The main goal of this study was to explore alterations in the plasma GFAP, NfL, and p-tau181 levels associated with cognitive impairment in a community-based adult cohort in Shenzhen City, China. The results revealed a trend for gradually increasing expression of plasma GFAP, NfL, and p-tau181 from normal cognitive function to MCI and AD, especially in the latter, which was associated with a significantly upregulated expression of these plasma biomarkers. The observed upregulation is consistent with the conclusions of previous research [14]. Furthermore, we found that plasma GFAP, NfL, and p-tau181 levels could distinguish between dementia and non-dementia but not between CU and MCI. In terms of a single plasma biomarker, p-tau181 levels could most strongly identify AD stage. These findings suggest that blood biomarkers can be used as convenient predictors of the AD continuum, providing important evidence for clinical performance in AD dementia.
Previous research has demonstrated that within the AD continuum, the plasma levels of GFAP, NfL, and p-tau181 exhibit a gradual increase, whereas the plasma Aβ1–42/Aβ1–40 ratio displays a progressive decline [14, 15]. In this study, the expression of plasma GFAP, NfL, and p-tau181 in patients with AD was significantly higher than those in the CU or MCI groups, with remarkable differences in pairwise comparisons (CU vs. AD, p < 0.0001; MCI vs. AD, p < 0.001). Although the levels of plasma GFAP, NfL, and p-tau181 in the MCI group were slightly higher than those in CU, the difference was not significant. Similarly, Ingannato et al. reported that the changes in plasma GFAP and p-tau181 levels between patients with MCI and SCD were not notable [14]. Analysis across the disease spectrum shows that plasma GFAP and p-tau181 levels are remarkably elevated during the AD phase. In this study, the plasma levels of Aβ1–42 and Aβ1–40 were measured using Somia technology to estimate the ratio of Aβ1–42 to Aβ1–40, but no significant differences were observed among the CU, MCI, and AD groups, consequently, the data pertaining to A-beta amyloid is not presented in the table. This further indicates that the plasma levels of GFAP, NfL, and p-tau181 demonstrate superior performance compared to plasma Aβ1–42 and Aβ1–40 levels and the Aβ1–42/Aβ1–40 ratio in the context of cognitive impairment.
Sex and age are important risk factors for AD, with women being more likely to present with the disease than men, and prevalence in women being almost twice that in men [16]. The incidence of AD gradually increases with increasing age, and is 6% in the population at 60 years old but can be as high as 35% in those aged 85 years and older. Both overweight and underweight can increase the risk of dementia [17]. This study revealed that the BMI of participants in the AD group was significantly lower than that in the CU and MCI groups (p = 0.026). Individuals with lower educational attainment are at a heightened risk for dementia, too. Although no statistically significant difference was found regarding the years of education among the three groups, our data showed that the AD group received fewer years of education than the MCI group, which, in turn, received fewer years of education than the CU group. It is well established that the major risk factors for developing AD are age and the status of the APOE ε4 allele [2, 18]. AD risk is determined by genetic factors at a level of 60–80%; the APOE ε4 allele explains an essential genetic variation in AD and its presence increases the risk of AD by 3–4 times [2, 19]. In this study, the prevalence of APOE ε4 in the AD group was significantly higher than that observed in the CU and MCI groups; however, prevalence in the MCI group was lower than that in the CU group, which may be related to the insufficient sample size in the study. Future research should aim to incorporate a larger sample size to further elucidate the relationship between APOE ε4 and cognitive function as well as BBB levels.
Interestingly, our correlation analysis of blood biomarkers and cognitive scores revealed that plasma GFAP and p-tau181 levels negatively correlated with the MMSE and MoCA scores to a certain extent, while no significant correlation between plasma NfL level and MoCA score was noted; these findings indicate a closer relationship between the former markers and cognitive function. GFAP levels represent an astrocytic reaction in the process of AD and are elevated in the brains of patients with AD. A systematic review and meta-analysis reported that plasma GFAP levels distinguished between patients with AD and individuals with cognitively normal function, and were strongly associated with brain Aβ pathology; these observations led to speculations for GFAP as a potential blood biomarker for AD [10]. Plasma p-tau181 expression is closely associated with the extensive aggregation of amyloid and tau proteins. The more severe the cognitive impairment, the higher the level of plasma p-tau181 levels, reflecting one of the important diagnostic and screening indicators of AD. In terms of diagnostic performance in AD, plasma p-tau181 showed a superior clinical value (AUC value > 0.8) compared to plasma GFAP and NfL. Bayoumy et al. described that plasma p-tau181 levels could discriminate AD from the control group and this measure has high diagnostic accuracy for AD, with an AUC value of 0.936–0.995 [22]. Moreover, a remarkable finding in this investigation was that the combined capacity of plasma GFAP, NfL, and p-tau181 to predict AD was stronger than individual contributions. When clinical indicators of AD such as age, sex, BMI, years of education, and APOE ε4 carrier status were added, the AUC value for predicting AD stage increased to a certain extent, showing a stronger predictive value for the AD spectrum (AUC > 0.9).
Our findings elucidated that plasma GFAP, NfL, and p-tau181 levels are effective diagnostic and predictive biomarkers for the AD continuum. Despite these observations, this study has some limitations. First, the sample size was small, especially of patients with AD. Second, cognitive status and the levels of plasma biomarkers in the participants were measured at a single point in time without further follow-up or assessment of dynamic changes. Lastly, the content of related proteins in PET–computed tomography and CSF was not included, and comparisons between disease biomarkers in the blood and brain could not be performed.