| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Drug Monitoring and Toxicology |
1
Abteilung Klinische Chemie, Georg-August-Universität Göttingen, D-37075 Göttingen, Germany.
2
Department of Clinical Laboratory, Medical University of
Sofia, 1341, Bulgaria.
a Address correspondence to this author at: Abteilung Klinische Chemie, Zentrum Innere Medizin, Georg-August-Universität, Robert Koch Strasse 40, D-37075 Göttingen, Germany. Fax 49-551-398551; e-mail eschuetz{at}med.uni-goettingen.de.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
10% at 50 µg/L and 5% at 300 µg/L; and
(b) comparison with the reference method (HPLC) should
yield a slope of 0.9–1.1, an intercept of -15 to 15 µg/L, and
Sy
x 15 µg/L. The newly developed CsA assays
for the AxSYM (Abbott) and the CEDIATM (Boehringer
Mannheim) as well as the EmitTM assay (Behring Diagnostica)
were evaluated. Results from samples of heart, kidney, and liver
recipients (100 specimens each) were compared with a validated
HPLC-ultraviolet detection method. Between-series imprecision (CV) with
commercial controls was 5.8% and 1.7% for AxSYM (70 and 300 µg/L),
11% and 5.5% for CEDIA (90 and 200 µg/L), and 8.1% and 4.5% for
Emit (63 and 172 µg/L). In the presence of 300 µg/L parent CsA,
cross-reactivities were (for AxSYM, CEDIA, and Emit, respectively) 7%,
4%, and none for AM1 (1 mg/L) and 12.6%, 25%, and 6% for AM9 (0.5
mg/L). Comparison with HPLC showed in heart and kidney recipients an
average overestimation with the Emit and the CEDIA of ~22%, with
overestimation in the AxSYM of 32%. In liver recipients, the most
challenging patient group, the CEDIA and the AxSYM showed a mean
overestimation of 43% and 47%, respectively, and the Emit differed by
31% compared with HPLC. None of the immunoassays fully satisfied the
performance criteria recommended in the consensus documents. In terms
of specificity, Emit ranks before CEDIA, which ranks before AxSYM.
Regarding imprecision, the ranking is AxSYM < Emit < CEDIA.
These limitations must be considered when using these assays for
therapeutic drug monitoring of CsA in clinical transplantation. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The CYP3A-dependent metabolism of CsA produces a variety of known
metabolites. Whether these metabolites are still toxic or whether they
possess substantial immunosuppressive efficacy is still controversial
(9)(10). On the other hand, these metabolites
are of substantial importance with regard to TDM of CsA because all
immunoassays currently used for whole blood CsA determination suffer
from cross-reactivity with these metabolites to varying degrees
(11)(12). Several consensus meetings have
established recommendations for TDM of CsA
(13)(14) during the last decade. At the most
recent meeting, the Lake Louise Consensus Conference on Cyclosporin
Monitoring, an overview from 35 transplant centers around the world
showed that 26% used a specific HPLC method for CsA determination,
whereas most other centers used relatively specific methods for CsA
measurement based on radioimmunoassays (14%) or nonisotopic
immunoassays (57%) (14). Because the reproducibility and
analytical specificity for the parent drug is critical for an
acceptable CsA immunoassay, the following performance criteria have
been recommended in the consensus documents
(13)(14): (a) imprecision 10% at
50 µg/L and 5% at 300 µg/L; and (b) for accuracy,
comparison with the reference method (a validated HPLC procedure)
should yield a slope of 0.9–1.1, an intercept of -15 to 15 µg/L,
and Syx 15 µg/L, calculated with bivariate,
preferably nonparametric procedures (e.g., Passing–Bablok method)
(15). Two newly specific CsA whole blood assays have become
available in 1997 and 1998. The monoclonal antibody-based fluorescence
polarization immunoassay (mFPIA) CsA assay has been adapted for the
AxSym immunoanalyzer (Abbott). A homogeneous enzyme immunoassay for CsA
quantification in whole blood, which does not require sample
extraction, has been developed using CEDIATM technology
(Boehringer Mannheim). The latter is not yet FDA-approved. We have
evaluated these assays in terms of analytical performance
characteristics and compared them to a validated HPLC method with
ultraviolet detection (16). In addition, the
EmitTM Assay (Dade Beh-ring Diagnostica), which has been
