Identifying Predictors of Repeat Colon Cancer Screening Following an Incomplete Colonoscopy

Abstract

Introduction: For a complete colonoscopy, a number of factors such as adequate prep and patient tolerance are involved in order to identify precursor lesions and potentially remove them. We reviewed patients who underwent incomplete colonoscopies with the purpose of identifying those who are at risk of incomplete colonoscopy and to predict those who undergo subsequent exams. Methods: This is an observational, retrospective, single-institution study of patients who underwent an incomplete colonoscopy, from 2005-2013, as determined by Current Procedural Terminology (CPT) coding. The following characteristics were reviewed: patient demographics, past medical and family history, colonoscopy findings, follow up exams, pathology, cancer staging, and mortality. Logistic regression analysis was used to identify predictors of repeat exams. Results: In our study, 902 incomplete colonoscopy exams were reviewed. Six patients were not of African American (AA) or White race/ethnicity and were excluded. Of the remaining 896 patients, 588 (66%; median age =59 yo, 30% male, 39% AA) underwent a subsequent complete exam of the colon, most within 1 year (85%). Patient characteristics include a history of abdominal surgery (62%), narcotic use (11%), and prior incomplete colonoscopy (7.7%). Twelve percent had a family history of colorectal cancer (CRC) and 6% had a personal history of CRC. The most common reasons for incomplete colonoscopy were inadequate prep (48%) and inability to pass the scope (53%). Ten percent of initial exams reached the cecum while 46% were limited to the sigmoid. Eighty-three patients (14%) had lesions identified on incomplete colonoscopy, of which 26 had advanced histology. Barium enema comprised the majority of subsequent exams (45%) followed by repeat colonoscopy (38%) and CT colonography (7%). On stepwise regression modeling, incomplete visualization was the highest predictor of repeat exam with sigmoid visualization being the most significant (OR 4.12 95% CI 2.56,6.67). In addition, lesions visualized (OR 1.84 95% CI 1.16,2.91), patients with IBD (OR 1.69 CI 0.82,3.48), and the inability to pass the scope (OR 1.40 95% CI1.02,1.94) were predictive of patients undergoing repeat exam. Conclusion: There are a number of potential characteristics that identify patients who are more likely to repeat colon cancer screening. With further prospective studies, we can better identify those patients who do not return for colon cancer screening and identify measures to improve screening outcomes.2756 Figure 1. Full Logistic Regression Model Examining Independent Predictors of Repeat Colorectal Cancer Screening After Incomplete Colonoscopy