Journal of Psychiatric Research 47 (2013) 726e732

Contents lists available at SciVerse ScienceDirect

Journal of Psychiatric Research

journal homepage: www.elsevier.com/locate/psychires

A 6-week randomized controlled trial with 4-week follow-up of acupuncture

combined with paroxetine in patients with major depressive disorder
Shan-Shan Qu a, Yong Huang a, **, Zhang-Jin Zhang b, *, Jun-Qi Chen a, Ren-Yong Lin a, Chong-Qi Wang a,
Gan-Long Li a, Hei Kiu Wong b, Cang-Huan Zhao c, Ji-Yang Pan c, Shen-Chang Guo d, Yan-Chi Zhang d
a
School of Traditional Chinese Medicine, Southern Medical University, 1023 Shatai Road, Guangzhou, Guangdong 510515, China
b
School of Chinese Medicine, LKS Faculty of Medicine, The University of Hong Kong, 10 Sassoon Road, Pokfulam, Hong Kong, China
c
Department of Psychiatry and Clinical Psychology, The First Affiliated Hospital of Jinan University, Guangzhou, Guangdong 510630, China
d
Department of Clinical Psychology and Behavior, Guangdong 999 Brain Hospital, Guangzhou, Guangdong 510510, China

a r t i c l e i n f o a b s t r a c t

Article history: Acupuncture possesses the antidepressant potential. In this 6-week randomized controlled trial with 4-
Received 22 November 2012 week follow-up, 160 patients with major depressive disorder (MDD) were randomly assigned to par-
Received in revised form oxetine (PRX) alone (n ¼ 48) or combined with 18 sessions of manual acupuncture (MA, n ¼ 54) or
5 February 2013
electrical acupuncture (EA, n ¼ 58). Treatment outcomes were measured mainly using the 17-item
Accepted 5 February 2013
Hamilton Depression Rating Scale (HAMD-17), Self-rating Depression Scale (SDS), clinical response
and remission rates. Average PRX dose taken and proportion of patients who required an increased PRX
Keywords:
dose due to symptom aggravation were also obtained. Both additional MA and EA produced a signifi-
Acupuncture
Paroxetine
cantly greater reduction from baseline in score on HAMD-17 and SDS at most measure points from week
Major depressive disorder 1 through week 6 compared to PRX alone. The clinical response was markedly greater in MA (69.8%) and
Clinical trial EA (69.6%) groups than the group treated with PRX alone (41.7%, P ¼ 0.004). The proportion of patients
Follow-up who required an increase dose of PRX due to symptom aggravation was significantly lower with MA
(5.7%) and EA (8.9%) than PRX alone (22.9%, P ¼ 0.019). At 4 weeks follow-up after completion of
acupuncture treatment, patients with EA, but not MA, continued to show significantly greater clinical
improvement. Incidence of adverse events was not different in the three groups. Our study indicates that
acupuncture can accelerate the clinical response to selective serotonin reuptake inhibitors (SSRIs) and
prevent the aggravation of depression. Electrical acupuncture may have a long-lasting enhancement of
the antidepressant effects (Trial Registration: ChiCTR-TRC-08000278).
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction
Depression is a serious mental illness that affects 8e20% of the
worldwide population (Ferrari et al., 2012). Although selective se-
rotonin reuptake inhibitors (SSRIs), such as paroxetine (PRX) and
fluoxetine (FLX), are a first-line pharmacotherapy for various
depressive disorders, there still remains a large portion of
depressed patients who do not make a full response and experience
relapse and adverse effects of treatment (Arroll et al., 2005). The
delay in the onset of the action of SSRIs further prolongs the
suffering of patients and exposes them to a substantial risk of
suicide (Adell et al., 2005). A high incidence of side effects also has
* Corresponding author. Tel.: þ852 2589 0445; fax: þ852 2872 5476.
** Corresponding author. Tel.: þ86 20 6164 8771.
E-mail addresses: nanfanglihuang@163.com (Y. Huang), zhangzj@hku.hk,
zhangz68@yahoo.com (Z.-J. Zhang).
hampered the clinical use of SSRIs (Arroll et al., 2005). These lim-
itations of treatment with SSRIs are thought to be due to the multi-
system pathogenesis of depressive disorders (Ward and Irazoqui,
2010). For example, it is well documented that, in addition to a
central serotonin (5-HT) deficiency, depressive disorders are asso-
ciated with hypothalamicepituitaryeadrenal axis dysfunction and
abnormalities in the brain regions associated with stress and
emotion processing (Rigucci et al., 2010; Ward and Irazoqui, 2010).
As an ancient therapeutic technique, acupuncture has been well
confirmed to be a generally safe and well tolerated therapy for
neuropsychiatric disorders (He et al., 2012; Lao et al., 2003; Zhang
et al., 2010). It has been increasingly introduced into the treatment
of various depressive disorders (Manber et al., 2010; Smith et al.,
2010; Wu et al., 2012; Zhang et al., 2010). Both manual and elec-
trical acupuncture stimulation considerably enhances the release of
5-HT from the brainstem raphe nuclei, the principal source of 5-HT
neuronal bodies sending axons to cortical and subcortical regions

