Safety of Repair for Severe Duodenal Injuries

Transcription

1 World J Surg (2008) 32:7 12 DOI /s Safety of Repair for Severe Duodenal Injuries George C. Velmahos Æ Constantinos Constantinou Æ George Kasotakis Published online: 22 October 2007 Ó Société Internationale de Chirurgie 2007 Abstract Background There is ongoing debate about the management of severe duodenal injuries (SDIs), and earlier studies have recommended pyloric exclusion. The objective of this study was to compare primary repair with pyloric exclusion to examine if primary repair can be safely used in SDIs. Methods The medical records of 193 consecutive patients who were admitted between August 1992 and January 2004 with duodenal injuries were reviewed. After excluding early deaths (n = 50), low-grade duodenal injuries (n = 81), and pancreatoduodenectomies for catastrophic trauma (n = 12), a total of 50 patients with SDIs (grade III, IV, or V) were analyzed. Primary repair (PR simple duodenorrhaphy or resection and primary anastomosis) was performed in 34 (68%) and pyloric exclusion (PE) in 16 G. C. Velmahos (&) C. Constantinou G. Kasotakis Division of Trauma and Critical Care, Department of Surgery, Los Angeles County and University of Southern California Medical Center, Los Angeles, California, USA gvelmahos@partners.org Present Address: G. C. Velmahos Department of Surgery, Massachusetts General Hospital, Harvard Medical School, 165 Cambridge Street, Suite 810, Boston, MA 02114, USA Present Address: C. Constantinou Department of Surgery, Geisinger Medical Center, Danville, PA, USA Present Address: G. Kasotakis Department of Surgery, Creighton University Medical Center, Omaha, NE, USA (32%). Characteristics and outcomes of these two groups were compared. Results PE and PR patients were similar for age, injury severity score, abdominal abbreviated injury score, physiologic status on admission, time to operation, and most abdominal organs injured. PE patients had more pancreatic injuries (63% vs. 24%, p \ 0.01), a higher frequency of injuries to the first and second part of the duodenum (79% vs. 42%, p = 0.02), and a nonsignificant trend toward more grade IV and V injuries (37% vs. 18%, p = 0.11). There was no difference in morbidity (including complications specific to the duodenal repair), mortality, and intensive care unit and hospital length of stay between the two groups. Conclusions Pyloric exclusion is not necessary for all patients with SDIs, as previously suggested. Selected SDI patients can be safely managed by simple primary repair. Level I evidence on the management of duodenal trauma is lacking, predominantly because this type of injury is infrequent. Influenced by the literature describing surgical techniques in nontraumatic duodenal perforations [1], surgeons have used similar techniques for duodenal injuries, following penetrating or blunt trauma. Although there is consensus that straightforward primary repair (PR) is adequate for simple traumatic perforations, the management of more complex injuries is controversial [2 7]. Pyloric exclusion (PE) has emerged as a simplified version of duodenal diverticulization [8] to manage injuries that are deemed to be at high risk for suture line dehiscence. Such severe duodenal injuries (SDIs) are classified as grade III, IV, and V by the American

2 8 World J Surg (2008) 32:7 12 Association for the Surgery of Trauma (Table 1) [9]. Catastrophic injuries of the pancreatoduodenal complex may require pancreatoduodenectomy and are frequently associated with poor outcomes [10 14]. PE is used as the intermediate solution for SDI that can be repaired without extensive resections but are still extensive enough for simple PR. However, PE is associated with prolonged operative times in patients who usually have multiple organs injured and are physiologically labile [15, 16]. The PE technique is far from standardized, [17 19] the rate and timing of spontaneous opening of the duodenum following closure is unknown, [16, 20] and the short- and long-term postoperative complications are poorly understood [7, 16, 21]. Although previous studies [7, 11, 16, 21 24] have recommended PE for SDIs, our ongoing experience with simple PR for many grade III, IV, and V injuries has been