Inefficacy of “Therapeutic”
Serum Levels of Digoxin in
Controlling the Ventricular Rate in Atrial Fibrillation
STEVEN GOLDMAN, MD PETER PROBST, MD ARTHUR SELZER, MD, FACC KEITH COHN, MD, FACC San Francisco, California
From the Division of Cardiology, Presbyterian Hospital, Pacific Medical Center, San Francisco, Calif. Manuscript accepted September 4, 1974. Address for reprints: Division of Cardiology, Presbyterian Hospital, Pacific Medical Center, P.O. Box 7999, San Francisco, Calif. 94120.
Although therapeutic and toxic serum concentrations of digoxin have been established, there is sparse information permitting correlation of drug level with cilnical effect. This study was undertaken to assess the radioimmunoassay serum digoxin levels in 30 patients with acute atriai fibrillation (38 determinations) and 30 patients with chronic atriai fibriiiation (54 determinations). Those with chronic fibrillation were subdivided into those in clinically stable condition (14 patients), and those seriously iii and in clinically unstable condition (18 patients). Slowing of ventricular rate in patients with stable, chronic atrial fibrillation was accomplished in 10 of 18 instances by “therapeutic” and “subtherapeutic” levels of digoxin (less than 2 ng/mi). Ventricular rate was “controlled” (85 to 85 beats/min) with therapeutic levels of serum digoxin in only five instances of acute atriai fibrillation and seven of unstable chronic atrial fibrillation. In 43 studies (23 of acute atriai fibriiiation, 20 of chronic atrial fibrillation), a rapid ventricular rate (95 to 140 beats/min) persisted in the presence of “therapeutic” or high levels of digoxin. Thirty-nine of these were in patients who were seriously iii with conditions such as infection, hypoxia or recent thoracotomy. Slowing of the ventricular rate required “toxic” concentrations of digoxin (2.5 to 8 ng/mi) in 15 instances. We conclude that sufficient amounts of digoxin to achieve “therapeutic” serum concentrations may fail to lower the ventricular rate in atriai fibrillation to less than 100 beats/min, especially when a serious, complicating illness coexists.
The digitalis glycosides have two principal actions in clinical medicine: (1) increasing myocardial contractility (hence their use in congestive heart failure); and (2) inducing atrioventricular (A-V) block, which serves to slow the ventricular response in the presence of atria1 fibrillation. When the ventricular response is rapid in atria1 fibrillation, the classic therapeutic approach has been to administer incremental doses of a digitalis preparation until the rate is brought to an acceptable level. This study was designed to document several problems inherent in the use of digitalis for atria1 fibrillation and to determine those clinical factors contributing to the inefficacy of digitalis in patients with fibrillation. Our data demonstrate that toxic levels of digitalis may be reached before adequate ventricular slowing is achieved and that under certain circumstances an appropriate degree of A-V block is sometimes unattainable using digitalis glycosides.
May 1975
The American
Journal 01 CARDIOLOGY
Volume
35
651
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
6.0
CHRONIC
1
ATRIAL
FIBRILLATION
ET AL.
-STABLE
so i
\E \ z?4.0 1 i g 3.0 1 n H 2.0 i? z
. .
1.0
II
0'
t
. .
1 I,
50
.
.
.
. .
be I
.
l
.
I
I,,
70
I
110
90 HEART
I,
I
130
150
I‘
170
RATE
FldUR+l. Relation between heart rate and serum digoxin levels in 14 patients in Group IA (chronic fibrillation, clinically stable condition). Dots falling within the box represent those with “therapeutic” serum digoxin concentrations (0.8 to 2.0 ng/ml) in combination with “controlled” ventricular rates (65 to 95/min).
The dose of digoxin given was determined by each patient’s primary physician. Patients in Group IA were receiving daily maintenance doses of digoxin: 0.25 mg in 12 cases, 0.375 mg in 1 case and 0.675 mg in 1. After initial or “booster” digitalization, digoxin was administered incrementally to patients in Groups IB and II. Initial determinations of serum digoxin level and heart rate were made after patients received 0.5 mg to 2.25 mg of digoxin intravenously or orally over 24 to 48 hours, whether or not the ventricular rate had slowed to less than lOO/min. Electrocardiographic rhythm strips were obtained at hourly intervals. Serum digoxin levels were determined by radioimmunoassay technique, and blood samples were drawn at least 5 to 6 hours after the last dose. In an effort to relate serum digoxin level to ventricular rates during incremental digitalization, two or three determinations were made in 13 patients. All patients had normal blood urea nitrogen and serum potassium levels at the time digoxin levels were measured. For purposes of discussion, serum digoxin concentrations of 0.8 to 2 ng/ml were considered to be in the “therapeutic” range. Concentrations above this level are thought to be potentially toxic. lp2A “controlled” ventricular rate was defined as one ranging from 65 to 95/min.
