Clinical Chemistry 45: 1449-1458, 1999;
(Clinical Chemistry. 1999;45:1449-1458.)
© 1999 American Association for Clinical Chemistry, Inc.
Analytical Validation of the PRO-Trac II ELISA for the Determination of Tacrolimus (FK506) in Whole Blood
Gordon D. MacFarlane1,a,
Daniel G. Scheller1,
Diana L. Ersfeld1,
Leslie M. Shaw2,
Raman Venkatarmanan3,
Laszlo Sarkozi4,
Richard Mullins5 and
Bonnie R. Fox1
1
DiaSorin Inc., Stillwater, MN 55082.
2
University of Pennsylvania Medical Center, Philadelphia,
PA 19104.
3
University of Pittsburgh Medical Center, Pittsburgh, PA
15261.
4
Mount Sinai Medical Center, New York, NY 10001.
5
Department of Clinical Chemistry, Emory
University, Atlanta, GA 30322.
a Address correspondence to this author at: DiaSorin Inc., PO Box 285, 1990 Industrial Blvd., Stillwater, MN 55082. Fax 651-351-5669; e-mail gordon.macfarlane{at}diasorin.com
 |
Abstract
|
|---|
Background: The analytical validation of multiple lots of the
PRO-TracTM II ELISA (DiaSorin) for the determination of
tacrolimus in whole blood is described.
Methods: The analytical parameters assessed included
analytical sensitivity, dilution linearity, functional sensitivity,
values in samples containing no tacrolimus, intra- and interassay
precision, supplementation and recovery, metabolite
cross-reactivity, interference studies, and method comparisons
HPLC-tandem mass spectrometry (HPLC/MS/MS) and the IMx®
Tacrolimus II multiparticle enzyme immunoassay. Where appropriate,
assessments were performed according to NCCLS guidelines.
Results: The mean analytical detection limit was <0.25
µg/L for all lots, whereas the functional sensitivity was 1.0 µg/L.
Excellent linear correlation (r = 0.985) was
observed for dilution linearity. The intraassay imprecision was
<7%, and the total imprecision by ANOVA was <10%. Recovery was
109% ± 11%. Metabolite cross-reactivity was consistent with previous
reports for this antibody. No interference was observed for 35 tested
drugs. Method comparison with HPLC/MS/MS showed no statistically
significant differences. Samples exhibited stability through four
freeze/thaw cycles and for 1 week at room temperature.
Conclusion: These data demonstrate that the PRO-Trac II
ELISA is a robust, accurate, and precise tool for the assessment and
management of tacrolimus blood concentrations in transplant
patients.
|
Introduction
|
|---|
Tacrolimus is a macrolide antibiotic of fungal origin with potent
immunosuppressive properties (1). Because of its narrow
therapeutic index, large inter- and intrapatient variability in
pharmacokinetics, and poor correlation between dose and trough blood
concentrations, tacrolimus concentrations in the blood must be
monitored regularly (2)(3)(4). Historically, methods to monitor
tacrolimus in the blood have been labor-intensive and lacked sufficient
analytical sensitivity.
A manual research ELISA for the determination of tacrolimus in whole
blood was developed by Fujisawa Pharmaceutical for use during drug
development (5). A manual ELISA based on this research ELISA
with ready-to-use reagents and a methanol extraction was introduced by
DiaSorin (INCSTAR) in 1994. A second-generation ELISA was developed in
1996 that eliminated the organic extraction techniques and shortened
the overall procedure to <4 h (6). The details of this
procedure have been reported previously (6). Here we report
the analytical validation of multiple production lots of this assay.
|
Materials and Methods
|
|---|
tacrolimus assays
The PRO-TracTM II FK506 ELISA assay
(DiaSorin Inc., Stillwater, MN) was performed according to the
procedures detailed in the product insert. This procedure has been
reported in detail previously (6). The only modification of
the reported procedure utilized in the current study was an increase in
the temperature (from 70 °C to 75 °C) required to halt
proteolysis. Unless otherwise noted, all performance studies were
performed on representative kits from three independent production kit
lots. The IMx® Tacrolimus II MEIA assay (Abbott
Laboratories) was performed according to the procedures detailed
in the product insert. This procedure has been reported previously
(7). HPLC-tandem mass spectrometry (HPLC/MS/MS) analysis of
clinical tacrolimus samples was performed by an independent reference
laboratory using a previously published method (8) with
a detection limit of 0.1 µg/L.
samples
Supplemented samples.
Powdered tacrolimus (>98%
purity; Fujisawa Pharmaceutical), was dissolved in methanol at a high
concentration to obtain a stock tacrolimus concentrate. Supplemented
samples were prepared in EDTA whole blood by the addition of stock
tacrolimus concentrate such that the final methanol concentration was
always <10 mL/L.
