- Email: [email protected]
Pediatric pancreatic trauma: predictors of nonoperative management failure and associated outcomes
Journal of Pediatric Surgery (2007) 42, 340 – 344
www.elsevier.com/locate/jpedsurg
Pediatric pancreatic trauma: predictors of nonoperative management failure and associated outcomes Kelly D. Mattixa,*, M. Tatariab, J. Holmesc, K. Kristoffersend, R. Browne, J. Gronerf, E. Scaifeg, D. Mooneyd, M. Nancec, L. Scherera a
Riley Hospital for Children, Indiana University, Indianapolis, IN 46202, USA Lucille Packard Children’s Hospital, Stanford University, Stanford, CA 94305, USA c Children’s Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA 19104, USA d Boston Children’s Hospital, Harvard University, Boston, MA 02115, USA e Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA f Children’s Hospital, Ohio State University, Columbus, OH 43207, USA g Primary Children’s Medical Center, University of Utah, Salt Lake City, UT 84113, USA b
Index words: Pancreatic injury; Nonoperative management; Pseudocyst; Pediatric trauma
Abstract Background: Nonoperative management (NOM) is an accepted treatment of pediatric solid organ injuries and is typically successful. Blunt pancreatic trauma tends to require operative intervention more frequently. We sought to identify predictors of failure of NOM and compare the outcome of operative management against NOM. Methods: A retrospective analysis was performed from January 1993 to December 2002 of all children with blunt pancreatic injuries from the trauma registries of 7 designated level 1 pediatric trauma centers. Failure of NOM was defined as the need for intraabdominal operative intervention. Injuries were graded I to V, and ductal injury was defined as grades III to V. Parameters included mechanism of injury, injury severity score (ISS), organ grade, Glasgow Coma Scale score, and outcome. Data were analyzed by Fisher exact test and Mann-Whitney U test, with mean values F SD and significance of P b .05. Results: Pancreatic injuries were present in 173 (9.2%) of 1823 patients. Of these, 43 (26.0% [43/173]) required an operation. Valid morbidity data was obtained in 118 of 173 patients. ISS was significantly higher in all patients treated operatively. Patients with an injury of grade III to V failed NOM more frequently than all patients with pancreatic injury ( P =.0169). Length of stay was longer, and the incidence of pseudocysts, drainage procedures, and pancreatitis was higher in NOM patients, although not significant. Conclusions: Patients with pancreatic injuries had a NOM failure rate of 26.0%. ISS and injury grades III to V were predictors of NOM failure. Patients with pancreatic ductal injury require more aggressive management. D 2007 Elsevier Inc. All rights reserved.
Presented at the British Association of Paediatric Surgeons 53rd Annual International Congress, Stockholm, Sweden, July 18 - 22, 2006. * Corresponding author. Tel.: +1 317 274 7262; fax: +1 317 274 8769. E-mail address: [email protected] (K.D. Mattix). 0022-3468/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2006.10.006
Worldwide, trauma is the leading cause of death and disability in children, with most owing to blunt mechanisms. Blunt abdominal trauma often leads to solid organ injury, primarily to the spleen, liver, and kidneys. Injury to the
Pediatric pancreatic trauma Table 1 centers
341
Participating institutions: Level I pediatric trauma
Boston Children’s Hospital, Harvard University, Boston, MA Children’s Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA Cincinnati Children’s Hospital Medical Center, University of Cincinnati, Cincinnati, OH Columbus Children’s Hospital, Ohio State University, Columbus, OH Lucille Packard Children’s Hospital, Stanford University, Palo Alto, CA Primary Children’s Hospital, University of Utah, Salt Lake City, UT Riley Hospital for Children, Indiana University, Indianapolis, IN
pancreas is the fourth most common, occurring in 3% to 12% of patients with blunt abdominal trauma [1]. Nonoperative management (NOM) has become the standard of care in most solid organ injuries, including pancreatic trauma [2-4]. However, treatment of major ductal disruption is controversial. Few reports describe the clinical outcome of operative management of pancreatic injuries, compared to NOM.
1. Methods A multiinstitutional retrospective case review was conducted from January 1993 to December 2002 of all patients with blunt pancreatic injuries from the trauma registries of 7 designated level I pediatric trauma centers. (Table 1) Failure of nonoperative management was defined as the need for an intraabdominal operative intervention. Injuries were graded I to V according to Organ Injury Scaling of the American Association for the Surgery of Trauma. Ductal Injuries were defined as grades III to V. Data were collected for age, sex, mechanism of injury, associated injuries, Glasgow Coma Scale and Injury Severity Score (ISS). Morbidity factors obtained were hospital length of stay (LOS), length of intensive care unit
(ICU) stay, number of days requiring total parenteral nutrition (TPN), development of pancreatitis, and formation of pancreatic fistulas or pseudocysts. Operative information included time to operative intervention; cause for intervention; and need for invasive procedures including external drainage, endoscopic retrograde cholangiopancreatography (ERCP), or internal drainage procedures. Data were analyzed by one-way Fisher exact test and Mann-Whitney U test with mean values F SD. Significance was defined as P V .05. Approval was obtained by the institutional review board of each institution.
