© 2008 American Thoracic Society
The INSPIRE Study: Influence of Prior Use and Discontinuation of Inhaled CorticosteroidsTo the Editor:In their article, Dr. Wedzicha and colleagues reported that the INSPIRE study found no difference in exacerbation rates of chronic obstructive pulmonary disease (COPD) between salmeterol/fluticasone (SFC) and tiotropium (1). They also reported more study withdrawals from the tiotropium group than from the SFC group. We wondered if withdrawing from the study may be related to changes in treatment. In patients with COPD, it has been shown that discontinuation of inhaled corticosteroids (ICS) "leads to a clinically significant deterioration" (2). This finding is pertinent because ICS are included in the SFC arm, but not in the tiotropium arm. Therefore, prior use of ICS would be discontinued only for patients randomized to tiotropium. To evaluate the impact of prior use and discontinuation of ICS on withdrawal from the study, we used data from Wedzicha and coworkers' Tables 1 and 2 to compute study withdrawal rates according to use of ICS at study entry (see Table 1 herein). The data in Table 1 show the following: (1) patients who were using ICS at study entry withdrew from the study at a higher rate than patients who were not using ICS at study entry and, among such patients, the greatest withdrawal rate was for those who had their ICS discontinued because they were randomized to tiotropium (49%); (2) among tiotropium patients who were not withdrawn from ICS at study entry, the withdrawal rate (35%) was similar to the withdrawal rate in the SFC group (35%), and slightly greater than the withdrawal rate for SFC patients who did not use ICS at study entry (31%).
In conclusion, outcomes during the study may be confounded by study design–induced changes in treatment at the outset of the study. Specifically, study withdrawals appear to be related to prior use and especially discontinuation of ICS. The data do not support the authors' conclusion that "Withdrawing from the study after receiving tiotropium was unrelated to the subjects' prior use of ICS" (1). Moreover, because ICS discontinuation would be expected to increase exacerbation rates as well as withdrawals, we encourage the authors to extend our table to include exacerbation rates and mortality rates according to use of ICS at study entry. Others have suggested that bias introduced by discontinuation of ICS for comparator patients may at least in part explain apparent benefits of ICS on mortality (3).
Boehringer Ingelheim Pharmaceuticals, Inc. FOOTNOTES Conflict of Interest Statement: S.F.L. is an employee of Boehringer Ingelheim. M.J. is an employee of Boehringer Ingelheim. REFERENCES
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
