ORIGINAL INVESTIGATION. Use of Urinary Albumin Excretion and Estimated Glomerular Filtration Rate

Transcription

1 ORIGINAL INVESTIGATION Definition of Kidney Dysfunction as a Cardiovascular Risk Factor Use of Urinary Albumin Excretion and Estimated Glomerular Filtration Rate Massimo Cirillo, MD; Maria Paola Lanti, MD; Alessandro Menotti, MD; Martino Laurenzi, MD; Mario Mancini, MD; Alberto Zanchetti, MD; Natale G. De Santo, MD Background: Urinary albumin excretion (UAE) and estimated glomerular filtration rate (egfr) have been used separately to provide information about cardiovascular risk. We analyzed whether UAE and egfr used together provide complementary information. Methods: We analyzed UAE, egfr, cardiovascular risk factors, and incidence of cardiovascular disease in 1665 men and women of the Gubbio Population Study (aged years). We designated UAE in the highest decile as high ( 18.6 µg/min in men and 15.7 µg/min in women) and egfr in the lowest decile as low ( ml/min/1.73 m 2 in men and ml/min/1.73 m 2 in women). Results: Kidney dysfunction defined using both markers was more frequent than using 1 marker (UAE alone or egfr alone) (P.001) because high UAE and low egfr clustered in different individuals and were weakly associated with each other (P=.12). The hazard ratio (HR) for incident cardiovascular disease was elevated for both markers, independently of each other (HR for high UAE, 2.15; 95% confidence interval [CI], ; HR for low egfr, 2.14; 95% CI, ). Kidney dysfunction defined by both markers predicted cardiovascular disease independently of sex, age, hypertension, hypercholesterolemia, smoking, diabetes mellitus, prior cardiovascular disease, left ventricular hypertrophy, and obesity (HR, 1.50; 95% CI, ). The discriminant power of dysfunction defined by both markers was statistically significant (area under the receiver operating characteristic curve, [P=.02]) and slightly higher than what was found with 1 marker of diabetes mellitus, prior cardiovascular disease, left ventricular hypertrophy, and obesity. Conclusions: High UAE and low egfr provide complementary information in defining kidney dysfunction because they cluster in different individuals. Concomitant evaluation of both markers should be considered to adequately assess kidney dysfunction and cardiovascular risk. Arch Intern Med. 2008;168(6): Author Affiliations: Unit of Nephrology, Second University of Naples (Drs Cirillo and De Santo), and Unit of Internal Medicine, Federico II University (Dr Mancini), Naples; Association for Cardiac Research, Rome (Drs Lanti and Menotti); Center for Preventive Medicine, Gubbio (Dr Laurenzi); and Centro Fisiologia Clinica e Ipertensione, Istituto Auxologico Italiano, Ospedale Maggiore, University of Milan, Milan (Dr Zanchetti), Italy. SEVERAL GUIDELINES INDICATE kidney dysfunction as a cardiovascular risk factor on the basis of data showing a relation with cardiovascular disease for 2 markers of kidney function: urinary albumin excretion (UAE) and glomerular filtration rate (GFR). 1-3 Longitudinal population-based studies support these statements For UAE, cardiovascular risk was elevated in persons with macroalbuminuria 4-6 and microalbuminuria. 7-9 For GFR, cardiovascular risk was elevated in persons with clinically elevated serum creatinine levels and by calculation of the estimated GFR (egfr) To our knowledge, no previous studies have reported complete data on both UAE and egfr. Thus, it is uncertain whether the 2 markers of kidney function provide complementary or overlapping information for cardiovascular risk. This point could be important for understanding the mechanisms linking kidney dysfunction to cardiovascular disease and for practical considerations regarding the optimal workup for the assessment of cardiovascular risk. The main objective of the present population-based study was to investigate prospectively the association with cardiovascular disease of kidney dysfunction defined using 2 markers (high UAE and low egfr) compared with kidney dysfunction defined using a single marker (high UAE alone or low egfr alone, alternatively). METHODS The Gubbio Population Study is an investigation ofapopulationsampleresidinginthecityofgubbio in central Italy. 22 Activities were approved by thelocalinstitutionalcommitteeandincludedan informed consent. Previous studies report information on response rates, responders and nonresponders, similaritiesofthegubbiopopulation to the Italian population, and the timing of examinations Thetargetcohortofthepresent analysisislimitedtoexamineesaged45to64years 617

