European Journal of Heart Failure (2017) 19, 615–623 RESEARCH ARTICLE

doi:10.1002/ejhf.778

The association of chronic kidney disease

and microalbuminuria with heart failure with
preserved vs. reduced ejection fraction
Matthew Nayor1,2, Martin G. Larson1,3, Na Wang4, Rajalakshmi Santhanakrishnan5,
Douglas S. Lee6,7, Connie W. Tsao1,8, Susan Cheng1,2, Emelia J. Benjamin1,9,
Ramachandran S. Vasan1,9, Daniel Levy1,10, Caroline S. Fox1,10,11, and
Jennifer E. Ho1,12*
1 NationalHeart, Lung, and Blood Institute’s and Boston University’s Framingham Heart Study, Framingham, MA, USA; 2 Division of Cardiovascular Medicine, Brigham and
Women’s Hospital, Boston, MA, USA; 3 Department of Biostatistics, Boston University School of Public Health, Boston, MA, USA; 4 Data Coordinating Center, Boston University
School of Public Health, Boston, MA, USA; 5 Section of Cardiovascular Medicine, Department of Medicine, Boston University, Boston, MA, USA; 6 Institute for Clinical Evaluative
Sciences, University of Toronto, Toronto, Canada; 7 Peter Munk Cardiac Centre, University Health Network, Toronto, Canada; 8 Cardiovascular Division, Beth Israel Deaconess
Medical Center, Boston, MA, USA; 9 Sections of Preventive Medicine & Epidemiology, and Cardiology, Department of Medicine, Boston University School of Medicine, and
Department of Epidemiology, Boston University School of Public Health, Boston, MA, USA; 10 Center for Population Studies of the National Heart, Lung, and Blood Institute,
Bethesda, MD, USA; 11 Division of Endocrinology, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA; and 12 Cardiology Division and Cardiovascular
Research Center, Department of Medicine, Massachusetts General Hospital, Boston, MA, USA

Received 18 August 2016; revised 31 October 2016; accepted 30 November 2016 ; online publish-ahead-of-print 20 February 2017

Aims Chronic kidney disease (CKD) and microalbuminuria are associated with incident heart failure (HF), but their relative
contributions to HF with preserved vs. reduced EF (HFpEF and HFrEF) are unknown. We sought to evaluate the
associations of CKD and microalbuminuria with incident HF subtypes in the community-based Framingham Heart
Study (FHS).
.....................................................................................................................................................................
Methods We defined CKD as glomerular filtration rate <60 mL/min/1.73 m2 , and microalbuminuria as a urine albumin to
and results creatinine ratio (UACR) ≥17 mg/g in men and ≥25 mg/g in women. We observed 754 HF events (324 HFpEF/326
HFrEF/104 unclassified) among 9889 FHS participants with serum creatinine measured (follow-up 13 ± 4 years).
In Cox models adjusted for clinical risk factors, CKD (prevalence = 9%) was associated with overall HF [hazard
ratio (HR) 1.24, 95% confidence interval (CI) 1.01–1.51], but was not significantly associated with individual HF
subtypes. Among 2912 individuals with available UACR (follow-up 15 ± 4 years), 192 HF events (91 HFpEF/93 HFrEF/8
unclassified) occurred. Microalbuminuria (prevalence = 17%) was associated with a higher risk of overall HF (HR
1.71, 95% CI 1.25–2.34) and HFrEF (HR 2.10, 95% CI 1.35–3.26), but not HFpEF (HR 1.26, 95% CI 0.78–2.03).
In cross-sectional analyses, microalbuminuria was associated with LV systolic dysfunction (odds ratio 3.19, 95% CI
1.67–6.09).
.....................................................................................................................................................................
Conclusions Microalbuminuria was associated with incident HFrEF prospectively, and with LV systolic dysfunction cross-sectionally
in a community-based sample. In contrast, CKD was modestly associated with overall HF but not differentially
associated with HFpEF vs. HFrEF. The mechanisms responsible for the relationship of microalbuminuria to future
development of HFrEF warrant further investigation.
..........................................................................................................
Keywords Heart failure • Epidemiology • Prevention • Kidney disease • Microalbuminuria •
Cardiorenal syndrome

*Corresponding author. Massachusetts General Hospital, CPZN 185 Cambridge Street, #3224, Boston, MA 02114, USA. Tel: +1 617 724 6411, Fax: +1 617 643 3451, Email:
jho1@mgh.harvard.edu
© 2017 The Authors
European Journal of Heart Failure © 2017 European Society of Cardiology
616 M. Nayor et al.

