ARTICLE IN PRESS

Pulmonary Pharmacology & Therapeutics 21 (2008) 234–238

www.elsevier.com/locate/ypupt

TNF-a antagonists and the prevention of hospitalisation for

chronic obstructive pulmonary disease$
Samy Suissaa,b,,1, Pierre Ernsta,c, Marie Hudsond
Division of Clinical Epidemiology, McGill University Health Centre2, Canada
a
b
Departments of Epidemiology and Biostatistics and of Medicine, McGill University, Montreal, Canada
c
Division of Respiratory Medicine, McGill University Health Centre, Canada
d
Division of Rheumatology, McGill University, Canada

Received 27 December 2006; received in revised form 22 February 2007; accepted 24 March 2007

Abstract

Background: TNF-a may be important in the pathogenesis of COPD. Consequently, the use of TNF-a antagonists has been advocated
for its treatment.
Methods: We conducted an observational study to evaluate the effectiveness of TNF-a antagonists in preventing COPD
hospitalisations in a cohort of patients diagnosed with both RA and COPD identified from a health claims database. A nested case-
control approach was used to match each case hospitalised to 10 controls on age and cohort entry date. Data on prescribed medications
during the year prior to the index date were obtained. Rate ratios (RR) of COPD hospitalisation were estimated by conditional logistic
regression, after adjustment for COPD severity and concomitant RA medication use.
Results: The cohort included 15,771 subjects with both RA and COPD, of which 1205 were hospitalised for COPD during follow-up.
The adjusted RR of COPD hospitalisation associated with the use of TNF-a antagonists was 0.62 (95% confidence interval (CI)
0.43–0.89). This rate reduction was due to etanercept (RR 0.49, 95% CI 0.29–0.82) but not infliximab (RR 0.95, 95% CI 0.59–1.52).
Conclusion: Our finding of a halving in the rate of COPD hospitalisation associated with the use of etanercept corroborates the
potential importance of TNF-a in the pathogenesis of COPD. This study supports the initiation of randomised controlled trials of this
TNF-a antagonist among COPD patients at high risk of severe exacerbations.
r 2007 Elsevier Ltd. All rights reserved.

Keywords: COPD; Drug effectiveness; Etanercept; Infliximab; Observational studies; Rheumatoid arthritis; Tumour necrosis factor a

1. Introduction for many years to come [1,2]. Treatments available to date

Chronic obstructive pulmonary disease (COPD) is a
major health problem whose impact is projected to grow
$
The study was funded by a grant from the Canadian Institutes of
Health Research (CIHR) and the database was provided thanks to a grant
from Sanofi-Aventis.
Corresponding author. Division of Clinical Epidemiology, McGill
University Health Centre, Canada. Tel.: +514 843 1564;
fax: +514 843 1493.
E-mail address: samy.suissa@clinepi.mcgill.ca (S. Suissa).
1
Recipient of a Distinguished Investigator award from the CIHR.
2
The McGill Pharmacoepidemiology Research Unit is funded by an
infrastructure grant from the Fonds de la recherché en santé du Québec
(FRSQ).
mainly target symptoms and appear to have little impact
on the natural history of the disease [3,4]. Tumor necrosis
factor a (TNF-a) is a multi-functional inflammatory
cytokine which is found in excess in the sputum of patients
with COPD [5] and appears to drive much of the
neutrophilic airway inflammation observed in COPD [6].
In addition, gene promoter polymorphisms for TNF-a are
associated with an increase in the incidence of COPD [7] as
well as a worsened prognosis [8]. It seems therefore
plausible that agents that block the effect of TNF-a might
provide benefits to patients with COPD. This inflammatory
target has been deemed promising among the novel
pharmacological agents to investigate in the treatment of
COPD, particularly in patients with systemic symptoms [9].

