Chronic kidney disease (CKD) reportedly affects

AJH 2006; 19:608 614 Kidney A Comparison of Serum Creatinine Based Methods for Identifying Chronic Kidney Disease in Hypertensive Individuals and Their Siblings Andrew D. Rule, Steven J. Jacobsen, Gary L. Schwartz, Thomas H. Mosley, Christopher G. Scott, Sharon L. R. Kardia, Eric Boerwinkle, and Stephen T. Turner Background: The Modification of Diet in Renal Disease (MDRD) equation is often used to determine an estimated glomerular filtration rate (egfr) from serum creatinine. This study compared kidney disease as defined by reduced egfr, elevated serum creatinine, or elevated urinary albumin-to-creatinine ratio (ACR). Methods: As part of the Genetic Epidemiology Network of Arteriopathy study, a community-based sample was ascertained through sibships having at least two members with essential hypertension. Kidney disease was defined by reduced egfr ( 60 ml/min/1.73 m 2 ), elevated serum creatinine ( 97.5 th percentile for sex-specific normal individuals), or elevated ACR ( 95 th percentile for sex-specific normal individuals). Results: The sample (n 2653) was 65% female, 61% African American, and 77% hypertensive, with a mean ( SD) age of 61 10 years. There was greater agreement between kidney disease defined by elevated ACR and an elevated serum creatinine level ( 0.19) than between kidney disease defined by elevated ACR and a reduced egfr ( 0.07). The multivariable-adjusted odds ratio of kidney disease for male versus female sex was 0.92 (95% CI, 0.75 to 1.12) by reduced egfr, but was 2.08 (95% CI, 1.62 to 2.67) by elevated serum creatinine and 2.11 (95% CI, 1.63 to 2.74) by elevated ACR. The multivariable-adjusted odds ratio of kidney disease for subjects of African American versus white ethnicity was 0.27 (95% CI, 0.22 to 0.33) by reduced egfr but was 1.17 (95% CI, 0.91 to 1.51) by elevated serum creatinine and 3.87 (95% CI, 2.89 to 5.25) by elevated ACR. Conclusion: In a predominantly hypertensive population, kidney disease identified by elevated ACR was more concordant with elevated serum creatinine than with reduced egfr. The MDRD equation, derived using kidney disease patients, may misrepresent the gender- and ethnicity-specific risk of kidney disease. Am J Hypertens 2006; 19:608 614 2006 American Journal of Hypertension, Ltd. Key Words: Hypertension, chronic kidney failure, creatinine, albuminuria, glomerular filtration rate. Chronic kidney disease (CKD) reportedly affects 11% of the adult general population. About 7% of CKD has been identified by an elevated urinary albumin-to-creatinine ratio (ACR), and the remaining 4% has been identified by a reduced estimated glomerular filtration rate (egfr). 1 Commonly GFR is estimated by the Modification of Diet in Renal Disease study (MDRD) equation that uses serum creatinine, age, sex, and ethnicity as predictor variables. 2,3 However, this equation was developed in subjects with a pre-existing diagnosis of kidney disease. Several studies have shown that the MDRD equation does not accurately model the relationship between Received July 12, 2005. First decision October 10, 2005. Accepted October 15, 2005. From the Division of Nephrology and Hypertension (ADR, GLS, STT), Department of Internal Medicine; Division of Epidemiology (ADR, SJJ), Department of Health Science Research, and the Division of Biostatistics (CGS), Mayo Clinic, Rochester, Minnesota; Department of Geriatric Medicine (THM), University of Mississippi Medical Center/ VAMC Consortium, Jackson, Mississippi; Department of Epidemiology (SLRK), University of Michigan, Ann Arbor, Michigan; and Human Genetics Center and Institute of Molecular Medicine (EB), University of Texas-Houston Health Science Center, Houston, Texas. This work was supported by United States Public Health Service Grants from the National Institutes of Health F32 DK 68996, U01 HL 54464, U01 HL 54457, U01 HL 54463, U01 HL 54481, and R01 AR 30582. Address correspondence and reprint requests to Dr. Stephen T. Turner, Mayo Clinic, 200 First Street SW, Rochester, MN, 55905; e-mail: turner. stephen@mayo.edu 0895-7061/06/$32.00 2006 by the American Journal of Hypertension, Ltd. Downloaded from doi:10.1016/j.amjhyper.2005.10.025 https://academic.oup.com/ajh/article-abstract/19/6/608/228786 Published by Elsevier Inc.

