Intervention Review

You have free access to this content

Antiepileptics other than gabapentin, pregabalin, topiramate, and valproate for the prophylaxis of episodic migraine in adults

  1. Mattias Linde1,*,
  2. Wim M Mulleners2,
  3. Edward P Chronicle3,
  4. Douglas C McCrory4,5

Editorial Group: Cochrane Pain, Palliative and Supportive Care Group

Published Online: 24 JUN 2013

Assessed as up-to-date: 15 JAN 2013

DOI: 10.1002/14651858.CD010608


How to Cite

Linde M, Mulleners WM, Chronicle EP, McCrory DC. Antiepileptics other than gabapentin, pregabalin, topiramate, and valproate for the prophylaxis of episodic migraine in adults. Cochrane Database of Systematic Reviews 2013, Issue 6. Art. No.: CD010608. DOI: 10.1002/14651858.CD010608.

Author Information

  1. 1

    Norwegian University of Science and Technology, Department of Neuroscience, Trondheim, Norway

  2. 2

    Canisius Wilhelmina Ziekenhuis, Department of Neurology, Nijmegen, Netherlands

  3. 3

    University of Hawaii at Manoa, (Deceased) Department of Psychology, Manoa, USA

  4. 4

    Duke University Medical Center, Department of Medicine, Durham, NC, USA

  5. 5

    Durham Veterans Affairs Medical Center, Center for Health Services Research in Primary Care, Durham, NC, USA

*Mattias Linde, Department of Neuroscience, Norwegian University of Science and Technology, Trondheim, Norway. mattias.linde@ntnu.no.

Publication History

  1. Publication Status: Edited (no change to conclusions)
  2. Published Online: 24 JUN 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Cady 2009

MethodsProspective, randomised, double-blind, parallel-group trial. Two months baseline period. Duration of double-blind treatment phase: 14 weeks (2 weeks dose titration, 12 weeks maintenance)

Discontinuation rate: carisbamate 100 mg 35%, carisbamate 300 mg 28%, carisbamate 600 mg 29%, placebo 28%

Compliance (adherence) data: not available

Rule for use of acute medication: not reported

Methodological quality score: 5


ParticipantsInclusion: migraine with or without aura according to ICHD-II; migraine frequency of 3 to 12 attacks/month; history of migraine at least 1 year; diagnosis of migraine before age 50. Ages 18 to 65

Exclusion: excessive use of acute medication ((> 15 days/month or beyond approved daily dose of ASA unless for cardiovascular indication, other NSAIDs, acetaminophen ± metheptane mucate) or (> 10 days/month or beyond approved daily dose of triptans, ergots opioids whether or not combined with other analgesics)). Acceptable exclusion of secondary headaches. CDH not a formal exclusion criterion, but baseline values + exclusion of medication overuse do not imply high proportion of CDH. Other exclusions: Failed > 2 adequate prophylactic courses. Requirement for continued use of β-blockers, TCAs, Ca-channel blockers, antiepileptics, 5-HT2-antagonists, MAO inhibitors, daily NSAIDs, high doses of magnesium or riboflavin, botulinum toxin, herbal preparations like St. John's Wort or feverfew
Setting: 42 centres
Countries: Germany, Spain, and USA
Intention-to-treat analysis of 318 patients. 47% (151/323) of randomised patients had migraine with aura. 276 females and 40 males randomised; mean age 41.3 ± 10.68, age range 18 to 64. 83 allocated to carisbamate 100 mg, 81 allocated to carisbamate 300 mg, 79 allocated to carisbamate 600 mg, 80 allocated to placebo


InterventionsCarisbamate 100 mg/day versus carisbamate 300 mg/day versus carisbamate 600 mg/day versus placebo (14 weeks). Dosing schedules not reported except for 2 weeks initial titration


OutcomesHeadache frequency using a 48-hour rule and a 24-hour rule (all migraine headaches starting within 48 or 24 hours, respectively, of the onset of the original migraine). 50% or greater reduction in migraine frequency. Frequency of migraine attacks lasting longer than 30 min. Migraine days with use of rescue medication
Time point(s) considered in the review: through double-blind-phase (14 weeks)


NotesFunders of the trial: Johnson & Johnson Pharmaceutical Research & Development


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low risk1:1:1:1 randomisation, balanced by using randomly permuted blocks, stratified by centre, computer-generated schedule

