Systematic Review and Meta-Analysis

Flunarizine as prophylaxis for episodic migraine:

a systematic review with meta-analysis
Anker Stubberuda,*, Nikolai Melseth Flaaena, Douglas C. McCroryb,c,d, Sindre Andre Pedersene, Mattias Lindea,f

Abstract
Based on few clinical trials, flunarizine is considered a first-line prophylactic treatment for migraine in several guidelines. In this
meta-analysis, we examined the pooled evidence for its effectiveness, tolerability, and safety. Prospective randomized
controlled trials of flunarizine as a prophylaxis against migraine were identified from a systematic literature search, and risk of
bias was assessed for all included studies. Reduction in mean attack frequency was estimated by calculating the mean
difference (MD), and a series of secondary outcomes—including adverse events (AEs)—were also analyzed. The database
search yielded 879 unique records. Twenty-five studies were included in data synthesis. We scored 31/175 risk of bias items
as “high,” with attrition as the most frequent bias. A pooled analysis estimated that flunarizine reduces the headache frequency
by 0.4 attacks per 4 weeks compared with placebo (5 trials, 249 participants: MD 20.44; 95% confidence interval 20.61 to 2
0.26). Analysis also revealed that the effectiveness of flunarizine prophylaxis is comparable with that of propranolol (7 trials,
1151 participants, MD 20.08; 95% confidence interval 20.34 to 0.18). Flunarizine also seems to be effective in children. The
most frequent AEs were sedation and weight increase. Meta-analyses were robust and homogenous, although several of the
included trials potentially suffered from high risk of bias. Unfortunately, reporting of AEs was inconsistent and limited. In
conclusion, pooled analysis of data from partially outdated trials shows that 10-mg flunarizine per day is effective and well
tolerated in treating episodic migraine—supporting current guideline recommendations.
Keywords: Migraine, Headache, Flunarizine, Sibelium, Pharmacological prophylaxis, Pooled analysis, Systematic review, Meta-
analysis

1. Introduction infrequent, with weight increase and somnolence as the most

Flunarizine is one of many prophylactic treatment options for common, whereas depression and extrapyramidal symptoms are
episodic migraine. The drug is a nonselective calcium entry the most feared.50,63 Flunarizine is recommended in headache
blocker acting on slow calcium channels.1,46 It was presented treatment guidelines in several countries12,41,57 and is also
some 30 years ago as a prophylactic drug in migraine treatment3 considered a first treatment choice by the European Federation
but has not attained the same popularity as other cardiovascular of Neurologic Societies.17 However, despite these recommen-
drugs prescribed for migraine prophylaxis. Still, flunarizine is dations, flunarizine is not readily available in many countries.62
largely regarded as effective, inexpensive, and easy to use with its In addition to the limited availability of the drug, the guidelines
once-daily dosing. Adverse events (AEs) are regarded as recommending flunarizine are primarily based on individual
clinical trials.4–7,9,13–15,19,22,35,44,52 All but one of these trials are
over 20 years old and many of them do not adhere to current
Sponsorships or competing interests that may be relevant to content are disclosed
at the end of this article.
guidelines of clinical trials and migraine diagnosis.25,59,60 To
address this problem, we believe that there is a need for
A. Stubberud and N.M. Flaaen contributed equally as co-first authors.
a
a systematic review of the topic—providing pooled estimates on
Department of Neuromedicine and Movement Sciences, NTNU Norwegian
University of Science and Technology, Trondheim, Norway, b Duke Evidence
effectiveness, tolerability, and safety, and describing the quality of
Synthesis Group, Duke Clinical Research Institute, Durham, NC, United States, trials and their risk of bias.
c
Department of Medicine, Duke University School of Medicine, Durham, NC, United The primary aims of this meta-analysis are: (1) to retrieve and
States, d Center for Health Services Research in Primary Care, Durham Veterans describe the scientific quality of randomized controlled trials
Affairs Medical Center, Durham, NC, United States, e Library Section for Medicine
(RCTs) investigating flunarizine as migraine prophylaxis; and (2) to
and Health Sciences, NTNU University Library, NTNU Norwegian University of
Science and Technology, Trondheim, Norway, f Norwegian Advisory Unit on assess the pooled evidence of effectiveness, tolerability, and
Headaches, St. Olavs Hospital, Trondheim, Norway safety in these trials.
*Corresponding author. Address: Department of Neuroscience and Movement
Sciences, NTNU Norwegian University of Science and Technology, Nevro Øst,
Edvard Griegs gt 8, Trondheim 7030, Norway. Tel.: 004745229174. E-mail address: 2. Methods
ankers@stud.ntnu.no (A. Stubberud).
2.1. Protocol
Supplemental digital content is available for this article. Direct URL citations appear
in the printed text and are provided in the HTML and PDF versions of this article on A protocol for the systematic review was registered at
the journal’s Web site (www.painjournalonline.com). PROSPERO international prospective register of systematic
PAIN 160 (2019) 762–772 reviews. Protocol number: CRD42017057670, available from
© 2018 International Association for the Study of Pain http://www.crd.york.ac.uk/PROSPERO/display_record.php?
http://dx.doi.org/10.1097/j.pain.0000000000001456 ID5CRD42017057670.

762
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 April 2019
· Volume 160
· Number 4 www.painjournalonline.com 763

