Pancreatic injuries constitute 0.2%–0.3% of all injuries [1]. Traumatic pancreatic injury (TPI) is rare, particularly as an isolated injury. Only 3.1% to 12% of patients with other abdominal injuries have additional TPI [2-4]. The prevalence of TPI according to each grade of the American Association for the Surgery of Trauma (AAST) scale has been reported to be 1.9% for grade II, 0.6% for grade III, 0.3% for grade IV, and 0.2% for grade V injuries [2,5]. The mechanism of injury mainly involves sudden and direct compression of the pancreas against the lumbar vertebra, especially in patients with thin retroperitoneal fat like children and thin adults [6]. Isolated injuries are difficult to diagnose due to the lack of clinical signs in the first hours after admission, leading to an increase in late phase mortality [4]. Measurement of amylase and lipase levels along with computer tomography usually could help establish its diagnosis [4,7,8]. The mortality of TPI ranges from 3% to 34%. Only 4% to 5% of deaths are directly related to pancreatic injury itself. More than 70% of pancreatic trauma deaths occur within the first 24 hours because of associated injuries [2,6,9,10].

The key predictor of morbidity and treatment decisions in conservative versus operative management of TPI is the integrity of the main pancreatic duct [6,9,11]. The AAST classification helps in treatment decision by grading the injury according to injury location and duct status [5]. Most TPI cases are grade I or II with conservative treatment including hemostasis and drainage [3,7]. Historically, surgical treatment is chosen for high-grade TPI, especially in patients with other abdominal injuries. Recently, there has been a clear paradigm and shift to conservative treatment in patients with complete duct dissection, leading to improved outcomes [2,12-15]. However, it has been suggested that nonstable patients should also be treated surgically regardless of nonsurgical treatment options [2,3,6,9]. An adjunct to conservative treatment is endoscopic retrograde cholangiography (ERCP) with stent insertion in selected cases to reduce endoscopic management failure [5,16]. Some studies have reported similar outcomes of conservative and operative treatments of TPI in pediatric patients [14,17].

The aim of this study was to review our experience over the past 20 years in the management of TPI in pediatric and adult groups and to assess patient outcomes according to age group and treatment strategy for each TPI grade.

This study was approved by the Ethics Committee of Hadassah Medical Center. The Institutional Review Board approved this study and waived the need for informed consent of patients (no. 0314-19-HMO). This was a retrospective analysis of patient records used to diagnose and treat TPI at a level I trauma center of our university teaching hospital between 2000 and 2019. Data retrieval used electronically medical records and international classification of disease coding. We double checked that no cases were missed via manual searches. Cases without a documented pancreatic injury were excluded. The following data were reviewed: demographics, mechanism and area of pancreatic injury, injury severity scale (ISS), Glasgow Coma Scale (GCS), type and time-to-surgery, length of hospital stay (LOHS), length of intensive care unit stay (ICU stay), imaging performed (including ERCP), injury to other organs, diastase level (maximal during hospitalization), procedures performed, and complications. Patients were divided into adults and pediatric groups in order to investigate if there were any differences in the feasibility of conservative management. Additionally, all patients were divided into non-operative management (NOM) and operative treated groups. Analysis of those who had NOM failure was done. Failed NOM was defined as surgically treating patients after a trial of conservative management due to either hemodynamic instability, clinical or laboratory deterioration; peritonitis, fever, or increases of inflammatory markers. We also noted if feeding jejunostomy or total parenteral nutrition was used. The grade of pancreatic injury was defined according to grading of AAST scale ranging from minor (grade I) to devastating (grade V).

A total of 77 patients were identified and included in this study. There were 52 (67.54%) in the adult group (≥ 18 years) and 25 (32.46%) in the pediatric group (< 18 years). The mean age ± SD of all patients was 24.89 ± 15.88 years. There were 54 (70.1%) male. A total of 56 (72.7%) patients had blunt injury. Motor-vehicle accident (MVA) was the predominant injury mechanism (54.5%, n = 42). Of the total cohort, 38 (49.4%), 24 (31.2%), and 15 (19.5%) had AAST grade I/II, III, and IV/V injuries, respectively. Diastase levels were elevated two-fold in 26 (41.3%) patients and three-fold in 24 (38.1%) patients. The most common injured area of the pancreas was the pancreatic tail (31.2%, n = 24). Most (n = 44, 57.1%) patients had an ISS of more than 25.

Table 1 shows baseline characteristic of patients. Most (n = 70, 90.9%) patients had concomitant injury to other organs. Of these organs, the most common injury was to the liver (n = 42, 54.5%) and the spleen (n = 40, 51.9%). Table 2 shows details of concomitant injuries. Thirty (39.0%) patients had conservative management. Forty-seven (61.0%) patients were operated directly. Fourteen out of 30 patients (18.2%) with conservative management underwent an operation during hospitalization. The mean time from admission to operation was 13.8 ± 31.09 hours. The mean LOHS and ICU stay were 25.4 ± 25.43 and 10.72 ± 16.62 days, respectively. Eight (10.4%) patients died. All of them had an operative management.

