| 2007 JASN IMPACT FACTOR 7.111 | HOME AUTHOR INFO EDITORIAL BOARD SUBSCRIBE FEEDBACK ALERTS HELP | |||
| CURRENT ISSUE | ARCHIVES | JASN Express | ONLINE SUBMISSION | |

,

,||
,#
*Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland; Departments of
Epidemiology,
Health Policy and Management, and #Biostatistics, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland;
National Center for Health Statistics, Centers for Disease Control and Prevention, Hyattsville, Maryland; ||Department of Sociology, Johns Hopkins University, Baltimore, Maryland; and ¶National Institutes of Health, Office of Behavioral and Social Sciences Research, Bethesda, Maryland.
Correspondence to Dr. Frederick L. Brancati, Welch Center for Prevention, Epidemiology, and Clinical Research, The Johns Hopkins Medical Institutions, 2024 East Monument St., Suite 2-600, Baltimore, MD 21205. Phone: 410-955-9843; Fax: 410-955-0476; E-mail: fbrancat{at}jhmi.edu
| Abstract |
|---|
|
|
|---|
| Introduction |
|---|
|
|
|---|
Previous studies that sought to explain the African-American/white differences in chronic kidney disease (CKD) rates demonstrated several limitations. Some studies used case-control study designs, which limit causal inferences (4), or relied on ecologic exposure data for information that was not available at the individual level (3,7,1921). Other studies either used study populations that were not typical of the United States (3,11,2224) or relied on intermediate markers of kidney disease as outcomes (6). With these considerations in mind, we conducted a population-based, prospective study to determine how much of the excess risk of CKD among African Americans could be explained on the basis of racial disparities in potentially modifiable risk factors.
| Materials and Methods |
|---|
|
|
|---|
This analysis included African-American and white subjects who were 30 to 74 yr of age at the time of the baseline evaluation. Of these 9087 individuals, five who were receiving renal replacement therapy (i.e., listed in the Medicare Registry) at the time of the baseline examination were excluded, as were those with missing baseline data on family income (n = 369) or other key variables (n = 155). A similar percentage of those excluded developed CKD, compared with those included in the analysis (2.2% versus 1.3%, P = 0.16). After exclusions, 894 African-American and 7664 white adults remained for analysis.
Data Collection
The race of the participants in NHANES II was determined by the interviewers and was classified as "white," "African American," or "other races" (25). Questions regarding participant age, gender, marital status, education, and household income were also asked during the baseline interviews. To judge the resources available to families as a function of their total income, all household incomes (assessed using income categories, to increase the response likelihood) were expressed relative to the federal poverty thresholds for equally sized households. A poverty income ratio of <1.00 is below the federal poverty level ("poor"), a poverty income ratio of 1.00 to 1.99 is at or near the poverty level ("near poor"), and a poverty income ratio of
2.00 is 200% above the federal poverty level ("not poor") (28).
Participants were also asked two questions regarding their level of physical activity in leisure and nonleisure activities, i.e., (1) whether they were getting "much," "moderate," or "little" exercise for recreation and (2) whether in their usual day, aside from recreation, they were "very active," "moderately active," or "quite inactive." Individuals reporting much exercise in recreation and/or a very physically active day were classified as being "very active." Those reporting little recreational exercise and a "quite inactive" daily routine were categorized as being "inactive." All other combinations were considered to be "moderately active." The frequency of beer, wine, and liquor consumption was obtained from a food frequency questionnaire and was categorized as never, seldom (<1 time/wk), weekly (1 to 6 times/wk), or daily (
1 time/d). Smoking history at baseline was used to classify participants as never, former, or current smokers.
The body mass index was calculated from the measured height (in meters) and weight (in kilograms) and was categorized as not overweight (<25 kg/m2), overweight (25 to 29.9 kg/m2), or obese (
30 kg/m2) (29). BP was measured twice in the right arm, by using a standard mercury sphygmomanometer, with the subject seated. The mean of the two readings was used for analysis. Persons were considered hypertensive if they demonstrated a systolic BP of
140 mmHg or a diastolic BP of
90 mmHg or if they reported being told by a doctor that they had hypertension. Persons who responded affirmatively to the question, "Has a doctor ever told you that you had diabetes?" were classified as having diabetes mellitus. A history of cardiovascular disease (CVD) was defined on the basis of a self-report of previous heart failure, heart attack, or stroke.
