Prevalence of vestibular disease in France: analysis of prescription data from a national health insurance database

Vestibular problems are frequent reasons for primary care consultations. However, there is considerable uncertainty about the prevalence and cost of vestibular disorders. Despite ambiguous effectiveness data, the histamine analogue betahistine is widely and almost exclusively used for treatment of vertigo. Prescription of betahistine can, therefore, be used as a proxy estimate for prevalence. We used openly available claims data from the French health insurance data warehouse, defining annual prevalence of vestibular disease as the number of people who received at least one betahistine prescription that year. Dosage and pack size of each prescribed formulation were extracted to calculate the sum of betahistine in mg and the Defined Daily Dose (DDD) for age and sex strata and in total. To estimate the relative impact of one landmark trial, the BEMED study, we compared prescriptions from the years 2014/2015 to prescriptions in 2019/2022. A total of 735,121 (2014), 694,705 (2015), 614,431 (2019), and 562,476 (2022) persons filled in a prescription of betahistine. Patients were predominantly older and female. Average amount dispensed per year and per person increased from 4422.54 mg during the pre-BEMED period to 4736.90 mg during the post-BEMED period. DDD decreased from 130 Mio per year in 2014/2015 to 116 Mio per year in 2019/2022. Total costs for betahistine decreased by 42% from 21,615,037 Euro in 2014 to 12,894,249 Euro in 2022. Vestibular disease is frequent in France and has a relevant impact on population health. Despite conflicting clinical evidence, betahistine continues to be prescribed widely in medical practice.


Introduction
In most industrialized countries, vertigo, dizziness, and balance problems are among the most frequent reasons for primary care consultations.Among the recurrent vestibular disorders that cause vertigo, Benign paroxysmal positional vertigo, Menière's Disease (MD), and Vestibular Migraine account for about 50% of all cases presenting with vertigo.One-year prevalence of vestibular disease has been estimated at 5% by a representative survey in Germany [1].Vestibular disease may be severely disabling and account for a considerable burden of disease and disability [2][3][4][5].A systematic review of international studies about primary care utilization found that between 1 and 8% of all listed patients had contacted a physician at least once because of vertigo or dizziness [6].For the United States of America, an average annual total of 48.1 billion USD direct costs for the years 2007 to 2015 were estimated [7], while another analysis of 2018 medical claims data reported an incremental increase of direct costs of 60 billion USD from patients diagnosed with episodic recurrent vestibular vertigo [8].Total annual direct and indirect costs of MD alone for the UK were between 541.30 million and 608.70 million pound per year [9].
Yet, there is considerable uncertainty around the data reported on prevalence, incidence, and costs of vestibular disease.Large population-based representative cohort studies hardly ever include a detailed workup of vestibular and balance functions.Self-reported occurrence of vertigo or dizziness is not specific for vestibular disease, and patient descriptions may be unclear, inconsistent, and unreliable [10].The second obvious place for representative data, medical claims data, has limited diagnostic validity, because vestibular disease is not well represented in the ICD-10; therefore, vestibular diagnoses in claims data tend to reflect these inaccuracies of the ICD system [11].
As a potential solution, information on medication from insurance claims data has successfully been used to estimate the prevalence of chronic disease [12].Thus, it may also be possible to estimate the prevalence of vestibular disease from prescription data.Betahistine is a histamine analogue that is widely [13] and almost exclusively used for treatment of vertigo in general, and specifically as a maintenance medication for the prevention of MD attacks [14].Using evidence from randomized-controlled trials, effectiveness of betahistine at currently recommended dosage regimes in Menière's disease seems to be weak [15] including, e.g., one clinical trial that could show no difference to placebo, the Medical Treatment of Menière's Disease with Betahistine (BEMED) trial [16].Betahistine at doses higher than the recommended dose was effective in relieving the symptoms in unilateral vestibulopathy [17].A meta-analysis of 17 studies found a reduction of symptoms in patients with vertigo of different origins; however, risk of bias and heterogeneity of the included studies were high [18].
Betahistine continues to be prescribed, as shown in a recent study from the United Kingdom that compared prescriptions before and after 2016, the year of the BEMED publication [19].Prescriptions of betahistine can therefore be a good indicator for the prevalence of vestibular disease, because betahistine is not regularly prescribed for other diagnoses.As France has an almost complete coverage of health insurance, we are using data from the French health insurance reporting system to estimate the annual prevalence of vestibular disease by the number of betahistine prescriptions.We were specifically interested in the amount of betahistine substance being prescribed per year, and in the sociodemographic characteristics of the patients who fill in betahistine prescriptions.Additionally, we wanted to investigate if there was any change in prescription practice following the BEMED trial in 2016.

