Practical Therapeutics
Drugs 40 (6): 841-853, 1990 0012-6667/90/0012-0841 /$06.50/0 © Adis International Limited All rights reserved. DRUG03411
Atrial Fibrillation
The Therapeutic Options
Richard V. Lewis Department of Cardiology, Pontefract General Infirmary, Yorkshire, England
Contents
Summary ................................................................................................................................... 841 I. Haemodynamic Consequences of AF ......................................................................... ....... 842 1.1 Absence of Atrial Systole .............................................................................................. 842 1.2 Rapidity of the Ventricular Response Rate ................................. ............................... 842 1.3 Irregularity of the Ventricular Response ..................................................................... 843 2. Clinical Consequences of AF .............................................................................................. 843 3. Treatment of AF ............................................................................................. ..................... 844 3.1 General Principles ......................................................................................................... 844 3.2 Treatment of Acute AF ................................................................................................ 844 3.2.1 Advantages of Cardioversion .................... .......................................................... 844 3.2.2 Pharmacological Conversion of AF .................................................................... 845 3.2.3 Maintenance of Sin.us Rhythm After Successful Cardioversion ........ .............. 846 3.3 Chronic, Stable AF ......... ............................................................................................... 846 3.3.1 Control of the Ventricular Response Rate ......................................................... 846 3.3.2 Digoxin .................................................................................................................. 846 3.3.3 Verapamil and Diltiazem ............................. ....................................................... 847 3.3.4 p-Adrenoceptor Antagonists ................................................................................ 848 3.3.5 Prevention of Embolisation ................................................................. ................ 848 4. AF in the 'Sick Sinus' Syndrome ....................................................................................... 849 5. Paroxysmal AF ..................................................................................................................... 849 6. AF in the Presence of Ventricular Pre-Excitation ................................................... ......... 850 6.1 Acute AF with Pre-Excitation ................................................................................. ..... 851 6.2 Prevention of Arrhythmia in Patients with Ventricular Pre-Excitation ................... 851 7. ·Current Recommendations ................................................................................................. 852
Summary
Atrial fibrillation (AF) is a common cardiac arrhythmia which is particularly prevalent among the elderly. In patients with AF of recent onset, restoration of sinus rhythm may be feasible and this can be achieved by DC cardioversion, or by the use of one of a number of drugs including amiodarone, flecainide or propafenone. Neither digoxin nor the calcium antagonists facilitate the restoration of sinus rhythm. Recurrence of AF is common after successful cardioversion and, although long term antiarrhythmic drug therapy may help to maintain sinus rhythm, all Slich drugs are potentially toxic and can have important proarrhythmic actions. In patients with chronic AF, rest? ration of sinus rhythm is rarely possible and treatment is directed towards control of Jlhe ventricular response rate, which may be achieved with digoxin and/or a rate-limiti~g calcium ant-
842
Drugs 40 (6) 1990
agonist such as verapamil or diltiazem ; iJ-blo~kers may also be used although they appear to impair effort tolerance. In addition, long term anticoagulation may be indicated to reduce the risks of systemic embolisation, even in patients with 'nonrheumatic' AF; antiplatelet drugs are of no apparent value in this context. A minority of patients present with AF associated with ventricular pre-excitation; in these individuals both digoxin and the calcium antagonists are contraindicated and the ventricular response rate should be controlled with tlecainide, amiodarone or propafenone.
Atrial fibrillation (AF) is a common cardiac arrhythmia which is particularly prevalent among the elderly (fig. 1) [Camm et al. 1980; Kulbertus et al. 1984]. It is sometimes regarded as a relatively trivial disorder, but many patients with chronic AF take long term treatment with potent antiarrhythmic and anticoagulant drugs which can have important adverse effects. Digoxin is traditionally regarded as the cornerstone of therapy in chronic stable AF, but other classes of drugs including the {3-blockers and the heart-rate-limiting calcium antagonists (such as verapamil and diltiazem) may have certain advantages over the cardiac glycosides in some patients. This article reviews the thera-
5
M
M
4
OF
t
Ii.
« 3 15 ~
c:
Q)
;;; >
Q)
a:
2
/
peutic options which are currently available for the treatment of AF in its various forms and presentations.
1. HaemQdynamic Consequences of AF The main haemodynamic disturbances seen in AF result from absence of atrial systole and from the rapidity and the irregularity of the ventricular response. 1.1 Absence of Atrial Systole In the normal heart, atrial systole augments ventricular filling by increasing both the rate and the extent of transmitral diastolic blood flow. However, the precise contribution of atrial systole to overall cardiac function is difficult to establish. Successful cardioversion of AF to sinus rhythm is associated with an eventual improvement in left ventricular function (Lipkin et a1.1988) but this may be due, in part, to changes in heart rate and rhythm and not just to the restoration of atrial systole. In a group of patients with complete heart block it was found that the mean cardiac index decreased from 2.7 L/min/m2 during correct atrioventricular sequencing to 2.3 L/min during ventricular pacing (Samet et al. 1965). , Thus, absence of atrial systole appears to reduce ~ardiac output, but it is likely that both the rate and the irregularity of the ventricular response also contribute to' the haemodynamic changes seen in atrial fibrillation. 1.2 Rapidity of the Ventricular Response Rate
45·54
55-64
65-74
> 74
Age group (years)
Fig. 1. Prevalence of AF in a sample of 43 809 subjects (reproduced with permission from Kulbertus et al. 1984).
