JOURNAL OF THE AMERICAN COLLEGE OF CARDIOLOGY

VOL. 64, NO. 8, 2014

2014 BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION

ISSN 0735-1097/$36.00

PUBLISHED BY ELSEVIER INC.

Letters
Lack of Negative
Interaction Between
Use of Beta-Blockers and
Statins on Cardiovascular
Outcomes Among
Patients With or At Risk
for Atherothrombosis

study participants enrolled in the REACH registry

were from 7 geographical regions in 44 countries.
Data were collected using standardized case report
forms. Patients were followed up prospectively for 2
years and in selected countries up to 4 years for the
occurrence of cardiovascular outcomes, hospitalization, or vascular interventions.
Patients were categorized on the basis of statin
and beta-blocker use at baseline. Interaction between
beta-blocker and statin use was then tested. Analyses
were performed in the entire cohort and in the subsets with established CAD, CVD, or PAD. The primary

For many years, some have suggested that beta-

endpoint was a composite of cardiovascular death,

blockers modify the metabolic prole of patients

nonfatal MI, or nonfatal stroke. Hazard ratios (HRs)

treated for high blood pressure. A pre-specied anal-

were estimated between groups (no statin versus

ysis of ASCOTAnglo-SCandinavian Outcomes Trial) (1)

statin use).

assessing the synergistic effects of lipid-lowering and

Because of differences in key baselines character-

blood-pressure-lowering therapies suggested a nega-

istics between comparison groups, we introduced

tive interaction between beta-blockers and statins, as

propensity score adjustment in HRs estimations. Pro-

compared with calcium-channel blockers (CCBs) and

pensity score was calculated using a multivariable

statins, with an increased rate of ischemic events

logistic regression model, with the dependent variable

among patients receiving both therapies. However,

of statin use, and 23 covariates describing baselines

this observation was not ofcially conrmed (2) and

characteristics (sex, age, region, BMI), medical history

therefore uncertainty persists. Given that many pa-

(smoking

tients receive both types of agents, particularly in

stroke, stroke, stable angina, carotid angioplasty or

secondary prevention after acute coronary syn-

stenting, carotid surgery, aortic valve stenosis, dia-

dromes, clarifying whether such an interaction exists

betes) and baseline medication (use of aspirin, anti-

or not is important. A recent analysis of beta-blocker

coagulant, anti-diabetics agent, lipid-lowering agent

use in stable atherothrombotic patients failed to show

other than statin, CCBs, nitrates, or other anti-anginal

a benet of beta-blockers in this population (3).

agents, diuretics, angiotensin-converting enzyme in-

Whether this lack of benet is due to a worsened

hibitors, angiotensin-II receptor antagonists, other

metabolic prole on beta-blockers remains unknown.

antihypertensive drugs, or peripheral arterial claudi-

We sought to study whether a negative interaction

cation drug). The same propensity scores were used

exists between beta-blocker use and the benet of

for adjustment in global cohort and in subgroups

statins in patients with atherothrombosis treated

analyses. Statistical analyses were all performed using

in primary or secondary prevention, using the REACH

SAS version 9.3 (SAS Institute, Cary, North Carolina).

status,

history

of

transient

ischemic

Continued

Based on the REACH registry (n 65,531), we

Health) registry. The design, methods, and main

assessed statins and beta-blockers at baseline in 65,181

results of REACH, an international, prospective,

patients (99.5%). Within this cohort, 45,312 patients

observational study, have been published (4,5).

(69.5%) were treated by statins and 30,971 patients

Briey, REACH enrolled consecutive patients, aged

(47.5%) by beta-blockers. Regarding the medical his-

45 years or older, with established coronary artery

tory, 38,758 patients (59.5%) had CAD, 18,102 patients

disease (CAD), cerebrovascular disease (CVD), or

(27.8%) had CVD, and 7,947 patients (12.2%) had PAD

peripheral arterial disease (PAD), or with at least 3

at baseline. Overall, median follow-up was 37 months

atherothrombotic risk factors. All patients provided

(interquartile range [IQR]: 21 to 45 months).