FDA-approved since 1997, was also included in the investigation.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
cedia
CEDIA is based on the spontaneous association of a short
recombinant NH2-terminal ß-galactosidase
fragment (
-peptide) and a recombinant ß-galactosidase monomer with
a deletion near the NH2 terminus to form active enzyme
tetramer. CsA is chemically attached to the -peptide, and an
anti-CsA antibody binds to both CsA in the sample and CsA coupled to
the -peptide. Because the latter process interferes with the
formation of active enzyme, the amount of CsA in the sample is directly
proportional to the residual enzyme activity. The pretreatment step for
the CEDIA is different to that of the other immunoassays, because a
separation step is not necessary. After lysis of the whole blood, CEDIA
(Boehringer) was carried out on a Hitachi 917 automated analyzer in
batch analyses, without concomitant measurements of other analytes. The
calibration was based on two calibration points (in duplicate), and the
system was recalibrated whenever the deviation of the controls was
>10% of the assigned value or whenever reagent bottles had to be
changed as recommended by the manufacturer. Mean calibration frequency
was once a week during the study period.
emit
The Emit procedure (Dade Behring) was preceded by a methanol
extraction according to the manufacturer's protocol. The closed
secondary cup was directly placed on the rack of one of two dedicated
Cobas Mira Plus analyzers (Roche). The calibration was based on five
calibration points (in duplicate) and was done whenever the
deviation of the controls was >10% of the assigned value. The
mean frequency of calibration is 2 weeks in our laboratory.
hplc
The HPLC procedure was carried out as described elsewhere
(16). Briefly, 1 mL of whole blood, calibrators, or controls
was mixed with internal standard solution. After hemolysis,
liquid–liquid extraction, and solid phase extraction, HPLC analyses
were carried out with a Supelcosil LC1 (50 x 4.5 mm, 5 µm,
Supelco) column, heated to 72 °C. Calibration was done within each
analysis, using in-house calibrators with precisely weighed
concentrations of CsA and cyclosporin D as internal standard. Peak
height from ultraviolet detection at 214 nm was used for the
construction of the calibration curve. Measurements were performed on a
Gynkotek HPLC system with a diode array detector (Gynkotek). HPLC
measurements were accepted if the ultraviolet spectra of the CsA and
cyclosporin D peaks showed an identity with standard CsA and
cyclosporin D spectra of >98%. The detection limit, defined as
signal-to-noise ratio of 5:1, was 25 µg/L. The within-run imprecision
(CV) of this HPLC method was <5.8%, whereas between-day imprecision
was <7.8% at CsA concentrations between 100 and 500 µg/L.
materials
The manufacturers' controls, pooled sample material, and
independent commercially available control materials (Lyphocheck Whole
Blood Control, Bio-Rad) were used for the evaluation of precision, both
within-run and between-day, and for reproducibility studies. Purified
metabolites (AM1, AM9, and AM4N; Novartis GmbH) were used for
estimation of cross-reactivity. Influences of the different
pretreatment procedures were excluded by using drug-free whole blood
specimens with CsA added as well as using the metabolites in methanol
solutions.
For recovery experiments, a CsA standard preparation (US Pharmacopeia) was added to a human whole blood pool to give concentrations of 50, 100, and 400 µg/L.
method comparison
For method comparisons of CsA, we used 100 samples each from
kidney, heart, and liver recipients. If the deviation of the CsA
concentrations determined with the AxSYM, CEDIA, and Emit was >15%,
samples were rerun on each systems to exclude random errors. Only
reproducible results were accepted for statistical calculations.
Values of CsA trough concentrations above the dilution limit of an
assay were excluded from calculations.
statistics
The detection limits of the immunoassays were defined and
estimated by the method according to Kaiser (17) as the
concentration that is above zero, with an error probability <2.5%
(The concentration that equals the mean plus 3 standard deviations of
the signal produced by analyte-free samples).
