Abstract
Aims
According to Kidney Disease: Improving Global Outcomes (KDIGO) guideline, the definition of chronic kidney disease (CKD) requires the presence of abnormal kidney structure or function for >3 months with implications for health. CKD in patients with heart failure (HF) has not been defined using this definition, and less is known about the true health implications of CKD in these patients. The objective of the current study was to identify patients with HF who met KDIGO criteria for CKD and examine their outcomes.
Methods and results
Of the 1,419,729 Veterans with HF not receiving kidney replacement therapy, 828,744 had data on ≥2 ambulatory serum creatinine >90 days apart. CKD was defined as estimated glomerular filtration rate (eGFR) <60mL/min/1.73m2 (n=185,821) or urinary albumin-creatinine ratio (uACR) >30mg/g (n=32,730) present twice >3 months apart. Normal kidney function (NKF) was defined as eGFR≥60, present for >3 months, without any uACR >30mg/g (n=366,963). Patients with eGFR<60 were categorized into 4 stages: 45–59 (n=72,606), 30–44 (n=74,812), 15–29 (n=32,077), and <15 (n=6326). 5-year all-cause mortality occurred in 40.4%, 57.8%, 65.6%, 73.3%, 69.7% and 47.5% of patients with NKF, 4 eGFR stages, and uACR>30 (albuminuria), respectively. Compared with NKF, HRs (95% CIs) for all-cause mortality associated with the 4 eGFR stages and albuminuria were 1.63 (1.62–1.65), 2.00 (1.98–2.02), 2.49 (2.45–2.52), 2.28 (2.21–2.35), and 1.22 (1.20–1.24), respectively. Respective age-adjusted HRs (95% CIs) were 1.13 (1.12–1.14), 1.36 (1.34–1.37), 1.87 (1.84–1.89), 2.24 (2.18–2.31), and 1.19 (1.17–1.21), and multivariable-adjusted HRs (95% CIs) were 1.11 (1.10–1.12), 1.24 (1.22–1.25), 1.46 (1.43–1.48), 1.42 (1.38–1.47), and 1.13 (1.11–1.16). Similar patterns were observed for associations with hospitalizations.
Conclusion
Data needed to define CKD using KDIGO criteria was available in six out of ten patients, and CKD could be defined in seven out of ten patients with data. HF patients with KDIGO-defined CKD had higher risks for poor outcomes, most of which was not explained by abnormal kidney structure or function. Future studies need to examine whether CKD defined using a single eGFR is characteristically and prognostically different from CKD defined using KDIGO criteria.
Keywords: Heart Failure, Chronic Kidney Disease, KDIGO, Outcomes
Graphical Abstract
X Post (formerly Tweet):
Patients with HF who met KDIGO definition of CKD (i.e., CKD present for >3 months) have higher risks of death and HF hospitalization, but many patients either lack necessary data or do not meet KDIGO criteria despite available data.
Introduction
Kidneys play an important role in the pathogenesis of heart failure (HF), and impaired kidney function is a risk factor for incident HF.1-4 Thus, kidney dysfunction is common in patients with HF and is also a predictor of poor outcomes.2, 5-7 Although chronic kidney dysfunction is often referred to as chronic kidney disease (CKD),8 prior studies of CKD in HF are based on different cutoffs of a single serum creatinine or a single estimated glomerular filtration rate (eGFR) value <60 mL/min/1.73m2. To establish the chronicity of kidney disease, the Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guideline for the Evaluation and Management of CKD recommends that abnormal kidney structure or function be present for >3 months.9, 10 According to KDIGO, an albumin to creatinine ratio >30 mg/g is the principal marker of kidney structure damage, and an eGFR <60 mL/min/1.73m2 represents a significant loss of kidney function. The objective of the current study was to identify patients with HF who met KDIGO eGFR criteria for CKD and examine their outcomes.
Methods
Study Design, Setting, and Participants
Using the electronic health record (EHR) of the U.S. Department of Veterans Affairs (VA) healthcare system, we identified 1,446,053 Veterans aged ≥20 years who had International Classification of Diseases (ICD) codes for HF between October 1, 1999, and December 31, 2017 (Figure 1, Graphical Abstract). We excluded patients who were receiving kidney replacement therapy (n=26,324), those who had no laboratory data on serum creatinine during a 3-year period before the first HF diagnosis in EHR (n=353,596), and those who did not have serum creatinine data necessary for KDIGO eGFR criteria for CKD (n= 237,389; Figure 1). The remaining 828,744 patients had data on at least 2 ambulatory serum creatinine measures >90 days apart. The study was approved by the VA Central Institutional Review Board.
Figure 1. Flowchart of Study Participants.
Of the 1,419,729 patients with a diagnosis of HF in VA national EHR, without ESRD or kidney transplantation, 58.3% (828,744 of 1,419,729) had data necessary to define CKD using KDIGO criteria. Of the 828,744 patients, 185,821 met KDIGO eGFR criteria (2 values <60 mL/min/1.73m2 >90 days apart). Using the eGFR proximal to the HF diagnosis date, we categorized these patients into 4 eGFR stages: 45–59 (3A), 30–44 (3B), 15–29 (4) and <15 (5). From the remaining 398,693 patients who did not meet KDIGO eGFR criteria and had data on uACR, we identified patients who met KDIGO albuminuria criteria (2 values of uACR >30 mg/g >90 days apart). Finally, we identified 365,963 patients with normal kidney function, defined as 2 values of eGFR ≥60 mL/min/1.73m2 >90 days apart, without any eGFR <60 or albuminuria during the 3-year period. The 244,230 patients who did not meet any of the above criteria were excluded. The final study sample included 70.5% (584,514of 828,744) of the patients who had data necessary to define CKD using KDIGO criteria. eGFR was calculated using the 2021 CKD-EPI (without race) formula to calculate eGFR.
Abbreviations: CKD, chronic kidney disease; CKD-EPI, Chronic Kidney Disease-Epidemiology Collaboration; EHR, electronic health record; eGFR, estimated glomerular filtration rate (ml/min/1.73m2); ESKD, end-stage kidney disease; HF, heart failure; KDIGO, Kidney Disease Improving Global Outcomes; LVEF, left ventricular ejection fraction; VHA, Veterans Health Administration.
ICD Codes for HF: ICD-9 codes 398.91, 402.01, 402.11, 402.91, 404.01, 404.03, 404.11, 404.13, 404.91, 404.93, 428.x and ICD-10 codes I09.81, I11.0, I13.0, I13.2, I50., I50.1, I50.20, I50.21, I50.22, I50.23, I50.3, I50.30, I50.31, I50.32, I50.33, I50.4, I50.40, I50.41, I50.42, I50.43, I50.9 were used to identify patients with HF.
