Circulation

Volume 92, Issue 10, 15 November 1995; Pages 3014-3024

https://doi.org/10.1161/01.CIR.92.10.3014

ARTICLE

Stereoselective Block of Cardiac Sodium Channels by Bupivacaine in Guinea Pig

Ventricular Myocytes

Carmen Valenzuela, Dirk J. Snyders, Paul B. Bennett, Juan Tamargo, and Luc M. Hondeghem

ABSTRACT: Background Bupivacaine is a potent local anesthetic widely used for prolonged local and
regional anesthesia. However, accidental intravascular injection of bupivacaine can produce severe
arrhythmias and cardiac depression. Although used clinically as a racemic mixture, S(−)-bupivacaine
appears less toxic than the R(+)-enantiomer despite at least equal potency for local anesthesia. If the
R(+)-enantiomer is more potent in blocking cardiac sodium channels, then the S(−)-enantiomer could
be used with less chance of cardiovascular toxicity. Therefore, we tested whether such
stereoselectivity existed in the bupivacaine affinity for the cardiac sodium channel.Methods and
Results The inhibitory effects on the cardiac sodium current (INa) of 10 μmol/L R(+)- and S(−)-
bupivacaine were investigated by use of the whole-cell voltage clamp technique in isolated guinea pig
ventricular myocytes. Both enantiomers produced similar but limited levels of tonic block (6% and
8%). During long depolarizations (5 seconds at 0 mV), R(+)-bupivacaine induced a significantly larger
inhibition of INa: 72±2% versus 58±3% for the S(−)-enantiomer (P<.01). Development of block was
slow, but its rate was faster for R(+)-bupivacaine [time constant, 1.84±0.16 versus 2.56±0.26 seconds
for the S(−)-enantiomer, P<.05]. The voltage dependence of the availability of the Na+ current was
shifted to more hyperpolarizing potentials compared with the control; R(+)-bupivacaine induced a
larger shift than S(−)-bupivacaine (37±2 versus 30±2 mV, P<.05). These data indicate stereoselective
interactions with the inactivated state. In addition, both enantiomers induced substantial use-
dependent block during 2.5-Hz pulse trains with medium (100-ms) and short (10-ms) depolarizations
but without stereoselective difference. A stepwise approach was used to model these experimental
results and to derive apparent affinities and rate constants. We initially assumed that bupivacaine
interacted only with the rested and inactivated states of the Na+ channel. The apparent affinities of
the inactivated state for S(−)- and R(+)-bupivacaine were 4.8 and 2.9 μmol/L, respectively. With the
derived binding and unbinding rate constants, this model reproduced the stereoselective block during
long depolarizations but failed to predict the use-dependent block induced by trains of short (10-ms)
depolarizations. To account for the observed use-dependent interactions, it was necessary to include
interactions with the activated state, which resulted in adequate reproduction of the experimental
results. The apparent affinities of the activated or open state for S(−)- and R(+)-bupivacaine were 4.3
and 3.3 μmol/L, respectively.Conclusions Both the large level of pulse-dependent block and the failure
of the pure inactivated-state block model indicate that bupivacaine interacts with the activated (or
open) state of the cardiac sodium channel in addition to its block of the inactivated state. The
bupivacaine-induced block of the inactivated state of the Na+ channel displayed stereoselectivity,
with R(+)-bupivacaine interacting faster and more potently. Both enantiomers also bind with high
affinity to the activated or open state of the channel, but this interaction did not display
stereoselectivity, although the binding to the activated or open state was faster for S(−)- than for R(+)-
bupivacaine. The higher potency of R(+)-bupivacaine to block the inactivated state of the cardiac Na+
channel may explain its higher toxicity because of the large contribution of the inactivated-state block
during the plateau phase of the cardiac action potential. These results would support the use of the
S(−)-enantiomer to reduce cardiac toxicity.

Key Words: antiarrhythmia agents ◼ electrophysiology ◼ receptors ◼ pharmacology ◼ sodium

Copyright © 1995 by American Heart Association

L
ocal anesthetics block the generation and conduction of nerve impulses by inhibiting the current
through voltage-gated Na+ channels in the nerve cell membrane.1 2 3 Nevertheless, all voltage-
gated Na+ channels exposed to sufficient concentrations of these agents will be affected. This
can explain why the accidental intravascular injection or the use of high concentrations of local
anesthetics can produce profound systemic effects, especially in the central nervous and
cardiovascular systems.4 Indeed, several local anesthetics also exhibit class I antiarrhythmic actions
on the myocardium at lower concentrations than those used for local anesthesia.5 6
Bupivacaine is a potent local anesthetic widely used for long-lasting regional local anesthesia. In
isolated cardiac tissues, bupivacaine decreases intracardiac conduction velocity and contractile force
and depresses spontaneous sinoatrial activity.7 8 In anesthetized animals, bupivacaine decreases
cardiac output, myocardial contractility, and intracardiac conduction velocity as evidenced by
increased PR and QRS durations.4 9 In both conscious and anesthetized animals, bupivacaine
induces ventricular arrhythmias that are in part preceded by a progressive widening of the QRS
complex.9 10 11 12 13 Furthermore, several studies have shown a correlation between cardiac sodium
channel (INa) inhibition and depression of the cardiovascular system.4 7 Specifically, several cases of
rapid and pronounced cardiovascular depression, cardiac arrest, or severe therapy-resistant
arrhythmias have been described after accidental intravenous administration or overdose of
bupivacaine.4 13 14 15 16 In guinea pig papillary muscles, bupivacaine depresses the maximum
upstroke velocity (Vmax) of the cardiac action potential in a use-dependent manner.17 With this
indirect index of the sodium conductance, these experiments were interpreted to indicate that
bupivacaine exhibits a low affinity for rested and activated Na+ channels and blocks only inactivated
channels with high affinity.2 5 6 18 In contrast, the interaction of bupivacaine with neuronal sodium
channels was found to involve interactions with both the activated and inactivated states.19
Bupivacaine is therapeutically used as a racemic mixture. The effects of S(−)- and R(+)-
bupivacaine on Na+ channels have been studied only with indirect indexes of the Na+ conductance. In
the isolated rabbit heart, the QRS widening and the occurrence of severe arrhythmias were much less
pronounced with S(−)- than with R(+)-bupivacaine or the racemic mixture despite similar
pharmacokinetics for both enantiomers.20 In guinea pig papillary muscles, R(+)-bupivacaine reduced
Vmax at different stimulation rates and shortened the action potential duration to a greater extent
than S(−)-bupivacaine.21 R(+)-bupivacaine also appeared to be more potent than S(−)-bupivacaine in
its ability to inhibit INa in nerve cells.22 23 Nevertheless, the potency and duration of local anesthesia
in vivo were equal or greater for S(−)- than for R(+)-bupivacaine.24 25 26 More importantly, these
studies found S(−)-bupivacaine to be less toxic than R(+)-bupivacaine, with LD50 values 30% to 40%
lower for R(+)-bupivacaine in different animal models.24 25 Thus, the S(−)-enantiomer has a
potentially beneficial profile both in efficacy and toxicity.
To better understand the mechanism and stereoselectivity for decreasing intracardiac conduction
velocity, we used the whole-cell voltage clamp technique to investigate the effects of S(−)- and R(+)-
bupivacaine on Na+ channels in isolated guinea pig ventricular myocytes. The results indicated that
both bupivacaine enantiomers displayed more complicated state-, time-, and voltage-dependent
interactions with cardiac Na+ channels than previously reported. Preliminary reports of this were
published in abstract form.27 28

