Abstract
IMPORTANCE: Anticholinergics have been associated with functional decline in older adults. Past studies have assumed constant effects over time and have not considered the etiologically relevant exposure window. OBJECTIVE: To examine the association of anticholinergic exposure with gait speed and grip strength assuming constant and time-varying effects of daily exposure. DESIGN, SETTING, AND PARTICIPANTS: This cohort study used data collected from February 1994 to March 2020 in the Adult Changes in Thought study at Kaiser Permanente Washington, an integrated health care delivery organization. Participants with at least 2 study visits and at least 10 years of enrollment prior to index were included. Data were analyzed from January 2023 to December 2024. EXPOSURE: Conventional anticholinergic exposures (10-year total standardized daily dose [TSDD] and 2-year mean SDD [mSDD]) assumed constant daily exposure effects. Weighted cumulative exposures (WCE) explored different exposure windows (T = 2, 4, 6, 8, or 10 years) and were quantified as T-year weighted mSDD to allow for time-varying effects of daily exposure. MAIN OUTCOMES AND MEASURES: Adjusted linear models with generalized estimating equations estimated mean differences (MDs) in change rates in gait speed or grip strength between anticholinergic TSDD and mSDD categories and per unit increase in weighted mSDD. Model fits were assessed by quasi-information criterion (QIC). RESULTS: The total sample included 4283 participants, with 4210 participants (2468 women [58.6%]; mean [SD] age,74.3 [6.1] years) with 8.2 (5.4) years of follow-up in the gait speed sample, and 4200 participants (2458 [58.5%] women, mean [SD] age 74.5 [6.1]) in the grip strength sample. Compared with nonusers, a greater decline rate in gait speed was found for those with 10-year TSDD 1096 or greater (MD per year, -0.0132 [95% CI, -0.0193 to -0.0070] m/s) and for 2-year mSDD 0.5 or greater (MD per year, -0.0101 [95% CI, -0.0174 to -0.0029] m/s). The 4-year WCE model had the lowest QIC and showed a significantly greater decline rate per 1-unit increase in weighted mSDD (MD per year, -0.0034 [95% CI -0.0048 to -0.0019] m/s). There were no significant associations between conventional exposures and grip strength, but the 6-year WCE model had the lowest QIC (MD per year, -0.0329 [95% CI -0.0612 to -0.0046] kg). CONCLUSIONS AND RELEVANCE: Is this cohort study, higher anticholinergic exposure was associated with accelerated decline in physical performance, consistent with clinically meaningful decline. These findings suggest that minimizing anticholinergic medications is important for healthy aging.