文章快速检索     高级检索
   复旦学报(医学版)  2020, Vol. 47 Issue (5): 715-722      DOI: 10.3969/j.issn.1672-8467.2020.05.011
0
Contents            PDF            Abstract             Full text             Fig/Tab
引用本文
王莉莉, 葛圣金, 金玲艳, 李倩倩. 术后早期血糖异常对食管癌根治术患者短期预后的预测价值[J]. 复旦学报医学版, 2020, 47(5): 715-722.
WANG Li-li, GE Sheng-jin, JIN Ling-yan, LI Qian-qian. The value of abnormal early postoperative blood glucose concentration in predicting the short-term outcome of patients after radical resection of esophageal cancer[J]. Fudan University Journal of Medical Sciences, 2020, 47(5): 715-722.
术后早期血糖异常对食管癌根治术患者短期预后的预测价值
王莉莉1,2 , 葛圣金1 , 金玲艳2 , 李倩倩3     
1. 复旦大学附属中山医院麻醉科 上海 200032;
2. 复旦大学附属上海市第五人民医院麻醉科 上海 200240;
3. 安徽省立医院麻醉科 合肥 230001
收稿日期：2019-07-21；网络首发时间: 2020-07-20 13:25:01
摘要：目的 分析食管癌根治术后早期血糖异常对术后短期预后的评估价值。方法 回顾性研究复旦大学附属中山医院2015年4月至2017年9月的308例食管癌根治术患者的住院病例，收集人口学特征、术前已知的夹杂症、术前空腹血糖、术后早期血糖值及术后住院期间的短期临床不良事件等数据。308例中1例术后早期发生低血糖(3.8 mmol/L)，未单独成组。307例患者按术后早期血糖分为正常血糖组(n=215)和高血糖组(n=92)。另以术前的空腹血糖为基础血糖，分为术后血糖升高较多组(≥ 4 mmol/L，n=52)及血糖升高较少组(< 4 mmol/L，n=255)。结果 采用分层χ2检验进行术后早期高血糖与各术后临床不良事件的风险评估，差异无统计学意义。多元Logistics回归处理混杂因素后发现，血糖波动大是术后不良事件总发生率的危险因素(P=0.003，OR=2.641，95% CI：1.402~4.976)，血糖波动大亦是病死率的危险因素(P=0.012，OR=7.539，95% CI：11.534~36.713)。结论 术后早期血糖较术前空腹血糖波动较大可以预测食管癌根治术后短期预后不良。
关键词：食管癌根治术    术后    血糖    短期预后    
The value of abnormal early postoperative blood glucose concentration in predicting the short-term outcome of patients after radical resection of esophageal cancer
WANG Li-li1,2 , GE Sheng-jin1 , JIN Ling-yan2 , LI Qian-qian3     
1. Department of Anesthesiology, Zhongshan Hospital, Fudan University, Shanghai 200032, China;
2. Department of Anesthesiology, Shanghai Fifth People's Hospital, Fudan University, Shanghai 200240, China;
3. Department of Anesthesiology, Anhui Provincial Hospital, Hefei 230001, Anhui Province, China
Corresponding author: GE Sheng-jin, E-mail:shengjinge@163.com.
Abstract: Objective To analyze the relationship between the short-term outcomes and abnormal early postoperative blood glucose concentration of patients after radical resection of esophageal cancer. Methods We conducted a retrospective cohort analysis of 308 patients who underwent radical resection of esophageal cancer between Apr.2015 and Sep.2017 in Zhongshan Hospital, Fudan University.These materials were including demographic characteristics, known preoperatively previous medical history (PMH), fasting blood glucose before operation, early postoperative blood glucose concentration and short-term clinical adverse events during hospitalization.Among the 308 cases, 1 patient developed hypoglycemia (3.8 mmol/L), which was not isolated into a group.The other 307 patients were divided into two groups according to the early postoperative blood glucose values:normal blood glucose group (n=215) and hyperglycemia group (n=92).In addition, based on the fasting blood glucose before operation, the patients were divided into the group with higher postoperative blood glucose fluctuation (≥ 4 mmol/L, n=52) and the group with lower blood glucose fluctuation(< 4 mmol/L, n=255). Results No statistical difference was found between early postoperative hyperglycemia and postoperative adverse events by stratified Person Chi-Square test. Using multivariate Logistics regression to deal with confounding factors, the blood glucose fluctuation was a risk factor.The incidence of postoperative adverse events was higher in the higher glucose fluctuation group than that in the lower glucose fluctuation group using preoperative fasting blood glucose concentration as baseline (P=0.003, OR=2.641, 95%CI:1.402-4.976).Mortality was higher in the higher glucose fluctuation group than in the lower glucose fluctuation groupusing preoperative fasting blood glucose as baseline (P=0.012, OR=7.539, 95%CI:1.534-36.713). Conclusion Early high postoperative blood glucose fluctuation may be a predictorfor poor short-term outcomes of patients after radical resection of esophageal cancer.
Key words: radical resection of esophageal cancer    postoperation    blood glucose    short-term outcome    

