Table of Contents Author Guidelines Submit a Manuscript
International Journal of Surgical Oncology
Volume 2017 (2017), Article ID 7526494, 18 pages
https://doi.org/10.1155/2017/7526494
Review Article

Pancreatogastrostomy versus Pancreatojejunostomy: An Up-to-Date Meta-Analysis of RCTs

1Department of Surgery, University Hospital of Larissa, Mezourlo, 41110 Larissa, Greece
2Postgraduate Programme (MSc): Research Methodology in Biomedicine, Biostatistics and Clinical Bioinformatics, University of Thessaly, Larissa, Greece
3Research Unit for General Medicine and Primary Health Care, Faculty of Medicine, School for Health Sciences, University of Ioannina, Ioannina, Greece
4Tufts University School of Medicine, Boston, MA, USA
5Department of Nephrology, Medical School, University of Thessaly, Larissa, Greece
6Department of Biomathematics, University of Thessaly School of Medicine, Larissa, Greece

Correspondence should be addressed to Konstantinos Perivoliotis; moc.liamg@91irepk

Received 14 January 2017; Revised 22 April 2017; Accepted 14 May 2017; Published 17 July 2017

Academic Editor: George H. Sakorafas

Copyright © 2017 Konstantinos Perivoliotis et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background. A meta-analysis was conducted in order to provide an up-to-date comparison of pancreatogastrostomy (PG) and pancreatojejunostomy (PJ), after pancreatoduodenectomy (PD), in terms of clinically significant postoperative pancreatic fistula (POPF) and other postoperative complications. Methods. This meta-analysis was conducted according to the PRISMA guidelines and the Cochrane Handbook for Systematic Reviews of Interventions. A systematic literature search in MEDLINE and Cochrane Central Register of Controlled Clinical Trials was performed. Fixed Effects or Random Effects model was used, based on the Cochran test. Results. In total, 10 studies (1629 patients) were included. There was no statistical significance between PG and PJ regarding the rate of clinically significant POPF (OR: 0.70, 95%CI: 0.46–1.06). PG was associated with a higher rate of postpancreatoduodenectomy haemorrhage (PPH) (OR: 1.52, 95%CI: 1.08–2.14). There was no difference between the two techniques in terms of clinically significant PPH (OR: 1.35, 95%CI: 0.95–1.93) and clinically significant postoperative delayed gastric emptying (DGE) (OR: 0.98, 95%CI: 0.59–1.63). Discussion. There is no difference between the two anastomotic techniques regarding the rate of clinically significant POPF. Given several limitations, more large scale high quality RCTs are required.

1. Introduction

1.1. Rationale

Pancreatoduodenectomy (PD) is still the gold standard of treatment for patients with resectable benign and malignant lesions of the head of the pancreas and the periampullary region. Although PD is considered a safe operative technique, with 30-day mortality rates in specialized, high volume centers currently estimated below 3% [1, 2], complications, such as postoperative pancreatic fistula (POPF), delayed gastric emptying (DGE), and postpancreatoduodenectomy haemorrhage (POPH), increase the overall morbidity to the rate of 45%, despite the application of enhanced recovery approaches after surgery [3].

Given the fact that the frequency of POPF, the most notorious postpancreatoduodenectomy complication, remains as high as 40% [4], researchers have focused on factors that may influence this rate, with the pancreatoenteric anastomosis being one of them. The anastomosis between the pancreatic stump and the GI is regarded as prone to leakage, due to exposure of the suture line to pancreatic juice. The two most widely adopted postpancreaticoduodenectomy anastomotic techniques are the pancreatogastrostomy (PG) and the pancreatojejunostomy (PJ), which combined with anastomotic reinforcing techniques, such as glue and intraductal stenting, are designed to provide a sealed and stable pancreatoenteric junction. In the current literature, a series of retrospective and prospective studies [510] have compared PG and PJ with inconclusive results. Keck et al. [11], in a large multicenter randomized controlled trial, reported no difference between the two techniques in terms of clinically significant POPF, which is in contrast with results from previous meta-analyses [1214], where it was suggested that PG was a safer and more effective method of reconstruction, with lower rates of POPF and other intra-abdominal complications and shorter length of hospital stay (LOS).

1.2. Objectives

In light of these conflicting evidences, we conducted a meta-analysis, in order to provide an up-to-date comparison of PG and PJ after PD, for benign or malignant diseases of the head of the pancreas and the periampullary region, in terms of clinically significant POPF and other postoperative complications.

2. Methods

2.1. Study Protocol

The conduction of this meta-analysis was completed according to the PRISMA [15] guidelines and the Cochrane Handbook for Systematic Reviews of Interventions. The present study was not registered in any database.

2.2. Primary Endpoint

The primary endpoint of this study was the rate of clinically significant postoperative pancreatic fistula (grade B/C according to ISGPF). POPF was defined by ISGPF [16] as a drain output of any measurable volume of fluid on or after POD 3 with an amylase content > 3 times the serum amylase activity. Classification to grades A, B, and C is based on the impact of POPF on the overall clinical course.

2.3. Secondary Endpoints

Secondary endpoints included overall postoperative POPF, postoperative delayed gastric emptying (DGE) [17], clinically significant DGE (grade B/C), postpancreatectomy haemorrhage (PPH) [18], clinically significant PPH (grade B/C), biliary fistula, intra-abdominal fluid collection, overall morbidity, mortality, reoperation rate, wound infection, intraoperative blood transfusion, operative time, and the length of hospital stay (LOS).

