OriginalArticleMulticenter,Double-Blind,Randomized,Placebo-Controlled,Parallel-GroupStudyoftheEfcacy,Safety,andTolerabilityofTHC:CBDExtractandTHCExtractinPatients With Intractable Cancer-Related PainJeremyR.Johnson,MBChB,MaryBurnell-Nugent,MBBChir,DominiqueLossignol,MBChB,MRCG,DRCOG,ElenaDoinaGanae-Motan,MD,RichardPotts,BSc(Hons),MICR,andMarieT.Fallon,MBChB,MD,FRCP(E),FRCP(Glasg)SevernHospice(J.R.J.),Shrewsbury,Shropshire,andSt.LukesHospice(M.B.-N.),Turnchapel,Plymouth,UnitedKingdom;AssociationHospitaliereDeBrussels(D.L.),CentredesTumeursdelULB, Brussels, Belgium; Emergency Department (E.D.G.-M.), Hospital Sf. Ioan cel Nou, Suceava,Romania;GWPharmaLtd.(R.P.),Ely,Cambridgeshire;andEdinburghCancerResearchCentre(M.T.F.),UniversityofEdinburgh,Edinburgh,UnitedKingdomAbstractObjectives. This study compared the efcacy of a tetrahydrocannabinol:cannabidiol (THC:CBD) extract, a nonopioid analgesic endocannabinoid systemmodulator, and a THC extract, with placebo, in relieving pain in patients withadvanced cancer.Methods. In total, 177 patients with cancer pain, who experienced inadequateanalgesia despite chronic opioid dosing, entered a two-week, multicenter, double-blind,randomized,placebo-controlled,parallel-grouptrial.Patientswererandomized toTHC:CBDextract(n 60),THCextract(n 58),orplacebo(n 59).Results. The primary analysis of change from baseline in mean pain NumericalRatingScale(NRS)scorewasstatisticallysignicantlyinfavorofTHC:CBDcomparedwithplacebo(improvementof 1.37 vs. 0.69),whereastheTHCgroup showed a nonsignicant change (1.01 vs. 0.69). Twice as many patientstaking THC:CBD showed a reduction of more than 30% from baseline pain NRSscore when compared with placebo (23 [43%] vs. 12 [21%]). The associated oddsratiowasstatisticallysignicant,whereasthenumberofTHCgroupresponderswassimilartoplacebo(12[23%]vs.12[21%])anddidnotreachstatisticalThis studywas sponsoredbyGWPharmaLtd. Allstudy medication was supplied by GWPharmaLtd., and it also funded all sites involved in the studyby means of per-patient payments based on recruit-ment. GWPharmaLtd. has fundedJ. R. Johnson(primary author) to attend two conferences to pres-enttheresultsofthisstudy.Address correspondence to: Marie T. Fallon, MD, St. Co-lumbas Hospice Chair of Palliative Medicine, Edin-burgh Cancer Research Centre, Crewe Road,Edinburgh EH4 2XR, United Kingdom. E-mail:marie.fallon@ed.ac.ukAcceptedforpublication:June22,2009.2009U.S. CancerPainReliefCommitteePublished by Elsevier Inc. All rights reserved.0885-3924/09/$eseefront matterdoi:10.1016/j.jpainsymman.2009.06.008Vol. -No. --2009 JournalofPainandSymptomManagement 1ARTICLEINPRESSsignicance.Therewasnochangefrombaselineinmediandoseofopioidbackgroundmedicationormeannumberofdosesofbreakthroughmedicationacross treatment groups. No signicant group differences were found in the NRSsleep quality or nausea scores or the pain control assessment. However, the resultsfromtheEuropeanOrganisationforResearchandTreatment ofCancerQualityofLifeCancerQuestionnaireshowedaworseninginnauseaandvomiting withTHC:CBDcomparedwithplacebo(P 0.02), whereasTHChadnodifference(P 1.0). Mostdrug-relatedadverseeventsweremild/moderateinseverity.Conclusion. ThisstudyshowsthatTHC:CBDextractis efcaciousforreliefofpain in patients with advanced cancer pain not fully relieved by strong opioids.JPainSymptomManage2009;-:-e-. 2009U.S.CancerPainReliefCommittee.PublishedbyElsevierInc.Allrightsreserved.KeyWordsCancer,pain,tetrahydrocannabinol,cannabidiol,SativexIntroductionCancer pain is a common problem, and70%e90%of patients withadvancedcancerexperiencesignicant pain.1Opioids remainthe keystone for the treatment of moderatetoseverecancerpain; however,somepatientsexperience inadequate pain relief with opioidsandstandardadjuvantanalgesicsdespitedoseadjustments, and unacceptable side effectsarecommon.2,3Cannabis contains 60 or more cannabinoids(CBs).Themainonesincludedelta-9-tetrahy-drocannabinol (THC) and cannabidiol(CBD).4,5There is evidence that bothTHCandCBDshowpromise inrelieving cancer-related pain.6,7Sativex(THC:CBD), an endo-cannabinoid system modulator, is produced byGWPharma Ltd, UnitedKingdom. It is de-rivedfromstrainsofCannabissativaL. plantsdevelopedtoproducehighandreproducibleyieldsofprincipal CBs(THCandCBD), withminoramountsofotherCBsandterpenesinasolutioncontainingethanol, propylenegly-col, and peppermint oil avoring.5The namedCBs constitute at least 90% of the total CB con-tentoftheextracts.CBs act primarily through specic CB recep-tors: CB1receptorsarepredominantlydistrib-utedinthecentral nervous system, andCB2receptorsarelocatedprimarilyintheperiph-ery (including the immune system). The prin-cipalpharmacologicaleffectsofTHCincludeanalgesia, muscle relaxation, antiemesis, appe-titestimulation, andpsychoactivity.8CBDhasshown anticonvulsant, muscle relaxant,anxiolytic, neuroprotective, antioxidant, andantipsychotic activity, and has also been showntoreducetheanxietyandpsychoactiveeffectsof THC.9,10Preliminarytestsofpharmacologyand behavioral activity support the similarity ofthe endogenous CBanandamide to THC.11Botharepartial agonistsattheCB1receptor.CBD, incontrast, binds weakly to CB1andCB2but does showpharmacological potencyas a neutral antagonist at each receptor,12that is, issilent at suchreceptorsbut canre-verse both agonist and inverse agonist re-sponses. CBDalsohas shownpowerful anti-inammatory, immunomodulatory,13and anti-oxidant propertiesinvitro.14It isatransientreceptor potential vanilloid receptor 1(TRPV1)vanilloidreceptoragonistinitsownright, while modulating anandamide by inhib-itingbothitsreuptakeandhydrolysis.15Addi-tionally, CBD increases adenosine A2Areceptorsignalingbyinhibitionoftheadeno-sinetransporter.16BothTHCandCBDhaveshown analgesic efcacy in animalmodels.10,17,18Inthisstudy, bothaTHC:CBDextract and a THC-only extract were comparedagainst placebo to ascertain if the inclusion ofCBDprovided a different efcacy or safetyprole.Campbell et al.19publishedaliteraturere-view of nine randomized controlled trialsperformed using CBs (any route of administra-tion) in patients with acute, chronic nonmalig-nant, or cancer pain. Five studies that weredescribed in four reports comprised 128patients with cancer pain.6,7,20,21All of the trials2 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESSconductedonpatientswithcancerpainwereplacebo-controlled trials. Four of the trialsfound CB as effective as codeine but with dose-limitingsideeffects. Thus, CBs havedemon-strated efcacy comparable to selected opioids.THC:CBD is the rst endocannabinoidsystem modulator to undergo clinical develop-ment for pain. It has been approved in Canadaforthereliefofneuropathicpaininmultiplesclerosisandpersistentbackgroundcancer-re-latedpain.Theformulationisanoromucosalspray that allows exible, individualizeddos-ing. Patientsself-titratetheiroverall doseandpatternofdosingaccordingtotheirresponseto, and tolerance of, the medicine, with admin-istration of approximately 8e12 sprays/day,that is, 22e32 mg/day THCand20e30 mg/day CBD. This study assessed the analgesic ef-cacy of THC:CBD and THC extracts comparedwiththat of placebointhe management ofpatients withat least moderately severecan-cer-relatedpaindespiteappropriatepharma-cologicalmanagement.MethodsThis two-week(twodays baselineandtwoweeks treatment), multicenter, double-blind,randomized, placebo-controlled, parallel-group study evaluated the efcacy ofTHC:CBD extract and THC extract in the anal-gesicmanagement of patients withmoderateto severe cancer-related pain. There wasa two-day baseline period. Adult male orfemale patients whohadbeenusing strongopioids for at least one week to relieve pain as-sociatedwithincurablemalignancy andwhogavewritteninformedconsentwerescreenedfor study entry. Eligible patients recordedapainseverityscoreof4oraboveona0e10Numerical RatingScale(NRS)onbothdaysof thetwo-daybaselineperiod. Patients wereexcluded if they had cancers affecting theoral cavity; radiotherapy tothe oor of themouth; major psychiatric or cardiovascular dis-orders;epilepsy;renalor hepaticimpairment;or if they were pregnant, lactating, or not usingadequatecontraception. Patientswhohadre-ceived therapies expected to confound thestudy outcome(epidural analgesiawithin48hours of screening; palliativeradio-, chemo-,or hormonal therapy within two weeks ofscreening; or CBs within seven days of random-ization) were also excluded. Patients takinglevodopa, sildenal, or fentanyl or patientswithahypersensitivity toCBs wereexcludedonsafetygrounds.Patientscompletedastudydiary, recordingpainscorethreetimes dailyand background medication and all additionalbreakthroughanalgesia oneachday duringthe baseline period. Patients thenreturnedfor assessment, randomization, and dose intro-ductiontooneof thethreetreatment arms:THC:CBDextract, THCextract, or placebo(Fig. 1)ina1:1:1treatment allocationratio.Patients werereviewedafter7e10days (Visit2)andat theendof study (14e20days) orwithdrawal (Visit 3). During the medicationdosing period, the patients continued to com-plete the daily study diaries with the aforemen-tionedinformationandthenumberofdosesofstudymedicationtaken.Therelevantregu-latoryauthoritiesandresearchethicscommit-teesapprovedthestudy.Thestudy medicationwas deliveredusinga pumpactionoromucosal spray. Each100-mL actuation of the pump containing theTHC:CBDextractdeliveredadosecontaining2.7 mg THC and 2.5 mg CBD. Each 100-mL ac-tuationofthepumpcontainingtheTHCex-tract delivered a dose containing 2.7 mgTHC, and each actuation of placebo deliveredonlyexcipientspluscolorants.Themaximumpermitted dose of all study medication waseightactuationsinanythree-hourperiodand48actuationsinany24-hourperiod.Patients self-titratedtotheir optimal doseoverthesevendaysofWeek1, basedonef-cacy, tolerability, and the maximum permitteddose. Patients could increase the total numberof sprays each day by a maximum of 50% untilthey either had satisfactory relief of their symp-toms ordevelopedunwantedeffects, suchasintoxication (high). The total number ofsprays was spread over the day with a minimumof 15minutesbetweenanytwosprays. If un-wantedeffects developedona newnumberofsprays, thepatientwouldnottakeanyfur-ther sprays for three to four hours. The patientwouldthengoback totaking their furthersprays at asimilar level totheprevious day.Once the patient hadfoundthe maximumnumberof spraysperdaythat theytoleratedwellorthenumberthatprovidedgoodsymp-tomrelief,they continued withapproximatelyVol. -No. --2009 3 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESSthe same number of sprays per day for theremainderofthestudy.Thecoprimaryendpoints werethechangefrombaselineinNRSpainscoreanduseofbreakthrough analgesia. The NRS, a widelyusedandvalidatedmeasureof painseverity,is capable of showing clinically and statisticallysignicantchangesinpaindisorders.22,23TheNRS question indicate your level of painwasansweredbypatientsthreetimesdaily(inthemorningonwaking,atlunchtime, andinthe evening before retiring), using the anchors0 nopainand10 verybadpain. Patientswere allowed to use their breakthrough analge-siaasrequired,andthiswasrecordeddailyinthe diary. Patients maintained