787

SENSITIVB RADIOIMMINOASSAY FOR DIGGXININ PLASMAAND URINE J. G. Wagner’, College

H. R. Hallmark,

E. Sakmar and J. W. Ayres*

of Pharmacy and Upjohn Center for Clinical The University of Michigan Ann Arbor, Michigan 48109

Pharmacology

2-8-77

Received :

ABSTRACT A reproducible and sensitive radioimmunoassay for digoxin in either Using 0.5 ml of serum or plasma, serum, plasma or urine is described. the assay sensitivity is 0.05 ng of digoxin/ml. The antiserum and tracer solutions employed are available in a kit sold in the United All other reagents were prepared in the laboratory. The assay States. allows measurement of digoxin in plasma or serum for 96 hours after single 0.5 mg doses of digoxin; this is necessary in human bioavailability studies to accurately estimate the total area under the digoxin concentration, time curve from zero to infinite time. In contrast, with the kit assay, employing 0.2 ml of plasma or serum, it has been reported that the 12 hr serum digoxin levels, after single 0.5 mg doses, are, in most subjects, below the sensitivity limit (about 0.5 ng/ml) of the assay. INTRODUCTION Attempts have been made to measure digoxin by gas-liquid involves

chromatography

seven steps

and is far

method based on digoxin cribed

(6-8)

and less

Butler

and Chen (10)

the antibodies

variable first

RIA kits

and developed

1

2

reported

but that digoxin, all

in our hands

(9);

as the tracer.

(11,12)

of digoxincharacterized

(RIA) for digoxin Subsequently,

It was later

in

digoxin

shown that anti-

antiserum were not specific

digoxigen-mono-digitoxoside

reacted

by the method.

on the preparation

a radioimmunoassay

in commercially-available

bis-digitoxoside

may be detected

and unreliable.

became commercially-available.

present

digoxin,

digoxin

Smith and associates

plasma or serum using ‘A-digoxin bodies

of red cell

for routine use. A 86 Rb uptake has been des-

assay has been described

Later,

antibodies.

too complicated

inhibition

assay was extremely

specific

The method of Watson and Kalmen (5)

than 1 ng of

A Na-K-ATPase displacement this

(l-5).

in human plasma or serum

with the antibody

for

and digoxigen-

to about the same degree

Send reprint requests to: Dr. John G. Wagner, Upjohn Center for Clinical Pharmacology, The University of Michigan, Ann Arbor, Michigan 48109 Present Oregon

address: 97331.

School

Volume 29, Number 6

of Pharmacy, Oregon State

S

-l?ElROXDS3

University,

Corvallis,

June,

2977

(13,141 L Digoxigiania also (13,14) * digoxin

Dihydrodigoxin in man (15),

with the antibody

cross-reacts,

was also

and this

sold

but to a much lesser

degree

shown to be a common metabolite

metabolite

was also

in commercially-available

of

shown to cross-react RIA kits

(16,17).

Gault -et al. (18) extracted digoxin and its metabolites from human urine and separaeed some of them by column chromatography with Sephadex I&l-20; they reported

that the highest

in one patient,

but that for

to 10X, and always occurred

percentage 5 other

patients

the possible

digitoxoside,

digoxigen-bie-digitoxoside,

presence

found was 22%

peak percentages

w&thin the first

They reported tives.

of metabolite day’s

of digoxin,

were 7

collections.

urine

digoxigen-mono-

digoxigenin

and dlhydro

deriva-

Subsequently,

Sugden -et al, (191, usLng columns containing diethylaminoethoxypropylated Sephadex IX-20, successfully separated tritiated

dihydrodigoxin

Although also

the digoxin

from tritiated RIA lacks

measures metabolites,

digoxin

bfoavailabflity

digoxin.

specificity

for dfgoxin

since

the method has been extensively studies

and in therapeutic

it

used in both

drug monitoring

of

serum digoxin concentrations during therapy. The original method (11,12) 3 usfng H-digoxin as thca tracer, has been employed by moat investigators, but has been modified by others (14,20-25). An assay, employing 125 I-labeled 3-0-succinyl digoxigenin Cyrosine, as the labeled antigen, has been reported clinically

(26) and kits,

RIA method (11,12)

to be sensitive

One modification,

to 0.2 &ml

laboratories,

have a sensitivity

limit

curves

of digoxin.

Recently,

bioavailability to obtain

were shown to yield With the kit

0.2 ml of

was r&ported

The method of Stoll -et al, (141, 0.5 ml of plasma, and was reported

area (0 tocc) under the concentretion,

truncated

utilizes

using 0.5 ml of plasma,

of 0.08 &ml..

in order

to about 0.5 ng of

procedure

ft

is necessary

for 96 hr after

an accurate time curves.

erroneous

assay,

Wagner and Ayres

studies

plasma or serum concentrations

OT5 mg doses of digoxin

bility

(25).

utilized

showed that tn human digoxin measure digoxin total

is sensitive

of plasma or serum, and the kit

plasma or serum. from these

method, have been used

(27-29).

The original digoxin/ml

based on this

estimatea

employing

to (30) to

single

estimate

of the

Areas from of bioavaila-

3H-digoxin

and 0.2 ml

S

.l?DOROID~

789

of plasma or serum, the 12 hr serum or plasma levelsafter single0.5 mg doses are below the sensitivitylimit of the assay. A sensitivitylimit of 0.2 ng/ml is still not sufficient,and a limit of 0.05 ng/ml is necessary in human digoxinbioavailability studies. Employingthe readily-available tracer solutionof 3U-digoxinand antiserumsolutionin a kit, which is cotmnercially-available in the United States,we have intensivelystudiedeach step in the digoxinRIA procedurein order to determineoptimumconditions. The resultantassay, coupledwith the new method of preparingcalibrationplots, gives good reproducibility and a sensitivitylimit of 0.05 nglml. It is shown that, at a concentration of 0.1 r&ml, the cornmanly-used logarithmic-logistic (logit-log)type of calibrationplot yields much largerstandarddeviations

for inversely-estimated concentrations, and, greaterbias across

the whole curve (O-5 &ml)

