Abstract
Background and objectives
AKI in the hospital is common and is associated with excess mortality. We examined whether AKI is also independently associated with a higher risk of different cardiovascular events in the first year after discharge.
Design, setting, participants, & measurements
We conducted a retrospective analysis of a cohort between 2006 and 2013 with follow-up through 2014, within Kaiser Permanente Northern California. We identified all adults admitted to 21 hospitals who had one or more in-hospital serum creatinine test result and survived to discharge. Occurrence of AKI was on the basis of Kidney Disease: Improving Global Outcomes diagnostic criteria. Potential confounders were identified from comprehensive inpatient and outpatient, laboratory, and pharmacy electronic medical records. During the 365 days after discharge, we ascertained occurrence of heart failure, acute coronary syndromes, peripheral artery disease, and ischemic stroke events from electronic medical records.
Results
Among a matched cohort of 146,941 hospitalized adults, 31,245 experienced AKI. At 365 days postdischarge, AKI was independently associated with higher rates of the composite outcome of hospitalization for heart failure and atherosclerotic events (adjusted hazard ratio [aHR], 1.18; 95% confidence interval [95% CI], 1.13 to 1.25) even after adjustment for demographics, comorbidities, preadmission eGFR and proteinuria, heart failure and sepsis complicating the hospitalization, intensive care unit (ICU) admission, length of stay, and predicted in-hospital mortality. This was driven by an excess risk of subsequent heart failure (aHR, 1.44; 95% CI, 1.33 to 1.56), whereas there was no significant association with follow-up atherosclerotic events (aHR, 1.05; 95% CI, 0.98 to 1.12).
Conclusions
AKI is independently associated with a higher risk of cardiovascular events, especially heart failure, after hospital discharge.
Keywords: Acute Coronary Syndrome, Acute Kidney Injury, Adult, Brain Ischemia, cardiovascular disease, Comorbidity, creatinine, Electronic Health Records, Follow-up Studies, glomerular filtration rate, heart failure, Hospital Mortality, hospitalization, Inpatients, Intensive Care Units, Kidney Function Tests, Length Of Stay, mortality, Outpatients, Patient Discharge, Peripheral Arterial Disease, proteinuria, Retrospective Studies, Sepsis, Stroke
Introduction
AKI frequently complicates hospitalizations and has been increasing over time (1,2). AKI contributes to the development and progression of CKD (3–6) and excess mortality (7,8). With cardiovascular disease being a leading cause of death, it is important to evaluate whether AKI contributes to excess cardiovascular events.
AKI leads to acute elevations in inflammatory cytokines (9), endothelial dysfunction (10), and dysregulation in mineral metabolism (11,12), and may contribute to increases in markers of cardiac ischemia (13). However, whether AKI contributes to an excess risk of cardiovascular events remains controversial as previous studies have focused primarily on selected patients such as those with coronary angiography (14), cardiac surgery (15–18), acute myocardial infarction (19), or acute severe hypertension (20), or in populations with limited diversity (21). Furthermore, methodological limitations of existing studies include lack of preadmission kidney function, reliance on administrative codes for AKI, not accounting for relevant confounders (including acute severity of illness) or not using contemporary definitions of AKI (22).
Among a large, contemporary matched cohort of hospitalized adults, we examined the association between AKI and risks of cardiovascular events during the first year after discharge. As seen in other acute medical conditions (23–27) linked to higher risks of cardiovascular events, we hypothesized that AKI would be independently associated with higher risks of postdischarge cardiovascular events.
Materials and Methods
Sample
Kaiser Permanente Northern California is a large integrated health care delivery system caring for more than 4.2 million persons that are highly representative of the local and statewide population (28,29).
We identified adult (≥20 years) members hospitalized at 21 Kaiser Permanente Northern California hospitals between January of 2006 and December of 2013 who had ≥12 months of continuous membership and pharmacy benefit at admission, one or more serum creatinine value during the index hospitalization, and surviving to discharge (Figure 1). Hospitalizations were excluded if the patient had prior ESKD (receipt of chronic dialysis or kidney transplant) identified from a health plan ESKD registry. If a patient had more than one hospitalization during the study period, only the first admission was analyzed.
Figure 1.
Cohort assembly of adults hospitalized between January 1, 2006 and December 31, 2013. AKI defined using modified KDIGO criteria: increase in serum creatinine by ≥0.3 mg/dl (≥26.5 μmol/l) within 48 hours or an increase in serum creatinine to ≥1.5 times baseline (defined as the most recent nonemergency department, outpatient value in the 7–365 days before admission). SCr, serum creatinine.
The study was approved by participating institutions’ institutional review boards. Waiver of consent was obtained because of the nature of the study.
AKI
We defined AKI patterned after Kidney Disease: Improving Global Outcomes (KDIGO) criteria demonstrating a difference between serial inpatient serum creatinine concentrations of ≥0.3 mg/dl within 48 hours and/or ≥50% above preadmission baseline serum creatinine concentration (30). Baseline kidney function was defined as the most recent outpatient, nonemergency department serum creatinine concentration between 7 and 365 days before admission (4). All serum creatinine measurements were performed at the regional health plan laboratory using an isotope dilution mass spectrometry-traceable assay. Urine output data were not systematically available. We characterized AKI severity on the basis of serum creatinine-based KDIGO criteria and assigned the highest level of severity that occurred during the index hospitalization (30).
Outcomes
Follow-up occurred through December of 2014 and up to 365 days postdischarge with censoring due to health plan disenrollment, death, or end of follow-up. Death was identified from health system databases, state death certificate files, and Social Security Administration files (31,32). At 365 days postdischarge, 6% of the study cohort were censored because of disenrollment.