backgroundmedicationforthedurationofthestudy.Thesecondaryendpointsincludedtheuseofopi-oid background medication, patient assess-ments of sleep quality, nausea, memory,concentration,andappetiteovertheprevious24 hours using diary NRSs. The Brief PainInventory-Short Form (BPI-SF) and The Euro-pean Organisation for Research andTreatment of Cancer (EORTC) Quality ofLife Questionnaire (QLQ-C30) Version 3werecompletedbypatients at Visit 1andattheendofthestudy.24,25TheBPI-SFconsistsof nine questions; eight have a single response,and Question 9 is subdivided into seven parts.The Total BPI (Questions 3e6) is the un-weighted sum of the four pain scores and rep-resents thepainintensity. TheTotal BPI forQuestion 9 is the unweighted sum of the sevenassessmentsandrepresentstheeffectofpain.The EORTCQLQ-C30 cancer questionnaireconsists of 30 questions that cover globalhealthstatus, functional scales (e.g., physicalfunctioning),andsymptoms(e.g.,fatigue).Adverseevents (AEs)anduseof concomi-tantmedicationswerereportedbypatientsatstudy visits throughout the trial. Predenedcategories for determiningtheintensity andthe relationship to study medication wereused. Theexpert clinical judgment fromtheinvestigatingstudyphysicianswasusedinde-terminingintensityandcausalrelationshipofAEsandseriousAEs.Fig. 1.Studydesign(ConsolidatedStandardsofReportingTrials[CONSORT]diagram).ITT intenttotreat.4 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESSThestudywaspoweredassuminganunder-lyingtreatment differenceof 1point onanNRS and a standard deviation(SD) of 1.6 (es-timated from previous studies), with 80%power and two-sided 5% signicance.26,27Afterallowing for 15%dropouts, 58 subjects pergroup were required. For the two coprimary ef-cacy variables (NRS painscore anduse ofbreakthrough medication), the Hochberg28methodwasusedtotesttheglobalhypothesisfor atreatment effect onpain. Thenull hy-pothesis was to be rejected if either coprimaryvariable produced two-sided P #0.025 or bothproducedP #0.05. ThedailypainNRSscorewas themeanof thethreedailyassessments.The change inmeanNRS painscore frombaseline (all days in run-in period) to theend of treatment (last three days on treat-ment) was analyzedusinganalysis of covari-ance (ANCOVA), with baseline pain asa covariate and grouped study center and treat-ment as factors.29,30The proportions of re-sponders (patients with$30%improvementfrom baseline to end of study NRS pain score)were compared betweentreatments. Use ofbreakthroughmedication(numberofdaysofuseduringlast threedaysontreatment)wasanalyzed using logistic regression with a cumu-lative logit model. In addition, the changefrom baseline in mean number of doses of es-cape medication was analyzed using ANCOVA.ResultsA total of 192 patients were screened over 25months, leading to 177 patients randomized totreatment (Fig. 1) at 28 European centers. Themean(SD) durationof cancer inthese pa-tients was 3.5 years (2.8 [3.27], 3.2 [4.27],and 4.5 [5.25] years, respectively, in theTHC:CBD,THC,andplacebogroups, respec-tively). The mean age, gender distribution,previous cannabis use, primary disease sites,and pain classication were similar amongthe three treatment groups (Table 1). Themost common type of cancer pain was ofmixedpathophysiology,followedbyboneandneuropathicpain(Table1). At baseline, themean daily dose of opioid background medica-tion in the whole study population was 271 mgof oral morphine equivalents. The median oralTable1PatientDemographicsDemographics THC:CBD THC Placebo TotalGender,n (%)Male 33(55) 30(52) 32(54) 95 (54)Female 27(45) 28(48) 27(46) 82 (46)Ethnic origin,n(%)Caucasian 59(98) 57(98) 58(98) 174 (98)Other 1(2) 1(2) 1(2) 3 (2)Previous cannabis use,n (%) 6(10) 6(10) 7(12) 19 (11)Age(years), mean (SD) 59.4 (12.1) 61.3(12.5) 60.1 (12.3) 60.2 (12.3)Duration ofcancer (years), mean (SD) 2.8(3.3) 3.2(4.3) 4.5(5.3) 3.5(4.4)BMI, mean (SD) 23.1 (4.2) 23.5(5.2) 24.1 (4.3) 23.6 (4.6)Primarycancer sites,n (%)Breast 12(20) 8(14) 9(15) 29 (16)Prostate 6(10) 8(14) 10(17) 24 (14)Lung 7(12) 9(16) 4(7) 20 (11)Painclassication,n(%)Mixed 31(52) 28(48) 30(51) 89 (50)Bone 16(27) 24(41) 25(42) 65 (37)Neuropathic 11(18) 11(19) 17(29) 39 (22)Visceral 14(23) 12(21) 11(19) 37 (21)Somatic/incident 7(11) 5(9) 6(10) 18 (10)Baseline morphine equivalentsaMedian (mg) 80.0 120.0 120.0 120.0Range 0e6,000 0e1,280 0e6,000 0e6,000Mean (SD) 258.4 (789.47) 188.2(234.49) 367.0 (886.38) 271.2 (698.98)1ste3rdquartile 30e180 50e213 40e240 40e240BMI bodymassindex.aOralmorphineequivalencedataaresourcedfromRefs.43e45.Vol. -No. --2009 5 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESSmorphine equivalent dose was slightly lower inthe THC:CBDgroup at baseline comparedwiththeTHCandplacebogroups(Table2).Forall threetreatment groups, thepredomi-nant primary reason for discontinuing thestudywasAEs(Fig.1).The mean (SD) number of sprays taken perday, which had stabilized by the end of the rstweek(Days 1e7)endingthetitrationphase,wereTHC:CBDextract, 8.75(5.14); THCex-tract, 8.34 (5.17); and placebo, 9.61 (4.67)(Fig. 2). Overall, fortheentiretreatmentpe-riod, themean(SD) number of sprays useddailyintheplacebogroup(10.88[5.81])washigher than those in the THC:CBD (9.26[5.53])andTHCgroups(8.47 [5.46]).EfcacyThemean(SD) baselineNRSpainscoreswere similar among treatment groups andwithin grouped centers (THC:CBD extrac-t 5.68 [1.24], range 2.33e8.25; THCex-tract 5.77 [1.33], range 2.87e9.33;placebo 6.05 [1.32], range 3.5e9.56).The adjusted mean reduction in NRS (ANCO-VA)forTHC:CBD, THC,andplacebogroupsat the end of the treatment were 1.37,1.01, and 