Analysis of Later Stage Morbidity and Mortality after Pancreatic Surgery Because of Abdominal Trauma

Abstract

Background:

Pancreatic trauma surgery is a complicated surgical procedure for severe pancreatic injuries, accompanied by a high incidence of complications and mortality. This study was designed to explore the long-term prognosis of pancreatic surgery because of abdominal trauma.

Patients and Methods:

The clinical data of 103 patients who were admitted to Jinling Hospital between August 2012 and August 2019 who had pancreatic trauma surgery were analyzed retrospectively.

Results:

All admissions involved pancreatic trauma surgery performed at an outside hospital network, which later transferred patients to our institution because of post-operative later-stage complications. Eight patients received American Association for the Surgery of Trauma (AAST) grade 1 or 2 pancreatic injuries and 95 received AAST grade 3, 4, or 5 pancreatic injuries. The primary surgical management of pancreatic injuries included drainage of the pancreatic injury (n = 28), repair of the pancreas (n = 35), partial pancreatectomy (n = 15), pancreaticojejunostomy (n = 6), and pancreaticoduodenectomy (n = 19). The most common mechanism of trauma was motor vehicle collision (n = 72), crush injury (n = 26), and stab wound (n = 5). Of 103 patients suffered varying degrees of gastrointestinal fistulae and intra-abdominal infections, there were 66 cases of pancreatic fistulae (64.1%), 49 cases of enteric fistulae (47.6%), 26 cases of colonic fistulae (25.2%), 14 cases of gastric or gastrointestinal anastomotic fistulae (13.6%), and 13 cases of biliary fistulae (12.6%). Ninety-five patients survived and eight patients died after therapy; the mean length of intensive care unit stay was 33 days. The number of patients who underwent emergency pancreaticoduodenectomy (EPD), the incidence of blood transfusion, the number of fistulae per patient, and the duration of mechanical ventilation and bacteremia in the mortality group were substantially higher than in the survival group (p < 0.05 each). The patients who underwent EPD had more grade 5 pancreatic injuries, more blood transfusions, higher peak total bilirubin, greater numbers of fistulae and open abdomen, and longer duration of mechanical ventilation and mortality than other patients (p < 0.05 each).

Conclusions:

The grade of pancreatic injury was associated with mortality and post-operative complications. The post-operative mortality and occurrence of complications of EPD because of abdominal trauma were significant; use of damage control surgery could potentially reduce the morbidity and mortality related to this procedure.

Traumatic injury to the pancreatic region is a complex problem for surgeons usually accompanied by injuries of adjacent organs and vessels [1]. Management options range from conservative treatment to closed drainage, operative repair, parenchyma-sparing procedures, distal pancreatectomy, and pancreaticoduodenectomy (PD) [2]. Each management strategy is characterized by high mortality and incidence of complications including necrotic acute pancreatitis, abscesses, bleeding, fistulae, and pancreatic pseudo-cysts. The overall prognosis depends on location of injury, time to intervention, amount of blood loss and duration of shock, and the experience of the operating surgeon [3].

Emergency pancreaticoduodenectomy (EPD) is an extremely uncommon procedure for severe pancreaticoduodenal injuries and is only performed in select hemodynamically stable patients because of high accompanying morbidity and mortality. In cases of massive damage or devitalization of the duodenum, head of pancreas, or ampulla because of trauma, EPD might be unavoidable [2]. On the other hand, the surgical procedure of EPD is complicated and requires special surgical skills and expertise. Furthermore, the unique and complex post-operative care involved with EPD creates a substantial burden for the hospital managing these patients. With the development of damage control surgery (DCS) [4,5], delayed reconstruction is a viable option, and the necessity and optimal time to perform EPD is debatable [6,7].

Few large series focused on the management of post-operative later stage complications after injuries to the pancreas have been reported. The purpose of this study was to explore the later stage morbidity and mortality after pancreatic surgery because of abdominal trauma, and the prognosis and feasibility of EPD. The role of damage control surgery is also assessed.

