Metabolic Syndrome and the Development of CKD in American Indians: The Strong Heart Study

Metabolic Syndrome and the Development of CKD in American Indians: The Strong Heart Study Jaime Lucove, MSPH,1 Suma Vupputuri, PhD,1,2 Gerardo Heiss, MD, PhD,1 Kari North, PhD,1 and Marie Russell, MD3 Background: Metabolic impairments that precede type 2 diabetes, such as metabolic syndrome, may contribute to the development of chronic kidney disease (CKD). This study documents the prevalence and incidence of CKD and the prospective association between metabolic syndrome and CKD in American Indians without diabetes in the Strong Heart Study. Study Design: Prospective cohort study. Setting & Participants: American Indians aged 45 to 74 years from 3 geographic regions were recruited by using tribal records and were assessed every 3 years from 1989 to 1999 as part of the Strong Heart Study. Participants with type 2 diabetes, on dialysis therapy, or who received a kidney transplant at baseline examination were excluded. Predictor: Metabolic syndrome, defined using Adult Treatment Panel III criteria. Outcomes & Measurements: CKD was measured by using estimated glomerular filtration rate (eGFR) and urinary albumin-creatinine ratio (ACR) dichotomized at conventional cutoff values. The association between metabolic syndrome and incident CKD was evaluated by using multivariable Cox proportional hazards models and binomial regression, with statistical adjustment for confounders (age, sex, study center, education, and smoking). Results: Metabolic syndrome was present in 896 (37.7%) and absent in 1,484 participants (62.3%) at baseline. The prevalence of ACR of 30 mg/g or greater at baseline examination was 12.1%, with 290 new cases and an incidence of 233/10,000 person-years. The prevalence of eGFR less than 60 mL/min/1.73 m2 was 7.8%, with 189 new cases and an incidence of 138/10,000 person-years. The prevalence of CKD was 17.8%, with 388 new cases and an incidence of 342/10,000 person-years. The adjusted hazard ratio for CKD associated with metabolic syndrome was 1.3 (95% confidence interval [CI], 1.1 to 1.6). Equivalent hazard ratios for ACR greater than 30 mg/g and eGFR less than 60 mL/min/1.73 m2 were 1.4 (95% CI, 1.0 to 1.9) and 1.3 (95% CI, 1.0 to 1.6), respectively. The relationship between metabolic syndrome and kidney outcomes was stronger in those who developed diabetes during follow-up. Limitations: Intraindividual variability in serum creatinine and ACR measures may have resulted in some misclassification of participants by outcome status. Conclusions: Metabolic syndrome is associated with an increased risk of developing CKD in American Indians without diabetes. The mechanism through which metabolic syndrome may cause CKD in this population likely is the development of diabetes. Am J Kidney Dis 51:21-28. © 2007 by the National Kidney Foundation, Inc. INDEX WORDS: Metabolic syndrome; glomerular filtration rate; albuminuria; chronic kidney disease; American Indians.

he prevalence of end-stage renal disease (ESRD) continues to increase in the United States and represents a growing clinical and public health problem.1,2 American Indians experience a greater burden of ESRD than whites. The prevalence of ESRD in American Indians is

T

2.5 times that of whites.1 Although it is difficult to attribute causes of ESRD, diabetes and hypertension are estimated to account for 44% and 27% of incident ESRD, respectively.1 American Indians also experience a disproportionate burden of type 2 diabetes mellitus.3

From the 1Department of Epidemiology, University of North Carolina-Chapel Hill; 2University of North Carolina Kidney Center, Chapel Hill, NC; and 3Medstar Research Institute, Phoenix, AZ. Received January 23, 2007. Accepted in revised form September 27, 2007. Originally published online as doi: 10.1053/j.ajkd.2007.09.014 on November 28, 2007. This work was performed at the University of North Carolina-Chapel Hill and all Strong Heart Study Centers: Medstar Research Institute, Aberdeen Area In-

dian Health Service, Aberdeen Area Tribal Chairmen’s Health Board, Center for American Indian Health Research, and University of Oklahoma Health Sciences Center. Address correspondence to Jaime C. Lucove, MSPH, Research Scientist, Allscripts, 27 East Delaware Ave, Pennington, NJ. E-mail: [email protected] © 2007 by the National Kidney Foundation, Inc. 0272-6386/07/5101-0004$34.00/0 doi:10.1053/j.ajkd.2007.09.014

