生理能力与手术应激评分系统和改良的生理能力与手术应激评分系统在预测胃癌术后死亡率和手术风险中的价值

Value of E-PASS and mE-PASS in predicting morbidity and mortality of gastric cancer surgery

目的：探讨生理能力与手术应激评分系统（ E-PASS）和改良的生理能力与手术应激评分系统（ mE-PASS）预测胃癌患者术后死亡率的价值及其相关参数与术后并发症的关系。方法收集2010年l月至2014年1月行手术切除的778例胃癌患者的临床资料，应用E-PASS和mE-PASS评分系统预测胃癌患者术后的死亡率，应用受试者工作特征（ ROC）曲线和拟合优度检验评价E-PASS和mE-PASS评分系统预测术后死亡率的效率，应用非条件Logistic回归分析评价E-PASS评分系统的相关参数与胃癌患者术后并发症的关系。结果 E-PASS和mE-PASS评分系统预测死亡组和非死亡组的死亡率差异均有统计学意义（均P＜0．05）。 E-PASS和mE-PASS评分系统预测胃癌术后死亡率风险的ROC曲线下面积分别为0．926和0．878。 E-PASS和mE-PASS评分系统预测胃癌患者术后死亡率的ROC曲线拟合优度检验差异均无统计学意义（均P＞0．05）。单因素分析显示，年龄、手术时间、严重心脏病、严重肺病、糖尿病、体能状态分级和美国麻醉医师协会（ ASA）分级为影响胃癌患者术后早期并发症的因素（均P＜0．05）；非条件Logistic回归分析结果显示，糖尿病、严重肺病、ASA分级和手术时间为影响胃癌患者术后并发症的独立因素（均P＜0．05）。结论 E-PASS和mE-PASS评分系统预测胃癌患者术后死亡率与实际死亡率具有较好的一致性，均适用于临床。与E-PASS评分系统比较， mE-PASS评分系统简化了术中参数。严重肺病、糖尿病、手术时间和ASA分级为影响胃癌患者术后并发症的独立危险因素。

Objective To investigate the clinical value of Physiologic Ability and Surgical Stress （ E-PASS） and modified Estimation of Physiologic Ability and Surgical Stress （ mE-PASS） scoring systems in predicting the mortality and surgical risk of gastric cancer patients, and to analyze the relationship between the parameters of E-PASS and early postoperative complications. Methods Clinical data of 778 gastric cancer patients who underwent elective surgical resection in Tianjin Medical University General Hospital from Jan. 2010 to Jan. 2014 were analyzed retrospectively. E-PASS and mE-PASS scoring systems were used to predict the mortality of gastric cancer patients, respectively. Univariate and unconditioned logistic regression analyses were performed to assess the relationships between nine parameters of E-PASS system and early postoperative complications. Results E-PASS and mE-PASS systems were used to predict the mortality in the death group and non-death group. The Z value was -5. 067 and -4. 492, respectively, showing a significant difference between the two groups （P〈0.05）. AUCs of mortality predicted by E-PASS and mE-PASS were 0.926 and 0.878 （P＆gt;0.05）, and the prediction calibration of postoperative mortality showed statistically non-significant difference （P〉0.05） between the E-PASS and mE-PASS prediction and actual mortality. Univariate analysis showed that age, operation time, severe heart disease, severe lung disease, diabetes mellitus, physical state index and ASA classification score are related to postoperative complications 〈br〉 （ P〈0. 05 for all ） . Unconditioned logistic regression analysis showed that severe lung disease, diabetes mellitus, ASA classification score and operation time are risk factors for early postoperative complications （ P〈0.05 for all） . Conclusions Both mE-PASS and E-PASS scoring system have good consistency in the predicting postoperative mortality and actual mortality, and both are suitable for clinical application. Moreover, the mE-PASS scoring system is clinically more simple and convenient than E-PASS scoring system. Preoperative severe lung disease, diabetes mellitus, ASA classification score and operation time are independent factors affecting the early postoperative complications.