Abstract
Background
Pancreatic cancer is the fourth‐leading cause of cancer death for both, men and women. The standard treatment for resectable tumours consists of a classic Whipple (CW) operation or a pylorus‐preserving pancreaticoduodenectomy (PPW). It is unclear which of these procedures is more favourable in terms of survival, postoperative mortality, complications, and quality of life.
Objectives
The objective of this systematic review was to compare the effectiveness of CW and PPW techniques for surgical treatment of cancer of the pancreatic head and the periampullary region.
Search methods
We conducted searches on 28 March 2006, 11 January 2011, 9 January 2014, and 18 August 2015 to identify all randomised controlled trials (RCTs), while applying no language restrictions. We searched the following electronic databases on 18 August 2015: the Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Database of Systematic Reviews (CDSR) and the Database of Abstracts of Reviews of Effects (DARE) from the Cochrane Library (2015, Issue 8); MEDLINE (1946 to August 2015); and EMBASE (1980 to August 2015). We also searched abstracts from Digestive Disease Week and United European Gastroenterology Week (1995 to 2010); we did not update this part of the search for the 2014 and 2015 updates because the prior searches did not contribute any additional information. We identified two additional trials through the updated search in 2015.
Selection criteria
RCTs comparing CW versus PPW including participants with periampullary or pancreatic carcinoma.
Data collection and analysis
Two review authors independently extracted data from the included trials. We used a random‐effects model for pooling data. We compared binary outcomes using odds ratios (ORs), pooled continuous outcomes using mean differences (MDs), and used hazard ratios (HRs) for meta‐analysis of survival. Two review authors independently evaluated the methodological quality and risk of bias of included trials according to the standards of The Cochrane Collaboration.
Main results
We included eight RCTs with a total of 512 participants. Our critical appraisal revealed vast heterogeneity with respect to methodological quality and outcome parameters. Postoperative mortality (OR 0.64, 95% confidence interval (CI) 0.26 to 1.54; P = 0.32), overall survival (HR 0.84, 95% CI 0.61 to 1.16; P = 0.29), and morbidity showed no significant differences, except of delayed gastric emptying, which significantly favoured CW (OR 3.03, 95% CI 1.05 to 8.70; P = 0.04). Furthermore, we noted that operating time (MD ‐45.22 minutes, 95% CI ‐74.67 to ‐15.78; P = 0.003), intraoperative blood loss (MD ‐0.32 L, 95% CI ‐0.62 to ‐0.03; P = 0.03), and red blood cell transfusion (MD ‐0.47 units, 95% CI ‐0.86 to ‐0.07; P = 0.02) were significantly reduced in the PPW group. All significant results were associated with low‐quality evidence based on GRADE (Grades of Recommendation, Assessment, Development and Evaluation) criteria.
Authors' conclusions
Current evidence suggests no relevant differences in mortality, morbidity, and survival between the two operations. However, some perioperative outcome measures significantly favour the PPW procedure. Given obvious clinical and methodological heterogeneity, future high‐quality RCTs of complex surgical interventions based on well‐defined outcome parameters are required.
Plain language summary
Comparison of the stomach‐preserving versus classic 'Whipple' operation for people with cancer of the pancreas or the periampullary region
Review question
It is unclear whether the pylorus‐preserving Whipple procedure results in a higher rate of recurrence of the tumour because it is less extensive and therefore may be less radical, but it is also unknown whether quality of life is decreased after removal of the lower portion of the stomach. Another uncertainty is whether the two methods are associated with different rates or types of complications.
Background
Pancreatic cancer is a leading cause of cancer death. Two surgical procedures can lead to a cure: the classic Whipple operation, in which part of the pancreas, the gallbladder, the duodenum, the pylorus (outlet of the stomach), and the distal (lower) part of the stomach are removed, and the so‐called pylorus‐preserving pancreaticoduodenectomy, or pylorus‐preserving Whipple operation, in which the stomach and the pylorus are not removed.
Study characteristics
We included eight randomised controlled trials with a total of 512 participants in this review. The included trials revealed vast differences in sample size as well as clinical and methodological quality.
Key results
We could identify no relevant differences in terms of main complications, long‐term survival, or death due to complications after the operation, but operating time, intraoperative blood loss, and need for blood transfusion seem to be less frequent in the group treated with the pylorus‐preserving Whipple operation. Our conclusion is that, at present, no relevant difference is evident between the two surgical procedures for the treatment of pancreatic or periampullary cancer.
Quality of the evidence
The quality of the body of evidence is still low since all trials revealed some shortcomings in terms of methodological quality or reporting.
Summary of findings
Background
Description of the condition
Pancreatic cancer is the fourth‐leading cause of cancer death for men and women. Whereas overall cancer death rates have been decreasing since the early 1990s, the incidence and death rate of pancreatic cancer continues to increase in recent years (Kohler 2015). Pancreatic cancer accounts for 6.6% of cancer deaths in men and 7.2% in women (Siegel 2015). In large retrospective studies and randomized controlled trials, the aggressive nature of these tumours and the high local recurrence rate, together with early metastatic spread, have resulted in disappointing five‐year survival rates of between 20% and 24% after resection together with adjuvant treatment strategies (Cameron 2015; Oettle 2013).
Description of the intervention
The current standard therapy for pancreatic tumours situated in the head of the pancreas is resection (Buchler 2003; Lillemoe 2000). Great improvements in pancreatic surgery have led to mortality rates of less than 5% at high‐volume centres (Buchler 2003; Cameron 2015; Trede 1990; Yeo 1997). Moreover, mortality and morbidity rates after resection have reached similar levels to those following the palliative bypass operation (Gouma 1999; Koslowsky 2001; Lillemoe 1996). Nevertheless, operative morbidity remains high, occasionally approaching 30% to 45% (Bassi 2001; Cameron 2015; Gouma 2000; Richter 2003), from causes including intra‐abdominal abscess, sepsis, pancreatic fistula, and delayed gastric emptying (DGE).
One of two surgical techniques is typically performed in the treatment of cancer of the pancreatic head: the classic Whipple (CW) operation developed by Kausch, in Kausch 1912, and later perfected by Whipple, in Whipple 1935, and the pylorus‐preserving Whipple (PPW) operation inaugurated by Watson, in Watson 1944, and popularised by Traverso and Longmire, in Traverso 1980.
How the intervention might work
The CW operation consists of an en bloc (as one unit) removal of the pancreatic head, the duodenum, the common bile duct, the gallbladder, and the distal portion of the stomach, together with adjacent lymph nodes (Trede 1993). This operation can lead to special complications such as early and late dumping (rapid emptying of the stomach), postoperative weight loss (Seiler 2000), and postoperative reflux (Williamson 1993).
The PPW operation preserves the stomach and the pylorus, the latter of which is needed for the physiologically timed transport of chyme (partially digested food passing into the intestine). This might lead to a more physiologic gastrointestinal passage, but also might lead to higher rates of specific complications such as DGE.
Why it is important to do this review
The extent of resection of the PPW is obviously less than that of the CW operation, therefore it is expected that operation time will decrease and intraoperative blood loss will be less. In addition, some authors report improved postoperative weight gain (Seiler 2000), higher quality of life (Wenger 1999), and better access to the biliary anastomosis for postoperative endoscopy in people with recurrent biliary obstruction. It is unclear whether the lesser extent of the resection is justified from the oncological perspective, and whether DGE occurs more often with the PPW than with the CW operation.
Objectives
The objective of this systematic review was to compare the effectiveness of CW and PPW techniques for surgical treatment of cancer of the pancreatic head and the periampullary region.
Methods
Criteria for considering studies for this review
Types of studies
Randomised controlled trials (RCTs) irrespective of publication status or language.
Types of participants
People undergoing an elective PPW or CW operation for periampullary or pancreatic carcinoma.
Types of interventions
Pylorus‐preserving pancreaticoduodenectomy (PPW).
Classic Whipple operation (CW).
Types of outcome measures
Primary outcomes
Postoperative pancreatic‐associated morbidity (disease‐specific), such as leakage from the pancreatic anastomosis/pancreatic fistula, DGE, and biliary leakage.
Secondary outcomes
Survival.
Postoperative mortality.
Intraoperative blood loss.
Red blood cell transfusion.
Operation time.
General morbidity, including postoperative gastrointestinal bleeding, wound infection, pulmonary complications, necessity for reoperation.
Duration of hospital stay.
Quality of life.
Status of resection margins (R0/R1* resection).
*R0 means that the tumour is resected without residual microscopic remnants. In an R1 resection, a microscopic remnant of the tumour is present, and in an R2 resection, macroscopic tumour parts can be seen.
Search methods for identification of studies
Electronic searches
We conducted searches on 28 March 2006, 11 January 2011, 9 January 2014, and 18 August 2015 to identify all published and unpublished RCTs, while applying no language restrictions.
We identified relevant trials by searching the following electronic databases on 18 August 2015:
the Cochrane Central Register of Controlled Trials (CENTRAL), the Cochrane Database of Systematic Reviews (CDSR), and the Database of Abstracts of Reviews of Effects (DARE) from the Cochrane Library (2015, Issue 8) (Appendix 1);
MEDLINE (1946 to 18 August 2015) (Appendix 2); and
EMBASE (1980 to 18 August 2015) (Appendix 3).
We also handsearched abstracts from the 1995 to 2010 American Digestive Disease Week, published in Gastroenterology, and the United European Gastroenterology Week, published in Gut. After 2011 we no longer updated the handsearch, as it did not contribute any additional relevant information.
We constructed the search strategy for this review by using a combination of medical subject headings (MeSH) and text words related to the use of pancreaticoduodenectomy or pylorus‐preserving pancreaticoduodenectomy in the surgical treatment of pancreatic cancer. To identify RCTs, we combined our searches with the Cochrane Highly Sensitive Search Strategy phases one, two, and three, as contained in the Cochrane Handbook for Systematic Reviews of Interventions (please see Appendix 4) (Higgins 2005).
Searching other resources
We contacted members of the Cochrane Upper Gastrointestinal and Pancreatic Diseases Group and experts in the field and asked them to provide details of outstanding clinical trials and any relevant unpublished or published materials. We furthermore screened reference lists of retrieved trials and used the 'related articles' function in PubMed.
Data collection and analysis
Selection of studies
Two review authors independently scanned the abstracts of all trials identified by the search to determine eligibility, selecting full articles for further assessment if the abstract suggested that the trial was relevant. If it was unclear from the abstract whether the article met the inclusion criteria, we retrieved and reviewed the full article for clarification. We excluded papers not meeting our inclusion criteria; we have listed these studies along with reasons for their omission in the Characteristics of excluded studies table. We resolved any disagreements by discussion with a third review author.
Data extraction and management
We extracted data from published and unpublished reports using standardised forms. We have presented a list of the data items that we extracted in Appendix 5.
Assessment of risk of bias in included studies
We assessed the methodological quality of the trials. Methodological quality is defined as confidence that the trial design and its report will restrict bias in the comparison of interventions (Moher 1998). We assessed methodological quality according to the Cochrane 'Risk of bias' tool by examining how the allocation sequence was generated, how allocation concealment was provided, and what risks of other bias were present (Figure 1; Figure 2) (Higgins 2011).
1.
Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.
2.
Risk of bias summary: review authors' judgements about each risk of bias item for each included study.
Measures of treatment effect
We measured the hazard ratio and determined the 95% confidence interval (CI) for survival. For dichotomous outcomes, we calculated the odds ratio along with 95% CI. For continuous outcomes, we calculated mean difference or standardised mean difference with 95% CI.
Unit of analysis issues
The unit of analysis was each participant recruited into the trials.
Dealing with missing data
We performed the analysis on an intention‐to‐treat basis when possible, otherwise adopting the per‐protocol analysis as presented in the original trials.
Assessment of heterogeneity
Using the random‐effects model (DerSimonian 1986), we calculated pooled estimates of effect and investigated the results for statistical heterogeneity (using the I2 statistic). If we identified an extreme level of heterogeneity, we interpreted summary effect measures with caution. However, we investigated the causes of heterogeneity as far as collected data would allow.
Assessment of reporting biases
We used funnel plots to explore publication bias and used asymmetry in a funnel plot of trial size against treatment effect to assess bias (Egger 1997a). We created funnel plots for overall survival, postoperative mortality, and the most representative cause of postoperative morbidity, which is occurrence of a pancreatic fistula. We planned a test for asymmetry if we included at least 10 trials in the review, otherwise using funnel plots for inspection and visual detection of potential publication bias only.
Data synthesis
The editorial base and the review authors' host institutions provided additional statistical guidance. We ensured data accuracy by performing double data entry. We combined data from different trials reporting the same or similar comparisons. We summarised dichotomous data using odds ratios, presented continuous data as weighted mean differences, and summarised survival or time‐to‐event data using hazard ratios (HRs). We calculated pooled HRs and estimated HRs from the data provided in the primary trials according to the methods described by Parmar et al and Tierney et al (Parmar 1998; Tierney 2007). If the primary trials provided only median and range, we used the methods described by Hozo et al to calculate mean and standard deviation where appropriate (Hozo 2005). We chose the appropriate method according to the sample size of the trial.
Subgroup analysis and investigation of heterogeneity
We performed subgroup analyses for survival in participants with periampullary cancer or pancreatic head cancer and in both groups together.
Sensitivity analysis
We performed sensitivity analysis in cases of differing definitions of outcome measures, analysing trials separately that defined outcome measures identically.
