Atrioventricular (AV) nodal reentrant tachycardias are the most common form of supraventricular tachycardias (1). Many supraventricular tachycardias are believed to result from reentrant pathways within the AV node and, possibly, from the adjacent atrial myocardium. The substrates for AV-node reentry may be the atrial myocardium, the AV node, and two or more atrionodal inputs (2). In recent years, radiofrequency modification of the atrionodal inputs has proved to be an effective treatment for this condition, and the results of these procedures have largely supported this dual atrionodal input model of AV-node reentry (3-5). Unfortunately, catheter ablation of right atrial structures is an invasive procedure with the potential for complications and untoward outcomes. In addition, some forms of AV-node reentry are transient. More effective pharmacologic alternatives to radiofrequency ablation would be useful in the care of some patients with AV-node reentry.
The electropharmacology of the AV node and associated atrium has not been well studied. The classic model of reentry would suggest that any drug that prolongs the refractory period of the tissue in the reentrant circuit (sufficiently to envelop the excitable gap) without a change in conduction velocity will make reentry impossible (6). A reduction in the conduction velocity of a reentrant wavefront with no change in refractoriness should, in general, increase the probability of reentry. Thus an ideal antiarrhythmic drug to terminate reentry would increase the refractory period of the reentrant circuit and either have no effect on conduction velocity or increase conduction velocity. Possibly, for that reason, class I antiarrhythmic drugs that typically increase steady-state atrio-His (AH) conduction are not particularly effective against AV-node reentry. Flecainide is the exception. Flecainide appears to increase nodal refractoriness to a greater extent than it affects conduction velocity (7).
The major activity of class III agents is to increase action-potential duration and refractory periods in cardiac tissue. Therefore it would be expected to have greater effects on refractoriness than on conduction velocity. We have observed in the adult perfused rabbit heart that d-sotalol, a first-generation class III agent, increased the steady-state AH conduction time and the effective refractory period of the AV node without fluencing the maximal AH conduction time produced by premature stimulation (2). Second-generation class III drugs such as the methanesulfonamides (of which dofetilide is a prototype) have a very different chemical and pharmacologic profile from that of d-sotalol (8,9) and therefore are likely to have different effects on the properties of the AV conduction axis.
The purpose of this study was to compare the effects of three second-generation class III antiarrhythmic drugs (clofilium, risotilide, and dofetilide) on the dynamic antegrade conduction properties of the AV node and His-Purkinje system (HPS) of the perfused rabbit heart. More specifically, we attempted to modulate antegrade AV refractory periods without affecting steady-state AV conduction (through careful selection of drug concentrations) and observed changes in the magnitude of delayed AV conduction resulting from premature responses.
Adult New Zealand white rabbits, weighing 1.5-2.0 kg, were anesthetized with 60 mg/kg pentobarbital sodium injected intravenously into a marginal ear vein along with 1,000 U/kg of heparin sulfate. The chest was opened with a midsternal incision, and the heart was removed and placed in cold modified-Krebs buffer solution having the following composition in millimoles per liter: NaCl, 141; KCl, 2.6; MgSO4, 1.2; KH2PO4, 1.2; HEPES, 10; and dextrose, 11. The right atrium was opened with an incision from the margin of the atrial appendage through the superior vena cava. Another incision was made across the lower margin of the right atrial appendage posterior to the junction of the lower crista terminalis and the coronary sinus. The upper part of the crista terminalis was not preserved because the subsidiary pacemakers residing there interfered with pacing studies. For the same reason, the sinoatrial node and associated intercaval region and the eustachian ridge were removed. An incision was made from the right atrium through the AV ring into the middle region of the right ventricular free wall, and the incision was continued to the apex of the right ventricular cavity. The cut ends of the right circumflex coronary artery were ligated, and traction was placed on the sutures to expose the right atrial and ventricular septum.
