CHEST

Original Research
COPD

Combination Antihypertensive Therapy

Among Patients With COPD
Melissa A. Herrin, BA; Laura Cecere Feemster, MD; Kristina Crothers, MD;
Jane E. Uman, MPH; Chris L. Bryson, MD; and David H. Au, MD

Background: COPD and hypertension both increase the risk of congestive heart failure (CHF).
Current clinical trials do not inform the selection of combination antihypertensive therapy among
patients with COPD. We performed a comparative effectiveness study to investigate whether choice
of dual agent antihypertensive therapy is associated with risk of hospitalization for CHF among
patients with these two conditions.
Methods: We identified a cohort of 7,104 patients with COPD and hypertension receiving care
within Veterans Administration hospitals between January 2001 and December 2006, with follow-up
through April 2009. We included only patients prescribed two antihypertensive medications. We
used Cox proportional hazard models for statistical analysis.
Results: Compared with b-blockers plus an angiotensin-converting enzyme inhibitor/angiotensin
II receptor blocker, patients prescribed a thiazide diuretic plus a b-blocker (adjusted hazard ratio
[HR], 0.49; 95% CI, 0.32-0.75), a thiazide plus an angiotensin-converting enzyme inhibitor/angiotensin II receptor blocker (adjusted HR, 0.50; 95% CI, 0.35-0.71), and a thiazide plus a calcium
channel blocker (adjusted HR, 0.55; 95% CI, 0.35-0.88) had a significantly lower risk of hospitalization for CHF. After stratification by history of CHF, we found that this association was isolated
to patients without a history of CHF. Adjustment for patient characteristics and comorbidities had
a small effect on risk of hospitalization. Choice of antihypertensive medication combination had
no significant association with risk of COPD exacerbation.
Conclusions: Among patients with comorbid hypertension and COPD requiring two antihypertensive agents, combination therapy that includes a thiazide diuretic was associated with a significantly lower risk of hospitalization for CHF among patients without a history of CHF.
CHEST 2013; 143(5):13121320
Abbreviations: ACE-I 5 angiotensin-converting enzyme inhibitor; ACS 5 acute coronary syndrome; ALLHAT 5 Antihypertensive and Lipid-Lowering to Prevent Heart Attack Trial; ARB 5 angiotensin II receptor blocker; CCB 5 calcium
channel blocker; CHF 5 congestive heart failure; ICD-9 5 International Classification of Diseases, Ninth Revision;
VA 5 Veterans Administration

United States, COPD is estimated to affect

In.the
23 million people. The majority of patients with
1

COPD have other tobacco-related comorbidities,2

which place them at particularly high risk for adverse
cardiovascular outcomes. Both COPD and hypertension are major risk factors for congestive heart failure
(CHF), a condition responsible for . 1 million hospitalizations and $35 billion in health-care expenditures
in the United States each year.3,4 Optimization of BP
control is an important preventive strategy to reduce
the incidence of CHF,5 yet there is little evidence from
clinical trials to guide the selection of antihypertensive
treatment among patients with COPD.
1312

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

Most patients with hypertension require more than

one antihypertensive agent to provide adequate hypertensive control.6 Clinical trials evaluating the efficacy
of single antihypertensive therapy on improving cardiovascular outcomes have been performed,7-13 but to
date, only one trial, to our knowledge, has focused on
combination therapy.14 None of these trials have been
able to inform how to approach antihypertensive therapy among patients with COPD who require more
than one antihypertensive medication. To address this
gap, we performed a comparative effectiveness study
to examine whether the selection of combination antihypertensive agents was associated with the risk of
Original Research

hospitalization for CHF among patients with COPD

treated with dual antihypertensive therapy.
Materials and Methods
Design and Data Sources
We performed a cohort study of patients treated within the
Veterans Integrated Service Network 20 for COPD and hypertension
who were identified between January 2001 and December 2006,
with follow-up data available through April 2009. We used data
obtained from the Veterans Integrated Service Network 20 data
warehouse, which has been used extensively for both research and
clinical applications. The warehouse regularly collects detailed
sociodemographic and clinical information, including administrative diagnostic (International Classification of Diseases, Ninth
Revision [ICD-9])codes for inpatient and outpatient visits; all
pharmacy medications dispensed from Veterans Administration
(VA) pharmacies; and health behaviors, including current tobacco
use status.
Setting and Patients
We used a validated algorithm that combines ICD-9 codes
and inhaled medication use to identify a cohort of patients with
COPD.15 Specifically, patients qualified for the cohort if they met
any of the following criteria: (1) two or more outpatient diagnoses
of COPD given in 2 consecutive years, (2) given a primary hospital discharge diagnosis of COPD, (3) having filled six or more
canisters of albuterol or three or more canisters of ipratropium
bromide in a 1-year period, or (4) having filled a prescription for
either an ipratropium bromide metered dose inhaler or an albuterol
or ipratropium bromide nebulizer within 1 year of an outpatient
visit for COPD exacerbation. The date of entry into the cohort
(index date) was defined as the day when the patient fulfilled entry
criteria. For this study, we included only those patients with comorbid hypertension treated with two antihypertensive medications
in the 12 months prior to the index date.
Primary Exposure
The antihypertensive medication classes examined were thiazides, calcium channel blockers (CCBs), b-blockers, and a combined
category of angiotensin-converting enzyme inhibitors (ACE-Is)
and angiotensin II receptor blockers (ARBs). We created six mutually exclusive categories based on the possible combination of
these medications.
Manuscript received July 16, 2012; revision accepted December 1,
2012.
Affiliations: From the Health Services Research and Development (Mss Herrin and Uman and Drs Feemster, Bryson, and Au),
VA Puget Sound Health Care System, Seattle, WA; and Division
of Pulmonary and Critical Care Medicine (Drs Feemster, Crothers,
and Au) and Division of General Internal Medicine (Dr Bryson),
Department of Medicine, University of Washington School of
Medicine, Seattle, WA.
Funding/support: This material is based on work supported by
the Department of Veterans Affairs, Health Services Research and
Development, and American Lung Association Grant CI-51755N.
Dr Feemster is supported by a Veterans Affairs Health Services
Research and Development fellowship (TPM 61-037).
Correspondence to: Melissa A. Herrin, BA, 1100 Olive Way, Ste
1400, Seattle, WA 98101; e-mail: herrinmelissa@gmail.com
2013 American College of Chest Physicians. Reproduction
of this article is prohibited without written permission from the
American College of Chest Physicians. See online for more details.
DOI: 10.1378/chest.12-1770
journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

