Download - Trabalho Parto Acupuntura
Acupuncture for pain relief in labour: asystematic review and meta-analysisS-H Cho,a H Lee,b E Ernstc
a Hospital of Korean Medicine, Kyung Hee University Medical Center, Kyung Hee University, Seoul, Korea b Acupuncture and Meridian
Science Research Center (AMSRC), College of Korean Medicine, Kyung Hee University, Seoul, Korea c Complementary Medicine, Peninsula
Medical School, Universities of Exeter and Plymouth, Exeter, UK
Correspondence: Dr H Lee, Acupuncture and Meridian Science Research Center (AMSRC), College of Korean Medicine, Kyung Hee University,
Hoegi-dong, Dongdaemoon-gu, 130-701 Seoul, Korea. Email [email protected]
Accepted 12 March 2010. Published Online 28 April 2010.
Background Acupuncture is frequently used for pain relief in
labour, but the evidence is not clear.
Objectives To critically evaluate the evidence for or against
acupuncture for labour pain management.
Search strategy Nineteen electronic databases, including English,
Korean, Japanese, and Chinese databases, were systematically
searched.
Selection criteria All randomised controlled trials (RCTs)
involving women receiving acupuncture alone, or as an adjunct to
conventional analgesia, for pain relief in labour were considered.
Data collection and analysis Pain intensity on a 100-mm visual
analogue scale (VAS; 0, no pain; 100, worst pain) and use of
other analgesic methods were used as primary outcomes, and for
statistical pooling. Maternal/fetal outcomes were secondary
outcomes, and adverse events were also recorded. Risk of bias was
assessed regarding randomisation, allocation concealment,
blinding, incomplete outcome data, selective outcome reporting,
and other biases.
Main results Ten RCTs involving 2038 women were included.
VAS for pain intensity data were available in seven studies; the
meta-analysis shows that acupuncture was not superior to
minimal acupuncture at 1 hour (pooled mean difference )8.02;
95% CI )21.88, 5.84; I2 = 94%) and at 2 hours ()10.15; 95% CI
)23.18, 2.87; I2 = 92%). Patients reported significantly reduced
pain by 4 and 6% during electroacupuncture (EA) treatment at 15
()4.09; 95% CI )8.05, )0.12) and 30 minutes ()5.94; 95% CI
)9.83, )2.06), compared with placebo EA, but the effect was not
maintained afterwards. Compared with no intervention,
acupuncture reduced pain by only 11% for the first 30 minutes
()10.56; 95% CI )16.08, )5.03). In trials where acupuncture was
compared with conventional analgesia, women receiving
acupuncture required less meperidine (pooled risk ratio 0.20;
95% CI 0.12, 0.33) and other analgesic methods (0.75; 95% CI
0.66, 0.85). No acupuncture-related adverse events were reported.
Most trials did not blind participants, care providers and/or
evaluators.
Author’s conclusions The evidence from RCTs does not support
the use of acupuncture for controlling labour pain. The primary
studies are diverse and often flawed. Further research seems
warranted.
Keywords Acupuncture, labour pain, randomised controlled trial,
systematic review.
Please cite this paper as: Cho S-H, Lee H, Ernst E. Acupuncture for pain relief in labour: a systematic review and meta-analysis. BJOG 2010;117:907–920.
Introduction
Pain relief is an important issue for women in labour.
Whatever method is used to reduce maternal discomfort, it
should be both effective and safe for the mother and child.
Several methods of relieving pain in labour and various
coping strategies have been advocated. They range from
intervention such as breathing exercises to medical tech-
niques like epidural analgesia (EDA). Several complemen-
tary or alternative methods have also been reported to
reduce pain during labour and delivery.1 Cochrane reviews
have demonstrated the value of continuous support and
non-pharmacological approaches to managing pain in
labour.2–4 A recent survey of certified nurse-midwives sug-
gested that the use of complementary and alternative medi-
cine (CAM) by pregnant women is common. More than
90% of midwives reported recommending CAM for their
pregnant women.5
Acupuncture and related techniques are increasingly
practiced in obstetrics and gynaecology for labour pain,
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 907
DOI: 10.1111/j.1471-0528.2010.02570.x
www.bjog.orgSystematic review
intraoperative analgesia, dysmenorrhea, pelvic pain in preg-
nancy, and pain during oocyte retrieval.6–8 The use of acu-
puncture to manage labour pain was introduced in the
1970s.9–11 Since then, acupuncture has been employed to
relieve pain and facilitate the process of labour. A previous
systematic review of three trials of acupuncture for labour
pain suggested that acupuncture might play a role in pain
control, but because of the paucity of trial data, the conclu-
sions remained uncertain.6 Several new studies have since
been published, including two placebo-controlled trials and
two sham-controlled trials. It is therefore timely to recon-
sider whether acupuncture reduces labour pain. This sys-
tematic review and meta-analysis are aimed at critically
evaluating the evidence for or against acupuncture as a
treatment of pain during labour.
Methods
Search strategyThe following sources were searched all from their incep-
tion to April 2009: The Cochrane Central Register of Con-
trolled Trials, MEDLINE, EMBASE, Allied and
Complementary Medicine Database (AMED), Cumulative
Index to Nursing and Allied Health Literature (CINAHL),
PsycInfo, Korean medical databases (the National Assembly
Library, KoreaMed, Korean Studies Information Service
System, DBpia, and Korea Institute of Science Technology
Information and Research Information Service System), a
Japanese database (Japan Science and Technology Informa-
tion Aggregator Electronic), and Chinese databases (the
China Academic Journal, Century Journal Project, China
Doctor/Master Dissertation Full Text Database, and China
Proceedings Conference Full Text Database). We also
searched current controlled trials (http://www.controlled-
trials.com), National Center for Complementary and Alter-
native Medicine at the National Institutes of Health (http://
nccam.nih.gov), and the Complementary and Alternative
Medicine Specialist Library at the National Health Service
National Library for Health (http://www.library.nhs.uk/
cam). The reference lists of articles were checked for fur-
ther relevant publications, and experts were asked for infor-
mation concerning any additional trials. A further manual
search was conducted for relevant journals, symposia, and
conference proceedings; all identified publications were
cross-referenced. Personal contact was made with the
authors of the published studies, if necessary, to request for
additional data.
The search terms used were (‘acupuncture’ OR ‘electro-
acupuncture’ OR ‘meridian’ OR ‘acupoint’ OR ‘scapular
acupuncture’ OR ‘auricular acupuncture’) AND (‘labour’
OR ‘childbirth’ OR ‘delivery’ OR ‘pregnancy’ OR ‘obstetric’
OR ‘parturi’ OR ‘labouring’). As all of the various databases
searched for this review possessed their own subject
headings, each database was searched independently. No
language restrictions were imposed.
Study selection
Types of studiesRandomised controlled trials (RCTs); quasi-randomised or
non-randomised trials were excluded.
Types of participantsWomen in labour; trials of induced labour were excluded.
Types of interventionTrials testing acupuncture as a sole analgesic method or as
an adjunct to conventional analgesia were considered.
