Figure 27.1 Pseuodcyst. Handlebar trauma. It was treated conservatively, but the patient developed vomiting and a substantial pseudocyst with fracture of the body of the pancreas.

A very important and specific cause of pancreatic injury in infants and children is the so-called nonaccidental injury (NAI). Usually, this occurs in those too small or young (i.e. toddlers or infants) to identify the perpetrator and is caused by fist or kick to the abdomen. More often, there is other organ damage, and when history and examination are scrutinised, other injuries become manifest. Duodenal injuries are particularly important in this respect. Gaines et al. [5] reported that of a series of 30 children with defined duodenal injury (haematoma or perforation), 8 were nonaccidental in origin. These could also be defined as they were all less than 4 years of age.

Jacombs et al. [6] reported a different, possibly hidden aspect of pancreatic injury, as they identified 22 deaths from a state registry where pancreatic injury was evident and usually identified postmortem and compared them with the more typical clinical series. Most died of head injuries, and all of these were motor vehicle injuries (MVIs), but nonsurvivors were more likely to have multiple abdominal, head and thoracic injuries in addition. There were two deaths in younger children under 2 years both from pancreatoduodenal bleeding.

27.3  CLINICAL FEATURES

There is often a latent period prior to the onset of significant symptoms, but these are usually central abdominal pain and vomiting. Pancreatic pain can be unrelenting and deep seated when established. Initiation of pancreatitis also sets off a systemic inflammatory response with untoward fluid shifts out of the circulation compartment into the retroperitoneum, peritoneal and pleural cavities. There may also be untoward effects of circulating inflammatory and immune complexes on basement membrane, leading to renal failure and respiratory distress with hypoxia, acidosis and so forth.

There may be little to see on the abdominal wall. Skin contusions caused by the lap belt, motor vehicle or handle-bar may be apparent. Sometimes, the latter may disrupt the fascia and muscles leaving the skin intact as an abdominal hernia [7]. In those presenting late with established pancreatitis, lumbar (Grey-Turner’s sign^*) or periumbilical bruising (Cullen’s sign^†) may be apparent. Pseudocysts may be felt in the epigastrium, but again, this is a late sign.

Table 27.1 Classification of pancreatic trauma

Source: After American Association for the Study of Trauma, ICD-9 coding 863.81 to 863.94.

^a The reference point for distinguishing proximal from distal is the superior mesenteric vein.

27.4  INVESTIGATIONS

The key to detection is recognition that it could happen and measurement of the blood amylase levels to confirm the event. Still, there are many caveats to this apparently straightforward process, and measurement of amylase (or lipase) does not seem to have any practical use in screening (i.e. in the absence of clinical suspicion [8]). While it is possible that too early an estimation may miss the rising level, this is unusual in practice. Monitoring of serial levels may also suggest progression of disease and development of complications. Multisystem trauma involving the face may also cause salivary gland damage and a factitious rise in salivary amylase levels. If this is a possibility in someone with possible pancreatic injury, then an isoenzyme profile can be distinguished [9]. Serum lipase levels tend to parallel amylase levels. Still, beyond diagnosis of injury, little else can be gleaned from the actual levels. In a retrospective review of 131 cases of confirmed pancreatic trauma, neither initial nor peak amylase or lipase correlated with grade of injury, and neither amylase nor lipase predicted length of stay or mortality [10]. The only positive association was that maximal amylase (>1100 IU/L) was highly predictive of developing a pseudocyst.

Ultrasound, particularly so-called focused abdominal sonogram for trauma (FASTscan), has no real role in the detection of acute injury, but it does have a role in the diagnosis and monitoring of complications such as pseuodcysts and fluid collections.

Beyond proving pancreatic involvement, the next task is to establish the degree of pancreatic injury and specifically degree of duct injury. The ubiquitous American Association for the Surgery of Trauma (AAST) has provided a grading scheme, replacing the older one (Table 27.1). Most (>60%) injuries in nonselected series are Grade I–II.

