Validation of a plasma GFAP immunoassay and establishment of age-related reference values: bridging analytical performance and routine implementation
-
Burak Arslan
,
Ulf Andreasson
Abstract
Objectives
Glial fibrillary acidic protein (GFAP) is a well-established biomarker of astrocytic activation associated with neurodegenerative diseases, neuroinflammatory disorders, and traumatic brain injury. With increasing interest in blood-based biomarkers, the need for analytically validated assays and reliable reference intervals is critical for routine clinical implementation. This study aimed to analytically validate the MSD S-Plex® GFAP immunoassay for plasma and to establish age-stratified reference intervals in an apparently healthy population.
Methods
This study was conducted in two phases. First, key analytical validation parameters – including repeatability, intermediate precision, measurement range, interferences, and sample stability – were evaluated following Clinical and Laboratory Standards Institute (CLSI) and published protocol guidelines. Second, reference intervals were derived from 579 apparently healthy individuals aged 17–91 years using a right-sided non-parametric percentile method. Age-specific upper reference limits were calculated for three predefined age groups, and a continuous age-dependent centile model was applied.
Results
MSD S-Plex® GFAP assay demonstrated strong analytical performance, with coefficients of variation for repeatability and intermediate precision below 12 %. After accounting for the 1:2 dilution ratio, the validated measurement range was 0.425–1760 ng/L, with all calibration residuals remaining within ±15 %. GFAP concentrations were unaffected by hemolysis (p=0.85) and remained stable for up to 7 days at 4 °C and under frozen storage conditions. Age-stratified upper reference limits for plasma GFAP were established as 38 pg/mL (18–<50 years), 73 pg/mL (≥50–<70 years), and 156 pg/mL (≥70 years). Additionally, sex-related differences were observed after age 50, with females showing higher absolute GFAP levels than males. A strong positive correlation between age and plasma GFAP levels was observed (Spearman’s r=0.832, p<0.0001).
Conclusions
This study demonstrates the robust analytical performance of the MSD S-Plex® GFAP assay and establishes age-related reference values for plasma GFAP. These findings support its suitability for routine clinical use and enhance its applicability in the diagnosis and monitoring of central nervous system (CNS) pathologies, such as neurodegenerative diseases, neuroinflammatory disorders, and acute brain injuries, within biomarker-supported clinical algorithms.
Introduction
Glial fibrillary acidic protein (GFAP) is a type III intermediate filament protein predominantly expressed in astrocytes, a type of glial cell that provides structural support and maintenance functions (e.g., maintaining the blood-brain barrier, responding to neuronal injury) within the central nervous system (CNS) [1]. Pathological events such as neurodegeneration and neuronal injury trigger astrocyte activation, resulting in increased GFAP expression, a well-established indicator of reactive astrogliosis. Beyond its normal turnover, GFAP is released into the extracellular space during astrogliosis, eventually reaching the cerebrospinal fluid (CSF) and bloodstream. GFAP concentrations in these biofluids can be quantified using various types of immunoassays, including enzyme-linked immunosorbent assay (ELISA) [2], ultra-sensitive single molecule array (Simoa) [3], microfluidic platforms like Ella [4], and fully-automated platforms such as Lumipulse®G immunoassay platform [5]. However, due to its limited sensitivity, ELISA is predominantly utilized for CSF measurements [6], [7], [8]. Consequently, the development of highly sensitive assays has enabled the reliable detection of low GFAP concentrations in blood, despite the analytical challenges posed by strong matrix effects in this biofluid.
Numerous studies have found that blood GFAP concentrations are elevated and associated with various pathological conditions, such as neurodegenerative diseases (e.g., Alzheimer’s disease (AD) [3]), neuroinflammatory disorders (e.g., multiple sclerosis (MS) [9]), and traumatic brain injury (TBI) [7]). In the context of AD, blood GFAP demonstrates superior discriminative capability across the AD continuum [3] compared to its CSF counterpart, likely due to the partial direct release of GFAP from astrocytic end feet into the bloodstream rather than the cerebrospinal fluid [9]. This interpretation is further supported by the observation that CSF and blood GFAP concentrations are only poorly to moderately correlated [3], 10], 11], suggesting partially distinct sources and kinetics of GFAP release between these compartments. Another plausible explanation is that, in most cohorts, the stored aliquots used for measuring plasma and CSF GFAP were subjected to multiple freeze–thaw cycles. On this basis, previous studies have shown that repeated freeze–thawing can negatively affect CSF GFAP levels [12], which may contribute to the weak correlation observed between plasma and CSF GFAP. Beyond AD, blood GFAP has also emerged as a marker of MS severity [9] and shows promise as a biomarker of disease activity and disability in neuromyelitis optica spectrum disorders (NMOSD) [13]. In addition, it is currently employed in clinical practice to aid in the diagnosis of mild TBI, with FDA approval supporting its use in this context [14], 15]. Collectively, these findings underscore the potential of blood GFAP as a valuable tool for supporting clinical diagnosis across a wide range of neurological disorders.
Given this context, since plasma GFAP is a valuable marker of astrogliosis and may support the biomarker-supported clinical diagnosis of various neurodegenerative and neuroinflammatory conditions, an important next step is to establish its concentration in a healthy reference population. Defining reference intervals is essential for distinguishing pathological elevations from normal biological variation. Moreover, for successful routine implementation, it is critical to identify which immunoassay platform is most suitable for a given laboratory, taking into account factors such as analytical performance, sensitivity, sample volume requirements, and operational feasibility. Although several sensitive immunoassays are available for quantifying GFAP in blood [4], [16], [17], [18], selecting and thoroughly evaluating one platform prior to clinical implementation is necessary. One such platform is the S-Plex Human GFAP assay developed by Meso Scale Discovery (MSD) – an ultrasensitive singleplex electrochemiluminescence (ECL)-based immunoassay that allows precise quantification of GFAP using minimal sample volumes.
We first conducted an analytical validation of MSD S-PLEX® Human GFAP assay to evaluate its performance prior to potential implementation for routine plasma GFAP measurement. Subsequently, we established a reference interval for plasma GFAP concentrations using a cohort of apparently healthy individuals spanning a broad age range.
Materials and methods
Study design
This study was conducted in two phases. In the first phase, we performed an analytical validation of MSD S-PLEX® Human GFAP assay. Key validation parameters – including repeatability, intermediate precision, measurement range, potential interferences, and sample stability – were evaluated in accordance with the intended clinical application of the biomarker. For this purpose, K2 ethylenediaminetetraacetic acid (EDTA) plasma leftover samples were randomly collected and processed at the Neurochemistry and Clinical Chemistry Laboratories of Sahlgrenska University Hospital. All samples were anonymized, and no clinical or demographic information was accessible. In the second phase, we established a reference range for plasma GFAP concentrations using samples from healthy individuals spanning a broad age range (17–91 years). These samples included leftover specimens from routine analyses at the Neurochemistry Laboratory and biobanked samples from various cohorts.
Ethics approval and consent to participate
This information is provided under the “Research ethics” section.
