Symposium on Childhood Trauma
Pancreatic and Gastrointestinal Trauma in Children Jay L. Grosfeld, M.D.,* and Donald R. Cooney, M.D.**
Blunt abdominal trauma in children presents a significant challenge to the responsible physician in regard to recognition, diagnosis, and treatment. While much attention has been given to the more common intra-abdominal injuries (liver, spleen, kidney), pancreatic and gastrointestinal trauma in children have received relatively little attention. Injuries to the pancreas and gastrointestinal tract continue to be a significant cause of morbidity and mortality in children and typify the diagnostic difficulties encountered in cases of blunt abdominal trauma. This article describes experiences with these injuries in 56 children.
PANCREATIC TRAUMA Pancreatic trauma was observed in 16 patients at the James Whitcomb Riley Hospital for Children, Indianapolis, Indiana. Twelve patients were boys (75 per cent) and four were girls. The average age was 9.5 years with a range of 10 months to 17 years. The cause of pancreatic trauma was related to automobile accidents (as passenger or pedestrian) in five patients, bicycle handle bar injury in four, falls against objects in three, "go-cart" accident in one, physical trauma by siblings or friends in two, and unknown in one (child abuse suspected). The injuries were often insidious in nature and considered trivial in a number of instances. Two children were sent home following emergency room evaluation at the time of initial injury. Three patients had appendectomy performed for suspected appendicitis at other institutions. All three had normal appendices and had associated excess peritoneal fluid at operation. All except one patient in this group were iniFrom the Section of Pediatric Surgery, Department of Surgery, lndiana University School of Medicine, and the James Whitcomb Riley Hospital for Children, Indianapolis, Indiana 'Professor and Director, Section of Pediatric Surgery, Indiana University School of Medicine; Surgeon-in-Chief, James Whitcomb Riley Hospital for Children, Indianapolis, Indiana **Chief Resident, Department of Surgery, Indiana University School of Medicine, Indianapolis, Indiana
Pediatric Clinics of North America- Vol. 22, No.2, May 1975
365
366
JAY
L. GROSFELD AND DONALD R. COONEY
tially evaluated elsewhere and subsequently transferred to the Riley Children's Hospital. Types of injuries included 11 patients with pancreatic pseudocysts, four with pancreatic contusion, laceration, or disruption, and one with complete transection. Associated injuries were relatively infrequent (three patients) and usually involved combined pancreatic and duodenal injuries. Clinical Patterns Persistent abdominal pain, bilious vomiting, fever, and leukocytosis were observed in the majority of patients. An elevated serum amylase was invariably present. A persistently elevated amylase level for prolonged periods of time was noted in all 11 patients with pseudocysts. The mean amylase level in patients with pseudocysts was 515 Somoygi units (upper limit of normal = 180). An epigastric mass was palpable in eight of 11 patients with pseudocysts. False negative paracentesis was frequently noted due to theretroperitoneallocation of the injury. Plain abdominal x-ray examination showed evidence of free fluid and a "sentinalloop" in the upper midabdomen in two of four patients with pancreatic contusion and laceration. One patient with pancreatic pseudocysts presented with bilateral pleural effusions on chest x-ray. Upper gastrointestinal contrast studies were suggestive of pseudocysts in 10 of 11 patients so studied (Fig. 1). Treatment and Results Patients with pseudocysts were managed by cyst-gastrostomy in five patients (Fig. 2), cyst-Roux-en-Y jejunostomy in five cases, and excision in one. Interval from time of injury to operation was greater than four weeks in six cases and more than four months in five. Two patients had operation done 11 months and 22 months respectively following trauma. One patient developed recurrent symptoms of pancreatitis follOwing cystojejunostomy and required sphincterotomy and drainage four years later. Gastrostomy was used as an adjunct in three patients and in addition two of these patients were managed postoperatively by total parenteral hyperalimentation for 10 to 14 days. Serum amylase returned to normal and symptoms resolved in each patient. A nine year old girl with pancreatic transection underwent distal pancreatectomy, splenectomy, and pancreatojejunostomy. One child underwent evacuation of a duodenal hematoma, gastrostomy, and drainage of the pancreas via the lesser sac. Two patients with lacerations of the pancreas were managed by pancreatic suture, drainage of the lesser sac, cholecystostomy, gastrostomy, antibiotics, and total parenteral hyperalimentation. A 22 month old boy with a combined pancreatoduodenal injury (duodenal laceration), craniocerebral injury, and extremity fractures died 12 hours following closure of the duodenal laceration and drainage as a result of the cerebral injury. This was the only death in this group of patients (6.2 per cent mortality). Complications included pancreatic abscess, wound infection, and
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
367
Figure 1. Upper gastrointestinal contrast study demonstrates a retrogastric mass compressing the posterior gastric wall. The findings are consistent with the diagnosis of pancreatic pseudocyst.
gram-negative sepsis in one patient each. One patient developed gastrointestinal bleeding following cystogastrostomy which subsided. Evidence of recurrent pancreatitis occurred in two cases following cystojejunostomy for pseudocyst in one and after distal pancreatectomy for pancreatic transection in the other. Each was managed successfully by sphincterotomy and drainage.
GASTROINTESTINAL TRAUMA This article also concerns experience with 40 infants and children with gastrointestinal injuries following blunt abdominal trauma. Twenty-six of the patients were treated at the Children's Hospital,
368
JAY
L. GROSFELD AND DONALD R. COONEY
Figure 2. Transgastric cystgastrostomy anastomosis between posterior gastric wall and adherent pseudocyst wall.
Columbus, Ohio,* and 14 at the Riley Children's Hospital in Indianapolis, Indiana. Thirty-three of the patients were boys (82 per cent), and seven were girls. The mean age was 7.3 years with a range of three days to 15 years. Presenting symptoms, cause, type, and location of injury, physical findings, and results of treatment were evaluated. Cause of Injury Miscellaneous falls against objects often during play resulted in approximately one-half of the injuries, and automobile accidents with the victim either as a passenger or pedestrian were responsible for one third. Four children were kicked by a horse (10 per cent), and malicious assault (child abuse) occurred in three. Location and Type of Injury There were 26 perforations (65 per cent), 12 obstructing hematomas (30 per cent), and two mesenteric avulsions (5 per cent) with devitalized bowel. The site of injury was duodenal in 17 children, jejunal in 14, and ileal in seven. In addition, there were two gastric perforations (Fig. 3). ·Previously reported.7
369
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
TRAUMA TO STOMACH AND SMALL BOWEL
Figure 3. trauma.
Location of 40 gastrointestinal injuries in childhood resulting from blunt
Fifteen of the 40 patients (37.5 per cent) had significant associated injuries involving the liver in five, musculoskeletal system in five, pancreas in five, brain in four, diaphragm, lung, spleen, and colon in two each, and the genitourinary tract in one (Table 1). Admission to the hospital was reasonably prompt (2 to 12 hours) in many cases, but a delay greater than four days from the onset of injury was observed in eight children, six of whom had duodenal hematoma.
Symptoms and Physical Findings Abdominal pain and tenderness were the only consistent findings and were observed in all conscious patients. Twenty-five patients (60
Table 1. Associated Injuries Noted in 35 Per Cent of Children with Gastrointestinal Trauma ASSOCIATED INJURIES
15 patients (15 of 40) (37.5 per cent) Liver Musculoskeletal Diaphragm Central Nervous System Pancreas Lung Kidney Spleen Colon
Total
5
5 2
4 5 2 1 2
2
28
370
JAY
L.
