- Email: [email protected]
Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis
ARTICLE IN PRESS
JID: JINJ
[m5G;February 20, 2020;19:33]
Injury xxx (xxxx) xxx
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis Mario Kopljar a,b, Stjepan Ivandic´ b, Marko Mesic´ c, Bore Bakota d, Tihomil Žiger b,e,∗, h ˇ Goran Kondža a,f, Roman Pavic´ a,b, Miloševic´ Milan g, Miran Coklo a
Medical Faculty Osijek, Josip Juraj Strossmayer University of Osijek, Osijek, Croatia University Hospital Center “Sestre milosrdnice”, Zagreb, Croatia Children’s Hospital, Zagreb, Croatia d Trauma and Orthopedic Department, Landesklinikum Horn, Austria e Faculty of Dental Medicine, Josip Juraj Strossmayer University of Osijek, Osijek, Croatia f University Hospital Center Osijek, Croatia g Polyclinic Ribnjak, Zagreb, Croatia h Institute for Anthropological Research, Zagreb, Croatia b c
a r t i c l e
i n f o
Article history: Accepted 9 February 2020 Available online xxx Keywords: Pancreatic injury Abdominal trauma Children Pancreatic trauma
a b s t r a c t AIM: Blunt abdominal trauma is the major cause of abdominal injury in children. No clear guidelines exist for the initial management of blunt pancreatic trauma in children. The aim of this study was to perform a systematic review and meta-analysis of initially non-operative versus initially operative treatment in children with blunt pancreatic injury. METHODS: Studies including children (<18 years) with blunt pancreatic injuries published in any language after year 1990 were included. Total of 849 studies were identified by searching PubMed, Scopus, CINAHL and Cochrane Database. After review, 42 studies met inclusion criteria and were included in this systematic review. There were 1754 patients, of whom 1095 were initially managed non-operatively (NOM), and 659 operatively (OM). Primary outcome was non-operative management success rate, and secondary outcomes were mortality, complications (including specifically pseudocysts and pancreatic fistulas), percent of patients and days on total parenteral nutrition (TPN), length of hospital stay and readmissions. RESULTS: There was no difference in mortality between NOM and OM groups. The incidence of pseudocysts was significantly higher in NOM group compared to OM (P<0.001), especially for AAST grade III or higher (P<0.0 0 0 01). Overall incidence of pancreatic fistulas was significantly lower for NOM group (p = 0.02) but no difference was observed for AAST grades III or higher (p = 0.49). There was no difference in the length of hospital stay (P = 0.31). Duration of total parenteral nutrition was not different for all AAST grades (P = 0.35), but was significantly shorter for OM group for AAST grades III and higher (p = 0.0 0 01). There was no overall difference in readmissions (P = 0.94). Overall success rate of initial non-operative treatment was 87%. CONCLUSIONS: Most patients with pancreatic trauma can initially be treated non-operatively, while early surgical treatment may benefit patients with lesions of the main pancreatic duct. ERCP offers both highly accurate diagnosis and potential treatment of ductal injuries. © 2020 Elsevier Ltd. All rights reserved.
Introduction Blunt abdominal trauma is the major cause of abdominal injury in children. Pancreatic trauma, although infrequent, can be associated with major mortality and morbidity including hemorrhage, abscesses, fistulae, pancreatic leaks and pancreatitis [1]. ∗
Corresponding author E-mail address: [email protected] (M. Kopljar).
Reported incidence of pancreatic injuries in children admitted for blunt abdominal trauma is less than 10% [2–4]. Pancreas is the fourth most commonly injured solid organ following the spleen, liver and kidneys [5]. Unlike pediatric splenic and hepatic injuries, in which non-operative management (NOM) is widely accepted standard of care, the optimal management strategy of pancreatic injuries remains a subject of debate [6]. Depending on the integrity of the main pancreatic duct pancreatic injuries can be divided into minor and major injuries based on
https://doi.org/10.1016/j.injury.2020.02.035 0020-1383/© 2020 Elsevier Ltd. All rights reserved.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
ARTICLE IN PRESS
JID: JINJ
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
2 Table 1 The AAST Pancreatic Injury Classification. Grade
∗
Type of injury
Description of injury
I I II
Hematoma Laceration Hematoma
II
Laceration
III
Laceration
IV
Laceration
V
Laceration
Minor contusion without duct injury Superficial laceration without duct injury Major contusion without duct injury or tissue loss Major laceration without duct injury or tissue loss Distal transection or parenchymal injury with duct injury Proximal† transection or parenchymal injury involving ampulla Massive disruption of pancreatic head
∗
Advance 1 grade for multiple injuries up to grade III. †Proximal pancreas is to the patients’ right of the superior mesenteric vein.
the American Association for the Surgery of Trauma (AAST) Organ Injury Scale (Table 1). AAST I and II injuries indicate parenchymal injuries without pancreatic duct involvement while AAST III, IV and V include ductal injuries with laceration or transection [7]. Management guidelines in adult blunt pancreatic trauma are fairly clear, recommending non-operative management for patients with grade I/II injuries and pancreatic resection with drainage for grade III or higher injuries. No guidelines exist for blunt pancreatic trauma in children and it is currently unclear whether the same approach should be applied in children [8]. There is a general agreement that grades I and II pancreatic injuries can be safely managed without surgery. For major injuries (AAST ≥ III), both operative and non-operative management strategies have been advocated. Advocates of operative management (OM) argue that surgery is associated with faster recovery, fewer complications and that delay in operative management leads to greater morbidity [9–12]. Proponents of non-operative management argue that it is safe and effective and that it avoids unnecessary surgery-related complications and mortality [13,14]. Operative treatment in all patients with pancreatic trauma is generally indicated if the patient’s clinical status worsens, showing signs of peritoneal irritation or patients become hemodynamically unstable [6]. Studies that have compared the outcomes of NOM and OM are retrospective studies with small sample sizes and a significant risk of selection bias. We therefore decided to perform a systematic review and meta-analysis of pertinent studies to determine benefits and harms of non-operative versus operative treatment in children with blunt pancreatic injury. Methods Study selection An electronic systematic literature search was performed on February 28th 2019, using PubMed, Scopus, CINAHL and the Cochrane Library. All titles and abstracts were independently screened by two authors for eligibility. Furthermore, references of studies considered for inclusion were scanned for and additional references. The search included studies published after year 1990. (Fig. 1). No language restrictions were applied. The following search strategy was used for PubMed: “(((trauma[tiab]) OR blunt injury [MeSH]) AND (pancreas[Title/Abstract] OR pancreatic[Title/Abstract])) AND children [MeSH Terms].” Search strategy was modified accordingly for other respective databases.
Fig. 1. Flow diagram (PRISMA) illustrating publication selection.
ies reporting on at least 5 children (age 18 or less) with nonoperative or operative management of blunt pancreatic trauma. Excluded were studies published prior to 1990 and those including adult patients, unless pediatric patients were described separately. Furthermore, publications were excluded if they were case reports, reviews or studies including penetrating trauma.
Eligibility criteria
Types of interventions
After title and abstract screening, the same two reviewers independently reviewed full text articles. Inclusion criteria were stud-
Operative management (OM) was defined as immediate surgery upon establishing the diagnosis of pancreatic injury. Nonoperative
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
management (NOM) was defined as observation, medical treatment, endoscopic or interventional radiology procedures (such as ERCP), as well as surgical procedures for injuries of other organ systems (e.g. head trauma).
Table 2 Characteristics of included studies.