Studies of classical acupuncture, electroacupuncture (EA),
and (electrical) auricular acupuncture were included. Trials
of acupuncture-related techniques (for example, auricular
seed, laser, injection, acupoint-embedding, acupressure,
magnetic device, or moxibustion) were excluded. Studies
that assessed the combined effect of acupuncture with
other related therapies (for example, acupuncture plus
moxibustion therapy) were excluded. Trials that compared
different forms of acupuncture were also excluded.
Types of controlAny form of pain relief, no pain relief, and placebo or
sham acupuncture, such as superficial needling and non-
penetrating sham needles, were included.
Types of outcome measuresOur primary outcome measures were pain relief in labour
and use of other analgesia. Pain intensity outcomes
included a 100-mm visual analogue scale (VAS) or numeric
rating scale, and number of women requiring meperidine,
EDA, or other pain relief. Trials reporting arbitrary or
undetermined grades for VAS scores of pain relief were
excluded to minimise selective outcome reporting. Second-
ary outcomes were maternal and fetal; maternal outcomes
included satisfaction with pain relief in labour, satisfaction
with childbirth experience, duration of labour, augmenta-
tion of labour, and mode of delivery. Fetal outcome mea-
sures were Apgar score at 5 minutes, birthweight, and
umbilical cord blood pH, and adverse events were
recorded.
Data extraction and quality assessmentTwo reviewers (S-HC and HL) independently evaluated
articles for inclusion, followed by data extraction regarding
author, year of publication, country, participants, acupunc-
ture treatment, control types, and primary and secondary
outcomes, using a pre-defined data extraction form. In the
case of duplicate publications and companion papers of a
Cho et al.
908 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
primary study, the yield of information was maximised by
simultaneously evaluating all available data. When in doubt
(for example, clinical trials performed in the same hospital
during the same period), the original publication (usually
the oldest version) was given priority.
Quality assessment was also independently performed by
the two reviewers (S-HC and HL) using the assessment tool
for ‘risk of bias’ from the Cochrane Handbook for System-
atic Reviews of Interventions.12 The following characteristics
were assessed. (1) Was the allocation sequence adequately
generated? (2) Was allocation adequately concealed? (3)
Was knowledge of the allocated interventions adequately
prevented during the study? (4) Were incomplete outcome
data adequately addressed? (5) Are reports of the study free
of suggestion of selective outcome reporting? (6) Was the
study free of other problems that could put it at a risk of
bias? This review used Y, U, and N to label judgments: the
answer ‘yes’ indicated a low risk of bias (Y); ‘unclear’ indi-
cated that the risk of bias is uncertain (U), and the answer
‘no’ indicated a high risk of bias (N). Prevention of knowl-
edge of the allocated interventions, i.e. blinding (patient,
care provider, and evaluator blinding), was assessed sepa-
rately. Disagreements were resolved by discussion between
all authors, and further information was sought from the
primary authors if necessary.
Statistical analysisStatistical analysis was carried out using Review Manag-
er v5.0.21 for Windows (The Nordic Cochrane Centre,
Copenhagen, Denmark). Studies were classified and com-
bined in the analysis according to their control types. We
used the mean difference (MD) for VAS for pain intensity,
and for dichotomous data, e.g. number of women requir-
ing meperidine, results were presented as risk ratios (RRs)
with 95% confidence intervals (CIs), using either a fixed or
random-effects model depending on the statistical hetero-
geneity between studies, ascertained by the I2 test. Studies
were combined to assess the net effect of all types of acu-
puncture treatments on labour pain only if statistical heter-
ogeneity was not evident. Heterogeneity was examined with
the I2 test, where I2 values of 50% or more were considered
to be indicators of a substantial level of heterogeneity.13
Results
Study characteristicsFigure 1 shows the flow of our search results.14 Ten studies
met our inclusion criteria and characteristics of these stud-
ies are presented in Table 1. We obtained further necessary
information on one study from the original author.15
Participants and settingsA total of 2038 participants were involved, ranging from 90
to 607 per study. Four studies recruited only nulliparous
women;16–19 five studies had both nulliparous and multipa-
rous women;15,20–23 and parity was not mentioned in one
study.24 The majority of the studies included women at
more than 37 weeks of gestation. All studies were con-
ducted in hospital settings: five in Europe,15,20–23 three in
China,16–18 and two in Iran.19,24
Potentially relevant articles reviewed and retrieved for more detailed evaluation (n = 92)
Excluded (n = 80)• Case studies (n = 14)• No analgesic outcomes (n = 12)• Reviews/surveys (n = 25)• Inadequate data reporting (n = 17)• Acupressure trials (n = 11)• Comparing different forms of acupuncture (n = 1)
Further evaluation regarding randomization (n =12)
Excluded (n = 2)• Not a randomized trial (n = 2)
RCTs included in the systematic review (n = 10)
Figure 1. Flowchart of search.
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 909
Tab
le1.
Char
acte
rist
ics
of
the
incl
uded
studie