Introduction incident HF, 3075 participants for cross-sectional analyses of eGFR

........................................................................................................................................................................
and echocardiographic traits, and 2652 for cross-sectional analyses
Kidney disease is associated with poor heart failure (HF)-related of albuminuria and echocardiographic traits (details of inclusion and
outcomes regardless of disease severity. In patients with estab- exclusion criteria are given in Supplementary material online, Figures
lished HF, reduced estimated glomerular filtration rate (eGFR) and S1 and S2). All study protocols were approved by the Boston Uni-
albuminuria are independent predictors of HF progression and versity Medical Center Institutional Review Board, and all participants
provided written informed consent.
mortality.1 – 6 Similarly, in community-based investigations, albu-
minuria and lower eGFR are associated with new-onset HF.7 – 13
The mechanisms responsible for cardiorenal interactions in HF are
incompletely understood and are likely to vary by underlying dis- Laboratory measures
ease process and HF syndrome.14,15 Fasting plasma samples were collected from FHS participants at each
Heart failure with preserved ejection fraction (HFpEF) and examination cycle, centrifuged, and stored at −80 ∘ C. Serum creatinine
HF with reduced ejection fraction (HFrEF) represent separate levels were assayed using the modified Jaffé method on a Roche
Hitachi 911 (Roche Diagnostics), and eGFR was calculated using
clinical subtypes, each accounting for approximately half of HF
the CKD-EPI equation.30 Chronic kidney disease (CKD) was defined
diagnoses.16 – 20 Although controversy exists regarding whether
as eGFR <60 mL/min/1.73 m2 . Spot urine samples were collected at
HFpEF and HFrEF represent overlapping or distinct pathophys-
the sixth offspring examination cycle and stored at −20 ∘ C. Urine
iological entities,21,22 substantial evidence supports distinct risk albumin concentration was assayed using immunoturbimetry (Tina
factor profiles23 – 25 and treatment responses26 for each subtype. Quant Albumin Assay, Roche Diagnostics) and urine creatinine was
Therefore, understanding how specific risk factors relate to the assayed using the modified Jaffé method (Colorimetric Assay, Roche,
development of HFpEF vs. HFrEF may allow for improved methods Diagnostics). Microalbuminuria was defined as a urine albumin to
to predict, and potentially prevent, the occurrence of HF. creatinine ratio (UACR) ≥17 mg/g in men and ≥25 mg/g in women.31
The relative contribution of kidney dysfunction to the devel-
opment of HFpEF vs. HFrEF is not well elucidated. Incident HF
subtypes have been adjudicated among participants of the Fram- Clinical assessment
ingham Heart Study (FHS), where kidney function has also been Systolic blood pressure was measured with a mercury sphygmo-
systematically assessed. We sought to evaluate the associations of manometer on seated participants, and the mean of two readings
albuminuria and reduced glomerular filtration with incident HFpEF was recorded for each individual. Weight in kilograms was divided by
and HFrEF prospectively. To understand further the mechanisms the square of height in metres to determine body mass index (BMI).
underlying these associations, we investigated the cross-sectional Valvular heart disease was defined as a systolic murmur ≥3/6 in inten-
relationships of albuminuria and reduced glomerular filtration with sity or any diastolic murmur. We defined diabetes as fasting glucose
echocardiographic measures of cardiac structure and function. ≥126 mg/dL, non-fasting glucose ≥200 mg/dL, or taking medications
for the treatment of elevated glucose. Total and HDL-cholesterol were
measured using standardized assays. Current smoking, antihypertensive
treatment status, and history of myocardial infarction were ascertained
Methods by review of relevant medical records and patient questionnaires.
Study samples
The design and enrolment of the FHS Original and Offspring cohorts
have been detailed previously.27,28 Participants underwent comprehen- Echocardiography
sive evaluation at the FHS clinic every 2 years (Original cohort) or Comprehensive two-dimensional transthoracic echocardiography with
4 years (Offspring cohort). At each examination cycle, a comprehen- Doppler colour flow imaging was performed on all attendees of the
sive medical history, anthropometry, a cardiovascular-targeted physical sixth examination cycle of the Framingham Offspring study. Echocar-
examination, and ECG were performed. diograms were read by a cardiologist or sonographer blinded to
To evaluate the associations of reduced eGFR and incident HF, participants’ clinical information. Measurements were made on dig-
we included Original cohort participants who attended examinations itized M-mode images using the leading edge-to-leading edge tech-
15 (1977–1979) or 24 (1995–1998), or Offspring cohort partici- nique and were averaged over at least three cardiac cycles. This
pants who attended examinations 2 (1979–1983) or 6 (1995–1998). technique was used to measure left atrial dimension (LAD), LV sep-
Follow-up was censored at 15 years or the date of the next exam- tal wall thickness, LV posterior wall thickness, LV diameter at the
ination, depending on which was the shorter follow-up time period. end of diastole (LVEDD), and the LV diameter at the end of sys-
Participants in each cohort were eligible for inclusion at both exam- tole (LVESD). Fractional shortening (FS) was then calculated using
ination cycles if no interim HF event occurred.29 Of 10 858 eligible the formula: [(LVEDD – LVESD)/LVEDD] × 100, and LV mass (LVM)
person-observations, we excluded 1 for missing follow-up time, 257 was calculated as: 0.8{1.04[(LVEDD + posterior wall thickness + septal
for off-site visits, 548 for missing clinical covariates, 108 for prevalent wall thickness)3 – (LVEDD)3 ]} + 0.6, based on American Society of
HF, and 55 for eGFR <15 mL/min/1.73 m2 or for being on haemodial- Echocardiography guidelines.32 Mild or greater reduction in LVEF based
ysis, yielding a final sample of 9889 person-observations. on qualitative assessment or FS <29% was used to define LV systolic
Urinary albumin and creatinine measures, and transthoracic dysfunction. LV hypertrophy (LVH) was determined by indexing LVM
echocardiography were performed at Offspring examination 6. Of to body surface area, and values >115 g/m2 for men and >95 g/m2 for
3532 participants who attended this examination cycle, we included women were considered abnormal.32 Left atrial enlargement (LAE) was