0.69points, respectively.31Theadjustedmeantreatmentdifferencefrompla-cebo was statistically signicant for a reductionin pain with the THC:CBD extract (0.67points, P 0.014) but not the THCextract(0.32 points, P 0.245). The ANCOVA didnot have normally distributedresiduals, butthenonparametricanalysis gaveasimilar re-sult. Themedianchanges frombaselineforTHC:CBD, THC, and placebo groups were1.36, 1.00, and 0.60, respectively. The me-diandifferencefromplacebowas statisticallysignicantforareductioninpain, whichwasin favor of THC:CBDextract (0.55 points,P 0.024)but not fortheTHCextract(0.24points, P 0.204). Sensitivity analyses of thechange from baseline in the mean NRS scoresconcurred withtheprimaryanalysis.Inchronic paintrials, it is recommendedthatthepercentagesofpatientsobtainingre-ductionsinpainintensityof at least 30%ona pain NRS (responders) should be docu-mented.32A reduction in pain NRS of approx-imately 30% is considered to representaclinicallyimportant difference.23Inthein-tent-to-treat responder analysis, approximatelytwice as many patients in the THC:CBD grouphadareductionfrombaselineNRSofatleast30%compared with the placebo and THCgroups (THC:CBD23[43%] vs. THC12[23%], placebo 12[21%]). Theodds ratiofor the comparison of responders betweenTHC:CBDandplacebowas 2.81(95%con-dence interval [CI] 1.22, 6.50; P 0.006),and between THC and placebo was 1.10(95%CI 0.44,2.73;P 0.28)(Fig.3).Thenumber of days onwhichany break-through medication was used was similaramongall treatment groups, withnosigni-cant differencesobservedinthisclinical trialof brief duration (THC:CBD vs. placebo:Table2ChangeinDoseofOpioidBackgroundMedication(OralMorphineEquivalents)andStrongOpioidBreakthroughMedicationOpioidcharacteristics THC:CBD THC Placebo AllOpioid background medication: change from baseline to last 3 days on study medication (patients with data available, excluding3patientsreceiving intrathecal opioids)n (% ITT population) 60 (100) 58(100) 58(98) 176(99)Median 0.0 0.0 0.0 0.0Range 627to300 27to1,088 1,200 to400 1,200 to1,088Mean (SD) 3.5(108.44) 26.9 (152.00) 41.4 (201.27) 6.4(160.60)Q1,Q3 0,0 0,0 0,0 0,0Opioidbackground medication: categorized change frombaseline inoral morphine equivalentsperdayIncrease,n(%) 6 (12) 6(12) 4(7) 16(10)No change,n(%) 41 (79) 40(77) 43(80) 124(78)Decrease,n (%) 5 (10) 6(12) 7(13) 18(11)Strong opioid breakthrough medication:categorized change frombaseline innumber ofdosestakenN (% ITT population) 22 (37) 18(31) 19(32) 59(33)Increase,n(%) 2 (9) 4(22) 7(37) 13(22)No change,n(%) 12 (56) 10(56) 12(63) 34(58)Decrease,n(%) 8 (36) 4(22) 0 12(20)ITT intenttotreat;SDstandarddeviation.6 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESSP 0.70). Therewasareductionobservedinthemeannumberofdailydosesofall break-through medication (THC:CBD extrac-t 0.19; THC extract 0.14; placebo 0.15) by the endof thestudy period, butthe difference in change frombaseline be-tween treatment groups was not statistically sig-nicant. More specically, there was no changefrombaselinetothelast threedays of treat-ment in the median oral morphine equivalentdose of opioid background medications in 124(78%)patientsforwhomthedatawereavail-able. Doses were increased for 16 patients(10%) and reduced for 18 (11%); thesechanges were evenly distributed across thethreetreatmentgroups(Table2).Duringthebaseline period or last three days ontreat-ment, strong opioid breakthrough medicationwas recorded by 59 patients (33%); of these, 34(58%)showednochangeinthenumber ofdoses taken when comparing baseline withlast three days of treatment, 13 (22%) in-creased the number of doses, and 12 (20%) re-ducedthenumberof doses taken. Agreaterproportion of patients in the THC:CBD group(eightpatients)reducedbreakthroughdoses;conversely, the highest proportion of increasesindosewas intheplacebogroup(sevenpa-tients), which was statistically signicantlygreater than those in the THC:CBDgroup(P 0.004).MostoftheNRSdiarysymptomscoresandinvestigator-assessedpaincontrol showednosignicant treatment differencesbetweenthethreegroups (Table3). Astatistically signi-cant differenceinimprovementwithplacebowasobservedinthediaryNRSconcentrationand memory scores, whereas the placebogroup showed a mean improvement frombaselineinconcentrationscore(0.35) andtheTHC:CBDgroupshowedadeterioration(0.33, P 0.02), asdidtheTHCgroup(0.29,P 0.03). The memory score showed nochange in the placebo group (0.01), but a dete-rioration in the THC:CBD group (0.63,P 0.045) and in the THC group (0.66,P 0.053). Similarly, theappetitediary NRSscore showed a mean improvement from base-line in the placebo group, and there wasa slight reductioninappetite score inbothTHC:CBDandTHCgroups (0.59vs. 0.24,P 0.016; and 0.59 vs. 0.06, P 0.056, re-spectively)(Table3).The QLQ-C30showed,asexpected,few dif-ferencesamongtreatmentgroupsinthetwo-weekfollow-up. Ofthe16itemsassessed, theonly statistically signicant observations were024680 12 14 16 14 1 3 1 2 1 1 1 0 1 9 8 7 6 5 4 3 2 1 L By a D t n e m t a e r T y d u t SMean Number of Sprays per Day (+ / - SE)D B C : C H To b e c a l PC H TFig. 2.Exposuretostudymedicationdmeannumberofspraysperday.SE standarderror.010203040506070>= 10% >= 30% >= 50%% Response Level% of Patients Achieving ResponseLevelTHC:CBDTHCPlacebo1.77 (0.83,3.80)a2.81 (1.22,6.50)a,P = 0.006b1.66 (0.44,6.25)aaOdds Ratio (95% CI) THC:CBD vs. PlacebobFishers Exact TestFig. 