| Background and Objective:Laparoscopic pancreaticoduodenectomy(LPD)can be effectively used in the treatment of periampullary tumors.Comparing with open pancreaticoduodenectomy(OPD),LPD has the advantages of early postoperative recovery,minimal wound extension for cosmetic effect and less intraoperative bleeding.While LPD is still complicated and time-consuming and it still needs to be carried out in a high-volume center to really benefit the patients.Postoperative pancreatic fistula(POPF)is still the most important and serious complication after LPD,which can lead to postoperative bleeding,abdominal infection,delayed gastric emptying and even death.How to effectively predict and reduce the occurrence of POPF is the focus and hot topic of pancreatic surgeons.On the one hand,previous studies have reported that the occurrence of POPF is affected by many factors,among which the quality of pancreaticojejunostomy is also an important factor.The Bing’s anastomosis reported by our center is a kind of duct-to-mucosa anastomosis,which is simple,easy to learn,safe and reliable,and is more suitable in laparoscopic anastomosis.According to the previous experience,we still routinely placed the pancreatic duct stent during the operation.In the long-term follow-up,it was found that the pancreatic duct stent could cause some complications such as displaced and intestinal perforation.Therefore,we review the literature and find that there were some studies including randomized controlled trials(RCTs)conserons about the questions in open surgery,but the results are still controversial and no corresponding reports in laparoscopic surgery.Therefore,we tried to explore whether the placement of pancreatic duct stent during LPD can reduce the incidence of POPF and increase the perioperative safety by RCT.However,there are still some limitations in the single center study.We hope to further explore the real relationship between pancreatic duct stent and the occurrence of POPF and perioperative safety through meta-analysis,so as to better guide clinical practice.On the other hand,if we can effectively predict the occurrence of POPF and identify highrisk patients,we can remind clinical doctors of the danger,change management strategies,and improve the prognosis of patients.Previous studies have reported that the current more common prediction models,such as the fistula risk score(FRS),the alternate pancreatic fistula risk score(a-FRS)and ultra alternate fistula risk score(uaFRS).In order to improve the prediction efficiency and provide dynamic prediction conclusions,our team reviewed 251 LPDs from April 2016 to January 2019,and successfully established the dynamic prediction model of clinical relevant pancreatic fistula(CR-POPF),The prediction probabilities on the 2nd,3rd and 5th day are as follows:P=exp(-5.726-1.780[ALB≥38.1g/L]+1.810[Cr≥64μmol/L]+2.839[AMS≥642/U/L]+1.496[CAR≥5.88])/1+exp(-5.726-1.780[ALB≥38.1g/L]+1.810[Cr≥64μmol/L]+2.839[AMS≥642/U/L]+1.496[CAR≥5.88]),P=exp(-5.714-1.887[ALB≥38.1g/L]+3.328[Cr≥64μmol/L]+1.856[AMS≥409/U/L]+1.815[CAR≥7.69])/1+exp(-5.714-1.887[ALB≥38.1g/L]+3.328[Cr≥64/μmol/L]+1.856[AMS≥409IU/L]+1.815[CAR≥7.69])and P=exp(-4.992-1.97 9[ALB≥38.1g/L]+2.465[Cr≥64μmol/L]+2.002[AMS≥162/U/L]+1.967[CAR≥3.37])/1+exp(-4.992-1.97 9[ALB≥38.1g/L]+2.465[Cr≥64μmol/L]+2.002[AMS≥162/U/L]+1.967[CAR≥3.37]).We applied this model to the RCT cohort of this study,in order to prospectively verify the discrimination ability and stability of it and help ensuring the safety and smooth development of RCT.,Materials and methods:For patients undergoing LPD surgery in our hospitawe were were randomly divided divided into intervention group(with pancreatic duct stent placement)and control group(without one).The perioperative data of all patients were collected prospectively.The primary outcomes were the incidence of CR-POPF and 90-day mortality.For each enrolled patient in the RCT study,we used the dynamic prediction model established by our team to predict CR-POPF on the 2nd,3rd and 5th day after operation.We will use the C-index,Specificity/Sensitivity and Integrated discrimination improvement(IDI)of the old and new predicted values to observe the discrimination and stability of the dynamic model.Finally,we design meta-analysis of our and published RCTs about the pancreatic duct stent.Results:From November 2019 to October 2020,we included 74 patients with periampullary tumor,including 39 patients in the intervention group and 35 patients in the control group.The baseline levels of the two groups before operation,including age,sex ratio,pancreatic texture,pancreatic duct diameter and intraoperative blood loss were comparable,and there was no significant difference in the incidence of postoperative complications between the two groups.Three(7.7%)patients in the intervention group and in 1(2.9%)patient in the control group had CR-POPF,respectively,P=0.708.In the intervention group,1 patient died of heart failure within 90 days after operation,with a mortality rate of 2.6%;in the control group,there was no death within 90 days after operation,P=1.000.The median postoperative hospital stay of the two groups were 15.5(12.0-23.8)days and 16.0(13.0-22.0)days,respectively,P=0.953.In this cohort,a total of 66 cases with complete data could be verified by the dynamic model.The C-index at most time points indicated that the predictive value of the model had good discrimination performance(2 days after operation:0.674(0.308-1.000);3 days after operation:0.894(0.772-1.000);5 days after operation:0.781.000).The cut-off values at different time points were 5.4%,21.9%and 6.1%,respectively.The corresponding specificity and sensitivity were 57.7-93.1%and 66.7-100%,respectively.For the cohort with stent placement,the C-index and 95%CI of the predicted value of the model on the 2nd,3rd and 5th day after operation were 0.870(0.685-1.000),0.956(0.885-1.000)and 0.702(0.451-0.953),respectively.The corresponding cut-off points were the same as those of the whole cohort,and the corresponding specificity and sensitivity were 53.6-93.3%and 100%,respectively.In the subgroup without stent placement,the C-index showed that except for the second day after operation(0.304(0.067-0.541)),the model had good discrimination at other time points(3 and 5 days after operation)(3 days after operation:0.821(0.659-0.984);5 days after operation:0.958(0.878-1.000)),and the corresponding cut-off points of probability predictive value were 5.2%and 36.7%,respectively,with the specificity of 0.958(0.878-1.000)75%and 95.8%,and the sensitivity was 100%.The IDI of the new and old predictive values suggested that:for the whole cohort and the subgroup with stent placement,the new predictive values of the risk factors involved in the model had no significant change compared with the original predictive values(P>0.05);for the patients without stent placement,the new predictive values on the third day after operation were higher than the original predictive values.There was no significant difference in comprehensive discrimination(P>0.05).A total of 10 studies were included in the meta-analysis.In the comparison of the studies with or without stent placement,we found that stent placement could reduce the incidence of postoperative pancreatic fistula(RR=0.3755,95%CI[0.1709-0.8250],P=0.01),but had no effect on in-hospital mortality,total complication rate and length of stay.When comparing external drainage with internal drainage,we found that external drainage did not reduce the incidence of postoperative pancreatic fistula(RR=0.87,95%CI[0.551.39],P=0.57),and did not affect the in-hospital mortality,total complication rate and length of stay.Conclusion:It is safe and effective to perform pancreaticojejunostomy without pancreatic duct stent in a high-volume LPD center even for high-risk glands.The dynamic model is more suitable for patients with pancreatic duct stent during LPD operation,which can more accurately screen the high-risk population of CR-POPF,and the screening results are stable.On the 3rd and 5th day after operation,the dynamic model can predict the occurrence of CR-POPF in patients without pancreatic duct stent in LPD.Through meta-analysis,it is found that the placement of pancreatic duct stent during PD can reduce the risk of POPF,but it cannot reduce the in-hospital mortality,the incidence of total surgical complications and the length of hospital stay.The above results need to be verified by multi-center and large-sample trial. |
PDF Full Text Request