.
Thirteen studies were identied for Ginkgo intervention, of which 2 were excluded
because they did not involve an ITT analysis (Kanowski et al., 1997; Maurer et al., 1997),
2 were repeated publications (Le Bars et al., 2000; 2002) and 1 lacked a placebo control
(Yancheva et al., 2009). Eight studies were nally included in quantitative synthesis,
namely meta-analysis for the comparison between Ginkgo and placebo, of which 6 studied
the effect of treatment (Le Bars et al., 1997; Kanowski and Hoerr, 2003; Schneider et al.,
2005; Mazza et al., 2006; Napryeyenko et al., 2009; Ihl et al., 2012) and two studied the
preventative effect of AD (DeKosky et al., 2008; Vellas et al., 2012).
Records idened through
database searching
(MEDLINE n=498
EMBASE n=675
Cochrane n=106
PsycINFO n=393)
S
c
r
e
e
n
i
n
g
I
n
c
l
u
d
e
d
E
l
i
g
i
b
i
l
i
t
y
I
d
e
n

c
a

o
n
Addional records idened
through other sources
(n=0)
Records aer duplicates removed
(n=524)
Records screened
(n=524)
Records excluded
(n=489
not complying to the
inclusion criteria in
Methods)
Full-text arcles assessed
for eligibility
(n=35, Jadads scores3)
Full-text arcles excluded,
not indicang demena of
AD type
(n=14)
Studies included in qualitave synthesis
(n=13, Ginkgo intervenon;
n=8, intervenons of Salvia ocinalis,
Melissa ocinalis, Crocus savus, curcumin,
huperzine A, Fuzhisan)
Studies included in quantave
synthesis (meta-analysis)
(n=6, Ginkgo for AD treatment;
n=2, Ginkgo for AD prevenon)
Records excluded for
Ginkgo intervenon
(n=2 not ITT analysis:
Maurer et al. (1997) and
Kanowski et al. (1997);
n=1 not versus placebo:
Yancheva et al. (2009);
n=2 repeated publicaons:
Le Bars et al. (2000, 2002)
Figure 1. Study ow diagram.
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 511
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Risks of Bias in Included Studies
Firstly, all trails were screened by Jadads 5-point criteria, which demonstrated whether
randomization, double-blind and dropouts were described. Figures 2 and 3 summarize the
general risk of bias of the included studies with Jadads scores 3, the former illustrates
the percentages of judgment across all included studies and the latter presents the judgment
for each domain in individual study shown as a cross-tabulation. If insufcient detail was
reported of what happened in the study, the judgment for a specic domain was considered
as unclear risk of bias.
Fourteen studies described a random component in the sequence generation process
using a computer for randomnumber generation (Kanowski et al., 1997; Le Bars et al., 1997;
Maurer et al., 1997; Akhondzadeh et al., 2003a,b; 2010; Kanowski and Hoerr, 2003;
Schneider et al., 2005; Mazza et al., 2006; DeKosky et al., 2008; Napryeyenko et al., 2009;
Burns et al., 2011; Yancheva et al., 2009; Ihl et al., 2012; Vellas et al., 2012). In 12 trials,
allocation concealment was achieved by central allocation, sequentially numbered drug
containers of identical appearance and sequentially numbered, opaque, sealed envelopes
(Kanowski et al., 1997; Le Bars et al., 1997; Maurer et al., 1997; Akhondzadeh et al., 2003a,b;
Figure 2. Risk of bias graph presented as percentages across all included studies.
Figure 3. Risk of bias summary presented as listed items for individual trial.
512 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
2010; Kanowski and Hoerr, 2003; Schneider et al., 2005; Mazza et al., 2006; DeKosky et al.,
2008; Napryeyenko et al., 2009; Burns et al., 2011; Yancheva et al., 2009; Ihl et al., 2012;
Vellas et al., 2012). Ten studies were assessed as appropriate double-blinding of participants
and personnel both for performance and outcome assessment (Kanowski et al., 1997; Le Bars
et al., 1997; Kanowski and Hoerr, 2003; Schneider et al., 2005; Mazza et al., 2006; DeKosky
et al., 2008; Napryeyenko et al., 2009; Yancheva et al., 2009; Burns et al., 2011; Vellas et al.,
2012). Three studies provided detailed information for double blinding during performance
but insufcient information for the blinding of outcome assessment (Akhondzadeh et al.,
