QS144 - Predicting a Difficult Colonoscopy

Purpose/Background: The ability to achieve a quality colonoscopy varies greatly, and prior studies have suggested variables such as body mass index (BMI), gender, co-morbidities and surgical history are associated with cecal intubation time and failure rate [1-3]. We present a retrospective chart review of colonoscopies performed at a single center to identify patient characteristics that contribute to increased colonoscopy difficulty and for predicting complexity prior to the procedure.

Methods/Interventions: IRB approval was obtained. We performed a retrospective chart review of patients 18 years and older undergoing lower endoscopy at a single tertiary academic center over a 12-month period to identify characteristics that increase colonoscopic difficulty. Sigmoidoscopy, pouchoscopy and colonoscopy through a colostomy were excluded. Demographic data were obtained from the medical record. Colonoscopy specific data were retrieved from the report generated with Provation® software. Endoscopists were stratified according to years of post-residency experience. Outcome variables included mean cecal intubation (insertion) time and procedure difficulty, listed as “without difficulty”, “somewhat difficult”, or “technically difficult.” For purposes of our analysis, “without difficulty” and “somewhat difficult” were coded together as “not difficult”. Multivariate and univariate analysis was performed. P-values were calculated using Fisher’s exact test for categorical data and Mann-Whitney test for non-categorical data.

Results/Outcomes: A total of 2342 colonoscopies met criteria for review. Three hundred and twenty one patients were excluded for missing information regarding difficulty and insertion time. Of the remaining 2285 procedures, 570 (25%) were characterized as difficult. The procedure was more likely to be deemed difficult by a less experienced endoscopist (33% vs 24%, p=.011). Insertion time was significantly longer in procedures that were deemed difficult (18 vs 9 minutes, p < .0001). Greater than 13 minutes was the upper quartile of insertion time. Among scopes in the upper quartile, multivariate analysis identified the following significant risk factors: female (OR 1.4, p=0.003), hysterectomy (OR 1.8, p=0.002), gastric bypass (OR 2.0, CI 1.2-3.29), senna use (OR 3.7, p= 0.043). Other race was protective (OR 0.5, p=.013), On multivariate analysis of endoscopist classification of the procedure, the following variables significantly increased difficulty: constipation (OR 1.4, p=0.006), females (OR 1.4, p=0.006), hysterectomy (OR 1.6, p=0.12), gastric bypass (OR 1.9, p=0.011), BMI < 27 or >43 (OR 1.4, p=0.0006). Race, other than white or black, predicted an easier colonoscopy (OR 0.52, p=0.015).

Conclusion/Discussion: These factors should be considered when selecting patients for screening/surveillance colonoscopy versus non-invasive methods. Consideration of these criteria may improve scheduling, endoscopy work flow, and identification of patients appropriate for a training environment.