The aorta was cannulated and retrogradely perfused with HEPES-buffered Krebs solution saturated with 100% oxygen and warmed to 37°C. The perfusion pressure was measured with a P23ID Statham pressure transducer (Statham Medical Instruments, Los Angeles, CA, U.S.A.), and the flow was provided by a roller pump adjusted to produce a mean perfusion pressure of ≈50 mm Hg. Bipolar stimulating and recording electrodes were constructed from 30-G (nominally 0.33 mm in diameter) and 34-G (nominally 0.18 mm in diameter) Teflon-coated silver wire (Cooner Wire, Chatsworth, CA, U.S.A.) tightly twisted together. A stimulus electrode was placed on the atrial septum superior to the foramen ovale. The atrial electrogram electrode was placed on the atrial septum superior to the coronary sinus. The His electrogram was recorded on the endocardial surface of the AV groove anterior to the coronary sinus. A ventricular electrogram was recorded on the upper ventricular septum. In most experiments, all conduction times were measured from the His electrogram. The atrial and ventricular electrograms were used to verify the His electrogram deflections. Stimulation pulses were produced by D/A output of a microcomputer-based I/O board and LabWindows software (AT-MIO-64F-5, National Instruments, Austin, TX, U.S.A.). In all experiments, the basic cycle interval (S1-S1) was either 400 or 250 ms, and the stimulus duration was 1 ms. The stimulus amplitude was adjusted to 2× threshold. Amplification and display of the electrograms were provided by an EVR-16 (Electronics for Medicine, Overland Park, KS, U.S.A.) with a bandwidth of 30-500 Hz. Data were digitized at 1 kHz and stored on a microcomputer hard disk for later analysis.
Stock solutions of risotilide (Wyeth-Ayerst, Princeton, NJ, U.S.A.) and clofilium (Research Biochemicals International, Natick, MA, U.S.A.) were prepared by dissolving the chemicals in distilled water. Dofetilide (Pfizer, Sandwich, UK) was dissolved in distilled water acidified to pH 3 with hydrochloric acid. All solutions were prepared fresh on the day of the experiment.
Only antegrade AV conduction properties were studied in these experiments. The premature pacing protocol was as follows: After 10 basic cycle stimuli (S1), a premature stimulus (S2) was delivered to the right atrium. The premature interval was initially decremented by 20 ms at the longest coupling interval. As the premature coupling intervals became shorter, the decrements were successively reduced to a minimum of 5 ms. During pacing protocols, functional parameters of the antegrade AV conduction were defined in the baseline state. Atrial, AV-node, and His-Purkinje refractory periods were measured at basic cycle lengths (BCLs) of 400 and 250 ms. Drugs were perfused at each concentration for 20 min, after which all measurements were repeated.
The atrial effective refractory period (A-ERP) was the longest S1-S2 interval in which the S2 failed to elicit an atrial response. The shortest A1-A2 interval that produced an A2 response was defined as the atrial functional refractory period (A-FRP). The AV-node effective refractory period (AVN-ERP) was the longest A1-A2 interval in which the A2 failed to elicit a response in the HPS. The shortest H1-H2 interval that produced an H2 response was defined as the AV-node functional refractory period (AVN-FRP). The longest H1-H2 interval in which the H2 failed to produce a ventricular response was referred to as the HPS effective refractory period (HPS-ERP). The HPS functional refractory period (HPS-FRP) was the shortest V1-V2 interval that produced a V2 response. In some cases, a proximal ERP was reached before a more distal ERP, making the more distal ERP measurement impossible. For example, if the AVN-ERP was longer than the HPS-ERP, the HPS-ERP could not be determined.
The AVN and HPS conduction curves were made by plotting the A2-H2 and H2-V2 intervals against their respective recovery intervals (H1-A2 and V1-H2). The minimal conduction times (AHmin, HVmin) were difined as the shortest conduction times measured at steady state. The maximal conduction times (AHmax, HVmax) were the longest conduction times measured during premature pacing. The difference between minimal and maximal conduction times of the AV node and the HPS were called ΔAH and ΔHV, respectively.
Differences between baseline and drug values were compared by using an analysis of variance. A Student's t test for paired observations with a Bonferroni adjustment also was used for individual comparisons (10). A Student's t test for independent groups was used to compare the baseline values between groups. A p value of <0.05 was considered significant.
For each drug, a range of concentrations was sought that progressively increased the refractory periods of each segment in the AV conduction axis with a minimal alteration of the basal conduction times (AHmin and HVmin). This was accomplished, for the most part, with all three drugs at a BCL of 400 ms, as shown in Tables 1, 2, and 3. AHmin was significantly increased at the highest drug concentration of dofetilide and risotilide.