Our primary outcome was the composite of hospitalization for heart failure, acute coronary syndromes (i.e., acute myocardial infarction or unstable angina), clinically significant peripheral artery disease (i.e., aortic dissection, rupture or repair, and peripheral artery revascularization), or ischemic stroke. Hospitalizations were ascertained from hospitalization databases using validated algorithms on the basis of International Classification of Diseases, Ninth Edition (ICD-9) discharge codes with high positive predictive values for these outcomes compared against medical record review in our health system (31–33). As a secondary outcome, we assessed a modified composite outcome of atherosclerotic events (acute coronary syndromes, peripheral artery disease, or ischemic stroke). In addition, we separately evaluated individual components of the composite outcome and death.
Covariates
Information on age, sex and self-reported race/ethnicity was obtained from electronic health records. Preadmission eGFR (eGFR in ml/min per 1.73 m2) was estimated using the CKD Epidemiology Collaboration equation (34) and proteinuria was determined using a urine dipstick result of ≥1+ (31).
We identified comorbid conditions up to 5 years before the index hospitalization using inpatient and ambulatory diagnoses and procedures, laboratory results, and drug dispensings from pharmacy databases (codes available upon request). Atrial fibrillation and/or atrial flutter was identified using primary discharge diagnoses. Diabetes mellitus was defined on the basis of one or more primary inpatient discharge diagnosis, two or more outpatient diagnoses, or receipt of an antidiabetic drug (35). Hypertension was defined as two or more outpatient diagnoses or one or more outpatient diagnosis plus receipt of an antihypertensive agent (32). Dyslipidemia was defined on the basis of outpatient diagnoses or receipt of antilipemic medications. Smoking status at the time of admission was ascertained from electronic medical records. Cancer was identified using one or more primary discharge diagnosis or two or more outpatient diagnoses of any malignancy other than nonmelanoma skin cancer. Coronary disease was defined as prior myocardial infarction, unstable angina, or coronary revascularization (31,32). Heart failure was defined as one or more primary discharge diagnosis and/or three or more outpatient visits for heart failure (36,37). Peripheral artery disease was defined as more than one primary hospital discharge or emergency department diagnosis or more than one revascularization procedure. Prior ischemic stroke was defined as one or more primary hospital discharge diagnosis (33).
To address potential confounders occurring during the index hospitalization, we identified occurrence of acute heart failure, sepsis, and coronary revascularization, using diagnosis and procedure codes for the index admission (31,36–38). As one measure of severity of acute illness, we identified patients who received care in the ICU and calculated the length of the stay in the ICU and the overall hospitalization. Finally, to further account for acute severity of illness, we calculated a validated laboratory-based acute physiology score, comorbidity point score, and the predicted mortality score, which are on the basis of automated inpatient, outpatient, and laboratory data (39,40).
We assessed for missingness in potential confounders listed above and found our matched study population had <15% missing data across any variable. For variables with significant missingness, we included a categorical value to reflect this.
Propensity Score and Matching
To control for imbalances in covariates between patients with and without AKI, we performed logistic regression using all variables in Table 1 to calculate the propensity of having AKI during the index hospitalization (model c-index =0.82). We then attempted to match five non-AKIs to each AKI, using age (±5 years), sex, year of index hospitalization, and a ≤0.001 difference in propensity score.
Table 1.
Baseline characteristics for adults hospitalized between 2006 and 2013 before and after propensity matching, stratified by AKI status
| Characteristic | Before Matching | After Matching | ||||
|---|---|---|---|---|---|---|
| AKI (n=39,153) | No AKI (n=391,006) | D Value | AKI (n=31,245) | No AKI (n=115,696) | D Value | |
| Age, yr | ||||||
| Mean (SD) | 69 (16) | 62 (18) | 0.38 | 69 (15) | 69 (16) | 0.04 |
| Age group, yr, n (%) | 0.37 | 0.03 | ||||
| 18–49 | 4663 (12) | 101,428 (26) | 3552 (11) | 14,205 (12) | ||
| 50–59 | 5463 (14) | 64,391 (17) | 4218 (14) | 15,995 (14) | ||
| 60–69 | 8359 (21) | 77,462 (20) | 6592 (21) | 24,593 (21) | ||
| 70–79 | 10,129 (26) | 75,320 (19) | 8244 (26) | 30,176 (26) | ||
| ≥80 | 10,539 (27) | 72,405 (19) | 8639 (28) | 30,727 (27) | ||
| Sex, n (%) | 0.17 | 0.02 | ||||
