Because of pharmacodynamic and pharmacokinetic changes associated with aging, there are many more risks attendant to treating elderly patients with psychotropics. The FDA has recently issued black box warnings for older adults who may be at higher risk for adverse events than their younger counterparts. This article reviews the concerns associated with the use of antipsychotics in older patients.
Concern about the inappropriateness of off-label prescribing of antipsychotics for older adults was magnified exponentially by the recent addition of an FDA black box warning for older adults with dementia to the prescribing information. In 2003, a warning was applied to risperidone about its increased risk of “cerebrovascular adverse events, including stroke.” Soon thereafter, similar black box warnings for cerebrovascular adverse events were issued for olanzapine and aripiprazole. While the absolute risk difference was 1% to 2% between antipsychotic- and placebo-treated patients, the relative risk was approximately twice as high in the medication group.1
Subsequently, the severity of the warnings escalated as data emerged of an increased risk of mortality in older persons with dementia treated with atypical antipsychotics. Specifically, the mortality black box warning reads, “Analyses of 17 placebo-controlled trials (modal duration of 10 weeks) in these patients [with dementia-related psychosis/agitation] revealed a risk of death in the drug-treated patients of 1.6 to 1.7 times higher than that of placebo-treated patients. Over the course of a typical 10-week controlled trial, the rate of death in drug-treated patients was about 4.5%, compared with a rate of about 2.6% in the placebo group.”
Retrospective data analyses later convinced the FDA to extend this warning to typical antipsychotics. Perhaps this was to prevent increased use of typical antipsychotics in the absence of supporting evidence or as a medicolegal safety choice. In fact, in our review of large retrospective data analyses, 4 of 6 studies indicate a higher risk of death with typical antipsychotics than with atypical antipsychotics.2
The next logical question is, “What is the long-term risk, because most older adults with dementia who take antipsychotics do so for more than 10 weeks?” Ballard and colleagues3 conducted a study in the United Kingdom to address this important question. Patients on various long-term regimens of typical antipsychotics or risperidone were randomly assigned to continue treatment with antipsychotics or to convert their regimen to supposedly indistinguishable placebo pills for 12 months of double-blind follow-up. At 12 months, the modified intent-to-treat analysis found cumulative probability of survival 70% for those maintained on antipsychotics versus 77% for those whose regimen was switched to placebo—a 7% absolute risk difference in favor of placebo, but not statistically significant. Although the trial ended there, participants were followed naturalistically from 24 to 54 months. During this period, the mortality rate for participants who had been in the placebo group during the 12-month controlled experimental phase was considerably lower than that for participants in the antipsychotic group (eg, 71% vs 46% at 24 months; 59% vs 30% at 36 months; 53% vs 26% at 42 months).
These numbers seem alarming and worthy of further study, but there are limitations to applying them to patients with dementia in the general population. Only 1 atypical antipsychotic (risperidone) was included in the analysis, and there was no statistical difference in mortality during the controlled phase of the study. After the initial 12 months, the investigators had no idea what factors were at play with patients in each group—including whether persons originally taking antipsychotics stopped them or persons in the placebo group started taking antipsychotics. Nonetheless, the trends reinforce the notion that trial tapers off antipsychotics are a good idea whenever possible. The actual statistics, however, should probably not be used in quoting risks to patients/families until such data emerge in controlled trials.
Amid the concerns about the safety of antipsychotics for patients with dementia, additional disappointing news emerged from the Clinical Antipsychotic Trial of Intervention Effectiveness for Alzheimer Disease (CATIE-AD). CATIE-AD compared the effects of risperidone, olanzapine, quetiapine, and placebo in patients with Alzheimer disease who had psychosis and/or agitation.4 There was no difference in overall efficacy (time to treatment discontinuation) between any of the active treatment groups and the placebo group. The superior efficacy in symptom reduction with risperidone and olanzapine was offset by early treatment discontinuation secondary to adverse effects.