Clinical samples.
Clinical samples were obtained from a
multisite clinical trial to be described in detail elsewhere. In this
study, 111 liver transplant patients from six clinical sites were
followed prospectively for 90 days posttransplantation. The clinical
sites involved in this study were the Mount Sinai Medical Center, New
York, NY; the University of Miami School of Medicine, Miami, FL; the
University of Pittsburgh School of Medicine, Pittsburgh, PA; the
University of Pennsylvania Medical Center, Philadelphia, PA; the
University of Wisconsin Hospitals and Clinics, Madison, WI; and
Emory University Hospital, Atlanta, GA. All patients provided informed
consent, and all procedures and protocols were approved by the
institutional review boards of the respective clinical sites. Whole
blood tacrolimus concentrations in patients were assessed 15 times
during this interval with both the PRO-Trac II ELISA and the IMx
Tac I microparticle enzyme immunoassay (MEIA). Blood samples
were collected three times per week during weeks 1 and 2, twice per
week during weeks 3 and 4, once per week during weeks 5 and 6, and one
sample every 2 weeks during weeks 7 through 12. These samples were
aliquoted and frozen at the clinical site.
Fifty subjects were randomly chosen from four clinical sites for
HPLC/MS/MS analysis. Three samples from each subject, for a total of
150 samples, were chosen to represent early, middle, and late time
points in the study. Samples were chosen without consideration of the
clinical laboratory results. A subset containing 31 samples from 22
patients with serum bilirubin concentrations 
30.0 mg/L was analyzed
separately. Method comparison studies with the IMx Tacrolimus II MEIA
were completed on 95 of the original 150 samples that contained
sufficient volume for reanalysis.
Nondosed normal samples.
Samples from healthy volunteers not
receiving tacrolimus were assessed as part of the interference studies.
The study was performed on EDTA whole blood from 25 healthy,
nonmedicated volunteers, assessed as unknowns with one kit lot of
materials. Twelve males and 13 nonpregnant females were included in
this study.
analytical methods
Detection limit.
The minimum detectable concentration was
assessed as the minimum analyte concentration that could be
discriminated from zero according to a previously reported procedure
(6). The minimum detectable concentration was determined
with fresh enzyme conjugate and conjugate near its labeled expiration.
Briefly, 10 replicates of the zero and A calibrators were assessed in a
single assay. The apparent concentration based on the absorbance at 2
SD below the mean absorbance of the zero calibrator was used to
determine the minimum detectable concentration according to a standard
equation (6)(9). The minimum detection limit was
determined on three independent production lots.
Dilution linearity.
Dilution linearity was assessed by
serially diluting nine supplemented whole blood samples (initial
concentrations 60, 50, 40, 30, 25, and 20 µg/L) and 20 clinical
samples with the zero calibrator as described in the package insert and
assaying the diluted samples as unknowns. All samples were diluted to
at least 1:16. The results were plotted as a least-squares linear
regression of the expected concentration vs the observed concentration.
The undiluted clinical sample values, as determined in the assay, were
used to establish the expected values for subsequent dilutions but were
not used in the construction of the regression plot. Dilution linearity
was determined for clinical and supplemented samples separately and
combined and for each lot of material.
Functional sensitivity.
Because the minimum detectable
concentration was less than the A calibrator and was within a portion
of the curve that is mathematically invalid for extrapolation
(10), a functional sensitivity was determined. The
functional sensitivity was defined as the concentration at which the
mean interassay CV exceeded 15%. This was determined by
diluting two clinical samples with the zero calibrator according to the
product insert into the range of 0.5–2.0 µg/L. These samples were
then assayed in three assays to establish an interassay CV. The data
were analyzed by nonlinear regression of the observed concentration vs
CV.
Supplementation and recovery.
Supplementation and
recovery studies were performed by the addition of kit calibrators C (3
µg/L), D (10 µg/L), and E (30 µg/L) to 15 clinical samples. The
values of samples without added calibrator as determined by HPLC/MS/MS
analysis (8) and the concentrations of the calibrators as
determined by HPLC/MS/MS were used to determine the expected values for
subsequent dilutions. Recoveries were calculated as the measured
concentrations divided by the expected concentrations and expressed as
percentages.
Precision.
