Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Sep 1.
Published in final edited form as: Ann Intern Med. 2024 Jan 30;177(3):269–279. doi: 10.7326/M23-1138

Association of Low Glomerular Filtration Rate With Adverse Outcomes at Older Age in a Large Population With Routinely Measured Cystatin C

Edouard L Fu 1,2,3, Juan-Jesus Carrero 4, Yingying Sang 5, Marie Evans 6, Junichi Ishigami 7, Lesley A Inker 8, Morgan E Grams 9, Andrew S Levey 10, Josef Coresh 11,12,*, Shoshana H Ballew 13,14,*
PMCID: PMC11079939  NIHMSID: NIHMS1988167  PMID: 38285982

Abstract

Background:

The commonly accepted threshold of glomerular filtration rate (GFR) to define chronic kidney disease (CKD) is less than 60 mL/min/1.73 m2. This threshold is based partly on associations between estimated GFR (eGFR) and the frequency of adverse outcomes. The association is weaker in older adults, which has created disagreement about the appropriateness of the threshold for these persons. In addition, the studies measuring these associations included relatively few outcomes and estimated GFR on the basis of creatinine level (eGFRcr), which may be less accurate in older adults.

Objective:

To evaluate associations in older adults between eGFRcr versus eGFR based on creatinine and cystatin C levels (eGFRcr-cys) and 8 outcomes.

Design:

Population-based cohort study.

Setting:

Stockholm, Sweden, 2010 to 2019.

Participants:

82 154 participants aged 65 years or older with outpatient creatinine and cystatin C testing.

Measurements:

Hazard ratios for all-cause mortality, cardiovascular mortality, and kidney failure with replacement therapy (KFRT); incidence rate ratios for recurrent hospitalizations, infection, myocardial infarction or stroke, heart failure, and acute kidney injury.

Results:

The associations between eGFRcr-cys and outcomes were monotonic, but most associations for eGFRcr were U-shaped. In addition, eGFRcr-cys was more strongly associated with outcomes than eGFRcr. For example, the adjusted hazard ratios for 60 versus 80 mL/min/1.73 m2 for all-cause mortality were 1.2 (95% CI, 1.1 to 1.3) for eGFRcr-cys and 1.0 (CI, 0.9 to 1.0) for eGFRcr, and for KFRT they were 2.6 (CI, 1.2 to 5.8) and 1.4 (CI, 0.7 to 2.8), respectively. Similar findings were observed in subgroups, including those with a urinary albumin–creatinine ratio below 30 mg/g.

Limitation:

No GFR measurements.

Conclusion:

Compared with low eGFRcr in older patients, low eGFRcr-cys was more strongly associated with adverse outcomes and the associations were more uniform.

Primary Funding Source:

Swedish Research Council, National Institutes of Health, and Dutch Kidney Foundation.

In routine clinical practice, glomerular filtration rate (GFR) is usually estimated from serum creatinine level (eGFRcr). An eGFRcr below 60 mL/min/1.73 m2 is very common in older persons, with prevalence estimates ranging from 23% to 44% for those aged 65 years or older (1–3). The rationale for using a GFR threshold of less than 60 mL/min/1.73 m2 to define chronic kidney disease (CKD) is that it represents a large decrease in GFR from the normal value in young persons and because of associations between low eGFRcr and adverse outcomes, such as kidney failure and all-cause mortality, in the general population (4, 5). However, eGFRcr levels of 60 mL/min/1.73 m2 or lower are less strongly associated with adverse outcomes in older adults than in young persons (6–10). This has led to substantial debate about the appropriateness of the current GFR threshold to define CKD in older adults, and several researchers have called for lower thresholds of GFR below 45 mL/min/1.73 m2 to define CKD in adults aged 65 years or older (6, 11–16).

The weaker associations between eGFRcr and adverse outcomes in older adults may be due in part to the limitations of creatinine for estimating GFR in the older population. Low muscle mass, which is frequently found with chronic disease or inactivity, may lead to low creatinine levels and consequently high eGFRcr that does not accurately represent true GFR (17–19). Patients with eGFRcr above 60 mL/min/1.73 m2 may therefore disproportionately include persons who are at high risk for adverse outcomes because of low muscle mass and frailty, which dilutes risk associations between low eGFRcr and adverse outcomes. Furthermore, previous studies that investigated the risks associated with CKD in older age focused primarily on the outcomes of kidney failure and mortality (8, 10). However, CKD increases the risk for many outcomes that are important to patients, and investigating these additional outcomes is important to appreciate the broad range of risks associated with CKD (20).

The limitations of eGFRcr may be overcome by using more accurate GFR estimates. Using both creatinine and cystatin C to calculate GFR (eGFRcr-cys) is the most accurate estimation method in both younger and older adults (21–23). Recent recommendations suggest that cystatin C testing should be available and widely used in the United States (24). In contrast to creatinine, cystatin C is minimally influenced by muscle mass (17, 18, 25). Consequently, associations between eGFR and outcomes would be expected to be more accurate for eGFRcr-cys than for eGFRcr.

To inform the evidence for the definition and staging of CKD in older adults, we evaluated associations between eGFRcr-cys versus eGFRcr with 8 outcomes in older adults receiving routine care in Sweden, where use of cystatin C testing as a supportive test in routine clinical practice was implemented more than a decade ago.

Methods

This study falls into the category of prognostic factor research (26), which studies factors whose values are associated with changes in the outcome’s risk. The prognostic factor framework is often used to define diseases (27) and has been a cornerstone in the definition and staging of CKD, on which clinical recommendations are based (5, 28).

Data Source and Study Population

We used data from the SCREAM (Stockholm CREAtinine Measurements) project, which contains health care utilization data from residents of Stockholm, Sweden, between 2006 and 2019 (29). SCREAM contains complete information on demographic characteristics, health care utilization, laboratory tests, dispensed drugs (30), diagnoses (31), and vital status (32). A single health care provider in the Stockholm region provides universal and tax-funded health care to 20% to 25% of the population of Sweden. The Regional Ethical Review Board in Stockholm approved the study (reference 2017/793–31); informed consent was not deemed necessary because all data were deidentified by the Swedish Board of Health and Welfare.

We included all persons aged 65 years or older who had same-day routine outpatient tests for creatinine and cystatin C between 1 January 2010 and 31 December 2018. Persons with a history of kidney failure with replacement therapy (KFRT) were excluded. If persons had multiple pairs of same-day creatinine and cystatin C measurements, we selected the first pair. Creatinine tests were standardized to isotope dilution mass spectrometry–traceable methods, and cystatin C measurements were also standardized (33). Performance of creatinine and cystatin C assays was monitored internally and by an external quality assessment program (Equalis). To investigate whether testing for cystatin C was random or directed at persons with certain characteristics, we also considered baseline characteristics for the complete cohort of older adults with a creatinine test available and compared their characteristics with those of the subset who had both a creatinine test and a cystatin C test.