than the new calibrationmethodwhich is

based on fittingnormalizedpercentbound, concentrationvalues to a biexponentialequationwith a digitalcomputerand a suitablenonlinear estimationprogram. MATRRIAIS Antiserumsolutionand tracersolutionwere pooled from Digoxin Radioinauunoassay Kits [3H]3. Phosphate-saline buffer (pR 7.4 phosphatebufferedsaline solution)was preparedby dissolving1.392 g of dipotassiumphosphate,0.276 g of monosodiumdihydrogenphosphate monohydrateand 0.77 g of sodiumchloridein sufficientwater to make one liter of solution. Charcoalsuspensionwas preparedby suspending125 mg of dextran4and 5 g of charcoal5in 100 1 of phosp ate-salinebuffer. 2 . For the 30% The liquid scintillationfluidwas Unogel(8 Emulsifier ethanol-wate standardsolutionsof digoxin the digoxinwas weighedon a f . Pipettingwas performedas follows:plasmaand buffer Cahn balance were pipettedwith an EppendorfMicroliterPipetS;30% ethanol-water standardsolutions,antiserumsolutionand tracer solutionwere pipetted with Lang-LevyMicro Pipetsg* , charcoalsuspensionwas measuredby syringe; 3 Schwart-MannCatalogNo. 0750-23,Divisionof Becton,Dickinsoncind Company,Orangbury,New York. 4 DextranT-70, Pharmacia,Uppsala,Sweden. 5 Norit A, Sigma ChemicalCompany,St. Louis,Missouri. 6 Schwars-Mann(addressabove). 7 Cahn Divisionof VentronInstrumentsCorp., Paramount,California. 8 BrinkmanInstrumentsInc., CentiagueRoad, Westbury,New York. 9 Arthur Ii.Thomas Company, Philadelphia,Pennsylvania.

10 pl

TRROXDI

S

790

aliquotr

General

&pore

of

urine

wer;omeaeured by Eppendorf pipet; . The rcintillatlon counter . Centrifugation was perfo rd in Automatic Refrigerated Centrifuge

and unogel 8 vaa a Packard a Sorvall RC-3

MET?tODS I.

Determination of Optimum Standards an& Effect of Volume. Calibration curves were prepared according to the method of Stoll et al. (14) using known digoxin concentrations of 0.1, 0.5, 1, 2 and 5 z z digoxln per ml of plasma. On one day analyst S obtained atendard curve data using 80 pl of kit (serum) standard and 420 pl of plasma. In addition, standard curve data were obtained from 30% ethanol-water standards where the volume and concentration of the standard were: 1 pl for 0.1 &ml, 5 pl for 0.5 ng/ml, 10 pl for 1 ng/ml, 20 pl for 2 ng/ml and SO ~1 for S &ml, with SO0 pl of plasma. On another day analyst S obtained standard curve data using both kit standards and 30% ethanolwater standards, with the total volume added being 50 pl in each case with 500 pl of plasma. II.

Determination of the Rinetica of Association and Ovtimum Pre-Incubation Time. There was no ore-incubation of unlabeled digoxln with antiserum in Phillips (22) claimed that pre-lncubathe method of Stoli et al. (14). tion increased both Gnztivlty and reproducibility of the digoxin RIA. For each temperature (25OC and 37OC) 16 tubes were prepared, each containing 500 pl of plasma, 500 pl of phosphate-saline buffer, 50 pl of 30X ethanol-water, 100 pl of antiserum solution and 100 pl of tracer solution. The mixtures were vortexed. Then SO0 pl of charcoal auapenaion (room temperature) was added to each of two tubes at 0, 5, 10, 15, 20, 30, 60 and 120 min. The charcoal contact time war 10 min for each. Tubes were centrifuged at 2000 r.p.m. for 20 min. Each aupernatant was decanted into 15 ml of liquid scintillation fluid in a scintillation vial and counted for 10 min. Hence, a series of duplicate B(0) valuer were obtained for different reaction timas. Plots of countalmin versus time wre made for each incubation temperature. III.

ppgeminatioa of the Kinetics of Q$arociatlonof the DiuoxinAntibody Comlex as a Function ol Temperature. For each temperature (OOC, 2%! and 37OC) 14 tubes were prepared as indicated under II above, except that the 50 pl of 30X ethanol-water standard contafned digoxin equivalent to 1 ng digoxin/ml plasma. After the tracer solution was added, the tubes were incubated at the desired temperature for 30 min. Then, 50 pl of a 30% ethanol-water solution, containing the equivalent of 0.5 pi digoxFn/ml, war added to each tube to provide a SOO-fold excess of digoxin. Two tubes were each incubated for 0, 5, 10, 20, 30, 60 or 120 min for each of the temperatures. After the desired incubation time, 500 pl of charcoal suspension (room temperature) was added, and the charcoal contact time was 10 min. Tubes were 10 11 12

Lebindustries, Packard Ivan

1802

Instruments

Sorvall,

Inc.,

Second Co.,

Street,

Inc.,

Newtown,

Berkeley,

Downers Connecticut.

Grove,

California. Illinois.

