Résumé
OBJECTIVE: To compare the accuracy of a surgeon's clinical diagnosis of acute appendicitis with Alvarado's predictive model and C-reactive protein (CRP) measurements. METHOD: The records of 231 adult patients between 14-75 years admitted to the hospital with suspected appendicitis from August 1999 to November 2001 were studied prospectively. Serum CRP measurements (217 patients) and Alvarado scores (231 patients) were performed before operations but were not taken into account prior to the decision to perform a laparotomy to compare the surgeon's clinical diagnosis. RESULTS: Based on the surgeon's clinical diagnosis, 193 patients underwent surgery, and 38 patients were observed. Histopathologic findings found acute appendicitis, confirming the surgeon's clinical impression, in 178 patients (positive predictive value = 92%) and normal appendix in 15 patients. Of the observed patients, 8 subsequently underwent operation for appendicitis (negative predictive value = 79%). Compared with the surgeon's clinical diagnosis (sensitivity 96% and specificity 67%), diagnosis based on an Alvarado score of > or = 7 had a lower sensitivity (79%) and that based on CRP of > 10 mg/l a much lower sensitivity (62%) and lower specificity (56%). Overall accuracy of these three diagnostic modalities were 90 per cent, 72 per cent and 61 per cent, respectively. However, median serum CRP value increased from 5 mg/l (range 3-188 mg/l) in patients with normal appendix, to 14 mg/l (range 3-222 mg/l) in patients with non-perforated appendicitis and 65 mg/l (range 3-213 mg/l) in patients with perforated or gangrenous appendicitis. CONCLUSION: The clinical assessment in diagnosing appendicitis by an experienced surgeon remains reliable and superior to either Alvarado score or CRP measurement. Nevertheless, Alvarado score and serum CRP measurements may be of value to the inexperienced surgeon, and a high Alvarado score and serum CRP should not be ignored.