- Email: [email protected]
Management of penetrating pancreatic trauma: an 11-year experience of a level-1 trauma center
Injury, Int. J. Care Injured 32 (2001) 753– 759 www.elsevier.com/locate/injury
Management of penetrating pancreatic trauma: an 11-year experience of a level-1 trauma center Julio C. Vasquez, Raul Coimbra *, David B. Hoyt, Dale Fortlage Department of Surgery, Di6ision of Trauma, Uni6ersity of California, UCSD Medical Center, 200 W. Arbor Dr., San Diego, CA 92103 -8896, USA Accepted 15 June 2001
Abstract Introduction: We present our experience in the management of penetrating pancreatic injuries, focusing on factors related to complications and death. Methods: Retrospective trauma registry-based analysis of 62 consecutive patients with penetrating pancreatic injuries during an 11-year period. Overall injury severity was assessed by the injury severity score (ISS) and the penetrating abdominal trauma index (PATI). Pancreatic injuries were graded according to the American Association for the Surgery of Trauma (AAST) Organ Injury Scaling (OIS). Complications were characterised using standardised definitions. Mortality was recorded as early (within 48 h after admission) and late (after 48 h). Results: Thirty patients suffered gunshot wounds and 24 had grade I pancreatic injuries. Shotgun and gunshot wounds were more destructive than stab wounds (higher PATI, number of intraabdominal injuries and mortality). Seventeen patients died. Most deaths occurred within 1 h after admission due to massive bleeding and severe associated injuries. Only one death was potentially related to the pancreatic injury. Mortality rate also correlated with pancreatic injury grading. Sixty-one patients had associated intraabdominal injuries. Combined pancreaticoduodenal injuries were present in 13 patients, and five died. Simple drainage was the most common procedure performed. Pancreas-related complications were found in 12 out of 47 patients who survived more than 48 h; intraabdominal abscess (n =7) that was associated with colon injuries, and pancreatic fistula (n=5). Conclusion: An approach based on injury grade and location is advised. Routine drainage is recommended; distal resection is indicated in the presence of main duct injury, and the management of severe injuries will be tailored according to the overall physiologic status, presence of associated injuries, and duodenal viability. Morbidity and mortality is mainly due to associated injuries. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction Pancreatic injury is infrequent following blunt or penetrating abdominal trauma [1]. The incidence of pancreatic injuries from penetrating mechanisms varies from 5 to 7% [2]. Isolated pancreatic injuries (2%) are uncommon due to its retroperitoneal location and the proximity of multiple organs and major vascular structures [3]. Mortality rates after penetrating trauma vary from 13 to 32%, however, most deaths are related to associated injuries [4]. The diagnosis of a penetrating pancreatic injury is usually made intraoperatively. * Corresponding author. E-mail address: [email protected] (R. Coimbra).
Several aspects in the management of penetrating pancreatic injuries are still controversial, particularly in the era of damage control operation. The optimal and definitive management of severe injuries involving the main pancreatic duct is still open to debate. There is a scarcity of reports solely dedicated to penetrating pancreatic trauma. We report our experience with the management of such injuries over a decade in a level-1 trauma centre, focusing on determinants of complications and deaths according to mechanism and injury severity. 2. Patients and methods We performed a retrospective review of 66 consecutive patients sustaining penetrating pancreatic injury
0020-1383/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 0 - 1 3 8 3 ( 0 1 ) 0 0 0 9 9 - 7
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
754
3. Results
treated during an 11-year period at the University of California San Diego Medical Centre, a level I Trauma Centre. Four patients were excluded because of missing records, leaving a total of 62 patients, which corresponds to 5.2% of all penetrating abdominal trauma patients (n=1192) treated during the study period. All clinical records, as well as autopsy data, when applicable, were reviewed. We recorded the following variables: age, gender, mechanism of injury, shock upon admission, anatomic location of the pancreatic injury, associated intra and extra abdominal injuries, classification of pancreatic injury, surgical management, hospital length of stay, complications and mortality. The penetrating abdominal trauma index (PATI) was calculated as described by Moore et al. [5] and Borlase et al. [6]. The injury severity score (ISS) was calculated according to Baker et al. [7]. Pancreatic injury classification was determined according to the Organ Injury Scaling (OIS) of the American Association for the Surgery of Trauma (AAST) [8]. In brief, grade I included minor contusion or superficial laceration without duct injury, grade II was defined as major contusion or laceration without duct injury or tissue loss, grade III included distal transection or parenchymal injury with duct injury, grade IV included proximal (right of the superior mesenteric vein) transection or parenchymal injury involving the ampulla, and grade V was a massive disruption of the pancreatic head. Pancreatic fistula was defined as drainage of more than 50 cc/day of fluid with elevated amylase level for at least 3 consecutive days [2]. Pancreatitis was defined as elevated serum amylase for more than 3 days associated with clinical findings [2]. Statistical analysis was performed using non-parametric Mann– Whitney, x2 or Student’s t-test when appropriate. A P value of B 0.05 was considered statistically significant.
Fifty-four patients (87.1%) were male and the mean age was 28.5 years (range from 16 to 56). Thirty patients (48.4%) sustained gunshot wounds, 27 (43.5%) stab wounds, and five (8.1%) shotgun wounds. The diagnosis of pancreatic injury was made during autopsy in nine patients, and during laparotomy in the remaining 53 patients. Twenty nine patients had systolic blood pressureB 90 mmHg on admission, and were considered in shock. Patients admitted in shock had significantly higher risk of dying when compared to haemodynamically stable patients (PB0.0001).
3.1. Pancreatic injury location and grading The body of the pancreas was the most common injury site (n=21), followed by the head (n= 19), and tail (n= 19). Two patients had injuries in more than one pancreatic region. Grade 1 pancreatic injuries were the most frequent (Table 1). Mortality rate correlated with pancreatic injury severity. Mortality rate for grade 1 injuries was 12.5%, increasing to 66.6% in grade 4 injuries and 100% in grade 5 injuries.
3.2. Trauma scores The anatomic injury severity scores correlated with mortality. PATI was calculated for a total of 53 patients. Mean PATI was 35.69 22.4 (range: 5– 104). Non-survivors had higher PATI values than survivors (57.8 vs. 31.6; PB 0.01). Mean ISS was 28.2917.3 (range: 4–75). Non-survivors had higher ISS than survivors (42.82 vs. 22.66; PB 0.01).
