Comparison of surgical outcomes between isolated pancreaticojejunostomy, isolated gastrojejunostomy, and conventional pancreaticojejunostomy after pancreaticoduodenectomy: a systematic review and meta-analysis

Background We aimed to compare the safety and effectiveness of the following procedures after pancreaticoduodenectomy: isolated pancreaticojejunostomy, isolated gastrojejunostomy, and conventional pancreaticojejunostomy. Methods We performed a systematic search of the following databases: PubMed, Embase, Web of Science, the Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov until 1 January 2020. Pooled odds ratios (OR) or weighted mean differences (WMD) with 95% confidence intervals (CIs) were calculated using STATA 12.0 statistical software. Results Thirteen studies involving 1942 patients were included in this study. Pooled analysis showed that reoperation rates following isolated pancreaticojejunostomy were lower reoperation than with conventional pancreaticojejunostomy (OR = 0.36, 95% CI: 0.15–0.86, p = 0.02, respectively), and that isolated pancreaticojejunostomy required longer operation time vs conventional pancreaticojejunostomy (WMD = 43.61, 95% CI: 21.64–65.58, P = 0.00). Regarding postoperative pancreatic fistula, clinically-relevant postoperative pancreatic fistula, delayed gastric emptying, clinically-relevant delayed gastric emptying, bile leakage, hemorrhage, reoperation, length of postoperative hospital stay, major complications, overall complications, and mortality, we found no significant differences for either isolated pancreaticojejunostomy versus conventional pancreaticojejunostomy or isolated gastrojejunostomy versus conventional pancreaticojejunostomy. Conclusions This study showed that isolated pancreaticojejunostomy was associated with a lower reoperation rate, but required longer operation time vs conventional pancreaticojejunostomy. Considering the limitations, high-quality randomized controlled trials are required.


Background
Pancreaticoduodenectomy, one of the most complex intra-abdominal operations, is widely used for benign and malignant disease located in the pancreatic head or periampullary region [1,2]. Despite developments in surgical techniques, pancreaticoduodenectomy is still accompanied by a high postoperative complication rate of 40-50% [3]. Previous studies demonstrated that the most common complications after pancreaticoduodenectomy were postoperative pancreatic fistula (POPF) and delayed gastric emptying (DGE) [3,4]. Several methods of digestive tract reconstruction have been recommended to reduce the main postoperative complications, namely conventional pancreaticojejunostomy, isolated pancreaticojejunostomy [5,6], and isolated gastrojejunostomy [7,8]. Isolated pancreaticojejunostomy, first described in 1976, was proposed to reduce complications such as POPF, based on the theory of separating bile and pancreatic enzymes [5]. In isolated gastrojejunostomy, a second loop is made to perform the gastroenteric anastomosis, which may favor digestive transit by separating the pancreatic enzymes and gastric sutures [9]. However, the debate regarding these three reconstructions is on-going. Some studies demonstrated that isolated pancreaticojejunostomy and isolated gastrojejunostomy may be associated with less postoperative complications, such as POPF [10] and DGE [11]. Kaman et al. suggested that morbidity and mortality could be reduced using isolated pancreaticojejunostomy to separate the bile from pancreatic secretions [12]; however, other studies reached different conclusions. Furthermore, studies comparing isolated pancreaticojejunostomy versus conventional pancreaticojejunostomy and isolated gastrojejunostomy versus conventional pancreaticojejunostomy involved low numbers of patients or had retrospective designs. With the recent publication of several new studies, we performed this systematic review and meta-analysis to compare the surgical outcomes of isolated pancreaticojejunostomy, isolated gastrojejunostomy, and conventional pancreaticojejunostomy.

Search strategy
We performed a systematic review and meta-analysis according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement. Two authors independently searched the electronic databases of PubMed, Embase, Web of Science, Cochrane Central Register of Controlled Trials (CENTRAL), and Clinical-Trials.gov. Published trials comparing the efficacy and safety of isolated pancreaticojejunostomy, isolated gastrojejunostomy, and conventional pancreaticojejunostomy after pancreaticoduodenectomy were evaluated in this study. We used the following English search terms: "pancreaticoduodenectomy," "pancreatoduodenectomy," "Whipple," "pylorus-preserving pancreaticoduodenectomy," "pancreaticojejunostomy," "Rou-en-Y," and "isolated Roux loop gastrojejunostomy." The search was restricted to human subjects, available full text, and Englishlanguage articles. The references of the articles identified after the initial search were also manually reviewed.

Inclusion and exclusion criteria
Studies were included based on the following criteria: (1) trials had to compare isolated pancreaticojejunostomy or isolated gastrojejunostomy versus conventional pancreaticojejunostomy in patients undergoing pancreaticoduodenectomy or pylorus-preserving pancreaticoduodenectomy; and (2) complete data were provided in English with available full text. Reviews, conference abstracts, and studies with unavailable full text were excluded.

Outcome measures
The analyzed outcome measures were POPF, clinicallyrelevant POPF (CR-POPF), DGE, clinically-relevant DGE (CR-DGE), operation time, bile leakage, hemorrhage, reoperation, length of postoperative hospital stay, major complications, overall complications, and mortality. The definition of POPF, CR-POPF, DGE and CR-DGE was according to the criteria of International Study Group for Pancreatic Surgery (ISGPS) [13,14]. Major complications were defined as Clavin-Dindo grade ≥ IIIa [15]. Isolated pancreaticojejunostomy was created to prevent pancreatic fistula by separating pancreatic juice from bile juice after pancreaticoduodenectomy. Isolated gastrojejunostomy was created to prevent delayed gastric emptying after pancreaticoduodenectomy. Then, subgroup analyses were conducted depending on the different construction method.

Data extraction and quality assessment
Data extraction was performed by two independent authors using a standardized selection form that included the first author, year of publication, type of study, country in which the study was performed, type of reconstruction, and general data. Conflicts in data abstraction were resolved by consensus and by referring to the original article. EndNote version X8 (Thomson Reuters, Toronto, ON, Canada) was used to remove duplicate studies. The methodological quality of all included studies was assessed using the validated Newcastle-Ottawa scale [16]. Studies scoring > 7 were considered of high quality.

Statistical analysis
All statistical analyses were performed using STATA/SE 12.0 (Stata Corp., College Station, TX, USA). We used odds ratios (OR) with 95% confidence intervals (CIs) for dichotomous outcomes, and weighted mean difference (WMD) with 95% CIs for continuous variables. A twotailed p value < 0.05 was considered statistically significant. Heterogeneity among studies was evaluated using the χ 2 test; values < 25, 25-50, and > 50 were classified as low, moderate, and high heterogeneity, respectively, and were treated as binary data. We created funnel plots and performed Egger's test [17] to evaluate the risk of publication bias. Sensitivity analyses were performed by removing individual studies from the data set and analyzing the effect on the overall results, to identify sources of signficant heterogeneity. This type of analysis is called "Jackknife analysis", named by John Tukey, which can improvise a solution for a variety of problems [18].

Study selection and trial characteristics
According to the search strategy, we identified 1563 studies. Of these, 420 duplicate articles were excluded, and we retrieved the remaining 1143 studies based on their titles and abstracts. After excluding irrelevant articles for various reasons, we included a final 14 trials involving a total of 2043 patients [6,[9][10][11][12][19][20][21][22][23][24][25][26][27]. A flowchart of the literature search process is shown in Fig. 1, and the characteristics and quality evaluation of the included studies are shown in Table 1. There were four randomized controlled trials (RCTs), one prospective study, and eight retrospective studies. The isolated pancreaticojejunostomy group comprised 482 patients, the isolated gastrojejunostomy group comprised 92 patients, and the isolated pancreaticojejunostomy + gastrojejunostomy group comprised 112 patients. The sample sizes among the studies ranged from 40 to 700 patients.

Bile leakage
Eleven studies provided data for the incidence of bile leakage, with 945 patients in the isolated pancreaticojejunostomy versus conventional pancreaticojejunostomy group and 259 patients in the isolated gastrojejunostomy versus conventional pancreaticojejunostomy group. We found no significant difference for the incidence of bile leakage between isolated pancreaticojejunostomy and conventional pancreaticojejunostomy (OR = 0.68, 95% CI: 0.27-1.69; P = 0.40) or between isolated gastrojejunostomy and conventional pancreaticojejunostomy (OR = 0.67, 95% CI: 0.28-1.63; P = 0.38) (Fig. 3a).

Publication bias and sensitivity analysis
The funnel plots for the parameters were symmetrical, and Egger's test revealed no significant publication bias. Sensitivity analyses were performed by removing individual studies from the data and analyzing the effect on the overall results. However, these exclusions did not alter the results.

Discussion
This meta-analysis compared isolated pancreaticojejunostomy and isolated gastrojejunostomy with conventional pancreaticojejunostomy after pancreaticoduodenectomy. Our results showed that isolated pancreaticojejunostomy was associated with fewer major complications and lower reoperation rates, but required longer operation time versus conventional pancreaticojejunostomy. However, the rates for overall complications, POPF, CR-POPF, DGE, CR-DGE, bile leakage, and hemorrhage, and the length of postoperative hospital stay and mortality rates with isolated pancreaticojejunostomy versus isolated gastrojejunostomy were similar to rates for conventional pancreaticojejunostomy. Considering the limitations, future high-quality RCTs are required. POPF, one of the most frequent complications after pancreaticoduodenectomy, is associated with intra-abdominal abscess, sepsis, and life-threatening hemorrhage. Many methods have been used to decrease the incidence of POPF such as using fibrin [28] or pancreatic stenting [29], and modifying the jejunal anastomosis [30]; however, the optimal technique is still debated. Our study revealed no significant difference between isolated pancreaticojejunostomy and conventional pancreaticojejunostomy, which was consistent with most previous studies. However, some studies demonstrated that isolated pancreaticojejunostomy was associated with lower rates of POPF after pancreaticoduodenectomy compared with conventional pancreaticojejunostomy [20,31]. Several reasons were revealed in these studies, including the separation of bile acids and pancreatic enzymes [5] and decreasing the reflux of bile into the pancreas [32]; however, these advantages were theoretical, and comparative studies are lacking. Of note, adding jejunojejunal anastomosis in isolated pancreaticojejunostomy could increase intestinal intraluminal pressure, which may affect the pancreaticojejunostomy anastomosis [19]. An RCT performed by Ke et al. showed that the rate of Grade B POPF in the isolated pancreaticojejunostomy group was higher than that in the conventional pancreaticojejunostomy group [6]. In addition to the 13 studies involving 1942 patients, our study showed that isolated pancreaticojejunostomy provides no advantage over conventional pancreaticojejunostomy regarding CR-POPF. Similar to previous studies evaluating isolated gastrojejunostomy, the incidences of POPF and CR-POPF were similar to those with conventional pancreaticojejunostomy.
DGE is also one of the most frequent complications after pancreaticoduodenectomy, with rates ranging from 13.5 to 40% [14,33]. Several surgical reconstruction procedures have been proposed to decrease the incidence of DGE, namely the Billroth I procedure, Braun enteroenterostomy, and isolated gastrojejunostomy. However, few studies have compared isolated gastrojejunostomy and conventional pancreaticojejunostomy. The debate regarding isolated gastrojejunostomy versus conventional pancreaticojejunostomy is on-going. Regarding DGE and CR-DGE, our results revealed no significant difference between isolated gastrojejunostomy and conventional pancreaticojejunostomy, and we found similar results when comparing isolated pancreaticojejunostomy and conventional pancreaticojejunostomy. A previous metaanalysis involving three studies revealed that conventional pancreaticojejunostomy was associated with lower rates of DGE versus isolated gastrojejunostomy [34]; however, the sample size in the study was small. In the current meta-analysis, we showed that the incidences of DGE and CR-DGE were similar for both isolated pancreaticojejunostomy and conventional pancreaticojejunostomy, indicating that the activation of pancreatic enzymes does not influence the occurrence of DGE.
Regarding major complications, our meta-analysis demonstrated that isolated pancreaticojejunostomy has comparable with conventional pancreaticojejunostomy. The definition of major complications in our included studies varied. Applied with Clavin-Dindo grade, there were three studies provided the data of major complications. Our study showed that there was no significantly difference between isolated pancreaticojejunostomy and conventional pancreaticojejunostomy. Interestingly, isolated pancreaticojejunostomy decreases the incidence of reoperation in this study. The study conducted by Aghalarov et al and Chhaidar et al showed that isolated pancreaticojejunostomy has less reoperation [10,20]. They showed that the reoperation was largely because of POPF-related complications. Nevertheless, there are many factors that affect the occurrence of reoperation after surgery, such as bleeding, gastrointestinal anastomotic leakage. Additionally, with the development of percutaneous drainage and other procedures, there was fewer reoperation. However, there was lack of enough data about the detail of reoperations which may lead to bias in this present study.

Conclusion
In conclusion, isolated pancreaticojejunostomy was associated with lower reoperation rates, but required longer operation times versus conventional pancreaticojejunostomy. The rate of major complications, overall complications, POPF, CR-POPF, DGE, CR-DGE, bile leakage, and hemorrhage, and the length of postoperative hospital stay and mortality rates with isolated pancreaticojejunostomy and isolated gastrojejunostomy were similar to the respective rates with conventional pancreaticojejunostomy. However, further randomized controlled trials are needed.