Clin Biochem, Vol. 24, pp. 43-48, 1991 Printed in Canada. All rights reserved.
0009-9120/91 $3.00 + .00 Copyright © 1991 The Canadian Society of Clinical Chemist&
Measurement of Cyclosporine A by a Specific Radioimmunoassay with a Monoclonal Antibody and 251 Tracer SUNG C. LEE, 1 ALICE M. BRUDZINSKI, 1 JOAN L. YASMINEH, 1 NANCY J. JOHANSON, 1 IRENE A. FERBER, 1 LORI K. MAAS, 1 P. Y. WONG, 2 PAUL A. KEOWN, 3 and JOHN W. ORF 1 I lNCSTAR Corporation, Stillwater, MN, USA, 2Toronto General Hospital, Toronto, Ontario, Canada, and 3British Columbia Transplant Society, Vancouver, BC, Canada We developed a sensitive radioimmunoassay (CYCLO-Trac ® SP) that specifically measures cyclosporine A in serum, plasma and whole blood of transplant patients. The specific monoclonal antibody was from Sandoz and the tracer was an 12Slderivative of cyclosporine C. The assay is performed at room temperature for 1 h followed by a 20 min centrifugation. The sensitivities of the assays are 2.6 ng/mL and 8.7 ng/mL for the serum/plasma assay and the whole blood assay, respectively. Within-run and between-run CVs for both types of assays using cyclosporine concentrations of 80 and 58 ng/ mL (serum) and 186 and 199 ng/mL (whole blood) were less than 5% and 9%, respectively. Averaged recovery of serum/plasma and whole blood assays at various levels ranged from 93% to 115%. Interferences by bilirubin, triglyceride, cholesterol, hemoglobin, OKT3, azathioprine, methylprednisolone and 20 other drugs were insignificant. Multicenter proficiency studies showed an excellent correlation between the CYCLO-Trac ® SP and the specific 3HSandimmune ® assay from Sandoz: whole blood assay (r = 0.998) and serum assay (r = 0.997).
KEY WORDS: cyclosporine A; drug monitoring; radioimmunoassay; transplantation; monoclonal antibody.
unclear how the immunosuppressive activity of some metabolites demonstrated in vitro (13) relate to patient care. For these reasons, the American Association of Clinical Chemistry task force on cyclosporine monitoring has recommended the specific measurement of the parent compound in whole blood specimens (14). Recently, a monoclonal antibody that binds specifically to the parent compound cyclosporine A was developed by Sandoz (15). The RIA using this specific monoclonal antibody (Sandimmune ®) closely matched HPLC results in samples from patients with kidney, heart, and lung transplants (16,17). However, the charcoal separation and the need to use liquid scintillation counting for the tritium labeled tracer made the assay inconvenient. We report here the development of an RIA using the same specific monoclonal antibody in the form of a preformed immune-complex and an z2sI-tracer.
Introduction
Materials a n d m e t h o d s ince the introduction of cyclosporine A by Sandoz Corporation (Basel, Switzerland) as an immunosuppressive drug in organ transplantation, the need to monitor the drug level in individual patients has been well recognized (1-3). Analytical methods using high-performance liquid chromatography (HPLC) (4-6) or radioimmunoassay (RIA) with polyclonal antibodies (7-9) have been developed. However, the two types of methods give very different values due to the difference in biotransformation among patients and high cross-reactivities of polyclonal antibodies with metabolites of cyclosporine A (10-11). Moreover, cyclosporine values vary according to the specimen matrix and the method of specimen preparation (12). It is still
S
Correspondence: Sung C. Lee, Ph.D., INCSTAR Corporation, 1990 Industrial Blvd., Stillwater, MN 55082, USA. Manuscript received May 21, 1990; revised July 9, 1990; accepted August 21, 1990. CLINICAL
BIOCHEMISTRY,
VOLUME
24, FEBRUARY
1991
MATERIALS 1 2 5 I labeled tracer was prepared by alkylating a derivative of cyclosporine C with radioiodohistamine. Cyclosporine A, cyclosporine C and the specific monoclonal antibody (#78-299) were from Sandoz Corporation (Basel, Switzerland). The USP reference standard of cyclosporine A (lot F, 98.9% purity) was from the United States Pharmacopeia, Rockville, MD, USA.
METHODS
Assay procedures Serum/plasma assays were performed by adding 100 ~L aliquots of the tracer (60,000 cpm) and i mL aliquots of Immunosep (preformed complex of the anticyclosporine monoclonal antibody, normal mouse sera and donkey antimouse antisera) to 25 ~L ali43
LEE, BRUDZINSKI, YASMINEH, ~r AL. TABLE 1 Precision of Cyclosporine A I m m u n o a s s a y U s i n g
CYCLO-Trac ® SP Without Outliers Deleted Whole Blood
Serum
10.7 (46)a 3.2 (186) 3.1 (625)
10.2 (29) 4.2 (80) 5.1 (247)
9.8 (81) 8.4 (199) 9.4 (547)
12.9 (28) 7.9 (58) 9.8 (226)
Within-run b
Low Medium High Between-run c Low Medium High
aMean values in ng/mL are shown in parentheses. bWithin-run CVs (%) were determined by running each specimen in 25 replicates as described in Methods. CBetween-run, within-center CVs (%) were the averages from the 5-center proficiency studies as described in Methods. quots of specimens. The mixture was incubated at room temperature for 1 h and centrifuged for 20 min at 1600 x g. B/Bo vs. log cyclosporine concentration in a spline fit by RIACALC software package (LKB, Bromma, Sweden) was used for data reduction. Whole blood assays were identical to the serum/plasma assay except that extraction of specimens and standards with methanol was carried out prior to RIA. The extraction involved adding 100 ~LL of whole blood specimen directly into 400 ~L of methanol with immediate mixing. The mixture was centrifuged for 5 min and 50 ~LL of the supernatant was assayed in the ensuing RIA. The 3HSandimmune ® assay from Sandoz was run according to their published protocol.
