Avagacestat, an oral γ-secretase inhibitor that selectively inhibits amyloid-beta (Aβ) synthesis, was investigated in Phase 2 studies to assess safety in mild-to-moderate Alzheimer's disease (AD) and predementia AD (PDAD). We report interim data from the first prospective, multicenter, randomized, controlled trial in PDAD (CN156–018). CN156–018 assessed the safety, tolerability and biomarker effects of avagacestat in patients with PDAD. Lumbar puncture demonstrating CSF profile consistent with AD pathology was required at entry. Patients were initially randomized to receive placebo or avagacestat (originally targeting a maximum dose of 125 mg/day but amended to 50 mg/day due to intolerability) for a duration of ≥2 years. Interim data are presented herein. 263 patients were randomized, and treatment groups were well-matched on entry criteria. 173 patients were randomized prior to dose amendment. Discontinuations were more frequent with avagacestat than placebo (53% vs. 35%), but were comparable (33% vs. 32%) after dose reduction. Serious adverse events were more frequent with avagacestat than placebo (35% vs. 23%), primarily attributable to avagacestat-associated, increased rates of non-melanoma skin cancers. Treatment-emergent adverse events (TEAEs) more frequent with avagacestat than placebo included diarrhea, nausea, vomiting, rash, fatigue, decreased weight, decreased appetite, and dizziness; these TEAEs were either less frequent or less severe with avagacestat 50 mg/day. Magnetic resonance imaging (MRI) safety readings showed numerically greater rates of amyloid-related imaging abnormalities for avagacestat- versus placebo-treated groups. Laboratory findings suggested non-progressive renal tubular dysfunction of unknown clinical significance with avagacestat. No clinically meaningful differences in cognition were observed between groups. Investigator-reported progression to dementia was similar with avagacestat (24/131) and placebo (27/132) [HR=1.17, 95% CI 0.68–2.03]. Post-baseline CSF amyloid measurements suggested a 10–15% reduction in avagacestat- versus placebo-treated groups. Volumetric MRI at Week 104 showed slightly greater ventricular volume increase and whole-brain volume loss for avagacestat-treated patients, with no differences in hippocampal volume. Safety and tolerability of avagacestat in the PDAD population was similar to that observed in a previous mild-to-moderate AD study (CN156–013). In aggregate, the study was terminated early as there was insufficient target engagement or trends for favorable pharmacodynamic effects to support continued development.