Criteria for considering studies for this review
Types of studies
Only studies where participants were either randomised or quasi-randomised (method of allocating participants to a treatment which is not strictly random, for example by date of birth, hospital record number, or alternation) into intervention groups will be included in this review.
Types of participants
There are no established criteria for FAI (Ayeni 2013). The diagnosis is generally made on the basis of symptoms of hip or groin pain, or both, restricted range of motion and a positive anterior impingement test (when the hip joint is placed in a position of flexion, adduction and internal rotation pain may be experienced), and the presence of abnormal hip shape morphology and abnormalities of the adjacent labrum and cartilage on imaging. The hip shape imaging should include cross-sectional studies. These may be: computed tomography (CT), magnetic resonance imaging (MRI), or magnetic resonance arthrography (MRA) (Ganz 2003; Beall 2005).
Trials that include participants with FAI as determined by the trial authors will be included provided they conform to the above criteria. We will systematically review the definitions of FAI used in the included trials. If definitions of FAI differ it might not be appropriate to pool data. We will explore whether differences in how FAI is defined in the trials (inclusion criteria) make any difference to the results.
If the trials have included people with established osteoarthritis (OA) then we will include the trials only if data for people without established OA are presented separately, or if people with OA make up 10% or less of the study population. We will include definitions of OA as per the included trials. Although current evidence suggests that FAI is a risk factor for subsequent OA (Gosvig 2010), there is no suggestion in the literature that surgery for FAI is a treatment for established OA.
We will exclude studies that have > 10% of participants with evidence of OA.
Types of interventions
Studies of all types of FAI surgery will be included. Surgery could be performed using open, mini-open, arthroscopic, assisted mini-open or arthroscopic approaches, and the interventions could consist of:
1. reshaping of the hip joint by removing bone or cartilage, or both (osteoplasty, osteochondroplasty) from either the femoral head-neck junction or rim of the acetabulum;
2. reorientating the hip joint by cutting the bones around the hip joint (osteotomy) and refixing the bones in a new orientation. The new orientation of the hip should reduce the risk of future FAI. The bony reorientation can be done for the femur or acetabulum, or both.
Comparators could be:
1. placebo (sham surgery);
2. no treatment;
3. non-operative treatment (e.g. physical therapy, analgesia, glucocorticoid injection, activity modification).
Types of outcome measures
We will not exclude studies on the basis of outcome measures. We will, however, use a hierarchy when multiple tools are used to measure the same functional outcomes. This hierarchy is based upon recent work and evidence supporting the use of a set of core outcomes for painful musculoskeletal conditions (Eccleston 2010; Lodhia 2010).
1. Efficacy: proportion of participants with 30% reduction in pain or greater.
2. The proportion of participants with any adverse events (AEs). AEs are defined as any untoward medical occurrence in a clinical trial patient and which do not necessarily have a causal relationship with the treatment. Possible AEs include deep vein thrombosis (DVT), numbness, and muscle soreness.
3. Pain reported as:
a. proportion with 50% reduction in pain or greater;
b. proportion below 30/100 mm on the visual analogue scale (VAS);
c. mean change or mean absolute pain score on a VAS or numerical rating scale;
d. if pain is not reported on its own but as a hip specific multidomain outcome assessment, an attempt will be made to obtain the breakdown of the score in order to extract the data on pain.
4. Proportion of participants with a serious adverse event (SAE). SAEs are defined as adverse events that are fatal, life-threatening, or require hospitalisation). Possible SAEs include: death; pulmonary embolism; fluid extravasation; fracture; and avascular necrosis.
5. Hip function, e.g. measured using hip specific multidomain outcome measures such as the Non-Arthritic Hip Score and International Hip Outcome Tool (iHOT-33) etc. If more than one measure of function is included in the studies we will use the following hierarchy:
a. Non-Arthritic Hip Score (NAHS);
d. Hip Outcome Score (HOS);
e. Modified Harris Hip Score (MHHS);
f. Vail Hip Score;
g. Western Ontario and McMaster Universities Arthritis Index (WOMAC).
A recent review of hip specific multidomain outcome measures recommended that, of established hip specific multidomain outcome measures, the NAHS and the HOS had the strongest clinimetric evidence to support their use as primary outcome instruments in studies for the effectiveness of treatment of FAI (Lodhia 2010). The iHOT-33 and iHOT-12 have been recently developed for active patients aged 18 to 60 years who present with a variety of hip conditions (Griffin 2012; Mohtadi 2012).