The effect of risotilide on AV conduction
Risotilide was perfused in three concentrations (1 × 10-6, 3 × 10-6, and 6 × 10-6M) in the first four experiments. In the final three experiments, only the highest concentration of drug (6 × 10-6M) was used to limit the duration of the experiments and the exposure of the tissue to drug. The response of the AVN to risotilide at a concentration of 6 × 10-6M was the same in both types of experiments.
The effect of risotilide on the AVN recovery curve recorded at a 400-ms BCL is shown in Fig. 1. For clarity, an experiment with only the highest concentration of risotilide was used in this example. This experiment was typical of all experiments conducted at a BCL of 400 ms. This rightward shift in the terminal portion of the AVN recovery curve during exposure to drug reflects the increase in AVN-ERP (vide infra). AHmin was unchanged by drug. The absence of drug-induced changes in AHmin and the reduction in AHmax required a concentration-dependent reduction in ΔAH. In one case, at the highest drug concentration, a single delayed response, similar in duration to the baseline and washout AHmax values (120 ms), was conducted at a short premature interval (H1 - A2 ≈90 ms). As shown in Table 1, risotilide produced a concentration-dependent reduction in AHmax across the range of drug concentrations used in these experiments at the 400-ms BCL. AHmax was not different from baseline at any concentration of risotilide at the 250 ms BCL.
Risotilide (3 × 10-6M) increased the A-FRP, the AVN-FRP, and HPS-FRP (Table 1) as compared with baseline measurements. That the refractory periods of all three of these elements of the AV axis are increased by the same concentration of drug indicates that all three structures are similarly sensitive to the drug.
The relations between A-FRP and the ΔAH at each BCL are depicted in Fig. 2. There was a reciprocal relation between the increase in A-FRP and the reduction in ΔAH at the 400-ms BCL but not at the 250-ms BCL. Because the drug increased refractory periods in all segments of the AV axis, it might be argued that the changes in AVN refractory periods or even the HPS refractory periods may be related to changes in ΔAH. However, as shown in Table 1, at the highest concentration of drug, there was an increase in both the AVN-FRP and HPS-FRP without a corresponding reduction in the ΔAH at the 250-ms BCL. However, as with ΔAH, there was no change in A-FRP at any concentration of risotilide at the 250-ms BCL (Fig. 2).
Figure 3 represents the HPS conduction curve obtained from the same experiment as shown in Fig. 1. Risotilide had no effect on either HVmax or ΔHV at any concentration of drug or BCL (Table 1). However, HVmin was increased by 6 × 10-6M risotilide at the 400-ms BCL. The large rightward shift of the terminal portion of the HPS conduction curve observed with risotilide, again, represents the increase in effective refractory period.
The effect of clofilium on AV conduction
The clofilium concentrations selected for these experiments were lower than those found by Argentieri et al. (11) to increase action potential duration of the rabbit atrium and AVN. This difference in sensitivity to the drug was likely the result of the fact that our preparations were perfused whereas their preparations were superfused. In a total of eight experiments, the clofilium concentrations of 5 × 10-8 and 1 × 10-7M were tested in five and seven of those experiments, respectively. In two experiments, at the highest concentration of drug, the atrium would not respond 1:1 to stimuli at the 250-ms BCL.
The effect of clofilium (1 × 10-7M) on the AVN conduction curve is depicted in Fig. 4. These results were similar to those of risotilide at the BCL of 400 ms, in that there was a significant reduction in ΔAH and AHmax at the highest drug concentration (Fig. 4 and Table 2). However, unlike with risotilide, AHmin was not significantly increased by clofilium at the 250-ms BCL. Clofilium produced a concentration-dependent increase in A-FRP that correlated with a reduction in ΔAH (Fig. 5) at both BCLs. This indicates that the effect of clofilium on ΔAH and A-FRP was not BCL related, as was found with risotilide. Clofilium significantly increased A-FRP (Fig. 5) and A-ERP (Table 2) at the highest concentration (1 × 10-7M). Clofilium at the lowest concentration (1 × 10-8M) increased the AVN-FRP at both BCLs. However, no concentration of clofilium increased the AVN-ERP from the baseline measurements at the 400-ms BCL (Table 2). This may be because of the low n, because the A-ERP was greater than the AVN-ERP in four of eight experiments. Both HPS refractory periods were increased at the longer BCL by all three concentrations of clofilium. These data suggest that the AVN and HPS are more sensitive to the effects of clofilium than is the atrial myocardium.