American Journal of Kidney Diseases, Vol 51, No 1 (January), 2008: pp 21-28

21

22

Chronic kidney disease (CKD), measured by using urinary albumin-creatinine ratio (ACR) or glomerular filtration rate (GFR),4 is present in 19.2 million Americans, produces no overt signs and symptoms,5 and is associated with cardiovascular complications.6 CKD shares many of the same risk factors as cardiovascular disease, such as high blood pressure,5,7,8 diabetes mellitus,5,8,9 low high-density lipoprotein (HDL) cholesterol level,10 and smoking. These factors also overlap with factors that define metabolic syndrome, a cluster of abnormalities of insulin and/or glucose levels, lipid metabolism, hypertension, and abdominal adiposity, present in 24% of US adults.11,12 Although it is well-established that overt diabetes is a major risk factor for kidney function decrease, it is unclear whether early metabolic changes in patients without diabetes, such as those associated with metabolic syndrome, are associated with the development of CKD. Some evidence shows a connection between metabolic syndrome and the development of CKD. A study based on the Atherosclerosis Risk in Communities cohort reported that in patients without diabetes, the 9-year odds of developing a GFR less than 60 mL/min/1.73 m2 (⬍1.0 mL/s/ 1.73 m2) was 1.43 (95% confidence interval [CI], 1.18 to 1.73) for participants with metabolic syndrome compared with those without.13 After adjusting for the development of diabetes during the course of the study, the odds ratio remained increased (1.24; 95% CI, 1.01 to 1.51). A study performed in the Données Epidémiologiques sur le Syndrome d’Insulino-Résistance cohort observed a prospective association between metabolic syndrome and microalbuminuria in patients without diabetes.14 Cross-sectional studies also supported these relationships.15-17 The association between metabolic syndrome and CKD was not evaluated in an American Indian population, although it is well known that overt diabetes is a strong risk factor for kidney disease in this population.1 The purpose of this study is to examine the relationship between metabolic syndrome and CKD in American-Indian participants without diabetes of the Strong Heart Study.18 In this report, we describe the burden of CKD in this population assessed by using 2 measures: estimated GFR (eGFR) and ACR. We also estimate

Lucove et al

the prospective association between metabolic syndrome and incident CKD. METHODS Study Population The Strong Heart Study population included 4,549 American Indians aged 45 to 74 years from 13 tribes and was described previously.19,20 Eligible participants were identified by 13 tribal governments located in 3 geographic regions and had to be legal residents of those communities. All eligible participants were recruited to participate in the study. The tribes are located in 3 geographically defined study areas (centers): Oklahoma, South and North Dakota, and Arizona. Phase I (baseline) examinations took place between 1989 and 1991; 2 subsequent examinations of the original cohort were performed (phase II, 1993-1995, and phase III, 1998-1999). The participation rate in phase I was 62% overall and ranged from 55% to 72% across centers.19,20 Participants varied from nonparticipants in that they were more likely to be women and hypertensive and less likely to smoke. They did not differ by age, body mass index, or self-reported diabetes status.20 Examinations consisted of a personal interview and medical history, fasting blood draw, urine specimen, and physician-administered clinical examination that included anthropometric measurements.18 Participants who at baseline reported being on kidney dialysis therapy (n ⫽ 53) and those who previously received a kidney transplant (n ⫽ 7) were excluded from these analyses. Participants who had type 2 diabetes at baseline (defined as either fasting glucose ⱖ 126 mg/dL [ⱖ7.0 mmol/L] or use of hypoglycemic agents)21 were also excluded (n ⫽ 2,128) because the aim of the analysis is to evaluate the effect of metabolic syndrome in patients without diabetes. This resulted in a baseline study population of 2,420, which was 53.2% of the original cohort. The end of follow-up was the date of the phase III examination or the censoring date. Mean follow-ups for participants included in the ACR, eGFR, and CKD analyses were 6.7, 6.9, and 6.5 years, respectively. Of the baseline study population, 439 participants were lost to follow-up by visit 2; an additional 156 were lost to follow-up by visit 3.

Study Measures Details of the clinical and laboratory measures and qualitycontrol techniques used in the Strong Heart Study were described elsewhere18,19,22; variables used in this analysis are briefly described here. Abdominal circumference was measured at the level of the umbilicus with the participant in the supine position. Blood pressure was measured as the mean of the second and third readings, consistent with the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure-V protocol.23 The following measures were collected during the interview by certified nurses and/or trained study staff: smoking status, current medication use for diabetes and hypertension, history of kidney transplantation or dialysis therapy, and birth date (age). The following laboratory values were obtained from a 12-hour overnight fasting blood sample: cholesterol, triglyc-

