急性复发性胰腺炎发生的影响因素及预测指标

Influencing factors and predictive indexes for acute recurrent pancreatitis

目的探讨急性复发性胰腺炎(acute recurrent pancreatitis,ARP)发生的影响因素及相关指标预测ARP发生的效能。方法急性胰腺炎(acute pancreatitis,AP)患者83例,采用EPS-G7连续检测法检测血清淀粉酶,酶比色速率法检测脂肪酶水平,采用双抗夹心免疫发光法检测血清降钙素原水平,采用ELISA法检测血清C反应蛋白、胰蛋白酶原激活肽、碱性磷酸酶(alkaline phosphates,ALP)、谷丙转氨酶、总胆红素和三酰甘油(triacylglycerol,TG)水平;记录患者年龄、性别比例、体质量指数、吸烟史等一般资料;多因素Cox回归分析ARP发生的影响因素;ROC曲线分析血清ALP、TG预测ARP发生的效能。结果 83例中失访3例,余80例中23例发生ARP(ARP组),57例未发生ARP(非ARP组);ARP组体质量指数[(35.28±4.39)kg/m^2],有吸烟史比率(73.9%),血清ALP[(173.52±82.16)u/L]、TG[(3.84±0.57)mmol/L]水平均高于非ARP组[(30.13±5.29)kg/m^2、42.1%、(116.39±63.74)u/L、(2.16±0.26)mmol/L](P<0.05),年龄、性别比例,有饮酒史比例,病因、CT分级及血清淀粉酶、脂肪酶、降钙素原、C反应蛋白、胰蛋白酶原激活肽、总胆红素和谷丙转氨酶水平与非ARP组比较差异均无统计学意义(P>0.05);多因素Cox回归分析结果显示,有吸烟史(HR=1.208,95%CI:1.012~2.369,P=0.023),高血清ALP(HR=1.013,95%CI:1.005~2.406,P=0.034)、高血清TG(HR=1.490,95%CI:1.335~1.756,P=0.006)是发生ARP的危险因素;ROC曲线分析结果显示,血清ALP以83.81u/L为最佳截断值,预测ARP发生的AUC为0.698(95%CI:0.566~0.810,P<0.001),灵敏度、特异度分别为96.7%、40.0%;血清TG以10.57 mmol/L为最佳截断值,预测ARP发生的AUC为0.716(95%CI:0.584~0.825,P<0.001),灵敏度、特异度分别为60.0%、76.7%;ALP联合TG预测ARP发生的AUC为0.869(95%CI:0.752~0.975,P<0.001),灵敏度、特异度分别为83.8%、89.2%。结论有吸烟史,高血清ALP、TG是ARP发生的危险因素,血清ALP、TG可作为预测ARP发生的分子标志物。

Objective To investigate the influencing factors for acute recurrent pancreatitis（ARP）and the related indexes for predicting ARP.Methods Eighty-three patients with acute pancreatitis were detected the levels of serum amylase by EPS-G7 continuous detection method,lipase by enzyme colorimetric rate method,serum procalcitonin by double-antibody Sandwich immunoluminescence assay,as well as the levels of C-reactive protein,trypsinogen activation peptide,alkaline phosphatase（ALP）,glutamic-pyruvic transaminase,total bilirubin and triacylglycerol（TG）by ELISA technique.The patient age,sex ratio,body mass index,smoking history and other general information were recorded.Cox regression analysis was used to study the influencing factors for ARP.ROC curve was used to analyze the efficacies of serum ALP and TG for predicting ARP occurrence.Results Among these 83 patients,3 were lost follow-up.In the remained 80 patients,ARP occurred in 23 patients（ARP group）,and the other 57 patients had no ARP（non-ARP group）.The body mass index（（35.28±4.39）kg/m^2）,proportion of smoking history（73.9%）,serum ALP level（（173.52±82.16）u/L）and TG level（（3.84±0.57）mmol/L）in ARP group were significantly higher than those in nonARP group（（30.13±5.29）kg/m^2,42.1%,（116.39±63.74）u/L,（2.16±0.26）mmol/L）（P〈0.05）.There were no significant differences in the age,sex ratio,proportion of drinking history,etiological classification,CT grading,and the serum levels of amylase,lipase,procalcitonin,C-reative protein,trypsionogen activation peptide,total bilirubin and glutamic-pyruvic transaminase between two groups（P〈0.05）.Multivariate Cox regression analysis showed that the history of smoking（HR=1.208,95%CI：1.012-2.369,P=0.023）,high serum ALP level（HR=1.013,95%CI：1.005-2.406,P=0.034）and high serum TG level（HR=1.490,95%CI：1.335-1.756,P=0.006）were the risk factors for ARP.ROC curve analysis showed when the optimal cut-offvalue of serum ALP was 83.81 u/L,the AUCfor predicting ARP was 0.698（95%CI：0.566-0.810,P〈0.001）,with the sensitivity and specificity of 96.7% and40.0%,respectively.When the optimal cut-offvalue of TG was 10.57 mmol/L,the AUCfor predicting ARP was 0.716（95%CI：0.584-0.825,P〈0.001）,with the sensitivity and specificity of 60.0%and 76.7%,respectively.The AUCof joint detection of ALP and TG for predicting ARP was 0.869（95%CI：0.752-0.975,P〈0.001）,and the sensitivity and specificity were 83.8% and 89.2%,respectively.Conclusion The history of smoking,high serum ALP level and high TG level are the risk factors for ARP,and serum ALP and TG can be used as molecular markers for predicting ARP.