GROSFELD AND DONALD
R.
COONEY
per cent) had bilious vomiting. Peritoneal irritation was elicited in 15 patients and abdominal distention was noted in nine. Eight patients presented in shock; all had associated major injuries. Pertinent laboratory data were limited to leukocytosis in 25 patients and anemia in 10. An elevated serum amylase was noted in 17 patients of whom only five actually had associated pancreatic trauma. Intraperitoneal free air was demonstrated on abdominal x-ray in less than 50 per cent of patients with perforations. Nine of 12 patients with duodenal hematomas had upper gastrointestinal contrast studies and demonstrated the obstructive lesion in each case (Fig. 4).
Treatment and Results Thirty-five of the 40 children underwent operation. Twenty of 26 perforations were treated with simple debridement and transverse closure of the defect, whereas four required limited bowel resection with primary anastomosis. Two patients with mesenteric avulsion also required resection and anastomosis. Seven of 12 patients with obstructing hematomas were treated primarily by operative evacuation. Five additional children with duodenal hematoma (documented by contrast swallow) were treated initially by nasogastric decompression and parenteral alimentation. Spontaneous resolution occurred in three patients
Figure 4. Upper gastrointestinal contrast study demonstrating obstruction due to duodenal hematoma.
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
371
within one week obviating the need for operation. The remaining two patients, however, required laparotomy and hematoma evacuation due to persistent evidence of obstruction. There were five deaths (12.5 per cent mortality rate). Two unoperated patients whose records dated back 30 years died within two and three days, respectively, due to unrecognized bowel perforation following injuries sustained by being kicked by a horse. Two other deaths occurred within 24 hours of operation and were associated with multiple trauma. A third death occurred 40 days after initial operation and was attributed to sepsis and peritonitis as a result of ileal perforation related to intestinal adhesions. Nonfatal complications occurred in 35 per cent of patients and included atelectasis, duodenal fistula, prolonged ileus, pancreatic pseudocyst formations, and wound infections. The mean duration of hospitalization in survivors was 18 days.
DISCUSSION Recognition and optimal treatment of pancreatic and gastrointestinal injuries in the pediatric age group following blunt abdominal injury are often hampered by a delay in diagnosis. Unlike the adult where easily identified penetrating injuries are commonplace, 90 per cent of abdominal injuries in children are the result of blunt abdominal trauma.5 Often the causative injury follows seemingly inconsequential trauma to the abdomen, an event which may not be related to the parent. In younger patients, nonproductive child-parent or childphysician communication, coupled with inappropriate physical examination, is a significant cause of delay in diagnosis and treatment. The location of the duodenum and pancreas high in the retroperitoneal space and the mobility of the rest of the intestine on its mesentery often protect these structures from injury. Similarly, symptoms and physical findings related to these injuries are often subtle in nature and compound the problems in establishing an early diagnosis. The mechanism of pancreatic injury resulting from blunt abdominal trauma is directly related to a compressing force which crushes the body of the gland against the vertebral columnY' 21 This may result in contusion, laceration, or disruption of the body of the pancreas and pancreatic duct. If the impact of force is to the right or left of the midline (tangential), the duodenum and head of the gland or the spleen and tail may be injured. Pancreatic injury is often overlooked as physical signs may be deceptively minimal due to the retroperitoneal position of the gland. Abdominal x-rays and peritoneal tap are often not helpfulY Diagnostic difficulties encountered (particularly in children) are typified by the fact that three children reported herein underwent appendectomy elsewhere for abdominal pain. Weitzman and Swenson,25 Laird and Clagett,9 and Stone2° have reported similar instances of pancreatic injury also operated upon with an incorrect preoperative diagnosis of acute appen-
372
JAY
L. GROSFELD AND DONALD R. COONEY
dicitis. Pancreatic injury may be overlooked even at the time of laparotomy done for trauma unless the lesser sac is opened to carefully examine the entire body and tail of the gland, and a Kocher maneuver performed to fully evaluate the head of the gland and the entire duodenum.2 In addition, if a "masking" hematoma over a paravertebral fracture of the body of the gland is not uncovered, a ductal injury may similarly be overlooked. Without recognition and appropriate operation, pancreatic disruption may be lethal or lead to subsequent fistula, pseudocyst, or scarring resulting in persistent pancreatitis.24 Sepsis and hemorrhage (two common sequelae of these injuries) often develop in the presence of necrotic tissue, persistent inflammation, and digestive pancreatic enzymes resulting from ductal disruption. The mortality rate following pancreatic trauma is reported to be 20 per cent, and when combined with duodenal injuries is as high as 45 per cent. 1, 11, 14, 17,21 Injuries to the head of the gland are more lethal than those involving the body or tail.H The mortality rate increases proportionately with the number of associated injuries to other organ systems and if the patient presents in shock.11 ,17 The only death in this group (6.2 per cent mortality) occurred in a two year old with associated duodenal and craniocerebral injuries. Serum amylase determination should be obtained in all patients with upper abdominal trauma. Although an elevated amylase level does not necessarily mean a pancreatic injury has occurred, when associated with persistent abdominal pain and tenderness, this observation suggests that laparotomy be performed. Acute pancreatic injuries were successfully managed by suture repair or distal resection and drainage. A sump catheter within a Penrose drain as described by Ranson et al. and others is a useful method of drainage.2 , 15 A treatment program that places the injured gland "at rest" by diverting the bile and gastric juice (by temporary cholecystectomy and gastrostomy), prohibiting oral intake, and maintaining caloric needs (by total parenteral hyperalimentation) is continued until amylase levels return to normal and symptoms subside. This was a very satisfactory method of therapy in two patients. Owens and Wolfman have found these above diversion adjuncts helpful in instances of pancreatic trauma and in addition performed a jejunostomy for feeding purposes.14 Others have suggested distal pancreatectomy be performed for ductal injuries, however, these procedures often carry a greater morbidity (particularly persistent fistula) and mortality rate than drainage aloneP' 19 The only patient in this study who underwent resection developed a fistula and eventually required reoperation for recurring symptoms of pancreatitis. Although pancreatoduodenectomy has been suggested for extensive injuries to the head of the pancreas and duodenum, this procedure carries with it a high operative mortality.l,17 In the present group of patients, we did not encounter a pancreatic injury in the childhood age group that required this procedure. Pseudocysts are the result of a collection of necrotic tissue, old blood, and secretions that escape from the injured pancreas and loculate
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
373
in the lesser sac. Occasionally, these contents may dissect into the mediastinum or retroperitoneum.9 The inflammatory reaction that follows encapsulates the pseudocyst with a fibrous wall. Over 80 per cent of pseudocysts in the adult are due to complications of alcoholic pancreatitis. In contrast, more than 70 per cent in childhood are a direct result of trauma. lO • 11 Pseudocysts are associated with a symptom complex of pain, fever, and vomiting. An epigastric mass is frequently palpable. Laboratory data often demonstrate an elevated serum amylase, leukocytosis, and occasionally hyperbilirubinemia. Thomford and Jesseph noted the triad of a mass, elevated serum amylase, and leukocytosis in half of their 50 patients with pseudocysts. 22 All the children in this study had persistently elevated amylase levels and eight of 11 had a palpable mass. X-ray examination of the chest and abdomen is often helpful in achieving a diagnosis. Twenty per cent of patients present with a pleural effusion. 24 One patient in this review demonstrated this finding on chest x-ray. The upper gastrointestinal contrast study was the most useful roentgenologic examination and was consistent with a diagnosis of pseudocyst in 10 of 11 cases. Reports of recent experiences with ultrasound echo grams and selective arteriograms suggest these studies may be useful in making the diagnosis of pseudocyst. 4. 24 Ultrasound discerns between a cystic (pseudocyst) or solid (tumor) mass, while arteriography usually demonstrates an avascular mass in instances of pseudocyst. Pseudocysts should all be treated operatively as they may result in duodenal obstruction, biliary obstruction, free rupture (peritonitis), hemorrhage (erosion into major vessels), splenic vein thrombosis, gastric ulceration, and pancreatic ascites.24 The operation of choice is somewhat controversial and is usually determined by the location and duration of the pseudocyst. Others on et al. reported successful use of external drainage of pancreatic pseudocysts by use of a mushroom catheter in five childrenP This type of modified marsupialization drained for 3 to 4 weeks in four patients and five months in another. There were no deaths or recurrences. The objections usually raised when this procedure is employed is a high recurrence rate and prolonged persistence of fistula. Anderson noted excessive losses of fluid and electrolytes and severe skin breakdown in 23 per cent of patients when external drainage was employed.3 These data, however, were obtained from observations in adults who ofte'n had underlying chronic pancreatitis and ductal abnormalities, and may not apply to pancreatic trauma in children with previously normal glands. Warren et al. have shown it takes from four to six weeks from the time of injury for a pseudocyst capsule to thicken enough to hold sutures effectively.23 As a general rule, we reserve external drainage for the rare patient in whom the cyst wall is thin and too friable to hold sutures. In most instances, internal drainage of the pseudocyst is preferred and the type of anastomosis is chosen according to the location of the pseudocyst. If the cyst is adherent to the stomach wall, cystgastrostomy has been quite effective therapy in our hands. Hutson et al. noted a 50
374
JAY
L.
