This study is to the best of our knowledge the largest published series to date of mGFRs in a healthy population, and we have used it as the basis for new age- and gender-specific reference ranges for GFR and to define advisory mGFR thresholds for living kidney donation that form part of the updated BTS guidelines [
8]. In 2974 prospective living kidney donors from 18 UK centres, we found that young adults had an mGFR of approximately 100 mL/min/1.73m
2 until 35 years of age, following which there was a linear decline that was faster in females. An age-related decline in GFR, which was faster in females, was confirmed in a subgroup of prospective donors selected on the basis of no proteinuria, normal renal imaging, and normal differential kidney function.
Ageing and GFR
Our finding that GFR is approximately 100 mL/min/1.73m
2 in young adults is consistent with previously published data [
3,
5,
12]. Published studies consistently show that GFR declines with age in healthy individuals, and it has been demonstrated that ageing is also associated with a decline in other physiological parameters, such as renal blood flow, and with structural changes such as a reduction in nephron number, glomerulosclerosis, and tubulointerstitial fibrosis [
13‐
15].
Over 10% of prospective donors older than 60 years had a mGFR < 60 mL/min/1.73m
2, and our model suggests that the lower limit of the normal range for GFR drops to 60 mL/min/1.73m
2 by the age of 50 and 55 years for females and males, respectively, meeting the GFR cut-off in the current definition of chronic kidney disease (CKD) [
16]. Our age- and gender-specific GFR reference ranges may be useful in the management of an older individual with a GFR < 60 mL/min/1.73m
2: if GFR is within their reference range and there is no other evidence of kidney disease, one may be more confident in attributing a GFR < 60 mL/min/1.73m
2 to normal ageing rather than disease, which may help to avoid unnecessary investigation and over-medicalisation. A meta-analysis which evaluated the interaction of age on the association between creatinine-based eGFR and end-stage renal disease (ESRD) and mortality suggested that a low eGFR is associated with increased risks across all age groups, although the relative mortality risk associated with a reduced eGFR decreased with increasing age [
17].
It should also be noted that men younger than 55 years and women younger than 50 years could have a GFR below the lower limit of their reference range but still ≥60 mL/min/1.73m2 and so, in the absence of another marker of renal disease, would be missed by the current definition of CKD. The available data on long-term outcomes is insufficient to know whether a young individual in this category is at an increased lifetime risk of adverse health outcomes.
The definition and diagnosis of CKD, particularly in the older population, is an area of ongoing debate, and a discourse on the use of a single GFR cut-off is beyond the scope of this paper. However, several reviews and opinion pieces have suggested that greater emphasis should be placed on age-related changes in GFR than current definitions of CKD allow for [
12,
18,
19]. Our data showing that over 10% of healthy adults over 60 years of age have a GFR < 60 mL/min/1.73m
2 could be used to support that position.
Gender and GFR
Although most studies have found no significant difference in GFR between males and females, our findings that males have a higher overall GFR and a less rapid decline with age are consistent with the results of several other reports [
6,
20‐
22]. These findings were also evident in the subgroup of prospective donors with no evidence of kidney pathology.
Several theories have been proposed to explain the more rapid GFR decline seen in females. First, it has been proposed that females may have a higher GFR than males in young adulthood that is masked by scaling to BSA, which may lead to a faster rate of GFR decline similar to that seen in hyperfiltration-related renal pathology [
9]. Second, as females age, the impact of oestrogens on renal haemodynamics and structure are lost due to a gradual decline in oestrogen levels even before the menopause [
23‐
25]. There is no evidence that this faster rate of decline in females is detrimental to health, and in the UK the prevalence of ESRD in women is actually lower than in men at all ages [
26].
Minimum advisory GFR thresholds for living kidney donation
The evaluation of prospective living kidney donors aims to identify those whom donation would put at an unacceptably high risk of long-term complications, including ESRD. Previous studies have suggested that the risk of ESRD after living kidney donation is not higher than in the general population [
27,
28], but there is a small absolute increased risk [
29‐
31]. A recent meta-analysis found the relative risk for ESRD was about 9-fold higher in donors compared to non-donors, but the estimated incidence rate was less than 1 case per 1000 person-years [
32]. Assessment of GFR in prospective kidney donors is an important factor in determining risk and living kidney donation guidelines have provided threshold GFRs above which the increased risk may generally be considered acceptable. For example, the 2017 KDIGO guideline suggests that a GFR ≥90 mL/min/1.73m
2 is acceptable for donation, while a GFR < 60 mL/min/1.73m
2 is not acceptable for donation and a GFR 60–89 mL/min/1.73m
2 may be acceptable depending on other risk factors [
33].
We recommended an advisory threshold GFR of > 80 mL/min/1.73 m
2 for prospective donors aged 30–45, because this appears safe based on long-term outcome studies showing only a very small absolute increased risk of ESRD in cohorts of donors with this level of renal function and in this age range [
30,
31]. However, because these studies contained only small numbers of younger donors, and another study showed an increased absolute lifetime risk in younger donors with a GFR < 90 mL/min/1.73m
2 [
34], we have recommended an advisory threshold GFR of > 90 mL/min/1.73m
2 for donors younger than 30 years of age.
For those over 45 years, we have recommended advisory GFR thresholds based on calculating the lowest pre-donation GFR that would leave the donor with a post-donation GFR within our age- and gender-specific GFR normal ranges, assuming a 25% loss of GFR [
27,
35,
36]. Based on data showing that age-related GFR decline after donation appears to be slower than in the general population, donors should remain within the healthy reference range up to the age of 80 years [
27,
35,
37]. Whilst it is acknowledged that these may be considered arbitrary thresholds, from first principles it would seem sensible to aim to keep GFR in the normal range. Our thresholds, unlike those in the KDIGO guideline, would also allow some older donors with a GFR < 60 mL/min/1.73m
2 to donate.
In our study population of 2974 prospective donors, these GFR thresholds would lead to the exclusion of an additional 5.0% (19.9% vs 14.8%) of prospective donors compared to the thresholds in the previous BTS living kidney donation guidelines if adhered to rigidly [
38].
However, whilst threshold GFRs provide useful guidance to clinicians assessing prospective living kidney donors, individualised decision-making is important, especially in cases where GFR may be just below the recommended advisory threshold, or where there are compounding risk factors for ESRD. This has recently been facilitated by the development of an online tool (
www.transplantmodels.com/esrdrisk) which provides a 15-year and lifetime pre-donation risk of ESRD in prospective donors, which was based on a meta-analysis of data from nearly 5 million healthy individuals, similar to kidney donor candidates, from general population cohorts [
34].
The strengths of this study include the large size of the cohort and the fact that it incorporates individuals from multiple centres which increases diversity and the generalisability of our reference ranges. However, we recognise that our study has some limitations. First, it is likely that there is some variation in practice between the 18 centres in conducting mGFRs. There was variation in the number of blood samples taken to calculate mGFR, one centre used a different glomerular filtration tracer, and there may have been variation in pre-procedure advice given to individuals, such as that pertaining to diet, medications, and fasting.
Second, our estimates of GFR in the general population may be biased by the use of prospective living kidney donors as our study population. Individuals who have volunteered to donate a kidney and have had a satisfactory initial medical assessment to reach the point of having their GFR measured are likely to be healthier than the unscreened general population, and we would therefore anticipate a higher reference range than in the general population. Conversely, many prospective kidney donors are related to the intended recipient with renal failure, and therefore the proportion of individuals in our study population with a family history of renal disease is almost certainly higher than in the general population. This may have resulted in a lower estimate of GFR than truly exists in the general population. We did not have data on donor-recipient relationship to examine this further. Another consequence of using prospective kidney donors was that there were relatively few individuals in the cohort over 70 years of age, and this is to be borne in mind when interpreting our GFR reference ranges in this age group.
Third, our data are cross-sectional rather than longitudinal. The change in GFR that we describe with age is the change in mean GFR at population level and does not necessarily describe the expected change in an individual’s GFR with ageing. Indeed, previous work has shown that there is considerable variation in GFR decline with age [
14,
39].
Finally, we did not have ethnicity data, which would have allowed us to validate previous work which has suggested that individuals of Asian ethnicity have a lower GFR [
20,
40,
41]. Whether or not ethnicity is a determinant of GFR is an important question which requires further study in large, ethnically diverse cohorts, with accurate measures of GFR.
A large multi-centre observational cohort study with prospective recruitment of potential living kidney donors, incorporating baseline and longitudinal demographic and bioclinical data with a standardized method for mGFR and other renal phenotyping would be desirable to provide robust data on the effects of age, gender, and ethnicity on GFR.