Refining Mild-to-Moderate Alzheimer Disease Screening: A Tool for Clinicians

Recent evidence suggests that a substantial minority of people clinically diagnosed with probable Alzheimer disease (AD) in fact do not fulfill the neuropathological criteria for the disease. A clinical hallmark of these phenocopies of AD is that these individuals tend to remain cognitively stable for extended periods of time, in contrast to their peers with confirmed AD who show a progressive decline. We aimed to examine the prevalence of patients clinically diagnosed with mild-to-moderate AD who do not experience the expected clinically significant cognitive decline and identify markers easily available in routine medical practice predictive of a stable cognitive prognosis in this population.

Data were obtained from two independent, longitudinal, observational multicenter studies in patients with mild-to-moderate AD.

The two studies were the European "Impact of Cholinergic Treatment Use" (ICTUS) and the French "REseau sur la maladie d'Alzheimer FRançais" (REAL.FR).

We used prospective data of 756 patients enrolled in ICTUS and 340 enrolled in REAL.FR.

A prediction rule of cognitive decline was derived on ICTUS using classification and regression tree analysis and then cross-validated on REAL.FR. A range of demographic, clinical and cognitive variables were tested as predictor variables.

Overall, 27.9% of patients in ICTUS and 20.9% in REAL.FR did not decline over 2 years. We identified optimized cut-points on the verbal memory items of the Alzheimer Disease Assessment Scale-Cognitive Subscale capable of classifying patients at baseline into those who went on to decline and those who remained stable or improved over the duration of the trial.

The application of this simple rule would allow the identification of dementia cases where a more detailed differential diagnostic examination (eg, with biomarkers) is warranted. These findings are promising toward the refinement of AD screening in the clinic. For a further optimization of our classification rule, we encourage others to use our methodological approach on other episodic memory assessment tools designed to detect even small cognitive changes in patients with AD.