Methods
Results
The 60 hospitalized patients with atria1 fibrillation selected for the study were divided into two groups: Group I was composed of 30 patients with chronic fibrillation and was further subdivided into Group IA, 14 patients in clinically stable condition, and Group IB, 16 patients with other complicating, serious conditions such as postthoracotomy (8), pulmonary disease with hypoxemia (4), sepsis (2) and pulmonary edema (2). In Group II (30 patients) were subjects manifesting acute atria1 fibrillation, all but two having concomitant serious conditions: postthoracotomy (13), myocardial infarction with angina (5), congestive heart failure (4), sepsis (2) and other (4). None of the patients had A-V nodal conduction disease as evidenced by previous documentation of a ventricular response during the fibrillatory state greater than 120/min within 1 year of the study. None had prior second or third degree A-V block.
CHRONIC
6.0
ATRIAL
FIBRILLATION
-UNSTABLE
Group I-Chronic
Atrial Fibrillation
Group IA (14 patients)
(Fig. 1): Nine of 16 determinations revealed “controlled” ventricular rates of 65 to 95/min with “therapeutic” serum digoxin levels of 0.8 to 2.0 ng/ml in 5 cases and with “subtherapeutic” levels in 4. Four additional determinations demonstrated uncontrolled ventricular rates, greater than 95/min, even after therapeutic serum digoxin levels had been achieved. One patient in this group was considered to have digitalis toxicity, manifesting high serum digoxin levels with a ventricular rate of 40/min. Group IB (16 patients) (Fig. 2): Only 7 of the 38 determinations revealed a controlled ventricular re-
6.0
ACUTE
7 5.0 \ p 4.0 0.0
. .
g 3.0 2.0
a
:
.
0 11
I
50
.
. . .
I,,
70
’
l
i. t..
.
.
... .. .*.. I. : :.: .
*
l
. .
l
.
.
0..
.
. I
90 HEART
I1
I
110
130
I
181
150
170
RATE
FIGURE 2. Relation between heart rate and serum digoxin levels in Group IB (chronic fibrillation with complicating, serious illness).
May 1975
. .
s
.
.
ti 1.0
652
FIBRILLATION
l*
i a 2 2
ATRIAL .
The American Journal of CARDIOLOGY
Volume 35
HEART
RATE
FIGURE 3. Relation between heart rate and serum digoxin levels in Group II (acute atrial fibrillation).
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
sponse associated with serum digoxin levels within the therapeutic range, whereas uncontrolled rates were seen in 12 at this same level. In 5 of 38 measurements, levels of digoxin between 2.0 and 5.0 ng/ml were required to control the ventricular response, and in 4, high serum digoxin levels were associated with a continued rapid ventricular response. Group II: Acute Atrial Fibrillation
Thirty-eight determinations were made in the 30 patients in Group II (Fig. 3). Eighteen revealed therapeutic digoxin levels with persistently uncontrolled ventricular rates; only five determinations within the therapeutic zone yielded controlled ventricular rates. In nine instances, high, potentially toxic levels of digoxin were required to slow the ventricular rate and, in four, similar levels of digoxin were ineffective in bringing about adequate slowing. All but two of these patients had complicating illnesses. Discussion Pharmacokinetics
of Digitalis
The results clearly demonstrate that levels of digoxin generally considered therapeutic are often ineffective in controlling the ventricular rate during atrial fibrillation. To evaluate these observations and place them in proper perspective, we should review several aspects of the pharmacokinetics of digitalis and the techniques for measuring serum concentration, and then attempt to explain those clinical and physiologic features that act upon the A-V node and influence digitalis activity. Measurement of serum digoxin levels: That the radioimmunoassay method of measuring serum digoxin is unusually sensitive and specific is beyond question; and this technique is capable of measuring levels as low as 0.2 ng/ml. The rationale for obtaining blood samples for digoxin at least 5 to 6 hours after administration of the last dose is based on data from several previous pharmacokinetic studies using tritium-labeled digoxin, the rubidium method and radioimmunoassay. 1-7 Immediately after intravenous digitalization and within 1 to 2 hours of oral ingestion, blood levels of digoxin begin to fall rapidly, a change related to tissue distribution and intracellular drug migration. Within 4 to 5 hours, however, the rate of reduction in digoxin blood levels is precipitously slowed, and the second phase of the disappearance curve is initiated, related primarily to renal excretion. The rate of decline of digoxin level is considerably diminished during this second, excretory phase (half-life approximately 33 hours in the absence of renal failure), and equilibrium between plasma and cellular sites presumably has taken place, so that concentrations of digoxin obtained from the serum after 4 hours are likely to be related to the clinically germane activities of the drug. A parallel relation between serum digoxin levels and the time
ET AL.
