Memantine for prophylaxis of episodic migraine: a systematic review and meta-analysis

Background Migraine is a common debilitating primary headache. Memantine is a non-competitive N-methyl D-aspartate (NMDA) antagonist that lowers neuronal excitability that could prevent migraine attacks. Objectives This study aimed to determine the efficacy and safety of memantine in patients experiencing episodic migraine attacks using a systematic review and meta-analysis. Methods We searched CENTRAL, MEDLINE, Scopus, Cochrane, LILACS, ClinicalTrials.gov, HERDIN and Google Scholar for relevant studies. Prespecified screening and eligibility criteria for inclusion were applied. Included studies underwent methodological quality assessment. Study design, patient characteristics, interventions given, and relevant outcomes were extracted and synthesized. Results This review included five relevant articles – 2 randomized controlled trials (RCT) and 3 observational studies (1 retrospective case-series, 2 prospective non-randomized open-label single arm trials). There were 109 patients included in the RCTs and 197 patients reported in the observational studies. Pooled data from the 2 RCTs showed that Memantine at 10mg/day decreased the monthly migraine days at 12 weeks compared to placebo with a mean difference of -1.58 [95% confidence interval (CI) -1.84, -1.32]. Observational studies also showed a decrease in migraine days per month with Memantine (5-20mg/day) after 12 weeks [95% CI]: -9.1 [-11, -7.23], -7.2 [-8.85, -5.55], and -4.9 [-6.29, -3.51]. There were no statistically significant adverse drug events (ADE) noted in patients treated with Memantine compared to placebo. Conclusion Memantine may be effective and well-tolerated as prophylaxis for episodic migraine.

studies; and case series and reports. We included studies that involved patients 18 years old and above who were clinically diagnosed with episodic migraine (having less than 15 migraine days per month) according to the International Headache Society's (IHS) International Classification of Headache Disorders (ICHD-II or III) and previous versions [12] [2]. Studies that enrolled patients who received migraine prophylaxis in the past 3 months as well as those diagnosed of chronic migraine and/or medication-overuse headache (MOH) were excluded. We considered studies that employed Memantine alone or in combination with other prophylactic drugs regardless of drug dosage and duration of treatment

Primary and secondary outcomes
The primary outcome for this review was headache frequency measured as mean number of migraine days per month. To determine treatment effect, the mean difference was computed between mean migraine days per month at a particular time period of exposure to treatment and baseline.

Secondary outcomes included: 1) intensity of headaches measured via pain Numerical
Rating Score (NRS), which is a scale from 0 (no pain) to 10 (most severe pain) and Head Impact Test (HIT) that ranges from 36 to 78 with higher scores indicating increase in severity of pain, 2) adverse drug events, and 3) Migraine Disability Assessment Scale (MIDAS), which is a 5-item tool that measures migraine-related disability and functional consequences where a higher score corresponds to more severe disability; Hospital Anxiety and Depression Scale (HADS), which is a 14-item questionnaire that assesses mood status on a 4-point scale wherein a higher score means higher level of anxiety and depression; Pittsburgh sleep quality index (PSQI), which is a 19-item questionnaire to assess sleep quality with higher scores indicating poorer sleep; and Short-form 12 (SF), which is a 12-item tool for mental and physical health-related quality of life [9].

Search methods for identification of studies
Literature search using medical subject headings (MeSH) and free texts related to ("migraine") and ("Memantine") were used. The following databases were systematically Titles and abstracts of studies were screened according to the set criteria. Full-text articles were obtained for relevant studies and were reviewed based on the eligibility criteria to determine inclusion in this review. Included studies were then subjected to qualitative and quantitative analyses.

Risk bias assessment, data collection and analysis
Risk of bias was assessed for each included study. The Cochrane Collaboration Tool was used for RCTs. For case reports/series, the tool developed by Murad [13] was used to assess for bias. This tool combined eight items that were categorized into four domains: selection, ascertainment, causality, and reporting. More than the numerical aggregate score representation of the binary choices, an overall judgement about methodological quality should be based on the questions deemed most critical in the specific clinical scenario [13]. For this review's clinical query, questions on selection, ascertainment, and follow-up are most important to determine risk of bias.
Year publication, study design, population size and baseline characteristics, dose and titration rate of Memantine, duration of treatment, follow-up, number of migraine days, intensity of headaches, measures for depression, cognitive performance, well-being, and quality of life, withdrawals, and adverse drug events were extracted from each included study.
For continuous outcomes, mean differences (MD) with 95% CI were employed as measure of treatment effect. For dichotomous outcomes, risk ratios (RR) with 95% CI were used.
Data were synthesized using the Review Manage (RevMan) 5.3 (Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). We pooled continuous outcomes using the inverse-variance method and dichotomous outcomes using the Mantel-Haenszel method. Meta-analysis was performed using the fixed-effects model. For continuous outcomes, statistically significant difference between the intervention and control groups were noted if the 95% CI of the mean difference (MD) did not include the number 0. For dichotomous outcomes, statistical significance was detected if the 95% of the RR did not include the number 1. Statistical heterogeneity was measured using the I 2 statistic with values 0 to 40% indicating unimportant heterogeneity 3 Results