Study Exposure

The primary study exposures were eGFRcr and eGFRcr-cys, calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2021 equations (22, 24). For comparisons with previous studies, we also evaluated eGFR based on cystatin C alone (eGFRcys), calculated using the CKD-EPI 2012 equation (34). All 3 CKD-EPI equations included age, sex, and filtration markers and did not include a coefficient for race.

Covariates

Covariates of interest included age, sex, hypertension, diabetes, history of cardiovascular disease, antihypertensive medication use, total and high-density lipoprotein cholesterol level, and urinary albumin–creatinine ratio (UACR) (definitions are provided in Supplement Table 1, available at Annals.org). We converted urinary protein–creatinine ratio and urine dipstick protein categories to UACR values using previously published conversion equations (35). We used the most recent value before the index date for laboratory values, with a maximum lookback period of 3 years (results were similar when we used a lookback period of 1 year).

Study Outcomes and Follow-up

The 8 study outcomes were all-cause mortality; cardiovascular mortality (32, 36); KFRT (37, 38); all-cause hospitalization; and hospitalization with infection, myocardial infarction or stroke, heart failure, or acute kidney injury (AKI) (definitions are provided in Supplement Table 1). The date of the first concurrent creatinine and cystatin C test after age 65 years was considered the index date and the start of follow-up. For the outcomes of all-cause mortality, cardiovascular mortality, and KFRT, patients were followed until occurrence of the outcome, death, or the end of available data (31 December 2019), whichever occurred first. For the other outcomes, patients were followed until death or 31 December 2019.

Statistical Analysis

We used cause-specific Cox regression models to estimate hazard ratios for the associations between eGFR and the outcomes of all-cause mortality, cardiovascular mortality, and KFRT. Negative binomial regression was used to estimate incidence rate ratios for the associations between eGFR and the recurrent outcomes, which could occur multiple times during follow-up. We modeled continuous eGFR (the exposure) using linear splines with knots at 30, 45, 60, 75, and 90 mL/min/1.73 m2. We selected 80 mL/min/1.73 m2 as the reference (hazard ratio or incidence rate ratio of 1.0), which is consistent with previous studies in older adults (6, 8). Of note, the reference point influences statistical significance at a given eGFR but does not influence the shape of the overall association between eGFR and outcomes. We adjusted for age, sex, hypertension, diabetes, cardiovascular disease, and log-transformed UACR. Analyses for cardiovascular mortality, all-cause hospitalization, myocardial infarction or stroke, and heart failure were also adjusted for total and high-density lipoprotein cholesterol levels and antihypertensive medication use. From the fitted models, we calculated hazard ratios and incidence rate ratios at eGFR values of 30, 45, and 60 mL/min/1.73 m2 (compared with the reference of 80 mL/min/1.73 m2), corresponding to current Kidney Disease: Improving Global Outcomes (KDIGO) GFR thresholds for CKD stages G3a, G3b, and G4 (5). Details on handling of missing data are provided in the Supplement Methods (available at Annals.org).

We also calculated incidence rates for specific ages (70 and 80 years) at specific eGFRcr-cys levels (30, 45, and 60 mL/min/1.73 m2) using Poisson regression for mortality and KFRT and negative binomial regression for the recurrent outcomes. These models included an interaction between the linear eGFR spline and age. To calculate incidence rates at specific ages and eGFRcr-cys levels, we used the mean value of each covariate in the study population with same-day creatinine and cystatin C measurements.

Many treatment recommendations for kidney disease are based on GFR thresholds (5). Thus, to investigate potential clinical implications of using eGFRcr-cys versus eGFRcr, we assessed reclassification across KDIGO GFR categories by cross-tabulating eGFR categories for eGFRcr versus eGFRcr-cys (5).

We also performed stratified analyses to assess consistency of associations for the following subgroups: age (≥75 vs. <75 years), sex, presence of cardiovascular disease, presence of diabetes, and UACR below 30 mg/g. In supporting analyses, we calculated eGFRcr and eGFRcr-cys using the CKD-EPI 2009 and 2012 equations with age, sex, and the non-Black race coefficient for all participants (39), which remain commonly used in Europe. We also calculated eGFRcr, eGFRcys, and eGFRcr-cys using the European Kidney Function Consortium (EKFC) equations with age and sex (40, 41), which were specifically developed for a European population. We note that our primary aim was to compare the association with adverse outcomes for eGFR based on different filtration markers (creatinine vs. cystatin C vs. both) rather than to compare newer versus older CKD-EPI equations (39, 42) or to compare CKD-EPI equations with eGFR equations developed by other research groups (21, 40, 41, 43–45). Finally, we also estimated standardized incidence rates instead of conditional incidence rates (additional details are provided in the Supplement Methods). Stata 16 MP was used for analyses.

Role of the Funding Source

The funder had no role in the design, conduct, or analysis of the study or the decision to submit the manuscript for publication.

Results

Study Population

In the Stockholm region, 19% of older adults with outpatient creatinine testing had a same-day cystatin C measurement (Supplement Figure 1, available at Annals.org). Compared with the total population tested for creatinine, those with both a creatinine test and a cystatin C test were older and had a higher prevalence of diabetes and cardiovascular disease (Table 1). In the cohort with creatinine and cystatin C tests (n = 82 154), the mean age was 77 years, 50% were female, 41% had a history of cardiovascular disease, and 25% had diabetes (Table 1). Mean eGFRs were 67 mL/min/1.73 m2 for eGFRcr, 61 mL/min/1.73 m2 for eGFRcr-cys, and 54 mL/min/1.73 m2 for eGFRcys (distributions are shown in Supplement Figure 2, available at Annals.org).

Table 1.

Baseline Characteristics of Persons Aged 65 Years or Older With Creatinine Testing and a Subset With Same-Day Creatinine and Cystatin C Testing in Stockholm During 2010–2019

Characteristic Population With Creatinine Testing Subset With Creatinine and Cystatin C Testing
Overall Aged 65–74 y Aged ≥75 y UACR <30 mg/g* UACR ≥30 mg/g*
Persons, n 432 198 82 154 39 562 42 592 29 998 11 214
Mean age (SD), y 73 (8) 77 (8) 70 (3) 83 (6) 76 (8) 76 (8)
Mean eGFRcr (SD), mL/min/1.73 m2 78 (18) 67 (22) 74 (21) 61 (21) 69 (21) 54 (25)
Mean eGFRcr-cys (SD), mL/min/1.73 m2 – 61 (24) 70 (23) 53 (21) 63 (23) 47 (24)
Mean eGFRcys (SD), mL/min/1.73 m2 – 54 (24) 64 (24) 45 (20) 56 (23) 40 (22)
Female, % 55.0 49.9 43.6 55.9 51.3 36.0
Hypertension, % 35.4 80.3 74.9 85.5 81.9 92.4
Antihypertensive medication use, % 19.9 75.9 71.1 80.6 77.8 88.3
Diabetes, % 10.4 24.6 26.3 23.2 31.2 52.1
History of cardiovascular disease, % 23.5 40.5 30.0 50.5 38.2 50.8
Mean total cholesterol level (SD)†
 mmol/L 5.1 (1.2) 5.1 (1.2) 5.1 (1.2) 5.0 (1.2) 5.0 (1.2) 4.7 (1.2)
 mg/dL 196 (46) 195 (46) 197 (47) 192 (46) 193 (46) 183 (48)
Mean HDL cholesterol level (SD)†
 mmol/L 1.4 (0.5) 1.4 (0.5) 1.4 (0.5) 1.5 (0.5) 1.5 (0.5) 1.3 (0.4)
 mg/dL 55 (18) 55 (18) 55 (18) 56 (18) 56 (18) 49 (17)
Median UACR (IQR), mg/g† 16 (6–58) 17 (6–68) 13 (5–58) 22 (8–79) 8.0 (4.4–14.2) 113 (52–362)
Open in a new tab

eGFRcr = estimated glomerular filtration rate using creatinine level; eGFRcr-cys = estimated glomerular filtration rate using creatinine and cystatin C levels; eGFRcys = estimated glomerular filtration rate using cystatin C level; HDL = high-density lipoprotein; UACR = urinary albumin–creatinine ratio.

*

Converted using dipstick values when UACR was missing.

†

In the population tested for creatinine, data on total cholesterol level, HDL cholesterol level, and UACR were missing in 29.4%, 44.3%, and 64.8% of persons, respectively. In the subset of persons tested for creatinine and cystatin C, the respective proportions of missing data were 24.5%, 32.7%, and 49.8% overall; 16.5%, 23.0%, and 46.3% among persons aged 65 to 74 years; 32.0%, 41.7%, and 53.1% among those aged ≥75 years; 15.2%, 21.4%, and 0% among those with UACR <30 mg/g; and 13.4%, 19.8%, and 0% among those with UACR ≥30 mg/g. To convert UACR from mg/g to mg/mmol, multiply by 0.113. The numbers shown are before multiple imputation.

Association Between eGFR and Outcomes

The median follow-up was 3.9 years (IQR, 2.0 to 6.2 years). During follow-up, 31 219 persons died (9654 due to cardiovascular causes) and 841 progressed to KFRT (Supplement Table 2, available at Annals.org). Furthermore, counting first events only, there were 51 096 hospitalizations, of which 26 754 involved infections, 16 074 involved heart failure, 8549 involved myocardial infarction or stroke, and 5014 involved AKI.

Figure 1 shows adjusted associations between eGFRcr, eGFRcr-cys, eGFRcys, and outcomes. Lower eGFRcr and eGFRcr-cys were associated with higher hazard and incidence rate ratios for all outcomes. We observed U-shaped associations for eGFRcr, with higher hazard and incidence rate ratios at values above 90 mL/min/1.73 m2 for all outcomes except KFRT and AKI. In contrast, associations of eGFRcr-cys and eGFRcys with outcomes were approximately linear. eGFRcr-cys and eGFRcys were more strongly associated with outcomes than eGFRcr. For example, adjusted hazard ratios and incidence rate ratios for 60 vs. 80 mL/min/1.73 m2 for eGFRcr, eGFRcr-cys, and eGFRcys were 1.0 (95% CI, 0.9 to 1.0), 1.2 (CI, 1.1 to 1.3), and 1.3 (CI, 1.2 to 1.4) for all-cause mortality; 1.0 (CI, 0.9 to 1.1), 1.3 (CI, 1.2 to 1.4), and 1.4 (CI, 1.2 to 1.6) for cardiovascular mortality; 1.4 (CI, 0.7 to 2.8), 2.6 (CI, 1.2 to 5.8), and 1.5 (CI, 0.6 to 3.9) for KFRT; 1.2 (CI, 1.1 to 1.4), 1.5 (CI, 1.4 to 1.7), and 1.7 (CI, 1.6 to 1.9) for heart failure; and 1.6 (CI, 1.4 to 1.9), 2.3 (CI, 2.0 to 2.6), and 2.3 (CI, 1.9 to 2.7) for AKI (Supplement Table 3, available at Annals.org). Differences between eGFRcr, eGFRcr-cys, and eGFRcys were also observed when eGFR levels of 30 or 45 vs. 80 mL/min/1.73 m2 were compared (Supplement Table 3).

Figure 1.

Figure 1.

Open in a new tab

Adjusted hazard ratios and incidence rate ratios and 95% CIs for eGFRcr vs. eGFRcr-cys vs. eGFRcys (per 1–mL/min/1.73 m2 increase) and outcomes. eGFR was modeled using a linear spline with knots at 30, 45, 60, 75, and 90 mL/min/1.73 m2. The reference value was 80 mL/min/1.73 m2. Adjusted hazard ratios for all-cause mortality, cardiovascular mortality, and KFRT were estimated using Cox regression, and adjusted incidence rate ratios for recurrent all-cause hospitalizations and hospitalizations with infection, MI or stroke, heart failure, and AKI were estimated using negative binomial regression. AKI = acute kidney injury; eGFR = estimated glomerular filtration rate; eGFRcr = estimated glomerular filtration rate using creatinine level; eGFRcr-cys = estimated glomerular filtration rate using creatinine and cystatin C levels; eGFRcys = estimated glomerular filtration rate using cystatin C level; KFRT = kidney failure with replacement therapy; MI = myocardial infarction.

Absolute incidence rates for all events except KFRT and AKI were higher at older age (Figure 2 and Table 2). For example, adjusted incidence rates per 100 person-years for all-cause mortality at eGFRcr-cys of 30, 60, and 80 mL/min/1.73 m2 were 6.6 (CI, 6.1 to 7.2), 3.3 (CI, 3.1 to 3.6), and 2.7 (CI, 2.6 to 2.8), respectively, for patients aged 70 years and 12 (CI, 11 to 12), 7.4 (CI, 7.0 to 7.9), and 6.7 (CI, 6.4 to 6.9) for patients aged 80 years. Furthermore, absolute incidence rates for all outcomes were higher with lower eGFRcr-cys. For example, incidence rates for recurrent hospitalizations at age 70 years for eGFRcr-cys levels of 30, 60, and 80 mL/min/1.73 m2 were 65 (CI, 61 to 70), 44 (CI, 43 to 46), and 39 (CI, 38 to 41) per 100 person-years, respectively, despite small adjusted hazard ratios (Figure 2 and Table 2).

Figure 2.

Figure 2.

Open in a new tab

Adjusted incidence rate per 100 person-years and 95% CIs for eGFRcr-cys for outcomes among persons aged 70 and 80 years. eGFR was modeled using a linear spline with knots at 30, 45, 60, 75, and 90 mL/min/1.73 m2. Adjusted incidence rates were estimated using Poisson regression for mortality and KFRT, and negative binomial regression was used for all other outcomes. AKI = acute kidney injury; eGFR = estimated glomerular filtration rate; eGFRcr = estimated glomerular filtration rate using creatinine level; eGFRcr-cys = estimated glomerular filtration rate using creatinine and cystatin C levels; eGFRcys = estimated glomerular filtration rate using cystatin C level; KFRT = kidney failure with replacement therapy; MI = myocardial infarction.

