SEX AND CHRONIC KIDNEY DISEASE , 655
Factors Related to Sex Differences, 657
Oral Contraceptives, Hormone Replacement Therapy, and Kidney Disease, 659
RACE, ETHNICITY, AND CHRONIC KIDNEY DISEASE , 660
Defining Race and Ethnicity, 660
Factors Related to Race and Ethnicity, 660
Potential Mechanisms of Racial and Ethnic Disparities, 662
SOCIOECONOMIC FACTORS AND CHRONIC KIDNEY DISEASE , 663
CONCLUSION , 666
Chronic kidney disease (CKD) prevalence in the United States has increased over the past decade, with 10% of the population having CKD stages 1 to 4 in 1994 compared to 13.1% in 2004, a trend only partly accounted for by the increased prevalence of diabetes and obesity. In addition, this increased prevalence of CKD has translated into a 32.4% increase in years of life lost between 1990 and 2010 on a U.S. population basis. Worldwide, the incidence of CKD has risen as well. A recent sampling of Chinese individuals found that 1.7% had an estimated glomerular filtration rate (eGFR) lower than 60 mL/min/1.73 m 2 and 9.4% had albuminuria, translating into an estimated 119.5 million individuals with CKD in that country alone. Worldwide incidence and prevalence estimates are presented in Figure 21.1 and reflect generally higher rates of CKD in industrialized nations. Patterns in the prevalence, incidence, and progression of chronic kidney disease vary by certain demographic characteristics, including sex, race and ethnicity, and socioeconomic status. This chapter will summarize what is known of these patterns, highlight consistent findings across sociodemographic groups, speculate on the genesis of variations across demographic groups, and highlight questions that still remain pertaining to kidney outcomes among these populations.
Sex and Chronic Kidney Disease
Differences between men and women in the incidence and prevalence of various kidney diseases and rate of kidney disease progression may be influenced by sex differences in glomerular mass, responses to hormones, cytokines, apoptosis, vasoactive, and other soluble circulating factors, and differences in the responses to aging and reductions in nephron mass. Women have been reported, on average, to have approximately 10% to 15% fewer glomeruli than men, but this is thought to be a function primarily of birth weight and body surface area (BSA) rather than sex. Glomerular volume tends to be similar in men and women. The glomerular filtration rate (GFR) is also similar in men and women when corrected for BSA and muscle mass. However, some have reported a somewhat lower BSA-adjusted GFR in women. Thus, although some subtle differences in renal mass and structure have been reported in men compared to women, these are probably of little or no clinical significance and are more likely related to factors other than sex.
Experimental animal models and human studies have described sex differences and sex hormone influences in the synthesis and plasma levels of, and biologic responses to, a variety of circulating factors involved in the regulation of normal renal function. These same factors may also be involved in responses to renal injury and susceptibility to kidney disease. Some of these include angiotensinogen, angiotensin II, prorenin, renin, angiotensin-converting enzyme (ACE), and angiotensin receptor expression. Sex differences have also been reported in nitric oxide and prostaglandin synthesis and responsiveness, lipid oxidation and oxidative stress, mesangial cell collagen synthesis and degradation, responses to transforming growth factor-β (TGF-β) and tumor necrosis factor-α (TNF-α), as well as in apoptotic and profibrotic signaling pathways. High adiponectin levels have also been reported to predict CKD progression in men, but not in women. Estradiol has been identified as having various effects on mesangial cells. Neither androgens nor estrogens directly influence GFR or renal blood flow in humans. The extent to which any of these factors are specifically and causally related to sex differences in kidney function or kidney disease incidence and progression is still uncertain.
The incidence rate of end-stage kidney disease (ESKD) in the United States is approximately 60% higher among men compared to women, whereas the estimated prevalence of moderate (i.e., stages 3 and 4) CKD is higher among women (8.0% in women vs. 5.4% in men). A recent Canadian study demonstrated similar rates of CKD by sex, however. A large international meta-analysis of more than 2 million individuals with CKD revealed that men with CKD had higher absolute risks of death and cardiovascular disease than women, whereas women had a steeper relationship of these outcomes with levels of eGFR.
Factors Related to Sex Differences
Glomerular Disease Incidence and Prevalence
Research assessments of glomerular disease incidence and prevalence in adults are made difficult because of uncertainty about the population base from which these figures are derived, variations in study participants’ ages, and varying indications for kidney biopsy. The overall incidence of primary glomerular diseases among residents of Olmsted County, Minnesota, a primarily white population, has been estimated based on kidney biopsy records to be 7.9/100,000 person-years in men and 5.4/100,000 person-years in women. A study from France reported a prevalence of primary glomerular disease of 8.2/1000 men and 5.1/1000 women during a 27-year period ending in 2002. Men tend to predominate in many series of adult patients with focal segmental glomerulosclerosis (FSGS) and immunoglobulin A (IgA) nephropathy, with a more variable sex mix for adults with minimal change disease and membranous nephropathy.
Progression of Chronic Kidney Disease
There is a paucity of data examining the effects of sex on CKD progression at the population level. A study by Eriksen and Ingebretsen examined a total of 3047 patients from a municipality in Norway and found that the rate of CKD progression was significantly slower among women. This study, which also reported a lower aggregate rate of kidney disease progression than is commonly seen in other industrialized nations, also revealed that women had better renal and patient survival. A study that modeled CKD incidence based on prevalence data also found lower rates of progression to ESKD among U.S. women.
