Skip to Main Content

An official website of the United States government

A Mendelian randomization study of diabetes mellitus and prostate cancer risk

Principal Investigator
Name
Brandon Pierce
Institution
University of Chicago
Email
About this CDAS Project
Study
PLCO (Learn more about this study)
Project ID
2009-0094
Initial CDAS Request Approval
Jan 21, 2010
Title
A Mendelian randomization study of diabetes mellitus and prostate cancer risk
Summary
Epidemiologic research suggests that diabetes mellitus is associated with reduced prostate cancer (PCa) risk; however, it is unknown whether this association is causal. In a previous analysis of a subset of PLCO data, we observed that susceptibility to type 2 diabetes (T2D), as measured by combining data on 18 T2D risk variants, was associated with PCa risk. We now propose to combine this genetic data on T2D susceptibility with data on self-reported diabetes status, to confirm that the effect of T2D susceptibility (on PCa) is at least partially medicated by the diabetes phenotype. Using available SNP data, we will generate a T2D genetic risk score for 1,171 non-Hispanic white, PSA-screened, PCa cases and 1,101 matched controls from the Cancer Genetic Markers of Susceptibility project (CGEMS – a study of individuals drawn from the PLCO Trial). We will compare the direct and indirect (i.e., adjusted for diabetes status) effects of T2D susceptibility on PCa risk and use instrumental variable methods to generate causal estimates of the effect of diabetes status on PCa risk.
Aims

Our primary hypothesis is that the inverse association between T2D genetic susceptibility and PCa represents a causal relationship that is at least partially mediated by diabetes status. We will test this hypothesis by completing the following aims. Primary Aims: Using case-control data drawn from the screening arm of the PLCO Cancer Screening Trial (i.e., the CGEMS genome-wide association study), we will generate genetic risk scores for T2D using data from ~20 T2D risk variants (or more if new T2D variants are identified in the near future) using available CGEMS data. We will link this information with data on diabetes status (and other covariates) from the PLCO Cancer Screening Trial. We will estimate the causal effect of diabetes status on PCa risk using data on PCa case-control status, self-reported history of diabetes, and T2D genetic risk scores (i.e., using instrumental variable analysis). This causal estimate will be compared to the association estimate derived from an ordinary logistic regression model, to determine to what degree the diabetes-PCa association is confounded. Secondary Aims:To assess the potential mediating effects of diabetes status, we will test associations between T2D susceptibility scores and PCa risk, both adjusted and unadjusted for T2D status, generating both direct and indirect effects T2D susceptibility on PCa. To explore the effects of diabetes and T2D genetic susceptibility among clinically relevant subgroups, we will conduct the analyses described in Aim 2 among PCa cases defined by clinical characteristics (stage, grade (i.e., Gleason score), and aggressiveness. We will explore the degree to which diabetes status and T2D susceptibly (and individuals T2D risk variants) influence baseline PSA concentrations. Because all cases are drawn from the screening arm of the PLCO trial, it is important to explore the effects of T2D susceptibility and diabetes status on PSA, a detection factor which has a large impact on the probability of a PCa diagnosis

Collaborators

Habibul Ahsan (University of Chicago)