Ciclesonide Noninferiority to Placebo on Growth Velocity: In Reply

Abstract

Listen

Translate article

In concordance with US Food and Drug Administration guidance,1 our study's main objective was to assess the effect of the inhaled corticosteroid ciclesonide on growth in children with asthma.2We agree that “the method of noninferiority trials is of delicate use and leads commonly to a false claim of noninferiority.” Naturally, a conclusion of noninferiority in a safety trial is valid only if patients complied with study medication use. Compliance was assessed in our study by 2 methods, one direct (diary cards, canister-weight changes) and one indirect (efficacy). The overall mean diary compliance was 97.4%. Highly concordant data were obtained by using canister-weight changes (94.9%). Thus, the possibility that parents of children enrolled in this trial overreported study-medication usage is highly unlikely.However, efficacy (forced expiratory volume in 1 second [FEV1]) was not observed. Although a positive signal would have provided additional support that patient adherence was good, the study was not designed to assess efficacy. On the basis of Food and Drug Administration guidance, patients with very mild asthma were recruited to minimize any confounding growth effect resulting from poor disease control or oral corticosteroid use. Thus, considering the high baseline FEV1 levels, an improvement in lung function would be difficult to detect.We also agree that both intention-to-treat (ITT) and per-protocol (PP) analyses are needed. These analyses were performed for nearly all parameters and demonstrated high concordance. For treatment withdrawals, the PP and ITT populations showed comparable rates (7.1% and 8.3%, respectively). Respective growth-velocity changes were also similar (mean ± SE: 5.76 ± 0.088 and 5.75 ± 0.080 cm/year for placebo, 5.69 ± 0.085 and 5.73 ± 0.080 cm/year for ciclesonide 40 μg/day, and 5.60 ± 0.090 and 5.60 ± 0.084 cm/year for ciclesonide 160 μg/day). In addition, with regard to concomitant treatments, >50% of the patients in each treatment group received montelukast in both populations. Thus, although it may have been helpful to include both the ITT and PP analyses in the article, only the primary PP analyses were displayed because of high concordance, a space limitation, and avoidance of confusion to readers.On the basis of the high concordance of the measures of compliance and the fact that the ITT and PP analyses were similar on the major end point, as well as possible confounding variables (eg, concomitant medications), we concluded that the relevant clinical outcome from this study, namely the linear regression estimate of the growth velocity of ciclesonide versus placebo over the year-long treatment period, was reported and demonstrated that ciclesonide at the 2 doses tested demonstrated no detectable effects on childhood growth velocity.