2 at the examination (hereinafter referred to as the baseline examination) because UAE was measured only for these individuals. 24 The baseline examination included the collection of timed overnight urine samples and morning blood samples under fasting conditions, the administration of questionnaires on cardiovascular disease and treatment, 28 a 12-lead electrocardiogram interpreted according to the Minnesota code, 28 and measurements of anthropometry and blood pressure. Measurements of laboratoryvalueswereperformedbymeansofautomatedproceduresand included levels of urinary albumin (using immunoturbidimetry after ultrafiltration), 24 serum glucose, total and high-density lipoprotein cholesterol, 22,24 and serum creatinine (using a kinetic alkaline picrate assay). 27 The intra-assay error in daily blind duplicates was less than 10% for urinary albumin levels and less than 5% forserumvariables. Trainedphysiciansadministeredquestionnaires and measured blood pressure after a 5-minute rest in the sitting position using mercury sphygmomanometers and cuffs of appropriate size. The mean of the second and third blood pressure measurements was used for analyses. KIDNEY FUNCTION The UAE was expressed in micrograms per minute, not as the ratio of urinary albumin to creatinine levels, to avoid the misclassification of a high ratio secondary to low urinary creatinine levels. 26 The egfr was calculated by the following abbreviated equation of the Modification Diet in Renal Disease Study (MDRD) 2 : 186 Serum Creatinine Level Age ( for Women). The ethnicity factor of the equation was not used because the Gubbio population is exclusively white. For the creatinine assay in the study, the MDRD equation was the best predictor of true GFR 29 and of disorders associated with low kidney function. 27 The MDRD equation has been questioned because it may underestimate the true GFR in the normal-high range 30 and in women. 27,29 To reduce these biases, main analyses were performed using a low egfr definition derived from data distribution within the study cohort, that is, using the lowest decile of the distribution analyzed separately for men and women. To avoid differences in the statistical power (ie, groups of different size), the main analyses for high UAE were conducted in a similar fashion. High UAE was defined as the highest decile of the distribution, which was analyzed separately for men and women to reduce the confounding of sex. 24,31 Three definitions were used for kidney dysfunction (categorical variable, yes or no). With the use of 1 marker, dysfunction was defined as a high UAE regardless of egfr (using UAE alone) or, alternatively, as a low egfr regardless of UAE (using egfr alone). With the use of both markers, dysfunction was defined as composite dysfunction and included high UAE without low egfr, low egfr without high UAE, or high UAE and low egfr. Other analyses focused on the thresholds suggested in clinical practice for UAE ( 20 µg/min) and egfr ( 60 ml/min/1.73 m 2 ). 1-3 CARDIOVASCULAR RISK FACTORS Hypertension was defined as a systolic pressure of 140 mm Hg or higher, a diastolic pressure of 90 mm Hg or higher, and/or reported drug treatment for hypertension. Hypercholesterolemia was defined as a serum total cholesterol level of 240 mg/dl or higher (to convert to millimoles per liter, multiply by ) and/or reported drug treatment for hypercholesterolemia. Smoking habit was self-reported information and had been validated by measurement of exhaled carbon monoxide level. 32 Diabetes mellitus was defined as a fasting serum glucose level of 126 mg/dl or higher (to convert to millimoles per liter, multiply by ) and/or reported treatment with 1 or more antidiabetic drugs or insulin. Obesity was defined as a body mass index (calculated as weight in kilograms divided by height in meters squared) of 30 or higher. Left ventricular hypertrophy was defined as high-amplitude R waves on the electrocardiogram in the absence of block patterns. PRIOR CARDIOVASCULAR DISEASE Prevalent cardiovascular disease was used as a covariate in analyses about incident events because preexisting disease predicts incident disease. 33 Prior cardiovascular disease was defined as preexisting cerebrovascular disease, ischemic heart disease, or peripheral artery disease. Cerebrovascular disease was defined as reported revascularization of a carotid artery or other cerebral arteries and/or as a reported diagnosis of ischemic or hemorrhagic stroke combined with stroke symptoms. 28 Ischemic heart disease was defined as reported revascularization of the coronary arteries and/or electrocardiographic abnormalities specific to myocardial infarction, 25,28 a criterion that aimed at reducing the misclassification of unrecognized or silent infarctions Peripheral artery disease was defined as reported revascularization of the aorta or of peripheral arteries. INCIDENT EVENTS The end point of the analyses was the incidence of events (hospitalizations and deaths) between the baseline examination and the censoring date (May 31, 2001, for living persons or the date of death for fatal events). Diagnoses from hospitalizations and mortality data (municipal registry) were systematically collected. This information was reviewed by an independent observer (A.M.) for coherence with reported data and coded according to the World Health Organization