298
Injury, 8, 298-302
Printed in Great Britain
Traumatic
lesions of the duodenum
N. Y. Wijemanne,
E. D. Rodrigo and Rudra Rasaretnam
Colombo
General Hospital,
Colombo,
Summary
Injuries of the duodenum are relatively uncommon on account of the organ’s size and position. Since most of it is retroperitoneal, lesions involving it give rise to such subtle physical and radiological signs thal the diagnosis is often overlooked in the early phase after injury. Twenty-six cases of duodenal injury are reviewed, 18 of which were due to penetrating wounds and the remaining 8 to blunt trauma. Anterior penetrating wounds were usually associated with other intraperitoneal lesions which caused more obvious physical signs and thus drew attention to the necessity for exploration. On the other hand, both blunt trauma and posterior stab wounds frequently caused isolated retroperitoneal duodenal lesions where the diagnosis was not evident on admission, but in which
the insidious and progressive development of symptoms and signs drew attention to the need for laparotomy. Early repair combined with drainage of the retroperitoneal space resulted in a good result in 23 of 26 cases, 4 of whom, however, developed a temporary lateral duodenal fistula. Two of the 3 deaths were in patients who presented late and had associated pancreatic injuries while the third was due to an abdominal vascular injury.
INTRODUCTION AU the abdominal viscera, the duodenum is least often affected by trauma partly on account of its small size, but more because of its situation on the posterior abdominal wall where the greater part of it is retroperitoneal. Injury of this part often leads to the delayed onset of symptoms and signs, and since the possibility of duodenal injury is often overlooked, there is an increased morbidity and mortality with such lesions, particularly when they occur alone. Penetrating abdominal wounds in such cases are almost invariably associated with injuries to other viscera which are more obvious. This may result in lacerations of the retroperitoneal duodenum OF
Sri Lanka being missed at operation. In this paper we review 26 cases of duodenal trauma treated over the past 10 years.
CLINICAL MATERIAL AND RESULTS Twenty-five of the 26 patients were males, aged between 13 and 52 years, the mean being 31.2 years. Eighteen had sustained penetrating injuries, 13 of which were due to stab and 5 to gunshot wounds. The remaining 8 injuries were caused by blunt trauma, but only one was due to the high-speed automobile accident often described in other series in which the driver sustains abdominal compression by the steering wheel. Of the other 7, 3 were due to falls from a height, 2 to blows and the other 2 occurred in cyclist and pedestrian accidents. Penetrating injuries Fifteen of the 18 patients with penetrating injuries were admitted within 4 hours of injury, 2 within 8 hours and the other at 25 hours. Sixteen of the 18 patients had anterior entry wounds; all had clinical signs of peritoneal irritation and underwent laparotomy within 3 hours of admission. All 5 patients with gunshot wounds had associated injuries to other abdominal viscera, but 2 of the 11 anterior stab wounds had caused injury confined to the second part of the duodenum, but even in these 2 there was sufficient contamination to give rise to the usual signs of peritoneal irritation. Two patients with posterior stab wounds had less obvious physical signs on admission, but both complained of persisting and even increasing abdominal pain, and this, with the development of peritonism led to operation being carried out respectively 6 and 20 hours after admission. The systolic blood pressure was over 100 mm Hg in 8 of the stabbed persons and in 2
299
Wijemanne et al. Duodenal Lesions Tab/e 1. Associated injuries Stabs Number of patients Survived Associated injuries Stomach Small intestine Large bowel Pancreas Liver Kidney Gall bladder Ureter Inferior vena cslva Heart
Gunshot
13 13
5 4
4 0 1 1 4 2 1 1
2 1 1 2 3 1
1
3 1
Blunt 8 6
1 1*
-
*Serosal tears in the colon in the patient who also had small intestinal rupture.
