Clinical Epidemiology

Past Decline Versus Current eGFR and Subsequent ESRD Risk

Csaba P. Kovesdy, Josef Coresh, Shoshana H. Ballew, Mark Woodward, Adeera Levin, David M.J. Naimark, Joseph Nally, Dietrich Rothenbacher, Benedicte Stengel, Kunitoshi Iseki, Kunihiro Matsushita and Andrew S. Levey
JASN August 2016, 27 (8) 2447-2455; DOI: https://doi.org/10.1681/ASN.2015060687

Abstract

eGFR is a robust predictor of ESRD risk. However, the prognostic information gained from the past trajectory (slope) beyond that of the current eGFR is unclear. We examined 22 cohorts to determine the association of past slopes and current eGFR level with subsequent ESRD. We modeled hazard ratios as a spline function of slopes, adjusting for demographic variables, eGFR, and comorbidities. We used random effects meta–analyses to combine results across studies stratified by cohort type. We calculated the absolute risk of ESRD at 5 years after the last eGFR using the weighted average baseline risk. Overall, 1,080,223 participants experienced 5163 ESRD events during a mean follow-up of 2.0 years. In CKD cohorts, a slope of −6 versus 0 ml/min per 1.73 m2 per year over the previous 3 years (a decline of 18 ml/min per 1.73 m2 versus no decline) associated with an adjusted hazard ratio of ESRD of 2.28 (95% confidence interval, 1.88 to 2.76). In contrast, a current eGFR of 30 versus 50 ml/min per 1.73 m2 (a difference of 20 ml/min per 1.73 m2) associated with an adjusted hazard ratio of 19.9 (95% confidence interval, 13.6 to 29.1). Past decline contributed more to the absolute risk of ESRD at lower than higher levels of current eGFR. In conclusion, during a follow-up of 2 years, current eGFR associates more strongly with future ESRD risk than the magnitude of past eGFR decline, but both contribute substantially to the risk of ESRD, especially at eGFR<30 ml/min per 1.73 m2.

CKD is characterized by poor outcomes and high costs, and its increasing worldwide prevalence represents a significant public health challenge.1 Although the vast majority of patients with CKD have early-stage disease,2–4 patients with late-stage disease and especially, those with ESRD suffer from an especially high burden of comorbid conditions, have extremely poor outcomes, and consume a disproportionate amount of health care resources.5 It is, thus, important to focus interventions, such as efforts to slow kidney progression and preparation for the transition to ESRD, on patients who are most prone to experience a progressive disease course. Recent studies have emphasized the importance of eGFR and albuminuria as measures of kidney disease severity, which can be assessed at the point of clinical contact, used to classify patients into various stages of CKD and form the basis of clinical interventions.2 However, many other factors influence the rate of progression, including age,6,7 comorbid conditions, such as diabetes mellitus or hypertension,7–9 race-ethnicity,10 and genetic mutations.11 Nonetheless, these factors do not account for the observed variability in kidney disease progression.12

A seminal study almost 40 years ago proposed that the trajectory of kidney disease progression could be predicted from the past rate of decline in kidney function.13 Current clinical practice guidelines continue to recommend assessing the future risk of kidney disease progression from the past slope of eGFR over time.2 Despite the widespread acceptance of this practice, relatively few studies have evaluated the past eGFR decline as a predictor of ESRD after taking into account the current level of eGFR.14 In clinical practice, both the level of eGFR at the point of assessment and its past trajectory over time are readily available for assessment, but the relative contribution of each to the risk of subsequent ESRD is not clear. We, thus, examined the magnitudes of associations of past decline in eGFR over 3 years versus current level of eGFR at the end of the slope evaluation period with subsequent progression to ESRD in 22 large and diverse cohort studies from across the globe to investigate the usefulness of both measures as predictors of CKD progression in clinical practice.

Results

We primarily provide results on the 3-year slope evaluation period and the absolute risk of ESRD at 5 years of follow-up in the CKD cohorts, and we present results for other baseline (1- and 2-year slopes) and follow-up periods (1, 3, and 10 years) and the other (general population and high risk) cohorts in the Supplemental Material. Twenty of 22 participating cohorts (11 CKD and nine other cohorts) provided data on change in eGFR for a slope evaluation period of 3 years. Among 1,080,223 participants (232,250 in CKD cohorts and 847,973 in other cohorts), approximately 10% and 5% had rapid decline (slope <−5 ml/min per 1.73 m2 per year) or rapid increase (slope >5 ml/min per 1.73 m2 per year), respectively, with the remaining 85% having less rapid changes (slope ≥−5 to ≤5 ml/min per 1.73 m2 per year) (Supplemental Table 1, Table 1). Individuals with rapid decline tended to have a poorer risk profile (higher prevalence of albuminuria, diabetes, and history of CVD) compared with those with rapid increase and less rapid changes, regardless of cohort types (Supplemental Tables 1–5, Table 1).

View this table:
Table 1.

Three-year baseline period characteristics by slope category in CKD cohorts

We observed a total of 5163 ESRD events (3256 in CKD cohorts and 1907 in other cohorts) during a mean subsequent follow-up period of 2.0 years after the 3-year baseline period (Supplemental Tables 2 and 6, Table 2). In CKD cohorts, the subsequent risk of ESRD showed overall higher adjusted hazard ratios (HRs) at both greater negative and positive slopes of eGFR compared with stable eGFR (slope of 0 ml/min per 1.73 m2 per year) (Figure 1). This finding was most pronounced for 3-year slopes, being slightly weaker for slopes assessed over shorter (1 and 2 years) baseline periods (Supplemental Figure 1). Slopes of −6 and −3 ml/min per 1.73 m2 per year over 3 years (−18 and −9 ml/min per 1.73 m2 per year over 3 years) were associated with adjusted HRs of ESRD of 2.28 (95% confidence interval [95% CI], 1.88 to 2.76) and 1.73 (95% CI, 1.50 to 2.00), respectively. Other cohorts displayed similar trends (Supplemental Figure 2). Additional adjustment for albuminuria did not alter the results substantially (Supplemental Figures 3 and 4). Furthermore, results were largely consistent across individual cohorts (Figure 2, Supplemental Figure 5). Associations were similar in CKD cohorts for patients exposed to renin-angiotensin-aldosterone system (RAAS) inhibitors and those not exposed to such agents. In other cohorts, the association of slope with ESRD was significant in patients who received RAAS inhibitors but not in patients who were not exposed to such agents (Supplemental Figure 6). Participants with a rapid rise in eGFR also had an elevated ESRD risk, but the number of events in this group was small (n=18).

