Abstract
Background:
The associations of kidney dysfunction and damage with heart failure with reduced and preserved ejection fraction (HFrEF and HFpEF), as well as adverse cardiac remodeling, in late-life remain incompletely understood.
Objective:
Define the associations between kidney dysfunction and damage and incident HFrEF and HFpEF and cardiac structure and function in late-life.
Methods:
This study included 5,170 adults initially free of a HF diagnosis who had estimated glomerular filtration rate (eGFR) and urine albumin-to-creatinine ratio (UACR) measured at Visit 5 (2011-13) of the Atherosclerosis Risk in Communities study. Multivariable Cox proportional hazards models were used to estimate the associations of eGFR and UACR with incident HF, HFrEF, and HFpEF through 2019. Multivariable linear regression models were used to investigate the associations of eGFR and UACR at Visit 5 with changes in cardiac structure and function between Visits 5 and 7 in 2,313 participants with available echocardiograms.
Results:
The mean age of participants was 76±5 years and 2,225 (43%) were men. The mean eGFR and median UACR were 66±18 mL/min per 1.73 m2 and 11 (25th, 75th percentile: 6-22) mg/g, respectively. In fully adjusted models, both lower eGFR and higher UACR were associated with greater risk of any HF, HFrEF and HFpEF. Lower eGFR associated with larger increases in left ventricular end-diastolic volume index and worsening of diastolic measures. UACR did not associate with changes in cardiac structure or function.
Conclusions:
Mild-moderate kidney dysfunction and damage associate with incident HF and adverse cardiac remodeling in late-life.
Keywords: Heart failure, chronic kidney disease, epidemiology, echocardiography, aging
Introduction
Both kidney dysfunction, as measured by glomerular filtration rate (GFR), and kidney damage, as measured by urine albumin-to-creatinine ratio (UACR), are associated positively and independently with incident heart failure (HF) and associate cross-sectionally with adverse cardiac remodeling in mid-life.(1-4) The prevalence of kidney dysfunction, kidney damage and HF, in particular HF with preserved ejection fraction (HFpEF), increase with age. One previous study of adults 65 years and older found reduced estimated GFR (eGFR) associated with greater risk of incident HF independent of UACR.(5) Yet, the associations of GFR and UACR with HF and HF phenotype (HFpEF and HF with reduced ejection fraction [HFrEF]) in late-life remain incompletely understood.
Several observations question the generalizability of mid-life associations between kidney function, heart structure and function and HF risk to late-life. Like the heart,(6) the kidneys can undergo structural and functional remodeling with advancing age in older individuals without sequelae of kidney damage.(7) On the other hand, single nephron GFR does increase to compensate for reduced total GFR beginning in the eighth decade of life, suggesting impaired kidney function.(8) Reduced GFR also remains a powerful predictor of all-cause and cardiovascular mortality even in adults aged 75 years or older, although the strength of the association attenuates with increasing age.(5, 9, 10)
Similarly, the heart undergoes characteristic changes during the progression from early adulthood to mid- and late-life in the absence of interval myocardial infarction or heart failure. Between young adulthood and middle age, left ventricular ejection fraction and measures of diastolic function, such as E/e’ ratio, worsen even in the absence of an interval myocardial infarction.(11) Among adults who remained free from myocardial infarction or HF between mid- and late-life, left ventricular mass and wall thickness increase while end-systolic and end-diastolic diameters decrease.(6, 12)
Furthermore, little is known about the relationships between kidney dysfunction and damage and longitudinal changes in cardiac structure and function. Studying longitudinal changes in cardiac structure and function may help disentangle the directionality of the associations between CKD and HF. We studied a biracial cohort of older, community-dwelling adults to define the associations of eGFR and UACR with risk of incident HFrEF and HFpEF and with longitudinal changes in cardiac structure and function in late-life.
Methods
The prospective Atherosclerosis Risk in Communities (ARIC) cohort study enrolled 15,792 community-dwelling adults between the ages of 45 and 64 years in 1987-1989 (Visit 1) from four communities in the United States: Forsyth County, North Carolina; Washington County, Maryland; suburban Minneapolis, Minnesota; Jackson, Mississippi. In 2011-2013, 6,538 participants returned for a fifth study visit (Visit 5), at which time they provided blood and urine samples and underwent transthoracic echocardiography. In 2018-2019, 3,408 participants attended the seventh study visit (Visit 7) at which time 2,935 underwent a repeat transthoracic echocardiogram. The ARIC study protocol was approved by the institutional review boards at all participating institutions and all participants provided written informed consent.
Study Sample
For analyses of the association of kidney function with risk of incident HFpEF and HFrEF, the study sample included participants who attended Visit 5, had no history of HF (438 excluded), and had available measures of kidney dysfunction (eGFR), kidney damage (UACR) (728 more excluded), and relevant covariates (159 more excluded). According to standard ARIC analytical procedures, Visit 5 participants were excluded if they reported race other than White or Black (n=18) due to low numbers of these individuals in ARIC or if their reported race was under-represented at their respective field center (n=25). A total of 5,170 participants were included in the analysis of HF outcomes. The analyses of the association of kidney measures with change in echocardiographic measures further excluded participants without paired echocardiograms at Visit 5 and Visit 7 and those who developed incident HF after Visit 5, leaving an analytical cohort of 2,313 participants.
Measurement of Kidney Dysfunction and Damage
The plasma creatinine- and cystatin C-based CKD-EPI 2012 equation was used to estimate GFR.(13) Plasma creatinine was measured using a creatinase enzymatic method standardized to isotope dilution mass spectrometry.(2) Cystatin C was measured using a turbidometric method standardized to international calibrator standard.(2) A random urine sample was collected to measure albumin with an immunoturbidometric method and creatinine with a creatinase enzymatic method.(2)
Echocardiography
Methods for echocardiographic assessment in ARIC at Visit 5 have been reported previously and echocardiographic methods at Visit 7 were similar.(14) Briefly, echocardiography was performed at all 4 study field centers at Visit 5 and Visit 7 by sonographers who were trained and certified in the study-specific imaging protocol. The same echocardiographic machines (Philips iE33, Koninklijke Philips, The Netherlands) and probes (Philips XMatrix) were used at all field centers at both study visits. The same echocardiography core laboratory at Brigham and Women’s Hospital (Boston, MA) performed all the quantitative measures in a blinded fashion on the basis of the recommendations of the American Society of Echocardiography.(14) The assessment of intraobserver variability and temporal drift has been reported previously for ARIC Visit 5.(14)
Outcomes
Death and HF event ascertainment and classification in ARIC have been described previously.(15) Surveillance of ARIC participants for potential events occurs through annual (or semi-annual since 2012) telephone follow-up calls, at study visits, and manual review of local hospital discharges and health department death certificates. Medical records from hospitalizations with potential HF-related ICD-9-CM or ICD-10-CM codes (as identified through participant report or surveillance) were abstracted comprehensively and subjected to physician adjudication. HFrEF was defined as an adjudicated hospitalization with HF with LVEF <50% at the time of hospitalization, while HFpEF was defined when the LVEF at the time of hospitalization was >=50%. LVEF at the time of hospitalization was not available in 69 (13%) HF events. Deaths are identified through ARIC surveillance, the National Death Index and hospital discharge lists for in-hospital deaths.
