w w w . e l s e v i e r . c o m / l o c a t e / p a i n

PAIN�

152 (2011) 1018–1023

Safety and efficacy of pregabalin in patients with central post-stroke pain

Jong S. Kim a,⇑, Guy Bashford b, T. Kevin Murphy c, Andrew Martin c, Vardit Dror d, Raymond Cheung c

a Department of Neurology, University of Ulsan, Asan Medical Center, Seoul, South Koreab Graduate School of Medicine, University of Wollongong, NSW, Australiac Pfizer Inc, New York, NY, USAd UBC Scientific Solutions, Southport, CT, USA

Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article.

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

Article history:Received 17 March 2010Received in revised form 8 November 2010Accepted 14 December 2010

Keywords:Central post-stroke painNeuropathic painPregabalinStroke

0304-3959/$36.00 � 2011 International Associationdoi:10.1016/j.pain.2010.12.023

⇑ Corresponding author. Tel.: +82 2 3010 3442; faxE-mail address: jongskim@amc.seoul.kr (J.S. Kim).

Pregabalin has demonstrated efficacy in several forms of neuropathic pain, but its long-term efficacy incentral post-stroke pain (CPSP) is unproven. We evaluated the efficacy and safety of pregabalin versusplacebo in patients with CPSP. A 13-week, randomized, double-blind, multicenter, placebo-controlled,parallel group study of 150 to 600 mg/day pregabalin was conducted in patients aged P18 years withCPSP. The primary efficacy endpoint was the mean pain score on the Daily Pain Rating Scale over the last7 days on study drug up to week 12 or early termination visit. Secondary endpoints included other painparameters and patient-reported sleep and health-related quality-of-life measures. A total of 219patients were treated (pregabalin n = 110; placebo n = 109). A mean pain score at baseline of 6.5 in thepregabalin group and 6.3 in the placebo group reduced at endpoint to 4.9 in the pregabalin group and5.0 in the placebo group (LS mean difference = –0.2; 95% CI = –0.7, 0.4; P = 0.578). Treatment with pregab-alin resulted in significant improvements, compared with placebo, on secondary endpoints includingMOS-sleep, HADS-A anxiety, and clinician global impression of change (CGIC) P < 0.05. Adverse eventswere more frequent with pregabalin than with placebo and caused discontinuation in 9 (8.2%) of pregab-alin patients versus 4 (3.7%) of placebo patients. Although pain reductions at endpoint did not differ sig-nificantly between pregabalin and placebo, improvements in sleep, anxiety, and CGIC suggest someutility of pregabalin in the management of CPSP.

� 2011 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved.

1. Introduction

Central post-stroke pain (CPSP) is a neuropathic pain syndromecharacterized by constant or intermittent pain in a body part occur-ring after stroke associated with sensory abnormalities [1]. Althoughpreviously known as thalamic pain syndrome, lesions at any level ofthe spinothalamocortical pathways may produce CPSP [5]. The vastmajority of patients develop CPSP within 6 months after a stroke;however, there are reports of pain onset as late as 2 to 3 years afterstroke [4,10]. Prevalence of CPSP after stroke has ranged between1% and 8% [14]. Secondary comorbidities such as depression, anxiety,and sleep disturbances contribute to the low quality of life of CPSPpatients [6].

Treatment options for CPSP are currently limited. Drugs includ-ing antidepressants, anticonvulsants, antiarrhythmics, opioids, andN-methyl-D-aspartate antagonists are used, but are only partlyeffective [14,15]. Amitriptyline was proved to be effective for CPSPand is currently considered the first-line treatment [16]. However,

for the Study of Pain. Published by

: +82 2 474 4691.

not all patients respond to amitriptyline and the treatment sideeffects are frequently intolerable.

Pregabalin binds the a2-d subunit of voltage-gated calcium chan-nels to modulate the release of several neurotransmitters that mayaffect pain pathways [26]. Pregabalin has demonstrated efficacy inseveral forms of peripheral neuropathic pain and centrally mediatedpain [8,23,30]. The objective of the current study was to investigatethe efficacy, safety, and tolerability of pregabalin (150–600 mg/day)for pain relief and other quality-of-life measures in patients withCPSP.

2. Methods

2.1. Study population

Enrolled patients were men or women P18 years of age withCPSP and otherwise in generally good health. The diagnosis of CPSPwas based on medical history, physical examination, imagingfindings, and conformance to criteria suggested by Treede et al.[27]: (1) painful symptoms with neuroanatomically plausible dis-tribution; (2) medical history consistent with relevant strokeaffecting the somatosensory system; (3) neurologic examination

Elsevier B.V. All rights reserved.