2912 participants for prospective analyses examining albuminuria and defined as LAD >4.0 cm in men and >3.8 cm in women.32

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Kidney disease and heart failure subtypes 617

Clinical assessment and outcome In cross-sectional analyses, we evaluated the associations between

........................................................................................................................................................................
the same predictors [CKD, log(eGFR), microalbuminuria, and
definitions log(UACR)] and echocardiographic traits using multivariable-adjusted
Participants in the FHS are under surveillance for cardiovascular linear and logistic regression models. These models adjusted for age,
outcomes, including hospitalized HF, which are ascertained after review sex, smoking, BMI, diabetes, valvular heart disease, systolic blood
of all medical records including physician visits and hospitalizations pressure, and hypertension treatment status.
related to cardiovascular disease. Outcome events were adjudicated We evaluated the association of two kidney function traits (CKD
by a committee of three FHS physicians using the FHS criteria,33 which and microalbuminuria) with two main outcomes (HFpEF and HFrEF),
define HF based on the presence of two major or one major and two and accordingly a Bonferroni corrected P-value of 0.0125 (0.05 divided
minor criteria. Major criteria are paroxysmal nocturnal dyspnoea or by 4) was used to determine statistical significance for our prospec-
orthopnoea, increased venous pressure, distended neck veins, rales, tive analyses, with a P-value of 0.0125–0.05 considered ‘suggestive’ of
cardiomegaly by radiograph, pulmonary oedema, a third heart sound, an association. For the echocardiographic analyses, we evaluated the
hepatojugular reflux, and weight loss on diuretic therapy. Minor criteria association of the two kidney function traits with three main echocar-
include the presence of ankle oedema, nocturnal cough, hepatomegaly, diographic variables (systolic function, LVM/LVH, and left atrial size)
dyspnoea on exertion, pleural effusion, decrease in vital capacity, and we therefore used a P-value of 0.008 to assess statistical signifi-
and tachycardia. The FHS HF criteria are used in a wide variety of cance. Analyses were performed using SAS Software version 9.3 (Cary,
study settings and have been shown to have adequate sensitivity and NC, USA).
specificity for the diagnosis of HF.34,35 Incident HF events occurring
after 1980 (after imaging was more widely used) were adjudicated into
HF subtypes using EF assessments by echocardiography, radionuclide Power calculations
angiography, or ventriculography during the index hospitalization or
within 1 year prior to the onset of HF if no myocardial infarction With the observed event rates, and assuming that covariates
occurred during the intervening period, as previously described.24 explain 25% of the variance of our outcomes, the present inves-
Thus, LVEF assessment for adjudication of HF subtypes was based tigation had 80% power to detect a hazard ratio (HR) of >1.24
on imaging reports and medical records obtained at the time of HF and >1.81 for the associations of HF subtypes with contin-
hospitalization. These assessments of LV function were distinct from uous eGFR and CKD, respectively. The power was 80% to
the echocardiograms systematically obtained at FHS study visits, which detect a HR of >1.50 for continuous UACR and >2.16 for
were used for cross-sectional analyses to examine cardiac structure microalbuminuria.
and function in relation to kidney function measures prior to HF onset.
HF events were classified as HFpEF if LVEF was >45%, and as HFrEF if
LVEF was ≤45%. Results
Baseline characteristics of the study samples by CKD and
microalbuminuria status are displayed in Table 1. The individ-
Statistical analysis uals with CKD were older, more likely to be women, had
Baseline characteristics for the samples were displayed according to the higher BMI and total cholesterol, and were more likely to have
presence of CKD or microalbuminuria. Age- and sex-standardized inci- hypertension and diabetes, when compared with those without
dence rates (by 10-year age strata) for HF and each HF subtype were
CKD. Among the 2912 participants included in the microal-
calculated according to the presence of CKD or microalbuminuria with
buminuria analyses, 499 (17%) had microalbuminuria. When
direct standardization using the internal study sample. To estimate the
cumulative incidence of HFpEF and HFrEF, we used a Kaplan–Meier-like compared with participants without, individuals with microalbu-
approach to account for competing risks of death, other HF subtype, minuria were older and more likely to have hypertension and
and unclassified HF.36 diabetes.
We used Cox proportional hazards regression models to assess
the associations of eGFR and albuminuria with incident HF and HF
subtypes after ensuring that the proportionality assumption was met. Chronic kidney disease is associated
To account for correlated observations within individuals, we used
robust sandwich covariance estimators.37 The predictors included
with incident overall heart failure
CKD (yes/no), eGFR (continuous, natural log-transformed), microal- Over a mean observation time of 13 ± 4 years, 598 (6.6%) individu-
buminuria (yes/no), and UACR (continuous, natural log-transformed). als without CKD and 156 (17.2%) individuals with CKD developed
Models were adjusted for age and sex, and then additionally for incident HF. The age- and sex-standardized incidence rate of HF
smoking, BMI, total/HDL-cholesterol, diabetes, valvular heart dis- was nearly three-fold higher for those with CKD compared with
ease, systolic blood pressure, hypertension treatment status, and those without CKD (17.4 vs. 5.9 events per 1000 person-years).
history of myocardial infarction. For modelling the associations of
Of 754 HF events, 324 (43%) were HFpEF, 326 (43%) HFrEF,
the predictors with each HF subtype, we used the cause-specific
and 104 (14%) were unclassified (baseline characteristics by HF
method to account for the competing risks of death, other HF sub-
type, and of unclassified HF. In secondary analyses, we evaluated
subtype are shown in Supplementary material online, Table S1).
whether the associations of CKD and albuminuria with HF sub- Incidence rates of both HFpEF and HFrEF were higher among
types were modified by systolic blood pressure or diabetes status individuals with CKD compared with those without CKD, although
by introducing multiplicative interaction terms into the multivariable rates of HFpEF and HFrEF appeared similar within the two groups
models. (Table 2; Figure 1A).

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
618 M. Nayor et al.

Table 1 Baseline clinical characteristics of the study samples

Characteristic No CKD CKD

(n = 8983) (n = 906)
...........................................................................................................................................
Age, years 56 ± 14 69 ± 10
Women, n (%) 4816 (54) 588 (65)
Body mass index, kg/m2 26.6 ± 4.7 27.0 ± 4.8
Systolic blood pressure, mmHg 128 ± 19 138 ± 21
Hypertension treatment, n (%) 1970 (22) 407 (45)
Current smoker, n (%) 2267 (25) 136 (15)
Diabetes mellitus, n (%) 501 (6) 102 (11)
Valvular heart disease, n (%) 231 (3) 55 (6)
History of myocardial infarction, n (%) 322 (4) 83 (9)
Total cholesterol, mg/dL 211 ± 41 220 ± 44
HDL-cholesterol, mg/dL 50 ± 16 50 ± 16
eGFR, mL/min/1.73 m2 94 ± 20 51 ± 8
...........................................................................................................................................
No microalbuminuria Microalbuminuria
(n = 2413) (n = 499)
...........................................................................................................................................
Age, years 58 ± 9 63 ± 10
Women, n (%) 1294 (54) 260 (52)
Body mass index, kg/m2 27.9 ± 5.1 27.9 ± 5.6
Systolic blood pressure, mmHg 127 ± 18 134 ± 21
Hypertension treatment, n (%) 595 (25) 197 (39)
Current smoker, n (%) 347 (14) 97 (19)
Diabetes mellitus, n (%) 163 (7) 103 (21)
Valvular heart disease, n (%) 58 (2) 22 (4)
History of myocardial infarction, n (%) 81 (3) 31 (6)
Chronic kidney disease, n (%) 156 (7) 78 (18)
Total cholesterol, mg/dL 206 ± 38 220 ± 44
HDL-cholesterol, mg/dL 51 ± 16 49 ± 17
eGFR, mL/min/1.73 m2 87 ± 18 81 ± 20
UACR, mg/g 4.98 (2.26, 9.33) 39.84 (27.40, 68.73)

Values represent the mean ± SD for continuous variables, except UACR, which is represented as median (25%, 75%), and number (%) for categorical variables.
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; UACR, urinary albumin to creatinine ratio.

Table 2 Age- and sex-standardized incidence of heart failure subtype by baseline chronic kidney disease and
microalbuminuria status

No CKD (n at risk = 8983) CKD (n at risk = 906)

...................................................... .....................................................
Outcome No. of events Incidence rate (95% CI) No. of events Incidence rate (95% CI)
...........................................................................................................................................
Overall HF 598 5.9 (5.4–6.4) 156 17.4 (14.5–20.2)
HFpEF 264 2.6 (2.3–3.0) 60 6.6 (4.9–8.3)
HFrEF 263 2.5 (2.2–2.8) 63 6.9 (5.1–8.6)
Unclassified HF 71 0.8 (0.6–0.9) 33 3.9 (2.5–5.3)
...........................................................................................................................................
No microalbuminuria (n at risk = 2413) Microalbuminuria (n at risk = 499)
...........................................................................................................................................
Overall HF 122 3.6 (3.0–4.3) 70 12.7 (9.6–15.8)
HFpEF 64 2.0 (1.5–2.4) 27 4.4 (2.7–6.0)
HFrEF 54 1.6 (1.2–2.0) 39 7.5 (5.0–9.9)
Unclassified HF 4 0.1 (0.0–0.2) 4 0.9 (0.0–1.8)

Incidence rate is per 1000 person-years.

