OUP user menu

Cardiac alternans: mechanisms and pathophysiological significance

David E Euler
DOI: http://dx.doi.org/10.1016/S0008-6363(99)00011-5 583-590 First published online: 1 June 1999

Time for primary review 31 days.

1 Introduction

Mechanical alternans (pulsus alternans) is a condition in which there is a beat-to-beat oscillation in the strength of cardiac muscle contraction at a constant heart rate. Since the first description of pulsus alternans by Traube in 1872 [1], there has been continuing interest in understanding the mechanisms and clinical manifestations of this phenomenon [2,3]. Initially observed in the hearts of laboratory animals, the phenomenon has been reported frequently in patients particularly those with severe heart failure and aortic valve disease. Although much is known about the cellular mechanisms of alternans in isolated cardiac muscle preparations, little is known about the mechanisms by which changes in preload or afterload evoke alternans in the intact heart. The purpose of this review article is to discuss the mechanisms of alternans at both the cellular level and in the intact heart. The relationship between mechanical and electrical alternans is reviewed. The role of sympathetic nervous system is discussed as a means of protecting the heart from alternans during accelerations of heart rate. Since pulsus alternans has been reported in patients with cardiac disease, the diagnostic and prognostic significance of mechanical and electrical alternans is also discussed.

2 Induction of alternans

The ability to induce mechanical alternans by rapid driving frequencies appears to be a fundamental property of mammalian ventricular muscle. Experimental studies have shown that by varying the pacing cycle length over a wide range, it is possible to define a critical cycle length (threshold) for the induction of sustained mechanical alternans [4–7]. Driving the heart at cycle lengths shorter than the threshold cycle length may increase the amplitude of the beat-to-beat oscillations in contraction strength (Fig. 1). Driving the heart at cycle lengths just longer than the threshold may produce transient alternans [4,6]. Transient alternans may also be observed after a single premature systole [8,9]. Since premature systoles may also be followed by postextrasystolic potentiation, the critical factor that seems to predict the occurrence of post-extrasystolic alternans is heart rate. Postextrasystolic potentiation disappears and is replaced by postextrasystolic alternans as heart rate increases [9].

Fig. 1

Tracings showing effect of shorter cycle lengths (CL) on mechanoelectrical alternans in a feline ventricular myocytes. (A) Concordant electrical (upper trace) and mechanical (cell length change; lower trace) alternans of a cell stimulated at a cycle length of 290 ms. Beats with longer action potential duration (APDs) were associated with larger contraction. Numbers under each action potential are APD in ms at 90% repolarization. Right inset shows two sequential action potentials superimposed. (B) Concordant electrical (upper trace) and mechanical (lower trace) alternans of the same cell at a shorter cycle length (250 ms). Reproduced from: Rubenstein DS, Lipsius SL. Premature beats elicit a phase reversal of mechanoelectrical alternans in cat ventricular myocytes. A possible mechanism for reentrant arrhythmias. Circulation 1995:91;201–214. Reprinted with permission.

The threshold cycle length for sustained mechanical alternans varies among different mammalian species [6,10] and is influenced by a number of intrinsic and extrinsic factors. An increase in the threshold cycle length (slower heart rate) for alternans is brought about by conditions such as hypothermia [4,10–13], hypocalcemia [5,10,12,14], hypercapnic acidosis [15,16], ischemia [17–21], hypertrophy [22], and congestive heart failure [23,24]. A decrease in the threshold cycle length for pulsus alternans has been reported in response to β-adrenergic agonists [5,7,10,25], hypercalcemia [5,10], digitalis [5] and calcium channel antagonists [13,19,26,27]. From this list of factors, it is evident that both positive and negative inotropic interventions have the capacity to abolish or suppress mechanical alternans.

3 Mechanism of alternans

Since it first description, two main mechanisms have been proposed to account for mechanical alternans. The first, based on the Frank–Starling relationship, proposes that the strong beat leaves a small residual end-systolic volume such that the end-diastolic volume before the next beat is reduced and there is reduced force development. The end-systolic volume is increased after the weak beat (due to decreased ejection) which leads to a greater end-diastolic volume and more force development in the next beat. Rapid heart rates would accentuate this process because of the effects of heart rate on the time available for diastolic filling. Studies in both intact and isolated working hearts have demonstrated an alternation of left ventricular end-diastolic volume that is consistent with this mechanism [25,28–31].

