Dementia: Update for the Practitioner

Antipsychotic medications remain the mainstay in the treatment of psychosis and agitation. I want to emphasize that, although we think of psychosis as being the target syndrome for antipsychotics, in Alzheimer's disease and other forms of dementia the antipsychotics seem to have equivalent efficacy for symptoms of both agitation and psychosis. We do not see the specificity one might expect.

Nowadays many atypical antipsychotics are used because of the neurological side effects. Several years ago we did a study at Columbia in which we showed that the commonly given dose of 0.5 to 0.75 milligrams per day of haloperidol, the typical antipsychotic drug, was no more effective than placebo. We needed to give about 2 to 3 milligrams a day to see an improvement. As we talk about the other antipsychotics, you will see this issue again; a very low dose is no different from placebo, a moderate dose is superior to placebo, and a slightly higher dose is toxic. Knowing the dose ranges is particularly important in dealing with these agents in Alzheimer's disease because there is a very narrow window of appropriate doses.

Another study across twenty sites looked at the use of risperidone in 625 demented patients in nursing homes. The majority of patients had Alzheimer's and some had other types of dementia. Patients were randomized to placebo and were given 0.5 milligrams, 1 milligram, or 2 milligrams of risperidone per day. The results showed that 1 and 2 milligrams were superior to either placebo or 0.5 milligrams. There was no difference between 0.5 milligrams of risperidone and placebo in either efficacy or side effects. If you look at the response rates, the difference is not huge but it is clear.

One striking feature of many of these studies of antipsychotics is that the placebo response is around 33 percent. You would not expect agitated or psychotic patients with Alzheimer's disease to have a placebo response, but they do. This is probably because these symptoms tend to fluctuate over time. During the course of a few weeks, the symptoms may naturally improve on their own, or the way in which the caregiver deals with the patient may change, whereby the symptom profile also changes.

Risperidone Study

30% dropout rate

1 mg and 2 mg superior to either placebo or 0.5 mg

placebo and 0.5 mg daily: no difference

Response rates: 50% reduction in BEHAVE-AD scores
placebo 33%	1 mg 45%	2 mg 50%

The response rates among 625 demented patients in nursing homes to placebo or varying amounts of risperidone.

Courtesy of Dr. Davangere P. Devanand. Source of data: Katz et al., Journal of Clinical Psychiatry 60 (1999): 107–15.

As Dr. Sano discussed, the Rotterdam study showed that if you give anti-inflammatory agents in a epidemiologic study they have some effect on outcome, but these effects are very hard to demonstrate in a controlled trial against placebo. This illustrates the difficulty in moving from general epidemiologic data, to some clinical experience, to controlled trials. However, in the patient group receiving 2 milligrams of risperidone, extrapyramidal signs were prominent, and there was a dose-dependent increase in EPS and sedation. Interestingly, they did not see orthostatic hypertension, which is a known side effect of risperidone, because the doses used are very low relative to those used in schizophrenia.

Because of these results, the recommended optimal dose of risperidone is 1 milligram per day. Half a milligram was no different from placebo, 1 milligram worked somewhat and without too many side effects, while 2 milligrams also worked but had many side effects. These recommendations do not mean that you must prescribe 1 milligram a day; individual patients are going to vary, but there is a narrow dose range in which you can operate.

Another fairly recent study involved 206 nursing-home patients who were assigned placebo, 5, 10, or 15 milligrams daily of olanzapine (Zyprexa) for six weeks. All the active doses were significantly superior to placebo. The slightly odd finding, which is difficult to explain, is that the 5-milligram dose was actually slightly better than the higher doses relative to placebo. There were no differences between the 5-, 10-, and 15-milligram doses relative to one another, but compared to placebo the 5 milligrams performed slightly better.

There was no difference in extrapyramidal signs between the groups; Zyprexa does not cause significant extrapyramidal side effects. Sedation and weight gain occur to a great extent with the higher doses in patients with schizophrenia whom you follow for a long period of time. Since this was a six-week trial, perhaps there was not enough time to see weight gain. Sedation can be a problem with Zyprexa in these patients, and if a patient gets too sedated one has to think about switching to other antipsychotics. Based on this study, it appears that the 5-milligram dose may be optimal.

Quetiapine is another atypical antipsychotic for which extrapyramidal signs are not a problem. There are reports of rare lenticular and corneal opacities in animal studies, but clinically it has been very hard to show that these occur in people. In the trials involving quetiapine in elderly patients with Alzheimer's disease it is impossible to even study this effect, because the ophthalmologists who examined all the patients found that more than 85 percent have some degree of cataract. Since the ophthalmologists do not have a good way of quantifying the change in the percentage of the lens affected by cataract, there is no way to know whether the lenticular and corneal opacities are getting slightly worse.

The dose range for this drug in terms of tolerability is unusually wide compared to the other drugs I have mentioned, from 25 to 400 milligrams. Many psychiatrists treat Alzheimer's patients with up to 500 to 600 milligrams a day. We do not know that yet whether 25 milligrams is too little to show efficacy, but some clinicians and researchers feel the dose should be on the order of 100 to 400 milligrams a day. I should mention that there are no published placebo control trials with quetiapine. There are large-scale studies using other designs, but the head-to-head trial compared to placebo has not been published.

PREVIOUS | NEXT