STRENGTHS AND LIMITATIONS OF THIS STUDY
-
Study selection, data extraction and quality assessment will be performed by 2–3 reviewers, which will minimise the chances of bias influencing the results.
-
Medical databases in other languages will not be searched because of language barriers, so language bias may exist.
-
A comparison of the three major pyeloplasty approaches ensures a thorough representation of the extant literature.
Background
Renal reconstruction surgery in the form of a dismembered pyeloplasty has been the gold standard of care for patients with ureteropelvic junction obstruction (PUJO) since Anderson and Hynes pioneered it in 1949.1
The introduction of minimally invasive procedures such as laparoscopic and robot-assisted laparoscopic pyeloplasty (RALP) for PUJO represents a natural progression from Anderson’s and Hynes’s open dismembered pyeloplasty due to a reduction in operative and postoperative complications and inpatient stay duration.2 As such, RALP is now considered a new gold standard in paediatric minimally invasive surgery3 4 and in all children, with the exception of small infants, the robotic approach appears to be very promising.5
However, in some regions (including the developing world), the financial implications of robotic pyeloplasty are prohibitive, and as such, prior established approaches (ie, open and laparoscopic) remain the surgical approaches of choice.3 It is important to appreciate the degree of disparity in clinical outcomes between different approaches to pyeloplasty in order to drive changes in the provision of paediatric surgical care.
The aim of this systematic review and meta-analysis is to provide a contemporary synthesis of the evidence surrounding paediatric PUJO surgery, comparing the key clinical outcomes between the dominant surgical approaches.
Methodology
This systematic review protocol was written in line with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) 2015 checklist (online supplemental material 1).6 The study was prospectively registered with PROSPERO review databases (CRD42023456779), and all methods described were established before implementation. Once identified, the included studies will undergo analysis and thematic synthesis to derive and compare the key outcomes of open, laparoscopic and RALP in children with PUJO.
Supplemental material
Database searches
A systematic search was conducted using PubMed, Ovid MEDLINE, Embase and Cochrane databases using the following search strategy using the following Medical Subject Headings (MeSH): (pyeloplasty) AND ((laparoscopic) OR (robotic) OR (open)) AND ((Ureteropelvic junction obstruction) OR (pelviureteric junction obstruction)) AND ((child) OR (p?ediatric)) AND (outcome). The search was conducted from inception to March 2024. No language filters were applied. To facilitate the initial screening process, Rayyan will be employed, an AI-powered application designed to improve the reporting accuracy and speed of systematic reviews. Identified articles will be uploaded to Rayyan to expedite the initial screening process and allow two reviewers to filter duplicate studies and then subsequently screen the articles for relevance.7 In addition, studies identified manually by the authors (PN, NC and JMN) will be retrieved and uploaded to Rayyan to be included in the screening process.
Study selection and data extraction
The title and abstract screening process will be completed independently by two researchers (PN and NC). Titles and abstracts of eligible studies will be assessed, and irrelevant articles will be removed. A full-text version of relevant articles will be downloaded for further eligibility review. The full-text review will be undertaken by two researchers (PN and NC), with disputes among researchers being discussed in a meeting and resolved by consensus or arbitrated by a third author (NZ). The reasoning for excluding articles at the full-text review stage will be documented within the PRISMA flow diagram, an exemplar of which is below. Data extraction will be undertaken by three authors (PN, NC and AC).
Inclusion and exclusion criteria
To be included in the analysis, studies must investigate children under the age of 16 with PUJO undergoing a pyeloplasty. Studies or case series with a sample size of fewer than 10 total patients will be excluded in order to minimise heterogeneity and increase the statistical power of the meta-analysis. Conference abstracts, letters to the editor, case reports, reviews and expert opinions will be excluded. Studies that include patients older than 16 years of age will be excluded. Unpublished studies will not be sought. In addition, studies identified manually by the authors (PN, NC and JMN) will be retrieved and uploaded to Rayyan to be included in the screening process. Complete details of the eligibility criteria can be found in table 1.
Eligibility criteria as outlined by the PICOS framework
Data extraction
The extracted data will be collated in a data sheet. The full details of the data extraction fields and outcomes we will extract, where possible, are given in online supplemental material 2. Our key outcomes of interest are numerous, including operative success (ie, procedures not requiring reoperation), length of stay, stent indwelling time, cost, estimated blood loss and complications. Our complications of interest include, but are not limited to, postoperative pain, subsequent haematuria or urinary tract infection, stent dislodgement and pyelonephritis. We will also extract data regarding complications as classified by the Clavien-Dindo criteria, if given.
Supplemental material
Data will be extracted by a minimum of one reviewer (PN, NC and AC), with any disagreements resolved by discussion. Relevant figures will be extracted from the data. If these are not provided, attempts will be made to calculate them from provided data. If this is not possible, the corresponding authors of each paper will be contacted to provide the relevant data. For studies not provided in English, an English language copy will be sought. If this is not successful, the authors will be contacted directly to obtain a translated version.
Risk of bias in individual studies
To assess the bias and quality of the included studies, a Newcastle-Ottawa score will be used, designed to assess the risk of bias in non-randomised studies.8 This scoring system is split into three main sections: selection, comparability and outcome. Each of the sections contains subquestions that assess the quality of the research methodology at the study level. For any identified randomised controlled trials, the Cochrane risk of bias tool V.2 for randomised trials will be used instead.9 Three of the reviewers (PN, AC and NC) will be involved with this process, and any disagreement will be solved by consensus. These results of the risk of bias assessment will be used to carry out a sensitivity analysis, excluding studies deemed to be at a high risk of bias.