AJH June 2006 VOL. 19, NO. 6 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE 609 these variables and GFR for persons without known kidney disease. 4 7 Essential or primary hypertension affects about 29% of the adult general population. 8 Although the cause-andeffect relationship between hypertension and CKD is debated, 9 hypertension has been consistently associated with early CKD 10 12 and end-stage renal disease. 13 Many studies 1,11,12,14,15 have shown an association between hypertension and a reduced egfr 60 ml/min/1.73 m 2. However, in populations of individuals among whom the diagnosis of kidney disease has not been established, it is important to consider the potential for misclassification with identification of CKD by egfr. Study of a predominantly hypertensive sample at increased risk for kidney disease facilitates the comparison between different definitions of CKD. There are two common ways serum creatinine has been used to identify CKD. The first uses linear regression to convert serum creatinine into an estimated GFR and defines CKD 16 by values less than 60 ml/min/1.73 m 2, the lower limit of normal (2.5 th percentile) for measured GFR in elderly individuals. 17 The second uses a normal value study of healthy volunteers and defines CKD by serum creatinine values above the upper limit of normal (97.5 th percentile). The objective of this study was to compare the prevalence of CKD as defined by these two serum creatinine based methods with elevated ACR (albuminuria). Furthermore, associations with predictors of CKD (eg, sex and ethnicity) were compared between these different definitions. Methods Study Population Individuals of non-hispanic African American ethnicity were recruited from Jackson, MS, and individuals of non- Hispanic white ethnicity were recruited from Rochester, MN, to participate in the Genetic Epidemiology Network of Arteriopathy (GENOA) study within the Family Blood Pressure Program (FBPP). 18 The GENOA subjects were ascertained from the community through sibships (ie, full biological offspring of the same mother and father) having two or more members with essential hypertension diagnosed before age 60 years as previously detailed. 19,20 Proband candidates with advanced CKD or end-stage renal disease per the medical record were excluded for having a potential cause of secondary hypertension. Normotensive siblings of the probands were also included in the study sample. Under institutional review board approval, subjects participated in a baseline visit (1997 to 2000) and a follow-up visit (2001 to 2004) a mean of 4.5 years later. Of 3471 participants in the baseline visit, 2653 (76%) returned for the follow-up visit. Subjects were more likely to return for the follow-up visit if they were older, female, African American, and were without hypertension or diabetes mellitus at baseline (P.05 for each in a multivariable model). Analyses were limited to the follow-up visit, as urine ACR was not measured at the baseline visit. Measurements The study visit consisted of a standardized medical history questionnaire administered by a trained examiner. Height, weight, and blood pressure were also measured by a standardized protocol at each site. 20 Fasting blood samples were drawn for measurements of glucose, creatinine, cholesterol, and high-sensitivity C-reactive protein (CRP). A spot urine sample was used to measure the ACR. The diagnosis of hypertension was based on a selfreported diagnosis, current use of antihypertensive medications, elevated blood pressure 140/90 mm Hg, and review of the medical record by a hypertension specialist. A diagnosis of diabetes mellitus was based on the subject s self-reported diagnosis, use of prescription antidiabetic medications, or fasting plasma glucose 7.0 mmol/l (126 mg/dl). Hyperlipidemia was defined by a total cholesterol 5.18 mmol/l (200 mg/dl). Obesity was defined as a body mass index (BMI, defined as weight/height 2 ) 30 kg/m 2. Inflammation was defined by a CRP 75 nmol/l (1.0 mg/l). Chronic kidney disease was defined using three different methods: 1) reduced estimated GFR, 2) elevated serum creatinine, and 3) elevated ACR. The ACR was dichotomized to define CKD using sex-specific thresholds representing the 95 th percentile for healthy men ( 17 mg/g) and women ( 25 mg/g), which estimate an albumin excretion rate 21,22 of 30 g/min. The GFR was estimated using the MDRD equation: egfr 186.3 (serum creatinine) 1.154 (age) 203 (0.742 if female) (1.212 if African American). To calibrate with the Cleveland Clinic assay used to develop the MDRD equation, 23 0.22 mg/dl (17 mol/l) was added to all serum creatinine values. This adjustment for calibration bias was based on 255 serum creatinine samples that were measured both with the GENOA examination rate Jaffe assay (Hitachi 911, Roche Diagnositcs Corp., Indianapolis, IN) and with the Cleveland Clinic rate Jaffe assay (Beckman CX3, Beckman Coulter, Fullerton, CA). Consistent with current guidelines, egfr 60 ml/min/1.73 m 2 was used to define CKD. 16 Serum creatinine was dichotomized to define CKD using levels greater than the 97.5 th percentile cutoff for healthy men: 115 mol/l (1.51 mg/dl) and women: 96 mol/l (1.26 mg/dl). These thresholds were determined by calibrating samples with the Mayo Clinic laboratory where normal values for serum creatinine in healthy persons of white ethnicity had been established. 17 The upper limit of normal for serum creatinine does not change with age, presumably because of concurrent age-related declines in both creatinine production (muscle mass) and in GFR. 17

610 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE AJH June 2006 VOL. 19, NO. 6 Table 1. Characteristics of study sample overall and by hypertension status Characteristic Overall (N 2653) Hypertensive (N 2037) Normotensive (N 616) Male 34.9% 34.2% 37.0% African American 55.9% 57.8% 49.5% Age (y) 61 10 63 9 56 10 Hypertension duration (y) NA 13.6 (5.4, 24.6) NA Diabetes mellitus 23.1% 27.3% 9.3% Current smoker 10.4% 9.3% 14.1% Total cholesterol (mg/dl) 5.18 1.02 5.17 1.03 5.22 0.98 Hyperlipidemia (Cholesterol 5.18 mmol/l) 46.9% 45.4% 51.8% BMI (kg/m 2 ) 31.2 6.5 31.8 6.5 29.5 6.2 Obesity (BMI 30 kg/m 2 ) 51.9% 56.0% 38.5% CRP (nmol/l) 23 (11, 47) 25 (12, 51) 16 (7, 35) Inflammation (CRP 75 nmol/l) 13.2% 14.7% 8.1% ACR* (mg/g) 4.3 (2.2, 10.8) 4.8 (2.3, 14.0) 3.6 (2.0, 7.3) Serum creatinine ( mol/l) 85 24 86 25 80 17 Estimated GFR by MDRD equation (ml/min/1.73 m 2 ) 70 18 69 18 75 17 ACR albumin-to-creatine ratio; BMI body mass index; CRP C-reactive protein; GFR glomerulofiltration rate; MDRD Modification of Diet in Renal Disease; NA not applicable. Data summarized as frequency (%), mean SD, or median (25 th,75 th percentile). * Spot urine ACR. Statistical Analyses The three definitions of CKD were compared using the Cohen statistic. The statistic measures the degree of agreement between two tests on a scale of 1 to 1, where 0 represents agreement no greater than expected from chance alone, 1 represents complete agreement, and 1 represents complete disagreement. The bootstrap standard error