CI, confience interval; CKD, chronic kidney disease; HF, heart failure; HFpEF, HF with preserved ejection fraction; HFrEF, HF with reduced ejection fraction.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Kidney disease and heart failure subtypes 619

HFpEF was not significantly associated with eGFR in age- and

........................................................................................................................................................................
A sex-adjusted or multivariable-adjusted models (Table 3).
8%
CKD HFrEF
CKD HFpEF
Cumulative incidence of HF

6%
No CKD HFrEF Microalbuminuria was associated
No CKD HFpEF
with the incidence of overall heart failure
4% and heart failure with reduced ejection
fracion but not heart failure
2% with preserved ejection fraction
Of the 2912 individuals with available UACR, there were 122
0% (5.0%) HF events among individuals without microalbuminuria
0 2 4 6 8 10 12
Number at risk and 70 (14.0%) events among those with microalbuminuria
Years
No CKD 8983 8845 8596 8318 7994 7644 7281 during 15 ± 4 years of follow-up. Participants with microalbumin-
CKD 906 854 783 711 648 596 511 uria had a 3.5-fold higher age- and sex-standardized incidence
rate of overall HF compared with those without microalbu-
B 8%
minuria (12.7 vs. 3.6 events per 1000 person-years). Of 192
MALB HFREF
Microalbuminuria HFrEF
HF events, 91 (47.3%) were classified as HFpEF, 93 (48.4%)
Cumulative incidence of HF

MALB HFPEF
Microalbuminuria HFpEF
6% No MALB HFREF
microalbuminuria HFrEF
microalbuminuria
No MALB HFPEF
4%
2%
0%
0 2 4 6
Years
Number at risk
No microalbuminuria 2413 2394 2351
Microalbuminuria 499 486 460
Figure 1 Cumulative incidence of heart failure (HF) subtype by
(A) chronic kidney disease (CKD) status and (B) microalbuminuria
status. HFpEF, HF with preserved ejection fraction; HFrEF, HF
with reduced ejection fraction.
In Cox regression models, the presence of CKD was associated
with a 52% increased risk for HF in age- and sex-adjusted analyses
[95% confidence interval (CI) 1.25–1.85, P < 0.0001], which was
partially attenuated upon adjustment for clinical HF risk factors
(HR 1.24, 95% CI 1.01–1.51, P = 0.04) (Table 3). CKD was not
associated with incident HFpEF, but was associated with a 66%
increased risk of HFrEF (95% CI 1.22–2.26, P = 0.001) in age- and
sex-adjusted analyses, which did not reach statistical significance
upon additional adjustment for clinical HF risk factors (HR 1.36,
95% CI 0.99–1.86, P = 0.06).
When examining eGFR as a continuous variable, we found simi-
lar associations. Each 1 SD higher log(eGFR) was associated with a
14% reduction in the risk of overall HF in age- and sex-adjusted
analyses (95% CI 0.79–0.93, P = 0.006), but the association did
not reach statistical significance after multivariable adjustment (HR
0.93, 95% CI 0.86–1.01, P = 0.08). Estimated GFR was inversely
associated with the incidence of HFrEF [HR 0.83 for each 1
SD higher log(eGFR), 95% CI 0.73–0.95, P = 0.005] in age- and
sex-adjusted analyses, which was not statistically significant upon
multivariable adjustment (HR 0.90, 95% CI 0.80–1.02, P = 0.10).
as HFrEF, and 8 (4.2%) were unclassified. Microalbuminuria
HFpEF
was associated with a nearly five-fold higher incidence rate of
HFrEF and an over two-fold higher incidence rate of HFpEF
(Table 2; Figure 1B).
In multivariable-adjusted Cox models, we observed a more
than two-fold higher risk of overall HF with microalbuminuria
in age- and sex-adjusted analyses (HR 2.22, 95% CI 1.64–2.99,
8 10 12 P < 0.0001). Microalbuminuria remained significantly associated
with risk of overall HF after accounting for potential clinical
2305 2240 2169 2072
confounders, including smoking, BMI, total/HDL-cholesterol, dia-
438 414 390 342
betes, valvular heart disease, systolic blood pressure, hyper-
tension treatment, and history of myocardial infarction (HR
1.71, 95% CI 1.25–2.34, P < 0.0001) (Table 3). When examin-
ing HF subtypes, microalbuminuria was associated with incident
HFpEF in age- and sex-adjusted (HR 1.62, 95% CI 1.02–2.55,
P = 0.04) but not multivariable-adjusted models (HR 1.26, 95%
CI 0.78–2.03, P = 0.34). In contrast, microalbuminuria was asso-
ciated with an over two-fold increased risk of HFrEF in both
age- and sex-adjusted (HR 2.89, 95% CI 1.89–4.41, P < 0.0001)
and multivariable-adjusted (HR 2.10, 95% CI 1.35–3.26, P = 0.01)
analyses.
When analysed as a continuous variable, UACR was associated
with an increased risk of overall HF and both HF subtypes in age-
and sex-adjusted and multivariable-adjusted analyses (Table 3).
Associations of chronic kidney disease
and microalbuminuria
with echocardiographic traits
We examined the cross-sectional associations of CKD with
echocardiographic traits for 3075 participants with available eGFR
and echocardiographic data (see Supplementary material online,
Table S2 for baseline characteristics). CKD and reduced eGFR were
not associated with LVM, LV systolic function, LVH, or LAE. A statis-
tically significant association of higher eGFR with larger LAD (beta
estimate for each 1 SD higher eGFR: 0.03 ± 0.01 cm, P = 0.005) was
observed.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
620 M. Nayor et al.

Table 3 Associations of incident heart failure subtypes with kidney function

Predictor Outcome No. of events/no. at risk

Age- and sex-adjusted Multivariable adjusteda
..................................... ......................................
HR (95% CI) P-value HR (95% CI) P-value
...........................................................................................................................................
CKD Overall HF 754/9889 1.52 (1.25–1.85) <0.0001 1.24 (1.01–1.51) 0.04
HFpEF 324/9889 1.32 (0.97–1.79) 0.08 1.10 (0.81–1.51) 0.53
HFrEF 326/9889 1.66 (1.22–2.26) 0.001 1.36 (0.99–1.86) 0.06
eGFR Overall HF 754/9889 0.86 (0.79–0.93) 0.006 0.93 (0.86–1.01) 0.08
HFpEF 324/9889 0.91 (0.81–1.03) 0.13 0.08 (0.86–1.10) 0.69
HFrEF 326/9889 0.83 (0.73–0.95) 0.005 0.90 (0.80–1.02) 0.10
Microalbuminuria Overall HF 192/2912 2.22 (1.64–2.99) <0.0001 1.71 (1.25–2.34) <0.0001
HFpEF 91/2912 1.62 (1.02–2.55) 0.04 1.26 (0.78–2.03) 0.34
HFrEF 93/2912 2.89 (1.89–4.41) <0.0001 2.10 (1.35–3.26) 0.01
UACR Overall HF 192/2912 1.78 (1.53–2.08) <0.0001 1.47 (1.27–1.71) <0.0001
HFpEF 91/2912 1.60 (1.27–2.02) <0.0001 1.39 (1.11–1.74) 0.004
HFrEF 93/2912 1.92 (1.56–2.36) <0.0001 1.48 (1.21–1.81) 0.001

Hazard ratios represent the relative hazard associated with having CKD or microalbuminuria, and the relative hazard associated with each 1 SD higher log-transformed eGFR
and UACR.
CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; HF, heart failure; HFpEF, HF with preserved ejection fraction; HFrEF, HF with
reduced ejection fraction; HR, hazard ratio; UACR, urinary albumin to creatinine ratio.
a Multivariable model is adjusted for age, sex, smoking, body mass index, total/HDL-cholesterol, diabetes, valvular heart disease, systolic blood pressure, hypertension treatment,

and history of myocardial infarction.

