Effects of acute kidney injury and chronic kidney disease on long-term mortality after coronary artery bypass grafting Seung Seok Han, MD, a,c Nara Shin, MD, a,c Seon Ha Baek, MD, a,b Shin Young Ahn, MD, a,b Dong Ki Kim, MD, PhD, a Sejoong Kim, MD, PhD, a,b Ho Jun Chin, MD, PhD, a,b Dong-Wan Chae, MD, PhD, a,b and Ki Young Na, MD, PhD a,b Seoul, Seongnam, Republic of Korea

Background Both acute kidney injury (AKI) and chronic kidney disease (CKD) are important issues in patients undergoing coronary artery bypass grafting (CABG), particularly with regard to mortality. However, their synergistic or discrete effects on long-term mortality remain unresolved. Methods A total of 1,899 patients undergoing CABG were retrospectively analyzed. The adjusted hazard ratios for allcause mortality were calculated after stratifying the timeframes. To evaluate the synergistic effects between AKI and CKD, the relative excess risk due to interaction was applied. Results The presence of AKI, CKD, or both increased the hazard ratios for mortality, compared with the absence of both: AKI alone, 1.84 (1.464-2.319); CKD alone, 2.46 (1.735-3.478); and AKI and CKD together, 3.21 (2.301-4.488). However, the relationships with mortality were different between AKI and CKD, according to the timeframes: AKI primarily affected early mortality, particularly within 3 years, whereas CKD had a relatively constant effect on both the early and late periods. When the parameters from the relative excess risk due to interaction were obtained, there was a synergistic additive effect on early mortality between AKI and CKD. Conclusions The relationships with mortality after CABG were different between AKI and CKD. However, their effects were not exclusive but synergistic. (Am Heart J 2015;169:419-25.)

Coronary artery bypass grafting (CABG) is one of the most common surgical procedures, and it has been proven to improve the survival of patients with severe angina. 1 The perioperative mortality rate has reached approximately 5%, 2 but this rate is highly dependent on comorbid disease and postoperative complications. 3 In this regard, current predictive models using these conditions have successfully predicted the short-term or mid-term mortality risks after CABG. 3,4 Regarding the comorbid conditions from predictive models, chronic kidney disease (CKD) strongly affects mortality; in particular, in the advanced stages of CKD, the

From the aDepartment of Internal Medicine, Seoul National University College of Medicine, Seoul, Republic of Korea, and bDepartment of Internal Medicine, Seoul National University Bundang Hospital, Seongnam, Republic of Korea. c Both authors contributed equally to this research. Submitted August 4, 2014; accepted December 20, 2014. Reprint requests: Ki Young Na, MD, PhD, Department of Internal Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173 Beon-gil, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-707, Republic of Korea. E-mail: [email protected] 0002-8703 © 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ahj.2014.12.019

short-term mortality rate reaches 10%. 5 As a postoperative complication, acute kidney injury (AKI) develops in up to 30% of cases after CABG, and it is a major cause of morbidity and mortality. 6 Despite a decreasing trend in its incidence, the current rate of AKI remains high. 7 CKD and AKI have often been studied with regard to mortality after CABG, but no studies have examined the effects of AKI and CKD together on overall mortality. Furthermore, studies with extended periods have been limited. 8-10 The relationship of AKI or CKD with mortality might be different according to the timeframe (eg, short term or long term), particularly in cases of AKI recovery or CKD progression. Therefore, we aimed to address the following issues from a CABG cohort: the difference between AKI and CKD in the correlations with short- and long-term mortality, possible synergistic effects between AKI and CKD, and the clinical significance of recovery after AKI.