Abbo, 2013 [13] Arkovitz, 1997 [42] Bass, 1991 [43] Beger, 2019 [17] Ben Hassine, 2010 [18] Beres, 2013 [19] Borkon, 2011 [44] Canty, 2001 [20] Cigdem, 2011 [21] Cuenca, 2012 [22] Dai, 2015 [47] de Blaauw, 2008 [14] Fabbro, 2001 [45] Firstenberg, 1999 [23] Garg, 2017 [24] Ghasoup, 2013 [25] Graham, 2000 [46] Holland, 1999 [26] Iqbal, 2014 [9] Jobst, 1999 [5] Juric, 2009 [27] Katz, 2018 [48] Keil, 2016 [28] Keller, 1997 [2] Klin, 2011 [29] Kouchi, 1999 [30] Loungnarath, 2001 [31] Mattix, 2007 [3] Meier, 2001 [11] Mora, 2016 [32] Nadler, 1999 [55] Paul, 2011 [33] Ruszinko, 2005 [34] Shaked, 2003 [35] Shilyansky, 1998 [36] Snajdauf, 2007 [10] Stringer, 2005 [12] Takishima, 1996 [37] Vane, 2009 [38] Wales, 2001 [39] Wood, 2010 [40] Yamana, 2012 [41] TOTAL
Quality assessment We assessed risk of bias at the study level using Cochrane’s ‘Risk of bias’ tool. Two review authors independently assessed the methodologic quality of each included study and resolved their disagreements by discussion. ‘Risk of bias’ criteria included: adequate sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, as well as other potential biases. Each criterion was rated as having ‘low’ risk of bias, ‘high’ risk of bias, or ‘unclear’ risk of bias, which indicated the lack of information or uncertainty regarding the potential for bias. Additionally, individual studies were graded for level of evidence using criteria from the Oxford Center for Evidence-Based Medicine. Quality assessment was performed using the NewcastleOttawa Scale, and studies with a score lower than 6 were considered low quality. Outcomes Primary outcomes were mortality and complications. Secondary outcomes were the success rate of nonoperative management, percent of patients and days on total parenteral nutrition (TPN), length of hospital stay and readmissions. Outcomes were analyzed separately for minor (AAST grades I and II) and major (AAST grades III or higher) (Table 1) whenever feasible. Statistical analysis Data obtained from eligible studies was collected into an electronic spread-sheet, and statistical analysis was performed using RevMan 5.3 software. Weighted treatment effects were calculated using random effects models. For dichotomous outcomes, results were expressed as odds ratios (OR) with 95% confidence intervals (CI). Where continuous scales of measurement were used to assess the outcomes of treatment, mean difference (MD) was used [15]. Assessment of heterogeneity Heterogeneity was assessed using the I2 test (16). An I2 statistic value of less than 40% was regarded as not showing any significant heterogeneity. Assessment of publication biases Funnel plots were used to assess reporting biases. Funnel plots were assessed qualitatively. Results Initial search of electronic databases revealed 849 potential studies, as well as 5 additional studies found in the references of these studies (Fig. 1). After duplicate studies were removed and full texts of studies scrutinized according to inclusion and exclusion criteria, there were 42 publications that met inclusion criteria for this systematic review (Table 2). All publications were retrospective, with high selection biases (Fig. 2). All of the studies were of low quality according to Newcastle-Ottawa Scale. In total, there
3
∗ ∗∗
NOP∗
OP∗∗
%NOP success∗∗∗
36 13 32 17 4 24 0 9 25 54 25 31 7 8 14 6 15 14 95 26 7 5 6 112 6 19 13 128 2 194 24 23 4 25 28 0 8 7 5 9 11 4 1095
0 13 8 0 1 15 25 9 6 25 8 3 2 1 1 4 1 0 72 30 0 8 18 42 4 1 8 45 9 202 27 20 3 0 7 13 1 1 9 0 14 3 659
100 – 100 82.4 – 91.3 – 22 96 96.3 100 90 86 87.5 71.4 100 80 79 92 80.8 100 – 100 – 100 84 100 74 0 – 79 87 100 80 100 – 38 100 100 100 – 100 87.0%
Non-operative management. Operative management. Percent of patients successfully managed non-operatively.
∗∗∗
were 1754 pediatric patients, of whom 1095 were managed nonoperatively (NOM), and 659 operatively (OM). Non-operative treatment was successful in 87% of the 1095 patients that were initially treated non-operatively.
Pseudocysts Information about pseudocyst formation was available in 32 studies (3, 5, 9–14, 17–40). Ten studies were available for metaanalysis. There were 858 children treated with NOM and 498 were treated with OM. Pseudocyst development was reported in 214 patients treated with NOM (24.9%) and 37 patients treated with OM (7.4%). Meta-analysis demonstrated a statistically significant higher risk of pseudocyst formation in the NOM group (OR 3.05, 95% CI 1.84 to 5.05; P<0.001). Heterogeneity was low (I2 = 14%, P = 0.28) and can be noted in the funnel plot (Fig. 3). When only patients with grade III or higher were analyzed, meta-analysis also demonstrated a statistically significant higher risk of pseudocyst formation in the NOM group (OR 12.46, 95% CI 4.60 to 33.77; P<0.0 0 0 01). Heterogeneity was low (I2 = 10%, P = 0.35).
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ 4
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
Fig. 2. Risk of bias graph: review authors’ judgments about each risk of bias item presented as percentages across all included studies.
Fig. 3. Funnel plot of comparison: non-operative vs. operative, outcome: pseudocysts.
Fisulas Information about pancreatic fistulas was available in seven studies (6, 23, 32–34, 38, 41). There were 289 children treated with NOM and 157 were treated with OM. Pancreatic fistula developed in 3 patients treated with NOM (1.0%) and 12 patients treated with OM (7.6%). Meta-analysis demonstrated a statistically significant lower risk of pancreatic fistula in the NOM group (OR 0.27, 95% CI 0.09 to 0.82; P = 0.02). Heterogeneity was low (I2 = 0%, P = 0.87) and can be noted in the funnel plot (Fig. 4). However, when only patients with grade III or higher were analyzed, meta-analysis did not demonstrate a difference in the risk of fistula (OR 0.56, 95% CI 0.11 to 2.91; P = 0.49). Heterogeneity was low (I2 = 0%, P = 0.77). Mortality Mortality due to pancreatic trauma was reported in 32 studies (1–4, 6–9, 11–15, 17, 18, 22, 27–32, 34, 36, 38, 40–46). Mortal-
ity from other causes (mostly severe head trauma or other causes not related to pancreatic injury) was not considered in this study. There were 741 children treated with NOM and 430 were treated with OM. There were 14 deaths reported in patients treated with NOM (1.9%) and 12 patients treated with OM (2.8%). Meta-analysis did not demonstrated a statistically significant difference in the risk of mortality between groups (OR 0.96, 95% CI 0.42 to 2.22; P = 0.93). Heterogeneity was low (I2 = 0%, P = 0.86) and can be noted in the funnel plot (Fig. 5). Length of hospital stay Information about the length of hospital stay was available in 18 studies (1, 2, 4, 6, 9, 12, 15, 22, 23, 27, 32–34, 36, 38, 40, 47, 48). There were 667 children treated with NOM and 403 were treated with OM. Meta-analysis demonstrated slightly shorter length of stay in the NOM group, although not significantly (MD −1.26, 95% CI −3.68 to 1.15; P = 0.31). Heterogeneity was high (I2 = 70%, P = 0.0 0 05) and can be noted in the funnel plot (Fig. 6).
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
5
Fig. 4. Funnel plot of comparison: non-operative vs. operative, outcome: fistulas.
Fig. 5. Funnel plot of comparison: non-operative vs. operative, outcome: mortality.
Seven studies reported hospital stay for patients with grade III or higher pancreatic injury (1, 6, 15, 23, 34, 40, 47), and only two studies provided sufficient data for meta-analysis (6, 23). Although meta-analysis is not reliable for only two studies, hospital stay was significantly shorter in OM group (MD 7.57, 95% CI 0.53 to 14.61; P = 0.04). Heterogeneity was moderate (I2 = 54%, P = 0.14). Total parenteral nutrition Information about the duration of total parenteral nutrition (TPN) was available in 9 studies (1, 6, 7, 22, 23, 33, 34, 38, 47), but only 4 were available for meta-analysis. There were 331 children treated with NOM and 186 were treated with OM. Meta-analysis demonstrated slightly shorter duration of TPN in the OM group, al-
though not significantly (MD 4.86, 95% CI −5.34 to 15.05; P = 0.35). Heterogeneity was high (I2 = 90%, P = 0.0 0 0 01) and can be noted in the funnel plot (Fig. 7). Five studies reported hospital stay for patients with grade III or higher pancreatic injury (6, 7, 23, 34, 47), and only three studies provided sufficient data for meta-analysis (6, 23, 34). Although meta-analysis is not reliable for only three studies, duration of TPN was significantly shorter in OM group (MD 10.98, 95% CI 5.48 to 16.48; P = 0.0 0 01). Heterogeneity was low (I2 = 15%, P = 0.31). Information about percentage of patients requiring total parenteral nutrition (TPN) was available in 5 studies (1, 12, 22, 33, 38). There were 232 children treated with NOM and 68 were treated with OM. Total parenteral nutrition was required in 84 patients treated with NOM (36.2%) and 20 patients treated with OM (29.4%). Meta-analysis did not demonstrate any difference in the
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ 6
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
Fig. 6. Funnel plot of comparison: non-operative vs. operative, outcome: length of hospital stay.
Fig. 7. Funnel plot of comparison: non-operative vs. operative, outcome: total parenteral nutrition (days).
risk of TPN (OR 1.17, 95% CI 0.11 to 12.17; P = 0.90). Heterogeneity was high (I2 = 83%, P = 0.002). Readmissions Information about readmissions was available in 3 studies (7, 23, 48) and two studies were available for meta-analysis. There were 106 children treated with NOM and 111 were treated with OM. Readmissions occurred in 19 patients treated with NOM (17.9%) and 26 patients treated with OM (23.4%). Meta-analysis did not demonstrate any difference in the risk of readmissions (OR 1.08, 95% CI 0.18 to 6.47; P = 0.94). Heterogeneity was high (I2 = 68%, P = 0.08) and can be noted in the funnel plot (Fig. 8).