Study Exposure
We calculated eGFR using the 2021 update of the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation from standardized ambulatory serum creatinine from a 3-year period before the first HF diagnosis.11, 12 Serum creatinine assays in the VA are calibrated to be traceable to an isotope dilution mass spectrometry (IDMS) reference standard,11, 12 and values before 2007 were converted to IDMS-traceable creatinine. CKD was defined by eGFR <60 mL/min/1.73m2 on 2 separate occasions >90 days apart. For patients who did not meet the KDIGO eGFR criteria for CKD and had data on uACR, CKD was defined by uACR >30mg/g (albuminuria) on 2 separate occasions >90 days apart, NKF was defined as eGFR ≥60 mL/min/1.73m2 on 2 occasions >90 days apart, without any eGFR <60 mL/min/1.73m2 or uACR >30mg/g during the 3-year period. Using the proximal eGFR, eGFR was categorized into 4 stages: 3A (eGFR 45–59), 3B (eGFR 30–44), 4 (eGFR 15–29) and 5 (eGFR <15). Of the 828,744 patients, 218,551 had CKD: 185,821 met the eGFR criteria and 32,730 met uACR criteria (Figure 1). Using the most recent of the two eGFR values, we categorized patients in the eGFR <60 group into 4 stages: Stage-3A (eGFR, 45–59; n=72,606), Stage-3B (eGFR, 30–44; n=74,812), Stage-4 (eGFR, 15–29; n=32,077) and Stage-5 (eGFR, <15; n=6326) (Figure 1). The 244,230 patients who did not meet the above criteria for CKD or NKF were excluded. The final study sample consisted of 584,514 patients.
Other Baseline Measurements
Data on baseline characteristics were collected from the VA EHR. Over 97% of the patients had ≥2 clinical encounters in the VA system in 3 years prior to index HF (Table 1). Baseline data on vital signs, laboratory values, and medications were collected from both inpatient and outpatient settings within ±90 days of index HF. Baseline comorbidities were defined using ICD codes documented any time before index HF. Left ventricular ejection fraction (EF) was extracted from echocardiogram reports using natural language processing and values 365 days before and 90 days after index HF were used as baseline EF.13 Data on baseline use of medications included key HF medications, namely, renin-angiotensin system inhibitors (RASIs), beta-blockers, mineralocorticoid receptor antagonists (MRAs), and loop diuretics. Few patients were on angiotensin receptor/neprilysin inhibitor as the drug was approved in 2015 and incorporated into HF guidelines in 2017.14
Table 1.
Baseline characteristics of 584,514 patients with heart failure with KDIGO-defined CKD and normal kidney function
| Normal kidney function |
eGFR Stage-3A (eGFR 45–59) |
eGFR Stage-3B (eGFR 30–44) |
eGFR Stage-4 (eGFR 15–29) |
eGFR Stage-5 (eGFR <15) |
Albuminuria (uACR >30mg/g) |
|
|---|---|---|---|---|---|---|
| N=365963 (62.6%) | N=72606 (13.2%) | N=74812 (13.6%) | N=32077 (5.8%) | N=6326 (1.1%) | N=32,730 (5.6%)† | |
| Age, years | 68.6 (±11.1) | 77.7 (±9.5) | 78.2 (±9.7) | 76.1 (±10.8) | 70.7 (±11.3) | 69.6 (±9.6) |
| Sex | ||||||
| Female | 9290 (2.5%) | 1548 (2.1%)* | 1671 (2.2%)* | 724 (2.3%)* | 144 (2.3%)* | 487 (1.5%)* |
| Male | 356673 (97.5%) | 71058 (97.9%)* | 73141 (97.8%)* | 31353 (97.7%)* | 6182 (97.7%)* | 32,243 (98.5%)* |
| Race | ||||||
| African American | 45089 (12.3%) | 10611 (14.6%) | 11480 (15.3%) | 6705 (20.9%) | 2357 (37.3%) | 4020 (12.3%)* |
| White | 305466 (83.5%) | 60057 (82.7%) | 61267 (81.9%) | 24144 (75.3%) | 3621 (57.2%) | 27216 (83.2%)* |
| Other1 | 15408 (4.2%) | 1938 (2.7%) | 2065 (2.8%) | 1228 (3.8%) | 348 (5.5%) | 1494 (4.6%)* |
| LVEF,2 % | 46.1 (±15.4) | 47.2 (±14.9)* | 47.6 (±14.8) | 48.5 (±14.6) | 49.8 (±14.0) | 47.6 (±14.6) |
| LVEF categories | ||||||
| ≤40% | 93032 (25.4%) | 15594 (21.5%) | 15486 (20.7%) | 6361 (19.8%) | 1241 (19.6%) | 8034 (24.5%) |
| 41-49% | 25323 (6.9%) | 4597 (6.3%) | 4441 (5.9%) | 1966 (6.1%) | 483 (7.6%) | 2587 (7.9%) |
| ≥50% | 114335 (31.2%) | 22275 (30.7%) | 23317 (31.2%) | 11025 (34.4%) | 2566 (40.6%) | 12337 (37.7%) |
| LVEF missing | 133273 (36.4%) | 30140 (41.5%) | 31568 (42.2%) | 12725 (39.7%) | 2036 (32.2%) | 9772 (29.9%) |
| eGFR, mL/min/1.73m2 | 86.6 (±13.8) | 51.7 (±4.2) | 38.2 (±4.2) | 23.9 (±4.2) | 11.0 (±2.9) | 76.0 (±17.9) |
| Albuminuria (any) † | 0 (0%) | 13205 (18.2%) | 16390 (21.9%) | 8057 (25.1%) | 1563 (24.7%) | 32730 (100%) |
| Albuminuria (CKD-defining)† | 0 (0%) | 6165 (8.5%) | 7875 (10.5%) | 4035 (12.6%) | 775 (12.3%) | 32730 (100%) |
| Smoking | 90884 (24.8%) | 9613 (13.2%) | 9523 (12.7%) | 4640 (14.5%) | 1185 (18.7%) | 7985 (24.4%)* |
| Comorbidities3 | ||||||
| Hypertension | 314995 (86.1%) | 68582 (94.5%) | 71502 (95.6%) | 31190 (97.2%) | 6227 (98.4%) | 31896 (97.5%) |
| Coronary artery disease | 223994 (61.2%) | 50487 (69.5%) | 52119 (69.7%) | 21564 (67.2%) | 3636 (57.5%)* | 22441 (68.6%) |
| Atrial fibrillation | 79105 (21.6%) | 20024 (27.6%) | 19700 (26.3%) | 6837 (21.3%)* | 874 (13.8%) | 8522 (26%) |
| Diabetes mellitus | 147784 (40.4%) | 38657 (53.2%) | 43518 (58.2%) | 20755 (64.7%) | 4349 (68.7%) | 31120 (95.1%) |
| Stroke | 56089 (15.3%) | 17061 (23.5%) | 18347 (24.5%) | 8056 (25.1%) | 1402 (22.2%) | 7387 (22.6%) |
| PAD | 71863 (19.6%) | 21656 (29.8%) | 24519 (32.8%) | 11309 (35.3%) | 2102 (33.2%) | 10686 (32.6%) |
| Asthma | 37560 (10.3%) | 6009 (8.3%)* | 5523 (7.4%) | 2142 (6.7%) | 409 (6.5%) | 3266 (10%)* |
| COPD | 105226 (28.8%) | 18345 (25.3%)* | 17718 (23.7%) | 6992 (21.8%) | 1329 (21%) | 10211 (31.2%)* |