2.2. Criteria for considering studies for this review converted precision and variance data where appropriate and
possible. We anticipated that endpoints such as “headache
2.2.1. Types of studies
index”60 would be reported in a variety of ways—often by
Eligible studies were required to be prospective, randomized, or combining outcomes. We used such endpoints only if they could
pseudo-RCTs, evaluating the use of flunarizine as a prophylactic be converted to one of our desired outcomes. We chose to focus
drug for episodic migraine. Studies without an explicit description analyses on the third month of treatment and onwards as
as randomized were excluded. Studies were also required to be recommended by guidelines.60 For continuous data, we pre-
published in papers and available in typing with Latin alphabet. ferred end-of-treatment values over change scores, and
extracted change scores only if final values were unavailable.26
2.2.2. Types of participants From cross-over trials, we extracted data from the pre–cross-
over period to analyze these as parallel group trials. In cases
Included studies were not required to have strictly applied the where data on variance were still unavailable after attempts to
International Headache Society diagnostic criteria24,25 as long as calculate estimated variances based on primary data, we
the migraine diagnoses were based on their list of distinctive imputed variance data as the median value of variance data from
features, such as nausea/vomiting, severe pain, pulsating pain, the other studies. Sensitivity analyses were conducted by
unilaterality, photophobia/phonophobia, or aura. Trials combin- excluding studies with missing data. In cases where different
ing migraine and other headache types were excluded. studies reported AE synonyms, these were pooled into preferred-
term categories as defined in the Medical Dictionary for
2.2.3. Types of interventions Regulatory Activities (MedDRA) by the World Health Organization.
Characteristics of included studies were summarized with
The included studies were required to have at least one treatment description of the study design, interventions, participants,
arm where participants received flunarizine regularly during outcomes, and risk of bias assessments.
headache-free intervals to reduce the migraine burden. Accept-
able comparison groups included placebo or other pharmaco-
logical and nonpharmacological treatments with proven efficacy. 2.5. Data synthesis
Overuse of acute medication among trial participants led to Meta-analyses and figures were made using Review Manager
exclusion of said trial. (RevMan 5.3; The Nordic Cochrane Centre, The Cochrane
Collaboration, Copenhagen, Denmark). For continuous outcome
2.2.4. Types of outcomes data, the mean difference (MD) with 95% confidence intervals
(CIs) was calculated using an inverse variance fixed-effects
The primary outcome of interest was mean reduction in migraine model. In cases where outcome scales varied within the same
frequency. Secondary outcomes included proportion of res- analysis, and were not feasible to convert, the standardized MD
ponders ($50% reduction in migraine frequency), intensity and (SMD) was used. For dichotomous data, we calculated odds
duration of migraine headache, doses of acute medication, ratios (ORs) with 95% CI, using a fixed-effects Mantel–Haenszel
disability, quality of life, and AEs. model. For AEs, we calculated the risk difference (RD) with 95%
CI. We additionally computed numbers needed to treat to benefit
2.3. Search strategies and numbers needed to treat for an additional harmful outcome
(NNTH) for dichotomous data. Subgroup analyses were made of
We conducted a database search across the databases MED- different drug doses. Statistical heterogeneity was also calculated
LINE, Embase, and CENTRAL with assistance of a medical for each meta-analysis and addressed in cases where it was
research librarian.48 The query involved a combination of deemed problematic.
thesaurus and free-text terms optimized to identify RCTs on In cases where only one study was available, we calculated the
patients with migraine receiving flunarizine treatment (Supple- MD in migraine frequency or OR for response to treatment (in
mental digital content 1, available at http://links.lww.com/PAIN/ case migraine frequency was not reported).
A702). A search filter optimized for detecting clinically sound
treatment studies was used when searching in MEDLINE23 and
Embase.67 To identify other potentially relevant studies, refer- 2.6. Risk of bias assessment
ences listed in reviews on flunarizine were also hand searched. All included studies were evaluated for risk of bias. We used the
Cochrane Collaboration risk assessment tool, assigning bias
2.4. Data collection and management categories to “low,” “unclear,” or “high” risk. The bias categories
were: sequence generation/randomization; allocation conceal-
Two of the authors independently screened the search results ment; blinding; blinding of outcome assessment; reporting of
through titles and abstracts and compared their finding. Full-text incomplete outcome data; evidence of selective outcome
files of the potentially eligible references were retrieved and reporting; and other potential risks of bias.
reviewed for inclusion. Near-eligible studies were reported with We also planned for creating and analyzing funnel plots, but
reasons for exclusion. Two of the reviewers extracted data such analyses were not deemed appropriate, as the number of
independently (using data collection forms from previous studies for each analysis was too low.
Cochrane reviews on antiepileptics in migraine32–34), before
comparing and reconciling their findings. Disagreements were
resolved through discussion. 3. Results
Migraine frequency was converted to number of days or
3.1. Search results
attacks per 28-day (4-week) period, and migraine intensity scores
were converted to a 4-point scale to facilitate comparison across Figure 1 shows the flow of study selection. The database search
studies. We extracted data from tables and figures, and updated to November 13, 2017, yielded 879 records after

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 764
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

associated with the outcome—these bias categories were rated

as “unclear risk.” Furthermore, 12 of the studies were assigned
a “high risk” of selective reporting. Finally, 2 studies were
assigned a “high risk” of other bias—one for only including
women and2 the other for only including previous responders to
migraine prophylactics.13

3.3. Data analysis

Frequency data were extracted from figures in 3 publica-
tions22,54,58 and converted to 28-day periods in 4 stud-
ies.13,35,45,53 Data for responders to treatment were extracted
from figures in 4 studies.20,35,37,42 Data on headache intensity
were required to be extracted from a figure in one study2 and
converted to a 4-point Likert-scale (0-3) in 3 other studies.22,37,66
Use of abortive medication was reported as number of attacks
treated with abortive medication in one study15 and as number of
analgesic doses in another study.37 This necessitated estimating
the SMD. On the other hand, 2 trials comparing flunarizine with
acupuncture2,66 reported drug consumption as number of
participants stopping abortive drug use, allowing for estimation
of OR. Quality-of-life measures were analyzed as SMD because
the 2 studies that reported quality of life used different scales.65,66
Eighteen of 25 studies provided 10-mg doses of flunarizine; 3
used different doses at 5,38 15,10 and 20 mg65; one15 investigated
5 mg vs 10 mg; whereas the 3 pediatric studies39,53,54 provided 5-
mg doses. All the placebo-controlled trials in adults used 10-mg
doses. Propranolol doses varied throughout studies. Subgroup
analyses depending on dosage of propranolol were made for
Figure 1. Study flow diagram. migraine frequency to ensure detailed analyses of the primary
outcome. On the other hand, propranolol doses were merged for
analyses of headache intensity, headache duration, and drug
removal of duplicates. Of these, 765 were excluded through consumption, as fewer studies reported these outcomes.
screening of titles and abstracts. After reviewing the remaining
114 full texts, 13 were identified as duplicates with different titles.
Twenty-five of the 101 unique studies met all the eligibility criteria 3.4. Results of analyses
and were included in data synthesis, whereas the remaining 76
were excluded with stated reason (Supplemental digital content 3.4.1. Flunarizine vs placebo
2, available at http://links.lww.com/PAIN/A702). A summary of Flunarizine was superior to placebo in reducing migraine
the characteristics of the included studies is presented in Table 1 frequency after 3 months of active treatment (MD 20.44; 95%
(comprehensive characteristics in Supplemental digital content 3, CI 20.61 to 20.26; Fig. 4) in the pooled analysis of 5 studies (249
available at http://links.lww.com/PAIN/A702). participants13,20,35,45,58). A sensitivity analysis ignoring trials with
imputed data20,58 produced a similar estimate (MD 20.43; 95%
CI 20.60 to 20.25). Flunarizine also showed higher responder
3.2. Risk of bias
proportion than placebo (OR 8.86; 95% CI 3.57-22.0; Fig. 5) in
Of 175 risk of bias items scored, 34.3% were deemed as low, the pooled analysis of 3 studies (113 participants20,35,42). The
48.0% as unclear, and 17.7% as high (Fig. 2). At least one “high number needed to treat to benefit was 3 (95% CI 2-4), based on
risk” score was assigned to 19 of the 25 studies (Fig. 3). A an assumed control risk of 0.28 calculated from the baseline
“low risk” of selection bias score was assigned to 6 stud- migraine frequency of the control groups.
ies2,15,42,47,65,66 providing a description of a computer-generated
randomization and 2 studies15,66 providing a description of
appropriate allocation concealment—the remaining selection 3.4.2. Flunarizine direct dose comparisons
bias judgments were of “unclear risk.” “Low risk” of performance A single study (524 participants15) comparing 5-mg vs 10-mg
bias was assigned to 3 studies52,54,66 providing an accurate doses of flunarizine revealed no difference in effect on headache
description of blinding procedures, whereas 6 stud- frequency after 4 months of active treatment (MD 0.20; 95% CI 2
ies2,37,38,43,53,65 were deemed to have insufficient blinding of 0.08 to 0.48).
participants and personnel, and thus a “high risk” of bias. Three
studies provided sufficient description of blinding of outcome
3.4.3. Flunarizine vs propranolol
assessors.2,37,66 Ten studies8,13,22,38,42,47,53–56 assigned a “high
risk” of attrition bias because they made completers-only No difference between 10-mg flunarizine and all doses of
analyses without reporting reasons for withdrawals, or because propranolol (60-160 mg) was observed after 4 months of active
reasons for withdrawal were associated with the outcome. Five treatment (MD 20.08; 95% CI 20.34 to 0.18; Fig. 6) in the pooled
additional studies2,10,36,43,58 provided completers-only analyses analysis of 7 studies (1151 participants8,15,22,37,51,55,56). A
with limited attrition, or the reported reasons for attrition were not sensitivity analysis ignoring trials with imputed data8,22,51 showed

