Validation of cost-effective, non-invasive methods to identify early (pre-clinical) Alzheimer's disease (AD) is increasingly becoming a key research challenge. We have developed two ELISA sandwich colorimetric tests for the accurate detection of amyloid-β (Aβ)1-40 and Aβ1-42: i) directly accessible (DA) in the plasma, ii) recovered from the plasma sample (RP) after diluting the plasma sample in a formulated buffer, and iii) associated with the remaining cellular pellet (CP). These tests were carried out on samples from healthy controls (n = 19) and individuals with mild cognitive impairment (MCI; n = 27) with amnestic-hippocampal syndrome to investigate whether this comprehensive approach may help to explain the association between blood Aβ levels and MCI. A logistic regression analysis detected seven direct or calculated markers (CP 40, DA 42, RP 42, DA/CP 40, DA/RP 42, DA/CP 42, and DA 42/40) with significant odds ratios (OR) after they were dichotomized with regard to the median of the pooled population. In particular, the likelihood [OR (95% CI)] of having MCI for patients with catCP 40, catDA/RP 42, catDA/CP 42, or catDA 42/40 below the corresponding population median ("positive test") was 11.48 (1.87-70.52), 22.09 (3.19-152.61), 11.48 (1.87-70.50), and 9.54 (1.77-51.38)-fold higher, respectively, than in those with a "negative test" after adjusting for the effect of the ApoE genotype. These results are congruent with the hypothesis that changes in blood Aβ levels may be associated with the initial stages of AD. Thus, these Aβ blood biomarkers might be useful tools for screening for those at increased risk of developing AD.