This case involves A 72-year-old man who had initially presented to his local hospital with symptoms and signs of biliary colic and cholestatic jaundice. Magnetic resonance cholangiopancreatography (MRCP) confirmed the presence of an obstructing common bile duct stone. He was transferred to a Dublin hospital for elective endoscopic retrograde cholangiopancreatography (ERCP) to remove the stone.

The procedure proved difficult. Deep cannulation of the common bile duct was not achieved, perhaps due to the obstructing stone. Contrast was injected into the duct and a filling defect in the distal common bile duct was noted on cholangiogram. The endoscopist noted swelling of the ampulla, and tried to perform a sphincterotomy to allow insertion of a stent. However, this too failed. Upon removal of the cannula, the endoscopist feared that the common bile duct had been perforated and the contrast injected submucosally. The procedure was abandoned and an urgent CT ordered. Fortunately, the CT showed no evidence of perforation, but showed persistent and marked dilatation of the common bile duct and the intrahepatic bile ducts.

The patient was transferred back to his local hospital, with a repeat ERCP scheduled for seven days later. However, he deteriorated rapidly. He became febrile, hypotensive, oliguric and developed respiratory distress. Unresponsive to ward-level resuscitative measures, he was transferred back to the Dublin hospital (two days after his first transfer there) and admitted to ICU for haemodynamic and ventilatory support. The patient had become septic secondary to ascending cholangitis (biliary sepsis), and had developed post-ERCP pancreatitis.

Post-ERCP pancreatitis

Pancreatitis occurs after 5-10% of cases of ERCP.^1,2 Pathophysiology is multifactorial involving the mechanical forces required for instrumentation, sphincter of Oddi oedema and spasm (obstructing the flow of pancreatic fluid), microbiological insults, and hydrostatic injury from contrast injection.^3,4 Risk of pancreatitis is reduced with pre-procedure rectal diclofenac⁵ and concomitant fluid administration.⁶

The patient was initially treated with piperacillin/tazobactam and gentamicin. To decompress the biliary system and ensure drainage of the septic fluid, a percutaneous biliary drain was inserted by interventional radiology. Samples of the drained fluid were sent to microbiology for analysis. Fluid culture showed a heavy growth of Citrobacter freundii, which was resistant to ciprofloxacin, piperacillin/tazobactam, aztreonam and third generation cephalosporins, but susceptible to meropenem, and Enterococcus faecium, which was resistant to vancomycin (VRE) but susceptible to daptomycin. The patient was thus started on meropenem and daptomycin, as well as caspofungin to cover fungal organisms.

Percutaneous drain insertion was a temporary measure. He needed a stent. Although this was unsuccessful on the first attempt, the presence of the percutaneous drain provided an interesting opportunity for a second. Six days later, a repeat ERCP was performed with the rendezvous technique. With this technique, the biliary system was successfully cannulated. Cholangiogram showed filling defects and so a balloon trawl was done, which removed sludge and small stones.

When biliary cannulation is unsuccessful during ERCP, the rendezvous technique may be employed.^7,8,9 In this case a guidewire was inserted into the biliary system antegrade through the percutaneous drain. The guidewire was advanced until it exited the ampulla, where it was met by an endoscope. A catheter could then be inserted into the common bile duct over the guidewire. If a percutaneous drain were not present, the bile ducts could be directly punctured via the stomach or duodenum.^8,9

During this second ERCP, a pigtail stent was inserted. However, satisfactory placement of this stent failed and it migrated upward into one of the intrahepatic ducts. Attempts at removing the stent with a trapezoid basket were unsuccessful. A second stent was placed to ensure biliary drainage, which achieved a satisfactory position. Interventional radiology was consulted and the stent that had migrated was successfully removed later the same day. With adequate drainage of the biliary system ensured, the percutaneous drain was removed.