2. Results There were 1823 patients identified with blunt abdominal trauma, 173 with injury to the pancreas (9.5%). Forty-three patients with pancreatic injury required operative intervention (26.0%). Ductal injuries were present in 53, with 23 requiring an operation (43.4%). The mean age of patients treated nonoperatively was 7.84 years (range, 1 month to 17 years) and was 7.56 years (range, 1-17 years) in patients requiring an operation. Boys accounted for 63.3% of operative patients and 62.2% of those nonoperatively managed. Mechanism of injury in all patients was bicycle in 21.1%, motor vehicle collision in 28.9%, and 31.1% and 22.2% of patients with a ductal injury, respectively. Mean ISS for all patients with pancreatic injury was not significantly different for those treated with NOM (15.5 F 13.7) and those managed operatively, (25.4 F 15.5) ( P = .917). However, those patients with a ductal injury who were treated operatively had a significantly higher ISS than those treated nonoperatively (24.0 F 12.6 vs 15.8 F 11.0, P = .025), but no association was seen with a specific organ or organ system. Indications for operative intervention included free intraperitoneal air in 4 children (8.9%), shock or refractory hypotension in 12 (26.7%), and 1 for other cause (2.2%). The pancreas was cited as the cause for operative intervention of 57.8% of all patients and 82.6% of patients with ductal injury.
Table 2 Data results: Comparison of patients treated nonoperatively and operatively. Patients with ductal injuries (Grade III-V) were also isolated and compared All pancreatic injuries Age (y) Sex (% male) ISS GCS Bicycle MVC LOS (d) ICU LOS (d)
Ductal injuries
NOM (n = 128)
Failure (n = 45)
7.7 F 63.3% 15 F 13 F 21.8% 26.6% 9.6 F 4.17 F
7.6 F 3.9 63.2% 27 F 18 10 F 6 32.6% 20.9% 12.1 F 12.0 3.58 F 5.0
4.6 14 4
19.8 19.5
P
NOM (n = 30) 9.2 F 4.1
.917 .916
.675 .675
16 F 14 F 40% 16.7% 13.8 F 1.31 F
11 3
10.1 2.1
Failure (n = 23)
P
6.9 F 3.7 29 F 16 11 F 5 30% 17.4% 9.7 F 8.4 1.60 F 2.5
.025 .884
.668 .139
342
K.D. Mattix et al.
Fig. 1 Morbidity data: Incidence of morbidity factors in patient with pancreatic injury treated by NOM or operatively.
Morbidity factors were obtained in 118 of 173 patients, and 33 of 53 patients with ductal injury. Mean LOS was 9.56 F 19.8 days for NOM patients and 12.08 F 12.0 days for patients managed operatively, whereas ICU stay was 4.17 F 19.8 and 3.58 F 5.0 days, respectively. Patients with ductal injury cared for with NOM had an LOS of 11.33 F 8.0 days, and those requiring an operation, 9.70 F 8.4 days, with a required ICU stay of 1.31 F 2.1 and 1.60 F 2.5 days, respectively (Table 2). Of the patients undergoing NOM for all pancreatic injuries, full morbidity data were available in 83 of 128 patients. Total parenteral nutrition was administered in 28 patients (30.1%) for a mean 4.60 days (range, 3-31 days). Eight patients required a drainage procedure (2 internal and 6 external), whereas 1 patient required an ERCP. Other complications observed were pancreatitis in 26 patients, pseudocyst formation in 12, and pancreatic fistula in 2. Similar data were obtained in 23 of the 53 patients managed operatively. Fifteen patients (60%) received TPN for a mean 13.21days (range, 3-66 days). No patients required an internal drainage procedure, 6 required operative external drainage, and ERCP was performed in 3. Related compli-
cations included 4 patients with pancreatitis; 4 developed pseudocysts, and 1 had a pancreatic fistula (Fig. 1). Analysis of patients with ductal injury demonstrated the following comparisons. Full morbidity data was obtained in 18 of 30 patients managed conservatively, with 1 patient requiring internal drainage and 4 managed by percutaneous external drainage; ERCP was not used in these patients. Pancreatitis developed in 6 patients, pseudocysts in 9, and pancreatic fistula in 1. Nine patients (50%) required TPN for a mean of 9.46 days (range, 2-28 days). Parallel data in the operative patients demonstrated no internal drainage procedures, 3 operative external drainage procedures, and the use of ERCP in 2. Complications included pancreatitis in 3 patients, 3 pseudocysts, and 1 fistula. Total parenteral nutrition was used in 9 patients (60%) for a mean 11.16 days (range, 3-34 days). Additional analysis of those patients with a pancreatic ductal injury was performed. (Table 3) Thirteen patients sustained a grade III injury and were managed nonoperatively, whereas an additional 13 patients with a grade III injury underwent operative intervention. The mean time to intervention was 52.11 hours (range, 0.75-216 hours), with 6 of the 13 patients taken to the operating room (OR) within 24 hours of their injury. Two episodes