Intra- and interassay imprecision was
determined for all three lots of components. Intraassay imprecision was
determined by assaying 10 extractions of the kit controls and three
EDTA-whole blood controls supplemented with tacrolimus at three
concentrations across the calibration range. Interassay imprecision was
determined by assaying two extractions of the three supplemented
samples in 20 assays over a minimum of 20 days according to guidelines
suggested in NCCLS Precision Performance Guideline EP5-T2
(11). Data from this study were analyzed by ANOVA. A second
interassay imprecision study was conducted among three of the clinical
study sites. Patient samples were aliquoted and distributed to
each site for proficiency testing as part of the study protocol. The
absolute number of samples varied among sites because the use of sample
materials varied among sites.
Method comparisons.
The PRO-Trac II ELISA was compared to
HPLC/MS/MS analysis and IMx Tacrolimus II MEIA analysis on clinical
samples from a multisite clinical trial. Comparisons between the IMx
Tac I MEIA assay and the PRO-Trac II ELISA are not presented because
the first- generation MEIA is no longer commercially available. ELISA
kit calibrators, kit controls, and clinical samples for HPLC/MS/MS
analysis were shipped frozen to a private reference laboratory for
analysis according to procedures developed and validated by that
laboratory in conjunction with Fujisawa Pharmaceutical. Details of
those methods are reported elsewhere (8). Comparisons of the
PRO-Trac II ELISA and the IMx Tacrolimus II MEIA were performed as
side-by-side assays on the same day with the same set of clinical
samples (n = 95) according to the manufacturers' instructions in
the product inserts. The resulting data were analyzed several ways. A
least-squares regression analysis, comparison by the Student
t-test, and Bland-Altman analysis were performed. A subset
of clinical samples (n = 31) obtained from patients with impaired
liver function, as determined by bilirubin concentrations 30.0 mg/L,
were similarly analyzed.
Metabolite cross-reactivity.
Tacrolimus metabolites MI through
MVII were the generous gift of Fujisawa Pharmaceutical, Osaka, Japan.
Metabolite cross-reactivity was assessed by adding 5 µg/L of each
metabolite (MI–MVII) into a whole blood control containing 3.0 µg/L
parent drug. Cross-reactivity was assessed according to the equation:
 |
Interference studies.
Compounds to be tested for interference
were obtained from commercial sources at the highest purities
available. Azithromycin and fluconazole were gifts from Pfizer, New
York, NY. Ketoconazole and phenytoin were obtained as US
Pharmacopiea reference material. Clarithromycin was a gift from
Abbott Laboratories, Abbott Park, IL. Itraconazole was a gift from
Janssen Pharmaceutica, Titusville, NJ. Prednisone was provided by
Seraloids. Ganciclovir was a gift from Roche Pharmaceuticals, Palo
Alto, CA. The remaining compounds were obtained from Sigma Chemical.
Interference by commonly coadministered drugs was examined by the
addition of the test compound to aliquots of EDTA whole blood with
tacrolimus added to a concentration of 3.8 µg/L. Compounds
were tested at concentrations ~2.5-fold higher than the expected
therapeutic concentrations. Interference studies were conducted on only
one lot of kits. Concentrated stock solutions of the test drugs were
prepared in appropriate solvent vehicles. Vehicle controls were also
prepared for each solvent. Control samples, vehicle control samples,
and test samples were assayed in replicates of five each. Compounds
were considered to be interfering if the mean assay result of the test
sample was outside the expected 2 SD range of the control samples
(12). When the vehicle controls demonstrated a vehicle
effect, the 2 SD range of the vehicle control was used.
Anticoagulant studies.
The influence of heparin as an
alternative anticoagulant was examined in split patient samples drawn
at one clinical site. Blood samples (n = 15) were drawn into both
EDTA and heparin Vacutainer Tubes and processed together on the same
day by the same technician.
Sample stability studies.
To examine the stability of
tacrolimus samples, several studies were conducted. Fresh split
clinical samples (n = 33) were compared with paired samples stored
frozen for at least 48 h. One aliquot was analyzed the same day it
was obtained, whereas the second aliquot was frozen for a minimum of
48 h before analysis. Frozen clinical samples (n = 15) were
aliquoted and subjected to one to four freeze/thaw cycles and analyzed
in a single assay. Finally, clinical samples (n = 20) were
maintained at room temperature for a period of 7 days. These samples
were analyzed on days 0, 1, 2, 3, 4, and 7.
|
Results
|
|---|
The analytical performance of the PRO-Trac II ELISA is summarized
in Table 1
. The mean (± SD) minimum detectable concentration observed
over the shelf-life of the kit was 0.17 ± 0.06 µg/L. The value
reported in the product insert (0.23 µg/L) represents the upper limit
of this 1 SD range. Both values are less than the A calibrator (0.3
µg/L) and represent extrapolations (Table 1
). As expected, nondosed
EDTA samples from healthy volunteers gave concentrations less than the
reportable lower limit of the assay (0.3 µg/L).