GROSFELD AND DONALD
R.
COONEY
per cent hemorrhage rate following cystgastrostomy and considered this related to the effects of gastric acid entering the cyst cavity.6 They further suggested that the edges of the stomach wall not be sutured so that they are not hindered from acting as a flap valve, thereby reducing the entry of gastric juices into the cavity.6 In contrast, Schumer et al. and Warshaw report much lower rates (3 per cent) of hemorrhage following cystgastrostomy and stress the importance of careful suture approximation to avoid hemorrhage. 16 ,24 We follow the latter suggestion and in addition employ a Stamm gastrostomy over the site of the cystgastrostomy anastomosis, prohibit oral intake, and use total parenteral hyperalimentation for 10 to 14 days postoperatively to prevent gastric juice and food particles from entering the cavity. The pseudocyst cavity is an informal structure without an epithelial lining. Once evacuated, the cavity rapidly shrinks and disappears within a couple of weeks, particularly if the pancreas is placed "at rest" as outlined above. Kilman et al. suggested that cystojejunostomy is the treatment of choice for pseudocyst as it avoids reflux of gastric contents into the cyst.s The Roux-en-Y loop technique was employed in five patients in this series with one recurrence observed. This procedure is advisable in instances in which the pseudocyst is not adherent to the stomach wall. Warshaw reported recurrence rates of 3 per cent with cyst gastrostomy, 5 per cent with cystjejunostomy, and 8 per cent with cystduodenostomy.24 Warren et al. found no difference between the effectiveness of cystgastrostomy or cystjejunostomy in treating experimentally produced pseudocysts.23 Low recurrence rates should be expected in children with pseudocysts resulting from pancreatic trauma due to an absence of underlying pancreatic disease and ductal obstructionP Recurrences in such cases rarely do occur and usually reflect the healing of the ductal injury by excessive scarring.
GASTROINTESTINAL TRAUMA Injuries to the stomach are quite rare due to its mobility and the protection afforded this viscus by the rib cage. Most retrospective reviews concerning blunt abdominal trauma usually fail to mention instances of gastric trauma. Both individuals in the present series sustained injury immediately following a large meal and had a distended stomach at the time of epigastric impact. Both responded to simple closure of perforation. Most gastric perforations occur on the anterior wall, however, careful examination of the posterior wall through the lesser sac must be done to avoid missing an associated posterior wall perforation as was observed in one of our patients. Small bowel injuries occur in from 5 to 15 per cent of children hospitalized for blunt abdominal trauma. 7 Despite improvements in diagnosis and management, these lesions continue to result in a significant morbidity and mortality. Mortality rates of 5 to 20 per cent and complication rates of 35 per cent have been recorded.5 , 7, 12 The complication rate in this group of patients was 35 per cent and the mortality rate was
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
375
12.5 per cent. Death following isolated injuries is unusual, however, the mortality may be high with an increasing number of associated organ injuriesP The small bowel is compressible and mobile and often escapes damage in its midportion. Williams and Sargent presented experimental evidence suggesting that a common mechanism of nonpenetrating intestinal trauma involves a crushing of the bowel against the vertebral column.26 In instances of duodenal hematoma, the hemorrhage originates where the duodenum is compressed by the vertebral column and spreads between the muscular layers of the duodenum about an equal distance in each direction. The second mechanism involves a tearing or shearing force of the bowel and its mesentery at points of intestinal fixation. The small bowel injuries in this review occurred either in the duodenum, in the jejunum in close proximity to the ligament of Treitz, and the distal ileum close to the cecum which is attached to the lateral abdominal wall. In these areas, the intestine is relatively fixed increasing its susceptibility to disruption, contusion, or avulsion. Pathologic fixation such as those related to intestinal adhesions from a previous operation may also increase the susceptibility of the small bowel near the adhesions to shearing forcesP Aside from the factors that tend to inhibit prompt recognition of these injuries that have previously been alluded to, inappropriate physical examination may also result in diagnostic delays. A meaningful examination is often difficult to achieve in an irritable and uncooperative child. The use of a short-acting barbiturate (seconal, 2 mg per pound of body weight, intramuscularly) will usually relieve the child's anxiety without masking true abdominal tenderness. 5• 7 Unwarranted reliance on laboratory results may also promote delays in diagnosis. Neither anemia nor leukocytosis was a specific finding. Serum amylase level was elevated in 17 of the 40 patients, but only five of the 17 patients actually had pancreatic trauma. Although upper gastrointestinal contrast radiographs are frequently diagnostic in patients with obstructing duodenal hematomas, plain roentgenograms demonstrated free-air in less than 50 per cent of children with perforations. The longest diagnostic delays are noted in instances of obstructing hematomas. Stewart et al. reported eight cases at the Boston Children's Hospital with duration of symptoms from 1 to 6 days prior to diagnosis. 18 Pain followed by bilious vomiting is the most frequent finding. Obstructing hematomas have been treated in a wide variety of fashions including bypassing gastroenterostomies, surgical decompression, and prolonged nonoperative nasogastric suction. Our approach to this injury is as follows: When the duodenal hematoma coexists with other intraabdominal injuries warranting early operation, it is evacuated at that time obviating the need for a second operation. An isolated injury resulting in hematoma is managed by nasogastric suction and parenteral alimentation for one week. At the end of this time, repeat contrast roentgenograms will often demonstrate partial resolution of the obstruction (in approximately one half of the cases) and avoid the need for operative intervention.7 If obstruction persists at this time, the hematoma has usually liquified and can easily be relieved by subserosal evacuation.
