Effects of Long-Term Acupuncture Treatment on Resting

13 Pages·2014·1.19 MB·English

Effects of Long-Term Acupuncture Treatment on Resting-
State Brain Activity in Migraine Patients: A Randomized
Controlled Trial on Active Acupoints and Inactive
Acupoints
Ling Zhao1, Jixin Liu2*, Fuwen Zhang3, Xilin Dong1, Yulin Peng1, Wei Qin2, Fumei Wu1, Ying Li1,
Kai Yuan2, Karen M. von Deneen2, Qiyong Gong4, Zili Tang5, Fanrong Liang1*
1Acupuncture and Tuina School, Chengdu University of Traditional Chinese Medicine, Chengdu, Sichuan, China, 2School of Life Science and Technology, Xidian
University,Xi’an,Shaanxi,China,3SchoolofClinicalMedicine,ChengduUniversityofTraditionalChineseMedicine,Chengdu,Sichuan,China,4DepartmentofRadiology,
TheCenterforMedicalImaging,HuaxiMRResearchCenter,WestChinaHospitalofSichuanUniversity,Chengdu,Sichuan,China,5GermanCancerConsortium(DKTK),
Heidelberg,Germany
Abstract
Background:Acupuncturehasbeencommonlyusedforpreventingmigraineattacksandrelievingpainduringamigraine,
althoughthereislimitedknowledgeonthephysiologicalmechanismbehindthismethod.Theobjectivesofthisstudywere
to compare the differences in brain activities evoked by active acupoints and inactive acupoints and to investigate the
possible correlationbetween clinicalvariables and brainresponses.
MethodsandResults:Arandomizedcontrolledtrialandresting-statefunctionalmagneticresonanceimaging(fMRI)were
conducted. A total of eighty migraineurs without aura were enrolled to receive either active acupoint acupuncture or
inactiveacupointacupuncturetreatmentfor8weeks,andtwentypatientsineachgroupwererandomlyselectedforthe
fMRI scan at the end of baseline and at the end of treatment. The neuroimaging data indicated that long-term active
acupoint therapy elicited a more extensive and remarkable cerebral response compared with acupuncture at inactive
acupoints. Most of the regions were involved in the pain matrix, lateral pain system, medial pain system, default mode
network, and cognitive components of pain processing. Correlation analysis showed that the decrease in the visual
analoguescale(VAS)wassignificantlyrelatedtotheincreasedaverageRegionalhomogeneity(ReHo)valuesintheanterior
cingulatecortexinthetwogroups.Moreover,thedecreaseintheVASwasassociatedwithincreasedaverageReHovalues
inthe insulawhichcould bedetected inthe activeacupoint group.
Conclusions:Long-termactiveacupointtherapyandinactiveacupointtherapyhavedifferentbrainactivities.Wepostulate
thatacupunctureattheactiveacupointmighthavethepotentialeffectofregulatingsomedisease-affectedkeyregionsand
the paincircuitry for migraine, andpromote establishing psychophysical pain homeostasis.
TrialRegistration:ChineseClinical TrialRegistry ChiCTR-TRC-13003635
Citation:ZhaoL,LiuJ,ZhangF,DongX,PengY,etal.(2014)EffectsofLong-TermAcupunctureTreatmentonResting-StateBrainActivityinMigrainePatients:A
RandomizedControlledTrialonActiveAcupointsandInactiveAcupoints.PLoSONE9(6):e99538.doi:10.1371/journal.pone.0099538
Editor:MarioD.Cordero,UniversityofSevilla,Spain
ReceivedSeptember28,2013;AcceptedMay14,2014;PublishedJune10,2014
Copyright: (cid:2) 2014 Zhao et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricteduse,distribution,andreproductioninanymedium,providedtheoriginalauthorandsourcearecredited.
Funding:ThisstudywassupportedbytheNationalBasicResearchProgramofChina(973Program,No.2012CB518501),NationalNaturalScienceFoundationof
China(Nos.30901900,30930112,81101108),theProjectofAdministrationofTraditionalChineseMedicineofSichuanProvince(No.2012-E-038),andtheProject
ofInnovativeResearchTeamResearchFundofSichuanProvincialEducationDepartment(No.12TD002).Thefundershadnoroleinstudydesign,datacollection
andanalysis,decisiontopublish,orpreparationofthemanuscript.
CompetingInterests:Theauthorshavedeclaredthatnocompetinginterestsexist.
*E-mail:[email protected](FL);[email protected](JL)
Introduction mechanism of migraine is still unclear, there is plenty of
neuroimagingevidenceshowingthatmigraineisacentralnervous
Migraine is a common neurological disorder that typically
system disorder [4–6]. Our research group involving migraine
manifests as repeated episodes of moderate or severe unilateral,
without aura patients showed that abnormal structure and
pulsating headache aggravated by routine physical activity and is
functionwaspossiblyassociatedwithanimpairedpainprocessing
associatedwithnauseaand/orphonophobiaandphotophobia[1].
and modulatory process, such as in the anterior cingulate cortex
Migraine has attracted more and more attention worldwide as a
(ACC),insula,basalganglia,thalamus,supplementarymotorarea
publichealthissuebecauseofitshighprevalence,frequentattack
(SMA),prefrontal cortex, etc.[7–9].
