Impact of Renal Dysfunction on Long-Term Survival After Isolated Coronary Artery Bypass Surgery

Transcription

1 Impact of Renal Dysfunction on Long-Term Survival After Isolated Coronary Artery Bypass Surgery Ye Lin, MD,* Zhe Zheng, MD,* Yan Li, MD,* Xin Yuan, MD, Jianfeng Hou, MD, Shiju Zhang, MD, Hongguang Fan, MD, Yang Wang, MD, Wei Li, MD, and Shengshou Hu, MD Department of Cardiovascular Surgery and Research Center for Cardiovascular Regenerative Medicine and Division of Biometrices National Center for Cardiovascular Disease, Cardiovascular Institute and Fuwai Hospital, Chinese Academy of Medical Science and Peking Union Medical College, Beijing, China Background. Preoperative renal dysfunction has been an important predictor for adverse cardiovascular events after coronary artery bypass grafting (CABG). In the past, serum creatinine was widely used to assess renal function. Until recently, estimated glomerular filtration rate (egfr) was recommended in evaluating renal function. The Cockcroft-Gault formula and the Modification of Diet in Renal Disease (MDRD) equation are two widely used formulas in clinical practice. Which method best predicts long-term outcome after CABG is still unknown. This study compared the predictive effectiveness of the Cockcroft-Gault formula, the MDRD equation, and serum creatinine level for in-hospital and long-term mortality. Methods. We retrospectively reviewed data collected from 5559 patients who underwent isolated CABG at Fuwai Hospital from January 1999 to December The main outcomes were in-hospital and long-term mortality. Receiver operating characteristic (ROC) curves and Cox analysis were used for the comparison. Results. Mean follow-up was months. ROC curve analysis showed that the Cockcroft-Gault formula had the greatest accuracy for predicting in-hospital mortality (area under the curve, 0.755; p < 0.001). Multivariate analysis confirmed that the egfr based on the Cockcroft- Gault formula was an independent predictor of in-hospital (odds ratio, 4.51, p < 0.001) and long-term (hazard ratio, 1.54; p 0.003) mortality. Both formulas were better than the serum creatinine level. Conclusions. Both formulas could provide a better measure of risk assessment than serum creatinine for in-hospital and long-term mortality. The Cockcroft-Gault formula was better than the MDRD equation for predicting in-hospital mortality. (Ann Thorac Surg 2009;87: ) 2009 by The Society of Thoracic Surgeons Coronary artery bypass grafting (CABG) is a successful surgical treatment for prolongation of life in selected patients with coronary artery disease [1]. Preoperative renal dysfunction has been an important predictor of adverse cardiovascular events after CABG [2 7]. Most previous studies have used the serum creatinine (SCr) concentration as an assessment of renal dysfunction. However, SCr concentration is affected by many factors other than glomerular filtration rate (GFR), such as age, sex, and body size. GFR is probably the best overall index of renal function [8]. The gold standard for determining GFR is to measure the clearance of exogenous substances such as inulin, chromium-51 ethylenediamine tetraacetic acid, technetium-99m labelled diethylenetriamine penta-acetate, and iodine-125 labeled iothalamate. However, the techniques used in determining GFR are time-consuming, laborintensive, and expensive, which make them unpractical for routine monitoring. For this reason, the estimated glomerular filtration rate (egfr) was recommended in evaluating renal function [8]. The formula that has been used most commonly is the Cockcroft-Gault formula [9]. More recently, the National Kidney Foundation of American recommended using the Modification of Diet in Renal Disease (MDRD) equation to estimate GFR [8, 10]. Even so, the method that best predicts in-hospital and long-term outcome after CABG is still unknown. This study retrospectively reviewed 5559 patients who underwent isolated CABG with an average follow-up of 4.7 years to compare the predictive effectiveness of the Cockcroft-Gault formula, the MDRD equation, and SCr concentration for in-hospital and long-term mortality after isolated CABG. Accepted for publication Jan 27, *The first three authors contributed equally to this work. Address correspondence to Dr Hu, Cardiovascular Institute and Fuwai Hospital, Peking Union Medical College and Chinese Academy of Medical Science, 167A Beilishi Rd, Xi Chen District, Beijing, , Peoples Republic of China; shengshouhu@yahoo.com. Patients and Methods Patients The study included 5559 patients who underwent isolated CABG between January 1999 and December A prespecified case report form was designed for data 2009 by The Society of Thoracic Surgeons /09/$36.00 Published by Elsevier Inc doi: /j.athoracsur

