- Email: [email protected]
Detection of traumatic pancreatic duct disruption in the modern era
The American Journal of Surgery xxx (2018) 1e5
Contents lists available at ScienceDirect
The American Journal of Surgery journal homepage: www.americanjournalofsurgery.com
Detection of traumatic pancreatic duct disruption in the modern era Morgan Schellenberg, Kenji Inaba*, James M. Bardes, Vincent Cheng, Kazuhide Matsushima, Lydia Lam, Elizabeth Benjamin, Demetrios Demetriades Division of Trauma and Surgical Critical Care, LACþUSC Medical Center, University of Southern California, Los Angeles, CA, 2051 Marengo Street, IPT C5L100, Los Angeles, CA 90033, United States
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 July 2017 Received in revised form 23 May 2018 Accepted 2 June 2018
Background: Pancreatic trauma management hinges upon the presence or absence of pancreatic duct injury, but the optimal method of assessment is unclear. This study endeavored to evaluate the methods of pancreatic duct assessment in modern practice. Methods: Patients presenting to LAC þ USC Medical Center (01/2008e06/2015) with a pancreatic injury were identified (ICD-9 codes). Demographics, clinical data, technique of duct evaluation, and outcomes were analyzed. Results: 71 patients with pancreatic injury were identified. 21 patients (30%) underwent CT scan (sensitivity 76%). Sixteen (76%) then underwent laparotomy while 5 (24%) were managed successfully nonoperatively. Most (n ¼ 50, 70%) underwent immediate laparotomy. Overall, 66 patients (93%) were managed operatively. The majority were assessed intraoperatively for ductal injury with visual inspection alone (n ¼ 62, 94%). Four (6%) underwent intraoperative pancreatography via duodenotomy/cholecystotomy, which were all inconclusive. Conclusion: In the evaluation of pancreatic duct injury, intraoperative pancreatography is frequently inconclusive and should have a limited role. Clinical suspicion for ductal injury based on intraoperative visual inspection alone should guide the management of pancreatic injuries. Summary for Table of Contents: The management of pancreatic trauma hinges upon the integrity of the main pancreatic duct. The optimal strategy for identification of pancreatic duct injury was previously unknown. This study found that intraoperative visual inspection alone is sufficient to guide the management of pancreatic injuries. Intraoperative pancreatography is unnecessary. © 2018 Published by Elsevier Inc.
Keywords: Trauma Pancreatic trauma Pancreatectomy Intraoperative pancreatography
Introduction Pancreatic injuries resulting from trauma are notoriously challenging to both diagnose and manage. They are uncommon, especially those that involve the pancreatic duct. They are also frequently associated with other injuries (>90% in modern series) and carry an inherent lethality (16e17% mortality) due to the anatomic region surrounding the pancreas.1 As a result, the literature on the optimal work-up of patients with suspected pancreatic duct injury is lacking. This is problematic because the management
* Corresponding author. E-mail addresses: [email protected] (M. Schellenberg), [email protected] (K. Inaba), [email protected] (J.M. Bardes), [email protected] (V. Cheng), [email protected] (K. Matsushima), [email protected] (L. Lam), [email protected]. edu (E. Benjamin), [email protected] (D. Demetriades).
of pancreatic injury is directed by the status of the duct. In order to secure the diagnosis of pancreatic duct injury, techniques such as computed tomography (CT) scan or magnetic resonance cholangiopancreatography (MRCP) can be performed outside of the operating room (OR) but are only suitable for stable patients. The current literature on CT scan and MRCP indicates that both of these methods have poor sensitivity in the diagnosis of duct injuries in trauma.2e4 Intraoperative techniques for duct evaluation include a variety of methods of pancreatography, including ontable endoscopic retrograde cholangiopancreatography (ERCP) and pancreatography via cholecystotomy or cannulation of the pancreatic duct, either through a duodenotomy to access the ampulla of Vater or by direct cannulation through injured pancreatic parenchyma. The ability of these tests to accurately define the status of the duct is unclear, as is the frequency with which these tests are utilized in modern practice. Consequently, at this time, the optimal method for diagnosing a pancreatic duct injury has not
https://doi.org/10.1016/j.amjsurg.2018.06.002 0002-9610/© 2018 Published by Elsevier Inc.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
2
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
been established. Therefore, the primary objective of this study was to examine the various investigations used to evaluate the pancreatic duct for injury, including the frequency of use, ability to correctly identify the presence or absence of ductal injury, and rates of associated complications. The secondary objectives of the study were to examine the patient demographics, injury data, surgical management, and outcomes associated with pancreatic injury in a highvolume trauma center. Our hypothesis was that invasive and time-consuming methods of duct evaluation, including all forms of intraoperative pancreatography, are rarely used in trauma patients and that simple visual inspection of the injured gland is sufficient to confirm or exclude the diagnosis of pancreatic duct injury in the majority of patients. Material and methods Institutional Review Board approval was granted from the University of Southern California. Between January 2008 and June 2015, all patients who sustained a pancreatic injury (AAST grades IV) as defined by ICD-9 codes (863.81e863.84, 863.91e863.94)5 were retrospectively identified from the LAC þ USC Medical Center trauma registry. Both blunt and penetrating mechanisms were included. There were no exclusion criteria. Patient charts were then reviewed. Patient demographics, injury characteristics (mechanism of injury, date of injury, associated injuries, AAST grade of pancreatic injury, Injury Severity Score [ISS]), investigations (CT scan, MRCP, ERCP), surgical procedures, technique of pancreatic duct evaluation, and outcomes (mortality, complications, failure of non-operative management, hospital length of stay, intensive care unit [ICU] length of stay, and need for mechanical ventilation) were abstracted. At LAC þ USC, patients who present after penetrating abdominal trauma are brought emergently to the operating room if they are hemodynamically unstable, have evidence of peritonitis or evisceration, or are unevaluable. Otherwise, these patients undergo CT scan (64-slice [0.5 mm] multidetector system; Aquilion 64 CFX Multislice CT Scanner; Toshiba Medical Systems Corporation, Japan). Intravenous contrast (Omnipaque 350; GE Healthcare, Princeton, NJ) is used routinely. Data collection was performed using a computerized spreadsheet (Microsoft Excel 2007; Microsoft Corporation; Redmond, WA) and analyzed using SPSS Statistics 23 (IBM Corporation; Armonk, NY). Descriptive statistics were calculated for the clinical variables defined above. Continuous variables are presented as mean ± standard deviation; median (range). Categorical variables are presented as n (%). Results Patient demographics and injury data Seventy-one patients with a pancreatic injury were identified over the study period. The mean age was 35 years (range 11e87) and 86% (n ¼ 61) were male (Table 1). The mechanism of injury was penetrating in 51 patients (72%), of which 41 (80%) were gunshot wounds (GSW) and 10 (20%) were stab wounds (SW). The remaining 20 patients (28%) were injured by blunt mechanisms: 9 (45%) as auto vs pedestrian, 7 (35%) in a motor vehicle collision, and 4 (20%) in a motorcycle crash. The mean AAST pancreatic injury grade was II (range I-V) (Table 1). There were 22 patients (31%) with AAST grade I injuries, 17 (24%) with grade II, 29 (41%) with grade III, 2 (3%) with grade IV, and 1 (1%) with grade V. Associated injuries were common (n ¼ 70,
Table 1 Patient Demographics, Clinical Data, Management, and Outcomes. Continuous variables are presented as mean ± standard deviation; median (range). Categorical variables are presented as n (%). Patients (n ¼ 71) Demographics Age, years Gender, male Mechanism of Injury4 (20%) Penetrating GSW SW Blunt AVP MVC MCC Injury Data Injury Severity Score (ISS) AAST Pancreatic Injury Grade I II III IV V Management Immediate Laparotomy CT Scan Other Pre-Operative Duct Assessment Nonoperative Management (NOM) Total Operative Management Intra-Operative Duct Evaluation VISUAL INSPECTION PANCREATOGRAPHY ERCP Duodenotomy Cholecystotomy Outcomes Mortality Hospital LOS ICU LOS Ventilator Days Need for ICU Need for Mechanical Ventilation Failed NOM Pancreatic Complication*
35 ± 17; 29 (11e87) 61 (86%) 51 (72%) 41 (80%) 10 (20%) 20 (28%) 9 (45%) 7 (35%) 4 (20%) 24 ± 13; 20 (4e75) 2 ± 1; 2 (1e5) 22 (31%) 17 (24%) 29 (41%) 2 (3%) 1 (1%) 50 (70%) 21 (30%) 0 (0%) 5 (7%) 66 (93%) 62 (94%) 4 (6%) 0 (0%) 1 (2%) 3 (5%) 14 (20%) 25 ± 26; 15 (1e131) 12 ± 16; 5 (0e78) 5 ± 9; 1 (0e39) 59 (83%) 40 (56%) 0 (0%) 10 (14%)
GSW, gunshot wound. SW, stab wound. AVP, auto vs. Pedestrian. MVC, motor vehicle collision. MCC, motorcycle crash. CT, computed tomography. ERCP, endoscopic retrograde cholangiopancreatography. LOS, length of stay in days. ICU, intensive care unit.*Pancreatic Complications included leak, abscess, fistula.
99%), most frequently the stomach (n ¼ 31, 44%), liver (n ¼ 24, 34%), colon (n ¼ 19, 27%), kidney (n ¼ 19, 27%), and spleen (n ¼ 19, 27%)
Table 2 Associated injuries. Injury
Patients (n ¼ 71)
Percentage (%)
Any Stomach Liver Colon Kidney Spleen Duodenum Diaphragm Small Bowel IVC PV/SMV/IMV Gallbladder Aorta Celiac Axis/SMA Extra-Abdominal
70 31 24 19 19 19 14 12 9 5 5 4 3 1 19
99 44 34 27 27 27 20 17 13 7 7 6 4 1 27
IVC, inferior vena cava. PV/SMV/IMV, portal vein/superior mesenteric vein/inferior mesenteric vein. SMA, superior mesenteric artery.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
(Table 2). The mean Injury Severity Score (ISS) was 24 (range 4e75) (Table 1).
Management Twenty-one patients (30%) underwent CT scan, of which 16 (76%) demonstrated pancreatic injury (Table 1, Fig. 1). The 5 patients (24%) who had a falsely negative CT scan for pancreatic injury were brought for laparotomy for the management of associated injuries and then found intraoperatively to have a pancreatic injury. Of the 16 patients who had a CT scan that was positive for pancreatic injury, 11 (69%) underwent laparotomy and 5 (31%) underwent non-operative management. These 5 patients all sustained blunt trauma and were managed successfully nonoperatively without complication or mortality. One of the patients managed non-operatively (20%) underwent MRCP after CT scan, which demonstrated pancreatic injury without ductal disruption. The remaining 4 patients did not undergo further workup after CT scan to rule out pancreatic duct injury. Most patients (n ¼ 50, 70%) underwent immediate laparotomy. When combined with the patients who underwent preoperative CT scan, a total of 66 patients (93%) were managed operatively. The overwhelming majority were assessed intraoperatively for ductal injury with visual inspection alone (n ¼ 62, 94%). Four patients (6%) underwent intraoperative pancreatography. Of these, one patient, who sustained a grade II injury to the head of the pancreas following GSW, had a duodenotomy created to cannulate the ampulla. Methylene blue was injected to interrogate the duct, which was inconclusive for duct injury. Three patients with grade II pancreatic injuries underwent intraoperative pancreatography using methylene blue via cholecystotomy, all of which were deemed inconclusive. All 4 of these patients were subsequently managed with drain placement alone and no pancreatic complications ensued. No patient underwent intraoperative ERCP or
3
intraoperative ultrasound. At the index operation, patients were most commonly managed with closed suction drainage alone (n ¼ 32, 45%) or distal pancreatectomy (n ¼ 28, 39%) (Table 3). Three patients (4%) had no intervention to the pancreas at laparotomy and 2 patients (3%) underwent a staged trauma Whipple. Patients managed with distal pancreatectomy or trauma Whipple resection typically had closed suction drains left in the surgical bed at the end of the procedure as well. Patients with low grade (I-II) injuries were most frequently managed with drain placement alone (n ¼ 31, 79%). Grade III pancreatic injuries (n ¼ 29) were all managed with distal pancreatectomy. This was typically performed at the index operation (n ¼ 28, 97%), except for 1 patient who initially underwent closed suction drain placement alone and then underwent distal pancreatectomy at the first take-back operation on post-operative day 1 (Table 3). Of the three patients with grade IV-V pancreatic injuries, 1 expired intraoperatively before the injury could be addressed. The remaining 2 patients underwent a staged trauma Whipple operation, with the resection performed at the index surgery. Reconstruction was undertaken in both patients on post-operative day 2.