Included studies
As shown in Fig. 1, identification, screening, and application eligibility criteria to relevant studies were done. Eight databases were searched, which generated a total of 698 records (PubMed: 24; Cochrane: 7; Google scholar: 105; review of references: 7; Scopus: 555; ClinicalTrials.gov, controlled-trials.com, LILACS, HERDIN: 0). Fifteen duplicates were removed. Screening of titles and abstracts was done for 683 reports from which 676 records were excluded. Full-text articles for 7 studies were obtained and evaluated for eligibility. Two studies were excluded since they were review articles. Five studies fulfilled the inclusion criteria. The two randomized controlled trials were analyzed for metaanalysis.

Study design and patient population
Five studies satisfied the inclusion criteria. All these studies were published in peerreviewed journals -3 of which in international journals (Charles 2007 [14]; Bigal 2008 [6]; Noruzzadeh 2015 [7]), one local journal in Iran (Assarzadegan 2017 [15]), and one local journal in India (Shanmugam 2019 [16]). Shown in Table 1  Noruzzadeh 2015 had stricter exclusion criteria that did not consider patients who had medication-overuse headaches, use of antipsychotics or antidepressants in the past 3 months, recent history of alcohol or drug abuse, allergy to Memantine, treatment-resistant migraine, pain disorder, severe psychiatric illness, severe infection, malignancy, low survival chance, severe cardiovascular disease, neurodegenerative disorders, pregnancy and lactation, and sexually active women. In comparison, Shanmugam 2019 excluded patients with headaches that did not respond to more than two migraine preventive medications, pregnant/breastfeeding women, medication-overuse headaches, severe medical illness, renal insufficiency, hepatic problems, and hypersensitivity.

Methodological quality assessment
Bias was evaluated using the Cochrane Collaboration's tool for the RCTs of Noruzzadeh 2015 and Shanmugam 2019. The trials were randomized with low risk for selection, performance, detection, attrition, and reporting bias indicating an excellent level of methodological quality as seen in Fig. 2 For the three included observational studies, bias was assessed using the tool proposed by Murad 2017 as seen in Table 2. All three observational studies had no comparators, thus alternative explanations for the noted effects could not be fully verified. There was also no challenge/re-challenge done in all studies. As for dose-response effect, the authors graded this to be "unclear" since maximum dose was determined by the dose at which patients were satisfied with pain control and no adverse effects were present. The remaining questions were satisfied by all three studies. Therefore, all observational studies satisfactorily addressed the necessary aspects for good methodological quality.  Table 3 are the primary and secondary outcomes from included studies.

