University of Southern California Dissertations and Theses

A biodemographic approach to understanding sociodemographic disparities in kidney functioning on three dimensions: individual, population, and cross-national

(USC Thesis Other)

A biodemographic approach to understanding sociodemographic disparities in kidney functioning on three dimensions: individual, population, and cross-national

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content A BIODEMOGRAPHIC APPROACH TO UNDERSTANDING SOCIODEMOGRAPHIC
DISPARITIES IN KIDNEY FUNCTIONING ON THREE DIMENSIONS:
INDIVIDUAL, POPULATION, AND CROSS-NATIONAL
By
Erfei Zhao
A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GERONTOLOGY)
August 2024
Copyright 2024 Erfei Zhao

ii
Dedication
I would like to thank my family for enduring the hardship they had to endure for me to be where I
am today.

iii
Acknowledgements
First, I would like to thank my mentor, Dr. Eileen Crimmins, for her immense support, patience,
and insights. I feel so incredibly fortunate that I got to receive professional training from a top
researcher in the field like her. I simply could not have completed this dissertation without her.
I would like to thank Dr. Jennifer Ailshire and Dr. Jung Ki Kim for their guidance throughout
this dissertation. You are the type of researchers that I aspire to become in the future.
I would also like to thank my cohort in this program whom I am lucky enough to call friends –
Qiao Wu, Margarita Osuna, Gillian Fennell, and Rachel Wilkie. Here is to 5-years of laughter,
tears, loyalty, and a bond that lasts a lifetime.
I would like to thank the members of our lab who have offered me tremendous support –
Eunyoung Choi, Calley Fisk, Eric Klopack, Yuan Zhang, Mateo Farina, Addam Reynolds,
Hyungmin Cha, and Kristina Dang.
Thank you, Dr. Elizabeth Zalinski, for offering me support when I needed it the most. And thank
you for coming to my dance competition.
Lastly, thank you Benny (my dog) for the emotional support and keeping me sane throughout the
pandemic.

iv
Table of Contents
Dedication………………………………………………………………………………………...ii
Acknowledgements……………………………………………………………………………...iii
Abstract…………………………………………………………………………………………..ix
Chapter 1: Introduction………………………………………………………………………....1
1.1 Kidney disease: A neglected public health emergency……………………………………....1
1.2 The sociodemographic disparities in kidney health………………………………………….3
1.3 A biodemographic approach to understanding differentials in kidney health……………….5
1.4 References…………………………………………………………………………………..10
Chapter 2: Associations Between Change in Kidney Functioning, Age, Race/Ethnicity,
and Health Indicators in the Health and Retirement Study……………….………………...20
2.1 Introduction…………………………………………………………………………………20
2.2 Methods……………………………………………………………………………………..23
2.2.1 Data…………………………………………………………………………………...23
2.2.2 Measures……………………………………………………………………………...24
2.2.3 Statistical Analysis…………………………………………………………………....26
2.3 Results………………………………………………………………………………………28
2.4 Discussion…………………………………………………………………………………..32
2.5 References…………………………………………………………………………………..42
Chapter 3: Discrepancies in age, sex, and racial/ethnic patterns of kidney functioning
using various indicators of kidney functioning…………………………………………….....48
3.1 Introduction…………………………………………………………………………………48
3.2 Methods……………………………………………………………………………………..51

v
3.2.1 Data………………………………………………………………………….………..51
3.2.2 Measures……………………………………………………………………………...52
3.2.3 Statistical Analysis……………………………………………………………………54
3.3 Results………………………………………………………………………………………55
3.3.1 Sample Characteristics………………………………………………………………..55
3.3.2 CKD Prevalence………………………………………………………………………55
3.3.3 Age, Sex, and Race/Ethnicity Patterns……………………………………………….56
3.3.4 Predicting Morality…………………………………………………………………...58
3.4 Discussion…………………………………………………………………………………..58
3.5 References…………………………………………………………………………………..70
Chapter 4: Comparisons of SES disparities in CKD prevalence and progression over
time in China and U.S..…………………………………………………………………………78
4.1 Introduction…………………………………………………………………………………78
4.2 Methods……………………………………………………………………………………..81
4.2.1 Data..………………………………………………………………………………….81
4.2.2 Measures……………………………………………………………………………...82
4.2.3 Statistical Analysis……………………………………………………………………86
4.3 Results………………………………………………………………………………………87
4.3.1 Comparison of sample characteristics between China and the U.S. …………………87
4.3.2 Associations between baseline kidney functioning and SES/related risk factors…….89
4.3.3 Associations between change in kidney health over 4-year and SES/related risk
factors………………………….………………………….……………………………………...92
4.4 Discussion……………………………………………………………………………..........93

vi
4.5 References…………………………………………………………………………………108
Chapter 5: Conclusions and Outlook………………………………………………………..114
5.1 References…….…………………………………………………………………………...119
References…………………………………………………………………………………..….123
Appendix A….…………………………………………………………………………………140

vii
List of Tables
Table 2.1 Weighted Mean Cystatin C at All Waves and Sample Characteristics at Baseline…..38
Table 2.2 Effects of Gender, Age, Health Conditions, Health Behaviors, and Polygenic Score
(PGS) on Baseline Cystatin C and Annual Cystatin C Change With 8 Years of Age…………...39
Table 2.3 Racial/Ethnic and Education Differential in Kidney Functioning Decline and Role
of Health Conditions and Health Behaviors in Explaining Disparities……………...…………..40
Table 2.4 Associations Between Kidney Functioning and Health Conditions and Health
Behaviors and Polygenic Score (PGS) Within Each Racial/Ethnic Group……………………...41
Table 3.1 Weighted Characteristics of Community-dwelling Population Aged 56 Years and
Older in 2016………..………..………..………..………..………..………..………..………….65
Table 3.2 Odds Ratios Indicating Sex/Racial/Ethnic Differentials from Logistic Regressions
on Six Measures of Impaired Kidney Functioning………………………...…………………….66
Table 3.3 Odds Rations Indicating from Multinomial Regressions Indicating Sex/Racial/
Ethnic Differentials in Severity Level of Kidney Functioning………….……………………….67
Table 3.4 Sensitivity and Specificity of Equations Predicting All-cause Mortality Among
Population with Impaired Kidney Functioning Defined by Each Kidney Measure, Stratified
by Sex and Race………………...………………………………………………………………..68
Table 4.1 Weighted Baseline Characteristics of Respondents with Both Waves of Cystatin C
in China and U.S. in 2011………………………………………………………………………..98
Table 4.2 Association between baseline kidney functioning, SES and CKD risk factors……..101
Table 4.3 Age-Controlled Probability of Kidney-related Health Outcomes Within 4-Year…..103
Table 4.4 Multinomial regression: Associations between change in kidney functioning, SES
and CKD risk factors…………………….……………………………………………………..104

viii
Table 4.5 Cox Proportional-Hazards Model on Mortality among People with Moderately to
Severely Impaired Kidney Functioning Over 4-Year…………………………………………..106

ix
List of Figures
Figure 3.1 Prevalence of impaired kidney impairment functioning with seven measures
(with 95% CI)…………………………………………………………………………………....66
Figure 4.1 Comparison of Age-adjusted Prevalence of CKD stages between China and the
U.S………………………………………………………………………………………...……100

x
Abstract
Chronic kidney disease (CKD) is one of most prevalent and expensive diseases and it can
severely impair people’s quality of life in older adulthood. In the U.S., CKD affects one in every
three Americans over the age of 65 and makes up for nearly one-quarter of all Medicare
spending; globally, the number of individuals with all-stages of CKD was nearly 700 million in
2017, surpassing some other major diseases (e.g., diabetes, asthma, or depressive disorders).
CKD affects disproportionally those that are socially disadvantaged, so that non-Hispanic Black
individuals and people with low socioeconomic status typically have worse kidney functioning at
older ages. However, despite the high prevalence, our understanding in this disease is lacking,
and the sociodemographic disparities in kidney functioning may be more nuanced and
complicated than previous research has indicated. This dissertation uses a biodemographic
approach, a method that integrates biomarker data into population-level studies, to expand our
understanding of the sociodemographic disparities in kidney functioning. Biomarkers, which are
characteristics that are objectively measured and evaluated as indicators of biological processes,
are increasingly used to clarify the physiological mechanisms behind how social factors “get
under the skin” to affect population health.
In Chapter 1 of my dissertation, I outline the significance of chronic kidney disease and
its impact on the growing older population both in the U.S. and across the globe. I then discuss
the aims of this dissertation in which I explore disparities in kidney functioning at three
dimensions – individual, population and cross-national. In Chapter 2, I examine how individual
trajectories of kidney function change with time across sociodemographic sub-groups. The
results show that non-Hispanic Black persons have worse baseline kidney functioning than non-

xi
Hispanic White persons, and Hispanic persons have faster decline in kidney functioning than
non-Hispanic White persons. Age remains a significant predictor of decline in kidney
functioning, and its association is not fully explained by health conditions/behaviors, or genetics.
Better management of diabetes, heart conditions, and obesity is effective in slowing this decline.
Baseline differences in kidney functioning (e.g., between non-Hispanic White and Black
persons; those with and without hypertension) suggest disparities occur early in the life course
and require early interventions.
In Chapter 3, I find differential sex and racial/ethnic patterns across indicators. Women
are more likely to have impaired functioning using serum level of creatinine (Scr), Glomerular
Filtration Rate (GFR) estimated by Cystatin C (CysC), and race-adjusted GFR estimated by both
CysC and Scr; men have worse kidney functioning using race-free GFRcysc_scr. Black and
Hispanic participants are more likely than White participants to have kidney failure on all
indicators. Race-free indicators generally indicate worse early-stage kidney disease among nonHispanic Black individuals than White individuals, whereas race-adjusted indictors generally
indicate the opposite conclusion, which is significant as early-stage CKD is a critical time for
implementing preventive care effectively. GFRcysc is the most predictive of all-cause mortality.
In Chapter 4, our finding suggests one’s socioeconomic status is closely related to both
baseline and progression of kidney health among older adults in China and the U.S., although the
mechanisms behind those associations are different. The education inequality in kidney health
cannot be explained by risk factors that were considered typical in the U.S., which points to the
need for further understanding of additional underlying mechanisms that link education
inequality with kidney in the Chinese context. The common risk factors between the two
countries are lack of effective control of diabetes and high BMI, and the rest are country-specific

xii
risk factors - presence of heart problems for the U.S. and hypertension for China. Our findings
also highlight the importance of availability of insurance regarding kidney health – having access
to private insurance plays a preventive role in both level and progression in kidney functioning
and having urban employee insurance even has a protective effect on CKD-related mortality in
China. By fully utilizing the longitudinal and nationally representative nature of the Health and
Retirement Study (HRS) and China Health and Retirement Longitudinal Study (CHARLS),
findings from these three studies could contribute to existing literature by presenting a more
complete picture on how and why kidney disparity exists across sub-populations, which in turn
led us to multiple major takeaways that have significant public health implications.

1
Chapter 1: Introduction
1.1 Kidney disease: A neglected public health emergency
The aging process is accompanied by progressive deterioration in the functioning of vital
organs. One of those age-related declines in organ functioning is characterized by Chronic
Kidney Disease (CKD), which is a syndrome defined as persistent alterations in kidney structure,
function or both with implications for the health of the individual (1). While some examples of
structural abnormalities are more visible and evident on imaging (e.g., cysts, tumors,
malformations and atrophy), kidney dysfunction can also manifest as hypertension, oedema,
changes in output or quality of urine, etc., changes that are most often recognized by increased
serum levels of creatinine and cystatin C (2-4), which are subsequently used as biomarkers to
indicate kidney functioning in research.
CKD is one of the most prevalent and expensive diseases. According to the 2019 annual
report of the United States Renal Data System, CKD accounts for nearly a quarter of all
Medicare spending, surpassing 120 billion dollars in the United States (5). CKD is extremely
common in the older population - on average, one in every three older Americans is at risk for
some degree of CKD (6) - in 2017, older adults over the age of 65 made up over half of the
patients with incident end-stage renal disease (ESRD), which is the most severe stage of CKD,
and are in need of renal replacement treatment (7). Globally, the total number of individuals with
all-stage CKD was nearly 700 million in 2017, surpassing those with diabetes, asthma, or
depressive disorders (8, 9). By 2024, about 850 million people worldwide are estimated to have
CKD, many of whom live in developing countries, in which a large proportion of these
individuals lack access to kidney disease diagnosis, prevention or treatment (10, 11). Despite the

2
fact that rapid population aging around the globe will translate to large increases in the
prevalence of CKD in the coming decades, awareness for CKD remains substantially lower as
compared with that of other chronic diseases such as hypertension and diabetes, and it often goes
undetected until its later stages (12 - 15) – according to centers for disease control and
prevention, in the U.S., the overall prevalence of being aware of having CKD was 24.8%; while
most people are aware (7 out of 10) of their conditions after they enter more severe stages of
CKD, only 16.3% are aware of their conditions among those with mildly to moderately impaired
kidney functioning (6). Some studies suggest the awareness of CKD is even lower in developing
societies, in which as many as 9 out of 10 individuals with CKD in resource-poor regions with
weak primary care infrastructure are unaware of their conditions and thus do not seek treatment
(16, 17).
Further, CKD is associated with very high mortality even in comparison to some of the
most prevalent causes of deaths - in 2017, CKD diagnoses resulted in 1.2 million deaths, which
are more deaths than tuberculosis or HIV - and this CKD-related mortality rate has continued to
be on the rise over the past 25 years (8, 18, 19). Currently, kidney disease is the third fastestgrowing cause of death globally and the only non-communicable disease to exhibit a continued
rise in age-adjusted mortality (18, 19). Studies project that, by 2040, CKD is projected to be the
5th highest cause of years of life lost globally (20).
Among those that live with CKD, the symptom burden is profound. Kidney disease is
associated with multiple adverse consequences, including disability, functional limitations,
reduced quality of life, poor life participation and mental illness (21). Those with kidney failure
may experience a similar or sometimes even greater symptom burden than those with terminal
malignancies (22). Even among patients who were not on dialysis or transplantation, many

3
reported a high symptom burden (e.g., poor mobility, bone and/or joint pain, insomnia, anxiety,
sexual dysfunction) (22). Among older adults, who typically have lower access to
dialysis/transplants than the younger adults, quality of life further decreases, and symptom
burden increases for years before starting dialysis/transplants (23, 24). Additionally, having a
person with CKD in the family severely jeopardizes the mental wellbeing of caregivers, half of
whom report symptoms of anxiety or depression (25, 26).
The most common underlying diseases associated with CKD are diabetes mellitus and
hypertension, particularly in developed countries (e.g., among those with diabetes, CKD
prevalence is estimated at 30–40% in the U.S.) (27, 28). However, in developing countries, CKD
may also be associated with infectious diseases, glomerulonephritis (a group of diseases that lead
to inflammation of the glomerulus) and inappropriate use of medications (such as traditional
remedies with potential nephrotoxins) (29, 30). In developing countries, with the current trends
in increasing disparities in socio-economic status and a growing aging population, the diabetes
and obesity epidemic might become major etiological causes for CKD and will further increase
the absolute number of people with CKD (31).
1.2 The sociodemographic disparities in kidney health
CKD are characterized by marked differences in incidence, prevalence, and/or
complications across sex, race/ethnicity, socioeconomic status (32-37). Specifically, studies have
shown that the overall prevalence of CKD is higher in women compared with men, but the
lifetime risk of end-stage-renal disease (ESRD) and kidney-related death are higher among men
(32, 33, 38). Men tend to experience a faster decline in kidney function than women, resulting in
a higher risk of kidney failure; this faster decline is partly caused by a greater prevalence of

4
unhealthy lifestyle behaviors (39, 40). Nonetheless, women with CKD are less likely than men to
be aware of their condition, screened for and diagnosed with CKD and referred to nephrologist
care (33, 41). Racially/ethnically disadvantaged populations exhibit a disproportionate burden of
kidney impairment, which results in higher risks of mortality, multiple morbidity, and substantial
impairment in quality of life for those who are socially marginalized (34, 35, 42). For instance, in
the U.S., while there is some evidence that the prevalence of early CKD is comparable across
racial/ethnic groups, the progression of CKD to ESRD is far more rapid among minority
populations, the burden of ESRD falls disproportionately on black Americans, as well as other
minority populations. Although Black Americans comprise only 13% of the United States
population, they made up over 30% of patients with ESRD in the U.S. (43). Those disparities
largely attributable to higher prevalence and greater severity of diabetes and hypertension, lower
socioeconomic status, lesser access to care, excess exposure to environmental toxins, and other
factors (44).
There is also abundant evidence suggesting CKD is patterned by socioeconomic status
(SES), which generally indicates an individual or group’s relative position in an economicsocial-cultural hierarchy (36, 37). The conceptualization of SES generally relates to two aspects:
1) access to resources and 2) class/rank in relation to others in society. In order to capture these
aspects of SES, studies in health research typically use objective and quantifiable measures such
as (e.g., income, education, wealth, occupation, etc.) (45). While there is no standardized SES
measure that is applicable across populations, studies have shown consistent associations
between SES with CKD, irrespective of how SES is measured (37). This may be interpreted as
that both access to resources and prestige or place in the social hierarchy within a society are
relevant to the likelihood of having CKD. It is believed that, in the U.S., health behaviors (e.g.,

5
smoking, physical activity, diet), comorbid conditions (e.g., hypertension, diabetes), and access
to care lie on the causal pathway between SES and poor kidney functioning (46-48). However, it
is believed that developing countries may face additional risks that translate to a greater kidney
disease burden that disproportionally affect people who are socially disadvantaged, such as
environmental toxins, air pollution, infectious diseases, etc. (49, 50)
However, CKD and its impact on different sociodemographic groups remains
understudied in many areas of the world, including the U.S. (9). First, the empirical work on
kidney progression is largely based on cross-sectional work and does not actually observe
progression, lacking empirical support for mechanisms behind racial/ethnic and SES differentials
in terms of progression in kidney disease over time (51); second, an increasing body of literature
has pointed out that differentials in CKD prevalence is not consistent across studies due to
reasons such as the unstandardized use of kidney measures, leading to potential complication in
informing health policy with their different or even conflicting results regarding CKD disparities
(52, 53); finally, despite differences in both epidemiological and policy contexts across countries
related to kidney diseases, there have been no studies, to our knowledge, that observe how
sociodemographic disparities in kidney functioning vary cross-nationally and how they relate to
common risk factors.
1.3 A biodemographic approach to understanding differentials in kidney health
For my dissertation, I take advantage of the available biomarker data in the Health and
Retirement Study (HRS) and the China Longitudinal Health and Retirement Study (CHARLS),
specifically serums such as cystatin C and creatinine, which are used to indicate kidney
functioning. A large body of biological, clinical, and epidemiological research recognizes the
importance of biomarkers as early indicators of physiological change and important determinants

6
of chronic diseases. Subsequently, many studies in the past decade have incorporated biomarkers
into population-level studies to clarify the physiological mechanisms behind how the social and
demographic variables “get under the skin” to affect health, especially in the context of chronic
diseases concentrated in the older population (54, 55). There are several advantages of this
biodemographic approach. First, the quantifiable nature of biomarkers becomes particularly
useful for us to accurately model and understand the process of physiological changes in kidney
functioning over time; second, biomarker-based diseases assessment is better than traditional
survey method as it does not rely on interaction with the healthcare system and respondent selfreports which often creates reporting bias (54, 56, 57); lastly, as early detection is the key
strategy in preventing kidney disease, its progression and related complications, there is
tremendous value in knowing the level of kidney functioning prior to disease diagnosis or before
symptoms arise, which can be a critical window for implementing preventive measures
effectively (58). By applying this biodemographic approach that directly assess biomarker data
to indicate kidney functioning (e.g., Cystatin C), we have access to more information regarding
the exact level of severity in kidney functioning than we would get from diagnosis and selfreporting, which in turn paints a more comprehensive picture of the sociodemographic disparities
in kidney functioning as we explore kidney functioning at three levels: individual, population,
and cross-national:
Dimension I: Individual – How and why do individual trajectories of
kidney deterioration vary by race/ethnicity and education?
First, at the individual level, though it has been accepted that non-Hispanic Blacks and
those with lower SES may experience faster kidney deterioration, researchers have recognized

7
that most existing data are cross-sectional and cannot clearly identify the epidemiologic forces
driving these disparities (59, 60). As most major causes of aging-related kidney deterioration are
preventable and modifiable, including comorbidities (e.g., diabetes, hypertension, and
cardiovascular diseases) and adverse health behaviors (e.g., physical inactivity/obesity, smoking
and problem drinking), (61), understanding those underlying mechanisms behind the differentials
in progression at older ages provides an opportunity to alleviate sociodemographic disparities in
kidney deterioration. However, to date, no longitudinal studies have examined how health
correlates influence the rate of kidney deterioration among older adults (51). Thus, the first aim
of this project is to examine how one’s trajectory of kidney deterioration is differentiated by
race/ethnicity and education attainment and determine the health correlates that associate with
those differentials among older Americans. This is the first study that uses a nationally
representative sample of aging individuals to observe longitudinal change in kidney health. This
sample includes large numbers of ethnically diverse respondents, which may provide insights
into and implications for treatment to help minimize kidney-related health disparities.
Dimension II: Population – Does using different indicators of kidney
functioning change how we view Population-level age, sex and
racial/ethnic disparities in kidney functioning?
It becomes even more complicated to understand the population-level demographic
disparities in kidney functioning when we examine a variety of indictors. There has been an
increasing body of literature that suggests that there are conflicting results in CKD prevalence
across studies when different kidney measures are being used (52, 62, 63). For instance, while
some population-level and epidemiologic studies used serum levels or at-risk cutoff thresholds of

8
Serum Creatinine (Scr) and Cystatin C (CysC) (64-66), clinical studies use glomerular filtration
rate (GFR) to identify CKD prevalence, which are essentially formulas that convert Scr and
CysC to estimate measured GFR with adjustment of age, gender, and sometimes race (67). There
is some evidence suggesting discrepancies in age, sex, and racial/ethnic differentials in kidney
functioning across various indicators (53, 64, 66, 68). These conflicting results have important
implications for understanding sociodemographic disparities in kidney functioning, but the
evidence of differentials has not been systematically examined across the commonly used
indicators of kidney functioning. Thus, the second aim of this dissertation is to determine and
compare racial/ethnic and educational disparities in kidney functioning among older Americans
using various kidney measures.
Dimension III: Cross-national – How and why do the socioeconomic
status disparities in kidney functioning differ between China and the
U.S.?
Lastly, cross-national studies in kidney functioning remain scarce. While one’s
socioeconomic status (SES) is a very important indicator of kidney functioning in both
developed countries like United States and developing countries like China, CKD is a
complicated disease and is associated with different risk factors in various epidemiological
contexts, resulting in differences in the mechanisms behind the associations with SES and kidney
functioning, thus requiring a variety of approaches for prevention and treatment (69). For
instance, as many as 1 in 3 people with diabetes and 1 in 5 with hypertension in developed
countries have CKD, which has led to the suggestion that focusing on the control of diabetes and
cardiovascular disease will alleviate the growing burden of CKD (70). However, CKD has

9
diverse causes and risk factors beyond comorbidities such as infections and toxins, especially in
developing countries, which account for two-thirds of the global burden of kidney disease.
Further understanding of the effect of SES and the related risk factors on kidney progression
requires both longitudinal and cross-national data between developed vs developing countries
that allow us to observe how SES/risk factors contribute to a more rapid trajectory of kidney
decline and which are modifiable, particularly at older ages across different epidemiological and
social context. Thus, in this chapter, we compare SES disparity in CKD among middle age and
older persons in the U.S. with that in China, aiming to examine the consistency of associations
between SES and CKD and the underlying mechanisms across two countries.
CKD is a global public health concern that disproportionally affect those who are socially
marginalized, leading to poor quality of life, financial burden and higher likelihood of mortality.
However, while there has been major progress in this line of research, awareness and the overall
understanding of CKD remain low especially when compared to other major non-communicable
diseases. This dissertation aims to strengthen our understanding in the sociodemographic patterns
in CKD by implementing novel biomarkers into population-level studies. Findings from these
three studies strengthen our understanding of the direction, extent, and causes of the
sociodemographic disparities in kidney diseases among older adults. providing potential policy
implications and interventions that help address those disparities on a national scale and
facilitating the adaptation and adoption of the kidney disease treatment and prevention guidelines
for the socially disadvantaged populations.

10
1.4 References
1. Weinstein JR, Anderson S. The aging kidney: physiological changes. Adv Chronic
Kidney Dis. 2010;17(4):302–307.
2. Romagnani, P., Remuzzi, G., Glassock, R., Levin, A., Jager, K. J., Tonelli, M., ... &
Anders, H. J. (2017). Chronic kidney disease. Nature reviews Disease primers, 3(1), 1-24.
3. Kashani, K., Rosner, M. H., & Ostermann, M. (2020). Creatinine: from physiology to
clinical application. European journal of internal medicine, 72, 9-14.
4. Onopiuk, A., Tokarzewicz, A., & Gorodkiewicz, E. (2015). Cystatin C: a kidney function
biomarker. Advances in clinical chemistry, 68, 57-69.
5. Saran, R., Robinson, B., Abbott, K. C., Bragg-Gresham, J., Chen, X., Gipson, D., ... &
Shahinian, V. (2020). US renal data system 2019 annual data report: epidemiology of
kidney disease in the United States. American journal of kidney diseases: the official
journal of the National Kidney Foundation, 75(1 Suppl 1), A6-A7.
6. Centers for Disease Control and Prevention. Chronic Kidney Disease in the United
States, 2021. Atlanta, GA: US Department of Health and Human Services, Centers for
Disease Control and Prevention; 2021.
7. United States Renal Data System. (2018). 2018 USRDS annual data report: epidemiology
of kidney disease in the United States.
8. James, S. L., Abate, D., Abate, K. H., Abay, S. M., Abbafati, C., Abbasi, N., ... & Briggs,
A. M. (2018). Global, regional, and national incidence, prevalence, and years lived with
disability for 354 diseases and injuries for 195 countries and territories, 1990–2017: a
systematic analysis for the Global Burden of Disease Study 2017. The
Lancet, 392(10159), 1789-1858.

11
9. Bikbov, B., Purcell, C. A., Levey, A. S., Smith, M., Abdoli, A., Abebe, M., ... &
Owolabi, M. O. (2020). Global, regional, and national burden of chronic kidney disease,
1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. The
lancet, 395(10225), 709-733.
10. Jager, K. J., Kovesdy, C., Langham, R., Rosenberg, M., Jha, V., & Zoccali, C. (2019). A
single number for advocacy and communication—worldwide more than 850 million
individuals have kidney diseases. Nephrology Dialysis Transplantation, 34(11), 1803-
1805.
11. Francis, A., Harhay, M. N., Ong, A., Tummalapalli, S. L., Ortiz, A., Fogo, A. B., ... &
Jha, V. (2024). Chronic kidney disease and the global public health agenda: an
international consensus. Nature Reviews Nephrology, 1-13.
12. Ritchie, H., & Roser, M. (2023). Does the news reflect what we die from?. Our world in
data.
13. Chu, C. D., Chen, M. H., McCulloch, C. E., Powe, N. R., Estrella, M. M., Shlipak, M. G.,
& Tuot, D. S. (2021). Patient Awareness of CKD: A Systematic Review and Metaanalysis of Patient-Oriented Questions and Study Setting. Kidney medicine, 3(4), 576–
585.e1.
14. Ene-Iordache, B., Perico, N., Bikbov, B., Carminati, S., Remuzzi, A., Perna, A., ... &
Remuzzi, G. (2016). Chronic kidney disease and cardiovascular risk in six regions of the
world (ISN-KDDC): a cross-sectional study. The Lancet Global Health, 4(5), e307-e319.
15. Evans, M., Lewis, R. D., Morgan, A. R., Whyte, M. B., Hanif, W., Bain, S. C., ... &
Strain, W. D. (2022). A narrative review of chronic kidney disease in clinical practice:
current challenges and future perspectives. Advances in therapy, 39(1), 33-43.

12
16. Gummidi, B., John, O., Ghosh, A., Modi, G. K., Sehgal, M., Kalra, O. P., ... & Jha, V.
(2020). A systematic study of the prevalence and risk factors of CKD in Uddanam,
India. Kidney International Reports, 5(12), 2246-2255.
17. Perico, N., & Remuzzi, G. (2016). Prevention programs for chronic kidney disease in
low-income countries. Internal and emergency medicine, 11, 385-389.
18. Foreman, K. J., Marquez, N., Dolgert, A., Fukutaki, K., Fullman, N., McGaughey, M., ...
& Murray, C. J. (2018). Forecasting life expectancy, years of life lost, and all-cause and
cause-specific mortality for 250 causes of death: reference and alternative scenarios for
2016–40 for 195 countries and territories. The Lancet, 392(10159), 2052-2090.
19. Roth, G. A., Abate, D., Abate, K. H., Abay, S. M., Abbafati, C., Abbasi, N., ... &
Borschmann, R. (2018). Global, regional, and national age-sex-specific mortality for 282
causes of death in 195 countries and territories, 1980–2017: a systematic analysis for the
Global Burden of Disease Study 2017. The lancet, 392(10159), 1736-1788.
20. Wang, H., Naghavi, M., Allen, C., Barber, R. M., Bhutta, Z. A., Carter, A., ... & Bell, M.
L. (2016). Global, regional, and national life expectancy, all-cause mortality, and causespecific mortality for 249 causes of death, 1980–2015: a systematic analysis for the
Global Burden of Disease Study 2015. The lancet, 388(10053), 1459-1544.
21. Kalantar-Zadeh, K., Lockwood, M. B., Rhee, C. M., Tantisattamo, E., Andreoli, S.,
Balducci, A., ... & Li, P. K. T. (2022). Patient-centred approaches for the management of
unpleasant symptoms in kidney disease. Nature Reviews Nephrology, 18(3), 185-198.
22. Fletcher, B. R., Damery, S., Aiyegbusi, O. L., Anderson, N., Calvert, M., Cockwell, P., ...
& Kyte, D. (2022). Symptom burden and health-related quality of life in chronic kidney
disease: A global systematic review and meta-analysis. PLoS medicine, 19(4), e1003954.

13
23. de Rooij, E. N., Meuleman, Y., de Fijter, J. W., Le Cessie, S., Jager, K. J., Chesnaye, N.
C., ... & Dekker, F. W. (2022). Quality of life before and after the start of dialysis in older
patients. Clinical Journal of the American Society of Nephrology, 17(8), 1159-1167.
24. de Rooij, E. N., Meuleman, Y., de Fijter, J. W., Jager, K. J., Chesnaye, N. C., Evans,
M., ... & Dekker, F. W. (2022). Symptom burden before and after dialysis initiation in
older patients. Clinical Journal of the American Society of Nephrology, 17(12), 1719-
1729.
25. Shukri, M., Mustofai, M. A., Md Yasin, M. A. S., & Tuan Hadi, T. S. (2020). Burden,
quality of life, anxiety, and depressive symptoms among caregivers of hemodialysis
patients: The role of social support. The International Journal of Psychiatry in
Medicine, 55(6), 397-407.
26. Adejumo, O. A., Iyawe, I. O., Akinbodewa, A. A., Abolarin, O. S., & Alli, E. O. (2019).
Burden, psychological well-being and quality of life of caregivers of end stage renal
disease patients. Ghana Medical Journal, 53(3), 190-196.
27. Jha, V., Garcia-Garcia, G., Iseki, K., Li, Z., Naicker, S., Plattner, B., ... & Yang, C. W.
(2013). Chronic kidney disease: global dimension and perspectives. The
Lancet, 382(9888), 260-272.
28. Afkarian, M., Zelnick, L. R., Hall, Y. N., Heagerty, P. J., Tuttle, K., Weiss, N. S., & De
Boer, I. H. (2016). Clinical manifestations of kidney disease among US adults with
diabetes, 1988-2014. Jama, 316(6), 602-610.
29. Garcia-Garcia, G., & Jha, V. (2015). Chronic kidney disease in disadvantaged
populations. Indian journal of nephrology, 25(2), 65–69.

14
30. Stanifer, J. W., Muiru, A., Jafar, T. H., & Patel, U. D. (2016). Chronic kidney disease in
low-and middle-income countries. Nephrology Dialysis Transplantation, 31(6), 868-874.
31. Agyei-Mensah, S., & de-Graft Aikins, A. (2010). Epidemiological transition and the
double burden of disease in Accra, Ghana. Journal of urban health, 87, 879-897.
32. Tomlinson, L. A., & Clase, C. M. (2019). Sex and the Incidence and Prevalence of
Kidney Disease. Clinical journal of the American Society of Nephrology :
CJASN, 14(11), 1557–1559.
33. Carrero, J. J., Hecking, M., Chesnaye, N. C., & Jager, K. J. (2018). Sex and gender
disparities in the epidemiology and outcomes of chronic kidney disease. Nature Reviews
Nephrology, 14(3), 151-164.
34. Norris, K., & Nissenson, A. R. (2008). Race, gender, and socioeconomic disparities in
CKD in the United States. Journal of the American Society of Nephrology, 19(7), 1261-
1270.
35. Norris, K. C., & Nissenson, A. R. (2020). Ethnicity and Chronic Kidney Disease—United
States. In Chronic Renal Disease (pp. 111-119). Academic Press.
36. Nicholas, S. B., Kalantar-Zadeh, K., & Norris, K. C. (2015). Socioeconomic disparities in
chronic kidney disease. Advances in chronic kidney disease, 22(1), 6–15.
37. Zeng, X., Liu, J., Tao, S., Hong, H. G., Li, Y., & Fu, P. (2018). Associations between
socioeconomic status and chronic kidney disease: a meta-analysis. J Epidemiol
Community Health, 72(4), 270-279.
38. Lewandowski, M. J., Krenn, S., Kurnikowski, A., Bretschneider, P., Sattler, M.,
Schwaiger, E., Antlanger, M., Gauckler, P., Pirklbauer, M., Brunner, M., Horn, S., Zitt,
E., Kirsch, B., Windpessl, M., Wallner, M., Aringer, I., Wiesholzer, M., Hecking, M., &

15
Hödlmoser, S. (2023). Chronic kidney disease is more prevalent among women but more
men than women are under nephrological care: Analysis from six outpatient clinics in
Austria 2019. Wiener klinische Wochenschrift, 135(3-4), 89–96.
39. Chesnaye, N. C., Dekker, F. W., Evans, M., Caskey, F. J., Torino, C., Postorino, M.,
Szymczak, M., Ramspek, C. L., Drechsler, C., Wanner, C., & Jager, K. J. (2021). Renal
function decline in older men and women with advanced chronic kidney disease-results
from the EQUAL study. Nephrology, dialysis, transplantation : official publication of the
European Dialysis and Transplant Association - European Renal Association, 36(9),
1656–1663.
40. Hecking, M., Hödlmoser, S., Ahmed, S. B., & Carrero, J. J. (2022, March). The other
way around: living with chronic kidney disease from the perspective of men. In Seminars
in Nephrology (Vol. 42, No. 2, pp. 122-128). WB Saunders.
41. Mayne, K. J., Sullivan, M. K., & Lees, J. S. (2023). Sex and gender differences in the
management of chronic kidney disease and hypertension. Journal of human
hypertension, 37(8), 649–653. https://doi-org.libproxy2.usc.edu/10.1038/s41371-023-
00843-9
42. Martins, D., Agodoa, L., & Norris, K. (2012). Kidney disease in disadvantaged
populations. International journal of nephrology, 2012, 469265.
43. Norton, J. M., Moxey-Mims, M. M., Eggers, P. W., Narva, A. S., Star, R. A., Kimmel, P.
L., & Rodgers, G. P. (2016). Social Determinants of Racial Disparities in CKD. Journal
of the American Society of Nephrology : JASN, 27(9), 2576–2595.