Linearity of dilution is an important assay characteristic for the
ability to assess out-of-range samples and maintain the accuracy of the
final determination. Dilution linearity was demonstrated across the
assay range of 1–30 µg/L (Table 1
and Fig. 1
). There was no significant difference between the dilution
linearity of clinical samples or whole blood samples with added
tacrolimus.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 1. Dilution linearity of whole blood samples.
Clinical (n = 20) and supplemented (n = 9) samples were
serially diluted, assessed with three lots of components, and analyzed
by linear regression of the expected concentration vs the observed
concentration. The resulting regression line is: Observed = 0.91
(expected) + 0.57; r = 0.985.
|
|
Because the minimum detectable concentration was below the reportable
range for the assay, a functional sensitivity was determined. The
dilution linearity of the clinical samples utilized to determine the
functional sensitivity was confirmed by least-squares linear regression
of the mean observed values vs the expected values based on the assay
values of undiluted samples. The resulting regression equation was:
Observed = 0.97 (expected) + 0.1; r = 0.997 (data
not shown). Analysis of these samples by nonlinear regression is shown
in Fig. 2
, with the resulting functional sensitivity between 0.8 and 1.0
µg/L.
View larger version (13K):
[in this window]
[in a new window]
|
Figure 2. Determination of functional sensitivity by nonlinear
regression of concentration vs CV.
The mean values of diluted clinical samples is plotted vs the CV and
assessed by nonlinear regression. The functional sensitivity is the
concentration at which this regression exceeds 15% (0.75–1.0
µg/L).
|
|
Intra- and interassay imprecision was determined as described above.
Intraassay imprecision was 1.9–6.2%, as shown in Table 1
. Total
imprecision evaluated at three concentrations across the assay range
was <10% (Table 1
). When analyzed by ANOVA, total imprecision of the
20-day precision studies was <10% (6.8–7.9%) as shown in Table 2
. The interassay precision data collected at individual clinical
sites with 10 aliquoted patient samples are listed in Table 3
. These data suggest that the expected imprecision in a clinical
laboratory setting is 10–15%.
Supplementation and recovery studies were performed to assess the
accuracy of the assay. These studies were performed utilizing clinical
samples and materials readily accessible to the clinical laboratory.
The mean recovery across the range of the assay from 2.5 to 30 µg/L
was 109% ± 11%.
Least-squares linear regression analysis (Fig. 3
) shows a good correlation between the two commercially
available methods [IMx Tacrolimus II = 1.04(PRO-Trac II) + 2.2;
r = 0.89], although they do not return statistically
equivalent values for the same samples. A mean difference of 2.7 µg/L
was seen between the two methods. This difference was statistically
significant by the Student paired t-test (P
<0.001). In addition, a Bland-Altman analysis was performed as shown
in Fig. 4
. HPLC/MS/MS values were available for these samples, and
comparisons of the HPLC/MS/MS values to the PRO-Trac II values were
performed (Fig. 5
and
Fig. 6
). The mean difference between the HPLC/MS/MS and PRO-Trac II
values was -0.8 µg/L. To provide clinical perspective, we calculated
the percentage of samples analyzed by both methods that varied from the
HPLC/MS/MS values by <3, 5, 8, and 10 µg/L. The 
5 µg/L
difference threshold comprised 95% of the PRO-Trac II samples and 70%
of the IMx Tacrolimus II samples These comparisons are summarized in
Tables 4 and
5.
View larger version (15K):
[in this window]
[in a new window]
|
Figure 3. Method comparison between PRO-Trac II ELISA and IMx
Tacrolimus II MEIA by linear regression.
Clinical samples (n = 95) were assessed in side-by-side assays on
the same day and analyzed by least-squares linear regression. The
resulting regression line is: IMx = 1.04 (PRO-Trac II) + 2.2;
r = 0.89.
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Figure 4. Bland-Altman analysis of the difference between the
PRO-Trac II ELISA and the IMx Tacrolimus II MEIA.
Clinical samples (n = 95) were assayed by the PRO-Trac II ELISA
and the IMx Tacrolimus II MEIA and analyzed by Bland-Altman analysis.
The mean value of the two methods is plotted against the difference
between the two values (PRO-Trac - IMx). The mean difference (±
SD) between the two methods was -2.7 ± 2.3 µg/L. The mean
and ± 2 SD lines (——–) are plotted for reference.
|
|
View larger version (15K):
[in this window]
[in a new window]
|
Figure 5. Method comparison between HPLC/MS/MS vs PRO-Trac II by
linear regression.
