Download β-Blockers in Atrial Fibrillation Patients With or Without Heart Failure

Original Article
β-Blockers in Atrial Fibrillation Patients With
or Without Heart Failure
Association With Mortality in a Nationwide Cohort Study
Peter Brønnum Nielsen, PhD; Torben Bjerregaard Larsen, PhD;
Anders Gorst-Rasmussen, PhD; Flemming Skjøth, PhD; Gregory Y.H. Lip, MD
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
Background—Recent data suggest that β-blockers are associated with prognostic advantages in heart failure (HF) patients
without concomitant atrial fibrillation (AF), but not in HF patients with concomitant AF. We aimed to investigate
associations between β-blocker treatment and cardiovascular outcome and mortality in AF patients with and without HF.
Methods and Results—Three nationwide registries were used to identify patients with nonvalvular AF patients with or
without concomitant HF. Patients were stratified into β-blocker users and β-blocker nonusers, and according to the
presence of a HF diagnosis. We followed the patients ≤5 years after baseline. Six different cardiovascular outcomes
were investigated, including all-cause mortality and fatal thromboembolic events. Crude event rates were ascertained
and propensity-matched Cox regression was used to compare event rates according to β-blocker usage status. A total of
205 174 patients were included, where 39 741 patients had prevalent HF. In the latter subgroup of patients, the 1-year
propensity-matched hazard ratio (HR) for all-cause mortality was 0.75 (95% confidence interval, 0.71–0.79; nontreated
used as reference). For patients without concomitant HF, the propensity-matched HR for all-cause mortality was 0.78
(95% confidence interval, 0.71–0.76).
Conclusions—In this large nationwide cohort study, evidence of a lower mortality with β-blocker therapy in AF patients with
concomitant HF was observed. In addition, this association was accompanied with indications that β-blocker treatment
is also associated with a better prognosis in AF patients without concomitant HF. (Circ Heart Fail. 2016;9:e002597.
DOI: 10.1161/CIRCHEARTFAILURE.115.002597.)
Key Words: atrial fibrillation
■
cohort studies
A
■
heart failure
■
mortality
■
prognosis
patients with HF. Also, treatment with β-blockers has been
shown effective in controlling the ventricular rate.9 Although
symptomatic treatment of the AF burden is of importance for
the individual patient, it remains of prognostic importance
to investigate if β-blocker treatment is also associated with
thromboembolism.
Recently, the β-Blockers in Heart Failure Collaborative
Group meta-analyzed individual-level data from 10 randomized controlled trials, and suggested that β-blockers were
associated with an improved prognostic advantage (with lower
mortality, cardiovascular events, etc.) in HF patients without
AF; however, no prognostic benefit was observed among HF
patients with AF.10 The findings were consistent across demographic features and comorbidities, and also consistent with
a previous meta-analysis.11 Large-scale randomized trials to
investigate the potential prognostic benefits of β-blockers
trial fibrillation (AF) and heart failure (HF) are 2 burdensome chronic cardiac diseases with an increasing
prevalence as the average life expectancy increases.1–3 HF can
lead to atrial remodeling, which increases the risk of developing AF. Conversely, AF can lead to HF because of the consequences of the arrhythmia itself and impairment in diastolic
filling.4 It has been estimated that ≤40% of patients with HF
also have concomitant AF.5
See Clinical Perspective
Current guidelines recommend treatment with β-blockers
for patients with HF irrespective of rhythm disorders.6,7
In patients with AF, strict versus lenient rate control with
β-blocker treatment were equally effective.8 Notwithstanding
their use for rate control, it is an open question whether
β-blocker usage is associated with prognostic benefits in AF
Received August 25, 2015; accepted December 30, 2015.
From the Atrial Fibrillation Study Group, Department of Cardiology (P.B.N., T.B.L.) and Unit of Clinical Biostatistics and Bioinformatics (A.G.-R.,
F.S.), Aalborg University Hospital, Aalborg, Denmark; Aalborg Thrombosis Research Unit, Department of Clinical Medicine, Faculty of Health,
Aalborg University, Aalborg, Denmark (P.B.N., T.B.L., A.G.-R., F.S., G.Y.H.L.); and Institute for Cardiovascular Sciences, University of Birmingham,
City Hospital, Birmingham, United Kingdom (G.Y.H.L.).
The Data Supplement is available at http://circheartfailure.ahajournals.org/lookup/suppl/doi:10.1161/CIRCHEARTFAILURE.115.002597/-/DC1.
Correspondence to Gregory Y.H. Lip, MD, University of Birmingham Centre for Cardiovascular Sciences City Hospital, Birmingham B187QH, United
Kingdom. E-mail [email protected] or Peter Brønnum Nielsen, PhD, Aalborg Thrombosis Research Unit, Department of Clinical Medicine, Aalborg
University, Forskningens Hus, Søndre Skovvej 15, DK-9100 Aalborg, Denmark. E-mail [email protected]
© 2016 American Heart Association, Inc.
Circ Heart Fail is available at http://circheartfailure.ahajournals.org
DOI: 10.1161/CIRCHEARTFAILURE.115.002597
1
2 Nielsen et al β-Blockers in AF Patients With or Without HF
are desirable, but with the expiration of the patent protection
of β-blockers, there is little incentive in the pharmaceutical
industry for sponsoring such trials. Thus, carefully conducted
large-scale register-based studies may add to the limited current evidence about β-blocker treatment in AF patients, with
or without concomitant HF.
This study is a propensity-adjusted analysis of the prognosis with β-blocker therapy among AF and HF patients
identified using Danish nationwide registries on prescription
purchases and hospital discharge diagnoses. We hypothesized
that β-blocker therapy would be associated with an improved
cardiovascular outcome and mortality in AF patients both with
and without HF.
Methods
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
In Denmark, all information on hospitalization and prescription
claims are linked to a unique personal identifier, which is provided
at date of birth or migration. We identified records from the National
Patient Registry to obtain hospitalization diagnoses by using codes
from the International Classification of Diseases (Eighth Revision
until 1994, hereafter Tenth Revision).12 Prescription claims were
identified from the National Prescription Registry, which uses codes
from the Anatomic Therapeutic Classification system to identify the
type of drug, and contains information on date of dispensation, dosage, and strength.13 Vital status, age, sex, and information on migrations were obtained from the Civil Registration System.14
Study Population and Follow-Up Information
We identified patients with a first-time hospital AF diagnosis between
January 1, 1998 and December 31, 2013. Both hospitalized AF patients and patients referred to an ambulatory setting for AF were included. All patients with valvular AF were excluded; valvular AF was
defined as presence of mitral stenosis or heart valve replacement at
the time of AF diagnosis. We stratified patients according to the presence of a previous hospital diagnosis of HF. We excluded patients if
they died or experienced a cardiovascular event within 7 days after
the discharge date of the AF diagnosis (baseline).
We categorized patients as β-blocker users if they had claimed
a β-blocker prescription within 4 months before and ≤7 days after
hospital discharge with AF, with no claims in the period from 1 year
before to 4 months before hospital discharge with AF. Patients were
categorized as nonusers if they had not claimed a prescription in the
year preceding the AF diagnosis. Finally, patients were excluded if
they had claimed a prescription of β-blocker in the period 1 year
before ≤4 months before hospital discharge with AF. The restriction
to recent β-blocker initiators was used to reduce potential bias that
might arise from inclusion long-time prevalent β-blocker users.15
We followed the patients in the Danish National Patient Register
from 7 days after discharge with an AF diagnosis and ≤5 years after baseline. Follow-up was censored at the time of death, migration,
end-of-study (December 31, 2013), or at the occurrence of an end
point, whichever came first. Definitions of comorbidities and medications are presented in Table I in the Data Supplement.
