Peripheral cytokine profiles contribute to machine learning model of frailty

Frailty is a multi-system syndrome causing increased susceptibility to health insults in older adults. Immune system dysregulation and inflammaging have emerged as mechanisms that may affect multiple organ systems in the frailty syndrome. This present study seeks to define the immune state in community-dwelling adults suffering from frailty. We evaluated two subgroups of individuals enrolled in the Gut-brain Alzheimer’s disease Inflammation and Neurocognitive Study (GAINS). Participants in the GAINS study were scored for frailty using the Clinical Frailty Scale (CFS). A panel of 27 inflammatory markers was analyzed from the serum of each participant using Bio-Plex Pro Human Cytokine Assay (Bio-Rad). Associations of inflammatory markers with frailty as a categorical outcome (CFS ≥ 4) were determined using conventional and machine learning analysis for 165 participants with data from the index study visit. To enhance model accuracy, a second machine learning analysis was performed using a group of 189 participants which included data from follow-up visits. Frailty was present in 31 (18.7%) of the cohort, and was correlated with age, malnutrition, and cognitive assessments. Frail participants had a median CFS of 4, indicative of mild frailty. Statistical analysis adjusting for clinical covariates (age, gender, ADAS-COG 13, and MNA score) revealed higher serum levels of IL-2 in frail participants. In addition, presence of IL-2 was associated with an odds ratio of 3.96 [1.76–8.92] for frailty. Machine learning models revealed inflammatory markers enhanced classification accuracy above clinical variables alone. Our data demonstrate that increased levels of IL-2 are associated with mild frailty. Additionally, peripheral inflammatory markers enhanced classification models of the presence of frailty. Together, these results suggest peripheral inflammation may be an important component of mild frailty. Not applicable.