METHODS

Isolation of Cardiac Myocytes, Solutions, and Drugs

Single ventricular myocytes were isolated from hearts of guinea pigs of either sex (200 to 300 g) by
enzymatic dissociation as previously described.29 The cells were stored in KB29 medium for 1 to 2
hours, and those used in the experiments reported here were rod shaped with well-defined cross
striations and quiescent in 1.8 mmol/L Ca2+. The “external” solution for electric recordings contained
(in mmol/L) NaCl 20, CsCl 110, CaCl2 1, MgCl2 2, CoCl2 3, glucose 10, and HEPES 10, adjusted to pH
7.35 (22°C) with CsOH. Cobalt, used to block Ca2+-dependent currents, reduced the Na+ current by
20% to 30%, similar to observations for a number of divalent cations.30 However, it did not detectably
modify the drug-channel interactions. The pipettes were filled with an “internal” solution containing (in
mmol/L) NaF 10, CsF 110, CsCl 20, MgCl2 2, EGTA 2, and HEPES 10, adjusted to pH 7.25 (22°C) with
CsOH. In all experiments, the concentration of bupivacaine enantiomers (kindly supplied by Astra
Hässle AB) was 10 μmol/L. The enantiomers were added to the external solution from an aqueous 10
mmol/L stock solution.

Electrophysiological Techniques
An aliquot of the cell suspension was transferred in a small-volume (0.5-mL) bath mounted on the
stage of an inverted microscope (model TMS, Nikon). The ventricular myocytes were allowed to
adhere to the bottom for 10 minutes, whereupon the chamber was perfused continuously (flow rate,
0.5 to 1.0 mL/min). The bath was cooled to 17±0.5°C by a Peltier device (Cambion/Midland Ross). INa
was measured in the whole-cell voltage clamp configuration of the patch clamp technique31 with an
Axopatch-1C patch clamp amplifier (Axon Instruments). Patch pipettes were pulled from capillary
tubes (Narishige, GD-1) on a horizontal puller (Sutter Instrument Co) and heat-polished with a
microforge (Narishige). With standard internal and external solutions, electrode DC resistances were
<1 MΩ (0.71±0.15 MΩ).
To minimize voltage control problems, small cells were selected (length, 88±20 μm; width, 20±4
μm; n=30), and cells with Na+ currents >10 nA were rejected. Cell capacitance was 73±15 pF (n=10),
measured by integration of uncompensated capacitive transients. In a separate study, cells prepared
by the same method had a mean resting potential of −84.5±3 mV (n=14) when superfused with
standard Tyrode’s solution ([K+]o=5.4 mmol/L).29 After formation of a GΩ seal (10±2 GΩ), the
electrode capacitance was compensated by an analog circuit, and the patch was disrupted with slight
additional suction. Cell capacitance and series resistance were compensated by analog circuits,
resulting in an effective series resistance of 0.4 MΩ or less. The worst-case voltage drop across the
residual series resistance was <4 mV (current <10 nA) and averaged 2.5 mV on the basis of a mean
current in the control of 5.7±0.7 nA (mean±SEM). The mean cell capacitance combined with a
maximum access resistance of 0.4 MΩ yielded a time constant of 29 μs for charging the membrane
capacitance. The observed time constants confirmed the above calculations because the
capacitance transient was usually complete within 100 μs. Together with the reduced temperature,
this resulted in a clear separation between the capacity transient and INa. The current-voltage relation
had a smooth descending limb that spanned at least 30 mV from threshold potential to the maximum
inward current. The currents obtained for h curves (availability curves for Na+ channels) could be
scaled to superimpose, which would not occur in the absence of adequate voltage control. Analog
linear leak subtraction was done on-line, and some additional digital correction was performed during
the analysis by subtraction of scaled averaged tracings from small depolarizations that resulted in no
activation of the Na+ current. Voltage steps were generated by a 12-bit digital-to-analog converter
(LabMaster, Scientific Solutions) controlled by pclamp 5.5.1 software (Axon Instruments). Voltage
clamp protocols are illustrated in the insets of Figs 1 through 5. Cells were maintained at a holding
potential of −140 mV between pulse protocols to allow full recovery from inactivation. Currents were
filtered at 5 kHz (−3 dB; four-pole Bessel filter) and sampled at 20 kHz with a 12-bit analog-to-digital
converter.

Data Analysis
Analysis of the experimental records was performed off-line with custom software, written in
fortran. A nonlinear least-squares error algorithm (modified Gauss-Newton) was used to fit
exponential functions to experimental data such as the time course of INa block development and
recovery kinetics. Results were displayed in linear and semilogarithmic format, together with a graph
of the residual deviations. Exponentials were of the form

\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb}

\usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \[y{=}A_{1}
{\times}exp(\mathit{B}_{1}{\times}t){+}A_{2}{\times}exp(\mathit{B}_{2}{\times}t){+}C\]
\end{document}

where A is amplitude, B is the rate constant, and C is baseline. Goodness of fit and the required
number of exponential components were judged by comparing χ2 values statistically (F test) and by
inspection for systematic, nonrandom deviations in the difference plot. Boltzmann equations were of
the form

\batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb}

\usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \
[y{=}1/{[}1{+}exp(E{-}E_{h})/k{]}\] \end{document}
 where E is the membrane potential, Eh is the membrane potential of midpoint of curve, and k is the
slope factor.

Modeling
To estimate the apparent rate constants for the drug-channel interactions, we used a state diagram
for the sodium channel consisting of three primary states: rested (R), activated (A), and inactivated
(I). Although each of these states may lump a set of closely related substates, the simplification
represents the widely used minimal model to capture the essence of sodium channel gating. To add
drug-channel interactions, it is assumed that these conformational changes affect the channel-
associated receptor, resulting in the corresponding RD, AD, and ID states with state-specific
association and dissociation rate constants k and l, respectively.1 2 After transient opening, the
sodium channel enters the inactivated state during sustained depolarization. The rate constants ki
and li for this continuously available receptor were derived from the rate of block development,
ki×[D]+li, and the steady state level of block, 1/(1+li/ki×[D]). To estimate the rate constants for the
briefly available activated state, we used a modified piecewise exponential approach as a
mathematical tool.32 33 In this approach, the apparent rate of block development during a train is
weighted by the rates and the available time for the interaction with each state. The time spent in the
activated state was set at 1 ms,32 34 with the remainder of the 10- or 100-ms step available for I-ID
interactions. The validity of this approach was discussed elsewhere.32 35 The formulas used were
similar to Equations 1 through 16 reported by Crumb and Clarkson.35

Statistics
Data are expressed as mean±SEM. Statistical significance (P<.05) of difference between two means
was judged with a paired or unpaired Student’s t test as appropriate.

RESULTS

Tonic Block Induced by R(+)- and S(−)-Bupivacaine

The inhibitory effects of bupivacaine enantiomers on cardiac INa can be divided into a tonic
component and a use-dependent component, similar to the action of other local anesthetic agents.1 2
6 18 Depolarization to −20 mV, after a 1-minute rest period at the holding potential of −120 mV,

resulted in a small reduction in INa amplitude compared with the control after exposure to 10 μmol/L
of either enantiomer. This tonic component of block was similar and averaged 8.1±2.3% (n=8) and
6.3±1.2% (n=8) for S(−)- and R(+)-bupivacaine, respectively. Hyperpolarization to −140 and −160 mV
further attenuated and depolarization to −100 mV augmented the tonic block, thus suggesting a very
low affinity of either enantiomer for the rested state of the Na+ channel and/or possible use-
dependent unblocking.36 37 Thus, at 10 μmol/L, tonic block was limited and did not display
stereoselectivity. The former is an important requirement to enable analysis of use-dependent block
without substantial contamination of tonic block.