of those that had been shot; the other 5 victims of stabbing had pressures of between 80 and 100 mm Hg, as did one of those with gunshot wounds. The remaining 2 patients with gunshot injuries had unrecordable pressures due to vena caval lacerations. Sixteen of the 18 patients were explored within 3 hours of admission; the 2 patients who were initially treated conservatively were those with posterior stab wounds but one of these was explored because of a progressive fall in blood pressure that was due to retroperitoneal haemorrhage. Nine of the 13 stab wounds and all 5 gunshot injuries had caused associated injuries involving the liver, stomach, kidney, pancreas, colon and gall bladder, in that order of frequency. Two patients also had penetrating cardiac wounds, which required repair (Table 1). A retroperitoneal haematoma was present in 8 cases. The duodenal lesion was in the first part in 5, the second part in 10 and the third and fourth parts in 3 each. One of the stab wounds and the 5 gunshot wounds had through-and-through injuries affecting one part of the duodenum, while 2 patients had entry wounds in the first part (intraperitoneal) and exit wounds in the second part (retroperitoneal), and one patient :had separate stab wounds of the extremities of the fourth part, both being almost circumferential. Sixteen were repaired by direct suture in two layers, 1 had resection of the fourth part and end-to-end anastomosis, and there was 1 intraoperative death due to exsanguination from a vena caval laceration. In 3 patients the suture line was reinforced with a pedicled omental graft. Three other patients developed lateral duodenal fistulae on the fourth, fifth and eighth days. Two of these resolved spontaneously
within a week, during which time they remained on intravenous fluids, while the third required a feeding jejunostomy on the eighteenth day and then healed after 4 weeks. One other patient underwent a second operation-nephrectomyfollowing failure of a ureteric anastomosis. Wound sepsis occurred in 6 patients, and 1 patient developed a burst abdomen. Hospital stay varied between 8 and 65 days, with an average of 23 days for stab wounds and 16 for gunshot wounds. There was 1 death in the 18 patients with penetrating injury. Blunt trauma
Eight patients had duodenal injuries following blunt trauma; 4 of them have been included in a previous report (Rasaretnam and Thavendran, 1974). The amount of force varied from mild blows in the abdomen to severe injury caused by steering wheel compression. The sequence of abdominal pain at the time of injury, followed by a temporary asymptomatic period and then by progressive development of pain, has been a feature in most of these cases. Four were admitted within 6 hours of injury and the other 4 between 14 and 120 hours. Some abdominal pain and guarding were present in all 8 patients at the time of admission, but only 4 of them were considered to have an indication for immediate exploration. Three of these patients were in the delayed admission group. The other 4 were initially treated conservatively. Increasing abdominal pain with the development of more guarding were the main indications for operation, which was carried out between 6 and 20 hours after admission. Two patients had positive abdominal taps and 2 had radiological evidence of gas in the retroperitoneal tissues. Only 2 of the 8 had other visceral intraperitoneal lesions (Table I), which was in marked contrast to the patients with penetrating injuries. Two patients had lesions in the first part of the duodenum, 3 in the second, 2 at the junction of the second and third parts and 1 in the fourth. A retroperitoneal haematoma was present in each of the patients with injury to the second, third or fourth parts of the duodenum. It was nearly always crepitant and sometimes contained bile. Only 2 patients had associated pancreatic injury; neither had disruption of the gland or the duct system, but both succumbed. The duodenal rent was nearly always circumferential and was more extensive than the lesions seen with penetrating trauma. All 8 patients had direct two-layer closure and 2 had omental reinforcements. Delayed repair was difficult as the sutures tended to cut through easily. One patient
300
developed a lateral duodenal fistula on the third day but after 14 days of intravenous therapy, spontaneous closure occurred. There were 6 survivors whose period of hospital stay averaged 13 days. At the time of operation all 8 patients had free fluid in the peritoneal cavity which was purulent in 4, bile-stained in 2 and bloodstained in 2 others. Corrugated drains down to the retroperitoneal duodenum were used in 18 of the 24 patients, who were returned to the intensive care unit, the main exceptions being patients with injuries to the first part. DISCUSSION Diagnosis Duodenal injuries are uncommon, and in a collected review of the cases reported in the literature, Kerry and Glas (1962) found only 314 cases, 192 of which followed blunt trauma and carried a mortality of 54 per cent. More recent reports indicate an increasing incidence of duodenal injuries due mainly to high-speed automobile accidents in which the driver suffers severe abdominal trauma owing to the impact of the steering wheel, which produces either shearing or blow-out lesions of the retroperitoneal duodenum (Roman et al., 1971; Lucas and Ledgerwood, 1975). Such accidents are unusual in Sri Lanka, but as shown by the cases presented in this report, even moderate blunt abdominal trauma may result in duodenal rupture. Awareness of the possibility of duodenal rupture and early definitive surgery with emphasis on the preoperative management of anaemia and hypovolaemia are essential for a successful result (Donovan and Hagen, 1966; Lucas and Ledgerwood, 1975). Although the presence of associated lesions will obviously have a bearing on the result, most authors agree that the presence of severe pancreatic injury is the most important single factor in determining the outcome. Penetrating injuries are almost always associated with damage to other structures, which dominates the clinical picture and usually provides definite evidence for immediate exploration. However, with posterior stab wounds, isolated lesions may occur in the retroperitoneal part of the duodenum and the signs of visceral injury then develop slowly, as they may do after blunt trauma. Abdominal pain and mild guarding are usually present on admission, but not of a degree to indicate immediate exploration. Both increase with time, and although initially their significance may be
