EP413 - Implementation of the "Colorectal Leakage App" for Early Detection of Anastomotic Leakage After Colorectal Surgery.

Image or Table (Do Not add text here):

Purpose/Background: Anastomotic leakage (AL) remains a major complication in colorectal surgery, leading to increased morbidity, prolonged hospitalization, and higher mortality rates. Early detection is crucial for improving patient outcomes. The Dutch Leakage Score (DLS) is a validated tool for AL diagnosis. We evaluated the feasibility and accuracy of the "Colorectal Leakage App," (CLA) a mobile application integrating the DLS for standardized postoperative monitoring.

Methods/Interventions: A prospective, single-center observational study was conducted at NROC, Kazakhstan. Patients undergoing colorectal surgery with anastomosis were included. The app records daily postoperative parameters, calculates DLS scores, and recommends further investigations when AL is suspected. A DLS score of 4 or higher is considered positive and requires re-evaluation within 12 hours, while a score of 8 or higher prompts a recommendation for CT with rectal contrast. Primary outcomes are feasibility of the app and AL incidence.

Results/Outcomes: A total of 200 patients were enrolled in the study, and the incidence of AL was 4.5% (9 out of 200 cases). The association between CLA test results and AL is presented as follows: among patients with AL, 5 tested positive and 4 tested negative for CLA, while among patients without AL, 22 tested positive and 169 tested negative. Based on this distribution, the preliminary analysis of CLA test performance, considering four or more leakage points as a positive result, showed a sensitivity of 55.56% (95% CI: 21.20%–86.30%), specificity of 88.48% (95% CI: 83.08%–92.64%), positive predictive value of 18.52% (95% CI: 10.10%–31.49%), negative predictive value of 97.69% (95% CI: 95.31%–98.87%), and overall accuracy of 87.00% (95% CI: 81.53%–91.33%). Fisher’s Exact Test, p = 0.0027. These results indicate that the CLA test demonstrates high specificity and an excellent negative predictive value, making it reliable for ruling out AL in negative cases.

Conclusion/Discussion: In this single-center study, the AL rate was 4.5% (9 out of 200 patients). CLA may facilitate early detection of AL. It is reliable for ruling out AL. We plan to conduct multicenter study to further evaluate the app's effectiveness across different healthcare settings in Kazakhstan, aiming to standardize postoperative care.