Abstract
Background: Recently, there has been substantial interest in establishing methods for identifying a disease modifying therapy for Alzheimer's disease. The concept of Disease Modification has been approached in many different ways including cross-over study designs (randomized withdrawal and staggered start), use of biomarkers or imaging outcomes, measuring clinically relevant milestones, comparing slopes of decline and comparing adjusted slopes (Natural History Staggered Start analysis - NHSS). These methods are able to distinguish between different patterns of clinical response. Methods: A definition of Disease Modification is proposed that defines a substantially different pattern of clinical response than that of a symptomatic therapy. Additional definitions are proposed to separate out long term and short term symptomatic responses, and permanent and temporary disease modification effects. Clinical approaches are discussed in the context of supporting these distinct types of mechanisms. Results: The Randomized Withdrawal and Staggered Start designs can separate between disease modifying effects and symptomatic effects. The Natural History Staggered Start analysis can also make this distinction without the difficulties of a two-phase study such as: the potential bias and loss of power associated with a high dropout rate, ethical concerns related to removing a potentially efficacious treatment and the long treatment duration required for these two phase studies. Unadjusted slopes analysis and use of clinically relevant milestones can identify long lasting effects without distinguishing between disease modification and symptomatic effects. Use of biomarkers and imaging outcomes can detect effects on the underlying disease process, but will need additional validation in order to demonstrate that these effects are due to disease modification. Conclusions: The pathological process of Alzheimer's disease is not understood well enough to be directly measurable with biomarkers or imaging outcomes, and is not tied to clinical outcomes except through the symptomatology. In this setting, demonstration of disease modification should rely on clinical evidence with support from several sources including: pre-clinical data supporting a disease modifying mechanism of action, biomarkers connecting that mechanism to the clinical efficacy, and imaging outcomes supporting a structural change consistent with modification of the underlying disease.