Beta-blockers for Heart Failure With Reduced, Mid-range, and Preserved Ejection Fraction

An Individual Patient-level Analysis of Double-blind Randomized Trials

John G.F. Cleland; Karina V. Bunting; Marcus D. Flather; Douglas G. Altman; Jane Holmes; Andrew J.S. Coats; Luis Manzano; John J.V. McMurray; Frank Ruschitzka; Dirk J. van Veldhuisen; Thomas G. von Lueder; Michael Böhm; Bert Andersson; John Kjekshus; Milton Packer; Alan S. Rigby; Giuseppe Rosano; HansWedel; A° ke Hjalmarson; John Wikstrand; Dipak Kotecha

Disclosures

Eur Heart J. 2018;39(1):26-35.

Abstract and Introduction

Abstract

Aims Recent guidelines recommend that patients with heart failure and left ventricular ejection fraction (LVEF) 40–49% should be managed similar to LVEF ≥ 50%. We investigated the effect of beta-blockers according to LVEF in double-blind, randomized, placebo-controlled trials.

Methods and results Individual patient data meta-analysis of 11 trials, stratified by baseline LVEF and heart rhythm (Clinicaltrials.gov: NCT0083244; PROSPERO: CRD42014010012). Primary outcomes were all-cause mortality and cardiovascular death over 1.3 years median follow-up, with an intention-to-treat analysis. For 14 262 patients in sinus rhythm, median LVEF was 27% (interquartile range 21–33%), including 575 patients with LVEF 40–49% and 244 ≥ 50%. Beta-blockers reduced all-cause and cardiovascular mortality compared to placebo in sinus rhythm, an effect that was consistent across LVEF strata, except for those in the small subgroup with LVEF ≥ 50%. For LVEF 40–49%, death occurred in 21/292 [7.2%] randomized to beta-blockers compared to 35/283 [12.4%] with placebo; adjusted hazard ratio (HR) 0.59 [95% confidence interval (CI) 0.34–1.03]. Cardiovascular death occurred in 13/292 [4.5%] with beta-blockers and 26/283 [9.2%] with placebo; adjusted HR 0.48 (95% CI 0.24–0.97). Over a median of 1.0 years following randomization (n = 4601), LVEF increased with beta-blockers in all groups in sinus rhythm except LVEF ≥50%. For patients in atrial fibrillation at baseline (n = 3050), beta-blockers increased LVEF when < 50% at baseline, but did not improve prognosis.

Conclusion Beta-blockers improve LVEF and prognosis for patients with heart failure in sinus rhythm with a reduced LVEF. The data are most robust for LVEF < 40%, but similar benefit was observed in the subgroup of patients with LVEF 40–49%.

Introduction

Double-blind, randomized, placebo-controlled trials (RCTs) show that beta-blockers increase left ventricular ejection fraction (LVEF) and reduce morbidity and mortality for a broad range of patients with a reduced LVEF in sinus rhythm.^[1,2] Until recently, international guidelines on heart failure have recognized two left ventricular phenotypes; heart failure with reduced LVEF (HFrEF) or preserved LVEF (HFpEF).^[3,4] Values for LVEF are continuously distributed but measurement precision is imperfect; differences of up to 10% for an individual patient may be attributed to measurement error^[5] and therefore precise cut-points of LVEF cannot reliably differentiate between phenotypes. Recently, the European Society of Cardiology (ESC) suggested there should be a third intermediate phenotype, called mid-range ejection fraction (HFmrEF; 40–49%), thereby creating a clear separation between HFrEF (<40%) and HFpEF (≥50%).^[4] These guidelines suggest that until more information becomes available, patients with HFmrEF should be managed similarly to those with HFpEF, for which no therapy has been shown to improve mortality.^[4]

The Beta-blockers in Heart Failure Collaborative Group (BB-meta-HF) was created to pool individual patient data (IPD) from the major heart failure RCTs comparing beta-blockers and placebo to address key issues in relevant patient subgroups.^[6] Most, but not all of these trials recruited patients with an LVEF ≤35% predominantly in sinus rhythm; IPD provides an opportunity to collate high-quality data from double-blind trials on the smaller number of patients with higher LVEF where the efficacy of beta-blockers is uncertain. Why beta-blockers appear ineffective in patients with heart failure and concomitant atrial fibrillation (AF),^[2,7,8] and whether this holds true regardless of LVEF is also unclear. In this paper, we investigate the effect of beta-blockers on LVEF and prognosis, stratified according to the baseline LVEF and heart rhythm.

