Once the pancreas is mobilized and the site of the injury fully defined, it is important to classify the extent of the injury. There are numerous classification systems, and they all have in common a measure of the extent of the parenchymal and main ductal injury. We prefer the American Association for the Surgery of Trauma (AAST) Committee on Organ Injury Scaling classification. See Table 47.1.

For minor parenchymal injuries without ductal disruption, we do not suture the gland but would leave a drain onto the site of the injury. We use soft silicone suction drains.

If the parenchyma is significantly divided or the main pancreatic duct (MPD) disrupted, for injuries to the left of the PV/SMV, we advocate a distal pancreatectomy.

The first step to perform resection of the mobilized distal pancreas should be to ligate the splenic artery and vein to decrease the possibility of extensive bleeding during the resection. Ligation of both vessels about 2 cm to the right of the injury site is performed so that they are not inadvertently damaged during the transection of the parenchyma. In the same way, the surgeon should continue the mobilization of the pancreas also for 2 cm to the right to the site of the proposed resection line. The surgeon then takes a soft bowel clamp and applies it on the pancreas as proximally as possible and divides the parenchyma with a scalpel. By intermittently releasing the soft bowel clamp, one will identify the superior and inferior pancreatic arteries and overrun them with a 5-0 Prolene figure-of-eight stitch. The bites of the needle are as close as possible to the bleeding vessel, including minimal pancreatic tissue. If it is applied further away from the vessel, there is a good possibility that the thin stitch will cut through the parenchyma while applying tension on throwing the knot causing small irritating bleeding.

Then, the surgeon tries to identify the very small main pancreatic duct. This identification, although difficult, is possible in most cases. For that reason, one should transect the pancreas with a blade; transecting it by electrocautery will make it very difficult to identify the opening of the duct through the cauterized, coagulated pancreatic tissue. However, with the modern electrocautery devises, this is much less likely and division of the pancreas is possible using electrocautery. The pancreatic duct when identified is closed using a nonabsorbable 5/0 suture using a figure-of-eight stitch. Although the different techniques of closure of the pancreatic stump aim at controlling the bleeding as well as a leak from the pancreatic duct by compressing them within the pancreatic tissue, applying the figure-of-eight stitch at the pancreatic duct separately can diminish the risk of fistula formation.

The pancreatic stump should be closed by inserting overlapping interrupted mattress sutures of polypropylene or silk. Which is the best way in inserting these mattress sutures? There is a tendency for the surgeon to move the needle holder, while he/she is inserting sutures, from away toward his/her body (in the right-handed surgeon, from the right to the left). So in the case of the pancreas, the mattress stitch will be inserted from the posterior pancreatic surface to the anterior and then from the anterior to the posterior. Consequently, the stitch will be knotted on the posterior surface of the pancreas. As the normal pancreatic tissue is very soft, it is important, when the surgeon puts tension on the knot, to do it in such a way that it compresses the occluded pancreatic tissue but does not cut through it. This can be better achieved when the knot is in front, so that more controlled tension is applied with the two index fingers. Therefore, the mattress stitch is inserted “backhanded” (remember the surgeon is standing on the patient’s right side) starting from the anterior surface to the posterior and back from the posterior to the anterior. The first knot – the one that counts – should be straight and double, so that appropriate tension can be applied and it does not give way while “relaxing” it on throwing the second knot.

Resection of the body of the pancreas can also be achieved with the linear stapler, but a 4.5–4.8 mm TA or GIA stapler (green) must be used. In our hands, handsewing of the pancreatic stump has achieved better results, but this is not reflected in the literature. If the pancreas is very swollen, as happens frequently in patients with blunt transection of the body of the pancreas especially when there is a delayed presentation – a stapler should not be used. The clips are too small to include the whole width of the transected pancreas. In this case, a thick stitch is used (usually stitches with high tensile strength like 0, 1, 2 which are also thicker). This helps the surgeon in two ways: firstly, the needle is longer, so that the surgeon includes in his/her bite the whole cut surface of the pancreas, and, secondly, as the stitch is thicker, there is less chance of cutting through the parenchyma – “like a hot wire through Swiss cheese.”

The distal pancreas should be removed together with the spleen. If the patient is physiologically stable (usually in isolated pancreatic injuries), an attempt can be made to preserve the spleen. This means additional time to dissect small perforating vessels originating from splenic vasculature on the posterior pancreas surface. As a rule, we do not attempt splenic preservation when we are performing distal pancreatectomy for trauma in adults.

If a splenectomy is performed, remember to give the patient pneumococcal vaccine 2–3 weeks after the procedure to help reduce the incidence of overwhelming postsplenectomy sepsis from encapsulated bacteria.

A parenchymal injury to the head without major duct disruption requires good suction drainage alone. If one is unable to classify the ductal injury in the head, then our recommendation is to err on the conservative side and only drain the head. Should an MPD injury be missed and a pancreatic fistula develop, endoscopic intervention with ERCP and MPD stenting works well in our experience to stop the leak. We do not advocate major resections of the head unless there is a defined major injury to the duodenal-ampullary complex or the missile has essentially divided the pancreas for you. When the head of the pancreas is shattered and there is a significant duodenal injury and minimal dissection is required to remove the head, a pancreaticoduodenectomy is performed. When dividing the pancreas as it runs posterior to the portal vein/superior mesenteric vein (PV/SMV), it is important to identify by palpation where the superior mesenteric artery runs to the left of the veins. Traction on the pancreas head during the process of dividing the uncinate process can tent up the artery, and injury to the SMA is possible at this point. As a matter of fact the uncinate process can be left on the portal vein making the resection of the head of the pancreas less tedious.

The remaining issue is the timing of the reconstruction. In experienced hands and with a stable patient without the need for damage control, an immediate reconstruction can be performed. However, this is the exception. We advocate that the reconstruction is performed at a second procedure, delayed for 48–72 h to address the patient’s general condition as per damage control guidelines. Again, it is the pancreatic reconstruction that is difficult due to the normal pancreas and small duct. We make no recommendation as to the type of pancreatic reconstruction that should be employed. Our own practice is to perform a posterior pancreaticogastrostomy. Then a proximal gastrojejunostomy is fashioned followed about 10 cm distally with a side-to-side hepaticojejunostomy. There is some suggestion that the gallbladder can be used as a conduit for biliary reconstruction. We think this is unnecessary and unwise and recommend routine cholecystectomy. In addition, the bile duct is also usually not dilated and care and experience should be employed to prevent postoperative strictures of the bile duct anastomosis. Should the bile duct be small, mobilize the common bile duct proximally and identify the left hepatic duct. Open the anterior wall of the common hepatic duct onto the left hepatic duct for at least 2 cm. During this dissection, a small branch of the hepatic artery crosses the left hepatic duct and is often divided at this point. This vessel is hard to identify before you cut it and can be controlled with a small suture or diathermy. A note of caution at this point is important, and preservation of the blood supply to the biliary tree is very important. Excessive mobilization can devascularize the bile ducts resulting in late ischemic strictures. A side-to-side hepaticojejunostomy using a 4/0 or 5/0 monofilament absorbable, as originally described by Hepp and Couninard, is then performed. The bowel loop is brought antecolic based on the recent evidence that this reduces the incidence of delayed gastric emptying it is worth mentioning that if a narrow common bile duct is ligated during the initial operation, it will be dilated at the time of the reconstruction, making the Hepp and Couninard anastomosis, technically easier.