1094-5539/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.pupt.2007.03.003

Descargado para Pedro Argos (pedrojose.argos@scsalud.es) en Marques de Valdecilla Foundation de ClinicalKey.es por Elsevier en agosto 31, 2023.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2023. Elsevier Inc. Todos los derechos reservados.
 ARTICLE IN PRESS
S. Suissa et al. / Pulmonary Pharmacology & Therapeutics 21 (2008) 234–238
Despite the general pathophysiological evidence, rando-
mised controlled trials to evaluate the effectiveness of
TNF-a antagonists in COPD have been few [10,11]. We
therefore carried out an observational study to indirectly
examine this hypothesis. We took advantage of the
increasing use of TNF-a antagonists in the treatment of
rheumatoid arthritis (RA) to evaluate the effectiveness of
TNF-a antagonists in reducing the risk of COPD
hospitalisation in a large cohort of patients with both
RA and COPD, some of whom were receiving anti-TNF-a
agents for their RA.
2. Methods
2.1. Data sources
The cohort for this study was formed using the
PharMetrics patient-centric outcomes database, a North
American insurance claims database [12]. This database
consists of standardized information on claims data from
75 different health insurance plans and currently encom-
passes more than 55 million unique patients. It includes
data on all physician visits, hospitalisations and prescribed
medications dispensed in these populations. Data spanning
the period from January 1995 to December 2003 were used
for this study. This database does not permit access to the
patients’ medical records in order to protect confidentiality.
This database has been used previously to assess the
hepatic and cardiovascular effects of disease-modifying
anti-rheumatic drugs (DMARDs) [13,14].
2.2. Cohort definition
We first identified the cohort of patients with RA from
all subjects with a diagnosis of RA (ICD-9 code 714), based
on data from physician visits between January 1, 1995 and
December 31, 2004. We then formed the sub-cohort of
patients who also had COPD by identifying from the RA
cohort all subjects with at least three physician visits with a
diagnosis of COPD within a one-year period. Cohort entry
was defined by the date of the third of these COPD visits or
January 1, 1999, if the date of the third COPD visit
occurred before this date. We chose January 1999 as the
start date because the drugs of interest in this study,
namely TNF-a antagonists, only became widely available
on the market after that time. All subjects were followed
from the date of cohort entry until the earliest of the date
of termination of enrolment in the health plan, the date of
death, the end of the study period (December 31, 2004) or
the date of the study outcome.
2.3. Outcome event
The outcome was the first occurrence of a hospitalisation
for COPD during follow-up. It was identified from all
inpatient encounters with a primary discharge diagnosis of
COPD (ICD-9 codes 490, 491, 492, 496).
Descargado para Pedro Argos (pedrojose.argos@scsalud.es) en Marques de Valdecilla Foundation de ClinicalKey.es por Elsevier en agosto 31, 2023.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2023. Elsevier Inc. Todos los derechos reservados.
235
2.4. Nested case-control design
Because of the complexity and time-varying nature of
drug exposures in this cohort, we used a nested case-
control approach within the cohort, which increases the
ease of analysis with inconsequential loss of power [15].
For each case of COPD hospitalisation identified in the
cohort, we randomly selected 10 controls, after matching
on age (within two years), month and year of cohort entry,
and ensuring that each control was still at risk for a
hospitalisation for COPD on the day the case occurred.
The date of hospitalisation for COPD was designated the
index date for each matched case-control set.
2.5. Exposure measurement
For all subjects, all drugs received during the year prior
to the index date were identified from dispensed prescrip-
tion data. The TNF-a antagonists available at the time
(etanercept and infliximab), other DMARDs, namely
anakinra, leflunomide, methotrexate, and the traditional
DMARDs (hydroxychloroquine, chloroquine, sulfasala-
zine, gold, azathioprine, minocycline, penicillamine, chlor-
ambucil, cyclophosphamide and cyclosporine) as well as
nonsteroidal anti-inflammatory drugs (NSAIDs) and
cyclooxygenase-2 (COX-2) inhibitors were identified.
Drugs commonly used for COPD, namely inhaled bronch-
odilators, as well as inhaled and oral corticosteroids, were
also identified.
2.6. Covariate information
The severity of COPD was assessed by the frequency of
use of short acting b-agonists, long acting b-agonists,
ipratropium bromide, inhaled and oral corticosteroids,
during the year prior to the index date. The number of
COPD hospitalisations during the year prior to cohort
entry was also used as marker of COPD disease severity.
2.7. Data analysis
Conditional logistic regression was used with the nested
case-control sample to estimate the rate ratio (RR) of
hospitalisation for COPD for each of the two TNF-a
antagonists, namely etanercept and infliximab, after
adjusting for the effects of covariates. Exposure to these
medications was defined as a prescription dispensed during
the year prior to the index date. The RRs were adjusted for
the potential confounding effects of sex, the severity of
COPD, as measured by the frequency of use of COPD
medications and the number of hospitalisations during the
year prior to the index date. We also adjusted for the
effects of the concurrent use of other medications used to
treat RA, including other DMARDs, NSAIDS and
COX-2 inhibitors. All analyses were carried out using
SAS (Cary N.C.).
 ARTICLE IN PRESS
236 S. Suissa et al. / Pulmonary Pharmacology & Therapeutics 21 (2008) 234–238

3. Results the cases and matched controls, respectively. The cases had
appreciably more use of medications for COPD but were
The cohort of RA patients included 235,272 subjects, quite similar in their use of traditional anti-rheumatic
from which we identified the study cohort of all 15,771 medications during the year prior to cohort entry.
COPD patients with at least 3 visits for COPD. During The adjusted RR of a hospitalisation for COPD with the
follow-up, 1205 subjects were hospitalised for COPD and use of TNF-a antagonists in the year prior to the index date
these subjects were matched to 12,050 controls. was 0.62 (95% confidence interval (CI) 0.43–0.89) relative
Table 1 shows that the cases hospitalised for COPD and to no use (Table 2). This effect was due exclusively to etan-
their matched controls were 67 years of age at the index ercept (RR 0.49, 95% CI 0.29–0.82) and not to infliximab
date and that the cases occurred on average 14 months (RR 0.95, 95% CI 0.59–1.52). Moreover, anakinra was not
after cohort entry. Women represented 69% and 67% of associated with a lower rate of hospitalisation for COPD.
We performed two sensitivity analyses. First, since
infliximab was approved only in combination with
Table 1
methotrexate in RA whereas etanercept was often used
Comparison of cases hospitalised for COPD and matched controls
alone, especially in earlier years, we estimated these effects
Cases Controls in the subset of patients who were not on methotrexate in
the year prior to the index date (88% of cases and 86% of
Number 1205 12,050
Age in years (mean7SD)a 67712 67712
controls). In this group, the RR of a COPD hospitalisation
Follow-up in months (mean7SD)a 14712 14712 with the use of etanercept was 0.30 (95% CI 0.14–0.65) and
Sex (% women) 69 67 with infliximab 1.03 (95% CI 0.56–1.93). Second, to
Hospitalisations for COPD in the year prior to cohort entry
address the uncertain validity of the RA diagnosis used
At least one (%) 3.9 2.2 to define the cohort, we restricted the analysis to the
Number (mean7SD) 0.0670.40 0.0370.23 homogeneous subgroup of RA patients who had already
Prescriptions for COPD medications in the year prior to index date (mean
been using DMARDs prior to cohort entry. In this group,
number7SD) the RR of a COPD hospitalisation with the use of
Short-acting beta-agonists (SABA) 4.879.7 2.276.8 etanercept was 0.47 (95% CI 0.26–0.85) and with
Long-acting beta-agonists (LABA) 0.772.7 0.572.5 infliximab 1.14 (95% CI 0.66–1.99).
Theophylline 0.973.7 0.472.6
Ipratropium bromide 0.773.2 0.271.8
Inhaled corticosteroids (ICS) 1.073.6 0.672.4 4. Discussion
Combination of ICS+LABA 4.879.7 2.276.8
Oral corticosteroids 1.772.9 1.172.6
We found using a large-scale observational approach
Use of other RA medications (%) that TNF-a inhibitors, agents directed against the deleter-
Methotrexate 12 14
Leflunomide 5 4
ious effects of the inflammatory cytokine TNF-a, were
Traditional DMARDSb 9 10 associated with a reduction in the rate of COPD
NSAIDS 21 23 hospitalisation among patients with COPD receiving these
COX-2 inhibitors 19 20 agents to treat their rheumatoid arthritis. This effect,
a however, was due exclusively to a reduction of 50% in the
Matching variables.
b
The traditional DMARDS include hydroxychloroquine, chloroquine, rate of COPD hospitalisation with etanercept, a soluble
sulfasalazine, gold, azathioprine, minocycline, penicillamine, chlorambu- receptor fusion protein that binds TNF. The other TNF-a
cil, cyclophosphamide and cyclosporine. inhibitor under study, namely infliximab, a humanised

