benzyl alcohol as an alternative local anesthetic

Lance Wilson, MD Steven Martin, MD

From the Department of EmergencyMedicine, Mount Sinai MedicalCenter and Case Western ReserveUniversity School of Medicine,Division of Internal Medicine,Cleveland, OH.

Received for publication October 31, 1997. Revisions received November 4 and November 24, 1998. Accepted forpublication December 10, 1998.

Presented at the Society for AcademicEmergency Medicine Annual Meeting,Denver, CO, May 1996.

Address for reprints: Lance Wilson,MD, Department of EmergencyMedicine, The Mount Sinai MedicalCenter, One Mt Sinai Drive,Cleveland, OH 44106; 216-421-4022, fax 216-421-4244.

Copyright © 1999 by the AmericanCollege of Emergency Physicians.

0196-0644/99/$8.00 + 047/1/97291

O R I G I N A L C O N T R I B U T I O N

M A Y 1 9 9 9 3 3 : 5 A N N A L S O F E M E R G E N C Y M E D I C I N E 4 9 5

Benzyl Alcohol as an Alternative Local

Anesthetic

Study objectives: Benzyl alcohol has been used as a localanesthetic for brief superficial skin procedures; however, itsefficacy for long-term cutaneous anesthesia has not beenestablished. We sought to compare the cutaneous anestheticeffects of benzyl alcohol with epinephrine with the effects oflidocaine with epinephrine and with placebo.

Methods: This study was a prospective, randomized, double-blind, placebo-controlled clinical trial of 30 healthy paid adultvolunteers. Subjects received 1-mL intradermal injections ofbenzyl alcohol .9% with 1:100,000 epinephrine, lidocaine 1%with 1:100,000 epinephrine, and physiologic saline solutionwithout benzyl alcohol as placebo in a randomized, double-blind fashion. Pain on injection and degree of anesthesia at 5,15, 30, and 45 minutes was assessed with a 10-cm gradedvisual analog scale (VAS). Statistical significance was deter-mined by repeated measures ANOVA between groups with aNeuman-Keuls test for post hoc comparison of means andStudent’s t test for paired means.

Results: Benzyl alcohol was 48% less painful on injectionthan placebo (P<.008) and 42% less painful on injection thanlidocaine with epinephrine (P<.05). Lidocaine with epinephrineand placebo were equally painful on injection. After the 5-minute measurement, benzyl alcohol provided significantly bet-ter anesthesia than placebo during the remaining observationperiod (VAS score 48%, 49%, and 51% decreased from base-line at 15, 30, and 45 minutes, respectively, all P<.02 versusplacebo). However, benzyl alcohol provided less effective anes-thesia than lidocaine with epinephrine (VAS score 72%, 76%,84%, and 88% decreased from baseline at 5, 15, 30, and 45minutes, respectively, all P<.001 versus placebo) throughoutmost of the observation period.

Conclusion: Benzyl alcohol with epinephrine provides pro-longed cutaneous anesthesia, although it is not as effective aslidocaine with epinephrine. However, benzyl alcohol is signifi-

ber generator from commercially available software(Labview, National Instruments Corporation, Austin,TX) to prepare the randomization/identification key.We randomly assigned both the location of injectionand order of drug administration. A single investigatoradministered study drugs in a double-blind fashion.We injected 1 mL of each of the study drugs intrader-mally with a 30-gauge needle at an injection rate of .1mL/s. Each injection sequence began near the wrist onthe volar aspect of the forearms and proceeded proxi-mally, allowing at least a 3-cm distance between injec-tions. The time interval between injections was 60 sec-onds. All study drugs were at room temperature whengiven. We assessed baseline pain with pinprick, painon injection, and pain to pinprick at 5, 15, 30, and 45minutes after anesthetic injection with a 10-cm visualanalog scale (VAS) with numerical gradations.9,10 Thewords “mild pain, moderate pain, severe pain” were alsolisted below the scale at the left end, under the midpoint,and at the right end of the scale, respectively. Beforeeach measurement, we asked subjects if they werehaving any reactions to study drugs and to report anydelayed problems that they experienced. The studyhad 80% power to detect a 13-mm difference in VASscores, assuming an SD of 25 mm and a value of Pless than .05. These assumptions were based on previ-ous work examining minimally significant clinicaldifferences in pain scores and previous work on localanesthetics.11-13

For evaluation of pain to pinprick during the mea-surement intervals, we determined statistical signifi-cance by repeated measures ANOVA among the 3group means with a Neuman-Keuls test for post hoccomparisons. We confirmed ANOVA assumptions withLevene’s test for homogeneity of variances. We alsoperformed a Kruskal-Wallis ANOVA confirming statis-tical significance between group means for pain to pin-prick, also supporting the use of the parametricANOVA. A 1-way ANOVA using location of injection asthe grouping variable also was performed. This wasnonsignificant, therefore allowing us to collapse thatgrouping variable and use the ANOVA. We usedStudent’s t test for paired means to compare pain oninjection of the study drugs.