Systole occupies a fairly consistent proportion of the cardiac cycle, irrespective of the heart rate, and during periods of tachycardia the duration of
Atrial Fibrillation: The Therapeutic Options
843
1.3 Irregularity of the Ventricular Response In some patients, atrial fibrillation is associated with mitral regurgitation but this appears to be a consequence of irregular cycle lengths, rather than atrial fibrillation itself (Naito et al. 1980).
2. Clinical Consequences of AF
/
/ Mitral stenosis
/
/ /
R-R interval
Fig.2. The relationship between left ventricular end-diastolic volume and duration ofR-R interval in patients with normal mitral valves and in mitral stenosis (adapted with permission from Meijler 1984).
diastole is necessarily reduced. In the presence of mitral stenosis, increases in diastolic dimensions are related to the duration of diastole (fig. 2) so that ventricular filling only becomes complete towards the end of a normal diastolic interval (Kulbertus et al. 1984). Thus, when the ventricular rate is rapid, ventricular filling is incomplete and this may lead to a decrease in cardiac output according to the Frank-Starling mechanism. However, in patients with atrial fibrillation and normal mitral valves, left ventricular end-diastolic dimensions are reached early in diastole and remain constant thereafter (fig. 2). Thus, a fast ventricular rate is unlikely to reduce cardiac output by a large amount in this situation. When R-R intervals are extremely short, mechanical alternans may be seen and this phenomenon may contribute further to the deterioration of cardiac haemodynamics (Edmands et al. 1970).
The sudden onset of fast atrial fibrillation may precipitate overt cardiac failure, particularly if left ventricular function is already compromised by coexisting valvular or ischaemic heart disease. More often, however, atrial fibrillation presents with less dramatic symptoms including palpitation, eXertional dyspnoea, ischaemic cardiac pain or general fatigue and lethargy. It is likely that haemodynamic disturbances in atrial fibrillation become more marked on exercise as disproportionate increases in the ventricular response rate may be seen during mild to moderate exertion. Symptoms such as dyspnoea, muscle fatigue or anginal pain (in those with ischaemic heart disease) may limit maximum exercise tolerance, and there is evidence to suggest that reversion to sinus rhythm is associated with improved exercise tolerance, maximum oxygen consumption and anaerobic threshold in some individuals. However, more objective symptoms may occur at rest or during mild exercise and it is possible that awareness of cardiac irregularity itself may affect the perceived sense of well-being. Thus, ' in assessing patients with atrial fibrillation, attention should be directed towards subjective, as well . as objective, parameters. While measurement of heart rate, exercise tolerance and cardiac output may reflect changes in cardiac function, these parameters may not be direct determinants of how well or unwell patients actually felt. For example, one study has shown that increasing doses of digoxin were associated with an improved sense .of general wellbeing although only small changes in effort tolerance were reported (Redfors 1971).
844
3. Treatment
Drugs 40 (6) 1990
0/ AF
3.1 General Principles AF can present in a number of ways, and the aims and modalities of treatment differ according to the nature of the presentation and the underlying cause (if any). Some patients present with rapid AF of recent onset, which can be associated with a marked disturbance ofhaemodynamic function. In this situation, treatment may be directed towards restoration and maintenance of sinus rhythm, particularly in the presence of a correctable underlying disorder. However, many patients present with chronic, relatively stable AF, when attempts to restore sinus rhythm may be neither appropriate nor successful. Here, the main aims of therapy are simply to control the ventricular response rate and to reduce the risks of systemic embolisation. Some patients have an underlying disorder of sinus node function, where periods of atrial fibrillation may alternate with runs of sinus bradycardia or episodes of sinus arrest; in this situation treatment is potentially difficult, as drugs which impair atrioventricular conduction may predispose to extreme bradycardia. Some patients have paroxysmal AF, where episodes of AF may frequently be associated with a particularly rapid ventricular response rate. Here, drug treatment should reduce both the frequency with which the arrhythmia recurs, and control the heart rate during those attacks which continue to occur. Finally, a minority of patients develop AF as a result of pre-excitation due to an accessory path~ay which can conduct impulses directly from the atria to the ventricles; such accessory pathways can conduct impulses at a higher rate than the AV ndde and may, therefore, be associated with a particularly rapid ventricular response rate. Moreover, maby of those drugs used to control 'normally conducted' AF are inappropriate in the presence of pre~excitation. 3.2 Treatment of Acute AF Rapid AF of recent onset may be associated with a marked disturbance of haemodynamic function and treatment may then be directed primarily to-
wards restoration and maintenance of sinus rhythm. Indeed, cardioversion should probably be attempted in most patients who present with AF of recent onset and is generally most successful in younger patients in whom there is no underlying structural cardiac disorder. In contrast, restoration of sinus rhythm is rarely possible in older patients with cardiomegaly and AF of more than 3 to 6 months duration. 3.2.1 Advantages of Cardioversion Sinus rhythm is presumably preferable to AF in virtually all patients, but the benefits of successful cardioversion (as opposed to 'controlled' AF) have not always been easy to document. Several studies have shown little improvement in either exercise tolerance or cardiac output following restoration of sinus rhythm, but recent work suggests that this may be because the atria take time to resume their normal contractile function. For example, Lipkin et al. (1988) found no improvement in haemodynamic function 24 hours after successful cardioversion, but 28 days later there was a 16% increase in maximal oxygen uptake and an 18% increase in the anaerobic threshold. These changes were associated with Doppler evidence of increased transmitral diastolic blood flow, indic~ting that the benefits were presumably a result of improved ventricular filling. Cardioversion is, therefore, of probable value in patients with AF of recent onset. and may be achieved by drug therapy, direct current (DC)countershock, or a combination . o~ the two. DC cardioversion is often effective at first but the long term results are disappointing and ,many patients revert back into AF during the follpwing months; only 15 to 50% of patients remain in sinus rhythm one year after successful cardioversion. Predictors of early relapse of AF include a lef,t atrial dimension > 45mm, duration of AF more than 3 months and failure of the A wave on DQPpier echocardiography to increase by more than 10% over the 24 hours following cardioversion . (Dethy et al. 1988).
