- Email: [email protected]
Comparison of the MDRD and the CKD-EPI equations to estimate the glomerular filtration rate in the general population
ARTICLE IN PRESS Med Clin (Barc). 2010;134(14):617 –623
www.elsevier.es/medicinaclinica
Original
Comparison of the MDRD and the CKD-EPI equations to estimate the glomerular filtration rate in the general population Alejandro Lo´pez-Sua´rez a,, Manuel Beltra´n-Robles a, Javier Elvira-Gonza´lez a, ˜ ana-Quirell a, Fernando Ferna´ndez-Palacı´n b,c, Antonio Bascun a Julio Benı´tez-Del-Castillo , Pablo Go´mez-Ferna´ndez d a
´diz, Spain ´ car de Barrameda, Ca Internal Medicine Department, Hospital Virgen del Camino, Sanlu ´diz, Ca ´diz, Spain Statistics and Operational Research Department, Universidad de Ca c ´diz, Spain ´ car de Barrameda, Ca Nephrology Department, Hospital Virgen del Camino, Sanlu d ´diz, Spain Nephrology Department, Hospital del SAS, Jerez de la Frontera, Ca b
ARTICLE INFO
A B S T R A C T
Article history: Received 11 June 2009 Accepted 30 July 2009 Available online 11 de marzo de 2010
Background and objective: The Chronic Kidney Disease Epidemiology Collaboration (CDK-EPI) equation has been proposed as a replacement for the Modification of Diet in Renal Disease (MDRD) equation to estimate the glomerular filtration rate, but this equation has not yet been evaluated in the general population. Patients and methods: Cross-sectional analysis of a random sample of 858 participants from the general population aged 50 – 75 years without known kidney disease. The prevalence of low eGFR (o 60 mL/min/ 1.73 m2) was assessed with the MDRD and the CKD-EPI equations in the overall sample and in normoalbuminuric individuals. Results: With the MDRD equation the median eGFRs (interquartile range) in men/women were 63.3(12.2)/ 56.7(9.4) mL/min/1.73 m2, and with the CKD-EPI equation 66.6(14.2)/61.3(11.6) mL/min/1.73 m2. The prevalence of low eGFR in men/women was 35.2%/68.5% and 25.1%/45.7% with the MDRD and the CKD-EPI equations, respectively. Normoalbuminuric women without risk factors for CKD experienced the most pronounced reduction in the number of cases with low eGFR with the CKD-EPI equation. The prevalence of renal impairment in this subgroup still remained even greater than that in men with diabetes, hypertension, or cardiovascular disease. Conclusions: Compared with the MDRD, the CKD-EPI equation generates a substantial reduction in the prevalence of renal impairment in subjects with diabetes, hypertension, cardiovascular disease, and in subjects without risk factors. The prevalence of renal impairment in normoalbuminuric females may be still overestimated with the CKD-EPI equation. ˜a, S.L. & 2010 Published by Elsevier Espan
Keywords: Albuminuria Cardiovascular risk factors Chronic kidney disease CKD-EPI equation Glomerular filtration rate MDRD
Comparacio´n de las ecuaciones MDRD y CKD-EPI en la estimacio´n de la tasa de filtrado glomerular en la poblacio´n general R E S U M E N
Palabras clave: Albuminuria factores de riesgo cardiovascular enfermedad renal cro´nica CKD-EPI filtrado glomerular MDRD
Fundamento y objetivo: La ecuacio´n Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) ha sido propuesta para sustituir a la actualmente recomendada Modification of Diet in Renal Disease (MDRD) para el ca´lculo de la tasa de filtrado glomerular (TFG), si bien no ha sido au´n evaluada en la poblacio´n general. Pacientes y Me´todo: Estudio transversal de una muestra aleatoria de 858 individuos de la poblacio´n ˜ os sin enfermedad renal conocida. Comparacio´n de la prevalencia de general con edades entre 50 – 75 an una TFGo 60 ml/min/1,73 m2 calculada con las ecuaciones MDRD y CKD-EPI en toda la muestra y en el subgrupo de individuos normoalbuminu´ricos (cociente albu´mina/creatinina o 30 mg/g). Resultados: Con la ecuacio´n MDRD la mediana de la TFG (rango intercuartı´lico) en varones/mujeres fue de 63,3 (12,2)/56,7 (9,4) ml/min/1,73 m2, y con la ecuacio´n CKD-EPI 66,6 (14,2)/61,3 (11,6) ml/min/1,73 m2, respectivamente. La prevalencia de una TFG baja en varones/mujeres fue del 35,2 – 68,5%, y del 25,1 – 45,7% con las ecuaciones MDRD y CKD-EPI, respectivamente. El grupo de mujeres normoalbuminu´ricas sin factores de riesgo para enfermedad renal cro´nica presento´ la mayor reduccio´n en la prevalencia de una TFG baja con la ecuacio´n CKD-EPI. Sin embargo, la prevalencia en este grupo se mantuvo incluso ma´s elevada que en el de los varones con diabetes, hipertensio´n o enfermedad cardiovascular.
DOI of original article: 10.1016/j.medcli.2009.09.051
Corresponding author.
E-mail address: [email protected] (A. Lo´pez-Sua´rez). ˜ a, S.L. 0025-7753/$ - see front matter & 2010 Published by Elsevier Espan doi:10.1016/j.medcli.2009.07.055
ARTICLE IN PRESS 618
´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
Conclusiones: Comparada con la ecuacio´n MDRD, la ecuacio´n CKD-EPI produce una importante reduccio´n en la prevalencia de insuficiencia renal en individuos con diabetes, hipertensio´n, enfermedad cardiovascular, y en individuos sin factores de riesgo. La prevalencia de insuficiencia renal en mujeres normoalbuminu´ricas calculada con la ecuacio´n CKD-EPI puede estar au´n sobreestimada. ˜ a, S.L. & 2010 Publicado por Elsevier Espan
Introduction The overall prevalence of chronic kidney disease (CKD) is estimated at up to 11% of adults in the United States and 6% in Europe.1 Patients with CKD are at risk for not only progression to end stage renal disease but, and far more frequently, for the development of cardiovascular disease (CVD).2 Therefore, early identification of CKD by reliable methods is an important part of preventive health strategies.3 The Modification of Diet in Renal Disease (MDRD) study equation has become the preferred method for estimating glomerular filtration rate (eGFR).4,5 This formula was initially developed in patients with known CKD, but has subsequently been widely used in the general population and in people with diverse cardiovascular risk profiles, even though there is no proof of its reliability as a screening test in people without known CKD.6 – 13 The MDRD equation provides unbiased and accurate estimates in a wide range of subgroups of patients when the eGFR is o60 mL/ min/1.73 m2.14 For estimates equal or greater than 60 mL/min/ 1.73 m2 the MDRD equation is not only less accurate but also less precise, especially in the absence of albuminuria and other markers of kidney damage.14 – 17 The MDRD has also been reported to be less precise in the absence of CKD than in cases with known disease, as eGFR values significantly overlap between patients with and without CKD and this compromises the use of eGFR to separate these two patients groups.9 However, clinicians continue using the MDRD equation for detecting renal impairment in everyday clinical practice, as measured GFR is not a suitable tool in this setting.18 In an attempt to overcome some of the drawbacks of the MDRD equation, it has been suggested that screening for CKD should be targeted at subjects at risk,19 in whom an eGFRo60 mL/min/1.73 m2 could be considered a true-positive result.4,5 However, the assumption that in the absence of markers of kidney damage a correct interpretation of low eGFR can be ascertained by taking into account the presence or absence of risk factors for CKD has not been evaluated thoroughly. Another way to improve the accuracy obtained with the MDRD equation is the development of new estimative equations. Levey et al.20 have recently proposed a new equation to replace the original MDRD equation for the routine clinical use in the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) study. But although the CKD-EPI equation has improved in bias, its precision remains suboptimal.20 Until now there is limited information about the performance of this new equation in people from the general population, especially in people whose renal function is unknown and who do not have markers of kidney damage. Our study compared the eGFR generated by the MDRD and the CKD-EPI equations in a random sample of individuals from the general population without known kidney disease. We also assessed whether there was a significant difference in the estimates of GFR among normoalbuminuric participants with and without risk factors for CKD.