Clinical samples (n = 95) were assayed by HPLC/MS/MS and PRO-Trac
II ELISA. The data were then analyzed by linear regression, with the
resulting regression equation: PRO-Trac II = 0.95(HPLC/MS/MS) +
1.3; r = 0.83.
|
|
View larger version (17K):
[in this window]
[in a new window]
|
Figure 6. Method comparison between HPLC/MS/MS and PRO-Trac II ELISA
by Bland-Altman analysis.
Clinical samples (n = 95) were assayed by HPLC/MS/MS and PRO-Trac
II ELISA and analyzed by Bland-Altman analysis. The mean value of the
two methods is plotted against the difference between the two values
(HPLC/MS/MS - PRO-Trac). The mean difference (± SD) between the
two methods was -0.8 ± 2.3 µg/L. The mean and ± 2 SD
lines (——–) are plotted for reference.
|
|
Method comparison analyses were also performed on a subset of 31
samples with high serum bilirubin concentrations (30.0 mg/L). The
mean (± SD) serum bilirubin concentration in this subset was 78
± 59 mg/L. Least-squares regression analysis produced the equation:
PRO-Trac II = 1.1(HPLC/MS/MS) + 1.3; r = 0.90.
When analyzed by the Student t-test, the resulting
P value was 0.109, indicating no statistically significant
difference between the HPLC/MS/MS and PRO-Trac values. The mean (± SD)
difference between the HPLC/MS/MS and PRO-Trac II values was -2.4
± 3.1 µg/L. The corresponding evaluation in the remaining 99
samples with bilirubin values within the reference interval produced
the regression equation: PRO-Trac II = 1.1(HPLC/MS/MS) + 0.04;
r = 0.92; P = 0.221, t-test;
mean difference, -1.0 ± 2.5 µg/L.
Metabolite cross-reactivity was assessed as described above (Table 6
). Metabolites MII, MIII, and MV exhibited cross-reactivities of
84%, 36%, and 42%, respectively. This metabolite cross-reactivity is
consistent with previously published reports for this antibody
(13).
Thirty-five commonly coadministered drugs were assessed for potential
interference in the PRO-Trac II assay system. None of the tested
compounds exhibited interference with the PRO-Trac II ELISA. These data
are summarized in Table 7
.
The effects of various sample storage conditions are summarized in
Table 8
. The comparison of fresh samples to frozen samples was
conducted with 33 clinical samples. The mean difference between fresh
and frozen samples was -0.5 ± 1.4 µg/L. Sample stability was
also assessed under repeated freeze/thaw conditions. Fifteen clinical
samples were subjected to one to four freeze/thaw cycles. The zero
freeze/thaw value was then used to calculate the difference from
baseline for each sample after each cycle. The mean differences from
baseline ranged from -0.6 to 0.7 µg/L.
Finally, samples were examined over the course of 1 week of storage at
room temperature. The difference from the day 0 value was calculated
for each sample on days 1, 2, 3, 4 and 7. The mean differences for all
samples ranged from -0.5 to 0.8 µg/L The mean difference for
individual samples over all time points ranged from -0.5 to 1.4
µg/L. There was no significant trend in the mean difference with time
of storage.
The use of heparin as an alternative anticoagulant was examined in
fresh clinical samples. The mean difference between EDTA and heparin
samples from clinical subjects was -0.7 ± 1.5 µg/L. These data
are summarized in Table 8
.
|
Discussion
|
|---|
At the present time, there are six generally used methods for the
measurement of tacrolimus in whole blood (14). These methods
include receptor-binding, bioassay, pentamer formation, HPLC with
various detection methods, MEIA, and ELISA technologies. Because of the
differences in methodologies, extraction procedures, and metabolite
recognition, results are not interchangeable among all methods.
Receptor-binding assays for tacrolimus have been developed using
purified immunophilins (FKBPs) (15). Although
receptor-binding assays show promise, the results indicate that this
method also measures substantial metabolite concentrations
(14). Various HPLC methods have been reported, although
these methods generally suffer a lack of sensitivity because there is
no chromophore or ultraviolet-absorbing structure in the molecule
(16)(17)(18)(19). HPLC/MS and HPLC/MS/MS methods that exhibit
thenecessary sensitivity have also been developed, but because of
equipment requirements, they may not be practical for routine
monitoring (8)(20)(21).
Differences in methodologies and values between methodologies makes the
choice and validation of an assay method an important necessity for
modern clinical laboratories. The data presented here are intended to
provide a baseline for validation studies for the PRO-Trac II ELISA for
the measurement of tacrolimus in whole blood samples.
