Autor(es): Querido, Sara ; Ramalhete, Luís ; Gomes, Perpétua ; Weigert, André
Data: 2025
Identificador Persistente: http://hdl.handle.net/10362/191502
Origem: Repositório Institucional da UNL
Assunto(s): immunosuppression; infection; kidney transplantation; machine learning; TTV; Infectious Diseases; SDG 3 - Good Health and Well-being
Publisher Copyright: © 2025 by the authors.
Background: Torque Teno Virus (TTV) viremia has been proposed as a marker for infection risk in kidney transplant (KT) recipients. This study aimed to evaluate the prognostic value of TTV levels for predicting infections post-KT. Methods: A cohort of 82 KT patients was analyzed. TTV loads were measured before KT and at the time of cutoff analysis (mean time since KT: 20.2 ± 10.3 months). Infections were tracked within six months following the time of cutoff analysis. Univariable analyses and a supervised machine learning approach (logistic regression with leave-one-out cross-validation) were conducted to rigorously assess TTV’s predictive ability for post-transplant infection. Results: Seventy-two patients (87.8%) had detectable TTV before KT. Of these, 30.5% developed infections, predominantly viral. TTV loads increased significantly from 3.35 ± 1.67 log10 cp/mL before KT to 4.53 ± 1.93 log10 cp/mL at the time of cutoff analysis. Infected patients had significantly higher TTV loads (5.39 ± 1.68 log10 vs. 4.16 ± 1.94 log10 cp/mL, p = 0.0057). The optimal TTV threshold for predicting infection at the time of cutoff analysis was 5.16 log10 cp/mL, with 60% sensitivity and 81% specificity. Machine learning models improved performance, with sensitivity and specificity 0.805 and 0.735, respectively. Conclusions: TTV viremia may serve as a biomarker for infection risk, particularly when used with other clinical variables. The identified TTV threshold of 5.16 log10 cp/mL offers a practical tool for clinical decision-making, particularly when integrated with a machine learning model. Further studies with larger cohorts are needed to validate these findings and refine clinical applications.