history, significant medical burden, and a serious reduction in
AcupuncturehasalonghistoryinChinaasoneofthetreatment
quality of life (QOL) and productivity [2,3]. Although the exact
modalities of Traditional Chinese Medicine (TCM) and is
PLOSONE | www.plosone.org 1 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
increasingly being adopted in the West as a complementary and 2000),andwasapprovedbytheethicscommitteeattheTeaching
alternative treatment to prevent migraine attacks and to relieve Hospital of Chengdu University of TCM. Based on the previous
pain during a migraine. The latest Cochrane meta-analysis reportaboutminimumsamplesizeinneuroimagingstudies[22],a
suggests that acupuncture as a migraine prophylaxis is safe and samplesizeof16pergroupwasneeded(totalN=32).Considering
at least as effective, if not more effective, than prophylactic drug a conservative dropout rate of 25%, a total sample size of 40
treatment[10].Duringthepastdecade,aconsiderablenumberof migraineurs was determined. However, during the period of
highqualityclinicalstudieshaveindicatedthatacupunctureisable recruitment,alargenumberofeligiblemigrainepatients(farmore
to alleviate headache degree and/or improve the QOL [11–13]. than the originally planned sample size) were willing to receive
However, despite the popularity of acupuncture in migraine acupuncture treatment. According to a previous study [11], the
therapy, there persists limited knowledge on the physiological differenceinmeanscoreofVASbetweentheacupuncturegroup
mechanisms behind this method, and some controversy on the and sham acupuncture group at 8 weeks was 1 (d=1). For this
superiorityofverumacupunctureovershamcontrol.Somestudies study, it was determined prospectively that a=0.05(two-sided),
suggested that the obvious influence of acupuncture on pain 1-b=0.9,andthatthestandarddeviationwouldbe1.2according
symptoms was eitherinsignificant or a placebo effect[12,14]. tothetwogroupsubsets.Thirty-oneparticipantswererequiredfor
With the development of neuroimaging techniques, this has eachgroup(1:1allocation).Thus,wedecidedtoenrollatotalof80
providedabrandnewviewtoexplorethecentral mechanismsof participants(afterattrition)andrandomlyselected40migraineurs
acupuncture,andhasbeenaglobaltrendinacupunctureresearch. toimplement thefMRIexperiment.
We detected cerebral glucose metabolism after short periods of
acupuncturestimulationinmigraineursthroughpositronemission 2.1 Participants
tomography(PET)withcomputedtomographyexamination,and All subjects gave written, informed consent after the experi-
found that transientapp:addword:transient acupuncture stimula- mental procedures had been fully explained. Subjects were
tion induced different levels of cerebral glucose metabolism in enrolledfromtheneurologydepartmentoftheTeachingHospital
some pain-related brain regions [15]. In fact, one session of of Chengdu University of TCM. Recruitment took place June
acupuncturestimulationdidnotfullymodeltheclinicalsituation, 2012 through March 2013. The diagnosis of migraine without
and was hardly enough to achieve the expected effect in clinical aura was established according to the classification criteria of the
practice.Therefore,thecumulativetherapeuticeffectoflong-term International Headache Society (IHS) [1]. The inclusion criteria
acupuncture would help to reveal the underlying mechanisms of wereasfollows:(1)allsubjectswereright-handed andhad2to6
acupuncture treatment inmoredepth. migraine attacks per month during the last 3 months and during
In the current study, we performed a ReHo approach [16] to the baseline period (4 weeks before enrollment); (2) all subjects
comparethebloodoxygenlevel-dependent(BOLD)signalsinthe were 18 to 55 years of age; in addition, the start of headache
brainsofmigrainepatientsduringtheresting-state.ReHoisbased neededtobebeforetheageof50;(3)receivededucationformore
on a data-driven approach andthus requires no prior knowledge than 6 years and completed the baseline headache diary; (4) had
andhasgoodtest-retestreliability[17].Itwasoriginallyproposed nottakenanyprophylacticheadachemedicineoranyacupuncture
formeasuringthedegreeofregionalsynchronizationoffunctional treatment during the last 3 months; (5) no record of long-term
magnetic resonance imaging (fMRI) time courses and focused on analgesics consumption; and (6) had no contraindications to
thesimilaritiesorcoherenceoftheintraregionalspontaneouslow- exposuretoahighmagneticfield.Generalexclusioncriteriawere:
frequency (,0.08Hz) BOLD signal, which enables a novel (1) existence of neurological diseases; (2) hypertension, diabetes
perspective to understand the functional regulation in particular mellitus, hypercholesteremia, vascular/heart disease, and major
brainregions.AnimportantadvantageofusingtheReHomethod systemicconditions;(3)pregnantorlactatingwomen;(4)alcoholor
overother methodsisthat itdetects changesormodulations that drug abuse; (5) any neuroimaging research study participation
are induced by different conditions across the whole brain in a during the last 6 months; and (6) inability to understand the
voxel-by-voxel manner. ReHo analysis has been used to study doctor’s instructions.
migraine inour group[7,18], andother diseaseslike Alzheimer’s
disease [19], Parkinson’s disease [20], attention-deficit/hyperac- 2.2 Study Design
tivity disorder [21], andsoon. We performed a single-blind, randomized controlled trial with
We hypothesized that if acupuncture therapy is effective, it two groups: active acupuncture group and inactive acupuncture
would modulate disease-affected brain regions and dysfunctional group. The primary objective of this study was to compare the
pain modulatory circuitry in migraine patients. In the current difference in resting-state brain activation patterns evoked by
study, a randomized controlled trial and resting-state fMRI were active acupoints and inactive acupoints in migraine patients via
adopted to compare the difference in brain activation patterns fMRI assessment. The secondary objective was to investigate the
evoked by active acupoints and inactive acupoints for migraine possiblecorrelationbetweenbrainresponsesandclinicalefficacy.
patients. Furthermore, a correlation analysis was performed to The total observation period within this study was 12 weeks for
investigate the possible correlation between clinical variables and each patient, including a baseline period of 4 weeks, and a
brain activity. treatment period of 8 weeks. Headache diaries were given to
recruited patients to record the details of migraine attacks for 4
Methods weeks(24to0weeks)duringthebaselineperiod.Aftertheinitial
assessment and screening, patients who met the inclusion criteria
TheprotocolforthistrialandsupportingCONSORTchecklist were randomly assigned into the active acupoint group or the
are available as supporting information; see Checklist S1 and inactive acupoint group in a 1:1 ratio. All patients were asked to
ProtocolS1.ThistrialwasperformedattheTeachinghospitalof document their headache diaries, and the outcome measurement
Chengdu University of TCM. The study protocol was registered was completed both for the baseline, 4 and 8 weeks after
with the Chinese Clinical Trial Registry (ChiCTR) (Identifier: randomization. Additionally, 20 migraineurs in each group were
ChiCTR-TRC-13003635).Thestudywasperformedaccordingto randomly selected to receive fMRI examinations at the end of
the principles of the Declaration of Helsinki (Edinburgh version, baseline andat theendof thetreatment period respectively.
PLOSONE | www.plosone.org 2 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
2.3 Randomization patients in both groups: they were informed that they would
Randomizationnumbersof80patientsweregeneratedthrough receive one of two types of acupuncture treatment, which
computerized block-randomization with the SAS procedure depended on different traditional Chinese acupuncture theories;
PROC PLAN in the SAS package (SAS Version 9.0, SAS acupuncturetreatmentwasachievedinalargeindependentsingle-
Institute, Inc., Cary, NC) by an independent statistician. In this room with screen dividers for patient blinding and privacy; and
study,theblocksizewassetto4,andthenumberofblockswas20. twogroupsofpatientsreceivedbilateralandequivalentnumberof
Opaque,sealedenvelopeswithconsecutivenumberswereusedfor acupoint stimulations eachtime.
allocation concealment. Investigators who selected the eligible
participants after baseline screening opened the envelopes 2.6 Outcome Measures in Clinical Efficacy
according to the patients’ screening sequence numbers, and Allpatientswererequiredtofilloutheadachediaryrecordsfor
placed the patients into either the active group or the inactive 12 weeks, including a 4-week baseline, and 4 and 8 weeks after
group. Additionally, we used Microsoft Excel’s sampling tool to randomization. The headache diary recorded the severity,
generate arandomsample of20from40eligiblemigraineursfor frequency and duration of headache according to the guidelines
eachgroup.Inthenewrandomlist,eachnumberrepresentedthe oftheIHSforClinicalTrialsinMigraine[25].VASscore0–10as
enrolledsequencenumberinthesubgroup.Next,thecorrespond- aprimaryclinicaloutcomemeasuredtheintensityofheadache.As
ing patients wereselected toperformthefMRI scans. secondary clinical outcome measures, the number of days with a
migraine per 4 weeks and frequency of migraines per 4 weeks
2.4 Intervention (definedasthenumberofmigraineseparatedbypainfreeintervals
In this study, traditional Chinese style acupuncture was used of at least 48hours) measured the duration and severity of
andtreatmentsweremanipulated bytwospecializedacupunctur- headache respectively. In addition, the HIT-6 questionnaire [26]
ists with at least five years of training and three years of was adopted to assess the severity and impact of headache on a
experience. They received special training prior to the study to patient’s life.
ensure they had consistent manual acupuncture therapy. The It is worth mentioning that the feelings of de-qi were collected
trainingprogramincludedsomestandardoperationprocedureson afterremovingneedlesduringthe8th,16th,24th,and32ndsessions.
the locations of the acupoints, acupuncture manipulation tech- Migraineurs were interviewed by an acupuncturist who did not
niques, and so on. They implemented acupuncture therapy in know the treatment allocation. Patients were asked to evaluate
bothgroupsbyturns.Theactiveacupuncturepointswereselected each component of the de-qi sensations they had experienced
according to traditional classic and systematic reviews of ancient duringtheacupuncturestimulationperiod,andtheintensityused
and modern literature of acupuncture for migraine upon several a VAS ranging from 0 (none) to 10 (max), which has been
consensusmeetingswithexpertsbasedontheexperiencefromour commonlyusedtomeasurethefeelingsofde-qi[27,28].Thescore
previous study [11,23]. Moreover, the control group was given fortheVASwasthesumofallcomponentscores.Theoverallde-qi
inactiveacupointswhichwerechosenaccordingtotheiranatom- score was themeanscore fromall sessions.