Preparation of standards Standards of cyclosporine A were prepared gravimetrically using a pool of processed h u m a n serum and calibrated by CYCLO-Trac ® SP assays to U S P reference material that had been made gravimetrically in h u m a n whole blood.
five centers: Vancouver General Hospital (Vancouver, BC, Canada), INCSTAR Corp. (Stillwater, MN, USA), St. Lukes Episcopal Hospital (Houston, TX, USA), Albany Medical Center (Albany, NY, USA), and Toronto General Hospital (Toronto, Ontario, Canada). Thirty-five whole blood specimens (ranging from 36 ng/mL to 977 ng/mL) and 35 serum specimens (ranging from 18 ng/mL to 512 ng/mL) were analyzed in 5 different assays by each of the above 5 centers. For the calculation of between-run, within-center CVs, 35 specimens were grouped into low, medium, and high concentration specimens based on their B/Bos and the results were averaged. Statistical analyses of the data were performed with Statistix H (NH Analytical Software, Roseville, MN, USA) and included a nested analysis of variance with and without outliers deleted (18).
Recovery and dilution parallelism Recovery studies were performed by spiking 5 whole blood patient specimens, 5 undosed serum specimens, and 5 undosed plasma specimens with three different levels (low, medium, and high) of cyclosporine A that had been determined by gravimetric procedures. Dilution parallelism studies were performed by diluting 30 specimens each of whole blood, serum, and plasma at 1:2, 1:4, 1:8, and 1:16 with the zero standard in the kit according to the recommendations in the kit protocol. Whole blood specimens were extracted with methanol at 1:9, 1:19, 1:39, and 1:79. Results REACTION CHARACTERISTICS
Standards prepared using undosed plasma and whole blood showed superimposing suppression lines when extracted with methanol prior to RIA (Figure 1). The reaction kinetics reached a plateau in both the whole blood and serum/plasma assays in 1 h at room temperature (data not shown). The sensitivities of the assays were 2.6 ng/mL and 8.7 ng/mL for the serum/plasma and the whole blood assay, respectively.
Sensitivity and precision PRECISION
Sensitivity was determined by running 20 replicates of the zero standard and the lowest concentration standard in the kit in each type of assay three times. The sensitivity was calculated by taking the value at 3 SD less than the mean counts for the zero standard. Within-run CVs were obtained by running 3 different levels of cyclosporine A in whole blood and serum matrices in 25 replicates. Low, medium, and high value specimens of about 70%, 50%, and 25% B/Bo were used for within-run CV determinations. Between-run CVs were obtained from a multicenter study conducted at the following
44
Precision profiles for within-run CVs and between-run, within-center CVs are shown in Table 1. Within-run CVs and between-run, within-center CVs for both types of assay at medium concentrations were less than 5% and 9%, respectively. The CVs for low and high values of cyclosporine A were in general greater than those observed for the medium range. The observed between-run, within-center CV of 12.9% in the ser,,m assay for the low concentration range was primarily the result of an unusually high 25.2% CV from one of the centers in the
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
MEASUREMENT
OF C Y C L O S P O R I N E A B Y R A D I O I M M U N O A S S A Y
90. 80-
706050-
Z(BIBo) 403020100 10
I 1O0
I 1000
I 1000
Cyclosporine A (ng/mL) Figure 1--Standard curves in the CYCLO-Trac® SP whole blood assay. Cyclosporine A standards prepared in undosed plasma (dashed line) and in undosed whole blood (solid line) were compared in the whole blood assay. Each standard was extracted with methanol by adding the sample into the methanol according to the kit protocol. (B/Bo) % is the percentage of bound cpm divided by cpm of the zero standard. study. When this CV was excluded, the average between-run CV was 9.8%, conforming to the general pattern observed for the rest of the data. The results of the nested analyses of variance of both serum/plasma and whole blood assays in the 5-center studies are shown in Table 2. Overall CVs were: 9.9% with outliers deleted (10.2% without outliers deleted) for the whole blood assay and 11.8% with outliers deleted (13.0% without outliers deleted) for the serum/plasma assay. I ~ T H O D COMPARISON
Comparisons of mean values by sH-Sandimmune ® and CYCLO-Trac ® SP assays from the 5center study are shown in Figures 2(a) and 2(b).
Recovery and dilution parallelism The recoveries ranged from 100-106%, 105-111%, and 93-115% for whole blood, serum, and plasma specimens, respectively. The overall averaged recoveries were 103%, 108%, and 104% for whole blood, serum, and plasma, respectively. The significance levels of the difference between the mean recovery percentages and 100% (perfect recovery) were p = 0.13, p < 0.01, and p = 0.50 for whole blood, serum, and plasma, respectively. Studies of dilution parallelism of whole blood, serum, and plasma specimens yielded values of the diluted specimens close to those expected.
Interference Interference by the following compounds was insignificant in both serum/plasma and whole blood
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
assays (the maximum concentrations of several levels tested are shown in parentheses): bilirubin (0.51 retool/L), triglyceride (13 retool/L), cholesterol (25.8 mmol/L), hemoglobin (0.078 mmol/L, serum/plasma assay only), OKT-3 (10 nmol/L), azathioprine (200 ng/mL), 6-methylprednisolone (60 ng/mL), cimetidine (1000 ng/mL). The following drugs tested at concentrations up to 100 tLg/mL also did not interfere in the assays: acetaminophen, cephalosporine, digoxin, ethosuximide, kanamycin, metoclopramide, penicillin, rifampin, tobramycin, N-nifedipine, Nacetylprocainamide, chloramphenicol, disopyramide, gentamycin sulfate, lidocaine, naficillin, propranolol, salicyclic acid, and verapamil. Discussion
Knepil and McPhillips (19) claimed 6% to 10% negative bias when comparing CYCLO-Trac ® SP standards to their in-house standards. Our studies showing a good agreement between the Sandoz method and CYCLO-Trac® SP method did not indicate 6 to 10% negative bias in CYCLO-Trac ® SP standards compared to Sandoz standards. Copeland and Yatscoff (20) claimed a positive bias of 35% compared to the Sandoz assay and inaccurate standardization in the CYCLO-Trac ® SP whole blood kit. However, after carefully following the kit procedures for data reduction, they found only 10 to 15% difference between the two methods (R. Yatscoff, personal communication, August 1990). McBride et al. (21) claimed that standards in the CYCLO-Trac ® SP plasma assay caused a positive bias of 23% when compared with their in-house standards. However, even when using their own
45
LEE, BRUDZINSKI, YASMINEH, ST AL.
(a) 1000. t
900.
_J
R= 0.998
E ¢-
J
SLOPE= 1.02
800.
/. ---
D
700.
v
0. b")
600500-
L,.
I--
I o
300-
._.J fJ >-. 200(.~ 1000
I
0
1O0
I 200
I 300
I 400
I 500
I 600
I 700
I 800
I 900
1ooo
S a n d i m r n u n e (ng/mL) (b) 550-
"E
500
R=
450
SLOPE= 0 . 9 9 3
0.997
•
..E. = 350, 13-
(/)
soo
(,,I C) s-
250. 200.
I
150
._J
loo.
t'--
0
>(_)
50. 0 0
l 50
I 100
I 150
I 200
I 250
I 300
I 350
I 400
I 450
I 500
Sondimmune (ng/mL) Figure 2--Comparison of mean values of 35 specimens by SancSmmune ® and CYCLO-Trac ® SP assays. The specimens were analyzed by the above two methods in five independent runs in each of the five centers as described in Methods. (a) Whole blood specimens; (b) Serum specimens. in-house standards, they still observed widely variable discrepancies between values for CYCLO-Trac ® SP and HPLC. They postulated the presence of interfering substances, such as metabolites crossreacting with the specific monoclonal antibody in the CYCLO-Trac ® SP assay. As noted in this report, there are no interfering substances which have been shown to significantly affect the CYCLO-Trac ® SP assays. The cross-reactivity of major metabolites in the CYCLO-Trac ® SP assays has been evaluated (22). Based on expected levels of cyclosporine metabolites and cyclosporine A in clinical specimens
46
(23,24), less than a 5% increase in cyclosporine levels can be attributed to the presence of metabolites. Johnston et al. (25) have compiled values of their Q.C. survey from many laboratories in Europe between February 1988 and August 1988. Total analytical error of the CYCLO-Trac ® SP whole blood assay was within 6% (between 100 ng/mL and 1000 ng/mL) and the values of patient pools were about 25% higher than those by HPLC. The CVs for HPLC methods are much greater than those for immunoassays, consistent with the widely different
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
MEASUREMENT
OF CYCLOSPORINE
TABLE2 Nested Analysis of Variance of CYCLO-Trac ® SP in the 5-center Proficiency Studies
A BY RADIOIMMUNOASSAY
1. Bowers LD, Canafax DM. Cyclosporine: experience with therapeutic monitoring. Ther Drug Monit 1984; 6: 142-7. 2. Kahan BD, ed. Cyclosporine-associated renal injury. Transplant Proc 1985; 17 (Suppl. 1). 3. Ptachcinsky RJ, Vankataramanan R, Burckart GJ. Clinical pharmacokinetics of cyclosporine. Clin Pharmacokinet 1986; 11: 107-32. 4. Carruthers SG, Freeman DJ, Koegler JC, et al. Simplified liquid-chromatographic analysis for cyclosporine A, and comparison with radioimmunoassay. Clin Chem 1983; 29: 180--4. 5. Moyer TP, Johnson P, Faynor SM, et al. Cyclosporine: a review of drug monitoring problems and presentation of a simple, accurate liquid chromatographic procedure that solves these problems. Clin Biochem