R E S U L T S

The 30 subjects ranged from 22 to 52 years of age (meanage 34.4 years, SD 7.7 years). Women comprised 53% ofthe subjects.

B E N Z Y L A L C O H O L A S A N A L T E R N A T I V E L O C A L A N E S T H E T I CWilson & Martin

4 9 6 A N N A L S O F E M E R G E N C Y M E D I C I N E 3 3 : 5 M A Y 1 9 9 9

cantly less painful on injection than lidocaine with epinephrine,and it may offer an alternative for local anesthesia.

[Wilson L, Martin S: Benzyl alcohol as an alternative localanesthetic. Ann Emerg Med May 1999;33:495-499.]

I N T R O D U C T I O N

Emergency physicians are sometimes faced with theproblem of patients claiming allergy to lidocaine or to “allthe -caine” anesthetics, making selection of a typicalamide or ester local anesthetic problematic.Diphenhydramine has been a proposed alternative localanesthetic; however, it has been shown to be significantlymore painful on injection than lidocaine and may have ashorter duration of action.1-3 Benzyl alcohol is an opiumalkaloid used as a popular antiseptic in hundreds of prod-ucts for its bacteriostatic properties.4 It is commonlyadded to physiologic saline solution (.9% benzyl alco-hol–bacteriostatic saline solution) and ubiquitouslyavailable. The anesthetic effect of benzyl alcohol was firstdescribed in 1918, and its safety has been clearly demon-strated.5,6 Benzyl alcohol has been used as a local anes-thetic for brief superficial skin procedures; however, itsefficacy for long-term dermal anesthesia has not beenestablished.7,8 The purpose of this study is to comparethe cutaneous anesthetic effects of benzyl alcohol withepinephrine to buffered lidocaine with epinephrine andplacebo.

M A T E R I A L A N D M E T H O D S

The institutional review board of our institution approvedthis study. Advertisements posted at our institutionrecruited 30 healthy paid adult volunteers, who were allhospital employees and the majority had a medical back-ground. We paid each volunteer $30. The first 30 respon-dents who were willing to participate in the study and didnot have exclusion criterion participated. As such, we didnot randomly select them. We excluded volunteers if theywere pregnant, breast-feeding, or had a history of allergyto any local anesthetics or study drugs.

Each person received each of the 3 study drugs: (1)benzyl alcohol .9% with epinephrine (1:100,000 dilu-tion), (2) 1% lidocaine with epinephrine (1:100,000dilution) buffered with sodium bicarbonate (1 mL of8.4% solution in 9 mL of 1% lidocaine), and (3) physio-logic saline solution without benzyl alcohol as placebo.The hospital pharmacy prepared tuberculin syringes,labeled with a numerical code. We used a random num-



D I S C U S S I O NThis study demonstrates that benzyl alcohol withepinephrine can produce prolonged cutaneous anesthe-sia to pinprick when compared with placebo. Benzyl alco-hol with epinephrine was significantly less painful oninjection than buffered lidocaine or placebo. However,benzyl alcohol was not as effective an anesthetic asbuffered lidocaine with epinephrine.

Although rare, anaphylaxis and other significantadverse reactions to lidocaine have been reported.14-16

Although most severe or anaphylactic reactions to localanesthetics are not truly allergic,17 most clinicians wouldnot risk anaphylaxis. The methylparaben additive used asa preservative in lidocaine for local anesthesia is the mostcommon cause of lidocaine “allergy.”17,18 Methylparabenalso may be found in ester anesthetics preparations, mak-ing the substitution of an ester for an amide problem-atic.19 A substitute anesthetic such as the antihistaminediphenhydramine is a further option. However, diphen-hydramine is not as effective as the other classes of localanesthetics, and it is painful on infiltration.1,2,3,20

Further, significant local reactions have been reportedwith diphenhydramine infiltration, and the manufacturerstates that it should not be used as a local anesthetic.2,21

Given the difficulty in determining true allergy to localamide anesthetics and the limited alternatives, anotheroption for local anesthesia in the ED setting could be useful.

There was no difference in baseline sensation to pin-prick for any of the study drugs (VAS scores 2.6, 2.5, and2.2 for placebo, lidocaine, and benzyl alcohol, respectively,P=.52).