(Reduction

in

Atherothrombosis

for

signed informed consent and the institutional review

Statin use was consistently associated with lower

board in each country approved the protocol. The

adjusted primary endpoint event rate. This was

Downloaded From: http://content.onlinejacc.org/ by Felix Malpartida-Torres on 08/19/2014

846

JACC VOL. 64, NO. 8, 2014

Letters

AUGUST 26, 2014:8459

Subgroups

No.of Events/No. at risk (KM estim.)

Hazard Ratios

No Statins use

Statins use

HR 95% CI

Non Beta-blockers Use (global)

1623/12851 (16.94%)

1877/21383 (12.47%)

0.73 [0.68 - 0.79]

Beta-blockers Use (global)

942/7194 (18.03%)

2346/23753 (13.64%)

0.79 [0.72 - 0.86]

Non Beta-blockers Use (CAD)

662/4611 (19.39%)

1036/10220 (14.49%)

0.72 [0.64 - 0.81]

Beta-blockers Use (CAD)

648/4672 (18.89%)

1926/19255 (13.69%)

0.74 [0.67 - 0.82]

Non Beta-blockers Use (CVD)

854/5658 (19.43%)

731/5911 (16.82%)

0.82 [0.73 - 0.92]

Beta-blockers Use (CVD)

403/2283 (23.97%)

668/4250 (21.38%)

0.89 [0.77 - 1.02]

Non Beta-blockers Use (PAD)

309/2053 (18.93%)

383/2698 (19.86%)

Beta-blockers Use (PAD)

138/807 (24.76%)

347/2389 (19.85%)

0.83 [0.66 - 1.05]

Non Beta-blockers Use (global)

1347/12851 (14.68%)

1475/21383 (10.2%)

0.66 [0.6 - 0.72]

Beta-blockers Use (global)

760/7194 (14.73%)

1774/23753 (10.62%)

0.71 [0.64 - 0.78]

Non Beta-blockers Use (CAD)

583/4611 (17.43%)

854/10220 (12.21 %)

0.64 [0.56 - 0.72]

Beta-blockers Use (CAD)

531/4672 (15.65%)

1472/19255 (10.77%)

0.68 [0.61 - 0.77]

Non Beta-blockers Use (CVD)

617/5658 (14.82%)

466/5911 (11.31%)

0.72 [0.62 - 0.83]

Beta-blockers Use (CVD)

271/2283 (16.5%)

438/4250 (14.44%)

0.81 [0.68 - 0.97]

Non Beta-blockers Use (PAD)

316/2053 (20.08%)

346/2698 (18%)

0.83 [0.69 - 1]

Beta-blockers Use (PAD)

115/807 (20.15%)

296/2389 (17.54%)

0.88 [0.69 - 1.14]

P Value for interaction

CV death/MI/Stroke
0.94

0.75

0.59

1 [0.83 - 1.2]

0.07

All death
0.76

0.76

0.39

0.84

0.50

0.75
Favor Statins

1.00

1.25

1.50

Favor No Statins

F I G U R E 1 Total Death Based on Beta-Blocker and Statin Usage

The image depicts the occurrence of primary or secondary endpoint and of total death according to the beta-blocker and statin use status.
CAD coronary artery disease; CV cardiovascular; CVD cerebrovascular disease; MI myocardial infarction; PAD peripheral artery disease.

observed among patients treated with beta-blockers

to 0.86]), and also among patients not on beta-blockers

Jrmie Abtan, MDy

Yedid Elbez, MScy
Deepak L. Bhatt, MD, MPHz
*Philippe Gabriel Steg, MDyx

(1,623 [16.9%] vs. 1,877 events [12.5%], HR: 0.73 [95%

*Hpital Bichat

CI: 0.68 to 0.79]), for patients receiving statins or not

46 rue Henri Huchard

respectively. There was no interaction between beta-

75018 Paris

blocker use and the benet of statins (p value for

France

interaction: 0.94). Consistent results were observed

E-mail: gabriel.steg@bch.aphp.fr

across all patient subsets (Fig. 1), with no evidence of

http://dx.doi.org/10.1016/j.jacc.2014.04.047

negative

were

From the yDpartement Hospitalo-Universitaire Fibrosis-

similar for each of the components of the primary

Inammation-REmodelling: Hpital Bichat (Assistance

(942 [Kaplan-Meier estimate: 18.0%] vs. 2,346 events

[Kaplan-Meier estimate: 13.6%], HR: 0.79 [95% CI: 0.72

interaction.

Furthermore,

results

outcome analyzed separately.

There are some limitations to this analysis. The

PubliqueHpitaux de Paris), and INSERM U-1148, Universit Paris-Diderot, Paris, France;

zBrigham

and

type of beta-blocker was not collected, and it has been

Womens Hospital and Harvard Medical School, Boston,

suggested that a negative interaction of beta-blockers

Massachusetts; and the xNational Heart and Lung Insti-

may be more marked for nonselective beta-blockers or

tute, Institute of Cardiovascular Medicine and Science,

for those with sympathomimetic activity. Although

Royal Brompton Hospital, Imperial College, London,

propensity score matching adjusts for baseline dif-

United Kingdom

ferences, the possibility of residual confounding

cannot be ruled out. Furthermore, unlike ASCOT, a
randomized trial with comparable groups, our data are
from a registry, probably ruling out a quantitative but
not qualitative interaction.
In this large international contemporary cohort for
those patients at risk for atherothrombosis or not,
there was no evidence of a negative interaction between beta-blockers and the benet of statins on
cardiovascular outcomes.

Downloaded From: http://content.onlinejacc.org/ by Felix Malpartida-Torres on 08/19/2014

Please note: Dr. Bhatt is a member of Elsevier Practice Update Cardiology,

Medscape Cardiology, and Regado Biosciences advisory boards; serves on the
boards of directors of Boston Veterans Affairs Research Institute and Society of
Cardiovascular Patient Care; is Chair of American Heart Association Get With
The Guidelines Steering Committee; has received honoraria from American
College of Cardiology (Editor, Clinical Trials, Cardiosource), Belvoir Publications
(Editor in Chief, Harvard Heart Letter), Duke Clinical Research Institute (clinical
trial steering committees), Population Health Research Institute (clinical trial
steering committee), Slack Publications (Chief Medical Editor, Cardiology
Todays Intervention), and WebMD (CME steering committees); serves as Senior
Associate Editor, Journal of Invasive Cardiology, and Data Monitoring Committee member of Duke Clinical Research Institute, Harvard Clinical Research
Institute, Mayo Clinic, and Population Health Research Institute; and has

JACC VOL. 64, NO. 8, 2014

Letters

AUGUST 26, 2014:8459

received research grants from Amarin, AstraZeneca, Bristol-Myers Squibb, Eisai,

Ethicon, Medtronic, Sano-Aventis, and The Medicines Company; and unfunded research from FlowCo, PLx Pharma, and Takeda. Dr. Steg has received
research grants through Unit INSERM U-1148 from Sano-Aventis and Servier;
has consulted with and served as speaker for Amarin, AstraZeneca, Bayer,
Boehringer-Ingelheim, Bristol-Myers-Squibb, Daiichi-Sankyo, GlaxoSmithKline,
Iroko Cardio, Lilly, Merck, Novartis, Otsuka, Pzer, Sano, Servier, The Medicines Company, and Vivus; and is a stockholder in Aterovax. All other authors
have reported they have no relationships relevant to the contents of this paper
to disclose.

