Background: Computed tomography (CT) is used increasingly to evaluate suspected cases of acute appendicitis (AA) in the emergency department (ED). This prospective study was performed to test the hypothesis that the evaluation of AA by CT in the ED remains suboptimal and that erroneous interpretation diminishes its utility.
Methods: Consecutive patients 18 years of age or older were enrolled prospectively if AA was among the first three differential diagnoses listed in the record of patients undergoing evaluation of abdominal pain in the ED. Imaging of the abdomen and pelvis was obtained at the discretion of the ED staff or consultant surgeon. Initial CT interpretation was by a radiology resident or fellow along with the surgical staff, but final review by an attending radiologist occurred later. Age, gender, presenting symptoms, white blood cell (WBC) count, final CT results, and final pathology (for patients undergoing operation) were recorded. X +/- SEM, p < 0.05 by chi(2), ANOVA, or MANOVA was used for statistical analysis as appropriate.
Results: A CT scan was performed in 104 patients (83% of those meeting entry criteria), 35 of whom were male (mean age, 37 +/- 2 years) and 69 of whom were female (mean age, 39 +/- 3 years). Thirty-five patients had pathologically proved appendicitis, 28 of whom were diagnosed prospectively by CT. There were seven false-negative scans. Sensitivity, specificity, and positive predictive value for the initial CT reading were 80%, 91%, and 82%, respectively. Gender (p < 0.03), WBC count (p < 0.0002), and a positive initial CT reading (p < 0.0001) correlated with operative management. However, although final CT interpretation did correlate with pathologic confirmation of AA (p < 0.0001), initial CT interpretation did not correlate with the presence of AA (p = 0.52).
Conclusion: The ability of CT to predict AA is dependent on the interpretative skill of the individual interpreting the images. Widespread use of CT in the evaluation of patients for AA should be implemented with caution until institution-specific protocols are validated.