Injury:
the British Journal
of Accident
Surgery Vol. ~/NO. 4
easily overestimated, particularly in inebriated patients and those with head injuries, repeated examination reveals their progressive nature, and should indicate the necessity for exploration. While extravasated duodenal and pancreatic secretions remain behind the peritoneum, the patient experiences relatively mild symptoms, but they slowly progress as more fluid collects. Ultimately, the posterior peritoneum is breached, giving rise to generalized peritonitis, which may then be easily recognized, if it is looked for. The place of peritoneal lavage in the diagnosis of retroperitoneal injuries is still controversial. Engrav et al. (1975) found that 3 of the 7 patients whose only injury was duodenal had a negative result. Although we did not use it routinely, the presence of free fluid in the peritoneal cavity at the time of operation in all 8 cases of blunt trauma suggests that the test might have been used profitably. However, since there was a minimal delay of 4% hours between injury and operation, it would perhaps be advisable to defer peritoneal lavage for 4 hours in patients with blunt trauma who are initially selected for observation. Persistent or increasing abdominal pain, and even mild increase in guarding, are suggestive of mischief, and delayed peritoneal lavage in these patients will probably yield more positive results. The classic radiological signs of retroperitoneal duodenal rupture have been well documented (Sperling and Rigler, 1937; Lucas and Ledgerwood, 1975), and confirmation of the diagnosis is easily made using a watersoluble contrast medium (Estes et al., 1952). Treatment The presence of a periduodenal haematoma at operation should arouse suspicion of the duodenum being injured and the presence of bile or air in it is confirmatory. Every such haematoma should be explored to identify a perforation of the bowel. Simple duodenal tears, even when associated with minor pancreatic injury, have been successfully treated by direct suture. However, the liability of these lesions to result in lateral duodenal fistulae, which carry an increased morbidity and mortality, has prompted some surgeons to reinforce the duodenal suture line with an on-lay jejunal serosal patch (Jones and Joergenson, 1963; Kobold and Thal, 1963). In a review of their cases, McInnis et al. (1975) found no difference in the results of those treated by direct suture and those by suture reinforced with a jejunal patch. The development of a lateral duodenal fistula is one of the most serious complications of duodenal rents, and
Wijemanne
et al.
: Duodenal Lesions
although it is said to occur infrequently with isolated injuries to the viscus, this is not borne out in our cases. The incidence of fistula increases when pancreatic injury coexists, and is probably due to the presence of activated pancreatic enzymes (Berne et al., 1974). It therefore seems prudent to use prophylactic measures to prevent its occurrence; having been impressed by the use of omentum in dealing with perforated peptic ulcers, even when seen late, and its use in the closure of the difficult duodenal stump (Bigg et al., 1964), we have used it in 5 patients, all of
whom made uneventful recoveries. Duodenal perforations should be repaired by two-layer closure in the transverse axis. This is easily achieved with the blow-out or shearing type of lesion caused by blunt trauma where the tear is usually circumferential. If, however, the repair narrows the lumen, either gastrojejunostomy should be added or an intraluminal tube passed beyond the site of repair (Roman et al., 1971). Lateral duodena.1 fistulae were associated with a mortality of 50-60 per cent before the era of intravenous hyperalimentation, because of excessive losses of fluids and electrolytes (Cukingnan et al., 1975). Three of the 4 patients in our series who developed this complication healed spontaneously after just a few days of intravenous therapy, while the fourth required a temporary feeding jejunostomy, as has been used by Welch and Edmunds (1962) and Donovan and Hagen (1966). Numerous operations have been described for the repair of a lateral duodenal fistula. Some are aimed at reducing gastric and biliary secretions, others involve patch repair of the duodenal laceration, internal drainage of the fistula into a loop of jejunum and the relief of distal obstruction, when present (Kobald and Thal, 1963; Wolfman et al., 1964; Donovan and Hagen, 1966; Cukingnan et al., 1975). In all of these procedures the surgeon is faced with the difficulty of dealing with friable tissues which hold sutures feebly, and by the poor general condition of most of these patients due to a high intestinal fistula. Feeding jejunostomy, which is a relatively simple procedure, has yielded gratifying results in some of these cases (Welch and Edmunds, 1962; Donovan and Hagen, 1966). Biliary drainage by either choledochostomy or cholecystostomy has been suggested for lesions near the ampulla of Vater or the intrapancreatic portion of the common bile duct (Berne et al., 1974). For the more complicated cases of combined pancreatic and duodenal injury, especially those with injury to the duct system which has a
301
poor prognosis when treated by direct suture, more extensive operative procedures have been described. Berne et al. (1974) used duodenal diverticulization by gastric antrectomy and end-to-side gastrojejunostomy, tube jejunostomy, closure of the duodenal wounds and drainage, and had a mortality of 16 per cent in 50 cases. Nance and DeLoach (1971) carried out pancreaticoduodenectomy with a 30 per cent mortality. Perhaps the most important factor in the treatment of duodenal trauma is early surgery. In a review of the literature, Miller (1916) found 100 per cent mortality when surgery was delayed for over tihours, and 50 years later there has been only minimal improvement in the results, Roman et al. (1971) reporting a 65 per cent mortality.
