Types of studies

Randomized or quasi‐randomized controlled trials were eligible for inclusion. Cross‐over and cluster‐randomized trials were not eligible for inclusion.

Types of participants

People with CF and ABPA. Diagnosis of CF should be in accordance with the criteria laid down by the Cystic Fibrosis Foundation Consensus Report (Farrell 2008); ABPA should be diagnosed using the Rosenberg‐Patterson criteria (Rosenberg 1977), Nelson’s criteria (Nelson 1979), Greenberger’s criteria (Greenberger 2002) or the Cystic Fibrosis Foundation Consensus Criteria (Stevens 2003). There was no limit to age or disease severity for participants included in the review.

Types of interventions

Anti‐IgE therapy compared to placebo or other therapies for ABPA in people with CF. We considered all doses of anti‐IgE therapy in the review.

Types of outcome measures

Electronic searches

We identified relevant studies from the Cochrane Cystic Fibrosis and Genetic Disorders Review Group's CF Trials Register using the term: allergic bronchopulmonary aspergillosis.

The CF Trials Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL) (updated each new issue of the Cochrane Library), weekly 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 Unpublished work is identified by searching the abstract books of three major cystic fibrosis conferences: the International Cystic Fibrosis Conference; the European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's website.

Date of last search of CF Trials Register: 29 September 2017.

We searched two ongoing trial registries (apps.who.int/trialsearch/; clinicaltrials.gov); details of these searches are provided in the appendices (Appendix 2).

Searching other resources

We also checked the reference lists of any identified studies for additional potentially relevant studies. We contacted authors of the included study to find any ongoing or unpublished relevant studies.

Selection of studies

All three authors independently assessed the results of the searches for any potentially relevant studies. The authors then retrieved and independently assessed the full text of the single study for inclusion, as per the criteria listed above. The authors did not identify any further studies and therefore did not list any studies in any other sections (Excluded studies; Studies awaiting classification; Ongoing studies). There were no disagreements to resolve by discussion.

Data extraction and management

Two review authors (KRJ and AK) independently extracted as much data as possible using a standardized data collection form in accordance with Chapter 7 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a), which included the following data: source; eligibility; study methods; participants and settings; interventions and comparisons; outcomes; results; adverse outcomes; and miscellaneous details (e.g. funding source of the study, or potential conflicts of interest). We resolved any disagreement which arose by discussion.

The authors planned to assess all primary and secondary outcomes at 'up to three months', 'over three and up to six months’ and 'over six and up to 12 months' of treatment and to consider any other time‐points reported. There are no efficacy study data available, but there are some adverse event data available on the website clinicaltrials.gov that were reported after six months of treatment.

Assessment of risk of bias in included studies

Two review authors (KRJ and AK) independently assessed the risk of bias in the included study using the criteria described in Chapter 8 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011b). The authors assessed the risk of bias for the domains listed below.

1. Allocation sequence generation

2. Concealment of allocation

3. Blinding of participants and investigators

4. Incomplete outcome data

5. Selective outcome reporting

6. Any other potential risk of bias

The authors judged each of these criteria to have a low, high or unclear risk of bias and resolved any disagreement by discussion. The authors produced figures for a 'Risk of bias graph' and 'Risk of bias summary' using Review Manager 5.1 (RevMan 2014).

Measures of treatment effect

The authors planned to analyse dichotomous variables by calculating the risk ratio (RR). They planned to express continuous variables as a mean difference (MD). They will express continuous variables as standardized mean difference (SMD) if different scales are used for an outcome in different studies in future updates of review. The authors planned to analyse time‐to‐event data using the hazard ratio (HR). They intended to express the overall results with 95% confidence intervals (CI). Finally, the authors planned to report the numbers of ABPA exacerbations and hospital admissions as a rate ratio (ratio of events per person years) in experimental and control group (Deeks 2011).

