Introduction

Approximately 80% of individuals with major neurocognitive disorder (dementia) experience behavioral and psychological symptoms of dementia (BPSD).^1,2 These symptoms are divided into 4 syndromes, which include hyperactivity (agitation, aggression, euphoria, disinhibition, irritability, aberrant motor activity), psychosis (hallucinations and delusions), mood liability (depression and anxiety), and instinctual (appetite disturbance, sleep disturbance, and apathy) clusters.³ The cause of BPSD is not completely understood, but the widespread deficit of serotonin in patients with dementia may account for many of the symptoms.⁴

Antipsychotics are used off label for BPSD, especially when agitation, aggression, and psychotic symptoms are present. However, these agents carry a Food and Drug Administration boxed warning for increased mortality when used for dementia-related psychosis.^5-8 Additional antipsychotic adverse effects include extrapyramidal symptoms, metabolic changes, and faster cognitive decline.^9-13

Due to the risks associated with antipsychotics, many clinicians search for other pharmacologic options for BPSD when nonpharmacologic therapies fail. Several studies show selective serotonin reuptake inhibitors (SSRIs) to be effective in treating BPSD. To date, citalopram, escitalopram, and sertraline are the most studied medications from this class.^14-19

Because of the risks associated with antipsychotics and the literature supporting SSRI therapy for BPSD, clinicians may wish to use SSRIs over antipsychotics. The purpose of this study was to evaluate changes in antipsychotic use after SSRI therapy was added for BPSD treatment.

Methods

Results

Two hundred six patients with dementia and concurrent prescriptions for an antipsychotic and SSRI were identified. After reviewing the patients' charts, 166 were excluded from the study. The most common reasons for exclusion were (1) a dementia diagnosis other than Alzheimer or (2) the SSRI prescription was started before the antipsychotic therapy. An additional 4 patients were excluded from analyses because they died during the study period.

Of the 36 patients included in the analyses, the mean age was 81 years, and most were white males. The majority of patients were prescribed second generation (n = 30, 83.3%) versus first-generation antipsychotics (Table 1). Citalopram and sertraline were the most prescribed SSRIs, and the median doses at the 6-month end point were 20 mg and 50 mg, respectively (Table 2). Half the study population (n = 18) was maintained on the same SSRI dose from baseline throughout the 6-month study period. Antidepressant dose increases occurred in 3 (27.2%) of the responders and 11 (44.0%) of the nonresponders.

TABLE 1 Baseline characteristics

View Fullscreen

TABLE 2 Selective serotonin reuptake inhibitors (SSRI) use

View Fullscreen

Overall, 11 patients (30.6%) responded (able to reduce antipsychotic use). Antipsychotic therapy was discontinued in 10 patients (27.8%), and the dose was reduced in 1 patient. The majority of nonresponders continued the same antipsychotic dose from baseline to the 6-month follow-up period (n = 16, 44.4%; Table 3).

TABLE 3 Antipsychotic use 6 months after selective serotonin reuptake inhibitors initiation

View Fullscreen

When comparing specific SSRIs, 72.7% of responders were prescribed citalopram compared to only 32% of nonresponders (P = .034). Adverse effects that led to SSRI discontinuation occurred in 8.3% of the study population, and all these patients belonged to the nonresponder group (Table 2).

Discussion

Several prospective studies found SSRIs to be effective for the treatment of BPSD. Gaber and colleagues¹⁴ evaluated the efficacy of sertraline versus haloperidol for treatment of agitated behavior due to dementia. Both sertraline (25-50 mg/d) and haloperidol (1-2 mg/d) showed a significant reduction in agitation based on the Cohen-Mansfield agitation inventory scores, and patients treated with sertraline showed less extrapyramidal symptoms compared to the haloperidol group. Sertraline was later studied in another trial to determine its efficacy in the treatment of BPSD in patients treated with donepezil.¹⁵ Participants received donepezil (5-10 mg) for 8 weeks and were then randomly assigned 12 weeks of adjunct sertraline (50-200 mg) or placebo. No statistical differences were seen between the two groups in regards to total neuropsychiatric inventory or clinical global impression scales. However, post hoc analyses of patients with moderate-to-severe BPSD found sertraline was associated with a greater improvement in the neuropsychiatric inventory behavioral and psychological symptom subscale compared to placebo (P = .04).¹⁵