Covariates
Participants self-reported date of birth, sex and race. Hypertension was defined as anti-hypertensive medication use or study blood pressure measurements indicative of hypertension (≥140/90 mmHg). Diabetes mellitus was defined as self-report of a physician diagnosis of diabetes mellitus, glucose-lowering medication use, fasting glucose level of at least 126 mg/dL or non-fasting glucose level of at least 200 mg/dL. Prevalent HF at Visit 5 was identified (for exclusion) by reviewing hospitalization surveillance data (described above).(16) Prevalent coronary heart disease (CHD) was defined as an adjudicated coronary heart disease event prior to Visit 5 or reported history of CHD at Visit 1.(16) Atrial fibrillation was identified through review of study electrocardiograms as well as hospitalization ICD codes from surveillance data.(17) Blood pressure was measured in triplicate after 5 minutes of rest using an automated sphygmomanometer and the average of the two final readings was recorded. Pulse pressure was defined as the difference between systolic and diastolic blood pressures. Body mass index (BMI) was calculated as body weight (kilograms) divided by height (meters) squared. NT-proBNP was measured with an immunoassay (Elecys proBNP II, Roche Diagnostics, Indianapolis, Indiana). High-sensitivity cardiac troponin T measured at Visit 5 using a standard immunoassay (Elecsys Troponin T; Roche Diagnostics).
Statistical Analysis
Categorical data were summarized as numbers and percentages, and continuous data as means and standard deviations if normally distributed or medians and [25th, 75th percentile] range if non-normally distributed. Spearman’s rank correlation coefficient was used to quantify the association between eGFR and UACR.
Cox proportional hazards regression models were used to estimate cause-specific hazard ratios for the associations between eGFR and UACR with incident HF, HFpEF, HFrEF, and the composite of HF or death from any cause. Follow-up time accrued from Visit 5 to either incident event, death, loss to follow-up, or else December 31, 2019 at the Forsyth County, Hagerstown and suburban Minneapolis Field Centers or December 31, 2017 at the Jackson Field Center. eGFR scaled to a mean of 0 and standard deviation of 1 then modeled per 1-standard deviation decrease. UACR was log-transformed and then the log-transformed value was scaled to a mean of 0 and standard deviation of 1. Scaled log-transformed UACR was modeled per 1-standard deviation increase. The following Cox regression models were generated: Model 1 adjusted for demographics (age, sex, and the combination of race and field center); Model 2 additionally adjusted for potentially confounding clinical HF risk factors selected based on a priori knowledge (body mass index, diabetes mellitus, coronary heart disease, hypertension, atrial fibrillation and pulse pressure); and Model 3 additionally adjusted for UACR in analyses with eGFR as primary predictor and vice versa. We considered non-linear associations between eGFR and UACR using restricted cubic splines. The number of knots was chosen based upon minimization of the Akaike information criterion (3 to 7 knots tested) and all models with restricted cubic splines were adjusted for demographics and clinical risk factors (Model 2 above). To assess the potential impact of HF events with unknown LVEF, we performed additional analyses assigning HF events with unknown LVEF as either HFpEF or HFrEF events. In a sensitivity analysis, subdistribution hazard ratios for the associations of eGFR and UACR with incident HF, HFrEF and HFpEF with death as a competing event were estimated using Fine-Gray regression models. Effect modification by age (younger and older than the median age) and sex were assessed using interaction terms within each Cox regression model. Models with interaction terms for race were censored at December 31, 2017 because of delays in event ascertainment at the Jackson, MS field center, which enrolled only Black participants. Lastly, additional sensitivity analyses censored HF events which were preceded by a myocardial infarction and adjusted for NTproBNP.
Linear regression models were used to estimate the associations of eGFR and UACR with longitudinal changes in cardiac structure and function between Visit 5 and Visit 7. Linear regression models of echocardiographic measures at Visit 5 were adjusted for Model 3 covariates above plus Visit 5 systolic blood pressure and heart rate. Linear regression models of changes in echocardiographic measures from Visit 5 to Visit 7 were adjusted for Model 3 covariates above, Visit 5 systolic blood pressure and heart rate, Visit 7 systolic blood pressure and heart rate and the standardized residual of the Visit 5 value of the echocardiographic measure of interest. The standardized residual of the Visit 5 value of the echocardiographic measure of interest was used to account for baseline value (instead of the unadjusted Visit 5 value) to minimize inflation of the beta coefficient.(18) Linear regression model covariates from Visit 5 were otherwise the same as those listed for Models 1-3 above. P-trend were calculated from linear regression models with a numeric covariate for eGFR or UACR category. We performed the following sensitivity analyses. We repeated models without adjustment for the residual of the baseline value. We modeled the primary exposures as continuous, rather than categorical variables. To address potential bias due to non-random non-attendance at Visit 7, models were repeated incorporating inverse probability of attrition weights (IPAW). Probability weights were based upon logistic regression models with Visit 7 attendance as an outcome among participants attending Visit 5 with age, sex, race, field center, diabetes, coronary heart disease, hypertension, atrial fibrillation and pulse pressure at Visit 5 as potential predictors among participants alive through Visit 7. Analyses of changes in cardiac structure and function were repeated after excluding participants with interval MI between Visits 5 and 7 and including participants who developed HF before Visit 7. Non-linear associations were considered using restricted cubic spline functions as described above.
A P-value < 0.05 was considered statistically significant. Stata version 15.1 (College Station, TX, United States) was used for all analyses.
Results
Associations of Kidney Dysfunction and Damage with Risk of Incident Heart Failure
Among the 5,170 participants initially free of HF included in the analyses of HF outcomes, the mean age was 76±5 years, 2,225 (43%) were men and 1,217 (24%) were Black (Table 1). The mean eGFR was 66±18 mL/min per 1.73 m2 and the median UACR was 11 (25th, 75th percentiles: 6, 22) mg/g (131 with UACR ≥300 mg/g and 864 with UACR 30 to <300 mg/g). There was a modest inverse correlation between eGFR and UACR at Visit 5 (Spearman’s rho = −0.14; P<.001) and mildly-moderately elevated albuminuria was prevalent across the spectrum of eGFR values (Supplemental Figure 1). Lower eGFR and greater UACR were associated with older age, a higher proportion of Black participants, a higher prevalence of diabetes, hypertension, coronary heart disease and atrial fibrillation, wider pulse pressure and higher NT-proBNP and troponin T concentration (Supplemental Tables 1 and 2). Lower eGFR, but not higher UACR, was associated with male gender and higher body mass index.
Table 1.
