�

AP

LL

TRanmbap

UN

tNl

P

Review Article

cute Toxicity of Local Anesthetics: Underlyingharmacokinetic and Pharmacodynamic Concepts

aurence E. Mather, Ph.D., Susan E. Copeland, MVetClinStud, andeigh A. Ladd, B.VSc.

The risk of accidental intravascular injection and consequent acute toxicity is ever-present with most neuralblockade techniques. The severity of cardiovascular and central nervous system (respectively, CVS and CNS)toxicity is directly related to the local anesthetic potency, dose, and rate of administration. Nonetheless, althoughthe anesthetic potency of ropivacaine and levobupivacaine is similar to that of bupivacaine, at usual clinicaldoses, ropivacaine and levobupivacaine are less likely than bupivacaine to cause convulsions or lethal dysrhyth-mias. Signs of CNS stimulation, ranging from tremors to convulsions and perhaps cardiac dysrhythmias, can bedescribed in terms of a chaos-derived state change in which the local anesthetic appears to act as an initiator.Both CNS and CVS effects are rather poorly correlated with arterial drug concentrations but better correlatedwith concentrations in the respective regional venous drainage. Lung uptake reduces the maximum drugconcentration by �40%. Prolonging intravenous administration from 1 to 3 minutes results in a similar decreasein maximum concentration. This is an underlying tenet of dose fractionation, but the main advantage of dosefractionation is that the anesthesiologist is able to cease administration with less of the dose given if signs orsymptoms of toxicity occur. Overall, it appears that the gains in safety from ropivacaine and levobupivacaine aredue more to favorable pharmacodynamic enantioselectivity than to pharmacokinetic factors. This essay presentssome pharmacokinetic aspects relevant to acute toxicity of local anesthetics, mainly using data from the authors’studies in a sheep model of simulated accidental intravenous administration. Reg Anesth Pain Med 2005;30:553-566.

af

(mtbfabsrt)nsctmtpqt

he matter of maximum recommended doses oflocal anesthetics was recently reviewed by

osenberg et al.1 They concluded that the presentpproach, which is based on individual drugs, wasot soundly evidence based and that new recom-endations should be based on specific nerve

locks with modifications according to patient char-cteristics to account for individual differences inharmacokinetics of systemically absorbed drug

See Editorial page 513

From the Department of Anaesthesia and Pain Management,niversity of Sydney at Royal North Shore Hospital, SydneySW, Australia.Accepted for publication July 8, 2005.Reprints: Laurence E. Mather, Ph.D., Department of Anaes-

hesia and Pain Management, University of Sydney at Royalorth Shore Hospital, Sydney NSW 2065 Australia. E-mail:

mather@med.usyd.edu.au© 2005 by the American Society of Regional Anesthesia and

ain Medicine.

i1098-7339/05/3006-0009$30.00/0doi:10.1016/j.rapm.2005.07.186

Regional Anesthesia and Pain Medicine, Vol 30, No

nd the concomitant risk of systemic toxic side ef-ects.

Implicit in any maximum recommended doseor minimum effective dose) are the dual require-ents for the dose to be large enough to produce

he desired effect, in this case neural blockade,ut small enough to reasonably preclude side ef-ects, in this case central nervous system (CNS)nd cardiovascular system (CVS) toxicity causedy systemically absorbed local anesthetic. Theseide effects, as pointed out in Rosenberg et al.’seview,1 are easier to understand when the sys-emic pharmacokinetics of the drugs, and (patho-physiologic factors that affect these pharmacoki-etics, are known. However, a more fearsomeide effect of local anesthetics is acute toxicityaused by accidental intravascular administra-ion, occasionally intra-arterial but more com-only intravenous. In such a situation, even

hese revised dosage recommendations may beroblematic because the local anesthetic dose re-uired for most nerve blocks could produce acuteoxicity if injected intravascularly. This topic of

ntravascular accidents and resultant acute toxic-

6 (November–December), 2005: pp 553–566 553

iitatpp

idmtttftmntmpAp

H

L

tisasepdJTliweis

aepdno

T

t

speFodssictoeidaf“rbohi“wnm

ubaoabocsmgsadcassaastsds

554 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

ty was not included in the previous review,1 butt has been reviewed recently, principally fromhe clinical viewpoints of incidence, prevention,nd resuscitation.2-4 However, it would appearhat a pharmacokinetic-pharmacodynamic view-oint of acute toxicity has not previously beenublished.Therefore, the purpose here was to add to the

nformation provided in the previous reviews byescribing some of the pharmacokinetic and phar-acodynamic research on acute local anesthetic

oxicity from accidental intravascular administra-ion. To do this, the authors have mainly drawn onheir own laboratory studies of systemic toxicityrom simulated clinical accidents with local anes-hetics in large animals (sheep). These studies wereainly concerned with preclinical evaluation of theewer long-acting local anesthetics and were in-ended to complement more limited studies in hu-ans. Some of these data have been previously

ublished but are re-presented here in new form.ll of the authors’ studies were approved by appro-riate ethics review panels.

uman Problems to Animal Research

ocal Anesthetic Toxicity and Its Investigation

Systemic toxicity from local anesthetics is rela-ively rare, and its incidence appears to be decreas-ng, corresponding, it would seem, to the wide-pread introduction of procedural safety steps2,3

ided by safer local anesthetics.3-5 Mulroy2 de-cribed an incidence of approximately 12/100,000pidural anesthetics and 200/100,000 brachiallexus blocks. More recent data from 4,291,925eliveries of anesthesia in the operating rooms ofapanese Society of Anesthesiologists Certifiedraining Hospitals indicated that the frequency ofocal anesthetic toxicity, as distinct from other crit-cal incidents and complications of neural blockade,as approximately 1.17/100,000 anesthetic deliv-

ries, with a fatality rate of 0.023/100,000.6 Despitets rarity, local anesthetic toxicity can be cata-trophic to the individual when it occurs.

Although many anesthesiologists may occasion-lly see mild manifestations of toxicity, most neverncounter serious intoxication. Nonetheless, it isrobable that many more “near misses” occur un-etected. Indeed, it was recently stated that “. . . it isot a question of ‘if’ an intravascular injection willccur, [it is] just a question of ‘when’. . .”7

he Need for Large-Animal “Models”