Unit of analysis issues

The authors planned to include only randomized and quasi‐randomized controlled parallel trials in the review; the included trial was randomized. They planned to exclude both cross‐over and cluster‐randomized trials.

Dealing with missing data

One author (KRJ) contacted study investigators via email to request some missing outcome data; in response, the authors received a small amount of additional information other than that available on website and have added this to the review. The authors also considered taking the following steps to deal with any missing data in future.

1. Perform sensitivity analyses to assess the sensitivity of any assumptions, if made in future updates.

2. Address the potential impact of missing data on the findings of the review in the 'Discussion' section.

3. For missing standard deviations (SDs) of continuous outcome data, to calculate the SD from study statistics (e.g. CIs, standard errors, T values, P values or F values); if it was still not possible to calculate the SD, then they planned to impute it from other studies in the meta‐analysis as per details given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011a). If data are still not available, the authors will generate analyses making assumptions to create the best and worst case scenarios. In this case, the authors planned to undertake a sensitivity analysis (see below).

4. If the review authors include more studies in future, they plan to explore the impact of including studies with missing outcome data in the overall assessment of results by a sensitivity analysis.

Where possible, the authors planned to extract data to allow an intention‐to‐treat (ITT) analysis, which aims to include all participants randomized into a trial irrespective of what happened subsequently.

Assessment of heterogeneity

The current review included only one study. in future updates of review if additional studies become available, the authors will assess clinical and methodological heterogeneity before pooling data in a meta‐analysis. We will carry out an assessment for statistical heterogeneity initially visually and then using a Chi² test and also using the I² statistic. Using the Chi² test, a low P value (P < 0.1) or a large Chi² statistic relative to its degree of freedom will provide evidence of heterogeneity of intervention effects (i.e. the variation in effect estimates beyond chance). The authors will interpret the value of the I² statistic as follows:

• 0% to 40% ‐ might not be important;

• 40% to 60% ‐ may represent moderate heterogeneity;

• 60% to 75% ‐ may represent substantial heterogeneity; and

• 75% to 100% ‐ considerable heterogeneity (Deeks 2011).

Assessment of reporting biases

One author (KRJ) contacted study investigators through email to provide missing outcome data, but no additional missing data were available as the trial was stopped prematurely.

If sufficient numbers of included studies are available in future updates of review, the authors will assess publication bias, by using funnel plots in Review Manager (RevMan 2014).

Data synthesis

If a sufficient number of included studies are available in future updates of review, the authors will perform meta‐analyses using Review Manager (RevMan 2014). We plan to use a fixed‐effect model for pooled data analysis; however, if there is important statistical heterogeneity identified (I² statistic is more than 50%) we will also use a random‐effects model in a sensitivity analysis among studies.

Subgroup analysis and investigation of heterogeneity

If a sufficient number of included studies (at least 10) are available in future updates of review and if the authors identify substantial heterogeneity (over 50%), we will explore it by performing the following subgroup analyses:

1. children (up to and including 18 years of age) versus adults (over 18 years of age);

2. different stages of ABPA (five stages according to Patterson (Patterson 1982): (1) acute; (2) remission; (3) exacerbation; (4) corticosteroid‐dependent asthma; and (5) fibrosis (end stage));

3. colonization of participants with Staphylococcus aureus or Pseudomonas aeruginosa or both or neither.

Sensitivity analysis

Further, the authors will perform a sensitivity analysis to explore whether the heterogeneity is a result of different risk of bias. If sufficient numbers of trials are available (at least 10) they will undertake sensitivity analyses as follows in future updates of review:

1. repeating meta‐analysis after exclusion of studies with a high risk of bias due to concealment of allocation;

2. repeating meta‐analysis after exclusion of studies in which the outcome evaluation was not blinded;

3. repeating meta‐analysis after imputing missing data as best possible and worst possible outcome; and

4. comparing the difference of pooling analysis results by using a fixed‐effect model and a random‐effects model.