Criteria for considering studies for this review
Types of studies
We will include RCTs of parallel design (published or unpublished). We will not include quasi-randomised studies. Additionally, we will not include cross-over studies as we do not feel this study design is appropriate given that the intervention aims to correct the underlying defect.
Types of participants
We will include studies involving children or adults with CF, as confirmed either by the presence of two disease-causing mutations, or by a combination of positive sweat test and recognised clinical features of CF. We will include studies that include patients with any level of disease severity. Participants should have at least one class II mutation.
Types of interventions
We will include studies in which CFTR correctors are compared with either placebo or another intervention. We will also include studies in which CFTR correctors are administered alongside another class of drug that also aims on improve CFTR function (e.g. potentiators).
A CFTR corrector is defined as a drug which aims to increase the amount of CFTR expressed at the epithelial cell apical membrane, by reducing or preventing degradation of CFTR by normal intracellular mechanisms.
Types of outcome measures
Quality of life (QoL) (measured using validated quantitative scales or scores (e.g. Cystic Fibrosis Questionnaire-Revised (CFQ-R) (Quittner 2009))
total QoL score
different sub-domains which may be reported
Physiological measures of lung function (litres or per cent predicted for age, sex and height)
forced expiratory flow rate at one second (FEV1) (relative change from baseline)
FEV1 absolute values (rather than change from baseline)
forced vital capacity (FVC) (absolute values and change from baseline)
graded by review authors as mild (therapy does not need to be discontinued)
graded by review authors as moderate (therapy is discontinued, and the adverse effect ceases)
graded by review authors as severe (life-threatening or debilitating, or which persists even after treatment is discontinued)
other adverse effects of therapy (of any severity) that are not classifiable according to these categories
number of days
number of episodes
time to next hospitalisation
School or work attendance (i.e. number of days missed)
Extra courses of antibiotics (measured as time-to the next course of antibiotics and the total number of courses of antibiotics)
Sweat chloride (change from baseline) as a measure of CFTR function
Radiological measures of lung disease (assessed using any scoring system)
chest radiograph scores
computerised tomogram (CT) score
Acquisition of respiratory pathogens
Pseudomonas aeruginosa (P. aeruginosa)
Staphylococcus aureus (S. aureus)
Haemophilus influenzae (H. influenzae)
other pathogen clinically relevant in CF
Eradication of respiratory pathogens (as defined by study authors)
other pathogen clinically relevant in CF
Nutrition and growth (measured as relative change from baseline) (including z scores or centiles)
body mass index (BMI)
Search methods for identification of studies
We will identify relevant studies from the Cochrane Cystic Fibrosis and Genetic Disorders Group's Cystic Fibrosis Trials Register using the terms: 'drugs that correct the molecular defect', and 'mutation-specific therapies'. Relevant studies have been tagged with these terms for indexing purposes in the Group's Cystic Fibrosis Trials Register.
The Cystic Fibrosis Trials Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL) (updated each new issue of The Cochrane Library), quarterly searches of MEDLINE, a search of Embase to 1995 and the prospective handsearching of two journals - Pediatric Pulmonology and the Journal of Cystic Fibrosis. Unpublished work is identified by searching the abstract books of three major cystic fibrosis conferences: the International Cystic Fibrosis Conference; the European Cystic Fibrosis Conference and the North American Cystic Fibrosis Conference. For full details of all searching activities for the register, please see the relevant sections of the Cystic Fibrosis and Genetic Disorders Group Module.
We will also search clinical trial registries maintained by the European Medicines Agency (https://www.clinicaltrialsregister.eu/), the US National Institute of Health (http://clinicaltrials.gov/) and the World Health Organisation (http://www.who.int/ictrp/en/).
Searching other resources
We will screen references of included studies to identify additional potentially relevant studies. We will also contact authors of included studies, leaders in the field, and companies known to be developing and investigating CFTR correctors, to identify any studies which may have been missed by this search. We will record response rates from this contact process.
Data collection and analysis
Selection of studies
Two authors (IS and SP) will independently assess the suitability of each potential study identified by the search. If disagreement arises on the suitability of a study for inclusion in the review, we will attempt to reach a consensus by discussion, failing which, a third author (KWS) will arbitrate.
Data extraction and management
Two authors (IS and SP) will independently extract relevant data from each included study. If disagreement arises on data extraction, we will attempt to reach a consensus by discussion, failing which, a third author (KWS) will arbitrate.
We will report our primary outcome 'survival' either as a binary outcome or a time-to-event outcome. We will extract QoL scores ideally as relative change from baseline ((measurement at end of treatment - measurement at baseline) / measurement at baseline) x 100). We will extract data presented as post-treatment values or change from baseline if this is not possible.
With regards to the secondary outcome 'Extra courses of antibiotics', this outcome may be measured in a number of ways, and we will ideally extract data for time-to the next course of antibiotics and the total number of courses of antibiotics. We will note whether these are physician-defined or protocol-defined.
We plan to report data as immediate (up to and including one month), short term (over one month and up to six months) and long term (over six months).
Assessment of risk of bias in included studies
Two authors (IS and SP) will assess the risk of bias for each trial using the Cochrane risk of bias tool (Higgins 2011a). This includes assessment of the following methodological aspects of the included studies:
procedure for randomisation (selection bias);
allocation concealment (selection bias);
masking (blinding) of the intervention from participants, clinicians, and trial personnel evaluating outcomes (performance bias);
missing outcome data (attrition bias);
selective outcome reporting (reporting bias);
other sources of bias (e.g. the influence of funding sources or industry on trial characteristics and presented results).
We will also assess whether all participants were included in an intention-to-treat analysis, regardless of whether they completed the treatment schedule or not. If disagreement arises on the assessment of risk of bias of a study, we will attempt to reach a consensus by discussion, failing which, a third author (KWS) will arbitrate.
Measures of treatment effect
For binary outcomes, such as survival, where participants will be reported as either deceased or alive, we will calculate a pooled estimate of the treatment effect for each outcome using the pooled odds ratio (OR) and 95% confidence intervals (CIs).
For continuous outcomes, such as weight (kg), we will calculate the mean change from baseline for each group or the mean post-intervention values and standard deviation (SD) for each group. We will convert standard errors to SDs. We will produce a pooled estimate of treatment effect by calculating the mean difference (MD) and 95% CIs. For QoL, we anticipate that the CFQ-R will be the most frequently used questionnaire and if studies only report CFQ-R, we will calculate the MD and 95% CIs; but if studies use other questionnaires, we will calculate the standardised mean difference (SMD) and 95% CIs to allow for the different scales. In a similar way, if studies report radiological scores using different scales, we will present these data using the SMD and 95% CIs.
For time-to-event outcomes, such as 'time to next extra course of antibiotics', we will use measures of survival analysis, and calculate hazard ratios (HR) and 95% CIs between different arms of the study.
When different studies present data for the same outcomes in different forms (e.g. absolute values of lung function measures, or change in these measures from a baseline), we will combine these in a meta-analysis.
If the studies do not report change data, but instead present absolute post-treatment data without baseline data so it is not possible to calculate change data, we will use absolute post-treatment data instead of change from baseline. However, if the report presents baseline and post-treatment data for any outcome, we will calculate SDs for the change from baseline, for example if the CI is available. When there is not enough information available to calculate the SDs for the changes, we can impute them from other studies in the review where the data are available and studies are similar (i.e. whether the studies used the same measurement scale, had the same degree of measurement error and had the same time periods between baseline and final value measurement). If neither of these methods are possible, we will impute a change-from-baseline SD in another study, making use of an imputed correlation coefficient (following methods described in section 126.96.36.199 in the Cochrane Handbook of Systematic Reviews of Interventions (Higgins 2011b)).
When reporting on outcomes we will use the following subheadings to describe the time points: immediate; short term; and longer term.
Unit of analysis issues
Within this review, we will only include results from RCTs of parallel design in which individual trial participants are randomised. We will exclude cross-over studies, as they are not appropriate for evaluating therapies in a progressive condition such as CF (Higgins 2011b). We will also exclude cluster RCTs, as the focus of this review relates to individual patients with CF.
Dealing with missing data
In order to allow an intention-to-treat analysis, we will extract data on the number of participants with each outcome event, by allocated treated group, irrespective of compliance and whether or not the participant was later thought to be ineligible or otherwise excluded from treatment or follow up. If any data are missing or unclear, we will contact the primary investigators for clarification.
Assessment of heterogeneity
We will assess heterogeneity through visual examination of the combined data presented in the forest plots, and by considering the I2 statistic (Higgins 2003) together with chi2 values (Deeks 2011) and their CIs. This reflects the likelihood that variation of results across studies are due to heterogeneity rather than by chance, and we will interpret this statistic using the following simple classification:
0% to 40%: might not be important;
30% to 60%: may represent moderate heterogeneity;
50% to 90%: may represent substantial heterogeneity;
75% to 100%: considerable heterogeneity.
Assessment of reporting biases
In order to identify selective outcome reporting, where possible, we will compare outcomes described in the protocol with those reported in the publication. We will request protocols for specific studies from the primary investigators, corresponding author, or relevant pharmaceutical company. We will record the proportion of protocols that were available to us. If a protocol is not available, information about outcomes may be available from trial registry databases. We will also compare outcomes listed in the 'Methods' section of the final paper with those presented in the 'Results' section. If the published papers reported negative findings either only partially, or not at all, we will contact primary investigators for these data.
We will assess publication bias by constructing and assessing the symmetry of a funnel plot. This will be possible if we include more than 10 studies in the review. We will plot the number of patients in the study against a measure of treatment effect. If the funnel plot is asymmetrical, we will consider whether this is due to publication bias, or whether methodology or small sample size has caused results of certain studies to show exaggerated treatment effects.
We will be assessing different CFTR correctors during this review, so we can not assume that there will be a single common true effect. In addition to this, participants in each study will vary due to different eligibility criteria employed by the authors. Therefore, regardless of the I2 value, we will use a random-effects model to analyse data from studies.
Subgroup analysis and investigation of heterogeneity
We will investigate any heterogeneity that we identify using subgroup analyses of potential confounding factors, if sufficient numbers (at least 10 studies) are available. For this review, these confounding factors will be:
As we will not seek individual patient data from study investigators, we will not undertake a subgroup analysis on the basis of disease severity. We may incorporate such an analysis in future updates of this review.
We will examine the impact of bias on the results examined by comparing meta-analyses including and excluding studies with concerns of high risk of selection or reporting bias due to issues relating to randomisation, allocation concealment, or masking of interventions from participants or study personnel.