Characteristics of participants free of heart failure at Visit 5 included in the incident heart failure analyses and the echocardiography analyses
| Characteristic1 | Overall Cohort (n=5170) |
Echocardiographic Cohort (n=2313) |
Excluded from Echocardiographic Cohort (n=2857) |
P-Value for Included vs. Excluded from Echocardiography Cohort |
|---|---|---|---|---|
| Age, years | 76 (5) | 74 (4) | 77 (5) | <.001 |
| Male sex, n (%) | 2225 (43%) | 1020 (44%) | 1205 (42%) | .15 |
| Black race, n (%) | 1217 (24%) | 555 (24%) | 662 (23%) | .61 |
| Diabetes, n (%) | 1956 (38%) | 761 (33%) | 1195 (42%) | <.001 |
| Hypertension, n (%) | 4334 (84%) | 1864 (81%) | 2470 (87%) | <.001 |
| Coronary heart disease, n (%) | 685 (13%) | 234 (10%) | 451 (16%) | <.001 |
| Atrial fibrillation, n (%) | 285 (6%) | 73 (3%) | 212 (7%) | <.001 |
| Body mass index, kg/m2 | 28.8 (5.7) | 28.8 (5.3) | 28.7 (6.1) | .79 |
| Systolic blood pressure, mm Hg | 131 (18) | 129 (17) | 133 (19) | <.001 |
| Diastolic blood pressure, mm Hg | 67 (11) | 67 (10) | 66 (11) | .003 |
| Pulse pressure, mm Hg | 64 (15) | 62 (13) | 67 (16) | <.001 |
| Serum creatinine, mg/dL | 0.99 (0.34) | 0.95 (0.26) | 1.01 (0.38) | <.001 |
| eGFR, mL/min per 1.73 m2 | 66 (18) | 70 (16) | 63 (18) | <.001 |
| Urine albumin-to-creatinine ratio, mg/g | 11 [6, 22] | 9 [6, 16] | 13 [7, 29] | <.001 |
| NT-proBNP, pg/mL | 129 [66, 249] | 97 [52, 174] | 164 [83, 332] | <.001 |
| Troponin T, ng/L | 1.1 [0.7, 1.6] | 1.0 [0.7, 1.3] | 1.2 [0.8, 1.8] | <.001 |
Entries are N (%), mean (SD), or median [25th percentile, 75th percentile]
ACE = angiotensin-converting enzyme; ARB = angiotensin receptor blocker; BNP = B-type natriuretic peptide; eGFR = estimated glomerular filtration rate
Over a median follow-up of 7.1 years, 519 incident HF events occurred (incidence rate [IR] 95% confidence interval [CI], 15.9 [14.6 to 17.3] per 1,000 person-years), including 204 HFrEF events (IR, 6.2 [5.4 to 7.2] per 1,000 person-years), 246 HFpEF events (7.5 [6.6 to 8.5] per 1,000 person-years) and 69 HF events with unknown LVEF. Table 2 summarizes unadjusted and adjusted associations between eGFR and incident HF outcomes. Figure 2 shows incidence rates for any heart failure across categories of eGFR and UACR. After adjustment for demographics and clinical risk factors (Model 2), each 1-standard deviation decrease in eGFR associated with an increased HR of incident HF (HR [95% CI], 1.27 [1.16 to 1.40]; P<.001), HFrEF (HR [95% CI], 1.19 [1.02-1.38]; P=.024) and HFpEF (HR [95% CI], 1.30 [1.13 to 1.49]; P<.001). Restricted cubic splines demonstrated linear relationships between eGFR and incidence of HF, incident HFrEF and incident HFpEF (Figure 1). The associations between eGFR and incident HF or HFpEF were independent of UACR and interval MI (Figure 3), whereas eGFR did not associate with HFrEF after adjustment for UACR or interval MI (Figure 3). Including HF events with an unknown LVEF in the HFrEF or HFpEF analysis or using a composite outcome of HF or all-cause death did not alter the primary results (Supplemental Table 3). Subdistribution hazard ratios for the association of eGFR with incident HF and HFpEF with death as a competing event were overall consistent with the main analysis with the exception of the association of eGFR with incident HFrEF which was no longer significant (Supplemental Table 4). Lower eGFR was no longer associated with incident HF, HFpEF or HFrEF, after further adjustment for NTproBNP (Supplemental Table 5). No effect modification of the association of eGFR with HF, HFrEF or HFpEF was observed by age category or gender (Supplemental Figure 2).
Table 2. Total events, incidence rates, unadjusted and fully adjusted hazard ratios of the association of estimated glomerular filtration rate and urine albumin-to-creatinine ratio with incident heart failure and heart failure subtypes.
Model 1 adjusted for Visit 5 age, gender, race and field center. Model 2 adjusted for Model 1 covariates plus Visit 5 body mass index, diabetes mellitus, coronary heart disease, hypertension, atrial fibrillation, and pulse pressure. Model 3 additionally adjusted for UACR (with eGFR as the primary exposure) or eGFR (with UACR as the primary exposure).
| Unadjusted | Model 1 | Model 2 | Model 3 | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Outcome | N events | IR# | HR (95% CI)* | P | HR (95% CI)* | P | HR (95% CI)* | P | HR (95% CI)* | P | |
| eGFR (per 1-SD decrease) | Any HF | 519 | 15.9 (14.6-17.3) | 1.55 (1.42-1.69) | <.001 | 1.43 (1.30-1.56) | <.001 | 1.27 (1.16-1.40) | <.001 | 1.18 (1.07-1.30) | .001 |
| HFrEF | 204 | 6.2 (5.4-7.2) | 1.43 (1.25-1.65) | <.001 | 1.32 (1.14-1.53) | .002 | 1.19 (1.02-1.38) | .024 | 1.11 (0.95-1.29) | .17 | |
| HFpEF | 246 | 7.5 (6.6-8.5) | 1.59 (1.40-1.81) | <.001 | 1.47 (1.28-1.68) | <.001 | 1.30 (1.13-1.49) | <.001 | 1.19 (1.04-1.37) | .011 | |
| Log-UACR (per 1-SD increase) | Any HF | 519 | 15.9 (14.6-17.3) | 1.55 (1.45-1.66) | <.001 | 1.47 (1.37-1.57) | <.001 | 1.34 (1.25-1.44) | <.001 | 1.30 (1.20-1.40) | <.001 |
| HFrEF | 204 | 6.2 (5.4-7.2) | 1.48 (1.33-1.65) | <.001 | 1.37 (1.23-1.54) | <.001 | 1.30 (1.16-1.46) | .001 | 1.27 (1.13-1.44) | <.001 | |
| HFpEF | 246 | 7.5 (6.6-8.5) | 1.61 (1.47-1.77) | <.001 | 1.54 (1.40-1.70) | <.001 | 1.39 (1.26-1.54) | <.001 | 1.35 (1.21-1.50) | <.001 | |
CI = confidence interval; HF = heart failure; HR = hazard ratio; HFrEF = heart failure with reduced ejection fraction; HFpEF = heart failure with preserved ejection fraction; HFuEF = heart failure with unknown ejection fraction; IPAW = inverse probability of attrition weight; IR = incidence rate; SD = standard deviation
Incidence rate per 1,000 person-years
Hazard ratio per each 1-standard deviation decrease in eGFR (−18 mL/min per 1.73 m2) or 1-standard deviation increase in log-UACR
Figure 2. Incidence rates for any heart failure across categories of estimated glomerular filtration rate and urine albumin-to-creatinine ratio.
This plot depicts additive incidence rates for any heart failure across categories of kidney function and damage.
CI = confidence interval; eGFR = estimated glomerular filtration rate; UACR = urine albumin-to-creatinine ratio
Figure 1. Estimated associations between estimated glomerular filtration rate and urine albumin-to-creatinine ratio with incident heart failure and heart failure subtypes.
Restricted cubic spline models were used to assess non-linear associations between estimated glomerular filtration rate and urine albumin-to-creatinine ratio with incident heart failure and heart failure subtypes. Models that minimized the Akaike information criterion with a P-value for non-linearity < 0.05 are shown. All models are adjusted for Visit 5 age, sex, race, field center, body mass index, diabetes mellitus, coronary heart disease, hypertension, atrial fibrillation, pulse pressure and UACR (for eGFR as primary exposure) or eGFR (for UACR as primary exposure).
eGFR = estimated glomerular filtration rate; HF = heart failure; HFpEF = heart failure with preserved ejection fraction; HFrEF = heart failure with reduced ejection fraction; UACR = urine albumin-to-creatinine ratio
Figure 3. Hazard ratios for the associations between estimated glomerular filtration rate and urine albumin-to-creatinine ratio with heart failure and heart failure subtypes with various model adjustments.
These forest plots depict the associations between estimated glomerular filtration rate and urine albumin-to-creatinine ratio at Visit 5 with incident heart failure (Panel A), heart failure with reduced ejection fraction (Panel B) and heart failure with preserved ejection fraction (Panel C) after adjustment for demographics and clinical characteristics, demographics, clinical characteristics and either eGFR or UACR and after excluding participants with interval myocardial infarction between Visit 5 and Visit 7. Hazard ratios modeled per 1-standard deviation increase in eGFR or decrease in UACR.
