Prediction models for infection after kidney transplantation: a systematic review

Infection after kidney transplantation is a leading cause of graft loss and mortality. Risk prediction models may facilitate early identification of high-risk recipients, yet their quality remains unclear. This study aimed to systematically evaluate the performance, risk of bias, and clinical applicability of prediction models for infection after kidney transplantation. It also sought to provide evidence-based guidance for model selection and future development.

We searched eight databases from inception to March 2026 for studies developing and validating prediction models for infection after kidney transplantation. Data extraction followed the CHecklist for critical Appraisal and data extraction for systematic Reviews of prediction Modelling Studies (CHARMS), and the Prediction model Risk Of Bias Assessment Tool (PROBAST) was used to assess risk of bias and applicability.

15 studies involving 20 models were included. All studies were retrospective and covered urinary tract infection, pulmonary infection, bloodstream infection (BSI), BK virus activation, and overall infection. The internally validated area under the curve (AUC) ranged from 0.63 to 0.925, with nine studies reporting an AUC above 0.80. However, these promising internal estimates should be interpreted with caution, as they may be inflated by optimism bias given the uniformly high risk of bias identified across all included studies. Only three studies undertook external validation. The PROBAST assessment revealed a high overall risk of bias in all studies. Frequently identified predictors were albumin, age, diabetes, kidney function, and donor type.

Current models show acceptable discrimination internally but have high risk of bias from methodological flaws. External validation and calibration assessment are insufficient. Future prospective multicenter studies should standardize predictor selection and missing-data handling while ensuring rigorous external validation.

https://www.crd.york.ac.uk/prospero/, identifier CRD420261336728.