Patients and Methods

This study was a retrospective review of patients treated at the Intestinal Fistula Center of Jinling Hospital between August 2012 and August 2019. Enrolled patients initially underwent pancreatic trauma surgery at outside hospitals and were then transferred to our institution for management of post-operative complications. Clinical data for each patient were collected, including demographic information, mechanism of injury, location of injury, types of surgery, hospital length of stay before referral, location of bleeding, blood transfusion, types of digestive fistula, presence of jaundice, ventilator support duration, initial time of enteral nutrition (EN), results of bacterial culture, length of intensive care unit (ICU) stay, hospital length of stay, mortality, and so on. Patients were followed until the end of hospitalization and were divided into a survival group and an in-hospital death group according to final treatment outcome; data were compared between the two groups. The clinical data were also compared between patients who underwent EPD and others.

Evaluation of injury severity

All patients in this study had pancreatic injuries of varying severity, therefore, the pancreatic injury grading system proposed by the Organ Injury Scaling Committee of the American Association for the Surgery of Trauma (AAST) was chosen to represent the severity of abdominal trauma (Table 1). Briefly, grade 1 was defined as mild contusion or superficial laceration without duct injury; grade 2 included major contusion or laceration without duct injury or tissue loss; grade 3 included distal transection or parenchymal injury with duct injury; grade 4 included proximal (right of the superior mesenteric vein) transection or parenchymal injury involving ampulla; and grade 5 was massive disruption of the pancreatic head. The grade of pancreatic injury was evaluated by our team based on the records from out of network surgeons during initial surgeries.

Table 1.

American Association for the Surgery of Trauma Classification of Pancreatic Trauma

Grade	Injury	Description of the pancreatic injury
1	Hematoma	Mild contusion without duct injury
	Laceration	Superficial laceration without duct injury
2	Hematoma	Major contusion without duct injury
	Laceration	Major laceration without duct injury or tissue loss
3	Laceration	Distal transection or parenchymal injury with duct injury
4	Laceration	Proximal transection or parenchymal injury involving ampulla
5	Laceration	Massive disruption of the pancreatic head

Management of common post-operative complications

Patients were transferred to our institution because of various post-operative complications including intra-abdominal infections (IAIs), digestive fistulae, intra-abdominal abscesses, bleeding, malnutrition, open abdomen, jaundice, renal failure, pneumonia, and acute respiratory distress syndrome (ARDS).

Most of the IAIs were caused by different types of digestive fistulae. Digestive fistulae consisted of pancreatic, biliary, and gastrointestinal fistulae. A double-catheterization cannula was placed around digestive fistulae to flush and drain digestive juice, instead of using original passive drainage with a latex tube during outside hospital stays. The double-catheterization cannula consists of a drip tube and a suction tube; physiologic saline is fed in from the drip tube and then drawn out by the suction tube with a negative pressure of 0.02 kpa after flushing around the fistula (Supplemental Figure S1. In our clinical practice, use of the double-catheterization cannula and octreotide injection has been proven to be particularly helpful to control digestive fistulae and IAIs.

Percutaneous drainage (PCD) was used to manage intra-abdominal abscesses. Laparotomy drainage was performed when other techniques to control IAIs failed. In addition, the double-catheterization cannula was used as much as possible for abscess drainage during the PCD process. Definitive surgery of resecting or repairing the fistulae was carried out to restore the continuity of the digestive tract several months later if the fistulae could not heal spontaneously.

Hemorrhage usually resulted from digestive tract or abdominal vessels in proximity to digestive fistulae. Satisfactory local drainage was able to prevent the corrosion of digestive juices released from fistulae and contributed to control of bleeding. Hemostatic and supplemental blood coagulation factors also helped to stop local bleeding. Angioembolization under digital subtraction angiography (DSA) was useful to deal with bleeding from large vessels. Laparotomy was considered when intra-abdominal bleeding was unable to be controlled by non-operative methods. Gauze compression packing hemostasis was the final treatment for massive hemorrhage of deep situation when stitching cannot be performed.

The presence of jaundice was mainly the result of liver dysfunction caused by IAIs, i.e., bacteriotoxins flowing into the liver via the portal vein. When IAIs were controlled, hepatic function generally recovered over time. In addition, the early initiation of enteral nutrition (EN) can promote enterohepatic circulation and decrease the level of bilirubin. Pneumonia and ARDS were managed with standard mechanical ventilation, and renal failure was managed by infection control and hemofiltration.