1986; 19: 83-9. 6. Annesley T, Matz K, Balogh L, Clayton L, Giacherio D. Liquid chromatographic analysis of cyclosporine utilizing a microbore column and reduced specimen requirement. Clin Chem 1986; 32: 1407-9. 7. Donatsch P, Abisch E, Homberger M, Traber K, Trapp M, Voges R. A radioimmunoassay to measure cyclosporine A in plasma and serum samples. J Immunoassay 1981; 2: 19-32. 8. Wong PY, Cheung M, Yip TK, Mee AV. Use of 12SI-labeled-histamine-cyclosporine C for monitoring serum cyclosporine concentrations in transplantation patients. Clin Chem 1986; 32: 492-5. 9. Mahoney WC, Orf JW. Derivatives of cyclosporine compatible with antibody-based assays: 1. The generation of 12SI-labeled cyclosporine. Clin Chem 1985; 31: 459-62. 10. Robinson WT, Schran HF, Barry EP. Methods to measure cyclosporine l e v e l s - high-pressure liquid chromatography, radioimmunoassay and correlation. Transplant Proc 1983; 15 (Suppl. 1/2): 2403-8. 11. Blyden GT, Franklin C, Cho SI, Kaplan MM, Hirsch CA, Greenblatt DJ. Cyclosporine blood concentrations determined by specific versus nonspecific assay methods. J Clin Pharmacol 1986; 26:367-71. 12. Yatscoff RW, Rush DN, Jeffery Jr. Effects of sample preparation on concentrations of cyclosporine A measured in plasma. Clin Chem 1984; 30: 1812-4. 13. Freed BM, Rosano TG, Lempert N. In vitro immunosuppressive properties of cyclosporine metabolites. Transplantation 1987; 43: 123-7. 14. Shaw WL, Bowers L, Demers L, et al. Critical issues in cyclosporine monitoring: report of the task force on cyclosporine monitoring. Clin Chem 1987; 33: 126988. 15. Quesniaux V, Tees R, Schreier MH, Maurer G, Regenmortel MHV. Potential of monoclonal antibodies to improve therapeutic monitoring of cyclosporine. Clin Chem 1987; 33:32-7. 16. Schran HF, Rosano TG, Hassell AE, Pell MA. Determination of cyclosporine concentrations with monoclonal antibodies. Clin Chem 1987; 33: 2225-9. 17. Ball PE, Munzer H, Keller HP, Abisch E, Rosenthaler J. Specific 3H radioimmunoassay with a monoclonal antibody for monitoring cyclosporine in blood. Clin Chem 1988; 34: 357-60. 18. Hawkins DM. Outliers from linear models. In: Hawkins DM, ed. Identification of outliers. Pp. 85-103. London: Chapman and Hall, 1980. 19. Knepil J, McPhillips M. Calibration of ~2sI immunoassay measuring cyclosporine A. Clin Chem 1989; 35: 181-2. 20. Copeland KR, Yatscoff RW. Calibration problems with the 12sI immunoassay for measuring cyclosporine A. Clin Chem 1989; 35: 1551-2. 21. McBride JH, Rodgerson SS, Reyes AF. Measurement of cyclospsorine in plasma from patients with various transplants: HPLC and radioimmunoassay with a specific monoclonal antibody compared. Clin Chem 1989; 35: 1726-30. 22. Wallemacq PE, Lee SC, Lhoest G, Hassoun A. Crossreactivity of cyclosporine metabolites in two different radioimmunoassays using the specific monoclonal antibody. Clin Chem 1990; 36: 385. 23. Rosano TG, Freed M.A, Lempert N. Cyclosporine metabolites in human blood and renal tissue. Transplant Proc 1986; 18: 35-40.
CLINICAL BIOCHEMISTRY,VOLUME 24, FEBRUARY 1991
47
Outliers Includeda Outliers Deleted~ Whole Blood Measurement b Center-sample¢ Run to run d Center to centere Total Serum Measurement ab Center-sample ~ Run to run d Center to centere Total
(0.074)2 (0.032)2 (0.051) 2 (0.037)2 (0.102) 2
(0.067) 2 (0.033)2 (0.055)2 (0.037)2 (0.099)2
(0.086)2 (0.033) 2 (0.052)2
(0.069)2 (0.014) 2 (0.054)2
(0.077) 2
(0.078) 2
(0.130)2
(0.118)2
~ h e squared numbers represent the variances in the nested analysis of variance using the log transformed cyclosporine values in ng/mL. Numbers inside parentheses can be transformed into approximate CVs (%) by multiplying them by 100.
b'c'd'eThese denote the variance due to measurement such as pipetting and counting, b center-sample interactions (relative variance between centers specific to cyclosporine concentrations of samples), c run-to-run, d and center-to-center,e HPLC methods for cyclosporine A q u a n t i t a t i o n in use (26). Man y of these methods have been used without adequate performance validation with pat i e n t samples. Several reports where investigators carefully validated their methods demonstrated an excellent comparison between the CYCLO-Trac ® SP assay and HPLC or the S a n d i m m u n e ® assay (27-29). In summary, CYCLO-Trac ® SP offers an excellent method for the specific m e a s u r e m e n t of cyclosporine A in whole blood, serum, and plasma specimens from t r a n s p l a n t patients on cyclosporine A therapy. This report and others (27-29) support the utility of CYCLO-Trac® SP in a clinical setting. References
LEE, BRUDZINSKI, Y A S M I N E H , ZT AL.
24. Lensmeyer GL, Wiebe DA, Carlson IH. Distribution of cyclosporine A metabolites among plasma and cells in whole blood: effect of temperature, hematocrit, and metabolite concentration. Clin Chem 1989; 35: 56-63. 25. Johnston A, Marsden JT, Holt DW. The continuing need for quality assessment of cyclosporine measurement. Clin Chem 1989; 35: 1309-12. 26. Vine W, Bowers LD. Cyclosporine: structure, pharmacokinetics, and therapeutic drug monitoring. CRC Crit Rev Clin Lab Sci 1987; 25: 275-311. 27. Wolf BA, Daft MC, Koenig JW, Flye MW, Turk JW,
48
Scott MG. Measurement of cyclosporine concentrations in whole blood: HPLC and radioimmunoassay with a specific monoclonal antibody in SH- or 12sIlabeled ligand compared. Clin Chem 1989; 35: 120--4. 28. Sgoutas DS, Hammarstrom M. Comparison of specific radioimmunoassays for cyclosporine. Transplantation 1989; 47: 668-70. 29. Plebani M, Masiero M, Paleari CD, Faggian D, Burlina A. Evaluation of the INCSTAR CYCLO-Trac SP kit for the determination of cyclosporine in blood. Clin Biochem 1989; 22: 345-7.
