Retinal thickness measured with optical coherence tomography has been proposed as a noninvasive biomarker for Alzheimer's disease (AD). We therefore measured retinal thickness in well-characterized AD and control participants, considering ophthalmological confounders. We included 57 amyloid-proven AD cases and 85 cognitively normal, amyloid-negative controls. All subjects underwent retinal thickness measurements with spectral domain optical coherence tomography and an ophthalmological assessment to exclude ocular disease. Retinal thickness did not discriminate cases from controls, including stratified analyses for early- versus late-onset AD. We found significant associations between macular thickness and global cortical atrophy β −0.358; P = .01 and parietal cortical atrophy on magnetic resonance imaging β −0.371; P < .01 in AD cases. In this study, representing the largest optical coherence tomography cohort with amyloid-proven AD cases, we show that retinal thickness does not discriminate AD from controls, despite evident changes on clinical, neuroimaging, and CSF measures, querying the use of retinal thickness measurements as an AD biomarker.