of the statistic is presented to account for the correlation between siblings. Prevalence estimates of CKD defined by each method were examined among groups based on hypertension, sex, ethnicity, and age using logistic regression with generalized estimating equations (GEE) assuming an exchangeable correlation structure. The GEE methods were used to adjust the standard errors to account for the non-independence of the siblings in the sample. Multivariable-adjusted odds ratios, also using GEE, were estimated for each definition of kidney disease with respect to predictors measured namely, sex, ethnicity, age, hypertension, diabetes mellitus, current smoker, hyperlipidemia, obesity, and inflammation. Analyses were performed with JMP 5.1 and SAS 8.2 (SAS Institute, Cary, NC). Results There were 2653 subjects from 659 sibships at the follow-up visit. The sample was 65% female, 56% African American, and 77% hypertensive, with a mean ( SD) age of 61 10 years (Table 1). As expected, hypertensive subjects had a lower egfr, a higher serum creatinine level, and a higher ACR than normotensive subjects (P.001, for each rank sum test). In addition, diabetes mellitus, obesity, and elevated CRP were more prevalent among hypertensive than normotensive subjects (P.001, for each, 2 test). Cigarette smoking was more prevalent among normotensive subjects than among hypertensive subjects (P.001, 2 test). The overall prevalence of CKD was 28.1% by egfr 60 ml/min/1.73 m 2, 13.9% by elevated serum creatinine, and 15.4% by elevated ACR. Almost all subjects with an elevated serum creatinine level had a reduced egfr (Fig. 1). The relative overlap between elevated serum creatinine and elevated ACR was greater than that between reduced egfr and elevated ACR. Accordingly, the agreement between elevated serum creatinine and elevated ACR ( SE 0.19 0.02) was stronger than the agreement between reduced egfr and 35.0% (341) egfr < 60 ml/min/1.73 m 2 24.3% (237) Elevated SCr 11.6% (113) 2.7% (26) Elevated ACR 25.9% (252) FIG. 1. Venn diagram comparing methods for defining chronic kidney disease: estimated glomerular filtration rate (egfr) 60 ml/ min/1.73 m 2 ; elevated serum creatinine (SCr) 115 mol/l (1.51 mg/dl) in men and 96 mol/l (1.26 mg/dl) in women; and elevated urinary albumin-to-creatinine ratio (ACR) 17 mg/g in men and 25 mg/g in women. Five subjects with elevated SCr and egfr 60 ml/min/1.73 m 2 are not labeled on the diagram.

AJH June 2006 VOL. 19, NO. 6 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE 611 elevated ACR ( SE 0.07 0.02). Among subjects with a normal serum creatinine level, reduced egfr was paradoxically associated with a decreased risk of elevated ACR ( SE 0.07 0.02). In other words, persons identified as having CKD by one method (reduced egfr) were less likely to be identified as having CKD by another method (elevated ACR). These unexpected findings were consistent across strata defined by hypertension status (data not shown). Overall, the prevalence of CKD as defined by reduced egfr, elevated serum creatinine, or elevated ACR was higher among hypertensive than normotensive subjects (P.001 for each) and increased with age for both groups (P.001 for each) (Table 2). Whether hypertensive or normotensive, there was more CKD by reduced egfr than by elevated serum creatinine or elevated ACR. In the hypertensive group, the prevalence of CKD increased with duration of hypertension (P.001 for each). Although men had significantly more CKD than women when defined by elevated serum creatinine or elevated ACR (P.05 