| Vital signs4 | ||||||
| Systolic BP, mmHg | 132.5 (±22.3) | 135.4 (±24.1) | 136.6 (±25.1) | 141.4 (±27.0) | 149.0 (±28.1) | 137.9 (±24.0) |
| Diastolic BP, mmHg | 74.5 (±13.5) | 71.9 (±13.6) | 71.2 (±13.8) | 72.3 (±14.6) | 76.1 (±15.7) | 74.5 (±13.8)* |
| Pulse, beats/min | 79.3 (±18.6) | 75.5 (±17.0) | 75.0 (±16.8) | 75.0 (±16.3) | 77.5 (±16.5) | 79.0 (±18.0)* |
| Body mass index, kg/m2 | 29.6 (±5.9) | 29.1 (±5.5)* | 29.0 (±5.5) | 28.9 (±5.6) | 28.6 (±5.6) | 31.9 (±6.0) |
| Laboratory values4 | ||||||
| Serum potassium, mEq/L | 4.2 (±0.5) | 4.3 (±0.5) | 4.4 (±0.5) | 4.5 (±0.6) | 4.4 (±0.7) | 4.3 (±0.5) |
| Serum sodium, mEq/L | 138.8 (±3.5) | 139.3 (±3.4) | 139.4 (±3.5) | 139.4 (±3.7) | 139.0 (±3.7) | 138.5 (±3.4)* |
| Serum albumin, mg/dL | 3.7 (±0.6) | 3.6 (±0.6) | 3.5 (±0.6) | 3.4 (±0.6) | 3.3 (±0.6)* | 3.6 (±0.6) |
| Hemoglobin, g/L | 13.3 (±2.0) | 12.5 (±1.9) | 12.0 (±1.9) | 11.2 (±1.8) | 10.5 (±1.8) | 12.8 (±2.0) |
| Medications4 | ||||||
| RAS inhibitors | 251431 (68.7%) | 49988 (68.8%) | 49259 (65.8%) | 18520 (57.7%) | 3130 (49.5%) | 29391 (89.8%) |
| ACE inhibitors | 213871 (58.4%) | 39309 (54.1%)* | 37335 (49.9%) | 13121 (40.9%) | 2161 (34.2%) | 21109 (64.5%) |
| ARBs | 37303 (10.2%) | 10609 (14.6%) | 11880 (15.9%) | 5394 (16.8%) | 969 (15.3%) | 6995 (21.4%) |
| ARNI | 257 (0.1%) | 70 (0.1%) | 44 (0.1%) | 5 (0%) | 0 (0%) | 1287 (3.9%)* |
| MRAs | ||||||
| Spironolactone | 36722 (10%) | 6030 (8.3%) | 6084 (8.1%) | 2146 (6.7%) | 207 (3.3%) | 3956 (12.1%)* |
| Eplerenone | 437 (0.1%) | 121 (0.2%)* | 114 (0.2%)* | 44 (0.1%)* | 14 (0.2%)* | 65 (0.2%)* |
| Loop diuretics | 201060 (54.9%) | 45254 (62.3%) | 51005 (68.2%) | 24593 (76.7%) | 5279 (83.4%) | 21697 (66.3%) |
| Beta-blockers | 240694 (65.8%) | 49744 (68.5%)* | 52667 (70.4%)* | 23973 (74.7%) | 5057 (79.9%) | 25508 (77.9%) |
| Digoxin | 46967 (12.8%) | 8210 (11.3%)* | 7650 (10.2%)% | 2493 (7.8%) | 279 (4.4%) | 3034 (9.3%) |
| Anti-hypertensives | ||||||
| Hydralazine | 15250 (4.2%) | 6907 (9.5%) | 11179 (14.9%) | 8841 (27.6%) | 2429 (38.4%) | 3791 (11.6%) |
| Thiazide diuretics | 81252 (22.2%) | 17659 (24.3%)* | 17854 (23.9%)* | 7133 (22.2%)* | 1408 (22.3%)* | 9581 (29.3%) |
| CCB | 109517 (29.9%) | 28446 (39.2%) | 34067 (45.5%) | 18992 (59.2%) | 4904 (77.5%) | 14633 (44.7%) |
Values are mean ±SD, n (%). Due to large sample size, all p values in table 1 were <0.001. ASD is unaffected by sample size and values <10% suggest adequate balance and 0% suggests complete balance. ASD values for each CKD categories are relative to the normal kidney function group.
Indicates that the ASD value for the baseline characteristic in that cell was <10% compared with the normal kidney function group. For example, the between-group difference in sex was statistically significant (p <0.001) due to large sample size, but the ASD of 2.2 suggests that sex is not a consequential confounder.
The % based on the overall population is likely an underestimation as data on uACR was not obtained from all patients.
Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin II receptor blocker; ARNI, angiotensin receptor/neprilysin inhibitor; ASD, absolute standardized difference; BP, blood pressure; CB, calcium channel blockers; CKD, chronic kidney disease; CKD-EPI, CKD-Epidemiology Collaboration; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate (ml/min/1.73m2); HF, heart failure; KDIGO, Kidney Disease Improving Global Outcomes; LVEF, left ventricular ejection fraction; MRA, , mineralocorticoid receptor antagonists; PAD, Peripheral arterial disease; RAS, renin-angiotensin system; uACR, urinary albumin creatinine ratio (mg/g); VA, Veterans Affairs.
CKD was defined by eGFR <60 mL/min/1.73m2 on 2 separate occasions >90 days apart, and among those with eGFR ≥60 mL/min/1.73m2 by albuminuria on 2 separate occasions >90 days apart. Albuminuria was defined as ≥1 uACR >30mg/g. Normal kidney function was defined as eGFR ≥60 mL/min/1.73m2 on 2 occasions >90 days apart, without any eGFR <60 or albuminuria during the 3-year period. Using the proximal eGFR, eGFR was categorized into 4 stages: 3A (eGFR 45–59), 3B (eGFR 30–44), 4 (eGFR 15–29) and 5 (eGFR <15). eGFR was calculated using the 2021 CKD-EPI (without race) using ambulatory serum creatinine calibrated to be traceable to an isotope dilution mass spectrometry (IDMS) reference standard.
Other races included Native American, Alaska Native, Asian or Pacific Islander, and other races.
LVEF data from 365 days before and 90 days after the index HF date was considered baseline (over half was available within ±30 days of the index HF date).
Morbidities were defined using International Classification of Diseases codes documented any time before the diagnosis of HF.
Medications, vital signs, and laboratory values are based on inpatient and outpatient data within ±90 days of index HF.
Study Outcomes
The outcomes were all-cause mortality, HF hospitalization, all-cause hospitalization, and the combined endpoint of HF hospitalization or all-cause mortality during 5 years of follow-up from study baseline, up to December 31, 2022. Over 95% of death dates in VHA electronic health records agree exactly with those from the National Death Index.15 Data on hospitalization in the VA healthcare system were obtained from VA EHR and those outside the VA from the VA-linked Medicare data.
