Real-world evidence to clarify risk-benefit of antihypertensive intensification vs. de-escalation and blood pressure thresholds to minimize harms and maximize brain and heart health in older adults - PROJECT SUMMARY Hypertension (HTN) affects 60% of U.S. adults aged ≥65 years and is a leading modifiable risk factor for Alzheimer’s disease and related dementias (AD/ADRD) and cardiovascular disease (CVD). Although intensive antihypertensive (anti-HTN) therapy reduces the risk of cognitive decline and CVD events, older and frailer adults are disproportionately vulnerable to adverse events (AEs), including injurious falls, hypotension, acute kidney injury (AKI), and electrolyte abnormalities. Current U.S. guidelines recommend systolic blood pressure (SBP) targets <130 mm Hg for older adults but lack a clear lower threshold for de-escalation, particularly in vulnerable older adults. While recent randomized controlled trial (RCT) evidence suggests that less-intensive SBP targets may benefit certain frail older adults, these populations remain underrepresented in large trials, and a new, fully inclusive RCT is not feasible. This project will generate real-world evidence to determine (1) when to intensify antihypertensive therapy, (2) when to de-escalate antihypertensive therapy, and (3) how age, frailty, comorbidities, and apolipoprotein E ε4 (APOE ε4) carrier status modify these decisions. Our Specific Aims are to: (1) compare the effects of two HTN treatment strategies—intensifying or maintaining anti-HTN therapy without vs. with de-escalation at varying SBP thresholds—on AD/ADRD and CVD risk in older adults, (2) evaluate the safety of these treatment strategies, assessing the risk of injurious falls and other AEs, and (3) establish the patient-specific risk-benefit balance of HTN management based on age, sex, frailty, comorbidities, rurality, and APOE ε4 carrier status. We will leverage large, longitudinal datasets from the Veterans Health Administration (VHA) and Intermountain Health (IH), integrating robust medication and clinical data to address time-varying confounding and missing BP measures. Our approach employs an inductive natural language processing (NLP) platform with large language models to enhance AD/ADRD case identification and utilizes a target trial emulation design with modern causal inference methods to evaluate dynamic treatment regimens. Our team has extensive expertise in causal inference, HTN management, pharmacoepidemiology, and AD/ADRD research, making us ideally suited to conduct this study. By comparing pragmatic SBP management strategies with and without de-escalation thresholds in a diverse older population, this project will provide clinicians with actionable guidance on individualizing antihypertensive therapy, striking the optimal balance between preventing cognitive and cardiovascular complications while minimizing overtreatment risks. In doing so, we will address a critical gap in real-world evidence for populations historically underrepresented in RCTs, ultimately informing clinical guidelines and improving patient-centered care for older adults.