376
JAY
L.
GROSFELD AND DONALD
R.
COONEY
At operation, the full extent of the injury can best be observed if the right colon is mobilized and elevated to the left and a Kocher maneuver is performed. The hematoma may extend to the proximal jejunum and this area must also be appreciated and the obstructing hematoma appropriately drained. Simple perforations of the jejunum and ileum can be managed by either debridement of the edges and direct closure (80 per cent) or by limited resection with primary anastomosis when the lesion is more extensive. A transverse closure of the perforation is suggested to avoid narrowing of the lumen. Instances of mesenteric avulsion require resection of the bowel supplied by the injured mesenteric vessels and primary end-to-end intestinal anastomosis.
SUMMARY Injuries to the pancreas and gastrointestinal tract following blunt abdominal trauma continue to be a significant cause of morbidity and mortality in the pediatric age group. Optimal treatment of these injuries is frequently hampered by considerable delays in diagnosis. Factors contributing to these delays include the location of much of the duodenum and the pancreas in the retroperitoneum resulting in an absence of initial symptoms and signs, the often trivial nature of some of the responsible blunt traumatic accidents, inappropriate child-parent or childphysician communication, failure to achieve a meaningful physical examination in uncooperative or unconscious patients, and false negative paracentesis. Eighty per cent of these injuries occurred in boys. Eleven of 16 patients with pancreatic trauma had pseudocysts. A persistently elevated serum amylase level was invariably noted and an epigastric mass was palpable in eight patients. Significant delays in diagnosis were prevalent and pseudocyst was misdiagnosed as appendicitis in three cases. Internal drainage by cyst gastrostomy or cystjejunostomy was effective operative treatment. In instances of acute pancreatic injuries, sump drains, gastrostomy, cholecystostomy, and total parenteral hyperalimentation were useful therapeutic adjuncts. There was one death for a 6.2 per cent mortality rate. Forty patients had gastrointestinal injuries involving the duodenum in 17, jejunum in 14, ileum in seven, and stomach in two. Perforations occurred in 65 per cent of cases, obstructing hematomas in 30 per cent, and mesenteric avulsions in 5 per cent. Associated injuries were observed in 15 patients (37.5 per cent). Pain and tenderness were the only consistent findings. Upper gastrointestinal contrast studies were diagnostic of duodenal hematomas. Eighty per cent of perforations were managed by simple closures and 20 per cent by resection and anastomosis. Obstructing hematomas unassociated with other injuries may be expected to resolve without requiring operation in 50 per cent of patients managed conservatively. Complications occurred in 35 per cent of patients and the mortality rate was 12.5 per cent (five deaths).
PANCREATIC AND GASTROINTESTINAL 1RAUMA IN CHILDREN
377
REFERENCES 1. Anderson, C. B., Weisz, D., Rodger, M. R., et al.: Combined pancreatico-duodenal trauma. Am. J. Surg., 125:530-534, 1973. 2. Anderson, C. B., Connors, J. P., Mejia, C. D., et al.: Drainage methods in the treatment of pancreatic injuries. Surg. GynecoI. Obstet., 138:587-590,1974. 3. Anderson, M. C.: Management of pancreatic pseudocysts. Am. J. Surg., 123:209-220, 1972. 4. Fu, W. R., and Stanton, L. W.: Angiographic study of pseudocysts of the pancreas. J. Canad. Assoc. Radio!., 20:176-179, 1969. 5. Grosfeld, J. L.: Blunt abdominal trauma in children. Pediat. Dig., 13:22-28, 1971. 6. Hutson, D. G., Zeppa, R., and Warren, W. D.: Prevention of postoperative hemorrhage after pancreatic cyst-gastrostomy. Ann. Surg., 177:689-692, 1973. 7. Kakos, G. S., Grosfeld, J. L., and Morse, T. S.: Small bowel injuries in children after blunt abdominal trauma. Ann. Surg., 174:238-241,1971. 8. Kilman, J. W., Kaiser, G. C., King, R. D., et a!.: Pancreatic pseudocysts in infancy and childhood. Surgery, 55:455-461, 1964. 9. Laird, C. A., and Clagett, O. T.: Mediastinal pseudocyst of the pancreas in a child: Report of a case. Surgery, 60:465-469, 1966. 10. Miller, R. E.: Pancreatic pseudocysts in infants and children. Arch. Surg., 89:517-521, 1964. 11. Northrup, W. F., III, and Simmons, R. L.: Pancreatic trauma: A review. Surgery, 71 :2743,1972. 12. Orloff, M. J., and Charters, A. C.: InjUries of the small bowel and mesentery and retroperitoneal hematoma. Surg. Clin. N. Amer., 52:729-734, 1972. 13. Otherson, H. B., Moore, F. T., and Boles, E. T., Jr.: Traumatic pancreatitis and pseudocyst in childhood. J. Trauma, 8:535-546,1968. 14. Owens, M. P., and Wolfman, E. F., Jr.: Pancreatic trauma: Management and presentation of a new technique. Surgery, 73:881-886,1973. 15. Ranson, J. H. C.: Safer intraperitoneal drainage. Surg. GynecoI. Obstet., 137:841-842, 1973. 16. Schumer, W., McDonald, G. 0., Nichols, R. L., et a!.: Transgastric cystgastrostomy. Surg. GynecoI. Obstet., 137:48-50,1973. 17. Steele, M., Sheldon, G. F., and Blaisdell, F. W.: Pancreatic injuries, methods, and management. Arch. Surg., 106:544-549, 1973. 18. Stewart, D. R., Byrd, C. L., and Schuster, S. R.: Intramural hematomas of the alimentary tract in children. Surgery, 68:550-557,1970. 19. Stone, H. H.: Pancreatic and duodenal trauma in children. J. Pediat. Surg., 7:670-675, 1972. 20. Stone, H. H., and Whitehurst, J. 0.: Pseudocysts of the pancreas in children. Am. J. Surg., 114:448-453,1967. 21. Sturim, H. S.: The surgical management of pancreatic injuries. Surg. GynecoI. Obstet., 122:133-138,1966. 22. Thomford, N. R., and Jesseph, J. E.: Pseudocyst of the pancreas: A review of 50 cases. Am. J. Surg., 118:86-94,1969. 23. Warren, W. D., Marsh, W. H., and Muller, W. H., Jr.: Experimental production of pseudocysts of the pancreas with preliminary observations of internal drainge. Surg. GynecoI. Obstet., 105:385-393, 1957. 24. Warshaw, A.: Inflammatory masses following acute pancreatitis. Surg. Clin. N. Amer., 54:621-636,1974. 25. Weitzman, J. J., and Swenson, 0.: Traumatic rupture of the pancreas in a toddler. Surgery, 57:309-312, 1965. 26. Williams, R. A., and Sargent, B. A.: The mechanism of intestinal injury in trauma. J. Trauma, 3:288-294,1963.
James Whitcomb Riley Hospital for Children noo West Michigan Street Indianapolis, Indiana 46202 (Dr. Grosfeld)
Pancreatic and Gastrointestinal Trauma in Children Jay L. Grosfeld, M.D.,* and Donald R. Cooney, M.D.**
Blunt abdominal trauma in children presents a significant challenge to the responsible physician in regard to recognition, diagnosis, and treatment. While much attention has been given to the more common intra-abdominal injuries (liver, spleen, kidney), pancreatic and gastrointestinal trauma in children have received relatively little attention. Injuries to the pancreas and gastrointestinal tract continue to be a significant cause of morbidity and mortality in children and typify the diagnostic difficulties encountered in cases of blunt abdominal trauma. This article describes experiences with these injuries in 56 children.
PANCREATIC TRAUMA Pancreatic trauma was observed in 16 patients at the James Whitcomb Riley Hospital for Children, Indianapolis, Indiana. Twelve patients were boys (75 per cent) and four were girls. The average age was 9.5 years with a range of 10 months to 17 years. The cause of pancreatic trauma was related to automobile accidents (as passenger or pedestrian) in five patients, bicycle handle bar injury in four, falls against objects in three, "go-cart" accident in one, physical trauma by siblings or friends in two, and unknown in one (child abuse suspected). The injuries were often insidious in nature and considered trivial in a number of instances. Two children were sent home following emergency room evaluation at the time of initial injury. Three patients had appendectomy performed for suspected appendicitis at other institutions. All three had normal appendices and had associated excess peritoneal fluid at operation. All except one patient in this group were iniFrom the Section of Pediatric Surgery, Department of Surgery, lndiana University School of Medicine, and the James Whitcomb Riley Hospital for Children, Indianapolis, Indiana 'Professor and Director, Section of Pediatric Surgery, Indiana University School of Medicine; Surgeon-in-Chief, James Whitcomb Riley Hospital for Children, Indianapolis, Indiana **Chief Resident, Department of Surgery, Indiana University School of Medicine, Indianapolis, Indiana
Pediatric Clinics of North America- Vol. 22, No.2, May 1975
365
366
JAY
L. GROSFELD AND DONALD R. COONEY
tially evaluated elsewhere and subsequently transferred to the Riley Children's Hospital. Types of injuries included 11 patients with pancreatic pseudocysts, four with pancreatic contusion, laceration, or disruption, and one with complete transection. Associated injuries were relatively infrequent (three patients) and usually involved combined pancreatic and duodenal injuries. Clinical Patterns Persistent abdominal pain, bilious vomiting, fever, and leukocytosis were observed in the majority of patients. An elevated serum amylase was invariably present. A persistently elevated amylase level for prolonged periods of time was noted in all 11 patients with pseudocysts. The mean amylase level in patients with pseudocysts was 515 Somoygi units (upper limit of normal = 180). An epigastric mass was palpable in eight of 11 patients with pseudocysts. False negative paracentesis was frequently noted due to theretroperitoneallocation of the injury. Plain abdominal x-ray examination showed evidence of free fluid and a "sentinalloop" in the upper midabdomen in two of four patients with pancreatic contusion and laceration. One patient with pancreatic pseudocysts presented with bilateral pleural effusions on chest x-ray. Upper gastrointestinal contrast studies were suggestive of pseudocysts in 10 of 11 patients so studied (Fig. 1). Treatment and Results Patients with pseudocysts were managed by cyst-gastrostomy in five patients (Fig. 2), cyst-Roux-en-Y jejunostomy in five cases, and excision in one. Interval from time of injury to operation was greater than four weeks in six cases and more than four months in five. Two patients had operation done 11 months and 22 months respectively following trauma. One patient developed recurrent symptoms of pancreatitis follOwing cystojejunostomy and required sphincterotomy and drainage four years later. Gastrostomy was used as an adjunct in three patients and in addition two of these patients were managed postoperatively by total parenteral hyperalimentation for 10 to 14 days. Serum amylase returned to normal and symptoms resolved in each patient. A nine year old girl with pancreatic transection underwent distal pancreatectomy, splenectomy, and pancreatojejunostomy. One child underwent evacuation of a duodenal hematoma, gastrostomy, and drainage of the pancreas via the lesser sac. Two patients with lacerations of the pancreas were managed by pancreatic suture, drainage of the lesser sac, cholecystostomy, gastrostomy, antibiotics, and total parenteral hyperalimentation. A 22 month old boy with a combined pancreatoduodenal injury (duodenal laceration), craniocerebral injury, and extremity fractures died 12 hours following closure of the duodenal laceration and drainage as a result of the cerebral injury. This was the only death in this group of patients (6.2 per cent mortality). Complications included pancreatic abscess, wound infection, and
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
367
Figure 1. Upper gastrointestinal contrast study demonstrates a retrogastric mass compressing the posterior gastric wall. The findings are consistent with the diagnosis of pancreatic pseudocyst.
gram-negative sepsis in one patient each. One patient developed gastrointestinal bleeding following cystogastrostomy which subsided. Evidence of recurrent pancreatitis occurred in two cases following cystojejunostomy for pseudocyst in one and after distal pancreatectomy for pancreatic transection in the other. Each was managed successfully by sphincterotomy and drainage.
GASTROINTESTINAL TRAUMA This article also concerns experience with 40 infants and children with gastrointestinal injuries following blunt abdominal trauma. Twenty-six of the patients were treated at the Children's Hospital,
368
JAY
L. GROSFELD AND DONALD R. COONEY
Figure 2. Transgastric cystgastrostomy anastomosis between posterior gastric wall and adherent pseudocyst wall.
Columbus, Ohio,* and 14 at the Riley Children's Hospital in Indianapolis, Indiana. Thirty-three of the patients were boys (82 per cent), and seven were girls. The mean age was 7.3 years with a range of three days to 15 years. Presenting symptoms, cause, type, and location of injury, physical findings, and results of treatment were evaluated. Cause of Injury Miscellaneous falls against objects often during play resulted in approximately one-half of the injuries, and automobile accidents with the victim either as a passenger or pedestrian were responsible for one third. Four children were kicked by a horse (10 per cent), and malicious assault (child abuse) occurred in three. Location and Type of Injury There were 26 perforations (65 per cent), 12 obstructing hematomas (30 per cent), and two mesenteric avulsions (5 per cent) with devitalized bowel. The site of injury was duodenal in 17 children, jejunal in 14, and ileal in seven. In addition, there were two gastric perforations (Fig. 3). ·Previously reported.7
369
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
TRAUMA TO STOMACH AND SMALL BOWEL
Figure 3. trauma.
Location of 40 gastrointestinal injuries in childhood resulting from blunt
Fifteen of the 40 patients (37.5 per cent) had significant associated injuries involving the liver in five, musculoskeletal system in five, pancreas in five, brain in four, diaphragm, lung, spleen, and colon in two each, and the genitourinary tract in one (Table 1). Admission to the hospital was reasonably prompt (2 to 12 hours) in many cases, but a delay greater than four days from the onset of injury was observed in eight children, six of whom had duodenal hematoma.
Symptoms and Physical Findings Abdominal pain and tenderness were the only consistent findings and were observed in all conscious patients. Twenty-five patients (60
Table 1. Associated Injuries Noted in 35 Per Cent of Children with Gastrointestinal Trauma ASSOCIATED INJURIES
15 patients (15 of 40) (37.5 per cent) Liver Musculoskeletal Diaphragm Central Nervous System Pancreas Lung Kidney Spleen Colon
Total
5
5 2
4 5 2 1 2
2
28
370
JAY
L.
