Date of Award

Summer 2021

Document Type

Open Access Dissertation


Epidemiology and Biostatistics

First Advisor

Matthew Lohman


The main objectives of this dissertation were to quantitatively assess whether there is an association between medications (e.g., antibiotics, PPIs, antipsychotics) and cognitive performance over time, conduct a causal mediation analysis to evaluate the role of inflammation (C-reactive protein) in the association between medications and subsequent cognitive decline, and examine if dysbiosis-related medications (DRM) are associated with a greater risk of dementia for participants that were cognitively normal and had MCI at baseline.

The Health and Retirement Study (HRS) was used to conduct a secondary data analysis that used longitudinal data from 2004 to 2016. The HRS is an ongoing, nationally representative prospective sample of individuals aged 50 years and older interviewed biannually since 1992. Beginning in 2005, the Prescription Drug Study was initiated, which captured usage of prescription medications. Additionally, in 2006, HRS began conducting enhanced face-to-face interviews, which included information on C-reactive protein as a marker of inflammation.

Chapter 4 examined the longitudinal association of specific DRM and grouped medications with cognitive performance over time. Our results showed that SSRI antidepressants, antipsychotics, antihistamines, and taking multiple medications were associated with worsening cognitive changes. Chapter 5 provided a more in-depth picture by which DRM may be associated with cognitive scores by understanding if inflammation mediates the association between DRM and cognitive scores. Our study did not provide any statistically significant results. Finally, chapter 6 explored how medications may differentially impact the association of those that were cognitively normal and had MCI at baseline with having dementia. This chapter found evidence that those who had MCI and were taking two or more medications had greater odds of dementia; however, these odds decreased over time compared to those not taking dysbiosis-related medications.

While these findings provide evidence to support and expand the current knowledge, it is critical to understand further how DRM may impact the gut microbiome over time and how these changes may be associated with cognitive changes. Future research should aim to understand how medications and their combinations directly impact the gut microbiome and how these changes may influence cognitive changes over time, specifically those in mid-to-late-life.

Included in

Epidemiology Commons