S

%=PIIROIDS

791

centrifuged and the supernatants counted as indicated formerly under II above. IV. Determination of Necessary Incubation Time After Tracer Solution Added. meen tubes were prepared containing 500 nl of plasma, 500 pl of phosphate-saline buffer, 50 ul of ethanol-water standard (8 tubes equivalent to 0.1 ng digoxinlml plasma and 8 tubes equivalent to 5 ng digoxinfml plasma) and 100 pl of antiserum solution. All tubes were pre-incubated at 25'C for 60 min. Then, 100 pl of tracer solution was added to each tube. Two tubes for each of the two concentrations were incubatad for 10, 20, 30 and 60 min at 25Oc. Then, 500 pl of charcoal suspension (pre-cooled in an ice-bath) was added to each tube, and the charcoal contact time was 10 min under ice-bath conditions. Tubes were centrifuged and the supernatants counted as indicated under II above. V. Determination of Optimum Charcoal Exposure Time Under Ice-Bath Conditions. Ten tubes were prepared, each containing 500 pl of plasma, 500 pl of phosphate-saline buffer, 50 ul of 30% ethanol-water standard equivalent to 1 ng digoxin/ml plasma and 100 nl of antiserum solution. Tubes were pre-incubated for 60 min at 25'C. Then, 100 1.21 of tracer solution was added to each tube, and the tubes were incubated for 60 min at 25'C. At the end of the incubation time, the tubes were placed in an ice-bath, and, after constant temperature had been attained, 500 pl of charcoal suspension (pre-cooled in an ice-bath) was added. Two tubes were used for each charcoal exposure time of 1, 5, 10, 30 and 60 min. At the end of the appropriate time, the tubes were centrifuged and the supernatants counted as indicated under II above. VI. Determination of the Minimum Amount of Liquid Scintillation Fluid Necessary Eight tubes were prepared as under V above, except thf)tthere was no pre-incubation and the incubation time was 30 min at 25 C aftsr tracer was added. The charcoal exposure time was 10 min. After centrifugation at 2000 r.p.m. for 20 min, the supernatants of each of two tubes were added to 8, 10, 12 or 15 ml c. liquid scintillation fluid, then counted for 10 min after visual observation. VII. guench correction. Fresh whole blood was collected and froren in dry ice-acetone mixture, then thawed and frozen again. This provided hemolyeed blood from which the plasma was collected by centrifugation. Various mixtures were made of this red-colored plasma and plasma from non-hemolyaed blood. Two groups of duplicates of 19 different mixtures were made--one group corresponding to 110,000 DPM H O/50 pl, and the other corresponding to 11,000 DPM 31iO/50 1.11.All via1s also contained 50 nl of the PH 7.4 phosphate-bufIered saline solution, 100 pl of antiserum and 50 pl of 30% ethanol-water (as used in the assay procedure) and 15 ml of liquid scintillation fluid. Instrument settings were: Red channel, 040-100 (A-B), 64% gain; Green channel, 040-1000 (C-D), 74% gain; and Blue channel, 300-1000 (E-P), 1.59% gain. The counts/min for e ch sample B20 present, were divfded by the DPM known to be present,from standard !J to obtain the X efficiency, then this value was plotted versus the

counts/min

in

the

Blue

channel

obtained

from AES counting.

VIIX.

Anticoagulant Study. Presh blood from a subject was drawn into Vacutainers containing sodium citrate and disodium ethylenediaminetetraacetic sodium heparin, acid (Na2EDTA). The original Vacutainer containing sodium citrate contained this anticcagulant as an aqueous solution. To avoid dilution of the blood we evaporated off the water, leaving the dried sodium citrate. The blood samples were centrifuged and the plasma samples from the three types of tubes were spiked with 0, 0.05, 0.1, 0.5, 1, 2 and 5 ng of digoxin per ml of plasma, then all samples were assayed by the new procedure. IX. Application of the Assay in Human Studies. The assay method has been successfully used in two human digoxin bioavailability studies--one a four-treatment crossover study in 12 subjects with 14 sampling times, the other a three-treatment crossover study in 15 subjects with 15 sampling times. The “unknown” plasma samples from these two studies were assayed independently in the same laboratory by each of two analysts (designated “B” and “S”). All samples for one subject from all phases of one study were assayed on the same day by both analysts. Each day that “unknowns” were assayed, each analyst also prepared calibration data by using pooled pre-dose plasma (digoxin-free) of the same subjects that were ured in the study. Also, a quality control sample, theoretically containing 0.9 ng of digoxin per ml of plasma, was prepared and aliquots were frozen; each day each analyst removed one of the vials from the freezer and aesayed the sample along with the “unknowns”. Before pooling the pre-dose plasmas for calibration purposes, each analyrt assayed all pre-dose plasmas for a given study on the same day as well as obtaining calibration data that day. X. Type of Calibration Plot. The calibration data were plotted various ways as described in the literature (31-35) to see if a suitable linear relationship could be obtained without undue bias in any segment of the plot and with similar coefficients of variation for inversely-estimated concentrations at various known digoxin concentrations in the range desired, namely 0.05 to 5 ng/ml plaema. No satisfactory linear relationship could be found Results will be given for the logarithmicwhich satisfied the criteria. logistic relationship (so-called “logit”-log plot) (31). For this type we plotted

In

[y]

versus

In x,

according

to equation

1. where

stands for natural logarithm, p - %?LL x 100, x - digoxin B(0) B(X) is the X bound at in ng of digoxin per ml of plasma, centration such

of

a plot

the

the where

x and B(0) the

slope

often-used

least

squares

In Q is

the

is

the

X bound

is

positive,

plot

of

line

obtained

intercept

In

in

but

the the

absence

of

[1+L-versus In x (31). may be represented as

(corresponding

to

the

value

concentration digoxin

digoxin.

same magnitude

In

as

conIn

the

slope

The equation in equation of

In [ l-y

of 1,

Y

I

S when x were

1 and In x = 0)

then

and s is

inversely-estimated

and programming

The most normalized the

an electronic

calculator

lnfti]

=lnQ+s*lnx

P

% bound,

successful

and P(4)

are

fitting,

performed

ted

with

program.

from a known set

In such final desired XI.

the parameters

squares

equation

3 using a case

answer error.

in

this

plot

Sullivan

in

obtained We used

et al. --

(36), 3,

were then

0.01

squares

nonlinear

inversely-estimaof

satisfied of

P(3)

Eq. (3)

least

x 100 and

@

an electronic

ng/ml

fitted

P(2),

lx

calculator.

and automatically, is

the equation

who

P(l),

and a suitable

with

gradually

of

by fitting

exponential

and a known value

when the equation an error

Eq* (2)

the nonlinear

x,

procedure

incremented

2

2.

sj

In equation

computer

parameters

equation

+ P(3) .e-‘(‘)

Concentrations, of

Concentrations

1 to equation

was obtained

manner.

estimated

line.

to a double

after

a digital

an iteration

x is is

In Q) /

x 100, 3,

the

equation

Eq. (1)

x 100 I P(1) .e-p(2)‘x

i%

least

calibration

9% shown as equation

[B(X) 1, x data

of

to solve

((WV1 -

% bound values

type

the slope

by rearranging

x _ e

of

'PEIROXDS

with

and the some

digoxin.