s
Firs
tau
tho
r
(year)
Part
icip
an
tsIn
terv
en
tio
n
(n)*
Typ
eo
f
con
tro
l(n
)*
Pri
mary
ou
tco
mes
Seco
nd
ary
ou
tco
mes
Ad
vers
e
even
ts
Mate
rnal
Feta
l
Boru
p(2
009)1
9607
wom
enw
ith
single
ton,
vert
ex
pre
senta
tion,
term
(37–4
2w
eeks
of
ges
tation)
AT
+A
Aan
d
acce
ss
toco
nve
ntional
care
(n=
314)
(a)
TEN
San
dac
cess
toco
nve
ntional
care
(n=
144)
(b)
Conve
ntional
anal
ges
ia(n
=149)
Nee
dfo
ran
yphar
mac
olo
gic
al
and/o
rin
vasi
vepai
nre
lief;
AT
bet
ter
than
conve
ntional
anal
ges
ia(5
9vs
83%
,
P<
0.0
01),
and
TEN
S
(59
vs69%
,P
=0.0
31)
NS
innee
dfo
rm
eper
idin
e
and
nee
dfo
rED
A
NS
indura
tion
of
labour,
use
of
oxy
toci
n,
mode
of
del
iver
y,post
par
tum
hae
morr
hag
e,an
d
birth
exper
ience
NS
inA
pgar
score
at1
min
ute
Signifi
cant
diffe
rence
in
cord
blo
od
pH
valu
e(P
=0.0
39),
and
med
ian
Apgar
score
at5
min
ute
s
(P=
0.0
4)
None
Han
toush
zadeh
(2007)1
8150
nulli
par
ous
wom
enw
ith
single
ton,
term
(>37
wee
ksof
ges
tation)
inac
tive
phas
e(c
ervi
caldila
tation
‡4
cman
d/o
r3
contr
actions
‡40
seco
nds/
10
min
ute
s)
AT
(n=
70)
Min
imal
AT
at
non-a
cupoin
ts
(n=
74);
no
de-
qi*
*
sensa
tion
sought;
nee
dle
sre
move
d
ifnec
essa
ry
VA
Sfo
rpai
n;
AT
bet
ter
than
contr
olat
the
beg
innin
g
(54
vs38,
P<
0.0
01)
and
afte
r2
hours
(57
vs70,
P<
0.0
01)
NS
inV
AS
for
pai
nat
30
min
ute
s,
and
1hour
AT
bet
ter
than
min
imal
AT
indura
tion
of
labour
(3.4
vs4.5
hours
,
P<
0.0
01)
and
use
of
oxy
toci
nunits
(5.6
vs
7.8
,P
=0.0
01)
AT
bet
ter
than
min
imal
AT
inpat
ients
’sa
tisf
action
(P<
0.0
01),
and
will
ingnes
sto
rece
ive
AT
agai
n(P
<0.0
01)
NS
inbirth
wei
ght,
and
Apgar
score
at1
min
ute
None
Huan
g(2
008)1
7324
nulli
par
ous
wom
en
with
single
ton,
vert
ex
pre
senta
tion
EA(n
=82)
(a)
TEN
S(n
=82);
2/1
00
Hz
(b)
EDA
(n=
79)
(c)
No
trea
tmen
t
(n=
81);
bre
athin
g
and
mas
sage*
**
VA
Sfo
rpai
n;
(a)
NS;
(b)
EDA
bet
ter
than
AT
at60
min
ute
s
(P=
0.0
06),
and
120
min
ute
s(P
<0.0
0001);
(c)
AT
bet
ter
than
no
trea
tmen
tat
30,
60,
and
120
min
ute
s(P
<0.0
0001)
Satisf
action
with
pai
nre
lief;
(a)
AT
bet
ter
than
TEN
S(P
=0.0
1);
(b)
EDA
bet
ter
than
AT
(P=
0.0
04);
(c)
NR
NS
inm
ode
of
del
iver
y
NR
NR
Mar
tenss
on
(2008)2
0128
wom
en,
term
(37–4
2w
eeks
of
ges
tation),
sponta
neo
us
phas
e(‡
3co
ntr
actions
/10
min
ute
s)
AT
(n=
62)
Subcu
taneo
us
ster
ile
wat
erin
ject
ion
(n=
66);
0.5
mlin
area
of
pai
n,
repea
ted
if
nec
essa
ry
VA
Sfo
rm
ean
pai
n;
contr
ol
bet
ter
than
AT
(57
vs76,
P<
0.0
01)
VA
Sfo
rm
ean
rela
xation;
contr
olbet
ter
than
AT
(57
vs69,
P<
0.0
03)
NS
inre
quirem
ent
for
additio
nal
pai
nre
lief
met
hods
NS
indura
tion
of
labour,
mode
of
del
iver
y,an
d
will
ingnes
sto
rece
ive
the
sam
ean
alges
ic
met
hod
agai
n
NS
inbirth
wei
ght,
Apgar
score
at
1,
5,
and
10
min
ute
s,
and
blo
od
stat
us
(pH
,BE)
NR
Cho et al.
910 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
Tab
le1.
(Continued
)
Firs
tau
tho
r
(year)
Part
icip
an
tsIn
terv
en
tio
n
(n)*
Typ
eo
f
con
tro
l(n
)*
Pri
mary
ou
tco
mes
Seco
nd
ary
ou
tco
mes
Ad
vers
e
even
ts
Mate
rnal
Feta
l
Nes
hei
m(2
003)2
1198
wom
enw
ith
regula
r
contr
actions
atte
rm
(37–4
2w
eeks
of
ges
tation)
AT
and
acce
ss
toco
nve
ntional
care
(n=
109)
Conve
ntional
anal
ges
ia(n
=92)
No.
of
pat
ients
usi
ng
mep
erid
ine;
AT
bet
ter
than
contr
ol(1
1vs
37%
,
P<
0.0
001)
No.
of
pat
ients
usi
ng
any
anal
ges
ia;
AT
bet
ter
than
contr
ol(6
6vs
82%
,
P<
0.0
5)
NS
indura
tion
of
labour,
and
mode
of
del
iver
y
NS
inbirth
wei
ght,
and
Apgar
score
at1
and
5m
inute
s
None
Ram
ner
o(2
002)2
2100
wom
enw
ith
single
ton,
vert
expre
senta
tion,
term
(‡37
wee
ksof
ges
tation)
in
sponta
neo
us
phas
ean
d
cerv
ical
dila
tation
(£6
cm)
atad
mis
sion
AT
and
acce
ss
toco
nve
ntional
care
(n=
46)
Conve
ntional
anal
ges
ia(n
=44)
NS
inN
RS
for
pai
n,
nee
dfo
r
mep
erid
ine,
and
nee
dfo
roth
er
phar
mac
olo
gic
anal
ges
ia
AT
bet
ter
than
contr
ol;
no.
of
pat
ients
requirin
gED
A
(26
vs50%
,
P<
0.0
5),
TEN
S
(P<
0.0
5),
war
m
rice
bag
(P<
0.0
1),
and
show
er(P
<0.0
5)
NRS
for
rela
xation;
AT
bet
ter
than
contr
ol(P
<0.0
5)
NS
inm
ode
of
del
iver
y,
use
of
oxy
toci
n,
dura
tion
of
labour,
and
ante
par
tum
hae
morr
hag
e
NS
inbirth
wei
ght,
and
Apgar
score
at1
and
5m
inute
s
None
Skiln
and
(2002)2
3210
wom
enw
ith
single
ton,
vert
expre
senta
tion,
term
(37–4
2w
eeks
of
ges
tation)
insp
onta
neo
us
phas
e
AT
(n=
106)
Min
imal
AT
at
non-a
cupoin
tsuntil
conve
rsio
nto
conve
ntional
anal
ges
iaor
del
iver
y
(n=
102)
VA
Sfo
rpai
n;
AT
bet
ter
than
contr
ol(P
<0.0
5)
Nee
dfo
rED
A;
AT
bet
ter
than
contr
ol(1
0vs
26.5
%,
P=
0.0
1)
Nee
dfo
rm
eper
idin
e;A
Tbet
ter
than
contr
ol(1
4vs
35%
,
P<
0.0
01)
Dura
tion
of
labour;
AT
bet
ter
than
contr
ol(P
=0.0
1)
Use
of
oxy
toci
n;
AT
bet
ter
than
contr
ol
(14
vs35%
,P
<0.0
01)
NS
inm
ode
of
del
iver
y
NS
inA
pgar
score
at5
min
ute
s
None
Zhan
g(2
006)1
6120
nulli
par
ous
wom
enw
ith
single
ton,
vert
expre
senta
tion,
term
(37–4
2w
eeks
of
ges
tation)
inre
gula
r
contr
actions
and
cerv
ical
dila
tation
(2–3
cm)
atad
mis
sion
EA(n
=39)
(a)
Plac
ebo
EA,
iden
tica
l
stim
ula
tion
with
non-p
enet
rating
nee
dle
atac
upoin
t
SP6
(n=
40)
(b)
No
trea
tmen
t
(n=
40)
VA
Sfo
rpai
n;
(a)
EAbet
ter
than
pla
cebo
EA(P
<0.0
5)
at15,
30
min
ute
s,an
d3
hours
afte
rEA
;
(b)
EAbet
ter
than
no
trea
tmen
t
(P<
0.0
1)
at15,
30
min
ute
s,
and
1an
d2
hours
afte
rEA
Dura
tion
of
labour;
EAbet
ter
than
pla
cebo
EA(P
=0.0
07)
NS
inA
pgar
score
at1
and
5m
inute
s
bet
wee
nEA
and
pla
cebo
EA
None
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 911
Tab
le1.