随着腔镜手术的开展和麻醉技术的不断更新，食管癌手术患者术后临床不良事件发生率明显下降[1-2]。围术期血糖管理仍是关系预后的重要环节，血糖异常(高血糖、血糖波动和低血糖)则是影响手术患者短期预后不良的因素[3-9]，有研究认为加强血糖控制有利于改善患者预后[10-11]。既往血糖与预后的研究多针对心血管疾病或急诊重症患者[11, 16]，目前关于食管癌患者围术期血糖管理的研究甚少。本研究旨在探讨术后早期血糖异常对食管癌根治术患者术后短期预后的预测价值，为进一步研究食管癌根治术患者术后血糖管理方案提供临床医学证据。

资料和方法

资料收集及分组   回顾性收集复旦大学附属中山医院2015年4月至2017年9月行食管癌根治术患者的住院病例，排除有严重系统疾病者、年龄 > 80岁及病史资料缺失的患者。收集的数据包括人口学特征和术前已知的夹杂症(呼吸系统疾病、心脏病、高血压、糖尿病等)，采集病史中术后入外科监护室后第1个血糖值，术后伤口恢复情况(包括有无感染)，肺部感染情况(具备影像学证据并有临床表现等)，食管瘘，乳糜胸，心血管意外发生情况(术后发生心律失常、心跳骤停、心衰等)，其他并发症(脏器功能衰竭、术后出血、喉返神经损伤等)，外科监护室住院天数，总住院天数，病死率，术后抗生素使用情况等。

高血糖评价标准根据2015年版围术期血糖管理专家共识[12]，推荐监护室患者血糖控制低于8.4 mmol/L。按照入外科监护室后第1个血糖值分为正常血糖(NG)组及高血糖(HG)组。

以术前空腹血糖为基础血糖，将患者按照手术前后血糖值差分组。根据2015年版围术期血糖管理专家共识[12]，术后血糖高于10.0 mmol/L需胰岛素治疗以稳定血糖，考虑术前血糖正常值上限为6.1 mmol/L，据此取差值4 mmol/L。将患者分为术后血糖升高较小组(< 4 mmol/L，A组)和术后血糖升高较大组(≥4 mmol/L，B组)。

统计学分析   本研究采用SPSS 20.0统计软件进行统计分析。连续变量表示为x±s，用t检验进行评价；分类变量表示为频率和百分比，用χ2检验和Wilcoxon秩和检验分析；用多元Logistic回归校正混杂因素。P < 0.05为差异有统计学意义。

结果

术后高血糖组与正常血糖组患者的基本资料   共308例患者入组，其中1例术后血糖值为3.8 mmol/L，未单独成组。其余307例患者分为正常血糖组(NG组，n=215)及高血糖组(HG组，n=92)，正常血糖组包括正常低值血糖(≤5.6 mmol/L，n=24)，高血糖组包括严重高血糖(≥11.1 mmol/L，n=5)(表 1)。两组间年龄(P < 0.001)、性别(P=0.001)及是否合并糖尿病(P=0.004)差异有统计学意义。高血糖组患者年龄较大，女性比例较高，合并糖尿病患者比例较高。

表 1 正常血糖组与高血糖组患者的基本资料 Tab 1 Characteristics of patients in hyperglycemia group and normal blood glucose groups 
[n (%) or x±s]
Item HG
(n=92)		 NG
(n=215)		 P
Age (y) 64.3±6.8 60.9±7.3 < 0.001
Male 63 (68.5) 183 (85.1) 0.001
BMI (kg/m2) 22.7±3.2 23.3±8.8 0.562
Other preoperative treatments(1) 12 (13.0) 26 (12.1) 0.817
PMH
  Diabetes mellitus 13 (14.1) 10 (4.7) 0.004
  Hypertension 26 (28.3) 56 (26.0) 0.688
  Cardiovascular disease 4 (4.3) 11 (5.1) 0.775
  Chronic bronchitis 1 (1.1) 1 (0.5) 0.536
  Stroke 5 (5.4) 4 (1.9) 0.090
(1)Including chemoradiotherapy and ESD surgery.HG:Hyperglycemia group; NG:Normal blood glucose group; BMI:Body mass index; ESD:Endoscopic submucosal dissection; PMH:Previous medical history.Wilcoxon rank sum test was used for classified variables, and t test was used for continuous variables.
表选项

