Analysis of value with early multi-indicators combined prediction of pancreatic fistula after pancreaticoduodenectomy in non-diabetic patients
10.3760/cma.j.issn.1673-4203.2019.09.005
- VernacularTitle: 非糖尿病患者胰十二指肠切除术后胰瘘早期多指标联合预测价值分析
- Author:
Hongtao TAN
1
;
Yue GAO
;
Meiling LIANG
;
Dexing GUO
;
Zhongqiang ZUO
;
Bei SUN
;
Hongchi JIANG
Author Information
1. Department of Pancreatic and Biliary Surgery, First Affiliated Hospital of Harbin Medical University, Harbin 150001, China
- Publication Type:Journal Article
- Keywords:
Pancreatoduodenectomy;
Postoperative complications;
Pancreatic fistula;
Procalcitonin;
Blood glucose;
C-reactive protein
- From:
International Journal of Surgery
2019;46(9):590-595,封3-1
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the relationship between procalcitonin (PCT), C-reactive protein (CRP) and blood glucose (BG) levels and postoperative pancreatic fistula (POPF) in non-diabetic patients after pancreaticoduodenectomy (PD), and evaluate the clinical value of early multi-indicators combined prediction of pancreatic fistula.
Methods:The clinical data of 68 non-diabetic patients who underwent PD surgery from April 2016 to June 2018 in the First Affiliated Hospital of Harbin Medical University were retrospectively analyzed, including 42 males and 26 females, with average age of 58 years, age range from 26 to 80 years. According to the diagnostic criteria of POPF, the patients were divided into the pancreatic fistula group (n=17) and the non-pancreatic fistula group (n=51). The surgical related indicators, health economic indicators and PCT, CRP, and BG levels on preoperative and postoperative day 1, 3 and 5(POD1, POD3, POD5) were compared between two groups. Multivariate logistic regression analysis was used to screen out independent risk factors associated with POPF, and a weighted predictive model (wScore M) for predicting pancreatic fistula in non-diabetic patients was established by combining independent risk factors. The receiver operating characteristic (ROC) curves of each independent risk factor and prediction model were plotted to determine the cutoff value, area under the ROC curve (AUC), sensitivity, specificity, positive predictive value, negative predictive value, and individual risk factors were compared. And the predictive value of the multi-indicator combined prediction model.
Results:Univariate analysis showed that the pathological types of the tumor, the tumor size, the blood glucose level of POD3, the PCT level of POD1, POD3, POD5 and the CRP level of POD3, POD5 were statistically different (P<0.05). Multivariate analysis showed that PCT (P=0.004), CRP (P=0.031), and BG (P=0.005) levels of POD3 were independent risk factors of POPF. The AUC of the ROC were 0.967, 0.692, and 0.698, respectively. The sensitivity were 0.588, 0.706, and 0.647, respectively. The specificities were 0.863, 0.686, and 0.765, respectively. The positive predictive values were 52.9%, 70.6%, and 64.7%, respectively. The negative predictive values were 88.2%, 70.6%, and 76.5%; the AUC of the wScore M ROC curve was 0.877, the sensitivity was 0.706, the specificity was 0.863, the positive predictive value was 70.6%, and the negative predictive value was 86.3%. Compared with the non-pancreatic fistula group, the patients of the pancreatic fistula group with pancreatic fistula complications had prolonged hospitalization time and increased the total hospitalization cost, which were statistically different (P=0.039, 0.010).
Conclusion:PCT, CRP and BG are independent risk factors for POPF, and early combined prediction of multiple indicators (POD3)has good predictive value, which has significant clinical reference for optimal and effective interventions at early stage.