(Continued
)
Firs
tau
tho
r
(year)
Part
icip
an
tsIn
terv
en
tio
n
(n)*
Typ
eo
f
con
tro
l(n
)*
Pri
mary
ou
tco
mes
Seco
nd
ary
ou
tco
mes
Ad
vers
e
even
ts
Mate
rnal
Feta
l
Zhou
(2007)1
5111
nulli
par
ous
wom
enw
ith
single
ton,
vert
ex
pre
senta
tion,
term
(37–4
2w
eeks
of
ges
tation)
inre
gula
r
contr
actions
and
cerv
ical
dila
tation
(2–3
cm)
at
adm
issi
on
EA(n
=38)
(a)
Plac
ebo
EA,
iden
tica
l
stim
ula
tion
with
non-p
enet
rating
nee
dle
atac
upoin
t
SP6
(n=
37)
(b)
No
trea
tmen
t
(n=
36)
VA
Sfo
rpai
n;
(a)
NS;
(b)
EAbet
ter
than
no
trea
tmen
t
(P<
0.0
1)
at15,
30
min
ute
s,
and
1,
2,
and
3hours
afte
rEA
NR
NR
NR
Ziae
i(2
006)2
490
wom
enw
ith
single
ton,
vert
expre
senta
tion,
term
(‡37
wee
ksof
ges
tation)
in
sponta
neo
us
phas
e,ce
rvic
al
dila
tation
(3–6
cm)
atad
mis
sion
AT
(n=
30)
(a)
Min
imal
AT
at
non-a
cupoin
ts
(n=
30)
(b)
No
trea
tmen
t
(n=
30)
(a)
NS
inV
AS
for
pai
n;
(b)
NS
inV
AS
for
pai
n
NS
inV
AS
for
rela
xation,
dura
tion
of
labour,
and
rate
of
caes
area
ndel
iver
y
Use
of
oxy
toci
n;
AT
bet
ter
than
contr
ol
(a)
(50
vs77%
,
P=
0.0
4),
and
contr
ol
(b)
(50
vs77%
,P
=0.0
4)
NS
in
birth
wei
ght
NR
AA
,au
ricu
lar
acupunct
ure
;A
T,ac
upunct
ure
;EA
,el
ectr
oac
upunct
ure
;ED
A,
epid
ura
lan
alges
ia;
NR,
not
report
ed;
NRS,
11-p
oin
tnum
eric
rating
scal
e;N
S,no
signifi
cant
bet
wee
n-g
roup
diffe
rence
;TE
NS,
tran
scuta
neo
us
elec
tric
alner
vest
imula
tion;
VA
S,vi
sual
anal
ogue
scal
e.
*N
um
ber
of
pat
ients
anal
ysed
.
**D
e-qi
usu
ally
refe
rsto
acupunct
ure
-evo
ked
spec
ific
sensa
tions
such
asso
renes
s,num
bnes
s,hea
vines
s,an
ddis
tention
atth
esi
teof
nee
dle
pla
cem
ent,
and
thes
ese
nsa
tions
may
spre
adto
oth
erpar
tsof
the
body.
***Bre
athin
gan
dlo
calm
assa
ge
was
giv
enfo
ral
lfo
ur
gro
ups.
Cho et al.
912 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
Acupuncture interventionsThe types of acupuncture used in the trials were classical
acupuncture and EA: individualised (50%), standardised
(30%), and semi-standardised (20%) acupuncture styles
were tested (semi-standardised acupuncture is defined as a
set of acupoints supplemented by some additional points
individually chosen for each patient). The SP6 acupoint
was most commonly used (Table 2).
Types of control groupsA range of comparison groups were used including
transcutaneous electrical nerve stimulation (TENS),18,20
conventional analgesia,20,22,23 minimal acupuncture at non-
acupoints,15,19,24 EDA,18 subcutaneous sterile water injec-
tion,21 placebo EA,16,17 and no intervention.16–18,24 Two
studies16,17 from China employed placebo EA where needle
penetrating was simulated, followed by needle fixing on an
acupoint with adhesive tapes and connected to an inactive
electric device. Minimal acupuncture was defined as super-
ficial needling in places that are not regarded as acupoints,
but are mainly used as injection sites, such as buttocks,
thighs, and arms.15,19,24
Outcome measuresSeven studies reported labour pain on a 100-mm
VAS.15–19,21,24 Five studies assessed the number of women
requiring other pain relief at any time.15,20–23
Methodological qualityOur ‘risk of bias’ assessments are shown in Figure 2:
sequence generation was adequately described by descrip-
tion in eight trials.15,17–23 One study referred to a ran-
dom number table,18 five studies used computer-
generated randomisation for sequence generation,17,19–22
and two studies assigned groups by drawing lots.15,23 The
other trials did not describe the sequence generation pro-
cess.16,24
Allocation concealment was adequate in seven studies;
six studies ensured that allocations were concealed by using
envelopes.15,17,19,21–23 One study used a code-generating
system based on the unique civil registration number in
the log.20 The other trials were rated ‘unclear’, as they did
not have clear descriptions of their method of allocation
concealment.16,18,24
Because of the nature of the intervention and the condi-
tion, it was difficult for the women, care-givers, and out-
come assessors to be blinded. Not a single RCT had
masked women, care providers, and evaluators. Two stud-
ies in which minimal acupuncture was used in the control
groups had adequate participant and assessor blinding.15,19
In three trials participants were blinded using minimal acu-
puncture or placebo EA, but it was not clear whether the
care providers or evaluators were blinded.16,17,24
Seven studies had no participant losses, or the missing
data were balanced in numbers across intervention
groups.15–19,21,23 In two studies,20,22 an ‘intention-to-treat’
analysis was performed with one drop-out datum. The risk
of bias in the remaining trial is unclear because the number
of participants analysed in each group was not reported.24
As far as we could see, all studies included were free of
selective outcome reporting and other potential sources of
bias.
Adequate sequence generation
Allocation concealment
Participant blinding
Care provider blinding
Evaluator blinding
Incomplete outcome data addressed
Free of selective outcome reporting
Free of other bias
0 25 50 75 100 %
Yes (low risk of bias) Unclear No (high risk of bias)
Figure 2. Risk of bias graph.
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 913
Outcomes
Primary outcomesFigure 3 shows the forest plots with MDs in VAS for pain
intensity and RRs for the number of women requiring
analgesic methods, with 95% CI.
Electroacupuncture versus placebo EA (two studies):
combining two studies (154 participants) that compared
EA with placebo EA yielded a significant pain reduction in
favour of EA at 15 and 30 minutes of treatment (pooled
MD )4.09; 95% CI )8.05, )0.12; pooled MD )5.94; )9.83,
)2.06; respectively).16,17 The effect was not sustained
1–3 hours after EA. There was no significant statistical het-
erogeneity between these trials (I2 = 0.7%).