Methods Participants and data collection Data on patients undergoing CABG were obtained retrospectively from a cohort, consisting of patients from 2 tertiary referral centers (Seoul National University

American Heart Journal March 2015

420 Han et al

Hospital and Seoul National University Bundang Hospital). A total of 2,278 patients consecutively underwent CABG from January 2004 to December 2010. Patients were excluded from the analysis if they were undergoing renal replacement therapy before surgery (n = 44), if they had end-stage renal disease (n = 30), if they underwent redo-CABG (n = 30), if they underwent concomitant valve surgery (n = 168), or if their electronic medical records were incomplete (n = 107). Therefore, 1,899 patients were ultimately analyzed in the present study. The clinical parameters recorded included the following: age; sex; height; weight; current smoking status; systolic blood pressure; diastolic blood pressure; hypertension; diabetes mellitus; histories of myocardial infarction or stroke; peripheral vascular disease; medications such as angiotensin-converting enzyme inhibitors/angiotensin receptor blockers, β-blockers, diuretics, and statins; history of coronary angiography within 1 week before surgery; intraoperative use of cardiopulmonary bypass; and perioperative use of an intra-aortic balloon pump. Body mass index was calculated as body weight (kilograms) divided by the square of body height (meters). Left ventricular ejection fraction was determined by Simpson's modified biplane method from the apical 2chamber and 4-chamber views on echocardiography. Preoperative blood findings, including creatinine, white blood cells, hemoglobin level, albumin, and cholesterol, were measured. The estimated glomerular filtration rate (eGFR) was calculated using the Modification of Diet in Renal Disease equation 11 and the validated coefficient for Korean subjects 12 as follows: 107.904 × serum creatinine (mg/dL) –1.009 × age (years) –0.02 (×0.667, if female). According to the guideline proposed by Kidney Disease Improving Global Outcomes (KDIGO), 13 CKD was defined and categorized from stage 3a to stage 5 using preoperative eGFR. The risk of AKI was measured from surgery to 48 hours. For the definition and staging of AKI, concentrations of both preoperative and postoperative serum creatinine were used. In adherence to the KDIGO guideline, 14 AKI was defined in the case with either an increase in serum creatinine by ≥0.3 mg/dL or ≥1.5 times from the baseline. Recovery from AKI was considered at 1 and 3 months after CABG. When the subsequent serum creatinine level was equal to or lower than the preoperative creatinine, this case was determined to have recovered from AKI. Cases in which the subsequent creatinine level was higher than the baseline level or a patient with AKI died within the decision time were defined as nonrecovery cases. The primary outcome was all-cause mortality after surgery. Except the death-censored cases, all subjects were followed till May 31, 2014. The mortality data were obtained from the national database of Statistics Korea; all Koreans have a resident registration number, and death events are recorded and stored at a national level. The study protocol complies with the Declaration of Helsinki and received full approval from the institutional

review boards at both the Seoul National University Bundang Hospital (no. B-1403/244-112) and the Seoul National University Hospital (H-1403-059-565). No extramural funding was used to support this work.

Statistical analysis All of the analyses and calculations were performed using STATA software (version 12.0; Stata Corp LP, College Station, TX). The data are presented as the means ± SDs for continuous variables and as proportions for categorical variables. Based on variable distributions using histograms, the variables with nonnormal distributions are expressed as the median (interquartile ranges). The χ 2 test was used to compare categorical variables. Comparisons between normally and nonnormally distributed continuous variables were performed using the Student t test and Mann-Whitney U test, respectively. Survival curves were drawn using the Kaplan-Meier method. To compare survival curves between groups, the log-rank test was initially applied. To calculate the hazard ratios (HRs) and CIs of mortality risk, a Cox proportional hazard model was used, with and without the adjustment of all covariates or according to the timeframes. When examining the log-minus-log plot, the assumption of proportional hazards was reasonable. The discrimination of predicting mortality was assessed by calculating the receiver operating characteristic (ROC) curve and the area under the curve (AUC), in which both the definitions and stages of AKI or CKD were considered. For AUCs of long-term outcomes (eg, ROC curve for the prediction of 7-year mortality), patients who were alive and followed up b7 years were excluded. To evaluate the synergistic effects on mortality between AKI and CKD, the relative excess risk due to interaction (RERI) was used, based on the logistic regression analyses. 15 The RERI is an approach for estimating the additive interaction of 2 variables. From the RERI method, we present 3 scales: RERI (part of the total effect that is due to interaction), attributable proportion (AP) (proportion of the combined effect that is due to interaction), and synergy index (ratio between combined effect and individual effects). Positive results for the RERI and AP and a value N1 for the synergy index indicate a positive interaction or more than additive between the variables. P b .05 was considered to be significant. The authors are solely responsible for the design and conduct of this study, all study analyses, and drafting and editing of the manuscript.

