- Email: [email protected]
Probability of Induction and Stabilization of Ventricular Fibrillation with Epinephrine
J Mol Cell Cardiol 30, 373–382 (1998)
Probability of Induction and Stabilization of Ventricular Fibrillation with Epinephrine Oscar H. Tovar, Paris P. Bransford and Janice L. Jones Cardiac Research Laboratory, Department of Veterans Affairs Medical Center and Department of Physiology and Biophysics, Georgetown University, Washington, DC 20422, USA (Received 17 June 1997, accepted in revised form 28 October 1997) O. H. T, P. P B J. L. J. Probability of Induction and Stabilization of Ventricular Fibrillation with Epinephrine. Journal of Molecular and Cellular Cardiology (1997) 30, 373–382. Clinical studies suggest that epinephrine facilitates ventricular fibrillation (VF) although mechanisms remain unclear. We tested the hypothesis that epinephrine increases the probability of inducing VF and stabilizes VF in association with shortening of fibrillation action potential duration. VF was induced in isolated, New Zealand White rabbit hearts (n=16) under control conditions and in the presence of 0.9 l/l epinephrine. Monophasic action potentials were recorded during sinus rhythm, pacing, and fibrillation. Epinephrine reduced fibrillation p80 by 80%, from 23±4 to 4.6±1 V (P<0.05); and reduced fibrillation p90 by 82%, from 29.3±5.4 to 5.4±1.9 V (P<0.05). Epinephrine also reduced the probability of spontaneous termination of VF during the first 5 s of VF from 29 to 8% (P<0.05). Epinephrine significantly decreased mean fibrillation cycle length from 104.5±2 to 75.7±2.3 ms (P<0.001). Mean action potential duration (60% repolarization) decreased from 76±3 to 40±3 ms (P<0.0003). Frequency analysis showed a mean dominant frequency during VF of 10.0±0.2 Hz under control conditions and 13.3±0.3 Hz with epinephrine (P<0.0001). These results suggest that epinephrine increases the probability of VF induction and decreases the probability of spontaneous defibrillation. Stabilization of fibrillation is associated with shortening of action potential duration and fibrillation cycle length, which may allow a greater number of fibrillation waves in the ventricle. 1998 Academic Press Limited
K W: Catecholamines; Epinephrine; Ventricular fibrillation; Ventricular fibrillation probability; Spontaneous defibrillation; Fibrillation cycle length; Action potential duration; Ventricular fibrillation threshold.
Introduction The role of epinephrine in early fibrillation is unclear. Clinical findings suggest that ventricular tachycardia or ventricular fibrillation in some patients is associated with and may be caused by a sustained increase in sympathetic activation (Meredith et al., 1991). Several studies suggest that spontaneous termination of fibrillation in humans may occur (Clayton et al., 1993; Josephson et al., 1979), although its frequency is unknown. However, a clinical study suggests that tachyarrhythmia
episodes with subsequent loss of aortic pressure (Peuhkurinen et al., 1994) produce sympathetic release that may accelerate the transformation of tachyarrhythmia into fibrillation or prevent the conversion to normal sinus rhythm. One human study, during the short fibrillation durations associated with ICD testing (Swartz et al., 1993), showed that fibrillation cycle length is controlled primarily by action potential duration. Another study in isolated cell aggregates showed that epinephrine shortens action potential duration, refractoriness and stimulation threshold at short
Please address all correspondence to: Oscar H. Tovar, Department of Veterans Affairs, Medical Center 151P, 50 Irving St, N.W., Washington, DC 20422, USA.
0022–2828/98/020373+10 $25.00/0
mc970600
1998 Academic Press Limited
374
O. H. Tovar et al.
cycle lengths characteristic of fibrillation (Tovar and Jones, 1997). Ventricular fibrillation threshold is also decreased in dogs treated with epinephrine (Papp and Szekeres, 1968b; Schwartz et al., 1977). These results suggest that if epinephrine decreases ventricular fibrillation threshold, then the probability of inducing ventricular fibrillation increases. Also, if action potential duration shortens, then more fibrillation wavefronts may be present in the heart thereby stabilizing fibrillation. Therefore the goal of this study was to test the hypothesis that epinephrine increases the probability of ventricular fibrillation and stabilizes fibrillation in association with shortening of the action potential duration at fibrillation cycle lengths. The present investigation is confined to the initial seconds of fibrillation before ischemia develops, thus the experiments were conducted during normal coronary flow.
Materials and Methods Isolated heart preparation Procedures for setting up the perfused isolated heart model are similar to those which have been described (Jones et al., 1990). Adult New Zealand White rabbits of either sex (n=16) weighing between 4.0 and 4.7 kg were anesthetized with pentobarbital (50 mg/kg). Heparin (1000 IU) was administered and allowed to circulate for 30 s. A median sternotomy was performed and the heart excised. The aorta was cannulated with a modified double lumen 11.5 French Mahurkar catheter (Quinton Instrument, Seattle, WA, USA). Langendorff type retrograde aortic perfusion was established within 30 s. The pulmonary veins were ligated close to the left atrium and the lungs removed. The right atrium was cannulated with an encapsulated type T thermocouple to monitor temperature which was maintained at 37.0±0.5°C. Aortic pressure was measured from the second lumen of the aortic cannula using a pressure transducer model P23XL (Spectramed, Inc., Oxnard, CA, USA). The heart was enclosed in an elastic sock with two platinum epicardial patch electrodes (12.7 mm diameter), Model FLO 21A (CPI, St Paul, MN, USA) to induce fibrillation and allow defibrillation.
Electrolyte solutions and drugs The perfusate was a modified Krebs–Henseleit bicarbonate buffer containing (in m/l) NaCl, 110;
NaHCO3, 32; NaH2PO4, 1.2; Na pyruvate, 2; KCl, 4; dextrose, 5.5; MgSO4, 1.2; CaCl2, 2.5; and 10 IU regular pork insulin (Nordisk A/S, Bagsvaerd, Denmark). All electrolytes were cell culture grade (Sigma Co., St Louis, MO, USA). Epinephrine HCl 0.9 l/l (Elkins–Sinn Inc., Cherry Hill, NJ, USA) was added to the Krebs–Henseleit buffer immediately prior to use. The concentration of 0.9 l/ l was comparable to the concentration of epinephrine used in canine Purkinje fibers (1 l/l epinephrine) (Danilo et al., 1978) and also in Tyrode superfused human papillary muscle samples (10 l/l epinephrine) (Jakob et al., 1988). The epinephrine concentration in our experiments is also comparable to the concentration used in the ACLS protocol. In two cardiac arrest patients in which epinephrine was administered (1 mg in 1 ml), the plasma concentration was 0.058 l/l in one patient and 1.6 l in the other patient (Little et al., 1985). The perfusate was monitored throughout the experiment and maintained within optimal values of pH 7.40, Po2 600 mmHg, and P2 40 mmHg.
Recording techniques A bipolar electrogram (EG) was recorded from the electrode patches. Endocardial monophasic action potentials (MAPs) were recorded using Ag–AgCl bipolar contact monophasic action potential recording catheters, model 225 (EP Technologies, Sunnyvale, CA, USA), placed through the pulmonary artery to the right ventricular endocardial apex. Epicardial MAPs were recorded using a similar electrode placed on the epicardial surface of the right ventricular wall. The MAPs were amplified through a DC-coupled amplifier, model 300 (EP Technologies, Sunnyvale, CA, USA). Data were recorded both on a Gould strip chart and on a computer for later frequency analysis.
Experimental protocol Three groups of rabbits (A, B and C) were used in this study to minimize the cumulative effects of multiple interventions on the same heart. The protocols for groups A and B were performed in each rabbit heart under control conditions (perfusate without epinephrine) and then again with Krebs solution containing 0.9 l/l epinephrine. Group C served as a control for time-dependent changes.
