Withdrawing Glucocorticoids in COPD: A Paradigm Shift?
This is Dr Andrew Shorr, in Washington, DC, with a pulmonary and critical care literature update.
One of the controversial questions in our management of COPD is the role of inhaled corticosteroids. Based on results of the TORCH study, we use inhaled corticosteroids predominantly as a tool to prevent exacerbations in patients who have frequent exacerbations. That is the current role of steroids in COPD according to the GOLD guidelines. But we have never adequately studied how much benefit the steroids provide and how long the benefit lasts. What happens if you stop the inhaled corticosteroid in patients who have been taking them as part of a combination therapy regimen for some time? Does discontinuing the steroids lead to an increase in respiratory exacerbations?
This brings us to the WISDOM study, which was published online in the New England Journal of Medicine; the primary data were presented at the ERS International Congress, held in September 2014 in Munich, Germany.
This study randomly assigned 2500 patients who were on triple therapy to withdrawal or continuation of the inhaled corticosteroid. All patients in the study had been receiving triple therapy, with a long-acting muscarinic antagonist (LAMA), a long-acting beta-agonist (LABA), and an inhaled corticosteroid. Moreover, all patients had experienced at least one exacerbation in the past year and thus were at increased risk for exacerbations.
In a systematic fashion over several weeks, the inhaled corticosteroid was withdrawn in patients in one arm and continued in patients in the other arm. The primary endpoint was time to moderate or severe exacerbation; the trial was designed as a noninferiority study. In other words, the goal was to show that the rate of exacerbations is similar regardless of whether the inhaled corticosteroid was withdrawn or continued—that withdrawal of steroids is noninferior to continuation.
These patients were followed for 1 year. If the increase in risk of exacerbation did not reach the hazard ratio of 1.2 (a 20% increase in the odds of having an exacerbation), the investigators would then conclude that withdrawal of inhaled corticosteroid is not inferior to continuation.
When they looked at the primary data from the study, they found a hazard ratio of 1.06, when adjusted for several covariates and over the year time period. The upper limit of the 95% confidence interval (CI) for that risk went up to 1.19. Because the upper boundary of the 95% CI was below their prespecified 1.2 hazard ratio, the investigators concluded that discontinuation of inhaled corticosteroids did not significantly increase the risk for exacerbations. They also conducted a sensitivity analysis of only severe exacerbations, and that observation was confirmed.
Of interest, when they measured lung function, they saw that patients whose inhaled corticosteroids were withdrawn had about a 60-cc decline in FEV1 over the course of the study, and that difference was statistically significant. I believe many of us would argue, however, that a 60-cc fall is not necessarily clinically significant.
The quality-of-life data were somewhat confusing. In some situations with some quality-of-life measures, it seemed that withholding the inhaled corticosteroid was associated with a worsening of quality of life, whereas in other situations it was not.
Overall, I believe the study shows that withdrawing an inhaled corticosteroid does not necessarily expose patients to a great deal of risk for exacerbations, as long as they have been on the regimen for some period of time and are on other drugs as part of a combination therapy—in this case, a LAMA and a LABA.
My concern arises because, with that upper boundary of the 95% CI reaching 1.19, it comes right up to the 1.2 line that they prespecified. Most noninferiority studies choose margins that are much tighter. For example, in anti-infective trials, to conclude that anti-infective A is noninferior to anti-infective B, we generally use 10%-15% noninferiority margins. A 20% noninferiority margin, actually, is a bit broad. I have no questions that the study was well conducted, the data are internally consistent, and the signal predominantly goes in the same direction when you look at their subgroup analyses. However, I believe we have to ask ourselves whether we are willing to accept a noninferiority margin that is that broad.
What would be very helpful is detailed subgroup analyses to look at—perhaps the population of patients with frequent exacerbations, based on baseline criteria or frequent exacerbations post-steroid withdrawal, in terms of not only time to first event but the intensity of the frequency or time from one exacerbation to another.
Overall this study is valuable because it suggests that we should not put people on inhaled corticosteroids reflexively as part of their COPD regimen, and then not consider withdrawing them. In patients with asthma, we think about stepping down all the time. This is a paradigm shift for us in COPD, and that is why I believe it is an important study.
This is Andy Shorr, from Washington, DC.
