- Email: [email protected]
Comparison of fluorescent polarization immunoassay (FPIA) versus HPLC to measure everolimus blood concentrations in clinical transplantation
Clinica Chimica Acta 380 (2007) 217 – 221 www.elsevier.com/locate/clinchim
Short communication
Comparison of fluorescent polarization immunoassay (FPIA) versus HPLC to measure everolimus blood concentrations in clinical transplantation GholamAli Khoschsorur ⁎, Franz Fruehwirth, Sieglinde Zelzer, Mariana Stettin, Gabriele Halwachs-Baumann Institute of Medical and Chemical Laboratory Diagnostics (LB I), Medical University of Graz, Auenbruggerplatz 29, A-8036 Graz, Austria Received 27 September 2006; received in revised form 17 January 2007; accepted 17 January 2007 Available online 25 January 2007
Abstract Clinical management of transplant patients depends on therapeutic drug monitoring (TDM) and regulation of immunosuppressive therapy. TDM of whole-blood concentrations is mandatory for everolimus (ERL) dosage individualisation. We compared the new semi-automated immunoassay (Innofluor® Certican® Assay System, Seradyn Inc) using FPIA technology on Abbott TDxFLx® analyzers with established HPLCUV as reference method. A total of 165 samples were analyzed from 52 transplant patients (40 kidney, 12 heart) receiving ERL or another immunosuppressive agent as part their routine care after organ transplantation. The correlation coefficient was r2 = 0.8229, and the regression equation (95% IC) yielded FPIA = 1.111 × (HPLC) + 0.378. FPIA compared to HPLC gave a positive bias of 1.19 ng/ml. The FPIA assay so appears to have a diagnostic efficacy comparable to HPLC for assessing the risk of acute rejection in transplant recipients. However, the values of the FPIA were higher than those calculated from HPLC measurements, because of the cross-reactivity of the antibody used in the FPIA assay with the ERL metabolite and/or with sirolimus; this cross-reactivity occurs frequently when transplant patients are switched from sirolimus to ERL. © 2007 Elsevier B.V. All rights reserved. Keywords: Comparative study; Everolimus; High-performance liquid chromatograph (HPLC-UV); Fluorescence polarization immunoassay (FPIA)
1. Introduction Everolimus (Certican®, SDZ-RAD, ERL, Novartis Pharmaceuticals, Basel, Switzerland), a macrolide immunosuppressant derived by chemical modification of the natural product sirolimus (SRL, Rapamycin), is a potent immunosuppressive agent used in combination with cyclosporine A or mycophenolate mofetil to prevent rejection of solid organ grafts [1,2]. Preclinical studies have demonstrated that the mechanism of action of ERL and the synergistic activity of ERL when it is combined with cyclosporine are quite similar to those of ⁎ Corresponding author. Institute of Medical and Chemical, Laboratory Diagnostics (LB I), Auenbruggerplatz 29, A-8036 Graz/Austria. Fax: +43 316 385 3419. E-mail address: [email protected] (G. Khoschsorur). 0009-8981/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2007.01.017
rapamycin [3–5]. Everolimus belongs to the same drug class as sirolimus, but relatively minor structural differences between the two agents confer different pharmacokinetic characteristics [6–10]. ERL is more hydrophilic and has a shorter half-life (the mean terminal half-life is 28 h) as well as a higher relative bioavailability than SRL [7,11,12]. SRL also has a long half-life of around 59 ± 19 h, which may be extended significantly to more than 100 h in liver dysfunction [9,13]. With increasing numbers of transplant patients and new protocols for immunosuppressant treatment (multi-drug strategy), ever more patients are switched from sirolimus to everolimus because of intolerance, and combinations with other immunosuppressants are common clinical practice. When a switch is made from SRL to ERL, SRL is still present and so the everolimus measurements in blood from these patients were in general higher with the immunoassay than with the chromatographic
218
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
method. Here, the cross-reactivities of the ERL metabolites and SRL with immunoassay antibodies play a major role. Therapeutic drug monitoring (TDM) plays a key role in maintaining blood ERL concentration within its narrow therapeutic range of 3–8 ng/ml [14] and may help to optimize dosage regimens, maintain immunosuppressive efficacy and minimize potential toxicity [15–17]. The generation of ERL metabolites could be attributed to cytochrome P450 3A4, the most abundant of the CYP enzymes in the liver and intestine; this is also the case with cyclosporine and tacrolimus [6,18–20]. The biotransformation of ERL by human liver microsomes produces at least 20 metabolites and those identified in the blood of human renal transplant recipients include hydroxyl-, dihydroxy-, demethyl-and ring-opened forms of everolimus [21–23]. The major metabolites, for the most part, are pharmacologically inactive and include hydroxyl-, (24/25 OH-RAD, 46 OH-RAD), ring-open compounds (RAD-SA, RAD-PSA) and 40-phosphatidylcholine-RAD (RAD-PC), whose immunosuppressive activity is currently unknown [22]. A number of high performance liquid chromatography (HPLC) methods with ultraviolet detection or mass spectrometry (LC/MS-MS) [23–28] are available for individual determination of ERL concentration in whole blood. In 2004, a new homogeneous fluorescence polarization immunoassay (FPIA) for the analysis of this drug became available. The aim of this study was to evaluate the possibility of estimating the concentration of blood ERL concentrations with the new semi-automated immunoassay (Innofluor® Certican®Assay System, Seradyn Inc, IN, USA) using FPIA technology on an Abbott TDxFLx analyzer and to analyze the degree of correlation with HPLC-UV as reference method. The chromatographic method is in routine clinical use in our lab. 2. Materials and methods For the comparison with FPIA and HPLC, a total of 165 whole-blood (EDTA) specimens were collected from 52 transplant patients; 112 samples (kidney, n = 40) and 53 samples (heart, n = 12) were from patients receiving ERL with other immunosuppressive agents as part of their routine care