Journal Logo


Hysteresis in Atrial Refractoriness in the Conscious Dog: Influence of Stimulation Parameters and Control by the Autonomic Nervous System

Boucher, Michel*; Chassaing, Claude; Chapuy, Eric; Lorente, Paco†‡

Author Information
Journal of Cardiovascular Pharmacology: December 1996 - Volume 28 - Issue 6 - p 842-847
  • Free


One of the most outstanding features of changes in cardiac refractoriness subsequent to an abrupt variation of the heart rate is the existence of a memory effect (1-3) [i.e., as proposed as early as 1913 by Mines (4), the “previous history of the cardiac tissue may influence its immediate behavior”]. Moreover, during various biologic parameter measurements, a hysteresis effect has very often been reported (5-8). Giorgi et al. (9) observed hysteresis during ventricular effective refractory period determinations in canine ventricular myocardium, depending on whether the diastole was scanned with a decreasing or increasing S1S2 interval. To our knowledge, no study had been carried out hitherto on hysteresis during atrial effective refractory period (AERP) measurements. Clinically, there has been increasing interest in atrial arrhythmias, particularly atrial fibrillation, since their occurrence and associated morbidity have proved higher than previously recognized (10). For a better understanding of the pathophysiology of these arrhythmias, atrial refractory period distribution has been studied inter alia by mapping (11), but these techniques are limited because the refractory periods are acquired sequentially and nonsimultaneously over an extended period. Therefore it is important to know to what extent the stimulation protocol and, more generally, the immediate past of the tissue may influence the AERP measurements. In addition, atrial refractory period alterations may be important in determining functional substrates prone to atrial arrhythmias. The autonomic nervous system is known to interfere with atrial refractoriness (12,13). It is therefore relevant to assess the influence of the autonomic nervous system on the memory effect. Thus the aim of this work was (a) to provide evidence for hysteresis in the AERP in the conscious dog, (b) to study the stimulation parameters that may affect this phenomenon, and (c) to determine the influence of the autonomic nervous system. The use of conscious dogs allows the control of cardiac activity to be fully exerted by the autonomic nervous system.


Six mongrel dogs of either sex, weighing between 13 and 22 kg, were used in this study. They were housed in individual cages in a large colony room, with food and water continuously available in their home cages. The study conformed to the National Institutes of Health (NIH) Guidelines for Care and Use of Laboratory Animals.

Surgical preparation and instrumentation

In the six dogs, atrioventricular block was induced under sodium pentobarbital anesthesia and aseptic conditions by crushing the His bundle with forceps introduced through the open right atrium during temporary occlusion of the venae cavae [modified Fredericq's technique (14,15)]. Two wired stainless steel electrodes were implanted 1.5 cm apart on the external surface of the right atrium near the sinoatrial node, and the leads were exteriorized through the neck. In addition, two ventricular electrodes were implanted 1.5 cm apart on the external surface of the right ventricle near the atrioventricular junction, and the leads were also exteriorized through the neck. Atrioventricular block allows AERP to be measured and ventricles simultaneously to be paced to minimize the possible interferences caused by either drug- or atrial pacing-induced ventricular chronotropic effects. All the dogs were left to recover for ≥8-10 days before experiments were performed.


ECG monitoring was performed with a Cardiopan III T instrument (Massiot-Philips, Clermont-Ferrand, France). AERP was measured by the extrastimulus method involving single premature atrial stimuli S2 applied during the increasing and decreasing phases of an S1S2 fixed protocol, in which the S1 stimuli were pacing stimuli. Hysteresis, as defined by Giorgi et al. (9), was calculated as the difference between the two values of AERP obtained during the two phases carried out at random (Fig. 1). Single premature atrial stimuli S2 were shifted toward (decreasing protocol) or away from (increasing protocol) the preceding S1 stimulus in 2-ms steps. In both protocols, S1S2 length plus S2S1 length was equal to 2 S1S1 lengths, and the number of S2 extrastimuli (≈10) applied before reaching the value of AERP was as far as possible identical (see also the number of trains). Atrial pacing was applied in 2-ms rectangular pulses from a Janssen programmable stimulator (Janssen Scientific Instruments, Paris, France); the stimulation voltage was 1.5 times the diastolic threshold voltage. These two parameters, stimulation intensity and pulse duration, were shown to be ineffective on hysteresis during the preliminary experiments, in agreement with the previous results of Giorgi et al. (9), and were thus not studied further. Pacing frequency (i.e., S1S1 basic cycle length) and number of basic cycles in trains before each S2 extrastimulus were studied as specified in the protocol. During recording, the previously trained dogs were placed on a table and lightly restrained. A microcatheter was fitted in a branch of the saphenous vein before tests in which pharmacologic agents were used, to allow painless drug administration.