Discussion Pancreatic injury represents 0.3% of all pediatric traumas and 0.6% of significant abdominal trauma [49]. Pancreatic trauma remains a morbid condition, with a mortality rate of 5% and major complications affecting approximately 1 in 4 children [50]. Pancreatic trauma is the fourth most common solid organ injury, right after the spleen, liver and kidneys, and it is very frequently associated with other significant injuries, such as traumatic brain injury, bone and joint trauma, urogenital injuries etc. [5,23]. The majority of pancreatic injuries arise from motor vehicle accidents, but bike injuries (handlebar injuries) and struck injuries are also frequent causes [49].
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
7
Fig. 8. Funnel plot of comparison: non-operative vs. operative, outcome: readmissions.
The diagnosis of pancreatic injury is difficult due to the fact that it is a retroperitoneal organ and symptoms of pancreatic injury maybe non-specific. Therefore, some authors estimate the incidence of pancreatic trauma to be as high as 9% of all injured children [3,40]. Furthermore, pancreatic trauma is often associated with other abdominal injuries, such as duodenal injuries, that have pronounced clinical picture that may mask pancreatic injury. Increased amylase and lipase levels may be helpful, and are often elevated in pancreatic trauma, although they may be normal in the first few hours after the trauma, and it may take several serial investigations to demonstrate and increase in serum amylase or lipase levels [3,16,43]. Increase serum amylase levels are also not specific to pancreatic trauma and may be elevated in cases of head and neck trauma that is often associated with pancreatic trauma). Serum lipase measurements may be used in these cases to exclude extrapancreatic trauma [5]. The use of ultrasound has been demonstrated to be inaccurate in the detection of pancreatic injury [3]. It can, however, be used for the detection of pancreatic pseudocysts. Computed tomography (CT scan) remains most utilized for the diagnosis of pancreatic trauma in both children and adults with accuracy rates as high as 91%. Direct or indirect signs of pancreatic injury such as complete gland transection, greater than 50% laceration through the gland, or the presence of early peripancreatic fluid, may be helpful in determining the presence of duct injury. Endoscopic retrograde cholangiopancreatography (ERCP) is routinely used to assess pancreatic duct pathology in adults and is an accurate method of investigating the pancreatic duct injury but also has drawbacks, especially in pediatric population, due to its invasive nature and associated risks of postprocedural pancreatitis [51]. Magnetic resonance cholangiopancreatography can also be used as a non-invasive alternative to ERCP. On the other hand, ERCP can be utilized not only for the diagnosis of pancreatic injury, but also in therapeutic setting, as it enables stenting of most pancreatic duct injuries [52]. The optimal management of pancreatic trauma continues to be the source of controversy among pediatric surgeons. Previously published Cochrane systematic review could not identify any randomized controlled studies comparing operative and non-operative
management of pancreatic trauma in children, and due to the nature of this condition it is unlikely that any such randomized trial could ever be performed [21]. Today, the standard of care for injuries to solid organs in hemodynamically stable patients following blunt abdominal trauma is non-operative management. Most minor pancreatic injuries go unrecognized and often heal spontaneously. It is therefore a common approach to treat minor pancreatic injuries (grades I and II) non-operatively. However, major pancreatic trauma (AAST grades III and higher) often requires operative approach and, if surgery is delayed, operative intervention may be rather complicated [10]. The results of this review confirms that non-operative management is justified in patients with grades I or II pancreatic trauma, as there was no significant difference in mortality or the formation of pseudocysts, although pseudocysts were more common in patients treated non-operatively. However, in most cases pseudocysts also be treated non-operatively by the means of endoscopic or percutaneous drainage [52-54] and the success rate of non-operative treatment was found to be as high as 87%. Published studies included in this review did not report significant morbidities related to delayed surgery for pseudocysts. Some studies reported longer hospital stay for non-operatively treated patients with minor pancreatic trauma but the quality of data did not allow meta-analysis [42]. For patients with major pancreatic trauma (grades III or higher according to AAST classification), there was also no difference in mortality regarding initial management strategy (operative versus non-operative). However, in these patients, significantly more pseudocysts were reported in patients who were initially treated non-operatively. Also, patients who were initially treated with surgery had shorter length of hospital stay and shorter duration of total parenteral nutrition. Finally, the frequency of readmission was similar between patients who were initially treated operatively and non-operatively. There is an obvious benefit in early surgical approach for patients with major pancreatic trauma with ductal injury. It is therefore important to tailor the diagnostic approach so that pancreatic duct injury can be accurately diagnosed as soon as possible. While CT scan and ultrasound may clearly indicate ductal lesions in some cases, ERCP may offer more definite diagnosis in over 97% of
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
8
patients, and if the ductal lesion is found, either early surgery or endoscopic stenting of the injured pancreatic duct can be performed (52). Conclusion Most patients with pancreatic trauma can initially be treated non-operatively, while early surgical treatment may benefit patients with lesions of the main pancreatic duct. This stresses the importance of accurate and timely diagnosis of pancreatic trauma. Initial diagnostic workup should include CT scan, primarily to exclude other serious or life-threatening injuries, and in cases of suspected pancreatic duct injuries, ERCP offers both highly accurate diagnosis and potential treatment of ductal injuries. Funding This paper is part of a supplement supported by The Croatian Trauma Society. Declaration of Competing Interest All authors declare no conflict of interest. References [1] Jacombs ASW, Wines M, Holland AJA, Ross FI, Shun A, Cass DT. Pancreatic trauma in children. J Pediatr Surg 2004;39(1):96–9. [2] Keller MS, Stafford PW, Vane DW. Conservative management of pancreatic trauma in children. J of Trauma 1997;42(6):1097–100. [3] Mattix KD, Tataria M, Holmes J, Kristoffersen K, Brown R, Groner J, et al. Pediatric pancreatic trauma: predictors of nonoperative management failure and associated outcomes. J Pediatr Surg 2007;42(2):340–4 Epub 2007/02/03. [4] Garvey EM, Haakinson DJ, McOmber M, Notrica DM. Role of ERCP in pediatric blunt abdominal trauma: a case series at a level one pediatric trauma center. J Pediatr Surg 2015;50(2):335–8 Epub 2015/02/02. [5] Jobst MA, Canty TG Sr, Lynch FP. Management of pancreatic injury in pediatric blunt abdominal trauma. J Pediatr Surg 1999;34(5):818–23 discussion 23-4. Epub 1999/06/08. [6] Haugaard MV, Wettergren A, Hillingso JG, Gluud C, Penninga L. Non-operative versus operative treatment for blunt pancreatic trauma in children. Cochrane Database Syst Rev 2014(2) CD009746. Epub 2014/02/14. [7] Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, Champion HR, Gennarelli TA, et al. Organ injury scaling, II: pancreas, duodenum, small bowel, colon, and rectum. J Trauma 1990;30(11):1427–9 Epub 1990/11/01. [8] Ho VP, Patel NJ, Bokhari F, Madbak FG, Hambley JE, Yon JR, et al. Management of adult pancreatic injuries: a practice management guideline from the eastern association for the surgery of trauma. J Trauma Acute Care Surg 2017;82(1):185–99 Epub 2016/10/28. [9] Iqbal CW, St Peter SD, Tsao K, Cullinane DC, Gourlay DM, Ponsky TA, et al. Operative vs nonoperative management for blunt pancreatic transection in children: multi-institutional outcomes. J Am Coll Surg 2014;218(2):157–62 Epub 2014/01/21. [10] Snajdauf J, Rygl M, Kalousova J, Kucera A, Petru O, Pycha K, et al. Surgical management of major pancreatic injury in children. Eur j of pediatr surg 2007;17(5):317–21 Epub 2007/10/31. [11] Meier DE, Coln CD, Hicks BA, Guzzetta PC. Early operation in children with pancreas transection. J Pediatr Surg 2001;36(2):341–4. [12] Stringer MD. Pancreatic trauma in children. Br J Surg 2005;92(4):467–70 Epub 2005/01/27. [13] Abbo O, Lemandat A, Reina N, Bouali O, Ballouhey Q, Carfagna L, et al. Conservative management of blunt pancreatic trauma in children: a single center experience. Eur J of Pediatr Surg 2013;23(6):470–3. [14] de Blaauw I, Winkelhorst JT, Rieu PN, van der Staak FH, Wijnen MH, Severijnen RSVM, et al. Pancreatic injury in children: good outcome of nonoperative treatment. J Pediatr Surg 2008;43(9):1640–3. [15] Thompson SG, Higgins JP. How should meta-regression analyses be undertaken and interpreted? Stat Med 20 02;21(11):1559–73 Epub 20 02/07/12. [16] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. Bmj 2003;327(7414):557–60 Epub 2003/09/06. [17] Beger B, Kizilyidiz BS, Simsek M, Akdeniz H, Sonmez B. Pancreatic injury due to blunt abdominal trauma in children. Med Sci (Turkey) 2019;8(1):225–8. [18] Ben Hassine L, Boughanmi N, Douira W, Mormeche J, Louati H, Hamzaoui M, et al. [Management of blunt duodenal and pancreatic injuries in children (about a series of 8 cases)]. Tunis Med 2010;88(8):586–92 Epub 2010/08/17. Prise en charge des traumatismes fermes du bloc duodeno- pancreatique chez l’enfant (a propos d’une serie de 8 cas).