course of inotropic effect measured by systolic time intervals has, in fact, been demonstrated after 3 to 4 hours have elapsed.6 Therapeutic vs. toxic serum digoxin levels: Definition of “therapeutic” and “toxic” levels of serum digoxin has been provided by Smith, Haber and their co-workers and others.‘,295 Their work demonstrated that approximately 90 percent of patients who have no evidence of digitalis toxicity manifest serum digoxin levels of 0.8 to 2 ng/ml, whereas a similar proportion of patients with clearly defined toxicity have serum levels greater than 2 ng/ml. Those patients were considered to have toxicity if they manifested supraventricular tachycardia with A-V block, frequent or multifocal premature ventricular contractions, ventricular tachycardia or A-V block and if the arrhythmias disappeared after discontinuation of digoxin. While an overlap between toxic and nontoxic groups was also demonstrated, serum digoxin levels above 3 ng/ml carried an unusually high probability of digitalis-induced rhythm disturbances. In light of these findings, it has been a recent clinical teaching that digoxin levels above 2 or 2.5 ng/ml are potentially dangerous and should, if possible, be avoided. Our study indicates that these “safe, therapeutic” levels are, in many instances, insufficient to evoke an adequate A-V blocking (negative dromotropic) effect; that is, a sizable number of patients manifested persistently rapid ventricular rates although appropriate serum digoxin levels were achieved. Even drug concentrations of 2 to 5 ng/ml did not bring about a sufficient degree of A-V block in eight instances and, of further note, failed to evoke signs or symptoms of digitalis toxicity. In all cases, we sought subtle evidence of toxicity, including gastrointestinal or mental dysfunction, ventricular extrasystoles, slight regularization of the ventricular response, progressive shortening of the R-R interval or group beating (suggesting forms of junctional Wenkebach phenomenon), periodic appearance of junctional escape beats and undue slowing of the ventricular rate, but such evidence was detected in only one case. Clinical Therapeutic
Implications
Several potentially useful clinical observations and conclusions can be drawn from this study. Patients in clinically stable condition: Levels of serum digoxin generally considered to be in the therapeutic, safe or nontoxic zone-below 2 nglml-are, in the majority of instances, sufficient to control the ventricular rate adequately in patients in clinically stable condition with chronic atria1 fibrillation. 2. Patients in clinically unstable condition: Under clinically unstable or precarious conditions (Groups IB and II), such as the postoperative state or conditions with fever, infection, pulmonary embolism or acute myocardial infarction, a ventricular response of less than 95/min is often not achievable in spite of therapeutic levels of digoxin. Approximately two
May 1975
The American Journal of CARDIOLOGY
Volume 35
653
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
ET AL.
thirds of patients having serious, complicating illnesses had inordinately rapid ventricular rates in the presence of therapeutic levels of digoxin, and in eight instances toxic concentrations were, in fact, ineffective in slowing the ventricular rate. In the previous studies relating serum digoxin levels with ventricular response in patients with atria1 fibrillation,3-5 there was little correlation between ventricular response and serum digoxin levels. However, in none of these studies were the patients separated on the basis of clinical status or duration of arrhythmia. The explanation for the lack of correlation between blood level and drug effect probably lies in other factors that influence A-V nodal conduction, particularly the amounts of circulating and neuronally mediated catecholamines. It is evident from previous studies that A-V nodal conduction is influenced by a variety of factors, including intrinsic A-V blockade, catecholamines, the parasympathetic system and the digitalis glycosides. It is enhanced by beta adrenergic stimulating catecholamines and slowed by the digitalis glycosides’ parasympathetic sensitizing activity and by their direct blocking effects on the A-V node. In the presence of serious, unstable clinical states, parasympathetic tone probably decreases, whereas hyperadrenergic activity and release of catecholamines becomes prominent. These factors tend to blunt and even override the A-V blocking activity of digoxin so that the ventricular rate in the presence of atria1 fibrillation remains rapid. The interplay between catecholamines and digitalis has been demonstrated in the experimental dog model of Ogden et at the nodal level beal.,s in which a competition tween beta stimulators and digitalis was shown, and the A-V blocking effects of digitalis were diminished in the presence of high circulating levels of beta adrenergic stimulators. The clinically important information emerging from these findings is that, in the presence of severe unstable clinical conditions, the effects of digitalis on the A-V node may be minimal and the drug remain incapable of normalizing the ventricular rate. This is not to say that it is necessarily desirable to decrease the ventricular rate below lOO/min, since even sinus tachycardia is a normal compensatory response, considered appropriate and even beneficial in patients with acute illnesses. Thus, the point is not that the ventricular rate need be slowed to below lOO/min, but rather that “therapeutic” levels of digoxin are incapable of doing so, and were excessive doses of the drug thoughtlessly employed, toxicity could eventuate. 3. Serum digoxin levels and signs of toxicity: Of additional note is the observation that high serum concentrations of digoxin-between 2.5 and 5 ng/ ml-failed in most instances to elicit gastrointestinal or cardiovascular signs or symptoms of digitalis toxicity. The explanation for the apparent disparity between the data gathered from these patients and the
