Risk Prediction Model for Vre Infection in Liver Transplant Recipients

Vancomycin-resistant Enterococcus spp. (VRE) are multidrug-resistant microorganisms frequently isolated in liver transplant (LT) patients. In this population, VRE infection is potentially associated with worse clinical outcomes. The objective of this study was to identify the main risk factors associated with VRE infection in the post-transplant period and to develop a risk prediction score for VRE infection. Historical cohort study including all adults who underwent LT between 2010 and 2022 in a tertiary-care hospital. Follow-up was from 90 days before transplant to 180 days after transplant. The outcome was VRE infection within the first 180 days after LT. Bivariate and multivariate analyses were performed using the Fine–Gray model, considering death as a competing risk. Prediction model metrics were calculated weekly, considering VRE colonization as a time-dependent variable, and described by ROC curves, sensitivity, specificity, accuracy, precision, and F1 score. A total of 1,209 patients were included; of these, 77 (6.4%) developed VRE infection post-transplant, most of which were intra-abdominal infections (58.4%). The main risk factors identified were: need for renal replacement therapy (OR 2.01; 95% CI 1.4–2.8), intraoperative bleeding (OR 4.01; 95% CI 2.7–6.0), post-transplant VRE colonization (OR 7.59; 95% CI 5.1–11.1) and pre-transplant colonization (OR 3.19; 95% CI 2.2–4.5), retransplantation (OR 3.85; 95% CI 2.6–5.6) and ICU length of stay (OR 1.01; 95% CI 1.0–1.02). Conversely, history of viral hepatitis (OR 0.54; 95% CI 0.3–0.7), autoimmune hepatitis (OR 0.33; 95% CI 0.1–0.6), Charlson comorbidity score (OR 0.84; 95% CI 0.7–0.9) and deceased-donor transplant (OR 0.23; 95% CI 0.1–0.3) were protective factors. All these variables were incorporated into the risk prediction model. The final score presented an area under the curve between 76.01% and 84.75%, with sensitivity from 75.13% to 89.33% and specificity from 74.6% to 85.87%. It was possible to develop a risk prediction score for VRE infection in liver transplant recipients with good discriminative performance. The proposed tool may support targeted surveillance and prevention strategies in higher-risk populations.