for both), the sex difference in prevalent CKD when defined by reduced egfr was not statistically significant (P.39). African Americans had more CKD than white individuals when defined by elevated serum creatinine or elevated ACR (P.01 for both), but the opposite was true when defined by reduced egfr (P.001). Figure 2 shows the prevalence of CKD stratified by age, sex, and ethnicity. For each definition of CKD, older subjects (lower graphs) had a higher prevalence of CKD than younger subjects (upper graphs). In both age and sex strata, there was more CKD among individuals of African American than among those of white ethnicity by elevated serum creatinine or by elevated ACR, but the opposite trend was present for reduced egfr. In both age and ethnicity strata, there was more CKD among men than among women by elevated serum creatinine or by elevated ACR. Although there was also more CKD among African American men than among African American women by reduced egfr, the opposite trend was observed between white men and white women. Each predictor (male sex, African American ethnicity, older age, hypertension, diabetes mellitus, current smoker, hyperlipidemia, obesity, and inflammation [CRP]) was independently associated with elevated ACR after multivariable adjustment (Table 3). Multivariable-adjusted predictor associations with elevated serum creatinine were Table 2. Prevalence of chronic kidney disease (%) Characteristic Elevated ACR* Elevated SCr Reduced egfr Hypertension Overall 18.2 16.2 31.5 Sex Male 20.9 (P.03) 21.6 (P.001) 33.0 (P.39) Female 16.8 13.4 30.7 Ethnicity African American 25.8 (P.001) 18.1 (P.005) 23.0 (P.001) White 7.2 13.6 42.8 Age (y) 20 39 12.5 (P.001) 12.5 (P.001) 12.5 (P.001) 40 59 14.5 7.8 20.0 50 69 17.7 15.5 32.9 70 24.3 29.5 45.6 Hypertension duration (y) 20 14.7 (P.001) 13.0 (P.001) 28.0 (P.001) 20 39 23.2 21.3 36.7 40 32.2 25.6 44.2 No hypertension Overall 6.2 6.2 16.9 Sex Male 10.6 (P.002) 8.8 (P.05) 18.0 (P.70) Female 3.7 4.7 16.3 Ethnicity African American 9.2 (P.004) 7.3 (P.34) 9.2 (P.001) White 3.1 5.1 24.4 Age (y) 20 39 0.0 (P.003) 0.0 (P.001) 5.4 (P.001) 40 59 4.6 2.8 12.9 50 69 9.2 9.5 22.0 70 11.8 23.1 36.5 ACR albumin-to-creatine ratio; egfr estimated glomerular filtration rate; SCr serum creatinine. * Elevated urine ACR 17 mg/g in men and 25 mg/g in women; Elevated SCr 115 mol/l (1.51 mg/dl) in men and 96 mol/l (1.26 mg/dl) in women; Estimated egfr 60 ml/min per 1.73 m 2 using the Modification of Diet in Renal Disease (MDRD) equation.

612 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE AJH June 2006 VOL. 19, NO. 6 FIG. 2. Prevalence of chronic kidney disease by three different definitions, stratified by age, sex, and ethnicity. Sex-related and ethnicityrelated trends were more consistent between elevated serum creatinine (SCr) and elevated urinary albumin-to-creatinine ratio (ACR) than they were with reduced estimated glomerular filtration rate (egfr). generally weaker (ie, odds ratio closer to 1) than the same predictor associations with elevated ACR. The notable exceptions were age, which had a stronger association with elevated serum creatinine (ie, odds ratio farther from 1), and current smoker status, which trended toward being protective against an elevated serum creatinine. The multivariable-adjusted predictor associations with reduced egfr were generally weaker and even discordant compared with the same associations with elevated serum creatinine. In particular, male sex and inflammation (CRP) were no longer statistically significant predictors of