The HPS conduction curve was unchanged by clofilium at any concentration or BCL studied except for a rightward shift of the terminal portion of the curve. These finding were virtually identical to the effect of risotilide (data not shown).
The effect of dofetilide on AV conduction
In the initial experiments with this drug, 1 × 10-8M was selected as the highest concentration of dofetilide. This was based on the previously published effects of dofetilide on canine Purkinje fibers (12). After three experiments, the maximum concentration of drug was increased to 1 × 10-7M. In one experiment, the highest concentration (1 × 10-7M) of drug caused first-degree AV block at both BCLs.
Dofetilide had uniform effects on the refractory periods of all three elements of the AV axis. The atrial, nodal, and HPS functional refractory periods were increased by a 5 × 10-8M concentration of dofetilide at the 250-ms BCL (Table 3). The A-FRP was increased at all concentrations of drug. The effect of this drug on ΔAH was very different between the two BCLs, as was the case with risotilide but not clofilium. A clear reduction in ΔAH is evident at the longer BCL, but no change was seen at the shorter BCL (Table 3). Figure 6 illustrates the AVN conduction curve at 400-ms BCL in one experiment. As with the other two drugs, dofetilide produced no change in the AHmin interval at the longer BCL. A concentration-dependent reduction in the AHmax and ΔAH was observed, which was similar to that seen with the other two drugs at the BCL of 400 ms. The highest concentration of dofetilide (1 × 10-7M) did increase AHmin at the 250-ms BCL (Table 3). As with risotilide at the 250-ms BCL, dofetilide produced no significant change in either ΔAH or AHmax.
Data from the same experiment as illustrated in Fig. 6 are shown in Fig. 7. Dofetilide produced changes in the HPS conduction curve similar to those of the other two drugs, that is, a rightward shift in the terminal portion of the conduction curve, indicating an increase in the ERP, but no statistically significant change in the ΔHV (Table 3). Intermittent HV conduction block was observed at drug concentrations of 5 × 10-8M in four of seven preparations at a BCL of 250 ms and in two of six preparations at a BCL of 400 ms. At the highest concentration (1 × 10-7M), 50% of the preparations displayed some degree of block in the HPS at the 400-ms BCL. The reduced n due to block in the HPS diminished the power of the t statistic and may have prevented the rejection of the null hypothesis. However, it is clear that dofetilide produced the highest rate of drug-induced HPS block of any of the three drugs; the HPS is more sensitive to dofetilide than either of the other two drugs.
The major findings of these experiments can be summarized as follows: (a) all three drugs produced a concentration-related increase in the refractory periods of all three elements of the AV axis; (b) the ΔAH parameter was significantly reduced by each drug at the 400-ms BCL (Figs. 2 and 5, Table 3), and only clofilium decreased ΔAH at the 250-ms BCL (Fig. 5), and (c) ΔHV was unaffected by any drug at any concentration or BCL.
The goal of these experiments was to use three class III antiarrhythmic drugs (with presumably different pharmacologic properties) to produce graded changes in the refractory periods of the atrium, AVN, and HPS and subsequently to assess the effect of these changes on AV conduction properties. There were commonalities in the effects of three drugs on AV conduction. For example, the methanesulfonamide-based drugs (risotilide and dofetilide) and the bretylium derivative, clofilium, had similar effects on the antegrade refractory periods of the AV axis. These findings are consistent with previous studies of class III drugs in this animal model (11,13) and in intact dogs (8,12). Regional differences in sensitivity to drugs were apparent. The refractory periods of the AVN and HPS were significantly increased at lower concentrations of risotilide and clofilium, whereas the refractory periods of the atrial myocardium at the lower concentrations of these drugs were not significantly increased. The lowest concentration of dofetilide (1 × 10-7M) produced uniform changes in the FRPs of the atrium, AVN, and HPS at the 400-ms BCL.
The relation between increased refractory periods and nodal conduction
All three drugs, at their highest concentration and at the longer BCL, produced a significant reduction in ΔAH (Figs. 2 and 5, Table 3). Concomitant with the reduction in ΔAH, there was a significant increase in A-FRP at the highest concentration of each drug at the longer BCL. Unlike risotilide and dofetilide, the same reciprocal relation between ΔAH and A-FRP was found at both BCLs in the clofilium experiments. This indicates that decreases in ΔAH and increases in A-FRP were BCL related with risotilide and dofetilide, but not clofilium.