1 2 3 4 5

Next Section

Abstract and Introduction
Methods
Results
Discussion
Conclusion
References

Table 1. Baseline characteristics for patients in sinus rhythm
Characteristic	Left ventricular ejection fraction at baseline
Characteristic	<20% n = 2553	20–25% n = 3885	26–34% n = 5076	35–39% n = 1929	40–49% n = 575	≥50% n = 244
LVEF, median (IQR)	0.15 (0.13–0.18)	0.23 (0.21–0.24)	0.30 (0.28–0.32)	0.36 (0.35–0.38)	0.40 (0.40–0.43)	0.58 (0.53–0.65)
Age, median years (IQR)	61 (51–69)	63 (54–71)	64 (55–71)	64 (56–72)	71 (61–75)	75 (72–78)
Women, n (%)	521 (20.4%)	886 (22.8%)	1272 (25.1%)	518 (26.9%)	198 (34.4%)	129 (52.9%)
Years with HF diagnosis, median (IQR)	3 (1–6)	3 (1–6)	2 (1–5)	2 (1–5)	2 (1–5)	2 (1–5)
Ischaemic HF aetiology, n (%)	1484 (58.1%)	2572 (66.2%)	3475 (68.5%)	1562 (81.0%)	522 (90.8%)	209 (85.7%)
Prior myocardial infarction, n (%)	1242 (48.7%)	2187 (56.4%)	2993 (59.2%)	1374 (71.4%)	412 (71.8%)	88 (36.1%)
Diabetes mellitus, n (%)	575 (25.1%)	956 (26.0%)	1153 (23.9%)	409 (22.2%)	135 (24.1%)	71 (29.1%)
NYHA class III/IV, n (%)	1624 (82.1%)	2045 (77.6%)	3265 (64.8%)	721 (37.7%)	136 (24.1%)	64 (26.6%)
Heart rate, median bpm (IQR)	84 (76–92)	80 (72–90)	78 (72–87)	76 (70–84)	76 (68–82)	75 (68–83)
Systolic BP, median mmHg (IQR)	114 (104–127)	120 (110–136)	127 (115–140)	130 (116–140)	131 (120–145)	147 (132–160)
Diastolic BP, median mmHg (IQR)	72 (66–80)	77 (70–82)	79 (70–83)	80 (70–83)	80 (70–85)	82 (78–90)
Body mass index, median kg/m² (IQR)	27 (24–32)	27 (24–31)	27 (24–31)	27 (25–30)	27 (25–30)	27 (24–31)
Estimated GFR, median mL/min (IQR)	62 (50–76)	61 (48–75)	66 (53–80)	65 (53–78)	66 (53–78)	69 (55–83)
Any diuretic therapy, n (%)	2410 (94.4%)	3547 (91.3%)	4331 (85.3%)	1273 (66.0%)	376 (65.4%)	199 (81.6%)
ACEi or ARB, n (%)	2304 (94.8%)	3490 (94.7%)	4643 (94.8%)	1774 (95.1%)	508 (90.6%)	203 (87.3%)
Aldosterone antagonists, n (%)	207 (8.8%)	381 (10.4%)	360 (7.5%)	85 (4.7%)	31 (5.8%)	27 (11.9%)
Digoxin, n (%)	1833 (73.8%)	2297 (60.4%)	2475 (49.9%)	555 (29.6%)	138 (25.6%)	48 (21.2%)

Missing data report: n = 2828 for years with HF diagnosis; n = 30 for prior myocardial infarction; n = 809 for diabetes mellitus; n = 1504 for NYHA class; n = 62 for systolic BP; n = 67 for diastolic BP; n = 8 heart rate; n = 123 for body mass index; n = 664 for GFR; n = 918 for aldosterone antagonists; n = 376 for digoxin. ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; b.p.m., beats/minute; GFR, glomerular filtration rate; HF, heart failure; IQR, interquartile range; LVEF, left-ventricular ejection fraction; NYHA, New York Heart Association.