654
May 1975
The American Journal of CARDIOLOGY
results reported by Smith et a1.l’ may reside in the populations studied. Two thirds of the patients from the Massachusetts General Hospital investigation had sinus rhythm and were receiving digitalis for congestive heart failure, ischemic heart disease, myocardial disease or pulmonary dysfunction. Patients with these conditions are considered among those most sensitive to the toxic effects of digitalis and, therefore, most prone to the development of digitalisinduced arrhythmias. The patients in our group with atria1 fibrillation probably as a whole had less myocardial disease and were thus at less risk of development of digitalis intoxication. This point is brought home cogently in a review of previous experience9 with digitalis-related suicidal attempts or poisoning in children which demonstrated that drug toxicity was relatively mild in face of ingestion of huge amounts of a digitalis preparation. A second factor influencing the low frequency of digitalis toxicity seen herein may relate to the initial rapid ventricular rates; such rates tend to suppress premature ventricular contractions or runs of ventricular tachycardia by overdriving the heart and, thus, a persistently rapid ventricular rate is, in effect, antiarrhythmic” and capable of masking early signs of digitalis-induced arrhythmias. 4. Serum digoxin levels as a reflection of myocardial concentration: Finally, there are many imponderables in our knowledge of the use of digitalis and digoxin blood levels. Just how closely serum levels reflect drug activity that is taking place either in the membrane or within the myocardial cell has never been adequately clarified. Certainly, it has been clearly shown that no simple relation exists between intra- and extracellular concentrations of the drug. Moreover, a marked inhomogeneity in concentrations of digoxin within different locations in the same heart has previously been demonstrated.‘l Although ‘Yoxic” and “nontoxic” concentrations of the drug have been defined these definitions represent mere averages and do not relate to the clinically relevant actions of the drug-inotropism and A-V blocking activity. A significant number of patients without digitalis toxicity (10 percent) have concentrations in excess of 2.0 ng/ml, usually in the range of 2 to 3 ng/ ml, and a significant number of patients with clinical evidence of toxicity have levels in the “normal” range of 1.4 to 2.0 ng/ml.i2 Most important, the time-tested procedure of “pushing” digitalis until the ventricular rate is adequately slowed in atria1 fibrillation must be decried since, in many clinical situations, potentially toxic and dangerous levels of the drug may build up. In clinical practice, if judicious amounts of digitalis do not effectively slow ventricular rate, other forms of therapy such as treatment of the underlying complicating condition, administration of antiarrhythmic agents, use of propranolol to elicit A-V blockade, or cardioversion seem to be rational alternatives to overdigitalization.
Volume 35
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
ET AL.
References I. Smith TW, Butler VP Jr, Haber E: Determination of therapeutic and toxic serum digoxin concentrations by radioimmunoassay. N Engl J Med 281:1212-1216, 1969 2. Smith TW, Haber E: Digoxin intoxication: the relationship of clinical presentation to serum digoxin concentration. J Clin Invest 49:2377-2386, 1970 3. Chamberlain DA, White RJ, Howard MR, et al: Plasma digoxin concentrations in patients with atrial fibrillation. Br Med J 32: 429-432, 1970 4. Redfors A: Plasma digoxin concentration-its relation to digoxin dosage and clinical effects in patients with atrial fibrillation. Br Heart J 34:383-391, 1972 5. Zener JC, Anggard EE, Harrison DC, et al: Persistence of digoxin effect in atrial fibrillation. JAMA 224:239-241, 1973 6. Shapiro W, Harahara K. Taubert K: Relationship of plasma digitoxin and digoxin to cardiac response following IV digitalization in man. Circulation 42:1065-1072. 1970
7. Doherty JE, Perkins WH, Mltchell GK: Tritiated digoxin studies in human subjects. Arch Intern Med 108:531-539, 1961 8. Ogden PC, Seizer A, Cohn KC: The relationship between inotropic and dromotropic effects of digitalis: the modulation of these effects by autonomic influences. Am Heart J 77:628-635, 1969 9. Smith TW, Wlllerson JT: Suicidal and accidental digoxin ingestion. Circulation 49:29-36, 1971 10. Cohn K, Agmon J, Gamble 0: The effect of glucagon on arrhythmias due to digitalis toxicity. Am J Cardiol 25:683-689, 1970 11. Doherty JE, Perkins WH, Flanigan WJ: Distribution and concentration of tritiated digoxin in human tissues. Ann Intern Med 66:116-124, 1966 12. Donoso E: Letter of the Council on Clinical Cardiology, American Heart Association 1:3, 1974
May 1975
The American Journal of CARDIOLOGY
Volume 35
655
Serum Levels of Digoxin in
Controlling the Ventricular Rate in Atrial Fibrillation
STEVEN GOLDMAN, MD PETER PROBST, MD ARTHUR SELZER, MD, FACC KEITH COHN, MD, FACC San Francisco, California
From the Division of Cardiology, Presbyterian Hospital, Pacific Medical Center, San Francisco, Calif. Manuscript accepted September 4, 1974. Address for reprints: Division of Cardiology, Presbyterian Hospital, Pacific Medical Center, P.O. Box 7999, San Francisco, Calif. 94120.