Metabolic Syndrome and CKD erides, fasting glucose, HDL cholesterol, and plasma creatinine. Urinary albumin and creatinine values were obtained from a random urine specimen. Technical error was calculated for blind duplicate laboratory measurements and was 10.2% for triglycerides, 9.5% for HDL cholesterol, 8.7% for urinary creatinine, and 6.4% for urinary albumin.20 Metabolic syndrome was the main exposure for this study and was defined by using the National Cholesterol Education Program criteria of 3 or more of the following 5 factors at the phase I examination: waist circumference greater than 40 inches in men or greater than 35 inches in women, serum triglyceride level of 150 mg/dL or greater (ⱖ1.7 mmol/L), HDL cholesterol level less than 50 mg/dL (⬍1.3 mmol/L) in women or less than 40 mg/dL (⬍1.0 mmol/L) in men, systolic blood pressure of 130 mm Hg or greater or diastolic blood pressure of 85 mm Hg or greater, and fasting glucose level of 110 mg/dL or greater (ⱖ6.1 mmol/L).12 This definition of metabolic syndrome was used in previous analyses of this cohort.22 Individuals administered antihypertensive medication were considered to have the “hypertension” factor, and individuals administered antidiabetic medication were considered to have the “fasting glucose” factor. Two clinically meaningful kidney measures were created. Albuminuria was defined as the presence of ACR of 30 mg/g or greater.24 Decreased kidney function was defined as an eGFR less than 60 mL/min/1.73 m2 (⬍1.0 mL/s/1.73 m2), which was estimated by using the 4-variable isotope dilution mass spectrometry Modification of Diet in Renal Disease Study equation25: eGFR (mL ⁄ min ⁄ 1.73 m2) ⫽ 186.3 ⫻ 共Serum creatinine兲⫺1.154 ⫻ 共Age兲⫺0.203 ⫻ 共0.742 if female兲 ⫻ 共1.210 if African American兲 CKD was defined as the presence of either the ACR or the eGFR outcome.4

Statistical Analyses For the 3 outcomes, baseline prevalence, cumulative incidence, incidence rate, and 95% CIs were calculated. For longitudinal analyses, participants with the outcome of interest at baseline (ACR ⱖ 30 mg/g for the ACR analysis, eGFR ⬍ 60 mL/min/1.73 m2 [⬍1.0 mL/s/1.73 m2] for the eGFR analysis, and either outcome for the CKD analysis) were excluded. This resulted in slightly different study populations for the ACR, eGFR, and CKD analyses. Some participants (N ⫽ 137) in the ACR analysis were excluded from the eGFR analysis, and some participants in the eGFR analysis (N ⫽ 238) were excluded from the ACR analysis. To make populations more comparable and decrease concerns about confounding of the eGFR analysis by baseline ACR, results of eGFR analyses were examined after excluding participants with ACR of 30 mg/g or greater at baseline. Baseline characteristics were calculated as mean ⫾ SD for continuous variables and frequencies and percents for categorical variables. Baseline characteristics were examined for their relationship to metabolic syndrome and kidney function outcomes. Cox proportional hazards regression was used to estimate the hazard ratio (HR) and 95% CI for the relationship

23 between metabolic syndrome and time to CKD. Only first events were considered. Time to event was calculated as the midpoint of the interval when the outcome was first observed and the previous visit. If the outcome was observed at visit 2, the time used was the midpoint of the interval between visits 1 and 2. If the outcome was observed at visit 3, the time used was the midpoint of the interval between visits 2 and 3. If the outcome was not observed, all available time was used. If a participant was lost to follow-up before the study was completed, their last known value was used. There was sufficient variation in outcomes; however, events were clustered around visit dates. The efron statement was included in regression models to adjust for clustering. The proportional hazard assumption was verified visually, using plots of the log (–log) survival curves, and was met. Binomial regression models were constructed to estimate risk differences to describe the absolute risk of CKD attributable to metabolic syndrome. Covariates considered as potential confounders (age, sex, center, smoking, and education) were included in multivariable models. Model 1 was unadjusted, model 2 included age and sex, and model 3 included age, sex, center, education, and smoking. Model 3 was stratified by the development of diabetes during follow-up (data not shown). For the eGFR analysis, model 4 includes the same covariates as model 3, but was performed in participants with baseline ACR less than 30 mg/g. For all outcomes, adjustment was also made for baseline kidney function (measured by using continuous ACR and continuous eGFR). Analyses were performed using SAS, version 8 for Windows (SAS Institute, Cary, NC).

