Clinical use of serum digoxin concentrations

Clinical use of serum digoxin concentrations

Clinical Use of Serum Digoxin Concentrations Richard P. Lewis, The development of the radkimmunoassayfor dlgoxln by Smith and coworkers in 1999 was a...

1MB Sizes 3 Downloads 83 Views

Clinical Use of Serum Digoxin Concentrations Richard P. Lewis,

The development of the radkimmunoassayfor dlgoxln by Smith and coworkers in 1999 was a landmark in diills therapy. since then, the complex pharmacokinetics of dtgoxln have been defined. As a result, the i&dence of digltalls toxkAty has markedly demeasd. To use the dlgoxln assay properly, howver, the relaUon of thii pharmacokinetic parameter to dioxin phannacodynam‘m must be known and the lMtat&ns of . Systok tlme intertheassayltselfuvals (9Il) are unkluely useful to quantitate the lnatropic effect oPdigitalii preparations. This technique can demonstrate the onse4 and magnitude of the lnotropic effwt for both oral and intravenous digltalls administraWn. By deflnlng the mathematical relation between STI and slmultaneous serum diioxin concentratknsfolowing intravenous administration of 1 mg digoxin, computersbnulatiicanbemadeoftheeffectof dosing changes on blood and tfssue concentrations. The serum dlgoxln assay has technical probiams relating to quality control, interference by metabolites, and cross-reactkns with endugenous dlgitalk+like sub&mces. Further, a standard time for measurementfol~dosinghas not been established. Physktal acttvity can signifScantly altar the serum dlgoxln concentratkns by increasingskeWalmusdebiwBng.Numerous chugs can interfere with dlgoxin absorptii or eliminat9on. Uslng the serum dfgoxln assay isthe only way to assess these interactions. Computer suweiltance (kleallywlth physician 01 pharmacist hteraction) has ken used to monitor digitalis buthasnotyetgained wklespread awnce. Thislsdea*amethodinneedoffufthertestJng. (Am J Cardioll99&69z97c;lO7G)

From the Division of Cardiology. the Ohio State University, Columbus, Ohio. Address for reprints: Richard P. Lewis, MD, The Ohio State University Division of Cardiology, Room 647, Means Hall. 1654 Upham Drive; Columbus, Ohio 43210.

MD

T

by radioimmunoassay he determination (RIA) of efficacious and toxic serum digoxin concentrations in 1969 dramatically improved digoxin therapy.’ The incidence of digoxin toxicity has diminished from 25% to 2%, Largely because serum digoxin concentrations within the generally accepted “therapeutic range” have been maintained.2J This knowledge has also allowed a more complete understanding of digoxin pharmacokinetics. Unlike most cardiovascular agents, digoxin has prolonged equilibration of serum and tissue concentrations and has a relatively long half-life. Proper interpretation of serum digoxin concentrations, therefore, requires an understanding not only of digoxin’s pharmacokinetics, but also of its pharmacodynamics (relation between serum concentrations and clinical effect).4 Fortunately, digoxin’s inotropic effect can be measured by the duration of electromechanical systole corrected for heart rate (Q&I) from systolic time intervals.s,” The QSJ has been shown7 to measure the onset and magnitude of this inotropic effect in humans. The changes in Q&I directly parallel changes in left ventricular dP/dt,,, (maximal rate of pressure increase in the ventricle), as measured by a highfidelity manometer tip catheter (Figure l).* DIGDXIN PHARMACOKINEMS PHARMACODYNAMICS

AND

In previous studies from our laboratory,’ digitalizing doses of 4 commonly used digitalis glycosides were administered intravenously to normal subjects. The shortening of the Q&I was measured serially and corrected for the molecular weight of the compound (AQSJ/mol). The response curve was exponential, as would be expected for an agent that binds to a receptori (Figure 2). Of interest, the predicted maximum AQS21 was not significantly different for any of these compounds, thus validating the empirically derived digitalizing doses7 (Table I). Exponential curves have a time constant representative of the point at which approximately 66% of peak effect has been attained. By comparing time constants, the relative speed of onset of A SYMPOSIUM:

CONGESTIVE

HEART

FAILURE

976

-70

-60

.

1

TABLE I lnotropic Response Used Digitalis Glycosides

.

-50

.

-40 . . . .

a-i cn'OJ -30

a0 -20

. .

- IO

f

l

.

for all pairs. C&l = duration of electromechanical Adapted from Circulation.7

.

I

I

I

I

I

,

3

4

5

6

7

A max dp/dt DP FIGURE I. Changes (Mm baseline)

(se&)

in the duration

of eiecand left ventricular dP/dt,, (maxlmal rate of pressure increase in the vent*ie) corrected for developed pressure (A,,,&P/@t/DP), obtained serlaiiy following administration of 1 mg digoxin intravenously to 6 patiants. k high-fidelity catheter system was employed. The direct relation between increased rate of pressure development and shortening of the Q&i is apparent. (Reprinted with permission from Cafxfiovasc C/in.*)

trom~anlcal~ystolecorreotedforheartrate.(AQS~l)

activity of the various digitalis compounds can be determined427 (Table I and Figure 2). Intravenous digoxin is an intermediate digitalis formulation in terms of rapidity of onset of action; most of the inotropic effect from an intravenous dose is reached within 60 minutes. Thus, only l-2 hours need elapse between intravenous digoxin treatments. In Figure 3 intravenous digoxin’s onset of inotropic effect is compared with its onset of slowing heart rate in patients with atria1 fibrillation.4 The atria1 fibrillation data ar,e replotted from the original data of Gold et a1.9Remarkably, the curves are

5.8 7.2 23 56

f f f r

0.06 3.3 2.3" lO.O*

systale corrected

for heart rate.