ical locations, corresponding to Chinese meridians, proximity to
verum acupoints and role in treating diseases [24]. The active 2.7 fMRI data acquisition
treatment (group A) was performed on bilateral SJ5 (Waiguan), Resting-statefMRIscanswereperformedoneachgroupatthe
GB20 (Fengchi), GB34 (Yanglingquan), and GB40 (Qiuxu); and baselineandafter8weeks’treatmenttodetectthelocalfeaturesof
the inactivecontrol (group B) was implemented on bilateral SJ22 spontaneousbrainactivity.Theimagingdatawerecarriedoutina
(Erheliao), PC7 (Daling), GB37 (Guangming), and SP3 (Taibai) 3 Tesla Siemens MRI system (Allegra, Siemens Medical System,
(figure 1). Erlangen, Germany) at the Huaxi MR Research Center, West
All acupoints were punctured bilaterally using single-use China Hospital of Sichuan University, Chengdu, China. A
stainless steel filiform needles (Hwato Needles, Sino-foreign Joint standard eight-channel phase-array head coil was used, along
Venture Suzhou Hua Tuo Medical Instruments Co., China), with restraining foam pads to minimize head motion and to
25mm–40 mm in length and 0.25mm–0.30mm in diameter. diminish scanner noise. The resting-state functional images were
Thedepthsoftheinsertedneedlesdifferedbutwereapproximately obtainedwithecho-planarimaging(EPI)(30continuoussliceswith
2.5 cm–3.5cm. Needles were twisted with rotation (90u,ampli- a slice thickness=5 mm, repetition time=2000ms, echo
tude,180u) at a frequency of 1–2Hz. Stimulation was repeated time=30ms, flip angle=90u, field of view=240mm6240mm,
1–3 times to acquire the de-qi sensation (‘‘de-qi sensation’’ is a matrix=64664). During the 6-min fMRI scanning, participants
complex feeling including soreness, numbness, heaviness, disten- were instructed to keep their eyes closed, relax, move as little as
tion and dull pain at the site of needle placement). Each group’s possible, and stay awake. It needs to be emphasized that if there
treatmentconsistedof32sessionsofacupunctureoveraperiodof was an attack for migraine patients during the scan or
8 weeks (once every other day, preferably 4 times a week), and examination, they could not be scanned and the scan would be
each sessionlasted 30minutes. postponed. In this study, records in the headache diary were
checked to ensure every patient did not suffer from a migraine
2.5 Blinding attack at least 72hoursprior tothebrain scan.
Due to the procedure of the acupuncture technique, two
acupuncturists in this study were not blinded. Investigators in 2.8 Data Analysis
chargeofpatientscreeningandrandomizeddistributionwerenot 2.8.1 Clinical data analysis. The statistical analysis was
involved in treatment and data analyses. They knew the group performed by an independent statistician blinded to treatment
assignment, but they did not know the corresponding treatment allocation in the Teaching Hospital of Chengdu University of
schedule. The outcome assessor, who was not involved in TCM.SPSSstatisticalpackageprogram(Version14.0,SPSSInc.,
acupuncturetreatmentanddataanalyses,wasblindedthroughout Chicago, IL, USA) was used. Baseline characteristics and clinical
thestudy. outcomeswereanalyzedbytheintention-to-treat(ITT)population
To guarantee that the patients were blinded during the which included all participants who had randomized allocation.
treatmentperiod,severalapproacheswereperformedformigraine Missingdataofdropped-outparticipantswerereplacedbythelast
PLOSONE | www.plosone.org 3 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
Figure1.Locationsofactiveacupointsandinactiveacupoints.Theactiveacupointswerelocatedasfollows:SJ5,onthedorsalaspectofthe
forearmonthelineconnectingSJ4andthetipoftheelbow,2cunabovethetransversecreaseofthewristbetweentheulnaandradius;GB20,ina
depressionbetweentheupperportionofthesternocleidomastoidmuscleandthetrapezius;GB34,onthelateralaspectofthelowerleginthe
depressionanteriorandinferiortotheheadofthefibula;GB40,anteriorandinferiortotheexternalmalleolusinadepressiononthelateralsideof
thetendonoftheextensordigitorumlongus.Theinactiveacupointswerelocatedasfollows:SJ22,onthesideoftheheadontheposteriorborderof
thehairlineofthetempleatthelevelwiththerootoftheauricle,posteriortothesuperficialtemporalartery;PC7,inthemiddleofthetransverse
creaseofthewristbetweenthetendonsofthepalmarislongusandflexorcarpiradialis;GB37,onthelateralaspectofthelowerleg5cunabovethe
tipoftheexternalmalleolusontheanteriorborderofthefibula;SP3,proximalandinferiortotheheadofthe1stmetatarsal-phalangealjointina
depressionatthejunctionoftheredandwhiteskin.
