J Minim Access Surg. 2022 Jul 20. doi: 10.4103/jmas.jmas_6_22. Online ahead of print.
CONTEXT: Gastrointestinal polyps are common gastrointestinal diseases that involve localised hyperplastic masses derived from gastrointestinal mucosa.
AIMS: To investigate the risk factors of delayed post-polypectomy bleeding (DPPB) after the treatment of gastrointestinal polyps with snare-assisted endoscopic sub-mucosal dissection (ESD) and to construct a nomogram model to predict the risk of DPPB.
SETTINGS AND DESIGN: A total of 226 patients who underwent snare-assisted ESD for gastrointestinal polyps from May 2018 to November 2020 were divided into DPPB group (n = 10) and non-DPPB group (n = 216).
SUBJECTS AND METHODS: The correlations of clinical data and endoscopic data with DPPB were compared. Univariate analysis was performed to screen the influencing factors of DPPB. Multivariate logistic regression analysis was used to screen the risk factors of DPPB, which was employed to construct a nomogram prediction model.
STATISTICAL ANALYSIS USED: SPSS 16.0 software was utilised for statistical analysis. Numerical data were expressed as percentage (n [%]), and Chi-square test was performed for univariate analysis. The significant factors (P < 0.05) in univariate analysis were included in multivariate logistic regression analysis, and the variables with statistical significance (P < 0.05) were considered as independent risk factors. The factors were used to construct a nomogram model for predicting the risk of DPPB. Bootstrap method was employed to perform repeated sampling 1000 times for internal verification. The consistency index (C-index) was used to evaluate the discrimination of the model, and C-index ≥0.70 represented a good discrimination. Two-tailed P < 0.05 indicated that a difference was statistically significant.
RESULTS: Univariate and multivariate logistic regression analyses revealed that hypertension, polyp location, polyp diameter, polyp morphology and intra-operative bleeding were the independent risk factors for DPPB (P < 0.05). The C-index of the nomogram model for predicting the risk of DPPB was 0.791, indicating a good discrimination. The calibration curve showed that the mean absolute error between predicted and actual DPPB occurrence risks was 0.014, indicating a high accuracy.
CONCLUSIONS: Hypertension, polyp location, polyp diameter, polyp morphology and intra-operative bleeding are the independent risk factors for DPPB, and the nomogram model established based on these factors for prediction has good discrimination and accuracy. Therefore, it is recommended to perform targeted intervention for high-risk groups to reduce the incidence of DPPB.