Avoiding EPS is key to realizing ‘atypical’ benefits
CATIE finding is not unique to one antipsychotic class.
Many findings of the Clinical Antipsychotic Trials of Intervention Effectiveness in schizophrenia (CATIE) were unexpected,1,2 but one was arguably the most surprising. It was that schizophrenia patients showed similar rates of extrapyramidal symptoms (EPS), whether treated with a first-generation antipsychotic (FGA) or any of four second-generation antipsychotics (SGAs).
This finding in CATIE phase 1 runs contrary to the understanding that SGAs, compared with FGAs, provide a broader spectrum of efficacy with significantly fewer motor side effects. A substantial body of evidence and virtually all schizophrenia treatment guidelines3-5 support this prevailing view.
Did earlier schizophrenia treatment studies misinform us, or was CATIE’s comparison of FGAs and SGAs “flawed”?6,7 This article attempts to reconcile the divergent findings about antipsychotics and EPS and reveals a clinical pearl that suggests how to provide optimum antipsychotic therapy to schizophrenia patients.
What did catie find?
CATIE was a three-phase, 18-month, randomized controlled clinical trial designed to evaluate the effectiveness of five SGAs (risperidone, olanzapine, quetiapine, ziprasidone, and clozapine) and two FGAs (perphenazine and fluphenazine) in treating schizophrenia. Findings from phases 1 and 2 have been published or presented (Table 1),2,8-9 and results from phase 3 are awaited.
Table 1
5 key findings from CATIE phases 1 and 2
|
CATIE phase 1 found no difference in efficacy, safety/tolerability, or effectiveness among perphenazine, risperidone, ziprasidone, and quetiapine. Soon-to-be-published data also will show no significant difference in cognitive effects among patients receiving perphenazine or any of four SGAs (risperidone, olanzapine, quetiapine, or ziprasidone).8 Because no FGA was used in CATIE phase 2,9-11 its results added little to phase 1 observations about how “typical” and “atypical” antipsychotics compare.
‘Atypicals’ and EPS. By definition, a reduced tendency to cause EPS (such as parkinsonism, dystonia, akathisia, and akinesia) distinguishes SGAs from FGAs. In fact, SGAs were called “atypical” because they disproved the belief that EPS are an unavoidable consequence of drugs that produce an antipsychotic effect.12,13 The CATIE trial’s inability to detect a difference in EPS rates between typical and atypical antipsychotics (Table 2)2 is therefore the study’s most surprising finding.
Table 2
CATIE: Similar EPS rates with perphenazine and SGAs*
| EPS measurement | Perphenazine-treated patients | SGA-treated patients |
|---|---|---|
| Increased mean Simpson-Angus Scale score | 6% | 4% to 8% |
| Increased AIMS global severity score | 17% | 13% to 16% |
| Increased Barnes Akathisia Rating Scale score | 7% | 5% to 9% |
| Anticholinergic added | 10% | 3% to 9% |
| * Differences were not statistically significant | ||
| EPS: extrapyramidal side effects | ||
| SGA: second-generation antipsychotic | ||
| AIMS: Abnormal Involuntary Movement Scale | ||
| Source: Reference 2 | ||
Making sense of catie
Most studies suggest consistent differences between FGAs and SGAs in risk of EPS and tardive dyskinesia.14-16 One explanation for CATIE’s discrepant findings may be that the use of high-dose, high-potency haloperidol as the typical comparator in pre-CATIE studies magnified differences between FGAs and SGAs.17,18
Conversely, CATIE researchers minimized this difference by studying a population of schizophrenia patients at an unusually low risk for EPS. The study design:
- assigned 231 patients with a history of tardive dyskinesia to an SGA, without the opportunity to be randomly assigned to an FGA
- excluded patients with first-episode schizophrenia
- enrolled patients who had been treated with antipsychotics for an average of 14 years without a history of significant adverse effects from study treatments.19
Just as prior studies might have exaggerated the EPS advantage for SGAs, CATIE might have minimized the FGA-SGA difference by studying a low-risk cohort in a way that reduced the trial’s ability to detect such differences.
Interpretation. How can we reconcile the absence of a difference between FGAs and SGAs in EPS liability in CATIE with the preponderance of data suggesting otherwise? It appears that SGAs may be less likely to cause EPS than FGAs, but this difference is not evident in all populations. Furthermore, SGAs and FGAs differ in their ability to provide an adequate antipsychotic effect without EPS.
Among FGAs, low-potency agents are less likely to cause EPS or require concomitant anticholinergics than high-potency agents. Among SGAs, the gradient of EPS liability appears to be risperidone > olanzapine, aripiprazole, ziprasidone > quetiapine > clozapine (Figure). Clinically, these pharmacologic differences interact with physiologic differences in EPS vulnerability—some patients are more liable to develop EPS than others. Individuals who are more susceptible to developing EPS are more likely to benefit from antipsychotics with lower EPS liability.
