This is a retrospective, propensity-score matched cohort study using prospective collected data from the NACC. The NACC started collecting prospective data from approximately 40 specialized Alzheimer’s research centers in the United Sates in 2005. Participants may volunteer to be included in the NACC dataset, or they may be offered enrollment by participating clinicians. We identified all participants who were enrolled between September 1 2005 and January 1 2020. We excluded people if: they did not have at least one follow-up visit, their medication list was not available, they were already on an OAB anticholinergic medication at enrollment, they had a cognitive condition other than normal cognition, mild cognitive impairment, or dementia, they were missing genetic testing results for APOE-e4 allele status, or if there was >4 years between their first and second visit.
Our primary exposure was defined as new reported use of an OAB anticholinergic medication (oxybutynin, tolterodine, solifenacin, trospium, darifenacin or fesoterodine) at a subsequent NACC visit. Medication use was determined by having the patient bring all medication bottles to the appointment, and these were reviewed by study staff. We considered APOE-e3/e4 or APOE-e4/e4 status as being positive for the APOE-e4 allele. Our co-primary outcomes were the clinical dementia rating (CDR, 0=no dementia, 4=severe cognitive impairment) and the mini-mental state examination (MMSE, which measures orientation, attention, memory, language and visual-spatial skills and is scored from 0-30; a lower score represents increased cognitive impairment). Secondary outcomes included the Boston naming test, Wechsler Adult Intelligence Scale-revised (WAIS-R), and the trail making test part B. Outcomes were compared between the baseline visit and the next follow-up visit.
We created a propensity score using 38 variables available in the NACC dataset that were relevant to cognition such as age, gender, language, years of education, marital status, vision and hearing impairment, and the prevalent use of several medications. Patients who were newly taking an OAB anticholinergic medication were matched 1:1 with participants who were not newly taking an OAB anticholinergic medication using the propensity score, data era (pre/post 2015), total number of NACC visits, and based on their cognitive status (normal, mild cognitive impairment, or dementia). Comparisons between primary and secondary outcomes were carried out using t-tests; a conditional logistic regression model (odds ratio, OR) was used to determine if OAB anticholinergic medication exposure predicted a clinically important cognitive decline based on our primary endpoints (≥1 point increase on the CDR, or a ≥3 point decrease on the MMSE). Means and standard deviation (SD) are reported. Statistical analysis was carried out using SAS 9.4, and a two-sided p<0.05 was considered significant. This analysis was exempt from ethics review.