Outcomes and Comorbidities
We examined 6 cardiovascular outcomes: the 2 primary end points
were all-cause mortality and fatal thromboembolic events. The latter was defined as death within the following 30 days of ischemic
stroke, systemic embolism (SE), pulmonary embolism (PE), or
myocardial infarction (MI). We also studied secondary end points:
a combined end point of ischemic stroke/SE and all-cause mortality;
a combined end point of ischemic stroke/SE/PE/MI. In addition, we
studied a combined end point of ischemic stroke/SE/PE; and finally
(lone) ischemic stroke. Both primary and secondary diagnoses were
included in all end point analyses (Table I in the Data Supplement);
emergency room diagnoses were not included. The outcomes for
individual patients were not adjudicated, but the National Patient
Registry have previously been validated on selected outcomes with
acceptable positive predictive values for this type of analyses.16,17
Cardiovascular comorbidities were assessed by the CHA2DS2-VASc
score18; information on concomitant pharmacotherapy ≤1 year before
baseline was also obtained. Ethical approval is not required for register-based studies in Denmark.
Statistical Methods
We performed 2 separate analyses, both according to β-blocker exposure: (1) among AF patients with concomitant HF and (2) among AF
patients without concomitant HF.
Propensity Models
For describing the association between β-blocker treatment and the
risk of a subsequent event, we used propensity matching to account for
baseline differences between β-blocker users and β-blocker nonusers
(controls). Propensity scores for β-blocker treatment were estimated
using a logistic regression model, which included information on cardiovascular comorbidities and use of concomitant pharmacotherapy
(binary indicators). Specifically, indicators from the CHA2DS2-VASc
score were extracted (sex, hypertension, diabetes mellitus, previous
stroke/transient ischemic attack, and vascular disease) and the following information on baseline medication: renin–angiotensin inhibitors,
antiplatelet therapy with aspirin, oral anticoagulant treatment, calcium channel blockers, loop diuretics, nonloop diuretics, statins, digoxin, and amiodarone. Age at AF was included as a restricted cubic
spline. When analyzing the cohort of patients with concomitant HF,
the regression model also included time because the incident HF diagnosis (categories of months ≤1 year before inclusion; categories of
years for time <1 year up to ≥5 years relative to inclusion). Propensity
matching was done one-to-one (no replacement) based on the nearest
neighbor method and a maximum difference (calliper) in the propensity score between 2 matched samples was set to 0.01 on a probability
scale. Balance of baseline characteristics after matching was assessed
by inspection of standardized mean differences; a standardized mean
difference numerically <0.1 was considered acceptable.19
Comparison of Event Rates
Unmatched and propensity-matched study population characteristics
were summarized according to β-blocker usage and to status of concomitant HF. Event rates for all the studied outcomes were calculated by dividing the number of events occurring during follow-up
with the total person-years of follow-up. HRs contrasting β-blocker
usage versus no β-blocker usage were estimated by means of Cox
proportional hazards model in the propensity-matched cohorts. To
describe the 1-year survival in the matched subgroups, we estimated
the Kaplan–Meier survival curves stratified according to β-blocker
usage and HF status. Stata version 13 was used for statistical analysis
(StataCorp LP, College Station, TX), and a 2-sided P<0.05 was considered statistically significant.
Sensitivity Analyses
With observational data, it is important to evaluate whether methodological limitations or decisions could bias findings. Accordingly,
various sensitivity analyses were performed to investigate the robustness of our findings.
The subgroup of patients who received a concomitant HF diagnosis long before the AF diagnosis and survived until inclusion into this
study may represent a selected group of HF survivors. Accordingly, to
investigate whether associations were different among patients recently diagnosed with HF, we performed a sensitivity analysis by restricting the propensity-matched cohort to patients with recently diagnosed
HF, that is, (first time) HF diagnosis within 4 months before baseline.
To investigate whether the definition of β-blocker usage by considering (recent) prevalent users might lead to healthy user bias, we
performed a sensitivity analysis along the lines of the new-user design.15 Specifically, we excluded all prevalent β-blocker users and instead defined β-blocker users as patients who claimed a prescription
within the first month after hospital discharge; accordingly, observation time was started 1 month after discharge from hospital.
3 Nielsen et al β-Blockers in AF Patients With or Without HF
We undertook a sensitivity analysis to assess the impact of the of
analytic strategy on findings by using inverse probability–weighted
Cox regression model.20,21 In detail, we constructed a logistic regression model using the same baseline variables used when constructing
the propensity-matched cohort. We used this model to predict the (inverse) probability of receiving β-blocker treatment.22 The probability
weights were entered in weighted Cox regression models, providing
estimates of the HR contrasting the hypothetical scenario where all
patients received β-blocker therapy, versus the scenario where no patients received β-blocker therapy.
Finally, given the increasing attention to patients with HF during
the past 5 to 10 years, we performed a sensitivity analysis by restricting the time period for inclusion, that is, from January 1, 2008 to
December 2013.
Results
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
We identified 227 431 patients with nonvalvular AF between
January 1, 1998 and December 31, 2013; of those 205 174 met
the inclusion criteria, with a mean follow-up time of 3.1 years.
The propensity-matched cohort of AF patients without concomitant HF comprised 110 768 patients (67%), whereas the
cohort of AF patients with concomitant HF comprised 23 896
patients (60%); see Figure 1 for flowchart and Table 1 for
patient characteristics. In general, patients with concomitant
HF were older than patients free from HF (mean age, 78 and
73 years, respectively), and more often had hypertension and
vascular diseases, including previous MI. Propensity score
models were able to balance the baseline covariates, as indicated by mean standardized difference <0.1. Further inspection revealed that those patients for whom no match could
be found were slightly older and more often hypertensive;
additionally, these patients more often received treatment with
statins as well as oral anticoagulant treatment.
β-Blocker Treatment in AF Patients With
Prevalent HF
In the original, unmatched population, event rates were higher
among AF patients with concomitant HF compared with those
without concomitant HF (Table 2). The 1-year mortality rate
(per 100 person-years) for β-blocker users with concomitant
HF was 23.2 versus 37.2 for β-blocker nonusers. Event rates
for fatal thromboembolic events were also slightly lower for
β-blocker users compared with nonusers, 2.6 versus 3.3, respectively. Event rates for any thromboembolic event were similar
in the 2 groups, 9.9 for treated and 9.7 for those not treated.
In the propensity-matched analysis, AF patients with concomitant HF who were β-blocker users had lower all-cause
mortality and fatal thromboembolic event rates compared with
nonusers, HR 0.75 (95% confidence interval [CI], 0.71–0.79)
and 0.87 (95% CI, 0.74–1.02; Figure 2; Table 3). Analyses of
secondary outcomes yielded HRs close to 1, with the upper
limits of CIs providing evidence against a harmful effect of
β-blockers.
β-Blocker Treatment in AF Patients Without HF
In the original, unmatched population, the all-cause mortality rate was lower for β-blocker users compared with nonusers (Table 2; Figure 2). In addition, the fatal thromboembolic
event rate was slightly lower among β-blocker users compared with nonusers, 1.2 versus 1.7, respectively. Analyzing
Figure 1. Flowchart of study population and propensity-matched
cohorts. AF indicates atrial fibrillation.
the combined end point of ischemic stroke/SE/PE/MI showed
event rates for β-blocker users of 6.5 versus 7.0 for nonusers.
β-blocker treatment was associated with lower all-cause
mortality, HR, 0.73 (95% CI, 0.71–0.76); and fatal thromboembolic events, 0.82 (95% CI, 0.74–0.92; Table 3). The HR
for any thromboembolic events was 0.99 (95% CI, 0.94–1.05)
for β-blocker users compared with nonusers, providing evidence against a harmful effect of β-blockers.