Atrial Fibrillation: The Therapeutic Options
3.2.2 Pharmacological Cardioversion of AF While electrical cardioversion is generally safe and effective, it does necessitate the use of sedation or light anaesthesia, and drug therapy might therefore have some advantages in certain situations. Moreover, the drugs used to achieve cardioversion may have a persisting therapeutic effect which might reduce the risks of early relapse. Contrary to widespread belief, there is no evidence to suggest that digitalisation increases the likelihood of reversion to sinus rhythm, and there is no reason to suppose t hat it should have such an effect; the cardiac glycosides have no demonstrable effect on fibrillatory activity within the atrium. Spontaneous reversion to sinus rhythm is relatively common in AF of recent onset, and a randomised, double-blind placebo-controlled study has shown that conversion to sinus rhythm occurred as often in the placebo group as in the dig" oxin-treated patients (Falk et al. 1987). Similarly, both verapamil and dilt~azem appear to be relatively unsuccessful at restoring sinus rhythm, with reported 'success' rates 01 between 0 and 29% (Salerno et al. 1989; Schamrbth et al. 1972); they are, however, like digoxin,1effective in controlling the ventricular response rate ,during established fibrillation unless there is an. accessory pathway. Quinidine has been extensively used in the past to achieve 'pharmacological' cardioversion. Similarly, treatment with procainamide may lead to restoration of sinus rhythm in between 43 and 88% of patients with atrial fibrillation, and disopyramide is effective in 55 to 71% of cases (Camm et al. 1979). However, nonb of these drugs depresses AV nodal function and they may, in fact, !facilitate conduction leading to an1increased ventricular rate if the fibrillation persists. These three agents are not now used routinely for 'pharmacplogical' cardioversion and the current trend is:,t o use newer drugs such as flecainide; propafenorie and amiodarone. Flecainide, given by intravenous infusion, has an overall efficacy of between 65 and 90% (Holmes & Heel 1985; Suttorp et :al. 1989) but tends to depress myocardial contractility .. anrlsh()uld be avoided in patients with cardiac dec~mpensation,
845
unless this is primarily a rate-related phenomenon. Flecainide, like other class lA antiarrhythmic agents, does not have any rate-limiting effect in AF and will not control the ventricular response rate if the fibrillation persists. Interestingly, pretreatment with flecainide may increase the threshold for electrical cardioversion if this is subsequently necessary (Van Gelder et al. 1989). Propafenone is a relatively new drug in British and North American cardiological practice, but may have certain advantages over flecainide. It inhibits the re-entry pathways, responsible for most cases of AF, by slowing conduction and blocking one limb of the re-entry circuit. It also has weak (3adrenoceptor antagonist properties (Harron & Brogden 1987) and will therefore suppress sympathetic stimulation of the heart. Bianconi et al. (1989) found that intravenous propafenone achieved successful cardioversion in 71 % of patients with AF of less than 48 hours duration, but in only 26% of those with more chronic fibrillation. Unlike flecainide, it will generally slow the ventricular response rate, even in the presence of persisting AF. Amiodarone (Faniel & Schoenfeld 1983; Strasberg et al. 1985) is reported to achieve restoration of sinus rhythm in between 48 and 81 % of patients, and is relatively nontoxic in the short term. Like propafenone, amiodarone has a rate-limiting action and will help to control the ventricular response rate even if atrial fibrillation persists. Intravenous amiodarone can depress left ventricular function and should be used with care in patients with impaired cardiac reserves. Nevertheless, it is probably less negatively inotropic than either flecainide or propafenone, and may be preferable to either of these agents in patients with cardiac failure. Sotalol, a (3-adrenoceptor blocking drug which, by delaying repolarisation, prolongs the duration of the action potential, has also been used to terminate acute atrial fibrillation (Singh et al. 1987). It has been found to be as effective as a combination of digoxin and disopyramide in patients who developed AFfollowing cardiopulmonary bypass (Campbell etal. 1985) but, like other agents, is relatively ineffective in patients with chronic AF.
