Background: The glomerular filtration rate (GFR) is the most important measure of kidney function and chronic kidney disease (CKD). This study aims to validate commonly used equations for estimated GFR (eGFR) based on creatinine (cr), cystatin C (cys), β-trace protein (BTP), and β2-microglobulin (B2M) in older adults. Method: We conducted a validation study with 126 participants aged between 72 and 98 with a mean measured GFR (mGFR) by iohexol clearance of 54 mL/min/1.73 m². The eGFR equations (CKD-Epidemiology collaboration [CKD-EPI], Berlin Initiative Study [BIS], Full Age Spectrum [FAS], Modification of Diet in Renal Disease [MDRD]_cr, Caucasian-Asian-Pediatric-Adult [CAPA]_cys, Lund-Malmö Revised [LM-REV]_cr, and MEAN-LM-CAPA_cr-cys), were assessed in terms of bias (median difference: eGFR-mGFR), precision (interquartile range of the differences), and accuracy (P30: percentage of estimates ±30% of mGFR). The equations were compared to a benchmark equation: CKD-EPI_cr-cys. Results: All cystatin C-based equations underestimated the GFR compared to mGFR, whereas bias was mixed for the equations based only on creatinine. Accuracy was the highest for CKD-EPI_cr-cys (98%) and lowest for MDRD (82%). Below mGFR 45 mL/min/1.73 m² only equations incorporating cystatin C reached P30 accuracy >90%. CKD-EPI_cr-cys was not significantly more accurate than the other cystatin C-based equations. In contrast, CKD-EPI_cr-cys was significantly more accurate than all creatinine-based equations except LM-REV_cr. Conclusion: This study confirms that it is reasonable to use equations incorporating cystatin C and creatinine in older patients across a wide spectrum of GFR. However, the results call into question the use of creatinine alone below mGFR 45 mL/min/1.73 m². B2M and BTP do not demonstrate additional value in eGFR determination in older adults.

BackgroundDifferences in cystatin C and creatinine-based estimated glomerular filtration rate (eGFR) can lead to clinical uncertainty. Existing eGFR equations perform poorly in a subset of individuals. This study aims to describe the prevalence of differences between cystatin C-based (eGFRcys) and creatinine-based (eGFRcreat) eGFR in older adults and to explore which subsets of individuals may be most affected by differing estimations.MethodsIn this cross-sectional study, participants from a cohort of community-dwelling older adults were examined at a baseline visit in 2001-2004 as part of the larger “Good Aging in Skåne” study. Exposure variables were obtained from questionnaires, interviews, examinations, and medical records. Blood samples were taken during the baseline visit, cryopreserved, and analyzed at a later time for biomarkers. The CKD-EPI equations were used to estimate GFR. Initial descriptive analyses were performed on 2931 individuals. A total of 2532 participants were included in the final multiple linear regression.ResultsNearly two-thirds of participants had eGFR differences exceeding 10%, with nearly 20 % of participants having eGFR differences exceeding 30%. Smoking, age, body mass index (BMI), C-reactive protein (CRP), glucocorticoid use, and mean eGFR were correlated with differences between eGFRcreat and eGFRcys.ConclusionsDifferences between eGFRcreat and eGFRcys are common and often of large magnitude in this community-dwelling population of older adults. The finding of multiple non-GFR determinants correlated to differences in GFR estimations can help direct future research to improve eGFR equations for subgroups prone to conflicting GFR estimations or to guide choice of biomarker for GFR estimation in these subgroups.Electronic supplementary materialThe online version of this article (10.1186/s12882-017-0759-3) contains supplementary material, which is available to authorized users.

No abstract