The n e w e ng l a n d j o u r na l of m e dic i n e

edi t or i a l s

Treating COPD — The TORCH Trial, P Values, and the Dodo

Klaus F. Rabe, M.D., Ph.D.

In the United States, the overall, age-standardized
death rate decreased from 1242 deaths per 100,000
population in 1970 to 845 deaths per 100,000 in
2002. This good news must be viewed against the
doubling during the same interval of the age-stan-
dardized death rate among persons with chronic
obstructive pulmonary disease (COPD),1 which
makes COPD a major cause of death (Fig. 1).
COPD is the diagnostic term for a group of
disorders that are characterized by respiratory
symptoms — dyspnea, cough, and sputum pro-
duction; airflow limitation; and chronic inflam-
mation of the lung.2 Risk factors include expo-
sure to a wide variety of inhaled particles and
gases, but in the Western world, inhaled cigarette
smoke is the most important known causative
factor.3 COPD is more than a pulmonary disorder
with a known effect on cardiovascular function
and on the risks of lung cancer, the metabolic syn­
drome, osteoporosis, cachexia, and depression.4
(Information for health care professionals and
patients with COPD is available on the Web site
of the National Heart, Lung, and Blood Institute,
at www.nhlbi.nih.gov/health/public/lung/copd.)
Patients with COPD die mainly from extrapul-
monary diseases, and COPD-related mortality is
probably underestimated because identifying the
precise cause of death is difficult in elderly pa-
tients with this disease, in whom cardiac arrhyth-
mias, ischemia and chronic pulmonary heart dis-
ease (cor pulmonale) or pulmonary embolism, or
both could be suspected. For example, 25% of
patients with COPD who were hospitalized for
severe exacerbations were shown to have pulmo-
nary embolisms.5
Treatment of COPD has been focused on im-
proving lung function and relieving symptoms
with the use of inhaled anticholinergic agents and
beta-adrenergic–receptor agonists. These drugs,
in principle, also reduce exacerbations, but there
is no convincing evidence so far to suggest that
this therapy substantially changes the course of
the disease or affects mortality.
Although the mechanism linking COPD and
coexisting diseases is uncertain, recent data sug-
gest that it might be chronic inflammation, in the
lung and systemically.6 Inhaled corticosteroids
have been reported to affect the frequency of ex-
acerbations,7 and, in retrospective analyses, to
reduce COPD-related mortality,8 although both
claims have been challenged in other statistical
analyses.9,10 Furthermore, database studies sug-
gest a beneficial effect of inhaled corticosteroids
on deaths from cardiovascular disease11 and lung
cancer.12
In this issue of the Journal, Calverley and col-
leagues report the results of the Towards a Revo-
lution in COPD Health (TORCH) trial, a large ran­
domized, prospective study involving patients with
COPD in which mortality was the primary end
point and the intervention consisted of therapy
with long-acting beta-agonists and inhaled corti-
costeroids, alone or in combination.13 Deaths from
any cause were analyzed with the use of a Cox
proportional-hazards model, and all efficacy
analyses were performed in the intention-to-
treat population — probably the most rigorous
approach that could have been taken. The study
was powered on the assumption of a mortality
rate in the placebo group equal to or greater than
17% and a 25% reduction in mortality in the
group receiving study treatments containing in-
haled corticosteroids, as compared with place-
bo.14 During the trial, the number of patients to
be enrolled had to be increased because the over-
all death rate was lower than expected.

n engl j med 356;8  www.nejm.org  february 22, 2007 851

Downloaded from www.nejm.org on June 11, 2010 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Causes of death in COPD

Cardiovascular
disease
27%
Pulmonary
disease
35%
Cancer
21%

Other
17%

Figure 1. Causes of Death in Patients with COPD.