Data sources
In France, health insurance is universally provided as part of the social security system.The statutory health insurance covers most of the costs of medical treatment including medication.Data on medication dispensed in France and fully or partially reimbursed are collected by the French health insurance data warehouse (SNDS) and openly available in aggregated form for download.Data from the SNDS cover almost all beneficiaries of the French health insurance, i.e., 99% of the French population [20].Betahistine is available only on prescription from retail pharmacies in France.

Variables
Data from the public aggregated version of the SNDS include information on the hierarchical anatomical therapeutic chemical (ATC) classification codes up to the chemical substance (5th level classification), the unique identifier for all drug presentations in France, the Code Identification Spécialité Pharmaceutique (CIP13), the CIP13 label containing dosage and pack size, the number of packs, the drug costs in Euro, and the amount of costs reimbursed.Costs are prices charged by pharmacies and do not include any costs for acquisition or administration.Sociodemographic information on patients is given on an aggregated level to preserve data protection, namely age in three brackets (0-19 years, 20-59 years, 60+ years), and gender.To avoid de-anonymisation, sociodemographic information is only given for strata with at least ten persons.This includes a total of 38 dose forms that were dispensed very infrequently, which might result in small discrepancies of our results to any summary measures reported elsewhere.

Statistical analyses
The ATC code N07CA01 was used to extract the prescriptions of betahistine in all trademark and generic formulations.We defined annual prevalence of vestibular disease as the number of people who received at least one betahistine prescription that year.From the CIP13 label, dosage of betahistine and pack size of each formulation were extracted to calculate the sum of prescribed betahistine in mg for each stratum and in total.To give an example, the dose form and package unit "BETAHISTINE ACCORD 08 MG CPR 30" would yield the information that each tablet contains 8 mg of betahistine and each package contains 30 tablets which adds a total of 240 mg betahistine per prescription.Total amount of betahistine in mg per person was then calculated by multiplying the package content in mg by the number of packages of the respective dose form and by dividing the sum of prescribed betahistine per stratum by the number of consumers of each stratum and in total.Defined Daily Dose (DDD) was calculated based on the standard dose of active substance of 24 mg.This dose is not equal to the actual prescribed or effective dose but serves as a standardization method for comparisons.
To compare prescriptions and dosage before and after the BEMED trial, we defined betahistine prescriptions from 2014 and 2015 as pre-BEMED and from 2019 and 2022 as post-BEMED.We hypothesized that prescriptions during the pandemic years 2020 and 2021 would differ from the previous years due to reasons unrelated to our objectives.
SAS (SAS Institute, version 9.4, NC, USA) and Excel Power Pivot (Microsoft Corp.) were used for all analyses.

Results
A total of 735,121 (2014), 694,705 (2015), 614,431 (2019), and 562,476 (2022) persons received and filled in a prescription of betahistine.Table 1 shows the total amount of mg betahistine and number of packages dispensed, the Defined Daily Dose (DDD), and the total amount of costs and the costs reimbursed, stratified by sociodemographic characteristics and year.
There was a notable decrease of the number of consumers and of most of the other summary measures in 2019 and 2022 as compared to 2014 and 2015, except for mg betahistine per person.Average amount of betahistine dispensed per year and per person was 4263.71

