Alzheimer's disease (AD) disproportionately burdens racial and ethnic minority populations, yet the extent to which clinical trials reflect this burden remains poorly understood. We review all US-based Phase III AD drug trials (1997-2023), using a multi-source approach that integrates the Trialtrove clinical trial database with PubMed, ClinicalTrials.gov, pharmaceutical reports, and conference abstracts. We document three main findings. First, nearly half of published trials reported no data on patient race or ethnicity. Among trials that did report, practices were highly inconsistent in terminology, categorization, and analytical depth. Second, White patients constituted shares of enrollment that are grossly disproportionate to AD prevalence rates. Third, almost no trials conducted any subgroup analyses by race or ethnicity. Critically, there is no evidence of improvement in reporting or representation. These patterns limit the generalizability of AD treatment evidence. Our findings support strengthening mandatory reporting standards, broadening eligibility criteria, and diversifying trial site selection to ensure emerging AD treatments are evaluated equitably across the populations most affected.