6. Quality of life, as measured by generic instruments such as the: Short Form (SH)-36 (SF-36), SF-12, EuroQol-5D (EQ-5D).
7. Participant global assessment of treatment success.
Our research question is on the efficacy and safety of FAI surgery. We regard an effect on progression to OA to be an important outcome in the longer term (Ganz 2003; Beck 2005). However, it is unlikely that if trials exist they will have the long term follow-up to answer this question. We believe the first and most important question is whether surgery in the short to medium term is both effective (relieves pain) and safe.
Timing of outcome assessment
Studies are likely to report the outcomes discussed at several time points. We will attempt to group these assessments into three categories: short (up to and including three months), medium (after three months and up to and including 12 months), and long term follow-up (greater than one year). We will extract AEs at the end of the trial.
Search methods for identification of studies
We will search the following electronic databases, unrestricted by date or language:
Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (current Issue);
MEDLINE (Ovid) (1946 to present); and
EMBASE (Ovid) (1980 to present).
In MEDLINE, we will combine a subject speciﬁc search with the Cochrane highly sensitive search strategy for identifying randomised trials, sensitivity-maximising version (Lefebvre 2011). There will be no language restriction. The strategy will be designed in MEDLINE and adapted to the other databases. The MEDLINE search strategy is shown in Appendix 1.
Trial registries (World Health Organization (WHO) International Clinical Trials Registry Platform at http://apps.who.int/trialsearch/; ClinicalTrials register at http://www.clinicaltrials.gov/; Current Controlled Trials register at http://www.controlled-trials.com/) will be searched to identify trials that are currently underway.
Searching other resources
Reference lists of all primary studies and review artciles will be searched for additional references. We will not search informally published written material (so called 'grey literature').
Data collection and analysis
Selection of studies
Two authors (PW and JB) will independently select the studies for inclusion in the review. Titles and abstracts obtained from the searches will be reviewed to determine potential eligibility and short listed if appropriate. The full text of each study in this shortlist will then be reviewed to determine which studies are eligible for inclusion in the review. Any disagreement between the two authors will be resolved by consensus or discussion with a third review author (MC or DG). Studies will be translated into English where necessary.
Data extraction and management
Two review authors (PW and JB) will independently extract the following data from the included trials and enter the data in RevMan 5:
1. trial characteristics including size and location of the trial, and source of funding;
2. characteristics of the study population, including age, and characteristics of the FAI including diagnostic criteria, type and duration of symptoms;
3. characteristics of the surgery and comparator treatment, including surgical approach, type of FAI being addressed (Cam, Pincer or mixed), type of intervention used to correct the FAI (osteochondroplasty or osteotomy);
4. risk of bias domains as outlined in Assessment of risk of bias in included studies, below;
5. outcome measures as mean and standard deviation, number of participants per treatment group for continuous outcomes (pain when reported as either a mean or change in pain score, hip function, and quality of life), and number of events and number of participants per treatment group for dichotomous outcomes (efficacy, pain when reported as a proportion, SAEs, AEs, and participant global assessment of treatment success).
If additional data are required, we will contact the trial authors to obtain these. Where data are imputed or calculated (Dealing with missing data) we will report this in the 'Characteristics of included studies' table. Any disagreements will be resolved by consensus and resolved by a third review author (RB). Extracted data from the studies will be managed and collated by the review statistician (NP).
Assessment of risk of bias in included studies
Studies that are included in the review will each be assessed for risk of bias using the recommended Cochrane Collaboration ’Risk of bias’ tool (Higgins 2011a). This tool incorporates assessment of randomisation (sequence generation and allocation concealment), blinding (participants, personnel and outcome assessors), completeness of outcome data, selection of outcomes reported, and other sources including any potential conflicts of interest for industry funded research. To determine the risk of bias of a study, we will assess each criterion for the presence of sufficient information and the likelihood of potential bias. Each criterion will be rated as ‘Low risk’ of bias, ‘High risk’ of bias or ‘Unclear risk’ of bias (uncertain of the potential for bias, or insufficient information reported to make an assessment). In a consensus meeting disagreements among the review authors will be discussed and resolved. A third review author (RB) will be available to make the final decision if no consensus can be reached.