Table 2.

Adjusted Incidence Rates per 100 Person-Years and 95% CIs for eGFRcr vs. eGFRcr-cys Among Persons Aged 70 and 80 Years at eGFR Levels of 30, 45, 60, and 80 mL/min/1.73 m2

Outcome Age Adjusted Incidence Rate per 100 Person-Years (95% CI), by eGFR Level
30 mL/min/1.73 m2 45 mL/min/1.73 m2 60 mL/min/1.73 m2 80 mL/min/1.73 m2
eGFR cr
 All-cause mortality 70 y 5.9 (5.4–6.5) 4.1 (3.8–4.5) 2.6 (2.4–2.8) 2.4 (2.3–2.5)
80 y 11 (10–12) 8.1 (7.7–8.7) 6.4 (6.0–6.9) 6.7 (6.4–7.0)
 Cardiovascular mortality 70 y 1.3 (1.1–1.6) 0.93 (0.80–1.1) 0.55 (0.47–0.65) 0.47 (0.44–0.51)
80 y 3.0 (2.6–3.4) 2.2 (1.9–2.4) 1.7 (1.5–1.9) 1.6 (1.5–1.8)
 KFRT 70 y 0.20 (0.15–0.26) 0.063 (0.047–0.086) 0.0130 (0.0082–0.021) 0.0077 (0.0050–0.012)
80 y 0.070 (0.052–0.096) 0.012 (0.0070–0.021) 0.0055 (0.0025–0.012) 0.0031 (0.0014–0.0070)
 Hospitalization 70 y 61 (57–66) 53 (50–56) 39 (37–41) 38 (37–39)
80 y 70 (67–73) 58 (56–60) 51 (49–53) 49 (48–51)
 Infection 70 y 23 (20–26) 17 (16–19) 8.9 (8.3–9.6) 8.0 (7.6–8.4)
80 y 29 (27–32) 22 (21–23) 16 (15–17) 16 (15–16)
 MI/stroke 70 y 3.7 (3.0–4.7) 3.9 (3.2–4.7) 2.3 (1.9–2.8) 2.1 (1.9–2.3)
80 y 5.6 (4.7–6.6) 5.1 (4.4–6.0) 3.8 (3.3–4.4) 3.8 (3.4–4.1)
 Heart failure 70 y 12 (10–14) 8.3 (7.2–9.4) 4.5 (4.0–5.1) 3.6 (3.4–3.9)
80 y 18 (16–21) 12 (11–14) 9.2 (8.3–10.3) 7.6 (7.1–8.1)
 AKI 70 y 5.2 (4.1–6.7) 2.8 (2.3–3.4) 1.1 (0.96–1.3) 0.68 (0.61–0.77)
80 y 4.5 (3.8–5.4) 2.7 (2.3–3.1) 1.5 (1.3–1.7) 0.93 (0.83–1.0)
eGFR cr-cys
 All-cause mortality 70 y 6.6 (6.1–7.2) 4.6 (4.2–4.9) 3.3 (3.1–3.6) 2.7 (2.6–2.8)
80 y 12 (11–12) 8.9 (8.3–9.4) 7.4 (7.0–7.9) 6.7 (6.4–6.9)
 Cardiovascular mortality 70 y 1.5 (1.3–1.7) 0.98 (0.85–1.1) 0.66 (0.56–0.78) 0.47 (0.43–0.51)
80 y 3.2 (2.8–3.6) 2.3 (2.1–2.6) 1.9 (1.6–2.1) 1.5 (1.4–1.7)
 KFRT 70 y 0.13 (0.10–0.17) 0.046 (0.033–0.064) 0.015 (0.0091–0.024) 0.0040 (0.0022–0.0073)
80 y 0.038 (0.028–0.053) 0.0087 (0.0047–0.016) 0.0041 (0.0016–0.011) 0.0037 (0.0014–0.0095)
 Hospitalization 70 y 65 (61–70) 57 (54–59) 44 (43–46) 39 (38–41)
80 y 70 (67–73) 60 (58–62) 52 (50–54) 51 (50–53)
 Infection 70 y 27 (24–29) 18 (17–20) 12 (11–13) 8.7 (8.3–9.1)
80 y 32 (30–34) 23 (22–25) 18 (17–19) 16 (15–17)
 MI/stroke 70 y 4.8 (3.9–5.9) 3.9 (3.3–4.7) 2.6 (2.2–3.1) 2.0 (1.8–2.3)
80 y 6.2 (5.3–7.3) 4.8 (4.1–5.5) 4.1 (3.6–4.8) 3.9 (3.6–4.3)
 Heart failure 70 y 15 (13–18) 8.3 (7.3–9.5) 5.6 (4.9–6.4) 3.2 (3.0–3.5)
80 y 20 (18–23) 12 (11–13.5) 9.3 (8.4–10) 7.1 (6.6–7.6)
 AKI 70 y 5.3 (4.3–6.5) 3.1 (2.6–3.7) 1.2 (0.98–1.3) 0.51 (0.45–0.57)
80 y 4.3 (3.7–4.9) 2.8 (2.4–3.1) 1.3 (1.1–1.4) 0.62 (0.54–0.72)
Open in a new tab

AKI = acute kidney injury; eGFR = estimated glomerular filtration rate; eGFRcr = estimated glomerular filtration rate using creatinine level; eGFRcr-cys = estimated glomerular filtration rate using creatinine and cystatin C levels; KFRT = kidney failure with replacement therapy; MI = myocardial infarction.

Use of eGFRcr-cys versus eGFRcr resulted in reclassification of 31.2% of older persons, predominantly to a more severe GFR category (Table 3). For example, 24.0% of persons with eGFRcr between 60 and 89 mL/min/1.73 m2 were reclassified to 45 to 59 mL/min/1.73 m2 with eGFRcr-cys, and 33.4% of those with eGFRcr between 30 and 44 mL/min/1.73 m2 were reclassified to 15 to 29 mL/min/1.73 m2 with eGFRcr-cys. Reclassification to more severe GFR categories was greater for persons aged 75 years or older than for those aged 65 to 74 years (Supplement Table 4, available at Annals.org).

Table 3.

Reclassification Across eGFR Categories for eGFRcr vs. eGFRcr-cys

eGFRcr Category Patients Reclassified Into Different eGFRcr-cys Category, n (%)*
 Total, n
≥90 mL/min/1.73 m2 60–89 mL/min/1.73 m2 45–59 mL/min/1.73 m2 30–44 mL/min/1.73 m2 15–29 mL/min/1.73 m2 <15 mL/min/1.73 m2
≥90 mL/min/1.73 m2 7588 (50.5) 7123 (47.4) 291 (1.9) 21 (0.1) 1 (0.01) – 15 024
60–89 mL/min/1.73 m2 2559 (7.0) 24 137 (65.9) 8796 (24.0) 1141 (3.1) 14 (0.04) – 36 647
45–59 mL/min/1.73 m2 5 (0.0) 1383 (9.0) 7683 (50.0) 6050 (39.4) 249 (1.6) 1 (0.01) 15 371
30–44 mL/min/1.73 m2 – 17 (0.2) 519 (5.3) 6049 (61.2) 3302 (33.4) 5 (0.1) 9892
15–29 mL/min/1.73 m2 – 1 (0.0) 8 (0.2) 182 (4.2) 3673 (84.7) 473 (10.9) 4337
<15 mL/min/1.73 m2 – – – 3 (0.3) 71 (8.0) 809 (91.6) 883
Total, n 10 152 32 661 17 297 13 446 7310 1288 82 154
Open in a new tab

eGFR = estimated glomerular filtration rate; eGFRcr = estimated glomerular filtration rate using creatinine level; eGFRcr-cys = estimated glomerular filtration rate using creatinine and cystatin C levels.

*

Row percentages are shown.

Supporting Analyses

eGFRcr-cys and eGFRcys were also more strongly associated with outcomes than eGFRcr within subgroups of age (65 to 74 and ≥75 years), although the magnitude of hazard and incidence rate ratios for all 3 GFR estimates was lower among those aged 75 years or older than among those aged 65 to 74 years (Supplement Figure 3, available at Annals.org). Similar findings were observed for subgroups of sex, cardiovascular disease, and diabetes (Supplement Figures 4 to 6, available at Annals.org). Among persons with UACR below 30 mg/g, eGFRcr-cys and eGFRcys were also more strongly associated with outcomes than eGFRcr (Supplement Table 3 and Supplement Figure 7, available at Annals.org). For example, adjusted hazard and incidence rate ratios for 60 versus 80 mL/min/1.73 m2 for eGFRcr, eGFRcr-cys, and eGFRcys were 1.0 (CI, 0.9 to 1.1), 1.2 (CI, 1.1 to 1.4), and 1.3 (CI, 1.2 to 1.5) for all-cause mortality and 2.2 (CI, 0.5 to 9.0), 7.3 (CI, 1.4 to 39.7), and 2.2 (CI, 0.3 to 14.2) for KFRT, respectively.

Similar findings were observed when the EKFC equations were used to calculate eGFR (Supplement Tables 5 and 6 and Supplement Figures 8 to 10, available at Annals.org). There were substantial differences in hazard and incidence rate ratios between eGFRcr and eGFRcr-cys for 60 versus 80 mL/min/1.73 m2: 0.8 (CI, 0.8 to 0.9) for eGFRcr versus 1.3 (CI, 1.3 to 1.4) for eGFRcr-cys for all-cause mortality; 1.0 (CI, 0.9 to 1.0) versus 1.7 (CI, 1.5 to 1.8) for cardiovascular mortality; 1.8 (CI, 1.0 to 3.3) versus 3.4 (CI, 1.6 to 7.2) for KFRT; and 2.0 (CI, 1.8 to 2.3) vs. 3.5 (CI, 3.1 to 4.1) for AKI. Results were also consistent when the CKD-EPI 2009 and 2012 equations were used (Supplement Tables 7 and 8 and Supplement Figures 11 and 12, available at Annals.org). Similar results were obtained when incidence rates standardized to the study population were used, with higher absolute incidence rates of all events except KFRT and AKI at older age (Supplement Table 9 and Supplement Figure 13, available at Annals.org).

Discussion

In this study, we examined 82 154 older adults from routine clinical practice with same-day outpatient measurements for creatinine and cystatin C. We found that eGFRcr-cys below 60 mL/min/1.73 m2 had stronger associations with clinical outcomes than eGFRcr. For eGFRcr, U-shaped associations were observed for most outcomes, whereas eGFRcr-cys and eGFRcys showed more linear associations. Findings were consistent across subgroups and when alternative eGFR equations (EKFC, CKD-EPI 2009 or CKD-EPI 2012 with non-Black race coefficient) were used.

These findings inform the discussion on the definition of CKD in older adults (6, 11–16). Age-adapted definitions for CKD have been proposed, with a GFR threshold of 45 mL/min/1.73 m2 for persons aged 65 years or older with albuminuria below 30 mg/g, due to inconsistent associations between eGFRcr levels of 45 to 60 mL/min/1.73 m2 and all-cause mortality (6). Our results show that the absence of associations is likely explained by the use of creatinine as a filtration marker rather than the GFR threshold per se, given that hazard ratios were increased for eGFRcr-cys of 60 mL/min/1.73 m2 compared with the reference of 80 mL/min/1.73 m2. Furthermore, we studied a wider range of outcomes than previous studies, which focused solely on kidney failure and mortality (8, 10). Indeed, we observed strong associations between eGFR and hospitalization with AKI or heart failure at the GFR threshold of less than 60 mL/min/1.73 m2. These data indicate that CKD stage G3+ (GFR <60 mL/min/1.73 m2) at older age is associated with a wider range of outcomes than previously recognized, even in the absence of albuminuria.

Our study found U-shaped associations between eGFRcr and outcomes, with higher risks with eGFRcr above 90 mL/min/1.73 m2, which is similar to findings from previous epidemiologic studies (9, 46). In contrast, associations for eGFRcr-cys and eGFRcys were more linear (47–49), which is more biologically plausible. These differences are due to non-GFR determinants that influence creatinine and cystatin C levels (17, 18, 25, 50). Besides GFR, creatinine is also influenced by muscle mass, diet, and physical activity, whereas cystatin C is influenced by inflammation, obesity, smoking, thyroid diseases, and glucocorticoid use (17, 18, 25, 51). The higher risk with eGFRcr above 90 mL/min/1.73 m2 is likely attributable to persons with low muscle mass and poorer health status. These persons have low creatinine levels leading to high eGFRcr that is an overestimate of their true GFR, and they are also at high risk for adverse outcomes. This also explains the weaker associations for eGFRcr below 60 mL/min/1.73 m2 compared with eGFRcr-cys. Even though the same reference value for eGFR (80 mL/min/1.73 m2) is applied for each eGFR measure, the characteristics of the persons who are included in the reference value differ. Thus, eGFRcr of 80 mL/min/1.73 m2 likely includes more people with low muscle mass and poor health status than eGFRcr-cys of 80 mL/min/1.73 m2. Similarly, the stronger associations observed with eGFRcys compared with eGFRcr-cys may not reflect GFR per se but rather the non-GFR determinants of cystatin C. Combining both markers in eGFRcr-cys improves precision by reducing errors that are due to variation in the non-GFR determinants of each marker (25).