3.3. Associated injuries Associated intraabdominal injuries were present in 61 patients (98.4%). The number of organs injured varied from 1 to 9, with a mean of 3.6 per patient. The stomach was the most frequent organ injured in associ-
Table 1 Grade of pancreatic injury and survival (n = 62) Grade
Grade Grade Grade Grade Grade N/A Total
Alive
1 2 3 4 5
21 12 11 1 0 0 45
Dead B48 h
\48 h
3 4 4 1 1 2 15
0 1 0 1 0 0 2
Total (%)
Mortality (%)
24 17 15 3 1 2 62
3 5 4 2 1 2 17
(38.7) (27.4) (24.2) (4.8) (1.6) (3.2) (100.0)
(12.5) (29.4) (26.6) (66.6) (100.0) (100.0) (27.4)
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759 Table 2 Associated intra abdominal injuries (n= 62) Intraabdominal organs
Stomach Liver Diaphragm Kidney Major vascular Minor vascular Spleen Small bowel Large bowel Duodenum Extrahepatic biliary Ureter
755
3.4. Surgical management
Mechanism
Total (%)
GSW
SW
SGW
16 21 20 15 14 6 7 9 6 7 1 0
12 6 5 7 7 11 7 2 5 3 2 1
5 5 2 2 2 2 2 3 3 3 2 0
33 32 27 24 23 19 16 14 14 13 5 1
(53.2) (51.6) (43.5) (38.5) (37.1) (30.6) (25,8) (22.6) (22.6) (21.0) (8.1) (1.6)
GSW, gunshot wound; SW, stab wound; SGW, shotgun wound.
ation with the pancreas (Table 2). The number of associated intra abdominal injuries was higher in patients sustaining gunshot wounds and shotgun injuries (4.0 and 6.2, respectively) than in patients with stab wounds (2.6; PB 0.01). Also, patients who died had a higher number of associated intraabdominal injuries than survivors (4.41 and 3.24, respectively, P B0.05). Combined pancreaticoduodenal injuries were present in 13 patients and 5 (38.5%) died. One patient died shortly after arrival and did not receive any surgical treatment. Most duodenal injuries were primarily repaired, and 2 patients underwent pyloric exclusion. Extra-abdominal associated injuries were found in 37 of 62 patients (59.7%), and included thoracic injuries (n= 24), soft tissue lesions (n =16), bone fractures (n= 5), spinal/vertebral column injuries (n = 4) and neck/facial injuries (n= 3).
Resuscitative thoracotomy was performed in ten patients admitted ‘in extremis’. All patients who underwent a resuscitative thoracotomy died, eight intraoperatively and two within 48 h postoperatively. Exploratory laparotomy was performed in 53 patients, 15 of whom required reexploration for various indications. Gunshot and shotgun wound victims underwent mandatory abdominal exploration. Stab wound patients underwent surgical exploration based on different indications, including evisceration of omentum (n= 11), hypotension (n= 7), positive diagnostic peritoneal lavage (n= 3), clinical diagnosis of peritonitis (n=3), blood in the nasogastric tube (n=2), and positive fascial penetration on digital wound exploration (n= 1). Surgical management varied depending on location and grade of the pancreatic injury. The most common procedure was simple drainage with closed suction drains (Table 3). The type of closure of the pancreatic stump following partial resection was with nonabsorbable suture (n = 10), staples with nonabsorbable suture (n =1), and staples alone (n=2). There was no difference in the development of pancreatic fistula or intraabdominal abscess when comparing different modalities of pancreatic stump closure. Only one out of 13 patients who underwent distal pancreatic resection did not undergo splenectomy.
3.5. Complications Pancreas-related complications were found in 12 out of 47 patients (29.7%) who survived more than 48 h (Table 4). These included six patients out of 27 with stab wounds (22.2%), five out of 17 with gunshot wounds (29.41%), and one out of 3 with shotgun wounds (33.33%). Patients who developed pancreas-re-
Table 3 Surgical management and grade of pancreatic injury (n = 62) Surgical procedure
No surgical procedure for pancreas Drainage Distal resection and Drainage Suture ligation and drainage Debridement and drainage Distal resection, duct ligation, and drainage Pyloric exclusion and drainagea Local bleeding control Pancreaticoduodenectomy Total a
Grade of injury N/A
1
2
3
4
5
Total
2 0 0 0 0 0 0 0 0 2
4 13 0 4 1 0 1 1 0 24
5 3 1 3 3 1 1 0 0 17
2 0 11 0 0 2 0 0 0 15
1 1 0 0 0 0 0 0 1 3
0 0 0 0 0 0 0 0 1 1
14 17 12 7 4 3 2 1 2 62
Pyloric exclusion performed because of the severity of duodenal injury.
756
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
Table 4 Pancreas-related complications and mechanism of injury in patients surviving more than 48 h (n =47) Complication
Intraabdominal abscess Pancreatic fistula Pancreatic hemorrhage Pseudocyst Pancreatitis
Mechanism
Mean PATI
GSW
SW
SGW
Total
3 2 1 0 2
3 2 2 2 0
1 1 1 0 0
7 5 4 2 2
53.8 57.4 59.3 18.5 32.0
Some patients had more than one complication. GSW, gunshot wound; SW, stab wound; SGW, shotgun wound; PATI, penetrating abdominal trauma index.
lated complications had longer hospital stay (31.4 vs. 15.8 days; PB0.05). Intraabdominal abscess was the most frequent postoperative complication. Pancreas-related complications according to injury severity are listed in Table 5. Non pancreatic-related complications were present in 22 (46.8%), and included atelectasis (n = 7), pneumonia (n = 7), urinary tract infection (n= 4) and wound infection (n = 3).
3.6. Intraabdominal abscess Seven patients developed intraabdominal abscesses between postoperative day 9 and 30, and the mean length of stay was 45 days. Mean PATI values in this group were higher (53.8 vs. 27.5, P B0.05), stressing the importance of associated injuries in the genesis of intraabdominal abscesses.. All cultures from the abscess were positive, many of them with more than one species, and included Enterobacter aerogenes, Enterococcus faecalis, Bacteroides sp., Pseudomonas aeruginosa, Candida sp., and Fusobacterium sp. A higher incidence of intraabdominal abscesses was found in the subgroup of patients with associated colonic injuries (PB 0.05). Treatment included percutaneous CT or ultrasound guided drainage in five and surgical drainage in two patients.
3.7. Pancreatic fistula Pancreatic fistula was diagnosed in five patients between postoperative day 1 and 25, (mean of 9 days), and the mean length of stay was 39.2 days. Two patients had sinograms that revealed communication with the pancreatic duct. External drainage was the treatment in all. Octreotide was used in four patients, with decrease in output. One patient, who was HIV positive, died 23 days after surgery due to sepsis and multiorgan failure. All other patients were discharged in a healthy state, and the fistulas closed in an average period of 38.3 days. Ligation of the main pancreatic duct following distal pancreatic resection did not influence the incidence of
pancreatic fistulas or intraabdominal abscesses (P= 0.20).
3.8. Postoperati6e hemorrhage Four patients developed postoperative pancreatic haemorrhage. In three, bleeding occurred within the first 24 h post surgery and one patient bled on postoperative day 19. Two patients presented with bloody drainage, hypotension and low haematocrit. Another patient became hypotensive and developed a coagulopathy shortly after the initial operation requiring reexploration. The remaining patient became hypotensive, had abdominal distention and low haematocrit. An abdominal angiogram was obtained but it did not reveal the source of bleeding. The patient was re-explored and bleeding from the pancreatic parenchyma was found. All these patients received 4–26 U of blood (mean of 15.8 U), and the bleeding site was controlled surgically. Intraoperative findings included bleeding from small pancreatic vessels in three and at the splenic artery stump in one.
3.9. Posttraumatic pancreatic pseudocyst Two patients developed pancreatic pseudocysts diagnosed on postoperative days 12 and 19, respectively. They had been discharged in good condition, but reTable 5 Grade of pancreatic injury and pancreas-related complications in patients surviving more than 48 h (n = 47) Grade of injury
Grade 1 Grade 2 Grade 3 Grade4 Total
Complications A
B
C
D
E
n/Total
0 2 0 0 2
0 1 2 1 4
1 1 0 0 2
1 2 4 0 7
0 0 4 1 5
2/21 5/12 4/12 1/2 12/47
Some patients had more than one complication. A, pseudocyst; B, hemorrhage; C, pancreatitis; D, abscess; E, fistula.