Study participants
For the technical validation, whole blood was collected in K2EDTA tubes from anonymized individuals at Sahlgrenska University Hospital in Mölndal, with no access to clinical or demographic data. Following centrifugation at 2000×g for 10 min, plasma was separated, aliquoted, and stored at −80 °C until analysis. Three plasma pools (high, intermediate, and low concentration) were prepared by combining leftover anonymized K2EDTA plasma samples at the Neurochemistry Laboratory at Sahlgrenska University Hospital. Prior to the validation experiments, the pools were analyzed to ensure their GFAP concentrations fell within the desired range. These verified pools were then used for precision assessment. Reference interval determination was conducted using de-identified/anonymized leftover plasma samples obtained from the Neurochemistry Laboratory at Sahlgrenska University Hospital and from various cohorts. Detailed descriptions of these cohorts are provided in Table 1. The total dataset for reference interval determination included EDTA plasma from 579 healthy individuals, aged 17–91 years. The samples were sourced as follows:
DDI study – A Norwegian national multicenter study: 100 individuals, aged 50–78 years, selected as cognitively normal and CSF Aβ-negative.
The Gothenburg Birth Cohort Studies, Sahlgrenska University Hospital – Population-based cohorts (H70, H79, H85, H88): 279 individuals, aged 70–88 years, selected with mini-mental state examination (MMSE) scores ≥29.
Neurochemistry Laboratory, Sahlgrenska University Hospital: 72 individuals, aged 71–91 years, selected based on non-pathological plasma concentrations of neurofilament light (NfL) and phosphorylated tau (p-tau217).
Neurology Clinic, Sahlgrenska University Hospital:128 symptomatic controls, aged 17–62 years, selected based on normal levels of neurodegeneration-associated biomarkers.
Cohort characteristics.
| Cohort | Name | Measurements, n | Age range, years |
|---|---|---|---|
| Cohort 1 | Neurochemistry, UGOT | 72 | 71–91 |
| Cohort 2 | DDI | 100 | 50–78 |
| Cohort 3 | The Gothenburg Birth Cohort Studies – SUH | 279 | 70–88 |
| Cohort 4 | Neurology Clinic – SUH | 128 | 17–62 |
-
UGOT, University of Gothenburg; SUH, Sahlgrenska University Hospital; DDI, Dementia Disease Initiation Study.
Prior to analysis, all samples and quality control (QC) materials were vortexed and centrifuged at 2000×g for 10 min.
MSD S-plex GFAP immunoassay and protocol
The S-PLEX® Human GFAP assay is a plate-based immunoassay that involves several manual pipetting steps prior to readout on the MESO QuickPlex SQ instrument, which is capable of reading up to five plates simultaneously. The manual procedure includes adding antibodies, samples, calibrators, and quality control materials. The details of the S-PLEX® Human GFAP assay are provided in the Supplementary Material.
Assay validation
The following assay verification experiments were conducted in alignment with the document published by Andreasson et al. [19].
Repeatability and intermediate precision
Plasma from de-identified patient samples was pooled to create three separate pools, representing low, medium, and high GFAP concentrations. The pool with a GFAP concentration of approximately 40 ng/L was used as the low QC, the pool with a concentration of approximately 340 ng/L was designated as the high QC, and the pool with a concentration of approximately 110 ng/L was assigned as the internal calibrator (IC). These pools were aliquoted into cryovials and stored in an ultra-freezer until analysis. Plasma samples with high and low levels of GFAP were analyzed in duplicate across five different occasions, using two different kit LOTs by two different technicians. Simultaneously, an IC was analyzed in duplicate twice during each of the five occasions. The IC was used in this verification for normalization, to compensate for variations between plates and kit batches, and to examine the effect of the IC on the precision of High and Low QC.
Measurement range
Calibration curves were generated using five independent experiments performed by two laboratory technicians across two different lot numbers. A five-parameter logistic regression model with 1/Y2 weighting was applied to fit the curves. Residuals were calculated for each calibrator point. The measurement range was defined as the span of the curve in which all calibrator concentrations deviated less than 15.2 % from their nominal values.
Interferences
Whole blood from 10 individuals was collected in two K2EDTA tubes per individual. One EDTA tube from each individual was centrifuged at 2000×g for 10 min, and plasma was transferred to a separate tube, aliquoted into four tubes, and stored in an ultra-freezer until analysis. The second EDTA tube was frozen without prior centrifugation to induce hemolysis. On the day of analysis, all samples were thawed, and the aliquots containing hemolyzed material were centrifuged at 4000×g for 10 min. Two of the four plasma aliquots were spiked with 3 μL of the corresponding hemolyzed sample, while the other two aliquots remained untreated. All tubes were centrifuged at 2000×g for 10 min at room temperature prior to analysis.
Sample stability
Sample stability was tested under three different conditions: room temperature, cold storage (4 °C), and frozen storage (−20 °C). All samples used were anonymized and collected from de-identified individuals.
Room temperature stability
To assess stability at room temperature, plasma from six individuals was aliquoted into nine tubes. One aliquot was immediately stored in an ultra-freezer at −80 °C, serving as the reference. Four aliquots were stored on a bench at room temperature and exposed to daylight for one, two, three, and seven days, respectively. After each specified storage period, the aliquots were transferred to an ultra-freezer at −80 °C, where they were kept until analysis.
Cold storage stability (4 °C)
Cold stability was assessed using the same six plasma samples. Four aliquots were stored at 4 °C for one, two, three, and seven days, respectively. After each corresponding storage period, the aliquots were transferred to an ultra-freezer at −80 °C until analysis. The same reference aliquot stored at −80 °C was used for comparison in this assessment.
Frozen storage stability (−20 °C)
Frozen storage stability at −20 °C was tested using plasma from five individuals. Plasma samples were aliquoted into five tubes per individual. One aliquot was immediately stored in an ultra-freezer at −80 °C, serving as the reference. The remaining four aliquots were stored at −20 °C for one, two, three, and seven days, respectively, before being transferred to an ultra-freezer at −80 °C for storage until analysis.
Sample type
The potential impact of sample type (i.e., lithium-heparin (Li-heparin) plasma, K2EDTA plasma, and serum) on GFAP concentrations was evaluated by collecting three paired blood samples with these tube additives from 24 individuals.
Statistical analysis
The normality of the data was assessed using the Shapiro–Wilk test. For non-normally distributed data, non-parametric tests were performed. The Mann–Whitney U test was performed for group comparisons. To evaluate the agreement between sample types, we used Bland–Altman [20] plots (showing percentage difference against the mean concentration), Passing-Bablok regression [21], and Spearman correlation analysis. A fixed (systematic) bias was considered present if the 95 % confidence interval for the regression intercept excluded 0, while proportional bias was defined as present when the 95 % CI for the slope did not include 1 [22]. For reference interval determination, the 579 individuals were stratified into three age groups: 18–<50 years, ≥50–<70 years, and ≥70 years. In the absence of a clinical reference standard, and given that most samples originated from population-based studies, it was not possible to fully confirm the health status of all participants. Therefore, to reduce the risk of including undiagnosed pathological cases, the 90th percentile was used to define the upper reference limit for the oldest age group, while the 95th percentile was used for the two younger groups. Accordingly, a right-sided reference interval approach was applied. The statistical analysis followed the direct non-parametric percentile method recommended by the Clinical and Laboratory Standards Institute (CLSI) guideline C28-A3, which does not assume a normal distribution of the data. To further investigate potential sex-related differences in cut-offs, we performed an additional analysis using available sex, age, and GFAP data. In some age groups, the number of participants was below 120, which limited the applicability of the direct non-parametric percentile method. Therefore, we applied the robust method recommended by the CLSI C28-A3 guideline, using a right-sided approach for both sexes. To visualize the age-related trajectory of plasma GFAP concentrations, individual GFAP values were plotted against age. A smoothed curve was fitted using the Fit>Spline>LOWESS function. The smoothing spline method was selected, with the number of knots set to 3, providing a balance between curve flexibility and overall trend clarity. This approach enabled the depiction of non-linear associations between age and plasma GFAP concentrations without imposing a predefined model structure Statistical significance was defined as an alpha of 0.05. All analyses were conducted using MedCalc (MedCalc Software Ltd, Belgium) and GraphPad Prism version 10.0.3 (GraphPad Software, Boston, MA, USA).
