Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection fraction spectrum

IJCA-27965; No of Pages 7 International Journal of Cardiology xxx (xxxx) xxx

Contents lists available at ScienceDirect

International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard

Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection fraction spectrum Gianluigi Savarese a,1, Åsa Jonsson b,1, Ann-Charlotte Hallberg c, Ulf Dahlström d, Magnus Edner a, Lars H. Lund a,e,⁎ a

Cardiology Unit, Department of Medicine, Karolinska Institutet, Stockholm, Sweden Department of Medicine, Division of Cardiology, County Hospital Ryhov, Jönköping, Sweden Department of Computer and Information Science, Linköping University, Linköping, Sweden d Department of Cardiology and Department of Medical and Health Sciences, Linköping University, Linköping, Sweden e Heart and Vascular Theme, Karolinska University Hospital, Stockholm, Sweden b c

a r t i c l e

i n f o

Article history: Received 13 March 2019 Received in revised form 11 July 2019 Accepted 26 August 2019 Available online xxxx Keywords: Heart failure Mid-range ejection fraction Preserved ejection fraction Anemia Outcomes Registry

a b s t r a c t Background: The role of anemia in heart failure with mid-range and preserved ejection fraction (HFmrEF, EF 40–49% and HFpEF, EF ≥50%) is unknown. We aimed to compare prevalence of, associations with, and prognostic role of anemia in HF across the EF spectrum. Methods: In patients from the Swedish HF Registry, we assessed the associations between clinical characteristics and anemia (hemoglobin b120 g/L in women and b130 g/L in men) by multivariable logistic regression, and between anemia, composite of all-cause death and HF hospitalization and all-cause death alone by multivariable Cox regression. Results: Of 49,985 patients with HF (anemia = 34%), 23% had HFpEF (anemia = 41%), 21% had HFmrEF (anemia = 35%) and 55% had HFpEF (anemia = 32%). Higher EF was independently associated with higher likelihood of concomitant anemia. Important predictors of anemia across the EF spectrum were male sex, older age, worse New York Heart Association class and renal function, lower systolic blood pressure, higher N-Terminal B-type natriuretic peptides levels, diabetes, valvular disease and in-patient status. Anemia had adjusted hazard ratios (95% CI) for mortality or HF hospitalization 1.24 (1.18–1.30) in HFpEF, 1.26 (1.19–1.34) in HFmrEF and 1.14 (1.10–1.19) in HFrEF; pinteractionEF = 0.003; and for mortality 1.28 (1.20–1.36) in HFpEF, 1.21 (1.13–1.29) in HFmrEF, and 1.30 (1.24–1.35) in HFrEF; pinteractionEF = 0.22. Conclusions: In this nation-wide registry, prevalence of anemia was higher in HFpEF vs. HFmrEF vs. HFrEF, but was associated with a similarly increased risk of death across the EF spectrum, with greater risk of death or HF hospitalization in HFpEF and HFmrEF vs. HFrEF. © 2019 Elsevier B.V. All rights reserved.

1. Introduction Heart failure (HF) affects 2–3% of the population and up to 20% of the elderly [1]. About half of the HF population has HF with preserved ejection fraction (HFpEF) [2]. It remains controversial whether HFpEF is a distinct syndrome or merely a representation of ageing and associated comorbidities, or more likely, a collection of heterogeneous phenotypes comprising overlapping subgroups [2]. The EF cut-offs for HFpEF are ⁎ Corresponding author at: Department of Medicine, Cardiology Unit, Karolinska Institutet, S1:02, 171 76 Stockholm, Sweden. E-mail address: [email protected] (L.H. Lund). 1 These Authors equally contributed as first author.

also controversial. The 2016 European Society of Cardiology (ESC) HF guidelines proposed a new HF phenotype, namely HF with mid-range EF (HFmrEF), to foster the characterization of patients with EF 40–49% [3]. Comorbidities are increasingly recognized as potential drivers of and confounders in understanding symptoms and outcomes in HFpEF [2,4,5]. Anemia is common and associated with poor outcomes in HFrEF [5,6]. In HFpEF, the prevalence of anemia ranges widely 19%– 68% in trials, observational cohorts and registries, and appears independently associated with mortality [2,7] and HF hospitalization [2]. However, a comprehensive assessment of the independent predictors of anemia, and of the independent association between anemia and outcomes in HFpEF vs. HFmrEF vs. HFrEF is lacking. Therefore, we assessed

https://doi.org/10.1016/j.ijcard.2019.08.049 0167-5273/© 2019 Elsevier B.V. All rights reserved.

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

2

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

(1) the prevalence of, (2) the independent predictors of, and (3) the prognostic impact of anemia in HFpEF vs. HFmrEF vs. HFrEF. 2. Methods 2.1. Study population and data The Swedish Heart Failure Registry (SwedeHF; www.SwedeHF.se) has been previously described [8,9] and provided the study population and baseline data. The only inclusion criteria are age ≥18 and clinician-judged HF, regardless of EF. Approximately 80 variables are recorded at discharge from hospital or after out-patient clinical visits and entered into a web-based case report form and database managed by Uppsala Clinical Research Center (www.UCR.se). The protocol, case report form and annual reports are available at www.SwedeHF.se. Patients registered in SwedeHF between 11 May 2000 and 31 December 2012 were considered for the current analysis. The index date was defined as either the date of an out-patient visit or hospital discharge (for in-patients). Exclusion criteria for this study were death during hospitalization which prompted the registration in SwedeHF, or EF or hemoglobin missing. If the same patient was registered more than once, the first registration without missing values for inclusion/exclusion criteria was selected (Supplement Fig. 1). The establishment of the registry, and registration and analysis of data for this study was approved by a multisite Ethics Committee and conform to the declaration of Helsinki. Individual patient consent was not required, but the patients were informed of entry into national registries and allowed to opt out. 2.2. Definitions of HFmrEF, HFpEF, anemia and renal function In SwedeHF, EF is categorized as b30% and 30%–39% (defined as HFrEF), 40%–49% (HFmrEF), and ≥50% (HFpEF). Anemia was defined according to the World Health Organization (WHO) as hemoglobin b120 g/L in women and b130 g/L in men. For in-patients, the closest hemoglobin value to the discharge (index data) was selected. For outpatients, the closest value to the date of the out-patient visit (index data) was considered. Renal function was assessed as estimated glomerular filtration rate (eGFR) and calculated by the MDRD equation. 2.3. Outcomes End of follow-up was 31 December 2012. Outcomes were time to allcause mortality or first HF hospitalization, with all-cause death obtained from the Causes of Death registry, and HF hospitalization obtained from the National Patient Registry (NPR) and defined as ICD-10 diagnoses I50, I42-I43, I25.5, K761, I11.0, I130, I132, J81 in the first position (main diagnosis). In the NPR, an HF diagnosis has previously been validated and verified in 86–91% of cases [10]. 2.4. Statistical analyses Baseline characteristics of patients according to the presence of concomitant anemia were compared by using Kruskal-Wallis test for continuous variables and by Chi-square for categorical variables (Table 1). For multivariable analyses, missing data were handled by multiple imputation using chained equations (10 imputed datasets generated, “MI impute chained” command on STATA was applied). Multiple imputation was separately executed in out-patients vs. in-patients with HFpEF, HFmrEF and HFrEF to allow the pre-specified investigation of interactions with HF type and caregiver. To assess the independent associations between baseline characteristics and prevalent anemia, multivariable logistic regressions were performed with anemia as the dependent variable (Table 2; Supplement Tables 1–3). All the variables included in the multivariable models and multiple imputation models are marked with $ in Table 2.

Crude outcomes (time to the occurrence of the composite of all-cause death and HF hospitalization, and time to all-cause death alone) were assessed by Kaplan-Meier analysis and compared in patients with vs. without anemia across the EF spectrum by a log-rank test. The crude and independent associations between anemia and outcomes were assessed by univariable and multivariable Cox regressions, including the same baseline variables as in the logistic regression analyses plus anemia itself as independent variables (Fig. 1; Table 3–4). Interactions between anemia and EF category as well as several other pre-specified clinically relevant variables were investigated in all the models. The proportional-hazards assumption was tested on the basis of Schoenfeld residuals and met. Statistical analyses were performed by Stata 14.2 (StataCorp LLC, College Station, TX, USA). A p-value b0.05 was considered as statistically significant. 3. Results 3.1. Patients (Supplement Fig. 1) Between 11 May 2000 and 31 December 2012, SwedeHF included 80,772 registrations. After applying exclusion criteria, 42,985 patients (23% with HFpEF, 21% with HFmrEF, and 55% with HFrEF) were considered for the current analysis and, of these, 34% had anemia. Prevalence of anemia was 41% in HFpEF, 35% in HFmrEF, and 32% in HFrEF (p b 0.001). As many as 24,707 (57%) patients were in-patients and 18,278 (43%) were out-patients. Prevalence of anemia in in-patients was 41%, with 47% prevalence in HFpEF, 43% in HFmrEF, and 38% in HFrEF (p b 0.001), whereas it was 25% in out-patients, with 28% prevalence in HFpEF, 25% in HFmrEF, and 24% in HFrEF (p b 0.001). 3.2. Baseline characteristics (Table 1) Baseline characteristics of patients with vs. without anemia are shown in Table 1. In the overall cohort, patients with anemia were more likely to have HFpEF vs. HFmrEF vs. HFrEF, were older, and more commonly male. Regardless of EF, patients with anemia also had characteristics reflecting more severe HF (i.e. more likely to be in-patients, higher NYHA class, higher NT-proBNP levels, longer HF duration, lower systolic blood pressure, more use of diuretics) and/or comorbidity (i.e. higher prevalence of chronic kidney disease, diabetes, ischemic heart disease, hypertension, atrial fibrillation, peripheral artery disease, history of stroke, valvular disease, pulmonary disease) and frailty (less obesity), less use of angiotensin converting enzyme inhibitors (ACEI)/angiotensin receptor blockers (ARB), beta blockers and mineralocorticoid receptor antagonists (in HFrEF). Patients with anemia were less likely to receive oral anticoagulants but more likely to receive antiplatelets, and they were more likely to be followed-up in primary vs specialty care. 3.3. Independent predictors of anemia (Table 2) After adjustments for several baseline characteristics, anemia was more likely to be observed in HFpEF (OR vs. HFrEF: 1.55; 95% CI: 1.46–1.65) vs. HFmrEF (OR vs. HFrEF: 1.24; 95% CI: 1.17–1.31) vs. HFrEF (ref). Important additional independent predictors of anemia across the EF spectrum were: higher age, male sex, lower systolic blood pressure, lower eGFR, diabetes, valvular disease and intervention, no use of betablockers, use of diuretics, and no use of oral anticoagulants. Being inpatient vs. out-patient was associated with higher likelihood of concomitant anemia in HFpEF vs. HFmrEF vs. HFrEF although it was a statistically significant predictor across all the phenotypes. Similarly, higher NTproBNP predicted concomitant anemia across the EF spectrum, but with significantly higher OR in HFmrEF. Being referral to HF nurse-led clinic was associated with less likelihood of concomitant anemia in HFrEF but was not a predictor in HFpEF and HFmrEF. Higher NYHA class, history of ischemic heart disease, coronary revascularization, and peripheral artery

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

3

Table 1 Baseline characteristics of patients with vs. without anemia across the EF spectrum. HFpEF

HFmrEF

HFrEF

Anemia No

Anemia Yes

p-Value Anemia No

Anemia Yes

5878 (59%)

4078 (41%)

6029 (65%)

3194 (35%)

2419 (41.2%) 3459 (58.8%) 76 (11)

2095 (51.4%) 1983 (48.6%) 79 (10)

3512 (58.3%) 2517 (41.7%) 73 (12)

2084 (65.2%) 1110 (34.8%) 77 (11)

2589 (50.2%) 2565 (49.8%)

1949 (50.9%) 1882 (49.1%)

0.55

2663 (47.9%) 2891 (52.1%)

1423 (46.8%) 1620 (53.2%)

3531 (60.1%) 2347 (39.9%)

3147 (77.2%) 931 (22.8%)

b0.001

2864 (47.5%) 3165 (52.5%)

2553 (46.7%) 2918 (53.3%) 1449 (26.5%)

1630 (43.5%) 2113 (56.5%) 892 (23.8%)

0.003

3364 (57.3%) 2506 (42.7%)

2393 (58.8%) 1680 (41.2%)

2917 (50.0%) 2913 (50.0%)

p-Value Anemia No

Anemia Yes

p-Value

16,299 (68%)

7507 (32%)

11,306 (69.4%) 4993 (30.6%) 70 (13)

5643 (75.2%) 1864 (24.8%) 75 (11)

b0.001

0.29

6875 (43.5%) 8912 (56.5%)

3192 (43.6%) 4130 (56.4%)

0.95

2130 (66.7%) 1064 (33.3%)

b0.001

8112 (49.8%) 8187 (50.2%)

4923 (65.6%) 2584 (34.4%)

b0.001

3532 (61.7%) 2197 (38.3%) 2179 (38.1%)

1640 (55.0%) 1340 (45.0%) 980 (32.9%)

b0.001

11,240 (73.0%) 4152 (27.0%) 7557 (49.2%)

4355 (61.9%) 2678 (38.1%) 2845 (40.5%)

b0.001

0.15

3016 (50.1%) 3007 (49.9%)

1677 (52.5%) 1515 (47.5%)

0.024

7872 (48.4%) 8380 (51.6%)

3744 (50.0%) 3743 (50.0%)

0.025

1834 (45.3%) 2213 (54.7%)

b0.001

3143 (52.3%) 2864 (47.7%)

1408 (44.4%) 1763 (55.6%)

b0.001

8798 (54.3%) 7403 (45.7%)

3322 (44.6%) 4126 (55.4%)

b0.001

708 (18.7%) 1717 (45.3%) 1254 (33.1%) 115 (3.0%) 130 (120, 148) 75 (67, 80) 72 (63, 81)