Although the 2012 KDIGO guideline for the evaluation and management of CKD recommends measuring cystatin C when eGFRcr may be less accurate due to low muscle mass (5), this practice remains limited in most countries, including the United States (52), and also among older adults. Sweden is unique in this regard because cystatin C is widely used as a supportive test in routine clinical practice (at a cost of about $3 per test), as evidenced by the fact that one fifth of older adults in our cohort who had an outpatient creatinine measurement also had a same-day cystatin C measurement. The National Kidney Foundation and the American Society of Nephrology recently recommended increased use of cystatin C to estimate GFR in the United States (24), based on the demonstration that eGFRcr-cys more accurately approximates measured GFR than eGFRcr or eGFRcys (18, 22, 53). We showed that more than 30% of older adults would be reclassified to a different GFR category if eGFRcr were to be replaced by eGFRcr-cys, predominantly to more severe GFR categories. Given that many treatment recommendations are based on GFR thresholds, implementation of cystatin C testing would have a significant effect on clinical practice. Future studies should investigate whether implementing cystatin C testing improves outcomes, such as through better drug dosing or more timely nephrologist referral.

Strengths of our study include its large sample size and contemporary data (2010 to 2019) involving more than 80 000 patients from routine care in a country with a long-standing history of cystatin C testing. We also restricted the study to the periods when measurements of creatinine and cystatin C were standardized and calibrated to traceable international standards. We investigated a wide range of clinically meaningful outcomes and analyzed recurrent events for outcomes that could occur more than once, which is especially relevant in older patients, who may have multiple hospitalizations.

Our study also has several limitations. First, we lacked GFR measurements, and future research should investigate associations between measured GFR and outcomes. Our analyses adjusted for cardiovascular risk factors but did not adjust for all non-GFR determinants that may influence creatinine and cystatin C levels. For example, muscle mass, diet, obesity status or body mass index, smoking, and inflammation were not available in our data sources. eGFR based on cystatin C alone may therefore overestimate risk associations with adverse outcomes compared with measured GFR. Second, only a subset of older adults with a creatinine test had a same-day cystatin C test (19%), and these persons were older and had a higher prevalence of diabetes and cardiovascular disease. It is reassuring to note that results were similar across subgroups based on age and comorbidities. Third, outcomes were based on diagnosis codes, which have high specificity but low sensitivity, particularly for AKI. Fourth, we investigated associations based on 1 eGFR measurement rather than 2 measurements more than 90 days apart, which is used in KDIGO guidelines to define CKD. However, selecting only persons with at least 2 measurements of creatinine and cystatin C would lead to highly selected samples in routinely collected health care data and would preclude studying eGFR as a continuous exposure (38, 54). Fifth, the measurement of cystatin C may have led to a change in subsequent treatment decisions. However, this would similarly affect the associations for eGFRcr, eGFRcys, and eGFRcr-cys and outcomes. Finally, we lacked information on race. According to Swedish government annual statistics, our study population consisted of predominantly White participants. Future studies should generalize our findings to populations with different racial or ethnic composition. For example, previous studies in the general population have indicated that differences between hazard ratios for eGFRcr vs. eGFRcr-cys may be even larger among Black persons (55).

In conclusion, low eGFRcr-cys was more strongly associated with a broad range of outcomes than low eGFRcr among older patients, including all-cause and cardiovascular mortality and many specific types of hospitalizations. eGFRcr above 90 mL/min/1.73 m2 was associated with higher risk, but not eGFRcr-cys or eGFRcys, suggesting the risk may be driven by low creatinine generation and low muscle mass rather than high GFR itself. The broad range of risks associated with CKD at older age is better appreciated when cystatin C is included in GFR estimation.

Supplementary Material

Supplement

Grant Support:

This study was supported by the Swedish Research Council (2023–01807), the Swedish Heart-Lung Foundation (20230371), National Institute of Diabetes and Digestive and Kidney Diseases grants R01DK100446 and R01DK115534, and a Junior Kolff Grant from the Dutch Kidney Foundation (22OK2026). Dr. Fu is supported by a Rubicon grant from the Netherlands Organisation for Scientific Research (452021107) and an internal funding grant from Karolinska Institute. Dr. Ishigami is supported by National Institutes of Health (NIH) grant K01DK125616. Dr. Grams is supported by NIH grant K24HL155861.

Footnotes

Note: Drs. Ballew and Coresh had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Reproducible Research Statement: Study protocol: Not available. Statistical code: Available in the Supplement. Data set: The data underlying this article cannot be shared publicly because of concerns about the privacy of the study participants. The data may be shared on reasonable request to Prof. Carrero ( juan.jesus.carrero@ki.se) for academic research collaborations that comply with the General Data Protection Regulation as well as national and institutional ethics regulations and standards.

Contributor Information

Edouard L. Fu, Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Stockholm, Sweden; Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Massachusetts; Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, the Netherlands.

Juan-Jesus Carrero, Department of Medical Epidemiology and Biostatistics, Karolinska Institute, and Division of Nephrology, Department of Clinical Sciences, Karolinska Institute, Danderyd Hospital, Stockholm, Sweden.

Yingying Sang, Optimal Aging Institute and Division of Epidemiology, Department of Population Health, New York University Grossman School of Medicine, New York, New York.

Marie Evans, Department of Clinical Intervention and Technology, Karolinska University Hospital and Karolinska Institute, Stockholm, Sweden.

Junichi Ishigami, Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland.

Lesley A. Inker, Division of Nephrology, Department of Internal Medicine, Tufts Medical Center, Boston, Massachusetts.

Morgan E. Grams, Division of Precision Medicine, Department of Medicine, New York University Grossman School of Medicine, New York, New York.

Andrew S. Levey, Division of Nephrology, Department of Internal Medicine, Tufts Medical Center, Boston, Massachusetts.

Josef Coresh, Optimal Aging Institute and Division of Epidemiology, Department of Population Health, New York University Grossman School of Medicine, New York, New York; Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland.

Shoshana H. Ballew, Optimal Aging Institute and Division of Epidemiology, Department of Population Health, New York University Grossman School of Medicine, New York, New York; Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, Maryland.