2003a,b; 2010). All studies mentioned the number of participants completing or withdrawing
fromthe trials, but eight studies reported the dropouts due to inefcacy of intervention in high
risk of attrition bias (Le Bars et al., 1997, 2000, 2002; Xu et al., 1999; Mazza et al., 2006;
Baum et al., 2008; Yancheva et al., 2009; Burns et al., 2011). Three studies were in high risk
of bias due to selective outcome reporting (Xu et al., 1999; Baumet al., 2008; Bi et al., 2011).
Eight studies had a potential source of bias related to the specic study design used (Maurer
et al., 1997; Xu et al., 1999; Akhondzadeh et al., 2003a,b; 2010; Bi et al., 2011; Yancheva
et al., 2009; Burns et al., 2011).
Description of the Reported Natural Medications
The reported natural medications in RCTs involve S. ofcinalis, Melissa ofcinalis, C.
sativus, curcumin, huperzine A, Fuzhisan (Bi et al., 2011) and Gingko biloba. Akhond-
zadeh et al. (2003a,b; 2010) reported that the extract of C. sativus, S. ofcinalis extract or
Melissa ofcinalis produced a signicantly better outcome on cognitive functions (ADAS-
Cog, CDR-SB) than the placebo over a 4-month period. Baum et al. (2008) reported
curcumin showed a non-signicant tendency of a slower decline rather than an improve-
ment in cognition (MMSE) in a 6-month pilot clinical trial. Huperzine A, extracted from
the Chinese herb Huperzia serrata, did not show cognitive benet (ADAS-Cog) at the dose
of 200 g BID in patients with mild to moderate AD, the 400 g BID dose showed non-
signicant tendency of cognitive enhancement as measured by the ADAS-Cog and the
MMSE (Rai et al., 2011). In the other trial for huperzine at the dose of 200 g BID (Xu
et al., 1999), also no signicant difference was found vs. the placebo for the psychological
assessment. It did, however, reduce the pathological changes of the free oxygen radicals in
the plasma and erythrocytes of patients with AD. Fuzhisan, a Chinese herbal formula, was
reported to signicantly improve ADAS-Cog scores, NPI scores and the regional cerebral
metabolic rate of glucose consumption (rCM-Rglc), which suggests that Fuzhisan treat-
ment may have a positive effect on cognition, behavior functions, and rCM-Rglc in mild-
to-moderate AD patients (Bi et al., 2011). Ginkgo was the most frequently studied natural
medicine for AD and was subjected to quantitative assessment with meta-analysis.
Quantitative Assessment of the Efcacy of Ginkgo Intervention
Two meta-analyses were performed for evaluating the treatment efcacy of Ginkgo vs. the
placebo. Six studies (N 1294) employed cognition scales as primary endpoint outcomes
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 513
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
for the assessment of Ginkgo intervention vs. the placebo. SMD in the meta-analysis for
combining continuous data of cognition scales showed Ginkgo had better cognition out-
comes than the placebo control using a random effect model. In total, the SMD between the
cognition scores of Ginkgo and placebo was 1.62 (95% CI: 2.69 to 0.56) shown in
Fig. 4. There was signicant heterogeneity in the results across the studies (test for het-
erogeneity, p < 0:00001, I
2
98%), however every individual study showed a positive
effect of Ginkgo treatment, except for the study of Schneider 2005. 4 studies (N 910)
showed the estimated effects showing favorable effects for the use of Ginkgo on activities
of daily living in AD patients. Overall, SMD of Ginkgo vs. placebo was 1.55 (95% CI:
2.55 to 0.55) and signicant heterogeneity was detected (test for heterogeneity,
p < 0:00001, I
2
96%), which is shown in Fig. 5. Each included study showed the effect-
favoring Ginkgo on activities of daily living in AD patients.
One meta-analysis was performed for assessing the preventative effect of Ginkgo
against AD (N 5884), considering subgroups of participants with different clinical de-
mentia ratings (CDR 0; N 3894 or CDR 0:5; N 1990) at baseline (Fig. 6). The
