Local Anesthetic-Induced Cardiac Toxicity: A Survey ofContemporary Practice Strategies Among AcademicAnesthesiology Departments

William Corcoran, MD*

John Butterworth, MD†

Robert S. Weller, MD*

Jonathan C. Beck, MD*

J. C. Gerancher, MD*

Timothy T. Houle, PhD*

Leanne Groban, MD

Though new local anesthetics (LA), effective test-dosing, and new regionalanesthetic techniques may have improved the safety of regional anesthesia, theoptimal management plan for LA-induced cardiac toxicity remains uncertain.Accordingly, we evaluated current approaches to LA cardiotoxicity among aca-demic anesthesiology departments in the United States. A 19-question surveyregarding regional anesthesia practices and approaches to LA cardiac toxicity wassent to the 135 academic anesthesiology departments listed by the Society ofAcademic Anesthesiology Chairs-Association of Anesthesiology Program Direc-tors. Ninety-one anonymously completed questionnaires were returned, at aresponse rate of 67%. The respondents were categorized into groups according tothe number of peripheral nerve blocks (PNBs) performed each month: �70 PNBs(38%), 51–70 PNBs (13%), 31–50 PNBs (20%), 11–30 PNBs (23%), and �10 PNBs(6%). Anesthesia practices administering �70 PNBs were 1.7-times more likely touse ropivacaine (NS), 3.9-times more likely to consider lipid emulsion infusions forresuscitation (P � 0.008), and equally as likely to have an established plan for useof invasive mechanical cardiopulmonary support in the event of LA cardiotoxicity(NS) than low-PNB volume centers. We conclude that there are differences in themanagement and preparedness for treatment of LA toxicity among institutions, butthe safety implications of these differences are undetermined.(Anesth Analg 2006;103:1322–26)

Improved regional anesthesia techniques and lesstoxic local anesthetics (LA) appear to be associatedwith a declining number of reported episodes ofsevere LA toxicity over the past quarter century. Thecurrent incidence of clinically important LA toxicity inadults may be subject to under-reporting, but esti-mates range from 7.5 to 20 per 10,000 peripheral nerveblocks (PNBs) and about 4 per 10,000 epidurals (1).Before the reports of cardiac arrest from labor epi-durals with 0.75% bupivacaine and etidocaine (2) andthe “black box” warning in the package insert (for theUnited States) of bupivacaine in 1983 (3), the rate ofLA systemic toxicity with epidurals ranged from 20 to320 per 10,000 (4–7). Although rare, LA toxicity maybe lethal, and clinicians should formulate provisions

for a rapid and educated response to such events.Unfortunately, despite more than 20 yr of animalresearch and case reports of potentially effective treat-ments, there appears to be no consensus for optimummanagement of catastrophic LA toxicity.

The primary aim of this survey was to determinehow U.S. academic anesthesiology departments pre-pare for and treat LA cardiotoxic events, and second-arily, to assess variation in practice patterns in PNBperformance. We hypothesized that we would en-counter a wide range of practice patterns and prepa-ration for treatment of toxicity, and that departmentswhere larger numbers of PNBs are performed wouldhave responses more in line with current literature.

METHODSThe 19-question survey with return envelope was

mailed to the chairs of the 135 academic anesthesiol-ogy departments listed by the Society of AcademicAnesthesiology Chairs-Association of AnesthesiologyProgram Directors. Questions were designed to iden-tify preferred LAs and typical dosage, monitoringduring PNB, level of training of those performingPNBs, location where PNBs were performed, volumeof PNBs performed, and choice of pharmacologic resus-citation used in the event of a cardiac complication (Fig.1). We also asked whether lipid emulsion infusionsand/or mechanical cardiopulmonary support would be

From the *Department of Anesthesiology, Wake Forest Univer-sity School of Medicine, Winston-Salem, North Carolina; and †De-partment of Anesthesia, Indiana University School of Medicine,Indianapolis, Indiana.

Accepted for publication August 3, 2006.Presented in part at the 2006 Annual Meeting of the International

Anesthesia Research Society, San Francisco, CA.Address correspondence to Leanne Groban, MD, Department of

Anesthesiology, Wake Forest University School of Medicine, Medi-cal Center Boulevard, Winston-Salem, NC 27157-1009. Addresse-mail to lgroban@wfubmc.edu. Reprints will not be available fromthe author.

Copyright © 2006 International Anesthesia Research SocietyDOI: 10.1213/01.ane.0000242515.03653.bb

Vol. 103, No. 5, November 20061322

used if initial attempts at pharmacologic resuscitationfailed. We asked that the department chair, the directorof regional anesthesia, or the director of acute painmanagement complete the survey. No follow-up surveywas sent to nonresponders.

