Mira 2003.

Study characteristics
Methods	Multicentre, parallel‐group, double‐blind RCT with 3 months total duration of treatment and follow‐up Note: this trial included participants with both BPPV and Ménière's disease. The trial was stratified by condition. Only data that were reported separately for Ménière's disease have been extracted for the review.
Participants	Setting:  Recruited from 11 hospital and university ENT centres across Italy. Both inpatients and outpatients were recruited. Trial conducted from January 1999 to June 2001.  Sample size: Number randomised: 81 participants with Ménière's disease (n = 144 total) Number completed: 81 participants Some participants are listed as not attending interim follow‐up, but the authors state that an ITT analysis was conducted, and do not state that any participants were missing from final follow‐up. We therefore assume that complete follow‐up was achieved.  Participant baseline characteristics Reported for participants with both Ménière's disease and BPPV Number in betahistine group = 75, placebo group = 69 Age:  Betahistine group: mean 46.9 years (SD 13.1) Placebo group: mean 48.8 years (SD 14.3) Gender:  Betahistine group: 33 males (44%): 42 females (56%) Placebo group: 27 males (39%): 42 females (61%) Probable/definite Ménière's disease:  Not reported Duration of disease: Betahistine group: mean 31.6 months (SD 55.0)  Placebo group: mean 32.5 months (SD 67.3) Attack frequency at baseline:  Not reported Hearing loss at baseline:  Not reported Measure of tinnitus at baseline:  Not reported Number of participants with bilateral disease: Not reported Inclusion criteria: Males and females aged 18 to 65 years old. Signed, informed, written consent. Interfering concomitant therapies withdrawn at least 7 days before the start of the study treatment. Normal laboratory‐documented renal and hepatic function. Adhering to the scheduled procedure. Exclusion criteria: Concomitant infectious or cerebrovascular diseases. Diseases that were not compatible with or contraindicated by betahistine. Concomitant therapy with anti‐vertigo drugs. Use of drugs that act on the cerebral circulation, antihistamines, calcium antagonists, anti‐aggregants, thiazide diuretics, corticosteroids and benzodiazepines. Having any major medical or surgical condition or having a terminal disease. Diagnosis of Ménière's disease: Stated to be according to the AAO‐HNS 1995 criteria
Interventions	Intervention (n = 41 randomised, n = 41 completed) Betahistine; 16 mg (2 tablets) twice a day, administered at 8 a.m. and 8 p.m. after meals for 3 months Comparator (n = 40 randomised, n = 40 completed) Placebo; 2 tablets twice a day, administered at 8 a.m. and 8 p.m. after meals for 3 months. The tablets were indistinguishable by colour, weight and flavour from betahistine tablets, but their composition was not reported. Background interventions administered to all participants None reported
Outcomes	Primary outcomes relevant to this review: Improvement in vertigo Article reports on the number of people who report an improvement in "intensity score" of vertigo at 1, 2 and 3 months. However, this scale does not apparently assess frequency of vertigo, and clearly does not capture duration of episodes therefore cannot be regarded as a global score. Therefore, no data are available for analysis.  Change in vertigo The frequency of attacks at 3 months was reported fully for the betahistine group, and partially for the placebo group (with no SD) Serious adverse events Adverse events are fully reported, but it is not possible to clearly distinguish which of these were "serious" Secondary outcomes relevant to this review: Disease‐specific health‐related quality of life The percentage change in DHI scores from baseline was reported, but without any estimate of variance, therefore these data cannot be included in the review  Hearing Not reported  Tinnitus  Not reported; tinnitus was only assessed as part of a composite score including aural fullness, nausea and vomiting  Other adverse effects Assessed and reported  Other outcomes reported in the study: Number of participants with improvement in associated symptoms Overall quality of life Overall "GISFaV" score ‐ including intensity, duration and associated symptoms of vertigo  Duration of attacks Dizziness Assessment Rating Scale Investigator assessment of overall treatment efficacy and acceptance "Doctor's preparedness to treat the patient again with the same treatment" Participant assessment of overall treatment efficacy and acceptance
Funding sources	Quote: "This study was supported by a grant from Grunenthal‐Formenti, Milan, Italy" (Pharmaceutical company funding)
Declarations of interest	No declaration is provided
Notes	Research integrity checklist: No retractions or expressions of concern were identified  The study was published prior to 2010 and no trial registration was required Baseline characteristics for participants with Ménière's disease cannot be established. However, no concerns according to the full cohort of participants with Ménière's disease and BPPV.  Plausible loss to follow‐up during the trial is reported, although all participants appear to have been assessed at the final 3‐month assessment point  No implausible results were noted No concerns over randomisation
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "The trial was [...] randomised for a series of four patients; it was balanced among the centres" "[...] 11 hospital and university ENT centres were involved in the study. Each of these was given a series of four patients to recruit. There was a total of at least 140 patients, who were stratified into groups having recurrent vertigo related to Meniere’s disease (MD) and positional paroxysmal vertigo (PPV) of probable vascular origin." Comment: indicative of block randomisation, stratified by condition (MD or PPV). However there is insufficient information about sequence generation.
Allocation concealment (selection bias)	Unclear risk	Quote: "The treatment was assigned to the patient according to the random list of the relevant diagnosis (MD or PPV)." Comment: there was insufficient detail about allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Quote: "The tablets were indistinguishable by colour, weight and flavour. They were supplied in identical packages with a fantasy name to keep the blindness"  Comment: the study is described as "double‐blind" suggesting the patients and either the personnel managing the trial, or those administering treatments, or those assessing outcomes were unaware of group assignment. It is clear that patients were blind to whether they received active drug or placebo (as above). No further details were reported regarding study personnel, therefore it is unclear whether blinding was ensured.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Comment: trial does state that it is "double‐blind" but no further details are reported. However, most outcomes of interest are reported by (blinded) participants, therefore considered low‐risk.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Quote: "Fifteen patients (ten in the BE group and five in the PL group) were excluded from the PP analysis because of protocol violations such as interfering concomitant therapies (four in the BE group and three in the PL), not coming to the 2‐month follow‐up visit (three in the BE group and two 2 in the PL) and adverse events (three in the BE group)."  "In the intention‐to‐treat analysis, all the randomised patients were considered who had taken at least one dose of the drug being studied, who had carried out at least one complete examination involving all the measurements specified in the protocol and who had not violated the protocol".  Comment: the number of patients lost to follow‐up is not clearly reported. No separate details are given for those with Ménière's disease.  It is not clear how the ITT analysis and PP analysis differ if participants with protocol violations were excluded from the ITT analysis.
Selective reporting (reporting bias)	High risk	Comment: primary outcome data (number of vertigo attacks per month) is reported incompletely for the placebo arm of this trial. Data are reported as a percentage change from baseline in attack frequency, which seems an unusual choice. No protocol is available to compare pre‐specified outcomes therefore we consider this high‐risk.
Other bias	Low risk	Comment: no other concerns identified.