Although therapeutic and toxic serum concentrations of digoxin have been established, there is sparse information permitting correlation of drug level with cilnical effect. This study was undertaken to assess the radioimmunoassay serum digoxin levels in 30 patients with acute atriai fibrillation (38 determinations) and 30 patients with chronic atriai fibriiiation (54 determinations). Those with chronic fibrillation were subdivided into those in clinically stable condition (14 patients), and those seriously iii and in clinically unstable condition (18 patients). Slowing of ventricular rate in patients with stable, chronic atrial fibrillation was accomplished in 10 of 18 instances by “therapeutic” and “subtherapeutic” levels of digoxin (less than 2 ng/mi). Ventricular rate was “controlled” (85 to 85 beats/min) with therapeutic levels of serum digoxin in only five instances of acute atriai fibrillation and seven of unstable chronic atrial fibrillation. In 43 studies (23 of acute atriai fibriiiation, 20 of chronic atrial fibrillation), a rapid ventricular rate (95 to 140 beats/min) persisted in the presence of “therapeutic” or high levels of digoxin. Thirty-nine of these were in patients who were seriously iii with conditions such as infection, hypoxia or recent thoracotomy. Slowing of the ventricular rate required “toxic” concentrations of digoxin (2.5 to 8 ng/mi) in 15 instances. We conclude that sufficient amounts of digoxin to achieve “therapeutic” serum concentrations may fail to lower the ventricular rate in atriai fibrillation to less than 100 beats/min, especially when a serious, complicating illness coexists.
The digitalis glycosides have two principal actions in clinical medicine: (1) increasing myocardial contractility (hence their use in congestive heart failure); and (2) inducing atrioventricular (A-V) block, which serves to slow the ventricular response in the presence of atria1 fibrillation. When the ventricular response is rapid in atria1 fibrillation, the classic therapeutic approach has been to administer incremental doses of a digitalis preparation until the rate is brought to an acceptable level. This study was designed to document several problems inherent in the use of digitalis for atria1 fibrillation and to determine those clinical factors contributing to the inefficacy of digitalis in patients with fibrillation. Our data demonstrate that toxic levels of digitalis may be reached before adequate ventricular slowing is achieved and that under certain circumstances an appropriate degree of A-V block is sometimes unattainable using digitalis glycosides.
May 1975
The American
Journal 01 CARDIOLOGY
Volume
35
651
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
6.0
CHRONIC
1
ATRIAL
FIBRILLATION
ET AL.
-STABLE
so i
\E \ z?4.0 1 i g 3.0 1 n H 2.0 i? z
. .
1.0
II
0'
t
. .
1 I,
50
.
.
.
. .
be I
.
l
.
I
I,,
70
I
110
90 HEART
I,
I
130
150
I‘
170
RATE
FldUR+l. Relation between heart rate and serum digoxin levels in 14 patients in Group IA (chronic fibrillation, clinically stable condition). Dots falling within the box represent those with “therapeutic” serum digoxin concentrations (0.8 to 2.0 ng/ml) in combination with “controlled” ventricular rates (65 to 95/min).
The dose of digoxin given was determined by each patient’s primary physician. Patients in Group IA were receiving daily maintenance doses of digoxin: 0.25 mg in 12 cases, 0.375 mg in 1 case and 0.675 mg in 1. After initial or “booster” digitalization, digoxin was administered incrementally to patients in Groups IB and II. Initial determinations of serum digoxin level and heart rate were made after patients received 0.5 mg to 2.25 mg of digoxin intravenously or orally over 24 to 48 hours, whether or not the ventricular rate had slowed to less than lOO/min. Electrocardiographic rhythm strips were obtained at hourly intervals. Serum digoxin levels were determined by radioimmunoassay technique, and blood samples were drawn at least 5 to 6 hours after the last dose. In an effort to relate serum digoxin level to ventricular rates during incremental digitalization, two or three determinations were made in 13 patients. All patients had normal blood urea nitrogen and serum potassium levels at the time digoxin levels were measured. For purposes of discussion, serum digoxin concentrations of 0.8 to 2 ng/ml were considered to be in the “therapeutic” range. Concentrations above this level are thought to be potentially toxic. lp2A “controlled” ventricular rate was defined as one ranging from 65 to 95/min.
