Plasma p‐tau181 for the differential diagnosis of cognitive impairment in a prospective, daily clinical practice cohort

Abstract Background Currently used biomarkers for the differential diagnosis of cognitive impairment are expensive and/or relatively invasive, limiting their availability to the general population. Blood protein biomarkers have showed promising results for screening, differential diagnosis, and prognosis. We aimed to study the diagnostic performance of plasma p‐tau181 in a daily clinical practice, prospective memory clinic cohort. Method All patients referred for a first clinical evaluation with suspected cognitive impairment between January 1 st , 2020 and March 30 th , 2021, were invited to participate in the study. Plasma p‐tau181 was measured using SIMOA technology (Quanterix). Clinical diagnoses were made following current diagnostic criteria and blinded to plasma p‐tau181 results. Result A total of 232 participants were recruited (mean age 69y, mean MMSE 24), including 25 cognitively unimpaired (CU) controls (mean age 67y, MMSE 28). Clinical diagnoses were AD (82 subjects, 43 of them with prodromal AD [CDR = 0.5]), 83 non‐neurodegenerative cognitive impairment (non‐ND, 75 of them CDR = 0.5), 22 frontotemporal dementia (FTD) and 20 Lewy body disease (LBD). Plasma p‐tau181 levels were statistically higher in AD (mean 2.39 pg/mL) compared with CU (mean 1.09 pg/mL), non‐ND (mean 1.43 pg/mL), FTD (mean 1.69 pg/mL) and LBD (mean 1.64 pg/mL) with a relatively good diagnostic performance for the differential diagnosis between AD and CU, non‐ND, LBD and FTD (AUC of 0.89, 0.80, 0.73 and 0.71, respectively). 128 subjects (55% of the cohort) had specific AD biomarkers (CSF or PET) available. In this subgroup, p‐tau181 differentiated AD from CU and non‐ND (AUC 0.91 and 0.92, respectively) and prodromal AD from CDR = 0.5 non‐ND participants (2.32 vs 1.02 pg/mL, p<0.001, AUC 0.90). Plasma p‐tau181 discriminated between a positive and negative amyloid beta status (defined by CSF/PET) with an AUC of 0.87 and AD subjects from those with a clinical diagnosis of LBD and FTD who had a negative amyloid beta status (AUC of 0.94 and 0.86, respectively). Plasma p‐tau181 correlated with CSF p‐tau181 (r s = 0.48, p<0.001). Conclusion In our cohort of everyday clinical practice, plasma p‐tau181 is an accurate biomarker for predicting the AD pathophysiological process and discriminating AD from other neurodegenerative and non‐neurodegenerative causes of cognitive impairment.