Fibrillation Induction and Stabilization by Epinephrine
Group A (n=6) The probability of inducing ventricular fibrillation and the probability of spontaneous defibrillation were determined. Ventricular fibrillation threshold was determined in each heart using an up and down protocol. Ventricular fibrillation was induced by 1-s 60 Hz pulses of 2, 3, 4, 5, 7.5, 10, 12.5, 15, 25, 30 and 35 V (peak to peak) delivered through the epicardial patches. Fibrillation was defined by electrograms of irregular morphology, which lasted longer than 1 s after the end of the stimulus. An average of 16 shocks per heart was used to determine the ventricular fibrillation threshold under control conditions (total n=64) and 11 shocks in the presence of epinephrine (total n= 46). A computerized error function fit routine (Jones et al., 1990) was used to construct sigmoidal probability of fibrillation success v shock voltage curves for each heart. The probability curves determined the voltages for 10, 20, 50, 80 and 90% ventricular fibrillation success. The ventricular fibrillation probabilities are reported as the mean±... from the individual curves. A two-way repeated measures analysis of variance on two factors was performed on the results. Differences between the groups were isolated using the Student–Newman–Keuls test. Ventricular fibrillation episodes terminating less than 5 s after the onset of fibrillation were defined as spontaneous defibrillation as shown by the MAP in Figure 1(a). To minimize stress on the heart, fibrillation episodes that did not spontaneously terminate by 5 s were defibrillated by a 10 ms asymmetric biphasic waveform of known effective voltage. In each heart, cycle length was determined manually as shown in Figure 1(b) for at least three episodes of ventricular fibrillation under control conditions and in the presence of epinephrine. Mean ventricular fibrillation cycle length in each heart was determined at the beginning of three successive 1-s intervals beginning 1 s after the termination of the 60 Hz fibrillation current. Action potential duration measurements at 60% repolarization (APD60) were used to determine whether epinephrine shortened action potential duration during fibrillation [Fig. 1(c)]. APD60 was defined as the time between the upstroke of the action potential and the repolarization potential corresponding to 60% of the upstroke amplitude. Mean action potential duration for each episode was defined as the average of the action potential durations measured during the three successive one second intervals. To directly show that epinephrine shortens action
375
potential duration at a specific cycle length, rather than solely by cycle length shortening, we determined action potential duration during pacing in three of the six hearts before and after epinephrine perfusion (this protocol was performed before the induction of fibrillation).
Group B (n=8) In this group, ventricular fibrillation was allowed to continue for 10 s before defibrillation was attempted, this longer time allowed us to perform fast Fourier transform analysis of the last 8 s of fibrillation. The effects of epinephrine on the mean dominant frequency of fibrillation were determined on well-stabilized ventricular fibrillation under control conditions and in the presence of epinephrine. This analysis was performed to verify the consistency of the results on cycle length from group A, because we were able to measure only a limited number of cycles manually. The fast Fourier transform analysis was carried out on epicardial monophasic action potentials using a custom computer program (Labview, National Instruments, Houston, TX, USA).
Group C (n=2) Time dependent changes on the action potential and cycle length were determined. To confirm that the changes in cycle length and action potential duration are due to epinephrine and not to timerelated effects, the identical protocol, which lasted up to 120 min was performed in two additional control rabbits hearts without epinephrine perfusion.
Data analysis Statistical analysis was performed using Sigmastat (Jandel Scientific, San Rafael, CA, USA). Results are expressed as mean±standard error of the mean (...). Differences were determined with paired Student’s t-test, v2 analysis, two way ANOVA and the Student–Newman–Keuls test and were considered significant at P<0.05.
376
O. H. Tovar et al.
Figure 1 (a) Monophasic action potentials recorded during fibrillation and spontaneous defibrillation. The regular pattern at the left of the panel is caused by the 60 Hz fibrillation stimulus. (b) Measurement of mean fibrillation cycle length from a sequence of six monophasic action potentials. (c) Measurement of fibrillation APD60. The mean action potential duration is calculated by measuring six consecutive monophasic action potentials at 60% repolarization, then averaging the measurements.
Results Group A (ventricular fibrillation probability) Figure 2 shows the probability of inducing VF (n=4) as a function of stimulus intensity, in the absence and presence of epinephrine. For the range of stimulus intensities tested, the overall probability of inducing ventricular fibrillation was significantly higher (P<0.03 by repeated measures ANOVA) when the hearts were exposed to epinephrine. A multiple comparison analysis showed that epinephrine significantly reduced fibrillation p80 by 80%, from 23±4 to 4.6±1 V (P<0.05) and reduced fibrillation p90 by 82%, from 29.3±5.4 to 5.4±1.9 V (P<0.05). The normalized curve width, defined by the expression (V80–V20)/V50, decreased from 1.45 under control conditions to 0.65 in the presence of epinephrine. Spontaneous defibrillation To determine the probability of spontaneous defibrillation between 3 and 5 s fibrillation, we ana-
lyzed 80 episodes of fibrillation in the absence of epinephrine and 75 episodes of fibrillation in the presence of epinephrine. Epinephrine significantly decreased (by v2 analysis) the probability of spontaneous defibrillation from 29% (23 out of 80 episodes) of the fibrillation episodes under control conditions to 8% (6 out of 75) of episodes in the presence of epinephrine (P<0.05).
Normal sinus rhythm Figure 3 shows endocardial monophasic action potentials (MAP) and electrograms (EG) recorded from one experiment during normal sinus rhythm and fibrillation. Figures 3(a) and (b) show MAPs recorded during normal sinus rhythm under control conditions and after epinephrine perfusion. Consistent with epinephrine’s known effects to increase heart rate, the cycle length shortened 20% from 300 to 240 ms. In this episode, the APD60 shortened 38%, from 130 to 80 ms. The APD100 shortened 24%, from 170 to 130 ms.
Fibrillation Induction and Stabilization by Epinephrine 100
377
Fibrillation cycle length
Fibrillation probability (%)
90 80 70 60 50 40 30 20 10 0
5
10
15 20 25 Voltage (V)
30
35
40
Figure 2 The probability of fibrillation induction as a function of stimulation voltage (V) in the absence (open circles) and presence of epinephrine (closed circles).
The mean cycle length during normal sinus rhythm in the six rabbits shortened from 324±12 ms (21 episodes under control conditions) to 256±6 ms (27 episodes in the presence of epinephrine) (P<0.05). Mean APD100 shortened from 178±5 ms in the absence of epinephrine to 140±3 ms (P<0.05).
Figures 3(c) and (d) show action potentials recorded 3 s after the onset of fibrillation during a sample control episode and after epinephrine perfusion, respectively. In this episode, epinephrine decreased fibrillation cycle length from 108 to 77 ms (29% decrease). APD60 during fibrillation decreased from 68 to 43 ms (37% decrease). The effect of epinephrine on cycle length and action potential duration during fibrillation is shown in Figure 4. Mean cycle length and mean APD60 during fibrillation were determined under control conditions and in the presence of epinephrine for the 3rd, 4th, and 5th seconds of fibrillation. Figure 4(a) shows mean fibrillation cycle length as a function of fibrillation duration in the absence and in the presence of epinephrine. Under neither condition did cycle length change significantly as a function of time. However, epinephrine significantly shortened mean fibrillation cycle length. For example, at 5 s fibrillation, epinephrine decreased mean fibrillation cycle length by 29% from 104.5±2 ms (control conditions) to 75.7±2.3 ms (P<0.001). Mean APD60 decreased 47% from 76±3 ms under control conditions to
Figure 3 MAPs from a typical experiment. MAPs with corresponding surface electrograms recorded during sinus rhythm in the absence (a) and presence (b) of epinephrine. MAPs with corresponding surface electrograms recorded during ventricular fibrillation in the absence (c) and presence (d) of epinephrine.