Time Course of Block Development During Depolarization

In guinea pig papillary muscle, racemic bupivacaine appears to preferentially block the inactivated
state of the INa.17 To test this under voltage clamp conditions and to determine whether bupivacaine
enantiomers exhibit a stereoselective affinity for the inactivated state of the Na+ channels, we studied
the extent and time course of block development during a depolarizing pulse using the double-pulse
protocol shown in the inset of Fig 1. The holding potential was maintained at −140 mV, and
conditioning clamp pulses of variable duration (from 50 to 5000 ms) to 0 mV were imposed. The level
of block produced by this conditioning pulse was then assessed from the reduction of the sodium
current during a 30-ms test pulse to −20 mV applied after a fixed recovery interval (500 ms). The 500-
ms interval allowed full recovery from inactivation of drug-free channels but was short enough to
allow only a minimal recovery from channel block (see the next section). This sequence was repeated
at 40-second intervals at a membrane potential of −140 mV to avoid buildup of use-dependent
effects. Fig 1 shows the time course of block development in the presence of 10 μmol/L S(−)-
bupivacaine (Fig 1A) and R(+)-bupivacaine (Fig 1B). In the absence of drug, conditioning pulses of up
to 5-second duration did not affect the amplitude of the current during the test pulse. In the presence
of either S(−)- or R(+)-bupivacaine, INa amplitude decreased progressively as the duration of the
conditioning pulse was increased. After 5-second depolarizing pulses to 0 mV, the degree of INa
inhibition induced by R(+)-bupivacaine was significantly higher than that produced by S(−)-
bupivacaine (Table 1). The time course of development of block was described well by a single
exponential function with a time constant that was significantly slower for the S(−)-enantiomer (Table
1). These results indicated that block of inactivated Na+ channels was stereoselective, with R(+)-
bupivacaine inducing faster and more extensive inhibition consistent with a higher affinity for the
inactivated state of the Na+ channel. Nevertheless, because of the need to use the test pulse
sequence to assess the level of block, additional possible explanations for this enantiomeric
difference in potency must be considered: compared with S(−)-bupivacaine, R(+)-bupivacaine might
(1) exhibit slower recovery from block, (2) induce more of the pulse-dependent block (ie,
stereoselective activated-state block), (3) shift the Na+ channel availability curve toward more
negative potentials, and/or (4) display less open channel unblocking. These additional possible
mechanisms underlying enantiomeric differences in potency were examined with the appropriate
pulse protocols as discussed below.

Kinetics of Recovery from Block

If S(−)-bupivacaine exhibited faster recovery kinetics than R(+)-bupivacaine, then the time between
the conditioning pulse and the test pulse shown in Fig 1 (500 ms) could be long enough to recover
more drug-bound Na+ channels and explain the differences in potency between S(−)- and R(+)-
bupivacaine. To test this possibility, the time course of recovery from Na+ channel block was
characterized with the double-pulse protocol shown in the inset of Fig 2. Steady state Na+ channel
block was achieved with a single 10-second conditioning pulse to 0 mV, and the time course of
recovery was determined with a standard test pulse (30 ms to −20 mV) at different coupling intervals,
ranging from 50 to 10 000 ms. The current for each test pulse was normalized to matching control
and plotted as a function of the recovery time, as shown in Fig 2. Under control conditions, the
reactivation of INa at −120 mV was a fast monoexponential process with a time constant (τre) of
20.1±1.2 ms (n=14). In the presence of either bupivacaine enantiomer, recovery from INa block was
markedly slowed and exhibited a fast and a slow phase. Table 2 shows the amplitudes and time
constants for each enantiomer. The fast and the slow time courses of recovery (τf and τs,
respectively) were similar for both enantiomers, but more importantly, most of the recovery
proceeded with the slow time constant (see Table 2). This marked slow phase of recovery in the
presence of drug was assumed to reflect the time course with which drug unbinds from closed
channels (RD or ID). These data suggest similar recovery kinetics from closed channels in the
presence of both S(−)- and R(+)-bupivacaine. The contribution of the slow amplitude component (As)
was statistically larger for R(+)-bupivacaine (P<.05), consistent with the larger amount of block
induced by R(+)-bupivacaine.

Use-Dependent Block
The enantiomeric differences in potency could also be due to more pronounced use-dependent block
by the R(+)-enantiomer. Therefore, we determined use-dependent INa block induced by either
enantiomer with 2.5-Hz trains of medium-duration (100-ms) and short-duration (10-ms) depolarizing
pulses from different holding potentials to a test potential of −20 mV. Fig 3 shows the use-dependent
inhibition of INa after application of trains of sixteen 100-ms pulses from different holding potentials
(−160, −140, and −120 mV) in the absence and presence of both enantiomers. In the absence of drug,
there was no noticeable effect of repetitive pulsing on the INa magnitude regardless of the holding
potential. In the presence of S(−)-bupivacaine (Fig 3A) and R(+)-bupivacaine (Fig 3B), there was a
small reduction of the current during the first pulse (reflecting tonic block), followed by a gradual
decrease during the train of depolarizing pulses until a steady state was reached. This steady state
use-dependent block was attenuated with more hyperpolarized holding potentials. However, the
degree of use-dependent block did not differ significantly between enantiomers, regardless of the
holding potentials tested (Table 3), although the rate at which this steady state level was achieved
was somewhat slower for the R(+)-enantiomer. More importantly, during sixteen 100-ms pulses, the
Na+ channels spent an aggregate time of 1.6 seconds at the depolarized potential, which promotes
inactivated-state block. Thus, the difference in time course could derive at least in part from
accumulated block of the inactivated state. Therefore, the level and onset kinetics of block were
determined by use of identical pulse trains with a 10-fold shorter depolarization time (10 ms). Fig 4
shows that such pulse trains produced a marked cumulative decrease in INa in the presence of 10
μmol/L S(−)-bupivacaine, which reached steady state within 16 beats. In six cells, the average steady
state use-dependent block was 19±2% and 33.0±4% with holding potentials of −140 and −120 mV,
respectively. In another six cells, the steady state INa block produced by R(+)-bupivacaine under
similar experimental conditions averaged 25±4% and 34±5.0% at holding potentials of −140 and −120
mV, respectively. Similar to the results with the 100-ms depolarizations, these levels of block were not
significantly different between both enantiomers. The time course of block during the pulse train from
a holding potential of −120 mV was fitted by a monoexponential function, which yielded rate
constants in the presence of S(−)- or R(+)-bupivacaine of 0.20±0.05 and 0.16±0.02 pulse−1,
respectively (P=NS). Thus, with a similar number of activations (n=16) but a reduction of total
depolarized time (160 versus 1600 ms), the level of use-dependent block was still significant, albeit
reduced and without stereoselectivity.