To account for health behaviors associated with differential

adherence to medications, we performed a secondary analysis of
patients adherent to both medications at a threshold of 0.80 during the 6-month period prior to entry into the cohort. This threshold was based on previous literature that suggested this level of
adherence to be adequate for preventing cardiovascular disease
outcomes.16,17 Medication adherence was assessed by ReComp
(refill compliance), a previously validated method that over longer
periods produces measures of adherence that are equivalent to
medication possession ratios.16
Outcome
The primary outcome was time to first CHF event requiring
hospitalization occurring after the index date. We defined a CHF
event as the first primary inpatient ICD-9 diagnosis code of CHF
recorded after the index date. If no CHF event occurred, patients
were censored either at the censoring date (April 28, 2009) or on
the date of death if this occurred before the censoring date. The
secondary outcome measure was time to admission for inpatient
exacerbation of COPD in the period following the index date. We
defined a COPD admission as a primary ICD-9 discharge diagnosis code of COPD. The use of a primary ICD-9 diagnosis code as
a method to define CHF and COPD has been previously examined and shown to perform reasonably well.18,19 Patients who did
not have an inpatient exacerbation were either censored at the
censoring date or on the date of death if this occurred first.
Potential Confounding and Effect-Modifying
Patient Characteristics
We collected demographic data, pharmacy records, and the
primary ICD-9 diagnosis codes for all outpatient and inpatient
visits during the 1-year period prior to entry into the cohort. Patient
characteristics considered to be potential confounders of the association between use of antihypertensive agents and the risk of hospitalization for CHF were (1) demographics and health behaviors
(age, BMI, race, smoking status, and the number of missed appointments within the past year), (2) prevalent comorbid conditions
(eg, CHF, acute coronary syndrome [ACS], diabetes mellitus, and
depression), and (3) markers of COPD severity (history of COPD
exacerbation defined as either a primary inpatient discharge diagnosis of COPD, an outpatient diagnosis of COPD, or both) and
canisters received in the past year of short-acting b-agonists and
ipratropium bromide.
Statistical Analysis
We used Cox proportional hazards models and stratification
to estimate the risk of hospitalization for CHF and to adjust for
potential confounding factors. The exposure of interest was the
combination of antihypertensive medication class. The largest
percentage of patients receiving two antihypertensive medications (32.0%) was taking a b-blocker and an ACE-I/ARB. Because
b-blockers and ACE-Is/ARBs are similarly indicated for cardiovascular disease, we chose this combination as the reference group
for all analyses. As part of our planned secondary analyses, we made
the decision a priori to stratify the cohort by a clinical history of
CHF to control for potential bias due to confounding by indication. We defined a history of CHF as either a primary inpatient
or a primary outpatient diagnosis of CHF in the year prior to the
patients date of enrollment into the cohort. We used Cox proportional hazard models to adjust for potential confounding factors.
We assessed potential effect-modifying variables on time to hospitalization for CHF and time to admission for inpatient exacerbation of COPD by adding interaction terms to the model.
Only variables that achieved P .10 in the preliminary models
were added to the final adjusted model. All statistical tests were
CHEST / 143 / 5 / MAY 2013

1313

two-tailed, and P , .05 was considered statistically significant.

Analyses were performed using SAS (SAS Institute Inc) and
STATA (StataCorp LP) statistical software.

Results
Baseline Characteristics by Medication Classes
We identified 7,104 patients with COPD and
hypertension who had been treated with two classes
of antihypertensive medication (Fig 1). Consistent with
veterans with COPD, the cohort comprised adults
aged (mean SD) 66.9 10.9 years with multiple comorbidities prevalent at baseline, including ACS, CHF,
depression, and diabetes mellitus (Table 1). Patients
within the six treatment groups had similar baseline
characteristics, including markers of COPD severity,
although the percentage of patients with ACS, CHF,
and diabetes varied among treatment groups (Table 1).
Risk of Incident Hospitalization for CHF
Median follow-up time for patients hospitalized
for CHF was 710 days. During the follow-up period,
511 (7.2%) patients were hospitalized for CHF, and
285 of these patients (55.8%) had a clinical history of
CHF. After adjustment for patient characteristics and
comorbidities, patients receiving a thiazide diuretic
plus a b-blocker, a thiazide plus an ACE-I/ARB and
a thiazide plus a CCB showed significant reduction in
risk of CHF hospitalization compared with those prescribed a b-blocker plus an ACE-I/ARB (Table 2).
Risk of CHF hospitalization among patients who were
prescribed a b-blocker plus a CCB or prescribed a
CCB plus an ACE-I/ARB was statistically similar to the
reference group (Table 2). We found similar results
when we imposed the criteria of being adherent to both
medications. Among adherent patients (n 5 2,982),

