Letter: Bowel Preparation Quality in Patients With Crohn's Disease—Authors' Reply

Abstract

We thank Dr. Schildkraut et al. for their thoughtful commentary and the acknowledgment of the importance of our study in demonstrating that existing bowel preparation quality assessment instruments, typically used in the general population, are also reliable and valid for patients with Crohn's disease (CD) [1]. The implications of these findings are that instrument selection for use in clinical practice should be based on familiarity and local practice. The ease of use and interpretability of these instruments are critical for effective clinical application and need further exploration. Instruments that are simple to administer and reduce ambiguity in score interpretation could significantly enhance clinician confidence, making them more suitable for routine practice and future guideline development. Our results also support the inclusion of patients with CD in studies to evaluate novel bowel preparation formulations, which are historically conducted on healthy subjects [2].

The 50 endoscopy videos evaluated in our study included recordings of 34 colonoscopies with terminal ileal evaluation, 14 colonoscopies, and two flexible sigmoidoscopies, involving a total of 40 patients with CD. We rigorously justified the sample size using formal sample size calculation as outlined by Zou [3]. We adhered to the fundamental principle that a study too small may lack the power to adequately address the research question, while an overly large study can result in unnecessary use of resources and may raise ethical concerns. In this study, assuming an intraclass correlation coefficient of 0.80, scoring 50 videos by three central readers provided over 86% probability of obtaining a one-sided 95% lower bound exceeding 0.65, meeting the “substantial” agreement threshold per Landis and Koch's criteria [4]. We also highlight that this estimate was conservative, as we ultimately analysed the data using a two-way random effects model, which is more efficient and yields stronger reliability estimates [3].

At diagnosis, disease location was ileocolonic in 38.5%, colonic in 15.4% and ileal in 12.8% of patients. Disease behaviour was non-stricturing/non-penetrating in 46.1%, penetrating in 20.5%, and stricturing in 7.7% of patients. We acknowledge that our study population was heterogeneous, reflecting the real-world diversity of patients with CD compared to the general population. This heterogeneity, while presenting interpretive challenges, is an inherent characteristic of CD and underscores the importance of evaluating novel bowel preparation formulations within this group [5]. We concur with the authors that there is a need for further research specifically targeting the CD population to better understand not only the reliability of bowel preparation instruments but also the effectiveness of novel bowel preparation formulations, particularly in subgroups such as those with penetrating, stricturing or perianal fistulising disease. This is especially important for these subgroups given their high disease burden and substantial treatment needs [6].

In summary, our study provided strong evidence for including patients with CD in clinical trials evaluating new bowel preparation formulations and supports the standardised assessment of bowel preparation quality in trials of novel therapies. We look forward to future studies that build on these findings and further explore their applicability in larger and more specific CD subgroups.