Long-Term Outcome Analyses
HRs for various outcomes using 5 years of follow-up were
overall consistent with results from the analysis at 1 year for
both the groups (concomitant HF or free from HF), with directions of associations unchanged (Table II in the Data Supplement). In contrast to the analysis at 1 year, there was a stronger
evidence for reduced event rate on β-blocker treatment for the
outcomes exploring thromboembolic events and ischemic
stroke.
Sensitivity Analyses
The restriction to patients with a recent diagnosis of HF
(within 4 months before baseline) yielded results which were
consistent with the main findings: the HR for all-cause mortality was 0.75 (95% CI, 0.70–0.80) and for fatal thromboembolism HR, 0.77 (95% CI, 0.62–0.95). The sensitivity analyses
in which prevalent users of β-blockers were excluded also
yielded materially similar findings as in the main analysis
(Table III in the Data Supplement).
The findings obtained from the inverse probability–
weighted Cox proportional hazard model were consistent
with the results obtained from the propensity-matched analysis (Table 3). Restricting the observation to most recent
5-year period did not alter the conclusions materially (data
not shown).
4 Nielsen et al β-Blockers in AF Patients With or Without HF
Table 1. Patient Characteristics of Total Populations and Propensity-Matched Cohorts Stratified According to β-Blocker Treatment
and Concomitant HF
AF Patients With Prevalent HF
Total Population
β-Blocker Treatment
n
Age, y*
39 741
78 (70–85)
AF Patients Free From HF
Propensity-Matched
Cohort
β-Blocker Treatment
Yes
No
Yes
19 464
20 277
80 (72–87)
76 (67–83)
β-Blocker Treatment
No
11 948
78 (70–85)
Propensity-Matched
Cohort
Total Population
11 948
165 433
78 (69–85) 73 (64–82)
β-Blocker Treatment
Yes
No
Yes
No
88 848
76 585
55 384
55 384
75 (64–83) 72 (63–80) 72 (63–81) 72 (63–81)
Sex (male)*
17 711 (45)
9413 (48)
8298 (41)
5377 (45)
5363 (45) 77 426 (47) 40 928 (46)
36 498 (48) 25 576 (46) 25 689 (46)
Age ≥65 y
33 492 (84) 17 267 (89)
16 225 (80)
10 069 (84)
9969 (83) 118 996 (72) 64 816 (73)
54 180 (71) 39 086 (71) 38 611 (70)
Age ≥75 y
24 210 (61) 13 293 (68)
10 917 (54)
7149 (60)
7099 (59) 74 702 (45) 43 319 (49)
31 383 (41) 23 186 (42) 23 405 (42)
Hypertension*
39 069 (51) 18 922 (34) 18 942 (34)
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
18 829 (47)
5915 (30)
12 914 (64)
5295 (44)
5202 (44) 59 665 (36) 20 596 (23)
Diabetes mellitus*
6926 (17)
3189 (16)
3737 (18)
1985 (17)
1973 (17) 17 360 (10)
8730 (10)
8630 (11)
5624 (10)
5700 (10)
Previous
thromboembolism*
6455 (16)
3317 (17)
3138 (15)
1864 (16)
1794 (15) 25 378 (15) 14 643 (16)
10 735 (14)
7778 (14)
7880 (14)
12 277 (31)
4960 (25)
7317 (36)
3384 (28)
3265 (27) 23 099 (14) 10 048 (11)
13 051 (17)
Previous myocardial 9565 (24)
infarction
3445 (18)
6120 (30)
2370 (20)
2651 (22) 14 428 (9)
Vascular disease*
5125 (6)
77 426 (47) 40 928 (46)
7161 (13)
7260 (13)
9303 (12)
3631 (7)
4990 (9)
36 498 (48)
1209 (2)
1353 (2)
1781 (3)
1679 (3)
Pacemaker/ICD
2341 (6)
791 (4)
1550 (8)
505 (4)
669 (6)
Abnormal renal
function
3101 (8)
1401 (7)
1700 (8)
944 (8)
876 (7)
4063 (2)
1963 (2)
2100 (3)
Nonloop diuretics
16 069 (40)
7590 (39)
8479 (42)
4803 (40)
4847 (41)
5392 (3)
2669 (3)
2723 (4)
Loop diuretics*
21 764 (55) 10 983 (56)
10 781 (53)
5984 (50)
5844 (49) 54 835 (33) 26 098 (29)
28 737 (38)
3149 (26) 29 831 (18) 16 092 (18)
13 739 (18) 13 165 (24) 13 600 (25)
Calcium channel
blockers*
9619 (24)
4568 (23)
5051 (25)
3099 (26)
Dipyridamole
1340 (3)
645 (3)
695 (3)
394 (3)
396 (3)
39 460 (24) 19 281 (22)
20 179 (26)
17 730 (32) 18 056 (33)
8874 (16)
1811 (3)
8993 (16)
1828 (3)
Renin–angiotensin 17 635 (44)
system inhibitors
(ACEi/ARBs)*
6929 (36)
10 706 (53)
4921 (41)
4859 (41)
Statins*
9560 (24)
2671 (14)
6889 (34)
2280 (19)
2260 (19) 47 228 (29) 20 504 (23)
26 724 (35)
9872 (18) 10 235 (18)
Digoxin*
9281 (23)
5387 (28)
3894 (19)
2488 (21)
2443 (20) 32 350 (20) 11 922 (13)
20 428 (27)
6191 (11)
6307 (11)
841 (2)
316 (2)
525 (3)
204 (2)
268 (>1)
266 (>1)
Amiodarone*
195 (2)
5609 (3)
885 (1)
2927 (3)
384 (>1)
2682 (4)
501 (1)
15 760 (28) 16 339 (30)
NSAID
10 726 (27)
5255 (27)
5471 (27)
3217 (27)
3214 (27) 44 817 (27) 23 772 (27)
21 045 (27) 15 147 (27) 14 922 (27)
Aspirin*
19 046 (48)
8328 (43)
10 718 (53)
5234 (44)
5162 (43) 44 817 (27) 23 772 (27)
21 045 (27) 17 499 (32) 17 970 (32)
6703 (17)
2408 (12)
4295 (21)
1721 (14)
1667 (14) 56 550 (34) 25 865 (29)
30 685 (40)
2.19 (4.03)
2.19 (4.03)
1.96 (3.90)
1.89 (3.81)
Oral anticoagulant*
Years since incident 2.19 (4.03)
HF (SD)*
NA
NA
NA
8360 (15)
NA
8782 (16)
NA
CHA2DS2-VASc score
0 (females=1)
1 (males)
≥2
NA
NA
2049 (9)
1113 (9)
21 847 (91)
10 835 (91)
NA
936 (8)
11 012 (92)
NA
29 824 (18) 11 817 (16) 18 007 (20)
11 484 (21) 11 405 (21)
29 505 (18) 14 639 (19) 14 866 (17)
11 022 (20) 10 741 (19)
10 835 (91) 11 012 (92) 106 104 (64) 50 129 (65 55 975 (63)
32 878 (59) 33 238 (60)
1113 (9)
NA
936 (8)
ACEi indicates renin-angiotensin system inhibitors; AF, atrial fibrillation; ARB, angiotensin receptor blockers; HF, heart failure; ICD, International Classification of
Diseases; NA, not applicable; and NSAID, non steroidal anti-inflammatory drug.
*Information used in propensity matching.