The second explanation is that there is beat-to-beat alternation of myocardial contractility. This theory is supported by the demonstration of alternans in isolated papillary muscles contracting isometrically [6,8,10,12,26] and in isolated ventricular myocytes [16,32]. Additional support for this explanation comes from experiments in isolated hearts were alternans was observed with ventricular volume held constant by an intraventricular balloon [23,24]. Evidence for an alternation of the inotropic state in the intact heart comes from measurements of the end-systolic pressure–volume relationship. The slope of the pressure–volume relationship is thought to provide an index of contractility that is independent of both preload and afterload. The strong beat during pulsus alternans has been shown to have a significantly elevated end-systolic pressure–volume slope compared to the weak beat [29–31]. Since mechanical alternans can be observed under a wide variety of different experimental and clinical conditions, it is possible that in the intact heart, both the Frank–Starling mechanism and contractility play a role. It may be that the relative contribution of each mechanism depends upon the conditions under which pulsus alternans is observed.

4 Alternation of intracellular Ca2+

The alternation of inotropic state that mediates the alternation in force of contraction is thought to be caused by an alternation of the intracellular Ca2+ transient. Direct measurements of the Ca2+ transient using Ca2+-sensitive indicators have shown an alternation of the Ca2+ transient during mechanical alternans in both isolated cardiac muscle preparations [15,16,22,33] and isolated perfused hearts [23,34]. The Ca2+ transient alternates in phase with the strength of the contraction. The alternation of the Ca2+ transient is load-independent since it was observed during both isometric and unloaded isotonic contractions [15]. Presumably, the alteration of the Ca2+ transient reflects an alternation of the release of Ca2+ from the sarcoplasmic reticulum (SR). The importance of the SR comes from the observation that ryanodine, a specific inactivator of the SR Ca2+ release channel, consistently abolishes mechanical alternans [15,22,32,33,35]. Although less specific in its effect on SR Ca2+ release, caffeine has also been shown to inhibit or abolish mechanical alternans [13,15,24,33,35].

Although the mechanism responsible for the alternation of SR Ca2+ release is not entirely clear, it can be explained by a model of excitation–contraction coupling that proposes that SR Ca2+ is stored in two separate compartments. An uptake compartment recovers Ca2+ from the cytoplasm and transports it to a release compartment which is responsible for most of the rise of intracellular Ca2+ when the cell is stimulated. The presence of a delay between uptake and release of SR Ca2+ has been postulated to account for the mechanical behavior associated with postextrasystolic potentiation [36]. It has been suggested that mechanical alternans occurs because there is not sufficient time between successive contractions for Ca2+ to be taken up into the uptake compartment and transported to the release compartment [15,33]. The weaker beat is caused by less Ca2+ release from the SR, since much of the Ca2+ released during the preceding beat did not have sufficient time to be recycled to the release compartment. The reduced amount of Ca2+ in the cytoplasm during the weak beat would require less time to be taken up and recycled to the release compartment. Furthermore, there now would have been two diastolic intervals to recycle the Ca2+ released during the strong beat that did not have time to be recycled and released during the weak beat.

Although the two-compartment model of SR Ca2+ cannot account for all of the experimental interventions that induce or suppress alternans, it does explain the ability of interventions that slow SR Ca2+ recycling such as hypothermia and ischemia to induce mechanical alternans. A rapid heart rate would decrease the diastolic time available for SR Ca2+ recycling and increase the probability of oscillations in SR Ca2+ release. The ratio of diastole to systole may be even more important than the absolute heart rate in the genesis of alternans, since the majority of SR Ca2+ recycling would occur during diastole. It is well known that a sudden increase in heart rate to a rate below the threshold for sustained alternans may induce transient alternans. A sudden increase in heart rate is accompanied by a gradual decrease in action potential duration that should shorten the duration of systole [37]. The reciprocal prolongation of diastole as systole shortens may act as an adaptive mechanism to increase the time available for SR Ca2+ recycling and damp the oscillations of SR Ca2+ release. The duration of diastole relative to systole may also be the critical factor that determines if transient alternans occurs in response to a single premature systole [9].