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 April 2019
· Volume 160
· Number 4 www.painjournalonline.com 765

Table 1
Characteristics of included studies in summary.
Allais et al.2 Methods Prospective, open RCT.
Participants Migraine without aura diagnosis according to IHS criteria. 160 participants; 150 completers; mean age 37.8 years; 160 females.
Interventions Flunarizine 10 mg/day vs acupuncture.
Outcomes 1, 3, 5, and 8.
Bordini et al.8 Methods Prospective, double-blind RCT.
Participants Migraine diagnosis according to IHS criteria. 45 participants; 38 completers; mean age 31.2 years; 41 females and 4 males.
Interventions Flunarizine 10 mg/day vs propranolol 60 mg/day vs flunarizine 1 propranolol 10 mg/day 1 60 mg/day.
Outcomes 1 and 8.
Cerbo et al.10 Methods Prospective, double-blind RCT.
Participants Characteristic migraine symptoms. 30 participants; 27 completers; age range 23 to 54 years; 14 females and 16 males.
Interventions Flunarizine 15 mg/day vs pizotifen 1.5 mg/day.
Outcomes 8.
Diamond and Methods Prospective, double-blind RCT.
Freitag13 Participants Two-year migraine history. 143 participants; 101 completers; mean age 35 years; 75 females and 26 males.
Interventions Flunarizine 10 mg/day vs placebo.
Outcomes 1.
Diener et al.15 Methods Prospective, double-blind RCT.
Participants Inclusion criteria: migraine as defined by IHS. 810 participants; 783 included in intention to treat analysis; median age 37 years;
658 females and 150 males.
Interventions Flunarizine 5 mg/day vs flunarizine 10 mg/day vs propranolol 160 mg/day.
Outcomes 1, 2, 4, 5, and 8.
Frenken and Nuijten20 Methods Prospective, double-blind RCT.
Participants Common or classic migraine as defined by IHS. 35 participants; 35 completers; age range 20 to 51 years; 29 females and 6
males.
Interventions Flunarizine 10 mg/day vs placebo.
Outcomes 1, 2, and 8.
Gawel et al.22 Methods Prospective, double-blind RCT.
Participants Migraine headache as defined by the World Federation of Neurology Research Group. 94 participants; 89 completers; mean age
35.7 years; 85 females and 9 males.
Interventions Flunarizine 10 mg/day vs propranolol 160 mg/day.
Outcomes 1, 3, 4, and 8.
Louis35 Methods Prospective, double-blind RCT.
Participants Classic or common migraine with throbbing or pulsating attacks. 58 participants; 58 completers; mean age 29 years; 29 females
and 29 males.
Interventions Flunarizine 10 mg/day vs placebo.
Outcomes 1, 2, and 8.
Louis and Spierings36 Methods Prospective, double-blind RCT.
Participants Classic or common migraine diagnosed according to IHS criteria. 75 participants; 72 completers; mean age 37 years; 40 females
and 32 males.
Interventions Flunarizine 10 mg/day vs pizotifen 2 to 3 mg/day.
Outcomes 1 and 8.
Ludin37 Methods Prospective, double-blind RCT.
Participants Headache attacks with characteristic features of migraine. 71 participants; 48 completers; mean age 34.3 years; 51 females and
20 males.
Interventions Flunarizine 10 mg/day vs propranolol 120 mg/day.
Outcomes 1, 2, 3, 4, 5, and 8.
Luo et al.38 Methods Prospective, open RCT.
Participants Migraine diagnosis according to IHS criteria. 150 participants; 126 completers; mean age 43 years; 90 females and 36 males.
Interventions Flunarizine 5 mg/day vs topiramate 25 to 100 mg/day vs flunarizine 1 topiramate 5 mg/day 1 25 to 100 mg/day.
Outcomes 1 and 8.
Lutschg and Methods Prospective, double-blind RCT.
Vassella39 Participants Children with classic or common migraine with characteristic migraine symptoms. 33 participants; 32 completers; mean age 10.5
years; 17 females and 16 males.
Interventions Flunarizine 5 to 10 mg/day vs propranolol 30 to 120 mg/day.
Outcomes 1 and 8.
(continued on next page)

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 766
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