GROSFELD AND DONALD
R.
COONEY
per cent) had bilious vomiting. Peritoneal irritation was elicited in 15 patients and abdominal distention was noted in nine. Eight patients presented in shock; all had associated major injuries. Pertinent laboratory data were limited to leukocytosis in 25 patients and anemia in 10. An elevated serum amylase was noted in 17 patients of whom only five actually had associated pancreatic trauma. Intraperitoneal free air was demonstrated on abdominal x-ray in less than 50 per cent of patients with perforations. Nine of 12 patients with duodenal hematomas had upper gastrointestinal contrast studies and demonstrated the obstructive lesion in each case (Fig. 4).
Treatment and Results Thirty-five of the 40 children underwent operation. Twenty of 26 perforations were treated with simple debridement and transverse closure of the defect, whereas four required limited bowel resection with primary anastomosis. Two patients with mesenteric avulsion also required resection and anastomosis. Seven of 12 patients with obstructing hematomas were treated primarily by operative evacuation. Five additional children with duodenal hematoma (documented by contrast swallow) were treated initially by nasogastric decompression and parenteral alimentation. Spontaneous resolution occurred in three patients
Figure 4. Upper gastrointestinal contrast study demonstrating obstruction due to duodenal hematoma.
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
371
within one week obviating the need for operation. The remaining two patients, however, required laparotomy and hematoma evacuation due to persistent evidence of obstruction. There were five deaths (12.5 per cent mortality rate). Two unoperated patients whose records dated back 30 years died within two and three days, respectively, due to unrecognized bowel perforation following injuries sustained by being kicked by a horse. Two other deaths occurred within 24 hours of operation and were associated with multiple trauma. A third death occurred 40 days after initial operation and was attributed to sepsis and peritonitis as a result of ileal perforation related to intestinal adhesions. Nonfatal complications occurred in 35 per cent of patients and included atelectasis, duodenal fistula, prolonged ileus, pancreatic pseudocyst formations, and wound infections. The mean duration of hospitalization in survivors was 18 days.
DISCUSSION Recognition and optimal treatment of pancreatic and gastrointestinal injuries in the pediatric age group following blunt abdominal injury are often hampered by a delay in diagnosis. Unlike the adult where easily identified penetrating injuries are commonplace, 90 per cent of abdominal injuries in children are the result of blunt abdominal trauma.5 Often the causative injury follows seemingly inconsequential trauma to the abdomen, an event which may not be related to the parent. In younger patients, nonproductive child-parent or childphysician communication, coupled with inappropriate physical examination, is a significant cause of delay in diagnosis and treatment. The location of the duodenum and pancreas high in the retroperitoneal space and the mobility of the rest of the intestine on its mesentery often protect these structures from injury. Similarly, symptoms and physical findings related to these injuries are often subtle in nature and compound the problems in establishing an early diagnosis. The mechanism of pancreatic injury resulting from blunt abdominal trauma is directly related to a compressing force which crushes the body of the gland against the vertebral columnY' 21 This may result in contusion, laceration, or disruption of the body of the pancreas and pancreatic duct. If the impact of force is to the right or left of the midline (tangential), the duodenum and head of the gland or the spleen and tail may be injured. Pancreatic injury is often overlooked as physical signs may be deceptively minimal due to the retroperitoneal position of the gland. Abdominal x-rays and peritoneal tap are often not helpfulY Diagnostic difficulties encountered (particularly in children) are typified by the fact that three children reported herein underwent appendectomy elsewhere for abdominal pain. Weitzman and Swenson,25 Laird and Clagett,9 and Stone2° have reported similar instances of pancreatic injury also operated upon with an incorrect preoperative diagnosis of acute appen-
372
JAY
L. GROSFELD AND DONALD R. COONEY
dicitis. Pancreatic injury may be overlooked even at the time of laparotomy done for trauma unless the lesser sac is opened to carefully examine the entire body and tail of the gland, and a Kocher maneuver performed to fully evaluate the head of the gland and the entire duodenum.2 In addition, if a "masking" hematoma over a paravertebral fracture of the body of the gland is not uncovered, a ductal injury may similarly be overlooked. Without recognition and appropriate operation, pancreatic disruption may be lethal or lead to subsequent fistula, pseudocyst, or scarring resulting in persistent pancreatitis.24 Sepsis and hemorrhage (two common sequelae of these injuries) often develop in the presence of necrotic tissue, persistent inflammation, and digestive pancreatic enzymes resulting from ductal disruption. The mortality rate following pancreatic trauma is reported to be 20 per cent, and when combined with duodenal injuries is as high as 45 per cent. 1, 11, 14, 17,21 Injuries to the head of the gland are more lethal than those involving the body or tail.H The mortality rate increases proportionately with the number of associated injuries to other organ systems and if the patient presents in shock.11 ,17 The only death in this group (6.2 per cent mortality) occurred in a two year old with associated duodenal and craniocerebral injuries. Serum amylase determination should be obtained in all patients with upper abdominal trauma. Although an elevated amylase level does not necessarily mean a pancreatic injury has occurred, when associated with persistent abdominal pain and tenderness, this observation suggests that laparotomy be performed. Acute pancreatic injuries were successfully managed by suture repair or distal resection and drainage. A sump catheter within a Penrose drain as described by Ranson et al. and others is a useful method of drainage.2 , 15 A treatment program that places the injured gland "at rest" by diverting the bile and gastric juice (by temporary cholecystectomy and gastrostomy), prohibiting oral intake, and maintaining caloric needs (by total parenteral hyperalimentation) is continued until amylase levels return to normal and symptoms subside. This was a very satisfactory method of therapy in two patients. Owens and Wolfman have found these above diversion adjuncts helpful in instances of pancreatic trauma and in addition performed a jejunostomy for feeding purposes.14 Others have suggested distal pancreatectomy be performed for ductal injuries, however, these procedures often carry a greater morbidity (particularly persistent fistula) and mortality rate than drainage aloneP' 19 The only patient in this study who underwent resection developed a fistula and eventually required reoperation for recurring symptoms of pancreatitis. Although pancreatoduodenectomy has been suggested for extensive injuries to the head of the pancreas and duodenum, this procedure carries with it a high operative mortality.l,17 In the present group of patients, we did not encounter a pancreatic injury in the childhood age group that required this procedure. Pseudocysts are the result of a collection of necrotic tissue, old blood, and secretions that escape from the injured pancreas and loculate
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
373
in the lesser sac. Occasionally, these contents may dissect into the mediastinum or retroperitoneum.9 The inflammatory reaction that follows encapsulates the pseudocyst with a fibrous wall. Over 80 per cent of pseudocysts in the adult are due to complications of alcoholic pancreatitis. In contrast, more than 70 per cent in childhood are a direct result of trauma. lO • 11 Pseudocysts are associated with a symptom complex of pain, fever, and vomiting. An epigastric mass is frequently palpable. Laboratory data often demonstrate an elevated serum amylase, leukocytosis, and occasionally hyperbilirubinemia. Thomford and Jesseph noted the triad of a mass, elevated serum amylase, and leukocytosis in half of their 50 patients with pseudocysts. 22 All the children in this study had persistently elevated amylase levels and eight of 11 had a palpable mass. X-ray examination of the chest and abdomen is often helpful in achieving a diagnosis. Twenty per cent of patients present with a pleural effusion. 24 One patient in this review demonstrated this finding on chest x-ray. The upper gastrointestinal contrast study was the most useful roentgenologic examination and was consistent with a diagnosis of pseudocyst in 10 of 11 cases. Reports of recent experiences with ultrasound echo grams and selective arteriograms suggest these studies may be useful in making the diagnosis of pseudocyst. 4. 24 Ultrasound discerns between a cystic (pseudocyst) or solid (tumor) mass, while arteriography usually demonstrates an avascular mass in instances of pseudocyst. Pseudocysts should all be treated operatively as they may result in duodenal obstruction, biliary obstruction, free rupture (peritonitis), hemorrhage (erosion into major vessels), splenic vein thrombosis, gastric ulceration, and pancreatic ascites.24 The operation of choice is somewhat controversial and is usually determined by the location and duration of the pseudocyst. Others on et al. reported successful use of external drainage of pancreatic pseudocysts by use of a mushroom catheter in five childrenP This type of modified marsupialization drained for 3 to 4 weeks in four patients and five months in another. There were no deaths or recurrences. The objections usually raised when this procedure is employed is a high recurrence rate and prolonged persistence of fistula. Anderson noted excessive losses of fluid and electrolytes and severe skin breakdown in 23 per cent of patients when external drainage was employed.3 These data, however, were obtained from observations in adults who ofte'n had underlying chronic pancreatitis and ductal abnormalities, and may not apply to pancreatic trauma in children with previously normal glands. Warren et al. have shown it takes from four to six weeks from the time of injury for a pseudocyst capsule to thicken enough to hold sutures effectively.23 As a general rule, we reserve external drainage for the rare patient in whom the cyst wall is thin and too friable to hold sutures. In most instances, internal drainage of the pseudocyst is preferred and the type of anastomosis is chosen according to the location of the pseudocyst. If the cyst is adherent to the stomach wall, cystgastrostomy has been quite effective therapy in our hands. Hutson et al. noted a 50
374
JAY
L.