positive. This report analyzes our experience with SDI and determines if PR is a valid option for definitive operative repair without risking a high rate of complications. Methods A total of 193 patients with duodenal injury and complete medical records admitted between August 1992 and January 2004 were identified. Fifty patients who died within 5 days of injury (1 111 hours) were excluded because the outcome of the duodenal repair could not be evaluated. Of the 143 remaining patients, 62 (43%) had grade III, IV, or V duodenal injuries. Of them, 12 patients who underwent pancreatoduodenectomy for catastrophic injuries of the Table 1 Grades of severity of duodenal injury Grade a Injury description Severity I Hematoma Involving single portion of duodenum Laceration Partial thickness, no perforation II Hematoma Involving more than one portion Laceration Disruption \50% of the circumference III Laceration Disruption 50 75% of circumference of D2 Disruption % of circumference of D1, D3, D4 IV Laceration Disruption [75% circumference of D2 Involving ampulla of distal common bile duct V Laceration Massive destruction of duodenopancreatic complex Vascular Devascularization of duodenum a According to the American Association for the Surgery of Trauma [9] pancreatoduodenal complex were excluded because this type of injury was not the focus of our study. Patients with grade V injuries who did not require pancreatoduodenectomy were included. The remaining 50 patients comprise the study population and were fully analyzed. Two groups were identified: (1) the PR group, with patients who underwent either duodenorrhaphy or resection and primary anastomosis; and (2) the PE group, with patients who underwent repair of the duodenal laceration, closure of the pylorus, and gastrojejunostomy. Closure of the pylorus was performed by a variety of techniques, including TA-stapling or suturing with absorbable or nonabsorbable sutures via gastrotomy. The gastrojejunostomy was performed as a hand-sewn, double-layer anastomosis in all cases. The stoma was created in a dependent portion of the stomach, usually at the site of gastrotomy in a retrocolic position. The repair of the duodenal laceration was hand-sewn in two layers with a variety of suture materials. A similar technique was used after resection and primary anastomosis. Wide drainage of the area was routinely added. Gastrostomy and feeding jejunostomy were added on occasion. Reverse tube duodenostomies were not performed except in 3 cases. Prophylactic octreotide was administered depending on the surgeon s preference. Postoperative long-term endoscopies were not performed routinely in patients with PE to identify the status of the pyloric opening. Data were collected regarding patient characteristics, type of injury (blunt or penetrating), injury severity [defined by the injury severity score (ISS) and abdominal abbreviated injury score (AIS)], presence of shock (systolic blood pressure\90 mmhg) on admission, extraabdominal injuries, operative findings including details about the grade and site of duodenal injury and other intraabdominal organs injured, surgical interventions, and outcomes. The duodenal injury was graded according to the Organ Injury Scaling Committee of the American Association for the Surgery of Trauma (Table 1) [9]. The following were recorded as outcomes. Complications related to the duodenal repair, including the following. Suture line leaks, as shown during reoperation or by drainage from intraoperatively placed drains or postoperatively placed computed tomography (CT)-guided percutaneous drains Duodenal stenosis or obstruction at the suture line Gastrojejunal anastomotic dehiscence, stenosis, obstruction, or marginal ulceration Bleeding at any suture repair site (duodenal, pyloric, gastrojejunal) Any other complication deemed to be related to the duodenal repair technique

