A systematic review of ethnic differences in the rate of renal progression in CKD patients

Chronic kidney disease (CKD) is a public health problem of increasing importance, consuming a growing proportion of health care resources. It is estimated that, in the United States, there are 19.2 million individuals with CKD [1], and this figure is expected to increase in parallel to the rising prevalence of hypertension and diabetes. By 2010, the projected number of end-stage renal disease (ESRD) patients will climb to over 650 000 [2], with a further predicted increase to 2.24 million by 2030 [3]. Therefore, improved understanding of the predictors of GFR decline is essential. Information about predictors would allow nephrologists to accurately predict those CKD patients at risk of progressing to ESRD, which would help to individualize patient care, allowing for optimal planning for renal replacement therapy (RRT), reduce the need for urgent dialysis and ultimately allow for more efficient allocation of scarce health care resources. In addition, the recognition of novel risk factors for progression may lead to the development of new therapies capable of altering the trajectory of the disease. It has been recognized that the incidence of ESRD is higher in ethnic minorities; however, the reasons for this have not been well defined. For example, American blacks are four times more likely to require dialysis than whites [4]. An increased burden of ESRD has also been shown to affect Hispanics and Asians [5]. Conversely, ethnic minorities tend, paradoxically, to have improved outcomes once started on haemodialysis [5,6]. Several mechanisms have been suggested to explain such differences, though these have not been proven. For example, a higher prevalence of co-morbidities, lower socioeconomic status and comparatively worse access to health care among ethnic minorities have been cited as reasons for the higher incidence of ESRD [7–9]. Such reasoning holds that the incidence of ESRD is dependent on the number of CKD patients at risk of progressing, and the higher prevalence of diabetes and hypertension in blacks may result in greater CKD [4]. Similarly, lower socioeconomic status in minorities may create inequalities in access to health care resources and thus reduce delivery of medical management known to slow the progression of renal dysfunction [10,11]. A second explanation is that a higher death rate amongst CKD patients in one ethnic group would leave fewer patients alive to require RRT, thus affecting the observed incidence rate of ESRD. Different thresholds for starting RRT would also affect the measured incidence of ESRD. Finally, the higher incidence of ESRD amongst ethnic minorities may in fact be due to a faster rate of GFR decline and more rapid progression of renal disease. The purpose of this review is to summarize the available evidence on ethnic differences in the rates of CKD progression towards ESRD. Ideally, available studies would directly observe rates of GFR decline in CKD cohorts of different races. This would avoid such confounders as CKD prevalence, the competing risk of death and varying thresholds for starting RRT. Alternatively, in studies examining the incidence rates of ESRD, inferences can be made regarding progression rates only if attempts are made to account for differences in the baseline prevalence of CKD and in longitudinal mortality rates. MEDLINE, EMBASE and the Cochrane Database of Systematic Reviews were searched by separately combining the keywords ‘chronic kidney disease’ with each of the keywords ‘race’ and ‘ethnicity’ using the Boolean operator ‘and’, and for the MEDLINE search, with the Medical Subject Heading term ‘ethnic groups’. The search was limited to English papers in adult humans, with no limits placed on the date range. Titles and abstracts were screened, and when indicated, the content of the papers was searched to meet our inclusion criteria. The reference lists of included papers were searched for appropriate studies. Our inclusion criteria were prospective cohort studies that directly observed the rates of renal progression in different ethnic groups with known all-cause CKD (as measured by rates of ESRD or changes in creatinine or measured or estimated GFR over time) or cross-sectional studies that compared the incidence of all-cause ESRD in different ethnic groups while adjusting for the prevalence of CKD. Studies were excluded if they addressed only specific causes of renal disease, were non-English, were case reports or case series or looked at paediatric populations. Study results were summarized in table format, including documentation of confounders and other potential limitations of the study designs. This search strategy was independently performed by two authors (S.B. and M.S.), and disagreements were resolved by consensus discussions. The selection criteria were chosen in order to specifically answer the question of whether there are ethnic differences in the rates of renal function decline in patients with all-cause kidney disease. Ethnic differences in the rate of progression of specific kidney diseases is a related, but different, question; therefore, the inclusion criteria were specifically chosen so as to address CKD in general and not specific subtypes of renal disease. In addition, the inclusion criteria were chosen so as to include studies with the pre-specified goal of comparing different ethnic groups. The authors