Results
Repeatability and intermediate precision
A summary of the precision experiments is presented in Table 2. The precision results remained within the pre-determined limit of 15 %, with the concentration of the IC determined to be 108 ng/L. When normalization to the IC was tested, the intermediate precision was significantly lower for the high QC, demonstrating improved precision. However, since the non-normalized measured uncertainty still fulfilled the acceptance criteria, and the variations anticipated in patient samples exceed this level of uncertainty, normalization to the IC was deemed unnecessary for this method. The expanded measurement uncertainty, calculated with a coverage factor of 2, is based on the highest observed repeatability (CVR) without IC normalization (7.6 %). This results in an expanded measurement uncertainty of approximately 15 %, calculated using the formula:
Repeatability and intermediate precision of plasma GFAP by the MSD S-plex GFAP kit.
| Samples | Mean concentration, ng/L | SDr, ng/L | CVr, % | SDRW, ng/L | CVRW, % |
|---|---|---|---|---|---|
| Low QC | 38.8 | 1.42 | 3.7 | 2.35 | 6.1 |
| Low QC normalized to IC | 38.1 | 1.45 | 3.8 | 2.43 | 6.4 |
| High QC | 343 | 26.2 | 7.6 | 40.8 | 11.9 |
| High QC normalized to IC | 336 | 27 | 8.0 | 27 | 8.0 |
-
SD, standard deviation; r, repeatability; RW, intermediate precision; IC, internal calibrator; CV, coefficient of variation; QC, quality control.
No significant differences were observed in lot-to-lot and operator comparisons (Figure 1).
Lot-to-lot and operator comparisons for GFAP measurements. (A, B) Lot comparison for GFAP measurements at low (A) and high (B) concentrations between two reagent lots. No significant differences were observed at low (p=0.82) or high (p=0.13) GFAP concentrations. (C, D) Operator comparison for GFAP measurements at low (C) and high (D) concentrations. Results generated by two operators show no significant differences at low (p=0.90) or high (p=0.30) GFAP concentrations, indicating good reproducibility across operators.
Measurement range
The endpoints of the calibration curve define the technical measurement range of this assay. All concentration residuals fall within the expanded measurement uncertainty of ±15 %. Therefore, the measurement range, defined by the lower and upper quantification limits, becomes 0.425–1,760 ng/L after considering the dilution factor of 1:2 (Figure 2A).
Residual plot for GFAP calibrators and impact of hemolysis on plasma GFAP. (A) Residual plot for GFAP calibrators across the measurement range. The residuals (%) for GFAP calibrators at various concentrations (0.21–880 ng/L) are shown to assess the measurement range. Residuals are calculated as the percentage difference between observed and expected values. The dotted lines represent the measurement uncertainty limits (±15 %). Residuals remain within these limits across the measurement range. (B) Impact of hemolysis on GFAP measurements. Comparison of GFAP concentrations in hemolyzed (red) and non-hemolyzed (black) samples. No significant difference is observed (p=0.85), indicating that hemolysis does not significantly affect GFAP measurements.
Interferences
The degree of hemolysis investigated in this experiment did not result in statistically significant differences in the measured GFAP concentrations (p=0.85, Figure 2B).
Sample stability
GFAP in plasma demonstrates robust stability across various storage conditions, including room temperature, cold storage (4 °C), and frozen storage (−20 °C). All measured concentrations were within the expanded measurement uncertainty of ±15 %, except for one data point (sample 5) observed during the experiment storage at −20 °C. Samples can reliably be stored at 4 °C for up to 7 days without significant changes in GFAP concentrations and can also withstand long periods at −20 °C, with all difference values remaining within the measurement uncertainty (Figure 3).
Stability of GFAP in plasma under various storage conditions. (A) Concentrations measured at room temperature (RT) for up to 7 days compared to −80 °C baseline storage. (B) Percentage difference from the average concentration during RT storage with a ±15 % uncertainty threshold (dashed lines). (C) Concentrations measured during storage at 4 °C for up to 7 days compared to −80 °C baseline storage. (D) Percentage difference from the average concentration during 4 °C storage with a ±15 % uncertainty threshold (dashed lines). (E) Concentrations measured during frozen storage at −20 °C for up to 7 days compared to −80 °C baseline storage. (F) Percentage difference from the average concentration during −20 °C storage with a ±15 % uncertainty threshold (dashed lines). Dashed lines in all percentage plots indicate the expanded measurement uncertainty of ±15 %. Notably, the largest differences were observed between the reference aliquot and the aliquot stored for one day at −20 °C, suggesting that the temperature during initial freezing may influence the measured concentrations. The single outlier (sample 5) exceeding the uncertainty threshold is reasonable, given the 95 % confidence interval and the limited number of data points (<20).
Sample type
GFAP concentrations showed very strong correlations between the three sample types: K2EDTA vs. Li-heparin plasma (Spearman’s r=0.967; 95 % CI: 0.924–0.986; p<0.0001) and K2EDTA vs. serum (Spearman’s r=0.977; 95 % CI: 0.946–0.990; p<0.0001). Despite these strong correlations, small differences were observed in the absolute concentrations between the sample types (Figure 4). Both comparisons indicated proportional bias, with K2EDTA samples generally yielding slightly higher GFAP values than those from Li-heparin plasma or serum, particularly at higher concentration levels (Figure 4).
