有核红细胞联合Ranson评分及APACHEⅡ评分构建重症急性胰腺炎患者的结局预测模型

Establishment of a predictive model for outcomes in patients with severe acute pancreatitis by nucleated red blood cells combined with Ranson score and APACHE Ⅱ score

目的探讨重症急性胰腺炎(SAP)患者的结局预测模型。方法回顾性病例对照研究。收集温州医科大学附属台州医院2016年1月1日至2018年4月30日102例SAP患者,将入选者按照入院后90天转归情况分为生存组及死亡组,其中生存组88例,男性57例,女性31例,年龄55.5(40.3~69.8)岁;死亡组14例,男性5例,女性9例,年龄63.0(50.8~80.8)岁。通过比较2组间临床资料及实验室指标,筛选出具有统计学差异的指标,采用卡方自动交互检测法(CHAID)构建SAP患者的结局预测模型。收集2018年8月1日至2019年7月1日的50例SAP患者的临床资料及实验室指标结果对得出的结局预测模型进行验证。结果(1)通过卡方检验及Mann-Whitney U检验进行生存组及死亡组临床资料比较得出组间并发心脑血管系统疾病比例、机械通气、感染性休克比例、Charlson合并症指数(CCI)、Ranson评分及APACHEⅡ评分,死亡组均高于生存组,差异具有统计学意义(χ^2=5.554,P=0.018;χ^2=5.585,P=0.018;P=0.008,Z=-3.007,P=0.003;Z=-2.982,P=0.003;Z=-3.257,P=0.001);(2)通过卡方检验及Mann-Whitney U检验进行实验室指标比较得出生存组及死亡组间MPV、CRP、MCHC、pH、PCO2及NRBC阳性率,差异具有统计学意义(Z=-2.466,P=0.014;Z=-2.689,P=0.007;Z=-2.238,P=0.025;Z=-1.977,P=0.048;Z=-2.239,P=0.025;P=0.000),死亡组NRBC阳性率较生存组高,其余指标均较生存组低;(3)决策树CHAID法得出预测方案:当SAP患者满足Ranson评分≤3分时判定存活;当满足Ranson评分>3分且外周血中NRBC呈阴性判定存活;当满足Ranson评分>3分、NRBC呈阳性且APACHEⅡ评分≤21分时判定存活;当满足Ranson评分>3分、NRBC呈阳性且APACHEⅡ评分>21分时,判定预后不良;(4)验证组共收集50例SAP患者,实际生存43例,死亡7例,用方案预测患者结局准确率为94.0%(47/50)。结论有核红细胞联合Ranson评分及APACHEⅡ评分可预测SAP患者的结局。

Objective To explore a predictive model for outcomes of severe acute pancreatitis(SAP)patients.Methods A retrospective study was conducted of 102 SAP patients from January 1,2016 to April 30,2018 from Taizhou Hospital in this study.The participants were divided into survival group and death group according to the outcome of 90 days after admission,88 cases were in survival group including 57 males and 31 females,aged 55.5(40.3-69.8)years;14 cases were in death group including 5 males and 9 females,aged 63.0(50.8-80.8)years.Clinical data and laboratory indicators were compared between the two groups.Statistical analyses were performed to compare categorical variables.Chi-square automatic interaction detector(CHAID)was used to construct the prediction model of SAP patients′outcomes.The study cohort consisted of SAP patients from August 1st 2018 to July 1st 2019 were collected to validate the prediction model.Results(1)Statistical analyses were performed by chi square test and Mann-Whitney U test.There were statistically significant differences in the proportion of cardiovascular and cerebrovascular diseases,mechanical ventilation and septic shock,and also the Charlson complication index(CCI),the Ranson score and APACHEⅡscore(χ^2=5.554,P=0.018;χ^2=5.585,P=0.018;P=0.008;Z=-3.007,P=0.003;Z=-2.982,P=0.003;Z=-3.257,P=0.001),death group were higher than survival group.(2)The MPV,CRP,MCHC,pH,pCO2 and positive rates of NRBC were statistically different between survival group and death group(Z=-2.466,P=0.014;Z=-2.689,P=0.007;Z=-2.238,P=0.025;Z=-1.977,P=0.048;Z=-2.239,P=0.025;P=0.000).The NRBC-positive rate in the death group was higher than that in the survival group,while the other indexes were lower in the survival group.(3)The decision tree CHAID method obtains the prediction scheme:when it meets the Ranson score≤3,the SAP patients were judged to be alive;when it meets the Ranson score>3 and the NRBC in peripheral blood was negative,the SAP patients were judged to be alive;when it meets the Ranson score>3,the NRBC was positive and the APACHEⅡscore≤21,the SAP patients were judged to be alive;when it meets the Ranson score>3,NRBC was positive and APACHEⅡ>21,the SAP patients were adverse prognosis.(4)50 SAP patients were collected in the validation group,with 43 actual survivors and 7 deaths.The accuracy rate of predicting the outcomes of SAP patients in validation group with the scheme was 94.0%(47/50).Conclusion The NRBC combined with Ranson score system and APACHE II score system can predict the outcomes of SAP patients.