Among patients with COPD, death can result from causes in a number of disease categories, in part, because of the strong association
between COPD and exposure to cigarette smoke. In the Towards a Revolution in COPD Health (TORCH) trial, 35% of deaths were adju-
COLOR FIGURE
dicated as due to pulmonary causes, 27% to cardiovascular disease, and 21% to cancer. Ten percent were attributed to other causes,
whereas the primary cause of death could not be determined 02/06/07
Draft 4 by the clinical end point committee in 7% of cases.
Author Rabe
Fig # 1
How are the findings
Titleof this trial to be inter- became intolerable (as might be expected in the
preted? The truly strong
ME part of the trial is the placebo group, for example) and left the study
meticulous recording DE of the causes of deaths. to obtain some relief. This weakness in the study
However, there are weaknesses
Artist in the study de- design probably could have been avoided if the
SBL
sign, namely the risk of skewed results
AUTHOR PLEASEdue NOTE:to primary comparison had been between patients
Figure has been redrawn and type has been reset
differential withdrawal from the Please assigned treat- receiving the combination therapy and those re-
check carefully
ment. This risk turnedIssue
intodatereality: 40% or more ceiving long-acting beta-agonists alone. The pla-
of the subjects enrolled in the study dropped out. cebo-controlled design could also have affected
Since all patients in the study clearly had an indi- enrollment: recruiting patients who would take
cation for therapy, clinically, it is conceivable that the chance of receiving placebo for 3 years might
some patients found that their COPD symptoms have excluded those with more severe disease —