Correlation and agreement between different blood collection tube types for GFAP. The first row presents the comparison between Li-heparin and K2EDTA plasma, while the second row shows the comparison between serum and K2EDTA plasma. Li-heparin vs. K2EDTA. (A) Correlation between GFAP concentrations measured in Li-heparin plasma (y-axis) and K2EDTA plasma (x-axis), both expressed in pg/mL, showing a strong positive correlation (Spearman’s r=0.967, p<0.0001). To better illustrate the data pattern, a LOESS (locally estimated scatterplot smoothing) curve with an 80 % span was applied to the trend line. (B) Bland-Altman plot showing the percentage differences (y-axis) between GFAP concentrations measured in Li-heparin and K2EDTA tubes, plotted against their mean concentration (x-axis). The solid blue line represents the mean bias, while the dashed brown lines indicate the limits of agreement (±1.96 SD). The orange dashed line marks the zero line, representing perfect agreement between the two sample types. The blue error bars denote the 95 % confidence intervals (CIs) for the limits of agreement, and the green error bar shows the 95 % CI for the mean bias. The mean difference was −4.5 % (95 % CI: −8.3 % to −0.7 %), with limits of agreement ranging from −22.2 % (95 % CI: −28.9 % to −15.6 %) to 13.2 % (95 % CI: 6.6–19.8 %). (C) Passing-Bablok regression plot showing GFAP concentrations measured in Li-heparin plasma (y-axis) and K2EDTA plasma (x-axis), demonstrating a proportional bias with a slope of 0.908 (95 % CI: 0.831 to 0.990) and an intercept of 0.339 (95 % CI: −1.690 to 5.628). The solid blue line represents the fitted regression line, the shaded blue area shows its 95 % confidence interval, and the dashed brown line indicates the identity line (y=x). Serum vs. K2EDTA. (A) Correlation between GFAP concentrations measured in serum (y-axis) and K2EDTA plasma (x-axis), both expressed in pg/mL, showing a strong positive correlation (Spearman’s r=0.977, p<0.0001). To better illustrate the data pattern, a LOESS (locally estimated scatterplot smoothing) curve with an 80 % span was applied to the trend line. (B) Bland-Altman plot showing the percentage differences (y-axis) between GFAP concentrations measured in serum and K2EDTA tubes, plotted against their mean concentration (x-axis). The solid blue line represents the mean bias, while the dashed brown lines indicate the limits of agreement (±1.96 SD). The orange dashed line marks the zero line, representing perfect agreement between the two sample types. The blue error bars denote the 95 % confidence intervals (CIs) for the limits of agreement, and the green error bar shows the 95 % CI for the mean bias. The mean difference was −5.7 % (95 % CI: −9.8 % to −1.7 %), with limits of agreement ranging from −24.3 % (95 % CI: −31.3 % to −17.4 %) to 12.9 % (95 % CI: 5.9–19.8 %). (C) Passing-Bablok regression plot showing GFAP concentrations measured in Li-heparin plasma (y-axis) and K2EDTA plasma (x-axis), demonstrating a proportional bias with a slope of 0.928 (95 % CI: 0.818 to 0.997) and an intercept of −0.249 (95 % CI: −2.591 to 5.587). The solid blue line represents the fitted regression line, the shaded blue area shows its 95 % confidence interval, and the dashed brown line indicates the identity line (y=x).
Derivation of age-stratified cut-offs
As age-related increases in plasma GFAP have been previously reported in adult populations [5], 23], we established age-specific reference limits by visually inspecting the distribution of plasma GFAP concentrations plotted against age (Figure 5). Based on this inspection, which revealed a non-linear increase in GFAP concentrations with inflection points around 50 and 70 years of age, three biologically distinct age groups were defined: 18–<50 years (n=121), ≥50–<70 years (n=101), and ≥70 years (n=357). Additional sensitivity analysis showed that the median GFAP concentrations of these predefined groups differed significantly from each other (Supplementary Table 2, Supplementary Figure 5). Using a right-sided percentile approach, the 95th percentile was applied to the two younger groups, while the 90th percentile was used for the oldest group due to the higher likelihood of subclinical pathology. The resulting age-specific upper reference limits for plasma GFAP were: 38 pg/mL for individuals aged 18–<50, 73 pg/mL for those aged ≥50–<70, and 156 pg/mL for the ≥70 age group (Table 3 and Figure 6).
Age-related trajectory of plasma GFAP concentrations in a healthy population. This Figure illustrates individual plasma GFAP concentrations plotted against age to visualize the age-related trajectory. A smoothed curve was fitted using the fit>spline > LOWESS function in GraphPad prism (version 10.0.3), with the smoothing spline method and three knots selected to balance curve flexibility and trend clarity. This approach enabled the depiction of non-linear associations between age and plasma GFAP levels without assuming a predefined model structure.
Age-stratified right-sided cut-offs for plasma GFAP concentrations this graph displays individual plasma GFAP concentrations plotted against age, illustrating age-related increases and the derivation of right-sided reference limits. A total of 579 individuals were stratified into three age groups: 18–<50 years (n=121, 95th percentile: 38 pg/mL, 90 % CI: 36–44), ≥50–<70 years (n=101, 95th percentile: 73 pg/mL, 90 % CI: 64–91), and ≥70 years (n=357, 90th percentile: 156 pg/mL, 90 % CI: 142–171). Solid red lines indicate the 95th percentile used for the two younger age groups, while the dotted red line represents the 90th percentile applied to the oldest group, reflecting the increased likelihood of subclinical pathology in this population.
Right-sided, age-stratified reference limits for plasma GFAP.
| Age group, years | Measurements, n | Plasma GFAP, pg/mL, median (range) | Cut-off (plasma GFAP) |
|---|---|---|---|
| 18 to<50 | 121 | 22 (7–46) | 38 pg/mL |
| 50 to<70 | 101 | 36 (13–132) | 73 pg/mL |
| >70 | 357 | 75 (17–372) | 156 pg/mL |
-
GFAP, glial fibrillary acidic protein.
Age-dependent centile model for plasma GFAP
In addition to the predefined age groups, we performed a continuous centile analysis to model the relationship between age and plasma GFAP levels. This approach generated lower and upper percentile estimates across the entire age range. The model revealed a non-linear increase in GFAP concentrations with advancing age, particularly above 60 years. The estimated 5th and 95th percentiles for selected age points, along with their 90 % confidence intervals, are presented in Supplementary Table 1, and the modeled age trajectory is visualized in Supplementary Figure 1.
Sex-specific, age-stratified cut-offs and age-dependent centile modeling of plasma GFAP
After obtaining sex information for the DDI study (n=100), the Gothenburg Birth Cohort Studies (n=275), and the Neurology Clinic at Sahlgrenska University Hospital (n=128), the dataset comprised 315 females and 188 males in total. Following the same approach used for the whole population, three age groups were defined for females and males, respectively: 18–<50 years (n=97, 23), ≥50–<70 years (n=60, 41), and ≥70 years (n=158, 124). Using a right-sided percentile approach, the 95th percentile was applied to the two younger groups, while the 90th percentile was used for the oldest group due to the higher likelihood of subclinical pathology. The resulting age-specific upper reference limits for plasma GFAP were, for females and males respectively: 36 and 37 pg/mL for individuals aged 18–<50 years, 74 and 61 pg/mL for those aged ≥50–<70 years, and 168 and 133 pg/mL for the ≥70-year group (Supplementary Figure 2).
In addition to the predefined age groups, we performed a continuous centile analysis to model the relationship between age and plasma GFAP levels for each sex. This approach generated lower and upper percentile estimates across the entire age range. For both sexes, the model captured the accelerating increase in GFAP concentrations with age, particularly after approximately age 60, consistent with the percentile trends observed in the whole cohort when only age (without sex information) was considered. However, males generally exhibited lower absolute GFAP concentrations than females, with the difference becoming more pronounced in older age groups (Supplementary Figures 3, 4).
Correlation between age and plasma GFAP
A significant positive correlation was observed between age and plasma GFAP concentration when considering the entire study population (Spearman’s r=0.832; 95 % CI: 0.8043–0.8561; p<0.0001; n=579). When analyzed by age group, the strength of the correlation varied. Among individuals aged 18–<50 years, the correlation was weak but statistically significant (r=0.2436; 95 % CI: 0.0627–0.4089; p=0.0071; n=121). In the 50–<70 years group, a moderate positive correlation was found (r=0.4346; 95 % CI: 0.2559–0.5846; p<0.0001; n=101). The strongest age-related increase in GFAP concentration was observed in individuals aged≥70 years, showing a moderate-to-strong correlation (r=0.5998; 95 % CI: 0.5267–0.6641; p<0.0001; n=357). These results support a progressive age-dependent increase in plasma GFAP levels, particularly pronounced in older individuals.
