Abstract
Eskelinen M, Ikonen J, Lipponen P. Sex-specific diagnostic scores for acute appendicitis. Scand J Gastroenterol 1994;29:59-66.
The role of clinical and computer-based decisions in the diagnosis of acute appendicitis was studied in connection with the survey of acute abdominal pain by the Research Committee of the World Organization of Gastroenterology (OMGE). One thousand three hundred and thirty-three patients presenting with acute abdominal pain were included in the study. Twenty-two preoperative clinical history variables, 14 clinical signs, and 3 tests were evaluated in a multivariate analysis to find the best combination of independent predictors of acute appendicitis for males and females. Independent predictors of acute appendicitis in males were tenderness, previous abdominal surgery, rebound, rigidity, location of pain at diagnosis, guarding, and body temperature. To sum up the contributions of the most significant diagnostic factors, a diagnostic score (DS) was built. When the male patients with a DS value between -2.00 and -0.48 were considered nondefined (n = 75, follow-up required before deciding to operate), the sensitivity of the computer-aided diagnosis in detecting acute appendicitis in males was 0.95, with a specificity of 0.89 and an efficiency of 0.91. In males whose leucocyte count was available (n = 476), previous abdominal surgery, leucocytosis, location of pain at diagnosis, tenderness, rigidity, rebound, guarding, rectal digital tenderness, and body temperature predicted significantly acute appendicitis. The DS reached a sensitivity of 0.94 (the cut-off level was -1.74), with a specificity of 0.80 and an efficiency of 0.84. When the male patients with a DS value between -1.74 and -0.14 were considered nondefined (n = 67, follow-up required before deciding to operate), the sensitivity of the computer-aided diagnosis in detecting acute appendicitis in males was 0.93, with a specificity of 0.93 and an efficiency of 0.93. The most important independent predictors of acute appendicitis in females were tenderness, rigidity, guarding, location of pain at diagnosis, and renal tenderness. When the patients with a DS value between -2.03 and -0.49 were considered nondefined (n = 123, follow-up required before deciding to operate), the sensitivity of the computer-aided diagnosis in detecting acute appendicitis in females was 0.93, with a specificity of 0.92 and an efficiency of 0.92. In females whose leucocyte count was available (n = 575) the most important independent predictors of acute appendicitis were tenderness, rigidity, guarding, leucocytosis, location of pain at diagnosis, and renal tenderness. When the patients with a DS value between -2.33 and -0.41 were considered nondefined (n = 77, follow-up required before deciding to operate), the sensitivity of the computer-aided diagnosis in detecting acute appendicitis in females was 0.93, with a specificity of 0.91 and an efficiency of 0.92. The results suggest that the use of computer-aided decisions improves diagnostic sensitivity, specificity, efficiency, and LR+ in detecting acute appendicitis in females with acute abdominal pain.