The nonparametric regression procedure developed by Passing and Bablok
(15) was used for the estimation of the structural
relationship between the immunoassays and the HPLC method in clinical
samples (EVAPAK, Ver. 2.08, Boehringer). The 95% confidence interval
for the estimates of slope and intercept is given, together with the
95% median distance of the residuals. In addition, the standard error
of the residuals (Syx) of the standard principal
component model is also presented for comparison with the
recommendations of the consensus documents
(13)(14).
The scatter of method comparison data was visualized according to the recommendations of Bland and Altman (18). The differences between the methods are plotted against the mean of the methods.
Analytical sensitivity was assessed using the method of critical
difference (19). The critical difference is defined as
2 x 
2 x SD (as assessed by imprecision profiles) for CsA
measurement with the particular method and represents the minimal
difference between two measurements that a method is able to
discriminate.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
imprecision and accuracy
Imprecision within series and analytical specificity.
Samples
of pooled whole blood were measured in a series of 20 within one
analytical run (Table 1
). A low CsA concentration was added to one pool; an identical
higher (290 µg/L) CsA concentration was added to the other three
(pools 2–4). Two of these latter whole blood pools were also
supplemented with AM1 (pool 3, 1000 µg/L) and AM9 (pool 4, 500
µg/L).
|
The calculated average cross-reactivity from the data in Table 2
for the AxSym was 8.6% for AM1, 14.5% for AM9, and not
detectable for AM4N. The CEDIA showed a cross-reactivity of 6.3% for
AM1, 27% for AM9, and 5% for AM4N. The Emit assay had no detectable
cross-reactivity with AM1 and AM4N; however, it had ~8%
cross-reactivity with AM9.
|
Imprecision between runs.
Imprecision data are given in Table 2
. The CsA procedures on the AxSYM as well as the Emit assay fulfill
the recommended consensus criteria of imprecision
(12)(13). In contrast, the imprecision of the
CEDIA determined in a control sample with ~90 µg/L CsA exceeded the
limit of 10% that was recommended for concentrations at the low end of
the therapeutic range.
Cross-reactivity with metabolites.
Cross-reactivity towards
the metabolites AM1, AM9, and AM4N is shown in Table 3
. Purified metabolites were added to drug-free whole blood
without CsA in the samples. All assays displayed a cross-reactivity
with AM9; the highest cross-reactivity was seen with the CEDIA and the
lowest with the Emit. In the case of AM1, the AxSYM assay showed the
highest cross-reactivity, followed by CEDIA, whereas the Emit did not
cross-react with AM1 at concentrations up to 1 mg/L. Only the CEDIA
procedure was found to have a detectable cross-reactivity towards AM4N.
The Emit assay showed the least cross-reactivity with all metabolites
tested. The use of the new nonvolatile pretreatment reagent for the
Emit (Cyclosporine Sample Pretreatment Reagent, Dade Behring),
which will serve as a substitute for methanol beginning in 1998
(anticipated launch date for United States is November 1998, Food and
Drug Administration filing in progress), did not alter the
cross-reactivity toward the metabolites (data not shown). The
cross-reactivity to AM1 and AM9 derived from CsA-free whole blood pools
with metabolites added (Table 3
) are in close agreement to those values
in the presence of ~290 µg/L CsA (pools 3 and 4, Table 1
), The
differences in the calculated cross-reactivities with and without
CsA present did not exceed the limits of the analytical sensitivity.
|
Dilution linearity.
Dilution linearity of the immunoassays was
investigated with the appropriate dilution protocol of each assay,
using either the highest calibrator or a patient sample. For dilutions
up to 1:8, the linearity was acceptable with all of the immunoassays
(deviation from target value <10%).
Recovery in human blood.
To assess the recovery of the CsA
whole blood assays, a human whole blood pool was prepared and
supplemented with a precisely weighed CsA standard preparation (US
Pharmacopeia). The HPLC method used for comparison in patients'
samples displayed the best recovery (mean, 100.2%), followed by the
Emit assay (mean, 98.0%), which was independent of the target CsA
concentration. In contrast, the percentage of recovery of both AxSYM
(mean, 96.6%) and particularly CEDIA (mean, 93.7%) showed a lower
recovery at the lower CsA concentrations. The details of the recovery
experiments are given in Table 4
.
|
comparison of methods
Kidney recipients.
The comparability of the CsA whole blood
concentration measured with the immunoassays vs results from HPLC in
samples from kidney recipients are visualized in Fig. 1
, using the Bland–Altman difference plot.
|
The structural relationship (estimates from the Passing–Bablok model
with 95% confidence intervals in parentheses) between the HPLC method
and the immunoassays as well the median distances of residuals (MD95),
Syx, and mean differences (d%) are given in
Table 5
. It is obvious that Emit and CEDIA show a very similar
overestimation of ~20%; however, the Emit has a much narrower
dispersion of residuals. Of the immunoassays evaluated, AxSYM displayed
the highest overestimation compared with HPLC.
|
Heart recipients.
The method comparison results with samples
derived from heart recipients are very similar to those from kidney
recipients. The scatters of the differences between HPLC and the
immunoassays are shown in Fig. 2
, and the statistical data are given in Table 5
.
|
Similar to the results for kidney recipients, all methods overestimate CsA in whole blood from heart recipients. The results from the CEDIA are comparable with those from the Emit; however, the latter displays the narrower dispersion of residuals. Again the AxSYM assay showed the highest discrepancy to HPLC.
Liver recipients.
The most challenging samples for CsA
assays are those from liver recipients because these samples are the
most likely to contain high concentrations of CsA metabolites. The
results are presented in Fig. 3
and Table 5
. It is obvious that the difference from HPLC is
indeed the highest of all patients groups. Of the three immunoassays,
the Emit procedure displayed the lowest overestimation compared with
HPLC and the smallest Syx (dispersion of
residuals). Comparable values were seen for the AxSYM and the CEDIA,
which both showed asubstantially higher overestimation compared with
HPLC than the Emit.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In view of our knowledge that in certain circumstances (e.g., cholestasis in liver recipients) large amounts of CsA metabolites may accumulate (20), analytical specificity towards the parent drug is very critical for effective CsA monitoring with immunoassays. Two immunoassays are thought to fulfill the specificity criteria recommended in the consensus documents (13)(14), the 1 H-radioimmunoassay originally produced by Sandoz and the Emit that is commercially available from Dade Behring (11). According to the data from the present evaluation in a larger number of samples, this interpretation should be regarded with caution. None of the commercial nonisotopic immunoassays available today totally fulfills the consensus recommendations for accuracy and specificity of CsA measurement.
Dilution linearity seems to be a less critical issue for CsA monitoring, because the measuring range of all assays spans the therapeutic ranges for CsA trough concentrations (AxSYM, 13–800 µg/L; CEDIA, 25–620 µg/L; Emit, 17–500 µg/L). On the other hand, it has recently been discussed that with the new microemulsion formulation of CsA (Neoral) that a reduction in acute rejection rate after kidney transplantation may be achieved if CsA concentrations measured 2 h after drug intake are maintained between 800 and 1200 µg/L (8). Because such 2-h samples will usually exceed the upper limit of the assays, dilution will be unavoidable. The linearity is acceptable with all immunoassays for dilutions up to 1:8.
Cross-reactivity toward CsA metabolites AM1 and AM4N remains higher with the new AxSYM assay than with the CEDIA, whereas the latter assay shows the highest cross-reactivity towards AM9. The low recovery for the US Pharmacopeia cyclosporin standard in human whole blood in the CEDIA assay at lower CsA concentrations may partially mask this lack of specificity. The lowest cross-reactivities and the best recoveries were found with the Emit. This is especially obvious in samples from liver recipients, where a 47% mean deviation from the HPLC method was observed with the AxSym and 43% with the CEDIA, as opposed to 31% deviation with the Emit. In kidney and heart recipients, both the Emit and the CEDIA displayed a similar mean deviation from HPLC of ~22%, whereas the results from the AxSym were on average ~34% higher than HPLC results in this patients.
The new AxSYM CsA assay has an advantage in that this immunoanalyzer is in widespread use, which makes the test readily available. In addition, it is easy to perform and is highly reproducible. This assay is, therefore, an attractive alternative for laboratories already using the established mFPIA running on the TDx/FLx system.
