EN
TÜRKÇE
Poster - 33
Factors associated with intestinal anastomosis failure in children: single-center retrospective cohort analysis
Laura Camila González-Villarreal 1, María Teresa Vallejo-Ortega 2, Juan Javier Valero-Halaby 1
1 HOMI Foundation, National University of Colombia
2 National University of Colombia
Purpose: The failure of an intestinal anastomosis is a significant event associated with high morbidity. However, few pediatric studies have focused on evaluating the potential risk factors for this outcome. This study aims to assess various preoperative, intraoperative, and postoperative variables possibly related to an increased risk of intestinal anastomotic leakage in children.
Methods: An observational, analytical, retrospective cohort study was conducted in patients who underwent intestinal anastomoses between January 2017 and July 2023 at a third-level pediatric reference center in Colombia. Demographic data were described, and risk factors for anastomotic failure were analyzed using backwards stepwise logistic regression on potential covariates with an inclusion threshold of p<0.1 and a significance definition of p<0.05.
Results: A total of 449 anastomoses in 404 patients were identified. The mean age was 4.95 years (SD 5.33), 55% were male, and the main indication for anastomosis was acute inflammatory pathology (n=211, 47%). The population was characterized by a high incidence of acute, high-complexity pathology (41.87% ASA > III, 44.77% postoperative ICU monitoring), and the incidence of anastomotic leakage was 13.14% (n=59), with a median time to diagnosis of 5.47 days (IQR 3). In the multivariate analysis, the only variable associated with a higher risk of anastomotic leak was the presence of laparostomy (OR 7.86, 95% CI 2.099 to 29.427, p = 0.002).
Conclusion: Since the only factor increasing the risk of anastomotic leakage in this cohort was the presence of an open abdomen, we conclude that the surgeon’s intraoperative judgment holds prognostic value. Failure to identify other statistically significant associations with other variables in the model may be attributed to the sample size. Therefore, further prospective multicenter studies are essential to create predictive tools that may aid in early diagnosis.