View this table:
Table 2.

Three-year baseline period events by slope category in CKD cohorts

Figure 1.
Figure 1.

Adjusted HR of ESRD associated with slope of eGFR during a 3-year baseline period and a histogram of the slope of eGFR in CKD cohorts. Values were trimmed at a −15-ml slope (0.3%) and a 10-ml slope (1.1%). Black dots indicate statistical significance compared with the reference (diamond) slope of eGFR=0 ml/min per 1.73 m2 per year. Open circles show slope of eGFR=−6 and −3 ml/min per 1.73 m2 per year.

Figure 2.
Figure 2.

Adjusted relative HRs of ESRD for a 6-ml/min per 1.73 m2 per year decline and a 3-ml/min per 1.73 m2 per year decline in eGFR (compared with a decline of 0 ml/min per 1.73 m2 per year) during a 3-year baseline period in CKD cohorts. The left panel shows adjusted relative HRs for a 6-ml/min per 1.73 m2 per year decline and the right panel shows adjusted relative HRs for a 3-ml/min per 1.73 m2 per year decline. AASK, African American Study of Kidney Disease and Hypertension; BC CKD, British Columbia CKD Study; CCF, Cleveland Clinic CKD Registry Study; Geisinger, Geisinger CKD Study; GLOMMS1, Grampian Laboratory Outcomes, Morbidity and Mortality Studies 1; MASTERPLAN, Multifactorial Approach and Superior Treatment Efficacy in Renal Patients with the Aid of a Nurse Practitioner; MDRD, Modification of Diet in Renal Disease Study; NephroTest, NephroTest Study; Sunnybrook, Sunnybrook Cohort; VA CKD, Veterans Administration CKD Study.

In analyses stratified by level of last eGFR, the risk of ESRD was always significantly associated with more rapid declines, but the magnitude of the excess risk was much less than that imparted by lower versus higher levels of last recorded eGFR (Figure 3A). The HRs of ESRD associated with eGFR levels of 20, 30, and 40 ml/min per 1.73 m2 (compared with 50 ml/min per 1.73 m2 and a slope of 0 ml/min per 1.73 m2 per year) were 216.8 (95% CI, 124.8 to 376.7), 46.4 (95% CI, 31.9 to 67.6), and 9.99 (95% CI, 8.03 to 12.44) in those with a slope of −6 ml/min per 1.73 m2 per year, respectively; 178.3 (95% CI, 92.9 to 342.2), 38.4 (95% CI, 23.8 to 61.9), and 8.11 (95% CI, 5.98 to 11.00) in participants with a slope of −3 ml/min per 1.73 m2 per year, respectively; and 88.7 (95% CI, 50.0 to 157.3), 19.9 (95% CI, 13.6 to 29.1), and 4.46 (95% CI, 3.69 to 5.40) in participants with a slope of 0 ml/min per 1.73 m2 per year, respectively (overall P for interaction between eGFR slope and last eGFR was 0.26). Consequently, both the slopes and the eGFR levels were independently associated with higher 5-year estimated absolute risk of ESRD (Figure 3B, Supplemental Table 7). The higher estimated absolute risk of ESRD associated with steeper declines in eGFR seemed to be more pronounced at lower levels of eGFR. Of note, at eGFR<30 ml/min per 1.73 m2, the estimated absolute risk of ESRD was substantial, even at a slope of 0 ml/min per 1.73 m2 per year. Results displayed similar trends when assessing 1- and 2-year slopes and when examining other cohorts (Supplemental Figures 7–11).

Figure 3.
Figure 3.

Adjusted HRs (95% CIs; reference: patients with eGFR=50 ml/min per 1.73 m2 and slope of 0 ml/min per 1.73 m2 per year) and absolute risks of ESRD associated with slope of eGFR and different levels of last eGFR during a 3-year baseline period in CKD cohorts. Panel A shows the adjusted HRs and panel B shows the absolute risks.

Discussion

In this international meta-analysis of 1,080,223 participants in 22 diverse cohorts, approximately 10% of participants had experienced past rapid eGFR declines (slopes) of <−5 ml/min per 1.73 m2 per year over 1–3 years before the current eGFR assessment. We observed a significant and independent association of both a lower current level of eGFR and a more rapid past decline in eGFR (slope) with higher subsequent risk of ESRD in CKD cohorts, especially when slopes were calculated from creatinine levels measured over 3 years. For example, a slope of −6- versus 0-ml/min per 1.73 m2 per year change over 3 years (an 18-ml/min per 1.73 m2 per year decline in total) was associated with adjusted HR of subsequent ESRD of 2.28 (95% CI, 1.88 to 2.76), whereas a current eGFR of 30 versus 50 ml/min per 1.73 m2 (a 20-ml/min per 1.73 m2 difference in final eGFR) was associated with an adjusted HR of subsequent ESRD of 19.9 (95% CI, 13.6 to 29.1) if the previous slope was 0 ml/min per 1.73 m2. Other cohorts displayed similar trends but with substantially higher heterogeneity and lack of statistical significance, indicating that our conclusions primarily refer to patients with CKD. The current level of eGFR seemed to be associated with a larger risk of ESRD, especially at very low eGFR, where the ESRD risk was substantial, even with past slopes of 0 ml/min per 1.73 m2 per year. However, more rapid past declines contributed substantially to significantly higher absolute risk of ESRD, especially in individuals with very low eGFR level. An additional finding in our study was that approximately 5% of participants had experienced a past rapid eGFR rise (>5 ml/min per 1.73 m2 per year) and that a more rapid past rise in eGFR was associated significantly and independently with a higher ESRD risk. The results are consistent with the current body of evidence indicating that point estimates of eGFR are one of the most robust predictors of ESRD, but they also provide new evidence that the past trajectory of eGFR slopes could be used in addition to contemporarily evaluated other risk factors to assess future risk of ESRD.

