CATIE support of atypical antipsychotics debated.

TORONTO -- The trial that was supposed to provide definitive guidance on which atypical antipsychotic or antipsychotics work best in schizophrenia apparently has not.

The months after the publication of the first-phase results from the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study, and a few months after the publication of the second-phase results, that question remains openly contested.

CATIE results were debated at the annual meeting of the American Psychiatric Association, and judging from the turnout at the special session, the debate was more than an academic exercise. Clinicians crowded the doorway to the lecture hall, sat in the aisles, and lined the walls to hear about the $42 million trial, which was supported entirely by the federal government.

The question as it came to be formulated in the debate was: Did the CATIE study's comparison of treatments inform clinical practice? Or was it so flawed that it ended up being meaningless?

The study was very carefully designed to mimic what actually happens in clinical practice, said Dr. Joseph P. McEvoy, one of the primary investigators, who spoke first in the debate.

Hence, the study was conducted at 57 different sites to capture diversity, and the participating physicians were left with relative freedom. Moreover, the primary measure of treatment success was chosen to be treatment continuation, because a composite measure of effectiveness and tolerance most accurately reflects how patients and physicians in the real world judge and use a treatment, said Dr. McEvoy, deputy clinical director at John Umstead Hospital, Butner, N.C.

Before CATIE, what was known about the antipsychotics and schizophrenia was that several studies had strongly suggested that the atypical antipsychotics olanzapine (Zyprexa) and risperidone (Risperdal) had greater efficacy than haloperidol (Haldol). And reviews of the data suggested that the lower the dose of haloperidol used (below 20 mg), the less the advantage of the atypicals, Dr. McEvoy said.

No good studies had compared a conventional neuroleptic with ziprasidone (Geodon) or quetiapine (Seroquel). And the studies that compared one atypical with another produced conflicting results with small differences, which appeared perhaps to be biased depending on who was conducting or sponsoring the study, he added.

Therefore, the first question the CATIE study attempted to answer in the first phase was: Are the newer atypicals more effective than conventional neuroleptics?

The investigators chose to use perphenazine (Trilafon, 8-32 mg/day) as their first-generation antipsychotic medication, representing the whole group of drugs, in part because haloperidol had already been compared unfavorably, and they did not want a drug that would be a "straw man" in the trial, Dr. McEvoy said.

"We selected [perphenazine] as a drug that was relatively smooth compared to the others--that you could get efficacy in many cases without [extrapyramidal symptoms] or any need for anticholinergics," he said.

The answer to CATIE's first question was that of the four atypical drugs included in the study, only olanzapine performed better than perphenazine. During the 18 months of the study, 64% of those taking olanzapine quit that drug.

That compared with 75% of those on perphenazine, 74% of those on risperidone, 79% of those on ziprasidone, and 82% of those on quetiapine (N. Engl. J. Med. 2005;353:1209-23).

The median time to discontinuation was also significantly longer among the patients taking olanzapine, while there were no significant differences in discontinuation time among the other agents. Olanzapine also was less likely to be discontinued because of lack of efficacy and was associated with fewer hospital admissions, Dr. McEvoy said.

Although other CATIE investigators have noted that olanzapine may have had more side effects, Dr. McEvoy did not.

The second question that CATIE asked, in the latter phase of the trial, was this: Is clozapine more effective than other atypicals in patients with some degree of treatment resistance (that is, those who discontinued their first drug in the initial part of the trial)?

The addition of a study arm to compare ziprasidone with the other four atypicals caused the only unforeseen difficulty of the study's execution that was mentioned by Dr. McEvoy. But, because of its favorable metabolic profile relative to the other atypicals, it deserved to be compared head-to-head with other atypicals as much as clozapine did, he said.

The study protocol called for patients who discontinued their drug in the first part of the trial to enter an arm comparing either clozapine or ziprasidone with the other four. The investigators intended that patients who discontinued their first drug for lack of efficacy would enter the clozapine arm and those who discontinued because of side effects would enter the ziprasidone arm. However, patients and physicians were given discretion, and few of the patients were willing to enter the clozapine arm.

Ninety-nine patients entered the clozapine arm, where they had a 50% chance of being assigned to clozapine and a 50% chance of being assigned to one of the other drugs. But 444 patients entered the ziprasidone arm, which had the same 50%-50% distribution.

Clozapine was the only drug not given blindly, so that patients who got assigned to another drug would not have to be subjected to the weekly blood testing they normally would have to undergo, because this was thought to be the kind of attention that could sway the results.

The answer to the second question was that clozapine was more effective than the other agents, Dr. McEvoy said (Am. J. Psychiatry 2006;163:600-10).

In the ziprasidone arm, olanzapine and risperidone had better effectiveness as measured by all-cause discontinuation and discontinuation for lack of efficacy, but, as expected, ziprasidone did have the least effect on body weight and lipids and the fewest extrapyramidal side effects (Am. J. Psychiatry 2006;163:611-22).

"If you are lucky enough to respond to ziprasidone, you get the best metabolic profile," Dr. McEvoy said.

The study results, however, run counter to conventional wisdom and conflict with much of the previous literature, said Dr. Herbert Y. Meltzer, who took the skeptical side of the debate.

Referring to the finding that perphenazine performed as well as the atypicals--except perhaps olanzapine--Dr. Meltzer noted that clinicians choose an atypical 95% of the time they prescribe. That suggests some disconnect with the truth, he said.

"If they were really no better than the earlier generation of drugs, I think they would not be dominating the market like they are," said Dr. Meltzer, director of the psychopharmacology division at Vanderbilt University, Nashville, Tenn.

Limitations of the study included the patient population. The discontinuation rate overall in the study was 74%, which was high, suggesting that the study execution favored switching, he said. Moreover, 60% of those enrolled had already been on an atypical prior to the study, and they would not have entered the study if they were easily manageable patients.

Dosing also appeared to be a problem, as a number of critics have noted, according to Dr. Meltzer. The allowed dosage of perphenazine used was too low, and the permitted dosage of olanzapine was higher than its package insert recommends.

"To put it simply, this was not a fair fight," Dr. Meltzer said.

"I think it is a real throwback, regardless of efficacy--which I do not concede--to start recommending typical antipsychotic drugs as first-line treatment," he added.

BY TIMOTHY F. KIRN

Sacramento Bureau