AbstractBackgroundThere are currently excellent fluid biomarkers of brain amyloidosis, including CSF Aβ42/Aβ40 and CSF pT217/T217 (%). However, many individuals with brain amyloidosis do not have symptoms of Alzheimer disease (AD). Biomarkers that are strongly associated not only with brain amyloidosis but also with symptomatic AD would be invaluable tool in clinical trials and clinical diagnosis. We evaluated whether CSF biomarkers in an immunoassay panel added value to existing biomarkers in prediction of cognitive impairment.MethodAn immunoassay panel from Rules Based Medicine was used to measure CSF levels of 52 proteins including plasminogen inhibitor 1 (PAI1). CSF Aβ42/Aβ40 was measured by Lumipulse, and CSF pT217/T217 (%) was measured by an immunoprecipitation‐mass spectrometry assay. Logistic regression and linear models were used to evaluate the association between biomarkers and cognitive status (cognitively unimpaired or symptomatic AD as defined by a Clinical Dementia Rating [CDR] of 0.5 or greater), the Mini‐Mental State Exam [MMSE] score, and CDR Sum of Boxes [CDR‐SB]. Models included the covariates of age, self‐identified gender, and years of education.ResultThe cohort included n = 179 older research participants (average age 73.0 years, standard deviation 6.6 years), of which 47 (26%) had cognitive impairment (CDR>0). Cognitively impaired individuals had higher levels of PAI1 (p = 1.18e‐05) (Fig. 1A). PAI1 levels predicted cognitive status in the entire cohort (p = 4e10−3) and the sub‐cohort with CSF Aβ42/Aβ40 levels consistent with brain amyloidosis (Aβ42/Aβ40<0.0673; p = 0.003). Even when CSF pT217/T217 (%) was included in the model of cognitive status, PAI1 was a significant independent predictor of cognitive status in both the entire cohort (p = 0.005) and the sub‐cohort with brain amyloidosis (p = 0.008) (Fig. 1B). Both the MMSE and CDR‐SB were associated with PAI1 in the entire cohort (p = 1.86e‐07 and p = 5.45e‐07), and the sub‐cohort with brain amyloidosis (p = 2.44e‐05 and p = 4.77E‐05) (Fig. 1C‐D). When CSF pT217/T217 (%) was included in the model, both the MMSE and CDR‐SB remained associated with PAI1 in the entire cohort (p = 8.85e‐06 and p = 3.02e‐05) and the sub‐cohort with brain amyloidosis (p = 0.0001 and p = 0.0003).ConclusionThese results demonstrate that CSF PAI1 is associated with symptomatic AD and adds value to CSF pT217/T217 in predicting AD symptoms.