This protocol has been designed according to the Preferred Reporting Items for Systematic review and Meta-Analysis Protocol (PRISMA-P) (27,28). The design and method has been formed through discussion between experts in the management of DDH and experts in the methodology of systematic reviews. The protocol has been registered in the International Prospective Register of Systematic Reviews (PROSPERO ID - CRD42022362325).
Eligibility criteria
Population
Children with idiopathic bilateral developmental dysplasia of the hip undergoing surgical management.
Exclusion criteria - children with teratologic bilateral developmental dysplasia of the hip, children undergoing revision surgery, surgery for acetabular dysplasia in adolescence.
Intervention
Medial approach open reduction of the hip (MAOR)
Comparison
Anterior approach open reduction of the hip (AOR)
Outcomes
1. Rate and severity of avascular necrosis of the femoral head at latest follow up using Kalamchi and MacEwen(29) or Bucholz and Ogden classification(30) or other appropriate scoring system.
and/or
2. Radiological outcome at latest follow-up using acetabular index measured in degrees, Severin Score(31) or other appropriate scoring system.
and/or
3. Clinical outcomes at latest follow up including Modified McKay criteria(32), Children's Hospital of Oakland Hip Evaluation Scale(33), Pediatric Outcomes Data Collection Instrument (PODCI)(34) or other appropriate scoring system.
and/or
4. Prevalence, event rate or time to event surgical complications assessed according to the Clavien-Dindo system(35,36) or other appropriate scoring system.
and/or
5. Prevalence, event rate or time to event of secondary surgery.
Study design
Inclusion criteria - Clinical studies, level IV (retrospective case series) and above, with a clear description of the operative management with clinical and/or radiological outcomes, published in English.
Exclusion criteria - case reports, technical or cadaveric studies, studies without a clear description of the operative management or where this is unobtainable, studies without a clear description of clinical and/or radiological outcomes or where this is unobtainable. Full text studies not available in English will be excluded.
Search Strategy
A search of the electronic medical and scientific databases; PubMed, MEDLINE, the Cochrane Library, Embase, Google Scholar, Web of Science and Scopus will be conducted from the date of first entry until date of search. The grey and difficult to locate literature (including theses and dissertations) will be searched via the Open Grey(37) and Open Access Theses and Dissertations(38) databases. The Medical Subject Headings (MeSH terms) “developmental dysplasia of the hip”, “congenital dysplasia of the hip”, “congenital hip dislocation”, “developmental hip dislocation”, and their abbreviations, “DDH” and “CDH” will be used, along with the qualifier “bilateral”. The search strategy will be developed in Medline and then applied to other databases. An example of the search strategy can be found in additional file 1. Only full text studies, published in English will be included. There will be no time limit imposed.
Study Selection
Two reviewers (EJ & GC) will independently screen the title and abstract of records for inclusion according to the eligibility criteria. Once preliminary screening has been performed, selected studies will be screened as full text. Researchers will be blinded to each other's decisions. Where there is disagreement a separate reviewer (CEB) will arbitrate. Screening decisions at the full text stage will be fully recorded. The results of screening will be presented in a PRISMA flow diagram(28).
Data Management
The selected studies will be collated in Zotero citation management system, screened for duplicates and exported to Systematic Review Data Repository-Plus(39). This database will be used to aid data extraction and management. Extracted data will be exported to RevMan software for analysis.
Data extraction
Data will be extracted in a predefined electronic data extraction form. Data on study design, population characteristics, details of operative intervention (intervention and comparison) and outcomes (clinical, radiological, complications and rate of secondary surgery) will be extracted. A summary of intended data items for extraction is shown in Table 1. Four reviewers (EJ, GC, MJC & AB) will be allocated the selected studies and will independently extract data. Each reviewer will be blinded to data extraction. Where possible corresponding authors will be contacted for unreported data. Data will be extracted to a secured anonymised form on Systematic Review Data Repository-Plus and then exported to RevMan for analysis.
Table 1. Items included for data extraction in selected studies
Data Item
|
Details
|
Study details
|
Reference, year of publication, geographical location of study, study design, ethical approval, funding, pre-registered protocol
|
Population details
|
Number of patients, age, sex, comorbidities, inclusion/exclusion critieria, duration of follow-up, DDH severity or classification, presence of ossific nucleus
|
Surgical intervention
|
Anaesthetic used (incl. Nerve blocks/ spinal), single stage or sequential surgery, Surgical approach, anatomical details of approach, blocks to reduction addressed and how addressed, presence or absence of ligamentum teres, surgical duration, volume of blood loss, details of additional procedures (e.g. pelvic osteotomy/femoral osteotomy), method of post-operative immobilisation, duration of post-operative immobilisation
|
Clinical outcomes - including hip specific outcome scores and general quality of life scores
|
Eg - Modified McKay criteria(32), Children's Hospital of Oakland Hip Evaluation Scale(33), Pediatric Outcomes Data Collection Instrument (PODCI)(34)
|
Radiological outcomes - including hip and age specific outcome scores
|
E.g. Rate and severity of AVN (Kalamchi and MacEwen(29) or Bucholz and Ogden classification(30)), acetabular index measured in degrees, Severin Score(31)
|
Surgical outcomes
|
Surgical compilation rate, surgical complications according to Clavien-Dindo classification,dislocation rate, secondary surgery
|
Data Synthesis
The extracted data will be summarised in a structured table format, grouped and ordered by study design (according to the hierarchy of evidence) or by risk bias if study designs are similar, and including the data items specific to the outcomes of interest. This will help to assess clinical and methodological heterogeneity across the studies and determine the feasibility of performing a meta-analysis. We do not expect the included studies to be of sufficient quality or consistency to allow a meta-analysis to be performed. In this instance, we will follow the Systematic Review without Meta-Analysis (SWiM) guidance(40) and analyse data according to this and the recommendations in the Cochrane Handbook Chapter 12(41). Studies will be grouped and tabulated as described. We expect that the key outcome data for radiological and clinical outcomes will be in short ordinal scales (e.g. Severin Score(31)). Where possible we will transform these data to dichotomous outcomes and present this as a relative risk with 95% confidence intervals for MAOR in comparison to AOR. Longer ordinal scales such the Pediatric Outcomes Data Collection Instrument(34) will be transformed to continuous data. For complications and secondary surgery data we will transform to an incidence estimate, event rate or time to event data. For non-comparative studies we will transform extracted data as described above and use this to generate a crude estimate of incidence, prevalence or event rate. Where possible we will pool this data using a random effects model as per the recommendation in Murad et al(42). Results will be reported according to the guidance in the Cochrane Handbook Chapter 12(41). Where sufficient information is available but synthesis cannot be performed a structured reporting of effects will be used. When effect estimates are available without measures of precision an illustrated synthesis of summary statistics will be used. If P values are available an illustrated synthesis of P values will be used. Where directions of effect are available an illustrated synthesis using vote-counting based on direction of effect will be used.
Meta-bias
We aim to limit publication bias by a thorough and systematic search of the literature including the grey literature as described in the search strategy. Where possible publication bias will be assessed across studies by generation of funnel plots. These will be inspected for asymmetry and analysed via Egger’s test(43).
Risk of Bias
Randomised trials will be assessed using the Cochrane Risk of Bias 2 (RoB 2) tool(44). However, included studies are most likely to be non-randomised, observational studies. For comparative studies (cohort or case-control) we will use the ROBINS-I tool to assess risk of bias(45). For case series, we will use Murad et al’s method for evaluating the methodological quality across four domains; selection, ascertainment, causality and reporting(42). Four reviewers (EJ, GC, MJC & AB) will assess included studies for risk of bias. A separate reviewer (CEB) will resolve disagreements through discussion. A summary figure of the risk of bias analysis will be included in the final manuscript.
Assessment of Quality
For each outcome measure a summary of findings will be presented in a table(46) with the overall quality of the recommendation assessed using the Grading of Recommendations Assessment Development and Evaluation approach (GRADE)(47). This approach uses five factors; risk of bias, inconsistency, indirectness, imprecision and publication bias to assess the quality of evidence and produce a rating of “high”, “moderate”, “low” or “very low”. GRADEpro GDT software(48) will be used to aid decision making when assessing quality of evidence.