A randomized non-inferiority trial investigating antibiotic adjustments based on blood culture in acute cholangitis.

Objective: This study evaluated the adequacy of using blood cultures alone for antibiotic therapy in mild-to-moderate acute cholangitis after adequate biliary drainage.

Methods: A prospective, multi-centre, non-inferiority, randomized trial was conducted from August 2015 to September 2023 across 12 tertiary hospitals in South Korea. Patients were randomly assigned 1:1 to groups. The control group was administered antibiotics based on blood and bile culture results, whereas the experimental group was treated based on blood culture results alone. The primary outcome was the rate of organ failure, assessed by the presence of shock, acute kidney injury, altered mental status, and acute respiratory distress. The secondary outcomes were all-cause mortality, frequency of re-interventions, antibiotic usage duration, and hospital stay. The non-inferiority margin was 10%.

Results: Patients in the control group (n = 215) and experimental group (n = 213) were analysed. Organ failure occurred in 28 (13.0%) in the control and 27 (12.7%) in the experimental group (difference: -0.350% [95% CI, -6.690% to 5.990%], p > 0.999). Mortality rates were 3.3% in the control (seven patients) and 2.3% in the experimental group (five patients) (-0.908% [-4.033% to 2.216%], p 0.782). The reintervention rates were 31 (14.9%) in the control and 26 (12.2%) in the experimental groups (-2.677% [-9.155% to 3.801%], p 0.504). The antibiotic therapy duration was different (-1.500 days [-2.840 to -0.160], p 0.037) between the control (12.8 ± 8.0 days) and the experimental groups (11.3 ± 6.5 days), and the hospital stay was different (-1.900 days [-0.380 to 0.000], p 0.046) between the control (14.1 ± 11.0 days) and the experimental groups (12.2 ± 9.0 days).

Conclusions: This study confirms that relying on blood culture results alone for managing mild-to-moderate cholangitis is reasonable and may help reduce the unnecessary use of medical resources.