The level of eGFR is a well established predictor of ESRD,2 but the role of past eGFR trajectories (slopes) in the assessment of patients with kidney disease has been less clear. Most previous studies have evaluated the association between slopes of eGFR and mortality,15–21 and a few assessed the role of future slopes (i.e., after the point of assessment adjusted for the initial level of eGFR) in predicting ESRD.22,23 The association of future slopes or percentage changes in eGFR with ESRD relates to their role as surrogate end points in clinical trials.24 In contrast, past slopes of eGFR are helpful to clinicians, who need information readily available at the point of contact for clinical decision making and future projection, and their association with ESRD events has not been previously extensively evaluated. Past slopes of eGFR provide an empirical measure of the disease process for an individual, which represents the aggregate effect of all known and unknown predictors of kidney disease progression for each patient. A prior community–based study found that the association between past slopes of eGFR and ESRD was attenuated and became nonsignificant after adjusting for the level of the last eGFR.14 Our study confirms the major association between eGFR level and the risk of ESRD but suggests that past slopes may also have an independent, albeit weaker association with this end point. Our study may have been better suited than the previous study for a detailed evaluation of the role of past eGFR trajectories because of the much higher statistical power imparted by the large number of outcomes, the availability of longer evaluation periods for slope estimations, and the diversity of patient populations with representation of higher–risk patient groups.

The finding of substantial risk of ESRD associated with very low eGFR in the absence of eGFR decline in the past 1–3 years suggests possible difficulty in accurate ascertainment of eGFR trajectory over a short interval, especially in patients with slowly progressive or nonlinear eGFR declines.25 The much stronger relative risk of ESRD associated with lower levels of eGFR versus steeper slopes may also be related to the fact that the studied end point (ESRD) is directly dependent on a very low eGFR but less dependent on rapid eGFR decline; hence, during the relatively short follow-up period of our study, it is much more likely for ESRD to be observed in patients who start follow-up with lower eGFR levels, and it is possible that more rapid declines in eGFR could have been stronger predictors of ESRD if patients were followed for a longer period of time. Nevertheless, in clinical practice, prediction of ESRD is most important for the immediately foreseeable future, because clinicians need to implement preparations, such as vascular access planning and referral for transplantation, during the 6–12 months preceding ESRD. These results suggest that interventions that slow kidney disease progression and preparations for ESRD should be continuously implemented in patients with CKD stages 4 and 5, even in the absence of demonstrable eGFR decline in the past 1–3 years.

The finding that a more rapid past eGFR rise is associated with higher ESRD risk is reminiscent of associations between positive slopes and higher mortality in previous studies.16,18 The explanation for these seemingly counterintuitive associations is unclear but could be because of confounding by loss of muscle mass, volume overload, presence of severe illness with an underlying heightened propensity for faster kidney disease progression, or recovery from previous AKI events. In our study, patients with positive slopes represented a minority of the study population (7% of patients) and experienced only a tiny fraction (n=18; 0.5%) of the total ESRD events, thereby limiting our ability to distinguish among these potential explanations. The demographic and comorbidity characteristics of those with positive eGFR slopes were similar to those seen in patients with stable eGFR, and significantly fewer of them had albuminuria, making it less likely that a heightened propensity for progressive CKD existed in this group. Recovery from a prior AKI might be associated with an increased risk for and subsequent development of ESRD, perhaps as a result of another AKI event. Many of our cohorts included unselected patients, and even the cohorts that included stable patients provided follow-up creatinine measurements during a subsequent time period when acute events could have occurred. We did not have detailed information about the cause of ESRD, characteristics, such as body composition and muscle mass, or other filtration markers, such as cystatin C, that would be necessary to evaluate these possibilities.

Our study is notable for its large size, international representation, and a diverse patient population. Despite its advantages, this study also has a number of limitations. Standardization of serum creatinine values may have varied across time and studies. The assumption that slopes of eGFR are uniform over time may be flawed. The least squares regression method for the calculation of slopes provides an average linear trajectory over the evaluation period but cannot account for nonlinear trajectories, and its results may be influenced by transient reversible changes in kidney function (e.g., episodes of AKI). Variation in design across cohorts introduces heterogeneity, but the consistency of our results across cohorts, despite the marked variation in design and populations, inspires confidence in them. The added benefit of using slopes over longer durations than 3 years, more frequent serum creatinine measurements, or measurement of other filtration markers, such as cystatin C, was not studied.

In summary, although the last eGFR level seems to be a robust predictor of future ESRD, past trajectory of eGFR over time is also independently associated with ESRD and adds significantly to the information provided by the single last eGFR level, especially in patients with lower eGFR in whom risk of progression to ESRD in the near future is greatest. The ubiquity of electronic medical records makes the evaluation of both single eGFR levels and past slopes of eGFR readily available to increasing numbers of physicians, and their incorporation in everyday clinical practice could improve risk prediction and allow for better strategic resource allocation. The result could be the delivery of better care for later stages of CKD with potential downstream advantages, such as lower ESRD incidence or a more seamless transition to ESRD.

Concise Methods

Study Selection Criteria

The Chronic Kidney Disease Prognosis Consortium (CKD-PC) has been described previously and is also described in Supplemental Appendices 1 and 2.26–30 Briefly, the CKD-PC incorporates cohorts with at least 1000 participants (not applied to cohorts predominantly enrolling persons with CKD [CKD cohorts]) with data on serum creatinine, albuminuria, and ≥50 events of outcomes of interest (mortality or kidney outcomes).26–30 This study included 22 cohorts (13 cohorts in which the presence of CKD was required for cohort entry [CKD cohorts] and nine cohorts in which entry was determined by factors other than CKD [general population and high–cardiovascular disease (CVD) risk cohorts; i.e., other cohorts]) with repeated measures of serum creatinine during baseline evaluation periods of 0.5–3.5 years to determine change in eGFR and data on subsequent ESRD. Meta-analyses were restricted to cohorts with a minimum of 10 ESRD events and participants ages ≥18 years old. This study was approved by the Institutional Review Board at the Johns Hopkins Bloomberg School of Public Health.