Discussion
In this large observational cohort reflecting clinical practice,
we analyzed the prognosis with β-blocker usage separately in
the subgroups of AF patients with and without concomitant
HF, and found that in both subgroups, β-blocker therapy at
baseline was associated with a better prognosis for various
cardiovascular outcomes. Second, there was a reduction of
all-cause mortality between 20% and 30% for patients on
5 Nielsen et al β-Blockers in AF Patients With or Without HF
Table 2. Event Rates Per 100 PY at 1 Year of Follow-Up According to Exposure of β-Blocker and Stratified by Prevalent Heart
Failure (Unmatched Populations)
Prevalent Heart Failure
Free From Heart Failure
No β-Blocker Treatment
β-Blocker Treatment
Events/PY
Rate
Events/PY
Rate
Events/PY
Rate
Events/PY
Rate
5639/15 164
37.2
3867/16 677
23.2
13 440/75 547
17.8
7242/67 160
10.8
507/15 164
3.3
433/16 677
2.6
1320/75 547
1.7
785/67 160
1.2
Ischemic stroke/SE and all-cause mortality
6069/14 862
40.8
4350/16 347
26.6
15 942/73 803
21.6
9209/65 877
14.0
Any thromboembolic event†
1409/14 589
9.7
1583/15 953
9.9
5131/73 054
7.0
4216/65 012
6.5
Thromboembolic events‡
934/14 798
6.3
884/16 280
5.4
4081/73 535
5.5
2960/65 698
4.5
Ischemic stroke
736/14 882
4.9
683/16 377
4.2
3397/73 861
4.6
2489/65 933
3.8
Outcome
All-cause mortality
Fatal thromboembolic event*
No β-Blocker Treatment
β-Blocker Treatment
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
PY indicates person-years; and SE, systemic embolism.
*Death within 30 days after an event of ischemic stroke, SE, pulmonary embolism, or myocardial infarction.
†Combined end point of ischemic stroke, SE, pulmonary embolism, or myocardial infarction.
‡Combined end point of ischemic stroke, SE, and pulmonary embolism.
β-blocker therapy. Our data suggest a potential benefit of
β-blocker treatment in patients with AF overall, corroborating current guideline recommendations on β-blocker
usage.6,7 These results should preferably be confirmed in
future studies.
The β-blockers in Heart Failure Collaborative Group
reported an individual-level meta-analysis involving 18 254
clinical trial patients.10 β-Blocker therapy was in the metaanalysis associated with a significant reduction in all-cause
mortality in patients with sinus rhythm (HR, 0.73; 95% CI,
0.67–0.80), but this was not seen in patients with AF and concomitant HF (HR, 0.97; 95% CI, 0.83–1.14). On the basis of
these findings, they recommended that β-blockers should not
be used preferentially for other rate-control medications and
not regarded as standard therapy to improve prognosis in AF
patients with concomitant HF, contrary to current treatment
guidelines. Of note, the AF group comprised only 17% of the
whole-patient cohort, and the obtained meta-analyzed results
might reflect an underpowered analysis. The lack of statistical significance for the primary outcome was also noted in all
subgroups of AF, including age, sex, left ventricular ejection
fraction (LVEF), New York Heart Association (NYHA) class,
heart rate, and baseline medical therapy. It is important to bear
in mind that, although statistical significance was not reached,
CIs from this meta-analysis were consistent with a reduction
in mortality of ≤17%: and for cardiovascular death ≤23%
Figure 2. Kaplan–Meier survival estimates (and corresponding 95% confidence interval bands) of atrial fibrillation patient with or without
heart failure stratified according to β-blocker treatment (solid line) versus no β-blocker treatment (dashed line) for 1 year of follow-up
(propensity-matched cohorts).
6 Nielsen et al β-Blockers in AF Patients With or Without HF
Table 3. HR for Different Outcomes in Propensity-Matched Cohorts Contrasting Treatment With
β-Blocker in a Stratum of Prevalent HF (Reference: No β-Blocker Treatment)
Propensity-Matched Cohort
Inverse Probability–Weighted Cohort
Outcome
Prevalent HF HR
(95% CI)
Free From HF HR
(95% CI)
Prevalent HF HR
(95% CI)
Free From HF HR
(95% CI)
All-cause mortality
0.75 (0.71–0.79)
0.73 (0.71–0.76)
0.78 (0.75–0.82)
0.72 (0.70–0.75)
Fatal thromboembolic event*
0.87 (0.74–1.02)
0.82 (0.74–0.92)
0.95 (0.81–1.11)
0.79 (0.72–0.88)
Ischemic stroke/SE and all-cause mortality
0.77 (0.74–0.82)
0.77 (0.75–0.80)
0.80 (0.77–0.84)
0.76 (0.74–0.79)
Any thromboembolic event†
1.02 (0.93–1.12)
0.99 (0.94–1.05)
1.04 (0.95–1.13)
0.98 (0.94–1.03)
Thromboembolic events‡
0.93 (0.83–1.05)
0.93 (0.87–0.98)
0.95 (0.85–1.06)
0.90 (0.86–0.95)
Ischemic stroke
0.94 (0.82–1.07)
0.96 (0.90–1.03)
0.95 (0.84–1.08)
0.93 (0.88–0.99)
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
CI indicates confidence interval; HF, heart failure; HR, hazard ratio; and SE, systemic embolism.
*Death within 30 days after an event of ischemic stroke, SE, pulmonary embolism, or myocardial infarction.
†Combined end point of ischemic stroke, SE, pulmonary embolism, and myocardial infarction.
‡Combined end point of ischemic stroke, SE, and pulmonary embolism.
reduction. Careful inspection of the subgroup of patients with
HF disclosed that 72% of the patients had an NYHA class of
III or IV, with the majority of the patients being treated with
an renin-angiotensin system inhibitors/angiotensin receptor
blocker and diuretics, and also 83% in digoxin treatment. We
cannot deduce whether these findings reflect a highly selected
clinical trial population of (very) severely ill subgroup of HF
and AF patients in the study by Kotecha et al,10 where benefit
from β-blocker treatment would be diminishing. Nonetheless,
our study corroborates a potential reduction in mortality with
CIs from propensity-adjusted analyses suggesting a 20%
to 30% reduction in mortality rates, consistently with other
reports.23–26 This reduction was observed among AF patients
both with and without concomitant HF.
We recognize that the discrepancy between obtained
results in our study compared with the study by Kotecha
et al10 may reflect differences in patient populations. Both
studies involve selected patient groups (selected trial populations versus selected populations in clinical practice) and
may not be directly comparable. In addition, we cannot rule
out unmeasured or residual confounding in our analyses.
However, sensitivity analyses, including analyses with more
restrictive inclusion criteria, did not alter our main conclusions. Kotecha et al10 investigated symptomatic HF patients
with reduced ejection fraction, whereas we did not have the
option to distinct patients according to ejection fraction. In
addition, the patients with AF included from the randomized
trials were stable or patients with permanent AF, whereas
this study started observation time at the onset of AF diagnosis. It has been previously shown that newly diagnosed
patients with AF are associated with higher mortality rates,
which could explain some of the discrepancies between
the present findings and that of Kotecha et al.27 Altogether,
definitive conclusions on the potential survival advantage
with β-blocker treatments are likely to require large-scale
randomized trials.
Our analyses of the secondary end point of thromboembolic events (end point including ischemic stroke, PE, SE,
and MI) indicated no effect of β-blocker treatment, with CIs
supporting neither a clear beneficial effect nor a clear adverse
effect. This could reflect drug characteristics: it has been suggested that selective β-blockers may confer a higher risk of
thromboembolism, compared with nonselective β-blockers.28 In
our population, 88% with prevalent HF and 86% free from HF
who received β-blocker treatment were prescribed a selective
β-blocking type. In addition, the observed rates of the secondary end points (Table 2) may be underestimated because of a
competing risk of death, which was not taking into account in
the statistical analyses.