Several more recent meta-analyses have also considered sex influences on kidney disease progression from a variety of underlying causes. Jafar and associates performed a patient-level meta-analysis using a pooled database for patients with nondiabetic kidney disease from 11 prospective randomized controlled trials of ACE inhibitors for slowing progression. Using doubling of serum creatinine levels or onset of ESKD as a composite primary end point, they concluded that the risk of kidney disease progression was not different in men and women in an unadjusted analysis, but that the risk was actually higher in women than men after adjustment for baseline variables, including urine protein excretion and treatment assignment (relative risk [RR], 1.30 to 1.36, depending on the model). They noted that most of the women in these trials were of postmenopausal age, limiting their applicability to younger premenopausal women.
Several additional studies considering sex influences on kidney disease progression have also been published. Two large population-based studies reported a more favorable prognosis for women compared to men with CKD in Norway and Sweden. The Modification of Diet in Renal Disease (MDRD) study, which enrolled patients with autosomal dominant polycystic kidney disease (ADPKD), glomerulonephritis, or other nondiabetic kidney diseases, reported the rate of kidney disease progression to be slower in women compared to men, particularly among younger premenopausal women. This difference was markedly diminished and no longer statistically significant after adjustment for level of proteinuria and blood pressure. A more recent report of the MDRD study participants’ long-term outcomes also found similar kidney failure event rates for men and women.
Progression of Primary Glomerular Disease
The prognosis—that is, the rate of progression of the underlying kidney disease—is generally considered to be worse in men than in women with membranous nephropathy, IgA nephropathy, FSGS, and lupus nephritis. However, studies using multivariable analysis to evaluate the effect of sex on kidney disease progression have produced variable findings. Although none have found female gender to be associated with more rapid kidney disease progression, the often cited association of male gender with more rapid disease progression has been inconsistent. Much of this literature was analyzed in a recent meta-analysis by Neugarten and coworkers. This meta-analysis considered eight studies, including over 2000 patients with nondiabetic “chronic kidney disease” for which no specific cause was identified, and concluded that kidney disease progression was statistically significantly associated with male gender ( Figure 21.2 ). Among five studies excluded from this analysis because of incomplete reporting of effect size, two found that kidney disease progression was more rapid in men, whereas three found no sex difference. Although not assessed in this meta-analysis, other studies have reported that the more favorable rate of progression in women is limited to the premenopausal period.
An association between male gender and progression of IgA nephropathy was demonstrated in the meta-analysis by Neugarten and colleagues, with 25 studies and over 3000 patients ( Figure 21.3 ). Of these 25 studies, 21 found more rapid progression in men. In all but a few of the studies in this meta-analysis, however, the association was not statistically significant. In addition, several studies suggested that men had better outcomes. Of 13 studies that were excluded from this meta-analysis because of the inability to calculate effect size, 12 found no sex differences. Combined, these data suggest that any association between sex and IgA nephropathy progression is likely to be weak.
Among 21 studies of almost 1900 patients with membranous nephropathy considered in the Neugarten and associates’ meta-analysis, male gender was significantly associated with disease progression ( Figure 21.4 ). However, five excluded studies (because of inability to calculate effect sizes) reported no sex association with progression. Other older, pooled analyses have also reported an association of male gender with poorer renal outcomes for membranous nephropathy.
Cattran and coworkers analyzed outcomes from over 1300 patients enrolled in the Toronto Glomerulonephritis Registry with membranous nephropathy, FSGS, and IgA nephropathy. After adjusting for blood pressure and proteinuria, disease progression and renal survival rates were not different between men and women, except among those with levels of proteinuria more than 7 g/day in which men had more rapid loss of GFR than women. Disease progression has also been found to be similar for men and women with IgA nephropathy in most other studies.
Lupus Nephritis Progression
Recent studies of sex influences on lupus nephritis outcomes in adults have reported discrepant findings but were limited by small numbers of patients, variable outcome measures, and varying assessment of other covariates such as histopathologic disease class, proteinuria, blood pressure, and immunosuppressive treatment.
Autosomal Dominant Polycystic Kidney Disease Progression
In the meta-analysis by Neugarten and colleagues mentioned above, of 12 studies with over 3000 patients with ADPKD, there was an apparent small protective effect of male gender on disease progression ( Figure 21.5 ). However, this conclusion was largely the result of inclusion of a single Italian study that reported a highly statistically significant favorable association with male gender and disease progression. Excluding this study resulted in the finding of a statistically significant association of male gender and more rapid progression, an effect seen in 10 of 12 studies (although all four excluded studies found no sex association with disease progression). More recent studies published since this meta-analysis have suggested that male gender is likely a risk factor for progression of CKD and ESKD in patients with ADPKD. Among two recent reports of magnetic resonance imaging of kidney and cyst growth, only one found male gender to be associated with more rapid growth.
It is worth noting that the association between ADPKD genotype and progression of kidney disease appears to interact with sex. Sex does not appear to influence renal outcomes significantly in the more common variant of ADPKD caused by mutation in the polycystin 1 gene. In contrast, women with ADPKD caused by mutations in the polycystin 2 gene tend to have more favorable renal outcomes compared to men. The biologic basis for this difference is not presently known.
