Objective: To examine the association of hyperglycemia and postoperative complications in non-diabetic patients underwent pancreaticoduodenectomy(PD). Methods: The clinical data of 209 non-diabetic patients who underwent PD from January 2012 to June 2018 at Department of Pancreatic and Biliary Surgery, the First Affiliated Hospital of Harbin Medical University and met the inclusion criteria were retrospectively analyzed. According to the diagnostic criteria of postoperative hyperglycemia, the patients were divided into postoperative hyperglycemia group (167 cases, 79.9%) and control group(42 cases, 20.1%). The propensity score matching(PSM) method was used to eliminate the difference between groups(caliper value=0.02; 38 cases in control group including 30 males and 8 females with age of 59.0 years;38 cases in postoperative hyperglycemia group including 32 males and 6 females with age of 61.0 years;37 cases of pancreatic head carcinoma,30 cases of periampullary carcinoma and 9 cases of benign diseases). A comparative analysis was applied for preoperative data, surgical related indicators and postoperative complication rates.The receiver operating characteristic(ROC) curve was used to calculate the area under the curve(AUC) of blood glucose values on postoperative day 1,3 and 5(POD1, POD3, POD5), to determine the high-risk blood glucose cutoff value of complications and to evaluate its sensitivity and specificity for the prediction of postoperative complications. Results: Univariate analysis showed that the differences in gender, body mass index, preoperative blood glucose, and serum urea nitrogen levels were statistically significant before PSM.There was no significant difference in the preoperative data between the two groups after PSM. Compared with the control group, the incidence of postoperative pancreatic fistula (31.6% vs. 5.3%), abdominal infection(29.0% vs. 7.9%) and Clavien-Dindo Ⅲ-Ⅴ complications(31.6% vs.7.9%) were statistically different(χ(2)=7.092,P=0.008; χ(2)=4.290,P=0.038; χ(2)=5.316,P=0.021), respectively. According to the AUC on POD3,the blood glucose value ≥8.860 mmol/L was an independent risk factor for pancreatic fistula with sensitivity of 58.3% and specificity of 76.9%,the blood glucose value ≥9.130 mmol/L was an independent risk factor for abdominal infection with sensitivity of 54.5% and specificity of 81.5% and the blood glucose value ≥7.685 mmol/L was independent risk factor of Clavien-Dindo Ⅲ-Ⅴ complications with sensitivity of 75.0% and specificity of 57.7%. Conclusions: Postoperative hyperglycemia in non-diabetic patients is associated with postoperative pancreatic fistula, abdominal infection, and Clavien-Dindo Ⅲ-Ⅴ complications.According to the early postoperative blood glucose value,the occurrence of postoperative pancreatic fistula, abdominal infection and Clavien-Dindo Ⅲ- Ⅴ complications can be effectively predicted.
目的： 探讨非糖尿病患者胰十二指肠切除术（PD）术后高血糖与术后并发症的关系。方法： 回顾性分析2012年1月至2018年6月于哈尔滨医科大学附属第一医院胰胆外科行PD且符合纳入和排除标准的209例非糖尿病患者的临床资料。参照术后高血糖相关诊断标准，将患者分为术后高血糖组（167例，79.9%）和对照组（42例，20.1%）。应用倾向性评分匹配（PSM）的方法消除组间差异（卡钳值=0.02；对照组38例，男性30例，女性8例，年龄59.0岁；术后高血糖组38例，男性32例，女性6例，年龄61.0岁；胰头癌37例，壶腹周围癌30例，良性疾病9例）。比较分析两组的术前资料、手术相关指标和术后并发症发生率等。采用受试者工作特征（ROC）曲线分别计算术后第1、3、5天血糖值的曲线下面积（AUC），确定并发症发生高风险的血糖截断值，并评价其对术后并发症预测的灵敏度和特异度。结果： 倾向性评分匹配前单因素分析结果显示，两组患者性别、体重指数、术前血糖、血清尿素氮水平方面的差异均有统计学意义（P值均<0.05）。倾向性评分匹配后两组患者一般资料的差异均无统计学意义（P值均>0.05），但术后高血糖组和对照组在术后胰瘘发生率（31.6%比5.3%，χ(2)= 7.092，P=0.008）、腹腔感染发生率（29.0%比7.9% ，χ(2)=4.290，P=0.038）、Clavien-Dindo Ⅲ~Ⅴ级并发症发生率（31.6%比7.9%，χ(2)=5.316，P=0.021）方面的差异均有统计学意义。依据非糖尿病患者PD术后第3天ROC的AUC可知，血糖值≥8.860 mmol/L是胰瘘发生的独立预后因素，其灵敏度为58.3%、特异度为76.9%；血糖值≥9.130 mmol/L是腹腔感染发生的独立预后因素，其灵敏度为54.5%、特异度为81.5%；血糖值≥7.685 mmol/L是Clavien-Dindo Ⅲ~Ⅴ级并发症发生的独立预后因素，其灵敏度为75.0%，特异度为57.7%。结论： 非糖尿病患者术后高血糖与术后胰瘘、腹腔感染、Clavien-Dindo Ⅲ~Ⅴ级并发症的发生率相关。依据术后早期血糖水平可预测胰瘘、腹腔感染等并发症的发生，对甄别高危患者并采取积极干预措施具有重要的临床意义。.
Keywords: Hyperglycemia; Pancreatoduodenectomy; Postoperative complications; Propensity score matching.