Most data documenting acute local anesthetic

oxicity in humans are gathered in retrospect, es- t

entially opportunistically, or from small cohortrospective studies with low power, so that levels ofvidence for most human data will never be high.or ethical reasons, human subjects can be givennly mildly toxic doses when local anesthetics areeliberately administered intravenously for re-earch, typically to the subjective onset of CNSymptoms.8-12 Information about more serious tox-city must therefore be derived either from clinicalircumstances in which the objectives are preserva-ion of life and well-being, rather than acquisitionf scientific data, or from laboratory animal “mod-ls.” Therefore, human studies can only be a “bluntnstrument” for such investigations. Laboratory ro-ent and isolated tissue models are widely used andre especially useful in elucidating mechanisms andor between-drug comparisons. These are generallysharp, but limited, instruments” because of theestricted types of data that can be obtained orecause of their destructive nature and/or isolationf tissues from their normal milieu. On the otherand, anatomically and physiologically sound stud-

es in large experimental animal models can be avery sharp instrument” to observe extensivehole-body pharmacokinetics and pharmacody-amics in a context quite similar to patient treat-ent.Research data from various large-animal models

sed to investigate local anesthetic intoxication haseen recently compiled by Groban.13 The presentrticle provides more comprehensive data from justne of these models: the adult female sheep prep-ration used by the authors. The major differencesetween this and other models, apart from speciesr gender, are that both CNS and CVS data areollected concurrently from discrete doses in con-cious, preprepared, closed-chested, chronicallyaintained animals, after their recovery from sur-

ical preparation. This model closely resembles theituation in which local anesthetic is accidentallydministered intravascularly to human patients un-ergoing neural blockade. It has allowed pharma-okinetic and dose-response relationship data to becquired for systemic toxic effects ranging fromlight to lethal in essentially unmedicated “normal”ubjects.14,15 Moreover, because intravenous localnesthetic administration provokes both direct CVSnd indirect (CNS-mediated) effects (at least in con-cious subjects), the direct and indirect effects ofhese agents on the brain and heart can be studiedeparately by administration of the local anestheticsirectly into the respective organ’s arterial bloodupply under the same experimental conditions as

hose used for simulated intravenous accidents.14-17

A

s(osepaiosmmccctontpt

fiva

bctBowrtmmt

ctbstCcnaaptbh

FbvqmHSiiiucacda

Acute Toxicity of Local Anesthetics • Mather et al. 555

Case of Acute Bupivacaine Toxicity

Figure 1 shows a time course of effects duringerious, but nonfatal, toxicity from bupivacaine HCl100 mg in 30 mL saline, infused intravenouslyver 3 minutes) in a conscious sheep. The panelshow electrocardiogram (ECG) and cortical electro-ncephalogram (EEG) traces, mean arterial bloodressure, left ventricular pressure and its first deriv-tive (dP/dtmax, an index of myocardial contractil-ty), pulmonary artery blood flow (pulsatile cardiacutput), left coronary artery blood flow, and sagittalinus blood flow; data were acquired before (0 to 10inutes), during (10-13 minutes), and after (13-70inutes) the infusion. Relevant blood samples were

ollected to correlate circulating bupivacaine con-entrations, shown in Figure 2, with the pharma-odynamic effects. The features of bupivacaine in-oxication shown here would also be expected toccur in a human patient after accidental intrave-ous injection of �2 mg/kg (i.e., a dose similar to

he present maximum recommended dose of bu-ivacaine), administered over a time period consis-ent with dose fractionation.

Acute hemodynamic changes began during therst 2 minutes with myocardial depression, re-ealed by decreased dP/dtmax, followed soon after,

ig 1. Physiographic data showing the time course ofupivacaine-induced CNS and CVS intoxication in a pre-iously prepared conscious adult sheep. Data were ac-uired digitally at 256 Hz for 10 minutes before, during 3inutes of intravenous infusion of 100 mg bupivacaineCl in 30 mL saline, and for 60 minutes afterwards.ections: 590 to 600 seconds (immediately before drugnfusion), 740 to 750 seconds (30 seconds before end ofnfusion), 820 to 830 seconds (40 seconds after ceasingnfusion), 1,750 to 1,760 seconds (approximately 20 min-tes after starting infusion). Traces in order from top:raniocaudal ECG, sagittal sinus blood flow (SSBF), meanrterial blood pressure (MABP), cardiac output (CO), leftoronary artery blood flow (CABF), left ventricular dP/tmax, left ventricular pressure (LVP), caudoventral ECG,nd total power (1-32 Hz) in the cortical (dural) EEG.

nd coinciding with the onset of CNS stimulation,

y an abrupt and marked increase in myocardialontractility, accompanied by increases in mean ar-erial blood pressure, heart rate, and cardiac output.y the time convulsions occurred (shown by burstsf EEG overactivity), disturbed cardiac conduction,ith widened QRS complexes and ventricular dys-

hythmias, was also present. Without treatment,he cardiovascular and CNS changes lasted for �20inutes, when there was a sudden return to nor-al heart rate and rhythm; the EEG activity re-

urned to normal �40 minutes later.A peak arterial blood (total) bupivacaine con-

entration of �8 �g/mL occurred, as expected, athe end of the 3-minute infusion; unbound druglood concentrations (see later) were not mea-ured. The arterial blood bupivacaine concentra-ion was �7 �g/mL at �2 minutes, when bothNS and CVS toxicity began (Fig 2). Bupivacaineoncentrations in coronary sinus and sagittal si-us blood (venous drainage of the myocardiumnd brain) reached maxima of �4 and �2 �g/mLfew minutes later, corresponding to the times ofeak concentrations of drug in those tissues. Byhe time that all effects had dissipated, bloodupivacaine concentrations at all 3 sampling sitesad decreased to �1 �g/mL.

Fig 2. Upper panel: bupivacaine concentrations (as race-mate) in arterial, coronary sinus, and sagittal sinus of thesubject shown in Figure 4. Lower panel: net flux ofbupivacaine between blood and myocardium and be-tween blood and brain, as determined from the productof the relevant respective drug arteriovenous concentra-tion gradient and blood flow. Time 0 marks the start and

t he solid circle marks the end of the drug infusion period.

arstawzafoqtt(olwatso

nliifcliaactoOmpaQdd

opiesmdatmr

LP

entnngcdbastvddladoiianbodrtwobtp

I

t

556 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

Figure 2 also shows the tissue (myocardiumnd brain) net flux; this is the instantaneous netate of drug exchange between blood and tissueserially calculated from the respective products ofhe arteriovenous drug concentration differencesnd regional blood flows. Net flux is positivehen there is net transfer from blood to tissue,

ero when the blood and tissue concentrationsre equal, and negative when the net transfer isrom tissue back to blood. The maximum influxccurred at the end of the infusion, thereafteruickly becoming negative and slowly returningoward zero as the arterial blood drug concentra-ions continuously decreased because of dilutiontissue uptake elsewhere) and clearance (metab-lism mainly and excretion). Thus, the initial

oading (uptake) of vital tissues with bupivacaineas very rapid, and unloading commenced soon

fter and continued, albeit slowly, for a very longime. This pattern is consistent with facile diffu-ion of this lipophilic drug and the high perfusionf these tissues.The acute hemodynamic changes (Fig 1) are the

et resultant of direct and indirect CVS effects, theatter mediated mainly via the CNS.18 This complexnteraction between the CNS and CVS is character-stic of the whole-body response to acute toxicityrom large intravenous doses of local anesthetics inonscious subjects. To study the cardiac effects ofocal anesthetics without the influence of CNS, var-ous investigators have used site-directed coronaryrterial administration in large animals, principallycutely prepared pigs19,20 and dogs21 and chroni-ally prepared conscious sheep.16 The results ofhese in vivo studies are essentially similar to thosef ex vivo studies using isolated heart tissue.22

verall, local anesthetics cause dose-related directyocardial depression with potency essentially pro-

ortional to potency for neural blockade, and theylso disturb cardiac conduction, shown primarily byRS complex widening and dysrhythmias, witheath caused by pump failure and/or malignantysrhythmias.16,19,20