REFERENCES
Berne C. J., Donovan A. J., White E. J. et al. (1974) Duodenal diverticulization for duodenal and pancreatic injury. Am. J. Surg. 127, 503. Bigg R. L., Silver J. M. and Kram D. D. (1964) Omental closure of the duodenal stump. Am. J. Surg. 108, 724. Cukingnan R. A., Culliford A. T. and Worth M. H. (1975) Surgical correction of a Iateral duodenal fistula with Roux-Y technique; report of a case. J. Trauma 15, 519. Donovan A. J. and Hagen W. E. (1966) Traumatic perforation of the duodenum. Am. J. Surg. 111,341. Engrav H., Benjamin C. I., S&ate R. G. et al. (1975) Diagnostic peritoneal lavage in blunt abdominal trauma. J. frauma 15, 854. _ Estes W. L.. Bowman T. L. and Meilicke F. F. (19521 Non-pen&rating abdominal trauma with s‘peciaj reference to lesions of the duodenum and pancreas. Am. J. Surg. 83,434. Jones S. A. and Joergenson E. J. (1963) Closure of duodenal wall defects. Surgery 53, 438. Kerry R. L. and Glas W. W. (1962) Traumatic lesions of the pancreas and duodenum: a clinical and experimental study. Arch. Surg. 85, 813. Kobald E. E. and Thal A. P. (1963) A simple method for the management of experimental wounds of the duodenum. Surg. Gynecol. Obstet. 116, 340. Lucas C. E. and Ledgerwood A. M. (1975) Factors influencing the outcome after blunt duodenal injury. J. Trauma 15, 839. McInnis W. D., Aust J.B. andCruzA. B. (1975) Traumatic injuries of the duodenum. A comparison of primary closure and the jejunal patch. J. Trauma 15, 847. Miller R. T. (1916) Retroperitoneal rupture of the duodenum by blunt force. Ann. Surg. 64, 550.
Injury: the British Journal of Accident
302
Nance F. C. and DeLoach D. H. (1971) Pancreaticoduodenectomy following abdominal trauma. J. Trauma 11, 577.
Rasaretnam R. and Thavendran A. (1974) Rupture of retroperitoneal duodenum after blunt the abdominal trauma. Br. J. Surg. 61, 893. Roman E., Silva Y. J. and Lucas C. (1971) The management of blunt duodenal injury. Surg. Gynecol.
Obstet.
132, 7.
Surgery Vol. ~/NO. 4
Sperling L. and Rigler L. G. (1937) Traumatic retroperitoneal rupture of the duodenum; description of valuable roentgen observation in its recognition. Radiology 29, 521. Welch C. E. and Edmunds L. H. (1962) Gastrointestinal fistula. Surg. Clin. North Am. 42, 1311. Wolfman E. F., Trevino G., Heals D. K. et al. (1964) An operative technic for the management of acute and chronic duodenal fistulas. Ann. Surg. 159, 563.
Requests for reprints should be addressed to: Rudra Rasaretnam, 58 Kynsey Road, Colombo 8, Sri Lanka.
THIRD CONFERENCE MEDICINE
ON MATERIALS AND BIOLOGY
FOR USE IN
A conference on the Mechanical Properties of Biomaterials will be held at Keele University, Staffordshire, on 13, 14 and 15 September, 1978. The aim of the meeting is to review the present state of knowledge about the mechanical aspects of human and implant materials relevant to orthopaedics, cardiovascular surgery, neurosurgery, plastic and reconstructive surgery and dentistry. It is hoped to have specialist topic sessions related to the special problems of the surgical disciplines and/or to the classes of materials. Anyone wishing to submit a paper should write immediately to the address given below outlining the subject matter. A preliminary programme and registration details will be published in early 1978. Communications concerning the conference should be addressed to: Dr G. W. Hastings, Bio-Medical Engineering Unit, c/o Medical Institute, Hartshill, Stoke-on-Trent, Staffordshire, ST4 7NY, England.
