Letters Section Editor: Robert M. Golub, MD, Senior Editor.
In Reply: Numerous studies have demonstrated an association between CRP levels and CVD events, including the Cardiovascular Health Study (CHS).1 Our goal was to confirm the association between CRP polymorphisms and CRP levels and to assess whether CRP polymorphisms are associated with increased risk of incident CVD.
As noted by Dr Lawlor and colleagues, mendelian randomization methods have the potential to tease out causal effects and differentiate these effects from associations due to confounding or “reverse causation.” However, the validity and utility of the mendelian randomization approach depends on a number of factors, including correct specification of the causal model and a strong association between genotype and intermediate phenotype (in this case, plasma CRP level).2 The causal model for CRP is likely to be complex. C-reactive protein may be both a marker and a mediator of CVD.3 Moreover, despite the accumulating evidence for a robust association between CRP genotypes and plasma CRP levels in large data sets, the proportion of variation in CRP levels explained by CRP polymorphisms is small. Although we agree that mendelian randomization methods may be informative, they may be most useful when performed in the context of extremely large data sets or meta-analyses, rather than in studies such as CHS.
Lange LA, Lange EM, Tracy RP, Reine AP. Polymorphisms in the. JAMA. 2007;297(12):1317-1318. doi:10.1001/jama.297.12.1317-b