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
turned to the hospital complaining of abdominal pain associated with nausea and vomiting. Abdominal CT scan confirmed the diagnosis. Surgical drainage was not required, and they resolved after total parenteral nutrition therapy for 3 weeks.
3.10. Pancreatitis Two patients developed pancreatitis on postoperative days 10 and 14. The first patient had persistent high serum amylase, with mild epigastric abdominal pain. Treatment included enteral nutrition through a jejunostomy tube and blood amylase levels returned to normal values within a week. The second patient presented with abdominal pain, nausea, vomiting and persistently elevated serum and urinary amylase levels. Abdominal CT demonstrated pancreatic oedema. Treatment included total parenteral nutrition, with good outcome. No endocrine or exocrine pancreas-related complications were identified.
3.11. Mortality Seventeen patients died (27.4%). Ten died within 1 h of arrival at the hospital. The mortality rate for gunshot, shotgun, and stab wounds was 43.3, 40.0, and 7.7%, respectively. Sixteen patients died from non-pancreatic causes, hypovolaemic shock (n = 14) being the most common culprit. Cardiac or major vessel injury was present in 13, and profuse liver bleeding in one patient. The remaining two patients died of multiorgan failure. Only one death was potentially related to the pancreatic injury. This was an HIV positive patient who sustained injuries to the inferior vena cava, spleen and head of the pancreas (Grade IV). In view of the injury severity, it was decided during laparotomy not to proceed with pancreaticoduodenectomy, but to place drains in order to have a controlled pancreatic fistula. On postoperative day 1, the patient became hypotensive requiring surgical intervention. Bleeding from minor pancreatic vessels was successfully controlled, however the patient subsequently died of sepsis and multiorgan failure on postoperative day 22. In retrospect, the repeated episodes of hypotension might have contributed to the development of sepsis and multiple organ failure in this patient. Thirteen patients died in the operating room, and four patients died during the postoperative period. Of note, a total of nine patients (14.5%) were found to have pancreatic injury during post-mortem examination. These patients had massive bleeding form associated injuries and all expired before a thorough evaluation of all intra abdominal injuries could be performed. The mean hospital stay for patients surviving more than 48 h (n =47) was 20.2 days.
757
4. Discussion The prevalence of pancreatic injury among penetrating abdominal trauma victims in our series was 5.20%, which is within the range (5– 7%) reported in the literature [9]. Shotgun and gunshot injuries were more destructive than stab wounds as reflected by higher PATI, ISS, number of intra abdominal organs injured and mortality. Due to its retroperitoneal location, the pancreas is surrounded by important organs and vessels, which contributes to the rarity of isolated pancreatic injury (1.6% in our series), and reported in 2% of cases of penetrating pancreatic trauma [10]. Death is mainly caused by associated injuries rather than by the pancreatic injury [11]. The diagnosis of a pancreatic injury is usually made intraoperatively, once most of these patients have associated injuries (3.6 injuries/patient in the present series), and present with clinical signs and symptoms of intraabdominal injuries requiring surgical intervention. Appropriate surgical management of penetrating pancreatic injuries depends on location (central vs. peripheral, head, body or tail, right or left of the mesenteric vessels), the extent of the injury, and the presence of main pancreatic duct injury. Injuries to adjacent organs should raise the suspicion of a pancreatic injury. A thorough evaluation of the pancreatic parenchyma and exploration of peripancreatic haematomas is mandatory to adequately assess ductal integrity and injury severity. Different modalities of surgical treatment for pancreatic injuries have been proposed, but the use of a graded approach to its management is associated with a significant reduction in morbidity and mortality [12]. In our series, grade 1 and 2 injuries (no ductal injuries) were usually treated with closed drainage with or without local debridement. This approach proved to be effective, safe, and is the most commonly performed [13,14]. It seems that closed suction drainage as opposed to sump or open drainage decreases the incidence of pancreas-related septic complications [13], and is the current method of choice. Duration of drainage, however, is still a matter of controversy. Some authors recommend drainage for 7– 10 days or until the patient is tolerating an oral diet [15]. It seems that this approach is appropriate for grades 3– 5 injuries; however, in minor injuries drains should be removed earlier, if drainage is minimal or ceases. All grade III injuries in the present series were treated by distal resection and drainage. The integrity of the pancreatic duct has been suggested as an important factor in determining the method of treatment [12]. This may be achieved by visual inspection in distal injuries of the pancreas, but it is difficult to assess in the pancreatic head where intraoperative pancreatography has been recommended [16]. Several methods have been
758
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
described for intra operative evaluation of the main pancreatic duct if direct visualisation is inconclusive: distal pancreatectomy and direct catheterisation of the main duct in the tail of the pancreas, cannulation of the major duodenal papilla through a surgically created duodenotomy, injection of contrast media in the gallbladder (cholecystopancreatography), and endoscopic retrograde cholangiopancreatography. All these techniques have limitations and the associated complication rate is significant. How aggressively one should pursue an intra operative diagnosis of main pancreatic duct injury is still debatable. We have adopted a more conservative approach. If by visually inspecting the pancreas and eventually stimulating pancreatic secretion with secretin intra operatively no duct injury is identified, we have opted for drainage rather than pursuing intra operative radiologic evaluation. Distal pancreatectomy is the treatment of choice for grade 3 injuries. A decrease in the number of postoperative complications following distal pancreatectomy has been shown when compared to drainage alone in haemodynamically stable patients [1,2,17,18]. Haemodynamically unstable patients with multiple associated injuries and physiologic derangement (acidosis, coagulopathy, and hypothermia) are best served by damage control operation and pancreatic drainage. Splenic preservation after partial pancreatic resection has been mainly proposed for children [2]. In our series, this was done in only one out of 13 patients who underwent distal pancreatectomy for grade 3 injuries. As reported previously by Degiannis et al., [19], we found that the type of suture used to approximate the edges of the pancreatic stump was unrelated to the development of postoperative complications. Grade 4 injuries should be treated by distal resection, when feasible, or drainage alone. In very special circumstances, such as in resection of more than 80% of the pancreatic parenchyma and associated diabetes mellitus, preservation of the distal pancreas and internal drainage by Roux-en-Y jejunal anastomosis should be considered. Grade 5 injuries characterised by massive devascularisation of the pancreatic head and adjacent duodenum are infrequent, usually require pancreaticoduodenectomy or duodenal diverticulisation, and are accompanied by elevated mortality rates [20,21]. A more conservative approach consisting of extensive peripancreatic drainage and primary repair of the duodenum is indicated in unstable critically ill patients. In our series, complex procedures such as pancreaticoduodenectomy were performed in only two patients. Pancreaticoduodenectomy entails very high mortality rates (30–40%) and is not recommended for acute situations unless the missile has effectively performed the resection and the operation is limited to the debridement of devitalised tissue [18]. Combined pancreaticoduodenal injuries carried a high mortality rate (38.5%) and were
