CLINICAL IMMUNOSUPPRESSION - KIDNEY EARLY

K.A. Barraclough¹, N.M. Isbel², D.W. Johnson³, G. Medley⁴, D. Leary⁵, S.B. Campbell¹, C.E. Staatz^6
1Nephrology, Princess Alexandra Hospital, Brisbane/AUSTRALIA, ²Department Of Nephrology, University of Queensland at Princess Alexandra Hospital, Brisbane/QLD/AUSTRALIA, ³Department Of Nephrology, University of Queensland at Princess Alexandra Hospital, Brisbane/AUSTRALIA, ⁴Therapeutics Research Unit, University of Queensland, Brisbane/AUSTRALIA, ⁵Nephrology, Princess Alexandra Hospital, Brisbane/QLD/AUSTRALIA, ⁶School Of Pharmacy, University of Queensland, Brisbane/AUSTRALIA

Body: Background: Prednisolone has generally been considered to have a wide therapeutic index, making therapeutic drug monitoring unnecessary. However, prednisolone results in substantial toxicities, sometimes at low doses, and marked between-subject variability in prednisolone pharmacokinetics has been demonstrated. This study examined the pharmacokinetics of free and total prednisolone and tested the predictive performance of limited sampling strategies (LSSs) for estimating prednisolone exposure in an independent cohort of adult kidney transplant recipients. Methods: Twenty pharmacokinetic profiles, comprising 13 concentration-time measurements of free and total prednisolone, were collected from 20 subjects. Total and free prednisolone was measured using high performance liquid chromatography. Two articles, presenting 10 multiple regression-derived LSSs for total prednisolone, were identified from the literature. Two of the LSSs estimated prednisolone area under the concentration-time curve (AUC) from 0 hours post-dose to ¥, while the remaining 8 estimated AUC from 0 to 6 hours post-dose. No free prednisolone LSSs were identified. Full prednisolone AUC_0-6, AUC_0-12 and AUC_0-¥ (AUCf) were calculated from all measured concentration-time points using non-compartmental analysis (trapezoidal rule). Predicted prednisolone AUC_0-6, AUC_0-12 and AUC_0-¥ (AUCp) were calculated by applying relevant concentration measurements within each of the multiple regression equations. A Pearson correlation coefficient test was applied to access the correlation between 1) total and free prednisolone, 2) each single time point concentration and AUC_0-12 and 3) AUCp and AUCf. Median percentage prediction error (MPPE) and median absolute percentage prediction error (MAPE) were calculated to evaluate bias and imprecision associated with the LSSs. Results: The median (IQR) dose adjusted AUC_0-12 for free and total prednisolone was 14.4 (8, 27) and 128.8 (103, 185) ng/h/mL/mg respectively. There was a 30-fold variation in free and 8-fold variation in total dose adjusted prednisolone exposure among study participants. Correlation (r²) between free and total prednisolone AUC_0-12 was 0.08. For both free and total prednisolone, the concentration at 6 hours post-dose (C₆) showed the highest correlation with AUC_0-12 (r² 0.84 and 0.94 respectively). Based on the 10 multiple regression derived LSSs, r² between AUCp and AUCf ranged from 0.77 to 0.99, MPPE varied from -11.2 to 10.4%, and MAPE varied from 2.3 to 11.6%. MAPE and MPPE were <15% for all of the equations. An equation that used 4 concentration measurements over the first 6 hours post-dose was superior to all other equations (AUC_0-6 = -0.985 + 1.501C₂ + 1.16C₆ + 1.095C₁ + 1.819C₄; r² 0.98, MAPE -1.8%, MAPE 2.3%). A LSS with measurements limited to the first 2 hours post-dose had acceptable predictive performance (AUC_0-6 = 9.356 + 3.522C₂ + 0.969C₁; r² 0.82, MAPE -8.7%, MAPE 8.7%). Conclusions: Kidney transplant recipients showed wide between-subject variation in prednisolone pharmacokinetics. High correlation existed between prednisolone single concentration-time measures and AUC_0-12 and all LSSs showed good predictive power. Poor correlation existed between free and total prednisolone exposure, suggesting free measurement may be preferable.

Disclosure: All authors have declared no conflicts of interest.