Clinical Psychopharmacology · Paper A
CATIE trial
CATIE was a large, real-world trial that challenged the assumption that newer antipsychotics are automatically better than older ones. A pragmatic trial like CATIE follows patients under everyday clinical conditions rather than the tightly controlled conditions of an efficacy trial. One arm used perphenazine, an older typical antipsychotic, as a direct comparator against several newer atypical drugs, e.g. olanzapine, quetiapine, and risperidone. Clinically, questions citing CATIE usually test whether a candidate wrongly assumes atypical antipsychotics are always superior.
CATIE, run in 2005, was a large pragmatic effectiveness trial in chronic schizophreniaA long-term mental illness involving episodes of psychosis such as hallucinations and delusions.Read more → (a severe, long-term mental illness causing loss of contact with reality). Its primary outcome was all-cause discontinuation, i.e. whether patients stopped their assigned drug for any reason, capturing both poor efficacy and poor tolerability in a single measure. Olanzapine had the lowest discontinuation rate, suggesting the best real-world effectiveness, but it also caused the greatest weight gain and metabolic problems such as high blood sugar and high cholesterol. Perphenazine, a mid-potencyHow small a dose of a drug is needed to produce a given effect.Read more → older 'typical' antipsychotic used as the comparator arm, performed comparably to several newer 'atypical' drugs, e.g. patients on perphenazine stayed on treatment about as long as those on quetiapine or risperidone. Overall discontinuation was very high, around 74%, highlighting real-world tolerability problems across the board regardless of drug class. At baseline, before any trial drug was given, roughly 41 to 43% of participants already met criteria for metabolic syndrome, a cluster of cardiovascular risk factors including abdominal obesity, raised blood pressure, and abnormal glucose or lipid levels. This baseline burden was separate from any weight gain caused by the drugs themselves, showing how much cardiometabolic risk this population already carried before treatment even began.
CATIE (Clinical Antipsychotic Trials of Intervention Effectiveness) was a large, US government-funded trial published in 2005, and it remains one of the most cited studies in psychiatry. It matters not because of one dramatic result but because it directly tested an assumption that shaped a decade of prescribing: that newer, more expensive antipsychotics were simply better than older ones. A fresh graduate who understands what CATIE actually showed can push back on that assumption with real evidence rather than habit.
If you are coming to this fresh, the first thing to understand is the difference between an efficacy trial (a tightly controlled study that tests whether a drug works under ideal, closely supervised conditions, usually excluding patients with other illnesses or substance use) and a pragmatic trial (a study designed to mimic real-world prescribing, keeping patients in the trial even when they have the messy, overlapping problems seen in everyday clinics). CATIE was pragmatic. It enrolled around 1,460 people with chronic (a severe, long-term psychotic illness marked by , , and disorganised thinking that causes loss of contact with reality), confirmed against standard diagnostic criteria, and followed them for up to 18 months.
The central teaching point of CATIE, and the trap most exam questions are built around, is that the drug patients stayed on longest was not the drug with the best side-effect profile. It is tempting to assume that "stayed on it longest" automatically means "safest" or "best tolerated." CATIE shows precisely why that assumption fails.
By the early 2000s, newer "atypical" antipsychotics (a class including olanzapine, quetiapine, risperidone, and ziprasidone, so named because they were assumed to cause fewer movement side effects than older drugs) had become far more expensive than older "typical" antipsychotics, yet were widely assumed to be simply better. CATIE was built to test that assumption directly. It randomised patients to one of several atypicals or to perphenazine, an older, mid- typical antipsychotic used deliberately as the comparator arm. The primary outcome was all-cause discontinuation, meaning the trial simply measured how long a patient stayed on their assigned drug before stopping it for any reason at all, whether that reason was the drug not working, intolerable side effects, or the patient's own choice. This outcome is clever precisely because it bundles efficacy and tolerability into one practical, real-world number: a drug that works brilliantly but that nobody can tolerate for long is not, in real life, a good drug.
Consider the trial's own numbers. Overall discontinuation across the whole trial was strikingly high, around 74%, showing that keeping patients on any antipsychotic for the long term is genuinely difficult regardless of which drug is chosen. Olanzapine had the lowest discontinuation rate of the drugs tested, meaning patients stayed on it longer than on the others, on the surface the best-performing drug. But olanzapine also caused the greatest weight gain and the worst metabolic side effects, for example raised blood glucose and raised cholesterol, of any drug in the trial. Separately, perphenazine performed comparably to several of the newer atypicals on discontinuation: patients on perphenazine stayed in treatment about as long as those on quetiapine or risperidone, directly challenging the idea that atypical antipsychotics are always superior to typicals. This does not mean typicals are free of problems; perphenazine still carries a higher risk of extrapyramidal side effects (movement problems such as stiffness, tremor, and restlessness caused by dopamine blockade) than most atypicals.
CATIE is often muddled with two adjacent but distinct concepts. It is not a trial about first-episode psychosis; its population was chronic, established , so its findings about tolerability over 18 months should not be extrapolated uncritically to a patient's very first antipsychotic choice, where different trials and different risk-benefit balances apply. It is also not evidence that typical antipsychotics are generally preferable, only that perphenazine specifically was not clearly worse than several atypicals on this one composite outcome; that is a narrower claim than "typicals are just as good as atypicals across the board," and a stem that over-generalises perphenazine's result to the whole typical class is testing whether you notice that overreach.
Entry into CATIE required a confirmed diagnosis of against standard diagnostic criteria, meaning the ICD or DSM criteria used to establish that a patient's presentation genuinely meets the threshold for the illness rather than a shorter-lived psychotic episode. This mattered for interpreting the results: because the trial population reflected real chronic, established disease rather than a first presentation, its conclusions about long-term tolerability and discontinuation are most directly applicable to patients with an already-confirmed, longstanding diagnosis.
Clinically, CATIE's lesson is that antipsychotic choice should be individualised rather than assumed from class membership alone. A patient who is already overweight or has a family history of diabetes may do worse long-term on olanzapine despite its low discontinuation rate, because the metabolic cost accumulates over years. Conversely, a patient sensitive to movement side effects may tolerate an atypical better than perphenazine even if headline discontinuation numbers look similar. In practice, prescribers weigh efficacy, metabolic risk, and movement-disorder risk together for each patient, guided by bodies such as NICE (the National Institute for Health and Care Excellence, which publishes UK clinical guidelines) and the Maudsley Prescribing Guidelines, rather than reaching automatically for whichever drug is newest or most expensive.
A well-written distractor will offer "olanzapine is the best choice because patients stay on it longest" as a tempting but incomplete answer, ignoring the metabolic cost that CATIE itself flagged as a major finding. The pearl worth carrying into the exam is that a single outcome measure, like discontinuation, never captures the whole clinical picture on its own; you must weigh it against the specific harm profile behind it.
CATIE showed that keeping a patient on treatment longest (olanzapine) is not the same as keeping them safest (worst metabolic burden), and that an older typical (perphenazine) can hold its own against newer atypicals, so antipsychotic choice must be individualised rather than assumed from class or price.