Self‐reported COVID‐19 and biomarkers of Alzheimer’s disease: interim analyses from an Alzheimer’s disease risk‐enriched cohort

Abstract

AbstractBackgroundCOVID‐19‐related SARS‐CoV‐2 infection has been linked with brain abnormalities and cerebrospinal fluid (CSF) biomarkers of Alzheimer’s disease (AD), neuroinflammation, and neuronal injury. Whether COVID‐19 increases the brain’s vulnerability to the development of AD is unclear. In participants at an increased risk of AD, we investigated the association of COVID‐19 with pre‐ and post‐pandemic AD biomarkers.MethodWe included 132 cognitively unimpaired adults from the ALFA+ cohort (52‐71 years, 61.4% women) who provided retrospective self‐reported data on SARS‐CoV‐2 infection (with COVID‐19 vs. without COVID‐19). As of February 2022, 125 (94.7%) participants had available CSF biomarkers (measured with exploratory Roche NeuroToolKit or Elecsys® immunoassays, Roche Diagnostics International Ltd, Rotkreuz, Switzerland) and 76 (57.6%) had [18F]flutemetamol‐PET data acquired before (pre‐pandemic) and after (post‐pandemic) the beginning of the COVID‐19 pandemic (average time lapse between the assessments: 3.5±0.6 years). [18F]flutemetamol PET SUVR maps were obtained using the whole cerebellum as the reference region. First, we performed ANCOVA models to investigate whether COVID‐19 was associated with post‐pandemic CSF biomarkers of core AD pathology (Aβ42/40 and p‐tau), neuronal injury (NfL), and neuroinflammation (sTREM2, GFAP, and IL‐6), while controlling for pre‐pandemic biomarker levels. Covariates were age, sex, years of education, APOE‐ε4 status, and time lapse between the biomarker assessments. Second, we performed a voxel‐wise analysis in SPM‐12 using a matched‐group comparison design (with COVID‐19 vs. without COVID‐19 matched by pre‐pandemic Centiloid values, age, and APOE‐ε4 status) with the post‐pandemic amyloid‐PET scans as the outcome of interest.ResultFifteen (12%) participants included in ANCOVA reported COVID‐19. Eleven (14.5%) with amyloid‐PET scans reported COVID‐19 and were matched with 11 participants without COVID‐19 for voxel‐wise analysis. In ANCOVA models, no significant association between COVID‐19 and post‐pandemic CSF biomarkers was found (Table 1). Results from the t‐test voxel‐wise analysis also did not show any differences between participants with vs. without COVID‐19 on [18F]flutemetamol‐PET uptake after correction with false discovery rate.ConclusionOur interim analyses did not show any association between self‐reported COVID‐19 and post‐pandemic biomarkers of AD pathology, neuronal injury, and neuroinflammation. The results are pending confirmation of COVID‐19 status using serology data (n = 181) and complete sample size.