J.S. Kim et al. / PAIN�

152 (2011) 1018–1023 1019

and quantitative assessment of thermal and mechanical sensationfindings showing distribution of sensory symptoms neuroanatom-ically consistent with the stroke; (4) CT or MRI findings showingthe relevant stroke disrupting the somatosensory pathway.Patients could have had P1 stroke provided they met all otherselection criteria. Patients were excluded if there were otherpotential causes of pain that could not be readily discriminatedfrom the CPSP.

To be eligible, the relevant stroke had to occur P4 months beforerandomization, and CPSP should have been present P3 months. Inaddition, patients must have scored P40 mm on the Short-FormMcGill Pain Questionnaire (SF-MPQ) Visual Analog Scale (VAS) tobe eligible. Patients were required to keep a daily pain diary, in whichthey recorded their daily pain score on an 11-point (0 = no pain to10 = worst possible pain) numeric rating scale (NRS). Patients musthave completed at least 4 daily pain entries and had an average dailypain score P4 over the 7 days before randomization.

Female participants could not be pregnant or lactating, and if ofchildbearing age, had to have a confirmed negative serum pregnancytest at baseline and practice appropriate methods of contraceptionthroughout the study period. Stable cardiovascular disease, diabe-tes, hypertension, and atherosclerosis were permitted provided theywere controlled. Patients were permitted to continue with stablecertain pharmacological therapies for pain or insomnia if used innormal routine >30 days before screening. However, patients hadto be willing not to change or initiate any of these therapies duringthe study. Other key exclusion criteria included skin conditions inthe affected dermatome that could have altered sensation, any se-vere cognitive impairment, and unstable psychological, medical, orpsychiatric conditions.

2.2. Study design

This 13-week, randomized, double-blind, placebo-controlled,parallel-group study was conducted at 32 centers in the Asia Pacificregion to evaluate the efficacy, safety, and tolerability of pregabalinfor the treatment patients with CPSP (evidence level, Class I). Enroll-ment began on August 23, 2006 and the last patient last visit wasSeptember 12, 2008. The study is compliant with the Declarationof Helsinki and the International Conference on HarmonizationGood Clinical Practice Guidelines. The final study protocol, anyamendments, and informed consent documentation were approvedby the Institutional Review Board(s) and/or Independent EthicsCommittee(s) at each investigational center. In addition, all localregulatory requirements were followed, particularly those provid-ing greater protection for the safety of the patients. The clinicalprotocol was conducted in accordance with Food and Drug Adminis-tration Regulations. The trial was registered at ClinicalTrials.gov:NCT00313820. The study consisted of 4 phases: 2-week screeningand washout, 4-week dose adjustment phase, 8-week maintenancephase, and 1-week taper phase with either pregabalin 150 mg/day ormatching placebo. There were 10 scheduled visits: 2 for the screen-ing and washout phase, 4 for the dose adjustment phase, 2 for thedose maintenance, 1 for the tapering phase, and 1 at the end of treat-ment. At screening, patients provided signed informed consent,were evaluated for inclusion/exclusion criteria, and were stabilizedon current pain or analgesic medication as defined in the protocol.Randomization to either pregabalin or placebo in a blinded 1:1 ratiowas done via a centralized telerandomization system (IMPALA)according to a computer-generated schedule with a randomizedpermuted block design. Access to the randomization code was re-stricted to authorized personnel. At the beginning of the dose adjust-ment phase, patients were randomized to treatment with pregabalin75 mg b.i.d. (150 mg/day) or matching placebo for 7 days. After1 week, patients were instructed to take pregabalin 150 mg b.i.d.,(total dose, 300 mg/day) or matching placebo for 7 days. Over the

next 2 weeks, patients were assessed weekly for dose adjustmentto either remain on the dose of 150 mg b.i.d. or to increase to themaximum allowed dose of 300 mg b.i.d. (total dose, 600 mg/day),based on the patient’s clinical response and tolerance. Patients expe-riencing intolerable adverse events (AEs) could have their dose re-duced by 1 level. During the maintenance phase, patientscontinued their final pregabalin dose from the end of the doseadjustment phase. At the tapering phase, patients were instructedto take pregabalin 75 mg b.i.d., (total dose of 150 mg/day) or match-ing placebo for 7 days.

