Pancreatic resection remains the effective treatment for both benign and malignant pathologies of the pancreas and the surroundings. The procedures include pancreaticoduodenectomy (PD), pylorus-preserving PD (PPPD), and distal pancreatectomy (DP). All these procedures require time-consuming, extensive tissue and vessel dissection with a high risk of increasing the quantity of intraoperative blood loss and the number of transfused blood units. Although the operative mortality rate has markedly declined to <5% with increasing experiences and advances in medical and surgical technology [
Ultrasonic dissection (UD) device, delivering high-frequency mechanical vibration onto the targeted tissue, denatures the proteins by disrupting the hydrogen bonds within the protein structure [
Thus, the aim of the present meta-analysis was to assess whether the use of UD in pancreatic surgery has clinical efficacy in improving perioperative outcomes when compared with CD.
A comprehensive literature search was conducted in PubMed, EMBASE, Web of Science, and the Cochrane Library to identify eligible studies. The following MeSH terms and text words were used in combination with Boolean operators AND or OR without language or geographical restrictions: “ultrasonics,” “ultrasonic dissection,” “harmonic scalpel,” “pancreatic surgery,” “pancreaticoduodenectomy,” and “pancreatectomy.” Reference lists of all relevant studies were screened to detect additional publications. Two of the investigators (Haiming Lei and Dong Xu) independently reviewed the titles and abstracts identified in the search. The latest search was conducted on June 1, 2015.
All published randomized controlled trials (RCTs) and non-RCTs that compared UD versus CD for pancreatic surgery were included. In case of duplicates, only the latest or the most detailed and informative article, or the one with the best quality in methodology, was selected, unless they were reports from different time periods or the data of overlapping patients could be subtracted. The exclusion criteria were in vitro experiments, animal studies, or studies with cointerventions. Case reports, reviews, letters, and conference abstracts which provided insufficient information were also excluded. Studies were considered for meeting the inclusion criteria by two reviewers (Haiming Lei and Dong Xu) with any disagreements resolved by discussion or arbitration by a third reviewer (Xinghua Shi). Cohen’s kappa statistic was used to evaluate the chance-corrected agreement between reviewers (SPSS, version 18.0) [
The following information regarding each eligible study was extracted using standardized data extraction forms: authors’ names, year of publication, country, study design, study interval, patients’ mean ages, cases per arm, type of surgery, and ultrasonic device used in each study. The outcomes of this meta-analysis were pancreatic fistula, abdominal abscess, postoperative hemorrhage, operation time, intraoperative blood loss, number of transfused blood units, postoperative hospital stay, and overall mortality and morbidity. Pancreatic fistula was defined according to the International Study Group on Pancreatic Fistula (ISGPF) [
Methodological quality of the eligible RCTs was assessed using the Jadad scoring system [
This study was conducted in accordance with the Statement of Preferred Reporting Items for Systematic Reviews and Meta-Analyses [
The initial search returned 168 potentially relevant citations. After screening of titles and abstracts, 152 citations were excluded for no relevance. Of the remaining 16 articles, 10 were excluded for not meeting the inclusion criteria. Ultimately, six studies including two RCTs [
Flow chart of search.
The major characteristics of the included studies, along with the quality assessment scores, are presented in Table
Characteristics of the included studies.
Author | Year | Country | Study design | Study interval | Age (years) | Cases | Type of surgery | Ultrasonic device | Definition of pancreatic fistula | Quality score | ||
---|---|---|---|---|---|---|---|---|---|---|---|---|
UD | CD | UD (M/F) | CD (M/F) | |||||||||
Suzuki et al. [ |
1999 | Japan | RCT | 1994–1998 | 57.7 | 58.5 | 27 (18/9) | 31 (18/13) | DP | CUSA | An external discharge of pancreatic fluid (>7 d) with an amylase level of >3 times the serum level | 2/5 |
|
||||||||||||
Sugo et al. [ |
2001 | Japan | Retro. | 1994–2000 | 61.7 | 56.7 | 11 (6/5) | 20 (12/8) | DP | HS | An external discharge of pancreatic fluid (>7 d) or a high level of amylase in the drainage fluid (>3 times the serum level) on day 7 after surgery | 15/24 |
|
||||||||||||
Satoi et al. [ |
2011 | Japan | PNR | 2009-2010 | 72 | 64 | 13 (8/5) | 13 (10/3) | PD, PPPD | UCS | ISGPF definition | 15/24 |
|
||||||||||||
Uzunoglu et al. [ |
2012 | Germany, Italy, Greece | RCT | 2009–2011 | 64.8 | 65.2 | 57 (33/24) | 44 (29/15) | PD, PPPD | HS | ISGPF definition | 3/5 |
|
||||||||||||
Heverhagen et al. [ |
2012 | Germany | PNR | 2005–2011 | 61 | 64.1 | 50 (37/13) | 50 (34/16) | PPPD | HS | ISGPF definition | 16/24 |
|
||||||||||||
Yui et al. [ |
2014 | Japan | Retro. | 2000–2010 | 66 | 65 | 57 (31/26) | 52 (29/23) | DP | USAD | UD group: ISGPF definition |
15/24 |
UD: ultrasonic dissection, CD: conventional dissection, M/F: male/female, RCT: randomized controlled trial, DP: distal pancreatectomy, CUSA: Cavitron Ultrasonic Surgical Aspirator, Retro.: retrospective observational study, HS: harmonic scalpel, PNR: prospective nonrandomized observational study, PD: pancreaticoduodenectomy, PPPD: pylorus-preserving PD, UCS: ultrasonically curved shear, ISGPF: International Study Group on Pancreatic Fistula, USAD: ultrasonically activated device.
Three studies [
Summary of meta-analysis.
