Clinica Chimica Acta, 99 (1979) 289-295 0 Elsevier/North-Holland Biomedical Press
A SIMPLE AUTOMATIC DIGOXIN
ALBERT0 CASTRO, and THOMAS NOT0
III, ALEX CID, PETER
Department of Pathology, Hormone Research Laboratory, Medicine, Miami, FL 33101 (U.S.A.) (Received
University of Miami School of
Introduction The digitalis glycoside digoxin is widely used in the treatment of chronic heart diseases. Several studies of the correlation between digoxin dosage, blood level, and clinical state have demonstrated that there is a narrow margin between therapeutic and toxic levels, and that rates of metabolism and excretion of the drug vary widely from patient to patient [l-4]. These combined effects make monitoring of serum levels during both initial digitalization and maintenance of great importance. The most commonly used technique for measurement of serum digoxin is radioimmunoassay [ 41. In 1975, an enzyme multiplied immunoassay technique (EMIT @) for digoxin was reported and has since been made commercially available [ 5,6]. As marketed, the method has several disadvantages. These include a great number of pipetting steps, crucial timing, and a relatively large sample volume of 200 ~1 per assay (of particular concern when dealing with pediatric or neonatal patients). Levy et al.  reported in abstract form an automated procedure for the enzyme multiplied immunoassay technique digoxin on the ABA-100 @. As described by Levy, the method has problems of precision. Moreover, the pretreatment of serum samples, an important part of the manufacturer’s protocol, was eliminated. We report the automation of the enzyme multiplied immunoassay technique digoxin system on the Abbott ABA-100 bichromatic analyzer with a technique which resolves the problems mentioned
Correspondence to: A. Castro, Ph.D., Department of Pathology Medicine, P.O. Box 016960. Miami, FL 33101, U.S.A.
R-40. University of Miami School of
above by utilizing a reaction system compatible with that proposed by Syva Corporation. The timing and most of the pipetting are performed by the instrument and the assay requires only 40 ~1 of serum, permitting duplicate assays from a capillary tube. Materials and methods Instrumentation (1) Abbott Bichromatic Analyzer-100 (ABA-loo). Instrument settings for the ABA were: mode: end point; reaction: up; decimal pt: 0.000; analysis time: 5 min; revolutions: M (infinite); temperature: 30°C; scale factor: 500 (machine prints out absorbance); filter: 340-380 nm; syringe plate: 1 : 26 dilution. Abbott Lab., South Pasadena, CA. (2) Syva Model Auxiliary dispenser with magnetic cable switch and dual probe assembly. Syva Corp., Palo Alto, CA 94304. (3) Micromedics Diluter Model with a 50 ~1 sampling pump and a 200 1_r1 dilution pump. Micromedic Systems, Horsham, PA 15041. Reagents and preparation (1) Materials for the digoxin enzyme immunoassay were purchased from Syva Corp., Palo Alto, CA 94304. The EMIT @ digoxin kit was used for both the manual and the automated assay. For the manual assay, all reagents were prepared as per manufacturer’s instructions. (2) Materials for the comparative RIA method IMMOPHASE12SI @ digoxin were purchased from Comin Medical, Medfield, MA 70549. (3) Two control sera with high and low values of digoxin were used for the precision study. (Assayed values of control sera reported for different RIA manufacturers were averaged and used. Value: Ortho I, 1.1 + 0.3, Ortho II, 3.1 f 0.6.) Ortho I and II RIA control sera were used, and were purchased from Ortho Diagnostics, Raritan, NJ 64312. The EMIT @ digoxin procedure was automa~d in an ABA-100 with a second reagent auxiliary dispenser. Reagent and sample volumes were cut to one-fifth of that required for the manual assay without altering the sample-to-reagent volume ratios. The buffer concentrations in the different reagents were changed so that the resulting reaction mixture had the same buffer concentration and pH as that of the manual assay. For the automated assay, all lyophilized reagents were reconstituted as indicated in the manufacturer’s package insert. Two buffers were prepared for use in the assay by diluting the buffer concentrate supplied with the kit. Buffer No. 1 is a 0.2475 mol/l Tris-HCl, pH 7.4, made by diluting 7 ml of the buffer concentrate with distiIIed water to a voIume of 23.3 mf. Buffer No. 2 is a 0.035 mol/l Tris-HCl, pH 7.4, made by diluting 7 ml of the buffer concentrate to 165 ml with distilled water. The serum pretreatment reagent (0.5 mol/l NaOH) was diluted 1 : 6 with distilled water, and 60 ~1 of the resulting mixture was used for pretreating each sample. The manual working reagent A was replaced with 1 : 5 dilution of reagent A with buffer No. 1. One ml of reagent A as the new working reagent (total
volume = 5 ml) was sufficient to run 3 X 32 well multicuvettes, allowing for purging of the lines. The working reagent was allowed to equilibrate for 8 h at room temperature before use and was found to be stable for up to 7 days after preparation when stored at room temperature. Due to the loss of linearity (over 5 pg/l) only calibrators 0 through 4 were used to construct the standard curve. The ratio and volumes of reactants for the manual and automated assays are as follows: Volume Component
Serum Reagent A Reagent B Pretreat.
4 1 1 1
Manual 200 50 50 50
/Ll /.&I /ll /.ll
Automated 40 10 10 10
@l /.ll /_ll /Jl
Assay protocol A 32-well multicuvette can accommodate 5 calibrators and 10 samples, run in duplicate (position No. 00 used to set zero absorbance and position No. 1 used as a reagent blank subtracted from all other absorbances). Pretreatment of serum samples was done directly in the multicuvette. Using the automatic dilutor, 40 ~1 of serum was diluted to 100 ~1 with the diluted pretreatment reagent. After all samples were placed in the multicuvette, it was mixed gently by hand and incubated for 5 min at room temperature. During this time the auxiliary dispenser was primed and set to dispense working reagent A. At the end of the 5 min, the multicuvet~ was placed in the machine water bath. The machine probe was not used in this run; therefore, only the auxiliary probe was in position. Using the auxiliary dispenser probe to dispense one well ahead of the regular machine probe position, the analyzer was allowed to run only the dispense cycle, stopping it before it dispensed into the last well. The multicuvette was then removed from the water bath, covered with an evaporation cover, gently mixed and let stand for a minimum of 15 min. After the end of this incubation, the multicuvette was returned to the water bath. Instrument settings were checked and zero absorbance set. The syringe plate probe was then used (auxiliary probe removed) in its normal position and the probe arm adjusted so it would sample reagent B from a glass mini-beaker set on the sensor post (the container for reagent B must be glass, due to the label’s affinity for plastics). Buffer No. 2 was used as the diluent in the syringe plate. The machine was set on run, and reagent B was dispensed into all but position No. 00. The first printout (5 min after the addition of reagent B) is used as Ainitial, the 4th printout is the 15 min final absorbance reading (Afinal 15), and the 7th printout is the 30 min final absorbance measurement (Afinal 30).
