Developing a Simple Score for Diagnosis of Acute Cholecystitis at the Emergency Department

Abstract

We aim to develop a diagnostic score for acute cholecystitis that integrates symptoms, physical examinations, and laboratory data to help clinicians for timely detection and early treatment of this disease. We retrospectively collected data from our database from 2010 to 2020. Patients with acute abdominal pain who underwent an ultrasound or computed tomography (CT) scan at the emergency department (ED) were included. Cases were identified by pathological, CT, or ultrasound reports. Non-cases were those who did not fulfill any of these criteria. Multivariable regression analysis was conducted to identify predictors of acute cholecystitis. The model included 244 patients suspected of acute cholecystitis. Eighty-six patients (35.2%) were acute cholecystitis confirmed cases. Five final predictors remained within the reduced logistic model: age < 60, nausea and/or vomiting, right upper quadrant pain, positive Murphy’s sign, and AST ≥ two times upper limit of normal. A practical score diagnostic performance was AuROC 0.74 (95% CI, 0.67−0.81). Patients were categorized with a high probability of acute cholecystitis at score points of 9−12 with a positive likelihood ratio of 3.79 (95% CI, 1.68−8.94). ED Chole Score from these five predictors may aid in diagnosing acute cholecystitis at ED. Patients with an ED Chole Score >8 should be further investigated.

Keywords: acute abdominal pain; acute cholecystitis; emergency department; prediction.

Figure 1

Study flowchart.

Figure 2

Receiver operating characteristic curve in the diagnostic prediction of acute cholecystitis of ED Chole Score within the development cohort.

Figure 3

Calibration plots visualizing agreement between the derived score (ED Chole Score) and the actual (observed) risk. The blue line depicts the predict risk of acute cholecystitis confirmed cases. The red circle defines the observed risk of acute cholecystitis confirmed cases.