Rouvick Mariano Gama, Tayeba Roper, Nikhil Johri, Hugh Gallagher, Mysore Keshavmurthy Phanish
For evaluation of renal safety under PRRT, clinically relevant KDIGO deterioration and CTCAE grading are not interchangeable. Clinically relevant KDIGO-derived eGFRcr worsening represents a stricter filtration-based endpoint, whereas early CTCAE eGFRcr worsening may help identify patients requiring closer post-treatment renal surveillance. Prospective studies integrating serial filtration markers, urine biomarkers, measured GFR, and individualized kidney dosimetry are warranted, especially in light of current ongoing trials investigating PRRT being moved to an earlier stage in the treatment course.
BACKGROUND AND HYPOTHESIS: Accurate assessment of glomerular filtration rate (GFR) is essential during evaluation of prospective kidney donors. While estimated GFR (eGFR) equations are widely used in clinical practice, evidence regarding their performance in individuals with preserved renal function remains limited. We evaluated the agreement of creatinine, cystatin-C, and combined biomarker eGFR equations against measured GFR (mGFR) calculated using iohexol plasma clearance in healthy prospective kidney donors.
METHODS: We conducted a retrospective observational study of adults undergoing kidney donor assessment at a tertiary UK centre between 2015 and 2018. Serum creatinine and cystatin-C were compared with iohexol plasma clearance. Ten validated eGFR equations were evaluated. Agreement was assessed using Lin's concordance correlation coefficient, bias, P10, P30, and Bland-Altman analyses. Sensitivity and specificity were calculated for each eGFR equation against British Transplant Society age- and sex-specific mGFR donor suitability thresholds.
RESULTS: One hundred seven participants were included; 50.5% were females, mean age was 48.5 ± 11.8 years and mean mGFR was 86.5 ± 14.6 ml/min/1.73 m2. The Lund-Malmo revised equation (LMR) had the most favourable agreement metrics, followed by European Kidney Function Consortium (EKFC). LMR had the highest correlation coefficient (r = 0.70; 95% confidence intervals (CI): 0.58 to 0.78), smallest bias (-1.3 ± 10.9 ml/min/1.73 m2), highest P30 (97.2%; 95% CI: 94.1-100), and highest P10 (57.0%; 95% CI 47.6-66.4). P30 exceeded 80% across most equations, whereas P10 was lower, ranging from 29.0% to 57.0%. Sensitivity for identifying donor suitability thresholds was generally high, ranging from 75.6% to 98.7%, whereas specificity was lower, ranging from 17.2% to 58.6%.
CONCLUSIONS: In healthy kidney donors with preserved kidney function, creatinine-based equations, particularly LMR and EKFC, demonstrated improved agreement with iohexol plasma clearance than cystatin-C-based approaches. Despite acceptable population-level performance, low P10 and substantial variability support continued use of mGFR during evaluation of prospective kidney donors.