Table 2
Crude and adjusted rate ratios of hospitalisation for COPD associated with use of biologic DMARDs including TNF-a antagonists

Cases Controls Crude rate Adjusteda

ratio
Rate ratio 95% CI

Number of subjects 1205 12,050

Use in the year prior to the index date
No use (reference) 1.00 1.00 Reference
Any TNF-a antagonist 34 523 0.64 0.62 0.43–0.89
Etanercept 16 310 0.51 0.49 0.29–0.82
Infliximab 21 224 0.94 0.95 0.59–1.52
Anakinra 57 481 1.20 1.15 0.86–1.55
a
Adjusted for one another as well as for age (by matching), sex, use of methotrexate, leflunomide, traditional DMARDs and COPD medications all in
the year prior to index date, and hospitalisation for COPD prior to cohort entry.

Descargado para Pedro Argos (pedrojose.argos@scsalud.es) en Marques de Valdecilla Foundation de ClinicalKey.es por Elsevier en agosto 31, 2023.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2023. Elsevier Inc. Todos los derechos reservados.
 ARTICLE IN PRESS
S. Suissa et al. / Pulmonary Pharmacology & Therapeutics 21 (2008) 234–238
monoclonal antibody also directed against TNF-a, did not
reduce the risk of COPD hospitalisation.
While the importance of TNF-a as an inflammatory
cytokine in COPD is increasingly recognised, data on the
efficacy of therapy with newer agents that block the effects
of TNF-a in this disease are limited. Van der Vaart and
colleagues reported on a randomised control trial of
infliximab in 22 patients with COPD who were also active
smokers and found no benefit on respiratory symptoms,
quality of life or lung function parameters [10]. More
recently Rennard and colleagues confirmed the lack of
efficacy of infliximab on several outcome measures includ-
ing exacerbations in more than two hundred patients with
COPD in a randomised controlled trial [11]. These findings
are consistent with our results for this agent. We could find
no trials of etanercept for COPD, but there are two recent
reports of the efficacy of this agent in severe asthma [16,17].
Severe asthma is often characterised by neutrophilic
inflammation not dissimilar from that found in COPD
[18] and therefore these studies might cautiously be
interpreted as also suggesting benefit for patients with
COPD. Other potential explanations for differences in
efficacy of infliximab and etanercept include the more
frequent occurrence of blocking antibodies with infliximab
[19] and differences between these agents in their pharma-
cological profile [20].
In this study, because TNF-a inhibitors have been used
in the treatment of rheumatoid arthritis since the late
1990s, we were able to use an indirect approach to evaluate
the potential effectiveness of these drugs in COPD. We
took advantage of a large cohort of patients with
rheumatoid arthritis that included over 15,000 patients
who also had COPD, some of whom were serendipitously
treated with TNF-a inhibitors for their RA, to assess the
effectiveness of these DMARDs on the patients’ COPD
condition.
This study is, nonetheless, observational in nature and
therefore subject to potential biases. The major concern
with such observational studies of drug effectiveness is
confounding by indication, which would be expected to
occur if patients with more severe COPD, and thus at
greater risk of being hospitalised for COPD, would be
preferentially prescribed anti-TNF-a agents. This is un-
likely to be the case in the present study since these
medications were being prescribed for RA, likely indepen-
dently of the severity of COPD. Moreover, no reduction in
the risk of COPD hospitalisation was seen with anakinra,
another new biologic agent approved for the treatment of
RA but with an altogether different mechanism of action,