Table 2. Beta-blockers vs. placebo according to left ventricular ejection fraction at baseline
Baseline heart rhythm and LVEF category	All-cause mortality		Cardiovascular death		Cardiovascular hospitalization		Composite of cardiovascular death or cardiovascular hospitalization
Baseline heart rhythm and LVEF category	Events/N	HR (95% CI); P-value	Events/N	HR (95% CI); P-value	Events/N	HR (95% CI); P-value	Events/N	HR (95% CI); P-value
Sinus rhythm
<20%	623/2531	0.70 (0.60–0.83); P < 0.001	517/2531	0.67 (0.56–0.80); P < 0.001	762/2407	0.70 (0.60–0.81); P < 0.001	990/2407	0.68 (0.60–0.77); P < 0.001
20–25%	619/3862	0.76 (0.65–0.89); P = 0.001	521/3862	0.78 (0.65–0.92); P = 0.004	1033/3807	0.75 (0.66–0.85); P < 0.001	1273/3807	0.75 (0.67–0.84); P < 0.001
26–34%	631/5043	0.75 (0.64–0.88); P < 0.001	504/5042	0.73 (0.61–0.87); P < 0.001	1118/4972	0.84 (0.74–0.94); P = 0.003	1384/4972	0.80 (0.72–0.88); P < 0.001
35–39%	189/1919	0.67 (0.50–0.90); P = 0.007	156/1919	0.72 (0.52–0.99); P = 0.041	401/1907	0.75 (0.61–0.91); P = 0.004	490/1907	0.74 (0.62–0.88); P = 0.001
40–49%	55/570	0.59 (0.34–1.03); P = 0.066	38/570	0.48 (0.24–0.97); P = 0.040	144/566	0.95 (0.68–1.32); P = 0.76	164/566	0.83 (0.60–1.13); P = 0.23
≥50%	24/241	1.79 (0.78–4.10); P = 0.17	15/241	1.77 (0.61–5.14); P = 0.29	50/241	0.66 (0.37–1.18); P = 0.16	54/241	0.66 (0.38–1.15); P = 0.14
Atrial fibrillation
<20%	143/492	1.23 (0.88–1.74); P = 0.23	124/492	1.16 (0.80–1.67); P = 0.44	148/471	0.97 (0.69–1.35); P = 0.85	201/471	0.96 (0.72–1.28); P = 0.79
20–25%	159/867	0.74 (0.54–1.02); P = 0.07	136/867	0.77 (0.54–1.08); P = 0.13	234/856	0.75 (0.58–0.98); P = 0.032	291/856	0.75 (0.59–0.95); P = 0.015
26–34%	208/1093	0.98 (0.74–1.29); P = 0.87	166/1093	0.98 (0.72–1.34); P = 0.92	321/1083	1.01 (0.81–1.26); P = 0.92	390/1083	0.93 (0.76–1.13); P = 0.47
35–39%	59/363	0.92 (0.53–1.58); P = 0.75	46/363	0.67 (0.35–1.25); P = 0.21	99/358	0.90 (0.60–1.36); P = 0.62	121/358	0.94 (0.65–1.37); P = 0.76
40–49%	32/146	1.30 (0.63–2.67); P = 0.48	22/146	0.86 (0.36–2.03); P = 0.73	34/143	1.15 (0.57–2.32); P = 0.69	46/143	1.06 (0.58–1.94); P = 0.84
≥50%	8/73	0.86 (0.19–3.94); P = 0.85	4/73	1.00 (0.10–9.91); P = 1.00	26/73	1.33 (0.56–3.16); P = 0.52	27/73	1.17 (0.51–2.71); P = 0.71

CI, confidence interval; HR, hazard ratio (adjusted for baseline characteristics and stratified by trial); n, number of individuals.

Table 3. Absolute mortality difference and observed change in left ventricular ejection fraction
Classification	'Reduced' LVEF				'Mid–range' LVEF	'Preserved' LVEF
LVEF at baseline	<20%	20–25%	26–34%	35–39%	40–49%	≥50%
Sinus rhythm: all aetiology^a
Change in absolute mortality; beta-blockers vs. placebo (95% CI)^b	n = 2552	n = 3885	n = 5076	n = 1929	n = 575	n = 244
	–6.9% (–10.3% to –3.5%)	–3.9% (–6.3% to –1.6%)	–3.2% (–5.1% to –1.4%)	–3.4% (–6.1% to –0.7%)	–5.2% (–10.0% to –0.3%)	n = 244
	–6.9% (–10.3% to –3.5%)	–3.9% (–6.3% to –1.6%)	–3.2% (–5.1% to –1.4%)	–3.4% (–6.1% to –0.7%)	–5.2% (–10.0% to –0.3%)	+2.3% (–5.3% to + 9.9%)
Change in LVEF from baseline to follow-up; mean difference (SE) beta-blockers vs. placebo^c	n = 1106	n = 1068	n = 1600	n = 375	n = 251	n = 201
	+4.7% (0.5%)	+4.0% (0.5%)	+4.2% (0.5%)	+4.9% (0.9%)	+1.9% (1.1%)	+0.1% (1.2%)
Atrial fibrillation: all aetiology
Change in absolute mortality; beta-blockers vs. placebo (95% CI)^a	n = 494	n = 867	n = 1101	n = 367	n = 146	n = 73
	+2.8% (–5.3% to + 10.9%)	–4.1% (–9.3% to + 1.1%)	–0.8% (–5.5% to + 3.9%)	–3.2% (–10.7% to + 4.3%)	+3.2% (–10.4% to + 16.7%)	+0.3% (–14.0% to + 14.6%)
Change in LVEF from baseline to follow-up; mean difference (SE) beta-blockers vs. placebo^b	n = 177	n = 200	n = 369	n = 98	n = 93	n = 59
	+4.6% (1.7%)	+3.4% (1.2%)	+1.5% (1.0%)	+0.1% (1.9%)	+4.8% (1.9%)	–2.2% (3.0%)

CI, confidence interval; LVEF, left ventricular ejection fraction; SE, standard error of the mean difference; IQR, interquartile range.
^aSee Supplementary material online, Table S6 for data according to ischaemic/non-ischaemic aetiology in sinus rhythm.
^bMedian follow-up of 1.3 years (IQR 0.8–1.9).
^cMedian 1.0 years after baseline assessment (IQR 0.3–2.0).

