INVITED REVIEW
A Mechanism of Seizure Induction by Electricity and its Clinical Implications Conrad M. Swartz, PhD, MD Abstract: A model of ECT seizure induction by rapid kindling is described. The electrical stimulus as a series of pulses progressively disrupts neuronal cell membranes, with corresponding progressive increases in intracellular concentrations of sodium, calcium, and voltage. Eventually, the intracellular voltage rises to trigger neuronal firing in waves from seizure foci. The quantity of seizure foci produced is expressed by the stimulus charge multiplied by the current cubed. Differences in implications are described between this model and the traditional model that extrapolates from an isolated single neuron undergoing immediate electrical depolarization by a single pulse. Total brain exposure to seizure neurotransmitter release in ECT is analogous to body exposure to medication in drug therapy and may be expressed by a physiological measurement such as electroencephalographic postictal suppression or peak seizure heart rate. Key Words: ECT, electroconvulsive, electricity, seizure, generation, induction, mechanism, stimulus, dose, heart rate (J ECT 2014;30: 94–97)
K
nowing how electroconvulsive therapy (ECT) electrical stimulus induces seizure is clinically useful by allowing us to state the stimulus dose with regard to the electrical quantities we know and can specify. Making analogies to medication dosing can help to understand the meaning and uses of the stimulus dose. Generally, a dose represents an amount of therapeutic agent we can choose and deliver, and it is related to but different from the extent of benefit. A medication dose does not completely state the amount of body exposure to a drug, but it is simpler to identify. In psychotropic medication therapy, daily drug exposure is represented by the 24-hour average blood level, but we routinely simplify by reckoning with a spot blood level instead. Here, we will consider the ECT electrical analogies to medication dose and exposure in a revised presentation of published concepts.1 Various studies indicate that ECT efficacy derives almost entirely from the generalized seizure.2 Efficacy of seizure also follows from the greater clinical improvement from some methods of ECT than others. Because transcranial magnetic brain stimulation and direct current stimulation do not include convulsions, any efficacy they provide implies that the ECT stimulus itself can bring therapeutic benefit, albeit markedly less than the seizure. Here, we will focus on the stimulus dose for inducing seizure.
MECHANISM OF SEIZURE INDUCTION The modern ECT stimulus consists of a series of electron pulses between 2 electrodes alternating in direction. This alternation is sometimes called bidirectional. Each brief pulse lasts 0.3 to 1.5 milliseconds (ms) in duration. Electrical silence between From the Oregon Health & Science University, Portland, OR and Southern Illinois University School of Medicine, Springfield, IL. Received for publication February 24, 2014; accepted March 17, 2014. Reprints: Conrad M. Swartz, MD, PhD, 12911 NW 25th Ct, Vancouver, WA 98685 (e‐mail: [email protected]). Dr Swartz has ownership interests in Medtronic Inc, Somatics LLC, Roche Holding, and the Vanguard Health Care Fund. Copyright © 2014 by Lippincott Williams & Wilkins DOI: 10.1097/YCT.0000000000000139
94
www.ectjournal.com
pulses typically lasts for 6 to 16 ms, approximately 10 times longer than the pulse. A typical ECT stimulus is a train of 100 to 1000 pulses and 1 to 3 times as long as the minimum needed to generate a seizure. The traditionally stated mechanism of seizure induction based on direct depolarization of an isolated single neuron by a single electrical pulse3 is not consistent with the common ECT stimulus of multiple separated pulses whose effects aggregate. Because this traditional mechanism has been discussed for decades, it is worth reviewing. In a single neuron at rest, the inner surface of the cell membrane has a negative voltage relative to the outer surface, approximately −70 mV. When this polarization is neutralized by current between a negative electrode outside the cell body and a positive interior electrode, the single neuron immediately fires. This transmits a wave of depolarization along the neuronal axon that concludes by extruding neurotransmitters onto adjacent neurons. The neuronal cell rapidly returns to the same state it had before the single pulse was applied. This model of neuronal depolarization and firing by a single electrical pulse cannot account for the aggregation of effects from a series of many pulses that comprise a modern brief pulse ECT stimulus. Moreover, the extrapolation from action potential at a single neuron to brain seizure threshold in this hypothesis vastly exceeds reasonability in both microscopic and macroscopic ways. It overlooks modification by other similarly microscopic aspects of brain structure and function such as glial cells. Macroscopically, a human brain is 11 powers of 10 (orders of magnitude) larger than a single neuron. An elephant is 5 orders of magnitude larger than a mouse, and pharmacologic extrapolation from mouse to elephant has been fatally unrealistic.4 Kindling is a means of seizure induction that differs from electrical depolarization. Kindling represents a progressive decrease of seizure threshold with continued exposure to an agent. Administered over several days, pentylenetetrazole (PTZ) or brain electrical shocks delivered in subconvulsive amounts to laboratory animals generate a persistent state of kindling,5 at least in the brain regions studied such as amygdala and frontal lobes. Rapid kindling is achieved by administering subconvulsive electrical shocks more frequently over several hours. The rapid kindling procedure induces both reversible and persistent effects.6 Electroconvulsive therapy seizure induction is presumably rapid kindling in which persistent changes predisposing to seizure are prevented by the seizure itself. This should resemble the mechanism by which a single large dose of PTZ induces seizure. In this rapid kindling mechanism, each pulse in the ECT electrical stimulus lowers seizure threshold until eventually seizure is induced. The ECT stimulus presumably disrupts the neuronal cell membrane and the functioning of its ion active transport channels, increasing leakage of ions across the membrane. In normal homeostasis, the large gradient between high extracellular sodium ion and very low intracellular sodium ion concentrations is maintained by active extrusion of sodium ions from the cell and low permeability of the cell membrane to sodium ions. When the ECT stimulus pulse reaches the membrane, it presumably increases permeability to cations, particularly sodium and calcium, increasing their flux from higher to lower concentrations extracellularly to intracellularly. This influx of (positively charged) Journal of ECT • Volume 30, Number 2, June 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Journal of ECT • Volume 30, Number 2, June 2014