of pancreatitis were seen in these 6 patients. In contrast, 1 pancreatic fistula and 3 pseudocysts developed in the 7 patients with grade III injuries who underwent operative treatment in delayed fashion (N24 hours). Patients with severe ductal injuries were taken to the OR in a more timely fashion. Eleven patients sustained a grade IV ductal injury and were treated with NOM, whereas 3 patients underwent operative intervention with a mean time to the OR of 42.00 hours (range, 2-100 hours); 1 of the 3 patients was delayed for more than 24 hours. Again, morbidity factors were noticeably different because no complications were seen in those patients with grade IV injury who were managed operatively, whereas 2 pseudocysts and 3 episodes of pancreatitis were noted in NOM patients. Grade V ductal injuries were treated nonopera-
Table 3 Ductal injury morbidity data: Incidence of morbidity factors in patients with ductal injury differentiated by grade and treatment. (No morbidity data was available for the patients with a grade IV injury treated operatively nor those with grade V injury treated with NOM.) No. of patients Morbidity data Mortality Time to OR (mean hours) N 24 h to OR Pseudocyst Pancreatitis Fistula TPN days TPN required
Grade III NOM
Grade III operative
Grade IV NOM
Grade IV operative
Grade V NOM
Grade V operative
13 6 2
13 5 1 52.11
11 8 0
3 0 0 42.00
3 0 0
7 6 1 15.18574
3 3 1 11.33333 3 of 6
6 3 2 1 8.4 4 of 5
1 2 3 8.0625 6 of 8
1 1 1 13.45833 5 of 6
Pediatric pancreatic trauma tively in 3 and operatively in 7 patients; the mean time to operation was 15.19 hours (range, 0.25-59 hours), with 1 patient delayed for more than 24 hours. Of the patients with grade V injuries, 2 patients were taken to the OR for free air and 1 patient for persistent hypotension; all 3 were found to have duodenal injury. Decision to undergo operative intervention in those patients with grade III or IV injuries was based solely on the presence of a pancreatic injury, with 1 exception of hypotension/shock.
3. Discussion Current standard of care for injury to solid organs after blunt trauma is NOM. The optimal management of injury to the pancreas has been more controversial. A small number of studies in both the adult and pediatric literature have assessed the use of ERCP in patients with pancreatic injury [5,6]. Several other studies have analyzed the treatment of pancreatic injuries, but few have compared operative treatment to NOM [1-8]. The data in this study were acquired from multiple pediatric level I trauma institutions and allowed us to accrue a larger number of patients to compare these 2 management strategies. The incidence of pancreatic injury in patients sustaining blunt trauma was 9.2%, which is similar to other reported data [1,3,9]. In addition, as in other series, these patients had significant morbidity related to their injury. Nonoperative management was successful in most (74%) of the entire cohort of patients with pancreatic injury. This is quite different from the reported NOM success rate of 5% in all pediatric patients with abdominal trauma [2]. Overall, there did not appear to be any advantage in one method of management compared with another in children with pancreatic injuries. However, our data suggest that there may be an advantage in early operation in those who have ductal injuries. Although pseudocysts occurred less frequently in our series compared with the reported literature (13%-30% vs 40%100%) [1,3,9,10] these data appear to indicate that this was more common in those with ductal injury, who were treated conservatively, compared with those who had early operation. This is now increasingly recognized, and there have been several reports of distal ductal injury (grades IIIIV or class II), which appear to demonstrate decreased morbidity where these patients are treated early with aggressive surgical management, most often a distal pancreatectomy [1,8,7]. Injury severity score will be expected to correlate with the comorbidities in patients with pancreatic injuries. In patients with a ductal injury, only increased ISS was found to be a significant factor correlating with the perceived need for operative treatment. Unfortunately, complete data were not recorded for all patients for the specific comorbidities. Despite this, it is clear from the ISS that those that came to surgery were subject to more multisystem injuries, including
