Introduction

Damage control laparotomy (DCL) has entered the mainstream surgical lexicon in modern trauma care. It represents one of the major advances in the surgical management of the severely injured patient. Stone described in 1983 the survival advantage of the early truncation of laparotomy when encountering exsanguinating traumatic hemorrhage to allow for physiologic resuscitation prior to definitive operative repairs 1. Rotondo applied the term “damage control laparotomy” to this approach as applied to the severely injured trauma patient, borrowing the vocabulary of the Naval service and the shipboard principle of maintaining the vital capacities of the ship and containing damage in the case of attack in order to maintain mission integrity and stay afloat.2 DCL can be a lifesaver for the trauma patient in the operating room who develops hypothermia, acidosis, and coagulopathy. Operative time is minimized, with attention to control of hemorrhage and containment of gastrointestinal contamination, employing stapled bowel closure, external tube drainage, abdominal packing, and rapid abdominal closure.1–4 This strategy, which has evolved in major trauma centers over the past two decades, is as applicable to patients who suffer operating room trauma as to those who suffer trauma on the streets.

It is alarming when patients undergoing elective and urgent abdominal surgery develop severe physiologic disturbances secondary to operative hemorrhage or sepsis. The accompanying intraoperative hypothermia, acidosis, and coagulopathy are harbingers of high morbidity and mortality. DCL principles are utilized successfully in general surgery patients in many situations. Elective and urgent pancreatic operations in particular have occasional need for utilization of DCL to salvage patients who are unexpectedly severely injured with operative trauma. In order to review and describe the use of DCL in operations on the pancreas, a retrospective review and analysis was undertaken of patients undergoing pancreatic surgery over a 6-year period.

Methods

An inpatient hospital database identified a cohort of 835 patients who underwent elective and urgent pancreatic operations at the Medical University of South Carolina from 2001 to 2007. Eight patients were identified who required DCL. With the approval of the Institutional Review Board for the evaluation of human subjects, the electronic medical records of these eight patients were reviewed. Data collected and analyzed included patient demographics, indications for surgery, operative conduct, intraoperative blood loss, pH, patient temperature and INR, damage control techniques, timing of definitive operation, and hospital outcome.

Results

There were five men and three women with a mean age of 51 years (range 37 to 65 years) who underwent an elective pancreas operation and subsequently required DCL during the study time period. The indication for surgery was chronic pancreatitis in six patients, chronic pancreatitis and intrapapillary mucinous neoplasm in one, and ampullary cancer in one. The index operation was pancreatoduodenectomy in four patients, distal pancreatectomy in three, and total pancreatectomy in one (Table 1).

Table 1 Demographics
Full size table

In four patients undergoing elective pancreatic resection intraoperative portal vein hemorrhage initiated damage control laparotomy. The site of portal vein injury was at the portosplenic confluence in two patients undergoing distal pancreatectomy, anteriorly under the neck of the pancreas in one patient undergoing total pancreatectomy, and laterally behind the common bile duct in one patient undergoing pancreatoduodenectomy.

While the injuries encountered were in variable locations, several common techniques were employed in the management of portal venous hemorrhage. Initial venous compression for temporary control was undertaken to facilitate identification of injury site and extent, as well as to allow for preparation of the operative team, including anesthesiologist and surgeon. Compression was accomplished via a Kocher maneuver for manual pressure in all cases. Spongesticks were also used for means of venous compression. Next adequate exposure of the injury was obtained via division of the pancreas and expeditious removal of the specimen.

In all patients primary repair of the identified portal venous injury was attempted. In two patients primary repair was successful. In one patient with a portosplenic confluence injury, temporary vascular control was obtained with vascular clamps. Due to the size of the injury, primary repair risked venous compromise and he therefore underwent venous repair using a patch of native splenic vein (harvested from the specimen). The final patient with portal injury had his portal vein ligated during attempts at hemostasis. The small bowel became immediately severely engorged and ischemic appearing. Due to concerns for bowel viability, an interposition graft of cadaver iliac vein was utilized at this primary operation. While classic DCL principles would entail temporizing measures such as vascular shunting with planned staged definitive repair, these latter two patients were managed well with revascularization at the primary operation (Table 2).

Table 2 Intraoperative Data
Full size table

Four patients had damage control utilized at reoperation for abdominal sepsis (two) and hemorrhage (two). Abdominal sepsis was due to a leak at the choledochojejunostomy in one patient. She presented with fever, hypotension, and extraluminal fluid and air on CT scan postoperative day 9, and on exploration was found to have anastomotic dehiscence. Another patient presented 53 months after pancreatoduodenectomy with obstructive symptoms and hemodynamic instability and was found to have an internal hernia with bowel infarction. As this patient demonstrates, the altered anatomy of the post pancreatectomy patient can lend to catastrophic complication. One patient developed hypotension, tachycardia, and profound anemia on postoperative day 1; bleeding was found in the bed of the resected pancreatic head on re-exploration. Another patient, on postoperative day 3, similarly became hemodynamically unstable and anemic and was found at reoperation to have bleeding from the splenic artery stump after distal pancreatectomy (Table 2).

DCL techniques included external tube drainage of biliary, pancreatic, or enteric contents (five), staple closure of open bowel (four), abdominal packing (seven), and rapid abdominal closure (seven; Table 2). External tube drainage consisted of the use of a red rubber catheter for control of biliary contents and a pediatric feeding tube for diversion of pancreatic exocrine secretions. Closed suction drains were placed in the operative bed and octreotide infusion was employed to decrease pancreatic exocrine output. Rapid abdominal closure employed a protective covering over the bowel with the application of a negative pressure dressing.