Table 1 (continued)
Mentenopoulos et Methods Prospective, double-blind RCT.
al.42 Participants Migraine diagnosis according to IHS criteria. 30 participants; 15 completers; median age 44 years; 16 females and 4 males.
Interventions Flunarizine 10 mg/day vs placebo
Outcomes 2 and 8.
Mitsikostas and Methods Prospective, double-blind RCT.
Polychronidis43 Participants Migraine diagnosis according to IHS criteria. 44 participants; 41 completers; mean age 36.1 years; 31 females and 13 males.
Interventions Flunarizine 10 mg/day vs sodium valproate 1000 mg/day.
Outcomes 2 and 8.
Pini et al.45 Methods Prospective, double-blind RCT.
Participants Diagnosis of classic or common migraine. 20 participants; 29 completers; mean age 39.5 years; 24 females and 5 males.
Interventions Flunarizine 10 mg/day vs placebo.
Outcomes 1.
Rascol et al.47 Methods Prospective, double-blind RCT.
Participants Migraine diagnosis according to IHS criteria. 35 participants; 32 completers; median age 38 years; 25 females and 10 males.
Interventions Flunarizine 10 mg/day vs pizotifen 2.19 mg/day.
Outcomes 1 and 8.
Shimell et al.51 Methods Prospective, double-blind RCT.
Participants Migraine diagnosis according to IHS criteria. 58 participants; 49 completers; mean 34.5 years; 40 females and 17 males.
Interventions Flunarizine 10 mg/day vs propranolol 180 mg/day.
Outcomes 1 and 8.
Sorge and Marano53 Methods Prospective, double-blind RCT.
Participants Children with migraine diagnosed according to the Valquist criteria. 48 participants; 42 completers; mean age 10.6 years; 27
females and 21 males.
Interventions Flunarizine 5 mg/day vs placebo.
Outcomes 1, 4, and 8.
Sorge et al.54 Methods Prospective, double-blind cross-over trial.
Participants Children with migraine diagnosed according to the Valquist criteria. 70 participants; 63 completers; mean age 10.6 years; 36
females and 34 males.
Interventions Flunarizine 5 mg/day vs placebo.
Outcomes 1, 4, and 8.
Soyka and Oestreich55 Methods Prospective, double-blind RCT.
Participants Classic or common migraine with characteristic features. 87 participants; 69 completers; mean age 42.5 years; 51 females and
18 males.
Interventions Flunarizine 10 mg/day vs propranolol 120 mg/day.
Outcomes 1, 4, and 8.
Soyka and Oestreich56 Methods Prospective, double-blind RCT.
Participants Classic or common migraine with characteristic features. 434 participants; 336 completers; mean age 42 years; 265 females and
61 males.
Interventions Flunarizine 10 mg/day vs propranolol 120 mg/day.
Outcomes 1, 4, and 8.
Sørensen et al.58 Methods Prospective, double-blind cross-over trial.
Participants Migraine diagnosis according to IHS criteria, modified by Olesen et al. 29 participants; 27 completers; median age 40 years; 23
females and 6 males.
Interventions Flunarizine 10 mg/day vs placebo.
Outcomes 1.
Sørensen52 Methods Prospective, double-blind RCT.
Participants Migraine diagnosis according to IHS criteria. 149 participants; 127 completers; median age 42 years; 118 females and 31 males.
Interventions Flunarizine 10 mg/day vs propranolol 120 mg/day.
Outcomes 1 and 8.
Vijayalakshmi et al.65 Methods Prospective, open RCT.
Participants Migraine diagnosis according to IHS criteria. 60 participants.
Interventions Flunarizine 20 mg/day vs acupuncture.
Outcomes 6.
Wang et al.66 Methods Prospective single-blind RCT.
Participants Migraine diagnosis according to IHS criteria. 140 participants; 120 completers; mean age 39.5 years; 119 females and 21 males.
Interventions Flunarizine 10 mg/day vs acupuncture.
Outcomes 1, 3, 5, 6, and 8.
1 5 migraine frequency; 2 5 responders to treatment; 3 5 migraine intensity; 4 5 headache duration; 5 5 drug consumption; 6 5 quality of life; 7 5 disability; 8 5 adverse events.
IHS, International Headache Society; RCT, randomized controlled trial.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 April 2019
· Volume 160
· Number 4 www.painjournalonline.com 767

(581 participants15,37) revealed no difference between the 2

drugs (OR 1.19; 95% CI 0.86-1.64). Using an assumed control
risk from the control groups in the included studies, at 0.19, the
number needed to treat to benefit in favor of flunarizine was 36 (CI
not defined).
For secondary outcomes in flunarizine vs propranolol trials, 2
studies (135 participants22,37) showed no difference in intensity of
migraine attacks after 4 months of treatment (MD 0.22; 95% CI 2
0.12 to 0.57); 5 studies (1063 participants15,22,37,55,56) showed
no difference in headache duration after 4 months of treatment
(MD 0.60; 95% CI 21.48 to 2.69); and 2 studies (583
participants15,37) demonstrated no difference in use of abortive
drugs between the groups (SMD 0.07; 95% CI 20.09 to 0.23).

3.4.4. Flunarizine vs pizotifen

Flunarizine was superior to pizotifen, with a larger percentage
reduction in migraine frequency after 4 months of treatment (MD
27.86; 95% CI 222.82 to 7.11) in the pooled analysis of 2
studies.36,47 A third10 study was excluded from the analysis due
to difference in flunarizine dosage and time point for data
reporting.

3.4.5. Flunarizine vs drugs other than propranolol or pizotifen

A single trial (127 participants52) comparing flunarizine with
metoprolol found no difference in migraine frequency after 3
months of treatment (MD 20.10; 95% CI 21.08 to 0.88). One
study (41 participants43) comparing flunarizine with sodium
valproate found no difference between the drugs (OR 1.07;
95% CI 0.28-4.12). A third parallel design and open trial (83
participants38) compared flunarizine with topiramate. At 3
months, no significant difference was found between the 2
treatments with respect to migraine frequency (MD 20.30; 95%
CI 20.97 to 0.37).

3.4.6. Flunarizine vs acupuncture

Acupuncture was superior to flunarizine in reducing migraine
frequency after 3 months of active treatment (MD 1.01; 95% CI
0.48-1.54) in the pooled estimate of 2 studies (290 partic-
ipants2,66). Acupuncture treatment also had a better effect on
migraine intensity (MD 0.26; 95% CI 0.06-0.46) and drug
consumption (OR 0.41; 95% CI 0.21-0.77).
Pooled analysis of 2 studies (200 participants65,66) showed
higher quality of life after one month of treatment in the
acupuncture group compared with the flunarizine group with
respect to both psychological (SMD 0.79; 95% CI 0.50-1.09) and
physical domains (SMD 0.76; 95% CI 0.45-1.06).

