Several communica-. Division of Clinical Chemistry, Vancouver General Hospital; and. Department of Pathology, University of British Columbia, Vancouver,.
use of detergent I instead of detergent III (2) permits the omission of enzymic pretreatment of hyperlipidemic serum. The correlation between results of apolipoprotein B determination by radial immunodiffusion and nephelometryin which detergent I is used was similar to that after the recently described procedure (2), as estimated from 255 serum samples: YRID = 0.79x + 0.200 (correlation coefficient, 0.91). Of these, 76 serum samples were from normolipemic subjects, and the others were from subjects with various forms of hyperlipoproteinemia. Triglyceride concentrations in the samples ranged from 0.49 to 29.5 g/L, cholesterol ranged from 150 to 9450 mgfL. Twenty-nine serum samples were ultracentrifuged before the analysis for apolipoprotein B in whole serum and in the subnatant fraction. The data correlated satisfactorily,
with almost identical lines of regression for the two series (Figure 4). In our experience the simplified procedure seems to be more suitable than the preincubation method for clinical and screening purposes because the cost for an assay is con-
siderably less and handling or storing of enzymes is avoided. This procedure can be also applied to nephelometric determination of other apolipoproteins.
We appreciate the skilled assistance of Mrs. I. Erbe and Mr. H. Meyer. We were encouraged by the scientific support of Dr. Schafer. The study was supported in part by SFB9O/K6 of the German Research Council.
References 1. Heuck, C. C., Middelhoff, G., and Schlierf, G., Improved direct determination of serum cholesterol in low-density lipoproteins with use of polycations. Clin. Che,n. 23, 1756-1759 (1977). 2. Heuck, C. C., and Schlierf, G., Nephelometric determination of apolipoprotein B in human serum. Clin. Chem. 25, 221-226 (1979). 3. Havel, R. J., Eder, H. A., and Bragdon, J. H., The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum. J. Clin. Invest. 34, 1345-1353 (1955).
CLIN. CHEM. 25/5, 785-787 (1979)
Evaluation of Automated Enzyme Immunoassaysfor Five Anticonvulsantsand Theophylline Adapted to a Centrifugal Analyzer Nadine Urquhart,1 William Godolphin,2 and Donald J. Campbell
We report a clinical evaluation of the enzyme immunoassay (EMIT) performed with the GEMSAEC centrifugal analyzer as compared to gas-liquid and liquid chromatography for anticonvulsant drugs and theophylline, respectively. A good correlation was obtained for au drugs,
although some difficulties were experienced with one lot of reagent for ethosuximide. The analyzer has an economic advantage
if many samples
are being analyzed
for few
drugs in each sample. Additional Keyphrases: monitoring therapy of laboratory operation pared
. economics gas- and liquid-chromatography com-
The “Enzyme Multiplied Immunoassay Technique” (EMIT’; Syva Corp., Palo Alto, CA 94304) is widely used for monitoring therapeutic drugs in serum. Several communica-
Division of Clinical Chemistry, Vancouver General Hospital; and Department of Pathology, University of British Columbia, Vancouver,
B.C., Canada. Present address: Department of Laboratory Medicine, Shaughnessy Hospital, Vancouver, B.C. 2 Address correspondence to this author at the Department of Pathology, Vancouver General Hospital, Vancouver, B.C., Canada V5Z 1M9. Received Nov. 6, 1978; accepted Feb. 6, 1979.
tions have dealt with adaptation of the EMIT reagents to various centrifugal analyzers for the determination of phenytoin, phenobarbital, and has been reported to be in
with other
correlations
theophylline (1-7), and precision the 4-15% range, with acceptable
methods.
We report here a clinical evaluation in which we used EMIT with the GEMSAEC centrifugal analyzer (Electro-Nucleonics, Inc., Fairfield, NJ 07006) as proposed by Syva Corp. for phenobarbital, phenytoin, primidone, carbamazepine, ethosuximide, and theophylline. Results for unselected routine patient specimens were compared with thoseobtainedby columnchromatographic
Materials
methods.
and Methods
Reagents EMIT
reagents
6B019-G02,
used
with
phenobarbital
the GEMSAEC
6D019-GO2B,
(phenytoin primidone
6C1 19-GO2B, carbamazepine 6F1 19-GO4A, ethosuximide 6E1 19-G03, and theophylline 6P019-GO3C) were kindly supplied by Syva Corp. An EMIT ethosuximide kit, Reagent A 6E118-H01 and Reagent B 6E148-H01, containing an improved antibody was also tested. The commercially available antiepileptic drug calibrators from Syva were used to calibrate the GEMSAEC and the gas-liquid chromatograph. Theophylline calibrators from Syva were used to calibrate the GEMSAEC. A standard containing 20mg of theophylline per CLINICAL CHEMISTRY,
Vol. 25, No. 5, 1979
785
Table 1.
Control8
Method
80
Between-Run Precision flb
Mean, mg/L
160
Pbenoborbitol
Phenytoin
CV, %
GIG
E E
22 37
27.8 31.4
4.1 7.0
E E
20 36
15.4 14.8
3.8 3.1
Icier,c Lpemic Hemolyzed
I
Phenobarbital
GEMSAEC
U 0
Phenytoin GEMSAEC GLC
Primidone GEMSAEC GLC
E E
22 37
11.8 11.6
4.5 6.1
E E
22 37
6.5 6.3
3.3 5.1
E E
22 33
74.0 76.1
6.7 8.3
7 10 61 48
7.8 34.2 19.7 15.0
3.8 8.1 4.1 5.7
a Control material used was: EMIT control (E), UTAK control (U), or controls prepared by us from serum pool (P). I
Determinations over about a month.
20
40
y.IO4
I
I
60
I
c-il
C,
I
80
40
80
Primidone
I
160
120
a Corbomozeplne/,,,,,,,,,,/”
y.098
cool
y#{149}094n#{149}04
V 0 320
/ 20
40
60
80
4
80
Morris Plains, NJ 07950).
12
/::
0 U
/
80
20
.o99
.
Is
.
Theophylhne
,
xl7
‘
,$s
litre of distilled water was used to calibrate the liquid chromatograph. The precision of various methods was determined by analyzing one or more of the following: EMIT aed control from Syva, two theophylline controls from Syva, a theophylline serum toxicology control (Utak Labs, Granada Hills, CA 91344), and a home-made control containing 20 mg of theophylline per litre of “Versatol Automated Lo” (General Di-
8
EIhoeuc,mde
y.l#{216}
c023
.C III
agnostics,
I
I,
E E P U
HPLC
I
a
Theophylline GEMSAEC
I
4
Ethosuximide GEMSAEC GLC
1’
In
Carbamazepine
GEMSAEC GLc
y.09503 4.
80
II.
160 Concentrolion
240
I
320 0 (nrg/I)
20
40
60
I
80
by Chro,natogrophy
Fig.1.Relationship between serum drug concentrations as determined by the EMIT method with the analyzer and by a chromatographic method (gas-liquid chromatography in the cases of phenytoin, phenobarbital, primidone, carbamazepine, and ethosuximide;
and liquid-chromatography for theophyluine) patients’ sera received for routine therapeutic drug
Specimens were unselected monitoring
Procedures Enzyme immunoassay. EMIT assays for the five anticonvulsant drugs and theophylline were performed with the centrifugal analyzer according to instructions outlined in the EMIT-aed Supplement approved for GEMSAEC use (8). Gas chromatography. The gas-chromatographic method consisted of a simple extraction and then chromatography without derivitization on a 1-rn column of SP-2510DA with a short pre-column of SP-2250DA (Supelco Inc., Bellefonte, PA 16823) (9). We used a Model HP-5830A gas chromatograph (Hewlett-Packard, Avondale, PA 19211) equipped with flame ionization detectors and an automatic sample injector. Liquid chromatography. A Model 1084A high-pressure liquid chromatograph with a Model 1030B variable-wavelength ultraviolet detector (Hewlett-Packard) was equipped with a 0.25 m X 4.6 mm (i.d.) reversed-phase column (LiChrosorb RP-8, 10 Am; Browndale Labs, Berkeley, CA 94710). The method for theophylline was based on minor modifications of published assay procedures (10, 11). The internal standard solution contained 20 mg of 7-(/3-hydroxyethyl)theophylline (Aldrich Chemical Co. Inc., Milwaukee, WI 53233) per litre of acetonitrile (HPLC grade; Fisher Scientific, Pittsburgh, PA 15219) in water (1/3 by vol). Serum specimens were vortex-mixed with equal volumes of internal standard solution, centrifuged for 2 mm at 15 000 X g (Eppendorf Microcentrifuge; Brinkmann Instruments, Rexdale, Ont. M9W 4Y5), and 20 AL of the clear supernate was injected onto the liquid chromatograph. The column was eluted with a mixture
786 CLINICALCHEMISTRY,Vol.
25,
No. 5,
1979
of water, acetonitrile, and acetic acid (950/50/0.2 by vol) at a column temperature of 45 #{176}C. Quantitation was by peak-area ratios of the peaks for theophylline and internal standard, measured at 275 nm.