RESULTS

Characteristics of study participants at baseline are listed in Table 1, stratified by the main exposure, metabolic syndrome at baseline. Missing data were minimal for covariates or the main exposure (⬍2%). Characteristics, such as baseline age, smoking status, eGFR, and ACR, as well as center, sex, and the development of diabetes during follow-up, varied by metabolic syndrome status (P ⬍ 0.05). Prevalence and incidence estimates for outcome measures are listed in Table 2. Baseline prevalence of ACR of 30 mg/g or greater was 12.1% (95% CI, 10.8 to 13.4) and was more common in those with (15.7%) than without metabolic syndrome (10.3%; data not shown). Baseline prevalence of eGFR less than 60 mL/min/1.73 m2 (⬍1.0 mL/s/ 1.73 m2) was 7.8% (95% CI, 6.8 to 9.0) and did not vary by metabolic syndrome status. Baseline prevalence of the CKD outcome was 17.8% (95% CI, 16.3 to 19.4) and was also more common in those with (24.3%) than without metabolic syndrome (13.8%; data not shown). Incidence estimates were measured by using

24

Lucove et al

Table 1. Baseline Characteristics of Nondiabetic Strong Heart Study Participants by Metabolic Syndrome Status

Characteristic

Age (y) 45-54 55-64 ⱖ65 Sex Women Center Arizona Oklahoma North and South Dakota Smoking status Current Former Never Educational status ⬍High school (v ⱖ high school) Develop diabetes over study period Develop (v do not develop) Components of metabolic syndrome* Large waist circumference Increased blood pressure Increased fasting glucose Low HDL cholesterol High triglycerides Estimated GFR (mL/min per 1.73 m2) ACR (mg/g)

Baseline Population (n ⫽ 2,420)

Metabolic Syndrome (n ⫽ 896)

No Metabolic Syndrome (n ⫽ 1,484)

53 30 17

51 29 20

55 30 15

56

66

51

22 38 41

26 40 34

19 37 44

43 27 29

34 34 33

44 29 27

42

43

41

23

30

14

65 45 25 44 26 82 ⫾ 18 54 ⫾ 386

92 69 49 80 53 80 ⫾ 18 85 ⫾ 555

49 31 11 30 9 83 ⫾ 17 36 ⫾ 238

P

0.01

⬍0.001 ⬍0.001

⬍0.001

0.6 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001 ⬍0.01 0.01

Note: Values expressed as column percent for categorical variables and mean ⫾ SD for continuous variables. To convert GFR in mL/min/1.73 m2 to mL/s/1.73 m2, multiply by 0.01667. Abbreviations: ACR, albumin-creatinine ratio; GFR, glomerular filtration rate; HDL, high-density lipoprotein. *Metabolic syndrome defined as having 3 or more of the metabolic syndrome components. See Methods section for additional detail.

cumulative incidence and incidence rate and were approximately 65% greater for the ACR outcome than the eGFR outcome. Table 3 lists HRs and 95% CIs from the Cox proportional hazards models and risk differences and 95% CIs from binomial regression models.

Outcome-free survival was better for those without versus those with metabolic syndrome throughout follow-up for all outcomes. In models adjusted for age, sex, education, center, and smoking (model 3), the incidence of ACR of 30 mg/g or greater in participants with metabolic

Table 2. Prevalence and Incidence of CKD

Prevalence and Incidence

ACR ⱖ 30 mg/g (N ⫽ 2,420)

eGFR ⬍ 60 mL/min/1.73 m2 (N ⫽ 2,420)

CKD (N ⫽ 2,420)

No. of prevalent cases Prevalence (%) No. of incident cases Cumulative incidence (%) Person time (person-years) Incidence rate (/10,000 person-years)

290 12.1 (10.8-13.4) 286 15.6 (14.0-17.3) 12,274 233 (206-260)

189 7.8 (6.8-9.0) 187 9.6 (8.4-11.0) 13,514 138 (116-158)

428 17.8 (16.3-19.4) 388 22.1 (20.2-24.0) 11,367 342 (308-376)

Note: Values expressed as percent (95% confidence interval) unless noted otherwise. CKD defined as the presence of either ACR of 30 mg/g or greater or eGFR less than 60 mL/min/1.73 m2. To convert GFR in mL/min/1.73 m2 to mL/s/1.73 m2, multiply by 0.01667. Abbreviations: ACR, albumin-creatinine ratio; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease.