852 n engl j med 356;8  www.nejm.org  february 22, 2007

 editorials
that is, patients known to have frequent exacer-
bations, which are associated with a higher mor-
tality rate.
These factors may have led to the failure of
the trial to demonstrate the expected mortality
in the placebo group. Setting total mortality as the
end point was an ambitious target, and it raises
the question of what the minimally clinically im-
portant difference for this end point would be.
One can only speculate that those designing the
study were (too) convinced of the effects of in-
haled corticosteroids on the basis of their own
retrospective data. These data might have per-
tained to a different group of patients, such as
those who may not have wanted either to enroll
or to stay in the trial. In the end, the trial failed
to meet its goal: the P value for death from any
cause was 0.052, which was higher than the pre-
specified value of 0.05. All clinical trials are a
gamble, and the TORCH investigators came close
to winning but did not win. Thus, the results of
this trial are difficult to interpret.
The real challenge, however, is what to make
of the trial for clinical practice. Lewis Carroll, in
Alice in Wonderland, offers a wonderful conclusion
that may be applicable here:
However, when they had been running half
an hour or so, and were quite dry again, the
Dodo suddenly called out “The race is over!”
and they all crowded round it, panting, and
asking “But who has won?” This question
the Dodo could not answer without a great
deal of thought, and it stood for a long time
with one finger pressed upon its forehead
.  .  .  while the rest waited in silence. At last
the Dodo said “Everyone has won, and all
must have prizes.”15
In the case of the TORCH trial, I am afraid
this conclusion is not an option. The interpreta-
tion of the data for the primary outcome as pub-
lished is conservative — and, in principle, it is
correct: the reduction in mortality in the com-
bined-therapy group did not reach the predeter-
mined level of statistical significance. On further
weighing these results, however, I think the treat-
ment with long-acting beta agonists was a win-
ner and that with inhaled corticosteroids was a
clear loser. The clinical guidance is obvious: mono-
therapy with corticosteroids should not be advo-
cated for patients with COPD, monotherapy with
n engl j med 356;8  www.nejm.org  february 22, 2007
Downloaded from www.nejm.org on June 11, 2010 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
a long-acting bronchodilator appears to be safe,
and the combination therapy offers no statisti-
cally significant additional survival benefit. In this
context, the results of the Understanding Poten-
tial Long-term Impacts on Function with Tiotro-
pium (UPLIFT) trial investigating the effect of an
anticholinergic drug over 4 years in patients with
COPD, which should be available in 2008, will be
of interest.16
Combination therapy, as compared with mono-
therapy with long-acting beta-agonists or inhaled
corticosteroids, offers statistically significant ad-
vantages for health status, frequency of exacer-
bations, use of oral steroids, and — probably
most important clinically — protection against a
decline in lung function. This finding confirms
the position of combination therapy in current
guidelines for the treatment of patients with COPD
that recommend its use for those with severe
COPD with frequent exacerbations2 but not for
patients with milder disease or without frequent
exacerbations. Caution in the use of combination
therapy is urged because of the finding in the
TORCH trial of an increased rate of pneumonia
among all patients receiving treatment contain-
ing inhaled corticosteroids. This finding urgently
requires further investigation, and I urge the
sponsor of this study to undertake a large trial
to determine its importance.
The TORCH trial is important, not only be-
cause it provides data on the natural history of
COPD but also because it clarifies, in part, the
role (and the shortcomings) of pharmacotherapy
for COPD in the overall mortality resulting from
this disease and highlights some of the inherent
problems of retrospective analyses of treatment
trials. Believe it or not, we still need more data,
from even larger trials.
Dr. Rabe is member of the Global Initiative for Obstructive
Lung Disease and chair of the science committee. He reports re-
ceiving consulting or lecture fees from GlaxoSmithKline, Astra­
Zeneca, Boehringer Ingelheim, Chiesi, Pfizer, Novartis, Altana
Pharma, Merck, and Almirall and grants from Altana Pharma,
AstraZeneca, Boehringer Ingelheim, and Pfizer. No other potential
conflict of interest relevant to this article was reported.
From the Leiden University Medical Center, Leiden, the Neth-
erlands.
1. Jemal A, Ward E, Hao Y, Thun M. Trends in the leading
causes of death in the United States, 1970-2002. JAMA 2005;294:
1255-9.
2. Global Initiative for Chronic Obstructive Lung Disease home
page. (Accessed February 1, 2007, at http://www.goldcopd.org.)
3. Pauwels RA, Rabe KF. Burden and clinical features of
chronic obstructive pulmonary disease (COPD). Lancet 2004;364:
613-20.
853
 The n e w e ng l a n d j o u r na l
4. Sin DD, Anthonisen NR, Soriano JB, Agusti AG. Mortality
in COPD: role of comorbidities. Eur Respir J 2006;28:1245-
57.
5. Tillie-Leblond I, Marquette C-H, Perez T, et al. Pulmonary
embolism in patients with unexplained exacerbation of chronic
obstructive pulmonary disease: prevalence and risk factors. Ann
Intern Med 2006;144:390-6.
6. Sevenoaks MJ, Stockley RA. Chronic obstructive pulmonary
