External validation of the bedside score for the diagnosis of acute cholecystitis

Abstract

Objective: Acute cholecystitis usually presents with right upper quadrant (RUQ) abdominal pain. However, there are other conditions with similar findings which make the diagnosis difficult. The objective of this study is to prospectively validate the performance of the bedside score for the diagnosis of cholecystitis in patients presenting to the emergency department (ED) with possible acute cholecystitis.

Study design: We performed a prospective observational study of a convenience sample of patients with RUQ pain admitted to the ED of three academic hospitals. Symptoms (post prandial symptoms), physical signs (RUQ tenderness, murphy's sign) and ultrasound findings (Murphy's sign, gallstone, and gallbladder thickening) were scoring system items. The final diagnosis of cholecystitis was confirmed with a surgical pathology and/or discharge diagnosis of the patient in a 30-day follow-up. The treating physicians' clinical gestalt of acute cholecystitis was also assessed by 5-point Likert scale.

Results: One hundred thirty patients were followed up and were included in the analysis. 42 patients (32 %) had cholecystitis. The bedside clinical score of less than 4 had a sensitivity of 100 % (CI95 %: 91.60 %-100 %), negative predictive value (NPV) of 100 % (CI 95 %: 41.35 %-63 %), and negative likelihood ratio (-LR) of 0. Score of 6 and above had a specificity of 90.91 % (CI 95 %: 82.87 %-95.99 %), positive predictive value (PPV) of 83.67 % (CI 95 %: 72.55 %-90.86 %), and positive likelihood ratio (+LR) of 10.74 (CI95 %: 5.54-20.83). Physicians' clinical gestalt at the scale of 4 and 5 showed a specificity of 95.45 % (CI 95 %: 88.77 %-98.75 %), PPV of 90.91 % (CI 95 %: 79.29 %-96.31 %), and +LR of 20.95 (CI95 %: 8.02-54.71). At the same time at the scale of 1 and 2, the sensitivity was 95.24 % (CI 95 %: 83.84 %-99.42 %), NPV was 97.22 % (CI 95 %: 90.01 %-99.27 %), and the -LR was 0.06 (CI 95 %: 0.02-0.423). The area under the curve of bedside clinical score was not significantly higher than clinical gestalt (97.58 (CI 95 %: 95.31-99.85) vs. 95.37 (CI 95: 99.24-100))(p-value = 0.35).

Conclusion: This study showed while the bedside score would be helpful to rule out and rule in acute cholecystitis, physicians' gestalt had similar diagnostic performance.

Keywords: Acute cholecystitis; Bedside score; Diagnosis; Emergency department.