the Lung Health Study, 1116 patients with COPD of
varying severities received triamcinolone 1200 µg
daily or placebo.75 Again, there was no difference
in the rate of FEV1 decline between the intervention
and placebo groups (44.2 ± 2. 9 versus 47.0 ± 3.0
mL/yr, P = 0.50). However, patients who received
triamcinolone had less respiratory symptomatology, less methacholine reactivity, and decreased
physician visits due to respiratory illness. Finally,
a mega-trial of 6112 patients (TORCH trial) with
COPD who had FEV1 below 60% included groups
that received fluticasone 1000 µg daily (n = 1534)
as well as placebo, salmeterol, and salmeterol-fluticasone combination.68 In the TORCH trial,
the primary endpoint was a reduction in mortality,
which did not materialize for any of the intervention groups. Fluticasone did reduce moderate or
severe exacerbation rates (rate ratio, 0.82; 95% CI,
0.76 to 0.89; P < 0.001) and improve SGRQ scores
(– 2.0; 95% CI, – 2. 9 to – 1.0; P < 0.001) compared
with placebo. In a post hoc analysis of the TORCH
trial, fluticasone was shown to reduce FEV1 decline
compared to placebo (42 mL/yr versus 55 mL/yr)
between 6 months and 3 years after randomiza-
tion.78 Importantly, a similar magnitude of reduction
in the rate of FEV1 decline was also shown for sal-
meterol and salmeterol-fluticasone combination.
In a recent Cochrane review of studies that involved ICS with placebo control, 47 trials (including
the aforementioned 5) with 13,139 patients were
evaluated in a meta-analysis.79 The study concluded that ICS had no statistically significant effect on
mortality and rate of FEV1 decline. However, there
was a favorable effect on exacerbations (weighted
mean difference, –0.26 exacerbations per patient
per year; 95% CI, 0.37 to –0.14) and the rate of
decline in quality of life as measured by SGRQ
scores (weighted mean difference, – 1. 22 units/
year; 95% CI, – 1.83 to –0.60).
More recently, in a meta-analysis of 7 randomized
trials (5997 patients) that compared the use of ICS
versus LABA in the treatment of stable COPD, no
difference was found in the number of patients experiencing exacerbations (OR, 1. 22; 95% CI, 0.89
to 1.67) or the rate of exacerbations per patient-
year (rate ratio, 0.96; 95% CI, 0.89 to 1.02).80 There
was a trend toward increased mortality (OR 1. 17;
95% CI, 0.97 to 1.42) for patients on ICS compared
with those on LABA. The incidence of pneumonia
was significantly higher for patients who received
ICS (OR, 1.38; 95% CI, 1. 10 to 1.73). There was a
small difference in SGRQ scores favoring ICS, but
the magnitude was not clinically significant (mean
difference, –0.74; 95% CI, – 1. 4 to –0.06).
The prescription of ICS for management of chronic stable COPD remains a controversial issue on
various accounts. A particular concern with the
presumed potential benefits of ICS relates to the
difficulty in reconciling the minor impact (if any) of
ICS on the rate of decline of lung function with its
reported benefits on the rate of COPD exacerbations. Critical review of those results by Suissa et
al69 uncovers several methodological concerns that
may have skewed the results towards a favorable
effect of ICS. These include:
1. An unweighted method of exacerbation assessment based on an average of individual COPD
exacerbation rates, with each COPD patient
contributing equally to the group mean of exacerbation rate.81
2. Failure to use an intent-to-treat principle such that
patients were followed only until discontinuation
of the study drug instead of the planned follow-up. Because of a generally higher withdrawal
rate in patients assigned to placebo, this flaw
would have the effect of magnifying the beneficial effect of ICS on mortality and rate of exacerbations. Specifically, in an analysis of several