CI = confidence interval; eGFR = estimated glomerular filtration rate; HF = heart failure; HFpEF = heart failure with preserved ejection fraction; HFrEF = heart failure with reduced ejection fraction; HR = hazard ratio; MI = myocardial infarction; UACR = urine albumin-to-creatinine ratio
Table 2 summarizes unadjusted and adjusted associations between UACR and incident HF outcomes. In models adjusting for demographics and clinical HF risk factors (Model 2), each 1-standard deviation increase in log-transformed UACR was associated with heightened HR of incident HF (HR [95% CI], 1.34 [1.25 to 1.44]; P<.001), incident HFrEF (HR [95% CI], 1.30 [1.16 to 1.46]; P=.001) and incident HFpEF (HR [95% CI], 1.39 [1.26 to 1.54; P<.001). Restricted cubic splines demonstrated non-linear relationships between UACR and incident HF, HFpEF and HFrEF with steeper increase in incidence with increasing UACR at lower UACR values (Figure 1). The associations between UACR and incident HF, HFrEF, and HFpEF persisted after additional adjustment for eGFR and in analyses censoring participants with interval MI (Figure 3). Models assigning HF events with unknown LVEF as either HFpEF or as HFrEF events or using the composite outcome of HF or all-cause death demonstrated similar results as the primary models (Supplemental Table 3). Subdistribution hazard ratios for the association of UACR with incident HF, HFrEF and HFpEF with death as a competing event were overall consistent with the primary analysis (Supplemental Table 4). Higher UACR remained associated with incident HF, HFrEF, and HFpEF after further adjustment for NTproBNP (Supplemental Table 5). No effect modification of the association of UACR with any HF outcome was observed by age or gender (Supplemental Figure 2).
Associations of Kidney Dysfunction and Damage with Longitudinal Changes in Cardiac Structure and Function
Among the 2,313 ARIC participants who completed echocardiograms at Visit 5 and Visit 7 and free of HF at Visit 7, the mean age at Visit 5 was 74±4 years and at Visit 7 was 81±4 years, 1,020 (44%) were men and 555 (24%) were Black Americans (Table 1). The mean eGFR was 70±16 mL/min per 1.73 m2 and the median (25th, 75th percentiles) UACR was 9 (6, 16) mg/g. Compared to participants without paired echocardiograms at Visit 5 and Visit 7, those with paired echocardiograms were at Visit 5 younger, had higher eGFR and lower UACR, and had lower prevalences of diabetes, hypertension, coronary heart disease and atrial fibrillation (Table 1). Lower eGFR and higher UACR at Visit 5 were each associated with a higher proportion of participants who died between Visit 5 and Visit 7 or who did not attend Visit 7 (Supplemental Table 6).
In fully adjusted models, lower eGFR associated with greater increases in LV end-diastolic volume index, E wave velocity and LA volume index between Visits 5 and 7 (Table 3). The association between lower eGFR and a greater increase in LV end-diastolic volume index was not significant after adjustment for interval MI. Lower eGFR associated with lower LVEF and higher E/e’ ratio in analyses that included participants with interval HF, but not without interval HF (Supplemental Table 7). Otherwise, the results were consistent in sensitivity analyses that excluded the adjustment for the residual of the baseline value, modeled eGFR as a continuous variable, adjusted for the probability of Visit 7 attendance or excluded participants with an interval MI (Supplemental Table 7). Modeling eGFR as a restricted cubic spline function supported non-linear associations with LV end-diastolic volume index, LV mass index, change in LV mass index, E/e’ ratio, pulmonary artery systolic pressure, change in pulmonary artery systolic pressure, and tricuspid annular plane systolic excursion (Supplemental Figure 3).
Table 3. Longitudinal changes in cardiac structure and function across categories of estimated glomerular filtration rate.
Models are adjusted for age, gender, race, field center, body mass index, diabetes mellitus, coronary heart disease, hypertension, atrial fibrillation, pulse pressure, UACR, Visit 5 systolic blood pressure and heart rate, and, for longitudinal analyses, Visit 7 systolic blood pressure and heart rate and the residual of the baseline value.
| eGFR (mL/min per 1.73 m2) | |||||||
|---|---|---|---|---|---|---|---|
| Outcome | ≥90 (n=241) | 60 to < 90 (n=1429) | 30 to < 60 (n=625) | < 30 (n=18) | Beta (95% CI) Per Lower eGFR Category |
P | |
| EDVI, mL/m2 | Visit 5 | 44.98 (9.82) | 44.11 (9.97) | 42.63 (10.08) | 47.04 (12.49) | −1.047 (−1.628 to −0.466) | <0.001 |
| Change (Absolute) | 1.48 (8.93) | 1.99 (7.67) | 2.87 (9.21) | 6.76 (7.79) | 0.598 (0.076 to 1.120) | 0.025 | |
| Change (%) | 0.05 (0.21) | 0.06 (0.18) | 0.09 (0.21) | 0.15 (0.18) | -- | -- | |
| MWT, cm | Visit 5 | 0.98 (0.14) | 0.97 (0.13) | 0.99 (0.13) | 1.04 (0.17) | −0.001 (−0.009 to 0.007) | 0.82 |
| Change (Absolute) | 0.04 (0.12) | 0.05 (0.11) | 0.05 (0.12) | 0.03 (0.17) | −0.001 (−0.008 to 0.007) | 0.84 | |
| Change (%) | 0.05 (0.13) | 0.06 (0.12) | 0.06 (0.13) | 0.04 (0.17) | -- | -- | |
| RWT | Visit 5 | 0.43 (0.08) | 0.42 (0.07) | 0.43 (0.07) | 0.45 (0.08) | 0.004 (−0.001 to 0.009) | 0.10 |
| Change (Absolute) | 0.03 (0.09) | 0.03 (0.08) | 0.03 (0.09) | −0.01 (0.10) | −0.002 (−0.007 to 0.004) | 0.51 | |
| Change (%) | 0.10 (0.23) | 0.08 (0.20) | 0.09 (0.22) | 0.01 (0.20) | -- | -- | |