378
O. H. Tovar et al. 120
(a)
100
100
80
80 APD60 (ms)
Cycle length (ms)
120
60
60
40
40
20
20
0
2
3
4 Time (s)
5
6
(b)
0
2
3
4 Time (s)
5
6
Figure 4 (a) Fibrillation cycle length as a function of fibrillation duration in the presence (open circles) and absence (closed circles) of epinephrine. (b) APD60 as a function of fibrillation duration in the presence (open circles) and absence (closed circles) of epinephrine.
40±3 ms (P<0.0003) in the presence of epinephrine as shown in Figure 4(b).
Action potential duration during pacing Figure 5 shows original MAP recordings from a typical pacing experiment at cycle lengths of 240 (a), 200 (b), and 150 ms (c), both before and after epinephrine perfusion. In the absence of epinephrine, as the cycle length shortened from 240 to 150 ms APD60 shortened from 115 to 80 ms in a rate-dependent manner. After epinephrine perfusion, the action potential duration also shortened in a rate-dependent manner from 60 to 50 ms as cycle length decreased. Figure 6 shows mean APD60 in the absence (closed symbols) and presence (open symbols) of epinephrine, during paced cycle lengths (squares) of 150, 175, 200 and 240/250 ms and at fibrillation cycle lengths (circles). The difference between the paced APD60 under control conditions and in the presence of epinephrine was statistically significant (P<0.001, by two way ANOVA). Logarithmic regression curves, obtained from the combined paced and fibrillation data, used to fit both groups were APD=−193+57 In (CL) under control conditions and APD=−74+26 In (CL) in the presence of epinephrine. Epinephrine shifted the regression curve downward so that, at a given cycle length, action potential duration was shorter in the epinephrine group.
Group B (frequency analysis of fibrillation) The measurements of fibrillation cycle length in Group A were consistent with the frequency content determined by fast Fourier transforms in the group of rabbits (n=8) in which fibrillation was allowed to persist for 10 s. Figure 7 shows the power spectra obtained from 8 s of fibrillation under control conditions [Figure 7(a)] and in the presence of epinephrine [Figure 7(b)]. The dominant frequency in this control episode was 9.9 Hz (cycle length≈101 ms); the dominant frequency in this epinephrine episode was 13.8 Hz (cycle length≈72 ms). The mean dominant frequency in the control group (n=10 episodes) was 10.0±0.2 Hz, which corresponds to a mean cycle length of 100 ms. The mean dominant frequency in the epinephrine group (n=10 episodes) was 13.3±0.3 Hz, which corresponds to a mean cycle length of 75 ms. This difference was statistically significant (P<0.0001).
Group C (control hearts) In the two control hearts for which the fibrillation protocol was carried out in the absence of epinephrine, Figure 8 shows that, in the absence of epinephrine, fibrillation cycle length does not decrease within the first 2 h of fibrillation. Rather, the fibrillation cycle length tends to increase with time. However, the change was not statistically
Fibrillation Induction and Stabilization by Epinephrine
379
Figure 5 MAPs recorded during pacing at basic cycle lengths of 240 (a), 200 (b), and 150 ms in the absence (left panel) and presence (right panel) of epinephrine.
150
Discussion Ventricular fibrillation probability
APD60 (ms)
100
50
0 50
100
150 200 Cycle length (ms)
250
300
Figure 6 Effect of epinephrine and cycle length on APD during pacing (squares) and fibrillation (circles) under control conditions (closed symbols) and in the presence of epinephrine (open symbols).
significant. This confirms that the shortening produced by epinephrine is not a time-dependent effect.
In contrast to ventricular pacing, which is described by a sharply defined “threshold”, induction of fibrillation by a stimulus is a probabilistic function similar to that describing defibrillation. Therefore, as shown in Figure 2, fibrillation threshold is a poor descriptor of fibrillation induction. Rather the probability of inducing fibrillation increases gradually with stimulus intensity. Under control conditions, a 5 V stimulus induced fibrillation in 10% of the episodes, while a 29 V stimulus induced fibrillation in 90% of episodes. Epinephrine not only shifted the fibrillation probability curve to the left but also markedly decreased the normalized width (increased the normalized slope) of the curve. Due to these two factors, the probability of inducing ventricular fibrillation with a 5 V stimulus increased from 10% under control conditions to 90% in the
380
O. H. Tovar et al.
Cycle length (ms)
120 100 80 60 40 20 0
20
40
60 80 Time (min)
100
Figure 8 Mean fibrillation cycle length as a function of time (min) during 10 s fibrillation episodes in control rabbit hearts. Each point represents the means of fibrillation episodes induced at a specific time in the experiment.
Figure 7 Fast Fourier transforms of MAPs during fibrillation in the absence (a) and presence (b) of epinephrine.
presence of epinephrine. These results agree with reports showing that sympathetic tone increases vulnerability to ventricular fibrillation (Han et al. 1964; Posel et al., 1989) and that epinephrine reduces ventricular fibrillation threshold (Papp and Szekeres, 1968b; Schwartz et al., 1977).
Epinephrine in early fibrillation The role of epinephrine in early fibrillation is unclear. However, clinical studies suggest that sympathetic activation can precipitate ventricular tachycardia or ventricular fibrillation (Meredith et al., 1991). In the case of tachyarrhythmia episodes with loss of aortic pressure (Peuhkurinen et al.,
1994), it may facilitate the degeneration of the tachyarrythmia into fibrillation and hinder spontaneous defibrillation. There is strong experimental and clinical evidence that epinephrine sensitizes the ischemic myocardium to ventricular fibrillation (Bertel et al., 1982; Little et al., 1985; Opie and Lubbe, 1979). In rats, the ventricular fibrillation threshold starts to decrease 6 min after regional ischemia induction and reaches the lowest point 14 min after ischemia (Opie and Lubbe, 1979). A study by Bertel showed that those patients with ventricular fibrillation as a complication of acute myocardial infarction showed the highest plasma levels of epinephrine. They suggested that this phenomenon represented a direct effect of catecholamines on myocardial irritability during the early phase of myocardial infarction (Bertel et al., 1982). This increase in myocardial irritability might be mediated at least partially by an increase in intracellular cAMP by epinephrine (Opie and Lubbe, 1979). Murnaghan (1975) found that, in isolated rabbit hearts, the lowering of ventricular fibrillation threshold was associated with shortening of refractoriness and was blocked by propranolol and pindolol (b-blockers) but it was not influenced by phentolamine (a-blocker). The increased vulnerability at the onset of fibrillation may be due to a decreased excitation threshold of cardiac cells at very short cycle lengths due to shortening of refractoriness (Tovar and Jones, 1997). As we showed in that study, the effects of epinephrine on action potential shortening and excitation threshold at short cycle lengths in isolated myocytes were reversed by propranolol. Again, this suggests that the undesirable effects of epinephrine are due to its b-effects.
Fibrillation Induction and Stabilization by Epinephrine
Spontaneous defibrillation The ability of a healthy heart to sustain fibrillation depends on its size. Spontaneous defibrillation occurs rapidly in healthy small hearts such as those from rats or guinea pigs (West and Laud, 1962; Wigger, 1929) because the small ventricular mass causes fibrillation wavefronts to reach refractory tissue so that fibrillation cannot be sustained. Until recently, spontaneous termination of fibrillation was considered a rare event in humans because of the size of the human heart. However, clinical reports suggest that spontaneous termination of fibrillation in humans may occur, (Clayton et al., 1993; Josephson et al., 1979) although its frequency is unknown (Clayton et al., 1993). This suggests that heart size is only one of several factors which may influence the ability for spontaneous defibrillation to occur (Manoach and Basat, 1990) and that other factors could be optimized in larger hearts to increase the probability of spontaneous defibrillation. Because spontaneous defibrillation occurs frequently in the rabbit heart, it is ideal for a study of drug-induced stabilization of fibrillation (Manoach and Basat, 1990; West and Laud, 1962). In the current study, epinephrine reduced the occurrence of spontaneous termination of fibrillation from 29% of fibrillation episodes under control conditions to 8%.