术后血糖值与各项结局及死亡率  分析高血糖与各项结局包括伤口感染、呼吸系统感染、吻合口瘘、乳糜胸、心血管不良事件、ARDS、死亡率的关系，未发现术后早期高血糖为食管癌根治术后预后不良的危险因素(表 2)。

表 2 高血糖与短期预后的相关性 Tab 2 Correlation between hyperglycemia and short-term outcomes 
[n (%) or x±s]
Item HG (n=92) NG (n=215) P OR 95%CI
Wound infection 7 (7.6) 20 (9.3) 0.631 0.803 0.327-1.970
Pulmonary infection 14 (15.2) 39 (18.1) 0.535 0.810 0.416-1.577
Anastomotic fistula 9 (9.8) 28 (13.0) 0.424 0.724 0.327-1.602
Chylothrax 0 (0) 3 (1.4) 0.557 1.014 0.998-1.030
MACE 6 (6.5) 10 (4.7) 0.499 1.430 0.504-4.058
RF 6 (6.5) 11 (5.1) 0.622 1.294 0.464-3.610
Total incidence of adverse events 38 (41.3) 77 (35.8) 0.363 1.261 0.765-2.079
Mortality 4 (4.3) 3 (1.4) 0.112 3.212 0.704-14.649
Days in SICU (d) 4.73±4.67 4.17±4.94 0.355 - -
Length of stay (d) 22.9±15.0 23.3±18.0 0.83 - -
MACE:Major adverse cardiovascular events; RF:Respiratory failure.Fisher’s exact test was used for frequency < 5, χ2 test was used for risk assessment, and Wilcoxon rank sum test was used for classified variables.
表选项