range of dabigatran to prevent both ischemia and

REFERENCES

of age, it would be interesting to analyze the results in

1. Sever P, Dahlf B, Poulter N, et al. Potential synergy between lipid-lowering

and blood-pressure-lowering in the Anglo-Scandinavian Cardiac Outcomes
Trial. Eur Heart J 2006;27:29828.
2. Sacks FM, Tonkin AM, Shepherd J, et al. Effect of pravastatin on coronary
disease events in subgroups dened by coronary risk factors: the Prospective
Pravastatin Pooling Project. Circulation 2000;102:1893900.
3. Bangalore S, Steg G, Deedwania P, et al. b-Blocker use and clinical outcomes in stable outpatients with and without coronary artery disease. JAMA
2012;308:13409.
4. Bhatt DL, Steg PG, Ohman EM, et al. International prevalence, recognition,
and treatment of cardiovascular risk factors in outpatients with atherothrombosis. JAMA 2006;295:1809.
5. Bhatt DL, Eagle KA, Ohman EM, et al. Comparative determinants of 4-year
cardiovascular event rates in stable outpatients at risk of or with atherothrombosis. JAMA 2010;304:13507.

bleeding?
Table 2 in the report by Reilly et al. (1) shows that
dabigatran plasma concentrations were higher in
patients 75 years of age or older, but no data are given
for patients older than 80 years of age. Because many
patients with atrial brillation are older than 80 years
this subgroup of patients more precisely.
What was the time interval between measurement
of the dabigatran level and occurrence of the bleeding
or ischemic event? Did patients with higher levels
develop bleeding earlier than patients with lower
levels, and did patients with low levels develop
ischemic events earlier than patients with higher
levels? Were there any patients in whom plasma
dabigatran levels were investigated at the time when
the bleeding or ischemic event occurred? Was the
volume of the bleeding or the size of the ischemic
stroke correlated with the dabigatran levels?
The authors report only the association of plasma
dabigatran levels with major bleeding. It would be of

Knowing Plasma
Concentrations Improves
Management of the
Effects of Dabigatran

interest to know if this association was also found in

patients with minor bleeding.
In the discussion, it is mentioned that an assay of
dabigatran concentration is not yet available. When
will this assay be available?
The RE-LY (Randomized Evaluation of Long-Term
Anticoagulation Therapy) study was published in

With interest we read the report by Reilly et al. (1) in a

recent issue of the Journal regarding the effect of
dabigatran plasma concentrations and patient characteristics on the frequency of ischemic stroke and
major bleeding in patients with atrial brillation. We
have the following comments and concerns.
Dabigatran is a substrate of the P-glycoprotein
system (P-gp). Many drugs are known to inhibit or
induce the activity of P-gp (2). Was there any association between comedication of drugs affecting P-gp
activity and plasma concentrations of dabigatran? It
is known that P-gp activity is dependent on polymorphisms (3). Did the authors investigate the prevalence of P-gp polymorphisms and its association with
dabigatran levels?
Because metabolization of a substance also de-

2009. Since that time, the drug has been used

in thousands of patients, and many bleeding and
ischemic episodes have been reported outside clinical
trials (4). Why were the results of the present study not
reported sooner than 5 years after publication of the
RE-LY study? Knowledge of the importance of dabigatran levels would have saved some of these patients.
Overall, this study provides evidence that prevention of ischemic stroke by dabigatran is dependent on
the plasma concentration of the drug and cannot be
managed simply by using 2 different types of dosages.
There is a strong need to determine the dose of this
drug according to the plasma concentration and to nd
the optimal plasma concentration that best prevents
the recurrence of stroke but also the occurrence of a
bleeding event.

know if the samples were all taken at the same time of

*Claudia Stllberger, MD
Josef Finsterer, MD, PhD

the day. Renal function may deteriorate over time,

*Krankenanstalt Rudolfstiftung

especially in elderly patients, due to comorbidities.

Steingasse 31/18

Did the authors observe an increase of plasma dabi-

A-1030 Vienna

gatran levels in patients whose renal function deteri-

Austria

pends on diurnal rhythms, it would be interesting to

orated over time? Were the plasma levels of dabigatran

E-mail: claudia.stoellberger@chello.at

correlated with renal function? What was the optimal

http://dx.doi.org/10.1016/j.jacc.2014.04.074

Downloaded From: http://content.onlinejacc.org/ by Felix Malpartida-Torres on 08/19/2014

847