Institutions were categorized by the number ofPNBs performed per month, and we arbitrarily de-fined high-volume centers as those performing �70per month, and low-volume as �70 month. Responseswere analyzed relative to the number of proceduresperformed in a responder’s medical setting using �2.Care was taken to artificially dichotomize the proce-dure volume categories so that �5 responses wereexpected in each cell. Odds ratios with 95% confidenceintervals (CI) were used to index the effect size of theobserved differences. The �-value was set at 0.05. Allanalyses were conducted using SPSS 13.0 (SPSS,Chicago, IL).

RESULTSSurvey Response and PNB Volume

Sixty-seven percent of surveys were completed andreturned. The distribution of the number of PNBs per-formed monthly was: �70 PNBs (38%), 51–70 PNBs(13%), 31–50 PNBs (20%), 11–30 PNBs (23%), and �10PNBs (6%). PNBs were performed by trainees (residentsor fellows) 88% of the time. At 9% of the programs, onlystaff anesthesiologists performed PNBs. The location ofblock administration and supervision included the pre-operative holding area (50%), a dedicated regional anes-thesia induction area (30%), and the operating room

(24%). Some programs reported that PNBs wereperformed in multiple locations.

LA Choice and DosageA substantial variation was observed in preferred

long-acting LAs. Although 19% of respondents selectedmultiple drugs, the preferred long-acting LA was eitherbupivacaine (55% of respondents) or ropivacaine (43%)(P � 0.26) (Fig. 2A). Though most centers preferredbupivacaine, there was a slight trend for high-volumeinstitutions to choose ropivacaine, odds ratio � 1.7 (95%CI 0.70–3.9, P � 0.24).

The choice of intermediate-acting LAs reflectedsimilar variability. The two most cited single drugswere mepivacaine (50%) and lidocaine (32%), with atrend for centers to prefer mepivacaine, �2 � 3.46, P �0.06 (Fig. 2B).

When performing an infraclavicular or axillary blockwith bupivacaine, most institutions chose doses of125–200 mg (71%). Few centers used �200 mg (10%) or�125 mg (19%). High-volume centers were 7.5-timesmore likely to use �200 mg (95% CI 1.4–40.3, P � 0.02).When performing infraclavicular or axillary blocks withropivacaine, most institutions chose 125–200 mg (66%).

Monitoring During Block PlacementStandard monitoring during PNB varied among

institutions. There was no apparent relationship be-tween monitoring procedures and the number ofPNBs performed per month or the level of training of

Figure 1. The questionnaire.

Vol. 103, No. 5, November 2006 © 2006 International Anesthesia Research Society 1323

the practitioner performing the block. Standard moni-toring consisted of pulse oximetry, noninvasive arte-rial blood pressure (BP), and electrocardiogram (ECG)for 69% of centers, while 15% used only pulse oxim-etry and BP, and 16% used only pulse oximetry whileperforming regional blocks.

Resuscitation PracticeCenters stored emergency drugs variously in a

“code blue” cart (63%), designated regional anesthesialock box (26%), or another location (11%) for emer-gency treatment of LA toxicity. In response to ventric-ular tachycardia from presumed bupivacaine toxicity,59% would use amiodarone, while 19% would choosebretylium, 13% esmolol, and 2% would use lidocaine.Low-volume centers, defined as �30 PNBs for thisanalysis only, were 3.6-times (95% CI 1.3–10.3) morelikely to choose bretylium for bupivacaine-induced ven-tricular tachycardia than high-volume centers (40% vs16%, P � 0.016).

For an episode of hypotension, 87% of respondentswould choose ephedrine and/or phenylephrine fol-lowed by epinephrine (8%) or vasopressin (3%). Forpersistent hypotension 71% of programs would

choose epinephrine, 20% ephedrine and/or phenyl-ephrine, and 11% vasopressin.

Lipid Emulsion: Choice and AvailabilityWith respect to the potential use of lipid emulsion

infusion to treat bupivacaine toxicity, 74% of centerswould not use it, and 26% would consider its use.High-volume centers were 3.9-times more likely (95%CI 1.4–10.6) than low-volume centers to use lipidinfusion for bupivacaine toxicity (44% vs 17%, P �0.008) (Fig. 3). Of those considering the use of lipidinfusion, storage location was the operating roompharmacy (39%), the hospital pharmacy (35%), thecode cart/lock box (22%), or a drug-dispensing ma-chine (4%). It could be obtained in �10 min for 59%,10–30 min for 26%, and �30 min for 15% of thecenters.

