From the Editor

What shall we do about CATIE?

Author and Disclosure Information

 

Six years in the making, phase 1 results of the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) in schizophrenia have confirmed my clinical observations:

  • One was the 74% discontinuation rate (though I admit I take some comfort in knowing I am not the only one who has trouble keeping patients with schizophrenia in treatment).
  • Another was that 42% of the patients had metabolic syndrome, and most were not being treated for diabetes, hyperlipidemia, and hypertension. I believe individuals with schizophrenia deserve better than that.

The first reported CATIE finding was that four tested atypical antipsychotics were similar in effectiveness to perphenazine, as measured by the discontinuation rate. How can this be? My experience—and probably yours, too—is that atypical antipsychotics are more effective and better tolerated than the older ones. Dr. Henry Nasrallah, a CATIE investigator, offers some explanations.

The CATIE study is sponsored by the National Institute of Mental Health. As Dr. Nasrallah describes, CATIE differs from FDA-required efficacy trials in its longer duration (18 months), larger size (nearly 1,500 patients), and less-stringent inclusion/exclusion criteria. Just about anyone with a diagnosis of schizophrenia, regardless of comorbidities, was eligible so that CATIE would reflect "real world" schizophrenia. And in that regard it succeeded; the study’s patients are much like those you and I treat.

I’m worried that this mega-study showed no agent to be clearly superior, but payers can clearly demonstrate higher costs for the newer agents. A study as huge and expensive as CATIE probably will never be repeated, so we may never have a comparable study to challenge its results.

Will our patients be deprived of treatments that our clinical experience shows are best for them? Not if we can help it; there is a lot to be learned from CATIE, and we need to be informed about its results.

Next Article: