Abstract
INTRODUCTION: There is evidence that postponing surgery in critically ill patients with severe acute pancreatitis (SAP) leads to improved survival, but previous reports included patients with both sterile and infected pancreatic necrosis who were operated on for various indications and with different degrees of organ dysfunction at the moment of surgery, which might be an important bias. The objective of this study is to analyze the impact of timing of surgery and perioperative factors (severity of organ dysfunction and microbiological status of the necrosis) on mortality in intensive care unit (ICU) patients undergoing surgery for SAP. METHODS: We retrospectively (January 1994 to March 2003) analyzed patients admitted to the ICU with SAP. Of 124 patients, 56 were treated surgically; these are the subject of this analysis. We recorded demographic characteristics and predictors of mortality at admission, timing of and indications for surgery, and outcome. We also studied the microbiological status of the necrosis and organ dysfunction at the moment of surgery. RESULTS: Patients' characteristics were comparable in patients undergoing early and late surgery, and there was a trend toward a higher mortality in patients who underwent early surgery (55% versus 29%, P = 0.06). In univariate analysis, patients who died were older, had higher organ dysfunction scores at the day of surgery, and had sterile necrosis more often; there was a trend toward earlier surgery in these patients. Logistic regression analysis showed that only age, organ dysfunction at the moment of surgery, and the presence of sterile necrosis were independent predictors of mortality. CONCLUSIONS: In this cohort of critically ill patients operated on for SAP, there was a trend toward higher mortality in patients operated on early in the course of the disease, but in multivariate analysis, only greater age, severity of organ dysfunction at the moment of surgery, and the presence of sterile necrosis, but not the timing of the surgical intervention, were independently associated with an increased risk for mortality.