Management of pancreatic necrosis and timing of surgery.
Introduction
Severe Acute Pancreatitis (SAP) is a common cause of admission to the ICU and presents many challenges for management. This case highlights the difficulty of if and when such patients should have surgery.
Clinical problem
A 65 year old admitted to hospital with abdominal pain and vomiting. He had a past medical history of a coronary artery bypass graft (CABG) in 2004 and was otherwise fit and well with a history confirming his ability to increase his oxygen consumption to greater than 4 times resting rate (>4 metabolic equivalents of task). He had no history of excess alcohol consumption.
Laboratory investigations revealed an amylase of 3428 and acute renal failure. A diagnosis of severe acute pancreatitis was made, the patient was transferred to a surgical ward, and fluid resuscitation commenced. An ultrasound revealed no gallstones. Urine output remained low and the following day he became acidotic (pH 7.21, lactate 4.4, BE -11.6), hyperkalaemic (6.1) and hypothermic. He was admitted to the ICU for monitoring, fluid management and renal replacement therapy.
Management
After commencement on CVVHDF he became hypotensive and was started on noradrenaline. Over the next few hours his respiratory function deteriorated with a hypercapnic acidosis and he was intubated and ventilated.
On day 5 Tazocin was started due to deteriorating gas exchange and increasing noradrenaline requirements. The noradrenaline was weaned off over the next 48 hours.
On day 8 he developed a pyrexia of 39 degrees and vasopressors were restarted. His central lines were changed and Tazocin changed to meropenem, vancomycin and caspofungin. His pyrexia settled within 24h and vasopressors were again able to be weaned off.
During the first days of admission thrombocytopenia developed with platelets falling from 160 to 44 by day 5. A HIT screen was positive and anticoagulation for RRT was changed to danaparide.
Nasogastric feeding was started on day 4 of admission but the patient failed to absorb the feed and TPN was commenced on day 8. A nasojejenal tube was inserted endoscopically by the gastroenterologists on day 13 and enteral feeding recommenced.
A percutaneous tracheostomy was sited on day 10 due to the likelihood of an extended period of ventilation.
CT scanning on day 6 revealed extensive pancreatic necrosis and a large fluid collection around the pancreas containing gas. A repeat CT on day 12 was done without contrast which added no new information. Surgery was considered with the decision to delay it to a later date (discussed below).
For the next week there was continual clinical improvement - renal function recovered and vasopressors were weaned off. Antibiotics were continued during this time. On day 21 a necrosectomy was performed with a drain left in for irrigation and drainage of the pancreatic bed. An NJ tube was sited during the operation (the previous one had come out) to facilitate enteral feeding.
Surgery led to an immediate deterioration with vasodilatory shock and acute kidney injury necessitating RRT once again. From this point on the patient made a steady recovery and was discharged to the ward 2 weeks later.
Discussion
50% of patients with SAP will develop necrosis and about half of these will develop shock and multi organ failure (MOF). Of the 50% who develop necrosis, 40% if these will develop pancreatic infection by week 3-4. Infection is thought to come from bowel translocation of gram –ve organisms. Prophylactic antibiotics (ABX) reduce infection in some studies but may increase infection with resistant bacteria or fungi. There is currently therefore insufficient evidence to support prophylactic ABX, antifungals or selective gut decontamination with or without necrosis. It is essential however to rigorously try to detect infection when it occurs. This is done by blood cultures and fine needle aspiration (radiologically guided) of pancreatic or peripancreatic collections. CT scanning may reveal retroperitoneal gas while clinical deterioration and pyrexia of >39 is suggestive of infection (Marik). Broad spectrum ABX (carbapenems ideal as excellent pancreatic penetration) should therefore be used if infection is suspected and the patient very unwell. It is essential to remember that the treatment for infected necrosis is surgical debridement, not just antibiotics.
Without debridement death is almost inevitable if infection does occur but with necrosis mortality is very high if surgery is performed in 1st 2 weeks. As stated above, however, infection of necrosis tends not to develop until week 3-4. By this time necrosis is well demarcated allowing a single procedure with good pancreatic preservation. Patients may need repeated laparotomies for washouts or a laparostomy. Open necrosectomy unfortunately has a high morbidity and mortality even after 2 weeks and there is now an increasing shift towards minimally invasive approaches (Navaneethan). Percutaneous drainage can be used as a temporary way of controlling sepsis until more definitive surgery. Endoscopic approaches though the stomach or duodenum for debridement and necrosectomy have now been described. Laparoscopic necrosectomy may have a role. Retroperitoneal drainage with a nephroscope seems to also be successful. All these techniques are dependent on the individual patient and the experience of the operator but seem destined to gain in popularity.
The patient in this case summary presented a challenge as signs of infection (peripancreatic gas) and MOF (not necessarily due to infection) were present earlier than week 3. While the incidence of infection peaks in week 3, infection occurs in the first 7 days in up to 25% of patients (Dugernier). On day 14 the ICU consultant and surgical consultant (upper GI specialist) discussed the timing of surgery. It was felt that while the patient was clinically stable with reducing inflammatory markers, surgery should wait until better demarcation of necrosis had developed with the proviso that any deterioration would prompt surgery. By the time of surgery on day 21, good demarcation had indeed occurred and the outcome was successful.
It is also worth noting that surgery can relieve biliary obstruction and ERCP +/- sphincterotomy in 1st 72 hours is recommended if obstructive jaundice or acute cholangitis is present.
Lessons Learnt
This case provides important lessons on the management of pancreatic necrosis in both the use of antibiotics and the timing of surgery. I realised in the early part of the patient’s stay that he needed surgery (after the CT scan showing necrosis and gas) but was unsure how to proceed also knowing that outcome is worse in the first 2 weeks of necrosis. It was here that the experience of a specialist surgeon proved invaluable. It seems that in this case the antibiotics, whilst not a definitive cure, allowed a window of opportunity for watching and waiting for definitive treatment.
References
25 Years of Progress and Innovation in Intensive Care Medicine. 2007. Kuhlen R, Moreno R, Ranieri M, Rhodes A
Nathens A et al. Management of the critically ill patient with severe acute pancreatitis. Crit Care Med 2004; 32: 2524-2536
Isenmann R et al. Prophylactic antibiotic treatment in patients with predicted severe acute pancreatitis: a placebo-controlled, double-blind trial. Gastroenterology 2004; 126: 997-1004
Paul E. Marik. Fever in the ICU. Chest 2000;117;855-869
Navaneethan et al. Minimally invasive techniques in pancreatic necrosis. Pancreas. 2009 Nov;38(8):867-75.
Dugernier et al. Pancreatitis. Organ specific problems. PACT module. ESICM