Operative blood loss ranged from 300 to 12,000 cc. Operative transfusions ranged from 0 to 44 U of packed red cells (mean 14.1 U). Intraoperative INR was a mean of 2.19 (range 1.43–3.3) and was greater than 1.5 in seven out of eight patients. Intraoperative pH ranged from 7.08 to 7.45 (mean 7.27), and temperature ranged from 34.8 to 38.8°C (mean 36.1; Table 3).

Table 3 Intraoperative Data Continued
Full size table

Laparotomy for pack removal and intestinal reconstruction was undertaken 1 to 3 days after DCL in seven of the patients. Three patients underwent more than one additional reoperation (range 0 to 6). One patient underwent external tube drainage of her bile duct with suture closure of the adjacent jejunum after complete dehiscence of her choledochojejunostomy. Reoperation for anastomotic revision was planned, but a fistula developed between the bile duct and jejunum, which was subsequently able to be dilated by the percutaneous transhepatic route in interventional radiology and she therefore avoided reoperation.

Six patients achieved fascial closure, while two underwent vicryl mesh closure. Perioperative morbidity was substantial, including intra-abdominal abscess (six), respiratory failure (five), pneumonia (two), acute renal failure (two), venous thromboembolism (two), cardiac arrhythmia (one), pancreatic fistula (one), and enterocutaneous fistula (one). Three patients developed ventral hernias requiring delayed reconstruction. Length of hospital stay ranged from 7 to 80 days. Five patients required hospital readmission. Hospital mortality and long-term mortality (median follow-up 58 months) were zero. Hospital charges averaged $275,627 per patient (range from $32,363 to $931,723)

Discussion

Elective pancreas surgery in the modern era is safe and effective when performed at high-volume centers. Despite the advantages which specialized centers offer, when surgeons work in and around the pancreas, the potential for catastrophic complication persists. Pancreatic cancer and chronic pancreatitis patients are often chronically malnourished, and may have resultant poor healing capabilities and limited physiologic reserve related to age and chronic illness.

The dense inflammatory fibrosis associated with chronic pancreatitis obliterates planes and distorts anatomy increasing risk for vascular injury. Pancreatic cancer can involve nearby major vascular structures requiring complex vascular reconstruction. Exsanguinating hemorrhage can quickly develop, particularly in the event of portal venous injury, which was the most common complication encountered in this series.

Postoperative pancreatic and biliary anastomotic leaks can lead to a noxious broth of pancreatic exocrine secretions and bile with consequent pancreatic enzyme activation with attendant autolysis of surrounding soft tissues. Activation of pancreatic enzymes is a potent stimulus for cytokine release and the vicious cascade of a severe systemic inflammatory response can result. This postoperative occurrence leads directly to renal, pulmonary, cardiovascular, hepatic, and central nervous system failure escalating the severity of illness. Additionally, severe hemorrhagic complications can occur as a result of enzymatic vessel destruction. The anastomotic leak after pancreatic surgery can be catastrophic.

Patients that have undergone pancreatic surgery with altered foregut anatomy are at risk for the development of small bowel obstruction and internal herniation. Afferent limb obstruction after pancreatoduodenectomy, efferent limb obstruction due to jejunal fixation for feeding tubes, and internal herniation of proximal jejunum at the base of the transverse mesocolon are potential causes of small bowel obstruction. These proximal obstructions are difficult to differentiate from postoperative pain and pain related to chronic pancreatitis, and history and physical examination will need to be supplemented by CT imaging.

In trauma, lag time leads to loss of physiologic reserve due to hemorrhage and visceral contamination. Time delay from injury to treatment is a large determinant of survival. In the 1980s, the recognition of the role of physiologic reserve in outcomes in critically injured trauma patients led to the emergence of damage control surgery. Once the deadly combination of hypothermia, acidosis, and coagulopathy has been incited in the setting of trauma, these factors become the priority and must be corrected, often delaying definitive operative injury repair.

In patients undergoing elective pancreas surgery, the physiologic disadvantage is usually not due to delay in time to operation. These patients are physiologically disadvantaged due to their underlying disease process and limited reserve. Portal venous hemorrhage and pancreatobiliary leak are significant physiologic stressors that can quickly lead to hypothermia, acidosis, and coagulopathy. As in trauma, damage control principles, with truncated laparotomy, stapled closure of bowel, external tube drainage, abdominal packing, and rapid abdominal closure with planned re-exploration after physiologic resuscitation are no different than principles used in trauma surgery.

All patients in the study group developed hypothermia, acidosis, and/or coagulopathy during or following pancreas surgery. Damage control techniques were successfully utilized with attention to controlling hemorrhage, maintaining intravascular volume, and draining pancreatobiliary secretions. Truncated laparotomy was employed to allow for correction of abnormal physiology with planned delayed definitive repair. While significant resource utilization was involved with multiple reoperations, intensive care therapy, and long hospital stay, there was no mortality.

Conclusions

Operations on the pancreas are known to be a risky undertaking with infrequent but well-known intraoperative and postoperative catastrophic complications. When pancreatic operations result in hemorrhage and sepsis with hypothermia, acidosis, and coagulopathy, DCL principles should be utilized. DCL can be life saving and should be an essential tool in the armamentarium of the modern pancreas surgeon.