From the authors:
We would like to thank S. Suissa for his letter, which appeared in a recent issue of the European Respiratory Journal 1. We share his interest in the methodology of chronic obstructive pulmonary disease (COPD) trials. The issues he raised are important and deserve a more comprehensive response than is possible here. We would, however, like to address the two major points regarding the design of the TRISTAN (Trial of Inhaled Steroids and Long-acting β2-agonists) and ISOLDE (Inhaled Steroids in Obstructive Lund Disease) trials, which lead him to claim that the trials “violate fundamental principles of randomised trial methodology” 1.
First, S. Suissa contends that intent-to-treat analysis is impossible when there is incomplete follow-up of patients who withdraw from the study 1; but this assertion is incorrect. Intent-to-treat analysis does require inclusion of all available subjects in the analysis but the principle allows for missing data. The CONSORT (Consolidated Standards of Reporting Trials) statement is the standard guideline for reporting randomised clinical trials adopted by major medical journals 2. An accompanying article to the 2001 revision of CONSORT states: “It is common for some patients not to complete a study – they may drop out or be withdrawn from active treatment – and thus are not assessed at the end. Although these patients cannot be included in the analysis, it is customary still to refer to analysis of all available participants as an intent-to-treat analysis” 3. In the analysis of TRISTAN and ISOLDE, all available patients were included and therefore the results presented are from a valid intent-to-treat analysis.
Secondly, S. Suissa states that the design of COPD trials needs to be stratified by prior use of inhaled corticosteroids (ICS) and suggests that this deficiency “cannot simply be corrected by data analysis” 1. Stratifying a design by important predictors of outcome can be helpful in terms of ensuring balance, but it is not essential for data analysis. In large trials, randomisation will lead to similar proportions of patients in each treatment arm with prior use of ICS. Indeed, the standard textbook on clinical trials, Clinical Trials: a Practical Approach, states: “if the trial is very large, say several hundred patients […] then stratification has little point” 4.
Thus, the designs of ISOLDE and TRISTAN conform to conventional clinical trial methodology and therefore have no major flaws. It remains valid to draw conclusions regarding the effectiveness of inhaled corticosteroids based on their outcome.
Statement of interest
Statements of interest for all authors of this manuscript can be found at www.erj.ersjournals.com/misc/statements.shtml
- © ERS Journals Ltd