| LVMI, g/m2 | Visit 5 | 76.80 (17.32) | 76.61 (17.39) | 78.27 (18.81) | 84.60 (24.80) | −0.897 (−2.005 to 0.211) | 0.11 |
| Change (Absolute) | 2.90 (14.66) | 4.36 (14.11) | 5.38 (17.31) | 12.85 (29.26) | 0.516 (−0.466 to 1.499) | 0.30 | |
| Change (%) | 0.05 (0.20) | 0.07 (0.19) | 0.08 (0.22) | 0.18 (0.30) | -- | -- | |
| EF | Visit 5 | 65.23 (5.85) | 65.32 (5.58) | 65.74 (6.08) | 65.33 (4.45) | 0.303 (−0.063 to 0.670) | 0.11 |
| Change (Absolute) | −0.73 (7.39) | −1.84 (6.78) | −2.33 (7.48) | −4.41 (10.11) | −0.291 (−0.707 to 0.124) | 0.17 | |
| Change (%) | −0.01 (0.12) | −0.03 (0.11) | −0.03 (0.13) | −0.07 (0.16) | -- | -- | |
| LS | Visit 5 | −17.87 (2.40) | −18.04 (2.39) | −18.00 (2.40) | −18.75 (1.94) | −0.113 (−0.269 to 0.044) | 0.16 |
| Change (Absolute) | −0.28 (2.80) | −0.52 (2.84) | −0.69 (3.03) | −1.63 (3.73) | −0.078 (−0.251 to 0.095) | 0.38 | |
| Change (%) | 0.00 (0.17) | 0.00 (0.66) | 0.03 (0.18) | 0.08 (0.21) | -- | -- | |
| CS | Visit 5 | −27.71 (3.52) | −27.80 (3.60) | −27.78 (3.86) | −28.10 (3.02) | 0.062 (−0.207 to 0.331) | 0.65 |
| Change (Absolute) | −0.41 (4.20) | −0.46 (4.14) | −0.97 (4.44) | −2.04 (5.10) | 0.045 (−0.232 to 0.323) | 0.75 | |
| Change (%) | 0.00 (0.16) | 0.01 (0.16) | 0.02 (0.17) | 0.06 (0.18) | -- | -- | |
| E wave, cm/s | Visit 5 | 65.98 (15.49) | 65.84 (16.97) | 67.15 (18.42) | 71.29 (19.99) | 0.058 (−1.078 to 1.193) | 0.92 |
| Change (Absolute) | 6.20 (17.09) | 8.30 (18.14) | 10.77 (19.98) | 14.56 (17.88) | 1.396 (0.201 to 2.592) | 0.022 | |
| Change (%) | 0.12 (0.27) | 0.16 (0.32) | 0.19 (0.32) | 0.27 (0.39) | -- | -- | |
| Septal e’, cm/s | Visit 5 | 5.75 (1.43) | 5.85 (1.46) | 5.65 (1.38) | 5.45 (1.25) | 0.025 (−0.074 to 0.124) | 0.62 |
| Change (Absolute) | −0.38 (1.66) | −0.57 (1.51) | −0.44 (1.56) | −0.25 (1.64) | −0.015 (−0.102 to 0.072) | 0.74 | |
| Change (%) | −0.03 (0.32) | −0.07 (0.25) | −0.05 (0.29) | −0.01 (0.30) | -- | -- | |
| E/e’ ratio | Visit 5 | 12.05 (3.93) | 11.73 (3.60) | 12.44 (4.34) | 13.51 (4.29) | −0.069 (−0.311 to 0.174) | 0.58 |
| Change (Absolute) | 2.33 (4.73) | 3.07 (4.62) | 3.37 (4.90) | 4.14 (5.84) | 0.292 (−0.009 to 0.593) | 0.057 | |
| Change (%) | 0.24 (0.40) | 0.31 (0.43) | 0.32 (0.42) | 0.37 (0.44) | -- | -- | |
| LAVI, mL/m2 | Visit 5 | 24.81 (6.32) | 25.13 (7.51) | 25.18 (7.61) | 27.57 (7.46) | −0.550 (−1.027 to −0.073) | 0.024 |
| Change (Absolute) | 1.33 (6.69) | 2.39 (6.86) | 3.67 (7.38) | 6.77 (8.44) | 0.713 (0.230 to 1.197) | 0.004 | |
| Change (%) | 0.09 (0.33) | 0.12 (0.28) | 0.17 (0.30) | 0.28 (0.35) | -- | -- | |
| RV FAC, % | Visit 5 | 0.52 (0.07) | 0.52 (0.07) | 0.52 (0.08) | 0.54 (0.07) | 0.003 (−0.002 to 0.008) | 0.24 |
| Change (Absolute) | −0.03 (0.10) | −0.03 (0.09) | −0.03 (0.09) | −0.05 (0.12) | −0.001 (−0.007 to 0.005) | 0.73 | |
| Change (%) | −0.04 (0.18) | −0.05 (0.18) | −0.05 (0.17) | −0.07 (0.23) | -- | -- | |
| PASP, mm Hg | Visit 5 | 27.34 (5.60) | 26.92 (4.49) | 27.71 (5.07) | 29.37 (6.86) | 0.096 (−0.332 to 0.523) | 0.66 |
| Change (Absolute) | 4.67 (7.21) | 4.75 (7.55) | 6.32 (8.33) | 4.31 (10.48) | 0.506 (−0.216 to 1.229) | 0.17 | |
| Change (%) | 0.19 (0.28) | 0.19 (0.29) | 0.24 (0.32) | 0.15 (0.32) | -- | -- | |
| Tricuspid s’, cm/s | Visit 5 | 12.01 (2.56) | 11.72 (2.67) | 11.62 (2.80) | 11.60 (3.50) | −0.037 (−0.222 to 0.149) | 0.70 |
| Change (Absolute) | 0.42 (2.83) | 0.11 (3.05) | −0.00 (3.11) | −0.12 (3.56) | −0.123 (−0.312 to 0.065) | 0.20 | |
| Change (%) | 0.06 (0.24) | 0.04 (0.27) | 0.03 (0.27) | 0.03 (0.32) | -- | -- | |
CI = confidence interval; CS = circumferential strain; Demo. = demographics; EDVI = end-diastolic volume index; EF = ejection fraction; eGFR = estimated glomerular filtration rate; HF = heart failure; IPAW = inverse probability of attrition weighting; LS = longitudinal strain; LAVI = left atrial volume index; LVMI = left ventricular mass index; MI = myocardial infarction; MWT = mean wall thickness; PASP = pulmonary artery systolic pressure; RV FAC = right ventricular fractional area change; RWT = relative wall thickness; UACR = urine albumin-to-creatinine ratio
In fully adjusted models, higher UACR associated with a greater increase in E wave velocity and septal e’ velocity (Table 4). There was a trend towards an association between higher UACR and greater change in LV mass index that was significant in analyses that included participants with interval HF (Supplemental Table 8). Otherwise, results were consistent in sensitivity analyses that excluded the adjustment for the residual of the baseline value, modeled eGFR as a continuous variable, adjusted for the probability of Visit 7 attendance, excluded participants with an interval MI or included participants with interval HF (Supplemental Table 8). Modeling UACR as a restricted cubic spline function supported non-linear associations with several echocardiographic measures (Supplemental Figure 4).
Table 4. Longitudinal changes in cardiac structure and function across categories of urine albumin-to-creatinine ratio.
Models are adjusted for age, gender, race, field center, body mass index, diabetes mellitus, coronary heart disease, hypertension, atrial fibrillation, pulse pressure, eGFR, Visit 5 systolic blood pressure and heart rate, and, for longitudinal analyses, Visit 7 systolic blood pressure and heart rate and the residual of the baseline value. Model 3 additionally adjusted for UACR (with eGFR as the primary exposure) or eGFR (with UACR as the primary exposure).
