Abstract
There are currently no accepted blood-based biomarkers of sporadic Alzheimer's disease (AD). Augmented oxidative stress has been implicated in both neural and peripheral AD tissues. In this study, we determined whether short-wavelength near-infrared (NIR) spectrophotometry of blood plasma differentiates mild sporadic AD from normal aging. NIR analysis was conducted on 75 μl plasma samples from 19 AD, 27 amnestic MCI, and 17 normal elderly control (NEC) persons using an optical fiber-coupled, holographic grating-based NIR spectrograph. Five spectral bands associated with heme, R-CH, R-OH, H2O, and R-NH functional groups were sensitive to oxidative modification in pre-clinical studies and were pre-selected to develop a logistic regression model for sample classification. This model differentiated AD from NEC samples with a sensitivity of 80% and specificity of 77%. Fifteen and twelve MCI patients were classified with the NEC and AD groups, respectively. The spectra were not influenced by age, gender, exposure to cholinesterase inhibitors or vitamin E, or sample storage time. The NIR data further implicate oxidative stress in the systemic pathophysiology of sporadic AD and differentiate mild (and possibly pre-clinical) AD from NEC individuals with moderate-high accuracy. The procedure is minimally-invasive, rapid, relatively-inexpensive, and may provide a useful biological marker of sporadic AD.