Methods Study population This study was performed using the data from the Sanlu´car Study, the design of which is described elsewhere.21 It was a
cross-sectional population-based study carried out in Sanlu´car de Barrameda, Spain, between July 2006 and February 2007. The study population was a random sample of 858 participants aged 50 – 75 years obtained from the local census, all of Caucasian origin. The investigation was carried out in accordance with the principles of the Declaration of Helsinki. Subjects who agreed to participate were included after being interviewed and informed about the nature, risks, and benefits of the study in the clinic and giving written consent. Measurements All participants were studied using a standardized interview, physical examination, and blood and urine testing. Patients reporting a known diagnosis of renal failure were not included (three cases with end-stage renal disease and one needing hemodialysis). Blood pressure was recorded with a validated sphygmomanometer (OMRON 705CP, Hoofddorp, The Netherlands). A cuff with a 12 40 cm bladder was used in obese participants with an arm circumference 435 cm. Blood pressure was measured at the clinic with the subject seated after resting for 5 min. The selected systolic (SBP) and diastolic (DBP) blood pressures were the means of three consecutive measurements not differing by more than 10 mmHg in the SBP. Participants who were unaware of their blood pressure and whose office mean blood pressure was Z140/90 mmHg were offered 24 h ambulatory blood pressure monitoring (BR-102, Schiller AG, Baar, Switzerland) or 1 week of home blood pressure self-measurement. To identify peripheral arterial disease, all participants underwent ankle – brachial index examination, using a handleheld vascular Doppler with an 8 MHz probe (Mini Dop ES-100VX, Hayashi Denki Co., Ltd., Kawasaki, Japan) and a calibrated mercury sphygmomanometer. The procedure used to obtain the SBP in the upper and the lower extremities has been previously described.22 Laboratory tests and estimates of GFR Blood and urine samples were obtained from all participants in the early morning after overnight fasting. All participants were asymptomatic and without intercurrent illnesses at the time of testing. An ILAB-600 analyzer and standardized reagents (ITC Diagnostics, Izasa S.A., Barcelona, Spain) were used for measuring plasma glucose, total cholesterol, HDL-cholesterol, triglycerides, and glycosylated hemoglobin. The Friedewald formula was used for calculating LDL-cholesterol. Serum creatinine was measured using the Jaffe method with an alkaline picrate assay, giving a normal range up to 1.10 mg/dL in women and 1.30 mg/dL in men. This method was calibrated to be traceable to an isotope dilution mass spectrometry reference method.23 Albumin was measured in an early morning urine sample using a colorimetric method, and the results were expressed as milligrams of urinary albumin per gram of serum creatinine (ACR). Estimated GFR was calculated with the re-expressed MDRD equation for standardized serum creatinine (1)24 and the CKD-EPI equation (2)20: GFR ¼ 175 Sc1:154 age0:203 0:742 ½if female;
ð1Þ
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
if female and Sc r 0:70 mg=dL
GFR ¼ 144ðSc=0:7Þ0:329 ð0:993Þage ;
if female and Sc4 0:70 mg=dL
ð2Þ
GFR ¼ 144ðSc=0:7Þ1:209 ð0:993Þage ;
if male and Sc r0:90 mg=dL
GFR ¼ 141ðSc=0:9Þ0:411 ð0:993Þage ;
if male and Sc 40:90 mg=dL
GFR ¼ 141ðSc=0:9Þ1:209 ð0:993Þage ;
where age is expressed in years and standardized serum creatinine (Sc) in mg/dL. An eGFR value less than 60 mL/min/1.73 m2 was selected to define an abnormal GFR.4,5 Albuminuria was defined as an ACRZ30 mg/g. Risk factors for CKD Type 2 diabetes, hypertension, and cardiovascular disease were the risk factors for CKD selected for the purposes of the study. Diabetes was diagnosed in participants being treated for diabetes and in those with fasting serum glucose Z7.0 mmol/L. Hypertension was defined as already being treated with antihypertensive drugs or having a mean blood pressure of Z135/85 mmHg following 24-h ambulatory blood pressure monitoring or 1 week of home blood pressure self-measurement. Patients with CVD were those with a documented diagnosis of a transient ischemic attack, ischemic stroke, stable angina, acute coronary syndrome, coronary revascularization, or peripheral arterial disease defined as an ankle – brachial index o0.90. Other cardiovascular and CKD risk factors were also considered. Subjects were classified as smokers (current smokers or smokers in the last year) or non-smokers (participants who never smoked or who stopped smoking a year ago). The Framingham risk score was used as an integrated tool of traditional cardiovascular risk factors.25 As a reference group for comparing the performance of eGFR in subjects with risk factors for CKD, we defined a group of participants with no risk factors as the non-smoking, nondiabetic, and non-hypertensive participants with no history of CVD and a Framingham risk score o10%. Statistical analysis Results are expressed as mean (95% confidence interval) for normally distributed continuous data, median (interquartile range) for continuous data without normal distribution, and frequencies (95% confidence intervals) for categorical data. Differences in the risk factor profile were compared between albuminuric and normoalbuminuric participants by categorizing
619
an eGFR lower vs. greater than 60 mL/min/1.73 m2 using the chi-square test for categorical variables, and ANOVA and Kruskal – Wallis test for normally and non-normally distributed continuous data respectively. A P-valueo0.05 was accepted as being statistically significant. SPSS for Windows, version 12.0 (Chicago, Illinois, USA) was used for the statistical analyses.
Results GFR calculated with the MDRD equation The general characteristics of the study sample are summarized in Table 1. The overall prevalence of renal impairment was 53%. Among them, 92.1% had an eGFR of 45 – 60 mL/min/1.73 m2. There was a marked difference in the median eGFR and the prevalence of low eGFR by gender (Table 2). An important proportion of all cases with renal impairment was normoalbuminuric women, in whom the percentage of low eGFR was as high as in women with an ACRZ30 mg/g (68.8% vs. 66.1%, respectively). Only 15.8% of cases with an eGFRo60 ml/min/1.73 m2 also had an ACRZ30 mg/g, whereas up to 58.5% of participants with albuminuria also had a low eGFR. The presence of albuminuria was associated with the most adverse risk factor profile regardless of the eGFR level. However, in the absence of albuminuria, no substantial difference in the risk factor profile was observed when comparing the groups with an eGFRZ or o60 mL/min/1.73 m2 (Tables 3 and 4). The values of the eGFR in the 735 normoalbuminuric participants, whether greater or lower than 60 mL/min/1.73 m2, closely overlapped between individuals with and without risk factors for CKD (Figs. 1 and 2). In the subgroup of normoalbuminuric women with diabetes, hypertension, or cardiovascular disease, the range of eGFR values and the proportion of cases with an eGFRo60 mL/min/1.73 m2 were very similar to those in women with no risk factors (Fig. 3). GFR calculated with the CKD-EPI equation Compared with the MDRD, the most noteworthy result obtained with the CKD-EPI equation was a 16.9% absolute reduction of participants with an eGFR o60 mL/min/1.73 m2, a reduction that was greater in women than in men (22.8% vs. 10.1%). Overall, in 91.6% of cases with renal impairment, the estimates fell in the range of 45 – 60 mL/min/1.73 m2. The subgroup of participants with the most pronounced reduction in the percentage of cases with low eGFR was the normoalbuminuric women with no CKD risk factors (29.8% absolute reduction).
Table 1 General characteristics of the study sample
Age (years) Smokers Diabetes Hypertension Body mass index Z30 kg/m2 Framingham risk score* o10% 420% Cardiovascular disease Participants with no risk factors Serum creatinine (mg/100 mL)
All (n= 858)
Women (n =460)
Men (n= 398)
61.5 15.1 27.6 49.0 55.1
(61.0 – 61.9) (12.8 – 17.7) (24.7 – 30.7) (45.6 – 52.4) (51.7 – 58.5)
61.1 7.9 26.1 52.4 56.1
(60.5 – 61.7) (5.5 – 10.7) (22.1 – 30.4) (47.7 – 57.0) (51.4 – 60.7)
61.9 23.4 29.4 45.0 54.0
(61.2 – 62.5) (19.3 – 27.8) (25.0 – 34.1) (40.0 – 50.0) (49.0 – 59.0)
59.6 11.2 15.6 24.6 1.09
(56.2 – 62.9) (9.2 – 13.4) (13.3 – 18.2) (21.7 – 27.6) (1.08 – 1.10)
90.7 1.5 13.3 34.3 1.01
(87.6 – 93.2) (0.6 – 3.1) (10.3 – 16.7) (30.0 – 38.9) (1.00 – 1.02)
23.6 22.4 18.3 13.3 1.18
(19.5 – 28.1) (18.4 – 26.8) (14.7 – 22.5) (10.1 – 17.1) (1.16 – 1.20)
Data are mean (95% confidence interval) or percentages (95% confidence interval).
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
620
Table 2 Mean eGFR and percentage of low eGFR Women (n= 460)
Mean eGFRa Low eGFRb Low eGFR-ACRo 30c Low eGFR-ACRZ 30 a b c
Men (n= 398)
MDRD
CKD-EPI
MDRD
CKD-EPI
56.8 68.5 11.7 88.3
61.4 45.7 13.3 86.7
64.4 35.2 24.3 75.7
67.4 25.1 30.0 70.0
(56.0 – 57.5) (64.0 – 72.7) (8.4 – 15.8) (84.2 – 91.6)
(60.6 – 62.3) (41.0 – 50.3) (9.0 – 18.7) (81.3 – 91.0)
(63.2 – 65.6) (30.5 – 40.1) (17.4 – 32.2) (67.8 – 82.6)
(66.2 – 68.5) (20.9 – 29.7) (21.2 – 40.0) (60.0 – 78.8)
Mean (95% confidence interval), mL/min/1.73 m2. eGFRo 60 ml/min/1.73 m2 (percentage [95% confidence interval]). ACR (mg/g).
