Abstract
Background and Aims
Studies of healthy individuals or candidates for living kidney donation, in various geographical areas and ethnic groups, describe a decline of glomerular filtration rate (GFR) with age. Most data on GFR are obtained from subjects in the general population or from candidates for kidney donation who are younger than 65 years. It is currently unknown whether the definition of normal GFR in subjects older than 65 years is similar to the definition in those younger than 65 years. Because the age of candidates for living kidney donation is increasing worldwide, lack of GFR references for older donors complicates the selection process. Moreover, older individuals are most likely to have a mildly decreased GFR that may be misinterpreted as chronic kidney disease. In this study, we calculated percentiles of measured GFR (mGFR) from a large cohort of effective kidney donors (EKD) younger than 65 years, and extrapolated them to subjects older than 65 years. Additionnaly, we collected mGFR data from different centers within Europe from EKD and/or healthy people (HP) from the general population older than 65 years. We tested if the distribution of mGFR of these older subjects fitted with the extrapolated percentiles.
Method
In this retrospective, observational, multi-center study, percentiles of mGFR in EKD were calculated from a development cohort of French and Belgian EKD younger than 65 years (n=1983). From the French kidney donor study, 147 EKD older than 65 years were considered as the internal validation cohort. In an external validation cohort, data on mGFR of subjects older than 65 years, either EKD or HP from the general population (from Germany, Sweden (2), Norway, Netherlands and France, n=2459) were included. Data were fully anonymized and this retrospective study was approved by the respective ethics committees. Percentiles were derived for the development database, using quantiles modeled as cubic splines with two linear parts joining at one age-knot of 40 years. The median quantile had a constant first part (slope of zero) and a second part with a negative slope of -0.88235 mL/min/1.73m² per year. To maintain consistency, all quantiles were adjusted to show the same shape as the medium quantile. Above 65 years, the percentile values were extrapolated using the same mathematical model. We then calculated the percentage of results from the internal and external validation cohorts that were within the 5th extrapolated percentile (P5) and 95th percentile (P95). A sensitivity analysis including the EKD only was performed.
Results
Individuals in the development cohort were younger than in the internal or external validation cohort (47.3±10.5 years vs. 68.8±2.9 years and 71.4±6.4 years; respectively, both p<0.001). Individuals in the development cohort had a higher mGFR than in the internal or external validation cohort (99.9±16.4 mL/min/1.73m2 vs. 86.4±14 mL/min/1.73m2 and 82.7±15.5 mL/min/1.73m2; respectively, both p < 0.001). Among the 147 EKD from the internal validation cohort, none (0%) had mGFR below the extrapolated P5 and12 (8.1%) had mGFR higher than the extrapolated P95. Consequently, 135/147 (91.2%) of subjects were between P5-P95. (Figure 1). Considering the whole external validation cohort (n=329), 5 subjects had mGFR lower than extrapolated P5 (1.5%), 25 were above P95, leaving 299 (90.9%) with mGFR between P5 and the extrapolated P95.
Conclusion
We demonstrate that extrapolated percentiles of mGFR (calculated in individuals younger than 65) fits well with the distribution of mGFR in individuals older than 65. Extrapolation of percentiles to individuals older than 65 is useful to define age-adapted GFR thresholds for older individuals.
Glomerular filtration rate: new age- and gender- specific reference ranges and thresholds for living kidney donation