16
44. Norris, K., Mehrotra, R., & Nissenson, A. R. (2008). Racial differences in mortality and
ESRD. American journal of kidney diseases : the official journal of the National Kidney
Foundation, 52(2), 205–208. https://doi.org/10.1053/j.ajkd.2008.06.004
45. Krieger, N., Williams, D. R., & Moss, N. E. (1997). Measuring social class in US public
health research: concepts, methodologies, and guidelines. Annual review of public
health, 18, 341–378. https://doi.org/10.1146/annurev.publhealth.18.1.341
46. Pampel, F. C., Krueger, P. M., & Denney, J. T. (2010). Socioeconomic Disparities in
Health Behaviors. Annual review of sociology, 36, 349–370.
47. Vart, P., Gansevoort, R. T., Crews, D. C., Reijneveld, S. A., & Bültmann, U. (2015).
Mediators of the association between low socioeconomic status and chronic kidney
disease in the United States. American journal of epidemiology, 181(6), 385-396.
48. White, S. L., McGeechan, K., Jones, M., Cass, A., Chadban, S. J., Polkinghorne, K. R., ...
& Roderick, P. J. (2008). Socioeconomic disadvantage and kidney disease in the United
States, Australia, and Thailand. American journal of public health, 98(7), 1306-1313.
49. Tannor, E. K., Chika, O. U., & Okpechi, I. G. (2022). The impact of low socioeconomic
status on progression of chronic kidney disease in low-and lower middle-income
countries. In Seminars in Nephrology (Vol. 42, No. 5, p. 151338). WB Saunders.
50. Luyckx, V. A., Tonelli, M., & Stanifer, J. W. (2018). The global burden of kidney disease
and the sustainable development goals. Bulletin of the World Health Organization, 96(6),
414–422D.
51. Bolignano, D., Mattace-Raso, F., Sijbrands, E. J., & Zoccali, C. (2014). The aging kidney
revisited: a systematic review. Ageing research reviews, 14, 65–80.
https://doi.org/10.1016/j.arr.2014.02.003

17
52. Tidman M, Sjöström P, Jones I. A Comparison of GFR estimating formulae based upon
s-cystatin C and s-creatinine and a combination of the two. Nephrol Dial Transplant.
2008;23(1):154-160.
53. Legrand H, Werner K, Christensson A, Pihlsgård M, Elmståhl S. Prevalence and
determinants of differences in cystatin C and creatinine-based estimated glomerular
filtration rate in community-dwelling older adults: a cross-sectional study. BMC
Nephrology. 2017;18(1).
54. Crimmins E, Vasunilashorn S, Kim JK, Alley D. Chapter 5 Biomarkers Related To
Aging In Human Populations. Advances in Clinical Chemistry. Published online
2008:161-216.
55. Crimmins, E. M., & Seeman, T. E. (2004). Integrating biology into the study of health
disparities. Population and Development Review, 30, 89-107.
56. Glei, D. A., Goldman, N., Rodríguez, G., & Weinstein, M. (2014). Beyond Self-Reports:
Changes in Biomarkers as Predictors of Mortality. Population and development
review, 40(2), 331–360.
57. Crimmins, E. M., Alley, D., Reynolds, S. L., Johnston, M., Karlamangla, A., & Seeman,
T. (2005). Changes in biological markers of health: Older Americans in the 1990s. The
Journals of Gerontology Series A: Biological Sciences and Medical Sciences, 60(11),
1409-1413.
58. Whaley-Connell, A., Nistala, R., & Chaudhary, K. (2011). The importance of early
identification of chronic kidney disease. Missouri medicine, 108(1), 25–28.
59. Norris, K. C., & Agodoa, L. Y. (2005). Unraveling the racial disparities associated with
kidney disease. Kidney international, 68(3), 914–924.

18
60. Derose, S. F., Rutkowski, M. P., Crooks, P. W., Shi, J. M., Wang, J. Q., Kalantar-Zadeh,
K., ... & Jacobsen, S. J. (2013). Racial differences in estimated GFR decline, ESRD, and
mortality in an integrated health system. American journal of kidney diseases, 62(2), 236-
244.
61. Stengel, B., Tarver-Carr, M. E., Powe, N. R., Eberhardt, M. S., & Brancati, F. L. (2003).
Lifestyle factors, obesity and the risk of chronic kidney disease. Epidemiology
(Cambridge, Mass.), 14(4), 479–487.
62. Stevens LA, Schmid CH, Greene T, et al. Comparative performance of the CKD
Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease
(MDRD) Study equations for estimating GFR levels above 60 mL/min/1.73 m2. Am J
Kidney Dis. 2010;56(3):486-495.
63. Åkerblom A, Helmersson-Karlqvist J, Flodin M, Larsson A. Comparison between
Cystatin C- and Creatinine-Estimated Glomerular Filtration Rate in Cardiology
Patients. Cardiorenal Med. 2015;5(4):289-296.
64. Jones C, McQuillan G, Kusek J, et al. Serum creatinine levels in the US population:
Third National Health and Nutrition Examination Survey. American Journal of Kidney
Diseases. 1998;32(6):992-999.
65. Nissenson AR, Pereira BJG, Collins AJ, Steinberg EP. Prevalence and characteristics of
individuals with chronic kidney disease in a large health maintenance organization.
American Journal of Kidney Diseases. 2001;37(6):1177-1183.
66. Köttgen A, Selvin E, Stevens LA, Levey AS, Van Lente F, Coresh J. Serum Cystatin C in
the United States: The Third National Health and Nutrition Examination Survey
(NHANES III). American Journal of Kidney Diseases. 2008;51(3):385-394.

19
67. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group (2024).
KDIGO 2024 Clinical Practice Guideline for the Evaluation and Management of Chronic
Kidney Disease. Kidney international, 105(4S), S117–S314.
68. Oliver III, J. D., Nee, R., Marneweck, H., Banaag, A., Koyama, A. K., Pavkov, M. E., &
Koehlmoos, T. P. (2024). Impact of Race-Free Glomerular Filtration Rate Estimations on
CKD Prevalence in the US Military Health System: A Retrospective Cohort Study.
Kidney Medicine, 100861.
69. Kovesdy C. P. (2022). Epidemiology of chronic kidney disease: an update 2022. Kidney
international supplements, 12(1), 7–11.
70. US National Institute of Diabetes and Digestive and Kidney Diseases. Kidney disease
statistics for the United States. https://www-niddk-nih-gov.libproxy2.usc.edu/healthinformation/health-statistics/kidney-disease (2023).

20
Chapter 2: Racial and Educational Disparities in Kidney
Deterioration and Their Associations with Health
Correlates: A Longitudinal Analysis over 8 years
2.1 Introduction
The aging process is accompanied by progressive decline in the functioning of vital
organs. Age-related decline in kidney functioning is recognized as significant for overall health
(1). People with chronic kidney disease (CKD) exhibit a higher risk of mortality, morbidity and
substantial impairment in quality of life (2,3). Decline in kidney functioning also imposes a
substantial socioeconomic burden on the health care system in the United States. The Centers for
Disease Control and Prevention estimated that one in every three older Americans had chronic
kidney disease (CKD) (4), and the overall Medicare expenditures for CKD surpassed 120 billion
dollars per year (5). As the older population in the United States increases, CKD will put an
increasingly severe burden on individuals and society.
Aging itself has long been considered a major risk factor for kidney decline. Previous
studies suggest the prevalence of diagnosed CKD is higher among older persons compared to
their young and middle-aged counterparts and continues to increase at advanced older age (6).
Age differences in kidney function have been explained by the fact that older people have more
comorbidities that are associated with decline in kidney functioning, including diabetes,
hypertension, and cardiovascular diseases (2,7-9). For instance, people with diabetes mellitus
may develop diabetic nephropathy, which is the most common cause of CKD (3,4), and
hypertension is associated with arteriolar nephrosclerosis and impaired kidney function (10).
Decline in kidney functioning is also related to adverse health behaviors including physical
inactivity/obesity and smoking, which are associated with earlier onset and more rapid

21
progression of CKD at older age (11). However, even among healthy people without renal
diseases or comorbidities, age-related decline in renal functioning has been observed (1,10). It is
generally believed that the kidneys undergo decline not only as a consequence of comorbidities
and unhealthy behaviors, but also normal age-related kidney senescence - a term that describes
age-related changes in physiological functioning, such as loss in renal mass, nephrons, renal
blood flow, etc., that occur even in the absence of disease or that cannot be attributed entirely to
disease (12,13). Genetic risk may also play a role in decline in kidney functioning among older
adults. Over 400 genes have been associated with accelerated renal aging (e.g., mortalin-2, IGF
receptor) (14-16).
On the one hand, it is important to determine to what extent age-related decline in kidney
functioning is attributable to modifiable health behaviors and comorbidities that can be avoided,
or delayed, by preventive health care and disease management. On the other hand, it is also
important to understand to what extent decline in kidney functioning is caused by age-related
senescence that may be a reflection of systemic aging and genetic factors, for example, at a
cellular or molecular level. However, to our knowledge, no study has differentiated these two
pathways (age vs. health conditions and behaviors) of decline in kidney functioning using a
nationally representative sample of older adults. Most previous analyses detailing kidney decline
with age are cross-sectional and performed among living kidney donors, a highly selective
population that do not have CKD or kidney-related comorbidities. This is the first study that
reflects the trajectory of kidney function changes for individuals and differentiates age-related
decline from disease-related decline in a representative sample of aging persons.
Both disease risks and biological aging vary across subgroups, which likely contribute to
disparities in decline in kidney functioning. Non-Hispanic Black and Hispanic individuals and

22
those with lower socioeconomic status (SES), for instance, have been shown to age biologically
faster than others (3,4,17,18). Previous data suggested Non-Hispanic Black and Hispanic
individuals, and those with a lower level of education are more likely to have more severe stages
of CKD that require treatment, due to their high levels of comorbidities and adverse behaviors
(3,19-21); there is also evidence for genetic differences that may contribute to the Black-White
difference - for example, level of transforming growth factor-beta (TGF-β) receptor type 2
activity appears to be higher among individuals with African ancestry, which increases rates of
severe stages of CKD (3,11,20,22). However, most existing research examining subgroup
differences uses cross-sectional data and cannot observe or compare trajectories/rates of change
over time between racial/ethnic groups (23). Longitudinal observations on individuals over years
are needed to assess differentials in change with aging and identify the association of
comorbidities and health behaviors with the rate of kidney decline across sociodemographic
groups (7,24,25). Thus, another aim of this paper is to examine differential kidney decline by
race/ethnicity and education among older Americans. This sample includes large numbers of
ethnically diverse respondents, which can provide important information on subgroup
differences that can be used in efforts to reduce kidney-related health disparities.
Using the Health and Retirement Study (HRS), a longitudinal and nationally
representative sample of Americans over the age of 50, we address the following research
questions: 1) How much do age-related health conditions, behaviors and genetic risk explain
decline in kidney functioning with age? and 2) Does the trajectory of decline in kidney
functioning differ by race/ethnicity and by education?

23
2.2 Method
2.2.1 Data
We used data from the HRS, a representative sample of U.S. adults over age 50 who are
surveyed every two years. The sample was selected using a stratified, multistage area probability
design with oversampling of African American and Hispanic individuals. HRS began to collect
blood-based biomarkers using dried blood spots (DBS) on a random half of the sample in 2006,
and on the other half of the sample in 2008; samples from 2006 and 2008 provided the baseline
for our analysis. Additional blood samples were then collected in 2010 and 2014 for the 2006
half-sample, and in 2012 and 2016 for the 2008 half-sample (26-29). We were able to examine
cystatin C at three time points spanning eight years.
There were 12,103 respondents who participated in the baseline interview. We dropped
242 participants that self-identify as non-Hispanic other and this analysis focuses on nonHispanic White, non-Hispanic Black and Hispanic older adults. Participants with missing data on
the following variables were also excluded: 588 on cystatin C; then 490 on hypertension
management; 202 on diabetes; 1 on heart conditions; 5 on BMI; 69 on smoking; 233 on drinking
behavior; 2 on race/ethnicity; and 14 on education. Our final analytic sample size is 10,257.
Excluded respondents are more likely to have higher cystatin C levels (included: 1.07; excluded:
1.16, p<0.000), be older (11.3% are over 80-year-old for included; 22.8% for excluded,
p<0.000), be identified as non-Hispanic Black race (included: 8.8%; excluded: 10.6, p=0.041),
and be less educated (18.2% does not finish high school for included; 22.4% for excluded,
p=0.003).

24
2.2.2 Measure
Kidney Functioning. Kidney function was determined by cystatin C level (mg/L), which
is considered by many researchers an ideal indicator of renal impairment for population-level
study as, in contrast to other measures, such as creatinine, its value is not affected by gender, age,
race/ethnicity, etc. (24,25). When kidney functioning declines, cystatin C becomes more
concentrated, and its level in blood rises.
The HRS cystatin C measure is based on dried blood spot (DBS) assays performed in a
series of labs. We use venous blood equivalent values for cystatin C provided by HRS, rather
than the DBS value itself (30). The equivalent value is constructed by HRS assuming that
distribution of DBS assays is similar to that of venous blood values in the National Health and
Nutrition Examination Survey (NHANES) sample of the same age while preserving the
variability in the HRS sample. Venous blood equivalent values are recommended for analytic use
because they allow comparison of biomarker results over time, across labs, and with other
population studies based on conventional assays. However, all years of the HRS study used the
same NHANES sample (1999-2002) to construct equivalent values (29); this means that there
was no population-level change with time reflected in cystatin C in HRS; however, relative
changes within-persons can be assessed.
Health Conditions. Information on presence of three age-related diseases related to
kidney functioning is based on self-reports of doctor diagnosis at each interview: hypertension,
diabetes, and heart conditions (heart attack, coronary heart disease, angina, congestive heart
failure, etc.). Measured levels of blood pressure and HbA1c are used to determine unreported or
undiagnosed hypertension and diabetes as well as disease that is controlled or not controlled by
drugs. These two conditions are divided into three mutually exclusive categories: 1) those that

25
had no self-report and did not measure high; 2) those with self-report but condition was now
controlled so they did not measure high; and 3) those whose conditions were uncontrolled.
Individuals are considered hypertensive if their average systolic blood pressure is greater
than or equal to 140 mmHg and/or if the average diastolic blood pressure is greater than or equal
to 90 mmHg. Diabetes is evaluated using HbA1c which gives a 2-3 month estimate of glucose
level. Values ≥ 6.5% are used to categorize respondents having controlled or uncontrolled
diabetes and to indicate presence of diabetes among those who do not report the condition. Heart
conditions is a binary variable based on participants’ self-reports of having been told by a doctor
they had one of the heart conditions listed above. Health conditions also include obesity and
overweight derived from the Body mass index (BMI): overweight BMI 25–30 kg/m2 and obese
BMI ≥ 30 kg/m2
. Health condition variables are treated as time-varying variables as information
was collected at each wave.
Health Behaviors. We included smoking (never smoked, former smokers and current
smokers) and problem drinking to indicate participants’ health behaviors. A binary variable,
problem drinking, indicates whether respondents reported indicators of alcohol misuse based on
the CAGE questionnaire, developed by Ewing (30), an accurate screening tool in detecting
problem drinking (31). “CAGE” is an acronym for its four questions – have your ever 1) felt the
need to Cut down your drinking; 2) felt Annoyed by criticism of your drinking; 3) had Guilty
feelings about drinking; and 4) taken a morning Eye opener? People with scores 2 or higher are
considered to have problem drinking and are coded as 1; people scored below 2 are coded as 0
(30, 31).
Genetic Risk for Kidney Disease. HRS constructed CKD polygenic risk scores (PGS) by
aggregating individual loci across the genome from respondents’ salivary DNA GWAS data to

26
produce an indicator of an individual’s inherited genetic risk of kidney disease (32). PGS for
kidney function phenotypes were created based on a 2019 study conducted by the Chronic
Kidney Disease Genetics consortium (32). We also adjusted for six principal components
obtained from the GWAS (PCs). These principal components are a standard method of
identifying and controlling for ancestry in the models to adjust for possible race/ancestry
differences in genetic characteristics. In our case, we use principal components to adjust for the
potential genetic differences between Black and White participants following previous literature
(3,11,20,22). PGS are only available for non-Hispanic White and Black participants in HRS so
Hispanics are not included.
Covariates. Gender was indicated using a binary variable for female. To account for age
differences, we categorized participants’ baseline age into groups: 52-59, 60-69, 70-79 and ≥80.
Race/ethnicity was self-reported and divided into three categories: those identified as nonHispanic White, non-Hispanic Black and Hispanic. Education was used as one indicator of
socioeconomic position: less than high school (<12 years), high school graduates (12 years),
some college or more (>12 years).
2.2.3 Statistical Analysis
We used growth curve models to estimate cystatin C trajectories over 8 years.
Respondents can be observed up to three waves. In our sample, participants were observed at 2.3
waves on average; 53.6% of the sample had data available at all three waves, 24.2% had two
waves of data and 22.2% had just one. We chose the mixed-effect approach to estimate cystatin
C trajectories over 8 years, which allowed time to be featured as a variable in the model and
made interactions with time easy to specify and interpret (33). A two-level linear model was used
- Level-1 represented individual trajectory, specified as:

27
Cystatin Cij = β0j +β1j(Timeij)+ β2j(Health Conditionsij)+ β3j(Health Behaviorsij)+ εij
where β0j is the baseline cystatin C for individual j in 2006/2008, β1j(TIMEij) captures rate of linear
change for person j at time i, and εij is the within-person error term. Health conditions and behaviors
were time-varying and were expressed as β2j(Health Conditionsij) and β3j(Health Behaviorsij).
At Level-2, we modeled a function that estimated between-person differences:
β0j = γ00 +γ01(Time-invariant Covariatesj) + u0j
β1j = γ10 +γ01(Time-invariant Covariatesj) + u1j
where γ00 and γ10 are sample means and between-person differences. u0j and u1j are residual errors
assumed to be uncorrelated with εij.
We added covariates in four sequential models to see how addition of variables changed
the association of age with health indicators. Model 1 started with age and gender. Model 2
included age, gender, and health conditions. Model 3 included age, gender, and health conditions
and behaviors. Model 4 included age, gender, health conditions and behaviors, and PGS.
We then ran a series of models to address race/ethnic differences. In model 5, we
observed whether there were race/ethnic differences in decline of kidney functioning in the total
sample, only controlling for age and gender. Model 6 included race/ethnicity, age, gender, and
education. Model 7 included race/ethnicity, age, gender, education, and health conditions and
behaviors. Model 8 included race/ethnicity, age, gender, education, health conditions and
behaviors, and PGS. To examine the trajectory within race/ethnic groups, models 9 and 10 ran
the full model (model 4), which included age, gender, health conditions and behaviors, and PGS

28
within non-Hispanic White and non-Hispanic Black participants, respectively. Model 11 ran
model 3 for Hispanic participants as they did not have genetic data for polygenic risk scores.
We found no multicollinearity, in correlations among variables in the models, and
variance inflation factors for all models were lower than 3. Baseline sample weights were
applied to all analyses. Statistical analyses were conducted using the xtmixed commands in
STATA version 16.
2.3 Results
Table 2.1 shows means and distributions of cystatin C at each wave and baseline sample
characteristics for 10,257 participants. Mean cystatin C was 1.07 mg/L in 2006/2008 and increased
to 1.15 mg/L in 2010/2012 and 1.18 mg/L in 2014/2016. About a third of the sample (34.3%) were
aged 52-59 at baseline; 33.6% were 60-69; 20.6% were 70-79; and 11.5% were 80 or over. The
sample was 83.7% non-Hispanic White, 9% non-Hispanic Black, and 7.3% Hispanic participants.
Over half (54.4%) of the sample was female. Almost half (48.7%) of the sample had at least some
college education; 33.2% had completed high school; and 18.1% did not graduate from high school.
14.5% were current smokers, 42.3% were former smokers, and 43.2% never smoked. 13.3% had
a CAGE score at least 2, indicating having some degree of problem drinking. More than a third
(35.1%) were considered overweight; 42.4% were considered obese. 32.7% had controlled
hypertension; 33.5% had uncontrolled hypertension; and 33.8% did not have any history of
hypertension. About 10% had controlled diabetes; 12.0% had uncontrolled diabetes; and 78% had
no history of diabetes. About 23.2% experienced some cardiovascular disease.
Table 2.2 presents results of growth curve models which indicate associations of individual
characteristics with both baseline differences as well as change over time. The results show the

29
intercepts, representing the association of variables with the baseline differences in 2006/2008,
and cystatin C change, representing the association of variables with change of cystatin C over 8
years. Model 1 examined age and gender difference in cystatin C baseline levels and also
trajectories. The association of age with a person’s cystatin C level and change was significantly
greater at older ages. For those in the 52-59 age group, the average initial cystatin C level was
0.943 mg/L; the baseline cystatin C level of the 60-69 age group was 1.042 mg/L, which was 0.099
higher than the 52-59 group; the 70-79 group had a baseline level of 1.185 mg/L (0.943+0.242)
and the 80+ group was 1.418 mg/L (0.943+0.475). Cystatin C increased yearly, on average, by
0.020 mg/L every year for those aged 52-59; those aged between 60-69 are 0.006 mg/L faster than
the 52-59 age group and thus had an annual increase of 0.026 mg/L (0.020 + 0.006); those aged
70-79 had annual increase of 0.033 (0.013+0.020) mg/L and those aged 80 or greater had 0.032
(0.012+0.020) mg/L.
Model 2 shows that with control of the health conditions the annual rate of change
attributed to age was somewhat reduced. After the inclusion of diseases and their management and
BMI, annual rate of change in cystatin C decreased by more than half from 0.020 to 0.007 mg/L
for the 52-59 group, from 0.026 to 0.011 mg/L for 60-69, from 0.033 to 0.019 mg/L for 70-79 and
from 0.032 to 0.017 mg/L for 80+. Participants with controlled diabetes had a higher level of
cystatin C at baseline, compared to those with no history of diabetes.
For those with uncontrolled diabetes, cystatin C increased at a significantly higher rate with
age (0.010, p=0.010) compared to people without diabetes. Participants with hypertension, either
controlled or uncontrolled, had higher baseline cystatin C (controlled: 0.102, p<0.000;
uncontrolled: 0.029, p=0.002), but there were no differences in change. People with self-reported
cardiovascular conditions had significantly higher baseline (0.102, p<0.000) and higher rate of

30
change (0.007, p=0.001). People who were obese had higher baseline levels (0.064, p<0.000) and
faster increase (0.006, p=0.006) compared to people with normal weight.
In model 3, after accounting for health behaviors, age-related change in cystatin C
remained significant. The addition of behaviors did not significantly change the associations
between cystatin C and comorbidities and BMI, although coefficients were slightly reduced.
Drinking did not have any association with cystatin C. Current smokers (0.064, p<0.000) had
higher baseline cystatin C levels than those who never smoked.
As the genetic data were not provided for Hispanic participants, we ran a separate model that
included only non-Hispanic White and Black participants (N= 8,509). Model 4 in Table 2.2 shows
that a higher PGS was associated with a higher baseline level of cystatin C (0.021, p<0.000) but it
was not related to rate of change over time.
We now address differences in cystatin C by race/ethnicity. We began with a model that
added race/ethnicity to a model with only age and gender controlled, indicating race differences in
the population before any variables that may be related to race were included (Model 5, Table 2.3).
Results suggest that non-Hispanic Black participants had higher baseline cystatin C than nonHispanic White participants (0.099, p<0.000) and that Hispanic participants had faster increase
than non-Hispanic White participants (0.007, p=0.039). We added education in Model 6 in order
to control for one aspect of the socioeconomic differences that may largely be responsible for
racial/ethnic differences. This shows that the difference in increase between Hispanic and nonHispanic White participants was explained by education; the difference between non-Hispanic
White and Black participants was reduced by 18%. Compared to those who did not complete high
school, high school graduates (-0.048, p=0.004) and those with some college experience (-0.098,

31
p<0.000) had lower baseline cystatin C. People with some college experience had slower increase
of cystatin C compared to those who did not finish high school (-0.009, p=0.004). When health
conditions and behaviors were included in the model with race/ethnicity, the coefficient reflecting
being Black was reduced from 0.082 to 0.063, indicating about one-fourth of the race difference
was due to health conditions and behaviors evaluated in this study (Model 7). The inclusion of the
polygenic risk score did not further change the association of race (0.065, p=0.017) (Model 8).
In Table 2.4, we took advantage of the relatively large samples of each of these race/ethnic
groups to see if the trajectory of change in cystatin C was similar in the three groups by examining
these relationships within each sample separately. However, the samples for non-Hispanic Black
and Hispanic participants were still considerably smaller than that of non-Hispanic White
participants and thus the samples were less powered to detect statistically significant differences.
The statistically significant increase in the rate of change in cystatin C only occurred among the
non-Hispanic White sample, which may be due to the lack of statistical power among the nonHispanic and Hispanic groups; however, the age pattern of increasing differences from the cystatin
C levels of those aged 52-59 was fairly similar in the three groups (Models 9-11). Non-Hispanic
Black women had faster decline (0.014, p=0.022) than non-Hispanic Black men, after controlling
for health conditions/behaviors and PGS. For non-Hispanic White participants, both controlled
(0.102, p<0.000) and uncontrolled hypertension (0.024, p=0.011) were associated with higher
baseline cystatin C, while only controlled hypertension (0.176, p<0.000) was associated with
higher baseline cystatin C for non-Hispanic Black participants, though the coefficients for nonHispanic Black and White participants were quite similar. For non-Hispanic White participants,
both controlled (0.086, p<0.000) and uncontrolled diabetes (0.049, p=0.006) were associated with
higher baseline cystatin C; diabetes management did not seem to be associated with kidney

32
functioning for non-Hispanic Black participants. For non-Hispanic White and Hispanic
participants, on the other hand, those with uncontrolled diabetes had faster decline than those with
no diabetes history. Cardiovascular disease was associated with both baseline (White: 0.101,
p<0.000; Black: 0.062, p<0.000) and change (White: 0.006, p=0.006; Black: 0.016, p=0.034) in
cystatin C for non-Hispanic White and Black participants. Health behaviors, including current
smoking (0.067, p<0.000), being obese (0.082, p<0.000), and polygenic score (0.030, p<0.000)
only had an association with the level of cystatin C for non-Hispanic White participants.
2.4 Discussion
This was the first nationally representative study to 1) quantify the relative importance of
aging versus health conditions and behaviors in changing kidney functioning among the older
population, and also 2) examine the differentials in trajectories of decline across racial/ethnic and
education groups. Our results suggest age remains a significant risk factor for increasing decline
in kidney functioning even after major risk factors were controlled. Although age-related decline
can be affected and accelerated by presence of comorbidities, decline in kidney functioning
appears on average even among healthy individuals. Similar to previous studies, we find
significant baseline level differences that confirm that people with diabetes, hypertension and
cardiovascular diseases tend to have worse kidney functioning as they enter the older ages
(3,20,34). We further suggest people with uncontrolled diabetes or cardiovascular conditions
experience more rapid decline with age compared to those with no history of such conditions.
These results match previous literature: a meta-analysis identified three longitudinal studies that
revealed a significantly adverse association of diabetes with CDK progression to ESRD (10);
similar results were found with myocardial ischemic events leading to accelerated decline in
kidney function (35). Our findings also suggest people with uncontrolled diabetes experience faster

33
decline than those with controlled diabetes, which supports previous studies indicating that
treatment for inadequately controlled diabetes can be effective in reducing CKD cases and renalrelated mortality (36). With a significant proportion of our nationally representative sample having
uncontrolled diabetes, our study suggests optimal management of diabetes and heart conditions is
one approach to slowing decline in kidney functioning in the older population. On the other hand,
hypertension did not appear to have any association with the rate of decline. Nevertheless, the
significantly worse baseline kidney functioning among those with hypertension indicates that
hypertension may have an influence on kidney functioning over the life course even though not
over this 8-year period of older adulthood. Thus, to attenuate the impact of hypertension on kidney
functioning, screening and modification of lifestyle are needed early in the life course. Obesity
was somewhat related to decline in kidney functioning in older age, which implies that physical
activity and modification of diet may be implemented to modulate decline in kidney functioning
(37,38).
The association between age and kidney functioning was not changed much after behaviors
and genetic risk were controlled. Current smokers had worse baseline kidney functioning,
reinforcing existing literature which shows smoking to be a risk factor for CKD (39). Our initial
model including alcohol misuse indicated no association between drinking problems and kidney
functioning; but when education was controlled, people who had drinking problems appeared to
have better baseline kidney functioning but have no association with the rate of change. Our
finding was similar to reports in a previous meta-analysis in which drinking may have a protective
or no effect on CKD (40).
The association between age and baseline level of kidney functioning is not changed much
by the inclusion of known risk factors; on the other hand, the coefficients representing the

34
association of aging with the annual change were reduced substantially with inclusion of health
conditions; however, they remained significant after risk factors were controlled. This points to
the need for future studies to 1) continue examining other socio-cultural-level factors that may be
related to renal health but are unavailable in the current data, such as stress management, nutrition
intake, environmental toxin exposure, poor hygiene, etc. (20), and 2) understand the more nuanced
mechanisms at a genetic level that explain kidney senescence in the absence of disease or adverse
behaviors. As one of the first longitudinal studies that incorporated the CKD PGS, our results
showed that individual genetic differences were associated with baseline but not decline in cystatin
C among older adults. Because the sample size differed when genetic factors were included, we
ran sensitivity analyses (see Table S2.1) repeating all the results with the same sample size
(n=8,509) for the first three models and found that changes in coefficients were the result of the
change in the size and demographics of the sample (smaller sample size and the exclusion of
Hispanic participants). The inclusion of a genetic indicator did not affect our assessment of race
differences in kidney functioning; however, this indicator of genetic risk was only significant
among the non-Hispanic White sample.
This study also observed sociodemographic differentials in the pace of decline in kidney
functioning, and how these differentials were influenced by risk factors. On average, Hispanic
older adults experienced the fastest decline in kidney functioning; though non-Hispanic Black
participants did not have faster decline than non-Hispanic White older adults, their baseline kidney
functioning was so much worse than the other two groups that they had the highest levels
throughout 8-year period. Because members of racial/ethnic minorities are more likely to reach
severe stages of kidney disease, many hypothesize that kidney function declines faster after the
onset of CKD among minority groups (17). However, there has been little longitudinal support for

35
this hypothesis (18,22). To our knowledge, this was the first study that used a nationally
representative sample to demonstrate that kidney function of Hispanic deteriorated faster than nonHispanic White participants, and, instead of having faster decline in kidney functioning at older
ages, non-Hispanic Black participants had worse baseline kidney functioning entering older
adulthood, probably resulting from disadvantages throughout the life course so that their kidney
functioning starts deteriorating at a younger age than non-Hispanic White participants. We also
observed that, within each racial/ethnic group, decline in kidney functioning was associated with
different factors. For non-Hispanic White and Black older adults, more rapid decline seemed to be
linked to presence of cardiovascular diseases, while, for Hispanic older adults, faster decline
seemed to be only related to uncontrolled diabetes. This is in line with the fact that diabetes is the
leading cause accounting for nearly 60% of end stage renal disease among Hispanic individuals
(41).
Lastly, we observed that people with higher education experienced a lower initial level and
slower decline in kidney functioning with age. Relative to people who did not finish high school,
people who had at least some college experienced a slower decline in kidney functioning, which
may be due to the fact that those with higher education have better awareness and management of
comorbidities (42). These findings offered insights into which groups could be the focus of
interventions aimed at reducing disparities in kidney disease progression. First, difference in rates
of decline in kidney functioning between non-Hispanic White and Hispanic participants was
statistically explained by including education, which was, in turn, a relationship “explained by
health conditions.” This indicated that better management of diabetes, heart problems and obesity
may effectively help slow kidney aging among the Hispanic population. Secondly, worse baseline
kidney functioning among non-Hispanic Black participants cannot be explained by the variables

36
included in this analysis indicating health conditions/behaviors or genetic predisposition. This
pointed to the need for future research that explores this racial disparity beyond health conditions
and behaviors – researchers have pointed out that the structural racism experienced by nonHispanic Black participants may be a fundamental reason for the persistent kidney disparities.
Black older adults are more likely to have delayed diagnosis of kidney problems due to systemic
reasons such as more limited access to primary care providers in their communities (43), difficulty
of primary care physicians in communicating CKD diagnosis to minority patients, low referral
rates to specialists (44), and even clinical measures of kidney functioning that have a racial
adjustment that could lead to underestimation of kidney problems among Black patients (45).
Our study has limitations. First, our data only observed change in kidney functioning over
8 years. We did not capture how certain risks affect kidney functioning over a longer time span as
this sample was only representative of older adults. Those risk factors may have an impact on
kidney functioning over the life course and the baseline differences between groups occurred
earlier in life. The significantly worse baseline kidney functioning among those with hypertension,
diabetes, heart disease, obesity or current smoking suggests disease screening and modification of
lifestyle needs to take place early in the adult life course. Second, our sample included participants
who dropped out in later waves due to death and poor health, meaning that we might have lost the
least healthy individuals in the later waves, so that the decline in kidney functioning over time may
have been underestimated in our analysis. Thirdly, studies have shown that cystatin C may be
independently associated with some health conditions such as diabetes and high BMI, in which
presence of diabetes and higher BMI are associated with a higher level of cystatin C even after
kidney functioning as indicated by GFR is controlled (46). Though previous literature has
suggested these independent associations between cystatin C and health conditions are relatively

37
small, future research is needed to understand the non-kidney related determinants of cystatin C in
a representative population to avoid systematic bias in estimating kidney functioning using serum
level of cystatin C. Finally, future studies should continue to examine and control for additional
renal-related social factors (e.g., stress management, nutrition intake, poor housing/environmental
toxin exposure, poor hygiene) and health indicators such as albuminuria/proteinuria, which are
strong predictors for kidney functioning.
While this analysis was conducted on an observational cohort and cannot determine
causality, we found significant associations that may have important clinical implications. First,
prevention of diseases remains one of the most effective ways to maintain kidney health - even in
later life, decline in kidney functioning among those with the presence of comorbidities can be
slowed through better management of diabetes and heart conditions and reduction in obesity.
Second, regarding racial/ethnic disparities, effective control of cardiovascular disease seems to be
particularly beneficial for non-Hispanic Black participants in slowing decline in kidney
functioning, whereas effective control of diabetes may be particularly beneficial for Hispanic
individuals. Decline in kidney functioning among Hispanic individuals may also be slowed by
addressing the education gap, which is, in turn, related to management of diabetes, heart conditions,
and obesity. The substantial disparities in kidney functioning at baseline also highlights the
importance of early screening and modification of lifestyle.

