Abstracts
1997 Digestive Disease Week

Debridement and closing packing for the treatment of necrotizing pancreatitis.

C Fernandez-del Castillo, DW Rattner, A Mostafavi, D McGrath, AL Warshaw. Department of Surgery, Massachusetts General Hospital and Harvard Medical School, Boston, MA.

The characteristics and outcomes of 64 consecutive patients with necrotizing pancreatitis requiring surgery between Feb 1990 and Nov 1996 were reviewed. Patients were treated with necrosectomy followed by "closed packing" of the cavity with the use of stuffed penrose and closed suction drains. The mean APACHE II score was 9, and 31% of patients had organ failure. The indication for surgery was persistent symptoms in 39%, sepsis syndrome in 25%, bacteria by needle aspiration in 25%, and pancreatic abscess by x-ray in 11%. Patients were operated on a median of 31 days after diagnosis; 56% of patients had infected necrosis with predominance of Gram positive bacteria.

Mortality was 6.2%, and no different between infected or sterile necrosis. Eleven patients required reoperation (two for bleeding and 9 for progressive necrosis); 13 needed percutaneous drainage for residual or recurrent collections. A single operation sufficed in 66%. Enteric fistulae complicated 10 patients (16%) and 34/64 developed a pancreatic fistula, of which only three ultimately required surgical closure. Long-term diabetes mellitus and exocrine insufficiency were documented in 9% and 25% of patients, respectively. Mean hospital stay was 41 days, and the interval until return to regular activities was 147 days.

Patients operated within the first six weeks of pancreatitis (n = 37, range 4-40 days, median 21) were compared to those operated after six weeks (n = 27, range 52-300 days, median 75). APACHE II scores (11.4 vs. 5.7, p < 0.001), presence of organ failure (43% vs. 15%, p < 0.05), indications for surgery (sepsis syndrome and infected necrosis vs lower-grade persistent symptoms), need for ICU care and TPN were all significantly different. However, mortality, reintervention and complication rates were similar, as was the time required for return to regular activities.

CONCLUSIONS: Outcomes for treatment of pancreatic necrosis have improved markedly: these based on debridement and closed packing are the best yet reported. Intervention is best delayed until demarcation of necrosisis is complete, but futher delay confers no additional advantage.