Information about indirect CNS-mediated effectsf blood-borne local anesthetics on CVS has beenrovided from site-directed carotid arterial infusionn chronically prepared conscious sheep.17 Thesexperiments have shown that CNS effects (convul-ions) of blood-borne local anesthetics result inyocardial stimulation (increased left ventricular

P/dtmax) but not necessarily cardiac dysrhythmias,pparently unlike introduction of drug directly intohe brain in which cardiac neurovascular controlay become deranged and in which cardiac dys-

hythmias may occur.23-25

ocal Anesthetic Chemistry andharmacodynamic “Potency”

The potency for CNS stimulatory (proconvulsant)ffects of local anesthetics parallels their potency foreural blockade but with substantial differences be-ween the enantiomers of bupivacaine.17,26 This isot surprising because of the conservation of ge-etic coding among the superfamily of voltage-ated sodium channels in various tissues.27 Be-ause, as mentioned earlier, direct myocardialepression also is proportional to potency for neurallockade, it is also not surprising that over a widerray of local anesthetics with different chemicaltructures, local anesthetic potency is also propor-ionally associated with lethality, judged by intra-enous median lethal dose values in laboratory ro-ents26,28,29 (Table 1). There are appreciableifferences between enantiomers of certain chiralocal anesthetics: these principally relate to differentffinities for ion channel receptors.30 Nevertheless,espite preservation of this rank order in a varietyf experimental designs, the mechanisms of lethalntoxication appear to differ among local anesthet-cs with different structures in both conscious andnesthetized subjects.31,32 On one hand, intrave-ous bupivacaine is more likely to produce fatalityy sudden onset of lethal dysrhythmias; on thether, intravenous lidocaine is more likely to pro-uce progressive contractile failure. Intravenousopivacaine and levobupivacaine have produced fa-alities by either mechanism, and it is not yet clearhat makes one outcome more likely than thether in individual cases. In addition, differencesetween local anesthetics in the cardiac contribu-ions of direct and indirect toxic effects remain per-lexing.

Table 1. Intrinsic Whole-Body Toxicity of Somemportant Local Anesthetic Agents as Determined

by the Median Lethal Dose After IntravenousInjection in the Mouse

Mather28 Aberg26Akerman

et al29

Prilocaine* 37.8 � 1.2Lidocaine 28.8 � 1.4 18.8 � 2.5Mepivacaine* 27.8 � 1.1 35 � 3.0

R(�)-Mepivacaine 32 � 2.5S(�)-Mepivacaine 34 � 3.5

Ropivacaine 11.0 � 0.2Bupivacaine* 5.2 � 0.2 7.3 � 1.0 7.9 � 0.4

R(�)-Bupivacaine 7.9 � 1.0S(�)-Bupivacaine† 9.6 � 1.0

NOTE. Dose is given in mg/kg of local anesthetic base; error ishe standard error of the mean.

*Clinically-used racemate.†Now also known as levobupivacaine.

When infused directly into the left coronary ar-

tlctiiecrlsvittei

PI

mrmst�gv�qs

ceopt

hdtvjcdrcthtbflpibe

aamiRuat

Frsdm2

Fcssa(aioa

Acute Toxicity of Local Anesthetics • Mather et al. 557

eries of conscious sheep16 and anesthetized pigs,20

ethal ventricular fibrillation occurred with bupiva-aine, levobupivacaine, and ropivacaine. Whereashere were no significant differences between drugsn lethal doses when infused as a small discrete dosento coronary arteries of conscious sheep,16 differ-nces between drugs were found when infused asumulative doses in anesthetized pigs such that theank order of lethal potency was bupivacaine �evobupivacaine � ropivacaine.20 Furthermore, ashown in Figure 3, lethal dysrhythmias and cardio-ascular collapse can occur also with intracoronarynfusion of lidocaine in both conscious16 and anes-hetized subjects.19,20 Such examples show clearlyhat sudden-onset cardiac collapse from a local an-sthetic need not be related to concurrent CNSntoxication.

harmacokinetic and Pharmacodynamicnterpretations of “Blood Drug Concentrations”

It can be seen from Figures 1 and 2 that a simpleonotonic drug dose or blood drug concentration-

esponse relationship cannot be shown for the he-odynamic changes in vivo, as can be normally

hown in preparations ex vivo. The data suggesthat a bupivacaine arterial blood concentration of7 �g/mL is both convulsant and dysrhythmo-

enic, but the convulsions and dysrhythmias bothanish when the concentration has decreased to �1g/mL. These observations raise a fundamentaluestion: what is the “toxic concentration?” The

ig 3. Cardiac effects of lidocaine infused into the leftoronary arteries of a previously prepared conscious adultheep. A dose of 40 mg/min was infused from time � 300econds. Data acquisition methods and abbreviations ares described for Figure 1. Sections: 290 to 300 secondsimmediately before drug infusion), 320 to 330 seconds,nd 370 to 390 seconds. The sheep was prepared accord-ng to our normal procedures. Death occurred by suddennset of dysrhythmias similar to those caused by bupiv-caine immediately followed by cardiovascular collapse.

ituation becomes especially complicated with a ra- i

emic drug such as bupivacaine because of differ-ntial toxicity and pharmacokinetics of the enanti-mers. Both systemic (or whole body) and regionalharmacokinetics are relevant in this context, andhey are complementary.

Systemic pharmacokinetics are time-averagedypothetical properties derived from curve-fittingrug concentration-time data sets: their relevanceo systemic absorption of local anesthetics was re-iewed by Rosenberg et al.1 After intravenous in-ection, the regional pharmacokinetics are espe-ially important: these are empirical propertieserived from products of the respective concurrentegional blood flows and arteriovenous drug con-entration gradients.33 For local anesthetic toxicity,he rates of drug uptake (influx) into the brain andeart are the most relevant. However, because ofhe requirement to obtain afferent and efferentlood drug concentrations and regional bloodows, regional pharmacokinetics are virtually im-ossible to obtain, except in large experimental an-mals. Fluxes so-measured in vivo are net fluxesecause they are the difference between influx andfflux.In addition to the previously described consider-

tions, general anesthesia is well documented tolter pharmacokinetics34 and drug effects and thusay affect interpretation of research data from var-

ous investigations of local anesthetic intoxication.egardless of the pharmacokinetic approachessed, the potential impact of concomitant generalnesthesia on various research models must beaken into consideration.35,36

ig 4. Physiological effects of intravenous infusion ofopivacaine HCl (150 mg over 3 minutes) in an adultheep. Data acquisition methods and abbreviations are asescribed for Figure 1. Convulsions commenced at 2.1inutes (at 726 seconds), and death occurred abruptly at

.6 minutes (at 756 seconds) after commencement of

nfusion.