Discussion
This study first focused on evaluating the analytical performance of the MSD S-PLEX® Human GFAP assay, which utilizes ECL technology, to support its potential use in routine plasma GFAP measurement. Key validation parameters were examined in accordance with established methodological guidelines [19]. Following this, we characterized plasma GFAP concentrations in a large cohort of apparently healthy individuals ranging in age from 17 to 91 years. Age-stratified reference limits were established across three predefined age groups to account for the well-recognized influence of aging on plasma GFAP levels.
Repeatability and intermediate precision assessments yielded maximum CVs% of 7.6 and 11.9 %, respectively. Normalizing measured concentrations with an IC reduced CVs for intermediate precision only at high QC levels. The manufacturer-defined acceptable ranges for the highest calibrator point (683–1,270 ng/L) and the lowest calibrator point (0.167–0.309 ng/L), along with a 1:2 dilution factor, were used to determine a final measurement range of 0.620–1,360 ng/L. This range is defined by the highest acceptable value of the lowest calibrator (0.309 ng/L) and the lowest acceptable value of the highest calibrator (683 ng/L). Values below the lower limit of quantification (LLoQ) were considered unmeasurable, and samples exceeding the upper limit of quantification (ULoQ) required additional dilution. It is important to note, however, that both lots used in the validation relied on the same lot of the calibrator, suggesting that greater deviations than those observed in this study may occur with future calibrator lots.
Given the nature of plasma samples, hemolysis was evaluated as a potential source of interference. A comparison of non-hemolyzed and highly hemolyzed samples from the same individuals revealed that hemolysis at a high degree had negligible effects on GFAP levels. Although interference from intermediate levels of hemolysis was not tested, the observed minimal impact at high degrees of hemolysis supports the assay’s robustness. GFAP concentrations in plasma remained stable for up to seven days when stored at 4 °C. For longer-term storage, freezing at −20 °C/−80 °C may be necessary to preserve protein stability. Notably, all aliquots, except the reference aliquot (−80 °C), were frozen at −20 °C, and the largest differences were observed between the reference aliquot and the aliquot stored for one day at −20 °C, suggesting that the freezing temperature during initial freezing may influence measured concentrations. One outlier (sample 5) was within expectations given the 95 % confidence level. Stability during freeze-thaw cycles (FTC) was not evaluated in this study, as previous publications on plasma GFAP [12], have already demonstrated that blood GFAP remains stable through up to five FTCs.
In parallel with evaluated pre-analytical factors, the impact of blood collection tube type on GFAP concentrations has also been examined in several previous studies [24], 25], primarily using the Simoa platform. Consistent with our findings, those studies reported strong correlations between different tube types, despite differences in absolute concentrations [24]. However, in contrast to their observations – where Li-heparin tubes yielded the highest GFAP values – we found that K2EDTA plasma yielded slightly higher concentrations than both Li-heparin plasma and serum, particularly at higher GFAP levels. This discrepancy may be attributed to differences in assay platforms, including potential effects of assay reagents or the specific capture antibodies used. Given these differences in absolute concentrations, matrix-specific validation and threshold values (e.g., for mild TBI) may be required. Further studies are needed to replicate these findings and to clarify the factors contributing to these variations.
To date, only a few validation studies have been published on the use of serum or plasma GFAP as a biomarker for astroglial activation. One of these studies validated the second-generation Ella assay and compared its performance to a homebrew Simoa GFAP assay [4]. The findings demonstrated that the Ella assay offers reliable results with high reproducibility, while the Simoa assay exhibited greater sensitivity. Another recent study introduced and validated a fully automated immunoassay for GFAP quantification, which also demonstrated high reproducibility [16]. A few studies [4], 12], 23], 24], 26] have also assessed the preanalytical properties of blood GFAP, with findings similar to ours.
Upon assessing the scatter plot of plasma GFAP concentrations across age visually, we defined three age intervals to account for the observed non-linear, age-dependent increase in GFAP levels over the adult lifespan. Pediatric samples were not included in this study due to the unavailability of such samples. The age-related increase observed in our dataset is consistent with previously published findings [5], 23], 27], although differences in absolute concentration values were noted. These discrepancies are likely attributable to variations in assay platforms and calibrators used across studies. It is also important to note that while some previous reference interval studies have been conducted using serum [5], 23], our analysis was performed using K2EDTA plasma samples, which may further complicate direct comparisons – even when the same assay and calibrator are employed. Additionally, we obtained sex information from the cohorts used in this study to determine sex-related plasma GFAP cut-offs, and the results were consistent with previous reports [5], 23].
When introducing plasma GFAP testing into routine clinical use, it becomes crucial for laboratories to ensure that their measurements are traceable to those used to define the reference intervals in this study. Although the same assay platform may be used, subtle differences in calibration, instrumentation, or lab conditions can still lead to variation in results. To achieve traceability, laboratories are welcome to send aliquots of previously measured samples to the Clinical Neurochemistry Laboratory at Sahlgrenska University Hospital, where the same validated method is used. This comparison enables the detection of any systematic differences and allows for adjustments if needed to align results. In our own internal precision assessment, we found that applying normalization resulted in a decrease in the intermediate precision from 11.9 to 8.0 % for the high QC. Such alignment efforts are key to ensuring that reference intervals can be applied reliably across different laboratories and patient populations.
Plasma GFAP, as an astrocytic biomarker, is increasingly being recognized and integrated into biomarker-supported clinical decision algorithms for conditions such as AD, MS, and TBI. The choice of assay for GFAP quantification often depends on factors such as laboratory infrastructure, budget constraints, and the volume of samples received from clinical settings. While fully automated immunoassays are ideal for minimizing human error and reducing turnaround time, the selection of an appropriate assay ultimately depends on availability and specific needs. In this study, MSD S-PLEX® Human GFAP assay was chosen for validation, given its good reproducibility and broad measurement range, making it a suitable candidate for future routine implementation.
There are now several platforms available on the market for the quantification of plasma or serum GFAP, including Simoa (Quanterix), Ella (ProteinSimple), the fully automated Lumipulse platform (Fujirebio), and the Alinity platform (Abbott) in combination with UCH-L1 for mild TBI testing, as well as Vidas for the same purpose. Some of these platforms, such as Lumipulse and Simoa, have been used for reference interval determination, while others have been applied mainly for analytical or clinical validation studies. Our results are consistent with previous reports [5], 16], 23], 27], in showing an age-related increase in blood GFAP concentrations irrespective of the platform used, with a strong correlation between GFAP levels and age. Similar to recent findings by Agnello et al. [5], our analysis of sex-related reference limits based on available samples showed that, from approximately age 50 onward, females had higher plasma GFAP concentrations than males. However, absolute concentrations and cut-offs differ across platforms, likely for several reasons. First, although we used plasma as the sample matrix in our study, most previous studies have used serum. Even though we demonstrated a strong correlation between serum and plasma GFAP, the absolute concentrations differ between the matrices. Second, variations in calibrators, immunoassay design, and analytical sensitivity contribute to differences in absolute concentration values. These findings underscore the need for harmonization between assays and the development of reference materials for blood GFAP to enable recalibration and comparability across platforms.