First, to evidence hysteresis and study the stimulation parameters able to affect this phenomenon, hysteresis, calculated as indicated previously, was measured with a S1S1 basic cycle length of either 400 or 300 ms set at random. For each of these two values, hysteresis was measured with trains of either six or 12 basic cycles before each S2 extrastimulus in random order. In both protocols (decreasing and increasing), the number of trains (≈10) applied before reaching the value of AERP was, as far as possible, identical.

Second, to test the influence of the autonomic nervous system, hysteresis in AERP was also determined after blockade of muscarinic cholinoceptors with 200 μg/kg/h i.v. atropine (sulfate), after blockade of β-adrenoceptors with 0.5 mg/kg i.v. propranolol (hydrochloride), after simultaneous blockade of β-adrenergic and muscarinic receptors (propranolol plus atropine), and after 10 μg/kg i.v. neostigmine (methylsulfate) and 0.02 μg/kg/min i.v. isoproterenol (hydrochloride); these latter two drugs mimic the cardiac effects of the autonomic nervous system. Hysteresis was measured before and after the different drugs with an S1S1 basic cycle length of 300 ms and trains of six basic cycles, the parameter values for which hysteresis was shown to be greatest. In addition, to avoid any interferences caused by drug-induced ventricular chronotropic effects, ventricular pacing (2-ms rectangular pulses, 1.25 times diastolic threshold voltage, frequency higher than the estimated ventricular rate after the administered drug) was applied from a Hugo Sachs Electronik 6512 stimulator (Hugo Sachs Electronik, March-Hugstetten, Germany) throughout a given experiment. The muscarinic receptors were blocked with an infusion of atropine rather than a single injection, to maintain a stable atrial cardioacceleration throughout the experiments. Propranolol was used because in this model, a dose of 0.5 mg/kg could be administered, inducing a very high degree of β-adrenoceptor blockade (as verified by antagonism of isoproterenol 1 μg/kg i.v.) without producing any significant atrial tachycardia, unlike some other β-blockers (16,17). The cardioaccelerator effects of the sympathetic nervous system were mimicked with an infusion of isoproterenol to maintain a stable atrial cardioacceleration throughout the experiments, and the cardiomoderator effects of the parasympathetic nervous system by using the acetylcholinesterase inhibitor neostigmine. Hysteresis was measured 15 min after the start of atropine infusion and 5 min after the start of isoproterenol infusion, when the atrial rate had achieved a stable plateau, and likewise 10 min after the injection of propranolol and 5 min after the injection of neostigmine. At least 72 h elapsed between successive tests performed on the same dog at random.


Atropine sulfate was purchased from Fluka (Buchs, Switzerland), isoproterenol HCl from Winthrop Laboratories (Clichy, France), neostigmine methylsulfate from Roche Laboratories (Neuilly-sur-Seine, France), pentobarbital Na from Abbott Laboratories (Saint-Rémy-sur-Arve, France), and propranolol HCl was supplied by I.C.I.-Pharma Laboratories (Cergy, France).

Data analysis

Results were expressed as arithmetic means ± SEM. Statistical analysis was performed by using Student's t test for paired samples.


Evidence of AERP hysteresis and influence of the stimulation parameter values

Whatever the stimulation parameters used, AERP was longer during the increasing phase of an S1S2 fixed protocol than during the decreasing phase, thus demonstrating the existence of hysteresis in the AERP (Figs. 2 and 3). Hysteresis was between 6 ± 1.5 and 10 ± 1.3 ms under our experimental conditions. When the results were analyzed to determine the effects of the S1S1 basic cycle length and of the number of basic cycles per train before each S2 extrastimulus, hysteresis was greater (p < 0.01) with a basic cycle length of 300 ms than with a basic cycle length of 400 ms (9 ± 0.9 and 7 ± 0.9 ms, respectively; Fig. 4), and greater (p < 0.01) with trains of six basic cycles than with trains of 12 basic cycles (9 ± 0.9 and 7 ± 1.0 ms, respectively; Fig. 4). To optimize the data obtained, the stimulation parameters of 300 ms and six cycles per train were used for the study of the autonomic nervous system effects.

Effects of the different drug pretreatments on AERP

For all the experiments carried out by using the increasing protocol with S1S1 equal to 300 ms and six cycles per train, control AERP values were between 130 and 149 ms with a mean value of 140 ± 2.2 ms. The effects of the different pretreatments are shown in Table 1. Blockade of muscarinic cholinoceptors (p < 0.05), β-adrenoceptors (p < 0.01), and both β-adrenergic and muscarinic receptors (p < 0.01) increased AERP, whereas inhibition of acetylcholinesterase (p < 0.05) and stimulation of β-adrenoceptors (p < 0.01) decreased it.

Influence of the autonomic nervous system on AERP hysteresis

After blockade of muscarinic cholinoceptors with atropine, hysteresis was reduced from 8 ± 0.6 to 4 ± 0.6 ms (p < 0.001) (Fig. 5), whereas after β-adrenoceptor blockade with propranolol, it was increased from 9 ± 0.9 to 14 ± 1.2 ms (p < 0.001; Fig. 5). After simultaneous blockade with propranolol and atropine, as after atropine alone, hysteresis was reduced from 9 ± 0.9 to 6 ± 0.8 ms (p < 0.01; Fig. 5). These results were entirely confirmed by those obtained after neostigmine and isoproterenol. Thus after neostigmine, hysteresis was increased from 9 ± 0.8 to 11 ± 0.8 ms (p < 0.01; Fig. 6), and after isoproterenol, it was decreased from 9 ± 0.6 to 4 ± 0.4 ms (p < 0.001; Fig. 6).


The results obtained in this study demonstrate, for the first time, the presence of a hysteresis phenomenon in the AERP in conscious dogs. It does not strictly comply with the classic definition of hysteresis, which for a given value of an input variable corresponds to two different stable states. However, as indicated in Methods, a phenomenon identical to ours has already been termed hysteresis (9). This hysteresis varied here between 6 and 10 ms according to the stimulation parameters used. A hysteresis phenomenon has already often been reported in various biologic parameter measurements (5-8,18). The clear evidence of AERP hysteresis is in the same line as the observations of Giorgi et al. (9), who found hysteresis of the ventricular effective refractory period in anesthetized open-chest dogs and in isolated ventricular muscle. Our hysteresis was greater for a shorter basic cycle length and smaller basic cycle trains, and so conflicts with the results of Giorgi et al. (9) that failed to demonstrate any effect of the basic cycle length on hysteresis. This discrepancy may be the result of the very small hysteresis (≈3 ms), most likely arising from the vagolytic effects of anesthesia plus thoracotomy (see discussion on atropine effects) and thus to the impossibility of obtaining significant variations under the experimental conditions of that study. The AERP hysteresis is therefore stimulation frequency-dependent and is very likely related to the restitution of the action potential duration. Many studies have reported data showing the influence of stimulation frequency on the action potential duration in multicellular preparations of either Purkinje tissue or ventricular muscle (19-22) or in enzymatically dissociated cells (23) or even in vivo in human ventricles (24). It is generally assumed that an inverse linear relation exists between action potential duration and stimulation frequency. The AERP hysteresis may result from the two basic processes involved in the electrical restitution: first, the recovery of membrane ionic currents lasting a few hundred ms only, and second, a memory effect related to the action of membrane pumps and intracellular ionic exchanges.

An important finding reported here is the role played by the autonomic nervous system in the control of the hysteresis. Atropine, which lengthened AERP, in agreement with reported data (12), decreased hysteresis by ≈50%, whereas propranolol, which as expected (13) also lengthened AERP, increased hysteresis by ≈50%. Eventually blockade of both vagal and cardioaccelerator tones led to a decrease in hysteresis by ≈35%. Therefore it can be assumed that cardiac vagal tone exerts a permanent increase effect on AERP hysteresis, whereas cardioaccelerator tone exerts a permanent decrease effect on this phenomenon, which means that vagal tone very likely increases the memory effect, whereas cardioaccelerator tone decreases it. Blockade of both systems also demonstrates that, like the other cardiac parameters, AERP hysteresis is mainly under the control of cardiac vagal tone. The results observed after neostigmine and isoproterenol, (i.e., hysteresis increase and decrease, respectively) indicate that AERP hysteresis and thus the memory effect also can be modulated by cardiac vagal or sympathetic stimulation or both.

