Gabapentin for the Treatment of Hot Flushes in Menopause

A Meta-analysis

Sang-Hee Yoon, MD, PhD; Ji Young Lee, MD, PhD; Chulmin Lee, MD, PhD; Hyojin Lee, MD; Soo-Nyung Kim, MD, PhD


Menopause. 2020;27(4):485-493. 

In This Article


Search Strategy

Three of the authors of the present study (S.H.Y., C.M.L., and J.Y.L.) designed the protocol and extraction forms in accordance with the Preferred Reporting Items for Systematic Review and Meta-analysis guidelines.[14,15] Reviews and original articles were searched for in MEDLINE, PubMed, and EMBASE databases and in the Cochrane Central Register for Controlled Trials (CENTRAL) in the Cochrane Library up to June 2018. The following search terms were used: "menopause," "hot flushes," "vasomotor symptoms," "gabapentin," and "non-hormonal therapy." All relevant reports were retrieved and reference lists were reviewed manually to identify further studies. A manual search of PubMed was also performed for related articles. No attempt was made to identify unpublished studies unless they had been released as online publications ahead of print. No reports from scientific meetings were included. The above-mentioned searches were performed by an accredited clinical librarian.

Selection criteria

Inclusion criteria for studies were as follows: randomized controlled trials (RCTs); participants of interest were postmenopausal women meeting the following criteria—amenorrhea for more than 12 months or amenorrhea for 6 to 12 months with a serum follicle-stimulating hormone level >40 mIU/mL and estrogen level <30 pg/mL, or postbilateral oophorectomy for 2 months; no estrogen, progestin, leuprolide, or tamoxifen therapy within the past 2 months; interventions of interest were gabapentin and identical-appearance placebo; primary outcomes were percentage reduction or mean difference in frequency, duration, or severity of hot flushes; and secondary outcomes were adverse events and dropout rate. Uncontrolled and open-label studies were reviewed but not included in the meta-analysis. Reviews, abstracts, editorials, letters to the editor, preliminary reports, and studies published in languages other than English were excluded. Study selection was performed independently by three reviewers (S.H.Y., H.J.L., and J.Y.L.). Any disagreement was resolved unanimously by consultation and discussion with a fourth author (S.N.K.).

Data Extraction

Three authors (S.H.Y., H.J.L., and J.Y.L.) scored the studies and collected information independently. The following data were recorded for each eligible study: name of the first author, study location, year of publication, study design, sample size, duration of treatment, dosage of gabapentin, hot-flush frequency (number of recorded episodes per day), hot-flush duration (average duration in minutes of all hot flushes experienced per day), hot-flush severity (expressed as a composite score calculated for each study), adverse events (dizziness or somnolence symptoms), and dropout rate. Continuous variables including duration, frequency, and composite score of hot flushes are described in percentage changes or mean difference from baseline to 1, 2, 3, or 6 months for each measure. Primary outcomes are reported as the standardized mean difference (SMD). Secondary outcomes are presented with odds ratios (ORs). Discrepancies in data extraction were jointly reviewed until a consensus was reached.

Quality Assessment

The Jadad score was used to assess the quality of RCTs.[16] Studies with a Jadad score of 2 or higher were considered to have high quality. Studies were included regardless of their quality; however, sensitivity analysis was performed based on quality. We assessed the risk of bias of included studies as low, unclear, or high using the Cochrane Collaboration tool,[17] and considered the method of sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting. Studies considered to have a high risk of bias were those graded as having a high risk for the method of sequence generation, allocation concealment, blinding procedures, or incomplete outcome data. Any disagreement was resolved after discussion and reevaluation with the fourth author (S.N.K.).

Data Synthesis and Analysis

Heterogeneity across studies was examined using I2 , which measures the percentage of total variation across studies,[18] and substantial heterogeneity defined as an I2 value >50%.[19] A random or fixed-effects model was used to estimate the SMD and combined OR for randomized studies depending on whether significant heterogeneity was present or not, respectively. Subgroup analysis was then conducted for the dosage of gabapentin (300, 900, or 1,800 mg). Subgroup analysis was planned a priori before data collection and analysis. Meta-regression analysis was carried out to evaluate the relationship between dosage of gabapentin and the mean difference in these variables.

Sensitivity analysis was performed to evaluate the influence of single studies on the overall estimate. Publication bias was evaluated using the Begg and Mazumdar rank correlation test,[20] Egger's test,[21] and fail-safe N test.[22] A funnel plot was constructed to assess this bias using the standard error and diagnostic OR.[23,24] Comprehensive Meta-Analysis version 3.3 (Biostat, Englewood, NJ) was used for all statistical tests. We considered P < 0.05 to indicate statistical significance. Data are presented according to a checklist created based on PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses).[25]