Duke's 180-Child Autism Stem Cell Trial: The Primary Result Was Null, But 'No Benefits' Isn't the Whole Story

The claim holds that a Duke University clinical trial on 180 children found stem cell treatments for autism produced only insignificant benefits. The verdict is partially false. The primary result is accurately described, but the blanket phrase 'insignificant benefits' misrepresents what the study actually reported, stripping out a secondary finding the researchers themselves flagged as potentially meaningful.

The trial itself is real and well-documented. According to ClinicalTrials.gov registration NCT02847182 and the published results in Nature Medicine (Kurtzberg et al., 2020), Duke enrolled 180 children aged 2–7 with autism spectrum disorder in a randomized, double-blind, placebo-controlled Phase II trial. Children received either a single infusion of their own banked cord blood or a placebo. The primary endpoint — change in Vineland Adaptive Behavior Scales-3 composite score at 12 months — showed no statistically significant difference between the two groups (p=0.36). Duke Health's own June 2020 press release stated plainly: 'A large clinical trial found that infusions of a child's own cord blood did not significantly improve autism symptoms overall.' On the primary question, the claim is correct.

Here is where the claim breaks down. The same Nature Medicine paper reported that a subgroup of higher-functioning children — those with a baseline Vineland score of 70 or above — showed a statistically significant improvement on the Vineland-3 socialization subscale (p=0.009). The authors flagged this as an exploratory finding requiring replication, but it is a real result from the same 180-child dataset. Saying the trial found 'insignificant benefits' across the board erases this entirely. The distinction matters: it is the difference between 'this treatment does nothing' and 'this treatment may help a specific subgroup, and we need more research to know.'

The strongest version of the claim draws on a real pattern: an earlier Duke open-label study of just 25 children (Dawson et al., Stem Cell Reports, 2017) had reported roughly 10-point improvements in Vineland scores at 6 and 12 months. That study, however, had no placebo arm — meaning there was no way to separate genuine treatment effects from natural development, parental attention, or regression to the mean. The Phase II trial was designed precisely to test whether those promising signals held up under controlled conditions. For the primary outcome across all children, they did not. Conceding that point is important: the broader claim that cord blood infusions are not a proven general treatment for autism is well-supported by the evidence.

The manipulation pattern here is selective summarizing. Taking a null primary result and presenting it as the complete story of a trial — when the same paper contains a significant secondary finding — is a form of cherry-picking in reverse. It is not fabrication; it is omission. Whether the motive is to dismiss stem cell research entirely or simply to simplify a complex paper into a shareable headline, the effect is the same: readers get a false sense of certainty about what the science settled.

When evaluating clinical trial claims, always ask two questions: Was this the primary endpoint or a secondary one? And was the result for all participants or a subgroup? A null primary result and a significant secondary subgroup result can coexist in the same paper — and both deserve to be reported accurately.