after transplantation. All specimens for trough concentrations were taken just before the morning administration of the respective immunosuppressant dosage and selected from routine clinical orders for ERL quantification. Specimens were analyzed with both methods on the same day or were stored at − 20 °C for further analysis within 48 h after collection. Everolimus was assayed on a TDxFLx analyzer (Abbott Diagnostics, IL, USA) with a semi-automated Certican immunoassay system (Innofluor® Certican®Assay System, Seradyn Inc, IN, USA), which is a homogeneous fluorescence polarization immunoassay. Specimen pretreatment prior to ERL-FPIA analysis followed the manufacturer's instructions. An aliquot (600 μl) of each calibrator, control, or patient sample was transferred with a positive pressure displacement fixed-volume pipette to an appropriate centrifuge tube containing 700 μl of methanol and 100 μl of Innofluor Certican precipitation reagent. Each tube was capped and vortexed vigorously at the highest speed for at least 10–15 s (up to complete suspension). Following centrifugation at 13,400 ×g for 8 min, 300 μl supernatant was transferred into sample cartridges and loaded into the carousel for analysis on the TDx/TDxFLx analyzer. The assay was calibrated using a calibration curve with 6 calibrators (A-F; 0.00, 2.24, 4.80, 9.82, 19.70, and 40.28 ng/ml). Calibration curve stability was measured on days 1, 7, 14, 21 and 28 during the study. We defined the stability of ERL calibration curves over four weeks on the TDxFLx analyzer; the
calibrators and three controls (low; 3.9 ng/ml, medium; 11.3 ng/ml, and high; 25.4 ng/ml) provided by the manufacturer were tested in duplicate, the same as patient samples and according to the manufacturer's instructions. The ERL patient samples were measured by HPLC using a previously described method [27] that can determine sirolimus and everolimus concentrations simultaneously. Calibrators and in-house control samples for HPLC were prepared in separate drug-free human whole-blood stock solutions: 6 non-zero concentrations (nominal values of 2.5, 5.0, 10.0, 20.0, 40.0, and 80.0 ng/ml), 4 quality controls with different concentrations (nominal values of 2.5, 7.5, 20.0, and 40.0 ng/ml ng/ml) for SRL and SDZ-ERL. Calibrators and controls were aliquoted into polypropylene tubes (Corning Inc.) and stored at approximately −20 °C before use [27]. The results were evaluated with the classical correlation methods of Passing–Bablok [29] (linear regression, sum of deviation squares) and the procedures recommended by Bland and Altman [30] applying SPSS (v 12.0).
3. Results and discussion FPIA assay precision was studied using clinical samples as well as quality control samples (no ERL metabolites present) at 3.9 (low), 11.30 (medium), and 25.40 ng/ml (high) provided by the manufacturer and prepared in the same synthetic matrix as the calibrating standards. The within-run precision was performed by analyzing each of the three controls (low, medium, and high) in replicates (n = 5) were 9.9% (low), 6.4% (medium), and 5.3% (high) controls. The between-day precision was performed by analyzing these controls over 10 consecutive days. The subsequent between-run analyses of the coefficients of variation (CVs) of low, medium and high controls of both methods were similar: 10.9%, 6.2%, and 5.8%, at 3.9, 11.3, and 25.4 ng/ml, respectively, for FPIA. Table 1 shows the results of analyses of these commercial controls compared with values provided by the manufacturer for ERL-FPIA and 11.4%, 8.7%, 7.8%, and 4.2% at 2.47 ng/ml, 4.43 ng/ml, 8.82 ng/ml, and 30.1 ng/ml for HPLC/UV in addition. Our data from FPIA compared well with the manufacturer's data of 6.0% and 9.0% for within run, and 12.0%, 5.0% and 5.0% for between run. The lower limit of detection of FPIA, determined as the mean + 3 SD of 10 replicates of drug-free EDTA blood, was 0.8 ng/ml. The lower limit of quantification was designated as 2.2 ng/ml, the value of the lowest calibrator. The latter was Table 1 Results of analyses of commercial quality control samples compared with values provided by manufacturer for ERL-FPIA (n = 10) and for ERL-HPLC (n = 5) Nominal value, range a (ng/ml)
Analyte found b (ng/ml, mean ± S.D.)
CV (%)
ERL-FPIA Control low Control medium Control high
3.9 (2.73–4.96) 11.3 (9.09–13.50) 25.4 (19.46–31.25)
3.31 ± 0.36 11.81 ± 0.73 21.56 ± 1.25
10.9 6.2 5.8
ERL-HPLC Level 1 Level 2 Level 3 Level 4
2.47 (1.61–3.33) 4.43 (2.88–5.98) 8.82 (5.73–11.9) 30.1 (21.1–39.1)
2.72 ± 0.31 3.92 ± 0.34 8.07 ± 0.63 31.8 ± 1.32
11.4 8.7 7.8 4.2
Sample analyte
a b
Values provided by the manufacturers. Results obtained in our laboratory.
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
similar to that for HPLC (2.5 ng/ml) and is acceptable for routine clinical monitoring. For intraday precision, the samples were stored at −20 °C and measured on the same day; for interday precision, we used the mean values of quality controls measured on 6 different days. The intra- and interday imprecision values (CVs) for HPLC were 3.2%–10% and 2.0%–9.6%, respectively, for SRL and 4.0%–9.3% and 3.4%–8.3% for everolimus (28). The limit of detection for HPLC, defined as the concentration of drug giving a signal-to-noise ratio N 3:1 was 1.0 ng/ml. The FPIA results (y-axis) were compared with those from the HPLC/UV (x-axis) in 165 whole-blood samples from kidney (n = 112) and heart (n = 53) transplant recipients, by Passing– Bablock estimation. The correlation function was FPIA = 1.11 (1.039 − 1.200) × (HPLC-UV) + 0.38 (− 0.080 − 0.807), with a correlation coefficient of r2 = 0.8229 for different types of transplant (Fig. 1A). Passing–Bablok regression analysis revealed that FPIA measured ERL concentrations that were significantly higher than with HPLC. Our results are similar to
219
Table 2 The cross-reactivity of the Innofluor® Certican® assay with the structurally related compound sirolimus, as well as metabolites of everolimus Everolimus metabolites
Concentration tested (ng/ml)
Percent (%) cross-reactivity
RAD SA RAD PSA RAD PC 25-OH RAD Sirolimus
5 5 5 50 10
5 12 72 6 74
As shown in the manufacturer's reagent package insert.