3. Pain 0e10 Numerical Rating Scale scores: re-sponder analysis (ITTanalysis).aOdds ratio(95%CI)THC:CBDvs.placebo;bFishersexacttest.Vol. -No. --2009 7 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESSTable3Primary and Secondary Endpoints Showing Baseline Score, Change from Baseline, Treatment Difference, andStatisticalSignicanceoftheDifferenceinChangeFromBaselineforCBD:THC,THC,andPlaceboEndpointTreatmentGroup BaselineChangeFromBaselineComparison WithPlaceboTreatmentDifferenceStatisticalSignicance,P ValueMeanpainseverityNRS score(coprimary)THC:CBD 5.68 1.37 0.67a0.014THC 5.77 1.01 0.32a0.245Placebo 6.05 0.67 d dBreakthroughmedication:no.ofdaysused (coprimary)THC:CBD d d OR0.96a0.697THC d d OR1.20b0.555Placebo d d d dBreakthroughmedication:meandailydoseTHC:CBD 0.91 0.19 0.04a0.688THC 1.10 0.14 0.01b0.899Placebo 0.80 0.15 d dMeansleep quality NRSscore THC:CBD 4.33 0.57 0.31a0.346THC 4.46 0.24 0.02b0.95Placebo 4.17 0.26 d dMeannausea NRS score THC:CBD 2.44 0.26 0.49b0.110THC 2.04 0.24 0.46b0.126Placebo 1.98 0.22 d dMeanmemoryNRSscore THC:CBD 3.02 0.63 0.65b0.045THC 2.98 0.66 0.62b0.053Placebo 2.90 0.01 d dMeanconcentrationNRS score THC:CBD 3.59 0.33 0.68b0.021THC 3.53 0.29 0.64b0.028Placebo 3.37 0.35 d dMeanappetiteNRSscore THC:CBD 4.83 0.24 0.83b0.016THC 4.58 0.06 0.66b0.056Placebo 4.98 0.59 d dPaincontrolassessmentproportionwith paincontrolledTHC:CBD 50% 1% OR1.70 0.488THC 54% 2% OR1.76 0.400Placebo 36% 4% d dMeanBPI-SF totalpaininlast24hoursTHC:CBD 20.88 0.17 1.04a0.619THC 21.29 3.20 4.07a0.048Placebo 23.48 0.87 d dMeanBPI-SF totalinterference bypaininlast24hoursTHC:CBD 46.63 3.53 4.84a0.325THC 39.39 4.50 5.81a0.275Placebo 51.05 1.31 d dMeanQLQ-C30 global healthstatusTHC:CBD 29.74 7.23 2.47a0.443THC 27.05 5.60 0.84a0.793Placebo 25.29 4.77 d dMeanQLQ-C30 physicalfunctioningTHC:CBD 40.34 6.92 4.23b0.108THC 35.56 3.94 1.25b0.631Placebo 34.14 2.69 d dMeanQLQ-C30 role functioning THC:CBD 29.02 0.02 3.31a0.415THC 28.65 0.12 3.21a0.434Placebo 25.00 3.33 d dMeanQLQ-C30 emotionalfunctioningTHC:CBD 24.44 7.70 6.73a0.084THC 22.41 6.19 5.22a0.174Placebo 25.37 0.98 d dMeanQLQ-C30 cognitivefunctioningTHC:CBD 50.57 5.33 9.01b0.022THC 56.32 6.77 10.46b0.008Placebo 50.85 3.68 d dMean QLQ-C30 social functioning THC:CBD 29.02 3.19 1.61a0.679THC 29.89 9.66 8.08a0.038Placebo 25.71 1.58 d dMeanQLQ-C30 fatigue THC:CBD 71.55 3.92 2.71a0.422THC 70.69 1.36 0.15a0.965Placebo 64.56 1.21 d dMeanQLQ-C30 nausea andvomitingTHC:CBD 25.57 5.13 8.56b0.020THC 22.13 3.41 0.02b0.997Placebo 21.75 3.43 d dMeanQLQ-C30 pain THC:CBD 83.62 15.64 6.34a0.107THC 79.60 15.71 6.41a0.103Placebo 81.64 9.30 d d(Continued)8 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESSreductions incognitivefunctionscore whencompared with placebo (THC:CBD extrac-t 5.33 vs. 3.68, P 0.02; THC extrac-t 6.77vs. 3.68, P 0.01)andaworseningof nausea and vomiting score in theTHC:CBD, although not in the THC onlygroup, when compared with placebo(THC:CBD5.13vs. 3.43, P 0.02; THC3.41 vs. 3.43; P 1.0). A trend towardimprovement was seeninboth active treat-ment groups in the QLQ-C30 pain assessmentscore (THC:CBDextract 15.64vs.9.30,P 0.11; THC extract 15.71 vs. 9.30,P 0.10) and in the constipation score(THC:CBD5.74vs. 2.23, P 0.08; THC3.11vs.2.23,P 0.23).SafetyandTolerabilityTheactivecompounds weregenerally welltolerated,andnosafetyconcernswereidenti-edduringthis study. Treatment-relatedAEswere reportedby 106(60%) patients. Com-mon treatment-related AEs (three or more pa-tients) were similar to those seen in otherTHC:CBD clinical trials: somnolence, dizzi-ness,andnausea,mostlyofmildormoderateseverity(Table4).26,27,33e38Theincidenceofdeathinthisadvancedcancerpopulationwassimilar across treatment groups (eightTHC:CBD, eight THC, sevenplacebo), andallwereconsideredbecauseofprogressionofunderlying disease. None of the cases fromthe10patientswhoreportednonfatalseriousAEs (SAEs) raised any concerns regardingthe safety of CBs. The nonfatal SAEs of urinaryretention, tumor-relatedpain, worsenednau-sea, weakness, tumor hemorrhage, and somno-lence were experienced by ve patients inTHC:CBDgroup, allofwhichwereunrelatedto study medication. Three events were moder-ateinseverity, andfoureventsweresevereinseverity. Five subjects who received THC expe-rienced the nonfatal SAEs of metastases tobrain, gastric ulcer hemorrhage, syncope,bronchopneumonia, hyperglycemia, confu-sion, oral candidiasis, somnolence, tremor,anddisorientation. All eventswereunrelatedto study medication, with the exception ofasingleepisodeofsyncope,whichwasproba-bly related to THC. Two events were moderateinseverity,andeighteventsweresevereinse-verity. No patients fromthe placebogroupre-portedanonfatalSAE.DiscussionUnrelieved cancer pain can result in signi-cant distress anddisability.1,39Theresults ofthis studyshowthat theTHC:CBDextract isan efcacious adjunctive treatment forTable3ContinuedEndpointTreatmentGroup BaselineChangeFromBaselineComparison WithPlaceboTreatmentDifferenceStatisticalSignicance,P ValueMeanQLQ-C30 dyspnea THC:CBD 40.23 1.09 0.80a0.846THC 43.27 4.21 4.49b0.282Placebo 34.46 0.28 d dMeanQLQ-C30 insomnia THC:CBD 52.30 6.15 1.05a0.833THC 51.15 0.41 4.69b0.347Placebo 51.41 5.10 d dMeanQLQ-C30 appetiteloss THC:CBD 60.34 3.69 0.88a0.857THC 54.60 1.19 1.62b0.743Placebo 59.32 2.81 d dMeanQLQ-C30 constipation THC:CBD 50.00 5.74 7.97a0.077THC 33.33 3.11 5.35a0.233Placebo 40.68 2.23 d dMeanQLQ-C30 diarrhea THC:CBD 13.22 2.15 1.57a0.615THC 8.62 0.56 1.15b0.713Placebo 12.99 0.58 d dMeanQLQ-C30 nancialdifcultiesTHC:CBD 58.05 5.58 1.70a0.714THC 59.20 8.93 5.05a0.276Placebo 57.06 3.88 d daInfavorofactivetreatment.bInfavorofplacebo.Vol. -No. --2009 9 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESScancer-relatedpaininpatients who are notachievinganadequateanalgesic responsetoopioids.This