Surrogate Outcomes for ESRD Risk: The Case for a 30% Reduction in Estimated GFR Over 2 Years

Abstract

Commentary on Coresh J, Turin TC, Matsushita K, et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality. JAMA. 2014;311(24):2518-2531. Commentary on Coresh J, Turin TC, Matsushita K, et al. Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality. JAMA. 2014;311(24):2518-2531. A highly meaningful clinical outcome for randomized clinical trials in chronic kidney disease (CKD) is the development of end-stage renal disease (ESRD), but the time needed to observe this event in sufficient numbers makes it impractical for most clinical investigators, funding agencies, and pharmaceutical developers. The US Food and Drug Administration accepts a doubling of serum creatinine level as a surrogate outcome for ESRD risk because it reflects a substantial decrease in kidney function and predicts the development of ESRD.1Levey A.S. Cattran D. Friedman A. et al.Proteinuria as a surrogate outcome in CKD: report of a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.Am J Kidney Dis. 2009; 54: 205-226Abstract Full Text Full Text PDF PubMed Scopus (242) Google Scholar However, even this surrogate outcome requires many years to develop, and for patients recruited into a trial at earlier stages of kidney disease, may take decades.2Inker L.A. Levey A.S. Pandya K. Stoycheff N. Okparavero A. Greene T. Early change in proteinuria as a surrogate end point for kidney disease progression: an individual patient meta-analysis.Am J Kidney Dis. 2014; 64: 74-85Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar, 3Coresh J. Turin T.C. Matsushita K. et al.Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality.JAMA. 2014; 311: 2518-2531Crossref PubMed Scopus (608) Google Scholar Thus, alternative surrogate outcomes of ESRD risk on which to evaluate new therapies within a shorter follow-up are needed. Such outcomes could reduce the duration, size, and cost of trials. Further, an intervention that has a demonstrable effect on a surrogate outcome could justify additional testing in a larger longer trial to directly observe the effect of the intervention on safety and efficacy outcomes, including the development of ESRD. In a recent issue of JAMA, Coresh et al3Coresh J. Turin T.C. Matsushita K. et al.Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality.JAMA. 2014; 311: 2518-2531Crossref PubMed Scopus (608) Google Scholar provide strong epidemiologic evidence that a 30% decline in estimated glomerular filtration rate (eGFR) could be this practical surrogate outcome. This international patient-level meta-analysis included 1.7 million participants from 35 cohorts. Baseline assessment occurred from 1975 to 2012. The analytic sample included a broad spectrum of participants, with and without CKD. Average baseline eGFR ranged from 48 to 92 mL/min/1.73 m2, average age was 51 to 74 years, 20% to 51% of participants were women, 1% to 7% were of black race, and 14% to 38% had diabetes. The primary analysis examined the hazard ratio of ESRD in relation to the percent change in eGFR measured over a 2-year baseline period. ESRD was defined as initiation of dialysis therapy or death due to kidney disease other than acute kidney injury. During a mean follow-up of 2.4 years, ESRD occurred in 8,532 of 1,341,193 (0.6%) participants, of whom 52% experienced a 30% decline in eGFR during the 2-year baseline period. In contrast, only 16% of those who reached ESRD had a 57% decline in eGFR (a percent decline consistent with the common study outcome of doubling of serum creatinine) in the 2-year baseline period. Declines in eGFR of 30% and 57% were associated with more than 5-fold and 30-fold increases in the risk of ESRD, respectively. Similar associations were seen in multiple sensitivity analyses (eg, at lower and higher strata of baseline eGFR, in the presence of albuminuria or diabetes, and when slope-based changes in eGFR were analyzed). In general, observing decrements in eGFR over a longer baseline period (which would represent the follow-up period when assessed in a clinical trial) strengthened the association with ESRD. Associations between changes in eGFR and all-cause mortality were similar, but weaker. The first step in validating a surrogate