Benzyl alcohol with epinephrine was 48% less painfulon injection than placebo (P<.008) and 42% less painfulon injection than lidocaine with epinephrine (P<.05).Lidocaine with epinephrine and placebo were equallypainful on injection (P=.4) (Figure 1).

By repeated measures analysis of variance, the grouptime interaction was significant among the 3 study drugsthroughout the observation period (P<.001). After the 5-minute measurement, benzyl alcohol provided signifi-cantly better anesthesia than placebo during the remain-ing observation period (VAS score 48%, 49%, and 51%decreased from baseline at 15, 30, and 45 minutes,respectively, all P<.02 versus placebo); however, it pro-vided less effective anesthesia than lidocaine withepinephrine during most of the observation period (VASscore 72%, 76%, 84%, and 88% decreased from baselineat 5, 15, 30, and 45 minutes, respectively, all P<.001 ver-sus placebo). At the 15-minute observation, the differ-ences between sensation to pinprick in the benzyl alcoholand lidocaine with epinephrine groups did not differ sig-nificantly (P=.07) (Figure 2). Lidocaine with epinephrineprovided significantly better anesthesia than placeboduring the entire observation period. We did not observeadverse reactions (ie, local skin reaction, allergic reac-tions, infections, or sedation) to any of the 3 study drugs.

Figure 1.Pain on injection of benzyl alcohol with epinephrine (BA), lido-caine with epinephrine (Lido), and placebo. Data are presentedas VAS score in centimeters, ±SD. *P<.05 from control group.

Figure 2.Pain to pinprick during study period in the 3 study groups. Dataare presented as % baseline VAS score. *P<.05 from the controlgroup.

alcohol and lidocaine groups. Epinephrine is a vasocon-strictor and could potentially cause some of the analgesiceffects observed with benzyl alcohol with epinephrine.

We observed no local or systemic adverse reactions tobenzyl alcohol or any of the study drugs. The toxicity ofbenzyl alcohol has been extensively studied and has beendetermined to be relatively nontoxic even at large par-enteral doses.6 A 1mL/kg dose of .9% benzyl alcohol (bac-teriostatic saline solution) produced no systemic effectsin anesthetized dogs or monkeys. At this concentration,benzyl alcohol was uniformly lethal at a volume of between88 and 113 mL/kg in dogs.6 The parenteral safety of ben-zyl alcohol seems to have been established, and only 1case report of an allergic dermatitis has been attributed tobenzyl alcohol in man.4,6 In our study, we asked patientsto report any potential adverse effects. Because no formalfollow-up was performed, it is possible that we couldhave missed potential reactions. However, all the subjectswere employees at our institution, and we were informedof no short- or long-term adverse effects related to benzylalcohol with epinephrine.

We used pain to pinprick as 1 of our main outcomevariables. The baseline pain to pinprick measured in thisstudy was reasonably low. This could have been becausethe stimulus itself is not especially painful, as well as thefact that most of our subjects were health care workers atthe authors’ institution, and may have tended to mini-mize their pain. It is also possible that the pinprick stimu-lus was not uniform throughout the study, although a sin-gle investigator provided it.

We also used only a 1-mL injection volume in this study,which was chosen on the basis of previous studies investi-gating the pain on infiltration of various agents. Numerousstudies using pinprick as an outcome variable have usedcomparable or smaller volumes of injection with volumesranging from .1 mL to 2 mL.2,3,23,27 A larger injectionvolume might have caused greater baseline pain on injec-tion, allowing better differentiation between the studydrugs. Furthermore, we cannot exclude the possibilitythat the brief intervals between each injection and pin-prick stimulus may have affected the subsequently mea-sured pain scores. However, we randomized the orderand location of study drug injection, which should haveminimized this limitation.

Todd and Funk11 and Todd et al11,12 have proposed dif-ferences less than 13 to 18 mm or a 23% change in a VASpain score may not be of clinical importance. In this study,the difference between benzyl alcohol and lidocaine onthe VAS score was approximately 9 mm, but benzyl alco-hol was 42% less painful. Thus this probably represents a


4 9 8 A N N A L S O F E M E R G E N C Y M E D I C I N E 3 3 : 5 M A Y 1 9 9 9

Benzyl alcohol is an opium alkaloid with intrinsic anes-thetic action. Its local anesthetic properties have beenreported as early as 1918.5 Because of its bacteriostaticproperties, benzyl alcohol is used in many injectable agents.8

Benzyl alcohol is typically used in concentrations between.5% and 2% and is most commonly found in bacteriostaticsaline solution in a concentration of .9%.4 Bacteriostaticsaline solution is found ubiquitously in health care facilities.Because of its ease of both availability and incorporationat many institutions, we used bacteriostatic saline solutionwith .9% benzyl alcohol in this investigation.