| Women | 19,322 (49) | 226,001 (58) | 15,686 (50) | 59,173 (51) | ||
| Men | 19,831 (51) | 165,005 (42) | 15,559 (50) | 56,523 (49) | ||
| Race/ethnicity, n (%) | 0.05 | 0.02 | ||||
| White | 27,941 (71) | 281,584 (72) | 22,756 (73) | 85,386 (74) | ||
| Black | 4506 (12) | 33,981 (9) | 3287 (11) | 11,754 (10) | ||
| Asian/Pacific Islander | 4748 (12) | 48,634 (12) | 3658 (12) | 13,093 (11) | ||
| Native American | 268 (0.7) | 2655 (0.7) | 206 (0.7) | 699 (0.6) | ||
| Other/unknown | 1690 (4) | 24,152 (6) | 1338 (4) | 4764 (4) | ||
| Hispanic ethnicity, n (%) | 5646 (14) | 58,425 (15) | 0.01 | 4462 (14) | 15,495 (13) | 0.03 |
| Current or former smoker, n (%) | 20,027 (51) | 178,832 (46) | 0.13 | 15,833 (51) | 58,126 (50) | 0.01 |
| Prior medical history, n (%) | ||||||
| Coronary heart disease | 3493 (9) | 24,097 (6) | 0.24 | 2690 (9) | 9754 (8) | 0.01 |
| Chronic heart failure | 7643 (20) | 28,762 (7) | 0.68 | 4779 (15) | 14,401 (12) | 0.14 |
| Ischemic stroke | 898 (2) | 10,118 (3) | 0.07 | 773 (3) | 2916 (3) | 0.01 |
| Peripheral artery disease | 2685 (7) | 15,285 (4) | 0.36 | 2071 (7) | 6925 (6) | 0.07 |
| Atrial fibrillation or flutter | 5266 (13) | 35,035 (9) | 0.28 | 4217 (13) | 14,708 (13) | 0.04 |
| Diabetes Mellitus | 15,555 (40) | 83,724 (21) | 0.54 | 11,256 (36) | 38,820 (34) | 0.07 |
| Hypertension | 30,834 (79) | 220,419 (56) | 0.64 | 23,015 (74) | 83,170 (72) | 0.05 |
| Dyslipidemia | 27,273 (70) | 207,974 (53) | 0.43 | 21,314 (68) | 77,189 (67) | 0.04 |
| Cancer | 9157 (23) | 70,462 (18) | 0.20 | 7498 (24) | 27,605 (24) | 0.00 |
| Proteinuria | 15,101 (39) | 77,139 (20) | 0.57 | 10,490 (34) | 32,290 (28) | 0.16 |
| Preadmission eGFR, ml/min per 1.73 m2, n (%) | 0.10 | 0.08 | ||||
| >150 | 38 (0.1) | 249 (0.1) | 32 (0.1) | 53 (0.0) | ||
| 90–150 | 6269 (16) | 80,484 (21) | 5131 (16) | 21,095 (18) | ||
| 60–89 | 11,929 (31) | 125,959 (32) | 10,256 (33) | 41,404 (36) | ||
| 45–59 | 6947 (18) | 41,133 (11) | 5549 (18) | 20,342 (18) | ||
| 30–44 | 6030 (15) | 21,162 (5) | 4301 (14) | 13,106 (11) | ||
| 15–29 | 3164 (8) | 6297 (2) | 1754 (6) | 4107 (4) | ||
| Unknown | 4776 (12) | 115,722 (30) | 4222 (14) | 15,589 (14) | ||
| Preadmission medication use, n (%) | ||||||
| Angiotensin-converting enzyme inhibitor | 16,823 (43) | 100,143 (26) | 0.47 | 12,753 (41) | 45,526 (39) | 0.04 |
| Angiotensin II receptor blocker | 4974 (13) | 31,458 (8) | 0.31 | 3812 (12) | 13,657 (12) | 0.02 |
| Diuretic | 19,823 (51) | 119,145 (31) | 0.52 | 14,998 (48) | 52,503 (45) | 0.06 |
| β-Blocker | 17,353 (44) | 112,333 (29) | 0.41 | 13,179 (42) | 46,701 (40) | 0.05 |
| Calcium channel blocker | 9520 (24) | 53,340 (14) | 0.43 | 6963 (22) | 23,662 (21) | 0.07 |
| Statin | 19,373 (50) | 133,614 (34) | 0.38 | 15,023 (48) | 53,697 (46) | 0.04 |
| Index hospitalization features, n (%) | ||||||
| Coronary bypass surgery | 77 (0.2) | 216 (0.1) | 0.77 | 58 (0.2) | 130 (0.1) | 0.30 |
| Percutaneous coronary intervention | 84 (0.2) | 842 (0.2) | 0.00 | 69 (0.2) | 338 (0.3) | 0.17 |
| Sepsis | 6945 (18) | 22,651 (6) | 0.76 | 3911 (13) | 11,130 (10) | 0.18 |
| Heart failure | 2746 (7) | 7972 (2) | 0.78 | 1832 (6) | 4793 (4) | 0.22 |
| ICU stay, n (%) | 12,273 (31) | 46,576 (12) | 0.74 | 7494 (24) | 23,024 (20) | 0.14 |
| ICU length of stay, h | ||||||
| Median (IQR) | 68 (35–139) | 7 (22–65) | <0.001a | 53 (29–99) | 40 (23–69) | <0.001a |
| Hospital length of stay, d | ||||||
| Median (IQR) | 6 (3–10) | 3 (2–5) | <0.001a | 5 (3–8) | 4 (2–6) | <0.001a |
| Laboratory-based acute physiology score | ||||||
| Median (IQR) | 27 (9–44) | 6 (0–20) | <0.001a | 24 (6–41) | 16 (1–30) | <0.001a |
| Missing, n (%) | 13 (0.0) | 179 (0.0) | 13 (0.0) | 32 (0.0) | ||
| Comorbidity point score | ||||||
| Median (IQR) | 77 (46–114) | 52 (26–86) | <0.001a | 75 (44–111) | 73 (43–108) | <0.001a |
| Missing, n (%) | 415 (1) | 4586 (1) | 366 (1) | 1319 (1) | ||
| Predicted mortality score | ||||||
| Median (IQR) | 3 (1–7) | 1 (0–2) | <0.001a | 2 (1–6) | 2 (1–4) | <0.001a |
| Missing, n (%) | 1446 (4) | 31,026 (8) | 1337 (4) | 5730 (5) | ||
| Predicted mortality score categorical, n (%) | 0.42 | 0.10 | ||||
| <0.1% | 1774 (5) | 69,083 (18) | 1702 (5) | 6442 (6) | ||
| 0.1%–0.4% | 4083 (10) | 90,413 (23) | 3809 (12) | 16,457 (14) | ||
| 0.5%–1.9% | 9422 (24) | 105,600 (27) | 8292 (27) | 35,639 (31) | ||
| 2.0%–4.9% | 9452 (24) | 53,410 (14) | 7503 (24) | 27,364 (24) | ||
| 5.0%–9.9% | 6654 (17) | 24,650 (6) | 4774 (15) | 14,369 (12) | ||
| 10.0%–14.9% | 2772 (7) | 8438 (2) | 1826 (6) | 4881 (4) | ||
| 15.0%–29.9% | 2621 (7) | 6927 (2) | 1590 (5) | 4018 (4) | ||
| ≥30.0% | 929 (2) | 1459 (0.4) | 412 (1) | 796 (0.7) | ||
| Unknown | 1446 (4) | 31,026 (8) | 1337 (4) | 5730 (5) | ||
The D value represents the standardized difference in means or proportions with a value >0.10 being significant. ICU, intensive care unit; IQR, interquartile range.
