Assessing safety and efficacy in subpopulations in Alzheimer's disease clinical trials: contextualizing representativeness.

Efficacy outcomes in clinical trials are based on well-powered analyes of the entire participating population. Trial populations will comprise many types of demographic and biological subgroups, including individuals of different sexes, groups of older and younger individuals, participants with or without the apolipoprotein E ε4 (APOE) genotype, ethnoracial groups, participants from urban versus rural communities, participants with lower and higher educational levels, or individuals who have or have not undergone previous therapies such as anti-amyloid monoclonal antibodies (MABs). Each subgroups is underpowered to draw definitive outcomes, and analyses can lead to inaccurate conclusions. Disciplined subgroup analysis can be hypothesis generating and can help guide drug development decision-making. The risks associated with subgroup analysis can be mitigated by using standard terminology, prespecifying outcomes of interest, stratifying randomization, conducting interaction analyses to identify confounds, and limiting the number of subgroup comparisons. Alternative efficacy and safety analyses such as the interaction test and non-inferiority analyses may yield important insights. Together, these design and analytic straegies may allow trialists to avoid spurious interpretations and derive more informative conclusions regarding the impact of therapy on subgroups in Alzheimer's disease (AD) clinical trials. Greater understanding of safety and efficacy in the subgroups participating in trials is crtically important for indicating what conclusions can be generalized if the candidate therapy is approved. Highlights: Clinical trials are sized to allow well-powered conclusions based on analysis of the entire participating population.Trial populations geared to be representative of the subgroups of the older population with AD are underpowered to allow drawing confident conclusions about efficacy or safety in subgroups.Strategies such as non-inferiority analysis combined with transparent reporting of the analytic framework may facilitate understanding treatment efficacy and safety in subgroups.