Research Article: Development and internal validation of a nomogram-based screening tool for concurrent autoimmune hepatitis in children with intestinal diseases: a retrospective cohort study
Abstract:
Autoimmune hepatitis (AIH) can complicate paediatric intestinal diseases, but practical tools for early screening and risk stratification are limited.
This retrospective cohort study included 346 children with intestinal diseases treated at a tertiary centre from 2015 to 2023, including 68 with concurrent AIH and 278 without AIH. Baseline clinical and laboratory variables measured at the initial gastroenterology evaluation were analysed. Candidate variables were selected using least absolute shrinkage and selection operator logistic regression and entered into multivariable logistic regression to construct a nomogram-based screening model. Model performance was assessed using receiver operating characteristic curve analysis, calibration metrics, bootstrap internal validation, and decision curve analysis.
Eight variables associated with concurrent AIH were included: disease duration ?12?months, C-reactive protein ?15?mg/L, immunoglobulin G ?16?g/L, erythrocyte sedimentation rate ?25?mm/h, antinuclear antibody positivity, corticosteroid use, pancolitis, and family history of autoimmune disease. The model showed an apparent area under the curve (AUC) of 0.974 and a bootstrap-corrected AUC of 0.893. Calibration was acceptable, with a slope of 0.87, intercept of ?0.03, Brier score of 0.098, and a non-significant Hosmer–Lemeshow test ( p =?0.398). Decision curve analysis indicated a greater net benefit for referral-oriented screening than single-predictor or default strategies.
The nomogram may help screen children with intestinal diseases who warrant early hepatological evaluation for concurrent AIH. It should not be used as a standalone diagnostic instrument, and external validation is required before routine clinical use.
Introduction:
Autoimmune hepatitis (AIH) can complicate paediatric intestinal diseases, but practical tools for early screening and risk stratification are limited.
Read more