The main strength of this study lies in several key aspects. First, selected analytical validation parameters – including precision, and sample stability – were systematically assessed to support the assay’s potential for routine clinical implementation. A further strength is the inclusion of a relatively large number of participants, particularly in the elderly age group. The use of samples from population-based cohorts enhances the generalizability of our findings, making the results broadly representative of the general population. However, this study also has some limitations. First, we did not assess the clinical performance of the GFAP assay, as the primary focus was on analytical validation. Second, due to the lack of access to complete demographic and clinical data, we were able to determine sex-specific reference intervals only for the subset of participants with this information available. Nevertheless, our findings were consistent with those of previous studies. Third, our reference interval determination was based on a relatively limited number of individuals aged 18–50 compared to the older age groups, which may reduce the precision of the cut-offs established for this younger population. Finally, we did not account for potential confounding factors such as impaired kidney function, comorbidities, or undiagnosed CNS conditions, which may influence plasma GFAP concentrations. In the ≥70-year group, we used the 90th percentile to help reduce the noise introduced by such factors. To further minimize this effect, and to reach an adequate sample size, we supplemented cohort samples – selected for healthy controls based on clinical records – with leftover samples from our neurochemistry laboratory, including only those with non-pathological plasma p-tau217 and NfL levels to exclude brain amyloid pathology and neurodegeneration. Nevertheless, despite these efforts, residual noise from unrecognized conditions in this age group may still be present [28]. Finally, a potential limitation of the S-PLEX® Human GFAP assay is its dependence on multiple manual steps, which may introduce variability if not performed consistently. Due to the complexity of the procedure, it may require trained and certified laboratory personnel to ensure reliable and reproducible results, which might limit its broader implementation in less experienced laboratories.
In conclusion, this study demonstrates the strong analytical and pre-analytical performance of a commercially available GFAP assay, supporting its suitability for routine clinical implementation. Furthermore, age-stratified reference limits for plasma GFAP were established, reflecting its dynamic changes across the adult lifespan. Together, these advancements will enhance the evaluation of patients across a range of neurological conditions and support the monitoring of disease-modifying therapies.
Funding source: Anna-Lisa och Bror Björnssons Stiftelse, Sweden
Funding source: Swedish Research Council
Award Identifier / Grant number: #2023–00356, #2022–01018 and #2019–02397
Funding source: the European Union’s Horizon Europe research and innovation programme
Award Identifier / Grant number: 101053962
Funding source: Swedish State Support for Clinical Research
Award Identifier / Grant number: #ALFGBG-71320
Funding source: The Alzheimer Drug Discovery Foundation (ADDF), USA
Award Identifier / Grant number: #201809–2016862
Funding source: The AD Strategic Fund and the Alzheimer’s Association
Award Identifier / Grant number: #ADSF-21–831376-C, #ADSF-21–831381-C, #ADSF-21–831377-C, and #ADSF-24–1284328-C
Funding source: The European Partnership on Metrology, co-financed from the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States
Award Identifier / Grant number: NEuroBioStand, #22HLT07
Funding source: The Bluefield Project, Cure Alzheimer’s Fund, the Olav Thon Foundation, the Erling-Persson Family Foundation, Familjen Rönströms Stiftelse, Stiftelsen för Gamla Tjänarinnor, Hjärnfonden, Sweden
Award Identifier / Grant number: #FO2022-0270
Funding source: The European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 860197
Award Identifier / Grant number: MIRIADE
Funding source: The European Union Joint Programme – Neurodegenerative Disease Research
Award Identifier / Grant number: JPND2021-00694
Funding source: The National Institute for Health and Care Research University College London Hospitals Biomedical Research Centre, and the UK Dementia Research Institute at UCL
Award Identifier / Grant number: UKDRI-1003
Funding source: The Swedish Research Council
Award Identifier / Grant number: #2017- 00915 and #2022–00732
Funding source: The Swedish Alzheimer Foundation
Award Identifier / Grant number: #AF-930351, #AF-939721 and #AF- 968270
Funding source: Hjärnfonden, Sweden
Award Identifier / Grant number: #FO2017-0243 and #ALZ2022-0006
Funding source: The Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement
Award Identifier / Grant number: #ALFGBG- 715986 and #ALFGBG-965240
Funding source: The European Union Joint Program for Neurodegenerative Disorders
Award Identifier / Grant number: JPND2019-466–236
Funding source: The Alzheimer’s Association 2021 Zenith Award
Award Identifier / Grant number: ZEN-21–848495
Funding source: The Alzheimer’s Association 2022–2025
Award Identifier / Grant number: SG-23–1038904 QC
Funding source: The Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement
Award Identifier / Grant number: ALFGBG-1005471, ALFGBG-965923, ALFGBG-81392, ALF GBG-771071
Funding source: The Alzheimerfonden
Award Identifier / Grant number: AF-842471, AF-737641, AF-929959, AF-939825
Funding source: The Swedish Research Council
Award Identifier / Grant number: 2019–02075, 2019–02075_15
Funding source: Stiftelsen Psykiatriska Forskningsfonden, The Swedish Brain Foundation
Award Identifier / Grant number: FO2024-0097
Funding source: The Swedish state under the agreement between the Swedish government and the county councils, the ALF agreement
Award Identifier / Grant number: ALFGBG-984092
Acknowledgments
The authors would like to express their most sincere gratitude to the cohort participants and their relatives, without whom this research would not have been possible. The authors also thank the staff at the study centers and universities involved in this study for their invaluable support and contributions.
-
Research ethics: For the anonymized samples the collection at the Clinical Chemistry Laboratory, Sahlgrenska University Hospital, was conducted in accordance with the Ethics Committee at University of Gothenburg (EPN140811). All participants provided written informed consent, and the study was conducted in accordance with the Declaration of Helsinki. The study was approved by the Regional Committees for Medical and Health Research Ethics in Norway (REK 2013/150) for the Dementia Disease Initiation (DDI) cohort; the Regional Ethical Review Board in Gothenburg, Sweden, for the Gothenburg Birth Cohort studies H70 (Dnr 869-13), H79 (Dnr 328-09), H85 (Dnr 131-15), and H88 (Dnr 278-18); and the Regional Ethical Review Board in Gothenburg, Sweden for the cohort from the Neurology Clinic at Sahlgrenska University Hospital (Dnr 460-13 and T217-15/Ad/460-13).
-
Informed consent: Written informed consent was obtained from all participants included in the cohorts used in this study, in accordance with the Declaration of Helsinki.
-
Author contributions: All authors have accepted responsibility for the entire content of this manuscript and approved its submission.
-
Use of Large Language Models, AI and Machine Learning Tools: None declared.