A“

aipfi(btpidtsppmcebbdtttvslwp

cd

tdeddfatddtaeacttbm

cfcsetam(bsrstth(mtspsc

DI

ibFc

fit

FcoBimpiv

558 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

natomical and Physiological Interpretation ofBlood Drug Concentrations”

An example showing sudden death of a sheepfter intravenous infusion with ropivacaine furtherllustrates the complicated interaction betweenharmacokinetics and pharmacodynamics and dif-culties in describing “toxic blood concentrations”Figs 4 and 5). The point of measuring circulatinglood (or plasma or serum) drug concentrations ishat, after initial distribution, they appear to be inseudo-equilibrium with the drug concentrationsn tissues37,38 so that any systemic response to therug would be more predictable from, and relatedo, blood drug concentrations than dose alone. Butoon after an intravenous administration, beforeseudo-equilibrium has been attained, even highlyerfused tissues do not necessarily behave as a “ho-ogeneous compartment”; thus, blood drug con-

entrations may be only roughly related to drugffect, especially if profound physiological distur-ances are also occurring. However, the total-bodyurden (or load) of drug will also be an impliciteterminant of the drug effect because it influenceshe magnitude of the relevant regional net flux andhe time over which drug concentrations remain inhe “active” region. Acute toxic effects from intra-enous administration are thereby likely to be moreevere, but shorter lived, than if caused by relent-ess systemic absorption from perineural injectionhen the dose and absorption rate exceed the ca-acity of the body for drug clearance.Pharmacodynamic interpretation of “blood drug

oncentrations” is influenced by the site, time after

ig 5. Arterial and coronary sinus blood, heart, and brainoncentrations of ropivacaine after intravenous infusionf 150 mg over 3 minutes in the study shown in Figure 4.lood was sampled until 4 minutes; the massive increase

n coronary sinus blood ropivacaine concentration postortem does not reflect the tissue concentrations and

resumably results from a combination of factors, includ-ng collection of infused drug solution with post mortemascular pressure changes.

rug administration, and conditions under which a

he blood samples are obtained. Whereas arterialrug concentrations are essentially the same wher-ver sampled, venous concentrations depend onrug solubility and metabolism in the region beingrained. However, venous blood samples, usuallyrom an antecubital vein in humans, are commonlynalyzed because of their ease of collection, butheir concentrations are damped and temporallyelayed because of the transit and solubility of therug in the (forearm) region being drained. Theime after drug administration is a “hidden” vari-ble in that it influences different drugs to differentxtents. Hence, if, as is sometimes done in humannd animal studies, comparison between pharma-ological tolerability of different drugs is made onhe basis of the tolerated blood drug concentration,hen the comparison can be skewed by differencesetween drugs in their rates of tissue uptake andetabolism.Moreover, both arterial and venous blood drug

oncentrations are influenced by hemodynamicactors,37 but the product of blood flow and drugoncentration will be constant because of the con-ervation of matter, with all other things beingqual.33 All of these factors impinge on interpreta-ion of blood drug concentration in relation to effectnd clearly cannot be studied systematically in hu-ans. For example, after death, sampled blood

particularly from the heart) may be contaminatedy drainage from different sites as vascular pres-ures fall, including those from blood vessels en-iched with injected drug solution (Fig 5). The con-equence is that the relationship between blood andissue drug concentrations and side effects becomesenuous and can make interpretation unexpectedlyazardous, especially if used for forensic purposese.g., to calculate whether a dosage error has beenade on the basis of post mortem drug concentra-

ion data). Hence, by offering a “calibration factor,”ystematic laboratory animal studies of regionalharmacokinetics have an important role in under-tanding human cases of acute intoxication thatannot be studied so invasively.

ose Dependence of Local Anestheticntoxication

Although it is generally held that serious cardiacntoxication from local anesthetics lags temporallyehind CNS intoxication, this was not observed inigure 1. Are there characteristic “toxic” blood drugoncentrations, as suggested earlier?The abrupt onset of convulsions is sometimes the

rst indication of serious local anesthetic intoxica-ion. The progression of acute CNS toxicity appears,

fter reaching some ill-defined threshold, to follow

acsuedstdts

v

ccsTiwcdoodCafwwat

Flmsscivo5

Fomavac

Acute Toxicity of Local Anesthetics • Mather et al. 559

chaotic model. As time progresses, the subjectonstantly reaches new CNS “state” levels thatummate temporally to either advance the effect,ltimately leading to convulsions, or extinguish theffects. Although the progression of effects clearlyepends to some extent on receptor occupancy, iteems that local anesthetics behave more as ahreshold initiator than as an agonist in a simpleeterministic manner whereby its degree of recep-or occupancy determines the intensity of CNStimulation.18

tk;4Attempts have been made to define the “con-

ig 6. Blood concentration-dose relationships forevobupivacaine after intravenous infusions made over 3

inutes on separate occasions in the same consciousheep subject. Left panels: arterial (closed circles) andagittal sinus (open circles) blood levobupivacaine con-entrations. Right panels: net flux (as % dose/min; netnflux is positive and net efflux is negative) of levobupi-acaine across the brain. Arrows indicate the time tonset of convulsions. There was no convulsion from the0 mg dose.

ulsant arterial blood drug concentration” so that it o

an be avoided. A significant problem is that thisoncentration is not a constant and varies with thepeed of injection of the drug, among other things.his is illustrated in Figure 6, using an example ofncreasing doses of levobupivacaine in a sheep, inhich the arterial drug concentration at onset of

onvulsions appears to increase with increasingose, with a corresponding decrease in the time ofnset. Figure 7 shows composite data from a cohortf animals infused intravenously with increasingiscrete doses of levobupivacaine and bupivacaine.onvulsions from levobupivacaine appear to begint a lower arterial blood drug concentration thanrom bupivacaine at the 75 mg/3 min dose (athich 5/6 convulsed with bupivacaine and 3/7ith levobupivacaine). However, this is partly an

rtifact of the duration of infusion in relation to theime taken for equilibration of drug between the

ig 7. Dose and arterial blood concentration data for thenset of convulsions from intravenous infusion over 3inutes of bupivacaine or levobupivacaine in a cohort of

dult female sheep. Upper panel: time of onset of con-ulsions plotted against administered dose, middle panel:rterial blood drug concentration at time of onset ofonvulsions, and lower panel: dose infused to the time of

nset of convulsions.