literature data for FPIA vs. LC/MS-MS from renal transplant patients' samples (n = 333) as measured by both methods using Deming regression analysis; FPIA = 1.19 (±0.03) × HPLC-MS + 0.51 (±0.21) μg/l [31]. The statistical analysis according to Altman and Bland provides further information on differences that can be expected with a 95% probability when two methods are used for the determination of ERL concentration in one blood sample. FPIA compared to HPLC gave a positive bias of 1.19 ng/ml, with 95% limits of agreement of − 1.37 to 3.76 ng/ml (Fig. 1B). Thus, both statistical procedures proved the analysis results for the two methods to be comparable. Overall, there was a tendency for FPIA assay to give higher ERL values than HPLC. The mean ERL concentration of patient samples using HPLC-UV was 5.66 ± 2.79 ng/ml (range 1.3 to 18.1 ng/ml) compared with 6.86 ± 3.11 ng/ml (range 2.05 to 20.2 ng/ml) using the immunoassay. This highlights the fact that the differences between the two methods are probably due to differences in the calibrators and likely to cross-reactivity between ERL metabolites, SRL and the antibody in the FPIA method (e.g. ERL-PC; 72%, and with sirolimus 74%). Table 2 shows the manufacturer's data for some of the ERL metabolites and SRL with their chosen antibody. It shows that the monoclonal antibody-based ERL immunoassay can demonstrate variable, and sometimes significant, cross-reactivity with various ERL metabolites. Strom et al. [32] report that the major hydroxy everolimus metabolites in the blood of patients pose insignificant cross-reactivity with the antibody used in the Innofluor Certican immunoassay (4–6% cross-reactivity with the 25-OH-RAD). With increasing numbers of transplant patients and new protocols for immunosuppressant treatment, patients may be switched from SRL to ERL because of intolerance. The degree Table 3 Results for patients switched from sirolimus to everolimus (HPLC and FPIA)
Fig. 1. Comparison of results of everolimus trough concentrations in 165 blood samples from transplant patients determined by HPLC and FPIA assay. (A) Passing–Bablok regression analysis between both methods shows the correlation function (solid line): FPIA = 1.111 × HPLC + 0.378. The solid and the dashed lines indicate, respectively, the regression line and unity. (B) The degree of agreement between methods was assessed using the Bland–Altman graphical technique. The mean difference between the methods was 1.19 ± 2.57.The solid line indicates the mean difference between the methods, and the 95% confidence intervals for the difference are indicated by dashed lines.
Patient 1 Patient 1 a Patient 2 Patient 3 Patient 4 Patient 5 Patient 6 a
ERL(FPIA)
ERL(HPLC)
SRL(HPLC)
ng/ml
ng/ml
ng/ml
10.1 7.7 10.6 7.4 4.6 4.1 10.7
4.7 4.8 7.6 4.0 2.0 2.2 5.1
5.1 1.2 2.8 2.8 1.9 1.6 5.4
Patient level after 5 days.
220
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
of cross-reactivity of the sirolimus metabolites with the FPIA assay could not be determined due to the lack of those substances in their pure forms. There is only the report of Jacobsen et al. [18] that in vitro metabolism of structurally similar ERL and SRL shows significant differences. For some laboratories, the relative technical simplicity and short run time of the immunoassay could make it an attractive alternative to HPLC for the measurement of ERL. The assay does not, however, measure the extent of accumulation of metabolites and their cross-reactivity with other immunosuppressants that may be present (e.g. SRL). This is often the case as ever more transplant patients are switched from sirolimus to everolimus, and requires more precise methods for determining ERL concentrations. For six patients who were switched from SRL to ERL, the immunological analysis produced a value that was more than 100% higher than with HPLC (e.g., a heart-transplant patient had an ERL value with the FPIA assay of 10.1 ng/ml and of 4.7 ng/ml with HPLC, with a measured SRL value of ng/ml). Table 3 shows the individual ERL and SRL results of 7 samples from 6 switched transplant patients with percentage differences. Detectable SRL values from patients whose medication had been changed were found for one week after the change. This means that daily monitoring of ERL from these patients with FPIA assay shows cross-reactivity with sirolimus, because of the long half-life of SRL. The percentage difference between the immunoassay and HPLC/UV shows that a monoclonal antibody-based assay is affected by drug metabolites. These studies show that the performance of some of the most used immunoassays is not satisfactory, and can lead to incorrect therapeutic decisions guided by erroneous immunosuppressive monitoring. The HPLC method is considered better suited for TDM, and so for application in clinical strategies to improve transplantation tolerance as well. In conclusion, the clinical interpretation of ERL concentrations determined with the FPIA assay is complicated by the ERL metabolite profile and co-medications that may be present. The main methodological problems with the measurement of immunosuppressive drug concentrations in the serum are generally due to the cross-reactivity of metabolites of the parent drug with antibodies used in immunoassays [33]. The effectiveness of ERL therapy may greatly depend on the method of determination, because the cross-reactivity is usually considered to be the major disadvantage of the FPIA immunoassay. Acknowledgements We thank Ms. Eugenia Lamont for editorial assistance and Ms. Andrea Prattes for technical assistance. References [1] Vitko S, Margreiter R, Weimar W, et al. Everolimus (certican) 12-month safety and efficacy versus mycophenolate mofetil in de novo renal transplant recipients. Transplantation 2004;78:1532–40. [2] Lorber MI, Mulgaonkar S, Butt KM, et al. Everolimus versus mycophenolate mofetil in the prevention of rejection in de novo renal transplant recipients: a 3-year randomized, multicenter, phase III study. Transplantation 2005;80:244–52.
