Published online Sep 19, 2026. doi: 10.5498/wjp.121407
Revised: April 20, 2026
Accepted: June 3, 2026
Published online: September 19, 2026
Processing time: 153 Days and 17 Hours
Anxiety-depression symptoms frequently coexist and greatly shape the life quality and disease course of individuals with Alzheimer’s disease (AD). While serum neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP), as well as neuronal and astrocytic injury markers may serve as relevant predictors in the development of such psychiatric symptoms among AD patients, the un
To determine the relationship of serum NfL and GFAP concentrations with anxi
We retrospectively analyzed clinical data from 270 patients with confirmed AD who were admitted to the neurology unit of our hospital from January 2022 until December of 2024. Participants were divided according to Hospital Anxiety and Depression Scale assessments as an anxiety-depression cohort (n = 122), or no such symptoms (control cohort, n = 148). Single-molecule array technology was used to quantify serum NfL, while enzyme-linked immunosorbent assay was used to measure serum GFAP. Age-stratified analyses were performed to examine the association of age groups and anxiety-depression prevalence. Using logistic regression to isolate risk factors, receiver operating characteristics curves were then used to quantify the model discriminative power.
Out of 270 enrolled patients (mean old: 73.1 ± 8.3 years; male proportion: 46.3%), we described a later encounter satisfaction in both anxiety and/or depressive symptoms which had been recognized in 122 individuals (45.2%). Using the log-transformed NfL and GFAP values, patients with anxiety-depression exhibited significantly higher levels of both proteins vs controls (32.8 ± 8.6 pg/mL vs 21.6 ± 6.4 pg/mL for NfL; μ = ± SD: 187.2 ± 43.2 pg/mL vs 127.8 ± 34.8 pg/mL for GFAP). Age-stratified analysis demonstrated a nonrandomized association between older age and anxiety-depression incidence (55-64: 32.8%, 65-74: 44.0%, 75-89: 54.5%; P for trend = 0.008). Two inde
Serum NfL and GFAP together would be particularly excellent predictors of anxiety and depression in people with AD. Age-stratified analysis shows an independent association between increasing age and vulnerability to anxiety-depression. Our composite biomarker panel provides a useful, robust and convenient means of identifying early AD patients at greater risk of anxiety and depression.
Core Tip: Anxiety and depression are common neuropsychiatric disorders that negatively influence prognosis and quality of life in patients with Alzheimer’s disease. Results from this study show that the combination of serum neuroaxonal damage marker and astrocytic activation marker, not the single markers, have improved prognostic value for anxiety and depression. Age-stratified analysis provides further evidence of a strong age-dependent gradient in anxiety-depression prevalence These findings highlight the role of thorough biomarker assessment in combination with age-related risk stratification in everyday clinical practice to identify and treat at-risk patients better and earlier for Alzheimer’s disease.
- Citation: Xu GW, Zhu XL, Yin CH. Serum NfL and GFAP levels and their association with anxiety-depression in Alzheimer’s disease patients: A clinical retrospective study. World J Psychiatry 2026; 16(9): 121407
- URL: https://www.wjgnet.com/2220-3206/full/v16/i9/121407.htm
- DOI: https://dx.doi.org/10.5498/wjp.121407
Alzheimer’s disease (AD) is the most common cause of dementia, affecting almost 50 million cases worldwide, and a major cause of disability in the elderly. Due to the rapidly rising number of AD patients, expected to reach 150 million in 2050 according to estimates from the World Health Organization[1], there will already be an enormous burden on healthcare systems globally. In China, the proportion of all-cause dementia among people aged 60 years and older is around 6% while specifically, AD accounts for about 3.9% in the same age stratum[1]; the absolute number of these cases has been estimated to grow by approximately 5%-6% per year due to aging population[2]. Although diagnostic and therapeutic approaches for both AD and NC have improved considerably over the last four decades, patients continue to face multiple neuropsychiatric complications, with anxiety and depression being two of the most common and important comorbidities in AD[3].
Anxiety and depression are mood disorders that occur in AD patients either concurrently or following the onset of cognitive impairment, with anxiety characterized by pervasive feelings of worry, depressed low mood (problems to fall asleep) and psycho-motor retardation[4]. Epidemiological studies show that overall rates of anxiety and depression in patients with AD vary considerably between studies, ranging from 30% to over 60%[5]. 25%-40% in early stage AD to 40%-70% in moderate to severe stages[6]. These symptoms not only greatly impact patients social functioning, but may contribute to a more rapid cognitive decline as well as increased caregiver burden and greater overall care needs[7]. Therefore, in order to improve the overall prognosis of patients by effectively targeting and managing depression during AD, understanding the complete landscape of factors associated with its development is essential. Of note is that the association between anxiety-depression and AD appeared to be bi- not unidirectional. Mid-life anxiety and depression are progressively being associated with an increased risk of subsequent cognitive decline and AD in that they are now framed as modifiable risk factors, while on the other hand, anxiety and depressive symptoms continue to be described as neuropsychiatric manifestations of evolving AD-related neurodegeneration itself, principally localised in limbic and prefrontal circuitry dedicated to affective regulation. This dual role as both earlier contributor and later expression of disease progression strengthens the argument for exploring circulating biomarkers indexing the neurodegenerative and neuroinflammatory substrate common to both processes. Standardized evaluation of anxiety and depressive symptoms in AD is typically done with instruments like the Hospital Anxiety and Depression Scale (HADS), the Geriatric Depression Scale, or scales based on caregiver input like neuropsychiatric inventory or cornell scale for depression in dementia, but each are confounded to some extent by reporter bias, overlap between cognitive impairment and symptom (in this case, self-report) vitality, as well as differences in regular caregiver availability; thus[8].
