LETTER TO THE EDITOR Is cinnarizine/betahistine combination superior to respective monotherapies in treatment of vertigo? Comment on a randomized triple-blind placebo-controlled trial Kiumarth Amini 1 & Maryam Taghizadeh-Ghehi 2 Received: 4 December 2019 /Accepted: 8 January 2020 # Springer-Verlag GmbH Germany, part of Springer Nature 2020 We read with great interest the report of a randomized, triple- blind, placebo-controlled trial (RCT) conducted by Asadi et al. [1], to compare efficacy of cinnarizine and betahistine combination with their monotherapies in patients with acute peripheral vertigo (APV). Despite abundant use of betahistine for vertigo, a Cochrane systematic review concluded that available supporting evidence had low quality [2, 3]. The study by Asadi et al. could provide valuable evidence if its robustness was assured. The authors concluded that cinnarizine and betahistine monotherapies were effective but the combination was superior; however, the conclusion would be questionable considering following points. First question emerged regarding the study inclusion and exclusion criteria. We faced ambiguity in the term “APV.” Whether it meant patients with acute vestibular vertigo or peripheral vertigo with unknown origin was unclear. It was not clear either, whether severity of the symptoms, history of vertigo, and recent or current treatment with anti-vertigo med- ications were taken into account in eligibility criteria. Using valid and reliable scales in assessment of subjective outcomes would be of great importance [3]. Vertigo intensity in terms of mean vertigo score (MVS) was considered as a primary outcome in this study. MVS has been used in previous RCTs while its reliability and validity was also examined [4, 5]. Validation of the Persian version of MVS in the present study was not reported. Visual analogue scale (VAS) has been used in responding to the MVS items in previous studies; however, Asadi et al. reported it as a separate primary outcome. Statistical analysis performed to compare outcome data might be the main challenging issue in this paper. Appropriate statistical methods were not used to compare out- come measures among three study groups, specifically, when there were three time points of data. Therefore, methods such as repeated measure, related post hocks or, corresponding non-parametric tests should have been applied. Reported sig- nificant p value obtained by Friedman test in comparing three groups, could only reveal difference among the groups, and could not show the superiority of one to others. Additionally, the change of the scores from baseline should have been con- sidered in comparisons. It was crucial when the baseline mea- sures were significantly different between groups (Table 4). Ignoring the effect of the baseline could have been resulted in wrong inferences. Other issues we faced in the statistical methods and results that arised concern regarding the study robustness include lack of standard deviation in the “Table 3,” reported p value with 4 decimals when SPSS could compute it up to 3 decimals, using Wilcoxon Mann-Whitney U test for data distribution assessment, and incorrect naming of a statis- tical test (e.g., covariance’ s ANOVA). Hence, making any conclusion regarding the superior treat- ment among compared groups would be incorrect and mis- leading with respect to the performed analysis. Moreover, it was not clear how the authors dealt with confounders. Among various demographic and clinical characteristics of the pa- tients, only age and sex distributions were compared among three groups. However, previous RCTs ascertained similarities of study groups in various aspects such as of patients’ weight, height, BMI, smoking status, concomitant medications, fre- quency and duration of previous episodes, and hearing loss. The final point is about assessment of safety. Adverse drug reactions (ADRs) should have been detected using a pre- defined protocol. Causality and severity of the detected * Maryam Taghizadeh-Ghehi taghizadehgm@sina.tums.ac.ir Kiumarth Amini drkiumarthamini1364@gmail.com 1 Department of Clinical Pharmacy, Faculty of Pharmacy, Tehran University of Medical Sciences, Tehran, Iran 2 Research Center for Rational Use of Drugs, Tehran University of Medical Sciences, 4th floor, No 92, Karimkhan Zand Ave., Hafte Tir Sq, Tehran, Iran European Journal of Clinical Pharmacology https://doi.org/10.1007/s00228-020-02833-z