Efficacy of Modified Technique in Pancreatojejunostomy to Prevent Postoperative Pancreatic Fistula after Pancreatoduodenectomy

Introduction: In pancreatoduodenectomy (PD), postoperative pancreatic fistula (POPF) remains the single most important cause of morbidity. We present a modification for duct to mucosa pancreaticojejunostomy. Materials and Methods: Total 134 patients, who had undergone PD or pylorus preserving PD (PPPD) between November 2007 and October 2013 at our institution, were analyzed. From April 2012 to December 2014, 53 consecutive patients underwent PD or PPPD by the new modified technique and 81 patients by the former technique before March 2012. The preoperative demographics and clinical information were retrospectively obtained from both groups and were analyzed along with risk factors of POPF. Moreover, risk factors for POPF grade B/C were analysed by univariate and multivariate analysis. Results: Operation procedures were pylorus preserving PD in 119 and PD in 15. Incidence of POPF grade B/C was 11% in the new method, which was significantly lower than in the former method (38%) (p=0.0135). Moreover, risk factors for POPF grade B/C in univariate analysis were texture of pancreas (p=0.0004), dilatation of pancreatic duct (p=0.0100), and anastomosis method (p=0.0135). In multivariate analysis, risk factors were texture of pancreas (p=0.0010) and anastomosis method (p=0.053). Conclusions: The new technique in pancreticojejunostomy was safe and reliable with low POPF grade B/C rate.


Introduction
Since the first en bloc resection of the pancreas head and duodenum in 1898 by Codvilla, pancreatic surgery has undergone major modifications. Milestones by Kausch in 1909, Whipple in 1935, and Traverso in 1978 led to the pylorus-preserving pancreatoduodenectomy (PPPD), which is currently performed as a standard procedure [1][2][3]. Advances in perioperative management and surgical techniques have made PD a relatively safe surgical procedure, with a mortality rate lower than 5% [4]. However, morbidity is still high, between 30% and 50%. Postoperative pancreatic fistula (POPF) (5-29%) remains the single most important cause of morbidity, which can lead to prolonged hospitalizations, the need for repeated surgical interventions, and increased mortality rates [5]. The surgical technique is one improvable aspect that might reduce the pancreatic leakage rate after PD [6]. Several pancreatic anastomotic techniques have been proposed and tested, including end-to-side, with or without duct-tomucosa anastomosis, end-to-end invagination and, arguably, anastomosis of the remnant pancreas with the stomach is an additional method, although it is still debated which of them has any clear advantage [7,8].
Pancreatojejunostomy has been the most commonly used method to restore pancreatoenteric continuity after PD. The main techniques to perform pancreatojejunal anastomosis are the invagination or ''dunking'' technique and the ''duct-to-mucosa'' anastomosis [9,10]. Many technical variations of the invagination technique exist, and they involve differences in the suture material, the number of layers, the number of sutures, running versus interrupted sutures, the binding versus the traditional suturing technique, and other modifications [11][12][13]. Moreover, it is also reported in a prospective randomized trial that external drainage of pancreatic duct with a stent reduced leakage rate of pancreatojejunostomy after PD [14]. Nevertheless, no consensus exists on which of these approaches represents the best way for reducing PF after PD. We present a modification for duct to mucosa end-to-side pancreatojejunostomy, with a seromuscular jejunal flap, in order to prevent POPF. In addition, we compared the operative and postoperative outcomes between new method and historical matched control and analyzed risk factors for POPF grade B/C.

Methods
Total 134 patients, who had undergone PD or PPPD between November 2007 and December 2014 at our institution by the same surgical team, were retrospectively analyzed. During the period April 2012 to December 2014, 53 consecutive patients underwent PD or PPPD by the new technique. 81 patients were underwent PD or PPPD by the former technique before March 2012. The following preoperative demographics and clinical information were retrospectively obtained from both groups' medical records and were analyzed along with risk factors of POPF: age, sex, underlying diseases, operative procedures, operative time, transfusion requirements, etc. Same group of surgeons performed the former and modified pancreatojejenostomy procedure.