Results
Description of studies
Results of the search
We retrieved 1235 abstracts from the 2006 search, and a further 129 abstracts from the 2011 search, 558 abstracts from the 2014 search, and 413 abstracts from the most recent search update in August 2015. We checked these abstracts for eligibility. The 2006 search produced six randomised controlled trials (RCTs) for inclusion, including data from seven publications, although we identified no additional trials in the 2011 and 2014 searches. The 2015 database search produced no new trials for inclusion in the review. However, we identified two additional trials through other resources during the 2015 update. We handsearched 619 abstracts from Digestive Disease Week and United European Gastroenterology Week (1995 to 2010), but found no additional trials.
Included studies
Eight RCTs were eligible (see Characteristics of included studies) (Bloechle 1999; Lin 1999; Paquet 1998; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004; Wenger 1999). The trial by Lin et al, published in 2005, is an update of the trial performed in 1999; the only data analysed from the older trial involved red blood cell transfusion, because this information was not reported in the later trial (Lin 1999). After this and the per‐protocol analysis of eight trials were accounted for, the quantitative data from 512 analysed participants comparing PPW (255 participants) and CW (257 participants) procedures were found suitable for the systematic review.
Besides obvious variation in sample sizes (range from 20 in Taher 2015 to 170 in Tran 2004), evaluation of trial population baseline characteristics revealed adequate consistency: all analysed participants were included because of resectable pancreatic or periampullary cancer. Moreover, In addition, we observed a similar distribution of positive lymph node status in four trials (mean positive lymph node status was 58.5% (SD = 8.6) in the PPW group and 66.5% (SD = 12.9) in the CW group) (Seiler 2005; Srinarmwong 2008; Tran 2004; Wenger 1999).
Regarding perioperative management, we observed non‐standardised administration of somatostatin with respect to dosage, route, frequency, and duration. Three of the eight RCTs reported the use of octreotide (100 to 200 µg 3 times per day) (Seiler 2005; Srinarmwong 2008; Tran 2004). No information about the application of erythromycin was available. In the trial by Tran et al (Tran 2004), 10 participants in the CW group and nine in the PPW group received postoperative radiochemotherapy, whereas no participants received adjuvant treatment in one trial by Lin et al (Lin 1999). No reports about adjuvant treatment were available for the other trials.
Excluded studies
We retrieved 2956 abstracts from the systematic literature search. After screening we excluded 2931 abstracts. Of the remaining 25 abstracts, for which we retrieved the full articles, 16 did not fulfil the inclusion criteria. Most of the excluded articles did not cover surgical aspects of the comparison of PPW and CW, nor did they provide quantitative data on the prespecified outcome parameters. We have presented a study flow diagram in Figure 3. See Characteristics of excluded studies for details.
3.
Study flow diagram.
Risk of bias in included studies
Our methodological appraisal of the individual trials revealed vast heterogeneity in trial design. Only in the two largest trials (130 participants in Seiler 2005 and 170 participants in Tran 2004) was the underlying sample size calculation described. Tran et al stated that at least 65 participants had to be included in each group to show 30% less intraoperative blood loss and a 20% decrease in operation time in the PPW group, with α = 0.05 and β = 0.05 (Tran 2004). The trial authors regarded this number as sufficient to demonstrate a reduction in hospital stay as well. The power calculation in the trial conducted by Seiler et al indicated that 58 participants needed to be enrolled for each procedure to detect a 25% difference in morbidity, with α = 0.05 and β = 0.2 (Seiler 2005). The number of participants included was not justified in the other six trials (Bloechle 1999; Lin 1999; Paquet 1998; Srinarmwong 2008; Taher 2015; Wenger 1999).
One trial performed statistical analysis according to the intention‐to‐treat principle (Tran 2004); the other trials did not state the analytic principle or were analysed per‐protocol.
We assessed publication bias by creating funnel plots for overall survival, postoperative mortality, and pancreatic fistula, the last of which is the most common cause of postoperative morbidity. As fewer than 10 trials were included in all analyses (overall survival N = 3; postoperative mortality N = 7; pancreatic fistula N = 7), we inspected funnel plots for potential publication bias only. We did not observe relevant asymmetry in any plot (Figure 4; Figure 5Figure 6).
4.
Funnel plot of comparison: 1 Survival, outcome: 1.1 Overall.
5.
Funnel plot of comparison: 2 Postoperative mortality, outcome: 2.1 Postoperative mortality.
6.
Funnel plot of comparison: 3 Pancreatic fistula, outcome: 3.1 Pancreatic fistula.
'Summary of findings' tables for survival, postoperative mortality, pancreatic fistula, delayed gastric emptying, operating time, intraoperative blood loss, and red blood cell transfusion, created on the basis of GRADE (Grades of Recommendation, Assessment, Development and Evaluation) criteria (GRADE 2008), showed low‐quality evidence for all of these results except for postoperative mortality, which showed moderate‐quality evidence (Table 2; Table 1; Table 3).
Summary of findings 2. Survival after surgical treatment for periampullary or pancreatic carcinoma.
| Survival after surgical treatment for periampullary or pancreatic carcinoma | ||||||
|
Patient or population: people with surgical treatment for periampullary or pancreatic carcinoma
Settings: surgical departments in Asia and Europe
Intervention: PPW Comparison: CW | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments | |
| Assumed risk | Corresponding risk | |||||
| CW | PPW | |||||
| Overall survival Follow‐up: 18 to 144 months | Medium‐risk population | HR 0.84 (0.61 to 1.16) | 284 (3 trials) | ⊕⊕⊝⊝ lowa,b | ||
| *The basis for the assumed risk (e.g. median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CW: classic Whipple; HR: hazard ratio; PPW: pylorus‐preserving Whipple | ||||||
| GRADE Working Group grades of evidence. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. | ||||||
aInadequate information about sequence generation and allocation concealment. No intention‐to‐treat analysis. bVery wide confidence intervals, unknown number of losses to follow‐up, low total number of events, no sample size calculations reported.
Summary of findings for the main comparison. Postoperative mortality and DGE after surgical treatment for periampullary or pancreatic carcinoma.
| Postoperative mortality and DGE after surgical treatment for periampullary or pancreatic carcinoma | ||||||
|
Patient or population: people with surgical treatment for periampullary or pancreatic carcinoma
Settings: surgical departments in Asia and Europe
Intervention: PPW Comparison: CW | ||||||
| Outcomes | Illustrative comparative risks* (95% CI) | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments | |
| Assumed risk | Corresponding risk | |||||
| CW | PPW | |||||
| Postoperative mortality | 60 per 1000 | 39 per 1000 (16 to 89) | OR 0.64 (0.26 to 1.54) | 464 (7 trials) | ⊕⊕⊕⊝ moderatea | |
| Pancreatic fistula | 93 per 1000 |
88 per 1000 (48 to 158) |
OR 0.95 (0.49 to 1.84) |
468 (7 trials) |
⊕⊕⊝⊝ lowa,b |
|
| DGE | 235 per 1000 |
482 per 1000 (244 to 728) |
OR 3.03 (1.05 to 8.70) |
459 (7 trials) |
⊕⊕⊝⊝ lowa,b | |
| *The basis for the assumed risk was the mean risk in the control group across studies. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; CW: classic Whipple; DGE: delayed gastric emptying; OR: odds ratio; PPW: pylorus‐preserving Whipple | ||||||
| GRADE Working Group grades of evidence. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. | ||||||
aConfidence intervals are wide due to small number of events. No sample size calculation was reported for trials except for Seiler 2005 and Tran 2004. bStrong heterogeneity due to differing outcome definitions and clinical heterogeneity between trials.
Summary of findings 3. Operating time, intraoperative blood loss, and red blood cell transfusion in surgical treatment for periampullary or pancreatic carcinoma.
| Operating time, intraoperative blood loss, and red blood cell transfusion in surgical treatment for periampullary or pancreatic carcinoma | |||||
|
Patient or population: people with surgical treatment for periampullary or pancreatic carcinoma
Settings: surgical departments in Asia and Europe
Intervention: PPW Comparison: CW | |||||
| Outcomes | Illustrative comparative risks (95% CI) | Relative effect (95% CI) | Number of participants (studies) | Quality of the evidence (GRADE) | Comments |
| CW vs PPW | |||||
| Intraoperative blood loss (millilitres) | Mean intraoperative blood loss in the intervention group was
320 millilitres lower than in the control group (620 to 30 lower) |
MD ‐0.32 (‐0.62 to ‐0.03) |
404 (5 trials) | ⊕⊕⊝⊝ lowa,b,c | |
| Red blood cell transfusion (units) | Mean red blood cell transfusion in the intervention group was 0.47 units lower than in the control group (0.86 to 0.07 lower) |
MD ‐0.47 (‐0.86 to ‐0.07) |
273 (5 trials) |
⊕⊕⊝⊝ lowa,b | |
|
Operating time (minutes) |
Mean operating time in the intervention group was
45 minutes lower than in the control group (75 to 16 lower) |
MD ‐45.22 (‐74.67 to ‐15.78) |
472 (7 trials) | ⊕⊕⊝⊝ lowa,b,c | |
| CI: confidence interval; CW: classic Whipple; MD: mean difference; PPW: pylorus‐preserving Whipple | |||||
| GRADE Working Group grades of evidence. High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate. | |||||
aSerious limitations in the study design of Bloechle 1999, Lin 1999, Taher 2015, and Wenger 1999 are a potential source of bias. All are characterised by small sample sizes, lack of blinding, and incomplete outcome reporting. bThe distribution of these continuous outcomes is known to be potentially skewed, which introduces a potential bias in the analyses. cStrong heterogeneity in the analyses.
Allocation
None of the trials provided description of the sequence generation process except for Srinarmwong 2008, which described a computer‐generated sequence generation. Srinarmwong 2008 and Tran 2004 described appropriate allocation concealment. None of the other trials provided information about allocation concealment.
Blinding
None of the trials reported blinding of the treating physicians, participants, or outcome assessors.
Incomplete outcome data
We evaluated follow‐up quality by assessing the follow‐up sequence and the duration of follow‐up. Three trials reported median follow‐up together with a range (Bloechle 1999; Seiler 2005; Tran 2004), which varied from 18 months, in Bloechle 1999, to 63.1 months, in Seiler 2005. The trial by Srinarmwong et al described three‐monthly follow‐up for a minimum of 24 months for all participants up to a maximum of 60 months (Srinarmwong 2008). Four trials provided insufficient information on follow‐up (Lin 1999; Paquet 1998; Taher 2015; Wenger 1999). None of the trials reported withdrawals or losses to follow‐up.
Selective reporting
Seiler 2005, Srinarmwong 2008, and Tran 2004 provided the most comprehensive definitions for analysed outcomes. The remaining trials did not adequately specify the criteria for the stated endpoints. A published protocol was not available for any of the trials.
Effects of interventions
See: Table 1; Table 2; Table 3
Primary outcomes
Data on pancreas‐associated morbidity were inconclusive owing to broad ranges for rates of occurrence for pancreatic fistula, delayed gastric emptying, and biliary leakage. For more specific differentiation of the presumed underlying heterogeneity, we performed a meta‐analysis of these parameters.
Pancreatic fistula
The fistula rate showed no differences between groups: 21 of 232 participants in the PPW group (9.1%) versus 22 of 236 participants in the CW group (9.3%) (OR 0.95, 95% CI 0.49 to 1.84; P = 0.88; I² = 0%). Tran et al defined a pancreatic fistula as drainage of more than 50 ml of amylase‐rich fluid per day through surgically placed drains on or after postoperative day 10, or as pancreatic anastomotic disruption demonstrated radiographically (Tran 2004). Similarily, Srinarmwong et al defined a pancreatic fistula as drainage of amylase‐rich fluid of more than 50 ml per day persisting after postoperative day 10 (Srinarmwong 2008). Seiler et al defined pancreatic fistula as secretion of 30 ml or more of amylase‐rich drainage fluid (more than 5000 units/L) per day for at least 10 days (Seiler 2005). No definition was given in the other trials (Bloechle 1999; Lin 1999; Paquet 1998; Taher 2015; Wenger 1999) (Analysis 1.1). Although the I² statistic showed no heterogeneity, the pooled results must be regarded with caution owing to the various definitions reported. Butturini et al showed that the incidence of pancreatic fistula varied from 10% to 28.5% within the same group of participants when different common definitions were applied (Butturini 2006).
1.1. Analysis.
Comparison 1 Pancreatic fistula, Outcome 1 Pancreatic fistula.
Delayed gastric emptying (DGE)
Seven trials provided data from 459 participants (229 PPW, 230 CW) on the occurrence of DGE (Bloechle 1999; Lin 1999; Paquet 1998; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004). Lin et al and Tran et al used the same definition for DGE (gastric stasis requiring nasogastric intubation for 10 or more days or the inability to tolerate a regular diet on the 14th postoperative day) (Lin 1999; Tran 2004). Srinarmwong et al defined DGE similarly as gastric stasis requiring nasogastric intubation for 10 or more days or the inability to tolerate a regular diet at three months after the operation (Srinarmwong 2008), whereas Seiler et al defined it differently (persistent drainage via the nasogastric tube of more than 500 ml per day for at least five days after surgery, or recurrent vomiting in combination with oedema of the gastrojejunostomy or duodenojejunostomy and proximal dilatation on contrast radiography) (Seiler 2005). No definition was given in the other trials.