Meta-analysis
The nature of our research question means that three distinct comparisons will be made: open versus laparoscopic pyeloplasty, laparoscopic versus robotic pyeloplasty and open versus robotic pyeloplasty. As such, three meta-analyses will be conducted for each outcome of interest, given sufficient homogeneity in the identified studies’ reported outcomes. We will extract the raw numbers of each relevant outcome in both groups. ORs (for binary outcomes) or standardised mean differences (for continuous outcomes) will be calculated and pooled. Between studies, heterogeneity will be assessed using τ2, Higgins and Thompson’s I2 statistic and Cochran’s Q. Given significant between-study heterogeneity, a random-effects model with the Knapp-Hartung (KH) adjustment will be used to calculate a pooled effect measure using the generic inverse-variance method. Otherwise, a fixed-effects model using the exact Mantel-Haenszel method will be used. τ2 will be calculated using the Paule-Mandel or restricted maximum likelihood estimator methods for binary or continuous outcome data, respectively.
For studies with only one experimental group, proportions of each outcome of interest will be extracted, and a meta-analysis of proportions will be carried out as per Wang et al if more than three studies report sufficiently homogenous outcomes in this manner.10 Proportions will first undergo logit transformation before being pooled using the generalised linear mixed-effects model, with the KH adjustment applied. τ2 will be obtained using the maximum-likelihood estimator. Studies with three experimental groups will be split into two separate ‘arms’ and their respective proportions of each outcome extracted, with the sample size of the control group halved in order to attribute half the weighting to each study. The validity of this assumption and its impact on pooled proportions will be tested in a sensitivity analysis.
For each meta-analysis, identification of outliers and influencer analysis will be undertaken if there is deemed to be significant between-study heterogeneity. Influencer analysis will take place according to the leave-one-out method. Influential studies will be identified using a random-effects model. Influencer analysis will be visualised using a Baujat plot and plots of influence diagnostics (including externally standardised residuals, DFFFITS value, Cook’s distance, covariance ratio, leave-one-out τ2 and Q values, hat values and study weights). The overall effect and I2 will be visualised in forest plots. Outliers and influential studies will be excluded from the meta-analysis as part of a sensitivity analysis. Publication bias will be assessed using funnel plots. Given sufficient homogeneity of reported outcomes and enough studies reporting outcomes on infants (children <1 year of age), a subgroup analysis will be undertaken to identify any differences in outcomes between infants and children over 1 year of age.
Data cleaning and visualisation will be undertaken in R using the tidyverse, dplyr and ggplot packages. Meta-analyses will be conducted in R using the meta package in accordance with Harrer et al.11 If there is insufficient data to conduct a meta-analysis, only thematic synthesis will be performed.
Patient and public involvement
The public and patients were not involved in the development of this systematic review and meta-analysis protocol.
Discussion
Dismembered pyeloplasty remains the gold standard of treatment of patients with PUJO.2 However, the optimal method of surgical access has not yet been determined based on key postoperative outcomes.12 Given the relatively recent development of RALP, this triplicate comparison has not been evaluated in the literature yet.2 Our review seeks to provide a comprehensive overview of the literature surrounding the three approaches to an Anderson-Hynes dismembered pyeloplasty and evaluate their efficacy both in isolation and when compared against each other on the basis of key postoperative outcomes.
The importance of this comparison cannot be understated. Determining the standard of care is a monumental undertaking, especially when the significant costs associated with robotic surgery (both in its undertaking and in the training of surgeons) are considered. For example, the standard of care for children with PUJO is a RALP in the UK; however, this is not yet the case in less affluent nations.3 If children undergoing RALP for PUJO are demonstrated to experience definitively better outcomes than children undergoing laparoscopic or open pyeloplasty, then children from poorer nations face worse surgical outcomes on the basis of their socioeconomic status. If, however, RALP is demonstrated to have similar or identical outcomes to open and laparoscopic techniques, then the additional cost of the procedure is negligible.13 A decision to switch to a more expensive procedure cannot be made without careful comparisons between the three groups, which our paper aims to highlight.
In the coming years, we are likely to see a rise in the number of RALPs performed on children around the world, and as such, we should aim to better understand the indications and outcomes of this procedure in children. Through systematic review and meta-analysis, we aim to identify commonality between studies that have investigated the outcomes of RALP and compared it to open or laparoscopic pyeloplasty.
Trial status
-
Preliminary searches: started.
-
Piloting of the study selection process: started.
-
Formal screening: started.
-
Data extraction: not started.
-
Risk of bias assessment: not started.
-
Data analysis: not started.
Draft of search strategy for MEDLINE, EMBASE, PubMed and Cochrane databases
Search terms used are (pyeloplasty) AND ((laparoscopic) OR (robotic) OR (open)) AND ((Ureteropelvic junction obstruction) OR (pelviureteric junction obstruction)) AND ((child) OR (p?ediatric)) AND (outcome).
Ethics Statement
Due to the nature of the present study, no relevant ethical concerns or informed consent will be required. The protocol and systematic review and meta-analysis will be disseminated through publication in a peer reviewed journal.
Ethics statements
Patient consent for publication
This post was originally published on https://bmjopen.bmj.com