0022-3956/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jpsychires.2013.02.004
 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732
(Kwon et al., 2000; Lee et al., 2004; Li et al., 2007; Zhao, 2008). It
robustly modulates neuroendocrine functions, in particular the
hypothalamicepituitaryeadrenal axis; it reduces stress-induced
behavior, and targets brain regions involved in emotion processing
(Zhang et al., 2012a). These observations suggest that acupuncture
may have an antidepressant potential. Indeed, numerous clinical
studies and observations have shown the therapeutic benefits of
acupuncture in treating depressed patients (Zhang et al., 2010).
Acupuncture is also beneficial in alleviating pain, autonomic, sleep,
and other mood symptoms (Zhang et al., 2012a). On the other hand,
many studies have shown superior effects of electroacupuncture
(EA) compared to manual acupuncture (MA) in alleviating tinnitus
(Wang et al., 2010), pain (Sator-Katzenschlager et al., 2003, 2004;
Schliessbach et al., 2011; Tsui and Leung, 2002), and modulating rat
neuroendocrine function (Feng et al., 2012). Most recently, we have
demonstrated that EA has an immediate and short-term effect in
enhancing the efficacy in the early phase of FLX treatment of patients
with major depressive disorder (MDD) (Zhang et al., 2012b) and
potential benefit in patients with postpartum depression (Chung
et al., 2012). Hence we hypothesize that adjuvant acupuncture, and
in particular EA, may provide a long-lasting enhancement of the
antidepressant efficacy of SSRIs.
To test this hypothesis we conducted a 6-week, randomized
controlled trial with 4-week follow-up of MA or EA combined with PRX
compared to PRX alone in patients with MDD. Adjuvant MA and EA
were both studied because a large body of evidence suggests different
neural pathways for these two most commonly used stimulation
modes (Zhang et al., 2012a). Here, EA is defined as an acupuncture
procedure in which inserted needles are manipulated manually at first,
followed by electrical stimulation. Paroxetine (PRX) was selected
because it is one of the most frequently prescribed SSRIs in China (Fang
et al., 2010, 2011) and its pharmacological and therapeutic properties
have been well delineated (Gibiino and Serretti, 2012).
2. Methods
2.1. Settings and subjects
This randomized controlled trial was conducted in Outpatient
Acupuncture Clinic of Southern Medical University, the First Affil-
iated Hospital of Jinan University, and Guangdong 999 Brain Hos-
pital between December 2008 and October 2010. The study
protocol was approved by Medical Ethical Committee of the First
Affiliated Hospital of Jinan University and registered in www.chictr.
org (Trial Registration: ChiCTR-TRC-08000278). All participants
gave voluntary, written, informed consent before entering the trial.
Outpatients were referred by psychiatrist from the First Affili-
ated Hospital of Jinan University and Guangdong 999 Brain Hos-
pital. Patients who met the following entry criteria were eligible for
the study: (1) either gender aged 18e60 years; (2) had a diagnosis
of MDD with the International Classification of Diseases (10th
version) (ICD-10); (3) moderate or severe illness, with a score of at
least 17 on the 17-item Hamilton Rating Scale for Depression
(HAMD-17) (Hamilton, 1960) and at least 4 on the Clinical Global
Impression-Severity (CGI-S); and (4) current either PRX or other
antidepressant treatment did not exceed one month. Patients who
had any of the following conditions were excluded from the study:
(1) unstable medical conditions; (2) a history of brain injury or
surgery; (3) suicidal attempts or aggressive behavior; (4) a history
of manic, hypomanic, or mixed episode illness; (5) comorbid with
other neuropsychiatric disorders; (6) a family history of mental
illnesses; (7) investigational drug treatment within the previous 6
months; (8) a history of alcohol or drug abuse within the previous
12 months; (9) pregnancy or lactation; (10) currently under
cognitive behavioral therapy or other behavioral therapies.
727
All participants were monitored closely on a daily check-up
basis by patient self-report, family member or doctor reports.
Those whose conditions developed significant suicidal or aggres-
sive behaviors that were a danger to themselves or others deter-
mined by psychiatrists, or had compliance rates with medication or
acupuncture regime less than 75% were discontinued from the
study and given standard clinical treatment.
2.2. Randomization and group allocation
Patients were randomly assigned to one of three groups: PRX
alone or combined with MA or EA. For randomization, simple,
complete, non-sequential random numbers were generated in
advance by a computer program (SPSS version II) and sealed in en-
velopes. The group allocation was done in a semi-blind manner, in
which random codes were known by only acupuncturists (J.Q.C. and
G.L.L.), but blind to other study personnel including the Principal
Investigator (Y.H.), the Associate Investigator (Z.J.Z.), psychiatrists
(C.H.Z. and S.C.G.), clinical assessors (J.Y.P., Y.C.Z. and C.Q.W.), data
collector and analysts (S.S.Q. and Y.R.L.). Clinical assessors and psy-
chiatrists communicated with patients separately and were
instructed not acquire information about their treatment conditions.
2.3. Paroxetine treatment
Patients in all three groups received PRX orally for 6 weeks in an
open manner. For patients who were not medicated at the time of
trial, the PRX dose was initiated at 10 mg/day and escalated to
20 mg/day within one week, based on individual patient response.
The FLX dose would be further increased if psychiatrists believed
patients’ symptom aggravated. The maximum dose was set at
40 mg/day. This PRX dosing regimen has been widely used in
Chinese patients (Fang et al., 2010, 2011). Patients who were
already taking PRX for less than one month would continue his/her
PRX regimen. Those who were taking other antidepressants for less
than one month would switch to PRX within one week. During 4
weeks of follow-up after acupuncture ended, patients are asked to
continue taking PRX; those whose medications were switched to
other antidepressants were excluded from follow-up analysis.
Concomitant use of other psychoactive agents was generally not
allowed, but the use of benzodiazepines and non-benzodiazepines
(e.g., zopiclone) for insomnia was permitted as long as these were
taken no more than 14 days cumulatively, as PRX has least in-
teractions with benzodiazepines and other anti-insomnia agents
(Calvo et al., 2004; Sproule et al., 1997).
Treatment outcomes included the average PRX dose taken over
6 weeks of treatment and over 4 weeks of follow-up, and the
proportion of patients who required an increase of at least 5 mg/
day PRX due to worsening symptoms.
2.4. Acupuncture intervention
Patients allocated to adjuvant MA or EA received 3 sessions per
week over 6 consecutive weeks. A brief introduction of acupunc-
ture procedure was given by acupuncturists during first visit. Based
on empirical evidence and previous studies (Zhang et al., 2010), the
ten commonly used acupoints in the treatment of depressive
symptoms were chosen for this study: Baihui (GV20), Yintang (EX-
HN3), Fengfu (GV16), Dazhui (GV14), bilateral Fengchi (GB20),
bilateral Neiguan (PC6), and bilateral Sanyinjiao (SP6). Disposable
acupuncture needles (0.30 mm in diameter and 25e40 mm in
length, Hwato) were perpendicularly or obliquely inserted into
acupoints at a depth of 10e30 mm. For MA, manual manipulation
was conducted and needling sensation was generally achieved
