Résumé
OBJECTIVE. The purpose of this study was to assess whether the availability of clinicobiologic findings would affect the diagnostic performance of CT of elderly emergency department patients with nontraumatic acute abdominal pain.
MATERIALS AND METHODS. The cases of 333 consecutively registered patients 75 years old or older presenting to the emergency department with acute abdominal pain and who underwent CT were retrospectively reviewed by two radiologists blinded or not to the patient's clinicobiologic results. Diagnostic accuracy was calculated according to the level of correctly classified cases in both the entire cohort and a surgical subgroup and was compared between readings performed with and without knowledge of the clinicobiologic findings. Agreement between each reading and the reference diagnosis and interobserver agreement were assessed with kappa statistics.
RESULTS. In both the entire cohort (87.4% vs 85.3%, p = 0.07) and the surgical group (94% vs 91%, p = 0.15), there was no significant difference in CT accuracy between diagnoses made when the radiologist was aware and those made when the radiologist was not aware of the clinicobiologic findings. Agreement between the CT diagnosis and the final diagnosis was excellent whether or not the radiologist was aware of the clinicobiologic findings.
CONCLUSION. In the care of elderly patients, CT is accurate for diagnosing the cause of acute abdominal pain, particularly when it is of surgical origin, regardless of the availability of clinical and biologic findings. Thus CT interpretation should not be delayed until complete clinicobiologic data are available, and the images should be quickly transmitted to the emergency physician so that appropriate therapy can be begun.