5 Relaxation alternans

Several studies have shown that an alternation of the rate of relaxation accompanies the alternation in force development during mechanical alternans [7,31,38,39]. In an in situ blood-perfused papillary muscle, where alternans was induced by ischemia–reperfusion and rapid pacing, the rate of force decay of the strong twitch was significantly slower than the rate of force decay of the weak twitch [39]. The incomplete relaxation after the strong twitch was thought to be directly responsible for the reduced force of the subsequent twitch. However, studies in intact hearts have shown opposite results. In isolated blood-perfused canine ventricles with cycle length-induced alternans, the diastolic pressure–volume relationship for the strong and weak beats were not significantly different [30]. In conscious dogs with alternans evoked by inferior vena cava occlusion, the time constant of isovolumic relaxation (τ) was significantly shorter (faster relaxation) during the strong beat [31]. Likewise, in anesthetized dogs with alternans induced by rapid atrial pacing, τ was shorter during the beat with greater pressure development [7]. These findings would argue against incomplete relaxation after the strong beat as a factor contributing to less pressure development during the weak beat. The observation that isovolumic relaxation is accelerated in the beat with higher pressure is contrary to predictions based on the effects of loading conditions on the rate of relaxation [40,41]. In contrast to the changes in τ observed experimentally, a study in 12 patients with severe aortic valve disease and pulsus alternans during sinus rhythm showed no significant difference in τ between the strong and weak beats [42]. Furthermore, in five patients with alternans induced by interior vena cava occlusion, there was no difference in peak negative dp/dt or τ between the strong and weak beats [43].

It has been proposed that abnormalities of left ventricular relaxation precede the development of pulsus alternans and thus may play a seminal role in the genesis of the alternation in systolic force. When mechanical alternans was induced by vena cava occlusion in the intact dog heart, an alternation of τ preceded an alternation of LV systolic pressure or peak positive dp/dt by several cardiac cycles [31]. Furthermore, studies in anesthetized pigs using strain gauges to measure regional contractile force showed that pacing-induced alternans of regional relaxation preceded the onset of regional alternans in contractility [38]. However, studies in anesthetized dogs using incremental atrial pacing to induce pulsus alternans reported that the threshold cycle length to induce an alternation of systolic pressure or peak positive dp/dt was not different form the threshold cycle length to evoke an alternation of τ [7]. Certainly additional studies will be required to further elucidate to role of relaxation alternans in the development of pulsus alternans.

6 Importance of preload and afterload

It has long been recognized that changes in both preload [31,43–46] and afterload [42,47–51] can lead to mechanical alternans. An elegant study in conscious dogs by Freeman et al. showed that occlusion of the inferior vena cava consistently led to pulsus alternans [31]. Although there was an alternation of end-diastolic volume in the left ventricle, the alternation of systolic performance was not due exclusively to the Frank–Starling mechanism since there was alternans of the end-systolic pressure–volume ratio. Because reflex changes in autonomic tone were eliminated by ganglionic blockade, the pulsus alternans occurred in the absence of any reflex tachycardia [31]. In the clinical setting, occlusion of the inferior vena cava was shown to induce pulsus alternans in 5/11 patients with nonischemic cardiomyopathy [43]. Fig. 2 shows left ventricular pressure recordings obtained before and after occlusion of the interior vena cava in one of the five patients that had inducible alternans. Echo-derived measures of ventricular performance demonstrated that the weak beat began at a smaller end-diastolic diameter than the strong beat [43].

Fig. 2

Left ventricular pressure hemodynamics before and during pulsus alternans induced by IVC balloon occlusion in a patient with nonischemic cardiomyopathy. The dramatic alteration in LV diastolic pressures are noted in association with a slight increase in the heart rate. Despite significant systolic pulsus alternans, there is little difference in the end-diastolic pressures between weak and stoning betas. In this patient both peak positive dp/dt and peak negative dp/dt alternans is present. Reproduced from: Bashore TM, Walker S, Van Fossen D, Shaffer PB, Fontana ME, Unverferth DV. Pulsus alternans induced by inferior vena caval occlusion in man. Cathet Cardiovasc Diagn 1988;14:24–32. Reprinted by permission of Wiley–Liss, a subsidiary of Wiley.