343 pancreatic ductal disruption, compared to those who were treated conservatively (Table 3). Patients with a ductal injury had a reduction in the incidence of complications when taken to the OR within 24 hours of injury. Therefore, early and accurate diagnosis of ductal injury and prompt operative intervention can be expected to result in fewer complications and decreased morbidity. There are several limitations to this study. The low incidence of pancreatic injury secondary to blunt trauma makes accurate assessment of this patient population difficult. Although the inclusion of multiple institutions increases the number of patients, there is a risk of variation in management, including the threshold for operative intervention, the indications for parenteral nutrition, or the availability of ERCP and helical computed tomography (CT). Similarly, objective and consistent description of pancreatic ductal injury may differ between institutions. Despite these limitations, our findings demonstrate that all children with pancreatic injury have significant morbidity. Initial diagnosis should ideally be made using helical CT imaging. If ductal injury is equivocal based on CT findings, ERCP has become an accepted diagnostic method to evaluate for pancreatic ductal injury [5-8]. Additionally, ERCP may also develop into a valuable option for therapeutic intervention, including stent placement, in proximal ductal injuries and should be considered in those patients [5,8]. Finally, although children who have suffered pancreatic injury (but without ductal disruption) do not appear to suffer increased morbidity after conservative management, these data suggest that patients with ductal disruption may benefit from early operative intervention.
4. Q&A Marven, Sheffield, UK In addition to CT, how many of these patients had MR or ERCP? A. Of the patients with ductal injury, 2 patients, of a total of 53, had ERCP as a diagnostic tool and then went on to operative intervention. Marven, Sheffield, UK What were the different operative interventions? A. All of them were distal pancreatectomies or subtotal pancreatectomies with the exception of one who had a Roux loop to the pancreatic duct for a very proximal injury. Marven, Sheffield, UK No patients had a Roux loop to the distal pancreatic duct. A. No. Langer, Toronto, Canada A number of years ago, we described 14 consecutive children with a complete transection of the pancreatic duct,
344 which were managed nonoperatively. Of your patients who had significant grade 3-5 injuries, were there any regional or within-institution differences with respect to surgeons’ preference or indications for operative interventions or what were considered failure of the NOM? A. I was unable to obtain that. We had a blinded study; once the patients were given to us, the institution was unknown. So I am unable to answer that question, I’m sorry. Ade-Aji, London, UK We recently also recommended the use of ERCP, a strategy which has evolved in the context of delayed referral pattern. Given that you are seeing patients acutely, what do you think is the role for ERCP and stents at this stage? A. I believe that that role becomes important when you have a patient who is brought in initially, and I also believe that role is limited to your experience within the institution. We have very skilled gastroenterologists who are able to perform ERCP for us in younger children. With that, I think that the evidence and a couple of recent papers have shown that ERCP can be a very effective diagnostic means. Also, there has been a recent paper published, which shows that stenting is a viable option. Again, I think that it
K.D. Mattix et al. becomes important when patients present and then the quality of your experience at the institution.
References [1] Jobst M, Canty T, Lynch F. Management of pancreatic injury in pediatric blunt abdominal trauma. J Pediatr Surg 1999;34:818 - 24. [2] Stylianos S. Evidence-based guidelines for resource utilization in children with isolated spleen of liver injury. J Pediatr Surg 2000; 35:164 - 9. [3] Shilyansky J, Sena L, Dreller M, et al. Non-operative management of pancreatic injuries in children. J Pediatr Surg 1998;33:343 - 9. [4] Loungnarath R, Blanchard H, Saint-Vil D, et al. Blunt pancreatic injuries in children. Ann Chir 2001;126:992 - 5. [5] Canty T, Weinman D. Treatment of pancreatic duct disruption in children by an endoscopically placed stent. J Pediatr Surg 2001;36:345 - 8. [6] Rescorla F, Plumley D, Sherman S, et al. The efficacy of early ERCP in pediatric pancreatic trauma. J Pediatr Surg 1995;30:336 - 40. [7] Gaines B, Ford H. Abdominal and pelvic trauma in children. Crit Care Med 2002;30:S416 - 23. [8] Canty T, Weinman D. Management of major pancreatic injuries in children. J Trauma 2001;50:1001 - 5. [9] Meier D, Coln C, Hicks B, et al. Early operation in children with pancreas transaction. J Pediatr Surg 2001;36:341 - 4. [10] Jacombs A, Wines M, Holland A, et al. Pancreatic trauma in children. J Pediatr Surg 2004;39:96 - 9.
www.elsevier.com/locate/jpedsurg
Pediatric pancreatic trauma: predictors of nonoperative management failure and associated outcomes Kelly D. Mattixa,*, M. Tatariab, J. Holmesc, K. Kristoffersend, R. Browne, J. Gronerf, E. Scaifeg, D. Mooneyd, M. Nancec, L. Scherera a
Riley Hospital for Children, Indiana University, Indianapolis, IN 46202, USA Lucille Packard Children’s Hospital, Stanford University, Stanford, CA 94305, USA c Children’s Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA 19104, USA d Boston Children’s Hospital, Harvard University, Boston, MA 02115, USA e Cincinnati Children’s Hospital Medical Center, Cincinnati, OH 45229, USA f Children’s Hospital, Ohio State University, Columbus, OH 43207, USA g Primary Children’s Medical Center, University of Utah, Salt Lake City, UT 84113, USA b
Index words: Pancreatic injury; Nonoperative management; Pseudocyst; Pediatric trauma
Abstract Background: Nonoperative management (NOM) is an accepted treatment of pediatric solid organ injuries and is typically successful. Blunt pancreatic trauma tends to require operative intervention more frequently. We sought to identify predictors of failure of NOM and compare the outcome of operative management against NOM. Methods: A retrospective analysis was performed from January 1993 to December 2002 of all children with blunt pancreatic injuries from the trauma registries of 7 designated level 1 pediatric trauma centers. Failure of NOM was defined as the need for intraabdominal operative intervention. Injuries were graded I to V, and ductal injury was defined as grades III to V. Parameters included mechanism of injury, injury severity score (ISS), organ grade, Glasgow Coma Scale score, and outcome. Data were analyzed by Fisher exact test and Mann-Whitney U test, with mean values F SD and significance of P b .05. Results: Pancreatic injuries were present in 173 (9.2%) of 1823 patients. Of these, 43 (26.0% [43/173]) required an operation. Valid morbidity data was obtained in 118 of 173 patients. ISS was significantly higher in all patients treated operatively. Patients with an injury of grade III to V failed NOM more frequently than all patients with pancreatic injury ( P =.0169). Length of stay was longer, and the incidence of pseudocysts, drainage procedures, and pancreatitis was higher in NOM patients, although not significant. Conclusions: Patients with pancreatic injuries had a NOM failure rate of 26.0%. ISS and injury grades III to V were predictors of NOM failure. Patients with pancreatic ductal injury require more aggressive management. D 2007 Elsevier Inc. All rights reserved.
Presented at the British Association of Paediatric Surgeons 53rd Annual International Congress, Stockholm, Sweden, July 18 - 22, 2006. * Corresponding author. Tel.: +1 317 274 7262; fax: +1 317 274 8769. E-mail address: [email protected] (K.D. Mattix). 0022-3468/$ – see front matter D 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jpedsurg.2006.10.006
Worldwide, trauma is the leading cause of death and disability in children, with most owing to blunt mechanisms. Blunt abdominal trauma often leads to solid organ injury, primarily to the spleen, liver, and kidneys. Injury to the
Pediatric pancreatic trauma Table 1 centers
341
Participating institutions: Level I pediatric trauma
Boston Children’s Hospital, Harvard University, Boston, MA Children’s Hospital of Philadelphia, University of Pennsylvania, Philadelphia, PA Cincinnati Children’s Hospital Medical Center, University of Cincinnati, Cincinnati, OH Columbus Children’s Hospital, Ohio State University, Columbus, OH Lucille Packard Children’s Hospital, Stanford University, Palo Alto, CA Primary Children’s Hospital, University of Utah, Salt Lake City, UT Riley Hospital for Children, Indiana University, Indianapolis, IN
pancreas is the fourth most common, occurring in 3% to 12% of patients with blunt abdominal trauma [1]. Nonoperative management (NOM) has become the standard of care in most solid organ injuries, including pancreatic trauma [2-4]. However, treatment of major ductal disruption is controversial. Few reports describe the clinical outcome of operative management of pancreatic injuries, compared to NOM.
1. Methods A multiinstitutional retrospective case review was conducted from January 1993 to December 2002 of all patients with blunt pancreatic injuries from the trauma registries of 7 designated level I pediatric trauma centers. (Table 1) Failure of nonoperative management was defined as the need for an intraabdominal operative intervention. Injuries were graded I to V according to Organ Injury Scaling of the American Association for the Surgery of Trauma. Ductal Injuries were defined as grades III to V. Data were collected for age, sex, mechanism of injury, associated injuries, Glasgow Coma Scale and Injury Severity Score (ISS). Morbidity factors obtained were hospital length of stay (LOS), length of intensive care unit
(ICU) stay, number of days requiring total parenteral nutrition (TPN), development of pancreatitis, and formation of pancreatic fistulas or pseudocysts. Operative information included time to operative intervention; cause for intervention; and need for invasive procedures including external drainage, endoscopic retrograde cholangiopancreatography (ERCP), or internal drainage procedures. Data were analyzed by one-way Fisher exact test and Mann-Whitney U test with mean values F SD. Significance was defined as P V .05. Approval was obtained by the institutional review board of each institution.