Acupuncture versus minimal acupuncture (three stud-
ies): three trials compared manual acupuncture with mini-
mal acupuncture, and reported VAS for pain.15,19,24 No
difference between the groups was found. One trial
reported a large difference in favour of acupuncture
30 minutes, 1 hour and 2 hours after acupuncture treat-
ment,15 whereas the other study failed to do so (MD 1.67;
95% CI )3.72, 7.06).19 No significant differences were also
found 1 and 2 hours after acupuncture (MD 4.30; 95% CI
)3.80, 12.40; MD 3.90; )3.57, 11.37; respectively).24 Con-
siderable heterogeneity was detected between studies for
this outcome, with I2 values ranging from 92 to 98%.
Acupuncture versus no intervention (four studies): four
trials compared EA,16,17,18 or manual acupuncture,24 with
no acupuncture intervention. Using a fixed-effects model
provided a significant pain reduction by EA (two studies)
at 15 and 30 minutes (pooled MD )6.81; 95% CI )10.77,
)2.81; pooled MD )10.56; )16.08, )5.03; respectively).16,17
Another study also reported a significant pain relief for the
EA group 30 minutes after acupuncture (MD )16.20;
95% CI )21.41, )10.99).18 No significant difference was
found between manual acupuncture and no acupuncture at
1, 2 and 3 hours after acupuncture. Substantial heterogene-
ity between studies was detected, except for VAS, after
15 minutes of EA (I2 = 0%).
Acupuncture versus conventional analgesia (three stud-
ies): less women receiving acupuncture required meperidine
compared with those receiving conventional analgesia alone
(pooled RR 0.20; 95% CI 0.12, 0.33; I2 = 0%).20,22,23
Whereas combined results from two studies did not reach
a statistical significance in the requirement for EDA (poo-
led RR 0.68; 95% CI 0.46, 1.00; I2 = 26%),20,23 the use of
any pharmacological and/or invasive methods (EDA,
nitrous oxide, meperidine, sterile water injection) was sig-
nificantly lower in the acupuncture groups (pooled RR
0.75; 95% CI 0.66, 0.85, I2 = 40%).20,22
Acupuncture versus other treatments (two studies): two
studies compared acupuncture or EA with TENS.18,20 In
the largest study comparing acupuncture plus conventional
care with TENS plus conventional care,20 the RR was 0.85
regarding the need for any pharmacological and/or invasive
pain relief (95% CI 0.74, 0.98). There was no significant
difference in the need for meperidine, and EDA (RR 0.92;
95% CI 0.17, 4.95; RR 1.06; 95% CI 0.68, 1.65; respec-
tively). In another study comparing EA with TENS,18 no
significant difference was reported after 30 minutes, or 1 or
2 hours after EA.
Secondary outcomesMore women in the acupuncture groups were satisfied with
their pain relief in two studies,18,19 were more relaxed,23 or
expressed a willingness to choose acupuncture for future
delivery.19 Women receiving acupuncture reported a signif-
icantly shorter duration of labour in three trials out of
eight reporting this outcome measure.15,17,19 Less oxytocin
use by the acupuncture group was reported in three out of
six studies.15,19,24 Acupuncture has no effect on the mode
of delivery or the extent of postpartum haemorrhage. Fetal
outcomes such as Apgar scores or birthweights were also
unaffected (Table 1).
Adverse eventsSix studies reported that there were no adverse events
caused by acupuncture treatment.15,17,19,20,22,23 The other
studies failed to mention adverse events.
Discussion and conclusion
Summary of main findingsThis systematic review of ten RCTs found little convincing
evidence that women receiving acupuncture are more likely
to experience less labour pain than those in the control
groups. Women rated their pain intensity on VAS in seven
studies; acupuncture was not effective compared with mini-
mal acupuncture in three studies. Patients reported signifi-
cantly less pain during EA treatment (after 15 and
30 minutes) compared with placebo EA, but, in two studies,
the effect was not maintained. When compared with no
treatment,16–18,24 acupuncture was effective only for the first
30 minutes. In trials where acupuncture was compared with
conventional analgesia women receiving acupuncture
required less meperidine and other analgesic methods.20,22,23
Acupuncture did not seem to have any impact on other
maternal or fetal outcomes, nor was it associated with harm.
Applicability of evidenceOur optimistic conclusion from the previous systematic
review was based on just three RCTs.6 The more negative
results of the present review are based on more and better
data. Before we accept that acupuncture has no effect on
labour pain, a careful look at the included studies is
required. The included studies originate from five different
Cho et al.
914 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
EA vs placebo EA: 100 mm VAS for pain
Study
A
B
/year
1.1.1 15 min for EA
Zhou 2007Zhang 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 0%Test for overall effect: Z = 2.02 (P = 0.04)
1.1.2 30 min for EA
Zhou 2007Zhang 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 0%Test for overall effect: Z = 3.00 (P = 0.003)
1.1.3 1 hour after EAZhou 2007
Zhang 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 0%Test for overall effect: Z = 1.17 (P = 0.24)
1.1.4 2 hours after EA
Zhou 2007Zhang 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 0%Test for overall effect: Z = 1.91 (P = 0.06)
1.1.5 3 hours after EA
Zhou 2007Zhang 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 7%Test for overall effect: Z = 1.55 (P = 0.12)
Weight
17.2%82.8%
100.0%
18.4%81.6%
100.0%
11.7%
88.3%100.0%
25.4%74.6%
100.0%
19.8%80.2%
100.0%
IV, fixed, 95% CI
–0.59 [–10.16, 8.98]–4.81 [–9.17, –0.45]–4.09 [–8.05, –0.12]
–5.70 [–14.76, 3.36]–6.00 [–10.30, –1.70]–5.94 [–9.83, –2.06]
–2.91 [–11.28, 5.46]
–1.56 [–4.61, 1.49]–1.72 [–4.59, 1.15]
–6.92 [–14.67, 0.83]–2.75 [–7.27, 1.77]–3.81 [–7.72, 0.10]
1.20 [–8.71, 11.11]–4.66 [–9.59, 0.27]–3.50 [–7.91, 0.91]
Mean difference Mean differenceIV, fixed, 95% CI
–10 –5 0 5 10Favours acupuncture Favours control
Acupuncture vs minimal acupuncture: 100 mm VAS for pain
Study/year
1.2.1 30 min after acupunctureHantoushzadeh 2007Skilnand 2002Subtotal (95% CI)Test for heterogeneity: I ² = 98%Test for overall effect: Z = 0.87 (P = 0.39)
1.2.2 1 hour after acupunctureHantoushzadeh 2007Skilnand 2002Ziaei 2006Subtotal (95% CI)Test for heterogeneity: I ² = 94%Test for overall effect: Z = 1.13 (P = 0.26)
1.2.3 2 hours after acupunctureHantoushzadeh 2007Skilnand 2002Ziaei 2006Subtotal (95% CI)Test for heterogeneity: I ² = 92%Test for overall effect: Z = 1.53 (P = 0.13)
Weight
49.9%50.1%
100.0%
34.1%33.9%31.9%
100.0%
34.2%33.2%32.6%
100.0%
IV, Random, 95% CI
1.67 [–3.72, 7.06]–23.20 [–28.12, –18.28]–10.79 [–35.16, 13.58]
–6.13 [–11.27, –0.99]–21.50 [–26.94, –16.06]
4.30 [–3.80, 12.40]–8.02 [–21.88, 5.84]
–13.40 [–18.92, –7.88]–20.60 [–27.39, –13.81]
3.90 [–3.57, 11.37]–10.15 [–23.18, 2.87]
Mean difference Mean differenceIV, Random, 95% CI
–20 –10 0 10 20Favours acupuncture Favours control
Figure 3. (A) EA versus placebo EA: 100-mm VAS for pain. (B) Acupuncture versus minimal acupuncture: 100-mm VAS for pain. (C) Acupuncture
vs. no intervention: 100-mm VAS for pain. (D) Acupuncture vs. conventional analgesia: use of meperidine, EDA, or any analgesia. Vertical line shows
no effect point; CI, confidence interval; EA, electroacupuncture; EDA, epidural analgesia; VAS, visual analogue scale.