846
3 .2.3 ' Maintenance of Sinus Rhythm After Successful Cardioversion Relapse of AF following successful cardioversion is common unless. a remedial underlying cause such as thyrotoxicosis can be identified and cor· rected. In other patients, long term antiarrhythmic drug treatment may be used to maintain sinus rhythm and a variety of agents have been evaluated in this context. There is little evidence to suggest that digoxin is of value in preventing recurrence of AF, and its vagotonic actions might be expected to promote the conditions which favour recurrence. Similarly, the calcium antagonists have no effect on the electrophysiological .properties of atrial tissue in the intact humiiln heart (Antman et al. 1980) and would not be expected to prevent recurrence of AF. Indeed, S~enasa et al. (1988) found that both verapamil and diltiazem enhanced the maintenance of electrically stimulated AF. Quinidine has been reported' by some, but not all, workers to reduce the relapse rate, but was one of the first antiarrhythmic agents to be associated with an important proarrhythmic potential and several deaths have been ass~ciated with the use of quinidine in AF. Since then; it has been realised that many antiarrhythmic agents can paradoxically predispose to serious ventricular arrhythmias; and disopyramide, propafenone, amiodarone and flecainide (Falk 1989) [among others] have been implicated. The recently reported CAST study (CAST Investigators Preliminary Report 1989) has increased concern about this phenomenon. All of the above drugs appear to help in the maintenance of sinus rhythm following successful cardioversion; disopyramide, for example, .has been shown to reduce the relapse rate from 70% to 46% after one year (Karlson et al. 1988). Nevertheless, AF is, for most patients, a relatively benign disorder and it is not by any means clear whether it is justifiable to start long term treatment with potentially harmful drugs simply to achieve the rather modest benefits of sinus rhythm as oppose;d to controlled AF.
3.3 Chronic, Stable AF . In patients with established AF; in whom a correctable underlying cause cannot be identified, cardioversion is unlikely to achieve a sustained res-
Drugs 40 (6) 1990
toration of sinus rhythm and is unlikely to be of great benefit to the patient. Treatment is therefore directed towards achieving adequate control of the ventricular response rate and minimising the risks of systemic embolisation. 3.3.1 Control of the Ventricular Response Rate Drug treatment may reduce both the rate and the . irregularity of the ventricular response. The ventricular rate can be readily reduced by giving drugs which increase the extent of the atrioventricular (A V) nodal conduction block, thereby improving ventricular ·filling by increasing diastolic filling times. Increasing heart rate is associated with increases in cardiac output until a critical value is reached, after which ventricular filling becomes a limiting factor. Once this point is reached, cardiac output will tend to decline with further increases in the heart rate. We might expect that rate reduction would be of particular value in patients with mitral stenosis, as increases in end-diastolic dimerisions are related, OVer a certain range, to the duration of diastole. However, there is little evidence to suggest that patients with mitral stenosis do derive any particular benefit from the 'better' rate control (e.g. exercise heart rates of less than 140 beats/min) which can be achieved with combinations of digoxin plus either calcium antagonists or /3-blockers (Lewis et al. I 988b). 3.3.2 Digoxin Digoxin, which has now been in clinical use for over 200 years, is widely regarded as the drug of choice for the long term treatment of chronic atrial fibrillation. In this context, the therapeutic effects of the cardiac glycosides are generally thought to be due primarily totheir rate~limitingeffects rather than their inotropic actions. Nevertheless; it has been known for many years that digoxin fails to control exercise-induced tachycardia, even when plasma drug concentrations are near the upper end of the accepted 'therapeutic' range (Beasley .et al. 1985). This is predictable from the Qasic pharmacology of the cardiac glycosides, as they have little direct effect on AV nodal conduction, and their
Atrial Fibrillation: The Therapeutic Options
major therapeutic effects are mediated via increased vagal tone and inhibition of sympathetic stimulation (Klein & Kaplinsky 1986). During exercise, intrinsic vagal tone is withdrawn and sympathetic activity increases so that digoxin appears to have relatively little effect on postexercise heart rate compared with placebo (Lang et al. 1983; Lewis et a1. 1988a). The reductions in heart rate observed during early experience with digoxin in severe congestive cardiac failure may, in fact, have been a secondary phenomenon reflecting an improvement in cardiac function brought about by some other mechanism. While the cardiac glycosides fail to control exercise-induced tachycardia, it is clear that digoxin can achieve marked reductions in the resting heart rate. This may predispose to periods of quite marked nocturnal bradycardia, because a diurnal variation in heart rate is frequently seen in subjects with atrial fibrillation (Pitcher et al. 1986). Digoxin has a narrow therapeutic ratio and toxicity may be associated with potentially lethal arrhythmias. The threshold for glycoside toxicity shows considerable interindividual variability and both ·hypokalaemia and hypomagnesaemia · predispose to cardiac toxicity. Conversely, serum concentrations of theglycoside·of less than about 1 nmoljL are thought to be largely ineffectual. Within the accepted 'therapeutic'range (1.3 to 2.6 nmoljL) there is evidence of a weak relationship between plasma drug concentration and pharmacological effect (Beasley et al. 1985). Digoxin is generally well tolerated and the more common adverse effects such as nausea or other gastrointestinal disturbances may respond to a decrease in drug dosage. However, there is some concern that even 'therapeutic' plasma concentrations of digoxin may increase the preva1enee of ventricular arrhythmias (Lewis & McDevitt .1988). There is also evidence suggesting that
Drugs 40 (6): 841-853, 1990 0012-6667/90/0012-0841 /$06.50/0 © Adis International Limited All rights reserved. DRUG03411
Atrial Fibrillation
The Therapeutic Options
Richard V. Lewis Department of Cardiology, Pontefract General Infirmary, Yorkshire, England
Contents