38
Table 2.1. Mean Cystatin C at All Waves and Sample Characteristics at Baseline (N=10,257)
Variables
Cystatin C (mg/L), mean (SD)
2006/2008 1.07 (0.49)
2010/2012 1.15 (0.49)
2014/2016 1.18 (0.52)
n (%)
Age at baseline
52-59 3518 (34.3)
60-69 3444 (33.6)
70-79 2117 (20.6)
≥80 1178 (11.5)
Race/Ethnicity
Non-Hispanic White 8589 (83.7)
Non-Hispanic Black 925 (9.0)
Hispanic 743 (7.3)
Female 5580 (54.4)
Education
Less than high school 1862 (18.1)
High school graduate 3402 (33.2)
Some college or more 4993 (48.7)
Hypertension
No history of hypertension 3472 (33.8)
Controlled 3351 (32.7)
Uncontrolled 3434 (33.5)
Diabetes
No history of diabetes 8000 (78.0)
Controlled 1024 (10.0)
Uncontrolled 1233 (12.0)
Heart conditions 2381 (23.2)
Smoking
Never smoked 4427 (43.2)
Former smoker 4344 (42.3)
Current smoker 1486 (14.5)
BMI (kg/m2
)
Normal 2307 (22.5)
Overweight (25-30) 3603 (35.1)
Obese (≥ 30) 4347 (42.4)
Problem drinking (CAGE≥2) 1367 (13.3)

39
Note: Model 4 controls for six principal components.
Table 2.2. Effects of Gender, Age, Health Conditions, Health Behaviors and Polygenic
Score on Baseline Cystatin C and Annual Cystatin C Change with Eight Years of Age
Model 1 Model 2 Model 3 Model 4
Gender, Age (n=10,257) Model 1 + Health Conditions
(n=10,257)
Model 2 + Health Behaviors
(n=10,257)
Model 3 + PGS (n=8,509)
Intercept P-value Annual
change
P-value Intercept P-value Annual
change
P-value Intercept P-value Annual
growth
P-value Intercept P-value Annual
growth
P-value
Time 0.943 <0.000 0.020 <0.000 0.839 <0.000 0.007 0.001 0.822 <0.000 0.007 0.005 0.007 0.007 0.007 0.007
Female -0.008 0.776 -0.003 0.108 0.004 0.699 0.002 0.351 0.005 0.648 0.002 0.279 0.011 0.268 0.002 0.364
Age group (ref=52-59)
60-69 0.099 <0.000 0.006 0.002 0.076 <0.000 0.004 0.034 0.076 <0.000 0.004 0.045 0.080 <0.000 0.003 0.182
70-79 0.242 <0.000 0.013 <0.000 0.208 <0.000 0.012 <0.000 0.211 <0.000 0.012 <0.000 0.204 <0.000 0.012 <0.000
≥80 0.475 <0.000 0.011 0.002 0.441 <0.000 0.010 0.021 0.447 <0.000 0.010 0.023 0.447 <0.000 0.009 0.032
Hypertension (ref=no history of hypertension)
Controlled hypertension 0.102 <0.000 0.000 0.676 0.102 <0.000 -0.001 0.660 0.110 <0.000 -0.002 0.298
Uncontrolled hypertension 0.029 0.002 0.000 0.885 0.028 0.003 0.000 0.845 0.026 0.006 0.000 0.946
Diabetes (ref=no history of diabetes)
Controlled diabetes 0.112 <0.000 -0.001 0.812 0.113 <0.000 -0.001 0.813 0.091 <0.000 0.002 0.567
Uncontrolled diabetes 0.041 0.005 0.010 0.001 0.042 0.005 0.010 0.001 0.041 0.013 0.008 0.009
Heart conditions 0.102 <0.000 0.007 0.001 0.100 <0.000 0.007 0.001 0.096 <0.000 0.007 <0.000
BMI (ref=normal)
Overweight 0.015 0.158 0.003 0.092 0.017 0.112 0.003 0.096 0.012 0.304 0.004 0.063
Obese 0.064 <0.000 0.006 0.006 0.068 <0.000 0.006 0.008 0.070 <0.000 0.005 0.033
Smoking (ref=never smoked)
Former smoker 0.018 0.097 0.001 0.667 0.018 0.089 0.002 0.229
Current smoker 0.064 <0.000 0.000 0.857 0.066 <0.000 0.001 0.668
Problem drinking (CAGE≥2) -0.024 0.066 0.001 0.645 -0.016 0.243 0.001 0.631
Polygenic score (PGS) for kidney disease 0.021 <0.000 0.000 0.745

40
Table 2.3. Racial/Ethnic and Education Differential in Kidney Functioning Decline and Role
of Health Conditions and Health Behaviors in Explaining Disparities
Model 5 Model 6 Model 7 Model 8
Model 1 + Race/Ethnicity (n =
10,471)
Model 5 + Education (n=10,471) Model 6 + Health Conditions +
Health Behaviors (n=10,471)
Model 7 + PGS (n=8,509)
for NH White and Black
Intercept P-value Annual
change
P-value Intercept P-value Annual
change
P-value Intercept P-value Annual
growth
P-value Intercept P-value Annual
growth
P-value
Time 0.931 <0.000 0.019 <0.000 1.010 <0.000 0.026 <0.000 0.876 <0.000 0.011 0.001 0.086 0.007 0.012 0.002
Female -0.005 0.635 -0.003 0.092 -0.009 0.389 -0.003 0.047 -0.001 0.892 0.001 0.417 0.006 0.553 0.001 0.527
Age group (ref=52-59)
60-69 0.100 <0.000 0.006 0.002 0.091 <0.000 0.005 0.006 0.072 <0.000 0.003 0.066 0.077 <0.000 0.002 0.289
70-79 0.246 <0.000 0.014 <0.000 0.228 <0.000 0.012 <0.000 0.202 <0.000 0.011 <0.000 0.197 <0.000 0.011 <0.000
≥80 0.479 <0.000 0.012 0.001 0.457 <0.000 0.010 0.006 0.435 <0.000 0.009 0.039 0.438 <0.000 0.008 0.067
Race/Ethnicity (ref=non-Hispanic White)
NonHispanic
Black
0.099 <0.000 0.002 0.559 0.082 0.002 0.000 0.959 0.063 0.012 -0.003 0.277 0.065 0.017 -0.003 0.332
Hispanic 0.031 0.251 0.007 0.039 -0.007 0.787 0.004 0.246 -0.002 0.934 0.003 0.437
Education (ref=less than high school)
High school graduate -0.048 0.004 -0.002 0.454 -0.030 0.068 -0.002 0.497 -0.022 0.203 -0.001 0.730
Some college or more -0.098 <0.000 -0.009 0.004 -0.065 <0.000 -0.004 0.087 -0.051 0.002 -0.005 0.080
Hypertension (ref=no history of hypertension)
Controlled hypertension 0.097 <0.000 -0.001 0.648 0.106 <0.000 -0.002 0.289
Uncontrolled hypertension 0.022 0.015 0.001 0.814 0.021 0.029 0.000 0.984
Diabetes (ref=no history of diabetes)
Controlled diabetes 0.107 <0.000 -0.001 0.740 0.087 <0.000 0.002 0.584
Uncontrolled diabetes 0.034 0.021 0.010 0.001 0.033 0.046 0.008 0.007
Heart conditions 0.099 <0.000 0.007 0.001 0.095 <0.000 0.007 0.001
BMI (ref=normal)
Overweight 0.016 0.124 0.003 0.149 0.012 0.306 0.003 0.095
Obese 0.064 <0.000 0.006 0.012 0.067 <0.000 0.005 0.049
Smoking (ref=never smoked)
Former smoker 0.016 0.138 0.001 0.735 0.016 0.133 0.002 0.267
Current smoker 0.054 <0.000 0.000 0.988 0.058 <0.000 0.000 0.890
Problem drinking (CAGE≥2) -0.029 0.026 0.001 0.738 -0.021 0.131 0.001 0.688
Polygenic score (PGS) for kidney disease 0.021 <0.000 0.000 0.774
Note. Model 8 controls for six principal components.

41
Table 2.4. Associations between Kidney Functioning and Health Conditions and Health
Behaviors and Polygenic Score (PGS) within Each Racial/Ethnic Group
Model 9 Model 10 Model 11
Model 1 + Health Conditions/Behaviors and
PGS among Non-Hispanic White Participants
(n=7,334)
Model 1 + Health
Conditions/Behaviors and PGS
among Non-Hispanic Black
Participants (n=1,175)
Model 1 + Health Conditions/Behaviors
among Hispanic Participants
(n=924)
Intercept P-value Annual
growth
P-value Intercept P-value Annual
growth
P-value Intercept P-value Annual
growth
P-value
Time 0.809 <0.000 0.009 0.001 0.910 <0.000 -0.014 0.141 0.928 <0.000 0.003 0.746
Female 0.014 0.142 0.000 0.935 -0.013 0.802 0.014 0.022 -0.071 0.321 0.007 0.252
Age group (ref=52-59)
60-69 0.070 <0.000 0.003 0.206 0.175 0.003 0.006 0.398 0.028 0.640 0.008 0.284
70-79 0.200 <0.000 0.012 <0.000 0.254 <0.000 0.019 0.018 0.342 <0.000 0.008 0.347
≥80 0.442 <0.000 0.008 0.062 0.532 <0.000 0.017 0.200 0.454 <0.000 0.017 0.272
Hypertension (ref=no history of hypertension)
Controlled
hypertension
0.102 <0.000 -0.002 0.397 0.176 <0.000 0.000 0.964 0.053 0.121 0.006 0.377
Uncontrolled
hypertension
0.024 0.011 0.000 0.884 0.030 0.399 0.010 0.206 0.036 0.342 -0.001 0.932
Diabetes (ref=no history of diabetes)
Controlled
diabetes
0.086 <0.000 0.004 0.340 0.102 0.108 -0.001 0.904 0.179 0.048 -0.014 0.216
Uncontrolled
diabetes
0.049 0.006 0.007 0.040 -0.005 0.916 0.015 0.099 0.030 0.490 0.030 0.004
Cardiovascular
disease
0.101 <0.000 0.006 0.006 0.062 0.150 0.016 0.034 0.067 0.405 0.020 0.171
BMI
(ref=normal)
Overweight 0.019 0.057 0.004 0.033 -0.060 0.352 -0.010 0.218 0.052 0.174 -0.004 0.614
Obese 0.082 <0.000 0.004 0.121 -0.032 0.634 0.009 0.239 0.028 0.543 0.009 0.256
Smoking (ref=never smoked)
Former smoker 0.015 0.157 0.002 0.293 0.035 0.523 0.008 0.247 -0.029 0.695 -0.003 0.694
Current smoker 0.067 <0.000 0.001 0.801 0.083 0.147 0.009 0.337 0.013 0.881 -0.001 0.949
Problem
drinking
(CAGE≥2)
-0.013 0.339 0.001 0.652 -0.060 0.307 0.002 0.845 -0.068 0.224 -0.003 0.709
Polygenic score
(PGS) for
kidney disease
0.030 <0.000 0.000 0.563 -0.012 0.928 0.001 0.583
Note. Models 9 and 10 control for six principal components.

42
2.5 References
1. Weinstein JR, Anderson S. The aging kidney: Physiological changes. Adv Chronic Kidney
Dis. 2010;17(4):302-307. doi:10.1053/j.ackd.2010.05.002
2. Martins D, Agodoa L, Norris K. Chronic kidney disease in disadvantaged populations. Int J
Nephrol. 2012;2012:1-6. doi:https://doi.org/10.1155/2012/469265
3. Yan Q, Zuo P, Cheng L, et al. Acute kidney injury is associated with in-hospital mortality in
older patients with COVID-19. J Gerontol A Biol Sci Med Sci. 2020;76(3):456-462. doi:
10.1093/gerona/glaa181
4. Centers for Disease Control and Prevention. Chronic Kidney Disease in the United States,
2021. Atlanta, GA: US Department of Health and Human Services, Centers for Disease
Control and Prevention; 2021. https://www.cdc.gov/kidneydisease/pdf/Chronic-KidneyDisease-in-the-US-2021-h.pdf
5. Saran R, Robinson B, Abbott KC, et al. US Renal Data System 2019 Annual Data Report:
Epidemiology of kidney disease in the United States. Am J Kidney Dis. 2020;75(1):A6-A7.
doi:10.1053/j.ajkd.2019.09.003
6. Shlipak MG, Katz R, Kestenbaum B, et al. Rate of kidney function decline in older adults: a
comparison using creatinine and cystatin C. Am J Nephrol. 2009;30(3):171-178.
doi:10.1159/000212381
7. Bolignano D, Mattace-Raso F, Sijbrands EJ, Zoccali C. The aging kidney revisited: a
systematic review. Ageing Res Rev. 2014;14:65-80. doi:10.1016/j.arr.2014.02.003
8. Silva FG. The aging kidney: a review--part II. Int Urol Nephrol. 2005;37(2):419-432.
doi:10.1007/s11255-004-0874-5
9. Choudhury D, Levi M. Kidney aging--inevitable or preventable? Nat Rev Nephrol.
2011;7(12):706-717. doi:10.1038/nrneph.2011.104

43
10. Tsai WC, Wu HY, Peng YS, et al. Risk factors for development and progression of chronic
kidney disease: A systematic review and exploratory meta-analysis. Medicine (Baltimore).
2016;95(11):e3013. doi:10.1097/MD.0000000000003013
11. Stengel B, Tarver-Carr ME, Powe NR, Eberhardt MS, Brancati FL. Lifestyle factors, obesity
and the risk of chronic kidney disease. Epidemiology. 2003;14(4):479-487.
doi:10.1097/01.EDE.0000071413.55296.c4
12. Kremers WK, Denic A, Lieske JC, et al. Distinguishing age-related from disease-related
glomerulosclerosis on kidney biopsy: the Aging Kidney Anatomy study. Nephrol Dial
Transplant. 2015;30(12):2034-2039. doi:10.1093/ndt/gfv072
13. Denic A, Glassock RJ, Rule AD. Structural and functional changes with the aging
kidney. Adv Chronic Kidney Dis. 2016;23(1):19-28. doi: 10.1053/j.ackd.2015.08.004
14. Wuttke M, Li Y, Li M, et al. A catalog of genetic loci associated with kidney function from
analyses of a million individuals. Nat Genet. 2019;51(6):957-972. doi:10.1038/s41588-019-
0407-x
15. Dietrich LG, Barceló C, Thorball CW, et al. Contribution of genetic background and data
collection on adverse events of anti-human immunodeficiency virus (HIV) drugs (D:A:D)
clinical risk score to chronic kidney disease in Swiss HIV-infected persons with normal
baseline estimated glomerular filtration rate. Clin Infect Dis. 2020;70(5):890-897.
doi:10.1093/cid/ciz280
16. Cañadas-Garre M, Anderson K, Cappa R, et al. Genetic susceptibility to chronic kidney
disease - some more pieces for the heritability puzzle. Front Genet. 2019;10:453.
doi:10.3389/fgene.2019.00453

44
17. Parsa A, Kao WHL, Xie D, et al. APOL1 risk variants, race, and progression of chronic
kidney disease. N Engl J Med. 2013;369(23):2183-2196. doi:10.1056/nejmoa1310345
18. Peralta CA, Vittinghoff E, Bansal N, et al. Trajectories of kidney function decline in young
black and white adults with preserved GFR: Results from the Coronary Artery Risk
Development in Young Adults (CARDIA) study. Am J Kidney Dis. 2013;62(2):261-266.
doi:10.1053/j.ajkd.2013.01.012
19. Westbrook A, Zhang R, Shi M, et al. Association between baseline buccal telomere length
and progression of kidney function: The Health and Retirement Study. J Gerontol A Biol Sci
Med Sci. 2022;77(3):471-476. doi:10.1093/gerona/glac004
20. Norris KC, Agodoa LY. Unraveling the racial disparities associated with kidney
disease. Kidney Int. 2005;68(3):914-924. doi:10.1111/j.1523-1755.2005.00485.x
21. Cobb RJ, Thorpe RJ, Norris KC. Everyday discrimination and kidney function among
older adults: Evidence from the Health and Retirement Study. J Gerontol A Biol Sci Med Sci.
2020;75(3):517-521. doi:10.1093/gerona/glz294
22. Friedman DJ, Pollak MR. APOL1 and kidney disease: From genetics to biology. Annu Rev
Physiol. 2020;82(1):323-342. doi:10.1146/annurev-physiol-021119-034345
23. Barbour SJ, Schachter M, Er L, Djurdjev O, Levin A. A systematic review of ethnic
differences in the rate of renal progression in CKD patients. Nephrol Dial Transplant.
2010;25(8):2422-2430. doi:10.1093/ndt/gfq283
24. Derose SF, Rutkowski MP, Crooks PW, et al. Racial differences in estimated GFR decline,
ESRD, and mortality in an integrated health system. Am J Kidney Dis. 2013;62(2):236-244.
doi:10.1053/j.ajkd.2013.01.019
25. Dharnidharka VR, Kwon C, Stevens G. Serum cystatin C is superior to serum creatinine as a

45
marker of kidney function: a meta-analysis. Am J Kidney Dis. 2002;40(2):221-226.
doi:10.1053/ajkd.2002.34487
26. Crimmins E, Kim J, Weir D., et al. HRS Documentation Report Documentation of
Biomarkers in the 2006 and 2008 Health and Retirement Study Report. 2013. Accessed
February 23, 2023.
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker2006and2008.pdf
27. Crimmins E, Faul J, Kim J, Weir D. HRS Documentation Report Documentation of
Biomarkers in the 2010 and 2012 Health and Retirement Study Report. 2015. Accessed
February 23, 2023.
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker2010and2012.pdf
28. Crimmins E, Faul J, Kim J, Weir D. Documentation of Blood-Based Biomarkers in the 2014
Health and Retirement Study.; 2017. Accessed February 23, 2023.
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker%202014_Dec2017.pdf
29. Crimmins E, Faul J, Kim J, Weir D. Documentation of DBS Blood-Based Biomarkers in the
2016 Health and Retirement Study.; 2020. Accessed February 23, 2023.
https://hrsdata.isr.umich.edu/sites/default/files/documentation/datadescriptions/HRS%20Data%20Documentation%20for%202016%20DBS%20Release.pdf
30. Ewing JA. Detecting alcoholism. The CAGE questionnaire. JAMA. 1984;252(14):1905-1907.
doi:10.1001/jama.252.14.1905
31. Dhalla S, Kopec JA. The CAGE questionnaire for alcohol misuse: a review of reliability and
validity studies. Clin Invest Med. 2007;30(1):33-41. doi:10.25011/cim.v30i1.447

46
32. Ware E, Gard A, Schmitz L. HRS Polygenic Scores – Release 4. 2020. Accessed February
23, 2023. https://hrsdata.isr.umich.edu/sites/default/files/documentation/datadescriptions/PGNSCORES4_3DD_0.pdf
33. McNeish D, Matta T. Differentiating between mixed-effects and latent-curve approaches to
growth modeling. Behav Res Methods. 2017;50(4):1398-1414. doi:10.3758/s13428-017-
0976-5
34. Esmeijer K, Geleijnse JM, de Fijter JW, Giltay EJ, Kromhout D, Hoogeveen EK.
Cardiovascular risk factors accelerate kidney function decline in post-myocardial
infarction patients: The Alpha Omega Cohort Study. Kidney Int Rep. 2018;3(4):879-888. doi:
10.1016/j.ekir.2018.03.005.
35. González KA, Stickel AM, Kaur SS, Ramos AR, González HM, Tarraf W. Serum Cystatin-C
is linked to increased prevalence of diabetes and higher risk of mortality in diverse middleaged and older adults. PLoS One. 2022;17(9):e0270289. doi: 10.1371/journal.pone.0270289
36. Howard K, White S, Salkeld G, et al. Cost-effectiveness of screening and optimal
management for diabetes, hypertension, and chronic kidney disease: a modeled
analysis. Value Health. 2010;13(2):196-208. doi:10.1111/j.1524-4733.2009.00668.x
37. Meyer TE, Kovács SJ, Ehsani AA, Klein S, Holloszy JO, Fontana L. Long-term caloric
restriction ameliorates the decline in diastolic function in humans. J Am Coll Cardiol.
2006;47(2):398-402. doi:10.1016/j.jacc.2005.08.069
38. Chen G, Bridenbaugh EA, Akintola AD, et al. Increased susceptibility of aging kidney to
ischemic injury: identification of candidate genes changed during aging, but corrected by
caloric restriction. Am J Physiol Renal Physiol. 2007;293(4):F1272-F1281.
doi:10.1152/ajprenal.00138.2007

47
39. Bleyer AJ, Shemanski LR, Burke GL, Hansen KJ, Appel RG. Tobacco, hypertension, and
vascular disease: risk factors for renal functional decline in an older population. Kidney Int.
2000;57(5):2072-2079. doi:10.1046/j.1523-1755.2000.00056.x
40. Yuan HC, Yu QT, Bai H, Xu HZ, Gu P, Chen LY. Alcohol intake and the risk of chronic
kidney disease: results from a systematic review and dose-response meta-analysis. Eur J Clin
Nutr. 2021;75(11):1555-1567. doi:10.1038/s41430-021-00873-x
41. Lora C, Daviglus M, Kusek J, et al. Chronic Kidney Disease in United States Hispanics: A
Growing Public Health Problem. Accessed February 23, 2023.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3587111/pdf/nihms-316436.pdf
42. Thio CHL, Vart P, Kieneker LM, Snieder H, Gansevoort RT, Bültmann U. Educational level
and risk of chronic kidney disease: longitudinal data from the PREVEND study. Nephrol
Dial Transplant. 2020;35(7):1211-1218. doi:10.1093/ndt/gfy361
43. Yearby R. Racial disparities in health status and access to healthcare: The continuation of
inequality in the United States due to structural racism. Am J Econ Sociol. 2018;77(3-
4):1113-1152. doi:10.1111/ajes.12230
44. Ephraim PL, Powe NR, Rabb H, et al. The providing resources to enhance African American
patients’ readiness to make decisions about kidney disease (PREPARED) study: protocol of a
randomized controlled trial. BMC Nephrology. 2012;13(1). doi:10.1186/1471-2369-13-135
45. Inker LA, Eneanya ND, Coresh J, et al. New creatinine- and cystatin C-based equations to
estimate GFR without race. N Engl J Med. 2021;385(19):1737-1749.
doi:10.1056/NEJMoa2102953
46. Stevens LA, Schmid CH, Greene T, et al. Factors other than glomerular filtration rate affect
serum cystatin C levels. Kidney Int. 2009;75(6):652-660. doi:10.1038/ki.2008.638

48
Chapter 3: Discrepancies in Age, Sex, and
Racial/Ethnic Patterns of Kidney Functioning Using
Various Indicators of Kidney Functioning
3.1 Introduction
Research studies use a number of measures to indicate kidney functioning. Some use
indicators of cystatin C (CysC), some use serum creatinine (Scr), and some use both (1).
Epidemiologic studies frequently use serum levels or at-risk cutoff thresholds of CysC and Scr
to indicate the prevalence, trends, and disparities of chronic kidney disease (CKD) (2-6). As
decline in kidney functioning is believed to be strongly associated with the aging process, serum
levels of CysC and Scr have also been used as one of the markers for estimating biological aging,
which measures an individual’s physiological status and define whether aging was accelerated or
delayed relative to chronological age (7-9). While the use of Scr is the most common in
population-level studies, CysC is considered a good indicator for population-level studies,
superior to Scr as a marker of kidney functioning in population studies, especially among older
people, because its value is not affected by sex, age, race/ethnicity, protein intake, or muscle
mass (10, 11).
On the other hand, in clinical settings and medical research, Scr and CysC are usually
converted to estimates of the glomerular filtration rate (GFR) to classify the severity of patients’
kidney problems (12, 13). Although GFR is regarded as the most accurate indicator of both
normal and impaired kidney function, measuring GFR (mGFR) directly can be extremely
complex and costly and requires the clearance of exogenous filtration markers (e.g., e.g. inulin
and iohexol), thus making it not commonly used in clinical practice (14). For practical reasons,
the glomerular filtration rate is generally estimated (eGFR) using equations based on Scr and

49
CysC that adjust for age, sex and sometimes race. These equations provide results that are widely
used for detection, diagnosis, and management of people with chronic kidney disease (CKD) (15,
16).
While all of these indicators are used to measure kidney health, different indictors
suggest different levels of kidney functioning problems and CKD in a population (17-20). For
instance, using data from the National Health and Nutrition Examination Survey (NHANES III),
one study found reduced kidney functioning in about 20% of those ages 60 to 69 based on level
of Scr (2), while another study found roughly 30% had reduced kidney functioning based on
serum levels of CysC (21). A third study using the same NHANES III data found more than 60%
had reduced kidney functioning according to a GFR equation based on Scr (22).
GFR estimates based on Scr or CysC can also produce significantly different estimates of
CKD presence and classifications of severity for the same individuals (17-20). Importantly, there
is some evidence suggesting discrepancies in age and sex differentials in kidney functioning
across various indicators. While all indicators suggest older age is associated with worse kidney
functioning on average, the size of the differences across age varies across indicators especially
at the oldest ages. For example, Legrand et al. found a difference of more than 30% in values of
GFRs estimated by different equations for persons over age 80 but only 12% for persons 60-69
(20). The sex disparity in kidney functioning also varies across indicators. When Scr is used men
have more kidney problems than women; but men and women have a similar level of kidney
problems when CysC is used as the indicator (2, 21). Despite some reports of inconsistencies in
the overall and also age- and sex-specific prevalence of and disparities in kidney functioning
across indicators, the evidence of differentials has not been systematically examined across the
commonly used indicators of kidney functioning.

50
Additionally, the race-specific CKD prevalence and the relative Black-White disparity
seem to differ across indictors as well, and it is believed that the inclusion of racial adjustments
in the eGFR equation plays a role in this (23, 24). For example, the CKD Epidemiology
Collaboration (CKD-EPI) 2009 creatinine equation, which has been widely used, included sex,
age, race (Black or non-Black), and serum creatinine level (Scr) in calculations of eGFR,
assigning a 16% higher eGFR for individuals identified as Black with the same age, sex, and Scr
(25, 26). This coefficient was originally included to account for perceived biological differences
between Black and White individuals. For instance, some older studies suggested that muscle
mass may be, on average, higher among Black than White individuals, which may in turn lead to
the phenomenon in which Black individuals “naturally” have a higher level independently of
kidney function - although this association between being Black and muscle mass was poorly
supported and was only noted by three studies with small and nonrepresentative samples that did
not control for any socioeconomic or comorbidity variables (25, 27-29). The most recent Kidney
Disease Improving Global Outcomes (KDIGO) clinical guideline advocates for the use of racefree GFR CKD-EPI 2021 estimating equations (30) because there is now consensus that race
should not be treated as a purely biological factor but also as a social construct (31-31). Research
has shown that, with the new race-free equation, the measured prevalence of Black individuals
with CKD increases and the prevalence of non-Black individuals with CKD decreases
significantly; Black adults are also likely to be reclassified to more advanced stages of CKD and
non-Black adults can be reclassified to less severe stages that could change clinical management
(24). Similarly to age and sex differentials, to our knowledge, no studies have systematically
examined the race-specific prevalence and disparities across all commonly used kidney
indicators.

51
In this study, we use data from the 2016 Health and Retirement Study (HRS) Venous
Blood Sample to compare the prevalence of kidney dysfunction and sex/race disparities in seven
indicators of kidney functioning among adults aged 56 and older. These include serum levels of
CysC, level of Scr, and GFR based on five equations that differ in the basic analyte(s) used and
whether they adjust for age, sex and race. Our study does not aim to indicate how CKD should
be diagnosed but to use indicators of kidney function used in the literature as estimative
measures to observe overall differences in population-level prevalence and disparities - we
examine whether different indicators a) yield consistent results in determining the overall
prevalence of CKD and the stages of kidney disease; b) reveal similar age, sex, and
race/ethnicity disparities in kidney functioning among older Americans; and, finally, 3) as kidney
disease is closely linked to mortality (35, 36), it is important to have a comprehensive assessment
of how the predictive power for mortality varies across kidney indicators and how this differs by
sex and racial groups.
3.2 Methods
3.2.1 Sample
The HRS is a nationally representative study of U.S adults over age 50 who are
interviewed every two years. It oversamples Black and Hispanic individuals to allow analysis of
ethnic differences. The data include weights to make the sample nationally representative. We
analyze an HRS sample of respondents age 56 or older who provided a blood sample in 2016
(37). Persons in nursing homes or interviewed by proxy did not provide a blood sample. Among
the 9,189 age-eligible respondents who provided samples, 7,104 were non-Hispanic White, 930
were non-Hispanic Black, 830 were Hispanic. We excluded 315 who identified as non-Hispanic

52
others and 10 with missing race/ethnicity. We also excluded 70 respondents who had missing Scr
and an additional 10 who had missing CysC values. Our final weighted analytic sample is 8,821.
The 80 respondents with missing Scr or CysC values did not significantly differ from our
analysis sample in age, sex, race/ethnicity, or mortality within 4 years of the interview.
3.2.2 Measures
Kidney functioning. We include seven indicators of kidney functioning: serum level of
cystatin C (CysC), serum level of creatinine (Scr), and five different equation-based estimates of
glomerular filtration rates (eGFR) that were validated using samples of U.S. older adults
according to the KDIGO 2024 Clinical Practice Guideline (27).
CysC is found throughout the body in tissue cells and body fluids and is consistently
produced at a steady rate. Its primary elimination route is via filtration through the glomerulus in
the kidney, making the kidney its exclusive metabolic pathway. When kidney functioning
declines, CysC becomes more concentrated, and its level rises. In the HRS CysC is measured in
serum using CysC reagent (Gentian, Moss Norway) on the Roche COBAS 6000 Chemistry
analyzer at the University of Minnesota. The laboratory inter-assay CV is 4.3% at a
concentration of 0.75 mg/L and 3.2% at a concentration of 3.83 mg/L. The Gentian CysC
Immunoassay (ERM-DA471/IFCC standardized) is an open channel turbidimetric test for
quantitative determination of CysC in human serum and plasma (37).
Scr is a waste product in the bloodstream that comes from normal wear and tear on
muscles of the body. Renal clearance of Scr is influenced by age, sex, and body size; when Scr is
not cleared by the kidney, its level increases (38). For HRS Scr is measured in serum using the
Roche enzymatic method (Roche Diagnostics, Indianapolis, IN 46250), on a Roche COBAS

53
6000 Chemistry Analyzer (Roche Diagnostics Corporation). The laboratory CV is 2.9% at a
concentration of 0.835 mg/dL and 2.8% at a concentration of 3.93 mg/dL.
Glomerular filtration rate (GFR) (mL/min/1.73m2) is the estimated flow rate of fluid
filtered through the kidney. It is widely considered to be the best standard for clinical assessment
of kidney functioning. To be comparable to existing literature, we include all GFR estimating
equations validated using samples of U.S. older adults according to the 2024 Clinical Practice
Guideline, including CKD-EPI 2009 based on Scr (eGFRScr), CKD-EPI 2012 based on CysC
(eGFRcysc), CKD-EPI 2012 based on both (eGFRScr_CysC), CKD-EPI 2021 based on Scr that
does not adjust for race (eGFRScr_w/o race), and CKD-EPI 2021 based on both CysC and Scr
that does not adjust for race (eGFRScr_CysC_w/o race) (30, 39). Our supplemental material
(Table S2.1) further describes these equations. All of those equations adjust for age and sex,
whereas CKD-EPI 2009 based on Scr (eGFRScr) and CKD-EPI 2012 based on both
(eGFRScr_CysC) also adjust for Black race (25, 40).
Impaired kidney functioning. We create binary indicators of impaired kidney functioning
based on cut-off values for each indicator of kidney functioning. While values in the literature
vary, we chose 1 mg/L as the cut-off value for CysC level as prior studies have indicated this
threshold is important for development of cardiovascular disease and all-cause mortality (41, 42,
43), and a Scr level higher than 1.3 mg/L for men and 1.0 mg/L for women (44), and a eGFR
lower than 60 (13).
Stage classification of eGFR. We also create a categorical variable used in clinical
settings to divide people into four stages of severity of kidney disease based on eGFR: stage 1
(eGFR≥90), which indicates normal kidney function and is the reference group in our study,

54
stage 2 (eGFR≥60 & <90), indicating mild impairment, stage 3 (eGFR≥30 & <60), indicating
moderate impairment, stage 4 (eGFR≥15, <30), indicating severe impairment and stage 5
(eGFR<15), which is kidney failure (13).
Demographic variables. Age is treated as a continuous variable. Race/ethnicity has three
categories: non-Hispanic White, non-Hispanic Black, and Hispanic. Sex is a dummy variable
where 1 indicates female. From here on we refer to non-Hispanic White individuals as White
individuals and non-Hispanic Black individuals as Black individuals.
Mortality. We construct an indicator of all-cause mortality over 4 years from the
interview in 2016 to the interview in 2020 using date of death from the HRS tracker file.
3.2.3 Statistical Analysis
First, we compare percentages of people with impaired kidney function using the cutoff
values above. Next, we use logistic regression models to examine age, sex, and race/ethnic
differentials in the likelihood of having impaired kidney function. Additionally, we use
multinomial logistic regression models to examine sex and race/ethnic differentials in CKD
stages using the eGFR stage classification. Finally, to assess the predictive power of each
indicator for mortality, and to examine variations in prediction among different demographic
groups, we analyze sensitivity, specificity, and their corresponding areas under the receiver
operating characteristic (ROC) curve (AUC) for each indicator of kidney functioning, stratified
by sex and race/ethnicity. We use the DeLong test to examine whether the ROC areas are
significantly different from each other in predicting all-cause mortality (45).