Represents P values and not D values.
Statistical Methods
Analyses were performed using SAS version 9.3 (Cary, NC). Continuous variables were reported as means with SDs or medians with interquartile ranges; discrete variables were reported as frequencies and proportions. Given the large sample size, for means and proportions, we compared characteristics between those with and without AKI, using Cohen’s D value, by taking the standardized difference of means or proportions between groups and dividing by the pooled estimate, with a value >0.10 considered significant (41,42). We compared median values using the Wilcoxon rank-sum test, with a P value <0.05 considered significant.
Outcomes through 365 days after discharge were reported as proportions with Cohen’s D value to assess differences between AKI and non-AKI groups, along with cumulative hazards for the primary composite outcome. We applied multivariable stratified Cox regression to assess the association of AKI on outcomes with accounting for matched sets. All variables in Table 1 were included as covariates except for those variables used for matching as well as the laboratory-based acute physiology score and comorbidity point scores, as they were incorporated into the summary predicted mortality score. We adjusted for onset of ESKD (i.e., initiation of RRT) during follow-up; in sensitivity analyses, we censored at ESKD and found that results were unchanged (data not shown), so only the main results are reported.
We assessed for a possible graded relationship between AKI severity with the primary composite outcome. Finally, we performed analyses separately by index age (<65, 65–74, and ≥75 years) and by sex.
Results
Study Population
During the study period, we identified 430,159 eligible adults who were discharged alive, with 39,153 (9%) experiencing AKI (70% stage 1, 14% stage 2, and 16% stage 3) (Figure 1). We successfully matched 31,245 (80%) patients with AKI with up to five patients without AKI on the basis of year of index hospitalization, age, sex, and propensity score, with a median of 5 (interquartile range, 2–5) matched patients without AKI for each patient with AKI. Our matching approach yielded successful balance across most patient characteristics, with the exception of patients with AKI being more likely than matched patients without AKI to have prior heart failure and proteinuria (Table 1). During the index hospitalization, patients with AKI were also more likely to receive coronary revascularization, be diagnosed with sepsis or heart failure, stay in the ICU, and have longer length of stay, but had no material differences in measures of severity of acute illness or predicted short-term mortality (Table 1).
Heart Failure and Atherosclerotic Outcomes
Overall, 10,605 patients experienced the composite outcome of heart failure and atherosclerotic events (4636 heart failure, 2399 acute coronary syndrome, 3042 peripheral artery disease, and 1410 ischemic stroke events) during the first 365 days after discharge (Table 2). Compared with those without AKI, those with AKI had higher crude risks for the primary composite outcome and for heart failure alone, but not for atherosclerotic events alone (Figure 2, Table 2). After adjustment for preadmission patient characteristics and key index hospitalization features (i.e., length of stay, ICU stay, sepsis, heart failure, receipt of coronary revascularization, and predicted mortality score), we found that AKI was associated with an 18% (95% confidence interval [95% CI], 13% to 25%) higher rate of the composite outcome of heart failure and atherosclerotic events (Figure 3). The multivariable association between AKI and the primary composite outcome was similar across age (adjusted hazard ratio [aHR], 1.27; 95% CI, 1.10 to 1.46 for age <65 years, aHR, 1.17; 95% CI, 1.04 to 1.31 for age 65–74 years, and aHR, 1.18; 95% CI, 1.11 to 1.27 for age ≥75 years) and sex (aHR, 1.20; 95% CI, 1.11 to 1.29 for women and aHR, 1.17; 95% CI, 1.09 to 1.25 for men).
Table 2.
Crude risks of the heart failure, atherosclerotic events, and death at 365 days after discharge among hospitalized adults, stratified by AKI status
| At 365 d postdischarge, n (%) | Overall (n=146,941) | AKI (n=31,245) | No AKI (n=115,696) | D Value |
|---|---|---|---|---|
| Composite of atherosclerotic events and heart failure | 10,605 (7) | 2765 (9) | 7840 (7) | 0.18 |
| Composite of atherosclerotic events | 6575 (5) | 1533 (5) | 5042 (4) | 0.08 |
| Heart failure | 4636 (3) | 1423 (5) | 3213 (3) | 0.31 |
| Acute coronary syndrome | 2399 (2) | 577 (2) | 1822 (2) | 0.10 |
| Peripheral artery disease | 3042 (2) | 725 (2) | 2317 (2) | 0.09 |
| Ischemic stroke | 1410 (1) | 307 (1) | 1103 (1) | 0.02 |
| Death | 22,759 (16) | 5963 (19) | 16,796 (15) | 0.20 |
A D value >0.10 is considered statistically significant.
Figure 2.
Cumulative incidence of heart failure or atherosclerotic events at 365 days after discharge higher among hospitalized adults who did or did not experience AKI.
Figure 3.
AKI is independently associated with all cardiovascular events, heart failure, and death, but not with individual atherosclerotic events at 365 days after discharge among hospitalized adults. All models adjusted for matching strata, race, prior medical history (coronary heart disease, heart failure, peripheral arterial disease, stroke, atrial fibrillation/flutter, hypertension, diabetes, and cancer), preadmission eGFR and documented proteinuria, diagnosis of selected medical conditions during the hospitalization (heart failure, sepsis), admission to the ICU, length of stay, predicted mortality score, and receipt of dialysis and/or transplant during follow-up.