most often treated by primary repair. Pyloric exclusion has been suggested as a valuable method in the treatment of severe duodenal injuries, leading to a reduction in the postoperative leakage rate, and it has been also considered a viable option for grade 4–5 pancreatic injuries associated with duodenal injuries [22]. If the duodenum or the main pancreatic duct is not involved, minor injuries to the head of the pancreas should be treated by simple drainage [23,24]. In general, the postoperative complication rate increases as the injury severity increases. Other factors that contribute to the development of pancreas-related complications include main duct injury and associated duodenal injuries [4,12,25]. Intra abdominal abscess formation was the most frequent postoperative complication (15%) in patients surviving more than 48 h. Increase PATI in this group indicates the importance of associated injuries in the genesis of this complication. As previously reported by others [17], we also found that combined pancreatic and colon injuries were associated with the development of intraabdominal abscesses. Adequate treatment of intra abdominal abscesses includes percutaneous or surgical drainage. The incidence of pancreatic fistula was 10.6% according to the definition used in this study (drainage of more than 50 cc/day of amylase-rich fluid for more than 3 days), and it is within the range of 5–30% reported in the literature [17,21]. All patients were treated non-operatively, one patient died, and in the remaining four patients, closure occurred within 6 weeks. Octreotide has been used for treatment of this complication, but no significant improvement in outcome has been found [26]. Four patients received octreotide in an attempt to decrease pancreatic exocrine secretion. In all 4 patients the fistulas resolved but a direct effect of this drug on fistula closure cannot be established. Pancreatic pseudocysts are more frequently seen following blunt trauma, and are often the result of inadequate pancreatic drainage [27]. The incidence was 4.5% in the present series. Usually patients with pseudocyst present late, as occurred in both patients in this series, who were initially discharged and then returned to the hospital with symptoms. Both patients had good outcome after receiving total parenteral nutrition and none required operative intervention. The overall mortality rate of 27.4% is in accordance with data previously reported in the literature [4,9,17,25], and it includes patients who arrived in extremis to the hospital and died shortly thereafter. The mortality was directly related to associated injuries, as reflected by a higher PATI, ISS and number of intraabdominal organs injured when compared to survivors. Most patients died within 24 h after admission, mainly due to associated cardiac or vascular injuries with
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
resultant hypovolaemia. In only one patient the cause of death was possibly related to the pancreatic injury. Pancreas-related deaths are uncommon, usually occur late, and are related to sepsis and multiorgan failure [3,4,25,28,29]. We conclude that penetrating pancreatic injuries are associated with elevated mortality and morbidity rates. An approach based on injury grade and location is advised. Routine drainage is recommended despite injury severity; distal resection is indicated in the presence of main duct injury, and the management of severe (grade 4–5) injuries will be tailored according to the overall physiologic status, presence of associated injuries, and duodenal viability. Most complications and deaths are related to associated injuries.
References [1] Wisner DH, Wold RL, Frey CF. Diagnosis and treatment of pancreatic injuries. Arch Surg 1990;125:1109 –13. [2] Cogbill TH, Moore EE, Morris JA Jr, Hoyt DB, et al. Distal pancreatectomy for trauma: a multicenter experience. J Trauma 1991;31:1600 – 6. [3] Jones RC. Management of pancreatic trauma. Am J Surg 1985;150:698 – 704. [4] Young PR, Meredith JW, Baker CC, et al. Pancreatic injuries resulting from penetrating trauma: a multi-institution review. Am Surg 1988;64:838 –44. [5] Moore EE, Dunn EL, Moore JB, Thompson JS. Penetrating abdominal trauma index. J Trauma 1981;21:439 –45. [6] Borlase BC, Moore EE, Moore FA. The abdominal trauma index: a critical reassessment and validation. J Trauma 1990;30:1340 – 4. [7] Baker SP, O’Neill B, Haddon W Jr, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14:187 – 96. [8] Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, et al. Organ Injury Scaling, II: pancreas, duodenum, small bowel, colon and rectum. J Trauma 1990;30:1427 – 9. [9] Jones RC, Foreman ML. Pancreas. In: Ivatury R, Cayten CG, editors. The Textbook of Penetrating Trauma. Media, Pennsylvania: Williams and Wilkins, 1996:631 –42. [10] Jones RC. Management of pancreatic trauma. Ann Surg 1978;187:555 – 64.
759
[11] Stone HH, Fabian TC, Satiani B, Turkleson ML. Experiences in the management of pancreatic trauma. J Trauma 1981;21:257 – 62. [12] Smego DR, Richardson JD, Flint LM. Determinants of outcome in pancreatic trauma. J Trauma 1985;25:771 – 6. [13] Fabian TC, Kudsk KA, Croce MA, Payne LW, Mangiante E, Voeller GR, Britt LG. Superiority of closed suction drainage for pancreatic trauma. Ann Surg 1990;211:724 – 30. [14] Madiba TE, Mokoena TR. Favorable prognosis after surgical drainage of gunshot, stab or blunt trauma of the pancreas. Br J Surg 1995;82:1236 – 9. [15] Northrup WF, Simmons RL. Pancreatic trauma: a review. Surgery 1972;71:27 – 43. [16] Berni GA, Bandyk DF, Oreskovich MR, Carrico CJ. Role of intraoperative pancreatography in patients with injury to the pancreas. Am J Surg 1982;143:602 – 5. [17] Ivatury RR, Nallathambi M, Rao P, Stahl W. Penetrating pancreatic injuries. Am Surg 1990;56:90 – 5. [18] Wilson RH, Moorehead RJ. Current management of trauma to the pancreas. Br J Surg 1991;78:1196 – 202. [19] Degiannis E, Levy RD, Potokar T, Lennox A, Rowse A, Saadia R. Distal pancreatectomy for gunshot injuries of the distal pancreas. Br J Surg 1995;82:1240 – 2. [20] Mansour MA, Moore JB, Moore EE, Moore FA. Conservative management of combined pancreatoduodenal injuries. Am J Surg 1989;158:531 – 5. [21] Wynn M, Hill D, Miller DR, Waxman K, Eisner ME, Gazzaniga AB. Management of pancreatic and duodenal trauma. Am J Surg 1985;150:327 – 32. [22] Degiannis E, Krawczykowski D, Velmahos GC, Levy RD, Souter I, Saadia R. Pyloric exclusion in severe penetrating injuries of the duodenum. World J Surg 1993;17:751 – 4. [23] Lewis G, Krottenbelt JD, Krige JEJ. Conservative surgery for trauma to the pancreatic head: is it safe? Injury 1991;22:372 –4. [24] Feliciano DV, Martin TD, Curse PA, Graham JM, Buch JM, Mattox KL, Bitondo CG, Jordan GL. Management of combined pancreatoduodenal injuries. Ann Surg 1987;205:673 – 80. [25] Akhrass R, Yaffe MB, Brandt CP, et al. Pancreatic trauma: A ten-year multi-institutional experience. Am Surg 1997;63:598 – 604. [26] Nwariaku FE, Terracina A, Mileski WJ, Minei JP, Carrico CJ. Is octreotide beneficial following pancreatic injury? Am J Surg 1995;170:582 – 5. [27] Graham JM, Mattox KL, Jordan GL. Traumatic injuries of the pancreas. Am J Surg 1978;136:744 – 8. [28] Flynn W Jr, Cryer HG, Richardson JD. Reappraisal of pancreatic and duodenal injury management based on injury severity. Arch Surg 1990;125:1539 – 41. [29] Sorensen VJ, Obeid FN, Horst HM, Bivins BA. Penetrating pancreatic injuries, 1978 – 1983. Am Surg 1986;52:354 – 8.
Management of penetrating pancreatic trauma: an 11-year experience of a level-1 trauma center Julio C. Vasquez, Raul Coimbra *, David B. Hoyt, Dale Fortlage Department of Surgery, Di6ision of Trauma, Uni6ersity of California, UCSD Medical Center, 200 W. Arbor Dr., San Diego, CA 92103 -8896, USA Accepted 15 June 2001
Abstract Introduction: We present our experience in the management of penetrating pancreatic injuries, focusing on factors related to complications and death. Methods: Retrospective trauma registry-based analysis of 62 consecutive patients with penetrating pancreatic injuries during an 11-year period. Overall injury severity was assessed by the injury severity score (ISS) and the penetrating abdominal trauma index (PATI). Pancreatic injuries were graded according to the American Association for the Surgery of Trauma (AAST) Organ Injury Scaling (OIS). Complications were characterised using standardised definitions. Mortality was recorded as early (within 48 h after admission) and late (after 48 h). Results: Thirty patients suffered gunshot wounds and 24 had grade I pancreatic injuries. Shotgun and gunshot wounds were more destructive than stab wounds (higher PATI, number of intraabdominal injuries and mortality). Seventeen patients died. Most deaths occurred within 1 h after admission due to massive bleeding and severe associated injuries. Only one death was potentially related to the pancreatic injury. Mortality rate also correlated with pancreatic injury grading. Sixty-one patients had associated intraabdominal injuries. Combined pancreaticoduodenal injuries were present in 13 patients, and five died. Simple drainage was the most common procedure performed. Pancreas-related complications were found in 12 out of 47 patients who survived more than 48 h; intraabdominal abscess (n =7) that was associated with colon injuries, and pancreatic fistula (n=5). Conclusion: An approach based on injury grade and location is advised. Routine drainage is recommended; distal resection is indicated in the presence of main duct injury, and the management of severe injuries will be tailored according to the overall physiologic status, presence of associated injuries, and duodenal viability. Morbidity and mortality is mainly due to associated injuries. © 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction Pancreatic injury is infrequent following blunt or penetrating abdominal trauma [1]. The incidence of pancreatic injuries from penetrating mechanisms varies from 5 to 7% [2]. Isolated pancreatic injuries (2%) are uncommon due to its retroperitoneal location and the proximity of multiple organs and major vascular structures [3]. Mortality rates after penetrating trauma vary from 13 to 32%, however, most deaths are related to associated injuries [4]. The diagnosis of a penetrating pancreatic injury is usually made intraoperatively. * Corresponding author. E-mail address: [email protected] (R. Coimbra).
Several aspects in the management of penetrating pancreatic injuries are still controversial, particularly in the era of damage control operation. The optimal and definitive management of severe injuries involving the main pancreatic duct is still open to debate. There is a scarcity of reports solely dedicated to penetrating pancreatic trauma. We report our experience with the management of such injuries over a decade in a level-1 trauma centre, focusing on determinants of complications and deaths according to mechanism and injury severity. 2. Patients and methods We performed a retrospective review of 66 consecutive patients sustaining penetrating pancreatic injury
0020-1383/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 0 - 1 3 8 3 ( 0 1 ) 0 0 0 9 9 - 7
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
754
3. Results
treated during an 11-year period at the University of California San Diego Medical Centre, a level I Trauma Centre. Four patients were excluded because of missing records, leaving a total of 62 patients, which corresponds to 5.2% of all penetrating abdominal trauma patients (n=1192) treated during the study period. All clinical records, as well as autopsy data, when applicable, were reviewed. We recorded the following variables: age, gender, mechanism of injury, shock upon admission, anatomic location of the pancreatic injury, associated intra and extra abdominal injuries, classification of pancreatic injury, surgical management, hospital length of stay, complications and mortality. The penetrating abdominal trauma index (PATI) was calculated as described by Moore et al. [5] and Borlase et al. [6]. The injury severity score (ISS) was calculated according to Baker et al. [7]. Pancreatic injury classification was determined according to the Organ Injury Scaling (OIS) of the American Association for the Surgery of Trauma (AAST) [8]. In brief, grade I included minor contusion or superficial laceration without duct injury, grade II was defined as major contusion or laceration without duct injury or tissue loss, grade III included distal transection or parenchymal injury with duct injury, grade IV included proximal (right of the superior mesenteric vein) transection or parenchymal injury involving the ampulla, and grade V was a massive disruption of the pancreatic head. Pancreatic fistula was defined as drainage of more than 50 cc/day of fluid with elevated amylase level for at least 3 consecutive days [2]. Pancreatitis was defined as elevated serum amylase for more than 3 days associated with clinical findings [2]. Statistical analysis was performed using non-parametric Mann– Whitney, x2 or Student’s t-test when appropriate. A P value of B 0.05 was considered statistically significant.
Fifty-four patients (87.1%) were male and the mean age was 28.5 years (range from 16 to 56). Thirty patients (48.4%) sustained gunshot wounds, 27 (43.5%) stab wounds, and five (8.1%) shotgun wounds. The diagnosis of pancreatic injury was made during autopsy in nine patients, and during laparotomy in the remaining 53 patients. Twenty nine patients had systolic blood pressureB 90 mmHg on admission, and were considered in shock. Patients admitted in shock had significantly higher risk of dying when compared to haemodynamically stable patients (PB0.0001).
3.1. Pancreatic injury location and grading The body of the pancreas was the most common injury site (n=21), followed by the head (n= 19), and tail (n= 19). Two patients had injuries in more than one pancreatic region. Grade 1 pancreatic injuries were the most frequent (Table 1). Mortality rate correlated with pancreatic injury severity. Mortality rate for grade 1 injuries was 12.5%, increasing to 66.6% in grade 4 injuries and 100% in grade 5 injuries.
3.2. Trauma scores The anatomic injury severity scores correlated with mortality. PATI was calculated for a total of 53 patients. Mean PATI was 35.69 22.4 (range: 5– 104). Non-survivors had higher PATI values than survivors (57.8 vs. 31.6; PB 0.01). Mean ISS was 28.2917.3 (range: 4–75). Non-survivors had higher ISS than survivors (42.82 vs. 22.66; PB 0.01).
3.3. Associated injuries Associated intraabdominal injuries were present in 61 patients (98.4%). The number of organs injured varied from 1 to 9, with a mean of 3.6 per patient. The stomach was the most frequent organ injured in associ-
Table 1 Grade of pancreatic injury and survival (n = 62) Grade
Grade Grade Grade Grade Grade N/A Total
Alive
1 2 3 4 5
21 12 11 1 0 0 45
Dead B48 h
\48 h
3 4 4 1 1 2 15
0 1 0 1 0 0 2
Total (%)
Mortality (%)
24 17 15 3 1 2 62
3 5 4 2 1 2 17
(38.7) (27.4) (24.2) (4.8) (1.6) (3.2) (100.0)
(12.5) (29.4) (26.6) (66.6) (100.0) (100.0) (27.4)
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759 Table 2 Associated intra abdominal injuries (n= 62) Intraabdominal organs
Stomach Liver Diaphragm Kidney Major vascular Minor vascular Spleen Small bowel Large bowel Duodenum Extrahepatic biliary Ureter
755
3.4. Surgical management
Mechanism
Total (%)
GSW
SW
SGW
16 21 20 15 14 6 7 9 6 7 1 0
12 6 5 7 7 11 7 2 5 3 2 1
5 5 2 2 2 2 2 3 3 3 2 0
33 32 27 24 23 19 16 14 14 13 5 1
(53.2) (51.6) (43.5) (38.5) (37.1) (30.6) (25,8) (22.6) (22.6) (21.0) (8.1) (1.6)
GSW, gunshot wound; SW, stab wound; SGW, shotgun wound.
ation with the pancreas (Table 2). The number of associated intra abdominal injuries was higher in patients sustaining gunshot wounds and shotgun injuries (4.0 and 6.2, respectively) than in patients with stab wounds (2.6; PB 0.01). Also, patients who died had a higher number of associated intraabdominal injuries than survivors (4.41 and 3.24, respectively, P B0.05). Combined pancreaticoduodenal injuries were present in 13 patients and 5 (38.5%) died. One patient died shortly after arrival and did not receive any surgical treatment. Most duodenal injuries were primarily repaired, and 2 patients underwent pyloric exclusion. Extra-abdominal associated injuries were found in 37 of 62 patients (59.7%), and included thoracic injuries (n= 24), soft tissue lesions (n =16), bone fractures (n= 5), spinal/vertebral column injuries (n = 4) and neck/facial injuries (n= 3).
Resuscitative thoracotomy was performed in ten patients admitted ‘in extremis’. All patients who underwent a resuscitative thoracotomy died, eight intraoperatively and two within 48 h postoperatively. Exploratory laparotomy was performed in 53 patients, 15 of whom required reexploration for various indications. Gunshot and shotgun wound victims underwent mandatory abdominal exploration. Stab wound patients underwent surgical exploration based on different indications, including evisceration of omentum (n= 11), hypotension (n= 7), positive diagnostic peritoneal lavage (n= 3), clinical diagnosis of peritonitis (n=3), blood in the nasogastric tube (n=2), and positive fascial penetration on digital wound exploration (n= 1). Surgical management varied depending on location and grade of the pancreatic injury. The most common procedure was simple drainage with closed suction drains (Table 3). The type of closure of the pancreatic stump following partial resection was with nonabsorbable suture (n = 10), staples with nonabsorbable suture (n =1), and staples alone (n=2). There was no difference in the development of pancreatic fistula or intraabdominal abscess when comparing different modalities of pancreatic stump closure. Only one out of 13 patients who underwent distal pancreatic resection did not undergo splenectomy.
3.5. Complications Pancreas-related complications were found in 12 out of 47 patients (29.7%) who survived more than 48 h (Table 4). These included six patients out of 27 with stab wounds (22.2%), five out of 17 with gunshot wounds (29.41%), and one out of 3 with shotgun wounds (33.33%). Patients who developed pancreas-re-
Table 3 Surgical management and grade of pancreatic injury (n = 62) Surgical procedure
No surgical procedure for pancreas Drainage Distal resection and Drainage Suture ligation and drainage Debridement and drainage Distal resection, duct ligation, and drainage Pyloric exclusion and drainagea Local bleeding control Pancreaticoduodenectomy Total a
Grade of injury N/A
1
2
3
4
5
Total
2 0 0 0 0 0 0 0 0 2
4 13 0 4 1 0 1 1 0 24
5 3 1 3 3 1 1 0 0 17
2 0 11 0 0 2 0 0 0 15
1 1 0 0 0 0 0 0 1 3
0 0 0 0 0 0 0 0 1 1
14 17 12 7 4 3 2 1 2 62
Pyloric exclusion performed because of the severity of duodenal injury.
756
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
Table 4 Pancreas-related complications and mechanism of injury in patients surviving more than 48 h (n =47) Complication
Intraabdominal abscess Pancreatic fistula Pancreatic hemorrhage Pseudocyst Pancreatitis
Mechanism
Mean PATI
GSW
SW
SGW
Total
3 2 1 0 2
3 2 2 2 0
1 1 1 0 0
7 5 4 2 2
53.8 57.4 59.3 18.5 32.0
Some patients had more than one complication. GSW, gunshot wound; SW, stab wound; SGW, shotgun wound; PATI, penetrating abdominal trauma index.
lated complications had longer hospital stay (31.4 vs. 15.8 days; PB0.05). Intraabdominal abscess was the most frequent postoperative complication. Pancreas-related complications according to injury severity are listed in Table 5. Non pancreatic-related complications were present in 22 (46.8%), and included atelectasis (n = 7), pneumonia (n = 7), urinary tract infection (n= 4) and wound infection (n = 3).
3.6. Intraabdominal abscess Seven patients developed intraabdominal abscesses between postoperative day 9 and 30, and the mean length of stay was 45 days. Mean PATI values in this group were higher (53.8 vs. 27.5, P B0.05), stressing the importance of associated injuries in the genesis of intraabdominal abscesses.. All cultures from the abscess were positive, many of them with more than one species, and included Enterobacter aerogenes, Enterococcus faecalis, Bacteroides sp., Pseudomonas aeruginosa, Candida sp., and Fusobacterium sp. A higher incidence of intraabdominal abscesses was found in the subgroup of patients with associated colonic injuries (PB 0.05). Treatment included percutaneous CT or ultrasound guided drainage in five and surgical drainage in two patients.
3.7. Pancreatic fistula Pancreatic fistula was diagnosed in five patients between postoperative day 1 and 25, (mean of 9 days), and the mean length of stay was 39.2 days. Two patients had sinograms that revealed communication with the pancreatic duct. External drainage was the treatment in all. Octreotide was used in four patients, with decrease in output. One patient, who was HIV positive, died 23 days after surgery due to sepsis and multiorgan failure. All other patients were discharged in a healthy state, and the fistulas closed in an average period of 38.3 days. Ligation of the main pancreatic duct following distal pancreatic resection did not influence the incidence of
pancreatic fistulas or intraabdominal abscesses (P= 0.20).
3.8. Postoperati6e hemorrhage Four patients developed postoperative pancreatic haemorrhage. In three, bleeding occurred within the first 24 h post surgery and one patient bled on postoperative day 19. Two patients presented with bloody drainage, hypotension and low haematocrit. Another patient became hypotensive and developed a coagulopathy shortly after the initial operation requiring reexploration. The remaining patient became hypotensive, had abdominal distention and low haematocrit. An abdominal angiogram was obtained but it did not reveal the source of bleeding. The patient was re-explored and bleeding from the pancreatic parenchyma was found. All these patients received 4–26 U of blood (mean of 15.8 U), and the bleeding site was controlled surgically. Intraoperative findings included bleeding from small pancreatic vessels in three and at the splenic artery stump in one.
3.9. Posttraumatic pancreatic pseudocyst Two patients developed pancreatic pseudocysts diagnosed on postoperative days 12 and 19, respectively. They had been discharged in good condition, but reTable 5 Grade of pancreatic injury and pancreas-related complications in patients surviving more than 48 h (n = 47) Grade of injury
Grade 1 Grade 2 Grade 3 Grade4 Total
Complications A
B
C
D
E
n/Total
0 2 0 0 2
0 1 2 1 4
1 1 0 0 2
1 2 4 0 7
0 0 4 1 5
2/21 5/12 4/12 1/2 12/47
Some patients had more than one complication. A, pseudocyst; B, hemorrhage; C, pancreatitis; D, abscess; E, fistula.
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
turned to the hospital complaining of abdominal pain associated with nausea and vomiting. Abdominal CT scan confirmed the diagnosis. Surgical drainage was not required, and they resolved after total parenteral nutrition therapy for 3 weeks.
3.10. Pancreatitis Two patients developed pancreatitis on postoperative days 10 and 14. The first patient had persistent high serum amylase, with mild epigastric abdominal pain. Treatment included enteral nutrition through a jejunostomy tube and blood amylase levels returned to normal values within a week. The second patient presented with abdominal pain, nausea, vomiting and persistently elevated serum and urinary amylase levels. Abdominal CT demonstrated pancreatic oedema. Treatment included total parenteral nutrition, with good outcome. No endocrine or exocrine pancreas-related complications were identified.
3.11. Mortality Seventeen patients died (27.4%). Ten died within 1 h of arrival at the hospital. The mortality rate for gunshot, shotgun, and stab wounds was 43.3, 40.0, and 7.7%, respectively. Sixteen patients died from non-pancreatic causes, hypovolaemic shock (n = 14) being the most common culprit. Cardiac or major vessel injury was present in 13, and profuse liver bleeding in one patient. The remaining two patients died of multiorgan failure. Only one death was potentially related to the pancreatic injury. This was an HIV positive patient who sustained injuries to the inferior vena cava, spleen and head of the pancreas (Grade IV). In view of the injury severity, it was decided during laparotomy not to proceed with pancreaticoduodenectomy, but to place drains in order to have a controlled pancreatic fistula. On postoperative day 1, the patient became hypotensive requiring surgical intervention. Bleeding from minor pancreatic vessels was successfully controlled, however the patient subsequently died of sepsis and multiorgan failure on postoperative day 22. In retrospect, the repeated episodes of hypotension might have contributed to the development of sepsis and multiple organ failure in this patient. Thirteen patients died in the operating room, and four patients died during the postoperative period. Of note, a total of nine patients (14.5%) were found to have pancreatic injury during post-mortem examination. These patients had massive bleeding form associated injuries and all expired before a thorough evaluation of all intra abdominal injuries could be performed. The mean hospital stay for patients surviving more than 48 h (n =47) was 20.2 days.
757
4. Discussion The prevalence of pancreatic injury among penetrating abdominal trauma victims in our series was 5.20%, which is within the range (5– 7%) reported in the literature [9]. Shotgun and gunshot injuries were more destructive than stab wounds as reflected by higher PATI, ISS, number of intra abdominal organs injured and mortality. Due to its retroperitoneal location, the pancreas is surrounded by important organs and vessels, which contributes to the rarity of isolated pancreatic injury (1.6% in our series), and reported in 2% of cases of penetrating pancreatic trauma [10]. Death is mainly caused by associated injuries rather than by the pancreatic injury [11]. The diagnosis of a pancreatic injury is usually made intraoperatively, once most of these patients have associated injuries (3.6 injuries/patient in the present series), and present with clinical signs and symptoms of intraabdominal injuries requiring surgical intervention. Appropriate surgical management of penetrating pancreatic injuries depends on location (central vs. peripheral, head, body or tail, right or left of the mesenteric vessels), the extent of the injury, and the presence of main pancreatic duct injury. Injuries to adjacent organs should raise the suspicion of a pancreatic injury. A thorough evaluation of the pancreatic parenchyma and exploration of peripancreatic haematomas is mandatory to adequately assess ductal integrity and injury severity. Different modalities of surgical treatment for pancreatic injuries have been proposed, but the use of a graded approach to its management is associated with a significant reduction in morbidity and mortality [12]. In our series, grade 1 and 2 injuries (no ductal injuries) were usually treated with closed drainage with or without local debridement. This approach proved to be effective, safe, and is the most commonly performed [13,14]. It seems that closed suction drainage as opposed to sump or open drainage decreases the incidence of pancreas-related septic complications [13], and is the current method of choice. Duration of drainage, however, is still a matter of controversy. Some authors recommend drainage for 7– 10 days or until the patient is tolerating an oral diet [15]. It seems that this approach is appropriate for grades 3– 5 injuries; however, in minor injuries drains should be removed earlier, if drainage is minimal or ceases. All grade III injuries in the present series were treated by distal resection and drainage. The integrity of the pancreatic duct has been suggested as an important factor in determining the method of treatment [12]. This may be achieved by visual inspection in distal injuries of the pancreas, but it is difficult to assess in the pancreatic head where intraoperative pancreatography has been recommended [16]. Several methods have been
758
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
described for intra operative evaluation of the main pancreatic duct if direct visualisation is inconclusive: distal pancreatectomy and direct catheterisation of the main duct in the tail of the pancreas, cannulation of the major duodenal papilla through a surgically created duodenotomy, injection of contrast media in the gallbladder (cholecystopancreatography), and endoscopic retrograde cholangiopancreatography. All these techniques have limitations and the associated complication rate is significant. How aggressively one should pursue an intra operative diagnosis of main pancreatic duct injury is still debatable. We have adopted a more conservative approach. If by visually inspecting the pancreas and eventually stimulating pancreatic secretion with secretin intra operatively no duct injury is identified, we have opted for drainage rather than pursuing intra operative radiologic evaluation. Distal pancreatectomy is the treatment of choice for grade 3 injuries. A decrease in the number of postoperative complications following distal pancreatectomy has been shown when compared to drainage alone in haemodynamically stable patients [1,2,17,18]. Haemodynamically unstable patients with multiple associated injuries and physiologic derangement (acidosis, coagulopathy, and hypothermia) are best served by damage control operation and pancreatic drainage. Splenic preservation after partial pancreatic resection has been mainly proposed for children [2]. In our series, this was done in only one out of 13 patients who underwent distal pancreatectomy for grade 3 injuries. As reported previously by Degiannis et al., [19], we found that the type of suture used to approximate the edges of the pancreatic stump was unrelated to the development of postoperative complications. Grade 4 injuries should be treated by distal resection, when feasible, or drainage alone. In very special circumstances, such as in resection of more than 80% of the pancreatic parenchyma and associated diabetes mellitus, preservation of the distal pancreas and internal drainage by Roux-en-Y jejunal anastomosis should be considered. Grade 5 injuries characterised by massive devascularisation of the pancreatic head and adjacent duodenum are infrequent, usually require pancreaticoduodenectomy or duodenal diverticulisation, and are accompanied by elevated mortality rates [20,21]. A more conservative approach consisting of extensive peripancreatic drainage and primary repair of the duodenum is indicated in unstable critically ill patients. In our series, complex procedures such as pancreaticoduodenectomy were performed in only two patients. Pancreaticoduodenectomy entails very high mortality rates (30–40%) and is not recommended for acute situations unless the missile has effectively performed the resection and the operation is limited to the debridement of devitalised tissue [18]. Combined pancreaticoduodenal injuries carried a high mortality rate (38.5%) and were