reduced egfr, and African American ethnicity appeared to be protective against reduced egfr. Discussion The results of this study challenge the validity of using estimated GFR (egfr) to identify chronic kidney disease (CKD). The recognized problem with serum creatinine is that levels are not only influenced by GFR but also by creatinine production (muscle mass). The MDRD equation was developed with regression analysis to estimate GFR by adjusting for creatinine production with surrogate markers (age, sex, and ethnicity). However, this equation was developed in a sample that had a clinical diagnosis of CKD, not in a sample in which the diagnosis of CKD was unknown such as this present sample of hypertensive sibships. Definition of CKD by a reduced egfr ( 60 ml/min/1.73 m 2 ) resulted in the unexpected finding of white subjects having a higher risk of CKD than African American subjects and no difference in risk of CKD between men and women. In contrast, when CKD was defined by elevated serum creatinine or elevated ACR, African American subjects and no difference in risk of CKD between men and women. Although the use of reduced egfr doubled the prevalence of CKD compared with that of elevated serum creatinine (31.5% v 16.2%, P.001), there was also much less agreement with elevated ACR ( 0.07 v 0.19, P.001). In fact, among subjects with normal serum creatinine levels, those with an egfr 60 ml/min/1.73 m 2 were less likely to have an elevated ACR than those with an egfr 60 ml/min/1.73 m 2 ( 0.07). It may be that most subjects with a reduced egfr but a normal serum creatinine level do not have kidney disease. These individuals could be misclassified as having CKD because of muscle mass (or protein intake) that is higher than average for their sex, ethnicity, and age but that is misinterpreted as a reduced GFR. Alternatively, higher-thanaverage muscle mass (or protein intake) increases urinary creatinine excretion, and this subsequently decreases the ACR. Thus subjects with higher than average muscle mass may be misclassified as not having CKD by an ACR. Although it is convenient to interpret egfr as true GFR and ACR as albuminuria, both are confounded by muscle mass and protein intake, especially in populations without a clinical diagnosis of kidney disease. Table 3. Adjusted* odds ratios for chronic kidney disease with respect to predictors Predictor Elevated ACR OR (95% CI) Elevated SCr OR (95% CI) Reduced egfr OR (95% CI) Sex: male 2.10 (1.62 2.73) 2.11 (1.63 2.72) 0.91 (0.74 1.13) Ethnicity: African American 3.75 (2.75 5.12) 1.19 (0.91 1.55) 0.27 (0.22 0.33) Age: per 10-year increment 1.26 (1.07 1.49) 2.10 (1.78 2.48) 1.97 (1.76 2.22) Hypertension 2.19 (1.45 3.29) 1.68 (1.16 2.43) 1.44 (1.12 1.85) Diabetes mellitus 3.85 (2.98 4.96) 1.91 (1.47 2.48) 1.53 (1.23 1.91) Current smoker 1.48 (1.01 2.16) 0.69 (0.44 1.08) 0.81 (0.58 1.11) Hyperlipidemia (cholesterol 5.18 mmol/l) 1.33 (1.03 1.70) 1.16 (0.91 1.47) 1.06 (0.88 1.27) Obesity (BMI 30 kg/m 2 ) 1.31 (1.01 1.68) 1.09 (0.84 1.40) 1.15 (0.95 1.39) Inflammation (CRP 75 nmol/l) 1.54 (1.13 2.10) 1.70 (1.24 2.32) 1.17 (0.85 1.53) CI confidence interval; OR odds ratio; other abbreviations as in Tables 1 and 2. * Adjusted for all other predictors (unadjusted and age-, sex-, and ethnicity-adjusted odds ratios showed similar trends between chronic kidney disease definitions); Elevated urine ACR 17 mg/g in men and 25 mg/g in women; Elevated SCr 115 mol/l (1.51 mg/dl) in men and 96 mol/l (1.26 mg/dl) in women; Reduced egfr 60 ml/min per 1.73 m 2 using the MDRD equation.

