Overview

Torrential bleeding from the region of the pancreatoduodenal complex (PDC), surging like a crimson tide, is one of the most intimidating sights in trauma surgery. This particularly lethal pancreatic injury is but one of the challenges posed by this unforgiving organ. Hidden deep in the upper recesses of the retroperitoneum, the injured pancreas exhibits a spectrum of clinical presentations, from that of occult injury to dramatic exsanguination. This chapter reviews the modern approach to the management of pancreatoduodenal trauma with emphasis on the need for a high index of suspicion, modern imaging technology, and safe operative principles.

Evolution of the Approach to Pancreatic Trauma

The modern approach to pancreatic trauma reflects advances in elective foregut surgery. Better confidence in surgery of the PDC, advanced imaging, and better physiologic insight have combined to shape the current management of the injured pancreas.

A patient with blunt pancreatic injury was first reported in the Lancet as early as 1827, by Tavers. Although several survivors were subsequently reported after ligation and excision of exteriorized penetrating pancreatic wounds (Otis, 1876), the major advance in the operative treatment came in 1923, when Walton described distal pancreatectomy (see Chapter 62A, Chapter 62B ) as the treatment of choice for pancreatic transection. The dramatic reduction in mortality among patients with pancreatic injuries—from 56% in World War II, mostly due to associated injuries, to 22% in the Korean War—was more a result of improvements in general perioperative care than of specific technical innovations (Sako et al, 1955).

In the final decades of the last century, advances in medical imaging and in surgical technique converged to improve the care of patients with pancreatic trauma. The most important of these was computerized tomography (CT) of the abdomen (see Chapter 16), which for the first time allowed detailed visualization of the pancreas (Udekwu et al, 1996). This enabled surgeons to diagnose hitherto unrecognized pancreatic injuries and to drain peripancreatic fluid collections without the need for a laparotomy.

At approximately the same time, the introduction of endoscopic retrograde cholangiopancreatography (ERCP; see Chapter 18) provided a minimally invasive approach to diagnose leakage from the pancreatic duct after trauma (Buccimazza et al, 2006; Gupta et al, 2004; Lin et al, 2004; Wind et al, 1999; Wong et al, 2002). By contrast, the role of magnetic resonance cholangiopancreatography (MRCP; see Chapter 17) in the evaluation of the injured pancreatic duct remains to be defined (Bhasin et al, 2009; Fulcher et al, 2000).

The presumed adverse effect of pancreatic enzymes on the suture lines of duodenal repairs in patients with injuries to both organs led to the development of the Berne duodenal diverticulization, essentially a Bilroth II gastrectomy combined with biliary diversion (Berne et al, 1968, 1974), to preemptively divert gastric secretions around the repaired duodenum. Later, Jordan introduced the pyloric exclusion procedure (Vaughan et al, 1977), which achieves the same result but with a much simpler operation. Because of its conceptual elegance and technical simplicity, pyloric exclusion rapidly became the buzzword of PDC trauma in the 1990s. This led to very liberal use of the procedure in reported series (Feliciano et al, 1987); however, pyloric exclusion has never been subjected to a prospective randomized trial, and most attempts to compare it retrospectively to simple duodenal repair have demonstrated increased morbidity without benefit (DuBose et al, 2008; Seamon et al, 2007).

Resection of the entire PDC (see Chapter 62A, Chapter 62B ), which is routinely done in elective pancreatic surgery, is very rarely needed in the trauma situation. When complete disruption of the head of the pancreas and surrounding duodenal loop leaves the surgeon with no other choice, a so-called trauma Whipple (TW) may be necessary, but only as a last resort to save an exsanguinating patient. When necessary, the procedure is performed in a staged fashion with resection during the original operation and reconstruction 24 to 48 hours later (Subramanian et al, 2007). This is a reflection of the move toward damage control surgery, the most important paradigm shift in trauma surgery in the last two decades of the last century (Hoey & Schwab, 2002).

Distal pancreatectomy remains the mainstay of treatment of patients with major ductal disruption (Balasegaram & Lumpur, 1976; Heitsch et al, 1976); however, reports of successful stenting of the injured pancreatic duct via ERCP (Lin et al, 2006, 2007) suggest that this minimally invasive intervention may evolve into a safe management option in selected patients. Pancreatic trauma continues to tax the judgment and operative skills of experienced surgeons, even as sophisticated, minimally invasive diagnostic and therapeutic adjuncts continue to evolve.

Surgical Anatomy of the Pancreas and Duodenum: A Trauma Surgeon’s Perspective (See Chapters 1A and 1B)

The injured pancreas is not immediately accessible during laparotomy for trauma and requires special maneuvers for exposure and access. In addition, the complex pancreatic anatomy dictates that the head and the body of the pancreas be managed operatively as two separate organs.

The closest anatomic relationship of the injured body and tail of the pancreas is with the spleen. The pancreatic body is typically approached by first mobilizing the spleen medially to enter the anatomic plane behind the spleen and pancreas. The splenic artery and vein are then easily exposed and controlled on the posterior surface of the gland; however, if necessary, control of the splenic vessels can be achieved by mobilizing the body of the gland along its superior and inferior aspects, with no need for splenic mobilization. Based on this anatomic arrangement, trauma surgeons view the body and tail of the pancreas as an organ with a single accessible vascular pedicle, similar to the kidney or spleen. When necessary, the injured pancreatic body can be easily resected by performing an en bloc distal pancreatectomy and splenectomy. The major concern in trauma is whether the main pancreatic duct is damaged, because this can cause a persistent fistula with its associated complications. In contrast, bleeding is usually not a major concern.

The head of the pancreas is densely adherent to the duodenal loop, and it is convenient to group them together as the PDC, which is approached from the right side. The PDC is supplied by multiple vascular arcades with ample collaterals that are not easily controlled. When injured, the PDC is difficult to resect and requires a high-risk reconstruction. The major concern after complex injuries to the PDC is a leaking duodenal suture line.

From the trauma surgeon’s perspective, most of the major vascular structures of the upper abdomen converge on a small area, not much larger than a silver dollar, around the PDC. The superior mesenteric vessels, portal vein, inferior vena cava, and hilum of the right kidney are all in close proximity to the pancreatic head. The vascular structures around the PDC are arranged in three overlapping layers. The deepest one consists of the inferior vena cava and right renal pedicle. Above it, just behind and below the gland, are the portal vein and superior mesenteric vessels. The uppermost layer consists of the gastroduodenal artery and pancreaticoduodenal arcades.

As a result of this unique multilayered anatomic arrangement, penetrating trauma to the PDC often involves associated injuries to more than one major vascular structure, some of which—for example, the retropancreatic portal vein—are not immediately accessible from the front. It is therefore not surprising that penetrating trauma to this area presents what is widely held by trauma surgeons to be the most formidable challenge in abdominal trauma.

The organ injury scales for the pancreas and duodenum, developed by the American Association for the Surgery of Trauma (AAST) (Tables 103.1 and 103.2; Moore et al, 1990) are used for grading injuries of these two organs. Unfortunately, these scales do not acknowledge the fundamental anatomic differences between the pancreatic body and tail and the PDC, and therefore their clinical usefulness is limited.

For practical purposes, trauma to the pancreas ranges from superficial injuries that require only drainage (Feliciano, 1989) to ductal disruption that mandates a distal pancreatectomy. Low-grade injuries to the duodenum can be repaired the same as any other penetrating injury to the gut; however, high-grade duodenal injuries, and especially those that combine duodenal and pancreatic trauma, require a different approach because of the high risk and unforgiving nature of a duodenal leak. Injuries that destroy the PDC or disrupt the ampulla are rare and carry very high mortality rates because of associated trauma to adjacent major vascular structures. This perspective on pancreatic anatomy forms the basis for a detailed discussion of the management of pancreatic injuries in the forthcoming sections.

Operative Management of Injuries to the Body and Tail of the Pancreas (See Chapter 62A, Chapter 62B )

The operative approach to injuries to the body and tail of the pancreas is similar for both blunt and penetrating injuries. Full exposure is achieved by either dividing the gastrocolic ligament or by freeing the greater omentum from the transverse colon, reflecting the stomach up and taking down any adhesions between the posterior wall of the stomach and the retroperitoneum. The organ is then thoroughly inspected to assess the damage. Any area of injury, contusion, or hematoma is explored by opening the posterior peritoneum over the anterior aspect of the pancreas, because a superficial and seemingly innocent hematoma may hide a deep laceration. Severe bleeding almost never originates from the pancreas itself but rather from adjacent structures.

In assessing the body and tail of the pancreas, the key question that the surgeon must address is whether a pancreatic duct injury is present (Bach & Frey, 1971; Bradley et al, 1998; Tyburski et al, 2001; Voeller et al, 1991). Unless there is an obvious complete transection, it is difficult to diagnose a duct injury by simple inspection. The integrity of the duct can be interrogated by intraoperative imaging, either by amputating the tail and cannulating the distal duct or by cannulating the gallbladder and injecting contrast into the common bile duct. Both these options are cumbersome and do not have high success rates (Berni et al, 1982).

A transduodenal pancreatogram is mentioned here only to be condemned, because it converts a pancreatic injury into a pancreaticoduodenal (PD) injury. Intraoperative ERCP, first reported in 1986, is another option, if it can be organized expeditiously (Laraja et al, 1986). The more practic option in the middle of the night is to simply proceed with empiric distal pancreatectomy, whenever a deep laceration is in the vicinity of the main pancreatic duct, or a large hematoma is found in the center of the gland. An alternative option for nonbleeding pancreatic injuries, particularly in the face of multiple associated injuries, is closed suction drainage of the lesser sac. Patton and colleagues (1997) demonstrated excellent results by simple drainage, regardless of the location of the injury. A disruption of the capsule should be drained without attempting to repair it. Drainage of an injured main pancreatic duct is an effective solution, but it often results in a high output pancreatic fistula, so distal pancreatectomy may be a better option.

The trauma literature of the 1970s and 1980s is replete with descriptions of pancreas-preserving techniques for complete pancreatic transections (Frey, 1982). The most popular of these was anastomosis of the distal pancreatic stump to a Roux-en-Y loop of jejunum to create a pancreatojejunostomy. The proximal stump, which drains into the duodenum, was then oversewn. The rationale for pancreas-preserving approaches is to maintain pancreatic function and to prevent potential long-term complications of exocrine insufficiency and/or diabetes mellitus; however, this elegant technical solution involves a high-risk jejunal anastomosis to a traumatized edematous pancreatic stump in an unstable patient who will not tolerate a leak. So in spite of the fact that it is sometimes technically feasible, preservation of the distal pancreatic stump is not in the best interests of the trauma patient; distal pancreatectomy is a much safer solution (Krige et al, 2005; Steele et al, 1973).

Postoperative Pancreatic Fistula

Approximately 25% of patients develop a fistula after distal pancreatectomy for trauma (Fahy et al, 2002). In most cases, the fistula is controlled by the drains placed at the time of surgery. Neither the technique of pancreatic closure (Ferrone et al, 2008; Sheehan et al, 2002) nor the use of sealants (Fisher et al, 2008; Suc et al, 2003; Yamamoto et al, 2009) significantly affects the risk of fistula. Nutritional support, preferably via jejunal feeding, is a crucial element of fistula management; however, the presence of a pancreatic fistula does not necessarily require abstinence from oral intake; indeed, depending on the nature and extent of associated injuries, most patients will be able to eat normally during the time it takes for the fistula to close. Endoscopically placed pancreatic stents can reduce the volume and duration of fistula drainage in problem cases (Goasguen et al, 2009; Grobmyer et al, 2009). Failure to identify and drain a contained leak can result in a pancreatic pseudocyst, which may on rare occasion require surgery.

Injuries in the Stable Patient

The first step in the operative management of the injured PDC is wide exposure of the area with a Kocher maneuver. Full mobilization of the PDC from the common bile duct to the superior mesenteric vein (SMV) exposes the head of the pancreas, the second and some of the third portions of the duodenum, and the underlying vena cava. Additional exposure of adjacent structures and an improved workspace can be developed by extending the incision in the posterior peritoneum caudad along the white line of Toldt and mobilizing the right colon (Buscaglia et al, 1969), the so-called extended Kocher maneuver. Complete visualization of the entire third and fourth portions of the duodenum is achieved by carrying the peritoneal incision around the cecum and along the line of fusion of the small bowel mesentery to the posterior peritoneum, all the way up to the ligament of Treitz. This Cattell-Braasch maneuver is a technique for the exposure of the third and fourth portions of the duodenum (Cattell & Braasch, 1960), which allows complete reflection of the midgut onto the chest wall and provides access to the entire duodenum; however, after performing the maneuver, the entire midgut remains suspended on the superior mesenteric vascular pedicle and should be handled with care.

After complete mobilization, the entire PDC must be thoroughly explored. The trajectory of the wounding missile must be reconstructed to avoid missing an injury, and the number of holes in the gut should be even; an uneven number should prompt a renewed search for another injury. It is especially easy to miss a hole on the medial side of the duodenal loop adjacent to the pancreas. In these cases, the safest option is to enlarge the hole on the free border of the duodenum in a transverse fashion to inspect and repair the medial wall from inside.

Low-grade duodenal injuries are repaired transversely when possible, to avoid narrowing the lumen. A well-vascularized tension-free repair is essential, but the specific technique and suture material are a matter of individual preference. Decompressive tube duodenostomy is unnecessary, but any injury to the PDC, no matter how simple, should be drained. If the duodenum cannot be closed without tension, it is not a low-grade injury.

High-grade duodenal injuries require sound judgment and safe technical solutions. An extensive laceration or loss of a segment of duodenal wall should never be closed under tension. Instead, the simple and versatile solution is a Roux-en-Y jejunal loop anastomosed to the injured portion of the duodenum.

All cases of complex duodenal repair benefit from establishing a route for enteral feeding distal to the duodenal suture line. Despite being recognized as a key principle, it is often forgotten in the excitement of the moment. Either a nasojejunal tube or a feeding jejunostomy will enable early postoperative feeding and greatly facilitates postoperative management if the duodenal repair leaks.

High-grade PDC injuries are an indication for pyloric exclusion (Degiannis & Boffard, 2000), the goal of which is to preemptively divert the gastric effluent around the injured duodenum by temporarily occluding the pylorus. If the duodenal repair leaks, a low-output end-duodenal fistula is much easier to manage than a high-output side fistula.

Pyloric exclusion begins with a generous longitudinal gastrotomy in the anterior wall of the antrum, close to the pylorus. The pylorus is grasped with two Babcock clamps, delivered through the gastrotomy and oversewn with size 0 PDS sutures. A side-to-side gastrojejunostomy with a short afferent limb is then constructed at the gastotomy site (Fig. 103.1). Some surgeons prefer to staple across the duodenum just distal to the pylorus, instead of suturing it from the inside. In most cases, the duodenum spontaneously opens several weeks later regardless of the method of closure; we prefer to use an absorbable synthetic suture.

FIGURE 103.1 Pyloric exclusion showing occlusion of the duodenum just distal to the pylorus followed by a gastrojejunostomy.

The gastrojejunostomy is the source of most postoperative morbidity after a pyloric exclusion. This led Ginzburg and colleagues (1997) to attempt pyloric exclusion without gastrojejunostomy, relying on the spontaneous reopening of the sutured pylorus. Pyloric exclusion is, in theory, an ulcerogenic operation, but the incidence of marginal ulcers is unknown (Buck et al, 1992; Feliciano et al, 1987; Martin et al, 1983). Routine vagotomy or delayed takedown of the gastrojejunostomy is not indicated. Routine pyloric exclusion has recently been criticized because of reported increased complications and no effect on outcome (Seamon et al, 2007); however, in our practice, we continue to perform pyloric exclusion in patients with tenuous duodenal closures and in those with high-grade pancreatic injuries associated with a duodenal injury of any grade.

Occasionally, a high-grade injury to the PDC is encountered during a trauma laparotomy with massive blood loss unrelated to the PDC. In this situation, a rapid damage control solution is the only feasible alternative (see Chapter 102). The damage-control option for the injured nonbleeding pancreas is external drainage, regardless of the grade of injury. Similarly, the “bail-out” solution for duodenal injuries is either rapid closure with external drainage or placement of a large Malecot catheter through the hole affixed in place with a purse-string suture to create a controlled fistula.

Management of A Leaking Duodenal Repair

A leak from the duodenal suture line after a PDC injury is unusual but can be devastating (Seamon et al, 2007). Leakage of bile from the drain in a Morison pouch, or any deviation from the patient’s expected clinical trajectory, especially signs of unexplained sepsis, mean a duodenal leak until proven otherwise. A CT scan with contrast in the gastrointestinal tract will demonstrate the leak and establish whether it is controlled. If the leak is uncontrolled—that is, if free fluid is present in the peritoneal cavity—an urgent laparotomy is indicated to achieve source control and effective drainage. Such leaks can be difficult to manage, especially if the pylorus is open; the duodenum cannot be resected or exteriorized. Proximal diversion, if not done during the original procedure, is rarely an option; in addition, the inflamed duodenal wall does not hold sutures well. Inserting an appropriately sized Malecot tube into the hole, combined with wide regional drainage, is often the only feasible option. A creative variation of tube drainage is insertion of a spiral three-channel tube for continuous intraluminal irrigation (Parc et al, 1999). Although we have no personal experience with this method, it nevertheless underscores the technical difficulty of controlling a duodenal leak.

Late discovery of a duodenal leak in a septic patient is an indication for open-abdomen management with planned reoperations for peritoneal toilet to optimize drainage; this is not the time for complex duodenal reconstructions. If the leak is discovered early, pyloric exclusion may perhaps still be feasible, but in most cases, it is a hazardous undertaking in the inflamed peritoneal cavity of a septic patient.

If the CT scan demonstrates a leak but not collection or free fluid in the peritoneum, the leak is controlled. The appropriate conservative approach in this situation hinges on early and aggressive nutritional support while awaiting spontaneous closure. If a pyloric exclusion was done, fluid management of the low-output fistula is usually straightforward. Most duodenal leaks close spontaneously; if the leak does not close, it is important to rule out distal obstruction. If the leak persists, but no distal obstruction is found, the approach is to wait at least 6 weeks or more and then perform a Roux-en-Y duodenojejunostomy to the fibrotic edges of the hole in the duodenum.

A complex PDC injury in the second portion of the duodenum resulting in destruction of the ampulla of Vater and/or the ductal systems draining both the biliary tree and the pancreas requires a pancreatoduodenectomy (PD). This situation is very rare—neither of us has encountered it in a hemodynamically stable patient. The technical details of a TW are discussed in the next section.

Pancreatoduodenal Injuries in the Unstable Patient

Severe bleeding from the area around the PDC is seen either immediately upon entering the peritoneal cavity, or it begins suddenly, when a periduodenal hematoma is explored. In either case, the surgeon must be prepared technically and mentally to deal with a major vascular injury, often more than one.

The first step is temporary packing of the bleeding area, until the rest of the peritoneal cavity is explored; however, each time the packs are removed, the operative field is immediately immersed in blood, which precludes accurate visualization of the injuries. The complex multilayered vascular anatomy around the PDC does not allow simple proximal and distal control of the injured vessels. The solution is to encircle the injured area with a ring of tightly rolled laparotomy pads held on sponge-holding forceps and firmly pressed downward by assistants. This “compressive ring” of pads occludes the major veins around the PDC, giving the surgeon an opportunity to identify and repair the injured vessels.

Injuries to the retroduodenal inferior vena cava (IVC) are typically associated with trauma to the PDC and sometimes also to the right renal pedicle. The key to rapid exposure is a complete Kocher maneuver. Once the IVC is exposed, bleeding is controlled by compressing it above and below the bleeding site. The pararenal segment of the IVC is especially difficult to access and control. Two helpful maneuvers to improve access are application of assertive retraction to the liver in a cranial direction, using a deep Deaver retractor, and rapid mobilization of the right kidney. Simple lacerations of the IVC are repaired using a running monofilament suture. A hole in the posterior wall must occasionally be repaired from within the vessel by enlarging the hole on the anterior wall of a through-and-through injury.

It is important to remember that complex IVC injuries can be treated with ligation. Lacerations of the pararenal segment are especially difficult to repair, even if both renal veins are first isolated and controlled, because bleeding continues from lumbar veins, and ligation of the pararenal IVC means loss of at least the right kidney. Unfortunately, in the massively bleeding patient with several sources of hemorrhage, this may be the only realistic option to save the patient’s life. Injuries to the renal hilum in the context of penetrating trauma to the PDC are best addressed by a rapid right nephrectomy. Rapid mobilization of the right kidney out of the Gerota fascia and medially also allows the surgeon to gain access to the lateral and posterior aspect of the pararenal IVC.

Exsanguinating hemorrhage into the lesser sac is most often due to an injury to the peripancreatic portal vein and is almost always associated with a penetrating injury to the neck of the pancreas. Effective circular compression as described above will dramatically reduce the bleeding, allowing the surgeon to perform a Kocher maneuver and widely open the lesser sac (Fig. 103.2).

FIGURE 103.2 Sponge-stick control of bleeding from the portal vein. The white circles indicate areas of compression.

One of the authors (A.H.) always performs a Cattell-Braasch maneuver in these situations to gain maximal exposure of the infrarenal retroperitoneum. Careful dissection along the inferior border of the mobilized duodenal loop is carried all the way to the point at which the SMV dives behind the pancreatic neck. The portal vein is then identified in the hepatoduodenal ligament along the upper border of the first portion of the duodenum. Bluntly developing the plane between the common bile duct and the portal vein from above allows the surgeon to pass a finger behind the pancreatic neck, divide it, and repair the portal vein directly from the front (Fig. 103.3). An early decision to transect the pancreas limits blood loss. A distal pancreatectomy is performed after the vascular repair, and the proximal stump is oversewn. In a damage-control situation, the SMV and the portal vein may be ligated (Sheldon, 1985); however, repair is preferable whenever possible, because massive bowel edema inevitably ensues that complicates fluid and wound management. The early fluid requirements during the first 24 hours after ligation of the SMV or portal vein can exceed 20 L. A second-look laparotomy is mandatory approximately 24 hours after the original operation to assess bowel viability.

FIGURE 103.3 A, Division of pancreas to expose portal vein. B, Portal vein injury with distal pancreatectomy and splenic preservation.

These vascular injuries usually occur in the context of multiple injuries to the PDC and adjacent structures. Damage-control solutions, rather than definitive repairs, are the rule in these situations, because the exsanguinating patient is racing toward a lethal physiologic insult. It is far better to stop the bleeding, close the holes in the bowel, and drain the entire area than to attempt a complex visceral reconstruction in an unstable, coagulopathic patient. Reconstruction of the anatomy can safely be done 36 to 48 hours after completion of resuscitation in the ICU.

Conclusion

Injuries to the pancreas and duodenum are uncommon. Most pancreatic injuries do not involve the pancreatic duct and can be managed by simple drainage. Ductal injuries to the body and tail of the pancreas should be treated by distal pancreatectomy. The emerging role of endoscopically placed stents in the treatment of pancreatic ductal injuries remains to be defined; however, isolated ductal injuries in the head of the pancreas are not an indication for a trauma Whipple, and these should be drained with or without ductal stenting.

Most injuries to the duodenum can be treated with simple closure. A preemptive pyloric exclusion should be considered in patients with PDC injuries at high-risk for a duodenal leak. Penetrating injuries to the medial wall of the second duodenum are best repaired from within the duodenum through an incision on the lateral duodenal wall. Complex injuries to the duodenum are best repaired with a Roux-en-Y duodenojejunostomy.

Major disruption of the PDC is very rare. Even a busy trauma surgeon can expect to see only a handful of these injuries in his or her career. In the overwhelming majority of cases, simple damage-control solutions are a much better option than heroic operations that involve long and elaborate resections; therefore, a TW should be considered only as a last, desperate resort.