The meta‐analysis of all 459 participants revealed a significant benefit for the CW group in terms of DGE, with 72 of 229 participants (31.4%) in the PPW group compared to 54 of 230 (23.5%) in the CW group (OR 3.03, 95% CI 1.05 to 8.70; P = 0.04; I² = 73%) (Analysis 2.1). Owing to the considerable statistical heterogeneity of 73% in the I² statistic, we performed sensitivity analyses that included trials only according to their definition of DGE, because it is well known that the rate of DGE varies with the underlying definition (Butturini 2006).
2.1. Analysis.
Comparison 2 Delayed gastric emptying (with sensitivity analysis), Outcome 1 All studies.
In contrast, the results of both sensitivity analyses were not significant if only trials with the same definition of DGE were included (OR 4.02, 95% CI 0.14 to 119.16; P = 0.42; I² = 80%) (Analysis 2.3), or if all trials that provided any definition of DGE were included (OR 2.00, 95% CI 0.55 to 7.22; P = 0.29; I² = 78%) (Analysis 2.2), but the analyses still presented substantial heterogeneity. In the first case, the data for 198 participants were analysed (99 PPW, 99 CW), and in the second case, the data for 355 participants were analysed (177 PPW, 178 CW).
2.3. Analysis.
Comparison 2 Delayed gastric emptying (with sensitivity analysis), Outcome 3 Studies with the same definitions of DGE.
2.2. Analysis.
Comparison 2 Delayed gastric emptying (with sensitivity analysis), Outcome 2 Studies in which DGE was defined (includes different definitions).
Another factor that could account for the observed heterogeneity might be the different applications of drugs. Participants in three trials received octreotide (Seiler 2005; Srinarmwong 2008; Tran 2004), whereas no participants in the trial by Lin et al were treated with octreotide (Lin 1999). No statement about the application of octreotide was provided for the remaining trials (Bloechle 1999; Paquet 1998; Taher 2015; Wenger 1999). None of the trials provided information on the application of prokinetic drugs such as erythromycin or metoclopramide. Owing to these restrictions, the above‐mentioned pooled results must be regarded with caution. Five trials favoured CW (Bloechle 1999; Lin 1999; Paquet 1998; Srinarmwong 2008; Taher 2015), and two trials reported comparable rates for DGE (Seiler 2005; Tran 2004).
Biliary leakage
The summarised effect size of five RCTs comparing biliary leakage after PPW and CW showed no differences between the groups (Lin 1999; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004): 5 of 191 (2.6%) in the PPW group versus 4 of 198 (2.0%) in the CW group (OR 0.96, 95% CI 0.15 to 6.17; P = 0.96; I² = 37%) (Analysis 3.1).
3.1. Analysis.
Comparison 3 Biliary leakage, Outcome 1 Biliary leakage.
Secondary Outcomes
Survival
Four trials provided survival data suitable for the meta‐analysis (Lin 1999; Paquet 1998; Seiler 2005; Tran 2004): long‐term results (60 months in Lin 1999, Paquet 1998, and Tran 2004, and 36 months in Seiler 2005, actuarial Kaplan‐Meier analysis) from a total of 317 participants (152 PPW, 165 CW) were evaluated by comparing hazard ratios (HRs) and corresponding 95% confidence intervals (CIs) using the random‐effects model. We calculated HRs from survival rates at different time points, as they were given in the Kaplan‐Meier curves according to the methods described by Parmar et al and Tierney et al (Parmar 1998; Tierney 2007). Three trials presented Kaplan‐Meier curves for periampullary and pancreatic head carcinoma together (Paquet 1998; Seiler 2005; Tran 2004). This analysis was based on the data provided for 284 participants (138 PPW, 146 CW). No significant differences could be observed in the pooled results (HR 0.84, 95% CI 0.61 to 1.16; P = 0.29; I² = 9%) (Analysis 4.1).
4.1. Analysis.
Comparison 4 Survival, Outcome 1 Overall.
Three trials with 203 participants (98 PPW, 105 CW) presented Kaplan‐Meier curves for those with pancreatic head carcinoma (Lin 1999; Seiler 2005; Tran 2004). Pooled results revealed no significant differences in survival (HR 0.73, 95% CI 0.43 to 1.22; P = 0.23; I² = 62%) (Analysis 4.2).
4.2. Analysis.
Comparison 4 Survival, Outcome 2 Pancreatic head carcinoma.
Two trials with 74 participants (37 PPW, 37 CW) presented Kaplan‐Meier curves for periampullary carcinoma (Seiler 2005; Tran 2004), but the pooled HR showed no significant difference in survival between the groups (HR 0.83, 95% CI 0.39 to 1.76; P = 0.63; I² = 0%) (Analysis 4.3).
4.3. Analysis.
Comparison 4 Survival, Outcome 3 Periampullary cancer.
The trial by Srinarmwong et al provided actual one‐ and two‐year survival rates, which showed no significant differences between the PPW and CW groups. However, as information needed to calculate HRs was insufficient, we could not incorporate these results in the meta‐analyses concerning survival.
Postoperative mortality
Postoperative mortality showed similar ranges in the PPW group (0% in Bloechle 1999 to 28.6% in Srinarmwong 2008) and the CW group (0% in Bloechle 1999 and Lin 1999 to 15.4% in Srinarmwong 2008). The overall effect estimate of 231 (PPW) versus 233 (CW) participants revealed no significant differences in postoperative mortality (odds ratio (OR) 0.64, 95% CI 0.26 to 1.54; P = 0.32; I² = 0%) (Analysis 5.1).
5.1. Analysis.
Comparison 5 Postoperative mortality, Outcome 1 Postoperative mortality.
Perioperative parameters
Intraoperative blood loss
The amount of intraoperative blood loss was analysed in 404 participants (202 PPW, 202 CW) of five trials (Bloechle 1999; Lin 1999; Seiler 2005; Srinarmwong 2008; Tran 2004). Intraoperative blood loss was significantly reduced in the PPW compared to the CW group (MD ‐0.32, 95% CI ‐0.62 to ‐0.03; P = 0.03; I² = 83%) (Analysis 6.1).
6.1. Analysis.
Comparison 6 Intraoperative blood loss, Outcome 1 Intraoperative blood loss (litres).
Red blood cell transfusion
Correspondingly, the summarised effect estimate of blood replacement (units) indicated a significantly reduced application of blood products in the PPW group (MD ‐0.47, 95% CI ‐0.86 to ‐0.07; P = 0.02; I² = 0%) (Analysis 7.1). This result was based on data from 273 participants in five trials (139 PPW, 134 CW) (Bloechle 1999; Lin 1999; Seiler 2005; Taher 2015; Wenger 1999).
7.1. Analysis.
Comparison 7 Red blood cell transfusion, Outcome 1 Red blood cell transfusion.
Operating time
A meta‐analysis of operating time for seven RCTs demonstrated a significant reduction in operating time (minutes) for the PPW operation (mean difference (MD) ‐45.22, 95% CI ‐74.67 to ‐15.78; P = 0.003; I² = 84%) (Bloechle 1999; Lin 1999; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004; Wenger 1999). This analysis was based on data from 472 participants (238 PPW, 234 CW) (Analysis 8.1).
8.1. Analysis.
Comparison 8 Operating time, Outcome 1 Operating time (minutes).
Postoperative bleeding
Five trials evaluated postoperative bleeding (Lin 1999; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004). Ten of 191 participants (5.2%) in the PPW group and 13 of 189 participants (6.9%) in the CW group were identified; this finding was not statistically significant (OR 0.74, 95% CI 0.32 to 1.74; P = 0.50; I² = 0%) (Analysis 9.1).
9.1. Analysis.
Comparison 9 Postoperative bleeding, Outcome 1 Postoperative bleeding.
Wound infection
Wound infection yielded no statistically significant differences, occurring in 10 of 119 participants in the PPW group (8.4%) and in 13 of 132 participants in the CW group (9.8%) (OR 0.85, 95% CI 0.35 to 2.05; P = 0.72; I² = 0%) (Analysis 10.1). Four trials reported incidences of wound infection but without definition (Lin 1999; Paquet 1998; Seiler 2005; Wenger 1999).
10.1. Analysis.
Comparison 10 Wound infection, Outcome 1 Wound infection.
Pulmonary complications
Three trials reported pulmonary complications. We analysed data for 218 participants (105 PPW, 113 CW). Wenger et al and Paquet et al specified pulmonary complications as pneumonia (Paquet 1998; Wenger 1999), and Seiler et al referred to pulmonary complications as "pulmonary morbidity" (Seiler 2000). We could identify no significant differences between PPW and CW groups (OR 0.67, 95% CI 0.29 to 1.58; P = 0.36; I² = 0%) (Analysis 11.1).
11.1. Analysis.
Comparison 11 Pulmonary complications, Outcome 1 Pulmonary complications.
Necessity for reoperation
Three trials reported this outcome (Seiler 2005; Taher 2015; Tran 2004). In total, 16 of 163 participants in the PPW group (9.8%) and 18 of 157 participants in the CW group (11.5%) required reoperation. This result did not show statistical significance (OR 0.80, 95% CI 0.38 to 1.68; P = 0.56; I² = 0%) (Analysis 12.1).
12.1. Analysis.
Comparison 12 Necessity for reoperation, Outcome 1 Necessity for reoperation.
Duration of hospital stay
The duration of hospital stay (days) showed similar results between groups (MD 0.26, 95% CI ‐2.04 to 2.56; P = 0.82; I² = 0%) in a meta‐analysis of five trials evaluating a total of 366 participants (188 in the PPW group and 178 in the CW group) (Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004; Wenger 1999) (Analysis 13.1).
13.1. Analysis.
Comparison 13 Duration of hospital stay, Outcome 1 Hospital stay (days).
Quality of life (QoL)
Owing to the use of different QoL questionnaires, it was impossible to pool these data quantitatively. We thus presented a qualitative overview of QoL results by operative procedure.
Three RCTs reported outcomes for QoL (Bloechle 1999; Seiler 2005; Wenger 1999). Bloechle 1999 and Wenger 1999 used the European Organisation for Research and Treatment of Cancer (EORTC) QLQ‐30 questionnaire in their RCTs, but only Wenger et al used the recommended pancreas‐specific module QLQ‐PAN 26 (Fitzsimmons 1998). The data from each RCT showed no significant differences in overall QoL by procedure.
In the trial by Wenger et al, QoL was assessed preoperatively and after 2, 4, 6, 24, 36, 48, and 60 weeks. Participants in the PPW group scored significantly better on the domains loss of appetite, nausea and vomiting, diarrhoea, and weight gain. No significant difference in overall QoL was noted (Wenger 1999).
Bloechle et al did not specify the assessment time points but, like Wenger, reported a significantly better score for participants in the PPW group in the domains loss of appetite and weight gain, but no difference in overall QoL, diarrhoea, nausea and vomiting (Bloechle 1999).
Seiler et al assessed QoL preoperatively using the Sickness Impact Profile and again after 6, 12, 18, and 24 months. This group found no differences in postoperative weight gain or overall QoL, but did report that ability to work after six months was better for participants in the PPW group than for those in the CW group (Seiler 2005).
Status of resection margins (R0/R1* resection)
Four trials provided data on resection margin status after PPW or CW, respectively (Paquet 1998; Seiler 2005; Tran 2004; Wenger 1999): all four trials showed a balanced interstudy and intrastudy distribution of R0 resections (resections without microscopical tumour remnants) with an overall R0 resection rate of 80.2% (142 of 177 participants) in the PPW group and 81.9% (149 of 182 participants) in the CW group. The meta‐analytic result was not statistically significant (OR 0.92, 95% CI 0.39 to 2.15; P = 0.84; I² = 0%) (Analysis 14.1).
14.1. Analysis.
Comparison 14 R0 resection rate, Outcome 1 R0 resection rate.
Discussion
Summary of main results
For the past decades preservation of the pylorus in people undergoing partial pancreaticoduodenectomy for cancer has been a controversial issue. Numerous studies have been performed, including several randomised controlled trials (RCTs), but the cumulative knowledge gained from these needed to be captured in a quantitative summary of the results to establish whether or not pylorus‐preserving pancreaticoduodenectomy (PPW) is a better technique than the classic Whipple (CW) operation.
We could find no evidence of superiority of one procedure over the other from a curative perspective in this systematic review and meta‐analysis. Pooled long‐term results of four RCTs showed no differences in terms of overall survival (hazard ratio (HR) 0.84; P = 0.29).
Differences in postoperative mortality between the two techniques were not statistically significant. As for surgical and clinically relevant parameters of morbidity, we found no significant differences for the occurrence of pancreatic fistula (odds ratio (OR) 0.95; P = 0.88), biliary leakage (OR 0.96; P = 0.96), or postoperative bleeding (OR 0.74; P = 0.50). The analysis of delayed gastric emptying (DGE) showed a significant reduction of this complication in the CW group (OR 3.03; P = 0.04), but the sensitivity analyses regarding the definitions of DGE did not corroborate this finding.
The PPW procedure showed several significant benefits in perioperative outcome measures such as operating time (mean difference (MD) ‐45.22 min; P = 0.003), intraoperative blood loss (MD ‐0.32 L; P = 0.03), and red blood cell transfusion (MD ‐0.47 units; P = 0.02). Shorter operating time is an advantage of the PPW operation in terms of cost, but the benefits of this reduced operating time for postoperative morbidity (for example pulmonary and cardiac complications) are unclear. Also, the reduction in intraoperative blood loss and blood transfusion seems not to impact postoperative morbidity and mortality. Furthermore, length of hospital stay was not significantly different between the two groups (MD 0.26 days; P = 0.82). The analyses on perioperative outcomes and length of hospital stay must be interpreted with caution as substantial heterogeneity (greater than 70%) was present in some of these analyses, and the estimated means and standard deviations might not be reliable due to a skewed distribution of the primary data.
