We read with interest the reply from Simms, Doecke and Radford-Smith. We welcome the opportunity to clarify questions about our previous studies,1^–3 because the details serve to emphasise the validity of our conclusions that differ from those of Simms et al (Gut 2008;57:903–10). First, our data from independent sample sets support that NOD2 SNP13 is associated with a decrease in HD5 mRNA levels. In the first study,1 the genotypes were mixed, but the majority (5/8, fig 1A) within the ileal Crohn’s disease (CD) group were SNP13 frameshift mutations; in the second study,2 larger sample sizes from US patients allowed the conclusion that only SNP13 samples were further decreased (as compared with wild-type NOD2) with ileal CD. Perhaps a sampling error resulting from their small sample size will explain the data …