166 Journal of Pain and Symptom Management Vol. 47 No. 1 January 2014

Brief Report

A Double-Blind, Placebo-Controlled,Crossover Pilot Trial With Extension Usingan Oral Mucosal Cannabinoid Extract forTreatment of Chemotherapy-InducedNeuropathic PainMary E. Lynch, MD, FRCPC, Paula Cesar-Rittenberg, MD, FRCPS, andAndrea G. Hohmann, PhDPain Management Unit (M.E.L.), Queen Elizabeth II Health Sciences Centre, Department of

Anesthesia, Psychiatry and Pharmacology (M.E.L.), and Division of Gynecologic Oncology (P.C.-R.),

Dalhousie University, Halifax, Nova Scotia, Canada; and Department of Psychological and Brain

Sciences (A.G.H.), Indiana University, Bloomington, Indiana, USA

Abstract

Context. Neuropathic pain caused by chemotherapy limits dosing and duration

of potentially life-saving anti-cancer treatment and impairs quality of life.Chemotherapeutic neuropathy responds poorly to conventional treatments, andthere is an urgent medical need for new treatments. Recent preclinical studiesdemonstrate that cannabinoid agonists suppress established chemotherapy-evoked neuropathy.

Objectives. This was a pilot trial to begin to investigate a currently availablecannabinoid agent, nabiximols (oral mucosal spray containing cannabinoids), inthe treatment of chemotherapy-induced neuropathic pain.

Methods. A randomized, placebo-controlled crossover pilot study was done in16 patients with established chemotherapy-induced neuropathic pain. A 0e10point numeric rating scale for pain intensity (NRS-PI) was used as the primaryoutcome measure.

Results. When examining the whole group, there was no statistically significantdifference between the treatment and the placebo groups on the NRS-PI. Aresponder analysis demonstrated that there were five participants who reporteda two-point or greater reduction in pain that trended toward statistical significanceand the number needed to treat was five.

Conclusion. Chemotherapy-induced neuropathic pain is particularly resistant tocurrently available treatments. This pilot trial found a number needed to treat offive and an average decrease of 2.6 on an 11-point NRS-PI in five ‘‘responders’’ (ascompared with a decrease of 0.6 with placebo) and supports that it is worthwhileto study nabiximols in a full randomized, placebo-controlled trial of

Address correspondence to:Mary E. Lynch, MD, FRCPC,Pain Management Unit, Queen Elizabeth II HealthSciences Centre, 4th Floor Dickson Centre, Room

4086, Dalhousie University, Halifax, Nova ScotiaB3H 1V7, Canada. E-mail: mary.lynch@dal.ca

Accepted for publication: February 22, 2013.

� 2014 U.S. Cancer Pain Relief Committee.Published by Elsevier Inc. All rights reserved.

0885-3924/$ - see front matterhttp://dx.doi.org/10.1016/j.jpainsymman.2013.02.018

Vol. 47 No. 1 January 2014 167Cannabinoids in Chemo-Induced Neuropathic Pain

chemotherapy-induced neuropathic pain. J Pain Symptom Manage2014;47:166e173. � 2014 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc.All rights reserved.

Key Words

Neuropathic pain, chemotherapy, cannabinoids, randomized controlled trial

IntroductionChemotherapeutic agents induce neuro-

toxic side effects that include severe neuro-pathic pain.1e3 This syndrome limits dosingand duration of potentially life-saving anti-cancer treatments and impairs quality of lifein affected patients. Chemotherapy-inducedneuropathy responds poorly to conventionaltreatments, making the identification of effec-tive alternative analgesics an urgent medicalneed.1,4 Neuropathic pain associated with che-motherapy (e.g., paclitaxel, vincristine, and cis-platin) is particularly resistant to treatmentand is a growing clinical problem as chemo-therapeutic regimens grow more successful inextending life.

Recent preclinical studies demonstrate thatcannabinoid agonists suppress establishedneuropathic pain induced by chemotherapytreatment. Agents such as vincristine and pacli-taxel sensitize neurons in the spinal cord thatcode information about pain,5,6 providing aneu-ral basis for chemotherapy-evoked neuropathicpain.Cannabinoids specifically suppress central7

as well as peripheral8 sensitization, effects postu-lated to contribute to pathological pain states(for review, see Rahn and Hohmann9). Thesestudies suggest a possiblemechanismunderlyingthe efficacy of cannabinoids in suppressingchemotherapy-evoked painful neuropathy.10,11

In preclinical studies, cannabinoids suppress es-tablished neuropathic pain evoked by all majorclasses of chemotherapeutic agents, includingvinca alkaloids (e.g., vincristine), taxol-derivedagents (e.g., paclitaxel),11,12 and platinum-derived agents (e.g., cisplatin).13e15 In fact, inanimal models, modulators of the endocannabi-noid system were more effective in suppressingcisplatin-inducedmechanical and cold allodyniacompared with conventional treatment withamitriptyline or gabapentin.14 Duration of anti-allodynic efficacyof the cannabinoidmodulatorsalso outlasted that of morphine.14 For this

reason, it was decided to examine the action ofa currently available cannabinoid in the treat-ment of chemotherapy-induced neuropathicpain.

MethodsDesign

A randomized, placebo-controlled, cross-over pilot study examining an oral mucosalspray containing cannabinoids (nabiximols[Sativex�, GW Pharmaceuticals, U.K.]) in 18patients with established chemotherapy-induced neuropathic pain was performed.

ParticipantsTo be included, participants had to have neu-

ropathic pain persisting for three months aftercompleting chemotherapy with paclitaxel, vin-cristine, or cisplatin. The average seven day in-tensity of pain had to be $4 on an 11-pointnumeric rating scale. Participants also exhibitedsensory abnormalities comprising allodynia, hy-peralgesia, or hypesthesia. Concurrent analge-sics had to be stable for 14 days before entryinto the trial. Participants were excluded ifthey suffered from ischemic heart disease, on-going epilepsy, a personal or family history ofschizophrenia, or psychotic disorder or sub-stance abuse or dependency within the previoustwo years. Exclusion criteria also included preg-nancy or other medical condition that mightcompromise safety in the trial.

Study MedicationThe study drug was nabiximols, an oromu-

cosal whole cannabis-based spray combiningD-9 tetrahydrocannabinol with cannabidiol,minor cannabinoids and terpenoids, plus eth-anol and propylene glycol excipients, and pep-permint flavoring. The placebo was packagedin exactly the same way, with a similar yellowishcolor and peppermint flavor.

168 Vol. 47 No. 1 January 2014Lynch et al.

Study ProcedureParticipants were recruited through adver-

tisements in the local paper and posters in on-cology clinics at the university teachinghospital (Capital District Health Authority,Halifax, Nova Scotia, Canada). Interested indi-viduals were interviewed by phone and bookedfor a screening appointment at the pain man-agement clinical trials unit.

The screening appointment included a dis-cussion with the potential participant aboutthe study to assure all questions were answered,at which point informed consent was obtained.A history and general physical examination,along with specific quantitative sensory testing(QST) of the painful area, was then completed,and a blood sample was taken for screening(complete blood count, liver function, and kid-ney function). An electrocardiogram also wasperformed.

Participants who met the study criteria werethen given a study number. The study numberswere assigned consecutively. A computer-generated randomization schedule determinedthe order of treatment (placebo-nabiximols ornabiximols-placebo) and was used at themanufacturing site where study numbers wereassigned to each participant’s supply of studymedication. Participants and study staff wereblinded to the randomization code, which wasnot broken until the completion of the study.

Participants completed a numeric ratingscale for pain intensity (NRS-PI) daily overthe next week and returned for a second visitto receive their first medication. They were in-structed to start with one spray of the studymedication under the tongue or inside of thecheek before bed the first night. Participantsthen increased the study medication by oneto two sprays per day until they reacheda dose that helped their pain; they were askednot to exceed 12 sprays per day and stop in-creases if limiting side effects, such as sedation,were encountered.

Once the dose was established, participantscontinued this stable dose for four weeks, re-turning after two and four weeks to completethe outcome measures. At the end of the firststudy period, they were instructed to decreasethe medication by one to two sprays per dayuntil discontinued and underwent a two weekwashout phase before returning for the second

study medication. The terminal eliminationhalf-lives of the cannabinoids in nabiximolsare 24e36 hours or longer. Thus, the twoweek washout was chosen to assure no carryover effect between study arms. Participantswere permitted to use concomitant medica-tions throughout the trial.If there was no effect of the medication (ei-

ther pain improvement or limiting side effectssuch as sedation) at the maximum of 12 spraysper day, an abbreviated one week stable dosephase was allowed subsequent to two weeks’ ti-tration (i.e., in contrast to the requirementthat it be taken for four weeks). This proce-dure allowed participants to proceed to thenext arm of the trial more quickly in the eventof no response at the maximum dose and pre-served blinding of the trial. Participants whocompleted the randomized, controlled, cross-over part of the study were eligible to enterthe extension trial where they were given thestudy drug (nabiximols) to use for up to sixmonths. During the extension phase, partici-pants returned to the study site at three andsix months to complete outcome measures.

Primary Outcome MeasureNeuropathic pain was measured using an 11-

point NRS-PI, with anchors ‘‘no pain’’ and‘‘pain as bad as you can imagine.’’ The primaryoutcome measure was change in NRS-PI frombaseline (average pain score over seven dayspretreatment) to the final week of stabledose treatment. Numerical rating scales havebeen widely used in pain research and havebeen demonstrated to be capable of identify-ing clinically meaningful changes.16

Secondary Outcome MeasuresThese included the Short Form-36 Health

Survey (SF-36�), QST, and collection of ad-verse events. Outcomes were collected beforetreatment and at midpoint and final week ofstable dose.

SF-36�. SF-36 is reliable and valid and themost commonly used generic measure ofhealth-related quality of life.17

QST. Sensory testing included assessment ofdynamic tactile allodynia and pinprick hyperal-gesia using the same protocol used in previous

Participants prescreened by telephone N = 54

36 excluded: travel issues (7);

decided not to take part (12); did not

have chemo NP diagnosis (7);

psychiatric history (2); lost contact

(2); wrong chemo (1); other (5)

N = 18 randomized to study

N = 9 Nabiximols then Placebo N = 9 Placebo then Nabiximols

N = 16 completed RCT

10 = entered extension

Withdrew

(N = 2)

Fig. 1. Screening, randomization, and dispositionof participants.

Vol. 47 No. 1 January 2014 169Cannabinoids in Chemo-Induced Neuropathic Pain

studies18 and was consistent with recommenda-tions presented at the annual meeting of theCanadian Pain Society.19

Other MeasuresBaseline demographics including name,

age, sex, diagnosis, medication, and previoustreatment also were recorded, as well as con-current medications and changes in medica-tion during the study.

Table 1Baseline Charac

Measure X (SD) S/P (N ¼ 9)

Age (years) 58 (11.34)Male:female 2:7NRS-PI baseline 6.56 (1.24)Cancer site

Ovary 2Cervix 1Lung 1Uterus 1Breast 2Testicle 0Blood 1Lymphoma 1

Chemotherapeutic agentCisplatin 1Oxaliplatin 1Paclitaxel 3Vincristine 1Combination 3

Number of cycles chemo 6.0Cycle number when pain started 2.33Chemo dose reduction because of pain (%) 28Duration of pain (months) 14Previous cannabinoid use 3Previous treatment with anticonvulsants 5Previous treatment antidepressants 1Previous treatment NSAIDS 0Previous treatment opioids 2

NRS-PI¼ numeric rating scale for pain intensity; NSAIDs¼ nonsteroidal anti

Statistical AnalysisA repeated measures analysis of variance

(ANOVA) for crossover data was performedwith time as the within-subjects factor andtreatment as the between-subjects factor.Where early release from a study phase waspermitted because of no effect, the last scorerecorded was carried forward and used as thefinal score, as it was assumed that if the partic-ipant had experienced no effect at the maxi-mum dose after two weeks’ titration anda week of stable maximum dose, they wouldbe unlikely to experience a response after this.

ResultsEighteen participants entered the trial and

16 completed the randomized, placebo- con-trolled, crossover protocol (Fig. 1). Therewere no significant differences between groupsbefore treatment (Table 1). The mean pre-treatment (seven day average) NRS-PI scorewas 6.75 (6.17e7.33). During active treatment,the mean NRS-PI score dropped to 5.5(4.43e6.57) at mid-treatment (during placebo

teristics

P/S (N ¼ 9) Total (N ¼ 18) P

55 (10.69) 56 (10.80) 0.591:8 3:15 1.0

7.00 (1.12) 6.78 (1.17) 0.44

3 5 1.01 2 1.00 1 1.02 3 1.02 3 1.01 2 0.470 1 1.00 1 1.0

2 3 1.00 1 1.04 7 1.00 1 1.03 6 1.05.44 5.72 0.373.22 2.78 0.27

33 1.019 17 0.562 5 1.05 10 1.00 1 12 2 0.470 2 0.47

-inflammatory drugs.

All Subjects

Baseline 1st Visit 2nd Visit Final Visit

0

2

4

6

8

10

Nabiximols

Placebo

Treatment Phase

NR

S-P

I S

co

re

s

Responders

Baseline 1st Visit 2nd Visit Final Visit

0

2

4

6

8

10

Nabiximols

Placebo

*

Treatment Phase

NR

S-P

I S

co

re

s

a

b

Fig. 2. Pain scores on numeric rating scale for painintensity. NRS-PI¼ numeric rating scale for painintensity.

170 Vol. 47 No. 1 January 2014Lynch et al.

treatment: 6.31 [5.58e7.04]) and to 6.00(6.98e5.02) at the end of four weeks ofactive treatment (placebo treatment: 6.38[5.67e7.09]). A repeated measures ANOVAdemonstrated that, for 16 participants whocompleted the study, there was a main effectfor time (P ¼ 0.007) but not for the interactionof time and treatment condition (P ¼ 0.29) orfor the between-subjects factor (P ¼ 0.52)(Fig. 2).

This was an unfunded pilot study intendedto collect data for a full randomized controlledtrial. We had enough funding to recruit ap-proximately 20 participants. We were able torecruit 18 within the time frame of approxi-mately one year. Using the comparison of thechange in NRS scores by condition, the power

Table 2Secondary Outcome Mea

Outcome Measure Pretreatment

SF-36 physical 32.68 (10.26)SF-36 mental 45.25 (10.21)Allodynia 25.94 (34.19)Hyperalgesia 46.67 (29.02)

SF-36¼ Short Form-36 Health Survey.

for this study was 13%; 183 participants wouldhave been required for statistical significance(mean difference placebo �0.37 [1.09] andfor treatment �0.78 [1.72]).

Responder AnalysisFive participants experienced a two-point or

greater drop in NRS-PI during treatment(Fig. 2). The mean baseline NRS-PI scores forthis group were 6.00 (4.92e7.08), with NRS-PIscores decreasing by 2.6 points to 3.40(2.40e4.40) during treatment with nabiximols,whereas dropping only 0.6 points to 5.40(4.07e6.73) with placebo. A repeated measuresANOVA demonstrated that the main effect fortime was significant (P < 0.001), and the inter-action of time and treatment condition was sig-nificant (P¼ 0.001); the between-subjects factorwas not (P ¼ 0.37).

Secondary MeasuresAnalysis of SF-36 and QST demonstrated no

statistically significant effect as compared withplacebo (Table 2).

Number Needed to TreatUsing the criterion of a drop of at least two

points on NRS-PI from baseline to the finalvisit in each phase as indicative of a clinicallysignificant change in pain, it was determinedthat there were five participants (31.5%) whodemonstrated pain reduction in the activetreatment phase and two participants who re-ported pain reduction in the placebo phase.The ‘‘risk reduction’’ is, therefore, 18.75%.The number needed to treat (NNT) ¼ 1/0.1875 ¼ 5.33, or five people.

Doses Used and Medication-Related AdverseEventsThe mean dose of medication used in the

treatment arm was eight sprays per day (range3e12) and 11 sprays in the placebo arm, withmost participants titrating to maximum dose

sure Means (SD)

Post-Treatment Post-Placebo

35.50 (9.19) 46.5 (8.5)44.86 (9.98) 33.9 (10.03)34.23 (22.71) 43.75 (26.01)45.0 (15.48) 50.62 (22.28)

Table 3Number of Participants Reporting Adverse

Events in Each Study Arm

Adverse EventsActive

Treatment Placebo

Fatigue 7 0Dry mouth 5 1Dizziness 6 0Nausea 6 1Increased appetite 2 0Diarrhea 2Decreased appetite 1 0Feeling ‘‘stoned’’ 1 0Anxiety 1 0Panic attack 1 0Headache 2 0Confusion 1 0‘‘Fuzzy thinking’’ or ‘‘foggy brain’’ 2 0

Vol. 47 No. 1 January 2014 171Cannabinoids in Chemo-Induced Neuropathic Pain

in the placebo arm (13/15). Medication-related side effects were reported by most par-ticipants. The majority of these were mild andtransient, did not lead to discontinuation ofthe medication or withdrawal from the trial,and improved with adjustment in dose. Themost common side effects reported were fa-tigue, dizziness, dry mouth, and nausea(Table 3). There were no serious medication-related adverse events.

Extension PhaseTen participants chose to continue into the

extension phase of the trial (Table 4). Painlevels dropped from a baseline of 6.9 to 5.0

Table 4Extension Phase Pain Scores, Doses,

Study PatientNumber

NRS-PI atBaseline(7 DayAverage)

NRS-PI at3 Months

NRS-PI at6 Months

DoP

101 8 7 d102 5 5 3105 6 5 d

106 7 6 6108 7 7 d

109 8 4 6112 7 3 d

113 5 1 0115 8 5 d

116 8 7 6Mean for the group 6.9 5.0 4.2

NRS-PI¼ numeric rating scale for pain intensity.

at three months and 4.2 at six months. The av-erage dose was 4.5 sprays per day with a rangefrom 2 to 10 sprays used per day. There werevery few medication-related adverse events re-ported by this group; one reported sleepinessand a second reported dizziness, nausea, andvisual scotoma on discontinuing the medica-tion, all of which were transient. Five withdrewafter three to four months. Reasons for with-drawing included return to work (1), recur-rent cancer (1), too busy (2), and wanted totry another treatment (1).

DiscussionWhen examining the whole group, there was

no statistically significant difference betweenthe treatment and the placebo groups. Re-sponder analysis nonetheless demonstratedthat five participants reported a two-point orgreater reduction in their pain according toNRS-PI, which trended toward statistical signif-icance. NNT was five.

Chemotherapy-induced neuropathic pain isa particularly resistant type of neuropathicpain to treat and responds poorly to currentlyavailable agents; thus, there is a critical needfor novel treatments.1,4 The fact that in this sta-tistically underpowered small pilot trial therewere five participants who experienced a clini-cally significant (two-point or greater)

and Length of Time in Study

se Sprayser Day

Length of Time in Extensionand Reason for Stopping

5e8 Returned to work and discontinued2e3 Completed 6 months3e4 Completed 3 months, withdrew because of

recurrence of cancer8e10 Completed 6 months6e8 Completed 3 months, withdrew because

too busy8e10 Completed 6 months3e4 Completed 4 months, withdrew because

too busy3e4 Completed 6 months3 Completed 3 months extension and then

decided to withdraw as was not gettingenough relief, wanted to be able to pursueother options

10 Completed 6 months extension4.5

172 Vol. 47 No. 1 January 2014Lynch et al.

reduction in their pain makes it worthwhile tocarry out a larger randomized controlled trialexamining nabiximols in this population. Tenparticipants who chose to continue into the ex-tension phase all noted modest improvements,with several noting further improvement be-yond the initial pain reduction. This trial alsodemonstrated that nabiximols is a safe medica-tion. There were no serious adverse eventsand the adverse events experienced weremild and transient and did not lead to with-drawal from the study or discontinuation ofthe medication.

This study is consistent with literature indi-cating that there is significant individual varia-tion in the response of pain to analgesics. Theconcept of NNT has been developed to assistwith comparing analgesic effectiveness acrossdifferent categories of drugs. A recent system-atic review identifies a wide variation in NNTfor the agents currently used to treat neuro-pathic pain.20 The best analgesics availablecurrently exhibit NNTs of 2.1 (tricyclic antide-pressants) and 2.6 (opioids). Other agents thatare commonly used include the gabapenti-noids, with NNTs of 6.4 (gabapentin) and 4.5(pregabalin), and the serotonin norepineph-rine reuptake inhibitor antidepressants, withan NNT of 5.0.13 With an NNT of 5.0 in thissmall pilot trial, nabiximols is in a similarrange to many other agents currently used rou-tinely for treatment of neuropathic pain.

In conclusion, this pilot trial supports that itwill be worthwhile to study nabiximols in a fullrandomized controlled trial of chemotherapy-induced neuropathic pain. Our studies alsoraise the possibility that nabiximols may beuseful as an adjunctive therapy for treatingchemotherapy-induced neuropathic pain.

Disclosures and AcknowledgmentsNo funding was received for this study and

the authors declare no conflicts of interest.The authors would like to acknowledge that

GW Pharmaceuticals kindly supplied the med-ication used in this study.

References1. Cata JP, Weng HR, Lee BN, Reuben JM,Dougherty PM. Clinical and experimental findings

in humans and animals with chemotherapy-induced peripheral neuropathy. Minerva Anestesiol2006;72:151e169.

