Are traditional prognostic criteria useful in pancreatic abscess?

Pancreatic abscess remains a potentially lethal disease. Efforts to relate outcome to the severity of associated pancreatitis or the type of surgical drainage employed have yielded conflicting results. This study was designed to test the validity of traditional prognostic criteria in the clinical setting of pancreatic abscess and to determine whether the technique of surgical drainage employed correlated with survival. The records of 40 consecutive patients with pancreatic abscess were reviewed. In each case the diagnosis was confirmed by operation. Prognostic factors analyzed included number of Ranson criteria, etiology, type, and number of microorganisms isolated, extent of abscess, time to diagnosis and operation, and technique of surgical drainage. Of the 11 Ranson criteria evaluated, only an elevation in blood urea nitrogen >5 mg/dl correlated with decreased survival (p < 0.001). Polymicrobial abscesses (three or more organisms) resulted in a higher mortality than abscesses wherefewer than three organisms were isolated (45.4 vs 13.8%; p < 0.05). Intraperitoneal extension of the abscess was associated with an increased mortality rate compared to those confined to the retroperitoneum (57.1 vs 15.2%; p < 0.01). In patients requiring unplanned reexploration, mortality was significantly increased (42.9 vs 11.5%; p < 0.05). The technique of surgical drainage employed (open versus closed) did not influence overall mortality (23.5 vs 21.7%; p = NS). Extent of disease at operation, polymicrobial abscess, reexploration for persistent or recurrent disease, and deterioration in renal function were all predictive of increased mortality in cases of pancreatic abscess. Traditional Ranson criteria were not useful predictors of survival. No correlation existed between type of surgical drainage employed and subsequent outcome.