Pulsus alternans has also been observed during sinus rhythm or after premature beats in patients with aortic stenosis [42,47–50] or subaortic stenosis [51]. Laskey et al. [50] used M-mode echocardiography to measure LV dimensions and found that end-diastolic minor-axis diameter was not different between the strong and weak beats. However, end-systolic minor-axis was greater for the weak beats at any given level of afterload suggesting that there was an alternation in myocardial contractility. Hess et al. [42] used cineangiography to measure left ventricular volume and reported that end-diastolic volume was not different between the strong and weak beats, but that the strong beats had a significantly greater ejection fraction than the weak beats. Pulsus alternans has also been reported in the right ventricle (in the absence of alternans in the left ventricle) in patients with elevated afterload due to pulmonary hypertension [52,53]. Mechanical alternans in the right ventricle has also been reported during the inflation of a valvuloplasty balloon in the pulmonary artery [54].

The development of pulsus alternans in response to preload or afterload changes is difficult to explain. Since there is an alternation of end-diastolic dimensions in pulsus alternans induced by preload reduction, it is possible that the Frank–Starling mechanism may play a role in this type of alternans. Alternatively, it is possible that loading conditions have some unrecognized influence on SR Ca2+ handling. Since systolic load is known to influence relaxation time in the intact heart [40,41], it may be that pulsus alternans develops in response to load-dependent changes in the rate of relaxation. Another possibility is that changes in ventricular wall stress or coronary perfusion pressure with alterations of preload or afterload may lead to subendocardial ischemia and ischemic-induced abnormalities in SR Ca2+ uptake and release. Additional work will be necessary to define the precise mechanisms by which altered loading conditions elicit pulsus alternans.

7 Role of the sympathetic nervous system

Sympathetic influences on mechanical alternans are complex since sympathetic activation increases both heart rate and has direct effects on ventricular contractility and relaxation. The intravenous infusion of epinephrine into intact dogs abolished pulsus alternans evoked by rapid atrial pacing [5]. The beneficial effects of epinephrine were attenuated by propranolol indicating a role for β-adrenergic receptors [5]. A reflex increase of sympathetic tone caused by a reduction of carotid sinus pressure in intact dogs was also shown to eliminate pulsus alternans induced by rapid atrial pacing [25]. Direct electrical stimulation of the left sympathetic nerves in intact dogs increased the atrial pacing rate that was necessary to elicit mechanical alternans (Fig. 3) [7]. The ability of sympathetic stimulation to suppress mechanical alternans was prevented by β-adrenergic blockade with timolol [7].

Fig. 3

Electrical and mechanical alternans induced by rapid atrial pacing in the intact canine heart. The records show a lead II ECG and left ventricular pressure (LVP, mmHg). The top two traces were obtained during control conditions at a pacing cycle length of 300 ms. The bottom two recordings were obtained at the same pacing cycle length in the presence of left stellate stimulation (LSS) at a frequency of 1 Hz. Reproduced from: Euler DE, Guo H, Olshansky B. Sympathetic influences on electrical and mechanical alternans in the canine heart. Cardiovasc Res 1996;32:854–860.

In addition to increasing the slow inward Ca+2 current and Ca2+ release from the SR, β-adrenergic receptor activation also accelerates the rate of relaxation and increases the diastolic filling period. All of these effects may be important in preventing or abolishing mechanical alternans. It is quite likely that sympathetic effects on Ca2+ recycling plays an important role in preventing mechanical alternans during supraventricular or ventricular tachycardia. Although sympathetic activation may protect the normal heart from the development of pulsus alternans, an opposite effect may occur during myocardial ischemia. In isolated perfused canine hearts, mechanical alternans occurs when aerobic demand exceeds oxygen supply [18]. The increase of heart rate and the direct effects of β-adrenergic activation on energy consumption in ischemic myocardium may increase the severity of ischemia such that there might be an increased likelihood of either regional or global mechanical alternans.

