Influence of Decreased Kidney Function on Plasma Biomarkers of Neurodegenerative Disorders in Routine Care: Confirmation of the Interest of Ratios

Abstract

Background and Objectives

Interest in plasma biomarkers for neurodegenerative disorders is growing, but their reliance on glomerular filtration makes kidney function a key potential confounder. This study assesses the effect of kidney function (eGFR) on plasma biomarkers of neurodegeneration and on their accuracy for detecting cerebral amyloidosis.

Methods

This observational study aims at studying the effect of kidney function on blood biomarkers through simultaneous measurements of creatinine, plasma, and CSF biomarkers (Aβ42, Aβ40, p-tau181, p-tau217, neurofilament light chain [NfL], glial fibrillary acidic protein [GFAP], and brain-derived tau [BD-tau]), as well as plasma biomarker ratios (Aβ42/Aβ40, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42), in the ALZAN cohort of patients. This prospective multicenter cohort (#NCT05427448), recruited across Montpellier, Nîmes, and Perpignan hospitals, included patients from November 2022 to July 2024 who met the following criteria: age ≥18 years, informed consent, and concomitant CSF and blood sampling. To determine the association between plasma biomarkers and kidney function, we performed univariable linear regression analyses.

Results

A total of 420 patients were included (mean age: 71.1 years, %female: 53.2) and subdivided into 3 groups according to eGFR value (ml/minute/1.73 m²): <60 (n = 36), [60–90] (n = 194), and >90 (n = 190). All mean plasma biomarker levels were significantly higher in the eGFR < 60 group. Except for p-tau217, a statistically significant inverse correlation was observed between eGFR and individual plasma biomarkers. Furthermore, age-adjusted univariable linear regression analysis revealed an association between eGFR and all plasma biomarkers. Kidney dysfunction significantly impaired the specificity of several biomarkers (p-tau181, GFAP, NfL, and BD-tau; p < 0.001) for detecting cerebral amyloidosis (CSF Aβ42/Aβ40 < 7%), whereas p-tau217 was unaffected. It is important to note that ratio-based plasma biomarkers were not influenced by reduced kidney function.

Discussion

Impaired kidney function was linked to increased plasma cerebral amyloidosis biomarkers, but ratio-based measures (especially p-tau217/Aβ42) showed stable sensitivity and specificity for detecting cerebral amyloidosis across all eGFR groups. Additional studies including more patients with lower eGFR values in diverse diagnostic settings are needed to clarify the influence of kidney function on these biomarkers.

Classification of Evidence

This Class II evidence shows that kidney function influences individual blood biomarkers used for cerebral amyloidosis detection, but not their ratios.

Trial Registration Information

NCT05427448.

Introduction

Over the past decade, CSF biomarkers, in particular amyloid peptides (Aβ40 and Aβ42), tau protein, and its phosphorylated isoforms (p-tau), have become central to the diagnosis of neurologic diseases (NDs), enabling precise patient stratification based on amyloidosis, tauopathy, and neurodegeneration (the ATN framework).¹ Recent advancements in analytical technologies now allow the detection of these biomarkers in the plasma, and the high performance of p-tau181 and p-tau217 isoforms to differentiate Alzheimer disease (AD) from other neurodegenerative conditions has been widely described.^2-5 The use of the plasma p-tau217/Aβ42 ratio, however, has been shown to yield greater diagnostic precision in detecting amyloid-beta (Aβ) positivity than assessing either biomarker independently.^6,7 The recent US Food and Drug Administration approval of the Lumipulse G1200 test, which measures the plasma p-tau217/Aβ42 ratio, marks a significant step forward in the development of noninvasive and broadly accessible diagnostic methods. This advancement highlights the promise of plasma biomarker ratios in enabling earlier and more precise detection of AD and supports their growing role in both clinical practice and research, offering a practical alternative to PET imaging and CSF analysis.

Other plasma biomarkers have also demonstrated strong utility for detecting and monitoring neurologic damage and ND, such as glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), and brain-derived tau (BD-tau). GFAP is an intermediate filament protein expressed by astrocytes, whose concentration increases in CSF and blood after astrocytic activation or injury, as observed in AD.⁸ NfL is a structural component of neuronal axons, released into CSF and blood in response to neuronal injury or degeneration. Increased NfL concentrations serve as a sensitive indicator of neuroaxonal damage across a wide range of NDs.⁹ BD-tau is a neuron-specific isoform of tau protein that originates exclusively from the CNS and can be measured in blood. It reflects brain tau pathology more accurately than total tau, making it a promising biomarker for AD and other tauopathies.¹⁰

Although plasma-based biomarkers for ND have shown strong reliability, their broader implementation depends on comprehensive standardization to overcome context-specific challenges.¹¹ Persistent challenges include accounting for confounding factors and comorbidities that influence biomarker levels, as well as the necessity of validating results across diverse and representative populations. In particular, various studies have indicated that individuals with chronic kidney disease (CKD) exhibited elevated plasma levels of p-tau217, p-tau181, Aβ40, and Aβ42.^12-15 Notably, the link between CKD and increased plasma p-tau drew particular interest, as the difference in p-tau levels between those with and without CKD was comparable in magnitude to the difference observed between individuals classified as amyloid-PET positive and negative.¹² However, the use of plasma protein ratios has been proposed as a strategy to minimize the impact of CKD on plasma biomarker concentrations, based on the assumption that both components of the ratio are similarly influenced by renal function. Supporting this approach, recent mass spectrometry studies have shown that ratios such as p-tau to total-tau provide more consistent markers of AD pathology across various clinical stages.¹⁶ This pattern also extends to other AD-related biomarkers, notably the Aβ42/40 ratio.¹⁷ Thus, plasma protein ratios seem to be a promising method for reducing CKD-related confounding in the interpretation of plasma-based ND biomarkers.

In this study, we used the prospective multicenter ALZAN cohort of patients from memory clinics to investigate the influence of renal function, measured by estimated glomerular filtration rate (eGFR), on a broad spectrum of plasma biomarkers of neurodegenerative disorders (including ratios). Another objective was to assess whether impaired kidney function influences the diagnostic performance of plasma biomarkers in detecting cerebral amyloidosis, which is central in AD diagnosis.

Methods

Study Population

Patients seen in the memory outpatient clinic of the ALZAN prospective and multicenter (Montpellier University Hospital, Nimes University Hospital, and Perpignan Hospital) cohort were included in this study from November 2022 to July 2024. Patients were mainly referred to a memory consultation for memory concern by their general practitioner, or in some rare cases by neurologists or by themselves. Inclusion criteria for this study and for participation in the ALZAN cohort comprised the following elements: age older than 18 years, presence of a memory concern, patient informed consent, and realization of concomitant CSF and blood sampling. CSF AD biomarkers and creatinine measurements were measured in clinical routine practice settings. Additional blood tubes were collected for measurement of blood neurodegeneration biomarker for research purposes, with informed patient consent. At the time of sampling, the following baseline medical information was collected from medical records and stored in a specific database according to the ALZAN cohort recruitment protocol: age, collected data on sex, body mass index, Mini-Mental State Examination (MMSE) score, and suspected diagnostics established by the senior neurologist at the end of the consultation. MMSE was performed by a senior neurologist or neuropsychologist at the time of the inclusion in the ALZAN cohort or at most 1 year before or after the inclusion, as national guidelines for memory center advise to perform 1 MMSE per year. The consensual French version of the MMSE test was used with a score out of 30 points.¹⁸ Diagnostic suspicions were established according to the following guidelines: National Institute on Aging–Alzheimer's Association (NIA-AA) criteria for AD,¹⁹ Petersen criteria for mild cognitive impairment (MCI),²⁰ international behavioral variant frontotemporal dementia (FTD) criteria consortium for FTD,²¹ Dementia with Lewy bodies consortium for Lewy body disease,²² Boston criteria version 2.0 for cerebral amyloid angiopathy,²³ and Working Group of the Subjective Cognitive Decline Initiative criteria for subjective cognitive impairment.²⁴

Laboratory Measurement

Blood samples were collected on tubes with bipotassic ethylenediaminetetracetic acid (EDTA-K2) anticoagulant (BD vacutainer), whereas CSF samples were collected in Starsted polypropylene low-bind tubes. All tubes were centrifuged at 2000g for 10 minutes and then divided into 0.5–1 µL aliquots conserved at −80°C until analysis in a storage facility with continuous temperature surveillance. Aliquots thawed at +4°C and gently homogenized were used for each assay. One aliquot was used per analysis, so that the samples would only undergo 1 freeze/thaw cycle. The delay between collection and processing in the laboratory (referred to as blood delay) was recorded and exploited in another study.²⁵ Simultaneously with the collection of the EDTA-K2 blood sample, a lithium heparin tube was obtained and sent to the university hospital's biochemistry laboratory within 24 hours at room temperature, as recommended by the preanalytical guidelines for creatinine measurement. Plasma creatinine was measured by a standardized traceable enzymatic colorimetric method on Roche Cobas. Estimated glomerular filtration rate (eGFR) was immediately calculated using this creatinine measurement, sex, and age at the time of sampling with Chronic Kidney Disease Epidemiology Collaboration (CKD-Epi) 2021 equation²⁶ (eTable 1). CSF and plasma phospho-tau217 (p-tau217) assays were performed on Fujirebio Lumipulse while research use only Elecsys assays on Roche Cobas were used for plasma amyloid beta-peptide (1–40 and 1–42) (Aβ40 and Aβ42), phospho-tau181 (p-tau181), NfL, and GFAP measurements. Plasma BD-tau levels were determined using the Quanterix Simoa BD-tau Advantage Plus assay on an HD-X platform. The following ratios expressed in percentages of plasma biomarkers were studied for each patient: Aβ42/Aβ40, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42.

Statistical Analysis

Patients were divided into 3 groups (eGFR < 60, eGFR [60–90], and eGFR > 60) according to their kidney function estimation based on creatinine, using the eGFR thresholds of 60 and 90 mL/minute/1.73 m², as described by international recommendations.²⁷

Quantitative data are expressed as mean ± SD in text. Review of density plots and mean vs median comparison were used to assess data normality. Means between the 3 eGFR groups were compared using the 1-way analysis of variance (ANOVA) test. If the ANOVA test highlighted statistically significant differences, post hoc Student t tests with Benjamini-Hochberg correction were used to compare means between each group. Two-way ANOVA was conducted to examine the effect of eGFR level (divided in the 3 groups previously mentioned) and presence of cerebral amyloidosis on the different quantitative variables. For ANOVA analysis, p values were reported with effect sizes (η²), representing the proportion of total variance explained by each factor, with the following interpretations: η² < 0.01 = very small effect, η² between 0.01-0.06 = small effect, η² between 0.06 and 0.14 = medium effect, and η² ≥ 0.14 = large effect. Proportions, sensitivities, and specificities were compared with the χ² test (with Yates correction for continuity when samples sizes were low). The Passing-Bablok procedure was used to determine the slope (a) and the intercept (b) of the linear equation: plasma biomarker = a*eGFR + b. Pearson determination coefficients (R²) were determined to compare plasma biomarker levels with eGFR, and t tests for Pearson correlation were used to evaluate correlation significance. Cerebral amyloidosis (defined by a CSF Aβ42/Aβ40 ratio <7%), detection capabilities of 5 plasma biomarkers (p-tau217, p-tau181, NfL, GFAP, and BD-tau), and 4 plasma biomarker ratios (Aβ42/Aβ40, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42) were assessed, with sensitivity and specificity expressed as 95% CI in the whole cohort and in the 3 eGFR groups. Optimal cutoffs were determined in the whole cohort using plasma biomarker levels that yielded maximum Youden index values. For all tests, alpha significance level was set at 5% and a p value under 0.05 was considered statistically significant.

To investigate the association between plasma biomarkers and clinical-biological parameters, we performed univariable linear regression analyses using the different plasma biomarkers as the outcome variables. Each parameter was standardized (z-score transformation) before analysis to allow for effect size comparison across variables with different measurement scales. Regression coefficients and their corresponding 95% CIs were extracted to estimate the strength and direction of associations. Missing data were handled by listwise deletion (complete case analysis).

Standard Protocol Approvals, Registrations, and Patient Consents

The ALZAN cohort (ClinicalTrials.gov identifier: #NCT05427448) has been registered with the French regulatory authorities after obtaining the necessary authorizations. The sponsor is the Montpellier University Hospital. Written informed consent was obtained from all patients, and the study was approved by the ethics committee. Additional information on the cohort is provided elsewhere.⁶ Our research is noninterventional and falls under the category of observational studies.

Data Availability

Data and informed consent forms are available on request (CHU Montpellier). Requests will be considered by each study investigator, based on the information provided by the requester, regarding the study and analysis plan. If the use is appropriate, a data-sharing agreement will be put in place before distributing a fully deidentified version of the data set, including the data dictionary used for analysis with individual participant data.

Results

Participants

A total of 423 patients were included in the ALZAN cohort and were eligible to participate in this study. Three patients were excluded, 1 patient with missing plasma creatinine value, 1 patient with missing plasma Aβ40 value, and another 1 patient with missing NfL value. A total of 420 patients were, therefore, included (mean age: 71.1 years, %female/male: 53.3/46.7%), with 36 patients having an eGFR under 60 mL/minute/1.73 m² (eGFR < 60 group), 194 patients having an eGFR between 60 and 90 mL/minute/1.73 m² (eGFR [60–90] group), and 190 patients having an eGFR above 90 mL/minute/1.73 m² (eGFR > 90 group). Baseline characteristics of the eGFR groups are presented in Table 1. The 3 groups were comparable regarding sex, body mass index (BMI) and MMSE score, CSF biomarker values, and suspected diagnostics (Alzheimer disease being the most prevalent). However, mean age was significantly lower in eGFR > 90 group than in the eGFR [60–90] and eGFR < 60 groups (66.5 ± 7.8 vs 74.6 ± 7.8 and 76.6 ± 6.9, p < 0.001 for both). Cerebral amyloidosis defined by a CSF Aβ42/Aβ40 ratio <7% was present in 66.7%, 58.8%, and 55.3% of patients in the eGFR < 60, eGFR [60–90], and eGFR > 90 groups, respectively (p = 0.420).

Table 1.

General Characteristics of the Population

	eGFR <60 (n = 36)	eGFR [60–90] (n = 194)	eGFR >90 (n = 190)
Age at baseline (y)	76.6 ± 6.9	74.6 ± 7.8	66.5 ± 11.5
Sex (%)
F	52.8	51	55.8
M	47.2	49	44.2
Body mass index (kg/m2)	24.7 ± 5.5	24.8 ± 7.8	24.6 ± 10.4
MMSE score	22.4 ± 7.9	23 ± 8	23.1 ± 10.4
Cerebrospinal fluid biomarkers
Aβ42/Aβ40 < 7% (%)	66.7	58.8	55.3
tau (pg/mL)	589.2 ± 436.3	479.1 ± 297.9	472 ± 363.4
p-tau181 (pg/mL)	83.5 ± 67.5	71.4 ± 52.3	69.3 ± 60
eGFRcreat (ml/minute/1.73 m2)	48.2 ± 2.1	78.7 ± 7.9	97.4 ± 6.1
Suspected diagnosis (n, (%))
Alzheimer disease	18 (50)	90 (46.4)	91 (47.9)
Mild cognitive impairment	3 (8.3)	19 (9.8)	23 (12.1)
Amyloid angiopathy	1 (2.8)	5 (2.6)	4 (2.1)
Subjective cognitive impairment	4 (11.1)	16 (8.2)	12 (6.3)
Vascular dementia	1 (2.8)	4 (2.1)	10 (5.3)
Lewy body disease	2 (5.6)	7 (3.6)	4 (2.1)
Mixed dementia	1 (2.8)	9 (4.6)	1 (0.5)
Hydrocephalus	2 (5.6)	8 (4.1)	4 (2.1)
Frontotemporal dementia	1 (2.8)	18 (9.3)	7 (3.7)
Parkinsonian syndrome	0 (0)	3 (1.5)	6 (3.2)
Others	3 (8.3)	15 (7.7)	28 (14.7)

Patients are classified in 3 groups according to their estimated glomerular filtration rate (eGFR) based on creatinine value (ml/minute/1.73 m²). Quantitative values are indicated as mean ± SD.

Plasma Biomarker Values in eGFR Groups and Interaction With Cerebral Amyloidosis

Density plots and mean-median comparison showed that all plasma biomarkers, creatinine, eGFR, and all other quantitative clinical data collected were normally distributed. Boxplot representation of plasma biomarker values in the 3 eGFR groups is illustrated in Figure 1, and numerical values are listed in Table 2. Plasma Aβ40 and Aβ42 peptides, plasma GFAP, plasma BT-tau, and serum NfL were affected by decreased eGFR starting at 90 mL/minute/1.73 m², with levels significantly higher for the eGFR < 60 group compared with eGFR [60–90] and eGFR > 90 groups as well as for eGFR [60–90] compared with eGFR > 90 groups. A significant increase in p-tau181 and p-tau217 plasma levels was observed only for eGFR values below 60 mL/minute/1.73 m², with levels significantly higher for the eGFR < 60 group compared with eGFR [60–90] and eGFR > 90 groups, but not for eGFR [60–90] and eGFR > 90 groups. Plasma biomarker ratios Aβ42/Aβ40, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42 were not statistically different between the 3 eGFR groups. The largest effect sizes of eGFR on plasma biomarkers were observed for plasma amyloid peptides (Aβ42, η² = 0.17, Aβ40, η² = 0.183), whereas the largest effect sizes of cerebral amyloidosis presence were observed for p-tau217, p-tau181, Aβ42/Aβ40 ratio, p-tau217/Aβ42 ratio, NfL/p-tau217 ratio, and p-tau181/Aβ42 ratio plasma levels (η² > 0.2 for all). Examination of the interaction between eGFR levels and cerebral amyloidosis presence on different blood biomarker levels revealed only small effect sizes for p-tau217, amyloid peptides, Aβ42/Aβ40 ratio, and GFAP plasma levels, with η² values ranging between 0.01 and 0.06.

Figure 1. Repartition of Blood Biomarkers According to eGFRcreat Value (Expressed in Ml/min/1.73m2) Groups: eGFR < 60 (Gray), eGFR 60–90 (Light Gray), eGFR > 90 (Dark).

Boxplot representation is used, values are represented by a point, and the following values are indicated by horizontal bars from bottom to top: minimum, first quartile (Q1), median, third quartile (3Q), and maximum. Minimum and maximum are defined as Q1-1.5*IQR and Q3+1.5*IQR, respectively (IQR: interquartile range). p Value of the 1-way ANOVA or Student t test with Benjamini-Hochberg correction is indicated for each biomarker. ***: p < 0.001, **: p < 0.01, *: p < 0.05, ns = nonsignificant.

Table 2.

Repartition of Blood Biomarker Levels (Means ± SDs) Across eGFR Groups and Cerebral Amyloidosis (CSF Aβ42 to Aβ40 < 7%)–Positive (Aβ+) and Negative (Aβ-) Patients

Blood biomarker	eGFR<60 (n = 36)	eGFR (60–90) (n = 194)	eGFR>90 (n = 190)	p Valuea	η2a	Aβ+ (n = 243)	Aβ- (n = 177)	p Valuea	η2a	p Valueb	η2b
p-tau217 (pg/mL)	0.63 ± 0.64	0.39 ± 0.35	0.38 ± 0.41	**	0.028	0.59 ± 0.43	0.16 ± 0.24	***	0.257	*	0.012
Aβ42 (pg/mL)	45 ± 10.1	35.4 ± 7.5	32.5 ± 7	***	0.17	32.3 ± 7.2	38.4 ± 8.3	***	0.132	*	0.011
Aβ40 (pg/mL)	362.1 ± 69.5	304.2 ± 48.3	279.9 ± 43.2	***	0.183	301.4 ± 50.2	293.7 ± 57.1	ns	0.005	**	0.022
Aβ42/40 (%)	12.5 ± 2	11.7 ± 2.1	11.7 ± 2.6	ns	0.01	10.7 ± 1.7	13.2 ± 2.2	***	0.282	*	0.011
p-tau217/Aβ42 (%)	1.5 ± 1.6	1.3 ± 1.4	1.3 ± 1.4	ns	0.003	1.9 ± 1.5	0.4 ± 0.6	***	0.272	ns	0.004
NfL/p-tau217 (%)	1962 ± 1778	2239 ± 2563	2421 ± 3059	ns	0.002	1133 ± 1229	3897 ± 3377	***	0.248	ns	0.002
p-tau181 (pg/mL)	1.9 ± 1.1	1.3 ± 0.7	1.1 ± 0.7	***	0.085	1.6 ± 0.8	0.9 ± 0.4	***	0.240	ns	0.008
p-tau181/Aβ42 (%)	4.6 ± 2.9	3.9 ± 2.5	3.8 ± 2.8	ns	0.006	5.2 ± 2.8	2.3 ± 1.2	***	0.286	ns	0.007
NfL (pg/mL)	6.1 ± 4.5	4.4 ± 2.5	3.7 ± 3.1	***	0.047	4.1 ± 2.1	4.3 ± 4.1	ns	0.001	ns	0.007
GFAP (pg/mL)	175.8 ± 77.4	143.7 ± 60.8	117.3 ± 58.3	***	0.079	158 ± 58.7	102.3 ± 55.8	***	0.186	*	0.011
BD-tau (pg/mL)	16 ± 7.8	11.3 ± 4.8	9.7 ± 4.7	***	0.109	12 ± 5.2	9.6 ± 5.3	***	0.051	ns	0.004

^ap Values of 1-way ANOVA with corresponding size effects η² are shown for biomarker comparison between eGFR and cerebral amyloidosis groups.

^bp Values of 2-way ANOVA with corresponding size effects η² examining the interaction between eGFR levels and cerebral amyloidosis presence on the different blood biomarkers are indicated.

For p values, ns: nonsignificant, *: p < 0.05, **: p < 0.01, ***: p < 0.001.

Regression and Correlation Between eGFR and Plasma Biomarkers

Passing-Bablok regression with R² coefficient determination in patients with (Aβ+, n = 243) and without (Aβ-, n = 177) cerebral amyloidosis is presented in Figure 2. In both Aβ+ and Aβ- patients, a clear negative correlation between eGFR and plasma Aβ40, Aβ42, p-tau181, GFAP, BD-tau, and serum NfL was observed. By contrast, no statistically significant correlation was observed in both groups between eGFR and plasma p-tau217, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42. A discrepancy was observed for Aβ42/Aβ40 because eGFR was significantly negatively correlated with the ratio for Aβ+ patients but not for Aβ− patients. Univariable linear regression analysis between eGFR (ml/minute/1.73 m²) and plasma biomarkers (Figure 3) showed that eGFR was significantly associated with Aβ40, Aβ42, p-tau181, GFAP, and BD-tau levels. This association was not found between eGFR and p-tau217, Aβ42/Aβ40, p-tau217, p-tau217/Aβ42, NfL/p-tau217, and p-tau181/Aβ42 ratios.

Figure 2. Regression and Correlation Between Estimated Glomerular Filtration Rate Based on Creatinine (eGFRcreat, x-Axis, Ml/min/1.7m2) and Blood Biomarkers (y-Axis, Pg/mL) of Patients With (Dark Gray) and Without Cerebral Amyloidosis (Light Gray).

The linear equation plasma biomarker = a*eGFR + b determined using the Passing-Bablok procedure is indicated and represented by thick lines. Thin lines represent the 95% CIs of the mean values. Pearson determination coefficient (R²) is also indicated, with a p value of the t test for Pearson correlation: ***: p < 0.001, **: p < 0.01, *: p < 0.05, ns = nonsignificant.

Figure 3. Univariable Linear Regression Analysis Between Blood Biomarkers and Clinical-Biological Parameters.

Regression coefficients and their corresponding 95% CIs are represented on forest plots. (A) Plasma Aβ40, (B) plasma Aβ42, (C) plasma Aβ42/40 ratio, (D) plasma p-tau181, (E) plasma p-tau181/Aβ42 ratio, (F) plasma p-tau217, (G) plasma p-tau217/Aβ42, (H) plasma p-tau217/NfL, (I) plasma NfL, (J) plasma GFAP, and (K) plasma BD-tau.

Diagnostic Performances of Plasma Biomarkers According to the eGFR Group

Cerebral amyloidosis detection ability of plasma biomarkers in the whole cohort and according to the eGFR group is summarized in Table 3. p-tau217/Aβ42 ratio was the best overall performing marker, given the highest combined sensitivity (0.89 [0.86–0.92]) and specificity (0.86 [0.83–0.89]). Specificity rates decreased gradually and significantly with worsening eGFR for plasma Aβ42/Aβ40, p-tau181, GFAP, BD-tau, and serum NfL. This was, however, not the case for plasma p-tau217, p-tau217/Aβ42 and NfL/p-tau217, and p-tau181/Aβ42 ratios.

Table 3.

Blood Biomarkers-eGFR Creatinine (ml/minute/1.73 m²), Cerebral Amyloidosis (CSF Aβ42 to Aβ40 Ratio <7%), Detection Performance Parameters, Sensitivity (Se), and Specificity (Sp), With 95% CIs

Blood biomarker	Threshold	WP		eGFR<60		eGFR (60–90]		eGFR>90		p Value
		Se	Sp	Se	Sp	Se	Sp	Se	Sp	Se	Sp
p-tau217 (pg/mL)	0.22	0.84 (0.8–0.88)	0.88 (0.85–0.91)	0.96 (0.89–1.02)	0.92 (0.83–1)	0.83 (0.78–0.89)	0.83 (0.77–0.88)	0.82 (0.76–0.87)	0.93 (0.89–0.97)	ns	ns
Aβ42/40 (%)	11.6	0.77 (0.73–0.81)	0.79 (0.75–0.82)	0.54 (0.38–0.70)	0.75 (0.61–0.89)	0.77 (0.71–0.83)	0.76 (0.7–0.82)	0.83 (0.77–0.88)	0.81 (0.76–0.87)	ns	ns
p-tau217/Aβ42 (%)	0.54	0.89 (0.86–0.92)	0.86 (0.83–0.89)	1 (1–1)	0.92 (0.83–1)	0.89 (0.84–0.93)	0.8 (0.74–0.86)	0.87 (0.82–0.92)	0.89 (0.85–0.94)	ns	ns
NfL/p-tau217 (%)	6.19	0.83 (0.8–0.87)	0.85 (0.82–0.89)	0.88 (0.77–0.98)	1 (1–1)	0.83 (0.78–0.89)	0.84 [0.79–0.89)	0.82 (0.76–0.87	0.85 (0.8–0.9)	ns	ns
p-tau181 (pg/mL)	1.03	0.83 (0.8–0.87)	0.76 (0.72–0.8)	0.92 (0.83–1)	0.42 (0.26–0.58)	0.83 (0.78–0.89)	0.69 (0.62–0.75)	0.81 (0.75–0.87)	0.88 (0.84–0.93)	ns	***
p-tau181/Aβ42 (%)	3.17	0.82 (0.79–0.86)	0.88 (0.84–0.91)	0.92 (0.83–1)	0.92 (0.83–1)	0.80 (0.74–0.85)	0.85 (0.8–0.9)	0.83 (0.77–0.88)	0.88 (0.84–0.93)	ns	ns
NfL (pg/mL)	2.48	0.89 (0.86–0.92)	0.35 (0.3–0.4)	0.96 (0.89–1)	0 (0–0)	0.88 (0.83–0.92)	0.18 (0.12–0.23)	0.89 (0.84–0.93)	0.56 (0.49–0.64)	ns	***
GFAP (pg/mL)	98.2	0.88 (0.85–0.91)	0.58 (0.53–0.63)	0.96 (0.89–1)	0.42 (0.26–0.58)	0.89 (0.85–0.94)	0.43 (0.36–0.49)	0.84 (0.79–0.89)	0.75 (0.69–0.81)	ns	***
BD-tau (pg/mL)	7.92	0.85 (0.82–0.89)	0.45 (0.4–0.5)	0.96 (0.89–1)	0.08 (0–0.17)	0.84 (0.79–0.89)	0.33 (0.26–0.4)	0.85 (0.8–0.9)	0.62 (0.55–0.69)	ns	***

Thresholds were determined on the whole population (WP) and were tested afterward on each eGFR group. The Yates χ² test was used to compare sensitivities and specificities across eGFR groups. ***: p < 0.001, **: p < 0.01,*: p < 0.05, ns = nonsignificant.

Discussion

In this multicenter study, we highlight precise impact of impaired kidney function on plasma biomarkers of ND around the CKD classification thresholds of 60 and 90 mL/minute/1.73 m²²⁷ in a large group of patients seen during routine clinical care. The sample size allowed for the formation of 3 eGFR-based groups, which showed no significant differences in baseline clinical characteristics, except for age (Table 1). Our findings confirm that for all tested analytes, plasma levels tend to increase as eGFR decreases. However, the extent of this impact varies by biomarker (Figure 1): Aβ40, Aβ42, NfL, GFAP, and BD-tau were significantly affected as soon as eGFR dropped below 90 mL/minute/1.73 m², whereas p-tau217 and p-tau181 were affected only when eGFR decreased below 60 mL/minute/1.73 m². eGFR status seemed to have a particular impact on β-amyloid peptides, whereas the presence of cerebral amyloidosis seemed to influence the levels of p-tau217, p-tau181, the Aβ42/Aβ40 ratio, the p-tau217/Aβ42 ratio, the NfL/p-tau217 ratio, and the p-tau181/Aβ42 ratio (Table 2). The interaction of these 2 variables had only a small effect on plasma levels of p-tau217, amyloid peptides, the Aβ42/Aβ40 ratio, and GFAP.

Furthermore, a significant negative correlation between eGFR and Aβ40, Aβ42, p-tau181, NfL, GFAP, and BD-tau was observed for both cerebral amyloid positive and negative groups (Figure 2). Remarkably, this pattern did not apply to p-tau217, as no significant correlation with eGFR was observed in either group. However, a stronger trend toward an association was noted in the Aβ+ group compared with the Aβ– group. Moreover, Aβ42/Aβ40 ratio was negatively correlated with eGFR only for cerebral amyloid–positive patients, which is surprising and could mean that the plasma Aβ42/Aβ40 ratio can be falsely elevated in case of kidney function decrease. Sensitivity and specificity analyses (Table 3) confirmed that p-tau217 and the p-tau217/Aβ42 ratio exhibited the highest clinical accuracy in detecting amyloid-β positivity in our cohorts. In addition, we demonstrate that an eGFR below 90 or 60 mL/minute/1.73 m² significantly affects the specificity of established thresholds for cerebral amyloidosis detection using p-tau181, NfL, GFAP, and BD-tau. Once again, p-tau217 seems to be less affected by decreased eGFR, with its sensitivity and specificity remaining consistent across eGFR < 60, eGFR [60–90], and eGFR > 90 groups.

