Diagnostic Performance of Creatinine-Based Equations for Estimating Glomerular Filtration Rate in Adults 65 Years and Older

Abstract

Importance: Estimating glomerular filtration rate (GFR) is useful in many clinical conditions. However, very few studies have evaluated the performance of GFR-estimating equations in older adults at various degrees of kidney impairment.

Objective: To determine the performance of plasma-creatinine-based equations Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI), Lund-Malmö Revised, (LMR), full age spectrum (FAS), and Berlin Initiative Study (BIS) 1 in older adults across a broad spectrum of GFRs.

Design, setting, and participants: Single-center cross-sectional study performed in France including 2247 participants aged 65 to 90 years who underwent inulin GFR measurements from July 1, 2003, to July 30, 2017, for suspected or established renal dysfunction, for renal risk, before kidney donation, or after kidney transplant.

Main outcomes and measures: The main outcome measure was GRF measured by inulin clearance. Equation performance criteria considered bias (difference between estimated and measured GFR), precision (interquartile range of the median difference), and accuracy P30 (percentage of estimated GFRs lying between [measured GFR - 30% of measured GFR] and [measured GFR + 30% of measured GFR]).

Results: The mean (SD) age of the 2247 participants was 71.5 (5) years and 1192 (53.0%) were male. The difference in median (95% CI) bias was significant between CKD-EPI vs LMR (-4.0 [-4.0 to -3.5 mL/min/1.73 m2; P < .001]) and CKD-EPI vs FAS (-2.0 [-3.5 to -2.5] mL/min/1.73 m2, P < .001) but not significant between CKD-EPI vs BIS 1 (0.0 [-1.5 to 0.5], P = .07, Mood test). In patients aged 65 to 74 years with measured GFR<45 mL/min/1.73 m2, the difference in median P30 (95% CI) was not significant between CKD-EPI vs LMR (P = .08) and CKD-EPI vs FAS (P = .48) but significant vs BIS 1 (P = .004, McNemar test). In subjects 75 years and older, with measured GFR less than 45 mL/min/1.73 m2, LMR and BIS 1 were more accurate than CKD-EPI and FAS (P30 = 74.5 [70.0-79.5] and 73.0 [68.0-78.0] vs 69.0 [64.5-74.0] and 69.0 [65.5-72.0]). In all patients, despite small statistical differences, the performance of CKD-EPI equation was not clinically different from that of LMR, FAS, or BIS 1.

Conclusions and relevance: In a referral group of patients 65 years and older who had GFR estimated using CDK-EPI, LMR, BIS 1, and FAS equations, a comparison with renal inulin clearance found that none of the equations had a superior diagnostic performance. Each had limitations regarding accuracy.

Figure 1.. Flowchart of the Study

BIS 1 indicates Berlin Initiative Study 1 equation; CKD-EPI, chronic kidney disease epidemiology collaboration equation; FAS, full age spectrum equation; GFR, glomerular filtration rate; and LMR, Lund-Malmö revised equation.

Figure 2.. Chronic Kidney Disease Epidemiology Collaboration Equation Scatter Plots of Equation-Estimated Glomerular Filtration Rates in Function of Inulin-Measured Glomerular Filtration Rates in the Whole Population

The solid line represents the regression line. The dotted lines represent the 95% limits of prediction. The concordance correlation coefficient (CCC) for panel A is 0.832 (95% CI, 0.827-0.838); B, 0.838 (95% CI, 0.832-0.843); C, 0.817 (95% CI, 0.812-0.823); D, 0.826 (95% CI, 0.820-0.832).