Migraine days
The main outcome for all studies focused on the mean difference for migraine days per month from baseline to 12 weeks follow-up.
Data used for meta-analysis is presented in Fig. 3 ADEs noted in the RCTs were pooled with subgroup analysis per event and reported as risk ratio (95% CI) as seen in Fig. 4.   According to the Global Burden of Disease Study last 2010, migraine ranked third in prevalence worldwide as well as the third-highest cause of disability across both sexes under the age of 50 [2]. According to the American Association of Neurology (AAN) last 2012, around 38% of migraineurs need preventive therapy, but only 3%-13% avail of preventive medication [17]. Prevention of migraine can lead to significant improvement in health-related quality of life.
Targeted therapy via monoclonal antibodies have also been used for migraine treatment.
Last 2018, Erenumab, Fremanezumab, and Galcanezumab were approved for prophylaxis for episodic migraine. However, for most of these drugs, intolerable side effects and cost are the common detriment for patient compliance.
Memantine may be an ideal drug for migraine prophylaxis due to its pharmacokinetics.
Additionally, Memantine has been widely used with a good safety profile in the treatment of Alzheimer's disease, wherein most of the elderly taking this drug have a variety of comorbid diseases -proof of its tolerability with most patient populations [18]. Most experience with this drug for migraine were clinical data from observational studies, which all showed positive results suggesting the effectivity of Memantine in migraine prophylaxis. Based on latest guidelines, Memantine has only Level C (weak) evidence for the prophylaxis of migraine since until recently, there was only one published RCT that had favorable results for Memantine [8] [7]. With recent available RCT evidence that supports Memantine use, a stronger level of recommendation may be given to Memantine [16].
Study design inclusion criterion for this review was deliberately expanded to include observational studies to generate a more comprehensive database due to the paucity of currently published evidence. Episodic rather than chronic migraine was prioritized since most drugs for migraine prophylaxis would initially be evaluated for acute attacks prior to proceeding to a less common population of chronic migraineurs. Frequency of headaches expressed as migraine days in a month was the primary outcome measured. Since prevention of attacks is the primary concern, the duration of each attack would be less appropriate to analyze as this is more of a measure migraine attack termination. Based on these parameters, evidence gathered from this review is applicable to patients experiencing migraine attacks less than 15 migraine days in a month.
Risk of bias was relatively low for the two RCTs. All included observational studies adhered to a common reference for migraine diagnosis, provided a reproducible methodology, and presented the necessary data to evaluate the efficacy of Memantine for migraine prophylaxis -showing a good methodological quality.
There were significant observations identified regarding characteristics of the studies included in this review. First, clinical heterogeneity in terms of the populations enrolled in the RCTs may be remarkable. No sub-group distinction between migraine with and without aura was made. Although both migraine types are usually treated similarly, it has been suggested that response to treatment may be different for migraine sub-types due their distinct pathophysiology [19]. Stricter inclusion criteria that will ensure homogeneity will thus benefit future studies to determine if sub-analysis of pure migraineurs with and without aura will have an effect. Lastly, the more practical approach is to determine if the positive outcomes that were measured in the included studies actually clinically significant. Disability measures like MIDAS may help define more meaningful results that could translate to clinically applicable data [9]. Not all included studies reported quality of life outcomes, thus it is highly recommended to expand analysis to include these measures to determine if statistically significant results on reduction of migraine days translate to clinically significant data. Furthermore, clinical trials involving head-on comparison of Memantine versus other active prophylactic drugs in patients with episodic migraine may be beneficial to determine the relative place in therapy of Memantine in this condition.
Currently, this is the only comprehensive consolidated study to systematically review and pool available data to provide good quality evidence regarding the efficacy and safety of Memantine for episodic migraine prophylaxis. This review provides a moderate level of evidence for the efficacy of memantine for the prevention of episodic migraine based on the pooled data from the 2 relevant studies with low risk of bias but with relatively low sample size. The point estimates with its confidence intervals of the outcomes measured did not cross statistical clinical threshold to dissuade from the use of memantine.
Thus, additional well-designed randomized placebo-controlled trials with: 1) a larger and more homogenous population, 2) include dose-effect analysis, 3) head-on comparison with other active drugs, and 4) measure more outcomes focused on disability and quality of life are recommended to validate the efficacy and safety of Memantine for patients with episodic migraine.

Conclusion
Our study showed that memantine has a beneficial effect on migraine prophylaxis compared to placebo as well as having a good tolerability profile based on pooled evidence from small clinical trials and single-arm observational studies. Memantine resulted to a decrease in migraine days per month and showed insignificant occurrence of adverse events. Further trials should have higher sample size and dose-response effect of memantine may be determined to substantiate the conclusions of this review. Effects of memantine on clinically significant outcome measures for patients with episodic migraine should be explored. Head-on comparison of memantine with other active drugs may be necessary to determine the place in therapy of memantine among the available prophylactic drugs for migraine.

Ethics approval and consent to participate
Ethics approval and consent to participate is not applicable to our study.

Consent for publication
There were no personal data of any individual included in our study, hence consent for publication is not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Authors' contribution
MWLM and AIE did the literature search, applied the inclusion and inclusion criteria, and gathered and analyzed pertinent data for the generation of the meta-analysis. MLB and AAR were major contributors in writing the manuscript. All authors read and approved the final manuscript. Flow diagram of information of the review as adapted from PRISMA guidelines [11].

Figure 1
Flow diagram of information of the review as adapted from PRISMA guidelines [11].        Forest plot of adverse drug events seen with administration of Memantine compared to placebo.