Effects of epinephrine on action potential duration Action potential duration shortens as a function of cycle length during pacing at sinus rates (Attwell et al., 1981). Restitution curves in pigs also show that, at a cycle length of 350 ms, epinephrine shortened action potential duration for a single extrasystole (Taggart et al., 1990). Cycle lengths shorter than 300 ms were not examined. Furthermore, it has been shown in dogs that epinephrine shortens refractory period at paced cycle length of ≈300 ms (Papp and Szekeres, 1968a). The results of our study also show cycle length dependent action potential shortening. Figure 6 showed that action potential duration decreased logarithmically with cycle length. This cycle length dependence occurred both in the absence and presence of epinephrine; however, epinephrine shifted the regression curve downward and changed its slope so that the greatest effects were observed at fibrillation cycle lengths. Epinephrine shortened action potential duration at the control fibrillation cycle length; this allowed cycle length to further
381
shorten so that a 29% decrease in the mean fibrillation cycle length and a 47% decrease in the mean fibrillation APD60 were produced. Similar results were obtained using myocardial cell aggregates where epinephrine reduced action potential duration at cycle lengths between 160 and 400 ms. Consistent with the downward and leftward shifting of the fibrillation APD60 v cycle length curve (Fig. 6), epinephrine also allowed pacing without alternans at a cycle length of 160 ms and permitted pacing at 140 ms, which could not be achieved in the absence of epinephrine even with high intensity stimuli of four times threshold (Tovar and Jones, 1997). The number of wavelets that can coexist in cardiac tissue has been inversely related to the duration of the refractory period and to the conduction velocity (Moe et al., 1964). Additionally, a study of monophasic action potentials recorded during clinical testing of internal defibrillators (Swartz et al., 1993) showed that during human ventricular fibrillation there is no period of diastole following the fibrillation action potentials. Consistently, during fibrillation episodes in this study, cells were restimulated by new wavefronts immediately following repolarization so that they never achieved a stable resting potential, neither under control conditions nor in the presence of epinephrine. These results suggest that local cycle length during fibrillation is controlled primarily by action potential duration, which, at short fibrillation durations, determines cellular refractory period. Because epinephrine does not appear to significantly alter conduction velocity (Smeets et al., 1986; Vodanovic et al., 1993), shortening of the refractory period by epinephrine allows the presence of more fibrillation wavefronts. Consequently, our present findings suggest that one mechanism through which epinephrine stabilizes fibrillation is by shortening fibrillation action potential duration and therefore refractoriness. This allows a greater number of fibrillation wavefronts to be present in the ventricle and decreases the probability that all wavefronts will become blocked and that fibrillation will be extinguished.
Clinical implications Although spontaneous defibrillation occurs in humans, it is rare. Epinephrine release in a deteriorated myocardium inducing tachycardia or during the initial tachyarrhythmic phase preceding fibrillation may decrease the probability of spontaneous conversion in humans and make it more
382
O. H. Tovar et al.
likely that the initial tachycardia will degenerate to fibrillation. In conclusion, epinephrine facilitates initiation of fibrillation as shown by the increased probability of inducing fibrillation with low intensity stimuli. It also decreases the probability of spontaneous defibrillation. The stabilization of fibrillation is associated with decreased fibrillation cycle length and action potential duration shortening during the early stages of fibrillation.
Acknowledgments We thank Ms Nettie Knight for her expert technical assistance. This work was supported by The Department of Veterans Affairs Medical Center and by USPHS HL 24606 and by HL 49089.
References A D, C I, E DA, 1981. The effects of heart rate on the action potential of guinea-pig and human ventricular muscle. J Physiol (London) 313: 439–461. B O, B¨ FR, B G, R R, B F, 1982. Plasma adrenaline and noradrenaline in patients with acute myocardial infarction. Relationship to ventricular arrhythmias of varying severity. Chest 82: 64– 68. C RH, M A, H PD, C RW, 1993. Self-terminating ventricular tachyarrhythmias—a diagnostic dilemma? The Lancet 341: 93–95. D P, V Y, V M, R MR, 1978. Epinephrine-induced automaticity of canine cardiac purkinje fibers and its relationship to the adenylate cyclase-adenosine 3′,5′-monophosphate system. J Pharmacol Exp Ther 205: 175–182. H J, J P, M GK, 1964. Adrenergic effects on ventricular vulnerability. Circ Res 16: 516–524. J H, N H, R J, 1988. Adrenoceptor-mediated changes of action potential and force of contraction in human ventricular heart muscle. Br J Pharmacol 94: 584–590. J JL, S JF, J RE, 1990. Increasing fibrillation duration enhances relative assymmetrical biphasic versus monophasic defibrillator waveform efficacy. Circ Res 67: 376–384. J ME, S SR, G AM, H LN, 1979. Mechanism of ventricular fibrillation in man. Observations based on electrode catheter recordings. Am J Cardiol 44: 623–631. L RA, F KN, R PE, S HB, Y DW, L GS, K S, B JM, 1985. Plasma
catecholamines in patients with acute myocardial infarction and in cardiac arrest. Quart J Med 54: 133– 140. M M, B MB, 1990. The structural–functional basis of spontaneous ventricular defibrillation. Int J of Cardiol 26: 129–136. M IT, B A, J GL, E MD, 1991. Evidence of a selective increase in cardiac sympathetic activity in patients with sustained ventricular arrhythmias. N Engl J Med 325: 618–624. M GK, R WC, A JA, 1964. A computer model of atrial fibrillation. Am Heart J 67: 200– 220. M MF, 1975. The effect of sympathomimetic amines on the ventricular fibrillation threshold in the rabbit isolated heart. Br J Pharmacol 53: 3–9. O LH, L WF, 1979. Catecholamine-mediated arrhythmias in acute myocardial infarction. Experimental evidence and role of beta-adrenoceptor blockade. S Afr Med J 56: 871–880. P JG, S L, 1968a. Analysis of the mechanism of adrenergic actions on ventricular vulnerability. Euro J of Pharmacol 3: 15–26. P JG, S L, 1968b. The arrhythmogenic action of sympathomimetic amines. Eur J of Pharmacol 3: 4–16. P KJ, H HV, L M, T JT. 1994. Changes in myocardial metabolism and transcardiac electrolytes during simulated ventricular tachycardia: effects of beta-adrenergic blockade. Am Heart J 128: 96–105. P D, N T, K P, 1989. Exercise training after experimental myocardial infarction increases the ventricular fibrillation threshold before and after the onset of reinfarction in the isolated rat heart. Circulation 80: 138–145. S PJ, V RL, L B, 1977. Effect of stellectomy and vagotomy on ventricular refractoriness in dogs. Circ Res 40: 536–540. S JL, A MA, L WJ, B FI, H J, 1986. The wavelength of the cardiac impulse and reentrant arrhythmias in isolated rabbit atrium. The role of heart rate, autonomic transmitters, temperature, and potassium. Circ Res 73: 96–108. S JF, J JL, F R, 1993. Characterization of ventricular fibrillation based on monophasic action potential morphology in the human heart. Circulation 87: 1907–1914. T P, S P, L M, D J, 1990. Interplay between adrenaline and interbeat interval on ventricular repolarization in intact heart in vivo. Cardiovasc Res 24: 884–895. T OH, J JL, 1997. Epinephrine facilitates cardiac fibrillation by shortening action potential refractoriness. J Mol Cell Cardiol 29: 1447–1455. V S, T LA, H RG, K JP, B ZJ, 1993. Transient negative dromotropic effects of catecholamines on canine purkinje fibers exposed to halothane and isoflurane. Anesth Analg 76: 592–597. W TC, L JF, 1962. Minimal mass required for induction of a sustained arrhythmia in isolated atrial segment. Am J Physiol 202: 232–236. W CJ, 1929. Studies of ventricular fibrillation caused by electrical shock. Am Heart J 5: 351–365.
