Genetic effect modification of cis-acting C-reactive protein variants in cardiometabolic disease status

Abstract

Mendelian randomization (MR) studies carried out among patients with a particular health condition should establish the genetic instrument influences the exposure in that subgroup, however this is normally investigated in the general population. Here, we investigated whether the genetic associations of four cis-acting C-reactive protein (CRP) variants differed between participants with and without three cardiometabolic conditions: obesity, type 2 diabetes, and cardiovascular disease. Associations of cis-genetic variants with CRP differed between obese and non-obese individuals. A multivariable analysis suggested strong independent associations of the gene-by-body mass index (BMI) interaction on CRP (P<1.18×10−8 for the CRP variants). Applying MR, we observed strong causal effect of BMI on CRP (P=2.14×10−65). In summary, our study indicates that genetic associations with CRP differ across disease sub-groups, with evidence to suggest that BMI is an effect modifier. MR studies of disease progression should report on the genetic instrument-exposure association in the disease subgroup under investigation.

Strategies to investigate and mitigate collider bias in genetic and Mendelian randomisation studies of disease progression

“…Additionally, analyses of disease progression require longitudinal data; therefore, they can be vulnerable to bias due to missing data and loss-to-follow-up. Equally, bias can arise in MR studies when the genetic variants differ in their association to the exposure in cases of the disease versus the general population [ 56 ]. An additional consideration when performing an MR analysis with disease progression as the outcome would be to verify that the SNP-exposure associations are the same in cases as in a healthy population as there may be effect modification by factors relating to having the disease.…”

Strategies to investigate and mitigate collider bias in genetic and Mendelian randomization studies of disease progression

“…Additionally, analyses of disease progression require longitudinal data, therefore they can be vulnerable to bias due to missing data and loss-to-follow-up. Equally, bias can arise in MR studies when the genetic variants differ in their association to the exposure in cases of the disease versus the general population (45). An additional consideration when performing a MR analysis with disease progression as the outcome would be to verify that the SNP-exposure associations are the same in cases as in a healthy population as there may be effect modification by factors relating to having the disease.…”

Proteome-wide Mendelian randomization in global biobank meta-analysis reveals multi-ancestry drug targets for common diseases

“…in males and females; in diseased patients and healthy controls). However, naïve use of genetic effects generated in a general population as a proxy in disease subgroups may yield biased estimates of causal relationships, which we have illustrated recently for C-reactive protein 67 . GBMI provides ancestry-specific and sex-specific GWAS data, and also clearly defines disease cases and controls.…”