GROSFELD AND DONALD
R.
COONEY
per cent hemorrhage rate following cystgastrostomy and considered this related to the effects of gastric acid entering the cyst cavity.6 They further suggested that the edges of the stomach wall not be sutured so that they are not hindered from acting as a flap valve, thereby reducing the entry of gastric juices into the cavity.6 In contrast, Schumer et al. and Warshaw report much lower rates (3 per cent) of hemorrhage following cystgastrostomy and stress the importance of careful suture approximation to avoid hemorrhage. 16 ,24 We follow the latter suggestion and in addition employ a Stamm gastrostomy over the site of the cystgastrostomy anastomosis, prohibit oral intake, and use total parenteral hyperalimentation for 10 to 14 days postoperatively to prevent gastric juice and food particles from entering the cavity. The pseudocyst cavity is an informal structure without an epithelial lining. Once evacuated, the cavity rapidly shrinks and disappears within a couple of weeks, particularly if the pancreas is placed "at rest" as outlined above. Kilman et al. suggested that cystojejunostomy is the treatment of choice for pseudocyst as it avoids reflux of gastric contents into the cyst.s The Roux-en-Y loop technique was employed in five patients in this series with one recurrence observed. This procedure is advisable in instances in which the pseudocyst is not adherent to the stomach wall. Warshaw reported recurrence rates of 3 per cent with cyst gastrostomy, 5 per cent with cystjejunostomy, and 8 per cent with cystduodenostomy.24 Warren et al. found no difference between the effectiveness of cystgastrostomy or cystjejunostomy in treating experimentally produced pseudocysts.23 Low recurrence rates should be expected in children with pseudocysts resulting from pancreatic trauma due to an absence of underlying pancreatic disease and ductal obstructionP Recurrences in such cases rarely do occur and usually reflect the healing of the ductal injury by excessive scarring.
GASTROINTESTINAL TRAUMA Injuries to the stomach are quite rare due to its mobility and the protection afforded this viscus by the rib cage. Most retrospective reviews concerning blunt abdominal trauma usually fail to mention instances of gastric trauma. Both individuals in the present series sustained injury immediately following a large meal and had a distended stomach at the time of epigastric impact. Both responded to simple closure of perforation. Most gastric perforations occur on the anterior wall, however, careful examination of the posterior wall through the lesser sac must be done to avoid missing an associated posterior wall perforation as was observed in one of our patients. Small bowel injuries occur in from 5 to 15 per cent of children hospitalized for blunt abdominal trauma. 7 Despite improvements in diagnosis and management, these lesions continue to result in a significant morbidity and mortality. Mortality rates of 5 to 20 per cent and complication rates of 35 per cent have been recorded.5 , 7, 12 The complication rate in this group of patients was 35 per cent and the mortality rate was
PANCREATIC AND GASTROINTESTINAL TRAUMA IN CHILDREN
375
12.5 per cent. Death following isolated injuries is unusual, however, the mortality may be high with an increasing number of associated organ injuriesP The small bowel is compressible and mobile and often escapes damage in its midportion. Williams and Sargent presented experimental evidence suggesting that a common mechanism of nonpenetrating intestinal trauma involves a crushing of the bowel against the vertebral column.26 In instances of duodenal hematoma, the hemorrhage originates where the duodenum is compressed by the vertebral column and spreads between the muscular layers of the duodenum about an equal distance in each direction. The second mechanism involves a tearing or shearing force of the bowel and its mesentery at points of intestinal fixation. The small bowel injuries in this review occurred either in the duodenum, in the jejunum in close proximity to the ligament of Treitz, and the distal ileum close to the cecum which is attached to the lateral abdominal wall. In these areas, the intestine is relatively fixed increasing its susceptibility to disruption, contusion, or avulsion. Pathologic fixation such as those related to intestinal adhesions from a previous operation may also increase the susceptibility of the small bowel near the adhesions to shearing forcesP Aside from the factors that tend to inhibit prompt recognition of these injuries that have previously been alluded to, inappropriate physical examination may also result in diagnostic delays. A meaningful examination is often difficult to achieve in an irritable and uncooperative child. The use of a short-acting barbiturate (seconal, 2 mg per pound of body weight, intramuscularly) will usually relieve the child's anxiety without masking true abdominal tenderness. 5• 7 Unwarranted reliance on laboratory results may also promote delays in diagnosis. Neither anemia nor leukocytosis was a specific finding. Serum amylase level was elevated in 17 of the 40 patients, but only five of the 17 patients actually had pancreatic trauma. Although upper gastrointestinal contrast radiographs are frequently diagnostic in patients with obstructing duodenal hematomas, plain roentgenograms demonstrated free-air in less than 50 per cent of children with perforations. The longest diagnostic delays are noted in instances of obstructing hematomas. Stewart et al. reported eight cases at the Boston Children's Hospital with duration of symptoms from 1 to 6 days prior to diagnosis. 18 Pain followed by bilious vomiting is the most frequent finding. Obstructing hematomas have been treated in a wide variety of fashions including bypassing gastroenterostomies, surgical decompression, and prolonged nonoperative nasogastric suction. Our approach to this injury is as follows: When the duodenal hematoma coexists with other intraabdominal injuries warranting early operation, it is evacuated at that time obviating the need for a second operation. An isolated injury resulting in hematoma is managed by nasogastric suction and parenteral alimentation for one week. At the end of this time, repeat contrast roentgenograms will often demonstrate partial resolution of the obstruction (in approximately one half of the cases) and avoid the need for operative intervention.7 If obstruction persists at this time, the hematoma has usually liquified and can easily be relieved by subserosal evacuation.
