Pitfalls in Using Estimated Glomerular Filtration Rate Slope as a Surrogate for the Effect of Drugs on the Risk of Serious Adverse Renal Outcomes in Clinical Trials of Patients With Heart Failure

Abstract

Both type 2 diabetes and chronic heart failure increase the risk of serious adverse renal outcomes, and the onset of kidney failure is generally preceded by a progressive decline in estimated glomerular filtration rate (eGFR). Long-term treatment with inhibitors of the reninangiotensin system, neprilysin, and the SGLT2 (sodiumglucose cotransporter 2) slows the rate of decline in eGFR, and these drugs decrease the risk of end-stage renal disease in certain populations. This apparent parallelism has led some investigators to propose that drug-induced changes in rate of decline in eGFR might represent a surrogate for the effect of a drug to reduce the risk of end-stage renal disease.1 The rate of decline in eGFR is commonly assessed by the measurement of eGFR slope, but this calculation is fraught with methodological difficulties. Estimation of slope relies on statistical models that make assumptions about linearity and are typically skewed by values measured late in the course of follow-up. However, the patients who provide data at late time points are those who were first recruited into the trial, and they represent a small fraction of those with baseline values (Figure). Furthermore, slope can be reliably calculated only over meaningful periods of time (about 3–5 years).1 However, many trials in heart failure have a median duration of follow-up of <18 months.