Linear regression prediction model for an outcome of Alzheimer's Disease in Down Syndrome including BPSD‐DSII sections and Alzheimer's Disease fluid biomarkers

TL;DR

This study creates a model to predict Alzheimer's Disease in people with Down Syndrome using behavior scores and blood biomarkers.

Contribution

A novel linear regression model combining behavioral and fluid biomarkers to predict Alzheimer's in Down Syndrome.

Findings

Behavioral changes like apathy and depression predict Alzheimer's in Down Syndrome.

Blood biomarkers pTau181 and GFAP are strong predictors of Alzheimer's progression.

The model uses baseline data to forecast future Alzheimer's severity in Down Syndrome individuals.

Abstract

By age 40, virtually all individuals with Down Syndrome (DS) will have developed the pathological symptoms of Alzheimer's Disease (AD) (McCarron, McCallion et al. 2017), making it the genetic leading cause of AD worldwide. Our lab developed a specialized behavior‐based scale to aid in the diagnosis of AD in DS, the BPSD‐DSII (Dekker, Ulgiati et al. 2021). Individuals with DS with or without (questionable) AD (total = 157) were tested with the BPSD‐DSII, and subsequently, we also performed immune‐based assays (Simoa) to further characterize blood biomarkers in this cohort. Three years after the original BPSD‐DSII was administered in this cohort, the individuals were, again, categorized in percentiles based on the BPSD‐DSII behavior frequency changes. The outcome of the different sections of the BPSD‐DSII at baseline and the levels of several blood‐based biomarkers at baseline were put in a statistical prediction model based on linear regression.

Linear regression prediction model coded in R with R studio

The outcome of several sections of the BPSD‐DSII questionnaire, as well as pTau181 levels and GFAP levels, were predictive of percentiles indicating AD in DS.

Specific behavioral changes at baseline such as in apathetic behavior, depressive behavior, and eating and drinking behavior, as well as pTau181 and GFAP levels were predictive for the development of AD in DS.