Health Services Research & Development

Veterans Crisis Line Badge
Go to the ORD website
Go to the QUERI website

2006 HSR&D National Meeting Abstract


1069 — The Use of Antipsychotics in Dementia: Is There An Increased Risk of Mortality?

Author List:
Kales HC (SMITREC)
Valenstein M (SMITREC)
McCarthy J (SMITREC)
Ganoczy D (SMITREC)
Blow FC (SMITREC)

Objectives:
Psychiatric symptoms are common among patients with dementia and associated with poor outcomes, including nursing home placement. Studies indicate clinically modest but significant reductions in psychiatric symptoms with antipsychotics. However, a recent FDA advisory warned of increased mortality risk with use of second-generation antipsychotics (SGAs) in patients with dementia, raising questions regarding the common clinical practice of using these agents to treat secondary psychiatric symptoms.

Methods:
We used national VA data to compare 12-month mortality in 12,906 patients with dementia aged >65 years with new psychiatric medication starts in FY 2002 and 2003 of: 1) first-generation antipsychotics (FGAs), 2) SGAs, or 3) other psychiatric medications (OPMs).

Results:
Among the sample, 55% (n=7113) were started on OPMs; 39% (n=5062) were started on SGAs; 4% (n=474) on FGAs; and 2% (n=257) on FGAs and SGAs. A new start of any type of antipsychotic was predictive of mortality (all p<0.0001) compared to a new start of OPMs. Age and medical comorbidity were also associated with mortality (both p<0.0001). Additional separate models controlled for specific comorbidity indicators (vascular, pulmonary and renal disease, diabetes, and delirium). In the first four comorbidity models, antipsychotic use remained predictive of mortality risk {odds ratios (ORs) between 2.2-2.3 for FGAs, and 1.7 for SGAs}. Initial antipsychotic prescription at inpatient hospital discharge was significantly related to a delirium diagnosis in the prior 6 months, as well as being predictive of mortality. Adding indicators for prescription at inpatient discharge or delirium diagnosis, resulted in smaller ORs for mortality associated with FGAs (OR=2.0) and SGAs (OR=1.5).

Implications:
In veterans with dementia, both first and second-generation antipsychotics were associated with increased mortality compared with other psychiatric medications. Findings relating antipsychotic prescription on discharge, delirium diagnosis, and mortality may suggest that the antipsychotic-mortality association is partly related to use of these agents in delirium, a syndrome with significant morbidity/mortality. Further analyses will examine the specific time period of mortality risk and causes of death among users of antipsychotics versus those using other psychiatric medications.

Impacts:
Based on these results and current expert recommendations, clinical decision-making regarding antipsychotic agents in veterans with dementia should be made on a case-by-case basis.