按年龄、性别、是否合并糖尿病病史分层进行风险评估，也未发现术后早期高血糖为食管癌根治术后预后不良的危险因素(表 3~5)。

表 3 按性别分层后高血糖与短期预后的相关性 Tab 3 Correlation between hyperglycemia and poor short-term outcomes after stratification by gender  
[n (%)]
Item HG (n=92) NG (n=215) P OR 95%CI
Wound infection
  Male 5 (7.9) 14 (7.3) 0.941 1.041 0.359-3.015
  Female 2 (6.9) 6 (18.8) 0.171 0.321 0.059-1.737
Pulmonary infection
  Male 10 (15.9) 32 (17.5) 0.769 0.890 0.410-1.934
  Female 4 (13.8) 7 (21.9) 0.412 0.571 0.148-2.199
Anastomotic fistula
  Male 7 (11.1) 25 (13.7) 0.604 0.790 0.324-1.927
  Female 2 (6.9) 3 (9.4) 0.725 0.716 0.111-4.620
Chylothrax
  Male 0 (0) 3 (1.6) 0.572 1.017 0.998-1.036
  Female 0 (0) 0 (0) - - -
MACE
  Male 4 (6.3) 8 (4.4) 0.530 1.483 0.431-5.104
  Female 2 (6.9) 2 (6.3) 0.919 1.111 0.146-8.441
RF
  Male 6 (9.5) 10 (5.5) 0.251 1.821 0.634-5.232
  Female 0 (0) 1 (3.1) 1.000 1.032 0.970-1.099
Total incidence of adverse events
  Male 27 (42.8) 64 (34.9) 0.264 1.395 0.778-2.010
  Female 11 (37.9) 13 (40.6) 0.830 0.893 0.319-2.501
Mortality
  Male 3 (4.8) 3 (1.6) 0.166 3.000 0.590-15.262
  Female 1 (3.4) 0 (0) 0.475 0.966 0.901-1.034
χ2 test was used for risk assessment, and Fisher’s exact test was used for frequency < 5.MACE:Major adverse cardiovascular events; RF:Respiratory failure.Male in HG group:n=63;Female in HG group:n=29;Male in NG group:n=183;Female in NG group:n=32.
表选项
表 4 按合并糖尿病与否分层后高血糖与短期预后的相关性 Tab 4 Correlation between hyperglycemia and short-term outcomes after stratification by diabetes mellitus 
[n (%)]
Item HG (n=92) NG (n=215) P OR 95%CI
Wound infection
  Diabetes 1 (7.7) 1 (10.0) 0.846 0.750 0.041-13.677
  Non-diabetes 6 (7.6) 19 (9.3) 0.656 0.321 0.309-2.095
Pulmonary infection
  Diabetes 1 (7.7) 4 (40.0) 0.063 0.125 0.011-1.379
  Non-diabetes 13 (16.5) 35 (17.1) 0.901 0.957 0.476-1.921
Anastomotic fistula
  Diabetes 1 (7.7) 1 (10.0) 0.846 0.750 0.041-13.677
  Non-diabetes 8 (10.1) 27 (13.2) 0.484 0.743 0.322-1.713
Chylothrax
  Diabetes 0 (0) 0 (0) - - -
  Non-diabetes 0 (0) 3 (1.5) 0.280 1.015 0.998-1.032
MACE
  Diabetes 1 (7.7) 0 (0) 0.370 0.923 0.789-1.080
  Non-diabetes 5 (6.3) 10 (4.9) 0.624 1.318 0.436-3.983
RF
  Diabetes 0 (0) 1 (10.0) 0.435 1.111 0.904-1.366
  Non-diabetes 6 (7.6) 10 (4.9) 0.374 1.603 0.562-4.567
Total incidence of adverse events
  Diabetes 4 (30.7) 5 (50.0) 0.349 0.444 0.080-2.457
  Non-diabetes 34 (43.0) 72 (35.1) 0.216 1.396 0.822-2.370
Mortality
  Diabetes 0 (0) 0 (0) - - -
  Non-diabetes 4 (5.1) 3 (1.5) 0.080 3.591 0.785-16.423
Diabetes in HG group:n=13;Non-diabetes in HG group:n=79;Diabetes in NG group:n=10;Non-diabetes in NG group:n=205.Others were as same as Tab 3.
表选项
表 5 按年龄分层后高血糖与短期预后的相关性 Tab 5 Correlation between hyperglycemia and short-term outcomes after stratification by age 
[n (%) or x±s]
Item HG (n=92) NG (n=215) P OR 95%CI
Wound infection
   < 60 y 2 (9.1) 5 (6.0) 0.635 1.560 0.282-8.641
  60-70 y 5 (9.1) 12 (10.7) 0.744 0.833 0.278-2.496
   > 70 y 0 3 (15.0) 0.244 1.176 0.979-1.414
Pulmonary infection
   < 60 y 4 (18.2) 14 (16.9) 1.000 1.095 0.321-3.733
  60-70 y 6 (10.9) 20 (17.9) 0.244 0.563 0.212-1.495
   > 70 y 4 (26.7) 5 (25.0) 1 1.091 0.237-5.027
Anastomotic fistula
   < 60 y 1 (4.5) 11 (13.3) 0.453 0.312 0.038-2.556
  60-70 y 6 (10.9) 13 (11.6) 0.894 0.932 0.334-2.602
   > 70 y 2 (13.3) 4 (20.0) 0.680 0.615 0.097-3.908
Chylothrax
   < 60 y 0 (0) 2 (2.4) 1.000 1.025 0.991-1.060
  60-70 y 0 (0) 0 (0) 1.000
   > 70 y 0 (0) 1 (5.0) 0.557 1.053 0.952-1.164
MACE
   < 60 y 0 (0) 3 (3.6) 1.000 1.038 0.995-1.082
  60-70 y 3 (5.5) 3 (2.7) 0.397 2.096 0.409-9.763
   > 70 y 3 (20.0) 4 (20.0) 1.000 1.000 0.188-5.332
RF
   < 60 y 2 (9.1) 2 (2.4) 0.193 4.050 0.537-30.534
  60-70 y 3(5.5) 6 (5.4) 1 1.019 0.245-4.238
   > 70 y 1 (6.7) 3 (15.0) 0.619 0.405 0.038-4.335
Total incidence of adverse events
   < 60 y 9 (40.9) 27 (32.5) 0.462 1.436 0.546-3.773
  60-70 y 22 (40.0) 42 (37.5) 0.755 1.111 0.574-2.153
   > 70 y 7 (46.7) 8 (40.0) 0.693 1.313 0.339-5.076
Mortality
   < 60 y 1 (4.5) 2 (2.4) 0.510 1.929 0.167-22.304
  60-70 y 2 (3.6) 0 (0) 0.107 0.964 0.915-1.014
   > 70 y 1 (6.7) 1 (5.0) 1.000 1.357 0.078-23.615
< 60 y in HG group:n=22;60-70 y in HG group:n=55; > 70 y in HG group:n=15; < 60 y in NG group:n=83;60-70 y in NG group:n=112；> 70 y in NG group:n=20.Others were as the same as Tab 3.
表选项