identical. Thus, both the inotropic and chronotropic effects of intravenous digitoxin have similar pharmacodynamics. Others, however, have reported a more rapid ,effect on atrioventricular conduction in patients with sinus rhythm.lO The onsets of inotropic effect for the oral glycosides digoxin, digoxin elixir, and digitoxin tablets are compared in Figure 4.4 Given orally, the 3 compounds have a similar onset of action, with peak effect reached at 4-S hours. The slower onset of pharmacodynamic effect is related to a delay in the absorption of oral compounds. In Figure 5 the effect of an intravenous digitalizing dose of digoxin on the AQSTI is compared in normal subjects and patients with congestive heart failure (CHF).+Jl The time course and extent of shortening of the Q&I are identical for the 2 groups, suggesting that there are no major pharmacokinetic or pharmacodynamic alterations when CHF is present. Although shortening of the Q&I indicates inotropic stimulation has occurred, it does not necessarily imply a resultant hemodynamic benefit. Improvement in the ratio of subintervals of Q&I (pre-ejection period/left ventricular ejection time) indicates improved left ventricular contraction occurs in the majority of patients with CHF.8J1 To define more precisely the relation between

RGURE?. Onset of [email protected] of 4 d#gitaiis glycosides administered intravenousiy. The @mtropic response is shbwn as the change in the duration of eiectromechanicai systoie dorrected for heart rate (AQW per mole ofgiycoside. (Reprlnted wRh permission from Cardiatonic Dnrgs.‘l)

-4.

0

20

40

60

120 TIME

980

-r 4.6 -r 1.8

Time Constant (min f SEMI

. l *

. I

16.6 15.5

with Four Commonly

*p
. .

.

I2

Maximum Response -f- SEM)

17.6 _+ 4.2 19.5 _+ 5.2

Ouabain Deslanoside Digoxin Digitoxin

0 E

Theoretic lnotropic (QS,l/mol

Digitalis Glycoside

to Treatment

THE AMERICAN

JOURNAL

360

( MINUTES)

OF CARDIOLOGY

VOLUME

I

69

JUNE

4, 1992

FIGURE 3. Comparkon of Inotropic and chronotropic effects of intravenous digoxln over a g-hour period. The chronotropic effect Is cakulated from the slowing of heart rate in patients with atria1 fibrlllatlon and the lnotropk effect from the shotienlng of change in the duration of electromechanical systole corrected for heart rate (AQ&l) per mole. (Reprinted with permission from Cardiatonic Dnfgs.q

-20

-

-16

-

CHRONOTROPIC

I NOTROPIC

1 -

HR SLOWING (beats /min)

--20

8

2’0

40

60 TIME

serum digoxin concentration, tissue digoxin concentration, and digoxin’s inotropic effect, we administered 1 mg digoxin intravenously to 10 normal subjects12; 22 serum digoxin samples and Q&I measurements were obtained simultaneously over a 4-day period. The serial digoxin concentrations best fit an open, 3-compartment, pharmacokinetic model noted in a prior study.13When the predicted tissue accumulation of digoxin in the deep compartment of this model was correlated with the AQS21, there was a highly significant, nonlinear relationship12(Figure 6). This nonlinear relation is consistent with a finite action for digoxin, and not with earlier theories that had postulated a linear dose-40

--30

EFFECT

120 ( MINUTES)

response curve for digitalis throughout the therapeutic range. As will be discussed, the nonlinear relation has major implications for digoxin dosing. Employing the raw data from the previous study,12we were able to make various computer simulations. Figure 7 shows a dose-response curve for the inotropic effect of digoxin in relation to serum concentration. l4 Figure 8 shows the relation between serum digoxin concentrations and tissue concentrations following dosing with digoxin.r5 The peak serum level gradually increases assteady state is approached. More importantly, the more slowly equilibrating tissue concentration shows much less day-to-day variation. This represents the major

1 DIGITOXIN

-2o-

DIGOXIN

TABS 1.6 mg ELIXIR

1.5mq

nos,1 DIGOXIN TABS

- io-

+

I

IO

0

I

2

1

1

I

I

I

4

8

12

24

TIME

1.5 mq

( HOURS)

FIGURE 4. Onset of hwtropic actMty over 12 hours for oral dlgitoxln tablets, dlgoxtn elixir, and dlgoxin tablets. Dlgoxin elbdr has the most rapid onset of acthrity; digoxln tablets have the slowest onset. The lesser effect on the change in duration of electromechanical systole corrected for heart rate (A-1) by digoxin tablets represents bwomplete absorption. Digitoxin requires up to I2 hours to attain Its maximum effect, but its onset of acth!Ry is similar to that of dlgoxin tablets. (Reprkted with permksion from Cardiotonic 0rugs.q

A SYMPOSIUM:

CONGESTIVE

HEART FAILURE

930

,

AQS21

-20 -

(msec) -30 -

o CHF \

-40 -

NL

FltURE 5. Shortening of the change in duration of electromechanical systole corrected for heart rate (AQSd)In normal subJects (0) versus eubJects wRh congestive hoart failure (0) after intravenous admbdstration of 16 mg deslanoside. There is no signifitint difference in onset or extent of shortening. (Reprhrted with permission from Cardiotonic Dntgsq

advantage of an agent with a long half-life. The phenomenon does not occur with most cardiovascular agents, for which rapid changes in serum concentration between dosing are directly manifested as diminishing tissue effect. Figure 9 depicts the relation between serum digoxin concentration and tissue response for a single dose of digoxin given intravenously or orally.i4 The response is more rapid with the intravenous dose, but at 12 hours there is minimal difference between formulations. Figure 10 illustrates the difference between starting digoxin with a single daily maintenance dose and using a digitalizing dosage regimen in