Modification of the Plasma Assay for Urine. The assay for digoxin in urine was run essentially the same as for plasma, except that 500 ul of the subject’s pre-dose (zero hour) plasma was used in place of the “unknown” plasma and the urine was introduced The validity of taking as a 10 ul aliquot using an Eppendorf pipet. such small aliquots was justified by repetitive pipettings of tracer solution with the same pipet, addition of liquid scintillation fluid and counting. For ten repetitive pipettings the coefficients of variation were 2.07% for the first counting and 2.64% for the second counting. Then two vials were each counted 10 times, with the resultant coefficients of variation being 1.13% and 1.50%. Experiments were also done to determine if the presence of plasma was necessary in the urine assay, and, if so, how much. The normalized X bound values corresponding to digoxin concentrations of 0.1, 0.5, 1, 2 and 5 ng/ml (total volume = 500 ul) were lower when buffer only was used, but not significantly different for each concentration when 10, 50, 100, 150, 300 and 500 pl of plasma were used. Hence, as little as 10 pl of plasma in a total volume of 500 ~1, has a maximal effect. Obviously then, changes in albumin concentrations do not affect the assay, as has been reported by Holtsman et al. (31), but challenged by Shaw (32). Why some minimum amount of-&&a is necessary is unknown.

Because of higher concentrations of digoxin in urine than in plasma, the concentrations used for the urine calibration curves were 5, 25, SO, 100 and 2M ng/ml--i.e. 50 timas those used routinely in the plasma assay. BESDLTS AND DISCUSSION I . l3 Determination

of Optimum Standards

Logarithmic-logistic equation

logarithms (see

1 1

2

(Methods section

experiment.

the least

1) and the slopes

of Volume.

squares

and intercepts

X,

Using the lines

were calcu-

of these

lines

are

1.

1. Parameters of Least Squares Lines of Logit-Log Plots Coefficients of Variation from Inversely Estimated Concentrations for Experiment I.

of

5pe

k

in this

of the variables

equation

shown in Table Table

were prepared

1) from the data collected

natural lated

plots

and Effect

Slope Intercept 0.927 0.164 1.15 0.162

Correlation

C.V.

Coefficient 0.999 0.999

Estimated

(2)

and

from Inversely Concentrations 6.99 5.18

Standards Kit 30% EtOW-h0 (Volume not constant) Kit

0.985

0.182

0.998

8.96

$1~~-B20

1.05

0.203

0.991

6.2

constant) When labeled antibody,

Bodbard -et al.

such a logit were held

and unlabeled

plot

constant,

(31)

digoxin indicated

would be unity. do not differ

1, when the volume was not held dards,

the slope,

ing a bias different

1.15,

amounts of ethanol,

When digoxin 2, the coefficients ethanol-water result

of variation

standards

is a reflection

somewhat more accurate

from unity constant

than for

the theoretical

for

slope

significantly.

of

But,

on day

the 30% ethanol-water

different

from unity,

probably

stan-

indicat-

caused by

than a volume effect.

were inversely

estimeted

using

were lower on both days for the kit

standards

(Table

that our own 30X ethanol-water than the kit

for

on day 2, when volumes

volumes added,

rather

concentrations

that

The slopes

was significantly

caused by non-uniform

have the same affinities

standards.

the 30X

1).

This

standards

On this

basis

equation

were

we decided

13 For ready reference, sections have seme romsn numerals as corresponding section in the Methods section.

S

-EDICOXD-

795

to use our own standards rather than the kit standards. II. Determination of the Kinetics of Association and Optimum Pre-Incubation Time.

”....... _r

.............8

ZiR. 1 - Kinetics of association (see Methods and Results II). Top curve (0), 25'C and bottom curve (A), 37'~. Tangent lines represent initial rates (see text).

Y I r

1 0

lllltllll

20

40

Bcl TIK

80

II’ 100

120

IN IIINUTES

2000

Figure 1 presents the results of this experiment. The optimum preincubation time is that time when the counts/min are first in the asymptotic region of such a curve. This time allows as much as possible of the unlabeled digoxin to associate with the antibody before the labeled digoxin is added to fill up the sites unoccupied by the unlabeled drug.

The results at 25OC indicated that a pre-incubation time of 60 min

is about optimum, whereas at 37'C only about 15 min is necessary. However, because of rapid dissociation at 37'C (later Figure 2), the asymptote of the 37OC curve is appreciably lower than the asymptote of the 25'C curve. Hence, in order to obtain higher B(0) values, the preincubation at 25'C is more desirable. The first six points obtained at each temperature (solid lines in Figure 1) were fitted to the equation of a parabola (equation 4).

The initial "reaction rate", dT/dt, could

TDRaOXDI

m

796

then be ecrtimated by differentiating yielding

the result

equation

shown an equation a0 + alt

Y -

(dT/dt),_,this

method, from the data

25’c and 37’C, III.

5.

+ a2t2

in Figure

Kq.(5)

1, were 331 and 364 couata/min2 for

of

the Kinetics

of Dissociation

-1 ++m \ ..-...._,.c i.

‘“-L... ...._..,_(.(_,,_

. ....

:j&\,

\‘,

s

‘...

Ysw

c

+

-4.

x 103

-m....,, -%...__,

m

I .‘h

i

fall-off

increases

3.6

3.4

3.2

‘L, “......_, ‘R...

--......

“.-*..

----...-...-_+

Figure 2 presents curves

of the Digoxin-

Figure 2 Kinetics of dirrociation of digoxin-antibody complex (see Methods and Results 411). Curve A, ice-bath (0 C); curve B, 2S°C; curve C, 37’C. Tangent lines represent initial ratee Inret: Arrheniue (@cc text). plot for k2.

Im

200

t,

by Eq.(4)

al

1100

3m

to time,

obtained

Complex as a Function of Temperature.

Antibody

‘Km

rater,

respectively.

Determination

;. g

4 with respect

Initial

the results

and approach

markedly with

of each curve were fitted

of thir

asymptotic

increase

in temperature.