2.4. Eligibility Criteria

Eligible trials were prospective human studies with a RCT design, comparing PG and PJ after PD for benign or malignant diseases of the head of the pancreas and the periampullary region, whose outcome data were reported in English and could be retrieved. Excluded studies included those not written in English or studies with no outcome of interest and no comparison group and observational, nonhuman, or nonrandomized studies. Moreover, studies reported in the form of editorials, letters, conference abstracts, expert opinion, or duplicate studies were excluded.

2.5. Literature Search

A systematic literature search in electronic databases (MEDLINE and Cochrane Central Register of Controlled Clinical Trials) was performed (search date: 20 July 2016) in order to identify the eligible RCTs.

In order to perform the literature search the following keywords were used:(i)MEDLINE: (Pancreaticoduodenectomy OR Pancreatoduodenectomy OR Whipple OR “pancreatoduodenal resection” OR “pancreaticoduodenal resection” OR pancreaticojejunostomy OR pancreatojejunostomy OR “pancreaticoenteric anastomosis” OR “pancreatoenteric anastomosis” OR pancreaticogastrostomy OR pancreatogastrostomy OR “pancreaticogastric anastomosis” OR “pancreatogastric anastomosis” OR “pancreaticojejunal anastomosis” OR “pancreatojejunal anastomosis”) AND (“Clinical Trials as Topic” OR “randomized controlled trial” OR “controlled clinical trial” OR randomized OR placebo OR randomly OR trial)(ii)Cochrane Central Register of Controlled Clinical Trials (Wiley): (Pancreaticoduodenectomy OR Pancreatoduodenectomy OR Whipple OR “pancreatoduodenal resection” OR “pancreaticoduodenal resection” OR pancreaticojejunostomy OR pancreatojejunostomy OR “pancreaticoenteric anastomosis” OR “pancreatoenteric anastomosis” OR pancreaticogastrostomy OR pancreatogastrostomy OR “pancreaticogastric anastomosis” OR “pancreatogastric anastomosis” OR “pancreaticojejunal anastomosis” OR “pancreatojejunal anastomosis”)

2.6. Study Selection and Data Collection

After duplicate removal, titles and abstracts of the studies were screened according to eligibility criteria. The next step included the full text review of the articles in order to assess that they are consistent with the inclusion criteria.

All electronic database search, study selection, data extraction, and methodological assessment of the studies were performed blindly and in duplicate by two independent investigators (PK and SE). Disagreements were resolved by mutual revision and discussion, in order to reach a consensus. In case of not resolving the discrepancies, the opinion of a third investigator (TA) was considered.

From all eligible studies, the data extracted included author’s name, study location and year, RCT type, sample size, the age and gender of the participants, primary outcome, follow-up duration, overall morbidity, underlying disease, operation type, rate of PD/pylorus preserving PD (PPPD), anastomotic technique, operative time, postoperative hospital stay, use of intraductal stent, glue and drains, postoperative administration of somatostatin, and information regarding the diameter of pancreatic duct and the texture of pancreas. Only results reported in the article of the studies were extracted.

All studies imported in this meta-analysis were submitted to rigorous quality and methodological evaluation for bias appraisal according to Cochrane’s risk of bias assessing tool [19]. Validity checkpoints included assessment of random sequence allocation, allocation concealment, blinding of participants and personnel and blinding of outcome assessment, incomplete outcome data, and selective reporting. Cohen’s statistic was also calculated.

2.7. Statistical Analysis

Data analysis was performed using the Cochrane Collaboration RevMan version 5.3. Dichotomous variables were reported in the form of Odds Ratio (OR), while for continuous variables Weighted Mean Differences (WMD) were used. Results of the analyses were presented with the corresponding 95% Confidence Interval (95% CI).

In the case of continuous variables, if the article did not provide the mean and the Standard Deviation (SD), these were calculated from the median and the Interquartile Range (IR), based on the formula by Hozo et al. [20]. To be more specific, if the sample size was >25, then the mean was considered equal to the median. For sample sizes <70, SD was regarded as IR/4. If the sample size was >70, then SD was equal to IR/6. For dichotomous variables, the statistical method used was the Mantel-Haenszel (MH) and for continuous variables the Inverse Variance (IV). Both Fixed Effects (FE) and Random Effects (RE) model were calculated and reported. The decision of which model to finally estimate was based on the Cochran test. If statistically significant heterogeneity was present ( test ), then RE model was applied. Moreover, heterogeneity was quantified with the use of . The studies were weighted on the basis of sample size. Statistical significance was considered at the level of .

2.8. Risk of Bias across Studies

The funnel plot of the primary outcome was also visually inspected, in order to determine the possible presence of publication bias. An Egger’s test was also performed for the primary outcome.

3. Results

3.1. Study Selection

From the literature search, 1240 citations (Figure 1) were retrieved, published up to 20 July 2016. After the removing of 236 duplicate records, the screening of the titles and the abstracts begun. From the 1004 studies submitted to the first phase of the screening, 993 were excluded. More specific, 10 were comments or conference abstracts, 5 did not have a RCT design, 5 did not have a comparison group, 18 were reviews of the current literature, 20 were meta-analysis, 3 articles were not written in English, 23 compared different techniques of PG or PJ instead, and 909 were irrelevant to the subject records. In full text review, 11 articles were submitted [9, 11, 2129]. At this step, 1 trial [9] was rejected due to a no RCT design. Finally, 10 studies [11, 2129] were included in qualitative and quantitative analysis.

