Diagnostic performance of plasma p-tau217 levels measured with different assays for Alzheimer’s disease

Abstract

Plasma tau phosphorylated at threonine 217 (p-tau217) has been recommended as a biomarker for the diagnosis of Alzheimer’s disease (AD). We evaluated the diagnostic and differential performance of plasma p-tau217 levels measured with three novel assays in a Chinese population. A total of 233 participants were recruited, including 39 cognitively unimpaired controls (CUCs), 28 individuals with mild cognitive impairment (MCI) due to AD, 57 individuals with AD dementia (ADD), 70 individuals with subcortical ischemic vascular dementia (SIVD), and 39 individuals with frontotemporal lobar degeneration (FTLD). Plasma p-tau217 levels were measured using one assay based on single-molecule techniques (DiSMS), one assay based on digital ELISA (LyMedivh™ AXL), and one assay based on flow cytometry (CBA), as well as a reference assay (ALZpath Simoa). Group differences in plasma p-tau217 levels were assessed using analysis of covariance, and the diagnostic and differential performance of the assays was evaluated via receiver operating characteristic analysis. Partial correlation analysis was used to examine the correlations between the measurements of the three novel assays and those of the reference assay. We found that plasma p-tau217 levels measured with all three novel assays were higher in the ADD group than in the CUC, SIVD, and FTLD groups (all p < 0.05) and effectively discriminated ADD patients from both CUCs and non-AD dementia patients. The diagnostic and differential performances did not significantly differ among the three assays (all p > 0.05). Both the DiSMS and LyMedivh™ AXL assays also revealed elevated plasma p-tau217 levels in the MCI group compared to the CUC group. Moreover, the measurements of the three novel assays demonstrated significant correlations with the ALZpath Simoa measurements (p < 0.01). When using their optimal cutoff values, both the DiSMS and LyMedivh™ AXL assays yielded a specificity of 100% and a sensitivity of 94.4%, and the CBA assay showed a specificity of 100% and a sensitivity of 88.9%. In conclusion, our study demonstrated the diagnostic and differential abilities of plasma p-tau 217 levels measured with three novel assays that can serve as potential alternatives to the currently available testing methods for AD diagnosis.

Introduction

Alzheimer’s disease (AD) is the most common cause of dementia in the elderly and is characterized by progressive neurodegeneration and cognitive and functional decline. Early and accurate diagnosis of AD is crucial for initiating timely clinical interventions and developing disease-modifying therapies [1]. However, it is not easy to determine whether cognitive impairment is caused by AD at early stages, e.g., at the mild cognitive impairment (MCI) stage, based on clinical information alone; moreover, the clinical manifestations of AD overlap with those of other common types of dementia, such as subcortical ischemic vascular dementia (SIVD) and frontotemporal lobar degeneration (FTLD) [2].

Core biomarkers of AD, particularly in the cerebrospinal fluid (CSF) and measured with positron emission tomography (PET), have been demonstrated to be highly consistent with neuropathological findings and are included in the revised criteria for the diagnosis and staging of AD [3]. Moreover, blood-based biomarkers offer a cost-effective and less invasive alternative to CSF and PET biomarkers for AD diagnosis [4–6]. In recent years, plasma tau phosphorylated at threonine 217 (p-tau217) has been recognized as an accurate biomarker for AD diagnosis and was emphasized as a key tool in AD diagnosis in the 2025 Alzheimer’s Association Clinical Practice Guideline [7–13]. There is evidence that the plasma p-tau217 level is strongly associated with cerebral amyloid-β (Aβ) pathology and can reflect early Aβ changes even in cognitively unimpaired individuals as well as identify AD-related neuropathological changes across a diverse spectrum of neurodegenerative diseases [7–12, 14, 15].

Although ultrasensitive assays for measuring plasma p-tau217 levels have been developed and widely used in research, such as single-molecule array (Simoa) assays, chemiluminescent enzyme immunoassays, mesoscale discovery (MSD) electrochemiluminescence immunoassays, and immunoprecipitation–mass spectrometry (IP-MS) [10, 16–20], novel methods that are easy to expand into clinical practice are worthy of investigation. The latest Clinical Practice Guideline from the Alzheimer’s Association has recommended blood-based biomarker tests with ≥90% sensitivity and specificity as credible diagnostic tools for AD patients [13]. This underscores the necessity for a rigorous evaluation of novel assays against this critical threshold before they can be translated into clinical practice. In addition, only a few studies have validated plasma p-tau217 assays in Chinese populations, and their ability to differentiate AD from other types of dementia has rarely been evaluated [21–23].

