The Journal of Practical Medicine ›› 2025, Vol. 41 ›› Issue (19): 3016-3025.doi: 10.3969/j.issn.1006-5725.2025.19.009

• Clinical Research • Previous Articles    

A predictive model for bowel preparation quality and mild adverse events in colonoscopy

Xiaoyi CHEN1,Chao WANG2,Yan. NIE1()   

  1. *.Digestive Endoscopy Center,the First Affiliated Hospital of PLA Air Force Medical University,Xi 'an 710032,Shanxi,China
  • Received:2025-07-18 Online:2025-10-10 Published:2025-10-10
  • Contact: Yan. NIE E-mail:18702550672@163.com

Abstract:

Objective To construct a predictive model to explore the factors influencing bowel preparation for colonoscopy and the risks of mild adverse events during colonoscopy. Methods A total of 573 patients undergoing colonoscopy at the digestive endoscopy center of the First Affiliated Hospital of the Air Force Military Medical University from July 2021 to June 2023 were enrolled in this prospective study. The patients were divided into an adequate group (n = 112) and an inadequate group (n = 461) based on bowel preparation assessed with the Boston Intestinal Readiness Score Scale (BBPS). Again, they were divided into an occurrence group (n = 106) and a non-occurrence group (n = 467) based on minor adverse events within 30 days after colonoscopy. Multivariable logistic regression was utilized to identify independent predictors of inadequate bowel preparation and minor adverse events after colonoscopy; model discrimination was quantified with receiver-operating characteristic (ROC) curve analysis, and the derived coefficients were used to construct a clinically applicable risk-scoring system. Results The adequate group achieved a significantly higher BBPS total score than the inadequate group (P < 0.05). The 30-day cumulative incidence of minor adverse events after colonoscopy was 18.5%. Inadequate bowel preparation was associated with a markedly higher event rate (36.61%) than adequate preparation (14.10%; P < 0.05). Multivariable logistic regression identified the following independent predictors of inadequate bowel preparation: age ≥ 60 years, body-mass index ≥ 28 kg/m2, history of constipation, diabetes mellitus, concomitant calcium-channel-blocker use, and Bristol Stool Form Scale types 1 ~ 2 (all P < 0.05). ROC analysis confirmed that age, constipation history, diabetes, calcium-channel-blocker use, Bristol stool type, and their combined model were all significant predictors (P < 0.05), whereas BMI alone was not (P > 0.05). Independent risk factors for post-colonoscopy minor adverse events were age ≥ 60 years, presence of ≥ 1 comorbidity, performance of biopsy or polypectomy, and warfarin use (all P < 0.05). ROC analysis demonstrated significant predictive value for age, number of comorbidities, biopsy/polypectomy, warfarin use, and their combined model (P < 0.05). Odds ratios derived from the multivariable models were converted into weighted scores; the resulting composite risk scale ranged from 0 to 100. In the validation cohort, this combined score predicted both inadequate bowel preparation and subsequent minor adverse events (P < 0.05), with an area under the curve of 0.880 (95% CI not shown), sensitivity of 0.829, and specificity of 0.707. Conclusion Inadequate bowel preparation and post-colonoscopy minor adverse events are inter-related. Independent predictors of inadequate preparation are age ≥ 60 years, BMI ≥ 28 kg/m2, constipation, diabetes mellitus, concurrent calcium-channel-blocker use, and Bristol Stool Form Scale types 1-2. Independent predictors of minor adverse events are age ≥ 60 years, ≥ 1 comorbidity, biopsy or polypectomy, and warfarin use. Targeted preventive measures against these factors should be implemented to improve clinical outcomes.

Key words: colonoscopy, bowel preparation, minor adverse events, influencing factors, forecast

CLC Number: