Sample sizes for comparative inhaled corticosteroid trials with emphasis on showing therapeutic equivalence

Abstract

In the near future it is to be expected that many new inhaled corticosteroids or formulations of these drugs will be compared with older ones, to discover whether they are therapeutically equivalent or not. The statistical evaluation of these trials differs from the classic methods. When two averages are similar or differ only slightly, power is very low. The regulatory bodies demand a power of at least 80%. This problem was initially solved by using the so-called power approach. Researchers included enough volunteers to enable them to detect a predefined difference, considered to be without any clinical significance, with a power of 80%. This approach, however, has been shown to be incorrect and has been replaced by the two one-sided tests procedure, where a new sample size equation is derived. Important elements of this new equation are the coefficient of variation of the parameter measured, the difference between the averages of the two groups and the equivalence limit (the difference between the means still tolerable). This equation was used in the present study to estimate the number of volunteers needed in a parallel inhaled corticosteroids equivalence trial. The end points chosen were the changes in FEV₁ and PC₂₀ due to the corticosteroid effect. Calculations were performed by extracting data from published placebo-controlled trials, and defining a range of equivalence limits and differences between the group averages. It was shown that a huge number of volunteers (500–1000) will be needed, as a result of the small corticosteroid effect and the high variance. In the case of inhaled corticosteroids, the equivalence limit is not known and needs defining to avoid discussions on the outcome. Due to the high number of patients who need to be included, the trial will most probably be multicentre and take place in several countries. Such a trial will suffer from several sources of bias. For instance, the definition of asthma can differ from country to country and from researcher to researcher, resulting in non-comparable groups of patients. The many sources of bias will make the outcome difficult to interpret. Therefore alternative methods to establish therapeutic equivalence are proposed and discussed.