Invaginated Duct to Mucosa Pancreaticojejunostomy Reduced Postoperative Pancreatic Leakage: A Matched Case-controlled Study

Background Postoperative pancreatic leakage was also the obstacle of pancreaticoduodenectomy (PD) which always followed the failure of pancreaticojejunostomy. Dozens of pancreaticojejunostomy had been reported. None showed superiority over others. To assess the potential advantages of invaginated duct to mucosa pancreaticojejunostomy (invaginated D-M PJ), we introduced this study. Methods Retrospectively analyzed the related data from the patients who had their pancreaticodedunostomy due to malignant tumor in The First Aliated Hospital to Henan University of Science and Technology during January 2017 to August 2019. According to the different procedure of pancreaticojejunostomy, the patients divided into custom D-M group and invaginated D-M group. Sex, age, pancreatic duct size and pancreatic texture were matched. Pancreatic leakage and other complications were compared. SPSS 16.0 was employed for analysis. Results 48 pairs of patients were involved. Both groups had almost the same baseline characterisitics, such as Sex (P=1.000), Age (P=0.897), ASA (P=0.575), BMI (P=0.873), pancreatic duct size (P=0.932), pancreatic texture (P=1.000) and tumor origin (P=0.686). No statistical difference was observed relation to operation outcomes, such as operation time (P=0.632), pancreaticojejunostomy time (P=0.748), blood loss (P=0.617) and number of required transfusion (P=0.523). Pancreatic leakage was signicantly declined for invaginated D-M group (P=0.005). The difference of other complications, such as, bleeding (P=0.617), biliary leakage (P=0.646), pneumonia (P=0.594) and thrombosis (P=0.714), didn’t reach statistical signicance. Postoperative hospitalization was almost the same for both groups

Background Pancreaticodudenectomy(PD) was one of the most problematic operations with high morbidity and mortality [1][2][3][4]. Most mortality was associated to pancreatic leakage which was mainly secondary to the failure of pancreaticojejunostomy (PJ). Dozens of anastomosises had been introduced; pancreatic leakage was still the obstacle of PD. Duct to mucosa (D-M) anastomosis and its modi cations might be the most popular ones, while, none was the standard. Kakita anastomosis [5] and Blumgart anastomosis [6] were the most widely used procedures with lower pancreatic leakage. Hong's single stitch anastomosis [7] was also popular in china. None showed the signi cant superiority over others [8]. As to the pancreaticojejunostomy, invaginated procedure was still popular since its introduction. Chen's Usuture PJ [9] and Peng's binding PJ [10,11] were still widely used. No consensus of the method of PJ was reached jet [12,13]. All the procedures in used now had some advantages, while disadvantages still existed. That was partially the reasons why none procedure of PJ was shown the superiority over others. Now, we practiced a novel procedure of PJ which theoretically involved all the advantages mentioned before. That is invaginated D-M PJ.