Metabolic Syndrome and CKD

25

Table 3. Hazard Ratios and Risk Differences for the Association Between Metabolic Syndrome and CKD Description of Model Covariates

ACR ⱖ 30 mg/g during follow-up Model 1: unadjusted Model 2: adjusted for age, sex Model 3: adjusted for age, sex, center, education, smoking eGFR ⬍ 60 mL/min/1.73 m2 during follow-up Model 1: unadjusted Model 2: adjusted for age, sex Model 3: adjusted for age, sex, center, education, smoking Model 4: model 3 for participants with ACR ⬍ 30 mg/g at baseline CKD Model 1: unadjusted Model 2: adjusted for age, sex Model 3: adjusted for age, sex, center, education, smoking

Hazard Ratio (95% CI)

Risk Difference (95% CI)

1.3 (1.0-1.6) 1.3 (1.0-1.6) 1.3 (1.0-1.6)

385 (25-745) 329 (⫺35-694) 384 (29-740)

1.6 (1.2-2.2) 1.4 (1.0-1.9) 1.4 (1.0-1.9) 1.4 (1.0-1.9)

465 (177-754) 267 (18-516) 279 (⫺19-577) 247 (⫺136-630)

1.4 (1.1-1.7) 1.3 (1.1-1.6) 1.3 (1.1-1.6)

661 (238-1,084) 492 (75-910) 516 (98-934)

Note: CKD defined as the presence of either ACR of 30 mg/g or greater or eGFR less than 60 mL/min/1.73 m2. Abbreviations: ACR, albumin-creatinine ratio; eGFR, estimated glomerular filtration rate; CKD, chronic kidney disease; CI, confidence interval.

syndrome at baseline was 1.3 times the incidence in those without metabolic syndrome at baseline (95% CI, 1.0 to 1.6). Adjusting for the same covariates, there were 384 more cases of ACR of 30 mg/g or greater per 10,000 person-years in those with compared with those without metabolic syndrome at baseline (95% CI, 29 to 740). The strength of the association observed between metabolic syndrome and the eGFR outcome was slightly stronger in magnitude, but not statistically significant. In similarly adjusted models for CKD, persons with metabolic syndrome showed 1.3 times greater risk of developing CKD than those without metabolic syndrome (95% CI, 1.1 to 1.6), and there were 516 more cases of CKD per 10,000 person-years in those with than without metabolic syndrome (95% CI, 98 to 934). On adjustment for both baseline continuous kidney function measures, Cox proportional hazards model results did not noticeably change: HRs were 1.3 (95% CI, 1.0 to 1.7) for the ACR model, 1.4 (95% CI, 1.0 to 1.9) for the eGFR model, and 1.3 (95% CI, 1.1 to 1.6) for the CKD model. In binomial regression models, adjustment for the baseline continuous kidney function measures resulted in model convergence problems (results not shown). Results listed in Table 3 were also calculated using the International Diabetes Federation (IDF) definition of metabolic syndrome, which requires the presence of a large waist circumference.26 Using the IDF definition of metabolic

syndrome, results did not vary significantly (data not shown). For example, the HR for CKD associated with metabolic syndrome using the Adult Treatment Panel III definition was 1.38 (95% CI, 1.1 to 1.7), whereas using the IDF definition, it was 1.3 (95% CI, 1.1 to 1.6). These models were also stratified by the development of new diabetes during follow-up. The relationship between metabolic syndrome and kidney function outcomes (measured by means of either HR or risk difference) increased in those who developed diabetes and was virtually absent for those who did not develop diabetes (data not shown). These results were not statistically significant. For example, the HR for the relationship between CKD and metabolic syndrome in those who developed diabetes was 1.4 (95% CI, 1.0 to 2.1), whereas in those who did not, it was 1.0 (95% CI, 0.8 to 1.4). When diabetes development was considered as a timedependent covariate, its inclusion reduced the relationship between kidney function measures and metabolic syndrome. For example, the HR for the relationship between CKD and metabolic syndrome, adjusting for diabetes development as a time-dependent covariate, was 1.1 (95% CI, 0.9 to 1.3). The contribution of individual metabolic syndrome components was examined in models adjusting for age, sex, center, and smoking (data not shown). Increased blood pressure showed the strongest effect on incident CKD (HR, 1.5; 95%

26

Lucove et al

CI, 1.2 to 1.9). A strong association with CKD was also observed for increased fasting glucose level (HR, 1.4; 95% CI, 1.1 to 1.8). Low HDL cholesterol level showed HRs that were positively associated with CKD, but not statistically significant (HR, 1.2; 95% CI, 0.9 to 1.5). Large waist circumference and increased triglyceride level were associated inversely with CKD, but the magnitudes were weak and not statistically significant. DISCUSSION

