摘要
目的:分析经皮肾镜碎石术(PCNL)后全身炎症反应综合征(SIRS)的相关危险因素,并建立Logistic回归预测模型。方法:回归性分析2017年1月-2018年3月期间在新疆医科大学第一附属医院泌尿中心结石病区行PCNL治疗上尿路结石的414例患者的临床资料。采用单因素及多因素Logistic回归分析筛选PCNL术后SIRS发生的危险因素,建立Logistic回归预测模型,最终用Hosmer and Lemeshow检验来对所建立的预测模型的拟合优度进行测验,用ROC曲线下面积来对该预测模型的效能进行评定。结果:414例PCNL患者当中术后发生SIRS的人数为62例,SIRS的发生率为15%,其中有3例患者发生感染性休克,占总数的0.72%。单因素Logistic回归分析的结果表明:性别、手术时间、通道数目、结石最大直径、术前尿培养阳性,术前尿白细胞阳性,术前尿亚硝酸盐阳性,结石成分等因素和PCNL术后SIRS的发生有着一定关系(P<0.05);多因素Logistic回归分析结果显示性别(OR=2.537,P<0.05)、手术时间(OR=5.025,P<0.05)、术前尿培养阳性(OR=4.524,P<0.05)、术前尿亚硝酸盐阳性(OR=2.247,P<0.05)是PCNL术后SIRS发生的独立危险因素(P<0.05);根据多因素Logistic回归分析有统计学意义的变量进入回归方程,获得的预测模型为P=1/1+EXP(-3.982+0.931×X1+1.614×X7+1.509×X14+0.810×X17)。Hosmer and Lemeshow检验提示预测模型拟合度为(χ2=5.724,P>0.05);ROC曲线下面积为0.857,95%CI(0.809~0.904)。结论:性别、手术时间、术前尿培养阳性、术前尿亚硝酸盐阳性是PCNL术后SIRS发生的独立危险因素,所建立的预测模型有一定的风险评估价值。
Objective:To analyze the risk factors of systemic inflammatory response syndrome(SIRS)after per?cutaneous nephrolithotomy(PCNL),and to establish a logistic regression-based prediction model.Methods:Regres?sion analysis was conducted on 414 patients with upper urinary tract calculi treated by PCNL in the Department of Urology of our hospital from January 2017 to March 2018. Univariate and multivariate logistic regression analyses were used to screen the risk factors of SIRS after PCNL,and to establish a logistic regression-based prediction mod?el. Finally,the Hosmer-Lemeshow test was used to assess the goodness of fit of the prediction model and the area un?der the ROC curve was used to evaluate its effectiveness.Results:Among 414 PCNL patients,62(15%)developed SIRS post-surgery and 3 patients had septic shock(0.72% of the total). Univariate logistic regression analysis showed that factors such as gender,operation time,number of channels,maximum diameter of stones,positive pre?operative urinary culture,positive preoperative urinary leukocytes,positive preoperative urinary nitrite and stone com?position were related to the occurrence of SIRS after PCNL(P<0.05). Multivariate logistic regression analysis showed that the gender(OR=2.537,P<0.05),operation time(OR=5.025,P<0.05),preoperative positive urine culture(OR=4.524,P<0.05),and preoperative positive urine nitrite(OR=2.247,P<0.05) were independent risk factors for SIRS after PCNL(P<0.05). According to the multivariate logistic regression analysis,statistically significant variables were added to a regression equation,and the resulting predictive model was P=1/1+ EXP(-3.982+0.931×X1+1.614×X7+1.509×X14+0.810×X17). The Hosmer-Lemeshow test indicated that the fit?ting degree of the prediction model was χ2=5.724(P>0.05),and the area under the ROC curve was 0.857 with95% CI[0,904].Conclusions:Patient gender,operation time,preoperative positive urine culture,and preopera?tive positive urine nitrite are independent risk factors for SIRS after PCNL,and the established predictive model has certain value in risk assessment.
作者
艾来提.吾甫尔
哈木拉提.吐送
安尼瓦尔.牙生
AILAITI Wufuer;HAMULATI Tusong;ANNIWAER Yasheng(Department of Urology,First Affiliated Hospital of Xinjiang Medical University,Urumqi 830054,China)
出处
《微创泌尿外科杂志》
2019年第1期40-45,共6页
Journal of Minimally Invasive Urology
关键词
经皮肾镜碎石术
全身炎症反应综合征
危险因素
预测模型
percutaneous nephrolithotomy
systemic inflammatory response syndrome
risk factors
prediction model