did not attempt to extract data from different studies to make novel comparisons. Prospective data that directly observed rates of disease progression were regarded as ideal; however, cross-sectional studies were also acceptable if both the prevalence of CKD and the incidence of ESRD were assessed in two or more discrete ethnic groups. Studies that cited ESRD prevalence data alone were discarded, as the incidence of ESRD is dependent on both the prevalence of CKD and the rate of renal progression through CKD. The MEDLINE search yielded 1252 results, EMBASE yielded 462 results and the Cochrane Database of Systematic Reviews yielded 7 results, for a total of 1721 potential studies. After review, 21 studies were selected. Of these, on careful review of the manuscripts, 16 studies were excluded. In total, five studies were found to have met our inclusion criteria, and are summarized in Table 1 [12–16]. Several of the studies were found to address the topic of ethnic progression, but did not meet our inclusion criteria due to methodological issues. Those studies that were excluded from the systematic review but are cited frequently in the literature are summarized in Table 2 [17–24] in order to ensure completeness in this paper. Summary of included studies GFR in millilitres per minute per 1.73 m2. VA, Veterans Affairs; DM, diabetes; KP, Kaiser Permanente. Summary of included studies GFR in millilitres per minute per 1.73 m2. VA, Veterans Affairs; DM, diabetes; KP, Kaiser Permanente. Summary of the studies found to address the topic of ethnic progression but did not meet the inclusion criteria, however, in the authors’ opinions, are still worthy of discussion GFR in millilitres per minute per 1.73 m2. KP, Kaiser Permanente. VA, Veterans Affairs; DM, diabetes; SES, socioeconomic status. Summary of the studies found to address the topic of ethnic progression but did not meet the inclusion criteria, however, in the authors’ opinions, are still worthy of discussion GFR in millilitres per minute per 1.73 m2. KP, Kaiser Permanente. VA, Veterans Affairs; DM, diabetes; SES, socioeconomic status. Brazy and Fitzwilliam provided one of only two studies that measured renal function decline over time in ethnic cohorts with CKD [12]. This small study analysed 200 black and white men with CKD who were followed up at a Veterans Affairs (VA) nephrology clinic from 1979 to 1988. The rate of renal function decline (assessed by the slope of the inverse creatinine concentration versus time) was similar in the black and white cohorts over a 38-month follow-up period, leading the authors to conclude that, once kidney disease has developed, similarly treated black and white patients demonstrate comparable rates of decline. This study design has several advantages in that all patients had the same access to health care, were all followed up by nephrologists and the results are independent of the mortality rate or thresholds for starting RRT. One major limitation to this study is that the authors included only those patients who had at least a 20% increase in their serum creatinine. Therefore, the results can only be generalized to male CKD patients who were specifically selected for having an established tendency toward progressive decline in renal function. The 2009 study by Choi et al. [13] used a cohort of white and black VA patients with at least one outpatient creatinine measurement. Although the majority of patients had normal renal function, 22% of whites and 13% of blacks had estimated GFR (eGFR) <60 mL/min/1.73 m2. Because of the large size of the overall cohort, this forms a significant group of CKD patients that were followed up for 3.7 years and showed a faster rate of eGFR decline in blacks compared to whites. In addition, the rate of ESRD was higher in blacks at all levels of initial eGFR, with a higher mortality rate in blacks arguing against a competing risk of death. This study suggests that, in a population of mostly male VA patients, the rate of renal function decline in those with CKD is faster in blacks compared to whites and that a competing risk of death does not explain differences in ESRD rates. This study design has the advantage of following up patients over time; however, the results may not be generalized outside of the specific population studied and the results may be confounded by both level of ancillary health determinants (socioeconomic factors) and additional health care services available to the white patients who were of higher socioeconomic status. The study by Walker et al. [22] also measured creatinine changes in cohorts of hypertensive black and non-black patients in a subgroup of the MRFIT trial, showing that black patients had a more rapid decline in renal function. However, the study included mostly patients with normal renal function at baseline and, therefore, the observed decline in renal function in blacks is dependent on both the rate of developing CKD and the rate of GFR decline once renal disease is established. This study cannot provide insight into relative CKD progression rates. A significant confounder in ethnic minority cohort studies is access to health care. If one cohort has poor access to specialists, health care resources or optimal medical management known to alter the rates of progressive GFR decline, the conclusion that ethnicity causes different rates of progression through CKD may be flawed. Several of the studies attempted to control for this by using cohorts of insured patients in the American health care system, such as the VA system [12,13,21,24], Medicare [20,23] or a group of Kaiser Permanente (KP)-insured patients in northern California [16–18]. However, it is not clear that patients in such groups actually had equal access to health care. In most of these studies, indices of lower socioeconomic status at baseline such as education level, household income and living below the poverty line were more prevalent in the ethnic minority groups [13,15–18,20,24]. Such observations allow for the possibility that a period of substandard access to care and/or uncontrolled disease activity prior to enrolment in insured cohorts may impact upon future progression rates. In the study by Li et al., baseline access to care was directly assessed by measurements of annual diabetic eye exams, other regular preventative measures and access to primary care physicians, all of which were worse in blacks [20]. Several of the studies did not measure baseline socioeconomic or health care access variables [21,23]. In addition, it is possible that, within insured groups during the follow-up periods of the studies, there is disparate access to health care in the ethnic minority cohorts. Hispanics insured by KP in northern California were more likely than their white counterparts to drop out of the health care plan during the follow-up period [16]. The impact of variable health care access among ethnicities represents a strong confounder in analysing such studies and only those that take careful measures to control for this may be interpreted correctly. Of the cohort studies that only used incidence rates of ESRD (as opposed to measured changes in GFR) as their outcome measure, only the study by Peralta et al. assessed a cohort of patients with established CKD [16]. Nearly 40 000 Hispanic and non-Hispanic whites insured by KP in northern California with creatinine measurements recorded between 1996 and 2002 and baseline eGFR between 15 and 59 mL/min/1.73 m2 were followed up for a mean of 3.83 years. The unadjusted hazard ratio (HR) for incident ESRD in Hispanics versus whites was 1.99. After sequential adjustments for age, gender, socioeconomic status, co-morbidities and known predictors of renal progression (proteinuria, hypertension, diabetes, baseline GFR and medication use), the higher risk of ESRD in Hispanics was attenuated but not completely explained (adjusted HR of 1.33). Baseline creatinine and eGFR distribution were qualitatively similar between the groups, suggesting that the study was not biassed because of a higher prevalence of CKD in one cohort. The authors concluded a faster rate of progression towards ESRD in Hispanic compared to white CKD patients. However, the mortality rate was lower among Hispanics, raising the possibility of a competing risk of death in whites leaving more Hispanics alive to progress to ESRD. The degree to which other potential predictors of renal progression that were not measured (such as BMI, smoking, haemoglobin A1C or acidosis) may explain the residual adjusted risk in Hispanics is not known. The insured group of KP patients in northern California is known to differ from the general population at the extremes of income and in unemployed individuals [17], which must be considered when generalizing the study results. Several cohort studies either did not measure baseline renal function or included patients with normal renal function, limiting their utility in describing ethnic differences in rates of progression through CKD. In the studies by Choi et al. [24] and Hall et al. [17], baseline renal function was assessed and included some patients with abnormal GFR; however, a large percentage of enrolled patients had normal renal function. Both studies found the incidence of ESRD to be higher in the ethnic minority groups (blacks and Asians) compared to whites. In such a study design, the incidence of ESRD is dependent on the risk of developing CKD, the rate of progression through CKD, the threshold for starting RRT and hence ‘counting’ as ESRD and the competing risks of death or loss to follow-up during the study period. It is not possible to differentiate the effects of developing CKD from the rates of progression once CKD has started. In the remaining cohort studies, which we excluded from the main analysis, baseline renal function was not assessed [18–21, 23]; therefore, the universal finding of higher ESRD incidence in the ethnic minority cohorts cannot be adjusted for differences in CKD prevalence, making conclusions about CKD progression rates impossible. There are two cross-sectional studies that looked at the incidence of ESRD while attempting to adjust for the prevalence of CKD. The first, by Hsu and colleagues [15], looked at black and white CKD patients (estimated GFR 15–59 mL/min/1.73 m2) aged 20–74 years from the NHANES III data (the midpoint of which was in 1991). To capture the rate of progression from CKD to ESRD, the authors assumed that it would take 5 years to progress to RRT and thus assessed the incidence of ESRD using the United States Renal Data System (USRDS) data from 1996 for patients aged 25–79 years. The analysis was repeated, assuming anywhere from 3 to 8 years to progress, and showed consistency of results. The study found no difference in the prevalence of CKD between blacks and whites (using the above eGFR definition based on the MDRD formula); however, the incidence