Total serum cholesterol levels were measured according to the protocol described for the Lipid Research Clinics Program (30). Serum creatinine determinations were performed by using the Jaffe reaction, and GFR were estimated by using the Cockcroft-Gault creatinine clearance equation (31). Because the nonresponse rate for serum creatinine levels was 28%, these measurements were considered only in subsidiary analyses (25).
Ascertainment of Outcomes
Incident all-cause CKD was defined as either treatment for ESRD or death related to kidney disease. Treated ESRD cases from February 1976 through December 31, 1992, were ascertained from the Medicare ESRD Registry, by matching participants in NHANES II on the basis of their full name, birth date, gender, and race. Those who were matched to the registry perfectly or nearly perfectly with respect to all of the matching factors were classified as definite matches (n = 33), whereas those with slight differences with respect to any of the matching factors were classified as probable matches (n = 4). All those who did not match in multiple fields or who did not match with respect to race alone were classified as possible or poor matches (n = 46). In our analyses, only definite and probable matches were counted as treated ESRD events.
Kidney-related deaths were identified through 1992 in the NHANES II Mortality Study. Deaths were ascertained by using the National Death Index and the Social Security Administration Death Master Files. Decedents who were not listed in the Medicare ESRD Registry but for whom any of the following kidney-related International Classification of Diseases (9th Revision) codes designated the underlying or contributing cause of death were also counted as CKD events: 250.4 (diabetes mellitus with nephropathy), 274.1 (gouty nephropathy), 275.4 (nephrocalcinosis), 403 (hypertensive renal disease), 404 (hypertensive heart and renal disease), 580 to 589 (nephritis, nephrotic syndrome, or nephrosis), or 593.9 (renal disease not otherwise specified). This broad definition was used to include individuals who received Medicare-funded renal replacement therapy, those who might have received it before being enrolled in Medicare, and those who might have chosen not to initiate therapy.
Analyses
All statistical analyses accounted for the complex survey design, providing nationally representative estimates. By using weighted Poisson models in Stata 6.0 (32) and the age distribution of the entire NHANES II cohort, directly age-adjusted CKD incidence rates were calculated per 100,000 person-yr of risk. Cumulative lifetime CKD incidence rates were estimated by using a weighted life-table method, and the cumulative risks of the two race groups were compared by using log-rank tests. Baseline variables were grouped as (1) sociodemographic factors, including poverty status, educational attainment, and marital status; (2), lifestyle factors, including smoking status, physical activity, alcohol use, and body mass index; and (3) clinical factors, including diabetes mellitus, hypertension, cardiovascular diseases, systolic BP, and serum cholesterol levels. To assess relative differences in CKD incidence rates for African Americans, compared with whites, we conducted a series of five Cox proportional-hazards model analyses with Sudaan 7.5 (33), adjusting for (1) age (continuously) and gender; (2) age, gender, and sociodemographic factors; (3) age, gender, and lifestyle factors; (4) age, gender, and clinical factors; and (5) all of the aforementioned variables simultaneously. The percentage of excess risk of CKD for African Americans that was explained by a set of risk factors was calculated by using the formula
|
| (1) |
where RR1 is the relative risk (RR) of CKD for African Americans versus whites adjusted only for age and gender and RR2 is the RR adjusted for age, gender, and the set of risk factors (34,35).
To test the robustness of the associations, three subsidiary analyses were performed. We first excluded individuals who developed CKD within the first 5 yr of follow-up monitoring, to determine whether the same associations between race and CKD were present. Next, we analyzed the subpopulation with serum creatinine measurements, adjusting for GFR. Finally, we limited the outcome to treated ESRD only. Because analyses performed with treated ESRD as the outcome yielded mostly similar results (data not shown), only analyses with the combined end point of CKD (defined as either treated ESRD or death related to kidney disease) are presented.
| Results |
|---|
|
|
|---|
|
|
|
|
|
| Discussion |
|---|
|
|
|---|
Results of this study are consistent with previous findings of a higher risk of CKD among African Americans. Since 1977, when Easterling first documented a 3.8-fold higher risk of CKD among African Americans versus whites, using the Michigan ESRD registry, there have been many studies focusing on African-American/white differences in CKD (211,19,23). All of those studies have documented an excess risk for African Americans, compared with whites, with estimates of the association ranging from 1.9 to 7.4. Most of those studies have been limited by reliance on data from the United States Renal Data System or other ESRD registries, which lack information on potential explanatory factors (5,7,9,10,19,3639).