This study found that metabolic syndrome was significantly and independently associated with a 30% increased risk of incident CKD during 9 years of follow-up in a population of American Indians without diabetes. This risk was greater for persons who developed diabetes during the follow-up period compared with those who did not develop diabetes. Limited information exists describing the prevalence and incidence of CKD in American Indians,1 and much of that information applies to diabetic populations.27-30 In the general US population, previous studies indicated that microalbuminuria was present in 17% to 23% of patients with newly diagnosed diabetes31-33 and 9% of those who had impaired fasting glucose.25 The prevalence of GFR less than 60 mL/min/1.73 m2 (⬍1.0 mL/s/1.73 m2) in a nationally representative sample of the US population older than 20 years was 4.7%.5 The prevalence in 60- to 69year-olds (a group similar in age to that studied here) was 10.5%.5 It is difficult to directly compare prevalence and incidence found in this study with other studies because of differences in race, inclusion criteria, definitions of kidney function, and age distributions. The last difference is particularly important because kidney function varies markedly with age.5 Our finding that metabolic syndrome was associated with incident CKD agrees with previous findings13,14 despite differences in follow-up times, populations, and methods. Unlike previous studies that evaluated these relationships, the current study was prospective, included 9 years of follow-up, and analyzed 3 kidney function measures: eGFR, ACR, and CKD. The magnitude of the association between metabolic syndrome and both measures of CKD was stronger for those who developed diabetes during fol-

low-up compared with those who did not. This finding may indicate commonalities in the precursors of diabetes and CKD, such as insulin resistance. Conversely, those who develop diabetes may be more likely to develop diabetic kidney disease during the subclinical phase of diabetes. Although insulin resistance is postulated by most to be the defining trait of metabolic syndrome, its absence is not inconsistent with metabolic syndrome. Other potential mechanisms include inflammation, hypertension, endothelial dysfunction, and increased uric acid levels.34-39 Although our analysis of individual metabolic syndrome components indicates that increased blood pressure (compared with other components) had the strongest relationship with the outcomes, more research is needed to determine whether one component has a stronger impact than others. The exclusion of patients with diabetes at baseline likely decreased the effect of the glucose and obesity factors. Additional research should take into consideration the interaction of these factors with each other and their persistence over time. This study has several strengths. Its prospective design allowed us to estimate the association of metabolic syndrome with incident CKD during a long follow-up period. Our use of 2 markers of subclinical CKD showed consistency in their association with metabolic syndrome. Finally, although it is known that American Indians experience high rates of ESRD, information about the early determinants of CKD in this population is limited. This study contributes new information about the role of a highly prevalent trait, such as metabolic syndrome, as an antecedent to kidney disease in this high-risk population. Among the limitations of this study, the following deserve mention. Differences between participants and nonparticipants may have resulted in selection bias, especially if selection factors were associated with the exposure status. ACR and eGFR are measures that show a high degree of intraindividual variability and thus require repeated measurements to ideally assess kidney function. For example, 3 readings greater than 30 mg/g are recommended for a diagnosis of microalbuminuria.24 The use of 3 repeated measurements (had they been available, as for other studies5) likely would have decreased the prevalence of albuminuria in the current study.

Metabolic Syndrome and CKD

According to a previous study, estimation bias was lower when the MDRD Study equation was used for American Indians who were calibrated as white as opposed to black; however, overestimation occurred.40 For the present study, we chose to calculate eGFR by using the MDRD Study equation and classified the American Indian population as white, resulting in a baseline prevalence of 7.8% (Table 2). When the baseline prevalence of eGFR was calculated for our study population using the MDRD Study equation and calibrating the population as black, the estimate was 2.5% (data not shown). The “true” prevalence likely lies between these numbers. Notably, calibrating this population as black or white in the MDRD Study estimating equation did not appreciably change HRs for the association between metabolic syndrome and CKD (data not shown). The equation used to estimate GFR in this study was validated in 2 studies of Pima Indians by using iothalamate as the gold-standard measurement.40,41 The equation performed reasonably well considering Pima Indians to be either white or black (Pearson r ⫽ 0.80 for both), with underestimation of GFR by 8% when participants were considered white and overestimation by 11% when participants were considered black.39 In comparing 4 indirect measures of GFR (MDRD Study equation, Cockcroft-Gault equation, 100/cystatin, and 100/creatinine) with iothalamate clearance in Pima Indians,40 all estimates showed upward bias. However, agreement was greatest for the MDRD Study equation and 100/cystatin C. Because creatinine assays may vary among different laboratories, to more accurately use the MDRD Study equation, it would have been appropriate to calibrate plasma creatinine values of the Strong Heart Study Laboratory with the Cleveland Clinic Laboratory (where serum creatinine was measured in the MDRD Study). A systematic difference in creatinine measurement between laboratories could have potentially affected the association between metabolic syndrome and CKD. In conclusion, in American Indians, early metabolic changes associated with increased blood pressure and/or insulin resistance appear to adversely influence temporal changes in kidney function measured by means of ACR and eGFR, mediated in part by the development of type 2