within 2 min after the manipulation. After needling sensation was
achieved, the needles were retained for 30 min and manipulated
728 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732
once again during retaining in order to maintain needling sensa-
tion. For EA, manual manipulation was also performed to achieve
needling sensation, and electrical stimulation was then delivered
between Baihui (Du20) and Yintang (EX-HN3) and between bilat-
eral Fengchi (GB20) with continuous waves at alternating low
(2 Hz) and high (100 Hz) frequency using Han’s Acupuncture Nerve
Stimulator (H.A.N.S., mode code: LH202H) as reported previously
(Han, 2003). The stimulation with alternating low and high fre-
quency is thought to produce broader central neurochemical effects
compared to unitary frequency (Han, 2003). The intensity of
stimulation was adjusted to a level at which patients felt most
comfortable. The stimulation lasted 30 min. Other acupoints that
were not stimulated electrically received an additional session of
manual manipulation during 30-min treatment. Those who were
currently having acupuncture treatment were required for at least a
one-week washout period.
Acupuncture procedure was performed in treatment room by
experienced acupuncturists (J.Q.C. and G.L.L.) who had received 5-
year undergraduate training in Chinese medicine and practiced Chi-
nese medicine over three years. To ensure consistency in acupuncture
procedure, the acupuncturists received a training workshop to
establish the acupuncture protocol before the study began.
2.5. Clinical assessment
Clinical assessment was conducted in examination room at each
site. The Clinical outcomes during 6 weeks of treatment and after 4-
week further follow-up were assessed. The primary outcomes
measured were the change from baseline in the total score on
HAMD-17 and the Chinese-version Self-rating Depression Scale
(SDS) (Lee et al., 1994) at baseline and week 1, 2, 4, and endpoint
(week 6) of treatment as well as at week 4 of follow-up. The
symptom severity was also measured using CGI-S at baseline and
endpoint of treatment. The secondary outcomes included clinical
response and remission. The clinical response is defined as 50%
reduction at endpoint from baseline on HAMD-17. The remission is
defined as an endpoint HAMD-17 score of 7. Safety and tolera-
bility were measured using Åsbergs’s Side Effects Rating Scale
(SERS) when patients visited study sites (Asberg et al., 1970).
To ensure consistency of assessments across the study, a training
workshop with video materials was conducted for clinical assessors
(J.Y.P., Y.C.Z. and C.Q.W.). An interrater reliability coefficients (k value)
of >0.80 was achieved after the completion of training workshop. In
most cases, assessments from baseline through 4-week follow-up
for each individual were conducted by the same assessor.
Compliance with PRX and acupuncture intervention was
calculated by dividing the number of PRX tablets actually taken and
acupuncture sessions completed by the number the patient should
have had and multiplying by 100.
2.6. Statistical analysis
Previous studies have shown that acupuncture intervention can
produce at least a 3-point reduction in HAMD-17 (Zhang et al.,
2010). Therefore, a sample size of 150 patients (n ¼ 50 per group)
should provide approximately 80% power at a statistical level of
0.05, with an estimated standard deviation of 4.5 and an estimated
dropout rate of 15% at the endpoint of 6-week treatment.
Efficacy analyzes were performed on the intention-to-treat
population, defined as participants who completed baseline and
at least one evaluation after treatment. Since measure time points
were not balanced in 6-week treatment, a linear mixed-effect
model was applied to compare treatment outcomes (HAMD-17
and SDS) over time among the three groups. The model was
established using time and group for categorical fixed factors and
random intercepts with scaled identity covariance matrix. Subject’s
age, gender, duration of the illness, and baseline HAMD-17 were
treated as covariates. Between-group differences at each measure
time point were further examined using one-way analysis of vari-
ance (ANOVA) based on observed data. One-way ANOVA was also
applied to detect between-group differences among the threes
groups in HAMD-17 and SDS at 4-week follow-up, CGI-S at
endpoint of the treatment, continuous baseline variables, and
average PRX doses. Categorical variables, including categorical
baseline variables, response and remission rates, the proportion of
patients who required an increased PRX dose due to worsening of
symptoms, and incidence of adverse events were analyzed using
Chi-square (c2) test. Statistical significance was defined as a two-
tailed P < 0.05. The analyzes were performed with SPSS version
16 software (Chicago, IL, USA).
3. Results
3.1. Disposition and characteristics of patients
Of 493 patients screened, 160 eligible patients were recruited
from Southern Medical University (n ¼ 50), the First Affiliated
Hospital of Jinan University (n ¼ 30), and Guangdong 999 Brain
Hospital (n ¼ 80). They were randomly assigned to PRX alone
(n ¼ 48), MA þ PRX (n ¼ 54) and EA þ PRX (n ¼ 58); 143 (89.4%) of
them completed the 6-week treatment and assessment. Three pa-
tients (1 in MA þ PRX and 2 in EA þ PRX) who did not complete the
first session of acupuncture treatment due to fainting of needling
and had no post-baseline assessment were excluded from data
analysis. Ninety-one patients (56.9%, 91/160) completed 4-week
follow-up per protocol (Fig. 1).
Baseline characteristics of the three groups of patients are sum-
marized in Table 1. No significant differences in baseline variables
were observed among the three groups. There were 55% (88/160) of
patients with first-onset MDD. Only 13.1% (21/160) of patients had
acupuncture treatment previously and 17.6% (28/160) were under
psychoactive medication when entering the study. The compliance
with acupuncture and PRX treatment was nearly 93% in the three
groups. No patients had to be discontinued due to the development
of significant suicidal or aggressive behavior or compliance rates
with medication or acupuncture regime less than 75%.
3.2. Clinical outcomes of 6-week treatment
The linear mixed-effect model revealed a significant difference
among the three groups in the slope of linear regression analysis
between measure time points and changes from baseline in score
on HAMD-17 (P ¼ 0.000, r2 ¼ 0.725 for PRX alone, r2 ¼ 0.655 on
MA þ PRX, and r2 ¼ 0.784 on EA þ PRX) and SDS (P ¼ 0.009,
r2 ¼ 0.414 for PRX alone, r2 ¼ 0.421 on MA þ PRX, and r2 ¼ 0.661 on
EA þ PRX). Between-group comparisons further revealed that both
MA and EA additional treatment produced a significantly greater
reduction in scores on HAMD-17 compared to PRX alone at week 1
through week 6 (P ¼ 0.000), but no significant differences were
observed between the two acupuncture groups. A significantly
greater reduction was also observed in SDS scores in patients with
EA þ PRX at week 1 through week 6 and in patients with MA þ PRX
at week 2 and week 6 (P  0.012) compared to PRX alone. At the
endpoint of treatment, both groups of patients treated with addi-
tional MA and EA had a significantly greater reduction in CGI-S
score compared to PRX alone (P  0.006) (Table 2, Fig. 2).
The response rates of MA þ PRX and EA þ PRX groups were
significantly higher than that of PRX alone (69.8% and 69.6% vs.
41.7%, c2 ¼ 11.043, P ¼ 0.004), but the remission rate across groups
 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732 729

