Wang Hai-Xu, Han Bing, Zhao Ying-Ying, Kou Lu, Guo Lu-Lu, Sun Tong-Wen, Song Lai-Jun
General ICU, The First Affiliated Hospital of Zhengzhou University, Henan Key Laboratory of Critical Care Medicine, Zhengzhou.
Department of Internal Medicine, People's Hospital of Jian'an District, Xuchang.
Medicine (Baltimore). 2020 Mar;99(13):e19563. doi: 10.1097/MD.0000000000019563.
Thrombotic thrombocytopenic purpura (TTP) is a life-threatening disease, and its mortality rate is 10% to 20%. However, there are currently only a few markers to predict the prognosis in patients with TTP. We aimed to identify several clinical indices and laboratory parameters for predicting the prognosis of TTP at admission.A single-centre observational cohort study that included patients with TTP from the First Affiliated Hospital of Zhengzhou University in China was conducted from January 1, 2012 to November 30, 2018. The primary outcome was prognosis, including in-hospital mortality, major thromboembolic events, or failure to achieve remission at discharge. We used the random forest method to identify the best set of predictors.Eighty-seven patients with TTP were identified, of whom 12 died during the treatment. The total number of patients within-hospital mortality, major thromboembolic events, and failure to achieve remission at discharge was 58. The machine learning method showed that the D-dimer level was the strongest predictor of the primary outcome. Receiver operating characteristic (ROC) analysis demonstrated that the sensitivity and specificity of the D-dimer level alone for identifying high-risk patients were 78% and 81%, respectively, with an optimum diagnostic cut-off value of 770 ng/mL. The area under the ROC curve (AUC) was 0.80, and the 95% confidence interval (CI) was 0.70 to 0.90.This study found that the D-dimer level exhibited a good predictive ability for prognosis in patients with TTP. These findings may aid in the development of new and intensive treatment strategies to achieve remission among high-risk patients. However, external validation is necessary to confirm the generalizability of our approach across populations and treatment practices.
血栓性血小板减少性紫癜（TTP）是一种危及生命的疾病，其死亡率为10%至20%。然而，目前仅有少数标志物可用于预测TTP患者的预后。我们旨在确定几种临床指标和实验室参数，以预测TTP患者入院时的预后。
我们进行了一项单中心观察性队列研究，纳入了2012年1月1日至2018年11月30日期间来自中国郑州大学第一附属医院的TTP患者。主要结局为预后，包括住院死亡率、主要血栓栓塞事件或出院时未实现缓解。我们使用随机森林方法来确定最佳预测指标集。
共纳入87例TTP患者，其中12例在治疗期间死亡。住院期间死亡、发生主要血栓栓塞事件以及出院时未实现缓解的患者总数为58例。机器学习方法显示，D-二聚体水平是主要结局的最强预测指标。受试者工作特征（ROC）分析表明，单独使用D-二聚体水平识别高危患者的敏感性和特异性分别为78%和81%，最佳诊断临界值为770 ng/mL。ROC曲线下面积（AUC）为0.80，95%置信区间（CI）为0.70至0.90。
本研究发现，D-二聚体水平对TTP患者的预后具有良好的预测能力。这些发现可能有助于制定新的强化治疗策略，以实现高危患者的缓解。然而，需要外部验证来确认我们的方法在不同人群和治疗实践中的通用性。