Although our findings are in line with the current recommendations on treatment with β-blocker, some caution in
interpretation is warranted: we did only include <70% of the
available populations into the propensity-matched cohorts.
Hence, these artificially constructed cohorts might not generalize into the entire population. However encouraging, when
applying a different methodological approach, that is, using
inverse probability–weighted in the Cox model to produce
contrasts, we observed materially equal results as those from
the propensity matching (Table 3).
Limitations
This was an observational study, which can be subject to
unmeasured and residual confounding. In addition, the
analysis might be subject to confounding by indication
despite the efforts made to resolve this by using propensity matching approach. Moreover, our definitions based
on prevalent medical conditions and medication usage may
have led to selection bias, including healthy user/survivor
bias. Although we investigated these issues by means of
various sensitivity analyses, and found virtually identical
results as in the main analysis, it is unlikely that we have
fully accounted for all sources of selection and confounding biases. In addition, we acquired data from administrative registries and misclassification bias including coding
errors, cannot be ruled out. The AF diagnosis has been
shown to be accurate in Danish registries.29 However, we
were not able to distinguish between new-onset AF from
carry-over diagnosis in relation to another condition (or
7 Nielsen et al β-Blockers in AF Patients With or Without HF
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
planned follow-up visit to an ambulatory), or if it indeed is
new onset of HF.
In addition, we studied patients with nonvalvular AF
by excluding those patients with mitral stenosis and valve
replacement. Recently, the precise definition on valvular/nonvalvular AF has been debated.30 For the HF diagnosis, a validation study involving a single Danish hospital showed this
diagnosis to be specific (99% specificity) but under-reported
(29% sensitivity)16; the extent to which these findings generalize to all hospitals in Denmark is unclear. In all, we cannot
completely rule out misdiagnosis of HF nor underdiagnosis of
AF (silent or asymptomatic).
An important clinical procedure in patients with HF is
to ascertain the left ventricular ejection fraction. In our data,
we have not been able to separate patients with preserved left
ventricular ejection fraction from those with reduced left ventricular ejection fraction. Yet, these 2 groups may exhibit 2 different HF populations with different prognosis.31 Furthermore,
it was not possible to obtain an NYHA classification from
patients because this information was not available from the
registries. However, as we only included patients with a hospital diagnosis of HF, we speculate that this subgroup comprises
patients with an NYHA class II and above.
In clinical practice, β-blocker dose administration is optimized for the individual patient, an approach which has been
shown to be associated with improved outcomes.32 Because
our registers do not contain this data, we could not investigate
if patients followed such recommendations or were attended
by, for example, an HF specialist nurse (or other healthcare
personnel). This may affect outcomes in terms of benefit versus no benefit from β-blocker treatment. Finally, it is important to note that our analysis uses as an exposure the β-blocker
treatment status at baseline only. Analyzing the prognosis with
continuous treatment by taking into account treatment discontinuation/initiation during follow-up may lead to difference
but is beyond scope of this article.
Conclusions
We observed significantly reduced all-cause mortality among
β-blocker users with and without HF compared with nonusers. The rate of thromboembolic events was similar between
users and nonusers. Although randomized studies are currently needed to confirm these observational findings, this
study is consistent with current guideline recommendations
of β-blocker treatment to patients with AF and concomitant
HF, as well as a clinical benefit for AF patients without concomitant HF.
Sources of Funding
The Obel Family Foundation partly funded this research by an unrestricted grant. The sponsor had no role the design and conduct of
the study; collection, management, analysis, and interpretation of the
data; and preparation, review, or approval of the article.
Disclosures
Dr Lip has served as a consultant for Bayer/Jensen J&J, Astellas,
Merck, Sanofi, BMS/Pfizer, Biotronik, Medtronic, Portola,
Boehringer Ingelheim, Microlife, and Daiichi-Sankyo, and has
been on the speaker bureaus for Bayer, BMS/Pfizer, Medtronic,
Boehringer Ingelheim, Microlife, Roche and Daiichi-Sankyo. Dr
Larsen has served as an investigator for Janssen Scientific Affairs,
LLC, and Boehringer Ingelheim, and has been on the speaker bureaus
for Bayer, BMS/Pfizer, Roche Diagnostics, Boehringer Ingelheim,
and Takeda Pharma. The other authors report no conflicts. All authors
read, revised, and approved the final manuscript; all authors confirm
that the manuscript is an honest, accurate, and transparent account of
the study being reported; that no important aspects of the study have
been omitted.
References
1. Go AS, Mozaffarian D, Roger VL, Benjamin EJ, Berry JD, Blaha MJ,
Dai S, Ford ES, Fox CS, Franco S, Fullerton HJ, Gillespie C, Hailpern
SM, Heit JA, Howard VJ, Huffman MD, Judd SE, Kissela BM, Kittner
SJ, Lackland DT, Lichtman JH, Lisabeth LD, Mackey RH, Magid
DJ, Marcus GM, Marelli A, Matchar DB, McGuire DK, Mohler ER
III, Moy CS, Mussolino ME, Neumar RW, Nichol G, Pandey DK,
Paynter NP, Reeves MJ, Sorlie PD, Stein J, Towfighi A, Turan TN,
Virani SS, Wong ND, Woo D, Turner MB; American Heart Association
Statistics Committee and Stroke Statistics Subcommittee. Heart disease and stroke statistics–2014 update: a report from the American
Heart Association. Circulation. 2014;129:e28–e292. doi: 10.1161/01.
cir.0000441139.02102.80.
2.Heidenreich PA, Trogdon JG, Khavjou OA, Butler J, Dracup K,
Ezekowitz MD, Finkelstein EA, Hong Y, Johnston SC, Khera A, LloydJones DM, Nelson SA, Nichol G, Orenstein D, Wilson PW, Woo YJ;
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;123:933–944. doi: 10.1161/
CIR.0b013e31820a55f5.
3. Krijthe BP, Kunst A, Benjamin EJ, Lip GY, Franco OH, Hofman A,
Witteman JC, Stricker BH, Heeringa J. Projections on the number of
individuals with atrial fibrillation in the European Union, from 2000 to
2060. Eur Heart J. 2013;34:2746–2751. doi: 10.1093/eurheartj/eht280.
4. Anter E, Jessup M, Callans DJ. Atrial fibrillation and heart failure: treatment considerations for a dual epidemic. Circulation. 2009;119:2516–
2525. doi: 10.1161/CIRCULATIONAHA.108.821306.
5. Neuberger HR, Mewis C, van Veldhuisen DJ, Schotten U, van Gelder IC,
Allessie MA, Böhm M. Management of atrial fibrillation in patients with
heart failure. Eur Heart J. 2007;28:2568–2577. doi: 10.1093/eurheartj/
ehm341.
6. Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE Jr, Drazner MH,
Fonarow GC, Geraci SA, Horwich T, Januzzi JL, Johnson MR, Kasper
EK, Levy WC, Masoudi FA, McBride PE, McMurray JJ, Mitchell
JE, Peterson PN, Riegel B, Sam F, Stevenson LW, Tang WH, Tsai
EJ, Wilkoff BL; WRITING COMMITTEE MEMBERS; American
College of Cardiology Foundation/American Heart Association Task
Force on Practice Guidelines. 2013 ACCF/AHA guideline for the
management of heart failure: a report of the American College of
Cardiology Foundation/American Heart Association Task Force on
practice guidelines. Circulation. 2013;128:e240–e327. doi: 10.1161/
CIR.0b013e31829e8776.
7. McMurray JJ V, Adamopoulos S, Anker SD, Auricchio A, Böhm M,
Dickstein K, Falk V, Filippatos G, Fonseca C, Gomez-Sanchez MA,
Jaarsma T, Køber L, Lip GYH, Maggioni AP, Parkhomenko A, Pieske
BM, Popescu BA, Rønnevik PK, Rutten FH, Schwitter J, Seferovic P,
Stepinska J, Trindade PT, Voors AA, Zannad F, Zeiher A. ESC Guidelines
for the diagnosis and treatment of acute and chronic heart failure 2012:
The Task Force for the Diagnosis and Treatment of Acute and Chronic
Heart Failure 2012 of the European Society of Cardiology. Developed in
collaboration with the Heart. Rev Port Cardiol. 2013;32:e641–e661.
8.Van Gelder IC, Groenveld HF, Crijns HJ, Tuininga YS, Tijssen JG,
Alings AM, Hillege HL, Bergsma-Kadijk JA, Cornel JH, Kamp O,
Tukkie R, Bosker HA, Van Veldhuisen DJ, Van den Berg MP; RACE
II Investigators. Lenient versus strict rate control in patients with
8 Nielsen et al β-Blockers in AF Patients With or Without HF
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
atrial fibrillation. N Engl J Med. 2010;362:1363–1373. doi: 10.1056/
NEJMoa1001337.
9. Khand AU, Rankin AC, Martin W, Taylor J, Gemmell I, Cleland JG.
Carvedilol alone or in combination with digoxin for the management
of atrial fibrillation in patients with heart failure? J Am Coll Cardiol.
2003;42:1944–1951.
10. Kotecha D, Holmes J, Krum H, Altman DG, Manzano L, Cleland JG,
Lip GY, Coats AJ, Andersson B, Kirchhof P, von Lueder TG, Wedel H,
Rosano G, Shibata MC, Rigby A, Flather MD; Beta-Blockers in Heart
Failure Collaborative Group. Efficacy of β blockers in patients with heart
failure plus atrial fibrillation: an individual-patient data meta-analysis.
Lancet. 2014;384:2235–2243. doi: 10.1016/S0140-6736(14)61373-8.
11. Rienstra M, Damman K, Mulder BA, Van Gelder IC, McMurray JJ,
Van Veldhuisen DJ. Beta-blockers and outcome in heart failure and
atrial fibrillation: a meta-analysis. JACC Heart Fail. 2013;1:21–28. doi:
10.1016/j.jchf.2012.09.002.
12.Lynge E, Sandegaard JL, Rebolj M. The Danish National Patient
Register. Scand J Public Health. 2011;39(suppl 7):30–33. doi:
10.1177/1403494811401482.
13.Kildemoes HW, Sørensen HT, Hallas J. The Danish National
Prescription Registry. Scand J Public Health. 2011;39(suppl 7):38–41.
doi: 10.1177/1403494810394717.
14. Thygesen LC, Daasnes C, Thaulow I, Brønnum-Hansen H. Introduction
to Danish (nationwide) registers on health and social issues: structure,
access, legislation, and archiving. Scand J Public Health. 2011;39(suppl
7):12–16.
15. Ray WA. Evaluating medication effects outside of clinical trials: newuser designs. Am J Epidemiol. 2003;158:915–920.
16. Kümler T, Gislason GH, Kirk V, Bay M, Nielsen OW, Køber L, TorpPedersen C. Accuracy of a heart failure diagnosis in administrative registers. Eur J Heart Fail. 2008;10:658–660. doi: 10.1016/j.
ejheart.2008.05.006.
17.Wildenschild C, Mehnert F, Thomsen RW, Iversen HK, Vestergaard
K, Ingeman A, Johnsen SP. Registration of acute stroke: validity in the
Danish Stroke Registry and the Danish National Registry of Patients.
Clin Epidemiol. 2013;6:27–36. doi: 10.2147/CLEP.S50449.
18. Olesen JB, Lip GY, Hansen ML, Hansen PR, Tolstrup JS, Lindhardsen
J, Selmer C, Ahlehoff O, Olsen AM, Gislason GH, Torp-Pedersen C.
Validation of risk stratification schemes for predicting stroke and thromboembolism in patients with atrial fibrillation: nationwide cohort study.
BMJ. 2011;342:d124.
19. Stuart EA. Matching methods for causal inference: a review and a look
forward. Stat Sci. 2010;25:1–21. doi: 10.1214/09-STS313.
20. Baser O. Too much ado about propensity score models? Comparing
methods of propensity score matching. Value Health. 2006;9:377–385.
doi: 10.1111/j.1524-4733.2006.00130.x.
21. Kurth T, Walker AM, Glynn RJ, Chan KA, Gaziano JM, Berger K,
Robins JM. Results of multivariable logistic regression, propensity
matching, propensity adjustment, and propensity-based weighting under
conditions of nonuniform effect. Am J Epidemiol. 2006;163:262–270.
doi: 10.1093/aje/kwj047.
22. Curtis LH, Hammill BG, Eisenstein EL, Kramer JM, Anstrom KJ. Using
inverse probability-weighted estimators in comparative effectiveness
analyses with observational databases. Med Care. 2007;45(10 suppl
2):S103–S107. doi: 10.1097/MLR.0b013e31806518ac.
23. Al-Gobari M, El Khatib C, Pillon F, Gueyffier F. β-Blockers for the prevention of sudden cardiac death in heart failure patients: a meta-analysis
of randomized controlled trials. BMC Cardiovasc Disord. 2013;13:52.
doi: 10.1186/1471-2261-13-52.
24.Dargie HJ, Lechat P. The Cardiac Insufficiency Bisoprolol Study II
(CIBIS-II): a randomised trial. Lancet. 1999;353:9–13.
25.Hjalmarson A, Goldstein S, Fagerberg B, Wedel H, Waagstein F,
Kjekshus J, Wikstrand J, Westergren G, Thimell M, El Allaf D, Vítovec J,
Aldershvile J, Halinen M, Dietz R, Neuhaus KL, Jánosi A, Thorgeirsson
G, Dunselman P, Gullestad L, Kuch J, Herlitz J, Rickenbacher P, Ball
S, Gottlieb S, Deedwania P, Vandenhoven G, Nováková I, Danker S,
Lundström M, Meyer-Sabellek W, Balla I, Sveinsdottir M, Dorhout B,
Hildebrandt A, Szczurko I, Larsson C, Bucher E, Scott E, Dwyer D,
Julian DG, DeMets DL, Chatterjee K, Feyzi J, Lehto S, Kárpáti P, Motz
W, Samuelsson O, Viersma JW, Andersson B, Berthe C, Boutefeu JM,
Boxho G, Decroly P, Derbaudrenghien JP, Pirlet J, Henry P, Heyndrickx
G, Missault L, Nannan M, Timmermans P, Vachiery JL, Van Mieghem
W, Vandenbossche JL, Dvorák K, Herold M, Hradec J, Kána A, Petr
P, Rybka J, Smíd J, Svítil P, Toman J, Agner E, Amtorp O, Egstrup K,
Eliasen P, Gotzsche CO, Hildebrandt P, Johannesen A, Kaiser-Nielsen
P, Nielsen H, Nielsen PE, Pedersen F, Rokkedal Nielsen J, Skagen K,
Honkanen T, Hussi E, Juvonen J, Jääskeläinen H, Rinne J, Salonen
T, Andresen D, Berwing H, Bethge, Beythien, Bischoff, Bundschu,
Daniel, Darius, Delius, Drude, Dingerkus, Dück, Eichler, Fach, Förster
A, Girth, Goss, Hahn, Hambrecht, Hauf GF, Heinemann, Hepp, Janka,
Klocke, Konz, Krosse B, Kühlkamp, Lewek, Lüderitz B, Löbe, Maier,
Melchior, Müller, Mäurer, Nast, Neuhaus, Obst, Odemar, von Olshausen
KE, Pomykaj, Schmailzl KJG, Schrader, Schröder, Schwimmbeck PL,
Sigmund M, Simon, Strasser, Thilo, Vöhringer, Völler H, Weibrodt,
Wirtz, Wunderlich, Czuriga I, Hetey M, Katona A, Lengyel M, Nyárádi
A, Rednik A, Sándori K, Szabó P, Tarján J, Tenczer J, Timár S, Vályi P,
Veress G, Zámolyi K, Öze B, Thorgeirsson G, Thorgeirsson G, Bernink
PJLM, Bredero AC, Breedveld RW, Breuls PNWM, Bucx JJJ, Cornel
JH, De Milliano PAR, Dunselman PHJM, Hamer BJB, Holwerda NL,
Hoogsteen J, Hoorntje JCA, Kragten JA, Liem AH, Linssen GCM,
Michels HR, Ramdat Misier AR, Schaafsma HJ, Sijbring P, Taverne
RJT, Van Kempen LHJ, Van Stralen R, Van Veldhuisen PJ, Veerhoek
MJ, Werter CJPJ, Wesdorp JCL, Willems AR, Withagen AJAM, Zwart
PAG, Bjornerheim R, Dahle M, Dickstein K, Froland GS, Gundersen
T, Hofsoy K, Hole TL, Johansen T, Mannsverk J, Nesje P, Omland
TM, Sjödin C, Smith P, Tjonndal HA, Vikesdal O, Waage K, Adamus,
Jaworska K, Kolodziej P, Kornacewicz-Jach Z, Krzemiñska-Pakula M,
Piotrowski JW, Piwowarska W, Stogowski A, Wodniecki J, Wrabec
K, Ahlström P, Ekdahl S, Hemmingson LO, Holmberg L, Lernfelt
B, Nilsson B, Widgren H, Ångman K, Erne P, Mohacsi P, Polikar R,
Schläpfer H, Batin P, Berkin KE, Callaghan TS, Forfar J, Frenneaux
M, Greenbaum RA, Maltz M, Murdoch D, Reynolds G, Stephens J,
Struthers A, Swan J, Tildesley G, Abbasi A, Alagona P, Alderman
J, Alipour M, Anderson JL, Ansari Z, Ashraf M, Beanblossom BT,
Bennett S, Benvenuti D, Berk MR, Bhalla R, Bilazarian SD, Browne KF,
MBuchter C, Carlson R, Carlson CJ, Danisa K, Dauber I, DeWood MA,
Dennish G, Denny DM, DiBianco R, Diller PM, Dunlap M, Dowd K,
Edmiston A, El Shahawy M, Elkayam U, Farnham J, Fenster P, Friedman
S, Heywood T, Galichia JP, Geller M, Ghali JK, Gheorghiade M, Giles
T, Gillespe R, Goldberg G, Goldberg MC, Goldscher DA, Gooden GP,
Goodman M, Goodman L, Gorwit J, Gottlieb SS, Gradman A, Grech
D, Hack T, Hall JH, Hattenhauer MT, Higginbotham MB, Hutchins S,
Imburgia M, Iteld BJ, Jackson B, Jafri S, Jauch W, Jennison S, Kahn BH,
Kao W, Kaplan K, Karlsberg R, Kennedy HL, Kennedy JJ, Kirkegaard
L, Kraus DH, LaBresh K, Lalonde L, Lesser MF, Levites R, Levy M,
Lewis RK, Loh IK, Madyoon H, Maislos F, Mann D, Maurice GL, Nisar
A, Old W, Pappas JD, Phadke K, Promisloff S, Rashkow AM, Reeves B,
Rosen JH, Rotman M, Saleem T, Savran SV, Shah R, Shalev Y, Shanes
JG, O’Shaughnessy M, Silverman B, Steingart RM, Swenson L, Syed
K, Thadani U, Thorsen RD, Tonkon MJ, Touchon R, Uhl G, Vaska KJ,
Wagner SG, Weaver CJ, Weiss RJ, Wickemeyer WJ, Willens HJ, Wilson
JR, Wright R, Yellen L. Effect of metoprolol CR/XL in chronic heart
failure: Metoprolol CR/XL Randomised Intervention Trial in Congestive
Heart Failure (MERIT-HF). Lancet. 1999;353:2001–2007.
26. Packer M, Coats AJ, Fowler MB, Katus HA, Krum H, Mohacsi P, Rouleau
JL, Tendera M, Castaigne A, Roecker EB, Schultz MK, DeMets DL;
Carvedilol Prospective Randomized Cumulative Survival Study Group.
Effect of carvedilol on survival in severe chronic heart failure. N Engl J
Med. 2001;344:1651–1658. doi: 10.1056/NEJM200105313442201.
27. Wang TJ, Larson MG, Levy D, Vasan RS, Leip EP, Wolf PA, D’Agostino
RB, Murabito JM, Kannel WB, Benjamin EJ. Temporal relations of atrial
fibrillation and congestive heart failure and their joint influence on mortality: the Framingham Heart Study. Circulation. 2003;107:2920–2925.
doi: 10.1161/01.CIR.0000072767.89944.6E.
28.de Peuter OR, Souverein PC, Klungel OH, Büller HR, de Boer A,
Kamphuisen PW. Non-selective vs. selective beta-blocker treatment and
the risk of thrombo-embolic events in patients with heart failure. Eur J
Heart Fail. 2011;13:220–226. doi: 10.1093/eurjhf/hfq176.
29.Rix TA, Riahi S, Overvad K, Lundbye-Christensen S, Schmidt EB,
Joensen AM. Validity of the diagnoses atrial fibrillation and atrial flutter
in a Danish patient registry. Scand Cardiovasc J. 2012;46:149–153. doi:
10.3109/14017431.2012.673728.
30. Molteni M, Polo Friz H, Primitz L, Marano G, Boracchi P, Cimminiello
C. The definition of valvular and non-valvular atrial fibrillation: results
of a physicians’ survey. Europace. 2014;16:1720–1725. doi: 10.1093/
europace/euu178.
9 Nielsen et al β-Blockers in AF Patients With or Without HF
31.Shah AM, Solomon SD. Phenotypic and pathophysiological heterogeneity in heart failure with preserved ejection fraction. Eur Heart J.
2012;33:1716–1717. doi: 10.1093/eurheartj/ehs124.
32. Fiuzat M, Wojdyla D, Kitzman D, Fleg J, Keteyian SJ, Kraus WE,
Piña IL, Whellan D, O’Connor CM. Relationship of beta-blocker
dose with outcomes in ambulatory heart failure patients with systolic dysfunction: results from the HF-ACTION (Heart Failure:
A Controlled Trial Investigating Outcomes of Exercise Training)
trial. J Am Coll Cardiol. 2012;60:208–215. doi: 10.1016/j.
jacc.2012.03.023.