The estimated minimum detectable concentration of the PRO-Trac II ELISA
was 0.18 µg/L, which is 27-fold lower than the 5 µg/L suggested as
the lower limit of the therapeutic range reported in the Lake Louise
Consensus Conference report (4) and ~10-fold lower than
the detection limit of the IMx Tacrolimus II MEIA (22). This
analytical value can only be an estimation because the minimum
detectable concentration is calculated on the assumption of a linear
extrapolation of the calibration curve between the zero and A
calibrators. Although this calculation gives a reproducible value, the
validity of extrapolations between the zero and A calibrators for
four-parameter logistic curve fits is questionable (10). For
this reason, we calculated a functional sensitivity of 1.0 µg/L based
on the concentration at which the mean CV exceeded 15%. This cutoff
was somewhat arbitrary because some proficiency testing schemes
consider CVs of 20% acceptable (23). The performance of the
PRO-Trac II ELISA in the subtherapeutic and low therapeutic ranges is
noteworthy (24). The current therapeutic range will likely
decrease as transplantation physicians balance organ rejection and
toxicity by using lower tacrolimus dosages and additional
immunosuppressants (25)(26).
Several estimates of precision were determined. The 20-day interassay
estimates performed at DiaSorin are likely to represent the most ideal
conditions and can be used to establish a baseline CV for assay
performance of ~8% across the range of clinically relevant
concentrations. The expected assay CV under typical clinical laboratory
conditions was shown to be ~12% across the range of clinically
relevant concentrations.
Supplementation and recovery studies were performed in a format that
could be repeated by any clinical laboratory, without accessing
pharmaceutical grade tacrolimus. The clinical samples, when diluted
with kit calibrators, exhibited recovery of ~109%. This is in
contrast to the results obtained by other investigators, who have
reported decreased recovery using supplemented samples (27).
Any differences in sample preparation, such as solvent content, sample
size, metabolite contribution, and the loss of parent drug on glass
surfaces, make direct comparisons equivocal.
The method comparisons demonstrate a statistical difference between the
PRO-Trac II ELISA and the IMx Tacrolimus II MEIA. For better
therapeutic drug monitoring, physicians should be aware of the
differences. The mean statistical difference of 2.7 µg/L suggests a
need for monitoring by a single method or notation of the method used
when results are reported. Some investigators have reported that the
difference between methods may have clinical significance
(6)(27). The basis of this difference between
methods is currently unclear because both assays utilize the same
monoclonal antibody. One could speculate that the differences in
soluble vs bound antibody and/or the presence of residual organic
solvents utilized for drug extraction may alter antibody affinity and,
therefore, contribute to the differences in observed values. Other
factors, such as assay incubation times and temperatures, that affect
assay kinetics may also play a role in the observed differences.
The agreement between the PRO-Trac II values and HPLC/MS/MS values
(mean difference of 0.8 µg/L; Table 4
) suggests that the
concentrations of cross-reactive metabolites in most samples is
minimal. The values reported here are consistent with the 96% of total
immunoreactivity reported for the parent drug fraction in patients with
mild hepatic dysfunction (28), and less than the mean 42%
reported for a study of 21 liver transplant patients (29).
The choice of difference thresholds in the comparison of methods is
somewhat arbitrary. The clinical significance of the difference between
HPLC/MS/MS and ELISA has not been determined. Both methods exhibit
random measurement error, and the "true" value for a given sample
when assessed by HPLC/MS/MS is determined only within a range of
values. Given the realistic CVs observed in clinical laboratories
(Table 3
), the expected differences between methods can be estimated to
be ~20% if both assays measured only parent drug. Metabolites may be
expected to contribute an additional 5–10% to that variation. The
clinical significance of such a difference needs to be examined
further.
Patients with impaired liver function may accumulate metabolites,
contributing to erroneous tacrolimus blood values when assayed by
either ELISA or MEIA (30). In the subset of the patient
population for whom HPLC/MS/MS sample values were available, samples
with bilirubin values 30.0 mg/L, an indication of impaired liver
function, showed a mean difference of 2.4 µg/L. In this case, the
difference did not reach statistical significance but may have clinical
significance. Although these values are increased relative to
HPLC/MS/MS and should probably be flagged to be monitored closely, the
general population of impaired liver function subjects in this study
did not seem to exhibit the degree of metabolite accumulation observed
by other investigators (29)(30).
The anticoagulant studies demonstrated that the PRO-Trac II ELISA
provides the laboratory with potential flexibility and economy in the
planning and acquisition of patient samples. The sample stability
studies were consistent with previous reports (31)(32)(33).
Overall this analytical validation demonstrates that the PRO-Trac II
ELISA is a robust assay for the assessment of tacrolimus whole blood
concentrations in liver transplant patients receiving tacrolimus for
immunosuppression.
|
Acknowledgments
|
|---|
This study was funded by DiaSorin Inc., Stillwater, MN. We
acknowledge the efforts of several individuals who contributed to the
completion of this work. Excellent technical assistance and clinical
coordination was provided by L. Fields, S. Mehta, B. Forrester, M.