-
Conflict of interest: KB has served as a consultant and at advisory boards for Abbvie, AC Immune, ALZPath, AriBio, BioArctic, Biogen, Eisai, Lilly, Moleac Pte. Ltd, Neurimmune, Novartis, Ono Pharma, Prothena, Roche Diagnostics, Sanofi and Siemens Healthineers; has served at data monitoring committees for Julius Clinical and Novartis; has given lectures, produced educational materials and participated in educational programs for AC Immune, Biogen, Celdara Medical, Eisai and Roche Diagnostics; and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program, outside the work presented in this paper. HZ has served at scientific advisory boards and/or as a consultant for Abbvie, Acumen, Alector, Alzinova, ALZPath, Annexon, Apellis, Artery Therapeutics, AZTherapies, CogRx, Denali, Eisai, Nervgen, Novo Nordisk, Optoceutics, Passage Bio, Pinteon Therapeutics, Prothena, Red Abbey Labs, reMYND, Roche, Samumed, Siemens Healthineers, Triplet Therapeutics, and Wave, has given lectures in symposia sponsored by Cellectricon, Fujirebio, Alzecure, Biogen, and Roche, and is a co-founder of Brain Biomarker Solutions in Gothenburg AB (BBS), which is a part of the GU Ventures Incubator Program (outside submitted work). BEK has served as a consultant for Biogen and medical advisory boards for Biogen and Eli Lilly. SK has served at scientific advisory boards, speaker and / or as consultant for Roche, Eli Lilly, Geras Solutions, Optoceutics, Biogen, Eisai, Merry Life, Triolab, Novo Nordisk and Bioarctic, unrelated to present study content. LN has received compensation for lectures from Biogen, Novartis, Teva, Sanofi and Merck, has served on advisory boards for Merck, Janssen and Sanofi, has received unconditional research grant from Novartis and Sanofi and participated in clinical trials as Principal Investigator sponsored by Amgen, Sanofi, ArgenX and Takeda.
-
Research funding: BA is supported by the Anna-Lisa och Bror Björnssons Stiftelse, Sweden. HZ is a Wallenberg Scholar and a Distinguished Professor at the Swedish Research Council supported by grants from the Swedish Research Council (#2023–00356, #2022–01018 and #2019–02397), the European Union’s Horizon Europe research and innovation programme under grant agreement No 101053962, Swedish State Support for Clinical Research (#ALFGBG-71320), the Alzheimer Drug Discovery Foundation (ADDF), USA (#201809–2016862), the AD Strategic Fund and the Alzheimer’s Association (#ADSF-21–831376-C, #ADSF-21–831381-C, #ADSF-21–831377-C, and #ADSF-24–1284328-C), the European Partnership on Metrology, co-financed from the European Union’s Horizon Europe Research and Innovation Programme and by the Participating States (NEuroBioStand, #22HLT07), the Bluefield Project, Cure Alzheimer’s Fund, the Olav Thon Foundation, the Erling-Persson Family Foundation, Familjen Rönströms Stiftelse, Stiftelsen för Gamla Tjänarinnor, Hjärnfonden, Sweden (#FO2022-0270), the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 860197 (MIRIADE), the European Union Joint Programme – Neurodegenerative Disease Research (JPND2021-00694), the National Institute for Health and Care Research University College London Hospitals Biomedical Research Centre, and the UK Dementia Research Institute at UCL (UKDRI-1003). KB is supported by the Swedish Research Council (#2017- 00915 and #2022–00732), the Swedish Alzheimer Foundation (#AF-930351, #AF-939721 and #AF- 968270), Hjärnfonden, Sweden (#FO2017-0243 and #ALZ2022-0006), the Swedish state under the agreement between the Swedish government and the County Councils, the ALF-agreement (#ALFGBG- 715986 and #ALFGBG-965240), the European Union Joint Program for Neurodegenerative Disorders (JPND2019-466–236), the Alzheimer’s Association 2021 Zenith Award (ZEN-21–848495), and the Alzheimer’s Association 2022–2025 Grant (SG-23–1038904 QC). SK was financed by grants from the Swedish state under the agreement between the Swedish government and the county councils, the ALF-agreement (ALFGBG-1005471, ALFGBG-965923, ALFGBG-81392, ALF GBG-771071). The Alzheimerfonden (AF-842471, AF-737641, AF-929959, AF-939825). The Swedish Research Council (2019–02075, 2019–02075_15),Stiftelsen Psykiatriska Forskningsfonden, The Swedish Brain Foundation FO2024-0097. AD was financed by the Swedish state under the agreement between the Swedish government and the county councils, the ALF agreement (ALFGBG-984092).
-
Data availability: The data supporting the findings of this study may be shared with qualified academic investigators for the purpose of result replication, upon reasonable request to the corresponding author and under a material transfer agreement.
References
1. Middeldorp, J, Hol, E. GFAP in health and disease. Prog Neurobiol 2011;93:421–43. https://doi.org/10.1016/j.pneurobio.2011.01.005.Suche in Google Scholar PubMed
2. Rosengren, LE, Wikkelsø, C, Hagberg, L. A sensitive ELISA for glial fibrillary acidic protein: application in CSF of adults. J Neurosci Methods 1994;51:197–204. https://doi.org/10.1016/0165-0270(94)90011-6.Suche in Google Scholar PubMed
3. Benedet, AL, Milà-Alomà, M, Vrillon, A, Ashton, NJ, Pascoal, TA, Lussier, F, et al.. Differences between plasma and cerebrospinal fluid glial fibrillary acidic protein levels across the Alzheimer disease continuum. JAMA Neurol 2021;78:1471–83. https://doi.org/10.1001/jamaneurol.2021.3671.Suche in Google Scholar PubMed PubMed Central
4. Fazeli, B, Gómez de San José, N, Jesse, S, Senel, M, Oeckl, P, Erhart, DK, et al.. Quantification of blood glial fibrillary acidic protein using a second-generation microfluidic assay. Validation and comparative analysis with two established assays. Clin Chem Lab Med 2024;62:1591–601. https://doi.org/10.1515/cclm-2023-1256.Suche in Google Scholar PubMed
5. Agnello, L, Gambino, CM, Ciaccio, AM, Giglio, RV, Scazzone, C, Tamburello, M, et al.. Establishing sex-and age-related reference intervals of serum glial fibrillary acid protein measured by the fully automated lumipulse system. Clin Chem Lab Med 2025;63:1402–8. https://doi.org/10.1515/cclm-2025-0093.Suche in Google Scholar PubMed
6. Abdelhak, A, Foschi, M, Abu-Rumeileh, S, Yue, JK, D’Anna, L, Huss, A, et al.. Blood GFAP as an emerging biomarker in brain and spinal cord disorders. Nat Rev Neurol 2022;18:158–72. https://doi.org/10.1038/s41582-021-00616-3.Suche in Google Scholar PubMed
7. Zetterberg, H, Blennow, K. Fluid biomarkers for mild traumatic brain injury and related conditions. Nat Rev Neurol 2016;12:563–74. https://doi.org/10.1038/nrneurol.2016.127.Suche in Google Scholar PubMed