In contrast, in the 2652 individuals with available echocardio- albuminuria may specifically be linked to systolic dysfunction and
................................................................................................

graphic data and UACR measured, UACR [odds ratio (OR) 1.48 future HFrEF.
for each 1 SD increase in UACR, 95% CI 1.14–1.90, P = 0.003] Multiple studies have demonstrated that albuminuria predicts
and microalbuminuria (OR 3.19, 95% CI 1.67–6.09, P = 0.0004) incident HF.12,13,39,40 In contrast, specific associations with HFrEF
were associated with higher odds of having LV systolic dysfunction vs. HFpEF remain less clear. We observed a statistically significant
(Tables 4 and 5). association between baseline microalbuminuria and incident HFrEF
but not HFpEF. In contrast to our findings, investigators from the
PREVEND (Prevention of Renal and Vascular End-stage Disease)
Secondary analyses cohort showed that higher urinary albumin excretion was more
In secondary analyses, we tested for statistical interactions with strongly associated with incident HFpEF compared with HFrEF.23
systolic blood pressure and diabetes in the primary analyses. We Several differences in study cohorts and design might account for
found no evidence that the effect of CKD or microalbuminuria the discrepant results between the PREVEND study and ours. In
on either HF subtype was modified by diabetes or systolic blood the present investigation, HFpEF and HFrEF each accounted for
pressure (P > 0.05 for all interactions). ∼50% of HF events; in PREVEND, 34% of events were charac-
In exploratory analyses, we formally tested for equality of the terized as HFpEF and 66% were HFrEF. The varying distribution
associations of CKD and microalbuminuria with incident HFrEF in HF types might reflect inherent differences between the two
using the Lunn–McNeil technique.38 The P-value for this test study samples, including older mean age in PREVEND. Further-
was 0.07 for the age- and sex-adjusted models and 0.13 for the more, the PREVEND cohort was designed to overselect individuals
multivariable model, and therefore the null hypothesis of equality with elevated urinary albumin excretion, and, although statistical
was not rejected. techniques were used to account for the probability of selection,
this overselection might contribute to our discordant findings.
The association of albuminuria with incident HFrEF was substan-
Discussion tiated in our study by the association with LV systolic dysfunction
In our community-based investigation, both CKD and microalbu- cross-sectionally. Indeed, numerous previous studies have demon-
minuria were associated with an increased risk of incident overall strated associations of albuminuria with changes to cardiac struc-
HF in adjusted analyses. When HF subtypes were examined ture and function including LV systolic dysfunction, higher LVM, and
separately, microalbuminuria was associated with a greater than LVH.41,42
two-fold increased risk of incident HFrEF but not HFpEF. Sup- The mechanisms linking albuminuria with HF are not well
porting our observation, we found that microalbuminuria (but not elucidated.41 However, because albuminuria is thought to reflect
CKD) was associated with LV systolic dysfunction cross-sectionally. endothelial dysfunction and inflammation,43 we may have expected
In contrast, CKD was not associated with either HF subtype stronger associations of albuminuria with HFpEF, which has been
in multivariable-adjusted models. Our findings suggest that postulated to result from these conditions.44 Our findings suggest

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Kidney disease and heart failure subtypes 621

Table 4 Multivariable-adjusted associations of kidney function measurements with binary echocardiographic traits

Predictor Left ventricular systolic dysfunction Left ventricular hypertrophy Left atrial enlargement
............................................
..................................... ...............................
OR (95% CI) P-value OR (95% CI) P-value OR (95% CI) P-value
...........................................................................................................................................
CKD 1.87 (0.86–4.06) 0.11 0.99 (0.69–1.40) 0.94 0.82 (0.61–1.10) 0.19
eGFR 0.86 (0.66–1.12) 0.27 1.07 (0.96–1.20) 0.22 1.09 (1.00–1.20) 0.06
Microalbuminuria 3.19 (1.67–6.09) 0.0004 1.25 (0.94–1.68) 0.13 0.92 (0.73–1.16) 0.47
UACR 1.48 (1.14–1.90) 0.003 1.06 (0.96–1.17) 0.22 0.98 (0.92–1.06) 0.68

Odds ratios represent the odds of having each binary echocardiographic trait for each 1 SD higher log-transformed eGFR and UACR, and for the presence of CKD and
microalbuminuria vs. their absence.
Analyses are adjusted for age, sex, smoking, body mass index, diabetes, valvular heart disease, systolic blood pressure, and hypertension treatment status.
CI, confidence interval; CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; OR, odds ratio; UACR, urinary albumin to creatinine ratio.

Table 5 Multivariable-adjusted associations of kidney function measurements with continuous echocardiographic

traits

Predictor Fractional shortening (%) Left ventricular mass (g) Left atrial dimension (cm)
................................... .................................. ....................................
Est 𝜷 ± SE P-value Est 𝜷 ± SE P-value Est 𝜷 ± SE P-value
...........................................................................................................................................
CKD −0.04 ± 0.43 0.92 1.93 ± 2.73 0.48 −0.07 ± 0.03 0.04
eGFR −0.10 ± 0.13 0.43 0.17 ± 0.84 0.84 0.03 ± 0.01 0.005
Microalbuminuria −0.52 ± 0.34 0.12 3.02 ± 2.13 0.16 −0.02 ± 0.03 0.53
UACR −0.22 ± 0.10 0.03 0.83 ± 0.66 0.21 0.00 ± 0.01 0.92

Beta estimates represent the mean change in echocardiographic trait for the presence of CKD and microalbuminuria vs. their absence, or for each 1 SD higher log-transformed
eGFR and UACR.
CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; SE, standard error; UACR, urinary albumin to creatinine ratio.

that additional mechanisms may contribute to the increased Previous reports of associations of CKD with echocardio-
...............................................................................