Previous anecdotal evidence has suggested that benzylalcohol is an excellent local anesthetic, but its short dura-tion of action makes it appropriate for only brief procedures,such as placement of intravenous catheters, punch biopsies,or small skin incisions.8,22 One double-blind study showedthat .9% benzyl alcohol provided anesthesia to pinprick,as measured on a 3-point pain scale, comparable to lidocaine;however, adequate anesthesia lasted only 2 to 3 minutes.23

It has been suggested that the addition of epinephrine tobenzyl alcohol will prolong anesthetic properties sufficientlyfor its use in longer procedures, such as laceration repair8;however, a prospective, double-blind fashion has neverdemonstrated this. Benzyl alcohol anesthesia has not beenstudied for repair of traumatic wounds. Benzyl alcohol .9%with epinephrine has been shown less painful on injectionthan lidocaine with epinephrine at volumes of .5 mL in 1nonblinded prospective study7; however, a direct com-parison of the long-term anesthetic effect of benzyl alcoholwith lidocaine was not reported for these subjects.

In this study, we compared benzyl alcohol withepinephrine to buffered lidocaine with epinephrine ondegree of pain on injection. We used buffered lidocainebecause buffering is considered the current “gold standard”to decrease pain on injection of lidocaine. We believe theincreased pain on injection is, at least in part, from theacidic pH of lidocaine, and thus provides one rationale forbuffering.24-27 Benzyl alcohol with epinephrine has a pH ofapproximately 5.3.7 Despite its acidic pH, this and otherstudies have shown benzyl alcohol with epinephrine rela-tively painless on injection.7,8,19,20 In this study, other fac-tors that might affect pain on infiltration, such as the rateof injection, temperature, and needle gauge used to injectthe study drugs, were held constant.28,29

We added epinephrine to the benzyl alcohol to prolongits anesthetic effect, and we used normal saline solutionwithout epinephrine to attempt to use a “true” placebo.However, we could have used epinephrine with normalsaline solution in this investigation, which might havecontrolled for the effects of epinephrine in the benzyl



R E F E R E N C E S1. Ernst AA, Anand P, Nick T, et al: Lidocaine versus diphenhydramine for anesthesia in therepair of minor lacerations. J Trauma 1993;34:354-357.

2. Dire DJ, Hogan DE: Double-blinded comparison of diphenhydramine verses lidocaine as alocal anesthetic. Ann Emerg Med 1993;22:1419-1422.

3. Green SM, Rothrock SG, Gorchynski J: Validation of diphenhydramine as a dermal localanesthetic. Ann Emerg Med 1994;23:1284-1289.

4. Shunes E: Allergic dermatitis to benzyl alcohol in an injectable solution. Arch Derm1984;120:1200-1201.

5. Macht DI: A pharmacological and therapeutic study of benzyl alcohol as a local anes-thetic. J Pharm Exp Therap 1918;11:263-279.

6. Kimura ET, Darby TD, Krause RA, et al: Parental toxicity studies with benzyl alcohol.Toxicol Appl Pharmacol 1971;18:60-68.

7. Lugo-Janer G, Padiel M, Sanchez JL: Less painful alternatives for local anesthesia. JDerm Surg Oncol 1993;19:237-240.

8. Holmes HS: Options for painless local anesthesia. Postgrad Med 1991;89:71-72.

9. Revill SI, Robinson JO, Rosen M, et al: The reliability of a linear analogue for evaluatingpain. Anaesthesia 1976;31:1191-1198.

10. Ho K, Spence J, Murphy MF: Review of pain-measurement tools. Ann Emerg Med 1996;27:427-432.

11. Todd KH, Funk JP: The minimum clinically important difference in physician-assignedvisual analog pain scores. Acad Emerg Med 1996;3:142-146.

12. Todd KH, Funk KG, Funk JP, et al: Clinical significance of reported changes in pain sever-ity. Ann Emerg Med 1996;27:485-489.

13. Ernst AA, Marvez-Valls E, Nick TG, et al: Comparison trial of four injectable anestheticsfor laceration repair. Acad Emerg Med 1996;3:228-233.

14. Holty G, Hood FJC: An anaphylactoid reaction to lidocaine. Dent Pract Dent Rec 1965;15:294-296.

15. Ross NM: Serum-type allergic reactions to local anesthetic injections—Report of twocases. Anesth Prog 1966;13:139-140.

