Racial Differences in Mortality Among Those with CKD
Year: 2008
Type: article
Abstract: Compared with white individuals, black individuals have a significantly higher risk for death in the general population but seem to have a survival advantage in the ESRD population. Data on the relationship of race to survival in early stages of chronic kidney disease (CKD) are inconsistent. This study evaluated racial differences in mortality among the adult participants of the Third National Health and Nutrition Examination Survey, a population-based survey of community-dwelling individuals. CKD was defined either by an estimated GFR <60 ml/min per 1.73 m2 or by the presence of albuminuria, and this status was determined for 14,611 individuals, 2892 of whom were found to have CKD. Adjusting for age, gender, and race, risk for all-cause mortality among individuals with CKD was more than double that of individuals with normal renal function. In the subgroup with CKD, adjusting for age and gender, black individuals had a significantly higher risk for death, and this risk was modified by age; specifically, black individuals who were younger than 65 yr were 78% more likely to die than white individuals, whereas no significant differences in mortality were observed among individuals who were ≥65 yr of age. Further adjustment for cardiovascular risk factors and CKD stage did not materially change the results, but the hazard ratios were significantly attenuated after adjustment for socioeconomic factors. In conclusion, these data demonstrate racial/ethnic disparities in mortality among individuals with CKD. This higher risk for death in early stages of CKD may explain the apparent survival advantage observed among black individuals who live long enough to reach stage 5 CKD. In the general population, the cardiovascular and noncardiovascular mortality rates for black individuals are persistently higher than those for white individuals.1–3 In contrast, studies among individuals who have stage 5 chronic kidney disease (CKD) and are undergoing maintenance dialysis have consistently shown that black individuals have a survival advantage compared with white individuals.4,5 Even though some studies in the general population have shown lower mortality rates among Mexican Americans despite a higher prevalence of cardiovascular risk factors and lower socioeconomic status (“Hispanic paradox”), studies with more complete follow-up demonstrated poorer outcomes in this ethnic group when compared with non-Hispanic white individuals.2,6,7 In contrast, Hispanic individuals with stage 5 CKD also seem to have a survival advantage.5,8 Despite various hypotheses, these paradoxic disparities with survival advantages for racial/ethnic minorities undergoing maintenance dialysis have not been satisfactorily explained. We undertook this study to test the hypothesis that black and Mexican American individuals may have a higher mortality than white individuals at earlier stages of CKD. Thus, the apparent survival advantage during stage 5 CKD may be a result of a higher mortality early during the course of CKD, wherein only the healthiest minority individuals survive long enough for their CKD to progress to require maintenance dialysis. RESULTS Patient Characteristics Of the 18,825 study participants who were ≥20 yr of age in the Third National Health and Nutrition Examination Survey (NHANES III), 715 did not identify themselves as white, black or Mexican American, and the data on follow-up interval was inadequate for 88 individuals (Figure 1). An additional 3411 individuals were excluded because they had missing data on urine albumin or serum creatinine. Thus, a total of 14,611 individuals formed the study cohort (Figure 1). Comparing the individuals included in the cohort, those with missing data on urine albumin or serum creatinine were slightly more likely to be black (12.6 versus 11.7%; P = 0.003), to be older (48.6 ± 0.6 versus 44.3 ± 0.5 yr; P < 0.001), and to have <12 yr of education (27.7 versus 23.1%; P = 0.005) and less likely to be obese (19.2 versus 22.8%; P = 0.04). Of the 14,611 individuals, 2892 had CKD, with a weighted prevalence of 14.1%. Comparisons of the risk factors among the three racial/ethnic groups with CKD are summarized in Table 1. The mean age of white individuals was 7 yr older than black individuals and 12 yr older than Mexican Americans. White individuals had the highest prevalence of hypercholesterolemia and a family history of premature heart disease and were most likely to go to a particular place for health care. Conversely, they were least likely to have diabetes, obesity, low income, or lower educational attainment or be uninsured (Table 1). Black individuals were significantly more likely to be hypertensive and had the highest prevalence of smoking (Table 1). Mexican Americans had the highest prevalence of diabetes, obesity, poverty, having less than a high school education, or being uninsured (Table 1). Outcomes The mean follow-up until the time to event or last follow-up for the cohort was 8.8 ± 0.2 to 9.0 ± 0.2 yr for individuals without CKD and 7.9 ± 0.2 yr for those with CKD. Among the 14,611 participants, 2171 died: 1127 with CKD and 1104 without CKD. The crude, all-cause mortality among individuals with CKD was 3.6 per 100 person-years compared with 0.6 per 100 person-years among those without CKD. Of these 2171 deaths, 969 (45%) were attributed to cardiovascular causes, 578 among those with CKD and 391 among those without CKD. Cardiovascular mortality among individuals with CKD was 1.8 per 100 patient-years compared with 0.2 per 100 patient-years without CKD. After adjustment for age, race, and gender, CKD was associated with 2.2-fold higher hazard ratio (HR; 95% confidence interval [CI] 1.9 to 2.5) for all-cause and 2.5-fold higher HR (95% CI 2.1 to 3.0) for cardiovascular mortality. There was no significant interaction between race and CKD for either all-cause or cardiovascular mortality. Among the 1127 participants who had CKD and died on follow-up, 707 were white, 243 were black, and 177 were Mexican American. Only five of these 1127 deaths were attributable to AIDS and three to homicide/suicide. After adjustment for age and gender, using white individuals as the reference group, black individuals with CKD had a substantially elevated risk for all-cause mortality; this risk was modified by age. The risk for all-cause mortality was only slightly attenuated upon further adjustment for traditional risk factors for cardiovascular disease (model 2) or for socioeconomic factors and access to care (model 3). Given the significant interaction of race with age on all-cause mortality, age-stratified analyses were conducted (Table 2). Among individuals who were younger than 65 yr, black individuals had a 78% higher risk for death compared with white individuals (Figure 2A). The higher risk for death in this age group remained significant even after adjustment for stage of CKD and cardiovascular risk factors (HR 1.53; 95% CI 1.01 to 2.31) but was attenuated to a NS level upon adjustment for socioeconomic and access-to-care variables (Table 2). Conversely, there was no significant difference in the risk for death in the older subgroups of black and white individuals (Table 2, Figure 2A). There was a trend toward a higher risk for all-cause mortality among Mexican Americans, with white individuals as a reference group; this trend, however, did not reach statistical significance. A similar trend for an interaction of age with ethnicity was observed among Mexican Americans (Table 2, Figure 2B). Entering the data on cardiovascular risk factors as continuous variables or adding laboratory data (hemoglobin, glycosylated hemoglobin, phosphorus, potassium, albumin, homocysteine, insulin, and C-reactive protein) did not change the HR for death among black or Mexican American individuals. Among the 578 deaths from cardiovascular causes, 384 events occurred among white individuals, 105 among black individuals, and 89 among Mexican Americans. Upon adjustment for age and gender, the risk for cardiovascular death was significantly and substantially higher among black compared with white individuals. The significant interaction of age with race was apparent for cardiovascular mortality as well such that the increased risk for cardiovascular mortality was limited to individuals who were younger than 65 yr (HR 2.11; 95% CI 1.00 to 4.47; Table 2). The same qualitative, albeit NS, trends were observed after adjustment for stage of CKD and cardiovascular risk factors or socioeconomic and access-to-care variables. There was a trend, albeit statistically NS, for a higher risk for cardiovascular death among younger Mexican Americans as well when compared with white individuals (Table 2). Sensitivity Analyses Three different sensitivity analyses were performed using modified definitions of CKD. First, only participants with estimated GFR (eGFR) <60 ml/min per 1.73 m2 were defined as having CKD (n = 996). The second analysis was performed to account for the fact that only 63% of the NHANES III participants had persistent microalbuminuria.9 For this analysis, CKD was defined as the presence of any one of the following: (1) Stage 3, 4, or 5 CKD; (2) stage 1 or 2 CKD with overt albuminuria (albumin/creatinine ratio >250 mg/g in men and >355 mg/g in women); and (3) a random selection of 54% of individuals with stage 1 and 73% of individuals with stage 2 CKD with microalbuminuria (total n = 2212).9 Finally, for the third analysis, all 3411 individuals with insufficient information to determine CKD status were considered to have CKD (n = 6207). The same qualitative trends that were observed in the main cohort were seen in each of the three sensitivity analyses performed (Table 3). Thus, there was a progressive decrease in the HR for all-cause mortality among black compared with white individuals, with increasing age in each of the three sensitivity analyses. The same qualitative trend for a higher risk for death was noted for Mexican Americans. DISCUSSION The key finding of this study may have important public health implications. Using a random sample of the population of the United States, black individuals with CKD were found to have a substantially higher risk for all-cause and cardiovascular mortality, a risk modified by age. Similar trends, albeit not statistically significant, were observed among Mexican Americans. In the general population, racial disparities have been reported with particularly inferior outcomes among black individuals with cardiovascular diseases and cancers.1–3 Disparities in several different outcomes also have been reported in populations with CKD. Thus, black individuals are less likely to receive erythropoietin in the predialysis stage of CKD, more likely to progress to ESRD, less likely to achieve clinical performance measure targets while undergoing maintenance hemodialysis, and less likely to be referred for transplantation and have a lower transplant allograft survival when compared with similar populations of white individuals.10–15 Our study adds to this list of disparities. Numerous studies have now documented the increased risk for death conferred by the presence of CKD, either by the presence of albuminuria or reduction in GFR16; however, up until recently, racial difference in mortality outcomes among non–dialysis-dependent patients with CKD has received only limited attention. Using pooled analyses of four community-based studies, Weiner et al.17 suggested that there may be a significant interaction between race and CKD. Thus, the presence of CKD conferred a greater increase in risk for the composite end point of death, nonfatal myocardial infarction, and stroke among black than among white individuals. The investigators attributed this interaction to residual confounding, with the differences possibly relating to greater severity of hypertension and/or diabetes in black individuals. Consistent with these findings, CKD has been reported to result in a significantly greater increase in risk for coronary artery disease mortality or incident congestive heart failure among blacks than white individuals.18,19 In our study, death was the only outcome measure studied, and we were unable to demonstrate any interaction between race and CKD. Our findings are in contrast to those reported among veterans with diabetes, wherein racial and ethnic minorities had a lower risk for death than white individuals, a survival advantage also seen among the subgroup with diabetic nephropathy.20 The differences in outcomes may be a result of better and equal access to care among veterans when compared with the general population. The modifying effect of age on the racial differences in outcomes has previously been reported in the general population but not among individuals with CKD.21 Thus, the health disparities seem to be greatest in the younger age groups. Similar outcomes among the elderly blacks and white individuals with CKD may reflect either survivor bias (such that only the healthiest black individuals survive into old age) or the benefit of federal/state programs for the elderly, such as Medicare, that provide improved access to care for participating individuals equally, thus minimizing racial disparities. The results of various studies that have examined the racial differences in mortality of elderly with CKD have come to differing conclusions. On the one hand, elderly black individuals who are hospitalized for congestive heart failure or survivors of acute myocardial infarction seem to have a survival advantage over white individuals.22–24 On the other hand, analyses using the Medicare 5% sample standard analytical files demonstrated that black individuals, with or without CKD, had a higher risk for death.25 Unlike previous investigations of the elderly, the cohort in this study was a random selection of noninstitutionalized population of United States, and the baseline assessment of renal function was comprehensive (including serum creatinine and albuminuria) and was done with individuals in a steady state and serum creatinine was calibrated to the Modification of Diet in Renal Disease (MDRD) laboratory measurements. Our study provides some insight into the reasons for the higher risk for death among younger black individuals with CKD, relative to white individuals. The higher risk for death among younger black compared with white individuals persisted despite multivariate adjustment for cardiovascular risk factors; however, the HR was attenuated to a NS level upon adjustment for socioeconomic and access-to-care variables. These data suggest that factors such as education, poverty, and lower probability of medical insurance may be more important in mediating the high risk for death among younger black individuals than are biologic differences. This, in turn, may help in prioritizing interventions aimed at reducing the disparities among younger black individuals with CKD. Recognition of this disparity among younger black individuals raises two important issues. First, the substantially higher mortality among black individuals with early stage CKD may suggest that the disparity in adverse CKD outcomes is far greater than previously suggested by studies restricted to ESRD progression. Second, the survival advantage of black individuals seen in studies of dialysis patients may be accounted for the higher risk for death seen in this racial group at early stages of CKD such that the black individuals who live long enough to reach stage 5 CKD are healthier than white individuals. This hypothesis is consistent with the recent observations from the Dialysis Outcomes and Practice Patterns Study that reported that minority individuals undergoing maintenance hemodialysis in the United States are healthier than their white counterparts, accounting for the former's survival advantage.5 In recent years, demographic shifts have made Latinos the largest minority group in the United States.26 Unlike the data for black individuals, the comparative data for Latinos are rather limited. Early studies in the general population indicated that despite a higher prevalence of cardiovascular risk factors (obesity and diabetes) and lower socioeconomic status, Latinos had lower all-cause and cardiovascular mortality when compared with non-Hispanic white individuals (“Hispanic paradox”).6 Subsequent data, however, suggested that outcomes among many subgroups of Latinos may actually be worse than that for white individuals.2,7,27 The data on outcomes for Latinos with CKD are far more limited. At least three studies suggested that Latinos, like black individuals, with ESRD may have a survival advantage when compared with white individuals.8,28,29 To our knowledge, this report is the first to describe racial differences in outcomes using a population that included a substantial proportion of non–dialysis-dependent Mexican Americans with CKD. The same qualitative trends, with a higher all-cause and cardiovascular mortality among Mexican Americans, were observed as those for black individuals; however, none of the trends reached statistical significance. This may reflect that there are no significant differences in outcomes of Mexican Americans and white individuals with CKD. Alternatively, this may be a result of relatively smaller sample size and fewer events, particularly cardiovascular, among Mexican Americans. Thus, larger studies with longer follow-up that allow more events to accrue will be needed to determine whether a similar disparity in outcomes exist for Mexican Americans with CKD. This study has several limitations. First, the sample size and the number of events for some subgroups were small, such as for Mexican Americans. This is also true for the age-stratified analyses and cardiovascular mortality. Second, it is possible that some Mexican Americans may have died in their country of origin and, therefore, their data may not have been captured by a review of the National Death Index (NDI). This may have led us to underestimate the risk for death in this ethnic population (salmon effect). Third, almost one fifth of eligible participants did not have enough data to determine the presence/absence of CKD; however, our sensitivity analyses suggest that this missing information is unlikely to have biased the study results. Fourth, data on risk factors were ascertained only at baseline; therefore, we could not systematically control for differences in severity of risk factors or perform time-dependent analyses. Finally, the baseline evaluation occurred before the benefits of contemporary therapies like blockers of the renin-angiotensin system were established. This study shows a higher all-cause and cardiovascular mortality for younger black individuals with early-stage CKD. The disparity in all-cause mortality remained after accounting for differences in prevalence of cardiovascular risk factors but was attenuated to NS levels after adjustment for socioeconomic status. Survivor bias, arising from the higher mortality during early stages of CKD, may explain the lower mortality seen among black individuals with ESRD. For determination of whether the noted trends for disparities exist for Mexican Americans with CKD, larger studies with longer follow-ups are needed. CONCISE METHODS NHANES III was a study conducted by the National Center for Health Statistics (NCHS) between 1988 and 1994.30 Study participants were selected using a complex, multistage, stratified, and clustered probability sampling of the noninstitutionalized population of the United States, and the baseline data were accrued between 1988 and 1994.31 The study design included a deliberate oversampling of black and Mexican American individuals and the elderly.31 The cohort for this study constituted white, black, and Mexican American individuals who were ≥20 yr of age and had CKD. For the survey, the participants self-identified as to a racial/ethnic using the and urine ratio were to determine The creatinine in the NHANES III were calibrated to the laboratory by as An was as having CKD when (1) was <60 ml/min per 1.73 m2 or (2) was but urine was mg/g in men and mg/g in of was defined as when the was or the was was defined as a total because only one third of the participants had data on was defined as a history of diabetes, being with for diabetes, or having a was defined as a were individuals who and had at least 100 during their The National was to stages of was as a or of were as having any health insurance or The effect of was by whether the was or of the to care was defined as whether a reported to a particular for care. of was determined by the baseline data to the mortality follow-up In this the the of status by the with a to be when a was made on the or on a death or For the the data in the were from the time of in the NHANES III until To the status to study participants using the the conducted a study to the for the of death was using the of up until and for and deaths before were by the into the for all deaths of the study participants, using are in the The of for the was to cardiovascular to and to This among of death to hypertensive heart coronary and heart and analyses were performed using sample that account for probability of and oversampling of elderly, black, and Mexican American individuals. were made using the variables are as ± individuals without a were to be on analysis was to determine the effect of CKD on survival using data from the cohort of 14,611 individuals. analyses were restricted to individuals with CKD, and three statistical were interaction with race were in the cohort with CKD; only one was statistically This interaction was included in all and age-stratified analyses were performed 65 to and Given the significant differences in age between the participants to the various racial/ethnic the survival analyses were for age, race, and (model 1). The second was to test whether the racial/ethnic differences in survival could be accounted for by differences in cardiovascular risk factors and thus in to race, and traditional cardiovascular risk factors diabetes, obesity, and family history of premature heart and stage of CKD (model 2). The third was to test whether the racial/ethnic differences in survival could be accounted for by differences in socioeconomic to care and thus in to socioeconomic factors and access to care to particular place for and stage of CKD (model 3). To the racial/ethnic differences in cardiovascular mortality, we performed similar analyses after participants who died of noncardiovascular of these analyses were on the three data for sensitivity analyses. the cohort was for analyses for this study from the 18,825 participants who were ≥20 yr of age in the NHANES of age on the differences in outcomes among and black individuals with CKD in the NHANES The HR for all-cause mortality, for age and gender, among black individuals in different age were as yr (95% CI to 65 to yr (95% CI to and yr (95% CI to of age on the differences in outcomes among white and Mexican American individuals with CKD in the NHANES The HR for all-cause mortality, for age and gender, among Mexican Americans in different age were as yr (95% CI to 65 to yr (95% CI to and yr (95% CI to of risk factors among white, black, and Mexican American individuals with of multivariate analyses of the of with all-cause and cardiovascular of the HR for all-cause mortality for black and Mexican American compared with white individuals using three different sensitivity is by a from the National of for this was also by National of Health and and and An on this was at the of the American of
Language: English
Subfield: Nephrology
Field: Medicine
Domain: Health Sciences
Open Access status: green
OpenAlex ID: W2142394319