Results and Discussion Between-run precision with the analyzer was clinically satisfactory and tended to be slightly better than that obtained by chromatography (Table 1). For patients’ sera, the correlation with chromatographic methods was excellent (Table 2 and Figure 1). An earlier batch of EMIT reagents for ethosuximide yielded much more scatter and a poorer correlation (Table 2, EMIT reagent A 6E119-G03). This was noted by Syva, who subsequently supplied an improved antibody (A 6E118-H01) with which precision and correlation were improved. Moderate lipemia, hemolysis, and icterus appeared not to alter the results significantly (Figure 1). Caffeine, dyphylline, 8-chlorotheophylline (20 mg/L, in Auto-Lo), or uric acid (100 to 130 mg/L) did not yield any measurable reaction with the EMIT theophylline assay. Interestingly, 3-isobutyl-1methylxanthine, sometimes used as an internal standard in chromatographic theophylline assays, was measured as identical to theophylline by the EMIT method. The EMIT assays certainly have some advantages over chromatographic techniques: a generally smaller sample is required and the analyses are fast. On the other hand, a smaller workload combined with the need to analyze several
Table 2. Comparison of EMIT vs. Chromatographic Methods for Analysis of Patients’ Sera
twiceas much assubsequentones per reportable patient result. Chromatographic methods are technically more de-
SO of duplicates for EMIT, mg/L8
tion of the analyzer. About 60% of the cost of operating the chromatographs was salary, as compared to about 20% for the analyzer. Thus we have found it most convenient to perform our high-volume analyses (phenytoin and phenobarbital) by EMIT on the analyzer and the other drugs by chromatography. This has the advantage that both techniques are kept active in the laboratory and, in case of reagent shortage or equipment malfunction, back-up is promptly available. The simultaneous multiple-drug chromatographic assay
ChromatoNo. patIents’ samples
graphic
EMIT
mean
mean
r2
mglL
Phenobarbital, 36 Phenytoin, 41 Primidone, 40 Carbamazepine, 40 Ethosuximide (EMIT reagent A 6E119G03), 40 Ethosuximide(EMIT reagentA 6E1 18HOl), 36 Theophylline, 76 8
30.0 18.6 10.4 5.5 79.6
30.0 18.0 10.3 5.5 72.3
0.990 0.995 0.990 0.976 0.958
± 1.5 ±0.6
±0.3 ±0.2
±5.4
manding
and require
a higher order of skill than does opera-
permits one to warn the physician of apparently unsuspected interactions (e.g., phenytoin metabolism altered by the ad-
59.0
56.7
0.98
±3.6
13.4
13.7
0.976
±0.4
______________
Standarddeviation of duplicates =
± V’(difference)2/(2n).