Screened (n = 493)

Excluded in pre-randomization (n = 333):

Did not meet the inclusion criteria (n = 237)
Refused to participate (n = 41)
Had other severe medical conditions (51)
Had needle phobia (n = 4)

Randomized (n = 160)

PRX alone (n =48) MA+PRX (n = 54) EA+PRX (n = 58)

Treatment dropouts (n = 5): Treatment dropouts (n = 9)

Treatment dropouts (n = 3):
Lost to follow-up (n = 2)
Lost to follow-up (n = 1)
Lack of efficacy (n = 2) Fainting of needling (n = 1)
Aggravation (n = 1)
Lack of efficacy (n = 1)
Lost to follow-up (n = 3)
Fainting of needling (n = 2)
Lack of efficacy (n = 2)
Aggravation (n = 1)
Pregnancy (n = 1)

Completed treatment Completed treatment Completed treatment

(n = 43, 89.6%) (n = 51, 94.4%) (n = 49, 84.5%)

Follow-up dropouts (n = 14): Follow-up dropouts (n = 23): Follow-up dropouts (n = 15):

Lost to follow-up (n = 6) Lost to follow-up (n = 5) Lost to follow-up (n = 5)
Switched to non-PRX Switched to non-PRX Switched to non-PRX
antidepressants (n = 8) antidepressants (n = 16) antidepressants (n = 10)

Completed 4-wk follow-up Completed 4-wk follow-up Completed 4-wk follow-up

(n = 29) (n = 28) (n = 34)

Fig. 1. Flowchart of screening and recruitment of study subjects. PRX, paroxetine; MA, manual acupuncture; EA, electrical acupuncture.

was not significantly different (22.6% and 28.6 vs. 22.9%, c2 ¼ 0.650,
P ¼ 0.723) (Table 3).
3.3. Clinical outcomes of 4-week follow-up
At week 4 of follow-up, EA þ PRX group displayed a significantly
greater improvement on both HAMD-17 (P ¼ 0.010) and SDS
(P ¼ 0.024) compared to PRX alone; the significantly greater
improvement on SDS was evident when compared with MA þ PRX
group (P ¼ 0.028), but no significant differences were observed
between MA þ PRX and PRX alone groups (Table 2, Fig. 2).
3.4. Medication profile
The average doses of PRX taken in the three groups were similar
during 6 weeks of acupuncture intervention and during 4 weeks of
follow-up, with approximately 19e21 mg/day. The proportion of
patients who needed to increase PRX dose due to worsening of
symptoms in 6-week treatment was significantly greater in group
with PRX alone (22.9%) compared to groups with MA þ PRX (5.7%)
and EA þ PRX (8.9%, c2 ¼ 7.875, P ¼ 0.019) (Table 4). The proportion
of patients co-medicated with benzodiazepines and non-
benzodiazepines for insomnia was not significantly different

Table 1
Baseline characteristics of patients with major depression.

Variables PRX alone (n ¼ 48) MA þ PRX (n ¼ 54) EA þ PRX (n ¼ 58) P values

Female, n (%)a 29 (60.4) 31 (57.4) 25 (60.3) 0.152
Age (yrs)b 34.4  10.8 32.3  9.6 33.2  9.0 0.551
Duration of the illness (month)b 20.8  26.1 23.8  30.1 19.8  22.4 0.701
No. (%) of patients with first-onset MDDa 29 (60.4) 28 (51.9) 31 (53.4) 0.656
No. (%) of patients with family members having mental illnessesa 4 (8.3) 5 (9.3) 5 (8.6) 0.986
No. (%) of patients previously having acupuncture treatmenta 7 (14.6) 6 (11.1) 8 (13.8) 0.859
No. (%) of patients under antidepressants when entrya,c 9 (18.8) 8 (14.8) 11 (19.0) 0.815
Baseline HAMD-17 scoreb 23.3  4.6 25.1  5.6 25.0  5.2 0.160
Baseline CGI-Sb 4.1  0.8 4.5  1.0 4.3  1.2 0.197
Baseline SDS scoreb 48.3  7.1 51.8  9.0 51.1  7.0 0.063

PRX, paroxetine; MA, manual acupuncture; EA, electroacupuncture; HAMD-17, 17-item Hamilton Rating Scale for Depression; CGI-S, Clinical Global Impression-Severity; SDS,
Self-rating Depression Scale.
a
Categorical data were analyzed using Chi-square (c2) test.
b
Continuous data are expressed as mean  SD and examined using one-way analysis of variance (ANOVA).
c
Most patients were medicated with selective serotonin reuptake inhibitors (SSRIs). The use of antidepressants did not exceed one month.
730 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732

Table 2
Changes in score on depression scales from baseline in patients with major depression.

