中华急诊医学杂志  2015, Vol. 24 Issue (10): 1118-1121
AGI分级系统引入SOFA评分对重症急性胰腺炎患者预后预测价值的研究
张东, 杨艺敏, 段奥凇, 王育珊, 赵淑杰     
130021 长春,吉林大学第一医院ICU
摘要目的 将急性胃肠损伤(acute gastrointestinal injury,AGI) 分级系统引入序贯器官衰竭评价(sequential organ failure assessment,SOFA)评分,探讨对重症急性胰腺炎(severe acute pancreatitis,SAP)患者预后的预测价值,为临床提供更可靠的评估SAP预后的工具。 方法 选择2012年7月至2014年7月吉林大学第一医院ICU收治的SAP患者为研究对象,排除年龄<18岁,入院时间<24 h,家属放弃治疗的患者。共计入选63例,其中男 37例,女26例,年龄为(47±15.3)岁,统计所有患者入科时急性生理和慢性健康状况(acute physiology and chronic health evaluation,APACHE)Ⅱ评分,一周内最高SOFA评分和AGI分级,以及28 d病死率。将AGI分级系统中无AGI定义为0分,AGI I级-IV级分别定义为1-4分。利用受试者工作特征曲线(receiver operator characteristic curve,ROC) 评价APACHE Ⅱ评分,SOFA评分,以及SOFA+AGI评分对SAP患者预后的预测价值。应用MedCalc软件比较各评分的ROC曲线下面积(area under roc curve,AUC)之间差异是否具有统计学意义,以P<0.05为差异具有统计学意义。 结果 (1)63例SAP患者28 d病死率为20.6%(13/63),其中存活组50例,死亡组13例。比较存活与死亡两组患者的APACHE Ⅱ评分、SOFA评分及SOFA+AGI评分,分别为(15.62±4.33 VS 12.10±3.74,P=0.0048)、(14.77±3.09 vs 9.24±2.88,P<0.01)、(18.77±3.09 vs. 10.74±3.17,P<0.01)。(2)APACHE Ⅱ评分的AUC为0.748±0.084(95%CI:0.622~0.849)、SOFA评分的AUC为0.902±0.059(95%CI:0.801~0.962)、SOFA+AGI评分的AUC为0.963±0.037(95%CI:0.882~0.994)。APACHE Ⅱ评分和SOFA评分的AUC比较差异无统计学意义(P=0.10),而APACHE Ⅱ评分与SOFA+AGI评分的AUC比较差异具有统计学意义(P=0.013),SOFA评分与SOFA+AGI评分的AUC比较差异具有统计学意义(P=0.008)。3种评分系统比较以SOFA+AGI的Youden指数和阳性似然比最大,分别为0.863和15.38。 结论 SOFA评分对重症急性胰腺炎患者预后有较好的预测价值,而将AGI分级系统引入SOFA评分,可以获得更好的预测能力。
关键词急性胃肠损伤分级系统     急性生理和慢性健康状况Ⅱ评分     序贯器官衰竭评价评分     重症急性胰腺炎     预后    
Predictive value of AGI grading system introduced into SOFA score in patients with severe acute pancreatitis
Zhang Dong,Yang Yimin,Duan Aosong,Wang Yushan, Zhao Shujie     
Intensive Care Unit, The First Hospital of Jilin University, Changchun 130021, China.
Corresponding author: Zhao Shujie,Email:zsjdr@sina.com
Abstract: Objective To study the predictive value of acute gastrointestinal injury (AGI) grading system introduced into Sequential Organ Failure Assessment (SOFA) score in patients with severe acute pancreatitis (SAP) in order to provideareliable clinical tool for the evaluation of prognosis of SAP. Methods Patients with acute pancreatitis admitted to ICU from July 2012 to July 2014 were enrolled for study. The criteria of exclusion were the age below 18 years old, pregnancy, or patients without consent to the treatment.Atotal of 63 patients with 37 males and 26 females aged (47±15.3) years were included. The data of their acute physiology and chronic health evaluation (APACHE) Ⅱ score, the highest SOFA score and AGI grade within the first week, and the 28-day mortality rate were collected. Patients without AGI were defined as zero point, and AGI gradeI-IV were defined as 1-4 points. The receiver operating characteristic curve (ROC) was used to evaluate the value of APACHE Ⅱ score, SOFA score, and SOFA+AGI score in predicting the prognosis of SAP. The areas under ROC curve (AUC) of the APACHE Ⅱ score, SOFA score, and SOFA+AGI score were compared with MedCalc software, andPvalue less than 0.01 was considered to be statistical significance. Results (1) The 28-day mortality of the 63 patients with SAP was 20.6% (13/63),in which 50 patients in the survival group, 13 patients in the death group. The APACHEⅡ scores of two groups were (15.62 ± 4.33 vs. 12.10 ± 3.74, P=0.0048), the SOFA scores were (14.77 ± 3.09 vs.9.24 ± 2.88, P<0.01), and the SOFA+AGI scores were (18.77 ± 3.09 vs.10.74 ± 3.17, P<0.01). (2) The AUC of APACHEⅡ score was 0.748 ± 0.084 (95% CI: 0.622-0.849), the AUC of SOFA score was 0.902 ± 0.059 (95% CI:0.801-0.962), and the AUC of SOFA+AGI score was 0.963 ± 0.037 (95% CI,0.882-0.994); There was no significant difference in AUC between APACHE Ⅱ score and SOFA score (P=0.10), and there was statistical significance between the AUC of APACHE Ⅱ score and that of SOFA+AGI score (P=0.013), and the difference in AUC between SOFA score and SOFA+AGI score was statistically significant (P= 0.008). The Youden index and the positive likelihood ratio of SOFA+AGI score system were the greatest to be 0.863 and 15.38, respectively. Conclusions SOFA scoring system has better predictive value in patients with SAP when AGI grading system was introduced into it.
Key words: Acute gastrointestinal injury grading system     APACHE Ⅱ score     SOFA score     Severe acute pancreatitis     Prognosis    

重症急性胰腺炎(severe acute pancreatitis,SAP)起病急骤而病情凶险复杂,病死率可高达36%~50%[1]。由于SAP患者常常合并多器官功能障碍,因此以器官功能衰竭为核心的评分系统如序贯器官衰竭评价(sequential organ failure assessment,SOFA)评分有可能对SAP患者预后有更高的预测价值[2]。胃肠道既是SAP诱发全身炎症反应过程中最易累及的器官之一[3],亦可能因“肠源性细菌移位”而成为多器官功能障碍综合征的启动器官,急性胃肠功能障碍对SAP患者的病情转归具有重要影响。2012年欧洲危重病学会统一定义了重症患者的急性胃肠损伤(acute gastrointestinal injury,AGI)及其分级系统[4],从而使临床评价重症患者胃肠功能障碍的严重程度成为可能。本研究拟将AGI分级系统引入SOFA评分,通过比较分析各评分系统的ROC曲线下面积(area under roc curve,AUC)来评价对SAP患者预后的预测价值和能力,进一步为临床提供更可靠的评估SAP患者预后的工具。

1 资料与方法 1.1 一般资料

选择2012年7月至2014年7月吉林大学第一医院ICU收治的重症急性胰腺炎患者为研究对象,SAP诊断标准参照中国急性胰腺炎诊治指南(2013年,上海)[5]。排除年龄<18岁,入院时间<24 h,家属放弃治疗的患者。共计入选SAP患者63例,其中男性37例,女性26例,年龄19~77岁,(47±15.3)岁。

1.2 研究方法

统计所有患者入ICU时APACHE Ⅱ评分,入院一周内最高SOFA评分和AGI分级,以及28 d病死率。将2012年欧洲危重病学会的AGI分级系统[4]中无AGI定义为0分,AGI I级-IV级分别定义为1~4分。利用受试者工作特征曲线(ROC)评价APACHE Ⅱ评分,SOFA评分,以及SOFA+AGI评分对SAP患者预后的预测价值。分析各评分系统ROC曲线数值中的最佳临界值、灵敏度、特异度、最大约登(Youden)指数、阳性似然比(likelihood ratio of positive test,LRPT)。

1.3 统计学方法

应用SPSS 17.0软件获得的数据进行统计学分析,计量资料以均数±标准差(x±s)表示,组间比较采用成组t检验,计数资料的比较采用卡方检验或Fisher确切概率法检验。应用MedCalc软件分别获得APACHE Ⅱ评分,SOFA评分,以及SOFA+AGI评分的ROC曲线下面积(AUC)和95%CI,同时比较各评分的AUC之间差异。以P<0.05为差异具有统计学意义。

2 结果 2.1 SAP患者的预后

63例SAP患者的28 d病死率为20.6%(13/63),其中存活组50例,死亡组13例。死亡组和存活组患者的APACHE Ⅱ评分、一周内最高SOFA评分及SOFA+AGI评分比较分别为(15.62±4.33 vs 12.10±3.74,P<0.01)、(14.77±3.09 vs 9.24±2.88,P<0.01)、(18.77±3.09 vs 10.74± 3.17,P<0.01),两组之间各评分比较差异具有统计学意义。