dichotomous data of AD incidence was analyzed with random-effects and Mantel-Haenszel
methods. In the subgroup of CDR 0 and CDR 0:5 at baseline, the risk ratio (RR) is
respectively 1.09 (95% CI: 0.89 to 1.35) and 1.02 (95% CI: 0.80 to 1.29). Overall, the RR
is 1.06 (95% CI: 0.92 to 1.22) between Gingko and the placebo. There is no signicant
heterogeneity across these two studies (test for heterogeneity: overall p 0:49, I
2
0%;
subgroup p 0:65, I
2
0%). Both trials reported that Ginkgo extract did not reduce the
risk of progression to AD compared with the placebo. Funnel plots were checked but
asymmetry measurement was not applicable since there were fewer than 10 studies in each
meta-analysis, because the test power was too low to distinguish chance from a real effect
(Sterne et al., 2011) (data not shown).
Figure 4. Forest plot of cognition outcomes (ADAS-cog, SKT): Ginkgo vs. placebo.
514 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Discussion
We performed this systematic review to show the natural medications that are involved in
documented RCTs up to date. The quality of literature was evaluated by Jadads score and
Cochrane assessing tool for risk of bias. Most of the clinical studies scored 2 or less
according to Jadads appraisal system, which are considered as low-quality trials. Even
though 21 studies with Jadads score above 3 (inclusive), 13 studies had at least one
domain with a high risk of bias in accordance with Cochrane appraisal criteria, of which 8
studies reported the dropouts due to inefcacy of intervention in high risk of attrition bias
(Le Bars et al., 1997; 2000; 2002; Xu et al., 1999; Mazza et al., 2006; Baum et al., 2008;
Yancheva et al., 2009; Burns et al., 2011), 3 studies were in high risk of reporting bias due
to selective outcome reporting (Xu et al., 1999; Baum et al., 2008; Bi et al., 2011), and 8
studies had a potential source of bias related to the specic study design used (Maurer
et al., 1997; Xu et al., 1999; Akhondzadeh et al., 2003a,b; 2010; Yancheva et al., 2009; Bi
Figure 5. Forest plot of activities of daily living in AD patients: Ginkgo vs. placebo.
Figure 6. Forest plot of incidence of AD: Ginkgo vs. placebo.
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 515
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
et al., 2011; Burns et al., 2011). The risk exists that the true intervention effect can be
overestimated or underestimated in these reports.
S. ofcinalis (Akhondzadeh et al., 2003a), M. ofninalis (Akhondzadeh et al., 2003b),
C. sativus (Akhondzadeh et al., 2010), curcumin (Baum et al., 2008), Fuzhisan (Bi et al.,
2011) were reported in RCTs for patients with AD, however of which the scientic basis
for clinical recommendation is questionable because minimal benets were observed and
the methodological quality of the available trials was poor due to a small sample size,
poorly dened participants, improper analysis of the impact of withdrawals, selective
reporting, etc. Burns et al. (2011) reported that the aromatherapy of M. ofninalis is not
superior to the placebo or donepezil in the treatment of agitation in people with AD.
Huperzine A is a natural cholinesterase inhibitor-derived from the Chinese herb Hperzia
serrata that may compare favorably in systematic efcacy to cholinesterase inhibitors
currently in use for AD. However, there is no demonstrable cognitive effect in patients with
mild to moderate AD (Xu et al., 1999; Rai et al., 2011). Of note, Rai et al. (2011) and
Xu et al. (1999) suggested that huperzine A at a high dose (400 g BID) showed limited
signs of cognitive enhancement as measured by the ADAS-Cog and MMSE, which in-
dicating that huperzine A may have a possible short-term symptomatic benet, but this
requires conrmation in additional studies. Apart from the RCT-related natural medica-
tions, Panax ginseng has been considered to have potential protective and neurotrophic
effects on the cognitive function for a long time. But no available RCTs with Jadad score
