目的 探讨血栓弹力图（TEG）参数联合血小板计数（PLT）对发热伴血小板综合征（SFTS）危重症患者血小板输注的指导作用。 方法 选取在感染科治疗的发热伴血小板减少综合征危重症患者62例，分为出血组（n = 30）和非出血组（n = 32），比较两组间PLT和TEG等相关指标的变化，采用二分类logistic回归分析预测危重症患者出血影响因素，用ROC曲线下面积分析相关指标预测患者出血的效能并计算最佳界值，用秩和检验和Fisher精确检验比较以MA界值分组后的PLT和死亡率。 结果 出血患者MA值和PLT低于非出血患者（P < 0.05），而K值、PT高于未出血患者（P < 0.05）。Logistic回归分析提示MA值和PLT是患者发生出血的影响因素（P < 0.05）。ROC分析显示PLT、MA对应的曲线下面积分别为0.884、0.890，P值均< 0.05，其中特异度最高的是PLT、灵敏度最高的是MA值，得出MA的最佳界值为48.85 mm，联合指标（K值、MA、PLT）得到的曲线下面积为0.941，约登指数0.77，均高于单独指标。秩和检验和Fisher精确检验显示，MA值< 48.85 mm组的PLT更低且死亡率更高（P < 0.05）。 结论 联合TEG和PLT，能有效预测SFTS危重症患者出血情况，对于指导SFTS危重症患者预防性血小板输注有较好的临床应用价值。

Objective To investigate the predictive value of thromboelastography （TEG） parameters in conjunction with platelet count （PLT） for guiding platelet transfusion in critically ill patients with fever associated with thrombocytopenia syndrome （SFTS）. Methods Sixty-two critically ill patients with fever and thrombocytopenia syndrome were selected from the infection department of a hospital. They were divided into two groups： the bleeding group （n = 30） and non-bleeding group （n = 32）. Changes in relevant indices， such as PLT and TEG， were compared between the two groups using t-test， chi-square test or non-parametric test. Binary logistic regression was employed to analyze factors predicting bleeding in critically ill patients. The efficacy of these indices in predicting bleeding was assessed by calculating the area under the receiver operating characteristic curve， determining optimal cut-off values， and comparing PLT levels with mortality rates using chi-square test after grouping based on MA cut-off value. Results First of all，the MA value and PLT were significantly lower in bleeding patients compared to non-bleeding patients （P < 0.05）， while the K value and PT were significantly higher in bleeding patients （P < 0.05）. Second of all，logistic regression analysis revealed that both MA value and PLT were significant influencing factors for hemorrhage in patients （P < 0.05）. Third of all， ROC analysis demonstrated that the area under the curve for PLT was 0.884， while for MA it was 0.890（P < 0.05）. Among these， PLT exhibited the highest specificity and MA value had the highest sensitivity， with a best cut-off value of 48.85 mm for MA. The combined index of K， MA， and PLT yielded an area under the curve of 0.941 and a Youden index of 0.77， which were higher than those obtained from individual indices. Finally， Rank sum test and Fisher exact test indicated that individuals with an MA value < 48.85 mm had lower PLT levels and higher mortality rates （P < 0.05）. Conclusion The integration of TEG and PLT demonstrates significant potential in predicting bleeding events among critically ill patients with SFTS， thereby offering valuable clinical implications for guiding prophylactic platelet transfusion in this patient population.