Objective: This study assessed the performance characteristics of a new microparticle enzyme immunoassay (MEIA) for the determination of sirolimus in whole blood. Background: In clinical investigatory studies, dose adjustments of the immunosuppressive drug sirolimus have been carried out using either high-performance liquid chromatography (HPLC) or, more recently, this investigational immunoassay kit based on the MEIA technique. Methods: Calibration was made over the linear range 0 to 30 ng/mL. Inaccuracy and imprecision were assessed by means of 3 control samples supplied with the kit (5, 11, and 22 ng/mL) and dilution of an above-quantitation-limit sample (154 ng/mL). Specificity was determined by the addition of 2 sirolimus metabolites to sirolimus-free human whole blood or to 1 of the control samples supplied with the kit. In addition, whole-blood samples from patients receiving either cyclosporine or tacrolimus (N = 24) were analyzed for sirolimus. A comparison of the MEIA and a validated HPLC/MS/MS assay analyzed both pooled samples from patients receiving sirolimus and spiked samples (sirolimus 2-60 ng/mL). In a more extensive comparison of patient samples measured by the MEIA assay, a validated HPLC assay with UV detection (HPLC-UV) was used (HPLC-UV sirolimus 7-64 ng/mL). Results: Inaccuracy (between-run) was less than or equal to 16.2% at all 4 concentrations (N = 5). Within-assay imprecision (repeatability) was <6% (N = 5), and between-assay imprecision (reproducibility) for the same samples was <11% (N = 5). Recovery, assessed by means of 3 inhouse control samples prepared in both fresh and previously frozen sirolimus-free human whole blood, ranged from 93.9% to 109.5%. The limit of detection, determined by dilution of the lowest nonzero calibrator (3 ng/mL), was set at 1 ng/mL, at which repeatability was 20.5% (N = 5). Five ng/mL of hydroxysirolimus cross-reacted with the assay by a mean of between 44% and 50% (N = 4); 5 ng/mL of 41-O-demethylsirolimus cross-reacted with the assay by a mean of between 86% and 127% (N = 4). Assay specificity was further challenged by ethylenediamine-tetraacetic acid (EDTA)-whole-blood samples from transplant patients not receiving sirolimus. These samples had tacrolimus and cyclosporine concentrations of 7.8 to 15.9 ng/mL and 38 to 485 mu g/L, respectively. The median result was 0 ng/mL (third quartile, 0.7 ng/mL; maximum, 1.4 ng/mL); no value was above the lowest nonzero calibrator. The results of the comparison between the MEIA and the HPLC/MS/MS assay showed mean positive biases of 21% and 8% for the MEIA in measuring sirolimus in pooled patient samples and spiked samples, respectively. The results of the comparison of the MEIA and HPLC-UV median sirolimus concentrations were 18.2 and 20.1. Whole-blood samples anticoagulated with EDTA and containing sirolimus were stable for analysis by MEIA for 3 freeze-thaw cycles when stored at -20 degrees C and for 10 days when stored at 4 degrees C or at ambient temperature. A decline in sirolimus concentration occurred when samples were stored at 37 degrees C. Conclusion: The MEIA showed suitable precision across a clinically relevant concentration range. In terms of patient management, the practical significance of cross-reactivity with sirolimus metabolites remains to be assessed.