Summary ................................................................................................................................... 841 I. Haemodynamic Consequences of AF ......................................................................... ....... 842 1.1 Absence of Atrial Systole .............................................................................................. 842 1.2 Rapidity of the Ventricular Response Rate ................................. ............................... 842 1.3 Irregularity of the Ventricular Response ..................................................................... 843 2. Clinical Consequences of AF .............................................................................................. 843 3. Treatment of AF ............................................................................................. ..................... 844 3.1 General Principles ......................................................................................................... 844 3.2 Treatment of Acute AF ................................................................................................ 844 3.2.1 Advantages of Cardioversion .................... .......................................................... 844 3.2.2 Pharmacological Conversion of AF .................................................................... 845 3.2.3 Maintenance of Sin.us Rhythm After Successful Cardioversion ........ .............. 846 3.3 Chronic, Stable AF ......... ............................................................................................... 846 3.3.1 Control of the Ventricular Response Rate ......................................................... 846 3.3.2 Digoxin .................................................................................................................. 846 3.3.3 Verapamil and Diltiazem ............................. ....................................................... 847 3.3.4 p-Adrenoceptor Antagonists ................................................................................ 848 3.3.5 Prevention of Embolisation ................................................................. ................ 848 4. AF in the 'Sick Sinus' Syndrome ....................................................................................... 849 5. Paroxysmal AF ..................................................................................................................... 849 6. AF in the Presence of Ventricular Pre-Excitation ................................................... ......... 850 6.1 Acute AF with Pre-Excitation ................................................................................. ..... 851 6.2 Prevention of Arrhythmia in Patients with Ventricular Pre-Excitation ................... 851 7. ·Current Recommendations ................................................................................................. 852
Summary
Atrial fibrillation (AF) is a common cardiac arrhythmia which is particularly prevalent among the elderly. In patients with AF of recent onset, restoration of sinus rhythm may be feasible and this can be achieved by DC cardioversion, or by the use of one of a number of drugs including amiodarone, flecainide or propafenone. Neither digoxin nor the calcium antagonists facilitate the restoration of sinus rhythm. Recurrence of AF is common after successful cardioversion and, although long term antiarrhythmic drug therapy may help to maintain sinus rhythm, all Slich drugs are potentially toxic and can have important proarrhythmic actions. In patients with chronic AF, rest? ration of sinus rhythm is rarely possible and treatment is directed towards control of Jlhe ventricular response rate, which may be achieved with digoxin and/or a rate-limiti~g calcium ant-
842
Drugs 40 (6) 1990
agonist such as verapamil or diltiazem ; iJ-blo~kers may also be used although they appear to impair effort tolerance. In addition, long term anticoagulation may be indicated to reduce the risks of systemic embolisation, even in patients with 'nonrheumatic' AF; antiplatelet drugs are of no apparent value in this context. A minority of patients present with AF associated with ventricular pre-excitation; in these individuals both digoxin and the calcium antagonists are contraindicated and the ventricular response rate should be controlled with tlecainide, amiodarone or propafenone.
Atrial fibrillation (AF) is a common cardiac arrhythmia which is particularly prevalent among the elderly (fig. 1) [Camm et al. 1980; Kulbertus et al. 1984]. It is sometimes regarded as a relatively trivial disorder, but many patients with chronic AF take long term treatment with potent antiarrhythmic and anticoagulant drugs which can have important adverse effects. Digoxin is traditionally regarded as the cornerstone of therapy in chronic stable AF, but other classes of drugs including the {3-blockers and the heart-rate-limiting calcium antagonists (such as verapamil and diltiazem) may have certain advantages over the cardiac glycosides in some patients. This article reviews the thera-
5
M
M
4
OF
t
Ii.
« 3 15 ~
c:
Q)
;;; >
Q)
a:
2
/
peutic options which are currently available for the treatment of AF in its various forms and presentations.
1. HaemQdynamic Consequences of AF The main haemodynamic disturbances seen in AF result from absence of atrial systole and from the rapidity and the irregularity of the ventricular response. 1.1 Absence of Atrial Systole In the normal heart, atrial systole augments ventricular filling by increasing both the rate and the extent of transmitral diastolic blood flow. However, the precise contribution of atrial systole to overall cardiac function is difficult to establish. Successful cardioversion of AF to sinus rhythm is associated with an eventual improvement in left ventricular function (Lipkin et a1.1988) but this may be due, in part, to changes in heart rate and rhythm and not just to the restoration of atrial systole. In a group of patients with complete heart block it was found that the mean cardiac index decreased from 2.7 L/min/m2 during correct atrioventricular sequencing to 2.3 L/min during ventricular pacing (Samet et al. 1965). , Thus, absence of atrial systole appears to reduce ~ardiac output, but it is likely that both the rate and the irregularity of the ventricular response also contribute to' the haemodynamic changes seen in atrial fibrillation. 1.2 Rapidity of the Ventricular Response Rate
45·54
55-64
65-74
> 74
Age group (years)
Fig. 1. Prevalence of AF in a sample of 43 809 subjects (reproduced with permission from Kulbertus et al. 1984).