Procedures

eGFR was calculated using the Chronic Kidney Disease Epidemiology Collaboration creatinine equation.31 In cohorts where the creatinine measurement was not standardized to isotope dilution mass spectrometry, creatinine concentrations were reduced by 5%.32

Our primary measure of change in eGFR was the annual change (slope), because this is the conventional approach to assess past trajectory in clinical practice. An average annual change in eGFR was estimated from a least squares regression model using all eGFR measurements during baseline periods of 1–3 years. For each baseline period, a 0.5 year of margin before and after the end of the period was allowed for determining the last available eGFR to calculate the change (e.g., eGFR between 0.5 and 1.5 years after the first available eGFR could be used for the 1-year baseline period analysis), but the eGFR closest to the baseline period of interest was selected for each participant. All covariates were assessed at the time of last eGFR (Supplemental Appendix 2 shows details for specific cohorts).

We defined diabetes as fasting glucose ≥7.0 mmol/L (126 mg/dl), nonfasting glucose ≥11.1 mmol/L (200 mg/dl), hemoglobin A1c ≥6.5%, use of glucose-lowering drugs, or self-reported diabetes. Participants with a history of myocardial infarction, coronary revascularization, heart failure, or stroke were considered to have a history of CVD. Albuminuria was not available in all cohorts at the time of last eGFR, and hence, we adjusted for its severity only in sensitivity analyses. Our primary measure of albuminuria was the urine albumin-to-creatinine ratio, but we also included studies with urine albumin excretion rate, urine protein-to-creatinine ratio, or semiquantitative dipstick protein.33

The primary outcome of interest was ESRD after the end of the baseline period. We defined ESRD as initiation of RRT or death caused by kidney disease other than AKI. Patients with ESRD before the baseline period were excluded from the analyses.

Statistical Analyses

We applied a two–stage meta-analysis, with each study first analyzed separately followed by a random effects meta–analysis. The overview of the analysis and analytic notes for individual studies are provided in Supplemental Appendix 2. We imputed missing values of covariates but not the main exposure (change in eGFR) using cohort–specific mean values. We quantified heterogeneity with the I2 statistic and Cochran Q test and explored sources of heterogeneity with random effects meta-regression analysis. Because the absolute risk of ESRD and the implication of change in eGFR vary substantially depending on the type of patient population, analyses were first stratified by type of cohort (CKD versus other).

We modeled the adjusted HRs of subsequent ESRD as a spline function of eGFR slopes. In each study, we fitted piecewise linear splines for eGFR slopes (knots were placed at −10, −5, −3, −1, 1, and 3 ml/min per 1.73 m2 per year). Cox models were adjusted for age, sex, race/ethnicity (black versus nonblack), systolic BP, total cholesterol, diabetes, history of CVD, and last eGFR used to calculate slopes for each evaluation period. Potential effect modifiers with change in eGFR were assessed by incorporating interaction terms. To assess the association of the past slope of eGFR with ESRD in the context of the level of the last eGFR, we present HRs according to eGFR slopes by prespecified levels of the last eGFR (20, 30, 40, and 50 ml/min per 1.73 m2) using no change in eGFR with last eGFR of 50 ml/min per 1.73 m2 as the reference. We selected these eGFR levels because of their relevance to progression to ESRD in the near future.

We translated meta–analyzed adjusted HRs for eGFR slopes stratified by level of last eGFR to absolute risk of ESRD at 1, 3, 5, and 10 years after the baseline period using the weighted average baseline risk; 1-year baseline risk in each cohort was calculated for the following combination of covariates: 60 years old, nonblack, men, no change in eGFR, last eGFR of 50 ml/min per 1.73 m2, systolic BP of 130 mmHg, total cholesterol of 5 mmol/L, and no history of diabetes or CVD. Risk was scaled for longer follow-up and pooled across cohorts using a weighted average (Supplemental Appendix 2). In sensitivity analyses, we applied the adjusted sub-HRs from competing risk models accounting for death as a competing end point.34 In additional sensitivity analyses, we examined the association between eGFR slopes and ESRD in subgroups of patients divided by exposure to RAAS inhibitor medications during the slope evaluation period.

Analyses were performed using Stata/SE 12 software (StataCorp., College Station, TX; www.stata.com). We considered P values <0.05 statistically significant.

Disclosures

None.

Acknowledgments

The Chronic Kidney Disease Prognosis Consortium (CKD-PC) Data Coordinating Center is funded, in part, by a program grant from the US National Kidney Foundation (funding sources include AbbVie and Amgen, Inc.) and National Institute of Diabetes and Digestive and Kidney Diseases Grant R01DK100446-01. A variety of sources have supported enrollment and data collection, including laboratory measurements, as well as follow-up in the collaborating cohorts of the CKD-PC. These funding sources include government agencies, such as the National Institutes of Health, and medical research councils as well as foundations and industry sponsors, and they are listed in Supplemental Appendix 3. Individual cohort and collaborator support is listed in Supplemental Appendix 3.

The funders had no role in the design, analysis, or interpretation of this study and did not contribute to the writing of this report or the decision to submit the article for publication. Some of the data reported here have been supplied by the US Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as an official policy or interpretation of the US Government.