Importantly, the adjusted rate of hospitalization for heart failure was 44% higher (95% CI, 33% to 56%) for those with AKI, but there were no significant multivariable associations between AKI and atherosclerotic events (composite or individual atherosclerotic event types) (Figure 3). Results were similar in analyses stratified by age group or sex (data not shown).
Of note, in evaluating the association between AKI severity and the primary outcome at 365 days after discharge, we found that AKI stage 1 was associated with an adjusted 25% (95% CI, 19% to 33%) higher rate of the composite outcome of heart failure or atherosclerotic events, a 54% (95% CI. 41% to 69%) higher adjusted rate of heart failure, and a 9% (95% CI, 1% to 18%) higher adjusted rate of atherosclerotic events. In contrast, there were no statistically significant multivariable associations for those with stage 2 or stage 3 AKI.
Death
Given the potential concern for competing risk of death associated with AKI in evaluating the strength of association between AKI and cardiovascular events, we separately examined postdischarge rates of death. More patients with AKI died compared with matched patients without AKI during the first 365 days after discharge (19% versus 15%, respectively; D-value =0.20) (Table 2). In multivariable analyses, AKI was associated with a higher adjusted rate of all-cause death postdischarge (aHR, 1.20; 95% CI, 1.16 to 1.24) (Figure 3).
Discussion
Within a large, diverse community-based population of hospitalized adults, AKI was associated with significantly higher rates of cardiovascular events during the first 365 days after hospital discharge, even after adjustment for a wide range of characteristics, measures of acute severity of illness, predicted short-term mortality and differences in medical therapy. This excess risk was driven by heart failure events, as AKI was not significantly associated with atherosclerotic events after accounting for potential confounders.
AKI is a known risk factor for early and late death, but it is unclear whether this may be explained, in part, through an excess risk of cardiovascular events. AKI can lead to higher levels of circulating inflammatory cytokines (9), may contribute to endothelial dysfunction (10), and promote or worsen cardiac ischemia in the short-term (13). AKI may also be associated with acute elevations in certain markers of abnormal mineral metabolism (e.g., fibroblast growth factor 23) which could induce or worsen left ventricular hypertrophy (11,43). However, there have been limited rigorous evaluations of the potential link between AKI and subsequent clinical cardiovascular events. Although there has recently been an increasing number of studies examining the potential cardiovascular-related complications of AKI in various clinical settings or patient populations (14–18,20,44–48), many published studies have focused on selected patients undergoing coronary angiography, with a meta-analysis (combined n=70,031) suggesting that AKI complicating coronary angiography is associated with a higher risk of major cardiovascular events (which could include cardiovascular death, myocardial infarction, target vessel reocclusion or need for revascularization, stroke, heart failure, or a composite) (14). However, there was substantial heterogeneity across studies as well as limitations in including “soft” events (e.g., receipt of coronary revascularization) in the analyses that are often the most frequent outcome and most susceptible to subjective clinical decision-making (14). A more recent meta-analysis that focused on cardiovascular death, major adverse cardiovascular events, myocardial infarction, heart failure, and stroke reported a higher risk of these events after AKI, but used varying definitions of AKI, did not incorporate peripheral artery disease events, did not evaluate a specific follow-up period postdischarge, and only adjusted for a modest number of possible confounders (22).
A large Veterans Affairs-based study focused on AKI in the setting of myocardial infarction demonstrated that compared with those who had myocardial infarction alone, those with AKI and myocardial infarction had a higher subsequent rate of a composite cardiovascular end point (49). However, that study did not examine individual types of cardiovascular outcomes or address peripheral artery disease events, and we found that the associations differed by type of outcome. A retrospective study using only inpatient claims data from Taiwan between 1999 and 2008 compared 4869 patients who recovered from dialysis-requiring AKI and survived ≥30 days postdischarge and had no readmission or reinitiation of dialysis with a group of 4869 patients without dialysis-requiring AKI matched on age, sex, calendar year, and propensity-matched on claims-based comorbidities (21). During mean follow-up of 3.4 years, dialysis-requiring AKI was associated with a higher adjusted rate (hazard ratio, 1.67; 95% CI, 1.36 to 2.04) of a composite of “coronary events” defined as nonfatal myocardial infarction, coronary bypass surgery, or diagnostic coronary angiography (21). This study was limited by only studying the most severe AKI, limited control for relevant confounders, inclusion of coronary angiography and bypass surgery as part of their primary outcome, and concerns about generalizability to other populations.
In a prospective cohort of 968 older adults undergoing cardiac surgery, Parikh et al. (18) studied the association of AKI with a composite end point of death or hospitalization for acute coronary syndrome, myocardial infarction, heart failure, or coronary revascularization. AKI Network stage 1 and stages 2 and 3 were associated with higher adjusted rates of the composite end point (hazard ratio, 1.99; 95% CI, 1.46 to 2.71 and hazard ratio, 3.52; 95% CI, 2.17 to 5.71, respectively), but this study was also limited by inclusion of coronary revascularization procedures and all-cause mortality as part of their primary outcome in addition to focusing only on selected cardiac surgery patients and adjustment for a limited number of potential confounding factors (18).
Our study materially expands on existing literature by examining a larger, contemporary, multicenter, and carefully matched population with a wide range of AKI severity, and clarifies the association between an AKI episode and primarily the risk of subsequent hospitalization for heart failure during the first year after discharge. The higher risk was present despite excess mortality from any cause after discharge associated with AKI. The reasons for why AKI was more strongly associated with subsequent heart failure than atherosclerotic events are not clear, but if the association is proven to be causal, it may be related to differential effects of aberrations in inflammatory, endothelial dysfunction, and other pathways associated with AKI or possible differences in therapy (e.g., use or dosage of renin-angiotensin system inhibitors or diuretics) after an AKI episode. Our population had broad demographic diversity with comprehensive ascertainment of AKI on the basis of recommended KDIGO criteria (rather than reliance on administrative diagnostic codes) as well as postdischarge, clinically important “hard” cardiovascular events (i.e., heart failure, acute coronary syndromes, peripheral artery disease, and ischemic stroke). Compared with previous studies, we were also able to carefully account for a much wider set of potential confounders, including relevant preadmission patient features, factors during the index hospitalization, longitudinal treatment, as well as a validated measure of acute illness severity (39).