doi:10.1371/journal.pone.0099538.g001
observationcarriedforward(LOCF)method.Thesignificantlevel first image of each session for motion correction and spatially
used for the statistical analysis with 2-tailed testing was 5%. normalizedtothestandardMontrealNeurologicalInstitute(MNI)
Continuous variables were presented as the mean (standard template in SPM5. We calculated the maximum excursion
deviation) with 95% confidence intervals (CI). Categorical movement values for each of the translation planes (x, y, and z)
variablesweredescribedasn(percentage).Treatmenteffectssuch and each of the rotation planes (roll, pitch, and yaw) for every
asVAS,frequencyofmigraineattackper4weeks,numberofdays participant.Noneofthemhadheadmovementsexceeding1 mm
with migraine per 4 weeks, and HIT-6 were evaluated using a on any axis and head rotation greater than 1u during the entire
repeated-measures analysis of variance (ANOVA) model with a fMRI scan. Finally, a band-pass filter (0.01Hz,f,0.08Hz) was
between-subjectsfactorTherapy(levels:activeandinactive)anda applied toremovephysiological andhigh-frequency noise.
within-subjects repeated measures factor TIME (levels: baseline, 2.8.3 MRI data analysis. ReHo, a method proposed by
1–4 weeks, and 5–8 weeks). For the change in VAS, analysis of Zang et al. [29], was performed in the Resting-state fMRI Data
covariancewithbaselineVASasacovariatewasusedtocompare Analysis Toolkit (http://www.restfmri.net) [16]. First, the Ken-
the difference between two groups at the end of treatment. The dall’s coefficient of concordance (KCC) of each voxel was
generallinearmodelrepeatedmeasuresprocedurewasusedtotest calculated by the time series of the voxel and its nearest 26
the differences in the repeated continuous variables (de-qi neighboring voxels (cluster size=27). Second, the KCC maps
sensations) between thetwogroups. were standardized by their own mean KCC within the whole
2.8.2Imagingdatapreprocessing. Inthefunctionalimage brain mask. Third, the resulting maps were smoothed with a
datapreprocessing,thefirstfivescanswerediscardedtoeliminate Gaussian kernel with a full-width at half-maximum (FWHM) of
nonequilibrium effects ofmagnetization andtoallow participants 4 mm.Inthestatisticalanalysis,anindependent-samplet-testwas
to become familiar with the scanning circumstances. Data usedtoexploreReHodifferencesbetweenthetwogroupswithage
preprocessing was done using Statistical Parametric Mapping asacovariate.Resultswereassumedtobestatisticallysignificantat
(SPM5, http://www.fil.ion.ucl.ac.uk/spm). The images were P,0.05 after false discovery rate (FDR) correction within the
corrected for the acquisition delay between slices, aligned to the whole brain. The correlation analysis was performed based on
PLOSONE | www.plosone.org 4 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
differentclustersinthebrainafteracupuncturetreatmentsrelative thattherewasnosignificantdifferencebetweenthetwogroupsin
to the baseline for each group. Within each cluster, we extracted de-qi sensations (P.0.05)(table 5).
theReHovaluesafteracupunctureandbaselinerespectively.The
mean of their subtraction (end of treatment-baseline) was 3.4 Correlation coefficients of the brain response and
correlated with the changes in the clinical variables, and clinical variable results
Bonferroni correction was used. Intheactiveacupointgroup,thedecreaseintheVASscorewas
significantly related to the increased average ReHo values in the
Results ACC (r=20.6619, P,0.05) and insula (r=20.7407, P,0.05,
Bonferronicorrected).Intheinactivecontrolgroup,thedecrease
3.1 Participants
in the VAS score was only significantly related to the increased
Eighty eligible patients were equally allocated into the active average ReHo values in the ACC (r=20.6611, P,0.05,
treatmentgroupandinactivetreatmentgroup(40ineachgroup). Bonferroni corrected) (figure 5).
Twopatientsfromtheactiveacupuncturegroupandfivefromthe
inactiveacupuncturegroupdroppedoutduringthestudybecause 3.5 Adverse events
of private reasons: noncompliance with treatment schedule or
Noseriousadverseeventshappenedduringthestudy.Onecase
inability to be contacted (figure 2). In total, 80 patients who in the active acupuncture group suffered acupuncture fainting
receivedacupuncturetherapieswereincludedintheITTanalysis during acupuncture treatment. The patient was told to lie down
oftheclinicaloutcomemeasures.Thebaselineanddemographics andrest.Thesymptomsofdizzinessandsweatingdisappearedin
withtheITTpopulationareshownintable1,whichshowedthat 15 minutes. Two cases in each group reported having minor
the two groups were comparable at baseline. Furthermore, 40 hemorrhageattheneedlingsite.Theyweretoldtoputpressureon
patients (20 in each group) finished the fMRI scans, and the the needling areas for about 5 minutes, and recovered in a short
baseline characteristics did not differ between the two groups time. All of the patients with adverse events completed the study
(table 2). process.