Group members are listed by study. African American Study of Kidney Disease and Hypertension: Jackson T. Wright Jr., Case Western Reserve University; Lawrence J. Appel, Johns Hopkins University; Tom Greene, University of Utah; and Brad C. Astor, University of Wisconsin. The Action in Diabetes and Vascular Disease Study: Preterax and Diamicron Modified Release Controlled Evaluation: Stephen MacMahon, George Institute; John Chalmers, George Institute; Hisatomi Arima, George Institute; and M.W., George Institute. Alberta Kidney Disease Network: Marcello Tonelli, University of Calgary; Brenda R. Hemmelgarn, University of Calgary; Matthew T. James, University of Calgary; and Tanvir Chowdhury Turin, University of Calgary. Atherosclerosis Risk in Communities Study: J.C., Johns Hopkins University; K.M., Johns Hopkins University; Morgan E. Grams, Johns Hopkins University; and Yingying Sang, Johns Hopkins University. British Columbia CKD Study: A.L., British Columbia Provincial Renal Agency and University of British Columbia; and Ognjenka Djurdjev, British Columbia Provincial Renal Agency and Provincial Health Services Authority. Cardiovascular and Renal Outcome in CKD 2–4 Patients - The Fourth Homburg evaluation (CARE for HOMe): Adam M. Zawada, Saarland University Medical Center; Kyrill S. Rogacev, Saarland University Medical Center; Sarah Seiler, Saarland University Medical Center; and Gunnar H. Heine, Saarland University Medical Center. Cleveland Clinic CKD Registry Study: Sankar D. Navaneethan, Cleveland Clinic; J.N., Cleveland Clinic; and Jesse D. Schold, Cleveland Clinic. Cardiovascular Health Study: Michael Shlipak, University of California, San Francisco and San Francisco Veterans Affairs Medical Center; Mark J. Sarnak, Tufts Medical Center; and Ronit Katz, University of Washington. Chronic Renal Impairment in Birmingham: David C. Wheeler, University College London; Jonathan Emberson, University of Oxford; Jonathan N. Townend, Queen Elizabeth Hospital Birmingham; and Martin J. Landray, University of Oxford. Epidemiologische Studie zu Chancen der Verhuetung, Frueherkennung und optimierten Therapie chronischer Erkrankungen in der aelteren Bevölkerung (ESTHER): D.R., German Cancer Research Center and University of Ulm; Hermann Brenner, German Cancer Research Center; Heiko Müller, German Cancer Research Center; and Ben Schöttker, German Cancer Research Center. Geisinger CKD Study: Jamie Green, Geisinger Medical Center; H. Lester Kirchner, Geisinger Medical Center; Robert Perkins, Geisinger Medical Center; and Alex R. Chang, Geisinger Medical Center. Grampian Laboratory Outcomes, Morbidity and Mortality Study 1: Corri Black, University of Aberdeen; Angharad Marks, University of Aberdeen; Nick Fluck, National Health Service (NHS) Grampian; and Gordon J. Prescott, University of Aberdeen. The Nord-Trøndelag Health Study: Stein Hallan, Norwegian University of Science and Technology and St. Olav University; Knut Aasarød, Norwegian University of Science and Technology and St. Olav University Hopstial; Cecilia M. Øien, Norwegian University of Science and Technology; and Maria Radtke, Norwegian University of Science and Technology and St. Olavs University Hospital. Kaiser Permanente Northwest: David H. Smith, Kaiser Permanente Northwest; Micah L. Thorp, Kaiser Permanente Northwest; and Eric S. Johnson, Kaiser Permanente Northwest. Kaiser Permanente Hawaii Cohort: Brian J. Lee, Kaiser Permanente Hawaii Region. Maccabi: Gabriel Chodick, Maccabi Healthcare Services; Varda Shalev, Maccabi Healthcare Services and Tel Aviv University; Yair C. Birnbaum, Maccabi Healthcare Services; and Bracha Shainberg, Maccabi Healthcare Services. Multifactorial Approach and Superior Treatment Efficacy in Renal Patients with the Aid of a Nurse Practitioner Study: Jack F.M. Wetzels, Radboud University Medical Centre; Peter J. Blankestijn, University Medical Center Utrecht; and Arjan D. van Zuilen, University Medical Center Utrecht. Modification of Diet in Renal Disease Study: Mark J. Sarnak, Tufts Medical Center; A.S.L., Tufts Medical Center; Lesley A. Inker, Tufts Medical Center; and Vandana Menon, Tufts Medical Center. Multiple Risk Factor Intervention Trial: Areef Ishani, Minneapolis Veterans Affairs Health Care System; and James D. Neaton, University of Minnesota. NephroTest Study: Marc Froissart, Paris Descartes University; B.S., l'Institut National de la Santé et de la Recherche Médicale U1018, University of Paris Sud, and l'Université Versailles Saint Quentin (USVQ); Marie Metzger, l′Institut National de la Santé et de la Recherche Médicale U1018 and University of Paris Sud-11; Jean-Philippe Haymann, Sorbonne Universités, Université Pierre et Marie Curie University of Paris 06, Assistance Publique-Hôpitaux de Paris; Pascal Houillier, Assistance Publique-Hôpitaux de Paris, Paris Descartes University; and Martin Flamant, Assistance Publique-Hôpitaux de Paris. New Zealand Diabetes Cohort Study: C. Raina Elley, University of Auckland; Timothy Kenealy, University of Auckland; Simon A. Moyes, University of Auckland; John F. Collins, Auckland District Health Board; and Paul L. Drury, Auckland District Health Board. Okinawa 83/93: K.I., University Hospital of the Ryukyus. Pima Indian Study: Robert G. Nelson, National Institute of Diabetes and Digestive and Kidney Diseases; and William C. Knowler, National Institute of Diabetes and Digestive and Kidney Diseases. Prevention of Renal and Vascular Endstage Disease Study: Ron T. Gansevoort, University Medical Center Groningen; Stephan J.L. Bakker, University Medical Center Groningen; Hans L. Hillege, University Medical Center Groningen; and Pim van der Harst, University Medical Center Groningen. Rancho Bernardo: Simerjot K. Jassal, University of California, San Diego and Veterans Affairs San Diego Healthcare; Jaclyn Bergstrom, University of California, San Diego; Joachim H. Ix, University of California, San Diego and Veterans Affairs San Diego Healthcare; and Elizabeth Barrett-Connor, University of California, San Diego. Reduction of Endpoints in Non-Insulin Dependent Diabetes Mellitus with the Angiotensin II Antagonist Losartan: Hiddo J. Lambers Heerspink, University of Groningen; Barry E. Brenner, Brigham and Women’s Hospital and Harvard School of Medicine; and Dick de Zeeuw, University of Groningen. Sunnybrook Cohort: Navdeep Tangri, University of Manitoba; Maneesh Sud, University of Toronto; and D.N., University of Toronto. Uppsala Longitudinal Study of Adult Men: Johan Ärnlöv, Uppsala University; Lars Lannfelt, Uppsala University Hospital; and Anders Larsson, Uppsala University. Veterans Administration CKD Study: C.P.K., Memphis Veterans Affairs Medical Center and University of Tennessee Health Science Center; and Kamyar Kalantar-Zadeh, University of California Irvine Medical Center. CKD-PC Steering Committee: J.C., Johns Hopkins University; Ron T. Gansevoort, University Medical Center Groningen; Morgan E. Grams, Johns Hopkins University; Paul E. de Jong, University Medical Center Groningen; K.I., University Hospital of the Ryukyus; A.S.L., Tufts Medical Center; K.M., Johns Hopkins University; Mark J. Sarnak, Tufts Medical Center; B.S., Inserm U1018, University of Paris Sud, and l′Université Versailles Saint Quentin (USVQ); David G. Warnock, University of Alabama at Birmingham; and M.W., George Institute. CKD-PC Data Coordinating Center: S.H.B. (Coordinator), Johns Hopkins University; J.C. (Principal Investigator), Johns Hopkins University; Morgan E. Grams (Director of Nephrology Initiatives), Johns Hopkins University; K.M. (Director), Johns Hopkins University; Yingying Sang (Lead Programmer), Johns Hopkins University; and M.W. (Senior Statistician), George Institute.