2.3. Efficacy and safety assessments

The primary endpoint was the mean pain score obtained fromthe last 7 available scores of the Daily Pain Rating Scale while onstudy medication, up to and including the day of the beginningof tapering phase or its early termination equivalent. The DailySleep Interference Scale (DSIS), a secondary endpoint, consists ofan 11-point NRS ranging from 0 (pain does not interfere withsleep) to 10 (pain completely interferes with sleep). Self-assess-ment of pain and DSIS was performed as part of the daily subjectdiary on awakening before taking study medication. Other second-ary endpoints included: weekly mean pain scores, proportion ofsubjects with at least a 30% and 50% reduction in mean pain score,Quantitative Assessment of Neuropathic Pain (QANeP; 11-pointNRS), Neuropathic Pain Symptom Inventory (NPSI, a self-adminis-tered questionnaire designed to evaluate the different symptomsof neuropathic pain), weekly mean sleep interference scores, Med-ical Outcome Study Sleep Scale (MOS-Sleep, a questionnaire to as-sess sleep quality and quantity), Hospital Anxiety and DepressionScale (HADS, a self-administered questionnaire), SF-MPQ VAS-PartB, The Euro Quality of Life (EQ-5D, a questionnaire designed to as-sess health-related quality of life), 7-point scales completed by thepatient (Patient Global Impression of Change, PGIC) and investiga-tor (Clinical Global Impression of Change, CGIC), with possible re-sponses ranging from 1 (very much improved) to 7 (very muchworse) at visit 9. The SF-MPQ, HADS, and QANeP were completedat visits 1, 2, and 9. The NPSI, MOS-Sleep, and EQ-5D were com-pleted at visits 1 and 9.

Vital signs, AEs, treatment discontinuation, laboratory data, andconcomitant medications were recorded from the safety popula-tion (randomized patients who received P1 dose of study medica-tion). All AEs that occurred after the start of treatment, volunteeredby the patient or observed by the investigator, were documentedthroughout the study for their nature, onset, duration, severity,outcome, and any relationship to study drug. Clinical laboratorytests (ie, hematology, chemistry, and urinalysis), an electrocardio-gram, a physical examination including a neurological examina-tion, and a peripheral edema assessment were performed atvisits 1 and 9; postural vital signs were collected at each visit;weight was measured at every visit except for visit 6.

2.4. Statistical analyses

A total of 226 patients in the double-blind treatment phase wasneeded to achieve 90% power to detect a difference of P1 point onthe mean pain score at endpoint assuming a 2-sided test at the 5%level of significance, and a common standard deviation of 2.3. Effi-cacy analyses were performed on the intent-to-treat population,defined as randomized patients who received P1 dose of studymedication and completed P 1 post-baseline assessment. Missingefficacy data were to be replaced using the last observation carriedforward when applying statistical tests.

The primary efficacy endpoint was analyzed using an analysis ofcovariance model with the mean pain score at baseline as thecovariate, and treatment and country as factors. Mean Sleep Inter-

Assessed for eligibility(n = 283)

Pregabalin 150–600 mg/day Treated (n = 110) Completed (n = 93)

Placebo Treated (n = 109) Completed (n = 90)

Discontinued (n = 17) Related to study drug (n = 5) Adverse event (n = 5) Lack of efficacy (n = 0) Not related to study drug (n = 12) Adverse event (n = 4) Lost to follow-up (n = 0) Other (n = 2) Patient choice (n = 6)

Discontinued (n = 19) Related to study drug (n = 4) Adverse event (n = 3) Lack of efficacy (n = 1) Not related to study drug (n = 15) Adverse event (n = 1) Lost to follow-up (n = 7) Other (n = 3) Patient choice (n = 4)

Randomized(n = 220)*

Fig. 1. Patient disposition. ⁄One subject was randomized to the pregabalin group,but did not receive any study drug.

1020 J.S. Kim et al. / PAIN�

152 (2011) 1018–1023

ference score, NPSI, HADS, Health State Profile, VAS, and SF-MPQ-VAS were analyzed with the same method. Treatment differencesin the proportion of patients with a P30% or P50% reduction inweekly mean pain score at endpoint were analyzed using a Coch-ran–Mantel–Haenszel test stratified by country. The Optimal SleepScale of the MOS was analyzed using a logistic regression withtreatment, country, and the baseline assessment fitted as explana-tory variables. Descriptive statistics on absolute scores and changefrom baseline were calculated for each item of the QANeP. CGICand PGIC were analyzed using an analysis of variance model withtreatment group and country fitted as factors.

3. Results

3.1. Patients

Of 283 patients screened, 219 met the entry criteria and weretreated with either pregabalin 150 to 600 mg/day (n = 110) or with

Table 1Demographic characteristics of study patients.

Characteristic

Gender, n (%) MaleAge, y, mean (SD)

RangeRace, n (%) White

Asian

Mean duration of CPSP, yMean (Range)

Common medical conditions at baseline, n (%)HypertensionDyslipidemiaDiabetes mellitusDepressionConstipationInsomnia

Type of sensory abnormalityMechanical allodynia (static or dynamSkin pinch hyperalgesiaCold allodynia or hyperalgesiaHeat allodynia or hyperalgesiaNone of the above

CPSP, central post-stroke pain; SD, standard deviation.

