In the article by Eyre et al published in the November 2006 issue of Arthritis & Rheumatism (pp. 3417–3422), the “A” and “G” were reversed in the assignment of allele calls for the rs3087456 single-nucleotide polymorphism. Therefore, throughout the text, “−168G/A” should have read “−168A/G.” The last 3 rows of the left column in Table 1 should have read “A/A,” “A/G,” and “G/G,” respectively. The last letter of the haplotypes shown in rows 1, 2, 3, 4, and 5 of the left column of Table 2 should have been “A,” “G,” “G,” “A,” and “G,” respectively. These changes do not affect any of the conclusions of the study

We regret the errors.