CATIE’s uncontrolled factors
CATIE study Phase 1 (Current Psychiatry, February 2006) is controversial for good reason. Critical details are problematic, such as the relatively low dosage of perphenazine that might have artificially reduced tardive dyskinesia incidence among patients taking that drug.1
A more fundamental limitation is the investigators’ use of “all-cause discontinuation” as the primary effectiveness measure. By contrast, an objective measure (Hamilton Rating Scale for Depression) was used in the STAR-D study to judge antidepressant effectiveness.
All-cause discontinuation measures are highly subjective, as patients stop taking medication for a variety of reasons:
- a spouse, parent, or friend pressured them into stopping
- they think the medication is making them weak, dependent, or vulnerable
- they don’t notice the drug’s therapeutic effects, or lack the mental skill to balance adverse and good effects
- or they are incapable of understanding that they might need to endure adverse effects to obtain a benefit.
These and other factors vary widely among patients and—in their decision process—could outweigh a medication’s objective effects on specific symptoms. These subjective factors probably obscured any objective differences among the five drugs studied in CATIE phase 1.
Other variables were uncontrolled, including:
- effects of other medications taken before and during the study. The authors state that 72% of subjects were taking other medications at base-line. Previous antipsychotics were washed out, but patients could remain on other medications. Did the investigators consider the effects of combining the study antipsychotics with these medications?
Even the washout periods seem to have been flexible (ie, not controlled) based on the study’s wording: “Overlap in the administration of (antipsychotics) that patients received before study entry was permitted for the first 4 weeks after randomization.”2 In other words, some patients stayed on their previous antipsychotics for 4 weeks, and others stopped taking the previous medication sooner. Do we know enough about variations in drug metabolism and effect duration to be sure that the overlap variable did not affect the results?
- nonpharmacologic treatment. According to the study, “No care was mandated across all sites other than the drug study.”3
If other types of treatment—such as group or individual therapy—were uncontrolled across all sites, did some patients receive such care whereas others did not? If so, were the potential effects of these treatments figured into the findings?
Arthur Mode, MD
Falls Church, VA
1. Nasrallah HA. CATIE’s surprises. Current Psychiatry 2006;5(2):48-65.
2. Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness of antipsychotic drugs in patients with chronic schizophrenia. N Engl J Med 2005;353:1209-23.
3. Conley RR. The CATIE outcome findings are relevant to clinical practice (CON). J Psychotic Disorders: Reviews & Commentaries 2006;9:3-16.