Four of the randomized trials, identically designed, each involve 11,700 adult and adolescent patients with asthma per study, for a total of 46,800 patients, while a smaller trial involves 6,200 children.9 Each trial compares the combination of a LABA with an ICS (LABA+ICS) vs monotherapy with the same ICS at the same dose. The primary end point is a composite of serious asthma outcomes, including asthma-related death, intubation, or hospitalization, observed over a 6-month follow-up. These trials are designed as noninferiority trials, which test whether the patients in the LABA+ICS group are “no worse” than those in the ICS-only group with respect to the risk of the serious asthma outcomes.11,12 For the four adult/adolescent trials, “no worse” is defined in terms of the relative risk for serious asthma outcomes comparing LABA+ICS vs ICS being ≤ 2, which, in practical terms to account for random variability, implies that the upper bound of the 95% CI for the relative risk must be < 2 to conclude that LABA+ICS is no worse than ICS alone. The sample size of 11,700 patients per trial was computed by fixing the power at 90%. That is, the trial is designed to rule out a relative risk of 2 for the composite outcome with a probability of 90% when, in fact, there is no increase in the risk. Under the hypothesized rate of 15 serious asthma events per 1,000 patients per year, the upper limit of 2 for the CI corresponds to a point estimate of 1.32 for the relative risk, so that only if the estimated relative risk is < 1.32 can LABA+ICS be established as noninferior to ICS. The fifth trial, involving 6,200 children, was designed to have 90% power to rule out a relative risk of 2.7, that is, for the upper bound of the 95% CI for the relative risk to be < 2.7.