A Polypharmacy Model and the Association of Polypharmacy with All-Cause Mortality and Incident Cognitive Impairment in the REGARDS Cohort.

摘要

Importance: Medications are a cornerstone of medicine. Americans frequently use many medications simultaneously. While medications are tested individually for safety and efficacy, such complex drug regimens may have many unintended effects, including direct drug toxicity, drug-drug interactions, and adverse drug reactions. The phenomenon of taking many drugs simultaneously is known as "polypharmacy" While polypharmacy can be appropriate and the standard of care, it often occurs unnecessarily and exposes the patient to pharmacologic risk.;Objective: This dissertation sought to fill some of the pharmacoepidemiologic knowledge gaps by exploring factors related to polypharmacy and assessing the associations between polypharmacy and 1) all-cause mortality and 2) cognitive impairment using data from the large REGARDS cohort.;Methods: We first transformed the very large REGARDS medication database by assigning generic names, drug classes, and prescription/OTC/supplement status to each manually recorded medication name. We documented the generic name assignments for over 99% of entries using internet queries of Drugs.com and Google..;The REGARDS Cohort data (total n= 30,183, comprised of blacks and whites ages ≥45 in the continental U.S.) were used. During an in-home study visit, pill-bottle inspections were conducted of all the medications used in the last two weeks. The cohort member's polypharmacy status was subsequently determined by summing the total number of generic (prescription or OTC) ingredients.;Study 1: A logistic model assessed whether polypharmacy status was associated with demographics, socioeconomic status, lifestyle, comorbidities, and biomarkers.;Study 2: Polypharmacy status (major [≥8 ingredients], minor [6-7 ingredients], none [0-5 ingredients]) was determined by counting the total number of generic (prescription or over-the-counter) ingredients. Cox Proportional Hazards models (using both time-on-study and age-time-scale methods to model time to event) were used to assess the relation of polypharmacy to mortality. Several alternative models were constructed to assess confounding by indication and to consider effect modification by CKD.;Study 3: Multiple logistic regression models (using both first follow-up and last follow-up Six Item Screener score to define incident impairment) were constructed to assess the association of polypharmacy and incident cognitive impairment.;Results: Overall, 171,573 in-home visits drug names were transcribed.;Study 1: The mean number of total generic ingredients was 4.12 (SE= 0.039), with 15.7% of the cohort using ≥8 total generic ingredients. White race and stroke belt/buckle or Southern residence were associated with a higher polypharmacy prevalence.;Study 2: Major polypharmacy was associated with increased mortality in all models, with hazard ratios and 95% confidence intervals ranging from 1.22 (1.07-1.40) to 2.35 (2.15-2.56). Minor polypharmacy was associated with mortality in some, but not all, models. The polypharmacy-mortality association did not differ in those with and without CKD.;Study 3: For all models constructed, the major polypharmacy-cognitive impairment odds ratios (ORs) were all greater than 1, but never with a point estimate exceeding 1.30, and most not statistically significant. Conversely, for minor polypharmacy-cognitive impairment, the associations were all near 1, with none of them statistically significant. The two-way polypharmacy-CKD status interactions assessed were not significant.;Conclusions: American adults are using a substantial number of medications. This may expose them to potential risks of drug toxicity, drug interactions, and adverse drug events. While residual confounding by indication cannot be ruled out, in this large US cohort, major polypharmacy was associated with mortality in all models. These findings suggest that a simple ingredient count sum is not strongly associated with incident cognitive impairment.;The racial and regional variation in polypharmacy merit further study. Moreover, the polypharmacy-mortality association should be replicated. However, if these associations are causal, then they could have major public health impacts.