(soowasotclc

smd

odrescticcrb“t

tpAdactc

Fbfca

Fdoc(dgd

560 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

rapidly changing) arterial concentration and tis-ues. At the lowest dose in particular, the onsetccurred later with levobupivacaine after cessationf the 3-minute infusion and when concentrationsere decreasing. Moreover, as shown in Figure 8,

rterial drug concentrations at the offset of convul-ions were always much less than those at thenset, another effect of the experimental condi-ions. Only a fairly broad range for blood drug con-entrations “toxic” to CNS can be described for eachocal anesthetic as it depends, at least partly, on theonditions under which toxicity occurs.In contrast, drug concentration in the sagittal

inus at the onset of convulsions does not changearkedly with dose rate. As shown in Figure 9,

ig 8. Levobupivacaine and bupivacaine arterial bloodrug concentrations resulting from intravenous infusionf the nominated doses over 3 minutes on separate oc-asions in the same conscious sheep. The points of onsetfilled upward triangle) and offset of convulsions (filledownward triangle) and of onset (empty upward trian-les) and offset (empty downward triangles) of cardiacysrhythmias are shown.

rug concentrations in sagittal sinus blood at the t

nset and offset of convulsions are similar for eachose and vary far less between doses than do arte-ial drug concentrations. This indicates that the av-rage “brain” levobupivacaine concentration is rea-onably similar at the onset and offset ofonvulsions (Fig 9). According to venous equilibra-ion theory, the regional venous drug concentrations presumed to better represent the pharmacologi-ally relevant concentration. Thus, the apparent in-rease in “convulsant drug concentration” for arte-ial blood with increasing dose appears to be causedy the rate of equilibration across the blood-brainbarrier” being slower than the distribution rate inhe rest of the body.

Figure 8 also shows the arterial drug concen-rations and the points of onset and offset oferiods in which cardiac dysrhythmias occurred.rterial drug concentrations at the onset of car-iac dysrhythmias differed and were more vari-ble as a function of dose than the correspondingoronary sinus concentrations. This is analogouso the correlation of CNS effects with drug con-entrations in sagittal sinus blood. We believe

ig 9. Levobupivacaine and bupivacaine sagittal sinuslood drug concentrations resulting from intravenous in-usion of the nominated doses over 3 minutes in the sameonscious sheep. The points of onset (filled upward tri-ngle) and cessation of convulsions (filled downward

riangle) are shown.

tCdmlttrTtstiae(dopolmds

SI

tttcbtlbtb

mbdWfsrihc

FMwlctofc

Acute Toxicity of Local Anesthetics • Mather et al. 561

hat a similar chaotic paradigm also prevails forVS effects, leading to the initiation of malignantysrhythmias. Dose-dependent depression ofyocardial contractility occurs in proportion to

ocal anesthetic potency and blood drug concen-ration until doses are large enough to stimulatehe CNS, whereupon the depression is abruptlyeversed with the onset of convulsions.13,14,31,32

hus, the matter of myocardial intoxication needso be studied in different settings including con-cious and anesthetized subjects, whole-body (in-ravenous) and site-directed (close arterial) dos-ng, and intact and isolated tissues to permit a fullnalysis to be made. This would enable closerxamination of the connection between directlocal cardiac) and indirect (mediated by CNS)rug effects. Overall, the role of the CNS by wayf sympathetic nervous system stimulation seemsaramount; the onset of seizures causes reversalf myocardial depression but predisposes to ma-

ignant dysrhythmias. Not surprisingly then, aarked difference in CVS outcome appears to

epend on cardiac resting state and state of con-ciousness.19,20,32,35,36,39-42

ig 10. Effect of prolonging the duration of administratioeasured arterial (open squares) and coronary sinus (opith the dose infused over 1 minute. (B) Measured art

evobupivacaine concentrations in the same sheep as inurve fitted for pharmacokinetic simulation purposes to the subject after 3 minutes of infusion. (D) Simulated artever 30 seconds (no fractionation). (E) Simulated arterialractionation as 3 equal portions, each over 30 seconds,

oncentrations for the dose infused by fractionation as 6 equal

peed of Local Anesthetic Injection and Uptakento Lungs

Slowing the speed of injection and/or dose frac-ionation is now routine clinical practice: it allowsime for reduction of arterial blood drug concentra-ions delivered to vital organs by concurrent pro-esses of distribution and elimination elsewhere buty how much? Similarly, it has long been knownhat the lungs exert a “protective” effect againstocal anesthetic intoxication by attenuating arteriallood drug concentrations,32 but the magnitude,ime course, and dose dependency of this effect haseen difficult to evaluate.It is intuitive that dose fractionation will reduceaximum arterial blood drug concentration (Cmax),

ut there does not yet appear to be experimentalemonstration of the magnitude of the outcome.e have found experimentally that prolonging

rom 1 to 3 minutes the intravenous infusion inheep of 37.5 mg levobupivacaine (as an example)educes its arterial Cmax by �40%, with correspond-ng reductions in other relevant regions such as theeart as shown by the coronary sinus levobupiva-aine concentrations (Fig 10). Computer simula-

7.5 mg levobupivacaine administered intravenously. (A)ngles) blood levobupivacaine concentrations in a sheep

open squares) and sagittal sinus (open triangles) bloodh the dose infused over 3 minutes. (C) Polyexponentialasured arterial blood levobupivacaine concentrations ofood levobupivacaine concentrations for the dose infusedlevobupivacaine concentrations for the dose infused by

ute apart. (F) Simulated arterial blood levobupivacaine

n of 3en triaerial (A withe merial blblood1 min

portions, each over 30 seconds, 1 minute apart.

tos1aTtnofldmlaCous

pa

etimlalinarsstferaoenarbb

PD

rtbbatataiacfwgseecbastid

Flmsatnvcac

562 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

ions of the same dose given as a bolus (30 seconds)r by spaced fractionation into 3 or 6 equal portionshows that the arterial blood Cmax is attenuated (Fig0), but this may not be a sufficient reduction inrterial drug concentration to avoid intoxication.he most important feature of dose fractionation ishat it gives the anesthesiologist an early opportu-ity to cease administering the drug if a side effectccurs. The latter aspect is pertinent, given the con-icting views on the suitability of a discrete testose to prevent intravenous administration of theain dose, particularly for ropivacaine and

evobupivacaine, which are less likely than bupiv-caine to provoke early CNS symptoms or signs.43,44

learly, in this context, and especially in the CNS-btunded patient, there is a case to be made forsing the response from an epinephrine-containingolution for the entire dose.

Because the lungs are a large organ with complexathways of transit and tissue uptake, they attenu-te arterial blood drug concentrations and, to some

ig 11. Blood concentration-dose relationships forevobupivacaine after intravenous infusions made over 3

inutes on separate occasions in the same consciousheep. Left panels: pulmonary arterial (open triangles)nd arterial (open circles) blood levobupivacaine concen-rations. Right panels: net flux (calculated as % dose/min;et influx is positive, net efflux is negative) of levobupi-acaine across the lungs calculated from the product ofardiac output and the difference between pulmonaryrterial and arterial drug concentrations. The time of

iessation of drug infusion is shown by an arrow.

xtent, consequent risk of local anesthetic intoxica-ion, whether from accidental intravenous admin-stration or systemic absorption after perineural ad-

inistration,37,45-47, but by how much and for howong? Some examples of arterial and pulmonaryrterial drug concentrations over a large range ofevobupivacaine doses in the same sheep are shownn Figure 11. These show that changes in pulmo-ary arterial drug concentrations precede those inrterial blood during infusion over a large doseange, as expected, and that the gradient reversesoon after cessation of infusion (i.e., net influx isoon followed by net efflux). The lungs attenuatehe arterial drug concentrations by �20%, and theraction of dose apparently lost into the lungs isssentially independent of dose; however, it is soonegained in an exponentially decreasing manner,nd this is a point often not brought out withoutbservation over an extended period of time. Inter-stingly, we have also found that there are no sig-ificant acute differences between bupivacaine en-ntiomers in sheep or humans in their rate ofegional uptake despite a slightly greater tissueinding of the R-bupivacaine enantiomer than S-upivacaine enantiomer.46

rotein Binding and the Interpretation of “Bloodrug Concentrations”