3 were documented for ginseng until now. Prospective open-label studies indicated that
ginseng treatment was found to be safe and feasible as a cognitive enhancer in AD patients
(Heo et al., 2008; 2011; Lee et al., 2008). Further double-blinded RCTs with larger
samples of patients and longer treatment duration are expected to conrm the efcacy of
ginseng intervention for AD.
The meta-analyses of pooling results from separate studies were performed on several
outcomes related to the treatment or prevention of AD patients with Ginkgo. The pooled
data was analyzed using random effects model to avoid dissimilar treatment effect across
studies. Heterogeneity determines the validity of the conclusions of meta-analyses, as the
level of heterogeneity increases the justication for an integrated result becomes more
difcult (Walker et al., 2008). Figures 4 and 5 show the forest plots for the outcome of
cognition and activities of daily living in the treatment of AD. In Fig. 4, the estimated
effects are on the same side of the unit line except for the study of Schneider et al.
(2005). The condence intervals of other studies overlap to an acceptable extent. These
studies reported a benet of Ginkgo but Schneiders study suggested that there was no
statistical signicance between the placebo and Ginkgo. Schneiders trial did not show
the efcacy of Ginkgo, however, the lack of decline of the placebo patients may have
compromised the sensitivity of the trial to detect a treatment effect. Thus, the study
remains inconclusive with respect to the efcacy of Ginkgo (Schneider et al., 2005).
Thus the signicant heterogeneity in Fig. 4 was caused largely due to Scheniders study.
A similar result of meta-analysis on the outcome of cognition was found in the subgroup
of patients with AD by Weinmann et al. (2010), which also showed a statistically
signicant advantage of Ginkgo compared to the placebo, despite a huge heterogeneity
516 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
of the clinical trials. In Fig. 5, the condence intervals of all studies are located on the
same side in favor of Ginkgo. Even though the test for heterogeneity is signicant, the
tendencies favoring Ginkgo on the outcome of activities of daily living in the studies
achieve consistency. It is only a matter of dissimilarity in the favoring degree of indi-
vidual results. There are limited systematic reviews published before in assessing the
efcacy of natural medicines on AD. Heterogeneity in types of natural medicines, di-
agnostic criteria, and outcome measures hindered comprehensive data analysis for nat-
ural medicines (Dos Santos-Neto et al., 2006; Man et al., 2008). This analysis is
consistent with the results of another meta-analysis, which draws the conclusion that
Ginkgo has a benecial effect on the outcome of activities of daily living, but the
potential effect size cannot be conrmed because of signicant heterogeneity across
included RCTs (Janssen et al., 2010). Further justication for the efcacy and dosage
requires more rigorous evidence with appropriate sample sizes, clinical settings, dosage
arms and experimental durations.
There are RCTs, which were reported for mixed dementia and not included in this
review specic for AD. However, the ndings of RCTs for mixed dementia containing AD
provide supporting evidence for better-off outcomes associated with AD. Herrschaft et al.
(2012) reported that the treatment with Ginkgo extract resulted in a signicant and clini-
cally relevant improvement in cognition, psychopathology, functional measures and
quality of life in 410 outpatients with mild to moderate dementia in a 24-week trial.
Scripnikov et al. (2007) reported that Ginkgo improved the neuropsychiatric symptoms
(apathy/indifference, anxiety, irritability/lability, depression/dysphoria and sleep/nighttime
behavior) of 400 dementia patients in a 22-week trial.
To indicate the underlying mechanism, two studies directly compared a cholinesterase
inhibitor donepezil with Ginkgo in the treatment of patients with mild to moderate AD in