AJH June 2006 VOL. 19, NO. 6 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE 613 Part of the sex and ethnicity discrepancy between CKD estimates based on reduced egfr versus the other methods may be caused by regression coefficients for the MDRD equation being inaccurate for the population sampled in this study. Recent studies have found that regression coefficients used in the MDRD equation for serum creatinine, sex, and ethnicity are substantially weaker in populations without kidney disease. 4 7 Poggio et al found that at the same serum creatinine level, healthy women had only a 5% lower measured GFR than healthy men, much less than the 26% lower egfr expected with the MDRD equation. Likewise, at the same serum creatinine level, healthy African American subjects had the same measured GFR as healthy white subjects, not the 21% higher egfr expected with the MDRD equation. 5 To explore the effect of using coefficients derived from a healthy population, the MDRD equation was modified with a different female coefficient (0.952 instead of 0.742) and African American ethnicity coefficient (0.983 instead of 1.21) (5). The multivariable-adjusted odds ratio of egfr 60 ml/min.1.73 m 2 for male versus female sex increased to 5.98 (95% CI, 4.80 to 7.49) instead of 0.92 and the multivariable-adjusted odds ratio for African American versus white ethnicity increased to 1.35 (95% CI, 1.08 to 1.68) instead of 0.29. These observations support the view that the regression coefficients in the MDRD equation may lead to an incorrect estimate of the sex-specific and ethnicity-specific risk of kidney disease. The correct coefficients and form of the equation for a sample of community hypertensive sibships is unknown without measurement of GFR (eg, iothalamate clearance). A more fundamental question is whether CKD should even be identified by estimated GFR, as estimating equations are population specific. 4 7 In contrast, a normal value study interprets serum creatinine as a marker of CKD probability instead of as a marker of GFR. For example, a serum creatinine greater than the 97.5 th percentile for normal individuals can be interpreted as having a specificity of 97.5% for the diagnosis of CKD. Moreover, predictors for an elevated serum creatinine (male sex and African American ethnicity) are more consistent with other methods of identifying CKD. A general population study found that African American subjects were more likely than white subjects to have an elevated ACR. 15 There is also a 50% higher incidence of end-stage renal disease among men than among women and a fourfold higher incidence of end-stage renal disease among individuals of African American ethnicity than among those of white ethnicity. 24 There are potential limitations to this study. First, sexspecific but not ethnicity-specific thresholds were used to define elevated serum creatinine and elevated ACR. Ethnicity-specific thresholds were not available but would be desirable, as there can be ethnicity-related differences in creatinine production and tubular secretion. 22,25 Second, without measured GFR (eg, iothalamate clearance), definitive conclusions cannot be made, particularly with regard to the sex- and ethnicity-specific risk for a reduction in renal function. However, these findings provide an impetus for more studies that measure rather than estimate GFR in populations without pre-existing diagnoses of kidney disease. In conclusion, among a sample of predominantly hypertensive siblings, elevated serum creatinine and elevated ACR had similar associations. The same associations with reduced estimated GFR were less consistent, particularly for gender and ethnicity. This is of particular importance for hypertensive patients given their increased risk of kidney disease. Estimated GFR by the MDRD equation is arguably appropriate for staging severity of CKD, but only after a diagnosis of CKD has been made by an elevated serum creatinine level or other evidence of kidney damage. Acknowledgments The authors appreciate the technical assistance provided by Jodie Van de Rostyne and Janell Hovey and the statistical advice provided by Kent R. Bailey. This work was supported by United States Public Health Service Grants from the National Institutes of Health MOI RR 00585. References 1. Coresh J, Astor BC, Greene T, Eknoyan G, Levey AS: Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis 2003;41:1 12. 2. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers 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;130:461 470. 3. Levey AS, Greene T, Kusek JW, Beck GJ: A simplified equation to predict glomerular filtration rate from serum creatinine (abstract). J Am Soc Nephrol 2000;11:A0828. 4. Ibrahim H, Mondress M, Tello A, Fan Y, Koopmeiners J, Thomas W: An alternative formula to the Cockcroft-Gault and the modification of diet in renal disease formulas in predicting GFR in individuals with type 1 diabetes. J Am Soc Nephrol 2005;16:1051 1060. 5. 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;16:459 466. 6. 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;141:929 937. 7. Lin J, Knight E, Hogan M, Singh A: A comparison of prediction equations for estimating glomerular filtration rate in adults without kidney disease. J Am Soc Nephrol 2003;14:2573 2580. 8. Hajjar I, Kotchen TA: Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988 2000. J Am Med Assoc 2003;290:199 206. 9. Rostand SG: Oligonephronia, primary hypertension and renal disease: is the child father to the man? Nephrol Dial Transplant 2003;18:1434 1438. 10. Segura J, Campo C, Gil P, Roldan C, Vigil L, Rodicio JL, Ruilope LM: Development of chronic kidney disease and cardiovascular prognosis in essential hypertensive patients. J Am Soc Nephrol 2004;15:1616 1622.