To address these gaps, we evaluated and compared the diagnostic performance of three novel p-tau217 assays in a Chinese cohort. Specifically, we aimed to 1) investigate plasma p-tau217 levels across different diagnostic groups, including cognitively unimpaired controls (CUCs), individuals with MCI due to AD, individuals with AD dementia (ADD), and individuals with non-AD dementia (e.g., SIVD and FTLD); 2) compare the ability of each assay to discriminate individuals with ADD from both CUCs and individuals with non-AD dementia, as well as MCI patients from CUCs, and determine whether any assay demonstrates superior accuracy in specific group comparisons; and 3) explore the consistency of the three assays with the reference assay (ALZpath Simoa) for discriminating between ADD patients and CUCs. This study provides evidence to support the clinical translation of alternative assays for plasma p-tau217 in AD diagnosis in Chinese healthcare settings.

Methods

Participants

This study was approved by the Ethics Committee of Tianjin Medical University General Hospital. A total of 233 participants were recruited from our longitudinal cohort at Tianjin Medical University General Hospital, including 39 CUCs and 194 patients with cognitive impairment. The patient group was further stratified into four subgroups: 28 individuals with MCI due to AD, 57 individuals with ADD, 70 individuals with SIVD, and 39 individuals with FTLD. All participants provided written informed consent. This was an exploratory study. Empirical and feasibility considerations determined the number of enrolled participants. No formal statistical estimation of sample size was performed. The inclusion criteria used for the CUCs were as follows: 1) aged 50–85 years; 2) the absence of subjective cognitive decline complaints; 3) normal performance in each cognitive domain on objective neuropsychological tests, with a Clinical Dementia Rating (CDR) [24] score of 0; and 4) the absence of clinically significant brain atrophy or cerebrovascular lesions on magnetic resonance imaging (MRI). All patients were diagnosed according to specific diagnostic criteria, with an age range of 50–85 years and CDR scores of 0.5 for MCI and 1–2 for dementia. Specifically, the AD patients (those with MCI due to AD or ADD) were diagnosed based on the National Institute on Ageing and the Alzheimer’s Association (NIA-AA) criteria (2011) [25, 26] and the International Working Group (IWG)-2 criteria [27], with a positive Aβ-PET result; the SIVD patients met the diagnostic criteria for vascular dementia according to the International Society of Vascular Behavioural and Cognitive Disorders [28]; and the FTLD patients were diagnosed based on the revised Frontotemporal Dementia Consensus criteria for behavioral variant frontotemporal dementia[29] or the classification recommendations for primary progressive aphasia and its variants [30]. Participants were excluded if their cognitive impairment was potentially caused by other neurological diseases, mental disorders or medical conditions, such as Parkinson’s disease dementia or dementia with Lewy bodies, multiple sclerosis, severe depression, alcohol or drug use disorders, vitamin B12 deficiency, or human immunodeficiency virus infection or syphilis.

Plasma sampling and analysis

Blood samples were collected in EDTA tubes in the morning. The samples were centrifuged at 2500 × g and 4 °C for 35 min within 2 h of collection to obtain plasma. The plasma samples were stored at −80 °C for analysis. Plasma p-tau217 levels were measured by three independent assays.

A single-molecule immunofluorescence assay based on a digital single-molecule system (DiSMS) was performed with a CJC - Di6000A instrument (Targeting Detection, Changsha, Hunan, China). The assay was performed with a sandwich immunoassay. First, magnetic beads coated with capture antibodies were incubated with diluted samples or standards, together with biotin‑labeled detection antibodies, and shaken at 30 °C for 30 min to form sandwich complexes. After magnetic separation and washing three times, streptavidin–β-galactosidase was added and incubated at 30 °C for 10 min. Following another washing step, resorufin-β-D-galactopyranoside was added. The fluorescence signal generated was measured by a CJC - Di6000A instrument, and the concentration of P-tau217 was calculated based on a standard curve.

