Risperidone or Aripiprazole for Conduct Disorder in a Teenager of Average Intelligence
A 14-year-old with conduct disorder and no intellectual disability needs an antipsychotic his family can trust. Most of the field's evidence for this drug class was built in a different population — and the two agents on the table don't carry the same kind of evidence at all.
T.J., a 14-year-old boy, spends most weekends stripping down and rebuilding an old dirt bike in his family's garage, a project his father half-jokingly calls the only thing that keeps him in one place for more than twenty minutes. School has gone the other direction. He was suspended twice this term — once for a fight in the cafeteria that started over a spilled tray, once for slashing the tires of a classmate's bicycle in what he later described only as “he had it coming” — on top of a longer pattern, since about age 11, of skipping class, shoplifting small items, and rule-breaking that has outpaced ordinary adolescent testing of limits. A psychoeducational evaluation done for school placement purposes eight months ago put his full-scale IQ at 104, squarely average, with nothing on the profile suggesting an autism spectrum presentation or an intellectual disability underneath the behavior.
That evaluation matters more to today's decision than it looks like it should. Eight months of individual and family therapy have produced real gains in how he talks about his anger afterward, but measurably little change in whether the fights and property destruction still happen. His parents have read enough about antipsychotics to arrive already worried about weight gain specifically, and his baseline labs — fasting glucose and lipids both normal, no family history of diabetes — are reassuring on that front without settling the actual question in front of the team: which of the two agents usually discussed for this kind of aggression, risperidone or aripiprazole, actually has evidence behind it for a kid who looks like T.J.
Most of what's published under “antipsychotics for pediatric aggression” was built in children with subaverage IQ or autism spectrum disorder — large, well-powered, placebo-controlled trials that don't describe him. The literature specific to conduct disorder at average intelligence is thinner for both drugs, but it isn't thin in the same way for both of them. Risperidone has Findling et al., 2000 — small, twenty patients, but randomized, blinded and placebo-controlled, in outpatients with conduct disorder and average intellectual functioning. Aripiprazole has two open-label studies in the same diagnosis and, as of now, no randomized controlled trial in it at all. That asymmetry is a difference in what kind of claim each drug can support, not a difference in how much each has been studied, and it is what the team actually has to reason through rather than a coin flip between two similar options.
Outpatient follow-up, eight months into therapy
I'd start with risperidone, and the reason is specifically about T.J.'s own diagnosis, not just the drug's general reputation. Findling and colleagues ran a small but genuinely randomized, double-blind, placebo-controlled trial in conduct disorder — twenty kids, ten weeks, no restriction to subaverage IQ — and risperidone beat placebo on most measures of the RAAPP, the aggression scale that trial ran on. That's a real trial in T.J.'s actual population, not a study built in a different one and extrapolated over.
I hear the trial-design point, but I want to put real weight on something the family already told us: they read about antipsychotic weight gain before this visit and came in worried. Two open-label studies in conduct disorder specifically — Ercan and colleagues, twenty patients, and Findling's own separate pharmacokinetic study, twenty-three patients — both found real improvement on aggression with aripiprazole, and the class-wide metabolic data for aripiprazole is genuinely more favorable than for risperidone.
If we start with the drug more likely to confirm their fear on day one, we may not get a fair trial of anything — adherence in a family already primed to distrust the plan is its own real clinical variable, not a soft consideration next to the trial data.
I want to name something both of you are talking past: those two evidence bases aren't actually comparable in kind, and that gap matters more than either drug's individual selling point. Findling's risperidone trial was randomized and placebo-controlled. Both aripiprazole studies were open-label — no blinding, no placebo arm. Ercan and colleagues say so themselves in their own conclusion: that placebo-controlled, double-blind studies are still needed before the drug's role here can be defined. To this day there is no randomized controlled trial of aripiprazole in conduct disorder at all.
An open-label improvement in a small, unblinded sample carries a real risk of expectation effect that a placebo arm exists specifically to rule out. The metabolic argument for aripiprazole is real, but it answers a different question than whether the drug actually works here — it doesn't retire the evidence-quality gap, it just sits next to it.
Agreed: risperidone starts at a low dose, with a metabolic panel drawn today and repeated at four weeks, and therapy continues unchanged alongside it.
Not agreed, and the reason the plan carries a real branch point rather than a single expectation:
The plan continues as started, with a formal aggression-outcome reassessment at eight weeks.
The switch to aripiprazole moves from a reserved option to the active plan, with the psychiatrist and pharmacist agreeing in advance that a genuine metabolic signal outweighs the efficacy-evidence gap the pharmacologist raised today.