[3] Schuurman HJ, Cottens S, Fuchs S, et al. SDZ RAD, a new rapamycin derivative: synergism with cyclosporine. Transplantation 1997;64:36–42. [4] Schuler W, Sedrani R, Cottens S, et al. SDZ RAD, a new rapamycin derivative: pharmacological properties in vitro and in vivo. Transplantation 1997;64:32–5. [5] Dumont FJ. Everolimus. Curr Opin Investig Drugs 2001;2:1220–34. [6] Kirchner GI, Meier-Wiedenbach I, Manns MP. Clinical pharmacokinetics of everolimus. Clin Pharmacokinet 2004;43:83–95. [7] Neumayer HH, Paradis K, Korn A, et al. Entry-into-human study with the novel immunosuppressant SDZ RAD in stable renal transplant recipients. Br J Clin Pharmacol 1999;48:694–703. [8] Kahan BD, Wong RL, Carter C, et al. A phase I study of a 4-week course of SDZ-RAD(RAD) in quiescent cyclosporine-prednisonetreated renal transplant recipients. Transplantation 1999;68:1100–6. [9] MacDonald A, Scarola J, Burke JT, Zimmerman JJ. Clinical pharmacokinetics and therapeutic drug monitoring of sirolimus. Clin Ther 2000;22: B101–21 [Suppl.]. [10] Gallant-Haidner HL, Trepanier DJ, Freitag DG, Yatscoff RW. Pharmacokinetics and metabolism of sirolimus. Ther Drug Monit 2000;22:31–5. [11] Sedrani R, Cottens S, Kallen J, Schuler W. Chemical modification of rapamycin: the discovery of SDZ RAD. Transplant Proc 1998;30:2192–4. [12] Kovarik JM, Kahan BD, Kaplan B, et al. Longitudinal assessment of everolimus in de novo renal transplant recipients over the first posttransplant year: pharmacokinetics, exposure-response relationships, and influence on cyclosporine. Clin Pharmacol Ther 2001;69:48–56. [13] Mahalati K, Kahan BD. Clinical pharmacokinetics of sirolimus. Clin Pharmacokinet 2001;40:573–85. [14] Kovarik JM, Eisen H, Dorent R, et al. Everolimus in de novo cardiac transplantation: pharmacokinetics, therapeutic range, and influence on cyclosporine exposure. J Heart Lung Transplant 2003;22:1117–25. [15] Kahan BD, Keown P, Levy GA, Johnston A. Therapeutic drug monitoring of immunosuppressant in clinical practice. Clin Ther 2002;24:330–50. [16] Schuurman HJ, Ringers J, Schuler W, Slingerland W, Jonker M. Oral efficacy of the macrolide immunosuppressant SDZ RAD and of cyclosporine microemulsion in cynomolgus monkey kidney allotransplantation. Transplantation 2000;69:737–42. [17] Serkova N, Hausen B, Berry GJ, et al. Tissue distribution and clinical monitoring of the novel macrolide immunosuppressant SDZ-RAD and its metabolites in monkey lung transplant recipients: interaction with cyclosporine. J Pharmacol Exp Ther 2000;294:323–32. [18] Jacobsen W, Serkova N, Hausen B, Morris RE, Benet LZ, Christians U. Comparison of the in vitro metabolism of the macrolide immunosuppressants sirolimus and RAD. Transplant Proc 2001;33:514–5. [19] Börnsen KO, Inverardi B, Gschwind HP. Electrospray ionization and collisional induced dissociation of RAD001 and related compounds and structural characterization of RAD001 metabolites by nano-spray and micro liquid chromatography mass spectrometry. Novartis Inter Rep 2001: R98–R709 [DMPK(CH)]. [20] Kovarik JM. Everolimus: a proliferation signal inhibitor targeting primary causes of allograft dysfunction. Drugs Today (Barc) 2004;40:101–9. [21] Vidal C, Kirchner GI, Sewing KF. Structural elucidation by electrospray mass spectrometry: an approach to the in vitro metabolism of the macrolide immunosuppressant SDZ RAD. J Am Soc Mass Spectrom 1998;9: 1267–74. [22] Kirchner GI, Winkler M, Mueller L, et al. Pharmacokinetics of SDZ RAD and cyclosporine including their metabolites in seven kidney graft patients after the first dose of SDZ RAD. Br J Clin Pharmacol 2000;50:449–54. [23] Kirchner GI, Vidal C, Winkler M, et al. LC/ESI-MS allows simultaneous and specific quantification of SDZ RAD and cyclosporine, including groups of their metabolites in human blood. Ther Drug Monit 1999;21:116–22. [24] Koal T, Deters M, Casetta B, Kaever V. Simultaneous determination of four immunosuppressants by means of high speed and robust on-line solid phase extraction–high performance liquid chromatography–tandem mass spectrometry. J Chromatogr, B, Biomed Sci Appl 2004;805:215–22. [25] Ceglarek U, Lembcke J, Fiedler GM, et al. Rapid simultaneous quantification of immunosuppressants in transplant patients by turbulent flow chromatography combined with tandem mass spectrometry. Clin Chim Acta 2004;346:181–90.
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221 [26] Streit F, Amstrong VW, Oellerich M. Rapid liquid chromatographytandem mass spectrometry routine method for simultaneous determination of sirolimus, everolimus, tacrolimus, and cyclosporin A. Clin Chem 2002;48:955–8. [27] Khoschsorur G. Simultaneous measurement of sirolimus and everolimus in whole blood by high-performance liquid chromatography with ultraviolet detection. Clin Chem 2005;9:1721–4. [28] Baldelli S, Zenoni S, Merlini S, Perico N, Cattaneo Dario. Simultaneous determination of everolimus and cyclosporine concentrations by HPLC with ultraviolet detection. Clin Chim Acta 2006;364:354–8. [29] Passing H, Bablok W. A new biometrical procedure for testing the equality of measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, part I. J Clin Chem Clin Biochem 1983;21:709–20.
221
[30] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [31] Salm P, Warnholtza C, Boydb J, Arabshahi L, Marbach P, Taylor PJ. Evaluation of a fluorescent polarization immunoassay for whole blood everolimus determination using samples from renal transplant recipients. Clin Biochem 2006;39:732–8. [32] Strom T, Boyd J, Arabshahi L, Marbachc P, Christiansa U. Cross-reactivity of the major everolimus hydroxy metabolites with the innofluor certican immunoassay for clinical Tdm of everolimus. Ther Drug Monit 2005;27:243 [Abstract]. [33] Holt DW, Johnston A. Monitoring new immunosuppressive agent. Are the methods adequate? Drug Metab Drug Interac 1997;14:5–15.
Short communication
Comparison of fluorescent polarization immunoassay (FPIA) versus HPLC to measure everolimus blood concentrations in clinical transplantation GholamAli Khoschsorur ⁎, Franz Fruehwirth, Sieglinde Zelzer, Mariana Stettin, Gabriele Halwachs-Baumann Institute of Medical and Chemical Laboratory Diagnostics (LB I), Medical University of Graz, Auenbruggerplatz 29, A-8036 Graz, Austria Received 27 September 2006; received in revised form 17 January 2007; accepted 17 January 2007 Available online 25 January 2007
Abstract Clinical management of transplant patients depends on therapeutic drug monitoring (TDM) and regulation of immunosuppressive therapy. TDM of whole-blood concentrations is mandatory for everolimus (ERL) dosage individualisation. We compared the new semi-automated immunoassay (Innofluor® Certican® Assay System, Seradyn Inc) using FPIA technology on Abbott TDxFLx® analyzers with established HPLCUV as reference method. A total of 165 samples were analyzed from 52 transplant patients (40 kidney, 12 heart) receiving ERL or another immunosuppressive agent as part their routine care after organ transplantation. The correlation coefficient was r2 = 0.8229, and the regression equation (95% IC) yielded FPIA = 1.111 × (HPLC) + 0.378. FPIA compared to HPLC gave a positive bias of 1.19 ng/ml. The FPIA assay so appears to have a diagnostic efficacy comparable to HPLC for assessing the risk of acute rejection in transplant recipients. However, the values of the FPIA were higher than those calculated from HPLC measurements, because of the cross-reactivity of the antibody used in the FPIA assay with the ERL metabolite and/or with sirolimus; this cross-reactivity occurs frequently when transplant patients are switched from sirolimus to ERL. © 2007 Elsevier B.V. All rights reserved. Keywords: Comparative study; Everolimus; High-performance liquid chromatograph (HPLC-UV); Fluorescence polarization immunoassay (FPIA)
1. Introduction Everolimus (Certican®, SDZ-RAD, ERL, Novartis Pharmaceuticals, Basel, Switzerland), a macrolide immunosuppressant derived by chemical modification of the natural product sirolimus (SRL, Rapamycin), is a potent immunosuppressive agent used in combination with cyclosporine A or mycophenolate mofetil to prevent rejection of solid organ grafts [1,2]. Preclinical studies have demonstrated that the mechanism of action of ERL and the synergistic activity of ERL when it is combined with cyclosporine are quite similar to those of ⁎ Corresponding author. Institute of Medical and Chemical, Laboratory Diagnostics (LB I), Auenbruggerplatz 29, A-8036 Graz/Austria. Fax: +43 316 385 3419. E-mail address: [email protected] (G. Khoschsorur). 0009-8981/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2007.01.017
rapamycin [3–5]. Everolimus belongs to the same drug class as sirolimus, but relatively minor structural differences between the two agents confer different pharmacokinetic characteristics [6–10]. ERL is more hydrophilic and has a shorter half-life (the mean terminal half-life is 28 h) as well as a higher relative bioavailability than SRL [7,11,12]. SRL also has a long half-life of around 59 ± 19 h, which may be extended significantly to more than 100 h in liver dysfunction [9,13]. With increasing numbers of transplant patients and new protocols for immunosuppressant treatment (multi-drug strategy), ever more patients are switched from sirolimus to everolimus because of intolerance, and combinations with other immunosuppressants are common clinical practice. When a switch is made from SRL to ERL, SRL is still present and so the everolimus measurements in blood from these patients were in general higher with the immunoassay than with the chromatographic
218
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
method. Here, the cross-reactivities of the ERL metabolites and SRL with immunoassay antibodies play a major role. Therapeutic drug monitoring (TDM) plays a key role in maintaining blood ERL concentration within its narrow therapeutic range of 3–8 ng/ml [14] and may help to optimize dosage regimens, maintain immunosuppressive efficacy and minimize potential toxicity [15–17]. The generation of ERL metabolites could be attributed to cytochrome P450 3A4, the most abundant of the CYP enzymes in the liver and intestine; this is also the case with cyclosporine and tacrolimus [6,18–20]. The biotransformation of ERL by human liver microsomes produces at least 20 metabolites and those identified in the blood of human renal transplant recipients include hydroxyl-, dihydroxy-, demethyl-and ring-opened forms of everolimus [21–23]. The major metabolites, for the most part, are pharmacologically inactive and include hydroxyl-, (24/25 OH-RAD, 46 OH-RAD), ring-open compounds (RAD-SA, RAD-PSA) and 40-phosphatidylcholine-RAD (RAD-PC), whose immunosuppressive activity is currently unknown [22]. A number of high performance liquid chromatography (HPLC) methods with ultraviolet detection or mass spectrometry (LC/MS-MS) [23–28] are available for individual determination of ERL concentration in whole blood. In 2004, a new homogeneous fluorescence polarization immunoassay (FPIA) for the analysis of this drug became available. The aim of this study was to evaluate the possibility of estimating the concentration of blood ERL concentrations with the new semi-automated immunoassay (Innofluor® Certican®Assay System, Seradyn Inc, IN, USA) using FPIA technology on an Abbott TDxFLx analyzer and to analyze the degree of correlation with HPLC-UV as reference method. The chromatographic method is in routine clinical use in our lab. 2. Materials and methods For the comparison with FPIA and HPLC, a total of 165 whole-blood (EDTA) specimens were collected from 52 transplant patients; 112 samples (kidney, n = 40) and 53 samples (heart, n = 12) were from patients receiving ERL with other immunosuppressive agents as part of their routine care after transplantation. All specimens for trough concentrations were taken just before the morning administration of the respective immunosuppressant dosage and selected from routine clinical orders for ERL quantification. Specimens were analyzed with both methods on the same day or were stored at − 20 °C for further analysis within 48 h after collection. Everolimus was assayed on a TDxFLx analyzer (Abbott Diagnostics, IL, USA) with a semi-automated Certican immunoassay system (Innofluor® Certican®Assay System, Seradyn Inc, IN, USA), which is a homogeneous fluorescence polarization immunoassay. Specimen pretreatment prior to ERL-FPIA analysis followed the manufacturer's instructions. An aliquot (600 μl) of each calibrator, control, or patient sample was transferred with a positive pressure displacement fixed-volume pipette to an appropriate centrifuge tube containing 700 μl of methanol and 100 μl of Innofluor Certican precipitation reagent. Each tube was capped and vortexed vigorously at the highest speed for at least 10–15 s (up to complete suspension). Following centrifugation at 13,400 ×g for 8 min, 300 μl supernatant was transferred into sample cartridges and loaded into the carousel for analysis on the TDx/TDxFLx analyzer. The assay was calibrated using a calibration curve with 6 calibrators (A-F; 0.00, 2.24, 4.80, 9.82, 19.70, and 40.28 ng/ml). Calibration curve stability was measured on days 1, 7, 14, 21 and 28 during the study. We defined the stability of ERL calibration curves over four weeks on the TDxFLx analyzer; the
calibrators and three controls (low; 3.9 ng/ml, medium; 11.3 ng/ml, and high; 25.4 ng/ml) provided by the manufacturer were tested in duplicate, the same as patient samples and according to the manufacturer's instructions. The ERL patient samples were measured by HPLC using a previously described method [27] that can determine sirolimus and everolimus concentrations simultaneously. Calibrators and in-house control samples for HPLC were prepared in separate drug-free human whole-blood stock solutions: 6 non-zero concentrations (nominal values of 2.5, 5.0, 10.0, 20.0, 40.0, and 80.0 ng/ml), 4 quality controls with different concentrations (nominal values of 2.5, 7.5, 20.0, and 40.0 ng/ml ng/ml) for SRL and SDZ-ERL. Calibrators and controls were aliquoted into polypropylene tubes (Corning Inc.) and stored at approximately −20 °C before use [27]. The results were evaluated with the classical correlation methods of Passing–Bablok [29] (linear regression, sum of deviation squares) and the procedures recommended by Bland and Altman [30] applying SPSS (v 12.0).
3. Results and discussion FPIA assay precision was studied using clinical samples as well as quality control samples (no ERL metabolites present) at 3.9 (low), 11.30 (medium), and 25.40 ng/ml (high) provided by the manufacturer and prepared in the same synthetic matrix as the calibrating standards. The within-run precision was performed by analyzing each of the three controls (low, medium, and high) in replicates (n = 5) were 9.9% (low), 6.4% (medium), and 5.3% (high) controls. The between-day precision was performed by analyzing these controls over 10 consecutive days. The subsequent between-run analyses of the coefficients of variation (CVs) of low, medium and high controls of both methods were similar: 10.9%, 6.2%, and 5.8%, at 3.9, 11.3, and 25.4 ng/ml, respectively, for FPIA. Table 1 shows the results of analyses of these commercial controls compared with values provided by the manufacturer for ERL-FPIA and 11.4%, 8.7%, 7.8%, and 4.2% at 2.47 ng/ml, 4.43 ng/ml, 8.82 ng/ml, and 30.1 ng/ml for HPLC/UV in addition. Our data from FPIA compared well with the manufacturer's data of 6.0% and 9.0% for within run, and 12.0%, 5.0% and 5.0% for between run. The lower limit of detection of FPIA, determined as the mean + 3 SD of 10 replicates of drug-free EDTA blood, was 0.8 ng/ml. The lower limit of quantification was designated as 2.2 ng/ml, the value of the lowest calibrator. The latter was Table 1 Results of analyses of commercial quality control samples compared with values provided by manufacturer for ERL-FPIA (n = 10) and for ERL-HPLC (n = 5) Nominal value, range a (ng/ml)
Analyte found b (ng/ml, mean ± S.D.)
CV (%)
ERL-FPIA Control low Control medium Control high
3.9 (2.73–4.96) 11.3 (9.09–13.50) 25.4 (19.46–31.25)
3.31 ± 0.36 11.81 ± 0.73 21.56 ± 1.25
10.9 6.2 5.8
ERL-HPLC Level 1 Level 2 Level 3 Level 4
2.47 (1.61–3.33) 4.43 (2.88–5.98) 8.82 (5.73–11.9) 30.1 (21.1–39.1)
2.72 ± 0.31 3.92 ± 0.34 8.07 ± 0.63 31.8 ± 1.32
11.4 8.7 7.8 4.2
Sample analyte
a b
Values provided by the manufacturers. Results obtained in our laboratory.
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
similar to that for HPLC (2.5 ng/ml) and is acceptable for routine clinical monitoring. For intraday precision, the samples were stored at −20 °C and measured on the same day; for interday precision, we used the mean values of quality controls measured on 6 different days. The intra- and interday imprecision values (CVs) for HPLC were 3.2%–10% and 2.0%–9.6%, respectively, for SRL and 4.0%–9.3% and 3.4%–8.3% for everolimus (28). The limit of detection for HPLC, defined as the concentration of drug giving a signal-to-noise ratio N 3:1 was 1.0 ng/ml. The FPIA results (y-axis) were compared with those from the HPLC/UV (x-axis) in 165 whole-blood samples from kidney (n = 112) and heart (n = 53) transplant recipients, by Passing– Bablock estimation. The correlation function was FPIA = 1.11 (1.039 − 1.200) × (HPLC-UV) + 0.38 (− 0.080 − 0.807), with a correlation coefficient of r2 = 0.8229 for different types of transplant (Fig. 1A). Passing–Bablok regression analysis revealed that FPIA measured ERL concentrations that were significantly higher than with HPLC. Our results are similar to
219
Table 2 The cross-reactivity of the Innofluor® Certican® assay with the structurally related compound sirolimus, as well as metabolites of everolimus Everolimus metabolites
Concentration tested (ng/ml)
Percent (%) cross-reactivity
RAD SA RAD PSA RAD PC 25-OH RAD Sirolimus
5 5 5 50 10
5 12 72 6 74
As shown in the manufacturer's reagent package insert.
literature data for FPIA vs. LC/MS-MS from renal transplant patients' samples (n = 333) as measured by both methods using Deming regression analysis; FPIA = 1.19 (±0.03) × HPLC-MS + 0.51 (±0.21) μg/l [31]. The statistical analysis according to Altman and Bland provides further information on differences that can be expected with a 95% probability when two methods are used for the determination of ERL concentration in one blood sample. FPIA compared to HPLC gave a positive bias of 1.19 ng/ml, with 95% limits of agreement of − 1.37 to 3.76 ng/ml (Fig. 1B). Thus, both statistical procedures proved the analysis results for the two methods to be comparable. Overall, there was a tendency for FPIA assay to give higher ERL values than HPLC. The mean ERL concentration of patient samples using HPLC-UV was 5.66 ± 2.79 ng/ml (range 1.3 to 18.1 ng/ml) compared with 6.86 ± 3.11 ng/ml (range 2.05 to 20.2 ng/ml) using the immunoassay. This highlights the fact that the differences between the two methods are probably due to differences in the calibrators and likely to cross-reactivity between ERL metabolites, SRL and the antibody in the FPIA method (e.g. ERL-PC; 72%, and with sirolimus 74%). Table 2 shows the manufacturer's data for some of the ERL metabolites and SRL with their chosen antibody. It shows that the monoclonal antibody-based ERL immunoassay can demonstrate variable, and sometimes significant, cross-reactivity with various ERL metabolites. Strom et al. [32] report that the major hydroxy everolimus metabolites in the blood of patients pose insignificant cross-reactivity with the antibody used in the Innofluor Certican immunoassay (4–6% cross-reactivity with the 25-OH-RAD). With increasing numbers of transplant patients and new protocols for immunosuppressant treatment, patients may be switched from SRL to ERL because of intolerance. The degree Table 3 Results for patients switched from sirolimus to everolimus (HPLC and FPIA)
Fig. 1. Comparison of results of everolimus trough concentrations in 165 blood samples from transplant patients determined by HPLC and FPIA assay. (A) Passing–Bablok regression analysis between both methods shows the correlation function (solid line): FPIA = 1.111 × HPLC + 0.378. The solid and the dashed lines indicate, respectively, the regression line and unity. (B) The degree of agreement between methods was assessed using the Bland–Altman graphical technique. The mean difference between the methods was 1.19 ± 2.57.The solid line indicates the mean difference between the methods, and the 95% confidence intervals for the difference are indicated by dashed lines.
Patient 1 Patient 1 a Patient 2 Patient 3 Patient 4 Patient 5 Patient 6 a
ERL(FPIA)
ERL(HPLC)
SRL(HPLC)
ng/ml
ng/ml
ng/ml
10.1 7.7 10.6 7.4 4.6 4.1 10.7
4.7 4.8 7.6 4.0 2.0 2.2 5.1
5.1 1.2 2.8 2.8 1.9 1.6 5.4
Patient level after 5 days.
220
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221
of cross-reactivity of the sirolimus metabolites with the FPIA assay could not be determined due to the lack of those substances in their pure forms. There is only the report of Jacobsen et al. [18] that in vitro metabolism of structurally similar ERL and SRL shows significant differences. For some laboratories, the relative technical simplicity and short run time of the immunoassay could make it an attractive alternative to HPLC for the measurement of ERL. The assay does not, however, measure the extent of accumulation of metabolites and their cross-reactivity with other immunosuppressants that may be present (e.g. SRL). This is often the case as ever more transplant patients are switched from sirolimus to everolimus, and requires more precise methods for determining ERL concentrations. For six patients who were switched from SRL to ERL, the immunological analysis produced a value that was more than 100% higher than with HPLC (e.g., a heart-transplant patient had an ERL value with the FPIA assay of 10.1 ng/ml and of 4.7 ng/ml with HPLC, with a measured SRL value of ng/ml). Table 3 shows the individual ERL and SRL results of 7 samples from 6 switched transplant patients with percentage differences. Detectable SRL values from patients whose medication had been changed were found for one week after the change. This means that daily monitoring of ERL from these patients with FPIA assay shows cross-reactivity with sirolimus, because of the long half-life of SRL. The percentage difference between the immunoassay and HPLC/UV shows that a monoclonal antibody-based assay is affected by drug metabolites. These studies show that the performance of some of the most used immunoassays is not satisfactory, and can lead to incorrect therapeutic decisions guided by erroneous immunosuppressive monitoring. The HPLC method is considered better suited for TDM, and so for application in clinical strategies to improve transplantation tolerance as well. In conclusion, the clinical interpretation of ERL concentrations determined with the FPIA assay is complicated by the ERL metabolite profile and co-medications that may be present. The main methodological problems with the measurement of immunosuppressive drug concentrations in the serum are generally due to the cross-reactivity of metabolites of the parent drug with antibodies used in immunoassays [33]. The effectiveness of ERL therapy may greatly depend on the method of determination, because the cross-reactivity is usually considered to be the major disadvantage of the FPIA immunoassay. Acknowledgements We thank Ms. Eugenia Lamont for editorial assistance and Ms. Andrea Prattes for technical assistance. References [1] Vitko S, Margreiter R, Weimar W, et al. Everolimus (certican) 12-month safety and efficacy versus mycophenolate mofetil in de novo renal transplant recipients. Transplantation 2004;78:1532–40. [2] Lorber MI, Mulgaonkar S, Butt KM, et al. Everolimus versus mycophenolate mofetil in the prevention of rejection in de novo renal transplant recipients: a 3-year randomized, multicenter, phase III study. Transplantation 2005;80:244–52.
[3] Schuurman HJ, Cottens S, Fuchs S, et al. SDZ RAD, a new rapamycin derivative: synergism with cyclosporine. Transplantation 1997;64:36–42. [4] Schuler W, Sedrani R, Cottens S, et al. SDZ RAD, a new rapamycin derivative: pharmacological properties in vitro and in vivo. Transplantation 1997;64:32–5. [5] Dumont FJ. Everolimus. Curr Opin Investig Drugs 2001;2:1220–34. [6] Kirchner GI, Meier-Wiedenbach I, Manns MP. Clinical pharmacokinetics of everolimus. Clin Pharmacokinet 2004;43:83–95. [7] Neumayer HH, Paradis K, Korn A, et al. Entry-into-human study with the novel immunosuppressant SDZ RAD in stable renal transplant recipients. Br J Clin Pharmacol 1999;48:694–703. [8] Kahan BD, Wong RL, Carter C, et al. A phase I study of a 4-week course of SDZ-RAD(RAD) in quiescent cyclosporine-prednisonetreated renal transplant recipients. Transplantation 1999;68:1100–6. [9] MacDonald A, Scarola J, Burke JT, Zimmerman JJ. Clinical pharmacokinetics and therapeutic drug monitoring of sirolimus. Clin Ther 2000;22: B101–21 [Suppl.]. [10] Gallant-Haidner HL, Trepanier DJ, Freitag DG, Yatscoff RW. Pharmacokinetics and metabolism of sirolimus. Ther Drug Monit 2000;22:31–5. [11] Sedrani R, Cottens S, Kallen J, Schuler W. Chemical modification of rapamycin: the discovery of SDZ RAD. Transplant Proc 1998;30:2192–4. [12] Kovarik JM, Kahan BD, Kaplan B, et al. Longitudinal assessment of everolimus in de novo renal transplant recipients over the first posttransplant year: pharmacokinetics, exposure-response relationships, and influence on cyclosporine. Clin Pharmacol Ther 2001;69:48–56. [13] Mahalati K, Kahan BD. Clinical pharmacokinetics of sirolimus. Clin Pharmacokinet 2001;40:573–85. [14] Kovarik JM, Eisen H, Dorent R, et al. Everolimus in de novo cardiac transplantation: pharmacokinetics, therapeutic range, and influence on cyclosporine exposure. J Heart Lung Transplant 2003;22:1117–25. [15] Kahan BD, Keown P, Levy GA, Johnston A. Therapeutic drug monitoring of immunosuppressant in clinical practice. Clin Ther 2002;24:330–50. [16] Schuurman HJ, Ringers J, Schuler W, Slingerland W, Jonker M. Oral efficacy of the macrolide immunosuppressant SDZ RAD and of cyclosporine microemulsion in cynomolgus monkey kidney allotransplantation. Transplantation 2000;69:737–42. [17] Serkova N, Hausen B, Berry GJ, et al. Tissue distribution and clinical monitoring of the novel macrolide immunosuppressant SDZ-RAD and its metabolites in monkey lung transplant recipients: interaction with cyclosporine. J Pharmacol Exp Ther 2000;294:323–32. [18] Jacobsen W, Serkova N, Hausen B, Morris RE, Benet LZ, Christians U. Comparison of the in vitro metabolism of the macrolide immunosuppressants sirolimus and RAD. Transplant Proc 2001;33:514–5. [19] Börnsen KO, Inverardi B, Gschwind HP. Electrospray ionization and collisional induced dissociation of RAD001 and related compounds and structural characterization of RAD001 metabolites by nano-spray and micro liquid chromatography mass spectrometry. Novartis Inter Rep 2001: R98–R709 [DMPK(CH)]. [20] Kovarik JM. Everolimus: a proliferation signal inhibitor targeting primary causes of allograft dysfunction. Drugs Today (Barc) 2004;40:101–9. [21] Vidal C, Kirchner GI, Sewing KF. Structural elucidation by electrospray mass spectrometry: an approach to the in vitro metabolism of the macrolide immunosuppressant SDZ RAD. J Am Soc Mass Spectrom 1998;9: 1267–74. [22] Kirchner GI, Winkler M, Mueller L, et al. Pharmacokinetics of SDZ RAD and cyclosporine including their metabolites in seven kidney graft patients after the first dose of SDZ RAD. Br J Clin Pharmacol 2000;50:449–54. [23] Kirchner GI, Vidal C, Winkler M, et al. LC/ESI-MS allows simultaneous and specific quantification of SDZ RAD and cyclosporine, including groups of their metabolites in human blood. Ther Drug Monit 1999;21:116–22. [24] Koal T, Deters M, Casetta B, Kaever V. Simultaneous determination of four immunosuppressants by means of high speed and robust on-line solid phase extraction–high performance liquid chromatography–tandem mass spectrometry. J Chromatogr, B, Biomed Sci Appl 2004;805:215–22. [25] Ceglarek U, Lembcke J, Fiedler GM, et al. Rapid simultaneous quantification of immunosuppressants in transplant patients by turbulent flow chromatography combined with tandem mass spectrometry. Clin Chim Acta 2004;346:181–90.
G. Khoschsorur et al. / Clinica Chimica Acta 380 (2007) 217–221 [26] Streit F, Amstrong VW, Oellerich M. Rapid liquid chromatographytandem mass spectrometry routine method for simultaneous determination of sirolimus, everolimus, tacrolimus, and cyclosporin A. Clin Chem 2002;48:955–8. [27] Khoschsorur G. Simultaneous measurement of sirolimus and everolimus in whole blood by high-performance liquid chromatography with ultraviolet detection. Clin Chem 2005;9:1721–4. [28] Baldelli S, Zenoni S, Merlini S, Perico N, Cattaneo Dario. Simultaneous determination of everolimus and cyclosporine concentrations by HPLC with ultraviolet detection. Clin Chim Acta 2006;364:354–8. [29] Passing H, Bablok W. A new biometrical procedure for testing the equality of measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, part I. J Clin Chem Clin Biochem 1983;21:709–20.
221
[30] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [31] Salm P, Warnholtza C, Boydb J, Arabshahi L, Marbach P, Taylor PJ. Evaluation of a fluorescent polarization immunoassay for whole blood everolimus determination using samples from renal transplant recipients. Clin Biochem 2006;39:732–8. [32] Strom T, Boyd J, Arabshahi L, Marbachc P, Christiansa U. Cross-reactivity of the major everolimus hydroxy metabolites with the innofluor certican immunoassay for clinical Tdm of everolimus. Ther Drug Monit 2005;27:243 [Abstract]. [33] Holt DW, Johnston A. Monitoring new immunosuppressive agent. Are the methods adequate? Drug Metab Drug Interac 1997;14:5–15.