Statistical Analysis
We compared baseline characteristics of the patients with NKF and the 5 CKD groups (4 eGFR stages 3A, 3B, 4 and 5, and albuminuria) by estimating absolute standardized differences (ASD).16 Unlike p values, which are significant when the sample size is large, ASD is not affected by sample size, and values <0.10 suggest adequate balance, despite p <0.05. Kaplan-Meier analysis was used to generate survival curves for unadjusted all-cause mortality and the combined endpoint of HF hospitalization or all-cause mortality for the 6 kidney function groups. Cox regression models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for all outcomes associated with the 5 CKD groups, using NKF as the reference. For all analyses, we used the following models: (1) unadjusted, to examine the natural history of clinical outcomes, (2) age-adjusted, to examine the role of between-group age differences on the outcome, (3) age-sex-race-adjusted, to examine the additional role of sex and race, and (4) multivariable-adjusted without medications, to examine the role of 21 other baseline characteristics listed in the Table 2 footnote. Then to examine the impact of between-group differences of medication use we added the following medications in each subsequent steps, starting with RASIs, MRAs, loop diuretics, beta-blockers, and anti-hypertensive medications. Missing values of continuous variables were replaced with values generated using multiple imputation program.17 We checked the proportional hazard assumption by visual examination of log-minus-log plots. All analytic statistical tests were 2-tailed and a p-value <0.05 was considered significant. Statistical analyses were conducted using SAS 9.4 for Windows.
Table 2.
All-cause mortality in 584,514 patients with heart failure by KDIGO-defined CKD
| Normal kidney function |
eGFR 45–59 mL/min/1.73m2 |
eGFR 30–44 mL/min/1.73m2 |
eGFR 15–29 mL/min/1.73m2 |
eGFR <15 mL/min/1.73m2 |
Albuminuria (UACR >30mg/g) |
|
|---|---|---|---|---|---|---|
| N=365963 (62.6%) |
N=72,606 (12.4%) |
N=74,812 (12.8%) |
N=32,077 (5.5%) |
N=6326 (1.1%) |
N=32,730 (5.6%) |
|
| All-cause mortality, n (%) | 147,771 (40.4%) | 41,986 (57.8%) | 49,053 (65.6%) | 23,499 (73.3%) | 4408 (69.7%) | 15,530 (47.5%) |
| Unadjusted | 1.00 | 1.63 (1.62–1.65) | 2.00 (1.98–2.02) | 2.49 (2.45–2.52) | 2.28 (2.21–2.35) | 1.22 (1.20–1.24) |
| Age-adjusted | 1.00 | 1.13 (1.12–1.14) | 1.36 (1.34–1.37) | 1.87 (1.84–1.89) | 2.24 (2.18–2.31) | 1.19 (1.17–1.21) |
| Age-sex-race-adjusted | 1.00 | 1.14 (1.13–1.15) | 1.37 (1.35–1.38) | 1.89 (1.86–1.91) | 2.28 (2.21–2.35) | 1.19 (1.17–1.21) |
| Multivariable-adjusted | ||||||
| Model 1: Adjusted for age, sex, race, and 21 other variables listed in the footnote below | 1.00 | 1.11 (1.10–1.13) | 1.26 (1.25–1.28) | 1.53 (1.51–1.56) | 1.54 (1.49–1.58) | 1.13 (1.11–1.16) |
| Model 2: Additional adjustment for RASIs | 1.00 | 1.11 (1.10–1.12) | 1.25 (1.24–1.26) | 1.48 (1.46–1.51) | 1.46 (1.41–1.50) | 1.14 (1.12–1.17) |
| Model 3: Additional adjustment for MRAs | 1.00 | 1.12 (1.11–1.13) | 1.25 (1.24–1.27) | 1.50 (1.48–1.51) | 1.47 (1.43–1.52) | 1.14 (1.12–1.16) |
| Model 4: Additional adjustment for loop diuretics | 1.00 | 1.10 (1.09–1.12) | 1.23 (1.22–1.24) | 1.43 (1.41–1.46) | 1.39 (1.34–1.43) | 1.13 (1.11–1.15) |
| Model 5: Additional adjustment for beta-blockers | 1.00 | 1.10 (1.09–1.12) | 1.23 (1.22–1.25) | 1.45 (1.43–1.47) | 1.41 (1.37–1.46) | 1.13 (1.11–1.16) |
| Model 6 (final): Additional adjustment for anti-hypertensive medication | 1.00 | 1.11 (1.10–1.12) | 1.24 (1.22–1.25) | 1.46 (1.43–1.48) | 1.42 (1.38–1.47) | 1.13 (1.11–1.16) |
Abbreviations: CI, confidence interval; CKD, chronic kidney disease; CKD-EPI, chronic kidney disease-epidemiology; eGFR, glomerular filtration rate (ml/min/1.73m2); HR, hazard ratio; KDIGO, Kidney Disease Improving Global Outcomes; MRA, mineralocorticoid receptor antagonists; RASI, renin-angiotensin system inhibitors.
CKD was defined by eGFR <60 mL/min/1.73m2 on 2 separate occasions >90 days apart, and among those with eGFR <60 mL/min/1.73m2 by albuminuria on 2 separate occasions >90 days apart. Albuminuria was defined as ≥1 uACR >30mg/g. Normal kidney function was defined as eGFR ≥60 mL/min/1.73m2 on 2 occasions >90 days apart, without any eGFR <60 or albuminuria during the 3-year period. Using the proximal eGFR, eGFR was categorized into 4 stages: 3A (eGFR 45–59), 3B (eGFR 30–44), 4 (eGFR 15–29) and 5 (eGFR <15). eGFR was calculated using the 2021 CKD-EPI (without race) using ambulatory serum creatinine calibrated to be traceable to an isotope dilution mass spectrometry (IDMS) reference standard.
Model 1: Adjusted for age, sex, race, and the 21 other variables: (1) left ventricular ejection fraction, (2) first HF diagnosis as principal hospital discharge diagnosis, (3) first HF diagnosis as primary outpatient encounter diagnosis, (4) albuminuria (only for the 4 eGFR groups), (5) smoking, (6) hypertension, (7) diabetes mellitus, (8) coronary artery disease, (9) atrial fibrillation, (10) stroke, (11) peripheral vascular disease, (12) asthma, (13) chronic obstructive pulmonary disease, (14) body mass index, (15) pulse, (16) systolic blood pressure, (17) diastolic blood pressure, (18) serum sodium, (19) serum potassium, (20) serum albumin, and (21) hemoglobin
Model 2: Additional adjustment for the use of RASI (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, angiotensin receptor/neprilysin inhibitor)
Model 3: Additional adjustment for the use of MRA (spironolactone and eplerenone)
Model 4: Additional adjustment for the use of loop diuretics (furosemide, torsemide, bumetanide, and ethacrynic acid)
Model 5: Additional adjustment for the use of beta blockers (bisoprolol, carvedilol, metoprolol, acebutolol, atenolol, betaxolol, carteolol, labetalol, nadolol, nebivolol, penbutolol, pindolol, propranolol, timolol, and sotalol)
Model 6: Additional adjustment for the use of anti-hypertensive medication (calcium channel blockers, hydralazine, and thiazide diuretics)
Results
Baseline Characteristics
Patients in the NKF and eGFR stages 3A, 3B, 4 and 5, and albuminuria groups had mean (SD) age of 68.6 (±11.1), 77.7 (±9.5), 78.2 (±9.7), 76.1 (±10.8), 70.7 (±11.3), and 69.6 (±9.6) years, respectively, with respective ASD values of 0.884, 0.925, 0.688, 0.190, and 0.096 (all relative to NKF), suggesting substantial imbalance for all eGFR stages (Table 1). Among those with NKF, the 4 eGFR stages, and albuminuria, 12.3%, 14.6%, 15.3%, 20.9%, 37.3% and 12.3% were African American, with respective ASD values of 0.067, 0.088, 0.232, 0.372, and 0.117 (Table 1). Albuminuria was present in 0.0%, 18.2%, 21.9%, 25.1%, 24.7%, and 100.0% of the patients with NKF, 4 eGFR stages, and albuminuria, respectively (Table 1). Between-group differences for all baseline characteristics were statistically significant at p <0.001, due to large sample size. However, the between-group differences are considered consequential for those with ASD values ≥10%.
