ASPO Abstracts
Type-2 diabetes mellitus and risk of colorectal polyps: a colonoscopy-based study using natural language processing
Category: Lifestyles Behavior, Energy Balance & Chemoprevention
Conference Year: 2022
Abstract Body:
Purpose: Although type-2 diabetes (T2D) has been associated with colorectal cancer in previous studies, the association of T2D with colorectal polyps is unknown. Methods: Using pathology reports from the University of Utah (UU) Enterprise Data Warehouse (EDW), we developed a rule-based natural language processing (NLP) pipeline to extract colorectal polyp diagnoses and features (site, shape, number, size) on 15,679 patients who underwent a colonoscopy at the UU Gastroenterology clinic from 2013-2016. The NLP pipeline was validated by manual abstraction of 350 pathology reports, and demonstrated excellent performance (accuracy 91%). Patient characteristics, including age, sex, race, diabetes status, smoking, BMI, and medication use, were abstracted from the EDW. Odds ratios (OR) and 95% confidence limits (95% CI) adjusted for abstracted variables were calculated using multivariable polytomous logistic regression. Results: Participants were on average 56 years old, 85% White, 50% male, with a mean BMI of 29 kg/m2. About 27% of the participants reported history of T2D; 71% of whom used anti-diabetes medication. Participants were classified as having adenomas (30%), serrated polyps (16%), synchronous adenomas and serrated polyps (19%) or as polyp-free controls (35%). T2D was associated with a statistically significant lower risk of colorectal polyps [0.83(0.73,0.92)]. When evaluated by polyp subtype, T2D was marginally associated with reduced adenoma risk [0.90(0.80,1.02)], and inversely associated with risk of serrated polyps [0.80(0.67,0.93)]. The associations did not vary by lesion severity within polyp subtypes. There was a statistically significant deceased risk for polyps among anti-diabetes medication users [0.84(0.69,0.99)].Conclusions: Overall, T2D was associated with a statistically significant reduced risk of colorectal polyps; this reduced risk was consistent for both adenomas and serrated polyps. As T2D has previously been shown to increase colorectal cancer risk, this differential association with colorectal polyps may possibly be due to a variable effect of anti-diabetes medication use. Further studies are needed to better understand the mechanisms through which diabetes and its treatment may be differentially associated with colorectal polyps.
Keywords: colorectal polyps, diabetes, adenomas, serrated polyps, anti-diabetes medications