Probability of Induction and Stabilization of Ventricular Fibrillation with Epinephrine Oscar H. Tovar, Paris P. Bransford and Janice L. Jones Cardiac Research Laboratory, Department of Veterans Affairs Medical Center and Department of Physiology and Biophysics, Georgetown University, Washington, DC 20422, USA (Received 17 June 1997, accepted in revised form 28 October 1997) O. H. T, P. P B J. L. J. Probability of Induction and Stabilization of Ventricular Fibrillation with Epinephrine. Journal of Molecular and Cellular Cardiology (1997) 30, 373–382. Clinical studies suggest that epinephrine facilitates ventricular fibrillation (VF) although mechanisms remain unclear. We tested the hypothesis that epinephrine increases the probability of inducing VF and stabilizes VF in association with shortening of fibrillation action potential duration. VF was induced in isolated, New Zealand White rabbit hearts (n=16) under control conditions and in the presence of 0.9 l/l epinephrine. Monophasic action potentials were recorded during sinus rhythm, pacing, and fibrillation. Epinephrine reduced fibrillation p80 by 80%, from 23±4 to 4.6±1 V (P<0.05); and reduced fibrillation p90 by 82%, from 29.3±5.4 to 5.4±1.9 V (P<0.05). Epinephrine also reduced the probability of spontaneous termination of VF during the first 5 s of VF from 29 to 8% (P<0.05). Epinephrine significantly decreased mean fibrillation cycle length from 104.5±2 to 75.7±2.3 ms (P<0.001). Mean action potential duration (60% repolarization) decreased from 76±3 to 40±3 ms (P<0.0003). Frequency analysis showed a mean dominant frequency during VF of 10.0±0.2 Hz under control conditions and 13.3±0.3 Hz with epinephrine (P<0.0001). These results suggest that epinephrine increases the probability of VF induction and decreases the probability of spontaneous defibrillation. Stabilization of fibrillation is associated with shortening of action potential duration and fibrillation cycle length, which may allow a greater number of fibrillation waves in the ventricle. 1998 Academic Press Limited
K W: Catecholamines; Epinephrine; Ventricular fibrillation; Ventricular fibrillation probability; Spontaneous defibrillation; Fibrillation cycle length; Action potential duration; Ventricular fibrillation threshold.
Introduction The role of epinephrine in early fibrillation is unclear. Clinical findings suggest that ventricular tachycardia or ventricular fibrillation in some patients is associated with and may be caused by a sustained increase in sympathetic activation (Meredith et al., 1991). Several studies suggest that spontaneous termination of fibrillation in humans may occur (Clayton et al., 1993; Josephson et al., 1979), although its frequency is unknown. However, a clinical study suggests that tachyarrhythmia
episodes with subsequent loss of aortic pressure (Peuhkurinen et al., 1994) produce sympathetic release that may accelerate the transformation of tachyarrhythmia into fibrillation or prevent the conversion to normal sinus rhythm. One human study, during the short fibrillation durations associated with ICD testing (Swartz et al., 1993), showed that fibrillation cycle length is controlled primarily by action potential duration. Another study in isolated cell aggregates showed that epinephrine shortens action potential duration, refractoriness and stimulation threshold at short
Please address all correspondence to: Oscar H. Tovar, Department of Veterans Affairs, Medical Center 151P, 50 Irving St, N.W., Washington, DC 20422, USA.
0022–2828/98/020373+10 $25.00/0
mc970600
1998 Academic Press Limited
374
O. H. Tovar et al.
cycle lengths characteristic of fibrillation (Tovar and Jones, 1997). Ventricular fibrillation threshold is also decreased in dogs treated with epinephrine (Papp and Szekeres, 1968b; Schwartz et al., 1977). These results suggest that if epinephrine decreases ventricular fibrillation threshold, then the probability of inducing ventricular fibrillation increases. Also, if action potential duration shortens, then more fibrillation wavefronts may be present in the heart thereby stabilizing fibrillation. Therefore the goal of this study was to test the hypothesis that epinephrine increases the probability of ventricular fibrillation and stabilizes fibrillation in association with shortening of the action potential duration at fibrillation cycle lengths. The present investigation is confined to the initial seconds of fibrillation before ischemia develops, thus the experiments were conducted during normal coronary flow.
Materials and Methods Isolated heart preparation Procedures for setting up the perfused isolated heart model are similar to those which have been described (Jones et al., 1990). Adult New Zealand White rabbits of either sex (n=16) weighing between 4.0 and 4.7 kg were anesthetized with pentobarbital (50 mg/kg). Heparin (1000 IU) was administered and allowed to circulate for 30 s. A median sternotomy was performed and the heart excised. The aorta was cannulated with a modified double lumen 11.5 French Mahurkar catheter (Quinton Instrument, Seattle, WA, USA). Langendorff type retrograde aortic perfusion was established within 30 s. The pulmonary veins were ligated close to the left atrium and the lungs removed. The right atrium was cannulated with an encapsulated type T thermocouple to monitor temperature which was maintained at 37.0±0.5°C. Aortic pressure was measured from the second lumen of the aortic cannula using a pressure transducer model P23XL (Spectramed, Inc., Oxnard, CA, USA). The heart was enclosed in an elastic sock with two platinum epicardial patch electrodes (12.7 mm diameter), Model FLO 21A (CPI, St Paul, MN, USA) to induce fibrillation and allow defibrillation.
Electrolyte solutions and drugs The perfusate was a modified Krebs–Henseleit bicarbonate buffer containing (in m/l) NaCl, 110;
NaHCO3, 32; NaH2PO4, 1.2; Na pyruvate, 2; KCl, 4; dextrose, 5.5; MgSO4, 1.2; CaCl2, 2.5; and 10 IU regular pork insulin (Nordisk A/S, Bagsvaerd, Denmark). All electrolytes were cell culture grade (Sigma Co., St Louis, MO, USA). Epinephrine HCl 0.9 l/l (Elkins–Sinn Inc., Cherry Hill, NJ, USA) was added to the Krebs–Henseleit buffer immediately prior to use. The concentration of 0.9 l/ l was comparable to the concentration of epinephrine used in canine Purkinje fibers (1 l/l epinephrine) (Danilo et al., 1978) and also in Tyrode superfused human papillary muscle samples (10 l/l epinephrine) (Jakob et al., 1988). The epinephrine concentration in our experiments is also comparable to the concentration used in the ACLS protocol. In two cardiac arrest patients in which epinephrine was administered (1 mg in 1 ml), the plasma concentration was 0.058 l/l in one patient and 1.6 l in the other patient (Little et al., 1985). The perfusate was monitored throughout the experiment and maintained within optimal values of pH 7.40, Po2 600 mmHg, and P2 40 mmHg.
Recording techniques A bipolar electrogram (EG) was recorded from the electrode patches. Endocardial monophasic action potentials (MAPs) were recorded using Ag–AgCl bipolar contact monophasic action potential recording catheters, model 225 (EP Technologies, Sunnyvale, CA, USA), placed through the pulmonary artery to the right ventricular endocardial apex. Epicardial MAPs were recorded using a similar electrode placed on the epicardial surface of the right ventricular wall. The MAPs were amplified through a DC-coupled amplifier, model 300 (EP Technologies, Sunnyvale, CA, USA). Data were recorded both on a Gould strip chart and on a computer for later frequency analysis.
Experimental protocol Three groups of rabbits (A, B and C) were used in this study to minimize the cumulative effects of multiple interventions on the same heart. The protocols for groups A and B were performed in each rabbit heart under control conditions (perfusate without epinephrine) and then again with Krebs solution containing 0.9 l/l epinephrine. Group C served as a control for time-dependent changes.