CLINICAL BIOCHEMISTRY,VOLUME 24, FEBRUARY 1991
0009-9120/91 $3.00 + .00 Copyright © 1991 The Canadian Society of Clinical Chemist&
Measurement of Cyclosporine A by a Specific Radioimmunoassay with a Monoclonal Antibody and 251 Tracer SUNG C. LEE, 1 ALICE M. BRUDZINSKI, 1 JOAN L. YASMINEH, 1 NANCY J. JOHANSON, 1 IRENE A. FERBER, 1 LORI K. MAAS, 1 P. Y. WONG, 2 PAUL A. KEOWN, 3 and JOHN W. ORF 1 I lNCSTAR Corporation, Stillwater, MN, USA, 2Toronto General Hospital, Toronto, Ontario, Canada, and 3British Columbia Transplant Society, Vancouver, BC, Canada We developed a sensitive radioimmunoassay (CYCLO-Trac ® SP) that specifically measures cyclosporine A in serum, plasma and whole blood of transplant patients. The specific monoclonal antibody was from Sandoz and the tracer was an 12Slderivative of cyclosporine C. The assay is performed at room temperature for 1 h followed by a 20 min centrifugation. The sensitivities of the assays are 2.6 ng/mL and 8.7 ng/mL for the serum/plasma assay and the whole blood assay, respectively. Within-run and between-run CVs for both types of assays using cyclosporine concentrations of 80 and 58 ng/ mL (serum) and 186 and 199 ng/mL (whole blood) were less than 5% and 9%, respectively. Averaged recovery of serum/plasma and whole blood assays at various levels ranged from 93% to 115%. Interferences by bilirubin, triglyceride, cholesterol, hemoglobin, OKT3, azathioprine, methylprednisolone and 20 other drugs were insignificant. Multicenter proficiency studies showed an excellent correlation between the CYCLO-Trac ® SP and the specific 3HSandimmune ® assay from Sandoz: whole blood assay (r = 0.998) and serum assay (r = 0.997).
KEY WORDS: cyclosporine A; drug monitoring; radioimmunoassay; transplantation; monoclonal antibody.
unclear how the immunosuppressive activity of some metabolites demonstrated in vitro (13) relate to patient care. For these reasons, the American Association of Clinical Chemistry task force on cyclosporine monitoring has recommended the specific measurement of the parent compound in whole blood specimens (14). Recently, a monoclonal antibody that binds specifically to the parent compound cyclosporine A was developed by Sandoz (15). The RIA using this specific monoclonal antibody (Sandimmune ®) closely matched HPLC results in samples from patients with kidney, heart, and lung transplants (16,17). However, the charcoal separation and the need to use liquid scintillation counting for the tritium labeled tracer made the assay inconvenient. We report here the development of an RIA using the same specific monoclonal antibody in the form of a preformed immune-complex and an z2sI-tracer.
Introduction
Materials a n d m e t h o d s ince the introduction of cyclosporine A by Sandoz Corporation (Basel, Switzerland) as an immunosuppressive drug in organ transplantation, the need to monitor the drug level in individual patients has been well recognized (1-3). Analytical methods using high-performance liquid chromatography (HPLC) (4-6) or radioimmunoassay (RIA) with polyclonal antibodies (7-9) have been developed. However, the two types of methods give very different values due to the difference in biotransformation among patients and high cross-reactivities of polyclonal antibodies with metabolites of cyclosporine A (10-11). Moreover, cyclosporine values vary according to the specimen matrix and the method of specimen preparation (12). It is still
S
Correspondence: Sung C. Lee, Ph.D., INCSTAR Corporation, 1990 Industrial Blvd., Stillwater, MN 55082, USA. Manuscript received May 21, 1990; revised July 9, 1990; accepted August 21, 1990. CLINICAL
BIOCHEMISTRY,
VOLUME
24, FEBRUARY
1991
MATERIALS 1 2 5 I labeled tracer was prepared by alkylating a derivative of cyclosporine C with radioiodohistamine. Cyclosporine A, cyclosporine C and the specific monoclonal antibody (#78-299) were from Sandoz Corporation (Basel, Switzerland). The USP reference standard of cyclosporine A (lot F, 98.9% purity) was from the United States Pharmacopeia, Rockville, MD, USA.
METHODS
Assay procedures Serum/plasma assays were performed by adding 100 ~L aliquots of the tracer (60,000 cpm) and i mL aliquots of Immunosep (preformed complex of the anticyclosporine monoclonal antibody, normal mouse sera and donkey antimouse antisera) to 25 ~L ali43
LEE, BRUDZINSKI, YASMINEH, ~r AL. TABLE 1 Precision of Cyclosporine A I m m u n o a s s a y U s i n g
CYCLO-Trac ® SP Without Outliers Deleted Whole Blood
Serum
10.7 (46)a 3.2 (186) 3.1 (625)
10.2 (29) 4.2 (80) 5.1 (247)
9.8 (81) 8.4 (199) 9.4 (547)
12.9 (28) 7.9 (58) 9.8 (226)
Within-run b
Low Medium High Between-run c Low Medium High
aMean values in ng/mL are shown in parentheses. bWithin-run CVs (%) were determined by running each specimen in 25 replicates as described in Methods. CBetween-run, within-center CVs (%) were the averages from the 5-center proficiency studies as described in Methods. quots of specimens. The mixture was incubated at room temperature for 1 h and centrifuged for 20 min at 1600 x g. B/Bo vs. log cyclosporine concentration in a spline fit by RIACALC software package (LKB, Bromma, Sweden) was used for data reduction. Whole blood assays were identical to the serum/plasma assay except that extraction of specimens and standards with methanol was carried out prior to RIA. The extraction involved adding 100 ~LL of whole blood specimen directly into 400 ~L of methanol with immediate mixing. The mixture was centrifuged for 5 min and 50 ~LL of the supernatant was assayed in the ensuing RIA. The 3HSandimmune ® assay from Sandoz was run according to their published protocol.