Allocation concealment (selection bias)Low riskCoded treatment allocation, code implemented via IVRS, codes maintained at IVRS centre.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskClinicians and patients blinded. Study drugs provided according to treatment code, drugs and packaging of identical appearance

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo concern among the review authors over incomplete outcome data

Selective reporting (reporting bias)High riskNo means (SD) for the treatment period provided, only percentages (range) for changes from baseline. Corresponding author requested twice for complementary data without replying

de Tommaso 2007

MethodsProspective, randomised, double-blind, parallel-group trial. Two months baseline period. Duration of treatment: two months

Discontinuation rate: levetiracetam 0%, topiramate 20%, placebo 27%

Compliance (adherence) data: "non-compliance" reported as the most common reason for dropping out (6/7), but the extent to which patients took medications as prescribed is not reported

Rule for use of acute medication: not reported

Methodological quality score: 3


ParticipantsInclusion: migraine without aura according to ICHD-II; attack frequency not specified. Ages 18 to 49. Consent to additional neurophysiological tests

Exclusion: no adequate rule reported for exclusion of secondary headaches, daily headache, or analgesic overuse headache. However, no included patient had CDH or other headache than migraine according to correspondence with the first author. Other exclusions: psychoactive drugs, general/neurological/psychiatric disorders

Setting: single-centre

Country: Italy

Intention-to-treat analysis of 39 patients. Patients with aura not recruited. 35 females and 10 males included; mean age 37.9 ± 12.4, age range 18 to 49. 15 received levetiracetam, 15 received placebo, and 15 received topiramate


InterventionsLevetiracetam 1000 mg/day (500 mg BID, presumably tablets) versus topiramate 100 mg/day (50 mg BID, presumably tablets) versus placebo (8 weeks). Dose escalation strategy not reported


OutcomesMigraine days per month. 50% or greater reduction in migraine frequency. Flicker frequency dependent α-rhythm phase synchronisation (phase synchronisation index). Mean iCNV amplitude and iCNV habituation index

Time point(s) considered in the review: through entire treatment period (2 months)


NotesStudy was not principally designed to obtain clinical outcome data of the interventions

Funders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information

Allocation concealment (selection bias)Unclear riskNo information

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskStudy was double-blind, but methodological description is lacking

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo concern among the review authors over incomplete outcome data

Selective reporting (reporting bias)Low riskMeans (SD) for migraine frequency during treatment period not presented in publications but provided by first author after request

Ghose 2002

MethodsProspective, randomised, double-blind cross-over trial. 4-week baseline period. Duration of treatment: 12 weeks of either vigabatrin or placebo, then 4-week placebo-only washout period, then cross-over medication for 12 weeks

Discontinuation rate: dropout: 4 of 23 patients

Compliance (adherence) data: compliance: 4 of 23 patients excluded by tablet count or serum vigabatrin levels

Rule for use of acute medication: triptans allowed, others not specified

Methodological quality score: 3


ParticipantsInclusion: IHS migraine criteria, no other relevant physical or psychiatric disorder, normal CT scan of the head, no EEG evidence of epilepsy, non-smokers

Exclusions: secondary headache and analgesic overuse headache were adequately excluded. 5 patients with possible daily headache were included

Setting: single public hospital headache clinic

Country: not reported (likely New Zealand)

Complete case analysis of 15 patients. 8 patients with aura; 7 without. 10 females and 5 males, age range 28 to 66


InterventionsVigabatrin (12 weeks) versus placebo (12 weeks); repeat with cross-over. Dosage titrated up to 1000 mg twice/day


OutcomesNumber of migraine attacks per week. Severity (3-point scale), duration (hours), and a migraine index (frequency x duration x severity)

Time point(s) considered in the review: last (12th) week of treatment phase


NotesFunders of the trial: Marion Merrell Dow Pharmaceuticals New Zealand Limited


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskNo information on method for random sequence generation

Allocation concealment (selection bias)Unclear riskNo information

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBoth participants and clinicians blinded. Placebo tablets were identical to verum. Serum samples were coded to ensure that the analysis of drug levels was performed blindly

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
High riskComplete case analysis of 15 patients

Selective reporting (reporting bias)Low riskNo suspicion of selective reporting of outcomes, time points, or analyses

Gupta 2007

MethodsProspective, randomised, double-blind, triple cross-over trial. Study duration 23 weeks (although stated 20 weeks); 4 weeks prospective baseline followed by 4 weeks with first intervention, 1 week washout, cross-over to 4 weeks with second intervention, 1 week washout, cross-over to 4 weeks with third intervention, 1 week washout, and finally cross-over to 4 weeks with fourth intervention. Each subject received all treatments in a specified order

