The Diet and BPPV Cohort Study (DaBC) is an ongoing multicenter prospective cohort study conducted at three specialized neuro-otology centers in Northwest China (Figure 1). Recruitment posters detailing study objectives were displayed in hospital outpatient clinics and community health centers. These posters featured a Quick Response (QR) code linking to a platform for collecting basic demographic information and self-reported dizziness symptoms. Initiated in July 2023, all potential patients involved in the recruitment process were invited to the department of the Otolaryngology-Head (ENT) of the involved centers for a comprehensive clinical evaluation, including an extensive medical history from an otolaryngology head and neck surgeon as well as detailed otologic and neurologic evaluations.

Participants are eligible if they are aged 18 years or older and diagnosed with BPPV by an otorhinolaryngologist following the Bàràny Society criteria (2). The diagnosis of BPPV is based on: (1) Presence of episodic positional vertigo of brief duration; (2) Presence of characteristic positioning nystagmus indicative of BPPV, elicited through Dix–Hallpike (posterior canal BPPV) or Supine head roll testing (horizontal canal BPPV) (11). Individuals are excluded if they have physician-diagnosed Meniere's disease, vestibular neuritis, vestibular migraine, or other vestibular diseases; have dizziness attributable to systemic diseases, such as cardiovascular and cerebrovascular disease, endocrine disorders; have a history of head trauma, inner ear surgery, exposure to ototoxic agents or related withdrawal syndromes; have communication impairments. Participants will also be excluded who are unavailable to attend the scheduled follow-up investigation. Initiated in May 2022, the DaBC successfully recruited 2,505 eligible BPPV patients by the closing date of October 30, 2024. Of these, 844 participants with complete baseline data were invited for subsequent follow-up, as outlined in the study protocol (Figure 2).

The study was approved by the Research Ethics Committees of Xi'an Jiaotong University (XJTU2021-1560). Detailed aims and protocol of DaBC will be delineated to all participants, and written informed consent will be obtained at enrollment. The cohort study adheres to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines (28) and the Declaration of Helsinki.

To ensure that the data collected reflects the natural course and initial presentation of BPPV, the participants in this study are restricted to first-diagnosed patients with BPPV. This study plans to enroll 1,200 people with complete baseline information through May 2025. At enrollment, participants are instructed to complete a detailed screening questionnaire, including sociodemographic data, medical history, lifestyle behaviors, anthropometrics, menopausal status and postmenopausal hormone use (for women). Following the screening, thorough history-taking and otolaryngologic, audiological, and neurotological evaluations, especially the Dix–Hallpike maneuver and the Supine head roll test, will be performed to further confirm BPPV diagnosis and determine eligibility of participants (29). Concurrently, participants are also requested to provide dietary information for the past year prior to diagnosis via a validated semi-quantitative Food Frequency Questionnaire (FFQ). Residual dizziness severity of BPPV patients is assessed using the validated 25-item Dizziness Handicap Inventory (DHI). Besides, participants are also asked to complete standardized scales including: the Vestibular Activities of Daily Living Scale (VADL) quantifying functional balance limitations; the Activities-specific Balance Confidence (ABC) Scale measuring self-efficacy in ambulation; the Hospital Anxiety and Depression Scale (HADS) screening psychological comorbidities; and the Pittsburgh Sleep Quality Index (PSQI) evaluating sleep disturbances—collectively measuring multidimensional BPPV-related symptom burdens and quality of life impacts. At the same time, fasting blood samples are collected for genotyping and quantification of circulating nutrients and other potentially bioactive substances; fecal specimens are also collected for gut microbiota profiling.

The follow-up of the DaBC study is performed based on the comprehensive consideration of the follow-up interval of BPPV progression in previous studies. Data were collected at baseline (time of diagnosis), 1 month, and annually through 5 years after the baseline examination (Figure 1). All patients in the cohort study will participate in routine follow-up visits after their initial visit. First, researchers will record the phone numbers and WeChat accounts of the participants for active surveillance of disease progression, dietary adherence, and other health-related information during the 5-year follow-up period. Second, patients will contact their physicians via WeChat upon symptom recurrence to initiate prompt clinical evaluations. Diagnostic tests and symptom assessments will be documented during each unscheduled visit.