Overall completeness and applicability of evidence
Our findings are in line with those of a recently published study from Bassi et al (Bassi 2005a), in which survival data were correlated with type of surgery by multivariate analysis. These results indicated that neither the type of surgery (PPW versus CW; HR 0.8; P = 0.078) nor the occurrence of postoperative complications significantly affected the hazard of death once tumour staging was taken into account (grade, nodal status, and maximum tumour size). As shown by Neoptolemos et al (Neoptolemos 2001), it was not resection margin status R0/R1 but rather tumour grade and lymph node status that were by far the most powerful prognostic factors of survival.
In terms of DGE, previous reports of non‐randomised cohort studies reporting higher rates after the PPW procedure were confirmed by the results of our meta‐analysis (Mosca 1997; Roder 1992). However, five of the eight included trials showing higher rates of DGE in the PPW group all had small sample sizes (Bloechle 1999; Lin 1999; Paquet 1998; Srinarmwong 2008; Taher 2015), with a maximum of 44 participants in one trial (Bloechle 1999). Given that a significant difference for the occurrence of DGE was found only when all included trials were considered (31.4% PPW versus 23.9% CW; OR 3.03; P = 0.04) and that there was no significant difference in the sensitivity analyses according to the definition of DGE, these results reinforce the importance of well‐defined endpoints in surgical trials and furthermore indicate that underpowered studies potentially overestimate the benefits of CW for the outcome of DGE (Butturini 2006; Kunz 1998).
If PPW and CW are equally effective regarding survival, postoperative mortality, and morbidity, quality of life (QoL) could become the crucial factor in comparisons between the two procedures. Although three of the eight trials reported QoL as an outcome, we could reach no conclusion because the trials used different questionnaires administered at different time points, and only sparse information assessing QoL was available in the trial by Bloechle et al (Bloechle 1999).
Quality of the evidence
Given obvious sources of potential bias in the included trials such as small sample sizes, varying or missing definitions of outcome parameters, the lack of blinding, and inexplicit methods of randomisation and allocation concealment, our results should be interpreted with caution. We observed no clear differences between PPW and CW techniques for survival, postoperative mortality, or main complications. Much larger and more rigorously designed trials than are currently available would be needed to demonstrate potential superiority of one or the other technique. This highlights the urgent need for the development of standardised RCTs for complex interventions. Moreover, these trials should rely on evidence‐based consensus definitions of surgical interventions and their outcome parameters.
Potential biases in the review process
Systematic reviews currently provide the best method of summarising existing evidence, but the quality of such reviews depends on the quality of the primary studies (Egger 1997b; Mulrow 1994). Thus, the trials included in this review are partially marred by bias and clinical heterogeneity, and this may distort results.
Our Cochrane Highly Sensitive Search Strategy was approved by an expert in the field (GA), but it may still be biased. In an attempt to identify all relevant trials, we contacted experts in pancreatic surgery and in clinical research to identify additional or ongoing trials. We identified two additional relevant trials by this procedure.
The fact that the randomisation process was carried out differently in the reviewed trials can be considered as a source of bias, because adequate randomisation and allocation concealment generate balanced groups for known, unknown, and unmeasured confounders. Only one of the included trials provided adequate descriptions of the random sequence generation process (Srinarmwong 2008), whereas the other trials failed to describe the method of sequence generation satisfactorily.
Moreover, only two trials specified maintenance of allocation concealment and blinding of the outcome assessor (Srinarmwong 2008; Tran 2004).
In addition, median follow‐up and follow‐up sequences varied significantly between individual trials (median follow‐up 18 months in Bloechle 1999 to 144 months in Paquet 1998). This insufficient reporting on inconsistent follow‐up sequences is a possible indicator of performance bias.
Furthermore, we found varying or even missing definitions of surrogate parameters such as pancreatic fistula and DGE. The trials by Seiler et al, Srinarmwong et al, and Tran et al provided the most comprehensive definitions (Seiler 2005; Srinarmwong 2008; Tran 2004). The other trials did not specify the criteria of their endpoints at all. The heterogeneous definitions of outcome parameters may have caused detection bias.
Internationally accepted scaled definitions for surgical outcome parameters in pancreatic cancer surgery have been introduced in recent years. Efforts such as the consensus conference of the International Study Group for Definition of Postoperative Pancreatic Fistula, Post‐Pancreatectomy Haemorrhage and DGE should be encouraged, to decrease interobserver variability (Bassi 2005b; Wente 2007a; Wente 2007b).
Regarding the use of antibiotics, somatostatin, drains, nasogastric tubes, etc., we observed a non‐standardised therapeutic concept or even non‐reporting with respect to dosage, route, frequency, and duration of administration. The considerable variation in perioperative management is thus a further indicator of interstudy heterogeneity, and this may influence the external validity of the summarised results.
Another possible source of bias lies in the statistical analysis. Only one trial used an intention‐to‐treat approach (Tran 2004), whereas the remaining trials applied per‐protocol analysis or did not specify their analysis approach. This distorted analysis of participants lost to follow‐up or missing information on excluded participants (attrition bias) makes it difficult to decide whether the remaining study population is representative (external validity). Moreover, statistical analysis of obviously underpowered trials (six trials, Bloechle 1999, Lin 1999, Paquet 1998, Srinarmwong 2008, Taher 2015, and Wenger 1999, represent only ˜40% of the total number of participants) reduces the explanatory power of these results, since small studies are known to potentially bias treatment effects (Kunz 1998). A power calculation was reported for only two trials (Seiler 2005; Tran 2004), and some of the results may not reach statistical significance owing to the lack of an appropriate number of participants.
In this context, we should point out that we did not investigate the influence of neoadjuvant and adjuvant treatment because relevant data were not reported in an interpretable way in the reviewed trials. Our results may therefore be biased, because adjuvant treatment (for example chemotherapy, chemoradiation) represents a possible confounder with an impact on survival, postoperative mortality, and morbidity.
In spite of the sources of clinical and methodological heterogeneity already mentioned, we still observed adequately balanced groups when we compared interstudy baseline population characteristics: all analysed participants were included because of suspected pancreatic or periampullary carcinoma. Our systematic review also revealed a balanced distribution of R0/R1 resections (mean R0 resection: 91.1% PPW, 90.4% CW) and lymph node status (mean positive lymph node status: 58.5% PPW, 66.5% CW). The calculated average level of statistical heterogeneity of all 15 meta‐analytical approaches indicated moderate heterogeneity of 22.8% (I² statistic). Nevertheless, given obvious sources of clinical and methodological heterogeneity, we decided to compute the statistics of our meta‐analysis using the random‐effects model, which allows for variation in the treatment effects of individual studies while providing a more conservative pooled effect estimate (DerSimonian 1986).
Given the obvious sources of bias, the results of our systematic review and meta‐analyses must be interpreted with caution. Small sample size, missing definitions of outcome parameters, and inexplicit randomisation and allocation concealment are some of the most relevant factors that can produce misleading results.
Eight trials providing high‐level evidence addressing identical surgical problems produced many examples of clinically relevant heterogeneity. This calls for future efforts to overcome current hurdles to achieving standardised RCTs for complex interventions. Moreover, evidence‐based consensus is mandatory to establish refined definitions for surgical interventions and their putative outcome parameters. This systematic review and meta‐analysis of the PPW versus the CW procedure could serve as an ideal basis for the pursuit of such endeavours.
Agreements and disagreements with other studies or reviews
This review agrees with the conclusions of the authors of all included trials (Bloechle 1999; Lin 1999; Paquet 1998; Seiler 2005; Srinarmwong 2008; Taher 2015; Tran 2004; Wenger 1999). It also agrees with the findings of a systematic review and meta‐analysis of randomised trials on this topic (Other published versions of this review). Our results differ slightly from the systematic reviews of Iqbal et al and Karanicolas et al (Iqbal 2008; Karanicolas 2007), which might be explained by our inclusion of two novel trials, Srinarmwong 2008 and Taher 2015, that were not available at the time of the former reviews. Furthermore, Iqbal et al included a number of non‐randomised studies, which could have introduced bias into their results (Iqbal 2008).
Authors' conclusions
Implications for practice.
In conclusion, current evidence on the basis of existing RCTs suggests no difference between the PPW and the CW operation in terms of survival, postoperative mortality, and main morbidity. There might be a potential benefit for the CW operation in terms of delayed gastric emptying, but sensitivity analyses could not corroborate this finding. However, it does appear that the PPW procedure may reduce operating time, intraoperative blood loss, and red blood cell transfusion.
Implications for research.
Much larger and more rigorously designed trials than those currently available are needed to demonstrate further differences between these two procedures. The ideal future trial would be a non‐inferiority (thousands of participants) multicentre RCT, which hardly seems feasible owing to the present lack of funding in this field.
Some trials have been started concerning a novel method of a pylorus‐resecting pancreaticoduodenectomy, in which the stomach is completely preserved, but the pyloric ring is resected (Hackert 2013; Kawai 2011). This novel technique might combine the positive effects of the CW (less delayed gastric emptying) with those of the PPW (shorter operating time, less intraoperative blood loss and blood transfusion), but high‐quality evidence is still lacking for this procedure.
What's new
| Date | Event | Description |
|---|---|---|
| 18 August 2015 | New citation required but conclusions have not changed | 2 new trials incorporated, conclusions were not changed. However, pooled result for red blood cell transfusion turned significant in favour of PPW. Furthermore, DGE turned significant in favour of CW due to the inclusion of the new trials. Hanns‐Peter Knaebel is no longer an author of this Cochrane review due to his commercial relationships. |
| 18 August 2015 | New search has been performed | Searches rerun and results incorporated. |
History
Protocol first published: Issue 2, 2006 Review first published: Issue 2, 2008
| Date | Event | Description |
|---|---|---|
| 3 June 2015 | Amended | Hanns‐Peter Knaebel's declaration of interest statement updated. |
| 18 November 2014 | Amended | Details of NIH funding held by review author included in acknowledgements. |
| 21 May 2014 | New citation required but conclusions have not changed | No new trials identified. |
| 21 May 2014 | New search has been performed | Study flow diagram added as a figure. Some minor errors revised. |
| 13 April 2011 | New citation required but conclusions have not changed | No new trials identified. |
| 22 February 2011 | New search has been performed | No new trials identified. |
| 4 January 2011 | Amended | Withdrawn while results of updated literature search were incorporated. |
| 5 October 2010 | Amended | Contact details updated. |
| 12 May 2008 | Amended | Converted to RevMan 5 format. 'Risk of bias' table added. |
| 29 April 2008 | Amended | Converted to new review format. |
Acknowledgements
This review was carried out on the basis of a systematic review and meta‐analysis published elsewhere (Diener 2007).
Christina Fitzmaurice holds a T32 NIH training grant (United States of America).
We thank Professor Dr Hanns‐Peter Knaebel for his contribution to the design and conduct of the formerly published versions of this review. He has since become Chief Executive Officer of Aesculap AG, a medical device company that produces surgical instruments, and as such is now in breach of Cochrane's commercial sponsorship policy. He is therefore no longer an author of the current and subsequent versions of this Cochrane review.
Appendices
Appendix 1. CENTRAL search strategy
exp Pancreatic Neoplasms/
(pancrea$ adj5 neoplas$).tw.
(pancrea$ adj5 cancer$).tw.
(pancrea$ adj5 carcin$).tw.
(pancrea$ adj5 tumo$).tw.
(pancrea$ adj5 malig$).tw.
or/1‐6
exp pancreatectomy/
pancreatectom$.tw.
exp pancreaticoduodenectomy/
pancreaticoduodenectom$.tw.
pancreatoduodenectom$.tw.
duodenopancreatectom$.tw.
exp pancreaticojejunostomy/
pancreaticojejunostom$.tw.
pancreatojejunostom$.tw.
pancrea$ resect$.tw.
whipple.tw.
(pylorus adj3 preserv$).tw.
or/8‐19
7 and 20
Appendix 2. MEDLINE search strategy
exp Pancreatic Neoplasms/
(pancrea$ adj5 neoplas$).tw.
(pancrea$ adj5 cancer$).tw.
(pancrea$ adj5 carcin$).tw.
(pancrea$ adj5 tumo$).tw.
(pancrea$ adj5 malig$).tw.
or/1‐6
exp pancreatectomy/
pancreatectom$.tw.
exp pancreaticoduodenectomy/
pancreaticoduodenectom$.tw.
pancreatoduodenectom$.tw.
duodenopancreatectom$.tw.
exp pancreaticojejunostomy/
pancreaticojejunostom$.tw.
pancreatojejunostom$.tw.
pancrea$ resect$.tw.
whipple.tw.
(pylorus adj3 preserv$).tw.
or/8‐19
7 and 20
randomized controlled trial.pt.
controlled clinical trial.pt.
randomized.ab.
placebo.ab.
drug therapy.fs.
randomly.ab.
trial.ab.
groups.ab.
or/22‐29
exp animals/ not humans.sh.
30 not 31
21 and 32
Appendix 3. EMBASE search strategy
exp Pancreatic Neoplasms/
(pancrea$ adj5 neoplas$).tw.