[19] Beres AL, Wales PW, Christison-Lagay ER, McClure ME, Fallat ME, Brindle ME. Non-operative management of high-grade pancreatic trauma: is it worth the wait? J Pediatr Surg 2013;48(5):1060–4 Epub 2013/05/25. [20] Canty TG Sr, Weinman D, Tepas JJ, Mullins RJ, Winthrop A, Powell RW. Management of major pancreatic duct injuries in children. J Trauma Inj Infect Crit Care 20 01;50(6):10 01–7. [21] Cigdem MK, Senturk S, Onen A, Siga M, Akay H, Otcu S. Nonoperative management of pancreatic injuries in pediatric patients. Surg Today 2011;41(5):655–9 Epub 2011/05/03. [22] Cuenca AG. Islam S. pediatric pancreatic trauma: trending toward nonoperative management? Am Surg 2012;78(11):1204–10. [23] Firstenberg MS, Volsko TA, Sivit C, Stallion A, Dudgeon DL, Grisoni ER. Selective management of pediatric pancreatic injuries. J Pediatr Surg 1999;34(7):1142–7 Epub 1999/08/12. [24] Garg RK, Mahajan JK. Blunt trauma pancreas in children: is non-operative management appropriate for all grades? Pediatr Gastroenterol Hepatol Nutr 2017;20(4):252–8. [25] Ghasoup A, Sadieh O, Mansor A, Alani Q, Sawaqed J. Management of pancreatic injuries with blunt abdominal trauma in children. Case Ser. J Trauma Treat 2013:2. [26] Holland AJ, Davey RB, Sparnon AL, Chapman M, LeQuesne GW. Traumatic pancreatitis: long-term review of initial non-operative management in children. J Paediatr Child Health 1999;35(1):78–81 Epub 1999/05/11. [27] Juric´ I, Pogorelic´ Z, Biocˇ ic´ M, Todoric´ D, Furlan D, Šušnjar T. Management of blunt pancreatic trauma in children. Surg. Today 2009;39(2):115–19. [28] Keil R, Drabek J, Lochmannova J, Stovicek J, Rygl M, Snajdauf J, et al. What is the role of endoscopic retrograde cholangiopancreatography in assessing traumatic rupture of the pancreatic in children? Scand J Gastroenterol 2016;51(2):218–24 Epub 2015/07/23. [29] Klin B, Abu-Kishk I, Jeroukhimov I, Efrati Y, Kozer E, Broide E, et al. Blunt pancreatic trauma in children. Surg Today 2011;41(7):946–54 Epub 2011/07/13. [30] Kouchi K, Tanabe M, Yoshida H, Iwai J, Matsunaga T, Ohtsuka Y, et al. Nonoperative management of blunt pancreatic injury in childhood. J Pediatr Surg 1999;34(11):1736–9 Epub 1999/12/11. [31] Loungnarath R, Blanchard H, Saint-Vil D. Blunt pancreatic injuries in children. Ann Chir 2001;126(10):992–5. [32] Mora MC, Wong KE, Friderici J, Bittner K, Moriarty KP, Patterson LA, et al. Operative vs nonoperative management of pediatric blunt pancreatic trauma: evaluation of the national trauma data bank. J Am Coll Surg 2016;222(6):977–82. [33] Paul MD, Mooney DP. The management of pancreatic injuries in children: operate or observe. J Pediatr Surg 2011;46(6):1140–3 Epub 2011/06/21. [34] Ruszinko V, Willner P, Olah A. Pancreatic injury from blunt abdominal trauma in childhood. Acta Chir Belg 2005;105(3):283–6 Epub 2005/07/16. [35] Shaked G, Kleiner O, Finally R, Mordechai J, Newman N, Cohen Z. Management of blunt pancreatic injuries in children. Eur J Trauma 2003;29(3):151–5. [36] Shilyansky J, Sena LM, Kreller M, Chait P, Babyn PS, Filler RM, et al. Nonoperative management of pancreatic injuries in children. J Pediatr Surg 1998;33(2):343–9 Epub 1998/03/14. [37] Takishima T, Sugimoto K, Asari Y, Kikuno T, Hirata M, Kakita A, et al. Characteristics of pancreatic injury in children: a comparison with such injury in adults. J Pediatr Surg 1996;31(7):896–900 Epub 1996/07/01. [38] Vane DW, Kiankhooy A, Sartorelli KH, Vane JL. Initial resection of potentially viable tissue is not optimal treatment for grades ii-iv pancreatic injuries. World J Surg 2009;33(2):221–7 Epub 2008/04/12. [39] Wales PW, Shuckett B, Kim PCW. Long-term outcome after nonoperative management of complete traumatic pancreatic transection in children. J Pediatr Surg 2001;36(5):823–7. [40] Wood JH, Partrick DA, Bruny JL, Sauaia A, Moulton SL. Operative vs nonoperative management of blunt pancreatic trauma in children. J Pediatr Surg 2010;45(2):401–6 Epub 2010/02/16. [41] Yamana I, Kawamoto S, Kamitani T, Ishikura H, Yamashita Y. Pancreatic injury in children: review of 7 cases and the pertinent literature. Hepatogastroenterology 2012;59(114):574–7 Epub 2011/09/24. [42] Arkovitz MS, Johnson N, Garcia VF. Pancreatic trauma in children: mechanisms of injury. J of Trauma 1997;42(1):49–53. [43] Bass DH, Lakhoo K. Pancreatic injuries in children. S Afr j of sur Suid-Afrik tydskr vir chirur 1991;29(2):39–40 Epub 1991/06/01. [44] Borkon MJ, Morrow SE, Koehler EA, Shyr Y, Hilmes MA, Miller RS, et al. Operative intervention for complete pancreatic transection in children sustaining blunt abdominal trauma: revisiting an organ salvage technique. Am Surg 2011;77(5):612–20. [45] Fabbro MA, D’Agostino S, Romanini MV, Chiarenza F, Fasoli L, Musi L, et al. Management of severe blunt pancreatic trauma in children. Personal exp La Pediatria medi e chirur 2001;23(3–4):179–82 Epub 2001/11/29. [46] Graham CA, O’Toole SJ, Watson AJ, Munro FD, Haddock G. Pancreatic trauma in scottish children. J R Coll Surg Edinb 20 0 0;45(4):223–6 Epub 20 0 0/12/29. [47] Dai LN, Chen CD, Lin XK, Wang YB, Xia LG, Liu P, et al. Abdominal injuries involving bicycle handlebars in 219 children: results of 8-year follow-up. Eur J of Trauma & Emerg Surg 2015;41(5):551–5. [48] Katz MG, Fenton SJ, Russell KW, Scaife ER, Short SS. Surgical outcomes of pancreaticoduodenal injuries in children. Pediatr Surg Int 2018;34(6):641–5. [49] Englum BR, Gulack BC, Rice HE, Scarborough JE, Adibe OO. Management of blunt pancreatic trauma in children: review of the national trauma data bank. J. Pediatr. Surg 2016;51(9):1526–31 Epub 2016/09/01.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx [50] Akhrass R, Yaffe MB, Brandt CP, Reigle M, Fallon Jr WF, Malangoni MA. Pancreatic trauma: a ten-year multi-institutional experience. Am Surg 1997;63(7):598–604. [51] Koh EY, van Poll D, Goslings JC, Busch OR, Rauws EA, Oomen MW, et al. Operative versus nonoperative management of blunt pancreatic trauma in children: a systematic review. Pancreas 2017;46(9):1091–7 Epub 2017/09/14. [52] Thomson DA, Krige JE, Thomson SR, Bornman PC. The role of endoscopic retrograde pancreatography in pancreatic trauma: a critical appraisal of 48 patients treated at a tertiary institution. J Trauma Acute Care Surg 2014;76(6):1362–6 Epub 2014/05/24.
