Background and aim: At times, the diagnosis of acute appendicitis may be difficult. However, for minimum morbidity to be obtained, early and accurate diagnosis is essential. This study aimed to validate a scoring system proposed by Eskelinen et al. as an aid in making the diagnosis of appendicitis.
Patients and methods: The prospectively documented data of a consecutive series of 2,359 patients admitted for suspicion of appendicitis were used for validation. Accuracy and positive predictive value were defined as the main overall performance parameters, as was the rate of unnecessary operations to assess changes of patient management. Overall performance was assessed by receiver-operator characteristics (ROC) analysis.
Results: Of 2,359 patients, 662 were proven to have acute appendicitis (prevalence of 28%). The overall sensitivity, specificity, positive and negative predictive value, and accuracy of the score were 0.79, 0.85, 0.68, 0.91 and 0.835 at a cut-off value of 55. Calibration of the score's cut-off value to 57 yielded more favourable results (0.72, 0.91, 0.76, 0.9 and 0.86), and the rate of unnecessary operations declined from 26.6% to 15.4% ( P<0.05, chi2). ROC analysis revealed an area index of 0.91.
Conclusion: The Eskelinen score delivered acceptable clinical results only after calibration to a cut-off value of 57. The data from this study suggest the investigation of whether a calibrated score might be particularly instrumental in the pre-admission evaluation of the patient in whom appendicitis is suspected.