All-Cause Mortality
During 5 years of follow-up from the first record of HF, all-cause mortality occurred in 40.4% (147,771 of 365,963) and 61.5% (134,476 of 218,551) of the patients with NKF and CKD, respectively (HR associated with CKD, 1.81; 95% CI, 1.79–1.82). Among patients with CKD, all-cause mortality occurred in 57.8% (41,986 of 72,606), 65.6% (49,053 of 74,812), 73.3% (23,499 of 32,077), 69.7% (4408 of 6326) and 47.5% (15,530 of 32,730) of those with eGFR stages 3A, 3B, 4 and 5, and albuminuria, respectively (Table 2). Unadjusted Kaplan Meier curves for 5-year total mortality for the 6 kidney function groups are displayed in Figure 2. Compared with NKF, HRs (95% CIs) for death associated with the 4 eGFR stages and albuminuria were 1.63 (1.62–1.65), 2.00 (1.98–2.02), 2.49 (2.45–2.52), 2.28 (2.21–2.35), and 1.22 (1.20–1.24), respectively (Table 2). Respective age-adjusted HRs (95% CIs) were 1.13 (1.12–1.14), 1.36 (1.34–1.37), 1.87 (1.84–1.89), 2.24 (2.18–2.31), and 1.19 (1.17–1.21). Multivariable-adjusted HRs (95% CIs) for death associated with eGFR stages 3A, 3B, 4 and 5, and albuminuria are displayed in Table 2.
Figure 2. Kaplan-Meier Plots for 5-Year Mortality by KDIGO-Defined CKD.
During 5 years of follow-up, patients with HF with KDIGO-defined CKD who had eGFR 45–59, 30–44, 15–29, <15, and albuminuria were associated with 63% (95% CI, 62–65%), 100% (95% CI, 98–102%), 149% (95% CI, 145–152%), 128% (95% CI, 121–135%) and 22% (95% CI, 22–24%) higher risk of all-cause mortality, when compared with normal kidney function. CKD was defined using KDIGO guideline recommendation as GFR <60 mL/min/1.73m2 or albuminuria on 2 separate occasions >90 days apart. eGFR was calculated using the 2021 CKD-EPI equation (without race). Using the proximal eGFR, patients were categorized into 4 eGFR stages: 3A (eGFR 45–59), 3B (eGFR 30–44), 4 (eGFR 15–29) and 5 (eGFR <15). Albuminuria was defined as ≥1 uACR >30mg/g. Normal kidney function was defined as eGFR ≥60 mL/min/1.73m2 on 2 occasions >90 days apart, without any eGFR <60 or uACR >30mg/g during the 3-year period. The apparently paradoxical lower risk in the eGFR <15 group is likely due to lower mean age of this group (701±11 vs. 76±11 years for CKD 4), which reversed when adjusted for age (as shown in Table 2) and initiation of dialysis during follow-up. Abbreviations: CI = confidence interval; CKD = chronic kidney disease; eGFR = estimated glomerular filtration rate (mL/min/1.73m2); HF = heart failure; KDIGO = Kidney Disease: Improving Global Outcomes; uACR = urinary albumin creatinine ratio (mg/g).
HF Hospitalization
HF hospitalization occurred in 13.7% (50,063 of 365,963) and 20.0% (43,651of 218,551) of the patients with NKF and CKD, respectively (HR associated with CKD, 1.72; 95% CI, 1.70–1.74). Respective HF hospitalization rates for patients with eGFR stages 3A, 3B, 4 and 5, and albuminuria were 18.5% (13,423 of 72,606), 21.0% (15,737 of 74,812), 22.8% (7307 of 32,077), 16.9% (1067 of 6326), and 18.7% (6117/32,730; Table 3). Compared with NKF, HRs (95% CIs) for HF hospitalization associated with the 4 eGFR stages and albuminuria were 1.53 (1.50–1.56), 1.88 (1.84–1.91), 2.25 (2.20–2.31), 1.59 (1.49–1.69), and 1.45 (1.41–1.48), respectively (Table 3). Respective age-adjusted HRs (95% CIs) were 1.37 (1.33–1.39), 1.67 (1.64–1.70), 2.07 (2.01–2.12), 1.57 (1.48–1.67), and 1.43 (1.39–1.46). Associations after multivariable adjustment for other baseline characteristics are displayed in Table 3.
Table 3.