Results Baseline characteristics The baseline characteristics of the patients are shown and compared between the groups in Table I. Among the subjects, 11.4% and 33.3% had baseline CKD and postoperative AKI, respectively. A total of 46.5% of the

American Heart Journal Volume 169, Number 3

Han et al 421

Table I. Baseline characteristics of the study subjects

Age (y) Male sex (%) Body mass index (kg/m 2) Current smoking (%) SBP (mm Hg) DBP (mm Hg) Comorbidities (%) Hypertension Diabetes mellitus History of MI History of stroke PVD Medications (%) ACEi or ARB β-Blocker Diuretics Statin CAG within 1 wk (%) CPB (%) IABP (%) LV ejection fraction (%) Laboratory findings Creatinine (mg/dL) eGFR (mL/min/1.73 m 2) White blood cells (/μL) Hemoglobin level (g/dL) Albumin (g/dL) Cholesterol (mg/dL) Follow-up duration (m)

Total (n = 1899)

Non-AKI (n = 1266)

AKI (n = 633)


Non-CKD (n = 1682)

CKD (n = 217)


64.9 ± 9.60 71.6 24.3 ± 3.01 30.3 123.1 ± 20.32 72.1 ± 12.49

64.2 ± 9.57 70.3 24.2 ± 2.88 30.0 122.3 ± 19.33 72.2 ± 11.87

66.5 ± 9.48 74.2 24.5 ± 3.24 31.0 124.7 ± 22.08 71.8 ± 13.64

b.001 .072 .090 .672 .019 .549

64.2 ± 9.53 75.6 24.3 ± 2.96 31.5 123.1 ± 20.21 72.4 ± 12.39

70.6 ± 8.16 41.0 24.3 ± 3.37 21.2 123.0 ± 21.13 69.2 ± 12.92

b.001 b.001 .926 .002 .943 b.001

51.3 45.1 9.8 24.5 8.3

49.6 42.1 8.3 23.7 7.4

54.7 51.0 12.8 26.1 10.0

.038 b.001 .002 .258 .059

50.0 43.0 9.5 23.6 7.5

61.3 61.3 12.4 31.3 14.3

.002 b.001 .163 .013 .001

30.9 38.2 18.1 29.8 31.8 23.5 10.7 54.7 ± 12.52

29.2 39.4 16.1 31.3 30.1 18.0 7.0 55.8 ± 11.79

34.1 35.9 22.1 26.9 35.1 34.6 18.2 52.5 ± 13.61

.029 .133 .001 .047 .028 b.001 b.001 b.001

30.0 38.1 16.3 29.9 32.1 22.9 10.0 55.1 ± 12.35

37.8 39.2 32.3 29.0 29.0 28.1 16.1 51.7 ± 13.45

.019 .762 b.001 .791 .360 .092 .006 b.001

1.1 ± 0.34 84.3 ± 20.61 8119.6 ± 3341.34 12.2 ± 2.18 3.6 ± 0.69 144.6 ± 45.20 71 (51-93)

1.1 ± 0.30 84.5 ± 19.06 8125.3 ± 3239.70 12.4 ± 2.18 3.6 ± 0.70 144.7 ± 45.10 73 (55-94)

1.2 ± 0.42 83.8 ± 23.42 8108.2 ± 3538.48 12.0 ± 2.18 3.6 ± 0.66 144.4 ± 45.44 65 (43-91)

.001 .540 .916 .003 .246 .922 b.001

1.0 ± 0.20 88.9 ± 16.76 8105.1 ± 3398.42 12.4 ± 2.17 3.7 ± 0.68 145.7 ± 45.19 72 (53-94)

1.7 ± 0.57 48.4 ± 9.96 8232.4 ± 2864.95 11.0 ± 1.82 3.3 ± 0.66 135.8 ± 44.39 58 (24-78)

b.001 b.001 .598 b.001 b.001 .002 b.001

Abbreviations: SBP, systolic blood pressure; DBP, diastolic blood pressure; MI, myocardial infarction; PVD, peripheral vascular disease; ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; CAG, coronary angiography; CPB, cardiopulmonary bypass; IABP, intra-aortic balloon pump; LV, left ventricular. ⁎ Difference between AKI and non-AKI groups. † Difference between CKD and non-CKD groups.

Figure 1

Kaplan-Meier survival curves according to the presence of AKI or CKD.

patients with CKD experienced AKI, whereas 31.6% of the patients without CKD had AKI. Several parameters, such as age, hypertension, diabetes mellitus, baseline medications, perioperative use of an intra-aortic balloon

pump, kidney function, and hemoglobin level, were different according to the presence of AKI or CKD. In addition, systolic blood pressure, histories of myocardial infarction and coronary angiography, and intraoperative use of cardiopulmonary bypass were different between the AKI and non-AKI groups. The CKD group had differences from the non-CKD group with additional covariates, such as sex, smoking, diastolic blood pressure, histories of stroke and peripheral vascular disease, albumin, and cholesterol.

Effects of AKI and CKD on mortality Over the median duration of 5 years (interquartile ranges 4-7, maximum 10 years), 434 patients died, and the mortality rates were 6.9%, 12.8%, 17.3%, and 22.9% at 1, 3, 5, and 10 years, respectively. Figure 1 shows the KaplanMeier survival curves according to the presence of AKI, CKD, or both. The 4 survival curves were separated, and the survival rates of each were significantly different (P b .001 by log-rank test). When the Cox model for overall mortality was conducted using AKI and CKD (Table II), these conditions, particularly severe stages of both, increased mortality rates more than in the counterpart

American Heart Journal March 2015

422 Han et al

Table II. Hazard ratios for all-cause mortality according to AKI and CKD Multivariate⁎

Univariate Group Presence of AKI Non-AKI AKI Stages of AKI Non-AKI AKI, stage 1 AKI, stage 2 AKI, stage 3 Presence of CKD Non-CKD CKD Stages of CKD Non-CKD CKD, stage 3a CKD, stage 3b CKD, stages 4 and 5 Presence of AKI and CKD Non-AKI, non-CKD AKI, non-CKD Non-AKI, CKD AKI, CKD

n (%)

HR (95% CI)


HR (95% CI)


1266 (66.7) 633 (33.3)

1 (reference) 2.35 (1.945-2.834)


1 (reference) 1.77 (1.444-2.171)


1266 (66.7) 571 (30.1) 42 (2.2) 20 (1.1)

1 2.15 3.43 7.62

b.001 b.001 b.001

1 1.67 1.91 5.74

b.001 .010 b.001

1682 (88.6) 217 (11.4)

1 (reference) 3.17 (2.550-3.950)


1 (reference) 2.13 (1.663-2.738)


1682 (88.6) 142 (7.5) 62 (3.3) 13 (0.7)

1 2.98 3.26 4.81

(reference) (2.275-3.892) (2.278-4.663) (2.564-9.029)

b.001 b.001 b.001

1 2.24 1.79 3.16

(reference) (1.669-3.012) (1.213-2.630) (1.630-6.106)

b.001 .003 .001

1150 532 116 101

1 2.35 3.41 5.88

(reference) (1.889-2.915) (2.475-4.709) (4.362-7.922)

b.001 b.001 b.001

1 1.84 2.46 3.21

(reference) (1.464-2.319) (1.735-3.478) (2.301-4.488)

b.001 b.001 b.001

(60.6) (28.0) (6.1) (5.3)

(reference) (1.770-2.623) (2.169-5.426) (4.510-12.860)

(reference) (1.351-2.060) (1.171-3.117) (3.266-10.085)

⁎ Adjusted for age; sex; body mass index; current smoking status; systolic blood pressure; diastolic blood pressure; hypertension; diabetes mellitus; histories of myocardial infarction and stroke; peripheral vascular disease; uses of angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers, β-blockers, diuretics, and statins; coronary angiography within 1 week; cardiopulmonary bypass; intra-aortic balloon pump; left ventricular ejection fraction; and laboratory findings, such as white blood cell count, hemoglobin level, albumin, and cholesterol.

Table III. Adjusted HRs for all-cause mortality in patients with AKI or CKD according to the timeframe AKI Timeframes 0-3 m 3-12 m 1-3 y 3-5 y 5-7 y 7-10 y

HR (95% CI)⁎ 3.74 1.96 1.49 1.31 1.59 1.27

(2.245-6.226) (1.065-3.603) (0.997-2.221) (0.830-2.076) (0.956-2.636) (0.548-2.957)

CKD HR (95% CI)⁎

P b.001 .031 .052 .245 .074 .575

2.70 2.19 2.60 1.96 2.43 0.82

(1.535-4.750) (1.034-4.634) (1.626-4.156) (1.085-3.538) (1.291-4.571) (0.236-2.844)

P .001 .041 b.001 .026 .006 .753

⁎ Adjusted for age; sex; body mass index; current smoking status; systolic blood pressure; diastolic blood pressure; hypertension; diabetes mellitus; histories of myocardial infarction and stroke; peripheral vascular disease; uses of angiotensin-converting enzyme inhibitors/angiotensin II receptor blockers, β-blockers, diuretics, and statins; coronary angiography within 1 week; cardiopulmonary bypass; intra-aortic balloon pump; left ventricular ejection fraction; and laboratory findings, such as white blood cell count, hemoglobin level, albumin, and cholesterol.

groups, irrespective of adjusting all of the covariates. The risk increase was greater in patients having both AKI and CKD than in patients having either AKI or CKD. We further analyzed the Cox model according to the timeframes (Table III). When the statistical significances or HRs were reviewed, the effect of AKI on mortality was significant, primarily during the early period and particularly within 3 months, compared with the late period. For CKD, the HRs were relatively invariable throughout the study period, ranging from 2 to 3, except for the period beyond 7 years.

Receiver operating characteristic curves for the prediction of mortality We evaluated the ROC curves of AKI, CKD, or both for the prediction of mortality (Figure 2). The AUCs (95% CI) of the ROC curves in the model using AKI alone decreased over time as follows: 1 year, 0.671 (0.6200.722); 3 years, 0.630 (0.591-0.669); 5 years, 0.622 (0.5870.657); and 10 years, 0.621 (0.590-0.652). For the model using CKD alone, the AUCs (95% CI) were relatively consistent over time as follows: 1 year, 0.586 (0.5310.641); 3 years, 0.593 (0.551-0.634); 5 years, 0.590 (0.553-

American Heart Journal Volume 169, Number 3

Figure 2

Han et al 423

survival curves of recovered and nonrecovered AKI cases were separate during the early period, but the curves crossed during the late period. When the multivariate Cox model using the decision for recovery at 3 months was applied, the HRs for overall mortality were 1.68 (1.349-2.097) and 2.06 (1.519-2.789) in the recovered and nonrecovered cases, respectively, compared with the non-AKI group (all P values b .001). Although the analyses were stratified by the timeframes, the recovery group was at a greater risk for mortality than the non-AKI group until 3 years, based on the following HRs: 0 to 3 months, 2.42 (1.358-4.314); 3 to 12 months, 1.89 (0.9903.627); and 1 to 3 years, 1.58 (1.039-2.413). When recovery from AKI was defined at 1 month, the overall trend was similar to the above results.

Area under the curves of the ROC regarding the predictability of mortality.

Discussion 0.626); and 10 years, 0.587 (0.554-0.619). When AKI and CKD were both incorporated into the model, the AUCs decreased during the early period but remained relatively constant during the later period, as follows: 1 year, 0.710 (0.663-0.757); 3 years, 0.686 (0.648-0.724); 5 years, 0.677 (0.643-0.710); and 10 years, 0.670 (0.639-0.700).

Synergistic effects between AKI and CKD on mortality There was a trend toward interaction between AKI and CKD for mortality within 4 years (P for interaction b .05), but after 4 years, there was no interaction for morality (P for interaction ≥ .05). We evaluated whether there was a synergistic effect for 1-year mortality between AKI and CKD using the RERI method (Table IV). For both AKI and CKD, the RERI and AP scales were positive, and the synergy index was N1, indicating a relative excess risk of mortality when both AKI and CKD were considered together. For the mortalities from 2 to 4 years, all of the scales of the RERI method gave appropriate reasons to determine a synergistic effect between AKI and CKD, as follows (eg, RERI, AP, and synergy index): 2-year mortality, 0.30 (−3.998 to 4.604), 0.04 (−0.539 to 0.621), 1.05 (0.296-1.803); 3-year mortality, 0.54 (−3.166 to 4.248), 0.07 (−0.425 to 0.569), 1.09 (0.4381.744); and 4-year mortality, 0.96 (−2.512 to 4.437), 0.13 (−0.342 to 0.605), 1.18 (0.481-1.877). Effects of recovery after AKI Based on the findings in which the effects of AKI were predominant during the early period, we further evaluated whether this feature was different between the recovered and nonrecovered cases of AKI. Accordingly, we followed serum creatinine after the AKI event and determined the cases that recovered at 1 month or 3 months. Figure 3 shows the survival curves of recovered and nonrecovered cases. As shown in this figure, the

Substantial numbers of patients with several comorbid conditions will undergo CABG in the future, and their outcomes might not be favorable due to postoperative complications. This fact underscores the importance of risk prediction and proper prevention of postoperative complications. In this regard, AKI and CKD are important issues both for clinicians and patients. The present study was the first to address AKI and CKD together in CABG patients. Similar to the previous reports, AKI and CKD increased the mortality of the study subjects. However, their features were different from one another: AKI predicted or was associated primarily with early mortality, whereas CKD had a relatively constant predictability throughout both early and late periods. In particular, these effects of AKI and CKD were synergistic rather than merely additive or exclusive. When recovery after AKI was evaluated, the recovered cases also had a higher risk of early mortality than the non-AKI group. Although AKI is a single event or results in perfect recovery, patients with AKI are susceptible to the risk of CKD. 16 Conversely, CKD is one of the significant risk factors for the development of AKI. 17 Therefore, the 2 diseases are interrelated, but many existing studies have not considered them together. As shown in the results presented above, AKI and CKD had synergistic additive effects on mortality. This intriguing effect can be understood with the following explanation. Patients with CKD can experience repeated AKI due to the nature of CKD, irrespective of recovery from AKI. 17,18 Acute kidney injury and CKD have certain common features in pathophysiology, such as decreased glomerular filtration rate. Furthermore, CKD can hamper or delay recovery after AKI due to immune dysregulation. 19 In these respects, patients with both AKI and CKD had a much lower survival rate than expected. Studies on predictive models of risk after CABG have been reported. 3,4 However, these models only focused on the short- or mid-term mortality and did not stratify the

American Heart Journal March 2015

424 Han et al

Table IV. Synergistic effects between AKI and CKD on 1-year mortality Non-AKI

Non-CKD CKD OR for CKD within strata of AKI status


OR for AKI within strata of CKD status

N with/without outcome

OR (95% CI)

N with/without outcome

OR (95% CI)

34/1116 15/101

1 (reference) 4.88 (2.569-9.252) 4.88 (2.569-9.252)

61/471 21/80

4.25 (2.757-6.555) 8.62 (4.779-15.534) 2.03 (1.170-3.512)

4.25 (2.757-6.555) 1.77 (0.856-3.648)

Measurement of interaction on additive scale: RERI (95% CI), 0.49 (−5.224 to 6.204); AP (95% CI), 0.06 (−0.604 to 0.718); synergy index (95% CI), 1.07 (0.157-1.980). Abbreviation: OR, odds ratio.

