Reply to Walker et al.

TO THE EDITOR—We thank Walker et al for their interest in our work [1]. Our study found no evidence of difference in clinical severity of Clostridium difficile infection (CDI) between patients infected with ribotype 027/078 strains compared with others, when adjusted for other co-variates [2]. Adjustment for factors on the hypothesized causal pathway between exposure (ribotype) and outcome (severity), such as leukocyte count, has been addressed elsewhere [3]. Removing the leukocyte count as a covariate or including it in the definition of severe disease did not alter our findings [3]. We appreciate the concern regarding the fact that the resultant P values were close to the standard α level of .05. We used the customary cut-point for α found in the scientific literature but appreciate its limitations. It is possible, with more patients , that an association may be found. However, a secondary analysis of our entire dataset (Supplementary Table 1 [2]) was conducted with a total of 413 CDI cases. The association between ribotype 027/078 and severe disease in this analysis yielded an odds ratio of 1.34 (95% confidence interval, .67–2.58) when only adjusted for age, sex, and CDI surveillance definition—factors that existed prior to CDI diagnosis. Contrary to Walker et al's assertion, we did not intend to claim " evidence of no difference " in disease severity for patients infected with ribotype 027/078 strains. We sought to test the hypothesis that infection with ribotype 027/078