A Statistical Summary: Treatment Of Schizophrenia With Newer Antipsychotic Drugs

British physicians developing guidelines for treating schizophrenia with a group of new drugs called “atypical antipsychotics” have found no clear evidence that they are more effective or better tolerated than older, conventional antipsychotics. The guidelines were developed by systematically reviewing randomized controlled clinical trials and using a statistical technique known as “meta-analysis” to assemble the results of many published studies on the treatment of schizophrenia. The recommendation from this extensive effort—published in the December 2, 2000, issue of the British Medical Journal (BMJ)—was that the older, conventional antipsychotics should be tried first.

The volume of scientific information is enormous, and health professionals, patients, scientific researchers and policy makers are often overwhelmed with the unmanageable amount of literature which sometimes yields contradictory results. For example, each month the U.S. National Library of Medicine adds 40,000 citations to its MEDLINE database, which contains research studies from more than 4,300 medical journals published worldwide.

This problem was recognized 30 years ago by Dr. Archie Cochrane, a British epidemiologist who drew attention to the great collective ignorance about the effects on health care, including drug treatments, that stemmed from the fact that published research had not been summarized in a useful format for physicians having to make treatment decisions.

Using the techniques of meta-analysis, the authors of the BMJ study included randomized trials of the atypical antipsychotics amisulpride (a drug not available in the U.S.), clozapine (CLOZARIL), olanzapine (ZYPREXA), quetiapine (SEROQUEL), risperidone (RISPERDAL), and sertindole (a drug withdrawn from the European market for safety reasons), compared to conventional antipsychotics, usually haloperidol (HALDOL) or chlorpromazine (THORAZINE).

In all, 52 clinical trials were identified, embracing results from 12,649 patients. Most of the studies were of short duration (median follow-up 6.5 weeks), but five trials lasted a year or more. Most compared the effectiveness of atypical antipsychotics to haloperidol. In a few studies, chlorpromazine was used as the benchmark. Perphenazine (TRILAFON) and two drugs not available in the U.S.—flupenthixol and zuclopenthixol—were used for comparison to an atypical drug in one trial each.

The dose of haloperidol significantly affected the outcome in a number of trials. The average dose of haloperidol ranged from 6 to 22.5 milligrams daily, and there was a significant advantage for atypical antipsychotics as the dose of haloperidol increased. But the evidence of superiority in efficacy or tolerability for the atypical antipsychotics as a class disappeared when the haloperidol dose (or the equivalent doses of other conventional antipsychotic drugs) was 12 milligrams or less.

Atypical antipsychotics are said to be more effective than the older ones against “negative” symptoms of schizophrenia, including emotional flatness or lack of expression, inability to start or follow through with activities, speech that is brief and lacks content, and lack of pleasure or interest in life. But overall, no evidence suggested that any of the atypical antipsychotics had a specific effect on either the negative or “positive” symptoms, which include delusions and hallucinations, that occur because the patient has lost touch with reality in important ways.

The review found that atypical antipsychotics cause fewer drug-induced nerve problems (so-called extrapyramidal adverse reactions); however, no difference was found in overall tolerability between the two groups of drugs.

All systematic reviews of published literature present problems. One is that the results are only as good as the trials themselves. Another is the fact that trials with positive results are more likely to be published than those with negative results—”publication bias,” which is a problem particularly with small trials. What this means is that in a systematic literature review it is likely that the result will be based mainly on good news because bad news sees print less often.

An editorial accompanying the review notes that in North America the atypical drugs account for nearly three out of four new prescriptions for antipsychotics. The editorial poses a sobering question in view of the review’s results: Is the prescribing of antipsychotics “largely a victory of clinical hope and marketing hype over hard evidence, or is there a truth that is missed in the meta-analysis”?

Truth was not missed in the BMJ review, which is simply a summary of available evidence. What is missing are the “gold standard” controlled trials comparing various antipsychotic drugs, both atypical and conventional, that could provide information on safety and effectiveness that would make prescribing more rational. In the absence of evidence, it seems more likely that the shift to the atypical drugs is largely “a victory of clinical hope and marketing hype.”

Comparative safety and effectiveness studies between old drugs and new ones are not required in the United States for approval of a new drug. This makes it possible for new drugs to be approved without a showing that they are more effective than those already on the market; in fact, they can be less effective or perhaps more dangerous.

The review’s final recommendation—that conventional antipsychotics should usually be used first in treating an episode of schizophrenia—is sound advice both scientifically and economically. But this recommendation does not, and should not, preclude the use of an atypical antipsychotic first if there is a sound reason for doing so.

What You Can Do

You should carefully discuss the adverse effects of various antipsychotic drugs with your doctor if you must make the decision for yourself or a family member to begin treatment. Because there is little evidence available to differentiate the effectiveness of these drugs, the choice may best be made on the basis of their potential toxicity.