Risk prediction models for renal injury in children with IgA vasculitis: a systematic review and meta-analysis.

Abstract

Aims: The goal of this systematic review and meta-analysis was to provide references for future researchers on how to develop and implement predictive models for renal injury in paediatric IgA vasculitis (IgAV).

Design: Systematic review and meta-analysis of observational studies.

Methods: We systematically searched databases including China National Knowledge Infrastructure (CNKI), Wanfang Database, China Science and Technology Journal Database (VIP), SinoMed, PubMed, Web of Science, Cochrane Library, and Embase for studies on the construction of predictive models for renal injury in children with IgAV, up until 24 November 2024. Two researchers independently screened the studies, extracted data, and assessed bias risk via the Prediction Model Risk of Bias Assessment Tool (PROBAST). STATA 16.0 software was used to conduct meta-analysis of the area under the curve (AUC) values of the models.

Results: A total of 1,157 studies were retrieved. And 11 studies met the inclusion criteria. The sample sizes ranged from 155 to 583, with a renal injury incidence of 26.7-63.8%. The most common predictors included age, recurrent or persistent purpura, immunoglobulin A (IgA), D-dimer, and serum albumin (ALB). The included studies showed good overall applicability, however all were highly biased, mainly because they used inadequate data sources and reported poorly in the area analyzed. The pooled AUC of the five models was 0.86 (95% CI: 0.80-0.92), demonstrating good predictive power.

Conclusion: In spite of the fact that the renal injury prediction model was found to be somewhat predictive in children with IgAV, all of them had a high risk of bias according to the PROBAST checklist. For these predictive tools to be more robust and clinically applicable, new models with larger sample sizes, rigorous designs, and external validation should be developed in the future.