Measures of treatment effect
Risk ratios with 95% confidence intervals (CIs) will be used to express the intervention effect for the following dichotomous outcomes:
pain, when reported as a proportion of participants within defined limits (i.e. reduction in pain of 30% or greater, 30/100 mm or less on VAS);
participant global assessment of treatment success.
Where dichotomous data from cross-over trials are combined with data from parallel-group trials, the odds ratio (OR) with 95% CI will be calculated, rather than relative risk (RR).
We will calculate mean difference (MD) or, where studies have used different measurement tools, standardised mean difference (SMD), both with 95% CIs, for the following continuous outcomes:
Unit of analysis issues
It is expected that most studies will report outcomes at a number of follow-up times, for example at three, six and 12 months. We will therefore aim to extract the data at three time points: ≤ three months; > three and < 12months; and ≥ 12months. If there are multiple time points within each category, we will extract the data closest to each time point (three, six and 12 months). It is expected that all studies will report simple parallel-group designs. However, if other designs are reported (for example cluster randomised designs), and where appropriate, generic inverse variance methods will be used to combine data. In the analysis we will use details of intraclass correlation coefﬁcients and cluster sizes (if available) to correct the sample sizes for trials of this type if reported effects have not been adjusted for clustering. For studies containing more than two intervention groups, making multiple pair-wise comparisons between all possible pairs of intervention groups possible, we plan to include the same group of participants only once in the meta-analysis.
Where studies have included either same day or delayed bilateral FAI hip surgery this information will be extracted and reported in the summary of findings along with the corresponding unit of analysis (either participants or hips). Studies that use the hip as the unit of analysis would be regarded as potentially inferior because the outcome variable would no longer be independent.
Dealing with missing data
We will seek additional information from the authors of the included studies where the published information or data are incomplete. In cases where individuals are missing from the reported results, we will assume that the missing value had a poor outcome. For dichotomous outcomes that measure SAEs and AEs (for example number of SAEs), the number of patients that received treatment will be used as the denominator (worst case analysis). For dichotomous outcomes that measure benefits, the worst case analysis will be calculated using the number of randomised participants as the denominator. For continuous outcomes (for example pain) we will calculate the mean difference (MD) or standardised mean difference (SMD) based on the number of patients at the time point. If the numbers of patients are not presented for each time point, the numbers of randomised patients in each group at baseline will be used. Sensitivity analysis will be conducted to test the effect of these assumptions.
Where possible, missing standard deviations will be computed from other statistics such as standard errors, confidence intervals (CI) or P values according to the methods recommended in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). If small amounts of outcome data are missing (for example standard deviations) we will consider imputing them (with the appropriate sensitivity analyses) from other studies (Higgins 2011b).
Assessment of heterogeneity
For any studies judged as clinically homogenous, the degree of statistical heterogeneity between studies will ﬁrst be assessed graphically using a forest plot and more formally using the I² statistic (Deeks 2011). The following will be used as a rough guide for interpretation: 0% to 40% might not be important, 30% to 60% may represent moderate heterogeneity, 50% to 90% may represent substantial heterogeneity, and 75% to 100% considerable heterogeneity. In cases of considerable heterogeneity (defined as I2 ≥ 75%) we plan to explore the data further, including undertaking subgroup analyses, in an attempt to explain the heterogeneity.
Assessment of reporting biases
In order to determine whether reporting bias is present, we will determine whether the protocol of the RCT was published before recruitment of patients into the study was started. For studies published after 1 July 2005, we will screen the clinical trial register in the WHO International Clinical Trials Registry Platform (http://apps.who.int/trialsearch) (DeAngelis 2004). We will evaluate whether selective reporting of outcomes was present (outcome reporting bias).
We will compare the fixed-effect model estimate against the random-effects model to assess the possible presence of small sample bias in the published literature (that is where the intervention effect is more beneficial in smaller studies). In the presence of small sample bias, the random-effects model estimate of the intervention is more beneficial than the fixed-effect model estimate (Sterne 2011).
The potential for reporting bias will be further explored by funnel plots if ≥ 10 studies are available.
Where studies are sufficiently homogeneous that it is clinically meaningful for them to be pooled, meta-analysis will be performed using a random-effects model, regardless of the I2 results. Analysis will be performed using Review Manager 5 and forest plots produced for all analyses. Where both change and mean pain are reported in studies, we will pool the results based upon whichever is reported by the majority of the included studies.