an explorative manner (Mazza et al., 2006; Yancheva et al., 2009). No signicant dif-
ference was found between Ginkgo and donepezil, suggesting that the efcacy of Ginkgo
may be explained by its modulatory inuence on the cholinergic system. The Ginkgo
extract, such as EGb761, used in RCTs, mainly contain avone glycosides and terpene
lactones including ginkgolides and bilobalide. Many lines of pre-clinical experiments in-
dicate the potential mechanisms. Ginkgo may exert effects on AD due to either a free
radical scavenging activity or the inhibition of pro-inammatory pathways or the poten-
tiation of the cell stress response (Mancuso et al., 2012). EGb761 has been shown in
several preclinical reports to increase nearly all aspects of impaired neuroplasticity (long-
term potentiation, spine density, neuritogenesis, neurogenesis). While all three fractions of
constituents (ginkgolides, avonoids, bilobalide) seem to be active, the avonoids and
specically the aglycone isorhamnetin seem to be most relevant (Muller et al., 2012).
Ginkgo extract was suggested to have long-term regulatory effects on mitochondria.
Ginkgo has a selective effect on the activities of mitochondrial enzymes that assemble the
electron transport system. The avonoids, bilobalide and some of the ginkgolides had a
high protective capacity, indicating that a combination of several compounds within
standardized Gingko extracts contribute disproportionately for these protective effects
(Eckert, 2012). Rendeiro et al. (2012) suggest that avonoids have a potential in
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 517
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
modulating neuronal function and thereby inuencing memory, learning and cognitive
function. Absorbed avonoids and their metabolites may interact with and modulate
critical signaling pathways, transcription factors and gene and/or protein expression, which
control memory and learning processes in the hippocampus.
Figure 6 represents the meta-analysis for assessing the prevention effect of Ginkgo
against AD, in which the RRs developing AD of these two included studies are on
the similar location (RR value is 1) of the unit line and condence intervals overlap to
a large extent. It suggests that there is no signicant effect between Ginkgo and
placebo when preventing AD occurrence. The included two studies for the effect esti-
mate of AD prevention were regarded as high-quality RCTs, between which there is no
signicant heterogeneity in a statistical test for heterogeneity. There is no link between
Ginkgo intervention and subsequent AD incidence, and the risks of AD in the Ginkgo
group compared with the placebo group were not proportional with time. Ginkgo cannot
be recommended for the purpose of preventing AD judging from the existing evidence.
The clinical trial performed by Snitz et al. (2009) reported that, compared with the
placebo, Ginkgo did not result in less cognitive decline in older adults with normal
cognition or with mild cognitive impairment. Dodge et al. (2008) reported that Ginkgo
neither altered the risk of progression from normal to Clinical Dementia Rating 0.5, nor
protected against a decline in memory function. These might imply that Ginkgo cannot
prevent AD progression because Ginkgo fails to hinder the cognitive decline.
It is seemingly contradictory that Ginkgo cannot decrease AD incidence and prevent
AD development yet has a benecial effect on the treatment outcomes of cognition and
activities of daily life. The two trials (DeKosky et al., 2008; Vellas et al., 2012) included in
the prevention effect are the most powerful trials and methodologically the most robust
studies on the use of Gingko in AD patients so far. Five trials (Le Bars et al., 1997;
Kanowski and Hoerr, 2003; Mazza et al., 2006; Napryeyenko et al., 2009; Ihl et al., 2012)
involved in the treatment analysis on cognition outcome with positive results are small
trials except for one neutral result (Schneider et al., 2005), thus we cannot exclude the