those receiving a thiazide plus a b-blocker and a thiazide plus an ACE-I/ARB demonstrated a reduced
risk of CHF hospitalization compared with the reference group (Table 2). Thiazide plus CCB had a similar
magnitude of association, but the association did not
achieve the a priori threshold of statistical significance
(Table 2). Of note, this model violated the proportional
hazards assumption but is presented here because it
was the intended primary analysis. When stratified by
previous diagnosis of CHF (planned secondary analysis presented next), this was no longer the case.
Stratification by Previous Diagnosis of CHF
Through stratified models, we assessed for effect
modification by whether the patient had previously
been given a diagnosis of CHF. We also tested to see
whether the addition of an interaction term between
previous diagnosis of CHF and dual agent antihypertensive drugs was significant, which it was not. Among
patients with previous CHF, no combination of dual
antihypertensive regimens was associated with a significant reduction in CHF hospital admission compared with the reference group receiving a b-blocker
and an ACE-I/ARB (Table 3). Among patients without previous CHF, those receiving a thiazide diuretic
plus an additional hypertensive agent were consistently
associated with a reduced risk of hospitalization for
CHF compared with the reference group (Table 3).
Adjustment for patient characteristics and comorbidities had a small effect on risk of hospitalization for
CHF (Table 3). After stratification by clinical history
of CHF among adherent patients, those without previous CHF and taking a thiazide plus b-blocker and
a thiazide plus an ACE-I/ARB showed a significant
reduction in risk for CHF (Table 3).
Risk of Incident Hospitalization
for COPD Exacerbation
To assess whether dual agent antihypertensive therapy had any association on risk of COPD exacerbation,
we evaluated the association between antihypertensive medication combination and risk of inpatient
COPD exacerbation. We found that even after adjustment for patient characteristics and comorbidities, the
addition of a multiplicative interaction term, and stratification by whether patients had previously been hospitalized for a COPD exacerbation, antihypertensive
medication combination choice appeared to have no
significant association on risk of COPD exacerbation
(Table 4).
Discussion

Figure 1. Study cohort.

1314

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

Of the estimated 23 million Americans who have

COPD,1 as many as two-thirds will have coexisting
Original Research

Table 1Baseline Characteristics of the 7,104 Patients With COPD and Hypertension by Antihypertensive
Medication Combination

Characteristic
Male sex
Age, y
BMI, kg/m2
Smoker in the past year
Race
White
Other
Unknown, declined, or missing
ACS
CHF
Depression
Diabetes mellitus
COPD exacerbation
Outpatient exacerbation
Inpatient exacerbation
Canisters received in the past year
Inhaled corticosteroids
SABA
Ipratropium bromide
Missed appointments in the past year

b-Blocker and
ACE-I/ARB
(n 5 2,275)

b-Blocker
and CCB
(n 5 867)

b-Blocker
and Thiazide
(n 5 786)

CCB and
ACE-I/ARB
(n 5 1,331)

CCB and
Thiazide
(n 5 552)

ACE-I/ARB
and Thiazide
(n 5 1,293)

2,224 (98.0)
66.7 10.4
30.4 (6.4)
897 (39.4)

839 (96.8)
68.4 10.8
29.6 (6.5)
311 (35.9)

739 (94.0)
65.6 11.2
30.4 (6.2)
311 (39.6)

1,292 (97.1)
68.4 10.5)
30.8 (6.7)
422 (31.7)

519 (94.0)
67.2 11.7
30.1 (6.6)
184 (33.3)

1,222 (94.5)
65.4 11.4
31.0 (6.5)
461 (35.7)

1,571 (69.1)
83 (3.7)
621 (27.3)
236 (10.4)
762 (33.5)
519 (22.8)
942 (41.4)
147 (6.5)
115 (5.1)
36 (1.6)

607 (70.0)
55 (6.3)
205 (23.6)
68 (7.8)
194 (22.4)
192 (22.2)
222 (25.6)
75 (8.7)
61 (7.0)
17 (2.0)

530 (67.4)
37 (4.7)
219 (27.9)
22 (2.8)
70 (8.9)
184 (23.4)
114 (14.5)
55 (7.0)
40 (5.1)
16 (2.0)

930 (69.9)
52 (3.9)
349 (26.2)
31 (2.3)
285 (21.4)
253 (19.0)
534 (40.1)
116 (8.7)
82 (6.2)
39 (2.9)

365 (66.1)
41 (7.4)
146 (26.5)
4 (0.7)
38 (6.9)
100 (18.1)
89 (16.1)
53 (9.6)
38 (6.9)
17 (3.1)

836 (64.7)
50 (3.9)
407 (31.5)
9 (0.7)
95 (7.4)
279 (21.6)
369 (28.5)
80 (6.2)
64 (5.0)
19 (1.5)

1.2 (3.4)
4.1 (5.9)
2.7 (5.0)
0.2 (0.6)

1.6 (4.3)
4.8 (6.4)
3.4 (6.6)
0.2 (0.6)

1.4 (3.8)
4.3 (5.9)
2.3 (4.4)
0.2 (0.8)

3.0 (6.0)
6.4 (8.6)
4.6 (8.0)
0.2 (0.7)

3.6 (6.6)
7.3 (9.1)
4.8 (8.7)
0.2 (0.7)

2.2 (5.0)
5.3 (7.7)
3.5 (6.9)
0.2 (0.5)

Data are presented as No. (%) or mean SD. ACE-I 5 angiotensin-converting enzyme inhibitor; ACS 5 acute coronary syndrome; ARB 5 angiotensin
receptor blocker; CCB 5 calcium channel blocker; CHF 5 congestive heart failure; SABA 5 short-acting b-agonists.