drugs in the same specimen (in the case of anticonvulsant polypharmacy) may give chromatography an economic advantage. A detailed tabulation of all costs-reagents, control and standard materials, including volume discounts, repeated assays due to mishaps or their being “out of control,” and technologist and professional time-was prepared for the alternative methods (EMIT by GEMSAEC vs. gas- or liquid chromatography), based on the analysis of approximately 7500 patients’ specimens for anticonvulsants and 1500 for theophylline per year. Capital costs, amortization, and major servicing of the instruments and laboratory overhead were not included. This analysis yielded a cost per reportable patient result of $6.10/anticonvulsant by EMIT vs. $4.85 by gas chromatography and $5.50 per theophylline by EMIT vs. $6.60 by liquid chromatography. Interpretation of these data must be tempered by the following facts and observations. Analyses performed by either gas-liquid or liquid chromatography were done singly, whereas those done on the analyzer were done in duplicate. If the anticonvulsants are analyzed singly by EM IT, the cost would drop to $4.50 per reportable patient result, and we believe the performance of only single assays, rather than duplicate assays, can probably be justified. For example, of 1000 consecutive specimens assayed for phenobarbital and (or) phenytoin, we found only 13 in which the difference between duplicates was unacceptable (0.006 A), and six of these were controls. In only one of the patients’ specimens would the clinical interpretation have been altered. An auto-sampler was used on the gas chromatograph, which allowed it to perform
analyses unattended. The liquid-chromatography samples were injected manually and hence required the nearly constant attention of a technologist. The workload required a dedicated gas chromatograph (i.e., no other analyses were performed with it), whereas the analyzer was used to perform several other sorts of routine clinical assays every day. The less frequently requested drug assays, such as ethosuximide and carbamazepine, are not run very efficiently on the analyzer, becausea complete standardization is performed (consuming reagent for nine analyses: two blanks, five calibrators, and the control in duplicate) before assay of even a relatively few patients’ specimens. Thus the costs for the first rotor load are
dition of phenobarbital) and metabolites (e.g., phenobarbital from primidone) (12). This advantage is lost when the EMIT method is used and only those drug analyses specifically requested by the physician are performed. The importance of this factor, of course, greatly diminishes if the treating physician is sufficiently aware of the pharmacology of the drug in question or if multiple-drug therapy is little used.
The
EMIT
reagents used in this evaluation
were a gift from Syva
Corp. We acknowledge the assistance of the staff of the GLC-Toxicology Section of the Division of Clinical Chemistry, and we thank Lourdes Garcia, who performed most of the analyses on the GEMSAEC.
References I. Finley,
P.R., Williams, R. J., and Byers, J. M.,
Adaptation of “EMIT” 91 1-914 (1976).
to the centrifugal
Assay of phenytoin: analyzer. Gun. Chem. 22,
2. Brunk, S. D., Hadjiioannou, T. P., Hadjiioannou, S. I., and Malmstadt, H. V., Adaptation of “EMIT” technique for serum phenobarbital and diphenylhydantoin assays to the miniature centrifugal analyzer. Clin. Chem. 22, 905-907 (1976). 3. Long, J. P., Enzyme immunoassay, with use of a centrifugal analyzer, of phenytoin, phenobarbital and theophylline. Clin. Chem. 24, 391 (1978).
Letter.
4. Haven, M., Phenytoin and phenobarbital measurement trifugal analyzer. Clin. Chem. 22, 2057 (1976). Letter. 5. London,
M., Sanabria,
D., and Yau, D., Enzyme
phenytoin and phenobarbital Chern. 23, 1362 (1977). Letter.
with
the centrifugal
by cen-
immunoassay analyzer.
of
Clin.
6. Shirey, T. L., Phenobarbital and phenytoin determined by enzyme immunoassay with a centrifugal analyzer. Clin. Chem. 23,611(1977). Letter. 7. Henry, V., Deutsch, J., and Lurn, G., Enzyme immunoassay of theophylline with a centrifugal analyzer, and comparison with an ultraviolet method. Clin. Chem. 24, 514 (1978). Letter. 8. EMIT-aed Supplement. Application of EMIT aed assays to EN! GEMSAEC centrifugal fast analyzer. Syva Corp., Palo Alto, CA (May 1978). 9. Godolphin, W., and Thoma, J., Quantitation of anticonvulsant drugs in serum by gas-chromatography on the stationary phase SP-2510. Clin. Chem. 24, 483-485 (1978).
10. Adams, R. F., Vandemark, F. L., and Schmidt, G. J., More sensitive high-pressure liquid-chromatographic determination of theophylline in serum. Gun. Chem. 22, 1903-1906 (1976). 11. Orcutt, .J. J., Kozak, P. P., Gillman, S. A., and Cummins, L. H., Micro-scale method for theophylline in body fluids by reversed-phase, high-pressure liquid chromatography. Clin. (‘hem. 23, 599-601 (1977). 12. Pippenger,
C. E., Pellock, J. M., and Gold, A. P., Antiepileptic in children on multiple therapy. In Antiepileptic
drug concentrations
Monitoring, C. Gardner-Thorpe, D. Janz, H. Meinardi, E. Pippenger, Eds., Pitman Med. Pub. Co. Ltd., Tunbridge Kent, U.K., 1977, pp 282-286. Drug
CLINICAL
CHEMISTRY,
Vol. 25,
No. 5,
1979
and C. Wells,
787
