Objective: Audit of the protocol for surgical treatment of patients with acute severe and necrotising pancreatitis (ANP).
Design: Prospective open study.
Settings: University hospital, Finland.
Patients: 33 patients treated for severe (Ranson score 3 or more) and necrotising (as judged on computed tomograms (CT)) pancreatitis between 1992-1993.
Protocol: Indications for antibiotic treatment (n = 25 patients) were: fulminant multiorgan disease; recurrent continual parallel increase in temperature, white cell count (WCC) and C-reactive protein concentration; or the presence of bacteria on Gram stain of a percutaneous fine needle aspiration smear of necrosis. Three of the 25 responded to antibiotics. They and eight others with ANP but without these indications were treated conservatively. Twenty-two patients underwent repeated necrosectomy by laparostomy.
Main outcome measures: Diagnosis of pancreatic infection, morbidity and mortality
Results: Of the 22 patients operated on 17 had contaminated necrosis at operation, and this had been predicted by the increasing inflammatory variable and the presence of bacteria in the Gram stain. Five patients operated on died (23%), four of the five having been operated on for fulminant multiorgan disease (80%). Recurrent sepsis developed in five patients, pancreatic fistulas in two, and there were no pseudocysts. Gastrointestinal fistulas developed in 12 patients, but not after we had changed the technique of wound packing. All 11 patients treated conservatively survived.
Conclusion: A third of patients with ANP can be selected for safe non-operative treatment. Infected ANP can be treated by repeated necrosectomy by laparostomy with low mortality (6%). Early fulminant multiorgan disease should not be treated with laparostomy.