A comparative study to validate the use of ultrasonography and computed tomography in patients with post-operative intra-abdominal sepsis

Purpose: To validate abdominal Ultrasonography and helical computed tomography in detecting causes for sepsis in patients after abdominal surgery and to determine improved criteria for its use. Materials and methods: Eighty-five consecutive surgical patients primarily operated for non-infectious disease were included in this prospective study. Forty-one patients were admitted to the intensive care unit. All patients were suspected of an intra-abdominal sepsis after abdominal surgery. Both ultrasonography (US) and helical abdominal computed tomography (CT) were performed to investigate the origin of an intra-abdominal sepsis. The images of both US and CT were interpreted on a four-point scale by different radiologists or residents in radiology, the investigators were blinded of each other's test. Interpretations of US and CT were compared with a reference standard which was defined by the result of diagnostic aspiration of suspected fluid collections, (re)laparotomy, clinical course or the opinion of an independent panel. Likelihood ratios and post-test probabilities were calculated and interobserver agreement was determined Using K statistics. Results: The overall prevalence of an abdominal infection was 0.49. The likelihood ratio (LR) of a positive test-result for US was 1.33 (95% CI: 0.8-2.5) and for CT scan 2.53 (95% CI: 1.4-5.0); corresponding post-test probabilities for US 0.57 (95% CI: 0.42-0.70) and for CT 0.71 (95% CI: 0.57-0.83). The LR of a negative test-result was, respectively, 0.60 (95% CI: 0.3-1.3) and 0.18 (95% CI: 0.06-0.5); corresponding post-test probabilities for US 0.37 (95% CI: 0.20-0.57) and for CT 0.15 (95% Cl: 0.06-0.32) were calculated. Conclusion: Computed tomography can be used as the imaging modality of choice in patients suspected of intra-abdominal sepsis after abdominal surgery. Because of the low discriminatory power ultrasonography should not be performed as initial diagnostic test. (c) 2004 Elsevier Ireland Ltd. All rights reserved.