A higher cecal intubation rate (CIR) was associated with increased detection of colorectal cancer precursors and a lower risk of post-colonoscopy colorectal cancer death, according to a retrospective cohort study published in Clinical Gastroenterology and Hepatology.
The findings support a CIR above 95% as a desirable target for screening colonoscopy and may help clarify the patient outcome tied to a common colonoscopy quality metric. Although complete visualization of the colon has long been considered central to high-quality colonoscopy, long-term outcome data have been limited.
The study included 349,782 screening participants in Austria between January 2012 and December 2022. Investigators linked individuals who participated in the Austrian Colonoscopy Quality Assurance Program to the Austrian death registry to identify deaths from post-colonoscopy colorectal cancer (PCCRC). They assessed the association between endoscopists’ CIRs and detection of adenomas, proximal serrated polyps, and PCCRC mortality.
The cecum was reached in 340,836 colonoscopies, or 97.44% of procedures. The leading reported reasons for incomplete colonoscopy were poor bowel preparation, pain, stenosis, complications, and other causes.
Higher CIR was tied to precursor detection
With each 1 percentage point increase in CIR, the probability of detecting an adenoma increased by 1 percentage point (odds ratio [OR], 1.01; P = .043). The association was stronger for proximal serrated polyps, with each 1 percentage point increase in CIR associated with a 3 percentage point increase in detection probability (OR, 1.03; P < .001).
Detection rates also differed across CIR categories. At least one adenoma was detected in 13.53% of colonoscopies performed by endoscopists with a CIR of 90% or lower, 18.66% of those performed by endoscopists with a CIR of 90% to 95%, and 24.06% of those performed by endoscopists with a CIR of 95% to 100%. For proximal serrated polyps, the corresponding detection rates were 1.22%, 1.93%, and 3.62%, respectively.
Mortality risk was lowest above 95%
Overall, 271 PCCRC deaths occurred after a median follow-up of 4.85 years. Each 1 percentage point increase in CIR was associated with a lower hazard of PCCRC death (hazard ratio [HR], 0.93; P < .001).
When investigators grouped endoscopists according to current guideline cutoffs, screening participants whose endoscopists had a CIR of 95% to 100% had a significantly lower risk of PCCRC death compared with those whose endoscopists had lower CIRs (HR, 0.44; P < .001). In a subgroup analysis limited to endoscopists with a CIR of at least 90%, participants whose endoscopists had a CIR of 90% to 95% had higher hazards for PCCRC mortality compared with those whose endoscopists had a CIR above 95% (HR, 1.91; P < .001).
The results tie a familiar quality metric to patient outcomes. In this cohort, patients whose endoscopists had CIRs below 95% had less favorable outcomes, suggesting that a 90% minimum standard may be too low for screening colonoscopy.
Limitations and quality implications
The study had several limitations. Cecal intubation was based on endoscopist reporting in the quality assurance database, though audits are performed. Investigators had access to PCCRC mortality through death registry linkage but not PCCRC incidence, so they could not calculate the World Endoscopy Organization–recommended PCCRC 3-year rate.
The study also lacked data on follow-up colonoscopies and patient-level risk factors such as body mass index, comorbidities, smoking, and diet. However, the association between CIR above 95% and lower PCCRC mortality persisted in sensitivity analyses excluding early PCCRC deaths.
The findings support using CIR in quality audits and benchmarking, with attention to case-mix factors such as age, sex, bowel preparation, and sedation use. The authors recommended a universal aspirational target of greater than 95% for screening colonoscopy.
Trauner disclosed advisory, speaker, research and travel ties to numerous pharmaceutical companies and is a co-inventor on Medical Universities of Graz and Vienna patents for norUDCA; the remaining authors reported no conflicts.