This interview took place on April 9, 2007, and was conducted by Norman Sussman, MD.


This interview is also available as an audio PsychCastTM at

Disclosure: Dr. Marder is consultant to Bristol-Myers Squibb, Merck, Otsuka, Pfizer, Roche, sanofi-aventis, and Solvay.


Dr. Marder is professor at the Semel Institute of Neuroscience and Human Behavior at the University of California, Los Angeles (UCLA); director of the Veterans Integrated Service Network 22 Mental Illness Research, Education, and Clinical Center at the Department of Veterans Affairs in Los Angeles; and director of the Section on Psychosis at the UCLA Semel Institute. Dr. Marder’s research has focused on the treatment of schizophrenia and the pharmacology of antipsychotics.



What is the Clinical Antipsychotic Trials of Intervention Effectiveness (CATIE) study and what lessons have we learned from it?

The CATIE study was a National Institute of Mental Health-sponsored trial that compared several antipsychotics. It is a complex study, with numerous stages. Phase I of the study, which is the phase most often quoted, compared numerous second-generation antipsychotics (SGAs), including risperidone, olanzapine, quetiapine, and ziprasidone, as well as the first-generation antipsychotic (FGA) perphenazine. There were other randomizations in later stages, when patients did not do well or needed their medications changed. The double-blind study was conducted at relatively typical clinical sites. Some of these sites were academic, a number were Veteran’s Administration hospitals, and many were community clinics. The idea behind CATIE was that the study should be easy to translate from research to the usual clinical sites.

As a result, outcome was defined in an unusual but interesting way. That is, it was not simply whether a patient improved on a medication, but rather whether the patient and his or her physicians thought that it was reasonable to continue use of that medication.


What were the major findings of the CATIE study?

When the CATIE study was first published,1 many were surprised that it showed that the FGA perphenazine exhibited similar effectiveness to the SGAs, suggesting that there was no huge advantage to prescribing the newer drugs.

There was a statistically significant advantage for olanzapine over the other drugs. This was interesting in light of the fact that olanzapine also had the most discontinuations for side effects. For example, although olanzapine was marginally better for patients staying on medications, patients treated with olanzapine gained substantial weight, approximately 2 pounds/month, and had substantial elevations in triglycerides and low-density lipoprotein cholesterol.

The CATIE study produced some interesting findings that confused many clinicians. For example, many clinicians had been persuaded that the SGAs were more effective and more easily tolerated than the FGAs. However, data from the CATIE study suggested that this was not clear and that the FGAs worked as well and were as tolerated as the SGAs.

Moreover, the most effective of the drugs, olanzapine, also had the worst side effects, clearly demonstrating how difficult it is to treat schizophrenia. If the most effective drug would also have the most mild side effects, schizophrenia or psychosis would be much easier to treat.


Did the CATIE study measure both the change in symptoms as well as the length of time patients stayed on treatment?

Yes, both were measured. The conducters of the study measured how long patients stayed on their particular medications until a decision was made to switch them to something else. In addition, there were no differences among the drugs over the course of the study in terms of the severity of positive and negative symptoms. The study did not suggest any large differences between the FGAs and SGAs, nor among each of the SGAs.


What were some limitations of the CATIE study?

One limitation was that most schizophrenia patients in the study, despite receiving an antipsychotic, were already severely ill and may not have been the typical patients who were being seen by primary care practitioners or even by many physicians who treat schizophrenia. These patients had a Positive and Negative Symptom Scale score similar to individuals who were acutely psychotic, and were relatively poor responders to antipsychotics. In addition, they had been ill and treated with antipsychotics for, on average, >14 years. Thus, one would predict that virtually all of these patients had taken at least one SGA and still had not found the appropriate medication. Patients who were currently taking antipsychotics were not included in the CATIE study.

A second limitation was lack of any good measure about how patients were feeling. Although SGAs, for pure efficacy, are not that different from FGAs, patients tend to like them better because they are more easily tolerated and cause fewer extrapyramidal symptoms (EPS). However, patients’ feelings on the matter were not measured as factors in the study.

A third limitation was that the study failed to accurately measure the side effect of tardive dyskinesia. There was insufficient time to measure the effect of tardive dyskinesia between the FGAs and SGAs because the time most patients stayed on each drug was too brief. For example, by 6 months, >50% of patients in the study had changed to another medication. It is a controversial fact that because SGAs cause fewer EPS they are also less likely to cause tardive dyskinesia. However, whether or not that advantage was there was not measured in the study.


Tardive dyskinesia was very prevalent in the past. Was this due to the mega doses of antipsychotics that were being used?

The issue of tardive dyskinesia is rather complex. Today, when I make rounds on an inpatient unit, I see relatively little tardive dyskinesia. Thus, something has changed over the years. Is it because we are using lower doses and treating more conservatively? Aside from this possibility of having used excessive doses in the past, one can also argue that SGAs cause less tardive dyskinesia than FGAs. The best evidence comes from studies of the elderly.2 For example, among elderly patients treated with high-potency FGAs, as many as 25% develop tardive dyskinesia after 1 year of treatment.

Research3 indicates that both quetiapine and risperidone are less likely to cause tardive dyskinesia. (Conversely, there is more evidence of EPS with perphenazine than with other drugs and more patients discontinued its use.)

I think this is still a controversial area. The best study is unlikely to be done. For the time being we are going to have to depend on the kind of evidence we have now.


Is there a difference between dosages of medications in the trials and those used in actual practice?

