Teach 2015a.

Methods	Study design: 3‐arm, randomised, double‐blind, double placebo‐controlled, multicentre clinical trial. Aim: to compare (1) omalizumab with placebo and (2) omalizumab with an ICS boost with regard to autumn exacerbation rates when initiated 4 to 6 weeks before return to school. Study centres and method of recruitment: 8 US urban clinical research centres, no recruitment method information given. Dates of study: October 2011 to November 2013. Run‐in period: 2‐ to 12‐week screening. Duration of participation: from 4 to 6 weeks before school return until 90 days after school return. Consent: approved by all 8 institutional review boards. Consent from guardians and assent according to local guidelines. Power: enrolment of 453 participants (223 in the omalizumab arm, 155 in the inhaled corticosteroid boost arm, and 75 in the placebo arm (52 in steps 2 to 4 and 23 in step 5)) estimated to provide greater than 90% power to compare the omalizumab and placebo arms (11.8% vs 35.9% estimated effect) and 80% power to compare the omalizumab and ICS boost arms (12.9% vs 25.8% estimated effect). Imputation of missing data, i.e. assumptions made for ITT analysis: main analysis was based on modified ITT (children who were randomised, began study treatment, and had 1 or more study contact during the 90‐day outcome period were included in mITT). Supplemental volume included sensitivity analyses of mITT, PP, complete‐case, best‐case, worst‐case, and multiple imputation models.
Participants	Age (mean, range): 10.2 years, 6 to 17 years. Gender: 63.4% male. Asthma severity: National Heart, Lung and Blood Institute Expert Panel Report‐3 based steps 2‐5 (mild‐severe). Diagnostic criteria: asthma diagnosis or symptoms for more than 1 year. Number recruited: 727 Number randomised steps 2‐4 (omalizumab, placebo, steroid boost): 133, 47, 138 Number randomised treatment step 5 (omalizumab, placebo): 145, 50 Number completed treatment: 439 total. Efficacy Number analysed steps 2‐4 (omalizumab, placebo, steroid boost): 121, 43, 130 Number analysed treatment step 5 (omalizumab, placebo): 138, 46 Safety Number analysed steps 2‐4 (steroid boost, placebo): 131, 45 Number analysed treatment steps 2‐5 (omalizumab, placebo): 268, 93 Withdrawals: 585 excluded pre‐enrolment, 214 excluded pre‐randomisation, 59 withdrew consent and were excluded pre‐enrolment, 35 withdrew consent and were excluded pre‐randomisation. Steps 2‐4: 12 excluded from omalizumab group: 5 lost to follow‐up, 4 missed injection, 2 anaphylaxis, 1 exclusionary condition. 4 excluded from placebo group: 3 lost to follow‐up, 1 scheduling issue. 8 excluded from ICS boost group: 3 withdrew consent, 2 lost to follow‐up, 1 anaphylaxis, 1 missed injection, 1 scheduling issue. Step 5: 7 excluded from omalizumab group: 7 lost to follow‐up. 4 excluded from placebo group: 1 anaphylaxis, 1 lost to follow‐up, 1 missed injection, 1 withdrew consent. Inclusion criteria: age 6 to 17 years asthma diagnosis or symptoms for more than 1 year 1 or more asthma exacerbations (requiring systemic corticosteroids) or hospitalisation within the prior 19 months positive skin test response to 1 or more perennial allergens body weight and total serum IgE levels suitable for omalizumab school attendance beginning the following August or September residence in a low‐income census tract in predefined inner‐city areas and insurance covering standard medications (Note: children requiring 500 μg of fluticasone or equivalent twice daily for control during the run‐in phase (step 5) were not entered into the ICS boost arm and instead were randomised at a ratio of 3:1 to omalizumab or injected placebo.) Exclusion criteria: not reported distinct from inclusion criteria.
Interventions	Intervention: omalizumab standard dosing based on IgE and weight 4 to 6 weeks before, until 90 days after school start. Comparison: 1) placebo, or 2) ICS boost (doubled dose). Concomitant medication: ongoing guidelines‐based management EPR‐3. Excluded medication: none reported.