Diabetic Nephropathy Progression
Influences of sex hormones on diabetes and diabetic nephropathy have been reviewed in detail. There are relatively few data on the association of sex with rate of progression of diabetic nephropathy, and the studies that have been performed reported inconsistent findings. Girls with type 1 diabetes are more likely to develop moderately increased albuminuria during puberty than boys, and there tends to be a more rapid loss of GFR following puberty in women compared to men. Among adults with childhood-onset type 1 diabetes mellitus (T1DM), several studies have suggested that males have a greater likelihood of developing albuminuria and, among those with diabetic nephropathy, the rate of decline in GFR is faster. One study found that compared to men, women with T1DM are more likely to develop diabetic nephropathy while under good metabolic control, a sex effect that was attenuated in the setting of poor metabolic control. Other studies have not found an independent adverse influence of male gender on rate of progression of diabetic kidney disease.
A large U.S. study of 10,290 patients with type 2 diabetes mellitus (T2DM) has shown that male gender is associated with a greater risk of progression than female gender in those with microalbuminuria but not in those who had macroalbuminuria or no albuminuria. The percentage of men and women with diabetic nephropathy starting dialysis is similar, although recent data from the U.S. Renal Data System (USRDS) have indicated that the incidence of ESKD is slightly higher among white men with T2DM compared to white women. Studies have also suggested a greater incidence and prevalence of microalbuminuria and macroalbuminuria in white men compared to white women, with the opposite sex predilection among blacks. Some studies have reported similar rates of disease progression and risk for development of renal end points, including ESKD across sexes.
Oral Contraceptives, Hormone Replacement Therapy, and Kidney Disease
Few studies have examined the influence of oral contraceptives and hormone replacement therapy (HRT) on kidney function in women, with and without recognized kidney disease. Oral contraceptives have been associated with a higher prevalence of microalbuminuria in women with diabetic nephropathy in some but not other studies. One study found a greater risk associated with higher estrogen strength and longer term use (>5 years). Oral contraceptives have also been associated with a higher risk of microalbuminuria and decline of GFR in premenopausal women without CKD.
In a small, short-term prospective study, administration of a combination of estradiol and norgestrel for 3.5 months to 16 postmenopausal women with diabetes mellitus and hypertension was associated with a statistically significant reduction in mean level of proteinuria from 452 to 370 mg/day and increase in creatinine clearance from 100.8 to 106.2 mL/min. In a community-based case-control study, postmenopausal HRT was associated with a twofold higher odds ratio (OR) for microalbuminuria after adjustment for several clinical variables. The OR for microalbuminuria was similar among women receiving HRT with and without progestins. The association between HRT use and microalbuminuria was limited to women using HRT for longer than 5 years. In contrast, in another study, postmenopausal HRT was associated with a lower mean urine albumin-to-creatinine ratio and a lower prevalence of microalbuminuria at a baseline examination and after 5 years of follow-up.
More recently, Ahmed and colleagues studied almost 6000 postmenopausal women for over 2 years to examine the effect of HRT on the eGFR using the abbreviated MDRD equation. After adjustment for age, diabetes, other comorbidities, and baseline eGFR, it was found that HRT use is associated with a more rapid decline in eGFR and a 19% greater risk for eGFR to fall by 4 mL/min/1.73 m 2 /yr or more. The higher rate of eGFR decline and risk of rapid GFR loss were limited to users of estrogen-only HRT; use of combined or progestin-only HRT was not associated with a decline in eGFR. There was also a linear relationship between cumulative dose of estrogen and decline in mean eGFR. In contrast, a smaller study of 443 patients with 10 years of follow-up demonstrated no association between postmenopausal estrogen use and decline in GFR. Thus, although women are considered to be at less risk for development and progression of many types of kidney disease, the influences of menopausal status and hormonal replacement have not been thoroughly investigated. The findings of Ahmed and associates need to be further explored with consideration of factors not evaluated in that study, such as blood pressure, level of proteinuria, and obesity, before concluding with any certainty that oral estrogen-based HRT accelerates the progression of CKD.
Patterns of the incidence of kidney disease across sexes are generally consistent, with higher rates occurring in men compared to women. Similarly, men are reported to have greater rates of progression of nondiabetic CKD for some specific types of kidney disease, especially compared to premenopausal women. More investigation into rates of progression of IgA nephropathy, lupus nephritis, and ADPKD across sexes and of overall progression rates in postmenopausal women is warranted. Additional study of the effects of HRT in women on the incidence and progression of kidney disease is also needed.
Race, Ethnicity, and Chronic Kidney Disease
Defining Race and Ethnicity
Population genetics studies have refuted the existence of biologic races in human populations; nevertheless, the term is still used widely in medicine in relation to continental population ancestry. The use of race classifications in medicine and epidemiology has been the subject of much debate, mainly because of the many ways this information can be captured and interpreted. Nonetheless, classifying race and ethnicity in biomedical research facilitates several important activities, including the characterization of health statistics, risk of adverse health outcomes, and examination of delivery of health care services across subpopulations. Also, these classifications can be used as a proxy for unmeasured biologic and social factors.