Fibrillation Induction and Stabilization by Epinephrine
Group A (n=6) The probability of inducing ventricular fibrillation and the probability of spontaneous defibrillation were determined. Ventricular fibrillation threshold was determined in each heart using an up and down protocol. Ventricular fibrillation was induced by 1-s 60 Hz pulses of 2, 3, 4, 5, 7.5, 10, 12.5, 15, 25, 30 and 35 V (peak to peak) delivered through the epicardial patches. Fibrillation was defined by electrograms of irregular morphology, which lasted longer than 1 s after the end of the stimulus. An average of 16 shocks per heart was used to determine the ventricular fibrillation threshold under control conditions (total n=64) and 11 shocks in the presence of epinephrine (total n= 46). A computerized error function fit routine (Jones et al., 1990) was used to construct sigmoidal probability of fibrillation success v shock voltage curves for each heart. The probability curves determined the voltages for 10, 20, 50, 80 and 90% ventricular fibrillation success. The ventricular fibrillation probabilities are reported as the mean±... from the individual curves. A two-way repeated measures analysis of variance on two factors was performed on the results. Differences between the groups were isolated using the Student–Newman–Keuls test. Ventricular fibrillation episodes terminating less than 5 s after the onset of fibrillation were defined as spontaneous defibrillation as shown by the MAP in Figure 1(a). To minimize stress on the heart, fibrillation episodes that did not spontaneously terminate by 5 s were defibrillated by a 10 ms asymmetric biphasic waveform of known effective voltage. In each heart, cycle length was determined manually as shown in Figure 1(b) for at least three episodes of ventricular fibrillation under control conditions and in the presence of epinephrine. Mean ventricular fibrillation cycle length in each heart was determined at the beginning of three successive 1-s intervals beginning 1 s after the termination of the 60 Hz fibrillation current. Action potential duration measurements at 60% repolarization (APD60) were used to determine whether epinephrine shortened action potential duration during fibrillation [Fig. 1(c)]. APD60 was defined as the time between the upstroke of the action potential and the repolarization potential corresponding to 60% of the upstroke amplitude. Mean action potential duration for each episode was defined as the average of the action potential durations measured during the three successive one second intervals. To directly show that epinephrine shortens action
375
potential duration at a specific cycle length, rather than solely by cycle length shortening, we determined action potential duration during pacing in three of the six hearts before and after epinephrine perfusion (this protocol was performed before the induction of fibrillation).
Group B (n=8) In this group, ventricular fibrillation was allowed to continue for 10 s before defibrillation was attempted, this longer time allowed us to perform fast Fourier transform analysis of the last 8 s of fibrillation. The effects of epinephrine on the mean dominant frequency of fibrillation were determined on well-stabilized ventricular fibrillation under control conditions and in the presence of epinephrine. This analysis was performed to verify the consistency of the results on cycle length from group A, because we were able to measure only a limited number of cycles manually. The fast Fourier transform analysis was carried out on epicardial monophasic action potentials using a custom computer program (Labview, National Instruments, Houston, TX, USA).
Group C (n=2) Time dependent changes on the action potential and cycle length were determined. To confirm that the changes in cycle length and action potential duration are due to epinephrine and not to timerelated effects, the identical protocol, which lasted up to 120 min was performed in two additional control rabbits hearts without epinephrine perfusion.
Data analysis Statistical analysis was performed using Sigmastat (Jandel Scientific, San Rafael, CA, USA). Results are expressed as mean±standard error of the mean (...). Differences were determined with paired Student’s t-test, v2 analysis, two way ANOVA and the Student–Newman–Keuls test and were considered significant at P<0.05.
376
O. H. Tovar et al.
Figure 1 (a) Monophasic action potentials recorded during fibrillation and spontaneous defibrillation. The regular pattern at the left of the panel is caused by the 60 Hz fibrillation stimulus. (b) Measurement of mean fibrillation cycle length from a sequence of six monophasic action potentials. (c) Measurement of fibrillation APD60. The mean action potential duration is calculated by measuring six consecutive monophasic action potentials at 60% repolarization, then averaging the measurements.
Results Group A (ventricular fibrillation probability) Figure 2 shows the probability of inducing VF (n=4) as a function of stimulus intensity, in the absence and presence of epinephrine. For the range of stimulus intensities tested, the overall probability of inducing ventricular fibrillation was significantly higher (P<0.03 by repeated measures ANOVA) when the hearts were exposed to epinephrine. A multiple comparison analysis showed that epinephrine significantly reduced fibrillation p80 by 80%, from 23±4 to 4.6±1 V (P<0.05) and reduced fibrillation p90 by 82%, from 29.3±5.4 to 5.4±1.9 V (P<0.05). The normalized curve width, defined by the expression (V80–V20)/V50, decreased from 1.45 under control conditions to 0.65 in the presence of epinephrine. Spontaneous defibrillation To determine the probability of spontaneous defibrillation between 3 and 5 s fibrillation, we ana-
lyzed 80 episodes of fibrillation in the absence of epinephrine and 75 episodes of fibrillation in the presence of epinephrine. Epinephrine significantly decreased (by v2 analysis) the probability of spontaneous defibrillation from 29% (23 out of 80 episodes) of the fibrillation episodes under control conditions to 8% (6 out of 75) of episodes in the presence of epinephrine (P<0.05).
Normal sinus rhythm Figure 3 shows endocardial monophasic action potentials (MAP) and electrograms (EG) recorded from one experiment during normal sinus rhythm and fibrillation. Figures 3(a) and (b) show MAPs recorded during normal sinus rhythm under control conditions and after epinephrine perfusion. Consistent with epinephrine’s known effects to increase heart rate, the cycle length shortened 20% from 300 to 240 ms. In this episode, the APD60 shortened 38%, from 130 to 80 ms. The APD100 shortened 24%, from 170 to 130 ms.
Fibrillation Induction and Stabilization by Epinephrine 100
377
Fibrillation cycle length
Fibrillation probability (%)
90 80 70 60 50 40 30 20 10 0
5
10
15 20 25 Voltage (V)
30
35
40
Figure 2 The probability of fibrillation induction as a function of stimulation voltage (V) in the absence (open circles) and presence of epinephrine (closed circles).
The mean cycle length during normal sinus rhythm in the six rabbits shortened from 324±12 ms (21 episodes under control conditions) to 256±6 ms (27 episodes in the presence of epinephrine) (P<0.05). Mean APD100 shortened from 178±5 ms in the absence of epinephrine to 140±3 ms (P<0.05).
Figures 3(c) and (d) show action potentials recorded 3 s after the onset of fibrillation during a sample control episode and after epinephrine perfusion, respectively. In this episode, epinephrine decreased fibrillation cycle length from 108 to 77 ms (29% decrease). APD60 during fibrillation decreased from 68 to 43 ms (37% decrease). The effect of epinephrine on cycle length and action potential duration during fibrillation is shown in Figure 4. Mean cycle length and mean APD60 during fibrillation were determined under control conditions and in the presence of epinephrine for the 3rd, 4th, and 5th seconds of fibrillation. Figure 4(a) shows mean fibrillation cycle length as a function of fibrillation duration in the absence and in the presence of epinephrine. Under neither condition did cycle length change significantly as a function of time. However, epinephrine significantly shortened mean fibrillation cycle length. For example, at 5 s fibrillation, epinephrine decreased mean fibrillation cycle length by 29% from 104.5±2 ms (control conditions) to 75.7±2.3 ms (P<0.001). Mean APD60 decreased 47% from 76±3 ms under control conditions to
Figure 3 MAPs from a typical experiment. MAPs with corresponding surface electrograms recorded during sinus rhythm in the absence (a) and presence (b) of epinephrine. MAPs with corresponding surface electrograms recorded during ventricular fibrillation in the absence (c) and presence (d) of epinephrine.