306 (12.3%) 1104 (44.2%) 976 (39.1%) 110 (4.4%) 130 (120, 146)

b0.001

260 (11.9%) 1049 (48.1%) 784 (35.9%) 88 (4.0%) 130 (115, 140)

b0.001

310 (5.7%) 2045 (37.8%) 2665 (49.3%) 385 (7.1%) 120 (110, 135)

b0.001

0.008

1323 (10.3%) 6169 (48.2%) 4853 (37.9%) 456 (3.6%) 120 (110, 140)

b0.001

70 (60, 80) 72 (63, 82)

b0.001 0.50

747 (16.6%) 2483 (55.3%) 1193 (26.6%) 69 (1.5%) 130 (120, 142) 75 (68, 80) 70 (61, 80)

70 (60, 80) 71 (63, 80)

b0.001 0.001

75 (65, 80) 72 (63, 82)

70 (60, 80) 72 (64, 83)

b0.001 b0.001

61 (47, 76) 3028 (51.7%) 2525 (43.1%) 302 (5.2%) 137 (130, 146) 1641 (726, 3330)

51 (37, 67) 1408 (34.6%) 2047 (50.3%) 612 (15.0%) 113 (106, 118)

b0.001 b0.001

53 (39, 71) 1252 (39.3%) 1514 (47.5%) 421 (13.2%) 115 (107, 121)

b0.001 b0.001

65 (50, 80) 9576 (58.9%) 6029 (37.1%) 649 (4.0%) 142 (134, 151)

53 (38, 71) 2980 (39.8%) 3429 (45.8%) 1074 (14.4%) 117 (110, 123)

b0.001 b0.001

2832 (1333, 5720)

b0.001

64 (50, 79) 3441 (57.3%) 2292 (38.2%) 272 (4.5%) 140 (132, 149) 1698 (726, 3768)

3442 (1572, 7400)

b0.001

2557 (1140, 5549)

5039 (2294, 11,000)

b0.001

Treatments ACEi or ARB Mineralocorticoid receptor antagonist Digoxin Diuretic Nitrate Platelet inhibitor Oral anticoagulant Statin Beta blocker ICD/CRT

4331 (74.5%) 1471 (25.2%) 1159 (19.8%) 4762 (81.5%) 967 (16.6%) 2568 (44.0%) 2401 (41.1%) 2241 (38.3%) 4634 (79.4%) 70 (1.2%)

2728 (67.7%) 1129 (27.9%) 636 (15.7%) 3652 (90.0%) 839 (20.7%) 2030 (50.1%) 1407 (34.8%) 1571 (38.8%) 3118 (77.1%) 34 (0.8%)

b0.001 0.003 b0.001 b0.001 b0.001 b0.001 b0.001 0.63 0.007 0.079

5197 (86.7%) 1341 (22.4%) 1003 (16.7%) 4198 (70.0%) 869 (14.5%) 3031 (50.5%) 2452 (40.9%) 2888 (48.1%) 5204 (86.7%) 136 (2.3%)

2449 (77.2%) 826 (26.0%) 481 (15.1%) 2643 (83.1%) 654 (20.7%) 1792 (56.6%) 1057 (33.3%) 1527 (48.0%) 2661 (83.7%) 64 (2.0%)

b0.001 b0.001 0.046 b0.001 b0.001 b0.001 b0.001 0.94 b0.001 0.43

15,039 (92.7%) 5377 (33.2%) 3103 (19.1%) 12,441 (76.7%) 2147 (13.3%) 8005 (49.4%) 6770 (41.8%) 7560 (46.6%) 14,780 (91.0%) 965 (6.0%)

6332 (84.9%) 2377 (31.9%) 1085 (14.5%) 6459 (86.5%) 1600 (21.4%) 4366 (58.5%) 2502 (33.5%) 3741 (50.1%) 6616 (88.5%) 498 (6.7%)

b0.001 0.043 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.030

Comorbidities Current smoker Obesity (body mass index ≥ 30 kg/m2) Hypertension Diabetes Ischemic heart disease Coronary revascularization Peripheral artery disease Stroke/TIA Atrial fibrillation Valvular disease Valvular intervention Pulmonary disease

464 (10.6%) 799 (31.1%) 4121 (70.1%) 1465 (24.9%) 2475 (43.5%) 1163 (19.8%) 487 (8.3%) 1047 (17.8%) 3699 (62.9%) 1740 (30.5%) 342 (5.9%) 1664 (28.3%)

240 (8.4%) 470 (26.1%) 2934 (71.9%) 1351 (33.1%) 2019 (50.9%) 1033 (25.3%) 540 (13.2%) 921 (22.6%) 2646 (64.9%) 1525 (38.2%) 430 (10.7%) 1332 (32.7%)

0.003 b0.001 0.047 b0.001 b0.001 b0.001 b0.001 b0.001 0.046 b0.001 b0.001 b0.001

613 (12.8%) 796 (28.6%) 3705 (61.5%) 1385 (23.0%) 3119 (53.4%) 1816 (30.1%) 508 (8.4%) 941 (15.6%) 3461 (57.4%) 1269 (21.7%) 297 (5.0%) 1410 (23.4%)

234 (9.8%) 323 (22.8%) 2118 (66.3%) 1094 (34.3%) 2010 (64.3%) 1184 (37.1%) 445 (13.9%) 610 (19.1%) 1913 (59.9%) 1022 (32.9%) 352 (11.2%) 915 (28.6%)

b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.021 b0.001 b0.001 b0.001

2165 (16.3%) 1815 (23.5%) 8553 (52.5%) 3864 (23.7%) 8138 (52.3%) 4607 (28.3%) 1226 (7.5%) 2263 (13.9%) 8303 (50.9%) 3258 (20.5%) 697 (4.3%) 3465 (21.3%)

771 (13.5%) 524 (15.4%) 4367 (58.2%) 2536 (33.8%) 5114 (69.9%) 3072 (40.9%) 1102 (14.7%) 1470 (19.6%) 3913 (52.1%) 2179 (29.8%) 586 (7.9%) 2098 (27.9%)

b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 b0.001 0.090 b0.001 b0.001 b0.001

Demographics/organizational/social Sex Male Female Age, years, mean (SD) Specialty Internal Medicine/Geriatrics Cardiology Caregiver In-patient Out-patient Follow-up referral specialty Specialty care Primary care/other Follow-up referral to out-patient HF nurse clinic Family type Living alone Married/cohabitating Clinical Duration of HF b6 months ≥6 months NYHA class I II III IV Systolic BP, mmHg, median (IQR) Diastolic BP, mmHg, median (IQR) Heart rate, bpm, median (IQR) Laboratory eGFR, ml/min/1.73 m2, median (IQR) ≥60 30–60 b30 Hemoglobin, g/l, median (IQR) NT-proBNP, pg/ml, median (IQR)

b0.001 b0.001

0.003

b0.001

b0.001 b0.001

b0.001

b0.001

b0.001

Categorical variables are reported as n (%), continuous variables as median (interquartile range). Abbreviations: ACEI, angiotensin converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CRT, cardiac resynchronization therapy; ICD, implantable cardioverter defibrillator; eGFR, estimated glomerular filtration rate calculated by the MDRD equation; MRA, mineralocorticoid receptor antagonist; NYHA, New York Heart Association, IQR, interquartile range; TIA, transient ischemic attack; BP, blood pressure; NT-proBNP, N-Terminal pro-B-type natriuretic peptide; HFpEF, heart failure with preserved ejection fraction; HFmrEF, heart failure with mid-range ejection fraction; HFrEF, heart failure with reduced ejection fraction.

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

4

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

Table 2 Independent predictors of concomitant anemia across the ejection fraction spectrum. Variables Demographics/organizational $ Age N80 vs. b65 Age 65–80 vs. b65 $ Sex, female vs. male $ Family type, married/cohabitating vs. living alone $ Caregiver, out-patients vs. in-patients $ Specialty, cardiology vs. int medicine/geriatrics $ Follow-up referral to out-patient HF nurse clinic $ Follow-up referral specialty, primary care/other vs. specialty care Clinical Duration of HF N6 vs. ≤6 months $ NYHA class III or IV vs. I or II $ SBP 100–140 vs. b100 mmHg SBP N140 vs. b100 mmHg $ Heart rate ≥ vs. b70 bpm $ eGFR 30–60 vs. ≥60 ml/min/1.73 m2 eGFR b30 vs. ≥60 ml/min/1.73 m2 $ NT-proBNP ≥ vs. bmedian $

Comorbidities Current smoking Obesity (body mass index ≥ 30 kg/m2) $ Atrial fibrillation/flutter $ Diabetes $ Hypertension $ IHD $ History of coronary revascularization $ Valvular disease $ Valvular intervention $ Pulmonary disease $ Stroke/TIA $ Peripheral artery disease $ $

HFpEF

HFmrEF

HFrEF

p-Interaction EF ∗ variable

1.50 (1.28–1.76)⁎⁎⁎ 1.38 (1.18–1.61)⁎⁎⁎ 0.53 (0.48–0.58)⁎⁎⁎ 0.93 (0.85–1.02) 0.59 (0.53–0.65)⁎⁎⁎

1.56 (1.34–1.82)⁎⁎⁎ 1.31 (1.14–1.52)⁎⁎⁎ 0.57 (0.52–0.63)⁎⁎⁎ 0.91 (0.83–1.01) 0.62 (0.56–0.68)⁎⁎⁎ 1.12 (1.01–1.23)⁎

1.69 (1.53–1.87)⁎⁎⁎ 1.41 (1.29–1.53)⁎⁎⁎ 0.60 (0.56–0.65)⁎⁎⁎ 0.95 (0.89–1.01) 0.74 (0.70–0.80)⁎⁎⁎

1.02 (0.92–1.13) 0.94 (0.84–1.04)

1.06 (0.99–1.13) 0.90 (0.84–0.96)⁎⁎ 1.05 (0.97–1.13)

0.34 0.72 0.06 0.78 b0.01 0.63 0.02 0.06

1.06 (0.97–1.16) 1.03 (0.93–1.15) 0.75 (0.58–0.98)⁎ 0.68 (0.52–0.89)⁎⁎ 1.03 (0.93–1.13) 1.44 (1.31–1.58)⁎⁎⁎ 2.99 (2.54–3.51)⁎⁎⁎ 1.56 (1.26–1.94)⁎⁎

1.10 (1.01–1.22)⁎ 1.24 (1.10–1.39)⁎⁎⁎ 0.73 (0.58–0.93)⁎ 0.66 (0.52–0.85)⁎⁎ 1.07 (0.97–1.18) 1.32 (1.19–1.46)⁎⁎⁎ 2.31 (1.92–2.76)⁎⁎⁎ 1.89 (1.61–2.20)⁎⁎⁎

1.06 (0.99–1.13) 1.18 (1.11–1.26)⁎⁎⁎ 0.83 (0.74–0.94)⁎⁎ 0.74 (0.65–0.85)⁎⁎⁎ 1.04 (0.97–1.10) 1.30 (1.21–1.38)⁎⁎⁎ 2.88 (2.56–3.24)⁎⁎⁎ 1.52 (1.37–1.69)⁎⁎⁎

0.71 0.02 0.54 0.66 0.83 0.15 0.06 0.01

0.81 (0.66–0.98)⁎ 0.84 (0.74–0.95)⁎⁎ 1.02 (0.93–1.12) 1.48 (1.34–1.63)⁎⁎⁎

0.84 (0.70–1.01) 0.73 (0.61–0.88)⁎⁎ 1.00 (0.90–1.10) 1.52 (1.37–1.69)⁎⁎⁎

0.93 (0.84–1.03) 0.68 (0.62–0.75)⁎⁎⁎ 0.95 (0.89–1.02) 1.44 (1.35–1.55)⁎⁎⁎

0.99 (0.90–1.09) 0.99 (0.90–1.08) 1.06 (0.95–1.19) 1.15 (1.05–1.27)⁎⁎ 1.85 (1.57–2.19)⁎⁎⁎

0.97 (0.88–1.07) 1.12 (1.01–1.24)⁎ 1.17 (1.05–1.31)⁎⁎ 1.24 (1.11–1.39)⁎⁎⁎ 2.30 (1.92–2.76)⁎⁎⁎

1.04 (0.94–1.14) 1.09 (0.98–1.22) 1.12 (0.97–1.29)

1.06 (0.96–1.18) 0.96 (0.85–1.09) 1.18 (1.02–1.37)⁎

0.98 (0.92–1.04) 1.36 (1.26–1.47)⁎⁎⁎ 1.37 (1.28–1.48)⁎⁎⁎ 1.12 (1.04–1.21)⁎⁎ 1.84 (1.61–2.10)⁎⁎⁎ 1.20 (1.12–1.29)⁎⁎⁎ 1.15 (1.06–1.24)⁎⁎ 1.37 (1.24–1.50)⁎⁎⁎

0.37 0.02 0.42 0.68 0.96 b0.001 b0.001 0.26 0.09 0.02 0.07 0.045

0.93 (0.84–1.03) 0.87 (0.78–0.97)⁎ 1.12 (1.01–1.23)⁎ 1.44 (1.26–1.64)⁎⁎⁎ 0.80 (0.71–0.91)⁎⁎⁎ 0.92 (0.83–1.01) 0.97 (0.87–1.09) 0.70 (0.63–0.77)⁎⁎⁎ 0.87 (0.79–0.96)⁎⁎⁎

0.85 (0.75–0.96)⁎ 0.81 (0.71–0.93)⁎⁎ 1.09 (0.98–1.22) 1.31 (1.16–1.48)⁎⁎⁎ 0.94 (0.83–1.07) 0.88 (0.80–0.97)⁎ 1.03 (0.91–1.17) 0.64 (0.57–0.71)⁎⁎⁎ 0.88 (0.80–0.98)⁎

–

–

1.05 (0.95–1.16) 1.04 (0.94–1.16) 0.93 (0.85–1.02)

$

Concomitant medications ACEI and/or ARB $ β-blocker $ MRA $ Diuretics $ Digoxin $ Statin $ Nitrates $ Oral anticoagulants $ Antiplatelet $ CRT/ICD $

