The main objective of this study was to measure the association between previous cumulative exposure to anticholinergic and sedative drugs and cognition in a longitudinal real-life cohort.

MEMORA is a multicentre prospective cohort study conducted throughout the patient's care pathway in Memory clinics of Lyon, France. MEMORA aims to investigate factors associated with changes in functional autonomy, cognitive performance, and Behavioral and Psychological Symptoms in Dementia (BPSD) over time in individuals receiving routine care [16]. MEMORA includes every patient who underwent a consultation at a memory clinic for a cognitive complaint, from November 2014.

The data of participants in the present study were extracted from a 6-year sample of MEMORA patients (2014–2020). Patients whose clinical data and medication data from the local branch of the Primary Health Insurance Fund (PHIF) database were available were included. Patients with missing data regarding one of these two criteria were excluded. This study followed the STROBE checklist from the EQUATOR guidelines [17].

The level of exposure to anticholinergic and sedative drugs for each patient was measured using the Drug Burden Index (DBI) [18], which was developed for older people based on pharmacological principles. The DBI is a daily score and is calculated according to the following formula:

$$DBI=_+_= \sum_^_}\frac_}_+_ }+ \sum_^_}\frac_}_+_}$$

where \(_\) represents the daily dose of medication \(i\) (\(i\) = 1, …, \(_\)) with anticholinergic and/or sedative proprieties and \(_\) represents the minimal effective dose of this medication according to the World Health Organization (WHO) Defined Daily Dose [19]. In the DBI calculation originally developed by Hilmer et al. [18], \(_\) represents the recommended minimum daily dose approved by the U.S. Food and Drug Administration (FDA). To enable the comparison of DBI across countries, a previous study demonstrated the equivalence between the two DBI calculation formulae [20]. The list of medications with anticholinergic or sedative properties was obtained from the literature and adapted according to French practices [20,21,22].

The medications received by the included patients was collected through a PHIF extraction, where all prescribed and reimbursed drugs are registered when dispensed. For each patient included, medication data were available from 2 years before the first memory consultation until the last one. PHIF data included the name, dosage and quantity of drugs dispensed, combined by semester. To calculate the daily DBI, a mean daily dose for all anticholinergic and sedative drugs was derived from the 6-month drug consumption. The cumulative DBI was computed by adding the daily DBI over the entire available period prior to each cognitive assessment (see Fig. 1).

Timeline description of the study. DBI: Drug Burden Index; MMSE: Mini-Mental State Examination; PHIF: Primary Health Insurance Fund

Patients were then divided into 3 groups based on the DBI’s standard daily classification: no exposure to anticholinergic and sedative drugs (DBI = 0), moderate exposure (0 > DBI > 1), high exposure (DBI ≥ 1). The usual daily thresholds were multiplied by the number of medication follow-up days available before each Mini-Mental State Examination (MMSE) measurement.

Comprehensive cognitive assessment was performed by a physician, a nurse, or a psychologist at baseline and at each consultation. Cognitive impairment was assessed using the standardized Mini-Mental State Examination (MMSE) [23] and range from 0 (severe cognitive impairment) to 30 (no impairment). A minimum of two MMSE measures separated by at least 6 months had to be available to include a patient in the study. In the following analysis, the term “baseline” refers to the first MMSE measurement of the patient.

Baseline characteristics, such as age, sex, educational level, functional autonomy level, and BPSD, were collected. Functional autonomy was assessed by the 8-item, version of the Lawton Instrumental Activities of Daily Living (IADL) score [24], with a higher score indicating greater functional autonomy. BPSD was measured using the Neuropsychiatric Inventory (NPI) score [25], which ranges from 0 to 144; a higher score indicates a greater number/severity of disorders. Anxiety and depressive disorders were collected only when they were suspected as etiological diagnoses for the cognitive complaint.

The participants’ characteristics are presented as numbers and percentages for qualitative variables and means and standard deviations (SD) for quantitative variables. Baseline characteristics were compared among the 3 groups at the anticholinergic and sedative exposure levels using the chi-squared test for categorical variables, and analysis of variance (ANOVA) for continuous variables.

A multivariable linear mixed model with a random intercept and slope was built to examine the longitudinal relationships between anticholinergic and sedative burden and cognitive function. This model allows time-series to vary between individuals and was adjusted for the baseline covariates age, educational level, anxiety disorders, depressive disorders, IADL, and NPI. The duration (in days) of the medication follow-up data available before each MMSE and DBI measurement was considered a time-dependent variable in the model. This model will produce an estimation of MMSE variation (stated as estimate and p-value) according to each outcomes tested in the analyses. The results will also present the interaction between natural MMSE variation during the follow-up length and anticholinergic and sedative burden.

Missing values were imputed only for covariates in the multivariate analysis, with Multiple Imputation by Chained Equations (MICE) methods.

Descriptive analyses were performed with SPSS Statistics for Windows (v21.0; IBM). The linear mixed model was performed in R Statistical Software (v4.1.3; R Core Team 2022) [26]. All tests were two tailed, and a priori p value less than 0.05 was considered to indicate statistical significance.