(pancrea$ adj5 cancer$).tw.
(pancrea$ adj5 carcin$).tw.
(pancrea$ adj5 tumo$).tw.
(pancrea$ adj5 malig$).tw.
or/1‐6
exp pancreatectomy/
pancreatectom$.tw.
exp pancreaticoduodenectomy/
pancreaticoduodenectom$.tw.
pancreatoduodenectom$.tw.
duodenopancreatectom$.tw.
exp pancreaticojejunostomy/
pancreaticojejunostom$.tw.
pancreatojejunostom$.tw.
pancrea$ resect$.tw.
whipple.tw.
(pylorus adj3 preserv$).tw.
or/8‐19
7 and 20
Clinical trial/
Randomized controlled trial/
Randomization/
Single‐Blind Method/
Double‐Blind Method/
Cross‐Over Studies/
Random Allocation/
Placebo/
Randomi?ed controlled trial$.tw.
Rct.tw.
Random allocation.tw.
Randomly allocated.tw.
Allocated randomly.tw.
(allocated adj2 random).tw.
Single blind$.tw.
Double blind$.tw.
((treble or triple) adj blind$).tw.
Placebo$.tw.
Prospective study/
or/22‐40
Case study/
Case report.tw.
Abstract report/ or letter/
or/42‐44
41 not 45
21 and 46
Appendix 4. Search strategy
(1) exp Pancreatic Neoplasms/ (2) (pancrea$ adj5 neoplas$).tw. (3) (pancrea$ adj5 cancer$).tw. (4) (pancrea$ adj5 carcin$).tw. (5) (pancrea$ adj5 tumo$).tw. (6) (pancrea$ adj5 malig$).tw. (7) or/2‐6 (8) exp pancreatectomy/ (9) pancreatectom$.tw. (10) exp pancreaticoduodenectomy/ (11) pancreaticoduodenectom$.tw. (12) pancreatoduodenectom$.tw. (13) duodenopancreatectom$.tw. (14) exp pancreaticojejunostomy/ (15) pancreaticojejunostom$.tw. (16) pancreatojejunostom$.tw. (17) pancrea$ resect$.tw. (18) whipple.tw. (19) (pylorus adj3 preserv$).tw. (20) or/8‐19 (21) 7 and 20
Appendix 5. Data extraction details
We extracted the following data.
Title, authors, contact address.
Source.
Published/unpublished.
Year of publication.
Trial sponsors.
Method of randomisation.
Blinding of outcome assessor, participant, and treating physician.
Criteria for patient inclusion and exclusion.
Sample size and sample size calculation.
Baseline characteristics and similarities of groups at baseline.
Withdrawals and losses to follow‐up.
Participant characteristics including mean/median age, age range, sex ratio.
Specific pancreatic diagnosis leading to surgical intervention.
Number of participants assigned to each treatment group.
Details of intervention regimens.
Survival related to each intervention (hazard ratios and their 95% confidence intervals, log‐rank Chi² values, log‐rank P values).
Postoperative mortality related to each intervention.
Number of lymph nodes removed in each group.
Status of resection margins.
Intraoperative blood loss related to each intervention.
Operation time related to each intervention.
Duration of hospital stay related to each intervention.
Definition of postoperative complications.
Number and frequency of postoperative complications related to each intervention.
Definitions of adverse reactions and outcomes.
Adverse reactions and outcomes.
Quality of life.
Data and analyses
Comparison 1. Pancreatic fistula.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Pancreatic fistula | 7 | 468 | Odds Ratio (M‐H, Random, 95% CI) | 0.95 [0.49, 1.84] |
Comparison 2. Delayed gastric emptying (with sensitivity analysis).
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 All studies | 7 | 459 | Odds Ratio (M‐H, Random, 95% CI) | 3.03 [1.05, 8.70] |
| 2 Studies in which DGE was defined (includes different definitions) | 4 | 355 | Odds Ratio (M‐H, Random, 95% CI) | 2.00 [0.55, 7.22] |
| 3 Studies with the same definitions of DGE | 2 | 198 | Odds Ratio (M‐H, Random, 95% CI) | 4.02 [0.14, 119.16] |
Comparison 3. Biliary leakage.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Biliary leakage | 5 | 380 | Odds Ratio (M‐H, Random, 95% CI) | 0.96 [0.15, 6.17] |
Comparison 4. Survival.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Overall | 3 | 284 | Hazard ratio (Random, 95% CI) | 0.84 [0.61, 1.16] |
| 2 Pancreatic head carcinoma | 3 | 203 | Hazard ratio (Random, 95% CI) | 0.73 [0.43, 1.22] |
| 3 Periampullary cancer | 2 | 74 | Hazard ratio (Random, 95% CI) | 0.83 [0.39, 1.76] |
Comparison 5. Postoperative mortality.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Postoperative mortality | 7 | 464 | Odds Ratio (M‐H, Random, 95% CI) | 0.64 [0.26, 1.54] |
Comparison 6. Intraoperative blood loss.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Intraoperative blood loss (litres) | 5 | 404 | Mean Difference (IV, Random, 95% CI) | ‐0.32 [‐0.62, ‐0.03] |
Comparison 7. Red blood cell transfusion.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Red blood cell transfusion | 5 | 273 | Mean Difference (IV, Random, 95% CI) | ‐0.47 [‐0.86, ‐0.07] |
Comparison 8. Operating time.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Operating time (minutes) | 7 | 472 | Mean Difference (IV, Random, 95% CI) | ‐45.22 [‐74.67, ‐15.78] |
Comparison 9. Postoperative bleeding.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Postoperative bleeding | 5 | 380 | Odds Ratio (M‐H, Random, 95% CI) | 0.74 [0.32, 1.74] |
Comparison 10. Wound infection.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Wound infection | 4 | 251 | Odds Ratio (M‐H, Random, 95% CI) | 0.85 [0.35, 2.05] |
Comparison 11. Pulmonary complications.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Pulmonary complications | 3 | 218 | Odds Ratio (M‐H, Random, 95% CI) | 0.67 [0.29, 1.58] |
Comparison 12. Necessity for reoperation.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Necessity for reoperation | 3 | 320 | Odds Ratio (M‐H, Random, 95% CI) | 0.80 [0.38, 1.68] |
Comparison 13. Duration of hospital stay.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 Hospital stay (days) | 5 | 366 | Mean Difference (IV, Random, 95% CI) | 0.26 [‐2.04, 2.56] |
Comparison 14. R0 resection rate.
| Outcome or subgroup title | No. of studies | No. of participants | Statistical method | Effect size |
|---|---|---|---|---|
| 1 R0 resection rate | 4 | 359 | Odds Ratio (M‐H, Random, 95% CI) | 0.92 [0.39, 2.15] |
Characteristics of studies
Characteristics of included studies [ordered by study ID]
Bloechle 1999.
| Methods | Method of randomisation: unknown Number randomly assigned: 44 (PPW = 23, CW = 21) Exclusion after randomisation (total and per group): none Losses to follow‐up: yes Method of allocation concealment: unknown Intention‐to‐treat analysis: no Description of sample size calculation: none | |
| Participants | Age, years: 69 (47 to 76) (median and range in the PPW group); 67 (43 to 78) (median and range in the CW group) Sex ratio (m/f): 1.6 (PPW), 1.5 (CW) Inclusion criteria: people with periampullary carcinoma (cT1‐4, cN0‐1, cM0) Exclusion criteria: none Equivalence of baseline characteristics: age and stage distribution similar for both groups | |
| Interventions | PPW and CW (no operation details available) Erythromycin application unknown Somatostatin application unknown | |
| Outcomes | Description of outcome parameters: insufficient Operation time (minutes): PPW: 239 ± 79; CW: 285 ± 91 Postoperative mortality: PPW: 0%; CW: 0% DGE: PPW: 8 (34%); CW: 2 (9%) | |
| Notes | Country: Germany Time of enrolment: unknown Duration of follow‐up: median 18 months (range 12 to 30) The trial investigator did not define the term ‘periampullary cancer’; we assume that it includes people with pancreatic head cancer | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Unclear risk | Comment: not reported |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | High risk | Comment: Postrandomisation dropouts could have influenced effect estimates |
| Selective reporting (reporting bias) | Unclear risk | Comment: no study protocol available |
| Other bias | High risk | Comment: sample size calculation not reported |
Lin 1999.
| Methods | Method of randomisation: unknown Number randomly assigned: 33 (PPW = 14, CW = 19) Exclusion after randomisation (total and per group): 3 (PPW = 3, CW = 0) Losses to follow‐up: yes Method of allocation concealment: unknown Intention‐to‐treat analysis: no Description of sample size calculation: none | |
| Participants | Age, years: 64.5 (48 to 77) (mean and range in the PPW group); 66.7 (46 to 84) (mean and range in the CW group) Sex: ratio (m/f): 2.5 (PPW), 2.2 (CW) Inclusion criteria: people with pancreatic head carcinoma Exclusion criteria: none Equivalence of baseline characteristics: Participants were equivalent in terms of age and sex distribution; the CW group included more participants with stage III cancer | |
| Interventions | PPW and CW (no operation details available) No somatostatin used Erythromycin application unknown | |
| Outcomes | Operation time (minutes): PPW: 221 ± 35; CW: 271 ± 65 Intraoperative blood loss (ml): PPW: 446 ± 342; CW: 1212 ± 194 Blood replacement (units): PPW: 1.0 ± 1.4; CW: 1.6 ± 2.6 Postoperative mortality: PPW: 1 (7.1%); CW: 0 (0%) DGE: PPW: 6 (42.8%); CW: 0 (0%) Bleeding: PPW: 0 (0%); CW: 1 (5.2%) Fistula: PPW: 1 (7.1%); CW: 1 (5.2%) Bile leak: PPW: 0 (0%); CW: 0 (0%) Wound infection: PPW: 1 (7.1%); CW: 1 (5.2%) Intra‐abdominal abscess: PPW: 0 (0%); CW: 1 (5.2%) | |
| Notes | Country: Taiwan
Time of enrolment: 156 weeks
Duration of follow‐up: unknown All operations performed by the same surgeon |
|
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Unclear risk | Comment: not reported |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | High risk | Comment: Postrandomisation dropouts could have influenced effect estimates |
| Selective reporting (reporting bias) | Unclear risk | Comment: no study protocol available |
| Other bias | High risk | Comment: sample size calculation not reported |
Paquet 1998.
| Methods | Method of randomisation: sealed envelopes (information from study author) Number randomly assigned: 40 (PPW = 17, CW = 23) Exclusion after randomisation (total and per group): unknown Losses to follow‐up: unknown Method of allocation concealment: unknown Intention‐to‐treat analysis: no Description of sample size calculation: none | |
| Participants | Age, years: unknown Sex: unknown Inclusion criteria: people with pancreatic adenocarcinoma or periampullary cancer and an R0 resection Exclusion criteria: none Equivalence of baseline characteristics: unknown | |
| Interventions | Anastomoses: retrocolic end‐to‐end pancreaticojejunostomy with a drain in the pancreatic duct, end‐to‐end hepaticojejunostomy, end‐to‐end duodenojejunostomy Erythromycin application unknown Somatostatin application unknown | |
| Outcomes | Description of outcome parameters: insufficient Postoperative mortality: PPW: 0 (0%); CW: 1 (4%) DGE: PPW: 6 (35%); CW: 1 (4%) Fistula: PPW: 1 (6%); CW: 2 (9%) Radicality (R0): PPW: 17 (100%); CW: 23 (100%) (condition for inclusion) | |
| Notes | Country: Germany Time of enrolment: 521 weeks Duration of follow‐up: minimum = 24 months, maximum = 144 months | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Unclear risk | Comment: "Sealed envelopes" were used (information obtained from the study author upon request) |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: no postrandomisation dropouts reported |
| Selective reporting (reporting bias) | Unclear risk | Comment: no study protocol available |
| Other bias | High risk | Comment: sample size calculation not reported |
Seiler 2005.
| Methods | Method of randomisation: sealed envelopes Number randomly assigned: 130 (PPW = 64, CW = 66) Exclusion after randomisation (total): 84 (group distribution unknown) Losses to follow‐up: unknown Method of allocation concealment: unknown Intention‐to‐treat analysis: no Description of sample size calculation: yes | |
| Participants | Age, years: 64.8 (26 to 83) (median and range in the PPW group); 65 (33 to 86) (median and range in the CW group) Sex: ratio (m/f): 1.29 (PPW), 1 (CW) Inclusion criteria: all people suitable for surgery with suspected pancreatic or periampullary cancer considered resectable on the basis of computed tomography or magnetic resonance imaging, with no history of previous gastric resection Exclusion criteria: direct tumour invasion of the proximal duodenum, pylorus, or stomach; peripyloric lymph node metastases confirmed by intraoperative frozen‐section examination; distant metastases or locally unresectable tumours due to major retroperitoneal infiltration; emergency resections; necessity for total pancreatectomy to achieve clear resection margins Equivalence of baseline characteristics: Groups were similar regarding age, sex, and stage distribution | |
| Interventions | Reconstruction performed by means of an interrupted 2‐layer end‐to‐side pancreaticojejunostomy, an end‐to side hepaticojejunostomy 10 to 15 cm distal to the pancreatic anastomosis, and an end‐to‐side gastrojejunostomy/duodenojejunostomy approximately 40 cm distal to the biliodigestive anastomosis, followed by a Braun jejunojejunostomy Somatostatin application: 100 to 200 µg three times a day for a total of seven days Erythromycin application: none Perioperative treatment: antibiotic prophylaxis | |
| Outcomes | Operation time (minutes): PPW: 382 (240 to 645); CW: 449 (240 to 780) Intraoperative blood loss (ml): PPW: 1198 (400 to 4000); CW: 1500 (400 to 6000) Blood replacement (units): PPW: 0.9 (0 to 6); CW: 1.9 (0 to 10) Hospital stay (days): PPW: 19.7 (10 to 61); CW: 20.8 (8 to 67) Postoperative mortality: PPW: 1 (2%); CW: 2 (3%) DGE: PPW: 20 (31%); CW: 30 (45%) Bleeding: PPW: 2 (3%); CW: 4 (6%) Fistula: PPW: 2 (3%); CW: 1 (2%) Bile leak: PPW: 0 (0%); CW: 1 (1.5%) Infection (wound or abscess): PPW: 4 (6%); CW: 4 (6%) Positive LNs: PPW: 33 (62%); CW: 41 (72%) Radicality (R0): PPW: 48 (91%); CW: 45 (79%) | |
| Notes | Country: Germany Time of enrolment: 274 weeks Duration of follow‐up: median = 63.1 months, minimum = 4 months, maximum = 93 months | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Unclear risk | Comment: envelopes used for randomisation but unclear whether envelopes were opaque |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: no postrandomisation dropouts |
| Selective reporting (reporting bias) | Unclear risk | Comment: not reported |
| Other bias | Low risk | — |
Srinarmwong 2008.
| Methods | Method of randomisation: computer‐generated random lists Number randomly assigned: 27 (PPW = 14, CW = 13) Exclusion after randomisation (total and per group): none Losses to follow‐up: unknown Method of allocation concealment: sealed, opaque envelopes Intention‐to‐treat analysis: unknown Description of sample size calculation: none | |
| Participants | Age, years: 61.8 (51 to 74) (median and range in the PPW group); 63.3 (52 to 72) (median and range in the CW group) Sex: ratio (m/f): 2.5 (PPW), 1.6 (CW) Inclusion criteria: all people with suspected pancreatic or periampullary cancer evaluated to have resectable disease Exclusion criteria: previous gastric resection; distant metastases; unresectable tumours; tumour invasion of pylorus or stomach; refusal to participate Equivalence of baseline characteristics: Groups were similar regarding age, sex, distribution of pancreatic and periampullary cancer, positive lymph nodes, and mesenteric vessel involvement | |
| Interventions | Resection line: 2 cm distal to the pylorus (PPW group); 20% to 40% of the stomach resected (CW group) Reconstruction performed by means of an end‐to‐side invaginated pancreaticojejunostomy, an end‐to side hepaticojejunostomy, and an end‐to‐side duodenojejunostomy in the PPW group or a side‐to‐side gastrojejunostomy in the CW group Somatostatin application: 100 µg three times a day for a total of seven days Erythromycin application: unknown Perioperative treatment: antibiotic prophylaxis (cefazolin + metronidazole); H2‐antagonists postoperatively |
|
| Outcomes | Operation time (minutes): PPW: 303 (245 to 367); CW: 316 (287 to 360)
Intraoperative blood loss (ml): PPW: 1850 (1300 to 2400); CW: 1946 (1200 to 2600) Hospital stay (days): PPW: 22 (12 to 28); CW: 20 (11 to 24) Postoperative mortality: PPW: 4 (28.6%); CW: 2 (15.4%) DGE: PPW: 9 (64.3%); CW: 2 (15.4%) Bleeding: PPW: 1 (7.1%); CW: 1 (7.7%) Fistula: PPW: 5 (35.7%); CW: 4 (30.8%) Bile leak: PPW: 2 (14.3%); CW: 0 (0%) Positive LNs: PPW: 10 (71.4%); 11 (84.6%) |
|
| Notes | Country: Thailand Time of enrolment: 260 weeks Duration of follow‐up: minimum = 24 months, maximum = 60 months; follow‐up every 3 months | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Low risk | Comment: computer‐generated random list |
| Allocation concealment (selection bias) | Low risk | Comment: sealed, opaque envelopes |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: no postrandomisation dropouts |
| Selective reporting (reporting bias) | Unclear risk | Comment: not reported |
| Other bias | High risk | Comment: sample size calculation not reported |
Taher 2015.
| Methods | Method of randomisation: unknown Number randomly assigned: 20 (PPW = 12, CW = 8) Exclusion after randomisation (total and per group): none Losses to follow‐up: unknown Method of allocation concealment: unknown Intention‐to‐treat analysis: unknown Description of sample size calculation: none | |
| Participants | Age, years: 50.3 ± 10.6 (mean and SD in the PPW group); 44 ± 10.9 (mean and SD in the CW group) Sex: ratio (m/f): 5.66 (total) Inclusion criteria: all people suitable for surgery with carcinoma of the head of the pancreas or the periampullary region Exclusion criteria: people with distant metastases Equivalence of baseline characteristics: only scarce data, age not significantly different; all participants in the CW group had periampullary cancer, whereas in the PPW group 8 of 12 participants had periampullary cancer | |
| Interventions | PPW and CW (no operation details available) Erythromycin application unknown Somatostatin application unknown Perioperative treatment details unknown |
|
| Outcomes | Operation time (minutes): PPW: 308 ± 46; CW: 356 ± 47
Blood replacement (units): PPW: 2.67 ± 0.65; CW: 2.88 ± 0.64
Hospital stay (days): PPW: 24.9 ± 6.7; CW: 28.1 ± 9.1
Postoperative mortality: PPW: 0 (0%); CW: 1 (12.5%)
DGE: PPW: 4 (33%); CW: 1 (12.5%)
Bleeding: PPW: 1 (8.3%); CW: 1 (12.5%)
Fistula: PPW: 1 (8.3%); CW: 0 (0%)
Bile leak: PPW: 1 (8.3%); CW: 3 (37.5%) Re‐laparotomy: PPW: 1 (8.3%); CW: 1 (12.5%) |
|
| Notes | Country: Bangladesh Time of enrolment: 104 weeks Duration of follow‐up: unknown | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: "total 20 patients [...] were enrolled [...] and randomly allocated" |
| Allocation concealment (selection bias) | Unclear risk | Comment: not reported |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: no postrandomisation dropouts |
| Selective reporting (reporting bias) | Unclear risk | Comment: not reported |
| Other bias | High risk | Comment: sample size calculation not reported; ambiguities in report of the trial |
Tran 2004.
| Methods | Method of randomisation: sealed envelopes Number randomly assigned: 170 (PPW = 87, CW = 83) Exclusion after randomisation (total and per group): 36 (PPW = 17, CW = 19) Losses to follow‐up: unknown Method of allocation concealment: sealed, opaque envelopes Intention‐to‐treat analysis: yes Description of sample size calculation: yes | |
| Participants | Age, years: 64 (43 to 78) (median and range in the PPW group); 62 (27 to 78) (median and range in the CW group) Sex: ratio (m/f): 2.32 (PPW), 1.35 (CW) Inclusion criteria: inclusion of consecutive patients with suspected pancreatic or periampullary cancer that was assumed to be resectable according to preoperative diagnostic imaging Exclusion criteria: previous gastric resection, distant metastasis or local unresectable tumours, direct invasion of the pylorus or stomach, positive peripyloric lymph nodes Equivalence of baseline characteristics: groups similar regarding age, sex, and stage distribution | |
| Interventions | End‐to‐side invaginated pancreaticojejunostomy, end‐to‐side hepaticojejunostomy, side‐to‐side gastroenterostomy, or end‐to‐side pylorus‐jejunostomy Somatostatin application: 100 µg preoperatively + three times a day for a total of seven days Erythromycin application: none Perioperative treatment: antibiotic prophylaxis, H2‐antagonists, drain in operation area | |
| Outcomes | Operation time (minutes): PPW: 300 (130 to 600); CW: 300 (160 to 480) Intraoperative blood loss (ml): PPW: 2000 (400 to 21,000); CW: 2000 (300 to 9500) Blood replacement (units): PPW: 2; CW: 2 Hospital stay (days): PPW: 18 (4 to 175); CW: 20 (11 to 138) Postoperative mortality: PPW: 3 (3%); CW: 6 (7%) DGE: PPW: 19/85 (22%); CW: 18/80 (23%) Bleeding: PPW: 6 (7%); CW: 6 (7%) Fistula: PPW: 11 (13%); CW: 12 (14%) Bile leak: PPW: 2 (2%); CW: 0 (0%) Intra‐abdominal abscess: PPW: 9 (10%); CW: 8 (10%) Re‐laparotomy: PPW: 13 (15%); CW: 16 (19%) Positive LNs: PPW: 37/72 (51.4%); 38/69 (55%) Radicality (R0): PPW: 53/72 (73.6%); CW: 57/69 (82.6%) | |
| Notes | Country: Netherlands Time of enrolment: 465 weeks Duration of follow‐up: median = 18.5 months, minimum = 1 month, maximum = 115 months | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Low risk | Quote: "An equal number of blind envelopes with protocols for the SW and the PPPD resection was prepared. The envelopes were used sequentially as patients were enrolled in the study. Therefore, there was strict randomization in both arms. Randomization was carried out in the operation room: a sealed envelope was opened only after it was ascertained that both operation techniques were feasible in the patient concerned" |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: Reasons for exclusion from long‐term survival analysis are given (e.g. non‐malignant disease) |
| Selective reporting (reporting bias) | Unclear risk | Comment: not reported |
| Other bias | Low risk | |
Wenger 1999.
| Methods | Method of randomisation: unknown Number randomly assigned: 48 (PPW = 24, CW = 24) Exclusion after randomisation (total and per group): none Method of allocation concealment: unknown Intention‐to‐treat analysis: no Description of sample size calculation: none | |
| Participants | Age, years: 61.2 ± 7.2 (mean and SD in the PPW group); 61.6 ± 8.9 (mean and SD in the CW group) Sex: ratio (m/f): 1 (PPW), 1 (CW) Inclusion criteria: all people with a preoperative diagnosis of a ductal carcinoma of the pancreatic head or a periampullary carcinoma, an R0 resection, and postoperative affirmation of the diagnosis Exclusion criteria: tumour infiltration of the stomach, the superior part of the duodenum, or the pylorus; age > 75; peritoneal carcinosis; reduced general condition; heart insufficiency; renal insufficiency; hepatic insufficiency; pulmonary insufficiency Equivalence of baseline characteristics: groups similar regarding age, sex, and stage distribution | |
| Interventions | PPW and CW (no description of procedures) Somatostatin application: unknown Erythromycin application: unknown | |
| Outcomes | Operation time (minutes): PPW: 206 ± 48; CW: 306 ± 54 Blood replacement (units): PPW: 5.5 ± 3.1; CW: 6.3 ± 5.2 Hospital stay (days): PPW: 19.1 ± 11.3; CW: 20.9 ± 13.8 Wound infection: PPW: 3 (12.5%); CW: 4 (16.6%) Positive LNs: PPW: 12 (50%); CW: 13 (54%) Radicality (R0): PPW: 24 (100%); CW: 24 (100%) | |
| Notes | Country: Germany Time period for enrolment: 208 weeks Follow‐up: assessed at 2, 6, 12, 24, 36, 48, and 60 weeks; median and range not available | |
| Risk of bias | ||
| Bias | Authors' judgement | Support for judgement |
| Random sequence generation (selection bias) | Unclear risk | Comment: not reported |
| Allocation concealment (selection bias) | Unclear risk | Comment: not reported |
| Blinding (performance bias and detection bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of participants and personnel (performance bias) All outcomes | Unclear risk | Comment: not reported |
| Blinding of outcome assessment (detection bias) All outcomes | Unclear risk | Comment: not reported |
| Incomplete outcome data (attrition bias) All outcomes | Low risk | Comment: no postrandomisation dropouts |
| Selective reporting (reporting bias) | Unclear risk | Comment: not reported |
| Other bias | High risk | Comment: sample size calculation not reported |
CW: classic Whipple DGE: delayed gastric emptying LNs: lymph nodes PPW: pylorus‐preserving Whipple SD: standard deviation
Characteristics of excluded studies [ordered by study ID]
| Study | Reason for exclusion |
|---|---|
| Bakkevold 1993 | Non‐randomised study design; no comparison of PPW versus CW |
| Bassi 1994 | No comparison of PPW versus CW |
| Bell 2005 | Narrative review of 1 included RCT |
| Brennan 1994 | No comparison of PPW versus CW |
| Buchler 1993 | No comparison of PPW versus CW |
| Chou 1996 | No comparison of PPW versus CW |
| Farnell 2005 | No comparison of PPW versus CW |
| Friess 1996 | No comparison of PPW versus CW |
| Johnstone 1993 | No comparison of PPW versus CW |
| Nguyen 2003 | No comparison of PPW versus CW |
| Pedrazzoli 1998 | No comparison of PPW versus CW |
| Seiler 2000 | Additional publication on the same trial (Seiler 2005). No further information |
| Shan 2003 | No comparison of PPW versus CW |
| van Berge Henegouwen | Non‐randomised study design; insufficient quantitative outcome parameters |
| Wagner 2004 | Non‐randomised study design |
| Yeo 1999 | Study not randomised for the comparison of PPW versus CW |
CW: classic Whipple PPW: pylorus‐preserving Whipple RCT: randomised controlled trial
Differences between protocol and review
The review differs in five points from the previously published protocol.