of age-, sex- and diabetes-adjusted ESRD was almost five times more common in blacks. The conclusion that blacks must progress faster through CKD is at best speculative. The patients were not observed to progress to ESRD; instead, this was deduced from data taken at two points in time. There are many confounders which may explain this trend, including unequal access to health care, the competing risk of death or inappropriate assumptions in the study design. In addition, the black CKD patients in the NHANES III data had worse hypertension and proteinuria, the extent to which this contributed to their higher ESRD rate is unknown. Other than age, sex and diabetes, no other known risk factors for renal disease progression were assessed for their effects on the observed trend. The second such study by Hallan et al. [14] used data from the HUNT II survey in Norway taken from 1995 to 1997, comprised largely of whites in a country with a fully universal health care system and compared them to American whites. The prevalence of CKD was based on a definition consistent with the K/DOQI guidelines [2] and was taken from the HUNT II data set (age and gender standardized to the US population in 1991) and the NHANES III data with a midpoint of 1991. The incidence of ESRD was taken from the USRDS data and the Norwegian Renal Registry from 1995 to 1997. The prevalence of CKD was similar between Norwegian and American whites, including when stratified for age, gender, diabetes, hypertension and CKD stage. Despite this, the incidence of ESRD was 2.5 times higher in American whites than their Norwegian counterparts. The eGFR at the time of RRT was similar in the two groups, suggesting similar thresholds for starting dialysis. However, markers of pre-dialysis care (such as visits to nephrologists, presence of AV fistulae, use of erythropoietin-stimulating agents and haemoglobin and albumin levels) all indicated better access to appropriate CKD care in Norwegian whites. This underscores the potential importance of health care access and its confounding in observational studies. Similar to the study by Hsu and colleagues, the progression to ESRD was deduced and not observed and, therefore, subject to the same limitations. In addition, the use of two different data sets in two different countries taken at different points in time to define each of CKD and ESRD introduces many sources of error. Therefore, this study cannot be used to conclude different intrinsic rates of progression through CKD in American versus Norwegian whites, but rather, succeeds in highlighting the importance of implementing measures which aim to slow progression. This review highlights that despite what may be a common assumption amongst nephrologists, the available evidence to date does not conclusively support the hypothesis of ethnic differences in the rates of progression through all-cause CKD. There are few properly designed studies that address this issue, and several often-cited studies have some methodological shortcomings that make interpretation difficult. An ideally designed study to answer this question would prospectively follow-up CKD ethnic cohorts, with equal access to care, and repeatedly measure or estimate GFR over time. This would avoid the competing risk of death, a different prevalence of CKD or a different threshold for starting RRT in one group. The closest study to such a design was by Brazy and Fitzwilliam [12] and did not suggest a different rate of progression in black versus white CKD patients. This is in contrast to many other cohort studies that looked at the incidence of ESRD and suggested a worse burden of disease in ethnic minorities, but had many potential confounders such as unmeasured co-morbidities or unequal access to care. Once again, when such factors are not quantified and compared among racial groups, it is difficult to make firm conclusions about race-mediated predilection to progression. It has been clearly established that ethnic minorities suffer a greater yearly incidence of ESRD compared to whites, but the evidence to date does not support a role for a more rapid rate of progression through CKD. Future studies will be to an between ethnicity and CKD progression. An of all variables with renal disease progression will be to out the of risk factors versus intrinsic of specific The Chronic Renal study may additional evidence on this It is a large prospective cohort study in the US that is predictors of CKD progression and of disease amongst an group of In a renal function will be estimated using the of the MDRD in different ethnic groups. However, ethnic differences in socioeconomic status and access to health care may to be a access to health care or optimal medical management in minority groups is a problem with a found in the problem and in changes to health care and The competing risk of mortality differences in CKD patients of different ethnicities must also be In the that ethnicities are to progress at reasons for it must be fully may be with differences in or renal to to and or other factors in the progression of renal disease. the increased risk of ESRD in compared to has been to a II it is this renal function decline such regarding risk factors for progression to ESRD will to and future It will also the allocation of health care resources to those patients who are more likely to progress to ESRD and those patients who are

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.