8 Alternation of cardiac action potentials

Recordings of transmembrane action potentials in isolated multicellular preparations or single cardiac myocytes (Fig. 1) has shown that mechanical alternans is usually accompanied by an alternation of the duration of the action potential [6,10,12,16,26,33,35,55]. Furthermore, the recording of monophasic action potentials from the epicardial surface of the intact heart have also shown that mechanical alternans is accompanied by an alternation of monophasic action potential duration [13,21,56]. In addition, it is possible to observe T-wave alternans in the ECG recorded from the body surface during mechanical alternans (Fig. 3). Although T-wave alternans has only been occasionally reported to accompany pulsus alternans in the intact heart [7,13,57], a quantitative beat-to-beat analysis of ST and T voltages is required to definitively establish the presence of electrical alternans [7]. The close association between mechanical and electrical alternans has led to the suggestion that electrical alternans may be responsible for mechanical alternans. However, mechanical alternans has been reported in the absence of electrical alternans in both isolated papillary muscles [8] and in isolated myocytes driven with voltage-clamped pulses of a constant duration [16]. In addition, interventions such as caffeine and ryanodine that suppress or abolish mechanical alternans in ventricular muscle also eliminate electrical alternans [32,35]. Since electrical alternans in ventricular muscle is abolished by all interventions that abolish mechanical alternans, it would appear that mechanical alternans is responsible for electrical alternans.

A discordant relationship between alternation of action potential duration and alternation of the strength of the contraction is frequently observed in isolated cardiac muscle preparations [6,10,12,26,33,35,55]. The action potential with a shorter duration is associated with the stronger contraction. However, in isolated myocytes from the ferret heart exposed to hypercapnic acidosis, a concordant relationship between mechanical and electrical alternans was observed [16]. Furthermore, in isolated guinea pig papillary muscles, raising the temperature above 24°C resulted in a concordant alternation between mechanical and electrical activity [6]. Studies in feline myocytes showed that the relationship between mechanical and electrical activity was concordant at a temperature of 35°C (Fig. 1), but changed to discordant as the temperature was reduced to 26°C [32]. In the intact heart, a discordant relationship between the duration of the monophasic action potential and force development is usually observed during mechanical alternans [13,21,56]. However, when myocardial ischemia was produced by acute coronary artery occlusion in pigs, the relationship between local electrical alternans in the ischemic zone and global mechanical alternans became concordant [56].

Since electrical alternans may be either concordant or discordant with mechanical alternans, it is quite likely that there is more than one mechanism responsible for the coupling of mechanical and electrical activity during alternans. It has been proposed that the alternation of intracellular Ca2+ concentration during mechanical alternans influences one or more transmembrane currents that causes the duration of the action potential to alternate [15,16,58]. Differences in Ca2+ released from the SR during the strong and weak contractions may lead to an alternation of the Na+/Ca2+ exchange current that might influence action potential duration [16]. It is also possible that there is mechanoelectric feedback such that the strength of a contraction influences the duration of the action potential [21,59,60]. In the intact pig heart, pulsus alternans simulated by clamping the aorta on alternate beats was accompanied by an alternation of the duration of the monophasic action potential that was similar in magnitude to that evoked by rapid atrial pacing [21].

9 Diagnostic and prognostic significance

Experimental studies have shown that mechanical alternans is easier to induce by rapid pacing in hearts from animals with congestive heart failure [23,24]. Thus, it is possible that the induction of alternans in patients with heart failure may have diagnostic or prognostic significance. In an early clinical study, White [61] reported that that the presence of continuous pulsus alternans during sinus rhythm in a group of 15 patients with heart failure was associated with a 1-year mortality of 53%. In 55 patients that displayed transient pulsus alternans after a spontaneous premature beat, the 1-year mortality was only 31% [61]. However, Ryan et al. [62] did not find that the presence of pulsus alternans during sinus rhythm in patients with heart failure was associated with an unfavorable prognosis. In a more recent clinical study, Schafer et al. [63] attempted to induce mechanical alternans by rapid atrial pacing or atrial premature stimulation in 104 patients undergoing cardiac catheterization. Only 29 of the patients had inducible alternans and only 38% of these patients had a history of congestive heart failure. Eight patients (28%) that were inducible had a normal left ventricular ejection fraction (>0.5) and a normal end-diastolic pressure (<13 mmHg) [63]. Although prospective long-term studies are lacking, the clinical observations published to date suggest that pulsus alternans lacks specificity as a diagnostic test. Alternans present at rest or induced by atrial electrical stimulation does not appear to have the ability to discriminate between hearts with depressed contractility from hearts with normal contractility and abnormal pressure or volume loads. However, in the absence of significant valvular heart disease, spontaneous or inducible pulsus alternans may have prognostic significance in patients with congestive heart failure. Additional studies will be required to determine the long-term prognostic implications of pulsus alternans in patients with congestive heart failure.