This is a real issue. When a drug is first marketed, clinicians look at clinical trials that are conducted in relatively homogeneous populations. Then, when clinicians try to use the drug, they need to use different dosages in order to properly treat their patients. In the CATIE study, patients on quetiapine were treated with substantially lower doses than are usually used in clinical practice. Many clinicians use quetiapine 600–800 mg, which is substantially more than was used in the CATIE study. The daily dose of olanzapine in the study was >20 mg, which is actually above what is printed on the label.
Oftentimes, the information from clinical trials is difficult to generalize for use in practice, where side effects tend to be more prominent and clinicians tend to use different doses.


Is there an emerging consensus on whether or not there is shared risk of diabetes and metabolic syndrome with all the atypical antipsychotics?

In the CATIE study, which has the best data comparing the metabolic risks on SGAs, there are very large differences among the drugs. Olanzapine causes substantial weight gain, elevations in hemoglobin A1C, and elevations in blood sugar. Ziprasidone and perphenazine are weight neutral. In fact, there was a second phase of the CATIE study in which patients who had gained weight during Phase I lost a substantial amount of weight when switched to ziprasidone. There are substantial differences in terms of weight, lipids, and glycemic control among the SGAs.

The Food and Drug Administration was concerned that the SGAs and other antipsychotics may have a direct effect in that they had the potential to decrease insulin resistance and actually cause diabetes. There are numerous case reports of patients who had no history of diabetes who suddenly had dangerously elevated blood-glucose levels.4

The FDA, being very conservative, put a warning on all of the antipsychotics. However, there are substantial differences among them, clozapine and olanzapine having the most risk of diabetes and metabolic syndrome. Probably aripiprazole (which was not studied in the CATIE study), ziprazadone, and perphenazine have substantially less risk. Risperidone and quetiapine fall somewhere in between.

What should a clinician prescribe to a patient who has never been treated before and for whom an antipsychotic is indicated?
For a patient who is anxious about drug treatment and about possible side effects, I would probably choose an SGA. The choice of drug is often driven by side effects. I would try to give the patient the lowest side-effect burden, based on what their worries are. If the patient is young and worried about weight gain, I would probably choose a drug that is relatively weight neutral. A high-potency FGA often causes EPS. Akathisia in particular occurs at such a high prevalence that almost every patient on a high-potency FGA will experience this side effect. Thus, most clinicians would continue to use an SGA.

If olanzapine, quetiapine, or risperidone is used, it is very important to monitor the patient for early signs of metabolic effects. The patient should be weighed frequently in order to look for early evidence of weight gain. I would recommend monitoring for an elevation in triglycerides, because that is an independent risk factor which could be a sign of insulin resistance.


What are the major differences between paliperidone and risperidone?

Paliperidone is actually the major metabolite of risperidone. From the studies that I have seen, it is clearly an effective antipsychotic. It is in an extended-release oral formulation, so it has relatively smooth kinetics. There may be fewer drug-drug interactions with it, because it is the metabolite. Regarding the side-effect profile and efficacy, I am not aware of any differences based on the limited data we have about the drug. It certainly elevates prolactin, which is a major concern with risperidone. It is an effective antipsychotic, but I cannot differentiate it from risperidone.


What are the clinical consequences of prolactin elevation, other than lactating?

Elevated prolactin can lead to menstrual irregularities and galactorrhea. It may play a role in sexual dysfunction in men such as ejaculatory and erectile problems. Some think that it leads to premature menopause in women and that it may have effects on bone density, perhaps leading to osteoporosis. Although that is not well substantiated, it is certainly a concern.


Are there any new antipsychotics in the pipeline?

Bifeprunox is a partial agonist, similar to aripiprazole. It does not appear to have metabolic side effects and it seems to be an effective antipsychotic. Asenapine is another effective antipsychotic with a good side-effect profile. These two drugs have particular advantages with regard to their side-effect profiles. However, I am not sure that studies will find them to be dramatically different than other drugs.


Is there any consensus on how these medications work?

It is clear that these drugs are all potent dopamine blockers. Even the partial agonists decrease dopamine. The antipsychotic potency of these drugs is closely related to their effect at dopamine (D)2 receptors, and it is quite clear that this is the mechanism for attenuating psychotic symptoms. There is not a single drug found to be effective for psychosis that does not have similar effects on dopamine, particularly D2. Why some drugs do not cause fewer EPS is less clear. Any differences in efficacy of these drugs, if any, is relatively small. Hopefully, new drugs will be coming out, with different mechanisms, that will have a different range of efficacy. There is a vigorous search for drugs to improve cognition in schizophrenia and to improve negative symptoms. The current drugs are an effective group but they have substantial limitations. PP



1. Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med. 2005;353(12):1209-1223.
2. Jeste DV, Okamoto A, Napolitano J, Kane JM, Martinez RA. Low incidence of persistent tardive dyskinesia in elderly patients with dementia treated with risperidone.  Am J Psychiatry. 2000;157(7):1150-1155.
3. Jeste DV. Tardive dyskinesia rates with atypical antipsychotics in older adults. J Clin Psychiatry. 2004;65(suppl 9):21-24.
4.  Reist C, Mintz J, Albers LJ, Jamal MM, Szabo S, Ozdemir V. Second-generation antipsychotic exposure and metabolic-related disorders in patients with schizophrenia: an observational pharmacoepidemiology study from 1988 to 2002. J Clin Psychopharmacol. 2007;27(1):46-51.