did riluzole beat placebo for treatment-resistant ocd in children? no, and the numbers mostly hold
“The Y-box and the Real ESAW group decreased from 27 to a little under 22.”
a guest, a professor of psychiatry at the University of Pittsburgh said on Psychopharmacology and Psychiatry Updates, 2026-10-14, at 07:52. hear it in context → the speaker cited: a double-blind randomized controlled trial in children.
the episode: Glutamatergic Augmentation in OCD (at 07:52). we link the show's own page; we do not re-host audio.
first, the part that holds
the trial is real and the core of the claim holds: a 12-week, double-blind, placebo-controlled trial of add-on riluzole in 60 treatment-resistant children and teens with OCD found that both groups improved and there was no significant difference between riluzole and placebo. the riluzole group's endpoint score of a little under 22 matches the registered result (21.72).
what we found when we went and looked
the trial was published in Neuropsychopharmacology in 2014. it randomized 60 treatment-resistant young people (mean age 14.5) with a mean baseline CY-BOCS of 28.2. the abstract says all subjects improved but there was "no significant difference between placebo and riluzole on any of the primary or secondary outcome measures." the results posted to the trial registry give week-12 CY-BOCS means of 21.72 for riluzole (n=22) and 23.30 for placebo (n=29). so the riluzole endpoint matches, but the placebo endpoint the speaker gave (22.5) does not match the registry's 23.30. we could not open the full text, and neither the abstract nor the registry gives a baseline for each arm, so we could not check the 27 and 29 starting points.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| the study was a double-blind randomized controlled trial in treatment-resistant children with OCD | confirmed | the abstract describes a 12-week, double-blind, placebo-controlled trial. it randomized 60 treatment-resistant children and adolescents (mean age 14.5±2.4) with moderate to severe OCD to riluzole (final dose 100 mg/day) or placebo, added on to their existing treatment. [1][2] |
| on riluzole, Y-BOCS fell from about 27 to just under 22 | unverifiable | the endpoint checks out: the registry reports a week-12 CY-BOCS of 21.72 ± 6.43 for riluzole among 22 completers. the starting point does not: we found no baseline score for each arm in anything we read. the abstract gives only the whole-sample baseline of 28.2±3.7, and the publisher does not release the full text in PMC's XML feed. [1][3] |
| on placebo, Y-BOCS fell from about 29 to 22.5 | unverifiable | the registry reports a week-12 CY-BOCS of 23.30 ± 4.64 for placebo among 29 completers, not 22.5. the paper may report an adjusted or model-based figure that we could not read, so we cannot call the 22.5 wrong, but the posted observed mean does not match it. we did not find a baseline of 29 for the placebo arm. |
| there was no significant difference between riluzole and placebo | confirmed | the abstract says "there was no significant difference between placebo and riluzole on any of the primary or secondary outcome measures" and that the study "failed to demonstrate superiority of riluzole over placebo." the registry also reports no significant difference at 12 weeks. [1][2] |
| context: who ran the trial and the safety signal | confirmed | all seven authors are listed at the NIMH pediatrics and developmental neuroscience branch, not at the University of Pittsburgh. the speaker described the trial and did not claim to have run it. the abstract reports riluzole was "fairly well tolerated" but notes one case of pancreatitis and five slight rises in transaminases (liver enzymes). 17 of the 60 participants also had autism spectrum disorder. [1][2] |
both directions
for the speaker: the speaker got the design right (double-blind, placebo-controlled, children, treatment-resistant) and the result right (both arms improved, no significant difference). the riluzole endpoint of "a little under 22" matches the registry's 21.72. the whole-sample baseline of 28.2 fits arm baselines of roughly 27 and 29 if the arms were close to equal in size; the registry says 30 started on each arm.
against: the registry's observed placebo mean at week 12 is 23.30, not 22.5. that makes the placebo group's drop look somewhat larger in the speaker's account than in the posted data, though the conclusion of no difference stands either way. the per-arm baselines of 27 and 29 do not appear in anything we could read. the endpoint means also come from completers only (22 of 30 on riluzole, 29 of 30 on placebo), so more people dropped out of the riluzole arm. simple before-and-after numbers can hide that.
what this cannot say
we could read the abstract, the publisher's abstract page and the registry's posted results, but not the full paper. PMC served a captcha, PMC's text-mining feed was rate-limited, and the publisher does not allow full-text XML download. the paper's tables may report per-arm baselines and adjusted endpoints that would confirm the 27, 29 and 22.5, and we cannot rule that out. the registry is also not on our usual list of hosts; its numbers are the sponsor's own posted results. one 60-person add-on trial cannot say whether riluzole helps some subgroup, or how it compares with other augmentation strategies.
sources, each one fetched and read
- NIMH pediatrics and developmental neuroscience branch. 12-week, placebo-controlled trial of add-on riluzole in the treatment of childhood-onset obsessive-compulsive disorder. Neuropsychopharmacology. 2014;39(6):1453-9. PMID 24356715, doi:10.1038/npp.2013.343 (PubMed record) What we read there: 60 treatment-resistant children (mean age 14.5) with moderate to severe OCD were randomized to riluzole 100 mg/day or placebo as an add-on; "there was no significant difference between placebo and riluzole on any of the primary or secondary outcome measures." Fetched 2026-10-16 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/24356715/
- Neuropsychopharmacology, publisher article page, doi:10.1038/npp.2013.343 What we read there: mean CY-BOCS at baseline 28.2±3.7; "all subjects showed significant reductions in CY-BOCS scores during treatment"; riluzole was "fairly well tolerated, although it was associated with one case of pancreatitis and five instances of slight increases in transaminases." Fetched 2026-10-16 (HTTP 200). https://www.nature.com/articles/npp2013343?error=cookies_not_supported&code=0cfce5e8-17a0-4882-b941-5c8644fa615e
- PubMed Central record PMC3988548 for the same trial (abstract only) What we read there: "The publisher of this article does not allow downloading of the full text in XML form"; the abstract repeats the baseline CY-BOCS of 28.2±3.7 and the finding of no significant difference. Fetched 2026-10-16 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC3988548/
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam