Effect of renal function on the diagnostic performance of plasma biomarkers for Alzheimer's disease

BackgroundSeveral blood-based biomarkers are promising to be used in the diagnosis of Alzheimer's disease (AD) including A beta 42/40, p-tau181, and neurofilament light (NfL). The kidney is associated with the clearance of proteins. It is crucial to evaluate the effect of renal function on the diagnostic performance of these biomarkers before clinical implementation, which is important for the establishment of reference ranges and the interpretation of results. MethodsThis study is a cross-sectional analysis based on ADNI cohort. Renal function was determined by the estimated glomerular filtration rate (eGFR). Plasma A beta 42/40 was measured by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Plasma p-tau181 and NfL were analyzed by Single Molecule array (Simoa) technique. [18F] florbetapir-PET (A beta-PET) was used as a reference standard to estimate the brain amyloid load. The cutoff of A beta-PET positivity was defined as >= 1.11. Linear regression models were used to investigate the associations of continuous eGFR with each plasma biomarker separately. The diagnostic accuracies of plasma biomarkers for positive brain amyloid across different renal function groups were analyzed by Receiver operating characteristic (ROC) curve. Youden-Index was used to determine the cutoff levels. ResultsA total of 645 participants were included in this study. The levels and diagnostic performance of A beta 42/40 were not affected by renal function. eGFR was only found negatively associated with p-tau181 levels in A beta-PET negetive sample (beta = -0.09, p = 0.039). eGFR was found negatively associated with NfL levels both in whole sample and A beta-PET stratified groups (beta = -0.27, p < 0.001 in whole sample; beta = -0.28, p = 0.004 in A(-); beta = -0.27, p < 0.001 in A(+)). The diagnostic accuracies of p-tau181 and NfL were not affected by renal function. But the cutoff values of p-tau181 and NfL changed in participants with mild to moderate eGFR decline compared to participants with normal eGFR. ConclusionPlasma A beta 42/40 was a robust biomarker for AD which was not affected by renal function. Plasma p-tau181 and NfL levels were affected by renal function, specific reference values of them should be considered in populations with different renal function stages.