possibility of publication bias. Considering the small trials were included and the existence
of heterogeneity across studies, the meta-analyses for treatment outcomes of Gingko
compared to placebo should be treated with caution for readers. However, since Ginkgo
was reported to have the potential treatment effect on the outcomes of AD in most available
RCTs, we speculated that Ginkgo may help established AD patients with cognitive
symptoms but cannot prevent AD from neurodegenerative progression.
One weakness of this review is that unpublished data was not specically included,
thus, despite a comprehensive search strategy employed to identify published studies, there
could be unpublished studies we did not identify. And this review includes only English
literature, which may not represent all of the evidence and can induce a language bias. Due
to the complicated and inconsistent nature of the documented RCTs, the other potential
limitation of this review is that conducting a meta-analysis does not overcome problems
that were inherent in the design and execution of the primary studies. We assessed the bias
of studies but this cannot be corrected in the meta-analysis for statistical results. And the
scope of categories of natural medicines and AD outcome measures are open to future
518 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
modications. More high-quality RCTs with rigorous methodology in exploring natural
medicines are desirable to achieve a solid conrmation for clinical decision-making.
Acknowledgments
M.Y. and D.D.X. contributed equally to the whole process including database searching,
data analysis and interpretation and paper preparation. W.C.C. carried out design of sys-
tematic review, scrutiny of the research methodology and preparation of this paper. Y.Z.
and X.L. were involved in RCT appraisal, approval of statistics. R.H. participated in
approval of RCT appraisal and critical revision of this paper. Financial support was pro-
vided by Tangdu Hospital of the Fourth Military Medical University, China.
References
Akhondzadeh, S., M. Noroozian, M. Mohammadi, S. Ohadinia, A.H. Jamshidi and M. Khani. Salvia
ofcinalis extract in the treatment of patients with mild to moderate Alzheimers disease: a
double blind, randomized and placebo-controlled trial. J. Clin. Pharm. Ther. 28: 5359, 2003a.
Akhondzadeh, S., M. Noroozian, M. Mohammadi, S. Ohadinia, A.H. Jamshidi and M. Khani.
Melissa ofcinalis extract in the treatment of patients with mild to moderate Alzheimers
disease: a double blind, randomised, placebo controlled trial. J. Neurol. Neurosurg. Psychiatr.
74: 863866, 2003b.
Akhondzadeh, S., M.S. Sabet, M.H. Harirchian, M. Togha, H. Cheraghmakani, S. Razeghi, S. Hejazi,
M.H. Youse, R. Alimardani, A. Jamshidi, F. Zare and A. Moradi. Saffron in the treatment of
patients with mild to moderate Alzheimers disease: a 16-week, randomized and placebo-
controlled trial. J. Clin. Pharm. Ther. 35: 581588, 2010.
Baum, L., C.W. Lam, S.K. Cheung, T. Kwok, V. Lui, J. Tsoh, L. Lam, V. Leung, E. Hui, C. Ng, J.
Woo, H.F. Chiu, W.B. Goggins, B.C. Zee, K.F. Cheng, C.Y. Fong, A. Wong, H. Mok, M.S.
Chow, P.C. Ho, S.P. Ip, C.S. Ho, X.W. Yu, C.Y. Lai, M.H. Chan, S. Szeto, I.H. Chan and V.
Mok. Six-month randomized, placebo-controlled, double-blind, pilot clinical trial of curcumin
in patients with Alzheimer disease. J. Clin. Psychopharmacol. 28: 110113, 2008.
Bi, M., S. Tong, Z. Zhang, Q. Ma, S. Zhang, Z. Luo, Y. Zhang, X. Li and D. Wang. Changes in
cerebral glucose metabolism in patients with mild-to-moderate Alzheimers disease: a pilot
study with the Chinese herbal medicine Fuzhisan. Neurosci. Lett. 501: 3540, 2011.
Burns, A., E. Perry, C. Holmes, P. Francis, J. Morris, M.J. Howes, P. Chazot, G. Lees and C. Ballard.
A double-blind placebo-controlled randomized trial of Melissa ofcinalis oil and donepezil for
the treatment of agitation in Alzheimers disease. Dement. Geriatr. Cogn. Disord. 31: 158
164, 2011.
DeKosky, S.T., J.D. Williamson, A.L. Fitzpatrick, R.A. Kronmal, D.G. Ives, J.A. Saxton, O.L.