risk of HF in those with albuminuria. As a marker of diffuse graphic traits have been similarly inconsistent. Increased LVM is
microvascular and endothelial injury, albuminuria may therefore present in up to 40% of individuals with CKD, but whether this
serve as an indicator of similar end-organ changes in the heart.45 association persists after accounting for shared risk factors is
Furthermore, albuminuria serves as a potent activator of the controversial.41 Whereas some studies have found independent
renin–angiotensin–aldosterone system, which is well known to associations between CKD and LVM,49,50 others have not.51 Inves-
contribute to the pathogenesis of HFrEF. The fact that inhibitors tigations evaluating the relationships between CKD and LVEF have
of the renin–angiotensin–aldosterone system (such as ACE consistently demonstrated no association,41,49,51,52 consistent with
inhibitors) both reduce albuminuria and improve HFrEF-related our findings. However, recent studies using more sensitive mark-
outcomes further supports a pathophysiological link. ers of LV systolic function, such as longitudinal strain imaging, have
Our observed associations between CKD and incident over- demonstrated more subtle abnormalities of LV systolic function to
all HF are consistent with numerous previous reports.7,9,10,46,47 be present in HFpEF patients with CKD.53 Although we did observe
Although several prior studies have examined the relationships of an association between eGFR and higher LAD (analysed as a contin-
CKD (or reduced eGFR) with HFpEF vs. HFrEF, results have been uous variable), no statistically significant association was observed
conflicting. A report from the FHS similarly concluded that eGFR between CKD (or eGFR) and LAE (analysed as a binary variable).
was not useful in predicting either HFpEF or HFrEF.25 However, The effect estimate for the observed association was small and the
using a different sample of participants from the FHS, Lam et al. clinical relevance is therefore uncertain. Furthermore, LAD is a rel-
reported that higher serum creatinine was associated with inci- atively crude measure of left atrial size, and future studies using left
dent HFrEF, but not HFpEF.8 Concordantly, a retrospective study atrial volumetric assessments or measurements of left atrial func-
of mostly male veterans with clinical HF found that CKD was more tion might be able to elucidate further relationships between CKD
common in patients with HFrEF than in those with HFpEF.48 Con- and left atrial size and function.
versely, using data from the PREVEND cohort, Brouwers et al. A number of mechanisms are likely to contribute to the
found higher cystatin C concentrations (another method of esti- observed association of CKD and incident HF. Cardiovascular and
mating glomerular filtration) to be associated with incident HFpEF, kidney disease share similar risk factors such as age, hypertension,
but not HFrEF.23 Therefore, although reduced glomerular filtration diabetes, and obesity, and their joint impact on the heart and kid-
(and CKD) was modestly associated with overall HF, whether it neys might partially explain the observed relationships. However, in
confers higher risk to either HF subtype remains unknown. our study, the associations between CKD and overall HF persisted

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
622 M. Nayor et al.

even after accounting for these clinical risk factors, suggesting that Funding

........................................................................................................................................................................
other pathways are likely to be involved. CKD leads to volume
This work was partially supported by the National Heart,
retention, altered calcium–phosphate metabolism, hyperparathy-
Lung and Blood Institute’s Framingham Heart Study (Contracts
roidism, vitamin D deficiency, anaemia, and the accumulation of
N01-HC-25195 and HHSN268201500001I). M.N. received
uraemic toxins. These and other mechanisms related to reduced
support from NIH grant T32-HL007604 and training grant
glomerular filtration may also contribute to the pathogenesis of
U10HL110337 from the National Heart, Lung, and Blood Insti-
overall HF, but the relative contributions to each HF subtype
tute. D.S.L. is supported by a clinician-scientist award from the
remain unknown.
Canadian Institutes of Health Research. C.W.T. is supported
Several limitations of the present investigation warrant men-
by NIH grant K23-HL118529. S.C. is supported in part by
tion. The adjudication of HFpEF and HFrEF was based on a single
R00-HL-107642 and a grant from the Ellison Foundation. J.H. is
assessment of EF, which might lead to misclassification as EF (and
supported by NIH grant K23-HL116780. The views expressed in
therefore HF subtype) may change over time. Given our restricted
this manuscript are those of the authors and do not necessarily
power, meaningful associations with lower effect sizes cannot be
represent the views of the National Heart, Lung, and Blood
definitively excluded. In particular, the formal testing of equality
Institute; the National Institutes of Health; or the US Department
between the associations of CKD and microalbuminuria with inci-
of Health and Human Services.
dent HFrEF should be interpreted in the context of limited power
Conflict of interest: none declared.
when examining separate HF subtypes. Albuminuria and eGFR are
commonly used but are relatively crude assessments of renal func-
tion; more sensitive measures of renal function could be used in References
future studies to explore our findings further. Furthermore, no 1. Hillege HL, Girbes AR, de Kam PJ, Boomsma F, de Zeeuw D, Charlesworth A,
EF assessment was available in ∼12% of individuals in the CKD Hampton JR, van Veldhuisen DJ. Renal function, neurohormonal activation, and
sample and 4% in the microalbuminuria sample. As elderly individ- survival in patients with chronic heart failure. Circulation 2000;102:203–210.
2. Dries DL, Exner DV, Domanski MJ, Greenberg B, Stevenson LW. The prognostic
uals may be less likely to have EF assessed during a HF event, the implications of renal insufficiency in asymptomatic and symptomatic patients with
unclassified individuals might be more likely to have HFpEF than left ventricular systolic dysfunction. J Am Coll Cardiol 2000;35:681–689.
HFrEF. Thus, ascertainment bias with regard to unclassified events 3. Damman K, Valente MAE, Voors AA, O’Connor CM, van Veldhuisen DJ, Hillege
HL. Renal impairment, worsening renal function, and outcome in patients with
might have influenced our findings. Our echocardiography analyses heart failure: an updated meta-analysis. Eur Heart J 2014;35:455–469.
were limited to the measurements obtained at the sixth examina- 4. Jackson CE, Solomon SD, Gerstein HC, Zetterstrand S, Olofsson B, Michelson EL,
tion cycle of the Offspring cohort. Left atrial volumes and diastolic Granger CB, Swedberg K, Pfeffer MA, Yusuf S, McMurray JJ, CHARM Investigators
and Committees. Albuminuria in chronic heart failure: prevalence and prognostic
function were therefore unavailable. The study sample consisted importance. Lancet 2009;374:543–550.
of mostly white individuals of European descent, potentially lim- 5. Masson S, Latini R, Milani V, Moretti L, Rossi MG, Carbonieri E, Frisinghelli A,
iting generalizability to other populations. Lastly, given the obser- Minneci C, Valisi M, Maggioni AP, Marchioli R, Tognoni G, Tavazzi L, GISSI-HF
Investigators. Prevalence and prognostic value of elevated urinary albumin excre-
vational and post-hoc nature of this investigation, we cannot infer tion in patients with chronic heart failure: data from the GISSI-Heart Failure trial.
a causal relationship, nor can we rule out that residual confound- Circ Heart Fail 2010;3:65–72.
ing may be present. Notwithstanding these limitations, our study 6. ter Maaten JM, Damman K, Verhaar MC, Paulus WJ, Duncker DJ, Cheng C, van
Heerebeek L, Hillege HL, Lam CSP, Navis G, Voors AA. Connecting heart failure
had several strengths. In particular, it was conducted in a large, with preserved ejection fraction and renal dysfunction: the role of endothelial
community-based sample free of HF at study onset, with extensive dysfunction and inflammation. Eur J Heart Fail 2016;18:588–598.
long-term follow-up and systematic adjudication of EF. 7. Köttgen A, Russell SD, Loehr LR, Crainiceanu CM, Rosamond WD, Chang PP,
Chambless LE, Coresh J. Reduced kidney function as a risk factor for incident
In conclusion, both CKD and albuminuria were associated heart failure: the Atherosclerosis Risk in Communities (ARIC) study. J Am Soc
with incident overall HF in the community. Notably, albuminuria Nephrol 2007;18:1307–1315.
was more strongly associated with HFrEF compared with HFpEF, 8. Lam CSP, Lyass A, Kraigher-Krainer E, Massaro JM, Lee DS, Ho JE, Levy
D, Redfield MM, Pieske BM, Benjamin EJ, Vasan RS. Cardiac dysfunction and
whereas CKD did not appear differentially associated with HF sub- noncardiac dysfunction as precursors of heart failure with reduced and preserved
type. Our results suggest different pathogenic relationships linking ejection fraction in the community. Circulation 2011;124:24–30.
mechanisms of kidney dysfunction with the development of HFrEF 9. Dhingra R, Gaziano JM, Djousse L. Chronic kidney disease and the risk of heart
failure in men. Circ Heart Fail 2011;4:138–144.
vs. HFpEF. Future studies are warranted to investigate the biological 10. Fried LF, Shlipak MG, Crump C, Kronmal RA, Bleyer AJ, Gottdiener JS, Kuller
pathways involved in the association of albuminuria with HFrEF. LH, Newman AB. Renal insufficiency as a predictor of cardiovascular outcomes
and mortality in elderly individuals. J Am Coll Cardiol 2003;41:1364–1372.
11. Chae CU, Albert CM, Glynn RJ, Guralnik JM, Curhan GC. Mild renal insufficiency
and risk of congestive heart failure in men and women ≥70 years of age. Am J
Supplementary Information Cardiol 2003;92:682–686.
12. Ingelsson E, Sundström J, Lind L, Risérus U, Larsson A, Basu S, Arnlöv J.
Additional Supporting Information may be found in the online Low-grade albuminuria and the incidence of heart failure in a community-based
version of this article: cohort of elderly men. Eur Heart J 2007;28:1739–1745.
Table S1. Baseline clinical characteristics by HF subtype. 13. Blecker S, Matsushita K, Kottgen A, Loehr LR, Bertoni AG, Boulware LE, Coresh
J. High-normal albuminuria and risk of heart failure in the community. Am J Kidney
Table S2. Baseline clinical and echocardiographic characteristics Dis 2011;58:47–55.
for the echocardiography sample. 14. Bock JS, Gottlieb SS. Cardiorenal syndrome: new perspectives. Circulation
Figure S1. Inclusion and exclusion criteria for CKD samples. 2010;121:2592–2600.
15. Ronco C, McCullough P, Anker SD, Anand I, Aspromonte N, Bagshaw SM,
Figure S2. Inclusion and exclusion criteria for microalbuminuria Bellomo R, Berl T, Bobek I, Cruz DN, Daliento L, Davenport A, Haapio M,
samples. Hillege H, House AA, Katz N, Maisel A, Mankad S, Zanco P, Mebazaa A,