Variables PRX alone (n) MA þ PRX (n) EA þ PRX (n) Statistical valuesa

F P
During 6-wk treatment
HAMD-17 7.708 0.000
Week 1 1.4  4.3 (48) 3.9  4.6 (53)* 4.8  3.9 (56)*
Week 2 4.9  4.8 (48) 7.5  6.2 (53)* 8.4  4.5 (54)*
Week 4 8.8  5.3 (47) 11.7  7.2 (53)* 12.5  5.2 (52)*
Week 6 11.3  4.6 (43) 14.1  6.8 (51)* 15.7  5.1 (49)*
SDS 4.720 0.009
Week 1 1.8  4.7 (48) 4.3  8.4 (53) 5.1  5.9 (56)*
Week 2 3.5  6.0 (48) 7.9  9.9 (53)* 8.1  7.6 (54)*
Week 4 9.7  7.1 (47) 11.8  11.8 (53) 13.8  9.1 (52)*
Week 6 11.2  6.6 (43) 15.4  13.1(51)* 17.8  10.3 (49)*
CGI-S
Week 6 1.8  0.9 (43) 2.3  1.1 (51)* 2.4  0.9 (49)* 5.694 0.004
At wk 4 of follow-up
HAMD-17 13.1  3.8 (29) 14.8  5.5 (28) 17.1  6.1 (34)* 4.553 0.013
SDS 13.9  6.6 (29) 15.0  10.3 (28) 21.0  12.7 (34)*,# 4.255 0.017

PRX, paroxetine; MA, manual acupuncture; EA, electroacupuncture; HAMD-17, 17-item Hamilton Rating Scale for Depression; CGI-S, Clinical Global Impression-Severity; SDS,
Self-rating Depression Scale.
*P < 0.05 vs. PRX alone group, #P < 0.05 vs. MA þ PRX group.
a
Overall F and P values of HAMD-17 and SDS during 6-wk acupuncture treatment were obtained from a linear mixed-effect model analysis. Statistical analysis among the
three groups and post hoc between-group comparisons at each time point were performed with one-way analysis of variance (ANOVA).

among the three groups [PRX alone, 64.6% (31/48); MA þ PRX, disturbance. No significant differences in the incidence of any
64.2% (34/53); EA þ PRX, 60.7% (34/56), c2 ¼ 0.207, P ¼ 0.902]. adverse events were observed among the three groups; but 3 pa-
tients (1 in MA þ PRX and 2 in EA þ PRX) did not complete the first
3.5. Adverse events session of acupuncture treatment and discontinued due to fainting
of needling (Fig. 1).
Adverse events occurred in at least 5% of the patients in all of the
three groups are summarized in Table 5. The three most common 4. Discussion
adverse events were physical tiredness, headache, and sleep
In the present study, compared to PRX alone, we found that
0
adjuvant treatment by both MA and EA produced a significantly
PRX alone
MA + PRX A greater improvement in depression symptoms in both “subjective”
* EA + PRX (HAMD-17) and “objective” (SDS) measures at most assessment
-5 time points. Moreover, the greater improvement was present as
*
HAMD-17

* early as the first week and through the endpoint of treatment.

-10 * Similar results were also observed in another trial we recently re-
* ported (Zhang et al., 2012b), showing that EA combined with FLX
* resulted in the significantly greater improvement at the first week
-15
*
and most measure time points in 3-week treatment of patients
* with MDD. Greater clinical improvement at the endpoint of treat-
-20
* ment with adjuvant acupuncture in this study was also confirmed
using the CGI-S measure. The clinical responses to adjuvant
0 acupuncture were significantly greater than those with PRX alone,
B
although the remission rate did not differ across treatment regimes.
-5 The average doses of PRX taken were similar in the three groups;
* * however in groups treated with acupuncture, a significantly lower
* proportion of patients needed to increase PRX doses to address
SDS

-10
worsening symptoms. These results clearly indicate that the addi-
-15 * * tion of acupuncture in both manual and electrical stimulation
modes is effective in augmenting the antidepressant efficacy and
-20 * *# reducing the incidence of exacerbation of depression in the early
phase of SSRI treatment. An important drawback of treatment with
week 0 2 4 6 10
SSRIs and other antidepressant agents is the slow onset of thera-
PRX alone (n) 48 48 47 43 29 peutic action, which may prolong patients’ suffering and expose
MA+PRX (n) 53 53 53 51 28 them to great risk of exacerbation and suicide (Adell et al., 2005).
EA+PRX (n) 56 54 52 49 34
Thus, our evidence that adjuvant acupuncture accelerates the onset
Fig. 2. Mean changes from baseline in score on the 17-item Hamilton rating scale for of the therapeutic response is a critical observation.
depression (HAMD-17) and self-rating depression scale (SDS) in 6-week treatment and It is well documented that the delayed onset of therapeutic
4-week follow-up. Overall statistical significance of 6-week treatment was determined action of SSRIs is associated with a decrease in the net amount of
using the linear mixed-effect model analysis, followed by one-way analysis of variance the extracellular 5-HT due to the suppression of the brainstem 5-HT
(ANOVA) to detect between-group differences at each measure point. One-way ANOVA
was also used to detect between-group differences at week 4 of follow-up. *P < 0.05 vs.
neuronal cell firing and terminal release via the 5-HT1A/1B autor-
PRX alone group, #P < 0.05 vs. MA þ PRX group. PRX, paroxetine; MA, manual eceptor negative feedback in the initial phase of SSRI treatment
acupuncture; EA, electrical acupuncture. (Adell et al., 2005; Artigas et al., 1996; Blier, 2003).
 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732 731