2. Results There were 1823 patients identified with blunt abdominal trauma, 173 with injury to the pancreas (9.5%). Forty-three patients with pancreatic injury required operative intervention (26.0%). Ductal injuries were present in 53, with 23 requiring an operation (43.4%). The mean age of patients treated nonoperatively was 7.84 years (range, 1 month to 17 years) and was 7.56 years (range, 1-17 years) in patients requiring an operation. Boys accounted for 63.3% of operative patients and 62.2% of those nonoperatively managed. Mechanism of injury in all patients was bicycle in 21.1%, motor vehicle collision in 28.9%, and 31.1% and 22.2% of patients with a ductal injury, respectively. Mean ISS for all patients with pancreatic injury was not significantly different for those treated with NOM (15.5 F 13.7) and those managed operatively, (25.4 F 15.5) ( P = .917). However, those patients with a ductal injury who were treated operatively had a significantly higher ISS than those treated nonoperatively (24.0 F 12.6 vs 15.8 F 11.0, P = .025), but no association was seen with a specific organ or organ system. Indications for operative intervention included free intraperitoneal air in 4 children (8.9%), shock or refractory hypotension in 12 (26.7%), and 1 for other cause (2.2%). The pancreas was cited as the cause for operative intervention of 57.8% of all patients and 82.6% of patients with ductal injury.
Table 2 Data results: Comparison of patients treated nonoperatively and operatively. Patients with ductal injuries (Grade III-V) were also isolated and compared All pancreatic injuries Age (y) Sex (% male) ISS GCS Bicycle MVC LOS (d) ICU LOS (d)
Ductal injuries
NOM (n = 128)
Failure (n = 45)
7.7 F 63.3% 15 F 13 F 21.8% 26.6% 9.6 F 4.17 F
7.6 F 3.9 63.2% 27 F 18 10 F 6 32.6% 20.9% 12.1 F 12.0 3.58 F 5.0
4.6 14 4
19.8 19.5
P
NOM (n = 30) 9.2 F 4.1
.917 .916
.675 .675
16 F 14 F 40% 16.7% 13.8 F 1.31 F
11 3
10.1 2.1
Failure (n = 23)
P
6.9 F 3.7 29 F 16 11 F 5 30% 17.4% 9.7 F 8.4 1.60 F 2.5
.025 .884
.668 .139
342
K.D. Mattix et al.
Fig. 1 Morbidity data: Incidence of morbidity factors in patient with pancreatic injury treated by NOM or operatively.
Morbidity factors were obtained in 118 of 173 patients, and 33 of 53 patients with ductal injury. Mean LOS was 9.56 F 19.8 days for NOM patients and 12.08 F 12.0 days for patients managed operatively, whereas ICU stay was 4.17 F 19.8 and 3.58 F 5.0 days, respectively. Patients with ductal injury cared for with NOM had an LOS of 11.33 F 8.0 days, and those requiring an operation, 9.70 F 8.4 days, with a required ICU stay of 1.31 F 2.1 and 1.60 F 2.5 days, respectively (Table 2). Of the patients undergoing NOM for all pancreatic injuries, full morbidity data were available in 83 of 128 patients. Total parenteral nutrition was administered in 28 patients (30.1%) for a mean 4.60 days (range, 3-31 days). Eight patients required a drainage procedure (2 internal and 6 external), whereas 1 patient required an ERCP. Other complications observed were pancreatitis in 26 patients, pseudocyst formation in 12, and pancreatic fistula in 2. Similar data were obtained in 23 of the 53 patients managed operatively. Fifteen patients (60%) received TPN for a mean 13.21days (range, 3-66 days). No patients required an internal drainage procedure, 6 required operative external drainage, and ERCP was performed in 3. Related compli-
cations included 4 patients with pancreatitis; 4 developed pseudocysts, and 1 had a pancreatic fistula (Fig. 1). Analysis of patients with ductal injury demonstrated the following comparisons. Full morbidity data was obtained in 18 of 30 patients managed conservatively, with 1 patient requiring internal drainage and 4 managed by percutaneous external drainage; ERCP was not used in these patients. Pancreatitis developed in 6 patients, pseudocysts in 9, and pancreatic fistula in 1. Nine patients (50%) required TPN for a mean of 9.46 days (range, 2-28 days). Parallel data in the operative patients demonstrated no internal drainage procedures, 3 operative external drainage procedures, and the use of ERCP in 2. Complications included pancreatitis in 3 patients, 3 pseudocysts, and 1 fistula. Total parenteral nutrition was used in 9 patients (60%) for a mean 11.16 days (range, 3-34 days). Additional analysis of those patients with a pancreatic ductal injury was performed. (Table 3) Thirteen patients sustained a grade III injury and were managed nonoperatively, whereas an additional 13 patients with a grade III injury underwent operative intervention. The mean time to intervention was 52.11 hours (range, 0.75-216 hours), with 6 of the 13 patients taken to the operating room (OR) within 24 hours of their injury. Two episodes