normal subjects.l4 Although nearly twice as much digoxin was given in the loading-dose protocol, tissue and blood concentration at the end of 96 hours are nearly identical in the 2 dosing groups. Thus, when rapid onset of action is not required, a loading dose of digoxin need not be given. In patients with chronic heart failure, however, renal function is often diminished and the half-life of digoxin is prolonged. Reaching steady state would therefore take up to twice as long in such patients. Figure 11 illustrates a more rapid decline of the serum digoxin concentration than of tissue effect when digoxin therapy is stopped.14In patients with a prolonged serum half-life, the effect is magnified. This observation has clinical implications, for example, when assessingwhether an arrhythmia is related to digitalis therapy. It may take several days for the tissuedigoxin concentration to fall substantially. Digoxin and digitoxin are the 2 most commonly used cardiac glycosides. Remarkably, these agents differ by only 1 hydroxyl group on the steroid nucleus, but have strikingly different pharmacokinetic properties due to differences in polarity.4 Digoxin is a polar compound, although not nearly as polar as ouabain and deslanoside, whereas digitoxin is a nonpolar compound. Polar compounds are less lipid soluble and, therefore, less well absorbed, less completely bound to serum albumin (leading to a lower serum concentration), and have less renal tubular reabsorption. This results in predominant urinary elimination of the drug and thus a shorter half-life. However, this also

AQS21 (msec)

0

40 Tissue Digoxin (percent of dose)

60

20 Predicted

ov 0.1, I 0.3, I 0.5, I 0.7, I 0.9, I 1.1,

Cont.

Concentration

in Serum

at End of Dosing

Interval

(rig/ml)

I FIWRE 6. Correlation of the change In the duration of elect&nechanical systoh corrected for heart rate (AQSzl) RRUR5 7. Nonlinear relation betwaen response (percent of maximum) and serum dlgoxin concentratfon at 24 hours se&l measurements Gplasma digoxin concentrations postdose. Data were obtalned from slmulatkms using eiafter admbdstratlon of 1 mg digoxln intravenously in 10 ther single oral doses (0325-6.75 mg, q ) or muMpIe oral normal subjects: 22 Q&l and plasma digoxin concentradoses (0.l25-0.75 mg, 0) every 24 hours for l2 days. The tlons were measured slmuRaneousiy over a 66-hour perelation does not depend on the number of doses given riod. Circfes represent points from onset to peak actMty; and app5es only to the postabsorption, postdistrlbutive f-Hang/es represent the decay portion. The nonlinear corphase of the serum concentration-thne curve (about relation was r = 0.67 (p < 0.01). (Reprbrted wRh permis1224 hours postdose). (Reprinted wtth permIssion from J don from J Phammawkinet Biophamky) Clin Phannawl.14)

1000

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

69

JUNE 4, 1992

FIGURE 8. Simulated time profile of serum dlgoxln concentrations

(solidline)andtissuerespoase

as judged by the duratlon of electromechanical systole corrected for heart rate (&I) (dashed Ime) over a 4-day period. The tissue response ls plotted as a percentage of maximum response. Pharmacoldnetic and pharmacodynamic constants for the slmulatlon were obtained from a slgle-dose study In normal humans. The dosage regbnen for the simulathm was 0.375 mg/day diioxin tablets wRh a bloavallablllty of 0.8. (Reprinted with perml&on from AppJkd Pfrarmecokh?fiCS.~)

0.8

TIME (hr)

leads to variability in elimination because of the many factors that can affect renal function (reviewed later). The basic pharmacokinetic parameters of digoxin are listed in Table II. In the early 1970s the poor bioavailability of several digoxin formulations (not including Lanoxin) was documented, which led to strict Food and Drug Administration guidelines for digoxin manufacturers to ensure adequate bioavailability.16More recently, Lanoxicaps have been developed. This digoxin formulation has improved oral bioavailability.17 Substances that can affect digoxin bioavailability are listed in Table III. Cathartics and antacids inhibit digoxin absorption, as does diarrhea.lsJ9 Cholesterol-binding agents bind digoxin in the gut and inhibit absorption. 2o Digoxin malabsorption can occur with most small bowel disorders, including the edematous bowel of patients with chronic CHF.21>22In all of these settings the Lanoxicap formulation of digoxin is more bioavailable and is, therefore, recommended.15 Up to 10% of the population in the United States carries the intestinal bacterium Eubacterium lentum, which hydrolyzes digoxin to the inactive metabolite dihydrodigoxin.23 Curiously, this bacterium is far more common in urban dwellers.24 Antibiotic therapy in these patients may abruptly increase the serum digoxin concentration.25 Omeprazole increases bioavailability by inhibiting the breakdown of digoxin by gastric acid.26 The distribution of digoxin is affected by a number of physiologic and pathophysiologic factors.4,24J27,28 Binding of digoxin to the sodiumpotassium adenosine triphosphatase (Na+,K+-

ATPase) receptor is directly related to serum potassium concentrations: it is enhanced in patients with hypokalemia and inhibited in those with hyperkalemia. Since this binding is reversible, potassium administration may cause unbinding. Fifty percent of digoxin in the body is bound to the Na+,K+-ATPase receptor in skeletal muscle.29>30 Physical activity increases binding and can produce a substantial lowering of the serum digoxin concentration (by up to 33%). Older individuals have relatively fewer binding sites for digoxin. Pregnancy and hyperthyroidism each increase the 100

IO .-s H F

8-

8 siOE

6-

80 IV IV --------_-__ --_____ ----------

60 40 20

I

0

I

I

1

3

6

9

12

Time (hr) RGURE 9. Plot of simulated serum dlgoxln concentration (solid lhes) and response intensity (dashed fines) as a function of time comparbrg a shrgle oral dose (0.75 mg) and an intravenous dose of equhmlent bloavallablllty (0.6 mg). (Reprinted wlth permission from J C/in PharnmcoL~)

A SYMPOSIUM:

