Pharmacological management in frontotemporal dementia and patient outcomes: a retrospective cohort study using the...

Symptomatic treatment remains the cornerstone of therapeutic management in frontotemporal dementia (FTD), as disease-modifying therapies are currently unavailable. However, several medications originally developed for other neuropsychiatric conditions may represent potentially harmful medications (PHMs) in FTD. This study aimed to describe prescribing patterns following an FTD diagnosis and to evaluate the association between PHM exposure and hospital-based acute care (HBAC) utilization. We conducted a retrospective cohort study using TriNetX, a global health research network comprising clinical data from real-world settings. We included 179 individuals with an electronic health record (EHR)-recorded diagnosis of FTD from 43 healthcare organizations across the United States. Propensity-score matching balanced baseline characteristics, with propensity scores estimated by logistic regression. Risk ratios were calculated from observed event proportions in the matched cohorts, and time to first event was compared using Kaplan-Meier estimates and a Cox proportional hazards model. Patients with at least one recorded PHM prescription after diagnosis formed the exposed cohort; those with none formed the unexposed cohort. The two cohorts were matched 1:1 using propensity scores to minimize confounding and to balance key demographic and clinical variables between groups. HBAC was defined as hospital inpatient or observation care, emergency department visits, or critical care services. Among the 179 individuals with FTD (mean age 67.9 ± 14.9 years; 48.0% women), 71 (39.7%) were exposed to PHMs. After propensity-score matching, baseline characteristics were well balanced (56 patients per group). Overall, 39.7% of patients received at least one PHM, most commonly benzodiazepines (69.5%). In matched analyses, PHM exposure was associated with a significantly higher risk of HBAC [risk ratio (RR) = 2.20; 95% confidence interval (CI), 1.35-3.57] and lower event-free survival (12.8% vs. 62.7%). This real-world study of post-diagnosis prescribing in FTD shows that PHM use is common and that a greater proportion of exposed patients experienced HBAC, with a shorter time to the first event; given residual confounding by indication, these findings are exploratory and hypothesis-generating rather than causal.