The first

of a parabola

with equation

The diseoclatlon

The rate

valuer.

to the equation

were estimated

experiment.

and initial

rates

5.

were 5.709,

26.32 and 56.47 counte/min2 for O’C,

of disrociatlon eight

(equation

There “initial 25’C and 37’C,

points 4), rates”

S respectively.

TD=OXDS

797

At O°C there was an initial

which there was a parabolic with the rate being incubation,tubes

fall-off

lag period

of counts/u&n

very much less

of 5

than 25’~ or 37’C.

should be immersed in an ice-bath

qin, after

from 5 to 30 min, Thus, after

and charcoal

treat-

ment at O°C is needed as reported rates”

indicated

the first

order

of tracer-labeled equation

by Samols -et al. (37). The “initial above may be taken as a measure of k2Ao, where k2 is

dissociation digoxin

1) were 1075,

in counts/min.

881.6

and 784.2

The estimate8 counts/a&n

amount

of A0 (a0 of

for O°C , 25°C and of k2 were made by

k2Ao value by the appropriate A0 value, then -1 to set . The k2 values obtained were -3 -1 for O’C, 25’C and set x 1o-4 and 1.196 x 10

by 60 to convert

x 10-5,

4.993

37OC, respectively.

Inset

these

The calculated

IV.

and A0 is the initial

the appropriate

multiplying 8.956

constant

From these values estimates

37OC, respectively. dividing

rate

values

of k2.

Determination

in Figure

of Necessary

1 is

the Arrhenius

activation

Incubation

plot

based

on

energy was 11.7 kcal/mol.

Time After

Tracer

Solution

Added. Figure 3 Determination of necessary incubation time after tracer added (see Methods and Results 0.1 IV) . Top curve (m), ;gL;;Albottom curve, (01, .

S

798

T=RIUTDDS

Resultsof this experimentare shown in Figure 3. This experkment indicatedthat the tfme requiredfor the labeleddigoxin to occupymost of the sites uncrccupfed by the unlabeleddigoxinwas about 60 min, since the curves are approachingtheir asymptoticvalues at that time. Hence, an incubationtime of 60 min at 25OC, the seme as the pre-incubation time, is satisfactory. V, peter&nation of Outimurn CharcoalExnosureTime Under Ice-Bath cOnditiOn8.

FfRure4 Determinationof optimum charcoalexposuretime under ice-bathconditions(see mthOd8 and Resultsv).

Resultsof this experimentare shown in Figure 4, There appeared to be an initial'linear drop in countdmin for the first 10 min. These six points gave a least squares equation:counts/tin= 8683 - 53.2 t (r - -0.958). At 10 m&n therewas an abrupt change in slope,and, in the LO-60 min period the drop again appearedto be linear;the six points in this period gave the equation:countdmin = 8204 - 12.4 t (r - -0.990). We chose 10 min as the charcoalexposuretime under fcebath conditions,since any slight increasein exposurebeyond 10 s&n

would not make a great VI. Determination

of

difference

in counts/min.

the Minimum Amount of Liquid

Scintillation

Fluid

Necessary. The counts/min

progressively

from 8 to 15 ml of liquid

scintillation

used the solution

was definitely

solutions

intermediate

tion

were of

was definitely

increased

clear.

fluid

cloudy;

was used.

When 8 ml were

when 10 or 12 ml were used the

clarity;

Hence,

from about 800 to 950 when

when 15 ml were used,

15 ml of scintillation

the solu-

fluid

was

necessary.

Final

asray procedure

500 nl of plasma,

- To each tube is added the following

500 ul of pH 7.4 phosphate-buffered these are vortexed,

and 50 pl of 30% ethanol-water; serum solution pre-incubated added,

is added and the mixture for 60 min at 25’C.

the mixture

is vortexed,

the end of the incubation 500 ul of charcoal the mixture trifuged

in a Sorval

supernatant into ing,

then

counted

8(X)

for

for

to provide

the desired

is used for

when the charcoal

that the minimum charcoal VII.

At then

is added and

for 20 min.

The

the charcoal, vial

by shak-

counter.

data,

digoxin-free

the B(0) values,

water contains number of tubes,

suspension contact

plasma is

but for

sufficient

the

digoxin

the timer should be

has been added to the last tube,

SO

time is 10 min.

Quench Correction. Linear

regression

analysis

of the ZYata from the two groups

DpI+!and 11,000 DP&f 3D20) indicated ences

is

concentration.

With batch assays with a large started

is

the tube is cen-

in a scintillation

the calibration

digoxln

time,

not to take any of

fluid

the 50 pl of 30% ethanol

solution

in an ice-bath,

at 2000 r.p.m.

10 min in a scintillation

and the same procedure values

careful

scintillation

tubes

tracer

in the ice bath,

(0-4’C)

being

This mixture

for 60 min at 25’C.

a 10 min charcoal

centrifuge

liquid

In preparing

pre-cooled

solution

then 100 ul of anti-

the tube is placed

After

is decanted,

15 ml of

used,

suspension,

saline

again.

Then, 100 ul of then incubated

time,

is vortexed.

vortexed

in sequence:

in either

countsfmin

plots,

the slopes indicating

quenching was not affected

(110,000

that there were no significant

or intercepts

of the X efficiency

that the % efficiency by the total

radioactivity

relationship present

differ-

versus to in the

m

800

range

studied.

through

Prom all

the origin,

data

of 7.559 rather

the least

- (6.683

in the same laboratory,

the mixtures

x 10-5) Stoll

x 1o-5 when hydrochloric

than

squares

regression

line,

forced

was:

X Efficiency Previously,

-J?TBROXDtS

acid,

of plasma

(AES)

Eq.(6)

et al. (14) obtained a slope -chloroform and water were used

from hemolyzed

and non-hemolgzed

blood. An electronic value tube

using for

calculator

the following

the plasma

the total

program was written

equations,

sample,

for the case

two tubes

for

to calculate

a 7. bound

where there

the blank

is one

and two tubes

count. S -8 X Bound = TC _ BC x 100

gq. (7)