Follow-up evaluations occur at 1 month, 1–5 years: 1-month post-baseline through telephone interviews focusing on residual dizziness via DHI updates and dietary adjustments captured by FFQ; Throughout the 1–4 years of follow-up study period, conduct annual post baseline interviews through WeChat or telephone to track changes in lifestyle, recurrent symptoms, FFQ diet, and assess persistent symptoms and related quality of life related to BPPV; and 5 years post-baseline through in-person revisits mirroring baseline procedures, as well as the blood and fecal sample collection. Throughout the 5-year study period, clinical endpoints, including the recurrence and intervals between episodes, are systematically extracted from hospital electronic medical records. Annual reminders via WeChat or phone ensure ongoing participant engagement. The complete study timeline, including enrollment phases, assessment schedules, and data collection protocols, is visually summarized in Figure 3 and operationally detailed in Table 1.

Dietary patterns promoted in the Dietary Guidelines that are intimately related to disease or crucial biological pathways will be summarized to evaluate the overall quality of the diet. Dietary information is obtained through the FFQ scale at baseline and each follow-up interval. A semi-quantitative FFQ (30) is used to collect detailed dietary intake on the usual consumption frequency of intake of standard portion sizes of each food item during the preceding year. The FFQ is a structured questionnaire on 108 specified food items assembled into 10 categories: cereals; vegetables; fruits; legumes and soybean products; mushrooms and seaweeds; nuts; red meat and processed meat; milk and dairy products; beverages; sweets and baked goods. Participants are required to report the frequency of each food item consumed, selecting from nine predefined frequency categories as follows: “never or less than one serving per month”; “1–3 servings per month”; “1 serving per week”; “2–4 servings per week”; “5–6 servings per week”; “1 serving per day”; “2–3 servings per day”; “4–6 servings per day”; and “6 or more servings per day”. For seasonal foods, participants also indicate the number of months these foods were consumed. The midpoints of the frequency options are utilized to convert consumption frequencies into daily intake frequency, from which the average daily intake of each food in grams (g/day) could be calculated. Furthermore, consumptions of cooking oil and condiments are also investigated. Total daily intakes of energy and nutrients are achieved by multiplying the intake of each food by the portions of nutrients per 100 grams and summing all foods consumed. The structure and content of the FFQ questionnaire are presented in Figure 4.

Dietary patterns promoted in the Dietary Guidelines that are intimately related to disease or crucial biological pathways will be summarized to evaluate the overall quality of the diet. Supplementary Material 1 detailing food mapping decisions, substitutions” and cutoffs for score computation in this cohort is provided.

The primary outcome is the number of BPPV relapses during the 5-year follow-up period. A relapse is defined as the recurrence of positional vertigo accompanied by characteristic nystagmus, confirmed through diagnostic maneuvers including the Dix–Hallpike test and supine head roll test (29). During each ENT clinic visit, specialized physicians conduct neurological evaluations and assess potential relapses by synthesizing clinical data: the temporal pattern and frequency of dizziness episodes, nystagmus characteristics observed during diagnostic testing, and resolution status of prior episodes. The secondary outcome is the average recurrence interval during the 5-year follow-up period, which will be calculated as the mean interval between consecutive confirmed relapse episodes based on electronically verified timestamps from hospital records and patient-reported symptom diaries (41, 42).

Moreover, exploratory outcomes are defined as the long-term patient-reported outcomes associated with BPPV, including dizziness handicap, vestibular dysfunction, fall risk, anxiety and depression, and sleep quality. Genome-wide genotyping and analysis of the gut microbiome will also be performed to investigate their potential influence on the association between foods or food groups, nutrients, dietary patterns, and the observed changes in outcomes.

Information on demographics (age, sex, educational level, occupation, marital status, residential status, and household income), medical history (cardiovascular disease, diabetes mellitus, liver and kidney function abnormalities, endocrine disorders, anxiety, depression, ocular and auditory conditions, history of malignancies, and traumatic events), family history of chronic disease (hypertension, coronary heart disease, diabetes mellitus, and cancer), anthropometric measures (height, weight), lifestyle behaviors (smoking, alcohol consumption, use of nutritional supplements, exposure to chemical toxins, and daily screen time) and menstruation will be captured by comprehensive questionnaires. The Body mass index (BMI; kg/m²) will be calculated as weight in kilograms divided by the square of height in meters. Additionally, physical activity will be assessed using the International Physical Activity Questionnaire (IPAQ) (56) and expressed as metabolic equivalent hours per week (MET hours per week). To mitigate confounding by baseline disease severity, the maximal duration of vertigo episodes was incorporated as a severity marker. This parameter was defined as the longest continuous period of positional vertigo in minutes during the 24 h preceding diagnosis. The selection of vertigo duration as a severity marker aligns with consensus recommendations for BPPV phenotyping (29, 31, 57), where prolonged episodes have demonstrated prognostic relevance for treatment resistance and recurrence risk in prior cohort studies (57).