The relation between A-FRP and ΔAH may be a critical determinant of nodal conduction properties for several reasons: The A-FRP is the output parameter of the atrium and the input parameter of the AVN. It may best represent changes in the refractory periods of the “atrionodal inputs” to the AVN. The role of the atrionodal inputs to the conduction properties of the AVN has been of great interest in recent years (3,14). From this point of view, it is reasonable to hypothesize that change in atrial rather than in AVN refractory periods might be critical to the AH conduction characteristics. More specifically, the posterior or “slow” atrionodal pathway has been shown to have a shorter refractory period than does the more anterior fast pathway, thus permitting the more premature impulses (15). Data to support this hypothesis are lacking, but the effect of these drugs may be to make the refractory periods of the two atrionodal inputs more homogeneous.
Adaniya and Hiraoka (13) studied the effect of E-4031 (a methanesulfonamide drug that is chemically similar to dofetilide and risotilide) on AVN reentrant tachycardia in a perfused rabbit heart model. They found that the drug increased the A-ERP and inhibited AVN reentrant tachycardia. In one of their experiments, a “jump” in the conduction curve suggesting dual-pathway physiology was observed. In this experiment, the three responses above the jump or discontinuity were eliminated by drug as was the AVN reentrant tachycardia. We have made similar observations with both dofetilide and clofilium (not shown). In our model, we believe that in the smooth, continuous nodal conduction curves presented here, the most delayed responses (at the shortest premature intervals) represent conduction through the posterior or slow pathways. The lack of a jump in the nodal conduction curves indicates only that three are not abrupt incongruities in the refractory periods between the “fast” and slow pathways (5,16,17). E-4031 blocked the delayed responses above the jump, in the experiment by Adaniya and Hiraoka, just as similar drugs did in our continuous nodal conduction curves. This finding supports our contention that drugs that abolish the long A2-H2 responses produced by short premature intervals in our model are also likely to abolish slow conduction in patients with dual-pathway physiology.
Our laboratory previously examined the effect of d-sotalol on the nodal conduction curves in neonatal and adult rabbit hearts (18). In these experiments, d-sotalol increased AHmin concomitant with the AVN refractoriness (atrial refractoriness was not measured in the d-sotalol experiments) in adults. ΔAH and AHmax were not altered by d-sotalol. The reason the three class III drugs used in this study reduced ΔAH and AHmax, whereas another class III drug, d-sotalol, did not is unclear. D-sotalol had effects on AVN and HPS refractory periods similar to those of the drugs used in this study. However, d-sotalol may not have sufficiently increased A-FRP, the refractory period change that, from these studies, appears to be crucial.
A limitation of this study is that we cannot directly measure changes produced by these drugs in the slow pathway itself. It is possible that only the access to the slow pathway has been eliminated by these drugs. They may increase the refractory period of atrial myocardium between the pacing site and the slow pathway, producing intraatrial block. Future experiments will focus on the effects of these drugs on the local refractory periods of the atrionodal inputs.
Basic cycle length-dependent changes on AV conduction
The effect of methanesulfonamide drugs may be diminished at faster heart rates because the action potential prolongation caused by these drugs at rapid heart rates is often small (19). This phenomenon has been termed “reverse use dependence.” Several aspects of these findings are consistent with the notion of reverse use dependence. Significant changes in the refractory periods were usually, but not always, observed at lower drug concentrations at the 400-ms BCL as compared with the 250-ms BCL. The percentage increase in refractory periods at any drug concentration was generally greater at the longer BCL. At the highest concentration of all three drugs, ΔAH was significantly decreased at the 400-ms BCL but unaffected at the 250-ms BCL.
We found that the normal basal heart rate of the animals studied in these experiments is between 160-200 beats/min (unpublished data), which is consistent with previously published observations (20). That is, the 400-ms BCL (150 beats/min) represents a slow to normal heart rate, and the 250-ms BCL (240 beats/min) represents a faster than normal heart rate in the rabbit.
It is logical to suspect that delayed AH conduction and echo beats reflect potential for AVN reentry (21). Therefore reduction or elimination of delayed nodal conduction should suppress AVN reentry. The results presented here suggest that all three drugs will increase atrial refractoriness and will reduce maximal AH conduction delay at slow heart rates. However, at higher heart rates (shorter BCLs), some class III drugs may not prolong atrial refractoriness and wil not affect delayed AH conduction. This would suggest that these drugs with rate-related effects on refractory periods might be less effective in younger patients or others with intrinsically high heart rates.
Acknowledgment: We appreciate Christine McKay for the statistical analysis and graphics and Barbara Roberts for assistance in the preparation of the manuscript. In addition, we thank Dr. Paul Gillette for his support of these experiments.
1. Josephson ME, Miller JM. Atrioventricular nodal reentry: evidence supporting and intranodal location. PACE Pacing Clin Electrophysiol
2. Case CL, Le FK, McKay CA, Gillette PC, Hewett KW. Developmental electrophysiologic effects of sotalol on antegrade and retrograde conduction of the rabbit atrioventricular node [Abstract]. Pediatr Res
3. McGuire MA, Janse MJ, Ross DL. “AV nodal” reentry: Part II: Nodal, AV junctional, or atrionodal reentry? J Cardiovasc Electrophysiol
4. Haïssaguerre M, Gaita F, Fischer B, et al. Elimination of atrioventricular nodal reentrant tachycardia using discrete slow potentials to guide application of radiofrequency energy. Circulation
5. Stein KM, Lerman BB. Evidence for functionally distinct dual atrial inputs to the human AV node. Am J Physiol
6. Janse MJ, Anderson RH, McGuire MA, Ho SY. “AV nodal” reentry: Part I: “AV nodal” reentry revisited. J Cardiovasc Electrophysiol
7. O'Hara G, Villemaire C, Talajic M, Nattel S. Effects of flecainide on the rate dependence of atrial refractoriness, atrial repolarization and atrioventricular node conduction in anesthetized dogs. J Am Coll Cardiol
8. Spinelli W, Parsons RW, Colatsky TJ. Effects of WAY-123,398, a new class III antiarrhythmic agent, on cardiac refractoriness and ventricular fibrillation threshold in anesthetized dogs: a comparison with UK-68798, E-4031, and dl
-sotalol. J Cardiovasc Pharmacol
9. Lynch JJ Jr, Baskin EP, Nutt EM, et al. Comparison of binding to rapidly activating delayed rectifier K+
, and effects on myocardial refractoriness for class III antiarrhythmic agents
. J Cardiovasc Pharmacol
10. Ludbrook J. Repeated measurements and multiple comparisons in cardiovascular research. Cardiovasc Res
11. Argentieri TM, Carroll MS, Sullivan ME. Cellular electrophysiological effects of the class III antiarrhythmic agents
sematilide and clofilium on rabbit atrial tissues. J Cardiovasc Pharmacol
12. Gwilt M, Arrowsmith JE, Blackburn KJ, et al. UK-68,798: a novel, potent and highly selective class III antiarrhythmic agent which blocks potassium channels in cardiac cells. J Pharmacol Exp Ther
13. Adaniya H, Hiraoka M. Effects of a novel class III antiarrhythmic agent, E-4031, on reentrant tachycardias in rabbit right atrium. J Cardiovasc Pharmacol
14. Csanadi Z, Klein GJ, Yee R, Thakur RK, Li H. Effect of dual atrioventricular node pathways on atrioventricular reentrant tachycardia. Circulation
15. Mendez C, Moe GK. Demonstration of a dual A-V nodal conduction system in the isolated rabbit heart. Circ Res
16. Rosen KM, Mehta A, Miller RA. Demonstration of dual atrioventricular nodal pathways in man. Am J Cardiol
17. Josephson ME. Supraventricular tachycardias. In: Josephson ME, ed. Clinical cardiac electrophysiology: techniques and interpretations
. 2nd ed. Philadelphia: Lea & Febiger, 1993:181-274.
18. Case CL, Le FK, McKay CA, Gillette PC, Hewett KW. Developmental effects of d
-sotalol on anterograde and retrograde atrioventricular conduction in the rabbit. J Cardiovasc Electrophysiol
19. Krafte DS, Volberg WA. Voltage dependence of cardiac delayed rectifier block by methanesulfonamide class III antiarrhythmic agents
. J Cardiovasc Pharmacol
20. Harkness JE, Wagner JE, eds. The biology and medicine of rabbits and rodents
. 2nd ed. Philadelphia: Lea & Febiger, 1983:10.
21. Rosenblueth A. Two processes for auriculo-ventricular and ventriculo-auricular propagation of impulses in the heart. Am J Physiol