Premium
P1‐280: CEREBROSPINAL FLUID Aβ42 AND TAU MEASUREMENT ON LUMIPULSE® G: ANALYTICAL VERIFICATION AND METHOD COMPARISON
Author(s) -
Leitão Maria João,
Santana Isabel,
Olmedo Veronica,
Nadal Alicia,
Le Bastard Nathalie,
Baldeiras Inês
Publication year - 2018
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2018.06.286
Subject(s) - cerebrospinal fluid , coefficient of variation , amyloid β , medicine , pathology , gastroenterology , chemistry , chromatography , disease
Background: The cerebrospinal fluid (CSF) Ab42/Ab40 ratio has been proposed a better biomarker for cerebral b-amyloidosis than CSFAb42 alone. Since 2014, the test has been available in our clinical laboratory practice upon request. Here, we evaluated the distribution of the CSF Ab42/Ab40 ratio, as well as its diagnostic accuracy in relation amyloid PET. Methods:The CSF Ab42/Ab40 ratio was measured on a weekly basis in clinical laboratory practice on consecutive samples from 3647 patients (1853 men, 1794 women, mean age 6 standard deviation, 71.5 6 8.2 years) over 3 years using the MSD Abeta Triplex assay according the manufacturer’s instructions (Meso Scale Discovery, Rockville, MD). Longitudinal stability in the measurements was maintained using an elaborate QC system. One thousand nineteen of the patients were from the specialized memory clinic at Sk ane University Hospital. One hundred and forty three of these had undergone amyloid ([18]F-flutemetamol) PET. Optimal cut-points to discriminate bimodally distributed groups were determined by mixture modelling. The optimal cut-point for differentiating Ab-positive from -negative patients according to amyloid PET was determined as the one that generated the highest Youden index. Results: The CSF Ab42/Ab40 ratio (all data) showed a bimodal normal distribution. The mixture modelling-derived optimal cut-point for differentiating the two groups was 0.076. More patients (56%) were found in the low (Ab-positive) group. When examining patients from the specialized memory clinic separately, a similar bimodal distribution was seen; the optimal cut-point for distinguishing the two groups was 0.080 and 49% of the patients had ratios below this limit. In patients who had undergone amyloid PET, the optimal mixture modelling-derived cut-point was 0.077, which was a little lower than the Youden index-derived cut-point (0.083) that best discriminated Ab-positive from -negative cases (diagnostic accuracy 97%). Conclusions: The CSF Ab42/Ab40 ratio is a bimodal biomarker. The striking lack of individuals with grey zone CSF Ab42/Ab40 ratios suggests that people change Ab status according to the ratio quite rapidly. It is not a gradual increase of Ab plaque pathology over years that slowly changes the ratio; rather, the ratio appears to reflect a switch-like shift in Ab homeostasis in the CSF.