Footnotes

References

  1. ↵
    1. Eckardt KU,
    2. Coresh J,
    3. Devuyst O,
    4. Johnson RJ,
    5. Köttgen A,
    6. Levey AS,
    7. Levin A
    : Evolving importance of kidney disease: From subspecialty to global health burden. Lancet 382: 158–169, 2013pmid:23727165
    OpenUrlCrossRefPubMed
  2. ↵
    1. 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 3: 1–150, 2013
    OpenUrlCrossRef
    1. Coresh J,
    2. Selvin E,
    3. Stevens LA,
    4. Manzi J,
    5. Kusek JW,
    6. Eggers P,
    7. Van Lente F,
    8. Levey AS
    : Prevalence of chronic kidney disease in the United States. JAMA 298: 2038–2047, 2007pmid:17986697
    OpenUrlCrossRefPubMed
  4. ↵
    1. Grams ME,
    2. Chow EK,
    3. Segev DL,
    4. Coresh J
    : Lifetime incidence of CKD stages 3-5 in the United States. Am J Kidney Dis 62: 245–252, 2013pmid:23566637
    OpenUrlCrossRefPubMed
  5. ↵
    1. US Renal Data System
    : USRDS 2010 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States, Bethesda, MD, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, 2010
  6. ↵
    1. Levin A,
    2. Djurdjev O,
    3. Beaulieu M,
    4. Er L
    : Variability and risk factors for kidney disease progression and death following attainment of stage 4 CKD in a referred cohort. Am J Kidney Dis 52: 661–671, 2008pmid:18805347
    OpenUrlCrossRefPubMed
  7. ↵
    1. Johnson ES,
    2. Thorp ML,
    3. Platt RW,
    4. Smith DH
    : Predicting the risk of dialysis and transplant among patients with CKD: A retrospective cohort study. Am J Kidney Dis 52: 653–660, 2008pmid:18585833
    OpenUrlCrossRefPubMed
    1. Fox CS,
    2. Matsushita K,
    3. Woodward M,
    4. Bilo HJG,
    5. Chalmers J,
    6. Heerspink HJL,
    7. Lee BJ,
    8. Perkins RM,
    9. Rossing P,
    10. Sairenchi T,
    11. Tonelli M,
    12. Vassalotti JA,
    13. Yamagishi K,
    14. Coresh J,
    15. de Jong PE,
    16. Wen C-P,
    17. Nelson RG, Chronic Kidney Disease Prognosis Consortium
    : Associations of kidney disease measures with mortality and end-stage renal disease in individuals with and without diabetes: A meta-analysis. Lancet 380: 1662–1673, 2012pmid:23013602
    OpenUrlCrossRefPubMed
  9. ↵
    1. Mahmoodi BK,
    2. Matsushita K,
    3. Woodward M,
    4. Blankestijn PJ,
    5. Cirillo M,
    6. Ohkubo T,
    7. Rossing P,
    8. Sarnak MJ,
    9. Stengel B,
    10. Yamagishi K,
    11. Yamashita K,
    12. Zhang L,
    13. Coresh J,
    14. de Jong PE,
    15. Astor BC, Chronic Kidney Disease Prognosis Consortium
    : Associations of kidney disease measures with mortality and end-stage renal disease in individuals with and without hypertension: A meta-analysis. Lancet 380: 1649–1661, 2012pmid:23013600
    OpenUrlCrossRefPubMed
  10. ↵
    1. Derose SF,
    2. Rutkowski MP,
    3. Crooks PW,
    4. Shi JM,
    5. Wang JQ,
    6. Kalantar-Zadeh K,
    7. Kovesdy CP,
    8. Levin NW,
    9. Jacobsen SJ
    : Racial differences in estimated GFR decline, ESRD, and mortality in an integrated health system. Am J Kidney Dis 62: 236–244, 2013pmid:23499049
    OpenUrlCrossRefPubMed
  11. ↵
    1. Freedman BI,
    2. Parekh RS,
    3. Kao WH
    : Genetic basis of nondiabetic end-stage renal disease. Semin Nephrol 30: 101–110, 2010pmid:20347640
    OpenUrlCrossRefPubMed
  12. ↵
    1. Tangri N,
    2. Kitsios GD,
    3. Inker LA,
    4. Griffith J,
    5. Naimark DM,
    6. Walker S,
    7. Rigatto C,
    8. Uhlig K,
    9. Kent DM,
    10. Levey AS
    : Risk prediction models for patients with chronic kidney disease: A systematic review. Ann Intern Med 158: 596–603, 2013pmid:23588748
    OpenUrlCrossRefPubMed
  13. ↵
    1. Mitch WE,
    2. Walser M,
    3. Buffington GA,
    4. Lemann J Jr.
    : A simple method of estimating progression of chronic renal failure. Lancet 2: 1326–1328, 1976pmid:63803
    OpenUrlCrossRefPubMed
  14. ↵
    1. Turin TC,
    2. Coresh J,
    3. Tonelli M,
    4. Stevens PE,
    5. de Jong PE,
    6. Farmer CK,
    7. Matsushita K,
    8. Hemmelgarn BR
    : Short-term change in kidney function and risk of end-stage renal disease. Nephrol Dial Transplant 27: 3835–3843, 2012pmid:22764191
    OpenUrlCrossRefPubMed
  15. ↵
    1. Turin TC,
    2. Coresh J,
    3. Tonelli M,
    4. Stevens PE,
    5. de Jong PE,
    6. Farmer CK,
    7. Matsushita K,
    8. Hemmelgarn BR
    : Change in the estimated glomerular filtration rate over time and risk of all-cause mortality. Kidney Int 83: 684–691, 2013pmid:23344477
    OpenUrlCrossRefPubMed
  16. ↵
    1. Al-Aly Z,
    2. Zeringue A,
    3. Fu J,
    4. Rauchman MI,
    5. McDonald JR,
    6. El-Achkar TM,
    7. Balasubramanian S,
    8. Nurutdinova D,
    9. Xian H,
    10. Stroupe K,
    11. Abbott KC,
    12. Eisen S
    : Rate of kidney function decline associates with mortality. J Am Soc Nephrol 21: 1961–1969, 2010pmid:20947634
    OpenUrlAbstract/FREE Full Text
    1. Cheng TY,
    2. Wen SF,
    3. Astor BC,
    4. Tao XG,
    5. Samet JM,
    6. Wen CP
    : Mortality risks for all causes and cardiovascular diseases and reduced GFR in a middle-aged working population in Taiwan. Am J Kidney Dis 52: 1051–1060, 2008pmid:18706747
    OpenUrlCrossRefPubMed
  18. ↵
    1. Matsushita K,
    2. Selvin E,
    3. Bash LD,
    4. Franceschini N,
    5. Astor BC,
    6. Coresh J
    : Change in estimated GFR associates with coronary heart disease and mortality. J Am Soc Nephrol 20: 2617–2624, 2009pmid:19892932
    OpenUrlAbstract/FREE Full Text
    1. Perkins RM,
    2. Bucaloiu ID,
    3. Kirchner HL,
    4. Ashouian N,
    5. Hartle JE,
    6. Yahya T
    : GFR decline and mortality risk among patients with chronic kidney disease. Clin J Am Soc Nephrol 6: 1879–1886, 2011pmid:21685022
    OpenUrlAbstract/FREE Full Text
    1. Rifkin DE,
    2. Shlipak MG,
    3. Katz R,
    4. Fried LF,
    5. Siscovick D,
    6. Chonchol M,
    7. Newman AB,
    8. Sarnak MJ
    : Rapid kidney function decline and mortality risk in older adults. Arch Intern Med 168: 2212–2218, 2008pmid:19001197
    OpenUrlCrossRefPubMed
  21. ↵
    1. Shlipak MG,
    2. Katz R,
    3. Kestenbaum B,
    4. Siscovick D,
    5. Fried L,
    6. Newman A,
    7. Rifkin D,
    8. Sarnak MJ
    : Rapid decline of kidney function increases cardiovascular risk in the elderly. J Am Soc Nephrol 20: 2625–2630, 2009pmid:19892934
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Coresh J,
    2. Turin TC,
    3. Matsushita K,
    4. Sang Y,
    5. Ballew SH,
    6. Appel LJ,
    7. Arima H,
    8. Chadban SJ,
    9. Cirillo M,
    10. Djurdjev O,
    11. Green JA,
    12. Heine GH,
    13. Inker LA,
    14. Irie F,
    15. Ishani A,
    16. Ix JH,
    17. Kovesdy CP,
    18. Marks A,
    19. Ohkubo T,
    20. Shalev V,
    21. Shankar A,
    22. Wen CP,
    23. de Jong PE,
    24. Iseki K,
    25. Stengel B,
    26. Gansevoort RT,
    27. Levey AS, CKD Prognosis Consortium
    : Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality. JAMA 311: 2518–2531, 2014pmid:24892770
    OpenUrlCrossRefPubMed
  23. ↵
    1. Lambers Heerspink HJ,
    2. Weldegiorgis M,
    3. Inker LA,
    4. Gansevoort R,
    5. Parving HH,
    6. Dwyer JP,
    7. Mondal H,
    8. Coresh J,
    9. Greene T,
    10. Levey AS,
    11. de Zeeuw D
    : Estimated GFR decline as a surrogate end point for kidney failure: A post hoc analysis from the Reduction of End Points in Non-Insulin-Dependent Diabetes With the Angiotensin II Antagonist Losartan (RENAAL) study and Irbesartan Diabetic Nephropathy Trial (IDNT). Am J Kidney Dis 63: 244–250, 2014pmid:24210590
    OpenUrlCrossRefPubMed
  24. ↵
    1. National Kidney Foundation
    : Research: GFR Decline as an Endpoint in Clinical Trials for CKD, New York, National Kidney Foundation, 2013
  25. ↵
    1. Li L,
    2. Astor BC,
    3. Lewis J,
    4. Hu B,
    5. Appel LJ,
    6. Lipkowitz MS,
    7. Toto RD,
    8. Wang X,
    9. Wright JT Jr.,
    10. Greene TH
    : Longitudinal progression trajectory of GFR among patients with CKD. Am J Kidney Dis 59: 504–512, 2012pmid:22284441
    OpenUrlCrossRefPubMed
  26. ↵
    1. Matsushita K,
    2. van der Velde M,
    3. Astor BC,
    4. Woodward M,
    5. Levey AS,
    6. de Jong PE,
    7. Coresh J,
    8. Gansevoort RT, 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 375: 2073–2081, 2010pmid:20483451
    OpenUrlCrossRefPubMed
    1. van der Velde M,
    2. Matsushita K,
    3. Coresh J,
    4. Astor BC,
    5. Woodward M,
    6. Levey A,
    7. de Jong P,
    8. Gansevoort RT,
    9. van der Velde M,
    10. Matsushita K,
    11. Coresh J,
    12. Astor BC,
    13. Woodward M,
    14. Levey AS,
    15. de Jong PE,
    16. Gansevoort RT,
    17. Levey A,
    18. El-Nahas M,
    19. Eckardt KU,
    20. Kasiske BL,
    21. Ninomiya T,
    22. Chalmers J,
    23. Macmahon S,
    24. Tonelli M,
    25. Hemmelgarn B,
    26. Sacks F,
    27. Curhan G,
    28. Collins AJ,
    29. Li S,
    30. Chen SC,
    31. Hawaii Cohort KP,
    32. Lee BJ,
    33. Ishani A,
    34. Neaton J,
    35. Svendsen K,
    36. Mann JF,
    37. Yusuf S,
    38. Teo KK,
    39. Gao P,
    40. Nelson RG,
    41. Knowler WC,
    42. Bilo HJ,
    43. Joosten H,
    44. Kleefstra N,
    45. Groenier KH,
    46. Auguste P,
    47. Veldhuis K,
    48. Wang Y,
    49. Camarata L,
    50. Thomas B,
    51. Manley T, Chronic Kidney Disease Prognosis Consortium
