Tel. (860) 408-4882  •  Fax (860) 408-4885
Valley Psychiatry Article

Fear Mongering and Vilification of Antipsychotics

Maybe you've seen the recent headlines proclaiming antipsychotic medication use in the elderly demented population being equated to... well to poison. A dubious study in the May 26th edition of the Archives of Internal Medicine (Antipsychotic Therapy and Short-term Serious Events in Older Adults With Dementia) has both the medical and lay press declaring that elderly patients with dementia who are treated with antipsychotics die at a three to four times higher rate within a month than similar patients who do not receive prescriptions for antipsychotics. Even a cursory analysis reveals that their data does not support this provocative conclusion.

This study is an example of how statistics can show a relationship between two items without shedding any light on their possible causal association, or lack thereof. Worse yet, it makes a conclusion about cause (antipsychotic use) and effect (subsequent hospitalization or death) that cannot be substantiated by its own data.

Where did it all go wrong? Two key areas were botched. First, the investigators did not consider the indication(s) for why antipsychotics were prescribed which caused the second key flaw of there being no true control group.

The study retrospectively divided over twenty thousand community residents and roughly the same number of nursing home residents in Ontario into three categories each. Thus ostensibly the only difference between the three groups of community dwelling residents was their exposure to typical antipsychotics, atypical antipsychotics, or no antipsychotics. The same allocation was made in the nursing home population.

They then compared the frequency of two categories of adverse events among the groups. The first category was known adverse effects from antipsychotics such as extrapyramidal symptoms (EPS: rigidity, dystonia, etc.) falls, and cerebrovascular events and the second category was conveniently labeled "other events."

While EPS and falls are known adverse events of antipsychotic use, the presence of these events without a control group of similar patients limits any conclusions. For instance EPS is caused by other medications and is even present in delirium in the absence of antipsychotics. Falls are also a risk of antipsychotic therapy but volumes can and have been written on the diverse and complex causes of falls. The risk of cerebrovascular events and antipsychotic therapy is a small but statistically significant phenomenon. Its clinical relevance is still debatable.

So, predictably, they found that in the community dwelling group individuals taking conventional antipsychotics were 3.8 times more likely to experience any serious event at 30 days follow-up and those taking atypical antipsychotics were 3.2 time more likely to experience the same. In the nursing home group the equivalent risk numbers were 2.4 for the conventional antipsychotics and 1.9 for the atypical antipsychotics compared to the 'control' group. And thus Rochon et. al. concludes that "serious adverse events, as indicated by a hospital admission or death, are frequent following the short-term use of antipsychotic therapy in older adults with dementia." Well, not so fast.

By not accounting for why antipsychotic medications were prescribed, the authors miss what is likely the most important cause of adverse events. Presumably antipsychotics were prescribed because of delusions and behavioral disturbance as well as delirium with its characteristic agitation frequently manifested by screaming, motor hyperactivity, and physical aggression. But this is only an assumption because this study failed to investigate this critical "unmeasured confounder."

Why is this so important? Delirium is an acute change in mental status with a variety of etiologies and is a medical emergency. Delirium independently predicts substantial morbidity and mortality. Thus it is not surprising that medications used to treat the symptoms of delirium in elderly patients with coexisting dementia are associated with high morbidity and mortality. These are very ill people.

Acute myocardial infarction is a medical emergency with high morbidity and mortality. It is often treated with aspirin. Thus the use of aspirin in this context would statistically be associated with heart attack deaths. Using this study's logic, aspirin causes adverse events and death in individuals who experience a heart attack. No one would argue that this logic is faulty.

Is there something unique about individuals with dementia and coexisting delusions and/or behavioral disturbances that predispose them to adverse events independent of treatment? This could also explain the presence of an elevated risk for adverse events in the antipsychotic treatment groups. This study never considered this possibility. It is commonly known that certain psychiatric illnesses predispose individuals to medical morbidity. Schizophrenia and diabetes is one example.

Even if the authors of this study considered the possibility that delirium, psychosis, or behavioral disturbance could independently increase morbidity and mortality, they could not begin to make any hypotheses or conclusions because the study did not have a valid control group. The study design precluded having a proper control group because it failed to address the reason antipsychotics were used.

This study lacked stratification based on behaviors, the presence or absence of psychosis or delirium. For instance, if acutely delirious and aggressive individuals with dementia were divided into different treatment groups based on use of a conventional antipsychotic, atypical antipsychotic, benzodiazepine, or no treatment, then conclusions could be made regarding both the efficacy and safety of these various treatments. The authors failed to even consider this type of stratification and thus made conclusions based on inadequate data.

One can't say their conclusions are wrong, it's just that the conclusions they made aren't supported by either their data collection or analysis.

Antipsychotic medications do have problematic side effects and are associated with cerebrovascular events and metabolic syndrome. Their use should always be a last resort and for as short a duration as possible, especially in a population as vulnerable as elderly individuals with dementia.

Yet most clinicians find these medications effective and they are widely used. The December 2007 American Psychiatric Association practice guidelines for Alzheimer's and other dementias states: "Antipsychotics are the primary pharmacological treatment available for psychotic symptoms in dementia. They are also the most commonly used and best-studied pharmacological treatment for agitation." And it's not as though there are many (or any) effective alternative medications available.

The benefits of antipsychotic treatment must be carefully and continually weighed against their known risks.

This study sheds no new light on this delicate balance and serves only to spread unwarranted feelings of fear and guilt amongst caregivers and families.

Valley Psychiatry Article