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 915
Acupuncture vs no intervention: 100 mm VAS for pain
Study/year
1.3.1 15 min for EAZhang 2006Zhou 2007Subtotal (95% CI)Test for heterogeneity: I ² = 0%Test for overall effect: Z = 3.37 (P = 0.0008)
1.3.2 30 min for EAZhang 2006Zhou 2007Subtotal (95% CI)Test for heterogeneity: I ² = 46%Test for overall effect: Z = 3.75 (P = 0.0002)
1.3.3 30 min after EA
Huang 2008Subtotal (95% CI)
Test for heterogeneity: Not applicableTest for overall effect: Z = 6.10 (P < 0.00001)
1.3.4 1 hour after acupunctureHuang 2008Zhang 2006Zhou 2007Ziaei 2006Subtotal (95% CI)Test for heterogeneity: I ² = 92%Test for overall effect: Z = 1.95 (P = 0.05)
1.3.5 2 hours after acupuncture
Huang 2008Zhang 2006Zhou 2007Ziaei 2006Subtotal (95% CI)
Test for heterogeneity: I ² = 84%Test for overall effect: Z = 1.83 (P = 0.07)
1.3.6 3 hours after EAZhang 2006Zhou 2007Subtotal (95% CI)Test for heterogeneity: I ² = 53%Test for overall effect: Z = 1.71 (P = 0.09)
Weight
77.7%22.3%
100.0%
63.0%37.0%
100.0%
100.0%100.0%
–16.20 [–21.41, –10.99]–16.20 [–21.41, –10.99]
27.3%28.0%23.8%20.9%
100.0%
28.6%27.2%24.0%20.2%
100.0%
58.6%41.4%
100.0%
IV, Random, 95% CI
–5.81 [–10.30, –1.32]–10.28 [–18.68, –1.88]
–6.81 [–10.77, –2.84]
–8.40 [–12.67, –4.13]–14.24 [–21.48, –7.00]–10.56 [–16.08, –5.03]
–15.20 [–18.69, –11.71]–3.53 [–6.10, –0.96]
–13.12 [–19.80, –6.44]2.60 [–6.29, 11.49]
–7.71 [–15.48, 0.06]
–11.60 [–15.14, –8.06]–3.65 [–8.19, 0.89]
–12.59 [–19.15, –6.03]5.50 [–3.34, 14.34]
–6.23 [–12.91, 0.46]
–2.84 [–7.80, 2.12]–9.46 [–16.79, –2.13]–5.58 [–11.97, 0.81]
Mean difference Mean differenceIV, Random, 95% CI
–20 –10 0 10 20Favours acupuncture Favours control
Acupuncture vs conventional analgesia: use of meperidine, EDA, or any analgesia
Study/year
2.1.1 Use of meperidine
Borup 2009Nesheim 2003Ramnero 2002Subtotal (95% CI)
Test for heterogeneity: I ² = 0%Test for overall effect: Z = 6.40 (P < 0.00001)
2.1.2 Use of EDA
Borup 2009Ramnero 2002Subtotal (95% CI)
Test for heterogeneity: I ² = 26%Test for overall effect: Z = 1.98 (P = 0.05)
2.1.3 Use of any analgesia
Borup 2009Nesheim 2003Subtotal (95% CI)
Test for heterogeneity: I ² = 40%Test for overall effect: Z = 4.43 (P < 0.00001)
Weight
17.4%79.3%3.3%
100.0%
63.1%36.9%
100.0%
60.3%39.7%
100.0%
IV, Random, 95% CI
0.24 [0.07, 0.78]0.18 [0.10, 0.31]0.96 [0.06, 14.83]0.20 [0.12, 0.33]
0.79 [0.53, 1.16]0.52 [0.30, 0.92]0.68 [0.46, 1.00]
0.71 [0.63, 0.80]0.81 [0.69, 0.96]0.75 [0.66, 0.85]
Risk ratio Risk ratioIV, Random, 95% CI
0.1 0.2 0.5 1 2 5 10Favours acupuncture Favours control
C
D
Figure 3. (Continued)
Cho et al.
916 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
Table 2. Acupuncture interventions in the included studies
First author (year) Type of
acupuncture
Acupoint
selection
Acupuncture
treatment
rationale
Treatment regimen
Borup (2009)19 MA + AA Individualised Previous reports
and pilot projects
Needle size: 0.20 · 15, 0.30 · 30,
or 0.35 · 50 mm
Needles left for 30 minutes–2
hours, or repeated; removed
if necessary
Unclear whether de-qi response
elicited
Selected from 34 specified acupoints
Hantoushzadeh (2007)18 MA Individualised Textbooks Bilateral insertion
Needles removed if necessary
De-qi sensation sought
LI4, LR3, and ST36 for analgesia
and general pain; BL32 and BL60
for back pain; SP6 for severe pain
during contraction; GB34 for
cervical rigidity; and HT7 for anxiety
and nervousness
Huang (2008)17 EA Basic + individualised NR 2/100 Hz
Basic acupoints: ST36, CV3, and CV4
Adjunctive points: BL32 and LR3 for
excessive pattern; LI4 and SP6 for
deficient pattern in TCM diagnosis
Martensson (2008)20 MA Basic + individualised Literature and
selection by
participating midwives
Needle size: 0.30 · 30,
or 0.35 · 50 mm
Needles left for 40 minutes, or
repeated if necessary; needles
were manually stimulated
every 10 minutes
De-qi sensation sought
GV20, LI4, and SP6, and between
four and seven individually chosen
points from BL23, BL24, BL54,
EX19, GB25–29, and KI11
Nesheim (2003)21 MA Individualised Selection by
participating midwives
Needle size: different needles for
different points (0.5–1.5 inches,
30/32 gauge)
Needles removed after 10–20 minutes,
or taped and left in place for the
entire duration of labour
De-qi sensation sought
GV20, HT7, and LR3 for tension,
nervousness, or tiredness; LR3
and GB34 for cervix rigidity; CV4
for symphysis pain in early labour;
LI10 and LI11 for pain during the
first stage of labour; BL23 and
BL60 for back pain in early labour;
BL27, BL28, and BL32 for back pain
later in labour; LI4 and SP6 for strong
pain during contractions; PC6 and
PC7 for nausea; and ST36 for
any pain
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 917
countries, they compared different acupuncture methods
with different controls, they had different inclusion and
exclusion criteria, and a range of outcome measures, and
routine care given to patients varied as well. It was difficult
to determine whether the acupuncture intervention given
was appropriate or optimal for labour pain relief in terms
of acupoint selection, stimulation, duration, or timing.
