Omental patch repair of large perforated peptic ulcers ≥25 mm is associated with higher leak rate

1Department of General Surgery, Khoo Teck Puat Hospital, Singapore

2Department of General Surgery, Tan Tock Seng Hospital, Singapore

2Department of General Surgery, Tan Tock Seng Hospital, Singapore

2Department of General Surgery, Tan Tock Seng Hospital, Singapore

Omental patch repair is the present gold-standard technique for patients with perforated peptic ulcers (PPUs). Data are lacking regarding the safe ulcer size for omental patch repair leak (OPL). We analyze our experience in managing PPU to identify an ulcer size cut-off for predicting OPL.

Patients who had undergone omental patch repair for PPU between Jan 2004 and Apr 2016 were included. Demographic data, the American Society of Anesthesiologists score, ulcer size, operative approach, post-operative complications, and length of stay were recorded. OPL, intra-abdominal collection, repeat surgery, and 30-day mortality were recorded. The relationship between ulcer size, pre-operative characteristics, and OPL were investigated with univariate and multivariate logistic regression. Receiver operating characteristic curve analysis derived the ulcer size cut-off to predict OPL. In addition, we analyzed if ulcer size predicted mortality or malignancy.

Six hundred and ninety patients with a mean age of 55.1 years (range 16-94) were managed for PPU during the study period. Free air on X-ray was evident in 417 (60.4%) patients. Mean ulcer size was 7.8 mm (range 1-50). OPL occurred in 15 patients (2.2%) and 30-day mortality was 7.4% (n=51). Multivariate analysis found ulcer size increase of 10 mm (OR 3.30, 95% CI 1.81-6.02, PDemographic and clinical profile of patients with perforated peptic ulcer treated with omental patch repairn=690 (%)Mean age, years55.1 (16-94)Risk factors Smoking221 (32) Previous peptic ulcer52 (7.5) NSAID use33 (4.8) Steroid use22 (3.2)Co-morbidities Diabetes mellitus81 (11.7) Ischemic heart disease34 (4.9) Chronic kidney disease23 (3.3) Chronic obstructive pulmonary disease16 (2.3)Investigations Hemoglobin, g/dL13.5 (4.8-20.4) White blood cell count, 109/L12.9 (1.2-68.9) Albumin, g/L34.8 (10-59) Creatinine, µmol/L110 (24-522) Free air on chest radiograph417 (60.4) Perforation on computed tomography scan322 (98.8)*Shock#41 (5.9)Delayed presentation of >24 h319 (46.2)Mortality prediction models ASA score, median (IQR)2 (2-3) Boey score, median (range)1 (0-3) Mannheim Peritonitis Index, mean (range)15.0 (0-41)Open in a separate window

All continuous variables are expressed as mean (range) unless stated otherwise.

All continuous variables are expressed as mean (range) unless otherwise stated. All categorical variables are expressed as n (%) unless otherwise stated.

Pre-operative factors and ulcer size were evaluated for an association with OPL with univariate logistic regression, as detailed in Table 3. Female gender, increasing age, NSAID use, glucocorticoid use, diabetes mellitus, chronic kidney disease, decreasing hemoglobin, increasing urea, increasing ASA score, increasing ulcer size, increasing Boey’s score, and Mannheim Peritonitis Index were significantly associated with an increased risk of OPL on univariate analysis. On multivariate analysis, however, only increasing ulcer size was independently associated with increased odds of OPL (OR 3.30, 95% CI: 1.81-6.02, PTable 3Risk factors for omental patch repair leakRisk factorUnivariate analysisMultivariate analysis Odds ratio (95% CI)P-valueOdds ratio (95% CI)P-valueGender (female)4.54 (1.56-12.5)0.0052.08 (0.49-8.92)0.322Increasing age (per year)1.05 (1.02-1.09)0.0031.01 (0.96-1.06)0.723Smoking0.15 (0.02-1.13)0.066--Previous peptic ulcer1.92 (0.42-8.76)0.40--NSAID use5.28 1.44-20.1)0.0125.45 (0.93-31.8)0.060Steroid use13.3 (3.86-45.7)Receiver-operating characteristics curve correlating ulcer size (mm) with (A) omental patch repair leak and (B) 30-day mortality.Table 4Ulcer size cut-offs for omental patch repair with respective leak rate, sensitivity, specificity, positive and negative LR predictive of omental patch repair leakUlcer size cut-off, mmLeak rate if ulcer size ≥ specified cut-off (%)Leak rate if ulcer size < specified cut-off (%)Sensitivity (%)Specificity (%)Positive LRNegative LR157/89 (7.9)8/601 (1.3)46.788.03.880.61205/50 (10.0)10/640 (1.6)33.393.35.000.71254/23 (17.4)11/667 (1.7)26.797.29.470.76303/17 (17.7)12/673 (1.8)20.097.99.640.82Open in a separate window

