AUTHORS' REPLY Comments on 'Mendelian randomization: Using genes as instruments for making causal inference in epidemiology': Authors' response

摘要

We thank Ziegler et al. for their positive comments on our paper . They make two criticisms of the paper. First, they question whether our ordinary least-squares and instrumental variables estimate of the association between circulating C-reactive protein (CRP) levels and homeostasis model insulin resistance (HOMA-R) might be biased as a result of the detection limit of (CRP). We think that this is unlikely since just 1.6 per cent of our cohort had values of CRP at or below the detection limit of 0.168mg/L and when these participants were removed from the analysesour findings did not differ from those presented in the Statistics in Medicine paper or in the original publication from which this example was taken .