3.2 Neuroimaging results Discussion
In the active acupoint group, migraine patients showed
In this RCT study, we focused on the difference in brain
significantlyhigherReHovaluesinthebilateralACC(Brodmman
activation patterns evoked by active acupoints and inactive
area (BA) 24, BA32), insula (BA13), thalamus, SMA (BA6),
acupoints in migraine patients via fMRI assessment, and
superior temporal gyrus (STG) (BA22), cuneus (BA17, BA18),
determined the potential physiological mechanism behind this
lingualgyrus(BA18),cerebellum,andbrainstemafteracupuncture
therapy.Aninactiveacupointisavalidatedshamcontrolmethod
treatment. A decrease in ReHo values was observed after
in acupuncture research [30–32] with the advantage of minimiz-
treatment in the bilateral posterior cingulate cortex (PCC)
ing bias from patients. By the way, non-acupoints are thought to
(BA31), middle frontal gyrus (MFG) (BA10), angular gyrus
havenotherapeuticinfluenceandareusuallyadoptedasaplacebo
(BA39), precuneus (BA7), middle temporal gyrus (MTG) (BA39),
control in previous clinical trials and neuroimaging studies. In
lefthippocampus,inferiorparietallobule(BA39),inferiortemporal
addition, minimal acupuncture or superficial insertion was often
gyrus (ITG) (BA20), and right postcentral gyrus (BA40) (P,0.05,
employedtostimulatenon-acupointsproducinginconspicuousde-
FDR corrected with a minimal cluster size of 20 voxels) (table 3
qisensations[11,23],butthismightsignificantlycausebiasamong
and figure3).
Chinesesubjects.Toensurecomparabilitybetweenthetwogroups
Inthecontrolgroup,anincreaseinReHovalueswasobserved
during acupuncture manipulation, de-qi sensations were assessed
afterinactivetreatmentintheleftACC(BA32)andmedialfrontal
several times during the treatment session. Needling at inactive
gyrus (MeFG) (BA10). A signal decrease in ReHo values was
acupoints could effectively reduce the aforementioned bias, and
detectedintherightMFG(BA6)(P,0.05,FDRcorrectedwitha
evenly control non-specific factors such as expectancy effects
minimal cluster size of20voxels) (table 3 andfigure3).
during the period of study. In this experiment, the overall de-qi
Additionally, we have performed a direct comparison of the sensations in the active acupoint group and inactive acupoint
ReHochangesbetweentheactiveandinactivegroup.Theactive groupwere comparable andhadno statistical difference.
acupointgroupshowedhigherReHointhethalamus,ACC,STG,
SMAandlowerReHointhehippocampus,MFG,andMTGthan 4.1 Similarities and differences in clinical efficacy
theinactive group(P,0.001,uncorrected) (as shownin figure4).
between active acupuncture and inactive acupuncture
Based on the clinical outcomes of this RCT, both active and
3.3Clinicaloutcomesandcomparisonofde-qisensations
inactive acupuncture methods were helpful in treating migraine
Comparisonwithineachgroup,boththeactiveacupointgroup after 8 weeks of therapy (P,0.05). Both treatments remarkably
andinactivegroupshowedsignificantdecreasesintheVASscore, alleviated the clinical symptoms of migraine (intensity of pain,
frequency of migraine attack per 4 weeks, number of days with attack frequency, and days with migraine) and improved the
migraine per 4 weeks and HIT-6 score after 8 weeks’ treatment QOL. Furthermore, acupuncture at active acupoints was signif-
(P,0.05).Basedonthisstudy,asignificantdifferencewasfoundin icantlysuperiortoacupunctureatinactiveacupointsinalleviating
theVASscoresbetweenthetwogroupsbyanalysisofvariancefor pain intensity (P=0.015) in the current study. This result was
repeated measures (P=0.015) (table 4). The difference in VAS similar with our previous RCT report on the efficacy of
betweentheactivegroupandinactivegroupwasmorethan0.9in acupuncture at true acupoints compared with non-acupoints for
week 8 (P=0.006). However, no significant differences were migraine prevention [11].We inferred that similar clinical effects
observed between the two groups for the frequency of migraine of both treatments might partly result from placebo and
attackper4weeks,numberofdayswithmigraineper4weeks,and psychological effects. Theplacebo response isan essential part of
HIT-6 score at the end of treatment (P.0.05) (table 4). pain treatment, especially in the improvement of headache
Furthermore, analysis ofvariance of repeated measures indicated sufferers. A systematic review has shown that when clinicians
PLOSONE | www.plosone.org 5 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
Figure2.Theflowchartofstudy.TheflowchartofthisstudyaccordingtotheCONSORTStatement.
doi:10.1371/journal.pone.0099538.g002
Table1. Baselineand demographics formigrainepatients withoutaura (ITT).
Items ActiveacupointGroup(n=40) InactiveacupointGroup(n=40)
Meanage(SD),(years) 33.35(11.69) 33.23(9.73)
Female,n(%) 28(70.0) 29(72.5)
Meaneducation(SD),(years) 12.70(3.29) 13.68(3.74)
Meandurationofillness(SD),(years) 10.58(7.40) 9.93(5.73)
Familyhistory(Y(%)/N(%)) 8(20.0)/32(80.0) 9(22.5)/31(77.5)
Notes:ITT,intention-to-treat;SD,Standarddeviation;Y,yes;N,no.
doi:10.1371/journal.pone.0099538.t001
PLOSONE | www.plosone.org 6 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
Table2. Baselinecharacteristics of 40migraineurs whoparticipated inthe fMRIscan.
