Résumé
Objectives: to assess the relevance of thoraco-abdominal-pelvic computed tomography (TAP-CT) in a context of sepsis of unclear origin, and to identify predictive factors for normal TAP-CT in the aim of reducing unnecessary exams. Methods: we consecutively reviewed CT-scan reports of 1204 adult patients who underwent TAP-CT in a context of sepsis of unclear origin over a 12-month period. Patients came from emergency department or general wards. CT-scan reports were categorized according to the presence and location of an infectious focus. Noninfectious findings that might be responsible for a septic syndrome were also noted. A multivariate logistic regression was performed to identify predictive factors of a normal TAP-CT.Results: of 634 consecutive TAP-CTs, 82% had been referred from the emergency department and 18% from general wards. 323 (51%) CT-scans identified an infectious focus. Most of the foci were located in the chest (47%), abdomen (25%), and genitourinary tract (23%). In 121 (19%) exams, a significant noninfectious pathological finding was noted. A CRP level <161.20 mg/L and the provenance from rheumatological wards were predictive of a normal CT-scan. However multivariate analyses could not provide any accurate model to predict a negative TAP-CT. Conclusion: in a context of sepsis of unclear origin, patients with a previously known rheumatological disorder and a low CRP-level (<161.20 mg/L) are more likely to have a normal TAP-CT. However, the use of CT-scan in that context can’t be avoided with an acceptable individual risk.