Internal pancreatic fistulas: proposal of a management algorithm based on a case series analysis

Internal pancreatic fistulas (IPF) are an uncommon but well-recognized complication of chronic pancreatitis (CP) that are associated with significant morbidity and mortality. Because of their low incidence, management is still controversial. The aims of this study are to report the 8-year experience with IPF management in a Brazil University-affiliated hospital and to propose a management algorithm.

A centralized diagnostic index was used to retrospectively identify all patients with IPF admitted to a teaching hospital from 1995 to 2003. The patient's medical records were reviewed for clinical features, diagnostic work-up, treatment strategies, response to therapy, and the length of hospital stay. All patients underwent contrast-enhanced computed tomography of the abdomen and endoscopic retrograde cholangiopancreatography, to guide the therapeutic modality to be offered. Conservative therapy included withholding of oral feedings in conjunction with total parenteral nutrition, octreotide subcutaneously, and multiple paracentesis or thoracentesis. Interventional therapy was either endoscopic or surgical.

IPF was identified in 11 (7.3%) of 150 patients with CP. They ranged in age from 24 to 47 years (mean 36.1), with a male to female ratio of 10:1. All patients had underlying alcoholic CP. The presentation was pancreatic ascites in 9 patients and pleural effusion in 2 cases. Five patients were undergoing the conservative treatment, all presenting main pancreatic duct (MPD) dilatation; endoscopic placement of transpapillary pancreatic duct stent was performed in 4 patients who presented partial MPD stricture or disruption; surgical therapy was performed in 2 patients exhibiting complete MPD obstruction or disruption. Stents were removed 3 to 6 weeks after initial placement. IPF resolved in 10 of 11 patients (90.9%) within 6 weeks. The resolution of IPF was faster (13 +/- 5 vs. 25 +/- 13 days, P < 0.01) and the length of hospital stay was significantly shorter (17.2 +/- 5.6 vs. 31.2 +/- 4.4 days, P < 0.01) in patients subject to interventional treatment compared with those treated conservatively. There was 1 death due to sepsis in a patient managed conservatively; no death was recorded in the interventional therapy group. There was no recurrence of IPF at a mean follow-up of 38 months.

Our results suggest that interventional therapy should be considered the best approach for the management of IPF in patients presenting MPD disruption or obstruction. Conservative therapy must be reserved for those showing MPD dilatation without ductal disruption or stricture. Early interventional therapy reduced hospital stay and convalescence, which likely resulted in lower healthcare overall costs.