Regarding the use of plasma biomarker ratios (Aβ42/40, p-tau181/Aβ42, p-tau217/Aβ42, or p-tau181/NfL), as suggested by several recent articles for differentiation of AD and FTD,^6,28 they significantly limit the impact of eGFR plasma-biomarker overestimation. Indeed, no statistical significant differences were observed between eGFR groups regarding mean levels, correlation, sensitivity, and specificity (Table 3). These results further support the use of biomarker ratios in this context, as they offer more stable and reliable diagnostic performance irrespective of kidney function, making them particularly valuable in populations with varying degrees of renal impairment.

In summary, p-tau217 clearly emerges as the best performing plasma biomarker for amyloidosis detection while it seems to be modestly affected by reduced kidney function (notably when eGFR <60 mL/minute/1.73 m²). However, although this decreased eGFR can slightly elevate blood p-tau217 levels, this does not meaningfully compromise the diagnostic precision of the biomarker. A potential explanation is that plasma p-tau217 level increase in case of amyloidosis (3.6-mean fold change in our cohort) is greater than the increase induced by eGFR decrease (1.6-mean fold change in our cohort). Furthermore, the use of ratio and especially p-tau217/Aβ42 mitigates the impact of decreased kidney function, as shown in Figure 3.

A primary hypothesis to explain this observation is that, in CKD, the filtration of plasma biomarkers (particularly proteins) is impaired because of reduced glomerular clearance. Aβ40 and Aβ42, with molecular weights below 5 kDa, are small enough to be filtered under normal conditions and are, therefore, likely to accumulate when renal function declines. By contrast, NfL, GFAP, and tau proteins have molecular weights exceeding 50 kDa and are generally too large to pass through intact glomerular pores. However, proteolysis can lead to the release of smaller protein fragments into the circulation, which may then be subject to renal clearance and thus influenced by kidney function. Of interest, it has been shown recently that tau proteins are likely cleaved between amino acids 181 and 231.²⁹ This could be a mechanistic rationale for why we observed that p-tau181 seemed more affected by decreased eGFR than p-tau217.

An alternative or complementary explanation is that CKD itself may contribute to cerebral injury, thereby leading to increased levels of neurodegeneration-related biomarkers in the blood. Supporting this hypothesis, there is growing evidence linking CKD to the development of neurocognitive impairment.³⁰ Two principal mechanisms have been proposed: (1) cerebrovascular damage associated with CKD and (2) direct neurotoxicity from uremic toxins accumulating in the brain.³¹ Although p values comparing CSF biomarkers were not < 0.05, a tendency was observed as p-tau181, tau levels and proportion of individuals with Aβ42/Aβ40 ratio <7% seemed to be higher in lower eGFR groups.

Our results are consistent with previous studies, most of which have reported increased levels of plasma biomarkers when eGFR drops: this reinforces the notion that reduced kidney function has a significant effect on the concentration of amyloid and neurodegeneration-related biomarkers in the bloodstream.

This was extensively studied for p-tau181 and p-tau217 levels, for which many studies found a strong association between CKD and increased plasma levels.^12-16,32 Individual levels of Aβ40 and Aβ42 are also subject to overestimation, but a previous study showed that the Aβ42/Aβ40 ratio seems to be independent of renal function.¹⁷ We confirmed this finding only in cerebral amyloid–negative patients, in whom a clear association between reduced eGFR and increased plasma biomarker levels was observed. Of interest, in amyloid-positive patients, we found a surprising negative correlation, suggesting that the relationship between kidney function and plasma biomarkers may differ depending on underlying cerebral amyloid status. This may suggest that plasma Aβ42 are more sensitive to decreased renal function than Aβ40, as a greater mean increase was observed for Aβ42 than for Aβ40 between eGFR < 60 and other eGFR groups. Similar to our findings, another study³³ demonstrated that kidney function was a major determinant for Aβ42, Aβ40, NfL, and GFAP levels, but not for p-tau217.

NfL, GFAP, and BD-tau were greatly affected by kidney function decrease as well. Serum NfL is increasingly being used as a biomarker across a wide range of NDs such as multiple sclerosis (MS), FTD, and amyotrophic lateral sclerosis (ALS).^34-36 We found a mean serum NfL level of 6.1 ± 4.5 pg/mL in eGFR < 60 group that is similar to the cutoff for ALS diagnosis of 6 pg/mL used in our settings.³⁴ Levels of NfL in MS and FTD are generally lower than for ALS and are also likely influenced by decreased eGFR under 60 mL/minute/1.73 m². Impact of kidney function on GFAP levels could be an important factor to take into account, for instance in the context of traumatic brain injury.³⁷ In addition, GFAP has been described as a promising marker for AD-related outcomes, but its blood levels depend not only on astrocytic release but also on blood-brain barrier (BBB) permeability, glucocorticoid drugs, or physical activity, making it possible to observe high CSF but low blood levels when the BBB is intact, or conversely elevated blood levels when BBB disruption or altered clearance predominates.³⁸

Finally, BD-tau has emerged as a promising predictive biomarker for AD and seems to be less influenced by renal function compared with other plasma biomarkers.³⁹ By contrast, our study revealed that BD-tau is among the blood biomarkers most significantly affected by renal function.