disease, inflammation and co-morbidity — a common inflam-
matory phenotype? Respir Res 2006;7:70-8.
7. Calverley P, Pauwels R, Vestbo J, et al. Combined salmeterol
and fluticasone in the treatment of chronic obstructive pulmonary
disease: a randomised controlled trial. Lancet 2003;361:449-56.
[Lancet 2003;361:1660.]
8. Sin DD, Wu L, Anderson JA, et al. Inhaled corticosteroids
and mortality in chronic obstructive pulmonary disease. Thorax
2005;60:992-7.
9. Suissa S. Statistical treatment of exacerbations in therapeutic
trials of chronic obstructive pulmonary disease. Am J Respir Crit
Care Med 2006;173:842-6.
Synergistic Copathogens — HIV-1 and HSV-2
Lawrence Corey, M.D.
The variability in both the clinical progression
and transmission of human immunodeficiency
virus (HIV) infection has prompted a search for
cofactors influencing replication of the virus. Al-
though it is clear that host immune and genetic
factors, as well as the replication kinetics of par-
ticular viral strains, influence the progression of
HIV disease, a variety of exogenously acquired
infectious agents also appear to influence the
pace of HIV replication, the destruction of CD4+
T cells, and HIV transmission to infants and sex-
ual partners. Transient bursts of HIV replication
occur after vaccination and during episodes of
acute systemic infection. More persistent eleva-
tions in plasma HIV levels have been seen in pa-
tients with chronic infections (such as those with
Mycobacterium tuberculosis and herpes and hepatitis
viruses), and such coinfected patients have a more
rapid loss of CD4+ T cells and an increased rate
of progression to AIDS and death.1
HIV replication is compartmentalized in ana-
tomic sites of the body, and the interactions be-
tween HIV type 1 (HIV-1) and microbes occupy-
ing these anatomic sites influence the amount
and strain of HIV-1 in these regions. Interactions
between the gut flora with HIV in gut lymphoid
tissue and between sexually acquired pathogens
and HIV-1 in the genital tract are perhaps the
two areas of greatest importance in influencing
the progression of disease and viral transmission.
Localized infections of the genital tract with sex-
854 n engl j med 356;8  www.nejm.org  february 22, 2007
Downloaded from www.nejm.org on June 11, 2010 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
10. Idem. Inhaled steroids and mortality in COPD: bias from un-
accounted immortal time. Eur Respir J 2004;23:391-5.
11. Huiart L, Ernst P, Ranouil X, Suissa S. Low-dose inhaled cor-
ticosteroids and the risk of acute myocardial infarction in COPD.
Eur Respir J 2005;25:634-9.
12. Parimon T, Chien JW, Bryson CL, McDonell MB, Udris EM,
Au DH. Inhaled corticosteroids and risk of lung cancer among
patients with COPD. Am J Respir Crit Care Med (in press).
13. Calverley PMA, Anderson JA, Celli B, et al. Salmeterol and
fluticasone propionate and survival in chronic obstructive pulmo-
nary disease. N Engl J Med 2007;356:775-89.
14. Vestbo J, TORCH Study Group. The TORCH (TOwards a
Revolution in COPD Health) survival study protocol. Eur Respir J
2004;24:206-10.
15. Carroll L. Alice’s adventures in wonderland. London: Mac-
millan, 1865.
16. Decramer M, Celli B, Tashkin DP, et al. Clinical trial design
considerations in assessing long-term functional impacts of
tiotropium in COPD: the UPLIFT trial. COPD 2004;1:303-12.
Copyright © 2007 Massachusetts Medical Society.
ually acquired bacterial infections such as Neisse-
ria gonorrhoeae and, to a lesser extent, Chlamydia
trachomatis are associated with higher amounts
of HIV in genital secretions; treatment of these
infections with antimicrobial agents is associated
with a lowering of the HIV load in these secre-
tions.2 Thus, the identification and treatment of
such infections have been important parts of the
medical care of patients with HIV infection.
The clinical management of herpes simplex
virus type 2 (HSV-2) in patients with HIV infec-
tion has lagged seriously behind the large body
of medical literature on the importance of the
interaction between these two pathogens.3 Per-
sistent HSV-2 infection was one of the original
opportunistic infections that resulted in the iden-
tification of HIV. Since the initial reports in 1988
studying men who have sex with men, many ad-
ditional studies have shown the association be-
tween prevalent and incident HSV-2 infection and
the risk of HIV acquisition.4,5 In this issue of the
Journal, a study by Nagot et al.6 underlines the
association of HSV-2 with significantly higher
amounts of HIV-1 in plasma and in genital secre-
tions in women with sexually acquired HIV-1. This
finding has direct clinical implications, suggest-
ing that HIV-1 replication can be reduced with anti­
viral therapy directed solely at HSV-2, since acy-
clovir has no direct antiviral activity against HIV.7
The conceptual importance of this observation
is high. HSV-2 is acquired rapidly after the onset
 New England Journal of Medicine

CORRECTION

Treating COPD — The TORCH Trial, P Values, and the

Dodo

Treating COPD — The TORCH Trial, P Values, and the Dodo . The
fifth sentence of the eighth paragraph (page 853) should have read
``In the end, the trial failed to meet its goal: the P value for death from
any cause was 0.052, which was higher than the prespecified value
of 0.05,´´ rather than ``prespecified value of 0.50.´´ The text has been
corrected on the Journal’s Web site at www.nejm.org.

N Engl J Med 2007;356:1692

Downloaded from www.nejm.org on June 11, 2010 . Copyright © 2007 Massachusetts Medical Society. All rights reserved.