hypertension,20 yet most efficacy studies of hypertension therapy do not provide high-quality evidence on
which to base decision-making among patients with
COPD. We found that among patients with comorbid
hypertension and COPD, dual treatment of hypertension that included a thiazide was associated with
a significant reduction in hospitalization for heart failure compared with a combination of a b-blocker and
an ACE-I/ARB. Moreover, we found that this reduction
in risk was found only among patients without CHF.
None of the treatment approaches were associated with
a difference in risk of COPD exacerbations. Finally,
the results were consistent when we imposed a criterion of being adherent to both medications, suggesting
that the effect was not related to a healthy behavior
effect. These results have face validity in efficacy studies and lead us to hypothesize that the combination
of medications that included a thiazide may lead to
better volume control when used in combination with
other medications, including ACE-Is, b-blockers, and
CCBs.
The difference in associated benefit of thiazide
diuretics and the risk of heart failure admission between
patients with and without a preexisting condition has
possible biologic and nonbiologic explanations. Thiazide
diuretics have been shown to be superior to other antihypertensive agents as first-line therapy to reduce
CHF.12 However, thiazides have been shown to be
most effective at hypertension control among nonedematous individuals,21 which may have limited the
journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

effectiveness of these medications among patients with

heart failure. Preexisting cardiovascular disease is a
strong risk factor for future heart failure.22 Although
we adjusted for these factors, confounding by indication for b-blocker and ACE-I/ARB was possible.
Patients receiving a b-blocker plus an ACE-I/ARB may
be more likely to have CHF or a higher risk for CHF.
If patients were prescribed a combination of antihypertensive medications, including hydrochlorothiazide, this suggests that their CHF was less severe or
that they did not have reduced left ventricular dysfunction. Furthermore, the rate of CHF admission may be
different between medication class combinations, or
there may be discrepancies in quality of medical care
received. The present findings are consistent with an
analysis of long-term follow-up of Antihypertensive
and Lipid-Lowering to Prevent Heart Attack Trial
(ALLHAT) participants with heart failure, which concluded that once CHF develops, choice of antihypertensive regimen has little overall effect on patient
outcomes, such as mortality.23
Currently, the Seventh Report of the Joint National
Committee on Prevention, Detection, Evaluation and
Treatment of High Blood Pressure recommends thiazides as the initial drugs of choice in the treatment
of essential hypertension.24 The present results suggest that the addition of an ACE-I/ARB, b-blocker, and
CCB to a thiazide produced similar effects on CHF
outcomes. Moreover, none of the treatment options
appeared to be associated with adverse pulmonary
CHEST / 143 / 5 / MAY 2013

1315

Data are presented as No. (%) or hazard ratio (95% CI). See Table 1 legend for expansion of abbreviations.
Violated proportional hazards assumption.
bFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, CHF in previous year, ACS in previous year, diabetes in previous year, depression in previous year, and number
of canisters of ipratropium bromide in previous year.
cFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, CHF in previous year, ACS in previous year, and diabetes in previous year.
dStatistically significant.

72 (14.1)
24 (4.7)
103 (20.2)
21 (4.1)
43 (8.4)
248 (48.5)
794 (12.0)
762 (11.6)
1,228 (18.6)
531 (8.1)
1,250 (19.0)
2,027 (30.8)
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I/ARB
CCB and thiazide
ACE-I/ARB and thiazide
b-blocker and ACE-I/ARB

0.94 (0.63-1.40)
0.25 (0.10-0.62)d
0.80 (0.56-1.14)
0.67 (0.34-1.31)
0.49 (0.28-0.84)d
Reference
0.73 (0.50-1.08)
0.12 (0.05-0.31)d
0.62 (0.44-0.88)d
0.32 (0.17-0.61)d
0.25 (0.15-0.42)d
Reference
34 (15.6)
5 (2.3)
49 (22.5)
10 (4.6)
16 (7.3)
104 (47.7)
366 (13.2)
305 (11.0)
593 (21.5)
224 (8.1)
483 (17.5)
793 (28.7)
0.99 (0.76-1.30)
0.49 (0.32-0.75)d
0.84 (0.66-1.06)
0.55 (0.35-0.88)d
0.50 (0.35-0.71)d
Reference

Adjusted Modela,b
Unadjusteda
Hospitalized for
CHF After Index
Date (n 5 511)
Not Hospitalized
for CHF After Index
Date (n 5 6,592)
Antihypertensive
Medication Combination

0.76 (0.59-0.99)d
0.25 (0.16-0.38)d
0.65 (0.52-0.82)d
0.30 (0.19-0.46)d
0.27 (0.20-0.38)d
Reference

Adjusted Modelc
Unadjusted
Hospitalized for
CHF After Index
Date (n 5 218)
Not Hospitalized
for CHF After Index
Date (n 5 2,764)

Adherent Patients
All Patients

Table 2Association of Dual Agent Antihypertensive Combination Therapy and Risk of Hospitalization for CHF
1316

