Should CATIE dictate choice of antipsychotic in chronic schizophrenia?

In December 2005, the New England Journal of Medicine published the first results of the CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) (1).  The 18-month, double-blind study, funded by the US National Institute of Mental Health (NIMH), compared an older and much cheaper typical or first-generation antipsychotic, perphenazine (Trilafon), against several newer atypical or second-generation antipsychotic drugs - olanzapine (Zyprexa), quetiapine (Seroquel), risperidone (Risperdal) and ziprasidone (Geodon), among 1460 patients with established, but not refractory, schizophrenia (1). 

CATIE was meant to reveal the truth as to how effective the different drugs were in the real world of everyday clinical practice.  Effectiveness was judged by the length of time patients remained on treatment before discontinuing on account of a perceived lack of efficacy, intolerable side effects or for other reasons.

The newer atypical antipsychotics are preferred among psychiatrists and patients because they spare most patients from extra-pyramidal side effects and the risk of developing tardive dyskinesia.  Theoretically, this should  encourage them to remain on treatment.  Despite their additional expense, the atypical drugs now account for around 95 per cent of the market for antipsychotic drugs and the cost of prescribing them has soared.  The worldwide market for antipsychotic drugs is close to $16 billion of which $10 billion is spent in the US. 

However, the CATIE results suggested the older cheaper antipsychotic perphenazine was, on balance, just as effective and tolerable as the rest. Overall 74 per cent of patients discontinued their medication before the end of the study.  Of those taking perphenazine, 75 per cent discontinued –about the same percentage as the others, excluding olanzapine.  Of patients on olanzapine, only 64 per cent discontinued before the study end, but this treatment group experienced the most intolerable effects – namely an average weight-gain of 2 lbs per month - and saw the most unfavourable metabolic changes.   Thirty per cent of patients on olanzapine gained 7 per cent or more of baseline body weight. Among the other treatment groups, much smaller proportions of patients increased body weight more than 7 per cent from baseline during drug therapy,  the range being only 7 to 16 per cent.

Significantly more patients discontinued perphenazine on account of extra-pyramidal side effects (8 per cent vs 2-4 per cent in the other groups) but the drug, costing a fraction of the atypicals’ price, was considered “a dominant choice with regard to cost-effectiveness”.  All out-patient costs remained the same, regardless of drug type used and there was no significant difference between drug types in the number of in-patient days.

In a second phase of the study, where patients who discontinued their first drug could be randomised to another, the older drug clozapine proved significantly more effective than the atypicals (2).  However, clozapine was open-label and patients randomised to it saw their psychiatrist weekly for blood monitoring whilst patients on the other atypicals did not. The drug proved an unpopular option with few patients willing to accept the chance of being randomised to it.  This part of the study therefore had small numbers (2).

Causes of concern with design

CATIE has caused considerable disquiet among psychiatrists, many of whom believe its design favoured olanzapine, perphenazine, and clozapine leading to an unfair comparison, and who find the interpretation worrying. The major points of contention are:

Dosing
• The drug that appeared more effective in phase 1, olanzapine, was the only one dosed above the approved upper limit.  Mean modal daily doses used throughout phase 1 were:  20.1mg olanzapine, 20.8mg perphenazine, 543.4mg quetiapine, 3.9mg risperidone and 112.8mg ziprasidone. 

Selection bias
 • Patients who had shown a tendency to tardive dyskinesia (TD) were kept away from perphenazine and randomised to the other drugs. The dose of perphenazine was kept low.  Critics also say the follow-up analyses were over too short a period to evaluate the longer-tem consequences of using drugs likely to lead to metabolic syndrome or TD after several years’ use. 

At the American Psychiatric Association’s annual meeting in Toronto in May 2006, more results from CATIE were presented with regard to neurocognitive outcomes showing perphenazine performed well.  However the difference between the older drug and second generation antipsychotics was not statistically significant.  Critics suggest this was because the study prevented patients with a tendency toTD from receiving perphenazine and randomised them to the atypicals instead.  Patients with TD may differ neurocognitively from those without the condition so at baseline the distribution of patients with TD could possibly have influenced the result on this point
 
Formularies could restrict choice

The effect of CATIE has been to exacerbate fears in the psychiatric community that formulary committees will now use the study as an evidence-base to justify restricting choice of antipsychotic medication to older and cheaper first generation antipsychotics and to the atypical drug that has the strongest potential to increase metabolic disease.  There are fears that prescribing advisers might generalise from the results for perphenazine to recommend use of other first generation antipsychotics with worse EPS profiles. Outside of CATIE there could be a tendency for psychiatrists using typical antipsychotics to increase the dose in order to boost efficacy over time and that this would cause more EPS and TD. Several months since publication, critics say the study is still generating more heat than light and its findings continue to be hotly debated.