CONGESTIVE

HEART

FAILURE

1010

--__ -._ t 60

24

0

48 TIME (Hr)

72

number of digoxin binding sites, whereas hypothyroidism decreases the number. Digoxin is not bound by adipose tissue, so the dose should be based on lean body mass.31Amiodarone may affect digoxin’s volume of distribution and therefore increase the serum concentration.32 Renal failure produces an increased serum digoxin concentration; it is unclear whether this represents a diminished volume of distribution for digoxin or an accumulation of hepatic digoxin metabolites.15 Digoxin elimination is predominantly renal, although newer evidence indicates hepatic metabolism is more extensive than was previously believed.15,26When renal function is reduced (e.g., via intrinsic renal disease, aging, excessivediuretics, or severe heart failure), there is more extensive hepatic metabolism and fecal excretion. Impaired renal elimination of digoxin can be induced by indomethacin, cyclosporine, or spironolactone, apparently by their direct effects on the kidney.33,34 Quinidine inhibits renal excretion of digoxin.35,36 Other antiarrhythmic agents also increase serum

t

I

0

102G

THE AMERICAN

I

48

JOURNAL

I

I

I

96 TIME (Hr)

OF CARDIOLOGY

VOLUME

96

digoxin concentrations, but the mechanisms are not exclusively renal. 3740Vasodilators, by increasing renal blood flow, can enhance renal excretion. Since the increased serum digoxin concentrations reflect biologically active compounds in these settings, the dose should be reduced, generally by half.41>42 Obviously, measurement of serum digoxin concentrations is critical to its safe and proper use. DIGOXIN

DOSING

In the early 1970s population-based nomograms were advocated for digoxin dosing.43 Because of marked individual variability, however, these nomograms did not achieve widespread acceptance. An alternative approach has been to select an empiric dose based on parameters like body mass, renal function, and drug interactions, and subsequently monitor serum digoxin concentrations. The goal is to achieve a desired body pool of 10 Fg/kg or a serum digoxin concentration of approximately 1.5 ng/mL.4J5 A 75% absorption of tablets or 90% absorption of Lanoxicaps is as-

I

1

FIGURE 10. Plot of simulated serum digoxhr concentration (so/M /ines) and response hrtensRy (dashefi /ines) as a function of time comparing protocols wRh and without an oral ioadhtg dose in normal subjects. Regimen I: 0.75 mg at 0 hours, 0.25 mg at 6, l2, and 18 hours, and 0.37s mg at 24,4S, and 72 hours; regimen 2: 0.375 mg at 0,24,48, and 72 hours. (Reprinted with permission from / C/in PhamracoLi4)

144

192

69

JUNE 4, 1992

FIGURE il. Simulation of the effect of Interrupting a muRipledoss digoxin regimen by omission of several doses. An oral regimen of 0.375 mg/day was given until steady state was attained and then stopped. Only resutts from the last 2 doses are shown. Simulations are for serum digoxin concentration (so/id /Me) and response (c/as/red line) as a function of time. (Reprinted with permission from I C/in PhamaCOI.~)

sumed. A baseline serum digoxin concentration should be obtained to determine the presence of digitalis-like immunoreactive substances, as will be discussed later. In many instances, a loading dose of digoxin is not required.44 A maintenance dose of 0.25 mg is started when creatinine clearance exceeds 20 mL/ min; when creatinine clearance is less than 20 mL/min, the maintenance dose is 0.125 mg. The serum digoxin concentration is measured 2 days later. This point generally represents 50% of the eventual steady-state concentration. If an exceedingly high or exceedingly low serum digoxin concentration is present, the maintenance dose can be adjusted accordingly. A repeat concentration is then obtained 5 half-lives after a dosing change. If a loading dose is preferred, the serum digoxin concentration should be measured 12 hours after the last loading dose. This will indicate whether the desired body pool has been achieved. The maintenance dose is then based on predicted half-life, and the serum digoxin concentration is remeasured in 2 days, as disc&sed above.

TABLE II Digoxin

Parameters

Total body clearance

180 mL/min/m*

Renal

140

clearance

Bioavailability Tablets Elixir Lanoxicaps

mL/min/m2 60-75% 75-85% 80-90%

Percent excreted in urine as native drug Intravenous dose Oral dose

72% 54%

Half-life Serum Tissue

36 hours 44 hours

INDICATIONS FOR SERUM DlGOXlN LEVELS Table IV lists the indications for the serum digoxin RIA. These include validating the dose, assessing patient compliance, evaluating bioavailability, and assessing the effects of renal function or of other drugs on elimination.s0 The assay also may be used when there has been a hemodynamic or clinical change (e.g., worsening heart failure, TABLE Ill

DlGOXlN ASSAYS Most clinically used digoxin assays are RIAs.~~>~~ These are sensitive to levels of 0.2-0.4 ng/mL and have a coefficient of variance of 5-15%. Errors are related to differences in assay kits as well as laboratory technique. It is, therefore, essential for the clinician to know the reproducibility of serum digoxin concentrations in the laboratory. Unfortunately, the RIA has relatively low specificity and cross-reacts with digitalis-like immunoreactive substances, which may be present in renal failure, hepatic failure, pregnancy, some hypertensive states, and CHF.4Js>28,46>47Whether these substances are truly digitalis is still unclear, but they can comprise up to 50% of the serum digoxin measured by RIA. Fortunately, most active digoxin metabolites, including digoxigenin and bisdigoxiside and monodigoxiside, react with most assays. Dihydrodigoxin, which is inactive, does not crossreact with most assays, but is usually present at low concentrations in the serum. Other assay methods include the enzymatic immunoassay and fluorescent polarization immunoassay. These have sensitivities and specificities similar to those of the RIA. The most specific assay for digoxin is that obtained with high-performance liquid chromatography. This method is tedious, however, and not currently available for clinical use.48 A promising new approach is to use digoxinspecific monoclonal antibodies.49

Pharmacokinetic

Digoxin

Pharmacokinetic

Interactions

Bioavailability Decreased Cathartics Antacids Cholesterol-binding agents Malabsorption syndromes Bowel edema Eubacterium lentum Gastric hyperacidity increased Lanoxicaps Antibiotics Omeprazole

or elixir LE. lenturn)

