Analysis of risk factors for duodenal leak after repair of a duodenal perforation

BMC Surg. 2023 May 10;23(1):116. doi: 10.1186/s12893-023-02005-7.

Abstract

Background: Repairing of a duodenal perforation is a well accepted procedure, but clinically, approximately 4% of patients develop duodenal leaks after perforation repair, increasing the risk of death. We retrospectively analyzed clinical data from 168 patients at our hospital to explore risk factors for duodenal leak after perforation repair and developed a nomogram for predicting postoperative duodenal leak.

Methods: This retrospective case-control study totalled 168 patients undergoing repair of a duodenal perforation with omentopexy at the General Surgery Department, Dongnan Hospital of Xiamen University, from January 2012 to January 2022. The patients were divided into the non-leak group and the leak group. Risk factors were evaluated by analyzing the patient's sex, shock, diameter and anatomic position of the ulcer, use of NSAIDS and Glucocorticoid, history of drinking, diabetes, chronic diseases, age, time of onset of symptoms and lab tests.

Result: One hundred fifty-six patients (92.9%) who did not develop leaks after repair of a duodenal perforation were included in the non-leak group, and 12 (7.1%) developed leaks were included in the leak group. In univariate analysis, there were significant differences between the two groups referring to age, shock, NSAIDs, albumin, and perforation size (P < 0.05). The area under the ROC curve for perforation diameter was 0.737, the p-value was 0.006, the optimal cutoff point was 11.5, sensitivity was 58.3%, and specificity was 93.6%, the positive predictive value is 41.1%, and the negative predictive value is 98.0%. In the internal validation of the performance of the nomogram, the C-index and AUC of the model were 0.896(95%CI 0.81-0.98), demonstrating that the nomogram model was well calibrated.

Conclusion: The study discussed the risk factors for postoperative duodenal leak in patients undergoing repair of a duodenal perforation, and a nomogram was constructed to predict the leak. Future prospective studies with large sample sizes and multiple centres are needed to further elucidate the risk of duodenal leak after repair of a duodenal perforation.

Keywords: Duodenal leak; Logistic regression; ROC curve; Repair of duodenal perforation; Risk factors.

MeSH terms

  • Case-Control Studies
  • Duodenal Ulcer* / surgery
  • Duodenum / surgery
  • Humans
  • Peptic Ulcer Perforation* / etiology
  • Peptic Ulcer Perforation* / surgery
  • Prospective Studies
  • Retrospective Studies
  • Risk Factors