| UACR (mg/g) | |||||||
|---|---|---|---|---|---|---|---|
| Outcome | <10 (n=1334) |
10 to <30 (n=684) |
30 to <300 (n=272) |
≥300 (n=23) | Beta (95% CI) Per Higher UACR Category |
P | |
| EDVI, mL/m2 | Visit 5 | 43.33 (9.89) | 43.98 (9.79) | 45.46 (10.73) | 46.22 (13.42) | 0.768 (0.295 to 1.241) | 0.001 |
| Change (Absolute) | 2.36 (8.17) | 1.70 (8.46) | 2.91 (8.19) | 1.99 (8.68) | −0.108 (−0.532 to 0.315) | 0.62 | |
| Change (%) | 0.07 (0.20) | 0.06 (0.19) | 0.08 (0.19) | 0.05 (0.22) | -- | -- | |
| MWT, cm | Visit 5 | 0.97 (0.12) | 0.98 (0.13) | 1.02 (0.14) | 1.07 (0.13) | 0.010 (0.003 to 0.016) | 0.003 |
| Change (Absolute) | 0.05 (0.12) | 0.05 (0.12) | 0.06 (0.13) | 0.08 (0.18) | 0.005 (−0.002 to 0.011) | 0.15 | |
| Change (%) | 0.05 (0.12) | 0.06 (0.12) | 0.07 (0.13) | 0.08 (0.19) | -- | -- | |
| RWT | Visit 5 | 0.42 (0.07) | 0.42 (0.07) | 0.44 (0.08) | 0.45 (0.08) | 0.003 (−0.001 to 0.007) | 0.13 |
| Change (Absolute) | 0.03 (0.08) | 0.03 (0.09) | 0.04 (0.09) | 0.04 (0.16) | 0.003 (−0.001 to 0.008) | 0.18 | |
| Change (%) | 0.08 (0.20) | 0.09 (0.22) | 0.10 (0.23) | 0.12 (0.40) | -- | -- | |
| LVMI, g/m2 | Visit 5 | 75.61 (17.28) | 77.82 (17.31) | 81.68 (20.37) | 89.66 (20.96) | 1.692 (0.787 to 2.597) | <0.001 |
| Change (Absolute) | 4.04 (14.62) | 4.52 (15.28) | 6.67 (17.19) | 10.82 (25.31) | 0.757 (−0.050 to 1.565) | 0.066 | |
| Change (%) | 0.07 (0.20) | 0.07 (0.20) | 0.10 (0.21) | 0.13 (0.27) | -- | -- | |
| EF | Visit 5 | 65.53 (5.69) | 65.64 (5.63) | 64.61 (6.30) | 63.70 (3.69) | −0.323 (−0.623 to −0.022) | 0.035 |
| Change (Absolute) | −2.01 (6.92) | −1.73 (7.19) | −1.55 (7.53) | −3.24 (8.18) | 0.185 (−0.156 to 0.525) | 0.29 | |
| Change (%) | −0.03 (0.11) | −0.02 (0.12) | −0.02 (0.13) | −0.05 (0.13) | -- | -- | |
| LS | Visit 5 | −18.14 (2.35) | −18.06 (2.40) | −17.33 (2.47) | −17.67 (2.27) | 0.210 (0.082 to 0.338) | 0.001 |
| Change (Absolute) | −0.56 (2.87) | −0.57 (2.83) | −0.50 (3.19) | −0.58 (3.19) | 0.115 (−0.027 to 0.257) | 0.11 | |
| Change (%) | 0.00 (0.68) | 0.02 (0.16) | 0.02 (0.19) | 0.02 (0.19) | -- | -- | |
| CS | Visit 5 | −27.74 (3.53) | −27.93 (3.81) | −27.62 (3.96) | −28.69 (2.67) | −0.027 (−0.251 to 0.197) | 0.82 |
| Change (Absolute) | −0.54 (4.17) | −0.59 (4.45) | −0.78 (4.07) | −2.40 (3.74) | 0.048 (−0.186 to 0.282) | 0.69 | |
| Change (%) | 0.01 (0.16) | 0.01 (0.18) | 0.01 (0.16) | 0.08 (0.13) | -- | -- | |
| E wave, cm/s | Visit 5 | 66.03 (16.62) | 66.07 (17.50) | 67.87 (18.87) | 66.38 (24.93) | −0.398 (−1.326 to 0.531) | 0.40 |
| Change (Absolute) | 8.44 (18.48) | 8.68 (18.47) | 10.79 (19.14) | 11.35 (22.54) | 1.116 (0.137 to 2.096) | 0.025 | |
| Change (%) | 0.16 (0.31) | 0.16 (0.32) | 0.19 (0.33) | 0.20 (0.33) | -- | -- | |
| Septal e’, cm/s | Visit 5 | 5.89 (1.41) | 5.71 (1.38) | 5.46 (1.60) | 5.09 (1.52) | −0.137 (−0.217 to −0.056) | <0.001 |
| Change (Absolute) | −0.55 (1.49) | −0.51 (1.60) | −0.39 (1.62) | 0.04 (1.77) | 0.085 (0.013 to 0.157) | 0.021 | |
| Change (%) | −0.07 (0.26) | −0.06 (0.28) | −0.04 (0.30) | 0.03 (0.38) | -- | -- | |
| E/e’ ratio | Visit 5 | 11.67 (3.61) | 12.02 (3.74) | 13.07 (4.39) | 14.25 (8.68) | 0.245 (0.047 to 0.444) | 0.015 |
| Change (Absolute) | 3.01 (4.56) | 3.08 (4.84) | 3.59 (5.13) | 1.98 (5.39) | 0.085 (−0.163 to 0.333) | 0.50 | |
| Change (%) | 0.30 (0.42) | 0.30 (0.42) | 0.33 (0.49) | 0.23 (0.37) | -- | -- | |
| LAVI, mL/m2 | Visit 5 | 24.75 (7.28) | 25.04 (7.18) | 27.04 (8.48) | 26.60 (6.06) | 0.458 (0.069 to 0.848) | 0.021 |
| Change (Absolute) | 2.58 (6.97) | 2.60 (7.07) | 3.10 (7.30) | 5.08 (7.18) | 0.168 (−0.228 to 0.564) | 0.41 | |
| Change (%) | 0.13 (0.28) | 0.14 (0.31) | 0.15 (0.30) | 0.22 (0.32) | -- | -- | |
| RV FAC, % | Visit 5 | 0.52 (0.08) | 0.53 (0.07) | 0.52 (0.08) | 0.54 (0.07) | −0.002 (−0.006 to 0.002) | 0.35 |
| Change (Absolute) | −0.03 (0.09) | −0.03 (0.09) | −0.03 (0.09) | −0.10 (0.11) | 0.001 (−0.004 to 0.006) | 0.68 | |
| Change (%) | −0.05 (0.18) | −0.04 (0.17) | −0.05 (0.17) | −0.18 (0.19) | -- | -- | |
| PASP, mm Hg | Visit 5 | 26.85 (4.58) | 27.48 (4.78) | 28.22 (5.64) | 26.92 (6.69) | 0.275 (−0.081 to 0.631) | 0.13 |
| Change (Absolute) | 5.07 (7.45) | 5.12 (7.95) | 5.60 (8.41) | 9.42 (13.18) | 0.175 (−0.423 to 0.774) | 0.57 | |
| Change (%) | 0.20 (0.29) | 0.20 (0.30) | 0.21 (0.30) | 0.37 (0.45) | -- | -- | |
| Tricuspid s’, cm/s | Visit 5 | 11.60 (2.66) | 11.97 (2.77) | 11.71 (2.66) | 11.25 (3.56) | 0.127 (−0.024 to 0.278) | 0.098 |
| Change (Absolute) | 0.16 (2.96) | 0.05 (3.14) | −0.02 (3.17) | 0.61 (3.97) | −0.052 (−0.205 to 0.101) | 0.50 | |
| Change (%) | 0.04 (0.26) | 0.03 (0.27) | 0.03 (0.30) | 0.10 (0.36) | -- | -- | |
CI = confidence interval; CS = circumferential strain; Demo. = demographics; EDVI = end-diastolic volume index; EF = ejection fraction; eGFR = estimated glomerular filtration rate; HF = heart failure; IPAW = inverse probability of attrition weighting; LS = longitudinal strain; LAVI = left atrial volume index; LVMI = left ventricular mass index; MI = myocardial infarction; MWT = mean wall thickness; PASP = pulmonary artery systolic pressure; RV FAC = right ventricular fractional area change; RWT = relative wall thickness; UACR = urine albumin-to-creatinine ratio
Discussion
This study of kidney function and damage, HF and adverse cardiac remodeling in a biracial, community-dwelling cohort of older adults includes the following principal findings. First, both lower eGFR and higher UACR were independently associated with a heightened risk of incident HF, HFrEF and HFpEF in late life. Second, reduced eGFR, but not increased UACR, predicted subsequent adverse left ventricular remodeling and worsening diastolic function. Notably, these associations between CKD measures and HF were independent of prevalent HF risk factors and intervening MI, whereas the associations with adverse cardiac remodeling were attenuated after censoring participants with an intervening MI.