ICU management

The patient’s stay in ICU was complicated. He required vasopressin and noradrenaline infusions for several weeks to ensure adequate perfusion, invasive ventilation to ensure adequate oxygenation and ventilation, and nasogastric feeding and later total parenteral nutrition to ensure nutrition. The patient had serial chest x-rays, which initially showed mild pulmonary oedema and later showed florid parenchymal changes consistent with acute respiratory distress syndrome (ARDS), which occurred as a consequence of the pancreatitis. The chest x-rays also showed bilateral pleural effusions and basal atelectasis. To allow for intubation, the patient was sedated, initially with a propofol infusion, which was stopped due to borderline high and rising triglyceride levels, and later with a midazolam infusion, atracurium, morphine and dexmedetomidine. Extubation failed on two occasions. To ventilate the patient without requiring sedation, a tracheostomy tube was inserted some 10 days after the repeat ERCP had been performed.

Cardiovascular complications

The patient developed several cardiovascular complications. He experienced multiple episodes of fast atrial fibrillation, which were treated with amiodarone. His INR initially rose to 1.87 secondary to septic coagulopathy. Serial troponin measurements showed a gradual rise up to 392. ECG showed new T wave inversion and ST depression in the inferolateral leads. 

Cardiology diagnosed type 2 non-ST elevation myocardial infarction, ie. myocardial infarction secondary to hypoperfusion,¹⁰ but thought the patient was not a candidate for invasive coronary interventions or for uninterrupted dual antiplatelet therapy. They recommended IV metoprolol if the patient developed ventricular tachycardia.

A CT abdomen and pelvis was performed, which revealed several interesting findings: 

• Active inflammation in the common and hepatic bile ducts, but no dilatation and an appropriately placed bile duct stent 
• Generalised intra-abdominal inflammation (anasarca) and an increasing volume of peritoneal fluid 
• A collection that had formed in the previous percutaneous drain site 
• An oedematous pancreas with a moderate volume of peripancreatic fluid, but no definite necrosis 
• Large bilateral pleural effusions and basal atelectasis with possible superimposed consolidation 
• Occlusion of the left portal vein branch with small locules of air, but patency of the splenic vein, the superior mesenteric vein and the splenoportal confluence.

The patient’s antibiotics were stopped gradually (caspofungin, daptomycin and meropenem). Following this, his C-reactive protein and white cell count rose and he had intermittently spiking temperatures. The findings in the bases of the lungs on the CT made respiratory infection the primary candidate. Peripheral and central blood cultures were sterile, the urine was clear, bronchoalveolar lavage was clear, and an echo showed no evidence of infective endocarditis. 

The findings of the CT were discussed at the gastroenterology radiology conference. The consultant radiologist felt that the portal vein occlusion and features of infection were due to an infected portal vein thrombosis. He recommended a prolonged course of antibiotics and anticoagulation.

Portal vein thrombosis with superimposed infection (pylephlebitis) is associated with appendicitis, cholecystitis, pancreatitis and other intra-abdominal infections, and occurs as a result of the infecting organisms draining from the gut into a thrombosed portal vein.^11,12 The infection is usually polymicrobial, but Bacteroides fragilis is the most commonly isolated organism.¹¹ The condition is treated with appropriate antibiotics and anticoagulation, the latter of which is most useful in preventing clot propagation and thus bowel ischaemia or infarction, rather than actually treating the primary condition.¹¹ Mortality rate with portal vein thrombosis infection is 25%.^11,12

The patient was started on a therapeutic dose of enoxaparin (70mg subcutaneously twice daily), and meropenem and daptomycin were recommenced. Microbiology recommended a four week course of antibiotics in total, followed by a CT to assess for interval changes. A week later (some 28 days from initial presentation), creatine kinase levels rose from 200 to 1,000, and so daptomycin, which is associated with myositis and rhabdomyolysis, was discontinued and replaced with linezolid.

The interventional radiology team was consulted about the collection in the right flank. They believed it was unnecessary to perform a diagnostic or therapeutic aspirate, as the patient’s inflammatory markers were downtrending. A repeat CT abdomen and pelvis was performed on nine days later. It revealed several further interesting findings: 

• Persistent occlusion of the left branch of the portal vein
• Probable necrosis along the anterior right and left lobes of the liver secondary to this occlusion
• An increase in the size of the peripancreatic fluid in keeping with a pancreatic pseudocyst, but reduced pancreatic inflammation
• Almost complete resolution of the right flank collection
• Normal bile duct diameters with a patent common bile duct stent.