Systole occupies a fairly consistent proportion of the cardiac cycle, irrespective of the heart rate, and during periods of tachycardia the duration of
Atrial Fibrillation: The Therapeutic Options
843
1.3 Irregularity of the Ventricular Response In some patients, atrial fibrillation is associated with mitral regurgitation but this appears to be a consequence of irregular cycle lengths, rather than atrial fibrillation itself (Naito et al. 1980).
2. Clinical Consequences of AF
/
/ Mitral stenosis
/
/ /
R-R interval
Fig.2. The relationship between left ventricular end-diastolic volume and duration ofR-R interval in patients with normal mitral valves and in mitral stenosis (adapted with permission from Meijler 1984).
diastole is necessarily reduced. In the presence of mitral stenosis, increases in diastolic dimensions are related to the duration of diastole (fig. 2) so that ventricular filling only becomes complete towards the end of a normal diastolic interval (Kulbertus et al. 1984). Thus, when the ventricular rate is rapid, ventricular filling is incomplete and this may lead to a decrease in cardiac output according to the Frank-Starling mechanism. However, in patients with atrial fibrillation and normal mitral valves, left ventricular end-diastolic dimensions are reached early in diastole and remain constant thereafter (fig. 2). Thus, a fast ventricular rate is unlikely to reduce cardiac output by a large amount in this situation. When R-R intervals are extremely short, mechanical alternans may be seen and this phenomenon may contribute further to the deterioration of cardiac haemodynamics (Edmands et al. 1970).
The sudden onset of fast atrial fibrillation may precipitate overt cardiac failure, particularly if left ventricular function is already compromised by coexisting valvular or ischaemic heart disease. More often, however, atrial fibrillation presents with less dramatic symptoms including palpitation, eXertional dyspnoea, ischaemic cardiac pain or general fatigue and lethargy. It is likely that haemodynamic disturbances in atrial fibrillation become more marked on exercise as disproportionate increases in the ventricular response rate may be seen during mild to moderate exertion. Symptoms such as dyspnoea, muscle fatigue or anginal pain (in those with ischaemic heart disease) may limit maximum exercise tolerance, and there is evidence to suggest that reversion to sinus rhythm is associated with improved exercise tolerance, maximum oxygen consumption and anaerobic threshold in some individuals. However, more objective symptoms may occur at rest or during mild exercise and it is possible that awareness of cardiac irregularity itself may affect the perceived sense of well-being. Thus, ' in assessing patients with atrial fibrillation, attention should be directed towards subjective, as well . as objective, parameters. While measurement of heart rate, exercise tolerance and cardiac output may reflect changes in cardiac function, these parameters may not be direct determinants of how well or unwell patients actually felt. For example, one study has shown that increasing doses of digoxin were associated with an improved sense .of general wellbeing although only small changes in effort tolerance were reported (Redfors 1971).
844
3. Treatment
Drugs 40 (6) 1990
0/ AF
3.1 General Principles AF can present in a number of ways, and the aims and modalities of treatment differ according to the nature of the presentation and the underlying cause (if any). Some patients present with rapid AF of recent onset, which can be associated with a marked disturbance ofhaemodynamic function. In this situation, treatment may be directed towards restoration and maintenance of sinus rhythm, particularly in the presence of a correctable underlying disorder. However, many patients present with chronic, relatively stable AF, when attempts to restore sinus rhythm may be neither appropriate nor successful. Here, the main aims of therapy are simply to control the ventricular response rate and to reduce the risks of systemic embolisation. Some patients have an underlying disorder of sinus node function, where periods of atrial fibrillation may alternate with runs of sinus bradycardia or episodes of sinus arrest; in this situation treatment is potentially difficult, as drugs which impair atrioventricular conduction may predispose to extreme bradycardia. Some patients have paroxysmal AF, where episodes of AF may frequently be associated with a particularly rapid ventricular response rate. Here, drug treatment should reduce both the frequency with which the arrhythmia recurs, and control the heart rate during those attacks which continue to occur. Finally, a minority of patients develop AF as a result of pre-excitation due to an accessory path~ay which can conduct impulses directly from the atria to the ventricles; such accessory pathways can conduct impulses at a higher rate than the AV ndde and may, therefore, be associated with a particularly rapid ventricular response rate. Moreover, maby of those drugs used to control 'normally conducted' AF are inappropriate in the presence of pre~excitation. 3.2 Treatment of Acute AF Rapid AF of recent onset may be associated with a marked disturbance of haemodynamic function and treatment may then be directed primarily to-
wards restoration and maintenance of sinus rhythm. Indeed, cardioversion should probably be attempted in most patients who present with AF of recent onset and is generally most successful in younger patients in whom there is no underlying structural cardiac disorder. In contrast, restoration of sinus rhythm is rarely possible in older patients with cardiomegaly and AF of more than 3 to 6 months duration. 3.2.1 Advantages of Cardioversion Sinus rhythm is presumably preferable to AF in virtually all patients, but the benefits of successful cardioversion (as opposed to 'controlled' AF) have not always been easy to document. Several studies have shown little improvement in either exercise tolerance or cardiac output following restoration of sinus rhythm, but recent work suggests that this may be because the atria take time to resume their normal contractile function. For example, Lipkin et al. (1988) found no improvement in haemodynamic function 24 hours after successful cardioversion, but 28 days later there was a 16% increase in maximal oxygen uptake and an 18% increase in the anaerobic threshold. These changes were associated with Doppler evidence of increased transmitral diastolic blood flow, indic~ting that the benefits were presumably a result of improved ventricular filling. Cardioversion is, therefore, of probable value in patients with AF of recent onset. and may be achieved by drug therapy, direct current (DC)countershock, or a combination . o~ the two. DC cardioversion is often effective at first but the long term results are disappointing and ,many patients revert back into AF during the follpwing months; only 15 to 50% of patients remain in sinus rhythm one year after successful cardioversion. Predictors of early relapse of AF include a lef,t atrial dimension > 45mm, duration of AF more than 3 months and failure of the A wave on DQPpier echocardiography to increase by more than 10% over the 24 hours following cardioversion . (Dethy et al. 1988).
