Tanzer raises 2 issues for discussion related to potential confounders in our study examining the association of CRP level with the risk of developing RA.1 First, as he indicates, smoking is a known risk factor for incident RA. In our cohort, 38% of RA cases were never smokers and 52% were past smokers at the time of diagnosis vs 52% and 37%, respectively, among women who did not develop RA. In our analyses, we adjusted for smoking status; thus, our findings were unlikely to reflect confounding by smoking. Second, regarding statin use, the blood samples in our study were collected between 1993 and 1995, a period when the prevalence of statin use was relatively low. We did not have data on statin use at baseline to adjust for this; however, there was unlikely to be major confounding because there was unlikely to be many women using statins.
Shadick NA, Karlson EW, Lee I. C-Reactive Protein Is Still a Potential Aid in Rheumatoid Arthritis Predictors—Reply. Arch Intern Med. 2007;167(14):1552. doi:10.1001/archinte.167.14.1552-b