Factors associated with septic shock and mortality in generalized peritonitis: comparison between community-acquired and postoperative peritonitis.

INTRODUCTION: The risk factors associated with poor outcome in generalized peritonitis are still debated. Our aim was to analyze clinical and bacteriological factors associated with the occurrence of shock and mortality in patients with secondary generalized peritonitis. METHODS: This was a prospective observational study involving 180 consecutive patients with secondary generalized peritonitis (community-acquired and postoperative) at a single center. We recorded peri-operative occurrence of septic shock and 30-day survival rate and analyzed their associations with patients characteristics (age, gender, SAPS II, liver cirrhosis, cancer, origin of peritonitis), and microbiological/mycological data (peritoneal fluid, blood cultures). RESULTS: Frequency of septic shock was 41% and overall mortality rate was 19% in our cohort. Patients with septic shock had a mortality rate of 35%, versus 8% for patients without shock. Septic shock occurrence and mortality rate were not different between community-acquired and postoperative peritonitis. Age over 65, two or more microorganisms, or anaerobes in peritoneal fluid culture were independent risk factors of shock. In the subgroup of peritonitis with septic shock, biliary origin was independently associated with increased mortality. In addition, intraperitoneal yeasts and Enterococci were associated with septic shock in community-acquired peritonitis. Yeasts in the peritoneal fluid of postoperative peritonitis were also an independent risk factor of death in patients with septic shock. CONCLUSIONS: Unlike previous studies, we observed no difference in incidence of shock and prognosis between community-acquired and postoperative peritonitis. Our findings support the deleterious role of Enterococcus species and yeasts in peritoneal fluid, reinforcing the need for prospective trials evaluating systematic treatment against these microorganisms in patients with secondary peritonitis.

Adrenal response in patients with septic shock of abdominal origin: relationship to survival.

OBJECTIVE: To evaluate the prognostic value of adrenocortical response to corticotropin in septic shock patients operated on exclusively for an intra-abdominal source of infection. DESIGN AND SETTING: Prospective, observational, single-center study in a surgical intensive care unit of a university hospital PATIENTS: 118 consecutive septic shock patients undergoing laparotomy or drainage for intra-abdominal infection. MEASUREMENTS AND RESULTS: Baseline cortisol (t (0)) and cortisol response to corticotropin test (Delta) were measured during the first 24 h following onset of shock. The relationship between adrenal function test results and survival was analyzed as well as the effect of etomidate anesthesia. Cortisol plasma level at t (0) was higher in nonsurvivors than in survivors (33 +/- 23 vs. 25 +/- 14 microg/dl), but the response to corticotropin test did not differ between these two subgroups. ROC analysis showed threshold values for t (0) (32 microg/dl) and Delta (8 microg/dl) that best discriminated survivors from nonsurvivors in our population. We observed no difference in survival at the end of hospital stay using log rank test when patients were separated according to t (0), Delta, or both. In addition, adrenal function tests and survival did not differ in patients who received etomidate anesthesia (n = 69) during the surgical treatment of their abdominal sepsis. CONCLUSIONS: In this cohort of patients with abdominal septic shock baseline cortisol level and the response to corticotropin test did not discriminate survivors from nonsurvivors. No deleterious impact of etomidate anesthesia on adrenal function tests and survival was observed in these patients.

Inflammatory cytokines, C reactive protein, and procalcitonin as early predictors of necrosis infection in acute necrotizing pancreatitis.

BACKGROUND: Infection of necrosis is a major risk factor in patients with acute pancreatitis. Systematic use of broad spectrum antibiotics has been recommended in these patients but may induce serious side effects. To better target patients in whom antibiotic prophylaxis could be beneficial, we evaluated whether early serum profiles of interleukin 6 (IL-6), tumor necrosis factor (TNF-alpha, C reactive protein (CRP) and procalcitonin (PCT) help to discriminate between patients who eventually develop infection of necrosis and those who do not. METHODS: Forty-eight patients with acute pancreatitis and a computed tomography (CT) severity index score of more than 3 were prospectively screened. They were then separated into infected and non-infected groups according to the occurrence of infected pancreatic necrosis. The severity of illness was assessed with Ranson's classification and Simplified Acute Physiologic Score II. Serum levels of IL-6, TNF-alpha, CRP, and PCT were measured during the first 3 days after admission. CT-guided fine needle aspiration of pancreatic necrosis was performed to prove infection when sepsis of abdominal origin was suspected. Using the methodology of receiver operating curves, we determined the presence of a threshold for markers that would be predictive of the development of infected necrosis. RESULTS: PCT and IL-6 were higher in the serum of patients who eventually developed infection of necrosis (P < 0.003 and < 0.04, respectively). No difference was noted between the 2 groups for TNF- alpha and CRP. The combination of IL-6 < 400 pg/l and PCT < 2 ng/L best identified patients who were not at risk for necrosis infection. The negative predictive value for these thresholds was 91%, whereas sensitivity and specificity were 75% and 84%. CONCLUSIONS: PCT and IL-6 serum levels were elevated very early in patients who eventually developed necrosis infection. A combination of PCT and IL-6 thresholds could be helpful in identifying a subgroup of patients in whom antibiotic prophylaxis is likely to be ineffective.