CLINICAL PERSPECTIVE
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
Atrial fibrillation (AF) and heart failure (HF) are 2 burdensome chronic cardiac diseases with an increasing prevalence as the
average life expectancy increases. Current guidelines recommend treatment with β-blockers for patients with HF irrespective of rhythm disorders. Recently, the β-Blockers in Heart Failure Collaborative Group meta-analyzed individual-level data
from 10 randomized controlled trials, and suggested that β-blockers were associated with an improved prognostic advantage
(with lower mortality, cardiovascular events, etc.) in HF patients without AF; however, no prognostic benefit was observed
among HF patients with AF. We aimed to investigate associations between β-blocker treatment and cardiovascular outcome
and mortality in AF patients with and without HF. The data were obtained from Danish nationwide registries and comprised
a total of 205 174 patients with AF where 39 741 had concomitant HF. We constructed 2 propensity-matched cohorts to balance (baseline) covariates observed in β-blockers users versus nonusers. After 1 year of follow-up, we found a propensitymatched hazard ratio for all-cause mortality of 0.75 (95% confidence interval, 0.71–0.79; nontreated used as reference). For
patients without concomitant HF, the propensity-matched hazard ratio for all-cause mortality was 0.78 (95% confidence
interval, 0.71–0.76). Hence, based on clinical practice data, a lower mortality with β-blocker therapy in AF patients with
concomitant HF was observed. In addition, this association was accompanied with indications that β-blocker treatment is
also associated with a better prognosis in AF patients without concomitant HF.
β-Blockers in Atrial Fibrillation Patients With or Without Heart Failure: Association With
Mortality in a Nationwide Cohort Study
Peter Brønnum Nielsen, Torben Bjerregaard Larsen, Anders Gorst-Rasmussen, Flemming
Skjøth and Gregory Y.H. Lip
Downloaded from http://circheartfailure.ahajournals.org/ by guest on May 2, 2017
Circ Heart Fail. 2016;9:e002597
doi: 10.1161/CIRCHEARTFAILURE.115.002597
Circulation: Heart Failure is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX
75231
Copyright © 2016 American Heart Association, Inc. All rights reserved.
Print ISSN: 1941-3289. Online ISSN: 1941-3297
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circheartfailure.ahajournals.org/content/9/2/e002597
Data Supplement (unedited) at:
http://circheartfailure.ahajournals.org/content/suppl/2016/01/29/CIRCHEARTFAILURE.115.002597.DC1
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation: Heart Failure can be obtained via RightsLink, a service of the Copyright Clearance Center, not
the Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation: Heart Failure is online at:
http://circheartfailure.ahajournals.org//subscriptions/
Supplemental Material
Supplemental Table 1. Definitions of outcomes and medication
International Classification of
Diseases 10th revision (ICD-10) code
Anatomical Therapeutic Chemical
(ATC) code
Congestive heart failure
I11.0; I13.0; I13.2; I42.0; I50
CO3C
Left ventricular dysfunction
I50.1; I50.9
Condition
Hypertension
See specified definition*
Diabetes mellitus
E10.0; E10.1; E10.9; E11.0; E11.1;
E11.9
Ischemic stroke
I63; I64
Systemic embolism
I74
Transient ischemic disease
G45
Aortic plaque
I70.0
Peripheral arterial disease
I70.2-I70.9; I71; I73.9; I74
Myocardial infarction
I21-I23
Abnormal renal function
I12; I13; N00-N05; N07; N11; N14;
N17-N19; Q61
Atrial fibrillation
I48
Abnormal renal function
I12 I13 N00 N01 N02 N03 N04 N05
N07 N11 N14 N17 N18 N19 Q61
Pacemaker/ICD
Z950
A10
581 582 583 584 59009 59320 59321
59322
Medication
Dabigatran
B01AE07
Rivaroxaban
B01AE07
Apixaban
B01AF02
Coumarin derivatives
B01AA
Aspirin
B01AC06
Beta-blockers
C07
Calcium channel blockers
C07F; C08; C09BB; C09DB
Renin-angiotensin system
inhibitors (ACEi/ARBs)
C09
Loop diuretics
C03C
Statin
C10
Digoxin
C01AA05
Amiodarone
C01BD01
Non-steroidal anti-inflammatory drugs
M01A
Non-loop diuretics
C02DA C02L C03A C03B C03D
C03E C03X C07C C07D C08G
C09BA C09DA C09XA52
Dipyridamole
B01AC07
Renin-angiotensin system inhibitors
C09
* We identified subjects with hypertension from combination treatment with at least two of the following classes of
antihypertensive Drugs:
I. Alpha adrenergic blockers (C02A, C02B, C02C)
II. Non-loop diuretics (C02DA, C02L, C03A, C03B, C03D, C03E, C03X, C07C, C07D, C08G, C09BA, C09DA,
C09XA52)
III. Vasodilators (C02DB, C02DD, C02DG, C04, C05)
IV. Beta blockers (C07)
V. Calcium channel blockers (C07F, C08, C09BB, C09DB)
VI. Renin-angiotensin system inhibitors (C09).
Supplemental Table 2. Hazard ratio for different outcomes in propensity matched cohorts contrasting treatment with beta-blocker in a stratum of prevalent HF
(reference: no beta-blocker treatment) using five years of follow-up.
Propensity matched cohort
Outcome
Inverse probability weighted cohort
Prevalent heart failure
Free from heart failure
Prevalent heart failure
Free from heart failure
HR (95% CI)
HR (95% CI)
HR (95% CI)
HR (95% CI)
All-cause mortality
0.78 (0.75 – 0.80)
0.77 (0.76 – 0.79)
0.81 (0.78 – 0.83)
0.77 (0.76 – 0.79)
Fatal thromboembolic event†
0.87 (0.78 – 0.97)
0.86 (0.80 – 0.92)
0.91 (0.82 – 1.01)
0.85 (0.80 – 0.91)
Ischemic stroke/SE and all-cause
mortality
0.78 (0.75 – 0.81)
0.81 (0.79 – 0.82)
0.81 (0.79 – 0.84)
0.80 (0.79 – 0.82)
Any thromboembolic event‡
0.91 (0.86 – 0.97)
0.99 (0.96 – 1.03)
0.93 (0.88 – 0.99)
0.97 (0.95 – 1.01)
Thromboembolic events*
0.86 (0.79 – 0.93)
0.96 (0.92 – 1.00)
0.88 (0.81 – 0.95)
0.94 (0.91 – 0.97)
Ischemic stroke
0.87 (0.80 – 0.95)
0.99 (0.95 – 1.04)
0.90 (0.83 – 0.97)
0.97 (0.93 – 1.01)
† Death within 30 days after an event of ischemic stroke, systemic embolism, pulmonary embolism or myocardial infarction.
‡ Combined endpoint of ischemic stroke, systemic embolism, pulmonary embolism and myocardial infarction.
* Combined endpoint of ischemic stroke, systemic embolism and pulmonary embolism.
Supplemental Table 3. Hazard ratio for different outcomes in propensity matched cohorts applying principals from a ‘new user’ design. Reference: no beta-blocker
treatment; one year of follow-up.
Propensity matched cohort applying ‘new user’ design
Outcome
Prevalent heart failure
Free from heart failure
HR (95% CI)
HR (95% CI)
All-cause mortality
0.66 (0.62 – 0.72)
0.71 (0.68 – 0.74)
Fatal thromboembolic event†
0.73 (0.58 – 0.92)
0.67 (0.59 – 0.76)
Ischemic stroke/SE and all-cause
mortality
0.68 (0.64 – 0.74)
0.75 (0.72 – 0.78)
Any thromboembolic event‡
0.97 (0.85 – 1.10)
1.04 (0.98 – 1.10)
Thromboembolic events*
0.85 (0.72 – 1.00)
0.92 (0.86 – 0.99)
Ischemic stroke
0.83 (0.68 – 1.00)
0.94 (0.87 – 1.01)
† Death within 30 days after an event of ischemic stroke, systemic embolism, pulmonary embolism or myocardial infarction.
‡ Combined endpoint of ischemic stroke, systemic embolism, pulmonary embolism and myocardial infarction.
* Combined endpoint of ischemic stroke, systemic embolism and pulmonary embolism.