We combined the outcomes of postoperative bleeding and postoperative gastrointestinal bleeding.
We discarded the outcomes of intra‐abdominal fluid collection/abscess, duration of intensive care unit stay, early and late dumping, postoperative reflux, number and status of removed lymph nodes, shock, sepsis, renal failure, weight loss, and endocrine and exocrine insufficiency because no usable data were available.
We performed an additional quality assessment on the basis of a checklist developed by Downs et al (Downs 1998).
We performed a subgroup analysis for pancreatic head cancer and periampullary cancer for survival.
We performed a sensitivity analysis for delayed gastric emptying by using different definitions.
We calculated means and standard deviations according to the methods by Hozo et al for trials that provided only medians and ranges (Hozo 2005).
Contributions of authors
MKD:
Designing the review.
Collecting data for the review.
Undertaking searches.
Screening search results.
Organising retrieval of papers.
Screening retrieved papers against inclusion criteria.
Appraising quality of papers.
Extracting data from papers.
Writing to authors of papers to ask for additional information.
Providing additional data about papers.
Obtaining and screening data on unpublished studies.
Managing data for the review.
Entering data into RevMan 2014.
Analysing data.
Interpreting data.
Providing a methodological perspective.
Providing a clinical perspective.
Providing a consumer perspective.
Writing the review.
Performing previous work that served as the foundation of the current study.
CF:
Collecting data for the review.
Designing search strategies.
Undertaking searches.
Screening search results.
Organising retrieval of papers.
Screening retrieved papers against inclusion criteria.
Appraising quality of papers.
Extracting data from papers.
Writing to authors of papers to ask for additional information.
Providing additional data about papers.
Obtaining and screening data on unpublished studies.
Managing data for the review.
Entering data into RevMan 2014.
Analysing data.
Interpreting data.
Writing the review.
Revising the review after peer review.
GS:
Extracting data from papers.
Analysing data.
Interpreting data.
CMS:
Providing a methodological perspective.
Providing general advice on the review.
Interpreting the data.
Providing a clinical perspective.
Providing a policy perspective.
FJH:
Collecting data for the review.
Undertaking searches.
Screening search results.
Organising retrieval of papers.
Screening retrieved papers against inclusion criteria.
Appraising quality of papers.
Extracting data from papers.
Writing to authors of papers to ask for additional information.
Providing additional data about papers.
Obtaining and screening data on unpublished studies.
Managing data for the review.
Entering data into RevMan 2014.
Analysing data.
Interpreting data.
Providing a clinical perspective.
Updating the review.
Writing the updated review.
GA:
Providing a methodological perspective.
Providing general advice on the review.
MWB:
Providing a clinical perspective.
Providing a policy perspective.
Providing general advice on the review.
Declarations of interest
MKD: none known.
CF: holds a T32 NIH training grant (United States of America).
GS: none known.
CMS: none known.
FJH: none known.
GA: none known.
MWB: none known.
New search for studies and content updated (no change to conclusions)
References
References to studies included in this review
Bloechle 1999 {published data only}
- Bloechle C, Broering DC, Latuske C. Prospective randomized study to evaluate quality of life after partial pancreatoduodenectomy according to Whipple versus pylorus preserving pancreatoduodenectomy according to Longmire‐Traverso for periampullary carcinoma. Deutsche Gesellschaft für Chirurgie 1999;Supplement 1, Forumband:661‐4. [Google Scholar]
Lin 1999 {published data only}
- Lin PW, Lin YJ. Prospective randomized comparison between pylorus‐preserving and standard pancreaticoduodenectomy. British Journal of Surgery 1999;86(5):603‐7. [DOI] [PubMed] [Google Scholar]
- Lin PW, Shan YS, Lin YJ, Hung CJ. Pancreaticoduodenectomy for pancreatic head cancer: PPPD versus Whipple procedure. Hepatogastroenterology 2005;52(65):1601‐4. [PubMed] [Google Scholar]
Paquet 1998 {published data only}
- Paquet K‐J. Comparison of Whipple's pancreaticoduodenectomy with the pylorus‐preserving pancreaticoduodenectomy ‐ a prospectively controlled, randomized long‐term trial [Vergleich der partiellen Duodenopankreatektomie (Whipple‐Operation) mit der pyloruserhaltenden Zephaloduodenopankreatektomie ‐ eine prospektive kontrollierte, randomisierte Langzeitstudie]. Chirurgische Gastroenterologie 1998;14:54‐8. [Google Scholar]
Seiler 2005 {published data only}
- Seiler CA, Wagner M, Bachmann T, Redaelli CA, Schmied B, Uhl W, et al. Randomized clinical trial of pylorus‐preserving duodenopancreatectomy versus classical Whipple resection ‐ long term results. British Journal of Surgery 2005;92(5):547‐56. [DOI] [PubMed] [Google Scholar]
Srinarmwong 2008 {published data only}
- Srinarmwong C, Luechakiettisak P, Prasitvilai W. Standard Whipple's operation versus pylorus preserving pancreaticoduodenectomy: a randomized controlled trial study. Journal of the Medical Association of Thailand 2008;91(5):693‐8. [PubMed] [Google Scholar]
Taher 2015 {published data only}
- Taher MA, Khan ZR, Chowdhury MM, Nur‐E‐Elahi M, Chowdhury AK, Faruque MS, et al. Pylorus preserving pancreaticoduodenectomy vs. standard Whipple's procedure in case of carcinoma head of the pancreas and periampullary carcinoma. Mymensingh Medical Journal 2015;24(2):319‐25. [PubMed] [Google Scholar]
Tran 2004 {published data only}
- Tran KT, Smeenk HG, Eijck CH, Kazemier G, Hop WC, Greve JW, et al. Pylorus preserving pancreaticoduodenectomy versus standard Whipple procedure: a prospective, randomized, multicenter analysis of 170 patients with pancreatic and periampullary tumors. Annals of Surgery 2004;240(5):738‐45. [DOI] [PMC free article] [PubMed] [Google Scholar]
Wenger 1999 {published data only}
- Wenger FA, Jacobi CA, Haubold K, Zieren HU, Muller JM. Gastrointestinal quality of life after duodenopancreatectomy in pancreatic carcinoma. Preliminary results of a prospective randomized study: pancreatoduodenectomy or pylorus‐preserving pancreatoduodenectomy. Chirurg 1999;70:1454‐9. [DOI] [PubMed] [Google Scholar]
References to studies excluded from this review
Bakkevold 1993 {published data only}
- Bakkevold KE, Kambestad B. Morbidity and mortality after radical and palliative pancreatic cancer surgery. Risk factors influencing the short‐term results. Annals of Surgery 1993;217(4):356‐68. [DOI] [PMC free article] [PubMed] [Google Scholar]
Bassi 1994 {published data only}
- Bassi C, Falconi M, Lobardi D, Briani G, Vesentini S, Camboni MG, et al. Prophylaxis of complications after pancreatic surgery: results of a multicenter trial in Italy. Italian Study Group. Digestion 1994;55(Suppl 1):41‐7. [DOI] [PubMed] [Google Scholar]
Bell 2005 {published data only}
- Bell RH Jr. Pancreaticoduodenectomy with or without pylorus preservation have similar outcomes. Cancer Treatment Reviews 2005;31(4):328‐31. [DOI] [PubMed] [Google Scholar]
Brennan 1994 {published data only}
- Brennan MF, Pisters PW, Posner M, Quesada O, Shike M. A prospective randomized trial of total parenteral nutrition after major pancreatic resection for malignancy. Annals of Surgery 1994;220(4):436‐41; discussion 441‐4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Buchler 1993 {published data only}
- Buchler M, Friess H. Prevention of postoperative complications following pancreatic surgery. Digestion 1993;54(Suppl 1):41‐6. [DOI] [PubMed] [Google Scholar]
Chou 1996 {published data only}
- Chou FF, Sheen‐Chen SM, Chen YS, Chen MC, Chen CL. Postoperative morbidity and mortality of pancreaticoduodenectomy for periampullary cancer. European Journal of Surgery 1996;162(6):477‐81. [PubMed] [Google Scholar]
Farnell 2005 {published data only}
- Farnell MB, Pearson RK, Sarr MG, DiMagno EP, Burgart LJ, Dahl TR, et al. A prospective randomized trial comparing standard pancreatoduodenectomy with pancreatoduodenectomy with extended lymphadenectomy in resectable pancreatic head adenocarcinoma. Surgery 2005;138(4):618‐28; discussion 628‐30. [DOI] [PubMed] [Google Scholar]
Friess 1996 {published data only}
- Friess H, Berberat P, Schilling M, Kunz J, Korc M, Buchler MW. Pancreatic cancer: the potential clinical relevance of alterations in growth factors and their receptors. Journal of Molecular Medicine 1996;74(1):35‐42. [DOI] [PubMed] [Google Scholar]
Johnstone 1993 {published data only}
- Johnstone PA, Sindelar WF. Lymph node involvement and pancreatic resection: correlation with prognosis and local disease control in a clinical trial. Pancreas 1993;8(5):535‐9. [DOI] [PubMed] [Google Scholar]
Nguyen 2003 {published data only}
- Nguyen TC, Sohn TA, Cameron JL, Lillemoe KD, Campbell KA, Coleman J, et al. Standard vs. radical pancreaticoduodenectomy for periampullary adenocarcinoma: a prospective, randomized trial evaluating quality of life in pancreaticoduodenectomy survivors. Journal of Gastrointestinal Surgery 2003;7(1):1‐9; discussion 9‐11. [DOI] [PubMed] [Google Scholar]
Pedrazzoli 1998 {published data only}
- Pedrazzoli S, DiCarlo V, Dionigi R, Mosca F, Pederzoli P, Pasquali C, et al. Standard versus extended lymphadenectomy associated with pancreatoduodenectomy in the surgical treatment of adenocarcinoma of the head of the pancreas: a multicenter, prospective, randomized study. Lymphadenectomy Study Group. Annals of Surgery 1998;228(4):508‐17. [DOI] [PMC free article] [PubMed] [Google Scholar]
Seiler 2000 {published data only}
- Seiler CA, Wagner M, Sadowski C, Kulli C, Buchler MW. Randomized prospective trial of pylorus‐preserving vs. classic duodenopancreatectomy (Whipple procedure): initial clinical results. Journal of Gastrointestinal Surgery 2000;4(5):443‐52. [PUBMED: 11077317] [DOI] [PubMed] [Google Scholar]
Shan 2003 {published data only}
- Shan YS, Sy ED, Lin PW. Role of somatostatin in the prevention of pancreatic stump‐related morbidity following elective pancreaticoduodenectomy in high‐risk patients and elimination of surgeon‐related factors: prospective, randomized, controlled trial. World Journal of Surgery 2003;27(6):709‐14. [DOI] [PubMed] [Google Scholar]
van Berge Henegouwen {published data only}
- Berge Henegouwen MI, Moojen TM, Gulik TM, Rauws EA, Obertop H, Gouma DJ. Postoperative weight gain after standard Whipple's procedure versus pylorus‐preserving pancreatoduodenectomy: the influence of tumour status. British Journal of Surgery 1998;85(7):922‐6. [DOI] [PubMed] [Google Scholar]
Wagner 2004 {published data only}
- Wagner M, Redaelli C, Lietz M, Seiler CA, Friess H, Buchler MW. Curative resection is the single most important factor determining outcome in patients with pancreatic adenocarcinoma. British Journal of Surgery 2004;91(5):586‐94. [DOI] [PubMed] [Google Scholar]
Yeo 1999 {published data only}
- Yeo CJ, Cameron JL, Sohn TA, Coleman J, Sauter PK, Hruban RH, et al. Pancreaticoduodenectomy with or without extended retroperitoneal lymphadenectomy for periampullary adenocarcinoma: comparison of morbidity and mortality and short‐term outcome. Annals of Surgery 1999;229(5):613‐22; discussion 622‐4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Additional references
Bassi 2001
- Bassi C, Falconi M, Salvia R, Mascetta G, Molinari E, Pederzoli P. Management of complications after pancreaticoduodenectomy in a high volume centre: results on 150 consecutive patients. Digestive Surgery 2001;18(6):453‐7. [DOI] [PubMed] [Google Scholar]
Bassi 2005a
- Bassi C, Stocken DD, Olah A, Friess H, Buckels J, Hickey H, et al. Influence of surgical resection and post‐operative complications on survival following adjuvant treatment for pancreatic cancer in the ESPAC‐1 randomized controlled trial. Digestive Surgery 2005;22(5):353‐63. [DOI] [PubMed] [Google Scholar]
Bassi 2005b
- Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery 2005;138(1):8‐13. [DOI] [PubMed] [Google Scholar]
Buchler 2003
- Buchler MW, Wagner M, Schmied BM, Uhl W, Friess H, Z'graggen K. Changes in morbidity after pancreatic resection: toward the end of completion pancreatectomy. Archives of Surgery 2003;138(12):1310‐4. [DOI] [PubMed] [Google Scholar]
Butturini 2006
- Butturini G, Marcucci S, Molinari E, Mascetta G, Landoni L, Crippa S, et al. Complications after pancreaticoduodenectomy: the problem of current definitions. Journal of Hepato‐Biliary‐Pancreatic Surgery 2006;13(3):207‐11. [DOI] [PubMed] [Google Scholar]
Cameron 2015
- Cameron JL, He J. Two thousand consecutive pancreaticoduodenectomies. Journal of the American College of Surgeons 2015;220(4):530‐6. [DOI] [PubMed] [Google Scholar]
DerSimonian 1986
- DerSimonian R, Laird N. Meta‐analysis in clinical trials. Controlled Clinical Trials 1986;7(3):177‐88. [DOI] [PubMed] [Google Scholar]
Downs 1998
- Downs SH, Black N. The feasibility of creating a checklist for the assessment of the methodological quality both of randomised and non‐randomised studies of health care interventions. Journal of Epidemiology & Community Health 1998;52(6):377‐84. [DOI] [PMC free article] [PubMed] [Google Scholar]
Egger 1997a
- Egger M, Davey SG, Schneider M, Minder C. Bias in meta‐analysis detected by a simple, graphical test. BMJ 1997;315(7109):629–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
Egger 1997b
- Egger M, Smith GD. Meta‐analysis. Potentials and promise. BMJ 1997;315(7119):1371‐4. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fitzsimmons 1998
- Fitzsimmons D, Johnson CD. Quality of life after treatment of pancreatic cancer. Langenbecks Archives of Surgery 1998;383(2):145‐51. [DOI] [PubMed] [Google Scholar]
Gouma 1999
- Gouma DJ, Nieveen van Dijkum EJ, Geenen RC, Gulik TM, Obertop H. Are there indications for palliative resection in pancreatic cancer?. World Journal of Surgery 1999;23(9):954‐9. [DOI] [PubMed] [Google Scholar]
Gouma 2000
- Gouma DJ, Geenen RC, Gulik TM, Haan RJ, Wit LT, Busch OR, et al. Rates of complications and death after pancreaticoduodenectomy: risk factors and the impact of hospital volume. Annals of Surgery 2000;232(6):786‐95. [DOI] [PMC free article] [PubMed] [Google Scholar]
GRADE 2008 [Computer program]
- Brozek J, Oxman A, Schünemann H. GRADEpro. Version 3.2 for Windows. GRADE Working Group, 2008.