Table 3 Table 5
Response and remission rates of depressed patients receiving acupuncture Adverse events occurred in at least 5% of patients in the three groups.
intervention.
Adverse event PRX alone MA þ PRX EA þ PRX P value
PRX alone MA þ PRX EA þ PRX P value (n ¼ 48) (n ¼ 54) (n ¼ 58) (c2 test)
(n ¼ 48, %) (n ¼ 53, %) (n ¼ 56, %) (c2 test) Physical tiredness 19 (39.6) 28 (51.9) 27 (46.6) 0.771
Response rate 20 (41.7) 37 (69.8) 39 (69.6) 0.004 Headache 11 (22.9) 11 (20.4) 5 (8.6) 0.103
Remission rate 11 (22.9) 12 (22.6) 16 (28.6) 0.723 Sleep disturbance 18 (37.5) 28 (51.9) 21 (36.2) 0.187
Vertigo 3 (6.3) 5 (9.3) 2 (3.4) 0.447
PRX, paroxetine; MA, manual acupuncture; EA, electroacupuncture.
Palpitations 4 (8.3) 2 (3.7) 4 (6.9) 0.608
Dry mouth 5 (8.6) 1 (1.9) 3 (5.2) 0.165
Constipation 5 (8.6) 4 (7.4) 4 (6.9) 0.782
Electrophysiological and immunohistochemical studies have Somnolence 3 (6.3) 3 (5.6) 0 0.167
confirmed that acupuncture, and in particular EA, can facilitate the Sexual problem 1 (2.1) 4 (7.4) 6 (10.3) 0.242
5-HT neuronal cell firing and stimulate 5-HT release at nerve ter- PRX, paroxetine; MA, manual acupuncture; EA, electroacupuncture.
minals (Kwon et al., 2000; Lee et al., 2004; Li et al., 2007).
Acupuncture is also effective in modulating norepinephrine- We note that several similar trials failed to demonstrate the su-
synthesizing neuronal cell activity in the locus coeruleus and the perior effects of acupuncture in the treatment of major depression
activity of brain regions involved in mood regulation, including the (Allen et al., 2006; Andreescu et al., 2011; Zhang et al., 2009). Two
hypothalamus, amygdala and the prefrontal cortex (Lee et al., 2004; possibilities could explain these negative results. First, trials with
Li et al., 2007; Zhang et al., 2012a). We therefore suggest that the negative results usually employed needling at non-meridian-based
rapid onset effects of acupuncture are likely explained by its direct acupoints as control acupoints, which are generally defined at a
facilitation of the brainstem 5-HT neuronal cell activity and its certain distances (usually 1e3 cm) from meridian-based acupoints
modulation of function in other key brain regions. (Allen et al., 2006; Andreescu et al., 2011; Zhang et al., 2009).
Our present study is first to show that, at 4 weeks of follow-up Although non-meridian-based acupoints are somewhat different in
after the completion of acupuncture treatment, patients treated histological profile from adjacent meridian-based acupoints, it might
with adjuvant EA, but not MA, continued to show a significantly be difficult to differentiate the clinical response to stimulation at the
greater improvement in both “subjective” and “objective” measures two types of acupoints (Zhang et al., 2012a,b). Second, most previous
compared to other treatment regimes. Thus, EA appears to produce trials did not consider quantitative and qualitative adequacy of acu-
a longer-lasting and more apparent enhancement of the antide- points and stimulation and sufficient treatment session frequency
pressant efficacy of SSRIs compared to MA. Similar results were also (Benham and Johnson, 2009). Insufficient “dosage” of acupuncture
observed in comparing the effects of EA and MA in treating tinnitus stimulation is thought to be an important factor determining
(Wang et al., 2010), pain (Sator-Katzenschlager et al., 2003, 2004; response to acupuncture trials (Benham and Johnson, 2009).
Schliessbach et al., 2011; Tsui and Leung, 2002), and modulating rat There are several limitations in the present study. First, the
neuroendocrine function (Feng et al., 2012). Like other brain stim- present study was an open trial in which acupuncturists and pa-
ulation therapies, such as electroconvulsive therapy (ECT) and re- tients were not blind to treatment conditions. However, the clinical
petitive transcranial magnetic stimulation (rTMS), the long-lasting assessors and psychiatrists rating response to treatment were blind
effects of EA also may be associated with its modification of syn- to treatment conditions. Moreover, the response of MA- and EA-
aptic plasticity at molecular and neurochemical levels (Kirkcaldie treated patients was distinct e we would have expected similar
et al., 1997). The treatment and follow-up results reported here follow-up outcomes if a similar perception of acupuncture influ-
suggest that EA is more efficacious in enhancing and maintaining enced clinical improvement. Second, similar to most previous
the therapeutic action of SSRIs than MA. studies (Zhang et al., 2010), the determination of acupoints used in
As SSRIs are known to have broad side effects which often result the present study was basically based on the doctrine of traditional
in discontinuation (Arroll et al., 2005), the current study did not find Chinese medicine (TCM) and empirical evidence. TCM-based indi-
significant differences in the incidence of adverse events between vidualized treatment regimens have resulted in a large variation in
patients treated with PRX alone and combination with acupuncture, acupuncture protocols and difficulties in comparing treatment
suggesting that acupuncture therapy could not cause additive outcomes among trials. Whether there are differences in the anti-
adverse effects. This, once again, confirms that acupuncture is depressant efficacy between the standardized and individualized
generally a tolerable and safe treatment modality (He et al., 2012; regimens remains to be further evaluated. Finally, although the
Lao et al., 2003). However, we noticed that there were 3 patients who present study demonstrated the antidepressant efficacy of
experienced fainting during needling and discontinued treatment. acupuncture therapy, the underlying mechanisms are not yet well
In acupuncture practice, it is not uncommon that needling stimu- delineated. Previous studies have suggested that the antidepres-
lation causes fainting (He et al., 2012; Lao et al., 2003). To avoid sant effects of acupuncture may be associated with the restoration
fainting of needling and other needling-related adverse effects, of decreased neuroimaging activity of brain regions involved in
great caution should be exercised, particularly on those who are first processing emotional signals (Zhang et al., 2012a). The determi-
time to receive acupuncture treatment or have chronic diseases. nation of neuroimaging correlates of the clinical improvement in

Table 4
Paroxetine doses taken during acupuncture intervention and follow-up.

PRX alone MA þ PRX EA þ PRX P value

Average dose during the 6-wk treatment (mg/day)a 20.8  4.3 (48) 20.2  3.7 (53) 20.3  4.1 (56) 0.274a
Average dose during the 4-wk follow-up (mg/day)a 20.7  5.3 (29) 18.9  3.2 (28) 20.3  1.7 (34) 0.160a
No. (%) of patients required to increase PRX dose due 11/48 (22.9) 3/53 (5.7) 5/56 (8.9) 0.019b
to the aggravation of symptoms during the 6-wk treatment

PRX, paroxetine; MA, manual acupuncture; EA, electroacupuncture.

a
Continuous data are expressed as mean  SD and examined using one-way analysis of variance (ANOVA). The number of patients is indicated in parenthesis.
b
Categorical data were analyzed using Chi-square (c2) test.
732 S.-S. Qu et al. / Journal of Psychiatric Research 47 (2013) 726e732
depressed patients would help gain some insights into antide-
pressant mechanisms of acupuncture.
Collectively, as most antidepressant agents have broad side ef-
fects, acupuncture in manual and electrical stimulation modes
provides a safe and effective treatment in augmenting the antide-
pressant efficacy and reducing the incidence of exacerbation of
depression in the early phase of SSRI treatment. Electrical
acupuncture even produced a long-lasting enhancement of the
antidepressant efficacy of SSRIs. We therefore recommend
acupuncture as an adjunctive therapy to accelerate the onset of the
antidepressant effect and clinical response.
Role of funding source
This study was supported by Key Project of the National
Eleventh-Five Year Research Program of China (2006BAI12B05-2,
Y.H.), Key Project of Phase III of Guangdong Provincial “211” Pro-
gram (Y.H.) and General Research Fund (GRF) of Research Grant
Council of HKSAR (786611, Z.J.Z.). All funding bodies had no role in
study design, data collection and processing, the preparation and
submission of the manuscript.
Contributors
YH and ZJZ were involved in conception and design of the study.
ZJZ, SSQ and YH conducted final data analyzes and drafted the
manuscript. JQC, RYL, CQW, GLL performed patients recruitment
and acupuncture treatment. CHZ, JYP, SCG, and YCZ performed
conventional treatment and clinical assessments. HKW performed
statistical analysis.
Conflict of interest
All authors declare no any conflict of interest in this study.
Acknowledgments
We thank Dr. Grainne M. McAlonan for critically reading the
paper.
References
Adell A, Castro E, Celada P, Bortolozzi A, Pazos A, Artigas F. Strategies for producing
faster acting antidepressants. Drug Discovery Today 2005;10:578e85.
Allen JJ, Schnyer RN, Chambers AS, Hitt SK, Moreno FA, Manber R. Acupuncture for
depression: a randomized controlled trial. Journal of Clinical Psychiatry 2006;
67:1665e73.
Andreescu C, Glick RM, Emeremni CA, Houck PR, Mulsant BH. Acupuncture for the
treatment of major depressive disorder: a randomized controlled trial. Journal
of Clinical Psychiatry 2011;72:1129e35.
Arroll B, Macgillivray S, Ogston S, Reid I, Sullivan F, Williams B, et al. Efficacy and
tolerability of tricyclic antidepressants and SSRIs compared with placebo for
treatment of depression in primary care: a meta-analysis. The Annals of Family
Medicine 2005;3:449e56.
Artigas F, Romero L, de Montigny C, Blier P. Acceleration of the effect of selected
antidepressant drugs in major depression by 5-HT1A antagonists. Trends in
Neurosciences 1996;19:378e83.
Asberg M, Cronholm B, Sjoqvist F, Tuck D. Correlation of subjective side effects with
plasma concentrations of nortriptyline. British Medical Journal 1970;4:18e21.
Benham A, Johnson MI. Could acupuncture needle sensation be a predictor of
analgesic response? Acupuncture in Medicine 2009;27:65e7.
Blier P. The pharmacology of putative early-onset antidepressant strategies. Euro-
pean Neuropsychopharmacology 2003;13:57e66.
Calvo G, García-Gea C, Luque A, Morte A, Dal-Ré R, Barbanoj M. Lack of pharma-
cologic interaction between paroxetine and alprazolam at steady state in
healthy volunteers. Journal of Clinical Psychopharmacology 2004;24:268e76.
Chung KF, Yeung WF, Zhang ZJ, Yung KP, Man SC, Lee CP, et al. Randomized non-
invasive sham-controlled pilot trial of electroacupuncture for postpartum
depression. Journal of Affective Disorders 2012;142:115e21.
Fang Y, Yuan C, Xu Y, Chen J, Wu Z, Cao L, et al. A pilot study of the efficacy and safety
of paroxetine augmented with risperidone, valproate, buspirone, trazodone, or
thyroid hormone in adult Chinese patients with treatment-resistant major
depression. Journal of Clinical Psychopharmacology 2011;31:638e42.
Fang Y, Yuan C, Xu Y, Chen J, Wu Z, Cao L, et al. Comparisons of the efficacy and
tolerability of extended-release venlafaxine, mirtazapine, and paroxetine in
treatment-resistant depression: a double-blind, randomized pilot study in a
Chinese population. Journal of Clinical Psychopharmacology 2010;30:357e64.
Feng Y, Johansson J, Shao R, Mannerås-Holm L, Billig H, Stener-Victorin E. Electrical
and manual acupuncture stimulation affect oestrous cyclicity and neuroendo-
crine function in an 5a-dihydrotestosterone-induced rat polycystic ovary syn-
drome model. Experimental Physiology 2012;97:651e62.