Supplementary Table S1. Baseline characteristics for patients in atrial fibrillation
Characteristic	Left ventricular ejection fraction at baseline
Characteristic	<20% N = 494	20–25% N = 868	26–34% N = 1101	35–39% N = 368	40–49% N = 146	≥50% N = 73
LVEF, median (IQR)	0.16 (0.14 -0.18)	0.23 (0.21 -0.24)	0.30 (0.28 -0.32)	0.35 (0.35 -0.37)	0.41 (0.40 -0.45)	0.56 (0.52 -0.64)
Age, median years (IQR)	66 (59 - 73)	67 (59 - 73)	69 (61 - 75)	70 (61 - 74)	75 (71 - 79)	76 (74 - 79)
Women, n (%)	63 (12.8%)	128 (14.7%)	204 (18.5%)	86 (23.4%)	72 (49.3%)	39 (53.4%)
Years with HF diagnosis, median (IQR)	5 (2 - 8)	4 (2 - 8)	3 (1 - 6)	3 (1 - 6)	3 (1 - 5)	1 (0 - 4)
Ischaemic HF aetiology, n (%)	267 (54.0%)	443 (51.0%)	581 (52.8%)	233 (63.3%)	104 (71.2%)	52 (71.2%)
Prior myocardial infarction, n (%)	218 (44.3%)	331 (38.1%)	417 (38.2%)	168 (46.0%)	43 (29.5%)	21 (28.8%)
Diabetes Mellitus, n (%)	108 (24.2%)	200 (23.9%)	243 (22.8%)	72 (20.7%)	32 (22.7%)	19 (26.0%)
NYHA class III/IV, n (%)	434 (88.0%)	765 (88.2%)	819 (74.7%)	209 (57.3%)	56 (38.6%)	31 (42.5%)
Heart rate, median bpm (IQR)	84 (73 - 95)	81 (72 - 92)	80 (72 - 91)	80 (72 - 89)	82 (73 - 92)	84 (78 - 96)
Systolic BP, median mmHg (IQR)	120 (107 - 130)	122 (110 - 138)	130 (118 - 143)	130 (120 - 145)	142 (128 - 155)	140 (130 - 153)
Diastolic BP, median mmHg (IQR)	73 (65–80)	78 (70–82)	80 (70–87)	80 (72–87)	80 (75–90)	80 (79–89)
Body mass index, median kg/m²(IQR)	27 (24–32)	27 (24–31)	28 (25–31)	28 (25–31)	26 (24–29)	28 (26–31)
Estimated GFR, median mL/min (IQR)	57 (45–69)	57 (46–70)	64 (52–77)	63 (50–75)	62 (49–76)	60 (49–78)
Any diuretic therapy, n (%)	480 (97.2%)	833 (96.0%)	1028 (93.4%)	312 (84.8%)	131 (89.7%)	68 (93.2%)
ACEi or ARB, n (%)	464 (93.9%)	833 (96.0%)	1036 (94.1%)	352 (95.7%)	133 (91.1%)	64 (87.7%)
Aldosterone antagonists, n (%)	78 (16.7%)	176 (20.7%)	136 (12.6%)	62 (17.6%)	27 (18.8%)	16 (21.9%)
Digoxin, n (%)	435 (88.1%)	749 (86.3%)	920 (83.6%)	272 (73.9%)	117 (80.1%)	52 (71.2%)

ACEi, angiotensin converting enzyme inhibitor; ARB, angiotensin receptor blocker; BP, blood pressure; bpm, beats/minute; GFR, glomerular filtration rate; HF, heart failure; IQR, interquartile range; LVEF, left-ventricular ejection fraction; NYHA, New York Heart Association. Missing data report: n=319 for years with HF diagnosis; n=13 for prior myocardial infarction; n=135 for diabetes mellitus; n=431 for NYHA class; n= 2 for diastolic BP; n=8 heart rate; n= 22 for body mass index; n=101 for GFR; n=85 for any diuretic therapy; n=85 for aldosterone antagonists.

Supplementary Table S2. Baseline characteristics according to randomised treatment allocation in sinus rhythm
Characteristic	LVEF <40%		LVEF 40–49%		LVEF ≥50%
Characteristic	PLACEBO N=6582	BETA-BLOCKER N=6861	PLACEBO N=283	BETA-BLOCKER N=292	PLACEBO N=121	BETA-BLOCKER N=123
LVEF, median (IQR)	0.26 (0.20–0.32)	0.26 (0.20–0.32)	0.40 (0.40–0.44)	0.40 (0.40–0.43)	0.58 (0.52–0.65)	0.58 (0.54–0.65)
Age, median years(IQR)	63 (54-71)	63 (54-71)	72 (61-75)	70 (60-74)	75 (72-78)	75 (71-78)
Women, %	23.6%	24.0%	33.6%	35.3%	47.9%	57.7%
Years with HF diagnosis, median (IQR)	3 (1-6)	3 (1-6)	2 (1-5)	2 (1-5)	2 (1-5)	2 (0-6)
Ischaemic HF aetiology, %	67.7%	67.5%	92.6%	89.0%	86.8%	84.6%
Prior myocardial infarction, %	58.0%	58.2%	72.3%	71.2%	39.7%	32.5%
Diabetes Mellitus, %	24.3%	24.6%	23.0%	25.1%	30.6%	27.6%
NYHA class III/IV, %	68.1%	67.7%	21.6%	26.5%	24.0%	29.3%
Heart rate, median bpm (IQR)	80 (72-88)	80 (72-88)	75 (68-82)	76 (70-82)	75 (68-84)	74 (68-80)
Systolic BP, median mmHg (IQR)	122 (110-138)	122 (110-138)	132 (120-148)	130 (120-145)	146 (132-161)	147 (132-160)
Diastolic BP, median mmHg (IQR)	77 (70-82)	77 (70-82)	80 (70-85)	80 (70-84)	80 (79-90)	82 (77-90)
Body mass index, median kg/m²(IQR)	27 (24-31)	27 (24-31)	27 (25-30)	27 (25-30)	26 (24-31)	27 (25-30)
Estimated GFR, median mL/min (IQR)	64 (52-78)	64 (51-77)	67 (53-78)	66 (53-78)	67 (54-81)	72 (56-85)
Any diuretic therapy, %	85.9%	86.1%	67.1%	63.7%	81.0%	82.1%
ACEi or ARB, %	95.2%	94.4%	90.1%	91.1%	87.6%	87.0%
Aldosterone antagonists, %	8.1%	8.3%	6.4%	5.3%	9.9%	13.9%
Digoxin, %	53.6%	55.5%	26.4%	24.7%	17.1%	25.2%