cations decreases the magnitude of the negative voltage on the inner membrane surface. As additional electrical stimulus pulses arrive, cations progressively accumulate inside the cell, and the internal voltage rises to reach the threshold for neuronal firing. These firing neurons are seizure foci. They trigger adjacent neurons to fire, and a wave of neuronal firing forms and sweeps across the brain. Supporting this mechanism are observations that ion flows across the neuronal membrane are involved in kindling. Kindling from PTZ is attenuated by amiloride, a sodium-hydrogen ion exchanger, indicating involvement of sodium flux into the cell.5 Several abnormalities of sodium channels are associated with predisposition to kindling.7 Increased sodium leakage into cells is associated with greater sensitivity to seizure.8 Moreover, suppression of induced seizure by some drugs is related to their inhibition of voltage-dependent sodium channels.9,10 Increased sodium influx may trigger other processes that then provoke seizure. When sodium ions accumulate intracellularly, the normal sodium-calcium exchange process reverses, exporting sodium while importing calcium. Inhibiting this reverse exchange diminishes PTZ-induced seizures,11 suggesting that increased intracellular calcium promotes seizure. Nevertheless, such additional details do not diminish the overall view that progressive changes in intracellular ionic composition during the electrical stimulus mediate seizure induction. How long does the kindling from an ECT electrical stimulus persist? An estimate of this is provided by the electroencephalographic (EEG) seizure duration itself, that is, typically a minute or 2 minutes. Measurements of the results of giving 2 separated identical individually subconvulsive ECT stimuli indicate that kindling from an electrical stimulus persists for at least 1 minute.12 The principle of homeostasis implies that seizure activity inhibits kindling, and so it seems likely that kindling resulting from subconvulsive ECT stimuli persists longer than kindling after stimuli that induce a grand mal seizure. The picture that ECT seizure persists primarily while substantial kindling persists is consistent with the model of seizure induction by kindling. In contrast, extrapolating from the model of direct depolarization of an isolated single neuron to seizure has no apparent implication about seizure duration. During the electrical stimulus, several successive waves of neuronal firing should form, with more waves forming during longer stimuli. It is not possible to observe these directly because the electrical stimulus obscures EEG observation of brain activity. A grand mal seizure involves many waves of neuronal firing, as it consists of 2 to 5 waves per second lasting 20 to 60 seconds, thus 40 to 300 waves of firing. This rate of wave generation does not seem consistent with the single neuron depolarization model in which each pulse triggers a wave of neuronal firing because typical rates of pulse delivery are 60/s (30 Hz) to 140/s (70 Hz). Single neuron depolarization by one electrical pulse is the conceptual foundation for the stimulus titration measurement of seizure threshold. It provides the rationale for counting only the final stimulus in a titration sequence. However, the rapid kindling mechanism indicates that this rationale is in error when 2 or more stimuli are administered in within a period of less than at least 1 minute. In stimulus titration, progressively larger electrical charges are administered until a convulsion develops. If a stimulus is not followed by a convulsion within approximately 20 seconds, another stimulus of greater charge is given. However, as previously noted, ECT stimuli contribute to kindling for at least a minute. Accordingly, with each additional stimulus administered, the “method of limits” measurement of seizure threshold further underestimates the seizure threshold. Other internal inconsistencies in using seizure threshold to select the stimulus dose are reviewed elsewhere.13 © 2014 Lippincott Williams & Wilkins
Electrical Seizure Induction Mechanism
IDENTIFYING THE NATURE OF STIMULUS DOSE The ECT stimulus dose corresponds to the seizure-inducing activity of the stimulus. We aim to identify this dose from the electrical aspects of the stimulus that we can choose or know. Traditionally, the stimulus dose has been said to be the charge (in mC) or the energy (in J). These traditions are not correct because, when given over several hours, even several times the amount of electrical charge and energy in a typical ECT stimulus would be imperceptible and plainly incapable of inducing seizure. Moreover, a minimum current (or voltage gradient) is needed to disrupt the neuronal cell membrane and produce kindling. Plainly, the stimulus dose must reflect the rate of delivery of charge and energy. Conceptually, the stimulus dose represents the quantity of seizure foci produced by the electrical stimulus. A model of ECT seizure generation has been described1,14 in which the volume of a hemisphere of radius E around the electrode site is (2/3)πE3. In this model, E represents both distance and voltage drop from the electrode. The amount of kindled brain equals the volume within this hemisphere multiplied by the number of voltage carriers (ie, the charge). Therefore, the kindled brain amount, which represents the stimulus dose, increases with the charge multiplied by the cube of the voltage drop. Using the Ohm’s law to replace the voltage drop with the current, the stimulus dose is proportional to the charge multiplied by the cube of the current. This proportionality is all we need to compare the doses of stimuli with different currents or charges. For simplicity, we can take the proportionality coefficient as unity (ie, one) so that stimulus dose equals charge times current cubed for bilateral ECT. Clinically, this expression of stimulus dose indicates that at the same charge, a stimulus of current I 3 has (I 3/I 2)3 times the seizure-inducing dose as a stimulus of current I 2, for bilateral ECT. As a specific example of what this means, an 800-mA stimulus requires 1.42 times the charge needed by a 900-mA stimulus to reach seizure threshold. This was confirmed by direct randomized comparisons measuring 88 patients receiving bitemporal ECT, all at 1-ms pulse width.15 In these comparisons, use of electrical stimuli 800 mA or less required 1.61 times as much charge to reach seizure threshold as stimuli of 900 mA. The observed ratio of 1.61 may be higher than the 1.42 ratio expected because