0.83 (0.76–0.92)⁎⁎⁎ 0.90 (0.81–0.99)⁎ 0.94 (0.88–1.01) 1.26 (1.15–1.37)⁎⁎⁎ 0.80 (0.74–0.87)⁎⁎⁎ 0.98 (0.92–1.04) 1.13 (1.04–1.22)⁎⁎ 0.67 (0.62–0.72)⁎⁎⁎ 0.93 (0.87–1.01) 1.04 (0.91–1.17)

0.25 0.48 0.005 0.23 0.09 0.15 0.09 0.45 0.39 –

NYHA, New York Heart Association Class; IHD, Ischemic Heart Disease; Pacemaker, patients implanted with a resynchronization device (CRT) or pacemaker; eGFR, estimated glomerular filtration rate calculated by the MDRD equation; ACEI/ARB, Angiotensin converting enzyme inhibitor/Angiotensin Receptor Blocker; ICD, Implantable Cardioverter Defibrillator; MRA, mineralocorticoid receptor antagonist, SBP Systolic Blood Pressure; NT-proBNP, N-Terminal pro-B-type natriuretic peptide; HFpEF, heart failure with preserved ejection fraction; HFmrEF, heart failure with mid-range ejection fraction; HFrEF, heart failure with reduced ejection fraction. $ Variables included in the multiple imputation models together with year of registration in SwedeHF and the composite outcome (all-cause death/HF hospitalization), and included in the logistic and Cox regression models together with year of registration in SwedeHF. ⁎ p b 0.05. ⁎⁎ p b 0.01. ⁎⁎⁎ p b 0.001.

disease were predictors of concomitant anemia in HFmrEF and HFrEF but not in HFpEF. Pulmonary disease was associated with concomitant anemia in HFrEF but not in HFpEF and HFmrEF. Use of mineralocorticoid receptor antagonists (MRA) was associated with concomitant anemia in HFpEF but not in HFmrEF and HFrEF. No use of ACEI or ARB significantly predicted anemia in HFrEF and HFmrEF but not in HFpEF, although no interaction between ACEI/ARB use and anemia was identified. Use of antiplatelets was predictor of anemia in HFpEF and HFmrEF but not in HFrEF, but no significant interaction between treatment and anemia was identified. There were not major differences in predictors of concomitant anemia in in-patients vs. out-patients (Supplement Tables 1–3).

5531 (55%) and 4260 (43%) patients with HFpEF, 4710 (51%) and 3470 (38%) patients with HFmrEF, and 13,470 (57%) and 9135 (38%) patients with HFrEF, respectively. One-year survival free of HF hospitalization in patients with vs. without anemia was 57% vs. 73% in HFpEF, 59% vs. 77% in HFmrEF, and 53% vs. 69% in HFrEF (p b 0.01 for all), whereas 1year survival was 73% vs. 86% in HFpEF, 76% vs. 89% in HFmrEF, and 74% vs. 88% in HFrEF (p b 0.01 for all). For the composite outcome, the unadjusted difference in risk between patients with vs. without anemia was greatest in HFmrEF, whereas for all-cause death it was highest in HFrEF, intermediate in HFmrEF and lowest in HFpEF.

3.4. Outcomes (Fig. 1)

3.5. Independent association between anemia and outcomes (Fig. 1, Table 3)

Over a median (IQR) follow-up of 2.2 (0.9–4.1) years, the composite outcome and all-cause mortality occurred with vs. without anemia in

The large differences in crude risk of outcomes may be confounded by e.g. comorbidities, severity of HF and medication use. Therefore, we

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

5

Fig. 1. Kaplan-Meier curves for survival free from HF hospitalization and survival in HFpEF, HFmrEF and HFrEF according to presence vs. absence of anemia.

conducted adjusted Cox regression models to assess the impact of patients' characteristics on the association between anemia and outcomes. The hazard ratio (HR, 95% CI) for the composite outcome gradually decreased, but remained significant, from 1.70 (1.61–1.79) to 1.24 (1.18–1.30) in HFpEF, from 1.91 (1.80–2.02) to 1.26 (1.19–1.34) in HFmrEF, and from 1.69 (1.64–1.75) to 1.14 (1.10–1.19) in HFrEF after adjusting for age, sex, eGFR, caregiver and finally all the baseline covariates, with a major role played by age, eGFR and caregiver. Anemia was associated with significantly higher risk of the composite outcome in HFpEF and HFmrEF vs. HFrEF (p-interaction = 0.003). For all-cause mortality, the HR decreased from 1.86 (1.75–1.97) to 1.28 (1.20–1.36) in HFpEF, from 2.04 (1.91–2.18) to 1.21 (1.13–1.29) in HFmrEF, and from 2.14 (2.05–2.23) to 1.30 (1.24–1.35) in HFrEF after adjustments, again with a major role as confounders for age, eGFR and caregiver. After adjustments, anemia was associated with a similarly increased risk of all-cause mortality across the EF spectrum. 3.6. Subgroup analysis (Supplement Table 4) A majority of the baseline characteristics had significant interactions with anemia at the outcome analysis. Anemia was overall associated

with greater risk of both all-cause death and the composite of allcause death and HF hospitalization in patients with overall better health status, i.e. younger age and out-patients vs. in-patients (expect for allcause mortality in HFrEF). In HFrEF, anemia was associated with greater risk of all-cause mortality in patients with NYHA class I–II vs. III–IV, in those with higher eGFR, no history of ischemic heart disease (same result also in HFmrEF), and those not receiving diuretics.

4. Discussion In this comprehensive assessment of anemia in patients with HFpEF vs. HFmrEF vs. HFpEF we found that (1) anemia was common regardless of EF, but with higher prevalence in HFpEF (41%) vs. HFmrEF (35%) vs. HFrEF (32%); (2) anemia was independently associated with several other patients' characteristics such as higher age, male sex, worse renal function, comorbidities, frailty and severity of HF; and (3) over long-term follow-up, anemia was independently associated with increased risk of death or HF hospitalization and of death alone in both HFpEF, HFmrEF and HFrEF, but with higher risk of all-cause death or HF hospitalization in HFpEF and HFmrEF vs. HFrEF and no differences in risk across the EF spectrum for all-cause mortality.