A cytometric bead array (CBA) assay (CellGene Biotech, Hangzhou, Zhejiang, China) was performed with a FACS Canto II flow cytometer (BD Biosciences, San Jose, CA, USA) [31]. The assay was conducted with a typical sandwich-type structure. First, capture beads coated with p‑tau217‑specific antibodies were incubated with samples or standards and biotin‑conjugated detection antibodies to form sandwich complexes for 18 h at 2–8 °C in the dark. Then streptavidin-PE was added and further incubated for 30 min in the dark at room temperature. Finally, the mixture was washed with washing buffer by centrifugal resuspension to remove the unbound antigen. The intensity of the fluorescence produced by the CBA beads was measured on a FACS Canto II flow cytometer.

A digital enzyme-linked immunosorbent assay (ELISA) was performed on the LyMedivh™ AXL platform (Lychix Bio, Suzhou, Jiangsu, China) [32, 33]. The assay was conducted with a sandwich immunoassay. First, the sample or standards was incubated with magnetic beads coated with capture antibodies and detection antibodies at room temperature for 30 min to form a “magnetic bead–capture antibody–antigen–detection antibody” sandwich complex. After magnetic separation and washing three times, streptavidin–β-galactosidase was added and incubated for 5 min at room temperature. Following another magnetic separation and extensive washing seven times, resorufin-β-D-galactopyranoside was added. Finally, fluorescence signals were measured on a bio‑chip reader using LyMedivh™ AXL, and P-tau217 concentrations were determined from a standard curve.

In addition, plasma p-tau217 was previously measured using the ALZpath single-molecule array (ALZpath Simoa) p-tau217 v2 EQC Kit (Quanterix, 104372) for all participants [7]. Owing to the limited volume of the plasma samples, the three assays were not completed for all participants, resulting in varying sample sizes across the assays. For the DiSMS assay, 39 samples were included in the CUC group, 26 in the MCI group, 54 in the ADD group, 68 in the SIVD group, and 37 in the FTLD group. For the CBA assay, the corresponding sample sizes were 26 (CUC), 19 (MCI), 47 (ADD), 49 (SIVD), and 32 (FTLD). For the LyMedivh™ AXL assay, the corresponding sample numbers in the groups were 15, 21, 40, 47 and 28.

Statistical analysis

Statistical analyses were performed with SPSS 26.0, GraphPad Prism version 9.0 and R statistical software version 4.5.1. Demographic data were compared between groups using one-way analysis of variance (ANOVA) with the post hoc Bonferroni correction for continuous variables and the chi-square test for dichotomous variables. Data were tested for normality and homoscedasticity using Shapiro Wilk and Levene’s tests, respectively. The plasma P-tau217 concentrations had a skewed distribution and were then log10 transformed for group comparisons and correlation analysis. The group comparisons were conducted with analysis of covariance (ANCOVA) and the post hoc Bonferroni correction, with adjustment for age and sex. Receiver operating characteristic (ROC) analysis was used to compare the discriminative performance of plasma p-tau217 levels between ADD patients and CUCs, SIVD patients, and FTLD patients, as well as between MCI patients and CUCs. The areas under the ROC curves (AUCs) for each assay were compared by performing the DeLong test.

To explore the consistency of the three novel assays with the reference assay (ALZpath Simoa) in discriminating between ADD patients and CUCs, a subset of samples (n = 51) was selected from the overall cohort. Specifically, only AD patients and CUCs with complete plasma p-tau217 data from all four assays were included. We first examined the associations between plasma p-tau217 levels measured by each of the three assays and those measured by the ALZpath Simoa assay using partial correlation analysis adjusted for age and sex. In addition, optimal cutoff values for each assay were determined using the Youden index determined via ROC analysis performed in R statistical software. These cutoff values were subsequently used to calculate the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy for each assay. Using 1000 bootstrap resamples, we estimated the 95% confidence intervals (CI) for the above metrics and performed statistical comparisons of these metrics between the three novel assays and the reference assay. In addition, the associations between plasma p-tau217 levels and cognitive function measured with MMSE were analyzed using partial correlation analysis adjusted for age and sex in patients on the AD continuum (including MCI due to AD and ADD). A two-tailed p value < 0.05 was considered to indicate statistical significance.

