Performance of diagnostic tools for acute cholangitis in patients with suspected biliary obstruction

Background: Acute cholangitis is an infection requiring endoscopic retrograde cholangiopancreatography (ERCP) and antibiotics. Several diagnostic tools help to diagnose cholangitis. Because diagnostic performance of these tools has not been studied and might therefore impose unnecessary ERCPs, we aimed to evaluate this. Methods: We established a nationwide prospective cohort of patients with suspected biliary obstruction who underwent an ERCP. We assessed the diagnostic performance of Tokyo Guidelines (TG18), Dutch Pancreatitis Study Group (DPSG) criteria, and Charcot triad relative to real-world cholangitis as the reference standard. Results: 127 (16%) of 794 patients were diagnosed with real-world cholangitis. Using the TG18, DPSG, and Charcot triad, 345 (44%), 55 (7%), and 66 (8%) patients were defined as having cholangitis, respectively. Sensitivity for TG18 was 82% (95% CI 74-88) and specificity 60% (95% CI 56-63). The sensitivity for DPSG and Charcot was 42% (95% CI 33-51) and 46% (95% CI 38-56), specificity was 99.7% (95% CI 99-100) and 99% (95% CI 98-100), respectively. Conclusions: TG18 criteria incorrectly diagnoses four out of ten patients with real-world cholangitis, while DPSG and Charcot criteria failed to diagnose more than half of patients. As the cholangitis diagnosis has many consequences for treatment, there is a need for more accurate diagnostic tools or work-up towards ERCP.