Voltage-Dependence of the Availability of the Na+ Current

The inhibition of Na+ channels by local anesthetics is a voltage-dependent process,1 6 and many local
anesthetics cause a shift of the inactivation curve. To test whether the enantiomers induced shifts of
the inactivation curve of different magnitudes, we determined the availability curves for INa using the
standard two-pulse protocol (Fig 5). A 10-second conditioning pulse from a holding potential of −140
to 0 mV was followed by a 500-ms depolarization to different potentials (between −180 and −60 mV)
and then by a 30-ms test pulse to −20 mV. This experimental protocol allowed us to evaluate not only
the voltage dependence of availability of the Na+ channel but also (indirectly) the degree of activation
unblocking37 induced by each enantiomer (see “Discussion”). The experimental values were well fit
with a Boltzmann equation (see “Methods”). Under control conditions, the mean values for midpoint
(E1/2) and slope factor (k) averaged −85±3 and 8.9±0.4 mV, respectively (n=12). S(−)-bupivacaine
reduced the maximum available INa by 28±4% and significantly altered the voltage dependence of Na+
channel availability, as illustrated in Fig 5A. The midpoint was shifted by 30±2 mV toward more
negative potentials (to −115±4 mV, n=6, P<.001), and k was increased to 15±1 mV (P<.001), indicating
a more shallow voltage dependence in the presence of S(−)-bupivacaine. Fig 5B shows that R(+)-
bupivacaine decreased the maximum available INa by 47±5%, shifted E1/2 by 37±2 mV (to −126±4 mV,
n=6, P<.001) in the hyperpolarizing direction, and increased the slope factor of the inactivation curve
to 16.0±0.8 mV (n=6, P<.001).

Mathematical Modeling of Bupivacaine Block and Stereoselectivity

The main experimental observations were low tonic block, substantial time-dependent block during
sustained depolarization, and intermediate levels of block with pulse trains. Rather than assuming a
specific model and globally fitting all parameters, we used a stepwise approach to examine whether
simpler models would be appropriate. Because most of the differences between the bupivacaine
enantiomers seem to be related to their ability to block inactivated Na+ channels, we first chose a
model that involved only the nonconducting (closed) states of the Na+ channel: rested and
inactivated (model 1, Fig 6A).
To account for low tonic block, the affinity of the closed state at negative potentials (rested state,
R, Fig 6) needs to be low, implying that the association rate kR×[D] should be substantially smaller
than the dissociation rate lR. From the recovery time constants at −120 mV, we obtained a recovery
rate {λR (=kR×[D]+lR)} of 0.23 and 0.19 seconds−1 for S(−)- and R(+)-enantiomers, respectively.
Because of the low affinity of the rested state and because we wanted to explore the basis for the
stereoselectivity of the high affinity states, we made the simplifying assumption that the association
rate is essentially nonexisting (kR=0), in which case lR=λR. During a depolarizing step, the sodium
channel briefly opens and then enters the inactivated state for the remainder of the sustained
depolarization. Because this state is continuously available, we can estimate the rate constants for
association (kI) and dissociation (lI) from the steady state level of block and the time constant of
block development (Table 4). This model reproduced the observed development of block for both
enantiomers when using a double-pulse protocol (Fig 6B and 6C) and the recovery from block at −120
mV. However, this model predicted hardly any block during a train of depolarizing pulses, especially
with short depolarizations (10 ms), which limit the time spent in the inactivated state. Thus, this
closed-state block model failed to reproduce with a single set of rate constants the amount and
kinetics of block induced by both pulse trains and single maintained depolarization.
 To maintain slow block during depolarization but increase the amount of block during repetitive
pulses, it was necessary to add interactions with the short-lived activated or open state, shown as
model 2 in Fig 6D. To maintain the closed-state interactions, we kept the set of rate constants of
model 1 and calculated only the two rate constants required for the activated-state interaction (see
“Methods”). Fig 6E and 6F shows that simulations with this expanded model satisfactorily
reproduced levels and the onset kinetics of block during pulse trains.
We also tested whether a full mathematical model based on the modulated receptor hypothesis
would reproduce these findings. In the voltage clamp version of this model,37 the mean open time
differs from the 1-ms value used in the stepwise approach. Therefore, the apparent interaction rate
constants for the open state (ka, la) had to be scaled down by a factor of three. However, this did not
modify the apparent affinity for the open state. All other rate constants were left unchanged. The
main features in Fig 6 were similarly reproduced by the full model. In addition, recovery from block
proceeded biexponentially, as observed in Fig 2. The fast component corresponded largely to
recovery from inactivation of the drug-free channels.

DISCUSSION
The main results of the present study are that (1) the bupivacaine-induced cardiac Na+ channel block
displayed stereoselectivity, with R(+)-bupivacaine being the more potent enantiomer; (2) the inhibitory
effects of bupivacaine on INa involve binding to the inactivated state of the Na+ channel and to the
activated or open state; (3) the stereoselectivity resided in the interaction with the inactivated state;
(4) both enantiomers induced the same low degree of tonic block, which was not stereoselective; and
(5) recovery from block at negative potentials was the same for both enantiomers, suggesting that
dissociation from rested or inactivated blocked states was not stereoselective.

Interaction With Nonconducting States

Determination of whether bupivacaine enantiomers block Na+ channels differently in the rested,
activated, and inactivated channel states may result in a better understanding of their
cardiodepressant effects. In the present study, both enantiomers of bupivacaine induced very limited
tonic block measured as the reduction of INa amplitude during the first pulse after a long rest period.
The amount of block measured from the reduction of peak current reflects both the rested-state block
and the activated (open)-state interaction between the Na+ channel and the drug that develops during
the depolarizing step before reaching peak current.18 Although tonic block is not necessarily an
accurate measure of drug binding to rested Na+ channels, the observed low level indicates that both
enantiomers must exhibit a low affinity for the rested state of the Na+ channel. Moreover, the
interaction of bupivacaine enantiomers with the rested state of the Na+ channel was not
stereoselective. Similar results were described previously for the enantiomers of another local
anesthetic, RAC-109.38
Previous studies in guinea pig papillary muscle that used the Vmax technique as an index of Na+
conductance suggested that racemic bupivacaine preferentially binds to the inactivated state of
cardiac Na+ channels17 and that the affinity of the inactivated channels was higher for the R(+)-
enantiomer,21 whereas no significant activated-state block could be resolved.17 This contrasts with
the present results showing significant block of the activated state. This should not come as a
surprise because several voltage clamp studies on cardiac Na+ channels provided evidence that use-
dependent block induced by tertiary amine local anesthetics results from drug binding to multiple
distinct channel states (eg, activated and inactivated).38 39 Some of the difference may derive from
the difference in experimental preparation (Vmax in multicellular cardiac preparations versus INa in
isolated cardiac myocytes). In addition, the time course for inactivated-state block is faster at 37°C
(τ=0.6 versus 1.8 seconds at 17°C), reflecting the higher association rate constant [550 000 (mol/L)−1
· s−1 at 37°C versus 30 000 (mol/L)−1 · s−1 at 17°C, Table 1].17 Because the stereoselective effects
are related to block of inactivated channels, the clinical implications discussed below should be even
more important at 37°C, at which the affinity of the R(+)-enantiomer for inactivated channels was also
found to be higher.21
The double-pulse protocol (Fig 1) in which the duration of the conditioning pulse was progressively
increased is a useful way to demonstrate interaction with the inactivated state.18 Relative INa
decreased progressively as the duration of the conditioning was lengthened in the presence of either
bupivacaine enantiomer. The simplest explanation for this component of block is that it reflects the
slow binding of the drug to the inactivated channels because virtually all Na+ channels are inactivated
within <100 ms at −20 mV. Although time constants of 1.8 to 2.5 seconds are slow compared with the
development of inactivation, these values are still fast compared with those for slow inactivation at
the same low temperature.34 Indeed, we observed no significant development of slow inactivation
within 5 to 10 seconds (Fig 1), consistent with earlier observations under similar conditions.37
The stereoselectivity of the interaction with the inactivated state on depolarization was reflected in
two ways. First, the level of INa inhibition induced by R(+)-bupivacaine was significantly higher than
that produced by S(−)-bupivacaine (Fig 1 and Table 1). This indicates a higher affinity of R(+)-
bupivacaine for the inactivated state compared with S(−)-bupivacaine. Second, the time course of
block development was faster for R(+)-bupivacaine, with a time constant of 1.8 seconds compared
with 2.5 seconds for the S(−)-enantiomer. The combination of higher affinity and faster binding
kinetics suggested that the association rate constant of R(+)-bupivacaine to this state of the Na+
channel was faster than that of S(−)-bupivacaine. An analysis based on a bimolecular reaction with
the receptor associated with the inactivated state supports this qualitative assessment (Table 4).
Thus, the stereoselectivity of the bupivacaine enantiomers in their interaction with the inactivated
state can be explained by differences in the apparent association rate constant for each enantiomer.
Although both bupivacaine enantiomers shifted the Na+ channel availability curve in
hyperpolarizing direction, this effect was more pronounced for R(+)-bupivacaine. According to Hille1
such shift can be attributed to a higher affinity of a drug for the inactivated state of the Na+ channel.
As such, the larger shift is consistent with the higher potency of R(+)-bupivacaine to block the
inactivated state of the Na+ channel.
Recovery from block did not display stereoselectivity. Although the fraction of channels recovering
with the slow time constant was larger for R(+)-bupivacaine, this can be explained on the basis of the
larger amount of block induced by this isomer. Indeed, the fractional contribution of the slowly
recovering component was proportional to the amount of block induced (compare Tables 1 and 2).
The slow time constants observed during recovery from block at negative potentials (−120 to −160
mV) were similar for both enantiomers. Because we have little evidence for any significant affinity of
the rested state, this time constant should reflect primarily the rate constant for dissociation from
closed blocked states. Interestingly, these values were similar to those for the dissociation from the
inactivated state at −20 mV. These results indicate that both enantiomers exhibit similar kinetics of
unbinding from the closed states of the Na+ channel (ID→I→R or ID→RD→R).