studyinvolvedpatients withadvancedcancer, whohadameandiseasedurationofmore than three years and moderate to severelevels of painat entry (>4onanNRSpainscale), despite ongoing opioid treatment. Aftertwoweeks of receivingstudy medicationad-junctive to all other treatments, the THC:CBDextractgroupshowedastatisticallysignicantreduction in pain severity when comparedwithplacebo, withareductioninmeanpainNRS scores from baseline of 1.37 points(22.6%). The painNRS data were not nor-mallydistributed;hence,parametric andnon-parametric analyses were conducted. This hadno inuence on the signicance of the results.Theheterogeneity inthedistributionof thepainscores (many large negative and largepositive results), combined with consensus-basedrecommendations,32highlight the im-portance of the responder analysis. These rec-ommendations are primarily based on theresultsofananalysisofrelationshipsbetweenchangesinpainintensityandpatient reportsofoverall improvementin10clinical trialsofchronicpain, withpatientsofdiversediagno-ses, in which a clinically relevant responsewasdenedasareductionofpainofatleast30%frombaselinetoendofstudy.In this current study population, 43% of pa-tients takingthe THC:CBDextract achieveda30%orgreaterimprovement intheirpainscore (equated to a mean improvement of2.71boxes), approximatelytwicethenumberofpatientswhoachievedthisresponseintheTHCandplacebogroups. Theresults of theresponder analysis and the mean changefrombaselinemustbeinterpretedremember-ing that the study medications were adjunctivetoexistingtreatments, includingstrongopi-oids, for the duration of the trial. Larger treat-mentdifferencesfromplacebomaybenotedina study of longer duration, as evident inotherconditions.23,24At baseline, themeandaily useof opioidbackground medication was relatively high(271 mgof oral morphineequivalents). Thechange innumber of daily doses of break-through medication between baseline andendofstudyshowedaslighttrendtowardre-ductionandnorelevant differences betweentreatment groups. Only a small number of pa-tients recorded taking strong opioid break-through medication in their daily diariesduringthebaselineperiodorlast threedaysontreatment. Of these, most showedconsis-tentdosingpatterns;thechangesthatdidoc-curshowedatrendtowardadecreaseinthenumber of doses taken in the THC:CBD groupandanincreaseintheplacebogroup. Therewasalargerangeinthedoseof backgroundoral morphine equivalent treatment. Thesendings maybe a reectionof different treat-ment models usedintheparticipatingcoun-tries and illustrates the need to includea morespeciceligibilitycriteriaof minimumopioidtreatmentinfuturestudies. Lessvaria-tion in the existing treatment regimens wouldenhancetheinterpretationoftheefcacyre-sultbutwouldmakerecruitmenttothestudymorechallenging.Nostatisticallysignicantdifferencesinpa-tient-assessed sleep quality or nausea NRSscores or investigator-assessed pain controlassessment were noted between the studymedicationsandplacebo. Therewasasigni-cant improvement inthe BPI-SFtotal scorefor THCbut not for THC:CBD. Studies oflonger duration in other indications haveregularly shownthat the quality of sleepinthe THC:CBD group needs to be im-proved.26,27,33,40The differences betweentreatment groups inthememory, concentra-tion,andappetiteNRSdiaryscoresarepartlyattributable toanapparent improvement intheplacebogroup.The QLQ-C30 showedfewdifferences be-tween study medications and placebo. Consid-eringthefollow-updurationandthepatientpopulation, this is unsurprising. There wereTable4MostCommonTreatment-RelatedAdverseEvents(ReportedbyThreeorMorePatients)Description ofEventTHC:CBD,n(%)THCextract,n(%)Placebo,n(%)Somnolence 8(13) 8(14) 6(10)Dizziness 7(12) 7(12) 3(5)Confusion 4(7) 1(2) 1(2)Nausea 6(10) 4(7) 4(7)Vomiting 3(5) 4(7) 2(3)Raisedgamma GT 2(3) 5(9) 1(2)Hypercalcemia 0 0 3(5)Hypotension 3(5) 0 0GT gammaglutamyltransferase.10 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESSmarginal improvements in QLQ-C30 painscores but signicantly reduced cognitive func-tion scores with the THC:CBD and THCgroups comparedwiththoseof theplacebogroup. The statistically signicant worseninginQLQ-C30nauseaandvomitingscoreseenwith the THC:CBD extract compared with pla-cebowasnotseeninthediaryscoresfornau-seaandis confoundedby amedianchangeof 0betweenthegroups, makinginterpreta-tion of this result difcult. Similarly, thechanges inappetite reportedinthe patientdiarieswerenotseenintheQLQ-C30.The few statistically signicant results of thesecondary endpoints should be interpretedwithcautionbecauseofthemultipleanalysesperformed on the questionnaire and the over-lap in content between some of the NRS scalesand questionnaire items. However, there isaconsistentimpairmentofcognitivefunctionreportedby patients inthis study. Althoughthe clinical signicance of this nding isunclear,itwarrantsfurthercarefulassessmentinlong-termstudies. Itisacceptedthattherewill belimitations andpotential inaccuraciesof patient-completed diary data, and futurestudieswilllooktorenethismethodofdatacompletion.TheAEs seeninthis studyweresimilartothoseseeninotherclinical trials26,27,33e38Ofthe AEs leading to permanent cessation ofstudymedication(17%, 12%and3%, respec-tively, for THC:CBD extract, THC extract,and placebo), approximately half were consid-ered to be related to study treatment. None ofthe33reportedSAEsraisedconcernsregard-ingtreatment safety. Theincidenceof deathwas comparable among treatment groups,andtherewerenotreatment-relateddeaths.Despite some uncertainty in the total mor-phine equivalent dose received by patients,thesafetyproleadds evidencethat this