Objective To establish a simple and convenient early warning scoring system by retrospectively analyzing the high risk factors of early intracranial hemorrhage (ICH) in acute non-promyelocytic leukemia (ANPL) patients.
Methods We enrolled 1281 adult patients with ANPL (model group) and 378 adult patients (validation group), all patients were followed up for 12 months respectively. We used the data of model group to summarize the relevant factors of ICH occurrence and established risk scoring system. The risk scoring system was validated by the data of the validation group (each patient in the validation group was scored by the scoring system and divided into high, medium and low risks groups). The prediction efficiency of the scoring system was evaluated by ROC curves.
Results Univariate and multivariate analyses showed that the independent risk factors of early ICH in ANPL were WBC≥30.00×109/L, Fibrinogen(FIB) < 2.92 g/L and PT≥12.91 s. A new early warning scoring system was established based on the weights of the above three factors. The area under ROC curve of the scoring system was 0.774 (95%CI: 0.688-0.859). The area under the ROC curve of the scoring system was 0.715 (95%CI: 0.559-0.872) confirmed by the validation group.
Conclusion The new early warning scoring system could predict the occurrence of early ICH in ANPL and guide the medical staff to identify potential critical patients as early as possible.