Latest evidence-based research regarding the neurobiology of anxiety and depression pathogenesis in AD largely focuses on three spheres; Neurobiological, neuroinflammatory and neurodegenerative[8]. Functional anomalies of serotonin and norepinephrine systems, and activation of the hypothalamic-pituitary-adrenal-axis represent neurobiological mechanisms chiefly associated with[9]. The rising recognition that neurodegeneration, and notably astrocytic activation, relates to the pathogenesis of AD-related psychiatric symptoms has emerged in recent years. Neurofilament light (NfL) chain is a specific neuroaxonal injury biomarker and has recently been identified as a potential blood biomarker of the severity of neurodegeneration[10]. Simultaneously, glial fibrillary acidic protein (GFAP), a marker of astrocytic activation and reactive gliosis, could also be implicated in the development of psychotic symptoms through neuroinflammatory mechanisms[11].
Neuronal injury, together with astrocytic activation, are important pathologic characteristics of AD pathogenesis. NfL is a neuronal cytoskeletal protein that becomes highly enriched in neurons and is released into the cerebrospinal fluid and blood following neuronal injury[12]. High plasma NfL levels have been consistently shown in patients with AD and correlate with the severity of disease and cognitive decline[13]. However, the role of NfL in the development of anxiety-depression is incompletely understood. GFAP is an intermediate filament protein that is specifically expressed in astrocytes and its expression increases during reactive gliosis and astrocytic activation[14]. Blood GFAP levels are associated with amyloid pathology and predict cognitive decline in AD patients[15]. The neuroinflammatory hypothesis of depression, as a dominant theoretical model in light of the pathology of mood disorders has been supported as a linking concept between depression and both degenerative diseases[16]. On integrating these observations, a consistent mechanistic chain can be proposed whereby neuroaxonal injury (as reflected in the releases of serum NfL from damaged neurons present within hippocampal, amygdalar and prefrontal circuits that mediate important aspects of affective regulation) and astrocytic activation (as assessed via upregulation of GFAP with downstream effects on pro-inflammatory cytokine release, glutamate homeostasis, and BDNF-mediated neurotrophic signaling) converge onto the underlying dysregulation of monoaminergic neurotransmission and function of hypothalamic–pituitary–adrenal axis which explains mood symptoms. This integrated pathway offers an a priori biological basis for the combined measurement of NfL and GFAP as potentially independent markers of the neurodegenerative and neuroinflammatory component, respectively, of anxiety-depression in AD.
While AD blood biomarker research is increasing rapidly, evidence relevant to the relationship between combined neurodegenerative and astrocytic markers with anxiety-depression in AD has been scarce[17]. Again, particularly few pedigree investigations have examined Chinese cohorts with some community-based studies having commenced mea
This is a retrospective study to evaluate serum NfL and GFAP as correlates of anxiety and depression in AD. The study was performed in compliance with the Declaration of Helsinki and was endorsed by the institutional Medical Ethics Committee, approval No. GB-HQ-00025.
Inpatient ADs admitted to the neurology department of our institution between January 2022 and December 2024 were reviewed. Eligibility criteria included: (1) Aged between 55 years and 90 years; (2) Probable AD as diagnosed by National Institute on Aging-Alzheimer’s Association criteria; (3) Mini-Mental State Examination (MMSE) score of 10-26; (4) Clinical and laboratory data with full records; (5) Availability of serum samples; and (6) A person who could undergo psychiatric assessment with sufficient assistance provided by a relative or family member. The exclusion criteria included: (1) Other forms of dementia; (2) Long-standing severe psychiatric disorders prior to the onset of AD; (3) Antidepressants or anxiolytic usage in the 3 months prior to the study visit; (4) Malignancies or any serious concomitant systemic diseases; (5) Ongoing infection, autoimmune process, or recent head trauma; (6) Very advanced clinical state precluding test administration or aphasia affecting verbal fluency tests; and (7) Absence of relevant clinical- or biomarker-related data.
Screening was conducted in 312 patients; 270 were eligible. We excluded those unable to continue HADS assessment (n = 18), with incomplete records (n = 14), without serum samples available (n = 7) or who refused (n = 3), leaving a study group of forty-two patients. Excluded patients were marginally older and had worse MMSE ratings than included ones, but age, sex, education and comorbidities were comparable across groups. The planned a priori sample-size calculation indicated that at least 236 patients were needed; thus, the final sample size was sufficient.
Serum NfL was measured using the single-molecule array NF-light Advantage Kit on the HD-X platform at Quanterix (Billerica, MA, United States) and serum GFAP was measured using an enzyme-linked immunosorbent assay kit from R&D Systems (Minneapolis, MN, United States). Samples were tested in duplicates by technicians who were blinded to the clinical data.
Established cut-offs for increased biomarkers were defined as NfL > 25 pg/mL and GFAP > 150 pg/mL according to previously reported studies in AD populations. We obtained similar optimal thresholds (27.1 pg/mL for NfL and 155.8 pg/mL for GFAP) via in-cohort receiver operating characteristic (ROC) analysis, and sensitivity analyses suggested comparative predictive performance between groups. During sample processing, venous blood in the fasting state was collected in the morning and centrifuged. The separated samples were then divided into aliquots for storage at -80 °C and no sample underwent more than one freeze-thaw cycle.
Anxiety and depressive symptom were measured at enrollment with the HADS. The scale consists of 14 items, including seven each for HADS-Anxiety subscale (HADS-A) and HADS-Depression subscale (HADS-D). A score between 0 and 3 is given to each item. A score of ≥ 8 in the subscale was deemed to be anxiety or depressive symptoms. Two trained neuropsychologists conducted assessments Convergent validity was examined with HADS-A and HADS-D demon
All statistical analyses were performed using SPSS 26.0. A two-sided P < 0.05 was determined a statistically significant result. Continuous variables are reported as mean ± SD or median (Q1, Q3) and categorical variables as n (%). For continuous variables and categorical variables, group comparisons were performed using the independent-samples t test or Mann-Whitney U test and χ2 test or Fisher’s exact test, respectively. Associations between biomarker levels and HADS scores were assessed using spearman correlation analysis. We grouped patients by three age brackets (55-64 years, 65-74 years and 75-89 years) and trend for anxiety-depression prevalence by age was evaluated using the Cochran-Armitage trend test. We used one-way analysis of variance with Bonferroni post hoc testing to compare age-stratified biomarker levels and HADS scores. Methods risk factors for anxiety and depression were defined using odds ratios (ORs) with 95% confidence intervals (CIs), calculated using logistic regression. Using ROC curves the predictive value of NfL, GFAP and their combination were evaluated. A combined biomarker model was created by linear-transforming continuous NfL and GFAP values into a logistic regression. Variance inflation factors were computed in order to diagnose multicollinearity. Hierarchical regression was also conducted to test the incremental predictive power of NfL and GFAP beyond cognitive severity. Sensitivity analysis excluding patients with previous psychotropic medication use dating back past the 3-month exclusion window was performed.
A total of 270 patients with AD were included, including 125 males (46.3%) and 145 females (53.7%) with an average age of 73.1 ± 8.3 years old. There were no differences in age (P > 0.05) or sex (P > 0.05) between groups. The overall mean MMSE score was 18.5 ± 4.2, the mean education duration was 9.8 ± 3.6 years, the mean disease duration was estimated at 2.83 ± 1.48 years and the mean clinical dementia rating (CDR) score was 1.2 ± 0.5 respectively. In total 146 patients (> 54.1%) had raised NfL (> 25 pg/mL), while also 142 (52.6%) were found to have raised GFAP (> 150 pg/mL). All biomarker levels were higher in the anxiety-depression group compared with the control group (all P < 0.001, Table 1). We called the two groups AD patients with coexisting anxiety-depression (n = 122) and AD patients without coexisting anxiety-depression (n = 148). While education and disease duration was different between groups but not statistically significant, the MMSE and CDR were both significantly different (both P < 0.001) across groups. In light of this, in the univariate analysis education and disease duration were included.
| Characteristics | Total (n = 270) | AD-Control (n = 148) | AD-AnxDep (n = 122) |
| Age (years) | 73.1 ± 8.3 | 72.3 ± 8.0 | 74.1 ± 8.7 |
| Male gender | 125 (46.3) | 70 (47.3) | 55 (45.1) |
| Education (years) | 9.8 ± 3.6 | 10.2 ± 3.8 | 9.3 ± 3.4 |
| Disease duration (years) | 2.8 ± 1.4 | 2.6 ± 1.3 | 3.1 ± 1.5 |
| MMSE score | 18.5 ± 4.2 | 19.5 ± 4.1 | 17.3 ± 4.2 |
| CDR score | 1.2 ± 0.5 | 1.1 ± 0.4 | 1.4 ± 0.5 |
| Hypertension | 156 (57.8) | 82 (55.4) | 74 (60.7) |
| Diabetes mellitus | 89 (33.0) | 46 (31.1) | 43 (35.2) |
| Hyperlipidemia | 112 (41.5) | 58 (39.2) | 54 (44.3) |
| NfL (pg/mL) | 26.7 ± 9.4 | 21.6 ± 6.4 | 32.8 ± 8.6 |
| GFAP (pg/mL) | 155.6 ± 47.2 | 127.8 ± 34.8 | 187.2 ± 43.2 |
Of the 270 AD patients, 122 met criteria for anxiety and/or depression resulting in an overall incidence of 45.2%. The scores of HADS-A ranged from 0 point to 18 points, and mean was 7.3 ± 3.8 points Scores for HADS-D were between 0 points and 19 points (mean = 7.9 ± 4.1 points). Among the 122 patients with anxiety-depression, 51 had isolated anxiety (Generalized Anxiety Disorder-7a ≥ 8), 40 isolated depression (Patient Health Questionnaire-9b ≥ 8) and 31 both (comorbidity; both measures ≥ 8). The mean total HADS score of anxiety-depression patients was 19.0 ± 5.5 points, significantly higher than that of non-anxiety-depression patients (8.6 ± 3.8 points, P < 0.001). The biomarker concentrations in anxiety-depression group were significantly increased (Figure 1).
Significant positive correlations were detected between both biomarkers and HADS total scores. NfL levels were moderately positively correlated with total HADS scores (r = 0.516, P < 0.001). Concentrations of GFAP correlated in a similar manner (r = 0.492; P < 0.001). At the single-subscale level, for both biomarkers HADS-D scores correlated more strongly than HADS-A scores, indicating a closer relationship with depressive compared to anxiety symptoms (Table 2).
| Biomarker | HADS-A (r) | P value | HADS-D (r) | P value | HADS-total (r) |
| NfL (pg/mL) | 0.448 | < 0.001 | 0.562 | < 0.001 | 0.516 |
| GFAP (pg/mL) | 0.422 | < 0.001 | 0.528 | < 0.001 | 0.492 |
| Age | 0.186 | 0.002 | 0.198 | 0.001 | 0.194 |
| MMSE score | -0.312 | < 0.001 | -0.346 | < 0.001 | -0.332 |
| CDR score | 0.298 | < 0.001 | 0.324 | < 0.001 | 0.314 |
| Disease duration | 0.224 | < 0.001 | 0.256 | < 0.001 | 0.242 |
In univariate logistic regression analyses, anxiety-depression was found to be related to various factors. In our univariate analysis, the significant predictors were elevated NfL [NfL > 25 pg/mL; OR = 3.28 (95%CI: 1.64-6.56), P = 0.001], elevated GFAP [GFAP > 150 pg/mL; OR = 2.94 (95%CI: 1.50-5.76), P = 0.002], MMSE ≤ 18 (OR = 2.28, P = 0.010), and CDR ≥ 1.5 (OR = 2.04, 95%CI: 1.06-4.35). Using multivariate analysis, high NfL (OR = 3.28, P = 0.001), elevated GFAP (OR = 2.92, P = 0.002) and MMSE ≤ 18 (OR = 2.12, P = 0.029) remained independent predictors (Table 3; Figure 2). Variance inflation factors were all < 2.1, indicating no critical multicollinearity. Multiplicative regression chains indicated that NfL and GFAP-based classification plummeted discrimination Δ area under the curve (AUC) = 0.094, P < 0.001 beyond MMSE and CDR alone. Variables in multivariate analysis included education ≤ 9 years and disease duration > 3 years, which were not independently significant after adjustment. The sensitivity analysis excluding patients with prior exposure to psychotropic medication demonstrated similar findings, with little change in ORs or AUC for the combined model.
| Risk factor | Odds ratio (95%CI) | P value |
| Age ≥ 75 years | 1.88 (0.98-3.62) | 0.058 |
| Female gender | 1.64 (0.88-3.06) | 0.119 |
| Education ≤ 9 years | 1.86 (0.96-3.60) | 0.066 |
| Disease duration > 3 years | 1.74 (0.92-3.30) | 0.089 |
| MMSE ≤ 18 | 2.28 (1.22-4.28) | 0.010 |
| CDR ≥ 1.5 | 2.04 (1.06-3.92) | 0.033 |
| Hypertension | 1.24 (0.68-2.26) | 0.481 |
| Diabetes mellitus | 1.20 (0.66-2.18) | 0.548 |
| Elevated NfL (> 25 pg/mL) | 3.28 (1.64-6.56) | 0.001 |
| Elevated GFAP (> 150 pg/mL) | 2.94 (1.50-5.76) | 0.002 |
Individual biomarker performance: NfL (AUC 0.772, 95%CI: 0.714-0.830, sensitivity 71.3%, specificity 76.4% at cut-off for positive ≥ 27.1 pg/mL), GFAP (AUC 0.758, 95%CI: 0.698-0.818, sensitivity: 72.1%, specificity: 73.6%, cut-off for positive ≥ 155.8 pg/mL). Performance of combined model was superior from individual markers (AUC 0.882; 95%CI: 0.836-0.928; P < 0.001 vs each marker) at the cost of 81.1% sensitivity and specificity of 84.5%.
Elevations in serum biomarker were stratified into three groups: No abnormalities (count = 0), 1 protein, or ≥ 2 proteins. The number of elevated anxiety-depression incidence rose sequentially: 0 (13.8%), 1 elevated (36.9%), and 2 elevated (76.2%). OR was 3.42 (1.38-8.48) for one elevated biomarker and 19.58 (7.32-52.34) for both compared with no elevated biomarkers (P for trend < 0.001; Table 4).
| Number of elevated biomarkers | n | Anxiety-depression incidence (%) | Mean HADS | OR (95%CI) | P value |
| 0 (low risk) | 58 | 13.8 | 6.2 ± 2.8 | Reference | - |
| 1 (intermediate risk) | 130 | 36.9 | 11.8 ± 4.6 | 3.42 (1.38-8.48) | 0.008 |
| 2 (high risk) | 82 | 76.2 | 18.4 ± 5.2 | 19.58 (7.32-52.34) | < 0.001 |
| P for trend | - | - | - | - | < 0.001 |
Results were also robust to other definitions and subgroups. Even with a more stringent criteria (HADS ≥ 10), accuracy remained high (AUC 0.868). Gender-stratified analysis showed no interaction (males: 0.876; females: 0.886; P = 0.758). Moderate dementia (AUC 0.902) had stronger performance than mild dementia (AUC 0.848). Results were consistent across stratifications of separately education level and duration of disease (Table 5).
| Analysis type | Subgroup | AUC (95%CI) | Sensitivity/specificity |
| Primary analysis | Overall (n = 270) | 0.882 (0.836-0.928) | 81.1%/84.5% |
| Stricter criteria | HADS ≥ 10 | 0.868 (0.818-0.918) | 79.6%/83.2% |
| Gender-stratified | Male (n = 125) | 0.876 (0.814-0.938) | 80.0%/84.3% |
| Gender-stratified | Female (n = 145) | 0.886 (0.834-0.938) | 82.1%/84.7% |
| Cognitive severity | Moderate (MMSE 10-18) | 0.902 (0.858-0.946) | 83.8%/86.2% |
| Cognitive severity | Mild (MMSE 19-26) | 0.848 (0.786-0.910) | 77.4%/81.9% |
| Education-stratified | Low education (≤ 9 years) | 0.874 (0.822-0.926) | 79.8%/83.6% |
| Education-stratified | High education (> 9 years) | 0.888 (0.836-0.940) | 82.4%/85.2% |
| Disease duration | Early stage (≤ 3 years) | 0.876 (0.820-0.932) | 80.3%/83.8% |
| Disease duration | Later stage (> 3 years) | 0.890 (0.832-0.948) | 81.9%/85.1% |
This retrospective study of 270 AD patients provides important evidence for the predictive value of serum NfL with GFAP in anxiety and depression prediction in AD. The implications of these results for clinical practice are considerable and call for a more holistic approach to AD management that considers both cognitive and neuropsychiatric dimensions of disease. We propose three possible mechanistic pathways: (1) Neurodegeneration and the activation of astrocytes may be common upstream triggers, where cognitive impairment by neurodegenerative processes consequently affects monoaminergic neurotransmission; (2) Both anxiety and depression can be drivers of AD where exacerbation of the two comorbidities leads to worsening cognitive function; and (3) AD-associated neurodegeneration might also impact limbic system structures in a more direct manner promoting a vicious cycle that exacerbates cognitive and emotional outcomes[21]. Based on these observations, and because biomarker measurement in this study was retrospective/cross-sectionally, we present these three pathways as reasonable hypotheses rather than proven mechanisms and denote our findings as associations with anxiety-depression in AD - not as causal or mechanistic relationships between NfL/GFAP elevation. There are two other explanations, which deserve to be made explicit. The association direction may be reversed (reverse causality), for example chronic anxiety and depressive symptoms might trigger neurodegeneration themselves via prolonged dysregulated hypothalamus-pituitary-adrenal axis, glucocorticoid-mediated hippocampal atrophy or a stronger neuroinflammatory tone thus rising serum NfL and GFAP secondarily. Second, apparent associations may be confounded by unmeasured variables influencing biomarker levels (e.g., cerebrovascular burden or sleep disturbance) and mood symptoms, which are not mutually exclusive categories. Longitudinal studies with multiple measurements of a serial biomarker will be needed to disentangle the direction of effect and adjudicate between these competing hypotheses.
High blood levels of NfL therefore reflect active neurodegeneration, potentially implicating networks involved with the regulation of emotion such as limbic and prefrontal systems[22]. These neurodegenerative changes can impact distributed circuits subserving emotional processing and stress-responses[23]. Additionally, NfL may reflect a more global injury to axons and white matter, thereby disconnecting brain regions that are involved in affective processing[24]. In the context of AD, astrogliosis and neuroinflammation are evidenced by increased expression levels of GFAP in astrocytes. Among the mechanisms in which activated astrocytes could participate to exacerbate anxiety-depression are: Pro-inflammatory cytokine release, glutamate homeostasis dysfunction and altered neurotrophic support[25]. The improved performance of the combined NfL-GFAP panel over that of either marker on its own supports the notion that they give complementary rather than redundant information[26].
A significant contribution of this study is the age stratum analysis[27]. Of note, the strong age group by biomarker status interaction provides evidence that that risk based on biomarker class may be especially informative for older age groups. This finding has implications for clinical risk stratification in biomarker-informed AD care[28].
The global rate of anxiety-depression in this study (45.2%) is consistent with earlier meta-analyses showing rates between 30% and 60%[5]. Our finding that combined biomarkers, instead of single markers, deliver better prediction, is also reminiscent of the field’s general push towards multi-marker biological characterization of AD[29]. In a clinical implementation perspective, the NfL-GFAP assessment as a joint screening. Accordingly, we suggest a risk stratification framework based on an increase in the number of elevated biomarkers consistent with those emerging multi-biomarker stratification models that have been proposed across AD research[30,31]. Structured anxiety-depression screening for high-risk patients at regular intervals, early psychological intervention, and possibly prophylactic treatment should be based on this. Early identification of patients at high risk may help optimize the allocation of mental health resources, which may be crucial in healthcare contexts with limited psychiatric resources.
There are various aspects that make the current work different from previous literature. This is one of the first studies, to our knowledge, that assesses the prediction capability of serum NfL and GFAP for the coexisting utilizing both biomarkers in a Chinese AD cohort. The NfL-GFAP combined panel was found to differentiate much more discriminatively (AUC = 0.882) than isolated biomarkers (ΔAUC = 0.110 vs NfL and ΔAUC = 0.124 vs GFAP; both P < 0.001), supporting the idea that neurodegeneration and neuroinflammation signals provide complementary rather than re
Practical implications: A management pathway with different implementation levels. Based on these findings, we established a novel tiered management pathway with diagnostic thresholds that are guided by the number of elevated biomarkers. For the high-risk patients (both NfL and GFAP elevated; anxiety-depression observed incidence 76.2%), we recommend structured HADS screening every 3 months, early referral to geriatric psychiatry, first-line non-pharmacological interventions including cognitive-behavioral therapy adapted for cognitive impairment, reminiscence therapy and caregiver psychoeducation, and where clinically indicated cautiously initiating selective serotonin reuptake inhibitors (preferentially sertraline or escitalopram given their favorable profile at least in elderly cognitively impaired patients). For our intermediate-risk patients (only one biomarker failed to meet the threshold for end-stage disease; incidence 36.9%), we would also recommend 6-monthly HADS screening as well as a low-intensity psychosocial support program be offered. Low-risk patients (none of the biomarkers elevated; incidence 13.8%) can be assumed to have an adequate standard of care and therefore are only routinely screened annually.
Several limitations should be acknowledged: (1) The retrospective nature of the design limits conclusions about temporality and causality; (2) Our single-center analysis performed in a Chinese population may limit the generalizability; (3) We used HADS to assess the psychiatric comorbidity but this may be rather specific and dismisses broader aspects; (4) There were no serial readings of biomarkers; and (5) Residual confounding by medications and other unmeasured factors may be possible. These results need to be confirmed by future prospective longitudinal studies with a larger sample size and heterogeneous populations of different sociodemographic characteristics[32].
We can classify these restrictions into three categories to better delineate what their ramifications are: (1) Study-design limitations: The retrospective, cross-sectional design prohibits inference about the temporal direction of the NfL/GFAP–mood association; a single biomarker time-point cannot capture trajectory; (2) Measurement limitations: HADS instead of structured psychiatric interview to assess the potential misclassification of borderline cases; and clinical case-by-case checking, while reducing potential variability from normal variation in homes due to caregiver-informant input assessing only a minority of items implicating some inter-rater agreement; and (3) Generalizability limitations: The cohort was selected from a single tertiary centre in northeast China, thus absolute estimates of prevalence and threshold per
Serum NfL and GFAP levels, both individually and in combination, have good predictive values for anxiety and depression in patients with AD. This study represents the first work to demonstrate this relationship. In the integrated biomarker model, overall discriminative performance was superior (AUC = 0.882) compared with using individual biomarkers alone. A strong age-dependent gradient for an anxiety-depression diagnosis was also confirmed on age-stratified analysis, as the positive predictive value of biomarker elevation is modified by patient age. Our risk stratification paradigm according to the number of elevated biomarkers provides a clinically translatable framework.
| 1. | Scheltens P, De Strooper B, Kivipelto M, Holstege H, Chételat G, Teunissen CE, Cummings J, van der Flier WM. Alzheimer's disease. Lancet. 2021;397:1577-1590. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 4504] [Cited by in RCA: 3880] [Article Influence: 776.0] [Reference Citation Analysis (0)] |
| 2. | Jia L, Du Y, Chu L, Zhang Z, Li F, Lyu D, Li Y, Li Y, Zhu M, Jiao H, Song Y, Shi Y, Zhang H, Gong M, Wei C, Tang Y, Fang B, Guo D, Wang F, Zhou A, Chu C, Zuo X, Yu Y, Yuan Q, Wang W, Li F, Shi S, Yang H, Zhou C, Liao Z, Lv Y, Li Y, Kan M, Zhao H, Wang S, Yang S, Li H, Liu Z, Wang Q, Qin W, Jia J; COAST Group. Prevalence, risk factors, and management of dementia and mild cognitive impairment in adults aged 60 years or older in China: a cross-sectional study. Lancet Public Health. 2020;5:e661-e671. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 1453] [Cited by in RCA: 1272] [Article Influence: 212.0] [Reference Citation Analysis (0)] |
| 3. | Pless A, Ware D, Saggu S, Rehman H, Morgan J, Wang Q. Understanding neuropsychiatric symptoms in Alzheimer's disease: challenges and advances in diagnosis and treatment. Front Neurosci. 2023;17:1263771. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 12] [Cited by in RCA: 68] [Article Influence: 22.7] [Reference Citation Analysis (0)] |
| 4. | Teixeira AL, Rocha NP, Gatchel J. Behavioral or neuropsychiatric symptoms of Alzheimer's disease: from psychopathology to pharmacological management. Arq Neuropsiquiatr. 2023;81:1152-1162. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 23] [Cited by in RCA: 19] [Article Influence: 6.3] [Reference Citation Analysis (0)] |
| 5. | Zhao QF, Tan L, Wang HF, Jiang T, Tan MS, Tan L, Xu W, Li JQ, Wang J, Lai TJ, Yu JT. The prevalence of neuropsychiatric symptoms in Alzheimer's disease: Systematic review and meta-analysis. J Affect Disord. 2016;190:264-271. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 768] [Cited by in RCA: 660] [Article Influence: 66.0] [Reference Citation Analysis (0)] |
| 6. | Lee CH, Giuliani F. The Role of Inflammation in Depression and Fatigue. Front Immunol. 2019;10:1696. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 514] [Cited by in RCA: 448] [Article Influence: 64.0] [Reference Citation Analysis (0)] |
| 7. | Connors MH, Seeher K, Teixeira-Pinto A, Woodward M, Ames D, Brodaty H. Dementia and caregiver burden: A three-year longitudinal study. Int J Geriatr Psychiatry. 2020;35:250-258. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 59] [Cited by in RCA: 170] [Article Influence: 28.3] [Reference Citation Analysis (0)] |