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

effects, including b-blockers. This finding suggests that

the addition of a second antihypertensive regimen
may be guided by other comorbid disease risk factors
not related to COPD. For example, multiple studies
have demonstrated that b-blockers are less consistently prescribed to patients with angina and COPD
perhaps because of a concern about worsening airflow obstruction.25,26 The present study provides additional evidence to help overcome this reticence.
The results are in general agreement with previously
conducted efficacy studies, including the ALLHAT12
and the PROGRESS (Perindopril Protection Against
Recurrent Stroke Study).27 The findings of ALLHAT
demonstrated that the reduction of incident CHF was
attributed to patients randomized to thiazides, whether
alone or in combination with other antihypertensive
medication classes.12 More recent analyses of ALLHAT
have confirmed the effectiveness of chlorthalidone compared with amlodipine and lisinopril and that an antihypertensive regimen that includes a diuretic remains
a fundamental component of heart failure prevention.23,28,29 The in-trial heart failure results of ALLHAT
showed that the risk of heart failure (defined as fatal,
hospitalized, or treated without hospitalization) was
higher in all comparator arms vs chlorthalidone (relative
risk, 1.38 [95% CI, 1.25-1.52] vs 1.19 [95% CI, 1.07-1.31]
for amlodipine and lisinopril, respectively).23 Furthermore, hospitalized heart failure risk decreased with
chlorthalidone compared with amlodipine during the
first year following randomization and thereafter, and
the hospitalized heart failure risk was decreased for
chlorthalidone compared with lisinopril during the
first year. It is unlikely that these differences can be
attributed significantly to antihypertensive medication
use prior to entry into ALLHAT, concomitant medications, or follow-up BP.29 ALLHAT findings concluded
that a regimen that includes a thiazide diuretic remains
important in the treatment of hypertension and in
the prevention of heart failure.
Likewise, although PROGRESS found that among
patients who had experienced a cardiovascular event,
an ACE-I alone did not prevent subsequent events,
but the addition of a diuretic with an ACE-I significantly reduced the incidence of a recurrent cardiovascular event.27 On the contrary, the ACCOMPLISH
(Avoiding Cardiovascular Events through Combination Therapy in Patients Living with Systolic Hypertension) trial found that among patients at high risk of
cardiovascular disease, the combination of benazepril
and amlodipine was superior to benazepril and hydrochlorothiazide on a pooled outcome of fatal and nonfatal cardiovascular events.14 In contrast to efficacy
studies such as ACCOMPLISH that exclude most
patients with prior cardiovascular disease at randomization, the present comparative effectiveness study
was based on real-world use of these medications in a
Original Research

journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

CHEST / 143 / 5 / MAY 2013

1317

1,159
156 (13.5)
60 (5.2)
239 (20.6)
32 (2.8)
82 (7.1)
590 (50.9)
5,433
638 (11.7)
702 (12.9)
989 (18.2)
499 (9.2)
1,168 (21.5)
1,437 (26.5)

Previous CHF, n
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I/ARB
CCB and thiazide
ACE-I/ARB and thiazide
b-blocker and ACE-I/ARB
No previous CHF, n
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I/ARB
CCB and thiazide
ACE-I/ARB and thiazide
b-blocker and ACE-I/ARB

285
38 (13.3)
10 (3.5)
46 (16.1)
6 (2.1)
13 (4.6)
172 (60.4)
226
34 (15.0)
14 (6.2)
57 (25.2)
15 (6.6)
30 (13.3)
76 (33.6)

Hospitalized for CHF

After Index Date
1.00 (0.70-1.42)
0.68 (0.36-1.28)
0.72 (0.51-1.01)
0.66 (0.29-1.50)
0.60 (0.33-1.09)
Reference
1.08 (0.71-1.63)
0.42 (0.24-0.76)c
1.01 (0.71-1.43)
0.54 (0.30-0.97)c
0.49 (0.31-0.76)c
Reference

1.01 (0.68-1.52)
0.36 (0.21-0.64)c
1.01 (0.72-1.43)
0.52 (0.30-0.90)c
0.47 (0.31-0.72)c
Reference

Adjusted Modela

0.96 (0.67-1.36)
0.59 (0.31-1.12)
0.68 (0.49-0.94)c
0.67 (0.30-1.51)
0.54 (0.31-0.95)c
Reference

Unadjusted
455
65 (14.3)
21 (4.6)
115 (25.3)
9 (2.0)
31 (6.8)
214 (47.0)
2263
293 (13.0)
280 (12.4)
470 (20.8)
212 (9.4)
442 (19.5)
566 (25.0)

Not Hospitalized for

CHF After Index Date

123
19 (15.5)
4 (3.3)
21 (17.1)
3 (2.4)
6 (4.9)
70 (56.9)
93
15 (16.1)
1 (1.1)
28 (30.1)
7 (7.5)
10 (10.8)
32 (34.4)

Hospitalized for CHF

After Index Date

Adherent Patients

1.02 (0.55-1.91)
0.08 (0.01-0.56)c
0.93 (0.55-1.55)
0.64 (0.28-1.48)
0.40 (0.19-0.81)c
Reference

0.95 (0.56-1.60)
0.62 (0.22-1.71)
0.67 (0.41-1.10)
0.89 (0.28-2.88)
0.76 (0.33-1.78)
Reference

Adjusted Modelb

Data are presented as No. (%) and hazard ratio (95% CI) unless otherwise indicated. See Table 1 legend for expansion of abbreviations.
aFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, ACS in previous year, diabetes in previous year, depression in previous year, and number of canisters of ipratropium
bromide in previous year.
bFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, ACS in previous year, and diabetes in previous year.
cStatistically significant.