2. Dougherty PM, Cata JP, Cordella JV, Burton A,Weng HR. Taxol-induced sensory disturbance ischaracterized by preferential impairment of myelin-ated fiber function in cancer patients. Pain 2004;109:132e142.

3. Polomano RC, Bennett GJ. Chemotherapy-evoked painful peripheral neuropathy. Pain Med2001;2:8e14.

4. Uhm JH, Yung WK. Neurologic complicationsof cancer therapy. Curr Treat Options Neurol1999;1:428e437.

5. Cata JP, Weng HR, Chen JH, Dougherty PM.Altered discharges of spinal wide dynamic rangeneurons and down-regulation of glutamatetransporter expression in rats with paclitaxel-induced hyperalgesia. Neuroscience 2006;138:329e338.

6. Weng HR, Cordella JV, Dougherty PM. Changesin sensory processing in the spinal dorsal horn ac-company vincristine-induced hyperalgesia and allo-dynia. Pain 2003;103:131e138.

7. Nackley AG, Zvonok AM, Makriyannis A,Hohmann AG. Activation of cannabinoid CB2 re-ceptors suppresses C-fiber responses and windup inspinal wide dynamic range neurons in the absenceand presence of inflammation. J Neurophysiol2004;92:3562e3574.

8. Khasabova IA, Khasabov S, Paz J, et al. Cannabi-noid type-1 receptor reduces pain and neurotoxicityproduced by chemotherapy. J Neurosci 2012;32:7091e7101.

9. Rahn EJ, Hohmann AG. Cannabinoids as phar-macotherapies for neuropathic pain: from thebench to the bedside. Neurotherapeutics 2009;6:713e737.

10. Rahn EJ, Makriyannis A, Hohmann AG. Activa-tion of cannabinoid CB1 and CB2 receptors sup-presses neuropathic nociception evoked by thechemotherapeutic agent vincristine in rats. Br JPharmacol 2007;152:765e777.

11. Rahn EJ, Zvonok AM, Thakur GA, et al. Selec-tive activation of cannabinoid CB2 receptors sup-presses neuropathic nociception induced bytreatment with the chemotherapeutic agent pacli-taxel in rats. J Pharmacol Exp Ther 2008;327:584e591.

12. Pascual D, Goicoechea C, Suardı́az M,Martı́n MI. A cannabinoid agonist, WIN 55,212-2,reduces neuropathic nociception induced by pacli-taxel in rats. Pain 2005;118:23e34.

13. Deng L, Guindon J, Vemuri VK, et al. The main-tenance of cisplatin- and paclitaxel-induced mechan-ical and cold allodynia is suppressed by cannabinoidCB(2) receptor activation and independent of

Vol. 47 No. 1 January 2014 173Cannabinoids in Chemo-Induced Neuropathic Pain

CXCR4 signaling in models of chemotherapy-induced peripheral neuropathy. Mol Pain 2012;8:71.

14. Guindon J, Lai Y, Takacs SM, Bradshaw HB,Hohmann AG. Alterations in endocannabinoidtone following chemotherapy-induced peripheralneuropathy: effects of endocannabinoid deactiva-tion inhibitors targeting fatty-acid amide hydrolaseand monoacylglycerol lipase in comparison to refer-ence analgesics following cisplatin treatment. Phar-macol Res 2013;67:94e109.

15. Vera G, Chiarlone A, Cabezos PA, et al. WIN55,212-2 prevents mechanical allodynia but notalterations in feeding behaviour induced bychronic cisplatin in the rat. Life Sci 2007;81:468e479.

16. Dworkin RH, Turk DC, Farrar JT, et al.IMMPACT. Core outcome measures for chronic

pain clinical trials: IMMPACT recommendations.Pain 2005;113:9e19.

17. Ware JE, Sherbourne CD. The MOS 36-itemshort form health survey (SF-36). Med Care 1992;30:473e483.

18. Lynch ME, Clark AJ, Sawynok J. Intravenousadenosine alleviates neuropathic pain, a doubleblind placebo controlled trial using an enriched en-rolment design. Pain 2003;103:109e115.

19. Squire P, Marchand S, Ware M. Using quantita-tive sensory testing (QST) in chronic pain evalua-tion: from research to clinical practice. Pain ResManag 2008;13:125.

20. Finnerup NB, Sindrup SH, Jensen TS. The evi-dence for pharmacological treatment of neuro-pathic pain. Pain 2010;150:573e581.