Discontinuation rate: lamotrigine 7%, topiramate 7%, lamotrigine placebo 7%, topiramate placebo 7%

Compliance (adherence) data: not available

Rule for use of acute medication: patients were allowed to take tablets with a combination of paracetamol and diclofenac potassium (supplied to them) at their choice

Methodological quality score: 2


ParticipantsInclusion: migraine with or without aura according to ICHD-I; migraine frequency of 4 to 10 attacks/month; history of migraine at least 1 year; debut of migraine before age 50; at least 48 h pain-free interval between attacks. Ages 18 to 65

Exclusion: headaches other than migraines. > 8 days/month of NSAIDs, ergots, or triptans. Paracetamol overuse and CDH not mentioned as exclusion criteria. Other exclusions: migraine prophylactic drug (or drug with such potential) last month, antipsychotic/antidepressant drug last 3 months, alcohol or other drug dependence, nephrolithiasis, participated in earlier study of lamotrigine or topiramate, used lamotrigine or topiramate 2 weeks or longer, used experimental drug last month

Setting: single-centre

Country: India

Intention-to-treat analysis of 57 patients. 32% (19/60) of included patients had migraine with aura. 47 females and 13 males; mean age 29.4 ± 7.7 years (range 16 to 48). 57 received lamotrigine, 57 received topiramate, 57 received lamotrigine placebo and 57 received topiramate placebo


InterventionsLamotrigine 50 mg/day, versus lamotrigine placebo, versus topiramate 50 mg/day, versus topiramate placebo (4 weeks). Lamotrigine and topiramate and were given as 25 mg tablet BID (stable dosage), and placebos as 1 tablet BID (stable dosage)


OutcomesChange in migraine frequency per 28 days compared to baseline. Proportion of responders (50% reduction in frequency). Responder rate migraine intensity. Attack frequency. Attack duration. Intensity on VAS. Phonophobia. Photophobia. Rescue medication use. Response to rescue medication. Aura frequency. "Reports of AEs communicated historically during visits, as transcribed on headache diaries"

Time point(s) considered in the review: through entire treatment period (4 weeks)


NotesUnclear if any participants had CDH. Shorter treatment periods (1 month) than recommended (3 months) by IHS for evaluation of efficacy in clinical trials. Since 2 potentially active drugs were used, there is an obvious risk of carry-over effect (analysis for order effects lacking)

Funders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated schedule allocating participants to one of four treatment arms: LTG – LPLAC-TOP-TPLAC or LPLAC-LTG-TPLAC-TOP or TOP-TPLAC-LTG-LPLAC or TPLAC-TOP-LPLAC-LTG

Allocation concealment (selection bias)Low riskPreprinted medication code labels. Sealed envelopes containing the code labels with a tear-off label concealing the randomisation number were provided to the investigator

Blinding of participants and personnel (performance bias)
All outcomes
High riskPatients and clinicians were blinded. Placebo was identical in appearance and packaging to active drug, but since the lamotrigine and topiramate tablets were different in appearance, 2 different placebos were used. For effective blinding a double-dummy design would be required. Study medication was packaged and labelled according to a medication code schedule generated before the trial. Each package had all 4 medications numbered according to the phase of the trial. Each bottle had a 2-part tear-off label; study medication identification was concealed and could be revealed only in case of emergency

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskInvestigators were blinded but not clearly stated that this included the stage of analysis

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo concern among the review authors over incomplete outcome data

Selective reporting (reporting bias)Low riskNo suspicion of selective reporting of outcomes, time points, subgroups, or analyses

Mohammadianinejad 2011

MethodsProspective, randomised, double-blind, parallel-group trial. Study duration: prospective (confirmed by correspondence with author) baseline of unclear duration followed by a 3 months treatment phase

Discontinuation rate: zonisamide 8%, topiramate 5%

Compliance (adherence) data: not available

Rule for use of acute medication: not described

Methodological quality score: 3


ParticipantsInclusion: migraine with or without aura according to ICHD (not stated which edition); migraine frequency of 4 to 15 attacks/month or < 4 attacks/month but so prolonged and debilitating that prophylaxis was required; history of migraine at least 1 year; debut of migraine before age 50. Age range not defined in publication, but corresponding author confirms that all were adults (17 to 58 years). Other inclusion: > 1 year since menopause (for postmenopausal women); negative pregnancy test and appropriate contraception; history of unsuccessful prophylactic treatment with one or more of the first-line migraine prophylaxis regimens; written informed consent obtained