wasa population with advanced disease; 13% of pa-tientsdiedduringthestudybecauseof theirunderlyingdisease.The THC:CBDandthe THCmedicationswere well tolerated. Patients were fully titratedat one week and maintained stable dosingthroughout the treatment period, that is, therewas no observed tendency to increase dosewithtime. This correspondedtoareductioninpainNRSscoreoverthesameperiod. Theclinical response to pain with THC:CBDextract oromucosal spray has not demon-strated tolerance in several clinical trials of lon-gerduration.26,27,33e38ThereisevidenceofsynergybetweenTHCandmorphineinpain, andTHCmaymodu-late endogenous opioidtone.41However, inthis study, the THC:CBD combination showeda more promising efcacy prole than theTHCextractalone. Thisndingissupportedby evidence of additional synergy betweenTHC and CBD. CBD may enhance the analge-sicpotential of THCbymeans of potent in-verseagonismat CB2receptors,14whichmayproduceanti-inammatoryeffects, alongwithitsabilitytoinhibit immunecell migration.42Additionally, CBD may modulate the potentialunwantedeffectsofTHCbymeansofantago-nismat CB1receptors,39which potentiallywouldprovideabetter safety prolefor theTHC:CBDmedicationinchronic use.Inconclusion, THC:CBDextract, anonop-ioid analgesic, endocannabinoid system modu-lator, has been shown to be a useful adjunctivetreatment for relief of paininpatients withadvancedcancer whoexperienceinadequateanalgesiadespitechronicopioidtherapy.Thereductions in pain scores were neither becauseof a change in opioid background medicationsnor becauseof anincreaseinuseof break-throughmedication. Therefore, wecancon-clude that the observed reduction in painscoresisattributabletothepositiveanalgesiceffectsofTHC:CBDextract. Theseresultsareveryencouraging andmeritfurther study.AcknowledgmentsTheauthorsthankallthepatients andtheirfamilies who participated in this trial. Addition-ally, the authors acknowledge GW Pharma Ltd.for sponsoring the trial and the followinginvestigators andtheir researchstaff for their par-ticipation: Dr. D. Brooks, Royal ChestereldHospital, Chestereld, UK; Dr. D. Feuer, St. Bar-tholomews Hospital, London, UK; Dr. R. Gaunt,Rowden Surgery, Chippenham,UK;Dr. A. Gri-gorescu, InstitutulOncologicProf. AlexTres-tioreanu Sos., Bucuresti, Romania; Dr. K.Gruffydd-Jones, The Box Surgery, Corsham,UK; Dr. E. Hall, St. Helena Hospice, Colchester,UK; Dr. J. Hardy, The Royal Marsden NHS Trust,Sutton, UK; Dr. C. Higgs, Dorothy HouseVol. -No. --2009 11 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESSFoundation, Bradford-Upon-Avon, UK; Dr. P.Hoskins, MountVernonHospital, Northwood,UK; Dr. S. Kelly, MarieCurieCentre, Belfast,UK;Dr.M.Leng,RoxburghHouse,GrampianUniversity Hospitals NHS Trust, Aberdeen, UK;Dr. D. McKeith, Townhead Surgery, Irvine, UK;Dr. D. Mosoiu, HospiceCasa Sperantei, Brasov,Romania; Dr. C. Nitipir, DiagnosticandTreat-ment Centre FoundationDr. Victor Babes281, Bucharest, Romania; Dr. W. Notcutt, JamesPaget Hospital, Norfolk, UK; Dr. I. Parker,BeehiveSurgery, Bath, UK; Dr. J. Riley, RoyalMarsden Hospital, Palliative Care Unit, London,UK; Dr. A. S. Thomson, Atherstone Surgery,Atherstone, UK; Dr. I. Trotman, Michael SobellHouse, Mount VernonHospital, Northwood,UK; Dr. L. Vata, District Hospital Dr. AlexandruSimionescu, Oncology Department, Hune-doara, Romania; Dr. C. Volovat, Center of Medi-cal Oncology, Lasi, Romania; Dr. A. Wijnberg,CornhillSurgery,Birmingham,UK;andDr.A.Wilcock, Hayward House Hospice, NottinghamCity Hospital, Nottingham, UK. Finally, theauthors acknowledge the assistance of V. Lane,Maclagan Healthcare Ltd., for her assistance inproducing the rst draft of this manuscript.The authors are grateful to Peter Clark, statisti-cian at GW Pharma Ltd., for acting as statisticaleditor for this article.References1. Glare P. Choice of opioids and the WHO ladder.In: Davis M, Glare P, Hardy J, eds. Opioids in cancerpain. Oxford, UK: OxfordUniversity Press, 2005:221e234.2. Scottish Intercollegiate Guidelines Network.Controlofpaininpatientswithcancer:ANationalClinical Guideline. Edinburgh, Scotland: SIGN44,2000.1e79.3. CleelandCS, GoninR, HateldAK, etal. Painand its treatment in outpatients with metastatic can-cer.NEnglJMed1994;330:592e596.4. HouseofLordsSelectCommitteeonScienceand Technology. Cannabis: The scientic andmedical evidence, 9th Report. Session 1997e98.London, UK: House of Lords, 1998. Availablefrom http://www.parliament.the-stationery-ofce.co.uk/pa/ld199798/ldselect/ldsctech/151/15101.htm. AccessedJanuary4,2008.5. GW, Pharma Ltd. Product monograph: Sativex,Ottawa, Canada. Health Canada, 2007. Available fromhttp://cpe0013211b4c6d-cm0014e88ee7a4.cpe.net.cable.rogers.com/dpdonline/searchRequest.do.Accessed March 4, 2008.6. NoyesR, BrunkSF, BaramDA, etal. Analgesiceffectofdelta9-tetrahydrocannabinol. JClinPhar-macol1975;15:139e143.7. Noyes R, Brunk SF, Avery DA, et al. The analge-sicpropertiesofdelta9-tetrahydro-cannabinol andcodeine.ClinPharmacolTher1975;18:84e89.8. Pertwee RG. Pharmacology of cannabinoid CB1and CB2receptors. Pharmacol Ther 1997;74:129e180.9. Pertwee RG. Neuropharmacology and therapeu-tic potential of cannabinoids. Addict Biol 2000;5:37e46.10. Russo EB, Guy GW. A tale of two cannabinoids:the therapeutic rationale for combining tetrahydro-cannabinol and cannabinol. Med Hypotheses 2006;66:234e246.11. Fride E, Mechoulam R. Pharmacological activityof thecannabinoidreceptoragonist, anandamide,a brain constituent. Eur J Pharmacol 1993;231:313e314.12. Thomas A, Baillie GL, Phillips AM, et al. Canna-bidiol displays unexpectedly high potency as an an-tagonist of