Mechanical Circulatory SupportFor failed resuscitation, 59% of respondents have

no established plan for mechanical cardiopulmonarysupport, with no difference by PNB volume (Fig. 4).For those with an established plan, most often it

0

25

50

75

100

High volumecenters

Low volumecenters

Use of Lipid Emulsion

noyes

Figure 3. Preferences for lipid emulsion use in the event oflocal anesthetic-induced cardiac toxicity. High-volume cen-ters are more likely to use lipid emulsion for the treatment ofbupivacaine toxicity than low-volume centers, �2(1) � 7.52,P � 0.008. Data represent percentage of respondents bycategory.

0

20

40

60

80

100

High volumecenters

Low volumecenters

No CP SupportCP Support

Figure 4. Presence of an established plan for mechanicalcardiopulmonary (CP) support in the event of failed resus-citation. High- and low-volume centers were similarly un-likely to have an established plan for CP support, �2(1) �2.74, P � 0.098. Data represent percentage of respondents bycategory.

0102030405060

%

Levo Bup Rop Mixtures Other

Preferred Long-acting Local Anesthetic

0

10

20

30

40

50

%

Lid Mep Mixtures Other

Preferred Intermediate-acting Local Anesthetic

A

B

Figure 2. (A) The preferred long-acting local anesthetic. Thelong-acting drug of choice is bupivacaine, although there isa trend for high-volume institutions to choose ropivacaineversus low-volume institutions. Using logistical regressionwe found that high-volume institutions are 1.7-times morelikely to choose ropivacaine (P � 0.24). Levo � levobupiva-caine; Bup � bupivacaine, Rop � ropivacaine. (B) Thepreferred intermediate-acting local anesthetic. The prefer-ence for mepivacaine (Mep) over lidocaine (Lid) nearlyreached statistical significance (P � 0.06).

1324 LA-Induced Toxicity: Contemporary Practice Strategies ANESTHESIA & ANALGESIA

included cardiopulmonary bypass (CPB) (89%), in-traaortic balloon pump counterpulsation (22%), andextracorporeal membrane oxygenation (14%). Car-diothoracic surgeons could respond in �10 min for33%, 10 –30 min for 51%, and �30 min for 15% ofrespondents.

DISCUSSIONOur results show a wide range of current practice

patterns for PNB in U.S. academic centers, and vari-ability in nearly all aspects of treatment strategies formanaging severe LA toxicity. Much of the observedvariability (PNB performance site, monitoring, pre-ferred intermediate-duration LA, resuscitation drugs,and so forth) was independent of PNB volume. Con-versely, the finding that high-PNB volume centersshowed a trend to favor ropivacaine, and to be signifi-cantly more likely to consider lipid emulsion infusionand less likely to list bretylium as a treatment forsevere LA toxicity than low-volume centers (�30PNBs) merits closer examination. It is possible thatprocess variation could lead to poorer outcome fromrelatively rare, but catastrophic, toxic reactions (8).

LA ChoiceThe findings of Aberg (9) and Akerman et al. (10), that

S-bupivacaine requires larger doses than R-bupivacaineto produce convulsions and mortality, led to the devel-opment of levobupivacaine and its S-enantiomer homo-logue ropivacaine (11,12). Subsequent animal studieshave demonstrated differences in mortality or inabilityto resuscitate between racemic bupivacaine-treated dogs(50% mortality) and ropivacaine-treated dogs (10% mor-tality) after anesthetized dogs received continuous,escalating infusions of LA. (13) Still, severe toxicity inpatients has been reported with ropivacaine (14–17), butropivacaine toxicity may be more amenable to treatmentthan reactions to bupivacaine, and no ropivacaine-induced deaths have been reported.

Despite the apparent increased safety margin withropivacaine, our study demonstrates that bupivacaineis widely, and even exclusively, used in many institu-tions. Factors other than safety during rare toxicevents, such as acquisition cost and perceived qualityor duration of PNB, may influence LA choice, butthese factors were not evaluated in our survey.

Monitoring During PNBBP, three-lead ECG, and pulse-oximetry are among

the standard monitors endorsed by the American Soci-ety of Anesthesiologists for use during all anesthetics(www.asahq.org/publicationsAndServices/standards/02.pdf), but only 69% of respondents used all threemonitors during PNB. Although the American Societyof Anesthesiologists Standards acknowledge thatpractitioners may circumvent standard monitoringwhen indicated (and documented), both BP and ECGmonitoring may supplement pulse oximetry to dem-onstrate early signs of LA toxicity (18). Specifically, the

ECG is necessary to characterize arrhythmias shouldsevere toxicity occur.