Table 3 Cox proportional hazard regression models for the association between anemia and the composite of all-cause mortality and HF hospitalization, and all-cause mortality alone. Stepwise adjustments are displayed. Composite outcome

Univariatea Adjusted only agea,b Adjusted only sexa,b Adjusted only eGFRa,b Adjusted only caregivera,b Adjusted age and sexa Adjusted age, sex, eGFRa Adjusted age, sex, eGFR, caregivera Adjusted alla

All-cause mortality

HFpEF

HFmrEF

HFrEF

HFpEF

HFmrEF

HFrEF

1.70 (1.61–1.79)⁎⁎⁎ 1.57 (1.49–1.66)⁎⁎⁎ 1.71 (1.62–1.80)⁎⁎⁎ 1.49 (1.41–1.57)⁎⁎⁎ 1.54 (1.46–1.62)⁎⁎⁎ 1.56 (1.48–1.65)⁎⁎⁎ 1.43 (1.36–1.51)⁎⁎⁎ 1.33 (1.26–1.41)⁎⁎⁎ 1.24 (1.18–1.31)⁎⁎⁎

1.91 (1.80–2.02)⁎⁎⁎ 1.70 (1.61–1.80)⁎⁎⁎ 1.91 (1.81–2.03)⁎⁎⁎ 1.68 (1.59–1.79)⁎⁎⁎ 1.72 (1.62–1.82)⁎⁎⁎ 1.70 (1.60–1.80)⁎⁎⁎ 1.57 (1.48–1.66)⁎⁎⁎ 1.47 (1.38–1.55)⁎⁎⁎ 1.26 (1.19–1.34)⁎⁎⁎

1.69 (1.64–1.75)⁎⁎⁎ 1.50 (1.45–1.56)⁎⁎⁎ 1.70 (1.64–1.76)⁎⁎⁎ 1.49 (1.43–1.54)⁎⁎⁎ 1.56 (1.51–1.62)⁎⁎⁎ 1.50 (1.45–1.55)⁎⁎⁎ 1.38 (1.33–1.43)⁎⁎⁎ 1.31 (1.26–1.35)⁎⁎⁎ 1.14 (1.10–1.19)⁎⁎⁎

1.86 (1.75–1.97)⁎⁎⁎ 1.67 (1.57–1.77)⁎⁎⁎ 1.88 (1.77–1.99)⁎⁎⁎ 1.56 (1.46–1.65)⁎⁎⁎ 1.67 (1.57–1.77)⁎⁎⁎ 1.65 (1.56–1.76)⁎⁎⁎ 1.48 (1.39–1.57)⁎⁎⁎ 1.39 (1.31–1.47)⁎⁎⁎ 1.28 (1.20–1.36)⁎⁎⁎

2.04 (1.91–2.18)⁎⁎⁎ 1.70 (1.59–1.82)⁎⁎⁎ 2.05 (1.92–2.19)⁎⁎⁎ 1.72 (1.61–1.84)⁎⁎⁎ 1.81 (1.70–1.94)⁎⁎⁎ 1.70 (1.59–1.81)⁎⁎⁎ 1.54 (1.44–1.65)⁎⁎⁎ 1.44 (1.35–1.54)⁎⁎⁎ 1.21 (1.13–1.29)⁎⁎⁎

2.14 (2.05–2.23)⁎⁎⁎ 1.77 (1.70–1.85)⁎⁎⁎ 2.16 (2.07–2.25)⁎⁎⁎ 1.80 (1.72–1.88)⁎⁎⁎ 1.97 (1.89–2.05)⁎⁎⁎ 1.76 (1.69–1.84)⁎⁎⁎ 1.59 (1.52–1.66)⁎⁎⁎ 1.51 (1.45–1.58)⁎⁎⁎ 1.30 (1.24–1.35)⁎⁎⁎

eGFR, estimated glomerular filtration rate calculated by the MDRD equation; HFpEF, heart failure with preserved ejection fraction; HFmrEF, heart failure with mid-range ejection fraction; HFrEF, heart failure with reduced ejection fraction. For HR (95% CI) p b 0.05; p b 0.01; ⁎⁎⁎p b 0.001. a p-Interaction anemia EF b0.05 for the composite outcome. b p-Interaction anemia EF b0.05 for all-cause mortality.

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

6

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

4.1. Prevalence of anemia

4.3. Associations between anemia and outcomes

Anemia has been extensively studied in patients with chronic and acute HF. Previous analyses showed that its prevalence ranges 10–68% in the overall HF population, 21–68% in HFpEF and 10–50% in HFrEF. Higher prevalence has been reported in real-world vs. trial populations [2,5,11–16]. In our large and unselected HF population, 34% of patients suffered from anemia. Specifically, 41% of patients with HFpEF, 35% of those with HFmrEF and 32% of those with HFrEF had anemia, which is more than what observed in other comparable cohorts. Indeed, in the Get With the Guidelines – HF (GWTG-HF) population, anemia was identified in 14% of patients with HFrEF, 20% of those with HFmrEF and 22% of those with HFpEF [4]. It is increasingly evident that anemia is more common with higher EF. This may be explained by higher age and more comorbidities which are related to anemia in patients with higher EF. However, in our analysis, after extensive adjustments for several patients' characteristics, higher EF was still an independent predictor of concomitant anemia. This raises the hypothesis that anemia is not only a marker of ageing, impaired renal function and other comorbidities (more common with higher EF), but it may also contribute to explain symptoms, physical inactivity, deconditioning and progressively worsening exercise tolerance [4]. Additionally, the higher prevalence of anemia in patients with HFpEF vs. HFmrEF vs. HFrEF may be explained by anemia being risk factor for HFpEF, but risk marker of more advanced HF, the cardiorenal syndrome, and potentially hypoxemic status due to low output in those with lower EF. We also observed higher prevalence of anemia in in-patients vs. out-patients across the EF spectrum, which is consistent with hemodilution and with more severe HF and higher burden of comorbidities in patients admitted for HF. The interaction between anemia and iron deficiency has been poorly studied, but the latter is common in prevalent HF [17] and may also be a risk factor for HF [18]. Studies aiming to investigate the distinct roles of anemia and iron deficiency in HF across the EF spectrum are needed.

In previous studies, anemia has been shown to be independently associated with worse outcomes regardless of EF, although the HFmrEF subgroup was poorly represented [4,16,24]. Many potential mechanisms have been suggested, including neurohormonal activation and inflammation, adverse cardiorenal interplay and myocardial remodeling, nutritional deficiency and anabolic/catabolic imbalance, and hemodilution. However, whether anemia may be a risk marker vs. a risk factor is still debated [4,25–28]. In our analysis, the association between anemia and outcomes was independent of several potential confounders, and importantly of CKD [29]. However, ferritin and other lab measurements needed to identify potential iron deficiency were not available and thus, not considered for adjustments. Notably, as previously observed [30,31], we showed that anemia was a predictor of all-cause mortality regardless of EF. However, we also report that anemia was associated with a greater risk of all-cause death or HF hospitalization in HFpEF and HFmrEF vs. HFrEF. Our results may be explained by anemia being associated with mortality independently of HF and thus of the HF phenotype [32]. However, in patients with HFpEF and HFmrEF, anemia may exacerbate HF-related symptoms more than in HFrEF, leading to higher risk of HF hospitalization. If our hypothesis is correct, treating anemia may be beneficial regardless of the EF phenotype, but may even more heavily contribute to reduce hospital admissions in patients with higher EF. Finally, we showed that anemia was associated with a greater risk of outcomes in younger patients and those with milder HF, in particular in HFrEF. These findings may suggest that anemia may be relatively more harmful when other factors, such as comorbidity and severity of HF, are less important.