9
[53] Akshintala VS, Saxena P, Zaheer A, Rana U, Hutfless SM, Lennon AM, et al. A comparative evaluation of outcomes of endoscopic versus percutaneous drainage for symptomatic pancreatic pseudocysts. Gastrointest Endosc 2014;79(6):921–8 quiz 83 e2, 83 e5. Epub 2013/12/10. [54] Pavic´ T, Kralj D, Hrabar D. Lumen-apposing metal stents for pancreatic fluid collection drainage: a review of current outcomes. BioMedicine and Surgery 2017;1(2):88–93. [55] Nadler EP, Gardner M, Schall LC, Lynch JM, Ford HR. Management of blunt pancreatic injury in children. Journal of Trauma 1999;47(6):1098–103.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
[m5G;February 20, 2020;19:33]
Injury xxx (xxxx) xxx
Contents lists available at ScienceDirect
Injury journal homepage: www.elsevier.com/locate/injury
Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis Mario Kopljar a,b, Stjepan Ivandic´ b, Marko Mesic´ c, Bore Bakota d, Tihomil Žiger b,e,∗, h ˇ Goran Kondža a,f, Roman Pavic´ a,b, Miloševic´ Milan g, Miran Coklo a
Medical Faculty Osijek, Josip Juraj Strossmayer University of Osijek, Osijek, Croatia University Hospital Center “Sestre milosrdnice”, Zagreb, Croatia Children’s Hospital, Zagreb, Croatia d Trauma and Orthopedic Department, Landesklinikum Horn, Austria e Faculty of Dental Medicine, Josip Juraj Strossmayer University of Osijek, Osijek, Croatia f University Hospital Center Osijek, Croatia g Polyclinic Ribnjak, Zagreb, Croatia h Institute for Anthropological Research, Zagreb, Croatia b c
a r t i c l e
i n f o
Article history: Accepted 9 February 2020 Available online xxx Keywords: Pancreatic injury Abdominal trauma Children Pancreatic trauma
a b s t r a c t AIM: Blunt abdominal trauma is the major cause of abdominal injury in children. No clear guidelines exist for the initial management of blunt pancreatic trauma in children. The aim of this study was to perform a systematic review and meta-analysis of initially non-operative versus initially operative treatment in children with blunt pancreatic injury. METHODS: Studies including children (<18 years) with blunt pancreatic injuries published in any language after year 1990 were included. Total of 849 studies were identified by searching PubMed, Scopus, CINAHL and Cochrane Database. After review, 42 studies met inclusion criteria and were included in this systematic review. There were 1754 patients, of whom 1095 were initially managed non-operatively (NOM), and 659 operatively (OM). Primary outcome was non-operative management success rate, and secondary outcomes were mortality, complications (including specifically pseudocysts and pancreatic fistulas), percent of patients and days on total parenteral nutrition (TPN), length of hospital stay and readmissions. RESULTS: There was no difference in mortality between NOM and OM groups. The incidence of pseudocysts was significantly higher in NOM group compared to OM (P<0.001), especially for AAST grade III or higher (P<0.0 0 0 01). Overall incidence of pancreatic fistulas was significantly lower for NOM group (p = 0.02) but no difference was observed for AAST grades III or higher (p = 0.49). There was no difference in the length of hospital stay (P = 0.31). Duration of total parenteral nutrition was not different for all AAST grades (P = 0.35), but was significantly shorter for OM group for AAST grades III and higher (p = 0.0 0 01). There was no overall difference in readmissions (P = 0.94). Overall success rate of initial non-operative treatment was 87%. CONCLUSIONS: Most patients with pancreatic trauma can initially be treated non-operatively, while early surgical treatment may benefit patients with lesions of the main pancreatic duct. ERCP offers both highly accurate diagnosis and potential treatment of ductal injuries. © 2020 Elsevier Ltd. All rights reserved.
Introduction Blunt abdominal trauma is the major cause of abdominal injury in children. Pancreatic trauma, although infrequent, can be associated with major mortality and morbidity including hemorrhage, abscesses, fistulae, pancreatic leaks and pancreatitis [1]. ∗
Corresponding author E-mail address: [email protected] (M. Kopljar).
Reported incidence of pancreatic injuries in children admitted for blunt abdominal trauma is less than 10% [2–4]. Pancreas is the fourth most commonly injured solid organ following the spleen, liver and kidneys [5]. Unlike pediatric splenic and hepatic injuries, in which non-operative management (NOM) is widely accepted standard of care, the optimal management strategy of pancreatic injuries remains a subject of debate [6]. Depending on the integrity of the main pancreatic duct pancreatic injuries can be divided into minor and major injuries based on
https://doi.org/10.1016/j.injury.2020.02.035 0020-1383/© 2020 Elsevier Ltd. All rights reserved.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
ARTICLE IN PRESS
JID: JINJ
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
2 Table 1 The AAST Pancreatic Injury Classification. Grade
∗
Type of injury
Description of injury
I I II
Hematoma Laceration Hematoma
II
Laceration
III
Laceration
IV
Laceration
V
Laceration
Minor contusion without duct injury Superficial laceration without duct injury Major contusion without duct injury or tissue loss Major laceration without duct injury or tissue loss Distal transection or parenchymal injury with duct injury Proximal† transection or parenchymal injury involving ampulla Massive disruption of pancreatic head
∗
Advance 1 grade for multiple injuries up to grade III. †Proximal pancreas is to the patients’ right of the superior mesenteric vein.
the American Association for the Surgery of Trauma (AAST) Organ Injury Scale (Table 1). AAST I and II injuries indicate parenchymal injuries without pancreatic duct involvement while AAST III, IV and V include ductal injuries with laceration or transection [7]. Management guidelines in adult blunt pancreatic trauma are fairly clear, recommending non-operative management for patients with grade I/II injuries and pancreatic resection with drainage for grade III or higher injuries. No guidelines exist for blunt pancreatic trauma in children and it is currently unclear whether the same approach should be applied in children [8]. There is a general agreement that grades I and II pancreatic injuries can be safely managed without surgery. For major injuries (AAST ≥ III), both operative and non-operative management strategies have been advocated. Advocates of operative management (OM) argue that surgery is associated with faster recovery, fewer complications and that delay in operative management leads to greater morbidity [9–12]. Proponents of non-operative management argue that it is safe and effective and that it avoids unnecessary surgery-related complications and mortality [13,14]. Operative treatment in all patients with pancreatic trauma is generally indicated if the patient’s clinical status worsens, showing signs of peritoneal irritation or patients become hemodynamically unstable [6]. Studies that have compared the outcomes of NOM and OM are retrospective studies with small sample sizes and a significant risk of selection bias. We therefore decided to perform a systematic review and meta-analysis of pertinent studies to determine benefits and harms of non-operative versus operative treatment in children with blunt pancreatic injury. Methods Study selection An electronic systematic literature search was performed on February 28th 2019, using PubMed, Scopus, CINAHL and the Cochrane Library. All titles and abstracts were independently screened by two authors for eligibility. Furthermore, references of studies considered for inclusion were scanned for and additional references. The search included studies published after year 1990. (Fig. 1). No language restrictions were applied. The following search strategy was used for PubMed: “(((trauma[tiab]) OR blunt injury [MeSH]) AND (pancreas[Title/Abstract] OR pancreatic[Title/Abstract])) AND children [MeSH Terms].” Search strategy was modified accordingly for other respective databases.
Fig. 1. Flow diagram (PRISMA) illustrating publication selection.
ies reporting on at least 5 children (age 18 or less) with nonoperative or operative management of blunt pancreatic trauma. Excluded were studies published prior to 1990 and those including adult patients, unless pediatric patients were described separately. Furthermore, publications were excluded if they were case reports, reviews or studies including penetrating trauma.
Eligibility criteria
Types of interventions
After title and abstract screening, the same two reviewers independently reviewed full text articles. Inclusion criteria were stud-
Operative management (OM) was defined as immediate surgery upon establishing the diagnosis of pancreatic injury. Nonoperative
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
management (NOM) was defined as observation, medical treatment, endoscopic or interventional radiology procedures (such as ERCP), as well as surgical procedures for injuries of other organ systems (e.g. head trauma).
Table 2 Characteristics of included studies.