Preparation of standards Standards of cyclosporine A were prepared gravimetrically using a pool of processed h u m a n serum and calibrated by CYCLO-Trac ® SP assays to U S P reference material that had been made gravimetrically in h u m a n whole blood.
five centers: Vancouver General Hospital (Vancouver, BC, Canada), INCSTAR Corp. (Stillwater, MN, USA), St. Lukes Episcopal Hospital (Houston, TX, USA), Albany Medical Center (Albany, NY, USA), and Toronto General Hospital (Toronto, Ontario, Canada). Thirty-five whole blood specimens (ranging from 36 ng/mL to 977 ng/mL) and 35 serum specimens (ranging from 18 ng/mL to 512 ng/mL) were analyzed in 5 different assays by each of the above 5 centers. For the calculation of between-run, within-center CVs, 35 specimens were grouped into low, medium, and high concentration specimens based on their B/Bos and the results were averaged. Statistical analyses of the data were performed with Statistix H (NH Analytical Software, Roseville, MN, USA) and included a nested analysis of variance with and without outliers deleted (18).
Recovery and dilution parallelism Recovery studies were performed by spiking 5 whole blood patient specimens, 5 undosed serum specimens, and 5 undosed plasma specimens with three different levels (low, medium, and high) of cyclosporine A that had been determined by gravimetric procedures. Dilution parallelism studies were performed by diluting 30 specimens each of whole blood, serum, and plasma at 1:2, 1:4, 1:8, and 1:16 with the zero standard in the kit according to the recommendations in the kit protocol. Whole blood specimens were extracted with methanol at 1:9, 1:19, 1:39, and 1:79. Results REACTION CHARACTERISTICS
Standards prepared using undosed plasma and whole blood showed superimposing suppression lines when extracted with methanol prior to RIA (Figure 1). The reaction kinetics reached a plateau in both the whole blood and serum/plasma assays in 1 h at room temperature (data not shown). The sensitivities of the assays were 2.6 ng/mL and 8.7 ng/mL for the serum/plasma and the whole blood assay, respectively.
Sensitivity and precision PRECISION
Sensitivity was determined by running 20 replicates of the zero standard and the lowest concentration standard in the kit in each type of assay three times. The sensitivity was calculated by taking the value at 3 SD less than the mean counts for the zero standard. Within-run CVs were obtained by running 3 different levels of cyclosporine A in whole blood and serum matrices in 25 replicates. Low, medium, and high value specimens of about 70%, 50%, and 25% B/Bo were used for within-run CV determinations. Between-run CVs were obtained from a multicenter study conducted at the following
44
Precision profiles for within-run CVs and between-run, within-center CVs are shown in Table 1. Within-run CVs and between-run, within-center CVs for both types of assay at medium concentrations were less than 5% and 9%, respectively. The CVs for low and high values of cyclosporine A were in general greater than those observed for the medium range. The observed between-run, within-center CV of 12.9% in the ser,,m assay for the low concentration range was primarily the result of an unusually high 25.2% CV from one of the centers in the
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
MEASUREMENT
OF C Y C L O S P O R I N E A B Y R A D I O I M M U N O A S S A Y
90. 80-
706050-
Z(BIBo) 403020100 10
I 1O0
I 1000
I 1000
Cyclosporine A (ng/mL) Figure 1--Standard curves in the CYCLO-Trac® SP whole blood assay. Cyclosporine A standards prepared in undosed plasma (dashed line) and in undosed whole blood (solid line) were compared in the whole blood assay. Each standard was extracted with methanol by adding the sample into the methanol according to the kit protocol. (B/Bo) % is the percentage of bound cpm divided by cpm of the zero standard. study. When this CV was excluded, the average between-run CV was 9.8%, conforming to the general pattern observed for the rest of the data. The results of the nested analyses of variance of both serum/plasma and whole blood assays in the 5-center studies are shown in Table 2. Overall CVs were: 9.9% with outliers deleted (10.2% without outliers deleted) for the whole blood assay and 11.8% with outliers deleted (13.0% without outliers deleted) for the serum/plasma assay. I ~ T H O D COMPARISON
Comparisons of mean values by sH-Sandimmune ® and CYCLO-Trac ® SP assays from the 5center study are shown in Figures 2(a) and 2(b).
Recovery and dilution parallelism The recoveries ranged from 100-106%, 105-111%, and 93-115% for whole blood, serum, and plasma specimens, respectively. The overall averaged recoveries were 103%, 108%, and 104% for whole blood, serum, and plasma, respectively. The significance levels of the difference between the mean recovery percentages and 100% (perfect recovery) were p = 0.13, p < 0.01, and p = 0.50 for whole blood, serum, and plasma, respectively. Studies of dilution parallelism of whole blood, serum, and plasma specimens yielded values of the diluted specimens close to those expected.