of occasional use of currents below 800 mA in the group of patients receiving lower current. With this formula, the physician can calculate the amount of charge needed for equal stimulus doses at different stimulus currents. This should be needed when changing the current used in a stimulus as when changing from one ECT device to another or adjusting the stimulus current. Similarly, a repeated-measures comparison of 900- and 1150-mA bitemporal stimuli at 0.5-ms pulse width in randomized balanced order observed that seizure induction required 30% greater charge at the lower current.16 This result simply confirms that the ECT stimulus dose depends on current separately from charge. By the expression of stimulus dose as charge times current cubed, a charge of 1150 mA should have twice the dose of a 900-mA charge. That the increase was only 30% suggests that under the treatment circumstances (ie, 0.5-ms pulse width), there was a ceiling on the effect of current beyond its contribution to charge. This should occur if the dynamic impedance is lower in brain regions with more intense current density and seizure. This is expected because the impedance to the convulsive current (“dynamic impedance”) is much lower than the static impedance to the much smaller current given to test the connection before treatment. An analogous ceiling effect in medication therapy is a limit on absorption rate sometimes www.ectjournal.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
95
Journal of ECT • Volume 30, Number 2, June 2014
Swartz
seen because of saturation of active transport across the intestinal wall or loss of intestinal motion (eg, ileus) from the medication. In this medication analogy, the limit on absorption does not change the dose; it decreases the efficiency of the dose. Of course, ECT stimulus efficiency likewise influences seizure induction. In analogy to gasoline in an automobile, dose is gallons of gasoline and efficiency is miles per gallon. Observations about the dependence of stimulus efficiency on wave shape, width of pulses or phases, frequency, and charge rate are reviewed elsewhere.1 There is no clear rationale for the electrical waveform characteristics that most efficiently depolarize an isolated single neuron to most efficiently induce rapid kindling, excepting the distant similarity that they are both processes that lead to neuronal firing. If cellular influx of calcium or another cation that is actively exchanged for sodium is important in rapid kindling, as noted above, the interpulse interval may provide an opportunity for this active cation exchange to occur. This interpulse opportunity may be influenced by electrical waveform characteristics such as pulse width and interpulse interval. The model of the amount of electrical seizure induction1,14 is a first approximation and does not account for the structure of the CNS or the skull. Introducing such detail adds the complexity of needing measurements of anatomical and histological details for individual patients, along with computations about their effects on current density. It is not clear that such additional large and costly efforts will be clinically useful. Still, one structural detail can be addressed without adding much complexity, an approximation of skull shape for standard right-unilateral electrode placement. In this, with electrodes at the right temple and vertex, the amount of seizure foci should increase proportionally to an exponential power of current between 2 and 3. Although we do not know a priori the most accurate value, we can take the middle value, 2.5, as an approximation. That is, for unilateral ECT stimulus dose equals charge times current raised to the 2.5 power.
AMOUNT OF BRAIN EXPOSURE TO THE STIMULUS Returning to the medication analogy mentioned at the beginning, the amount of an effect a drug has on the body is more closely related to the body exposure to the drug than to the drug dose. As a shortcut to measure body exposure, we use trough blood drug levels for lithium or anticonvulsants. For drugs of half-life several days or longer, the trough blood level consistently expresses the body exposure per day. With a shorter half-life, the trough level is strongly influenced by variations in half-life and dose division. Nevertheless, it is common practice to use trough blood drug levels and not calculate the body exposure to lithium or anticonvulsants, although this can be done with a graph by using one blood drug level, the dose schedule, and an age-based estimate of drug half-life.17 Making an ECT analogy to body drug exposure and expressing the amount of brain exposure to the stimulus involve understanding ECT mechanism beyond seizure induction. In the “reboot mechanism” of ECT, replacement of pathological patterns of neurotransmitters is involved.1 In an initial step of describing this process, the amount of neurotransmitter release toward therapeutic effect is analogous to body drug exposure. This macroscopic concept of neurotransmitter release hopes to sidestep the vast complexity of microscopic neurotransmitter regulation and interactions. Presumably, this complexity is partially reflected by the varieties of different patterns seen in the EEG during ECT seizure, and we do not yet know about associations between these EEG patterns and specific neurotransmitter changes. Postictal
96
www.ectjournal.com
suppression is an EEG pattern that appears to represent a summation of neurotransmitter effects. This should be expressed in the suppression of baseline EEG. Its measurement can be interfered with by continuing seizure activity in the brain, and this interference seems common. Measurements of the time course of ECTinduced hormones may reflect a summation of neurotransmitter effects, but the complexities involved are formidable and have yet to be overcome. The complexities include extremely large interindividual variations in hormone responses and needs for several accurately timed blood samplings and reliable prompt hormone assays.18 A simple, free, physiological, and easily accessible apparent measure of exposure to seizure neurotransmitter release is heart rate acceleration during ECT, when it is not obscured by medications or cardiac disease. It presumably represents the exposure of cardioacceleratory brain and brain stem regions to seizure activity. Clinical study has reported initial evidence that peak heart rate during ECT seizure is associated with efficacy.19 In this study, the peak seizure heart rate was compared to the individual patients’ maximum. This use of peak heart rate was explained in a step-by-step illustration of the “benchmark method” for setting ECT stimulus dose and regulating it along the course of treatment.20 This method is similarly applicable with postictal EEG suppression or total prolactin release instead.