376
JAY
L.
GROSFELD AND DONALD
R.
COONEY
At operation, the full extent of the injury can best be observed if the right colon is mobilized and elevated to the left and a Kocher maneuver is performed. The hematoma may extend to the proximal jejunum and this area must also be appreciated and the obstructing hematoma appropriately drained. Simple perforations of the jejunum and ileum can be managed by either debridement of the edges and direct closure (80 per cent) or by limited resection with primary anastomosis when the lesion is more extensive. A transverse closure of the perforation is suggested to avoid narrowing of the lumen. Instances of mesenteric avulsion require resection of the bowel supplied by the injured mesenteric vessels and primary end-to-end intestinal anastomosis.
SUMMARY Injuries to the pancreas and gastrointestinal tract following blunt abdominal trauma continue to be a significant cause of morbidity and mortality in the pediatric age group. Optimal treatment of these injuries is frequently hampered by considerable delays in diagnosis. Factors contributing to these delays include the location of much of the duodenum and the pancreas in the retroperitoneum resulting in an absence of initial symptoms and signs, the often trivial nature of some of the responsible blunt traumatic accidents, inappropriate child-parent or childphysician communication, failure to achieve a meaningful physical examination in uncooperative or unconscious patients, and false negative paracentesis. Eighty per cent of these injuries occurred in boys. Eleven of 16 patients with pancreatic trauma had pseudocysts. A persistently elevated serum amylase level was invariably noted and an epigastric mass was palpable in eight patients. Significant delays in diagnosis were prevalent and pseudocyst was misdiagnosed as appendicitis in three cases. Internal drainage by cyst gastrostomy or cystjejunostomy was effective operative treatment. In instances of acute pancreatic injuries, sump drains, gastrostomy, cholecystostomy, and total parenteral hyperalimentation were useful therapeutic adjuncts. There was one death for a 6.2 per cent mortality rate. Forty patients had gastrointestinal injuries involving the duodenum in 17, jejunum in 14, ileum in seven, and stomach in two. Perforations occurred in 65 per cent of cases, obstructing hematomas in 30 per cent, and mesenteric avulsions in 5 per cent. Associated injuries were observed in 15 patients (37.5 per cent). Pain and tenderness were the only consistent findings. Upper gastrointestinal contrast studies were diagnostic of duodenal hematomas. Eighty per cent of perforations were managed by simple closures and 20 per cent by resection and anastomosis. Obstructing hematomas unassociated with other injuries may be expected to resolve without requiring operation in 50 per cent of patients managed conservatively. Complications occurred in 35 per cent of patients and the mortality rate was 12.5 per cent (five deaths).
PANCREATIC AND GASTROINTESTINAL 1RAUMA IN CHILDREN
377
REFERENCES 1. Anderson, C. B., Weisz, D., Rodger, M. R., et al.: Combined pancreatico-duodenal trauma. Am. J. Surg., 125:530-534, 1973. 2. Anderson, C. B., Connors, J. P., Mejia, C. D., et al.: Drainage methods in the treatment of pancreatic injuries. Surg. GynecoI. Obstet., 138:587-590,1974. 3. Anderson, M. C.: Management of pancreatic pseudocysts. Am. J. Surg., 123:209-220, 1972. 4. Fu, W. R., and Stanton, L. W.: Angiographic study of pseudocysts of the pancreas. J. Canad. Assoc. Radio!., 20:176-179, 1969. 5. Grosfeld, J. L.: Blunt abdominal trauma in children. Pediat. Dig., 13:22-28, 1971. 6. Hutson, D. G., Zeppa, R., and Warren, W. D.: Prevention of postoperative hemorrhage after pancreatic cyst-gastrostomy. Ann. Surg., 177:689-692, 1973. 7. Kakos, G. S., Grosfeld, J. L., and Morse, T. S.: Small bowel injuries in children after blunt abdominal trauma. Ann. Surg., 174:238-241,1971. 8. Kilman, J. W., Kaiser, G. C., King, R. D., et a!.: Pancreatic pseudocysts in infancy and childhood. Surgery, 55:455-461, 1964. 9. Laird, C. A., and Clagett, O. T.: Mediastinal pseudocyst of the pancreas in a child: Report of a case. Surgery, 60:465-469, 1966. 10. Miller, R. E.: Pancreatic pseudocysts in infants and children. Arch. Surg., 89:517-521, 1964. 11. Northrup, W. F., III, and Simmons, R. L.: Pancreatic trauma: A review. Surgery, 71 :2743,1972. 12. Orloff, M. J., and Charters, A. C.: InjUries of the small bowel and mesentery and retroperitoneal hematoma. Surg. Clin. N. Amer., 52:729-734, 1972. 13. Otherson, H. B., Moore, F. T., and Boles, E. T., Jr.: Traumatic pancreatitis and pseudocyst in childhood. J. Trauma, 8:535-546,1968. 14. Owens, M. P., and Wolfman, E. F., Jr.: Pancreatic trauma: Management and presentation of a new technique. Surgery, 73:881-886,1973. 15. Ranson, J. H. C.: Safer intraperitoneal drainage. Surg. GynecoI. Obstet., 137:841-842, 1973. 16. Schumer, W., McDonald, G. 0., Nichols, R. L., et a!.: Transgastric cystgastrostomy. Surg. GynecoI. Obstet., 137:48-50,1973. 17. Steele, M., Sheldon, G. F., and Blaisdell, F. W.: Pancreatic injuries, methods, and management. Arch. Surg., 106:544-549, 1973. 18. Stewart, D. R., Byrd, C. L., and Schuster, S. R.: Intramural hematomas of the alimentary tract in children. Surgery, 68:550-557,1970. 19. Stone, H. H.: Pancreatic and duodenal trauma in children. J. Pediat. Surg., 7:670-675, 1972. 20. Stone, H. H., and Whitehurst, J. 0.: Pseudocysts of the pancreas in children. Am. J. Surg., 114:448-453,1967. 21. Sturim, H. S.: The surgical management of pancreatic injuries. Surg. GynecoI. Obstet., 122:133-138,1966. 22. Thomford, N. R., and Jesseph, J. E.: Pseudocyst of the pancreas: A review of 50 cases. Am. J. Surg., 118:86-94,1969. 23. Warren, W. D., Marsh, W. H., and Muller, W. H., Jr.: Experimental production of pseudocysts of the pancreas with preliminary observations of internal drainge. Surg. GynecoI. Obstet., 105:385-393, 1957. 24. Warshaw, A.: Inflammatory masses following acute pancreatitis. Surg. Clin. N. Amer., 54:621-636,1974. 25. Weitzman, J. J., and Swenson, 0.: Traumatic rupture of the pancreas in a toddler. Surgery, 57:309-312, 1965. 26. Williams, R. A., and Sargent, B. A.: The mechanism of intestinal injury in trauma. J. Trauma, 3:288-294,1963.
James Whitcomb Riley Hospital for Children noo West Michigan Street Indianapolis, Indiana 46202 (Dr. Grosfeld)