New surgical technique
In this report, we describe a new pancreatojejunostomy technique that has been in use since April 2012. Basically, the pancreatojejunostomy procedure can be separated into a duct-tomucosa pancreatojejunostomy and an anastomosis between the pancreatic parenchyma and jejunum. Figure 1: A) An incision is made on the serosa at the planned location of the jejunal anastomosis (arrows), up to the muscularis, with a width that is tailored to that of the pancreatic arenchyma. B) In duct-to-mucosa pancreaticojejunostomy, the anastomosis between the posterior wall of the pancreatic duct and jejunal mucosa is performed without initially creating a small opening in the jejunum. C) A small opening is created in the jejunum, and the anterior wall of the pancreatic duct and the jejunal mucosa are anastomosed together. D) In the anastomosis between the pancreatic parenchyma and jejunum, two needles in the caudal part of the head are tightly sutured with all the layers of the pancreatic parenchyma, whereas for the others, adhesive anastomosis was performed with two layers of the anterior and posterior walls of the pancreatic parenchyma.
First, an incision is made on the serosa at the planned location of the jejunal anastomosis, up to the muscularis, with a width that is tailored to that of the pancreatic parenchyma ( Figure 1A). In duct-tomucosa pancreatojejunostomy, the anastomosis between the posterior wall of the pancreatic duct and jejunal mucosa is performed without initially creating a small opening in the jejunum ( Figure 1B); thereafter, a small opening is created in the jejunum, and the anterior wall of the pancreatic duct and the jejunal mucosa are anastomosed together ( Figure 1C). In the anastomosis between the pancreatic parenchyma and jejunum, two needles in the cranial and caudal part of the head are tightly sutured with all the layers of the pancreatic parenchyma, whereas for the others, adhesive anastomosis was performed with two layers of the anterior and posterior walls of the pancreatic parenchyma ( Figure 1D). Pancreatic duct drainage is usually performed using the incomplete drainage method. With regard to the pancreatic duct tube, lost tubes have been used only in patients with a cirrhotic pancreas and main pancreatic duct dilatation since April 2012. However, before April 2012, procedures involving the use of lost tubes have been generally performed.
The differences of the new method from the earlier anastomosis methods are as follows: 1)In duct-to-mucosa pancreaticojejunostomy, the timing of the creation of a small opening in the jejunum is different. In the conventional method, a small opening was first created in the jejunum, and then the pancreatic duct and the jejunal mucosa were anastomosed. In the new method implemented since April 2012, the posterior wall is first anastomosed and a small opening is created before the anterior wall of the pancreatic duct is anastomosed to the jejunal mucosa. 2)In the new method, the anastomosis is preceded by an incision of the jejunal serosa and opening of the muscle layer. 3) The anastomosis between the pancreatic parenchyma and jejunum had previously been performed by creating an adhesive anastomosis between the pancreatic parenchyma and jejunum; however, since April 2012, the procedure consisted of an adhesive anastomosis with all the layers of the pancreatic parenchyma with two needles on the caudal side of the head, and with two layers of the anterior and posterior walls of the pancreatic parenchyma for all the others.

Standard postoperative care
Amylase level was monitored daily in the serum and in the intraoperatively-placed abdominal drains on the first and fourth postoperative day. Computed tomography was performed on the fourth postoperative day. In the absence of signs of POPF, abdominal drains were removed and oral food intake was begun on the seventh postoperative day.
The diagnosis of a POPF was based on the definition of the International Study Group on Pancreatic Fistula (ISGPF) [15]. In this study, According to the ISGPF definitions, grades B and C fistulas were considered as POPF.

Statistical analysis
Data was expressed as mean (SD). Statistical calculations were performed using SPSS (version 13.0 for Windows; SPSS, Inc). Comparisons between groups were tested using the Pearson χ 2 test. For continuous variables, independent samples t-test was used to compare the 2 groups. Odds ratios were used to estimate relative risk for POPF. Logistic regression was used for univariate analysis, while multiple logistic regression analysis was used for multivariate analysis. For multivariate analysis, variables possibly significant (p<0.05) on univariate analysis were evaluated. P values less than 0.05 were considered significant.

Clinical characteristics
The indications for operation were: adenocarcinoma of the head of the pancreas in 78 patients, distal bile duct cancer in 29, and the other tumor in 27 ( Table 1)    No differences were noted in age, sex, lying disease and neoadjuvant chemotherapy between the two groups without PD or PPPD procedure ( Table 2). And the two groups were not statistically different for operation-related factors, including texture of pancreas, diameter of pancreatic duct, operative time, operative blood loss, need for transfusion, combined vascular resection. New method group (those who underwent new surgical technique of pancreatojejunostomy) had a significantly lower incidence of pancreatic fistula over grade B (P=. 00135). Pancreatic fistulas over grade B occurred in 31 of 81 patients (38%) in the former method group and in 6 of the 53 patients (11%) in the new method group. No statistically significant difference was noted in the occurrence of other complications except pancreatic fistula between the two groups (data not shown).