614 METHODS FOR IDENTIFYING CHRONIC KIDNEY DISEASE AJH June 2006 VOL. 19, NO. 6 11. Fox CS, Larson MG, Leip EP, Culleton B, Wilson PW, Levy D: Predictors of new-onset kidney disease in a community-based population. J Am Med Assoc 2004;291:844 850. 12. Domrongkitchaiporn S, Sritara P, Kitiyakara C, Stitchantrakul W, Krittaphol V, Lolekha P, Cheepudomwit S, Yipintsol T: Risk factors for development of decreased kidney function in a southeast Asian population: a 12-year cohort study. J Am Soc Nephrol 2005;16:791 799. 13. Haroun MK, Jaar BG, Hoffman SC, Comstock GW, Klag MJ, Coresh J: Risk factors for chronic kidney disease: a prospective study of 23,534 men and women in Washington County, Maryland. J Am Soc Nephrol 2003;14:2934 2941. 14. Chen J, Muntner P, Hamm LL, Jones DW, Batuman V, Fonseca V, Whelton PK, He J: The metabolic syndrome and chronic kidney disease in U.S. adults. Ann Intern Med 2004;140:167 174. 15. Coresh J, Byrd-Holt D, Astor BC, Briggs JP, Eggers PW, Lacher DA, Hostetter TH: Chronic kidney disease awareness, prevalence, and trends among U.S. adults, 1999 to 2000. J Am Soc Nephrol 2005;16:180 188. 16. Levey AS, Coresh J, Balk E, Kausz AT, Levin A, Steffes MW, Hogg RJ, Perrone RD, Lau J, Eknoyan G: National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003;139:137 147. 17. Rule AD, Gussak HM, Pond GR, Bergstralh EJ, Stegall MD, Cosio FG, Larson TS: Measured and estimated GFR in healthy potential kidney donors. Am J Kidney Dis 2004;43:112 119. 18. FBPP Investigators: Multi-Center Genetic Study of Hypertension: The Family Blood Pressure Program (FBPP). Hypertension 2002;39:3 9. 19. Daniels PR, Kardia SL, Hanis CL, Brown CA, Hutchinson R, Boerwinkle E, Turner ST: Familial aggregation of hypertension treatment and control in the Genetic Epidemiology Network of Arteriopathy (GENOA) study. Am J Med 2004;116:676 681. 20. O Meara JG, Kardia SL, Armon JJ, Brown CA, Boerwinkle E, Turner ST: Ethnic and sex differences in the prevalence, treatment, and control of dyslipidemia among hypertensive adults in the GENOA study. Arch Intern Med 2004;164:1313 1318. 21. Warram JH, Gearin G, Laffel L, Krolewski AS: Effect of duration of type I diabetes on the prevalence of stages of diabetic nephropathy defined by urinary albumin/creatinine ratio. J Am Soc Nephrol 1996;7:930 937. 22. Mattix HJ, Hsu CY, Shaykevich S, Curhan G: Use of the albumin/ creatinine ratio to detect microalbuminuria: implications of sex and ethnicity. J Am Soc Nephrol 2002;13:1034 1039. 23. Coresh J, Astor BC, McQuillan G, Kusek J, Greene T, Van Lente F, Levey AS: Calibration and random variation of the serum creatinine assay as critical elements of using equations to estimate glomerular filtration rate. Am J Kidney Dis 2002;39:920 929. 24. US Renal Data System: Excerpts from the USRDS 2004 Annual Data Report. Am J Kidney Dis 2005;45(Suppl 1):S60 S61. 25. Hsu CY, Chertow GM, Curhan GC: Methodological issues in studying the epidemiology of mild to moderate chronic renal insufficiency. Kidney Int 2002;61:1567 1576.