按血糖波动分组患者的基本资料307例患者以术前、术后血糖值差分组，血糖升高 < 4 mmol/L (A)组255例，血糖升高≥4 mmol/L(B)组52例。两组间年龄(P=0.020)和性别(P=0.017)差异均有统计学意义。B组患者年龄较大，女性占比更高(表 6)。

表 6 血糖波动较大组与血糖波动较小组患者的基本资料 Tab 6 Characteristics of patients in higher and lower blood glucose fluctuation groups 
[n (%) or x±s]
Item Group A (n=255) Group B (n=52) P
Age (y) 61.5±7.5 64.1±6.18 0.020
Male 208 (81.6) 38 (73.1) 0.017
BMI (kg/m2) 23.2±8.2 22.5±2.8 0.505
Other preoperative treatments⑴ 30 (11.8) 12 (23.1) 0.471
PMH
Diabetes 16 (6.3) 7 (13.5) 0.073
Hypertension 67 (26.2) 15 (28.8) 0.703
Cardiovascular disease 14 (5.5) 1 (1.9) 0.278
Chronic bronchitis 1 (0.4) 1 (1.9) 0.212
Stroke 7 (2.7) 2 (3.8) 0.668
(1)Including chemoradiotherapy and ESD surgery.Group A:Higher blood glucose fluctuation group (≥4 mmol/L); Group B:Lower blood glucose fluctuation group (< 4 mmol/L).Wilcoxon rank sum test was used for classified variables, and t test was used for continuous variables.
表选项

血糖波动与各项结局采用χ2检验分析食管癌根治术后短期预后的各项结局与血糖波动值的关系，发现不良事件总发生率(P=0.001，OR=2.727，95%CI：1.484~5.013)和病死率(P=0.004，OR=7，95%CI：1.518~32.277)差异均有统计学意义(表 7)。针对不良事件总发生率和病死率这两个结局，将各种危险因素放入多元Logistics回归模型中，发现血糖波动大这个危险因素的OR值分别为2.641(P=0.003，95%CI：1.402~4.976)和7.539(P=0.012，95%CI：1.534~36.713)(表 8~9)。