Atrial Fibrillation: The Therapeutic Options
3.2.2 Pharmacological Cardioversion of AF While electrical cardioversion is generally safe and effective, it does necessitate the use of sedation or light anaesthesia, and drug therapy might therefore have some advantages in certain situations. Moreover, the drugs used to achieve cardioversion may have a persisting therapeutic effect which might reduce the risks of early relapse. Contrary to widespread belief, there is no evidence to suggest that digitalisation increases the likelihood of reversion to sinus rhythm, and there is no reason to suppose t hat it should have such an effect; the cardiac glycosides have no demonstrable effect on fibrillatory activity within the atrium. Spontaneous reversion to sinus rhythm is relatively common in AF of recent onset, and a randomised, double-blind placebo-controlled study has shown that conversion to sinus rhythm occurred as often in the placebo group as in the dig" oxin-treated patients (Falk et al. 1987). Similarly, both verapamil and dilt~azem appear to be relatively unsuccessful at restoring sinus rhythm, with reported 'success' rates 01 between 0 and 29% (Salerno et al. 1989; Schamrbth et al. 1972); they are, however, like digoxin,1effective in controlling the ventricular response rate ,during established fibrillation unless there is an. accessory pathway. Quinidine has been extensively used in the past to achieve 'pharmacological' cardioversion. Similarly, treatment with procainamide may lead to restoration of sinus rhythm in between 43 and 88% of patients with atrial fibrillation, and disopyramide is effective in 55 to 71% of cases (Camm et al. 1979). However, nonb of these drugs depresses AV nodal function and they may, in fact, !facilitate conduction leading to an1increased ventricular rate if the fibrillation persists. These three agents are not now used routinely for 'pharmacplogical' cardioversion and the current trend is:,t o use newer drugs such as flecainide; propafenorie and amiodarone. Flecainide, given by intravenous infusion, has an overall efficacy of between 65 and 90% (Holmes & Heel 1985; Suttorp et :al. 1989) but tends to depress myocardial contractility .. anrlsh()uld be avoided in patients with cardiac dec~mpensation,
845
unless this is primarily a rate-related phenomenon. Flecainide, like other class lA antiarrhythmic agents, does not have any rate-limiting effect in AF and will not control the ventricular response rate if the fibrillation persists. Interestingly, pretreatment with flecainide may increase the threshold for electrical cardioversion if this is subsequently necessary (Van Gelder et al. 1989). Propafenone is a relatively new drug in British and North American cardiological practice, but may have certain advantages over flecainide. It inhibits the re-entry pathways, responsible for most cases of AF, by slowing conduction and blocking one limb of the re-entry circuit. It also has weak (3adrenoceptor antagonist properties (Harron & Brogden 1987) and will therefore suppress sympathetic stimulation of the heart. Bianconi et al. (1989) found that intravenous propafenone achieved successful cardioversion in 71 % of patients with AF of less than 48 hours duration, but in only 26% of those with more chronic fibrillation. Unlike flecainide, it will generally slow the ventricular response rate, even in the presence of persisting AF. Amiodarone (Faniel & Schoenfeld 1983; Strasberg et al. 1985) is reported to achieve restoration of sinus rhythm in between 48 and 81 % of patients, and is relatively nontoxic in the short term. Like propafenone, amiodarone has a rate-limiting action and will help to control the ventricular response rate even if atrial fibrillation persists. Intravenous amiodarone can depress left ventricular function and should be used with care in patients with impaired cardiac reserves. Nevertheless, it is probably less negatively inotropic than either flecainide or propafenone, and may be preferable to either of these agents in patients with cardiac failure. Sotalol, a (3-adrenoceptor blocking drug which, by delaying repolarisation, prolongs the duration of the action potential, has also been used to terminate acute atrial fibrillation (Singh et al. 1987). It has been found to be as effective as a combination of digoxin and disopyramide in patients who developed AFfollowing cardiopulmonary bypass (Campbell etal. 1985) but, like other agents, is relatively ineffective in patients with chronic AF.
846
3 .2.3 ' Maintenance of Sinus Rhythm After Successful Cardioversion Relapse of AF following successful cardioversion is common unless. a remedial underlying cause such as thyrotoxicosis can be identified and cor· rected. In other patients, long term antiarrhythmic drug treatment may be used to maintain sinus rhythm and a variety of agents have been evaluated in this context. There is little evidence to suggest that digoxin is of value in preventing recurrence of AF, and its vagotonic actions might be expected to promote the conditions which favour recurrence. Similarly, the calcium antagonists have no effect on the electrophysiological .properties of atrial tissue in the intact humiiln heart (Antman et al. 1980) and would not be expected to prevent recurrence of AF. Indeed, S~enasa et al. (1988) found that both verapamil and diltiazem enhanced the maintenance of electrically stimulated AF. Quinidine has been reported' by some, but not all, workers to reduce the relapse rate, but was one of the first antiarrhythmic agents to be associated with an important proarrhythmic potential and several deaths have been ass~ciated with the use of quinidine in AF. Since then; it has been realised that many antiarrhythmic agents can paradoxically predispose to serious ventricular arrhythmias; and disopyramide, propafenone, amiodarone and flecainide (Falk 1989) [among others] have been implicated. The recently reported CAST study (CAST Investigators Preliminary Report 1989) has increased concern about this phenomenon. All of the above drugs appear to help in the maintenance of sinus rhythm following successful cardioversion; disopyramide, for example, .has been shown to reduce the relapse rate from 70% to 46% after one year (Karlson et al. 1988). Nevertheless, AF is, for most patients, a relatively benign disorder and it is not by any means clear whether it is justifiable to start long term treatment with potentially harmful drugs simply to achieve the rather modest benefits of sinus rhythm as oppose;d to controlled AF.
3.3 Chronic, Stable AF . In patients with established AF; in whom a correctable underlying cause cannot be identified, cardioversion is unlikely to achieve a sustained res-
Drugs 40 (6) 1990
toration of sinus rhythm and is unlikely to be of great benefit to the patient. Treatment is therefore directed towards achieving adequate control of the ventricular response rate and minimising the risks of systemic embolisation. 3.3.1 Control of the Ventricular Response Rate Drug treatment may reduce both the rate and the . irregularity of the ventricular response. The ventricular rate can be readily reduced by giving drugs which increase the extent of the atrioventricular (A V) nodal conduction block, thereby improving ventricular ·filling by increasing diastolic filling times. Increasing heart rate is associated with increases in cardiac output until a critical value is reached, after which ventricular filling becomes a limiting factor. Once this point is reached, cardiac output will tend to decline with further increases in the heart rate. We might expect that rate reduction would be of particular value in patients with mitral stenosis, as increases in end-diastolic dimerisions are related, OVer a certain range, to the duration of diastole. However, there is little evidence to suggest that patients with mitral stenosis do derive any particular benefit from the 'better' rate control (e.g. exercise heart rates of less than 140 beats/min) which can be achieved with combinations of digoxin plus either calcium antagonists or /3-blockers (Lewis et al. I 988b). 3.3.2 Digoxin Digoxin, which has now been in clinical use for over 200 years, is widely regarded as the drug of choice for the long term treatment of chronic atrial fibrillation. In this context, the therapeutic effects of the cardiac glycosides are generally thought to be due primarily totheir rate~limitingeffects rather than their inotropic actions. Nevertheless; it has been known for many years that digoxin fails to control exercise-induced tachycardia, even when plasma drug concentrations are near the upper end of the accepted 'therapeutic' range (Beasley .et al. 1985). This is predictable from the Qasic pharmacology of the cardiac glycosides, as they have little direct effect on AV nodal conduction, and their
Atrial Fibrillation: The Therapeutic Options
major therapeutic effects are mediated via increased vagal tone and inhibition of sympathetic stimulation (Klein & Kaplinsky 1986). During exercise, intrinsic vagal tone is withdrawn and sympathetic activity increases so that digoxin appears to have relatively little effect on postexercise heart rate compared with placebo (Lang et al. 1983; Lewis et a1. 1988a). The reductions in heart rate observed during early experience with digoxin in severe congestive cardiac failure may, in fact, have been a secondary phenomenon reflecting an improvement in cardiac function brought about by some other mechanism. While the cardiac glycosides fail to control exercise-induced tachycardia, it is clear that digoxin can achieve marked reductions in the resting heart rate. This may predispose to periods of quite marked nocturnal bradycardia, because a diurnal variation in heart rate is frequently seen in subjects with atrial fibrillation (Pitcher et al. 1986). Digoxin has a narrow therapeutic ratio and toxicity may be associated with potentially lethal arrhythmias. The threshold for glycoside toxicity shows considerable interindividual variability and both ·hypokalaemia and hypomagnesaemia · predispose to cardiac toxicity. Conversely, serum concentrations of theglycoside·of less than about 1 nmoljL are thought to be largely ineffectual. Within the accepted 'therapeutic'range (1.3 to 2.6 nmoljL) there is evidence of a weak relationship between plasma drug concentration and pharmacological effect (Beasley et al. 1985). Digoxin is generally well tolerated and the more common adverse effects such as nausea or other gastrointestinal disturbances may respond to a decrease in drug dosage. However, there is some concern that even 'therapeutic' plasma concentrations of digoxin may increase the preva1enee of ventricular arrhythmias (Lewis & McDevitt .1988). There is also evidence suggesting that