Injury, 8, 298-302
Printed in Great Britain
Traumatic
lesions of the duodenum
N. Y. Wijemanne,
E. D. Rodrigo and Rudra Rasaretnam
Colombo
General Hospital,
Colombo,
Summary
Injuries of the duodenum are relatively uncommon on account of the organ’s size and position. Since most of it is retroperitoneal, lesions involving it give rise to such subtle physical and radiological signs thal the diagnosis is often overlooked in the early phase after injury. Twenty-six cases of duodenal injury are reviewed, 18 of which were due to penetrating wounds and the remaining 8 to blunt trauma. Anterior penetrating wounds were usually associated with other intraperitoneal lesions which caused more obvious physical signs and thus drew attention to the necessity for exploration. On the other hand, both blunt trauma and posterior stab wounds frequently caused isolated retroperitoneal duodenal lesions where the diagnosis was not evident on admission, but in which
the insidious and progressive development of symptoms and signs drew attention to the need for laparotomy. Early repair combined with drainage of the retroperitoneal space resulted in a good result in 23 of 26 cases, 4 of whom, however, developed a temporary lateral duodenal fistula. Two of the 3 deaths were in patients who presented late and had associated pancreatic injuries while the third was due to an abdominal vascular injury.
INTRODUCTION AU the abdominal viscera, the duodenum is least often affected by trauma partly on account of its small size, but more because of its situation on the posterior abdominal wall where the greater part of it is retroperitoneal. Injury of this part often leads to the delayed onset of symptoms and signs, and since the possibility of duodenal injury is often overlooked, there is an increased morbidity and mortality with such lesions, particularly when they occur alone. Penetrating abdominal wounds in such cases are almost invariably associated with injuries to other viscera which are more obvious. This may result in lacerations of the retroperitoneal duodenum OF
Sri Lanka being missed at operation. In this paper we review 26 cases of duodenal trauma treated over the past 10 years.
CLINICAL MATERIAL AND RESULTS Twenty-five of the 26 patients were males, aged between 13 and 52 years, the mean being 31.2 years. Eighteen had sustained penetrating injuries, 13 of which were due to stab and 5 to gunshot wounds. The remaining 8 injuries were caused by blunt trauma, but only one was due to the high-speed automobile accident often described in other series in which the driver sustains abdominal compression by the steering wheel. Of the other 7, 3 were due to falls from a height, 2 to blows and the other 2 occurred in cyclist and pedestrian accidents. Penetrating injuries Fifteen of the 18 patients with penetrating injuries were admitted within 4 hours of injury, 2 within 8 hours and the other at 25 hours. Sixteen of the 18 patients had anterior entry wounds; all had clinical signs of peritoneal irritation and underwent laparotomy within 3 hours of admission. All 5 patients with gunshot wounds had associated injuries to other abdominal viscera, but 2 of the 11 anterior stab wounds had caused injury confined to the second part of the duodenum, but even in these 2 there was sufficient contamination to give rise to the usual signs of peritoneal irritation. Two patients with posterior stab wounds had less obvious physical signs on admission, but both complained of persisting and even increasing abdominal pain, and this, with the development of peritonism led to operation being carried out respectively 6 and 20 hours after admission. The systolic blood pressure was over 100 mm Hg in 8 of the stabbed persons and in 2
299
Wijemanne et al. Duodenal Lesions Tab/e 1. Associated injuries Stabs Number of patients Survived Associated injuries Stomach Small intestine Large bowel Pancreas Liver Kidney Gall bladder Ureter Inferior vena cslva Heart
Gunshot
13 13
5 4
4 0 1 1 4 2 1 1
2 1 1 2 3 1
1
3 1
Blunt 8 6
1 1*
-
*Serosal tears in the colon in the patient who also had small intestinal rupture.
of those that had been shot; the other 5 victims of stabbing had pressures of between 80 and 100 mm Hg, as did one of those with gunshot wounds. The remaining 2 patients with gunshot injuries had unrecordable pressures due to vena caval lacerations. Sixteen of the 18 patients were explored within 3 hours of admission; the 2 patients who were initially treated conservatively were those with posterior stab wounds but one of these was explored because of a progressive fall in blood pressure that was due to retroperitoneal haemorrhage. Nine of the 13 stab wounds and all 5 gunshot injuries had caused associated injuries involving the liver, stomach, kidney, pancreas, colon and gall bladder, in that order of frequency. Two patients also had penetrating cardiac wounds, which required repair (Table 1). A retroperitoneal haematoma was present in 8 cases. The duodenal lesion was in the first part in 5, the second part in 10 and the third and fourth parts in 3 each. One of the stab wounds and the 5 gunshot wounds had through-and-through injuries affecting one part of the duodenum, while 2 patients had entry wounds in the first part (intraperitoneal) and exit wounds in the second part (retroperitoneal), and one patient :had separate stab wounds of the extremities of the fourth part, both being almost circumferential. Sixteen were repaired by direct suture in two layers, 1 had resection of the fourth part and end-to-end anastomosis, and there was 1 intraoperative death due to exsanguination from a vena caval laceration. In 3 patients the suture line was reinforced with a pedicled omental graft. Three other patients developed lateral duodenal fistulae on the fourth, fifth and eighth days. Two of these resolved spontaneously
within a week, during which time they remained on intravenous fluids, while the third required a feeding jejunostomy on the eighteenth day and then healed after 4 weeks. One other patient underwent a second operation-nephrectomyfollowing failure of a ureteric anastomosis. Wound sepsis occurred in 6 patients, and 1 patient developed a burst abdomen. Hospital stay varied between 8 and 65 days, with an average of 23 days for stab wounds and 16 for gunshot wounds. There was 1 death in the 18 patients with penetrating injury. Blunt trauma
Eight patients had duodenal injuries following blunt trauma; 4 of them have been included in a previous report (Rasaretnam and Thavendran, 1974). The amount of force varied from mild blows in the abdomen to severe injury caused by steering wheel compression. The sequence of abdominal pain at the time of injury, followed by a temporary asymptomatic period and then by progressive development of pain, has been a feature in most of these cases. Four were admitted within 6 hours of injury and the other 4 between 14 and 120 hours. Some abdominal pain and guarding were present in all 8 patients at the time of admission, but only 4 of them were considered to have an indication for immediate exploration. Three of these patients were in the delayed admission group. The other 4 were initially treated conservatively. Increasing abdominal pain with the development of more guarding were the main indications for operation, which was carried out between 6 and 20 hours after admission. Two patients had positive abdominal taps and 2 had radiological evidence of gas in the retroperitoneal tissues. Only 2 of the 8 had other visceral intraperitoneal lesions (Table I), which was in marked contrast to the patients with penetrating injuries. Two patients had lesions in the first part of the duodenum, 3 in the second, 2 at the junction of the second and third parts and 1 in the fourth. A retroperitoneal haematoma was present in each of the patients with injury to the second, third or fourth parts of the duodenum. It was nearly always crepitant and sometimes contained bile. Only 2 patients had associated pancreatic injury; neither had disruption of the gland or the duct system, but both succumbed. The duodenal rent was nearly always circumferential and was more extensive than the lesions seen with penetrating trauma. All 8 patients had direct two-layer closure and 2 had omental reinforcements. Delayed repair was difficult as the sutures tended to cut through easily. One patient
300
developed a lateral duodenal fistula on the third day but after 14 days of intravenous therapy, spontaneous closure occurred. There were 6 survivors whose period of hospital stay averaged 13 days. At the time of operation all 8 patients had free fluid in the peritoneal cavity which was purulent in 4, bile-stained in 2 and bloodstained in 2 others. Corrugated drains down to the retroperitoneal duodenum were used in 18 of the 24 patients, who were returned to the intensive care unit, the main exceptions being patients with injuries to the first part. DISCUSSION Diagnosis Duodenal injuries are uncommon, and in a collected review of the cases reported in the literature, Kerry and Glas (1962) found only 314 cases, 192 of which followed blunt trauma and carried a mortality of 54 per cent. More recent reports indicate an increasing incidence of duodenal injuries due mainly to high-speed automobile accidents in which the driver suffers severe abdominal trauma owing to the impact of the steering wheel, which produces either shearing or blow-out lesions of the retroperitoneal duodenum (Roman et al., 1971; Lucas and Ledgerwood, 1975). Such accidents are unusual in Sri Lanka, but as shown by the cases presented in this report, even moderate blunt abdominal trauma may result in duodenal rupture. Awareness of the possibility of duodenal rupture and early definitive surgery with emphasis on the preoperative management of anaemia and hypovolaemia are essential for a successful result (Donovan and Hagen, 1966; Lucas and Ledgerwood, 1975). Although the presence of associated lesions will obviously have a bearing on the result, most authors agree that the presence of severe pancreatic injury is the most important single factor in determining the outcome. Penetrating injuries are almost always associated with damage to other structures, which dominates the clinical picture and usually provides definite evidence for immediate exploration. However, with posterior stab wounds, isolated lesions may occur in the retroperitoneal part of the duodenum and the signs of visceral injury then develop slowly, as they may do after blunt trauma. Abdominal pain and mild guarding are usually present on admission, but not of a degree to indicate immediate exploration. Both increase with time, and although initially their significance may be