The clinical relevance of moderately reduced eGFR and moderately increased UACR in older adults remains debated.(7) While studies of older kidney donors indicate that some reduction of GFR and kidney remodeling can occur during healthy aging, individuals eligible to donate a kidney have unusually good health for their age.(7) Indeed, the incidence of cardiovascular disease amongst kidney donors is comparable to, and perhaps lower than, that of the general population.(19) In contrast, one analysis of non-malignant nephrectomy samples, which included patients with comorbidities such as diabetes and hypertension, suggest that age does not modify the association between eGFR and histological kidney damage.(20) Furthermore, the studies that demonstrate the prognostic value of eGFR and UACR to all-cause death and heart failure in late-life, including the present study, also enrolled adults with comorbidities such as diabetes and hypertension.(5, 10) The mechanisms that lead to reductions in eGFR and increases in UACR therefore may exert strong influence over the prognostic value of these measures in late-life.
Our results indicate that reduced eGFR and increased UACR each contribute to the risks of both HFrEF and HFpEF in older adults, although reduced eGFR may associate more strongly with HFpEF than HFrEF. These findings agree with prior epidemiologic studies assessing the associations between eGFR and UACR with HF phenotypes in mid-life.(21-26) Our results now extend these observations to older adults with moderately reduced eGFR and mildly-moderately increased UACR. This segment of the population is poised to contribute immensely to the increasing burden of HF.
In previous cross-sectional studies, measures of CKD associated with abnormalities in cardiac structure and function, including left ventricular concentric hypertrophy, higher E/e’ ratio, a correlate of left ventricular filling pressure, and E/A ratio, a measure of abnormal left ventricular relaxation.(2-4) UACR, but not eGFR, associated with lower LVEF and impaired longitudinal strain.(2-4) Overall, these cross-sectional associations between CKD and cardiac structure and function appear stronger for UACR than eGFR. The interpretation of these cross-sectional analyses, however, should consider the bidirectional nature of the heart-kidney relationship. CKD can lead to HF (e.g., cardio-renal syndrome type 4), and HF can lead to CKD (e.g., cardio-renal syndrome type 2).(27) Longitudinal studies have shown that prevalent HF, as well as measures of cardiac structure and function in persons free of HF, predict incident CKD.(28-30) In this context, the present longitudinal study of older adults provides an important addition to the existing literature. Our findings suggest that, among older adults free of HF, lower eGFR promotes LV enlargement, as reflected in increases in left ventricular end-diastolic volume, and increases in LV filling pressure, as reflected in increases in left atrial volume index and possibly E/e’ ratio. Importantly, larger LV size and higher filling pressure characterize both HFpEF and HFrEF in late-life – consistent with the association of lower eGFR with both incident HFpEF and HFrEF. Healthy survivor bias potentially led to underestimates of the aforementioned associations, even for measures such as longitudinal and circumferential strain, which have utility for the identification of subclinical cardiac dysfunction. There was no association between reduced eGFR and left ventricular size or systolic function. While right ventricular dysfunction and volume overload may contribute to renal venous congestion and kidney dysfunction, this study suggests that kidney dysfunction itself does not impact right ventricular function. Interestingly, eGFR did not associate with LV end-diastolic volume index after excluding participants with an interval MI between the Visit 5 and Visit 7 echocardiograms, which suggests that eGFR associates with greater increases in LV end-diastolic volume index through ischemic cardiomyopathy.
UACR associated cross-sectionally with cardiac structure and function and longitudinally with incident HFrEF and HFpEF, did not associate with longitudinal changes in cardiac structure or function even after including adults with interval HF. This null finding contrasts with observations from prior cross-sectional studies of the associations between UACR and cardiac structure and function.(2-4) Most participants had UACR within the normal range and UACR was the most frequently missing covariate (728 or 11%), which may have reduced our power to detect significant associations. The impact of differential survival and non-attendance on our analyses also may contribute to the null findings in this study, as the rates of death and non-attendance increased across categories of UACR. Furthermore, although we observed similar findings in analyses incorporating inverse probability weights to account for Visit 7 non-attendance among those alive at Visit 7, predicting future visit attendance is difficult and residual healthy attendance bias likely persisted. Further studies of UACR and longitudinal changes in cardiac structure and function should be conducted in younger cohorts with lower overall rates of death or by combining multiple cohorts to increase statistical power. An additional potential explanation for these null findings is that greater UACR influences HF risk through unmeasured risk factors, which are more common in people with rather than without albuminuria.(31)
This study has certain limitations. Serum creatinine, urine albumin and urine creatinine measurements were obtained at a single timepoint. Clinico-pathologic diagnoses of CKD, which inform risk of CKD progression and possibly cardiovascular risk, were unavailable.(32, 33) ARIC enrolled a biracial cohort from four locations and these results may not generalize to other populations.
Conclusions
In late-life, measures of kidney dysfunction and damage each associate independently with both heart failure with reduced and preserved ejection fraction. Lower eGFR associates with LV enlargement and increases in LV filling pressure over 6 years, whereas UACR did not associate with longitudinal changes in cardiac structure or function. This study demonstrates the impact of kidney dysfunction and damage on cardiac structure and function and heart failure in older adults with mild-moderate kidney disease (Central Illustration).
Central Illustration. Independent and Additive Associations Between Kidney Dysfunction and Damage with Incident Heart Failure.
This study demonstrates the importance of kidney dysfunction and damage to cardiac structure and function and heart failure in older adults with mild-moderate kidney disease.
Supplementary Material
Clinical Perspectives.
Competency in Medical Knowledge:
Older adults with either mild-moderate kidney dysfunction or kidney damage have an increased risk of incident heart failure
Kidney dysfunction and damage confer independent risks of incident heart failure
Translational Outlook:
eGFR or UACR may identify older adults with a higher likelihood of benefitting from novel cardio-renal therapies
Acknowledgements
The authors thank the staff and participants of the ARIC study for their important contributions.
Financial Disclosures:
The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under Contract nos. (HHSN268201700001I, HHSN268201700002I, HHSN268201700003I, HHSN268201700004I, HHSN268201700005I). Dr. Buckley was supported by NIH/NHLBI grant K23HL150311. Dr. Shah was supported by NIH/NHLBI grants R01HL135008, R01HL143224, R01HL150342, R01HL148218 and K24HL152008. The remaining authors have nothing to disclose.
Abbreviations:
- ARIC
Atherosclerosis Risk in Communities
- BMI
body mass index
- CHD
coronary heart disease
- CKD
chronic kidney disease
- eGFR
estimated glomerular filtration rate
- HF
heart failure
- HFpEF
heart failure with preserved ejection fraction
- HFrEF
heart failure with reduced ejection fraction
- LVEF
left ventricular ejection fraction
- MI
myocardial infarction
- UACR
urine albumin-to-creatinine ratio
Footnotes
Relationships with Industry: Dr. Shah reports research support not related to this study from Novartis and Philips Ultrasound, and consulting fees from Philips Ultrasound. The remaining authors have nothing to disclose.
References
- 1.Matsushita K, Coresh J, Sang Y, et al. Estimated glomerular filtration rate and albuminuria for prediction of cardiovascular outcomes: A collaborative meta-analysis of individual participant data. The Lancet Diabetes and Endocrinology 2015;3:514–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Matsushita Kunihiro, Kwak Lucia, Sang Yingying, et al. Kidney Disease Measures and Left Ventricular Structure and Function: The Atherosclerosis Risk in Communities Study. Journal of the American Heart Association 2017;6:e006259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Park M, Hsu CY, Li Y, et al. Associations between kidney function and subclinical cardiac abnormalities in CKD. Journal of The American Society Of Nephrology 2012;23:1725–1734. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Katz DH, Selvaraj S, Aguilar FG, et al. Association of low-grade albuminuria with adverse cardiac mechanics: findings from the hypertension genetic epidemiology network (HyperGEN) study. Circulation 2014;129:42–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Hallan S, Astor B, Romundstad S, Aasarød K, Kvenild K, Coresh J. Association of kidney function and albuminuria with cardiovascular mortality in older vs younger individuals: The HUNT II Study. Arch Intern Med 2007;167:2490–6. [DOI] [PubMed] [Google Scholar]
- 6.Cheng S, Xanthakis V, Sullivan LM, et al. Correlates of Echocardiographic Indices of Cardiac Remodeling Over the Adult Life Course. Circulation 2010;122:570–578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hommos MS, Glassock RJ, Rule AD. Structural and Functional Changes in Human Kidneys with Healthy Aging. J Am Soc Nephrol 2017;28:2838–2844. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Denic A, Mathew J, Lerman LO, et al. Single-Nephron Glomerular Filtration Rate in Healthy Adults. N Engl J Med 2017;376:2349–2357. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hallan SI, Matsushita K, Sang Y, et al. Age and Association of Kidney Measures With Mortality and End-stage Renal Disease. JAMA 2012;308:2349–2360. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rifkin DE, Katz R, Chonchol M, et al. Albuminuria, impaired kidney function and cardiovascular outcomes or mortality in the elderly. Nephrol Dial Transplant 2010;25:1560–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Perak AM, Khan SS, Colangelo LA, et al. Age-Related Development of Cardiac Remodeling and Dysfunction in Young Black and White Adults: The Coronary Artery Risk Development in Young Adults Study. J Am Soc Echocardiogr 2021;34:388–400. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Lieb W, Xanthakis V, Sullivan LM, et al. Longitudinal Tracking of Left Ventricular Mass Over the Adult Life Course. Circulation 2009;119:3085–3092. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Levey AS, Stevens LA, Schmid CH, et al. A new equation to estimate glomerular filtration rate. Annals of Internal Medicine 2009;150:604–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Shah AM, Cheng S, Skali H, et al. Rationale and Design of a Multicenter Echocardiographic Study to Assess the Relationship Between Cardiac Structure and Function and Heart Failure Risk in a Biracial Cohort of Community-Dwelling Elderly Persons. Circulation: Cardiovascular Imaging 2014;7:173–181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rosamond WD, Chang PP, Baggett C, et al. Classification of heart failure in the atherosclerosis risk in communities (ARIC) study a comparison of diagnostic criteria. Circulation: Heart Failure 2012;5:152–159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.White AD, Folsom AR, Chambless LE, et al. Community surveillance of coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) Study: Methods and initial two years’ experience. Journal of Clinical Epidemiology 1996;49:223–233. [DOI] [PubMed] [Google Scholar]
- 17.Alonso A, Agarwal SK, Soliman EZ, et al. Incidence of atrial fibrillation in whites and African-Americans: the Atherosclerosis Risk in Communities (ARIC) study. Am Heart J 2009;158:111–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Glymour MM, Weuve J, Berkman LF, Kawachi I, Robins JM. When is baseline adjustment useful in analyses of change? An example with education and cognitive change. Am J Epidemiol 2005;162:267–78. [DOI] [PubMed] [Google Scholar]
- 19.O’Keeffe LM, Ramond A, Oliver-Williams C, et al. Mid- and Long-Term Health Risks in Living Kidney Donors: A Systematic Review and Meta-analysis. Ann Intern Med 2018;168:276–284. [DOI] [PubMed] [Google Scholar]
- 20.Li P, Gupta S, Mothi SS, et al. Histopathologic Correlates of Kidney Function: Insights From Nephrectomy Specimens. Am J Kidney Dis 2021;77:336–345. [DOI] [PubMed] [Google Scholar]
- 21.De Boer RA, Nayor M, DeFilippi CR, et al. Association of cardiovascular biomarkers with incident heart failure with preserved and reduced ejection fraction. JAMA Cardiology 2018;3:215–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Bansal N, Zelnick L, Bhat Z, et al. Burden and Outcomes of Heart Failure Hospitalizations in Adults With Chronic Kidney Disease. Journal of the American College of Cardiology 2019;73:2691–2700. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Brouwers FP, de Boer RA, van der Harst P, et al. Incidence and epidemiology of new onset heart failure with preserved vs. reduced ejection fraction in a community-based cohort: 11-year follow-up of PREVEND. European Heart Journal 2013;34:1424–1431. [DOI] [PubMed] [Google Scholar]
- 24.Lam CSP, Lyass A, Kraigher-Krainer E, et al. Cardiac dysfunction and noncardiac dysfunction as precursors of heart failure with reduced and preserved ejection fraction in the community. Circulation 2011;124:24–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Bailey LN, Levitan EB, Judd SE, et al. Association of Urine Albumin Excretion With Incident Heart Failure Hospitalization in Community-Dwelling Adults. JACC: Heart Failure 2019;7:394–401. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Nayor M, Larson MG, Wang N, et al. The association of chronic kidney disease and microalbuminuria with heart failure with preserved vs. reduced ejection fraction. European Journal of Heart Failure 2017;19:615–623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Rangaswami J, Bhalla V, Blair JEA, et al. Cardiorenal Syndrome: Classification, Pathophysiology, Diagnosis, and Treatment Strategies: A Scientific Statement From the American Heart Association. Circulation 2019. [DOI] [PubMed] [Google Scholar]
- 28.George LK, Koshy SKG, Molnar MZ, et al. Heart Failure Increases the Risk of Adverse Renal Outcomes in Patients with Normal Kidney Function. Circulation: Heart Failure 2017;10:e003825–e003825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Ishigami J, Cowan LT, Demmer RT, et al. Incident Hospitalization with Major Cardiovascular Diseases and Subsequent Risk of ESKD: Implications for Cardiorenal Syndrome. Journal of The American Society Of Nephrology 2020;31:405–414. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Ishigami J, Mathews L, Hishida M, et al. Echocardiographic measures and subsequent decline in kidney function in older adults: the Atherosclerosis Risk in Communities Study. Eur Heart J Cardiovasc Imaging 2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Foster MC, Hwang SJ, Larson MG, et al. Cross-classification of microalbuminuria and reduced glomerular filtration rate: associations between cardiovascular disease risk factors and clinical outcomes. Arch Intern Med 2007;167:1386–92. [DOI] [PubMed] [Google Scholar]
- 32.Srivastava A, Palsson R, Kaze AD, et al. The Prognostic Value of Histopathologic Lesions in Native Kidney Biopsy Specimens: Results from the Boston Kidney Biopsy Cohort Study. Journal of The American Society Of Nephrology 2018;29:ASN.2017121260–ASN.2017121260. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.O’Shaughnessy MM, Liu S, Montez-Rath ME, Lafayette RA, Winkelmayer WC. Cause of kidney disease and cardiovascular events in a national cohort of US patients with end-stage renal disease on dialysis: a retrospective analysis. European Heart Journal 2019;40:887–898. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.