An open abdomen is a serious complication that usually is the result of incision dehiscence and intra-abdominal hypertension. The wound surface of an open abdomen should be well protected to prevent discretionary broadening of the wound and the development of enteroatmospheric fistulae. Intra-abdominal pressure was controlled by limiting intravenous fluid, increasing colloid osmotic pressure, obtaining infection source control, and reducing intestinal secretions. Subsequently, the abdomen could be closed by skin-only suture or skin grafting in a staged approach.

Data analysis

The data were analyzed using SPSS Statistics, version 19 (IBM Corp, Armonk, NY). Categorical variables were expressed by absolute numbers with percentages, and continuous data were presented as mean with standard deviation. Categorical variables were compared by χ² test or Fisher exact test when appropriate. Continuous variables were analyzed by t-test; p < 0.05 was considered statistically significant.

Results

Involved abdominal organs after trauma

A total of 103 patients who underwent pancreatic surgery because of trauma were included in this study; 95 (92.2%) patients were men, with a mean age of 36.8 years. Seventy-two patients had had motor vehicle collisions, 26 patients suffered crush injuries from heavy weights, and five patients had stab wounds.

All patients suffered different degrees of pancreatic injuries. According to the diagnostic criteria of pancreatic injury proposed by the AAST (Table 1), three patients sustained grade 1 injuries, and there were five grade 2 injuries, 60 grade 3 injuries, 21 grade 4 injuries, and 14 grade 5 injuries. The three grade 1 injuries and five grade 2 injuries were all in the group of survivors (n = 95), with 37 grade 3 injuries, 19 grade 4 injuries, and 11 grade 5 injuries. In the death subgroup (n = 8), there were three grade 3 injuries, two grade 4 injuries, and three grade 5 injuries.

Duodenal injuries were found in 43 (41.7%) patients. In the survival group, 23 patients suffered AAST grade 3–5 duodenal injuries and 15 patients had AAST grade 1 or 2 duodenal injuries. However, in the death group, four patients suffered AAST grade 3–5 duodenal injuries and one patient sustained AAST grade 1 or grade 2 duodenal injuries. In addition, there were 33 (32.0%) cases of AAST grade 1–5 hepatic injuries, 26 (25.2%) patients with minor vascular, 26 (25.2%) patients with colon injuries, 19 (18.4%) small bowel injuries, 17 (16.5%) patients with AAST grade 1–4 splenic injuries, 15 (14.6%) lacerations of major vascular, 10 (9.7%) cases of extrahepatic biliary injuries, and four (3.9%) cases of AAST grade 1–3 renal injuries. An average of 3.1 organs and large vessels were injured per patient (Table 2).

Table 2.

Associated Intra-Abdominal Injuries (n = 103)

Intra-abdominal organs	Mechanism			Total
Intra-abdominal organs	Motor vehicle collision (n = 72)	Crush injury (n = 26)	Stab wound (n = 5)	Total
Duodenum	27	13	3	41.7%
Liver	24	8	1	32.0%
Minor vascular	16	6	4	25.2%
Colon	15	9	2	25.2%
Small bowel	10	7	2	18.4%
Spleen	12	5	0	16.5%
Major vascular	10	4	1	14.6%
Stomach	9	1	1	10.7%
Extrahepatic biliary	7	2	1	9.7%
kidney	1	3	0	3.9%

Only seven of the 103 patients had an isolated pancreatic injury (none died), with 24 cases of pancreas plus one associated organ (two died); 42 cases of pancreas plus two associated organs (three died); 17 cases of pancreas plus three associated organs (two died); 13 cases of pancreas plus four or more associated organs (one died; Table 3). Mortality did not increase as the number of associated injuries increased, however, this correlation is not significant because of the small number of deceased patients.

Table 3.

Number of Associated Organ Injuries and Mortality

Injury	Survival group (n = 95)	Death group (n = 8)	Total	Mortality
Pancreas only	7	0	7	0
Pancreas +1 associated organ	22	2	24	8.3%
Pancreas +2 associated organs	39	3	42	7.1%
Pancreas +3 associated organs	15	2	17	11.8%
Pancreas +4 or more associated organs	12	1	13	7.7%

Timing and operative methods

Sixty-four patients from the survival group underwent pancreatic surgery within eight hours of injury; 15 patients within 24 hours; and in 16 patients the time exceeded 24 hours, ranging from two days to 33 days. In the death group, five patients underwent surgery within eight hours of injury; for the other three patients the time ranged from two to three days. The percentage of patients who underwent surgery within eight hours of injury was not significantly different between the survival and death group (p > 0.05).