Subgroup analysis and investigation of heterogeneity
The inclusion of intervention effects at a number of time points (for example three, six and 12 months) will provide some sensitivity to the selection of an appropriate follow-up time for assessment of the treatment effect.
Where sufficient data are available, the following subgroup analyses are planned.
1. Cam versus Pincer type FAI. These two shape abnormalities arise from different aspects of the hip joint and therefore the results of surgery may differ between the two types. Subgroup analysis will measure the results of surgery using pain as the outcome.
2. Open versus arthroscopic surgery. Previous research suggests that the outcome in terms of pain and function is comparable for open and arthroscopic surgery, but that the risk of AEs are greater from open surgery (Botser 2011).
Subgroup analysis will measure the results of surgery using pain and the proportion of participants with any AEs as the outcomes. The subgroup analysis will also informally compare the magnitudes of effect to assess possible differences in response to treatment by considering the overlap of the confidence intervals of the summary estimates in the two subgroups; non-overlap of the confidence intervals indicates statistical significance.
If it is necessary to exclude any studies because they appear to differ markedly (that is if the outcome is different, the effect goes in the opposite direction) from the majority of studies then all main analyses will be reported with and without these studies.
Where sufficient studies exist, sensitivity analyses are planned to assess the impact on the primary outcome of any bias attributable to inadequate or unclear treatment allocation (including studies with quasi-randomised designs) and blinding.
Summary of findings tables
We will present the main results of the review in 'Summary of findings' (SoF) tables, which provide key information concerning the quality of evidence, the magnitude of effect of the interventions examined, and the sum of the available data on the outcomes (efficacy, SAEs, pain, AEs, hip function, participant global assessment, quality of life), as recommended by The Cochrane Collaboration (Schünemann 2011a). The SoF table includes an overall grading of the evidence related to each of the main outcomes, using the GRADE approach (Schünemann 2011b).
In addition to the absolute and relative magnitude of effect provided in the SoF table, for dichotomous outcomes the number needed to treat to benefit (NNTB) or the number needed to treat to harm (NNTH) will be calculated from the control group event rate (unless the population event rate is known) and the risk ratio using the Visual Rx NNT calculator (Cates 2008). For continuous outcomes, the NNT will be calculated using the Wells calculator software available at the CMSG editorial office (www.cochranemsk.org). The minimal clinically important difference (MCID) for each outcome will be determined for input into the calculator.
Minimal clinically important differences (MCIDs) have been defined (where known) for each continuous outcome, see below.
i. VAS: no published references when applied to FAI. However, when used for patients with hip and knee OA treated with non-steroidal anti-inflammatory medications an MCID of 15.3 mm has been reported (Tubach 2005).
i. NAHS: no published references.
ii. iHOT-33: no published references when applied to FAI. However, when used for patients undergoing hip arthroscopy an MCID of 6.1 points (100 point scale) is reported in the original paper and more recently < 11 points has been reported (Mohtadi 2012; Kemp 2013).
iii. iHOT-12: not published but authors suggest similar performance to iHOT-33 (Griffin 2012).
iv. HOS: no published references.
v. MHHS: no published references when applied to FAI. However, when used for patients undergoing hip arthroscopy an MCID of < 11 points (100 point scale) is reported (Kemp 2013).
vi. Vail Hip Score: no published references.
vii. WOMAC: no published references when applied to FAI. However, when used for patients undergoing total hip arthroplasty at six months the MCID exceeds 25 points (100 point scale) for all domains (pain, functional limitation, and stiffness) (Quintana 2005).
Quality of life
i. SF-36: no published references when applied to FAI. However, when applied to patients undergoing total hip arthroplasty the following are reported for each domain, 11 points for physical role and 20 points for physical function (using a 100 point scale) (Keurentjes 2012).
ii. SF-12: no published references when applied to FAI. However, SF-12 is a shortened derivative of SF-36 and as a result correlates closely (Ostendorf 2004). A published report for unrelated surgery (cervical decompression) reports an MCID of 2.5 points for the physical component and 10.1 points for the mental component (Parker 2012).
iii. EQ-5D: no published references when applied to FAI. However, more empirical analysis across several applications suggests (across a continuous scale from -0.59 to 1) a mean of 0.074 (range -0.011 to 0.140) (Walters 2005).