重症发热伴血小板减少综合征实验室指标预测死亡风险分析

目的 分析重症发热伴血小板减少综合征患者实验室指标与预后的相关性,并建立死亡模型加以验证。方法 收集重症发热伴血小板减少综合征病例,以2011年1月至2015年12月病例数据为建模组,建立死亡风险模型,以2016年1月至12月的病例数据为验证组,用于验证死亡模型。选取入住ICU后首次11项实验室检查结果进行分析,通过ROC曲线,筛选出曲线下面积(AUC)＞0.6的实验室指标,确定各项指标的最佳cutoff值。通过Logistic回归分析筛选出对预后有影响的实验室指标,并对其进行β值原始赋分,同时β值取整后行简化赋分,建立风险模型并验证,并对不同总评分的敏感度和特异度进行分析。结果 血ALT、AST、CKMB、Cr、UA、BUN,血清Ca²⁺、K⁺、Na⁺的AUC均＞0.60。logistic回归分析筛选出 AST、K⁺、BUN指标升高与预测死亡直接相关。测死亡的ROC曲线下面积分别为0.908,0.907,以简化赋分为3分为界值时预测死亡的敏感度和特异度分别为97.6%和95.2%。结论 AST、K、BUN的简化赋分对预测重症发热伴血小板减少综合征患者的死亡风险具有较高的价值,总评分≥3分的患者应引起临床医生的高度警惕。

Analysis of laboratory indicators predicting the risk of death for severe fever with thrombocytopenia syndrome

Objective To analyze the correlation between laboratory indexes and prognosis in patients with severe fever and thrombocytopenia syndrome, and to establish a death model for verification. Methods Cases of severe fever with thrombocytopenia syndrome were collected. The case data from January 2011 to December 2015 were used as the modeling group to establish the death risk model. The case data from January 2016 to December 2016 were used as the validation group to verify the death model. The first time 11 laboratory examination results were selected after admission to ICU for analysis. Through ROC curve, laboratory indexes with area under the curve (AUC)＞0.6 were selected to determine the best cutoff value of each index. Through the logistic regression analysis, the laboratory indexes that had influence on the prognosis were selected, and the original β value was assigned to them. At the same time, after the β value was rounded, the risk model was established and verified, and the sensitivity and specificity of different total scores were analyzed.Results The AUC of ALT, AST, CKMB, Cr, UA, BUM, Ca²⁺, K⁺ and Na⁺ were all >0.60. Logistic regression analysis showed that the increase of AST, K⁺, and BUN was directly related to the predicted death, respectively. The area under ROC curve was 0.908 and 0.907 respectively. The sensitivity and specificity were 97.6% and 95.2%.Conclusion The simplified scores of AST, K and BUN are of high value in predicting the risk of death in patients with severe fever and thrombocytopenia syndrome. Patients with a total score≥3 should be highly alert to the clinicians.