3.4.7. Flunarizine in children

Two placebo-controlled trials (105 participants53,54) of flunarizine
in children showed a reduction in migraine frequency after 3
months of active treatment (MD 21.14; 95% CI 21.51 to 20.77).
The same 2 studies (105 participants) also found a shorter
duration of headache in the flunarizine group (MD 20.46; 95% CI
Figure 2. Distribution of risk of bias assessments, presented as percentages 20.77 to 20.16).
across all included studies. A single study (32 participants39) found flunarizine to be more
efficient than propranolol in reducing migraine frequency in
a similar result (MD 20.07; 95% CI 20.33 to 0.20). Figure 6 children after 3 months of treatment (MD 22.00; 95% CI 23.05 to
shows the effect estimates for different doses of propranolol. A 20.95). However, after 4 months of treatment, the children
pooled analysis of 2 trials comparing responders to treatment responded better to propranolol (MD 0.96; 95% CI 0.53-1.39).

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 768
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

Figure 3. Judgment for each risk of bias item for each included study.

3.4.8. Safety and tolerability depression. Extrapyramidal symptoms were reported in 1 of 25

studies52—among 2.7% (2/74) of the flunarizine users during the
Adverse events were reported in 3 of 6 placebo-controlled trials.
run-in phase. No extrapyramidal symptoms were observed
Flunarizine users did not have higher risk of experiencing any one
during or after flunarizine treatment in any of the included studies.
or more AEs, compared with placebo (RD 0.04; 95% CI 20.08 to
The reported data on AEs in the 2 placebo-controlled trials of
0.17; Fig. 7) in the pooled analyses of these trials.20,35,42 The
flunarizine in children were insufficient for meta-analysis. One of
following mild-to-moderate AEs were reported in the placebo-
these (48 participants53) reported that 3 of 24 participants
controlled trials: Weight gain (NNTH 6; CI not defined); daytime
discontinued due to AEs, whereas the other study (70 partic-
sedation (NNTH 8; 95% CI 4-50); stomach complaints (NNTH not
ipants54) reported weight gain in 14 and drowsiness in 6 of all
defined); and dry mouth (NNTH not defined).
analyzed participants.
No serious AEs were reported in any of the placebo-controlled
trials and only one flunarizine-treated participant withdrew due to
AEs.58 The single study15 comparing doses of flunarizine found
4. Discussion
that 88 of 263 (33.5%) participants in the 5-mg group
experienced one or more AEs, whereas 88 of 275 (32%) Our meta-analysis shows that active flunarizine treatment
participants in the 10-mg group experienced one or more AEs. reduces the migraine frequency by approximately 0.4 attacks
None of the trials comparing flunarizine with active treatment per month compared with placebo (results from 5 studies with
reported any serious AEs. Six studies (1133 partici- a median baseline migraine frequency of 3.4 attacks per month).
pants8,15,22,51,55,56) of flunarizine vs propranolol found no To confirm our assumptions on imputing data, a sensitivity
difference in the occurrence of any AEs (RD 20.04; 95% CI 2 analysis was conducted, which showed no significant alteration
0.09 to 0.02). Figure 8 gives a summary of the frequency of AEs in the result of the analysis. Migraine sufferers treated with
reported in more than one of the flunarizine vs propranolol trials. flunarizine were also more likely to experience a 50% or greater
Two combined AE categories were created, the first including reduction in headache frequency compared with placebo—one
synonyms for sedation and somnolence, and the second out of every 3 participants showed this response. Neither 5-mg
including synonyms for fatigue and asthenia. The flunarizine vs nor 10-mg doses of flunarizine were superior in terms of
pizotifen trials had insufficient reporting of AEs to allow for meta- effectiveness and tolerability. Flunarizine seems to be noninferior
analysis. Finally, 2 trials of flunarizine vs acupuncture (270 to propranolol, pizotifen, metoprolol, valproate, and topiramate in
participants2,66) found a higher proportion of AEs among reduction in migraine frequency, but acupuncture seems to be
flunarizine users (RD 0.15; 95% CI 0.07-0.23). more effective. Flunarizine seems to have the same risk of AEs as
Depression was only reported in 3 of 25 studies2,15,52—in total placebo, but the pooled data for this analysis were limited and
2.9% (20/683) of the flunarizine users. In one of these studies, included trials with potential bias. The most prominent AEs from
a flunarizine vs propranolol trial,15 7/263 of 5-mg dose flunarizine the placebo-controlled trials were weight increase and daytime
users and 2/275 of 10-mg flunarizine users experienced sedation, with NNTHs at 6 and 8, respectively.

Figure 4. Forest plot of flunarizine vs placebo for migraine frequency. 95% CI, 95% confidence interval; (1), SDs imputed; (2), SD calculated from individual patient
data; (3), point estimates extracted from figures; IV, inverse variance; SD, standard deviation.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 April 2019
· Volume 160
· Number 4 www.painjournalonline.com 769

Figure 5. Forest plot of flunarizine vs placebo for responders to treatment ($50% reduction in migraine frequency). 95% CI, 95% confidence interval; (1), data
extracted from figures; (2), data extracted from figures; (3), data extracted from figures; M-H, Mantel-Haenszel.

Our main findings are in line with those of other systematic propranolol. Consequently, this study was weighed at 87.0% in
reviews. A Spanish meta-analysis from 2003,49 including 4 RCTs, the meta-analysis for headache frequency and highlights the
found that flunarizine reduced monthly migraine frequency with importance of conducting sufficiently powered studies.
0.55 compared with placebo. A network meta-analysis from Flunarizine has acquired a reputation to induce depression and
201528 also found an advantage of flunarizine over placebo, but extrapyramidal symptoms.11,16,18 Despite this, depression was
because this study combined data on migraine frequency and rarely reported and extrapyramidal symptoms were not reported
headache index these findings are not directly comparable with in any included studies during or after flunarizine treatment. We
ours. In addition, the meta-analysis also included a nonrandom- found daytime sedation and weight increase to be the most
ized trial.61 common AEs. This is in line with the results of a large open study
Despite positive findings, most of the placebo-controlled trials with 1435 participants.40 However, of the 6 placebo-controlled
currently available lack sufficient power to properly assess the trials, only one reported weight increase20 and 2 reported daytime
effect size of the intervention. In fact, several of the studies are sedation20,35—it is therefore possible that similar AEs may have
underpowered in their sample size, and none provides sample gone unreported in other studies. In the propranolol analyses, we
size calculations. A power analysis reveals that a sample size of decided to analyze AEs reported by 2 or more trials. According to
64 participants is required in each treatment arm to identify the hierarchical categorization of the MedDRA, several low-level
a significant difference given an effect size of 0.5 and a power of term synonyms for AEs were reported in the included studies. We
0.8 at the 0.05 significance level.27 Only one of the placebo- therefore chose to pool these into categories encompassing
controlled parallel trials recruited more participants.13 Similarly, preferred terms within the same high-level categories. This
the sample sizes for most individual trials investigating flunarizine resulted in 2 combined preferred-term categories. The first
vs active comparators were far too low, for noninferiority included somnolence and sedation synonyms representing AEs
analysis.30 Only one study15 provided sample size calculations, related to disturbances in consciousness, and the other included
concluding with a necessary sample size of over 260 participants fatigue and asthenia synonyms representing asthenic conditions.
per arm to prove that flunarizine was at least as effective as Together, this serves to give a direct comparison of flunarizine

