Predictive factors associated with bile culture positivity: a model development and diagnostic test accuracy study

The role of antibiotics in the preoperative and perioperative management of patients with benign biliary pathology is still not clear. This study aims to develop a predictive model for identifying patients with benign biliary disease who are at low risk of positive bile cultures, optimizing antibiotic administration.

Prospective cohort study conducted between April and October 2024 in a single center from a hospital network with two locations. Consecutive adult patients (aged ≥ 18 years) diagnosed with benign biliary disease who underwent cholecystectomy with intraoperative bile culture sampling. The primary validation outcome was positive bile cultures; predictors included demographic, clinical, and paraclinical variables.

Overall, 703 patients were included in the study. The incidence of positive cultures was 32.1%. The rate of major complications (Clavien-Dindo ≥ 3), comorbidities, age, and higher cholecystitis severity was significantly higher in patients with positive bile cultures. The multivariable prediction model for positive cultures included age, ERCP, and time from admission to procedure, showing fair discrimination (c-statistic 0.75, 95% CI 0.70-0.80) with no evidence of poor calibration. Patients classified as low risk for positive bile cultures had a sensitivity of 92% and a negative likelihood ratio of 0.21, indicating that classification in the low-risk group effectively rules out the possibility of a positive bile culture.

This study found our predictive model useful in assessing bile culture positivity during cholecystectomy to rule out the need for antibiotic initiation in low-risk patients, both preoperatively while awaiting surgery (in cases of acute cholecystitis) and perioperatively (prophylaxis).