Abstract

Background

The presence of chronic kidney disease is an independent prognostic factor in patients with myocardial infarction (MI). We compared the Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI) equation and the Modification of Diet in Renal Disease (MDRD) study equation with regard to prognostic value in patients with MI.

Methods

This study analyzed a retrospective cohort of 11 050 consecutive patients who had MI and were enrolled in the Korea Acute Myocardial Infarction Registry from November 2005 to August 2008. We applied the CKD-EPI equation and the MDRD study equation to determine the estimated glomerular filtration rate (eGFR) in a cohort of patients with MI.

Results

The mean eGFRCKD-EPI was slightly higher than that of eGFRMDRD (73.16 versus 72.23 mL/min/1.73 m2; P < 0.001). The prevalence of eGFRCKD-EPI <60 mL/min/1.73 m2 was 26.9%, whereas that of eGFRMDRD was 28.5%. The area under the receiver operating characteristic curve was significantly larger for predicting the 1-year major adverse cardiovascular event (MACE) and 1-year all-cause mortality with CKD-EPI equation (0.648 versus 0.641, 0.768 versus 0.753, respectively; P < 0.001). The net reclassification index for improvement in risk of 1-year MACE and 1-year all-cause mortality were 4.09% (P< 0.001) and 9.25% (P< 0.001), respectively.

Conclusions

The application of the eGFRCKD-EPI demonstrated better predictive values for clinical outcomes than eGFRMDRD in a cohort of patients with MI.

Introduction

Serum creatinine is used to determine renal function and estimate its severity. However, the accuracy of the serum creatinine level as a marker of renal function is controversial, because it is easily influenced by various factors such as age, gender, race and lean body mass [1, 2]. Therefore, estimating the glomerular filtration rate (GFR) or creatinine clearance is a preferable method for accessing renal function [3].

The Modification of Diet in Renal Disease (MDRD) study equation is the most commonly used equation for estimating GFR in clinical practice and is the basis for many kidney disease guidelines [4, 5]. Despite its widespread use, the MDRD study equation has some considerable limitations. The MDRD study equation was developed using a population of 1628 people with chronic kidney disease (CKD), with a mean measured GFR of 40 mL/min/1.73 m2. The equation systematically underestimates GFR in individuals without known CKD. This leads to misclassification into a lower GFR category and results in overdiagnosis of CKD [6, 7]. The CKD-Epidemiology Collaboration (CKD-EPI) recently published an equation to improve GFR estimation, particularly in individuals with GFR ≥60 mL/min/1.73 m2 [8]. This equation was developed using a population of 8254 individuals from 10 studies with a mean measured GFR of 68 mL/min/1.73 m2 and was validated in 16 additional studies including 3896 individuals. The CKD-EPI equation was more accurate than the MDRD study equation, especially at higher values of GFR [9] and was demonstrated to have stronger prognostic value in the general population [10, 11].

Even mild-to-moderate CKD is independently associated with a high risk for mortality and morbidity in patients with myocardial infarction (MI) [12, 13]. Apart from the prognostic impact of CKD, the patients with CKD were less likely to receive evidence-based medical therapies or reperfusion therapies as those with or without CKD [14, 15]. Therefore, assessment of renal function is an important issue in patients with MI in the aspect of risk prediction, drug prescription and planning diagnostic or therapeutic procedures. However, until recently, limited studies have compared various equations with regard to their prognostic value in patients with MI [16–18].

In the present study, the MDRD study and CKD-EPI equations were applied in a cohort of patients with MI, and clinical outcomes were compared with the aim of establishing the prognostic value of the two equations. The hypothesis of this study was that the estimated GFR using the CKD-EPI equation would more accurately predict clinical outcomes in a population of patients with MI.

Materials and methods

Study population

The study population was enrolled in a nationwide prospective Korea Acute Myocardial Infarction Registry (KAMIR) from November 2005 to August 2008. KAMIR is a Korean prospective, open, observational, multi-center on-line registry of patients with acute MI started in November 2005 with support from the Korean Society of Cardiology. Details of the KAMIR have been published [19]. This study was carried out retrospectively in a cohort of 11,050 consecutive patients who were available for the estimation of GFR (eGFR) and who had completed at least 1 year of follow-up. The diagnosis of MI was based on the triad of chest pain, ECG changes and raised serum cardiac enzyme level [20].

Data collection

The baseline variables included age, sex, body mass index (BMI), coronary risk factors which included hypertension, diabetes mellitus (DM), hyperlipidemia, history of smoking, previous history of ischemic heart disease (IHD), total cholesterol, low-density lipoprotein (LDL), high-sensitivity C-reactive protein (hs-CRP), N-terminal prohormone of brain natriuretic peptide (NT-pro BNP) and Killip class.