55
In all analyses, we use sample weights that adjust for differential sampling and
nonresponse probabilities to make the results representative of the older U.S. communitydwelling population. We conducted statistical analyses using STATA version 16.
3.3 Results
3.3.1 Sample Characteristics
Table 3.1 shows the mean of each indicator of kidney functioning and sample
characteristics in 2016. Over half (54%) of our sample is female; about 80% of the sample are
White, 10.4% are Black, and 9.5% are Hispanic. The mean age is 68.8 years. About a quarter
(27.0%) of the sample are considered to have normal weight, 2% are considered underweight,
37.6% are overweight and 33.4% are considered obese. We also examined the presence of
comorbidities associated with kidney disease. About one quarter (25.2%) of the sample had
reported ever having diabetes, 59.3% reported every having hypertension and 26.0% reported
every having some kind of cardiovascular problems.
3.3.2 CKD Prevalence
Figure 3.1 indicates the prevalence of impaired kidney functioning varies across our
indicators. It ranges from 15.0% using the serum level of Scr to 58.8% using the serum level of
CysC. We also compare the prevalence of older adults with impaired kidney function between
eGFR equations with and without the racial adjustment (see Figure 3.1 and Figure S3.1). For
eGFR derived from Scr, the overall prevalence is 3.3 percentage points higher with racial
adjustment in comparison to the race-free 2021 equation. Removing the racial adjustment results
in lower prevalence for the White and the overall sample but slightly higher prevalence for Black

56
sample. About 5% of Black participants moved from the normal functioning group to the
impaired group when we removed the racial adjustment (Figure S3.1). Similarly, prevalence
based on GFRcysc_scr is 2.3 percentage points higher with racial adjustment. About 4% of
Black participants moved from the normal functioning group to the impaired group when we
removed the racial adjustment (Figure S3.1).
3.3.3 Age/Sex/Race/Ethnicity Patterns
Table 3.2 presents differences by age, sex, race/ethnicity in prevalence of impaired
kidney functioning, controlling for all demographic factors. For all indicators, older age is
associated with more impaired kidney functioning, with the odds ratios ranging from 1.07 to
1.14. Women are more likely to have impaired kidney functioning than men based on Scr,
eGFRcysc and race-adjusted eGFRcysc_scr but less likely to have impaired functioning based on
eGFRcysc_scr unadjusted for race. We find no difference by sex using the CysC and eGFRscr,
adjusted or unadjusted for race. There are no differences between Hispanic and White
participants using any of the indicators. Black participants are more likely to have impaired
kidney functioning than White participants by all indicators except for CysC (p=0.976) and
eGFRcysc (p=0.057). The magnitude of the Black-White difference varies by indicator. For
instance, using the race-adjusted eGFRscr equation, the odds ratio for Black participants having
impaired kidney functioning is 1.28 (p=0.018), while based on the eGFRscr indicator unadjusted
for race it is 2.52 (p<0.000), when non-Hispanic White participants are the reference group.
Hispanic participants are less likely to have impaired kidney functioning than Black participants
on all indicators except for CysC and eGFRcysc (See Table S3.2).

57
Our multinomial logistic regression results in Table 3.3 show that the likelihood of
having moderately impaired kidney functioning (stage 3) and severely impaired kidney
functioning (stage 4) is significantly higher among women than men on the eGFRcysc and raceadjusted eGFRcysc_scr indicators. Women are less likely to have moderately impaired kidney
functioning on eGFRcysc_scr unadjusted for race. There is no sex difference for kidney failure.
Black and Hispanic participants are generally more likely than White participants to have
kidney failure across all indicators. Hispanic participants are less likely than White participants
to have mild and moderate impairment on all indicators except for eGFRcysc. Black participants
are more likely to have mild impairment than participants on eGFRscr adjusted for race, and
eGFRcysc_scr adjusted for race. However, when eGFRscr is unadjusted for race, White
participants are more likely to have mild impairment than Black participants. Interestingly, Black
participants are more likely to have moderate impairment than White participants on eGFRscr
and eGFRcysc_scr when they are adjusted for race, but these associations both change direction
when they are adjusted for race. Compared to White participants, Hispanic participants are more
likely to have severely impaired kidney functioning only on eGFRscr and Black participants are
more likely to have severely impaired kidney functioning across all indicators. Hispanic
participants are more likely than Black participants to have stage 3 CKD on all indicators but
eGFRcysc. Black participants are only more likely than White participants to have severely
impaired kidney functioning on indicators that do not adjust for race. There is no difference
between Black and Hispanic participants in kidney failure.

58
3.3.4 Predicting Mortality
One way to assess the utility of each indicator for understanding kidney health is to
determine how it relates to subsequent health outcomes. We examine mortality within 4 years as
predicted by the at-risk cut-offs for each indicator of kidney functioning. Table 4 shows how the
sensitivity and specificity of these predictions varies across indicators. For example, the
sensitivity of Scr is 35.2%, indicating that poor kidney function defined by Scr correctly
identifies 35.2% of people who died within 4 years. On the other hand, eGFRcysc seems to be
much more accurate, correctly identifying 77.0% of people who died. Specificity, which is how
well each indicator predicts the proportion of respondents who would not die, ranges across
indicators from 44.2% to 85.8%. Of all indicators, eGFRcysc has the highest area under a
receiver operating characteristic curve (AUC) (0.71) with a decent sensitivity (77%) and, which
according to the DeLong test indicates significantly greater accuracy than the other indicators.
The predictive power of indicators appears to be similar for both sexes and races. The sensitivity
is highest for both men and women on eGFRcysc. eGFRcysc is also the only indicator with an
AUC more than 0.70 for both men and women. Table 4 also shows that eGFRcysc has the high
level of sensitivity for both White and Black participants. CysC-based indicators are generally
better at predicting mortality than Scr-based indicators. The AUC are highest for Black
participants when eGFRcysc and eGFRcysc_scr unadjusted for race are used.
3.4 Discussion
Our study uses a nationally representative sample of older adults to investigate how the
use of a variety of indicators of kidney functioning affects the population prevalence of kidney
dysfunction and the extent of disparities by sex and race based on different indicators of kidney

59
functioning. We found widely varying conclusions about level, and age, sex and racial/ethnic
patterns.
First, various indicators suggest contradictory sex disparities in kidney impairment at a
population-level. Our logistic models revealed that three of the seven indicators—Scr, eGFRcysc
and eGFRcysc_scr (adjusted for race)—suggest women are more likely to have impaired kidney
functioning; however, eGFRcysc_scr unadjusted for race suggests men are more likely to have
impaired kidney functioning. Our multinomial model further concluded that this conflict is
evident for moderately impaired kidney functioning, meaning that such differences have
implications for early detection of kidney problems and adopting preventive measures, such as
medication and health behavior change, to prevent kidney disease from progressing to more
severe stages.
It is noteworthy that, while indicators included in this study do not lead to a definitive
statement on whether men or women are more likely to develop kidney diseases, they are
consistent in suggesting men and women are equally likely to develop kidney failure. This
finding differs from that of earlier research suggesting men have a higher risk of entering end
stage renal disease (ESRD) (46, 47). One explanation may be the reliance of many previous
studies on self-reported data as previous research has shown women have less access to
nephrology care and less awareness of impaired kidney functioning, leading to underestimates in
studies based on self-reported CKD data of women (48-50). GFR based on CysC has high
sensitivity levels for both men and women as well as the highest overall AUC. Altogether, our
research suggests that sex disparities in CKD may be more nuanced than suggested by previous
research. Future work should examine which kidney indicators are most suitable for assessing

60
sex disparities in kidney functioning, as our results suggest different indicators generate
significantly different patterns by sex.
Black participants are more likely than White participants to have impaired kidney
functioning on all indicators, except for CysC and that eGFRcysc has a marginal association
(p=0.057). While we find no overall difference between Hispanic and White participants in all
indicators of impaired kidney functioning, Hispanic participants have more kidney failure. This
may be a result of the relative lack of access to health care for Hispanic participants, particularly
for those seeking pre-ESRD care (51). Black participants have worse kidney functioning than
both White and Hispanic participants across most kidney indicators. However, when broken
down into specific CKD stages, the picture becomes particularly unclear when different
indicators were being used and is somewhat related to the inclusion of racial adjustment in the
eGFR equations. For mildly to moderately impaired kidney functioning, the disadvantage of
Black participants is only evident in eGFRscr and eGFRcysc_scr unadjusted for race; when the
racial adjustments are present in eGFRscr and eGFRcysc_scr, Black persons became less likely
to have mildly to moderately impaired kidney functioning than White individuals. Thus, whether
racial adjustment are included in eGFRscr and eGFRcysc_scr indeed affects the direction of
population-level racial disparities between White and Black older adults, particularly in the early
stages of CKD, which is an important finding as detecting kidney problems at an early stage can
lead to cost-effective treatment that prevents development of more severe disease.
We also find that the age pattern for the increase in risk of impaired kidney functioning
varies by kidney indicators. eGFRcysc suggests the odds of having impaired kidney functioning
increase by 14% with every year of age, while Scr suggests a lower increase of 7% each year.
Previous literature showed that varying kidney indicators often yield more different estimates of

61
kidney functioning problems for older age groups (10-14). Our study further confirms this
variation, which may lead to different diagnosis decisions, especially for older persons. To better
understand the variation in indicators by age, we re-ran our models but included age as a
categorical variable (Table S3.3). We find that differences in prevalence across indicators are
particularly pronounced in older age groups. For instance, the odds ratio relative to those
younger than 60 for the 60-69 group only ranges from 1.5 (Scr) to 3.0 (eGFRcysc_scr without
race), whereas for the 80+ group it ranges from 6.5 (Scr) to 34.7 (eGFRcysc).
Our findings suggest that Scr-based kidney indicators have a less pronounced age
gradient than CysC-based indicators. One possible explanation for this may be that older adults
have lower levels of Scr due to age-related loss of muscle mass. Hence, their overall Scr levels
are not as elevated relative to other biomarkers, such as CysC, which are less affected by muscle
mass (14). Our observations about differentials in age patterns suggest the need for further
research to understand which indicators are best to use for the oldest age groups.
Lastly, we found that eGFR estimated from CysC alone has the highest predicting power
for mortality within 4-years among all indicators we assessed. To our knowledge, our study is
the first to use a nationally representative sample to confirm this advantage of CysC. The
advantage of CysC over Scr in predicting mortality may be due to the unreliability of Scr as a
GFR-biomarker among severely ill people given their ongoing loss of muscle mass, altered
distribution volume, etc. (52, 53). Nonetheless, some research have found that CysC may have
better predictive power for mortality due to its independent relationship to cardiovascular health
(54, 55). To further explore the advantages of CysC in predicting mortality, we ran a sensitivity
analysis in which we examine the odds ratio of death using each kidney indicator. We controlled
for health correlates including cardiovascular diseases, hypertension, and diabetes. We observe a

62
significantly stronger association between mortality and GFR equations that include CysC as an
analyte. At the same time, GFR estimated by both CysC and Scr without the racial adjustment
has the same predictive power for mortality as eGFRcysc among Black individuals. Therefore,
among the race-free equations for kidney indicators, GFR estimated by CysC and both CysC and
Scr are slightly better at predicting all-cause mortality for Black individuals than the Scr racefree equation.
Among Black participants, sensitivity levels are higher for eGFRcysc (66.9%) and
eGFRcysc_scr unadjusted for race (59.5%) than for measures that adjust for race (43.6–54.1%),
although, interestingly, eGFRscr unadjusted for race does not necessarily have high sensitivity
among Black older adults (49.4%). Higher sensitivity levels suggest that indicators more
accurately identify individuals with impaired kidney function who are at a greater risk of
experiencing severe illness and mortality. This has significant implications for nephrology, as
people in advanced stages of kidney disease can deteriorate rapidly if assessed with indicators
characterized by low sensitivity and that do not accurately reflect renal functioning.
This study has limitations. First, the sample we use is this study excluded proxy
respondents and nursing home residents, which means our analytic sample may be healthier than
the overall United Stated population. In fact, those not included may have the worst kidney
functioning, as those who require proxy respondents or live in nursing home typically have the
worst overall health. Our results may therefore be somewhat conservative regarding the level of
at-risk kidney functioning among older adults. Second, for our mortality analysis, we examined
all-cause death data. While we limited our sample in the mortality analysis to those who have
impaired kidney functioning, CKD is not the cause of death for most decedents. We ran a
sensitivity analysis (Table S3.5) in which we limited our sample to respondents who died from

63
kidney diseases, heart diseases, diabetes, and hypertension. The sensitivity analysis suggests the
overall conclusions regarding the predictive power for mortality across indicators do not change
with this more selected sample. We could not conduct an analysis limited to those who died from
kidney disease, because the sample size would be too small (N=11). Additionally, we could not
include Hispanic respondents in our mortality analysis due to small sample size. Finally, our
results are based on one-time samples. According to KDIGO guidelines, an ideal scenario would
determine CKD based on two readings 90 days apart. Still, our study does not aim to diagnose
CKD but to use kidney indicators as estimating measures to examine population-level prevalence
and disparities.
This study used a nationally representative sample to paint a comprehensive picture of
CKD prevalence and demographic disparities across kidney indicators. Our approach of using
biomarker data instead of relying on self-reported data alleviates the reporting bias in existing
literature examining CKD prevalence and disparities, especially as individuals who are already
socially disadvantaged tend to be less aware of their conditions. Our findings brought awareness
to the stark differences and/or sometimes conflicting conclusions about population-level
disparities. In terms of implications, while our study does not aim to determine which kidney
indicators are “the best”, this study provided insights that may be useful from a public health
policymaking perspective. We observed that while the indicators are quite consistent in
identifying demographic disparities for the more severe stages of CKD, conflicting conclusions
are more evident for moderately impaired kidney functioning, which is an important finding as
early detection of kidney problems and adaptation of preventive measures may play a key role in
addressing the increasing financial burden of late stage kidney disease and its overall negative
impact on the wellbeing of the growing older population. Secondly, we found that indicators that

64
include CysC as an analyte may be more closely associated with mortality and just the aging
process in general. It is clearly reasonable for labs to implement the race-free GFRscr 2021
equation as Scr remains the most common and available analyte with its decent performance, at
the same time, our study also supports the long-term goal of the National Kidney Foundation
(NKF) and the American Society of Nephrology (ASN) Task Force, which is to recommend
national efforts to facilitate increased, routine, and timely use of cystatin C and development of
next-generation race-independent markers of kidney function, which can be useful in confirming
eGFR in adults who are at risk for or have CKD (56).

65
Table 3.1. Characteristics of community-dwelling population aged 56 years and older: US Health and Retirement
Study (weighted), 2016 (n=8,821)
Variables Mean (SD) %
Cystatin C (mg/L) 1.17 (0.51)
Creatinine (mg/L) 0.95 (0.47)
eGFR CysC-only (mL/min/1.73m2
) 67.1 (22.4)
eGFR Scr-only w/Race (mL/min/1.73m2
) 76.2 (19.2)
eGFR Scr-only w/o Race (mL/min/1.73m2
) 79.1 (19.2)
eGFR CysC_Scr w/ Race (mL/min/1.73m2
) 71.4 (20.5)
eGFR CysC_Scr w/o Race
(mL/min/1.73m2
) 73.7 (21.5)
Race/ethnicity
Non-Hispanic White 80.1
Non-Hispanic Black 10.4
Hispanic 9.5
Age 68.8 (9.41)
Female 54.3
BMI
Normal 27.0
Underweight 2.0
Overweight 37.6
Obese 33.4
History of comorbidities
Ever had diabetes 25.2
Ever had hypertension 59.3
Ever had cardiovascular disease 26.0
Note: “SD” stands for standard deviation; “cysc” stands for cystatin C; “scr” stands for creatinine

66
Figure 3.2. Prevalence of impaired kidney impairment functioning with seven measures (with 95% CI)
Table 3.2. Odds ratios indicating Sex/racial/ethnic differentials from logistic regressions on
seven indicators of impaired kidney functioning (n=8,821)
Cystatin C Creatinine eGFR cysc eGFR scr eGFR scr w/o
race eGFRcysc_scr eGFRcysc_scr
w/o race
OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value
Age 1.12 <0.000 1.07 <0.000 1.14 <0.000 1.10 <0.000 1.10 <0.000 1.12 <0.000 1.12 <0.000
Female 0.85 0.007 1.29 0.001 1.32 <0.000 1.02 0.73 1.10 0.180 1.22 0.003 0.86 0.028
Race/Ethnicity (ref=Non-Hispanic White)
Non-Hispanic
Black 1.00 0.976 2.5 <0.000 1.2 0.057 1.28 0.018 2.52 <0.000 1.22 0.031 1.91 <0.000
Hispanic 0.98 0.826 0.81 0.119 1.01 0.896 0.79 0.075 0.92 0.531 0.87 0.24 0.9 0.401
Pseudo R2 0.136 0.079 0.2 0.135 0.142 0.184 0.178
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
Cystatin C Creatinine GFR cysc GFR_scr w race GFR_scr w/o
race
GFR_cysc_scr
w race
GFR_cysc_scr
w/o race

67
Table 3.3 Odds rations indicating from multinomial regressions indicating sex/racial/ethnic
differentials in severity level of kidney functioning (n=8,821)
Reference = Stage 1 - Normal kidney functioning eGFR cysc eGFR scr w race eGFR scr w/o race eGFRcysc_ scr w
race
eGFRcysc_ scr w/o
race
OR P-value OR P-value OR P-value OR P-value OR P-value
Stage 2 - Mildly impaired Age 1.09 <0.000 1.13 <0.000 1.1 <0.000 1.12 <0.000 1.11 <0.000
Female 1.11 0.182 0.9 0.196 1.01 0.783 1.12 0.169 0.9 0.2
Race
Race/Ethnicity (ref=Non-Hispanic
White)
Non-Hispanic
Black 0.89 0.291 0.49 <0.000 1.87 <0.000 0.59 <0.000 1.18 0.088
Hispanic 0.94 0.638 0.51 <0.000 0.54 <0.000 0.65 <0.000 0.73 0.003
Stage 3 - Moderately impaired Age 1.22 <0.000 1.22 <0.000 1.18 <0.000 1.24 <0.000 1.22 <0.000
Female 1.4 <0.000 0.96 0.646 1.13 0.156 1.32 0.003 0.79 0.01
Race/Ethnicity (ref=Non-Hispanic
White)
Non-Hispanic
Black 0.98 0.93 0.66 0.001 3.49 <0.000 0.71 0.005 1.92 <0.000
Hispanic 0.89 0.457 0.4 <0.000 0.53 <0.000 0.55 <0.000 0.62 0.001
Stage 4 - Severely impaired Age 1.28 <0.000 1.24 <0.000 1.19 <0.000 1.69 0.007 1.25 <0.000
Female 1.91 <0.000 0.97 0.912 1.16 0.571 1.54 0.017 1.05 0.777
Race/Ethnicity
Non-Hispanic
Black 1.7 0.018 1.91 0.054 8.62 <0.000 1.58 0.077 4.03 <0.000
Hispanic 1.38 0.17 1.22 0.573 1.3 0.54 1.12 0.674 1.35 0.291
Stage 5 - Kidney failure Age 1.23 <0.000 1.15 <0.000 1.09 <0.000 0.9 0.745 1.18 <0.000
Female 1.15 0.609 0.67 0.256 0.76 0.445 0.82 0.533 0.63 0.162
Race/Ethnicity
Non-Hispanic
Black 5.29 <0.000 2.79 0.006 10.96 <0.000 3.3 0.001 7.02 <0.000
Hispanic 3.6 0.001 2.36 0.051 3.39 0.007 2.54 0.019 3.04 0.005
Pseudo R2 0.133 0.131 0.114 0.137 0.128
Note: “cysc” stands for cystatin C; “scr” stands for creatinine

68
Table 3.4 Sensitivity and specificity of equations predicting all-cause mortality by each kidney
functioning indicator, stratified by sex and race
Kidney measure Sensitivity % Specificity % AUC
Creatinine 35.2 85.8 0.60 [0.59, 0.62]
Cystatin C 62.4 75.8 0.69 [0.67, 0.71]
eGFRcysc 77.0 64.6 0.71 [0.69, 0.72]
All eGFRscr adjusted
for race 45.6 82.4 0.64 [0.62, 0.66]
eGFRscr
unadjusted for race 40.2 84.6 0.62 [0.61, 0.64]
eGFRcysc_scr
adjusted for race 61.8 74.9 0.68 [0.67, 0.70]
eGFRcysc_scr
unadjusted for race 59.7 76.9 0.68 [0.67, 0.70]
Creatinine 36.5 84.2 0.60 [0.58, 0.62]
Cystatin C 58.3 75.9 0.67 [0.65, 0.69]
eGFRcysc 76.7 62.4 0.70 [0.68, 0.72]
Women eGFRscr adjusted
for race 44.7 82.0 0.63 [0.61, 0.66]
eGFRscr
unadjusted for race 39.6 83.8 0.62 [0.60, 0.64]
eGFRcysc_scr
adjusted for race 61.9 73.4 0.68 [0.65, 0.70]
eGFRcysc_scr
unadjusted for race 56.3 77.2 0.67 [0.65, 0.69]
Creatinine 33.6 88.1 0.61 [0.59, 0.63]
Cystatin C 67.0 75.6 0.71 [0.69, 0.74]
eGFRcysc 77.4 67.8 0.73 [0.71, 0.75]
eGFRscr adjusted
for race 46.6 82.9 0.65 [0.62, 0.67]
Men eGFRscr
unadjusted for race 40.9 85.7 0.63 [0.61, 0.66]

69
eGFRcysc_scr
adjusted for race 61.5 77.0 0.69 [0.67, 0.72]
eGFRcysc_scr
unadjusted for race 63.5 76.6 0.70 [0.68, 0.72]
Creatinine 31.9 86.6 0.59 [0.58, 0.61]
Cystatin C 63.9 73.7 0.69 [0.67, 0.71]
eGFRcysc 79.6 61.0 0.70 [0.69, 0.72]
eGFRscr adjusted
for race 46.0 80.2 0.63 [0.61, 0.65]
NonHispanic
White
eGFRscr
unadjusted for race 37.6 84.7 0.61 [0.59, 0.63]
eGFRcysc_scr
adjusted for race 64.2 71.9 0.68 [0.66, 0.70]
eGFRcysc_scr
unadjusted for race 60.3 75.3 0.68 [0.66, 0.70]
Creatinine 48.8 78.1 0.63 [0.60, 0.67]
Cystatin C 54.1 77.7 0.66 [0.62 0.70]
eGFRcysc 66.9 70.0 0.68 [0.65, 0.72]
eGFRscr adjusted
for race 43.6 84.3 0.64 [0.60, 0.68]
NonHispanic
Black
eGFRscr unadjusted
for race 49.4 78.8 0.64 [0.60, 0.68]
eGFRcysc_scr
adjusted for race 52.9 78.3 0.66 [0.62, 0.69]
eGFRcysc_scr
unadjusted for race 59.3 75.9 0.68 [0.64, 0.71]
Note: “cysc” stands for cystatin C; “scr” stands for creatinine

70
3.5 References
1. Crimmins E, Vasunilashorn S, Kim JK, Alley D. Chapter 5 Biomarkers Related
To Aging In Human Populations. Advances in Clinical Chemistry. Published
online 2008:161-216. doi:https://doi.org/10.1016/s0065-2423(08)00405-8
2. Jones C, McQuillan G, Kusek J, et al. Serum creatinine levels in the US population:
Third National Health and Nutrition Examination Survey. American Journal of Kidney
Diseases. 1998;32(6):992-999. doi:https://doi.org/10.1016/s0272-6386(98)70074-5
3. Nissenson AR, Pereira BJG, Collins AJ, Steinberg EP. Prevalence and characteristics of
individuals with chronic kidney disease in a large health maintenance organization.
American Journal of Kidney Diseases. 2001;37(6):1177-1183.
doi:https://doi.org/10.1053/ajkd.2001.24520
4. Jain A, Sutton K, Roxburgh S, et al. SUN-094 A single centre review of end-of-life
discussions and location of death in a dialysis population between 2016 and 2017. Kidney
International Reports. 2019;4(7):S195. doi:https://doi.org/10.1016/j.ekir.2019.05.491
5. Stephan Y, Sutin AR, Terracciano A. Subjective Age and Cystatin C Among Older
Adults.J Gerontol B Psychol Sci Soc Sci. 2019;74(3):382-388.
doi:10.1093/geronb/gbx124
6. Bishop N, Zhu J. Investigating Racial/Ethnic Disparities in the Association Between
AHEI-2010 and Change in Cystatin C in a Representative Sample of Older Americans.
Current Developments in Nutrition. 2021;5:1015.
doi:https://doi.org/10.1093/cdn/nzab053_008
7. Smith, H. W. (1951). The kidney: structure and function in health and disease.
Oxford University Press, USA.

71
8. Levey AS, de Jong PE, Coresh J, et al. The definition, classification, and prognosis of
chronic kidney disease: a KDIGO Controversies Conference report. Kidney International.
2011;80(1):17-28. doi:https://doi.org/10.1038/ki.2010.483
9. Fishbane S, Nissenson AR. Anemia management in chronic kidney disease. Kidney Int
Suppl. 2010;(117):S3-S9. doi:10.1038/ki.2010.188
10. Tidman M, Sjöström P, Jones I. A Comparison of GFR estimating formulae based upon
s-cystatin C and s-creatinine and a combination of the two. Nephrol Dial Transplant.
2008;23(1):154-160. doi:10.1093/ndt/gfm661
11. Stevens LA, Schmid CH, Greene T, et al. Comparative performance of the CKD
Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease
(MDRD) Study equations for estimating GFR levels above 60 mL/min/1.73 m2. Am J
Kidney Dis. 2010;56(3):486-495. doi:10.1053/j.ajkd.2010.03.026
12. Mombelli CA, Giordani MC, Imperiali NC, et al. Single-Center Long-Term Follow-Up
of Kidney Donors in Argentina (Hospital Italiano de Buenos Aires). Transplant Proc.
2018;50(2):418-422. doi:10.1016/j.transproceed.2017.12.046
13. Åkerblom A, Helmersson-Karlqvist J, Flodin M, Larsson A. Comparison between
Cystatin C- and Creatinine-Estimated Glomerular Filtration Rate in Cardiology
Patients. Cardiorenal Med. 2015;5(4):289-296. doi:10.1159/000437273
14. Legrand H, Werner K, Christensson A, Pihlsgård M, Elmståhl S. Prevalence and
determinants of differences in cystatin C and creatinine-based estimated glomerular
filtration rate in community-dwelling older adults: a cross-sectional study. BMC
Nephrology. 2017;18(1). doi:https://doi.org/10.1186/s12882-017-0759-3

72
15. Köttgen A, Selvin E, Stevens LA, Levey AS, Van Lente F, Coresh J. Serum Cystatin C in
the United States: The Third National Health and Nutrition Examination Survey
(NHANES III). American Journal of Kidney Diseases. 2008;51(3):385-394.
doi:https://doi.org/10.1053/j.ajkd.2007.11.019
16. Coresh J, Astor BC, Greene T, Eknoyan G, Levey AS. Prevalence of chronic kidney
disease and decreased kidney function in the adult US population: Third national health
and nutrition examination survey. American Journal of Kidney Diseases. 2003;41(1):1-
12. doi:https://doi.org/10.1053/ajkd.2003.50007
17. Florkowski, C. M., & Chew-Harris, J. S. (2011). Methods of estimating GFR–different
equations including CKD-EPI. The Clinical Biochemist Reviews, 32(2), 75.
18. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to
estimate glomerular filtration rate from serum creatinine: a new prediction equation.
Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130(6):461-
470. doi:10.7326/0003-4819-130-6-199903160-00002
19. Levey AS, Stevens LA, Schmid CH, et al. A New Equation to Estimate Glomerular
Filtration Rate. Annals of Internal Medicine. 2009;150(9):604.
doi:https://doi.org/10.7326/0003-4819-150-9-200905050-00006
20. Carrero JJ, Hecking M, Chesnaye NC, Jager KJ. Sex and gender disparities in the
epidemiology and outcomes of chronic kidney disease. Nat Rev Nephrol. 2018;14(3):151-
164. doi:10.1038/nrneph.2017.181
21. Cohn SH, Brennan BL, Yasumura S, Vartsky D, Vaswani AN, Ellis KJ. Evaluation of
body composition and nitrogen content of renal patients on chronic dialysis as determined

73
by total body neutron activation. Am J Clin Nutr. 1983;38(1):52-58.
doi:10.1093/ajcn/38.1.52
22. Harsha DW, Frerichs RR, Berenson GS. Densitometry and anthropometry of black and
white children. Hum Biol. 1978;50(3):261-280.
23. Worrall JG, Phongsathorn V, Hooper RJ, Paice EW. Racial variation in serum creatine
kinase unrelated to lean body mass. Br J Rheumatol. 1990;29(5):371-373.
doi:10.1093/rheumatology/29.5.371
24. Diao, J. A., Wu, G. J., Taylor, H. A., Tucker, J. K., Powe, N. R., Kohane, I. S., & Manrai,
A. K. (2021). Clinical implications of removing race from estimates of kidney function.
Jama, 325(2), 184-186. doi:10.1001/jama.2020.22124
25. Ahmed, S., Nutt, C. T., Eneanya, N. D., Reese, P. P., Sivashanker, K., Morse, M., ... &
Mendu, M. L. (2021). Examining the potential impact of race multiplier utilization in
estimated glomerular filtration rate calculation on African-American care outcomes.
Journal of General Internal Medicine, 36, 464-471. https://doi.org/10.1007/s11606-020-
06280-5
26. Parekh RS, Perl J, Auguste B, Sood MM. Elimination of race in estimates of kidney
function to provide unbiased clinical management in Canada. CMAJ.
2022;194(11):E421-E423. doi:10.1503/cmaj.210838
27. Tsai JW, Cerdeña JP, Goedel WC, et al. Evaluating the Impact and Rationale of RaceSpecific Estimations of Kidney Function: Estimations from U.S. NHANES, 2015-2018.
EClinicalMedicine. 2021;42:101197. Published 2021 Nov 19.
doi:10.1016/j.eclinm.2021.101197

74
28. Levey, A. S., Titan, S. M., Powe, N. R., Coresh, J., & Inker, L. A. (2020). Kidney
disease, race, and GFR estimation. Clinical journal of the American Society of
Nephrology: CJASN, 15(8), 1203. doi: 10.2215/CJN.12791019
29. Tonelli M, Wiebe N, Culleton B, et al. Chronic kidney disease and mortality risk: a
systematic review. J Am Soc Nephrol. 2006;17(7):2034-2047.
doi:10.1681/ASN.2005101085
30. Hansrivijit P, Chen YJ, Lnu K, et al. Prediction of mortality among patients with chronic
kidney disease: A systematic review. World J Nephrol. 2021;10(4):59-75.
doi:10.5527/wjn.v10.i4.59
31. Crimmins E, Faul J, Thyagarajan B, Weir D. Venous Blood Collection and Assay
Protocol in the 2016 Health and Retirement Study. hrs.isr.umich.edu. Published 2017.
Accessed April 20, 2023. https://hrs.isr.umich.edu/publications/biblio/9065
32. Dharnidharka VR, Kwon C, Stevens G. Serum cystatin C is superior to serum creatinine
as a marker of kidney function: a meta-analysis. Am J Kidney Dis. 2002;40(2):221-226.
doi:10.1053/ajkd.2002.34487
33. Murty MS, Sharma UK, Pandey VB, Kankare SB. Serum cystatin C as a marker of renal
function in detection of early acute kidney injury. Indian J Nephrol. 2013;23(3):180-183.
doi:10.4103/0971-4065.111840
34. Peiris H, Chandrasena LG, Lanerolle RD. Serum cystatin C as a marker to identify
patients with moderately impaired renal function. Indian J Clin Biochem.
2008;23(2):163-166. doi:10.1007/s12291-008-0036-0

75
35. Duru OK, Vargas RB, Kermah D, Nissenson AR, Norris KC. High prevalence of stage 3
chronic kidney disease in older adults despite normal serum creatinine. J Gen Intern Med.
2009;24(1):86-92. doi:10.1007/s11606-008-0850-3
36. Inker LA, Eneanya ND, Coresh J, et al. New Creatinine- and Cystatin C-Based Equations
to Estimate GFR without Race. The New England Journal of Medicine.
2021;385(19):1737-1749. doi:https://doi.org/10.1056/NEJMoa2102953
37. Albertus P, Morgenstern H, Robinson B, Saran R. Risk of ESRD in the United States.
American Journal of Kidney Diseases. 2016;68(6):862-872.
doi:https://doi.org/10.1053/j.ajkd.2016.05.030
38. Cobo G, Hecking M, Port FK, et al. Sex and gender differences in chronic kidney
disease: progression to end-stage renal disease and haemodialysis. Clinical Science.
2016;130(14):1147-1163. doi:https://doi.org/10.1042/cs20160047
39. Coresh J, Byrd-Holt D, Astor BC, et al. Chronic Kidney Disease Awareness, Prevalence,
and Trends among U.S. Adults, 1999 to 2000. Journal of the American Society of
Nephrology. 2004;16(1):180-188. doi:https://doi.org/10.1681/asn.2004070539
40. Lewandowski MJ, Krenn S, Kurnikowski A, et al. Chronic kidney disease is more
prevalent among women but more men than women are under nephrological care. Wiener
klinische Wochenschrift. Published online August 31, 2022.
doi:https://doi.org/10.1007/s00508-022-02074-3
41. Markides KS, Coreil J. The health of Hispanics in the southwestern United States: an
epidemiologic paradox. Public Health Rep. 1986;101(3):253-265. PMID: 3086917;
PMCID: PMC1477704.