2.2 APACHE Ⅱ评分、SOFA评分及SOFA+AGI评分对SAP患者预后的预测价值

各评分系统预测SAP患者预后的ROC曲线见图 1,其中APACHE Ⅱ评分的ROC曲线下面积(AUC)为0.748±0.084(95%CI为0.622-0.849)、SOFA评分的AUC为0.902±0.059(95%CI为0.801~0.962)、以及SOFA+AGI评分的AUC为0.963±0.037(95%CI为0.882~0.994);APACHE Ⅱ评分和SOFA评分的AUC比较差异无统计学意义(P=0.10),而APACHE Ⅱ评分与SOFA+AGI评分的AUC比较差异具有统计学意义(P=0.013),SOFA评分与SOFA+AGI评分的AUC比较差异具有统计学意义(P=0.008)。各评分系统诊断的最佳临界值、灵敏度、特异度、最大Youden 指数、以及LRPT,其中以SOFA+AGI评分系统的Youden 指数和LRPT最大,分别为0.863和15.38,见表 1

图 1 三种评分系统预测SAP患者预后的ROC曲线 Fig 1 ROC curve of the three scoring systems in predicting the prognosis of patients with SAP
表 1 三种评分系统预测SAP患者预后的ROC数值分析 Table 1 Analysis of ROC value of the three scoring systems in predicting the prognosis of patients with SAP
评分最佳临界值灵敏度特异度最大Youden指数LRPTAUC
APACHEⅡ≥15.50.5380.880.4184.480.748
SOFA ≥11.50.920.780.7034.180.902
SOFA+AGI ≥15.50.9230.940.86315.380.963
3 讨论

重症急性胰腺炎常常合并多器官功能衰竭,已成为ICU收治的主要重症之一。虽然有机械通气、血液滤过等多种器官功能支持的手段或措施,但SAP患者的病死率仍较高。本研究结果显示63例SAP患者28 d病死率为20.6%(13/63),这与相关文献报道是相近的[6, 7]。临床上如何进一步降低SAP患者的病死率已成为ICU医师面临的重要课题之一。准确而有效评价SAP患者的严重程度和病情转归,有利于及时开展或加强有效的器官支持治疗、指导临床合理使用医疗资源、以及客观评估疗效,可最大限度的提高救治水平,从而改善患者的预后。

APACHE Ⅱ评分是目前临床使用最普遍的病情分类和预测预后的综合性指标,主要由急性生理学评分(A)、年龄评分(B)、慢性健康评分(C)三项评分之和综合得出,所得分值越高,表示病情越重,预后越差。该系统包括了12项基本生理参数并考虑了患者年龄、慢性疾病的因素对预后影响,具有简便可靠、设计合理、预测准确的优点。一些研究结果显示APACHE Ⅱ评分可以预测急性胰腺炎患者的预后,但对SAP患者预后的预测并不可靠,本研究显示APACHE Ⅱ评分预测SAP预后的AUC值为0.748±0.084(95%CI为0.622~0.849),与一些研究报道相近[8, 9],然而只有当AUC值达到0.8以上时,评分系统才具有较可靠的预测价值。由此提示,APACHE Ⅱ评分系统并不能很好地预测SAP患者预后,这可能与该系统中年龄、慢性疾病等因素占有过多的权重相关[10]

SOFA评分在临床常用于评价器官功能衰竭的严重程度,目前对呼吸、凝血、肝脏、循环、神经等五个系统的评价指标,在评价患者循环系统状态时还引入了血管活性药物(多巴胺、肾上腺素等)的临床应用情况作为评分标准。已有研究表明SOFA评分系统预测ICU重症患者的病死率要优于Marshall的MODS评分[11]。本研究结果显示SAP患者在入院后一周内最高SOFA评分的AUC值为0.902±0.059(95%CI为0.801~0.962),但与APACHE Ⅱ的AUC值比较差异无显著统计学意义(P=0.10)。为进一步增加SOFA评分的预测能力,本研究将2012欧洲危重病学会的AGI分级系统作为急性胃肠功能障碍/衰竭评分引入SOFA评分系统,结果显示SOFA+AGI的AUC值为0.963±0.037(95%CI为0.882~0.994),而其分别与APACHE Ⅱ和SOFA的AUC值比较差异均有统计学意义,由此提示将AGI分级系统引入SOFA评分可进一步增强其对SAP患者预后的预测能力。

