Real-World Data Show Trazodone Rarely Prescribed for FTD, Despite Trial Support

A retrospective analysis of United States health records finds that benzodiazepines and antipsychotics, not trazodone, dominate real-world prescribing for behavioral symptoms of frontotemporal dementia, even though the drugs linked to more emergency and hospital visits lack disease-specific trial evidence.

Frontotemporal dementia (FTD) patients who were prescribed medication for behavioral symptoms received benzodiazepines far more often than any other drug class, 69.5 percent of the treated group, followed by antipsychotics at 30.4 percent and Alzheimer's disease medications at 15.9 percent, according to a retrospective cohort study published September 3, 2026 in Frontiers in Neurology. Only four patients in the cohort were prescribed trazodone, one of the few drugs to have been tested against placebo specifically in FTD. Patients exposed to the first group of drugs, which the researchers classed as "potentially harmful medications" because of their documented ties to falls, pneumonia, hospitalization and death in older adults generally, had a 2.20-fold higher rate of emergency department visits, inpatient hospitalizations and critical care admissions during follow-up than patients who were not exposed.

There is no approved disease-modifying treatment for FTD, so clinicians manage its behavioral symptoms, agitation, disinhibition, apathy and compulsive behaviors, with drugs developed for other conditions. The study, led by researchers at the Pontifical Catholic University of Rio Grande do Sul in Brazil with coauthors at the University of Pittsburgh and the University of California, San Francisco, drew on TriNetX, a health research network that aggregates de-identified electronic health records from dozens of United States healthcare organizations. The authors describe it as the first study to use that database specifically to examine FTD, a disease whose rarity has made it hard to assemble large enough patient groups outside a handful of specialist centers.

Trazodone, along with SSRIs such as paroxetine, citalopram and fluvoxamine, and oxytocin, was tracked in the study's descriptive prescribing data but was not placed in the "potentially harmful" category the authors applied to benzodiazepines, older antipsychotics and Alzheimer's drugs. The same research group previewed related work on FTD prescribing and hospitalization outcomes using the TriNetX platform in a conference abstract in December 2025, underscoring how new this line of real-world research still is for the disease.

The authors frame the higher acute-care rate among patients on the flagged drug classes as a signal that this group may already be a higher-risk subset of patients, not necessarily proof that the medications themselves caused the emergency visits and hospitalizations.