The primary surgery to manage pancreatic injuries included drainage of the pancreatic injury (n = 28), repair of pancreas (n = 35), partial pancreatectomy (n = 15), pancreaticojejunostomy (n = 6), and PD (n = 19). In the survival group, 61 patients underwent repair of pancreas or peripancreatic drainage, 15 patients underwent partial pancreatectomy, 13 patients underwent EPD, and six patients underwent pancreaticojejunostomy. However, two patients in the group underwent peripancreatic drainage and the other six underwent EPD. The percentage of patients who underwent EPD in the death group was significantly higher than that of survival group (p < 0.01).

Emergency pancreaticoduodenectomy was performed in 16 patients within eight hours of injury, one patient within 24 hours, and two patients within 48 hours, both of whom died. A PD (Whipple procedure) alone was performed in nine patients, a total PD alone in one patient, and nine patients underwent a Whipple procedure plus repair of hepatic injury, bowel resection and anastomosis, repair of intestinal perforation, or right nephrectomy. All patients underwent a classic PD (resection of the pylorus); no patient had pylorus preserving pancreaticoduodenectomy (PPPD).

Six patients from the survival group were performed with damage control procedure. None of the death group received damage control surgery.

Common post-operative complications

A total of 99 patients developed digestive fistulae and subsequent IAIs after surgery. Of 103 patients there were 66 cases of pancreatic fistulae (64.1%), 49 cases of enteric fistulae (47.6%), 26 cases of colonic fistulae (25.2%), 14 cases of gastric or gastrointestinal anastomotic fistulae (13.6%), and 13 cases of biliary fistulae (12.6%). The incidence of pancreatic fistulae was 62.1% (59/95); enteric fistulae, 49.5% (47/95); and colonic fistulae, 23.2% (22/95) in the survival group, which was not significantly different from those of the death group (p > 0.05, respectively). However, the incidence of gastric or gastrointestinal anastomotic fistulae (50.0%) and biliary fistulae (62.5%) in death group was significantly higher than those of survival group (p < 0.05, respectively). Higher grade of pancreatic injuries was associated with more complications and greater mortality. The comparison of common complications according to different grades of pancreatic injury is listed in Table 4.

Table 4.

Grade of Pancreatic Injury and Post-Operative Complications (n = 103)

Grade of injury	n (%)	Complications (n, %)	Mortality
Grade 1	3 (2.9%)	Jaundice (3; 100%)	0
Grade 1	3 (2.9%)	DF numbers per patient (0.7)	0
Grade 2	5 (4.9%)	Jaundice (4; 80.0%)	0
		Respiratory failure (4,8; 0.0%)
		Bleeding (2; 40.0%)
		Open abdomen (1; 20.0%)
		DF numbers per patient (1.4)
Grade 3	60 (58.3%)	Jaundice (45; 75.0%)	5%
		Respiratory failure (26; 43.3%)
		Open abdomen (18; 30.0%)
		Bleeding (17; 28.3%)
		Renal failure (1; 1.7%)
		DF numbers per patient (1.6)
Grade 4	21 (20.4%)	Jaundice (16; 76.2%)	9.5%
		Respiratory failure (11; 52.4%)
		Bleeding (9; 42.9%)
		Open abdomen (5; 23.8%)
		Renal failure (3; 14.3%)
		DF numbers per patient (1.7)
Grade 5	14 (13.6%)	Jaundice (12; 85.8%)	21.4%
		Respiratory failure (10; 71.4%)
		Open abdomen (9; 64.3%)
		Bleeding (7; 50%)
		Renal failure (2; 14.3%)
		DF numbers per patient (2.6)

The incidence of post-operative complications categorized by severity of pancreatic injury.

DF = digestive fistula.

Substantial bleeding occurred in 31 of 95 patients in the survival group (32.6%), consisting of four cases of gastrointestinal hemorrhage and 27 cases of intra-peritoneal vascular hemorrhage. Gastrointestinal hemorrhage in six of these 27 cases all gradually stopped with the use of hemostatic drugs. Among the patients with intra-peritoneal vascular hemorrhage, four patients received angioembolization under digital subtraction angiography (DSA), 15 patients required laparotomy to manage vascular hemorrhage, and two patients required both angioembolization and laparotomy. Four cases of hemorrhage in the death group were not effectively controlled, which led to mortality. The packed red blood cell and plasma transfusion volume in the death group were significantly higher than that of survival group (15,916 mL vs 1,915 mL; p < 0.05). The cause of death in the other four cases in the death group was multiple organ failure caused by unsatisfactory control of IAIs.