Figure 1: Study flow diagram.
3.2. Study Characteristics

Table 1 summarizes the characteristics of the included studies. The publication date ranges from 1995 up to 2016. Four studies were multicentered while the other six were single-centered. Fernàndez-Cruz et al. [24] were the first to adopt the ISGPS definition and classification of POPF. Since then, heterogeneity existed in the definition and diagnosis of POPF. The overall amount of patients included in this meta-analysis is 1629 (Table 2). A total of 826 PGs and 803 PJs were performed. The age of the participants extended from 12 to 87 years. Regarding the gender allocation between the two comparison groups, data are shown in Table 2. El Nakeeb et al. [23] compared the results of PG and an isolated Roux loop pancreatojejunostomy while Fernàndez-Cruz et al. [24], respectively, compared PJ and PG with gastric partition. In the rest of the studies, PG was considered the intervention and PJ the control. All studies, except Duffas et al. [22], had the rate of POPF as primary outcome. Four studies [21, 24, 26, 29] did not report the duration of follow-up. In the other six studies, follow-up varied from 30 days to 12 months. Regarding the underlying disease, carcinoma of the pancreatic head was the most frequent (Table 3). The PD and PPPD ratio is shown in Table 3. There was a lack of uniformity between the studies regarding the technique of PG and PJ anastomoses. Both PG and PJ could be performed in either a telescoped or a duct-to-mucosa manner. Table 4 reports a summary of the studies implementing the use of stents in the pancreatic duct, anastomotic glue reinforcement, and the overall drain use. Postoperative octreotide was administered in 7 studies [2123, 2528]. All studies reported data regarding the main pancreatic duct diameter. Similarly, only Topal et al. [27] did not provide the allocation of the patients regarding pancreatic texture.

Table 1: Included studies.
Table 2: Study characteristics.
Table 3: Operative characteristics.
Table 4: Intraoperative characteristics.
3.3. Risk of Bias within Studies

Figure 2 represents a summary of the included studies quality assessment. More specifically, as shown in Figure 3, all studies included a random sequence generation procedure in their protocol. Allocation concealment was also applied in all studies except one [29]. Only two trials [11, 22] reported the blinding of participants and personnel and the blinding of outcome assessment. Only in the study of Grendar et al. [26], incomplete outcome data and possible selective reporting were detected. There was almost perfect agreement between the two investigators (Cohen’s statistic: 82.3%  ).

Figure 2: Risk of bias graph: review authors’ judgments about each risk of bias item presented as percentages across all included studies.
Figure 3: Risk of bias summary: review authors’ judgments about each risk of bias item for each included study.
3.4. Primary Endpoint

(i)All ten studies (Figure 4(a)) compared the two anastomotic techniques regarding the clinically significant POPF. More specifically, 108 patients from a total of 826 in the PG group developed clinically significant POPF, whereas in the PJ group the same ratio was 144/803. Meta-analysis of these data showed no statistically significant () difference between the two groups regarding clinically significant POPF (OR: 0.70, 95% CI: 0.46–1.06). Since there was significant heterogeneity between the studies ( test : 0.04, : 48% (95% CI: 0–75%)), a RE model was applied. Estimation of FE model did not wield consistent results (OR: 0.68, 95% CI: 0.51–0.89) with the RE model.

Figure 4: (a) Clinically significant postoperative pancreatic fistula, (b) postoperative pancreatic fistula, (c) delayed gastric emptying, and (d) clinically significant delayed gastric emptying.
3.5. Secondary Endpoints

