A NEW study has developed a predictive model to identify which children with immune thrombocytopenia (ITP) are at risk of progressing to chronic disease at the point of diagnosis.
ITP is a disorder characterised by low platelet counts, leading to an increased risk of bleeding. While many paediatric cases resolve spontaneously within 12 months, up to 30% develop chronic ITP, creating uncertainty for clinicians and families early in the disease course.
Predicting Chronic Immune Thrombocytopenia at Diagnosis
In this retrospective study, researchers analysed data from 611 children with ITP to build a multivariable logistic regression model, before validating it in two independent cohorts totalling 161 patients. The aim was to improve early prediction of chronic ITP using routinely available clinical and laboratory parameters.
The model identified several key predictors present at diagnosis, including age, sex, immunoglobulin levels (IgG, IgA, and IgM), platelet count, lymphocyte count, presence of a known secondary cause, and positive direct antiglobulin test results. Together, these variables enabled accurate risk stratification for chronic ITP.
External validation demonstrated consistent discriminative performance, suggesting that the model could reliably distinguish between children likely to experience transient versus persistent disease.
Clinical Implications for ITP Management
The ability to predict chronic ITP at onset represents a significant advance in paediatric haematology. Currently, limited predictors exist, making it difficult to guide treatment decisions or provide clear prognostic information.
By identifying high-risk patients earlier, clinicians may be better equipped to tailor monitoring strategies and consider earlier therapeutic interventions where appropriate. Equally, families of children with a low predicted risk of chronic ITP may be reassured about the likelihood of spontaneous resolution.
Limitations and Future Directions
Despite its strengths, including external validation, the study was based on retrospective data, which may introduce inherent biases. Additionally, while the model incorporates widely available variables, further prospective studies are needed to confirm its utility in diverse clinical settings.
Future research may also explore integrating additional biomarkers or machine learning approaches to further refine prediction accuracy. Nonetheless, this model represents a practical and accessible tool that could enhance early decision-making in children with ITP, addressing a long-standing challenge in predicting chronic disease progression.
Reference
Hillier K et al. Predicting development of pediatric chronic immune thrombocytopenia at disease onset using a statistical risk model. Blood. 2026;DOI:10.1182/blood.2025028563.
Featured image: rdkcho on Adobe Stock