The patients in the survival group all received EN uneventfully and the mean time to goal rate was 14 days after referral. However, only four patients in the death group achieved goal EN by a mean of 14.5 days; moreover, the other four patients in the death group were never able to tolerate EN.

Twenty-eight of 95 patients in the survival group (29.5%) were ultimately managed with an open abdomen. The wound was protected by skin grafting in the later stages and abdominal wall reconstruction was ultimately performed. Furthermore, an open abdomen was required in five of eight patients in the death group (62.5%) and the abdomen remained open until death.

The mean duration of mechanical ventilation in the survival group was 9.4 days, much shorter than that of the death group (29.8 days; p < 0.05). The mean peak level of serum total bilirubin during hospital stay in the survival group was 75.4 mcmol/L, much lower than that of the death group (176.8 mcmol/L; p < 0.05). The mean length of ICU stay in the survival and the death groups were 30.6 days and 59 days, respectively, with a significant difference (p < 0.05). Nineteen patients in the survival group were found to have positive blood cultures, the most common cultures in turn were fungus (4/19), Klebsiella pneumoniae (3/19), and Escherichia coli (2/19). Three patients in the death group had bacteremia, including three strains of Klebsiella pneumoniae, two strains of Escherichia coli, one strain of Pseudomonas aeruginosa, Enterobacter cloacae, and Enterococcus faecium. The comparison of common complications and other factors between the survival group and the death group is provided in Table 5.

Table 5.

Demographic and Clinical Characteristics

	Survival group	Death group	p
	(n = 95)	(n = 8)	p
Gender (male), n (%)	87 (91.6)	8 (100.0)	0.867
Age (y)	36.6 ± 11.8	39.6 ± 10.6	0.481
Smoking, n (%)	30 (31.6)	3 (37.5)	0.503
Drinking, n (%)	31 (32.6)	4 (50)	0.265
Hypertension, n (%)	2 (2.1)	1 (12.5)	0.217
Diabetes, n (%)	4 (4.2)	2 (25)	0.068
Hospital stay before referral (d)	40.7 ± 34.9	38.1 ± 55.9	0.903
Grade 5 pancreatic injury, n (%)	11 (11.6)	3 (37.5)	0.075
Patients underwent surgery within 8 h, n (%)	64 (67.4)	5 (62.5)	0.528
Patients underwent EPD, n (%)	13 (13.7)	6 (75)	<0.001
Blood transfusion (mL)	1,915 ± 2,809	15,916 ± 10,671	0.007
Pancreatic fistula, n (%)	59 (62.1)	7 (87.5)	0.253
Number of injured organs per patient (n)	3.1 ± 1.2	3.3 ± 1.0	0.640
Peak total bilirubin (mcmol/L)	75.4 ± 103.9	1,76.8 ± 131.7	0.011
Numbers of fistulae per patient (n)	1.6 ± 0.8	2.8 ± 1.2	<0.001
Open abdomen, n (%)	28 (29.5)	5 (67.5)	0.107
Duration of mechanical ventilation (d)	9.4 ± 11.6	29.8 ± 20.1	<0.001
Renal failure, n (%)	3 (3.2)	3 (37.5)	0.006
EN start after admission (d)	14.0 ± 12.1	30.8 ± 22.0	0.001
Length of ICU stay (d)	30.6 ± 18.4	59.0 ± 56.3	0.001
Number of operations, n	1.6 ± 1.1	2.3 ± 1.3	0.141
Bacteremia, n (%)	19 (20.0)	5 (62.5)	0.016
Post-operative hemorrhage, n (%)	31 (32.6)	6 (75.0)	0.024

EPD = emergency pancreaticoduodenectomy; EN = enteral nutrition; ICU = intensive care unit.