Lead investigator Professor Jeffrey Lieberman of Columbia University, New York has attempted to calm fears and cool feelings of indignation.
“The main message of the CATIE studies” he emphasises, “ is that treatments should be individualised.”    All antipsychotic drugs are generally effective, he insists, but the variation in their side effects can be substantial for individual patients.  “The results of CATIE speak for maximising choices.”

Professor Lieberman reiterated this message at the APA meeting.  He has gone on record before to deliver the same recommendation and to speak out against the media interpretation of CATIE’s results as a horse race in which there were winners and losers.  Lieberman’s view is one of horses for courses.  Choice of antipsychotic should be governed by efficacy and an individual patient’s profile, he says.“ For example, If a patient is overweight and has a family history of diabetes you wouldn’t prescribe a drug to control schizophrenia symptoms that is associated with the most weight gain and other metabolic side effects. “

Yet for all Professor Lieberman’s advice, the perception remains among psychiatrists that CATIE has unleashed a genie out of the bottle suggesting that older, typical or first generation antipsychotics offer the same benefits as newer drugs at a substantial discount.   In the original CATIE publication, the investigators wrote:  “How clinicians, patients, families and policymakers evaluate the trade-offs between efficacy and side effects, as well as drug prices, will determine future patterns of use.”  In other words, anything could happen.

At APA psychiatrists debated whether or not CATIE has really informed psychiatry to the extent that it might alter practice.  Arguing that it has, Professor Joseph McEvoy, Associate Professor of Psychiatry at Duke University, North Carolina, pointed out that the investigators consulted widely before finalising the study’s design and that there had been no dissent before the study.  The results provided much data to inform clinical practice, he maintained, telling clinicians which drugs offered greater efficacy, their respective tolerability and which had the best metabolic profiles.

Arguing against the usefulness of CATIE Professor Herbert Meltzer of Vanderbilt University, Tenessee, agreed much had been learned but said he remained sceptical. Just because CATIE was not sponsored by the pharmaceutical industry did not mean there was no conflict of interest, he asserted.   “The results challenge everything we have learned from evidence-based medicine.” Reviews and meta-analyses all suggest the atypical antipsychotics have the same efficacy, he noted.

The 74 per cent discontinuation rate was not consistent with clinical practice where patients often remain on their original prescribed drug for five years or more, he added ”Patients in CATIE were treatment-resistant patients, looking for something better who were told they could have a choice of drugs then switch if they wanted to.  This is not a prescription for remaining on a medication, ”  he pointed out.  “The study suggested olanzapine beat all comers but there was a fundamental flaw.  Olanzapine came out on top because it was given a privileged position.  For unclear reasons it was allowed to be used above the licensed dose.  So this was not a fair fight, “ he suggested. “Once you get the dose of an atypical up, these drugs work much better in treatment-resistant patients.”

The prescribing of perphenazine was also unrepresentative of its everyday use, he suggested, making it  compare well against the atypicals.  Not allowing patients with TD to get it was disingenuous, he suggested.  “We know there is a  2-3 per cent risk per year of extra-pyramidal symptoms with the typical antipsychotics.”  This study excluded the patients most likely to experience EPS from the drug most likely to produce them,hhe claimed.  “Older patients, in their 40s, are ripe for TD.  I think it would be a real throwback to consider using drugs like perphenazine in this population.”
. 
As for the finding regarding neurocognitive effects, he found this astonishing.  A meta-analysis of studies has shown they are in complete agreement that atypicals have a significant advantage over typicals in neurotransmission.  “If CATIE got this right it is a finding of enormous significance.  However, I don’t think it did.  We need a lot more research.”

Overall the study was disappointing and its findings dubious, he concluded, suggesting a fixed dose study needed to be done.   After CATIE, world-wide leaders in psychiatry met to decide if any action should be taken to review clinical guidelines on the management of schizophrenia. Had the study really informed clinical practice they would have done so. “They decided not to change anything,” he said.  Formulary committees should take note and heed Professor Lieberman’s and others’ advice.  Psychiatry needs to offer as many choices of antipsychotic therapy as it can.
References

1. Lieberman JA et al.  Effectiveness of antipsychotic drugs in patients with chronic schizophrenia.  New Engl J Med 2005; 353: 1209-1223.
2. McEvoy JP et al.  Effectiveness of clozapine versus olanzepine, quetiapine and risperidone in patients with chronic schizophrenia who did not respond to prior atypical antipsychotic treatment.  Am J Psychiat 2006; 163: 600-610.

Olwen Glynn Owen

To see more features by Olwen Glynn Owen follow this link to his first feature article and scroll to the end for a list:
Glivec improves long-term survival prospects for chronic myeloid leukaemia patients