Abbo, 2013 [13] Arkovitz, 1997 [42] Bass, 1991 [43] Beger, 2019 [17] Ben Hassine, 2010 [18] Beres, 2013 [19] Borkon, 2011 [44] Canty, 2001 [20] Cigdem, 2011 [21] Cuenca, 2012 [22] Dai, 2015 [47] de Blaauw, 2008 [14] Fabbro, 2001 [45] Firstenberg, 1999 [23] Garg, 2017 [24] Ghasoup, 2013 [25] Graham, 2000 [46] Holland, 1999 [26] Iqbal, 2014 [9] Jobst, 1999 [5] Juric, 2009 [27] Katz, 2018 [48] Keil, 2016 [28] Keller, 1997 [2] Klin, 2011 [29] Kouchi, 1999 [30] Loungnarath, 2001 [31] Mattix, 2007 [3] Meier, 2001 [11] Mora, 2016 [32] Nadler, 1999 [55] Paul, 2011 [33] Ruszinko, 2005 [34] Shaked, 2003 [35] Shilyansky, 1998 [36] Snajdauf, 2007 [10] Stringer, 2005 [12] Takishima, 1996 [37] Vane, 2009 [38] Wales, 2001 [39] Wood, 2010 [40] Yamana, 2012 [41] TOTAL
Quality assessment We assessed risk of bias at the study level using Cochrane’s ‘Risk of bias’ tool. Two review authors independently assessed the methodologic quality of each included study and resolved their disagreements by discussion. ‘Risk of bias’ criteria included: adequate sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, as well as other potential biases. Each criterion was rated as having ‘low’ risk of bias, ‘high’ risk of bias, or ‘unclear’ risk of bias, which indicated the lack of information or uncertainty regarding the potential for bias. Additionally, individual studies were graded for level of evidence using criteria from the Oxford Center for Evidence-Based Medicine. Quality assessment was performed using the NewcastleOttawa Scale, and studies with a score lower than 6 were considered low quality. Outcomes Primary outcomes were mortality and complications. Secondary outcomes were the success rate of nonoperative management, percent of patients and days on total parenteral nutrition (TPN), length of hospital stay and readmissions. Outcomes were analyzed separately for minor (AAST grades I and II) and major (AAST grades III or higher) (Table 1) whenever feasible. Statistical analysis Data obtained from eligible studies was collected into an electronic spread-sheet, and statistical analysis was performed using RevMan 5.3 software. Weighted treatment effects were calculated using random effects models. For dichotomous outcomes, results were expressed as odds ratios (OR) with 95% confidence intervals (CI). Where continuous scales of measurement were used to assess the outcomes of treatment, mean difference (MD) was used [15]. Assessment of heterogeneity Heterogeneity was assessed using the I2 test (16). An I2 statistic value of less than 40% was regarded as not showing any significant heterogeneity. Assessment of publication biases Funnel plots were used to assess reporting biases. Funnel plots were assessed qualitatively. Results Initial search of electronic databases revealed 849 potential studies, as well as 5 additional studies found in the references of these studies (Fig. 1). After duplicate studies were removed and full texts of studies scrutinized according to inclusion and exclusion criteria, there were 42 publications that met inclusion criteria for this systematic review (Table 2). All publications were retrospective, with high selection biases (Fig. 2). All of the studies were of low quality according to Newcastle-Ottawa Scale. In total, there
3
∗ ∗∗
NOP∗
OP∗∗
%NOP success∗∗∗
36 13 32 17 4 24 0 9 25 54 25 31 7 8 14 6 15 14 95 26 7 5 6 112 6 19 13 128 2 194 24 23 4 25 28 0 8 7 5 9 11 4 1095
0 13 8 0 1 15 25 9 6 25 8 3 2 1 1 4 1 0 72 30 0 8 18 42 4 1 8 45 9 202 27 20 3 0 7 13 1 1 9 0 14 3 659
100 – 100 82.4 – 91.3 – 22 96 96.3 100 90 86 87.5 71.4 100 80 79 92 80.8 100 – 100 – 100 84 100 74 0 – 79 87 100 80 100 – 38 100 100 100 – 100 87.0%
Non-operative management. Operative management. Percent of patients successfully managed non-operatively.
∗∗∗
were 1754 pediatric patients, of whom 1095 were managed nonoperatively (NOM), and 659 operatively (OM). Non-operative treatment was successful in 87% of the 1095 patients that were initially treated non-operatively.
Pseudocysts Information about pseudocyst formation was available in 32 studies (3, 5, 9–14, 17–40). Ten studies were available for metaanalysis. There were 858 children treated with NOM and 498 were treated with OM. Pseudocyst development was reported in 214 patients treated with NOM (24.9%) and 37 patients treated with OM (7.4%). Meta-analysis demonstrated a statistically significant higher risk of pseudocyst formation in the NOM group (OR 3.05, 95% CI 1.84 to 5.05; P<0.001). Heterogeneity was low (I2 = 14%, P = 0.28) and can be noted in the funnel plot (Fig. 3). When only patients with grade III or higher were analyzed, meta-analysis also demonstrated a statistically significant higher risk of pseudocyst formation in the NOM group (OR 12.46, 95% CI 4.60 to 33.77; P<0.0 0 0 01). Heterogeneity was low (I2 = 10%, P = 0.35).
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ 4
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
Fig. 2. Risk of bias graph: review authors’ judgments about each risk of bias item presented as percentages across all included studies.
Fig. 3. Funnel plot of comparison: non-operative vs. operative, outcome: pseudocysts.
Fisulas Information about pancreatic fistulas was available in seven studies (6, 23, 32–34, 38, 41). There were 289 children treated with NOM and 157 were treated with OM. Pancreatic fistula developed in 3 patients treated with NOM (1.0%) and 12 patients treated with OM (7.6%). Meta-analysis demonstrated a statistically significant lower risk of pancreatic fistula in the NOM group (OR 0.27, 95% CI 0.09 to 0.82; P = 0.02). Heterogeneity was low (I2 = 0%, P = 0.87) and can be noted in the funnel plot (Fig. 4). However, when only patients with grade III or higher were analyzed, meta-analysis did not demonstrate a difference in the risk of fistula (OR 0.56, 95% CI 0.11 to 2.91; P = 0.49). Heterogeneity was low (I2 = 0%, P = 0.77). Mortality Mortality due to pancreatic trauma was reported in 32 studies (1–4, 6–9, 11–15, 17, 18, 22, 27–32, 34, 36, 38, 40–46). Mortal-
ity from other causes (mostly severe head trauma or other causes not related to pancreatic injury) was not considered in this study. There were 741 children treated with NOM and 430 were treated with OM. There were 14 deaths reported in patients treated with NOM (1.9%) and 12 patients treated with OM (2.8%). Meta-analysis did not demonstrated a statistically significant difference in the risk of mortality between groups (OR 0.96, 95% CI 0.42 to 2.22; P = 0.93). Heterogeneity was low (I2 = 0%, P = 0.86) and can be noted in the funnel plot (Fig. 5). Length of hospital stay Information about the length of hospital stay was available in 18 studies (1, 2, 4, 6, 9, 12, 15, 22, 23, 27, 32–34, 36, 38, 40, 47, 48). There were 667 children treated with NOM and 403 were treated with OM. Meta-analysis demonstrated slightly shorter length of stay in the NOM group, although not significantly (MD −1.26, 95% CI −3.68 to 1.15; P = 0.31). Heterogeneity was high (I2 = 70%, P = 0.0 0 05) and can be noted in the funnel plot (Fig. 6).
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
5
Fig. 4. Funnel plot of comparison: non-operative vs. operative, outcome: fistulas.
Fig. 5. Funnel plot of comparison: non-operative vs. operative, outcome: mortality.
Seven studies reported hospital stay for patients with grade III or higher pancreatic injury (1, 6, 15, 23, 34, 40, 47), and only two studies provided sufficient data for meta-analysis (6, 23). Although meta-analysis is not reliable for only two studies, hospital stay was significantly shorter in OM group (MD 7.57, 95% CI 0.53 to 14.61; P = 0.04). Heterogeneity was moderate (I2 = 54%, P = 0.14). Total parenteral nutrition Information about the duration of total parenteral nutrition (TPN) was available in 9 studies (1, 6, 7, 22, 23, 33, 34, 38, 47), but only 4 were available for meta-analysis. There were 331 children treated with NOM and 186 were treated with OM. Meta-analysis demonstrated slightly shorter duration of TPN in the OM group, al-
though not significantly (MD 4.86, 95% CI −5.34 to 15.05; P = 0.35). Heterogeneity was high (I2 = 90%, P = 0.0 0 0 01) and can be noted in the funnel plot (Fig. 7). Five studies reported hospital stay for patients with grade III or higher pancreatic injury (6, 7, 23, 34, 47), and only three studies provided sufficient data for meta-analysis (6, 23, 34). Although meta-analysis is not reliable for only three studies, duration of TPN was significantly shorter in OM group (MD 10.98, 95% CI 5.48 to 16.48; P = 0.0 0 01). Heterogeneity was low (I2 = 15%, P = 0.31). Information about percentage of patients requiring total parenteral nutrition (TPN) was available in 5 studies (1, 12, 22, 33, 38). There were 232 children treated with NOM and 68 were treated with OM. Total parenteral nutrition was required in 84 patients treated with NOM (36.2%) and 20 patients treated with OM (29.4%). Meta-analysis did not demonstrate any difference in the
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ 6
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
Fig. 6. Funnel plot of comparison: non-operative vs. operative, outcome: length of hospital stay.
Fig. 7. Funnel plot of comparison: non-operative vs. operative, outcome: total parenteral nutrition (days).
risk of TPN (OR 1.17, 95% CI 0.11 to 12.17; P = 0.90). Heterogeneity was high (I2 = 83%, P = 0.002). Readmissions Information about readmissions was available in 3 studies (7, 23, 48) and two studies were available for meta-analysis. There were 106 children treated with NOM and 111 were treated with OM. Readmissions occurred in 19 patients treated with NOM (17.9%) and 26 patients treated with OM (23.4%). Meta-analysis did not demonstrate any difference in the risk of readmissions (OR 1.08, 95% CI 0.18 to 6.47; P = 0.94). Heterogeneity was high (I2 = 68%, P = 0.08) and can be noted in the funnel plot (Fig. 8).
Discussion Pancreatic injury represents 0.3% of all pediatric traumas and 0.6% of significant abdominal trauma [49]. Pancreatic trauma remains a morbid condition, with a mortality rate of 5% and major complications affecting approximately 1 in 4 children [50]. Pancreatic trauma is the fourth most common solid organ injury, right after the spleen, liver and kidneys, and it is very frequently associated with other significant injuries, such as traumatic brain injury, bone and joint trauma, urogenital injuries etc. [5,23]. The majority of pancreatic injuries arise from motor vehicle accidents, but bike injuries (handlebar injuries) and struck injuries are also frequent causes [49].