Table 3 Risk factor profile according to the albumin/creatinine ratio and eGFR categories (MDRD equation) mL/min/1.73 m2
n (%) Age (years) Gender (female) Glucose (mg/dL) Haemoglobin A1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) Body mass index (kg/m2) Waist circumference (cm) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) Triglycerides (mg/dL)a Framingham risk score o 10% 4 20%
ACRo 30 mg/g
ACR Z30 mg/g
eGFRZ 60
eGFRo 60
eGFR Z60
eGFRo 60
351 60.0 36.2 110 6.0 143 79 30.4 105 225 56 142 116
384 62.0 72.4 112 6.0 142 80 31.0 103 238 60 150 124
51 63.1 37.3 121 6.5 153 81 31.3 108 229 55 145 129
72 64.6 52.1 122 6.5 157 84 32.0 107 233 53 147 150
(40.9) (59.4 – 60.7) (31.1 – 41.5) (107 – 114) (5.9 – 6.1) (141 – 145) (78 – 81) (29.9 – 30.9) (104 – 106) (221 – 230) (55 – 57) (138 – 146) (83)
52.3 (47.1 – 57.7) 9.2 (6.3 – 12.6)
(44.7) (61.3 – 62.6) (67.6 – 76.8) (108 – 115) (5.9 – 6.1) (140 – 144) (79 – 81) (30.5 – 31.5) (102 – 104) (234 – 242) (58 – 61) (146 – 154) (83)
72.1 (67.4 – 76.6) 8.7 (6.0 – 11.9)
(6.0) (61.0 – 65.3) (24.1 – 51.9) (106 – 137) (6.1 – 6.9) (148 – 158) (78 – 84) (30.0 – 32.7) (103 – 112) (215 – 243) (51 – 59) (133 – 157) (116)
37.3 (24.1 – 51.9) 21.6 (11.3 – 35.3)
(8.4) (63.2 – 66.4) (40.7 – 64.7) (111 – 133) (6.1 – 6.8) (151 – 163) (81 – 87) (30.7 – 33.2) (104 – 110) (222 – 244) (50 – 56) (137 – 156) (106)
43.7 (31.4 – 55.3) 26.8 (16.7 – 38.1)
Data are percentages (95% confidence interval) or mean (95% confidence interval). Po 0.001. P o 0.01. Po 0.05. a Median (interquartile range).
Table 4 Risk factor profile according to the albumin/creatinine ratio and eGFR categories (CKD-EPI) mL/min/1.73 m2
n (%) Age (years) Gender (female) Glucose (mg/dL) Haemoglobin A1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) Body mass index (kg/m2) Waist circumference (cm) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) Triglycerides (mg/dL)a Framingham risk score o 10% 4 20%
ACRo 30 mg/g
ACR Z30 mg/g
eGFRZ 60
eGFRo 60
eGFR Z60
eGFRo 60
484 60.0 46.3 111 6.0 143 80 30.4 104 229 57 145 116
252 63.7 72.2 111 6.0 142 79 31.3 104 237 59 149 127
64 62.2 43.8 120 6.5 154 82 31.8 108 234 55 149 133
58 66.2 48.3 125 6.4 157 84 31.6 106 229 53 142 158
(56.4) (59.1 – 60.2) (40.5 – 52.1) (108 – 114) (5.9 – 6.1) (141 – 145) (79 – 81) (30.0 – 30.9) (103 – 105) (225 – 233) (56 – 59) (141 – 148) (80)
58.5 (52.6 – 64.1) 8.8 (5.9 – 12.6)
(29.3) (62.9 – 64.6) (64.6 – 79.0) (107 – 116) (5.9 – 6.1) (140 – 144) (78 – 80) (30.7 – 31.9) (102 – 105) (232 – 243) (57 – 61) (145 – 154) (82)
70.6 (64.6 – 76.2) 9.3 (5.9 – 13.4)
Data are percentages (95% confidence interval) or mean (95% confidence interval). Po 0.001. P o 0.01. Po 0.05. a Median (interquartile range).
(7.5) (60.4 – 64.4) (31.4 – 56.7) (107 – 132) (6.2 – 6.9) (149 – 159) (79 – 85) (30.5 – 33.1) (105 – 112) (221 – 247) (51 – 58) (139 – 160) (104)
42.2 (29.9 – 55.2) 18.8 (10.1 – 30.5)
(6.8) (64.5 – 67.9) (35.0 – 61.8) (112 – 138) (6.0 – 6.8) (150 – 163) (81 – 87) (30.3 – 33.0) (103 – 110) (218 – 241) (49 – 56) (132 – 153) (111)
39.7 (27.0 – 53.4) 31.0 (19.5 – 44.5)
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
621
Normoalbuminuric men 120
*
120
MDRD equation
100 2
eGFR (ml/min/1,73 m )
eGFR (ml/min/1,73 m2)
100
CKD-EPI equation
80
60
80
60
40
40
20
20 No risk factors
Diabetes
Hypertension
CVD
No risk factors
Diabetes
Hypertension
CVD
Fig. 1. Box plots comparing eGFR within different risk factors for CKD in normoalbuminuric men.
Normoalbuminuric women 120
120
MDRD equation
100 eGFR (ml/min/1,73 m2)
eGFR (ml/min/1,73 m2)
100
CKD-EPI equation
80
60
40
20
80
60
40
20 No risk factors Diabetes
Hypertension
CVD
No risk factors Diabetes
Hypertension
CVD
Fig. 2. Box plots comparing eGFR within different risk factors for CKD in normoalbuminuric women.
However, the prevalence of a low eGFR in this subgroup remained almost double that of men, irrespective of the presence or absence of risk factors for CKD (Fig. 3). Unlike with the MDRD equation, with the CKD-EPI equation normoalbuminuric women with diabetes, hypertension, or CVD had significantly lower levels of eGFR than women with no CKD risk factors (Fig. 3). However, the range of values of eGFR also overlapped between normoalbuminuric participants (men or women) with and without risk factors for CKD (Figs. 1 and 2). In the absence of albuminuria, no significant difference was observed between participants with an eGFRZ and o60 mL/min/1.73 m2 in systolic or diastolic blood pressure, body mass index, waist circumference, serum glucose, or hemoglobin A1C. The level of HDL-cholesterol and the percentage of cases with a Framingham risk score o10% were even greater among normoalbuminuric participants with low eGFR than among those with an eGFRZ60 mL/min/1.73 m2.
Discussion The main finding of this study is that the CKD-EPI equation, applied to the general population, generates higher eGFR values than the MDRD equation, decreasing the overall prevalence of low eGFR by 16.9% (36.1% vs. 53.0%). Normoalbuminuric women with no risk factors were the subgroup with the greatest reduction in the prevalence of renal impairment. However, the prevalence of apparent renal impairment in this subgroup still remained very high, especially in women with no CKD risk factors, in whom the rate of low eGFR was even higher than in normoalbuminuric men with any risk factor for CKD. The first practical consequence derived using the new equation is likely to be a significant reduction in the number of people previously labeled as having stage 3 CKD. These people would not undergo further nephrological evaluation or overaggressive cardiovascular risk factor
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
622
Normoalbuminuric men
Normoalbuminuric women
90
80
80
70
70
60
60 Porcentaje
Porcentaje
90
50 40
50 40
30
30
20
20
10
10
0
0 No risk factors
Diabetes
Hypertension
CVD
eGFR < 60 ml/min/1,73 m2 (MDRD)
No risk factors
Diabetes
Hypertension
CVD
eGFR < 60 ml/min/1,73 m2 (CKD-EPI)
Fig. 3. Percentage of low eGFR (95% confidence interval) within different risk factors for CKD.