4.2. Associations with anemia Our analysis confirmed previous studies showing that HF patients with anemia, regardless of EF, are older, with worse renal function, more comorbidities and are frailer as compared with those without anemia [19]. However, as for other comorbidities and biomarkers [20], whether these patients' characteristics are independent predictors of anemia is less well studied [5]. We observed that age was a strong independent predictor of anemia within each HF phenotype, suggesting that ageing itself or ageing-related characteristics other than those considered in this analysis, e.g. poor nutrition and cognitive decline [21], may have contributed. Notably, diabetes was associated with a 48% increased risk of anemia in HFpEF, 51% in HFmrEF and 44% in HFrEF, independently of renal function and all the other patients' characteristics. Furthermore, as expected, also chronic kidney disease (CKD) was an independent predictor of anemia. These data highlight the importance of the interplay between anemia, CKD and diabetes in both HFpEF and HFmrEF, and may support a key role for comorbidities in HFpEF [22,23]. Important predictors which differed across the EF spectrum were NYHA class and ischemic heart disease, with higher NYHA class and history of ischemic heart disease being associated with concomitant anemia in HFrEF and HFmrEF but not in HFpEF. These results may be explained by 1) anemia as consequence of the cardiorenal syndrome and hypoxemic status due to low output in patients with more advanced HFrEF (i.e. NYHA III-IV vs I-II) and those with HFmrEF transiting toward HFrEF, and 2) coronary artery disease more frequent in HFrEF vs. HFmrEF vs. HFpEF and thus anemia more likely to contribute to type 2 myocardial infarctions in patients with lower EF. Finally, being inpatient vs. out-patient was more likely to be associated with concomitant anemia in HFpEF and HFmrEF vs. HFrEF, which may be linked with anemia having a role in explaining the HF symptoms which lead to hospital admission in HFpEF and HFmrEF vs. HFrEF.

4.4. Limitations Although our observational study was extensively adjusted for several potential confounders, we cannot rule out potential residual confounding. Indeed, laboratory data needed for the diagnosis of iron deficiency were not available, and thus the relationship/interaction among iron deficiency, anemia and HF has not been captured in our analyses. Some variables had missing data, which were handled by multiple imputation. Although this statistical method reduces the chance of bias due to data not missing at random and thus increases external validity, internal validity may be affected. Although our registry is nationwide, it does not have complete coverage, and thus selection bias may be still a limitation [33]. Moreover, hemoglobin levels were obtained as a single measurement at the hospital discharge/clinical visit, and therefore, we cannot exclude that some patients had only transitory anemia or hemodilution. Finally, potential underlying causes of anemia in these patients were not considered. 5. Conclusions In this nation-wide registry, prevalence of anemia was higher in HFpEF vs. HFmrEF vs. HFrEF, but was associated with a similarly increased risk of death across the EF spectrum, with greater risk of death or HF hospitalization in HFpEF and HFmrEF vs. HFrEF. These findings suggest that while anemia may be one among many other comorbidities contributing to HFpEF and, although to a lesser extent, to HFmrEF pathogenesis, it may be a risk marker in HF, leading to identify those patients with more severe HF and cardiorenal syndrome who are more likely to experience adverse outcomes. Sources of funding The Swedish Heart Failure Registry (SwedeHF) is funded by the Swedish National Board of Health and Welfare, the Swedish Association of Local Authorities and Regions, and the Swedish Society of Cardiology. This study was supported in part by a research grant from Linköping

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049

G. Savarese et al. / International Journal of Cardiology xxx (xxxx) xxx

University (LIO 19211) and financial support from the Swedish HF Registry foundation. No funding agency had any role in the design and conduct of the study, in the collection, management, analysis, or interpretation of the data, or in the preparation, review, or approval of the manuscript. Disclosures There are no disclosures related to the work submitted. The following are the authors' declaration of the potential conflicts of interest outside the work submitted: AJ: None declared. ACH: None declared. ME: None declared. LHL: Research grants to authors' institution: Astra Zeneca, Novartis. Speaker's and/or consulting honoraria from AstraZeneca, Novartis, Bayer, Vifor Pharma, Relypsa. UD: research grants to authors' institution from AstraZeneca Inc., speakers and/or consulting honoraria, Novartis Inc., Vifor Pharma. GS. None related to the current work. Outside: research grants from Boehringer Ingelheim and Merck Sharp & Dohme; honoraria from Vifor, Servier, Roche, AstraZeneca. Author contributions AJ, LHL, GS, ACH, and UD and had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. UD is also the founder of and registrar for research for the Swedish Heart Failure Registry. Study concept and design: LHL, GS, AJ, UD, ME (co-founder of the Swedish Heart Failure Registry). Acquisition, analysis, or interpretation of data: AJ, GS, LHL, ACH, UD, ME, GS. Drafting of the manuscript: AJ, LHL, UD, GS. Critical revision of the manuscript for important intellectual content: AJ, LHL, UD, ME, GS. Statistical analysis: GS. Obtained funding: UD. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.ijcard.2019.08.049. References [1] D. Mozaffarian, E.J. Benjamin, A.S. Go, D.K. Arnett, M.J. Blaha, M. Cushman, et al., Heart disease and stroke statistics—2015 update: a report from the American Heart Association, Circulation. 131 (2015) e29–322. [2] L.H. Lund, E. Donal, E. Oger, C. Hage, H. Persson, I. Haugen-Lofman, et al., Association between cardiovascular vs. non-cardiovascular co-morbidities and outcomes in heart failure with preserved ejection fraction, Eur. J. Heart Fail. 16 (2014) 992–1001. [3] P. Ponikowski, A.A. Voors, S.D. Anker, H. Bueno, J.G. Cleland, A.J. Coats, et al., 2016 ESC guidelines for the diagnosis and treatment of acute and chronic heart failure: the Task Force for the diagnosis and treatment of acute and chronic heart failure of the European Society of Cardiology (ESC). Developed with the special contribution of the Heart Failure Association (HFA) of the ESC, Eur. J. Heart Fail. 18 (2016) 891–975. [4] R.J. Mentz, J.P. Kelly, T.G. von Lueder, A.A. Voors, C.S. Lam, M.R. Cowie, et al., Noncardiac comorbidities in heart failure with reduced versus preserved ejection fraction, J. Am. Coll. Cardiol. 64 (2014) 2281–2293. [5] F. Triposkiadis, G. Giamouzis, J. Parissis, R.C. Starling, H. Boudoulas, J. Skoularigis, et al., Reframing the association and significance of co-morbidities in heart failure, Eur. J. Heart Fail. 18 (2016) 744–758. [6] A. Jonsson, A.C. Hallberg, M. Edner, L.H. Lund, U. Dahlstrom, A comprehensive assessment of the association between anemia, clinical covariates and outcomes in a population-wide heart failure registry, Int. J. Cardiol. 211 (2016) 124–131. [7] C. Berry, K.K. Poppe, G.D. Gamble, N.J. Earle, J.A. Ezekowitz, I.B. Squire, et al., Prognostic significance of anaemia in patients with heart failure with preserved and reduced ejection fraction: results from the MAGGIC individual patient data meta-analysis, QJM 109 (2016) 377–382.