Items ActiveacupointGroup(n=20) InactiveacupointGroup(n=20)
Meanage(SD),(years) 32.90(10.99) 37.25(9.68)
Female,n(%) 14(70.0) 12(60.0)
Meaneducation(SD),(years) 12.95(3.52) 13.35(4.12)
Meandurationofillness(SD),(years) 8.55(6.49) 10.40(7.40)
Familyhistory(Y(%)/N(%)) 2(10.0)/18(90.0) 0(0)/20(100.0)
VASscore(SD) 5.28(2.03) 5.44(1.48)
Frequencyofmigraineattacksper4weeks* 7.90(4.88) 5.45(4.33)
Numberofdayswithmigraine(days)per4weeks 11.45(9.30) 8.75(9.21)
HIT-6score 60.45(8.13) 61.55(7.98)
Notes:SD,Standarddeviation;Y,yes;N,no.;
*Frequencyofmigraineattack,thenumberofepisodesofmigraineattacksseparatedbypain-freeintervalsofatleast48hours.
doi:10.1371/journal.pone.0099538.t002
stated positive outcome expectancies as opposed to uncertain 4.2 The similarities in resting-state brain activity evoked
expectancies, most studies found improvements in patient self- by active and inactive treatment
reports on pain, anxiety, and distress [33]. During the process of
Basedontheresting-statefMRIresults,commonbrainregions
the study, two acupuncturists were responsible for the treatments
responding to the acupuncture active treatment and inactive
alternately,andanotherexperienceddoctorwhodidnotknowthe
treatment included the ACC, MFG, and MTG. Among these
treatmentallocationtookchargeoftheefficacyevaluation.Aswe
areas, there was a significant negative correlation between the
know, acupuncture treatment could create enhanced placebo
increasedaverageReHovaluesoftheACCandadecreaseinthe
effects, such as patient expectations, longer patient-doctor VAS score in both groups (P,0.05, corrected). The results
appointments, and the power of touch and suggestion, so both
suggestedthattheincreaseinReHovaluesintheACCmightbe
theactivetreatmentandinactivetreatmentevenlyamelioratedthe
the common mechanism of acupuncture treatment for migraine
headache degree and frequency originating from patients’ self-
patients,despitetheneedledactiveacupointsorinactiveacupoints.
reports which may be explained by the aforementioned nonspe-
TheACCisakeyregioncomposedofthe‘‘painmatrix’’andis
cific effects.
involvedinthemedialpainsystem.Itisoneofthecommon‘‘brain
signature’’structuresinchronicpaindiseases,andisthoughttobe
Figure3.Brainactivityinmigraineurswithoutauraafterdifferentacupuncturetreatment.Long-termactiveacupointtherapyeliciteda
moreextensiveandremarkablecerebralresponsecomparedwithacupunctureatinactiveacupoints.
doi:10.1371/journal.pone.0099538.g003
PLOSONE | www.plosone.org 7 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
e
z
si
er
st
u
Cl 23 32 26
n
g
Si q Q q
2 0
A A3 A6 A1
B B B B
p(n=20) tvalue 3.47 24.43 3.69
u
o
Gr
nt z 25 44 25
oi
p
ent. acu h y 46 2 49
m e c
v a
treat Inacti Talair x 26 45 26
e
ur
unct size
p er
eacu Clust 33 52 39 61 21 21 23 26 21 22 27 40 79 26 47 67 87 30 53 51 85 16 71
v
cti
na gn
ori Si q q Q Q Q q q q q q q Q Q q q q q Q Q Q q q q q
e
v
cti
a
uraafter BA BA24/32 BA31 - BA13 VPM VPL BA10 BA6 BA22 BA39 BA20 BA17/18 BA18
a
ut e
o u 8 8 4 7 7 9
with 0) tval 3.52 3.86 23.7 23.7 23.8 3.26 4.04 3.95 2.89 3.05 3.34 23.3 24.4 2.9 3.78 4.76 5.38 3.56 4.2 23.2 3.71 3.94 2.84 3.55
s 2
patient oup(n= z 12 23 41 38 214 6 20 4 7 4 4 22 14 50 53 22 0 28 28 227 9 21 4 4
e Gr
n
ai nt
migr upoi y 30 33 245 248 218 211 214 220 220 220 217 62 59 29 29 218 29 263 263 210 278 286 273 273
c h
ngesin Activea Talairac x 23 6 23 3 230 242 36 215 15 218 15 221 42 227 33 245 50 250 50 256 218 21 215 12
a
h
c
o
H mi
e e
R H L R L R L R L R L R L R L R L R L R L R L R L R L R L R
al
br
e
er
ThecTable3. Region LimbicSystem ACC PCC Hippocampus Insula Thalamus FrontalCortex MFG MeFG SMA TemporalCortex STG MTG ITG OccipitalCortex Cuneus Lingualgyrus
PLOSONE | www.plosone.org 8 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
R,
ersize ely;L,left;
ust ctiv
Cl pe
es
r
e
Sign ecreas
d
or
e
as
e
cr
n
BA nali
g
si
oup(n=20) tvalue ureshoweda
Gr uct
nt z str
poi the
Inactiveacu Talairach xy cateswhether
di
n
i
ze )Q
si /q
er (
Clust 37 39 31 28 153 47 73 37 51 arrow
n
w
o
d
gn or
Si Q Q Q Q Q Q q q q p
U
a;
e
Ar
n
n
a
m
d
BA BA39 BA39 BA40 BA7 - - BA,Bro
e;
er
e h
0) tvalu 23.94 24.21 23.67 23.11 23.72 24.58 3.89 3.61 3.96 Hemisp
=2 mi,
oup(n z 42 31 31 49 46 36 212 213 231 xels;He
Gr vo
nt 20
oi 65 71 65 33 44 54 53 47 34 of
cup h y 2 2 2 2 2 2 2 2 2 size
Activea Talairac x 245 248 50 56 23 3 215 12 6 alcluster
m
ni
mi 3
0
Hemi eL R L R L R L R L R L R orrectedwitha one.0099538.t0
ul c p
Cont.Table3. Region ParietalLobe Inferiorparietallob Angulargyrus Postcentralgyrus Precuneus Cerebellum Brainstem ,Notes:P0.05,FDRright.doi:10.1371/journal.
