The Phase III HARBOR study of delpacibart etedesiran (del-desiran) in myotonic dystrophy type 1 (DM1) has failed its primary endpoint, dealing a major setback to one of the most advanced programs in a disease that still has no approved disease-modifying therapy. The trial did not demonstrate statistically significant improvement versus placebo on video hand opening time (vHOT), a novel machine learning-derived measure of hand myotonia. Novartis said evidence of clinical activity was observed in secondary endpoints and exploratory analyses, and that safety findings were consistent with prior data, but provided no numerical results. The company said it will evaluate the full dataset and engage health authorities to determine the most appropriate development path.
Del-desiran is an antibody oligonucleotide conjugate (AOC) that uses a transferrin receptor 1 (TfR1)-targeting monoclonal antibody to deliver a small interfering RNA (siRNA) into muscle cells, where it degrades toxic DMPK messenger RNA — the product of the CTG repeat expansion that drives DM1. The mechanism had generated substantial clinical optimism: Phase I/II MARINA trial results published in the New England Journal of Medicine in February 2026 showed approximately 40% mean reduction in DMPK mRNA and improvements across functional measures including vHOT, muscle strength, and activities of daily living. Long-term MARINA open-label extension data had also shown apparent reversal of disease progression relative to natural history comparators, which formed the basis for regulatory agreement on the HARBOR endpoints.
HARBOR enrolled approximately 150 adults and adolescents with DM1 across roughly 40 global sites, randomizing participants 1:1 to del-desiran or placebo every eight weeks over 54 weeks. The FDA granted del-desiran Breakthrough Therapy designation in May 2024, adding to previously granted Orphan Drug and Fast Track designations. Enrollment completed in July 2025, with a readout originally anticipated in Q2 2026 before slipping to H2 2026.
The failure raises questions about vHOT as a primary endpoint. The measure — which uses machine learning to quantify hand opening speed — was novel and agreed upon with regulators based on its performance in MARINA.