of pancreatitis were seen in these 6 patients. In contrast, 1 pancreatic fistula and 3 pseudocysts developed in the 7 patients with grade III injuries who underwent operative treatment in delayed fashion (N24 hours). Patients with severe ductal injuries were taken to the OR in a more timely fashion. Eleven patients sustained a grade IV ductal injury and were treated with NOM, whereas 3 patients underwent operative intervention with a mean time to the OR of 42.00 hours (range, 2-100 hours); 1 of the 3 patients was delayed for more than 24 hours. Again, morbidity factors were noticeably different because no complications were seen in those patients with grade IV injury who were managed operatively, whereas 2 pseudocysts and 3 episodes of pancreatitis were noted in NOM patients. Grade V ductal injuries were treated nonopera-
Table 3 Ductal injury morbidity data: Incidence of morbidity factors in patients with ductal injury differentiated by grade and treatment. (No morbidity data was available for the patients with a grade IV injury treated operatively nor those with grade V injury treated with NOM.) No. of patients Morbidity data Mortality Time to OR (mean hours) N 24 h to OR Pseudocyst Pancreatitis Fistula TPN days TPN required
Grade III NOM
Grade III operative
Grade IV NOM
Grade IV operative
Grade V NOM
Grade V operative
13 6 2
13 5 1 52.11
11 8 0
3 0 0 42.00
3 0 0
7 6 1 15.18574
3 3 1 11.33333 3 of 6
6 3 2 1 8.4 4 of 5
1 2 3 8.0625 6 of 8
1 1 1 13.45833 5 of 6
Pediatric pancreatic trauma tively in 3 and operatively in 7 patients; the mean time to operation was 15.19 hours (range, 0.25-59 hours), with 1 patient delayed for more than 24 hours. Of the patients with grade V injuries, 2 patients were taken to the OR for free air and 1 patient for persistent hypotension; all 3 were found to have duodenal injury. Decision to undergo operative intervention in those patients with grade III or IV injuries was based solely on the presence of a pancreatic injury, with 1 exception of hypotension/shock.
3. Discussion Current standard of care for injury to solid organs after blunt trauma is NOM. The optimal management of injury to the pancreas has been more controversial. A small number of studies in both the adult and pediatric literature have assessed the use of ERCP in patients with pancreatic injury [5,6]. Several other studies have analyzed the treatment of pancreatic injuries, but few have compared operative treatment to NOM [1-8]. The data in this study were acquired from multiple pediatric level I trauma institutions and allowed us to accrue a larger number of patients to compare these 2 management strategies. The incidence of pancreatic injury in patients sustaining blunt trauma was 9.2%, which is similar to other reported data [1,3,9]. In addition, as in other series, these patients had significant morbidity related to their injury. Nonoperative management was successful in most (74%) of the entire cohort of patients with pancreatic injury. This is quite different from the reported NOM success rate of 5% in all pediatric patients with abdominal trauma [2]. Overall, there did not appear to be any advantage in one method of management compared with another in children with pancreatic injuries. However, our data suggest that there may be an advantage in early operation in those who have ductal injuries. Although pseudocysts occurred less frequently in our series compared with the reported literature (13%-30% vs 40%100%) [1,3,9,10] these data appear to indicate that this was more common in those with ductal injury, who were treated conservatively, compared with those who had early operation. This is now increasingly recognized, and there have been several reports of distal ductal injury (grades IIIIV or class II), which appear to demonstrate decreased morbidity where these patients are treated early with aggressive surgical management, most often a distal pancreatectomy [1,8,7]. Injury severity score will be expected to correlate with the comorbidities in patients with pancreatic injuries. In patients with a ductal injury, only increased ISS was found to be a significant factor correlating with the perceived need for operative treatment. Unfortunately, complete data were not recorded for all patients for the specific comorbidities. Despite this, it is clear from the ISS that those that came to surgery were subject to more multisystem injuries, including