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
7
Fig. 8. Funnel plot of comparison: non-operative vs. operative, outcome: readmissions.
The diagnosis of pancreatic injury is difficult due to the fact that it is a retroperitoneal organ and symptoms of pancreatic injury maybe non-specific. Therefore, some authors estimate the incidence of pancreatic trauma to be as high as 9% of all injured children [3,40]. Furthermore, pancreatic trauma is often associated with other abdominal injuries, such as duodenal injuries, that have pronounced clinical picture that may mask pancreatic injury. Increased amylase and lipase levels may be helpful, and are often elevated in pancreatic trauma, although they may be normal in the first few hours after the trauma, and it may take several serial investigations to demonstrate and increase in serum amylase or lipase levels [3,16,43]. Increase serum amylase levels are also not specific to pancreatic trauma and may be elevated in cases of head and neck trauma that is often associated with pancreatic trauma). Serum lipase measurements may be used in these cases to exclude extrapancreatic trauma [5]. The use of ultrasound has been demonstrated to be inaccurate in the detection of pancreatic injury [3]. It can, however, be used for the detection of pancreatic pseudocysts. Computed tomography (CT scan) remains most utilized for the diagnosis of pancreatic trauma in both children and adults with accuracy rates as high as 91%. Direct or indirect signs of pancreatic injury such as complete gland transection, greater than 50% laceration through the gland, or the presence of early peripancreatic fluid, may be helpful in determining the presence of duct injury. Endoscopic retrograde cholangiopancreatography (ERCP) is routinely used to assess pancreatic duct pathology in adults and is an accurate method of investigating the pancreatic duct injury but also has drawbacks, especially in pediatric population, due to its invasive nature and associated risks of postprocedural pancreatitis [51]. Magnetic resonance cholangiopancreatography can also be used as a non-invasive alternative to ERCP. On the other hand, ERCP can be utilized not only for the diagnosis of pancreatic injury, but also in therapeutic setting, as it enables stenting of most pancreatic duct injuries [52]. The optimal management of pancreatic trauma continues to be the source of controversy among pediatric surgeons. Previously published Cochrane systematic review could not identify any randomized controlled studies comparing operative and non-operative
management of pancreatic trauma in children, and due to the nature of this condition it is unlikely that any such randomized trial could ever be performed [21]. Today, the standard of care for injuries to solid organs in hemodynamically stable patients following blunt abdominal trauma is non-operative management. Most minor pancreatic injuries go unrecognized and often heal spontaneously. It is therefore a common approach to treat minor pancreatic injuries (grades I and II) non-operatively. However, major pancreatic trauma (AAST grades III and higher) often requires operative approach and, if surgery is delayed, operative intervention may be rather complicated [10]. The results of this review confirms that non-operative management is justified in patients with grades I or II pancreatic trauma, as there was no significant difference in mortality or the formation of pseudocysts, although pseudocysts were more common in patients treated non-operatively. However, in most cases pseudocysts also be treated non-operatively by the means of endoscopic or percutaneous drainage [52-54] and the success rate of non-operative treatment was found to be as high as 87%. Published studies included in this review did not report significant morbidities related to delayed surgery for pseudocysts. Some studies reported longer hospital stay for non-operatively treated patients with minor pancreatic trauma but the quality of data did not allow meta-analysis [42]. For patients with major pancreatic trauma (grades III or higher according to AAST classification), there was also no difference in mortality regarding initial management strategy (operative versus non-operative). However, in these patients, significantly more pseudocysts were reported in patients who were initially treated non-operatively. Also, patients who were initially treated with surgery had shorter length of hospital stay and shorter duration of total parenteral nutrition. Finally, the frequency of readmission was similar between patients who were initially treated operatively and non-operatively. There is an obvious benefit in early surgical approach for patients with major pancreatic trauma with ductal injury. It is therefore important to tailor the diagnostic approach so that pancreatic duct injury can be accurately diagnosed as soon as possible. While CT scan and ultrasound may clearly indicate ductal lesions in some cases, ERCP may offer more definite diagnosis in over 97% of
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx
8
patients, and if the ductal lesion is found, either early surgery or endoscopic stenting of the injured pancreatic duct can be performed (52). Conclusion Most patients with pancreatic trauma can initially be treated non-operatively, while early surgical treatment may benefit patients with lesions of the main pancreatic duct. This stresses the importance of accurate and timely diagnosis of pancreatic trauma. Initial diagnostic workup should include CT scan, primarily to exclude other serious or life-threatening injuries, and in cases of suspected pancreatic duct injuries, ERCP offers both highly accurate diagnosis and potential treatment of ductal injuries. Funding This paper is part of a supplement supported by The Croatian Trauma Society. Declaration of Competing Interest All authors declare no conflict of interest. References [1] Jacombs ASW, Wines M, Holland AJA, Ross FI, Shun A, Cass DT. Pancreatic trauma in children. J Pediatr Surg 2004;39(1):96–9. [2] Keller MS, Stafford PW, Vane DW. Conservative management of pancreatic trauma in children. J of Trauma 1997;42(6):1097–100. [3] Mattix KD, Tataria M, Holmes J, Kristoffersen K, Brown R, Groner J, et al. Pediatric pancreatic trauma: predictors of nonoperative management failure and associated outcomes. J Pediatr Surg 2007;42(2):340–4 Epub 2007/02/03. [4] Garvey EM, Haakinson DJ, McOmber M, Notrica DM. Role of ERCP in pediatric blunt abdominal trauma: a case series at a level one pediatric trauma center. J Pediatr Surg 2015;50(2):335–8 Epub 2015/02/02. [5] Jobst MA, Canty TG Sr, Lynch FP. Management of pancreatic injury in pediatric blunt abdominal trauma. J Pediatr Surg 1999;34(5):818–23 discussion 23-4. Epub 1999/06/08. [6] Haugaard MV, Wettergren A, Hillingso JG, Gluud C, Penninga L. Non-operative versus operative treatment for blunt pancreatic trauma in children. Cochrane Database Syst Rev 2014(2) CD009746. Epub 2014/02/14. [7] Moore EE, Cogbill TH, Malangoni MA, Jurkovich GJ, Champion HR, Gennarelli TA, et al. Organ injury scaling, II: pancreas, duodenum, small bowel, colon, and rectum. J Trauma 1990;30(11):1427–9 Epub 1990/11/01. [8] Ho VP, Patel NJ, Bokhari F, Madbak FG, Hambley JE, Yon JR, et al. Management of adult pancreatic injuries: a practice management guideline from the eastern association for the surgery of trauma. J Trauma Acute Care Surg 2017;82(1):185–99 Epub 2016/10/28. [9] Iqbal CW, St Peter SD, Tsao K, Cullinane DC, Gourlay DM, Ponsky TA, et al. Operative vs nonoperative management for blunt pancreatic transection in children: multi-institutional outcomes. J Am Coll Surg 2014;218(2):157–62 Epub 2014/01/21. [10] Snajdauf J, Rygl M, Kalousova J, Kucera A, Petru O, Pycha K, et al. Surgical management of major pancreatic injury in children. Eur j of pediatr surg 2007;17(5):317–21 Epub 2007/10/31. [11] Meier DE, Coln CD, Hicks BA, Guzzetta PC. Early operation in children with pancreas transection. J Pediatr Surg 2001;36(2):341–4. [12] Stringer MD. Pancreatic trauma in children. Br J Surg 2005;92(4):467–70 Epub 2005/01/27. [13] Abbo O, Lemandat A, Reina N, Bouali O, Ballouhey Q, Carfagna L, et al. Conservative management of blunt pancreatic trauma in children: a single center experience. Eur J of Pediatr Surg 2013;23(6):470–3. [14] de Blaauw I, Winkelhorst JT, Rieu PN, van der Staak FH, Wijnen MH, Severijnen RSVM, et al. Pancreatic injury in children: good outcome of nonoperative treatment. J Pediatr Surg 2008;43(9):1640–3. [15] Thompson SG, Higgins JP. How should meta-regression analyses be undertaken and interpreted? Stat Med 20 02;21(11):1559–73 Epub 20 02/07/12. [16] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. Bmj 2003;327(7414):557–60 Epub 2003/09/06. [17] Beger B, Kizilyidiz BS, Simsek M, Akdeniz H, Sonmez B. Pancreatic injury due to blunt abdominal trauma in children. Med Sci (Turkey) 2019;8(1):225–8. [18] Ben Hassine L, Boughanmi N, Douira W, Mormeche J, Louati H, Hamzaoui M, et al. [Management of blunt duodenal and pancreatic injuries in children (about a series of 8 cases)]. Tunis Med 2010;88(8):586–92 Epub 2010/08/17. Prise en charge des traumatismes fermes du bloc duodeno- pancreatique chez l’enfant (a propos d’une serie de 8 cas).