在急性胰腺炎发病初期,可根据APACHE Ⅱ评分、Ranson评分、BISAP评分、CT Balthazar分级、降钙素原等指标判断其严重程度和初步评估预后[12, 13, 14, 15]。然而SAP是在急性胰腺炎的基础上,伴有持续性器官功能衰竭,受累器官的个数和衰竭的程度将决定患者的预后,因此本研究将评价急性胃肠损伤的分级系统引入SOFA评分,比较了存活组和死亡组患者的APACHE Ⅱ评分、入院一周内最高SOFA评分和SOFA+AGI评分的差异,结果显示死亡组的三种评分均高于存活组(P<0.01),由此提示这三种评分系统分值越高,SAP患者预后可能越差。通过ROC曲线评价评分系统的准确性和预测能力的指标还包括Youden 指数和阳性似然比(LRPT),其中Youden 指数是评分系统的灵敏度与特异度之和减去1,取值范围在(-1,1)之间,Youden 指数越接近+1,说明该评分系统预测的准确性越好;而LRPT是真阳性率与假阳性率之比,其数值越大代表评分系统预测预后的能力越强。本研究结果显示在选取最佳临界值时,SOFA+AGI评分系统预测SAP患者预后的灵敏度、特异度、Youden 指数和LRPT均最大,分别为0.923、0.94、0.863、15.38,由此提示将AGI分级系统引入SOFA评分可提高对SAP 患者预后的预测准确性和能力。

参考文献
[1] VegeSS,Gardner TB,Chari ST,et al. Low mortality and high morbidity in severe acute pancreatitis without organ failure:acase for revising the Atlanta classification to include mode rately severe acute pancreatitis [J].AmJ Gastro-enterol,2009,104(3):710-715.
[2] Adam F,Bor C,Uyar M,et al.Severe acute pancreatitis admitted to intensive care unit: SOFA is superior to Ranson’ s criteria and APACHE Ⅱ in determining prognosis[J].TurkJGastroenterol,2013,24(5):430-435.
[3] 孙鑫国,肖帅,张树友.重症急性胰腺炎对胃肠动力影响的研究进展[J].中华急诊医学杂志,2013,22(3):327-330.
[4] Reintam Blaser A,Malbrain ML,Starkopf J,et al.Gastrointestinal function in intensive care patients:terminology, definitions and manage ment.Recommen-dations of the ESICM Working Group on Abdominal Problems[J].Intensive Care Med, 2012,38(3):384-394.
[5] 中华医学会消化病学分会胰腺疾病学组,中华胰腺病杂志编辑委员会,中华消化杂志编辑委员会.中国急性胰腺炎诊治指南(2013年,上海)[J].胃肠病学,2013,18(7):428-433.
[6] Xu H,Ebner L,Jiang S,et al. Retrocrural space involvement on computed tomography asapredictor of mortality and disease severity in acutepancreatitis[J].PLoS One,2014,9(9):e107378.
[7] 王春亭,曲鑫.重症急性胰腺炎的诊治进展[J].中华急诊医学杂志,2012,21(10):1080-1082.
[8] Papachristou GI,Muddana V,Yadav D,et al.Comparison of BISAP,Ranson's,APACHE-Ⅱ,and CTSI scores in predicting organ failure,complications,and mortality in acute pancreatitis[J].AmJGastroenterol,2010,105(2):435-441.
[9] 金洲祥,刘海斌,王向昱,等.四种评分系统对急性胰腺炎严重程度的预测价值[J].温州医学院学报,2012,42(5):449-451.
[10] 骆晓攀,方俊标,陈龙,等.APACHE Ⅱ、MODS、SOFA和LODS评分系统对重症急性胰腺炎预后评估的比较[J].医学研究杂志,2012,41(7):104-108.
[11] Khwmqnimit B.A comparison of three organ dysfunction scores:MODS,SOFA and LODS for predicting ICU mortality in critically Ill patients[J].J Med Assoc Thai,2007,90:1074-1081.
[12] Woo SM,Noh MH,Kim BG,et al.Comparison of serum procalcitonin with Ranson, APACHE-Ⅱ, Glasgow and Balthazar CT severity index scores in predicting severity of acute pancreatitis[J].KoreanJGastroenterol,2011,58(1):31-37.
[13] Khanna AK, Meher S, Prakash S,et al.Comparison of Ranson, Glasgow, MOSS, SIRS, BISAP, APACHE-Ⅱ, CTSI Scores, IL-6, CRP, and procalcitonin in predicting severity, organ failure, pancreatic necrosis,and mortality in acute panc-reatitis[J].HPB Surg,2013,24:367581.
[14] Senapati D,Debata PK,Jenasamant SS,et al.Aprospective study of the bedside index for severity in acute pancreatitis (BISAP) score in acute panc-reatitis: an Indian perspective[J].Pancreatology,2014,14(5):335-339.
[15] 高明,项和平,张长乐,等.急性胰腺炎患者外周血α-MSH、TNF-α、PCT的动态检测及临床意义[J].中华急诊医学杂志,2015, 24(4): 431-434.