Estimation of GFR

The serum creatinine was measured using the Jaffe method and standardized to isotope dilution mass spectrometry values according to each center's protocol. We calculated eGFR using the MDRD study equation (eGFRMDRD) [21] and also calculated eGFR using the CKD-EPI equation (eGFRCKD-EPI) [8]. The eGFR was categorized as ≥120, 90–119, 60–89, 30–59 and <30 mL/min/1.73 m2, which are based on NKF Kidney Disease Outcomes Quality Initiative (KDOQI) CKD stages, but modified for the following reasons: (i) individuals with eGFR ≥120 mL/min/1.73 m2 were separated from those with eGFR of 90–119 mL/min/1.73 m2 under assumption that high eGFR may result from a low creatinine level and many of them were associated with muscle loss or malnutrition, which might contribute to worse outcomes, (ii) individuals corresponding to CKD stages 4 and 5 were combined because there were relatively too small numbers of people in this category.

Statistical analysis

Continuous variables with normal distributions are expressed as mean ± SD, and categorized eGFR groups were compared using one-way analysis of variance. Continuous data with a skewed distribution are presented as median (with 25th and 75th percentiles) and were compared using the Kruskal–Wallis test. Categorical variables were compared using the χ2 test or the Fisher's exact test if the expected value of the variable was <5 in at least one group. The paired Wilcoxon rank-sum test was used to compare the eGFRCKD-EPI and eGFRMDRD within age categories.

To compare the incremental prognostic value of using CKD-EPI equation over MDRD study equation, two statistical concepts were used: area under the receiver operating characteristic (ROC) curves (AUC) and net reclassification improvement. In the ROC curve analysis, the non-parametric approach was used to compare AUC, as described by DeLong et al. [22]. The net reclassification index was calculated as the sum of the proportion of participants reclassified downward to a lower eGFR category for people who experienced a major adverse cardiovascular event and the proportion of participants reclassified upward to a higher eGFR category for people who did not experience an event minus the sum of the proportion of participants reclassified upward for people who experienced an event and the proportion of participants reclassified downward for people who did not experience an event [23].

A P-value < 0.05 was considered significant. The statistical analyses were performed using the Statistical Package for Social Sciences software, version 18.0 (IBM, Armonk, NY) except to compare AUC, which used SAS software version 9.12 (SAS Institute Inc., Cary, NC).

Results

A total of 11 050 patients [age (mean ± SD), 63 ± 13 years; men, 70.5%] were included in the present study. The baseline characteristics and biochemical parameters of patients divided according to the CKD-EPI equation category are listed in Table 1. A lower GFR was associated with older age, female gender, lower BMI, higher prevalence of hypertension, DM, history of IHD and a lower prevalence of current smoking. At the time of arrival at the hospital, a lower GFR associated with lower blood pressure, higher heart rate, lower left ventricular ejection fraction (LVEF) and higher Killip class was observed. In patients with lower GFR, the total cholesterol, LDL, troponin-I and creatine kinase-MB (CK-MB) levels were lower, whereas the glucose, hs-CRP and NT-pro BNP levels were higher in patients with higher GFR.

Table 1.

Baseline characteristics

 ≥120 (n = 79) 90–119 (n = 2908) 60–89 (n = 5098) 30–59 (n = 2318) <30 (n = 647) P-value 
Age, years 44 ± 14 55 ± 10 64 ± 12 72 ± 11 70 ± 12 <0.001 
Male 69 (87.3) 2327 (80.0) 3723 (73.0) 1338 (57.7) 334 (51.6) <0.001 
BMI, kg/m2 25.1 ± 3.7 24.3 ± 3.2 24.0 ± 3.2 23.6 ± 3.3 23.1 ± 3.2 <0.001 
Past History 
 Hypertension 24 (30.4) 1107 (38.1) 2311 (45.4) 1472 (63.7) 479 (74.3) <0.001 
 DM 24 (30.4) 640 (22.1) 1223 (24.0) 860 (37.3) 354 (55.1) <0.001 
 Previous IHD 5 (6.3) 310 (10.7) 747 (14.7) 482 (20.9) 185 (28.7) <0.001 
 Hyperlipidemia 5 (6.3) 316 (10.9) 448 (8.8) 209 (9.1) 74 (11.5) 0.449 
 Smoking 50 (63.3) 1662 (57.4) 2271 (44.8) 630 (27.4) 138 (21.7) <0.001 
At admission 
 SBP (mmHg) 131 ± 26 132 ± 26 129 ± 28 123 ± 33 126 ± 38 <0.001 
 DBP (mmHg) 80 ± 17 81 ± 16 79 ± 17 74 ± 19 75 ± 21 <0.001 
 Heart rate 83 ± 17 77 ± 16 77 ± 20 80 ± 25 82 ± 26 <0.001 
 Killip class 1.25 ± 0.72 1.23 ± 0.61 1.38 ± 0.78 1.76 ± 1.04 2.08 ± 1.11 <0.001 
 LVEF (%) 53.2 ± 12.7 53.5 ± 11.1 51.9 ± 12.0 48.3 ± 13.6 46.1 ± 13.6 <0.001 
Biochemical parameters 
 Creatinine (mg/dL) 0.5 ± 0.2 0.8 ± 0.1 1.0 ± 0.2 1.4 ± 0.3 5.0 ± 6.4 <0.001 
 Troponin-I (ng/mL) 22 (6.71) 22 (4.53) 20 (4.50) 18 (3.50) 15 (4.45) 0.001 
 CK-MB (U/L) 82 (12 224) 79 (21 204) 79 (19 210) 61 (15 181) 32 (11 118) <0.001 
 TC (mg/dL) 173 ± 45 188 ± 44 183 ± 43 176 ± 47 170 ± 53 <0.001 
 LDL (mg/dL) 108 ± 36 122 ± 42 117 ± 41 112 ± 44 105 ± 56 <0.001 
 Glucose (mg/dL) 173 ± 93 156 ± 64 163 ± 70 198 ± 100 219 ± 131 <0.001 
 hs-CRP (mg/dL) 0.6 (0.2, 4.2) 0.6 (0.2, 3.0) 0.8 (0.2, 3.9) 1.5 (0.3, 7.3) 2.9 (0.7, 11.1) <0.001 
 NT-proBNP (pg/mL) 356 (90 752) 246 (66 844) 426 (103 1464) 1673 (347 5329) 10160 (2776 31436) <0.001 
 eGFRCKD-EPI (mL/min/1.73 m2129 ± 13 100 ± 7 76 ± 9 48 ± 8 17 ± 9 <0.001 
 eGFRMDRD (mL/min/1.73 m2186 ± 82 100 ± 16 73 ± 9 48 ± 8 18 ± 9 <0.001 
 ≥120 (n = 79) 90–119 (n = 2908) 60–89 (n = 5098) 30–59 (n = 2318) <30 (n = 647) P-value 
Age, years 44 ± 14 55 ± 10 64 ± 12 72 ± 11 70 ± 12 <0.001 
Male 69 (87.3) 2327 (80.0) 3723 (73.0) 1338 (57.7) 334 (51.6) <0.001 
BMI, kg/m2 25.1 ± 3.7 24.3 ± 3.2 24.0 ± 3.2 23.6 ± 3.3 23.1 ± 3.2 <0.001 
Past History 
 Hypertension 24 (30.4) 1107 (38.1) 2311 (45.4) 1472 (63.7) 479 (74.3) <0.001 
 DM 24 (30.4) 640 (22.1) 1223 (24.0) 860 (37.3) 354 (55.1) <0.001 
 Previous IHD 5 (6.3) 310 (10.7) 747 (14.7) 482 (20.9) 185 (28.7) <0.001 
 Hyperlipidemia 5 (6.3) 316 (10.9) 448 (8.8) 209 (9.1) 74 (11.5) 0.449 
 Smoking 50 (63.3) 1662 (57.4) 2271 (44.8) 630 (27.4) 138 (21.7) <0.001 
At admission 
 SBP (mmHg) 131 ± 26 132 ± 26 129 ± 28 123 ± 33 126 ± 38 <0.001 
 DBP (mmHg) 80 ± 17 81 ± 16 79 ± 17 74 ± 19 75 ± 21 <0.001 
 Heart rate 83 ± 17 77 ± 16 77 ± 20 80 ± 25 82 ± 26 <0.001 
 Killip class 1.25 ± 0.72 1.23 ± 0.61 1.38 ± 0.78 1.76 ± 1.04 2.08 ± 1.11 <0.001 
 LVEF (%) 53.2 ± 12.7 53.5 ± 11.1 51.9 ± 12.0 48.3 ± 13.6 46.1 ± 13.6 <0.001 
Biochemical parameters 
 Creatinine (mg/dL) 0.5 ± 0.2 0.8 ± 0.1 1.0 ± 0.2 1.4 ± 0.3 5.0 ± 6.4 <0.001 
 Troponin-I (ng/mL) 22 (6.71) 22 (4.53) 20 (4.50) 18 (3.50) 15 (4.45) 0.001 
 CK-MB (U/L) 82 (12 224) 79 (21 204) 79 (19 210) 61 (15 181) 32 (11 118) <0.001 
 TC (mg/dL) 173 ± 45 188 ± 44 183 ± 43 176 ± 47 170 ± 53 <0.001 
 LDL (mg/dL) 108 ± 36 122 ± 42 117 ± 41 112 ± 44 105 ± 56 <0.001 
 Glucose (mg/dL) 173 ± 93 156 ± 64 163 ± 70 198 ± 100 219 ± 131 <0.001 
 hs-CRP (mg/dL) 0.6 (0.2, 4.2) 0.6 (0.2, 3.0) 0.8 (0.2, 3.9) 1.5 (0.3, 7.3) 2.9 (0.7, 11.1) <0.001 
 NT-proBNP (pg/mL) 356 (90 752) 246 (66 844) 426 (103 1464) 1673 (347 5329) 10160 (2776 31436) <0.001 
 eGFRCKD-EPI (mL/min/1.73 m2129 ± 13 100 ± 7 76 ± 9 48 ± 8 17 ± 9 <0.001 
 eGFRMDRD (mL/min/1.73 m2186 ± 82 100 ± 16 73 ± 9 48 ± 8 18 ± 9 <0.001 

Continuous data are expressed as the mean ± SD or the median and interquartile range (25th and 75th percentiles). Categorical data were expressed as percentages.

BMI, body mass index; DM, diabetes mellitus; IHD, ischemic heart disease; SBP, systolic blood pressure; DBP, diastolic blood pressure; LVEF, left ventricular ejection fraction; CK-MB, creatine kinase-MB; TC, total cholesterol; LDL, low-density lipoprotein; hs-CRP, high-sensitivity C-reactive protein; NT-proBNP, N-terminal pro B-type natriuretic peptide; eGFRCKD-EPI, estimated GFR by using CKD-EPI equation; eGFRMDRD, estimated GFR by using MDRD study equation.

There was a significant correlation between the eGFRCKD-EPI and eGFRMDRD (r = 0.922, P< 0.001). The mean eGFRCKD-EPI was slightly higher than eGFRMDRD (73.16 ± 24.36 mL/min/1.73 m2 versus 72.23 ± 27.70 mL/min/1.73 m2, P< 0.001). The CKD-EPI equation classified a smaller proportion of patients as having eGFR <60 mL/min/1.73 m2 compared with the MDRD study equation (eGFRCKD-EPI, n = 2965, 26.9%; eGFRMDRD, n = 3147, 28.5%).

The major adverse cardiovascular event (MACE) occurred in 1624 (14.7%) patients within 1 year and 1056 patients (9.6%) died within 1 year. Although the rate of 1-year MACE and 1-year all-cause mortality was higher in groups with lower eGFR determined by both the CKD-EPI equation and the MDRD study equation, the decrease in eGFRCKD-EPI was more proportional to the increase in the rate of 1-year MACE and 1-year all-cause mortality compared with the eGFRMDRD (Table 2). Patients who were classified as having an eGFRCKD-EPI <90 mL/min/1.73 m2 showed a higher rate of 1-year MACE and 1-year all-cause mortality than those classified with an eGFRMDRD <90 mL/min/1.73 m2.

Table 2.

One-year clinical outcomes according to eGFRCKD-EPI and eGFRMDRD categories

 ≥120 (n = 79) 90–119 (n = 2908) 60–89 (n = 5098) 30–59 (n = 2318) <30 (n = 647) P-value 
eGFRCKD-EPI 
 1-year MACE, % (n8.9% (7) 9.4% (273) 11.3% (578) 23.1% (536) 35.5% (230) <0.001 
 1-year all-cause mortality, % (n3.8% (3) 2.7% (79) 5.5% (280) 20.0% (463) 35.7% (231) <0.001 
eGFRMDRD 
 1-year MACE, % (n10.3% (43) 10.1% (184) 10.8% (610) 22.6% (569) 34.8% (218) <0.001 
 1-year all-cause mortality, % (n5.5% (23) 3.6% (65) 4.8% (269) 19.1% (481) 34.8% (218) <0.001 
 ≥120 (n = 79) 90–119 (n = 2908) 60–89 (n = 5098) 30–59 (n = 2318) <30 (n = 647) P-value 
eGFRCKD-EPI 
 1-year MACE, % (n8.9% (7) 9.4% (273) 11.3% (578) 23.1% (536) 35.5% (230) <0.001 
 1-year all-cause mortality, % (n3.8% (3) 2.7% (79) 5.5% (280) 20.0% (463) 35.7% (231) <0.001 
eGFRMDRD 
 1-year MACE, % (n10.3% (43) 10.1% (184) 10.8% (610) 22.6% (569) 34.8% (218) <0.001 
 1-year all-cause mortality, % (n5.5% (23) 3.6% (65) 4.8% (269) 19.1% (481) 34.8% (218) <0.001 

MACE, major adverse cardiovascular event; eGFRCKD-EPI, estimated GFR by using CKD-EPI equation; eGFRMDRD, estimated GFR by using MDRD study equation.

ROC analysis results comparing eGFRCKD-EPI and eGFRMDRD are listed in Table 3. The 1-year MACE was better predicted by the CKD-EPI equation than the MDRD study equation [AUC, 0.648, 95% confidence interval (CI) 0.632–0.663 versus AUC, 0.641, 95% CI 0.626–0.657]. The 1-year all-cause mortality was also better predicted by the CKD-EPI equation than the MDRD study equation (AUC 0.768, 95% CI 0.752–0.784; AUC 0.753, 95% CI 0.736–0.769, respectively) (Table 3).

Table 3.

Comparison of the area under age receiver operator characteristic curves for the eGFRCKD-EPI and eGFRMDRD

 eGFRCKD-EPI eGFRMDRD P-value 
1-year MACE 0.648 (0.632–0.663) 0.641 (0.626–0.657) <0.001 
1-year all-cause mortality 0.768 (0.7528–0.784) 0.753 (0.736–0.769) <0.001 
 eGFRCKD-EPI eGFRMDRD P-value 
1-year MACE 0.648 (0.632–0.663) 0.641 (0.626–0.657) <0.001 
1-year all-cause mortality 0.768 (0.7528–0.784) 0.753 (0.736–0.769) <0.001 

MACE, major adverse cardiovascular event; eGFRCKD-EPI, estimated GFR by using CKD-EPI equation; eGFRMDRD, estimated GFR by using MDRD study equation.

When the study population was reclassified by eGFRCKD-EPI, 1208 (10.9%) patients in this cohort were reclassified into higher GFR categories and 639 (5.6%) reclassified into lower GFR categories. The patients who were reclassified into a higher eGFR on the basis of the CKD-EPI equation had a lower rate of 1-year MACE and 1-year all-cause mortality than those who were reclassified into lower eGFR categories (1-year MACE, 8.9 versus 12.7%, P= 0.007; 1-year all-cause mortality, 2.4 versus 9.9%, P< 0.001). The net reclassification index for improvement in risk of 1-year MACE was calculated (Table 4). Of 1624 patients who had a MACE within 1 year, 80 patients were correctly reclassified into a lower eGFR category and 107 incorrectly reclassified into a higher eGFR category. In contrast, of 9426 patients who did not have a MACE within 1 year, 1101 patients were correctly reclassified into a higher eGFR category and 558 incorrectly reclassified into a lower eGFR category. In total, the net reclassification index for 1-year MACE was 4.09% (P< 0.001). Similarly, the net classification index for improvement in risk of 1-year all-cause mortality was calculated (Table 5). Of 1056 patients who died within 1 year, 63 patients were correctly reclassified into a lower eGFR category and 29 incorrectly reclassified into a higher eGFR category. In contrast, of 9994 patients who did not die of MACE within 1 year, 1056 patients were correctly reclassified into a higher eGFR category and 576 incorrectly reclassified into a lower eGFR category. 2The net reclassification index for the 1-year all-cause mortality was 9.25% (P< 0.001). The net reclassification index was calculated again using different categories, eGFR ≥60, 30–59 and <30 mL/min/1.73 m2, which were the more commonly used settings in clinical practice. The net reclassification index of 1-year MACE and all-cause mortality was 1.03% (P = 0.02) and 2.42% (P < 0.001), respectively.

Table 4.

Net reclassification index for 1-year MACE

eGFRMDRD (mL/min/1.73 m2eGFRCKD-EPI (mL/min/1.73 m2)
 
Total no. 
≥120 90–119 60–89 30–59 <30 
People who experience a MACE (n = 1624) 
 ≥120 36    43 
 90–119  157 27   184 
 60–89  80 524  610 
 30–59   27 530 12 569 
 <30     218 218 
 Total no. 273 578 536 230  
People who did not experience a MACE (n = 9426) 
 ≥120 64 310    374 
 90–119 1426 207   1641 
 60–89  899 4123 29  5051 
 30–59   190 1749 12 1951 
 <30    405 409 
 Total no. 72 2635 4520 1782 417  
eGFRMDRD (mL/min/1.73 m2eGFRCKD-EPI (mL/min/1.73 m2)
 
Total no. 
≥120 90–119 60–89 30–59 <30 
People who experience a MACE (n = 1624) 
 ≥120 36    43 
 90–119  157 27   184 
 60–89  80 524  610 
 30–59   27 530 12 569 
 <30     218 218 
 Total no. 273 578 536 230  
People who did not experience a MACE (n = 9426) 
 ≥120 64 310    374 
 90–119 1426 207   1641 
 60–89  899 4123 29  5051 
 30–59   190 1749 12 1951 
 <30    405 409 
 Total no. 72 2635 4520 1782 417  

Net reclassification index for 1-year MACE; clinically correct reclassification [proportion of participants reclassified upward who did not experience a MACE: (1101/9426) + proportion of participants classified downward who experienced a MACE (80/1624)]—clinically incorrect reclassification [proportion of participants reclassified upward who experienced a MACE: (107/1624) + proportion of participants classified downward who did not experience a MACE (558/9426)].

MACE, major adverse cardiovascular event; eGFRCKD-EPI, estimated GFR by using CKD-EPI equation; eGFRMDRD, estimated GFR by using MDRD study equation.

Table 5.

Net reclassification index for 1-year all-cause mortality

eGFRMDRD (mL/min/1.73 m2eGFRCKD-EPI (mL/min/1.73 m2)
 
Total no. 
≥120 90–119 60–89 30–59 <30 
People who died (n = 1056)       
 ≥120 20    23 
 90–119  43 22   65 
 60–89  16 245  269 
 30–59   13 455 13 481 
 <30     218 218 
 Total no. 79 280 463 231  
People who did not die (n = 9994) 
 ≥120 68 326    394 
 90–119 1540 212   1760 
 60–89  963 4402 27  5392 
 30–59   204 1824 11 2039 
 <30    405 409 
 Total no. 76 2829 4818 1855 416  
eGFRMDRD (mL/min/1.73 m2eGFRCKD-EPI (mL/min/1.73 m2)
 
Total no. 
≥120 90–119 60–89 30–59 <30 
People who died (n = 1056)       
 ≥120 20    23 
 90–119  43 22   65 
 60–89  16 245  269 
 30–59   13 455 13 481 
 <30     218 218 
 Total no. 79 280 463 231  
People who did not die (n = 9994) 
 ≥120 68 326    394 
 90–119 1540 212   1760 
 60–89  963 4402 27  5392 
 30–59   204 1824 11 2039 
 <30    405 409 
 Total no. 76 2829 4818 1855 416  

Net reclassification index for 1-year all-cause mortality; clinically correct reclassification [proportion of participants reclassified upward who did not die: (1179/9994) + proportion of participants classified downward who died (63/1056)]—clinically incorrect reclassification [proportion of participants reclassified upward who died: (29/1056) + proportion of participants classified downward who did not die (576/9994)].

eGFRCKD-EPI, estimated GFR by using CKD-EPI equation; eGFRMDRD, estimated GFR by using MDRD study equation.

Discussion

The presence of CKD is a more significant, independent prognostic factor in patients with MI than in patients with preserved renal function [12]. We compared the ability of the MDRD study equation and the CKD-EPI equation, as an estimate of renal function, to predict clinical outcomes.

It has been well recognized that the MDRD study equation underestimates eGFR for subjects with a measured GFR approximately or more than 60 mL/min/1.73 m2 [6, 7]. Recently, the CKD-EPI equation and the MDRD study equation have been compared in the general population. In the US National Health and Nutrition Examination Surveys (NHANES), the prevalence of CKD was 11.1% by eGFRCKD-EPI and 13.2% by eGFRMDRD [3]. The Kidney Early Evaluation Program (KEEP) study also showed that the prevalence of eGFR <60 mL/min/1.73 m2 was 14.3% compared with 16.8% by eGFRCKD-EPI and eGFRMDRD, respectively.

In the present study, compared with the MDRD study equation, use of the CKD-EPI equation resulted in a lower prevalence of eGFR <60 mL/min/1.73 m2, which was consistent with the results of previous studies. In addition, we observed a higher prevalence of eGFR <60 mL/min/1.73 m2 by both equations than in previous studies, which might reflect comorbidities of high-risk study populations. The underestimation of the GFR by the MDRD study equation is an important issue when managing patients, because underestimation of the GFR might lead to insufficient drug dosing or to withholding or delay of an essential diagnostic work-up. A more accurate estimate of GFR might aid in providing the appropriate medical care.

The rates of adverse outcomes were higher with a lower eGFR by both the CKD-EPI equation and the MDRD study equation. To compare the ability of the two equations to predict clinical outcomes in patients with MI, we used AUC analysis and net reclassification improvement. Analysis of the AUC is the most popular and classic method, in which data are analyzed as continuous variables [24]. The net reclassification index, a recently proposed statistical method, focuses on reclassification tables constructed separately for participants with and without events and quantifies the correct movement between categories (i.e. reclassifications) [23]. In contrast to AUC analysis, the net reclassification index requires categorization, which is expressed as discontinuous variables. In the present study, comparison of AUCs demonstrated that the CKD-EPI equation had better predictive ability for clinical outcomes than the MDRD study equation. Although the AUC of both eGFRs is relatively not powerful and little difference exists between the AUC of eGFRCKD-EPI and eGFRMDRD (1-year MACE, 0.648 and 0.641; 1-year all-cause mortality, 0.768 and 0.753, respectively), a very small increase in magnitude is often important in AUC analysis. Pepe et al. [25] and Ware [26]  showed simple examples in which enormous odds ratios are required to meaningfully increase the AUC. Of note, the results of AUC were similar to the previous report, which compared Cockcroft–Gault equation and the MDRD study equation in MI patients [16].

Analysis from the KEEP study showed that patients reclassified into higher eGFR categories had a lower incidence of mortality than those reclassified into lower categories or those not reclassified. The Australian Diabetes, Obesity and Lifestyle (AusDiab) study also showed similar results in that people reclassified into higher eGFR categories had a lower risk of development of cardiovascular disease. Both studies applied the net reclassification index as the statistical method for identifying the predictive ability of the CKD-EPI equation over the MDRD study equation and showed significant net reclassification improvement. Our data are consistent with the results of these studies and extend the findings to special populations, e.g. patients with MI. Accordingly, we demonstrated that eGFRCKD-EPI showed a better predictive ability for clinical outcomes than eGFRMDRD by two statistical methods; AUC and the net reclassification index.

The prognostic availability of the CKD-EPI equation has never been investigated in Asian populations. Because creatinine-based equations mainly based on American white populations, it can lead to inaccuracy of eGFR in Asian populations. The ethnic coefficient for the MDRD study equation proposed by Japanese and Chinese are markedly different (0.808 and 1.233, respectively) [27, 28]. Unlike the MDRD study equation, the CKD-EPI equation was based on data containing both Asians and whites, and Teo et al. [29] reported that the CKD-EPI equation does not require ethnic adjustment for estimating GFR in Asian populations. We did not consider ethnic coefficient for each equation, due to confusing results of previous studies in the MDRD study equation and a small difference in the CKD-EPI equation [27–30]. In the present study, the eGFRCKD-EPI showed a better predictive ability for clinical outcomes than eGFRMDRD in Asian populations.

Despite the advantage of the CKD-EPI equation over the MDRD study equation, the CKD-EPI equation cannot overcome the inherent limitation of being a creatinine-based estimation. Additional markers, such as cystatin C, might be necessary to estimate the GFR more accurately [31]. However, cystatin C has not been widely used at present, and eGFR based on serum creatinine will continue to be used in most clinical practices for the time being.

The limitations of the present study include assessment of kidney function on the basis of a single serum creatinine value obtained at the time of admission to the hospital. This value could have been affected by hemodynamic or metabolic status.