[19] Beres AL, Wales PW, Christison-Lagay ER, McClure ME, Fallat ME, Brindle ME. Non-operative management of high-grade pancreatic trauma: is it worth the wait? J Pediatr Surg 2013;48(5):1060–4 Epub 2013/05/25. [20] Canty TG Sr, Weinman D, Tepas JJ, Mullins RJ, Winthrop A, Powell RW. Management of major pancreatic duct injuries in children. J Trauma Inj Infect Crit Care 20 01;50(6):10 01–7. [21] Cigdem MK, Senturk S, Onen A, Siga M, Akay H, Otcu S. Nonoperative management of pancreatic injuries in pediatric patients. Surg Today 2011;41(5):655–9 Epub 2011/05/03. [22] Cuenca AG. Islam S. pediatric pancreatic trauma: trending toward nonoperative management? Am Surg 2012;78(11):1204–10. [23] Firstenberg MS, Volsko TA, Sivit C, Stallion A, Dudgeon DL, Grisoni ER. Selective management of pediatric pancreatic injuries. J Pediatr Surg 1999;34(7):1142–7 Epub 1999/08/12. [24] Garg RK, Mahajan JK. Blunt trauma pancreas in children: is non-operative management appropriate for all grades? Pediatr Gastroenterol Hepatol Nutr 2017;20(4):252–8. [25] Ghasoup A, Sadieh O, Mansor A, Alani Q, Sawaqed J. Management of pancreatic injuries with blunt abdominal trauma in children. Case Ser. J Trauma Treat 2013:2. [26] Holland AJ, Davey RB, Sparnon AL, Chapman M, LeQuesne GW. Traumatic pancreatitis: long-term review of initial non-operative management in children. J Paediatr Child Health 1999;35(1):78–81 Epub 1999/05/11. [27] Juric´ I, Pogorelic´ Z, Biocˇ ic´ M, Todoric´ D, Furlan D, Šušnjar T. Management of blunt pancreatic trauma in children. Surg. Today 2009;39(2):115–19. [28] Keil R, Drabek J, Lochmannova J, Stovicek J, Rygl M, Snajdauf J, et al. What is the role of endoscopic retrograde cholangiopancreatography in assessing traumatic rupture of the pancreatic in children? Scand J Gastroenterol 2016;51(2):218–24 Epub 2015/07/23. [29] Klin B, Abu-Kishk I, Jeroukhimov I, Efrati Y, Kozer E, Broide E, et al. Blunt pancreatic trauma in children. Surg Today 2011;41(7):946–54 Epub 2011/07/13. [30] Kouchi K, Tanabe M, Yoshida H, Iwai J, Matsunaga T, Ohtsuka Y, et al. Nonoperative management of blunt pancreatic injury in childhood. J Pediatr Surg 1999;34(11):1736–9 Epub 1999/12/11. [31] Loungnarath R, Blanchard H, Saint-Vil D. Blunt pancreatic injuries in children. Ann Chir 2001;126(10):992–5. [32] Mora MC, Wong KE, Friderici J, Bittner K, Moriarty KP, Patterson LA, et al. Operative vs nonoperative management of pediatric blunt pancreatic trauma: evaluation of the national trauma data bank. J Am Coll Surg 2016;222(6):977–82. [33] Paul MD, Mooney DP. The management of pancreatic injuries in children: operate or observe. J Pediatr Surg 2011;46(6):1140–3 Epub 2011/06/21. [34] Ruszinko V, Willner P, Olah A. Pancreatic injury from blunt abdominal trauma in childhood. Acta Chir Belg 2005;105(3):283–6 Epub 2005/07/16. [35] Shaked G, Kleiner O, Finally R, Mordechai J, Newman N, Cohen Z. Management of blunt pancreatic injuries in children. Eur J Trauma 2003;29(3):151–5. [36] Shilyansky J, Sena LM, Kreller M, Chait P, Babyn PS, Filler RM, et al. Nonoperative management of pancreatic injuries in children. J Pediatr Surg 1998;33(2):343–9 Epub 1998/03/14. [37] Takishima T, Sugimoto K, Asari Y, Kikuno T, Hirata M, Kakita A, et al. Characteristics of pancreatic injury in children: a comparison with such injury in adults. J Pediatr Surg 1996;31(7):896–900 Epub 1996/07/01. [38] Vane DW, Kiankhooy A, Sartorelli KH, Vane JL. Initial resection of potentially viable tissue is not optimal treatment for grades ii-iv pancreatic injuries. World J Surg 2009;33(2):221–7 Epub 2008/04/12. [39] Wales PW, Shuckett B, Kim PCW. Long-term outcome after nonoperative management of complete traumatic pancreatic transection in children. J Pediatr Surg 2001;36(5):823–7. [40] Wood JH, Partrick DA, Bruny JL, Sauaia A, Moulton SL. Operative vs nonoperative management of blunt pancreatic trauma in children. J Pediatr Surg 2010;45(2):401–6 Epub 2010/02/16. [41] Yamana I, Kawamoto S, Kamitani T, Ishikura H, Yamashita Y. Pancreatic injury in children: review of 7 cases and the pertinent literature. Hepatogastroenterology 2012;59(114):574–7 Epub 2011/09/24. [42] Arkovitz MS, Johnson N, Garcia VF. Pancreatic trauma in children: mechanisms of injury. J of Trauma 1997;42(1):49–53. [43] Bass DH, Lakhoo K. Pancreatic injuries in children. S Afr j of sur Suid-Afrik tydskr vir chirur 1991;29(2):39–40 Epub 1991/06/01. [44] Borkon MJ, Morrow SE, Koehler EA, Shyr Y, Hilmes MA, Miller RS, et al. Operative intervention for complete pancreatic transection in children sustaining blunt abdominal trauma: revisiting an organ salvage technique. Am Surg 2011;77(5):612–20. [45] Fabbro MA, D’Agostino S, Romanini MV, Chiarenza F, Fasoli L, Musi L, et al. Management of severe blunt pancreatic trauma in children. Personal exp La Pediatria medi e chirur 2001;23(3–4):179–82 Epub 2001/11/29. [46] Graham CA, O’Toole SJ, Watson AJ, Munro FD, Haddock G. Pancreatic trauma in scottish children. J R Coll Surg Edinb 20 0 0;45(4):223–6 Epub 20 0 0/12/29. [47] Dai LN, Chen CD, Lin XK, Wang YB, Xia LG, Liu P, et al. Abdominal injuries involving bicycle handlebars in 219 children: results of 8-year follow-up. Eur J of Trauma & Emerg Surg 2015;41(5):551–5. [48] Katz MG, Fenton SJ, Russell KW, Scaife ER, Short SS. Surgical outcomes of pancreaticoduodenal injuries in children. Pediatr Surg Int 2018;34(6):641–5. [49] Englum BR, Gulack BC, Rice HE, Scarborough JE, Adibe OO. Management of blunt pancreatic trauma in children: review of the national trauma data bank. J. Pediatr. Surg 2016;51(9):1526–31 Epub 2016/09/01.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035
JID: JINJ
ARTICLE IN PRESS
[m5G;February 20, 2020;19:33]
M. Kopljar, S. Ivandi´c and M. Mesi´c et al. / Injury xxx (xxxx) xxx [50] Akhrass R, Yaffe MB, Brandt CP, Reigle M, Fallon Jr WF, Malangoni MA. Pancreatic trauma: a ten-year multi-institutional experience. Am Surg 1997;63(7):598–604. [51] Koh EY, van Poll D, Goslings JC, Busch OR, Rauws EA, Oomen MW, et al. Operative versus nonoperative management of blunt pancreatic trauma in children: a systematic review. Pancreas 2017;46(9):1091–7 Epub 2017/09/14. [52] Thomson DA, Krige JE, Thomson SR, Bornman PC. The role of endoscopic retrograde pancreatography in pancreatic trauma: a critical appraisal of 48 patients treated at a tertiary institution. J Trauma Acute Care Surg 2014;76(6):1362–6 Epub 2014/05/24.
9
[53] Akshintala VS, Saxena P, Zaheer A, Rana U, Hutfless SM, Lennon AM, et al. A comparative evaluation of outcomes of endoscopic versus percutaneous drainage for symptomatic pancreatic pseudocysts. Gastrointest Endosc 2014;79(6):921–8 quiz 83 e2, 83 e5. Epub 2013/12/10. [54] Pavic´ T, Kralj D, Hrabar D. Lumen-apposing metal stents for pancreatic fluid collection drainage: a review of current outcomes. BioMedicine and Surgery 2017;1(2):88–93. [55] Nadler EP, Gardner M, Schall LC, Lynch JM, Ford HR. Management of blunt pancreatic injury in children. Journal of Trauma 1999;47(6):1098–103.
Please cite this article as: M. Kopljar, S. Ivandic´ and M. Mesic´ et al., Operative versus non-operative management of blunt pancreatic trauma in children: Systematic review and meta-analysis, Injury, https://doi.org/10.1016/j.injury.2020.02.035