Patients characteristic
All the patients administrated in The First A liated Hospital to Henan University of Science and Technology during January 2017 to August 2019 who needed PD for malignant tumor. All the operations were completed by the same team. This was a retrospectively designed trial. This clinical trial was approved by the ethics committee of The First A liated Hospital to Henan University of Science and Technology.
According to the procedure of PJ, we divide the patients into two groups: Custom D-M group and invaginated D-M group. All patients included were aged 18 to 80 years old; the American society of anesthesiologists (ASA) classi cation should be at 1 to 3; the diagnosis was approved by pathology; enhanced CT scan and MRCP were routinely performed. All patients included should be proved by enhanced CT or / and enhanced MR of no tumor invasion of vessels. All pancreatic duct size was dilated more than 2 mm. In case of bias control, we matched some factors. The matched factors were age (± 5 y), sex, pancreatic texture and pancreatic duct size (± 1 mm). Those factors were reported as important risk factors of pancreatic leakage [9]. All the characteristics of the two groups were list in the table 1.
Operation procedure All the operations were performed under general anesthesia. Pancreaticoduodenectomy was performed with local lymph nodes dissection. Jejunal limb was brought up through retrocolic root. After end to side PJ, cholangiojejunostomy was performed with 4-0 knotless suture (Strata x, Ethicon) by a running suture about 8 cm distal to the PJ. Gastrojejunostomy was performed by a circular stapler (Panther) about 50 cm distal to cholangiojejunostomy. Additional jejunojejunostomy was performed by 4-0 knotless suture (Strata x, Ethicon) about 10 cm distal to gastrojejunostomy. The difference between the two groups was the procedure of PJ. Custom D-M group was performed as modi ed Blumgart anastomoses descript by Shoji Kawakatsu [8]. Invaginated D-M group was performed as follow: a 3-0 polypropylene suture (Prolene, Ethicon) was employed to nish the anastomosis. The rst stitch trans xed pancreas from anterior to posterior at about 1 cm from the superior board of pancreatic neck where located 1.5 cm distal to the pancreatic cutting end. The needle gone through the seromuscular of the jejunum limb at the antimesenteric margin. Then, the rst knot was made anteriorily shown as gure 1. The second stitch was about 1 cm departed from the rst knot which was also 1.5 cm distal to the pancreatic cutting end. The second stitch trans xed the pancreatic neck from anterior to posterior ( gure 2 ), then, it gone through the seromuscular layer of the jejunum 1 cm from the rst stitch ( gure 2 ). The needle gone through the pancreas from posterior to anterior about 1 cm distal to the pancreatic cutting end ( gure 2 ), then, it gone through the seromuscular layer of the jejunum 1 cm from the rst knot ( gure 2 ). The suture was strengthened, and duplicated the second suture ( gure 2 ). When the suture reached to the superior of pancreatic duct, it suspended. Then, duct to mucosa anastomosis was performed as custom D-M by 4-0 polyglactin suture (Vicryl Plus, Ethicon). After duct to mucosa anastomosis, pancreaticojejunostomy continued as descript above until to the inferior board of pancreatic neck. The last stitch was just like the rst one. When the needle gone through the pancreatic neck about 1.5 cm distal to the pancreatic cutting end from anterior to posterior, it gone through the seromuscular layer of the jejunum limb; then, the second knot was made ( gure 3). Before making the knot, what was emphasized was to strengthen every stitch moderately. Others should be emphasized were following: the jejunal serosa was destructed by electrocoagulation to accelerate healing of pancreatic cutting end and jejunum limb as descript by Hong [7]; pancreatic duct stent should be placed when duct to mucosa anastomosis; pancreatic duct should be separated about 2 mm when transaction of pancreas, this separation facilitated the invigation of pancreatic duct to mucosa of jejunum. Drainage tubes should be placed around cholangiojejunostomy and pancreaticojejunostomy with additional pelvic drainage when necessary. Peritoneal lavage with distilled water was emphasized before abdominal closure.

Perioperative management
Percutaneous transhepatic cholangial drainage (PTCD) were perform when cholangitis existed. Anemia correction should be performed routinely to 90 g/L or above. Serum albumin should be maintained at normal level. Breathing exercises were performed at least 2 days before operation. The pressure of O 2 and CO 2 should be at normal level by arterial blood gas analysis.
Postoperative management was performed according to the principle of Enhanced Recovery after Surgery (ERAS) [14]. Gastric tube was removed after operation when the patient recovered from anesthesia. Abdominal drainage amylase test was routinely performed at 1, 3, 7 day postoperatively with additional test when pancreatic leakage was suspected. Abdominal drainage tubes were removed when the drainage was less than 20 ml/d, which should be retained at least 7 days postoperatively. Abdominal CT scan was routinely performed to ensure no ascites before drainage removal. Discharged criteria was set as recovery to semi uid diet without symptomatic pancreatic leakage or biliary leakage at least 7 days postoperatively.

De nition
Primary endpoint was pancreatic leakage. Pancreatic leakage was detected following the International Study Group of Pancreatic Fistula (ISGPF) [15].

