Determinants of the intercept and slope of glomerular filtration rate in recipients of a live donor kidney transplant

Live donor (LD) kidney transplantation is considered the best treatment option for patients with end-stage renal disease (ESRD), allowing superior outcomes in terms of patient survival, quality of life and health-related expenses. Careful LD selection, however, is essential to ensure best possible treatment results and maximum safety for both donors and recipients.

To facilitate the work-up of potential kidney donors, a variety of national and international guidelines have been formulated, most of them agreeing that glomerular filtration rate (GFR) should be evaluated with direct measurements of exogenous filtration markers, in addition to serum creatinine-based estimations of GFR [1, 2]. Levels of kidney function accepted for donation need to be adapted to the individual risk profile, but for individuals with a GFR of >90 mL/min per 1.73 m² it is generally considered safe to donate. In addition, isotope nephrography (ING) may help to determine the relative function of the kidneys supporting the choice of the nephrectomy side [1, 2].

In recent years, there has been a trend towards the acceptance of significant LD comorbidities, especially in older individuals, provided that the lifetime risk for the development of chronic kidney disease is low [3]. This may also include donors with a GFR below the commonly accepted thresholds [4]. The use of marginal donor kidneys, however, may impact substantially on allograft performance. A prominent risk factor in this respect was shown to be LD age. Large cohort studies have revealed inferior short-term and long-term outcomes for organs originating from older donors [5‐10]. Results are consistent with observations made in deceased donor (DD) kidney transplantation, where organ allocation is supported by age-matching algorithms, which take the metabolic demand of recipients into account [11].

The relative outcome effect of predonation kidney function in LD transplantation is less well studied. Norden et al. [12] observed an increased risk for graft loss in a population of 344 LD kidney transplant recipients, when donors had an unadjusted GFR below 80 mL/min. This finding is supported by a systematic review of seven studies, demonstrating associations of higher donor GFR with superior allograft function and transplant survival [13]. Definitions of GFR, however, were heterogeneous, there was no adjustment for relevant confounders, and a possible influence of unequal functional distribution between the donated and remaining kidneys was not taken into account. Moreover, none of these studies included detailed analyses of the slope of recipient estimated glomerular filtration rate (eGFR), which, confounded by a variety of immunological and non-immunological factors, may serve as a useful surrogate endpoint predicting long-term renal allograft survival [14, 15].

In this retrospective cohort study the independent impact of LD kidney function on recipient eGFR at baseline (intercept) and its slope calculated from serial eGFR measurements was investigated during the first 3 post-transplantation years. To add accuracy to the analysis the eGFR of the donated kidneys was separately calculated based on the results of Tc-99m-mercaptoacetyltriglycine acid (^99mTc-MAG3) scintigraphy. Mixed linear models were applied to quantify the impact of LD kidney function on allograft performance, in the context of other potentially outcome-related variables.

The primary objective of this study was to dissect the relative impact of LD kidney function on recipient baseline eGFR (intercept) at 6 months and eGFR slope. Major results of multivariable analysis were that the eGFR of donated kidneys, and in addition LD age, had an independent effect on allograft function at baseline, while there was no meaningful effect on eGFR slope. In line with earlier studies [23‐25], ABMR was the dominant cause of transplant functional decline, with an associated mean eGFR slope of approximately −6 mL/min/1.73 m² per year as compared to −0.2 mL/min/1.73 m² per year in the overall cohort.

LD kidney transplantation is the best available treatment option for patients with ESRD, allowing for excellent clinical outcomes, with 1‑ and 5‑year graft survival rates of 96% and 87%, respectively shown for Europe [26]. Due to substantial demographical changes, however, the demand for donor organs is constantly rising. In recent years, there has been a progressive increase in the use of older LD, who often present with additional risk factors, such as obesity, hypertension or subnormal GFR levels (even below 60 mL/min/1.73 m²) [4]. A trend towards the use of marginal donors raises major safety concerns regarding long-term LD outcomes. In addition, such variables, most prominently LD age and renal function, may also be important independent correlates of recipient allograft function. Previous studies have shown that recipients of kidneys from older donors are at an increased risk of delayed graft function, graft failure and death [5, 6]. Similar associations have been observed for subnormal predonation GFR, but in smaller recipient cohorts and in less well-designed studies [13].

For our study, we have chosen baseline eGFR‑r at 6 months and eGFR‑r slope calculated from 6‑monthly measurements as dependent variables in mixed linear models. There is increasing evidence that the extent of eGFR decline over time may serve as a valuable surrogate endpoint for long-term renal survival, in both transplantation [14, 15] and native kidney disease [27, 28]. For example, evaluating a consecutive cohort of 508 non-sensitized DD or LD renal allograft recipients, Wiebe et al. [23] described a close interrelationship between eGFR and long-term graft survival. Focusing on a specific subgroup of renal allograft recipients who developed de novo donor-specific antibodies (dnDSA), a highly significant 6% increase in the risk of post-dnDSA graft loss was calculated for each 1 mL/min/1.73 m² decease in eGFR at 3 years postsubclinical dnDSA onset [23].

We identified donor kidney function and, in accordance with previous studies [9, 10], donor age as independent predictors of baseline eGFR‑r, reinforcing the usefulness of these parameters for risk stratification of organs from potential kidney donors. In our cohort of LD kidney transplants, eGFR-dk was associated with a mean estimated increase in recipient baseline eGFR of 0.6 mL/min/1.73 m² per unit and increasing donor age was associated with a marginal decrease in recipient baseline eGFR. In contrast, we found no significant effect of total kidney function in the eGFR-dt-adjusted model. This result indicates a diagnostic benefit of ING for assessment of functional side distribution in the context of LD evaluation; however, we are aware of the limited sample size which may have precluded detection of subtle differences. For another ING-based parameter—MTT to quantify the dynamics of parenchymal tracer transit—we found no association with any of the endpoints, suggesting that this parameter may have a limited diagnostic value in the evaluation of normal functioning kidneys; however, impaired renal transit may help to dissect certain disease states, such as acute tubular injury or cyclosporine toxicity in renal transplants [29].

Interestingly, while there was a marginal effect of BMI, our study did not reveal a significant effect of LD kidney function (and age) on the slope of eGFR‑r. This finding was unexpected considering the potential functional impact of a limited renal functional reserve associated with lower donor GFR, which may ultimately cause injury due to hyperfiltration in remaining nephrons [30]; however, we want to point out that our local standard did not accept donors with an adjusted measured GFR (or a urinary creatinine clearance) <80 mL/min and/or an unequal distribution of kidney function detected by ING (>20% side difference), and this policy resulted in the overall inclusion of donated kidneys with favorable baseline function (median eGFR-dk: 43 (IQR: 38–50) mL/min/1.73 m²; median relative function: 51 (48–54) %). Our results are in line with an earlier analysis of 4488 patients, mostly DD recipients, where donor age had a significant impact on recipient eGFR at 12 months, but no influence on eGFR slope [9]. Perhaps as a result of inherent differences in case selection, which may also include marked differences in donor characteristics, other studies have revealed controversial results. For example, in a study by Issa et al. [8] changes in eGFR of LD kidney recipients over a period of 2 years after transplantation were estimated at −8.76 mL/min/1.73 m², if donors were aged ≥45 years and at −7.40 mL/min/1.73 m², if donors had an unadjusted predonation eGFR of <110 mL/min. Moreover, also in two other larger studies [7, 10], donor age was reported to be a significant determinant of progressive functional deterioration of renal allografts, in one of these studies [10], however, only beyond the first post-transplantation year.

A major finding of our study was that ABMR (10 recipients in our cohort) turned out to be the strongest predictor of annual eGFR‑r decline. The diagnosis of acute or chronic active ABMR, the leading cause of graft failure in our cohort (six of nine recorded allograft losses), was found to be associated with a mean eGFR‑r slope of approximately −6 mL/min/1.73 m² per year. This observation is consistent with the previous literature reinforcing a deleterious impact of ABMR on kidney allograft outcomes [31]. Few studies have analyzed the finding of dnDSA or diagnosis of ABMR in relation to the dynamics of eGFR decline. For example, Wiebe et al. [23] found an eGFR decline of −3.15 and −5.61 mL/min/1.73 m² per year in patients with subclinical (n = 19) and clinical (n = 45) dnDSA, respectively. Moreover, in a recent randomized controlled trial evaluating bortezomib in 44 subjects with late ABMR, eGFR slopes were about −5 mL/min/1.73 m² per year in both placebo and treatment groups [24]. Similar results (eGFR slope of approximately −7 mL/min/1.73 m² per year among 25 randomized subjects) were reported in a trial evaluating the effect of combined IVIG and rituximab in ABMR with transplant glomerulopathy [25]. The unfavorable course of allograft function in patients with ABMR, as opposed to TCMR, may reflect the current unavailability of effective therapeutic measures to counteract this type of rejection, in particular late ABMR associated with chronic irreversible injury [24, 25]. Our data reinforce the need for the establishment of effective measures to prevent or treat ABMR.

Our study has several inherent limitations. One major limitation is the comparatively small sample size, which was due to the monocentric study design and limited availability of ING data in our cohort. While we were able to dissect strong independent predictors of graft function evolution, our study may have not been sufficiently powered to detect subtle effects of some other potentially confounding variables, such as baseline immunosuppression (e.g. calcineurin inhibitors versus belatacept, which may delay progressive functional deterioration [32]). Another limitation is the intermediate-term follow-up (median 7 years), which in our cohort of LD kidney transplant recipients coincided with a low rate of graft loss (10%). Therefore, based on previous studies, we have selected eGFR slope as a surrogate endpoint, which allowed us to detect relevant outcome differences even in (i) a smaller cohort and (ii) after a shorter follow-up period. Finally, it may also be considered a limitation that our analysis was based on serum creatinine-based estimations of donor kidney function. Measured GFR was available only for half of the included LD and the resulting sample size would have been too small to detect meaningful effects. For our study, we have chosen the CKD-EPI equation, which, in contrast to other equations, such as the MDRD equation, may more accurately reflect the GFR in subjects with normal renal function [16].

While our results support that LD kidney function and age independently predict allograft function at baseline, we were not able to demonstrate a significant effect of these variables on the slope of recipient eGFR. In contrast, occurrence of ABMR turned out to be the strongest risk factor for accelerated loss of allograft function after LD kidney transplantation.