Discussion
Our study on the magnitude of filled in prescriptions found that between 2014 and 2022, 560,000 to 740,000 persons insured by the French statutory health insurance received betahistine.Patients with betahistine were predominantly older and female.While package count, total mg prescribed and DDD decreased when comparing years before 2016 to years after, mg per person per year increased steadily in all age groups.
As 99% of the French population is currently inscribed in the statutory health insurance [20], the number of persons with filled in prescriptions of betahistine corresponds to roughly 0.9% of the population of 67Mio persons contained in the SNDS.This number aligns well with the estimated 1% of annual consultations for vestibular disease from other studies [6], albeit at the lower end.The total annual number of 142 Mio prescribed tablets found in our study for 2022 also approximates the monthly 11 Mio tablets prescribed in the UK for an insured population of 57 Mio [19].We are likely to underestimate the true prevalence of vestibular disease, because not all patients who seek consultation for vestibular symptoms will receive a prescription of betahistine, but the magnitudes we found are in the expected range of persons with moderate-to-severe vertigo.
While betahistine is mainly licenced for Menière's disease (MD), our estimates of prescriptions largely exceed the prevalence of MD of 40 to 200 per 100,000 in the general population estimated by other studies [21][22][23].Partly, this can be explained by the tendency to overdiagnose MD in medical practice [21,23].Nevertheless, our findings also confirm that betahistine is prescribed not only for MD but also for a wide range of other vestibular diagnoses.These prescription patterns have also been reported elsewhere [19].Data from Germany show that about 20% of patients with benign paroxysmal positional vertigo, 63% of MD patients, and 26% of vestibular migraine patients received betahistine in primary care before presenting at a tertiary care clinic [24].
According to our data, prescriptions of betahistine in France decreased considerably between 2014 and 2022.There are several potential explanations for this: first, a decrease of vestibular disease in France, second, an improved, more evidence-based prescription practice following the BEMED trial, or third, reasons related to the pharmaceutical market.To start with disease prevalence, it is unlikely that diagnosis or true prevalence of vestibular disease decreased within 8 years by almost 25%.While fluctuations in the prevalence of vestibular disorders in different regions have been reported, there is no clear indication of a consistent increase or decrease in prevalence over the last decade.Second, decline in prescriptions could be the sign of some improvement of prescribing practice.Interestingly, the use of betahistine in MD or more generally in vestibular disease has no strong base of evidence [15,18].Likewise, the French Otorhinolaryngology-Head and Neck Surgery Society does not recommend betahistine as a first line treatment of MD [25].It can be argued that prescriptions did decrease after 2016, probably also following the publication results

Table 1 Prescriptions of betahistine in France
To avoid de-anonymisation, sociodemographic information is only given for strata with at least ten persons a Mean mg betahistine prescribed per person per year; DDD: Defined Daily Dose (basis 24 mg betahistine) from the BEMED clinical trial for betahistine in MD in 2016 [16] and a Cochrane review on the utilization of betahistine in vertigo [18].At the same time, we found that the amount of betahistine dispensed per person increased.Both findings may be a consequence of the discussions around the low bioavailability of betahistine due to the first-pass elimination [26], which may be counteracted by increasing the oral dose.As patient safety precludes a substantial dose increase, and as parenteral application is hardly feasible, the addition of a pharmacological booster such as selegiline is hypothesized to improve the effectiveness of betahistine [27].This tendency for higher dosage regimes is also reflected in the decreasing market presence of the low dosage forms found in our data.A third reason for the decline in prescriptions is likely to be market-related, as betahistine formulations have been out of stock or difficult to obtain in France for some time.Prescriptions of betahistine in France are still on a high level.To give an example, the decrease of DDD in France from 131 Mio mg in 2014 to 114 Mio mg in 2022 found in our data is well above the 68 Mio DDD reported in Germany in a population of 74 Mio insured persons [28].Of note, the German DDD for betahistine in 2022 was higher as compared to 2019 (62 Mio DDD).
Our study has its strengths in the rigorous database of the SNDS that gives access to a complete, reproducible, and unbiased analysis of medications in the French health care system.Limitations include the lack of individual information on patients including detailed diagnoses and prescription trajectories.
In summary, the prescription of betahistine each year seems to be a good indicator for the prevalence of vestibular disease in France.Vestibular disease is frequent in France and has a relevant impact on population health.Despite conflicting clinical evidence, betahistine continues to be prescribed widely in medical practice.
mg in 2014, 4482.75 mg in 2015, 4619.62 mg in 2019, and 4864.90 mg in 2022, increasing from 4422.54 mg during the pre-BEMED period to 4736.90 mg during the post-BEMED period.DDD decreased from an average of 130 Mio per year in 2014/2015 to an average of 116 Mio per year in 2019/2022.Patients aged 60 and over received the highest amount of betahistine per year in all years, with 5206.46 mg in 2014 to 5750.17 mg in 2022.Mg per person increased in both men and women.Total costs for betahistine decreased by 42% from 21,615,037 Euro in 2014 to 12,894,249 Euro in 2022, of which less than 50% was reimbursed.The number of available betahistine dose forms in France decreased from 49 in 2014 to 38 in 2022.Most dose forms were either 8 mg or 24 mg.The percentage of dose forms with 8 mg decreased from 61% in 2014 to 47% in 2022.