Overall our results show that in the conscious dog, the hysteresis in AERP is stimulation frequency-dependent and strongly modulated by the autonomic nervous system. This phenomenon, the relevance and the implications of which in cardiac physiology and pathophysiology are as yet unknown, deserves further investigation to assess its possible clinical impact.

Acknowledgment: We thank Danièle Hosmalin for her valuable assistance in the preparation of this manuscript.

This work was partly supported by the University of Clermont-Ferrand I (BQR).

FIG. 1.
FIG. 1.:
Stimulation protocol and atrial effective refractory period (AERP) measurement by the extrastimulus method during the increasing (top) and decreasing (bottom) phases of an S1S2 fixed protocol. Hysteresis (H) was defined as the difference between the two values of AERP obtained. All the S1 or S2 stimuli depicted with a large arrow were followed by an atrial depolarization; S2 stimuli depicted with a small arrow were not. S1, atrial pacing stimuli; S2, atrial premature extrastimuli.
FIG. 2.
FIG. 2.:
Typical recordings of an experiment showing the existence of hysteresis in the atrial effective refractory period (AERP) in conscious dogs with chronic atrioventricular block. All traces are standard electrocardiographic leads II and III. Pulse width was 2 ms, and stimulation intensity 1.5 times the diastolic threshold. During the increasing (nearr S1S2) and decreasing (searr S1S2) phases of the protocol, AERP values were 132 and 122 ms and 136 and 130 ms, respectively.
FIG. 3.
FIG. 3.:
Atrial effective refractory period (AERP) values in six conscious dogs with chronic atrioventricular block during an increasing protocol (left of each panel) and a decreasing protocol (right of each panel). Stimulation parameters were S1S1 basic cycle length of 300 or 400 ms with trains of either six or 12 basic cycles at random. Hysteresis (H) was calculated as the difference between the values of AERP obtained during the two protocols. H = 8 ± 1.3 ms (A), H = 10 ± 1.3 ms (B), H = 6 ± 1.5 ms (C), and H = 8 ± 1.2 ms (D).
FIG. 4.
FIG. 4.:
Influence of the stimulation parameter values: S1S1 basic cycle length of 400 vs. 300 ms and trains of 12 basic cycles vs. six on hysteresis in atrial effective refractory period in conscious dogs with chronic atrioventricular block. Values are means for groups of six dogs. Vertical lines show SEM. *p < 0.05 and **p < 0.01.
FIG. 5.
FIG. 5.:
Influence of atropine, propranolol, and propranolol plus atropine on hysteresis in atrial effective refractory period in conscious dogs with chronic atrioventricular block. Stimulation parameters were S1S1 basic cycle length of 300 ms and trains of six basic cycles. Ventricles were paced throughout the experiments. Values are means for groups of six dogs. Vertical lines show SEM. *p < 0.05 and ***p < 0.001 in comparison with control values.
FIG. 6.
FIG. 6.:
Influence of neostigmine and isoproterenol on hysteresis in atrial effective refractory period in conscious dogs with chronic atrioventricular block. Stimulation parameters were S1S1 basic cycle length of 300 ms and trains of six basic cycles. Ventricles were paced throughout the experiments. Values are means for groups of six dogs. Vertical lines show SEM. **p < 0.01 and ***p < 0.001 in comparison with control values.