is to establish a causal connection between a change in the surrogate outcome and the meaningful clinical outcome, and this study provides strong epidemiologic evidence for this with data from a broad array of patient populations and study designs. The authors tested and verified their findings in multiple sensitivity analyses, and the associations they observed are unlikely to be the result of confounding or selection bias. This study also highlights the trade-off in using a surrogate outcome with a lower prevalence and higher relative risk for ESRD (eg, doubling of serum creatinine) versus a smaller decline in eGFR, which is more prevalent, but for which the associated ESRD risk is attenuated. If the relative risk reduction of an intervention on ESRD is greater than the relative risk reduction of the same intervention on a surrogate outcome, this could outweigh the statistical benefit of an increased number of surrogate outcome events.4Lambers Heerspink H.J. Weldegiorgis M. Inker L.A. et al.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.Am J Kidney Dis. 2014; 63: 244-250Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar As the authors note, a strong and consistent association between the surrogate outcome and ESRD is a necessary, but insufficient, condition to support use of the surrogate outcome in clinical trials. The key empirical criterion for validating a surrogate is to determine whether the effect of the treatment on the surrogate outcome reliably predicts its effect on the more meaningful clinical outcome (in this case, ESRD and/or death).1Levey A.S. Cattran D. Friedman A. et al.Proteinuria as a surrogate outcome in CKD: report of a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.Am J Kidney Dis. 2009; 54: 205-226Abstract Full Text Full Text PDF PubMed Scopus (242) Google Scholar, 5Stevens L.A. Greene T. Levey A.S. Surrogate end points for clinical trials of kidney disease progression.Clin J Am Soc Nephrol. 2006; 1: 874-884Crossref PubMed Scopus (110) Google Scholar Researchers typically test this criterion in 2 ways: (1) by meta-analysis of clinical trials to compare the relative risk reduction on the surrogate outcome and the meaningful clinical outcome, and (2) by determining the proportion of the treatment effect on the clinical outcome that is explained by the surrogate outcome (the Prentice-Freedman criterion).1Levey A.S. Cattran D. Friedman A. et al.Proteinuria as a surrogate outcome in CKD: report of a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.Am J Kidney Dis. 2009; 54: 205-226Abstract Full Text Full Text PDF PubMed Scopus (242) Google Scholar, 2Inker L.A. Levey A.S. Pandya K. Stoycheff N. Okparavero A. Greene T. Early change in proteinuria as a surrogate end point for kidney disease progression: an individual patient meta-analysis.Am J Kidney Dis. 2014; 64: 74-85Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar, 6Prentice R.L. Surrogate endpoints in clinical trials: definition and operational criteria.Stat Med. 1989; 8: 431-440Crossref PubMed Scopus (1464) Google Scholar Ideally, the change in the clinical outcome resulting from at least one type of intervention, and preferably many, should be fully explained by the change in the surrogate outcome.7Bucher H.C. Guyatt G.H. Cook D.J. Holbrook A. McAlister F.A. Users’ guides to the medical literature: XIX. Applying clinical trial results. A. How to use an article measuring the effect of an intervention on surrogate end points. Evidence-Based Medicine Working Group.JAMA. 1999; 282: 771-778Crossref PubMed Scopus (283) Google Scholar Unfortunately, this rarely occurs in practice given the complexity of kidney disease, measurement error in the surrogate, variation in analytic technique, and most importantly, a requirement for multiple large lengthy clinical trials, ideally of different interventions, that measure both the surrogate outcome and the important clinical outcome (in this case, ESRD).1Levey A.S. Cattran D. Friedman A. et al.Proteinuria as a surrogate outcome in CKD: report of a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.Am J Kidney Dis. 2009; 54: 205-226Abstract Full Text Full Text PDF PubMed Scopus (242) Google Scholar Nonetheless, an intermediate outcome will fail as a surrogate if the intervention affects the clinical outcome by a separate causal pathway, and thus it remains crucial to consider the natural history of the specific kidney disease under investigation.1Levey A.S. Cattran D. Friedman A. et al.Proteinuria as a surrogate outcome in CKD: report of a scientific workshop sponsored by the National Kidney Foundation and the US Food and Drug Administration.Am J Kidney Dis. 2009; 54: 205-226Abstract Full Text Full Text PDF PubMed Scopus (242) Google Scholar, 5Stevens L.A. Greene T. Levey A.S. Surrogate end points for clinical trials of kidney disease progression.Clin J Am Soc Nephrol. 