Predictive Factors for Reoperation After Graham Patch Repair of Perforated Duodenal Ulcer: a NSQIP Analysis
Mehida Alexandre, Raymond A Jean, *Alexander S Chiu, *Kevin Y Pei
Yale University School of Medicine, New Haven, CT
Objective: To identify predictive factors for reoperation after Graham patch repair for perforated duodenal ulcers
Design: Retrospective case-control study
Setting: Pyloric and duodenal channel perforations are the most common site of ulcer perforation and are functionally grouped as duodenal perforations. Graham patch closure is the most common method of surgical repair of perforated ulcers. Emergency surgery for perforated ulcers has a 6-30% risk of mortality and is associated with severe morbidity and repair failure. Clinical predictors of reoperation after surgical management of perforated ulcers have not been elucidated.
Patients: Using data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) between 2012 and 2015, patients with a diagnosis code for perforated duodenal ulcer and underwent Graham patch repair were selected.
Interventions: None
Main Outcome Measures: Predictors of unplanned reoperation
Results: A total of 603 patients met inclusion criteria. There were 59 (9.8%) unplanned reoperations. Stepwise logistic regression for re-operation resulted in four statistically significant covariates remaining in the model: bleeding disorder (OR 2.3, 95%CI [1.0, 4.9]), preoperative sepsis/septic shock (OR 2.1, 95%CI [1.1, 4.1]), preoperative ventilator dependence (OR 5.2, 95%CI [2.2, 12.6]), and morbid obesity (OR 2.2, 95%CI [1.2, 4.1]).
Conclusions: Omental patch repair of duodenal perforation is associated with a 9.8% re-operation rate. Bleeding disorder, sepsis/shock, ventilator dependence, and morbid obesity predicted increased risk of unplanned reoperation.
Table 1:
| Preoperative Risk Factor | Odds Ratio | 95% Confidence Interval | |
| Bleeding Disorder | 2.3 | 1.0 | 4.9 |
| Sepsis/Septic Shock | 2.1 | 1.1 | 4.1 |
| Ventilator Dependence | 5.2 | 2.2 | 12.6 |
| Morbid Obesity (BMI>30) | 2.2 | 1.2 | 4.1 |
Appendix A:
| Preoperative Covariates for Predictive Model | |
| Age | RBC Transfusion within 72 hours before surgery |
| Sex | Renal Failure |
| Ascites | Sepsis or Septic Shock |
| Bleeding Disorder | Steroid use at time of surgery |
| Chronic Obstructive Pulmonary Disease | Tobacco use at time of surgery |
| Congestive Heart Failure | Underweight (BMI>18) |
| Diabetes (Insulin or Non-Insulin dependent) | Ventilator Dependence |
| Dialysis | Weight Loss >10% |
| Disseminated Cancer | Serum Creatinine >2 within 72 hours before operation |
| Dyspnea on Exertion or at Rest | INR > 1.5 within 72 hours before operation |
| Hypertension requiring medication | WBC >16,000 or <4000 within 72 hours before operation |
| Morbid Obesity (BMI>30) | Dysnatremia (>150 or <130) within 72 hours before operation |
| Partially or Totally Dependent Functional Status | Anemia (HCT <30) within 72 hours before operation |
| Preoperative Wound Infection | |
Back to 2017 Posters