Our study also had several limitations. Even though we carefully matched patients and further adjusted for a wide range of confounders, we cannot confirm a causal association. We conducted this study in a large, community-based population with comprehensive longitudinal follow-up, which should enhance generalizability, but our results may not be fully applicable to uninsured patients or to all settings. Although we defined AKI on the basis of recommended KDIGO criteria using changes in serum creatinine concentration, additional diagnostic information on the basis of urine output was unavailable. Given the known difficulty in determining cause(s) of an AKI episode, we were unable to evaluate potential differences in cardiovascular risk by AKI etiology. Information on potential specific pathways that may mediate the association between AKI and subsequent cardiovascular events was unavailable. Also, because there was no systematic follow-up testing within the index hospitalization and postdischarge in this clinical database study, we were unable to address the rates of AKI recovery and the potential mediating effect on subsequent outcomes. Finally, although we relied on validated algorithms to detect the occurrence of heart failure and atherosclerotic events in our health system, we were unable to address the negative predictive value of these algorithms on the basis of hospitalization records, as well as potentially missed cardiovascular events among observed deaths as the underlying cause was unavailable.
In conclusion, an episode of AKI complicating a hospitalization was independently associated with a higher risk of subsequent heart failure but not atherosclerotic events. If these associations are confirmed, additional research is needed to delineate underlying mechanisms driving the excess risk of heart failure events, to clarify longer-term risks, and to determine whether prevention of AKI or more aggressive screening and risk reduction measures after AKI improves overall outcomes.
Disclosures
A.S.G. has received a research grant through his institution from Astra-Zeneca. None of the other authors have any relevant disclosures.
Acknowledgments
This study was supported by research grants R01DK098233, U01DK082223, and R01DK067126 from the National Institute of Diabetes, Digestive and Kidney Disease, National Institutes of Health.
The funders had no role in the design or conduct of the study, collection, management, analysis or interpretation of the data, or preparation, review, or approval of the manuscript.
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
See related editorial, “Ptolemy and Copernicus Revisited: The Complex Interplay between the Kidneys and Heart Failure,” on pages 825–828.
References
- 1.Hsu RK, McCulloch CE, Dudley RA, Lo LJ, Hsu CY: Temporal changes in incidence of dialysis-requiring AKI. J Am Soc Nephrol 24: 37–42, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hsu CY, McCulloch CE, Fan D, Ordoñez JD, Chertow GM, Go AS: Community-based incidence of acute renal failure. Kidney Int 72: 208–212, 2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Lo LJ, Go AS, Chertow GM, McCulloch CE, Fan D, Ordoñez JD, Hsu CY: Dialysis-requiring acute renal failure increases the risk of progressive chronic kidney disease. Kidney Int 76: 893–899, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hsu CY, Chertow GM, McCulloch CE, Fan D, Ordoñez JD, Go AS: Nonrecovery of kidney function and death after acute on chronic renal failure. Clin J Am Soc Nephrol 4: 891–898, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Amdur RL, Chawla LS, Amodeo S, Kimmel PL, Palant CE: Outcomes following diagnosis of acute renal failure in U.S. veterans: Focus on acute tubular necrosis. Kidney Int 76: 1089–1097, 2009 [DOI] [PubMed] [Google Scholar]
- 6.Wald R, Quinn RR, Luo J, Li P, Scales DC, Mamdani MM, Ray JG; University of Toronto Acute Kidney Injury Research Group : Chronic dialysis and death among survivors of acute kidney injury requiring dialysis. JAMA 302: 1179–1185, 2009 [DOI] [PubMed] [Google Scholar]
- 7.Palevsky PM, Zhang JH, O’Connor TZ, Chertow GM, Crowley ST, Choudhury D, Finkel K, Kellum JA, Paganini E, Schein RM, Smith MW, Swanson KM, Thompson BT, Vijayan A, Watnick S, Star RA, Peduzzi P; VA/NIH Acute Renal Failure Trial Network : Intensity of renal support in critically ill patients with acute kidney injury. N Engl J Med 359: 7–20, 2008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bellomo R, Cass A, Cole L, Finfer S, Gallagher M, Lo S, McArthur C, McGuinness S, Myburgh J, Norton R, Scheinkestel C, Su S; RENAL Replacement Therapy Study Investigators : Intensity of continuous renal-replacement therapy in critically ill patients. N Engl J Med 361: 1627–1638, 2009 [DOI] [PubMed] [Google Scholar]
- 9.Hoke TS, Douglas IS, Klein CL, He Z, Fang W, Thurman JM, Tao Y, Dursun B, Voelkel NF, Edelstein CL, Faubel S: Acute renal failure after bilateral nephrectomy is associated with cytokine-mediated pulmonary injury. J Am Soc Nephrol 18: 155–164, 2007 [DOI] [PubMed] [Google Scholar]