Treatment Strategies for LA ToxicityThere is no consensus strategy for how best to treat

severe LA toxicity. Indeed, there is no ethical way toconduct meaningful clinical trials in this area. Oursurvey showed, however, that high-PNB volume in-stitutions are more likely to choose amiodarone (asopposed to lidocaine or bretylium) for ventriculartachycardia. The stated preference of some centers forbretylium is particularly alarming, since this drug isno longer available in North America or included inAdvanced Cardiac Life Support Guidelines.

Lipid Emulsion Infusion and Mechanical CirculatorySupport for Severe LA Toxicity

We presumed that all centers would institute ven-tilatory support and chest compression per AdvancedCardiac Life Support Guidelines for complete cardio-pulmonary collapse after PNB, and we did not includesurvey questions for those interventions. Lipid emul-sion infusion may be a consideration after institutionof cardiopulmonary resuscitation in such severe toxicevents. Animal studies by Weinberg and co-workers(19–21) have shown lipid emulsion infusions to beeffective as a means of resuscitation from bupivacainetoxicity after chest compression (rat) or cardiac massage(dog) failed to resuscitate the animals. The mechanism oflipid resuscitation remains under investigation, but maybe due to the migration of amphiphilic LA moleculesfrom binding sites in the heart into the plasma-bornlipid. Dog studies show that after circulatory collapsesecondary to bupivacaine toxicity, lipid treatment in-creased the survival rate from 0% in the control group to100% in the lipid-treated group (19). Enthusiasm forlipid treatment must be tempered by the absence ofpublished reports of successful use of lipid to resuscitatehumans with LA toxicity. Results of this survey showthat the high-volume centers are significantly more re-ceptive to the use of lipid emulsion for refractory LAcardiotoxicity than the low-volume centers. One casereport documents the successful use of lipid for bupiv-acaine cardiac toxicity (22).

CPB has resulted in survival in human cases ofbupivacaine cardiovascular toxicity. Bupivacaine mayhave a prolonged duration of action in cases ofcardiovascular collapse; thus, patients may requirecardiopulmonary support for �45 min (23–25). In theevent of failed pharmacologic resuscitation, our re-sults show that only 41% of centers have a plan formechanical cardiopulmonary support (Fig. 4). Mostinstitutions report that cardiothoracic surgeons couldarrive in 30 min or less, although the most importantvariable, time from identification of the need untilonset of bypass, was not ascertained. There are no

Vol. 103, No. 5, November 2006 © 2006 International Anesthesia Research Society 1325

documented reports of the successful use of extracor-poreal membrane oxygenation for LA toxicity; never-theless, if available, it may be instituted more rapidlythan CPB.

Study LimitationsThis study has a number of limitations. The inabil-

ity to show differences between high and low-volumecenters can indicate that differences were not presentor that there was simply a lack of statistical power todetect them. Although a response rate of 67% is good,our findings should be interpreted cautiously becauseof limited statistical power, which enabled us to detectonly moderate to large effect sizes in the differencesbetween categories (26). In addition, only academicanesthesiology departments were surveyed, and theresults may not reflect the much broader practice ofregional anesthesia in non-academic departments.Furthermore, there may be a range of practice withinresponding departments not reflected by the survey,and an assumption was made that a department chairwould provide the same responses as the director ofregional anesthesia in a given institution, which maynot be accurate. This limitation, however, would mostlikely lead to our survey results under-estimating thevariability in PNB practice and preparedness for LAtoxicity.

Our finding of wide variability in preparedness forLA toxicity and lack of consensus for treatment isnoteworthy. Lessons learned from malignant hyper-thermia show that well-established treatment guide-lines could allow for early intervention to effectivelysave lives (27). LA cardiac toxicity is likely at least ascommon as malignant hyperthermia, and potentiallyas frequently fatal. Our survey results support effortsto determine and disseminate optimal treatment strat-egies for severe LA toxicity.

REFERENCES

1. Mulroy MF. Systemic toxicity and cardiotoxicity from localanesthetics: incidence and preventive measures. Reg AnesthPain Med 2002;27:556–61.

2. Albright GA. Cardiac arrest following regional anesthesia withetidocaine or bupivacaine. Anesthesiology 1979;51:285–7.

3. Horlocker TT, Wedel DJ. Local anesthetic toxicity—does productlabeling reflect actual risk? Reg Anesth Pain Med 2002;27:562–7.