A rigorous and standardized research protocol has been scrupulously established, featuring strict standard operating procedures and robust quality control mechanisms. ENT clinicians, as well as neurologists, attend the standardization meetings from the outset and are involved in the diagnosis and the assessment of the prognosis of the BPPV throughout the study. During each visit, study personnel reiterate the study details and follow-up procedures to all participants.

The research team has undergone extensive training to ensure strict adherence to protocol and to enhance the efficiency of the research process. Quality control checks are performed weekly by at least two researchers to identify and rectify any errors, outliers, or missing data. In instances of incomplete or improperly filled questionnaires, the researchers promptly contact participants within a week via telephone to complete and correct the submissions.

Baseline, follow-up, and outcome information are securely stored, accessible only to authorized personnel. Moreover, clinical laboratory measurements and bio-banked samples are managed through a dedicated bio-bank system, which meticulously records details of sample usage and measurements.

The distributions of continuous variables for normality are assessed by the Kolmogorov-Smirnov normality test. Descriptive analyses present baseline characteristics of participants by sex as means (standard deviations) for normally distributed variables, medians (interquartile ranges) for skewed variables, and frequency (percentage) for categorical variables. Comparisons between groups at baseline are performed using Student's t-test or Wilcoxon rank-sum tests for the continuous variables and the chi-square test for the categorical variables. Sample groupings were trichotomized, with AHEI scores of 42 and 46.4 being used as cutoffs (58, 59), as the AHEI provides a comprehensive measure of overall diet quality, enabling a broader assessment of chronic disease risk (31). Similar analyses will be performed using other dietary indices in subsequent analyses.

Further analysis will categorize participants into tertiles based on dietary pattern scores. Demographic and clinical characteristics will be compared across these categories for class variables, while continuous variables will be analyzed using the Wilcoxon rank-sum test or analysis of variance with subsequent 2 × 2 post-hoc comparisons. Multivariate ordinal logistic regression will be used to examine the association between exposure variables and outcome variables (the number of recurrences). Odds ratios (ORs) will be estimated for dietary pattern scores considered as a categorical variable. Linear trends will be assessed using the Wald test by incorporating median values of each tertile of dietary pattern scores as continuous variables in the regression model (60). To explore the dose-response relationships between dietary pattern scores with the average recurrence interval, restricted cubic spline regression with four knots will be used, with scores and outcome analyzed as continuous variables and truncated at the 0.5th and 99.5th centiles. Non-linearity will be evaluated by a likelihood ratio test comparing models with only linear terms against models including both linear and cubic spline terms. Changes in the scoring of long-term patient-reported scale will be calculated as the difference between follow-up and baseline values, and analyzed using linear mixed effects models to assess variations over time across dietary pattern score groups (61). Longitudinal associations between changes in dietary pattern scores and BPPV outcomes are analyzed using generalized estimating equations (GEE) with repeated measures. Dietary pattern scores are categorized into tertiles at each follow-up interval (baseline, 1-year, 5-year), and time-updated tertiles are incorporated as time-varying covariates in the models (62).

Subgroup analyses will stratify participants by baseline covariates, including sex, age, smoking status, BMI, affected semicircular canals, and maximal vertigo episode duration during the 24 h pre-diagnosis (categorized into tertiles: < 2 min, 2–5 min, >5 min) (63).

Genotyping data will be used to estimate the proportion of variance displayed by all single-nucleotide polymorphisms (SNPs) after imputation by restricted maximum-likelihood (64). For microbiome analysis, microbial diversity will be assessed using alpha diversity indices (Shannon and Chao1 index) and beta diversity, which will be visualized through principal component and coordinate analysis. Linear mixed effects models will explore the impacts of genetic variations and gut microbiota on the change in level of related life quality, considering potential interactions (65). Additionally, key neurotransmitters and bioactive substances will be analyzed to explore their mediating effects on the relationships between dietary pattern scores and BPPV prognosis. Multi-omics techniques (including glycomics, lipidomics, proteomics, and microbiomics) and machine learning algorithms will be further integrated into the sample and data analysis (66).

All statistical analyses will be performed with SPSS version 22.0 (SPSS Inc., Chicago, Illinois) or R software, version 4.3.0. A two-sided p-value of less than 0.05 will be considered statistically significant.