reduction and would not receive an ineffective reduction in the dosage of drugs excreted by the kidneys.20 The authors of the CKD-EPI equation reported a lower bias of the estimates higher than 60 mL/min/1.73 m2; this could explain the rise in the proportion of normoalbuminuric people without renal impairment, which we found when comparing the CKD-EPI and MDRD equations. However, the authors also stated that the precision of the equation was limited, and this may explain the extensive overlap in the values of the estimates between normoalbuminuric people with and without risk factors for CKD observed in our study. One of the main reasons for using the CKD-EPI equation is to reduce the rate of false-positive diagnoses of stage 3 CKD obtained with the MDRD equation.20 This is a matter of special concern when screening people from the general population who lack markers of kidney damage.11 The MDRD in our study population produced very low estimates that fell below 60 mL/min/1.73 m2 in more than a third of the men and in more than two thirds of the women. Paradoxically, these proportions were even greater among normoalbuminuric women regardless of the presence or absence of risk factors for CKD. Although we did not perform a direct measurement of the GFR, these results strongly suggest that the MDRD equation underestimated the real GFR in the individuals without albuminuria. The significant reduction in the prevalence of renal impairment we obtained with the CKD-EPI equation supports this idea. Consequently, the results of our study argue against the use of the MDRD study equation as estimate of the GFR in the general population when there are no markers of kidney damage. At least in women, focusing screening for CKD on individuals with risk factors for CKD will not help clinicians interpret low estimates generated with the MDRD equation or avoid false-positive diagnoses. The new CKD-EPI equation increases the values of the estimates and decreases the rate of apparent renal impairment compared with the MDRD equation. However, some results in our study suggest that also there may be a particularly inaccurate correction factor for females in the CKD-EPI equation, and it is likely that the bias in this equation might not have improved in females to the same extent as in males. First, the prevalence of
renal impairment among normoalbuminuric women remained much greater than in men, despite the women’s better cardiovascular risk factor profile. Second, the rate of renal impairment of normoalbuminuric women with no CKD risk factors was very high and even greater than in men with diabetes, hypertension, or CVD. Finally, the estimates extensively overlapped between normoalbuminuric women with no risk factor and women with diabetes, hypertension, or CVD. The CKD-EPI equation shares another feature with the MDRD equation. In the absence of albuminuria no substantial difference in the cardiovascular risk factor profile was observed between the groups with an eGFR lower or greater than 60 mL/min/1.73 m2. The association between risk factors for CKD and renal impairment depends not only on the accuracy of the equations, but also on the cut-off points selected to identify low eGFR.26 An isolated mildly depressed eGFR may confer a low cardiovascular risk that can be indistinguishable from that of an eGFR just over 60 mL/min/1.73 m2 because the threshold of 60 mL/min/1.73 m2 has been arbitrarily defined as identifying stage 3 CKD.11,27,28 In our sample, the large number of normoalbuminuric participants with an eGFR 445 mL/min/1.73 m2 significantly accounted for the very high prevalence of renal impairment, but most of these subjects had a very low cardiovascular risk profile. Using the threshold of 60 mL/min/1.73 m2, 30 – 50% of older people could be misclassified as having renal impairment.29 To avoid this, it has been proposed that clinicians should focus on eGFR o45 mL/min/ 1.73 m2.30 The performance of the CKD-EPI equation for this lower threshold could not be analyzed in our cohort because of the low number of participants with this value. Our study has several limitations. First, the lack of a direct measurement of GFR, which would have allowed us to compare the accuracy of the different formulas, prevents us from making a statement about the accuracy of either formula in the general population. It was not within our scope to compare the diagnostic performance of the estimates against a direct measurement of GFR, but rather to compare their performance in an unselected sample from the general population with different risk factor profiles and to investigate the factors that could affect such performance. Second, we only made one determination of serum
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
creatinine and albuminuria, without further confirmation, and this might affect a rigorous classification of the participants into the categories of eGFR and albuminuria. However, we think that the impact of a single determination on our results is likely to be insignificant, considering that all samples were obtained in asymptomatic subjects in the early morning and given the relatively large sample size. Third, our population has one of the highest prevalences of obesity and diabetes in Spain.21 Our results may not be comparable with those of other populations with lower prevalence of obesity and diabetes. Lastly, in our crosssectional analysis we described a very poor association between prevalent cardiovascular risk factor profile and normoalbuminuric renal insufficiency. The renal and cardiovascular outcomes of normoalbuminuric individuals with low eGFR defined with the new CKD-EPI equation should be assessed in prospective studies. In conclusion, compared with the MDRD equation, using the CKD-EPI equation to estimate GFR generates a substantial reduction in the prevalence of renal impairment that has important practical consequences. Normoalbuminuric individuals, especially women, previously labeled as having stage 3 CKD with the MDRD equation would be the people most affected by the new estimations. The persistence of a disproportionately elevated prevalence of low eGFR among normoalbuminuric women, even without risk factors for CKD, suggests that an inaccurate correction factor for females may persist in the CKD-EPI equation. Conflict of interest The authors declare no conflicts of interest.
Acknowledgements We would like to thank the Doctor Pascual Foundation and the Investigacio´n Clı´nica Sanlu´car Association for providing financial support to our investigation. References 1. Zhang QL, Rothenbacher D. Prevalence of chronic kidney disease in population-based studies: systematic review. Br Med J Public Health. 2008;8: 117 – 30. 2. Sarnak MJ, Levey AS, Schoolwerth AC, Coresh J, Culleton B, Hamm LL, et al., American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease: a statement from the American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Circulation. 2003;108:2154 – 69. 3. Brosius FC, Hostetter TH, Kelepouris E, Mitsnefes MM, Moe SM, Moore MA, et al., American Heart Association Kidney and Cardiovascular Disease Council; Council on High Blood Pressure Research; Council on Cardiovascular Disease in the Young; Council on Epidemiology and Prevention; Quality of Care and Outcomes Research Interdisciplinary Working Group. Detection of chronic kidney disease in patients with or at increased risk of cardiovascular disease: a Science Advisory from the American Heart Association Kidney and Cardiovascular Disease Council; the Councils on High Blood Pressure Research, Cardiovascular Disease in the Youth, and Epidemiology and Prevention; and the Quality of Care and Outcomes Research Interdisciplinary Working Group: developed in collaboration with the National Kidney Foundation. Circulation. 2006;114: 1083 – 7. 4. Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function—measured and estimated glomerular filtration rate. N Engl J Med. 2006;354:2473 – 83.
623
5. Vassalotti JA, Stevens LA, Levey AS. Testing for chronic kidney disease: a position statement from the National Kidney Foundation. Am J Kidney Dis. 2007;50:169 – 80. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N. Roth D for the Modification on Diet in Renal Disease Study Group. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med. 1999;130:461 – 70. 7. Ruilope LM, Salvetti A, Jamerson K, Hansson L, Warnold I, Wedel H, et al. Renal function and intensive lowering of blood pressure in hypertensive participants of the Hypertension Optimal Treatment (HOT) Study. J Am Soc Nephrol. 2001;12:218 – 25. 8. Sarafidis PA, Bakris GL. Microalbuminuria and chronic kidney disease as risk factors for cardiovascular disease. Nephrol Dial Transplant. 2006;21: 2366 – 74. 9. Giles PD, Fitzmaurice DA. Formula estimation of glomerular filtration rate: have we gone wrong? Br Med J. 2007;334:1198 – 200. 10. Delanaye P, Cohen EP. Formula-based estimates of the GFR: equations variable and uncertain. Nephron Clin Pract. 2008;110:48 – 53. 11. Glassock RJ, Winearls C. An epidemic of chronic kidney disease: fact or fiction? Nephrol Dial Transplant. 2008;23:1117 – 21. 12. Coresh J, Stevens AL, Levey AS. Chronic kidney disease is common: what do we do next? Nephrol Dial Transplant. 2008;23:1122 – 5. 13. Feehally J, Griffit KE, Lamb EJ, O’Donoghe DJ, Tomson CR. Early detection of chronic kidney disease. Br Med J. 2008;337:845 – 7. 14. Stevens LA, Coresh J, Feldman HI, greene T, Lash JP, Nelson RG, et al. Evaluation of the modification of diet in renal disease study equation in a large diverse population. J Am Soc Nephrol. 2007;18:2749 – 57. 15. Poggio ED, Wang X, Greene T, Van Lente F, Hall PM. Performance of the Modification of Diet in Renal Disease and Cockcroft – Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol. 2005;16:459 – 66. 16. Froissart M, Rossert J, Jacquot C, Paillard M, Houillier P. Predictive performance of the Modification of Diet in Renal Disease and Cockcroft – Gault equations for estimating renal function. J Am Soc Nephrol. 2005;16:763 – 73. 17. Lin J, Knight E, Hogan MI, Singh AK. A comparison of prediction equations for estimating glomerular filtration rate in adults without kidney disease. J Am Soc Nephrol. 2003;14:2573 – 80. 18. Poggio ED, Rule AD. A critical evaluation of chronic kidney disease—should isolated reduced estimated glomerular filtration rate be considered a ‘disease’? Nephrol Dial Transplant. 2009;24:698 – 700. 19. Jaar BG, Khatib R, Plantinga L, Boulware LE, Powe NR. Principles of screening for chronic kidney disease. Clin J Am Soc Nephrol. 2008;3:601 – 9. 20. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, et al., for the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150: 604 – 12. 21. Lo´pez Sua´rez A, Elvira Gonza´lez J, Beltra´n Robles M, Alwakil M, Saucedo JM, ˜ ana Quirell A, et al. Prevalence of obesity, diabetes, hypertension, Bascun hypercholesterolemia and metabolic syndrome in over-50-year-old in Sanlu´car de Barrameda, Spain. Rev Esp Cardiol. 2008;61:1150 – 8. 22. Moler ER. Peripheral artery disease. Identification and implications. Arch Intern Med. 2003;163:2306 – 14. 23. Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, et al. Chronic Kidney Disease Epidemiology Collaboration: expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem. 2007;53:766 – 72. 24. Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, et al. Chronic Kidney Disease Epidemiology Collaboration: using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med. 2006;145:247 – 54. 25. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D. Predictors on newonset kidney disease in a community-based population. J Am Med Assoc. 2004;18:844 – 50. 26. Poggio ED, Rule AD. Can we do better than a single estimated GFR threshold when screening for chronic kidney disease? Kidney Int. 2007;72:534 – 6. 27. Hsu CY, Chertow GM. Chronic renal confusion: insufficiency, failure, dysfunction, or disease. Am J Kidney Dis. 2000;36:415 – 8. 28. Eckardt KU, Berns JS, Rocco MV, Kasiske BL. Definition and classification of CKD: the debate should be about patient prognosis—a position statement from KDOQI and KDIGO. Am J Kidney Dis. 2009;53:915 – 20. 29. Hallan S, Mutsert R, Carlsen S, Dekker FW, Aasarød K, Holmen J. Obesity, smoking, and physical inactivity as risk factor for CKD: are men more vulnerable? Am J Kidney Dis. 2006;47:396 – 405. 30. Jong PE, Gansevoort RT. Fact or fiction of the epidemic of chronic kidney disease—let us not squabble about estimated GFR only, but also focus on albuminuria. Nephrol Dial Transplant. 2008;23:1092 – 5.