References

  • 1.Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney disease in the United States. JAMA. 2007;298:2038–2047. [ 10.1001/jama.298.17.2038] [17986697] [DOI] [PubMed] [Google Scholar]
  • 2.Zhang QL, Rothenbacher D. Prevalence of chronic kidney disease in population-based studies: systematic review. BMC Public Health. 2008;8:117. [ 10.1186/1471-2458-8-117] [18405348] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Stevens LA, Li S, Wang C, et al. Prevalence of CKD and comorbid illness in elderly patients in the United States: results from the Kidney Early Evaluation Program (KEEP). Am J Kidney Dis. 2010;55:S23–33. [ 10.1053/j.ajkd.2009.09.035] [20172445] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Levey AS, de Jong PE, Coresh J, et al. The definition, classification, and prognosis of chronic kidney disease: a KDIGO Controversies Conference report. Kidney Int. 2011;80:17–28. [ 10.1038/ki.2010.483] [21150873] [DOI] [PubMed] [Google Scholar]
  • 5.Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int Suppl. 2013;3:1–150. doi: 10.1038/kisup.2012.76 [DOI] [Google Scholar]
  • 6.Delanaye P, Jager KJ, Bökenkamp A, et al. CKD: a call for an age-adapted definition. J Am Soc Nephrol. 2019;30:1785–1805. [ 10.1681/ASN.2019030238] [31506289] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Denic A, Glassock RJ, Rule AD. Structural and functional changes with the aging kidney. Adv Chronic Kidney Dis. 2016;23:19–28. [ 10.1053/j.ackd.2015.08.004] [26709059] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Hallan SI, Matsushita K, Sang Y, et al. ; Chronic Kidney Disease Prognosis Consortium. Age and association of kidney measures with mortality and end-stage renal disease. JAMA. 2012;308:2349–2360. [ 10.1001/jama.2012.16817] [23111824] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Matsushita K, van der Velde M, Astor BC, et al. ; Chronic Kidney Disease Prognosis Consortium. Association of estimated glomerular filtration rate and albuminuria with all-cause and cardiovascular mortality in general population cohorts: a collaborative meta-analysis. Lancet. 2010;375:2073–2081. [ 10.1016/S0140-6736(10)60674-5] [20483451] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Liu P, Quinn RR, Lam NN, et al. Accounting for age in the definition of chronic kidney disease. JAMA Intern Med. 2021;181:1359–1366. [ 10.1001/jamainternmed.2021.4813] [34459844] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Glassock R, Delanaye P, El Nahas M. An age-calibrated classification of chronic kidney disease. JAMA. 2015;314:559–560. [ 10.1001/jama.2015.6731] [26023760] [DOI] [PubMed] [Google Scholar]
  • 12.Levey AS, Inker LA, Coresh J. Chronic kidney disease in older people. JAMA. 2015;314:557–558. [ 10.1001/jama.2015.6753] [26023868] [DOI] [PubMed] [Google Scholar]
  • 13.Glassock RJ, Winearls C. CKD—fiction not fact. Nephrol Dial Transplant. 2008;23:2695–2621. [ 10.1093/ndt/gfn331] [18567889] [DOI] [PubMed] [Google Scholar]
  • 14.Coresh J, Stevens LA, Levey AS. Chronic kidney disease is common: what do we do next? Nephrol Dial Transplant. 2008;23:1122–1125. [ 10.1093/ndt/gfn117] [18359871] [DOI] [PubMed] [Google Scholar]
  • 15.Glassock RJ, Delanaye P, Rule AD. Should the definition of CKD be changed to include age-adapted GFR criteria? YES. Kidney Int. 2020;97:34–37. [ 10.1016/j.kint.2019.08.033] [31901354] [DOI] [PubMed] [Google Scholar]
  • 16.Levey AS, Inker LA, Coresh J. “Should the definition of CKD be changed to include age-adapted GFR criteria?”: Con: the evaluation and management of CKD, not the definition, should be age-adapted. Kidney Int. 2020;97:37–40. [ 10.1016/j.kint.2019.08.032] [31901355] [DOI] [PubMed] [Google Scholar]
  • 17.Inker LA, Titan S. Measurement and estimation of GFR for use in clinical practice: core curriculum 2021. Am J Kidney Dis. 2021;78:736–749. [ 10.1053/j.ajkd.2021.04.016] [34518032] [DOI] [PubMed] [Google Scholar]
  • 18.Levey AS, Grams ME, Inker LA. Uses of GFR and albuminuria level in acute and chronic kidney disease. N Engl J Med. 2022;386:2120–2128. [ 10.1056/NEJMra2201153] [35648704] [DOI] [PubMed] [Google Scholar]
  • 19.Foley RN, Wang C, Ishani A, et al. Kidney function and sarcopenia in the United States general population: NHANES III. Am J Nephrol. 2007;27:279–286. [ 10.1159/000101827] [17440263] [DOI] [PubMed] [Google Scholar]
  • 20.Grams ME, Coresh J, Matsushita K, et al. ; Writing Group for the CKD Prognosis Consortium. Estimated glomerular filtration rate, albuminuria, and adverse outcomes: an individual-participant data meta-analysis. JAMA. 2023;330:1266–1277. [ 10.1001/jama.2023.17002] [37787795] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Schaeffner ES, Ebert N, Delanaye P, et al. Two novel equations to estimate kidney function in persons aged 70 years or older. Ann Intern Med. 2012;157:471–481. [ 10.7326/0003-4819-157-7-201210020-00003] [23027318] [DOI] [PubMed] [Google Scholar]
  • 22.Inker LA, Eneanya ND, Coresh J, et al. ; Chronic Kidney Disease Epidemiology Collaboration. New creatinine- and cystatin C–based equations to estimate GFR without race. N Engl J Med. 2021;385:1737–1749. [ 10.1056/NEJMoa2102953] [34554658] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Fu EL, Levey AS, Coresh J, et al. Accuracy of GFR estimating equations in patients with discordances between creatinine and cystatin C–based estimations. J Am Soc Nephrol. 2023;34:1241–1251. [ 10.1681/ASN.0000000000000128] [36995139] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Delgado C, Baweja M, Crews DC, et al. A unifying approach for GFR estimation: recommendations of the NKF-ASN task force on reassessing the inclusion of race in diagnosing kidney disease. J Am Soc Nephrol. 2021;32:2994–3015. [ 10.1681/ASN.2021070988] [34556489] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Levey AS, Coresh J, Tighiouart H, et al. Measured and estimated glomerular filtration rate: current status and future directions. Nat Rev Nephrol. 2020;16:51–64. [ 10.1038/s41581-019-0191-y] [31527790] [DOI] [PubMed] [Google Scholar]
  • 26.Riley RD, Hayden JA, Steyerberg EW, et al. ; PROGRESS Group. Prognosis Research Strategy (PROGRESS) 2: prognostic factor research. PLoS Med. 2013;10:e1001380. [ 10.1371/journal.pmed.1001380] [23393429] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Riley RD, Van der Windt D, Croft P, et al. Prognosis research in healthcare: concepts, methods, and impact. Oxford Univ Pr; 2019. [Google Scholar]
  • 28.Brotman DJ, Bash LD, Qayyum R, et al. Heart rate variability predicts ESRD and CKD-related hospitalization. J Am Soc Nephrol. 2010;21:1560–1570. [ 10.1681/ASN.2009111112] [20616169] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Carrero JJ, Elinder CG. The Stockholm CREAtinine Measurements (SCREAM) project: fostering improvements in chronic kidney disease care. J Intern Med. 2022;291:254–268. [ 10.1111/joim.13418] [35028991] [DOI] [PubMed] [Google Scholar]
  • 30.Wettermark B, Hammar N, Fored CM, et al. The new Swedish Prescribed Drug Register—opportunities for pharmacoepidemiological research and experience from the first six months. Pharmacoepidemiol Drug Saf. 2007;16:726–735. [ 10.1002/pds.1294] [16897791] [DOI] [PubMed] [Google Scholar]
  • 31.Ludvigsson JF, Andersson E, Ekbom A, et al. External review and validation of the Swedish National Inpatient Register. BMC Public Health. 2011;11:450. [ 10.1186/1471-2458-11-450] [21658213] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Brooke HL, Talbäck M, Hörnblad J, et al. The Swedish cause of death register. Eur J Epidemiol. 2017;32:765–773. [ 10.1007/s10654-017-0316-1] [28983736] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Grubb A, Blirup-Jensen S, Lindstrom V, et al. ; IFCC Working Group on Standardisation of Cystatin C (WG-SCC). First certified reference material for cystatin C in human serum ERM-DA471/IFCC. Clin Chem Lab Med. 2010;48:1619–1621. [ 10.1515/CCLM.2010.318] [21034257] [DOI] [PubMed] [Google Scholar]
  • 34.Inker LA, Schmid CH, Tighiouart H, et al. ; CKD-EPI Investigators. Estimating glomerular filtration rate from serum creatinine and cystatin C. N Engl J Med. 2012;367:20–29. [ 10.1056/NEJMoa1114248] [22762315] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Sumida K, Nadkarni GN, Grams ME, et al. ; Chronic Kidney Disease Prognosis Consortium. Conversion of urine protein-creatinine ratio or urine dipstick protein to urine albumin-creatinine ratio for use in chronic kidney disease screening and prognosis: an individual participant-based meta-analysis. Ann Intern Med. 2020;173:426–435. [ 10.7326/M20-0529] [32658569] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36.Eriksson A, Stenlund H, Ahlm K, et al. Accuracy of death certificates of cardiovascular disease in a community intervention in Sweden. Scand J Public Health. 2013;41:883–889. [ 10.1177/1403494813499653] [23982462] [DOI] [PubMed] [Google Scholar]
  • 37.Emilsson L, Lindahl B, Köster M, et al. Review of 103 Swedish healthcare quality registries. J Intern Med. 2015;277:94–136. [ 10.1111/joim.12303] [25174800] [DOI] [PubMed] [Google Scholar]
  • 38.Carrero JJ, Fu EL, Vestergaard SV, et al. Defining measures of kidney function in observational studies using routine health care data: methodological and reporting considerations. Kidney Int. 2023;103:53–69. [ 10.1016/j.kint.2022.09.020] [36280224] [DOI] [PubMed] [Google Scholar]
  • 39.Levey AS, Stevens LA, Schmid CH, et al. ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration). A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. [ 10.7326/0003-4819-150-9-200905050-00006] [19414839] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Pottel H, Björk J, Rule AD, et al. Cystatin C–based equation to estimate GFR without the inclusion of race and sex. N Engl J Med. 2023;388:333–343. [ 10.1056/NEJMoa2203769] [36720134] [DOI] [PubMed] [Google Scholar]
  • 41.Pottel H, Björk J, Courbebaisse M, et al. Development and validation of a modified full age spectrum creatinine-based equation to estimate glomerular filtration rate: a cross-sectional analysis of pooled data. Ann Intern Med. 2021;174:183–191. [ 10.7326/M20-4366] [33166224] [DOI] [PubMed] [Google Scholar]
  • 42.Fu EL, Coresh J, Grams ME, et al. Removing race from the CKD-EPI equation and its impact on prognosis in a predominantly White European population. Nephrol Dial Transplant. 2023;38:119–128. [ 10.1093/ndt/gfac197] [35689668] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Grubb A, Horio M, Hansson LO, et al. Generation of a new cystatin C–based estimating equation for glomerular filtration rate by use of 7 assays standardized to the international calibrator. Clin Chem. 2014;60:974–986. [ 10.1373/clinchem.2013.220707] [24829272] [DOI] [PubMed] [Google Scholar]
  • 44.Björk J, Grubb A, Sterner G, et al. Revised equations for estimating glomerular filtration rate based on the Lund-Malmö Study cohort. Scand J Clin Lab Invest. 2011;71:232–239. [ 10.3109/00365513.2011.557086] [21391777] [DOI] [PubMed] [Google Scholar]
  • 45.Horio M, Imai E, Yasuda Y, et al. ; Collaborators Developing the Japanese Equation for Estimated GFR. GFR estimation using standardized serum cystatin C in Japan. Am J Kidney Dis. 2013;61:197–203. [ 10.1053/j.ajkd.2012.07.007] [22892396] [DOI] [PubMed] [Google Scholar]
  • 46.Shlipak MG, Matsushita K, Ärnlöv J, et al. ; CKD Prognosis Consortium. Cystatin C versus creatinine in determining risk based on kidney function. N Engl J Med. 2013;369:932–943. [ 10.1056/NEJMoa1214234] [24004120] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Shlipak MG, Katz R, Sarnak MJ, et al. Cystatin C and prognosis for cardiovascular and kidney outcomes in elderly persons without chronic kidney disease. Ann Intern Med. 2006;145:237–246. [ 10.7326/0003-4819-145-4-200608150-00003] [16908914] [DOI] [PubMed] [Google Scholar]
  • 48.Sarnak MJ, Katz R, Stehman-Breen CO, et al. ; Cardiovascular Health Study. Cystatin C concentration as a risk factor for heart failure in older adults. Ann Intern Med. 2005;142:497–505. [ 10.7326/0003-4819-142-7-200504050-00008] [15809461] [DOI] [PubMed] [Google Scholar]
  • 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;352:2049–2060. [ 10.1056/NEJMoa043161] [15901858] [DOI] [PubMed] [Google Scholar]
  • 50.Tangri N, Inker LA, Tighiouart H, et al. Filtration markers may have prognostic value independent of glomerular filtration rate. J Am Soc Nephrol. 2012;23:351–359. [ 10.1681/ASN.2011070663] [22173699] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51.Odden MC, Chertow GM, Fried LF, et al. ; HABC Study. Cystatin C and measures of physical function in elderly adults: the Health, Aging, and Body Composition (HABC) Study. Am J Epidemiol. 2006;164:1180–1189. [ 10.1093/aje/kwj333] [17035344] [DOI] [PubMed] [Google Scholar]
  • 52.Ebert N, Shlipak MG. Cystatin C is ready for clinical use. Curr Opin Nephrol Hypertens. 2020;29:591–598. [ 10.1097/MNH.0000000000000638] [32868529] [DOI] [PubMed] [Google Scholar]
  • 53.Han J, Zhao Y, Canney M, et al. Are patients with primary glomerular disease at increased risk of malignancy? Nephrol Dial Transplant. 2023:gfad261. [ 10.1093/ndt/gfad261] [38070875] [DOI] [PubMed] [Google Scholar]
  • 54.Mazhar F, Sjölander A, Fu EL, et al. Estimating the prevalence of chronic kidney disease while accounting for nonrandom testing with inverse probability weighting. Kidney Int. 2023;103:416–420. [ 10.1016/j.kint.2022.10.027] [36462535] [DOI] [PubMed] [Google Scholar]
  • 55.Gutiérrez OM, Sang Y, Grams ME, et al. ; Chronic Kidney Disease Prognosis Consortium. Association of estimated GFR calculated using race-free equations with kidney failure and mortality by Black vs non-Black race. JAMA. 2022;327:2306–2316. [ 10.1001/jama.2022.8801] [35667006] [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement
NIHMS1988167-supplement-Supplement.docx (19.5MB, docx)

RESOURCES