CB(1)andCB(2)receptor agonists invitro.BrJPharmacol2007;150:613e623.13. Malfait AM, Gallily R, Sumariwalla PF, et al. Thenonpsychoactive cannabis constituent cannabidiolisan oralanti-arthritictherapeuticinmurinecolla-gen-induced arthritis. Proc Natl Acad Sci USA 2000;97:9561e9566.14. HampsonAJ, Grimaldi M, AxelrodJ, WinkD.Cannabidiol and (-)delta 9-tetrahydrocannabinolare neuroprotective antioxidants. Proc Natl AcadSciUSA1998;95:8268e8273.15. Bisogno T, Hanus L, De Petrocellis L, et al. Mo-lecular targets for cannabidiol and its synthetic ana-logues: effect on vanilloid VR1 receptors and on thecellular uptake and enzymatic hydrolysis of ananda-mide.BrJPharmacol2001;134:845e852.16. CarrierEJ,AuchampachJA,HillardCJ.Inhibi-tionof anequilibrativenucleosidetransporter bycannabidiol: a mechanism of cannabinoid immuno-suppression. Proc Natl Acad Sci USA 2006;103:7895e7900.17. Pertwee RG. Cannabinoidpharmacology: therst 66years. Br J Pharmacol 2006;147(Suppl 1):163e171.18. PertweeRG. Cannabinoidreceptors andpain.ProgNeurobiol2001;63:569e611.19. Campbell FA, TamerMR, Carroll D, etal. Arecannabinoids an effective and safe treatment optioninthemanagementofpain?Aqualitativesystemicreview.BMJ2001;323:1e6.20. Jochimsen PR, Lawton RL, VerSteeg K, Noyes R.Effectofbenzopyranoperidine,adelta-9THCcon-gener, onpain. ClinPharmacol Ther 1978;24(2):223e227.12 Vol. -No. --2009 Johnsonetal.ARTICLEINPRESS21. Staquet M, Gantt C, Machin D. Effect of a nitro-gen analog of tetrahydrocannabinol on cancer pain.ClinPharmacolTher1978;23(4):297e401.22. McCormackHM, HorneDJ, SheatherS. Clini-cal applicationsofvisual analoguescales: acriticalreview.PsycholMed1988;18:1007e1019.23. Farrar JT, Young JP Jr, LaMoreaux L, et al. Clin-ical importance of changes in chronic pain intensitymeasured on an 11-point numerical pain ratingscale.Pain2001;94:149e158.24. Cleeland CS, Ladinsky JL, Serlin RC, et al. Mul-tidimensional measurement of cancer pain: com-parisons of US and Vietnamese patients. J PainSymptomManage1988;3:23e27.25. AaronsonNK, Ahmedzai S, BergmanB, et al.The European Organisation for Research and Treat-ment of Cancer QLQ-C30: aquality if lifeinstru-ment for use in international clinical trials inoncology.JNatlCancerInst1993;85:365e376.26. Rog DJ, Nurmikko TJ, Friede T, et al. Random-ized, controlled trial of cannabis-based medicine incentral paininmultiplesclerosis. Neurology2005;65:812e819.27. Berman JS, Symonds C, Birch R. Efcacy of twocannabisbasedmedicinalextractsforreliefofcen-tral neuropathic pain from brachial plexus avulsion:resultsofarandomisedcontrolledtrial.Pain2004;112:299e306.28. HochbergY. Asharper Bonferroni procedureformultipletestsofsignicance. Biometrika1988;75:800e802.29. Conover WJ, InamRL. Analysis of covarianceusingtheranktransformation.Biometrics1982;38:715e724.30. HodgesJLJr, LehmannEL. Estimatesof loca-tionbasedonranktests.AnnMathStatist1963;34:598e611.31. JohnsonJR, Wright ST. Cannabis-basedmedi-cines onthe treatment of cancer pain: a rando-mised, double blind, parallel group, placebocontrolled, comparative study of the efciency,safety, andtolerabilityofSativexandTetranabinexin patients with cancer-related pain. J SupportOncol2005;3(Suppl3):21.32. DworkinRH, TurkDC, Farrar JT, et al. Coreoutcomemeasures forchronicpainclinical trials:IMMPACT recommendations.Pain2005;113:9e19.33. CollinC, Davies P, MutibokoI, et al. Rando-misedcontrolledtrial of cannabisbasedmedicinein spasticity caused by multiple sclerosis. Eur J Neu-rol2007;14:290e296.34. WadeDT,MakelaP, RobsonP, etal.Docanna-bis-based medicinal extracts have general or speciceffects onsymptoms inmultiplesclerosis? Adou-ble-blind, randomized, placebo-controlled studyon160patients.MultScler2004;10:1e8.35. WadeDT, MakelaPM, HouseH, et al. Long--term use of a cannabis-based medicine in the treat-mentofspasticityandothersymptomsinmultiplesclerosis.MultScler2006;12:639e645.36. Barnes MP. Sativexclinical efcacy and tolera-bility in the treatment of symptoms of multiple scle-rosis and neuropathic pain. Expert OpinPharmacother2006;7:607e615.37. Nurmikko TJ, Serpell MG, Hoggart B, et al. Sat-ivexsuccessfullytreatsneuropathicpaincharacter-ised by allodynia: a randomised, double-blind,placebo-controlled clinical trial. Pain 2007;133:210e220.38. RogD, NurmikkoTJ, YoungCA. Oromucosaldelta-9-tetrahrdrocannabinol/cannabidiol for neu-ropathicpainassociatedwithmultiplesclerosis: anuncontrolled, open-label, 2-year extension trial.ClinTher2007;20:2068e2079.39. Ripamonti C, DickersonED. Strategiesforthetreatment of cancer paininthenewmillennium.Drugs2001;61:955e977.40. RussoE,GuyG,RobsonP.Cannabis,painandsleep: lessons from therapeutic clinical trials ofSat-ivex cannabis based medicine. Chem Biodivers2007;4:1729e1743.41. Cox ML, Haller VL, Welch SP. Synergy betweendelta9-tetrahydrocannabinol andmorphineinthearthriticrat.EurJPharmacol2007;12:125e130.42. Pertwee RG. The diverse CB1 and CB2 receptorpharmacology of three plant cannabinoids:D9-tetrahydrocannabinol, cannabidiol and D9-tetra-hydrocannabivarin. Br J Pharmacol 2008;153:199e215.43. Twycross R, WilcockA, CharlesworthS, et al,eds. Palliativecareformulary, 2ndedOxford, UK:RadcliffePublishing,2002.44. FallonM. Opioidswitching androtation. In:Sykes N, Fallon MT, Patt RB, eds. Clinical pain man-agement: Cancer pain. London, UK: Hodder Ar-nold,2003:157e165.45. Hanks G, CherneyN, FallonM. Themanage-ment of pain. In: DoyleD, Hanks G, Cherny N,Calman K,eds.Oxfordtextbookofpalliativemedi-cine, 3rdedOxford, UK: OxfordUniversityPress,2005.Vol. -No. --2009 13 THC:CBDforTreatmentofCancer-RelatedPainARTICLEINPRESS