[Construction and evaluation of a thrombosis risk prediction model in patients with primary immune thrombocytopenia].

The overall incidence of thrombosis among hospitalized ITP patients was 12.0%.

• 334 ITP patients hospitalized at Qilu Hospital of Shandong University from January 2018 to December 2022 were retrospectively analyzed.
• 40 of 334 patients (12.0%) developed thrombosis, including 18 males and 22 females.
• Arterial thrombosis occurred in 9.58% (32/334) of patients.
• Venous thrombosis occurred in 1.80% (6/334) of patients.
• Mixed thrombosis occurred in 0.60% (2/334) of patients.

Univariate analysis identified age, ITP duration >1 year, comorbid hypertension, coronary heart disease, diabetes, and platelet count (PLT) as risk factors for thrombosis in ITP patients.

• All identified univariate risk factors were statistically significant (all P<0.05).
• Six factors were identified: age, ITP duration >1 year, comorbid hypertension, coronary heart disease, diabetes, and PLT.
• Analysis was performed using logistic regression on retrospectively collected clinical data.

Multivariate analysis identified age, ITP duration >1 year, comorbid coronary heart disease, and PLT as independent risk factors for thrombosis.

• All four independent risk factors were statistically significant (all P<0.05).
• Comorbid hypertension and diabetes, significant on univariate analysis, were not retained as independent predictors in multivariate analysis.
• Multivariate logistic regression was used to construct the prediction model.

The nomogram prediction model demonstrated good discriminative ability with an AUC of 0.80.

• The area under the ROC curve for the nomogram was 0.80 (95% CI: 0.72–0.88).
• The calibration curve showed good consistency between predicted and actual thrombosis rates.
• The Hosmer-Lemeshow goodness-of-fit test showed χ²=5.838, P=0.665, indicating good model fit.
• The model was constructed using the four independent risk factors identified by multivariate analysis.