Injury:
the British Journal
of Accident
Surgery Vol. ~/NO. 4
easily overestimated, particularly in inebriated patients and those with head injuries, repeated examination reveals their progressive nature, and should indicate the necessity for exploration. While extravasated duodenal and pancreatic secretions remain behind the peritoneum, the patient experiences relatively mild symptoms, but they slowly progress as more fluid collects. Ultimately, the posterior peritoneum is breached, giving rise to generalized peritonitis, which may then be easily recognized, if it is looked for. The place of peritoneal lavage in the diagnosis of retroperitoneal injuries is still controversial. Engrav et al. (1975) found that 3 of the 7 patients whose only injury was duodenal had a negative result. Although we did not use it routinely, the presence of free fluid in the peritoneal cavity at the time of operation in all 8 cases of blunt trauma suggests that the test might have been used profitably. However, since there was a minimal delay of 4% hours between injury and operation, it would perhaps be advisable to defer peritoneal lavage for 4 hours in patients with blunt trauma who are initially selected for observation. Persistent or increasing abdominal pain, and even mild increase in guarding, are suggestive of mischief, and delayed peritoneal lavage in these patients will probably yield more positive results. The classic radiological signs of retroperitoneal duodenal rupture have been well documented (Sperling and Rigler, 1937; Lucas and Ledgerwood, 1975), and confirmation of the diagnosis is easily made using a watersoluble contrast medium (Estes et al., 1952). Treatment The presence of a periduodenal haematoma at operation should arouse suspicion of the duodenum being injured and the presence of bile or air in it is confirmatory. Every such haematoma should be explored to identify a perforation of the bowel. Simple duodenal tears, even when associated with minor pancreatic injury, have been successfully treated by direct suture. However, the liability of these lesions to result in lateral duodenal fistulae, which carry an increased morbidity and mortality, has prompted some surgeons to reinforce the duodenal suture line with an on-lay jejunal serosal patch (Jones and Joergenson, 1963; Kobold and Thal, 1963). In a review of their cases, McInnis et al. (1975) found no difference in the results of those treated by direct suture and those by suture reinforced with a jejunal patch. The development of a lateral duodenal fistula is one of the most serious complications of duodenal rents, and
Wijemanne
et al.
: Duodenal Lesions
although it is said to occur infrequently with isolated injuries to the viscus, this is not borne out in our cases. The incidence of fistula increases when pancreatic injury coexists, and is probably due to the presence of activated pancreatic enzymes (Berne et al., 1974). It therefore seems prudent to use prophylactic measures to prevent its occurrence; having been impressed by the use of omentum in dealing with perforated peptic ulcers, even when seen late, and its use in the closure of the difficult duodenal stump (Bigg et al., 1964), we have used it in 5 patients, all of
whom made uneventful recoveries. Duodenal perforations should be repaired by two-layer closure in the transverse axis. This is easily achieved with the blow-out or shearing type of lesion caused by blunt trauma where the tear is usually circumferential. If, however, the repair narrows the lumen, either gastrojejunostomy should be added or an intraluminal tube passed beyond the site of repair (Roman et al., 1971). Lateral duodena.1 fistulae were associated with a mortality of 50-60 per cent before the era of intravenous hyperalimentation, because of excessive losses of fluids and electrolytes (Cukingnan et al., 1975). Three of the 4 patients in our series who developed this complication healed spontaneously after just a few days of intravenous therapy, while the fourth required a temporary feeding jejunostomy, as has been used by Welch and Edmunds (1962) and Donovan and Hagen (1966). Numerous operations have been described for the repair of a lateral duodenal fistula. Some are aimed at reducing gastric and biliary secretions, others involve patch repair of the duodenal laceration, internal drainage of the fistula into a loop of jejunum and the relief of distal obstruction, when present (Kobald and Thal, 1963; Wolfman et al., 1964; Donovan and Hagen, 1966; Cukingnan et al., 1975). In all of these procedures the surgeon is faced with the difficulty of dealing with friable tissues which hold sutures feebly, and by the poor general condition of most of these patients due to a high intestinal fistula. Feeding jejunostomy, which is a relatively simple procedure, has yielded gratifying results in some of these cases (Welch and Edmunds, 1962; Donovan and Hagen, 1966). Biliary drainage by either choledochostomy or cholecystostomy has been suggested for lesions near the ampulla of Vater or the intrapancreatic portion of the common bile duct (Berne et al., 1974). For the more complicated cases of combined pancreatic and duodenal injury, especially those with injury to the duct system which has a
301
poor prognosis when treated by direct suture, more extensive operative procedures have been described. Berne et al. (1974) used duodenal diverticulization by gastric antrectomy and end-to-side gastrojejunostomy, tube jejunostomy, closure of the duodenal wounds and drainage, and had a mortality of 16 per cent in 50 cases. Nance and DeLoach (1971) carried out pancreaticoduodenectomy with a 30 per cent mortality. Perhaps the most important factor in the treatment of duodenal trauma is early surgery. In a review of the literature, Miller (1916) found 100 per cent mortality when surgery was delayed for over tihours, and 50 years later there has been only minimal improvement in the results, Roman et al. (1971) reporting a 65 per cent mortality.
