Reply to Are we certain that chlorhexidine gluconate bathing is not beneficial in deducing central line associated blood stream infections among children with cancer or undergoing hematopoietic stem cell transplantation?

Abstract

We appreciate the thoughtful comments provided by Hord and Dandoy regarding our article titled “Chlorhexidine Gluconate Bathing in Children With Cancer or Those Undergoing Hematopoietic Stem Cell Transplantation: A DoubleBlinded Randomized Controlled Trial From the Children’s Oncology Group.” The issues they raised and our responses follow below. There was a concern that we excluded patients with implanted ports. We targeted patients with external central lines because they are a population that generally has a higher rate of central line– associated bloodstream infections (CLABSIs), whether from their underlying diagnosis or from the line itself. Although we agree that our findings may not be applicable to patients at lower risk for CLABSIs, such patients may also have less of an opportunity to benefit from the intervention, especially when they are outpatients and their port is not frequently accessed. Regarding adherence to the chlorhexidine gluconate (CHG) and control bathing regimen, we agree that discontinuation of the study protocol was disappointingly high. As for satisfaction with the cloths, only 73% of the subjects were very or extremely satisfied with the cloths; this meant that 27% were not. Perhaps more importantly, 32 of the 43 subjects (74%) who did not complete the survey discontinued protocol bathing early. Thus, when we consider the number of “less satisfied” survey respondents and the number of subjects who did not complete the survey but were likely to be dissatisfied, many subjects found the cloths challenging. This suggests that compliance may be a concern in this population. Hord and Dandoy pointed to a potential imbalance of acute myeloid leukemia between randomized groups. It is for this reason that we performed a post hoc analysis adjusting for underlying disease as reported in our publication. The estimated effect was similar after adjustments for the underlying diagnosis but was not statistically significant (see paragraph 2 of the Results section). Hord and Dandoy also noted that poor accrual was a limit of our study. We recognized this limitation and performed a post hoc conditional power analysis (see the supporting information for methodological details) to assess the chance of rejecting the null hypothesis assuming the following best scenario conditions for the unenrolled patients: full followup, 100% adherence, and the designed CLABSI rates rather than the observed rates (in other words, an optimistic effect of the intervention). This analysis showed a 0.2% chance that CHG would have demonstrated a statistically significant reduction in the CLABSI rate in comparison with the control group if the study had been fully enrolled. Thus, if had been fully enrolled, it is unlikely that the trial would have found reduced CLABSI rates in the CHG arm. Finally, we agree with Hord and Dandoy that CHG bathing as an intervention to reduce CLABSIs should not be “hastily” discarded, but nor should the decision to use it be made without careful consideration; instead these decisions should be deliberated, with available evidence for effectiveness and other relevant information, such as expense and opportunity costs, taken into account. The results of our study do not support the effectiveness of CHG bathing in pediatric patients with cancer or those undergoing allogeneic hematopoietic cell transplantation. In fact, little data currently support CHG bathing as a means of preventing CLABSIs in patients (of any age) with cancer. We stand by our conclusions and hope that our data provide the impetus for others to carefully consider the utility of CHG bathing in children with cancer.