A further obstacle in the interpretation of theelationship between circulating drug concentra-ion and drug effects arises from plasma proteininding and blood cell uptake of the drug. This cane even more complex if the drug is administered asracemate, such as bupivacaine, where the enan-

iomers can have different protein binding affinitiess well as pharmacodynamics. Overall, plasma pro-ein binding opposes blood cell uptake so that bloodnd plasma drug concentrations will diverge withncreased plasma binding.48 Therefore, the questionrises as to whether unbound (or free) drug con-entrations (in plasma water) need to be measuredor correlation with pharmacological effects orhether total blood or total plasma concentrations

ive essentially the same information, regardless ofpecies.11,12,40,49-52 The former are assumed to bequivalent to tissue interstitial fluid and thus inquilibrium with receptor concentrations; unboundoncentrations are therefore intuitively attractive,ut whole plasma or whole blood concentrationsre normally measured because of the lack of auitable technique for measuring free drug concen-rations in most laboratories. In all of the foregoing,t is emphasized that it is the plasma unboundrug concentration, not percentage, that is the crit-

cal issue in driving the pharmacodynamics; there

is

pcaerct(msbtdcea(r

lctttcghfiatAfi

vsasiocgwt

dabwupaatcfLcLhcrarasaubp

A

S

siatwdeasah

da

Faelbal

Acute Toxicity of Local Anesthetics • Mather et al. 563

s sometimes confusion on this point and its con-equences.48

Measurement of unbound drug concentration inlasma now features in many research articles con-erned with local anesthetic intoxication in humansnd experimental animals. As the current knowl-dge has accrued over the past few decades, theeporting of plasma-unbound local anesthetic con-entrations has been useful to better characterizehe pharmacokinetic differences between the drugse.g., in transplacental distribution). However, inost cases, the resultant data are a set of numbers

caled differently according to unbound fraction orlood cell uptake that remain closely related to theotal plasma drug concentrations of the particularrug, but in a nonlinear manner, and are subject toumulative methodological errors. With some res-rvations depending on the use to which the datare being put, it would seem that the total plasmaor blood, as appropriate) drug concentration is aeasonable compromise for most applications.

The most significant feature about the binding ofocal anesthetics is that its extent is markedly con-entration dependent in all species (Fig 12). Thus,he unbound fraction at “toxic” blood drug concen-rations can be several times greater than at “non-oxic” concentrations, although it is the latter con-entrations at which most drug binding data areenerated experimentally. Various plasma proteinsave different affinities and capacities to bind dif-

erent local anesthetics,53 and the extent of bindings altered by changes of their concentrations, as wells by acid-base changes that affect both the ioniza-ion of the drug and conformation of the protein(s).s a further complication, there is some evidence

or species-related quantitative differences in bind-

ig 12. An example of concentration dependence of localnesthetic plasma protein binding showing the individualnantiomers of bupivacaine (from racemate) andevobupivacaine in an individual sheep, showing the un-ound fraction is several-fold greater at concentrationsssociated with central nervous system and cardiovascu-ar system toxicity.

ng, which may relate to differential binding to a

arious proteins and their abundance in differentpecies.52 All such measures are determined gener-lly ex vivo, and they may not reflect well theituation in vivo during a toxic event, when phys-ology may be altered (e.g., in the immediate post-perative period51 when concentration of the prin-ipal binding protein of local anesthetic [�1 acid-lycoprotein] changes as a response to stress, andhen acid-base and/or fluid balances may be al-

ered).54

Sometimes the value of the average blood:plasmarug concentration ratio (B/P, sometimes abbrevi-ted �) is used to provide a factor for conversionetween blood and plasma drug concentrationshen one is measured and the other is required forse in a pharmacokinetic calculation.37 A measuredlasma bupivacaine concentration of 3.0 �g/mL insubject with a B/P ratio of 0.67 would be equiv-

lent to a blood concentration of 2.0 �g/mL becausehe blood cells occupy a disproportionate volumeompared to their drug concentration. It thereforeollows that a value of total body clearance as 0.6/min determined from measured plasma drug con-entrations would be equivalent to a value of 0.9/min if calculated from blood samples. A potentialazard of this approach is that B/P itself can beoncentration dependent and could depend, theo-etically, on other variables such as acid-base bal-nce in the samples assayed, which can alter theatio of unionized to ionized drug concentrationsnd protein binding.54 Nevertheless, it gives a rea-onable approximation, as long as concentrationsre neither very high nor very low. B/P is alsoseful as a surrogate for the average drug plasmainding on the premise that greater binding tolasma protein decreases its value.

nimal Research to Human Problems

ome Implications

Prospective studies in humans are mainly used totudy pharmacokinetics of local anesthetics after per-neural administration, and there are numerous ex-mples of such studies. Intravascular administrationo humans for research purposes is permissible onlyith relatively small doses, with the intention of pro-ucing only mild intoxication, and there are a fewxamples of this type of study. The main purpose ofnimal research in this context, therefore, is to under-tand how other mammalian species respond to localnesthetic agents and to use the relevant findings inuman medicine.It has been proposed that maximum recommended

oses of local anesthetics should be specified for eachdministration site rather than for each agent.1 This is

sensible proposal. Nevertheless, it does not, and

paaecbccdgouf

inewbdmtncdpScdmsabdbwt

tathmcpereucpicsmi

itisabtmdtm

1

1

1

564 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

robably cannot, take account of the possibility ofccidental intravascular injection and consequentcute local anesthetic intoxication, which remains anver present risk of neural blockade. Although somelinical reports of acute intoxication include measuredlood drug concentrations to verify the probableause, the amount of pharmacokinetic and pharma-odynamic data available in such circumstances, un-erstandably, is typically sparse. Therefore, to providereater detail on these matters, this article has focusedn some findings from the authors’ research into sim-lated accidental intoxication in conscious sheep and

rom other models used for the same purpose.The potency of local anesthetics for causing acute

ntoxication essentially parallels that for producingeural blockade, with the exceptions of the newernantiopure agents, ropivacaine and levobupivacaine,hich have anesthetic potency similar to bupivacaineut are less toxic. However, the use of these agentsoes not diminish the risk of their intravascular ad-inistration. By-and-large, the gains in safety with

hese agents are due to more favorable pharmacody-amic profiles than from improved pharmacokineticharacteristics. The risk to the individual from acci-ental intravascular administration is thus largely pro-ortional to the dose of local anesthetic administered.afety is promoted by use of the least toxic drugonsistent with achieving the desired outcomes ofensity and duration of nerve block. It is also pro-oted by safer administration techniques, especially

lowing the speed of a local anesthetic injectionnd/or dose fractionation. Although reduction inlood drug concentrations by these means is un-oubtedly beneficial, the main gain in safety may beecause of the ability to cease drug administrationith less of the dose having been given at the onset of

oxic symptoms or signs.Formal pharmacokinetic-pharmacodynamic inves-

igation of local anesthetic acute intoxication is virtu-lly impossible in humans. Studies in large animalshat can be applied to humans can be of immenseelp in understanding the underlying principles andaking comparisons between drugs and between re-

ipients of drugs, especially when pathophysiologicalerturbations are present. Traditional models of drugffect based on receptor occupancy and monotonicelationships between blood drug concentration andffect do not apply well to local anesthetics, evennder controlled laboratory conditions, especially inonscious subjects. Despite this, whenever possible,harmacokinetic investigation of human intoxications valuable, but measured blood local anesthetic con-entrations in clinical cases of suspected intoxicationhould be interpreted with some caution and postortem samples present special hazards, particularly

f used to deduce a dosage error.