REFERENCES
Berne C. J., Donovan A. J., White E. J. et al. (1974) Duodenal diverticulization for duodenal and pancreatic injury. Am. J. Surg. 127, 503. Bigg R. L., Silver J. M. and Kram D. D. (1964) Omental closure of the duodenal stump. Am. J. Surg. 108, 724. Cukingnan R. A., Culliford A. T. and Worth M. H. (1975) Surgical correction of a Iateral duodenal fistula with Roux-Y technique; report of a case. J. Trauma 15, 519. Donovan A. J. and Hagen W. E. (1966) Traumatic perforation of the duodenum. Am. J. Surg. 111,341. Engrav H., Benjamin C. I., S&ate R. G. et al. (1975) Diagnostic peritoneal lavage in blunt abdominal trauma. J. frauma 15, 854. _ Estes W. L.. Bowman T. L. and Meilicke F. F. (19521 Non-pen&rating abdominal trauma with s‘peciaj reference to lesions of the duodenum and pancreas. Am. J. Surg. 83,434. Jones S. A. and Joergenson E. J. (1963) Closure of duodenal wall defects. Surgery 53, 438. Kerry R. L. and Glas W. W. (1962) Traumatic lesions of the pancreas and duodenum: a clinical and experimental study. Arch. Surg. 85, 813. Kobald E. E. and Thal A. P. (1963) A simple method for the management of experimental wounds of the duodenum. Surg. Gynecol. Obstet. 116, 340. Lucas C. E. and Ledgerwood A. M. (1975) Factors influencing the outcome after blunt duodenal injury. J. Trauma 15, 839. McInnis W. D., Aust J.B. andCruzA. B. (1975) Traumatic injuries of the duodenum. A comparison of primary closure and the jejunal patch. J. Trauma 15, 847. Miller R. T. (1916) Retroperitoneal rupture of the duodenum by blunt force. Ann. Surg. 64, 550.
Injury: the British Journal of Accident
302
Nance F. C. and DeLoach D. H. (1971) Pancreaticoduodenectomy following abdominal trauma. J. Trauma 11, 577.
Rasaretnam R. and Thavendran A. (1974) Rupture of retroperitoneal duodenum after blunt the abdominal trauma. Br. J. Surg. 61, 893. Roman E., Silva Y. J. and Lucas C. (1971) The management of blunt duodenal injury. Surg. Gynecol.
Obstet.
132, 7.
Surgery Vol. ~/NO. 4
Sperling L. and Rigler L. G. (1937) Traumatic retroperitoneal rupture of the duodenum; description of valuable roentgen observation in its recognition. Radiology 29, 521. Welch C. E. and Edmunds L. H. (1962) Gastrointestinal fistula. Surg. Clin. North Am. 42, 1311. Wolfman E. F., Trevino G., Heals D. K. et al. (1964) An operative technic for the management of acute and chronic duodenal fistulas. Ann. Surg. 159, 563.
Requests for reprints should be addressed to: Rudra Rasaretnam, 58 Kynsey Road, Colombo 8, Sri Lanka.
THIRD CONFERENCE MEDICINE
ON MATERIALS AND BIOLOGY
FOR USE IN
A conference on the Mechanical Properties of Biomaterials will be held at Keele University, Staffordshire, on 13, 14 and 15 September, 1978. The aim of the meeting is to review the present state of knowledge about the mechanical aspects of human and implant materials relevant to orthopaedics, cardiovascular surgery, neurosurgery, plastic and reconstructive surgery and dentistry. It is hoped to have specialist topic sessions related to the special problems of the surgical disciplines and/or to the classes of materials. Anyone wishing to submit a paper should write immediately to the address given below outlining the subject matter. A preliminary programme and registration details will be published in early 1978. Communications concerning the conference should be addressed to: Dr G. W. Hastings, Bio-Medical Engineering Unit, c/o Medical Institute, Hartshill, Stoke-on-Trent, Staffordshire, ST4 7NY, England.