The patient’s inflammatory markers had reduced to normal by this time, and so microbiology recommended stopping the antibiotics, after 32 days in total. It was determined that the pancreatic pseudocyst and the minor liver necrosis would be managed conservatively. Two days later, the patient was discharged from ICU after a 35 day admission, and was transferred back to his local hospital to complete a six-month course of therapeutic doses of enoxaparin in total, and for post-critical care rehabilitation.

Patient consent

This case illustrates how a relatively simple and commonly-performed procedure can be complicated, in this case resulting in myocardial infarction, acute respiratory distress syndrome, pleural effusions, biliary infection, sepsis, coagulopathy, portal vein thrombosis, intravascular infection and liver infarction.

I have consented patients for an ERCP many times and have explained that the procedure is complicated by pancreatitis in 5-10% of cases, but often patients don’t fully understand how severely ill they can become if they develop it. All of them agreed and signed the form, eager to have the yellow go from their eyes or to relieve their abdominal pain. 

Informed consent is defined as “permission granted in full knowledge of the possible consequences” by Oxford Languages. It is difficult to imagine how one could convey full knowledge of the possible consequences of post-ERCP pancreatitis. In my opinion, it should be a thorough and detailed discussion, held in a private room rather than by the bedside, and led by a senior doctor who has experience in managing the complications.

References

1. Forsmark CE, Vege SS, Wilcox CM. Acute Pancreatitis. N Engl J Med 2016 Nov 17; 375(20):1972-81
2. Fogel EL, Sherman S. ERCP for gallstone pancreatitis. N Engl J Med 2014 Jan 9; 370(2):150-7
3. Pekgöz M. Post-endoscopic retrograde cholangiopancreatography pancreatitis: A systematic review for prevention and treatment. World J Gastroenterol 2019 Aug 7; 25(29):4019-42
4. Tryliskyy Y, Bryce GJ. Post-ERCP pancreatitis: pathophysiology, early identification and risk stratification. Adv Clin Exp Med 2018 Jan; 27(1):149-54
5. Geraci G, Palumbo VD, D’Orazio B, Maffongelli A, Fazzotta S, Lo Monte AI. Rectal diclofenac administration for prevention of post-endoscopic retrograde cholangiopancreatography (ERCP) acute pancreatitis. Randomized prospective study. Clin Ter 2019 Sep-Oct; 170(5):e332-e336
6. Smeets XJ, da Costa DW, Besselink MG et al. Systematic review: periprocedural hydration in the prevention of post-ERCP pancreatitis. Aliment Pharmacol Ther 2016 Sep; 44(6):541-53
7. Sunada F, Morimoto N, Tsukui M, Kurata H. The rendezvous technique involving insertion of a guidewire in a percutaneous transhepatic gallbladder drainage tube for biliary access in a case of difficult biliary cannulation. J Rural Med 2017; 12(1):46–9
8. Tsuchiya T, Itoi T, Sofuni A, Tonozuka R, Mukai S. Endoscopic ultrasonography-guided rendezvous technique. Dig Endosc. 2016 Apr; 28 Suppl 1:96-101
9. Huang L, Yu QS, Zhang Q, Liu JD, Wang Z. The Rendezvous Technique for Common Bile Duct Stones: A Meta-Analysis. Surg Laparosc Endosc Percutan Tech 2015 Dec; 25(6):462-70
10. Sandoval Y, Jaffe AS. Type 2 Myocardial Infarction: JACC Review Topic of the Week. J Am Coll Cardiol. 2019 Apr 16;73(14):1846- 1860. doi: 10.1016/j.jacc.2019.02.018. PMID: 30975302.
11. Wong K, Weisman DS, Patrice KA. Pylephlebitis: a rare complication of an intra-abdominal infection. J Community Hosp Intern Med Perspect. 2013;3(2):10.3402/jchimp.v3i2.20732. Published 2013 Jul 5
12. Plemmons RM, Dooley DP, Longfield RN. Septic thrombophlebitis of the portal vein (pylephlebitis): diagnosis and management in the modern era. Clin Infect Dis. 1995 Nov; 21(5):1114-20