Results

Demographic and clinical characteristics

The demographic and clinical characteristics of all the participants are presented in Table 1. The SIVD group was significantly older than both the ADD group and the FTLD group (both p < 0.05), and had a greater proportion of males compared to the other patient groups (all p  < 0.05). The overall analysis revealed significant differences in the mean years of education across the groups (p < 0.05); however, post hoc pairwise comparisons indicated no significant differences between any two groups. The Mini-Mental State Examination (MMSE) scores were lower in all patient groups than in the CUC group (all p < 0.05) and in all dementia groups than in the MCI group (all p < 0.05). When comparing between each dementia group, the SIVD group had higher MMSE scores than the ADD group (p < 0.05).

Table 1.

Demographic and clinical characteristics of all participants.

	CUC (n=39)	MCI (n=28)	ADD (n=57)	SIVD (n=70)	FTLD (n=39)	χ2/F	p
Age, years	67.64 (6.96)	68.86 (5.28)	65.21 (8.05)	70.40 (7.06)	65.00 (7.63)	5.673	< 0.001
Sex, male	19 (48.72%)	5 (17.86%)	20 (35.09%)	50 (71.43%)	16 (41.03%)	30.15	< 0.001
Education, years	12.44 (3.12)	12.68 (3.04)	10.88 (3.94)	10.89 (3.11)	10.97 (3.72)	2.736	0.030
MMSE score	27.85 (1.53)	24.54 (2.80)	16.60 (5.67)	21.27 (4.52)	19.10 (6.49)	38.081	< 0.001

Data are presented as the means (SD) for continuous data and n (%) values for dichotomous data. Abbreviations: CUC, cognitively unimpaired control; MCI, mild cognitive impairment; ADD, Alzheimer’s disease dementia; SIVD, subcortical ischemic vascular dementia; FTLD, frontotemporal lobar degeneration; MMSE, Mini-Mental State Examination.

Differences in plasma p-tau217 levels across all groups

We investigated differences in the plasma p-tau217 levels measured by the three assays across the diagnostic groups (Table 2 and Fig. 1). All three assays showed significantly higher plasma p-tau217 levels in the ADD group than in the CUC, SIVD, and FTLD groups (all p < 0.05). Higher plasma p-tau217 levels were also observed in the MCI group than in the CUC, SIVD, and FTLD groups for the DiSMS and LyMedivh™ AXL assays and in the MCI group than in the FTLD group for the CBA assay (all p < 0.05). The results of the CBA assay also revealed significantly higher plasma p-tau217 levels in the ADD group than in the MCI group (p < 0.05); this difference was not observed in the other two novel assays. In the comparisons between the non-AD dementia groups and the CUC group, both the DiSMS and LyMedivh™ AXL assays showed significantly elevated plasma p-tau217 levels in the SIVD group compared with the CUC group (both p < 0.05); this difference was not observed for the CBA assay. In contrast, no significant difference between the FTLD group and the CUC group was detected by any assay. Finally, all three assays showed higher plasma p-tau217 levels in the SIVD group than in the FTLD group (all p < 0.05).

Table 2.

Plasma p-tau217 levels across all groups.

	CUC	MCI	ADD	SIVD	FTLD	χ2/F	p
n	39	26	54	68	37
DiSMS, pg/ml	0.22 (0.19, 0.35)	0.96 (0.66, 1.37)	1.11 (0.87, 1.94)	0.36 (0.25, 0.60)	0.23 (0.20, 0.31)	36.241	< 0.001
n	26	19	47	49	32
CBA, pg/ml	0.16 (0.11,0.24)	0.37 (0.30, 0.54)	0.91 (0.59, 1.49)	0.22 (0.14, 0.51)	0.11 (0.06, 0.29)	27.265	< 0.001
n	15	21	40	47	28
LyMedivh™ AXL, pg/ml	0.07 (0.05, 0.13)	0.39 (0.22, 0.53)	0.45 (0.29, 0.68)	0.11 (0.08, 0.27)	0.06 (0.04, 0.10)	37.548	< 0.001

The data are presented as medians (IQR).

