Redefining Alzheimer's: From Symptoms to Biomarkers
Source:vignettes/articles/history.Rmd
history.RmdMcKhann et al. (1984): The clinical turn
In a landmark consensus paper, McKhann et al. (1984) established the first widely adopted clinical diagnostic criteria for Alzheimer’s Disease (AD) under the joint auspices of the National Institute of Neurological and Communicative Disorders and Stroke (NINCDS)1 and the Alzheimer’s Disease and Related Disorders Association (ADRDA)2.
Briefly, these criteria introduced a three-tier classification system, reflecting increasing levels of diagnostic certainty:
Possible AD: cases with atypical features or the presence of comorbid conditions that could contribute to dementia, but where Alzheimer’s disease remains a likely diagnosis;
Probable AD: a clinical diagnosis based on insidious onset and progressive memory and cognitive decline, in the absence of alternative explanations3;
Definite AD: requires all the criteria for Probable AD, combined with histopathological confirmation of Alzheimer’s-specific brain lesions at autopsy or biopsy.
Although grounded in clinical observation, the framework acknowledged the absence of specific biomarkers and the high misclassification rate4 in clinical practice.
Blacker et al. (1994): Toward refining diagnostic accuracy
Blacker et al. (1994) conducted a multisite study to evaluate the reliability and validity of the NINCDS-ADRDA diagnostic criteria for Alzheimer’s disease (AD), originally proposed in (McKhann et al. 1984). Using autopsy-confirmed diagnoses as the gold standard, the study assessed how well clinicians across three academic centers could classify 60 cases (AD and non-AD) based on detailed clinical summaries.
The findings supported moderate inter-rater reliability and good post-consensus validity of the criteria, but also revealed important diagnostic limitations—especially in atypical presentations.
Key recommendations for refining the criteria:
Early behavioral symptoms (e.g. personality change, disinhibition) were often associated with non-AD pathologies, such as Pick’s disease or hippocampal sclerosis. The authors suggested that such cases should not be classified as probable or possible AD under future criteria.
Prominent early language dysfunction (e.g. progressive aphasia) also frequently corresponded to non-AD pathology. The authors proposed these cases be diagnosed as possible AD only, and only when otherwise typical.
The distinction between probable and possible AD showed little difference in predictive value, questioning the usefulness of treating them as hierarchically distinct in terms of diagnostic certainty.
The study highlighted the need for greater clarity in temporal relationships between cognitive decline and co-occurring neurological or psychiatric symptoms, as this was a common source of diagnostic error.
5 The authors advocated for future formal revisions of the criteria to incorporate these observations—many of which were already reflected in the clinical judgment of experienced diagnosticians.
This work laid the groundwork for future revisions of AD diagnostic frameworks, including the NIA-AA criteria and the AT(N) biomarker model, by underscoring the limits of symptom-based classification and the need to refine criteria for atypical and mixed presentations.