Aims
Gastrointestinal transmural defects, especially postoperative leaks and fistulas, are associated with substantial morbidity and increasing incidence due to more complex surgery. Endoscopic therapy is the preferred minimally invasive alternative to rescue surgery, but clinical success varies widely and depends on anatomical, physiological, and technical factors. This study aimed to identify predictors of technical and clinical success in endoscopic management of gastrointestinal wall defects and to develop a prognostic nomogram for personalized decision-making.
Methods
A retrospective cohort study was conducted at a tertiary referral center including all patients who underwent endoscopic treatment of postsurgical or iatrogenic gastrointestinal transmural defects from 2014 to 2024. Demographic, anatomical, and clinical variables were collected, along with defect size, location, and distance from natural orifices. Endoscopic therapies were classified as mechanical closure (clips or suturing) or coverage/drainage approaches (stents, vacuum therapy).Technical success was defined as complete closure or successful placement of the selected technique, and clinical success as definitive resolution without further intervention or surgery. Temporal trends were evaluated comparing early and recent periods, corresponding to wider adoption of drainage-based strategies. Univariate logistic regressions identified candidate predictors for inclusion in a multivariate logistic regression model, which was assessed for calibration and goodness of fit and used to generate a prognostic nomogram.
Results
A total of 95 procedures were performed in 83 patients, mostly postoperative defects (81.1%) and predominantly located in the upper gastrointestinal tract (94.7%). Technical success was 93.7%, while clinical success reached 52.6%. Outcomes improved over time, increasing from 45.1% in 2014–2021 to 61.4% in 2022–2024 with greater use of drainage and negative-pressure techniques (p=0.149). In the multivariate analysis, lower gastrointestinal location (OR 81.2; p = 0.012), greater distance from natural orifices (OR 1.13 per cm; p = 0.011), and procedures performed from 2022 onwards (OR 7.16; p = 0.0047) independently predicted higher clinical success. Mechanical closure techniques (OR 0.05; p = 0.001), larger defect size (OR 0.95 per mm; p = 0.024), and ASA III status (OR 0.26; p = 0.045) were associated with lower success. The final model demonstrated good calibration and allowed development of a nomogram for individualized prognostic assessment.
| Variable | Effect direction | Odds Ratio (95% CI) | p-value |
|---|---|---|---|
| Low anatomical location | ↑ Higher success | 81.2 (3.6–1834) | 0.012 |
| Distance to natural orifices (per cm) | ↑ Higher success | 1.13 (1.03–1.25) | 0.011 |
| Procedures performed ≥2022 | ↑ Higher success | 7.16 (1.89–27.2) | 0.0047 |
| Mechanical techniques vs coverage/drainage | ↓ Lower success | 0.05 (0.009–0.27) | 0.001 |
| Defect size (per mm) | ↓ Lower success | 0.95 (0.91–0.99) | 0.024 |
| ASA III | ↓ Lower success | 0.26 (0.07–0.97) | 0.045 |
Conclusions
Endoscopic treatment of gastrointestinal transmural defects achieves high technical success but only moderate clinical success, although outcomes have improved significantly with the adoption of drainage-based and negative-pressure approaches. Clinical success is influenced by anatomical complexity, patient physiological status, and appropriate technique selection. The resulting predictive nomogram may support individualized risk stratification and guide optimal endoscopic management of gastrointestinal leaks and fistulas.