LR: Likelihood ratio

LR: Likelihood ratio

Peptic ulcer disease occurs due to an imbalance between mucosal defensive factors (such as the mucus-bicarbonate layer, prostaglandins, cellular regeneration, and healthy mucosal blood flow) and aggravating factors (such as H. pylori infection, acidity, pepsin, smoking, ethanol, NSAIDs, and steroids) [16]. Peptic ulcer disease is common, and hospitalization rates as high as 160-210 cases per 100,000 person-years have been reported [16,17]. PPU is a complication of peptic ulcer disease and requires emergency surgical intervention. Omental patch repair is the standard surgical technique for almost a century as it is safe with acceptable peri-operative risks [3,4,18,19] In this single-center retrospective study including a large sample of PPU patients, we have shown that increasing ulcer size is significantly associated with a higher risk of OPL, and ulcer size ≥25 mm strongly predicts leak rate. Kumar et al. identified ulcer size ≥5 mm as predictor of leak following omental patch repair of perforated duodenal ulcer [8]. Lee et al. similarly demonstrated that ulcer size >5 mm is associated with significantly higher leak rate following laparoscopic fibrin glue repair of PPU (leak rate for ulcer size >5 mm: 29%, ≤5 mm: 10%) [20]. They, however, failed to demonstrated any difference in leak rate for laparoscopic suture repair, though this was attributed to small sample size. The recently published the World Society of Emergency Surgery (WSES) 2020 guidelines suggest ulcer size 2.5 cm in size as a giant and advocate that such ulcers should not be repaired by simple techniques [23]. They report a series of seven PPU patients with giant ulcers who were treated with a free omental plug with excellent outcomes. The definition of giant ulcer is inconsistent in literature, with variable definitions of ulcers being larger than 2-3 cm in size [9,24,25]. It is also unclear if gastric resection procedures for giant ulcers reduce the morbidity burden [26]. In a study including 62 PPU patients with gastric resections, Seow et al. have reported a malignancy risk of 3%, morbidity of 27.7%, and mortality of 24.2% [10], resulting in the conundrum of the choice of omental patch repair versus gastric resection: omental patch repair for large ulcers has high leak rates, yet gastric resections are associated with inferior outcomes. A recent retrospective observational study by Chan et al. in 2019 on 110 patients with PPU demonstrated that the outcomes of omental patch repair are comparable to gastric resection in patients with PPU of ≥20 mm [26]. In cases of large gastric ulcers with suspicion of malignancy, resection with intra-operative frozen pathologic examination is proposed. For large duodenal ulcers, resection or repair with or without pyloric exclusion and external bile drainage is advised. These are, however, grade 2D recommendations which are weak recommendations based on very low-quality evidence [21]. Therefore, more evidence is required to guide the modality of treatment based on ulcer size. To date, however, there is no literature detailing the optimal ulcer size cut-off for safe omental patch repair. While we obtained similar positive LR for cut-off of 25 mm and 30 mm, we propose the use of 25 mm as a cut-off in clinical practice, as a liberal policy of gastric resections is also fraught with high morbidity and mortality [10,27].

Our study reported low incidence of malignancy of 2.3% compared to existing literature. Hodnett et al. reported 7.6% incidence of malignancy in 202 patients with perforated gastric ulcer [28], while a review by Roviello et al. reported 10-16% incidence of gastric cancer in patients with gastric perforation [29]. Perforation of gastric cancer occurs more frequently at advanced stage of disease [30]. In addition, our data reported location of ulcer in the stomach, duodenum and jejunum. The incidence of gastric cancer is significantly low in duodenal ulcers [13]. Low malignancy rates reported by our institution may be attributed due to different patient demographics, increased population willingness to learn about H. pylori infection, and liberal use of esophagogastroduodenoscopy in patients with epigastric pain symptoms [31]. In local context, majority of gastric cancers are identified through esophagogastroduodenoscopy done for evaluation of patients with epigastric pain, constitutional symptoms and anemia.

Our series also reported low 30-day mortality of 7.4% which is consistent with existing literature ranging 1.3% to 20% [9,32]. Delay in surgery has been associated with higher mortality and has been postulated to be due to the extent of peritoneal contamination [33]. A large nationwide cohort study by Boyd-Carson et al. in 2020 on 3809 patients with PPU who underwent emergency laparotomy within 24 h showed an adjusted 4% increase in mortality per every hour delay to surgery, and an adjusted 6% increase in mortality per every hour delay in patients with shock [34]. Median time to surgery from admission in their study was 7.5 h (interquartile range 5-11.6 h) [34]. Svanes et al. reported marked delay of ≥12 h to surgery resulted in higher mortality (22.8% vs. 5.9%, P