Ferrari AJ, Somerville AJ, Baxter AJ, Norman R, Patten SB, Vos T, et al. Global variation
in the prevalence and incidence of major depressive disorder: a systematic re-
view of the epidemiological literature. Psychological Medicine 2012;25:1e11.
Gibiino S, Serretti A. Paroxetine for the treatment of depression: a critical update.
Expert Opinion on Pharmacotherapy 2012;13:421e31.
Hamilton M. A rating scale for depression. Journal of Neurology, Neurosurgery &
Psychiatry 1960;23:56e62.
Han JS. Acupuncture: neuropeptide release produced by electrical stimulation of
different frequencies. Trends in Neurosciences 2003;26:17e22.
He W, Zhao X, Li Y, Xi Q, Guo Y. Adverse events following acupuncture: a systematic
review of the Chinese literature for the years 1956e2010. Journal of Alternative
and Complementary Medicine 2012;18:892e901.
Kirkcaldie M, Pridmore S, Reid P. Bridging the skull: electroconvulsive therapy (ECT)
and repetitive transcranial magnetic stimulation (rTMS) in psychiatry.
Convulsive Therapy 1997;13:83e91.
Kwon YB, Kang MS, Son SS, Kim JT, Lee YH, Han HJ, et al. Different frequencies of
electroacupuncture modified the cellular activity of serotonergic neurons in
brainstem. The American Journal of Chinese Medicine 2000;28:435e41.
Lao L, Hamilton GR, Fu J, Berman BM. Is acupuncture safe? A systematic review of
case reports. Alternative Therapies in Health and Medicine 2003;9:72e83.
Lee HC, Chiu HF, Wing YK, Leung CM, Kwong PK, Chung DW. The Zung Self-rating
Depression Scale: screening for depression among the Hong Kong Chinese
elderly. Journal of Geriatric Psychiatry and Neurology 1994;7:216e20.
Lee HJ, Lee B, Choi SH, Hahm DH, Kim MR, Roh PU, et al. Electroacupuncture reduces
stress-induced expression of c-fos in the brain of the rat. The American Journal
of Chinese Medicine 2004;32:795e806.
Li A, Wang Y, Xin J, Lao L, Ren K, Berman BM, et al. Electroacupuncture suppresses
hyperalgesia and spinal Fos expression by activating the descending inhibitory
system. Brain Research 2007;1186:171e9.
Manber R, Schnyer RN, Lyell D, Chambers AS, Caughey AB, Druzin M, et al.
Acupuncture for depression during pregnancy: a randomized controlled trial.
Obstetrics & Gynecology 2010;115:511e20.
Rigucci S, Serafini G, Pompili M, Kotzalidis GD, Tatarelli R. Anatomical and func-
tional correlates in major depressive disorder: the contribution of neuro-
imaging studies. World Journal of Biological Psychiatry 2010;11:165e80.
Sator-Katzenschlager SM, Scharbert G, Kozek-Langenecker SA, Szeles JC, Finster G,
Schiesser AW, et al. The short- and long-term benefit in chronic low back pain
through adjuvant electrical versus manual auricular acupuncture. Anesthesia &
Analgesia 2004;98:1359e64.
Sator-Katzenschlager SM, Szeles JC, Scharbert G, Michalek-Sauberer A, Kober A,
Heinze G, et al. Electrical stimulation of auricular acupuncture points is more
effective than conventional manual auricular acupuncture in chronic cervical
pain: a pilot study. Anesthesia & Analgesia 2003;97:1469e73.
Schliessbach J, van der Klift E, Arendt-Nielsen L, Curatolo M, Streitberger K. The
effect of brief electrical and manual acupuncture stimulation on mechanical
experimental pain. Pain Medicine 2011;12:268e75.
Smith CA, Hay PP, Macpherson H. Acupuncture for depression. Cochrane Database
of Systematic Reviews 2010;20. CD004046.
Sproule BA, Naranjo CA, Brenmer KE, Hassan PC. Selective serotonin reuptake in-
hibitors and CNS drug interactions. A critical review of the evidence. Clinical
Pharmacokinetics 1997;33:454e71.
Tsui P, Leung MC. Comparison of the effectiveness between manual acupuncture
and electro-acupuncture on patients with tennis elbow. Acupuncture and
Electro-therapeutics Research 2002;27:107e17.
Wang K, Bugge J, Bugge S. A randomised, placebo-controlled trial of manual and
electrical acupuncture for the treatment of tinnitus. Complementary Therapies
in Medicine 2010;18:249e55.
Ward MP, Irazoqui PP. Evolving refractory major depressive disorder diagnostic and
treatment paradigms: toward closed-loop therapeutics. Frontiers in Neuro-
engineering 2010;3:7.
Wu J, Yeung AS, Schnyer R, Wang Y, Mischoulon D. Acupuncture for depression: a
review of clinical applications. Canadian Journal of Psychiatry 2012;57:397e405.
Zhang WJ, Yang XB, Zhong BL. Combination of acupuncture and fluoxetine for
depression: a randomized, double-blind, sham-controlled trial. Journal of
Alternative and Complementary Medicine 2009;15:837e44.
Zhang ZJ, Chen HY, Yip KC, Ng R, Wong VT. The effectiveness and safety of
acupuncture therapy in depressive disorders: systematic review and meta-
analysis. Journal of Affective Disorders 2010;124:9e21.
Zhang ZJ, Ng R, Man SC, Li TY, Wong W, Tan QR, et al. Dense cranial electro-
acupuncture stimulation for major depressive disorderea single-blind, ran-
domized, controlled study. PLoS One 2012b;7:e29651.
Zhang ZJ, Wang XM, McAlonan GM. Neural acupuncture unit: a new concept for
interpreting effects and mechanisms of acupuncture. Evidence-based Comple-
mentary and Alternative Medicine 2012a;2012. 429412.
Zhao ZQ. Neural mechanism underlying acupuncture analgesia. Progress in
Neurobiology 2008;85:355e75.