Interference Interference by the following compounds was insignificant in both serum/plasma and whole blood
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
assays (the maximum concentrations of several levels tested are shown in parentheses): bilirubin (0.51 retool/L), triglyceride (13 retool/L), cholesterol (25.8 mmol/L), hemoglobin (0.078 mmol/L, serum/plasma assay only), OKT-3 (10 nmol/L), azathioprine (200 ng/mL), 6-methylprednisolone (60 ng/mL), cimetidine (1000 ng/mL). The following drugs tested at concentrations up to 100 tLg/mL also did not interfere in the assays: acetaminophen, cephalosporine, digoxin, ethosuximide, kanamycin, metoclopramide, penicillin, rifampin, tobramycin, N-nifedipine, Nacetylprocainamide, chloramphenicol, disopyramide, gentamycin sulfate, lidocaine, naficillin, propranolol, salicyclic acid, and verapamil. Discussion
Knepil and McPhillips (19) claimed 6% to 10% negative bias when comparing CYCLO-Trac ® SP standards to their in-house standards. Our studies showing a good agreement between the Sandoz method and CYCLO-Trac® SP method did not indicate 6 to 10% negative bias in CYCLO-Trac ® SP standards compared to Sandoz standards. Copeland and Yatscoff (20) claimed a positive bias of 35% compared to the Sandoz assay and inaccurate standardization in the CYCLO-Trac ® SP whole blood kit. However, after carefully following the kit procedures for data reduction, they found only 10 to 15% difference between the two methods (R. Yatscoff, personal communication, August 1990). McBride et al. (21) claimed that standards in the CYCLO-Trac ® SP plasma assay caused a positive bias of 23% when compared with their in-house standards. However, even when using their own
45
LEE, BRUDZINSKI, YASMINEH, ST AL.
(a) 1000. t
900.
_J
R= 0.998
E ¢-
J
SLOPE= 1.02
800.
/. ---
D
700.
v
0. b")
600500-
L,.
I--
I o
300-
._.J fJ >-. 200(.~ 1000
I
0
1O0
I 200
I 300
I 400
I 500
I 600
I 700
I 800
I 900
1ooo
S a n d i m r n u n e (ng/mL) (b) 550-
"E
500
R=
450
SLOPE= 0 . 9 9 3
0.997
•
..E. = 350, 13-
(/)
soo
(,,I C) s-
250. 200.
I
150
._J
loo.
t'--
0
>(_)
50. 0 0
l 50
I 100
I 150
I 200
I 250
I 300
I 350
I 400
I 450
I 500
Sondimmune (ng/mL) Figure 2--Comparison of mean values of 35 specimens by SancSmmune ® and CYCLO-Trac ® SP assays. The specimens were analyzed by the above two methods in five independent runs in each of the five centers as described in Methods. (a) Whole blood specimens; (b) Serum specimens. in-house standards, they still observed widely variable discrepancies between values for CYCLO-Trac ® SP and HPLC. They postulated the presence of interfering substances, such as metabolites crossreacting with the specific monoclonal antibody in the CYCLO-Trac ® SP assay. As noted in this report, there are no interfering substances which have been shown to significantly affect the CYCLO-Trac ® SP assays. The cross-reactivity of major metabolites in the CYCLO-Trac ® SP assays has been evaluated (22). Based on expected levels of cyclosporine metabolites and cyclosporine A in clinical specimens
46
(23,24), less than a 5% increase in cyclosporine levels can be attributed to the presence of metabolites. Johnston et al. (25) have compiled values of their Q.C. survey from many laboratories in Europe between February 1988 and August 1988. Total analytical error of the CYCLO-Trac ® SP whole blood assay was within 6% (between 100 ng/mL and 1000 ng/mL) and the values of patient pools were about 25% higher than those by HPLC. The CVs for HPLC methods are much greater than those for immunoassays, consistent with the widely different
CLINICAL BIOCHEMISTRY, VOLUME 24, FEBRUARY 1991
MEASUREMENT
OF CYCLOSPORINE
TABLE2 Nested Analysis of Variance of CYCLO-Trac ® SP in the 5-center Proficiency Studies
A BY RADIOIMMUNOASSAY
1. Bowers LD, Canafax DM. Cyclosporine: experience with therapeutic monitoring. Ther Drug Monit 1984; 6: 142-7. 2. Kahan BD, ed. Cyclosporine-associated renal injury. Transplant Proc 1985; 17 (Suppl. 1). 3. Ptachcinsky RJ, Vankataramanan R, Burckart GJ. Clinical pharmacokinetics of cyclosporine. Clin Pharmacokinet 1986; 11: 107-32. 4. Carruthers SG, Freeman DJ, Koegler JC, et al. Simplified liquid-chromatographic analysis for cyclosporine A, and comparison with radioimmunoassay. Clin Chem 1983; 29: 180--4. 5. Moyer TP, Johnson P, Faynor SM, et al. Cyclosporine: a review of drug monitoring problems and presentation of a simple, accurate liquid chromatographic procedure that solves these problems. Clin Biochem