CUC cognitively unimpaired control, MCI mild cognitive impairment, ADD Alzheimer’s disease dementia, SIVD subcortical ischemic vascular dementia, FTLD frontotemporal lobar degeneration, DiSMS digital single-molecule system, CBA cytometric bead array.

Fig. 1. Comparison of plasma p-tau217 levels measured with the three assays among all groups.

Group differences were assessed using ANCOVA and the post hoc Bonferroni correction with adjustment for age and sex. (A) DiSMS; (B) CBA; (C) LyMedivh^TM AXL. * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001. Abbreviations: CUC, cognitively unimpaired control; MCI, mild cognitive impairment; ADD, Alzheimer’s disease dementia; SIVD, subcortical ischemic vascular dementia; FTLD, frontotemporal lobar degeneration; DiSMS, digital single-molecule system; CBA, cytometric bead array.

Discriminative performance of plasma p-tau217

We evaluated the performance of plasma p-tau217 levels measured with three novel assays for discriminating between MCI patients and CUCs and for discriminating between ADD patients and CUCs, SIVD patients, and FTLD patients (Fig. 2). ROC analyses revealed that all three assays exhibited excellent performance in discriminating between ADD patients and both CUCs and FTLD patients. The DiSMS assay achieved AUC values of 0.977 (95% CI, 0.953–1.000) for ADD vs. CUC and 0.973 (95% CI, 0.944–1.000) for ADD vs. FTLD; the CBA assay achieved AUC values of 0.921(95% CI, 0.858–0.984) for ADD vs. CUC and 0.909 (95% CI, 0.836–0.981) for ADD vs. FTLD; the LyMedivh™ AXL achieved AUC values of 0.971 (95% CI, 0.928–1.000) for ADD vs. CUC and 0.947 (95% CI, 0.892–1.000) for ADD vs. FTLD. The differences in the AUC values for corresponding discrimination between the three assays were not statistically significant (DeLong test, p > 0.05). All three assays also exhibited similar good performance in discriminating between ADD patients and SIVD patients (DiSMS, AUC = 0.882, 95% CI = 0.822–0.941; CBA, AUC = 0.829, 95% CI = 0.740–0.910; LyMedivh™ AXL, AUC = 0.834, 95% CI = 0.749–0.920). For discriminating between MCI patients and CUCs, both the DiSMS assay (AUC = 0.941, 95% CI = 0.865–1.000) and the LyMedivh™ AXL assay (AUC = 0.957, 95% CI = 0.749–0.920) yielded higher values than the CBA assay did (AUC = 0.807, 95% CI = 0.740–0.910), although these AUC values did not differ significantly (DeLong test, p > 0.05).

Fig. 2. Diagnostic and differential performance of plasma p-tau217 for AD.

ROC analysis for p-tau217 measured with three novel assays in discriminating ADD from CUC (A), ADD from SIVD (B), ADD from FTLD (C), and MCI from CUC (D), and corresponding AUC values (E). CUC, cognitively unimpaired control; MCI, mild cognitive impairment; ADD, Alzheimer’s disease dementia; SIVD, subcortical ischemic vascular dementia; FTLD, frontotemporal lobar degeneration; DiSMS, digital single-molecule system; CBA, cytometric bead array.

Consistency of the three assays with the reference assay (ALZpath Simoa)

We examined the associations between plasma p-tau217 levels measured by each of the three novel assays and those measured by the ALZpath Simoa assay in a subset of samples (Fig. 3). The measurements of all three assays demonstrated significant correlations with those of the ALZpath Simoa assay (DiSMS assay, r = 0.890, p < 0.001; LyMedivh™ AXL assay, r = 0.737, p < 0.001; CBA assay, r = 0.407, p = 0.004)). The cutoff values for the plasma p-tau217 level were established for each assay as follows: ALZpath Simoa, 0.475; DiSMS, 0.451; CBA, 0.381; and LyMedivh™ AXL, 0.151. These cutoff values were subsequently used to calculate the sensitivity, specificity, PPV, NPV, and accuracy in discriminating between ADD patients and CUCs (Table 3). All four assays achieved a specificity and PPV of 100%. The ALZpath Simoa, DiSMS and LyMedivh™ AXL assays exhibited the same sensitivity (94.4%), NPV (88.2%), and accuracy (96.1%), whereas the CBA assay exhibited lower sensitivity (88.9%), NPV (79.0%), and accuracy (92.2%). No significant differences in these metrics were observed between the reference assay (ALZpath Simoa) and the other three assays (p > 0.05).