The CEDIA assay has a very easy sample preparation that requires less technical time and can be performed, in principle, on several clinical chemistry analyzer models. The comparatively poor precision of the test even in batch analyses seems to be a critical issue, and it can be speculated that running the test in a random access mode may even lead to deterioration. The Emit test has a sample preparation that requires separation but uses only one reagent, which is available in each laboratory. If the new nonvolatile pretreatment reagent is used, this test can also be adapted to clinical chemistry analyzers and so would require a technical time between that needed for the other immunoassays. Concerning the performance criteria, the Emit has the best specificity towards the parent drug, and it shows the lowest overestimation of the investigated immunoassays compared with the HPLC method. The between-day imprecision also fulfills the consensus criteria (13)(14).
Taken together, the data clearly show that, currently, the performance of nonisotopic immunoassays for CsA monitoring is still not satisfactory. In particular, analytical specificity has to be substantially improved. In liver recipients with unexpectedly high trough concentrations because of the accumulation of CsA metabolites, a validated HPLC assay is still the method of choice and should be available in centers dealing with such samples. From an economic point of view, the additional costs for confirmatory HPLC analyses in such cases must be calculated against the follow-up costs arising from incorrect therapeutic decisions guided by erroneous CsA monitoring. In the majority of the other clinical situations, the use of an immunoassay with the performance characteristics of the Emit seems to be a practical compromise.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  R. G Morris Immunosuppressant Drug Monitoring: Is the Laboratory Meeting Clinical Expectations? Ann. Pharmacother., January 1, 2005; 39(1): 119 - 127. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. M. Juenke, P. I. Brown, F. M. Urry, and G. A. McMillin Specimen Dilution for C2 Monitoring with the Abbott TDxFLx Cyclosporine Monoclonal Whole Blood Assay Clin. Chem., August 1, 2004; 50(8): 1430 - 1433. [Full Text] [PDF]
|
|
|
|
|
|
A. R. Terrell, T. M. Daly, K. G. Hock, D. C. Kilgore, T. Q. Wei, S. Hernandez, D. Weibe, L. Fields, L. M. Shaw, and M. G. Scott Evaluation of a No-Pretreatment Cyclosporin A Assay on the Dade Behring Dimension RxL Clinical Chemistry Analyzer Clin. Chem., July 1, 2002; 48(7): 1059 - 1065. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. Streit, V. W. Armstrong, and M. Oellerich Rapid Liquid Chromatography-Tandem Mass Spectrometry Routine Method for Simultaneous Determination of Sirolimus, Everolimus, Tacrolimus, and Cyclosporin A in Whole Blood Clin. Chem., June 1, 2002; 48(6): 955 - 958. [Full Text] [PDF]
|
|
|
|
|
|
B. G. Keevil, D. P. Tierney, D. P. Cooper, and M. R. Morris Rapid Liquid Chromatography-Tandem Mass Spectrometry Method for Routine Analysis of Cyclosporin A Over an Extended Concentration Range Clin. Chem., January 1, 2002; 48(1): 69 - 76. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. von Ahsen, M. Richter, C. Grupp, B. Ringe, M. Oellerich, and V. W. Armstrong No Influence of the MDR-1 C3435T Polymorphism or a CYP3A4 Promoter Polymorphism (CYP3A4-V Allele) on Dose-adjusted Cyclosporin A Trough Concentrations or Rejection Incidence in Stable Renal Transplant Recipients Clin. Chem., June 1, 2001; 47(6): 1048 - 1052. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. W. Holt, A. Johnston, B. D. Kahan, R. G. Morris, M. Oellerich, and L. M. Shaw New Approaches to Cyclosporine Monitoring Raise Further Concerns about Analytical Techniques Clin. Chem., June 1, 2000; 46(6): 872 - 874. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
CsA (µg/L) from the HPLC method
is plotted against the mean CsA value for HPLC and the indicated
method. The mean difference is displayed by a solid line,
the SD of differences by a dotted line, and the line of
identity by a dashed line. (A) AxSYM,
(B) CEDIA, (C) Emit.
