Concordance of CSF AD biomarkers between intraventricular and lumbar CSF in iNPH

Abstract

AbstractBackgroundThe core Alzheimer’s Disease cerebrospinal fluid (CSF) biomarkers, amyloid β 1‐42 (Aβ42), total tau (T‐tau) and phosphorylated tau 181 (P‐tau181), have presented added value for differential diagnostics in the field of neurodegenerative diseases. Moreover, the concentration gradient of these biomarkers between intraventricular (V‐CSF) and lumbar region CSF (L‐CSF) has been partly established for idiopathic normal pressure hydrocephalus (iNPH) patients. In this study, we aim to provide conversion factors between V‐CSF and L‐CSF.MethodAltogether 138 patients were diagnosed for possible iNPH and received ventriculoperitoneal shunt. Lumbar CSF samples were obtained 3‐months prior shunt surgery. Additional intraoperative interventricular CSF samples were obtained from 97 patients. Furthermore, follow‐up samples of 41 patients were obtained 0‐, 3‐, 6‐ and 18‐months (0M, 3M, 6M, 18M) postoperatively by shunt valve and lumbar punctures. Aβ42, T‐tau and P‐tau181 CSF concentrations of samples collected were analyzed using Elecsys (Cohort of 97 patients) and Innotest (Cohort of 41 patients) assays.ResultPreoperative L‐CSF Aβ42, T‐tau and P‐tau181 associated to intraoperative V‐CSF (Sprearman ρ=0.54 Aβ42, r=0.34 T‐tau, r=0.55 P‐tau, all p<0.001). Strong associations were seen between postoperative L‐ and V‐CSF for all biomarkers in the every follow‐up sampling point (Spearman ρ Aβ42: 0M=.77, 3M=.88, 6M=.83, 18M=.84) (T‐tau: 0M=.92, 3M=.91, 6M=.92, 18M=.94) (P‐tau181: 0M=.96, 3M=.95, 6M=.94, 18M=.94) (all p<0.0001). Regression equations were determined for intraoperative V‐CSF and preoperative L‐CSF (Aβ42: V‐CSF=185+0.34*L‐CSF, T‐tau: Ln(V‐CSF)=3.11+0.49*Ln(L‐CSF) and P‐tau181: V‐CSF=8.2+0.51*L‐CSF, all p<0.01), and for postoperative V‐CSF and L‐CSF (Aβ42: V‐CSF=86.7+0.75*L‐CSF, T‐tau: V‐CSF=86.9+0.62*L‐CSF and P‐tau181: V‐CSF=1.9+0.75*L‐CSF all p<0.0001).ConclusionThe Aβ42, T‐tau and P‐tau181 showed linear dependency between ventricular and lumbar CSF for preoperative L‐CSF and intraoperative V‐CSF as well as for postoperatively collected L‐CSF and V‐CSF. The regression equations based on these findings provide a novel tool to use V‐CSF for diagnostics and prognostics entities that rely on the lumbar CSF concentrations.