Methods
Results
The 60 hospitalized patients with atria1 fibrillation selected for the study were divided into two groups: Group I was composed of 30 patients with chronic fibrillation and was further subdivided into Group IA, 14 patients in clinically stable condition, and Group IB, 16 patients with other complicating, serious conditions such as postthoracotomy (8), pulmonary disease with hypoxemia (4), sepsis (2) and pulmonary edema (2). In Group II (30 patients) were subjects manifesting acute atria1 fibrillation, all but two having concomitant serious conditions: postthoracotomy (13), myocardial infarction with angina (5), congestive heart failure (4), sepsis (2) and other (4). None of the patients had A-V nodal conduction disease as evidenced by previous documentation of a ventricular response during the fibrillatory state greater than 120/min within 1 year of the study. None had prior second or third degree A-V block.
CHRONIC
6.0
ATRIAL
FIBRILLATION
-UNSTABLE
Group I-Chronic
Atrial Fibrillation
Group IA (14 patients)
(Fig. 1): Nine of 16 determinations revealed “controlled” ventricular rates of 65 to 95/min with “therapeutic” serum digoxin levels of 0.8 to 2.0 ng/ml in 5 cases and with “subtherapeutic” levels in 4. Four additional determinations demonstrated uncontrolled ventricular rates, greater than 95/min, even after therapeutic serum digoxin levels had been achieved. One patient in this group was considered to have digitalis toxicity, manifesting high serum digoxin levels with a ventricular rate of 40/min. Group IB (16 patients) (Fig. 2): Only 7 of the 38 determinations revealed a controlled ventricular re-
6.0
ACUTE
7 5.0 \ p 4.0 0.0
. .
g 3.0 2.0
a
:
.
0 11
I
50
.
. . .
I,,
70
’
l
i. t..
.
.
... .. .*.. I. : :.: .
*
l
. .
l
.
.
0..
.
. I
90 HEART
I1
I
110
130
I
181
150
170
RATE
FIGURE 2. Relation between heart rate and serum digoxin levels in Group IB (chronic fibrillation with complicating, serious illness).
May 1975
. .
s
.
.
ti 1.0
652
FIBRILLATION
l*
i a 2 2
ATRIAL .
The American Journal of CARDIOLOGY
Volume 35
HEART
RATE
FIGURE 3. Relation between heart rate and serum digoxin levels in Group II (acute atrial fibrillation).
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
sponse associated with serum digoxin levels within the therapeutic range, whereas uncontrolled rates were seen in 12 at this same level. In 5 of 38 measurements, levels of digoxin between 2.0 and 5.0 ng/ml were required to control the ventricular response, and in 4, high serum digoxin levels were associated with a continued rapid ventricular response. Group II: Acute Atrial Fibrillation
Thirty-eight determinations were made in the 30 patients in Group II (Fig. 3). Eighteen revealed therapeutic digoxin levels with persistently uncontrolled ventricular rates; only five determinations within the therapeutic zone yielded controlled ventricular rates. In nine instances, high, potentially toxic levels of digoxin were required to slow the ventricular rate and, in four, similar levels of digoxin were ineffective in bringing about adequate slowing. All but two of these patients had complicating illnesses. Discussion Pharmacokinetics
of Digitalis
The results clearly demonstrate that levels of digoxin generally considered therapeutic are often ineffective in controlling the ventricular rate during atrial fibrillation. To evaluate these observations and place them in proper perspective, we should review several aspects of the pharmacokinetics of digitalis and the techniques for measuring serum concentration, and then attempt to explain those clinical and physiologic features that act upon the A-V node and influence digitalis activity. Measurement of serum digoxin levels: That the radioimmunoassay method of measuring serum digoxin is unusually sensitive and specific is beyond question; and this technique is capable of measuring levels as low as 0.2 ng/ml. The rationale for obtaining blood samples for digoxin at least 5 to 6 hours after administration of the last dose is based on data from several previous pharmacokinetic studies using tritium-labeled digoxin, the rubidium method and radioimmunoassay. 1-7 Immediately after intravenous digitalization and within 1 to 2 hours of oral ingestion, blood levels of digoxin begin to fall rapidly, a change related to tissue distribution and intracellular drug migration. Within 4 to 5 hours, however, the rate of reduction in digoxin blood levels is precipitously slowed, and the second phase of the disappearance curve is initiated, related primarily to renal excretion. The rate of decline of digoxin level is considerably diminished during this second, excretory phase (half-life approximately 33 hours in the absence of renal failure), and equilibrium between plasma and cellular sites presumably has taken place, so that concentrations of digoxin obtained from the serum after 4 hours are likely to be related to the clinically germane activities of the drug. A parallel relation between serum digoxin levels and the time
ET AL.
course of inotropic effect measured by systolic time intervals has, in fact, been demonstrated after 3 to 4 hours have elapsed.6 Therapeutic vs. toxic serum digoxin levels: Definition of “therapeutic” and “toxic” levels of serum digoxin has been provided by Smith, Haber and their co-workers and others.‘,295 Their work demonstrated that approximately 90 percent of patients who have no evidence of digitalis toxicity manifest serum digoxin levels of 0.8 to 2 ng/ml, whereas a similar proportion of patients with clearly defined toxicity have serum levels greater than 2 ng/ml. Those patients were considered to have toxicity if they manifested supraventricular tachycardia with A-V block, frequent or multifocal premature ventricular contractions, ventricular tachycardia or A-V block and if the arrhythmias disappeared after discontinuation of digoxin. While an overlap between toxic and nontoxic groups was also demonstrated, serum digoxin levels above 3 ng/ml carried an unusually high probability of digitalis-induced rhythm disturbances. In light of these findings, it has been a recent clinical teaching that digoxin levels above 2 or 2.5 ng/ml are potentially dangerous and should, if possible, be avoided. Our study indicates that these “safe, therapeutic” levels are, in many instances, insufficient to evoke an adequate A-V blocking (negative dromotropic) effect; that is, a sizable number of patients manifested persistently rapid ventricular rates although appropriate serum digoxin levels were achieved. Even drug concentrations of 2 to 5 ng/ml did not bring about a sufficient degree of A-V block in eight instances and, of further note, failed to evoke signs or symptoms of digitalis toxicity. In all cases, we sought subtle evidence of toxicity, including gastrointestinal or mental dysfunction, ventricular extrasystoles, slight regularization of the ventricular response, progressive shortening of the R-R interval or group beating (suggesting forms of junctional Wenkebach phenomenon), periodic appearance of junctional escape beats and undue slowing of the ventricular rate, but such evidence was detected in only one case. Clinical Therapeutic
Implications
Several potentially useful clinical observations and conclusions can be drawn from this study. Patients in clinically stable condition: Levels of serum digoxin generally considered to be in the therapeutic, safe or nontoxic zone-below 2 nglml-are, in the majority of instances, sufficient to control the ventricular rate adequately in patients in clinically stable condition with chronic atria1 fibrillation. 2. Patients in clinically unstable condition: Under clinically unstable or precarious conditions (Groups IB and II), such as the postoperative state or conditions with fever, infection, pulmonary embolism or acute myocardial infarction, a ventricular response of less than 95/min is often not achievable in spite of therapeutic levels of digoxin. Approximately two
May 1975
The American Journal of CARDIOLOGY
Volume 35
653
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
ET AL.
thirds of patients having serious, complicating illnesses had inordinately rapid ventricular rates in the presence of therapeutic levels of digoxin, and in eight instances toxic concentrations were, in fact, ineffective in slowing the ventricular rate. In the previous studies relating serum digoxin levels with ventricular response in patients with atria1 fibrillation,3-5 there was little correlation between ventricular response and serum digoxin levels. However, in none of these studies were the patients separated on the basis of clinical status or duration of arrhythmia. The explanation for the lack of correlation between blood level and drug effect probably lies in other factors that influence A-V nodal conduction, particularly the amounts of circulating and neuronally mediated catecholamines. It is evident from previous studies that A-V nodal conduction is influenced by a variety of factors, including intrinsic A-V blockade, catecholamines, the parasympathetic system and the digitalis glycosides. It is enhanced by beta adrenergic stimulating catecholamines and slowed by the digitalis glycosides’ parasympathetic sensitizing activity and by their direct blocking effects on the A-V node. In the presence of serious, unstable clinical states, parasympathetic tone probably decreases, whereas hyperadrenergic activity and release of catecholamines becomes prominent. These factors tend to blunt and even override the A-V blocking activity of digoxin so that the ventricular rate in the presence of atria1 fibrillation remains rapid. The interplay between catecholamines and digitalis has been demonstrated in the experimental dog model of Ogden et at the nodal level beal.,s in which a competition tween beta stimulators and digitalis was shown, and the A-V blocking effects of digitalis were diminished in the presence of high circulating levels of beta adrenergic stimulators. The clinically important information emerging from these findings is that, in the presence of severe unstable clinical conditions, the effects of digitalis on the A-V node may be minimal and the drug remain incapable of normalizing the ventricular rate. This is not to say that it is necessarily desirable to decrease the ventricular rate below lOO/min, since even sinus tachycardia is a normal compensatory response, considered appropriate and even beneficial in patients with acute illnesses. Thus, the point is not that the ventricular rate need be slowed to below lOO/min, but rather that “therapeutic” levels of digoxin are incapable of doing so, and were excessive doses of the drug thoughtlessly employed, toxicity could eventuate. 3. Serum digoxin levels and signs of toxicity: Of additional note is the observation that high serum concentrations of digoxin-between 2.5 and 5 ng/ ml-failed in most instances to elicit gastrointestinal or cardiovascular signs or symptoms of digitalis toxicity. The explanation for the apparent disparity between the data gathered from these patients and the