Lopez, G. Burke, M.C. Carlson, L.P. Fried, L.H. Kuller, J.A. Robbins, R.P. Tracy, N.F.
Woolard, L. Dunn, B.E. Snitz, R.L. Nahin and C.D. Furberg. Ginkgo biloba for prevention of
dementia: a randomized controlled trial. JAMA 300: 19: 22532262, 2008.
Dodge, H.H., T. Zitzelberger, B.S. Oken, D. Howieson and J. Kaye. A randomized placebo-con-
trolled trial of Ginkgo biloba for the prevention of cognitive decline. Neurology 70: 1809
1817, 2008.
Dos Santos-Neto, L.L., M.A. de Vilhena Toledo, P. Medeiros-Souza and G.A. de Souza. The use of
herbal medicine in Alzheimers disease a systematic review. Evid. Based Complement.
Alternat. Med. 3: 441445, 2006.
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 519
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Eckert, A. Mitochondrial effects of Ginkgo biloba extract. Int. Psychogeriatr. 24(Suppl 1): S18S20,
2012.
Heo, J.H., S.T. Lee, K. Chu, M.J. Oh, H.J. Park, J.Y. Shim and M. Kim. An open-label trial of
Korean red ginseng as an adjuvant treatment for cognitive impairment in patients with Alz-
heimers disease. Eur. J. Neurol. 15: 865868, 2008.
Heo, J.H., S.T. Lee, M.J. Oh, H.J. Park, J.Y. Shim, K. Chu and M. Kim. Improvement of cognitive
decit in Alzheimers disease patients by long term treatment with Korean red ginseng. J.
Ginseng Res. 35: 457461, 2011.
Herrschaft, H., A. Nacu, S. Likhachev, I. Sholomov, R. Hoerr and S. Schlaefke. Ginkgo biloba
extract EGb 761(R) in dementia with neuropsychiatric features: a randomised, placebo-con-
trolled trial to conrm the efcacy and safety of a daily dose of 240 mg. J. Psychiatr. Res. 46:
716723, 2012.
Ihl, R., M. Tribanek and N. Bachinskaya. Efcacy and tolerability of a once daily formulation of
Ginkgo biloba extract EGb 761(R) in Alzheimers disease and vascular dementia: results from
a randomised controlled trial. Pharmacopsychiatry 45: 4146, 2012.
Jadad, A.R., R.A. Moore, D. Carroll, C. Jenkinson, D.J.M. Reynolds, D.J. Gavaghan and H.J.
McQuay. Assessing the quality of reports of randomized clinical trials: is blinding necessary?
Control. Clin. Trials 17: 112, 1996.
Janssen, I.M., S. Sturtz, G. Skipka, A. Zentner, M. Velasco Garrido and R. Busse. Ginkgo biloba in
Alzheimers disease: a systematic review. Wien. Med. Wochenschr. 160: 539546, 2010.
Kanowski, S., W.M. Herrmann, K. Stephan, W. Wierich and R. Horr. Proof of efcacy of the Ginkgo
biloba special extract EGb 761 in outpatients suffering from mild to moderate primary degen-
erative dementia of the Alzheimer type or multi-infarct dementia. Phytomedicine 4: 313, 1997.
Kanowski, S. and R. Hoerr. Ginkgo biloba extract EGb 761 in dementia: intent-to-treat analyses of a
24-week, multi-center, double-blind, placebo-controlled, randomized trial. Pharma-
copsychiatry 36: 297303, 2003.
Le Bars, P.L., M.M. Katz, N. Berman, T.M. Itil, A.M. Freedman and A.F. Schatzberg. A placebo-
controlled, double-blind, randomized trial of an extract of Ginkgo biloba for dementia. JAMA
278: 13271332, 1997.
Le Bars, P.L., M. Kieser and K.Z. Itil. A 26-week analysis of a double-blind, placebo-controlled trial
of the Ginkgo biloba extract EGb 761 in dementia. Dement. Geriatr. Cogn. Disord. 11: 230
237, 2000.
Le Bars, P.L., F.M. Velasco, J.M. Ferguson, E.C. Dessain, M. Kieser and R. Hoerr. Inuence of the
severity of cognitive impairment on the effect of the Ginkgo biloba extract EGb 761 in
Alzheimers disease. Neuropsychobiology 45: 1926, 2002.
Lee, S.T., K. Chu, J.Y. Sim, J.H. Heo and M. Kim. Panax ginseng enhances cognitive performance in
Alzheimer disease. Alzheimer. Dis. Assoc. Disord. 22: 222226, 2008.
Man S.C., S.S. Durairajan, W.F. Kum, J.H. Lu, J.D. Huang, C.F. Cheng, V. Chung, M. Xu and M. Li.
Systematic review on the efcacy and safety of herbal medicines for Alzheimers disease. J.
Alzheimer Dis. 14: 209223, 2008.
Mancuso, C., R. Siciliano, E. Barone and P. Preziosi. Natural substances and Alzheimers disease:
from preclinical studies to evidence based medicine. Biochim. Biophys. Acta 1822: 616624,
2012.