表 7 血糖波动与短期预后的相关性 Tab 7 Correlation between blood glucose fluctuation and short-term outcomes 
[n (%) or x±s]
Item Group B (n=52) Group A (n=255) P OR 95%CI
Wound infection 5 (10.0) 22 (8.6) 0.819 1.127 0.406-3.126
Pulmonary infection 13 (25.0) 40 (15.7) 0.105 1.792 0.879-3.654
Anastomotic fistula 9 (17.3) 28 (11.0) 0.202 1.697 0.748-3.848
Chylothrax 0 (0) 3 (1.2) 1.000 1.012 0.998-1.026
MACE 4 (7.7) 12 (4.7) 0.377 1.688 0.522-5.454
RF 5 (9.6) 12 (4.7) 0.158 2.154 0.725-6.401
Total incidence of adverse events 30 (57.6) 85 (33.3) 0.001 2.727 1.484-5.013
Mortality 4 (7.7) 3 (1.2) 0.004 7.000 1.518-32.277
Days in SICU(d) 5.58±5.862 4.08±4.597 0.088 - -
Length of stay(d) 26.44±18.547 22.53±16.786 0.163 - -
Fisher’s exact test was used for frequency < 5, and χ2 test was used for risk assessment.Wilcoxon rank sum test was used for classified variables, and t test was used for continuous variables..Group A:Higher blood glucose fluctuation group (≥4 mmol/L); Group B:Lower blood glucose fluctuation group (≥4 mmol/L).MACE:Major adverse cardiovascular events; RF:Respiratory failure.
表选项
表 8 按年龄分层后血糖波动与短期预后的相关性 Tab 8 Correlation between blood glucose fluctuation and short-term outcomes after stratification by age 
[n (%)]
Item Group B (n=52) Group A (n=255) P OR 95%CI
Wound infection
   < 60 y 0 (0) 7 (7.4) 1.000 1.080 1.020-1.143
  60-70 y 5 (14.3) 12 (9.1) 0.366 1.667 0.545-5.095
   > 70 y 0 (0) 3 (10.7) 1.000 1.120 0.985-1.273
Pulmonary infection
   < 60 y 4 (40) 14 (14.7) 0.066 3.857 0.964-15.432
  60-70 y 6 (17.1) 20 (15.2) 0.773 1.159 0.426-3.148
   > 70 y 3 (42.9) 6 (21.4) 0.340 2.750 0.479-15.794
Anastomotic fistula
   < 60 y 1 (10.0) 11 (11.5) 1.000 0.848 0.098-7.353
  60-70 y 6 (17.1) 13 (9.8) 0.227 1.894 0.663-5.407
   > 70 y 2 (28.6) 4 (14.3) 0.576 2.400 0.341-16.899
Chylothrax
   < 60 y 0 (0) 2 (2.1) 1.000 1.022 0.992-1.052
  60-70 y 0 (0) 0 (0) 1.000 1.037 0.966-1.114
   > 70 y 0 (0) 1 (3.6) 1.000 1.012 0.998-1.026
MACE
   < 60 y 0 (0) 3 (3.2) 1.000 1.033 0.996-1.071
  60-70 y 3 (8.6) 3 (2.3) 0.107 4.031 0.777-20.916
   > 70 y 1 (14.3) 6 (21.4) 1.000 0.611 0.061-6.104
RF
   < 60 y 2 (20.0) 2 (2.1) 0.044 11.625 1.440-93.871
  60-70 y 2 (5.7) 7 (5.3) 1.000 1.082 0.215-5.456
   > 70 y 1 (14.3) 3 (10.7) 1.000 1.389 0.122-15.812
Total incidence of adverse events
   < 60 y 6 (60.0) 30 (31.6) 0.088 3.250 0.853-12.376
  60-70 y 19 (54.3) 45 (34.1) 0.029 2.296 1.078-4.890
   > 70 y 5 (71.4) 10 (35.7) 0.112 4.500 0.734-27.577
Mortality
   < 60 y 1 (10.0) 2 (2.1) 0.262 5.167 0.426-62.695
  60-70 y 2 (5.7) 0 (0) 0.043 0.943 0.869-1.023
   > 70 y 1 (14.2) 1 (3.6) 0.365 4.500 0.245-82.568
Group A:Higher blood glucose fluctuation group (≥4 mmol/L); Group B:Lower blood glucose fluctuation group (≥4 mmol/L). < 60 y in group B:n=10;60-70 y in group B:n=35; > 70 y in group B:n=7; < 60 y in group A:n=95;60-70 y in group A:n=132；> 70 y in group A:n=28.Others were as the same as Tab 3.
表选项
表 9 多元Logistic回归校正混杂因素后的术后不良反应事件总发生率 Tab 9 Total incidence of adverse events with multivariate Logistic regression correcting for confounding factors 
(n=308)
Factors β (parameter estimate) P OR 95%CI
Age -0.006 0.746 0.994 0.961-1.029
BMI 0.002 0.923 1.002 0.968-1.037
Sex 0.041 0.895 1.042 0.567-1.917
Other preoperative treatments -0.254 0.486 0.776 0.380-1.585
PMH
  Hypertension -0.475 0.089 0.622 0.359-1.075
  Cardiovascular disease -0.129 0.823 0.879 0.283-2.727
  Stroke 0.021 0.978 1.021 0.244-4.275
  Diabetes 0.2 0.679 1.221 0.475-3.139
  Glucose fluctuations 0.971 0.003 2.641 1.402-4.976
(1)Including chemoradiotherapy and ESD surgery.BMI:Body mass index; PMH:Previous medical history.
表选项

按年龄分层做风险评估对于60岁以下患者中呼吸功能衰竭致术后使用呼吸机这个结局，血糖波动是一个显著相关危险因素(P=0.044，OR=11.625，95%CI：1.440~93.871)；对于60~70岁患者中不良事件总发生率这个结局，血糖波动是一个中等相关危险因素(P=0.029，OR=2.296，95%CI：1.078~4.890)(表 10)。

表 10 多元Logistic回归校正混杂因素后的病死率 Tab 10 Mortality with multivariate Logistic regression correcting for confounding factors 
(n=308)
Factors β (parameter estimate) P OR 95%CI
Age -0.080 0.222 0.923 0.812-1.050
BMI 0.117 0.427 1.127 0.842-1.500
Sex -0.565 0.616 0.568 0.062-5.176
Other preoperative
treatmentsa		 0.435 0.714 1.546 0.150-15.891
Hypertension -1.028 0.233 0.358 0.066-1.936
Glucose fluctuations 2.020 0.012 7.539 1.548-36.713
aIncluding chemoradiotherapy and ESD surgery.BMI:Body mass index.
表选项
讨论