Supplementary Table S3. Baseline LVEF and hazard for all-cause and cardiovascular mortality
	All-cause mortality		Cardiovascular death
	N (events/patients	HR, 95% CI; p-value	N (events/patients	HR, 95% CI; p-value
Sinus rhythm; per 5% lower LVEF at baseline	2,160/14,261	1.24, 1.21–1.28; p<0.0001	1,768/14,260	1.20, 1.22–1.30; p<0.0001
Atrial fibrillation; per 5% lower LVEF at baseline	609/3,034	1.09, 1.03–1.15; p=0.002	498/3,034	1.10, 1.05–1.18; p<0.0001

Adjusted hazard ratio (HR) analysed using a one-stage Cox regression model, with studies as strata. See also Supplementary Figure B.

Supplementary Table S4. Mode of death by baseline LVEF category in sinus rhythm
Baseline LVEF	<20%		20–25%		26–34%		35–39%		40–49%		≥50%
Randomised allocation	PLC	BB	PLC	BB	PLC	BB	PLC	BB	PLC	BB	PLC	BB
Ischaemic cardiomyopathy
All-cause mortality*	229/727 (31.5%)	180/756 (23.8%)	254/1258 (20.2%)	208/1314 (15.8%)	248/1697 (14.6%)	214/1778 (12.0%)	97/776 (12.5%)	62/786 (7.9%)	29/262 (11.1%)	19/260 (7.3%)	9/105 (8.6%)	13/104(12.5%)
CV death	201/727 (27.6%)	156/756 (20.6%)	220/1258 (17.5%)	181/1314 (13.8%)	207/1697 (12.2%)	172/1778 (9.7%)	78/776 (10.1%)	53/786 (6.7%)	22/262 (8.4%)	12/260 (4.6%)	6/105 (5.7%)	9/104 (8.7%)
Sudden death	103/727 (14.2%)	83/756 (11.0%)	117/1258 (9.3%)	86/1314 (6.5%)	121/1697 (7.1%)	88/1778 (4.9%)	36/776 (4.6%)	29/786 (3.7%)	10/262 (3.8%)	4/260 (1.5%)	2/105 (1.9%)	3/104 (2.9%)
HF-related death	73/727(10.0%)	45/756 (6.0%)	66/1258 (5.2%)	63/1314 (4.8%)	52/1697 (3.1%)	48/1778 (2.7%)	20/776 (2.6%)	9/786 (1.1%)	6/262 (2.3%)	3/260 (1.2%)	1/105 (1.0%)	1/104 (1.0%)
Non-CV death	10/727 (1.4%)	18/756 (2.4%)	16/1258 (1.3%)	14/1314 (1.1%)	21/1697 (1.2%)	19/1778 (1.1%)	9/776 (1.2%)	4/786 (0.5%)	4/262 (1.5%)	3/260 (1.2%)	1/105 (1.0%)	1/104 (1.0%)
Non-ischaemic cardiomyopathy
All-cause mortality*	136/508 (26.8%)	118/561 (21.0%)	99/644 (15.4%)	82/669 (12.3%)	113/782 (14.5%)	81/819 (9.9%)	16/189 (8.5%)	18/178 (10.1%)	6/21 (28.6%)	2/32 (6.3%)	2/16 (12.5%)	1/19 (5.3%)
CV death	99/508 (19.5%)	80/561 (14.3%)	72/644 (11.2%)	61/669 (9.1%)	83/782 (10.6%)	54/819 (6.6%)	11/189 (5.8%)	15/178 (8.4%)	4/21 (19.0%)	1/32 (3.1%)	1/16 (6.3%)	0/19 (0.0%)
Sudden death	48/508 (9.4%)	39/561 (7.0%)	38/644 (5.9%)	27/669 (4.0%)	48/782 (6.1%)	27/819 (3.3%)	9/189 (4.8%)	5/178 (2.8%)	2/21 (9.5%)	1/32 (3.1%)	1/16 (6.3%)	0/19 (0.0%)
HF-related death	38/508 (7.5%)	28/561 (5.0%)	26/644 (4.0%)	17/669 (2.5%)	22/782 (2.8%)	15/819 (1.8%)	2/189 (1.1%)	6/178 (3.4%)	1/21 (4.8%)	0/32 (0.0%)	0/16 (0.0%)	0/19 (0.0%)
Non-CV death	8/508 (1.6%)	9/561 (1.6%)	8/644 (1.2%)	7/669 (1.0%)	11/782 (1.4%)	13/819 (1.6%)	3/189 (1.6%)	1/178 (0.6%)	0/21 (0.0%)	0/32 (0.0%)	0/16 (0.0%)	1/19 (5.3%)

BB, beta-blockers; CV, cardiovascular; LVEF, left-ventricular ejection fraction; PLC, placebo. * Includes deaths due to an unknown cause. Note that some deaths were ascribed to unknown causes and therefore are attributed neither to cardiovascular or non-cardiovascular deaths.

Supplementary Table S5. Mode of death by baseline LVEF category in atrial fibrillation
Baseline LVEF	<20%		20–25%		26–34%		35–39%		40–49%		≥50%
Randomised allocation	PLC	BB	PLC	BB	PLC	BB	PLC	BB	PLC	BB	PLC	BB
Ischaemic cardiomyopathy
All-cause mortality*	47/141 (33.3%)	47/126 (37.3%)	49/224 (21.9%)	44/219 (20.1%)	60/302 (19.9%)	72/279 (25.8%)	29/120 (24.2%)	18/113 (15.9%)	9/50 (18.0%)	14/54 (25.9%)	3/26 (11.5%)	3/26 (11.5%)
CV death	45/141 (31.9%)	39/126 (31.0%)	41/224 (18.3%)	37/219 (16.9%)	51/302 (16.9%)	60/279 (21.5%)	26/120 (21.7%)	14/113 (12.4%)	7/50 (14.0%)	9/54 (16.7%)	1/26 (3.8%)	2/26 (7.7%)
Sudden death	17/141 (12.1%)	16/126 (12.7%)	22/224 (9.8%)	18/219 (8.2%)	19/302 (6.3%)	28/279 (10.0%)	15/120 (12.5%)	7/113 (6.2%)	1/50 (2.0%)	3/54 (5.6%)	0/26 (0.0%)	0/26 (0.0%)
HF-related death	25/141 (17.7%)	21/126 (16.7%)	13/224 (5.8%)	11/219 (5.0%)	24/302 (7.9%)	18/279 (6.5%)	6/120 (5.0%)	5/113 (4.4%)	2/50 (4.0%)	1/54 (1.9%)	0/26 (0.0%)	0/26 (0.0%)
Non-CV death	2/141 (1.4%)	5/126 (4.0%)	4/224 (1.8%)	5/219 (2.3%)	4/302 (1.3%)	4/279 (1.4%)	0/120 (0.0%)	2/113 (1.8%)	1/50 (2.0%)	2/54 (3.7%)	1/26 (3.8%)	1/26 (3.8%)
Non-ischaemic cardiomyopathy
All-cause mortality*	24/106 (22.6%)	31/121 (25.6%)	42/213 (19.7%)	28/211 (13.3%)	50/255 (19.6%)	31/265 (11.7%)	5/70 (7.1%)	8/64 (12.5%)	5/17 (29.4%)	5/25 (20.0%)	1/11 (9.1%)	1/10 (10.0%)
CV death	18/106 (17.0%)	24/121 (19.8%)	34/213 (16.0%)	25/211 (11.8%)	34/255 (13.3%)	22/265 (8.3%)	4/70 (5.7%)	3/64 (4.7%)	4/17 (23.5%)	2/25 (8.0%)	1/11 (9.1%)	0/10 (0.0%)
Sudden death	11/106 (10.4%)	7/121 (5.8%)	18/213 (8.5%)	16/211 (7.6%)	20/255 (7.8%)	8/265 (3.0%)	2/70 (2.9%)	1/64 (1.6%)	1/17 (5.9%)	0/25 (0.0%)	1/11 (9.1%)	0/10 (0.0%)
HF-related death	6/106 (5.7%)	14/121 (11.6%)	9/213 (4.2%)	6/211 (2.8%)	7/255 (2.7%)	8/265 (3.0%)	2/70 (2.9%)	1/64 (1.6%)	2/17 (11.8%)	2/25 (8.0%)	0/11 (0.0%)	0/10 (0.0%)
Non-CV death	0/106 (0.0%)	0/121 (0.0%)	2/213 (0.9%)	1/211 (0.5%)	4/255 (1.6%)	3/265 (1.1%)	0/70 (0.0%)	1/64 (1.6%)	0/17 (0.0%)	2/25 (8.0%)	0/11 (0.0%)	1/10 (10.0%)