1. Janse MJ, Van Der Steen ABM, Van Dam RT, Durrer D. Refractory period of the dog's ventricular myocardium following sudden changes in frequency. Circ Res 1969;24:251-62.
2. Denker S, Shenasa M, Gilbert CJ, Akhtar M. Effects of abrupt changes in cycle length on refractoriness of the His-Purkinje system in man. Circulation 1983;67:60-8.
3. Lehmann MH, Denker S, Mahmud R, Akhtar M. Patterns of human atrioventricular nodal accommodation to a sudden acceleration of atrial rate. Am J Cardiol 1984;53:71-6.
4. Mines GR. On dynamic equilibrium in the heart. J Physiol (Lond) 1913;46:349-83.
5. Katchalski A, Neumann E. Hysteresis and molecular memory record. Int J Neurosci 1972;3:175-82.
6. Cukierman S, Paes de Carvalho A. Frequency-dependent excitability of “membrane” slow responses of rabbit left atrial trabeculate in the presence of Ba2+ and high K+. J Gen Physiol 1982;79:1017-39.
7. Chu PL, Freedman AR, Fleming S, Malik M, Camm AJ, Ward DE. Hysteresis of the ventricular paced QT interval in response to abrupt changes in pacing rate. Cardiovasc Res 1988;22:67-72.
8. Lorente P, Delgado C, Delmar M, Henzel D, Jalife J. Hysteresis in the excitability of isolated guinea pig ventricular myocytes. Circ Res 1991;69:1301-15.
9. Giorgi C, Vermeulen M, Cardinal R, Savard P, Nadeau R, Shenasa M. Demonstration of hysteresis in refractoriness of canine ventricular myocardium. Am J Physiol 1988;255:H1342-8.
10. Schuessler RB, Boineau JP, Bromberg BI, Hand DE, Yamauchi S, Cox JL. Normal and abnormal activation of the atrium. In: Zipes DP, Jalife J, eds. Cardiac electrophysiology: from cell to bedside. Philadelphia: WB Saunders, 1995:543-62.
11. Schoels W, Kuebler W, Yang H, Gough WB, el-Sherif N. A unified functional/anatomic substrate for circus movement atrial flutter: activation and refractory patterns in the canine right atrial enlargement model. J Am Coll Cardiol 1993;21:73-84.
12. Duchêne-Marullaz P, Lavarenne J, Combre A, Lapalus P, Boucher M. Modification de la période réfractaire auriculaire par suppression de l'innervation cardiomodératrice chez le chien chloralosé. C R Soc Biol 1971;165:1908-10.
13. Boucher M, Chapuy E, Duchêne-Marullaz P. Membrane stabilizing activity and beta-adrenoceptor antagonist-induced bradycardia in conscious dogs. Eur J Pharmacol 1992;211:343-9.
14. Fredericq L. L'atriotomie temporaire, procédé nouveau d'exploration des fonctions du coeur. Arch Int Physiol 1904;1:83-5.
15. Boucher M, Duchêne-Marullaz P. Methods for producing experimental complete atrioventricular block in dogs. J Pharmacol Methods 1985;13:95-107.
16. Duchêne-Marullaz P, Combre A, Lavarenne J, Lapalus P, Schaff G. Comparaison des effets du propranolol, de l'alprénolol, du pindolol et du practolol sur les rythmes cardiaques de chiens non narcosés en dissociation auriculo-ventriculaire chronique. J Pharmacol 1975;6:441-52.
17. Boucher M, Duchêne-Marullaz P. Acebutolol, metoprolol and propranolol in conscious dogs with chronic heart-block: chronotropic effects and relation between depression of ventricular activity and beta-adrenoceptor blocking potency. Br J Pharmacol 1980;70:335-40.
18. Mobbs CV. Molecular hysteresis: residual effects of hormones and glucose on genes during aging. Neurobiol Aging 1994;15:523-34.
19. Boyett MR, Jewell BR. A study of the factors responsible for rate-dependent shortening of the action potential in mammalian ventricular muscle. J Physiol (Lond) 1978;285:359-80.
20. Colatsky TJ, Hogan PM. Effects of external calcium, calcium channel-blocking agents, and stimulation frequency on cycle length-dependent changes in canine cardiac action potential duration. Circ Res 1980;46:543-52.
21. Elharrar V, Surawicz B. Cycle length effect on restitution of action potential duration in dog cardiac fibers. Am J Physiol 1983;249:H782-92.
22. Elharrar V, Atarashi H, Surawicz B. Cycle length-dependent action potential duration in canine cardiac Purkinje fibers. Am J Physiol 1984;247:H936-45.
23. Robinson RB, Boyden P, Hoffman BF, Hewett KW. Electrical restitution process in dispersed canine cardiac Purkinje and ventricular cells. Am J Physiol 1987;253:H1018-25.
24. Franz MR, Swerdlow CD, Bing Liem L, Schaefer J. Cycle length dependence of human action potential duration in vivo. J Clin Invest 1988;82:972-9.

Atrial effective refractory period; Extrastimulus method; Conscious dogs; Atropine and neostigmine; Propranolol and isoproterenol; Memory effect

© Lippincott-Raven Publishers