s International Classification of Diseases, Ninth Revision (ICD-9). For the definition of preexisting cardiovascular disease, an incident cardiovascular disease was defined as an event that signaled cerebrovascular disease, ischemic heart disease, or peripheral artery disease. Incident cerebrovascular disease was defined as a hospitalization with revascularization of a carotid artery or other cerebral arteries or with a diagnosis of ischemic or hemorrhagic stroke (ICD-9 codes , 436, or 433, nonfatal event) or as death due to ischemic or hemorrhagic stroke (ICD-9 codes , fatal event). Incident ischemic heart disease was defined as a hospitalization with revascularization of the coronary arteries or with a diagnosis of myocardial infarction (ICD-9 codes or 414, nonfatal event) or as death due to myocardial infarction (ICD-9 codes , fatal event). Incident peripheral artery disease was defined as a hospitalization with revascularization of the aorta or of other peripheral arteries (ICD-9 code 441, nonfatal event) or with a diagnosis of aortic aneurysm or as death due to aortic aneurysm (ICD-9 code 441, fatal event). An incident event for other diseases was defined as death (not hospitalization) for causes other than those defined as cardiovascular disease. In the case of a nonfatal event followed by a fatal event, priority was given to the nonfatal event. In the case of multiple causes of death, a hierarchical system was used giving priority to violent causes, advanced cancer, ischemic heart disease, and cerebrovascular disease, in that order. STATISTICAL ANALYSIS The main analyses were designed to compare the definitions of dysfunction based on the population-derived thresholds of kidney function (highest decile for UAE and lowest decile for egfr). Ancillary analyses focused on the data for the thresholds of kidney function suggested in clinical practice (UAE, 20 µg/min; egfr, 60 ml/min/1.73 m 2 ) and on the relation of UAE or egfr to cardiovascular events in individuals without kidney dysfunction. 618

3 160 Men Women egfr, ml/min/1.73 m UAE, µg/min Figure 1. Plot of estimated glomerular filtration rate (egfr) over urinary albumin excretion (UAE) in men and women aged 45 to 64 years. Estimated GFR and UAE were not correlated with each other (correlation coefficients, 0.035; P.15 in partial correlation analysis for men and women combined with control for sex, and in Pearson correlation analyses done separately for men and women). Statistical procedures included 2 analysis with calculation of odds ratio (OR), McNemar test, correlation analysis, proportional hazard ratio (HR) by Cox regression, and 95% confidence interval (CI). The discriminant power for incident events was assessed by receiver operating characteristic curves with calculation of the area under the curve (AUC). RESULTS Nineteen of the 1684 individuals with baseline ages of 45 to 64 years in the Gubbio population were excluded because of missing electrocardiographic data. Thus, the study cohort was composed of 1665 individuals. Descriptive statistics in the whole cohort for baseline values of cardiovascular risk factors and kidney function were previously reported with and without inclusion of diabetic individuals The mean (SD) for the interval between the baseline examination and the censoring date was 10.4 (2.2) years. The number of person-years was The number of events was 110 for cardiovascular disease and 86 for other diseases. The incidence of events was 0.64% per year for cardiovascular disease and 0.50% per year for other diseases. UAE AND egfr AS SEPARATE MARKERS OF KIDNEY DYSFUNCTION The cutoff for high UAE (highest decile) was µg/ min or higher in men and µg/min or higher in women. The cutoff for low egfr (lowest decile) was less than ml/min/1.73 m 2 in men and less than ml/min/1.73 m 2 in women. With the use of a single marker (high UAE alone or low egfr alone, alternatively), 1498 individuals (90.0%) were defined as having no dysfunction and 167 individuals (10.0%) as having dysfunction (74 men and 93 women). The number of individuals with dysfunction was identical using UAE or egfr, as expected according to the decile-based definition. With the use of both markers (composite dysfunction), only 1354 individuals (81.3%) were defined as having no dysfunction and 311 (18.7%) as having dysfunction (1.86 times higher than the number using a single marker, P.001). These changes in the number of individuals with or without dysfunction reflected the weak association between high UAE and low egfr (P=.12) that clustered in different subgroups as shown by the following cross-tabulation: 1354 individuals without high UAE and without low egfr; 144 individuals with high UAE and without low egfr (isolated high UAE); 144 individuals with low egfr and without high UAE (isolated low egfr); and 23 individuals with high UAE and low egfr. Thus, when we used high UAE alone, 144 individuals were defined as having no dysfunction despite having a low egfr. When we used low egfr alone, 144 individuals were defined as having no dysfunction despite having a high UAE. The markers UAE and egfr were also not associated with each other in analyses using clinical thresholds of kidney function (association between a UAE of 20 µg/ min and an egfr of 60 ml/min/1.73 m 2, n=77 and n=177, respectively [P=.94]) and in analyses across the whole range of UAE and egfr values (Figure 1). 619

4 Table 1. Indexes of Kidney Function and Other Variables or Prevalence by Presence and Type of Baseline Kidney Dysfunction in Individuals Aged 45 to 64 Years Variable No Kidney Dysfunction (n=1354) Isolated High UAE b (n=144) Composite Kidney Dysfunction a High UAE and Low egfr (n=23) Isolated Low egfr c (n=144) UAE, mean (SD), µg/min 7.8 (4.4) 50.5 (179.2) (210.4) 7.7 (4.6) Serum creatinine level, mean (SD), mg/dl 0.93 (0.14) 0.93 (0.15) 1.26 (0.13) 1.21 (0.13) egfr, mean (SD), ml/min/1.73 m (12.7) 77.7 (14.5) 57.2 (5.6) 56.0 (4.5) Women, % Age, mean (SD), y 54.5 (5.6) 55.5 (5.1) 57.7 (5.0) 56.9 (5.5) Prevalent cardiovascular risk factors Systolic pressure, mean (SD), mm Hg (17.0) (21.6) (17.5) (17.9) Diastolic pressure, mean (SD), mm Hg 77.7 (9.3) 85.5 (11.0) 84.4 (6.7) 78.8 (8.8) Antihypertensive drug use, % Hypertension, % Serum total cholesterol level, mean (SD), mg/dl (37.6) (43.0) (32.7) (37.0) Serum HDL cholesterol level, mean (SD), mg/dl 50.7 (12.8) 50.0 (13.1) 51.1 (15.1) 48.1 (11.1) Hypercholesterolemia, % Current smokers, % Serum glucose level, mean (SD), mg/dl 94.3 (20.2) 99.1 (27.9) 96.9 (20.0) 92.1 (14.4) Diabetes mellitus, % Left ventricular hypertrophy, % BMI, mean (SD) 27.8 (4.2) 30.1 (4.6) 31.2 (5.7) 28.0 (4.2) Obesity, % Prior cardiovascular disease, % Cerebrovascular disease Ischemic heart disease Peripheral artery disease Any of the above Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); egfr, estimated glomerular filtration rate; HDL, high-density lipoprotein; UAE, urinary albumin excretion. SI conversion factors: To convert total and HDL cholesterol to millimoles per liter, multiply by ; creatinine to micromoles per liter, multiply by 88.4; and glucose to millimoles per liter, multiply by a Indicates with high UAE (highest sex-specific decile) or with low egfr (lowest sex-specific decile). b Indicates with high UAE and without low egfr. c Indicates with low egfr and without high UAE. KIDNEY DYSFUNCTION AND BASELINE CORRELATES Table 1 shows the baseline values of kidney function and other variables in individuals without dysfunction and in individuals with dysfunction according to the type of dysfunction (isolated high UAE, high UAE and low egfr, or isolated low egfr). Cardiovascular risk factors were more prevalent in individuals with dysfunction than in individuals without dysfunction. The difference was significant for hypertension, hypercholesterolemia, and obesity in individuals with isolated high UAE (OR, 2.6 [P.001]); for hypertension, hypercholesterolemia, obesity, and left ventricular hypertrophy in individuals with high UAE and low egfr (OR, 3.0 [P.01]); and only for hypertension in individuals with isolated low egfr (OR, 1.46 [P=.04]). Prior cardiovascular disease was more prevalent in all subgroups with dysfunction than in individuals without dysfunction (OR, 2.2 [P.04]). Prior cardiovascular disease was more prevalent in individuals with high UAE and low egfr than in individuals with isolated high UAE and isolated low egfr combined (OR, 3.58 [P=.03]). KIDNEY DYSFUNCTION AND INCIDENT EVENTS The incidence of events was higher in individuals with high UAE than in individuals without high UAE for cardiovascular disease (12.0% vs 6.0%; OR, 2.13 [P=.003]) but not for other diseases (7.2% vs 4.9%; OR, 1.49 [P=.21]). The HR for cardiovascular events in individuals with high UAE was 2.15 (95% CI, ) and did not change when we excluded the 9 individuals with macroalbuminuria (UAE, 200 µg/min) (HR, 2.14; 95% CI, ). Results were similar using the clinical threshold for UAE (UAE, 20 µg/min vs 20 µg/min) (HR, 2.81; 95% CI, ). The incidence of events in individuals with low egfr was higher than in individuals without low egfr for cardiovascular disease (12.0% vs 6.0%; OR, 2.13 [P=.003]) but not for other diseases (4.8% vs 5.2%; OR, 0.92 [P=.82]). The HR for cardiovascular events in individuals with low egfr was 2.14 (95% CI, ). Results were similar using the clinical threshold for egfr ( 60 vs 60 ml/ min/1.73 m 2 ) (HR, 1.64; 95% CI, ). The incidence of cardiovascular events by the presence and type of kidney dysfunction is shown as a per- 620

5 Table 2. Incidence of Cardiovascular Disease and Other Disease Events by Presence and Type of Baseline Kidney Dysfunction in Individuals Aged 45 to 64 Years Cardiovascular and Other Disease Events, % of Patients No Kidney Dysfunction Isolated High UAE b Composite Kidney Dysfunction a High UAE and Low egfr Isolated Low egfr c Cerebrovascular disease Ischemic heart disease Peripheral artery disease Any of the above Nonfatal Fatal Other diseases Event-Free Survival No kidney dysfunction Isolated low egfr Isolated high UAE High UAE and low egfr Time, y Abbreviations: egfr, estimated glomerular filtration rate; UAE, urinary albumin excretion. a Indicates with high UAE (highest sex-specific decile) or with low egfr (lowest sex-specific decile). b Indicates with high UAE and without low egfr. c Indicates with low egfr and without high UAE. centage rate in Table 2 and as a Kaplan-Meier plot in Figure 2. Compared with individuals without dysfunction, the HR for a cardiovascular event was 1.85 in individuals with isolated high UAE (95% CI, ), 1.84 in individuals with isolated low egfr (95% CI, ), and 5.93 in individuals with high UAE and low egfr (95% CI, ). The elevated HR in individuals with high UAE and low egfr suggested an interaction between high UAE and low egfr. The interaction was significant because the HR for cardiovascular events was higher in individuals with high UAE and low egfr than individuals with isolated high UAE and isolated low egfr combined (HR, 3.19; 95% CI, ). When kidney dysfunction was defined using both markers (composite dysfunction), the HR for cardiovascular events was 2.09 compared with no kidney dysfunction (95% CI, ). Findings were similar when this HR was calculated separately for men and women (2.22 and 2.10, respectively), for nonfatal and fatal events (1.89 and 3.97, respectively), and for ischemic heart disease and other cardiovascular disease (2.10 and 1.84, respectively). MULTIVARIATE ANALYSES Table 3 shows the HR for cardiovascular events associated with kidney dysfunction and other variables in multivariate Cox models. Analyses were limited to the kidney dysfunction defined using both markers (composite dysfunction) compared with no kidney dysfunction because definitions using 1 marker misclassified individuals with isolated low egfr (definition by high UAE alone) or with isolated high UAE (definition by low egfr alone). The following 3 models were analyzed for investigation about different presentations of incident cardiovascular disease: model 1 for first events (nonfatal and fatal events in individuals without prior cardiovascular disease); model 2 for first or recurrent events (nonfatal and Figure 2. Kaplan-Meier plot of incident cardiovascular events in individuals aged 45 to 64 years. No kidney dysfunction (n=1354 without high urinary albumin excretion [UAE] and without low estimated glomerular filtration rate [egfr]) compared with kidney dysfunction divided subjects into the following 3 subgroups: isolated high UAE (n=144 with high UAE and without low egfr), isolated low egfr (n=144 with low egfr and without high UAE), and high UAE and low egfr (n=23). High UAE was defined as a UAE in the highest sex-specific decile, and low egfr was defined as an egfr in the lowest sex-specific decile. fatal events in the whole cohort including individuals with prior cardiovascular disease); and model 3 for fatal events (fatal events only in the entire cohort). The HR of kidney dysfunction was significant in all models. Three similar additional models were analyzed using categorical variables as follows: obesity replacing body mass index, hypertension replacing systolic pressure and antihypertensive treatment, and hypercholesterolemia replacing serum total cholesterol level. The HR of composite dysfunction was elevated in these models (HRs, 1.41, 1.45, and 2.49, respectively [P.05]). For the additional categorical variables, HR was never significant for obesity (P.5), significant for hypercholesterolemia in models on nonfatal and fatal events in the entire cohort (HRs, 1.55 and 1.58, respectively [P.05]), and significant for hypertension in all models (HRs, 1.80, 1.92, and 1.98, respectively [P.05]). The area under the receiver operating characteristic curve of composite dysfunction was (P=.02), a value that ranked between the AUCs of hypertension, smoking, and hypercholesterolemia (0.601, 0.572, and 0.570, respectively [P.01]) and the AUCs of prior cardiovascular disease, left ventricular hypertrophy, diabetes mellitus, and obesity (0.541, 0.534, 0.524, and 0.517, respectively [P.15]). The AUC of composite dysfunction was higher for fatal events than for nonfatal events (0.639 [P=.054] and [P=.07], respectively). RELATION OF UAE AND egfr TO EVENTS IN INDIVIDUALS WITHOUT KIDNEY DYSFUNCTION In analyses limited to individuals without high UAE ( µg/min in men and µg/min in women), the incidence of cardiovascular events did not differ among UAE tertiles (6.2%, 7.5%, and 6.1% [P=.35]). The relation between UAE and events was not signifi- 621

6 Table 3. Hazard Ratios for Incident Cardiovascular Disease of Kidney Dysfunction Defined by 2 Markers (Composite Dysfunction a ) and Other Variables in Multivariate Cox Models for Individuals Aged 45 to 64 Years Incident Cardiovascular Disease b Variable All (Nonfatal and Fatal) All (Nonfatal and Fatal) Fatal Only No. of individuals 1611 c 1665 d 1665 d No. of events Composite kidney dysfunction, a (yes vs no) 1.47 ( ) e 1.50 ( ) e 3.03 ( ) f Sex (men vs women) 4.23 ( ) g 3.88 ( ) g 8.69 ( ) g Age, y (55-64 vs y) 1.27 h 1.32 h 1.47 h Systolic pressure 1 SD (18 mm Hg) 1.25 ( ) e 1.12 h 1.24 ( ) i Antihypertensive drug treatment (yes vs no) 1.91 ( ) f 2.21 ( ) g 1.60 h Serum total cholesterol 1 SD (40 mg/dl) 1.17 h 1.21 ( ) e 1.19 h Serum HDL cholesterol 1 SD (13 mg/dl) 0.85 h 0.90 h 1.01 h Smoking (yes vs no) 1.63 ( ) e 1.66 ( ) f 2.40 ( ) e Diabetes (yes vs no) 1.03 h 1.27 h 2.49 ( ) Left ventricular hypertrophy (yes vs no) 2.24 ( ) e 2.48 ( ) f 1.41 h Prior cardiovascular disease (yes vs no) Not included in the model 2.59 ( ) f 4.81 ( ) g BMI 1 SD (4.3) 1.13 h 1.11 h 1.44 ( ) e Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); HDL, high-density lipoprotein. SI conversion factor: To convert total and HDL cholesterol to millimoles per liter, multiply by a With high urinary albumin excretion (highest sex-specific decile) or with low estimated glomerular filtration rate (lowest sex-specific decile). b Unless otherwise indicated, data are expressed as hazard ratio (95% confidence interval). c Indicates without prior cardiovascular disease. d Indicates whole cohort. e P.05. f P.01. g P.001. h P.10; therefore, no 95% confidence interval is given. i P.10 and P.05. cant in any analysis limited to these individuals, even when using UAE as a continuous variable (data not shown). In analyses limited to individuals without low egfr ( ml/min/1.73 m 2 in men and ml/min/ 1.73 m 2 in women), the incidence of cardiovascular events did not differ among egfr tertiles (8.2%, 6.3%, and 6.3% [P=.13]). The relation between egfr and events was not significant in any analysis limited to these individuals, even when using egfr as a continuous variable (data not shown). COMMENT This population-based study reports several new observations in support of the idea that the measurement of 2 markers of kidney function UAE and egfr provides additional and independent information for assessing cardiovascular risk. First, the 2 markers were not associated with each other as proved by analyses using the highest UAE decile and the lowest egfr decile, by analyses using clinical definitions of altered UAE and egfr, and by analyses on the entire range of UAE and egfr values. The lack of the association between these 2 markers implied that high UAE and low egfr tended to mark separate groups in the population and that the number of individuals defined as having kidney dysfunction was higher when using both markers than when using one. Second, high UAE and low egfr were associated with high cardiovascular risk independently of each other, as proved by data for the subgroups with isolated high UAE or isolated low egfr. The association with cardiovascular disease was of similar strength for high UAE and low egfr, significant longitudinally (as well as crosssectionally), and independent of the few individuals carrying both markers of kidney dysfunction (high UAE and low egfr). Third, for prediction of cardiovascular disease, the definition of dysfunction based on both markers had a discriminant power similar to that of traditional factors. In addition, the study provides further information on the independence of the association between kidney dysfunction and cardiovascular disease showing that, when defined by both markers, the association of dysfunction with cardiovascular disease is statistically independent of several other factors. Caution should be used in extrapolating these observations to other populations given that the study involved only middleaged white subjects with a low prevalence of diabetes mellitus. The present results agree with reports of a relation between microalbuminuria and incident cardiovascular disease, 7-9 confirming that the relation exists also with nonfatal events. 9 For low egfr and incident cardiovascular disease, the results support the large number of studies reporting an independent relation. 10,12,15-21 The sample size and the number of events were the major limitations of this study. These limitations affected the power of the study but were unlikely to cause a substantial bias because the results for traditional predictors of cardiovascular risk were in agreement with expectations for middle-aged persons. For UAE, the absolute urinary excretion was preferred to the ratio of urinary albumin to creatinine levels to avoid the misclassifica- 622

7 tion of individuals with a high ratio secondary to low muscular mass. 26 For egfr, the method used in the study was the best predictor of true GFR in clinical settings 29 and of disorders associated with low kidney function in the population. 27 The use of cutoffs derived from the sexspecific distribution of kidney function in the population was designed to achieve groups of similar size and to reduce the bias secondary to laboratory-specific factors 37 or to the effects of sex on UAE 24,31 and egfr. 27,38 These results have practical implications on the usefulness of the assessment of kidney function in the estimation of cardiovascular risk. The findings appear meaningful from a clinical viewpoint because the excess of cardiovascular disease associated with kidney dysfunction was of a magnitude similar to that associated with traditional factors. Analyses of the mechanisms linking kidney dysfunction to cardiovascular disease were beyond the aims of this study. High UAE is considered a sign of endothelial damage secondary to uncontrolled cardiovascular risk factors. 1-3 For egfr, several factors could represent a link between a low GFR and increased cardiovascular risk (eg, activation of the renin-angiotensin system, anemia, and hyperhomocysteinemia). 1-3 The data support the view that high UAE and low egfr are separate entities in relation to cardiovascular risk because the 2 markers clustered in different groups and predicted cardiovascular disease independently of each other. The interaction between high UAE and low egfr in the prediction of cardiovascular risk suggests some interdependence between the underlying mechanisms, although the small number of individuals with high UAE and low egfr suggests that further investigation in larger samples is needed. Altogether these results agree with the concept that kidney dysfunction may be regarded as a quantifiable index of cardiovascular risk. The lack of a continuous relation across the whole range of UAE and egfr supports the view that the relation is nonlinear and significant only beyond a given threshold. 39 CONCLUSIONS This study shows that high UAE and low egfr do not run together in the population. The use of only 1 of these 2 markers underscores the potential to misclassify patients as having no kidney dysfunction, resulting in less intensive management of their disease. Also, the association of kidney dysfunction with cardiovascular disease is highlighted because high UAE and low egfr mark different groups of persons with high cardiovascular risk. Further studies are needed to explore the factors responsible for the different expression of kidney dysfunction and the possible countermeasures. Accepted for Publication: October 15, Correspondence: Massimo Cirillo, MD, Nefrologia (Ed 17) Policlinico, via Sergio Pansini 5, Naples, Italy (massimo.cirillo@unina2.it). Author Contributions: Study concept and design: Cirillo, Laurenzi, Mancini, and Zanchetti. Acquisition of data: Cirillo, Lanti, Laurenzi, and De Santo. Analysis and interpretation of data: Cirillo, Menotti, and De Santo. Drafting of the manuscript: Cirillo and Laurenzi. Critical revision of the manuscript for important intellectual content: Cirillo, Lanti, Menotti, Mancini, Zanchetti, and De Santo. Statistical analysis: Cirillo, Lanti, and Menotti. Obtained funding: Laurenzi. Administrative, technical, and material support: Laurenzi. Study supervision: Mancini, Zanchetti, and De Santo. Financial Disclosure: None reported. Funding/Support: This study was supported by Merck Sharp & Dohme Italy for study design, implementation, and data collection. Additional Contributions: The Gubbio Study was made possible thanks to the people and the local authorities of Gubbio in collaboration with Merck Sharp & Dohme Italy; Center for Preventive Medicine, Gubbio; Gubbio Hospital; Federico II University; University of Milan; Northwestern University of Chicago, Illinois; Istituto Superiore di Sanità, Rome, Italy; and Second University of Naples. REFERENCES 1. Sarnak MJ, Levey AS, Schoolwerth AC, 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(17): Levey AS, Coresh J, Balk E, et al; National Kidney Foundation. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification [published correction appears in Ann Intern Med. 2003;139(7):605]. Ann Intern Med. 2003;139(2): European Society of Hypertension European Society of Cardiology Guidelines Committee European Society of Hypertension European Society of Cardiology guidelines for the management of arterial hypertension [published corrections appear in J Hypertens. 2003;21(11): and J Hypertens. 2004;22(2):435]. J Hypertens. 2003;21(6): Culleton BF, Larson MG, Parfrey PS, Kannel WB, Levy D. Proteinuria as a risk factor for cardiovascular disease and mortality in older people: a prospective study. Am J Med. 2000;109(1): Madison JR, Spies C, Schatz IJ, et al. Proteinuria and risk for stroke and coronary heart disease during 27 years of follow-up: the Honolulu Heart Program. Arch Intern Med. 2006;166(8): Halbesma N, Kuiken DS, Brantsma AH, et al. Microalbuminuria is a better risk marker than low estimated GFR to identify individuals at risk for accelerated GFR loss in population screening. J Am Soc Nephrol. 2006;17(9): Hillege HL, Fidler V, Diercks GFH, et al; Prevention of Renal and Vascular End Stage Disease (PREVEND) Study Group. Urinary albumin excretion predicts cardiovascular and noncardiovascular mortality in general population. Circulation. 