7

[8] A. Jonsson, M. Edner, U. Alehagen, U. Dahlstrom, Heart failure registry: a valuable tool for improving the management of patients with heart failure, Eur. J. Heart Fail. 12 (2010) 25–31. [9] G. Savarese, P. Vasko, A. Jonsson, M. Edner, U. Dahlstrom, L.H. Lund, The Swedish Heart Failure Registry: a living, ongoing quality assurance and research in heart failure, Ups. J. Med. Sci. (2018) 1–5. [10] E. Ingelsson, J. Arnlov, J. Sundstrom, L. Lind, The validity of a diagnosis of heart failure in a hospital discharge register, Eur. J. Heart Fail. 7 (2005) 787–791. [11] S. Hamaguchi, M. Tsuchihashi-Makaya, S. Kinugawa, T. Yokota, A. Takeshita, H. Yokoshiki, et al., Anemia is an independent predictor of long-term adverse outcomes in patients hospitalized with heart failure in Japan. A report from the Japanese Cardiac Registry of Heart Failure in Cardiology (JCARE-CARD), Circ. J. 73 (2009) 1901–1908. [12] B.A. Steinberg, X. Zhao, P.A. Heidenreich, E.D. Peterson, D.L. Bhatt, C.P. Cannon, et al., Trends in patients hospitalized with heart failure and preserved left ventricular ejection fraction: prevalence, therapies, and outcomes, Circulation 126 (2012) 65–75. [13] V.M. van Deursen, R. Urso, C. Laroche, K. Damman, U. Dahlstrom, L. Tavazzi, et al., Co-morbidities in patients with heart failure: an analysis of the European Heart Failure Pilot Survey, Eur. J. Heart Fail. 16 (2014) 103–111. [14] A. Makubi, C. Hage, J. Lwakatare, B. Mmbando, P. Kisenge, L.H. Lund, et al., Prevalence and prognostic implications of anaemia and iron deficiency in Tanzanian patients with heart failure, Heart 101 (2015) 592–599. [15] A. Makubi, C. Hage, J. Lwakatare, P. Kisenge, J. Makani, L. Ryden, et al., Contemporary aetiology, clinical characteristics and prognosis of adults with heart failure observed in a tertiary hospital in Tanzania: the prospective Tanzania Heart Failure (TaHeF) study, Heart 100 (2014) 1235–1241. [16] S. von Haehling, D.J. van Veldhuisen, M. Roughton, D. Babalis, R.A. de Boer, A.J. Coats, et al., Anaemia among patients with heart failure and preserved or reduced ejection fraction: results from the SENIORS study, Eur. J. Heart Fail. 13 (2011) 656–663. [17] V. Melenovsky, J. Petrak, T. Mracek, J. Benes, B.A. Borlaug, H. Nuskova, et al., Myocardial iron content and mitochondrial function in human heart failure: a direct tissue analysis, Eur. J. Heart Fail. 19 (2017) 522–530. [18] O.M. Silvestre, A. Goncalves, W. Nadruz Jr., B. Claggett, D. Couper, J.H. Eckfeldt, et al., Ferritin levels and risk of heart failure-the Atherosclerosis Risk in Communities Study, Eur. J. Heart Fail. 19 (2017) 340–347. [19] J.B. Young, W.T. Abraham, N.M. Albert, W. Gattis Stough, M. Gheorghiade, B.H. Greenberg, et al., Relation of low hemoglobin and anemia to morbidity and mortality in patients hospitalized with heart failure (insight from the OPTIMIZE-HF registry), Am. J. Cardiol. 101 (2008) 223–230. [20] L.H. Lund, A. Gabrielsen, Biomarkers in advanced heart failure—pathophysiology leading to clinical use? J. Heart Lung Transplant. 33 (2014) 1213–1214. [21] P. Cermakova, L.H. Lund, S.M. Fereshtehnejad, K. Johnell, B. Winblad, U. Dahlstrom, et al., Heart failure and dementia: survival in relation to types of heart failure and different dementia disorders, Eur. J. Heart Fail. 17 (2015) 612–619. [22] W.J. Paulus, C. Tschope, A novel paradigm for heart failure with preserved ejection fraction: comorbidities drive myocardial dysfunction and remodeling through coronary microvascular endothelial inflammation, J. Am. Coll. Cardiol. 62 (2013) 263–271. [23] C.S. Lam, L.H. Lund, Microvascular endothelial dysfunction in heart failure with preserved ejection fraction, Heart 102 (2016) 257–259. [24] T. Thorvaldsen, L. Benson, M. Stahlberg, U. Dahlstrom, M. Edner, L.H. Lund, Triage of patients with moderate to severe heart failure: who should be referred to a heart failure center, J. Am. Coll. Cardiol. 63 (2014) 661–671. [25] A.S. Androne, S.D. Katz, L. Lund, J. LaManca, A. Hudaihed, K. Hryniewicz, et al., Hemodilution is common in patients with advanced heart failure, Circulation 107 (2003) 226–229. [26] C. Hage, E. Michaelsson, C. Linde, E. Donal, J.C. Daubert, L.M. Gan, et al., Inflammatory biomarkers predict heart failure severity and prognosis in patients with heart failure with preserved ejection fraction: a holistic proteomic approach, Circ. Cardiovasc. Genet. 10 (2017). [27] U.L. Faxen, C. Hage, L. Benson, S. Zabarovskaja, A. Andreasson, E. Donal, et al., HFpEF and HFrEF display different phenotypes as assessed by IGF-1 and IGFBP-1, J. Card. Fail. 23 (2017) 293–303. [28] L.H. Lund, J. Stehlik, Risk scores and biomarkers in heart failure: a journey to predictive accuracy and clinical utility, J. Heart Lung Transplant. 35 (2016) 711–713. [29] I. Lofman, K. Szummer, I. Hagerman, U. Dahlstrom, L.H. Lund, T. Jernberg, Prevalence and prognostic impact of kidney disease on heart failure patients, Open Heart 3 (2016), e000324. [30] K.W. Streng, J.F. Nauta, H.L. Hillege, S.D. Anker, J.G. Cleland, K. Dickstein, et al., Noncardiac comorbidities in heart failure with reduced, mid-range and preserved ejection fraction, Int. J. Cardiol. 271 (2018) 132–139. [31] A. Iorio, M. Senni, G. Barbati, S.J. Greene, S. Poli, E. Zambon, et al., Prevalence and prognostic impact of non-cardiac co-morbidities in heart failure outpatients with preserved and reduced ejection fraction: a community-based study, Eur. J. Heart Fail. 20 (2018) 1257–1266. [32] M.J. Sarnak, H. Tighiouart, G. Manjunath, B. MacLeod, J. Griffith, D. Salem, et al., Anemia as a risk factor for cardiovascular disease in The Atherosclerosis Risk in Communities (ARIC) study, J. Am. Coll. Cardiol. 40 (2002) 27–33. [33] L.H. Lund, J.J. Carrero, B. Farahmand, K.M. Henriksson, A. Jonsson, T. Jernberg, et al., Association between enrolment in a heart failure quality registry and subsequent mortality-a nationwide cohort study, Eur. J. Heart Fail. 19 (2017) 1107–1116.

Please cite this article as: G. Savarese, Å. Jonsson, A.-C. Hallberg, et al., Prevalence of, associations with, and prognostic role of anemia in heart failure across the ejection..., International Journal of Cardiology, https://doi.org/10.1016/j.ijcard.2019.08.049