The utility of describing race and ethnicity and relating it to outcomes of interest lies in the ability to capture information about differences in genetics and biology, behavior, exposure to environmental factors, and social and physical environments. However, the imperfect nature of the relationship between race and these factors highlights the importance of supplementing race and ethnicity data, when possible, with those on individual-level factors that are often meant to be represented by race. To reflect factors related to social, cultural, and physical environments and exposures most accurately, individual race and ethnicity is often self-designated. This approach to classification was first adopted by the U.S. Census Bureau in 1960, followed by the opportunity to self-designate Hispanic ethnicity in 1970 and finally, in 2000, the ability to designate more than one race category. The increasing percentage of the U.S. population that can trace its roots to multiracial or multiethnic sources has motivated researchers and demographers to collect and analyze self-designated racial and ethnic data so as to reflect this racial and ethnic admixture. However, limited knowledge of ancestry, and the large and increasing frequency of migration, creates additional challenges for valid race classification. Despite these limitations, when race is used as an explanatory factor to represent genetic and biologic determinants of disease, self-designated race may be informative as long as there is enough additional information on important socioeconomic, behavioral, and physical environment factors. Finally, it has been suggested that ethnic groups that share a unique history, language, customs, ancestry, geography, religion, and/or specific genetic markers should replace traditional race classifications in biomedical research. However, these approaches may limit the usefulness of race as an explanatory factor in research and may not be suitable across all types of investigations.
Factors Related to Race and Ethnicity
Incidence of Chronic Kidney Disease
An emphasis on defining and investigating CKD prior to dependence on dialysis has only emerged in more recent years. National guidelines were first established in 2002 to define and stage prevalent CKD. To date, no such guidance has been provided to define and capture information on the incidence of CKD nationally or in the research setting. To address the lack of national data on the burden, awareness, risk factors, and health consequences of CKD, the National Chronic Kidney Disease Surveillance System has been under development after a pilot and feasibility phase as part of the CKD Initiative of the Centers for Disease Control and Prevention. This system uses passive surveillance strategies that incorporate a broad network of data sources and aims to disseminate information through fact sheets, reports, and a website. To date, the only U.S. CKD data on incidence available in the surveillance system come from the Department of Veterans Affairs, documenting an incidence of CKD stages 3 to 5 of 46.55 (95% confidence interval [CI], 46.35 to 46.75)/1000 person-years.
In research, large longitudinal studies are necessary to provide reliable estimates of the incidence of disease. International guidelines have suggested that incidence be defined by a combination of eGFR, proteinuria, and cause of kidney disease (e.g., allowing for individuals with polycystic kidney disease but normal GFR and no proteinuria to be included in CKD incidence data). Other published definitions include data on serum creatinine, International Classification of Diseases (ICD) codes, and/or death records, but do so in varied ways. Variations in the definition of incident CKD can modify the relationship between race and CKD occurrence. For example, incident CKD among participants aged 45 to 64 years at baseline in the Atherosclerosis Risk in Communities (ARIC) study occurred at a higher rate among blacks than whites using a definition of CKD based on a rise in serum creatinine levels (8.0 vs. 3.2/1000 person-years, respectively) and ICD codes (2.0 vs. 1.0/1000 person-years), but at a lower rate when CKD was defined as a low eGFR (8.9 vs. 10.8/1000 person-years). A composite definition requiring a low eGFR and at least a 25% drop in eGFR resulted in similar incidence rates for blacks and whites (6.9 vs. 6.6/1000 person-years, respectively).
Despite the lack of consistency in defining incident CKD, several estimates of CKD incidence rates across different racial and ethnic groups have been reported. The 2009 USRDS Annual Data Report cited the incidence of CKD, based on diagnostic codes, among the general Medicare population (mean age, 75.5 years) at 5.6% in African Americans compared to 3.8% in whites in 2007. A 1999 publication that included ARIC participants reported a CKD incidence of 28.4/1000 person-years among black participants with diabetes compared to 9.6/1000 person-years among whites with diabetes, and an age-, sex-, and baseline serum creatinine level–adjusted odds ratio of early kidney function decline among blacks compared to whites of 3.2 (95% CI, 1.9 to 5.3). After further adjustment for potentially modifiable risk factors related to socioeconomic status and health behaviors, including education, household income, health insurance, fasting glucose level, mean systolic blood pressure, smoking history, and physical activity level, the odds ratio (95% CI) decreased to 1.4 (0.7 to 2.7), an 82% reduction in excess risk. Although this and other studies have attributed a substantial proportion of the excess risk for kidney disease in black Americans to these nonracial factors, a difference in risk across race still remains in adjusted analyses. A more recent study, which modeled CKD incidence based on published prevalence data, has suggested that 59.1% of U.S. individuals would suffer from CKD stage 3A or worse in their lifetime, with blacks of both sexes experiencing higher rates of CKD stages 4 or 5 and ESKD.
Prevalence of Chronic Kidney Disease
Estimates of CKD prevalence are much more frequently reported than estimates of incidence because of the ability to do so with a single assessment of renal function. These estimates are associated with certain limitations (see Chapter 20 ). Racial disparities in CKD were examined at entry into the Reasons for Geographic and Racial Differences in Stroke (REGARDS) study, a population-based cohort study of adults older than 45 years. An eGFR of less than 60 mL/min/1.73 m 2 (i.e., CKD stages 3 to 5) was found in 43.3% overall and was more prevalent among whites compared to blacks (49.9% vs. 33.7%, respectively; Table 21-1 ). However, although blacks were less likely than whites to have an eGFR between 30 and 60 mL/min/1.73 m 2 , the reverse was true in the eGFR range of less than 30 mL/min/1.73 m 2 . Using data from the National Health and Nutrition Examination Surveys (NHANES) from 1988 to 1994 and 1999 to 2004, and the MDRD GFR estimating equation, the prevalence of CKD increased from 10.5% to 13.8% among non-Hispanic whites, from 10.2% to 11.7% among non-Hispanic blacks, and from 6.3% to 8.0% among Mexican Americans over this time period. Emphasizing the impact of the tool used to assess kidney function, the 1999 to 2004 estimates of the prevalence of CKD stages 1 to 4 among non-Hispanic whites and blacks in the United States were no longer significantly different when made using the more accurate Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation to estimate GFR. However, a racial difference persists in the prevalence of CKD stages 3 to 4; estimates shifted from 9.6% and 5.2% using the MDRD GFR estimating equation to 7.8% and 5.4% using the CKD-EPI equation among whites and blacks, respectively. Another analysis of NHANES 1988 to 1994 data revealed a higher likelihood, even after multivariable adjustment, of U.S. blacks and Mexican Americans with and without diabetes to have albuminuria compared to U.S. whites with or without diabetes (odds ratio [OR], 1.8 to 2.8). Finally, in a study of adult Navajo Indians, 3% to 6% of nondiabetics and 10% to 11% of diabetics had an elevated serum creatinine level consistent with a creatinine clearance of 65 mL/min or higher for men and 53 mL/min or higher for women.
|eGFR (mL/min/1.73 m 2 )||N (%)||OR (95% CI)||aOR † (95% CI)|
|Black (n = 8139)||White (n = 11,620)|
|>60||5394 (66.3)||5817 (50.1)||Reference||Reference|
|50-59||1541 (18.9)||3611 (31.1)||0.46 (0.43-0.49)||0.42 (0.40-0.46)|
|40-49||693 (8.5)||1506 (13.0)||0.50 (0.45-0.55)||0.37 (0.33-0.41)|
|30-39||287 (3.5)||521 (4.5)||0.59 (0.51-0.67)||0.38 (0.32-0.45)|
|20-29||116 (1.4)||131 (1.1)||0.95 (0.74-1.22)||0.48 (0.36-0.64)|
|10-19||60 (0.7)||25 (0.2)||2.56 (1.62-4.13)||1.73 (1.02-2.94)|
|<10||48 (0.6)||9 (0.08)||5.75 (2.82-11.7)||4.19 (1.90-9.24)|
Hispanic ethnicity is often aggregated into one group, despite a wide variety of national origins and races represented by this classification. Rodriguez and coworkers examined data from the Hispanic Health and Nutrition Examination Survey (HHANES) on differences in serum creatinine levels and estimated creatinine clearances across Hispanic subgroups, including Mexican Americans, mainland Puerto Ricans, and Cuban Americans. Cuban Americans had the highest mean serum creatinine levels, and both Puerto Ricans (OR 1.7; 95% CI, 1.2 to 2.6) and Cuban Americans (OR 4.6; 95% CI, 2.5 to 8.3) were more likely than Mexican Americans to have estimated creatinine clearances lower than 60 mL/min/1.73 m 2 . These observations further highlight the heterogeneity of physiology within currently used race and ethnicity categorizations.
Progression of Chronic Kidney Disease
Rates of progression of CKD to ESKD are higher among African American, Hispanic, and American Indian adults compared to white U.S. adults, as described in Chapter 20 . For example, Hispanic ethnicity was associated with a significantly increased risk (hazard ratio [HR] 1.33; 95% CI, 1.17 to 1.52) for ESKD among individuals with CKD compared to non-Hispanic whites in large study of Kaiser Permanente of Northern California health plan enrollees. Among African Americans, the rate of decline of GFR is greater (i.e., 1.4 to 1.5 mL/min/yr greater decline in blacks compared to whites), and the risk of developing ESKD was twofold higher (RR [95% CI], 2.0 [1.1 to 3.6]) compared with whites. This finding was reinforced by a longitudinal study of Medicare recipients aged 65 years or older followed for up to 10 years. After adjustment for age and sex, black patients with diabetes were 2.4 to 2.7 times and other races and ethnicities were 1.6 to 1.7 times more likely to develop ESKD compared to whites. Similar elevations in risk were noted among black and other racial and ethnic minorities with hypertension. Finally, among patients with neither diabetes nor hypertension, black patients were still 3.5 times more likely, and those with “other” designated race, were twice as likely to develop ESKD than whites.
Potential Mechanisms of Racial and Ethnic Disparities
Racial disparities in kidney disease may partially be explained by a higher prevalence and lower levels of control of hypertension among African American adults ( Figure 21.6 ). The onset of hypertension appears to occur earlier and with more severity in African Americans, leading to greater end-organ damage. Additional potential causes of kidney disease disparities between whites and blacks may be differences in the prevalence and control of diabetes, prevalence and severity of obesity, physiologic differences in cytokine production, renal hemodynamics, and electrolyte regulation, genetic factors, and differences in socioeconomic status, access to health care, behavioral factors, and physical environments. One such explanation was provided in a publication in which Hung and colleagues reported a higher risk for CKD progression among participants of the African American Study of Kidney Disease and Hypertension (AASK) with certain C-reactive protein polymorphisms. Similar disparities exist between Hispanic adults compared to whites that may be partially explained by a higher prevalence, earlier onset, and increased severity of diabetes in this ethnic minority population, lower rates of awareness, treatment, and control of hypertension, and higher prevalence of obesity, among other biologic, social, behavioral, and communication factors.
Several studies have also identified factors related to access to health care as being strongly predictive of the development of ESKD. One ecologic study comparing regions in California defined by zip codes reported a higher incidence of ESKD caused by diabetes in areas with a higher proportion of hospitalizations of those with no insurance or with Medicaid, and a lower incidence in areas with more hospitalizations for those with managed care insurance plans. Surprisingly, the unadjusted incidence rates were lower in zip codes with known shortages of health professionals compared to zip codes with ample health professional populations (adjusted rates were not presented). Also, incidences of ESKD were higher in areas with more hospitalizations for hyperglycemic complications, suggesting a role for ineffective or poor access to treatment in the development of ESRD caused by diabetes. Similarly, a report cited more abnormal laboratory values at the onset of ESKD treatment in those without medical insurance.
The persistence of variations in risk of CKD progression by race after adjustment for many of these factors has motivated investigation of previously unexplained genetic variation using new investigative tools. Two independent groups performed genome-wide analyses for ESKD risk loci within incident African American ESKD patients. Both used admixture linkage dysequilibrium analysis, which is based on the premise that when two genetically diverse populations mix, the admixed population receives chromosomal regions from either ancestry that can be identified by genotyping markers with different allelic frequencies between the ancestral populations. The groups screened the genome of African Americans with ESKD to identify ESKD susceptibility loci, regions of the genome where individuals with ESKD have more or less African ancestry than their nondiseased counterparts. The results from the Family Investigation of Nephropathy and Diabetes Research Group identified an association between excess African ancestry and nondiabetic ESKD on chromosome 22q12, but not with diabetic ESKD. In contrast, the presence of an allele of European origin at this locus was found to be protective, with a relative risk of 0.5 compared with an allele of African origin. In this study, most of the excess ESKD risk in those of African ancestry was correlated with a number of common single-nucleotide polymorphisms (SNPs) in the gene encoding nonmuscle myosin heavy chain type II isoform A (MYH9). Similarly, Kopp and associates found this same gene to be associated with biopsy-proven, idiopathic FSGS and human immunodeficiency virus type 1 (HIV-1)–associated FSGS in African Americans. In this study, the OR associated with the recessive E-1 haplotype for MYH9 SNPs in African Americans was 4.7 (95% CI, 3.1 to 7.0) in idiopathic FSGS and 5.9 (95% CI, 2.9 to 12.9) in HIV-associated FSGS. However, subsequent studies have suggested that certain polymorphisms of an adjacent gene, APOL1, are responsible for different rates of renal progression across races. This gene, which codes for apolipoprotein L-1 (apo L-1), is in linkage disequilibrium with MYH9, which likely accounts for the findings of prior studies. In 2010, Genovese and coworkers and Tzur and colleagues demonstrated two sequence variants in apo L-1 that were strongly associated with the development of FSGS and hypertensive nephrosclerosis in African Americans. These variants were also considered to have risen in allele frequency under positive evolutionary selection and were shown to have an increased ability to lyse the human pathogen Trypanosoma brucei rhodesiense, suggesting that positive-selection pressure may account for the increased prevalence of these isoforms among African Americans.
In a study of 2955 participants in the Chronic Renal Insufficiency Cohort, black individuals homozygous for a high-risk apo L-1 allele had an increased risk of doubling of creatinine level or progression to ESKD compared to nonhomozygous blacks or whites. Among diabetics, the HR for adverse renal outcomes was 3.1 (2.1 to 4.5) times higher among high-risk blacks compared to whites. Among nondiabetics, the risk was 2.5 (1.8 to 3.4) times higher. Similarly, among AASK participants, those with a high-risk, apo L-1 genotype had a rate of doubling of creatinine level or ESKD 1.9 times higher than in those without a high-risk genotype.
The ARIC study of 3067 African Americans without CKD at baseline revealed a significantly elevated risk of developing incident CKD (HR 1.5; 95% CI, 1.0 to 2.2) and progressing to ESKD (HR 2.2; 95% CI, 1.0 to 4.8) among those homozygous for the apo L-1 risk alleles compared to heterozygotes and those with the wild-type alleles. The mechanisms through which APOL1 mutations accelerate kidney disease progression remain unclear, but are an area of very active research.
There is little variation in reports of the patterns of incidence, prevalence, and progression of kidney disease across race and ethnicity. In general, despite a lower prevalence of CKD among African Americans and Hispanics compared to whites, the incidence of CKD is higher in African Americans and rates of progression are faster in African Americans, Hispanics, and American Indians compared to non-Hispanic whites.
Socioeconomic Factors and Chronic Kidney Disease
Earlier in this chapter, we discussed racial disparities in the incidence and progression of CKD. Racial differences in kidney disease risk are partially mediated by factors related to socioeconomic status and social deprivation (see also Chapter 84 ). This portion of the excess risk is potentially modifiable and, therefore, of particular interest for targeting prevention strategies. Socioeconomic status has been described as a distal risk factor for kidney disease that acts through several proximal factors, including poverty and low income, lack of nutrition, low educational levels, exposure to heavy metals, substance abuse, and limited access to health care. These factors can be examined at the individual and neighborhood level at any given point in time. As a result, several analyses have accounted for both individual- and area-level socioeconomic status. A framework for considering the numerous social and cultural determinants of disparities in CKD is shown in Figure 21.7 . In addition, as in many disease processes that develop over protracted periods of time, past and present exposures are responsible for increases in risk. For this reason, life course (i.e., the cumulative effect of social environments over the course of a lifetime) and parental socioeconomic factors have also been investigated as contributors to the incidence and progression of kidney disease. Full evaluation of these individual- and area-level factors in biomedical research studies is key to explaining the observed racial and ethnic disparities in kidney disease.
Relationship to Socioeconomic Factors
Incidence of Chronic Kidney Disease
Very few studies have examined risk factors for the incidence of CKD and fewer yet have investigated socioeconomic factors. Krop and associates observed that blacks with diabetes mellitus were three times more likely than whites with diabetes mellitus to develop CKD in unadjusted analyses. Subsequent adjustment for additional covariates revealed that 6% of the excess risk for development of CKD in black adults compared to white adults with diabetes was explained by income and education level. Suboptimal health behaviors and poor control for glucose level and blood pressure accounted for a substantial proportion of the remaining risk. Given the strong relationship among socioeconomic status, health behaviors, and glycemic and hypertensive control, the overall effect of socioeconomic factors on the incidence of CKD is understated in the aforementioned 6% excess risk.
A similar constellation of risk factors has been described in conjunction with renal involvement in systemic lupus erythematosus (SLE). A higher incidence of lupus nephritis has been noted among minority populations. One report, which included participants with recently diagnosed SLE from the Lupus in Minorities: Nature vs. Nurture (LUMINA) study, noted the development of lupus nephritis among 44.6% of Texas Hispanics, 11.3% of Puerto Rican Hispanics, 45.8% of African Americans, and 18.3% of whites. When examined further, a composite socioeconomic status factor, including information on education, insurance, and poverty status, accounted for 14.5% of the variance because of ethnicity after adjustment; socioeconomic status and genetic admixture together accounted for an additional 12.2% of the variance. Of note, an additional 36.8% of the variance in this model could be attributed to genetic admixture, underscoring the importance of genetic factors over socioeconomic factors in explaining the racial disparities in renal involvement in SLE.
A population-based case-control study in Sweden has provided additional evidence of the risk of incident CKD with low socioeconomic status. ORs of incident CKD associated with families of solely unskilled workers were 2.1 (95% CI, 1.1 to 4.0) and 1.6 (95% CI, 1.0 to 2.6) among women and men, respectively, compared to families with at least one professional worker. In addition, Swedish adults with 9 years or less of education were 30% more likely (OR, 1.3; 95% CI, 1.0 to 1.7) to develop CKD than adults with a college education.
Finally, in parts of the United Kingdom, an approximate 50% increase in the incidence of CKD was reported among those in the highest quintile of social deprivation (i.e., living in an electoral ward with a high proportion of households without a car, unemployed, overcrowded, and not owner-occupied) compared to the lowest quintile. A more recent analysis of the ARIC study demonstrated an increased incidence of CKD among blacks compared to whites (14.7 vs. 12.0 cases/1000 person years). Demographic, socioeconomic, lifestyle, clinical factors, and access to health care accounted for 74% of this increased risk.
Prevalence of Chronic Kidney Disease
The relationship of measures of socioeconomic status and social deprivation with the prevalence of CKD has been better characterized than with the incidence of CKD. Using data from the ARIC study, Shoham and coworkers found that being a member of the working class across some or all of their life course was associated with an increased OR of CKD among whites and blacks (OR [95% CI], 1.4 [0.9 to 2.0] and 1.9 [1.3 to 2.9], respectively). Martins and colleagues have cited a significant association (OR 1.2; 95% CI, 1.1 to 1.3) between living at less than 200% of the federal poverty level and microalbuminuria using data from NHANES III after multivariable adjustment. Another analysis of data from NHANES III has confirmed these findings; it also revealed a higher prevalence of CKD in those with less than 12 years of education among non-Hispanic whites and blacks and with an income equivalence level (i.e., total household income divided by the square root of the number of people dependent on that income) less than $12,000 compared to $28,000 or more among non-Hispanic whites only. An even stronger relationship was observed between unemployment and prevalent CKD, but only among non-Hispanic blacks and Mexican Americans. Interestingly, these same associations could not be detected using similar surveys of adults in Australia and Thailand.
Some effects of socioeconomic status on CKD are mediated through more proximal factors, such as nutritional status, health behaviors, and environmental exposures. Nutritional deprivation or physiologic insults in utero cause intrauterine growth restriction (IUGR). Building on the developmental origins of health and disease hypothesis proposed by Barker, Brenner and Chertow hypothesized that IUGR causes a decrease in nephron number, leading to a susceptibility to hypertension and reduced kidney function later in life. A systematic review and meta-analysis of observational studies evaluating the relationship between birth weight and CKD, assessed at age older than 12 months, included 32 studies. Combined weighted estimates of effect provided an OR for CKD associated with low birth weight of 1.7 (95% CI, 1.4 to 2.1). These results were consistent across multiple definitions of CKD, including albuminuria, ESKD, and low eGFR. (For further discussion, see Chapter 23 .)