76
42. Desai N, Lora CM, Lash JP, Ricardo AC. CKD and ESRD in US Hispanics. American
Journal of Kidney Diseases. 2019;73(1):102-111.
doi:https://doi.org/10.1053/j.ajkd.2018.02.354
43. Bentley JN, Viswanathan A, Rosenberg WS, Patil PG. Treatment of medically refractory
cancer pain with a combination of intrathecal neuromodulation and neurosurgical
ablation: case series and literature review. Pain Med. 2014;15(9):1488-1495.
doi:10.1111/pme.12481
44. Kevin Tucker J. Social Justice as a Tool to Eliminate Inequities in Kidney Disease.
Seminars in Nephrology. 2021;41(3):203-210.
doi:https://doi.org/10.1016/j.semnephrol.2021.05.001
45. Willey JZ, Moon YP, Husain SA, et al. Creatinine versus cystatin C for renal functionbased mortality prediction in an elderly cohort: The Northern Manhattan Study.
Shimosawa T, ed. PLOS ONE. 2020;15(1):e0226509.
doi:https://doi.org/10.1371/journal.pone.0226509
46. Helmersson‐Karlqvist J, Miklós Lipcsey, Johan Ärnlöv, et al. Addition of cystatin C
predicts cardiovascular death better than creatinine in intensive care. Heart.
2021;108(4):279-284. doi:https://doi.org/10.1136/heartjnl-2020-318860
47. Puthucheary ZA, Rawal J, McPhail M, et al. Acute Skeletal Muscle Wasting in Critical
Illness. JAMA. 2013;310(15):1591. doi:https://doi.org/10.1001/jama.2013.278481
48. Prowle JR, Chua HR, Bagshaw SM, Bellomo R. Clinical review: Volume of fluid
resuscitation and the incidence of acute kidney injury - a systematic review. Crit Care.
2012;16(4):230. Published 2012 Aug 7. doi:10.1186/cc11345

77
49. Shlipak MG, Sarnak MJ, Katz R, et al. Cystatin C and the risk of death and
cardiovascular events among elderly persons. N Engl J Med. 2005;352(20):2049-2060.
doi:10.1056/NEJMoa043161
50. Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J
Thorac Oncol. 2010;5(9):1315-1316. doi:10.1097/JTO.0b013e3181ec173d

78
Chapter 4: Comparisons of SES disparities in CKD
prevalence and progression over time in China and U.S.
4.1 Introduction
As the prevalence of chronic kidney disease (CKD) is on the rise globally and is putting a
significant financial burden on healthcare systems, understanding causes and management of
renal health has become an increasingly important topic. Globally, the number of individuals
with CKD was nearly 700 million in 2017, surpassing those with diabetes, asthma, or depressive
disorders (1). By 2040, chronic kidney disease is estimated to become the fifth leading cause of
death globally—one of the largest projected increases of any major cause of death (2, 3).
Without proper screening, detection, and management, older adults experiencing kidney decline
may exhibit high risk of morbidities, substantial impairment in quality of life and even mortality
within a short period of time (4, 5). Further understanding of decline in kidney functioning
requires longitudinal data that allow us to observe which risk factors contribute to a more rapid
trajectory of kidney decline and which are modifiable, particularly at older ages.
Research suggests that one’s socioeconomic status (SES) is one of the most significant
predictors for deterioration in kidney functioning, with lower SES related to faster decline at
older ages (6, 7). This SES gap in CKD is likely linked to differentials in 1) comorbidities, (e.g.,
diabetes, hypertension and cardiovascular conditions), 2) adverse health behaviors (e.g.,
smoking, lack of physical activity, unhealthy diet), and 3) lack of access to medical insurance,
preventative care and treatment (8-10). Further, SES is linked to kidney-related death as well,
with research suggesting lower education attainment is associated with higher risk of mortality
among people with kidney diseases (11). Advancing our understanding of the role of

79
socioeconomic factors in older individuals with or at risk for CKD can lead to improved
strategies for disease prevention/management, and even higher chance of survival for those with
severe kidney problems.
However, CKD is a complicated disease and is associated with different risk factors in
countries/regions with various epidemiological contexts, resulting in differences in the
mechanisms behind the associations with SES and kidney functioning and thus requiring a
variety of approaches for prevention and treatment (12). First, CKD has various causes, and
those causes vary across contexts. For instance, for developed societies like the United States,
comorbidities such as diabetes, hypertension, and cardiovascular diseases are the leading causes
of chronic kidney disease and are heavily influenced by health behaviors and lifestyles (13);
however, for developing countries such as China, while comorbidities that traditionally drive
disease in developed country are growing rapidly, they can face a constellation of additional
risks that translate to a greater kidney disease burden, such as glomerulonephritis resulting from
infections and toxins (14, 15). This could mean that, while low socioeconomic status is an
important social driver contributing to one’s vulnerability to CKD, the underlying associations
and/or the strength of those associations linking SES to CKD may be different – management of
comorbidities and change in modifiable behaviors/lifestyles may be a more effective preventive
measure in addressing the SES disparities in kidney health in the U.S., while improvement of
environment and overall prevention of infectious diseases might also be useful in addressing SES
disparities of kidney health in China.
Second, socioeconomic factors such as income level and insurance types, which have
effects on both access to and quality of treatments, may influence CKD differently between
China and the U.S. (16, 17). For example, in China, while a large portion of prevention and

80
treatment expenses are reimbursed by public insurance plans, there is some variation in the
reimbursement rate for CKD treatment (18), which then require different levels of disposable
income and savings. In China, insurance plans can differ based on one’s residential status as
urban residents are believed to be entitled to better healthcare coverage and services compared to
rural residents due to lack of medical resources (19, 20). On the other hand, in the U.S.,
healthcare services related to CKD are mostly covered by public insurance such as Medicare,
including preventive care, immunosuppressant medicines, and also procedures for severe kidney
problems like dialysis and transplants (21). Nevertheless, there has been some evidence that
having access to private insurance, which is associated with higher SES, may be related to better
health outcomes than those associated with having only public insurance (22). Therefore,
healthcare access may play an important yet distinct role in contributing to SES disparities in
CKD in the two countries.
Additionally, this cross-national comparison is particularly meaningful because, in
contrast to many other chronic diseases, people may die very quickly if they do not receive
immediate treatment after a kidney fails. Those who do not have access to renal transplant face a
high risk of death (23). As healthcare policies and available prevention and treatments regarding
CKD work through very different mechanisms between the two countries, we expect that the
SES disparities in kidney-related mortality are likely to differ across the two countries.
In this paper, we compare SES disparity in CKD among middle age and older persons in
the U.S. with that in China, aiming to examine the consistency of associations between SES and
CKD and the underlying mechanisms across two countries. We take advantage of two waves of
available biomarker data in the Health and Retirement Study (HRS) and the China Health and
Retirement Longitudinal Study (CHARLS), both of which are nationally representative, to

81
examine and compare 1) the strengths of associations between baseline kidney functioning and
SES, comorbidity, and adverse behaviors in China and the U.S. in 2011, 2) the strengths of
associations between the progression of deterioration in kidney functioning over 4-years and SES
and risk factors between China and the U.S., 3) and whether the SES disparity in mortality
differs between China and U.S. among those with reduced kidney functioning.
4.2 Data and Methods
4.2.1 Data
The China Health and Retirement Longitudinal Study (CHARLS) was designed as a
biennial nationally representative longitudinal survey for households with members aged 45
years and above. The baseline national wave of CHARLS is being fielded in 2011 and includes
about 10,000 households and 17,708 individuals in 150 counties/districts and 450
villages/resident committees. It followed stratified random sampling procedures with multi-stage
(counties-villages-households) PPS sampling. The individuals are followed up every two years
(24). The survey is harmonized with the HRS family of surveys. Our analysis uses the venous
blood samples collected in wave 1 (2011) and wave 3 (2015) to assess kidney functioning.
Bioassays were performed at the Youanmen Center for Clinical Laboratory of Capital Medical
University in 2011 and at KingMed Diagnostics, a leading testing laboratory, in 2015 (25).
Cystatin C (CysC) is used to assess kidney functioning, which is considered a good indicator for
population-level studies and is superior to Scr as a marker of kidney functioning in population
studies, especially among older people, because its value is not affected by sex, age,
race/ethnicity, protein intake, or muscle mass (26). For the U.S. we use data from the Health and
Retirement Study (HRS) which is a longitudinal study of Americans over age 50. HRS collected
dried blood spots (DBS) in 2010/2012 for the first half of the sample, and in 2014 and 2016 for

82
the second half (27). Interviewers mailed the DBS cards directly to the University of Michigan,
where they were sorted, frozen and shipped to the University of Washington where cystatin C
was assayed. Therefore, we are able to examine kidney functioning at two time points and
observe change in both China and U.S. For the Chinese sample, among the 17,708 individuals
who were interviewed, 3,374 of them did not have biomarker data at baseline; 979 passed away
by 2015, 1,927 were excluded due to losses of follow-up by 2015. 6,834 did not have CysC at
baseline and 5,922 did not have CysC at in 2015. After excluding those had missing data on
covariates, the final sample size is 4,016 eligible participants for China. In terms of the U.S.
sample, 15,378 individuals were interviewed in 2010/2012, and 6,985 of them did not have
biomarker data. By 2014/2016, 968 have passed away, and 2,987 were excluded due to loss of
follow-up. 414 individuals did not have CysC data in 2010/2012, and 1,873 did not have CysC in
2014/2016. After excluding those with missingness on covariates, the final sample for the U.S. is
5,497.
For analysis of risk factors, we only include people with CysC data available at both
waves. To examine SES disparities in mortality among those with reduced kidney functioning,
we run a hazard analysis on mortality over the 4-year period among those who have moderately
to severely impaired kidney functioning to determine the factors associated with dying in the two
countries. This analysis includes 1,449 people for China and 2,948 people for the U.S.
4.2.2 Measures
Cystatin C. CysC (mg/L) is ubiquitous in the tissue cells/body fluid and has a regular rate
of production; it is entirely filtered by glomerulus, which makes the kidney its only metabolic
pathway. When kidney functioning declines, CysC becomes more concentrated, and its level rises

83
(28). The HRS cystatin C measure is based on dried blood spot (DBS) assays performed in a series
of labs over the ten-year period. We use the venous blood equivalent values for cystatin C provided
by HRS, rather than the DBS value itself (29). The equivalent value is constructed assuming that
the distribution of the DBS assays is similar to the venous blood values in the National Health and
Nutrition Examination Survey (NHANES) national sample of the same age while preserving the
variability in the HRS sample. However, because cystatin C has not been regularly done in
NHANES, all years of the HRS study use the same NHANES sample (1999-2002) to construct
serum equivalent values. This means that there is no population-level change with time reflected
in the cystatin C measure in the HRS; however, within-person change can still be assessed.
Similarly, for CHARLS, we constructed the 2015 CysC data to have the same distribution as 2011
so that the data at two waves are comparable. This is because assay kits for CHARLS changed
between 2011 and 2015.
Kidney Functioning: While cystatin C provides the best measure of kidney functioning for
populations as it is not affected by gender, age, race/ethnicity, protein intake, or muscle mass (26,
30), cystatin C can be used to estimate the glomerular filtration rate (eGFR), which is a measure
used in clinical settings for staging kidney disease. Our eGFR is based on assayed levels of
Cystatin C using the following equations, which are believed to have good accuracy for CKD
diagnosis and GFR staging (31).
• eGFR = 133 * CysC-0.499 * 0.996Age * 0.932 [if female], if CysC ≤ 0.8
• eGFR = 133 * CysC-1.328 * 0.996Age * 0.932 [if female], if CysC > 0.8
We use the eGFR values to make a four-stage classification of kidney functioning – normal (≥90
mL/min/1.73m2), mildly impaired kidney function (60-89 mL/min/1.73m2), moderately impaired
(30-59) and severely impaired kidney function (0-30 mL/min/1.73m2). We also generated a

84
categorical variable indicating how a respondent’s CKD stage changed over the 4-year period.
This variable has four categories: Continuously good, Continuously bad, Improved, and Worsened.
For instance, if a participant’s eGFR changed from mildly impaired to moderately/severely
impaired kidney function, then they would be classified as Worsened; if a participant went from
mildly impaired to normal, then this person would be considered Improved. They would be
considered Continuously good if they had normal or mildly impaired kidney functioning in both
waves and would be subsequently considered Continuously bad if they had moderately to severely
impaired kidney functioning in both waves.
Education. For both U.S. and China, Education is categorized into lower, middle and
higher education. However, because Chinese older adults received less education compared to
American older adults, the education grouping for China is no formal education, at least some
primary school experience and literate, and more than primary school, while the U.S. categories
are less than high school completion, high school graduates, and at least some college.
Insurance. For the U.S., insurance plans are divided into four categories – 1) Public
insurance only, 2) Private insurance only, 3) both Public and private, and 4) no insurance. For
China, the most prevalent types of insurances are Urban employee, Urban resident, Rural new
cooperative, Private, and other public plans. Urban employee and private plans are generally
available to those who have formal employment or those with some purchasing power and are
typically associated with better benefits and access to health services relative to Urban resident
and other supplemental public plans, which target the unemployed and older population (20).
People with rural new cooperative are also typically considered to have less access to and lower
quality of healthcare services in comparison to the urban plans (20). Therefore, we categorized the

85
Chinese insurance types into four groups which reflect four types of access to care: 1) Urban
employee and/or private plans, 2) Urban resident and/or other public plans, 3) Rural new
cooperative, and 4) no insurance.
Comorbidities. Reports of prior doctor diagnosis of hypertension and diabetes are reported
at each interview. In addition, measured levels of blood pressure and HbA1c are used to indicate
unreported hypertension and diabetes as well as to divide reported disease into controlled or not
controlled. The variables for these two diseases are divided into three categories based on the
following criteria: 1) No history - no self-report, not measured high, and no medication, 2)
Controlled – self-report, not measured high, 3) Uncontrolled – self-report, measured high.
Individuals are considered hypertensive if their average systolic blood pressure is equal
or greater than 140 mmHg and/or if the average diastolic blood pressure is equal or greater than
90 mmHg. Diabetes was evaluated using HbA1c which gives a 2-3 month estimate of glucose
control and is considered an accurate reflection of long-term glucose regulation. Values ≥6.5%
are used to indicate unreported diagnosis and to divide diabetes into controlled and uncontrolled.
Health Behaviors. We include smoking (Never smoked, former smokers and current
smokers) and drinking (Never drinker, former drinker, and current drinker) to indicate a
respondent’s health behaviors.
Death. We constructed an all-cause mortality variable using data on death year and date
from HRS and CHARLS tracker files to determine who died by 2015 for China, or 2014/2016
for the U.S. For China, people with a confirmed death after the beginning of 2011 and before the
end of 2015 are coded as 1; people who are alive at the end of 2015 are coded as 0. For the U.S.,

86
those interviewed in 2010 and confirmed dead by the end of 2014 or those interviewed in 2012
and then confirmed dead by 2016 are coded as 1; those that are alive are coded as 0.
Covariates: Age in 2011 for CHARLS and 2010/2012 for HRS and gender are included
in all equations. For the U.S., we have four age groups based on their age at baseline, 50-59, 60-
69, 70-79 and 80+. For China, we combined the 70-79 and 80+ groups as there are not enough
Chinese respondents over age 80 in the sample. Gender is measured using a dummy variable for
female.
4.2.3 Analytic Strategy
We first compare the characteristics of the two samples from the U.S. and China. We then
determine the age-controlled prevalence of CKD by severity for both countries, for this we
standardize the American population to the same age structure as China to eliminate the effect of
age differences. To understand baseline level differences, we run a cross-sectional multinomial
regression model that examines associations between baseline stage of CKD and education,
insurance types, comorbidity management, and health behaviors. The outcome variable for this
has three categories, 1) normal, 2) mildly impaired and 3) moderately/severely impaired. We
combined moderately and severely impaired because there are too few people with severe stage of
CKD in China. As we lost people due to missing data on the second wave, we ran a sensitivity
analysis in which we did not limit our samples to people that were in both waves, and thus the
sample size expanded to 6,686 for the U.S. and 5,574 for China to examine whether the
associations change.
We then estimate a multinomial regression model examining change in kidney functioning
over 4 years where the outcomes are continuously good, continuously bad, improving, worsened,

87
and death. This model is used to determine how change is related to SES factors including
education, insurance types and income, and whether those associations are further explained by
health correlates. Lastly, we observe how the age-controlled probability of death over 4 years is
patterned by SES among those with moderately to severely impaired kidney functioning using a
hazard model. This model also examines factors associated with their death.
4.3 Results
4.3.1 Comparison of sample characteristics between China and the U.S.
Table 4.1 presents the sample characteristics for both countries. Overall, the Chinese
sample is younger and also less likely to have history of adverse health conditions (e.g., history
of diabetes/hypertension, heart problems, BMI) and/or behaviors (e.g., drinking) than the U.S.
sample. There is some exception, in which people in China have more people with uncontrolled
hypertension than U.S., which may indicate lack of effective management. More specifically,
22.1% of the Chinese sample have normal kidney functioning, 52.1% are mildly impaired,
24.5% are moderately impaired and 1.3% are severely impaired. 31.5% of the sample have never
attended school and are illiterate, 43.9% have at least some experiences with primary school and
are literate, and 24.6% are primary school graduates and beyond. Regarding age structure, the
Chinese sample consists of 40.1% between 50-59, 34.9% aged 60-69, 21.2% aged 70-79, and
3.8% aged over 80. A little over half (50.2%) of the sample is female. Regarding insurance
types, 5.4% of the sample do not have any insurance; 13.9% have urban employee-based public
insurance and/or private insurances; 7.7% have urban resident plan and/or other public
supplemental plans; 73% have rural new cooperative. The average household per capita income
is 13,110 RMB. 92.5% of the sample have no history of diabetes; 4.7% have controlled diabetes,

88
meaning that they are aware of their conditions and have the condition controlled, while 2.9%
have uncontrolled diabetes. Over half of the sample (53.6%) have no history of hypertension;
only 10.9% have controlled hypertension, while the rest (35.5%) have uncontrolled hypertension.
14.8% have reported history of cardiovascular problems. In terms of health behaviors, in the
Chinese sample, more than 59.1% of the respondents have never smoked before, 10.4% used to
smoke but have now quit, and 30.5% are still smoking. 58.2% of the sample have never
consumed alcohol; about 9.3% have had alcohol but have quit; and 32.5 are active drinkers.
Lastly, regarding BMI, 59.7% of the Chinese sample have normal weight; 29.3% are considered
overweight, and 11.0% are considered obese.
For the U.S. sample, 22.9% have normal kidney function, 38.9% are mildly impaired,
32.9% are moderately impaired and 5.3% have severely impaired kidney functioning. Regarding
education, 14.4% of the sample did not complete high school, about one-third (32.6%) are high
school graduates, and over half (53.0%) the sample have at least some college experiences.
12.2% of the sample age between 50-59, 45.3% between 60-69, 27.2% between 70-79, and
15.3% over 80. 55.7% of the sample are female. In terms of insurance, 4.1% of the sample do not
have any insurance, about one-third (33.2%) only have public insurance including Medicare and
Medicaid, 29.8% only have private plans, and the other one-third (32.9%) of the sample have
both private and public insurance. The average household per capita income is 74,700 dollars.
About three quarters (75.1 %) of the U.S. sample do not have any history of diabetes, 10.5% and
14.4% have controlled and uncontrolled diabetes, respectively. 30.2% of the sample have never
had hypertension, 38.7% have the condition adequately controlled and 31.2% uncontrolled. Over
one-quarter (25.8%) of the sample have had reported having some cardiovascular diseases.
Finally, for health behaviors, less than half of the sample (45.7%) have never smoked before,

89
43.9% are former smokers and 10.3% are current smokers. 41.3% of the U.S. sample have never
drunk alcohol; 17.3% are former drinkers and 41.4% are active drinkers. About one-fifth
(21.8%) of the sample have normal weight, 36.1% are consider overweight and 42.1% are
considered obese.
By comparing the sample characteristics of the two countries, we learned that our
Chinese sample is much younger than U.S. sample - as CKD is an age-related disease, it is more
meaningful to compare prevalence adjusted for age for the two countries, especially when the
age structures of the two samples are quite different. Figure 4.1 shows that, if the U.S. were to
have the same age structure as China., there would be significantly more older Chinese (53.8%)
who have mildly impaired kidney functioning than that of older Americans (43.1%, p<0.05).
About one-third of the U.S. older adults have normal renal health, which is significantly more
than that of the Chinese sample (23.2%, p<0.05). However, American older adults (2.2%) are
more likely to have severely impaired kidney health than older Chinese (0.84%, p<0.05).
4.3.2 Associations between baseline kidney functioning and SES/related risk factors
Table 4.2 shows associations between risk factors and being in the categories of having
mild, moderate or severe status relative to having normal kidney functioning at baseline. For the
U.S., we found that people in the older age groups are more likely to have kidney problems, and
this age gradient becomes increasingly more apparent as the stages of kidney problems become
more severe. Women in the U.S. are also more likely to have kidney problems than men, and this
sex difference is larger at more severe stages. For China, people in the older age groups are
generally more likely to kidney problems at baseline, and, similar to the U.S., this age gradient
becomes more apparent as the later stages of kidney problems. However, in contrary to the U.S.,

90
women in China are actually less likely to have kidney problems at any stages (mildly impaired:
0.700, p<0.05; moderately impaired: 0.486, p<0.001; severely impaired: 0.317, p<0.01).
Americans with higher education are generally less likely than those with a low level of
education to have moderately impaired (OR: 0.750, p<0.05) kidney function, and this education
differences appear to be explained by presence of hypertension (controlled OR: 1.777, p<0.001;
uncontrolled OR: 1.341, p<0.01), heart problems (OR: 1.548, p<0.001), smoking differences
(OR: 1.882, p<0.001), overweight (1.586, p<0.001) and obesity (2.470, p<0.001). Americans
with higher education (OR: 0.543, p<0.01) or higher income (OR: 0.994, p<0.01) are also less
likely to have severely impaired kidney functioning. These SES differences are explained by
CKD health risk factors, including controlled diabetes (OR: 1.731, p<0.05), presence of
hypertension (controlled OR: 3.880 p<0.001; uncontrolled OR: 2.426, p<0.001) , heart problems
(2.278, p<0.001), being a current smoker 2.216, p<0.01), being a drinker (former drinker OR:
0.623, p<0.05; current drinker OR: 0.367, p<0.001) and being obese (OR: 2.116, p<0.001). We
also found that people with private insurance tend to be less likely to have mildly (OR: 0.807,
p<0.05), moderately (OR: 0.699, p<0.01), and severely impaired kidney functioning (OR: 0.382,
p<0.01) than people with only public insurance (Table S4.2).
Interestingly, for China, we do not observe any SES gradient in kidney functioning at
baseline, although it is related to various risk factors. Current smoking is associated with a higher
likelihood of having mildly impaired kidney functioning (1.353, p<0.05); obesity is associated
with higher likelihood of having moderately impaired kidney functioning (2.325, p<0.01);
controlled hypertension (4.885, p<0.01), and uncontrolled diabetes (4.248, p<0.01) are
associated with a higher likelihood of having severely impaired kidney functioning.

91
Comparing the two countries, we observed that differences by age in baseline kidney
functioning appear greater in China than the U.S. for mildly to moderately impaired kidney
functioning; the sex difference is in the opposite direction for the two countries; women are more
at risk than men in the U.S. whereas men are more at risk in China. Regarding risk factors, it
appears that the presence and management of comorbidities and other risk factors are more
closely related to mild to moderate kidney problems in the U.S. than in China, though
comorbidities also play a more important role in the more severe stage of kidney problem in
China.
We also find that SES differences are observed in the U.S. but not in China. There may
be two explanations for this – 1) there is some evidence that the relationship between SES,
including educational attainment, and health can be quite complicated in developing countries.
People with highest SES do not necessarily have better health behaviors than those with lower
education level; 2) it is possible that, since our sample is particularly selective as we only
included people with both waves of biomarker data. For this concern, we ran a sensitivity
analysis in which we did not limit our samples to people that were in for both waves, and thus
the sample size expanded to 6,686 for the U.S. and 5,574 for China. We now find that people
with higher education level have significantly better kidney functioning in China as well. People
with higher education level are less likely to have all stages of kidney problems. People with
urban employee and/or private insurance are less likely to have mildly impaired kidney
functioning than people with no insurance (0.542, p<0.05) (Table S4.1). Through analyzing
those who were excluded from the analysis, we found that people who drop out by 2015 in China
are actually more educated – while the analytic group (n=4016) is less educated, but according to

92
our data they do necessarily have worse kidney functioning, which may thus explain why the
smaller sample did not show education gradient.
4.3.3 Associations between change in kidney health over 4-year and SES/related risk
factors
In terms of transitions, we compared the age-adjusted probabilities of how people’s
kidney functioning changed over 4-year in both countries. Interestingly, we found that the
probabilities of all types of transitions are similar in both countries (See Table 4.3). According to
the multinomial regression results that examine the likelihood of decline or improvement over 4-
years (Table 4.4), when compared to “continuously good” as the outcome, decline in kidney
functioning over time is related to heart problems (OR: 1.873, p<0.01) and obesity (OR: 1.786,
p<0.01). Mortality is related to age, lower income (OR: 0.995, p<0.05), being female (OR:
0.619, p<0.05), uncontrolled diabetes (OR: 2.129, p<0.001), controlled hypertension (OR: 1.417,
p<0.05), heart problems (OR: 1.459, p<0.05), being a former smoker (OR: 1.613, p<0.01), being
a current smoker (OR:2.063, p<0.01), being a current drinker (OR: 0.671, p<0.05) and
overweight (0.611, p<0.01).
For China, unlike the U.S., there is some education gradient in decline and also transition
to death, in which people with middle level of education are more likely to have worsened
kidney functioning over 4-year (OR: 1.455, p<0.01), while people with highest education level
are less likely to die over 4-year (OR: 0.524, p<0.05). Income does not seem to be related to
transitions in China. There are much clearer age group differences in China. Women in China
seem to be more likely than men to have decline in kidney functioning (OR: 1.307, p<0.001).
People who are obese are more likely to be continuously bad than continuously good (OR: 2.152,

93
p<0.01). People who used to drink have higher chance of improvement over 4-year (OR: 1.706,
p<0.05). Worsened kidney health among Chinese older adults is also associated with
uncontrolled diabetes (OR: 1.750, p<0.05), uncontrolled hypertension (OR: 1.483, p<0.01), and
obesity (OR: 1.467, p<0.05). Finally, older Chinese are more likely to die after 4-year if they had
at baseline uncontrolled diabetes (OR: 4.827, p<0.01), uncontrolled hypertension (OR: 1.580,
p<0.05), or were former smokers (OR: 2.341, p<0.01).
Finally, the Cox proportional-hazards model showed that, among those with moderately
to severely impaired kidney function at baseline, women are less likely to die within 4-years than
men in the U.S. but not in China. The SES differences in mortailty were not significant in the
U.S. However, in China, people with no insurance or with rural insurance are more likely to die
(OR: 2.837, p<0.05; OR: 2.727, p<0.05) over the 4-year period compared to those with urban
employee/private insurance,even after comorbidities and health behaviors are controlled. Among
those with kidney problems, in both countries, the likelihood of death within 4-years is higher
among those that were fomrer smokers. People with uncontrolled diabetes or who were
overweight/obese are more liekly to die in the U.S.; people with uncontrolled hypertension are
more likely to die in China.
4.4 Discussion
This is the first cross-national comparison study that utilized longitudinal biomakrer data
to directly assess and compare kidney health profiles between China and the U.S and their
associations with risk factors and links to SES. Our results suggest that, overall, one’s SES plays
a critical role in renal health in both China and the U.S. and is associated with people’s baseline
kidney functioning for both U.S. and China but only for change over time for China.

94
Specifically, at baseline, education and having access to good quality insurance both seem to be
protective against kidney problems in both countries, though income level is only associated with
kidney impairment at the more severe stage in the U.S. Over time, older adults in the U.S. are
more likely to improve their kidney functioning with better access to good quality insurance and
higher income, whereas in China people are less likely to experience kidney decline if they had
high level of education. By comparing the two countries, we found that the education gap in the
U.S. is somewhat explained by better management of comorbidities and more behaviors that are
protective of renal health - however, this education gap cannot be entirely explained by known
risk factors that are typically associated with education in China. This suggests that the education
inequality in renal health is not only pronounced in China but also cannot be explained by risk
factors that were considered typical in the U.S., which points to the need for further
understanding in additional underlying mechanisms that link education inequality with kidney
management in the Chinese context. As some researchers have pointed out, China has
experienced more than three decades of unprecedentedly rapid economic growth, significant
urbanization, and medical advancement, the mechanism behind its education disparities in
kidney health may be vastly different from that of most developed countries like the U.S. that
went through economic growth more smoothly (32). Nevertheless, access to good quality
insurance, and effective management of hypertension, diabetes and obesity still play a role in
protecting kidney health in China.
Our results also revealed that insurance, and we assume the care it is linked to, plays a
critical role in renal health in both countries. First, having good quality insurance, such as private
insurance in the U.S., or urban employee/private insurance in China, is somewhat associated
with better baseline kidney functioning and/or better kidney outcomes by the second wave.

95
These findings are in agreement with previous findings, in which people with private insurance
tend to be at lower risk of progressing to more severe stages of CKD as they have better access
to medication, kidney-related procedures and preventive services, management of comorbidities,
etc. (33, 34). We did not observe the same difference between people with public insurance and
uninsured people, possibly due to the fact that public insurance tends to be conditioned by
presence of morbidity and disability, most often associated with poor health status that may have
not been picked up in our model. Importantly, in China, insurance may even be somewhat related
to mortality. Among those with kidney problems, the probability of death appears to be lower
among those that have urban employee insurance, in comparison to people with no insurance or
rural insurance. This finding is consistent with the fact that rural health insurance may generally
be linked to inferior health care coverage, treatment quality, and a lower reimbursement rate,
which may lead to much higher costs due to co-payments or excluded services (35, 36).
Another aspect that makes this international comparison meaningful is that it allows us to
examine and compare the strength of associations between various risk factors and kidney health
across the two countries. First, while age plays a role in baseline kidney functioning and
progression in both countries, it appears that magnitude of age effect is greater in China - one
explanation could be that the age groups reflect a cohort effect – as China has been going
through rapid changes in its epidemiological environment and also people’s lifestyles in the past
decades more so than the U.S., there may be a bigger shift in the kidney function between
cohorts (37). The sex difference at baseline goes opposite directions for the two countries, in
which women are more at risk than men in the U.S. whereas men are more at risk in China in all
stages of CKD. While this finding seems different from previous studies, in which women
typically have higher risk of CKD than men in China (38). For transition, however, women are

96
more likely to have worsened kidney functioning over time in China, and this sex disparity may
be addressed by proper management of diabetes, hypertension and obesity.
We also observed that, though sharing some common risk factors, China and the U.S.
have differentials in impact potentials in the progression of renal health. For example, our results
showed that the presence of cardiovascular disease and obesity play a more important role in
older Americans entering more severe stages of CKD, while uncontrolled hypertension
contributes to older Chinese entering more severe stages of CKD in China within 4-year. These
observations both further solidify our current understanding in which hypertension may have a
stronger association with worse kidney functioning in China than the U.S. (39), and that the
underlying etiology of CKD in China is rapidly becoming diabetes-related, which could be
caused the unhealthy lifestyle and diets that come with rapid economic growth (40, 41). These
findings show that, though sharing some common risk factors, China and the U.S. have
differentials in impact potentials in the progression and management of renal health. Thus, CKDrelated health policies and prevention programs should take into account the country-specific
importance of each risk factor. Overall, it appears that the presence and management of
comorbidities and other risk factors are more closely related to mild to moderate kidney
problems in the U.S. than in China, though comorbidities do also play a more important role in
the more severe stage of kidney problem in China. This may be due to the differences in the
epidemiological environments of the two countries – in which earlier stages of CKD in China
may be less related to comorbidities and behaviors in comparison to the U.S. but more related to
other factors that were not examined in this study, such as environmental toxin, access to clean
water, etc.

97
This study has limitations. First, there is no consensus regarding how to compare
education groups across countries with different levels of economic development such as China
and the U.S. Though we tried our best to match the percentages of people that fall into the
lowest, middle, and highest education group, the percentages of each group are not exactly the
same in two countries. Secondly, there was no information regarding treatment or management
of CKD in HRS or CHARLS, which is an important reason for progression of CKD over 4-years.
Thirdly, although we were interested in fully examining stages of disease in the hazard model,
the sample size of the most severe stage, particularly for China, becomes too small. Lastly,
infection-related kidney problems that are believed to be relatively important in China are not
examined in this study due to lack of data on environmental factors, toxin, herbal use, etc. Future
studies should continue to expand risk factors are unique to developing countries like China to
develop effective preventive measures and effective management over time for CKD.