27.4.1  CT and MRCP scans

Multi-detector-row computed tomography (CT) scan using double contrast (IV and oral) is the investigative tool of choice, and its diagnostic accuracy for injury nowadays approaches 100%. Features include haematomas (decreased attenuation and reduced contrast), peripancreatic fluid, parenchymal disruption and visualisation of a laceration or fracture, generalised organ enlargement, mesocolic oedema and fluid-separating dorsal pancreas and splenic vein. The key elements which should change initial conservative management are twofold:

1.  Is the duct intact?

2.  Is the duodenum intact?

The former takes more effort to establish than the latter but determines outcome. Magnetic resonance cholangiopancreatography (MRCP) may help, but if there is doubt, then endoscopic retrograde cholangiopancreatography (ERCP) should be definitive. of course, if a nihilistic approach to management of pancreatic trauma has been adopted (see below), then both of these are redundant investigations, as nothing further will be done. If oral contrast leaks or there is free intraperitoneal air (~50% of cases), then duodenum integrity is compromised and leakage of intestinal contents will occur – a laparotomy is indicated. A duodenal haematoma may also be seen, but it is not an absolute indication for surgery [11].

27.4.2  ERCP

ERCP is the definitive investigation for the presence of duct injury, although its adoption has been slow to gain momentum in paediatric practice (Figure 27.2). Still, there is some evidence that it is becoming more widespread [12]. It was first advocated by Hall et al. [13] in a report describing four cases in 1986, and since that time, there have been several published series [14–16]. Our initial experience was published in 2007 with 12 ERCPs, of which 9 were stented [4].

Critics would suggest that the endoscopic skills required for ERCP in this age group are not immediately available in even the largest of children’s hospitals, and that there is a higher risk of actual injury and initiation of pancreatitis particularly in inexperienced hands. The other criticism is that knowledge of duct injury must mean a change in management. Thus, it might only be indicated if you are actually considering surgical (or endoscopic) intervention.

27.5  MANAGEMENT

This remains one of the more controversial areas in paediatric trauma practice, with a distinct geographical variation in practice.

Figure 27.2 ERCP in pancreatic trauma. Eight-year-old boy with innocuous trauma from bicycle handlebar presenting late with substantial pseudocyst. (a) ERCP shows partial tear (arrowed) of the duct (in front of vertebral column). It was successfully stented (b) with a Zimmon stent (c).

The options are detailed and then a suggested algorithm is offered.

27.5.1  Conservative management

All Grade I and II injuries can be managed conservatively with the expectation of resolution. This involves restriction of oral intake, nasogastric intubation (if vomiting is problematic), analgesia and intravenous fluid resuscitation. The child should be monitored, in a surgical environment, with the expectation of resolution of pain and resumption of an oral diet in 2–4 days. Serum amylase levels should parallel the improvement in clinical signs. Failure of conservative management may occur, possibly because of underestimation of grade, but also because pseudocysts may simply occur even in the absence of duct injury [17].

A number of centres also adopt this approach for the more severe injuries (Grade III+) irrespective of the perceived degree of injury or presence of duct laceration. There is an expectation of complications, such as acute fluid collections and later pseudocysts, perhaps in the majority (Table 27.2), but these can be treated and most will ultimately settle. As it is likely that an early resumption of enteral feeding will not be possible, the child should be started on parenteral nutrition. Antibiotics, and perhaps gastric acid suppressants, are usually prescribed, but evidence that they reduce septic complications or shorten the period of pancreatitis is poor. Similarly, the administration of octreotide, while common, lacks actual evidence of benefit, certainly in adult practice [24,25].

27.5.2  Early operative intervention

This may be indicated for Grade III–V injuries, as defined on imaging or for other coexisting intra-abdominal injuries. The surgical approaches differ according to the nature of the injury. By far the most common is a Grade III or IV injury near the midline, and the easiest option is to accept the injury and remove the potential for later complications by resecting the distal pancreas and oversewing the fractured proximal stump and duct. Mostly, this can be achieved by dissecting the gland off the splenic vein, therefore preserving the spleen. There is a parenchymal-sparing alternative whereby the distal gland is anastomosed into a Roux loop [26]. Comparative studies are rare, but one retrospective study of 15 distal pancreatectomies and 10 Roux pancreatojejunostomies for trauma from Nashville, Tennessee, showed equal outcomes in terms of length of stay and complications. The former technique tended to be used in lower-grade injuries in older patients and tolerated feeds significantly earlier.