namely interleukin-1 receptor antagonism. This suggests
that being prescribed newer biologic agents or factors
associated with this level of care are not likely to be linked
to COPD severity and the risk of a COPD hospitalisation.
Such differences in care are therefore unlikely to explain
the protective effect of etanercept against COPD hospita-
lisation. Confounding by contraindication could also have
occurred if COPD patients with more severe RA, who are
Descargado para Pedro Argos (pedrojose.argos@scsalud.es) en Marques de Valdecilla Foundation de ClinicalKey.es por Elsevier en agosto 31, 2023.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2023. Elsevier Inc. Todos los derechos reservados.
237
more likely to receive TNF-a antagonists, are less likely to
be labelled as COPD when they are hospitalised. Here
again, such a bias would have also led to lower RRs for
infliximab and anakinra, which was not the case. While we
selected patients who had been labelled as having COPD
on several health care contacts and subjects were on
average 67 years of age, we cannot be sure that we have not
included patients with asthma rather than COPD. If this
were so we may have exaggerated the benefit of etanercept
in COPD since this medication has been shown to be
effective in severe asthma [16,17].
TNF-a antagonists have transformed the treatment of
rheumatoid arthritis [21,22]. They have been shown not
only to improve signs and symptoms of the disease but to
dramatically change the course of the disease by slowing
radiographic damage and improving function in patients
with RA. In contrast, there has only been modest progress
in the development of drug therapy in COPD. It would
thus be important and urgent to assess whether these TNF-
a antagonists, or even other drugs, might also have similar
disease-modifying properties in COPD [9].
COPD hospitalisation is a marker of a severe exacerba-
tion of COPD and such exacerbations are associated with
significant morbidity and costs [23,24]. Our results suggest
that a treatment strategy aimed at reducing the adverse
effects of the pro-inflammatory cytokine TNF-a is effective
at preventing such exacerbations. This study thus supports
the initiation of randomised controlled trials of TNF-a
antagonists, particularly etanercept, among patients with
COPD at high risk of severe exacerbations.
Acknowledgements
The sponsors did not participate in the design, conduct,
analysis or interpretation of the data; and preparation of
the manuscript. The authors had full access to all of the
data in the study and the first author takes responsibility
for the integrity of the data and the accuracy of the data
analysis. We thank Dr Juhaeri Juhaeri of Sanofi-Aventis
for his assistance with the database acquisition.
References
[1] Lopez AD, Murray CC. The global burden of disease, 1990–2020.
Nat Med 1998;4(11):1241–3.
[2] Gerardi DA, Lovett L, Benoit-Connors ML, Reardon JZ, Zuwallack
RL. Variables related to increased mortality following out-patient
pulmonary rehabilitation. Eur Respir J 1996;9(3):431–5.
[3] Burge PS, Calverley PM, Jones PW, Spencer S, Anderson JA, Maslen
TK. Randomised, double blind, placebo controlled study of
fluticasone propionate in patients with moderate to severe chronic
obstructive pulmonary disease: the ISOLDE trial. Br Med J 2000;
320(7245):1297–303.
[4] The lung health study research group. Effect of inhaled triamcinolone
on the decline in pulmonary function in chronic obstructive
pulmonary disease. N Engl J Med 2000;343(26):1902–9.
[5] Keatings VM, Collins PD, Scott DM, Barnes PJ. Differences in
interleukin-8 and tumor necrosis factor-alpha in induced sputum
 ARTICLE IN PRESS
