The vast majority of patients with Alzheimer’s disease will show agitation, delusion, and other disruptive behaviors during the course of their illnesses. These behaviors cause suffering and distress for patients and for caregivers and families, often diminishing quality of life and making care more difficult.

The behaviors are variable and hard to classify. The phenotypes are often unstable and overlapping. We may diagnose depression when we see apathy and aggression when we see psychomotor agitation. The term “agitation” commonly subsumes a wide range of behaviors.

Because certain phenotypes improve with certain drugs across psychiatric disorders – for example, antipsychotics for delusions in both schizophrenia and mood disorder – we often assume that the behavioral phenotypes seen in dementia, such as depression or delusions, will also improve with antidepressants and antipsychotics as they do in other disorders.

Patients with dementia receive psychotropic medications to try and control these behaviors. But are antipsychotics the best choice? Are they effective and safe? The Food and Drug Administration and clinical researchers agree: These drugs are largely ineffective and are associated with significant adverse events in patients with dementia.

Antipsychotics have been advocated for disruptive behavior in dementia since their inception. Manufacturers promoted this use into the mid-1980s, even as concern mounted about their overuse. For example, a 1956 journal advertisement pictured an elderly man aggressively raising his cane, his daughter on the phone to his physician asking, “What can you do for Pop?” The tag line touted Thorazine [chlorpromazine] “for the prompt control of the agitated, belligerent senile.”

The Federal Nursing Home Reform Act was enacted in 1987 (OBRA ’87), and regulations were established in 1991. It required, in part, restriction, justification, and monitoring of antipsychotics to address their potential unnecessary and inappropriate use as chemical restraints.

The advent of atypical antipsychotics in the early 1990s – risperidone (Risperdal) in particular – led to their promotion for treating aggression and psychosis in dementia. It’s fair to say as well that atypicals were enthusiastically adopted by physicians as alternatives to conventional antipsychotics; they were viewed as different, safer, less sedative, and had fewer motor system effects than did the first generation antipsychotics. Indeed, advertisements from 1998 depicted an elderly woman entitled “Hostile outside. … Fragile inside,” calling Risperdal the most prescribed antipsychotic in long-term care, improving a “broad range of psychotic symptoms,” and having few adverse events. The advertisements were remarkably similar to the Thorazine advertisements 40 years earlier.

By the early 2000s, reports of increased cardiovascular events and stroke surfaced. In 2003, the drugs began to carry a warning of cerebrovascular events, including stroke, in elderly patients with dementia-related psychosis. In 2005, this was followed by a black box warning: “Elderly patients with dementia-related psychosis treated with antipsychotic drugs are at an increased risk of death,” adding that the drugs were “not approved for the treatment of dementia-related psychosis.” The latter phrase expressed the FDA’s view that there was insufficient randomized controlled trials evidence of benefit for this indication.

Efficacy of antipsychotics from controlled clinical trials is indeed marginal. Many trials yielded insignificant outcomes, and those that were significant showed only a few points difference on scales of uncertain clinical significance. Notably, when the trials are statistically combined, there is no efficacy evidence for olanzapine (Zyprexa) or quetiapine (Seroquel) and only slight evidence for risperidone and aripiprazole (Abilify), but adverse events with these drugs are considerable, compared with placebo. A characteristic of the trials is that both the drug and placebo groups improve considerably and to such an extent that any statistically significant effect for antipsychotics is marginal or incremental over placebo.

The warnings, however, generated only a relatively small change in prescribing patterns. Risperidone use decreased somewhat, but quetiapine use increased, all the more remarkable because of the lack of efficacy data and the very low quetiapine doses that are used clinically (Arch. Gen. Psychiatry 2011;68:190-7). A 2007 look at Canadian prescribing rates showed a substantial and highly variable amount of antipsychotic use in nursing homes. Rates ranged from 20% to 44%, suggesting that we are not all clear on how to use them (Arch. Intern. Med. 2007;167:676-83).

A recent retrospective study enumerates the risks for death that antipsychotics pose for Alzheimer’s patients. In a large cohort of elderly veterans with dementia, the death rate per 100 person-years was 46 for haloperidol, 27 for olanzapine and risperidone, and 19 for quetiapine, rates that are noticeable in busy practices (Am. J. Psychiatry 2012;169:71-9).

Lawyers and state departments of justice have entered the picture as well with malpractice pleadings and enforcement against elder abuse, questioning whether we physicians are doing our job in protecting patients. Pharmaceutical companies have been sued, prosecuted, settled, and pled to criminal and civil offenses involving past misleading or fraudulent promotion of the drugs. Companies have paid out billions, with the latest being Johnson & Johnson’s anticipated $2.2 billion settlement over misleading marketing claims for risperidone.

Antipsychotics, when used appropriately and for a short time, can be effective for some. But, in general, they are most commonly used chronically. The continuing long-term prescribing of these drugs underlines the need to be able to treat behavioral disorders.

We need to examine how we treat dementia patients with disruptive behaviors and what we are doing with these medications, and consider alternatives. If we don’t, patients may continue to be harmed and to a greater extent than we think. We can do better in treating behavioral symptoms.

Dr. Schneider is a professor of psychiatry, neurology, and gerontology at the University of Southern California, Los Angeles. He is also director of the university’s Alzheimer’s Disease Research and Clinical Center. Dr. Schneider reports being an editor on the Cochrane Collaboration Dementia and Cognitive Improvement Group, which oversees systematic reviews of drugs for cognitive impairment and dementia. He has multiple financial relationships with a number of companies that manufacturer drugs for Alzheimer’s disease.