Four prospective studies used geographically aggregated measures of exposure as indirect estimates of individual-level measures (2,3,8,11). For example, Whittle et al. (8) used prevalence estimates of explanatory factors (such as hypertension and diabetes mellitus) obtained from a regional survey to examine the racial differences in CKD risk. They observed that adjustment for age, prevalence and severity of hypertension, diabetes mellitus, and level of education in the population reduced the risk of hypertensive ESRD in African-American areas, but the racial association remained strong and statistically significant.
Although they had individual measurements of chronic medical conditions and behaviors at baseline, Brancati et al. (11) and Klag et al. (3) lacked individual measurements of income for Multiple Risk Factor Intervention Trial screenees, which hindered accurate quantification of the effects of socioeconomic factors on the individual risk of kidney disease. Nevertheless, those groups both concluded that higher BP, lower income, and higher prevalence of diabetes mellitus and hypertension among African-American adults explained some of the racial differences in CKD risk, with a significant amount remaining unexplained.
Like Rostand et al. (36) and Lopes et al. (20,37,38), we observed that racial differences in CKD risk were modified by age, with middle-age adults (30 to 59 yr) exhibiting a greater racial disparity than older adults (60 to 74 yr). The reasons for this age interaction are unclear; however, previous authors have speculated regarding accelerated kidney damage attributable to poorer control of BP, lower potassium intake, sustained higher levels of psychologic stress, and underdeveloped kidneys (38).
In our study, the large residual excess risk observed could possibly be explained on the basis of suboptimal measurement of exposures in NHANES II. In particular, factors such as BP were measured on only one occasion, which could result in an underestimation of the strength of the attenuation between race and CKD. Additionally, better characterization of potentially modifiable risk factors such as diabetes mellitus (e.g., using hemoglobin A-1c levels) might have yielded greater reductions in the excess risk. By grossly assessing socioeconomic status (using education and poverty levels), we might have inadequately adjusted for the local environment or access to and quality of health care. It is possible that these factors in combination could account for all of the residual excess risk among African Americans.
Another potential explanation for the unexplained excess risk among African Americans might be unmeasured environmental, behavioral, sociocultural, or developmental factors that were beyond the scope of NHANES II data. Literature reports suggest that undernutrition in fetal life imparts a higher risk of CKD in adulthood (40,41). Because African Americans exhibit much higher rates of low birth weights, compared with whites (42), and low birth weights are associated with kidney underdevelopment, the low birth weight theory has been advanced to help explain the racial differences in CKD rates (13). Additionally, African Americans are more likely to be exposed to occupational and environmental toxins such as lead (43), to experience viral infections (44), and to have less access to preventive medical care, as well as being referred to treatment for CKD late in the course of their disease (45). Enhanced susceptibility of African-American kidneys to injury resulting from hypertension (46,47) and racial differences in renal vascular hemodynamics (48,49) have also been cited as explanations for the racial disparity in CKD risk.
In addition to previously mentioned limitations related to the assessment of exposures, other limitations deserve comment. First, because NHANES II was restricted to the noninstitutionalized population, the absolute risks of kidney disease among whites and especially African Americans were likely underestimated (15,50,51). Second, the NHANES II Mortality Study determined vital status only for adults of age 30 to 74 yr at baseline, limiting inferences to the middle-age United States population. Third, race was determined not by self-report but by interviewer assessment, leading to potential misclassification of African Americans, Hispanics, and Native Americans as whites. A number of studies have demonstrated discrepancies between the interviewer-observed race and the self-reported race, with approximately 6% of persons self-identifying as African American being classified as white (52,53). Fourth, renal function was not assessed at baseline times for all participants. Therefore, we could not completely establish whether renal insufficiency was already present in some of the subjects who later developed CKD.