CONCLUSIONS A model of ECT seizure induction is presented consistent with the ECT stimulus as a series of pulses whose effects aggregate over a period much longer than the duration of each pulse, typically for up to 2 minutes. This mechanism is a form of rapid kindling, and it differs from concepts modeled on the depolarization of a single isolated neuron. Each pulse presumably disrupts the neuronal cell membrane, opening channels to sodium and calcium ions. This mechanism is a basis for identifying the ECT stimulus dose as the expected quantity of seizure foci. This dose is proportional to the charge multiplied by the cube of the current. Two separate experimental studies confirmed the strong dependence of dose on current besides charge. Total brain exposure to seizure neurotransmitter release in ECT is analogous to body exposure to medication in drug therapy. Possible reflections of such brain exposure may be seen in postictal suppression of baseline EEG, induced hormone release, or heart rate elevation during the seizure. With the benchmark method, these physiological measurements may be used to regulate the stimulus dose along the course of ECT treatment. REFERENCES 1. Swartz CM. Electricity and electroconvulsive therapy. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York, NY: Cambridge University Press; 2009:1–16. 2. Cronholm B, Ottosson JO. Experimental studies of the therapeutic action of electroconvulsive therapy in endogenous depression. The role of the electrical stimulation and of the seizure studied by variation of stimulus intensity and modification by lidocaine of seizure discharge. Acta Psychiatr Scand Suppl. 1960;35:69–101. 3. Sackeim HA, Long J, Luber B, et al. Physical properties and quantification of the ECT stimulus: I. Basic principles. Convuls Ther. 1994;10:93–123. 4. West LJ, Pierce CM, Thomas WD. Lysergic acid diethylamide: its effects on a male Asiatic elephant. Science. 1962;138:1100–1103. 5. Ali A, Ahmad FJ, Pillai KK, et al. Amiloride protects against pentylenetetrazole-induced kindling in mice. Br J Pharmacol. 2005;145:880–884.
© 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Journal of ECT • Volume 30, Number 2, June 2014
Electrical Seizure Induction Mechanism
6. Lothman EW, Williamson JM. Closely spaced recurrent hippocampal seizures elicit two types of heightened epileptogenesis: a rapidly developing, transient kindling and a slowly developing, enduring kindling. Brain Res. 1994;649:71–84.
14. Swartz CM. ECT stimulus dose expressed as volume of seizure foci. J.ECT. 2006;22:54–58.
7. Blumenfeld H, Lampert A, Klein JP, et al. Role of hippocampal sodium channel Nav1.6 in kindling epileptogenesis. Epilepsia. 2009;50:44–55.
15. Chanpattana W. Seizure threshold in electroconvulsive therapy: effect of instrument titration schedule. German J Psychiatry. 2001;4:51–56.
8. Marley R, Baines RA. Increased persistent Na+ current contributes to seizure in the slamdance bang-sensitive Drosophila mutant. J Neurophysiol. 2011;106:18–29.
16. Swartz CM, Krohmer R, Michael N. ECT stimulus dose dependence on current separately from charge. Psychiatry Res. 2012; 198:164–165.
9. Fischer W, Kittner H, Regenthal R, et al. Anticonvulsant and sodium channel blocking activity of higher doses of clenbuterol. Naunyn Schmiedebergs Arch Pharmacol. 2001;363:182–192.
17. Swartz CM. Correction of blood drug levels for changes over the day. J Clin Pharmacol. 1989;29:234–239.
10. Large CH, Kalinichev M, Lucas A, et al. The relationship between sodium channel inhibition and anticonvulsant activity in a model of generalised seizure in the rat. Epilepsy Res. 2009;85:96–106. 11. N’Gouemo P. Probing the role of the sodium/calcium exchanger in pentylenetetrazole-induced generalized seizures in rats. Brain Res Bull. 2013;90:52–57. 12. Swartz CM. Repeated ECT stimuli and the seizure threshold. Convuls Ther. 1990;6:181–182.
© 2014 Lippincott Williams & Wilkins
13. Swartz CM. Stimulus dosing in electroconvulsive therapy and the threshold multiple method. J ECT. 2001;17:87–90.
18. Swartz CM. Hormonal effects of ECT. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York, NY: Cambridge University Press; 2009:149–163. 19. Swartz CM. Heart rate and ECT. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York: Cambridge University Press; 2009:477–484. 20. Swartz CM. ECT dosing by the benchmark method. German J Psychiatry. 2002;5:1–4. Available at: http://www.gjpsy.uni-goettingen. de/gjp-article-swartz.pdf. Accessed March 16, 2014.
www.ectjournal.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
97
A Mechanism of Seizure Induction by Electricity and its Clinical Implications Conrad M. Swartz, PhD, MD Abstract: A model of ECT seizure induction by rapid kindling is described. The electrical stimulus as a series of pulses progressively disrupts neuronal cell membranes, with corresponding progressive increases in intracellular concentrations of sodium, calcium, and voltage. Eventually, the intracellular voltage rises to trigger neuronal firing in waves from seizure foci. The quantity of seizure foci produced is expressed by the stimulus charge multiplied by the current cubed. Differences in implications are described between this model and the traditional model that extrapolates from an isolated single neuron undergoing immediate electrical depolarization by a single pulse. Total brain exposure to seizure neurotransmitter release in ECT is analogous to body exposure to medication in drug therapy and may be expressed by a physiological measurement such as electroencephalographic postictal suppression or peak seizure heart rate. Key Words: ECT, electroconvulsive, electricity, seizure, generation, induction, mechanism, stimulus, dose, heart rate (J ECT 2014;30: 94–97)
K
nowing how electroconvulsive therapy (ECT) electrical stimulus induces seizure is clinically useful by allowing us to state the stimulus dose with regard to the electrical quantities we know and can specify. Making analogies to medication dosing can help to understand the meaning and uses of the stimulus dose. Generally, a dose represents an amount of therapeutic agent we can choose and deliver, and it is related to but different from the extent of benefit. A medication dose does not completely state the amount of body exposure to a drug, but it is simpler to identify. In psychotropic medication therapy, daily drug exposure is represented by the 24-hour average blood level, but we routinely simplify by reckoning with a spot blood level instead. Here, we will consider the ECT electrical analogies to medication dose and exposure in a revised presentation of published concepts.1 Various studies indicate that ECT efficacy derives almost entirely from the generalized seizure.2 Efficacy of seizure also follows from the greater clinical improvement from some methods of ECT than others. Because transcranial magnetic brain stimulation and direct current stimulation do not include convulsions, any efficacy they provide implies that the ECT stimulus itself can bring therapeutic benefit, albeit markedly less than the seizure. Here, we will focus on the stimulus dose for inducing seizure.
MECHANISM OF SEIZURE INDUCTION The modern ECT stimulus consists of a series of electron pulses between 2 electrodes alternating in direction. This alternation is sometimes called bidirectional. Each brief pulse lasts 0.3 to 1.5 milliseconds (ms) in duration. Electrical silence between From the Oregon Health & Science University, Portland, OR and Southern Illinois University School of Medicine, Springfield, IL. Received for publication February 24, 2014; accepted March 17, 2014. Reprints: Conrad M. Swartz, MD, PhD, 12911 NW 25th Ct, Vancouver, WA 98685 (e‐mail: [email protected]). Dr Swartz has ownership interests in Medtronic Inc, Somatics LLC, Roche Holding, and the Vanguard Health Care Fund. Copyright © 2014 by Lippincott Williams & Wilkins DOI: 10.1097/YCT.0000000000000139
94
www.ectjournal.com
pulses typically lasts for 6 to 16 ms, approximately 10 times longer than the pulse. A typical ECT stimulus is a train of 100 to 1000 pulses and 1 to 3 times as long as the minimum needed to generate a seizure. The traditionally stated mechanism of seizure induction based on direct depolarization of an isolated single neuron by a single electrical pulse3 is not consistent with the common ECT stimulus of multiple separated pulses whose effects aggregate. Because this traditional mechanism has been discussed for decades, it is worth reviewing. In a single neuron at rest, the inner surface of the cell membrane has a negative voltage relative to the outer surface, approximately −70 mV. When this polarization is neutralized by current between a negative electrode outside the cell body and a positive interior electrode, the single neuron immediately fires. This transmits a wave of depolarization along the neuronal axon that concludes by extruding neurotransmitters onto adjacent neurons. The neuronal cell rapidly returns to the same state it had before the single pulse was applied. This model of neuronal depolarization and firing by a single electrical pulse cannot account for the aggregation of effects from a series of many pulses that comprise a modern brief pulse ECT stimulus. Moreover, the extrapolation from action potential at a single neuron to brain seizure threshold in this hypothesis vastly exceeds reasonability in both microscopic and macroscopic ways. It overlooks modification by other similarly microscopic aspects of brain structure and function such as glial cells. Macroscopically, a human brain is 11 powers of 10 (orders of magnitude) larger than a single neuron. An elephant is 5 orders of magnitude larger than a mouse, and pharmacologic extrapolation from mouse to elephant has been fatally unrealistic.4 Kindling is a means of seizure induction that differs from electrical depolarization. Kindling represents a progressive decrease of seizure threshold with continued exposure to an agent. Administered over several days, pentylenetetrazole (PTZ) or brain electrical shocks delivered in subconvulsive amounts to laboratory animals generate a persistent state of kindling,5 at least in the brain regions studied such as amygdala and frontal lobes. Rapid kindling is achieved by administering subconvulsive electrical shocks more frequently over several hours. The rapid kindling procedure induces both reversible and persistent effects.6 Electroconvulsive therapy seizure induction is presumably rapid kindling in which persistent changes predisposing to seizure are prevented by the seizure itself. This should resemble the mechanism by which a single large dose of PTZ induces seizure. In this rapid kindling mechanism, each pulse in the ECT electrical stimulus lowers seizure threshold until eventually seizure is induced. The ECT stimulus presumably disrupts the neuronal cell membrane and the functioning of its ion active transport channels, increasing leakage of ions across the membrane. In normal homeostasis, the large gradient between high extracellular sodium ion and very low intracellular sodium ion concentrations is maintained by active extrusion of sodium ions from the cell and low permeability of the cell membrane to sodium ions. When the ECT stimulus pulse reaches the membrane, it presumably increases permeability to cations, particularly sodium and calcium, increasing their flux from higher to lower concentrations extracellularly to intracellularly. This influx of (positively charged) Journal of ECT • Volume 30, Number 2, June 2014
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Journal of ECT • Volume 30, Number 2, June 2014
cations decreases the magnitude of the negative voltage on the inner membrane surface. As additional electrical stimulus pulses arrive, cations progressively accumulate inside the cell, and the internal voltage rises to reach the threshold for neuronal firing. These firing neurons are seizure foci. They trigger adjacent neurons to fire, and a wave of neuronal firing forms and sweeps across the brain. Supporting this mechanism are observations that ion flows across the neuronal membrane are involved in kindling. Kindling from PTZ is attenuated by amiloride, a sodium-hydrogen ion exchanger, indicating involvement of sodium flux into the cell.5 Several abnormalities of sodium channels are associated with predisposition to kindling.7 Increased sodium leakage into cells is associated with greater sensitivity to seizure.8 Moreover, suppression of induced seizure by some drugs is related to their inhibition of voltage-dependent sodium channels.9,10 Increased sodium influx may trigger other processes that then provoke seizure. When sodium ions accumulate intracellularly, the normal sodium-calcium exchange process reverses, exporting sodium while importing calcium. Inhibiting this reverse exchange diminishes PTZ-induced seizures,11 suggesting that increased intracellular calcium promotes seizure. Nevertheless, such additional details do not diminish the overall view that progressive changes in intracellular ionic composition during the electrical stimulus mediate seizure induction. How long does the kindling from an ECT electrical stimulus persist? An estimate of this is provided by the electroencephalographic (EEG) seizure duration itself, that is, typically a minute or 2 minutes. Measurements of the results of giving 2 separated identical individually subconvulsive ECT stimuli indicate that kindling from an electrical stimulus persists for at least 1 minute.12 The principle of homeostasis implies that seizure activity inhibits kindling, and so it seems likely that kindling resulting from subconvulsive ECT stimuli persists longer than kindling after stimuli that induce a grand mal seizure. The picture that ECT seizure persists primarily while substantial kindling persists is consistent with the model of seizure induction by kindling. In contrast, extrapolating from the model of direct depolarization of an isolated single neuron to seizure has no apparent implication about seizure duration. During the electrical stimulus, several successive waves of neuronal firing should form, with more waves forming during longer stimuli. It is not possible to observe these directly because the electrical stimulus obscures EEG observation of brain activity. A grand mal seizure involves many waves of neuronal firing, as it consists of 2 to 5 waves per second lasting 20 to 60 seconds, thus 40 to 300 waves of firing. This rate of wave generation does not seem consistent with the single neuron depolarization model in which each pulse triggers a wave of neuronal firing because typical rates of pulse delivery are 60/s (30 Hz) to 140/s (70 Hz). Single neuron depolarization by one electrical pulse is the conceptual foundation for the stimulus titration measurement of seizure threshold. It provides the rationale for counting only the final stimulus in a titration sequence. However, the rapid kindling mechanism indicates that this rationale is in error when 2 or more stimuli are administered in within a period of less than at least 1 minute. In stimulus titration, progressively larger electrical charges are administered until a convulsion develops. If a stimulus is not followed by a convulsion within approximately 20 seconds, another stimulus of greater charge is given. However, as previously noted, ECT stimuli contribute to kindling for at least a minute. Accordingly, with each additional stimulus administered, the “method of limits” measurement of seizure threshold further underestimates the seizure threshold. Other internal inconsistencies in using seizure threshold to select the stimulus dose are reviewed elsewhere.13 © 2014 Lippincott Williams & Wilkins
Electrical Seizure Induction Mechanism
IDENTIFYING THE NATURE OF STIMULUS DOSE The ECT stimulus dose corresponds to the seizure-inducing activity of the stimulus. We aim to identify this dose from the electrical aspects of the stimulus that we can choose or know. Traditionally, the stimulus dose has been said to be the charge (in mC) or the energy (in J). These traditions are not correct because, when given over several hours, even several times the amount of electrical charge and energy in a typical ECT stimulus would be imperceptible and plainly incapable of inducing seizure. Moreover, a minimum current (or voltage gradient) is needed to disrupt the neuronal cell membrane and produce kindling. Plainly, the stimulus dose must reflect the rate of delivery of charge and energy. Conceptually, the stimulus dose represents the quantity of seizure foci produced by the electrical stimulus. A model of ECT seizure generation has been described1,14 in which the volume of a hemisphere of radius E around the electrode site is (2/3)πE3. In this model, E represents both distance and voltage drop from the electrode. The amount of kindled brain equals the volume within this hemisphere multiplied by the number of voltage carriers (ie, the charge). Therefore, the kindled brain amount, which represents the stimulus dose, increases with the charge multiplied by the cube of the voltage drop. Using the Ohm’s law to replace the voltage drop with the current, the stimulus dose is proportional to the charge multiplied by the cube of the current. This proportionality is all we need to compare the doses of stimuli with different currents or charges. For simplicity, we can take the proportionality coefficient as unity (ie, one) so that stimulus dose equals charge times current cubed for bilateral ECT. Clinically, this expression of stimulus dose indicates that at the same charge, a stimulus of current I 3 has (I 3/I 2)3 times the seizure-inducing dose as a stimulus of current I 2, for bilateral ECT. As a specific example of what this means, an 800-mA stimulus requires 1.42 times the charge needed by a 900-mA stimulus to reach seizure threshold. This was confirmed by direct randomized comparisons measuring 88 patients receiving bitemporal ECT, all at 1-ms pulse width.15 