Supplementary Table S6. Absolute mortality difference and observed change in LVEF according to aetiology in sinus rhythm
Classification	'Reduced' LVEF				'Mid-range' LVEF	'Preserved' LVEF
LVEF at baseline	<20%	20–25%	26–34%	35–39%	40–49%	≥50%
Sinus rhythm: Ischaemic aetiology
Change in absolute mortality;beta-blockers vs placebo (95% CI)†	n=1483 -7.7% (-12.2% to -3.1%)	n=2572 -4.4% (-7.3% to -1.4%)	n=3475 -2.6% (-4.8% to -0.3%)	n=1562 -4.6% (-7.6% to -1.6%)	n=522 -3.8% (-8.7% to +1.2%)	n=209 +3.9% (-4.4% to +12.2%)
Change in LVEF from baseline to follow-up; mean difference (SE)beta-blockers vs placebo‡	n=593 +3.1% (0.6%)	n=667 +3.3% (0.6%)	n=1070 +3.0% (0.5%)	n=277 +4.4% (1.0%)	n=227 +2.5% (1.2%)	n=177 +0.6% (1.3%)
Sinus rhythm: Non-ischaemic aetiology
Change in absolute mortality;beta-blockers vs placebo (95% CI)†	n=1069 -5.7% (-10.9% to -0.6%)	n=1313 -3.1% (-6.8% to +0.6%)	n=1601 -4.6% (-7.8% to -1.4%)	n=367 +1.6% (-4.3% to +7.6%)	n=53 -22.3% (-43.4% to -1.3%)	n=35 -7.2% (-26.3% to +11.8%)
Change in LVEF from baseline to follow-up; mean difference (SE) beta-blockers vs placebo‡	n=513 +6.2% (0.9%)	n=401 +5.6% (1.0%)	n=530 +6.2% (0.9%)	n=98 +6.3% (2.0%)	n=24 -4.2% (4.3%)	n=24 -4.4% (4.5%)

^†Median follow-up of 1.3 years (IQR 0.8–1.9)
^‡Median 1.0 years after baseline assessment (IQR 0.3–2.0) CI = confidence interval; LVEF = left ventricular ejection fraction; SE = standard error of the mean difference.

Authors and Disclosures

John G.F. Cleland¹, Karina V. Bunting², Marcus D. Flather³, Douglas G. Altman⁴, Jane Holmes⁴, Andrew J.S. Coats⁵, Luis Manzano⁶, John J.V. McMurray⁷, Frank Ruschitzka⁸, Dirk J. van Veldhuisen⁹, Thomas G. von Lueder^10,11, Michael Böhm¹², Bert Andersson¹³, John Kjekshus¹⁴, Milton Packer¹⁵, Alan S. Rigby¹⁶, Giuseppe Rosano^17,18, HansWedel¹⁹, A° ke Hjalmarson¹³, John Wikstrand²⁰ and Dipak Kotecha^2,11* on behalf of the Beta-blockers in Heart Failure Collaborative Group

¹Robertson Institute of Biostatistics and Clinical Trials Unit, University of Glasgow, University Avenue, Glasgow G12 8QQ, UK; ²Institute of Cardiovascular Sciences, University of Birmingham, Vincent Drive, Birmingham B15 2TT, UK; ³Norwich Medical School, Faculty of Medicine and Health Science, University of East Anglia, Norwich NR4 7TJ, UK; ⁴Centre for Statistics in Medicine, Nuffield Department of Orthopaedics, Rheumatology and Musculoskeletal Sciences, University of Oxford, Oxford OX1 2JD, UK; ⁵San Raffaele Pisana Scientific Institute, Via della Pisana, 235, 00163 Rome, Italy; ⁶Internal Medicine Department, Hospital Universitario Ramón y Cajal, Universidad de Alcalá (IRYCIS), Plaza de San Diego, 28801 Alcalá de Henares, Madrid, Spain; ⁷Institute of Cardiovascular and Medical Sciences, University of Glasgow, University Avenue, Glasgow G12 8QQ, UK; ⁸Klinik für Kardiologie, UniversitätsSpital Zürich, Universitätstrasse 8, 8006 Zürich, Switzerland; ⁹Department of Cardiology, University Medical Centre Groningen, University of Groningen, PO box 30.001 9700 RB Groningen, The Netherlands; ¹⁰Department of Cardiology, Oslo University Hospital, PO Box 4950 Nydalen N-0424 Oslo, Norway; ¹¹Centre of Cardiovascular Research and Education in Therapeutics, Monash University, 99 Commercial Road, Melbourne, Victoria 3004, Australia; ¹²Kardiologie, Angiologie und internistische Intensivmedizin, Universitätsklinikum des Saarlandes, Kirrberger Str. 100, 66421 Homburg/Saar, Germany; ¹³Department of Cardiology, Sahlgrenska University Hospital and Gothenburg University, Bla° stra°ket 5, 413 45 Gothenburg, Sweden; ¹⁴Rikshospitalet University Hospital and Faculty of Medicine, University of Oslo, Problemveien 7, 0315 Oslo, Norway; ¹⁵Baylor Heart and Vascular Institute, Baylor University Medical Center, 621 Hall St, Dallas TX 75226, USA; ¹⁶Hull York Medical School, Faculty of Health Sciences, University of Hull, Kingston-upon-Hull, HU6 7RX, UK; ¹⁷Cardiovascular and Cell Science Institute, St George's University of London, SW17 0RE, UK; ¹⁸Department of Medical Sciences, IRCCS San Raffaele Pisana, Via della Pisana, 235, 00163 Roma, Italy; ¹⁹Health Metrics, Sahlgrenska Academy, University of Gothenburg, Box 100, S-405 30 Gothenburg, Sweden; and ²⁰Wallenberg Laboratory for Cardiovascular Research, Sahlgrenska Academy, Gothenburg University, Bruna Stra°ket 16, 413 45 Gothenburg, Sweden