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology
Kidney disease and heart failure subtypes 623

Palazzuoli A, Ronco F, Shaw A, Sheinfeld G, Soni S, Vescovo G, Zamperetti the American Society of Echocardiography and the European Association of

................................................................................................................................................
N, Ponikowski P, Acute Dialysis Quality Initiative (ADQI) consensus group. Cardiovascular Imaging. J Am Soc Echocardiograph 2015;28:1–39.e14.
Cardio-renal syndromes: report from the consensus conference of the acute 33. McKee PA, Castelli WP, McNamara PM, Kannel WB. The natural history of con-
dialysis quality initiative. Eur Heart J 2010;31:703–711. gestive heart failure: the Framingham study. N Engl J Med 1971;285:1441–1446.
16. Bhatia RS, Tu JV, Lee DS, Austin PC, Fang J, Haouzi A, Gong Y, Liu PP. Outcome 34. Mosterd A, Deckers JW, Hoes AW, Nederpel A, Smeets A, Linker DT, Grobbee
of heart failure with preserved ejection fraction in a population-based study. N DE. Classification of heart failure in population based research: an assessment of
Engl J Med 2006;355:260–269. six heart failure scores. Eur J Epidemiol 1997;13:491–502.
17. Bursi F, Weston SA, Redfield MM, Jacobsen SJ, Pakhomov S, Nkomo VT, 35. Maestre A, Gil V, Gallego J, Aznar J, Mora A, Martín-Hidalgo A. Diagnostic
Meverden RA, Roger VL. Systolic and diastolic heart failure in the community. accuracy of clinical criteria for identifying systolic and diastolic heart failure:
JAMA 2006;296:2209–2216. cross-sectional study. J Eval Clin Pract 2009;15:55–61.
18. Devereux RB, Roman MJ, Liu JE, Welty TK, Lee ET, Rodeheffer R, Fabsitz RR, 36. Gaynor JJ, Feuer EJ, Tan CC, Wu DH, Little CR, Straus DJ, Clarkson BD, Brennan
Howard BV. Congestive heart failure despite normal left ventricular systolic MF. On the use of cause-specific failure and conditional failure probabilities:
function in a population-based sample: the Strong Heart Study. Am J Cardiol examples from clinical oncology data. J Am Stat Assoc 2012;88:400–409.
2000;86:1090–1096. 37. Lin DY, Wei LJ. The robust inference for the Cox proportional hazards model. J
19. Fonarow GC, Stough WG, Abraham WT, Albert NM, Gheorghiade M, Green- Am Stat Assoc 2012;84:1074–1078.
berg BH, O’Connor CM, Sun JL, Yancy CW, Young JB, OPTIMIZE-HF Inves- 38. Lunn M, McNeil D. Applying Cox regression to competing risks. Biometrics
tigators and Hospitals. Characteristics, treatments, and outcomes of patients 1995;51:524–532.
with preserved systolic function hospitalized for heart failure: a report from the 39. Velagaleti RS, Gona P, Larson MG, Wang TJ, Levy D, Benjamin EJ, Selhub J, Jacques
OPTIMIZE-HF Registry. J Am Coll Cardiol 2007;50:768–777. PF, Meigs JB, Tofler GH, Vasan RS. Multimarker approach for the prediction of
20. Redfield MM, Jacobsen SJ, Burnett JC, Mahoney DW, Bailey KR, Rodehef- heart failure incidence in the community. Circulation 2010;122:1700–1706.
fer RJ. Burden of systolic and diastolic ventricular dysfunction in the com- 40. Bahrami H, Bluemke DA, Kronmal R, Bertoni AG, Lloyd-Jones DM, Shahar E,
munity: appreciating the scope of the heart failure epidemic. JAMA 2003;289: Szklo M, Lima JA. Novel metabolic risk factors for incident heart failure and their
194–202. relationship with obesity. J Am Coll Cardiol 2008;51:1775–1783.
21. De Keulenaer GW, Brutsaert DL. Systolic and diastolic heart failure 41. Matsushita K, Ballew SH, Coresh J. Influence of chronic kidney disease on cardiac
are overlapping phenotypes within the heart failure spectrum. Circulation structure and function. Curr Hypertens Rep 2015;17:1–9.
2011;123:1996–2004. 42. Lieb W. Association of low-grade urinary albumin excretion with left ventricular
22. Borlaug BA, Redfield MM. Diastolic and systolic heart failure are distinct pheno- hypertrophy in the general population: the MONICA/KORA Augsburg Echocar-
types within the heart failure spectrum. Circulation 2011;123:2006–2013. diographic Substudy. Nephrol Dial Transplant 2006;21:2780–2787.
23. Brouwers FP, de Boer RA, van der Harst P, Voors AA, Gansevoort RT, Bakker 43. Delles C, Currie G. Proteinuria and its relation to cardiovascular disease. Int J
SJ, Hillege HL, van Veldhuisen DJ, van Gilst WH. Incidence and epidemiol- Nephrol Renovasc Dis 2014;7:13–24.
ogy of new onset heart failure with preserved vs. reduced ejection fraction 44. Paulus WJ, Tschope C. A novel paradigm for heart failure with preserved ejection
in a community-based cohort: 11-year follow-up of PREVEND. Eur Heart J fraction. J Am Coll Cardiol 2013;62:263–271.
2013;34:1424–1431. 45. Dobre D, Nimade S, de Zeeuw D. Albuminuria in heart failure: what do we really
24. Ho JE, Lyass A, Lee DS, Vasan RS, Kannel WB, Larson MG, Levy D. Predictors of know? Curr Opin Cardiol 2009;24:148–154.
new-onset heart failure: differences in preserved versus reduced ejection fraction. 46. Peralta CA, Katz R, Sarnak MJ, Ix J, Fried LF, De Boer I, Palmas W, Siscovick D,
Circ Heart Fail 2013;6:279–286. Levey AS, Shlipak MG. Cystatin C identifies chronic kidney disease patients at
25. Ho JE, Gona P, Pencina MJ, Tu JV, Austin PC, Vasan RS, Kannel WB, D’Agostino higher risk for complications. J Am Soc Nephrol 2011;22:147–155.
RB, Lee DS, Levy D. Discriminating clinical features of heart failure with preserved 47. Sarnak MJ, Katz R, Stehman-Breen CO, Fried LF, Jenny NS, Psaty BM, New-
vs. reduced ejection fraction in the community. Eur Heart J 2012;33:1734–1741. man AB, Siscovick D, Shlipak MG, Cardiovascular Health Study. Cystatin C
26. Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE, Drazner MH, Fonarow concentration as a risk factor for heart failure in older adults. Ann Intern Med
GC, Geraci SA, Horwich T, Januzzi JL, Johnson MR, Kasper EK, Levy WC, 2005;142:497–505.
Masoudi FA, McBride PE, McMurray JJ, Mitchell JE, Peterson PN, Riegel B, Sam F, 48. Ather S, Chan W, Bozkurt B, Aguilar D, Ramasubbu K, Zachariah AA, Wehrens
Stevenson LW, Tang WH, Tsai EJ, Wilkoff BL. 2013 ACCF/AHA Guideline for the XH, Deswal A. Impact of noncardiac comorbidities on morbidity and mortality in
Management of Heart Failure: a report of the American College of Cardiology a predominantly male population with heart failure and preserved versus reduced
Foundation/American Heart Association Task Force on Practice Guidelines. ejection fraction. J Am Coll Cardiol 2012;59:998–1005.
Circulation 2013;128:e240–e327. 49. Park M, Hsu CY, Li Y, Mishra RK, Keane M, Rosas SE, Dries D, Xie D, Chen
27. Dawber TR, Meadors GF, Moore FE. Epidemiological approaches to heart dis- J, He J, Anderson A, Go AS, Shlipak MG, Chronic Renal Insufficiency Cohort
ease: the Framingham Study. Am J Public Health Nations Health 1951;41:279–281. (CRIC) Study Group. Associations between kidney function and subclinical
28. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation cardiac abnormalities in CKD. J Am Soc Nephrol 2012;23:1725–1734.
of coronary heart disease in families. The Framingham offspring study. Am J 50. Patel PC, Ayers CR, Murphy SA, Peshock R, Khera A, de Lemos JA, Balko JA,
Epidemiol 1979;110:281–290. Gupta S, Mammen PP, Drazner MH, Markham DW. Association of cystatin C
29. D’Agostino RB, Lee ML, Belanger AJ, Cupples LA, Anderson K, Kannel WB. with left ventricular structure and function: the Dallas Heart Study. Circ Heart Fail
Relation of pooled logistic regression to time dependent Cox regression analysis: 2009;2:98–104.
the Framingham Heart Study. Stat Med 1990;9:1501–1515. 51. Agarwal S, Thohan V, Shlipak MG, Lima J, Bluemke DA, Siscovick D, Gomes
30. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, Kusek JW, A, Herrington DM. Association between cystatin C and MRI measures of left
Eggers P, Van Lente F, Greene T, Coresh J. A new equation to estimate glomerular ventricular structure and function: multi-ethnic study of atherosclerosis. Int J
filtration rate. Ann Intern Med 2009;150:604–612. Nephrol 2011;2011:153868.
31. McMahon GM, Hwang S-J, Tanner RM, Jacques PF, Selhub J, Muntner P, Fox CS. 52. Ix JH, Shlipak MG, Chertow GM, Ali S, Schiller NB, Whooley MA. Cystatin C,
The association between vitamin B12, albuminuria and reduced kidney function: Left ventricular hypertrophy, and diastolic dysfunction: data from the Heart and
an observational cohort study. BMC Nephrol 2015;16:7–8. Soul Study. J Card Fail 2006;12:601–607.
32. Lang RM, Badano LP, Mor-Avi V, Afilalo J, Armstrong A, Ernande L, Flachskampf 53. Unger ED, Dubin RF, Deo R, Daruwalla V, Friedman JL, Medina C, Beussink L,
FA, Foster E, Goldstein SA, Kuznetsova T, Lancellotti P, Muraru D, Picard MH, Freed BH, Shah SJ. Association of chronic kidney disease with abnormal cardiac
Rietzschel ER, Rudski L, Spencer KT, Tsang W, Voigt JU. Recommendations for mechanics and adverse outcomes in patients with heart failure and preserved
cardiac chamber quantification by echocardiography in adults: an update from ejection fraction. Eur J Heart Fail 2016;18:103–112.

© 2017 The Authors

European Journal of Heart Failure © 2017 European Society of Cardiology