4. Blundell AE, Bodell B, Andorko JE, et al. Clinical evaluation ofdrugs used in obtaining lumbar epidural anesthesia. Anesthe-siology 1955;16:386–93.

5. Bonica JJ, Backup PH, Anderson CE, et al. Peridural block: analysisof 3,637 cases and a review. Anesthesiology 1957;18:723–84.

6. Moore DC. Toxic effects of local anesthetics. JAMA 1978;240:434.7. Kenepp NB, Gutsche BB. Inadvertent intravascular injections

during lumbar epidural anesthesia. Anesthesiology 1981;54:172–3.8. Shewhart WA. Statistical methods from the viewpoint of quality

control. Mineola, NY: Dover, 1986.9. Aberg G. Toxicological and local anaesthetic effects of optically

active isomers of two local anaesthetic compounds. Acta Phar-macol Toxicol (Copenh) 1972;31:273–86.

10. Akerman B, Hellberg IB, Trossvik C. Primary evaluation of thelocal anaesthetic properties of the amino amide agent ropiva-caine (LEA 103). Acta Anaesthesiol Scand 1988;32:571–8.

11. Mather LE, Chang DH. Cardiotoxicity with modern local anaes-thetics: is there a safer choice? Drugs 2001;61:333–42.

12. Groban L. Central nervous system and cardiac effects fromlong-acting amide local anesthetic toxicity in the intact animalmodel. Reg Anesth Pain Med 2003;28:3–11.

13. Groban L, Deal DD, Vernon JC, et al. Cardiac resuscitation afterincremental overdosage with lidocaine, bupivacaine, levobupi-vacaine, and ropivacaine in anesthetized dogs. Anesth Analg2001;92:37–43.

14. Chazalon P, Tourtier JP, Villevielle T, et al. Ropivacaine-inducedcardiac arrest after peripheral nerve block: successful resuscita-tion. Anesthesiology 2003;99:1449–51.

15. Huet O, Eyrolle LJ, Mazoit JX, Ozier YM. Cardiac arrest afterinjection of ropivacaine for posterior lumbar plexus blockade.Anesthesiology 2003;99:1451–3.

16. Klein SM, Pierce T, Rubin Y, et al. Successful resuscitation afterropivacaine-induced ventricular fibrillation. Anesth Analg2003;97:901–3.

17. Polley LS, Santos AC. Cardiac arrest following regional anesthesiawith ropivacaine: here we go again. Anesthesiology 2003;99:1253–4.

18. Takahashi S, Tanaka M, Toyooka H. The efficacy of hemody-namic and T-wave criteria for detecting intravascular injectionof epinephrine test dose in propofol-anesthetized adults. AnesthAnalg 2002;94:717–22.

19. Weinberg GL. Current concepts in resuscitation of patients withlocal anesthetic cardiac toxicity. Reg Anesth Pain Med 2002;27:568–75.

20. Weinberg G, Ripper R, Feinstein DL, Hoffman W. Lipid emul-sion infusion rescues dogs from bupivacaine-induced cardiactoxicity. Reg Anesth Pain Med 2003;28:198–202.

21. Weinberg GL, VadeBoncouer T, Ramaraju GA, et al. Pretreatmentor resuscitation with a lipid infusion shifts the dose-response tobupivacaine-induced asystole in rats. Anesthesiology 1998;88:1071–5.

22. Rosenblatt MA, Abel M, Fischer GW, et al. Successful use of a20% lipid emulsion to resuscitate a patient after a presumedbupivacaine-related cardiac arrest. Anesthesiology 2006;105:217–8.

23. Soltesz EG, van Pelt F, Byrne JG. Emergent cardiopulmonarybypass for bupivacaine cardiotoxicity. J Cardiothorac VascAnesth 2003;17:357–8.

24. Tsai MH, Tseng CK, Wong KC. Successful resuscitation of abupivacaine-induced cardiac arrest using cardiopulmonary by-pass and mitral valve replacement. J Cardiothorac Anesth1987;1:454–6.

25. Long WB, Rosenblum S, Grady IP. Successful resuscitation ofbupivacaine-induced cardiac arrest using cardiopulmonary by-pass. Anesth Analg 1989;69:403–6.

26. Cohen JC. Statistical power analysis for the behavioral sciences,2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.

27. Halliday NJ. Malignant hyperthermia. J Craniofac Surg 2003;14:800–2.

1326 LA-Induced Toxicity: Contemporary Practice Strategies ANESTHESIA & ANALGESIA