The other major role of studies in animals is tonvestigate treatments for intoxication. In the questo deal with acute local anesthetic intoxication, var-ous strategies have been/are being investigated,ome involving pretreatment with pharmacologicalgents, resuscitation strategies, and even “silverullet”/selective treatments. This topic is beyondhe scope of this article, in which we have outlinedore of the relationships between drug structure,

ose, pharmacokinetics, and pharmacodynamicshat assist in making recommendations about theaximum doses of local anesthetics.

References

1. Rosenberg PH, Veering BT, Urmey WF. Maximumrecommended doses of local anesthetics: A multifac-torial concept. Reg Anesth Pain Med 2004;29:564-575.

2. Mulroy MF. Systemic toxicity and cardiotoxicity fromlocal anesthetics: Incidence and preventive measures.Reg Anesth Pain Med 2002;27:556-561.

3. Faccenda KA, Finucane BT. Complications of re-gional anaesthesia. Incidence and prevention. DrugSafety 2001;24:413-442.

4. Weinberg GL. Current concepts in resuscitation ofpatients with local anesthetic cardiac toxicity. RegAnesth Pain Med 2002;27:568-575.

5. Ruetsch YA, Boni T, Borgeat A. From cocaine toropivacaine: The history of local anesthetic drugs.Curr Top Med Chem 2001;1:175-182.

6. Irita K, Kawashima Y, Morita K, et al. [Critical inci-dents during regional anesthesia in Japanese Societyof Anesthesiologists-Certified Training Hospitals: Ananalysis of responses to the annual survey conductedbetween 1999 and 2002 by the Japanese Society ofAnesthesiologists]. Masui 2005;54:440-449.

7. Buckenmaier CC, Bleckner LL. Anaesthetic agentsfor advanced regional anaesthesia: A North Americanperspective. Drugs 2005;65:745-759.

8. Mather LE, Long GJ, Thomas J. The intravenoustoxicity and clearance of bupivacaine in man. ClinPharmacol Ther 1971;12:935-943.

9. Mather LE, Tucker GT, Murphy TM, Stanton-HicksMD, Bonica JJ. Cardiovascular and subjective centralnervous system effects of long-acting local anaesthet-ics in man. Anaesth Intensive Care 1979;7:215-221.

0. Scott DB, Lee A, Fagan D, Bowler GM, Bloomfield P,Lundh R. Acute toxicity of ropivacaine comparedwith that of bupivacaine. Anesth Analg 1989;69:563-569.

1. Knudsen K, Beckman Suurkula M, Blomberg S, Sjo-vall J, Edvardsson N. Central nervous and cardiovas-cular effects of i.v. infusions of ropivacaine, bupiva-caine and placebo in volunteers. Br J Anaesth 1997;78:507-514.

2. Stewart J, Kellett N, Castro D. The central nervoussystem and cardiovascular effects of levobupivacaineand ropivacaine in healthy volunteers. Anesth Analg

2003;97:412-416.

1

1

1

1

1

1

1

2

2

2

2

2

2

2

2

2

2

3

3

3

3

3

3

3

3

3

3

4

4

Acute Toxicity of Local Anesthetics • Mather et al. 565

3. Groban L. Central nervous system and cardiac effectsfrom long-acting amide local anesthetic toxicity inthe intact animal model. Reg Anesth Pain Med 2003;28:3-11.

4. Huang YF, Pryor ME, Mather LE, Veering BT. Car-diovascular and central nervous system effects of in-travenous levobupivacaine and bupivacaine insheep. Anesth Analg 1998;86:797-804.

5. Chang DH, Ladd LA, Wilson KA, Gelgor L, MatherLE. Tolerability of large-dose intravenous levobupi-vacaine in sheep. Anesth Analg 2000;91:671-679.

6. Chang DH, Ladd LA, Copeland S, Iglesias MA, Plum-mer JL, Mather LE. Direct cardiac effects of intracoro-nary bupivacaine, levobupivacaine and ropivacainein the sheep. Br J Pharmacol 2001;132:649-658.

7. Ladd LA, Chang DH, Wilson KA, Copeland SE, Plum-mer JL, Mather LE. Effects of CNS site-directed ca-rotid arterial infusions of bupivacaine, levobupiva-caine, and ropivacaine in sheep. Anesthesiology 2002;97:418-428.

8. Ladd LA, Mather LE. Central Effects Index—A semi-quantitative method for the assessment of CNS tox-icity of local anaesthetic agents in sheep. J PharmacolToxicol Methods 2000;44:467-476.

9. Nath S, Häggmark S, Johansson G, Reiz S. Differen-tial depressant and electrophysiologic cardiotoxicityof local anesthetics: An experimental study with spe-cial reference to lidocaine and bupivacaine. AnesthAnalg 1986;65:1263-1270.

0. Morrison SG, Dominguez JJ, Frascarolo P, Reiz S. Acomparison of the electrocardiographic cardiotoxiceffects of racemic bupivacaine, levobupivacaine, andropivacaine in anesthetized swine. Anesth Analg2000;90:1308-1314.

1. Fujita Y, Kimura K, Mihira M, Yasukawa T. De-creased coronary blood flow is not responsible formyocardial dysfunction during bupivacaine-inducedcardiotoxicity. Acta Anaesth Scand 1996;40:216-221.

2. Heavner JE. Cardiac toxicity of local anesthetics inthe intact isolated heart model: A review. Reg AnesthPain Med 2002;27:545-555.

3. Pickering AE, Waki H, Headley PM, Paton JF. Inves-tigation of systemic bupivacaine toxicity using the insitu perfused working heart-brainstem preparation ofthe rat. Anesthesiology 2002;97:1550-1556.

4. Heavner JE. Cardiac dysrhythmias induced by infu-sion of local anesthetics into the lateral cerebral ven-tricle of cats. Anesth Analg 1986;65:133-138.

5. Denson DD, Behbehani MM, Gregg RV. Enantiomer-specific effects of an intravenously administered ar-rhythmogenic dose of bupivacaine on neurons of thenucleus tractus solitarius and the cardiovascular sys-tem in the anesthetized rat. Reg Anesth 1992;17:311-316.

