Influence of ultrasound on clinical decision making in acute appendicitis: a prospective study.

Abstract

To assess the efficacy of ultrasound (US) as part of an algorithm to establish the indication for laparotomy in patients with suspected acute appendicitis. Prospective investigation. University department of surgery, Germany. 669 unselected patients admitted with suspected acute appendicitis. Clinicopathological and procedural diagnoses of the algorithm were evaluated by correlating clinical and US findings with the results of laparotomy in 171 patients of whom 143 had acute appendicitis (prevalence 21%), and clinical as well as follow up data in the remainder. The major clinicopathological variables were accuracy and positive predictive value; the rate of negative laparotomies and that of bad diagnostic errors served as the main procedural variables. The overall sensitivity, specificity, and accuracy of the clinical diagnosis were 0.503, 0.950, and 0.855, respectively (positive predictive value: PPV 0.734, negative predictive value: NPV 0.875), those of ultrasound: 0.797, 0.967, and 0.931 (PPV 0.870, NPV 0.946); and 0.853, 0.927, and 0.940 at the end of the algorithm (PPV 0.762, NPV 0.958). However, the algorithm would have resulted in a significant increase in the rate of unnecessary laparotomies (from 13% to 16%). A revised clinical algorithm gave an overall diagnostic accuracy of 0.940 (p < 0.001) together with a low rate of negative laparotomies (11%, p < 0.01) and a significantly reduced number of diagnostic errors (from 71 to 21, p < 0.001). Ultrasonography enabled us to diagnose acute appendicitis in more patients more often and more quickly than clinical evaluation alone, suggesting that US may produce a better outcome. The revised clinical algorithm may be helpful in the study of US in patients with suspected acute appendicitis in prospective randomised controlled clinical trials.