www.elsevier.es/medicinaclinica
Original
Comparison of the MDRD and the CKD-EPI equations to estimate the glomerular filtration rate in the general population Alejandro Lo´pez-Sua´rez a,, Manuel Beltra´n-Robles a, Javier Elvira-Gonza´lez a, ˜ ana-Quirell a, Fernando Ferna´ndez-Palacı´n b,c, Antonio Bascun a Julio Benı´tez-Del-Castillo , Pablo Go´mez-Ferna´ndez d a
´diz, Spain ´ car de Barrameda, Ca Internal Medicine Department, Hospital Virgen del Camino, Sanlu ´diz, Ca ´diz, Spain Statistics and Operational Research Department, Universidad de Ca c ´diz, Spain ´ car de Barrameda, Ca Nephrology Department, Hospital Virgen del Camino, Sanlu d ´diz, Spain Nephrology Department, Hospital del SAS, Jerez de la Frontera, Ca b
ARTICLE INFO
A B S T R A C T
Article history: Received 11 June 2009 Accepted 30 July 2009 Available online 11 de marzo de 2010
Background and objective: The Chronic Kidney Disease Epidemiology Collaboration (CDK-EPI) equation has been proposed as a replacement for the Modification of Diet in Renal Disease (MDRD) equation to estimate the glomerular filtration rate, but this equation has not yet been evaluated in the general population. Patients and methods: Cross-sectional analysis of a random sample of 858 participants from the general population aged 50 – 75 years without known kidney disease. The prevalence of low eGFR (o 60 mL/min/ 1.73 m2) was assessed with the MDRD and the CKD-EPI equations in the overall sample and in normoalbuminuric individuals. Results: With the MDRD equation the median eGFRs (interquartile range) in men/women were 63.3(12.2)/ 56.7(9.4) mL/min/1.73 m2, and with the CKD-EPI equation 66.6(14.2)/61.3(11.6) mL/min/1.73 m2. The prevalence of low eGFR in men/women was 35.2%/68.5% and 25.1%/45.7% with the MDRD and the CKD-EPI equations, respectively. Normoalbuminuric women without risk factors for CKD experienced the most pronounced reduction in the number of cases with low eGFR with the CKD-EPI equation. The prevalence of renal impairment in this subgroup still remained even greater than that in men with diabetes, hypertension, or cardiovascular disease. Conclusions: Compared with the MDRD, the CKD-EPI equation generates a substantial reduction in the prevalence of renal impairment in subjects with diabetes, hypertension, cardiovascular disease, and in subjects without risk factors. The prevalence of renal impairment in normoalbuminuric females may be still overestimated with the CKD-EPI equation. ˜a, S.L. & 2010 Published by Elsevier Espan
Keywords: Albuminuria Cardiovascular risk factors Chronic kidney disease CKD-EPI equation Glomerular filtration rate MDRD
Comparacio´n de las ecuaciones MDRD y CKD-EPI en la estimacio´n de la tasa de filtrado glomerular en la poblacio´n general R E S U M E N
Palabras clave: Albuminuria factores de riesgo cardiovascular enfermedad renal cro´nica CKD-EPI filtrado glomerular MDRD
Fundamento y objetivo: La ecuacio´n Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) ha sido propuesta para sustituir a la actualmente recomendada Modification of Diet in Renal Disease (MDRD) para el ca´lculo de la tasa de filtrado glomerular (TFG), si bien no ha sido au´n evaluada en la poblacio´n general. Pacientes y Me´todo: Estudio transversal de una muestra aleatoria de 858 individuos de la poblacio´n ˜ os sin enfermedad renal conocida. Comparacio´n de la prevalencia de general con edades entre 50 – 75 an una TFGo 60 ml/min/1,73 m2 calculada con las ecuaciones MDRD y CKD-EPI en toda la muestra y en el subgrupo de individuos normoalbuminu´ricos (cociente albu´mina/creatinina o 30 mg/g). Resultados: Con la ecuacio´n MDRD la mediana de la TFG (rango intercuartı´lico) en varones/mujeres fue de 63,3 (12,2)/56,7 (9,4) ml/min/1,73 m2, y con la ecuacio´n CKD-EPI 66,6 (14,2)/61,3 (11,6) ml/min/1,73 m2, respectivamente. La prevalencia de una TFG baja en varones/mujeres fue del 35,2 – 68,5%, y del 25,1 – 45,7% con las ecuaciones MDRD y CKD-EPI, respectivamente. El grupo de mujeres normoalbuminu´ricas sin factores de riesgo para enfermedad renal cro´nica presento´ la mayor reduccio´n en la prevalencia de una TFG baja con la ecuacio´n CKD-EPI. Sin embargo, la prevalencia en este grupo se mantuvo incluso ma´s elevada que en el de los varones con diabetes, hipertensio´n o enfermedad cardiovascular.
DOI of original article: 10.1016/j.medcli.2009.09.051
Corresponding author.
E-mail address: [email protected] (A. Lo´pez-Sua´rez). ˜ a, S.L. 0025-7753/$ - see front matter & 2010 Published by Elsevier Espan doi:10.1016/j.medcli.2009.07.055
ARTICLE IN PRESS 618
´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
Conclusiones: Comparada con la ecuacio´n MDRD, la ecuacio´n CKD-EPI produce una importante reduccio´n en la prevalencia de insuficiencia renal en individuos con diabetes, hipertensio´n, enfermedad cardiovascular, y en individuos sin factores de riesgo. La prevalencia de insuficiencia renal en mujeres normoalbuminu´ricas calculada con la ecuacio´n CKD-EPI puede estar au´n sobreestimada. ˜ a, S.L. & 2010 Publicado por Elsevier Espan
Introduction The overall prevalence of chronic kidney disease (CKD) is estimated at up to 11% of adults in the United States and 6% in Europe.1 Patients with CKD are at risk for not only progression to end stage renal disease but, and far more frequently, for the development of cardiovascular disease (CVD).2 Therefore, early identification of CKD by reliable methods is an important part of preventive health strategies.3 The Modification of Diet in Renal Disease (MDRD) study equation has become the preferred method for estimating glomerular filtration rate (eGFR).4,5 This formula was initially developed in patients with known CKD, but has subsequently been widely used in the general population and in people with diverse cardiovascular risk profiles, even though there is no proof of its reliability as a screening test in people without known CKD.6 – 13 The MDRD equation provides unbiased and accurate estimates in a wide range of subgroups of patients when the eGFR is o60 mL/ min/1.73 m2.14 For estimates equal or greater than 60 mL/min/ 1.73 m2 the MDRD equation is not only less accurate but also less precise, especially in the absence of albuminuria and other markers of kidney damage.14 – 17 The MDRD has also been reported to be less precise in the absence of CKD than in cases with known disease, as eGFR values significantly overlap between patients with and without CKD and this compromises the use of eGFR to separate these two patients groups.9 However, clinicians continue using the MDRD equation for detecting renal impairment in everyday clinical practice, as measured GFR is not a suitable tool in this setting.18 In an attempt to overcome some of the drawbacks of the MDRD equation, it has been suggested that screening for CKD should be targeted at subjects at risk,19 in whom an eGFRo60 mL/min/1.73 m2 could be considered a true-positive result.4,5 However, the assumption that in the absence of markers of kidney damage a correct interpretation of low eGFR can be ascertained by taking into account the presence or absence of risk factors for CKD has not been evaluated thoroughly. Another way to improve the accuracy obtained with the MDRD equation is the development of new estimative equations. Levey et al.20 have recently proposed a new equation to replace the original MDRD equation for the routine clinical use in the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) study. But although the CKD-EPI equation has improved in bias, its precision remains suboptimal.20 Until now there is limited information about the performance of this new equation in people from the general population, especially in people whose renal function is unknown and who do not have markers of kidney damage. Our study compared the eGFR generated by the MDRD and the CKD-EPI equations in a random sample of individuals from the general population without known kidney disease. We also assessed whether there was a significant difference in the estimates of GFR among normoalbuminuric participants with and without risk factors for CKD.