Virji, V. Esquanazi, S. Babischkin, L. Maxwell, L. Ramanathan, E.
Weiszmann, I. Fernandez, R. Miller, A. Reyes, E. Culligan, C. Janus,
and D. Weppler.
|
References
|
|---|
-
Kino T, Hatanaka H, Hashimoto M, Nishiyama M, Goto T, Okuhara M, et al. FK506, a novel immunosuppressant isolated from a Streptomyces. I. Fermentation, isolation, and physio-chemical and biological characteristics. J Antibiotics 1987;40:1249-1255.
[Medline]
[Order article via Infotrieve]
-
Venkataramanan R, Swaminathan A, Prasad T, Jain A, Zuckerman S, Warty V, et al. Clinical pharmacokinetics of tacrolimus. Clin Pharmacokinet 1995;29:404-430.
[ISI][Medline]
[Order article via Infotrieve]
-
Wallemacq PE, Reding R. FK506 (tacrolimus), a novel immunosuppressant in organ transplantation: clinical, biochemical, and analytical aspects. Clin Chem 1993;39:2219-2228.
[Abstract]
-
Jusko WJ, Thomson AW, Fung JJ, McMaster P, Wong SH, Zylber-Katz E, et al. Consensus document: therapeutic monitoring of tacrolimus (FK506). Ther Drug Monit 1995;17:606-614.
[ISI][Medline]
[Order article via Infotrieve]
-
Tamura K, Kobayashi M, Hashimoto K, Kojima K, Nagase K, Iwasaki K, et al. A highly sensitive method to assay FK506 levels in plasma. Transplant Proc 1987;19:23-29.
[ISI][Medline]
[Order article via Infotrieve]
-
MacFarlane GD, Scheller D, Ersfeld D, Jensen T, Jevans A, Kobayashi M. A simplified whole blood ELISA [PRO-TracTM II] for tacrolimus (FK506) using a proteolytic extraction in place of organic solvents. Ther Drug Monit 1996;18:698-705.
[ISI][Medline]
[Order article via Infotrieve]
-
Grenier FC, Luczkiw J, Bergmann M, Lunetta S, Morrison M, Blonski D, et al. A whole blood FK506 assay for the IMx analyzer. Transplant Proc 1991;23:2745-2747.
[ISI][Medline]
[Order article via Infotrieve]
-
Alak AM, Moy S, Cook M, Lizak P, Niggebiugge A, Menard S, Chilton A. An HPLC/MS/MS assay for tacrolimus in patient blood samples. Correlation with results of an ELISA assay. J Pharm Biomed Anal 1997;16:7-13.
[ISI][Medline]
[Order article via Infotrieve]
-
Smith M. Method evaluation. Wild D eds. The immunoassay handbook 1994:256-262 Stockton Press New York. .
-
Maciel RJ. Standard curve fitting in immunodiagnostics: a primer. J Clin Immunoassay 1985;8:98-106.
-
National Committee for Clinical Laboratory Standards.
Evaluation of precision performance of clinical chemistry devices,
second edition. NCCLS Document EP5–T2. Villanova, PA: NCCLS, 1992..
-
National Committee for Clinical Laboratory Standards.
Interference testing in clinical chemistry. NCCLS Document EP7-P.
Villanova, PA: NCCLS, 1986..
-
Iwasaki K, Shiraga T, Matsuda H, Nagase K, Tokuma Y, Hata T, et al. Further metabolism of FK506 (tacrolimus). Identification and biological activities of the metabolites oxidized at multiple sites of FK506. Drug Metab Dispos 1995;23:28-34.
[Abstract]
-
Alak A. Measurement of tacrolimus (FK506) and its metabolites: a review of assay development and application in therapeutic drug monitoring and pharmacokinetic studies. Ther Drug Monit 1997;19:338-351.
[ISI][Medline]
[Order article via Infotrieve]
-
Murthy JN, Chen Y, Warty VS, Venkataramanan R, Donnelly JG, Zeevi A, Soldin S. Radioreceptor assay for quantifying FK506 immunosuppressant in whole blood. Clin Chem 1992;38:1307-1310.
[Abstract/Free Full Text]
-
Vincent SH, Karanam BV, Painter SK, Chiu S-HL. In vitro metabolism of FK506 in rat, rabbit, and human liver microsomes: identification of a major metabolite and of cytochrome P450 3A as the major enzymes responsible for its metabolism. Arch Biochem Biophys 1992;294:454-460.