More difficult to treat is the Grade IV or even V injuries involving the head. A pylorus-sparing Whipple resection of the head is possible, but a formidable prospect in the acute situation [27]. A more prudent approach might be placing of multiple drains with later endoscopic intervention to determine anatomy of duct injury and relationship to subsequent fistula or pseudocyst. Alternatively, creating a Roux loop and enveloping the injured organ provides an internal drainage conduit.

Laparoscopic distal pancreatectomy for trauma is possible with a number of case reports [28,29], a single-centre series of three cases demonstrating its feasibility [30] and a multi-institutional review [31]. Most restrict its use to <72 h postinjury to give the best chance of success and staple the body proximal to the site of injury to achieve duct control. Compromises have been made, with division of splenic vein and artery during dissection in one case and leaving the proximal gland unsutured for fear of ‘damaging the bile duct’ [30]. A ‘handport’ has also been described to aid dissection [30]. The length of time needed to complete resection laparoscopically can also be quite considerable, with a mean time of 3–4 h in the review of six cases [31] and a morbidity rate of >50%.

Table 27.2 Contrasting management in Grade III+ injuries in children

Note: Ps, pseudocyst; Lap, laparotomy; DP, distal pancreatectomy; NOM, non-operative management.

Early precipitate laparotomy with an unstable child or for multitrauma has its own problems. This is actually an unusual scenario, but the advisability of resection and certainly reconstructive pancreatic surgery should be questioned. The principles of damage control surgery emphasise cessation of bleeding and control of gastrointestinal peritoneal contamination. In a large retrospective adult study of 42 cases, abdominal packing alone, packing and draining and resection were compared [32]. The mortality in all groups was formidable (~50%) but significantly less for the pack and drain group.

27.5.3  Endoscopic stenting

This is a stratagem which was designed as a minimally invasive approach to pancreatic trauma and follows on from the definitive investigation for duct injury [15]. It has become the norm in many adult centres, although the driving force (as in our centre) has been access to competent endoscopists.

Many of our children were being referred relatively late in their natural history, with the original units adopting a non-operative approach until complications supervened. ERCP was able to show the underlying ductal injury for the complication, and placement of a stent achieved a measure of internal drainage. More than 20 children have now been treated with stenting, our first report being published in 2007 [4]. The usual limiting factor is age, and therefore duct size. The smallest Zimmon stent is a 6 Fg single pigtail. In the first instance, we tried to pass a guidewire across the laceration into the distal pancreas. This then allowed a transductal stent to be passed. Alternatively, and particularly if the laceration is complete, it may only be possible to pass the tip of the stent into the retroperitoneal collection. This seems to be acceptable. If there is rapid resolution of symptoms and signs, we then leave the stent in situ for 2–3 months before endoscopic removal. More frequently, we will repeat the ERCP 1–2 weeks later and, if possible, substitute a larger gauge stent to improve drainage. Later pseudocysts do still occur, but at a much reduced rate compared with a completely nonoperative approach. Our only real failure occurred after three attempts at stenting in a 7-year-old child, resulting in a distal pancreatectomy at 30 days postinjury.

27.5.4  Summary position

The largest series so far reported is a huge retrospective multi-institutional review of 167 children treated in 14 North American centres [23]. They state unequivocally that in those with duct involvement, distal pancreatectomy is superior to nonoperative management, with more rapid resumption of diet, fewer repeat interventions and a shorter period to complete resolution. A similar series, again retrospective and a multi-institutional review of 43 (20 in surgical cohort) children treated in North America for perceived Grade II and III injuries, did not show much evidence of benefit [33]. Thus, there was no difference in outcome (length of stay), although the incidence of pseudocysts was clearly higher in those treated nonoperatively.

Nonetheless, there are many advocates of nihilism, and the rarity of the condition makes it difficult to adjudicate one way or the other. A recent Cochrane review could also find no comparison or convincing evidence one way or the other for Grade III+ injuries [34].

Box 27.2 illustrates the experience of the two largest centres in Paris and London, with contrasting philosophies. The former were distinctly ‘let us wait and see’ (laissez faire), while the latter were defiantly for the initial endoscopic stenting approach. There were much less complications, particularly pseudocysts, in the English cohort, but the overall length of stay and duration of parenteral nutrition were actually less in the French group.

Stay updated, free articles. Join our Telegram channel

Join