(i)All the included studies (Figure 4(b)) provided comparison between the two anastomotic techniques regarding POPF. In summary, 138 patients from a total of 826 submitted to PG developed POPF, instead of 175 and 803, respectively, in the PJ group. Meta-analysis of these data showed a statistically significant () lower ratio of POPF (OR: 0.71, 95% CI: 0.55–0.91) for the PG group. Since there was no significant heterogeneity between the studies ( test : 0.27, : 19% (95% CI: 0–59.8%)), a FE model was applied. Estimation of RE model wielded consistent results (OR: 0.73, 95% CI: 0.54–0.98) with the FE model.(ii)Eight studies (Figure 4(c)) provided data for DGE. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding DGE (OR: 1.08, 95% CI: 0.68–1.70). Heterogeneity was significant between the studies ( test : 0.04, : 53% (95% CI: 0–78.9%)), so a RE model was used. Estimation of FE model wielded consistent results (OR: 1.07, 95% CI: 0.81–1.40) with the RE model.(iii)Eight studies (Figure 4(d)) provided data for clinically significant DGE. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding clinically significant DGE (OR: 0.98, 95% CI: 0.59–1.63). Heterogeneity was significant between the studies (test : 0.03, : 55% (95% CI: 1.7%–79.8%)), so a RE model was used. Estimation of FE model wielded consistent results (OR: 1.03, 95% CI: 0.76–1.40) with the RE model.(iv)Eight studies (Figure 6(a)) provided data for PPH. Meta-analysis of the data showed statistically significant () difference between the two groups regarding PPH (OR: 1.52, 95% CI: 1.08–2.14) in favor of PJ group. Heterogeneity was not significant between the studies ( test : 0.85, : 0% (95% CI: 0–80.3%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 1.52, 95% CI: 1.08–2.14) with the FE model.(v)Eight studies (Figure 6(b)) provided data for clinically significant PPH. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding clinically significant PPH (OR: 1.35, 95% CI: 0.95–1.93). Heterogeneity was not significant between the studies ( test : 0.96, : 0% (95% CI: 0–75.9%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 1.35, 95% CI: 0.94–1.94) with the FE model.(vi)Seven studies (Figure 6(c)) provided data for biliary fistula. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding biliary fistula (OR: 0.58, 95% CI: 0.31–1.06). Heterogeneity was not significant between the studies ( test : 0.14, : 38% (95% CI: 0–73.7%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 0.58, 95% CI: 0.23–1.48) with the FE model.(vii)Nine studies (Figure 6(d)) provided data for intra-abdominal fluid collection. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding intra-abdominal fluid collection (OR: 0.64, 95% CI: 0.40–1.02). Heterogeneity was significant between the studies ( test : 0.07, : 45% (95% CI: 0–74.6%)), so a RE model was used. Estimation of FE model wielded consistent results (OR: 0.64, 95% CI: 0.47–0.87) with the RE model.(viii)Eight studies (Figure 7(a)) provided data for morbidity. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding morbidity (OR: 0.97, 95% CI: 0.77–1.23). Heterogeneity was not significant between the studies ( test : 0.21, : 28% (95% CI: 0–67.5%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 0.97, 95% CI: 0.73–1.28) with the FE model.(ix)Ten studies (Figure 7(b)) provided data for mortality. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding mortality (OR: 0.98, 95% CI: 0.60–1.61). Heterogeneity was not significant between the studies ( test : 0.94, : 0% (95% CI: 0–76.8%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 0.99, 95% CI: 0.60–1.64) with the FE model.(x)Eight studies (Figure 7(c)) provided data for reoperation rate. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding reoperation rate (OR: 0.84, 95% CI: 0.59–1.20). Heterogeneity was not significant between the studies ( test : 0.79, : 0% (95% CI: 0–83%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 0.83, 95% CI: 0.58–1.20) with the FE model.(xi)Four studies (Figure 7(d)) provided data for wound infection. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding wound infection (OR: 1.08, 95% CI: 0.66–1.76). Heterogeneity was not significant between the studies ( test : 0.86, : 0% (95% CI: 0–90%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 1.08, 95% CI: 0.66–1.76) with the FE model.(xii)Six studies (Figure 8(a)) provided data for blood transfusion. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding blood transfusion (OR: 1.03, 95% CI: 0.72–1.47). Heterogeneity was not significant between the studies ( test : 0.39, : 5% (95% CI: 0–91.4%)), so a FE model was used. Estimation of RE model wielded consistent results (OR: 1.04, 95% CI: 0.72–1.51) with the FE model.(xiii)Ten studies (Figure 8(b)) provided data for operative time. Meta-analysis of the data showed no statistically significant () difference between the two groups regarding operative time (MWD: −5.73, 95% CI: −19.3, 7.85). Heterogeneity was significant between the studies ( test : <0.001, : 97% (95% CI: 0–98.1%)), so a RE model was used. Estimation of FE model did not wield consistent results (MWD: −16, 95% CI: −17.24, −14.76) with the RE model.(xiv)Ten studies (Figure 8(c)) provided data for LOS. Meta-analysis of the data showed no statistically significant () difference between the two groups LOS (MWD: −0.74, 95% CI: −2.24, 0.76). Heterogeneity was significant between the studies ( test : <0.001, : 91% (95% CI: 0–94.6%)), so a RE model was used. Estimation of FE model wielded consistent results (MWD: −0.06, 95% CI: −0.35, 0.23) with the RE model.

3.6. Risk of Bias across Studies

Funnel plot of primary outcome (POPF) is shown in Figure 5. No study resides beyond the limits of 95% CI. Egger’s test showed that there was no statistically significant publication bias ().

Figure 5: Funnel plot of comparison: postoperative pancreatic fistula.
Figure 6: (a) Postpancreatectomy haemorrhage, (b) clinically significant postpancreatectomy haemorrhage, (c) biliary fistula, and (d) intra-abdominal fluid collection.
Figure 7: (a) Morbidity, (b) mortality, (c) reoperation, and (d) wound infection.
Figure 8: (a) Blood transfusion, (b) operative time, and (c) length of hospital stay.

4. Discussion

4.1. Summary of Evidence

Pancreaticoduodenectomy remains the most widely used surgical modality for the treatment of pancreatic head and periampullary tumors. Failure of the pancreatic anastomosis resulting in POPF has been identified as one of the most important factors of postoperative morbidity. It must also be mentioned that POPF is assumed to have a close relationship with other post-PD complications, such as IAC, DGE, and PPH [30, 31]. As a result, surgeons, in an attempt to minimize post-PD complications have meticulously compared the available anastomotic techniques.