16. Waldman HB, Binkley G: Lidocaine hypersensitivity: Report of case. J Am Dent Assoc1967;74:747-749.

17. Adriani J, Zepernick R: Allergic reactions to local anesthetics. South Med J 1981;74:694-699.

18. Swanson JG: Assessment of allergy to local anesthetic. Ann Emerg Med 1983;12:316-318.

19. Physicians’ Desk Reference. Montvale, NJ: Medical Economics Data Production, 1996:554, 2303.

20. Ernst AA, Marvez-Valls E, Mall G, et al: 1% lidocaine versus 0.5% diphenhydramine forlocal anesthesia in minor laceration repair. Ann Emerg Med 1994;23:1328-1332.

21. Diphenhydramine HCl [package insert]. Morris Plains, NJ: Parke-Davis Co, August 1991.

22. Thomas DV: Saline with benzyl alcohol prevents pain of needle insertion [letter). AnesthAnalg 1984;63:883.

23. Wightman MA, Vaughan RW: Comparison of compounds used for intradermal anesthesia.Anesthesiology 1976;45:687-689.

24. Bartfield: Buffered versus plain lidocaine as a local anesthetic for simple laceration repair.Ann Emerg Med 1990;19:51-53.

25. McKay W, Morris R, Mushlin P: Sodium bicarbonate attenuate pain on skin infiltrationwith lidocaine, with or without epinephrine. Anesth Analg 1987;66:572-574.

26. Christoph RA, Buchanan L, Begalia K, et al: Pain reduction in local anesthetic administra-tion through pH buffering. Ann Emerg Med 1988;17:117-120.

27. Singer AJ, Hollander JE: Infiltration pain and local anesthetic effects of buffered versusplain 1% diphenhydramine. Acad Emerg Med 1995;2:884-888.

28. Mader TJ, Playe SJ, Garb JL: Reducing the pain of local anesthetic infiltration: Warmingand buffering have a synergistic effect. Ann Emerg Med 1994;23:550-554.

29. Brogan GX, Hollander JE, Giarrusso E, et al: Comparison of plain, warmed, and bufferedlidocaine for anesthesia of traumatic wounds. Ann Emerg Med 1995;26:121-125.

30. Stahmer SA, Shofer FS, Marino A, et al: Do Quantitative changes in pain untensity corre-late with pain relief and satisfactions? Acad Emerg Med 1998;5:851-857.

clinical difference in terms of pain on injection (Figure 1).The difference in VAS scores between benzyl alcohol andplacebo were all less than 13 mm, but the percentagedecrease in VAS score from baseline was always greater than40% after 5 minutes, and probably represents a clinicallysignificant difference (Figure 2). Furthermore, pinprick is amildly painful stimulus, and Todd et al determined their cri-terion for a minimally significant clinical difference inpatients with moderate to severe pain. It may be true thatthese criteria for minimally significant differences cannot begeneralized to less painful stimuli. In 1 study, patients whoexperienced less baseline pain had clinically meaningfulimprovement of pain with smaller improvement in painscores than patients with more severe pain.30 The true abil-ity to clinically generalize VAS pain scores is still a matter ofcontroversy.11,12 Furthermore, we used a graded VAS,which may not be completely applicable to the work ofTodd et al. Other investigators have used this type of VAS,and some believe it is equivalent to a nongraded VAS.10,30

Benzyl alcohol did not show a statistically significantdifference in VAS score from placebo 5 minutes after injec-tion, but over time, offered superior cutaneous anesthesia.This was unexpected because benzyl alcohol producedimmediate, but relatively short-lived, anesthesia in all pre-vious reports.8,22,23 The 1 study of benzyl alcohol withepinephrine did not present time of onset of anesthesia.7 Itis possible that epinephrine-induced vasoconstrictioncould have caused a delay in the anesthetic effect of benzylalcohol. We also observed a statistically significant placeboeffect at 5 minutes in this study, which may account for thelack of a statistically significant effect of benzyl alcohol rela-tive to control at this measurement interval. Normal salinesolution may cause mild, short-lived analgesia as a result ofskin distention.

Although lidocaine was a significantly better anestheticthan benzyl alcohol throughout most of the observationperiod, benzyl alcohol may still produce a degree ofanesthesia suitable for laceration repair and other EDprocedures. Given that benzyl alcohol can sustain cuta-neous anesthesia for at least 45 minutes as this studydemonstrated, further investigation is warranted todetermine whether benzyl alcohol can offer an alterna-tive for local anesthesia for wound repair and other EDprocedures.

We thank Stephen Zyzanski, PhD, for his help in statistical preparation, and MikeRitzenthaller for his help in preparing the study drugs.

benzyl alcohol as an alternative local anesthetic

Documents