Not Hospitalized for

CHF After Index Date

Antihypertensive Medication
Combination

All Patients

Table 3Association of Dual Agent Antihypertensive Combination Therapy and Risk of Hospitalization for CHF Stratified by Previous CHF

Table 4Association of Dual Agent Antihypertensive Combination Therapy and Risk of Hospitalization for
COPD Exacerbation
Antihypertensive Medication
Combination
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I
CCB and thiazide
ACE-I and thiazide
b-blocker and ACE-I
Previous inpatient exacerbation,a,e No.
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I
CCB and thiazide
ACE-I and thiazide
b-blocker and ACE-I
No previous exacerbation,e No.
b-blocker and CCB
b-blocker and thiazide
CCB and ACE-I
CCB and thiazide
ACE-I and thiazide
b-blocker and ACE-I

No Inpatient
Exacerbation
(n 5 6,569)

Inpatient
Exacerbation
(n 5 441)

Unadjusted

Adjusted Modela,b

Final Adjusted
Model With
Interaction Termc

798
736
1,213
502
1,217
2,103
371
55
36
78
34
63
105
6,198
743
700
1,135
468
1,154
1,998

52
38
100
40
67
144
62
4
7
20
9
8
14
379
48
31
80
31
59
130

0.94 (0.68-1.29)
0.69 (0.48-0.99)d
1.07 (0.83-1.38)
0.96 (0.67-1.36)
0.74 (0.55-0.99)d
Reference

0.92 (0.66-1.27)
0.89 (0.62-1.29)
1.03 (0.79-1.35)
0.96 (0.66-1.38)
0.86 (0.63-1.18)
Reference

0.37 (0.11-1.20)
1.98 (0.78-5.05)
1.62 (0.78-3.38)
1.64 (0.66-4.03)
0.83 (0.31-2.19)
Reference

0.97 (0.69-1.36)
0.78 (0.52-1.17)
0.95 (0.71-1.27)
0.87 (0.58-1.31)
0.85 (0.61-1.18)
Reference

0.98 (0.70-1.37)
0.80 (0.54-1.19)
0.95 (0.71-1.27)
0.88 (0.59-1.32)
0.86 (0.62-1.19)
Reference

Data are presented as counts and hazard ratio (95% CI). See Table 1 legend for expansion of abbreviations.
aProportional hazards assumption is violated.
bFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, smoker in previous year, missed appointments in previous
year, CHF in previous year, ACS in previous year, COPD exacerbation in previous year, number of canisters of SABA, and number of canisters of
ipratropium bromide.
cFinal model includes dual agent antihypertensive therapies, age at index date, race, BMI, smoker in previous year, missed appointments in previous
year, CHF in previous year, ACS in previous year, COPD exacerbation in previous year, number of canisters of SABA, number of canisters of
ipratropium bromide, and interaction of dual agent antihypertensive drugs with COPD exacerbation in previous year .
dStatistically significant.
eAdjusted model includes dual agent antihypertensive therapies, age at index date, race, BMI, smoker in previous year, missed appointments in
previous year, CHF in previous year, ACS in previous year, number of canisters of SABA, and number of canisters of ipratropium bromide.

population-based approach that was free of multiple

inclusion and exclusion criteria commonly found in
efficacy studies.14
This study had several strengths. We studied a
large cohort of patients with COPD, using a validated
approach to define the cohort. Moreover, this approach
identified all patients within a large geographic region,
which minimized the chances for patient selection bias.
Furthermore, we obtained the data from a diverse
population of patients within an extensive, integrated
health-care network that included patients from rural
and urban environments as well as from communitybased and academic medical centers. This reduced the
chance that bias from individual physician diagnosis
patterns or treatment practices might unduly influence
the results. We were also able to use a computerized
pharmacy database to obtain complete antihypertensive medication exposure data, including an estimation
of adherence using a validated algorithm. The VA provides medications at a reduced cost, and the majority
of veterans registered at VA primary-care clinics fill
their prescriptions at VA pharmacies,30,31 increasing
1318

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

the likelihood that patient pharmacologic exposure

data were comprehensive and comparable. Finally, we
assessed the outcome measure consistently across all
facilities within a region, enhancing the likelihood of
complete ascertainment of outcome.
This study also had several important limitations.
First, although we performed rigorous statistical methods to correct for potential confounders, including
adjusting for patient sociodemographic factors, comorbidities, and behaviors that might influence treatment
outcomes, it is possible that we were unable to completely correct for confounding factors, including
severity of comorbid illness and hypertension. This
population may be at risk for important comorbidities
that are potential causes of heart failure in the presence of COPD, such as obstructive sleep apnea. Furthermore, factors such as alcohol abuse and use of
other medications, including nonsteroidal antiinflammatory drugs, may have influenced the risk of COPD
and CHF. We did not have information on these factors available to us within the data set and, therefore,
could not include them in our models. To control for
Original Research

severity of hypertension, we restricted the cohort to

patients on two treatments. Nonetheless, patients may
be receiving dual therapy for other comorbidities, and
we may have been unable to adequately adjust for confounding by indication for other conditions with similar indications for the selected medications, such as
the prescription of ACE-Is to treat proteinuria among
patients with diabetes. Second, although we were able
to confirm that patients filled their prescriptions from
VA pharmacies, we did not assess whether they actually consumed their medications. Not taking prescribed
medications would not invalidate the results unless
medication consumption occurred disproportionally
across medication class combination. Moreover, it
seems unlikely that patients would consistently fill
medications without taking them regularly. It is also
possible that patients obtained additional medications
from other pharmacies. Third, the cohort was predominantly male, and although it has been shown that heart
failure and hypertension occur less commonly in male
patients than in female patients,32 we may not be able
to generalize conclusions from the results to female
patients. Furthermore, data on race/ethnicity other than
white were unavailable, precluding an analysis of differences in antihypertensive efficacy among race/ethnic
groups. Fourth, we cannot make inferences about
whether the use of thiazides were associated with
reductions in disease-specific mortality, although we
believe that this information would be valuable in future
studies. Finally, the outcome measure was based on
the primary ICD-9 discharge diagnosis code. Although
this approach is common and valid in observational
studies,33 we did not have echocardiography assessment
of left ventricular function, which may or may not have
been otherwise clinically defined by the attending physician or by cardiologist consultation. Without echocardiographic studies, we were unable to distinguish
whether heart failure diagnosis was attributable to
systolic or diastolic heart failure or the possible influence associated with the use or selection of b-blockers
or ACE-Is/ARBs.
In conclusion, for practical reasons such as costs and
the number of patients in such a trial, the comparative effectiveness of antihypertensive therapies among
patients with common comorbid conditions are unlikely
to be directly addressed by randomized comparative
efficacy trials. Although efficacy studies provide the
strongest internal validity, most exclude a large proportion of patients, thereby limiting their generalizability to real-world populations. COPD and hypertension
are two of the most common conditions treated in
primary-care settings, yet no data exist to inform the
treatment of hypertension among patients with COPD.
Although found no cause-and-effect relationship, the
study suggests that among patients requiring two antihypertensive agents, a combination treatment that
journal.publications.chestnet.org

