Brief Report on a Novel Reconstruction Technique Following Pancreaticoduodenectomy

Abstract

Mini-abstract: Although pancreaticoduodenectomy (PD) has significantly improved and is considered a safe procedure, it is still associated with increased rates of morbidity. The aim of the present study is to evaluate an alternative reconstruction technique following PD. This technique uses an isolated hepaticojejunal anastomosis and presents low postoperative morbidity rates.

INTRODUCTION

Pancreaticoduodenectomy (PD) or Whipple’s procedure remains the standard of treatment for periampullary neoplasms.¹ During the last few years, the technique has significantly improved, and it is now considered a safe procedure with a hospital mortality rate less than 5% in high-volume centers.² Nonetheless, PD is still associated with increased rates of morbidity and major postoperative complications, such as delayed gastric emptying (DGE) and postoperative pancreatic fistula (POPF), which are thought to be the Achilles heel of this procedure.³

The aim of the present study is to evaluate an alternative reconstruction technique following Whipple’s procedure. This technique uses a pancreaticojejunal anastomosis in the same loop with the duodenojejunal anastomosis and presents remarkably lower postoperative morbidity rates compared with the classic Whipple procedure. The theory behind this approach is that, in case of anastomotic disruption, the activation of pancreatic enzymes by bile salt and alkalized pH could be avoided by the placement of the pancreatic anastomosis far from the biliary and closer to the low pH of the gastric fluid.⁴ This is probably due to the fact that the activity of the pancreatic amylase and lipase is dependent on the pH of the bile and the gastric juice.⁵ When the hepaticojejunostomy is distanced from the gastrojejunostomy and pancreaticojejunostomy in a Roux-en-Y formation, the alkali from the bile duct drain away from the pancreaticojejunostomy and cannot be affected by the gastric juice’s pH, the increase of which would cause the activation of pancreatic enzymes. Although not completely avoidable, leakage from the pancreatic anastomosis, in this case, is less harmful and easier to manage postoperatively.

PATIENTS

During the last decade (from January 2010 to December 2019), 123 consecutive patients were operated with this novel technique in our surgical department by a single surgical team. Median age was 66 years with 55.6% males. Among them, 91.9% presented with malignancy (61.1% pancreatic, 12.6% ampullary, 10.6% bile duct) and 8.1% with a benign neoplasm or intraductal papillary mucinous neoplasm.

TECHNIQUE

The operation begins with a bilateral Kocher incision. After initial evaluation of tumor resectability, the head of the pancreas, the duodenum, the gallbladder, and the bile duct are resected en bloc with preservation of the pylorus. A typical lymph node resection is also performed.

The pancreaticojejunal anastomosis is performed with a retrocolic, end-to-side, duct-to-mucosa anastomosis, using a short jejunal limb. This starts with a posterior continuous row of seromuscular sutures securing the jejunum to the pancreas. A pancreatic duct-to-mucosa anastomosis is performed to a 3 mm enterotomy in the jejunum, with interrupted sutures. A second circumferential anterior continuous layer of seromuscular sutures between the jejunum and the pancreatic capsule reinforces the anastomosis. In most cases, we use an internal pancreatic duct stent (size 4F) and, if feasible, a wrapping of omentum around the anastomosis.

In addition, approximately 30 cm distally in the same jejunal loop above the transverse colon, the upper gastrointestinal is reconstructed by performing a pylorus-preserving duodenojejunostomy using a 2-layer, interrupted, end-to-side suturing. We emphasize that a mechanical dilatation of the pylorus is performed prior to construction of the anastomosis to a diameter of approximately 5 cm.

Subsequently, the bile duct is reunited to the distal main long limb of the gastrointestinal tract by performing an end-to-side retrocolic hepaticojejunostomy with 1 layer of interrupted absorbable sutures. Finally, a side-to-side jejunojejunal anastomosis with continuous sutures in 2 layers is created 35 cm distally from the hepaticojejunal anastomosis, in order to complete the Roux-en-Y reconstruction (Fig. 1). Drains are put below the anastomoses (bile/pancreatic) in the upper abdomen. The abdomen is closed in a single, continuous layer.

FIGURE 1.

Modified Whipple technique. Created by “Medical illustrations ©Medical-Artist.com.”

RESULTS

Our records show that median hospitalization time was 11 days. Clinically significant POPF (grades B and C) complicated 4.9% of the cases, while only 7.3% patients suffered from DGE. Postoperative hemorrhage presented in 4 patients (3.3%) and 8 patients (6.5%) required reoperation due to postoperative complications. Overall morbidity (Clavien-Dindo > 2) was 14.7%. Perioperative mortality was 4.1% (2/5 patients died from nonprocedure-related complications). Median overall survival was 35 months, while 1-, 2-, and 5-year survival was 79.4%, 61.5%, and 37.9%, respectively. There is no specific and systematic follow-up data on weight loss. However, on 1-year follow-up, most of our patients self-reported their weight as “normal.”

DISCUSSION

In this series, the rates of POPF are comparable with results from high-volume hospitals.^2,6–8 Historical control data of our team performing standard pylorus-preserving Whipple on more than 100 consecutive patients during the prior decade (2000–2010) show that median hospitalization time for these patients was approximately 17 days, clinically significant POPF rate was around 10% and DGE rate was around 16%.

On the contrary, our DGE rates are significantly lower than most series reported in the literature. This may be due to several reasons, including the nonactivation of pancreatic proenzymes close to the pancreaticojejunal anastomoses, a factor which may be associated with less local inflammation and potential gastroparesis. The mechanical dilation of the pyloric muscle fibers, which is a standard part of our procedure, may also play an important role in DGE prevention. Adding to that, placement of the hepaticojejunostomy distally from the gastrojejunostomy prevents bile salt reflux into the stomach.

The only drawback of the described technique is the prolongation of operative time, compared with the classic PD, due to an extra anastomosis.

CONCLUSIONS

In our technique, the different arrangement of the anastomoses compared with the “traditional” Whipple may alter the mixing of pancreatic, gastric, and biliary fluids, resulting in reduced incidence of serious complications in our patients. In addition, pylorus preservation with dilation possibly leads to diminished rates of DGE. The proposed technique, based on our clinical results, deserves evaluation from other centers of pancreatic surgery.