PLOSONE | www.plosone.org 9 June2014 | Volume 9 | Issue 6 | e99538
AfMRIStudyonAcupunctureforMigraineurs
Figure4.DirectcomparisonoftheReHochangesbetweentheactiveandinactivegroup.TheactiveacupointgroupshowedhigherReHo
inthethalamus,ACC,superiortemporalgyrus,SMAandlowerReHointhehippocampus,middlefrontalgyrus,andmiddletemporalcortexthanthe
inactivegroup(P,0.001,uncorrected).
doi:10.1371/journal.pone.0099538.g004
engaged with both cognitive-attentional and affective dimensions induced reduction in pain intensity ratings was negatively
ofpain.TheACChasbeenrecognizedinplayingadeterministic associatedwithincreasedaverageReHovaluesintheACCwhich
roleinendogenouspaincontrol,whichismediatedbyendogenous illustrated that acupuncture treatment could promote pain
opioid systems [34]. In previous neuroimaging studies, the ACC reduction successfully by modulating the migraine-affected
was the most consistently deactivated region in PET and fMRI dysfunction region, theACC,tosome extent.
migraine studies [35,36], and also had a decrease in gray matter On the other hand, we inferred that the similarities in both
[37,38]. Our research group verified that compared with healthy clinical improvements and cerebral responses between active
controls, migraineurs showed a significant decrease in ReHo treatmentandinactivetreatmentwerepossiblyduetotheplacebo
values and amplitude of low-frequency fluctuation (ALFF) in the effect. During the process of treatment, migraineurs had positive
ACC [7,9], and showed aberrant functional connectivity which expectationstowardsacupuncturetherapyindependentofwheth-
hadtheACCinvolved[8,39].Inthepresentstudy,acupuncture- er or not the treatments were active or inactive, and moreover,
Table4. Clinicaloutcomemeasures ineach group (ITT).
ActiveacupointGroup(n=40) InactiveacupointGroup(n=40)
Outcomemeasure Mean(SD) 95%CI Mean(SD) 95%CI P" P{
VASscore
24–0weeks 5.11(1.75) (4.55–5.67) 5.23(1.78) (4.66–5.80) 0.7484 P =0.0000
T
1–4weeks 3.80(1.62) (3.28–4.32) 4.64(1.17) (4.27–5.01) 0.0094 P =0.0888
T*G
5–8weeks 3.07(1.57) (2.57–3.57) 4.07(1.54) (3.58–4.56) 0.0052 PG=0.0150
DifferencefrombaselineinVASI 2.096(0.25) (1.61–2.58) 1.110(0.25) (0.62–1.60) 0.006| -
Frequencyofmigraineattacksper4weeks*
24–0weeks 6.83(4.21) (5.48–8.17) 5.98(3.72) (4.79–7.16) 0.3412 PT=0.0000
1–4weeks 4.35(2.63) (3.51–5.19) 3.92(1.69) (3.38–4.46) 0.3802 P =0.3168
T*G
5–8weeks 2.85(2.19) (2.15–3.55) 3.1062.00 (2.46–3.74) 0.5983 P =0.4742
G
Numberofdayswithmigraine(days)per4weeks
24–0weeks 9.85(7.94) (7.31–12.39) 9.73(7.62) (7.29–12.16) 0.9429 P =0.0000
T
1–4weeks 5.56(4.25) (4.20–6.92) 4.91(2.36) (4.16–5.66) 0.4043 P =0.6459
T*G
5–8weeks 3.51(2.66) (2.66–4.36) 3.91(2.82) (3.01–4.82) 0.5122 PG=0.8835
HIT-6score
24–0weeks 58.10(6.81) (55.92–60.28) 58.13(7.12) (55.85–60.40) 0.4224 P =0.0000
T
1–4weeks 47.25(9.55) (44.20–50.30) 49.69(9.35) (46.70–52.68) 0.2515 PT*G=0.3834
5–8weeks 47.86(8.42) (45.17–50.55) 50.39(6.67) (48.26–52.52) 0.1395 P =0.2232
G
Notes:ITT,intention-to-treat;
CI,confidenceinterval;
*Frequencyofmigraineattack,thenumberofepisodesofmigraineattacksseparatedbypain-freeintervalsofatleast48hours;
"Pvaluesbasedont-testbetweenthetwogroups;
{Pvaluesbasedonrepeatedmeasures;
Ibasedonanalysisofcovarianceanalysis;
P,valuesforcomparisonbetweendifferenttimepoints;
T
P ,valuesforTime*Groupinteraction;
T*G
PG,valuesforcomparisonbetweendifferentgroups.
doi:10.1371/journal.pone.0099538.t004
PLOSONE | www.plosone.org 10 June2014 | Volume 9 | Issue 6 | e99538