| 8. | Jy Kang M, Grewal J, Eratne D, Malpas C, Chiu WH, Katisko K, Solje E, Santillo AF, Mitchell PB, Hopwood M, Velakoulis D. Neurofilament light and glial fibrillary acidic protein in mood and anxiety disorders: A systematic review and meta-analysis. Brain Behav Immun. 2025;123:1091-1102. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 1] [Cited by in RCA: 11] [Article Influence: 11.0] [Reference Citation Analysis (0)] |
| 9. | Miller AH, Raison CL. The role of inflammation in depression: from evolutionary imperative to modern treatment target. Nat Rev Immunol. 2016;16:22-34. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 3262] [Cited by in RCA: 2964] [Article Influence: 296.4] [Reference Citation Analysis (4)] |
| 10. | Bao Y, Zhang Z, Yu H, Huo X, Lu Y, Xu T. Development of Intelligent Acupuncture Applications and Related Technologies. World J Tradit Chin Med. 2023;9:21-28. [RCA] [DOI] [Full Text] [Cited by in RCA: 4] [Reference Citation Analysis (0)] |
| 11. | Yang Z, Sreenivasan K, Toledano Strom EN, Osse AML, Pasia LG, Cosme CG, Mugosa MRN, Chevalier EL, Ritter A, Miller JB, Cordes D, Cummings JL, Kinney JW. Clinical and biological relevance of glial fibrillary acidic protein in Alzheimer's disease. Alzheimers Res Ther. 2023;15:190. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 1] [Cited by in RCA: 30] [Article Influence: 10.0] [Reference Citation Analysis (0)] |
| 12. | Gaetani L, Blennow K, Calabresi P, Di Filippo M, Parnetti L, Zetterberg H. Neurofilament light chain as a biomarker in neurological disorders. J Neurol Neurosurg Psychiatry. 2019;90:870-881. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 394] [Cited by in RCA: 1017] [Article Influence: 145.3] [Reference Citation Analysis (0)] |
| 13. | Mattsson N, Andreasson U, Zetterberg H, Blennow K; Alzheimer’s Disease Neuroimaging Initiative. Association of Plasma Neurofilament Light With Neurodegeneration in Patients With Alzheimer Disease. JAMA Neurol. 2017;74:557-566. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 482] [Cited by in RCA: 848] [Article Influence: 94.2] [Reference Citation Analysis (0)] |
| 14. | Abdelhak A, Foschi M, Abu-Rumeileh S, Yue JK, D'Anna L, Huss A, Oeckl P, Ludolph AC, Kuhle J, Petzold A, Manley GT, Green AJ, Otto M, Tumani H. Blood GFAP as an emerging biomarker in brain and spinal cord disorders. Nat Rev Neurol. 2022;18:158-172. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 29] [Cited by in RCA: 591] [Article Influence: 147.8] [Reference Citation Analysis (0)] |
| 15. | Pereira JB, Janelidze S, Smith R, Mattsson-Carlgren N, Palmqvist S, Teunissen CE, Zetterberg H, Stomrud E, Ashton NJ, Blennow K, Hansson O. Plasma GFAP is an early marker of amyloid-β but not tau pathology in Alzheimer's disease. Brain. 2021;144:3505-3516. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 195] [Cited by in RCA: 478] [Article Influence: 95.6] [Reference Citation Analysis (0)] |
| 16. | Köhler CA, Freitas TH, Maes M, de Andrade NQ, Liu CS, Fernandes BS, Stubbs B, Solmi M, Veronese N, Herrmann N, Raison CL, Miller BJ, Lanctôt KL, Carvalho AF. Peripheral cytokine and chemokine alterations in depression: a meta-analysis of 82 studies. Acta Psychiatr Scand. 2017;135:373-387. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 1266] [Cited by in RCA: 1139] [Article Influence: 126.6] [Reference Citation Analysis (3)] |
| 17. | Wu H, Dai W, Wang L, Zhang J, Wang C. Comprehensive Analysis of the Molecular Mechanism for Gastric Cancer Based on Competitive Endogenous RNA Network. World Journal of Traditional Chinese Medicine. World J Tradit Chin Med. 2023;9:29-42. [RCA] [DOI] [Full Text] [Cited by in RCA: 10] [Reference Citation Analysis (0)] |
| 18. | Liang C, Yan X, Tian J, Yang Y, Xiao X, Huang Y, Wang T. Plasma GFAP, NfL, and p-tau181 levels as early biomarkers of dementia in Chinese adults: Shenzhen community cohort study. Aging Clin Exp Res. 2025;37:105. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in RCA: 3] [Reference Citation Analysis (0)] |
| 19. | Kjaergaard D, Simonsen AH, Waldemar G, Nielsen TR. Ethnic and racial influences on blood biomarkers for Alzheimer's disease: A systematic review. J Alzheimers Dis. 2025;103:81-91. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 20] [Reference Citation Analysis (0)] |
| 20. | Teunissen CE, Verberk IMW, Thijssen EH, Vermunt L, Hansson O, Zetterberg H, van der Flier WM, Mielke MM, Del Campo M. Blood-based biomarkers for Alzheimer's disease: towards clinical implementation. Lancet Neurol. 2022;21:66-77. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 64] [Cited by in RCA: 718] [Article Influence: 143.6] [Reference Citation Analysis (0)] |
| 21. | Leng F, Edison P. Neuroinflammation and microglial activation in Alzheimer disease: where do we go from here? Nat Rev Neurol. 2021;17:157-172. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 640] [Cited by in RCA: 2347] [Article Influence: 469.4] [Reference Citation Analysis (4)] |
| 22. | Chen Y, Dang M, Zhang Z. Brain mechanisms underlying neuropsychiatric symptoms in Alzheimer's disease: a systematic review of symptom-general and -specific lesion patterns. Mol Neurodegener. 2021;16:38. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 183] [Cited by in RCA: 157] [Article Influence: 31.4] [Reference Citation Analysis (0)] |
| 23. | Zhang NK, Zhang SK, Zhang LI, Tao HW, Zhang GW. The neural basis of neuropsychiatric symptoms in Alzheimer's disease. Front Aging Neurosci. 2024;16:1487875. [RCA] [PubMed] [DOI] [Full Text] [Cited by in RCA: 18] [Reference Citation Analysis (0)] |
| 24. | Upreti S, Prusty JS, Kumar A, Samant M. Identification of SARS-CoV-2 Spike Protein Inhibitors from Urtica Dioica to Develop Herbal-Based Therapeutics Against COVID-19. World J Tradit Chin Med. 2023;9:61-70. [RCA] [DOI] [Full Text] [Cited by in RCA: 5] [Reference Citation Analysis (0)] |
| 25. | Kumar A, Fontana IC, Nordberg A. Reactive astrogliosis: A friend or foe in the pathogenesis of Alzheimer's disease. J Neurochem. 2023;164:309-324. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 15] [Cited by in RCA: 106] [Article Influence: 35.3] [Reference Citation Analysis (0)] |
| 26. | Suresh S, Maffei L, Bauermeister S, Raymont V. Blood biomarkers for diagnosis and differential diagnosis of Alzheimer's disease in real-world clinical populations: A systematic review. J Alzheimers Dis. 2026;109:1574-1589. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 4] [Cited by in RCA: 2] [Article Influence: 2.0] [Reference Citation Analysis (0)] |
| 27. | Rose DK, Lyketsos CG, Rosenberg PB, Nowrangi MA. Neuropsychiatric symptoms in Alzheimer's disease: Bridging mechanisms, management, and emerging innovations. Neurotherapeutics. 2026;23:e00788. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 2] [Cited by in RCA: 2] [Article Influence: 2.0] [Reference Citation Analysis (0)] |
| 28. | Hansson O, Edelmayer RM, Boxer AL, Carrillo MC, Mielke MM, Rabinovici GD, Salloway S, Sperling R, Zetterberg H, Teunissen CE. The Alzheimer's Association appropriate use recommendations for blood biomarkers in Alzheimer's disease. Alzheimers Dement. 2022;18:2669-2686. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 234] [Cited by in RCA: 473] [Article Influence: 118.3] [Reference Citation Analysis (0)] |
| 29. | Jack CR Jr, Bennett DA, Blennow K, Carrillo MC, Dunn B, Haeberlein SB, Holtzman DM, Jagust W, Jessen F, Karlawish J, Liu E, Molinuevo JL, Montine T, Phelps C, Rankin KP, Rowe CC, Scheltens P, Siemers E, Snyder HM, Sperling R; Contributors. NIA-AA Research Framework: Toward a biological definition of Alzheimer's disease. Alzheimers Dement. 2018;14:535-562. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 9052] [Cited by in RCA: 7978] [Article Influence: 997.3] [Reference Citation Analysis (0)] |
| 30. | Pan FF, Huang L, Wang Y, Huang Q, Guan YH, Li YH, Xie F, Guo QH. Plasma P-tau217, GFAP, and NfL as biomarkers for Alzheimer's disease: role in disease stratification, pathological progression, and cognitive decline. Alzheimers Dement. 2025;21:e70987. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 1] [Cited by in RCA: 4] [Article Influence: 4.0] [Reference Citation Analysis (0)] |
| 31. | Livingston G, Huntley J, Sommerlad A, Ames D, Ballard C, Banerjee S, Brayne C, Burns A, Cohen-Mansfield J, Cooper C, Costafreda SG, Dias A, Fox N, Gitlin LN, Howard R, Kales HC, Kivimäki M, Larson EB, Ogunniyi A, Orgeta V, Ritchie K, Rockwood K, Sampson EL, Samus Q, Schneider LS, Selbæk G, Teri L, Mukadam N. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet. 2020;396:413-446. [RCA] [PubMed] [DOI] [Full Text] [Cited by in Crossref: 8647] [Cited by in RCA: 7338] [Article Influence: 1223.0] [Reference Citation Analysis (4)] |
| 32. | Angioni D, Delrieu J, Hansson O, Fillit H, Aisen P, Cummings J, Sims JR, Braunstein JB, Sabbagh M, Bittner T, Pontecorvo M, Bozeat S, Dage JL, Largent E, Mattke S, Correa O, Gutierrez Robledo LM, Baldivieso V, Willis DR, Atri A, Bateman RJ, Ousset PJ, Vellas B, Weiner M. Blood Biomarkers from Research Use to Clinical Practice: What Must Be Done? A Report from the EU/US CTAD Task Force. J Prev Alzheimers Dis. 2022;9:569-579. [RCA] [PubMed] [DOI] [Full Text] [Full Text (PDF)] [Cited by in Crossref: 18] [Cited by in RCA: 79] [Article Influence: 19.8] [Reference Citation Analysis (0)] |