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

includes a thiazide diuretic is associated with a reduced

risk of incident CHF hospital admission among patients
with COPD. The results suggest that clinicians may
consider including a thiazide diuretic when prescribing antihypertensive therapies among patients with
COPD.

Acknowledgments
Author contributions: Dr Au had full access to all of the data in
the study and takes responsibility for the integrity of the data and
the accuracy of the data analysis.
Ms Herrin: contributed to the study conception and design; data
analysis and interpretation; and writing and final approval of the
manuscript.
Dr Feemster: contributed to the data analysis and interpretation
and final approval of the manuscript.
Dr Crothers: contributed to the data analysis and interpretation
and final approval of the manuscript.
Ms Uman: contributed to the data analysis and interpretation and
final approval of the manuscript.
Dr Bryson: contributed to the data analysis and interpretation and
final approval of the manuscript.
Dr Au: contributed to the study conception and design; collection
and assembly of the data; data analysis and interpretation; and
writing and final approval of the manuscript.
Financial/nonfinancial disclosures: The authors have reported
to CHEST the following conflicts of interest: Drs Bryson and Au
are coinvestigators on a grant from Gilead for work unrelated
to this manuscript. Dr Au is a research consultant for Robert
Bosch LLC. Mss Herrin and Uman and Drs Feemster and Crothers
have reported that no potential conflicts of interest exist with any
companies/organizations whose products or services may be discussed in this article.
Role of sponsors: The views expressed in this article are those of
the authors and do not necessarily reflect the position or policy of
the Department of Veterans Affairs.

References
1. Mannino DM, Braman S. The epidemiology and economics
of chronic obstructive pulmonary disease. Proc Am Thorac
Soc. 2007;4(7):502-506.
2. Barr RG, Celli BR, Mannino DM, et al. Comorbidities, patient
knowledge, and disease management in a national sample of
patients with COPD. Am J Med. 2009;122(4):348-355.
3. Roger VL, Go AS, Lloyd-Jones DM, et al; American Heart
Association Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics2012 update:
a report from the American Heart Association. Circulation.
2012;125(1):e2-e220.
4. Heidenreich PA, Trogdon JG, Khavjou OA, et al; American
Heart Association Advocacy Coordinating Committee; Stroke
Council; Council on Cardiovascular Radiology and Intervention; Council on Clinical Cardiology; Council on Epidemiology
and Prevention; Council on Arteriosclerosis; Thrombosis and
Vascular Biology; Council on Cardiopulmonary; Critical Care;
Perioperative and Resuscitation; Council on Cardiovascular
Nursing; Council on the Kidney in Cardiovascular Disease;
Council on Cardiovascular Surgery and Anesthesia, and Interdisciplinary Council on Quality of Care and Outcomes Research.
Forecasting the future of cardiovascular disease in the United
States: a policy statement from the American Heart Association. Circulation. 2011;123(8):933-944.
5. Cohn JN. Preventing congestive heart failure. Am Fam
Physician. 1998;57(8):1901-1904.
CHEST / 143 / 5 / MAY 2013

