Year of Publication



Public Health

Degree Name

Dr. of Public Health (Dr.P.H.)

Committee Chair

Erin Abner, PhD

Committee Member

Steve Fleming, PhD

Committee Member

Daniela Moga, MD, PhD


Background: Dementia and sleep disorders are two significant health problems in older adults. Studies suggest that sleep disorders and their treatment might associate with the risk of dementia. Z-drugs are one of the most widely used prescription hypnosedatives in the United States (US) for insomnia in addition to benzodiazepines (BZDs). Studies have shown that BZDs are associated with dementia among senior users. Study Aim: The purpose of this study is to evaluate the association between Z-drug initiation and cognitive impairment among the elderly in the US. Study Design and Data Source: We conducted a retrospective cohort study with a new user design using the Uniform Data Set (UDS) from the National Alzheimer’s Coordinating Center (NACC) from 2005-2017. Methods: We performed Inverse Probability of Treatment Weighting (IPTW) based on propensity scores generated from data obtained at the study baseline and Incidence Density Sampling (IDS) for the non-user control selection. Descriptive analysis, repeated measures general linear regression, and survival analysis were performed to estimate the crude and adjusted risk of Z-drug use and cognitive impairment. Results: The neuropsychological outcomes were measured by the Mini-Mental State Examination (MMSE), Clinical Dementia Rating (CDR) sum of boxes, and Trail Making Test (TMT) A and B. The decreased MMSE, increased CDR summary score, and the increased Trail A and Trail B test scores within user groups indicated that cognitive function declined over time among the UDS participants regardless of the hypnosedative initiation (P<0.05). The significant difference in the change in the Trail B test score among Z-drug users compared to non-users suggested that Z-drug users had a slower decline in the Trail B test one and two years after initiation (P<0.05). Z-drug users also had a worse survival rates than BZD-users from one year after initiation in the Kaplan-Meier analysis (P<0.05). Conclusions: Most neuropsychological tests showed decline in global cognitive and executive function among users and non-users during follow-up. Z-drug users had a slower decline in executive function than non-users over time but worse survival rates than BZD-users one year after initiation. Future studies with a prospective study design may further explore the drug-outcome association and the dose-response relationship.

Included in

Public Health Commons