Prognosis and feasibility of EPD

The patients were divided into EPD group and control group (no PD) based on the performance of EPD to explore the prognosis and feasibility of EPD. The grade 5 pancreatic injuries in the EPD group was significantly more than those of control group (p < 0.05). However, the incidence of pancreatic fistula was not different between the two groups (p > 0.05). The mortality in the EPD group was 31.6% (6/11), substantially higher than that of control group (2.4%; 2/85; p < 0.05). The blood transfusion, peak total bilirubin, numbers of fistulae per patient, and incidence of open abdomen in the EPD group was substantially higher than that of the control group (p < 0.05). The duration of mechanical ventilation and time to EN start after admission in the EPD group was substantially longer than that of the control group (p < 0.05). Post-operative hemorrhage in the EPD group was more frequent than that of the control group (p < 0.05). In contrast, the length of ICU stay and incidence of bacteremia was not different between the EPD and control group (Table 6).

Table 6.

Demographic and Clinical Characteristics

	Control group (n = 84)	EPD group (n = 19)	p
Gender (male), n (%)	78 (92.9)	17 (89.5)	0.638
Age (y)	37.2 ± 11.9	36.4 ± 12.5	0.801
Hospital stay before referral (d)	42.9 ± 38.0	29.7 ± 27.6	0.156
Grade 5 pancreatic injury, n (%)	7 (8.3)	7 (36.8)	0.004
Blood transfusion (mL)	1,919 ± 2,895	7,749 ± 9,910	<0.001
Pancreatic fistula, n (%)	55 (65.5)	12 (63.2)	0.523
Peak total bilirubin (mcmol/L)	65.5 ± 98.8	161.9 ± 119.4	<0.001
Numbers of fistulae per patient (n)	1.5 ± 0.7	2.8 ± 0.9	<0.001
Open abdomen, n (%)	22 (26.2)	11 (57.9)	0.01
Duration of mechanical ventilation (d)	9.2 ± 12.7	17.8 ± 15.4	0.012
Renal failure, n (%)	3 (3.6)	3 (15.8)	0.075
EN start after admission (d)	13.4 ± 11.3	23.9 ± 19.7	0.002
Number of operations, n	1.5 ± 0.9	2.5 ± 1.8	<0.001
Length of ICU stay (d)	32.4 ± 26.0	31.1 ± 19.0	0.837
Mortality, n (%)	2 (2.4)	6 (31.6)	<0.001
Bacteremia, n (%)	17 (20.2)	7 (36.8)	0.109
Post-operative hemorrhage and	22 (26.2)	15 (78.9)	<0.001
hemostasis, n (%)
Laparotomy and stitching	16	0
Angioembolization	6	1
Medicine	12	1
Stuffing	1	4

EPD = emergency pancreaticoduodenectomy; EN = enteral nutrition; ICU = intensive care unit.

Discussion

The management of pancreatic injuries is diverse and complex. Most pancreatic injuries are classified as low grade (grade 1 or 2) and can be approached by non-surgical treatment [8]. Options for surgical intervention include simple repair, drainage, and complicated EPD. Injuries to the pancreas are often accompanied by contiguous organs damage. The most frequently affected organs in turn were duodenum, liver, colon, small bowel, spleen, stomach, and extrahepatic biliary in this study. The mortality rates directly related to pancreatic injuries ranged from 2%–17%, and late mortality generally occurred because of sepsis and associated organ failure [2].

The patients included in this study all suffered different levels of IAIs. Digestive fistulae were the most common causes of IAIs in this study, including pancreatic, biliary, and gastrointestinal fistulae. Source control is considered fundamental to the treatment of most patients with IAIs [9]. Unobstructed drainage is the key source control to treat IAIs caused by digestive fistulae. In the past decades, the double catheterization cannula has been proven better than passive drainage to cure IAIs and digestive fistulae in thousands of cases in our institute, most of whom needed somatostatin analogue therapy [10]. Some cases of organ dysfunction and bleeding usually improve spontaneously after control of IAIs. Some bleeding required vascular embolization digital subtraction angiography or laparotomy. Pancreatic fistula was the most common morbidity seen, with an incidence of 64.1% (66/103). Post-operative endoscopic retrograde cholangiopancreatography (ERCP) with stenting of the pancreatic duct was reported [10,11]. This method was applied occasionally with the absence of IAIs in our institute. The rate of pancreatic fistula in the death group was numerically but not statistically significantly higher than that of the survival group.