Only three out of ten included studies specified how many
hours of acupuncture training the study midwives had
received,20,22,23 and it seems safe to assume that their acu-
puncture skills and experience varied greatly, not only
across studies but also within trials. Some studies included
women with cervical dilatation of 2–3 cm at admission,16,17
and in one study women in labour were asked to partici-
pate if they requested pain relief.20 In addition to this high
level of clinical heterogeneity, most of our analyses detected
a significant level of statistical heterogeneity. All this means
that our findings must be interpreted with caution.
Table 2. (Continued)
First author (year) Type of
acupuncture
Acupoint
selection
Acupuncture
treatment
rationale
Treatment regimen
Ramnero (2002)22 MA Individualised Selection by
participating midwives
Needles removed after 1–3 hours,
or when necessary; taped to allow
women to move freely
De-qi sensation sought
GV20, Yintang, LU7 for relaxation;
BL25–36, BL54, BL60, GB25–29,
GB41, CV2, CV3, subcutaneous
points, LI4, LR3, KI3, and SP6
for analgesia
Skilnand (2002)23 MA Individualised Treatment protocol from
the Norwegian School
of Acupuncture
2–12 needles (average 7) used;
removed after 20 minutes or
left until conversion to conventional
analgesia or delivery; taped to
allow women to move freely
De-qi sensation sought
LU7, HT7, ST29, ST30, ST36, GB34,
SP6, SP8, KI3, LR3, GB41, BL32,
BL34, BL60, BL67, LI4, and GV20
Zhang (2006)16 EA Fixed Literature No manipulation, 25-mm insertion
depth, 2/100 Hz, 20 mA,
for 30 minutes
Rt SP6
Zhou (2007)15 EA Fixed Literature No manipulation, 25–30-mm insertion
depth, 2/100 Hz, electric current
threshold individually adjusted,
for 30 minutes
Rt SP6
Ziaei (2006)24 MA Fixed NR Needles inserted when VAS > 3;
taped and left until delivery
De-qi sensation sought
GV20, Yintang for relaxation;
ST36, SP6, LI3, LI4, CV2,
and CV3 for analgesia
De-qi refers to acupuncture-evoked specific sensations such as soreness, numbness, heaviness, and distention at the site of needle placement,
and these sensations may spread to other parts of the body. Acupoint ST36 refers to 36th point of the stomach meridian. Some extra points have
Chinese names.
AA, auricular acupuncture; BL, bladder meridian; CV, conception vessel; EA, electroacupuncture; EX19, extra point 19 (Shiqizhui); GB, gallbladder
meridian; GV, governor vessel; HT, heart meridian; KI, kidney meridian; LI, large intestine meridian; LR, liver meridian; LU, lung meridian; MA,
manual acupuncture; NR, not reported; PC, pericardium meridian; Rt, right; SP, spleen meridian; ST, stomach meridian; TCM, traditional Chinese
medicine; VAS, visual analogue scale.
Cho et al.
918 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology
It may be confusing that women receiving acupuncture
required less meperidine and other analgesic methods than
the women receiving conventional analgesia alone, whereas
no difference was detected between real and sham acupunc-
ture. It should be noted that acupuncture versus conven-
tional analgesia studies also reported similar pain intensity
ratings between groups, and this is in agreement with the
results from acupuncture versus sham acupuncture studies
(Table 1). Acupuncture is well known for its powerful non-
specific/placebo effect.25 One possible interpretation could
be that the analgesic effect of acupuncture did not reach a
significantly different level from that of the control group,
but that women receiving acupuncture may have felt that
they are being given something, thus did not require more
analgesic methods.
Quality of studiesWe used the ‘risk of bias’ assessment tool by the Cochra-
ne collaboration to evaluate the quality of the primary
studies.12 Although quality rating scales offer simplicity by
calculating a summary score, weights given to different
items of the scale are difficult to justify.26 Instead of
adopting quality scales, we tried to fully report how each
trial has been rated for six critical domains, i.e. randomi-
sation, allocation concealment, blinding, incomplete out-
come data reporting, selective outcome reporting, and
other biases. Yet the ‘risk of bias’ assessment tool may
not be entirely free from subjectivity.27 We tried to mini-
mise this source of bias by working independently to eval-
uate each domain for each study, and solving
discrepancies through discussions. The allocation sequence
generation and concealment were assessed to be adequate
in most studies included. Two studies rated as ‘unclear’
for allocation concealment,18,24 against the notion that
inadequate or unclear allocation concealment is associated
with overestimates of effect,28 did not demonstrate acu-
puncture was better than the sham control. Probably the
highest risk of bias in the included studies lies in blind-
ing. As most studies used subjective outcome measures,
blinding may have a larger impact on the result than the
other domains. None of the included studies blinded the
three parties of participant, care provider, and outcome
assessor. Considering the condition itself and the nature
of acupuncture treatment, blinding all three parties may
not be practical. However, separating the roles of care
provider and outcome assessor would have minimised the
measurement bias: only two of the ten trials blinded
the outcome assessor.15,19 Reviewers agreed that there
were no major flaws in outcome reporting and other
potential biases in the included studies. Consequently,
incomplete blinding of the included trials seems to bring
the highest risk to their overall methodological quality.
Potential biases of the reviewBias can be introduced in every step of reviewing, and our
systematic review is no exception. Heterogeneity may be
caused by known methodological differences between trials,
or it can be related to unknown study characteristics. The
included trials greatly varied in the characteristics of partic-
ipants, labour management protocols, and control types. In
addition, quite different interventions were tested under
the label of ‘acupuncture’, and all these factors may have
influenced the course of labour, analgesic requirements,
including EDA, and accordingly, the outcomes. The statisti-
cal heterogeneity found in the meta-analyses of our pri-
mary outcomes may thus be inevitable. A large number of
studies would be necessary to investigate potential sources
of heterogeneity.
The potential of missing eligible trials always exists in
any systematic review. Even though empirical evidence
tells us that including publications in languages other
than English appears to have little impact on the conclu-
sion of the review,29 we tried to be comprehensive by
searching all available databases across countries without
language restrictions. The three Chinese language articles
do not appear to deviate from the overall findings of our
review.
Implications for practice and researchAcupuncture does not appear to afford more effective
pain relief than sham/placebo acupuncture, conventional
analgesia, or no analgesia at all. Because of the heteroge-
neity of the studies, the findings in this review are diffi-
cult to interpret as firm evidence against acupuncture for
pain relief in labour. Acupuncture appears to reduce the
requirement for other forms of pain relief compared with
conventional analgesia, but the evidence is not strong. No
evidence emerged for any harm caused by acupuncture.