We acknowledge that our study has several limitations. First, age is known to be a major confounding variable of plasma biomarker levels, notably for NfL.⁴⁰ Mean age was significantly higher in eGFR < 60 and eGFR [60–90] groups and may have accentuated the gap in concentrations with the eGFR > 90 group. Second, our population presented a relatively preserved renal function, with a small proportion of patients (8.6%) in the eGFR < 60 group, which defines CKD. This study was conducted in real-life routine care settings and included mainly patients aged 65 years or older. It is well known that lower muscle mass is commonly observed in geriatric populations⁴¹ and, therefore, eGFR based on creatinine that we used can be overestimated. Use of cystatin C, which is far less influenced by muscular mass,⁴² could be an interesting alternative that we unfortunately could not test in our study. In the end, use of reference glomerular filtration rate measurement with exogenous molecule injection such as iohexol would be necessary to precisely determine the effect of CKD on plasma biomarkers of neurodegeneration. Another point is that we chose to determine cutoffs for cerebral amyloidosis in the whole population before applying them in the eGFR subgroups as it would be performed in clinical practice. This could lead to overestimation of sensitivity and specificity of blood biomarkers, especially in the eGFR < 60 group, which is probably statistically disadvantaged. Lack of apolipoprotein E4 status was also a limitation because it can affect the interpretation of cerebral amyloïdosis biomarkers.⁴³ Finally, given that the biomarkers were quantified using different analytical platforms (Lumipulse for p-tau217 and Cobas for Aβ peptides and p-tau181), concerns could be raised regarding the potential impact on the p-tau217/Aβ42 ratio. Nonetheless, our findings show that all ratios exhibited robust reliability. In addition, despite inherent analytical differences across assays, p-tau217 maintains excellent performance in detecting cerebral amyloidosis on automated platforms, and the ratio's diagnostic value does not seem to be affected by this analytical consideration.

In conclusion, we found that even a moderate decrease in kidney function that we observed in our cohorts is associated with increased levels of all tested plasma biomarkers. Use of plasma biomarker ratio is particularly valuable because it remains unaffected by kidney function and can be used without adaptation to eGFR. In the context of cerebral amyloidosis detection, p-tau217 and above all the p-tau217/Aβ42 ratio are the most powerful biomarkers, even when eGFR falls below 60 mL/minute/1.73 m². Further studies including more patients with glomerular filtration rates below 60 and 30 mL/minute/1.73 m², as well as across diverse diagnostic contexts, are needed to fully clarify and rule out the effect of kidney function on these plasma biomarkers.

Glossary

AD

Alzheimer disease

ALS

amyotrophic lateral sclerosis

BBB

blood-brain barrier

CKD

chronic kidney disease

EDTA-K2

ethylenediaminetetracetic acid

eGFR

estimated glomerular filtration rate

FTD

frontotemporal dementia

GFAP

glial fibrillary acidic protein

MMSE

Mini-Mental State Examination

MS

multiple sclerosis

ND

neurologic disease

NfL

neurofilament light chain

Author Contributions

E. Mondesert: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; analysis or interpretation of data. J-P. Cristol: drafting/revision of the manuscript for content, including medical writing for content. A-S. Bargnoux: drafting/revision of the manuscript for content, including medical writing for content. M. Duchiron: major role in the acquisition of data. G.U.Busto: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data. C. Hirtz: drafting/revision of the manuscript for content, including medical writing for content. G. Barnier-Figue: drafting/revision of the manuscript for content, including medical writing for content. F. Perrein: drafting/revision of the manuscript for content, including medical writing for content. C. Turpinat: drafting/revision of the manuscript for content, including medical writing for content. S. Jurici: drafting/revision of the manuscript for content, including medical writing for content. A. Gabelle: drafting/revision of the manuscript for content, including medical writing for content. K. Bennys: drafting/revision of the manuscript for content, including medical writing for content. S. Lehmann: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data. C. Delaby: drafting/revision of the manuscript for content, including medical writing for content; major role in the acquisition of data; study concept or design; analysis or interpretation of data.

Study Funding

The work was funded by the Fondation Research Alzheimer (ALZAN and RENALZ projects), the AXA Mécénat Santé (INTERVAL Project), and the Fondation pour la Recherche Médicale (FRM, team Proteinopathies). None of the funding bodies had any role in study design, in the collection, analysis, and interpretation of data, in the writing of the report, or in the decision to submit the paper for publication.

Disclosure

S. Lehmann reports serving as Advisory Board/Consultant for Roche diagnostics, Biogen, Lilly, and Fujirabio. A. Gabelle reports serving as Advisory Board/Consultant for Biogen, Lilly, and Esai. No other authors report relevant disclosures. Go to Neurology.org/N for full disclosures.