8. Arslan, B, Zetterberg, H, Ashton, NJ. Blood-based biomarkers in Alzheimer’s disease–moving towards a new era of diagnostics. Clin Chem Lab Med 2024;62:1063–9. https://doi.org/10.1515/cclm-2023-1434.Suche in Google Scholar PubMed
9. Abdelhak, A, Huss, A, Kassubek, J, Tumani, H, Otto, M. Serum GFAP as a biomarker for disease severity in multiple sclerosis. Sci Rep 2018;8:14798. https://doi.org/10.1038/s41598-018-33158-8.Suche in Google Scholar PubMed PubMed Central
10. Pereira, JB, Janelidze, S, Smith, R, Mattsson-Carlgren, N, Palmqvist, S, Teunissen, CE, et al.. Plasma GFAP is an early marker of amyloid-β but not tau pathology in Alzheimer’s disease. Brain 2021;144:3505–16. https://doi.org/10.1093/brain/awab223.Suche in Google Scholar PubMed PubMed Central
11. Baiardi, S, Quadalti, C, Mammana, A, Dellavalle, S, Zenesini, C, Sambati, L, et al.. Diagnostic value of plasma p-tau181, NfL, and GFAP in a clinical setting cohort of prevalent neurodegenerative dementias. Alzheimers Res Ther 2022;14:153. https://doi.org/10.1186/s13195-022-01093-6.Suche in Google Scholar PubMed PubMed Central
12. Simrén, J, Weninger, H, Brum, WS, Khalil, S, Benedet, AL, Blennow, K, et al.. Differences between blood and cerebrospinal fluid glial fibrillary acidic protein levels: the effect of sample stability. Alzheimer’s Dement 2022;18:1988–92. https://doi.org/10.1002/alz.12806.Suche in Google Scholar PubMed PubMed Central
13. Watanabe, M, Nakamura, Y, Michalak, Z, Isobe, N, Barro, C, Leppert, D, et al.. Serum GFAP and neurofilament light as biomarkers of disease activity and disability in NMOSD. Neurology 2019;93:e1299–311. https://doi.org/10.1212/wnl.0000000000008160.Suche in Google Scholar
14. Kobeissy, F, Arja, RD, Munoz, JC, Shear, DA, Gilsdorf, J, Zhu, J, et al.. The game changer: UCH-L1 and GFAP-Based blood test as the first marketed in vitro diagnostic test for mild traumatic brain injury. Expert Rev Mol Diagn 2024;24:67–77. https://doi.org/10.1080/14737159.2024.2306876.Suche in Google Scholar PubMed
15. Keski-Pukkila, M, Karr, JE, Posti, JP, Berghem, K, Kotilainen, A-K, Blennow, K, et al.. Preliminary evaluation of the Scandinavian guidelines for initial management of minimal, mild, and moderate head injuries with glial fibrillary acidic protein. Neurotrauma Rep 2024;5:50–60. https://doi.org/10.1089/neur.2023.0077.Suche in Google Scholar PubMed PubMed Central
16. Kamada, J, Hamanaka, T, Oshimo, A, Sato, H, Nishii, T, Fujita, M, et al.. Glial fibrillary acidic protein’s usefulness as an astrocyte biomarker using the fully automated LUMIPULSE® system. Diagnostics (Basel) 2024;14:2520. https://doi.org/10.3390/diagnostics14222520.Suche in Google Scholar PubMed PubMed Central
17. Kang, H, Woo, S-Y, Shin, D, Yim, S, Lee, EH, Ryu, H, et al.. Reproducibility of plasma biomarker measurements across laboratories: insights into ptau217, GFAP, and NfL. Dement Neurocogn Disord 2025;24:91. https://doi.org/10.12779/dnd.2025.24.2.91.Suche in Google Scholar PubMed PubMed Central
18. Yue, JK, Yuh, EL, Korley, FK, Winkler, EA, Sun, X, Puffer, RC, et al.. Association between plasma GFAP concentrations and MRI abnormalities in patients with CT-negative traumatic brain injury in the TRACK-TBI cohort: a prospective multicentre study. Lancet Neurol 2019;18:953–61. https://doi.org/10.1016/s1474-4422(19)30282-0.Suche in Google Scholar
19. Andreasson, U, Perret-Liaudet, A, van Waalwijk van Doorn, LJ, Blennow, K, Chiasserini, D, Engelborghs, S, et al.. A practical guide to immunoassay method validation. Front Neurol 2015;6:179. https://doi.org/10.3389/fneur.2015.00179.Suche in Google Scholar PubMed PubMed Central
20. Bland, JM, Altman, DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135–60. https://doi.org/10.1191/096228099673819272.Suche in Google Scholar
21. Passing, H, Bablok, W. Comparison of several regression procedures for method comparison studies and determination of sample sizes application of linear regression procedures for method comparison studies in clinical chemistry, part II. J Clin Chem Clin Biochem 1984;22:431–45. https://doi.org/10.1515/cclm.1984.22.6.431.Suche in Google Scholar PubMed
22. Bilic-Zulle, L. Comparison of methods: passing and bablok regression. Biochem Med (Zagreb) 2011;21:49–52. https://doi.org/10.11613/bm.2011.010.Suche in Google Scholar PubMed
23. Tybirk, L, Hviid, CVB, Knudsen, CS, Parkner, T. Serum GFAP–Reference interval and preanalytical properties in Danish adults. Clin Chem Lab Med 2022;60:1830–8. https://doi.org/10.1515/cclm-2022-0646.Suche in Google Scholar PubMed
24. Ashton, NJ, Suárez-Calvet, M, Karikari, TK, Lantero-Rodriguez, J, Snellman, A, Sauer, M, et al.. Effects of pre-analytical procedures on blood biomarkers for Alzheimer’s pathophysiology, glial activation, and neurodegeneration. Alzheimers Dement (Amst) 2021;13:e12168. https://doi.org/10.1002/dad2.12168.Suche in Google Scholar PubMed PubMed Central
25. Verberk, IM, Misdorp, EO, Koelewijn, J, Ball, AJ, Blennow, K, Dage, JL, et al.. Characterization of pre-analytical sample handling effects on a panel of Alzheimer’s disease–related blood-based biomarkers: results from the standardization of Alzheimer’s blood biomarkers (SABB) working group. Alzheimer’s Dement 2022;18:1484–97. https://doi.org/10.1002/alz.12510.Suche in Google Scholar PubMed PubMed Central
26. van Lierop, ZY, Verberk, IM, van Uffelen, KW, Koel-Simmelink, MJ, in’t Veld, L, Killestein, J, et al.. Pre-analytical stability of serum biomarkers for neurological disease: neurofilament-light, glial fibrillary acidic protein and contactin-1. Clin Chem Lab Med 2022;60:842–50. https://doi.org/10.1515/cclm-2022-0007.Suche in Google Scholar PubMed
27. Cooper, JG, Stukas, S, Ghodsi, M, Ahmed, N, Diaz-Arrastia, R, Holmes, DT, et al.. Age specific reference intervals for plasma biomarkers of neurodegeneration and neurotrauma in a Canadian population. Clin Biochem 2023;121:110680. https://doi.org/10.1016/j.clinbiochem.2023.110680.Suche in Google Scholar PubMed
28. Arslan, B, Brum, WS, Pola, I, Therriault, J, Rahmouni, N, Stevenson, J, et al.. The impact of kidney function on Alzheimer’s disease blood biomarkers: implications for predicting amyloid-β positivity. Alzheimers Res Ther 2025;17:48. https://doi.org/10.1186/s13195-025-01692-z.Suche in Google Scholar PubMed PubMed Central
Supplementary Material
This article contains supplementary material (https://doi.org/10.1515/cclm-2025-0870).
© 2025 the author(s), published by De Gruyter, Berlin/Boston
This work is licensed under the Creative Commons Attribution 4.0 International License.