Fig. 3. Associations between plasma p-tau217 levels measured by each of the three novel assays (DiSMS, CBA, and LyMedivh™ AXL) and those measured by the ALZpath Simoa assay in a subset of samples.

(A) DiSMS; (B) CBA; (C) LyMedivh^TM AXL. Plasma p-tau217 values were log-transformed. The partial correlation coefficients (r) were adjusted for age and sex. DiSMS, digital single-molecule system; CBA, cytometric bead array.

Table 3.

Performance of plasma p-tau217 levels measured with the four assays for discriminating ADD patients from CUCs.

	Sensitivity	Specificity	PPV	NPV	Accuracy
ALZpath Simoa	94.4 (71.4–100.0)	100.0	100.0	88.2 (71.4–100.0)	96.1 (90.2–100.0)
DiSMS	94.4 (86.1–100.0)	100.0	100.0	88.2 (71.4–100.0)	96.1 (90.2–100.0)
CBA	88.9 (78.1–97.4)	100.0	100.0	79.0 (60.0–95.2)	92.2 (84.3–98.0)
LyMedivh™ AXL	94.4 (85.4–100.0)	100.0	100.0	88.2 (70.6–100.0)	96.1 (90.2–100.0)

The ALZpath Simoa assay served as the reference method. The differences in all the metrics between the three novel methods and the reference were not statistically significant (all p > 0.05) when they were analyzed on the basis of bootstrap resampling (n = 1000). The accuracy indicates the percentage of correctly classified participants. The 95% CI is presented in parentheses. ADD, Alzheimer’s disease dementia; CUCs, cognitively unimpaired controls; PPV, positive predictive value; NPV, negative predictive value. DiSMS, digital single-molecule system; CBA, cytometric bead array.

Association between plasma p-tau217 level and MMSE score

No statistically significant correlation was observed between the plasma p-tau217 concentration and MMSE score in AD patients (DiSMS: r = −0.129, p = 0.261; CBA: r = −0.221, p = 0.079; LyMedivh™ AXL: r = −0.074, p = 0.575).

Discussion

In this study, the diagnostic and differential performance of three novel plasma p-tau217 assays (DiSMS, CBA, and LyMedivh™ AXL) for MCI and dementia due to AD was evaluated in a Chinese cohort. All three assays effectively discriminated between ADD patients and both CUCs and patients with non-AD dementias, and the measurements were significantly correlated with those of the ALZpath Simoa reference assay. Although the DiSMS and LyMedivh™ AXL assays were more accurate, no significant differences were observed between the three assays, suggesting that they could be applied for the diagnosis and differentiation of AD in clinical practice.

Consistent with the results of previous studies using the Simoa, Lumipulse, MSD and IP-MS assays (9,12,21), we found that the plasma p-tau217 levels measured with the three novel assays were substantially higher in ADD patients than in CUCs or patients with non-AD dementias [9, 12] and exhibited excellent diagnostic performance for AD. Moreover, both the DiSMS and LyMedivh™ AXL assays could detect elevated plasma p-tau217 levels in MCI patients and discriminate between those individuals and both CUCs and patients with non-AD dementias, suggesting their high sensitivity for identifying early stages of AD-related neuropathological changes. In contrast, the CBA assay showed only moderate diagnostic performance for MCI due to AD. Thus, the DiSMS and LyMedivh™ AXL assays appear to be advantageous for screening for AD patients at prodromal stages.

Elevated plasma p-tau217 levels measured by the DiSMS and LyMedivh™ AXL assays were also observed in SIVD patients compared with CUCs. While all three novel assays demonstrated a moderate ability to discriminate between ADD patients and SIVD patients, with AUC values ranging from 0.825–0.882, a previous study reported that plasma p-tau217 levels measured using immunoassays achieved an AUC value of 0.97 (95% CI, 0.94–0.99) in distinguishing ADD from subcortical vascular dementia [9]. Our current results could be attributed to the presence of mixed AD pathology in SIVD patients because we did not use PET or CSF biomarkers to exclude the presence of possible AD pathology in these patients. In addition, the SIVD patients included in this study were significantly older and carried a greater burden of vascular risk factors such as hypertension and diabetes. Previous studies have shown that higher plasma p-tau217 levels were correlated with older age and presence of hypertension and diabetes [10, 34]. Furthermore, although current evidence directly linking plasma p-tau217 levels to pure vascular dementia remains limited, studies in patients with AD have suggested that subcortical cerebrovascular lesions (e.g., white matter hyperintensities and reduced white matter microstructural integrity) are associated with elevated plasma p-tau217 levels [35, 36]. Thus, higher plasma p-tau217 concentrations in SIVD patients may be attributed to age- and vascular-related effects on AD pathology, in addition to comorbidities.

Furthermore, we compared the accuracy of the three novel assays with the reference method of ALZpath Simoa in discriminating between ADD patients and CUCs. The ALZpath Simoa assay has demonstrated excellent performance in identifying amyloid PET positivity [18, 37, 38]. When using their optimal cutoff values, all four assays achieved a specificity and PPV of 100%, which means that individuals classified as “positive” by any assay were consistently true AD patients. These results supported the use of these assays for determining eligibility for clinical trials and disease-modifying therapies and for confirming an AD diagnosis in patients with complex situations, such as those with early onset or atypical manifestations. In terms of sensitivity, the DiSMS and LyMedivh™ AXL assays were highly and precisely consistent with the reference ALZpath Simoa assay (94.4%), whereas the CBA assay showed slightly a lower sensitivity (88.9%) that was not significantly different from that of the other assays. These results suggest that the three novel assays are potential alternatives to the ALZpath Simoa assay in the context of measuring plasma p-tau 217 levels in the Chinese population.

In our study, plasma p‑tau217 levels did not significantly correlate with MMSE scores in AD patients. It has been recognized that an increase in plasma p‑tau217 levels is an early pathological change in AD. Previous studies reporting a significant correlation between plasma p‑tau217 levels and cognitive function primarily in individuals at the preclinical or early symptomatic stages of AD; however, the present study included not only MCI patients but a considerable proportion of patients with mild to moderate dementia. Furthermore, the MMSE serves as a screening tool for global cognition and is not sensitive to changes in specific cognitive domains, such as executive function, information processing, and visuospatial function, which usually worsen in the later stages of AD and were associated with elevated plasma p-tau217 levels in our previous study [7].

This study included individuals with various cognitive statuses and patients with various types of dementia, thereby addressing a significant gap in the evaluation of plasma p-tau217 levels in the Chinese population. This study has several limitations. First, the modest sample size compromised statistical power, increasing the risk of false-negative results and limiting the generalizability of the findings to broader populations. In addition, the sample size varied across the different assays, undermining the fairness of intermethod comparisons. Our results still require validation via studies with larger sample sizes. Lastly, as mentioned above, PET and CSF biomarkers were not employed in this study to exclude possible comorbid AD pathology in all individuals from the CUC group and the non-AD dementia groups, may leading to spuriously elevated plasma p‑tau217 levels in these groups. Therefore, the differential performance might be underestimated in this study (e.g., all the AUC values of the three methods were less than 0.9 for AD vs. SIVD).

Conclusions

In conclusion, our study demonstrated the diagnostic and differential performance of plasma p-tau 217 levels measured with three novel assays, which are potential alternatives to the currently available testing methods for the screening and diagnosis of AD.

Author contributions

YH drafted the manuscript and performed data analyses; YR D assisted with data collection, organization and analyses; DQ L, PC, YW, JL, and CX W contributed to data collection, organization and curation; NZ conceptualized the study and reviewed and edited the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by Tianjin Public Health Science and Technology Major Project (24ZXGZSY00060), Science and Technology Innovation 2030—Major Projects (2021ZD0201805, 2022ZD0211605), and the Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-004A).

Data availability

The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Tianjin Medical University General Hospital. Written informed consent was obtained from all participants.