Other evidence also supports modest overall efficacy of atypical antipsychotics for psychosis and/or agitation in patients with dementia. When data for individual drugs were combined in a 2006 review of 15 randomized, controlled trials (RCTs), psychosis scores improved significantly in trials using risperidone; also, global neuropsychiatric symptoms improved with aripiprazole and risperidone.1
Yet, even when antipsychotic therapy is effective, once a patient is clinically stable for a reasonable duration (perhaps 2 to 3 months—there is no evidence-based answer), a trial taper off the medication should be initiated. This suggestion is supported by recent data from the dementia antipsychotic withdrawal trial (DART-AD), in which long-term–care residents with Alzheimer disease were randomized to continue their current antipsychotic regimen or to switch their regimen to placebo and were followed for 12 months.4 There were no significant differences in cognitive symptoms or global neuropsychiat-ric symptoms—both of which worsened—between the groups. The only suggestion of between-group differences was less severe neuropsychiatric symptoms for participants maintained on antipsychotics who had relatively high baseline neuropsychiatric symptom severity.
The “graying” of America is no demographic secret, and as many as 50% of persons older than 85 years have some form of dementia (Alzheimer disease being the cause in at least two-thirds of cases). Neuropsychiatric symptoms are extremely common in dementia and often are the reasons caregivers/family members seek treatment—especially from psychiatrists. Psychosis occurs in approximately 40% of persons with Alzheimer disease, and agitation occurs in 80% or more of persons with dementia at some point.5 Although dementia is diagnosed on the basis of cognitive deficits, behavioral and psychological symptoms are often the primary cause of adverse outcomes in dementia—including increased caregiver morbidity, decreased patient and caregiver quality of life, and hastened patient institutionalization.6
What clinicians, patients, and caregivers are left with is an unenviable quandary. While there are some alternative medications that have been tried off-label to target psychosis or agitation in patients with dementia, their overall evidence base is weak and their safety profiles in dementia (while they do not yield the legally ominous black box) are unknown. Although psychosocial/behavioral treatments are likely to be underused because they are time-intensive and not well reimbursed, they are also generally more difficult to study than medications. This leaves many unanswered questions about the empirical efficacy of many proposed psychosocial/behavioral therapies. A 2005 review of RCTs reported the best empirical evidence for caregiver psychoeducation/support, music therapy, cognitive stimulation therapy, Snoezelen therapy, behavioral management–based techniques, and staff training/education.7
Table 1 outlines the medications that have shown at least some promise in efficacy for psychosis or agitation in patients with dementia. However, these studies in the aggregate still pale in comparison to the number of trials and overall strength of evidence (albeit modest) for at least efficacy (if not effectiveness) of antipsychotics for such symptoms.2
Clinicians must digest this dizzying accumulation of data and act. What should occur, if possible, is a process of informed consent and shared decision making about the options for intervention. These include nonpharmacological management strategies, no intervention, using off-label medications with their available risk-benefit information, or institutionalizing the patient.
In a previous publication, my colleague and I2 proposed a potential algorithm for decision making regarding the off-label use of antipsychotics for patients with dementia-related psychosis or agitation (Figure). It may provide some guidance in admittedly murky waters. Table 2 outlines suggested starting and target doses of atypical antipsychotics if used in patients with dementia. One important point to consider in discussions of informed consent is the means in which one communicates numbers related to risks and benefits. Numerical literacy is notably lacking in many persons and often clouded by emotion. A helpful hint is to recall that citing absolute risks (versus relative risks) usually conveys more realistic and accessible information.
The illness for which the most evidence exists for the efficacy of antipsychotics in older adults is schizophrenia. While industry-sponsored trials make a nominal nod for the need of clinical trial data in older adults, trials for FDA approval for schizophrenia may exclude the old-old (85 years or older) or include a small nonrepresentative subsample of older adults. This is why trials designed specifically to assess the efficacy of antipsychotics in older adults were major milestones in recent years.
In 1 such trial, among older adults aged 60 to 86 years with schizophrenia, haloperidol (mean dose, 9.4 mg/d) was compared with olanzapine (mean dose, 11.9 mg/d).8 Olanzapine proved superior to haloperidol on global positive and negative symptoms and also caused fewer extrapyramidal symptoms. Although olanzapine has more inherent anticho-linergic potency than haloperidol, this was offset by the use of anticholinergics to treat haloperidol-related extrapyramidal symptoms. Olanzapine caused more weight gain than haloperidol and tended to increase glucose levels, whereas haloperidol increased prolactin levels to a greater extent.
Another RCT of antipsychotics among older adults compared olanzapine (median dose, 10 mg/d) with risperidone (median dose, 2 mg/d).9 The 2 medications showed equal efficacy in reducing positive and negative symptoms associated with schizophrenia or schizoaffective disorder. Weight gain, observed in both treatment groups, was significantly greater in the olanzapine-treated group. Extrapolating from these studies and others that have examined real-world prescribing patterns, suggests that antipsychotic doses in older adults be one-third to one-half those used in comparable younger patients.
With reasonably established efficacy in treating older adults with schizophrenia, the question of possible adverse effects arises. The risk of tardive dyskinesia (TD) is 4 to 5 times higher in older patients with schizophrenia than in their younger cohorts.10 This risk is diminished with the use of atypical versus typical antipsychotics (eg, 4 times higher relative risk of TD with the use of haloperidol 1 mg/d vs risperidone 1 mg/d for 9 months).11 Thus, the American Psychiatric Association (APA) treatment guidelines for schizophrenia suggest more frequent monitoring for TD in older patients (ie, every 3 months for older adults receiving typical antipsychotics and every 6 months for those receiving atypical antipsychotics). Older adults are also particularly susceptible to antipsychotic-induced parkinsonism, especially from typical agents and, relative to other atypicals, risperidone.
However, atypical agents come with their own problem—namely, the metabolic syndrome. Atypical antipsychotics may cause substantial abdominal weight gain, hyperglycemia, and dyslipidemia, which increase the risk of cardiovascular disease. The metabolic effects of atyp-ical antipsychotics, which were first observed in younger adults, also apply to older adults, although some evidence suggests that these adverse effects decrease with increasing age. The APA and the American Diabetes Association (ADA) published a joint consensus on monitoring metabolic adverse effects of antipsychotics (Table 3), and for now, this extrapolates to older adults as well.11
Warnings regarding adverse metabolic effects extend across the spectrum of atypical antipsychotics, although evidence suggests the risk is somewhat drug-specific. Clinical experience along with published data (eg, from the APA/ADA guidelines) suggest that metabolic risks are highest with clozapine and olanzapine, intermediate with quetiapine and risperidone, and lowest with aripipra-zole and ziprasidone. All atypicals (except clozapine) have shown the same efficacy in RCTs, but they are not created equal. Their different adverse-effect profiles can help guide treatment choices (eg, use of quetiapine in persons with parkinsonism because of low risk of extrapyrami-dal symptoms, use of olanzapine in someone with anorexia and weight loss, or avoidance of higher doses of olanzapine in those susceptible to anticholinergic effects).
The antipsychotic-associated risk of stroke and death in dementia raises the obvious question of whether this is truly diagnosis-specific or more a phenomenon of risk inherent to age. This remains an unanswered question, and FDA warnings are thus far specific to dementia. Some retrospective data indicate an increased risk of stroke in older users of antipsychotics—even in those without dementia; however, the relative risk was higher in patients with dementia than in those with other symptoms (3.50 vs 1.41).12
A large study reported that persons aged 30 to 74 years with varied diagnoses taking antipsychotics had a doubled incidence rate (absolute rates) of sudden cardiac death compared with antipsychotic-naive persons. However, this remained a rare event (nonusers, 0.143%; typical antipsychotic users, 0.294%; atypical antipsychotic users, 0.28%).13 The risk appeared dose-related, and while rates of sudden death increased with age, this was not a medication-age interaction but rather a reflection of baseline higher rates of sudden cardiac death with increasing age. A postulated mechanism is antipsychotic interference with cardiac potassium channels that predisposes to prolonged QT intervals, which, in turn, increases the risk of arrhythmia.
A safe practice might include pretreatment and posttreatment ECGs with attention to corrected QT (QTc) intervals when prescribing antipsychotics for older adults, especially those with cardiac disease. While precise guidelines do not exist, a baseline QTc interval greater than 450 milliseconds would warrant cause for reconsidering antipsychotic treatment or at least close follow-up ECGs and coordinated care with a cardiologist. A QTc greater than 500 milliseconds or an increase of more than 50 milliseconds after starting an antipsychotic would probably prohibit further treatment with that agent. It may be fruitful to look for reversible causes that could be contributing to a prolonged QTc interval, such as untreated cardiac illness; electrolyte disturbances (especially hypokale-mia or hypomagnesemia); or other medications, such as tricyclic antidepressants, macrolide antibiotics, or methadone.14
Research on antipsychotics for mood disorders among older adults is quite sparse. Various atypical antipsychotics are FDA-approved for a spectrum of indications in bipolar disorder (eg, acute manic or mixed episodes, bipolar depression, and/or maintenance mood stabilizer therapy). The same issues exist with studies submitted for FDA approval as with those previously described in reference to schizophrenia (ie, exclusion of older adults or including a small, nonrepresentative subsample). A post hoc analysis that examined the older adult subsample in a trial of quetiapine for bipolar mania reported similar positive efficacy among older adults compared with younger adults.15 No RCTs specifically designed for older adults with bipolar disorder were found.
In the absence of data, antipsychotics remain a reasonable choice for treating various phases of bipolar disorder in older adults—especially psychotic mania or depression. The same safety precautions, dosing suggestions, and monitoring protocols discussed in reference to schizophrenia apply in mood disorders.
Antipsychotics in combination with antidepressants are the mainstay of treatment of patients who have major depression with psychotic features, despite scant empirical studies. Older adults with psychotic depression respond well to electroconvulsive therapy, but many barriers may impede its implementation.
A recent RCT examined the efficacy of olanzapine (15 to 20 mg/d) with placebo or with sertraline (150 to 200 mg/d) for psychotic unipolar depression.16 This study included many older adults and compared response by age—older adults responded as well as younger adults, and combination treatment was found to be superior to monotherapy. Dyslipidemia was an adverse effect in older adults, but significant hyperglycemia occurred in the younger sample only, and weight gain was more prominent in the younger adults; this finding is consistent with accumulating evidence that atypi-cal antipsychotic-related metabolic changes are more prominent in younger adults.
Aripiprazole and quetiapine recently received FDA approval for augmentation in treatment-resistant major depression. Again, little is known about how to extrapolate these findings to older adults. The only relevant data were reported from an open-label trial of aripiprazole augmentation (mean dose, 9 mg/d) for adults 65 years or older with major depression without remission despite adequate trials of escitalopram followed by a trial of either duloxetine or venlafaxine.17 Aripiprazole augmentation yielded a remission rate of 50% after 12 weeks, although interpretation is clearly limited by the open-label trial design. However, a multisite clinical trial funded by the NIMH is currently under way to test the benefits—and risks—of aripiprazole augmentation in treatment-resistant late-life major depression.
Prescribing antipsychotics for older adults has never been under more scrutiny than now. Due diligence to considering and discussing the risks and benefits of antipsychotic treatment with older patients and/or their caregivers is critical. There is no perfectly safe or completely effective treatment for the symptoms of older adults that prompt consideration of antipsychotic therapy. Nevertheless, these agents remain reasonable treatment choices for patients with schizophrenia and bipolar disorder. Their use for dementia is more precarious, but on occasion, their use is within reasonable clinical practice, especially given the dearth of alternatives. The use of antipsychotics in patients with unipolar major depression or anxiety disorders requires further study before reasonable recommendations can be made. Antipsychotics can improve symptoms and quality of life for select older adults, but older adults’ sensitivity to the adverse effects of these medications warrants considerable vigilance.
References1. Schneider LS, Dagerman K, Insel PS. Efficacy and adverse effects of atypical antipsychotics for dementia: meta-analysis of randomized, placebo-controlled trials. Am J Geriatr Psychiatry. 2006;14:191-210.