Figure 3

Survival curves according to the recovery after AKI, which was defined at 1 month (A) and 3 months (B) after surgery.

mortality risk by timeframes. According to the present results, the effects of AKI were primarily restricted to the early period after surgery. This feature was not different according to recovery from AKI. If AKI developed repeatedly or became CKD, the effects of AKI should have been persistent. Although the present study did not collect these subsequent data, all of the AKI cases might not experience repeated AKI, particularly when insults, such as reoperation, were not encountered. Whether patients recover from AKI completely, the risk of CKD does not disappear. 18,20 However, the risk of CKD could not be conferred on all of the subjects with AKI, and the effects of later CKD might have been smaller than those of initial AKI. These differences might have contributed to the effects of AKI in the present results. The incorporation of present results into the previous models could increase the predictability of mortality risk after CABG. Based on these, we urge clinicians to monitor AKI and CKD before and after surgery. Despite complete recovery, patients with AKI showed higher mortality than those without AKI, particularly during the early follow-up period. This finding could be

explained by the following issues. Recovery after AKI does not guarantee both histologic and physiologic normality, 21 although serum creatinine itself has limitations in defining recovery. Furthermore, AKI can reflect the functioning or failure of other organs. 22 In this regard, patients who recovered from AKI had poorer conditions than those without AKI, and these baseline conditions contributed to the discrepancy in mortality, which is why some previous studies reported no independent correlation between AKI and mortality. 20,23 However, even in studies of recovery from AKI, it is necessary to consider analyses stratified according to timeframe, but previous studies have not undertaken such considerations. Although the present results are informative, this study has some limitations. First, the observational study design limits the drawing of conclusion because of causality. This limitation is inevitable, but reverse causation is less plausible. Furthermore, we could not conclude definitively that the reduction or prevention of postoperative AKI improves survival, as AKI might merely be a sign of other organ failures. Second, the treatment of postoperative AKI can differ depending on the clinician or center.

American Heart Journal Volume 169, Number 3

However, the general management of AKI did not differ much among subjects because none of the subjects were enrolled in trials using novel or comparable treatments. Third, if the sample size is larger or the disease subset is different, the duration of the effects of AKI may not be exactly the same as the present results. Lastly, some valuable parameters, such as the etiology of AKI, severity or duration of diabetes, and hourly urine data, were not available in the current data set. Irreversible AKI might have a different etiology than reversible AKI. In particular, we did not use the urine output criteria for AKI due to the lack of data on hourly urine collection, although those criteria could improve the sensitivity in defining AKI. Further studies addressing these limitations will be necessary in the future.

Conclusions Acute kidney injury and CKD are both prevalent conditions during CABG surgery. Apparently, these conditions have strong effects on mortality, but the features of these conditions are simultaneously discrete and synergistic additive when the patients are followed up for a long period. To predict the outcomes of CABG precisely, these features should be noted in clinical practice. Furthermore, recovery from AKI should not be considered a reassuring state. The present study could be helpful in guiding follow-up studies on the issues of AKI and CKD in patients undergoing CABG.

References 1. ACC/AHA guidelines and indications for coronary artery bypass graft surgery. A report of the American College of Cardiology/American Heart Association Task Force on Assessment of Diagnostic and Therapeutic Cardiovascular Procedures (Subcommittee on Coronary Artery Bypass Graft Surgery). Circulation 1991;83(3):1125-73. 2. Hannan EL, Racz MJ, Walford G, et al. Long-term outcomes of coronary-artery bypass grafting versus stent implantation. N Engl J Med 2005;352(21):2174-83. 3. Nilsson J, Algotsson L, Hoglund P, et al. Comparison of 19 preoperative risk stratification models in open-heart surgery. Eur Heart J 2006;27(7):867-74. 4. Nilsson J, Algotsson L, Hoglund P, et al. Early mortality in coronary bypass surgery: the EuroSCORE versus The Society of Thoracic Surgeons risk algorithm. Ann Thorac Surg 2004;77(4):1235-9. 5. Cooper WA, O'Brien SM, Thourani VH, et al. Impact of renal dysfunction on outcomes of coronary artery bypass surgery: results from the Society of Thoracic Surgeons National Adult Cardiac Database. Circulation 2006;113(8):1063-70. 6. Rosner MH, Okusa MD. Acute kidney injury associated with cardiac surgery. Clin J Am Soc Nephrol 2006;1(1):19-32.