Finally, numerous environmental exposures that affect kidney health occur disproportionately among certain populations, including those of low socioeconomic status. Exposure to cadmium and lead, even at low levels, is associated with a significantly increased prevalence of kidney disease. Data on the renal effects of drinking water or residential environmental exposure to uranium are scarce, and very few studies have found a significant association between this type of exposure to uranium and renal outcomes, including kidney stones, chronic nephritis, and microalbuminuria. Overcrowding in housing poses more risk for streptococcal infections, which may lead to poststreptococcal glomerulonephritis. Reductions in renal functional reserve, chronic kidney disease, and development of ESKD have been reported in patients who have recovered from poststreptococcal glomerulonephritis.
The effect of income on CKD prevalence may vary by country. An analysis of individuals who participated in NHANES (1999 to 2002) demonstrated that low income was strongly associated with CKD prevalence. Applying the same analysis to a Dutch cohort, low income was only weakly associated with CKD prevalence. This may be a result of the fact that access to health care is more income-dependent in the United States than in the Netherlands. A study of an English cohort revealed significant associations between low income and CKD prevalence, but these disappeared after adjustment for lifestyle and clinical factors.
The impact of income on CKD prevalence may differ by sex and race in U.S. studies, but evidence is sparse. In the Jackson Heart Study, an all–African American cohort, high socioeconomic status was statistically significantly protective against CKD among men (OR 0.5; 95% CI, 0.2 to 1.0), but of only borderline significance among women (OR 0.6; 95% CI, 0.4 to 1.0). A study of 2375 community-living adults in Baltimore, Maryland, has demonstrated that low socioeconomic status is associated with CKD in blacks (OR 1.9; 95% CI, 1.5 to 2.4), but not in whites (OR 1.0; 95% CI, 0.6 to 1.6).
Progression of Chronic Kidney Disease
The association between race and ethnicity and the progression of CKD was discussed earlier in this chapter. As noted, socioeconomic factors highly correlated with race and ethnicity account for a proportion of this excess risk. In particular, Tarver-Carr and colleagues reported a 2.7-fold increased risk of ESKD among African Americans compared with whites. Of the excess risk of developing ESLD among African Americans, 12% was explained by sociodemographic factors, including education, poverty status, and marital status. A large longitudinal study of over 170,000 members of Kaiser Permanente of Northern California reported a significant risk of developing ESKD among members who never attended college (HR 1.6; 95% CI, 1.2 to 2.0) or who completed some college (HR 1.5; 95% CI, 1.1 to 1.9) compared to college graduates after multivariable adjustment.
Further studies have integrated neighborhood-level socioeconomic factors in the examination of the progression of CKD. One study, using data on the incidence of ESKD in Georgia, North Carolina, and South Carolina, defined neighborhood poverty using the percentage of the census tract population living below the poverty level. Unadjusted incidence rate ratios of ESKD increased from 1.5 to 4.5 in a dose-response manner for census tracts with 5% to 9.9%, 10% to 14.9%, 15% to 19.9%, 20% to 24.9%, and 25% or more of the population below the poverty level, respectively, compared to tracts with fewer than 5% below the poverty level. This marker of neighborhood poverty was significantly associated with a higher incidence of ESKD among blacks and whites. However, there was a stronger relationship between the incidence of ESKD and census tract poverty among blacks compared to whites. Finally, Merkin and associates analyzed data from the ARIC and Cardiovascular Health Studies separately to assess the relationship between individual- and neighborhood-level socioeconomic status and progressive CKD. In the ARIC study, age- and center-adjusted incidence rates of progressive CKD increased with declining area-level socioeconomic scores for African American women and white men, but not for African American men and white women. Living in the lowest quartile of area-level socioeconomic status was significantly associated with a 60% greater risk (HR 1.6; 95% CI, 1.0% to 2.5) for progressive CKD only among white men after multivariable adjustment, including individual-level socioeconomic status. Among participants in the Cardiovascular Health Study, age- and center-adjusted incidence rates of progressive CKD were inversely related to area-level socioeconomic scores and to individual levels of income and education. After multivariable adjustment, including individual-level socioeconomic status, living in the lowest quartile of area-level socioeconomic status was associated with an HR for progressive CKD of 1.5 (95% CI, 1.0 to 2.0) compared to the highest quartile. The association with individual-level socioeconomic status no longer remained significant after adjustment in this older population.
A study of 4735 older Cardiovascular Health Study participants demonstrated that the risk of progressive CKD was 50% higher among those living in an area with low socioeconomic status, even after adjustment for personal socioeconomic status. These findings suggest that local factors may be more important than individual income in the relationship between low socioeconomic status and progression of CKD in older adults.
Socioeconomic factors, especially income, education, and environmental factors, appear to explain a large proportion of the excess incidence and progression of CKD among African American and other racial and ethnic groups compared to whites. Additional studies are needed to clarify the influence of access to health care and early-life socioeconomic status on kidney health.