10 Risk factor for arrhythmias

The importance of mechanical alternans as a risk factor for the development of serious ventricular arrhythmias is unknown. A local alternation of action potential duration caused by mechanical alternans could lead to increased spatial and temporal dispersion of refractoriness and increase the likelihood of the fractionation of conduction and reentry. The alternation of the ST-segment and T-wave during acute myocardial ischemia in experimental animals has been associated with an increased probability of ventricular tachycardia and fibrillation [64–66]. Alternation of the ST-segment and T-wave during acute myocardial ischemia may be caused by the same oscillations of SR Ca2+ release that elicit mechanical alternans [19,20,34]. Furthermore, an alternation of intracellular Ca2+ concentration may occur during ischemia in the absence of a detectable alternation of local contractile force. This is because severe ischemia leads to a loss of force development and often results in paradoxical systolic lengthening (bulging) in the ischemic zone. However, there are probably other mechanisms that mediate electrical alternans in ischemic myocardium in addition to an alternation of intracellular Ca2+ handling [56]. Ischemia is known to inhibit the fast Na+ current resulting in slowed conduction and postrepolarization refractoriness [67,68]. Studies in isolated blood-perfused pig hearts have shown that ST-segment and T-wave alternans during ischemia may be caused by an alternation of conduction into the ischemic zone [67].

Recent clinical studies have shown that T-wave alternans in the microvolt range serves an independent predictor of both inducible ventricular arrhythmias [69] and sudden cardiac death [70]. The amplitude of the T-wave alternans is such that it can only be detected using spectral signal-processing techniques. Furthermore, T-wave alternans is observed in the absence of acute myocardial ischemia [69,70]. Although it is unlikely that T-wave alternans in the microvolt range is accompanied by global mechanical alternans, it is possible that there may be local areas of ventricular myocardium with mechanical and electrical alternans. Alternatively, T-wave alternans in the microvolt range may be related to an alternation of the duration of action potentials in Purkinje fibers. Microelectrode studies in isolated canine Purkinje fibers have shown that alternans of the duration of the action potential is determined by the process controlling action potential duration during electrical restitution and is not linked to an alternation of SR Ca2+ release [35]. The precise role of local mechanical alternans in producing repolarization alternans in the surface ECG in patients at risk of malignant ventricular arrhythmias will require further investigation.


  1. [1]
  2. [2]
  3. [3]
  4. [4]
  5. [5]
  6. [6]
  7. [7]
  8. [8]
  9. [9]
  10. [10]
  11. [11]
  12. [12]
  13. [13]
  14. [14]
  15. [15]
  16. [16]
  17. [17]
  18. [18]
  19. [19]
  20. [20]
  21. [21]
  22. [22]
  23. [23]
  24. [24]
  25. [25]
  26. [26]
  27. [27]
  28. [28]
  29. [29]
  30. [30]
  31. [31]
  32. [32]
  33. [33]
  34. [34]
  35. [35]
  36. [36]
  37. [37]
  38. [38]
  39. [39]
  40. [40]
  41. [41]
  42. [42]
  43. [43]
  44. [44]
  45. [45]
  46. [46]
  47. [47]
  48. [48]
  49. [49]
  50. [50]
  51. [51]
  52. [52]
  53. [53]
  54. [54]
  55. [55]
  56. [56]
  57. [57]
  58. [58]
  59. [59]
  60. [60]
  61. [61]
  62. [62]
  63. [63]
  64. [64]
  65. [65]
  66. [66]
  67. [67]
  68. [68]
  69. [69]
  70. [70]