食管癌是我国恶性肿瘤中死亡顺位排名第4的恶性肿瘤，是预后较差的消化系统肿瘤[13]，通过外科技术、麻醉管理的提高及术后监护室的支持，有助于改善食管癌患者短期预后。其中一项重要组成部分就是围术期的血糖管理。虽有文献阐述围术期血糖值的异常(高血糖、低血糖和血糖波动)与患者预后不良的相关性[4, 9, 11, 14]，但尚缺乏食管癌方面的资料。

本研究同时纳入糖尿病患者和非糖尿病患者，其中7例住院期间死亡的患者均为非糖尿病患者。有研究在其他手术围术期也发现非糖尿病患者的风险高于糖尿病患者，并解释可能是由于实际临床工作中术前已知的糖尿病患者更容易得到关注及治疗[4]。故而，非糖尿病患者的围术期血糖管理同样值得我们重视。

关于早期高血糖与患者预后相关性的研究结果并不一致。Kutz等[11]的研究纳入了7 132名急诊科患者，认为早期高血糖与患者预后不良相关。van den Berghe等[15]的应用胰岛素控制血糖，认为严格控制危重患者的血糖能够降低发病率和死亡率，从反面论证了高血糖与预后的关系。而Smith等[16]关于高血糖与脑外伤预后的研究结果则显示，外伤后持续48 h以上的高血糖与预后差有关系，但早期高血糖与预后的相关性不强。本研究将患者按术后早期血糖分为正常血糖组与高血糖组，按照性别、年龄、糖尿病病史等因素分层进行风险评估，差异无统计学意义，与Smith等[16]的结果较一致。此结果提示，除注意单次高血糖外，更应重视患者是否存在持续性高血糖。

血糖波动的研究起始于一些热衷于用胰岛素控制高血糖以期达到改善预后的研究，在研究过程中意外发现血糖较大波动会增加死亡率和发病率[17]。有关血糖波动的临床研究中，患者多为急症或危重患者，重症患者的血糖控制目标已有报道，而择期手术患者中血糖波动的研究尚不多见，目前尚无理想的血糖控制目标，故采用适当宽松的血糖控制方案[12]。本研究发现以术前空腹血糖值为基础，术后早期血糖升高幅度较大(≥4 mmol/L)是术后短期不良事件总发生率、病死率的危险因素。在校正性别、年龄等混杂因素后，针对术后不良事件总发生率这个结局，术后血糖升高幅度较大是中等相关的危险因素，而针对病死率这个结局，则是强相关的危险因素。另外，我们发现术后早期血糖升高幅度较大是60岁以下患者术后发生呼吸功能衰竭而使用呼吸机的强相关危险因素。

根据Quagliaro等[6-7]的研究结论：在体外环境中稳定的高血糖与间歇性高血糖(即血糖波动)同样造成内皮细胞损伤，而间歇性高血糖环境中内皮功能障碍、氧化应激反应及凋亡表现更为明显。本研究选择食管癌根治术患者作为研究对象，手术创伤大、患者焦虑紧张、麻醉和术后疼痛等均可引起糖代谢异常[3]，促使血管内皮功能障碍，加速脂肪分解，产生过量的游离脂肪酸[4]，引起白细胞吞噬能力和伤口愈合速度明显下降等[5]，其与术后短期并发症及死亡有关。根据以上研究，我们认为术后早期血糖较术前基础血糖升高幅度较大的情况下，食管癌根治术后不良事件发生率更高，对于预测术后短期预后不良有一定价值，为下一步研究食管癌根治术患者围术期血糖控制目标，以减少术后患者血管内皮损伤等障碍及降低并发症发生率提供了临床依据。

本研究为回顾性队列研究，不足之处在于未能将术后胰岛素使用情况纳入。进一步研究需将实验设计为前瞻性研究，还需控制纳入病例的肿瘤位置、分期、分型及术者等影响预后的混杂因素，且可在术前检测糖化血红蛋白水平，发现隐匿性糖尿病患者，则结果将更为可靠。

参考文献
[1]
刘宝兴, 李印, 秦建军, 等. 胸腹腔镜联合与常规三切口食管次全切除术治疗食管癌的比较研究[J]. 中华胃肠外科杂志, 2012, 15(9): 938-942. [URI]
[2]
茅腾, 方文涛, 谷志涛, 等. 腔镜微创与开放食管癌根治术围手术期并发症和淋巴结清扫的比较研究[J]. 中华胃肠外科杂志, 2012, 15(9): 922-925. [URI]
[3]
BARTH E, ALBUSZIES G, BAUMGART K, et al. Glucose metabolism and catecholamines[J]. Crit Care Med, 2007, 35(9 Suppl): S508-S518.
[4]
DUNGAN KM, BRAITHWAITE SS, PREISER JC. Stress hyperglycaemia[J]. Lancet, 2009, 373(9677): 1798-1807. [DOI]
[5]
MC MANUS LM, BLOODWORTH RC, PRIHODA TJ, et al. Agonist-dependent failure of neutrophil function in diabetes correlates with extent of hyperglycemia[J]. J Leukoc Biol, 2001, 70(3): 395-404. [URI]
[6]
QUAGLIARO L, PICONI L, ASSALONI R, et al. Intermittent high glucose enhances apoptosis related to oxidative stress in human umbilical vein endothelial cells:the role of protein kinase C and NAD(P)H-oxidase activation[J]. Diabetes, 2003, 52(11): 2795-2804. [DOI]
[7]
QUAGLIARO L, PICONI L, ASSALONI R, et al. Intermittent high glucose enhances ICAM-1, VCAM-1 and E-selectin expression in human umbilical vein endothelial cells in culture:the distinct role of protein kinase C and mitochondrial superoxide production[J]. Atherosclerosis, 2005, 183(2): 259-267. [DOI]
[8]
WRIGHT RJ, FRIER BM. Vascular disease and diabetes:is hypoglycaemia an aggravating factor?[J]. Diabetes Metab Res Rev, 2008, 24(5): 353-363. [DOI]
[9]
KOTAGAL M, SYMONS RG, HIRSCH IB, et al. Perioperative hyperglycemia and risk of adverse events among patients with and without diabetes[J]. Ann Surg, 2015, 261(1): 97-103. [DOI]
[10]
JEREMITSKY E, OMERT LA, DUNHAM CM, et al. The impact of hyperglycemia on patients with severe brain injury[J]. J Trauma, 2005, 58(1): 47-50. [DOI]
[11]
KUTZ A, STRUJA T, HAUSFATER P, et al. The association of admission hyperglycaemia and adverse clinical outcome in medical emergencies:the multinational, prospective, observational TRIAGE study[J]. Diabet Med, 2017, 34(7): 973-982. [DOI]
[12]
中华医学会麻醉学分会. 围术期血糖管理专家共识(快捷版)[J]. 临床麻醉学杂志, 2016, 32(1): 93-95. [URI]
[13]
郑荣寿, 孙可欣, 张思维, 等. 2015年中国恶性肿瘤流行情况分析[J]. 中华肿瘤杂志, 2019, 41(1): 19-28. [URI]
[14]
RICHARDS JE, KAUFFMANN RM, ZUCKERMAN SL, et al. Relationship of hyperglycemia and surgical-site infection in orthopaedic surgery[J]. J Bone Joint Surg Am, 2012, 94(13): 1181-1186. [DOI]
[15]
VAN DEN BERGHE G, WOUTERS P, WEEKERS F, et al. Intensive insulin therapy in critically ill patients[J]. N Engl J Med, 2001, 345(19): 1359-1367. [DOI]
[16]
SMITH RL, LIN JC, ADELSON PD, et al. Relationship between hyperglycemia and outcome in children with severe traumatic brain injury[J]. Pediatr Crit Care Med, 2012, 13(1): 85-91. [DOI]
[17]
FINFER S, CHITTOCK DR, SU SY, et al. Intensive versus conventional glucose control in critically ill patients[J]. N Engl J Med, 2009, 360(13): 1283-1297. [DOI]

文章信息

王莉莉, 葛圣金, 金玲艳, 李倩倩
WANG Li-li, GE Sheng-jin, JIN Ling-yan, LI Qian-qian
术后早期血糖异常对食管癌根治术患者短期预后的预测价值
The value of abnormal early postoperative blood glucose concentration in predicting the short-term outcome of patients after radical resection of esophageal cancer
复旦学报医学版, 2020, 47(5): 715-722.
Fudan University Journal of Medical Sciences, 2020, 47(5): 715-722.
Corresponding author
GE Sheng-jin, E-mail:shengjinge@163.com.

工作空间