The aim of an equivalence trial is to show the therapeutic equivalence of two treatments, usually a new drug under development and an existing drug for the same disease used as a standard active comparator. Unfortunately the principles that govern the design, conduct, and analysis of equivalence trials are not as well understood as they should be. Consequently such trials often include too few patients or have intrinsic design biases which tend towards the conclusion of no difference. In addition the application of hypothesis testing in analysing and interpreting data from such trials sometimes compounds the drawing of inappropriate conclusions, and the inclusion and exclusion of patients from analysis may be poorly managed. The design of equivalence trials should mirror that of earlier successful trials of the active comparator as closely as possible. Patient losses and other deviations from the protocol should be minimised; analysis strategies to deal with unavoidable problems should not centre on an "intention to treat" analysis but should seek to show the similarity of results from a range of approaches. Analysis should be based on confidence intervals, and this also carries implications for the estimation of the required numbers of patients at the design stage.

Equivalence trials aim to show that two treatments have equivalent therapeutic effects. The approach is to define, in advance, a range of equivalence -d to +d for the treatment difference such that any value in the range is clinically unimportant. If the confidence interval for the difference, calculated after the trial, lies entirely within the interval, then equivalence is claimed. Glaxo Wellcome has carried out a series of trials using this methodology to assess new formulations of inhaled beta-agonists and inhaled steroids in asthma. Eleven of these trials are used to review some practical issues in equivalence trials. For the series of asthma trials, a range for peak expiratory flow rate (PEF) from -15 to +15 l/min was chosen to be the range of equivalence. This fitted well with physicians' opinions and with previously demonstrated differences between active and placebo. The choice of the size of the confidence interval should depend on the medical severity of the clinical endpoints under consideration and the level of risk acceptable in assuming equivalence if a difference of potential importance exists. From this point of view, a recommendation in the CPMP Note for Guidance on Biostatistics that 95 per cent confidence intervals should be used is inappropriate. Intent-to-treat (ITT) and per-protocol (PP) analyses were compared for the eleven asthma trials. Confidence intervals were always wider for the PP analysis and this was entirely due to the smaller number of subjects included in the PP analysis. There was no evidence that the ITT analyses were more conservative in their estimates of treatment difference. The need to demonstrate equivalence in both an ITT and a PP analysis in a regulatory trial increases the regulatory burden on drug developers. The relative importance of the two analyses will depend on the definitions used in particular therapeutic areas. Demonstrating equivalence in one population with strong support from the other would be preferred from the Industry viewpoint. In trials with regulatory importance, prior agreement with regulators on the role of ITT and PP populations should be sought. Trial designs will need to take account of the estimated size of the PP population if adequate power is needed for both analyses. Careful design in the series of asthma trials, particularly identifying a population of patients with potential to improve, resulted in notable increases in lung function during the course of the trials for both treatments. This provided reassurance that equivalence was not due to a lack of efficacy for both treatments. In one trial equivalence was demonstrated overall but a treatment by country interaction was noted. However, this interaction could not be attributed to differences in patient characteristics or baseline data between the countries. Study conduct was also similar in the different countries. The conclusion was that the interaction was spurious and that the trial provided good evidence of equivalence.

Salmeterol 50 micrograms bd is the appropriate dose for the treatment of children with mild to moderate asthma.

The use of three periods in the two-treatment crossover design for clinical trials is considered. It is proposed that a series of such trials in a particular therapeutic area may establish the relevance of the crossover design in that area. Treatment sequences to be used in three-period two-treatment trials are discussed. Two possible designs which allow carryover effects to be tested against the within-subject variability are compared. A design involving four treatment sequences is recommended.

We carried out a randomized double-blind controlled secondary-prevention trial of oxprenolol over seven years. Forty milligrams of oxprenolol or placebo was given twice daily to 1103 men 35 to 65 years old who had an acute myocardial infarction between 1 and 90 months previously. Overall, there was no difference in mortality or cardiac events between the placebo and oxprenolol groups. The major influence on prognosis was the time at which treatment was started after infarction. In 417 patients in whom treatment was started within four months of infarction oxprenolol increased the six-year cumulative survival rate from 77 to 95 per cent (P less than 0.001). In 274 patients with treatment starting between 5 and 12 months of infarction the survival rate was similar in the two groups, but in 412 patients entered between 1 and 7 1/2 years after their first infarction oxprenolol reduced the six-year survival rate from 92 to 79 per cent (P = 0.002). The increased mortality in this latter group mainly occurred late after withdrawal from active treatment. The value of low-dose oxprenolol in secondary prevention appears to be confined to patients treated relatively soon after myocardial infarction.