Outcomes Overall mortality was 20% (n ¼ 14) (Table 1). Mean hospital length of stay was 25 days (range 1e131), with 59 patients (83%) requiring ICU admission and 40 (56%) needing mechanical ventilation. Pancreatic complications, defined as the development of a pancreatic leak, fistula, pancreatitis, or peripancreatic abscess, occurred in 10 patients (14%). The mean grade of pancreatic injury among these patients was II (range I-III). Of the 10 patients who developed a pancreatic complication, 4 (40%) underwent drain
Fig. 1. Flow of patients through study. CT, computed tomography. NOM, nonoperative management.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
4
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
Table 3 Surgical management at the index operation.
Nonoperative Management No Intervention at Laparotomy Drain Placement Distal Pancreatectomy Whippleb a b
All Patients (n ¼ 71a)
AAST I-II (n ¼ 39)
AAST III (n ¼ 29)
AAST IV-V (n ¼ 3a)
5 (8%) 3 (4%) 32 (45%) 28 (39%) 2 (3%)
5 (13%) 3 (8%) 31 (79%) 0 (0%) 0 (0%)
0 (0%) 0 (0%) 1 (3%) 28 (97%) 0 (0%)
0 0 0 0 2
(0%) (0%) (0%) (0%) (67%)
One patient expired before the injury could be addressed. A trauma Whipple resection without reconstruction was performed.
placement alone at the index operation and 6 (60%) underwent distal pancreatectomy. The complication was diagnosed on CT scan in 5 patients (50%), based on drain output and biochemistry (i.e. elevated drain fluid lipase) in 3 patients (30%), and using ERCP in 2 (20%) patients. The 2 patients who underwent ERCP were both investigated for a leak with MRCP prior to ERCP. However, both of these MRCPs were inconclusive. In terms of treatment of pancreatic complications, 7 patients (70%) were managed with the drain placed at the index OR, a period of NPO, and a combination of TPN (n ¼ 4) and antibiotics (n ¼ 5); 2 (20%) with a stent placed in the pancreatic duct during ERCP; and 1 (10%) with a period of NPO, TPN, and antibiotics alone. There were no mortalities among the patients with pancreatic complications. Discussion Pancreatic injuries can be challenging to diagnose and manage due to their infrequent occurrence, the retroperitoneal nature of the pancreas, the proximity to major vascular structures, and the severity of concomitant injuries. The AAST injury grading system clearly defines pancreatic injury on the basis of duct injury and the location of the injury along the pancreas.6 The management principles for pancreatic trauma are then based upon this AAST grading.7e10 Although it is critical for proper AAST injury grading and injury management, the optimal method for diagnosing a pancreatic duct injury is not well defined by the current literature. Earlier studies defined intraoperative criteria for the diagnosis of main pancreatic duct injury, emphasizing the importance of intraoperative visual inspection of the gland. These criteria included direct visualization of duct injury, complete transection or disruption of >50% of the gland, central gland perforation, or severe maceration of the pancreas.11 However, early studies also argued for the liberal use of intraoperative pancreatography to interrogate the pancreatic duct for injury.11,12 More recently, literature has begun to emerge suggesting that intraoperative visual inspection of the pancreas without intraoperative pancreatography is sufficient to rule out duct injury.8,13 These studies, performed on patients with pancreatic injury from 1996 to 2009 and 1990e1995, respectively, showed that visual inspection for duct injury and management on the basis of these findings resulted in acceptable rates of pancreatic complications. In both of these studies, however, no patient underwent intraoperative pancreatography and thus the utility of these techniques could not truly be assessed. As the accessibility and resolution of CT scans have improved, and techniques such as MRCP and ERCP have evolved over time, the diagnostic armamentarium for potential pancreatic duct injury has expanded. Furthermore, with significant improvements in prehospital care and shortened transport time since the original studies, more patients with pancreatic injury are presenting alive to hospital now, requiring evaluation of their injuries. In the current study, patients typically sustained a pancreatic injury after penetrating mechanisms, especially gunshot wounds.
They were severely injured with a high mortality rate, and associated injuries occurred almost universally. Therefore, when pancreatic injury is detected, a thorough search for concomitant injuries must ensue. Most patients were brought for immediate laparotomy. A minority of patients, however, underwent CT scan, which was not a sensitive test (76%) for diagnosing pancreatic injuries. The current literature defines the sensitivity of CT scan for diagnosis of pancreatic injuries in the range of 47e76%,2,7,14 including a AAST multicenter study which published a sensitivity of 52% for 64-slice CT scan.2 CT scan is therefore insufficient to rule out pancreatic injury. In our study, approximately a quarter of patients who were found intraoperatively to have a pancreatic injury had a falsely negative CT scan demonstrating a normal pancreas immediately preoperatively. In this series, intraoperative visual inspection alone was the most frequently used method to investigate the pancreas for potential duct injury. However, visual inspection alone may not be perfect. Four patients who were assessed for ductal injury using visual inspection were deemed not to have a ductal injury and were managed with drain placement alone, but subsequently developed pancreatic leaks or abscesses. This may simply be due to the fact that the soft pancreas of a young trauma patient is prone to leak after injury, even if the main pancreatic duct is intact. However, the possibility that a duct injury was missed with visual inspection must be considered. This would place the failure rate of visual inspection for pancreatic duct injury at a maximum of 6% in this series. The majority of patients in this series who experienced a pancreatic complication were managed successfully using a combination of the closed suction drain left at the index surgery, antibiotics, and TPN. Intraoperative drains should therefore be placed routinely after pancreatic injury in order to mitigate the need for additional intervention should a complication occur. Intraoperative pancreatography is rarely used in contemporary practice and, in all cases within this series, was found to be inconclusive. It is invasive, time consuming, and of limited utility in the work-up of pancreatic trauma. In this study, MRCP and ERCP were also used infrequently and played a limited role in patients with pancreatic injury. They were almost exclusively reserved for the diagnosis and management of postoperative complications. One patient who was diagnosed with a low grade pancreatic injury after blunt trauma based on CT scan underwent MRCP to confirm the absence of duct injury. Otherwise, MRCP was only used among two postoperative patients who were suspected clinically of having a pancreatic leak. Both tests failed to show the leak that was then demonstrated and stented with ERCP. No patient underwent intraoperative ERCP. Although there is literature supporting the use of intraoperative ERCP in pancreatic trauma,15 it is a resourceintensive undertaking and this limits its use, particularly in a practice setting such as ours. The limitations of this study include its retrospective, registrybased design. This is a single-center study and the external validity of these results, particularly for lower volume trauma centers,
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
is unknown. It is, however, a large contemporary analysis of current practice patterns at one of the highest volume trauma centers in North America. Future multi-center prospective study is warranted. Conclusions This study demonstrates that the clinician must be cautious when using CT scan to screen trauma patients for pancreatic injury since the sensitivity is poor. Intraoperatively, pancreatography for the evaluation of the pancreatic duct should have a very limited role. In addition to being time consuming and invasive, it was inconclusive in all cases in which it was used. Clinical suspicion for ductal injury based on intraoperative visual inspection alone is sufficient for the management of pancreatic injuries and should be the preferred strategy for evaluation of the pancreatic duct. Funding sources This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflicts of interest The authors have no conflicts of interest to disclose. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References
5
Trauma. seventh ed. New York: The McGraw Hill Companies; 2013:603e619. 2. Phelan HA, Velmahos GC, Jurkovich GJ, et al. An evaluation of multidetector computed tomography in detecting pancreatic injury: results of a multicenter AAST study. J Trauma. 2009;66(3):641e646. 3. Ragozzino A, Manfredi R, Scaglione M, De Ritis R, Romano S, Rotondo A. The use of MRCP in the detection of pancreatic injuries after blunt trauma. Emerg Radiol. 2003;10(1):14e18. 4. Fulcher AS, Turner MA, Yelon JA, et al. Magnetic resonance cholangiopancreatography (MRCP) in the assessment of pancreatic duct trauma and its sequelae: preliminary findings. J Trauma. 2000;48(6):1001e1007. 5. Centers for Disease Control and Prevention. International Classification of Diseases, Ninth Revision (ICD-9); 2009. Available at: http://www.cdc.gov/nchs/icd/ icd9.htm [accessed 28 March 2017]. 6. Moore EE, Cogbill TH, Malangoni MA, et al. Organ injury scaling. Surg Clin North Am. 1995;75(2):293e303. 7. Ho VP, Patel NJ, Bokhari F, et al. Management of adult pancreatic injuries: a practice management guideline from the eastern association for the surgery of trauma. J Trauma. 2017;82(1):185e199. 8. Sharpe JP, Magnotti LJ, Weinberg JA, et al. Impact of a defined management algorithm on outcome after traumatic pancreatic injury. J Trauma. 2012;72: 100e105. 9. Fisher M, Brasel K. Evolving management of pancreatic injury. Curr Opinion Crit Care. 2011;17:613e617. 10. Stawicki SP, Schwab CW. Pancreatic trauma: demographics, diagnosis, and management. Am Surg. 2008;74:1133e1145. 11. Berni GA, Bandyk DF, Oreskovich MR, Carrico CJ. Role of intraoperative pancreatography in patients with injury to the pancreas. Am J Surg. 1982;143: 602e605. 12. Jones RC, Shires GT. Pancreatic trauma. Arch Surg. 1971;102:424e430. 13. Patton Jr JH, Lyden SP, Croce MA, et al. Pancreatic trauma: a simplified management guideline. J Trauma. 1997;43:234e239. 14. Velmahos GC, Tabbara M, Gross R, et al. Blunt pancreaticoduodenal injury: a multicenter study of the research consortium of new england centers for trauma (ReCONECT). Arch Surg. 2009;144(5):413e419. 15. Rogers SJ, Cello JP, Schecter WP. Endoscopic retrograde cholangiopancreatography in patients with pancreatic trauma. J Trauma. 2010;68(3):538e544.
1. Duodenum Biffl WL. Pancreas. In: Mattox KL, Moore EE, Feliciano DV, eds.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
Contents lists available at ScienceDirect
The American Journal of Surgery journal homepage: www.americanjournalofsurgery.com
Detection of traumatic pancreatic duct disruption in the modern era Morgan Schellenberg, Kenji Inaba*, James M. Bardes, Vincent Cheng, Kazuhide Matsushima, Lydia Lam, Elizabeth Benjamin, Demetrios Demetriades Division of Trauma and Surgical Critical Care, LACþUSC Medical Center, University of Southern California, Los Angeles, CA, 2051 Marengo Street, IPT C5L100, Los Angeles, CA 90033, United States
a r t i c l e i n f o
a b s t r a c t
Article history: Received 16 July 2017 Received in revised form 23 May 2018 Accepted 2 June 2018
Background: Pancreatic trauma management hinges upon the presence or absence of pancreatic duct injury, but the optimal method of assessment is unclear. This study endeavored to evaluate the methods of pancreatic duct assessment in modern practice. Methods: Patients presenting to LAC þ USC Medical Center (01/2008e06/2015) with a pancreatic injury were identified (ICD-9 codes). Demographics, clinical data, technique of duct evaluation, and outcomes were analyzed. Results: 71 patients with pancreatic injury were identified. 21 patients (30%) underwent CT scan (sensitivity 76%). Sixteen (76%) then underwent laparotomy while 5 (24%) were managed successfully nonoperatively. Most (n ¼ 50, 70%) underwent immediate laparotomy. Overall, 66 patients (93%) were managed operatively. The majority were assessed intraoperatively for ductal injury with visual inspection alone (n ¼ 62, 94%). Four (6%) underwent intraoperative pancreatography via duodenotomy/cholecystotomy, which were all inconclusive. Conclusion: In the evaluation of pancreatic duct injury, intraoperative pancreatography is frequently inconclusive and should have a limited role. Clinical suspicion for ductal injury based on intraoperative visual inspection alone should guide the management of pancreatic injuries. Summary for Table of Contents: The management of pancreatic trauma hinges upon the integrity of the main pancreatic duct. The optimal strategy for identification of pancreatic duct injury was previously unknown. This study found that intraoperative visual inspection alone is sufficient to guide the management of pancreatic injuries. Intraoperative pancreatography is unnecessary. © 2018 Published by Elsevier Inc.
Keywords: Trauma Pancreatic trauma Pancreatectomy Intraoperative pancreatography
Introduction Pancreatic injuries resulting from trauma are notoriously challenging to both diagnose and manage. They are uncommon, especially those that involve the pancreatic duct. They are also frequently associated with other injuries (>90% in modern series) and carry an inherent lethality (16e17% mortality) due to the anatomic region surrounding the pancreas.1 As a result, the literature on the optimal work-up of patients with suspected pancreatic duct injury is lacking. This is problematic because the management
* Corresponding author. E-mail addresses: [email protected] (M. Schellenberg), [email protected] (K. Inaba), [email protected] (J.M. Bardes), [email protected] (V. Cheng), [email protected] (K. Matsushima), [email protected] (L. Lam), [email protected]. edu (E. Benjamin), [email protected] (D. Demetriades).
of pancreatic injury is directed by the status of the duct. In order to secure the diagnosis of pancreatic duct injury, techniques such as computed tomography (CT) scan or magnetic resonance cholangiopancreatography (MRCP) can be performed outside of the operating room (OR) but are only suitable for stable patients. The current literature on CT scan and MRCP indicates that both of these methods have poor sensitivity in the diagnosis of duct injuries in trauma.2e4 Intraoperative techniques for duct evaluation include a variety of methods of pancreatography, including ontable endoscopic retrograde cholangiopancreatography (ERCP) and pancreatography via cholecystotomy or cannulation of the pancreatic duct, either through a duodenotomy to access the ampulla of Vater or by direct cannulation through injured pancreatic parenchyma. The ability of these tests to accurately define the status of the duct is unclear, as is the frequency with which these tests are utilized in modern practice. Consequently, at this time, the optimal method for diagnosing a pancreatic duct injury has not
https://doi.org/10.1016/j.amjsurg.2018.06.002 0002-9610/© 2018 Published by Elsevier Inc.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
2
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
been established. Therefore, the primary objective of this study was to examine the various investigations used to evaluate the pancreatic duct for injury, including the frequency of use, ability to correctly identify the presence or absence of ductal injury, and rates of associated complications. The secondary objectives of the study were to examine the patient demographics, injury data, surgical management, and outcomes associated with pancreatic injury in a highvolume trauma center. Our hypothesis was that invasive and time-consuming methods of duct evaluation, including all forms of intraoperative pancreatography, are rarely used in trauma patients and that simple visual inspection of the injured gland is sufficient to confirm or exclude the diagnosis of pancreatic duct injury in the majority of patients. Material and methods Institutional Review Board approval was granted from the University of Southern California. Between January 2008 and June 2015, all patients who sustained a pancreatic injury (AAST grades IV) as defined by ICD-9 codes (863.81e863.84, 863.91e863.94)5 were retrospectively identified from the LAC þ USC Medical Center trauma registry. Both blunt and penetrating mechanisms were included. There were no exclusion criteria. Patient charts were then reviewed. Patient demographics, injury characteristics (mechanism of injury, date of injury, associated injuries, AAST grade of pancreatic injury, Injury Severity Score [ISS]), investigations (CT scan, MRCP, ERCP), surgical procedures, technique of pancreatic duct evaluation, and outcomes (mortality, complications, failure of non-operative management, hospital length of stay, intensive care unit [ICU] length of stay, and need for mechanical ventilation) were abstracted. At LAC þ USC, patients who present after penetrating abdominal trauma are brought emergently to the operating room if they are hemodynamically unstable, have evidence of peritonitis or evisceration, or are unevaluable. Otherwise, these patients undergo CT scan (64-slice [0.5 mm] multidetector system; Aquilion 64 CFX Multislice CT Scanner; Toshiba Medical Systems Corporation, Japan). Intravenous contrast (Omnipaque 350; GE Healthcare, Princeton, NJ) is used routinely. Data collection was performed using a computerized spreadsheet (Microsoft Excel 2007; Microsoft Corporation; Redmond, WA) and analyzed using SPSS Statistics 23 (IBM Corporation; Armonk, NY). Descriptive statistics were calculated for the clinical variables defined above. Continuous variables are presented as mean ± standard deviation; median (range). Categorical variables are presented as n (%). Results Patient demographics and injury data Seventy-one patients with a pancreatic injury were identified over the study period. The mean age was 35 years (range 11e87) and 86% (n ¼ 61) were male (Table 1). The mechanism of injury was penetrating in 51 patients (72%), of which 41 (80%) were gunshot wounds (GSW) and 10 (20%) were stab wounds (SW). The remaining 20 patients (28%) were injured by blunt mechanisms: 9 (45%) as auto vs pedestrian, 7 (35%) in a motor vehicle collision, and 4 (20%) in a motorcycle crash. The mean AAST pancreatic injury grade was II (range I-V) (Table 1). There were 22 patients (31%) with AAST grade I injuries, 17 (24%) with grade II, 29 (41%) with grade III, 2 (3%) with grade IV, and 1 (1%) with grade V. Associated injuries were common (n ¼ 70,
Table 1 Patient Demographics, Clinical Data, Management, and Outcomes. Continuous variables are presented as mean ± standard deviation; median (range). Categorical variables are presented as n (%). Patients (n ¼ 71) Demographics Age, years Gender, male Mechanism of Injury4 (20%) Penetrating GSW SW Blunt AVP MVC MCC Injury Data Injury Severity Score (ISS) AAST Pancreatic Injury Grade I II III IV V Management Immediate Laparotomy CT Scan Other Pre-Operative Duct Assessment Nonoperative Management (NOM) Total Operative Management Intra-Operative Duct Evaluation VISUAL INSPECTION PANCREATOGRAPHY ERCP Duodenotomy Cholecystotomy Outcomes Mortality Hospital LOS ICU LOS Ventilator Days Need for ICU Need for Mechanical Ventilation Failed NOM Pancreatic Complication*
35 ± 17; 29 (11e87) 61 (86%) 51 (72%) 41 (80%) 10 (20%) 20 (28%) 9 (45%) 7 (35%) 4 (20%) 24 ± 13; 20 (4e75) 2 ± 1; 2 (1e5) 22 (31%) 17 (24%) 29 (41%) 2 (3%) 1 (1%) 50 (70%) 21 (30%) 0 (0%) 5 (7%) 66 (93%) 62 (94%) 4 (6%) 0 (0%) 1 (2%) 3 (5%) 14 (20%) 25 ± 26; 15 (1e131) 12 ± 16; 5 (0e78) 5 ± 9; 1 (0e39) 59 (83%) 40 (56%) 0 (0%) 10 (14%)
GSW, gunshot wound. SW, stab wound. AVP, auto vs. Pedestrian. MVC, motor vehicle collision. MCC, motorcycle crash. CT, computed tomography. ERCP, endoscopic retrograde cholangiopancreatography. LOS, length of stay in days. ICU, intensive care unit.*Pancreatic Complications included leak, abscess, fistula.
99%), most frequently the stomach (n ¼ 31, 44%), liver (n ¼ 24, 34%), colon (n ¼ 19, 27%), kidney (n ¼ 19, 27%), and spleen (n ¼ 19, 27%)
Table 2 Associated injuries. Injury
Patients (n ¼ 71)
Percentage (%)
Any Stomach Liver Colon Kidney Spleen Duodenum Diaphragm Small Bowel IVC PV/SMV/IMV Gallbladder Aorta Celiac Axis/SMA Extra-Abdominal
70 31 24 19 19 19 14 12 9 5 5 4 3 1 19
99 44 34 27 27 27 20 17 13 7 7 6 4 1 27
IVC, inferior vena cava. PV/SMV/IMV, portal vein/superior mesenteric vein/inferior mesenteric vein. SMA, superior mesenteric artery.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
(Table 2). The mean Injury Severity Score (ISS) was 24 (range 4e75) (Table 1).
Management Twenty-one patients (30%) underwent CT scan, of which 16 (76%) demonstrated pancreatic injury (Table 1, Fig. 1). The 5 patients (24%) who had a falsely negative CT scan for pancreatic injury were brought for laparotomy for the management of associated injuries and then found intraoperatively to have a pancreatic injury. Of the 16 patients who had a CT scan that was positive for pancreatic injury, 11 (69%) underwent laparotomy and 5 (31%) underwent non-operative management. These 5 patients all sustained blunt trauma and were managed successfully nonoperatively without complication or mortality. One of the patients managed non-operatively (20%) underwent MRCP after CT scan, which demonstrated pancreatic injury without ductal disruption. The remaining 4 patients did not undergo further workup after CT scan to rule out pancreatic duct injury. Most patients (n ¼ 50, 70%) underwent immediate laparotomy. When combined with the patients who underwent preoperative CT scan, a total of 66 patients (93%) were managed operatively. The overwhelming majority were assessed intraoperatively for ductal injury with visual inspection alone (n ¼ 62, 94%). Four patients (6%) underwent intraoperative pancreatography. Of these, one patient, who sustained a grade II injury to the head of the pancreas following GSW, had a duodenotomy created to cannulate the ampulla. Methylene blue was injected to interrogate the duct, which was inconclusive for duct injury. Three patients with grade II pancreatic injuries underwent intraoperative pancreatography using methylene blue via cholecystotomy, all of which were deemed inconclusive. All 4 of these patients were subsequently managed with drain placement alone and no pancreatic complications ensued. No patient underwent intraoperative ERCP or
3
intraoperative ultrasound. At the index operation, patients were most commonly managed with closed suction drainage alone (n ¼ 32, 45%) or distal pancreatectomy (n ¼ 28, 39%) (Table 3). Three patients (4%) had no intervention to the pancreas at laparotomy and 2 patients (3%) underwent a staged trauma Whipple. Patients managed with distal pancreatectomy or trauma Whipple resection typically had closed suction drains left in the surgical bed at the end of the procedure as well. Patients with low grade (I-II) injuries were most frequently managed with drain placement alone (n ¼ 31, 79%). Grade III pancreatic injuries (n ¼ 29) were all managed with distal pancreatectomy. This was typically performed at the index operation (n ¼ 28, 97%), except for 1 patient who initially underwent closed suction drain placement alone and then underwent distal pancreatectomy at the first take-back operation on post-operative day 1 (Table 3). Of the three patients with grade IV-V pancreatic injuries, 1 expired intraoperatively before the injury could be addressed. The remaining 2 patients underwent a staged trauma Whipple operation, with the resection performed at the index surgery. Reconstruction was undertaken in both patients on post-operative day 2.
Outcomes Overall mortality was 20% (n ¼ 14) (Table 1). Mean hospital length of stay was 25 days (range 1e131), with 59 patients (83%) requiring ICU admission and 40 (56%) needing mechanical ventilation. Pancreatic complications, defined as the development of a pancreatic leak, fistula, pancreatitis, or peripancreatic abscess, occurred in 10 patients (14%). The mean grade of pancreatic injury among these patients was II (range I-III). Of the 10 patients who developed a pancreatic complication, 4 (40%) underwent drain
Fig. 1. Flow of patients through study. CT, computed tomography. NOM, nonoperative management.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
4
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
Table 3 Surgical management at the index operation.
Nonoperative Management No Intervention at Laparotomy Drain Placement Distal Pancreatectomy Whippleb a b
All Patients (n ¼ 71a)
AAST I-II (n ¼ 39)
AAST III (n ¼ 29)
AAST IV-V (n ¼ 3a)
5 (8%) 3 (4%) 32 (45%) 28 (39%) 2 (3%)
5 (13%) 3 (8%) 31 (79%) 0 (0%) 0 (0%)
0 (0%) 0 (0%) 1 (3%) 28 (97%) 0 (0%)
0 0 0 0 2
(0%) (0%) (0%) (0%) (67%)
One patient expired before the injury could be addressed. A trauma Whipple resection without reconstruction was performed.
placement alone at the index operation and 6 (60%) underwent distal pancreatectomy. The complication was diagnosed on CT scan in 5 patients (50%), based on drain output and biochemistry (i.e. elevated drain fluid lipase) in 3 patients (30%), and using ERCP in 2 (20%) patients. The 2 patients who underwent ERCP were both investigated for a leak with MRCP prior to ERCP. However, both of these MRCPs were inconclusive. In terms of treatment of pancreatic complications, 7 patients (70%) were managed with the drain placed at the index OR, a period of NPO, and a combination of TPN (n ¼ 4) and antibiotics (n ¼ 5); 2 (20%) with a stent placed in the pancreatic duct during ERCP; and 1 (10%) with a period of NPO, TPN, and antibiotics alone. There were no mortalities among the patients with pancreatic complications. Discussion Pancreatic injuries can be challenging to diagnose and manage due to their infrequent occurrence, the retroperitoneal nature of the pancreas, the proximity to major vascular structures, and the severity of concomitant injuries. The AAST injury grading system clearly defines pancreatic injury on the basis of duct injury and the location of the injury along the pancreas.6 The management principles for pancreatic trauma are then based upon this AAST grading.7e10 Although it is critical for proper AAST injury grading and injury management, the optimal method for diagnosing a pancreatic duct injury is not well defined by the current literature. Earlier studies defined intraoperative criteria for the diagnosis of main pancreatic duct injury, emphasizing the importance of intraoperative visual inspection of the gland. These criteria included direct visualization of duct injury, complete transection or disruption of >50% of the gland, central gland perforation, or severe maceration of the pancreas.11 However, early studies also argued for the liberal use of intraoperative pancreatography to interrogate the pancreatic duct for injury.11,12 More recently, literature has begun to emerge suggesting that intraoperative visual inspection of the pancreas without intraoperative pancreatography is sufficient to rule out duct injury.8,13 These studies, performed on patients with pancreatic injury from 1996 to 2009 and 1990e1995, respectively, showed that visual inspection for duct injury and management on the basis of these findings resulted in acceptable rates of pancreatic complications. In both of these studies, however, no patient underwent intraoperative pancreatography and thus the utility of these techniques could not truly be assessed. As the accessibility and resolution of CT scans have improved, and techniques such as MRCP and ERCP have evolved over time, the diagnostic armamentarium for potential pancreatic duct injury has expanded. Furthermore, with significant improvements in prehospital care and shortened transport time since the original studies, more patients with pancreatic injury are presenting alive to hospital now, requiring evaluation of their injuries. In the current study, patients typically sustained a pancreatic injury after penetrating mechanisms, especially gunshot wounds.
They were severely injured with a high mortality rate, and associated injuries occurred almost universally. Therefore, when pancreatic injury is detected, a thorough search for concomitant injuries must ensue. Most patients were brought for immediate laparotomy. A minority of patients, however, underwent CT scan, which was not a sensitive test (76%) for diagnosing pancreatic injuries. The current literature defines the sensitivity of CT scan for diagnosis of pancreatic injuries in the range of 47e76%,2,7,14 including a AAST multicenter study which published a sensitivity of 52% for 64-slice CT scan.2 CT scan is therefore insufficient to rule out pancreatic injury. In our study, approximately a quarter of patients who were found intraoperatively to have a pancreatic injury had a falsely negative CT scan demonstrating a normal pancreas immediately preoperatively. In this series, intraoperative visual inspection alone was the most frequently used method to investigate the pancreas for potential duct injury. However, visual inspection alone may not be perfect. Four patients who were assessed for ductal injury using visual inspection were deemed not to have a ductal injury and were managed with drain placement alone, but subsequently developed pancreatic leaks or abscesses. This may simply be due to the fact that the soft pancreas of a young trauma patient is prone to leak after injury, even if the main pancreatic duct is intact. However, the possibility that a duct injury was missed with visual inspection must be considered. This would place the failure rate of visual inspection for pancreatic duct injury at a maximum of 6% in this series. The majority of patients in this series who experienced a pancreatic complication were managed successfully using a combination of the closed suction drain left at the index surgery, antibiotics, and TPN. Intraoperative drains should therefore be placed routinely after pancreatic injury in order to mitigate the need for additional intervention should a complication occur. Intraoperative pancreatography is rarely used in contemporary practice and, in all cases within this series, was found to be inconclusive. It is invasive, time consuming, and of limited utility in the work-up of pancreatic trauma. In this study, MRCP and ERCP were also used infrequently and played a limited role in patients with pancreatic injury. They were almost exclusively reserved for the diagnosis and management of postoperative complications. One patient who was diagnosed with a low grade pancreatic injury after blunt trauma based on CT scan underwent MRCP to confirm the absence of duct injury. Otherwise, MRCP was only used among two postoperative patients who were suspected clinically of having a pancreatic leak. Both tests failed to show the leak that was then demonstrated and stented with ERCP. No patient underwent intraoperative ERCP. Although there is literature supporting the use of intraoperative ERCP in pancreatic trauma,15 it is a resourceintensive undertaking and this limits its use, particularly in a practice setting such as ours. The limitations of this study include its retrospective, registrybased design. This is a single-center study and the external validity of these results, particularly for lower volume trauma centers,
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002
M. Schellenberg et al. / The American Journal of Surgery xxx (2018) 1e5
is unknown. It is, however, a large contemporary analysis of current practice patterns at one of the highest volume trauma centers in North America. Future multi-center prospective study is warranted. Conclusions This study demonstrates that the clinician must be cautious when using CT scan to screen trauma patients for pancreatic injury since the sensitivity is poor. Intraoperatively, pancreatography for the evaluation of the pancreatic duct should have a very limited role. In addition to being time consuming and invasive, it was inconclusive in all cases in which it was used. Clinical suspicion for ductal injury based on intraoperative visual inspection alone is sufficient for the management of pancreatic injuries and should be the preferred strategy for evaluation of the pancreatic duct. Funding sources This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Conflicts of interest The authors have no conflicts of interest to disclose. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. References
5
Trauma. seventh ed. New York: The McGraw Hill Companies; 2013:603e619. 2. Phelan HA, Velmahos GC, Jurkovich GJ, et al. An evaluation of multidetector computed tomography in detecting pancreatic injury: results of a multicenter AAST study. J Trauma. 2009;66(3):641e646. 3. Ragozzino A, Manfredi R, Scaglione M, De Ritis R, Romano S, Rotondo A. The use of MRCP in the detection of pancreatic injuries after blunt trauma. Emerg Radiol. 2003;10(1):14e18. 4. Fulcher AS, Turner MA, Yelon JA, et al. Magnetic resonance cholangiopancreatography (MRCP) in the assessment of pancreatic duct trauma and its sequelae: preliminary findings. J Trauma. 2000;48(6):1001e1007. 5. Centers for Disease Control and Prevention. International Classification of Diseases, Ninth Revision (ICD-9); 2009. Available at: http://www.cdc.gov/nchs/icd/ icd9.htm [accessed 28 March 2017]. 6. Moore EE, Cogbill TH, Malangoni MA, et al. Organ injury scaling. Surg Clin North Am. 1995;75(2):293e303. 7. Ho VP, Patel NJ, Bokhari F, et al. Management of adult pancreatic injuries: a practice management guideline from the eastern association for the surgery of trauma. J Trauma. 2017;82(1):185e199. 8. Sharpe JP, Magnotti LJ, Weinberg JA, et al. Impact of a defined management algorithm on outcome after traumatic pancreatic injury. J Trauma. 2012;72: 100e105. 9. Fisher M, Brasel K. Evolving management of pancreatic injury. Curr Opinion Crit Care. 2011;17:613e617. 10. Stawicki SP, Schwab CW. Pancreatic trauma: demographics, diagnosis, and management. Am Surg. 2008;74:1133e1145. 11. Berni GA, Bandyk DF, Oreskovich MR, Carrico CJ. Role of intraoperative pancreatography in patients with injury to the pancreas. Am J Surg. 1982;143: 602e605. 12. Jones RC, Shires GT. Pancreatic trauma. Arch Surg. 1971;102:424e430. 13. Patton Jr JH, Lyden SP, Croce MA, et al. Pancreatic trauma: a simplified management guideline. J Trauma. 1997;43:234e239. 14. Velmahos GC, Tabbara M, Gross R, et al. Blunt pancreaticoduodenal injury: a multicenter study of the research consortium of new england centers for trauma (ReCONECT). Arch Surg. 2009;144(5):413e419. 15. Rogers SJ, Cello JP, Schecter WP. Endoscopic retrograde cholangiopancreatography in patients with pancreatic trauma. J Trauma. 2010;68(3):538e544.
1. Duodenum Biffl WL. Pancreas. In: Mattox KL, Moore EE, Feliciano DV, eds.
Please cite this article in press as: Schellenberg M, et al., Detection of traumatic pancreatic duct disruption in the modern era, The American Journal of Surgery (2018), https://doi.org/10.1016/j.amjsurg.2018.06.002