C

C

S sc=

l

100 t(AWs *1

100 WWB

x Sll

and total

blank

tively;

AES represents

1~6.683~10

#l,

blank

sample; -5 .

obtained

T2 1 100 [(AWT

+

S, Bl,

2 (next

the counts

page)

in the anticoagulant

with

satisfactory has

of doing bias

Eq.W

Eq. (10)

l/2 2

x Sll sample,

blank

B2, Tl and T2 are counts/z&n #1 and total

obtained

count

for

#2, respec-

from the Blue channel

the slope

for

of the quench plot,

inversely-estimated

study.

results,

less

2

l/2 x S11

the

name-

Study. lists

from the biexponential

the advantage

count

and Sl represents

data

usually

100 t(AWB

#2, total

and X. Anticoagulant Table

*2

1

respectively;

sample, particular

+

10, SC, Bc and Tc are the corrected

7 through

counts,

x Sll

1

T1 Tc = [ 1 100 [(AXS)T x Sll 1 In equations

Xq (8)

1

*c * [l

VIII.

for

fittings All

as the data

duplicate than either

assays,

concentrations

(equation three

anticoagulants

since

may be used

These data

indicate.

individual

3) of the calibration also

show

the mean of duplicates value

on the average.

S

801

TDEOXDS

Table 2 Inversely-Estimated Concentrations from Plasma CalibrationCurve Data (Biexponential Fits) in AnticoagulantStudy Inversely-Estimated Concentrations of Digoxin (&ml) Known Digoxin sodium Dried Na2ED!CA Citrate Concentration(r&ml) Heparin 0.04 0.07 0.05a 0.06 0.04 0.05 0.05 0.05 0.055 0.05b 0.10

0.07

0.10

0.41

0.53

0.52

1.0

1.13 1.10 1.115

0.98 0.94 0.96

0.95 1.03 0.99

1.91 1.93 1.92

2.12

2.02

2.0

5.0

_.v--

4.95 5.13 KG

5.28 5.17 5.225

0.10

0.50

_w--

aDuplicateassay values by same analyston same day. b Average of the duplicates. Table 3 Data for SensitivityLimit of PlasmaAssay NormalizedX Bound IIEQQ g o x 100 Parameter Theoreticalfor C I 0 computer-fittedfor c I 0a Observedaverage for C - 0.05 Computer-fittedfor C - 0.05

Sodium Hevarin 100.0

Dried Citrate 100.0

Na2EDTA 1oo.o

116.75 100.2

107.3 98.1

100.1 92.05

100.0

99.2

92.9

aOalue of P(1) + P(3) from fittedequation3, where C is the digoxin concentrationin ng/ml.

S Comparing indicates of

that

the

assay

116.75

is

ng/ml

IX.

2 and 3 in Table

with

biexponential

0.05

&ml,

significantly

the

by data

rows

la

apparent

unreliable.

TIOEOXDI

fitting

since,

for

different

Ability

to estimate

(equation

100.2.

In the

range

accurately

ng/ml

0.05

renaitivity heparin

or positive

of

the

Aaaay

0 to 0.05 and

ia

also

is

supported

to Humen Studier.

Table

No. of Sample8 per Subjects Subject

4

Total Samples

by Two Analyst8 (Ii and S) Each Phase of the Studies X Bound

Parameter

A

S

1

12

4

48

Nean Range C.V. (X)

44.0 41.4-46.7 3.34

44.0 41.1-46.1 2.90

2

15

3

45

Mean Range C.V. (X)

45.8 44.0-48.0 2.62

43.1 40.8-45.8 2.63

Table analysts

4 auwnarizee with

studies. table

the

was higher day,

that

The coefficients

Table and bias the

the

table

ent

days.

logistic logistic

ia

of

aeta plot

since

indicative

the

B(0)

value

pre-dose

of

in

also

comparea

the

for

namely

of

very 0.1

and

non-linearity.

large

produced

contained

digox-

2.90,

2.62

of

variation

concentrations the

subjects

achieved

with

out

standard for

the

logarithmicof

the

logarithmic-

deviations analyst

considerable

mean in

on 12 differ-

fits

the

for

Each

studies.

biexponential

and

the

calibration

3.34,

coefficient8

We ruled

ng/ml,

the

in

aaaay.

12 different

the

humen

groups

plasmas 4,

one of

results and

data.

produced

(38)

of

plot

calibration it

four

inversely-estimated

plaamea

teat of

the

each

in Table

from

concentration

a linearity

of

collected

calibration of

in

mean concentrations, from

two independent

two different

reproducibility

data

The table

known digoxin failed

the

of

by the in

variation

obtained

derived

type

individual

none

5 aumaerizea

calibration

value

obtained samples

corresponding

of to

valuer

all

the

indicating

attest

reaults plasma

X bound

than

in.

2.63,

the

pre-dose

The lower

data

trend

the

sodium

can be negative

Sutmmry of Results Obtained Independently in Two Studies Using Pre-Doae Plaamaa of

and

3)

for

separately

2.

and X. Application

Study

anticoagulant

example,

from

concentration

in Table

3 for

at

S the biaa

with

a

data a

Table Sunnnary of

Calibration

5

Mean, Coefficient of Variation and Bias Estimated Digoxin Concentrations

Method

Mean Known Estimated

Analyst

Logarithmic-logistic with pooled data’

of InverselyBiasb ng/ml

C.V.(%)

X

Ii

0.1 0.5 1.0 2.0 5.0

0.100 0.516 1.036 2.00 4.85

29.8 7.00 7.54 5.38 5.16

0 0.016 0.036 0 -0.15

0 3.2 3.6 0 -3.0

8

0.1 0.5 1.0 2.0 5.0

0.095 0.546 1.09 1.97 4.66

24.6 8.54 5.50 5.35 5.44

-0.005 0.046 0.09 -0.03 -0.34

-5.0 9.2 9.0 -1.5 -6.8

Individual subject biexponential fits

R

0.1 0.5 1.0 2.0 5.0

0.095 0.498 0.988 2.10 4.82

7.10 4.18 4.08 4.48 8.78

-0.005 -0.002 -0.013 0.10 -0.18

-5.0 -0.4 -1.3 5.0 -3.6

Individual subjact biexponential fits

8

0.1 0.5 1.0 2.0 5.0

0.0967 0.500 0.994 2.04 4.94

5.09 2.26 3.60 3.53 3.14

-0.0033 .:06 -0 0.04 -0.06

-3.3 0 -0.6 2.0 -1.2

with Logarithmic-logiitic pooled data

asoth analysts spiked digoxin-free pre-dose plasma of twelve subjects on The normalized X bound values were used each of twelve different days. to prepare both types of calibration plots. b The bias expressed in ng/ml is the difference between the mean estimated This value expressed concentration and the known digoxin concentration. as a percentage of the known concentration is the X bias. ‘These

data

passed

the

linearity

test

(38).

dThese

data

failed

the linearity

test

(38).

Table

6 summarizes

results

obtained

in one of the human studies. biexponential logistic

fits,

plots

of performing

biexponential

of pooled

fits

value

assays

than either

average,

tion

(i.e.

0.90

individual

obtained

ng/ml,

value

Results

on any given alone, with

Also,

control

individual

data

on each sample,

obtained

subject

with

of pooled

are compared.

of the two assays

the overall method

data

two independent

shown) the average to the +zue”

Results

wLth the quality

samples

subject

and logarlthmicshow the advantage sinca

(although

day is usually as seen

closer

in Table

the recommended calibra-

biexponential

fits)

is

not

the actual

6,

concentration.

It should be noted that the coefficient8 of variation

obtained from the quality control samples, namsly 6.23 and 8.182, by the recosmmnded method, are somewhat higher than mOst of those obtained from inverrely-estimsted concentrations from calibration curve data (Table 5). Results obtained in the second human study are not reported, but were very similar. Table 6 Results Obtained with Quality Control Sample of Spiked Plasms Assayed Each Day a Gallbratlon Curve Was Prepared Independently by Both Analysts (?land S)

Paranvter *an

gange C.V. (X) Pooled data Mean C.V. (D

Concentration of Digoxin (an/ml) Individual Bubject Biexponential Fits Logit-Log with Blexponential Pits of Poolad Data Pooled Data Ii S H S H S 0.94 0.85 0.93 0.84 0.99 0.89 0.88-1.00 0.71-0.93 0.86-1.00 0.64-0.94 0.92-1.07 0.84-0.96 6.23 8.18 4.25 8.65 4.66 9.14 0.89 8.17

0.98 8.57

XI. Modification of the Plasma Assay

0.94 8.63

for Urine

Prom inversely-estimsted concentrations of both analysts on three separate days (i.e. 6 values per concentration), the coefficients of variation were 5.77, 1.97, 1.68, 3.04 and 3.44% for known urine digoxin concentrations of 5, 25, !IO,100 and 250 ng/ml, raspectlvely. Prom six quality control samples, assayed on the sm

three days, sn overall

coefficient of variation of 6.19% was obtained, which is very comparable to those obtained in the plasms assay. GENBPAL DI8GUSSION Various measures of precision of radioimnrnoassayshave been discussed in the literature. Among these are: (a) the coefficients of variation calculated from X bound values; (b) the coefficients of variation calculated from norusllsed X bound values; (c) the coefficients of variation calculated from the transforrsstionon the ordinate of the calfbration plot, which is used to linearise the data; (d) the coefficients of variation calculated from inversely-estimatedconcentrations; and (e) the coefficients of variation calculated from assays of quality control et al. (39); it samples. These were discussed by Tembo --

was shown that

the relative (e)

> (d)

comparing of

magnitudes > (c)

different

and (e)

nected the

of

power of

one dose

From quality

tions

3.62

of

6.6

Table

6 after

for

application

a quality

Our value

6.19X

instructions

the

sample

developed

containing

of

are

“Precision closely

control

related

to

from zero

the or one

the distiguishing

al.

on precision.” (41)

reported concentraof

well of

urine

con-

the higher

8.572,

1.88

of

digoxin,

(c)

(a)

rather

variation

reported with

their

variation

ng of

sample

in

of

8.97.

digoxiniml. compares

0.5

ml of

the kit

plasma

the kit.

(d)

0.1,

1, 2 and 5 ng/ml

than

ng/ml.

(e)

those

We pre-incubate

1 hr at 25’~, We incubate

recommended

whereas for

under

the kit

1 hr after

room temperature whereas 0.05

ice-bath

- 5 ng/ml

time

is

of

rather

curve

(and 0.5,

digoxin

is

1.5,

2,

whereas

(h)

We use

recommends

10 ml.

or serum,

the kit

15 ml of Since

whereas

0.05

the antiserum step.

for

(f)

in the

kit

We do the charcoal method

liquid

recommends

scintillation

the new assay the kit

ng/ml),

3 and 5

whereas

(g)

We

those

we use digoxin

with

added,

(b)

than the 0.2

sometimes 1,

recommended.

conditions,

plasma

standard

stan-

than using

method has no pre-incubation

the 3h-digoxin

conditions. the kit

in the kit

the unlabeled

a 30 min incubation

treatment

For the

rather

or serum,

in

rather

in the kit

serum standards.

suspension,

of

0.5,

recommended

We use our own 30% ethanol-water than

and charcoal

We use

from that

concentrations

range

et --

agrees

ml recommended

fluid,

any assay

mean digoxin

a coefficient

differs

ways:

our own buffer

the kit.

method

of

to coefficients

new assay,

the quality

in several

solutions

prepare in

for

the value

reported

types

dependent

Fraser

ng/ml

similar

that

of

in

for

a dose

clearly

sera

usually

.

The new assay

dard

of

Hence,

urine

is say

sense

corresponding

(42)

for

also

and 0.20

of

al.

control

of

favorably

et --

curve

in the

samples

&ml,

that:

to distinguish

is

respectively.

Halkin

out

One can generally

0.24

and 1.65

and 12X,

values.

of

the precision

a standard

are

must be excercised

variation

pointed

sensitivity,

control

deviations

of

of

from another,

standard

of

Cekan (40)

variation

caution of

indicators

the ability

from another . ..The

dose

of

great

coefficients

an assay.

the higher

coefficients

hence

the determination

sensitivity

precision,

such

valid Also,

above.

with

of > (b),

authors’

The swst

assays.

(d)

> (a)

assay

covers

covers

the the

806

6c

T~EIROSD~

0.5 - 5 ng/ml,we saw little reason to compareresultsin the common range of 0.5 - 5 ngfml. The advantageof the new assay is accuracy range

and reproducibility in the range 0.05 - 0.5 ng/ml,which is not covered by the kit assay. REPRR&NCES 1. Jelliffe,R. W. and Blankenhorn,D. H., J. Chromatogr.12, 268-270 (1963). 2. Wilson,W. E., Johnson,9. A., Perkins,W. H. and Ripley,J. E., Anal. Chem. 39, 40-44 (1967). 3. Maume, B., Wilson,W. E. and Horning,E. C., Anal. Letters4 401415 (1968). 4. Wilson,W. E. and Ripley,J. E., Anal. Chem. 41, 810-815 (1969). Watson, E. and Kalman, S. M., J. Chromatogr.56, 209-218 (1971). 65: Lowenstein,J. M., Circulation3l, 228-233 (1965). 7. Lowenstein,J. M. and Corrill,E. M., J. Lab. Clin. Med. 67, 10481052 (1966). a. Gjerdrum,K. Acta Med. Stand.187, 371-379 (1970). 9. Marcus, F. I., Ryan, J. H. and Stafford,M. G., J. Lab. Clin. Med. 85, 610-620 (1975). 10. Butler,V. P.,Jr. and Ghan, J. P., Proc. Natl. Acad. Sci. 97. 71-78 (1967). 11. Smith, T. W., Butler,V. P.,Jr. and Haber, E., Biochemistrys 331337 (1970). 12. Smith, T+ W., Butler,V. P., Jr. and Haber, E., N. Engl. J. Med. 281, 1212-1216(1969). 13. Larbig, D. and Kochsiek,K., Dtsch. Med. Wochenschr.97, 1310-1312 (1972). 14. Stoll, R. G., Christensen,M. S., Sakmar,E. and Wagner,J. G., Res. Corn.Chem. Pathol.Pharmacol.k, 503*510 (1972). 15. Clark, D. R. and Kalmar,S. M., Drug Metab. Dispos.& 148-150(1974). 16. Greenwood,II.,Snedden,W., Hayward,R. P. and Landon,J., Clin, Chim. Acta 62. 213-224 (1975). 17. Kramer,W. G., Bathala,M. 8. and Reuning,R. H., Res. Corn.Chem. Path. Pharmecol.& 83-88 (1976). 18. Gault, M. H., Ahmed, M., Symes,A. L. and Vance, J., Clin. Biochem. s 46-52 (1976). 19. Sugden,I).,Ahmed, M. and Gault, M. Ii.,J. Chromatogr.121. 401-404 1976). 20. Barrett,M. J. and Cohen, P. S., Clin. Chem. L8, 1339-1342(1972). 21. Meade, R. C. and Kleist,T. J., J. Lab. Clin. Med. 80. 748-754 (1972). 22. Phillips,A. P., Clin. chin. Aeta 44. 333-340 (1973). 23. Zeegers, J. J. W., Mass, A. H. J., Willebrands,A. F., Kruyrwijk, H. H. and Jambroes,G., Clin. Chim. Acta 44. 109-117 (1973). 24. Beach,H. R. and Watanabe, A. M., Clin. Chea. 2& 1315-1329(1975). 25. MarCUS, P. I., Diokerson,J., Pippin,S., Stafford,M, and Bresaler, 253-259 (1976). R ., Clin. Pharmacol.Ther. a 26. Lader, 9. R., Court,G., Johnson,B. F. and Hurn, B. A. L., Stand. J. Clin. Lab. Invest.29, 14.10 (1972),Suppl. 126. 27. Weintraub,H., Au, W. T. W. and Lasagna,L., J. Am. Med. Assoc. 224, 481-485 (1973).

S 28. 29. 30.

31. 32. 33.

35. 36. 37. 38. 39. 40. 41. 42.

807

Park, H. M., Chen, I. W., PIanitasas,G. T., Lowey, A. and Sacnger, E. L., J. Nucl. Med. 14, 531-533 (1973). Calesnick, B. and Dinan, A., Clin. Chem. 22, 903-905 (1976). Wagner, J. G. and Ayres, J. W., Bioavailability assassment: Methods to estimate total area (AUC-c-00) and total amount excreted (Aw) and importance of blood and urine sampling schemes with applfcation to digoxin, J. Phannacok. Biopharm., submitted January 12, 1977. Holtzman, J., Shafer, R. and Erickson, R., Clin. Chem. 20, 11941198 (1974). Shaw, W., Clin. Chem. 2l, 636 only (1975). Rodbard, D., Bridson, W. and Rayford, P. L., J. Lab. Clin. Med. 14, 770-781

34.

TSIROSDS

(1969).

Harding, B. R., Thompson, R. and Curtis, A. R., .I.Clin. Pathol. 26, 973-976 (1973). Rodbard, D., Clin. Chem. 20, 1255-1270 (1974). Sullivan, T. J., Stoll, R. G., Sakmar, E., Blair, D. C. and Wagner, J. G., J. Pharmecok. Biopharm. 2, 29-41 (1974). San&s, E., Lewis, M. and Bell, D., S. Med. J. 67, 1392 only (1974). Documents Geigy, Scientific Tables, Sixth Edition, 1962, pp. 173178. Tembo, A. V., Schork, M. A. and Wagner, J. G., Steroids 2S, 387403 (1976). Cekan, Z., J. Steroid Biochem. 5 271-275 (1975). Fraser, B. J., Leach, R. H., Poston, J. W., Bold, A. H. Culank, C. S. and Lepedo, A. B., J. Pharm. Pharmac. a 968-973 (1973). Halkin, II.,Sheiner, L. B., Peck, C. C. and Melmon, K. L., Clin. Pharmecok. Ther. 1z, 385-394 (1975).