654
May 1975
The American Journal of CARDIOLOGY
results reported by Smith et a1.l’ may reside in the populations studied. Two thirds of the patients from the Massachusetts General Hospital investigation had sinus rhythm and were receiving digitalis for congestive heart failure, ischemic heart disease, myocardial disease or pulmonary dysfunction. Patients with these conditions are considered among those most sensitive to the toxic effects of digitalis and, therefore, most prone to the development of digitalisinduced arrhythmias. The patients in our group with atria1 fibrillation probably as a whole had less myocardial disease and were thus at less risk of development of digitalis intoxication. This point is brought home cogently in a review of previous experience9 with digitalis-related suicidal attempts or poisoning in children which demonstrated that drug toxicity was relatively mild in face of ingestion of huge amounts of a digitalis preparation. A second factor influencing the low frequency of digitalis toxicity seen herein may relate to the initial rapid ventricular rates; such rates tend to suppress premature ventricular contractions or runs of ventricular tachycardia by overdriving the heart and, thus, a persistently rapid ventricular rate is, in effect, antiarrhythmic” and capable of masking early signs of digitalis-induced arrhythmias. 4. Serum digoxin levels as a reflection of myocardial concentration: Finally, there are many imponderables in our knowledge of the use of digitalis and digoxin blood levels. Just how closely serum levels reflect drug activity that is taking place either in the membrane or within the myocardial cell has never been adequately clarified. Certainly, it has been clearly shown that no simple relation exists between intra- and extracellular concentrations of the drug. Moreover, a marked inhomogeneity in concentrations of digoxin within different locations in the same heart has previously been demonstrated.‘l Although ‘Yoxic” and “nontoxic” concentrations of the drug have been defined these definitions represent mere averages and do not relate to the clinically relevant actions of the drug-inotropism and A-V blocking activity. A significant number of patients without digitalis toxicity (10 percent) have concentrations in excess of 2.0 ng/ml, usually in the range of 2 to 3 ng/ ml, and a significant number of patients with clinical evidence of toxicity have levels in the “normal” range of 1.4 to 2.0 ng/ml.i2 Most important, the time-tested procedure of “pushing” digitalis until the ventricular rate is adequately slowed in atria1 fibrillation must be decried since, in many clinical situations, potentially toxic and dangerous levels of the drug may build up. In clinical practice, if judicious amounts of digitalis do not effectively slow ventricular rate, other forms of therapy such as treatment of the underlying complicating condition, administration of antiarrhythmic agents, use of propranolol to elicit A-V blockade, or cardioversion seem to be rational alternatives to overdigitalization.
Volume 35
DIGITALIS AND ATRIAL FIBRILLATION-GOLDMAN
ET AL.
References I. Smith TW, Butler VP Jr, Haber E: Determination of therapeutic and toxic serum digoxin concentrations by radioimmunoassay. N Engl J Med 281:1212-1216, 1969 2. Smith TW, Haber E: Digoxin intoxication: the relationship of clinical presentation to serum digoxin concentration. J Clin Invest 49:2377-2386, 1970 3. Chamberlain DA, White RJ, Howard MR, et al: Plasma digoxin concentrations in patients with atrial fibrillation. Br Med J 32: 429-432, 1970 4. Redfors A: Plasma digoxin concentration-its relation to digoxin dosage and clinical effects in patients with atrial fibrillation. Br Heart J 34:383-391, 1972 5. Zener JC, Anggard EE, Harrison DC, et al: Persistence of digoxin effect in atrial fibrillation. JAMA 224:239-241, 1973 6. Shapiro W, Harahara K. Taubert K: Relationship of plasma digitoxin and digoxin to cardiac response following IV digitalization in man. Circulation 42:1065-1072. 1970
7. Doherty JE, Perkins WH, Mltchell GK: Tritiated digoxin studies in human subjects. Arch Intern Med 108:531-539, 1961 8. Ogden PC, Seizer A, Cohn KC: The relationship between inotropic and dromotropic effects of digitalis: the modulation of these effects by autonomic influences. Am Heart J 77:628-635, 1969 9. Smith TW, Wlllerson JT: Suicidal and accidental digoxin ingestion. Circulation 49:29-36, 1971 10. Cohn K, Agmon J, Gamble 0: The effect of glucagon on arrhythmias due to digitalis toxicity. Am J Cardiol 25:683-689, 1970 11. Doherty JE, Perkins WH, Flanigan WJ: Distribution and concentration of tritiated digoxin in human tissues. Ann Intern Med 66:116-124, 1966 12. Donoso E: Letter of the Council on Clinical Cardiology, American Heart Association 1:3, 1974
May 1975
The American Journal of CARDIOLOGY
Volume 35
655