    : Lower estimated glomerular filtration rate and higher albuminuria are associated with all-cause and cardiovascular mortality. A collaborative meta-analysis of high-risk population cohorts. Kidney Int 79: 1341–1352, 2011pmid:21307840
    OpenUrlCrossRefPubMed
    1. Gansevoort RT,
    2. Matsushita K,
    3. van der Velde M,
    4. Astor BC,
    5. Woodward M,
    6. Levey AS,
    7. de Jong PE,
    8. Coresh J, Chronic Kidney Disease Prognosis Consortium
    : Lower estimated GFR and higher albuminuria are associated with adverse kidney outcomes. A collaborative meta-analysis of general and high-risk population cohorts. Kidney Int 80: 93–104, 2011pmid:21289597
    OpenUrlCrossRefPubMed
    1. Astor BC,
    2. Matsushita K,
    3. Gansevoort RT,
    4. van der Velde M,
    5. Woodward M,
    6. Levey AS,
    7. Jong PE,
    8. Coresh J,
    9. Astor BC,
    10. Matsushita K,
    11. Gansevoort RT,
    12. van der Velde M,
    13. Woodward M,
    14. Levey AS,
    15. de Jong PE,
    16. Coresh J,
    17. El-Nahas M,
    18. Eckardt KU,
    19. Kasiske BL,
    20. Wright J,
    21. Appel L,
    22. Greene T,
    23. Levin A,
    24. Djurdjev O,
    25. Wheeler DC,
    26. Landray MJ,
    27. Townend JN,
    28. Emberson J,
    29. Clark LE,
    30. Macleod A,
    31. Marks A,
    32. Ali T,
    33. Fluck N,
    34. Prescott G,
    35. Smith DH,
    36. Weinstein JR,
    37. Johnson ES,
    38. Thorp ML,
    39. Wetzels JF,
    40. Blankestijn PJ,
    41. van Zuilen AD,
    42. Menon V,
    43. Sarnak M,
    44. Beck G,
    45. Kronenberg F,
    46. Kollerits B,
    47. Froissart M,
    48. Stengel B,
    49. Metzger M,
    50. Remuzzi G,
    51. Ruggenenti P,
    52. Perna A,
    53. Heerspink HJ,
    54. Brenner B,
    55. de Zeeuw D,
    56. Rossing P,
    57. Parving HH,
    58. Auguste P,
    59. Veldhuis K,
    60. Wang Y,
    61. Camarata L,
    62. Thomas B,
    63. Manley T, Chronic Kidney Disease Prognosis Consortium
    : Lower estimated glomerular filtration rate and higher albuminuria are associated with mortality and end-stage renal disease. A collaborative meta-analysis of kidney disease population cohorts. Kidney Int 79: 1331–1340, 2011pmid:21289598
    OpenUrlCrossRefPubMed
  30. ↵
    1. Matsushita K,
    2. Mahmoodi BK,
    3. Woodward M,
    4. Emberson JR,
    5. Jafar TH,
    6. Jee SH,
    7. Polkinghorne KR,
    8. Shankar A,
    9. Smith DH,
    10. Tonelli M,
    11. Warnock DG,
    12. Wen CP,
    13. Coresh J,
    14. Gansevoort RT,
    15. Hemmelgarn BR,
    16. Levey AS, Chronic Kidney Disease Prognosis Consortium
    : Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. JAMA 307: 1941–1951, 2012pmid:22570462
    OpenUrlCrossRefPubMed
  31. ↵
    1. Levey AS,
    2. Stevens LA,
    3. Schmid CH,
    4. Zhang YL,
    5. Castro AF 3rd,
    6. Feldman HI,
    7. Kusek JW,
    8. Eggers P,
    9. Van Lente F,
    10. Greene T,
    11. Coresh J, CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration)
    : A new equation to estimate glomerular filtration rate. Ann Intern Med 150: 604–612, 2009pmid:19414839
    OpenUrlCrossRefPubMed
  32. ↵
    1. Levey AS,
    2. Coresh J,
    3. Greene T,
    4. Marsh J,
    5. Stevens LA,
    6. Kusek JW,
    7. Van Lente F, Chronic Kidney Disease Epidemiology Collaboration
    : Expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 53: 766–772, 2007pmid:17332152
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Miller WG,
    2. Bruns DE,
    3. Hortin GL,
    4. Sandberg S,
    5. Aakre KM,
    6. McQueen MJ,
    7. Itoh Y,
    8. Lieske JC,
    9. Seccombe DW,
    10. Jones G,
    11. Bunk DM,
    12. Curhan GC,
    13. Narva AS, National Kidney Disease Education Program-IFCC Working Group on Standardization of Albumin in Urine
    : Current issues in measurement and reporting of urinary albumin excretion. Clin Chem 55: 24–38, 2009pmid:19028824
    OpenUrlAbstract/FREE Full Text
  34. ↵
    1. Grams ME,
    2. Coresh J,
    3. Segev DL,
    4. Kucirka LM,
    5. Tighiouart H,
    6. Sarnak MJ
    : Vascular disease, ESRD, and death: Interpreting competing risk analyses. Clin J Am Soc Nephrol 7: 1606–1614, 2012pmid:22859747
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

View Selected Citations (0)
Print
Download PDF
Email Article
Citation Tools
Request Permissions
Past Decline Versus Current eGFR and Subsequent ESRD Risk
Csaba P. Kovesdy, Josef Coresh, Shoshana H. Ballew, Mark Woodward, Adeera Levin, David M.J. Naimark, Joseph Nally, Dietrich Rothenbacher, Benedicte Stengel, Kunitoshi Iseki, Kunihiro Matsushita, Andrew S. Levey
JASN Aug 2016, 27 (8) 2447-2455; DOI: 10.1681/ASN.2015060687
del.icio.us logo Digg logo Reddit logo Twitter logo Facebook logo Google logo Mendeley logo

Jump to section

Keywords