Medicines and falls

The best-known trial is small and old. In 93 people aged 65 and over living at home in New Zealand, gradual double-blind withdrawal of psychotropic medicines reduced falls over 44 weeks (relative hazard 0.34, 95% CI 0.16 to 0.74). The authors also reported that permanent withdrawal was very difficult to achieve ²⁵.

In residential care the trial evidence is weaker. The 2025 Cochrane review found that medication review and deprescribing, tested as a single intervention, had a very uncertain effect on falls (rate ratio 0.94, 95% CI 0.76 to 1.18; 12 trials) ².

That is an absence of good evidence. It is not evidence that sedatives are safe. Medication optimisation was a component of all seven tailored multifactorial programs that did reduce falls ².

A US cohort study of 29,648 older nursing home residents, designed to emulate a trial, found that starting an antihypertensive was associated with more fractures over the following 30 days (adjusted hazard ratio 2.42, 95% CI 1.43 to 4.08), more severe falls (1.80, 95% CI 1.53 to 2.13) and more syncope (1.69, 95% CI 1.30 to 2.19). The association was numerically stronger in residents with dementia. The cohort was 97.7% male and the study was observational ²⁶.

A French randomised trial, RETREAT-FRAIL, tested the reverse. In 1,048 nursing home residents aged 80 and over who were taking more than one antihypertensive and had a systolic pressure below 130 mm Hg, stepping treatment down made no difference to deaths from any cause over a median of about three years (adjusted hazard ratio 1.02, 95% CI 0.86 to 1.21), and adverse events did not differ between the groups. Systolic pressure was about 4 mm Hg higher in the step-down group ²⁷.

The trial did not show harm from reducing treatment in this group. It also did not show a benefit, and it was not designed to test falls.

Falls are often given as a reason to stop an anticoagulant in atrial fibrillation. A decision analysis from 1999 concluded that, for people aged 65 and over, the risk of falling was not an important factor in choosing antithrombotic treatment ²⁸.

That analysis was a model. It rested on assumptions drawn from people living in the community, compared warfarin with aspirin and no treatment, and predates the newer anticoagulants. For a very frail resident with a limited life expectancy the balance may be different. It is a decision to make with the resident, against their goals of care.

STOPPFall, a European expert consensus, lists 14 classes of medicine to consider in older people at high risk of falls, most of them psychotropic, with guidance on how each might be withdrawn. Its authors note that its effect on falls still needs to be tested in intervention studies ²⁴.

The Australian guidelines for residential care recommend a review of all of a resident’s medicines at least yearly, after a fall and after a new medicine is started, using a structured tool to identify medicines that increase fall risk. They state that psychotropic medicines should only be considered when changed behaviours are causing significant distress or risk of harm to the person or others ¹.