238 S. Suissa et al. / Pulmonary Pharmacology & Therapeutics 21 (2008) 234–238
from patients with chronic obstructive pulmonary disease or asthma.
Am J Respir Crit Care Med 1996;153(2):530–4.
[6] Churg A, Wang RD, Tai H, Wang X, Xie C, Wright JL. Tumor
necrosis factor-alpha drives 70% of cigarette smoke-induced emphy-
sema in the mouse. Am J Respir Crit Care Med 2004;170(5):492–8.
[7] Sakao S, Tatsumi K, Igari H, Shino Y, Shirasawa H, Kuriyama T.
Association of tumor necrosis factor alpha gene promoter poly-
morphism with the presence of chronic obstructive pulmonary
disease. Am J Respir Crit Care Med 2001;163(2):420–2.
[8] Keatings VM, Cave SJ, Henry MJ, et al. A polymorphism in the
tumor necrosis factor-alpha gene promoter region may predispose to
a poor prognosis in COPD. Chest 2000;118(4):971–5.
[9] Barnes PJ, Hansel TT. Prospects for new drugs for chronic
obstructive pulmonary disease. Lancet 2004;364(9438):985–96.
[10] Van der Vaart H, Koeter GH, Postma DS, Kauffman HF, ten
Hacken NH. First study of infliximab treatment in patients with
chronic obstructive pulmonary disease. Am J Respir Crit Care Med
2005;172(4):465–9.
[11] Rennard SI, Fogarty C, Kelsen S, et al. The safety and efficacy of
infliximab in moderate-to-severe chronic obstructive pulmonary
disease. Am J Respir Crit Care Med 2007;175:926–34.
[12] Pharmetrics Inc. Pharmetrics patient-centric database 2005. Water-
town, MA, USA.
[13] Suissa S, Ernst P, Hudson M, Bitton A, Kezouh A. Newer disease-
modifying antirheumatic drugs and the risk of serious hepatic adverse
events in patients with rheumatoid arthritis. Am J Med 2004;117(2):
87–92.
[14] Bernatsky S, Hudson M, Suissa S. Anti-rheumatic drug use and risk
of hospitalization for congestive heart failure in rheumatoid arthritis.
Rheumatology (Oxford) 2005;44(5):677–80.
Descargado para Pedro Argos (pedrojose.argos@scsalud.es) en Marques de Valdecilla Foundation de ClinicalKey.es por Elsevier en agosto 31, 2023.
Para uso personal exclusivamente. No se permiten otros usos sin autorización. Copyright ©2023. Elsevier Inc. Todos los derechos reservados.
[15] Suissa S. Novel approaches to pharmacoepidemiology study design
and statistical analysis. In: Strom BL, editor. Pharmacoepidemiology.
3rd ed. New York, NY: Wiley; 2000. p. 785–805.
[16] Howarth PH, Babu KS, Arshad HS, et al. Tumour necrosis factor
(TNFalpha) as a novel therapeutic target in symptomatic corticoster-
oid dependent asthma. Thorax 2005;60(12):1012–8.
[17] Berry MA, Hargadon B, Shelley M, et al. Evidence of a role of tumor
necrosis factor alpha in refractory asthma. N Engl J Med 2006;
354(7):697–708.
[18] Wenzel S. Severe asthma in adults. Am J Respir Crit Care Med
2005;172(2):149–60.
[19] Baert F, Noman M, Vermeire S, et al. Influence of immunogenicity
on the long-term efficacy of infliximab in Crohn’s disease. N Engl J
Med 2003;348(7):601–8.
[20] Jit M, Henderson B, Stevens M, Seymour RM. TNF-alpha
neutralization in cytokine-driven diseases: a mathematical model to
account for therapeutic success in rheumatoid arthritis but ther-
apeutic failure in systemic inflammatory response syndrome. Rheu-
matology (Oxford) 2005;44(3):323–31.
[21] Haraoui B. The anti-tumor necrosis factor agents are a major
advance in the treatment of rheumatoid arthritis. J Rheumatol Suppl
2005;72:46–7.
[22] Weaver AL. The impact of new biologicals in the treatment of
rheumatoid arthritis. Rheumatology (Oxford) 2004;43(Suppl 3):
iii17–23.
[23] Chapman KR, Mannino DM, Soriano JB, et al. Epidemiology and
costs of chronic obstructive pulmonary disease. Eur Respir J 2006;
27(1):188–207.
[24] Donaldson GC, Wedzicha JA. COPD exacerbations. 1: epidemiol-
ogy. Thorax 2006;61(2):164–8.