There are currently over two dozen agents targeting β-amyloid (Aβ) in human clinical trials. More than a dozen of these are forms of anti-amyloid immunotherapy. Although other anti-amyloid interventions are further along in the development process, thus far only immunotherapy has provided post-mortem evidence that it can alter elements of the underlying pathology of Alzheimer’s disease (AD) in actual patients.

In the past 30 years, there have been many attempts to develop treatments for AD. Early therapies were developed based on a limited understanding of the disease (Slide 1). Prior to the 1980s, a clear pathophysiologic mechanism for AD was not known; instead, symptomatic therapies targeted associated symptoms, such as agitation, insomnia, and psychosis. In the 1970s, several preclinical studies pointed toward synaptic transmission abnormalities, particularly neurochemical abnormalities, as the root cause of AD, and treatments with cholinesterase inhibitors grew out of that theory. Today, the cholinergic hypothesis has been largely discredited in the primary pathogenesis of AD. Another theory based on neurotransmitter abnormalities, the glutaminergic hypothesis, has also gone out of favor as a causal explanation for AD. This did not stop medications based on these mechanisms from finding a meaningful place in the clinical pharmacopeia for treatment of AD.

In the 1990s, many clinical trials followed up on epidemiologic studies suggesting systemic causes of AD. These clinical trials focused on anti-inflammatories, hormone replacement, and antioxidants. The trials performed have largely failed, with the possible exception of the trials of vitamin E, an antioxidant. None of these agents have proven useful as disease-modifying therapies for symptomatic AD.

The introduction of the amyloid hypothesis, which identifies amyloid as the primary target for treating Alzheimer’s disease (AD), was a watershed in the field of AD research. Treatment approaches have since focused on how amyloid might be manipulated. In fact, there are challenges to amyloid manipulation. First, removal of amyloid has shown minimal clinical effects. Second, some data indicate that secretase modulation shows hints of stimulating feedback systems, which cause concern over how to effectively manipulate amyloid. Third, side-effect profiles of the direct manipulation of amyloid are not currently optimal. Finally, evidence suggests that cell and synapse loss may be better markers of dementia and its severity. Alternative approaches to direct manipulation of amyloid include antioxidant mechanisms for prevention, anti-inflammatory mechanisms for neuroprotection, manipulation of metabolic risks, modification of tau accumulation, and neuroregeneration.

These alternative mechanisms of action provide support for a wide range of potential agents for clinical trials. The exact connection between mechanism of action and clinical outcomes is not well defined, but typically trial designs using these mechanisms attempt to measure affects on disease progression, often with longer studies that include established clinical outcomes accompanied by putative biomarkers of disease. One trend that has provided much information has been to examine the effects of these drugs in subjects at risk for AD, thereby expanding the intervention to prevention of disease. These studies include both secondary prevention studies, for example deferring onset in those with prodromal disease such as mild cognitive impairment (MCI), and primary prevention directed at those with no specific signs or symptoms.

As disease-modifying therapies near realization, there are concerns about the criteria by which these therapies will be judged. It is not yet clear what kind of evidence (clinical, biomarker, or otherwise) will be required to support a disease-modifying claim. When such a drug is approved, regulators must be certain that the claims in the label are factual and unambiguous, yet a definition for “disease modification” remains to be established. One strong potential definition is: “a therapy that affects the underlying pathology and structure of the disease”. However, this is only one possibility, and a consensus definition must be codified before criteria to evaluate it can be determined.

There is room for informed speculation, however. Criteria to evaluate disease-modifying effects have been proposed, and typically involve one of two approaches (neither of which has yet been endorsed by regulators). The first is a clinical approach, in which clinical designs are employed that would ideally force a conclusion that a drug has a disease-modifying effect. In one proposed design, patients would be randomized to drug or placebo for an appropriate duration. At the end of that period, and if a difference in outcome between drug and placebo on an appropriate clinical measure or measures has been achieved, patients originally randomized to drug would then be treated with placebo, while patients originally treated with placebo would remain on placebo.