98
Table 4.1. Baseline Sample Characteristics of China and U.S. in 2011
China (n=4,016) US (n=5,497)
% Mean 95% CI % Mean 95% CI
Kidney functioning
Normal 22.1 [20.8, 23.4] 22.9 [21.8, 24.1]
Mildly impaired 52.1 [50.6, 53.7] 38.9 [37.6, 40.2]
Moderately impaired 24.5 [23.2, 25.9] 32.9 [31.6, 34.1]
Severely impaired 1.3 [1.0, 1.7] 5.3 [4.5, 6.0]
Educational attainment
Lower 31.5 [30.1, 32.9] 14.4 [13.5, 15.4]
Middle 43.9 [42.4, 45.4] 32.6 [31.3, 33.8]
Higher 24.6 [23.3, 25.9] 53 [51.7, 54.3]
Age
50-59 40.1 [40.6, 43.7] 12.2 [11.4, 13.1]
60-69 34.9 [34.4, 37.4] 45.3 [44.0, 46.6]
70-79 21.2 [20.6, 23.3] 27.2 [26.0, 28.4]
80+ 3.8 [3.2, 4.3] 15.3 [14.3, 16.2]
Female 50.2 [48.7, 51.8] 55.7 [54.4, 57.1]
Insurance type
No insurance 5.4 [4.7, 6.1] 4.1 [3.5, 4.6]
Urban employee and/or private 13.9 [12.9, 15.0]
Urban resident and/or other public 7.7 [6.8, 8.5]
Rural cooperative 73 [71.6, 74.4]
Public 33.2 [32.0, 34.5]
Private 29.8 [28.6, 31.1]
Public + private 32.9 [31.6, 34.1]
Income (thousands) 13.1 [12.1, 14.2] 74.7 [72.0, 77.4]
Diabetes
No history 92.5 [91.6, 93.3] 75.1 [73.9, 76.2]
Controlled 4.7 [4.0, 5.3] 10.5 [9.7, 11.4]
Uncontrolled 2.9 [2.4, 3.4] 14.4 [13.4, 15.3]
Hypertension
No history 53.6 [52.0, 55.1] 30.2 [28.9, 31.4]
Controlled 10.9 [10.0, 11.9] 38.7 [37.4, 40.0]
Uncontrolled 35.5 [34.0, 37.0] 31.2 [29.9, 32.4]
Heart problems 14.8 [13.7, 15.9] 25.8 [24.6, 26.9]
Smoking
Never smoked 59.1 [57.6, 60.6] 45.7 [44.4, 47.1]
Former smoker 10.4 [9.4, 11.3] 43.9 [42.6, 45.2]
Current smoker 30.5 [29.1, 31.9] 10.3 [9.5, 11.1]

99
Drinking
Never drunk 58.2 [56.7, 59.7] 41.3 [40.0, 42.7]
Former drinker 9.3 [8.4, 10.2] 17.3 [16.3, 18.3]
Current drinker 32.5 [31.0, 33.9] 41.4 [40.1, 42.7]
BMI
Normal 59.7 [58.2, 61.2] 21.8 [20.7, 22.9]
Overweight 29.3 [27.9, 30.8] 36.1 [34.9, 37.4]
Obese 11 [10.0, 11.9] 42.1 [40.7, 43.4]

100
30% 23.2%
43.1% 53.8%
24.6% 22.1%
2.2% 0.84%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
US China
Figure 4.1 Comparison of Age-adjusted prevalence of
CKD stages
Normal Mildly impaired Moderately impaired Severely impaired

101
Table 4.2. Association between baseline kidney functioning, SES and CKD risk
factors (n = 5,497 for U.S.; n = 4,016 for China) Base
outcome:
Normal
kidney
function
Mildly impaired Moderately impaired Severely impaired
US China US China US China
Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age group (ref=50-59)
60-69 1.163 1.189 2.272*** 2.256*** 1.965*** 1.918*** 6.058*** 6.100*** 5.080* 4.484* 3.18 3.204*
70-79 1.584** 1.705*** 4.006*** 4.168*** 4.267*** 4.901*** 41.392*** 44.41*** 14.363*** 16.17*** 18.683*** 17.972***
80+ 3.398*** 3.841*** 0.765 0.792 17.278*** 22.769*** 41.593*** 46.75*** 93.749*** 121.3*** 76.675*** 81.685***
Female 1.160* 1.155 0.700* 0.768 1.307** 1.345*** 0.486*** 0.493* 1.538** 1.643* 0.317** 0.32
Educational attainment (reference: Lower)
Middle 0.974 1.018 1.044 1.024 0.931 1.069 0.838 0.982 0.951 1.17 0.873 0.844
Higher 0.936 1.039 0.805 0.81 0.750* 0.981 0.699 0.775 0.543** 0.799 0.924 0.684
Income 1 1 1 1 0.999 1 0.998 0.999 0.994** 0.996 1.002 0.996
Insurance (ref=no insurance)
Public 1.039 1.044 1.321 1.258 3.630* 3.782*
Private 0.798 0.843 0.778 0.88 1.034 1.445
Public +
private 0.86 0.891 1.079 1.086 3.077 3.378
Urban
employee
and/or
private
0.604 0.595 0.612 0.586 0.666 0.548
Urban
resident
and/or other
public
0.764 0.727 0.793 0.748 1.707 0.993
Rural
cooperative 0.816 0.799 0.721 0.718 0.542 0.474
Diabetes (ref
= no history)
Controlled 0.838 0.873 1.018 1.178 1.731* 0.721
Uncontrolled 0.756* 0.340** 0.948 0.445* 1.091 4.248**
Hypertension
(ref = no
history)
Controlled 1.123 1.051 1.777*** 1.221 3.880*** 4.885**
Uncontrolled 1.173 0.894 1.341** 1.035 2.426*** 1.738
Heart
Problems 1.177 1.41 1.548*** 1.358 2.278*** 1.371
Smoking (ref
= never
smoked)
Used to
smoke 0.906 1.17 0.958 1.171 1.187 2.41
Current
smoker 1.373* 1.353* 1.882*** 1.426 2.216** 1.576
Drinking (ref
= never
drunk)
Used to
drink 1.096 1.03 0.936 1.291 0.623* 1.194

102
Current
drinker 0.778** 0.967 0.562*** 0.959 0.367*** 0.353*
Obesity (ref
= normal)
Overweight 1.264* 0.897 1.586*** 0.824 1.172 0.984
Obese 1.549*** 1.253 2.470*** 2.325** 2.116*** 0.471
Pseudo R2 0.0911 0.1179 0.1305 0.1460 0.0911 0.1179 0.1305 0.1460 0.0911 0.1179 0.1305 0.1460
Notes: *p<0.05, ** p<0.01, *** p<0.001

103
Table 4.3 Age-adjusted kidney-related health outcomes within 4-year
U.S. % 95% CI China % 95% CI
Continuously good 30.8 [28.1, 33.5] 34.6 [33.1, 36.1]
Continuously bad 14.0 [12.1, 15.8] 14.4 [13.3, 15.4]
Improved 17.0 [14.8, 19.2] 14.9 [13.8, 16.0]
Worsened 30.3 [27.7, 33.0] 28.2 [26.8, 29.6]
Death 7.9 [6.4, 9.3] 7.9 [7.1, 8.7]

104
Table 4.4 Multinomial regression: Associations between change in kidney functioning, SES and CKD risk factors
Base outcome:
Continuously good Continuously bad Improved Worsened Death
US China US China US China US China
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
Age, sex,
SES
risk
factors
GFR at baseline 0.878*** 0.878*** 0.877*** 0.876*** 0.870*** 0.867*** 0.950*** 0.949*** 1.001 1.003 1.015** 1.015** 0.879*** 0.883*** 0.925*** 0.926***
Age (ref = 50-59)
60-69 0.675 0.684 2.246*** 2.313*** 0.735 0.772 0.638** 0.639** 1.316 1.259 1.850*** 1.783*** 1.058 1.075 1.302 1.204
70-79 0.608 0.6 4.504*** 4.757*** 0.545 0.544 0.596* 0.604* 1.664** 1.702** 3.302*** 3.142*** 1.358 1.946 3.960*** 3.508***
80+ 1.136 1.172 17.632*** 19.490*** 0.596 0.6 0.537 0.533 2.970*** 3.181*** 22.298*** 21.316*** 8.410*** 9.067*** 56.174*** 47.040***
Female 0.992 1.004 0.885 0.76 0.871 0.852 0.985 0.997 1.031 1.142 1.307* 1.343 0.569*** 0.619** 0.721 0.943
Education
(ref=lower ed)
Middle ed 1 0.989 1.177 1.138 1.22 1.143 1.315 1.306 0.967 1.03 1.455** 1.411** 0.889 0.91 1.172 1.141
Higher ed 1.166 1.17 1.099 1.163 1.495 1.353 1.094 1.079 0.875 0.984 0.932 0.911 0.943 1.045 0.581* 0.524*
Household per
capita income 0.999 0.999 1.003 1.003 1 1 1.002 1.002 0.999 0.999 0.998 0.998 0.994* 0.995* 1.002 1.002
Insurance (ref = no
insurance)
Public 1.892 1.429 1.045 1.049 1.109 1.059 1.409 1.473
Private 0.932 0.674 0.777 0.767 0.771 0.782 0.512 0.635
Both public and
private 1.72 1.376 0.939 0.939 1.011 0.978 1.25 1.334
Urban employee
and/or private 2.014 1.696 1.269 1.186 2.112 1.877 0.485 0.591
Urban resident and
other public 1.568 1.557 1.655 1.539 1.811 1.635 0.858 0.866
Rural cooperative 1.546 1.546 1.856 1.778 1.168 1.169 0.924 0.981
Diabetes (ref=no
history)
Controlled 0.641 0.774 1.052 0.663 1.238 0.552 1.289 0.975
Uncontrolled 0.491 0.754 0.796 1.767 1.546** 1.750* 2.129*** 4.827**
Hypertension
(ref=no history)
Controlled 1.147 0.992 0.893 0.977 1.106 1.13 1.417* 0.712

105
Uncontrolled 1.225 1.016 0.693* 0.834 1.117 1.483** 1.347 1.580*
Heart problems 1.332 1.202 0.82 1.18 1.873*** 1.431 1.459* 1.197
Smoking
(ref=never
smoked)
Used to smoke 0.877 0.665 0.991 1.097 1.014 1.211 1.613** 2.341**
Current smoker 0.856 0.948 1.048 0.998 1.045 1.171 2.063** 1.406
Drinking (ref=
never drink)
Used to drink 1.276 1.286 0.939 1.706* 0.908 1.055 0.836 1.316
Current drinker 0.822 1.22 1.027 1.309 0.799* 0.822 0.671* 1.087
BMI (ref=normal)
Overweight 1.227 1.272 1.145 0.91 1.353 1.286 0.611** 0.746
Obese 1.449 2.152** 1.067 1.042 1.786** 1.467* 0.613 0.818
Pseudo R2

106
Table 4.5. Cox Proportional-hazards model on mortality among people with moderately
to severely impaired kidney functioning over 4-year (n=1,415 for China; n=2,848 for
U.S.)
Model 1 Model 2
U.S. China U.S. China
Age (ref = 50-59)
60-69 1.834 1.339 3.232 1.213
70-79 2.987 2.611** 5.692* 2.185*
80+ 8.908*** 4.828*** 15.004** 3.661**
Female 0.554*** 0.811 0.641*** 0.964
Education (baseline: lower
education)
Middle education 0.943 1.099 0.898 1.084
Higher education 0.909 0.652 0.938 0.62
Income 0.996* 0.996 0.998 0.997
Insurance
U.S.
Public 1.624 2.233
Private 1.463 2.272
Public + private 1.594 2.302
China (baseline = Urban
employee or private)
No insurance 3.168* 2.837*
Urban resident and/or other public 2.196 1.821
Rural new corporative 3.149** 2.727*
Diabetes (ref = no history)
Controlled 1.092 1.598
Uncontrolled 1.478* 1.606
Hypertension (ref = no history)
Controlled 1.035 1.022
Uncontrolled
0.97 1.698**
Smoking (Baseline = never
smoked)
Former smoker 1.301* 1.788*
Current smoker 1.466 1.212
Drinking (Baseline = never
drank)
Former drinker 0.782 0.933
Current drinker 0.774 0.98
BMI (ref = normal)
Overweight 0.491*** 0.721

107
Obese 0.457*** 0.497
Notes: *p<0.05, ** p<0.01, *** p<0.001

108
4.5 References
1. Bikbov, B., Purcell, C. A., Levey, A. S., Smith, M., Abdoli, A., Abebe, M., ... &
Owolabi, M. O. (2020). Global, regional, and national burden of chronic kidney disease,
1990–2017: a systematic analysis for the Global Burden of Disease Study 2017. The
lancet, 395(10225), 709-733.
2. Foreman, K. J., Marquez, N., Dolgert, A., Fukutaki, K., Fullman, N., McGaughey, M., ...
& Murray, C. J. (2018). Forecasting life expectancy, years of life lost, and all-cause and
cause-specific mortality for 250 causes of death: reference and alternative scenarios for
2016–40 for 195 countries and territories. The Lancet, 392(10159), 2052-2090.
3. Francis, A., Harhay, M. N., Ong, A., Tummalapalli, S. L., Ortiz, A., Fogo, A. B., ... &
Jha, V. (2024). Chronic kidney disease and the global public health agenda: an
international consensus. Nature Reviews Nephrology, 1-13.
4. Mallappallil, M., Friedman, E. A., Delano, B. G., McFarlane, S. I., & Salifu, M. O.
(2014). Chronic kidney disease in the elderly: evaluation and management. Clinical
practice (London, England), 11(5), 525–535.
5. Tonelli, M., Wiebe, N., Culleton, B., House, A., Rabbat, C., Fok, M., McAlister, F., &
Garg, A. X. (2006). Chronic kidney disease and mortality risk: a systematic
review. Journal of the American Society of Nephrology : JASN, 17(7), 2034–2047.
6. Vart, P., Powe, N. R., McCulloch, C. E., Saran, R., Gillespie, B. W., Saydah, S., &
Crews, D. C. (2020). National trends in the prevalence of chronic kidney disease among
racial/ethnic and socioeconomic status groups, 1988-2016. JAMA network open, 3(7),
e207932-e207932.

109
7. Vart, P., Gansevoort, R. T., Joosten, M. M., Bültmann, U., & Reijneveld, S. A. (2015).
Socioeconomic disparities in chronic kidney disease: a systematic review and metaanalysis. American journal of preventive medicine, 48(5), 580-592.
8. Vart, P., Gansevoort, R. T., Crews, D. C., Reijneveld, S. A., & Bültmann, U. (2015).
Mediators of the association between low socioeconomic status and chronic kidney
disease in the United States. American journal of epidemiology, 181(6), 385-396.
9. Patzer, R. E., & McClellan, W. M. (2012). Influence of race, ethnicity and socioeconomic
status on kidney disease. Nature reviews. Nephrology, 8(9), 533–541.
10. Thio, C. H. L., Vart, P., Kieneker, L. M., Snieder, H., Gansevoort, R. T., & Bültmann, U.
(2020). Educational level and risk of chronic kidney disease: longitudinal data from the
PREVEND study. Nephrology, dialysis, transplantation : official publication of the
European Dialysis and Transplant Association - European Renal Association, 35(7),
1211–1218.
11. Morton, R. L., Schlackow, I., Staplin, N., Gray, A., Cass, A., Haynes, R., Emberson, J.,
Herrington, W., Landray, M. J., Baigent, C., Mihaylova, B., & SHARP Collaborative
Group (2016). Impact of Educational Attainment on Health Outcomes in Moderate to
Severe CKD. American journal of kidney diseases : the official journal of the National
Kidney Foundation, 67(1), 31–39.
12. Kovesdy C. P. (2022). Epidemiology of chronic kidney disease: an update 2022. Kidney
international supplements, 12(1), 7–11.
13. GBD Chronic Kidney Disease Collaboration (2020). Global, regional, and national
burden of chronic kidney disease, 1990-2017: a systematic analysis for the Global
Burden of Disease Study 2017. Lancet (London, England), 395(10225), 709–733.

110
14. Chen, T. K., Knicely, D. H., & Grams, M. E. (2019). Chronic Kidney Disease Diagnosis
and Management: A Review. JAMA, 322(13), 1294–1304.
15. Jha, V., Garcia-Garcia, G., Iseki, K., Li, Z., Naicker, S., Plattner, B., ... & Yang, C. W.
(2013). Chronic kidney disease: global dimension and perspectives. The
Lancet, 382(9888), 260-272.
16. Chang, T. I., Lim, H., Park, C. H., Rhee, C. M., Kalantar-Zadeh, K., Kang, E. W., Kang,
S. W., & Han, S. H. (2020). Association Between Income Disparities and Risk of
Chronic Kidney Disease: A Nationwide Cohort Study of Seven Million Adults in
Korea. Mayo Clinic proceedings, 95(2), 231–242.
17. Shen, Q., Jin, W., Ji, S., Chen, X., Zhao, X., & Behera, T. R. (2019). The association
between socioeconomic status and prevalence of chronic kidney disease: A crosssectional study among rural residents in eastern China. Medicine, 98(11), e14822.
18. Liu, Z. H. (2013). Nephrology in china. Nature Reviews Nephrology, 9(9), 523-528.
19. Liu K. (2016). Insuring against health shocks: Health insurance and household
choices. Journal of health economics, 46, 16–32.
20. Meng, Y., Zhang, X., & Han, J. (2020). The impact of medical insurance programmes on
the health of the insured: Evidence from China. The International journal of health
planning and management, 35(3), 718–734. https://doi.org/10.1002/hpm.2931
21. Mendu, M. L., & Weiner, D. E. (2020). Health policy and kidney care in the United
States: core curriculum 2020. American Journal of Kidney Diseases, 76(5), 720-730.
22. Jerant, A., Fiscella, K., & Franks, P. (2012). Health characteristics associated with
gaining and losing private and public health insurance: a national study. Medical
care, 50(2), 145-151.

111
23. Johansen, K. L., Chertow, G. M., Foley, R. N., Gilbertson, D. T., Herzog, C. A., Ishani,
A., ... & Wetmore, J. B. (2021). US renal data system 2020 annual data report:
epidemiology of kidney disease in the United States. American journal of kidney
diseases, 77(4), A7-A8.
24. Zhao, Y., Hu, Y., Smith, J. P., Strauss, J., & Yang, G. (2014). Cohort profile: the China
health and retirement longitudinal study (CHARLS). International journal of
epidemiology, 43(1), 61-68.
25. Chen, X., Crimmins, E., Hu, P., Kim, J. K., Meng, Q., Strauss, J., ... & Zhao, Y. (2019).
Venous blood-based biomarkers in the China health and retirement longitudinal study:
rationale, design, and results from the 2015 wave. American Journal of
Epidemiology, 188(11), 1871-1877.
26. Murty, M. S., Sharma, U. K., Pandey, V. B., & Kankare, S. B. (2013). Serum cystatin C
as a marker of renal function in detection of early acute kidney injury. Indian journal of
nephrology, 23(3), 180–183.
27. Kim, J. K., Faul, J., Weir, D. R., & Crimmins, E. M. (2024). Dried blood spot based
biomarkers in the Health and Retirement Study: 2006 to 2016. American journal of
human biology : the official journal of the Human Biology Council, 36(2), e23997.
28. Shlipak, M. G., Mattes, M. D., & Peralta, C. A. (2013). Update on cystatin C:
incorporation into clinical practice. American Journal of Kidney Diseases, 62(3), 595-
603.
29. Crimmins, E. M., Zhang, Y. S., Kim, J. K., Frochen, S., Kang, H., Shim, H., ... & Faul, J.
(2020). Dried blood spots: Effects of less than optimal collection, shipping time, heat,
and humidity. American Journal of Human Biology, 32(5), e23390.

112
30. Dharnidharka, V. R., Kwon, C., & Stevens, G. (2002). Serum cystatin C is superior to
serum creatinine as a marker of kidney function: a meta-analysis. American journal of
kidney diseases : the official journal of the National Kidney Foundation, 40(2), 221–226.
https://doi.org/10.1053/ajkd.2002.34487
31. Inker, L. A., Eneanya, N. D., Coresh, J., Tighiouart, H., Wang, D., Sang, Y., ... & Levey,
A. S. (2021). New creatinine-and cystatin C–based equations to estimate GFR without
race. New England Journal of Medicine, 385(19), 1737-1749.
32. Yang, G., Wang, Y., Zeng, Y., Gao, G. F., Liang, X., Zhou, M., ... & Murray, C. J.
(2013). Rapid health transition in China, 1990–2010: findings from the Global Burden of
Disease Study 2010. The lancet, 381(9882), 1987-2015.
33. Jurkovitz, C. T., Li, S., Norris, K. C., Saab, G., Bomback, A. S., Whaley-Connell, A.
T., ... & Keep Investigators. (2013). Association between lack of health insurance and
risk of death and ESRD: results from the Kidney Early Evaluation Program
(KEEP). American journal of kidney diseases, 61(4), S24-S32.
34. Fu, X. Z. (2021). Financial protection effects of private health insurance: experimental
evidence from Chinese households with resident basic medical insurance. International
journal for equity in health, 20(1), 122.
35. Fu, D., Liu, L., Zhang, X., Yu, C., Luo, H., & Li, N. (2021). The relationship between
urban and rural health insurance and the self-rated health of migrant workers in
Southwest China. BMC Health Services Research, 21, 1-13.
36. Yang, D., Acharya, Y., & Liu, X. (2022). Social health insurance consolidation and
urban-rural inequality in utilization and financial risk protection in China. Social Science
& Medicine, 308, 115200.

113
37. Yang, C., Wang, H., Zhao, X., Matsushita, K., Coresh, J., Zhang, L., & Zhao, M. H.
(2020). CKD in China: evolving spectrum and public health implications. American
Journal of Kidney Diseases, 76(2), 258-264.
38. Hockham, C., Bao, L., Tiku, A., Badve, S. V., Bello, A. K., Jardine, M. J., ... & Jun, M.
(2022). Sex differences in chronic kidney disease prevalence in Asia: a systematic review
and meta-analysis. Clinical Kidney Journal, 15(6), 1144-1151.
39. Wang, F., He, K., Wang, J., Zhao, M. H., Li, Y., Zhang, L., ... & Bragg-Gresham, J. L.
(2018). Prevalence and risk factors for CKD: a comparison between the adult populations
in China and the United States. Kidney international reports, 3(5), 1135-1143.
40. Du, S., Mroz, T. A., Zhai, F., & Popkin, B. M. (2004). Rapid income growth adversely
affects diet quality in China—particularly for the poor!. Social science &
medicine, 59(7), 1505-1515.
41. Li, F. (2016). Physical activity and health in the presence of China's economic growth:
meeting the public health challenges of the aging population. Journal of sport and health
science, 5(3), 258-269.

114
5. Conclusions and Outlook
Kidney disease is a global public health problem that disproportionally affect vulnerable
and marginalized population, resulting in poor quality of life, financial burden and even elevated
mortality. While there has been major advancement in this field of kidney research, in comparison
to other major non-communicable diseases, awareness for CKD remained low - while it is
estimated that approximately 700 million people have CKD worldwide, this number is likely an
underestimate, owing to a severe lack of early kidney disease detection and screening programs in
both developed and developing countries (1-4). As data used in previous literature largely relied
on clinical diagnoses, national registry, and/or self-reported kidney data, researchers have pointed
out that there may be significant reporting bias, which limited our understanding in the
sociodemographic disparities and mechanisms behind those differentials in CKD at a populationlevel (5-9). The three chapters in this dissertation are examples of how implementation of
biomarkers in population-level studies can deepen our understanding in the sociodemographic
patterns in CKD, which in turn led us to multiple major takeaways that have significant public
health implications.
First, being able to observe the level change and trajectory of kidney health indicated by
biomarkers and its associations with risk factors provides us with useful insights that have public
health implications – first, it helps us identify the effective measures that slow down decline in
kidney functioning. This is significant because there has been an historical argument against CKD
screening due to the belief that there is a lack of effective measures to slow disease progression
(4), this dissertation provided evidence that disease management and lifestyles changes can indeed
slow progression of kidney decline even in older adulthood, which in turn helps address the
sociodemographic disparities in CKD-related morbidity and mortality in both U.S. and China.

115
Specifically, successful management of conditions including diabetes, hypertension,
cardiovascular disease, and obesity remain effective in slowing decline in kidney functioning even
during older adulthood. Avoidance of smoking may be useful in slowing kidney decline.
Race/ethnicity-specific interventions, in which control of cardiovascular disease seems to be
particularly useful for non-Hispanic Black participants in slowing progression, whereas effective
control of diabetes is particularly beneficial for Hispanic individuals may also be worth
considering. This dissertation also highlighted that while China and U.S. shared some common
ground regarding effective ways to slow progression in CKD (e.g., effective management of
diabetes), some risk factors play more important roles in one country than the other (e.g., presence
of cardiovascular disease is more closely related to kidney functioning in the U.S. than China,
whereas hypertension is the opposite). We also observed that, for China, the SES disparities in
kidney problems are not entirely explained by risk factors that were considered typical in the U.S.,
which indicated the need for further understanding in additional underlying mechanisms that link
SES inequality with kidney health in the context of developing societies. As developing countries
are aging at a faster pace than developed countries, there is a pressing need for future research to
explore how risk factors that are more specific to developing countries contribute to the SES
disparities in kidney functioning and the overall progression of kidney decline, such as infection,
environmental toxin, consumption of herbal medicine, etc. In addition to the modifiable risk
factors discussed in this dissertation, research has also shown ways to alleviate progression of
kidney disease through therapeutic agents such as SGLT2 inhibitors, glucagon-like peptide-1
(GLP-1) receptor agonists, endothelin receptor antagonists, selective mineralocorticoid receptor
antagonists, and new glomerulonephritis-targeted therapies, meaning that early recognition of
disease can translate into important health improvements (10-12).

116
Secondly, we realized disparities in kidney functioning may be more nuanced that we
expected after the incorporation of biomarker-based data – while much of what we found is in
fact in agreement with previous knowledge, there are findings that are novel, which seem to add
more detail and nuance to earlier research. For instance, we found that, in our sample, women are
as likely as men to enter end stage renal disease, which is in contrary to the previous notion that
being male is a typical risk factor for faster progression to kidney failure (9). One explanation for
this may be the reliance of many previous studies on self-reported data as previous research has
shown women have less access to nephrology care and less awareness of impaired kidney
functioning, even among those with severely impaired kidney function (13, 14). Further,
previous research generally has only found a significant difference between Black and White
individuals in terms of end stage renal disease but not for earlier stages. Our findings however
did find a significant disadvantage among Black participants for early-stage CKD, but only when
the indicators do not adjust for race. This association may be gone or even flip in direction when
indicators do adjust for race. This finding may partially explain why the conclusion regarding
Black-White disparity in kidney functioning remained unclear in the literature (15), as different
studies have used different equations/indicators. While this dissertation does not aim to
determine which indictors is “best”, it raised awareness regarding the stark differences and/or
sometimes conflicting conclusions about population-level disparities that is particularly evident
at early-stage CKD, which is a critical window for implementing preventive measures to address
the increasing financial burden of late-stage kidney disease and the overall negative impact on
quality of life that disproportionally affect those who are socially disadvantaged. It is also worth
mentioning that our current understanding in the sex disparities of kidney problems remained
binary - more research is needed in the future regarding how best to estimate kidney functioning

117
in transgender, gender-diverse, or nonbinary individuals in which a person’s gender identity is
different from their sex assigned at birth, and the effect of gender hormone-affirming therapy or
puberty-blocking therapies (16-18).
Thirdly, this dissertation emphasizes the key importance of early disease detection and a
life course approach in reducing sociodemographic disparities in CKD. Our findings highlighted
that gap in kidney functioning across racial and SES groups are already evident at baseline age,
which is prior to the age of 60 in our sample. This suggests the attempt to alleviate disparities
among older adults may be too late. As CKD is often silent and patients may be asymptomatic at
early stages, which results in large-scale unawareness of the burden and prevalence of earlier
stages of CKD (19, 20). This lack of awareness resulted in delayed diagnoses and enormous
socioeconomic burdens for individuals, families, society, and the health care system across the
globe (21). This dissertation provided evidence that supports the Kidney Disease Improving Global
Outcomes global multidisciplinary expert panel’s recommendation of early detection in
individuals with known risk factors such as diabetes, hypertension and cardiovascular disease
before CKD symptoms arise (11). Nonetheless, as noted above, developing countries may face
additional risk factors that should be studied and documented at a deeper level (4, 22).
Finally, the findings from this dissertation calls for greater awareness and education
campaigns to help people understand the importance of early diagnosis and management of CKD.
Failure to identify people at a high risk of kidney disease development and/or progression is a
missed opportunity to intervene and prevent kidney failure and its stratospheric health, economic
and psychosocial costs, particularly in developing countries in which the majority of population
with CKD is not aware of their condition, and many more are not aware of the steps that they can
take to slow its progression. Raised awareness of CKD, in combination with recommended

118
treatment options, could have a huge impact on the future health of the world’s population, the
health care system, the economy and people’s overall happiness. We must act now.

119
5.1 References
1. Centers for Disease Control and Prevention. (2021). Chronic kidney disease in the United
States, 2021. Atlanta, GA: US Department of Health and Human Services, Centers for
Disease Control and Prevention.
2. Lewington, A. J., Cerdá, J., & Mehta, R. L. (2013). Raising awareness of acute kidney
injury: a global perspective of a silent killer. Kidney international, 84(3), 457-467.
3. Tummalapalli, S. L., Shlipak, M. G., Damster, S., Jha, V., Malik, C., Levin, A., ... &
Bello, A. K. (2021). Availability and affordability of kidney health laboratory tests
around the globe. American journal of nephrology, 51(12), 959-965.
4. Francis, A., Harhay, M. N., Ong, A., Tummalapalli, S. L., Ortiz, A., Fogo, A. B., ... &
Jha, V. (2024). Chronic kidney disease and the global public health agenda: an
international consensus. Nature Reviews Nephrology, 1-13.
5. Althubaiti A. (2016). Information bias in health research: definition, pitfalls, and
adjustment methods. Journal of multidisciplinary healthcare, 9, 211–217.
https://doi.org/10.2147/JMDH.S104807
6. Norton, J. M., Moxey-Mims, M. M., Eggers, P. W., Narva, A. S., Star, R. A., Kimmel, P.
L., & Rodgers, G. P. (2016). Social determinants of racial disparities in CKD. Journal of
the American Society of Nephrology, 27(9), 2576-2595.
7. Brück, K., Jager, K. J., Dounousi, E., Kainz, A., Nitsch, D., Ärnlöv, J., ... & StojcevaTaneva, O. (2015). Methodology used in studies reporting chronic kidney disease
prevalence: a systematic literature review. Nephrology Dialysis
Transplantation, 30(suppl_4), iv6-iv16.

120
8. Vart, P., Gansevoort, R. T., Joosten, M. M., Bültmann, U., & Reijneveld, S. A. (2015).
Socioeconomic disparities in chronic kidney disease: a systematic review and metaanalysis. American journal of preventive medicine, 48(5), 580-592.
9. Carrero, J. J., Hecking, M., Chesnaye, N. C., & Jager, K. J. (2018). Sex and gender
disparities in the epidemiology and outcomes of chronic kidney disease. Nature Reviews
Nephrology, 14(3), 151-164.
10. Vanholder, R., Annemans, L., Bello, A. K., Bikbov, B., Gallego, D., Gansevoort, R.
T., ... & Wieringa, F. (2021). Fighting the unbearable lightness of neglecting kidney
health: the decade of the kidney. Clinical kidney journal, 14(7), 1719-1730.
11. Shlipak, M. G., Tummalapalli, S. L., Boulware, L. E., Grams, M. E., Ix, J. H., Jha, V., ...
& Zomer, E. (2021). The case for early identification and intervention of chronic kidney
disease: conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO)
Controversies Conference. Kidney international, 99(1), 34-47.
12. Li, P. K. T., Garcia-Garcia, G., Lui, S. F., Andreoli, S., Fung, W. W. S., Hradsky, A., ...
& Kalantar-Zadeh, K. (2020). Kidney health for everyone everywhere–from prevention
to detection and equitable access to care. Nephrologia. 40 (2): 133–141.
13. Lewandowski, M. J., Krenn, S., Kurnikowski, A., Bretschneider, P., Sattler, M.,
Schwaiger, E., Antlanger, M., Gauckler, P., Pirklbauer, M., Brunner, M., Horn, S., Zitt,
E., Kirsch, B., Windpessl, M., Wallner, M., Aringer, I., Wiesholzer, M., Hecking, M., &
Hödlmoser, S. (2023). Chronic kidney disease is more prevalent among women but more
men than women are under nephrological care : Analysis from six outpatient clinics in
Austria 2019. Wiener klinische Wochenschrift, 135(3-4), 89–96.

121
14. Tong, A., Evangelidis, N., Kurnikowski, A., Lewandowski, M., Bretschneider, P.,
Oberbauer, R., Baumgart, A., Scholes-Robertson, N., Stamm, T., Carrero, J. J., PecoitsFilho, R., & Hecking, M. (2021). Nephrologists' Perspectives on Gender Disparities in
CKD and Dialysis. Kidney international reports, 7(3), 424–435.
15. Eneanya, N. D., Boulware, L. E., Tsai, J., Bruce, M. A., Ford, C. L., Harris, C., ... &
Norris, K. C. (2022). Health inequities and the inappropriate use of race in
nephrology. Nature Reviews Nephrology, 18(2), 84-94.
16. Ahmed, S. B., Saad, N., & Dumanski, S. M. (2021). Gender and CKD: beyond the
binary. Clinical Journal of the American Society of Nephrology, 16(1), 141-143.
17. Collister, D., Saad, N., Christie, E., & Ahmed, S. (2021). Providing care for transgender
persons with kidney disease: a narrative review. Canadian Journal of Kidney Health and
Disease, 8, 2054358120985379.
18. Chesnaye, N. C., Carrero, J. J., Hecking, M., & Jager, K. J. (2024). Differences in the
epidemiology, management and outcomes of kidney disease in men and women. Nature
Reviews Nephrology, 20(1), 7-20.
19. Whaley-Connell, A., Nistala, R., & Chaudhary, K. (2011). The importance of early
identification of chronic kidney disease. Missouri medicine, 108(1), 25–28.
20. Evans, M., Lewis, R. D., Morgan, A. R., Whyte, M. B., Hanif, W., Bain, S. C., Davies,
S., Dashora, U., Yousef, Z., Patel, D. C., & Strain, W. D. (2022). A Narrative Review of
Chronic Kidney Disease in Clinical Practice: Current Challenges and Future
Perspectives. Advances in therapy, 39(1), 33–43.

122
21. Luyckx, V. A., Tonelli, M., & Stanifer, J. W. (2018). The global burden of kidney disease
and the sustainable development goals. Bulletin of the World Health Organization, 96(6),
414–422D.
22. Nugent, R. A., Fathima, S. F., Feigl, A. B., & Chyung, D. (2011). The burden of chronic
kidney disease on developing nations: a 21st century challenge in global health. Nephron
Clinical Practice, 118(3), c269-c277.

123
References
Adejumo, O. A., Iyawe, I. O., Akinbodewa, A. A., Abolarin, O. S., & Alli, E. O. (2019). Burden,
psychological well-being and quality of life of caregivers of end-stage renal disease patients.
Ghana Medical Journal, 53(3), 190-196.
Afkarian, M., Zelnick, L. R., Hall, Y. N., Heagerty, P. J., Tuttle, K., Weiss, N. S., & De Boer, I.
H. (2016). Clinical manifestations of kidney disease among US adults with diabetes, 1988-2014.
JAMA, 316(6), 602-610.
Agyei-Mensah, S., & de-Graft Aikins, A. (2010). Epidemiological transition and the double
burden of disease in Accra, Ghana. Journal of Urban Health, 87, 879-897.
Åkerblom, A., Helmersson-Karlqvist, J., Flodin, M., & Larsson, A. (2015). Comparison between
Cystatin C- and Creatinine-Estimated Glomerular Filtration Rate in Cardiology Patients.
Cardiorenal Med, 5(4), 289-296.
Albertus, P., Morgenstern, H., Robinson, B., & Saran, R. (2016). Risk of ESRD in the United
States. American Journal of Kidney Diseases, 68(6), 862-872.
doi:https://doi.org/10.1053/j.ajkd.2016.05.030
Althubaiti, A. (2016). Information bias in health research: definition, pitfalls, and adjustment
methods. Journal of Multidisciplinary Healthcare, 9, 211–217.
https://doi.org/10.2147/JMDH.S104807
Barbour, S. J., Schachter, M., Er, L., Djurdjev, O., & Levin, A. (2010). A systematic review of
ethnic differences in the rate of renal progression in CKD patients. Nephrology Dialysis
Transplantation, 25(8), 2422-2430. doi:10.1093/ndt/gfq283
Bentley, J. N., Viswanathan, A., Rosenberg, W. S., & Patil, P. G. (2014). Treatment of medically
refractory cancer pain with a combination of intrathecal neuromodulation and neurosurgical
ablation: case series and literature review. Pain Medicine, 15(9), 1488-1495.
doi:10.1111/pme.12481
Bikbov, B., Purcell, C. A., Levey, A. S., Smith, M., Abdoli, A., Abebe, M., ... & Owolabi, M. O.
(2020). Global, regional, and national burden of chronic kidney disease, 1990–2017: a systematic
analysis for the Global Burden of Disease Study 2017. The Lancet, 395(10225), 709-733.
Bishop, N., & Zhu, J. (2021). Investigating Racial/Ethnic Disparities in the Association Between
AHEI-2010 and Change in Cystatin C in a Representative Sample of Older Americans. Current
Developments in Nutrition, 5, 1015. doi:https://doi.org/10.1093/cdn/nzab053_008

124
Bleyer, A. J., Shemanski, L. R., Burke, G. L., Hansen, K. J., & Appel, R. G. (2000). Tobacco,
hypertension, and vascular disease: risk factors for renal functional decline in an older
population. Kidney International, 57(5), 2072-2079. doi:10.1046/j.1523-1755.2000.00056.x
Bolignano, D., Mattace-Raso, F., Sijbrands, E. J., & Zoccali, C. (2014). The aging kidney
revisited: a systematic review. Ageing Research Reviews, 14, 65-80.
doi:10.1016/j.arr.2014.02.003
Brück, K., Jager, K. J., Dounousi, E., Kainz, A., Nitsch, D., Ärnlöv, J., ... & Stojceva-Taneva, O.
(2015). Methodology used in studies reporting chronic kidney disease prevalence: a systematic
literature review. Nephrology Dialysis Transplantation, 30(suppl_4), iv6-iv16.
Cañadas-Garre, M., Anderson, K., Cappa, R., et al. (2019). Genetic susceptibility to chronic
kidney disease - some more pieces for the heritability puzzle. Frontiers in Genetics, 10, 453.
doi:10.3389/fgene.2019.00453
Carrero, J. J., Hecking, M., Chesnaye, N. C., & Jager, K. J. (2018). Sex and gender disparities in
the epidemiology and outcomes of chronic kidney disease. Nature Reviews Nephrology, 14(3),
151-164.
Centers for Disease Control and Prevention. (2021). Chronic kidney disease in the United States,
2021. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control
and Prevention.
Chang, T. I., Lim, H., Park, C. H., Rhee, C. M., Kalantar-Zadeh, K., Kang, E. W., Kang, S. W.,
& Han, S. H. (2020). Association Between Income Disparities and Risk of Chronic Kidney
Disease: A Nationwide Cohort Study of Seven Million Adults in Korea. Mayo Clinic
Proceedings, 95(2), 231–242.
Chen, G., Bridenbaugh, E. A., Akintola, A. D., et al. (2007). Increased susceptibility of aging
kidney to ischemic injury: identification of candidate genes changed during aging, but corrected
by caloric restriction. American Journal of Physiology-Renal Physiology, 293(4), F1272-F1281.
doi:10.1152/ajprenal.00138.2007
Chen, T. K., Knicely, D. H., & Grams, M. E. (2019). Chronic Kidney Disease Diagnosis and
Management: A Review. JAMA, 322(13), 1294–1304.
Chen, X., Crimmins, E., Hu, P., Kim, J. K., Meng, Q., Strauss, J., ... & Zhao, Y. (2019). Venous
blood-based biomarkers in the China health and retirement longitudinal study: rationale, design,
and results from the 2015 wave. American Journal of Epidemiology, 188(11), 1871-1877.
Chesnaye, N. C., Carrero, J. J., Hecking, M., & Jager, K. J. (2024). Differences in the
epidemiology, management, and outcomes of kidney disease in men and women. Nature
Reviews Nephrology, 20(1), 7-20.