Methods Study population This study was performed using the data from the Sanlu´car Study, the design of which is described elsewhere.21 It was a
cross-sectional population-based study carried out in Sanlu´car de Barrameda, Spain, between July 2006 and February 2007. The study population was a random sample of 858 participants aged 50 – 75 years obtained from the local census, all of Caucasian origin. The investigation was carried out in accordance with the principles of the Declaration of Helsinki. Subjects who agreed to participate were included after being interviewed and informed about the nature, risks, and benefits of the study in the clinic and giving written consent. Measurements All participants were studied using a standardized interview, physical examination, and blood and urine testing. Patients reporting a known diagnosis of renal failure were not included (three cases with end-stage renal disease and one needing hemodialysis). Blood pressure was recorded with a validated sphygmomanometer (OMRON 705CP, Hoofddorp, The Netherlands). A cuff with a 12 40 cm bladder was used in obese participants with an arm circumference 435 cm. Blood pressure was measured at the clinic with the subject seated after resting for 5 min. The selected systolic (SBP) and diastolic (DBP) blood pressures were the means of three consecutive measurements not differing by more than 10 mmHg in the SBP. Participants who were unaware of their blood pressure and whose office mean blood pressure was Z140/90 mmHg were offered 24 h ambulatory blood pressure monitoring (BR-102, Schiller AG, Baar, Switzerland) or 1 week of home blood pressure self-measurement. To identify peripheral arterial disease, all participants underwent ankle – brachial index examination, using a handleheld vascular Doppler with an 8 MHz probe (Mini Dop ES-100VX, Hayashi Denki Co., Ltd., Kawasaki, Japan) and a calibrated mercury sphygmomanometer. The procedure used to obtain the SBP in the upper and the lower extremities has been previously described.22 Laboratory tests and estimates of GFR Blood and urine samples were obtained from all participants in the early morning after overnight fasting. All participants were asymptomatic and without intercurrent illnesses at the time of testing. An ILAB-600 analyzer and standardized reagents (ITC Diagnostics, Izasa S.A., Barcelona, Spain) were used for measuring plasma glucose, total cholesterol, HDL-cholesterol, triglycerides, and glycosylated hemoglobin. The Friedewald formula was used for calculating LDL-cholesterol. Serum creatinine was measured using the Jaffe method with an alkaline picrate assay, giving a normal range up to 1.10 mg/dL in women and 1.30 mg/dL in men. This method was calibrated to be traceable to an isotope dilution mass spectrometry reference method.23 Albumin was measured in an early morning urine sample using a colorimetric method, and the results were expressed as milligrams of urinary albumin per gram of serum creatinine (ACR). Estimated GFR was calculated with the re-expressed MDRD equation for standardized serum creatinine (1)24 and the CKD-EPI equation (2)20: GFR ¼ 175 Sc1:154 age0:203 0:742 ½if female;
ð1Þ
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
if female and Sc r 0:70 mg=dL
GFR ¼ 144ðSc=0:7Þ0:329 ð0:993Þage ;
if female and Sc4 0:70 mg=dL
ð2Þ
GFR ¼ 144ðSc=0:7Þ1:209 ð0:993Þage ;
if male and Sc r0:90 mg=dL
GFR ¼ 141ðSc=0:9Þ0:411 ð0:993Þage ;
if male and Sc 40:90 mg=dL
GFR ¼ 141ðSc=0:9Þ1:209 ð0:993Þage ;
where age is expressed in years and standardized serum creatinine (Sc) in mg/dL. An eGFR value less than 60 mL/min/1.73 m2 was selected to define an abnormal GFR.4,5 Albuminuria was defined as an ACRZ30 mg/g. Risk factors for CKD Type 2 diabetes, hypertension, and cardiovascular disease were the risk factors for CKD selected for the purposes of the study. Diabetes was diagnosed in participants being treated for diabetes and in those with fasting serum glucose Z7.0 mmol/L. Hypertension was defined as already being treated with antihypertensive drugs or having a mean blood pressure of Z135/85 mmHg following 24-h ambulatory blood pressure monitoring or 1 week of home blood pressure self-measurement. Patients with CVD were those with a documented diagnosis of a transient ischemic attack, ischemic stroke, stable angina, acute coronary syndrome, coronary revascularization, or peripheral arterial disease defined as an ankle – brachial index o0.90. Other cardiovascular and CKD risk factors were also considered. Subjects were classified as smokers (current smokers or smokers in the last year) or non-smokers (participants who never smoked or who stopped smoking a year ago). The Framingham risk score was used as an integrated tool of traditional cardiovascular risk factors.25 As a reference group for comparing the performance of eGFR in subjects with risk factors for CKD, we defined a group of participants with no risk factors as the non-smoking, nondiabetic, and non-hypertensive participants with no history of CVD and a Framingham risk score o10%. Statistical analysis Results are expressed as mean (95% confidence interval) for normally distributed continuous data, median (interquartile range) for continuous data without normal distribution, and frequencies (95% confidence intervals) for categorical data. Differences in the risk factor profile were compared between albuminuric and normoalbuminuric participants by categorizing
619
an eGFR lower vs. greater than 60 mL/min/1.73 m2 using the chi-square test for categorical variables, and ANOVA and Kruskal – Wallis test for normally and non-normally distributed continuous data respectively. A P-valueo0.05 was accepted as being statistically significant. SPSS for Windows, version 12.0 (Chicago, Illinois, USA) was used for the statistical analyses.
Results GFR calculated with the MDRD equation The general characteristics of the study sample are summarized in Table 1. The overall prevalence of renal impairment was 53%. Among them, 92.1% had an eGFR of 45 – 60 mL/min/1.73 m2. There was a marked difference in the median eGFR and the prevalence of low eGFR by gender (Table 2). An important proportion of all cases with renal impairment was normoalbuminuric women, in whom the percentage of low eGFR was as high as in women with an ACRZ30 mg/g (68.8% vs. 66.1%, respectively). Only 15.8% of cases with an eGFRo60 ml/min/1.73 m2 also had an ACRZ30 mg/g, whereas up to 58.5% of participants with albuminuria also had a low eGFR. The presence of albuminuria was associated with the most adverse risk factor profile regardless of the eGFR level. However, in the absence of albuminuria, no substantial difference in the risk factor profile was observed when comparing the groups with an eGFRZ or o60 mL/min/1.73 m2 (Tables 3 and 4). The values of the eGFR in the 735 normoalbuminuric participants, whether greater or lower than 60 mL/min/1.73 m2, closely overlapped between individuals with and without risk factors for CKD (Figs. 1 and 2). In the subgroup of normoalbuminuric women with diabetes, hypertension, or cardiovascular disease, the range of eGFR values and the proportion of cases with an eGFRo60 mL/min/1.73 m2 were very similar to those in women with no risk factors (Fig. 3). GFR calculated with the CKD-EPI equation Compared with the MDRD, the most noteworthy result obtained with the CKD-EPI equation was a 16.9% absolute reduction of participants with an eGFR o60 mL/min/1.73 m2, a reduction that was greater in women than in men (22.8% vs. 10.1%). Overall, in 91.6% of cases with renal impairment, the estimates fell in the range of 45 – 60 mL/min/1.73 m2. The subgroup of participants with the most pronounced reduction in the percentage of cases with low eGFR was the normoalbuminuric women with no CKD risk factors (29.8% absolute reduction).
Table 1 General characteristics of the study sample
Age (years) Smokers Diabetes Hypertension Body mass index Z30 kg/m2 Framingham risk score* o10% 420% Cardiovascular disease Participants with no risk factors Serum creatinine (mg/100 mL)
All (n= 858)
Women (n =460)
Men (n= 398)
61.5 15.1 27.6 49.0 55.1
(61.0 – 61.9) (12.8 – 17.7) (24.7 – 30.7) (45.6 – 52.4) (51.7 – 58.5)
61.1 7.9 26.1 52.4 56.1
(60.5 – 61.7) (5.5 – 10.7) (22.1 – 30.4) (47.7 – 57.0) (51.4 – 60.7)
61.9 23.4 29.4 45.0 54.0
(61.2 – 62.5) (19.3 – 27.8) (25.0 – 34.1) (40.0 – 50.0) (49.0 – 59.0)
59.6 11.2 15.6 24.6 1.09
(56.2 – 62.9) (9.2 – 13.4) (13.3 – 18.2) (21.7 – 27.6) (1.08 – 1.10)
90.7 1.5 13.3 34.3 1.01
(87.6 – 93.2) (0.6 – 3.1) (10.3 – 16.7) (30.0 – 38.9) (1.00 – 1.02)
23.6 22.4 18.3 13.3 1.18
(19.5 – 28.1) (18.4 – 26.8) (14.7 – 22.5) (10.1 – 17.1) (1.16 – 1.20)
Data are mean (95% confidence interval) or percentages (95% confidence interval).
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
620
Table 2 Mean eGFR and percentage of low eGFR Women (n= 460)
Mean eGFRa Low eGFRb Low eGFR-ACRo 30c Low eGFR-ACRZ 30 a b c
Men (n= 398)
MDRD
CKD-EPI
MDRD
CKD-EPI
56.8 68.5 11.7 88.3
61.4 45.7 13.3 86.7
64.4 35.2 24.3 75.7
67.4 25.1 30.0 70.0
(56.0 – 57.5) (64.0 – 72.7) (8.4 – 15.8) (84.2 – 91.6)
(60.6 – 62.3) (41.0 – 50.3) (9.0 – 18.7) (81.3 – 91.0)
(63.2 – 65.6) (30.5 – 40.1) (17.4 – 32.2) (67.8 – 82.6)
(66.2 – 68.5) (20.9 – 29.7) (21.2 – 40.0) (60.0 – 78.8)
Mean (95% confidence interval), mL/min/1.73 m2. eGFRo 60 ml/min/1.73 m2 (percentage [95% confidence interval]). ACR (mg/g).
Table 3 Risk factor profile according to the albumin/creatinine ratio and eGFR categories (MDRD equation) mL/min/1.73 m2
n (%) Age (years) Gender (female) Glucose (mg/dL) Haemoglobin A1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) Body mass index (kg/m2) Waist circumference (cm) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) Triglycerides (mg/dL)a Framingham risk score o 10% 4 20%
ACRo 30 mg/g
ACR Z30 mg/g
eGFRZ 60
eGFRo 60
eGFR Z60
eGFRo 60
351 60.0 36.2 110 6.0 143 79 30.4 105 225 56 142 116
384 62.0 72.4 112 6.0 142 80 31.0 103 238 60 150 124
51 63.1 37.3 121 6.5 153 81 31.3 108 229 55 145 129
72 64.6 52.1 122 6.5 157 84 32.0 107 233 53 147 150
(40.9) (59.4 – 60.7) (31.1 – 41.5) (107 – 114) (5.9 – 6.1) (141 – 145) (78 – 81) (29.9 – 30.9) (104 – 106) (221 – 230) (55 – 57) (138 – 146) (83)
52.3 (47.1 – 57.7) 9.2 (6.3 – 12.6)
(44.7) (61.3 – 62.6) (67.6 – 76.8) (108 – 115) (5.9 – 6.1) (140 – 144) (79 – 81) (30.5 – 31.5) (102 – 104) (234 – 242) (58 – 61) (146 – 154) (83)
72.1 (67.4 – 76.6) 8.7 (6.0 – 11.9)
(6.0) (61.0 – 65.3) (24.1 – 51.9) (106 – 137) (6.1 – 6.9) (148 – 158) (78 – 84) (30.0 – 32.7) (103 – 112) (215 – 243) (51 – 59) (133 – 157) (116)
37.3 (24.1 – 51.9) 21.6 (11.3 – 35.3)
(8.4) (63.2 – 66.4) (40.7 – 64.7) (111 – 133) (6.1 – 6.8) (151 – 163) (81 – 87) (30.7 – 33.2) (104 – 110) (222 – 244) (50 – 56) (137 – 156) (106)
43.7 (31.4 – 55.3) 26.8 (16.7 – 38.1)
Data are percentages (95% confidence interval) or mean (95% confidence interval). Po 0.001. P o 0.01. Po 0.05. a Median (interquartile range).
Table 4 Risk factor profile according to the albumin/creatinine ratio and eGFR categories (CKD-EPI) mL/min/1.73 m2
n (%) Age (years) Gender (female) Glucose (mg/dL) Haemoglobin A1C (%) Systolic BP (mmHg) Diastolic BP (mmHg) Body mass index (kg/m2) Waist circumference (cm) Total cholesterol (mg/dL) HDL cholesterol (mg/dL) LDL cholesterol (mg/dL) Triglycerides (mg/dL)a Framingham risk score o 10% 4 20%
ACRo 30 mg/g
ACR Z30 mg/g
eGFRZ 60
eGFRo 60
eGFR Z60
eGFRo 60
484 60.0 46.3 111 6.0 143 80 30.4 104 229 57 145 116
252 63.7 72.2 111 6.0 142 79 31.3 104 237 59 149 127
64 62.2 43.8 120 6.5 154 82 31.8 108 234 55 149 133
58 66.2 48.3 125 6.4 157 84 31.6 106 229 53 142 158
(56.4) (59.1 – 60.2) (40.5 – 52.1) (108 – 114) (5.9 – 6.1) (141 – 145) (79 – 81) (30.0 – 30.9) (103 – 105) (225 – 233) (56 – 59) (141 – 148) (80)
58.5 (52.6 – 64.1) 8.8 (5.9 – 12.6)
(29.3) (62.9 – 64.6) (64.6 – 79.0) (107 – 116) (5.9 – 6.1) (140 – 144) (78 – 80) (30.7 – 31.9) (102 – 105) (232 – 243) (57 – 61) (145 – 154) (82)
70.6 (64.6 – 76.2) 9.3 (5.9 – 13.4)
Data are percentages (95% confidence interval) or mean (95% confidence interval). Po 0.001. P o 0.01. Po 0.05. a Median (interquartile range).
(7.5) (60.4 – 64.4) (31.4 – 56.7) (107 – 132) (6.2 – 6.9) (149 – 159) (79 – 85) (30.5 – 33.1) (105 – 112) (221 – 247) (51 – 58) (139 – 160) (104)
42.2 (29.9 – 55.2) 18.8 (10.1 – 30.5)
(6.8) (64.5 – 67.9) (35.0 – 61.8) (112 – 138) (6.0 – 6.8) (150 – 163) (81 – 87) (30.3 – 33.0) (103 – 110) (218 – 241) (49 – 56) (132 – 153) (111)
39.7 (27.0 – 53.4) 31.0 (19.5 – 44.5)
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
621
Normoalbuminuric men 120
*
120
MDRD equation
100 2
eGFR (ml/min/1,73 m )
eGFR (ml/min/1,73 m2)
100
CKD-EPI equation
80
60
80
60
40
40
20
20 No risk factors
Diabetes
Hypertension
CVD
No risk factors
Diabetes
Hypertension
CVD
Fig. 1. Box plots comparing eGFR within different risk factors for CKD in normoalbuminuric men.
Normoalbuminuric women 120
120
MDRD equation
100 eGFR (ml/min/1,73 m2)
eGFR (ml/min/1,73 m2)
100
CKD-EPI equation
80
60
40
20
80
60
40
20 No risk factors Diabetes
Hypertension
CVD
No risk factors Diabetes
Hypertension
CVD
Fig. 2. Box plots comparing eGFR within different risk factors for CKD in normoalbuminuric women.
However, the prevalence of a low eGFR in this subgroup remained almost double that of men, irrespective of the presence or absence of risk factors for CKD (Fig. 3). Unlike with the MDRD equation, with the CKD-EPI equation normoalbuminuric women with diabetes, hypertension, or CVD had significantly lower levels of eGFR than women with no CKD risk factors (Fig. 3). However, the range of values of eGFR also overlapped between normoalbuminuric participants (men or women) with and without risk factors for CKD (Figs. 1 and 2). In the absence of albuminuria, no significant difference was observed between participants with an eGFRZ and o60 mL/min/1.73 m2 in systolic or diastolic blood pressure, body mass index, waist circumference, serum glucose, or hemoglobin A1C. The level of HDL-cholesterol and the percentage of cases with a Framingham risk score o10% were even greater among normoalbuminuric participants with low eGFR than among those with an eGFRZ60 mL/min/1.73 m2.
Discussion The main finding of this study is that the CKD-EPI equation, applied to the general population, generates higher eGFR values than the MDRD equation, decreasing the overall prevalence of low eGFR by 16.9% (36.1% vs. 53.0%). Normoalbuminuric women with no risk factors were the subgroup with the greatest reduction in the prevalence of renal impairment. However, the prevalence of apparent renal impairment in this subgroup still remained very high, especially in women with no CKD risk factors, in whom the rate of low eGFR was even higher than in normoalbuminuric men with any risk factor for CKD. The first practical consequence derived using the new equation is likely to be a significant reduction in the number of people previously labeled as having stage 3 CKD. These people would not undergo further nephrological evaluation or overaggressive cardiovascular risk factor
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
622
Normoalbuminuric men
Normoalbuminuric women
90
80
80
70
70
60
60 Porcentaje
Porcentaje
90
50 40
50 40
30
30
20
20
10
10
0
0 No risk factors
Diabetes
Hypertension
CVD
eGFR < 60 ml/min/1,73 m2 (MDRD)
No risk factors
Diabetes
Hypertension
CVD
eGFR < 60 ml/min/1,73 m2 (CKD-EPI)
Fig. 3. Percentage of low eGFR (95% confidence interval) within different risk factors for CKD.
reduction and would not receive an ineffective reduction in the dosage of drugs excreted by the kidneys.20 The authors of the CKD-EPI equation reported a lower bias of the estimates higher than 60 mL/min/1.73 m2; this could explain the rise in the proportion of normoalbuminuric people without renal impairment, which we found when comparing the CKD-EPI and MDRD equations. However, the authors also stated that the precision of the equation was limited, and this may explain the extensive overlap in the values of the estimates between normoalbuminuric people with and without risk factors for CKD observed in our study. One of the main reasons for using the CKD-EPI equation is to reduce the rate of false-positive diagnoses of stage 3 CKD obtained with the MDRD equation.20 This is a matter of special concern when screening people from the general population who lack markers of kidney damage.11 The MDRD in our study population produced very low estimates that fell below 60 mL/min/1.73 m2 in more than a third of the men and in more than two thirds of the women. Paradoxically, these proportions were even greater among normoalbuminuric women regardless of the presence or absence of risk factors for CKD. Although we did not perform a direct measurement of the GFR, these results strongly suggest that the MDRD equation underestimated the real GFR in the individuals without albuminuria. The significant reduction in the prevalence of renal impairment we obtained with the CKD-EPI equation supports this idea. Consequently, the results of our study argue against the use of the MDRD study equation as estimate of the GFR in the general population when there are no markers of kidney damage. At least in women, focusing screening for CKD on individuals with risk factors for CKD will not help clinicians interpret low estimates generated with the MDRD equation or avoid false-positive diagnoses. The new CKD-EPI equation increases the values of the estimates and decreases the rate of apparent renal impairment compared with the MDRD equation. However, some results in our study suggest that also there may be a particularly inaccurate correction factor for females in the CKD-EPI equation, and it is likely that the bias in this equation might not have improved in females to the same extent as in males. First, the prevalence of
renal impairment among normoalbuminuric women remained much greater than in men, despite the women’s better cardiovascular risk factor profile. Second, the rate of renal impairment of normoalbuminuric women with no CKD risk factors was very high and even greater than in men with diabetes, hypertension, or CVD. Finally, the estimates extensively overlapped between normoalbuminuric women with no risk factor and women with diabetes, hypertension, or CVD. The CKD-EPI equation shares another feature with the MDRD equation. In the absence of albuminuria no substantial difference in the cardiovascular risk factor profile was observed between the groups with an eGFR lower or greater than 60 mL/min/1.73 m2. The association between risk factors for CKD and renal impairment depends not only on the accuracy of the equations, but also on the cut-off points selected to identify low eGFR.26 An isolated mildly depressed eGFR may confer a low cardiovascular risk that can be indistinguishable from that of an eGFR just over 60 mL/min/1.73 m2 because the threshold of 60 mL/min/1.73 m2 has been arbitrarily defined as identifying stage 3 CKD.11,27,28 In our sample, the large number of normoalbuminuric participants with an eGFR 445 mL/min/1.73 m2 significantly accounted for the very high prevalence of renal impairment, but most of these subjects had a very low cardiovascular risk profile. Using the threshold of 60 mL/min/1.73 m2, 30 – 50% of older people could be misclassified as having renal impairment.29 To avoid this, it has been proposed that clinicians should focus on eGFR o45 mL/min/ 1.73 m2.30 The performance of the CKD-EPI equation for this lower threshold could not be analyzed in our cohort because of the low number of participants with this value. Our study has several limitations. First, the lack of a direct measurement of GFR, which would have allowed us to compare the accuracy of the different formulas, prevents us from making a statement about the accuracy of either formula in the general population. It was not within our scope to compare the diagnostic performance of the estimates against a direct measurement of GFR, but rather to compare their performance in an unselected sample from the general population with different risk factor profiles and to investigate the factors that could affect such performance. Second, we only made one determination of serum
ARTICLE IN PRESS ´rez et al / Med Clin (Barc). 2010;134(14):617 –623 ´pez-Sua A. Lo
creatinine and albuminuria, without further confirmation, and this might affect a rigorous classification of the participants into the categories of eGFR and albuminuria. However, we think that the impact of a single determination on our results is likely to be insignificant, considering that all samples were obtained in asymptomatic subjects in the early morning and given the relatively large sample size. Third, our population has one of the highest prevalences of obesity and diabetes in Spain.21 Our results may not be comparable with those of other populations with lower prevalence of obesity and diabetes. Lastly, in our crosssectional analysis we described a very poor association between prevalent cardiovascular risk factor profile and normoalbuminuric renal insufficiency. The renal and cardiovascular outcomes of normoalbuminuric individuals with low eGFR defined with the new CKD-EPI equation should be assessed in prospective studies. In conclusion, compared with the MDRD equation, using the CKD-EPI equation to estimate GFR generates a substantial reduction in the prevalence of renal impairment that has important practical consequences. Normoalbuminuric individuals, especially women, previously labeled as having stage 3 CKD with the MDRD equation would be the people most affected by the new estimations. The persistence of a disproportionately elevated prevalence of low eGFR among normoalbuminuric women, even without risk factors for CKD, suggests that an inaccurate correction factor for females may persist in the CKD-EPI equation. Conflict of interest The authors declare no conflicts of interest.