378
O. H. Tovar et al. 120
(a)
100
100
80
80 APD60 (ms)
Cycle length (ms)
120
60
60
40
40
20
20
0
2
3
4 Time (s)
5
6
(b)
0
2
3
4 Time (s)
5
6
Figure 4 (a) Fibrillation cycle length as a function of fibrillation duration in the presence (open circles) and absence (closed circles) of epinephrine. (b) APD60 as a function of fibrillation duration in the presence (open circles) and absence (closed circles) of epinephrine.
40±3 ms (P<0.0003) in the presence of epinephrine as shown in Figure 4(b).
Action potential duration during pacing Figure 5 shows original MAP recordings from a typical pacing experiment at cycle lengths of 240 (a), 200 (b), and 150 ms (c), both before and after epinephrine perfusion. In the absence of epinephrine, as the cycle length shortened from 240 to 150 ms APD60 shortened from 115 to 80 ms in a rate-dependent manner. After epinephrine perfusion, the action potential duration also shortened in a rate-dependent manner from 60 to 50 ms as cycle length decreased. Figure 6 shows mean APD60 in the absence (closed symbols) and presence (open symbols) of epinephrine, during paced cycle lengths (squares) of 150, 175, 200 and 240/250 ms and at fibrillation cycle lengths (circles). The difference between the paced APD60 under control conditions and in the presence of epinephrine was statistically significant (P<0.001, by two way ANOVA). Logarithmic regression curves, obtained from the combined paced and fibrillation data, used to fit both groups were APD=−193+57 In (CL) under control conditions and APD=−74+26 In (CL) in the presence of epinephrine. Epinephrine shifted the regression curve downward so that, at a given cycle length, action potential duration was shorter in the epinephrine group.
Group B (frequency analysis of fibrillation) The measurements of fibrillation cycle length in Group A were consistent with the frequency content determined by fast Fourier transforms in the group of rabbits (n=8) in which fibrillation was allowed to persist for 10 s. Figure 7 shows the power spectra obtained from 8 s of fibrillation under control conditions [Figure 7(a)] and in the presence of epinephrine [Figure 7(b)]. The dominant frequency in this control episode was 9.9 Hz (cycle length≈101 ms); the dominant frequency in this epinephrine episode was 13.8 Hz (cycle length≈72 ms). The mean dominant frequency in the control group (n=10 episodes) was 10.0±0.2 Hz, which corresponds to a mean cycle length of 100 ms. The mean dominant frequency in the epinephrine group (n=10 episodes) was 13.3±0.3 Hz, which corresponds to a mean cycle length of 75 ms. This difference was statistically significant (P<0.0001).
Group C (control hearts) In the two control hearts for which the fibrillation protocol was carried out in the absence of epinephrine, Figure 8 shows that, in the absence of epinephrine, fibrillation cycle length does not decrease within the first 2 h of fibrillation. Rather, the fibrillation cycle length tends to increase with time. However, the change was not statistically
Fibrillation Induction and Stabilization by Epinephrine
379
Figure 5 MAPs recorded during pacing at basic cycle lengths of 240 (a), 200 (b), and 150 ms in the absence (left panel) and presence (right panel) of epinephrine.
150
Discussion Ventricular fibrillation probability
APD60 (ms)
100
50
0 50
100
150 200 Cycle length (ms)
250
300
Figure 6 Effect of epinephrine and cycle length on APD during pacing (squares) and fibrillation (circles) under control conditions (closed symbols) and in the presence of epinephrine (open symbols).
significant. This confirms that the shortening produced by epinephrine is not a time-dependent effect.
In contrast to ventricular pacing, which is described by a sharply defined “threshold”, induction of fibrillation by a stimulus is a probabilistic function similar to that describing defibrillation. Therefore, as shown in Figure 2, fibrillation threshold is a poor descriptor of fibrillation induction. Rather the probability of inducing fibrillation increases gradually with stimulus intensity. Under control conditions, a 5 V stimulus induced fibrillation in 10% of the episodes, while a 29 V stimulus induced fibrillation in 90% of episodes. Epinephrine not only shifted the fibrillation probability curve to the left but also markedly decreased the normalized width (increased the normalized slope) of the curve. Due to these two factors, the probability of inducing ventricular fibrillation with a 5 V stimulus increased from 10% under control conditions to 90% in the
380
O. H. Tovar et al.
Cycle length (ms)
120 100 80 60 40 20 0
20
40
60 80 Time (min)
100
Figure 8 Mean fibrillation cycle length as a function of time (min) during 10 s fibrillation episodes in control rabbit hearts. Each point represents the means of fibrillation episodes induced at a specific time in the experiment.
Figure 7 Fast Fourier transforms of MAPs during fibrillation in the absence (a) and presence (b) of epinephrine.
presence of epinephrine. These results agree with reports showing that sympathetic tone increases vulnerability to ventricular fibrillation (Han et al. 1964; Posel et al., 1989) and that epinephrine reduces ventricular fibrillation threshold (Papp and Szekeres, 1968b; Schwartz et al., 1977).
Epinephrine in early fibrillation The role of epinephrine in early fibrillation is unclear. However, clinical studies suggest that sympathetic activation can precipitate ventricular tachycardia or ventricular fibrillation (Meredith et al., 1991). In the case of tachyarrhythmia episodes with loss of aortic pressure (Peuhkurinen et al.,
1994), it may facilitate the degeneration of the tachyarrythmia into fibrillation and hinder spontaneous defibrillation. There is strong experimental and clinical evidence that epinephrine sensitizes the ischemic myocardium to ventricular fibrillation (Bertel et al., 1982; Little et al., 1985; Opie and Lubbe, 1979). In rats, the ventricular fibrillation threshold starts to decrease 6 min after regional ischemia induction and reaches the lowest point 14 min after ischemia (Opie and Lubbe, 1979). A study by Bertel showed that those patients with ventricular fibrillation as a complication of acute myocardial infarction showed the highest plasma levels of epinephrine. They suggested that this phenomenon represented a direct effect of catecholamines on myocardial irritability during the early phase of myocardial infarction (Bertel et al., 1982). This increase in myocardial irritability might be mediated at least partially by an increase in intracellular cAMP by epinephrine (Opie and Lubbe, 1979). Murnaghan (1975) found that, in isolated rabbit hearts, the lowering of ventricular fibrillation threshold was associated with shortening of refractoriness and was blocked by propranolol and pindolol (b-blockers) but it was not influenced by phentolamine (a-blocker). The increased vulnerability at the onset of fibrillation may be due to a decreased excitation threshold of cardiac cells at very short cycle lengths due to shortening of refractoriness (Tovar and Jones, 1997). As we showed in that study, the effects of epinephrine on action potential shortening and excitation threshold at short cycle lengths in isolated myocytes were reversed by propranolol. Again, this suggests that the undesirable effects of epinephrine are due to its b-effects.
Fibrillation Induction and Stabilization by Epinephrine
Spontaneous defibrillation The ability of a healthy heart to sustain fibrillation depends on its size. Spontaneous defibrillation occurs rapidly in healthy small hearts such as those from rats or guinea pigs (West and Laud, 1962; Wigger, 1929) because the small ventricular mass causes fibrillation wavefronts to reach refractory tissue so that fibrillation cannot be sustained. Until recently, spontaneous termination of fibrillation was considered a rare event in humans because of the size of the human heart. However, clinical reports suggest that spontaneous termination of fibrillation in humans may occur, (Clayton et al., 1993; Josephson et al., 1979) although its frequency is unknown (Clayton et al., 1993). This suggests that heart size is only one of several factors which may influence the ability for spontaneous defibrillation to occur (Manoach and Basat, 1990) and that other factors could be optimized in larger hearts to increase the probability of spontaneous defibrillation. Because spontaneous defibrillation occurs frequently in the rabbit heart, it is ideal for a study of drug-induced stabilization of fibrillation (Manoach and Basat, 1990; West and Laud, 1962). In the current study, epinephrine reduced the occurrence of spontaneous termination of fibrillation from 29% of fibrillation episodes under control conditions to 8%.