Maurer, K., R. Ihl, T. Dierks and L. Frlich. Clinical efcacy of Ginkgo biloba special extract EGb
761 in dementia of the Alzheimer type. J. Psychiatr. Res. 31: 645655, 1997.
Mazza, M., A. Capuano, P. Bria and S. Mazza. Ginkgo biloba and donepezil: a comparison in the
treatment of Alzheimers dementia in a randomized placebo-controlled double-blind study.
Eur. J. Neurol. 13: 981985, 2006.
Muller, W.E., J. Heiser and K. Leuner. Effects of the standardized Ginkgo biloba extract EGb 761(R)
on neuroplasticity. Int. Psychogeriatr. 24(Suppl 1): S21S24, 2012.
520 M. YANG et al.
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.
Napryeyenko, O., G. Sonnik and I. Tartakovsky. Efcacy and tolerability of Ginkgo biloba extract
EGb 761 by type of dementia: analyses of a randomised controlled trial. J. Neurol. Sci. 283:
224229, 2009.
Perry, E.K., A.T. Pickering, W.W. Wang, P.J. Houghton and N.S.L. Perry. Medicinal plants and
Alzheimers disease: from ethnobotany to phytotherapy. J. Pharm. Pharmacol. 51: 527534,
1999.
Rai, M.S., S. Walsh, J.T. Little, K. Behan, B. Reynolds, C. Ward, S. Jin, R. Thomas and P.S. Aisen.
A phase II trial of huperzine A in mild to moderate Alzheimer disease. Neurology 76: 1389
1394, 2011.
Rendeiro, C., J.D. Guerreiro, C.M. Williams and J.P. Spencer. Flavonoids as modulators of memory
and learning: molecular interactions resulting in behavioural effects. Proc. Nutr. Soc. 71: 246
262, 2012.
Schneider, L.S., S.T. DeKosky, M.R. Farlow, P.N. Tariot, R. Hoerr and M. Kieser. A randomized,
double-blind, placebo-controlled trial of two doses of Ginkgo biloba extract in dementia of the
Alzheimers type. Curr. Alzheimer. Res. 2: 541551, 2005.
Scripnikov, A., A. Khomenko and O. Napryeyenko. Effects of Ginkgo biloba extract EGb 761 on
neuropsychiatric symptoms of dementia: ndings from a randomised controlled trial. Wien.
Med. Wochenschr. 157: 295300, 2007.
Selkoe, D.J. Alzheimers disease: Genes, proteins, and therapy. Physiol. Rev. 81: 741766, 2001.
Snitz, B.E., E.S. OMeara, M.C. Carlson, A.M. Arnold, D.G. Ives, S.R. Rapp, J. Saxton, O.L. Lopez,
L.O. Dunn, K.M. Sink and S.T. DeKosky. Ginkgo biloba for preventing cognitive decline in
older adults: a randomized trial. JAMA 302: 26632670, 2009.
Sterne, J.A.C., A.J. Sutton, J.P.A. Ioannidis, N. Terrin, D.R. Jones, J. Lau, J. Carpenter, G. Rcker,
R.M. Harbord, C.H. Schmid, J. Tetzlaff, J.J. Deeks, J. Peters, P. Macaskill, G. Schwarzer, S.
Duval, D.G. Altman, D. Moher and J.P.T. Higgins. Recommendations for examining and
interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. BMJ 343:
d4002, 2011.
Vellas, B., N. Coley, P.J. Ousset, G. Berrut, J.F. Dartigues, B. Dubois, H. Grandjean, F. Pasquier, F.
Piette, P. Robert, J. Touchon, P. Garnier, H. Mathiex-Fortunet and S. Andrieu. Long-term use
of standardised Ginkgo biloba extract for the prevention of Alzheimers disease (GuidAge): a
randomised placebo-controlled trial. Lancet Neurol. 11: 851859, 2012.
Walker E., A.V. Hernandez and M.W. Kattan. Meta-analysis: its strengths and limitations. Cleve.
Clin. J. Med. 75: 431439, 2008.
Weinmann, S., S. Roll, C. Schwarzbach, C. Vauth and S.N. Willich. Effects of Ginkgo biloba in
dementia: systematic review and meta-analysis. BMC Geriatr. 10: 14, 2010.
Xu, S.S., Z.Y. Cai, Z.W. Qu, R.M. Yang, Y.L. Cai, G.Q. Wang, X.Q. Su, X.S. Zhong, R.Y. Cheng,
W.A. Xu, J.X. Li and B. Feng. Huperzine-A in capsules and tablets for treating patients with
Alzheimer disease. Zhongguo Yao Li Xue Bao 20: 6: 486490, 1999.
Yancheva, S., R. Ihl, G. Nikolova, P. Panayotov, S. Schlaefke and R. Hoerr. Ginkgo biloba extract
EGb 761(R), donepezil or both combined in the treatment of Alzheimers disease with neu-
ropsychiatric features: a randomised, double-blind, exploratory trial. Aging Ment. Health 13:
183190, 2009.
META-ANALYSES OF GINKGO ON ALZHEIMERS DISEASE 521
A
m
.

J
.

C
h
i
n
.

M
e
d
.

2
0
1
4
.
4
2
:
5
0
5
-
5
2
1
.

D
o
w
n
l
o
a
d
e
d

f
r
o
m

w
w
w
.
w
o
r
l
d
s
c
i
e
n
t
i
f
i
c
.
c
o
m
b
y

F
U
D
A
N

U
N
I
V
E
R
S
I
T
Y

o
n

0
6
/
0
6
/
1
4
.

F
o
r

p
e
r
s
o
n
a
l

u
s
e

o
n
l
y
.