placebo (n = 109), and 183 (83.5%) patients completed the study(Fig. 1). The mean dose (range) of pregabalin received by thepatients was 356.8 (125.0–539.7). Twenty-one (19%) patientsreceived pregabalin 150 to <300 mg/day, 39 patients (35.5%) 300to <450 mg/day, and 50 patients (45.5%) 600 mg/day. Of the pre-gabalin-treated patients who discontinued the study, 9 (8.2%) dis-continued because of AEs (5 [4.5%] considered treatment related),and 12 (10.9%) for reasons not related to the study drug; in the pla-cebo-treated group, 4 (3.7%) discontinued because of AEs (3 [2.8%]considered treatment related), and 15 (13.7%) for reasons not re-lated to the study drug. Patient demographic and clinical charac-teristics are summarized in Table 1. Of the pregabalin-treatedpatients, 109 (99.1%) patients were taking concomitant medica-tions during the study, as were 107 (98.2%) of the placebo-treatedpatients. Of all patients, 96% had at least 1 disease/syndrome inaddition to CPSP.

3.2. Efficacy

The weekly mean pain score decreased from baseline to endpointfor patients in both treatment groups (from 6.5 at baseline to 4.9 atendpoint for the pregabalin group, and from 6.3 at baseline to 5.0at endpoint for the placebo group) (Fig. 2). The least-squares (LS)mean difference between groups was –0.2 (95% confidence interval[CI] = –0.7 to 0.4) favoring pregabalin; however, this difference wasnot statistically significant (P = .578). A similar result was obtainedwhen the primary analysis was repeated for the per-protocol popu-lation comprising members of the intent-to-treat populationwithout any major protocol deviation. Country was not found to bea statistically significant effect modifier, nor were baseline insomniastatus, neuropathic pain status, or pain severity at baseline. At weeks3, 5, 6, and 8, the pregabalin group reported significantly less painthan the placebo group (Fig. 2). The majority of patients treated withpregabalin did not achieve a 30% or 50% reduction in mean pain scoreat endpoint compared with baseline, and the differences betweentreatment groups were not statistically significant.

For secondary outcomes (Table 2), pregabalin was associatedwith significant improvements compared with placebo for themean HADS-A anxiety scores—the LS mean difference betweenthe treatment groups at endpoint was –1.0 (95% CI = –1.8 to –0.2) favoring pregabalin (P = .015). For the PGIC and CGIC mea-sures, both patients and their treating physicians reportedimprovements; however, at endpoint, only the CGIC improvement

Pregabalin (N = 110) Placebo (N = 109)

67 (60.9) 70 (64.2)59.4 (9.8) 57.1 (10.2)34–85 35–849 (8.2) 10 (9.2)101 (91.8) 99 (90.8)

2.2 (0.1–17.7) 2.5 (0.2–14.1)

91 (82.7) 87 (79.8)40 (36.4) 41 (37.6)26 (23.65) 32 (29.4)7 (6.4) 10 (9.2)6 (5.5) 7 (6.4)3 (2.7) 7 (6.4)(N = 95) (N = 99)

ic) 46 (48.4) 51 (51.5)38 (40.0) 41 (41.4)62 (65.3) 58 (58.6)27 (28.4) 30 (30.3)11 (11.6) 13 (13.1)

654

Mea

n sc

ore

3210

0 1 2 3 4 5 6Week

Pregabalin (n = 110)Placebo (n = 109)

*p < 0.057 8 9 10 11 12

7

* * * *

Fig. 2. Weekly mean pain score.

J.S. Kim et al. / PAIN�

152 (2011) 1018–1023 1021

was significantly better in the pregabalin group (LS mean differ-ence –0.3 [95% CI = –0.6 to 0.0] favoring pregabalin, P = .049). Thephysicians reported 74.7% of patients improved with pregabalincompared with 66.0% in the placebo group. Statistically insignifi-cant differences favoring pregabalin over placebo were noted inmost of the other secondary efficacy measures such as mean sleepinterference score at endpoint, SF-MPQ VAS, 5 clinically relevantNPSI dimensions (Table 3), and EQ-5D VAS. Among the 9 QANePscales the pregabalin group experienced numerically greaterimprovements from baseline to endpoint compared to placebo on5 scales (static mechanical allodynia [�0.7 vs �0.5], dynamicmechanical allodynia [�0.8 vs �0.6], punctate hyperalgesia testarea [�0.7 vs �0.0], cold allodynia [�1.1 vs �0.3], and temporalsummation to tactile stimuli [�1.0 vs �0.5]). No difference fromplacebo on 2 (punctate hyperalgesia reference area [�0.2], and coldhyperalgesia test area [�0.9]), and placebo did somewhat numeri-cally better than pregabalin on 2 (sensory threshold [0.0 vs �0.1],and cold hyperalgesia reference area [0.0 vs �0.2]). On the MOSsleep scale (Table 4), pregabalin treatment resulted in significantdifferences in sleep quantity, sleep adequacy, snoring, and reducedthe sleep problems index compared with placebo. In addition, thepregabalin treated group reported significant improvements in theweekly mean sleep interference scores at weeks 1 through 8(P < .05).

3.3. Safety

In all, 77 patients (70.0%) in the pregabalin group and 60 pa-tients (55.0%) in the placebo group experienced AEs. The mostcommon AEs reported were dizziness, somnolence, peripheral ede-ma, and increased weight (Table 5). The dose of pregabalin was re-duced or temporarily discontinued in 20 patients because of AEs.Of those 20 patients, 17 had treatment-related AEs. Serious AEswere observed in 6 pregabalin patients versus 2 placebo patients.Only 1 patient in the pregabalin group experienced a treatment-re-

Table 2Secondary outcome measures.

Pregabalin PlaceboMeans

Baseline Week 12/LOCF Baseline Wee

SF-MPQ VAS 66.2 48.5 68.0 51.0NPSI 37.0 25.7 37.2 28.1HADS-A 7.7 5.8 7.5 6.7HADS-D 8.3 7.1 7.6 6.5EQ-5D utility 0.4 0.6 0.4 0.5EQ-5D VAS 56.9 64.1 58.8 61.4PGIC 2.9 3.1CGIC 2.8 3.1

CI, confidence interval; LS, least squares; LOCF, last observation carried forward; VAS, vpathic Pain Symptom Inventory; HADS, Hospital Anxiety and Depression Subscale; EQ-Global Impression of Change.

lated serious AE of edema. One patient receiving pregabalin expe-rienced severe serious AEs of hypoglycemia and dizziness thatwere considered to be unrelated to treatment. A total of 13 patientsdiscontinued the study because of AEs, of whom 8 discontined be-cause of treatment-emergent AEs. Overall, there were no clinicallymeaningful laboratory abnormalities, no remarkable physicalexaminations findings, or changes in vital signs over the courseof the study. There were no deaths during this study.

4. Discussion

In this trial, improvements in pain score from baseline were ob-served in both treatment groups with no significant differencebetween pregabalin and placebo arms at endpoint. There was anunexpectedly high placebo response—20.4% of patients receivingplacebo reported a P50% reduction in mean pain score frombaseline to endpoint. The high placebo response (and accompany-ing low rate of drop-outs because of lack of efficacy) have also beennoted in other recent trials in neuropathic pain (8–36% of patientsreporting a P50% reduction in mean pain score) [9,13,21].

A recent study in patients with peripheral neuropathic painidentified specific sensory profiles which have the potential toidentify subpopulations with better treatment response [2]. We at-tempted to discern whether any baseline characteristics influencedresponses, and found that baseline insomnia status, country,neuropathic pain status, and mean pain scores were not statisti-cally significant effect modifiers. Further analysis of baseline paincharacteristics might help to identify pregabalin responders in pa-tients with CPSP. For example, the presence, intensity or combina-tion of neuropathic symptoms based on the NPSI, or the presenceor absence of allodynia or hyperalgesia at baseline examinationmay affect the results outcome.

Although the primary endpoint was not met with statistical sig-nificance, we found that pregabalin produced greater pain relief upto 8 weeks versus placebo, and no loss of pain reduction was seenthereafter. However, whereas the pain score remained stable be-yond 8 weeks in the pregabalin group, pain reductions with pla-cebo increased gradually over time, contributing to the loss ofstatistical separation between the 2 groups at endpoint. In this re-spect, our study is consistent with the 4- to 8-week studies of ami-triptyline [16], lamotrigine [29], and pregabalin [30] reportedsignificant pain benefit; however, with the 12-week long durationof this study the significant difference of pregabalin versus placebowas lost. Evidence from other trials suggests a tendency for theplacebo response to increase in magnitude with longer study dura-tion [21]. Factors influencing placebo responses included symp-tomatic variations, expectation of the physicians and patients,conditioning, regression to the mean and environmental factors[3,21,28]. The magnitude of the increased placebo response was

Difference between LSmeans at week 12

95% CI P value

k 12/LOCF

�1.0 �7.0 to 5.0 .741�2.8 �6.5 to 0.9 .138�1.0 �1.8 to �0.2 .015

0.2 �0.6 to 1.0 .6000.0 �0.1 to 0.1 .5663.0 �1.8 to 7.9 .220�0.2 �0.5 to 0.1 .144�0.3 �0.6 to 0.0 .049

isual analogue scale; SF-MPQ, Short-Form McGill Pain Questionnaire; NPSI, Neuro-5D, Euro Quality of Life; PGIC, Patient Global Impression of Change; CGIC, Clinical

Table 3Neuropathic Pain Symptom Inventory.

Pregabalin Placebo Difference between LSmeans at week 12

95% CI P valueMeans

Baseline Week 12/LOCF Baseline Week 12/LOCF

Burning (superficial) spontaneous pain 3.6 2.6 3.3 2.9 �0.4 �1.0 to 0.2 .216Pressing (deep) spontaneous pain 3.7 2.7 3.8 2.6 0.1 �0.4 to 0.6 .760Paroxysmal pain 2.9 1.8 3.1 2.0 �0.2 �0.7 to �0.3 .360Evoked pain 4.0 2.9 3.9 3.1 �0.3 �0.8 to 0.2 .239Paresthesia/dysesthesia 4.1 2.8 4.3 3.3 �0.4 �0.1 to 0.1 .118Total Score 37.0 25.7 37.2 28.1 �2.8 �6.5 to 0.9 .138

CI, confidence interval; LS, least squares; LOCF, last observation carried forward.

Table 4Medical outcomes study sleep scale.

Sleep score Pregabalin Placebo Difference between LSmeans at week 12

95% CI P valueMeans

Baseline Week 12 Baseline Week 12

Sleep disturbance 41.6 27.5 42.2 32.7 –4.8 –10.3 to 0.7 .086Snoring 38.7 40.8 39.1 32.6 7.7 0.4 to 15.1 .039Short of breath/headache 17.5 10.9 19.4 14.4 –3.7 –9.1 to 1.6 .169Sleep quantity 6.3 6.9 6.5 6.6 0.4 0.0 to 0.7 .030Sleep adequacy 54.5 66.6 61.0 60.6 8.6 1.8 to 15.4 .013Sleep somnolence 41.2 40.3 38.0 36.9 2.1 –2.8 to 7.1 .399Sleep problems index 38.6 28.5 37.2 32.1 –4.2 –8.4 to 0.0 .049

CI, confidence interval; LS, least squares.

Table 5Frequently occurring adverse events (P5% patients in either treatment group).

Adverse event % of Patients with AEs, all causality(treatment related)

Pregabalin (N = 110) Placebo (N = 109)

Any AE 70.0 (51.8) 55.0 (22.9)Dizziness 28.1 (23.6) 7.3 (6.4)Somnolence 21.8 (20.9) 4.6 (3.7)Edema peripheral 10.0 (8.2) 2.8 (1.8)Headache 6.3 (2.7) 7.3 (1.8)Diarrhea 5.5 (1.8) 1.8 (0)Edema 5.5 (4.5) 0Increased weight 5.5 (5.5) 1.8 (1.8)Upper respiratory tract infection 2.7 (0) 5.5 (0)

AE, adverse event.

1022 J.S. Kim et al. / PAIN�

152 (2011) 1018–1023

unexpected. A much larger study would be required to have suffi-cient statistical power to detect the diluted treatment effect. Thelack of statistical significance between pregabalin and placebocould represent a limitation of the study rather than drug failure.The increased placebo response is now recognized by the Initiativeon Methods, Measurement, and Pain Assessment in Clinical Trials(IMMPACT) group that recommends research on the methodologi-cal aspects of chronic pain clinical trials to enhance their sensitivityand provide more meaningful evaluations of treatments forchronic pain [7].

Notably, a study reporting that pregabalin was effective in thetreatment of central pain [30]. In addition to the shorter durationof the study (4 weeks), this study was different from ours in thatonly a small portion of the patients had CPSP (n = 12, 30%). Anotherrecent study showed that pregabalin was effective on central paincaused by spinal cord lesions up to 12 weeks [23]. Our resultstherefore suggest a different degree of pregabalin efficacy accord-ing to the level of lesions. The presumed pathophysiology of CPSPis complex and includes hyperexcitation in the damaged sensorypathways, loss of inhibitory mechanisms in the sensory thalamusor in pathways projecting into the thalamus, or combination of

the 2 [12,14]. The variety and complexity of lesions at higher levelsof the spinothalamocortical pathway resulting in CPSP may help toexplain the challenge of effectively treating this condition—a singlemodality or mechanism of action may not be sufficient to treat it.In support of this observation, it has been shown that cortical stim-ulation is differentially effective in peripherally and centrally med-iated pain [17–19,22]. If processing of pain at the spinal level iswhere pregabalin has its effect, this could potentially explain itslesser effect on pain due to brain lesions compared with spinal cordor peripheral nerve pathologies. However, although the sites ofgabapentin and pregabalin action to modulate pain have not beenfully elucidated, studies in animal models [20,24,25] and in humanbeings [11] indicate both spinal and superaspinal sites of action. Fi-nally, although we attempted to select patients with central painonly, it is possible that pain resulting from spasticity and musculo-skeletal pathology was misdiagnosed as CPSP, limiting the poten-tial for pregabalin to show an effect on the primary efficacyendpoint.

Although interpretation of the pain effect is complicated by theplacebo response, we found that several secondary efficacy mea-sures improved significantly with pregabalin. Pregabalin producedstatistically significant differences compared with placebo in anx-iety (mean HADS-A scores) and sleep disturbances (MOS sleepscale measures of snoring, sleep quantity, sleep adequacy, sleepproblems index, and weekly mean sleep interference scores atweeks 1–8). In addition, the physicians’ and patients’ assessmentsof general status (CGIC and PGIC) correlated with these improve-ments although only the CGIC improvement was significantly bet-ter in the pregabalin group at endpoint. As patients with CPSPfrequently report concomitant anxiety and sleep disturbances[14], the significant effect of pregabalin on these 2 conditions is apossible explanation for the general impression of improvementexperienced by the patients and their physicians.

Generally, pregabalin was well tolerated by patients with CPSPand the most common AEs (dizziness, somnolence, peripheral ede-ma, and increased weight) were consistent with those reported inother patient populations. Although pain relief at endpoint with

J.S. Kim et al. / PAIN�

152 (2011) 1018–1023 1023

pregabalin was not significantly better than with placebo in thisstudy, pregabalin has demonstrated improvements in comorbidconditions such as sleep and anxiety, which contribute to thelow quality of life of CPSP patients, suggesting some clinical utilityof pregabalin in the management of CPSP.

Conflict of interest statement

Jong S. Kim and Guy Bashford received honoraria from servingas a consultant and on the speakers’ bureau of Pfizer Inc. Kevin T.Murphy and Raymond Cheung are employees and shareholdersof Pfizer Inc. Andrew Martin, the study statistician, was an employ-ee of Pfizer Inc when the study was conducted. Vardit Dror is anemployee of UBC Scientific Solutions. Authors had full access tothe data. Editorial support for the development of the manuscriptwas provided by UBC Scientific Solutions and funded by Pfizer Inc.

Acknowledgements

This study was sponsored by Pfizer Inc. Co-investigators were asfollows: Philippines: Raymond L. Rosales, Maria Salome N. Vios;Indonesia: Aboe Amar Joesoef, Jusuf Misbach; Hong Kong: Law-rence K. S. Wong; Taiwan: Han-Hwa Hu, Ming-Hong Chang,Shan-Jin Ryu; Australia: Robert D. Helme, Guy M. Bashford, Geof-frey K. Herkes, Bronwyn Williams, Kathryn Brooke, Stephan Alex-ander Schug, Denis S. Crimmins; India: Usha Kant Misra,Rangasetty Srinivasa, Sunder Subramaniam, Man Mohan Mehndir-atta, Ajit Kumar Roy, Avathvadi Venkatesan; China: Zhen Hong,ZhongXin Zhao, Liying Cui, Dongsheng Fan, XiaoPing Pan; Malay-sia: Mary Suma Cardosa, Ramani Vijayan, Azizi Ahmad; Pakistan:Farrukh Shahab, Nadir Ali Syed, Shoukat Ali; Thailand: SomchaiTowanabut, Wasuwat Kitisomprayoonkul, Yotin Chinvaru; SouthKorea: Jong S. Kim, Sun U. Kwon.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.pain.2010.12.023.

References

[1] Andersen G, Vestergaard K, Ingeman-Nielsen M, Jensen TS. Incidence of centralpost-stroke pain. Pain 1995;61:187–93.

[2] Baron R, Tolle TR, Gockel U, Brosz M, Freynhagen R. A cross-sectional cohortsurvey in 2100 patients with painful diabetic neuropathy and postherpeticneuralgia: differences in demographic data and sensory symptoms. Pain2009;146:34–40.

[3] Benedetti F, Pollo A, Lopiano L, Lanotte M, Vighetti S, Rainero I. Consciousexpectation and unconscious conditioning in analgesic, motor, and hormonalplacebo/nocebo responses. J Neurosci 2003;23:4315–23.

[4] Bowsher D. The management of central post-stroke pain. Postgrad Med J1995;71:598–604.

[5] Cassinari V, Pagni CA. Central pain: a neurosurgical survey. Cambridge,MA: Harvard University Press; 1969.

[6] Choi-Kwon S, Chung CK, Lee SK, Choi J, Han K, Lee EH. Quality of life afterepilepsy surgery in Korea. J Clin Neurol 2008;4:116–22.

[7] Dworkin RH, Turk DC, Peirce-Sandner S, Baron R, Bellamy N, Burke LB, ChappellA, Chartier K, Cleeland CS, Costello A, Cowan P, Dimitrova R, Ellenberg S, FarrarJT, French JA, Gilron I, Hertz S, Jadad AR, Jay GW, Kalliomaki J, Katz NP, KernsRD, Manning DC, McDermott MP, McGrath PJ, Narayana A, Porter L, Quessy S,Rappaport BA, Rauschkolb C, Reeve BB, Rhodes T, Sampaio C, Simpson DM,Stauffer JW, Stucki G, Tobias J, White RE, Witter J. Research designconsiderations for confirmatory chronic pain clinical trials: IMMPACTrecommendations. Pain 2010;149:177–93.