Figure 6. Forest plot of flunarizine vs propranolol for migraine frequency. 95% CI, 95% confidence interval; (1), data extracted from figures; IV, inverse variance; SD,
standard deviation.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 770
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

Figure 7. Forest plot of flunarizine vs placebo for adverse events. 95% CI, 95% confidence interval; M-H, Mantel-Haenszel.

AEs with those of the commonly used propranolol. Interestingly, studies. Another strength is also that we translated papers
flunarizine gives a higher risk of fatigue, whereas propranolol gives from several other languages than English (16 different
a higher risk of somnolence. Ultimately, flunarizine seems to be countries). Only a few Chinese papers were not translated,
a well-tolerated alternative for patients with contraindications for yet none of these studies had placebo or active drugs as
beta blockers, such as obstructive pulmonary disease and controls, and we can conclude that the pooled estimates were
bradyarrhythmias, but such diagnoses were excluded from unaffected. Finally, we have made comprehensive reviews,
several flunarizine vs propranolol trials. descriptions, and thorough assessments of risk of bias for all
Data on AEs in the pediatric trials were limited and lacked included studies.
transparency. This makes us reluctant to draw conclusions and A limitation of this review is the variability and incompleteness of
compare it with AE findings in the adult trials. Still, the most data in the included studies. This required us to complete a series
frequent AEs in children were, similarly as for adults, weight gain of conversions and calculations from scarce primary data to allow
and drowsiness. Furthermore, many earlier systematic reviews for pooled analysis of the eligible studies. In some studies, we also
and guidelines recommend flunarizine for children, with the same had to impute missing variance data. This is hypothesized not to
source of tolerability findings as we present in this arti- introduce bias21 but still makes the pooled estimate less certain.
cle.17,31,63,64 On the other hand, it is possible that these estimates Nonetheless, omitting all studies with missing variance data could
are somewhat high as a recent retrospective study observed AEs have yielded a biased point estimate because these studies may
in 10/166 (6.0%) pediatric flunarizine users.29 not be a random subset of all studies.21 However, the sensitivity
We took several steps to reduce between-study heteroge- analyses indicate that the assumptions made on imputing data
neity issues in this study. First, we used strict criteria for are valid. One should also keep in mind the limitations of the AE
inclusion and avoided merging different drugs, populations, analyses due to heterogeneous and often incomplete reporting in
interventions, comparisons, and outcomes. We made the many studies. For example, 2 studies55,56 analyzed effectiveness
choice to strictly compare flunarizine to treatments with proven of data only from participants with “accepted rating sheets” but
efficacy, which excluded studies with comparators such as still reported AEs from all participants. If we assume all dropouts
dihydroergocryptine, dihydroergotamine, and calcium channel were due to ineffectiveness, there could potentially be a large
blockers.17 In addition, the use of pragmatic criteria to define mismatch between the reported effect and the number of AEs.
migraine in included studies, despite the fact that specific Similar attrition bias might also have been present in several of the
diagnose criteria have changed over the years, justifies pooling included studies.

Figure 8. Distribution of adverse events reported in more than one study for trials of flunarizine vs propranolol. AEs, adverse events.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 April 2019
· Volume 160
· Number 4 www.painjournalonline.com 771