2 1080 LIN ET AL Ann Thorac Surg CABG AND RENAL DYSFUNCTION 2009;87: Table 1. Clinical Characteristics in Terms of Preoperative Renal Function Characteristics Mean SD, or No. (%) Severe egfr 30 (n 21) egfr Levels by Cockcroft-Gault Formula (ml/min/1.73 m 2 ) Moderate egfr (n 1031) Mild egfr (n 2765) Normal egfr 90 (n 1742) p Value Age, y Female gender 10 (47.6) 301 (29.3) 424 (15.3) 178 (10.2) Current smoker 1 (4.8) 119 (11.5) 421 (15.2) 416 (23.9) Hypertension 15 (71.4) 658 (63.8) 1618 (58.5) 1023 (58.7) Diabetes 9 (42.9) 274 (26.6) 666 (24.1) 457 (26.2) Hyperlipidemia 6 (28.6) 350 (33.9) 1017 (36.8) 669 (38.4) Previous MI 4 (19) 211 (20.5) 499 (18) 326 (18.7) Left main stenosis 50% 12 (57.1) 340 (33) 799 (28.9) 486 (27.9) Triple-vessel disease 16 (76.2) 884 (85.7) 2298 (83.1) 1379 (79.2) COPD 5 (23.8) 186 (18)) 431 (15.6) 205 (11.8) Previous stroke 5 (23.8) 95 (9.2) 177 (6.4) 79 (4.5) Ejection fraction Last creatinine value, mol/l Emergency/urgent operation 1 (4.8) 55 (5.3) 96 (3.5) 59 (3.4) 0.04 On-pump 11 (52.4) 547 (53.1) 1612 (58.3) 896 (51.4) CAD coronary artery disease; COPD chronic obstructive pulmonary disease; egfr estimated glomerular filtration rate; MI myocardial infarction; On-pump cardiopulmonary bypass; SD standard deviation. collection. Definitions of preoperative characteristics were consistent with those of Society of Thoracic Surgeons (STS) database. Data were 100% complete for critical preoperative risk factors of interest as well as for each major postoperative hospital outcome. Estimation of Renal Function No patient in this study was receiving dialysis before CABG. The last single SCr levels (within 72 hours before operation) in mol/l were obtained for all patients. Estimated GFR was calculated with the Cockcroft-Gault formula and the MDRD equation. The Cockcroft-Gault formula was adjusted with body surface area (BSA). Renal dysfunction was defined according to the guideline from the National Kidney Foundation [8]. Normal renal function was defined as egfr of 90 ml/min/1.73 m 2 or more. Mild, moderate, and severe renal dysfunction were defined as egfr of 60 to 90, 30 to 60, and less than 30 ml/min/1.73 m 2, respectively. Because the Cockcroft-Gault formula had the greatest accuracy for predicting in-hospital mortality in this study, the four major study groups were divided by the Cockcroft-Gault formula (Tables 1 and 2). Zakeri and colleagues [11] used preoperative SCr of 130 mol/l or higher in their article to define mild to moderate renal dysfunction and stated that mild renal dysfunction was an important predictor of outcome in terms of inhospital death. For this reason, we used SCr 130 mol/l or higher to compare with egfr of less than 60 ml/min/1.73 m 2 for predicting in-hospital and longterm mortality. Table 2. Clinical Outcomes in Terms of Preoperative Renal Function Outcomes No. (%) or Mean SD Severe egfr 30 (n 21) egfr Levels by Cockcroft-Gault Formula (ml/min/1.73 m 2 ) Moderate egfr (n 1031) Mild egfr (n 2765) Normal egfr 90 (n 1742) p Value In-hospital mortality 2 (9.5) 47 (4.6) 21 (0.8) 4 (0.2) Hemofiltration 1 (4.8) 9 (0.9) 7 (0.3) 8 (0.1) Postoperative coma 4 (19) 26 (2.5) 21 (0.8) 4 (0.2) Pneumonia 3 (14.3) 39 (3.8) 26 (0.9) 14 (0.8) Mechanical ventilation time, h Reoperation for bleeding 0 20 (1.9) 42 (1.6) 15 (0.9) Atrial fibrillation 6 (28.6) 350 (33.9) 1017 (36.8) 669 (38.4) IABP use 4 (19) 133 (12.9) 252 (9.1) 104 (6) Long-term mortality 4 (23.5) 68 (7.3) 121 (4.6) 55 (3.2) egfr estimated glomerular filtration rate; IABP intraaortic balloon pump; SD standard deviation.

3 Ann Thorac Surg LIN ET AL 2009;87: CABG AND RENAL DYSFUNCTION 1081 Table 3. Comparison of Area Under Receiver Operating Curve for Modification of Diet in Renal Disease Equation, Cockcroft-Gault Formula and Serum Creatinine Serum Creatinine Cockcroft-Gault Formula MDRD Equation Mortality AUC 95% CI p Value AUC 95% CI p Value AUC 95% CI p Value In-hospital Long-term AUC area under the receiver operating characteristic curve; CI confidence interval; MDRD Modification of Diet in Renal Disease. The Cockcroft-Gault formula [9] is as follows: egfr (140 age) weight 1.73 m 2 72 SCr BSA (m 2 ) for men egfr (140 age) weight 1.73 m SCr BSA (m 2 ) for women The MDRD equation [8, 10, 12] is as follows: egfr (SCr) (age) for men egfr (SCr) (age) for women where egfr is the estimated glomerular filtration rate in ml/min/1.73 m 2, the SCr concentration is in mg/dl, age is in years, and weight is given in kg. Patient Follow-Up and Outcomes According to institutional follow-up protocol, patients discharged alive were required to visit our outpatient clinic 6 months after CABG and then once every year. Patients were followed up through two cross-sectional investigations in 2007 and 2008, during which 92% of all patients were contacted by telephone and 5% by mail. The information of 0.6% patients was retrieved from their medical records. Follow-up was completed for 5485 patients (97.6%). The mean follow-up was months. The main events analyzed in this study were in-hospital mortality, long-term mortality, and major morbidity, including hemofiltration, postoperative coma, postoperative intraaortic balloon pump use, stroke, pneumonia, and atrial fibrillation. Statistical Analysis Values were expressed as means standard deviation or percentages, as appropriate. Comparisons between means were performed using the t test, whereas differences in categoric variables were assessed using the Fisher exact test or the 2 test. The sensitivity and specificity of both formulas were assessed from nonparametric receiver operating characteristic (ROC) curves generated by plotting sensitivity vs 1 specificity, giving the ideal test a sensitivity 1 and specificity 1. Areas under the curve (AUC) were calculated and compared according to the procedure of Hanley and McNeil [13]. Survival curves of the four groups were generated using the Kaplan-Meier method, and group differences were assessed by the log-rank test. Multivariate logistic regression and Cox proportional hazard models were created with the use of perioperative variables to identify independent predictors of in-hospital and long-term mortality. We developed three models to compare the Cockcroft- Gault formula, the MDRD equation, and SCr for predicting mortality. Renal dysfunction was defined as egfr of less than 60 ml/min/1.73 m 2 by the Cockcroft-Gault formula, egfr of less than 60 ml/min/1.73 m 2 by the MDRD equation, and SCr of 130 mol/l or more. Other candidate variables included age older than 65 years, female gender, chronic obstructive pulmonary disease (COPD), history of stroke, hypertension, hyperlipidemia, diabetes, low ejection fraction, previous myocardial infarction, emergency operation, left main disease, triplevessel disease, and off-pump technique. Only variables with a value of p 0.25 at univariate analysis were included in the multivariate Cox regression model. All statistical tests were 2-tailed, and a significant level of 0.05 was used throughout. Odds ratios (OR) and hazard ratios (HR) with 95% confidence intervals (CI) were Fig 1. (A) In-hospital mortality based on increasing estimated glomerular filtration rate (egfr) comparing the Cockcroft-Gault formula (solid line) and the Modification of Diet in Renal Disease (MDRD, dashed line). (B) In-hospital mortality based on decreasing serum creatinine. (C) Long-term mortality based on increasing egfr comparing the Cockcroft-Gault formula (solid line) and the MDRD, dashed line). (D) Long-term mortality based on decreasing serum creatinine.

