Why Did a Failed Trial Show a Treatment Signal Only in Women?

Progressive supranuclear palsy (PSP) progressively robs patients of balance, eye movement, swallowing, and cognitive abilities, and no effective disease-modifying drug is currently available. Now, a reanalysis of davunetide, an experimental drug that was declared a failure in a large trial more than a decade ago, has revealed an unexpected signal: its benefits may be concentrated among women.

The study, published in *Molecular Psychiatry*, reexamined data from the NCT01110720 trial. This Phase 2/3, randomized, double-blind trial enrolled 313 participants, who received 30 milligrams of intranasal davunetide or placebo twice daily for 52 consecutive weeks. Davunetide is an experimental peptide targeting pathology related to microtubules and the Tau protein; PSP is a neurodegenerative disease caused by the abnormal accumulation of Tau protein.

The original trial analyzed outcomes using the preregistered full PSP Rating Scale and an activities of daily living scale. None of the primary, secondary, or exploratory endpoints showed that davunetide was superior to placebo. The new study instead used the 10-item version of the PSP Rating Scale (PSPRS-10) currently preferred by the U.S. Food and Drug Administration and applied a mixed-effects model to analyze interactions among time, sex, and treatment simultaneously; the statistical result for the three-way interaction was P = 0.0005.

When assessed separately, women in the treatment group deteriorated more slowly on measures including balance, fine limb movements, and daily functioning, while no corresponding signal was observed in men. Women also showed significant differences from the placebo group in language, letter-number sequencing, and overall RBANS cognitive scores. A small amount of cerebrospinal fluid data further suggested that the direction of correlations between Tau markers and some clinical measures may be opposite depending on sex; however, this portion of the analysis included only 24 participants and is better suited to generating mechanistic hypotheses than to proving causality.

The most important limitation of the finding is precisely that this was not a new trial conducted exclusively in women. Sex stratification, the PSPRS-10, and the related statistical tests were not prespecified as primary efficacy analyses more than a decade ago. The study also examined multiple clinical and cognitive measures and may therefore have been affected by multiple comparisons, missing data, and chance differences between subgroups. The paper’s authors likewise characterized it as a post hoc analysis and argued that the next step should be confirmation in a trial with prespecified sex stratification.

Safety cannot be overlooked either. The original trial reported 54 serious adverse events in each group, with 11 deaths in the davunetide group and 10 in the placebo group; investigators at the time did not consider the difference in deaths to be caused by the drug. Nosebleeds, runny nose, and nasal discomfort were more common in the treatment group. The new analysis also noted that 21 participants in the davunetide group and 12 in the placebo group discontinued treatment because of adverse events, mostly related to nosebleeds or nasal congestion. In addition, the davunetide patent is currently licensed to ExoNavis Therapeutics, and two authors hold positions with or serve as consultants to the company. These data are sufficient to support a prospective confirmatory trial, but they remain insufficient to regard davunetide as a proven treatment for women with PSP.