Use-Dependent Block
The use-dependent block produced by S(−)- and R(+)-bupivacaine by a train of short (10-ms)
depolarizing pulses suggests that the enantiomers also interact with the activated state of the
cardiac Na+ channels. In fact, if S(−)- and R(+)-bupivacaine bound only to the inactivated state of the
Na+ channel with the time constants shown in Table 1, one would expect only a slight decrease of the
recorded INa during the application of a train of short (10-ms) pulses that activates the Na+ channels
repeatedly without allowing much time for inactivated state block. Moreover, the magnitude of use-
dependent block induced by both enantiomers was similar, which suggests that, in contrast to the
interaction of bupivacaine enantiomers with the inactivated state, their interaction with the activated
state of Na+ channels does not display stereoselectivity, or at least it is not detectable with this
approach.
An apparent discrepancy is the similar levels of block during the pulse trains. R(+)-bupivacaine
would be expected to induce more accumulated block unless offset by some property of S(−)-
bupivacaine. A likely explanation is offered by the difference in the rate constants for the interaction
with the activated state (Table 4). Despite a similar predicted level of A-AD block, the kinetics are
faster for S(−)-bupivacaine (τ=10 ms) compared with R(+)-bupivacaine (τ=12.5 ms). This results in a
difference in the amount of activated-state block induced per pulse (6.7% versus 5.7% per pulse). This
15% difference accumulates during the pulse train and apparently offsets the higher affinity of R(+)-
bupivacaine for the inactivated state. This kinetic difference would also contribute to the slower
accumulation of block with R(+)-bupivacaine during the pulse train.

Mathematical Modeling
Although model 1 did not reproduce all observations, it provided an adequate description of the
interactions with the closed states. The rate constants for dissociation from the inactivated state
were similar for both enantiomers; moreover, they were similar to the rate constants for dissociation
from the rested state. Thus, bupivacaine appears to dissociate as slowly from the nonconducting
channels whether they are rested or inactivated. This may indicate a similar route of escape, probably
through a hydrophobic pathway.1
Whereas the resulting (apparent) dissociation rate constants were similar for both enantiomers,
this model predicted different association rate constants, suggesting that the S(−)-enantiomer is less
likely to approach the receptor in the most favorable conformation for binding. The difference in
apparent affinity for the inactivated state is similar to the differences observed between RAC-109
stereoisomers and to those inferred from the Vmax study in papillary muscle.21 38 In contrast, the rate
constants for the interaction with the activated state did not display marked stereoselectivity, with
apparent affinities of 4.3 and 3.3 μmol/L for S(−)- and R(+)-bupivacaine, respectively. From these rate
constants, the time constant of block of the open (activated) state would be around 10 ms (for 10
μmol/L), which is slow compared with the Na+ channel kinetics. Presumably, this would preclude an
accelerated decline of current in the presence of drug because the channels inactivate faster than
they become blocked in the open state. The reduction of the amount of block with hyperpolarization
was similar for both enantiomers, which can be explained by the fact that this most likely reflects use-
dependent unblocking. For agents with slow closed-state recovery kinetics, this phenomenon has
been shown to result from activated-state unblock.36 37 The lack of stereoselectivity of the activated-
state interaction is therefore consistent with such process. Thus, both the large level of
experimentally observed pulse-dependent block and the failure of the inactivated-state block model
indicate that bupivacaine also interacts with the activated or open state of the cardiac sodium
channel, in addition to its block of the inactivated state.
The voltage dependence of channel availability (inactivation curve) in the presence of drug cannot
be considered a true steady state measure because of the kinetics of recovery of normal and blocked
channels involved. However, microscopic reversibility would predict a negative shift of the
inactivation curve. Considering the loop R→I→ID→RD→R, the voltage shift would be ΔV=k×ln(KR/KI),
where k represents the slope factor.40 The observed hyperpolarizing shifts of about 30 mV require
that the rested state would have at least a 20-fold lower affinity. This is fully in accordance with the
low “tonic” block observed in these experiments. Moreover, the difference between both enantiomers
would be ΔΔV=k×ln(KI,S/KI,R). A 7-mV shift would correspond to a twofold difference in inactivated-
state affinity between the enantiomers (assuming identical rested-state affinity). This is also
consistent with the estimated affinities of 4.8 and 2.9 μmol/L.