4 1082 LIN ET AL Ann Thorac Surg CABG AND RENAL DYSFUNCTION 2009;87: Table 4. Multivariate Analysis of the Risk of Mortality a Serum Creatinine Cockcroft-Gault Formula MDRD Equation Mortality OR b /HR c 95% CI p Value OR b /HR c 95% CI p OR b /HR c 95% CI p In-hospital Long-term a Candidate variables included renal dysfunction, defined as estimated glomerular filtration rate (egfr) 60 ml/min/1.73 m 2 or serum creatinine 130 mol/l; age 65 years, female gender, chronic obstructive pulmonary disease, history of stroke, hypertension, hyperlipidemia, diabetes, ejection fraction value 0.50, previous myocardial infarction, emergency operation, left main disease, triple-vessel disease, off-pump technique. b OR (egfr 60 ml/min/1.73 m 2 or serum creatinine 130 mol/l) for in-hospital mortality by logistic regressions. c HR (egfr 60 ml/min/1.73 m 2 or serum creatinine 130 mol/l) for long-term mortality by Cox regressions. CI confidence interval; HR hazard ratio; OR odds ratio. calculated. Statistical analyses were performed using SPSS 13.0 software (SPSS Inc, Chicago, IL). Results The study included 5559 consecutive patients who underwent isolated CABG. These patients were divided by the Cockcroft-Gault formula. There were 21 patients (0.38%) with severe renal dysfunction (egfr 30 ml/min/1.73 m 2 ), 1031 (18.5%) with moderate renal dysfunction (egfr 30 to 59 ml/min/1.73 m 2 ), 2765 (49.7%) with mild renal dysfunction (egfr 60 to 89 ml/min/1.73 m 2 ), and 1742 (31.3%) with normal renal function (egfr 90 ml/min/1.73 m 2 ). Perioperative characteristics for patients in the four main study groups are listed in Table 1. Patients with poorer renal function were older, were more often female, and had a higher prevalence of hypertension. ROC Curve Comparisons for Predicting Mortality To the MDRD equation, the Cockcroft-Gault formula, and SCr for predicting mortality, the ROC curves were constructed using all egfr and SCr measures. The AUC were calculated and compared (Table 3). The Cockcroft- Gault formula had the greatest accuracy for predicting in-hospital mortality (AUC, 0.755; 95% CI, to 0.815; p 0.001). ROC curve analysis showed a similar AUC for egfr and SCr in predicting long-term mortality (AUC, for Cockcroft-Gault formula, p 0.001; AUC, for MDRD equation, p 0.001; AUC, for SCr, p 0.001). In-Hospital Outcomes There were 74 in-hospital all-cause deaths (1.33%). The patients with moderate to severe renal dysfunction (egfr 60 ml/min/1.73 m 2 ) had a higher rate of postoperative complications (Table 2). The in-hospital mortality rates increased with increasing preoperative SCr levels and decreasing egfr (Fig 1). Three multivariate logistic regression analysis models were created to identify independent predictors of in-hospital mortality. Renal dysfunction was defined as egfr of less than 60 ml/min/1.73 m 2 by the Cockcroft-Gault formula, egfr of less than 60 ml/min/1.73 m 2 by the MDRD equation, and SCr of 130 mol/l or more. The Cockcroft-Gault formula had the maximum OR of 4.51 for predicting in-hospital mortality; the ORs were 3.43 for the MDRD equation and 2.86 for SCr (p Table 4). Survival Analysis There were 248 (4.52%) late all-cause deaths during follow-up. Figure 2 shows the Kaplan-Meier survival analyses according to the three renal assessment methods. The log-rank test was used to describe the differences among four groups. The groups with moderate to severe renal dysfunction had higher long-term mortality rates. Both formulas were better than SCr for predicting long-term mortality. Cox Multivariate Analysis We used Cox proportional hazard models to examine the relationship between renal dysfunction and other known preoperative risk factors for predicting long-term mortality. We also created three models by using the three renal assessment methods. Table 4 presents HRs of the three methods for predicting long-term mortality. There were no significant differences between the two formulas, and both formulas were more accurate in predicting longterm mortality than SCr (HR, 1.60 for the MDRD equation, p 0.002; 1.54 for the Cockcroft-Gault formula, p 0.003; 1.40 for SCr, p 0.108). Long-term mortality rates increased with decreasing egfr. In contrast, the relationship between SCr and long-term mortality rates was nonlinear (Fig 1). Comment Principle Findings We compared the predictive effectiveness of the Cockcroft-Gault formula, the MDRD equation, and SCr for in-hospital and long-term mortality in patients after isolated CABG at our institution. Both the Cockcroft- Gault formula and the MDRD equation could provide a better measure of risk assessment than SCr for inhospital and long-term mortality. The Cockcroft-Gault formula was better than the MDRD equation for predicting in-hospital mortality. There was no significant difference between the MDRD equation and the Cockcroft- Gault formula for predicting long-term mortality.

5 Ann Thorac Surg LIN ET AL 2009;87: CABG AND RENAL DYSFUNCTION 1083 Fig 2. Kaplan-Meier estimated survival according to estimated glomerular filtration rate (egfr; ml/min/1.73 m 2 ) calculated by the (top panel) Cockcroft-Gault formula, (middle panel) the Modification of Diet in Renal Disease (MDRD), and (bottom panel) and serum creatinine ( mol/l). Comparison of MDRD Equation, Cockcroft-Gault Formula, and SCr A number of renal function tests are available, and each focuses on a unique aspect of kidney function. SCr is the most widely used assessment of renal function; however, it can be affected by many factors, such as muscle mass, dietary intake, changes in tubular secretion, and interference by other substances in the serum. As a result, SCr concentration is suboptimal for estimating renal function. Later on, egfr was recommended in evaluating renal function [8]. The Cockcroft-Gault formula is probably one of the most widely used prediction equations [9]. The MDRD equation was recently recommended by the National Kidney Foundation [8, 10]. Its advantage over the Cockcroft-Gault formula was documented in some [14] but not all recent reports [15, 16]. Neither formula was developed for patients with cardiac disease. Moreover, they differ in variables and coefficients. The best method to predict in-hospital and long-term outcome after CABG is still unknown. In this study, the Cockcroft-Gault formula was found to be better than the MDRD equation in terms of predictive power for in-hospital mortality. The fact that the former can incorporate more variables, such as BSA and weight, serves as the underlying reason. We assume that these factors indirectly reflect body mass index (BMI), which is generally identified as a risk factor for inhospital mortality after CABG [17, 18]. He and colleagues [19] reported that a BMI of lower than 18.5 was an independent risk factor of death in the Chinese population. The Cockcroft-Gault formulas are adjusted using weight and BSA and therefore are more powerful for such prediction purposes. Previous studies showed that the bias and precision of the MDRD equation were poorer and tended to underestimate true GFR in patients with levels of renal function nearer to normal [20, 21]. In our cohort, renal function in most patients ranged from normal to moderate. This may be another reason explaining the superiority of the Cockcroft-Gault formula over the MDRD equation for predicting in-hospital mortality. Although the two formulas varied in their accuracy for predicting in-hospital mortality, we found a similar trend of decreased mortality rates with increased egfr by both the Cockcroft-Gault formula and the MDRD equation (Fig 1). There was also no statistical difference between the two formulas for predicting in-hospital mortality using ROC curves and multivariate logistic regression analysis (p 0.001), so both the Cockcroft-Gault formula and the MDRD equation could be used for predicting in-hospital and long-term mortality after CABG. Previous studies showed an increase in all-cause and cardiovascular mortality in patients with renal dysfunction [4 7]. Our study results are similar to those reported. The recent analysis had indicated mild renal dysfunction as a major risk factor for cardiovascular complications and death [7, 11, 22]. Our study confirmed and extended these findings with long-term follow-up results.