2006; 1: 874-884Crossref PubMed Scopus (110) Google Scholar, 7Bucher H.C. Guyatt G.H. Cook D.J. Holbrook A. McAlister F.A. Users’ guides to the medical literature: XIX. Applying clinical trial results. A. How to use an article measuring the effect of an intervention on surrogate end points. Evidence-Based Medicine Working Group.JAMA. 1999; 282: 771-778Crossref PubMed Scopus (283) Google Scholar For example, serum creatinine concentration is affected by a number of non–GFR-related factors, and creatinine-based GFR estimates should be avoided as outcomes in trials testing interventions that affect creatinine generation, secretion, or extrarenal elimination.5Stevens L.A. Greene T. Levey A.S. Surrogate end points for clinical trials of kidney disease progression.Clin J Am Soc Nephrol. 2006; 1: 874-884Crossref PubMed Scopus (110) Google Scholar Also, surrogate outcomes based on smaller declines in eGFR may be more prone to confounding in trials of drugs that exert acute hemodynamic effects on eGFR, for example, angiotensin-converting enzyme inhibitors and angiotensin receptor blockers, which cause an initial decrease in eGFR, but appear beneficial in the long run.4Lambers Heerspink H.J. Weldegiorgis M. Inker L.A. et al.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.Am J Kidney Dis. 2014; 63: 244-250Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar For trials of losartan and irbesartan, Lambers Heerspink et al4Lambers Heerspink H.J. Weldegiorgis M. Inker L.A. et al.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.Am J Kidney Dis. 2014; 63: 244-250Abstract Full Text Full Text PDF PubMed Scopus (48) Google Scholar showed that adopting an eGFR decline less than the standard 57% decrease as a surrogate outcome did not consistently improve statistical efficiency unless eGFR decline was computed from 3 months after baseline (which violates the principles of randomized design). Since 2008, nine studies8Matsushita K. Selvin E. Bash L.D. Franceschini N. Astor B.C. Coresh J. Change in estimated GFR associates with coronary heart disease and mortality.J Am Soc Nephrol. 2009; 20: 2617-2624Crossref PubMed Scopus (204) Google Scholar, 9Shlipak M.G. Katz R. Kestenbaum B. et al.Rapid decline of kidney function increases cardiovascular risk in the elderly.J Am Soc Nephrol. 2009; 20: 2625-2630Crossref PubMed Scopus (213) Google Scholar, 10Turin T.C. Coresh J. Tonelli M. et al.Short-term change in kidney function and risk of end-stage renal disease.Nephrol Dial Transplant. 2012; 27: 3835-3843Crossref PubMed Scopus (76) Google Scholar, 11Turin T.C. Coresh J. Tonelli M. et al.One-year change in kidney function is associated with an increased mortality risk.Am J Nephrol. 2012; 36: 41-49Crossref PubMed Scopus (45) Google Scholar, 12Turin T.C. Coresh J. Tonelli M. et al.Change in the estimated glomerular filtration rate over time and risk of all-cause mortality.Kidney Int. 2013; 83: 684-691Crossref PubMed Scopus (90) Google Scholar, 13Rifkin D.E. Shlipak M.G. Katz R. et al.Rapid kidney function decline and mortality risk in older adults.Arch Intern Med. 2008; 168: 2212-2218Crossref PubMed Scopus (270) Google Scholar, 14Perkins R.M. Bucaloiu I.D. Kirchner H.L. Ashouian N. Hartle J.E. Yahya T. GFR decline and mortality risk among patients with chronic kidney disease.Clin J Am Soc Nephrol. 2011; 6: 1879-1886Crossref PubMed Scopus (93) Google Scholar, 15Al-Aly Z. Zeringue A. Fu J. et al.Rate of kidney function decline associates with mortality.J Am Soc Nephrol. 2010; 21: 1961-1969Crossref PubMed Scopus (148) Google Scholar, 16Al-Aly Z. Balasubramanian S. McDonald J.R. Scherrer J.F. O’Hare A.M. Greater variability in kidney function is associated with an increased risk of death.Kidney Int. 2012; 82: 1208-1214Crossref PubMed Scopus (60) Google Scholar have reported an association between eGFR change and mortality; however, only 2 studies have specifically examined the risk of ESRD in relation to eGFR change.10Turin T.C. Coresh J. Tonelli M. et al.Short-term change in kidney function and risk of end-stage renal disease.Nephrol Dial Transplant. 2012; 27: 3835-3843Crossref PubMed Scopus (76) Google Scholar, 17Skupien J. Warram J.H. Smiles A.M. et al.The early decline in renal function in patients with type 1 diabetes and proteinuria predicts the risk of end-stage renal disease.Kidney Int. 2012; 82: 589-597Crossref PubMed Scopus (107) Google Scholar In one, a Canadian cohort of 598,397 patients with CKD, a 25% decline in eGFR over 1 year was associated with a 5-fold increased risk of ESRD.10Turin T.C. Coresh J. Tonelli M. et al.Short-term change in kidney function and risk of end-stage renal disease.Nephrol Dial Transplant. 2012; 27: 3835-3843Crossref PubMed Scopus (76) Google Scholar In the second, a study of 216 individuals with type 1 diabetes and stages 1 to 2 CKD, the risk of ESRD increased by 31% for each 5-mL/min/1.73 m2 decrease in eGFR.17Skupien J. Warram J.H. Smiles A.M. et al.The early decline in renal function in patients with type 1 diabetes and proteinuria predicts the risk of end-stage renal disease.Kidney Int. 2012; 82: 589-597Crossref PubMed Scopus (107) Google Scholar Finally, the National Kidney Foundation and the US Food and Drug Administration recently hosted a workshop that examined data on alternative definitions of GFR decline as end points in randomized controlled trials18National Kidney Foundation. Research: GFR decline as an endpoint in clinical trials for CKD. http://www.kidney.org/professionals/research/research_info.cfm. Accessed July 3, 2014.Google Scholar; however, preliminary results are available only as abstracts.19Inker L. Lambers Heerspink H.J. Mondal H. Coresh J. Greene T. Levey A.S. GFR decline as an endpoint for clinical trials in CKD.J Am Soc Nephrol. 2013; 24 ([abstract]): 12AGoogle Scholar, 20Greene T.H. Teng C.-C. Redd A.M. et al.Validity and statistical power of alternative eGFR-based endpoints: a report from an NFK FDA Workshop.J Am Soc Nephrol. 2013; 24 ([abstract]): 151AGoogle Scholar In addition to discussing the validity of eGFR change as a surrogate outcome, issues relating to eGFR measurement require consideration. By definition, GFR must decline for patients to develop kidney failure; however, many patients have nonprogressive or slowly progressive disease and never reach ESRD, instead dying prematurely of comorbid conditions such as cardiovascular disease that disproportionately affect individuals with kidney diseases.21O’Hare A.M. Batten A. Burrows N.R. et al.Trajectories of kidney function decline in the 2 years before initiation of long-term dialysis.Am J Kidney Dis. 2012; 59: 513-522Abstract Full Text Full Text PDF PubMed Scopus (176) Google Scholar eGFR change frequently is nonlinear, with many patients exhibiting a variable pattern of decreasing, stable, and increasing eGFR.16Al-Aly Z. Balasubramanian S. McDonald J.R. Scherrer J.F. O’Hare A.M. Greater variability in kidney function is associated with an increased risk of death.Kidney Int. 2012; 82: 1208-1214Crossref PubMed Scopus (60) Google Scholar, 22Chang A. Kramer H. CKD progression: a risky business.Nephrol Dial Transplant. 2012; 27: 2607-2609Crossref PubMed Scopus (9) Google Scholar, 23Turin T.C. Hemmelgarn B.R. Improvement in kidney function: a real occurrence.J Am Soc Nephrol. 2012; 23: 575-577Crossref PubMed Scopus (3) Google Scholar The prognostic implications of an increasing eGFR based on creatinine-based estimating equations are unclear. Many studies, including this one, show that an increasing or highly variable eGFR is associated with a 30% to 120% increased risk of death (but not ESRD).3Coresh J. Turin T.C. Matsushita K. et al.Decline in estimated glomerular filtration rate and subsequent risk of end-stage renal disease and mortality.JAMA. 2014; 311: 2518-2531Crossref PubMed Scopus (608) Google Scholar, 8Matsushita K. Selvin E. Bash L.D. Franceschini N. Astor B.C. Coresh J. Change in estimated GFR associates with coronary heart disease and mortality.J Am Soc Nephrol. 2009; 20: 2617-2624Crossref PubMed Scopus (204) Google Scholar, 11Turin T.C. Coresh J. Tonelli M. et al.One-year change in kidney function is associated with an increased mortality risk.Am J Nephrol. 2012; 36: 41-49Crossref PubMed Scopus (45) Google Scholar, 12Turin T.C. Coresh J. Tonelli M. et al.Change in the estimated glomerular filtration rate over time and risk of all-cause mortality.Kidney Int. 2013; 83: 684-691Crossref PubMed Scopus (90) Google Scholar, 16Al-Aly Z. Balasubramanian S. McDonald J.R. Scherrer J.F. O’Hare A.M. Greater variability in kidney function is associated with an increased risk of death.Kidney Int. 2012; 82: 1208-1214Crossref PubMed Scopus (60) Google Scholar, 24Perkins R.M. Tang X. Bengier A.C. Kirchner H.L. Bucaloiu I.D. Variability in estimated glomerular filtration rate is an independent risk factor for death among patients with stage 3 chronic kidney disease.Kidney Int. 2012; 82: 1332-1338Crossref PubMed Scopus (35) Google Scholar However, others show that a stable or improving eGFR is associated with a favorable risk profile (eg, younger age or lower urine protein excretion).23Turin T.C. Hemmelgarn B.R. Improvement in kidney function: a real occurrence.J Am Soc Nephrol. 2012; 23: 575-577Crossref PubMed Scopus (3) Google Scholar, 25Hu B. Gadegbeku C. Lipkowitz M.S. et al.Kidney function can improve in patients with hypertensive CKD.J Am Soc Nephrol. 2012; 23: 706-713Crossref PubMed Scopus (38) Google Scholar, 26Eriksen B.O. Tomtum J. Ingebretsen O.C. Predictors of declining glomerular filtration rate in a population-based chronic kidney disease cohort.Nephron Clin Pract. 2010; 115: c41-c50Crossref PubMed Scopus (18) Google Scholar High variability in eGFR may reflect changes in glomerular hemodynamics, impaired renal homeostasis, acute sarcopena, or developing and resolving acute kidney injury.10Turin T.C. Coresh J. Tonelli M. et al.Short-term change in kidney function and risk of end-stage renal disease.Nephrol Dial Transplant. 2012; 27: 3835-3843Crossref PubMed Scopus (76) Google Scholar, 16Al-Aly Z. Balasubramanian S. McDonald J.R. Scherrer J.F. O’Hare A.M. Greater variability in kidney function is associated with an increased risk of death.Kidney Int. 2012; 82: 1208-1214Crossref PubMed Scopus (60) Google Scholar Given the substantial heterogeneity in patient eGFR trajectories and associated outcomes, researchers should be cautious about using a dichotomized change in eGFR as their sole assessment of kidney function. The evaluation of new surrogate outcomes for ESRD is an important endeavor. As history has shown, the premature adoption of interventions based on their effects on untested surrogates can waste health care resources and may even cause patient harm.2Inker L.A. Levey A.S. Pandya K. Stoycheff N. Okparavero A. Greene T. Early change in proteinuria as a surrogate end point for kidney disease progression: an individual patient meta-analysis.Am J Kidney Dis. 2014; 64: 74-85Abstract Full Text Full Text PDF PubMed Scopus (79) Google Scholar, 5Stevens L.A. Greene T. Levey A.S. Surrogate end points for clinical trials of kidney disease progression.Clin J Am Soc Nephrol. 2006; 1: 874-884Crossref PubMed Scopus (110) Google Scholar, 7Bucher H.C. Guyatt G.H. Cook D.J. Holbrook A. McAlister F.A. Users’ guides to the medical literature: XIX. Applying clinical trial results. A. How to use an article measuring the effect of an intervention on surrogate end points. Evidence-Based Medicine Working Group.JAMA. 1999; 282: 771-778Crossref PubMed Scopus (283) Google Scholar, 27Echt D.S. Liebson P.R. Mitchell L.B. et al.Mortality and morbidity in patients receiving encainide, flecainide, or placebo. The Cardiac Arrhythmia Suppression Trial.N Engl J Med. 1991; 324: 781-788Crossref PubMed Scopus (2621) Google Scholar Conversely, effective therapies may be discarded if the observed risk reductions in the surrogate outcome mis-specify what would be seen on the clinically important outcome. Thus, further analyses from clinical trials are needed to validate a smaller change in eGFR as a surrogate outcome and to fully understand its limitations. Support: None. Financial Disclosure: The authors declare that they have no relevant financial interests.