2002;106(14): Yuyun MF, Khaw K-T, Luben R, et al. A prospective study of microalbuminuria and incident coronary heart disease and its prognostic significance in a British population. Am J Epidemiol. 2004;159(3): Wang TJ, Gona P, Larson MG, et al. Multiple biomarkers for the prediction of first major cardiovascular events and death. N Engl J Med. 2006;355(25): Wannamethee SG, Shaper AG, Perry IJ. Serum creatinine concentration and risk of cardiovascular disease: a possible marker for increased risk of stroke. Stroke. 1997;28(3): Culleton BF, Larson MG, Wilson PW, Evans JC, Parfrey PS, Levy D. Cardiovascular disease and mortality in a community-based cohort with mild renal insufficiency. Kidney Int. 1999;56(6): Henry RM, Kostense PJ, Bos G, et al. Mild renal insufficiency is associated with increased cardiovascular mortality: the Hoorn Study. Kidney Int. 2002;62(4): Garg AX, Clark WF, Haynes RB, House AA. Moderate renal insufficiency and the risk of cardiovascular mortality: results from the NHANES I. Kidney Int. 2002; 61(4): Jurkovitz CT, Abramson JL, Vaccarino V, Weintraub WS, McClellan WM. Asso- 623

8 ciation of high serum creatinine and anemia increases the risk of coronary events: results from the Prospective Community-Based Atherosclerosis Risk in Communities (ARIC) Study. J Am Soc Nephrol. 2003;14(11): Fried LF, Shlipak MG, Crump C, et al. Renal insufficiency as a predictor of cardiovascular outcomes and mortality in elderly individuals. J Am Coll Cardiol. 2003; 41(8): Abramson JL, Jurkovitz CT, Vaccarino V, Weintraub W, McClellan W. Chronic kidney disease, anemia, and incident stroke in a middle-aged, communitybased population: the ARIC Study. Kidney Int. 2003;64(2): Manjunath G, Tighiouart H, Coresh J, et al. Level of kidney function as a risk factor for cardiovascular outcomes in the elderly. Kidney Int. 2003;63(3): Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004; 351(13): Shlipak MG, Fried LF, Cushman M, et al. Cardiovascular mortality risk in chronic kidney disease. JAMA. 2005;293(14): Meisinger C, Döring A, Löwel H; KORA Study Group. Chronic kidney disease and risk of incident myocardial infarction and all-cause and cardiovascular disease mortality in middle-aged men and women from the general population. Eur Heart J. 2006;27(10): Weiner DE, Tabatabai S, Tighiouart H, et al. Cardiovascular outcomes and allcause mortality: exploring the interaction between CKD and cardiovascular disease. Am J Kidney Dis. 2006;48(3): Laurenzi M, Cirillo M, Angeletti M, et al. Gubbio Population Study: baseline findings. Nutr Metab Cardiovasc Dis. 1991;1(4)(suppl):S1-S Laurenzi M, Cirillo M, Panarelli W, et al. Baseline sodium-lithium countertransport and 6-year incidence of hypertension: the Gubbio Population Study. Circulation. 1997;95(3): Cirillo M, Senigalliesi L, Laurenzi M, et al. Microalbuminuria in non-diabetic adults: relation of blood pressure, body mass, plasma cholesterol, and smoking: the Gubbio Population Study. Arch Intern Med. 1998;158(17): Cirillo M, Laurenzi M, Panarelli P, Mancini M, Zanchetti A, De Santo NG. Relation of urinary albumin excretion to coronary heart disease and low renal function: role of blood pressure. Kidney Int. 2004;65(6): Cirillo M, Laurenzi M, Mancini M, Zanchetti A, De Santo NG. Low muscular mass and overestimation of microalbuminuria by urinary albumin/creatinine ratio. Hypertension. 2006;47(1): Cirillo M, Laurenzi M, Mancini M, Zanchetti A, De Santo NG. Low glomerular filtration in the population: prevalence, associated disorders, and awareness. Kidney Int. 2006;70(4): Rose GA, Blackburn H, Gillman RF, Prineas RJ. Cardiovascular Survey Methods. Geneva, Switzerland: World Health Organization; Cirillo M, Anastasio P, Santo NG. Relationship of gender, age, and body mass index to error in predicted kidney function. Nephrol Dial Transplant. 2005;20 (9): 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(12): Mattix HJ, Hsu CY, Shaykevich S, Curhan G. Use of the albumin/creatinine ratio to detect microalbuminuria: implications of sex and race. J Am Soc Nephrol. 2002; 13(4): Cirillo M, Laurenzi M, Trevisan M, Stamler J. Hematocrit, blood pressure, and hypertension: the Gubbio Population Study. Hypertension. 1992;20(3): Anderson KM, Wilson PW, Odell PM, Kannel WB. An updated coronary risk profile: a statement for health professionals. Circulation. 1991;83(1): Sheifer SE, Manolio TA, Gersh BJ. Unrecognized myocardial infarction. Ann Intern Med. 2001;135(9): Boland LL, Folsom AR, Sorlie PD, et al. Occurrence of unrecognized myocardial infarction in subjects aged 45 to 65 years (the ARIC study). Am J Cardiol. 2002; 90(9): Cohn PF, Fox KM, Daly C. Silent myocardial ischemia. Circulation. 2003;108(10): Coresh J, Astor BC, McQuillan G, et al. 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(5): 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(3): Hostetter TH. Chronic kidney disease predicts cardiovascular disease. N Engl J Med. 2004;351(13):