6 1084 LIN ET AL Ann Thorac Surg CABG AND RENAL DYSFUNCTION 2009;87: Limitations This study has several limitations. This study was performed in a single center. The GFR was estimated by using the Cockcroft-Gault formula and the MDRD equation. The formula provides an acceptable estimate of GFR but it is not the gold standard for determining GFR. Also, because renal function measurements (egfr) were based on a single preoperative SCr value, which might fluctuate, particularly in patients with unstable hemodynamics and various medical therapies, we were unable to examine the effect of a rising or falling creatinine trajectory. Renal function in most patients in our cohort was normal to moderate. This might also affect our findings. Conclusions This study showed that an egfr of less than 60 ml/min/ 1.73 m 2 was an independent predictor of in-hospital and long-term mortality in Chinese patients after CABG. The Cockcroft-Gault formula was better than the MDRD equation for predicting in-hospital mortality, and both formulas were superior to SCr for predicting in-hospital and long-term mortality. Funding for this study was supported by the Key Project in the National Science and Technology Pillar Program during the Eleventh Five-Year Plan Period (2006BAI01A09) and the Key Project of Beijing Municipal Science and Technology Commission (D ). References 1. Eagle KA, Guyton RA, Davidoff R, et al. ACC/AHA guidelines for coronary artery bypass graft surgery: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Revise the 1991 Guidelines for Coronary Artery Bypass Graft Surgery). J Am Coll Cardiol 1999;34: Liu JY, Birkmeyer NJ, Sanders JH, et al. Risks of morbidity and mortality in dialysis patients undergoing coronary artery bypass surgery. Circulation 2000;102: Anderson RJ, O brien M, MaWhinney S, et al. Renal failure predisposes patients to adverse outcome after coronary artery bypass surgery. VA Cooperative Study #5. Kidney Int 1999;55: O Connor GT, Plume SK, Olmstead EM, Coffin LH, Martin JR. Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Circulation 1992;85: Shroyer AL, Grover FL, Edwards FH Coronary artery bypass risk model: The Society of Thoracic Surgeons Adult Cardiac National Database. Ann Thorac Surg 1998;65: 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: Weerasinghe A, Hornick P, Smith P, Taylor K, Ratnatunga C. Coronary artery bypass grafting in non-dialysis-dependent mild-to-moderate renal dysfunction. J Thorac Cardiovasc Surg 2001;121: Levey AS, Coresh J, Balk E, Kausz AT, et al; National Kidney Foundation. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003;139: Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16: National Kidney Foundation: K/DOQI Clinical Practice Guidelines for Chronic Kidney Disease: evaluation, classification, and stratification: Am J Kidney Dis 2002;39(suppl): S Zakeri R, Freemantle N, Lipkin GW, et al. Relation between mild renal dysfunction and outcomes after coronary artery bypass grafting. Circulation 2005;112(9 suppl):i Levey AS, Bosch JP, Lewis JB, Greene T, Rodgers N, Roth D. 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: Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143: Hallan S, Asberg A, Lindberg M, Johnsen H. Validation of the Modification of Diet in Renal Disease formula for estimating GFR with special emphasis on calibration of the serum creatinine assay. Am J Kidney Dis 2004;44: Vervoort G, Willems HL, Wetzels JF. Assessment of glomerular filtration rate in healthy subjects and normal albuminuric diabetic patients: validity of a new (MDRD) prediction equation. Nephrol Dial Transplant 2002;17: Lamb EJ, Webb MC, Simpson DE, Coakley AJ, Newman DJ, O Riordan SE. Estimation of glomerular filtration rate in older patients with chronic renal insufficiency: is the Modification of Diet in Renal Disease formula an improvement? J Am Geriatr Soc 2003;51: Romero-Corral A, Montori VM, Somers VK, et al. Association of bodyweight with total mortality and with cardiovascular events in coronary artery disease: a systematic review of cohort studies. Lancet 2006;368: Wagner BD, Grunwald GK, Rumsfeld JS, et al. Relationship of body mass index with outcomes after coronary artery bypass graft surgery. Ann Thorac Surg 2007;84: He J, Gu D, Wu X, Reynolds K, et al. Major causes of death among men and women in China. N Engl J Med 2005;353: Bostom AG, Kronenberg F, Ritz E. Predictive performance of renal function equations for patients with chronic kidney disease and normal serum creatinine levels. J Am Soc Nephrol 2002;13: Rule AD, Larson TS, Bergstralh EJ, Slezak JM, Jacobsen SJ, Cosio FG. Using serum creatinine to estimate glomerular filtration rate: accuracy in good health and in chronic kidney disease. Ann Intern Med 2004;141: Van de Wal RM, Van Brussel BL, Voors AA, et al. Mild preoperative renal dysfunction as a predictor of long-term clinical outcome after coronary bypass surgery. J Thorac Cardiovasc Surg 2005;129:330 5.