Comparison study One hundred clinical samples were used for the comparative study. These were random monitoring samples received for assaying by manual EMIT in our laboratory. All comparative methods were performed as per manufacturer’s suggested protocol. Results The results of the automated procedure were calculated at 15-min and 30min reaction time. The 30-min reaction results were calculated in the same manner as the manual procedure and plotted on the graph paper provided with the kit. For the 15-min assay, the A-A0 for calibrators and samples were calculated and multiplied by 2 and the resulting number graphed vs concentration on the kit graph paper. Concentration of digoxin in pg/l at 15 and 30 min was read from the graphs for those reaction times respectively. The precision of the automated assay was determined using values obtained for two commercial control sera. The within run precision was calculated using values from a cuvette containing only calibrators and 20 control replicates. The coefficient of variation for the low range control at 15 and 30 min reaction time was 3.1% and 3.3% respectively, and for the high range control at 15 and 30 min, respectively, 4.7% and 3.5%. The between run precision was calculated by using the average of duplicate values for both controls from 10 separate runs. The C.V.‘s for low and high range controls at 15 and 30 min were 3.1% and 2.6%, and 5.3% and 4.1%, respectively. As a measure of the run-to-run reproducibility, 10 samples were run on two consecutive days and a correlation was done. The results are shown in Fig. 7. The complete results of the precision study are shown in Table I. The correlation of the automated digoxin assay at both 15- and 30-min reaction time with the manual digoxin assay, a commercially available radio-
( g! hter)
Fig. 1. Comparison of results for automated EMIT assay using 15-min and 3%min reaction time for 100 clinical samples. Fig. 2. Comparison of results for the 15min ples.
automated assay with the manual EMIT for 100 clinical sam-
293 4y = 0.918 Y - 0099
;2A $ c E
Fig. 3. Comparison ples.
emit ( pg I liter)
of results for the 30-min
Fig. 4. Correlation
of the 15.min
assay with the manual EMIT for 100 clinical sam-
EMIT assay with RIA for 29 clinical samples.
4y : 0.927xr
( pgI Net- )
Radiolmmuncassay ( pg
Fig. 5. Correlation
of the 30-mln automated
Fig. 6. Correlation
of manual EMIT assay with RIA for 30 clinical samples.
y: 1060xr = 0.969 n= 10
EMIT assay with RIA for 29 clinical samples.
Fig. 7. Run-to-run
Rod~olmmunoassoy ( pg / Her)
of 10 clinical samples expressed
as a correlation.
Ortho I ABA 30 ABA 15’ Ortho
ABA 30’ ABA 15’
Ortho I ABA 30’ ABA 15’ Ortho
ABA 30’ ABA 15’
immunoassay method, and to each other, were calculated. Results are shown in Table II and in Figs. l-6. Discussion This automation of the enzyme multiplied immuno~say technique for digoxin provides several attractive advantages over the manual assay. The reduction in sample volume makes duplicate determinations possible with less than 100 ~1 of serum. The reduction in reagent volumes increases the test capacity of one kit five-fold, thereby reducing cost. Technologist work time is reduced and crucial timing steps are done by the instrument. Our data has shown that the assay system is compatible with either a 15- and 30-min reaction time. Shortening the reaction time has the advantage of reduced turn-around time. However, the 30-min assay shows slightly better precision. The comparison of the automated digoxin system with the RIA showed that, for both the 30- and 15-min assay, the correlation coefficients were higher than those of the manual method with the RIA. Other reports comparing the manual technique to RIA showed comparable or better correlation [S-10]. Rosenthal et al. [S] showed a correlation of 0.979 when the manual digoxin method was compared to RIA; however, this study included certain modifications to the manual procedure that are not easily adaptable for use in a routine clinical laboratory (i.e. recentrifuging all serum samples prior to assaying, de-gasing the reagent water used to reconstitute the lyophilized reagents, and filtering all reconstituted reagents through a Gelman cellulose acetate filter before using). The specimens used in this study were random routine clinical samples and were assayed regardless of their appearance, which ranged from clear normal appearance to extremes of icterus, hemolysis and lipemia.
The precision of the reported method was comparable to, or better than, that obtained by other investigators using the manual assay and RIA for digoxin [ 8-101. The ABA-100 @ is a commonly used instrument in the clinical chemistry laboratory. The adaptation of the EMIT @ digoxin system onto such an instrument is simple, eliminates sources of human error inherent in the manual assay, and provides an economical routine assay for digoxin. It combines discrete sampling and multi assay capabilities, it has small reagent requirements, and excellent reproducibility and precision. The method as presented is adaptable to emergency situations as well as to large volume workloads. References 1
J.J.. S. and
L. and R.H..
E. (1970) J. Am.