TCR: Dr. Hsiao, you were NIMH’s project officer for the CATIE trial. What was your goal when you came up with the idea for CATIE?
Dr. Hsiao: Well, our discussions about the CATIE trial began back in 1998, and at that time there really wasn’t much research in comparative trials of antipsychotics. There were trials comparing atypical antipsychotics with placebo or with Haldol (haloperidol), but there were very few head-to-head comparisons. NIMH wanted to change that.
TCR: Did researchers come to NIMH with the idea of doing the CATIE trial?
Dr. Hsiao: No. In fact, there seemed to be very little interest in the research community in doing such a trial, so we had to go looking for someone who would do it! We put out an RFP (Request for Proposal), and we chose Jeffrey Lieberman as the lead investigator.
TCR: What were some of the challenges of putting together such a huge research endeavor?
Dr. Hsiao: One challenge was that we couldn’t afford to pay subjects as much as the drug companies, so we were afraid that we wouldn’t be able to recruit enough patients to make the trial work. But in the end, we achieved our recruitment goals. At the outset of the trial, one of the big questions was whether to include a typical antipsychotic as a comparator and I advocated strongly against it.
Dr. Hsiao: Because at that point, most of us thought that it was clear that atypicals were better than typicals; that was the general opinion in the psychiatric community. Luckily, however, several of the other researchers argued that we needed to include a typical antipsychotic, and they prevailed. And ultimately, we were very clever in choosing Trilafon (perphenazine) instead of Haldol (haloperidal), because Trilafon, being a medium-potency antipsychotic, was tolerated very well.
TCR: Talk to us a bit about how you decided on the dosing in the trial, because that has turned out to be a controversial issue.
Dr. Hsiao: We basically went to the drug companies, and said, “Here’s your shot at making your drug shine, so what dose do you want to use?” Our methodology was to give patients up to 4 capsules a day, so we asked the companies what unit dose they wanted us to put in each capsule. Janssen decided to go with 1.5 mg, which meant a maximum of 6 mg, and they may regret that decision now because they feel that Risperdal was dosed too low – the average was 3.9 mg. We also chose the fairly low dose of 40 mg of Geodon (ziprasidone) per capsule, and this was partly due to concerns about QT intervals. This meant that the maximum allowable dose was 160 mg, and in retrospect I think we should have gone up to 200 or 240 mg.
TCR: Did it turn out that QT interval increases were not a problem in the trial?
Dr. Hsiao: No, they were not. There were no significant increases in the QT interval with Geodon or any of the other drugs.
TCR: What about the dosing of Trilafon?
Dr. Hsiao: We clearly underdosed Trilafon, only going up as high as 32 mg, whereas the maximum in the PDR is 64 mg. We kept it at this dose in order to minimize EPS.
TCR: What about the Seroquel (quetiapine) dosing? There’s certainly a lot of variability among practicing psychiatrists in dosing this drug.
Dr. Hsiao: AstraZeneca put 200 mg in each capsule, so that the maximum allowable dose was 800 mg. Still, the average dose used by researchers was 543 mg, which is definitely a low dose for Seroquel.
TCR: I’ve spoken to AstraZeneca reps who complain that Seroquel was disadvantaged because if you increase the dose too quickly, it causes a lot of sedation, making it less likely that an adequate dose was achieved.
Dr. Hsiao: I don’t really buy that. We took pains to deal with this potential problem by using a blinded starter pack for Seroquel in which patients were started at 100 mg QD for the first few days. And in fact, there weren’t any more early dropouts in the Seroquel group than in most other groups.
TCR: Dr. Hsiao, speaking frankly, when you look at the results of the CATIE trial, what do you think? At first glance it looks like Zyprexa (olanzapine) comes out ahead, but then you see that Zyprexa was dosed so much higher relative to its PDR maximum than the others.
Dr. Hsiao: Right, and at times when I look at the results I wonder if it was entirely dosing that drove them. However, that really isn’t the question. Everyone tends to approach CATIE as a horse race, asking “Which horse won?” But that’s not a good way to understand the results from CATIE. The main thing CATIE shows is that the drugs are really quite different from one another. Zyprexa might be more “effective,” but it also causes much more weight gain. Geodon actually causes weight loss, but it’s not such a great antipsychotic. Treatment choices have to be individualized. That’s not news for any clinician. What CATIE can add is actual data to help patients and docs pick which antipsychotic to try next. There’s a ton of data yet to be mined from the CATIE trial.
TCR: What about Trilafon’s good performance? This is certainly a result that got a lot of media airplay.
Dr. Hsiao: That’s true. We were really surprised by how well Trilafon did – remember I wanted to exclude typicals from the trial but was fortunately outvoted. Amazingly, the rate of EPS for patients on Trilafon was not different from that of the atypicals, even though Trilafon had more dropouts due to EPS. I think what that tells us is that some people might benefit from the typicals and clinicians probably should be prescribing them more. In Great Britain about 50% of antipsychotic prescriptions are for typicals, and in the U.S. it’s only about 10%. The “pie in the sky” dream, which I don’t know how to achieve, would be for us to use more typicals and use the money saved to hire more case workers and add more psychosocial interventions.
TCR: I know that you’re not at liberty to preview the results of Phase 2 and 3 of CATIE, but can you at least outline the methodology of those phases and what kinds of questions you hope they might answer?
Dr. Hsiao: In Phase 2, people who discontinued their Phase 1 medication could go into either of two pathways: the effectiveness pathway or the tolerability pathway. In the effectiveness pathway, patients who were originally assigned to one of the newer atypicals who discontinued due to efficacy failure are randomly assigned either to double-blind treatment with one of the atypicals that they had not received (50%), or to open-label treatment with clozapine (50%). This pathway will help clinicians decide whether it’s best to go to clozapine right away after a treatment failure with an atypical, or whether it’s just as good to go with an alternative atypical that doesn’t pose as many side effect problems as clozapine.
TCR: And the “tolerability” pathway?
Dr. Hsiao: In the tolerability pathway, patients in Phase 1 who discontinued an atypical because of side effect issues are randomly assigned to double-blind treatment with one of three atypicals (Zyprexa, Risperdal, or Seroquel) or to treatment with the low-side-effect atypical Geodon. Hopefully, this will guide clinicians on which of the atypicals are most tolerable.
TCR: Finally, what’s Phase 3?
Dr. Hsiao: This is an open treatment phase in which patients who discontinued the drug they took during Phase 2 will have the option of choosing one of several different antipsychotics, including some combinations of typicals and atypicals. This phase is not blinded or controlled, but may provide some interesting clinical hints to guide future research.