Heart failure hospitalization in 584,514 patients with heart failure by KDIGO-defined CKD
| Normal kidney function |
eGFR 45–59 mL/min/1.73m2 |
eGFR 30–44 mL/min/1.73m2 |
eGFR 15–29 mL/min/1.73m2 |
eGFR <15 mL/min/1.73m2 |
Albuminuria (UACR >30mg/g) |
|
|---|---|---|---|---|---|---|
| N=365963 (62.6%) |
N=72,606 (12.4%) |
N=74,812 (12.8%) |
N=32,077 (5.5%) |
N=6326 (1.1%) |
N=32,730 (5.6%) |
|
| HF hospitalization, n (%) | 50,063 (13.7%) | 13,423 (18.5%) | 15,737 (21.0%) | 7307 (22.8%) | 1067 (16.9%) | 6117 (18.7%) |
| Unadjusted | 1.00 | 1.53 (1.50–1.56) | 1.88 (1.84–1.91) | 2.25 (2.20–2.31) | 1.59 (1.49–1.69) | 1.45 (1.41–1.48) |
| Age-adjusted | 1.00 | 1.37 (1.34–1.39) | 1.67 (1.64–1.70) | 2.07 (2.01–2.12) | 1.57 (1.48–1.67) | 1.43 (1.39–1.46) |
| Age-sex-race-adjusted | 1.00 | 1.34 (1.31–1.37) | 1.63 (1.60–1.66) | 2.00 (1.95–2.05) | 1.47 (1.39–1.56) | 1.42 (1.38–1.46) |
| Multivariable-adjusted | ||||||
| Model 1: Adjusted for age, sex, race, and 21 other variables listed in the footnote below | 1.00 | 1.25 (1.23–1.28) | 1.45 (1.42–1.48) | 1.59 (1.55–1.63) | 1.06 (1.00–1.13) | 1.33 (1.29–1.39) |
| Model 2: Additional adjustment for RASIs | 1.00 | 1.26 (1.24–1.29) | 1.49 (1.46–1.52) | 1.71 (1.67–1.78) | 1.19 (1.12–1.27) | 1.32 (1.27–1.37) |
| Model 3: Additional adjustment for MRAs | 1.00 | 1.26 (1.24–1.28) | 1.50 (1.47–1.53) | 1.74 (1.69–1.78) | 1.23 (1.15–1.31) | 1.31 (1.26–1.36) |
| Model 4: Additional adjustment for loop diuretics | 1.00 | 1.24 (1.21–1.26) | 1.42 (1.39–1.44) | 1.56 (1.52–1.60) | 1.05 (0.99–1.12) | 1.28 (1.24–1.33) |
| Model 5: Additional adjustment for beta-blockers | 1.00 | 1.24 (1.21–1.26) | 1.41 (1.39–1.44) | 1.55 (1.51–1.60) | 1.04 (0.98–1.11) | 1.28 (1.24–1.33) |
| Model 6 (final): Additional adjustment for anti-hypertensive medication | 1.00 | 1.21 (1.19–1.23) | 1.36 (1.34–1.39) | 1.46 (1.42–1.50) | 0.96 (0.90–1.02) | 1.26 (1.22–1.31) |
Abbreviations: CI, confidence interval; CKD, chronic kidney disease; CKD-EPI, chronic kidney disease-epidemiology; eGFR, glomerular filtration rate (ml/min/1.73m2); HR, hazard ratio; KDIGO, Kidney Disease Improving Global Outcomes; MRA, mineralocorticoid receptor antagonists; RASI, renin-angiotensin system inhibitors.
CKD was defined by eGFR <60 mL/min/1.73m2 on 2 separate occasions >90 days apart, and among those with eGFR <60 mL/min/1.73m2 by albuminuria on 2 separate occasions >90 days apart. Albuminuria was defined as ≥1 uACR >30mg/g. Normal kidney function was defined as eGFR ≥60 mL/min/1.73m2 on 2 occasions >90 days apart, without any eGFR <60 or albuminuria during the 3-year period. Using the proximal eGFR, eGFR was categorized into 4 stages: 3A (eGFR 45–59), 3B (eGFR 30–44), 4 (eGFR 15–29) and 5 (eGFR <15). eGFR was calculated using the 2021 CKD-EPI (without race) using ambulatory serum creatinine calibrated to be traceable to an isotope dilution mass spectrometry (IDMS) reference standard.
Model 1: Adjusted for age, sex, race, and the 21 other variables: (1) left ventricular ejection fraction, (2) first HF diagnosis as principal hospital discharge diagnosis, (3) first HF diagnosis as primary outpatient encounter diagnosis, (4) albuminuria (only for the 4 eGFR groups), (5) smoking, (6) hypertension, (7) diabetes mellitus, (8) coronary artery disease, (9) atrial fibrillation, (10) stroke, (11) peripheral vascular disease, (12) asthma, (13) chronic obstructive pulmonary disease, (14) body mass index, (15) pulse, (16) systolic blood pressure, (17) diastolic blood pressure, (18) serum sodium, (19) serum potassium, (20) serum albumin, and (21) hemoglobin
Model 2: Additional adjustment for the use of RASI (angiotensin-converting enzyme inhibitors, angiotensin II receptor blockers, angiotensin receptor/neprilysin inhibitor)
Model 3: Additional adjustment for the use of MRA (spironolactone and eplerenone)
Model 4: Additional adjustment for the use of loop diuretics (furosemide, torsemide, bumetanide, and ethacrynic acid)
Model 5: Additional adjustment for the use of beta blockers (bisoprolol, carvedilol, metoprolol, acebutolol, atenolol, betaxolol, carteolol, labetalol, nadolol, nebivolol, penbutolol, pindolol, propranolol, timolol, and sotalol)
Model 6: Additional adjustment for the use of anti-hypertensive medication (calcium channel blockers, hydralazine, and thiazide diuretics).
Addition adjustment for the use of digoxin had no effect on the association of CKD with HF hospitalization, and was not included in the final model.
All-Cause Hospitalization
All-cause hospitalization occurred in 70.6% (258,293 of 365,963) and 76.9% (168,077 of 218,551) of the patients with NKF and CKD, respectively (HR associated with CKD, 1.23; 95% CI, 1.23–1.24). Respective rates for any hospitalization for patients with the 4 eGFR stages and albuminuria were 75.2% (54,616 of 72,606), 77.3% (57,860 of 74,812), 80.4% (25,803 of 32,077), 83.0% (5249 of 6326), and 75.0% (24,549 of 32,730; Online Table 1). Compared with NKF, HRs (95% CIs) for all-cause hospitalization associated with eGFR stages 3A, 3B, 4, 5 and albuminuria were 1.21 (1.20–1.22), 1.37 (1.36–1.38), 1.71 (1.69–1.73), 2.08 (2.03–2.14), and 1.18 (1.16–1.19) respectively (Online Table 1). Respective age-adjusted HRs (95% CIs) were 1.10 (1.09–1.11), 1.24 (1.23–1.25), 1.58 (1.56–1.60), 2.05 (1.99–2.11), and 1.16 (1.15–1.18). Associations after multivariable adjustment for other baseline characteristics are displayed in Online Table 1.
Combined Endpoint of HF Hospitalization or Death
The combined endpoint of HF hospitalization or all-cause mortality occurred 47.1% (172,284 of 365,963) and 68.1% (148, 886 of 218,551) of the patients with NKF and CKD, respectively (HR associated with CKD, 1.36; 95% CI, 1.35–1.37). Respective rates for any hospitalization for patients with the 4 eGFR stages and albuminuria were 64.3% (46,648 of 72606), 71.8% (53,748 of 74,812), 79.0% (25,349 of 32,077), 74.8% (4729 of 6326), and 56.3% (18,412 of 32,730; Online Table 2).Unadjusted Kaplan Meier curves for the 5-year combined endpoint for NKF, 4 eGFR groups and albuminuria are displayed in the Graphical Abstract. Compared with NKF, HRs (95% CIs) for the combined endpoint of HF hospitalization or death due to any cause associated with CKD stages 3A, 3B, 4, 5 and albuminuria were 1.58 (1.56–1.59), 1.93 (1.91–1.95), 2.39 (2.36–2.42), 2.13 (2.07–2.19), and 1.28 (1.26–1.30), respectively (Online Table 2). Respective age-adjusted HRs (95% CIs) were 1.16 (1.15–1.18), 1.39 (1.38–1.41), 1.88 (1.85–1.90), 2.09 (2.03–2.15), and 1.24 (1.15–1.26). Associations after multivariable adjustment for other baseline characteristics are presented in Online Table 2.
Discussion
The findings of our study based on nearly one and a half million US Veteran patients with HF demonstrate that data necessary to define CKD using KDIGO guideline criteria was available from 57% of the patients, of whom 26% had CKD, 44% had NKF, and 30% had neither (Graphical Abstract). These findings also demonstrate that compared with NKF, CKD defined by KDIGO criteria was associated with a higher risk for death and hospitalization (Graphical Abstract). These findings provide new information about the true estimates of the associations between CKD and outcomes in patients with HF than those reported in prior studies based on a single eGFR that may have misclassified acute kidney disease as CKD. However, the exclusion of nearly two-thirds of the patients who either did not have data necessary for KDIGO eGFR criteria for CKD or did not meet those criteria despite available data highlights the need for future studies to examine the nature and severity of kidney disease in patients with HF.
Poor outcomes in HF patients with CKD would be expected to be due to neurohormonal and hemodynamic changes associated with kidney dysfunction.18 However, the findings from our study suggest that impaired kidney function only explains a small part of the risk. Age explained 79% (63–13/63= 0.79; the risk dropped from 63% to 13% after adjustment for age) of the risk of death in patients with CKD-3A and 64% (100–36/100= 0.79) in CKD-3B. This is not surprising given that patients with CKD stages 3A and 3B were on an average 9.1 and 9.6 years older than those with NKF, respectively. However, age played a lesser role as the eGFR stage advanced, so that after adjustment for age the risk of death in patients with CKD-5 dropped slightly from 128% to 124%. All baseline characteristics explained 83% (63–11/63= 0.83), 76% (100–24/100= 0.76), 64% (129–46/129= 0.64), 67% (128–42/128= 0.67), and 41% (22–13/22= 0.41) of the overall risk of death in patients with the four eGFR stages and albuminuria, respectively, suggesting that kidney dysfunction explained the respective remaining 17%, 24%, 36%, 33%, and 59% of the risk of death in these patients. Taken together, these findings suggest that compared with HF patients with NKF, those with functional abnormalities of the kidney had a higher risk of death (unadjusted, 63% to 149%), but impaired kidney function explained only up to one-third of those risks. In contrast, compared with NKF, those with structural abnormalities of the kidney had a modest 22% higher risk of death, but damaged kidney structure (i.e., albuminuria) explained nearly two-thirds of that risk. The numerically lower risk of death in patients with CKD-5 (vs. CKD-4) is likely due to a survivor cohort effect and initiation of dialysis in these patients.
Age plays a lesser role in explaining the higher risk of HF hospitalization in all stages of CKD, explaining 30% (53–37/53= 0.30), 24% (88–67/88= 0.24), 14% (125–107/125= 0.14), 3% (59–57/59= 0.03) and 4% (45–43/45= 0.04) of the risk in patients with eGFR stages 3A, 3B, 4, 5, and albuminuria, respectively. All measured baseline characteristics explained 60% (53–21/53= 0.60), 59% (88–36/88= 0.59), 64% (125–46/125= 0.64), and 42% (45–26/45= 0.42) of the risk of HF hospitalization in patients with eGFR stages 3A, 3B, 4, and albuminuria, respectively, suggesting that kidney dysfunction explained the respective remaining 40%, 41%, 36%, and 58% of the risk of HF hospitalization in these patients. The association of eGFR stage 5 and HF hospitalization completely disappeared after adjustment for all measured baseline characteristics. Damaged kidney structure explained about 60% of the higher risk of HF hospitalization in HF patients with albuminuria, which suggest that albuminuria with a normal eGFR has a more independent or intrinsic association with outcomes than that observed with reduced eGFR.
Interestingly, while declining eGFR increasingly explained the risk of death (17%, 24%, and 36% for CKD 3A, 3B, and 4, respectively), a similar progression was not seen for the risk of HF hospitalization (40%, 41%, and 36% for CKD 3A, 3B, and 4, respectively). One potential explanation for this is the use of loop diuretics in more patients and/or in higher doses with declining eGFR. A stronger intrinsic association of CKD 3A with HF hospitalization (40%) than with death (17%) suggest that impaired renal sodium excretion and fluid retention in CKD 3A have a greater effect on HF hospitalization than on death. A lack of incremental increase in the risk of death and absence of risk of HF hospitalization in patients with CKD 5 (kidney failure) is likely due to correction of fluid overload by dialysis during the follow-up.
The above observations from our study are likely representative of the true absolute and relative associations of CKD stages with outcomes in HF for several reasons. First, CKD was defined using internationally recognized KDIGO criteria. Second, only ambulatory creatinine values were used to calculate eGFR, reducing the chance of misclassification due to acute kidney injury/disease. Third, the relative risks are also likely true estimates because we defined NKF using strict criteria. These findings from our study suggest that in patients with CKD stage 3A, who represent nearly 4 out of 10 patients with CKD, impaired kidney function is causally associated with a small increase in the risk of death and HF hospitalization, and no association with all-cause hospitalization. This is consistent with prior observations of no association of CKD stage 3A with outcomes,2, 3 albeit based on a single eGFR. Although the large sample size of the current study makes the associations of CKD Stage-3 with outcomes statistically significant, their clinical significance is likely modest.
Because no prior studies of CKD in HF have used KDIGO definition of CKD, we are unable to compare our findings with those in the literature. In one study of 118,465 hospitalizations due to HF in which CKD was defined using a single admission eGFR, 64% had had eGFR <60.5, which is much higher than the 34% observed in our study. It is possible that the prevalence of CKD is higher in patients with acutely decompensated HF. However, it is also possible that the prevalence of CKD is overestimated due to the inclusion of patients with acute kidney disease. Despite the intrinsic validity of the findings of our study, the external validity might be limited because over a third of the patients did not have data necessary for assessment of CKD using KDIGO criteria. Furthermore, 19% of the patients could not be defined as CKD, despite available data on 2 eGFR values >90 days apart because one of the values was ≥60. Therefore, future studies are needed to better define kidney disease in these patients.
Limitations
As in any observational study, bias due to unmeasured confounders is possible. The KDIGO criteria were not developed or validated for patients with established HF. It is possible that kidney function in patients with HF is affected by non-kidney factors, such as neurohormonal activation, congestion, medications, sarcopenia, and cachexia.19, 20 Because some of the baseline characteristics are also mediators, adjustment for those may result in over- or under-estimation of these associations. For example, systolic blood pressure, was lower in patients with NKF, which has been shown to be associated with poor outcomes in patients with HF.21, 22 On the other hand, anemia, associated with poor outcomes in patients with HF,23 was more common in patients with lower eGFR. Finally, results of our study based on predominantly male Veterans may limit generalizability to other populations.
Conclusions
The findings of our study provide new information about the identification of CKD using KDIGO guidelines and the health implications of KDIGO-defined CKD. We observed that data needed to define CKD using KDIGO criteria was available in nearly six out of ten patients, and CKD could be defined in seven out of ten patients with data. We also observed that KDIGO-defined CKD is a marker of poor outcomes, although abnormal kidney structure or function does not explain all the higher risk associated with CKD. Future studies need to examine whether CKD defined in routine clinical practice is characteristically and prognostically different from CKD defined using KDIGO criteria. Future studies also need to define and categorize kidney disease in patients who did not meet KDIGO definition despite available data and in those who did not have data necessary for KDIGO definition.
Supplementary Material
Funding:
This work was supported by a grant from the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development Service (I01HX002422) to the Washington DC VA Medical Center. Support for CMS and USRDS Data provided by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development Service, VA Information Resource Center (Project Numbers SDR 02-237 and 98-004). The funding organization or sponsor played no role in the design, analysis, or interpretation of the current study or in the preparation, review, approval, or the decision to submit the manuscript for publication.
Footnotes
Disclaimer: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
References:
- 1.Ahmed A, Campbell RC. Epidemiology of chronic kidney disease in heart failure. Heart Fail Clin 2008;4:387–99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Campbell RC, Sui X, Filippatos G, Love TE, Wahle C, Sanders PW, Ahmed A. Association of chronic kidney disease with outcomes in chronic heart failure: a propensity-matched study. Nephrol Dial Transplant 2009;24:186–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Bowling CB, Feller MA, Mujib M, Pawar PP, Zhang Y, Ekundayo OJ, Aban IB, Love TE, Sanders PW, Anker SD, Fonarow GC, Ahmed A. Relationship between stage of kidney disease and incident heart failure in older adults. Am J Nephrol 2011;34:135–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Metra M, Cotter G, Gheorghiade M, Dei Cas L, Voors AA. The role of the kidney in heart failure. Eur Heart J 2012;33:2135–42. [DOI] [PubMed] [Google Scholar]
- 5.Heywood JT, Fonarow GC, Costanzo MR, Mathur VS, Wigneswaran JR, Wynne J, Committee ASA, Investigators. High prevalence of renal dysfunction and its impact on outcome in 118,465 patients hospitalized with acute decompensated heart failure: a report from the ADHERE database. J Card Fail 2007;13:422–30. [DOI] [PubMed] [Google Scholar]
- 6.Ahmed A, Rich MW, Sanders PW, Perry GJ, Bakris GL, Zile MR, Love TE, Aban IB, Shlipak MG. Chronic kidney disease associated mortality in diastolic versus systolic heart failure: a propensity matched study. Am J Cardiol 2007;99:393–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hillege HL, Nitsch D, Pfeffer MA, Swedberg K, McMurray JJ, Yusuf S, Granger CB, Michelson EL, Ostergren J, Cornel JH, de Zeeuw D, Pocock S, van Veldhuisen DJ, Candesartan in Heart Failure: Assessment of Reduction in M, Morbidity I. Renal function as a predictor of outcome in a broad spectrum of patients with heart failure. Circulation 2006;113:671–8. [DOI] [PubMed] [Google Scholar]
- 8.Hsu CY, Chertow GM. Chronic renal confusion: insufficiency, failure, dysfunction, or disease. Am J Kidney Dis 2000;36:415–8. [DOI] [PubMed] [Google Scholar]
- 9.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1–150. [Google Scholar]
- 10.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2023 clinical practice guideline for the evaluation and management of chronic kidney disease (public review draft, July 2023). https://kdigo.org/wp-content/uploads/2017/02/KDIGO-2023-CKD-Guideline-Public-Review-Draft_5-July-2023.pdf 2023. [Google Scholar]
- 11.Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, Van Lente F, Chronic Kidney Disease Epidemiology C. Expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007;53:766–72. [DOI] [PubMed] [Google Scholar]
- 12.Miller WG, Kaufman HW, Levey AS, Straseski JA, Wilhelms KW, Yu HE, Klutts JS, Hilborne LH, Horowitz GL, Lieske J, Ennis JL, Bowling JL, Lewis MJ, Montgomery E, Vassalotti JA, Inker LA. National Kidney Foundation Laboratory Engagement Working Group Recommendations for Implementing the CKD-EPI 2021 Race-Free Equations for Estimated Glomerular Filtration Rate: Practical Guidance for Clinical Laboratories. Clin Chem 2022;68:511–520. [DOI] [PubMed] [Google Scholar]
- 13.Garvin JH, DuVall SL, South BR, Bray BE, Bolton D, Heavirland J, Pickard S, Heidenreich P, Shen S, Weir C, Samore M, Goldstein MK. Automated extraction of ejection fraction for quality measurement using regular expressions in Unstructured Information Management Architecture (UIMA) for heart failure. J Am Med Inform Assoc 2012;19:859–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Yancy CW, Jessup M, Bozkurt B, Butler J, Casey DE Jr., Colvin MM, Drazner MH, Filippatos GS,Fonarow GC, Givertz MM, Hollenberg SM, Lindenfeld J, Masoudi FA, McBride PE, Peterson PN, Stevenson LW, Westlake C. 2017. ACC/AHA/HFSA Focused Update of the 2013 ACCF/AHA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Failure Society of America. Circulation 2017;136:e137–e161. [DOI] [PubMed] [Google Scholar]
- 15.Sohn MW, Arnold N, Maynard C, Hynes DM. Accuracy and completeness of mortality data in the Department of Veterans Affairs. Popul Health Metr 2006;4:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Austin PC. Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples. Stat Med 2009;28:3083–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Li P, Stuart EA, Allison DB. Multiple Imputation: A Flexible Tool for Handling Missing Data. JAMA 2015;314:1966–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Rangaswami J, Bhalla V, Blair JEA, Chang TI, Costa S, Lentine KL, Lerma EV, Mezue K, Molitch M, Mullens W, Ronco C, Tang WHW, McCullough PA, American Heart Association Council on the Kidney in Cardiovascular D, Council on Clinical C. Cardiorenal Syndrome: Classification, Pathophysiology, Diagnosis, and Treatment Strategies: A Scientific Statement From the American Heart Association. Circulation 2019;139:e840–e878. [DOI] [PubMed] [Google Scholar]
- 19.Springer J, Springer JI, Anker SD. Muscle wasting and sarcopenia in heart failure and beyond: update 2017. ESC Heart Fail 2017;4:492–498. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Hartupee J, Mann DL. Neurohormonal activation in heart failure with reduced ejection fraction. Nat Rev Cardiol 2017;14:30–38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Arundel C, Lam PH, Gill GS, Patel S, Panjrath G, Faselis C, White M, Morgan CJ, Allman RM, Aronow WS, Singh SN, Fonarow GC, Ahmed A. Systolic Blood Pressure and Outcomes in Patients With Heart Failure With Reduced Ejection Fraction. J Am Coll Cardiol 2019;73:3054–3063. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Tsimploulis A, Lam PH, Arundel C, Singh SN, Morgan CJ, Faselis C, Deedwania P, Butler J, Aronow WS, Yancy CW, Fonarow GC, Ahmed A. Systolic Blood Pressure and Outcomes in Patients With Heart Failure With Preserved Ejection Fraction JAMA Cardiol 2018;3:288–297. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bello NA, Lewis EF, Desai AS, Anand IS, Krum H, McMurray JJ, Olson K, Solomon SD, Swedberg K, van Veldhuisen DJ, Young JB, Pfeffer MA. Increased risk of stroke with darbepoetin alfa in anaemic heart failure patients with diabetes and chronic kidney disease. Eur J Heart Fail 2015;17:1201–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.