Management of postoperative pancreatic fistula after pancreaticoduodenectomy

A postoperative pancreatic fistula (POPF) is the main complication after cephalic pancreaticoduodenectomy (CPD). Unlike its prevention, the curative management of POPFs has long been poorly codified. This review seeks best practices for managing POPFs after CPD. The diagnosis of a POPF is based on two signs: (i) an amylase level in drained fluid more than 3 times the upper limit of the blood amylase level; and (ii) an abnormal clinical course. In the standardised definition of the International Study Group of Pancreatic Surgery, a purely biochemical fistula is no longer counted as a POPF and is treated by gradual withdrawal of the drain over at most 3 weeks. POPF risk can be scored using pre- and intraoperative clinical criteria, many of which are related to the quality of the pancreatic parenchyma and are common to several scoring systems. The prognostic value of these scores can be improved as early as Day 1 by amylase assays in blood and drained fluid. Recent literature, including in particular the Dutch randomised trial PORSCH, argues for early systematic detection of a POPF (periodic assays, CT-scan with injection indicated on standardised clinical and biological criteria plus an opinion from a pancreatic surgeon), for rapid minimally invasive treatment of collections (percutaneous drainage, antibiotic therapy indicated on standardised criteria) to forestall severe septic and/or haemorrhagic forms, and for the swift withdrawal of abdominal drains when the risk of a POPF is theoretically low and evolution is favourable. A haemorrhage occurring after Day 1 always requires CT angiography with arterial time and monitoring in intensive care. Minimally invasive treatment of a POPF (radiologically-guided percutaneous drainage or, more rarely, endoscopic drainage, arterial embolisation) should be preferred as first-line treatment. The addition of artificial nutrition (enteral via a nasogastric or nasojejunal tube, or parenteral) is most often useful. If minimally invasive treatment fails, then reintervention is indicated, preserving the remaining pancreas if possible, but the expected mortality is higher.