*Purpose: Predicting post-donation GFR remains a challenge during living kidney donor (LKD) evaluation. Iothalamate split scans have been employed during LKD evaluation to document measured GFR (mGFR) and split renal function. However, the utility of these factors to predict post-donation eGFR is unknown. Therefore, we compared equations using pre-donation split scan results versus pre-donation estimated GFR (eGFR) values to predict 6-month post-donation kidney function.

*Methods: All LKD nephrectomies from 06/2015-02/2018 at a single institution were identified (N=221). All participants underwent protocol pre-donation iothalamate split scan and serum creatinine measurement. The CKD-EPI formula was used for all eGFR values. Participants were excluded if lost to follow-up within 6 months of donation (n=15). A training (n=137) and validation set (n=69) were randomized. Stepwise linear regression was used to construct equations that were based on pre-donation split scan results (Equation 1) or pre-donation eGFR (Equation 2) to predict 6-month post-donation eGFR in the training set. Equations were compared by multiple correlation coefficient (R) and explained variance (R²) in the training cohort; root mean square error (RMSE) comparison was used to assess for overfitting.

*Results: Of 207 study participants, 83 (40.3%) were male and 15 (7.3%) were black with mean age 44±13 years and mean BMI 27.0±4.3 kg/m². The mean pre-donation mGFR was 104.4±17.6 mL/min/1.73m², mean pre-donation eGFR was 101.4±14.8 mL/min/1.73m², and mean 6-month post-donation eGFR was 66.5±14.5 mL/min/1.73m². The mean split proportion of the non-donated kidney was 48.9% (range 38.7-61.0%). Derived equations are shown in Fig 1. In the training cohort, Equation 1 had an R=0.56 (R²=0.32) and Equation 2 had an R=0.76 (R²=0.58) (all P<.001). Comparison of the RMSE in each cohort did not suggest overfitting [Fig 1]. In the validation cohort, Equation 1 over-predicted 6-month eGFR by >20 mL/min/1.73m² in 4/69 (5.8%) patients compared to none with Equation 2 [Fig 2].

*Conclusions: The equation derived from pre-donation eGFR correlated better with 6-month post-donation eGFR than that derived from the split scan, with few clinically-relevant over-prediction episodes. These findings question the clinical utility of split-scan determined GFR in our study. There are subsets of patients, such as those with disproportionate kidney sizes or borderline GFR, that require further study.

« Back to 2020 American Transplant Congress