Exclusion: any other primary headache, for example TTH or cluster headache. Overuse of acute medication and CDH not mentioned as exclusion criteria, but corresponding author confirms that no participant had CDH. Other exclusions: history of topiramate or zonisamide consumption; any other cause of pain that needs medications regularly, for example neuropathy or arthritis

Setting: single neurological clinic

Country: Iran

80 migraine patients participated. No information on proportion with migraine with aura. 63 females and 17 males; mean age zonisamide group 35.3 ± 9.5 years; mean age topiramate group 33.0 ± 9.2 years. 40 received zonisamide; 40 received topiramate


InterventionsZonisamide 200 mg/day versus topiramate 100 mg/day (12 weeks). Zonisamide capsules gradually titrated up from 50 mg/day to 200 mg/day within 1 month. Topiramate capsules gradually titrated up from 25 mg/day to 100 mg/day within 1 month


OutcomesHeadache frequency (number of migraine attacks per month). Response to treatment (decrease in frequency by more than 50%). Headache severity (VAS 1 to 10). Frequency of acute medication. MIDAS score. AEs

Time point(s) considered in the review: third month of treatment


NotesFunders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy was randomised (1:1), but method of sequence generation not described

Allocation concealment (selection bias)Low riskAn independent pharmacist packaged the drugs into numbered containers that were dispensed on the day of randomisation

Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskParticipants and clinicians were reported as blinded. Both drugs were capsules. Not stated that they were similar in appearance

Blinding of outcome assessment (detection bias)
All outcomes
Low riskAll active assessors were completely blinded

Incomplete outcome data (attrition bias)
All outcomes
High riskComplete case analysis, ie, patients discontinuing prematurely were excluded from efficacy analysis

Selective reporting (reporting bias)High riskThe percentages of 50% responders are not internally consistent. Absolute numbers requested twice from corresponding author but not provided. AEs are not adequately reported

Rompel 1970

MethodsProspective, randomised, double-blind, single cross-over trial. No baseline or washout periods. Total duration: 12 weeks

Discontinuation rate: dropout 6% for active treatment; 0% for placebo

Compliance (adherence) data: no compliance data reported

Rule for use of acute medication not reported

Methodological quality score: 4


ParticipantsInclusion: idiosyncratic criteria for migraine. Mixed and combination headaches may have been included, but subgroups cannot be distinguished

Exclusion: secondary headaches were adequately excluded. Neither daily headache nor analgesic overuse headache were adequately excluded

Setting: single migraine clinic

Country: South Africa

48 migraine patients participated; numbers with and without aura are not reported. 33 females, 15 males; age range 14 to 60. 45 received active treatment and 48 received placebo


InterventionsCarbamazepine (6 weeks) versus placebo (6 weeks). Precise dosage not reported


OutcomesNumber of migraine attacks in a 6-week period

Time point(s) considered in the review: through entire treatment phase (6 weeks)


NotesNot all participants were adults over 16 years of age

Funders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskA randomised table of numbers was used for sequence generation

Allocation concealment (selection bias)Low riskRandomisation codes were unavailable to the trialist

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBoth participants and clinicians were blinded. Placebo and verum were indistinguishable (supplied by Geigy South Africa (Pty) Limited)

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
High riskWithdrawals were excluded from statistical analysis

Selective reporting (reporting bias)Low riskNo suspicion of selective reporting of outcomes, time points, or analyses

Silberstein 2008

MethodsProspective, randomised, double-blind, parallel-group trial. Study duration: 4-week single-blind baseline phase (placebo given BID), thereafter a 15-week double-blind phase (6-week titration, 8-week maintenance, 1-week down-titration), followed by a 13-week open-label extension phase for patients who chose to continue or switch to oxcarbazepine therapy

Discontinuation rate: oxcarbazepine 34%, placebo 19%

Compliance (adherence) data: not available

Rule for use of acute medication: not reported

Methodological quality score: 5


ParticipantsInclusion: migraine with or without aura according to ICHD-I; migraine frequency of 3 to 9 attacks/month; history of migraine at least 1 year; debut of migraine before age 50; at least 48 h pain-free interval between attacks. Ages 16 to 65. Serum sodium levels ≥ 135 mEq/L at visit 1. Ability to read, write, and understand English. Capability of satisfying the requirements of the protocol and willingness and ability to give informed consent/assent according to legal requirements. No risk of pregnancy