Effects of epinephrine on action potential duration Action potential duration shortens as a function of cycle length during pacing at sinus rates (Attwell et al., 1981). Restitution curves in pigs also show that, at a cycle length of 350 ms, epinephrine shortened action potential duration for a single extrasystole (Taggart et al., 1990). Cycle lengths shorter than 300 ms were not examined. Furthermore, it has been shown in dogs that epinephrine shortens refractory period at paced cycle length of ≈300 ms (Papp and Szekeres, 1968a). The results of our study also show cycle length dependent action potential shortening. Figure 6 showed that action potential duration decreased logarithmically with cycle length. This cycle length dependence occurred both in the absence and presence of epinephrine; however, epinephrine shifted the regression curve downward and changed its slope so that the greatest effects were observed at fibrillation cycle lengths. Epinephrine shortened action potential duration at the control fibrillation cycle length; this allowed cycle length to further
381
shorten so that a 29% decrease in the mean fibrillation cycle length and a 47% decrease in the mean fibrillation APD60 were produced. Similar results were obtained using myocardial cell aggregates where epinephrine reduced action potential duration at cycle lengths between 160 and 400 ms. Consistent with the downward and leftward shifting of the fibrillation APD60 v cycle length curve (Fig. 6), epinephrine also allowed pacing without alternans at a cycle length of 160 ms and permitted pacing at 140 ms, which could not be achieved in the absence of epinephrine even with high intensity stimuli of four times threshold (Tovar and Jones, 1997). The number of wavelets that can coexist in cardiac tissue has been inversely related to the duration of the refractory period and to the conduction velocity (Moe et al., 1964). Additionally, a study of monophasic action potentials recorded during clinical testing of internal defibrillators (Swartz et al., 1993) showed that during human ventricular fibrillation there is no period of diastole following the fibrillation action potentials. Consistently, during fibrillation episodes in this study, cells were restimulated by new wavefronts immediately following repolarization so that they never achieved a stable resting potential, neither under control conditions nor in the presence of epinephrine. These results suggest that local cycle length during fibrillation is controlled primarily by action potential duration, which, at short fibrillation durations, determines cellular refractory period. Because epinephrine does not appear to significantly alter conduction velocity (Smeets et al., 1986; Vodanovic et al., 1993), shortening of the refractory period by epinephrine allows the presence of more fibrillation wavefronts. Consequently, our present findings suggest that one mechanism through which epinephrine stabilizes fibrillation is by shortening fibrillation action potential duration and therefore refractoriness. This allows a greater number of fibrillation wavefronts to be present in the ventricle and decreases the probability that all wavefronts will become blocked and that fibrillation will be extinguished.
Clinical implications Although spontaneous defibrillation occurs in humans, it is rare. Epinephrine release in a deteriorated myocardium inducing tachycardia or during the initial tachyarrhythmic phase preceding fibrillation may decrease the probability of spontaneous conversion in humans and make it more
382
O. H. Tovar et al.
likely that the initial tachycardia will degenerate to fibrillation. In conclusion, epinephrine facilitates initiation of fibrillation as shown by the increased probability of inducing fibrillation with low intensity stimuli. It also decreases the probability of spontaneous defibrillation. The stabilization of fibrillation is associated with decreased fibrillation cycle length and action potential duration shortening during the early stages of fibrillation.
Acknowledgments We thank Ms Nettie Knight for her expert technical assistance. This work was supported by The Department of Veterans Affairs Medical Center and by USPHS HL 24606 and by HL 49089.
References A D, C I, E DA, 1981. The effects of heart rate on the action potential of guinea-pig and human ventricular muscle. J Physiol (London) 313: 439–461. B O, B¨ FR, B G, R R, B F, 1982. Plasma adrenaline and noradrenaline in patients with acute myocardial infarction. Relationship to ventricular arrhythmias of varying severity. Chest 82: 64– 68. C RH, M A, H PD, C RW, 1993. Self-terminating ventricular tachyarrhythmias—a diagnostic dilemma? The Lancet 341: 93–95. D P, V Y, V M, R MR, 1978. Epinephrine-induced automaticity of canine cardiac purkinje fibers and its relationship to the adenylate cyclase-adenosine 3′,5′-monophosphate system. J Pharmacol Exp Ther 205: 175–182. H J, J P, M GK, 1964. Adrenergic effects on ventricular vulnerability. Circ Res 16: 516–524. J H, N H, R J, 1988. Adrenoceptor-mediated changes of action potential and force of contraction in human ventricular heart muscle. Br J Pharmacol 94: 584–590. J JL, S JF, J RE, 1990. Increasing fibrillation duration enhances relative assymmetrical biphasic versus monophasic defibrillator waveform efficacy. Circ Res 67: 376–384. J ME, S SR, G AM, H LN, 1979. Mechanism of ventricular fibrillation in man. Observations based on electrode catheter recordings. Am J Cardiol 44: 623–631. L RA, F KN, R PE, S HB, Y DW, L GS, K S, B JM, 1985. Plasma
catecholamines in patients with acute myocardial infarction and in cardiac arrest. Quart J Med 54: 133– 140. M M, B MB, 1990. The structural–functional basis of spontaneous ventricular defibrillation. Int J of Cardiol 26: 129–136. M IT, B A, J GL, E MD, 1991. Evidence of a selective increase in cardiac sympathetic activity in patients with sustained ventricular arrhythmias. N Engl J Med 325: 618–624. M GK, R WC, A JA, 1964. A computer model of atrial fibrillation. Am Heart J 67: 200– 220. M MF, 1975. The effect of sympathomimetic amines on the ventricular fibrillation threshold in the rabbit isolated heart. Br J Pharmacol 53: 3–9. O LH, L WF, 1979. Catecholamine-mediated arrhythmias in acute myocardial infarction. Experimental evidence and role of beta-adrenoceptor blockade. S Afr Med J 56: 871–880. P JG, S L, 1968a. Analysis of the mechanism of adrenergic actions on ventricular vulnerability. Euro J of Pharmacol 3: 15–26. P JG, S L, 1968b. The arrhythmogenic action of sympathomimetic amines. Eur J of Pharmacol 3: 4–16. P KJ, H HV, L M, T JT. 1994. Changes in myocardial metabolism and transcardiac electrolytes during simulated ventricular tachycardia: effects of beta-adrenergic blockade. Am Heart J 128: 96–105. P D, N T, K P, 1989. Exercise training after experimental myocardial infarction increases the ventricular fibrillation threshold before and after the onset of reinfarction in the isolated rat heart. Circulation 80: 138–145. S PJ, V RL, L B, 1977. Effect of stellectomy and vagotomy on ventricular refractoriness in dogs. Circ Res 40: 536–540. S JL, A MA, L WJ, B FI, H J, 1986. The wavelength of the cardiac impulse and reentrant arrhythmias in isolated rabbit atrium. The role of heart rate, autonomic transmitters, temperature, and potassium. Circ Res 73: 96–108. S JF, J JL, F R, 1993. Characterization of ventricular fibrillation based on monophasic action potential morphology in the human heart. Circulation 87: 1907–1914. T P, S P, L M, D J, 1990. Interplay between adrenaline and interbeat interval on ventricular repolarization in intact heart in vivo. Cardiovasc Res 24: 884–895. T OH, J JL, 1997. Epinephrine facilitates cardiac fibrillation by shortening action potential refractoriness. J Mol Cell Cardiol 29: 1447–1455. V S, T LA, H RG, K JP, B ZJ, 1993. Transient negative dromotropic effects of catecholamines on canine purkinje fibers exposed to halothane and isoflurane. Anesth Analg 76: 592–597. W TC, L JF, 1962. Minimal mass required for induction of a sustained arrhythmia in isolated atrial segment. Am J Physiol 202: 232–236. W CJ, 1929. Studies of ventricular fibrillation caused by electrical shock. Am Heart J 5: 351–365.