1986; 19: 83-9. 6. Annesley T, Matz K, Balogh L, Clayton L, Giacherio D. Liquid chromatographic analysis of cyclosporine utilizing a microbore column and reduced specimen requirement. Clin Chem 1986; 32: 1407-9. 7. Donatsch P, Abisch E, Homberger M, Traber K, Trapp M, Voges R. A radioimmunoassay to measure cyclosporine A in plasma and serum samples. J Immunoassay 1981; 2: 19-32. 8. Wong PY, Cheung M, Yip TK, Mee AV. Use of 12SI-labeled-histamine-cyclosporine C for monitoring serum cyclosporine concentrations in transplantation patients. Clin Chem 1986; 32: 492-5. 9. Mahoney WC, Orf JW. Derivatives of cyclosporine compatible with antibody-based assays: 1. The generation of 12SI-labeled cyclosporine. Clin Chem 1985; 31: 459-62. 10. Robinson WT, Schran HF, Barry EP. Methods to measure cyclosporine l e v e l s - high-pressure liquid chromatography, radioimmunoassay and correlation. Transplant Proc 1983; 15 (Suppl. 1/2): 2403-8. 11. Blyden GT, Franklin C, Cho SI, Kaplan MM, Hirsch CA, Greenblatt DJ. Cyclosporine blood concentrations determined by specific versus nonspecific assay methods. J Clin Pharmacol 1986; 26:367-71. 12. Yatscoff RW, Rush DN, Jeffery Jr. Effects of sample preparation on concentrations of cyclosporine A measured in plasma. Clin Chem 1984; 30: 1812-4. 13. Freed BM, Rosano TG, Lempert N. In vitro immunosuppressive properties of cyclosporine metabolites. Transplantation 1987; 43: 123-7. 14. Shaw WL, Bowers L, Demers L, et al. Critical issues in cyclosporine monitoring: report of the task force on cyclosporine monitoring. Clin Chem 1987; 33: 126988. 15. Quesniaux V, Tees R, Schreier MH, Maurer G, Regenmortel MHV. Potential of monoclonal antibodies to improve therapeutic monitoring of cyclosporine. Clin Chem 1987; 33:32-7. 16. Schran HF, Rosano TG, Hassell AE, Pell MA. Determination of cyclosporine concentrations with monoclonal antibodies. Clin Chem 1987; 33: 2225-9. 17. Ball PE, Munzer H, Keller HP, Abisch E, Rosenthaler J. Specific 3H radioimmunoassay with a monoclonal antibody for monitoring cyclosporine in blood. Clin Chem 1988; 34: 357-60. 18. Hawkins DM. Outliers from linear models. In: Hawkins DM, ed. Identification of outliers. Pp. 85-103. London: Chapman and Hall, 1980. 19. Knepil J, McPhillips M. Calibration of ~2sI immunoassay measuring cyclosporine A. Clin Chem 1989; 35: 181-2. 20. Copeland KR, Yatscoff RW. Calibration problems with the 12sI immunoassay for measuring cyclosporine A. Clin Chem 1989; 35: 1551-2. 21. McBride JH, Rodgerson SS, Reyes AF. Measurement of cyclospsorine in plasma from patients with various transplants: HPLC and radioimmunoassay with a specific monoclonal antibody compared. Clin Chem 1989; 35: 1726-30. 22. Wallemacq PE, Lee SC, Lhoest G, Hassoun A. Crossreactivity of cyclosporine metabolites in two different radioimmunoassays using the specific monoclonal antibody. Clin Chem 1990; 36: 385. 23. Rosano TG, Freed M.A, Lempert N. Cyclosporine metabolites in human blood and renal tissue. Transplant Proc 1986; 18: 35-40.
CLINICAL BIOCHEMISTRY,VOLUME 24, FEBRUARY 1991
47
Outliers Includeda Outliers Deleted~ Whole Blood Measurement b Center-sample¢ Run to run d Center to centere Total Serum Measurement ab Center-sample ~ Run to run d Center to centere Total
(0.074)2 (0.032)2 (0.051) 2 (0.037)2 (0.102) 2
(0.067) 2 (0.033)2 (0.055)2 (0.037)2 (0.099)2
(0.086)2 (0.033) 2 (0.052)2
(0.069)2 (0.014) 2 (0.054)2
(0.077) 2
(0.078) 2
(0.130)2
(0.118)2
~ h e squared numbers represent the variances in the nested analysis of variance using the log transformed cyclosporine values in ng/mL. Numbers inside parentheses can be transformed into approximate CVs (%) by multiplying them by 100.
b'c'd'eThese denote the variance due to measurement such as pipetting and counting, b center-sample interactions (relative variance between centers specific to cyclosporine concentrations of samples), c run-to-run, d and center-to-center,e HPLC methods for cyclosporine A q u a n t i t a t i o n in use (26). Man y of these methods have been used without adequate performance validation with pat i e n t samples. Several reports where investigators carefully validated their methods demonstrated an excellent comparison between the CYCLO-Trac ® SP assay and HPLC or the S a n d i m m u n e ® assay (27-29). In summary, CYCLO-Trac ® SP offers an excellent method for the specific m e a s u r e m e n t of cyclosporine A in whole blood, serum, and plasma specimens from t r a n s p l a n t patients on cyclosporine A therapy. This report and others (27-29) support the utility of CYCLO-Trac® SP in a clinical setting. References
LEE, BRUDZINSKI, Y A S M I N E H , ZT AL.
24. Lensmeyer GL, Wiebe DA, Carlson IH. Distribution of cyclosporine A metabolites among plasma and cells in whole blood: effect of temperature, hematocrit, and metabolite concentration. Clin Chem 1989; 35: 56-63. 25. Johnston A, Marsden JT, Holt DW. The continuing need for quality assessment of cyclosporine measurement. Clin Chem 1989; 35: 1309-12. 26. Vine W, Bowers LD. Cyclosporine: structure, pharmacokinetics, and therapeutic drug monitoring. CRC Crit Rev Clin Lab Sci 1987; 25: 275-311. 27. Wolf BA, Daft MC, Koenig JW, Flye MW, Turk JW,
48
Scott MG. Measurement of cyclosporine concentrations in whole blood: HPLC and radioimmunoassay with a specific monoclonal antibody in SH- or 12sIlabeled ligand compared. Clin Chem 1989; 35: 120--4. 28. Sgoutas DS, Hammarstrom M. Comparison of specific radioimmunoassays for cyclosporine. Transplantation 1989; 47: 668-70. 29. Plebani M, Masiero M, Paleari CD, Faggian D, Burlina A. Evaluation of the INCSTAR CYCLO-Trac SP kit for the determination of cyclosporine in blood. Clin Biochem 1989; 22: 345-7.
CLINICAL BIOCHEMISTRY,VOLUME 24, FEBRUARY 1991