The highly corrosive nature of pancreatic secretions resulting in necrosis and bleeding of tissue around the fistula is probably the reason for the high associated mortality. However, the biliary fistulae and gastrointestinal anastomotic fistulae in death group were more than those of survival group. The possible reason maybe that death group had more patients underwent EPD with more bilioenteric anastomosis and gastrointestinal anastomosis, and EPD was proven to be a risk factor of mortality (p < 0.05).

All cases of mortality in this study suffered digestive fistulae and subsequent severe IAIs. Most of the patients survived with the help of changes in drainage methods when transferred to our institute. The double-catheterization cannula could control IAIs more effective than passive drainage. Therefore, double catheterization cannula was suggested to be placed alongside anastomosis during initial surgery to control IAIs timely when anastomosis leak happened. The mean blood transfusion in the death group was 15,916 mL, significantly higher than the 1,915 mL in the survival group. A higher grade of pancreatic injuries usually indicates more severe abdominal trauma and greater mortality [11]. In this study, the incidence of grade 5 pancreatic injuries in the death group was 37.5% (3/8), higher than the survival group (11.6%; 11/95). Although this difference was not significant, the finding is probably caused by small sample size. In addition, the peak total bilirubin, numbers of fistulae per patient, duration of mechanical ventilation, and post-operative hemorrhage in the death group were higher than those of the survival group. The mean level of peak total bilirubin in the death group was higher than that of the survival group. In our experience, an elevated bilirubin usually represents persistent IAI leading to organ dysfunction and failure; thus, persistently elevated bilirubin was a predictor of poor prognosis. These complications were more common in the patients who underwent EPD. These results illustrate the association of high morbidity after EPD because of abdominal trauma that can subsequently result in higher mortality.

There were eight deaths, for an overall mortality of 7.8%, slightly lower than previous reports [11 –13]. partly because most of the patients enrolled in this study were late post-traumatic cases. The early mortality caused by uncontrolled hemorrhage or severe sepsis was not included. It has been reported that the overall morbidity and mortality of pancreatic trauma is dependent on the degree of injury to adjacent organs. No patients with only isolated pancreatic injury resulted in death, which is similar to another report [13]. However, no increase of mortality was observed as the number of associated injuries increased in this study. There was no difference in mortality between patients with injuries of the pancreas only or two associated injuries and patients with three or more associated injuries. The interpretation of these data is confounded by selection bias, whereby the less associated injuries were also followed by severe IAIs when transferred to our institute. Our institute is the national-level center for IAIs and digestive fistulae treatment, admitting patients with various post-operative complications. Moreover, patients with more associated injures may die before being transferred. Therefore, there was no significant difference in the mortality of patients with minor injuries and those with severe injuries during a later stage.

The mortality rate after EPD because of abdominal trauma is 20%–50% [7,14], and the mortality in our study reached 31.6%. Therefore, the indication for EPD should continue to be strictly limited. With regard to the pancreaticoduodenal injuries involving the pancreatic duct, especially grade 5 injuries, EPD is considered only when there is little hope for conservative treatment or the use of less invasive surgical procedures [3]. Furthermore, EPD must be done by experienced hepato-bilio-pancreatic (HBP) surgeons [15].

Only seven patients in this study underwent EPD for grade 5 pancreatic injuries, indicating that the patients for were not specially selected. We believed that the high incidence of complications in this study reflects failure to strictly follow indications for EPD. High morbidity and mortality was observed in patients who underwent EPD when it was not absolutely necessary. Some of these complex operations were performed in local small-scale hospitals rather than abdominal trauma centers, raising the possibility that the procedure was performed by surgeons who may not have been extensively trained in liver, biliary, and pancreatic surgeries.

Severe abdominal trauma frequently should be managed by DCS principles including abbreviated surgery aimed at controlling ongoing hemorrhage and containing gastrointestinal contamination, followed by definitive reconstruction after physiological stabilization in the ICU, leading to decreased mobility and mortality [4]. Therefore, DCS has been advocated to manage pancreaticoduodenal injuries. Thompson et al. [16] reported 15 cases of EPD because of abdominal trauma, 80% of which were given staged EPD using a DCS approach, resulting in a low mortality of 13%. Gupta et al. [7] suggested placement of T-tube drainage at the common bile duct and external drainage of the main pancreatic duct, reducing contamination and the corrosive effects of bile and pancreatic secretions. Late EPD after inflammation could decrease fatal intra-operative or post-operative complications. The overall mortality of our patient cohort was high at 35.3%, although it should be noted that all of our patients were referred to a specialty center due to complications of the original procedure. Of note, only two patients had undergone DCS management and survived. Although we have limited guidance on the treatment of severe pancreaticoduodenal injury with DCS, a strategic incorporation of DCS at smaller hospitals followed by transfer and definitive management at a referral centers could result in improved outcomes.