Effects of S(−)- and R(+)-Bupivacaine on the Action Potential Characteristics

It has been demonstrated that bupivacaine decreased Vmax and shortened the action potential
duration in guinea pig and rat ventricles.17 41 Furthermore, in guinea pig papillary muscles, S(−)-
bupivacaine affected Vmax and action potential duration much less than R(+)-bupivacaine.21 The
effects of bupivacaine racemate and its enantiomers on phase 0 characteristics of the cardiac action
potential can be attributed to their potent inhibitory effects on INa. This effect can also explain the
decrease in intracardiac conduction velocity, which may be primarily responsible for the
cardiodepressant effects of bupivacaine.4 7

Clinical Implications
The accidental intravascular injection of bupivacaine is associated with the development of
cardiovascular collapse and ventricular arrhythmias, which were accompanied by PR prolongation
and widening of the QRS complex, indicative of slowed conduction.9 11 14 16 After intravenous
injection, the whole-blood concentration of bupivacaine ranged from 3 to 11 μg/mL in conditions
where cardiac conduction was seriously depressed and difficult to reverse.14 With a blood-to-plasma
concentration ratio of 0.73 assumed,42 this should correspond to a plasma concentration of 4 to 15
μg/mL bupivacaine. Because bupivacaine is 66% to 88% bound to plasma proteins over this
concentration range,43 the free concentration should be approximately 1.5 to 13 μmol/L. The
concentration used in this study (10 μmol/L or 3 μg/mL) is therefore in the clinically relevant range for
toxicity. These concentrations produced clinical toxicity in sheep11 and marked depression of Vmax in
guinea pig ventricular muscle.17
Because reentrant arrhythmias are favored by slow conduction,44 the mechanism of bupivacaine-
induced cardiac arrhythmias may result from its inhibitory effect on INa. In this context, the present
results would suggest that S(−)-bupivacaine may be less toxic because of its lower potency to block
cardiac sodium channels compared with R(+)-bupivacaine. Hypoxia, acidosis, and hyperkalemia
commonly occur during local anesthetic toxicity45 46 and greatly potentiate the cardiotoxicity of
bupivacaine.4 12 These conditions all result in partial depolarization, which increases the fraction of
Na+ channels in the inactivated state during diastole. Because of its lower affinity for the inactivated
state, this also would represent a potential benefit for the S(−)-enantiomer over racemic bupivacaine.
Previous studies24 25 26 demonstrated that the potency and duration of local anesthesia in vivo
were the same as or greater for S(−)- than for R(+)-bupivacaine. Nevertheless, the mechanism of
block of neuronal sodium channels is similar to that observed here. Both activated- and inactivated-
state interactions were observed,19 with interaction rate constants for R(+)-bupivacaine of ki=44 000
(mol/L)−1 · s−1 and ka=25 (mol/L)−1 · s−1, which are very similar to the values in Table 4. Other
observations also indicate that neuronal sodium channels display a stereoselectivity similar to that of
the cardiac channels, with R(+)-bupivacaine being the more potent enantiomer.23 However, the local
tissue concentration during local anesthesia may be high enough (to ensure total block) that the
intrinsic enantiomeric difference would be lost. In addition, block of other ion channels may
contribute to the anesthetic effect because bupivacaine has also been shown to block potassium
channels.22 41 47 48 Importantly, S(−)-bupivacaine appears to be less toxic than R(+)-bupivacaine: the
LD50 was about 30% to 40% lower for R(+)-bupivacaine in mice, rats, and rabbits.20 24 25 The results
from our study demonstrate that the EC50 for block of the inactivated state of the Na+ channel is 39%
lower for R(+)-bupivacaine compared with the S(−)-enantiomer. Because part of the cardiotoxicity
appears to be related to block of cardiac Na+ channels, our results provide support for the use of
S(−)-bupivacaine in local anesthesia because it is highly efficacious for local anesthesia but with a
lower potential for adverse effects on inadvertent access to the cardiovascular system.

Figure 1. Plots showing time course of block induction by bupivacaine enantiomers. A, Sodium currents obtained
at −20 mV under control conditions and in the presence of 10 μmol/L S(−)-bupivacaine. B, Time course of block
development in the presence of S(−)-bupivacaine obtained from a typical experiment. C, Original records
obtained under control conditions and in the presence of 10 μmol/L R(+)-bupivacaine. D, Time course of block
development observed in the presence of R(+)-bupivacaine. Cells were held at a holding potential (Vh) of −140
mV. A conditioning pulse to 0 mV of variable duration was applied and cardiac sodium current (INa) was recorded
in a test pulse (*) to −20 mV after 500 ms at −120 mV to allow recovery of drug-free channels. Under control
conditions, the conditioning prepulses had little effect. In the presence of either bupivacaine enantiomer, a
progressive increase in INa inhibition was observed as a function of prepulse duration. Note that the R(+)-
enantiomer induced more block with a faster time course.

Figure 2. Plots showing recovery from block induced by S(−)-bupivacaine (A) and R(+)-bupivacaine (B). Under
control conditions, recovery from inactivation could be fitted with a single exponential with a time constant (τ) of
20 ms. In the presence of S(−)- or R(+)-bupivacaine, the recovery process displayed biexponential time course,
dominated by a slow component with time constants of 4.4 and 5.4 seconds, respectively, although the
contribution of this component to the total reactivation was different (see Table 2 for average values).

Figure 3. Plots showing use-dependent effects of S(−)-bupivacaine (A) and R(+)-bupivacaine (B) during 2.5-Hz
trains of sixteen 100-ms pulses from different holding potentials (Vh, −160, −140, and −120 mV). In the absence
of drug, the cardiac sodium current (INa) amplitude remained unchanged. In the presence of either enantiomer,
the INa amplitude declined exponentially to a steady state level. This level of drug-induced inhibition depended on
the holding potential and was attenuated at more hyperpolarized levels. The solid line represents the best
monoexponential fit for the experimental data. The level of block induced by both bupivacaine enantiomers at the
end of the train was similar, although the kinetics of development of block was different (see Table 3).
Figure 4. Plots showing use-dependent effects of both bupivacaine enantiomers during 2.5-Hz trains of 10-ms
depolarizations from different holding potentials (Vh,−140 and −120 mV). A, Effects of S(−)-bupivacaine; B,
effects of R(+)-bupivacaine. In the absence of drug, the pulse trains did not affect the INa amplitude. In the
presence of both S(−)- and R(+)-bupivacaine, the INa amplitude decayed exponentially to a steady state level. The
latter depended on the holding potential (larger with the more depolarized level) but was similar for both
bupivacaine enantiomers.

Figure 5. Plots showing voltage dependence of the availability of the Na+ current in the presence of S(−)- and
R(+)-bupivacaine. A, Sodium currents recorded in the absence and presence of S(−)-bupivacaine. B, Availability
curve for INa obtained with the pulse protocol shown in the inset. A 10-second prepulse to 0 mV was followed by
500-ms conditioning step membrane potentials between −180 and −60 mV in 10-mV steps. Current was
recorded during the subsequent 30-ms test pulse to −20 mV. This pulse protocol was applied every 40 seconds.
The solid lines represent the best-fitting Boltzmann equation (see “Methods”), with the following parameters in
this experiment for control and S(−)-bupivacaine, respectively: midpoint, Eh=−79.4 and −102.1 mV; slope factor,
k=7.3 and 14.5 mV. The dotted line represents the curve in drug scaled to control, illustrating both the shift and
reduced slope. C, Sodium currents in the absence and in the presence of R(+)-bupivacaine. D, Availability curve
for INa obtained with the same pulse protocol as described in B in the absence and presence of R(+)-bupivacaine.
Parameters for the fitted curves were Eh=−82.5 and −124.2 mV and k=8.2 and 16.4 mV for control and R(+)-
bupivacaine, respectively.
Figure 6. Plots showing stepwise mathematical modeling of sodium channel block by bupivacaine enantiomers.
A through C, Model 1 incorporated interactions with rested and inactivated states of the cardiac Na+ channel (A).
The simulated time course of block in the presence of S(−)- and R(+)-bupivacaine is shown in B and C,
respectively. The time course of block during single long depolarizations (solid line) corresponded well with the
experimental results (Fig 1B and 1D). However, during 10-ms pulse trains (○), the predicted time course deviated
substantially from the experimental observations, as indicated by the dotted line. D through F, To account for the
use-dependent block, the model was expanded (model 2, D) to incorporate interaction of bupivacaine
enantiomers with the activated state of the Na+ channel. E and F show results for S(−)- and R(+)-bupivacaine,
respectively. The dashed lines in E and F represent the average experimental data for 10- and 100-ms pulse
trains. Contrary to model 1, the predicted time course of block, indicated by the open symbols and the solid
connecting lines, corresponded reasonably well with the experimental data for both pulse durations. This model
reproduced block induced by a single prepulse as well as model 1 (omitted for clarity).

Table 1. Percentage of Na+ Channel Block After 5-Second Depolarizing Pulses and Time Course of Block
Development in the Presence of S(−)- and R(+)-Bupivacaine (Table view)

  Percentage of Block, % Time Constant, s

S(−)-bupivacaine 57.8±3.0 2.56 ±0.26
R(+)-bupivacaine 72.2±2.12 1.84 ±0.161

Data are mean±SEM of eight experiments.

1 P<.05,
2 P<.01. Comparisons were performed between data from S(−)- and R(+)-bupivacaine.
Table 2. Parameters Defining the Time Course of Recovery From Na+ Channel Block (Table view)

  S(−)-Bupivacaine R(+)-Bupivacaine
Af 0.25±0.02 0.18±0.021
τf, s 0.14±0.01 0.13±0.01
As 0.58±0.02 0.69±0.041
τs, s 4.40±0.90 5.40±0.80
As/(Af+As) 0.70 ±0.02 0.79±0.031

Af and As indicate the amplitudes of the fast and the slow components of the recovery process, respectively; τf and τs, the
fast and the slow time constants of recovery, respectively. Data are mean±SEM from seven experiments.
1 P<.05. Comparisons were performed between data from S(−)- and R(+)-bupivacaine.

Table 3. Extent and Rate of Na+ Channel Block Induced by S(−)- and R(+)-Bupivacaine After Application of Trains
of Sixteen 100-ms Depolarizing Pulses (Table view)

Holding potential, mV Block, % Onset Rate, pulse−1

S(−) R(+) S(−) R(+)
−120 58±5 55±5 0.25±0.03 0.15 ±0.012
−140 43±8 39±4 0.31±0.04 0.19±0.031
−160 32±6 33±4 0.37±0.05 0.23±0.04

Data are mean±SEM from five experiments.

1 P<.05;
2 P<.01. Comparisons were performed between data from S(−)-bupivacaine [S(−)] and R(+)-bupivacaine [R(+)].

Table 4. Calculated Association and Dissociation Rate Constants for the Rested, Inactivated, and Activated
States for Models 1 and 2 (Table view)

  S(−)-Bupivacaine R(+)-Bupivacaine
ki 26 300 42 000
li 0.13 0.12
kr 0 0
lr 0.23 0.19
ka 7×106 6×106
la 30 20

k indicates the association rate constant [(mol/L)−1 · s−1]; l, dissociation rate constant (seconds−1); and subscripts R, I, and A
refer to rested, inactivated, and activated states, respectively. Model 1 incorporated the rate constants for rested and
inactivated states; model 2 is an expanded model incorporating activated-state interactions.

ARTICLE INFORMATION
Received July 5, 1994; revision received May 16, 1995; accepted July 5, 1995.

Correspondence
Correspondence to Carmen Valenzuela, PhD, Institute of Pharmacology and Toxicology, CSIC, School of
Medicine, Universidad Complutense, 28040 Madrid, Spain. E-mail carmenva@eucmvx.sim.ucm.es.

Affiliations
From the Institute of Pharmacology and Toxicology, CSIC, School of Medicine, Universidad Complutense, Madrid,
Spain (C.V., J.T.); the Departments of Pharmacology and Medicine, Vanderbilt University School of Medicine,
Nashville, Tenn (D.J.S., P.B.B.); and the Department of Pharmacology, HPC NV, Oostende, Belgium (L.M.H.).

Acknowledgments
This study was supported by CYCIT SAF92-0157, FIS 95/0318, Salud 2000, and CAM 157/92 and NIH grants HL-
46681, HL-47599, and HL-51197. We want to thank Drs E. Delpón and O. Pérez for valuable comments on the
manuscript; we also thank G. Pablo and R. Vara for their excellent technical assistance. Dr P.B. Bennett is an
Established Investigator of the American Heart Association. We thank Astra Hässle AB, Mölndal, Sweden, for
providing bupivacaine enantiomers.

