Can classic markers of Alzheimer's disease help with early screening?

Table 1 summarizes the demographic characteristics of possible AD patients and CN participants, including 277 possible AD patients and 153 CN participants
.
The mean age of onset (AAO) was 62.
23 years, and the average duration of the disease was about 2.
91 years
.
The sex ratio (F/M) for probable AD patients was 172/105, matching the CN group (99/54) (P>0.
05).

Meanwhile, 160 (57.
8%) AD patients carried at least one APOE4 allele
.
There was no clear difference between the age at diagnosis (referring to the age at which blood was drawn) between the likely AD patients and CN participants, but there was a clear difference in the educational level between the two groups (P < 0.
001).

Table 1 Demographic and clinical characteristics of possible AD patients and healthy controls

Before correcting for confounders, all detected plasma biomarkers, including Aβ42, Aβ40, Aβ42/Aβ40, t-tau, and NfL, differed
significantly between AD and CN groups.
Among them, plasma Aβ42 and Aβ42/40 levels in the AD group were significantly lower than those in the CN group (P, respectively = 0.
02 and P < 0.
001).

At the same time, in patients with probable AD, levels of Aβ42/Aβ40, t-tau, and NfL showed an upward trend (p < 0.
05; Table 1).

However, after correcting for age, sex and APOE alleles and corrected by Bonferroni, plasma t-tau and NfL in the AD group increased significantly compared with the CN group (p<0.
001), while other plasma biomarkers, including Aβ42 and Aβ40, showed no significant difference (p=0.
308 and p=0.
062, respectively), and Aβ42/Aβ40 showed a downward trend, but the difference was not significant (p=0.
056; Table 1; Figure 1).
<b10> 。

Fig.
1 Comparison of plasma biomarkers between AD group and CN group

Next, the authors analyzed correlations between plasma biomarkers and demographic data, including age of onset (AAO), age at diagnosis, disease course, and level of
education.
Partial correlation analysis was used to study associations
between plasma biomarkers and demographic data after correcting for age, sex, and APOE alleles at diagnosis.

In addition, differences in plasma biomarkers were compared according to sex, family history, and distribution of APOE4 alleles (Table 2).

After adjusting for the diagnosis of age, sex, and APOE alleles, Aβ42, Aβ40, Aβ42/Aβ40, and t-tau were not significantly associated with AAO (P>0.
05) in the probable AD group, but NfL showed a significant association (r=-0.
183, P<0.
001), indicating that the earlier the AAO, the higher
the level of NfL in plasma.
In contrast, there was a significant positive correlation between age at diagnosis and plasma NfL levels (r = 0.
235，p < 0.
001).

At the same time, plasma NfL is positively correlated with the course of the disease (r = 0.
199，p < 0.
001).

In addition, plasma Aβ42 and Aβ42/Aβ40 were significantly lower
in patients with dementia without a positive family history compared with patients with a positive family history.
Regarding the APOE allele, patients carrying the APOE4 allele had significantly lower Aβ42/Aβ40 ratios and NfL than non-carriers (p, respectively = 0.
033 and p = 0.
034), while other plasma biomarkers did not differ significantly between the two subgroups (Table 2).

The authors further analyzed the correlation between plasma biomarkers and neuropsychological assessments, including the Miniature Mental Status Examination (MMSE), the Montreal Cognitive Assessment Scale (MoCA), Activities of Daily Living (ADL), and the Neuropsychiatric Questionnaire (NPI) (Table 2).

Partial correlation analysis was used to study the association
between plasma biomarkers and neuropsychological assessments, after adjusting for age, sex, APOE alleles, and education level.

The results showed that all plasma biomarkers were not associated with neuropsychological assessments, including MMSE, MoCA, ADL, and NPI
.
To further analyze the correlation between plasma biomarker levels and disease severity, patients who may have AD were divided into three subgroups
based on CDR scores.
There were no differences
in plasma biomarkers in patients of different disease severities.

Table 2 Correlation of plasma biomarkers in AD patients with demographic and neuropsychological assessments

According to the diagnostic framework of ATN, not all clinically diagnosed patients with suspected AD match
the biological definition of AD.
Specifically, of the 89 patients who underwent a lumbar puncture, 70 (78.
7%) were diagnosed with the AD continuum (A +T+N+: 39 cases, A + T + N-: 8 cases, A +T - N+: 9 cases, A + T - N -: 14 cases).

To assess the ability of plasma biomarkers to reflect pathological changes in the brain, a correlation between plasma and CSF biomarker levels in possible AD patients was determined, as shown
in Table 3.
After adjusting for age, sex, and APOE alleles, only plasma t-tau was positively correlated with t-tau in CSF (r = 319，p = 0.
003）
。

Table 3 Correlation between plasma biomarkers and CSF biomarkers in AD patients (n=89)

Finally, ROC curves are generated to assess the differential performance of plasma biomarkers for possible AD patients and controls (Figure 2).

The cut-off and its corresponding sensitivity and specificity were calculated using the maximum Youden index (Table 4).

As a single plasma biomarker, NfL showed the best diagnostic efficacy (AUC = 0.
85, sensitivity = 73.
28%, specificity = 83.
00%)
.

Based on the results of differences in plasma biomarkers, models combining plasma t-tau and NfL had the best diagnostic power (AUC: 0.
86, sensitivity: 83.
75%, specificity: 76.
47%)
.
In addition, the joint model containing age, sex, and APOE alleles showed the best performance in differentiating possible AD and CN (AUC: 0.
89, sensitivity: 82.
31%, specificity: 83.
66%) and significantly outperformed other models (P < 0.
05).

In addition, the authors extracted 89 AD patients who underwent CSF biomarker testing and analyzed the diagnostic performance
of their plasma biomarkers by ROC curves.
The results showed that when plasma t-tau and NfL were included in the model, their diagnostic efficiency was better than any single plasma biomarker
.
After including age, sex, and APOE alleles into the model, there was no statistically significant difference in diagnostic power between the two models, but the AUC of the latter was optimal (AUC=0.
89, sensitivity =78.
65%, specificity=88.
88%; Table 4).

Fig.
2 ROC curve analysis of plasma biomarkers for diagnosing AD

Table 4 Plasma biomarkers for possible diagnostic manifestations of AD