Two studies led by Pollock and colleagues^16,17 compared citalopram to antipsychotic therapy for treatment of dementia-related psychosis and behavioral disturbances. The first study¹⁶ had a placebo arm in addition to comparing citalopram to perphenazine. Eighty-five hospitalized patients were included in this 17-day study. Patients treated with citalopram and perphenazine showed significant improvement on several neurobehavioral rating scale subscores, and citalopram users showed significantly more improvement on the total neurobehavioral rating scale compared to placebo.¹⁶ A second study¹⁷ published later compared citalopram to risperidone. Agitation and psychosis scores improved in both treatment groups by the end of the 9-week trial although there were no statistical differences between the two groups. Significant improvement in agitation scores was observed with citalopram treatment (–12.5%) compared to baseline. Psychosis scores improved significantly with both citalopram (–32.3%) and risperidone (–35.2%) compared to baseline.¹⁷

The CitAD randomized clinical trial found citalopram 30 mg daily to be more effective than placebo in improving agitation at 9 weeks based on the neurobehavioral rating scale agitation subscale (P = .04) and the Cohen-Mansfield agitation inventory (P = .008). However, concerning effects, such as worsening cognition (decline of –1.05 on the mini mental status examination) and prolonged QTc interval were also observed in the citalopram group.¹⁸

Last, escitalopram was studied in a randomized, double-blind, comparison trial with risperidone. Neuropsychiatric scores improved significantly at 6 weeks with both risperidone 1 mg/d and escitalopram 10 mg/d. There were no statistical differences in regards to efficacy between the 2 groups. While risperidone appeared to have a quicker onset of symptom relief, escitalopram patients were more likely to complete the study due to fewer adverse effects.¹⁹

To the authors' knowledge, this was the first study to observe antipsychotic discontinuation rates when SSRIs were prescribed for BPSD. Study results showed antipsychotic use decreased in 30% of patients who were prescribed an SSRI. When comparing individual SSRIs, more responders were prescribed citalopram (72.7%). In clinical practice, the use of citalopram may be limited due to a recommend maximum daily dose of 20 mg in those over 60 years due to risk of prolonged QTc interval. However, 7 of 8 citalopram responders in this study were prescribed no more than 20 mg daily, and none of the patients prescribed citalopram discontinued the medication, which suggests the drug was well tolerated.

Sixteen patients (44.4%) in this study were continued on the same antipsychotic dose during the 6-month period. It is possible SSRI therapy was still effective in these patients, but the prescriber chose not to change the medication regimens due to stability of symptoms. A potential future study is to observe the effects of SSRIs in a population in which antipsychotic reduction attempts are mandated, such as a long-term care setting.

The study's inclusion and exclusion criteria were developed to decrease confounding factors that could have influenced the outcomes. Only patients with Alzheimer dementia were included in the study. Patients with other dementias, psychotic disorders, or bipolar disorder were excluded. Furthermore, included patients had to be prescribed the SSRI and antipsychotic for dementia-related symptoms and not for psychiatric symptoms that were diagnosed before the dementia onset. Patients were also excluded if they were prescribed other psychiatric medications that could have influenced the outcomes, such as non-SSRI antidepressants and mood stabilizers.

Despite the authors' efforts to develop a good trial design, there were still several weaknesses to the study. After screening patients for study criteria, the authors were left with a smaller patient population than anticipated. By identifying study patients based on pharmacy data (those prescribed a cholinesterase inhibitor and/or memantine), all patients with an Alzheimer dementia diagnosis were not captured. The authors attempted to reduce this bias by searching for additional patients enrolled in the geriatric psychiatry clinic, but even with this additional step, potential study patients were likely missed. Further, eligible patients may have been missed because they had a misdiagnosed or nonspecific dementia listed in the chart instead of Alzheimer dementia.

Because the trial was retrospective, prescription data were used to determine whether antipsychotic therapy was discontinued or reduced, which could have led to overestimation or underestimation of the results. For example, it could not be determined exactly why a patient's antipsychotic dose was decreased or not refilled. In some cases, the reduced use could have been due to adverse effects from the antipsychotic and not necessarily because the patient had improvement in symptoms. In addition, if patients were prescribed any behavioral medications outside the Veterans Affairs system, this could not be captured. Last, the authors could not be certain of the amount of benzodiazepine medications specific patients used during the study period. Patients prescribed scheduled doses of benzodiazepines were excluded from the study, but those with an “as needed” prescription were still included. It was possible these “as needed” prescriptions were used on a consistent basis for some patients.

Conclusion

Approximately 30% of patients with Alzheimer dementia who were prescribed antipsychotics for BPSD were able to discontinue the medication or had a dose reduction after starting SSRI therapy. In this study, more responders were prescribed citalopram compared to other SSRIs.