[ISI][Medline]
[Order article via Infotrieve]
-
Perotti BYT, Prueksaritanont T, Benet LZ. HPLC assay for FK506 and two metabolites in isolated rat hepatocytes and rat liver microsomes. Pharm Res 1994;11:844-847.
[ISI][Medline]
[Order article via Infotrieve]
-
Takada K, Oh-Hashi M, Yoshikawa H, Muranishi S, Nishiyama M, Yoshida H, et al. Determination of a novel immunosuppressant (FK506) in rat serum and lymph by high-performance liquid chromatography with chemiluminescence detection. J Chromatogr 1990;530:212-218.
[ISI][Medline]
[Order article via Infotrieve]
-
Beysens AJ, Beuman GH, van der Heijden JJ, Hootgtanders KEJ, Steijger OM, Lingeman H. Determination of tacrolimus (FK506) in whole blood using liquid chromatography and fluorescence detection. Chromatographia 1994;39:490-496.
-
Taylor PJ, Jones A, Balderson GA, Lynch SV, Norris RLG, Pond SM. Sensitive, specific quantitative analysis of tacrolimus (FK506) in blood by liquid chromatography-electrospray tandem mass spectrometry. Clin Chem 1996;42:279-285.
[Abstract/Free Full Text]
-
Christians U, Braun F, Schmidt M, Kosian N, Schiebel HM, Ernst L, et al. Specific and sensitive measurement of FK506 and its metabolites in blood and urine of liver grafted patients. Clin Chem 1992;38:2025-2032.
[Abstract]
-
Garg UC, Austin G, Barnes C, Hamilton M. Comparison of the Abbott IMx tacrolimus I and tacrolimus II assays. Clin Chem 1998;44:1783-1785.
[Free Full Text]
-
Alak A, Lizak P. Quality assurance procedure for monitoring tacrolimus (FK506) concentrations in whole blood by IMx assay. Ther Drug Monit 1996;18:267-272.
[ISI][Medline]
[Order article via Infotrieve]
-
Lee JW, Sukovaty RL, Farmen RH, Dressler DE, Alak AM, Bekersky I. Tacrolimus (FK506). Validation of a sensitive enzyme-linked immunosorbent assay kit for and application to a clinical pharmacokinetic study. Ther Drug Monit 1997;19:201-207.
[ISI][Medline]
[Order article via Infotrieve]
-
Cai TH, Esterl RM, Jr, Nichols L, Cigarroa F, Speeg KV, Halff GA. Improved immunosuppression with combination tacrolimus (FK506) and mycophenolic acid in orthotopic liver transplantation. Transplant Proc 1998;30:1413-1414.
[ISI][Medline]
[Order article via Infotrieve]
-
Eckhoff DE, McGuire BM, Frenette LR, Contreras JL, Hudson SL, Bynon JS. Tacrolimus (FK506), mycophenolate mofetil combination therapy versus tacrolimus in adult liver transplantation. Transplantation 1998;65:180-187.
[ISI][Medline]
[Order article via Infotrieve]
-
Cao TZ, Jevans A, Brown G, Linder MW. Valdes R Jr. Discrepancies between tacrolimus concentrations measured using the MEIA and ELISA methods are due to differences in drug recovery [Abstract]. Clin Chem 1998;44:A87.
-
Tokunaga Y, Alak AM. FK506 (tacrolimus), its immunoreactive metabolites in whole blood of liver transplant patients and subjects with mild hepatic dysfunction. Pharm Res 1996;13:137-140.
[ISI][Medline]
[Order article via Infotrieve]
-
Gonschior AK, Christians U, Winkler M, Linck A, Baumann J, Sewing KF. Tacrolimus (FK506) metabolite patterns in blood from liver and kidney transplant patients. Clin Chem 1996;42:1426-1432.
[Abstract/Free Full Text]
-
Braun F, Schutz E, Christians U, Lorf T, Schiffmann JH, Armstrong VW, et al. Pitfalls in monitoring tacrolimus (FK506). Ther Drug Monit 1997;19:628-631.
[ISI][Medline]
[Order article via Infotrieve]
-
Alak AM, Lizak P. Stability of FK506 in blood samples. Ther Drug Monit 1996;18:209-211.
[ISI][Medline]
[Order article via Infotrieve]
-
Freeman DJ, Stawecki M, Howson B. Stability of FK506 in whole blood samples. Ther Drug Monit 1995;17:266-267.
[ISI][Medline]
[Order article via Infotrieve]
-
Ingels SC, Koenig J, Scott MG. Stability of FK506 (tacrolimus) in whole blood specimens. Clin Chem 1995;41:1320-1321.
[Free Full Text]
Top
Abstract
Introduction