Current evidence indicates that 10-mg flunarizine is as effective [12] Dansk Hovedpine Selskab. Referenceprogram, Diagnostik og behandling af
as other well-established alternatives, such as propranolol, but hovedpinesygdomme og ansigtssmerter. Dansk Hovedpine Selskab 2.
utgave, 2010.
with an AE profile focused on fatigue, somnolence, and weight [13] Diamond S, Freitag F. A double blind trial of flunarizine in migraine
increase. Guidelines give grade A recommendation to flunarizine prophylaxis. Headache Q 1993;4:169–72.
as migraine prophylaxis, derived from results presented in [14] Diamond S, Schenbaum H. Flunarizine, a calcium channel blocker, in the
individual and, to a large extent, old studies. This review supports prophylactic treatment of migraine. Headache 1983;23:39–42.
this recommendation, but our conclusion is mainly based on the [15] Diener H, Matias-Guiu J, Hartung E, Pfaffenrath V, Ludin H, Nappi G,
Beukelaar F. Efficacy and tolerability in migraine prophylaxis of flunarizine
same sources. Methodological quality issues in the included in reduced doses: a comparison with propranolol 160 mg daily.
studies—several of them involves substantial risks of bias— Cephalalgia 2002;22:209–21.
hamper us from concluding whether today’s limited use of [16] Donnet A, Vuillaume De Diego E, Lanteri-Minet M. Possible precipitation
flunarizine represents healthy skepticism or a neglect of a sub- of migraine attacks with prophylactic treatment. Eur Neurol 2009;61:
23–6.
group of patients in need of additional prophylactic drug options.
[17] Evers S, Afra J, Frese A, Goadsby PJ, Linde M, May A, Sandor PS. EFNS
To avoid simply putting a new timestamp on something that is guideline on the drug treatment of migraine—revised report of an EFNS
outdated, new placebo-controlled RCTs meeting the latest task force. Eur J Neurol 2009;16:968–81.
methodological standards are required. [18] Fabiani G, Pastro PC, Froehner C. Parkinsonism and other movement
disorders in outpatients in chronic use of cinnarizine and flunarizine. Arq
Neuropsiquiatr 2004;62:784–8.
[19] Freitag F, Diamond S, Diamond M. A placebo controlled trial of flunarizine
Conflict of interest statement in migraine prophylaxis. Cephalalgia 1991;11(suppl 11):157–8.
The authors have no conflict of interest to declare. [20] Frenken C, Nuijten S. Flunarizine, a new preventive approach to migraine.
A double-blind comparison with placebo. Clin Neurol Neurosurg 1984;
86:17–20.
[21] Furukawa TA, Barbui C, Cipriani A, Brambilla P, Watanabe N. Imputing
Acknowledgements missing standard deviations in meta-analyses can provide accurate
results. J Clin Epidemiol 2006;59:7–10.
The authors thank the personnel at the Library for Health and [22] Gawel MJ, Kreeft J, Nelson RF, Simard D, Arnott WS. Comparison of the
Medicine at NTNU Norwegian University of Science and efficacy and safety of flunarizine to propranolol in the prophylaxis of
Technology for assistance with identifying and retrieving papers. migraine. Can J Neurol Sci 1992;19:340–5.
No specific funding was provided for the systematic review. [23] Haynes RB, McKibbon KA, Wilczynski NL, Walter SD, Werre SR. Optimal
search strategies for retrieving scientifically strong studies of treatment
from Medline: analytical survey. BMJ 2005;330:1179.
[24] Headache Classification Committee of the International Headache
Appendix A. Supplemental digital content Society (IHS). Classification and diagnostic criteria for headache
disorders, cranial neuralgias and facial pain. Cephalalgia 1988;8(suppl
Supplemental digital content associated with this article can be 7):1–96.
found online at http://links.lww.com/PAIN/A702. [25] Headache Classification Committee of the International Headache
Society (IHS). The International Classification of Headache Disorders,
Article history: 3rd edition (beta version). Cephalalgia 2013;33:629–808.
Received 15 October 2018 [26] Higgins JPT, Green S (editors). Cochrane Handbook for Systematic
Reviews of Interventions Version 5.1.0 [updated March 2011]: The
Received in revised form 24 November 2018 Cochrane Collaboration, 2011. Available from http://
Accepted 29 November 2018 handbook.cochrane.org.
Available online 3 December 2018 [27] Houle TT, Penzien DB, Houle CK. Statistical power and sample size
estimation for headache research: an overview and power calculation
tools. Headache 2005;45:414–18.
[28] Jackson JL, Cogbill E, Santana-Davila R, Eldredge C, Collier W, Gradall A,
References Sehgal N, Kuester J. A comparative effectiveness meta-analysis of drugs for
[1] Agnoli A. The classification of calcium antagonists by the WHO expert the prophylaxis of migraine headache. PLoS One 2015;10:e0130733.
committee: relevance in neurology. Cephalalgia 1988;8(suppl 8):7–10. [29] Kim H, Byun SH, Kim JS, Lim BC, Chae JH, Choi J, Kim KJ, Hwang YS,
[2] Allais G, Lorenzo C, Quirico P, Airola G, Tolardo G, Mana O, Benedetto C. Hwang H. Comparison of flunarizine and topiramate for the prophylaxis of
Acupuncture in the prophylactic treatment of migraine without aura: pediatric migraines. Eur J Paediatr Neurol 2013;17(1):45–9.
a comparison with flunarizine. Headache 2002;42:855–61. [30] Laster LL, Johnson MF. Non-inferiority trials: the “at least as good as”
[3] Amery WK. Flunarizine, a calcium channel blocker: a new prophylactic criterion. Stat Med 2003;22:187–200.
drug in migraine. Headache 1983;23:70–4. [31] Lewis D, Ashwal S, Hershey A, Hirtz D, Yonker M, Silberstein S; American
[4] Amery WK, Caers LI, Aerts TJ. Flunarizine, a calcium entry blocker in Academy of Neurology Quality Standards Subcommittee; Practice
migraine prophylaxis. Headache 1985;25:249–54. Committee of the Child Neurology Society. Practice parameter:
[5] Balkan S, Aktekin B, Zülküf Ö. Efficacy of flunarizine in the prophylactic pharmacological treatment of migraine headache in children and
treatment of migraine. Gazi Med J 1994;5:81–4. adolescents report of the American Academy of Neurology Quality
[6] Bassi P, Brunati L, Rapuzzi B, Alberti E, Mangoni A. Low dose flunarizine Standards Subcommittee and the Practice Committee of the Child
in the prophylaxis of migraine. Headache 1992;32:390–2. Neurology Society. Neurology 2004;63:2215–24.
[7] Bono G, Manzoni GC, Martucci N, Baldrati A, Farina S, Cassabgi F, De [32] Linde M, Mulleners WM, Chronicle EP, McCrory DC. Antiepileptics other
Carolis P, Nappi G. Flunarizine in common migraine: Italian cooperative than gabapentin, pregabalin, topiramate, and valproate for the
trial. II. Long-term follow-up. Cephalalgia 1985;5(suppl 2):155–8. prophylaxis of episodic migraine in adults. Cochrane Database Syst
[8] Bordini C, Arruda M, Ciciarelli M, Speciali J. Propranolol vs flunarizine vs Rev 2013;6:CD010608.
flunarizine plus propranolol in migraine without aura prophylaxis. A [33] Linde M, Mulleners WM, Chronicle EP, McCrory DC. Gabapentin or
double-blind trial. Arq Neuropsiquiatr 1997;55:536–41. pregabalin for the prophylaxis of episodic migraine in adults. Cochrane
[9] Centonze V, Magrone D, Vino M, Caporaletti P, Attolini E, Campanale G, Database Syst Rev 2013;6:CD010609.
Albano O. Flunarizine in migraine prophylaxis: efficacy and tolerability of 5 [34] Linde M, Mulleners WM, Chronicle EP, McCrory DC. Topiramate for the
mg and 10 mg dose levels. Cephalalgia 1990;10:17–24. prophylaxis of episodic migraine in adults. Cochrane Database Syst Rev
[10] Cerbo R, Casacchia M, Formisano R. Double-blind clinical trial of 2013;6:CD010610.
flunarizine against pizotifen in a single evening dose in patients with [35] Louis P. A double-blind placebo-controlled prophylactic study of
migraine. Riv Neurol 1985;55:139–46. flunarizine (Sibelium) in migraine. Headache 1981;21:235–9.
[11] Chouza C, Scaramelli A, Caamano JL, De Medina O, Aljanati R, Romero [36] Louis P, Spierings E. Comparison of flunarizine (Sibelium) and pizotifen
S. Parkinsonism, tardive dyskinesia, akathisia, and depression induced (Sandomigran) in migraine treatment: a double-blind study. Cephalalgia
by flunarizine. Lancet 1986;1:1303–4. 1982;2:197–203.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.
 772
·
A. Stubberud et al. 160 (2019) 762–772 PAIN®