In these comparisons, use of electrical stimuli 800 mA or less required 1.61 times as much charge to reach seizure threshold as stimuli of 900 mA. The observed ratio of 1.61 may be higher than the 1.42 ratio expected because of occasional use of currents below 800 mA in the group of patients receiving lower current. With this formula, the physician can calculate the amount of charge needed for equal stimulus doses at different stimulus currents. This should be needed when changing the current used in a stimulus as when changing from one ECT device to another or adjusting the stimulus current. Similarly, a repeated-measures comparison of 900- and 1150-mA bitemporal stimuli at 0.5-ms pulse width in randomized balanced order observed that seizure induction required 30% greater charge at the lower current.16 This result simply confirms that the ECT stimulus dose depends on current separately from charge. By the expression of stimulus dose as charge times current cubed, a charge of 1150 mA should have twice the dose of a 900-mA charge. That the increase was only 30% suggests that under the treatment circumstances (ie, 0.5-ms pulse width), there was a ceiling on the effect of current beyond its contribution to charge. This should occur if the dynamic impedance is lower in brain regions with more intense current density and seizure. This is expected because the impedance to the convulsive current (“dynamic impedance”) is much lower than the static impedance to the much smaller current given to test the connection before treatment. An analogous ceiling effect in medication therapy is a limit on absorption rate sometimes www.ectjournal.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
95
Journal of ECT • Volume 30, Number 2, June 2014
Swartz
seen because of saturation of active transport across the intestinal wall or loss of intestinal motion (eg, ileus) from the medication. In this medication analogy, the limit on absorption does not change the dose; it decreases the efficiency of the dose. Of course, ECT stimulus efficiency likewise influences seizure induction. In analogy to gasoline in an automobile, dose is gallons of gasoline and efficiency is miles per gallon. Observations about the dependence of stimulus efficiency on wave shape, width of pulses or phases, frequency, and charge rate are reviewed elsewhere.1 There is no clear rationale for the electrical waveform characteristics that most efficiently depolarize an isolated single neuron to most efficiently induce rapid kindling, excepting the distant similarity that they are both processes that lead to neuronal firing. If cellular influx of calcium or another cation that is actively exchanged for sodium is important in rapid kindling, as noted above, the interpulse interval may provide an opportunity for this active cation exchange to occur. This interpulse opportunity may be influenced by electrical waveform characteristics such as pulse width and interpulse interval. The model of the amount of electrical seizure induction1,14 is a first approximation and does not account for the structure of the CNS or the skull. Introducing such detail adds the complexity of needing measurements of anatomical and histological details for individual patients, along with computations about their effects on current density. It is not clear that such additional large and costly efforts will be clinically useful. Still, one structural detail can be addressed without adding much complexity, an approximation of skull shape for standard right-unilateral electrode placement. In this, with electrodes at the right temple and vertex, the amount of seizure foci should increase proportionally to an exponential power of current between 2 and 3. Although we do not know a priori the most accurate value, we can take the middle value, 2.5, as an approximation. That is, for unilateral ECT stimulus dose equals charge times current raised to the 2.5 power.
AMOUNT OF BRAIN EXPOSURE TO THE STIMULUS Returning to the medication analogy mentioned at the beginning, the amount of an effect a drug has on the body is more closely related to the body exposure to the drug than to the drug dose. As a shortcut to measure body exposure, we use trough blood drug levels for lithium or anticonvulsants. For drugs of half-life several days or longer, the trough blood level consistently expresses the body exposure per day. With a shorter half-life, the trough level is strongly influenced by variations in half-life and dose division. Nevertheless, it is common practice to use trough blood drug levels and not calculate the body exposure to lithium or anticonvulsants, although this can be done with a graph by using one blood drug level, the dose schedule, and an age-based estimate of drug half-life.17 Making an ECT analogy to body drug exposure and expressing the amount of brain exposure to the stimulus involve understanding ECT mechanism beyond seizure induction. In the “reboot mechanism” of ECT, replacement of pathological patterns of neurotransmitters is involved.1 In an initial step of describing this process, the amount of neurotransmitter release toward therapeutic effect is analogous to body drug exposure. This macroscopic concept of neurotransmitter release hopes to sidestep the vast complexity of microscopic neurotransmitter regulation and interactions. Presumably, this complexity is partially reflected by the varieties of different patterns seen in the EEG during ECT seizure, and we do not yet know about associations between these EEG patterns and specific neurotransmitter changes. Postictal
96
www.ectjournal.com
suppression is an EEG pattern that appears to represent a summation of neurotransmitter effects. This should be expressed in the suppression of baseline EEG. Its measurement can be interfered with by continuing seizure activity in the brain, and this interference seems common. Measurements of the time course of ECTinduced hormones may reflect a summation of neurotransmitter effects, but the complexities involved are formidable and have yet to be overcome. The complexities include extremely large interindividual variations in hormone responses and needs for several accurately timed blood samplings and reliable prompt hormone assays.18 A simple, free, physiological, and easily accessible apparent measure of exposure to seizure neurotransmitter release is heart rate acceleration during ECT, when it is not obscured by medications or cardiac disease. It presumably represents the exposure of cardioacceleratory brain and brain stem regions to seizure activity. Clinical study has reported initial evidence that peak heart rate during ECT seizure is associated with efficacy.19 In this study, the peak seizure heart rate was compared to the individual patients’ maximum. This use of peak heart rate was explained in a step-by-step illustration of the “benchmark method” for setting ECT stimulus dose and regulating it along the course of treatment.20 This method is similarly applicable with postictal EEG suppression or total prolactin release instead.