125
Chesnaye, N. C., Dekker, F. W., Evans, M., Caskey, F. J., Torino, C., Postorino, M., Szymczak,
M., Ramspek, C. L., Drechsler, C., Wanner, C., & Jager, K. J. (2021). Renal function decline in
older men and women with advanced chronic kidney disease-results from the EQUAL study.
Nephrology, Dialysis, Transplantation: official publication of the European Dialysis and
Transplant Association - European Renal Association, 36(9), 1656–1663.
Choudhury, D., & Levi, M. (2011). Kidney aging--inevitable or preventable? Nature Reviews
Nephrology, 7(12), 706-717. doi:10.1038/nrneph.2011.104
Chu, C. D., Chen, M. H., McCulloch, C. E., Powe, N. R., Estrella, M. M., Shlipak, M. G., &
Tuot, D. S. (2021). Patient Awareness of CKD: A Systematic Review and Meta-analysis of
Patient-Oriented Questions and Study Setting. Kidney Medicine, 3(4), 576–585.e1.
Cobb, R. J., Thorpe, R. J., & Norris, K. C. (2020). Everyday discrimination and kidney function
among older adults: Evidence from the Health and Retirement Study. Journal of Gerontology:
Series A, Biological Sciences and Medical Sciences, 75(3), 517-521. doi:10.1093/gerona/glz294
Cobo, G., Hecking, M., Port, F. K., et al. (2016). Sex and gender differences in chronic kidney
disease: progression to end-stage renal disease and haemodialysis. Clinical Science, 130(14),
1147-1163. doi:https://doi.org/10.1042/cs20160047
Cohn, S. H., Brennan, B. L., Yasumura, S., Vartsky, D., Vaswani, A. N., & Ellis, K. J. (1983).
Evaluation of body composition and nitrogen content of renal patients on chronic dialysis as
determined by total body neutron activation. American Journal of Clinical Nutrition, 38(1), 52-
58. doi:10.1093/ajcn/38.1.52
Collister, D., Saad, N., Christie, E., & Ahmed, S. (2021). Providing care for transgender persons
with kidney disease: a narrative review. Canadian Journal of Kidney Health and Disease, 8,
2054358120985379.
Coresh, J., Astor, B. C., Greene, T., Eknoyan, G., & Levey, A. S. (2003). Prevalence of chronic
kidney disease and decreased kidney function in the adult US population: Third national health
and nutrition examination survey. American Journal of Kidney Diseases, 41(1), 1-12.
doi:https://doi.org/10.1053/ajkd.2003.50007
Coresh, J., Byrd-Holt, D., Astor, B. C., et al. (2004). Chronic Kidney Disease Awareness,
Prevalence, and Trends among U.S. Adults, 1999 to 2000. Journal of the American Society of
Nephrology, 16(1), 180-188. doi:https://doi.org/10.1681/asn.2004070539
Crimmins, E. M., Alley, D., Reynolds, S. L., Johnston, M., Karlamangla, A., & Seeman, T.
(2005). Changes in biological markers of health: Older Americans in the 1990s. The Journals of
Gerontology Series A: Biological Sciences and Medical Sciences, 60(11), 1409-1413.
Crimmins, E. M., & Seeman, T. E. (2004). Integrating biology into the study of health
disparities. Population and Development Review, 30, 89-107.

126
Crimmins, E. M., Zhang, Y. S., Kim, J. K., Frochen, S., Kang, H., Shim, H., ... & Faul, J. (2020).
Dried blood spots: Effects of less than optimal collection, shipping time, heat, and humidity.
American Journal of Human Biology, 32(5), e23390.
Crimmins, E., Faul, J., Kim, J., & Weir, D. (2015). Documentation of Blood-Based Biomarkers
in the 2014 Health and Retirement Study. Retrieved from
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker%202014_Dec2017.pdf
Crimmins, E., Faul, J., Kim, J., & Weir, D. (2015). HRS Documentation Report Documentation
of Biomarkers in the 2010 and 2012 Health and Retirement Study Report. Retrieved from
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker2010and2012.pdf
Crimmins, E., Faul, J., Kim, J., & Weir, D. (2017). Documentation of Blood-Based Biomarkers
in the 2014 Health and Retirement Study. Retrieved from
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker%202014_Dec2017.pdf
Crimmins, E., Faul, J., Kim, J., & Weir, D. (2020). Documentation of DBS Blood-Based
Biomarkers in the 2016 Health and Retirement Study. Retrieved from
https://hrsdata.isr.umich.edu/sites/default/files/documentation/datadescriptions/HRS%20Data%20Documentation%20for%202016%20DBS%20Release.pdf
Crimmins, E., Faul, J., Thyagarajan, B., & Weir, D. (2017). Venous Blood Collection and Assay
Protocol in the 2016 Health and Retirement Study. hrs.isr.umich.edu. Accessed April 20, 2023.
Retrieved from https://hrs.isr.umich.edu/publications/biblio/9065
Crimmins, E., Kim, J., Vasunilashorn, S., & Alley, D. (2008). Chapter 5 Biomarkers Related To
Aging In Human Populations. Advances in Clinical Chemistry, 161-216.
doi:https://doi.org/10.1016/s0065-2423(08)00405-8
Crimmins, E., Kim, J., Weir, D., & et al. (2013). HRS Documentation Report Documentation of
Biomarkers in the 2006 and 2008 Health and Retirement Study Report. Retrieved from
https://hrs.isr.umich.edu/sites/default/files/biblio/Biomarker2006and2008.pdf
de Rooij, E. N., Meuleman, Y., de Fijter, J. W., Le Cessie, S., Jager, K. J., Chesnaye, N. C.,
Evans, M., ... & Dekker, F. W. (2022). Quality of life before and after the start of dialysis in
older patients. Clinical Journal of the American Society of Nephrology, 17(8), 1159-1167.
de Rooij, E. N., Meuleman, Y., de Fijter, J. W., Jager, K. J., Chesnaye, N. C., Evans, M., ... &
Dekker, F. W. (2022). Symptom burden before and after dialysis initiation in older patients.
Clinical Journal of the American Society of Nephrology, 17(12), 1719-1729.
Denic, A., Glassock, R. J., & Rule, A. D. (2016). Structural and functional changes with the
aging kidney. Advances in Chronic Kidney Disease, 23(1), 19-28. doi:
10.1053/j.ackd.2015.08.004

127
Derose, S. F., Rutkowski, M. P., Crooks, P. W., Shi, J. M., Wang, J. Q., Kalantar-Zadeh, K., ... &
Jacobsen, S. J. (2013). Racial differences in estimated GFR decline, ESRD, and mortality in an
integrated health system. American Journal of Kidney Diseases, 62(2), 236-244.
Desai, N., Lora, C. M., Lash, J. P., & Ricardo, A. C. (2019). CKD and ESRD in US Hispanics.
American Journal of Kidney Diseases, 73(1), 102-111.
doi:https://doi.org/10.1053/j.ajkd.2018.02.354
Dhalla, S., & Kopec, J. A. (2007). The CAGE questionnaire for alcohol misuse: a review of
reliability and validity studies. Clinical Investigation in Medicine, 30(1), 33-41.
doi:10.25011/cim.v30i1.447
Dharnidharka, V. R., Kwon, C., & Stevens, G. (2002). Serum cystatin C is superior to serum
creatinine as a marker of kidney function: a meta-analysis. American Journal of Kidney
Diseases, 40(2), 221-226. doi:10.1053/ajkd.2002.34487
Diao, J. A., Wu, G. J., Taylor, H. A., Tucker, J. K., Powe, N. R., Kohane, I. S., & Manrai, A. K.
(2021). Clinical implications of removing race from estimates of kidney function. JAMA, 325(2),
184-186. doi:10.1001/jama.2020.22124
Dietrich, L. G., Barceló, C., Thorball, C. W., et al. (2020). Contribution of genetic background
and data collection on adverse events of anti-human immunodeficiency virus (HIV) drugs (D
) clinical risk score to chronic kidney disease in Swiss HIV-infected persons with normal
baseline estimated glomerular filtration rate. Clinical Infectious Diseases, 70(5), 890-897.
doi:10.1093/cid/ciz280
Ene-Iordache, B., Perico, N., Bikbov, B., Carminati, S., Remuzzi, A., Perna, A., ... & Remuzzi,
G. (2016). Chronic kidney disease and cardiovascular risk in six regions of the world (ISNKDDC): a cross-sectional study. The Lancet Global Health, 4(5), e307-e319.
Eneanya, N. D., Boulware, L. E., Tsai, J., Bruce, M. A., Ford, C. L., Harris, C., ... & Norris, K.
C. (2022). Health inequities and the inappropriate use of race in nephrology. Nature Reviews
Nephrology, 18(2), 84-94.
Ephraim, P. L., Powe, N. R., Rabb, H., et al. (2012). The providing resources to enhance African
American patients’ readiness to make decisions about kidney disease (PREPARED) study:
protocol of a randomized controlled trial. BMC Nephrology, 13(1). doi:10.1186/1471-2369-13-
135
Esmeijer, K., Geleijnse, J. M., de Fijter, J. W., Giltay, E. J., Kromhout, D., & Hoogeveen, E. K.
(2018). Cardiovascular risk factors accelerate kidney function decline in post-myocardial
infarction patients: The Alpha Omega Cohort Study. Kidney International Reports, 3(4), 879-
888. doi: 10.1016/j.ekir.2018.03.005

128
Evans, M., Lewis, R. D., Morgan, A. R., Whyte, M. B., Hanif, W., Bain, S. C., Davies, S.,
Dashora, U., Yousef, Z., Patel, D. C., & Strain, W. D. (2022). A Narrative Review of Chronic
Kidney Disease in Clinical Practice: Current Challenges and Future Perspectives. Advances in
Therapy, 39(1), 33–43.
Ewing, J. A. (1984). Detecting alcoholism. The CAGE questionnaire. JAMA, 252(14), 1905-
1907. doi:10.1001/jama.252.14.1905
Fishbane, S., & Nissenson, A. R. (2010). Anemia management in chronic kidney disease. Kidney
International Supplements, 117, S3-S9. doi:10.1038/ki.2010.188
Fletcher, B. R., Damery, S., Aiyegbusi, O. L., Anderson, N., Calvert, M., Cockwell, P., ... &
Kyte, D. (2022). Symptom burden and health-related quality of life in chronic kidney disease: A
global systematic review and meta-analysis. PLoS Medicine, 19(4), e1003954.
Florkowski, C. M., & Chew-Harris, J. S. (2011). Methods of estimating GFR–different equations
including CKD-EPI. The Clinical Biochemist Reviews, 32(2), 75-79.
Foreman, K. J., Marquez, N., Dolgert, A., Fukutaki, K., Fullman, N., McGaughey, M., ... &
Murray, C. J. (2018). Forecasting life expectancy, years of life lost, and all-cause and causespecific mortality for 250 causes of death: reference and alternative scenarios for 2016–40 for
195 countries and territories. The Lancet, 392(10159), 2052-2090.
Francis, A., Harhay, M. N., Ong, A., Tummalapalli, S. L., Ortiz, A., Fogo, A. B., ... & Jha, V.
(2024). Chronic kidney disease and the global public health agenda: an international consensus.
Nature Reviews Nephrology, 1-13.
Friedman, D. J., & Pollak, M. R. (2020). APOL1 and kidney disease: From genetics to biology.
Annual Review of Physiology, 82(1), 323-342. doi:10.1146/annurev-physiol-021119-034345
Fu, D., Liu, L., Zhang, X., Yu, C., Luo, H., & Li, N. (2021). The relationship between urban and
rural health insurance and the self-rated health of migrant workers in Southwest China. BMC
Health Services Research, 21, 1-13.
Fu, X. Z. (2021). Financial protection effects of private health insurance: experimental evidence
from Chinese households with resident basic medical insurance. International Journal for Equity
in Health, 20(1), 122.
Garcia-Garcia, G., & Jha, V. (2015). Chronic kidney disease in disadvantaged populations.
Indian Journal of Nephrology, 25(2), 65–69.
GBD Chronic Kidney Disease Collaboration (2020). Global, regional, and national burden of
chronic kidney disease, 1990-2017: a systematic analysis for the Global Burden of Disease Study
2017. The Lancet, 395(10225), 709–733.

129
Glei, D. A., Goldman, N., Rodríguez, G., & Weinstein, M. (2014). Beyond Self-Reports:
Changes in Biomarkers as Predictors of Mortality. Population and Development Review, 40(2),
331–360.
González, K. A., Stickel, A. M., Kaur, S. S., Ramos, A. R., González, H. M., & Tarraf, W.
(2022). Serum Cystatin-C is linked to increased prevalence of diabetes and higher risk of
mortality in diverse middle-aged and older adults. PLoS One, 17(9), e0270289. doi:
10.1371/journal.pone.0270289
Gummidi, B., John, O., Ghosh, A., Modi, G. K., Sehgal, M., Kalra, O. P., ... & Jha, V. (2020). A
systematic study of the prevalence and risk factors of CKD in Uddanam, India. Kidney
International Reports, 5(12), 2246-2255.
Hansrivijit, P., Chen, Y. J., Lnu, K., et al. (2021). Prediction of mortality among patients with
chronic kidney disease: A systematic review. World Journal of Nephrology, 10(4), 59-75.
doi:10.5527/wjn.v10.i4.59
Harsha, D. W., Frerichs, R. R., & Berenson, G. S. (1978). Densitometry and anthropometry of
black and white children. Human Biology, 50(3), 261-280.
Hecking, M., Hödlmoser, S., Ahmed, S. B., & Carrero, J. J. (2022, March). The other way
around: living with chronic kidney disease from the perspective of men. In Seminars in
Nephrology (Vol. 42, No. 2, pp. 122-128). WB Saunders.
Helmersson‐Karlqvist, J., Miklós Lipcsey, Johan Ärnlöv, et al. (2021). Addition of cystatin C
predicts cardiovascular death better than creatinine in intensive care. Heart, 108(4), 279-284.
doi:https://doi.org/10.1136/heartjnl-2020-318860
Hockham, C., Bao, L., Tiku, A., Badve, S. V., Bello, A. K., Jardine, M. J., ... & Jun, M. (2022).
Sex differences in chronic kidney disease prevalence in Asia: a systematic review and metaanalysis. Clinical Kidney Journal, 15(6), 1144-1151.
Howard, K., White, S., Salkeld, G., et al. (2010). Cost-effectiveness of screening and optimal
management for diabetes, hypertension, and chronic kidney disease: a modeled analysis. Value
in Health, 13(2), 196-208. doi:10.1111/j.1524-4733.2009.00668.x
Inker, L. A., Eneanya, N. D., Coresh, J., et al. (2021). New creatinine- and cystatin C-based
equations to estimate GFR without race. New England Journal of Medicine, 385(19), 1737-1749.
doi:10.1056/NEJMoa2102953
Jager, K. J., Kovesdy, C., Langham, R., Rosenberg, M., Jha, V., & Zoccali, C. (2019). A single
number for advocacy and communication—worldwide more than 850 million individuals have
kidney diseases. Nephrology Dialysis Transplantation, 34(11), 1803-1805.

130
Jain, A., Sutton, K., Roxburgh, S., et al. (2019). A single centre review of end-of-life discussions
and location of death in a dialysis population between 2016 and 2017. Kidney International
Reports, 4(7), S195. doi:https://doi.org/10.1016/j.ekir.2019.05.491
James, S. L., Abate, D., Abate, K. H., Abay, S. M., Abbafati, C., Abbasi, N., ... & Briggs, A. M.
(2018). Global, regional, and national incidence, prevalence, and years lived with disability for
354 diseases and injuries for 195 countries and territories, 1990–2017: a systematic analysis for
the Global Burden of Disease Study 2017. The Lancet, 392(10159), 1789-1858.
Jerant, A., Fiscella, K., & Franks, P. (2012). Health characteristics associated with gaining and
losing private and public health insurance: a national study. Medical Care, 50(2), 145-151.
Jha, V., Garcia-Garcia, G., Iseki, K., Li, Z., Naicker, S., Plattner, B., ... & Yang, C. W. (2013).
Chronic kidney disease: global dimension and perspectives. The Lancet, 382(9888), 260-272.
Johansen, K. L., Chertow, G. M., Foley, R. N., Gilbertson, D. T., Herzog, C. A., Ishani, A., ... &
Wetmore, J. B. (2021). US renal data system 2020 annual data report: epidemiology of kidney
disease in the United States. American Journal of Kidney Diseases, 77(4), A7-A8.
Jones, C., McQuillan, G., Kusek, J., et al. (1998). Serum creatinine levels in the US population:
Third National Health and Nutrition Examination Survey. American Journal of Kidney Diseases,
32(6), 992-999. doi:https://doi.org/10.1016/s0272-6386(98)70074-5
Jurkovitz, C. T., Li, S., Norris, K. C., Saab, G., Bomback, A. S., Whaley-Connell, A. T., ... &
Keep Investigators. (2013). Association between lack of health insurance and risk of death and
ESRD: results from the Kidney Early Evaluation Program (KEEP). American Journal of Kidney
Diseases, 61(4), S24-S32.
Kalantar-Zadeh, K., Lockwood, M. B., Rhee, C. M., Tantisattamo, E., Andreoli, S., Balducci,
A., ... & Li, P. K. T. (2022). Patient-centred approaches for the management of unpleasant
symptoms in kidney disease. Nature Reviews Nephrology, 18(3), 185-198.
Kashani, K., Rosner, M. H., & Ostermann, M. (2020). Creatinine: from physiology to clinical
application. European Journal of Internal Medicine, 72, 9-14.
Kevin, T. J. (2021). Social Justice as a Tool to Eliminate Inequities in Kidney Disease. Seminars
in Nephrology, 41(3), 203-210. doi:https://doi.org/10.1016/j.semnephrol.2021.05.001
Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group (2024). KDIGO
2024 Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney
Disease. Kidney International, 105(4S), S117–S314.
Kim, J. K., Faul, J., Weir, D. R., & Crimmins, E. M. (2024). Dried blood spot based biomarkers
in the Health and Retirement Study: 2006 to 2016. American Journal of Human Biology, 36(2),
e23997.

131
Köttgen, A., Selvin, E., Stevens, L. A., Levey, A. S., Van Lente, F., & Coresh, J. (2008). Serum
Cystatin C in the United States: The Third National Health and Nutrition Examination Survey
(NHANES III). American Journal of Kidney Diseases, 51(3), 385-394.
Kovesdy, C. P. (2022). Epidemiology of chronic kidney disease: an update 2022. Kidney
International Supplements, 12(1), 7–11.
Kremers, W. K., Denic, A., Lieske, J. C., et al. (2015). Distinguishing age-related from diseaserelated glomerulosclerosis on kidney biopsy: the Aging Kidney Anatomy study. Nephrology
Dialysis Transplantation, 30(12), 2034-2039. doi:10.1093/ndt/gfv072
Krieger, N., Williams, D. R., & Moss, N. E. (1997). Measuring social class in US public health
research: concepts, methodologies, and guidelines. Annual Review of Public Health, 18, 341–
378. https://doi.org/10.1146/annurev.publhealth.18.1.341
Legrand, H., Werner, K., Christensson, A., Pihlsgård, M., & Elmståhl, S. (2017). Prevalence and
determinants of differences in cystatin C and creatinine-based estimated glomerular filtration rate
in community-dwelling older adults: a cross-sectional study. BMC Nephrology, 18(1).
Levey, A. S., Bosch, J. P., Lewis, J. B., Greene, T., Rogers, N., & Roth, D. (1999). A more
accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction
equation. Modification of Diet in Renal Disease Study Group. Annals of Internal Medicine,
130(6), 461-470. doi:10.7326/0003-4819-130-6-199903160-00002
Levey, A. S., de Jong, P. E., Coresh, J., et al. (2011). The definition, classification, and prognosis
of chronic kidney disease: a KDIGO Controversies Conference report. Kidney International,
80(1), 17-28. doi:https://doi.org/10.1038/ki.2010.483
Levey, A. S., Stevens, L. A., Schmid, C. H., et al. (2009). A New Equation to Estimate
Glomerular Filtration Rate. Annals of Internal Medicine, 150(9), 604-612.
doi:https://doi.org/10.7326/0003-4819-150-9-200905050-00006
Levey, A. S., Titan, S. M., Powe, N. R., Coresh, J., & Inker, L. A. (2020). Kidney disease, race,
and GFR estimation. Clinical Journal of the American Society of Nephrology, 15(8), 1203.
doi:10.2215/CJN.12791019
Lewandowski, M. J., Krenn, S., Kurnikowski, A., Bretschneider, P., Sattler, M., Schwaiger, E.,
Antlanger, M., Gauckler, P., Pirklbauer, M., Brunner, M., Horn, S., Zitt, E., Kirsch, B.,
Windpessl, M., Wallner, M., Hecking, M., & Hödlmoser, S. (2023). Chronic kidney disease is
more prevalent among women but more men than women are under nephrological care: Analysis
from six outpatient clinics in Austria 2019. Wiener klinische Wochenschrift, 135(3-4), 89–96.
Lewington, A. J., Cerdá, J., & Mehta, R. L. (2013). Raising awareness of acute kidney injury: a
global perspective of a silent killer. Kidney International, 84(3), 457-467.

132
Li, F. (2016). Physical activity and health in the presence of China's economic growth: meeting
the public health challenges of the aging population. Journal of Sport and Health Science, 5(3),
258-269.
Li, P. K. T., Garcia-Garcia, G., Lui, S. F., Andreoli, S., Fung, W. W. S., Hradsky, A., ... &
Kalantar-Zadeh, K. (2020). Kidney health for everyone everywhere–from prevention to detection
and equitable access to care. Nephrologia, 40(2), 133–141.
Liu, K. (2016). Insuring against health shocks: Health insurance and household choices. Journal
of Health Economics, 46, 16–32.
Liu, Z. H. (2013). Nephrology in China. Nature Reviews Nephrology, 9(9), 523-528.
Lora, C., Daviglus, M., Kusek, J., et al. (n.d.). Chronic Kidney Disease in United States
Hispanics: A Growing Public Health Problem. Retrieved from
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3587111/pdf/nihms-316436.pdf
Luyckx, V. A., Tonelli, M., & Stanifer, J. W. (2018). The global burden of kidney disease and
the sustainable development goals. Bulletin of the World Health Organization, 96(6), 414–422D.
Mallappallil, M., Friedman, E. A., Delano, B. G., McFarlane, S. I., & Salifu, M. O. (2014).
Chronic kidney disease in the elderly: evaluation and management. Clinical Practice (London,
England), 11(5), 525–535.
Mandrekar, J. N. (2010). Receiver operating characteristic curve in diagnostic test assessment.
Journal of Thoracic Oncology, 5(9), 1315-1316. doi:10.1097/JTO.0b013e3181ec173d
Markides, K. S., & Coreil, J. (1986). The health of Hispanics in the southwestern United States:
an epidemiologic paradox. Public Health Reports, 101(3), 253-265. PMID: 3086917; PMCID:
PMC1477704.
Martins, D., Agodoa, L., & Norris, K. (2012). Kidney disease in disadvantaged populations.
International Journal of Nephrology, 2012, 469265.
Mayne, K. J., Sullivan, M. K., & Lees, J. S. (2023). Sex and gender differences in the
management of chronic kidney disease and hypertension. Journal of Human Hypertension,
37(8), 649–653. https://doi-org.libproxy2.usc.edu/10.1038/s41371-023-00843-9
McNeish, D., & Matta, T. (2017). Differentiating between mixed-effects and latent-curve
approaches to growth modeling. Behavior Research Methods, 50(4), 1398-1414.
doi:10.3758/s13428-017-0976-5
Mendu, M. L., & Weiner, D. E. (2020). Health policy and kidney care in the United States: core
curriculum 2020. American Journal of Kidney Diseases, 76(5), 720-730.

133
Meng, Y., Zhang, X., & Han, J. (2020). The impact of medical insurance programmes on the
health of the insured: Evidence from China. The International Journal of Health Planning and
Management, 35(3), 718–734. https://doi.org/10.1002/hpm.2931
Meyer, T. E., Kovács, S. J., Ehsani, A. A., Klein, S., Holloszy, J. O., & Fontana, L. (2006).
Long-term caloric restriction ameliorates the decline in diastolic function in humans. Journal of
the American College of Cardiology, 47(2), 398-402. doi:10.1016/j.jacc.2005.08.069
Mombelli, C. A., Giordani, M. C., Imperiali, N. C., et al. (2018). Single-Center Long-Term
Follow-Up of Kidney Donors in Argentina (Hospital Italiano de Buenos Aires). Transplantation
Proceedings, 50(2), 418-422. doi:10.1016/j.transproceed.2017.12.046
Morton, R. L., Schlackow, I., Staplin, N., Gray, A., Cass, A., Haynes, R., Emberson, J.,
Herrington, W., Landray, M. J., Baigent, C., Mihaylova, B., & SHARP Collaborative Group
(2016). Impact of Educational Attainment on Health Outcomes in Moderate to Severe CKD.
American Journal of Kidney Diseases, 67(1), 31–39.
Murty, M. S., Sharma, U. K., Pandey, V. B., & Kankare, S. B. (2013). Serum cystatin C as a
marker of renal function in detection of early acute kidney injury. Indian Journal of Nephrology,
23(3), 180–183.
Nicholas, S. B., Kalantar-Zadeh, K., & Norris, K. C. (2015). Socioeconomic disparities in
chronic kidney disease. Advances in Chronic Kidney Disease, 22(1), 6–15.
Nissenson, A. R., Pereira, B. J. G., Collins, A. J., & Steinberg, E. P. (2001). Prevalence and
characteristics of individuals with chronic kidney disease in a large health maintenance
organization. American Journal of Kidney Diseases, 37(6), 1177-1183.
Norris, K. C., & Agodoa, L. Y. (2005). Unraveling the racial disparities associated with kidney
disease. Kidney International, 68(3), 914–924.
Norris, K. C., & Nissenson, A. R. (2008). Race, gender, and socioeconomic disparities in CKD
in the United States. Journal of the American Society of Nephrology, 19(7), 1261-1270.
Norris, K. C., & Nissenson, A. R. (2020). Ethnicity and Chronic Kidney Disease—United States.
In Chronic Renal Disease (pp. 111-119). Academic Press.
Norris, K. C., Mehrotra, R., & Nissenson, A. R. (2008). Racial differences in mortality and
ESRD. American Journal of Kidney Diseases, 52(2), 205–208.
https://doi.org/10.1053/j.ajkd.2008.06.004
Norton, J. M., Moxey-Mims, M. M., Eggers, P. W., Narva, A. S., Star, R. A., Kimmel, P. L., &
Rodgers, G. P. (2016). Social determinants of racial disparities in CKD. Journal of the American
Society of Nephrology, 27(9), 2576-2595.

134
Nugent, R. A., Fathima, S. F., Feigl, A. B., & Chyung, D. (2011). The burden of chronic kidney
disease on developing nations: a 21st century challenge in global health. Nephron Clinical
Practice, 118(3), c269-c277.
Oliver III, J. D., Nee, R., Marneweck, H., Banaag, A., Koyama, A. K., Pavkov, M. E., &
Koehlmoos, T. P. (2024). Impact of Race-Free Glomerular Filtration Rate Estimations on CKD
Prevalence in the US Military Health System: A Retrospective Cohort Study. Kidney Medicine,
100861.
Onopiuk, A., Tokarzewicz, A., & Gorodkiewicz, E. (2015). Cystatin C: a kidney function
biomarker. Advances in Clinical Chemistry, 68, 57-69.
Pampel, F. C., Krueger, P. M., & Denney, J. T. (2010). Socioeconomic Disparities in Health
Behaviors. Annual Review of Sociology, 36, 349–370.
Parekh, R. S., Perl, J., Auguste, B., & Sood, M. M. (2022). Elimination of race in estimates of
kidney function to provide unbiased clinical management in Canada. CMAJ, 194(11), E421-
E423. doi:10.1503/cmaj.210838
Parsa, A., Kao, W. H. L., Xie, D., et al. (2013). APOL1 risk variants, race, and progression of
chronic kidney disease. New England Journal of Medicine, 369(23), 2183-2196.
doi:10.1056/nejmoa1310345
Patzer, R. E., & McClellan, W. M. (2012). Influence of race, ethnicity and socioeconomic status
on kidney disease. Nature Reviews Nephrology, 8(9), 533–541.
Peiris, H., Chandrasena, L. G., & Lanerolle, R. D. (2008). Serum cystatin C as a marker to
identify patients with moderately impaired renal function. Indian Journal of Clinical
Biochemistry, 23(2), 163-166. doi:10.1007/s12291-008-0036-0
Peralta, C. A., Vittinghoff, E., Bansal, N., et al. (2013). Trajectories of kidney function decline in
young black and white adults with preserved GFR: Results from the Coronary Artery Risk
Development in Young Adults (CARDIA) study. American Journal of Kidney Diseases, 62(2),
261-266. doi:10.1053/j.ajkd.2013.01.012
Perico, N., & Remuzzi, G. (2016). Prevention programs for chronic kidney disease in lowincome countries. Internal and Emergency Medicine, 11, 385-389.
Prowle, J. R., Chua, H. R., Bagshaw, S. M., & Bellomo, R. (2012). Clinical review: Volume of
fluid resuscitation and the incidence of acute kidney injury - a systematic review. Critical Care,
16(4), 230. doi:10.1186/cc11345
Puthucheary, Z. A., Rawal, J., McPhail, M., et al. (2013). Acute Skeletal Muscle Wasting in
Critical Illness. JAMA, 310(15), 1591-1592. doi:https://doi.org/10.1001/jama.2013.278481
Ritchie, H., & Roser, M. (2023). Does the news reflect what we die from?. Our World in Data.

135
Romagnani, P., Remuzzi, G., Glassock, R., Levin, A., Jager, K. J., Tonelli, M., ... & Anders, H.
J. (2017). Chronic kidney disease. Nature Reviews Disease Primers, 3(1), 1-24.
Roth, G. A., Abate, D., Abate, K. H., Abay, S. M., Abbafati, C., Abbasi, N., ... & Borschmann,
R. (2018). Global, regional, and national age-sex-specific mortality for 282 causes of death in
195 countries and territories, 1980–2017: a systematic analysis for the Global Burden of Disease
Study 2017. The Lancet, 392(10159), 1736-1788.
Saran, R., Robinson, B., Abbott, K. C., Bragg-Gresham, J., Chen, X., Gipson, D., ... &
Shahinian, V. (2020). US renal data system 2019 annual data report: epidemiology of kidney
disease in the United States. American Journal of Kidney Diseases: the official journal of the
National Kidney Foundation, 75(1 Suppl 1), A6-A7.
Shen, Q., Jin, W., Ji, S., Chen, X., Zhao, X., & Behera, T. R. (2019). The association between
socioeconomic status and prevalence of chronic kidney disease: A cross-sectional study among
rural residents in eastern China. Medicine, 98(11), e14822.
Shlipak, M. G., Katz, R., Kestenbaum, B., et al. (2009). Rate of kidney function decline in older
adults: a comparison using creatinine and cystatin C. American Journal of Nephrology, 30(3),
171-178. doi:10.1159/000212381
Shlipak, M. G., Mattes, M. D., & Peralta, C. A. (2013). Update on cystatin C: incorporation into
clinical practice. American Journal of Kidney Diseases, 62(3), 595-603.
Shlipak, M. G., Sarnak, M. J., Katz, R., et al. (2005). Cystatin C and the risk of death and
cardiovascular events among elderly persons. New England Journal of Medicine, 352(20), 2049-
2060. doi:10.1056/NEJMoa043161
Shlipak, M. G., Tummalapalli, S. L., Boulware, L. E., Grams, M. E., Ix, J. H., Jha, V., ... &
Zomer, E. (2021). The case for early identification and intervention of chronic kidney disease:
conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies
Conference. Kidney International, 99(1), 34-47.
Shukri, M., Mustofai, M. A., Md Yasin, M. A. S., & Tuan Hadi, T. S. (2020). Burden, quality of
life, anxiety, and depressive symptoms among caregivers of hemodialysis patients: The role of
social support. The International Journal of Psychiatry in Medicine, 55(6), 397-407.
Silva, F. G. (2005). The aging kidney: a review--part II. International Urology and Nephrology,
37(2), 419-432. doi:10.1007/s11255-004-0874-5
Smith, H. W. (1951). The kidney: structure and function in health and disease. Oxford University
Press, USA.
Stanifer, J. W., Muiru, A., Jafar, T. H., & Patel, U. D. (2016). Chronic kidney disease in low-and
middle-income countries. Nephrology Dialysis Transplantation, 31(6), 868-874.

136
Stengel, B., Tarver-Carr, M. E., Powe, N. R., Eberhardt, M. S., & Brancati, F. L. (2003).
Lifestyle factors, obesity and the risk of chronic kidney disease. Epidemiology (Cambridge,
Mass.), 14(4), 479–487.
Stephan, Y., Sutin, A. R., & Terracciano, A. (2019). Subjective Age and Cystatin C Among
Older Adults. Journal of Gerontology: Series B, Psychological Sciences and Social Sciences,
74(3), 382-388. doi:10.1093/geronb/gbx124
Stevens, L. A., Schmid, C. H., Greene, T., et al. (2009). Comparative performance of the CKD
Epidemiology Collaboration (CKD-EPI) and the Modification of Diet in Renal Disease (MDRD)
Study equations for estimating GFR levels above 60 mL/min/1.73 m². American Journal of
Kidney Diseases, 56(3), 486-495. doi:10.1053/j.ajkd.2010.03.026
Stevens, L. A., Schmid, C. H., Greene, T., et al. (2009). Factors other than glomerular filtration
rate affect serum cystatin C levels. Kidney International, 75(6), 652-660.
doi:10.1038/ki.2008.638
Tannor, E. K., Chika, O. U., & Okpechi, I. G. (2022). The impact of low socioeconomic status
on progression of chronic kidney disease in low-and lower middle-income countries. In Seminars
in Nephrology (Vol. 42, No. 5, p. 151338). WB Saunders.
Thio, C. H. L., Vart, P., Kieneker, L. M., Snieder, H., Gansevoort, R. T., & Bültmann, U. (2020).
Educational level and risk of chronic kidney disease: longitudinal data from the PREVEND
study. Nephrology Dialysis Transplantation, 35(7), 1211-1218. doi:10.1093/ndt/gfy361
Tidman, M., Sjöström, P., & Jones, I. (2008). A Comparison of GFR estimating formulae based
upon s-cystatin C and s-creatinine and a combination of the two. Nephrology Dialysis
Transplantation, 23(1), 154-160.
Tomlinson, L. A., & Clase, C. M. (2019). Sex and the Incidence and Prevalence of Kidney
Disease. Clinical Journal of the American Society of Nephrology, 14(11), 1557–1559.
Tonelli, M., Wiebe, N., Culleton, B., et al. (2006). Chronic kidney disease and mortality risk: a
systematic review. Journal of the American Society of Nephrology, 17(7), 2034-2047.
doi:10.1681/ASN.2005101085
Tong, A., Evangelidis, N., Kurnikowski, A., Lewandowski, M., Bretschneider, P., Oberbauer, R.,
Baumgart, A., Scholes-Robertson, N., Stamm, T., Carrero, J. J., Pecoits-Filho, R., & Hecking,
M. (2021). Nephrologists' Perspectives on Gender Disparities in CKD and Dialysis. Kidney
International Reports, 7(3), 424–435.
Tsai, J. W., Cerdeña, J. P., Goedel, W. C., et al. (2021). Evaluating the Impact and Rationale of
Race-Specific Estimations of Kidney Function: Estimations from U.S. NHANES, 2015-2018.
EClinicalMedicine, 42, 101197. doi:10.1016/j.eclinm.2021.101197

137
Tsai, W. C., Wu, H. Y., Peng, Y. S., et al. (2016). Risk factors for development and progression
of chronic kidney disease: A systematic review and exploratory meta-analysis. Medicine, 95(11),
e3013. doi:10.1097/MD.0000000000003013
Tummalapalli, S. L., Shlipak, M. G., Damster, S., Jha, V., Malik, C., Levin, A., ... & Bello, A. K.
(2021). Availability and affordability of kidney health laboratory tests around the globe.
American Journal of Nephrology, 51(12), 959-965.
United States Renal Data System. (2018). 2018 USRDS annual data report: epidemiology of
kidney disease in the United States.
US National Institute of Diabetes and Digestive and Kidney Diseases. Kidney disease statistics
for the United States. https://www-niddk-nih-gov.libproxy2.usc.edu/health-information/healthstatistics/kidney-disease (2023).
Vanholder, R., Annemans, L., Bello, A. K., Bikbov, B., Gallego, D., Gansevoort, R. T., ... &
Wieringa, F. (2021). Fighting the unbearable lightness of neglecting kidney health: the decade of
the kidney. Clinical Kidney Journal, 14(7), 1719-1730.
Vart, P., Gansevoort, R. T., Crews, D. C., Reijneveld, S. A., & Bültmann, U. (2015). Mediators
of the association between low socioeconomic status and chronic kidney disease in the United
States. American Journal of Epidemiology, 181(6), 385-396.
Vart, P., Gansevoort, R. T., Joosten, M. M., Bültmann, U., & Reijneveld, S. A. (2015).
Socioeconomic disparities in chronic kidney disease: a systematic review and meta-analysis.
American Journal of Preventive Medicine, 48(5), 580-592.
Vart, P., Powe, N. R., McCulloch, C. E., Saran, R., Gillespie, B. W., Saydah, S., & Crews, D. C.
(2020). National trends in the prevalence of chronic kidney disease among racial/ethnic and
socioeconomic status groups, 1988-2016. JAMA Network Open, 3(7), e207932-e207932.
Wang, F., He, K., Wang, J., Zhao, M. H., Li, Y., Zhang, L., ... & Bragg-Gresham, J. L. (2018).
Prevalence and risk factors for CKD: a comparison between the adult populations in China and
the United States. Kidney International Reports, 3(5), 1135-1143.
Wang, H., Naghavi, M., Allen, C., Barber, R. M., Bhutta, Z. A., Carter, A., ... & Bell, M. L.
(2016). Global, regional, and national life expectancy, all-cause mortality, and cause-specific
mortality for 249 causes of death, 1980–2015: a systematic analysis for the Global Burden of
Disease Study 2015. The Lancet, 388(10053), 1459-1544.
Ware, E., Gard, A., & Schmitz, L. (2020). HRS Polygenic Scores – Release 4. Retrieved from
https://hrsdata.isr.umich.edu/sites/default/files/documentation/datadescriptions/PGNSCORES4_3DD_0.pdf
Weinstein, J. R., & Anderson, S. (2010). The aging kidney: physiological changes. Advances in
Chronic Kidney Disease, 17(4), 302–307.

138
Westbrook, A., Zhang, R., Shi, M., et al. (2022). Association between baseline buccal telomere
length and progression of kidney function: The Health and Retirement Study. Journal of
Gerontology: Series A, Biological Sciences and Medical Sciences, 77(3), 471-476.
doi:10.1093/gerona/glac004
Whaley-Connell, A., Nistala, R., & Chaudhary, K. (2011). The importance of early identification
of chronic kidney disease. Missouri Medicine, 108(1), 25–28.
White, S. L., McGeechan, K., Jones, M., Cass, A., Chadban, S. J., Polkinghorne, K. R., ... &
Roderick, P. J. (2008). Socioeconomic disadvantage and kidney disease in the United States,
Australia, and Thailand. American Journal of Public Health, 98(7), 1306-1313.
Willey, J. Z., Moon, Y. P., Husain, S. A., et al. (2020). Creatinine versus cystatin C for renal
function-based mortality prediction in an elderly cohort: The Northern Manhattan Study.
Shimosawa T, ed. PLOS ONE, 15(1), e0226509.
Worrall, J. G., Phongsathorn, V., Hooper, R. J., & Paice, E. W. (1990). Racial variation in serum
creatine kinase unrelated to lean body mass. British Journal of Rheumatology, 29(5), 371-373.
doi:10.1093/rheumatology/29.5.371
Wuttke, M., Li, Y., Li, M., et al. (2019). A catalog of genetic loci associated with kidney
function from analyses of a million individuals. Nature Genetics, 51(6), 957-972.
doi:10.1038/s41588-019-0407-x
Yan, Q., Zuo, P., Cheng, L., et al. (2020). Acute kidney injury is associated with in-hospital
mortality in older patients with COVID-19. Journal of Gerontology: Series A, Biological
Sciences and Medical Sciences, 76(3), 456-462. doi:10.1093/gerona/glaa181
Yang, C., Wang, H., Zhao, X., Matsushita, K., Coresh, J., Zhang, L., & Zhao, M. H. (2020).
CKD in China: evolving spectrum and public health implications. American Journal of Kidney
Diseases, 76(2), 258-264.
Yang, D., Acharya, Y., & Liu, X. (2022). Social health insurance consolidation and urban-rural
inequality in utilization and financial risk protection in China. Social Science & Medicine, 308,
115200.
Yang, G., Wang, Y., Zeng, Y., Gao, G. F., Liang, X., Zhou, M., ... & Murray, C. J. (2013). Rapid
health transition in China, 1990–2010: findings from the Global Burden of Disease Study 2010.
The Lancet, 381(9882), 1987-2015.
Yearby, R. (2018). Racial disparities in health status and access to healthcare: The continuation
of inequality in the United States due to structural racism. American Journal of Economics and
Sociology, 77(3-4), 1113-1152. doi:10.1111/ajes.12230
Yuan, H. C., Yu, Q. T., Bai, H., Xu, H. Z., Gu, P., & Chen, L. Y. (2021). Alcohol intake and the
risk of chronic kidney disease: results from a systematic review and dose-response meta-

139
analysis. European Journal of Clinical Nutrition, 75(11), 1555-1567. doi:10.1038/s41430-021-
00873-x
Zeng, X., Liu, J., Tao, S., Hong, H. G., Li, Y., & Fu, P. (2018). Associations between
socioeconomic status and chronic kidney disease: a meta-analysis. J Epidemiol Community
Health, 72(4), 270-279.
Zhao, Y., Hu, Y., Smith, J. P., Strauss, J., & Yang, G. (2014). Cohort profile: the China health
and retirement longitudinal study (CHARLS). International Journal of Epidemiology, 43(1), 61-
68.

140
Appendix A
Table S2.1 Effects of Gender, Age Group, Race/Ethnicity, Education, Health Conditions and Behaviors,
and Polygenic Score on Baseline Cystatin C and Annual Cystatin C Change with Eight Years of Age with
the Sample Size for the Polygenic Score Model (n=8,509)
Model 1 Model 2 Model 3 Model 4
Gender, Age Model 1 + Health Conditions Model 2 + Health Behaviors Model 3 + PGS
Intercept P-value Annual
change
P-value Intercept P-value Annual
change
P-value Intercept P-value Annual
growth
P-value Intercept P-value Annual
growth
P-value
Time 0.998 <0.000 0.026 <0.000 0.880 <0.000 0.013 <0.000 0.860 <0.000 0.012 0.002 0.860 <0.000 0.012 0.002
Female -0.001 0.936 -0.004 0.026 0.005 0.623 0.001 0.703 0.006 0.574 0.001 0.533 0.006 0.553 0.001 0.527
Age group (ref=52-59)
60-69 0.094 <0.000 0.006 0.004 0.077 <0.000 0.002 0.222 0.076 <0.000 0.002 0.287 0.077 <0.000 0.002 0.289
70-79 0.220 <0.000 0.012 <0.000 0.194 <0.000 0.011 <0.000 0.196 <0.000 0.011 <0.000 0.197 <0.000 0.011 <0.000
≥80 0.456 <0.000 0.011 0.006 0.432 <0.000 0.008 0.064 0.438 <0.000 0.008 0.066 0.438 <0.000 0.008 0.067
Race (ref=non-Hispanic White)
NonHispanic
Black
0.086 0.002 0.000 0.942 0.066 0.014 -0.003 0.331 0.065 0.017 -0.003 0.335 0.065 0.017 -0.003 0.332
Education (ref=less than high school)
High
school
graduate
-0.041 0.021 -0.002 0.639 -0.027 0.120 -0.001 0.699 -0.024 0.170 -0.001 0.727 -0.022 0.203 -0.001 0.730
Some
college
or more
-0.086 <0.000 -0.009 0.014 -0.059 <0.000 -0.005 0.076 -0.052 0.001 -0.005 0.078 -0.051 0.002 -0.005 0.080
Hypertension (ref=no history of
hypertension)
Controlled hypertension 0.106 <0.000 -0.002 0.311 0.106 <0.000 -0.002 0.295 0.106 <0.000 -0.003 0.289
Uncontrolled hypertension 0.029 0.022 0.019 0.000 0.945 0.021 0.023 0.000 0.972 0.021 0.029 0.000
Diabetes (ref=no history of diabetes)
Controlled diabetes 0.088 <0.000 0.002 0.624 0.088 <0.000 0.002 0.614 0.087 <0.000 0.002 0.584
Uncontrolled diabetes 0.033 0.044 0.008 0.008 0.034 0.040 0.008 0.008 0.033 0.046 0.008 0.007
Heart conditions 0.096 <0.000 0.007 0.001 0.095 <0.000 0.007 0.001 0.095 <0.000 0.007 0.001
BMI (ref=normal)
Overweight 0.010 0.371 0.004 0.086 0.012 0.289 0.003 0.093 0.012 0.306 0.003 0.095
Obese 0.064 <0.000 0.005 0.038 0.068 <0.000 0.005 0.049 0.067 <0.000 0.005 0.049
Smoking (ref=never smoked)
Former smoker 0.016 0.127 0.002 0.266 0.016 0.133 0.002 0.267
Current smoker 0.059 <0.000 0.000 0.896 0.058 <0.000 0.000 0.890
Problem drinking (CAGE≥2) -0.022 0.104 0.001 0.686 -0.021 0.131 0.001 0.688
Polygenic score (PGS) for kidney disease 0.021 <0.000 0.000 0.774

141
Table S3.1 CKD-EPI Formulas
CKD-EPI Formula: eGFR=μ × (Scr/k)a × (CysC/0.8)b × cAge × d[if female] × e[if Black]
*k=0.7 for female and 0.9 for male
Intercept μ Coefficients a for Scr Coefficients b for CysC
Coefficient
c for age
Coefficient
d for female
Coefficient
e for black
Female Male
Scr≤0.7 Scr>0.7 Scr≤0.9 Scr>0.9 CysC≤0.8 CysC>0.8
Scr 141 -0.329 -1.209 -0.411 -1.209 0.9929 1.018 1.16
Scr without
race 142 -0.241 -1.200 -0.302 -1.200 0.9938 1.012 n/a
CysC 133 -0.499 -1.328 0.9962 0.932 n/a
Scr_CysC
with race 135 -0.248 -0.601 -0.207 -0.601 -0.375 -0.711 0.9952 0.969 1.08
Scr_CysC
without race 135 -0.219 -0.544 -0.144 -0.544 -0.323 -0.778 0.9961 0.963 n/a
Example: For a non-Hispanic Black woman age 65, Scr level = 1 mg/dL; CysC = 1 mg/dL, using equation based on Scr_CysC with race
eGFR = 135× (1/0.7)-0.601 × (1/0.8)-0.711 × 0.995265 × 0.969 × 1.08 = 71.2 mL/min/1.73 m2.
Note. “cysc” stands for cystatin C; “scr” stands for creatinine
Notes on the CKD-EPI formulas:
The basic CKD-EPI formula for calculating eGFR expressed as a single equation is shown below.
eGFR=μ × (Scr/k)a × (CysC/0.8)b × cAge × d[if female] × e[if Black].
In the CKD-EPI formula, coefficient k equals 0.7 for women and 0.9 for men; the values of intercept μ and
coefficients c, d, and e vary depending on which one of the four equations is being used. The values of coefficient a
and b vary by equation as well as by sex and serum level of cystatin C/creatinine. For instance, for a non-Hispanic
Black woman age 65, with a Scr level of 1 mg/dL and a CysC level of 1 mg/dL, the assigned value for μ is 135; k is
0.7; a is -0.601; b is -0.711; c is 0.9952; d is 0.969; and e is 1.08. Her GFR value estimated by the CysC_Scr
equation would be 135×(1/0.7)-0.601 × (1/0.8)-0.711 × 0.995265 × 0.969 × 1.08 = 71.2 mL/min/1.73m2
. All values of
intercepts and coefficients can be found in Supplemental eTable 1 above.

142
Figure S3.1. Comparison of prevalence in Non-Hispanic White and Black individuals with
impaired kidney function with and without racial adjustment
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
with race without race with race without race with race without race with race without race
GFR_scr GFR_cysc_scr GFR_scr GFR_cysc_scr
Non-Hispanic White Non-Hispanic Black

143
Table S3.2 Odds ratios indicating Sex/racial/ethnic differentials from logistic regressions on
seven indicators of impaired kidney functioning with non-Hispanic Black individuals as the
reference group (n=8,821)
Cystatin C Creatinine eGFR cysc eGFR scr eGFR scr w/o race eGFRcysc_scr eGFRcysc_scr w/o
race
OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value
Age 1.12 <0.000 1.07 <0.000 1.14 <0.000 1.10 <0.000 1.10 <0.000 1.12 <0.000 1.12 <0.000
Female 0.85 0.007 1.29 0.001 1.32 <0.000 1.02 0.73 1.10 0.180 1.22 0.003 0.86 0.028
Race/Ethnicity (ref=Non-Hispanic Black)
Non-Hispanic
White 1.00 0.976 0.40 <0.000 0.83 0.057 0.77 0.018 0.40 <0.000 0.82 0.031 0.53 <0.000
Hispanic 0.98 0.826 0.33 <0.000 0.84 0.203 0.62 0.002 0.37 <0.000 0.72 0.014 0.48 <0.000
Pseudo R2 0.136 0.079 0.2 0.135 0.142 0.184 0.178
Note. “cysc” stands for cystatin C; “scr” stands for creatinine

144
Table S3.3 Odds ratios indicating age group differentials from logistic regressions on seven
indicators of impaired kidney functioning (n=8,821)
Cystatin C Creatinine eGFR cysc eGFR scr eGFR scr w/o
race eGFRcysc_scr eGFRcysc_scr
w/o race
OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value OR P-value
Age (ref = 56-59)
60-69 2.04 <0.000 1.52 0.008 2.68 <0.000 2.24 <0.000 2.3 <0.000 2.6 <0.000 2.99 <0.000
70-79 5.74 <0.000 3.23 <0.000 8.1 <0.000 6.22 <0.000 6.71 <0.000 7.95 <0.000 8.73 <0.000
80+ 20 <0.000 6.48 <0.000 34.72 <0.000 16.83 <0.000 18.25 <0.000 27.24 <0.000 30.22 <0.000
Pseudo R2 0.1194 0.0747 0.1761 0.1241 0.1334 0.1647 0.1621
Note. “cysc” stands for cystatin C; “scr” stands for creatinine

145
Table S3.4. Odds ratios from logistic regression of 4-year mortality (2016-2020) on various
indicators of kidney function (all indicators are standardized and converted to z-score) cysc scr eGFR_cysc eGFR_scr eGFR_scr w/o race eGFR_cysc_scr eGFR_cysc_scr w/o
race
OR OR OR OR OR OR OR
Age,
sex
+Health
correlates
Age,
sex
+Health
correlates
Age,
sex
+Health
correlates
Age,
sex
+Health
correlates
Age,
sex
+Health
correlates
Age,
sex
+Health
correlates
Age, sex +Health
correlates
cysc 0.686
[0.613,
0.767]
0.726
[0.654,
0.804]
scr 0.806
[0.744,
0.872]
0.841
[0.782,
0.904]
eGFR_cysc 0.473
[0.408,
0.547]
0.516
[0.446,
0.597]
eGFR_scr 0.750
[0.673,
0.837]
0.812
[0.727,
907]
eGFR_scr_
w/o race
0.742
[0.669,
0.824]
0.803
[0.723,
0.892]
eGFR_cysc_
scr
0.573
[0.507,
0.648]
0.627
[0.553,
0.712]
eGFR_cysc_
scr w/o race
0.573
[0.507,
0.648]
0.627
[0.553,
0.711]
Pseudo R2 0.192 0.205 0.177 0191 0.209 0.214 0.177 0.189 0.177 0.190 0.195 0.203 0.195 0.203
Note. All at p<.0001; “cysc” stands for cystatin C; “scr” stands for creatinine

146
Table S3.5 Sensitivity and specificity of equations predicting kidney-related mortality by each
kidney functioning indicator, stratified by sex and race
Kidney indicator Sensitivity % Specificity % AUC
Creatinine 42.9 84.0 0.63 [0.59, 0.68]
Cystatin C 91.4 41.3 0.66 [0.64, 0.69]
eGFRcysc 84.8 60.8 0.73 [0.69, 0.76]
All eGFRscr adjusted for
race
56.2 79.9 0.68 [0.63, 0.73]
eGFRscr unadjusted
for race 54.3 82.5 0.68 [0.64, 0.73]
eGFRcysc_scr adjusted
for race 71.4 71.6 0.72 [0.67, 0.76]
eGFRcysc_scr
unadjusted for race 70.5 73.7 0.72 [0.68, 0.76]
Creatinine 53.6 82.7 0.68 [0.62, 0.75]
Cystatin C 91.1 42.7 0.67 [0.63, 0.71]
eGFRcysc 83.9 59.3 0.72 [0.67, 0.76]
Women eGFRscr adjusted for
race
62.5 80.0 0.71 [0.65, 0.78]
eGFRscr unadjusted
for race 60.7 82.1 0.71 [0.65, 0.78]
eGFRcysc_scr adjusted
for race 76.8 70.6 0.74 [0.68, 0.79]
eGFRcysc_scr
unadjusted for race 71.4 74.6 0.73 [0.67, 0.79]
Creatinine 30.6 85.8 0.58 [0.52, 0.65]
Cystatin C 91.8 39.3 0.66 [0.62, 0.70]
eGFRcysc 85.7 63.1 0.74 [0.69, 0.79]
eGFRscr adjusted for
race
49.0 79.8 0.64 [0.57, 0.71]
Men eGFRscr unadjusted
for race 46.9 83.0 0.65 [0.58, 0.72]

147
eGFRcysc_scr adjusted
for race 65.3 73.0 0.69 [0.62, 0.76]
eGFRcysc_scr
unadjusted for race 69.4 72.4 0.71 [0.64, 0.77]
Creatinine 42.1 84.8 0.63 [0.58, 0.69]
Cystatin C 94.7 37.0 0.66 [0.63, 0.68]
eGFRcysc 90.8 56.9 0.74 [0.71, 0.77]
eGFRscr adjusted for
race
60.5 77.7 0.69 [0.64, 0.75]
NonHispanic
White
eGFRscr unadjusted
for race 54.0 82.6 0.68 [0.63, 0.74]
eGFRcysc_scr adjusted
for race 75.0 68.3 0.72 [0.67, 0.77]
eGFRcysc_scr
unadjusted for race 72.4 71.8 0.72 [0.67, 0.77]
Creatinine 47.6 75.7 0.62 [0.51, 0.73]
Cystatin C 76.2 48.3 0.62 [0.53, 0.72]
eGFRcysc 61.9 66.7 0.64 [0.54, 0.75]
eGFRscr adjusted for
race
38.1 81.7 0.60 [0.49, 0.71]
NonHispanic
Black
eGFRscr unadjusted
for race 57.1 76.4 0.67 [0.56, 0.78]
eGFRcysc_scr adjusted
for race 57.1 75.6 0.66 [0.55, 0.77]
eGFRcysc_scr
unadjusted for race 66.7 72.9 0.70 [0.59, 0.80]
Note. “cysc” stands for cystatin C; “scr” stands for creatinine

148
Figure S4.1a Flow chart of the sample selection for the Chinese sample
17,708 individuals interviewed
3,374 did not have biomarker data at
baseline.
13,974 eligible individuals at
baseline 979 passed away by 2015
1,927 loss of follow-up by 2015
Outcomes
6,834 did not have cystatin C data at baseline;
5,922 did not have cystatin C in 2015
Covariates
85 did not have data on education level; 162 did
not have data on insurance; 1 did not have age;
233 did not have income data; 4,536 did not have
data on diabetes management (4,445 missing on
hba1c); 2,961 missing on hypertension
management; 162 did not have data on heart
problems; 576 did not have smoking data; and
2,957 did not have BMI data.
Final sample size:
4,016 eligible participants

149
Figure S4.1b Flow chart of sample selection for the U.S. sample
15,378 individuals interviewed in
2010/2012
6,985 individuals did not have
biomarker
8,394 eligible individuals at
baseline
968 passed away by 2014/2016
2,987 loss of follow-up by 2014/2016
Outcomes
414 did not have cystatin C data at baseline; 1,873
did not have cystatin C in 2014/2016
Covariates
12 did not have data on education level; 61 did
not have data on insurance; 7 did not have age;
356 did not have income data; 312 did not have
data on diabetes management; 413 missing on
hypertension management; 10 did not have data
on heart problems; 50 did not have smoking data;
and 494 did not have BMI data.
Final sample size:
5,497 eligible participants

150
Table S4.1 Association between baseline kidney functioning and SES and CKD risk factors (n
=5,574 for China)
Base outcome:
Normal
kidney
function
Mildly impaired Moderately impaired Severely impaired
China China China
Model 1 Model 2 Model 1 Model 2 Model 1 Model 2
Age, Sex, SES +Risk factors Age, Sex, SES +Risk factors Age, Sex, SES +Risk factors
Age group (ref=50-59)
60-69 2.243*** 2.190*** 6.042*** 5.675*** 6.042*** 5.675***
70-79 3.783*** 3.656*** 42.75*** 39.115*** 42.75*** 39.115***
80+ 1.384 1.296 70.256*** 62.537*** 70.256*** 62.537***
Female 0.737* 0.817 0.550*** 0.598** 0.550*** 0.598**
Educational attainment (reference: Lower)
Middle 0.914 0.9 0.838 0.838 0.838 0.838
Higher 0.709 0.733* 0.688* 0.687* 0.688* 0.687*
Income 1 1 0.998 0.999 0.998 0.999
Insurance (ref=no insurance)
Urban
employee
and/or
private
0.538* 0.542* 0.552 0.604 0.552 0.604
Urban
resident
and/or other
public
0.8 0.803 0.859 0.897 0.859 0.897
Rural
cooperative 0.752 0.742 0.697 0.682 0.697 0.682
Diabetes (ref = no history)
Controlled 0.946
Uncontrolled 0.435***
Hypertension (ref = no history)
Controlled 1.038 1.515* 1.515*
Uncontrolled 1.03 1.212 1.212
Heart Problems 1.342 1.327 1.327
Smoking (ref = never smoked)
Used to smoke 1.17 1.171 1.171
Current smoker 1.305 1.426 1.426
Drinking (ref = never drunk)
Used to drink 1.03 1.291 1.291
Current drinker 0.967 0.959 0.959

151
Obesity (ref = normal)
Overweight 0.897 0.824 0.824
Obese 1.253 1.689* 1.689*
Notes: *p<0.05, ** p<0.01, *** p<0.001

152
Table S4.2 Association between baseline kidney functioning and SES/CKD risk factors with
different reference groups (n = 5,497 for U.S.; n =4,016 for China)
Base
outcome:
Normal
kidney
function
Mildly impaired Moderately impaired Severely impaired
US China US China US China
Model 1 Model 2 Model 1 Model
2 Model 1 Model 2 Model 1 Model
2 Model 1 Model 2 Model 1 Model 2
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES
+Risk
factors
Age, Sex,
SES +Risk factors Age, Sex,
SES
+Risk
factors
Age, Sex,
SES +Risk factors Age, Sex,
SES
+Risk
factors
Age group (ref=50-59)
60-69 1.163 1.189 2.272*** 2.25*** 1.965*** 1.918*** 6.058*** 6.10*** 5.080* 4.484* 3.18 3.20*
70-79 1.584** 1.705*** 4.006*** 4.16*** 4.267*** 4.901***
41.392*** 44.4*** 14.363*** 16.17*** 18.683*** 17.97***
80+ 3.398*** 3.841*** 0.765 0.792 17.278*** 22.769*** 41.593*** 46.7*** 93.749*** 121.3*** 76.675*** 81.68***
Female 1.160* 1.155 0.700* 0.768 1.307** 1.345*** 0.486*** 0.493* 1.538** 1.643* 0.317** 0.32
Educational attainment (reference: Middle)
Lower 1.026 1.981 1.958 1.024 1.073 0.935 1.014 0.977 1.050 0.854 1.144 1.818
Higher 0.961 1.020 0.769 0.81 0.805* 0.918 0.782 0.788 0.571** 0.683* 1.059 1.798
Income 1 1 1 1 0.999 1 0.998 0.999 0.994** 0.996 1.002 0.863
Insurance (ref=Public for U.S.)
No insurance 0.961 0.957 1.321 1.794 0.275* 0.264*
Private 0.768* 0.807* 0.589*** 0.699** 0.285*** 0.382**
Public +
private 0.828 0.853 0.817* 0.862 0.848 0.893
Insurance (ref= Urban employee/private)
No insurance 1.653 0.595 1.635 1.676 1.500 0.548
Urban resident and/or other public 1.251 0.727 1.281 1.205 2.564 0.993
Rural cooperative 1.349 0.799 1.175 1.340 0.812 0.474
Diabetes (ref = controlled)
No history 1.193 0.873 0.981 0.871 0.577** 1.370
Uncontrolled 0.902 0.340** 0.930 0.340** 0.630 5.830
Hypertension (ref = controlled)
No history 0.890 1.051 0.562*** 0.948 0.257*** 0.217**
Uncontrolled 1.044 0.894 0.754** 0.851 0.625** 0.380
Heart Problems 1.177 1.41 1.548*** 1.358 2.278*** 1.371
Smoking (ref
= used to
smoke)
Never
smoked 1.102 1.17 0.958 0.971 0.842 0.415
Current
smoker 1.515** 1.353* 1.882*** 1.313 1.866* 0.651
Drinking
(ref = used to
drink)
Never drank 0.912 1.03 1.067 0.938 1.833 1.194
Current
drinker 0.710** 0.967 0.599*** 0.855 0.589* 0.296*

153
Obesity (ref
= normal)
Overweight 1.264* 0.897 1.586*** 0.824 1.172 0.984
Obese 1.549*** 1.253 2.470*** 2.325** 2.116*** 0.471
Pseudo R2 0.0911 0.1179 0.1305 0.1460 0.0911 0.1179 0.1305 0.1460 0.0911 0.1179 0.1305 0.1460
Notes: *p<0.05, ** p<0.01, *** p<0.001

154
Table S4.3 Multinomial regression: Associations between change in kidney functioning and SES, and CKD risk factors with different
references (n = 5,497 for U.S.; n =4,016 for China)
Base
outcome:
Continuously
good
Continuously bad Improved Worsened Death
US China US China US China US China
Age, sex,
SES risk factors Age, sex,
SES
risk
factors
Age, sex,
SES risk factors Age, sex,
SES
risk
factors
Age, sex,
SES risk factors Age, sex,
SES
risk
factors
Age, sex,
SES risk factors Age, sex,
SES
risk
factors
GFR at baseline 0.878*** 0.878*** 0.877*** 0.876*** 0.870*** 0.867*** 0.950*** 0.949*** 1.001 1.003 1.015** 1.015** 0.879*** 0.883*** 0.925*** 0.926***
Age (ref = 50-
59)
60-69 0.675 0.684 2.246*** 2.313*** 0.735 0.772 0.638** 0.639** 1.316 1.259 1.850*** 1.783*** 1.058 1.075 1.302 1.204
70-79 0.608 0.6 4.504*** 4.757*** 0.545 0.544 0.596* 0.604* 1.664** 1.702** 3.302*** 3.142*** 1.358 1.946 3.960*** 3.508***
80+ 1.822* 2.037* 17.632*** 19.490*** 0.596 0.6 0.537 0.533 2.970*** 3.181*** 22.298*** 21.316*** 8.410*** 9.067*** 56.174*** 47.040***
Female 0.992 1.004 0.885 0.76 0.871 0.852 0.985 0.997 1.031 1.142 1.307* 1.343 0.569*** 0.619** 0.721 0.943
Education
(ref=middle ed)
Lower ed 0.797 0.989 1.177 1.138 1.22 0 .764 1.315 1.306 0.967 1.03 1.455** 0.699** 0.889 0.942 1.172 1.141
Higher ed 0.906 1.17 1.099 1.163 1.495 0.974 1.094 1.079 0.875 0.984 0.932 0.648* 0.943 0.947 0.581* 0.524*
Household per
capita income 0.999 0.999 1.003 1.003 1.001* 1.001* 1.002 1.002 0.999 0.999 0.998 0.998 0.994* 0.995* 1.002 1.002
Insurance (ref =
public)
No insurance 0.567 0.605 1.045 0.980 0.882 0.919 0.732 0.687
Private 0.522** 0.529** 0.777 0.741 0.686** 0.741* 0.371** 0.451*
Both public and
private 0.937 0.929 0.939 0.893 0.907 0.925 0.903 0.928
Insurance (ref=
Urban employee
and/or private )
No insurance 0.525 0.603 0.791 0.827 0.465 0.494 2.098 1.725
Urban resident
and other public 0.796 0.956 1.308 1.319 0.826 0.826 1.803 1.525
Rural
cooperative 0.799 0.931 1.470 1.508 0.543 0.580 1.935 1.691
Diabetes
(ref=controlled)
No history 1.075 1.307 0.974 1.772 0.817 1.772 0.815 0.975
Uncontrolled 1.038 1.092 0.760 2.715** 1.221 3.093** 1.628 5.010**

155
Hypertension
(ref=controlled)
No history 0.858 1.041 1.140 0.977 0.905 0.850 0.705* 0.712
Uncontrolled 0.825 1.066 0.785 0.834 1.035 1.266 0.968 2.293**
Heart problems 1.098 1.202 0.82 1.18 1.873*** 1.431 1.459* 1.197
Smoking
(ref=used to
smoke)
Never smoked 0.939 1.080 1.014 1.097 0.966 1.211 0.590*** 0.720
Current smoker 0.983 0.680 1.076 0.998 1.020 1.171 1.279 1.642
Drinking (ref=
used to drink)
Never drank 0.830 1.286 1.053 1.706* 1.096 1.055 1.196 1.316
Current drinker 0.699* 1.22 1.102 1.309 0.875 0.822 0.802 1.087
BMI
(ref=normal)
Overweight 1.227 1.272 1.145 0.91 1.353 1.286 0.611** 0.746
Obese 1.449** 2.152** 1.067 1.042 1.786** 1.467* 0.613** 0.818
Pseudo R2 0.2063 0.2233 0.2017 0.2153 0.2063 0.2233 0.2017 0.2153 0.2063 0.2233 0.2017 0.2153 0.2063 0.2233 0.2017 0.2153
Notes: *p<0.05, ** p<0.01, *** p<0.001

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Biomarkers of age-related health changes: associations with health outcomes and disparities

PDF

Social determinants of physiological health and mortality in China

PDF

Statistical algorithms for examining gene and environmental influences on human aging

PDF

Is stress exposure enough? Race/ethnic differences in the exposure and appraisal of chronic stressors among older adults

PDF

Internet communication use, psychological functioning and social connectedness at older ages

PDF

International sex and age differences in physical function and disability

PDF

Three essays on modifiable determinants of shingles: risk factors for shingles incidence and factors affecting timing of vaccine uptake

PDF

Understanding the role of population and place in the dynamics of seasonal influenza outbreaks in the United States

PDF

Self-perceptions of Aging in the Context of Neighborhood and Their Interplay in Late-life Cognitive Health

PDF

Estimating survival in the face of pain: evidence from the health and retirement study

PDF

Health differences between the elderly in Japan and the United States

PDF

Residential care in Los Angeles: policy and planning for an aging population

PDF

Marital quality, gender, and biomarkers of disease risk in the MIDUS cohort

PDF

Cognitive health and self-rated memory in later life: linkages to race/ethnicity, multimorbidity, and survival

PDF

Age integration in late life: sociodemographic & psychosocial correlates of intergenerational-only, peer-only, and age-integrated social networks

PDF

Exploring end-of-life care experiences and disparities among Hispanic populations in the United States

PDF

Examination of the long-term psychosocial and functioning effects of a healthy living, life management behavior intervention for older adults

PDF

More knowledge, better plans: a study of heterogeneity in dementia prevalence and mid-life cognitive changes

PDF

Palliative care: breaking the knowledge barrier among community-based older adults

PDF

Psychological and physiological pathways from social support exchanges to health: a lifespan perspective

Asset Metadata

Creator Zhao, Erfei (author)

Core Title A biodemographic approach to understanding sociodemographic disparities in kidney functioning on three dimensions: individual, population, and cross-national

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Degree Conferral Date 2024-08

Publication Date 08/31/2024

Defense Date 02/21/2024

Publisher Los Angeles, California (original), University of Southern California (original), University of Southern California. Libraries (digital)

Tag aging,biomarkers,disparities,Kidney

Format theses (aat)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Crimmins, Eileen (committee chair), Ailshire, Jennifer (committee member), Kim, Jung Ki (committee member)

Creator Email erfeizha@usc.edu,philemonerfei@gmail.com

Unique identifier UC11399A0AP

Identifier etd-ZhaoErfei-13463.pdf (filename)

Legacy Identifier etd-ZhaoErfei-13463

Document Type Dissertation

Format theses (aat)

Rights Zhao, Erfei

Internet Media Type application/pdf

Type texts

Source 20240831-usctheses-batch-1205 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu