Neuropsychiatric symptom clusters and plasma proteomic signatures across cognitive ageing and dementia in Kinshasa, Democratic Republic of the Congo: a cross-sectional analysis of the EVCD-RDC cohort

Background: Neuropsychiatric symptoms (NPS) are common across cognitive ageing and dementia, but their molecular correlates are poorly understood in sub Saharan African populations. We examined associations between individual NPS, empirically derived NPS dimensions, and plasma proteomic profiles in older adults in Kinshasa, Democratic Republic of the Congo. Methods: In this cross sectional analysis of the EVCD RDC cohort, NPS were assessed using the Neuropsychiatric Inventory (NPI) and Geriatric Depression Scale (GDS). Plasma proteins were quantified using the NULISAseq CNS Disease Panel 120. Of 1044 participants in the full cohort, 779 had available proteomic data. Multivariable regression models adjusted for age, sex, education, and estimated glomerular filtration rate, with false discovery rate (FDR) correction. Exploratory factor analysis identified latent NPS dimensions. Findings: The full cohort included 1044 participants, of whom 779 comprised the proteomics cohort. Higher NEFL was associated with greater NPI symptom burden (Beta=0.349, 95% CI 0.201 to 0.497; FDR<0.001; N=674). Higher IL6 (Beta=0.805, 0.362 to 1.248) and CCL13 (Beta=0.434, 0.191 to 0.678) were associated with higher GDS scores (both FDR=0.032; N=693). Higher APOE protein was associated with nighttime behaviour (OR 2.11, 95% CI 1.42 to 3.11; FDR=0.025). Three NPS dimensions were identified. NEFL was the only factor level association surviving FDR correction (behavioural/motor: Beta=0.222, 95% CI 0.127 to 0.318; FDR<0.001; N=644). Interpretation: NPS showed heterogeneous plasma protein correlates, with the most robust findings implicating neuroaxonal injury and inflammatory signalling. Longitudinal replication in African cohorts is needed.