Abstract
OBJECTIVE: Our study aims to identify risk factors that predict an abnormally prolonged duration of abdominal pain relief (DAPR) in patients with abdominal Henoch-Schönlein purpura (A-HSP) and to construct a nomogram for early prediction. METHODS: We reviewed data of all patients (n = 375) with confirmed A-HSP from the Chongqing Medical University platform, from 22 January 2011 to 18 November 2022. After applying rigorous inclusion and exclusion criteria, 220 patients were ultimately enrolled. We split them into two groups by the DAPR: < 1 week and ≥ 1 week. Multivariate relogit regression was performed to select factors associated with DAPR lasting ≥ 1 week, including demographics, symptoms, laboratory results, and treatment. We then constructed a nomogram to estimate risk probability and internally validated its performance via bootstrapping using discrimination, calibration, and clinical utility. RESULTS: There were 220 patients in the training. Multivariate relogit regression analysis demonstrated that age, initial onset, neutrophil-to-lymphocyte ratio (NLR), and bowel wall thickening were independent risk factors for DAPR ≥ 1 week. The Area Under the Curve (AUC) of the nomogram constructed based on the above factors was 0.849. Both the Calibration curve and Decision Curve Analysis (DCA) of the nomogram showed that the model exhibited good fitness. CONCLUSION: The nomogram can effectively predict the prolonged duration of abdominal pain (≥ 1 week) in A-HSP patients, helping clinicians distinguish high-risk patients at an early stage and optimize treatment plans. However, external validation remains essential before clinical implementation.