6. Åberg G. Toxicological and local anaesthetic effects ofoptically active isomers of two local anaesthetic com-pounds. Acta Pharmacol Toxicol 1972;31:273-286.

7. Clare JJ, Tate SN, Nobbs M, Romanos MA. Voltage-gated sodium channels as therapeutic targets. Drug

Discov Today 2000;5:506-520.

8. Mather LE. Studies in the absorption, distribution,metabolism and excretion of drugs by adult and neo-natal humans [PhD Thesis]. Sydney, Australia: TheUniversity of Sydney, 1972.

9. Åkerman B, Hellberg IB, Trossvik C. Primary evalu-ation of the local anaesthetic properties of the aminoamide agent ropivacaine (LEA 103). Acta AnaesthScand 1988;32:571-578.

0. Vladimirov M, Nau C, Mok WM, Strichartz G. Potencyof bupivacaine stereoisomers tested in vitro and in vivo:Biochemical, electrophysiological, and neurobehavioralstudies. Anesthesiology 2000;93:744-755.

1. Nancarrow C, Rutten AJ, Runciman WB, Mather LE,Carapetis RJ, McLean CF, Hipkens SF. Myocardialand cerebral drug concentrations and the mecha-nisms of death after fatal intravenous doses of lido-caine, bupivacaine, and ropivacaine in the sheep.Anesth Analg 1989;69:276-283.

2. Groban L, Deal DD, Vernon JC, James RL, Butter-worth J. Does local anesthetic stereoselectivity orstructure predict myocardial depression in anesthe-tized canines? Reg Anesth Pain Med 2002;27:460-468.

3. Upton RN, Mather LE, Runciman WB, Nancarrow C.The use of mass balance principles to determine re-gional drug distribution and elimination. J Pharmaco-kinet Biopharm 16:13-29, 1988.

4. Runciman WB, Myburgh J, Upton RN, Mather LE.Effects of anaesthesia on drug disposition. In: Feld-man SA, Scurr CF, Paton W, eds. Mechanisms of Actionof Drugs in Anaesthetic Practice. 2nd ed. London: Ed-ward Arnold;1993:83-128.

5. Ladd LA, Copeland SE, Mather LE. Neurally-medi-ated cardiotoxicity of local anesthetics: Direct effectof seizures or of local anesthetics? Anesthesiology2003;98:1296-1297.

6. Mather LE, Duke CC, Ladd LA, Copeland SE, Gal-lagher G, Chang DH. Direct cardiac effects of coro-nary site-directed thiopental and its enantiomers: Acomparison to propofol in conscious sheep. Anesthe-siology 2004;101:354-364.

7. Tucker GT, Mather LE. Pharmacology of local anaes-thetic agents. Pharmacokinetics of local anaestheticagents. Br J Anaesth 1975;47(suppl):213-224.

8. Feldman HS, Hartvig P, Wiklund L, Doucette AM,Antoni G, Gee A, Ulin J, Langstrom B. Regionaldistribution of 11C-labeled lidocaine, bupivacaine,and ropivacaine in the heart, lungs, and skeletalmuscle of pigs studied with positron emission tomog-raphy. Biopharm Drug Dispos 1997;18:151-164.

9. Huang YF, Upton RN, Runciman WB. IV bolus ad-ministration of subconvulsive doses of lignocaine toconscious sheep: Relationships between myocardialpharmacokinetics and pharmacodynamics. Br J An-aesth 1993;70:556-561.

0. Santos AC, DeArmas PI. Systemic toxicity oflevobupivacaine, bupivacaine, and ropivacaine dur-ing continuous intravenous infusion to nonpregnantand pregnant ewes. Anesthesiology 2001;95:1256-1264.

1. Lefrant JY, de La Coussaye JE, Ripart J, et al. The

comparative electrophysiologic and hemodynamic

4

4

4

4

4

4

4

4

5

5

5

5

5

566 Regional Anesthesia and Pain Medicine Vol. 30 No. 6 November–December 2005

effects of a large dose of ropivacaine and bupivacainein anesthetized and ventilated piglets. Anesth Analg2001;93:1598-1605.

2. Aya AG, de la Coussaye JE, Robert E, et al. Compar-ison of the effects of racemic bupivacaine, levobupi-vacaine, and ropivacaine on ventricular conduction,refractoriness, and wavelength: An epicardial map-ping study. Anesthesiology 2002;96:641-650.

3. McCartney CJ, Murphy DB, Iagounova A, Chan VW.Intravenous ropivacaine bolus is a reliable marker ofintravascular injection in premedicated healthy vol-unteers. Can J Anaesth 2003;50:795-800.

4. Owen MD, Gautier P, Hood DD. Can ropivacaine andlevobupivacaine be used as test doses during regionalanesthesia? Anesthesiology 2004;100:922-925.

5. Kietzmann D, Foth H, Geng WP, Rathgeber J, Gun-dert-Remy U, Kettler D. Transpulmonary dispositionof prilocaine, mepivacaine, and bupivacaine in hu-mans in the course of epidural anaesthesia. Acta An-aesth Scand 1995;39:885-890.

6. Sharrock NE, Mather LE, Go G, Sculco TP. Arterialand pulmonary arterial concentrations of the enan-tiomers of bupivacaine after epidural injection in el-derly patients. Anesth Analg 1998;86:812-817.

7. Ohmura S, Sugano A, Kawada M, Yamamoto K.Pulmonary uptake of ropivacaine and levobupiva-

caine in rabbits. Anesth Analg 2003;97:893-897.

8. Tucker GT. Is plasma binding of local anestheticsimportant? Acta Anaesth Belg 1988;39:147-150.

9. Veering BT, Burm AGL, Feyen HM, Olieman W,Souverijn JHM, Van Kleef JW. Pharmacokinetics ofbupivacaine during postoperative epidural infusion:Enantioselectivity and role of protein binding. Anes-thesiology 2002;96:1062-1069.

0. Groban L, Deal DD, Vernon JC, James RL, Butter-worth J. Cardiac resuscitation after incremental over-dosage with lidocaine, bupivacaine, levobupivacaine,and ropivacaine in anesthetized dogs. Anesth Analg2001;92:37-43.

1. Rutten AJ, Mather LE, Plummer JL, Henning EC.Postoperative course of plasma protein binding oflignocaine, ropivacaine and bupivacaine in sheep.J Pharm Pharmacol 1992;44:355-358.

2. Coyle DE, Denson DD, Thompson GA, Myers JA,Arthur GR, Bridenbaugh PO. The influence of lacticacid on the serum protein binding of bupivacaine:Species differences. Anesthesiology 1984;61:127-133.

3. Mather LE, Thomas J. Bupivacaine binding to plasmaprotein fractions. J Pharm Pharmacol 1978;30:653-654.

4. Mather LE, Ladd LA, Chang DH-T, Copeland SE.The effects of imposed acid-base derangement onthe cardio-activity and pharmacokinetics of bupiv-acaine and thiopental. Anesthesiology 2004;100:

1457-1468.