- 10.Ko GJ, Grigoryev DN, Linfert D, Jang HR, Watkins T, Cheadle C, Racusen L, Rabb H: Transcriptional analysis of kidneys during repair from AKI reveals possible roles for NGAL and KIM-1 as biomarkers of AKI-to-CKD transition. Am J Physiol Renal Physiol 298: F1472–F1483, 2010 [DOI] [PubMed] [Google Scholar]
- 11.Christov M, Waikar SS, Pereira RC, Havasi A, Leaf DE, Goltzman D, Pajevic PD, Wolf M, Jüppner H: Plasma FGF23 levels increase rapidly after acute kidney injury. Kidney Int 84: 776–785, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Leaf DE, Waikar SS, Wolf M, Cremers S, Bhan I, Stern L: Dysregulated mineral metabolism in patients with acute kidney injury and risk of adverse outcomes. Clin Endocrinol (Oxf) 79: 491–498, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Song D, de Zoysa JR, Ng A, Chiu W: Troponins in acute kidney injury. Ren Fail 34: 35–39, 2012 [DOI] [PubMed] [Google Scholar]
- 14.James MT, Samuel SM, Manning MA, Tonelli M, Ghali WA, Faris P, Knudtson ML, Pannu N, Hemmelgarn BR: Contrast-induced acute kidney injury and risk of adverse clinical outcomes after coronary angiography: A systematic review and meta-analysis. Circ Cardiovasc Interv 6: 37–43, 2013 [DOI] [PubMed] [Google Scholar]
- 15.Hansen MK, Gammelager H, Jacobsen CJ, Hjortdal VE, Layton JB, Rasmussen BS, Andreasen JJ, Johnsen SP, Christiansen CF: Acute kidney injury and long-term risk of cardiovascular events after cardiac surgery: A population-based cohort study. J Cardiothorac Vasc Anesth 29: 617–625, 2015 [DOI] [PubMed] [Google Scholar]
- 16.Hansen MK, Gammelager H, Mikkelsen MM, Hjortdal VE, Layton JB, Johnsen SP, Christiansen CF: Post-operative acute kidney injury and five-year risk of death, myocardial infarction, and stroke among elective cardiac surgical patients: A cohort study. Crit Care 17: R292, 2013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Olsson D, Sartipy U, Braunschweig F, Holzmann MJ: Acute kidney injury following coronary artery bypass surgery and long-term risk of heart failure. Circ Heart Fail 6: 83–90, 2013 [DOI] [PubMed] [Google Scholar]
- 18.Parikh CR, Puthumana J, Shlipak MG, Koyner JL, Thiessen-Philbrook H, McArthur E, Kerr K, Kavsak P, Whitlock, Garg AX, Coca SG: Relationship of kidney injury biomarkers with long-term cardiovascular outcomes after cardiac surgery. J Am Soc Nephrol 28: 3699–3707, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Parikh CR, Coca SG, Wang Y, Masoudi FA, Krumholz HM: Long-term prognosis of acute kidney injury after acute myocardial infarction. Arch Intern Med 168: 987–995, 2008 [DOI] [PubMed] [Google Scholar]
- 20.Szczech LA, Granger CB, Dasta JF, Amin A, Peacock WF, McCullough PA, Devlin JW, Weir MR, Katz JN, Anderson FA Jr., Wyman A, Varon J; Studying the Treatment of Acute Hypertension Investigators : Acute kidney injury and cardiovascular outcomes in acute severe hypertension. Circulation 121: 2183–2191, 2010 [DOI] [PubMed] [Google Scholar]
- 21.Wu VC, Wu CH, Huang TM, Wang CY, Lai CF, Shiao CC, Chang CH, Lin SL, Chen YY, Chen YM, Chu TS, Chiang WC, Wu KD, Tsai PR, Chen L, Ko WJ; NSARF Group : Long-term risk of coronary events after AKI. J Am Soc Nephrol 25: 595–605, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Odutayo A, Wong CX, Farkouh M, Altman DG, Hopewell S, Emdin CA, Hunn BH: AKI and long-term risk for cardiovascular events and mortality. J Am Soc Nephrol 28: 377–387, 2017 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Warren-Gash C, Hayward AC, Hemingway H, Denaxas S, Thomas SL, Timmis AD, Whitaker H, Smeeth L: Influenza infection and risk of acute myocardial infarction in England and Wales: A CALIBER self-controlled case series study. J Infect Dis 206: 1652–1659, 2012 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Smeeth L, Thomas SL, Hall AJ, Hubbard R, Farrington P, Vallance P: Risk of myocardial infarction and stroke after acute infection or vaccination. N Engl J Med 351: 2611–2618, 2004 [DOI] [PubMed] [Google Scholar]
- 25.Corrales-Medina VF, Alvarez KN, Weissfeld LA, Angus DC, Chirinos JA, Chang CC, Newman A, Loehr L, Folsom AR, Elkind MS, Lyles MF, Kronmal RA, Yende S: Association between hospitalization for pneumonia and subsequent risk of cardiovascular disease. JAMA 313: 264–274, 2015 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Corrales-Medina VF, Madjid M, Musher DM: Role of acute infection in triggering acute coronary syndromes. Lancet Infect Dis 10: 83–92, 2010 [DOI] [PubMed] [Google Scholar]
- 27.Elkind MS, Carty CL, O’Meara ES, Lumley T, Lefkowitz D, Kronmal RA, Longstreth WT Jr.: Hospitalization for infection and risk of acute ischemic stroke: The Cardiovascular Health Study. Stroke 42: 1851–1856, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Krieger N: Overcoming the absence of socioeconomic data in medical records: Validation and application of a census-based methodology. Am J Public Health 82: 703–710, 1992 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Gordon NP: Characteristics of Adult Health Plan Members in the Northern California Region Membership, as Estimated from the 2011 Member Health Survey, Oakland, CA, Division of Research, Kaiser Permanente Medical Care Program, 2013 [Google Scholar]
- 30.Kidney Disease: Improving Global Outcomes (KDIGO) Work Group: KDIGO clinical practice guideline for acute kidney injury. Kidney Int [Suppl 2]: 1–138, 2012 [Google Scholar]
- 31.Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY: Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med 351: 1296–1305, 2004 [DOI] [PubMed] [Google Scholar]
- 32.Yeh RW, Sidney S, Chandra M, Sorel M, Selby JV, Go AS: Population trends in the incidence and outcomes of acute myocardial infarction. N Engl J Med 362: 2155–2165, 2010 [DOI] [PubMed] [Google Scholar]
- 33.Go AS, Hylek EM, Chang Y, Phillips KA, Henault LE, Capra AM, Jensvold NG, Selby JV, Singer DE: Anticoagulation therapy for stroke prevention in atrial fibrillation: How well do randomized trials translate into clinical practice? JAMA 290: 2685–2692, 2003 [DOI] [PubMed] [Google Scholar]
- 34.Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF 3rd, Feldman HI, Kusek JW, Eggers P, Van Lente F, Greene T, Coresh J; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) : A new equation to estimate glomerular filtration rate. Ann Intern Med 150: 604–612, 2009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Selby JV, Ray GT, Zhang D, Colby CJ: Excess costs of medical care for patients with diabetes in a managed care population. Diabetes Care 20: 1396–1402, 1997 [DOI] [PubMed] [Google Scholar]
- 36.Go AS, Lee WY, Yang J, Lo JC, Gurwitz JH: Statin therapy and risks for death and hospitalization in chronic heart failure. JAMA 296: 2105–2111, 2006 [DOI] [PubMed] [Google Scholar]
- 37.Go AS, Yang J, Ackerson LM, Lepper K, Robbins S, Massie BM, Shlipak MG: Hemoglobin level, chronic kidney disease, and the risks of death and hospitalization in adults with chronic heart failure: The Anemia in Chronic Heart Failure: Outcomes and Resource Utilization (ANCHOR) Study. Circulation 113: 2713–2723, 2006 [DOI] [PubMed] [Google Scholar]
- 38.Gaieski DF, Edwards JM, Kallan MJ, Carr BG: Benchmarking the incidence and mortality of severe sepsis in the United States. Crit Care Med 41: 1167–1174, 2013 [DOI] [PubMed] [Google Scholar]
- 39.Liu V, Kipnis P, Gould MK, Escobar GJ: Length of stay predictions: Improvements through the use of automated laboratory and comorbidity variables. Med Care 48: 739–744, 2010 [DOI] [PubMed] [Google Scholar]
- 40.Escobar GJ, Greene JD, Scheirer P, Gardner MN, Draper D, Kipnis P: Risk-adjusting hospital inpatient mortality using automated inpatient, outpatient, and laboratory databases. Med Care 46: 232–239, 2008 [DOI] [PubMed] [Google Scholar]
- 41.Austin PC: Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simulat Comput 38: 1228–1234, 2009 [Google Scholar]
- 42.Austin PC: An introduction to propensity score methods for reducing the effects of confounding in observational studies. Multivariate Behav Res 46: 399–424, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Faul C, Amaral AP, Oskouei B, Hu MC, Sloan A, Isakova T, Gutiérrez OM, Aguillon-Prada R, Lincoln J, Hare JM, Mundel P, Morales A, Scialla J, Fischer M, Soliman EZ, Chen J, Go AS, Rosas SE, Nessel L, Townsend RR, Feldman HI, St John Sutton M, Ojo A, Gadegbeku C, Di Marco GS, Reuter S, Kentrup D, Tiemann K, Brand M, Hill JA, Moe OW, Kuro-O M, Kusek JW, Keane MG, Wolf M: FGF23 induces left ventricular hypertrophy. J Clin Invest 121: 4393–4408, 2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Giacoppo D, Madhavan MV, Baber U, Warren J, Bansilal S, Witzenbichler B, Dangas GD, Kirtane AJ, Xu K, Kornowski R, Brener SJ, Généreux P, Stone GW, Mehran R: Impact of contrast-induced acute kidney injury after percutaneous coronary intervention on short- and long-term outcomes: Pooled analysis from the HORIZONS-AMI and ACUITY trials. Circ Cardiovasc Interv 8: e002475, 2015 [DOI] [PubMed] [Google Scholar]
- 45.James MT, Ghali WA, Knudtson ML, Ravani P, Tonelli M, Faris P, Pannu N, Manns BJ, Klarenbach SW, Hemmelgarn BR; Alberta Provincial Project for Outcome Assessment in Coronary Heart Disease (APPROACH) Investigators : Associations between acute kidney injury and cardiovascular and renal outcomes after coronary angiography. Circulation 123: 409–416, 2011 [DOI] [PubMed] [Google Scholar]
- 46.Liotta M, Olsson D, Sartipy U, Holzmann MJ: Minimal changes in postoperative creatinine values and early and late mortality and cardiovascular events after coronary artery bypass grafting. Am J Cardiol 113: 70–75, 2014 [DOI] [PubMed] [Google Scholar]
- 47.Thakar CV, Parikh PJ, Liu Y: Acute kidney injury (AKI) and risk of readmissions in patients with heart failure. Am J Cardiol 109: 1482–1486, 2012 [DOI] [PubMed] [Google Scholar]
- 48.Warren J, Mehran R, Baber U, Xu K, Giacoppo D, Gersh BJ, Guagliumi G, Witzenbichler B, Magnus Ohman E, Pocock SJ, Stone GW: Incidence and impact of acute kidney injury in patients with acute coronary syndromes treated with coronary artery bypass grafting: Insights from the Harmonizing Outcomes With Revascularization and Stents in Acute Myocardial Infarction (HORIZONS-AMI) and Acute Catheterization and Urgent Intervention Triage Strategy (ACUITY) trials. Am Heart J 171: 40–47, 2016 [DOI] [PubMed] [Google Scholar]
- 49.Chawla LS, Amdur RL, Shaw AD, Faselis C, Palant CE, Kimmel PL: Association between AKI and long-term renal and cardiovascular outcomes in United States veterans. Clin J Am Soc Nephrol 9: 448–456, 2014 [DOI] [PMC free article] [PubMed] [Google Scholar]