CONCLUSIONS A model of ECT seizure induction is presented consistent with the ECT stimulus as a series of pulses whose effects aggregate over a period much longer than the duration of each pulse, typically for up to 2 minutes. This mechanism is a form of rapid kindling, and it differs from concepts modeled on the depolarization of a single isolated neuron. Each pulse presumably disrupts the neuronal cell membrane, opening channels to sodium and calcium ions. This mechanism is a basis for identifying the ECT stimulus dose as the expected quantity of seizure foci. This dose is proportional to the charge multiplied by the cube of the current. Two separate experimental studies confirmed the strong dependence of dose on current besides charge. Total brain exposure to seizure neurotransmitter release in ECT is analogous to body exposure to medication in drug therapy. Possible reflections of such brain exposure may be seen in postictal suppression of baseline EEG, induced hormone release, or heart rate elevation during the seizure. With the benchmark method, these physiological measurements may be used to regulate the stimulus dose along the course of ECT treatment. REFERENCES 1. Swartz CM. Electricity and electroconvulsive therapy. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York, NY: Cambridge University Press; 2009:1–16. 2. Cronholm B, Ottosson JO. Experimental studies of the therapeutic action of electroconvulsive therapy in endogenous depression. The role of the electrical stimulation and of the seizure studied by variation of stimulus intensity and modification by lidocaine of seizure discharge. Acta Psychiatr Scand Suppl. 1960;35:69–101. 3. Sackeim HA, Long J, Luber B, et al. Physical properties and quantification of the ECT stimulus: I. Basic principles. Convuls Ther. 1994;10:93–123. 4. West LJ, Pierce CM, Thomas WD. Lysergic acid diethylamide: its effects on a male Asiatic elephant. Science. 1962;138:1100–1103. 5. Ali A, Ahmad FJ, Pillai KK, et al. Amiloride protects against pentylenetetrazole-induced kindling in mice. Br J Pharmacol. 2005;145:880–884.
© 2014 Lippincott Williams & Wilkins
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Journal of ECT • Volume 30, Number 2, June 2014
Electrical Seizure Induction Mechanism
6. Lothman EW, Williamson JM. Closely spaced recurrent hippocampal seizures elicit two types of heightened epileptogenesis: a rapidly developing, transient kindling and a slowly developing, enduring kindling. Brain Res. 1994;649:71–84.
14. Swartz CM. ECT stimulus dose expressed as volume of seizure foci. J.ECT. 2006;22:54–58.
7. Blumenfeld H, Lampert A, Klein JP, et al. Role of hippocampal sodium channel Nav1.6 in kindling epileptogenesis. Epilepsia. 2009;50:44–55.
15. Chanpattana W. Seizure threshold in electroconvulsive therapy: effect of instrument titration schedule. German J Psychiatry. 2001;4:51–56.
8. Marley R, Baines RA. Increased persistent Na+ current contributes to seizure in the slamdance bang-sensitive Drosophila mutant. J Neurophysiol. 2011;106:18–29.
16. Swartz CM, Krohmer R, Michael N. ECT stimulus dose dependence on current separately from charge. Psychiatry Res. 2012; 198:164–165.
9. Fischer W, Kittner H, Regenthal R, et al. Anticonvulsant and sodium channel blocking activity of higher doses of clenbuterol. Naunyn Schmiedebergs Arch Pharmacol. 2001;363:182–192.
17. Swartz CM. Correction of blood drug levels for changes over the day. J Clin Pharmacol. 1989;29:234–239.
10. Large CH, Kalinichev M, Lucas A, et al. The relationship between sodium channel inhibition and anticonvulsant activity in a model of generalised seizure in the rat. Epilepsy Res. 2009;85:96–106. 11. N’Gouemo P. Probing the role of the sodium/calcium exchanger in pentylenetetrazole-induced generalized seizures in rats. Brain Res Bull. 2013;90:52–57. 12. Swartz CM. Repeated ECT stimuli and the seizure threshold. Convuls Ther. 1990;6:181–182.
© 2014 Lippincott Williams & Wilkins
13. Swartz CM. Stimulus dosing in electroconvulsive therapy and the threshold multiple method. J ECT. 2001;17:87–90.
18. Swartz CM. Hormonal effects of ECT. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York, NY: Cambridge University Press; 2009:149–163. 19. Swartz CM. Heart rate and ECT. In: Swartz Conrad, ed. Electroconvulsive and Neuromodulation Therapies. New York: Cambridge University Press; 2009:477–484. 20. Swartz CM. ECT dosing by the benchmark method. German J Psychiatry. 2002;5:1–4. Available at: http://www.gjpsy.uni-goettingen. de/gjp-article-swartz.pdf. Accessed March 16, 2014.
www.ectjournal.com
Copyright © 2014 Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
97