Table 3 shows patient outcomes for all patients as well as for age subgroups (adults and pediatric groups). Eight patients died (all were adults), including four who had penetrating trauma and four who had a blunt mechanism of injury. Of these eight patients, five had AAST grade IV injury, two had an AAST grade III injury, and one had a grade I injury. One patient with AAST grade I TPI died from associated kidney and spleen injury. Of three patients who died due to gunshot wounds (GSW) and AAST grade III/IV TPI, one died due to associated injury to the superior mesenteric artery and bowel ischemia. Four other patients died due to injury to other vessels and associated organs.

As shown in Table 4, 59 (76.6%) patients underwent laparotomy. Of them, 12 underwent the operation mainly because of other injuries or reasons. However, during the surgery, an exploration of the pancreas and lesser sac was performed with drainage mainly. No resection was performed. The remaining 47 patients had an operative intervention on the pancreas.

TPI is an elusive diagnosis that can be easily overlooked. Nevertheless, it is a diagnosis that harbors a bad prognosis for the patient. Therefore, immediate and appropriate management is the key [6]. The current study investigated outcomes of various management options and compared conservative and operative management outcomes.

In accordance with other studies, this study showed that the leading mechanism of pancreatic injuries was blunt injury, with MVA being the most common. The pancreas is crushed between the steering wheel and the first and second vertebra [6,18].

Our study detailed conservative versus operative treatment. Our data showed that conservative management was relatively successful, especially in grade I and II pancreatic injuries. However, operative management was consequently needed in 14 of 30 patients. Such cases were mainly defined as grades III and IV. In addition, no deaths were reported in the conservative group. However, previous studies showed higher rates of mortality following conservative treatment [19]. This might be due to a relatively small study population analyzed in the current study. These findings might be misguiding since the vast majority of pancreatic injuries are concomitant with other organ injuries. Only 9.1% of cases have isolated pancreatic injuries, making it especially challenging to isolate the effect of pancreatic injuries on patients’ outcomes per se. However, our study showed that conservative treatment was especially successful in low grade ASST and in pediatric patients. In addition, pediatric patients were less likely to have laparotomy than adult patients. This might be related to the fact that adult patients were more likely to be involved in penetrating trauma such as gunshot and stab wound.

Operative treatment included distal pancreatectomy, the Whipple procedure, and hemostasis and drainage. Distal pancreatectomy was the most frequent operation. In general, operative management showed higher mortality rates in our study along with a number of previous studies that similarly found higher mortality rates, especially in patients who had higher grades of injuries [19]. This might be explained by the fact that higher ASST score usually implies more concomitant organ injuries, higher ISS, and subsequently worse outcomes regardless whether the injury is directly related to the pancreas or not. This shows the importance of a multidisciplinary team in making decisions.

As addressed in recent reports, there is a huge controversy on whether to manage pancreatic trauma surgically or conservatively [12-14]. Thus, prompt and reasonable decision-making is still at the top of the list because no rigorous guidelines have been published on this issue.

Early clinical imaging of pancreatic injury may be subtle. Thus, the history and mechanism of injury should be strictly obtained. The current study found that amylase levels were normal in 20.6% of cases and minimally elevated (two-fold) in 41.3% of cases. Thus, we could conclude that amylase was not sensitive enough to detect pancreatic injury or its severity. This is because amylase levels are not usually sufficiently elevated within the first six hours of injury if at all. This finding is consistent with previous studies [8].

Several diagnostic modalities have been used to assess the severity of pancreatic injuries. In the present study, CT was used when pancreatic injuries were suspected along with other abdominal injuries. However, the accuracy of this modality depends on findings themselves. Peripancreatic fluid is highly sensitive but not specific for pancreatic injury. Pancreatic lacerations and hemorrhage in CT are less sensitive but more specific finding for TPI [20].

ERCP was also used in special cases when other diagnostic modalities had failed to assess the pancreatic duct status. Thus, it has a therapeutic role where a stent can be inserted. In our study, 15.6% of patients underwent an ERCP [21].

Finally, our study found higher failure rates of NOM in those with a female sex, those aged less than 18 years, those with AAST scores III and IV, and those with a high ISS. However, the pediatric subgroup showed significantly a higher success rate following conservative treatment than the adult group. This finding is consistent with a study by Mora et al. [14]. The higher rate of failure of NOM in female patients found in this study should be very carefully interpreted because of the small number of female patients used in this study.