This is Dr Andrew Shorr, in Washington, DC, with a pulmonary and critical care literature update.
One of the controversial questions in our management of COPD is the role of inhaled corticosteroids. Based on results of the TORCH study, we use inhaled corticosteroids predominantly as a tool to prevent exacerbations in patients who have frequent exacerbations. That is the current role of steroids in COPD according to the GOLD guidelines. But we have never adequately studied how much benefit the steroids provide and how long the benefit lasts. What happens if you stop the inhaled corticosteroid in patients who have been taking them as part of a combination therapy regimen for some time? Does discontinuing the steroids lead to an increase in respiratory exacerbations?
This brings us to the WISDOM study, which was published online in the New England Journal of Medicine; the primary data were presented at the ERS International Congress, held in September 2014 in Munich, Germany.
WISDOM Design
This study randomly assigned 2500 patients who were on triple therapy to withdrawal or continuation of the inhaled corticosteroid. All patients in the study had been receiving triple therapy, with a long-acting muscarinic antagonist (LAMA), a long-acting beta-agonist (LABA), and an inhaled corticosteroid. Moreover, all patients had experienced at least one exacerbation in the past year and thus were at increased risk for exacerbations.
In a systematic fashion over several weeks, the inhaled corticosteroid was withdrawn in patients in one arm and continued in patients in the other arm. The primary endpoint was time to moderate or severe exacerbation; the trial was designed as a noninferiority study. In other words, the goal was to show that the rate of exacerbations is similar regardless of whether the inhaled corticosteroid was withdrawn or continued—that withdrawal of steroids is noninferior to continuation.
These patients were followed for 1 year. If the increase in risk of exacerbation did not reach the hazard ratio of 1.2 (a 20% increase in the odds of having an exacerbation), the investigators would then conclude that withdrawal of inhaled corticosteroid is not inferior to continuation.
Study Results
When they looked at the primary data from the study, they found a hazard ratio of 1.06, when adjusted for several covariates and over the year time period. The upper limit of the 95% confidence interval (CI) for that risk went up to 1.19. Because the upper boundary of the 95% CI was below their prespecified 1.2 hazard ratio, the investigators concluded that discontinuation of inhaled corticosteroids did not significantly increase the risk for exacerbations. They also conducted a sensitivity analysis of only severe exacerbations, and that observation was confirmed.
Of interest, when they measured lung function, they saw that patients whose inhaled corticosteroids were withdrawn had about a 60-cc decline in FEV1 over the course of the study, and that difference was statistically significant. I believe many of us would argue, however, that a 60-cc fall is not necessarily clinically significant.
The quality-of-life data were somewhat confusing. In some situations with some quality-of-life measures, it seemed that withholding the inhaled corticosteroid was associated with a worsening of quality of life, whereas in other situations it was not.
Overall, I believe the study shows that withdrawing an inhaled corticosteroid does not necessarily expose patients to a great deal of risk for exacerbations, as long as they have been on the regimen for some period of time and are on other drugs as part of a combination therapy—in this case, a LAMA and a LABA.
Noninferiority Margin Too Broad?
My concern arises because, with that upper boundary of the 95% CI reaching 1.19, it comes right up to the 1.2 line that they prespecified. Most noninferiority studies choose margins that are much tighter. For example, in anti-infective trials, to conclude that anti-infective A is noninferior to anti-infective B, we generally use 10%-15% noninferiority margins. A 20% noninferiority margin, actually, is a bit broad. I have no questions that the study was well conducted, the data are internally consistent, and the signal predominantly goes in the same direction when you look at their subgroup analyses. However, I believe we have to ask ourselves whether we are willing to accept a noninferiority margin that is that broad.
What would be very helpful is detailed subgroup analyses to look at—perhaps the population of patients with frequent exacerbations, based on baseline criteria or frequent exacerbations post-steroid withdrawal, in terms of not only time to first event but the intensity of the frequency or time from one exacerbation to another.
Overall this study is valuable because it suggests that we should not put people on inhaled corticosteroids reflexively as part of their COPD regimen, and then not consider withdrawing them. In patients with asthma, we think about stepping down all the time. This is a paradigm shift for us in COPD, and that is why I believe it is an important study.
This is Andy Shorr, from Washington, DC.
Source...