Acknowledgements We would like to thank the Doctor Pascual Foundation and the Investigacio´n Clı´nica Sanlu´car Association for providing financial support to our investigation. References 1. Zhang QL, Rothenbacher D. Prevalence of chronic kidney disease in population-based studies: systematic review. Br Med J Public Health. 2008;8: 117 – 30. 2. Sarnak MJ, Levey AS, Schoolwerth AC, Coresh J, Culleton B, Hamm LL, et al., American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Kidney disease as a risk factor for development of cardiovascular disease: a statement from the American Heart Association Councils on Kidney in Cardiovascular Disease, High Blood Pressure Research, Clinical Cardiology, and Epidemiology and Prevention. Circulation. 2003;108:2154 – 69. 3. Brosius FC, Hostetter TH, Kelepouris E, Mitsnefes MM, Moe SM, Moore MA, et al., American Heart Association Kidney and Cardiovascular Disease Council; Council on High Blood Pressure Research; Council on Cardiovascular Disease in the Young; Council on Epidemiology and Prevention; Quality of Care and Outcomes Research Interdisciplinary Working Group. Detection of chronic kidney disease in patients with or at increased risk of cardiovascular disease: a Science Advisory from the American Heart Association Kidney and Cardiovascular Disease Council; the Councils on High Blood Pressure Research, Cardiovascular Disease in the Youth, and Epidemiology and Prevention; and the Quality of Care and Outcomes Research Interdisciplinary Working Group: developed in collaboration with the National Kidney Foundation. Circulation. 2006;114: 1083 – 7. 4. Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function—measured and estimated glomerular filtration rate. N Engl J Med. 2006;354:2473 – 83.
623
5. Vassalotti JA, Stevens LA, Levey AS. Testing for chronic kidney disease: a position statement from the National Kidney Foundation. Am J Kidney Dis. 2007;50:169 – 80. 6. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N. Roth D for the Modification on Diet in Renal Disease Study Group. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Ann Intern Med. 1999;130:461 – 70. 7. Ruilope LM, Salvetti A, Jamerson K, Hansson L, Warnold I, Wedel H, et al. Renal function and intensive lowering of blood pressure in hypertensive participants of the Hypertension Optimal Treatment (HOT) Study. J Am Soc Nephrol. 2001;12:218 – 25. 8. Sarafidis PA, Bakris GL. Microalbuminuria and chronic kidney disease as risk factors for cardiovascular disease. Nephrol Dial Transplant. 2006;21: 2366 – 74. 9. Giles PD, Fitzmaurice DA. Formula estimation of glomerular filtration rate: have we gone wrong? Br Med J. 2007;334:1198 – 200. 10. Delanaye P, Cohen EP. Formula-based estimates of the GFR: equations variable and uncertain. Nephron Clin Pract. 2008;110:48 – 53. 11. Glassock RJ, Winearls C. An epidemic of chronic kidney disease: fact or fiction? Nephrol Dial Transplant. 2008;23:1117 – 21. 12. Coresh J, Stevens AL, Levey AS. Chronic kidney disease is common: what do we do next? Nephrol Dial Transplant. 2008;23:1122 – 5. 13. Feehally J, Griffit KE, Lamb EJ, O’Donoghe DJ, Tomson CR. Early detection of chronic kidney disease. Br Med J. 2008;337:845 – 7. 14. Stevens LA, Coresh J, Feldman HI, greene T, Lash JP, Nelson RG, et al. Evaluation of the modification of diet in renal disease study equation in a large diverse population. J Am Soc Nephrol. 2007;18:2749 – 57. 15. Poggio ED, Wang X, Greene T, Van Lente F, Hall PM. Performance of the Modification of Diet in Renal Disease and Cockcroft – Gault equations in the estimation of GFR in health and in chronic kidney disease. J Am Soc Nephrol. 2005;16:459 – 66. 16. Froissart M, Rossert J, Jacquot C, Paillard M, Houillier P. Predictive performance of the Modification of Diet in Renal Disease and Cockcroft – Gault equations for estimating renal function. J Am Soc Nephrol. 2005;16:763 – 73. 17. Lin J, Knight E, Hogan MI, Singh AK. A comparison of prediction equations for estimating glomerular filtration rate in adults without kidney disease. J Am Soc Nephrol. 2003;14:2573 – 80. 18. Poggio ED, Rule AD. A critical evaluation of chronic kidney disease—should isolated reduced estimated glomerular filtration rate be considered a ‘disease’? Nephrol Dial Transplant. 2009;24:698 – 700. 19. Jaar BG, Khatib R, Plantinga L, Boulware LE, Powe NR. Principles of screening for chronic kidney disease. Clin J Am Soc Nephrol. 2008;3:601 – 9. 20. Levey AS, Stevens LA, Schmid CH, Zhang YL, Castro AF, Feldman HI, et al., for the CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150: 604 – 12. 21. Lo´pez Sua´rez A, Elvira Gonza´lez J, Beltra´n Robles M, Alwakil M, Saucedo JM, ˜ ana Quirell A, et al. Prevalence of obesity, diabetes, hypertension, Bascun hypercholesterolemia and metabolic syndrome in over-50-year-old in Sanlu´car de Barrameda, Spain. Rev Esp Cardiol. 2008;61:1150 – 8. 22. Moler ER. Peripheral artery disease. Identification and implications. Arch Intern Med. 2003;163:2306 – 14. 23. Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, et al. Chronic Kidney Disease Epidemiology Collaboration: expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem. 2007;53:766 – 72. 24. Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, et al. Chronic Kidney Disease Epidemiology Collaboration: using standardized serum creatinine values in the modification of diet in renal disease study equation for estimating glomerular filtration rate. Ann Intern Med. 2006;145:247 – 54. 25. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D. Predictors on newonset kidney disease in a community-based population. J Am Med Assoc. 2004;18:844 – 50. 26. Poggio ED, Rule AD. Can we do better than a single estimated GFR threshold when screening for chronic kidney disease? Kidney Int. 2007;72:534 – 6. 27. Hsu CY, Chertow GM. Chronic renal confusion: insufficiency, failure, dysfunction, or disease. Am J Kidney Dis. 2000;36:415 – 8. 28. Eckardt KU, Berns JS, Rocco MV, Kasiske BL. Definition and classification of CKD: the debate should be about patient prognosis—a position statement from KDOQI and KDIGO. Am J Kidney Dis. 2009;53:915 – 20. 29. Hallan S, Mutsert R, Carlsen S, Dekker FW, Aasarød K, Holmen J. Obesity, smoking, and physical inactivity as risk factor for CKD: are men more vulnerable? Am J Kidney Dis. 2006;47:396 – 405. 30. Jong PE, Gansevoort RT. Fact or fiction of the epidemic of chronic kidney disease—let us not squabble about estimated GFR only, but also focus on albuminuria. Nephrol Dial Transplant. 2008;23:1092 – 5.