27

diabetes. This agrees with an analysis performed in the Atherosclerosis Risk in Communities cohort13 and other literature that suggest CKD often occurs before or closely after the development of diabetes.31-33 These findings suggest the need for additional research to assess whether screening for early signs of kidney disease in those at high risk of overt diabetes, such as those with metabolic syndrome, may identify persons with undiagnosed CKD. ACKNOWLEDGEMENTS The opinions expressed in this report are those of the authors and do not necessarily reflect the views of the Indian Health Service. Support: This study was supported by cooperative agreement Grants U01HL-41642, U01HL-41652, and U01HL41654 from the National Heart, Lung, and Blood Institute (NHLBI) and by NHLBI Cardiovascular Disease Epidemiology Training Grant 5-T32-HL007055-28. Financial Disclosure: None.

REFERENCES 1. US Renal Data System: USRDS Annual Data Report. The National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2004 2. Arias EAR, Kung HC, Murphy SL, Kochanek KD: Deaths: final data for 2001. Natl Vital Stat Rep 52:1-115, 2003 3. Lee ET, Howard BV, Savage PJ, et al: Diabetes and impaired glucose tolerance in three American Indian populations aged 45-74 years. The Strong Heart Study. Diabetes Care 18:599-610, 1995 4. Levey AS, Coresh J, Balk E, et al: National Kidney Foundation practice guidelines for chronic kidney disease: Evaluation, classification, and stratification. Ann Intern Med 139:137-147, 2003 5. 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 41: 1-12, 2003 6. Weiner DE, Tighiouart H, Stark PC, et al: Kidney disease as a risk factor for recurrent cardiovascular disease and mortality. Am J Kidney Dis 44:198-206, 2004 7. Whelton PK, Perneger TV, He J, Klag MJ: The role of blood pressure as a risk factor for renal disease: A review of the epidemiologic evidence. J Hum Hypertens 10:683-689, 1996 8. 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 14:2934-2941, 2003 9. Brancati FL, Whelton PK, Randall BL, Neaton JD, Stamler J, Klag MJ: Risk of end-stage renal disease in diabetes mellitus: A prospective cohort study of men screened

28 for MRFIT. Multiple Risk Factor Intervention Trial. JAMA 278:2069-2074, 1997 10. Muntner P, Coresh J, Smith JC, Eckfeldt J, Klag MJ: Plasma lipids and risk of developing renal dysfunction: The Atherosclerosis Risk in Communities Study. Kidney Int 58:293-301, 2000 11. Ford ES, Giles WH, Dietz WH: Prevalence of the metabolic syndrome among US adults: Findings from the Third National Health and Nutrition Examination Survey. JAMA 287:356-359, 2002 12. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 285:2486-2497, 2001 13. Kurella M, Lo JC, Chertow GM: Metabolic syndrome and the risk for chronic kidney disease among nondiabetic adults. J Am Soc Nephrol 16:2134-2140, 2005 14. Bonnet F, Marre M, Halimi JM, et al, for the DESIR Study Group: Waist circumference and the metabolic syndrome predict the development of elevated albuminuria in non-diabetic subjects: The DESIR Study. J Hypertens 24: 1157-1163, 2006 15. Hoehner CM, Greenlund KJ, Rith-Najarian S, Casper ML, McClellan WM: Association of the insulin resistance syndrome and microalbuminuria among nondiabetic native Americans. The Inter-Tribal Heart Project. J Am Soc Nephrol 13:1626-1634, 2002 16. Chen J, Muntner P, Hamm LL, et al: The metabolic syndrome and chronic kidney disease in US adults. Ann Intern Med 140:167-174, 2004 17. Chen J, Muntner P, Hamm LL, et al: Insulin resistance and risk of chronic kidney disease in nondiabetic US adults. J Am Soc Nephrol 14:469-477, 2003 18. Lee ET, Welty TK, Fabsitz R, et al: The Strong Heart Study. A study of cardiovascular disease in American Indians: Design and methods. Am J Epidemiol 132:1141-1155, 1990 19. Welty TK, Lee ET, Yeh J, et al: Cardiovascular disease risk factors among American Indians. The Strong Heart Study. Am J Epidemiol 142:269-287, 1995 20. Howard BV, Lee ET, Cowan LD, et al: Coronary heart disease prevalence and its relation to risk factors in American Indians. The Strong Heart Study. Am J Epidemiol 142:254-268, 1995 21. American Diabetes Association: Diagnosis and classification of diabetes mellitus. Diabetes Care 27:S5-S10, 2004 (suppl 1) 22. Resnick HE, Jones K, Ruotolo G, et al: Insulin resistance, the metabolic syndrome, and risk of incident cardiovascular disease in nondiabetic American Indians: The Strong Heart Study. Diabetes Care 26:861-867, 2003 23. The Fifth Report of the Joint National Committee on Detection, Evaluation, and Treatment of High Blood Pressure (JNC V). Arch Intern Med 153:154-183, 1993 24. American Diabetes Association: Standards of medical care for patients with diabetes mellitus. Diabetes Care 26:S33-S50, 2003 (suppl 1) 25. 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 equa-