In our study, after a systematic literature search, a meta-analysis of available RCTs was performed. In the qualitative and quantitative analysis, 10 studies with a total of 1629 patients were included. Regarding the primary outcome, PG was not superior to PJ. However, this result was different when the two techniques were compared on the basis of overall POPF, where a significant difference was found. Heterogeneity in clinically significant POPF could possibly be the result of nonuniformity in the definition of POPF. Although the included studies after 2005 were consistent with the 2005 ISGPS POPF definition, the remaining defined POPF in an inconsistent way. DGE and clinically significant DGE were found to have no difference between PG and PJ, with a high level of heterogeneity though. As the operation type was not determined in most eligible studies, surgeons performed either PD or PPPD. The above-mentioned heterogeneity could be explained in the light of lack of stratification regarding the operation type.

Respectively, results from pooled data showed a lower rate of PPH for PJ, but no difference for clinically significant PPH. Heterogeneity for both of them was 0%, increasing thus the validity of these findings. The rate of biliary fistula and the intra-abdominal fluid collection was not significantly different between PG and PJ, which diverges from the results of previous studies [3235], due to inclusion of the recent RCTs [11, 26]. Moreover, overall postoperative morbidity for both techniques was estimated at the level of 49%, complying with current literature [4]. Similarly, no difference was found in terms of mortality, reoperation rate, wound infection, and perioperative blood transfusion. Finally, PG was not superior to PJ in terms of operative time and LOS. Heterogeneity was significantly high in these comparisons, possibly due to the approximate calculation of the mean and SD.

Risk factors for development of POPF are the age, gender of the patient, preoperative jaundice and malnutrition, underlying pathology, cirrhotic liver, BMI, soft pancreas, pancreatic diameter, pancreatic duct size, operative time, resection type, anastomotic technique, and intraoperative blood loss [36]. El Nakeeb et al. [31], however, in a retrospective study of 471 patients, suggested that risk factors for POPF include the cirrhotic liver, BMI, soft pancreas, pancreatic diameter < 3 mm, and pancreatic duct near the posterior border.

The superiority of PG over PJ in terms of POPF can be justified by some theoretical advantages. Firstly, due to the fact that the posterior wall of the stomach lies just above the pancreatic remnant, the tension between the stomach and the pancreatic stump is minimized. Secondly, the acidic gastric content prevents the activation of pancreatic enzymes and consequently the anastomotic lysis. Moreover, compared to a jejunum loop, the stomach wall is thicker, thus stabilizing the anastomosis. Finally, the abundant stomach wall vascularization decreases the chance of an anastomotic ischaemia. This may also be the reason of increased post-PD PPH in the PG group, rendering perioperative meticulous haemostasis of utmost importance.

As far as postoperative exocrine pancreatic function is concerned, data are scarce and inconsistent, thus making further analysis very difficult. More specifically, a higher stool elastase level and a significant lesser weight loss were reported in the PG group [25]. Comparing PG and IRPJ, El Nakeeb et al. [23] concluded that postoperative steatorrhea and need for pancreatic enzyme supplements were higher in the PG group, while post-PD serum albumin was in a lower level in patients submitted to PG. On the contrary, the need for oral enzyme supplements, six months after surgery, was lower in the PG group, with the rate of reported steatorrhea further decreasing after 12 months [11]. In a study of 99 patients, Hirono et al. [37] identified hard texture of pancreas and PG reconstruction as individual risk factors for postoperative pancreatic exocrine function insufficiency.

Regarding the pancreatic endocrine function, El Nakeeb et al. [23] showed that, although there was no difference in the overall rate of postoperative diabetes mellitus between PG and IRPJ, postoperative fasting blood sugar was higher in the PG group. Furthermore, fasting blood sugar increased postoperatively in the PG group, unlike IRPJ, where fasting blood sugar was significantly lower after surgery. However, two studies claimed that there was no statistically significant difference between PG and PJ in the rate of de novo diabetes mellitus [11, 25].

Morphological outcomes were not systematically provided and therefore a pooled analysis could not be reported. Data show that pancreatic duct tended to be more dilated in the PG group, even after a median of 32 months and the pancreatic parenchyma density is significantly decreased [38, 39]. A significant higher impact of postoperative atrophy of the pancreatic parenchyma was recorded in the PG group [39]. However, in a study by Fang et al. [40], no significant differences between PG and PJ regarding postoperative pancreatic duct diameter were reported.

Our meta-analysis provides an up-to-date pooled, published only data, estimation of the rate of POPF, and other postoperative complications between the two most popular anastomotic techniques. Compared to other recent studies [12, 41], it reports results not only in overall morbidity but also in clinically significant complications, such as DGE and PPH.

4.2. Limitations

Several limitations should be taken into account before appraising the results of this meta-analysis. First of all, the between studies heterogeneity was substantial, limiting, in this way, the significance of the results. Furthermore, there is a diversity in the POPF definition among the included studies. It must be noted, though, that all studies after 2005 use the ISGPS definition. The included trials have also incorporated both PD and PPPD in their study groups and there was, also, no stratification on the basis of the underlying pathology. Moreover, a lack of uniformity exists, regarding the surgical anastomotic technique that may possibly result in biased results. Factors like the texture of pancreas and the pancreatic duct diameter might also influence the results. Another source of bias could be the perioperative use of glue and stents and the postoperative administration of somatostatin, since not all studies reported this information. Another factor that contributes to heterogeneity is the surgical experience in the applied anastomotic technique. Last literature search was performed 20 July 2016. The new refinement of the ISGPS POPF definition [42] was published later; thus, it was not incorporated.

