Slowing the progression of Alzheimer disease: methodologic issues.

The evidence to support a claim that a new drug will slow the progression of Alzheimer disease (AD) must derive from epistemologically valid research methods. Although agency regulations do not specify the magnitude of an effect that a drug must possess to be granted a claim as a treatment for AD, the evidence to support any claim must be adduced in adequate and well-controlled clinical investigations and must meet the standard of "substantial evidence." Because a claim presented in drug product labeling may not be false or misleading in any particular, a distinction must be made between treatments that provide a "symptomatic" benefit and those that alter the course of dementia. Examples of some of the difficulties likely to be encountered by sponsors seeking to develop evidence to support a claim that a new drug slows the progression of dementia are presented. A suggestion is made for a clinical trial design, designated as the "randomized start design," that may be useful in such a question. Why this design might overcome many of the difficulties, both practical and ethical, present in the "discontinuation" design, the design ordinarily proposed to assess a drug's effect on disease progression, is discussed.