343 pancreatic ductal disruption, compared to those who were treated conservatively (Table 3). Patients with a ductal injury had a reduction in the incidence of complications when taken to the OR within 24 hours of injury. Therefore, early and accurate diagnosis of ductal injury and prompt operative intervention can be expected to result in fewer complications and decreased morbidity. There are several limitations to this study. The low incidence of pancreatic injury secondary to blunt trauma makes accurate assessment of this patient population difficult. Although the inclusion of multiple institutions increases the number of patients, there is a risk of variation in management, including the threshold for operative intervention, the indications for parenteral nutrition, or the availability of ERCP and helical computed tomography (CT). Similarly, objective and consistent description of pancreatic ductal injury may differ between institutions. Despite these limitations, our findings demonstrate that all children with pancreatic injury have significant morbidity. Initial diagnosis should ideally be made using helical CT imaging. If ductal injury is equivocal based on CT findings, ERCP has become an accepted diagnostic method to evaluate for pancreatic ductal injury [5-8]. Additionally, ERCP may also develop into a valuable option for therapeutic intervention, including stent placement, in proximal ductal injuries and should be considered in those patients [5,8]. Finally, although children who have suffered pancreatic injury (but without ductal disruption) do not appear to suffer increased morbidity after conservative management, these data suggest that patients with ductal disruption may benefit from early operative intervention.
4. Q&A Marven, Sheffield, UK In addition to CT, how many of these patients had MR or ERCP? A. Of the patients with ductal injury, 2 patients, of a total of 53, had ERCP as a diagnostic tool and then went on to operative intervention. Marven, Sheffield, UK What were the different operative interventions? A. All of them were distal pancreatectomies or subtotal pancreatectomies with the exception of one who had a Roux loop to the pancreatic duct for a very proximal injury. Marven, Sheffield, UK No patients had a Roux loop to the distal pancreatic duct. A. No. Langer, Toronto, Canada A number of years ago, we described 14 consecutive children with a complete transection of the pancreatic duct,
344 which were managed nonoperatively. Of your patients who had significant grade 3-5 injuries, were there any regional or within-institution differences with respect to surgeons’ preference or indications for operative interventions or what were considered failure of the NOM? A. I was unable to obtain that. We had a blinded study; once the patients were given to us, the institution was unknown. So I am unable to answer that question, I’m sorry. Ade-Aji, London, UK We recently also recommended the use of ERCP, a strategy which has evolved in the context of delayed referral pattern. Given that you are seeing patients acutely, what do you think is the role for ERCP and stents at this stage? A. I believe that that role becomes important when you have a patient who is brought in initially, and I also believe that role is limited to your experience within the institution. We have very skilled gastroenterologists who are able to perform ERCP for us in younger children. With that, I think that the evidence and a couple of recent papers have shown that ERCP can be a very effective diagnostic means. Also, there has been a recent paper published, which shows that stenting is a viable option. Again, I think that it
K.D. Mattix et al. becomes important when patients present and then the quality of your experience at the institution.
References [1] Jobst M, Canty T, Lynch F. Management of pancreatic injury in pediatric blunt abdominal trauma. J Pediatr Surg 1999;34:818 - 24. [2] Stylianos S. Evidence-based guidelines for resource utilization in children with isolated spleen of liver injury. J Pediatr Surg 2000; 35:164 - 9. [3] Shilyansky J, Sena L, Dreller M, et al. Non-operative management of pancreatic injuries in children. J Pediatr Surg 1998;33:343 - 9. [4] Loungnarath R, Blanchard H, Saint-Vil D, et al. Blunt pancreatic injuries in children. Ann Chir 2001;126:992 - 5. [5] Canty T, Weinman D. Treatment of pancreatic duct disruption in children by an endoscopically placed stent. J Pediatr Surg 2001;36:345 - 8. [6] Rescorla F, Plumley D, Sherman S, et al. The efficacy of early ERCP in pediatric pancreatic trauma. J Pediatr Surg 1995;30:336 - 40. [7] Gaines B, Ford H. Abdominal and pelvic trauma in children. Crit Care Med 2002;30:S416 - 23. [8] Canty T, Weinman D. Management of major pancreatic injuries in children. J Trauma 2001;50:1001 - 5. [9] Meier D, Coln C, Hicks B, et al. Early operation in children with pancreas transaction. J Pediatr Surg 2001;36:341 - 4. [10] Jacombs A, Wines M, Holland A, et al. Pancreatic trauma in children. J Pediatr Surg 2004;39:96 - 9.