Classification of non-demented patients attending a memory clinic using the new diagnostic criteria for Alzheimer's disease with disease-related biomarkers.

Abstract

New diagnostic criteria for predemential Alzheimer's disease (AD) advocate the use of biomarkers. However, the benefit of using biomarkers has not been clearly demonstrated in clinical practice. To investigate whether a combination of biomarkers may be helpful in classifying a population of non-demented patients attending a Memory Clinic. Sixty non-demented patients were compared with 31 healthy elderly subjects. All subjects underwent a neuropsychological examination, brain 3T magnetic resonance imaging, [F18]-fluorodeoxyglucose and [F18]-flutemetamol positron emission tomography. According to their performance on memory, language, executive, and visuo-spatial domains, the patients were classified as mild cognitive impairment (amnestic, non-amnestic, single, or multiple domain) or subjective cognitive impairment. Patients were then classified according to the National Institute on Aging-Alzheimer's Association (NIA-AA) criteria, using the normalized mean hippocampal volume (Freesurfer), [F18]-FDG PALZAD, and [F18]-flutemetamol standard uptake value ratio (SUVr) (cut-off at the 10th percentile of controls). The standard of truth was the clinical status at study entry (patient versus control). The sensitivity/specificity of the clinical classification was 65/84%. The NIA-AA criteria were applicable in 85% of patients and 87% of controls. For biomarkers the best sensitivity (72%) at a fixed specificity of 84% was achieved by a combination of the three biomarkers. The clinical diagnosis was reconsidered in more than one third of the patients (42%) as a result of including the biomarker results. Application of the new NIA-AA AD diagnostic criteria based on biomarkers in an unselected sample of non-demented patients attending a Memory Clinic was useful in allowing for a better classification of the subjects.