Outcome | Number of studies | Number of participants | Heterogeneity | Overall effect size | 95% CI of overall effect |
|
|
---|---|---|---|---|---|---|---|
UD | CD | ||||||
|
|||||||
Pancreatic fistula | 3 | 95 | 103 |
|
RR = 0.46 | 0.27–0.76 | 0.003 |
Abdominal abscess | 3 | 95 | 103 |
|
RR = 0.24 | 0.08–0.71 | 0.01 |
Postoperative hemorrhage | 2 | 68 | 72 | NA | RR = 0.58 | 0.03–13.22 | 0.73 |
Operation time (min) | 1 | 57 | 52 | NA | MD = −63.00 | −116.41 to −9.59 | 0.02 |
Intraoperative blood loss (mL) | 1 | 57 | 52 | NA | MD = −215.00 | −695.96 to 265.96 | 0.38 |
Hospital stay (days) | 1 | 57 | 52 | NA | MD = −9.00 | −17.78 to −0.22 | 0.04 |
Mortality | 3 | 95 | 103 | NA | RR = 0.91 | 0.06–14.22 | 0.95 |
Morbidity | 1 | 57 | 52 | NA | RR = 0.81 | 0.55–1.20 | 0.30 |
|
|||||||
Pancreatic fistula | 3 | 120 | 107 |
|
RR = 0.79 | 0.48–1.29 | 0.34 |
Abdominal abscess | 1 | 13 | 13 | NA | RR = 0.33 | 0.01–7.50 | 0.49 |
Postoperative hemorrhage | 2 | 70 | 57 | NA | RR = 1.80 | 0.49–6.57 | 0.37 |
Operation time (min) | 3 | 120 | 107 |
|
MD = −5.98 | −31.98 to 20.03 | 0.65 |
Intraoperative blood loss (mL) | 3 | 120 | 107 |
|
MD = −183.08 | −346.01 to −20.16 | 0.03 |
Transfused blood unit | 2 | 107 | 94 |
|
MD = −0.69 | −1.28 to −0.09 | 0.02 |
Hospital stay (days) | 3 | 120 | 107 |
|
MD = 1.04 | −2.92 to 5.00 | 0.61 |
Mortality | 3 | 120 | 107 |
|
RR = 0.29 | 0.08–1.08 | 0.07 |
Morbidity | 2 | 70 | 57 |
|
RR = 1.00 | 0.72–1.38 | 0.99 |
CI: confidence interval, UD: ultrasonic dissection, CD: conventional dissection, RR: risk ratio, DP: distal pancreatectomy, PD: pancreaticoduodenectomy, PPPD: pylorus-preserving PD, NA: not applicable, MD: mean difference.
Meta-analyses of pancreatic fistula for DP and PD/PPPD.
Three studies [
The funnel plot, based on the incidence of pancreatic fistula, revealed visual asymmetry (Figure
Funnel plot for publication bias.
To our knowledge, this is the first meta-analysis to date that evaluates the clinical efficacy of UD in improving perioperative outcomes in pancreatic surgery when compared with CD. Our meta-analysis suggests that, in comparison with CD, UD is associated with reduced pancreatic fistula, abdominal abscess, operation time, and hospital stay following DP and related with less intraoperative blood loss and transfused blood unit after PD/PPPD. There was no difference between two groups in overall mortality or morbidity after pancreatic surgery.
Pancreatic fistula is the most common and serious complication after pancreatic resection and can further cause other complications [
Nevertheless, such advantage of UD in decreasing fistula rate was only remained in our analysis of DP procedure. The analysis of PD/PPPD did not show any difference of fistula rate between two groups. It has to be mentioned that in our analysis the devices used to cut the pancreatic parenchyma in UD groups were not consistent. For DP, the UD device was applied to transect the parenchyma in all three studies [
Many studies have reported the superiority of UD in reducing operation time and intraoperative blood loss as compared with CD. Inoue et al. showed in a prospective randomized study that ultrasonic scalpel could significantly shorten operation time and decrease intraoperative blood loss for open gastric cancer surgery [
Although the UD device is disposable and may potentially increase the costs of surgery, the savings of blood transfusion and suture material in pancreatic surgery might have compensated the extra costs of the UD device [
There are some limitations in the present meta-analysis. First, we realized that inclusion of non-RCTs could not avoid an inherent selection bias in the treatment groups and may exaggerate the effect magnitude of an intervention. However, the number of RCTs comparing UD with CD for pancreatic surgery is really limited. We had to pool data from non-RCTs to reach a relatively larger sample size for evaluation of the interested outcomes in this analysis. For example, the data regarding operation time, blood loss, and hospital stay was reported by only one RCT. However, the number of such studies was increased to four and the overall sample size was turned to be nearly threefold when non-RCTs were considered. Second, clinical heterogeneity across studies was noted, which is common to all meta-analytic studies. Types of UD device, the experience of surgeons, and the slightly different definition of outcome might have influenced the results of this study, although statistical heterogeneity was not significant in all outcomes except one. Third, the publication bias may be presented in our meta-analysis due to the visual asymmetry of the funnel plot. However, we searched several databases according to the standards of the Cochrane Collaboration. Moreover, Egger’s regression did not reveal any evidences for the presence of such bias. Despite these, the results of our study should be interpreted with caution.
The present meta-analysis shows that, in comparison with CD, UD is associated with better perioperative outcomes in pancreatic surgery, especially in DP procedure. Due to the limited RCTs in this study, future larger randomized trials are necessary for reevaluation of the clinical outcomes of the UD in pancreatic surgery.
The authors declare that there is no conflict of interests regarding the publication of this paper.
Haiming Lei and Dong Xu contributed equally to this work.