Han et al 425

7. Amin AP, Salisbury AC, McCullough PA, et al. Trends in the incidence of acute kidney injury in patients hospitalized with acute myocardial infarction. Arch Intern Med 2012;172(3):246-53. 8. Holzmann MJ, Hammar N, Ahnve S, et al. Renal insufficiency and long-term mortality and incidence of myocardial infarction in patients undergoing coronary artery bypass grafting. Eur Heart J 2007;28(7): 865-71. 9. Hobson CE, Yavas S, Segal MS, et al. Acute kidney injury is associated with increased long-term mortality after cardiothoracic surgery. Circulation 2009;119(18):2444-53. 10. Holzmann M, Jernberg T, Szummer K, et al. Long-term cardiovascular outcomes in patients with chronic kidney disease undergoing coronary artery bypass graft surgery for acute coronary syndromes. J Am Heart Assoc 2014;3(2):e000707. 11. Levey AS, Bosch JP, Lewis JB, et al. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999;130(6):461-70. 12. Lee CS, Cha RH, Lim YH, et al. Ethnic coefficients for glomerular filtration rate estimation by the Modification of Diet in Renal Disease study equations in the Korean population. J Korean Med Sci 2010;25 (11):1616-25. 13. 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. 14. Kidney Disease: Improving Global Outcomes (KDIGO) Acute Kidney Injury Work Group. KDIGO Clinical Practice Guideline for Acute Kidney Injury. Kidney Int Suppl 2012;2:1-138. 15. Richardson DB, Kaufman JS. Estimation of the relative excess risk due to interaction and associated confidence bounds. Am J Epidemiol 2009;169(6):756-60. 16. Ishani A, Xue JL, Himmelfarb J, et al. Acute kidney injury increases risk of ESRD among elderly. J Am Soc Nephrol 2009;20(1):223-8. 17. James MT, Hemmelgarn BR, Wiebe N, et al. Glomerular filtration rate, proteinuria, and the incidence and consequences of acute kidney injury: a cohort study. Lancet 2010;376(9758):2096-103. 18. Heung M, Chawla LS. Predicting progression to chronic kidney disease after recovery from acute kidney injury. Curr Opin Nephrol Hypertens 2012;21(6):628-34. 19. Kato S, Chmielewski M, Honda H, et al. Aspects of immune dysfunction in end-stage renal disease. Clin J Am Soc Nephrol 2008;3(5):1526-33. 20. Jones J, Holmen J, De Graauw J, et al. Association of complete recovery from acute kidney injury with incident CKD stage 3 and allcause mortality. Am J Kidney Dis 2012;60(3):402-8. 21. Zager RA, Johnson AC, Becker K. Acute unilateral ischemic renal injury induces progressive renal inflammation, lipid accumulation, histone modification, and “end-stage” kidney disease. Am J Physiol Renal Physiol 2011;301(6):F1334-45. 22. de Mendonca A, Vincent JL, Suter PM, et al. Acute renal failure in the ICU: risk factors and outcome evaluated by the SOFA score. Intensive Care Med 2000;26(7):915-21. 23. Wald R, Quinn RR, Luo J, et al. Chronic dialysis and death among survivors of acute kidney injury requiring dialysis. JAMA 2009;302 (11):1179-85.

Effects of acute kidney injury and chronic kidney disease on long-term mortality after coronary artery bypass grafting.

Both acute kidney injury (AKI) and chronic kidney disease (CKD) are important issues in patients undergoing coronary artery bypass grafting (CABG), pa...
377KB Sizes 2 Downloads 8 Views