Per-protocol analyses: Prevalence and effect estimates of primary and secondary outcomes at age 6 years follow-up by allocation group.

*Adjusted for randomisation stratum, and baseline prevalence of outcomes. Care as usual is the reference group.aData are numerator/denominator (%).bMean ± standard deviation.cMedian (range).dNo baseline measurement available.eApplying Bonferroni correction: we performed 20 comparisons. At p = 0.0025 (i.e. 0.05/20), the decreased risk on ETS exposure at home ever in the intervention group remained statistically significant.Numbers of children does not equal the sum of the denominators in each subgroup because only those with baseline and follow-up data are included. Measurements on FeNO and Rint were available for respectively 3497 (45%) and 3828 (49%) of the participating children. FeNO = Fractional exhaled Nitric Oxide, Rint = airway resistance, ETS = Environmental Tobacco Smoke.