*Corresponding author
Tel: +44 121 371 8122, Fax: +44 121 371 4175, Email: d.kotecha@bham.ac.uk

Conflict of interest
All authors have completed the ICMJE uniform disclosure form (www.icmje.org/coi_disclosure.pdf) and declare: J.G.F.C. reports grants and personal fees from Amgen, grants and personal fees from Novartis and Stealth Biotherapeutics, non-financial support from Medtronic and Boston Scientific, all outside the submitted work. K.V.B. has nothing to disclose. M.D.F. reports grants from Novarts and personal fees from AstraZeneca, all outside the submitted work. D.G.A. has nothing to disclose. J.H. has nothing to disclose. A.J.S.C. reports grants and personal fees from Menarini, during the conduct of the study and personal fees from Servier, Lone Star, Vifor and Respicardia, all outside the submitted work. L.M. has nothing to disclose. J.J.V.M. reports payments for trial-related activities to the University of Glasgow from Novartis, Cardiorentis, Amgen, Oxford University/Bayer, GlaxoSmithKline, Theracos, Abbvie, DalCor, Pfizer, Merck, AstraZeneca, Bristol Myers Squibb, and Kidney Research UK (KRUK)/Kings College Hospital, London/Vifor-Fresenius Pharma, all outside the submitted work. F.R. reports grants or personal fees from SJM, Servier, Zoll, AstraZeneca, Sanofi, Cardiorentis, Amgen, BMS, Pfizer, Fresenius, Vifor, Roche, Bayer, Abbott and Novartis, all outside the submitted work. D.J.V. has nothing to disclose. T.G.L. reports personal fees from Novartis, St Jude Medical, and Vifor, all outside the submitted work. M.B. reports personal fees from Boehringer-Ingelheim, Medtronic, Servier, Abbot, Astra-Zeneca and BMS, all outside the submitted work. B.A. has nothing to disclose. J.K. has nothing to disclose. M.P. reports personal fees from Amgen, Boehringer Ingelheim, Cardiorentis and Sanofi, all outside the submitted work. A.S.R. has nothing to disclose. G.R. has nothing to disclose. H.W. reports personal fees from AstraZeneca during the conduct of the study. A.H. has nothing to disclose. J.W. reports grants and personal fees from AstraZeneca Sweden, during the conduct of the study; and previously a Medical Adviser on cardiovascular research for AstraZeneca, Sweden. D.K. reports grants from Menarini, during the conduct of the study (unrestricted grant for administration of the Beta-Blockers in Heart Failure Collaborative Group); professional development support from Daiichi Sankyo and personal fees from Atricure, both outside the submitted work; and Chief Investigator of the RAte control Therapy Evaluation in permanent Atrial Fibrillation trial (RATE-AF; NCT02391337).

Authors' contributions
J.G.F.C., K.V.B., and D.K. drafted the manuscript. D.K. also participated in the design of the study, leads the collaborative group and performed data management and statistical analysis. J.H. and D.G.A. performed independent statistical analyses and also manuscript revision. M.D.F. participated in the design and coordination of the study, and manuscript revision. All other named authors read, revised, and approved the final manuscript.

Comments

processing....

Beta-blockers for Heart Failure With Reduced, Mid-range, and Preserved Ejection Fraction

An Individual Patient-level Analysis of Double-blind Randomized Trials

Abstract and Introduction

Abstract

Introduction

Comments

Most Popular Articles

What to Read Next on Medscape

About

Membership

Apps

WebMD Network

Editions

Beta-blockers for Heart Failure With Reduced, Mid-range, and Preserved Ejection Fraction

An Individual Patient-level Analysis of Double-blind Randomized Trials

Abstract and Introduction

Abstract

Introduction

Tables

Tables

Tables

Tables

Tables

Tables

Tables

Tables

Tables

References

Authors and Disclosures

Authors and Disclosures

Comments

Most Popular Articles

What to Read Next on Medscape

More from This Journal

Latest News & Perspective

Recommended Reading

Drugs Related to Heart Failure (HF)

Most Popular Articles

Recommendations

Email This

Find Us On

About

Membership

Apps

WebMD Network

Editions