In conclusion, the CKD-EPI equation resulted in reclassification into higher eGFR categories and a decreased prevalence of CKD than the MDRD equation did. The application of the eGFRCKD-EPI demonstrated better predictive values for clinical outcomes than eGFRMDRD in a cohort of patients with MI. More accurate estimation of GFR will lead to improved risk assessment and provision of appropriate treatment.

Acknowledgements

This research was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2011-0009743), and by the Korea Science and Engineering Foundation through the Medical Research Center for Gene Regulation grant (2011-0030732) at Chonnam National University.

Conflict of interest statement. None declared.

References

1
Shlipak
MG
Pharmacotherapy for heart failure in patients with renal insufficiency
Ann Intern Med
 , 
2003
, vol. 
138
 (pg. 
917
-
924
)
2
Walsh
CR
O'Donnell
CJ
Camargo
CA
Jr
, et al.  . 
Elevated serum creatinine is associated with 1-year mortality after acute myocardial infarction
Am Heart J
 , 
2002
, vol. 
144
 (pg. 
1003
-
1011
)
3
Levey
AS
Coresh
J
Balk
E
, et al.  . 
National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification
Ann Intern Med
 , 
2003
, vol. 
139
 (pg. 
137
-
147
)
4
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
, vol. 
130
 (pg. 
461
-
470
)
5
K/DOQI clinical practice guidelines for chronic kidney disease: evaluation, classification, and stratification
Am J Kidney Dis
 , 
2002
, vol. 
39
 (pg. 
S1
-
266
)
6
Rule
AD
Larson
TS
Bergstralh
EJ
, et al.  . 
Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in chronic kidney disease
Ann Intern Med
 , 
2004
, vol. 
141
 (pg. 
929
-
937
)
7
Rule
AD
Gussak
HM
Pond
GR
, et al.  . 
Measured and estimated GFR in healthy potential kidney donors
Am J Kidney Dis
 , 
2004
, vol. 
43
 (pg. 
112
-
119
)
8
Levey
AS
Stevens
LA
Schmid
CH
, et al.  . 
A new equation to estimate glomerular filtration rate
Ann Intern Med
 , 
2009
, vol. 
150
 (pg. 
604
-
612
)
9
Stevens
LA
Schmid
CH
Greene
T
, et al.  . 
Comparative performance of the CKD Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease (MDRD) Study equations for estimating GFR levels above 60 mL/min/1.73 m2
Am J Kidney Dis
 , 
2010
, vol. 
56
 (pg. 
486
-
495
)
10
Matsushita
K
Selvin
E
Bash
LD
, et al.  . 
Risk implications of the new CKD Epidemiology Collaboration (CKD-EPI) equation compared with the MDRD Study equation for estimated FR: the Atherosclerosis Risk in Communities (ARIC) Study
Am J Kidney Dis
 , 
2010
, vol. 
55
 (pg. 
648
-
659
)
11
White
SL
Polkinghorne
KR
Atkins
RC
, et al.  . 
Comparison of the prevalence and mortality risk of CKD in Australia using the CKD Epidemiology Collaboration (CKD-EPI) and Modification of Diet in Renal Disease (MDRD) Study GFR estimating equations: the AusDiab (Australian Diabetes, Obesity and Lifestyle) Study
Am J Kidney Dis
 , 
2010
, vol. 
55
 (pg. 
660
-
670
)
12
Anavekar
NS
McMurray
JJ
Velazquez
EJ
, et al.  . 
Relation between renal dysfunction and cardiovascular outcomes after myocardial infarction
N Engl J Med
 , 
2004
, vol. 
351
 (pg. 
1285
-
1295
)
13
Bae
EH
Lim
SY
Cho
KH
, et al.  . 
GFR and cardiovascular outcomes after acute myocardial infarction: results from the Korea Acute Myocardial Infarction Registry
Am J Kidney Dis
 , 
2012
, vol. 
59
 (pg. 
795
-
802
)
14
Beattie
JN
Soman
SS
Sandberg
KR
, et al.  . 
Determinants of mortality after myocardial infarction in patients with advanced renal dysfunction
Am J Kidney Dis
 , 
2001
, vol. 
37
 (pg. 
1191
-
1200
)
15
Coca
SG
Krumholz
HM
Garg
AX
, et al.  . 
Underrepresentation of renal disease in randomized controlled trials of cardiovascular disease
J Am Med Assoc
 , 
2006
, vol. 
296
 (pg. 
1377
-
1384
)
16
Kontos
MC
Jamal
SM
Ornato
JP
, et al.  . 
Comparison of the modification of diet in renal disease and the Cockcroft-Gault equations for predicting mortality in patients admitted for exclusion of myocardial ischemia
Am J Cardiol
 , 
2008
, vol. 
102
 (pg. 
140
-
145
)
17
Szummer
K
Lundman
P
Jacobson
SH
, et al.  . 
Cockcroft-Gault is better than the Modification of Diet in Renal Disease study formula at predicting outcome after a myocardial infarction: data from the Swedish Web-system for Enhancement and Development of Evidence-based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART)
Am Heart J
 , 
2010
, vol. 
159
 (pg. 
979
-
986
)
18
Skali
H
Uno
H
Levey
AS
, et al.  . 
Prognostic assessment of estimated glomerular filtration rate by the new Chronic Kidney Disease Epidemiology Collaboration equation in comparison with the Modification of Diet in Renal Disease Study equation
Am Heart J
 , 
2011
, vol. 
162
 (pg. 
548
-
554
)
19
Chen
KY
Rha
SW
Li
YJ
, et al.  . 
Triple versus dual antiplatelet therapy in patients with acute ST-segment elevation myocardial infarction undergoing primary percutaneous coronary intervention
Circulation
 , 
2009
, vol. 
119
 (pg. 
3207
-
3214
)
20
Alpert
JS
Thygesen
K
Antman
E
, et al.  . 
Myocardial infarction redefined–a consensus document of The Joint European Society of Cardiology/American College of Cardiology Committee for the redefinition of myocardial infarction
J Am Coll Cardiol
 , 
2000
, vol. 
36
 (pg. 
959
-
969
)
21
Levey
AS
Coresh
J
Greene
T
, et al.  . 
Using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate
Ann Intern Med
 , 
2006
, vol. 
145
 (pg. 
247
-
254
)
22
DeLong
ER
DeLong
DM
Clarke-Pearson
DL
Comparing the areas under two or more correlated receiver operating characteristic curves: a nonparametric approach
Biometrics
 , 
1988
, vol. 
44
 (pg. 
837
-
845
)
23
Pencina
MJ
D'Agostino
RB
Sr
D'Agostino
RB
Jr
, et al.  . 
Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond
Stat Med
 , 
2008
, vol. 
27
 (pg. 
157
-
172
)
24
Hanley
JA
McNeil
BJ
The meaning and use of the area under a receiver operating characteristic (ROC) curve
Radiology
 , 
1982
, vol. 
143
 (pg. 
29
-
36
)
25
Pepe
MS
Janes
H
Longton
G
, et al.  . 
Limitations of the odds ratio in gauging the performance of a diagnostic, prognostic, or screening marker
Am J Epidemiol
 , 
2004
, vol. 
159
 (pg. 
882
-
890
)
26
Ware
JH
The limitations of risk factors as prognostic tools
N Engl J Med
 , 
2006
, vol. 
355
 (pg. 
2615
-
2617
)
27
Matsuo
S
Imai
E
Horio
M
, et al.  . 
Revised equations for estimated GFR from serum creatinine in Japan
Am J Kidney Dis
 , 
2009
, vol. 
53
 (pg. 
982
-
992
)
28
Ma
YC
Zuo
L
Chen
JH
, et al.  . 
Modified glomerular filtration rate estimating equation for Chinese patients with chronic kidney disease
J Am Soc Nephrol
 , 
2006
, vol. 
17
 (pg. 
2937
-
2944
)
29
Teo
BW
Xu
H
Wang
D
, et al.  . 
GFR estimating equations in a multiethnic Asian population
Am J Kidney Dis
 , 
2011
, vol. 
58
 (pg. 
56
-
63
)
30
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
, vol. 
25
 (pg. 
1616
-
1625
)
31
Shlipak
MG
Praught
ML
Sarnak
MJ
Update on cystatin C: new insights into the importance of mild kidney dysfunction
Curr Opin Nephrol Hypertens
 , 
2006
, vol. 
15
 (pg. 
270
-
275
)

