Risk of falls and fractures in older adults using antipsychotic agents: a propensity-matched retrospective cohort study

Abstract

Background: Antipsychotics, especially atypical agents, are widely used in the elderly population to treat behavioural and psychiatric symptoms. Very few studies have compared the risk of falls and fractures among older adults using typical and atypical agents and none of the studies have evaluated differential risk across antipsychotic classes.

Objective: To examine the risk of falls and fractures associated with atypical antipsychotic use and typical antipsychotic use in community-dwelling older adults in the US.

Methods: The study involved a retrospective population-based cohort design matched on propensity scores involving older adults (aged ≥50 years) using atypical or typical antipsychotic agents in the IMS LifeLink™ Health Plan Claims Database. Patients taking atypical antipsychotics were matched with patients taking typical antipsychotics using the Greedy 5 → 1 matching technique. The study evaluated the relative risk of hospitalization/emergency room (ER) visits due to falls/fractures in a 1-year follow-up period, and patients treated with atypical antipsychotics were compared with those treated with typical antipsychotics using the Cox proportional-hazards regression model stratified on matched pairs. The covariates adjusted for in the regression model included duration of therapy and exposure to other psychotropic medications that increase the risk of falls and fractures.

Results: From July 2000 to December 2007, 11 160 (5580 atypical and 5580 typical) users of antipsychotics were obtained after matching on propensity scores. A total of 825 cases of falls/fractures with at least one hospitalization/ER visit following the use of antipsychotic agents were identified. The number of cases with falls/fractures was 450 in atypical antipsychotic users and 375 in typical antipsychotic users. Cox regression model analysis revealed no statistically significant difference between atypical users and typical users with respect to risk of falls/fractures (hazard ratio [HR] 1.01; 95% CI 0.83, 1.22). However, duration of therapy with any antipsychotic medication for >90 days was significantly (HR 1.81; CI 1.35, 2.43) associated with increased risk of falls/fractures compared with <30 days of treatment.

Conclusions: No statistically significant difference was found between atypical antipsychotic agents and typical antipsychotic agents with regards to the likelihood of falls/fractures in a large cohort of older adults. However, there is a need to be cautious while prescribing atypical and typical antipsychotics in older adults for long periods of time.