Use of Multivariable Mendelian Randomization to Address Biases Due to Competing Risk Before Recruitment.

The Q-statistics were highly significant for lipid modifiers and BMI on MEGASTROKE stroke and AF ( Table 1 ), indicating that these estimates were likely still biased by pleiotropy probably from selection bias given the same instruments gave estimates apparently unbiased by genetic pleiotropy for stroke in the UK Biobank.

also in 132,142 other papers