Objective: Factors influencing the prognosis of hemophagocytic lymphohistiocytosis (HLH) in adults were analyzed based on multicentric data. Methods: Clinical data of 124 adult patients with HLH diagnosed in eight medical centers in the Huaihai Lymphoma Working Group from March 2014 to July 2020 were collected. The optimal truncation value of continuous variables was obtained based on the Maxstat algorithm, X-Tile software, and restricted cubic spline. Cox proportional risk regression model was used to construct the adult HLH risk prediction model, and the visualization of the model was realized through the histogram. The bootstrap resampling method was used to verify the model, C-index and calibration curve was used to verify the histogram, and the prediction accuracy was checked. Kaplan-Meier analysis was used to calculate the survival rate and draw the survival curve. Furthermore, the differences between groups were tested by log-rank. Results: The median age of the 124 patients was 55 (18-84) years, including 61 (49.19%) males. The most common etiology was infection. Serum ferritin increased in 110 cases (88.71%) , hepatosplenomegaly in 57 cases (45.97%) . Of the 124 patients, 77 (62.10%) died, and the median survival time of the patients was 7.07 months. Univariate results showed that the prognosis of adult HLH was influenced by sex, age, fibrinogen, serum creatinine, alanine aminotransferase, and albumin (P<0.05) . The results of multivariate analysis showed that gender, platelet, albumin, alanine aminotransferase, and treatment regimens were independent influencing factors for prognosis. Based on the above five risk factors, the prediction model of the histogram was established, and the C-index of the model was 0.739. Finally, the calibration chart showed good consistency between the observed and predicted values of HLH. Conclusion: The prognosis of the adult hemophagocytic syndrome is influenced by many factors. Gender, platelet, albumin, alanine aminotransferase, and treatment regimens are independent risk factors. Therefore, the established histogram provides a visual tool for clinicians to evaluate the prognosis of adult HLH.
目的： 基于多中心数据，分析成人噬血细胞综合征（HLH）预后的影响因素。 方法： 收集2014年3月至2020年7月淮海淋巴瘤协作组8个医疗中心确诊的124例成人HLH患者的临床资料。基于Maxstat算法、X-Tile软件和限制立方样条获取连续变量的最佳截断值。采用Cox比例风险回归模型构建成人HLH风险预测模型并通过列线图实现模型的可视化，采用Bootstrap重抽样的方法进行验证，使用一致性指数（C-index）和校正曲线验证列线图，检查预测精度。采用Kaplan-Meier法分析计算生存率并绘制生存曲线，组间比较使用Log-rank检验。 结果： 124例患者的中位年龄为55（18~84）岁，男性61例（49.19%）。最常见的病因是感染，血清铁蛋白增高者110例（88.71%），肝脾肿大者57例（45.97%）。124例患者中77例（62.10%）死亡，患者的中位生存期为7.07个月，单因素分析结果表明成人HLH的预后受性别、年龄、纤维蛋白原、血肌酐、ALT和白蛋白的影响（P<0.05）。多因素分析结果表明性别、PLT、白蛋白、ALT和治疗方案是预后的独立影响因素，基于以上5个危险因素建立列线图预测模型，模型的C-index为0.739，校准图显示HLH的观测值和预测值之间有较好的一致性。 结论： 成人HLH的预后受多方面因素影响，性别、PLT、白蛋白、ALT和治疗方案是独立危险因素，基于上述危险因素建立的列线图为临床医师评估成人HLH预后提供了一个可视化工具。.
Keywords: Adult; Hemophagocytic lymphohistiocytosis; Prognostic analysis.