[37] Ludin H. Flunarizine and propranolol in the treatment of migraine. [54] Sorge F, Simone R, Marano E, Nolano M, Orefice G, Carrieri P. Flunarizine
Headache 1989;29:219–24. in prophylaxis of childhood migraine. A double-blind, placebo-controlled,
[38] Luo N, Di W, Zhang A, Wang Y, Ding M, Qi W, Zhu Y, Massing M, Fang Y. crossover study. Cephalalgia 1988;8:1–6.
A randomized, one-year clinical trial comparing the efficacy of topiramate, [55] Soyka D, Oestreich W. Flunarizine versus propranolol in migraine
flunarizine, and a combination of flunarizine and topiramate in migraine prophylaxis—A multicenter double-blind study in 12 hospitals.
prophylaxis. Pain Med 2012;13:80–6. Nervenheilkunde 1987;6:177–83.
[39] Lutschg J, Vassella F. Flunarizine and propranolol in the treatment of migraine [56] Soyka D, Oestreich W. Therapeutic effectiveness of flunarizine and
in children [in German]. Schweiz Med Wochenschr 1990;120:1731–6. propranolol in the interval therapy of migraine. Cephalalgia 1987;7(suppl
[40] Martinez-Lage JM. Flunarizine (Sibelium) in the prophylaxis of migraine. 6):467–8.
An open, long-term, multicenter trial. Cephalalgia 1988;8(suppl 8):15–20. [57] Steiner TJ, Paemeleire K, Jensen R, Valade D, Savi L, Lainez MJ, Diener
[41] Matre C, Linde M, Ljøstad U, Mygland Å. Migrene. 2018. Available at: HC, Martelletti P, Couturier EG. European principles of management of
http://nevro.legehandboka.no/handboken/sykdommer/hodepine/ common headache disorders in primary care. J Headache Pain 2007;
migrene/migrene2/. Accessed 15 October 2018. 8(suppl 1):S3–S47.
[42] Mentenopoulos G, Manafi T, Logothetis J, Bostantzopoulou S. [58] Sørensen P, Hansen K, Olesen J. A placebo-controlled, double-blind, cross-
Flunarizine in the prevention of classical migraine: a placebo-controlled over trial of flunarizine in common migraine. Cephalalgia 1986;6:7–14.
evaluation. Cephalalgia 1985;5(suppl 2):135–40. [59] Tfelt-Hansen P, Bjarnason NH, Dahlof C, Derry S, Loder E, Massiou H;
[43] Mitsikostas D, Polychronidis I. Valproate versus flunarizine in migraine Task Force on Adverse Events in Migraine Trials of Subcommittee on
prophylaxis: a randomized, double-open, clinical trial. Funct Neurol 1997; Clinical Trial of International Headache Society. Evaluation and
12:267–76. registration of adverse events in clinical drug trials in migraine.
[44] Nappi G, Sandrini G, Savoini G, Cavallini A, de Rysky C, Micieli G. Cephalalgia 2008;28:683–8.
Comparative efficacy of cyclandelate versus flunarizine in the prophylactic [60] Tfelt-Hansen P, Pascual J, Ramadan N, Dahlof C, D’Amico D, Diener HC,
treatment of migraine. Drugs 1987;33(suppl 2):103–9. Hansen JM, Lanteri-Minet M, Loder E, McCrory D, Plancade S, Schwedt
[45] Pini L, Ferrari A, Guidetti G, Galetti G, Sternieri E. Influence of flunarizine T; International Headache Society Clinical Trials Society. Guidelines for
on the altered electronystagmographic (ENG) recordings in migraine. controlled trials of drugs in migraine: third edition. A guide for
Cephalalgia 1985;5(suppl 2):173–5. investigators. Cephalalgia 2012;32:6–38.
[46] Quarterman CP, Kendall MJ, Jack DB. The effect of age on the [61] Thomas M, Behari M, Ahuja G. Flunarizine in migraine prophylaxis: an
pharmacokinetics of metoprolol and its metabolites. Br J Clin Pharmacol Indian trial. Headache 1991;31:613–15.
1981;11:287–94. [62] UpToDate. Flunarizine (United States: Not available): Drug information,
[47] Rascol A, Montastruc J, Rascol O. Flunarizine versus pizotifen: a double- 2018. Available at: https://www.uptodate.com/contents/flunarizine-
blind study in the prophylaxis of migraine. Headache 1986;26:83–5. united-states-not-available-drug-information?source5search_result&search5
[48] Rethlefsen ML, Murad MH, Livingston EH. Engaging medical librarians to Flunarizine&selectedTitle51;12 - F11233776. Accessed 29 November
improve the quality of review articles. JAMA 2014;312:999–1000. 2018.
[49] Reveiz-Herault L, Cardona AF, Ospina EG, Carrillo P. Effectiveness of [63] Vecsei L, Majlath Z, Szok D, Csati A, Tajti J. Drug safety and tolerability in
flunarizine in the prophylaxis of migraine: a meta-analytical review of the prophylactic migraine treatment. Expert Opin Drug Saf 2015;14:667–81.
literature [in Spanish]. Rev Neurol 2003;36:907–12. [64] Victor S, Ryan SW. Drugs for preventing migraine headaches in children.
[50] Schmidt R, Oestreich W. Flunarizine in migraine prophylaxis: the clinical Cochrane Database Syst Rev 2003:CD002761.
experience. J Cardiovasc Pharmacol 1991;18(suppl 8):S21–26. [65] Vijayalakshmi I, Shankar N, Saxena A, Bhatia M. Comparison of
[51] Shimell C, Fritz V, Levien S. A comparative trial of flunarizine and effectiveness of acupuncture therapy and conventional drug therapy on
propranolol in the prevention of migraine. S Afr Med J 1990;77:75–7. psychological profile of migraine patients. Indian J Physiol Pharmacol
[52] Sørensen PS, Larsen BH, Rasmussen MJ, Kinge E, Iversen H, Alslev 2014;58:69–76.
T, Nohr P, Pedersen KK, Schroder P, Lademann A. Flunarizine versus [66] Wang L, Zhang X, Guo J, Liu H, Zhang Y, Liu C, Yi J, Wang L, Zhao J, Li S.
metoprolol in migraine prophylaxis: a double-blind, randomized Efficacy of acupuncture for migraine prophylaxis: a single-blinded,
parallel group study of efficacy and tolerability. Headache 1991;31: double-dummy, randomized controlled trial. PAIN 2011;152:1864–71.
650–7. [67] Wong SSL, Wilczynski NL, Haynes RB. Developing optimal search
[53] Sorge F, Marano E. Flunarizine v. placebo in childhood migraine. A strategies for detecting clinically sound treatment studies in EMBASE.
double-blind study. Cephalalgia 1985;5(suppl 2):145–8. J Med Libr Assoc 2006;94:41.

Copyright © 2018 by the International Association for the Study of Pain. Unauthorized reproduction of this article is prohibited.