Dementia: Update for the Practitioner

In choosing a drug and treatment algorithm, the first order of business is side effects. If you have an elderly patient who is walking a bit slowly, has some bradykinesia and a bit of a tremor or signs of Parkinson's, you want to avoid typical antipsychotics. Even among the atypicals, you might want to choose something like olanzapine or quetiapine rather than risperidone, which even at these doses can cause some extrapyramidal side effects.

If orthostatic hypertension or sedation are a concern, you can still use Haloperidol in the doses I mentioned, but it has become the overall second-line choice. If your patient is sedated or not walking well and you are worried about sedation, olanzapine may not be the ideal choice. You might think of quetiapine, although that also causes a bit of sedation. Conversely, if a patient is not sleeping at all or is wandering agitated at night, you want to use one of the sedating antipsychotics. Rather than think of one drug as being better than the other, because no one has shown that one is better in terms of efficacy, look at the individual patient, think of the potential side effects that would be a concern, and use the drugs accordingly.


	Choice of Antipsychotic If neurologic side effects are a concern, use atypical. Risperidone 1 to 2 mg daily. Olanzapine 5 to 10 mg daily. Quetiapine 25 to 400 mg daily If orthostatic hypotension or sedation is a concern, use typical. Typical antipsychotic Haloperidol 1 to 2 mg daily

	In choosing a drug and treatment algorithm, the first order of business is side effects. Courtesy of Dr. Davangere P. Devanand

We do not have data on ziprasidone and aripiprazole in Alzheimer's patients, though some studies are being conducted with aripiprazole. Ziprasidone can cause some QTc prolongation on the electrocardiogram; though it causes less QTc than thioridazine, there is still some apprehension about using it in older patients. If you are going to use ziprasidone in older patients, you need to monitor cardiograms. It is not clear how often you need to do this, but certainly at baseline and once later on, perhaps a couple of weeks to a month later.

Among anticonvulsants, carbamazepine was shown to be efficacious in a study by Pierre Tariot using a complicated crossover design and a small sample size. Other studies have not shown this to be as effective. Valproate (divalproex) is widely used, but we do not have convincing evidence from placebo control trials. In fact, some trials which were supported by the company that makes valproate struggled to find a difference, and there are no published data showing that valproate is significantly superior to placebo for this condition. In some studies valproate is used as commonly as antipsychotics because its side-effect profile shows that it can be used safely at relatively low doses. This does not necessarily mean that it is working very well.

Other anticonvulsants have not yet been systematically studied. They can be considered for nonpsychotic, agitated, aggressive patients. Titrate doses and monitor for side effects, but be aware that we do not have solid data like we have for the atypical antipsychotics.

Several years ago there were early studies suggesting that benzodiazepines may be almost as effective as antipsychotics in comparison trials. We now know more about tolerance and withdrawal in relation to their addictive potential, and we know that they can worsen cognition even in the normal elderly. In studies done by Nunzio Pomara at the Nathan Kline Institute in Rockland County, New York, normal elderly subjects were given diazepam and lorazepam. Neuropsychological tests for attention and memory showed a negative impact, even with short-term use. If that is happening with the normal elderly, in people with dementia the effect will probably be even worse. We use benzodiazepines at most for short-term crisis management. Lorazepam can be given for a few days for an acute crisis when antipsychotics are not having enough effect.

Behavioral symptoms have a large impact on caregivers and predispose patients to institutionalization. Agitation and aggression are major problems; psychosis is also frequent but less of a problem. You want to consider using low-dose antipsychotics, particularly the atypical drugs. Consider the tradeoff between efficacy and side effects, and in treatments of Alzheimer's disease, monitor target symptoms, extrapyramidal signs, cognition, and ADLs.

There are no good data on whether to withdraw medication after a few months, but we are doing some studies to look at how long people need to be on medication. The recommendation is often to withdraw medication after about six months to see if symptoms return, but this decision should be based on an individual patient. If the patient's agitation or aggression is severe and persistent you may choose to wait longer, while for mild symptoms you may taper the medication more quickly. Medications can always be restarted if symptoms recur.

PREVIOUS | NEXT