The effect of genetic predisposition to Alzheimer's disease and related traits on recruitment bias in study of cognitive ageing

The recruitment of participants for research studies may be subject to bias due to an overrepresentation of those more willing to participate voluntarily. No study has analysed the effect of genetic predisposition to Alzheimers disease (AD) on study participation. The Prospective Imaging Study of Ageing (PISA), aims to characterise the phenotype and natural history of healthy adult Australians at high future risk of AD. Participants approached to take part in PISA were selected from existing cohort studies with available genome-wide genetic data for both successfully and unsuccessfully recruited participants, allowing us to investigate the genetic contribution to voluntary recruitment. From a recruitment pool of 13,432 individuals (age 40-80), 64% of participants were successfully recruited into the study. Polygenic risk scores (PRS) were computed in order to test to what extent the genetic risk for AD, and related risk factors (including educational attainment, household income and IQ), predicted participation in PISA. We examined the associations between PRS and APOE {varepsilon}4 with recruitment using logistic regression models. We found significant associations of age and sex with study participation, where older and female participants were more likely to complete the core module. We did not identify a significant association of genetic risk for AD with study participation. Nonetheless, we identified significant associations with genetic scores for key causal risk factors for AD, such as IQ, household income and years of education. Our findings highlight the importance of considering bias in key risk factors for AD in the recruitment of individuals for cohort studies.