3 World J Surg (2008) 32: Other intraperitoneal complications (including abscess, bleeding, or obstruction) not directly attributable to the duodenal repair. Extraperitoneal complications, including surgical wound infection, bullet tract infection, necrotizing soft tissue infection, pulmonary infections, systemic sepsis, and venous thromboembolism, as well the complications listed above, are reported together here as all postoperative complications ). Intensive care unit and hospital lengths of stay. Mortality. The two groups were compared for all group characteristics and outcomes. Subanalyses were performed according to the duodenal grade of injury because grade IV and V injuries may be more likely to be managed by PE, whereas grade III injuries are managed by PR. Continuous variables were compared by t-test and categorical variables by the chi-squared or Fisher s exact test. Statistical significance was assigned at p \ Results Of the 50 eligible patients, 43 (86%) were male, 42 (84%) were injured by a penetrating mechanism (38 gunshot wounds, 4 stab wounds), and only 2 (4%) had isolated SDIs. The mean age was 26 ± 9 years, ISS 22 ± 8.5, and abdominal AIS 4.1 ± 0.8. There were 38 grade III injuries (76%) and 12 grade IV and V injuries (24%). A total of 34 patients (68%) underwent PR (23 duodenorraphy, 11 resection with primary anastomosis), and 16 (32%) were subjected to PE. One patient died (2%) on postoperative day (POD) 6 owing to sepsis unrelated to the duodenal repair. Table 2 shows the comparison of the two groups. PE and PR patients were similar for general and abdominal injury severity as well as associated intra- and extraabdominal organs injured, except the pancreas. PR patients had a higher frequency of injuries in the first two portions of the duodenum (79% vs. 42%, p = 0.02) and a nonsignificant trend toward more grade IV and V injuries (37% vs. 18%, p = 0.11). None of the outcomes were different in the two groups. Six (18%) PR and four (25%) PE patients developed complications deemed to be related to the duodenal procedure (p = 0.4). These patients are presented in Table 3. All 10 had duodenal leaks, and one of them developed a marginal ulcer at the gastrojejunostomy site. No cases of bleeding or obstruction at the repair site were recorded. Of those with duodenal leaks, only two required reoperation. In the remaining eight patients, the leaks were controlled by intraoperatively placed drains (placed at the initial procedure) or CT-guided drains placed postoperatively; the leaks healed without further operation. Table 2 Comparison of patients managed by primary repair (PR) and patients managed by pyloric exclusion (PE) Parameter Specifically among patients with associated pancreatic injury, the duodenal leak rate was 30% (3/10) in the PE group and 11% (1/9) in the PR group (p = 0.33). Tables 4 and 5 present the comparison of PR and PE patients grouped according to the grade of duodenal injury (grade III or grade IV/V). No statistical difference was identified in any of the outcomes, but the number of patients in these subgroups was low. Discussion Primary repair (n = 34) Pyloric exclusion (n = 16) Age (years) 26 ± 8 27 ± Age [30 years 9 (26%) 6 (38%) 0.43 Male 32 (94%) 11 (69%) 0.01 AIS abdomen ISS ISS [25 11 (33%) 6 (38%) 0.72 GCS GCS \15 9 (29%) 3 (20%) 0.39 Shock 14 (41%) 10 (67%) 0.14 Grade III 28 (82%) 10 (63%) 0.11 Grade IV & V 6 (18%) 6 (37%) 0.11 Portion I and II 13 (42%) 11 (79%) 0.02 Abdominal organs 4 ± ± injured (no.) Pancreas 8 (24%) 10 (63%) 0.01 Colon 13 (38%) 4 (25%) 0.28 Small bowel 17 (50%) 6 (38%) 0.41 Complications 22 (65%) 11 (69%) 0.78 Duodenal leak 6 (18%) 4 (25%) 0.4 Intraabdominal sepsis 8 (35%) 5 (31%) 0.56 Pancreatic complications 5 (15%) 2 (12%) 0.6 Abdominal complications 18 (53%) 10 (63%) 0.31 Systemic complications 18 (53%) 9 (56%) 0.83 ICU LOS (days) 11 ± ± Hospital LOS (days) 32 ± ± ISS: injury severity score; AIS: abbreviated injury score; GCS: Glasgow Coma Scale; ICU: intensive care unit; LOS: length of stay The duodenum is infrequently injured owing to its protected retroperitoneal location [25]; duodenal injury is reported to be the reason for 3.7% of all laparotomies for trauma [26] and is rarely an isolated injury [5, 27]. An average of four abdominal organs were injured in our series. Overall, the major mechanism of injury is p

4 10 World J Surg (2008) 32:7 12 Table 3 Ten patients who developed duodenal suture line leaks after repair Age (years) Mechanism Time from admission to OR (min) Organs injured Duodenal procedure Additional complications Hospital stay (days) 31 Gunshot wound 45 Pancreas, colon, stomach, liver, spleen, diaphragm 28 Gunshot wound 6 Spleen, small bowel, colon, diaphragm, superior mesenteric artery and vein 16 Gunshot wound 25 Pancreas, liver, common bile duct, gallbladder, inferior vena cava 33 Gunshot wound 20 Pancreas, liver, colon, superior mesenteric vein 19 Gunshot wound 25 Pancreas, liver, small bowel, kidney, ureter 15 Stab wound 18 Pancreas, liver, small bowel, colon, inferior vena cava 50 Motor vehicle crash 66 Small bowel, colon, superior mesenteric artery and vein 33 Gunshot wound 195 Liver, small bowel, diaphragm Pyloric exclusion Abdominal abscess, wound infection, bacteremia, liver failure, ARDS, DVT, marginal ulcer Resection-anastomosis Pancreatic fistula, abdominal abscess, wound infection, bacteremia Pyloric exclusion Abdominal abscess 31 Duodenor-rhaphy Pyloric exclusion Abdominal abscess, wound infection, bacteremia, pulmonary failure, extremity compartment syndrome Abdominal abscess, wound infection, bacteremia, extremity compartment syndrome Resection-anastomosis Abdominal abscess, wound infection, bacteremia Pyloric exclusion Duodenor-rhaphy Abdominal abscess, bacteremia, wound infection, ARDS, MOF, gastric bleeding, small bowel infarction Pancreatic fistula, gastric ulcer, small bowel obstruction 30 Gunshot wound 20 Small bowel, colon Resection-anastomosis Abdominal abscess, wound infection, wound dehiscence, gastric bleeding, bacteremia, endocarditis 25 Gunshot wound 20 Small bowel, colon, kidney Resection-anastomosis Pancreatitis, abdominal sepsis OR: operating room; ARDS: adult respiratory distress syndrome; DVT: deep venous thrombosis; MOF: multiple organ failure penetrating trauma and especially gunshot wounds [5], which accounted for 76% of the SDIs in this study. The optimal management of SDI remains controversial. In rare circumstances and for catastrophic injuries of the pancreatoduodenal region, pancreatoduodenectomy is unavoidable [28, 29]; but for most grade III, IV, and V injuries, the trauma surgeon faces the dilemma of performing PR or PE. The theoretical argument for PE is based on the diverting gastric contents, decreasing the fluid volume (which is distending the duodenum), and avoiding activation of pancreatic enzymes. All these alterations are thought to act protectively on the tenuous duodenal repair and reduce the likelihood of suture line dehiscence. In a 1993 study from South Africa, [23], the anastomotic leak rate was 43% in 14 SDI patients who underwent PR (6 patients) and 12% in 17 SDI patients who had PE (2 patients). It was concluded that pyloric exclusion was the procedure of choice for SDI. Other authors [6] found no difference in duodenum-related complications between PR and PE and concluded that complex operations are rarely necessary. An analysis of risk factors of morbidity and mortality in 52 patients with traumatic duodenal injuries

5 World J Surg (2008) 32: Table 4 Grade III duodenal injuries: comparison of patients who underwent primary repair or pyloric exclusion Table 5 Grade IV or V duodenal injuries: comparison of patients who underwent primary repair or pyloric exclusion Parameter Primary repair (n = 28) Pyloric exclusion (n = 10) P Parameter Primary repair (n =6) Pyloric exclusion (n =6) p Age (years) 26 ± 8 2 ± Age [30 years 8 (29%) 4 (40%) 0.38 Male 27 (96%) 8 (80%) 0.16 AIS abdomen 4.3 ± ± ISS 21.5 ± ± ISS [25 10 (36%) 1 (10%) 0.13 GCS 14 ± 1 14 ± GCS \15 6 (22%) 1 (11%) 0.41 Shock 9 (32%) 5 (56%) 0.23 Abdominal organs 4 ± 2 4 ± injured (no.) Pancreas 5 (18%) 5 (50%) 0.05 Colon 10 (36%) 1 (10%) 0.13 Small bowel 14 (50%) 4 (40%) 0.43 Complications 17 (61%) 5 (50%) 0.56 Duodenal leak 4 (14%) 1 (10%) 0.6 Intra-abdominal sepsis 5 (18%) 3 (30%) 0.35 Pancreatic complications 2 (7%) Abdominal complications 12 (43%) 4 (40%) 0.59 Systemic complications 14 (50%) 4 (40%) 0.43 ICU LOS (days) 9 ± ± Hospital LOS (days) 23 ± ± Age (years) 26 ± 5 26 ± Age [30 years 1 (17%) 2 (33%) 0.5 Male 5 (83%) 3 (50%) 0.27 AIS abdomen 4.0 ± ± ISS ISS [25 1 (17%) 5 (83%) 0.04 GCS 14 ± 1 14 ± GCS \15 2 (33%) 3 (60%) 0.61 Shock 4 (80%) 5 (83%) 0.73 Abdominal organs 4.5 ± ± injured (no.) Pancreas 3 (50%) 5 (83%) 0.27 Colon 3 (50%) 3 (50%) 1.0 Small bowel 3 (50%) 2 (33%) 0.5 Complications 5 (83%) 6 (100%) 0.5 Duodenal leak 2 (33%) 3 (50%) 0.5 Intraabdominal sepsis 3 (50%) 1 (17%) 0.27 Pancreatic complications 3 (50%) 2 (33%) 0.5 Abdominal complications 6 (100%) 6 (100%) 1.0 Systemic complications 4 (67%) 5 (83%) 0.5 ICU LOS (days) 15 ± ± Hospital LOS (days) 60.5 ± ± failed to correlate the type of operation with outcome and concluded that the physiologic presentation of the patient is the most important factor when predicting mortality [30]. Similarly, among 167 patients with duodenal and/or pancreatic injury, hypothermia and the presence of pancreatic injury, but not the type of operation, were independent predictors of infectious complications [31]. In a multicenter study, duodenal fistula and anastomotic dehiscence were more common in patients undergoing PE [5]. In a previous study of duodenal injuries from our institution, the use of complex procedures, including PE, increased for injuries of high severity, resulting in favorable outcomes [24]. In the present study we focused on patients with grade III, IV, and V duodenal injuries. We excluded patients with injuries of lower grade because it is generally accepted that simple repair is adequate. Similarly, we excluded patients with catastrophic pancreatoduodenal injuries who underwent a Whipple procedure as PR may not be an option in these cases. The incidence of postoperative complications related directly to the duodenal repair was 20% in the current study and consistent with other reports of SDI [32, 33]. Although the anastomotic leak rate was associated with the severity of duodenal injury (12% in grade III injuries vs. 43% in grades IV and V injuries, p = 0.031), there was no detectable relation with the surgical technique chosen for repair. The high overall incidence of postoperative complications (66%) confirms results from other studies [6] and literature reviews [32, 33] about duodenal trauma being a marker of severe injury. In contrast, mortality was low among patients who survived the crucial first 48 hours. Only one patient died from multiple system organ failure on the sixth postoperative day. Despite the lack of difference in most variables examined between the PR and PE groups, including ISS and abdominal AIS, a similarity of these groups cannot be assumed. The two procedures were not used randomly. As shown by the higher frequency of pancreatic and part I/II duodenal injuries in the PE group, bias may have existed toward using PE for more severe duodenal injuries. At the same time, PE may have been avoided even for complex injuries if the patient was physiologically unstable and the operation needed to be abbreviated. These intraoperative decisions, which were made by experienced and dedicated trauma surgeons, cannot be extracted retrospectively. Of note, among patients with associated pancreatic injuries, 1 of 9 had a duodenal leak following PR compared to 3 of 10 following PE.

6 12 World J Surg (2008) 32:7 12 To date, the body of literature supports more complex procedures for more complex duodenal injuries [32, 33]. However, a recent analysis from South Africa of lowvelocity civilian gunshot injuries to the duodenum supports the liberal use of PR even for SDIs [34]. PR was used in all 46 patients with grade III/IV duodenal injuries who survived the initial 24 hours (40 duodenorrhaphies and 6 resections with primary anastomosis); two patients with grade V injuries underwent pancreatoduodenectomy. Among these 46 patients duodenum-related morbidity was recorded in 9 patients (19.5%), including 6 duodenal leaks and duodenum-related mortality in 3 (6.5%). Notably, among PR patients the duodenum-related morbidity was not different between those with simple repair of lacerations (8/40, 20%) and those with resection and anastomosis (1/7, 14%). The current study is the largest review with an exclusive focus on SDIs, but it is limited by the nature of its retrospective design. Although PR remains an option in the armamentarium of trauma surgeons, increasingly more SDIs can be managed by PR. The precise criteria for doing so are difficult to discern retrospectively. The recommendation of preferring PE in SDI should be reconsidered. As ideal as it may be, a multicenter prospective randomized trial among SDI patients