#3337 ACCURACY OF ESTIMATED GLOMERULAR FILTRATION RATE EQUATIONS IN PATIENTS WITH DISCORDANCES BETWEEN CREATININE AND CYSTATIN C-BASED ESTIMATIONS

Cystatin C is recommended for use along with creatinine in estimating glomerular filtration rate (eGFR) when precise estimates are needed for clinical decision-making. Although eGFR based on both creatinine and cystatin (eGFRcr-cys) is the most accurate in research studies, it is uncertain as to whether this remains true in real-world settings, particularly when there are large discordances between eGFR based on creatinine (eGFRcr) and that based on cystatin C (eGFRcys).

We included 6185 adults from the Stockholm Creatinine Measurements (SCREAM) project referred for plasma clearance of iohexol in Stockholm, Sweden, who had 9404 concurrent measurements of creatinine, cystatin C and iohexol clearance. The performance of eGFRcr, eGFRcys and eGFRcr-cys was assessed against mGFR with bias, P30 and correct classification of GFR strata. We stratified analyses within three categories: eGFRcys at least 20% lower than eGFRcr (eGFRcys < eGFRcr), eGFRcys within 20% of eGFRcr (eGFRcys ≈ eGFRcr) and eGFRcys at least 20% higher than eGFRcr (eGFRcys > eGFRcr). Bias was expressed as the median difference in estimated GFR minus mGFR, with negative biases indicating underestimation of mGFR. P30 described the percentage of individuals with eGFR within 30% of mGFR. Correct classification of GFR categories was defined as agreement of eGFR and mGFR categories using the KDIGO GFR categories (<15, 15-29, 30-44, 45-59, 60-89 and ≥90 ml/min/1.73 m²).

eGFRcr and eGFRcys were similar in 4226 (45%) of samples, and there all three estimating equations displayed similar performance (Table 1). In contrast, eGFRcr-cys was much more accurate in cases of discordance. For example, when eGFRcys < eGFRcr (47% of samples), median biases were 15.0 (overestimation), −8.5 (underestimation) and 0.8 ml/min/1.73 m² for eGFRcr, eGFRcys and eGFRcr-cys, respectively; P30 was 50%/73%/84%, respectively; and correct classification was 38%/45%/62%, respectively. When eGFRcys > eGFRcr (8% of samples), median biases were −4.5, 8.4, and 1.4 ml/min/1.73 m². Findings were consistent among individuals with cardiovascular disease, heart failure, diabetes mellitus, liver disease and cancer (Figure 1).

When large discordances between eGFR_cr and eGFR_cys are found in clinical practice, eGFR_cr-cys is more accurate than either eGFR_cr or eGFR_cys.

Open in new tabDownload slide

Figure 1:

Bias of eGFR_cr, eGFR_cys and eGFR_cr-cys across subgroups, stratified by the extent of discordance between eGFR_cr and eGFR_cys.

Open in new tabDownload slide