Finally, the use of passive follow-up techniques without social security numbers to determine vital status and treatment for ESRD might have led to some misclassification of outcomes. Previous studies that evaluated the effectiveness of the National Death Index and the Social Security Administration Death Master Files suggested that underascertainment of vital status occurs more commonly for African Americans than for whites (27,54,55), whereas other work proposed that there is no racial difference in ascertainment (56). Additionally, whites are more likely to have multiple causes of death coded on their death certificates (57), which could increase the ascertainment of CKD events for whites while underestimating the number for African Americans. Because the majority of our events were deaths related to kidney disease, this differential ascertainment could significantly affect our results. In addition to deaths, only individuals treated through the Medicare program were included in our analyses. Therefore, possible racial differences in enrollment in Medicare versus private insurance could affect our findings. We might have underestimated the true racial disparity in CKD by relying on these methods. Lastly, our use of CKD attributable to any cause as the primary outcome instead of cause-specific CKD reduced the possibility of misclassification bias, which is known to be influenced by race (58,59).
In light of the exorbitant costs, to individuals and to society, of treatment for CKD, the identification of modifiable risk factors for CKD is an important public health priority, as indicated in Healthy People 2010. Our results suggest that interventions aimed at reducing the racial disparity in CKD risk should focus on primary prevention and improved treatment of diabetes mellitus and hypertension, lifestyle modification, and elimination of health disparities attributable to socioeconomic status. Our data further suggest that the benefits of these improvements might be greatest for middle-age adults. Further prospective studies are needed to identify novel environmental, developmental, and genetic causes of the large African-American/white disparity in the incidence of CKD.
| Acknowledgments |
|---|
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. Lambie, N. Richards, and S. Smith Ethnicity, age and incidence rates for renal replacement therapy (RRT) in Birmingham, UK: 1990-2004 Nephrol. Dial. Transplant., December 1, 2008; 23(12): 3983 - 3987. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Xu, L. Zhang, P. Zhang, F. Wang, L. Zuo, and H. Wang Comparison of the prevalence of chronic kidney disease among different ethnicities: Beijing CKD survey and American NHANES Nephrol. Dial. Transplant., November 5, 2008; (2008) gfn609v1. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. M. Salem, P. E. Cadman, Y. Chen, F. Rao, G. Wen, B. A. Hamilton, B. K. Rana, D. W. Smith, M. Stridsberg, H. J. Ward, et al. Chromogranin A Polymorphisms Are Associated With Hypertensive Renal Disease J. Am. Soc. Nephrol., March 1, 2008; 19(3): 600 - 614. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. L. Davis, W. E. Harmon, J. Himmelfarb, T. Hostetter, N. Powe, P. Smedberg, L. A. Szczech, P. S. Aronson, and for the American Society of Nephrology Public Poli World Kidney Day 2008: Think Globally, Speak Locally J. Am. Soc. Nephrol., March 1, 2008; 19(3): 413 - 416. [Full Text] [PDF] |
||||
![]() |
N. Volkova, W. McClellan, M. Klein, D. Flanders, D. Kleinbaum, J. M. Soucie, and R. Presley Neighborhood Poverty and Racial Differences in ESRD Incidence J. Am. Soc. Nephrol., February 1, 2008; 19(2): 356 - 364. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. L. Schwartz, K. R. Bailey, T. Mosley, D. S. Knopman, C. R. Jack Jr, V. J. Canzanello, and S. T. Turner Association of Ambulatory Blood Pressure With Ischemic Brain Injury Hypertension, June 1, 2007; 49(6): 1228 - 1234. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. L. Xue, P. W. Eggers, L. Y. Agodoa, R. N. Foley, and A. J. Collins Longitudinal Study of Racial and Ethnic Differences in Developing End-Stage Renal Disease among Aged Medicare Beneficiaries J. Am. Soc. Nephrol., April 1, 2007; 18(4): 1299 - 1306. [Abstract] [Full Text] [PDF] |