We believe the evidence presented in this review should
be made available to both doctors and women to help
them make informed decisions on whether or not to use
acupuncture for pain control. A number of questions
should be answered in future studies. This review found
the data to be limited on the optimal acupuncture inter-
vention and corresponding sham/placebo controls, the
best timing of outcome measurements, and additional
staff and necessary education. We have no sound data on
the costs associated with providing acupuncture during
labour, but, according to our estimate, they are not exces-
sive. Adverse events were frequently not monitored.
Future studies should be planned and conducted to
address all these issues.
In conclusion, current evidence does not appear to sup-
port the use of acupuncture for labour pain. Acupuncture
might reduce the use of other forms of pain relief, but the
Current evidence does not support acupuncture for labour pain
ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology 919
evidence is limited. The heterogeneity of the studies avail-
able to date prevents us making a firm conclusion. Further
research seems to be warranted.
Disclosure of interestsNone.
Contribution to authorshipS-HC formulated the idea for updating the systematic
review; S-HC and HL performed literature searches, study
selection, data extraction, risk of bias assessment, meta-
analysis, and wrote the initial draft; S-HC, HL, and EE
resolved disagreements regarding study selection and risk
of bias assessment, and critically revised the manuscript.
Details of ethics approvalNo ethics approval was required.
FundingHyangsook Lee was supported by the Basic Science
Research Programme through the National Research
Foundation, funded by the Korean Ministry of Education,
Science and Technology (R11-2005-014).
AcknowledgementWe thank Ditlev Fossen, MD, for providing us with addi-
tional data. j
References
1 Tournaire M, Theau-Yonneau A. Complementary and alternative
approaches to pain relief during labor. Evid Based Complement
Alternat Med 2007;4:409–17.
2 Hodnett E, Gates S, Hofmeyr G, Sakala C. Continuous support for
women during childbirth. Cochrane Database Syst Rev
2007;CD003766.
3 Cluett E, Burns E. Immersion in water in labour and birth. Cochrane
Database Syst Rev 2009;CD000111.
4 Smith C, Collins C, Cyna A, Crowther C. Complementary and alter-
native therapies for pain management in labour. Cochrane Database
Syst Rev 2006;CD003521.
5 Allaire A, Moos M, Wells S. Complementary and alternative
medicine in pregnancy: a survey of North Carolina certified nurse-
midwives. Obstet Gynecol 2000;95:19–23.
6 Lee H, Ernst E. Acupuncture for labor pain management: a system-
atic review. Am J Obstet Gynecol 2004;191:1573–9.
7 Ewies A, Olah K. The sharp end of medical practice: the use of acu-
puncture in obstetrics and gynaecology. BJOG 2002;109:1–4.
8 Ee C, Manheimer E, Pirotta M, White A. Acupuncture for pelvic and
back pain in pregnancy: a systematic review. Am J Obstet Gynecol
2008;198:254–9.
9 Wallis L, Shnider S, Palahniuk R, Spivey H. An evaluation of acu-
puncture analgesia in obstetrics. Anesthesiology 1974;41:596–601.
10 Abouleish E, Depp R. Acupuncture in obstetrics. Anesth Analg
1975;54:83–8.
11 Hyodo M, Gega O. Use of acupuncture anesthesia for normal deliv-
ery. Am J Chin Med 1977;5:63–9.
12 Higgins JPT, Altman D. Chapter 8: assessing risk of bias in included
studies. In: Higgins J, Green S, editors. Cochrane Handbook for Sys-
tematic Reviews of Interventions. The Cochrane Collaboration,
2008. Available from www.cochrane-handbook.org.
13 Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-
analysis. Stat Med 2002;21:1539–58.
14 Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche P, Ioannidis J,
et al. The PRISMA statement for reporting systematic reviews and
meta-analyses of studies that evaluate healthcare interventions:
explanation and elaboration. BMJ 2009;339:b2700.
15 Skilnand E, Fossen D, Heiberg E. Acupuncture in the management
of pain in labor. Acta Obstet Gynecol Scand 2002;81:943–8.
16 Zhou PJ. Clinical observation on relieving the labor pain of electro-
acupuncture on SP6. Thesis, Beijing University of Chinese Medicine;
2007.
17 Zhang Y. Clinical observation on therapeutic effect of electroacu-
puncture on SP6 for labor pain. Thesis, Beijing University of Chinese
Medicine; 2006.
18 Huang T, Yang Y, Huang X. Selection of acupoints and opportunity
for acupuncture analgesia in delivery. J Tradit Chin Med
2008;49:625–8.
19 Hantoushzadeh S, Alhusseini N, Lebaschi AH. The effects of
acupuncture during labour on nulliparous women: a randomised
controlled trial. Aust N Z J Obstet Gynaecol 2007;47:26–30.
20 Borup L, Wurlitzer W, Hedegaard M, Kesmodel US, Hvidman L.
Acupuncture as pain relief during delivery: a randomised controlled
trial. Birth 2009;36:5–12.
21 Martensson L, Stener-Victorin E, Wallin G. Acupuncture versus sub-
cutaneous injections of sterile water as treatment for labour pain.
Acta Obstet Gynecol Scand 2008;87:171–7.
22 Nesheim BI, Kinge R, Berg B, Alfredsson B, Allgot E, Hove G, et al.
Acupuncture during labor can reduce the use of meperidine: a con-
trolled clinical study. Clin J Pain 2003;19:187–91.
23 Ramnero A, Hanson U, Kihlgren M. Acupuncture treatment dur-
ing labour – a randomised controlled trial. BJOG 2002;109:637–
44.
24 Ziaei S, Hajipour L. Effect of acupuncture on labor. Int J Gynaecol
Obstet 2006;92:71–2.
25 Kaptchuk T, Stason W, Davis R, Legedza A, Schnyer R, Kerr C, et al.
Sham device v inert pill: randomized controlled trail of two placebo
treatments. BMJ 2006;332:391–7.
26 Moher D, Jadad A, Nichol G, Penman M, Tugwell P, Walsh S.
Assessing the quality of randomized controlled trials: an annotated
bibliography of scales and checklists. Control Clin Trials 1995;16:62–
73.
27 Ernst E, Resch K. Reviewer bias: a blinded experimental study. J Lab
Clin Med 1994;124:178–82.
28 Pildal J, Hrobjartsson A, Jørgensen K, Hilden J, Altman D, Gøtzsche
P. Impact of allocation concealment on conclusions drawn from
meta-analyses of randomized trials. Int J Epidemiol 2007;36:847–
57.
29 Egger M, Juni P, Bartlett C, Holenstein F, Sterne J. How important
are comprehensive literature searches and the assessment of trial
quality in systematic reviews? Empirical study. Health Technol Assess
2003;7:1–76.
Cho et al.
920 ª 2010 The Authors Journal compilation ª RCOG 2010 BJOG An International Journal of Obstetrics and Gynaecology