ABSTRACT
BACKGROUND: Cerebrospinal fluid (CSF) biomarkers are increasingly used to confirm the accuracy of a clinical diagnosis of mild cognitive impairment or dementia due to Alzheimer disease (AD). Recent evidence suggests that fully automated assays reduce the impact of some preanalytical factors on the variability of these measures. This study evaluated the effect of several preanalytical variables common in clinical settings on the variability of CSF ß-amyloid 1-42 (Aß1-42) concentrations. METHODS: Aß1-42 concentrations were measured using the LUMIPULSE G1200 from both freshly collected and frozen CSF samples. Preanalytic variables examined were: (1) patient fasting prior to CSF collection, (2) blood contamination of specimens, and (3) aliquoting specimens sequentially over the course of collection (i.e., CSF gradients). RESULTS: Patient fasting did not significantly affect CSF Aß1-42 levels. While assessing gradient effects, Aß1-42 concentrations remained stable within the first 5 1-mL aliquots. However, there is evidence of a gradient effect toward higher concentrations over successive aliquots. Aß1-42 levels were stable when fresh CSF samples were spiked with up to 2.5% of blood. However, in frozen CSF samples, even 0.25% blood contamination significantly decreased Aß1-42 concentrations. CONCLUSIONS: The preanalytical variables examined here do not have significant effects on Aß1-42 concentrations if fresh samples are processed within 2 h. However, a gradient effect can be observed on Aß1-42 concentrations after the first 5 mL of collection and blood contamination has a significant impact on Aß1-42 concentrations once specimens have been frozen.