is unlikely, and treatment choices will continue to be based on non-level I evidence. Mounting evidence on the safety of PR allows confidence in using it more liberally in SDI. References 1. Berg AA (1923) Duodenal fistula: its treatment by gastrojejunostomy and pyloric exclusion. Ann Surg 78: Adkins RB Jr, Keyser JE 3rd (1985) Recent experiences with duodenal trauma. Am Surg 51: Ivatury RR, Nallathambi M, Gaudino J, et al. (1985) Penetrating duodenal injuries: analysis of 100 consecutive cases. Ann Surg 202: Flynn WJ, McCoy M, Richardson DJ (1990) Reappraisal of pancreatic and duodenal injury management based on injury severity. Arch Surg 125: Cogbill TH, Moore EE, Feliciano DV, et al. (1990) Conservative management of duodenal trauma: a multicenter prospective. J Trauma 30: Nassoura ZE, Ivatury RR, Simon RJ, et al. (1994) A prospective reappraisal of primary repair of penetrating duodenal injuries. Am Surg 60: Kline G, Lucas CE, Ledgerwood AM, et al. (1994) Duodenal organ injury severity score and outcome. Am Surg 60: Berne CJ, Donovan AJ, White EJ, et al. (1974) Duodenal diverticulization for duodenal and pancreatic injury. Am J Surg 127: Moore EE, Cogbill TH, Malangoni MA, et al. (1990) Organ injury scaling. II. Pancreas, duodenum, small bowel, colon, and rectum. J Trauma 30: Yellin AE, Rosoff L (1975) Pancreatoduodenectomy for combined pancreatoduodenal injuries. Arch Surg 110: Orescovich MR, Carrico CJ (1984) Pancreativoduodenectomy for trauma: a viable option? Am J Surg 147: Feliciano DV, Martin TD, Cruse PA, et al. (1987) Management of combined pancreatoduodenal injuries. Ann Surg 205: McKone TK, Bursch LR, Scholten DJ (1988) Pancreaticoduodenectomy for trauma: a life-saving procedure. Am Surg 54: Heimansohn DA, Canal DF, McCarthy MC, et al. (1990) The role of pancreaticoduodenectomy in the management of traumatic injuries to the pancreas and duodenum. Am Surg 56: Morton Jr, Jordan GL Jr (1968) Traumatic duodenal injuries: review of 131 cases. J Trauma 8: Stone HH, Fabian TC (1979) Management of duodenal wounds. J Trauma 19: DeSantis M, Dvereux D, Thompson D (1987) Pyloric exclusion: suture material of choice. Am Surg 53: Kashuk JL, Moore EE, Cogbill TH (1982) Management of the intermediate severity duodenal injury. Surgery 92: Feliciano D, Moore EE, Pachter HL, et al. (1986) Symposium pancreaticoduodenal trauma. Contemp Surg 29: Martin TD, Feliciano DV, Mattox KL (1983) Severe duodenal injuries: treatment with pyloric exclusion and gastrojejunostomy. Arch Surg 118: Ginzburg E, Carrillo EH, Sosa JL, et al. (1997) Pyloric exclusion in the management of duodenal trauma: is concomitant gastrojejunostomy necessary? Am Surg 63: Buck JR, Sorensen JJ, Fath HM, et al. (1992) Severe pancreatoduodenal injuries: the effectiveness of pyloric exclusion with vagotomy. Am Surg 58: Degiannis E, Krawczykowski D, Velmahos GC, et al. (1993) Pyloric exclusion in severe penetrating injuries of the duodenum. World J Surg 17: Velmahos GC, Kamel E, Chan LS, et al. (1999) Complex repair for the management of duodenal injuries. Am Surg 10: Jurkovich GJ (2000) The duodenum and pancreas. In: Mattox KL, Feliciano DV, Moore EE (eds) Trauma. 4th edition. New York: McGraw-Hill. p Weigelt JA (1990) Duodenal injuries. Surg Clin North Am 70: Jansen M, Du Toit DF, Warren BL (2002) Duodenal injuries: surgical management adapted to circumstances. Injury 33: Lopez PP, Benjamin R, Cockburn M, et al. (2005) Recent trends in the management of combined pancreatoduodenal injuries. Am Surg 71: Asensio JA, Petrone P, Roldan G, et al. (2003) Pancreaticoduodentomy: a rare procedure for the management of complex pancreaticoduodenal injuries. J Am Coll Surg 197: Huerta S, Bui T, Porral D, et al. (2005) Predictors of morbidity and mortality in patients with traumatic duodenal injuries. Am Surg 71: Tyburski JG, Dente CH, Wilson RF, et al. (2001) Infectious complications following duodenal and/or pancreatic trauma. Am Surg 67: Asensio JA, Petrone P, Roldan G, et al. (2002) Pancreatic and duodenal injuries: complex and lethal. Scand J Surg 91: Degiannis E, Boffard K (2000) Duodenal injuries. Br J Surg 87: Talving P, Nicol AJ, Navsaria PH (2006) Civilian duodenal gunshot wounds: surgical management made simpler. World J Surg 30: