Antipsychotic Augmentation in PTSD After a Large Negative VA Trial
A large VA trial found risperidone augmentation no better than placebo for PTSD symptoms not controlled by an SSRI alone. Real-world augmentation with second-generation antipsychotics continues anyway, often for reasons the original trial was never designed to test.
K.W. is a 39-year-old man, a former infantry sergeant now working in warehouse logistics, who has been on sertraline 150 mg daily for eight months following two deployments, with meaningful improvement in his mood and general anxiety but persistent, near-nightly hyperarousal and irritability that he describes as "still living like something's about to happen." He has no cardiac or metabolic history, is a nonsmoker, and has gained no weight over the past year, all directly relevant given the medication being discussed.
His prescriber raises risperidone augmentation for the residual hyperarousal, a common real-world practice despite a large VA-sponsored trial — published in JAMA in 2011 — that found risperidone added to an antidepressant produced no significant benefit over placebo on total PTSD severity, nor on remission rates or any of its other main measures. That trial is genuinely difficult to argue around: it was well-powered, rigorously conducted, and directly on point. What it does not settle, and what keeps augmentation alive in practice anyway, is that its primary endpoint was a total severity scale rather than the narrower, largely sleep-and-arousal-driven symptom pattern many prescribers are actually reaching for these drugs to treat. A post hoc subscale analysis did find small improvements in the hyperarousal and re-experiencing clusters — but the investigators themselves judged the effect too small to be clinically detectable, which makes it thin ground to build a prescription on — and quetiapine, the more commonly used agent in real-world augmentation, was not the drug tested at all. Whether that distinction is a genuine basis for departing from the trial's conclusion, or a convenient reason to keep prescribing something the best available evidence says does not work, is the actual disagreement.
At the follow-up visit
The 2011 VA trial is one of the more rigorous negative results in this whole literature, and it tested exactly this scenario — risperidone added to an antidepressant for PTSD not fully controlled by the antidepressant alone. Reaching for it here means going against a well-powered, directly applicable negative trial, and I think that needs to be said plainly before anything else. I'd also head off the obvious rejoinder: yes, the post hoc subscale analysis nudged on hyperarousal, but the authors were explicit that the size of it was below what anyone would notice in clinic.
I don't disagree with the trial's finding on total PTSD severity, but that scale is not the same as what K.W. is actually describing — a specific, sleep- and arousal-driven residual symptom cluster, not a global relapse. The trial was never designed to isolate that narrower question, and quetiapine, which is what most clinicians reach for in practice, was not even the drug studied.
That is a real distinction, not just a convenient one — but I want to be honest that it is also exactly the kind of reasoning that lets a negative trial get quietly worked around without ever being tested itself.
Given his clean metabolic profile, a time-limited trial of low-dose quetiapine specifically targeted at sleep and arousal, with an explicit stop-or-continue decision point at six weeks based on that narrow symptom cluster rather than global improvement, is a reasonable way to test the distinction honestly rather than either dismissing it or accepting it uncritically.
A six-week trial of low-dose quetiapine started specifically for residual hyperarousal and sleep disruption, with sertraline continued unchanged and an explicit follow-up scheduled to decide whether to continue.
Not resolved, and named as such: whether targeting a narrower symptom cluster with a different agent than the one tested genuinely escapes the 2011 trial's negative finding, or simply reframes the same practice the trial argued against. The six-week checkpoint was set specifically so the answer would come from K.W.'s actual response, not from either side winning the argument today.