Conclusions

Injuries to the pancreas usually result in various post-operative complications and high mortality. Only grade 5 injury involving pancreatic head fracture demands EPD by experienced HBP surgeons. The incidence of pancreaticoduodenal injury may be gradually increasing in China because of increasing accidental injuries caused by traffic accidents. Abdominal trauma centers should be established in various places to specialize in dealing with such injuries. Damage control surgery principles are important in managing these patients, although their exact role is currently unknown. Subsequent organ function failure and hemorrhage usually results from uncontrolled IAI. Effective source control including placement of double catheterization cannula is the key factor in treatment when digestive fistula and IAI occurs.

Footnotes

Funding Information

General Research Project of Jiangsu Provincial Health Commission, Grant Number H2018058.

Author Disclosure Statement

All authors declare no conflicts of interest.

Supplementary Material

References

Gulla

, Tan

, Pucci

, et al. Emergent pancreaticoduodenectomy: A dual institution experience and review of the literature. J Surg Res, 2014; 186:1–6.

Iacono

, Zicari

, Conci

, et al. Management of pancreatic trauma: A pancreatic surgeon's point of view. Pancreatology, 2016; 16:302–308.

Krige

, Nicol

, Navsaria

. Emergency pancreatoduodenectomy for complex injuries of the pancreas and duodenum. HPB (Oxford), 2014; 16:1043–1049.

Weber

, Bendinelli

, Balogh

. Damage control surgery for abdominal emergencies. Br J Surg, 2014; 101:e109–e118.

Roberts

, Bobrovitz

, Zygun

, et al. Indications for use of damage control surgery in civilian trauma patients: A content analysis and expert appropriateness rating study. Ann Surg, 2016; 263:1018–1027.

Vasquez

, Coimbra

, Hoyt

, et al. Management of penetrating pancreatic trauma: An 11-year experience of a level-1 trauma center. Injury, 2001; 32:753–759.

Gupta

, Wig

, Garg

. Trauma pancreaticoduodenectomy for complex pancreaticoduodenal injury. Delayed reconstruction. JOP, 2008; 9:618–623.

, Patel

, Bokhari

, 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; 82:185–199.

Mazuski

J E

, Tessier J

, May A

, et al. The Surgical Infection Society Revised Guidelines on the Management of Intra-Abdominal Infection. Surg Infect, 2017; 18:1–76.

10.

Krige

, Kotze

, Setshedi

, et al. Surgical management and outcomes of combined pancreaticoduodenal injuries: Analysis of 75 consecutive cases. J Am Coll Surg, 2016; 222:737–749.

11.

Kao

, Bulger

, Parks

, et al. Predictors of morbidity after traumatic pancreatic injury. J Trauma, 2003; 55:898–905.

12.

Tan

K K

, Chan

, Vijayan

, et al. Management of pancreatic injuries after blunt abdominal trauma. Experience at a single institution. JOP, 2009; 10:657–663.

13.

Krige

, Kotze

, Hameed

, et al. Pancreatic injuries after blunt abdominal trauma: An analysis of 110 patients treated at a level 1 trauma centre. S Afr J Surg 2011;49:58, 60, 62–64 passim.

14.

Lin

, Chen

, Fang

, et al. Management of blunt major pancreatic injury. J Trauma, 2004; 56:774–778.

15.

Lissidini

, Prete

, Piccinni

, et al. Emergency pancreaticoduodenectomy: When is it needed? A dual non-trauma centre experience and literature review. Int J Surg, 2015; 21(Suppl 1):S83–S88.

16.

Thompson

, Shalhub

, DeBoard

, et al. Revisiting the pancreaticoduodenectomy for trauma: A single institution's experience. J Trauma Acute Care Surg, 2013; 75:225–228.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.29 MB

0.00 MB