REFERENCES
1. Hille B. Local anesthetics: hydrophilic and hydrophobic pathways for the drug-receptor reaction. J Gen
Physiol. 1977;69:497-575. Crossref. PubMed.
2. Hondeghem L, Katzung B. Time- and voltage-dependent interactions of antiarrhythmic drugs with cardiac
sodium channels. Biochim Biophys Acta. 1977;472:373-398. Crossref. PubMed.
3. Strichartz G, Ritchie J. The action of local anesthetics on ion channels of excitable tissues. In: Strichartz G,
ed. Local Anesthetics Handbook of Experimental Pharmacology. Heidelberg, Germany: Springer-Verlag;
1987;81:21-52.
4. Covino B. Toxicity and systemic effects of local anesthetic agents. In: Strichartz G, ed. Local Anesthetics
Handbook of Experimental Pharmacology. Heidelberg, Germany: Springer-Verlag; 1987;81:187-212.
5. Grant A, Starmer C, Strauss H. Antiarrhythmic drug action: blockade of the inward sodium current. Circ Res.
1984;55:427-439. Crossref. PubMed.
6. Tamargo J, Valenzuela C, Delpón E. New insights into the pharmacology of sodium channel blockers. Eur
Heart J. 1992;13(suppl F):2-13.
7. Block A, Covino B. Effect of local anesthetic agents on cardiac conduction and contractility. Reg Anesth.
1982;6:55-61.
8. Moller R, Covino B. Cardiac electrophysiologic effects of lidocaine and bupivacaine. Anesth Analg.
1988;67:107-114. Crossref. PubMed.
9. Liu P, Feldman H, Covino B, Giasi R, Covino BG. Acute cardiovascular toxicity of intravenous amide local
anesthetics in anesthetized ventilated dogs. Anesth Analg. 1982;61:317-322. PubMed.
10. DeJong R, Ronfeld R, DeRosa R. Cardiovascular effects of convulsant and supraconvulsant doses of amide
local anesthetics. Anesth Analg. 1982;61:3-9. PubMed.
11. Kotelko D, Shnider S, Dailey P, Brizgys R, Levinson G, Shapiro A, Koike M, Rosen M. Bupivacaine-induced
cardiac arrhythmias in sheep. Anesthesiology. 1984;60:10-18. Crossref. PubMed.
12. Rosen M, Thigpen J, Shnider S, Foutz S, Levinson G, Koike M. Bupivacaine-induced cardiotoxicity in hypoxic
and acidotic sheep. Anesth Analg. 1985;64:1089-1096. PubMed.
13. Nath S, Häggmark S, Johansson G, Reiz S. Differential depressant and electrophysiological cardiotoxicity of
local anesthetics: an experimental study with special reference to lidocaine and bupivacaine. Anesth Analg.
1986;65:1263-1270. PubMed.
14. Albright G. Cardiac arrest following regional anesthesia with etidocaine or bupivacaine. Anesthesiology.
1979;51:285-287. Crossref. PubMed.
15. Buffington C. The magnitude and duration of direct myocardial depression following intracoronary local
anesthetics: a comparison of lidocaine and bupivacaine. Anesthesiology. 1989;70:280-287. Crossref.
PubMed.
16. Prentiss J. Cardiac arrest following caudal anesthesia. Anesthesiology. 1979;50:51-53. Crossref. PubMed.
17. Clarkson C, Hondeghem L. Mechanism for bupivacaine depression of cardiac conduction: fast block of
sodium channels during the action potential with slow recovery from block during diastole. Anesthesiology.
1985;62:396-405. Crossref. PubMed.
18. Snyders D, Bennett P, Hondeghem L. Mechanisms of drug-channel interaction. In: Fozzard HA, Haber E,
Jennings RB, Katz AM, Morgan HE, eds. The Heart and Cardiovascular System. New York, NY: Raven Press,
Ltd; 1992:2165-2193.
19. Chernoff D. Kinetic analysis of phasic inhibitory of neuronal sodium currents by lidocaine and bupivacaine.
Biophys J. 1990;58:53-68. Crossref. PubMed.
20. Mazoit J, Boı̈co O, Samii K. Myocardial uptake of bupivacaine, II: pharmacokinetics and pharmacodynamics
of bupivacaine enantiomers in the isolated perfused rabbit heart. Anesth Analg. 1993;77:477-482. PubMed.
21. Vanhoutte F, Vereecke J, Verbeke N, Carmeliet E. Stereoselective effects of the enantiomers of bupivacaine
on the electrophysiological properties of the guinea-pig papillary muscle. Br J Pharmacol. 1991;103:1275-
1281. Crossref. PubMed.
22. Guo X, Castle N, Chernoff D, Strichartz G. Comparative inhibition of voltage-gated cation channels by local
anesthetics. Ann N Y Acad Sci. 1991;625:181-199. Crossref. PubMed.
23. Lee-Son S, Wang G, Concus A, Crill E, Strichartz G. Stereoselective inhibition of neuronal sodium channels by
local anesthetics: evidence for two sites of action? Anesthesiology. 1992;77:324-335. Crossref. PubMed.
24. Luduena F, Bogado E, Tullar B. Optical isomers of mepivacaine and bupivacaine. Arch Int Pharmacodyn Ther.
1972;200:359-369. PubMed.
25. Åberg G. Toxicological and local anaesthetic effects of optically active isomers of two local anaesthetic
compounds. Acta Pharmacol Toxicol. 1972;31:273-286. Crossref. PubMed.
26. Aps C, Reynolds F. An intradermal study of the local anaesthetic and vascular effects of the isomers of
bupivacaine. Br J Clin Pharmacol. 1978;6:63-68. Crossref. PubMed.
27. Valenzuela C, Bennett P, Hondeghem L. Stereospecific block of cardiac Na+ channels by bupivacaine.
Biophys J. 1990;57:108a. Abstract.
28. Valenzuela C, Delpón E, Tamargo J, Hondeghem L, Bennett P, Snyders D. State-dependent interactions of
bupivacaine stereoisomers with cardiac sodium channels. Biophys J. 1993;64:A90. Abstract.
29. Valenzuela C, Sánchez-Chapula J, Delpón E, Elizalde A, Pérez O, Tamargo J. Imipramine blocks rapidly
activating and delays slowly activating K+ current activation in guinea pig ventricular myocytes. Circ Res.
1994;74:687-699. Crossref. PubMed.
30. Sheets M, Hanck D, Fozzard H. Divalent block of sodium current in single canine cardiac Purkinje cells.
Biophys J. 1988;53:535a. Abstract.
31. Hamill O, Marty A, Neher E, Sakmann B, Sigworth F. Improved patch-clamp techniques for high resolution
current recording from cells and cell-free membrane patches. Pflugers Arch. 1981;391:85-100. Crossref.
PubMed.
32. Starmer C, Grant A. Phasic ion channel blockade: a kinetic model and parameter estimation procedure. Mol
Pharmacol. 1985;28:348-356. PubMed.
33. Kojima M, Ban T. Nicorandil shortens action potential duration and antagonizes the reduction of Vmax by
 lidocaine but not by disopyramide in guinea-pig papillary muscles. Naunyn Schmiedebergs Arch Pharmacol.
1988;337:203-212. Crossref. PubMed.
34. Kunze D, Lacerda A, Wilson D, Brown A. Cardiac Na currents and the inactivating, reopening and waiting
properties of single cardiac Na channels. J Gen Physiol. 1985;86:691-719. Crossref. PubMed.
35. Crumb WJ, Clarkson CW. Characterization of cocaine-induced block of cardiac sodium channels. Biophys J.
1990;57:589-599. Crossref. PubMed.
36. Carmeliet E. Activation block and trapping of pentacainide, a disopyramide analogue, in the Na+ channel of
rabbit cardiac Purkinje fibers. Circ Res. 1988;63:50-60. Crossref. PubMed.
37. Snyders D, Hondeghem L. Effects of quinidine on the sodium current of ventricular guinea pig myocytes:
evidence for a drug-associated rested state with altered kinetics. Circ Res. 1990;66:565-579. Crossref.
PubMed.
38. Clarkson C. Stereoselective block of cardiac sodium channels by RAC109 in single guinea pig ventricular
myocytes. Circ Res. 1989;65:1306-1323. Crossref. PubMed.
39. Clarkson C, Follmer C, Ten Eick R, Hondeghem L, Yeh J. Evidence for two components of sodium channel
block by lidocaine in isolated cardiac myocytes. Circ Res. 1988;63:869-878. Crossref. PubMed.
40. Bean B. Nitrendipine block of cardiac calcium channels: high affinity binding to the inactivated state. Proc
Natl Acad Sci U S A. 1984;81:6388-6392. Crossref. PubMed.
41. Castle N. Bupivacaine inhibits the transient outward K+ current but not the inward rectifier in rat ventricular
myocytes. J Pharmacol Exp Ther. 1990;255:1038-1046. PubMed.
42. Tucker G, Mather L. Pharmacology of local anesthetic agents: pharmacokinetics of local anesthetic drugs.
Br J Anaesth. 1975;47:213-224. PubMed.
43. Coyle D, Denson D, Thompson G, Myres J, Arthur G, Bridenbaugh P. The influence of lactic acid on the serum
protein binding of bupivacaine: species differences. Anesthesiology. 1984;61:127-133. Crossref. PubMed.
44. Janse M, Wit A. Electrophysiological mechanisms of ventricular arrhythmias resulting from myocardial
ischemia and infarction. Physiol Rev. 1989;69:1049-1169. Crossref. PubMed.
45. Moore D, Thompson G, Crawford R. Long-acting local anesthetic drugs and convulsions with hypoxia and
acidosis. Anesthesiology. 1982;56:230-232. Crossref. PubMed.
46. Nancarrow C, Runciman W, Mather L, Upton R, Plummer J. The influence of acidosis on the distribution of
lidocaine and bupivacaine into the myocardium and brain of the sheep. Anesth Analg. 1987;66:925-935.
PubMed.
47. Sánchez-Chapula J. Effects of bupivacaine on membrane currents of guinea-pig ventricular myocytes. Eur J
Pharmacol. 1988;156:303-308. Crossref. PubMed.
48. Valenzuela C, Delpón E, Tamkun MM, Tamargo J, Snyders DJ. Stereoselective block of a human cardiac
potassium channel (Kv1.5) by bupivacaine enantiomers. Biophys J. 1995;69:418-427. Crossref. PubMed.