Lucove et al tion. Modification of Diet in Renal Disease Study Group. Ann Intern Med 130:461-470, 1999 26. Alberti KG, Zimmet P, Shaw J, for the IDF Epidemiology Task Force Consensus Group: The metabolic syndrome—A new worldwide definition. Lancet 366:1059-1062, 2005 27. Nelson RG, Bennett PH, Beck GJ, et al: Development and progression of renal disease in Pima Indians with non-insulin-dependent diabetes mellitus. Diabetic Renal Disease Study Group. N Engl J Med 335:1636-1642, 1996 28. Nelson RG, Newman JM, Knowler WC, et al: Incidence of end-stage renal disease in type 2 (non-insulindependent) diabetes mellitus in Pima Indians. Diabetologia 31:730-736, 1988 29. Nelson RG, Kunzelman CL, Pettitt DJ, Saad MF, Bennett PH, Knowler WC: Albuminuria in type 2 (noninsulin-dependent) diabetes mellitus and impaired glucose tolerance in Pima Indians. Diabetologia 32:870-876, 1989 30. Tapp RJ, Shaw JE, Zimmet PZ, et al: Albuminuria is evident in the early stages of diabetes onset: Results from the Australian Diabetes, Obesity, and Lifestyle Study (AusDiab). Am J Kidney Dis 44:792-798, 2004 31. Sosenko JM, Hu D, Welty T, Howard BV, Lee E, Robbins DC: Albuminuria in recent-onset type 2 diabetes: The Strong Heart Study. Diabetes Care 25:1078-1084, 2002 32. Spijkerman AM, Dekker JM, Nijpels G, et al: Microvascular complications at time of diagnosis of type 2 diabetes are similar among diabetic patients detected by targeted screening and patients newly diagnosed in general practice: The Hoorn Screening Study. Diabetes Care 26:2604-2608, 2003 33. Kohler KA, McClellan WM, Ziemer DC, Kleinbaum DG, Boring JR: Risk factors for microalbuminuria in black Americans with newly diagnosed type 2 diabetes. Am J Kidney Dis 36:903-913, 2000 34. Sonnenberg GE, Krakower GR, Kissebah AH: A novel pathway to the manifestations of metabolic syndrome. Obes Res 12:180-186, 2004 35. Feldt-Rasmussen B: Microalbuminuria, endothelial dysfunction and cardiovascular risk. Diabetes Metab 26:S64S66, 2000 (suppl 4) 36. Greenberg AS: The expanding scope of the metabolic syndrome and implications for the management of cardiovascular risk in type 2 diabetes with particular focus on the emerging role of the thiazolidinediones. J Diabetes Complications 17:218-228, 2003 37. Kambham N, Markowitz GS, Valeri AM, Lin J, D’Agati VD: Obesity-related glomerulopathy: An emerging epidemic. Kidney Int 59:1498-1509, 2001 38. Hall JE, Crook ED, Jones DW, Wofford MR, Dubbert PM: Mechanisms of obesity-associated cardiovascular and renal disease. Am J Med Sci 324:127-137, 2002 39. Cirillo P, Sato W, Reungjui S, et al: Uric acid, the metabolic syndrome, and renal disease. J Am Soc Nephrol 17:S165-S168, 2006 (suppl 3) 40. Nelson RG, GT, Beck GJ, et al: Estimating GFR by the MDRD and Cockroft-Gault Equations in Pima Indians. J Am Soc Nephrol 14:134A, 2003 (abstr) 41. Perkins BA, Nelson RG, Ostrander BE, et al: Detection of renal function decline in patients with diabetes and normal or elevated GFR by serial measurements of serum cystatin C concentration: Results of a 4-year follow-up study. J Am Soc Nephrol 16:1404-1412, 2005