Potential Surrogate Outcomes for Kidney Failure in Advanced CKD: Evaluation of Power and Predictive Ability in CKDopps

Abstract

Rationale & objective: Potential surrogate end points for kidney failure have been proposed in chronic kidney disease (CKD); however, they must be evaluated to ensure accurate, powerful, and harmonized research, particularly among patients with advanced CKD. The aim of the current study was to investigate the power and predictive ability of surrogate kidney failure end points in a population with moderate-to-advanced CKD.

Study design: Analysis of longitudinal data of a large multinational CKD observational study (Chronic Kidney Disease Outcomes and Practice Patterns Study).

Setting & participants: CKD stage 3-5 patients from Brazil, France, Germany, and the United States.

Outcomes: Reaching an estimated glomerular filtration rate (eGFR) < 15 mL/min/1.73 m² or eGFR decline of ≥40%, and composite end points of these individual end points.

Analytical approach: Each end point was used as a time-varying indicator in the Cox model to predict the time to kidney replacement therapy (KRT; dialysis or transplant) and was compared by the number of events and prediction accuracy.

Results: 8,211 patients had a median baseline eGFR of 27 mL/min/1.73 m² (interquartile range, 21-36 mL/min/1.73 m²) and 1,448 KRT events over a median follow-up of 2.7 years (interquartile range, 1.2-3.0 years). Among CKD stage 4 patients, the eGFR < 15 mL/min/1.73 m² end point had higher prognostic ability than 40% eGFR decline, but the end points were similar for CKD stage 3 patients. The combination of eGFR < 15 mL/min/1.73 m² and 40% eGFR decline had the highest prognostic ability for predicting KRT, regardless of the CKD stage. Including KRT in the composite can increase the number of events and, therefore, the power.

Limitations: Variable visit frequency resulted in variable eGFR measurement frequency.

Conclusions: The composite end point can be useful for CKD progression studies among patients with advanced CKD. Harmonized use of this approach has the potential to accelerate the translation of new discoveries to clinical practice by identifying risk factors and treatments for kidney failure.

Keywords: Chronic kidney disease progression; composite outcomes; estimated glomerular filtration rate; kidney failure; surrogate end points.

Graphical abstract

Figure 1

Scatterplots showing number of events versus ρ² for each primary potential surrogate end point, separated by CKD stage at study enrollment. ρ² represents the proportion of variability in KRT that can be explained by the potential surrogate, with 0 representing no prognostic value of the potential surrogate and 1.0 representing a perfect prediction. A higher number of events implies higher statistical power. Therefore, points closer to the top right are generally optimal. ∗Potential surrogate end points that include eGFR < 15 mL/min/1.73 m² are limited to those with enrollment eGFR of 15 mL/min/1.73 m² or greater. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; KRT, kidney replacement therapy.

Figure 2

Scatterplots showing number of events versus ρ² for each primary potential surrogate end point, separated by CKD stage at study enrollment and age tertiles. Participants of age 66 are included in the interval [18, 66] and not in the interval (66, 76], and participants of age 76 are included in the interval (66, 76] and not in the interval (76, 98] as indicated by the parentheses and square brackets. ρ² represents the proportion of variability in KRT that can be explained by the potential surrogate, with 0 representing no prognostic value of the potential surrogate and 1.0 representing a perfect prediction. A higher number of events implies higher statistical power. Therefore, points closer to the top right are generally optimal. ∗Potential surrogate end points that include eGFR < 15 mL/min/1.73 m² are limited to those with enrollment eGFR of 15 mL/min/1.73 m² or greater. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; KRT, kidney replacement therapy.

Figure 3

Scatterplots showing number of events versus ρ² for each primary potential surrogate end point, separated by CKD stage at study enrollment and sex. ρ² represents the proportion of variability in KRT that can be explained by the potential surrogate, with 0 representing no prognostic value of the potential surrogate and 1.0 representing a perfect prediction. A higher number of events implies higher statistical power. Therefore, points closer to the top right are generally optimal. ∗Potential surrogate end points that include eGFR < 15 mL/min/1.73 m² are limited to those with enrollment eGFR of 15 mL/min/1.73 m² or greater. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; KRT, kidney replacement therapy.

Figure 4

Scatterplots showing number of events versus ρ² for each primary potential surrogate end point, separated by CKD stage at study enrollment and diabetes status. ρ² represents the proportion of variability in KRT that can be explained by the potential surrogate, with 0 representing no prognostic value of the potential surrogate and 1.0 representing a perfect prediction. A higher number of events implies higher statistical power. Therefore, points closer to the top right are generally optimal. ∗Potential surrogate end points that include eGFR < 15 mL/min/1.73 m² are limited to those with enrollment eGFR of 15 mL/min/1.73 m² or greater. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; KRT, kidney replacement therapy.

Figure 5

Scatterplots showing number of events versus ρ² for each primary potential surrogate end point, separated by CKD stage at study enrollment and country. ρ² represents the proportion of variability in KRT that can be explained by the potential surrogate, with 0 representing no prognostic value of the potential surrogate and 1.0 representing a perfect prediction. A higher number of events implies higher statistical power. Therefore, points closer to the top right are generally optimal. ∗Potential surrogate end points that include eGFR < 15 mL/min/1.73 m² are limited to those with enrollment eGFR of 15 mL/min/1.73 m² or greater. Abbreviations: CKD, chronic kidney disease; eGFR, estimated glomerular filtration rate; KRT, kidney replacement therapy.