most often treated by primary repair. Pyloric exclusion has been suggested as a valuable method in the treatment of severe duodenal injuries, leading to a reduction in the postoperative leakage rate, and it has been also considered a viable option for grade 4–5 pancreatic injuries associated with duodenal injuries [22]. If the duodenum or the main pancreatic duct is not involved, minor injuries to the head of the pancreas should be treated by simple drainage [23,24]. In general, the postoperative complication rate increases as the injury severity increases. Other factors that contribute to the development of pancreas-related complications include main duct injury and associated duodenal injuries [4,12,25]. Intra abdominal abscess formation was the most frequent postoperative complication (15%) in patients surviving more than 48 h. Increase PATI in this group indicates the importance of associated injuries in the genesis of this complication. As previously reported by others [17], we also found that combined pancreatic and colon injuries were associated with the development of intraabdominal abscesses. Adequate treatment of intra abdominal abscesses includes percutaneous or surgical drainage. The incidence of pancreatic fistula was 10.6% according to the definition used in this study (drainage of more than 50 cc/day of amylase-rich fluid for more than 3 days), and it is within the range of 5–30% reported in the literature [17,21]. All patients were treated non-operatively, one patient died, and in the remaining four patients, closure occurred within 6 weeks. Octreotide has been used for treatment of this complication, but no significant improvement in outcome has been found [26]. Four patients received octreotide in an attempt to decrease pancreatic exocrine secretion. In all 4 patients the fistulas resolved but a direct effect of this drug on fistula closure cannot be established. Pancreatic pseudocysts are more frequently seen following blunt trauma, and are often the result of inadequate pancreatic drainage [27]. The incidence was 4.5% in the present series. Usually patients with pseudocyst present late, as occurred in both patients in this series, who were initially discharged and then returned to the hospital with symptoms. Both patients had good outcome after receiving total parenteral nutrition and none required operative intervention. The overall mortality rate of 27.4% is in accordance with data previously reported in the literature [4,9,17,25], and it includes patients who arrived in extremis to the hospital and died shortly thereafter. The mortality was directly related to associated injuries, as reflected by a higher PATI, ISS and number of intraabdominal organs injured when compared to survivors. Most patients died within 24 h after admission, mainly due to associated cardiac or vascular injuries with
J.C. Vasquez et al. / Injury, Int. J. Care Injured 32 (2001) 753–759
resultant hypovolaemia. In only one patient the cause of death was possibly related to the pancreatic injury. Pancreas-related deaths are uncommon, usually occur late, and are related to sepsis and multiorgan failure [3,4,25,28,29]. We conclude that penetrating pancreatic injuries are associated with elevated mortality and morbidity rates. An approach based on injury grade and location is advised. Routine drainage is recommended despite injury severity; distal resection is indicated in the presence of main duct injury, and the management of severe (grade 4–5) injuries will be tailored according to the overall physiologic status, presence of associated injuries, and duodenal viability. Most complications and deaths are related to associated injuries.
References [1] Wisner DH, Wold RL, Frey CF. Diagnosis and treatment of pancreatic injuries. Arch Surg 1990;125:1109 –13. [2] Cogbill TH, Moore EE, Morris JA Jr, Hoyt DB, et al. Distal pancreatectomy for trauma: a multicenter experience. J Trauma 1991;31:1600 – 6. [3] Jones RC. Management of pancreatic trauma. Am J Surg 1985;150:698 – 704. [4] Young PR, Meredith JW, Baker CC, et al. Pancreatic injuries resulting from penetrating trauma: a multi-institution review. Am Surg 1988;64:838 –44. [5] Moore EE, Dunn EL, Moore JB, Thompson JS. Penetrating abdominal trauma index. J Trauma 1981;21:439 –45. [6] Borlase BC, Moore EE, Moore FA. The abdominal trauma index: a critical reassessment and validation. J Trauma 1990;30:1340 – 4. [7] Baker SP, O’Neill B, Haddon W Jr, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974;14:187 – 96. [8] Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, et al. Organ Injury Scaling, II: pancreas, duodenum, small bowel, colon and rectum. J Trauma 1990;30:1427 – 9. [9] Jones RC, Foreman ML. Pancreas. In: Ivatury R, Cayten CG, editors. The Textbook of Penetrating Trauma. Media, Pennsylvania: Williams and Wilkins, 1996:631 –42. [10] Jones RC. Management of pancreatic trauma. Ann Surg 1978;187:555 – 64.
759
[11] Stone HH, Fabian TC, Satiani B, Turkleson ML. Experiences in the management of pancreatic trauma. J Trauma 1981;21:257 – 62. [12] Smego DR, Richardson JD, Flint LM. Determinants of outcome in pancreatic trauma. J Trauma 1985;25:771 – 6. [13] Fabian TC, Kudsk KA, Croce MA, Payne LW, Mangiante E, Voeller GR, Britt LG. Superiority of closed suction drainage for pancreatic trauma. Ann Surg 1990;211:724 – 30. [14] Madiba TE, Mokoena TR. Favorable prognosis after surgical drainage of gunshot, stab or blunt trauma of the pancreas. Br J Surg 1995;82:1236 – 9. [15] Northrup WF, Simmons RL. Pancreatic trauma: a review. Surgery 1972;71:27 – 43. [16] Berni GA, Bandyk DF, Oreskovich MR, Carrico CJ. Role of intraoperative pancreatography in patients with injury to the pancreas. Am J Surg 1982;143:602 – 5. [17] Ivatury RR, Nallathambi M, Rao P, Stahl W. Penetrating pancreatic injuries. Am Surg 1990;56:90 – 5. [18] Wilson RH, Moorehead RJ. Current management of trauma to the pancreas. Br J Surg 1991;78:1196 – 202. [19] Degiannis E, Levy RD, Potokar T, Lennox A, Rowse A, Saadia R. Distal pancreatectomy for gunshot injuries of the distal pancreas. Br J Surg 1995;82:1240 – 2. [20] Mansour MA, Moore JB, Moore EE, Moore FA. Conservative management of combined pancreatoduodenal injuries. Am J Surg 1989;158:531 – 5. [21] Wynn M, Hill D, Miller DR, Waxman K, Eisner ME, Gazzaniga AB. Management of pancreatic and duodenal trauma. Am J Surg 1985;150:327 – 32. [22] Degiannis E, Krawczykowski D, Velmahos GC, Levy RD, Souter I, Saadia R. Pyloric exclusion in severe penetrating injuries of the duodenum. World J Surg 1993;17:751 – 4. [23] Lewis G, Krottenbelt JD, Krige JEJ. Conservative surgery for trauma to the pancreatic head: is it safe? Injury 1991;22:372 –4. [24] Feliciano DV, Martin TD, Curse PA, Graham JM, Buch JM, Mattox KL, Bitondo CG, Jordan GL. Management of combined pancreatoduodenal injuries. Ann Surg 1987;205:673 – 80. [25] Akhrass R, Yaffe MB, Brandt CP, et al. Pancreatic trauma: A ten-year multi-institutional experience. Am Surg 1997;63:598 – 604. [26] Nwariaku FE, Terracina A, Mileski WJ, Minei JP, Carrico CJ. Is octreotide beneficial following pancreatic injury? Am J Surg 1995;170:582 – 5. [27] Graham JM, Mattox KL, Jordan GL. Traumatic injuries of the pancreas. Am J Surg 1978;136:744 – 8. [28] Flynn W Jr, Cryer HG, Richardson JD. Reappraisal of pancreatic and duodenal injury management based on injury severity. Arch Surg 1990;125:1539 – 41. [29] Sorensen VJ, Obeid FN, Horst HM, Bivins BA. Penetrating pancreatic injuries, 1978 – 1983. Am Surg 1986;52:354 – 8.