Distribution Decreased Renal failure Hyperkalemia Aging Hypothyroidism Amiodarone increased Hypokalemia Hyperthyroidism Pregnancy Physical activity Elimination Decreased Renal failure Excessive diuretics Aging lndomethacin Cyclosporine Spironolactone Verapamil Quinidine Propafenone Increased Diarrhea Vasodilators

A SYMPOSIUM:

CONGESTIVE

HEART

FAILURE

1036

TABLE IV Indications for Serum Radioimmunoassay of Digoxin Validate

dose

Assess

compliance

Assess

bioavailability

Assess

effect of renal function

Assess

effect of other

Assess

effect of hemodynamic

Assess

patients

Prevent

toxicity

Diagnose

on elimination

pharmacologic

agents

on elimination

change

who fail to respond

toxicity

acute myocardial infarction), apparent failure to respond to digoxin, or in the diagnosis and prevention of digoxin toxicity. Once steady state has been reached, measurement more than annually is seldom required. Computer-assisted surveillance of digoxin dosing has been advocated to reduce toxicity and improve efficacy.51-55Although this approach appears to have merit, it remains a “hypothesis in need of testing.“56 A major error in the use of serum digoxin assays is the failure to wait for equilibration, which requires 12 hours following an oral dose. In our institution, digoxin is administered in the evening so that the morning blood sample is taken 12 hours later. Because there can be a substantial difference between the serum digoxin concentration at 8 hours and at 24 hours postdose, a standard collection time should be specified and implemented. In outpatients there should be at least a l-hour rest period prior to obtaining the sample, since physical activity can reduce the blood leve1.57When ordering a digoxin assay, the clinician must always I

I

0

0.5

1.0

1.5

2.0

2.5

I

consider whether a steady state has been reached, and remember that this takes 5 half-lives (range, 10-20 days) following any change in the dose. The digoxin assayis one of the most widely used drug assaysand represents a considerable medical care cost.58Unfortunately, a significant percentage of digoxin assaysare not necessary and generally reflect a poor understanding of digoxin pharmacokinetics. This is clearly an area where education (and perhaps computer surveillance) could produce both cost savings and improved digoxin usage. Finally, there is the question of the role of the serum digoxin concentration to assessthe clinical response to digoxin. Traditional clinical parameters (slowing of heart rate, initiation of diuresis, improvement in dyspnea, disappearance of an S3 gallop, improvement in chest radiograph, etc.) used to assessresponse to digoxin continue to be most important. When atria1 fibrillation is present, the ventricular rate is used as a therapeutic end point. As noted elsewhere in this symposium, however, the usual therapeutic concentration of digoxin may not always be sufficient to slow ventricular rate, especially in acutely ill patients. In such cases,adding another agent is generally preferred over increasing the serum digoxin concentration to the toxic range. When a patient in sinus rhythm has had a suboptimal clinical response to chronic digoxin administration, the serum concentration is uniquely important for determining whether the dose should be increased. Several studies 59-62have suggested that patients with steady-state serum digoxin concentrations < 1 ng/mL show subjective and objective improvement when the serum concentration is raised closer to 1.5 ng/mL. Because of the current practice to initiate therapy with 0.25 mg (rather than 0.375 mg) digoxin in all cases, a significant number of heart failure patients have sustained subtherapeutic serum digoxin concentrations. Figure 12 schematically illustrates the relation between serum digoxin concentration, physiologic effect, and toxicity.4 It is based on empiric observations of digitalis toxicity and the dose-response curve of digoxin. If a steady-state therapeutic concentration between 1.0 and 1.5 ng/mL is achieved, then near-maximal digoxin effect will be achieved with a low risk of toxicity.

Serum digoxin level ng/mL)

REFERENCES RGURE l2. Schemetk illustration of the reletlon between the therapeutic effect of digoxln, the toxic effect, and the serum digoxin concentration. Above a seam digoxin level of2.0 rig/ml, there Is minimal further therapeotk effect but a drametk Increase in the toxk effect. (Reprinted with permission from Cardiotonfc Dnrgs.4) 1044

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

69

l. Smith TW, Butler VP Jr, Haber E. Determination of therapeutic and totic semm dig&n concentrations by immunoassay. N Engl J Med 1969;281:12121217. 2. Belier GA, Smith TW, Abelmann W Haber E, Bcod WB Jr. Digitalis intoxication: a prospective clinical study with serum level correlations. N I?& J Mea 1971;284:989-997.

JUNE

4, 1992

3. Mahdyoon H, Battilana G, Rosman H, Goldstein S, Gheorghiade M. The evolving pattern of digoxin intoxication: observations at a large urban hospital from 198%1988.Am Heart J 1990;120:1189-1194. 4. Lewis RP. Digitalis. In: Leier CV, ed. Cardiotonic Drugs: A Clinical Sutvey. New York: Marcel Dekker, 1991;107-177. 5. Weissler AM, Snyder JR, Schoenfeld DC, Cohen S. Assay of digitalis glycosides in man. Am J Cardiol 1%6;17:768-780. 6. Weissler AM, Lewis RP, Leighton RF, Bush CA. Comparative responses to the digitalis glycosides in man. In: Marks BH Weissler AM, eds. Basic and Clinical Pharmacology of Digitalis. Springfield, IlIz Charles C Thomas, 1972;26& 280. 7. Forester W, Lewis RP, Weissler AM, Wiie TA The onset and magnitude of the contractile response to commonly used digitalis glycosides in normal subjects. Circulation 1974;49:517-521. 6. Lewis RP. Use of systolic time intervals for evaluation of left ventricular function. Cardiovasc C[in 1983;13:335-353. 9. Gold H, Camel1 McK, Greiner T, Hanlon LW, Kwit NT, Model1 W, C&love E, Benton J, Otto HL. Clinical pharmacology of digoxin. J Phannacol Elp Ther 1953;109:45-57. 10. Powell AC, Horowitz JD, Hasin Y, Syrjanen ML, Horomidis S, Louis WJ. Acute myocardial uptake of digoxin in humans: correlation with hemodynamic and electrocardiographic effects. JAm Coil Car&l 1990;15:123&1247. il. Weissler AM, Schoenfeld CD. Effect of digitalis on systolic time intervals in heart failure. Am J Med 1970;259:4-20. 12. Kramer WG, Kolibash AI, Lewis RP, Bathala MS, Viiconti JA Reuning RH. Pharmacokinetics of digoxim relationship between response intensity and predicted compartmental drug levels in man. JPharmacohinet Biophavn 1979;7: 4761. W. Kramer WG, Lewis RP, Cobb TC, Forester WF, Visconti JA, Wanke LG Boxenbaum HG, Reunning RH. Pharmacokinetics of digoxin comparison of a two- and a three-compartment model in man. J Pharmacokinet Biophann 1974;2:299-312. 14. Kolibash AJ Jr, Lewis RP, Bourne DW, Kramer WG, Reunning RH. Extension of the serum digoxin concentration-response relationship to patient management. J Clin Pharmacol1989;29:300-306. is. Reunning RH, Geraets DR. Digoxin. In: Evans WE, Schentag JJ, Juske WF, eds. Applied Pharmacokinetics: Principles of Therapeutic Drug Monitoring, 2nd ed. Vancouver, WA Applied Therapeutics, 1986;[email protected] l6. Lmdenbaum J, Mellow MH Blackstone MO, Butler VP Jr. Variation in biologic availability of digoxin from four preparations. N En& J Med 1971;285: 1344-1347. 17.Astorri E, Bianchi G, LaCamra G, Assanelli D, Visioli 0, Marzo A. Bioavailability and related heart function index of digoxin capsules and tablets in cardiac patients. J Pharm Sci 1979;68:104-1% is. Brown DD, Juhl RP. Decreased bioavailabihty of digoxin due to antacids and kaolinpectin. N Engl J Med 1976;295:1034-1037. 19. Kolibash AJ, Kramer WG, Retmning RH, Caldwell JH. Marked decline in serum digoxin concentrations during an episode of severe diarrhea. Am Heart J 1977;94:806807. 20. Brown DD, Juhl RP, Warner SL. Decreased bioavailability of digoxin due to hypocholesterolemic interventions. Circulation 1978;58:164-172, 21. Heizer WD, Smith TW, Goldlinger SE. Absorption of digoxin in patients with malabsorption syndromes. N Engl J Med 1971;285:257-259. 22. Greenberger NJ, Caldwell JH. Studies on the intestinal absorptions of 3H-digitalis glycosides in experimental animals and man. In: Marks BH, Weissler AM, eds. Basic and Clinical Pharmacology of Digitalis. Springfield, Ilk Charles C Thomas, 1972~1547. 23. Dobkin JF, Saha JR, Butler VP, Neil HC, Lindenbaum J. Digoxininactivating bacteria: identitication in human gut flora. Science 1983:220:325327. 24. Machan VI, Wiederman J, Dobkin JF, Lindenbaum J. Geographic differences in digoxin activation. A metabolic activity of the human anaerobic gut flora. Gut 1989;30:971-977. 25. Lindenbaum J, Rund DG, Butler VP Jr, Tse-Eng D, Saha JR. Inactivation of digoxin by the gut flora: reversal by antibiotic therapy. N Engl J Med 1981;305:79&794. 26. Cohen AF, Droon R, Schoemaker R, Hoogkamer A, Van Vliet A. IntIuence of gastric acidity on the bioavailability of digoxin. Arut Intern Med 1991;115: 54&546. 27. Marcus FI. Digitalis pharmacokinetics and metabolism. Am J Med 1975;58: 452-459. 26. Smith TW. Digitalis Glycosides. New York: Grune & Stratton, 198&l-348. 29. Hall PD, Garnett WR, Kolb KW, Rock WL, Stanley H. The effect of everyday exercise on steady state digoxin concentrations. J Cl& Phamzacol 1989;29:108>1088.

30. Jogestrand T, Andersson K. The effect of physical exercise on the pharmacokinetics of digoxin during maintenance treatment. J Cardiovasc Phamzacol 1989;14:73-76. 31. Ewy GA, Grwes BM, Ball MF, Nimmo L, Jackson B, Marcus F. Digoxin metabolism in obesity. Circulation 1971;44810814. 32. Robinson K Johnston A, Walker S, Mulrow JP, McKenna WJ, Holt DW. The digoxin-amiodarone interaction. Cardiovcrsc Drugs Ther 1989;3:25-28. 33. Jorgenson HS, Christensen HR, Kampmann JP. Interaction between digoxin and indomethacin or ibuprofen. BrJ Chin Pharmaco11991;31:108-110. 34. Foukaridis GN. Intluence of spironolactone and its metabolite camenone on serum digoxin assays. Ther Drug Monif 19%;12:82-84. 35.Leahey EB Jr, ReifIel JA, Giardina E-GV, Bigger JT Jr. The effect of quinidme and other oral antiarrhythmic drugs on serum digoxin. Ann Intern Med 1980;92:605608.

36. Leahey EB Jr, Bigger JT Jr, Butler VP Jr, Reilfel JA, O’Connell GC, Scaflidi LE, Rottman JN. Quinidiiedigoxin interaction: time course and pharmacokinetics. Am J Cardiol1981;48:1141-1146. 37. Klein HO, Land R, Weiss E, Segni EDI, Libhaber C, Buerrero J, Kaplinsky E. The intluencc of verapamil on serum digoxin concentration. C~rc~&~tion 198&65:9961002. 38. Nademanee K, Kamran R, Hendrickson J, Ookhtens M, Kay I, Smgh BN. Amiodarone-digoxin interaction: clinical significance, time course of develop merit, potential pharmacokinetic mechanisms and therapeutic implications. J Am Coil Cardiol1984,4:111-116. 39. Nolan PE Jr, Marcus FI, Erstad BL, Hoyer GL, Furman C, Kirsten EB. Effects of coadministration of propafenone on the pharmacokinetics of digoxin in healthy volunteer subjects. J Clin Phamuzcol1989;29:46-52. 40. Hedman A, Angelm B, Arvidsson A, Beck 0, Dahlquest R, Nelsson B, Olsson M, Schenck-Gustafsson K. Digoxinverapamil interaction: reduction of biliary but not renal digoxin clearance in humans. C&z Phavnacol Ther 1991;49: 256-262. 41. Warner NJ, Barnard JT, Bigger JT Jr. Tissue digoxin concentrations and digoxin effect during the quinidinedigoxin interaction. J Am Coil Cardiof 1985;5:68&&36. 42. C&in RA, Ashavaid TF, Katy AM, Messineo FC. Amiodarone, verapamil, and quinidine do not affect equilibrium binding of digoxin. J Cardiovasc Ph.armac011990$6:519-522.

43. JeUifIe RW, Brooker G. A nomogram for digoxin therapy. Am J Med 1974;57:63-68. 44. Marcus FI, Burkhalter L, Cuccia C, Pavlovich J, Kapadia GG. Administration of tritiated digoxin with and without a loading dose. Circulation 1966,34:865874. 45. Stone JA, Soldin SJ. An update on digoxin. Clin Chin 1989;35:1326-1331. 48. Stone J, Bentur Y, Zalstein E, Soldin S, Geistrecht E, Karen G. Effect of endogenous digoxin-like substances on the interpretation of high concentrations of digoxin in children. J Pediatr 1990,117:321-325. 47. Dasgupta A, Saldana S, Heiman P. Monitoring free digoxin instead of total digoxin in patients with congestive heart failme and high concentrations of digoxin like immunoreactive substances. Clin C/tern 1990;36:2121-2123. 48. Kwong E, Embell L, McErlane KM. Comparison of analysis of digoxin in cardiac patients by HPLC and RLA. Res Cormnun Chem Path01 Pharmacol 1990;68;121-124. 49. Papanastasion-Diamandi A, Conway K, Diamandis EP. Digoxin immunoassay with monoclonal and polyclonal antibodies using time resolved fluorometry. J Pharm Sci 1989;78:617-621. 50. Okada RD, Hager WD, Graves PE, Mayer&n M, Perrier DG, Marcus FL Relationship between plasma concentration and dose of digoxin in patients with and without renal impairment. Cirnr[ation 1978;58:1196-1203. 51.Sheiner LB, Halkin H, Peck C, Rosenberg B, Melmon KL. Improved computer-assisted digoti therapy: a method using feedback of measured seturn digoxin concentrations. Ann Intern Med 1975;82:619-627. 52. White KS, Lindsay A, Pryor TA, Brown WF, Walsh K. Application of a computerized medical decision-making process to the problem of digoxin intoxication. JAm Coil CardioZ 1984;4:571-576. 53. Weiner F, Groth T, Mottimer 0, Hallquest I, Rane A A knowledge-based information system for monitoring drug levels. Cornput Methods proSrams Biomed 1989;29:115-128. !!4. D’Angio RG, Stevenson RG, Lively BT, Morgan JE. Therapeutic drug monitoring: improved performance through educational intervention. Ther Dung Monit1990$2:73-81. 55. Dobbs RJ, O’NeiU CI, Deshmakh AA, Nicholson PW, Dobbs SM. Serum concentration monitoring of cardiac glycosides. How helpful is it for adjusting dosage regimens? Clin Pharmacokinet 1991;20:175-193. 56. McImtes GT. The value of therapeutic drug monitoring to the practicing

A SYMPOSIUM:

CONGESTIVE

HEART FAILURE

u)50

physician-an hypothesis in need of resting. Br J C/in Amrwcol 1989;27:2812&i. S7. Jogstrand T, Edner M, Haverling M. Clinical value of strum digoxin a&says in outpatients. Am Hem J 1989;117:107~~1083. 58. Fleg JI, Hinron PC, Iakatta EG, Marcus T;I, Smith I’W, Strauss HC, Heatky MA. Physician utilization of laboratory procedures to monitor ourparicnts with congestive heart failure. Arch It&m Md 1989;2:3Y~3!%. 59. Carliner NH, Golberl CA. Pruiu AW, Goldberg Ll. ISects of maintenance digoxin therapy on systolic time intervals and serum digoxin cnnccntralions. Ci~ulolri~~t I Y7k;f;O:~98. 60. Eklr (;G, Erbel R, Shumann K, Gilfrich HJ. Dose-response relationships and plasma concentrations of digiralk in man. EwJ C&J I’humu~cd 1!978;13:1(% III. 81 Buch I-‘, WaldorfT S. Cla.ssical concentration-reyunte relationship between strum digoxin level and contractility indicts. Dan ~Med Bull l%lk27:287-2w. 62. C;ritiiths SF- Penny WJ, Lewis MJ, Henderson All. Maintenance of the inotropic eliect of digoxin on long-term treatment. Br ,kf J 3982;284:181% 1822.

DISCUSSION Dr. Leier: What are your thoughts about cndogenous digitalis-like substances? 1 doubt if these are measurable with the techniques you talked about. Dr. Lewis: They are measurable by the radioimmunoassay. Actually, they have been measured in concentrations of 3-5 ng/mL in neonates. Interestingly, they seem to appear in volume-overload states. Some investigators think that digitalis is a primitive hormone, just as opiates are primitive hormones. Like opium, it is found in plants, but whether the substance we are measuring is a primitive hormone is just not clear. Dr. Leier: Why do we. have digoxin receptors? Do you think it is because of this endogenous digitalis-like substance, or is it another receptor that digoxin happens to borrow? Dr. Lewis: I do not know. I think digitalis may be a primitive hormone. Dr. [email protected] It is a misnomer to say it is a receptor. There is a binding site in one of the subunits that binds cardiac glycosides, but I do not think it is any different from any other highly conserved orientation of a membrane ion-transporting protein. It is not a receptor, it is not hooked to a guaninc protein (G protein). It is not at all like opiate receptors, for example. It is more like calcium channels, to which are bound dihydropyridines that affect cardiac function. Digoxin binds to one conformation of the sodium pump and affects the pumping cycle, but it is not like a rcccptor. Why that conformation has been so highly conserved is the question. That is why people continue to look for endogenous digoxin. An article by Hamlyn and colleagues,’ a reputable group, reports that they have looked at it for many years and believe it is ouabain. Others of us believe that this is just not the case. Dr. McCaU: They are the second group to demonstrate, by high-performance liquid chroma1066

THE AMERICAN

JOURNAL

OF CARDIOLOGY

VOLUME

69

tography and by chemical analysis, that it is ouabain. It had been described by another group previously. Dr. Kelly: There is disagreement, anyway. Dr. [email protected]: In practice, most physicians will not measure serum digoxin concentrations to maximize the dose. Interestingly, approximately 14,000 serum digoxin concentrations were measured at our institution over a 2-year period, and in many were < 0.9 ng/mL.2 Dr. Lewis: It is the same at our institution. Dr. G-ala&: The question then is, are we really optimizing therapy? For example, in the Digoxin-Captopril study, the mean serum digoxin concentration was approximately 0.9 ng/mL. Dr. Lewis: When I was a medical student, I was given a handout that listed digitalizing doses; the doses on that are twice what we use today. The digitalizing dose of digoxin was 3 mg; the maintcnance dose was 0.5 mg. Today, everybody gets 0.25 mg whether they are 280 lb or not. It is amazing. People do not realize that the dose should be based on body size, and the average dose should be 0.375 mii5 Dr. GheorghSade: You are right. It is difficult to assess efficacy of a drug if the right dose is not being used. Dr.Yusd But what is the right dose? b. Gheorghiade: In a recently completed study we have seen a significant increase in rest and exercise ejection fraction in patients with CHF when the serum concentration of digoxin was increased from 0.7 to 1.2 ng/mL. Dr. Yusuf: What is the basis for that statement? I can understand using that dose range in patients with atria1 fibrillation, but what is the clinical basis for using it, say, in cases of heart failure in sinus rhythm? Dr. Gheorm: In a study by Johnston and McDevitt3 in which patients with CHF were evaluated following withdrawal of digoxin therapy, only 1 of the 34 patients with a serum level <0.8 ng/mL deteriorated when digoxin was discontinued. In contrast, 7 of 22 patients with a serum level that was considered therapeutic (0.8-2.0 ng/mL) deteriorated when digoxin was withdrawn. However, I agree with you that there are no good studies to suggest that the digoxin dose for “optimal” clinical and hemodynamic effects is one that provides a serum concentration of 1-2 ng/mL. Dr. Ywuf: Response is not a yes-or-no phenomenon. So there is probably a substantial degree of response, even in patients receiving a dose that gives a serum digoxin concentration of 0.8 ng/mL. There is also very good anecdotal evidence-and it JUNE

4, 1992

is unusual for me to call anecdotal evidence very good-that digoxin toxicity is relatively rare compared with 3 decades ago. This may be explained partially by the fact that we have been using lower doses.4 Dr. Leier: Anecdotally, I do not remember ever improving a patient’s cardiac function by doubling the dose or by increasing the serum digoxin concentration from 0.7 to 1.4 ng/mL, unless the patient had atria1 fibrillation. Dr. Lewis: A number of heart failure patients report to us with serum digoxin concentration <0.4 ng/mL. I do not know whether that is because they are not taking it or because they are not absorbing it due to bowel edema. Personally, I try to keep the serum digoxin concentration around 1.2 ng/mL. Dr. McCall: Maybe the old physicians were right: initiate therapy with toxic levels of digoxin and then taper the dose. Dr. Leier: I would go by the other philosophy, that is, a little bit of digoxin is better than no digoxin, and it is better than too much digoxin. Dr. McCall: What we are talking about is where do you get the maximum effect, whatever that is, for the individual patient? Dr. Gheorghiae: Now that we know more about drug interactions and serum concentrations,

maybe it would be reasonable and safer to increase the dose in patients who continue to have symptoms and signs of heart failure and the serum digoxin concentration is < 0.9 ng/mL. Dr. Lewis: I think that is one of the reasons many people have stated digoxin is a mild inotrope. Just parenthetically, the maximum effect you get with digoxin on systolic intervals is about the equivalent of 2 pg/kg isoproterenol. So, intravenous digoxin is not a mild inotropic drug. Pr. Gheorghiadez Ware et al5 examined 20 older patients, mean age 83.8 years. They reported that ejection fraction increased with digoxin and the maximum improvement in ejection fraction occurred at the level of 1.0 ng/mL. Dr. &ewis: And very rarely produced toxicity.

REFERENCES I. Han&n JM, Blaustein MP, Bow S. Identilication and characterization of a ouabair-like compound from human plasma. I’roc N&l Acad Sci USA 1991;88: 62596263. 2. Gheorgbiade

M, Rosman H, Mahdyoon H, Goldstein S. Incidence of digitalis intoxication. plima~ Cardiol 1988;1:5. 3. Johnston GD, McDevitt DG. Is maintenance digoxin necessary in patients with sinus rhythm? Lancet 1979$567-570. 4. Duhme DW, Greenblatt DJ, Koch-Weser .I. Reduction of digoxin toxicity associated with measurement of serum levels. A report from the Boston Collaborative Drug Surveillance Program. ATVI h&m Med 1974;80:516-519. 5. Ware LA, Snow E, Luchi JM, Luchi R. Effect of digoxin on ejection fraction in elderly patients with congestive heart failure. JAm Geriatr Sot 1984;32:631.

A SYMPOSIUM:

CONGESTIVE

HEART

FAILURE

1070