||||
![]() |
B. B. Newsome, W. M. McClellan, C. S. Coffey, J. J. Allison, C. I. Kiefe, and D. G. Warnock Survival Advantage of Black Patients with Kidney Disease after Acute Myocardial Infarction Clin. J. Am. Soc. Nephrol., September 1, 2006; 1(5): 993 - 999. [Abstract] [Full Text] [PDF] |
||||
![]() |
N. R. Powe Reverse Race and Ethnic Disparities in Survival Increase with Severity of Chronic Kidney Disease: What Does This Mean? Clin. J. Am. Soc. Nephrol., September 1, 2006; 1(5): 905 - 906. [Full Text] [PDF] |
||||
![]() |
S. I. Hallan, J. Coresh, B. C. Astor, A. Asberg, N. R. Powe, S. Romundstad, H. A. Hallan, S. Lydersen, and J. Holmen International Comparison of the Relationship of Chronic Kidney Disease Prevalence and ESRD Risk J. Am. Soc. Nephrol., August 1, 2006; 17(8): 2275 - 2284. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Bibbins-Domingo, G. M. Chertow, L. F. Fried, M. C. Odden, A. B. Newman, S. B. Kritchevsky, T. B. Harris, S. Satterfield, S. R. Cummings, and M. G. Shlipak Renal function and heart failure risk in older black and white individuals: the health, aging, and body composition study. Arch Intern Med, July 10, 2006; 166(13): 1396 - 1402. [Abstract] [Full Text] [PDF] |
||||
![]() |
W. McClellan, D. G. Warnock, L. McClure, R. C. Campbell, B. B. Newsome, V. Howard, M. Cushman, and G. Howard Racial Differences in the Prevalence of Chronic Kidney Disease among Participants in the Reasons for Geographic and Racial Differences in Stroke (REGARDS) Cohort Study J. Am. Soc. Nephrol., June 1, 2006; 17(6): 1710 - 1715. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. C.-c. Hsu, M. S. Bray, W.H.L. Kao, J. S. Pankow, E. Boerwinkle, and J. Coresh Genetic Variation of the Renin-Angiotensin System and Chronic Kidney Disease Progression in Black Individuals in the Atherosclerosis Risk in Communities Study J. Am. Soc. Nephrol., February 1, 2006; 17(2): 504 - 512. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Jones-Burton, S. L. Seliger, J. Brown, L. Stackiewicz, V. Doren Hsu, and J. C. Fink Racial variations in erythropoietic response to epoetin alfa in chronic kidney disease and the impact of smoking Nephrol. Dial. Transplant., December 1, 2005; 20(12): 2739 - 2745. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. A. Stevens, G. Fares, J. Fleming, D. Martin, K. Murthy, J. Qiu, P. C. Stark, K. Uhlig, F. Van Lente, and A. S. Levey Low Rates of Testing and Diagnostic Codes Usage in a Commercial Clinical Laboratory: Evidence for Lack of Physician Awareness of Chronic Kidney Disease J. Am. Soc. Nephrol., August 1, 2005; 16(8): 2439 - 2448. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. L. Smith, M. G. Shlipak, E. P. Havranek, F. A. Masoudi, W. M. McClellan, J. M. Foody, S. S. Rathore, and H. M. Krumholz Race and Renal Impairment in Heart Failure: Mortality in Blacks Versus Whites Circulation, March 15, 2005; 111(10): 1270 - 1277. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. S. M. Buist, P. L. Porter, C. Lehman, S. H. Taplin, and E. White Factors Contributing to Mammography Failure in Women Aged 40-49 Years J Natl Cancer Inst, October 6, 2004; 96(19): 1432 - 1440. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. G. Gharavi, T. Ahmad, R. D. Wong, R. Hooshyar, J. Vaughn, S. Oller, R. Z. Frankel, L. A. Bruggeman, V. D. D'Agati, P. E. Klotman, et al. Mapping a locus for susceptibility to HIV-1-associated nephropathy to mouse chromosome 3 PNAS, February 24, 2004; 101(8): 2488 - 2493. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. Giannini and R. Testa The Metabolic Syndrome: All Criteria Are Equal, but Some Criteria Are More Equal Than Others Arch Intern Med, December 8, 2003; 163(22): 2787 - 2788. [Full Text] [PDF] |
||||
![]() |
C.-y. Hsu, F. Lin, E. Vittinghoff, and M. G Shlipak Racial Differences in the Progression from Chronic Renal Insufficiency to End-Stage Renal Disease in the United States J. Am. Soc. Nephrol., November 1, 2003; 14(11): 2902 - 2907. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. J Karter Commentary: Race, genetics, and disease-- in search of a middle ground Int. J. Epidemiol., February 1, 2003; 32(1): 26 - 28. [Full Text] [PDF] |
||||
| ||||||||||||||||||||||