Discussion
Pancreatic leakage was following the failure of the anastomosis of pancreatic stump. It was still the most fatal morbidity after pancreaticoduodenectomy (PD). Duct to mucosa pancreaticojejunostomy (PJ) might be the most popular method worldwide. Though a serial of modi cations had been introduced [7][8][9], pancreatic leakage didn't decrease. Each modi cation had some advantages over others theoretically, while, defect still existed. Kakita duct to mucosa anastomosis [5] had penetrating sutures to close the pancreatic cutting end and the serosa of jejunum limb. This penetrating suture was prone to progress tangential shear following failure of anastomosis, which led to pancreatic leakage. Blumgart anastomosis [8,16] adopt transpancreatic U-suture which avoided tangential shear and the pancreatic stump could be covered by jejunum serosa to protect the knots from cutting through the pancreatic tissue. This U-suture might cause ischemia of the pancreatic stump, which hampered the healing of duct to mucosa anastomosis, and increased the risk of pancreatic leakage from the pancreatic cutting end followed by hemorrhage. Although the modi ed Blumgart anastomosis [8] adopted U-suture to avoid shear force, the pancreatic leakage was detected without any statistical difference. Maybe the potential ischemia of pancreatic stump offset the potential bene t from less tangential shear force. Hong's single stitch duct to mucosa PJ [7] had something like Kakita anastomosis. It was also prone to progress tangential shear of the pancreatic stump. It emphasized destruction of the jejunal serosa by electrocoagulation which might accelerate healing of pancreatic cutting end and jejunum limb. Chen's anastomosis was in fact an invagination PJ. Chen's U-suture was something different from Blumgart Usuture. Its shear force was parallel the vessels, this change greatly reduced the risk of ischemia of pancreatic stump. While, the pancreatic stump was invaginated into jejunum serosa by chen's U-suture, the seromuscular layer of jejunum could protect the pancreas from rupture by the suture. As invagination asastomosis, chen's anastomosis had a potential risk. Even one suture was failure, a big defect of the anastomosis formed which may lead to high volume of pancreatic leakage. That may be lethal. That was why we adopt D-M PJ recent years.
The invaginated D-M PJ anastomosis which we introduced was indeed a modi cation of custom D-M PJ. Duct to mucosa suture was performed as usual. What should be emphasized was the pancreatic duct stent implantation, this procedure may reduce the risk of pancreatic leakage potentially [17,18], though controversy existed [19]. Four to six sutures might be suitable. Too many sutures might increase the risk of ischemia. Another thing should be emphasized was separation of pancreatic duct. 2 to 3 mm separation of pancreatic duct might facilitate the pancreatic duct invaginated into mucosa of jejunum, which might accelerate healing. The suturing of pancreatic stump and jejunum seromuscular was just like Chen's U-suture which ensured the pancreatic stump wholly invaginated into jejunal seromuscular. This procedure greatly decreased the tangential shear force of duct to mucosa anastomosis followed by decreased risk of pancreatic leakage. Our destruction by electrocoagulation of the jejunal serosa accelerated healing of pancreatic stump and jejunal serosa which also decreased the risk of pancreatic leakage. By our procedure, what should be emphasized was avoidance to suture through the pancreatic duct to reduce pancreatic leakage from pinhole. Another thing should be emphasized was the strengthening of every stitch moderately before the last knot when suturing pancreatic stump and jejunal serosa. As a running suture, potential time saving was obvious. Theoretically, the invaginated D-M PJ integrated all the advantages mentioned above without obvious defect, it reduced the risk of pancreatic leakage.
Our study showed no grade C pancreatic leakage in both groups. A decreased risk of grade B pancreatic leakage was observed by invaginated D-M PJ, which resulted in statistically decreased risk of whole pancreatic leakage. Other complications were the same between the two groups. This reality was just according to the theory above.
The limitation of this method was its adaptability. Not all the patients adapted this method. When the pancreatic duct didn't dilate, invaginated D-M PJ wasn't suitable. As a retrospectively designed trial, bias might exist. Although some risk factors were matched, those were not the all. The small volume might also lead to some bias.      Figure 1 The rst knot was made anteriorily shown as gure 1.

Conclusions
Page 13/14 Figure 2 The suture was strengthened, and duplicated the second suture