Mesh-reinforced pancreaticojejunostomy versus conventional pancreaticojejunostomy after pancreaticoduodenectomy: a retrospective study of 126 patients

From August 2005 to November 2016, 126 patients who underwent mesh-reinforced PJ or conventional PJ after PD in our institution were included in this study. Patients’ data, including demographics, operative procedures, postoperative complications, and mortality, were retrospectively compared between 65 consecutive patients with mesh-reinforced PJ and 61 consecutive patients with conventional PJ. The perioperative management, including antibiotics, perioperative Octreotide administration, and enteral and parenteral nutrition, was the same in both groups. Definitions of pancreatic fistula (PF), postpancreatectomy hemorrhage (PPH), and delayed gastric emptying (DGE) were followed according to the International Study Group of Pancreatic Surgery (ISGPS) [16‐18]. Medical morbidity was termed as conditions not related to surgical complications including cardiac, pulmonary, and renal-related complications. Follow-up results were obtained from patients’ medical records and telephone calls that were made. The diameter of the main pancreatic duct (MPD) 1 year after PD was measured on computed tomography scans and recorded. The last follow-up day was July 10, 2017. The study was approved by the Committee of Ethics of Sir Run Run Shaw Hospital of Zhejiang University. All patients signed a written informed consent acknowledging potential surgical risks.

In both groups, pancreatoduodenectomy (PD) was performed using the standard procedure [19]. Hemostatic management of the cut surface was done by electric coagulation or suture ligatures with 4–0 prolene stitches (Ethicon, Somerville, NJ). Our technique of mesh-reinforced PJ has been reported previously [14, 15]. In brief, pancreatic remnant was isolated 3 cm in length from the pancreatic transection. A mesh (polypropylene mesh, Ethicon, New Jersey, USA) strip of 1.5 cm in width was tightly wrapped over the pancreatic stump in one circle roughly 1.0 cm from the transection edge. The main pancreatic duct was identified, and subsequently, a stent tube was inserted (Fig. 1). The posterior part of jejunal stump and the inner part of mesh in the pancreas were anastomosed using continuous 4–0 prolene stitches (Fig. 2). The pancreatic stump invagination was performed by tightening the posterior sutures after all of the posterior stitches were completed. The continuous sutures were subsequently extended to fix the anterior part of jejunal loop and the inner part of mesh in the pancreatic stump using the same principle (Fig. 3). Conventional pancreaticojejunostomy (CPJ) with end-to-end anastomosis was performed between the pancreatic stump and the jejunum loop by two layers. The outer layer consisted of the remnant of pancreatic capsule and the seromuscular layer of the jejunum. The inner layer consisted of the pancreas and the full thickness of the jejunum. In both groups, a stent tube was placed in the main pancreatic duct.

Statistical analysis was carried out using SPSS version 20.0 (IBM, Armonk, New York, USA). Categorical variables were presented as numbers and percentages; continuous variables were expressed as mean ± standard deviation (SD) or median (range). Categorical variables were compared using the χ² test, or Fisher’s exact test when necessary, and continuous variables with Student’s t test. The risk factors of POPF were investigated by using logistic regression analysis. Parameters that were significant on univariable analysis (p < 0.100) and/or expected to be important clinically were included in the multivariable logistic regression model. Results were expressed as odds ratios (ORs) with 95% confidence intervals (CI). A p value < 0.05 was considered statistically significant.

Patient baseline parameters are shown in Table 1. There were no significant differences between mesh-reinforced PJ and conventional PJ with regard to age (p = 0.312), gender (p = 0.726), body mass index (p = 0.186), and the American Society of Anesthesiologists (ASA) Classification (p = 0.899). Intraoperative parameters are shown in Table 2. No significant differences between mesh-reinforced PJ and conventional PJ groups with regard to pathologic findings (p = 0.839), pancreatic texture (p = 0.688), and blood loss (p = 0.818) were observed. There was also no statistical difference between the two groups concerning operative time (359 ± 69 vs. 351 ± 63, p = 0.487).

Table 3 describes the postoperative complications and economic outcomes. The overall in-hospital mortality was recorded as 4.0% (5 of 126). One fatality was recorded in the mesh-reinforced PJ group, and four fatalities occurred in the conventional PJ group. The cause of death was intraperitoneal hemorrhage in mesh-reinforced PJ group (n = 1). And in conventional PJ group, the causes of death were multiorgan failure (n = 2), intraperitoneal abscess (n = 1), and intraperitoneal hemorrhage (n = 1). There were no substantial differences in mortality between the two groups (p = 0.197). For patients who survived the operation, seven patients required reoperation due to pancreatic anastomosis leakage (n = 3), gastrojejunostomy leakage (n = 2), or intraperitoneal hemorrhage (n = 2). The reoperation rate was 3% (2 of 65) in the mesh-reinforced PJ group and 8% (5 of 61) in the conventional PJ group, with no significant differences (p = 0.262). Postpancreatectomy hemorrhage (PPH) was observed in nine patients. All mentioned cases occurred more than 24 h after PD and originated from the peripancreatic region. The treatment for PPH included blood transfusion (9/9), arterial embolization (4/9), and re-laparotomy for severe PPH (2/9). Postpancreatectomy hemorrhage revealed no substantial difference between the two groups (p = 0.454). Delayed gastric emptying (DGE), biliary leakage, and gastrojejunostomy leakage, as well as intra-abdominal fluid collection, also showed no significant differences between the two groups (Table 3). The incidence of POPF was significantly different between mesh-reinforced PJ and the conventional PJ group (15 vs. 34%, respectively; p = 0.013) (Table 3). Eleven cases of pancreatic fistula of grade B were managed by antibiotics and total parenteral nutrition, and all recovered well after conservative treatment. Treatment for grade C pancreatic fistula included percutaneous drainage (4/4) and re-laparotomy (3/4). One case of grade C in the mesh-reinforced PJ group was treated by re-laparotomy and ultimately recovered well. Two cases of grade C in the conventional PJ group received re-laparotomy and died from POPF-induced sepsis and bleeding. Total surgical morbidity (25 versus 43%; p = 0.032) and postoperative length of hospital stay (18 ± 9 days vs. 23 ± 12 days; p = 0.016) were significantly different between mesh-reinforced PJ and conventional PJ groups (Table 3). There was no substantial difference in the total hospital costs (106,265 ± 1231 RMB vs. 114,265 ± 1349 RMB; p = 0.142) between mesh-reinforced PJ and conventional PJ groups (Table 3). Forty-one cases in the mesh-reinforced PJ and 45 cases in the conventional PJ group received computed tomography scans in a yearly follow-up (Table 4). No statistical differences between the two groups concerning dilated pancreatic duct were observed (p = 0.578).

Predictors of PF for all patients are shown in Table 5. Factors significantly increasing the risk of PF by univariate logistic regression analysis included soft pancreatic texture, ampullary or duodenal disease, and type of conventional pancreaticojejunostomy (p < 0.05). A multivariate logistic regression analysis revealed that the independent factors that were highly associated with PF were soft pancreatic texture and type of conventional pancreaticojejunostomy (p < 0.05).