Appendix

Korea Acute Myocardial infarction Registry (KAMIR) Investigators

Myung Ho Jeong MD, Young Keun Ahn MD, Sung Chull Chae MD, Jong Hyun Kim MD, Seung Ho Hur MD, Young Jo Kim MD, In Whan Seong MD, Dong Hoon Choi MD, Jei Keon Chae MD, Taek Jong Hong MD, Jae Young Rhew MD, Doo Il Kim MD, In Ho Chae MD, Jung Han Yoon MD, Bon Kwon Koo MD, Byung Ok Kim MD, Myoung Yong Lee MD, Kee Sik Kim MD, Jin Yong Hwang MD, Myeong Chan Cho MD, Seok Kyu Oh MD, Nae Hee Lee MD, Kyoung Tae Jeong MD, Seung Jea Tahk MD, Jang Ho Bae MD, Seung Woon Rha MD, Keum Soo Park MD, Chong Jin Kim MD, Kyoo Rok Han MD, Tae Hoon Ahn MD, Moo Hyun Kim MD, Ki Bae Seung MD, Wook Sung Chung MD, Ju Young Yang MD, Chong Yun Rhim MD, Hyeon Cheol Gwon MD, Seong Wook Park MD, Young Youp Koh MD, Seung Jae Joo MD, Soo Joong Kim MD, Dong Kyu Jin MD, Jin Man Cho MD, Byung Ok Kim MD, Sang-Wook Kim MD, Jeong Kyung Kim MD, Tae Ik Kim MD, Deug Young Nah MD, Si Hoon Park MD, Sang Hyun Lee MD, Seung Uk Lee MD, Hang-Jae Chung MD, Jang Hyun Cho MD, Seung Won Jin, MD, Yang Soo Jang MD, Jeong Gwan Cho, MD and Seung Jung Park MD.

Author notes

*
See the appendix.

Comments

0 Comments