4.3. Conclusions

The present meta-analysis of RCTs demonstrates that there is no difference between the two anastomotic techniques regarding clinically significant POPF. PG has lower overall incidence of POPF and higher rate of PPH against PJ. Moreover, PG and PJ did not differ in terms of overall DGE, clinically significant DGE, clinically significant PPH, biliary fistula, intra-abdominal fluid collection, overall morbidity, mortality, reoperation rate, wound infection, intraoperative blood transfusion, operative time, and LOS. Therefore, selection of proper pancreatic reconstruction should be according to the risk of patients, in order to reduce POPF, postoperative complications, and mortality. PG is superior to PJ regarding short term outcomes, while PJ provides better pancreatic function. Given several limitations, more large scale high quality RCTs are required for the effect of the anastomotic technique on the incidence of POPF to be clarified.

Disclosure

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.

Authors’ Contributions

Konstantinos Perivoliotis, Eleni Sioka, Athina Tatsioni, Ioannis Stefanidis, Elias Zintzaras, and Dimitrios Zacharoulis contributed study conception and design. Konstantinos Perivoliotis and Eleni Sioka were responsible for acquisition of data. Konstantinos Perivoliotis and Eleni Sioka performed analysis and interpretation of data. Konstantinos Perivoliotis, Eleni Sioka, and Athina Tatsioni drafted the manuscript. Athina Tatsioni, Ioannis Stefanidis, Elias Zintzaras, and Dimitrios Zacharoulis conducted critical revision.