1319

6. Gradman AH, Basile JN, Carter BL, Bakris GL; American

Society of Hypertension Writing Group. Combination therapy
in hypertension. J Am Soc Hypertens. 2010;4(1):42-50.
7. Brown MJ, Palmer CR, Castaigne A, et al. Morbidity and
mortality in patients randomised to double-blind treatment
with a long-acting calcium-channel blocker or diuretic in the
International Nifedipine GITS study: Intervention as a Goal in
Hypertension Treatment (INSIGHT). Lancet. 2000;356(9227):
366-372.
8. Dahlf B, Devereux RB, Kjeldsen SE, et al; LIFE Study Group.
Cardiovascular morbidity and mortality in the Losartan Intervention For Endpoint reduction in hypertension study (LIFE):
a randomised trial against atenolol. Lancet. 2002;359(9311):
995-1003.
9. Hansson L, Lindholm LH, Niskanen L, et al. Effect of
angiotensin-converting-enzyme inhibition compared with conventional therapy on cardiovascular morbidity and mortality
in hypertension: the Captopril Prevention Project (CAPPP)
randomised trial. Lancet. 1999;353(9153):611-616.
10. Hansson L, Hedner T, Lund-Johansen P, et al. Randomised
trial of effects of calcium antagonists compared with diuretics
and beta-blockers on cardiovascular morbidity and mortality in hypertension: the Nordic Diltiazem (NORDIL) study.
Lancet. 2000;356(9227):359-365.
11. Nissen SE, Tuzcu EM, Libby P, et al; CAMELOT Investigators.
Effect of antihypertensive agents on cardiovascular events in
patients with coronary disease and normal blood pressure:
the CAMELOT study: a randomized controlled trial. JAMA.
2004;292(18):2217-2225.
12. ALLHAT Officers and Coordinators for the ALLHAT
Collaborative Research Group; Antihypertensive and LipidLowering Treatment to Prevent Heart Attack Trial. Major
outcomes in high-risk hypertensive patients randomized to
angiotensin-converting enzyme inhibitor or calcium channel
blocker vs diuretic: the Antihypertensive and Lipid-Lowering
Treatment to Prevent Heart Attack Trial (ALLHAT). JAMA.
2002;288(23):2981-2997.
13. Yusuf S, Pfeffer MA, Swedberg K, et al; CHARM Investigators and Committees. Effects of candesartan in patients with
chronic heart failure and preserved left-ventricular ejection
fraction: the CHARM-Preserved Trial. Lancet. 2003;362(9386):
777-781.
14. Jamerson K, Weber MA, Bakris GL, et al; ACCOMPLISH
Trial Investigators. Benazepril plus amlodipine or hydrochlorothiazide for hypertension in high-risk patients. N Engl
J Med. 2008;359(23):2417-2428.
15. Cooke CR, Joo MJ, Anderson SM, et al. The validity of using
ICD-9 codes and pharmacy records to identify patients with
chronic obstructive pulmonary disease. BMC Health Serv Res.
2011;11:37.
16. Bryson CL , Au DH, Young B, McDonell MB, Fihn SD.
A refill adherence algorithm for multiple short intervals to
estimate refill compliance (ReComp). Med Care. 2007;45(6):
497-504.
17. Bryson CL, Au DH, Sun H, Williams EC, Kivlahan DR,
Bradley KA. Alcohol screening scores and medication nonadherence. Ann Intern Med. 2008;149(11):795-804.
18. Goff DC Jr, Pandey DK, Chan FA, Ortiz C, Nichaman MZ.
Congestive heart failure in the United States: is there more
than meets the I(CD code)? The Corpus Christi Heart Project.
Arch Intern Med. 2000;160(2):197-202.
19. Stein BD , Bautista A, Schumock GT, et al. The validity of
International Classification of Diseases, Ninth Revision,
Clinical Modification diagnosis codes for identifying patients

1320

Downloaded From: http://journal.publications.chestnet.org/ on 08/17/2015

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.
31.
32.

33.

hospitalized for COPD exacerbations. Chest. 2012;141(1):

87-93.
Almagro P, Cabrera FJ, Diez J, et al. Comorbidities and shortterm prognosis in patients hospitalized for acute exacerbation of COPD: the EPOC en Servicios de medicina interna
(ESMI) study. Chest. 2012;142(5):1126-1133.
Materson BJ, Reda DJ, Cushman WC, et al; The Department
of Veterans Affairs Cooperative Study Group on Antihypertensive Agents. Single-drug therapy for hypertension in men.
A comparison of six antihypertensive agents with placebo.
N Engl J Med. 1993;328(13):914-921.
He J, Ogden LG, Bazzano LA, Vupputuri S, Loria C, Whelton
PK. Risk factors for congestive heart failure in US men and
women: NHANES I epidemiologic follow-up study. Arch
Intern Med. 2001;161(7):996-1002.
Piller LB, Baraniuk S, Simpson LM, et al; ALLHAT Collaborative
Research Group. Long-term follow-up of participants with
heart failure in the antihypertensive and lipid-lowering treatment to prevent heart attack trial (ALLHAT). Circulation.
2011;124(17):1811-1818.
Chobanian AV, Bakris GL, Black HR, et al; National Heart,
Lung, and Blood Institute Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood
Pressure; National High Blood Pressure Education Program
Coordinating Committee. The Seventh Report of the Joint
National Committee on Prevention, Detection, Evaluation,
and Treatment of High Blood Pressure: the JNC 7 report.
JAMA. 2003;289(19):2560-2572.
Olenchock BA, Fonarow GG, Pan W, Hernandez A, Cannon CP;
Get With The Guidelines Steering Committee. Current use
of beta blockers in patients with reactive airway disease who
are hospitalized with acute coronary syndromes. Am J Cardiol.
2009;103(3):295-300.
Wei L, Flynn R, Murray GD, MacDonald TM. Use and
adherence to beta-blockers for secondary prevention of
myocardial infarction: who is not getting the treatment?
Pharmacoepidemiol Drug Saf. 2004;13(11):761-766.
PROGRESS Collaborative Group. Randomised trial of a
perindopril-based blood-pressure-lowering regimen among
6,105 individuals with previous stroke or transient ischaemic
attack. Lancet. 2001;358(9287):1033-1041.
Proschan M, Ford CE, Cutler JA, et al; ALLHAT Collaborative
Research Group. How much effect of different antihypertensive medications on cardiovascular outcomes is attributable to their effects on blood pressure? Stat Med. In press.
doi:10.1002/sim.5580.
Grimm RH, Davis BR, Piller LB, et al; ALLHAT Collaborative
Research Group. Heart failure in ALLHAT: did blood pressure medication at study entry influence outcome? J Clin
Hypertens (Greenwich). 2009;11(9):466-474.
Steiner JF, Koepsell TD, Fihn SD, Inui TS. A general method
of compliance assessment using centralized pharmacy records.
Description and validation. Med Care. 1988;26(8):814-823.
Steiner JF, Prochazka AV. The assessment of refill compliance using pharmacy records: methods, validity, and applications. J Clin Epidemiol. 1997;50(1):105-116.
Cui H, Miao DM, Wei ZM, et al. Prevalence of cardiovascular disease in subjects hospitalized due to chronic obstructive
pulmonary disease in Beijing from 2000 to 2010. J Geriatr
Cardiol. 2012;9(1):5-10.
Au DH, Udris EM, Curtis JR, McDonell MB, Fihn SD;
ACQUIP Investigators. Association between chronic heart
failure and inhaled beta-2-adrenoceptor agonists. Am Heart J.
2004;148(5):915-920.

Original Research