[8] Freeman R, Durso-Decruz E, Emir B. Efficacy, safety, and tolerability ofpregabalin treatment for painful diabetic peripheral neuropathy: findingsfrom seven randomized, controlled trials across a range of doses. Diabetes Care2008;31:1448–54.

[9] Gordh TE, Stubhaug A, Jensen TS, Arner S, Biber B, Boivie J, Mannheimer C,Kalliomaki J, Kalso E. Gabapentin in traumatic nerve injury pain: a randomized,double-blind, placebo-controlled, cross-over, multi-center study. Pain2008;138:255–66.

[10] Hansson P. Post-stroke pain case study: clinical characteristics, therapeuticoptions and long-term follow-up. Eur J Neurol 2004;11:22–30.

[11] Hayashida K, DeGoes S, Curry R, Eisenach JC. Gabapentin activates spinalnoradrenergic activity in rats and humans and reduces hypersensitivity aftersurgery. Anesthesiology 2007;106:557–62.

[12] Henry JL, Lalloo C, Yashpal K. Central poststroke pain: an abstruse outcome.Pain Res Manag 2008;13:41–9.

[13] Katz J, Finnerup NB, Dworkin RH. Clinical trial outcome in neuropathic pain:relationship to study characteristics. Neurology 2008;70:263–72.

[14] Kim JS. Post-stroke pain. Expert Rev Neurother 2009;9:711–21.[15] Kumar B, Kalita J, Kumar G, Misra UK. Central poststroke pain: a review of

pathophysiology and treatment. Anesth Analg 2009;108:1645–57.[16] Leijon G, Boivie J. Central post-stroke pain–a controlled trial of amitriptyline

and carbamazepine. Pain 1989;36:27–36.[17] Nandi D, Smith H, Owen S, Joint C, Stein J, Aziz T. Peri-ventricular grey

stimulation versus motor cortex stimulation for post stroke neuropathic pain. JClin Neurosci 2002;9:557–61.

[18] Owen SL, Green AL, Nandi DD, Bittar RG, Wang S, Aziz TZ. Deep brainstimulation for neuropathic pain. Acta Neurochir Suppl 2007;97:111–6.

[19] Owen SL, Green AL, Stein JF, Aziz TZ. Deep brain stimulation for the alleviationof post-stroke neuropathic pain. Pain 2006;120:202–6.

[20] Partridge BJ, Chaplan SR, Sakamoto E, Yaksh TL. Characterization of the effectsof gabapentin and 3-isobutyl-gamma-aminobutyric acid on substance P-induced thermal hyperalgesia. Anesthesiology 1998;88:196–205.

[21] Quessy SN, Rowbotham MC. Placebo response in neuropathic pain trials. Pain2008;138:479–83.

[22] Saitoh Y, Yoshimine T. Stimulation of primary motor cortex for intractabledeafferentation pain. Acta Neurochir Suppl 2007;97:51–6.

[23] Siddall PJ, Cousins MJ, Otte A, Griesing T, Chambers R, Murphy TK. Pregabalinin central neuropathic pain associated with spinal cord injury: a placebo-controlled trial. Neurology 2006;67:1792–800.

[24] Takeuchi Y, Takasu K, Ono H, Tanabe M. Pregabalin, S-(+)-3-isobutylgaba,activates the descending noradrenergic system to alleviate neuropathic pain inthe mouse partial sciatic nerve ligation model. Neuropharmacology2007;53:842–53.

[25] Tanabe M, Ono K, Honda M, Ono H. Gabapentin and pregabalin amelioratemechanical hypersensitivity after spinal cord injury in mice. Eur J Pharmacol2009;609:65–8.

[26] Taylor CP, Angelotti T, Fauman E. Pharmacology and mechanism of action ofpregabalin: the calcium channel alpha2-delta (alpha2-delta) subunit as atarget for antiepileptic drug discovery. Epilepsy Res 2007;73:137–50.

[27] Treede RD, Jensen TS, Campbell JN, Cruccu G, Dostrovsky JO, Griffin JW,Hansson P, Hughes R, Nurmikko T, Serra J. Neuropathic pain: redefinition and agrading system for clinical and research purposes. Neurology 2008;70:1630–5.

[28] Turner JA, Deyo RA, Loeser JD, Von Korff M, Fordyce WE. The importance ofplacebo effects in pain treatment and research. Jama 1994;271:1609–14.

[29] Vestergaard K, Andersen G, Gottrup H, Kristensen BT, Jensen TS. Lamotriginefor central poststroke pain: a randomized controlled trial. Neurology2001;56:184–90.

[30] Vranken JH, Dijkgraaf MG, Kruis MR, van der Vegt MH, Hollmann MW, HeesenM. Pregabalin in patients with central neuropathic pain: a randomized,double-blind, placebo-controlled trial of a flexible-dose regimen. Pain2008;136:150–7.