Abstract and Introduction
Background: In 2007, more than 18,000 adults aged 65 or older died from injuries related to falls, with up to 30% experiencing severe injuries such as hip fracture or head trauma. The economic impact of falls and fractures among older people is substantial, with direct economic consequences totaling $19 billion in 2000.
Objective: To evaluate the association between antipsychotic and antidepressant agents and the risk of hip fracture in older adults, across multiple studies.
Methods: An English-language PubMed/MEDLINE search for studies from January 1966 to January 2011 was conducted, using key words including aged, hip fracture, fractures, antidepressive agents, and antipsychotic agents, as well as individual drug names. Criteria for study inclusion were mean subject age greater than or equal to 65 years, adjusted for age and sex, hip fracture-specific results provided, data specific to a drug class, subclass, or single agents, and cohort or case-controlled study design. Two authors reviewed all studies for inclusion/exclusion. A random effects model was used to calculate summary odds ratios.
Results: A total of 166 studies were identified in the initial search. Ten antipsychotic-related and 14 antidepressant-related studies, representing more than 70,000 hip fracture cases and approximately 270,000 subjects from 4 continents, met the inclusion criteria. Summary odds ratios (95% CI) were first-generation (conventional) antipsychotics 1.68 (1.43 to 1.99), second-generation (atypical) antipsychotics 1.30 (1.14 to 1.49), first-generation (tricyclic) antidepressants 1.71 (1.43 to 2.04), and second-generation (selective serotonin reuptake inhibitors, serotonin-norepinephrine reuptake inhibitors, and unique agents such as bupropion, mirtazapine, and trazodone) antidepressants 1.94 (1.37 to 2.76). Clear evidence of heterogeneity was noted among all antidepressant study analyses (I > 87%; Q statistic p < 0.05).
Conclusions: All drug classes studied—first- and second-generation antipsychotics and antidepressants—were associated with an increased risk of hip fracture in predominantly older adult populations.
Up to one third of community-dwelling adults aged 50 years and older fall each year. Although most of these events do not result in injury, falls are the leading cause of accidental death and the seventh leading cause of death in adults aged 65 years or older. Of note, 20–30% of falls result in severe injuries such as hip fracture or head trauma. In 2008, more than 19,700 adults aged 65 and older died from unintentional injuries related to falls. The economic impact of these falls and fractures among older people is substantial, with direct economic consequences totaling $19 billion in 2000, equivalent to $28.2 billion in 2010 dollars.
Fall prevention has received considerable attention in research and public policy, resulting in a large body of findings on risks for falling as well as clinical guidelines for fall prevention. Across studies and guidelines, polypharmacy and psychotropic drugs are 2 of the most consistent risk factors for falls, commonly coexisting in the same patient. Procyshyn et al. found that 25.7% of mental health outpatients in British Columbia were exposed to persistent antipsychotic polypharmacy. Two classes of psychotropic drugs—antidepressants and antipsychotics—are among agents that have been strongly associated with falls and fall-related injuries. In a classic meta-analysis, Leipzig and associates confirmed the association of tricyclic antidepressants (eg, amitriptyline, imipramine, desipramine, nortriptyline) and antipsychotics (eg, haloperidol, thioridazine, chlorpromazine, thiothixene) with falls. At the time of the Leipzig et al. meta-analysis, which included studies conducted up to 1996, many of the currently available second-generation psychotropic agents were unavailable or had not been studied sufficiently. These second-generation antidepressants included selective serotonin reuptake inhibitors (SSRIs) (eg, citalopram, escitalopram, fluoxetine, fluvoxamine, paroxetine, sertraline), serotonin-norepinephrine reuptake inhibitors (SNRIs) (eg, duloxetine, venlafaxine, desvenlafaxine), and unique mechanism agents such as bupropion, mirtazapine, and trazodone. Second-generation antipsychotics include aripiprazole, clozapine, olanzapine, risperidone, quetiapine, and ziprasidone. It is unclear whether second-generation psychotropic agents carry the same risk as the first-generation agents, which often have greater anticholinergic, hypotensive, and sedative properties.
However, the ultimate concern is not the falls, but rather the injuries that result from falls, including fractures and other sequelae such as activity limitation and fear of falling. One of the most common and serious injuries is hip fracture. Hip fractures have significant consequences for both the patient and the public in terms of morbidity, mortality, and health care costs, and they account for many of the long-term effects and expenses related to falling. The 1-year mortality rate after a hip fracture is about 20%, and 20% of people living in the community at the time of the hip fracture must relocate to an institutional residence. Twenty-five percent of individuals who experience fractures remain institutionalized 1 year after the fall, and most of those who return to community living never regain their prefracture level of activities.
In rare cases, a hip fracture is the initiating event that results in a fall, but most often the impact of the fall is the cause of the fracture. Thus, the assumption underlying most falls research and fall prevention guidelines is that while only a small fraction of falls results in fractures, prevention of all falls can eliminate hip fractures and other injuries. However, the risks for injury and noninjury falls are not identical. Drugs may affect both the risk for the fall and the risk that the fall will result in a fracture. For example, SSRIs might disproportionately increase injury secondary to a fall because of the role of serotonin receptors in bone metabolism.
In recent years, there have been studies, mostly in Europe, that have used large databases to evaluate the association between drugs and falls and between drugs and hip fractures. The results of studies concerning the relationship of psychotropic drugs, particularly the second-generation agents with distinct pharmacologic mechanisms, with falls and fractures have been contradictory. Although many clinicians have assumed that the second-generation (atypical) antipsychotics and second-generation antidepressants pose less risk than first-generation agents for falls, and by inference, less risk for hip fracture, there have been no systematic reviews and meta-analytic studies focused on the association of hip fracture with first- and second-generation psychotropic agents. Takkouche and colleagues reported a meta-analysis on the association of 7 classes of neuropsychiatric drugs with all types of fractures but did not exclude studies with predominately younger subjects, nor did they distinguish effects of second-generation agents. Because of the impact of hip fracture on life trajectory and quality of life, a review specific to this injury in older patients taking first- or second-generation psychotropic agents is indicated. The aim of our study was to evaluate the association of both first- and second-generation antipsychotic and antidepressant agents with risk of hip fracture in older adults, as reported in the literature.