BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS评分系统对MODS患者病情评估价值的比较

The comparison of BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS scoring systems in evaluating severity and predicting prognosis of patients with MODS

目的评价BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS四种评分系统对多器官功能障碍综合征(MODS)患者的病情评估及预后的价值。方法前瞻性、多中心搜集MODS患者191例,选取资料完整的MODS患者141例,分别进行BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS评分,比较存活组和死亡组之间各种评分以及同一组不同时间段的分值差异;以及四种评分不同分数段患者病死率。然后分别绘制患者入组第1天的BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS受试者工作特征曲线,计算ROC曲线下面积,评价各评分预测MODS患者预后的准确性。结果141例患者中,存活84例(59.6%),死亡44例(32.2%),放弃治疗12例(8.5%),其他1例(0.7%)。总住院病死率为31.2%,ICU病死率为29.1%。存活组和死亡组在入组第1天APACHE Ⅱ评分,有统计学差异(P<0.05),而BJ-MODS、SOFA、Marshall-MODS评分均无统计学差异(P>0.05),入组第7天的BJ-MODS、APACHE Ⅱ、SOFA、Marshall-MODS评分均有统计学差异(P<0.05)。存活组和死亡组四种评分在两个时间段的差值,有统计学意义(P<0.05)。BJ-MODS和Marshall-MODS评分随着病死率增加而增加,入组第7天差异有统计学意义(P<0.001);APACHE Ⅱ和SOFA入组第1天评分随着病死率增加而有所下降,入组第7天差异有统计学意义(P<0.001)。BJMODS评分入组第1天、入组第7天及第1天与第7天差值的曲线下面积分别为0.600、0.865、0.835,APACHE Ⅱ评分为0.618、0.869、0.821,SOFA评分为0.556、0.872、0.893,Marshall-MODS评分为0.551、0.870、0.871。BJ-MODS入组第1天评分随着器官障碍数目的增加而增加,有统计学差异(P<0.05),而BJ-MODS入组第7天、APACHE Ⅱ、SOFA、Marshall-MODS评分增加不明显。ICU病死率随着器官障碍数目的增加而增加,有统计学差异(P<0.05)。结论 APACHE Ⅱ评分对MODS患者病情严重程度的预后评估能力最好,BJ-MODS与SOFA、Marshall评分相当。

Objective To evaluate the value of Beijing multiple organ dysfunction score （ BJ - MODS）, acute physiology and chronic healthy condition score II （ APACHE II ）, the sequential organ failure assessment（SOFA） score and Marshall - MODS in analyzing the severity and the prognosis of MODS. Methods A total of 191 patients with MODS were collected into this study. All the complete data of 141 patients were analyzed prospectively. The BJ - MODS, APACHE II , SOFA and Marshall - MODS scores were calculated and compared , as well as the mor- tality of patients with different scores segment. The areas of under receiver operating curves for these four scoring systems were calculated to evalu- ate the accuracy of each score systems in predicting the prognosis of patients with MODS. Results In the 141 patients, 84 （59.6%） died, 44 （32.2%） survived, 12 （8.5%） given up treatment and 1 （0.7%） for other reasons. The total hospital mortality was 31.2%, the ICU mortality rate was 29.1%. The score of APACHE II in the non - survivors was higher than those of the survivors in the first day into the group. The differ- ence was statistically significant （ P 〈 0.05 ）. However, there was no significant difference of BJ - MODS, SOFA and Marshall - MODS （ P 〉 0. 05 ）. The scores of BJ - MODS ,APACHE II ,SOFA and Marshall - MODS were statistically significant different （ P 〈0.001 ） in seventh day into the group. In survival group, the first day of these four scores were higher than those of the seventh day （ P 〈 0. 001 ）. In the non - survival group, the first day of these four scores were lower than those of the seventh day （ P 〈0. 001 ）. The mortality increased along with the increase of scores of BJ - MODS and Marshall - MODS. Only BJ - MODS at both the first and seventh days were statistically significant different （ P 〈 0. 05 ）. Only the seventh day of Marshall - MODS was statistically significant different （ P 〈 0.001 ）. The mortality increased along with the de- crease of the first day APACHE I1 and SOFA. The difference was not statistically significant （ P 〉 0. 05）, while the mortality increased with the in- crease of the seventh day of the two scores. There was a statistically significant difference （ P 〈0. 001 ）. The AUC of the first day were O. 600 for BJ - MODS, 0. 618 for APACHE II , 0. 556 for SOFA, 0. 551 for Marshall - MODS. The AUC of the seventh day were O. 865 for BJ - MODS, 0. 869 for APACHE II , 0. 872 for SOFA, 0. 870 for Marshall - MODS. And the AUC of the difference between the first day of the seventh day were 0. 835 for BJ - MODS, 0. 821 for APACHE II , 0. 893 for SOFA, 0.871 for Marshall - MODS. Conclusion APACHE II seorirtg system is the best method-in predicting the severity and prognosis of patients with MODS. BJ - MODS is similar with SOFA and Marshall. therefore, BJ - MODS scoring systems can be used to predict the severity and prognosis of the MODS.