References

  1. M. E. Lidsky, Z. Sun, D. P. Nussbaum, M. A. Adam, P. J. Speicher, and D. G. Blazer, “Going the Extra Mile: Improved Survival for Pancreatic Cancer Patients Traveling to High-volume Centers,” Annals of Surgery, 2016. View at Publisher · View at Google Scholar · View at Scopus
  2. E. M. Gleeson, M. F. Shaikh, P. A. Shewokis et al., “WHipple-ABACUS, a simple, validated risk score for 30-day mortality after pancreaticoduodenectomy developed using the ACS-NSQIP database,” Surgery, vol. 160, no. 5, pp. 1279–1287, 2016. View at Publisher · View at Google Scholar
  3. J. Xiong, P. Szatmary, W. Huang et al., “Enhanced recovery after surgery program in patients undergoing pancreaticoduodenectomy a prisma-compliant systematic review and meta-analysis,” Medicine (United States), vol. 95, no. 18, Article ID e3497, 2016. View at Publisher · View at Google Scholar · View at Scopus
  4. G. Conzo, C. Gambardella, E. Tartaglia et al., “Pancreatic fistula following pancreatoduodenectomy. Evaluation of different surgical approaches in the management of pancreatic stump. Literature review,” International Journal of Surgery, vol. 21, no. 1, pp. S4–S9, 2015. View at Publisher · View at Google Scholar · View at Scopus
  5. J. A. Khalil, N. Mayo, S. Dumitra et al., “Pancreatic fistulae after a pancreatico-duodenectomy: Are pancreatico-gastrostomies safer than pancreatico-jejunostomies? An expertise-based trial and propensity-score adjusted analysis,” HPB, vol. 16, no. 12, pp. 1062–1067, 2014. View at Publisher · View at Google Scholar · View at Scopus
  6. J. P. Arnaud, J. J. Tuech, C. Cervi, and R. Bergamaschi, “Pancreaticogastrostomy compared with pancreaticojejunostomy after pancreaticoduodenectomy,” European Journal of Surgery, vol. 165, no. 4, pp. 357–362, 1999. View at Publisher · View at Google Scholar · View at Scopus
  7. K. Heeger, V. Fendrich, J. Waldmann, P. Langer, V. Kanngießer, and D. K. Bartsch, “Reduced complication rate after modified binding purse-string-mattress sutures pancreatogastrostomy versus duct-to-mucosa pancreaticojejunostomy,” Surgeon, vol. 11, no. 5, pp. 246–252, 2013. View at Publisher · View at Google Scholar · View at Scopus
  8. E. Oussoultzoglou, P. Bachellier, J.-M. Bigourdan, J.-C. Weber, H. Nakano, and D. Jaeck, “Pancreaticogastrostomy decreased relaparotomy caused by pancreatic fistula after pancreaticoduodenectomy compared with pancreaticojejunostomy,” Archives of Surgery, vol. 139, no. 3, pp. 327–335, 2004. View at Publisher · View at Google Scholar · View at Scopus
  9. S. Takano, Y. Ito, Y. Watanabe, T. Yokoyama, N. Kubota, and S. Iwai, “Pancreaticojejunostomy versus pancreaticogastrostomy in reconstructions following pancreaticoduodenectomy,” British Journal of Surgery, vol. 87, no. 4, pp. 423–427, 2000. View at Publisher · View at Google Scholar · View at Scopus
  10. J. J. Tuech, P. Pessaux, R. Duplessis, F. Villapadierna, J. Ronceray, and J. P. Arnaud, “Pancreatico-jejunal vs pancreatico-gastric anastomosis after pancreatico-duodenectomy. A retrospective comparative study,” Chirurgie, vol. 123, no. 5, pp. 450–455, 1998. View at Publisher · View at Google Scholar · View at Scopus
  11. T. Keck, U. F. Wellner, and M. Bahra, “Pancreatogastrostomy versus pancreatojejunostomy for RECOnstruction After PANCreatoduodenectomy (RECOPANC, DRKS 00000767),” Annals of Surgery, vol. 263, no. 3, pp. 440–449, 2015. View at Publisher · View at Google Scholar
  12. G. P. Guerrini, P. Soliani, G. D'Amico et al., “Pancreaticojejunostomy versus pancreaticogastrostomy after pancreaticoduodenectomy: an up-to-date meta-analysis,” Journal of Investigative Surgery, vol. 29, no. 3, pp. 175–184, 2016. View at Publisher · View at Google Scholar · View at Scopus
  13. F.-B. Liu, J.-M. Chen, W. Geng et al., “Pancreaticogastrostomy is associated with significantly less pancreatic fistula than pancreaticojejunostomy reconstruction after pancreaticoduodenectomy: A meta-analysis of seven randomized controlled trials,” HPB, vol. 17, no. 2, pp. 123–130, 2015. View at Publisher · View at Google Scholar · View at Scopus
  14. W. Que, H. Fang, B. Yan et al., “Pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy: a meta-analysis of randomized controlled trials,” American Journal of Surgery, vol. 209, no. 6, pp. 1074–1082, 2015. View at Publisher · View at Google Scholar · View at Scopus
  15. D. Moher, A. Liberati, J. Tetzlaff, and D. G. Altman, “Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement,” PLoS Medicine, vol. 6, no. 7, Article ID e1000097, 2009. View at Publisher · View at Google Scholar
  16. C. Bassi, C. Dervenis, G. Butturini et al., “Postoperative pancreatic fistula: an international study group (ISGPF) definition,” Surgery, vol. 138, no. 1, pp. 8–13, 2005. View at Publisher · View at Google Scholar · View at Scopus
  17. M. N. Wente, C. Bassi, C. Dervenis et al., “Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS),” Surgery, vol. 142, no. 5, pp. 761–768, 2007. View at Publisher · View at Google Scholar · View at Scopus
  18. M. N. Wente, J. A. Veit, C. Bassi et al., “Postpancreatectomy hemorrhage (PPH): an International Study Group of Pancreatic Surgery (ISGPS) definition,” Surgery, vol. 142, no. 1, pp. 20–25, 2007. View at Publisher · View at Google Scholar · View at Scopus
  19. J. P. T. Higgins, D. G. Altman, P. C. Gøtzsche et al., “The Cochrane Collaboration's tool for assessing risk of bias in randomised trials,” The British Medical Journal, vol. 343, no. 7829, Article ID d5928, 2011. View at Publisher · View at Google Scholar · View at Scopus
  20. S. P. Hozo, B. Djulbegovic, and I. Hozo, “Estimating the mean and variance from the median, range, and the size of a sample,” BMC Medical Research Methodology, vol. 5, article 13, 2005. View at Publisher · View at Google Scholar · View at Scopus
  21. C. Bassi, M. Falconi, E. Molinari et al., “Reconstruction by pancreaticojejunostomy versus pancreaticogastrostomy following pancreatectomy: results of a comparative study,” Annals of Surgery, vol. 242, no. 6, pp. 767–773, 2005. View at Publisher · View at Google Scholar · View at Scopus
  22. J.-P. Duffas, B. Suc, S. Msika et al., “A controlled randomized multicenter trial of pancreatogastrostomy or pancreatojejunostomy after pancreatoduodenectomy,” American Journal of Surgery, vol. 189, no. 6, pp. 720–729, 2005. View at Publisher · View at Google Scholar · View at Scopus
  23. A. El Nakeeb, E. Hamdy, A. M. Sultan et al., “Isolated Roux loop pancreaticojejunostomy versus pancreaticogastrostomy after pancreaticoduodenectomy: a prospective randomized study,” HPB, vol. 16, no. 8, pp. 713–722, 2014. View at Publisher · View at Google Scholar · View at Scopus
  24. L. Fernàndez-Cruz, R. Cosa, L. Blanco, M. A. López-Boado, and E. Astudillo, “Pancreatogastrostomy with gastric partition after pylorus-preserving pancreatoduodenectomy versus conventional pancreatojejunostomy a prospective randomized study,” Annals of Surgery, vol. 248, no. 6, pp. 930–937, 2008. View at Publisher · View at Google Scholar · View at Scopus
  25. J. Figueras, L. Sabater, P. Planellas et al., “Randomized clinical trial of pancreaticogastrostomy versus pancreaticojejunostomy on the rate and severity of pancreatic fistula after pancreaticoduodenectomy,” British Journal of Surgery, vol. 100, no. 12, pp. 1597–1605, 2013. View at Publisher · View at Google Scholar · View at Scopus
  26. J. Grendar, J.-F. Ouellet, F. R. Sutherland, O. F. Bathe, C. G. Ball, and E. Dixon, “In search of the best reconstructive technique after pancreaticoduodenectomy: pancreaticojejunostomy versus pancreaticogastrostomy,” Canadian Journal of Surgery, vol. 58, no. 3, pp. 154–159, 2015. View at Publisher · View at Google Scholar · View at Scopus
  27. B. Topal, S. Fieuws, R. Aerts et al., “Pancreaticojejunostomy versus pancreaticogastrostomy reconstruction after pancreaticoduodenectomy for pancreatic or periampullary tumours: a multicentre randomised trial,” The Lancet Oncology, vol. 14, no. 7, pp. 655–662, 2013. View at Publisher · View at Google Scholar · View at Scopus
  28. U. F. Wellner, O. Sick, M. Olschewski, U. Adam, U. T. Hopt, and T. Keck, “Randomized controlled single-center trial comparing pancreatogastrostomy versus pancreaticojejunostomy after partial pancreatoduodenectomy,” Journal of Gastrointestinal Surgery, vol. 16, no. 9, pp. 1686–1695, 2012. View at Publisher · View at Google Scholar · View at Scopus
  29. C. J. Yeo, J. L. Cameron, M. M. Maher et al., “A prospective randomized trial of pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy,” Annals of Surgery, vol. 222, no. 4, pp. 580–592, 1995. View at Publisher · View at Google Scholar · View at Scopus
  30. B. Topal, R. Aerts, T. Hendrickx, S. Fieuws, and F. Penninckx, “Determinants of complications in pancreaticoduodenectomy,” European Journal of Surgical Oncology, vol. 33, no. 4, pp. 488–492, 2007. View at Publisher · View at Google Scholar · View at Scopus
  31. A. El Nakeeb, T. Salah, A. Sultan et al., “Pancreatic anastomotic leakage after pancreaticoduodenectomy. Risk factors, clinical predictors, and management (single center experience),” World Journal of Surgery, vol. 37, no. 6, pp. 1405–1418, 2013. View at Publisher · View at Google Scholar · View at Scopus
  32. P. Lei, J. Fang, Y. Huang, Z. Zheng, B. Wei, and H. Wei, “Pancreaticogastrostomy or pancreaticojejunostomy? Methods of digestive continuity reconstruction after pancreaticodudenectomy: a meta-analysis of randomized controlled trials,” International Journal of Surgery, vol. 12, no. 12, pp. 1444–1449, 2014. View at Publisher · View at Google Scholar · View at Scopus
  33. B. Menahem, L. Guittet, A. Mulliri, A. Alves, and J. Lubrano, “Pancreaticogastrostomy is superior to pancreaticojejunostomy for prevention of pancreatic fistula after pancreaticoduodenectomy: an updated meta-analysis of randomized controlled trials,” Annals of Surgery, vol. 261, no. 5, pp. 882–887, 2015. View at Publisher · View at Google Scholar · View at Scopus
  34. J. J. Xiong, C. L. Tan, P. Szatmary et al., “Meta-analysis of pancreaticogastrostomy versus pancreaticojejunostomy after pancreaticoduodenectomy,” British Journal of Surgery, vol. 101, no. 10, pp. 1196–1208, 2014. View at Publisher · View at Google Scholar · View at Scopus
  35. Y. Zhou, J. Yu, L. Wu, and B. Li, “Meta-analysis of pancreaticogastrostomy versus pancreaticojejunostomy on occurrences of postoperative pancreatic fistula after pancreaticoduodenectomy,” Asian Journal of Surgery, vol. 38, no. 3, article no. 199, pp. 155–160, 2015. View at Publisher · View at Google Scholar · View at Scopus
  36. N. O. MacHado, “Pancreatic fistula after pancreatectomy: definitions, risk factors, preventive measures, and management—review,” International Journal of Surgical Oncology, vol. 2012, Article ID 602478, 10 pages, 2012. View at Publisher · View at Google Scholar · View at Scopus
  37. S. Hirono, Y. Murakami, M. Tani et al., “Identification of risk factors for pancreatic exocrine insufficiency after pancreaticoduodenectomy using a13C-labeled mixed triglyceride breath test,” World Journal of Surgery, vol. 39, no. 2, pp. 516–525, 2015. View at Publisher · View at Google Scholar · View at Scopus
  38. E. Lemaire, D. O'Toole, A. Sauvanet, P. Hammel, J. Belghiti, and P. Ruszniewski, “Functional and morphological changes in the pancreatic remnant following pancreaticoduodenectomy with pancreaticogastric anastomosis,” British Journal of Surgery, vol. 87, no. 4, pp. 434–438, 2000. View at Publisher · View at Google Scholar · View at Scopus
  39. Y. Tomimaru, Y. Takeda, S. Kobayashi et al., “Comparison of postoperative morphological changes in remnant pancreas between pancreaticojejunostomy and pancreaticogastrostomy after pancreaticoduodenectomy,” Pancreas, vol. 38, no. 2, pp. 203–207, 2009. View at Publisher · View at Google Scholar · View at Scopus
  40. W. Fang, C. Su, Y. Shyr et al., “Functional and morphological changes in pancreatic remnant after pancreaticoduodenectomy,” Pancreas, vol. 35, no. 4, pp. 361–365, 2007. View at Publisher · View at Google Scholar
  41. S. Crippa, R. Cirocchi, J. Randolph et al., “Pancreaticojejunostomy is comparable to pancreaticogastrostomy after pancreaticoduodenectomy: an updated meta-analysis of randomized controlled trials,” Langenbeck's Archives of Surgery, vol. 401, no. 4, pp. 427–437, 2016. View at Publisher · View at Google Scholar · View at Scopus
  42. C. Bassi, G. Marchegiani, C. Dervenis et al., “The 2016 update of the International Study Group (ISGPS) definition and grading of postoperative pancreatic fistula: 11 Years After,” Surgery, vol. 161, no. 3, pp. 584–591, 2017. View at Publisher · View at Google Scholar