Exclusion: ≥ 14 headache days with each headache lasting > 4 hours (of either migraine or non-migraine type) during the last 28 days of the single-blind baseline phase. Required symptomatic (acute) therapy more than 3 days/7 consecutive day period for a non-migraine headache during the last 28 days of the single-blind baseline phase. Patients requiring acute treatment for > 3 days/week in any 28-day period were discontinued from the study. Acceptable exclusion of secondary headaches. Other exclusions: missed > 20% of their expected doses of placebo during the last 28 days of the single-blind baseline phase. Missed ≥ 3 migraine diary entries during the last 28 days of the single-blind baseline phase. Previously failed > 3 standard courses of a commonly effective preventive migraine treatment. Having taken antidepressants (except SSRIs), beta-blockers, verapamil, diuretics, other AEDs, magnesium, herbal supplements, or > 50 mg/day of vitamin B2 within 1 month of study entry

Setting: 23 centres

Country: USA

Intention-to-treat analysis of 170 patients. Not reported how many had migraine with aura and how many had migraine without aura. 144 females and 26 males; mean age: allocated to oxcarbazepine 40.6 years (SD not reported, range 17 to 63); allocated to placebo 40.3 years (SD NR, range 17 to 68). 85 allocated to oxcarbazepine, 85 allocated to placebo


InterventionsOxcarbazepine 1200 mg/day versus placebo (15 weeks). During the 6-week titration period, oxcarbazepine was initiated at 150 mg/day and increased by 150 mg/day every 5 days to a to maximum tolerated dose of 1200 mg/day. At the investigator's discretion (based on poor tolerability), the dose could then be tapered downward, if necessary. Following step-down, the patient could be maintained at that dose level for the remainder of the titration phase, or the dose could be titrated up so the patient could reach his or her optimal dose. No further dose increases were allowed after the end of the 6-week titration period. At the end of the titration period, the median dose was 1200 mg/day (range 300 to 1200 mg/day). Upon completion of the 8-week maintenance period, or at premature discontinuation, patients were gradually withdrawn from study medication in a 1-week down-titration phase. Placebo tablets were administered BID during the baseline phase (not described for the double-blind period)

Time point(s) considered in the review: last 28-day period of the double-blind phase


OutcomesNumber of migraine attacks per 28-day period. Proportion of patients who responded to treatment with at least a 50% reduction in migraine attacks relative to baseline. Number of migraine days. Peak severity of migraine attacks. Acute therapy consumption. Number of non-migraine headache days during per 28 days. Migraine-related disability (MIDAS). Quality of life (SF-36). Clinical Global Impressions and Patient Global Impressions scales. AEs

Time point(s) considered in the review: last 4 weeks of the 15-week treatment phase


NotesFunders of the trial: Novartis Pharmaceuticals Corporation


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomisation was performed by a contracted outside clinical research organisation using a validated system that automates the random assignment of treatment groups to randomisation numbers

Allocation concealment (selection bias)Low riskAs patients qualified for entry, the investigator called into the IVRS for assignment of randomisation numbers

Blinding of participants and personnel (performance bias)
All outcomes
Low riskDuring the single-blind baseline-phase, only the patients were blinded to the treatment being administered. During the double-blind phase, treatment assignments were not revealed to patients, investigators, clinical staff, or study monitors until all patients completed therapy and the database was finalised. The study drug was packaged and labelled according to a medication code generated before the trial. Each bottle had a two-part tear-off: study medication was concealed and only revealed in case of emergency. Placebo tablets were matched in size to verum

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo concern among the review authors over incomplete outcome data

Selective reporting (reporting bias)High riskSDs for changes in least squares means of migraine frequencies lacking in publication (only SEs given). Supplementary information requested twice from corresponding author, but no reply. Change from baseline in number of non-migraine headache days during the last 28 days of the double-blind phase and the entire double-blind phase stated as a secondary outcome, but data not given

Steiner 1997

MethodsProspective, randomised, double-blind, parallel trial. Total duration: 4 months. 1-month single-blind, placebo-only baseline period. 3-month treatment period

Discontinuation rate: dropout: 31%

Compliance (adherence) data: compliance was assessed by count of returned tablets, but data are not reported

Rule for use of acute medication: Co-codamol encouraged, ergotamine discouraged, but some other medication also allowed

Methodological quality score: 5


ParticipantsInclusion: IHS migraine criteria; migraine onset at least 2 years prior to study; 2 to 8 attacks per month in the previous 3 months

Exclusion: secondary headaches, daily headaches, and analgesic overuse headaches were adequately excluded. Other exclusions: chronic or recurrent pain, hepatic or renal disease, depression, other migraine prophylactic medication in last 2 months, pregnancy, change in use of oral contraception, previous participation in more than 2 clinical trials

Setting: 2 secondary-referral centres

Country: UK

77 patients were randomised; 31 with aura and 46 without. 63 females and 14 males; age range 18 to 60


InterventionsLamotrigine (3 months) versus placebo (3 months). Target dosage 200 mg/day, but see notes


OutcomesNumber of migraine attacks per 28 days. Number of migraine days per 28 days. Severity (3-point scale). Number of doses of escape medication

Time point(s) considered in the review: last (third) month of treatment phase


NotesInitial protocol specified 200 mg dosage immediately, but adverse events led to an amendment in which dosage was titrated for the first 4 weeks

Funders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskRandom sequence generation was generated separately at each centre, but method not described

Allocation concealment (selection bias)Unclear riskNo information

Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants and clinicians were blinded. All medication was dispensed as matching hard-gelatin capsules manufactured by the Wellcome Foundation Ltd. To maintain blindness, equally many capsules of verum and placebo were given

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo information

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll randomised patients were included in the analysis

Selective reporting (reporting bias)Low riskNo suspicion of selective reporting of outcomes, time points, subgroups, or analyses

Stensrud 1979

MethodsProspective, randomised, double-blind, single cross-over trial. No washout period. 4-week baseline period. Total duration of double-blind period: 8 weeks. In an additional 4-week period, all participants received active treatment in higher dose

Discontinuation rate: dropout: 10.5% overall

Compliance (adherence) data: no compliance data reported

Rule for use of acute medication: analgesics or ergotamine

Methodological quality score: 1


ParticipantsInclusion: Ad Hoc Committee criteria; at least 3 attacks per month

Exclusion: no criteria reported

Setting: single outpatient clinic

Coutnry: not reported (likely Norway)

38 migraine patients participated; numbers with and without aura are not reported. 27 females and 11 males; age range 20 to 60


InterventionsClonazepam (4 weeks) versus placebo (4 weeks). Dosage 1 mg/day, then 2 mg/day for additional 4-week higher dose period


OutcomesNumber of headache days per 28 days. Headache index: sum of individual headache severity grades (1 to 3 scale) per 28 days

Time point(s) considered in the review: first month of treatment phase


NotesFunders of the trial: not reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy was randomised (to sequence A or B), but method of sequence generation not described

Allocation concealment (selection bias)Unclear riskNo information

Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants and clinicians were blinded. Placebo was used

Blinding of outcome assessment (detection bias)
All outcomes
High riskNo information on blinding of analysts. Both sequences incorporated a final 4-week period in which all patients received same dose (2 mg) of clonazepam

Incomplete outcome data (attrition bias)
All outcomes
Low riskNo concern among the review authors over incomplete outcome data

Selective reporting (reporting bias)Low riskNo suspicion of selective reporting of outcomes, time points, subgroups, or analyses

Vahedi 2002

MethodsProspective, randomised, double-blind parallel trial. 28-day baseline period. Duration of treatment: 12 weeks

Discontinuation rate: dropout 34% overall, 22% in placebo group and 46% in treatment group

Compliance (adherence) data: compliance data not reported

Rule for use of acute medication: patients' usual medication, except that more than 1 g aspirin daily was not permitted

Methodological quality score: 5


ParticipantsInclusion: IHS migraine criteria; migraine for more than 1 year; migraine frequency of 2 to 8 per month; no more than 6 days per month of interval headache; age of onset of migraine 50 years or younger; discontinuation of other migraine prophylaxis at least 6 weeks prior to the trial

Exclusion: secondary headaches, daily headaches, and analgesic overuse headaches were adequately excluded. Other exclusions: depression, use of antidepressant or antipsychotic drugs; women of childbearing potential not using contraception; relevant allergies; use of medication likely to interfere with potassium or acid-base balances; history of renal lithiasis, renal or hepatic insufficiency, hyperuricaemia, diabetes, hypokalaemia

Setting: multicentre

Country: 7 sites in France

53 patients enrolled. 48 migraine without aura; 5 with aura. 40 females and 13 males, age range 19 to 60. 27 received placebo and 26 active drug


InterventionsAcetazolamide 250 mg twice/day versus placebo (12 weeks). Dose reduction to 125 mg twice/day permitted in the case of bothersome side effects


OutcomesHeadache frequency in the final 28 days of treatment. Proportion of responders (50% reduction in frequency of attacks). Severity and duration of attacks also recorded

Time point(s) considered in the review: last (third) month of treatment phase


NotesPower calculations indicated the need for 90 patients in each arm of the trial. The study was discontinued prematurely because of the high frequency of adverse events. Hence only 53 patients were enrolled

Funders of the trial: Assistance Publique – Hôpitaux de Paris, GERMED GNE96003 at the Délégation à la Recherche Clinique of Saint-Louis Hospital and PHRCAOM97096 (Programme Hospitalier de Recherche Clinique of theFrench Ministry of Health)


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskStudy was randomised, but method for sequence generation not described

Allocation concealment (selection bias)Low riskRandomisation was performed centrally by fax

Blinding of participants and personnel (performance bias)
All outcomes
Low riskBoth participants and clinicians were blinded. Each package of trial medication contained 3 boxes of 5 blisters of 14 placebo or verum tablets. These were delivered by a pharmacist at each centre

Blinding of outcome assessment (detection bias)
All outcomes
Low riskNo information

Incomplete outcome data (attrition bias)
All outcomes
Low riskIntention-to-treat analysis of randomised patients

Selective reporting (reporting bias)High riskVariance measures lacking for migraine attack frequency during treatment

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Anthony 1972Insufficient evidence of randomisation or pseudo-randomisation

Chen 2001Reports case studies only

Cochran 2004Letter to editor

D'Andrea 1999No control group

Delvaux 2001No control group

Hedri 1975Reports case studies only

Lampl 1999Reports case studies only

Skupin 1975No control group

Smirne 1979No control group

Stieg 1977No arm of trial in which antiepileptic alone was given

Voto Bernales 1974No control group

 
Comparison 1. Acetazolamide versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 2. Carbamazepine versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 OR for responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 2 RR for responders (patients with ≥ 50% reduction in headache frequency)1Risk Ratio (M-H, Random, 95% CI)Totals not selected

 3 Any adverse event1Risk Difference (M-H, Random, 95% CI)Totals not selected

 4 Drowsiness1Risk Difference (M-H, Random, 95% CI)Totals not selected

 5 Nausea1Risk Difference (M-H, Random, 95% CI)Totals not selected

 6 Vertigo/giddiness1Risk Difference (M-H, Random, 95% CI)Totals not selected

 
Comparison 3. Carisbamate versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

    1.1 Carisbamate titrated to 100 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

    1.2 Carisbamate titrated to 300 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

    1.3 Carisbamate titrated to 600 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 
Comparison 4. Carisbamate dose comparisons

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

    1.1 Carisbamate 300 mg/day versus 100 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

    1.2 Carisbamate 600 mg/day versus 100 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

    1.3 Carisbamate 600 mg/day versus 300 mg/day
1Odds Ratio (M-H, Random, 95% CI)0.0 [0.0, 0.0]

 
Comparison 5. Clonazepam versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (post-treatment; headache days per month)1Mean Difference (IV, Random, 95% CI)Totals not selected

 
Comparison 6. Lamotrigine versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (change from baseline to post-treatment, or post-treatment alone)2190Mean Difference (IV, Random, 95% CI)-0.49 [-1.83, 0.85]

    1.1 Lamotrigine 50 mg/day
1113Mean Difference (IV, Random, 95% CI)-1.17 [-2.17, -0.17]

    1.2 Lamotrigine titrated to 200 mg/day
177Mean Difference (IV, Random, 95% CI)0.20 [-0.81, 1.21]

 2 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 7. Lamotrigine versus topiramate

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (change from baseline to post-treatment)1Mean Difference (IV, Random, 95% CI)Totals not selected

 2 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 8. Levetiracetam versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (post-treatment; headache days per month)1Mean Difference (IV, Random, 95% CI)Totals not selected

 2 OR for responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 3 RR for responders (patients with ≥ 50% reduction in headache frequency)1Risk Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 9. Levetiracetam versus topiramate

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (post-treatment; headache days per month)1Mean Difference (IV, Random, 95% CI)Totals not selected

 2 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 10. Oxcarbazepine versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (change from baseline to post-treatment)1Mean Difference (IV, Random, 95% CI)Totals not selected

 2 Responders (patients with ≥ 50% reduction in headache frequency)1Odds Ratio (M-H, Random, 95% CI)Totals not selected

 
Comparison 11. Vigabatrin versus placebo

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (post-treatment)1Mean Difference (IV, Random, 95% CI)Totals not selected

 
Comparison 12. Zonisamide versus topiramate

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Headache frequency (change from baseline to post-treatment)1Mean Difference (IV, Random, 95% CI)Totals not selected

 
Table 1. Numbers (percentages) of adverse events (AEs) in single dosage placebo-controlled studies

StudyType of AEAntiepilepticPlacebo

Vahedi 2002Acetazolamide (n = 26)Placebo (n = 27)

Paresthesia21 (81)2 (7)

Fatigue/drowsiness/memory problems/malaise15 (58)4 (15)

Gastrointestinal intolerance3 (12)2 (7)

Hypokalaemia1 (4)0 (0)

Hyperuricaemia1 (4)0 (0)

Skin eruption0 (0)2 (7)

Fever and shivering0 (0)1 (4)

Dry mouth1 (4)1 (4)

Breast tension0 (0)1 (4)

Rhinitis1 (4)2 (7)

Tinnitus0 (0)1 (4)

Miscellaneous1 (4)3 (11)

Rompel 1970Carbamazepine (n = 45)Placebo (n = 48)

Any AE30 (67)11 (23)

Vertigo, giddiness23 (51)2 (4)

Drowsiness5 (11)1 (2)

Nausea4 (9)3 (6)

Dry mouth2 (4)0 (0)

Heavy eyes2 (4)0 (0)

Constipation2 (4)0 (0)

Vomiting1 (2)0 (0)

Weight gain1 (2)1 (2)

Sweating1 (2)0 (0)

Transient rash1 (2)0 (0)

Dysuria1 (2)0 (0)

Blocked nose0 (0)1 (2)

Flushing0 (0)1 (2)

Blunted feeling0 (0)1 (2)

Heavy head0 (0)1 (2)

Cady 2009Carisbamate (n = 238)Placebo (n = 78)

Any AE183 (77)58 (74)

Fatigue41 (17)5 (6)

Nasopharyngitis30 (13)9 (12)

Nausea21 (9)6 (8)

Back pain12 (5)1 (1)

Diarrhoea11 (5)2 (3)

Dizziness11 (5)3 (4)

Somnolence8 (3)4 (5)

Upper abdominal pain7 (3)2 (3)

Vomiting7 (3)2 (3)

Hepatic enzyme increased6 (3)0

Migraine6 (3)5 (6)

Insomnia4 (2)4 (5)

Stensrud 1979Clonazepam (n = 38)Placebo (n = 38)

Sedation23 (61)Unclear

Dizziness10 (26)Unclear

Irritability2 (5)Unclear

Steiner 1997Lamotrigine (n = 37)Placebo (n = 40)

Rash11 (30)1 (3)

Dizziness4 (11)2 (5)

Asthenia2 (5)2 (5)

Somnolence2 (5)0 (0)

Thinking abnormality1 (3)0 (0)

Tremor0 (0)1 (3)

Leukopenia1 (3)1 (3)

Nausea0 (0)3 (8)

Weight gain0 (0)3 (8)

Allergy0 (0)1 (3)

Gupta 2007Lamotrigine (n = 57)Placebo (n = 57)

Sleepiness and concentration difficulty2 (4)0

Paresthesias2 (4)Unclear

Gastrointestinal intolerance2 (4)1

Anorexia1 (2)2 (4)

Giddiness2 (4)1 (2)

Rash2 (4)1 (2)

Palpitations02 (4)

Menorrhagia00

Hair loss00

Pain in lower limbs01 (2)

de Tommaso 2007Levetiracetam (n = 15)Placebo (n = 15)

Sedation and dizziness5 (33)Unclear

Silberstein 2008Oxcarbazepine (n = 85)Placebo (n = 85)

Any AE68 (80)55 (65)

Fatigue17 (20)6 (7)

Dizziness15 (18)6 (7)

Nausea14 (16)4 (5)

Somnolence7 (8)6 (7)

Balance disorder5 (6)2 (2)

Insomnia5 (6)6 (7)

Migraine5 (6)2 (2)

Paresthesias5 (6)1 (1)

Sinusitis2 (2)5 (6)