Hackert 2013
- Hackert T, Bruckner T, Dörr‐Harim C, Diener MK, Knebel P, Hartwig W, et al. Pylorus resection or pylorus preservation in partial pancreatico‐duodenectomy (PROPP study): study protocol for a randomized controlled trial. Trials 2013;14:44. [DOI: 10.1186/1745-6215-14-44] [DOI] [PMC free article] [PubMed] [Google Scholar]
Higgins 2005
- Appendix 5b. Highly sensitive search strategies for identifying reports of randomized controlled trials in MEDLINE. In: Higgins JPT, Green S, editors. Cochrane Handbook of Systematic Reviews of Interventions 4.2.5. The Cochrane Library, Issue 3, 2005. Chichester, UK: John Wiley and Sons, Ltd.
Higgins 2011
- Higgins JP, Altman DG, Gøtzsche PC, Jüni P, Moher D, Oxman AD, et al. The Cochrane Collaboration's tool for assessing risk of bias in randomised trials. BMJ 2011;343:d5928. [DOI: 10.1136/bmj.d5928] [DOI] [PMC free article] [PubMed] [Google Scholar]
Hozo 2005
- Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical Research Methodology 2005;5:13. [DOI] [PMC free article] [PubMed] [Google Scholar]
Iqbal 2008
- Iqbal N, Lovegrove RE, Tilney HS, Abraham AT, Bhattacharya S, Tekkis PP, et al. A comparison of pancreaticoduodenectomy with pylorus preserving pancreaticoduodenectomy: a meta‐analysis of 2822 patients. European Journal of Surgical Oncology 2008;34(11):1237‐45. [DOI] [PubMed] [Google Scholar]
Karanicolas 2007
- Karanicolas PJ, Davies E, Kunz R, Briel M, Koka HP, Payne DM. The pylorus: take it or leave it? Systematic review and meta‐analysis of pylorus‐preserving versus standard whipple pancreaticoduodenectomy for pancreatic or periampullary cancer. Annals of Surgical Oncology 2007;14(6):1825‐34. [DOI] [PubMed] [Google Scholar]
Kausch 1912
- Kausch W. Carcinoma of the duodenal papilla and its radical excision [Das Carcinom der Papilla duodeni und seine radikale Entfernung]. Beitrage zur Klinischen Chirurgie 1912;78:439‐86. [Google Scholar]
Kawai 2011
- Kawai M, Tani M, Hirono S, Miyazawa M, Shimizu A, Uchiyama K, et al. Pylorus ring resection reduces delayed gastric emptying in patients undergoing pancreatoduodenectomy: a prospective, randomized, controlled trial of pylorus‐resecting versus pylorus‐preserving pancreatoduodenectomy. Annals of Surgery 2011;253(3):495‐501. [DOI] [PubMed] [Google Scholar]
Kohler 2015
- Kohler BA, Sherman RL, Howlader N, Jemal A, Ryerson AB, Henry KA, et al. Annual report to the nation on the status of cancer, 1975‐2011, featuring incidence of breast cancer subtypes by race/ethnicity, poverty, and state. Journal of the National Cancer Institute 2015;107(6):djv048. [DOI: 10.1093/jnci/djv048] [DOI] [PMC free article] [PubMed] [Google Scholar]
Koslowsky 2001
- Koslowsky TC, Wilke J, Voiss W, Michaelis S, Balta D, Siedek M. Surgical palliation of pancreatic carcinoma. Results of a 7 year period. Chirurg 2001;72(6):704‐9. [DOI] [PubMed] [Google Scholar]
Kunz 1998
- Kunz R, Oxman AD. The unpredictability paradox: review of empirical comparisons of randomised and non‐randomised clinical trials. BMJ 1998;317:1185‐90. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lillemoe 1996
- Lillemoe KD, Cameron JL, Yeo CJ, Sohn TA, Nakeeb A, Sauter PK, et al. Pancreaticoduodenectomy. Does it have a role in the palliation of pancreatic cancer?. Annals of Surgery 1996;223(6):718‐25. [DOI] [PMC free article] [PubMed] [Google Scholar]
Lillemoe 2000
- Lillemoe KD, Yeo CJ, Cameron JL. Pancreatic cancer: state‐of‐the‐art care. CA: A Cancer Journal for Clinicians 2000;50(4):241‐68. [DOI] [PubMed] [Google Scholar]
Moher 1998
- Moher D, Pham B, Jones A, Cook DJ, Jadad AR, Moher M, et al. Does quality of reports of randomised trials affect estimates of intervention efficacy reported in meta‐analyses?. The Lancet 1998;352(9128):609‐13. [DOI] [PubMed] [Google Scholar]
Mosca 1997
- Mosca F, Giulianotti PC, Balestracci T, Candio G, Pietrabissa A, Sbrana F, et al. Long‐term survival in pancreatic cancer: pylorus‐preserving versus Whipple pancreatoduodenectomy. Surgery 1997;122(3):553‐66. [DOI] [PubMed] [Google Scholar]
Mulrow 1994
- Mulrow CD. Rationale for systematic reviews. BMJ 1994;309(6954):597‐9. [DOI] [PMC free article] [PubMed] [Google Scholar]
Neoptolemos 2001
- Neoptolemos JP, Stocken DD, Dunn JA, Almond J, Beger HG, Pederzoli P, et al. Influence of resection margins on survival for patients with pancreatic cancer treated by adjuvant chemoradiation and/or chemotherapy in the ESPAC‐1 randomized controlled trial. Annals of Surgery 2001;234(6):758‐68. [DOI] [PMC free article] [PubMed] [Google Scholar]
Oettle 2013
- Oettle H, Neuhaus P, Hochhaus A, Hartmann JT, Gellert K, Ridwelski K, et al. Adjuvant chemotherapy with gemcitabine and long‐term outcomes among patients with resected pancreatic cancer: the CONKO‐001 randomized trial. JAMA 2013;310(14):1473‐81. [DOI] [PubMed] [Google Scholar]
Parmar 1998
- Parmar MK, Torri V, Stewart L. Extracting summary statistics to perform meta‐analyses of the published literature for survival endpoints. Statistics in Medicine 1998;17(24):2815‐34. [DOI] [PubMed] [Google Scholar]
RevMan 2014 [Computer program]
- The Nordic Cochrane Centre, The Cochrane Collaboration. Review Manager (RevMan). Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014.
Richter 2003
- Richter A, Niedergethmann M, Sturm JW, Lorenz D, Post S, Trede M. Long‐term results of partial pancreaticoduodenectomy for ductal adenocarcinoma of the pancreatic head: 25‐year experience. World Journal of Surgery 2003;27(3):324‐9. [DOI] [PubMed] [Google Scholar]
Roder 1992
- Roder JD, Stein HJ, Huttl W, Siewert JR. Pylorus‐preserving versus standard pancreatico‐duodenectomy: an analysis of 110 pancreatic and periampullary carcinomas. British Journal of Surgery 1992;79(2):152‐5. [DOI] [PubMed] [Google Scholar]
Siegel 2015
- Siegel RL, Miller KD, Jemal A. Cancer statistics, 2015. CA: A Cancer Journal for Clinicians 2015;65(1):5‐29. [DOI] [PubMed] [Google Scholar]
Tierney 2007
- Tierney JF, Stewart LA, Ghersi D, Burdett S, Sydes MR. Practical methods for incorporating summary time‐to‐event data into meta‐analysis. Trials 2007;8:16. [DOI] [PMC free article] [PubMed] [Google Scholar]
Traverso 1980
- Traverso LW, Longmire WP Jr. Preservation of the pylorus in pancreaticoduodenectomy: a follow‐up evaluation. Annals of Surgery 1980;192(3):306‐10. [DOI] [PMC free article] [PubMed] [Google Scholar]
Trede 1990
- Trede M, Schwall G, Saeger HD. Survival after pancreatoduodenectomy. 118 consecutive resections without an operative mortality. Annals of Surgery 1990;211(4):447‐58. [DOI] [PMC free article] [PubMed] [Google Scholar]
Trede 1993
- Trede M, Carter DC. Surgery of the Pancreas. New York: Churchill Livingstone, 1993. [Google Scholar]
Watson 1944
- Watson K. Carcinoma of the ampulla of Vater. Successful radical resection. British Journal of Surgery 1944;31:368‐73. [Google Scholar]
Wente 2007a
- Wente MN, Veit JA, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, et al. Postpancreatectomy hemorrhage (PPH): an International Study Group of Pancreatic Surgery (ISGPS) definition. Surgery 2007;142(1):20‐5. [DOI] [PubMed] [Google Scholar]
Wente 2007b
- Wente MN, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, Izbicki JR, et al. Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS). Surgery 2007;142(5):761‐8. [DOI] [PubMed] [Google Scholar]
Whipple 1935
- Whipple AO, Parsons WB, Mullins CR. Treatment of carcinoma of the ampulla of Vater. Annals of Surgery 1935;102:763‐79. [DOI] [PMC free article] [PubMed] [Google Scholar]
Williamson 1993
- Williamson RC, Bliouras N, Cooper MJ, Davies ER. Gastric emptying and enterogastric reflux after conservative and conventional pancreatoduodenectomy. Surgery 1993;114(1):82‐6. [PubMed] [Google Scholar]
Yeo 1997
- Yeo CJ, Cameron JL, Sohn TA, Lillemoe KD, Pitt HA, Talamini MA, et al. Six hundred fifty consecutive pancreaticoduodenectomies in the 1990s: pathology, complications, and outcomes. Annals of Surgery 1997;226(3):248‐57. [DOI] [PMC free article] [PubMed] [Google Scholar]
References to other published versions of this review
Diener 2007
- Diener MK, Knaebel HP, Heukaufer C, Antes G, Buchler MW, Seiler CM. A systematic review and meta‐analysis of pylorus‐preserving versus classical pancreaticoduodenectomy for surgical treatment of periampullary and pancreatic carcinoma. Annals of Surgery 2007;245(2):187‐200. [DOI] [PMC free article] [PubMed] [Google Scholar]
Diener 2008
- Diener MK, Heukaeufer C, Schwarzer G, Seiler CM, Antes G, Knaebel H‐P, et al. Pancreaticoduodenectomy (classic Whipple) versus pylorus‐preserving pancreaticoduodenectomy (pp Whipple) for surgical treatment of periampullary and pancreatic carcinoma. Cochrane Database of Systematic Reviews 2008, Issue 2. [DOI: 10.1002/14651858.CD006053.pub2] [DOI] [PubMed] [Google Scholar]
Diener 2011
- Diener MK, Fitzmaurice C, Schwarzer G, Seiler CM, Antes G, Knaebel HP, et al. Pylorus‐preserving pancreaticoduodenectomy (pp Whipple) versus pancreaticoduodenectomy (classic Whipple) for surgical treatment of periampullary and pancreatic carcinoma. Cochrane Database of Systematic Reviews 2011, Issue 5. [DOI: 10.1002/14651858.CD006053.pub4] [DOI] [PubMed] [Google Scholar]
Diener 2014
- Diener MK, Fitzmaurice C, Schwarzer G, Seiler CM, Hüttner FJ, Antes G, et al. Pylorus‐preserving pancreaticoduodenectomy (pp Whipple) versus pancreaticoduodenectomy (classic Whipple) for surgical treatment of periampullary and pancreatic carcinoma. Cochrane Database of Systematic Reviews 2014, Issue 11. [DOI: 10.1002/14651858.CD006053.pub5] [DOI] [PMC free article] [PubMed] [Google Scholar]