3种量表预测急性髓系白血病患者有症状性PICC相关血栓形成风险效果的比较

Predictive power comparison of three different risk assessment models for symptomatic peripherally inserted central catheter related venous thromboembolism among patients with acute myeloblastic leukemia

目的比较Padua量表、Autar量表与Caprini量表在预测急性髓系白血病患有症状性PICC相关血栓风险的预测效能。方法采用便利抽样法,选取2019年3月—2021年3月在广州市南方医科大学南方医院血液科留置PICC的265例急性髓系白血病患者作为研究对象,采用一般资料调查表、Padua量表、Caprini量表、Autar量表和PICC导管维护随访表收集资料,计算3种评估量表的接受者操作特征曲线下面积并进行两两比较,计算量表的最佳临界值、灵敏度、特异度和约登指数。结果265例急性髓系白血病患者中,37例患者发生有症状性PICC相关血栓,发生率约14.0%;血栓发生时间为置管后第19.00(9.00,61.00)天。Padua量表、Caprini量表和Autar量表预测该类患者有症状性PICC相关血栓风险时ROC曲线下面积分别是0.784,0.633和0.648,且Padua量表ROC曲线下面积显著高于另外2种量表(P<0.05)。Padua量表的最佳临界值为4分,灵敏度为0.757,特异度为0.680,约登指数为0.437;Caprini量表最佳临界值为8分,灵敏度为0.676,特异度为0.548,约登指数为0.224;Autar量表最佳临界值为10分,灵敏度为0.595,特异度为0.693,约登指数为0.288。结论相较于Caprini量表和Autar量表,Padua量表能够更好地预测急性髓系白血病患者有症状性PICC相关血栓风险,但未来仍需针对具体疾病人群制定特异性的导管相关血栓风险评估工具以提高预测的准确性。

Objective To compare the predictive power of the Padua,Caprini and Autar scale for venous thromboembolism associated with symptomatic peripherally inserted central catheter(PICC)in patients with acute myeloid leukemia.Methods From March 2019 to March 2021,265 patients with acute myeloid leukemia with PICC at the Hematology Department of Nanfang Hospital of Southern Medical University were selected using convenient sampling.The General Information Questionnaire,the Padua scale,the Caprini scale,the Autar scale,and the PICC Maintenance Follow-up Tables were used to collect data.The area under the receiver operating characteristic curve(ROC)of the three scales was calculated and pairwise comparison was performed.The best cut-off value,sensitivities,specificity,and Youden exponents of the scales were calculated.Results Of the 265 patients with acute myeloid leukemia,37 had symptomatic PICC-associated thrombosis,with an incidence of approximate 14.0%;the median interval between PICC insertion and thrombotic episode was the 19.00th(9.00,61.00)day.The area under the ROC curve of Padua scale,Caprini scale and Autar scale was 0.784,0.633 and 0.648,respectively,and the area under the ROC curve of Padua scale was significantly higher than that of the other two scales(P<0.05).The sensitivity,specificity,and Youden index of the Padua scale were 0.757,0.680,and 0.437,respectively,with an optimal cutoff of 4.The sensitivity,specificity,and Youden index of the Caprini scale were 0.676,0.548,and 0.224,respectively,with an optimal cutoff of 8.The sensitivity,specificity,and Youden index of the Autar scale were 0.595,0.693,and 0.288,respectively,with an optimal cutoff of 10.Conclusion Compared to the Caprini and Autar scale,the Padua scale provides better prediction for the risk of symptomatic PICC-associated thrombosis in patients with acute myeloid leukemia.However,it is still necessary to develop specific catheter-associated thrombosis risk assessment tools for specific disease groups in the future to improve prediction accuracy.