Identification of prognostic risk factors for POPF
In a logistic regression univariate analysis, soft pancreas (P=0.0004), diameter of pancreatic duct<3.0 mm (P=0.0100), and new method of pancreaticojejunostomy (P=0.0135) were identified as prognostic risk factors for POPF (Table 2). Other factors such as patient gender, age, neoadjuvant chemotherapy, operative time, operative blood loss, need for transfusion, and combined vascular resection were not significantly correlated with the incidence of POPF. And in a multivariate analysis, new surgical technique (OR=0.099, 95%CI=0.020 -0.503, p-value=0.0053) were identified as prognostic risk factors for POPF ( Table 3).

Variables
Odds ratio 95%Cl p-value

Discussion
In the pancreaticojejunostomy, several technical variations have been proposed, in an effort to minimize postoperative pancreatic fistula rates [16][17][18]. Postoperative pancreatic fistula (POPF) is the most common and severe complication after PD, and its incidence varies widely in the reported series, between 2% and over 31%. Because POPF may strongly associate with other complications and affect the short-and long-term outcomes, an uncomplicated course is particularly important for the patients who undergo PD or PPPD [19,20]. The most important risk factors identified are technique, soft pancreatic texture and main pancreatic duct diameter of 3 mm or less [21][22][23]. In this report, we describe a new pancreaticojejunostomy technique that has been in use since April 2012. Basically, the pancreaticojejunostomy procedure can be separated into a duct-tomucosa pancreaticojejunostomy and an anastomosis between the pancreatic parenchyma and jejunum. Advantages of the new methodas follows; 1) Duct-to-mucosa pancreaticojejunostomy allows for a posterior wall anastomosis by facilitating the achievement of a reliable anastomosis between the pancreatic duct and jejunal mucosa by avoiding the need for the creation of a small opening in the jejunumif the anastomosis is performed after creating a small opening, the jejunal mucosa may prolapse outside the serosa, making the anastomosis possibly difficult to achieve. Nevertheless, all of this can be prevented using the new method. 2) By making an incision in advance in the jejunal mucosa at the site of the planned anastomosis, a duct-tomucosa pancreaticojejunostomy can be achieved without applying excessive tension. When the pancreatic duct and all the layers of the jejunum are anastomosed, the excessive tension exerted by the jejunal serosa on the anastomotic site can be prevented. 3) In the anastomosis between the pancreatic parenchyma and jejunum, the adhesion intensity can be strengthened by attaching two needles on the caudal side of the head to all the layers of the pancreatic parenchyma. In addition, damage to the main pancreatic duct can be prevented by creating an adhesive anastomosis between the pancreatic parenchyma around the main pancreatic duct, and anterior and posterior walls. 4) In the pancreatic duct drainage method, it remains controversial whether an external drainage or a lost tube should be chosen; however, in the present study, the lost tube was used only in the patients with a cirrhotic pancreas and main pancreatic duct dilation.
Based on clinical relevance, POPFs were classified as grade A, B, or C. Our study reflected the results of the retrospective clinical data; pancreatic fistulas over grade B occurred in 31 of 81 patients (38%) in the former method group and in 6 of the 53 patients (11%) in the new method group, which underlines the severe influence of "soft" pancreatic tissue on postoperative outcome. Other clinical and surgical factors that have been reported to correlate with the POPF rate include age>65 years, preoperative jaundice, operation time, amount of intraoperative blood loss, and intraoperative transfusion of erythrocyte concentrates [24,25]. In our study, soft pancreas and new method were identified by multivariate analysis as single prognostic risk factors for POPF.
Several limitations affect the interpretation of the present findings. This study was based on a retrospective analysis in a single center and a small number of patients were included. In conclusion, this modified new technique appears to be safe and reliable. Because this is a preliminary report of a small series, it is of essential importance that it is evaluated via a prospective study in a larger series, before firm conclusions can be drawn.

Ethics Approval
Collecting of the retrospective data from patients was approved by Tokyo Medical University Ethics Committee.