Outcomes	Primary outcome: asthma exacerbation in the 90‐day period beginning on the first day of each child’s school year, defined as worsening of asthma control requiring systemic corticosteroids or hospitalisation. Secondary outcome: 11 prespecified, non‐mechanistic secondary outcomes (analysed exacerbation during 90‐day intervention according to subgroups based upon: exacerbation during run‐in, eosinophil count, total IgE, roach IgE, age, fraction FeNO, FEV1, BMI, ethnicity, and gender). IFNα responses to rhinovirus were measured in PBMCs from a subset of participants. Time points measured: 2 to 4 weekly during intervention. Primary outcome result: asthma exacerbation in the 90‐day period beginning on the first day of each child’s school year: omalizumab vs placebo arm: 11.3% vs 21.0%; OR 0.48, 95% CI 0.25 to 0.92 omalizumab vs ICS boost arm: 8.4% vs 11.1%; OR 0.73, 95% CI 0.33 to 1.64 Secondary outcome results: exacerbation during 90‐day intervention according to subgroups. The following results differed significantly according to group: in those with an exacerbation during run‐in omalizumab vs placebo OR 0.12, 95% CI 0.02 to 0.64 (steps 2‐5), omalizumab vs ICS boost OR 0.05, 95% CI 0.002 to 0.98 (step 2‐4); in those with BMI centile ≥ 85 omalizumab vs ICS boost OR 0.13, 95% CI 0.03 to 0.61, (steps 2‐4); in those with BMI percentile < 85 ICS boost vs placebo OR 0.19, 95% CI 0.04 to 0.84 (steps 2‐4); in those with IgE < 255 kU/L omalizumab vs ICS boost OR 0.24, 95% CI 0.06 to 0.93 (steps 2‐4); in those with IgE 255 kU/L ICS boost vs placebo OR 0.24, 95% CI 0.06 to 0.87 (steps 2‐4); IFN‐α responses to rhinovirus were significantly increased in the omalizumab‐treated group (P = 0.03); among the omalizumab‐treated group, children with increases in ex vivo IFN‐α responses to rhinovirus to greater than the median value had a significantly lower rate of exacerbations during the outcome period OR 0.14, 95% CI 0.01 to 0.88. Adverse events: adverse events were reported by 54.5% of children in the omalizumab arm and 54.8% of children in the placebo arm (P > 0.99, steps 2–5) during the intervention phase. 1 or more adverse events were reported by 43.5% of children in the ICS boost arm and 53.3% of children in the placebo arm (P = 0.30, steps 2–4). 3 cases of grade 1 anaphylaxis occurred in the ICS boost, 2 in the placebo, and 3 in the omalizumab arm. Two serious AEs occurred during the intervention period, 1 each in the placebo (seventh nerve palsy) and ICS boost (anaphylaxis) arm. There were no deaths and no non–asthma‐related hospitalisations during the intervention phase.
Notes	Funding: National institute for Allergy and Immune Diseases and an unrestricted grant from Novartis. Omalizumab and matching placebo were donated by Novartis. The ICS boost and matching placebo were donated by GlaxoSmithKline. Both companies had the opportunity to comment on the study design, but they had no role in the trial’s performance, data analysis, manuscript preparation, or decision to submit the manuscript for publication. Adrenaline auto injectors were provided by Mylan. Subgroups: 11 subgroups were based on: exacerbation during run‐in, eosinophil count, total IgE, roach IgE, age, FeNO, FEV1, BMI, ethnicity, and gender. A prespecified subgroup analysis was conducted considering children with an exacerbation during the run‐in phase. Omalizumab was more efficacious than both placebo (6.4% vs 36.3%; OR 0.12, 95% CI 0.02 to 0.64) and ICS boost (2.0% vs 27.8%; OR 0.05, 95% CI 0.002 to 0.98).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Centralised, computer‐based random allocation scheme
Allocation concealment (selection bias)	Low risk	Described as centralised. No information on allocation concealment in report, but study authors confirmed that allocation was concealed using a third party and identical containers.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Placebo, inhalers and injections. No evidence that adverse events differed between placebo and interventions, and no other reasons to suspect participants could identify to which group they had been assigned. Participants and other staff blinded. Unblinded nurses administered injections but not involved in outcome measurement.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Mix of objective and subjective outcomes, but assessors all blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Primary analysis was modified intention‐to‐treat restricted to children who were randomised, began study treatment, and had more than or equal to 1 study contact during the 90‐day outcome period. There was good retention (94%) and similar dropout rates and reasons between groups.
Selective reporting (reporting bias)	Low risk	Secondary outcomes predefined. All reported in online supplement.
Other bias	Low risk	Groups balanced according to baseline characteristics.