Pii: s0196-0644(98)70062-9

O R I G I N A L C O N T R I B U T I O N
Randomized Trial of Diphenhydramine VersusBenzyl Alcohol With Epinephrine as an Alternative to Lidocaine Local Anesthesia Joel M Bartfield, MD
Study objectives: We compared the pain of infiltration and
Stacy Weeks Jandreau, MD
anesthetic effects of .9% benzyl alcohol with epinephrine, 1% Nancy Raccio-Robak, RN, MPH
diphenhydramine, and .9% buffered lidocaine.
Received for publication December 28, 1997. Revision Methods: A prospective, randomized, double-blind study compar-
ing benzyl alcohol, diphenhydramine, and lidocaine was carried out on adult volunteers. Each subject received all 3 injections in Address for reprints: Joel M
Bartfield, MD, Department of
a standardized manner. Pain of infiltration was measured on a 100-mm visual analog pain scale and analyzed with a Kruskal- Wallis test. Duration of anesthesia was assessed at 5-minute intervals for a maximum of 45 minutes and compared with the use of survival analysis techniques by a log-rank test. Return of sensation by 45 minutes was evaluated with an exact χ2 test.
All tests were 2-tailed, with significance defined as P<.05.
0196-0644/98/$5.00 + 0
47/1/94122
Results: Thirty subjects were enrolled. The diphenhydramine
median pain score was 55 mm, compared with 12.5 mm for
lidocaine and 5 mm for benzyl alcohol (P=.001). Pairwise compar-
isons showed that all possible combinations were statistically
significant. The 3 anesthetics were different with respect to
duration of anesthesia (P<.001). Pairwise comparisons revealed
a longer duration of anesthesia for lidocaine than for diphenhy-
dramine or benzyl alcohol, but no significant difference was
found between diphenhydramine and benzyl alcohol. Pain sen-
sation returned within the 45-minute study period in only 3 of
30 lidocaine injections, compared with 11 of 30 benzyl alcohol
injections and 19 of 30 diphenhydramine injections (P=.001).
Conclusion: Benzyl alcohol is a better alternative than
diphenhydramine as a local anesthetic for lidocaine-allergic
patients.
[Bartfield JM, Jandreau SW, Raccio-Robak N: Randomized trialof diphenhydramine versus benzyl alcohol with epinephrine asan alternative to lidocaine local anesthesia. Ann Emerg MedDecember 1998;32:650-654.] 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 2 : 6 D E C E M B E R 1 9 9 8
A L T E R N A T I V E S T O L I D O C A I N E A N E S T H E S I A
Bartfield, Jandreau & Raccio-Robak Emergency physicians often administer local anesthetics.
A prospective, randomized, double-blind study compar- Lidocaine is the most commonly used local anesthetic.1 ing 3 local anesthetics was carried out on adult volunteers.
Occasionally patients report a history of allergy to lidocaine, The anesthetics were compared for pain of infiltration and often stating that they are allergic to “all numbing medicines.” duration of effect. The study was approved by our institu- True anaphylaxis to local anesthetics, particularly amides, tion’s committee on research involving human subjects.
is extremely rare.1-4 Skin testing among patients with a The study was performed in a nonpatient care area of reported lidocaine allergy have shown that very few have an urban, university hospital emergency department.
true allergies.1-3 Nonetheless, anaphylaxis is a potentially Subjects were volunteers who were 18 years of age or older lethal complication that is best avoided.
and had no known allergy to any of the study solutions.
Even among those patients with a true lidocaine allergy, Study subjects were recruited from the physician, nursing, most often the reaction is not to the anesthetic itself but clerical, and housekeeping staff of our ED, none of whom rather to the preservative commonly used in multidose were in any other way involved with the study. Subjects vials, methylparaben.1,4,5 The traditional alternatives to the were paid $15 for participating; those who had abnormal amide anesthetics (eg, lidocaine, bupivacaine, mepivacaine) pain sensation in their upper extremities were excluded.
are the ester anesthetics (eg, procaine, tetracaine). The degra- Written informed consent was obtained.
dation product of these agents is para-amino benzoic acid Data collection occurred on July 24, 1997. Volunteers (PABA), a chemical that is closely related to methylparaben received 3 injections of 3 different solutions. The 3 solutions and could possibly induce the same allergic reaction.1 If a were given in a previously determined, computer-generated, patient were known to be in fact allergic to methylparaben, a randomized order. The 3 solutions were 1% diphenhy- safe alternative to multidose lidocaine would be single-dose dramine, .9% benzyl alcohol with 1:100,000 epinephrine, lidocaine, which contains no preservatives. However, and .9% buffered lidocaine. The diphenhydramine was patients and physicians alike are usually unable to make this prepared by making a 4:1 dilution of single-use normal saline distinction. Therefore, alternative agents to local anesthet- solution (containing no preservatives) and 5% diphenhy- dramine. The benzyl alcohol solution was prepared by Diphenhydramine has been shown to be a potential adding .2 mL epinephrine 1:1,000 to a 20-mL vial of multi- substitute for lidocaine.1,5-8 The chemical structure of dose normal saline solution containing benzyl alcohol, .9%.
antihistamines is closely related to that of local anesthetics The buffered lidocaine was prepared by making a 10:1 but dissimilar enough that the antigenicity is not the same.5 dilution of 1% lidocaine and sodium bicarbonate (1 However, diphenhydramine is more painful to administer mEq/mL). All solutions were prepared on the day of data than lidocaine,5-7 and side effects ranging from drowsiness to skin sloughing have been reported from this agent.5,6,8,9 All injections were given by 1 of the investigators (SWJ).
Benzyl alcohol has been studied as a possible local anes- Injections were given as follows: .5 mL of solution was thetic in volunteers and as an anesthetic used before needle administered subcutaneously over a period of 5 seconds insertion.10-13 Although the agent causes relatively little pain in 3 sites along the volar aspect of 1 forearm (chosen by on infiltration, its short duration of activity limits its clinical the subject). The needles were oriented in a longitudinal utility.10 Martin and Wilson11 compared benzyl alcohol axis and directed proximally, and the injections were given with epinephrine versus lidocaine with epinephrine. They at a single point. The sites were marked at 5-cm intervals, found that the former agent was less painful on infiltration with the first mark being made 5 cm proximal to the ulnar and, although it provided better anesthesia than placebo, styloid. Immediately after each injection, pain of infiltra- it was less effective than the lidocaine.11 Benzyl alcohol tion was measured with the use of a previously validated with epinephrine has never been compared with diphen- visual analog pain scale.14 The pain scale used was a 100- mm unmarked horizontal line with “Pain as bad as it can In this study we investigated alternative anesthetics for be” written on the left side, “No pain” on the right side, and patients allergic to lidocaine. We performed a double-blind “Severe,” “Moderate,” and “Mild” equally spaced across the clinical trial comparing buffered lidocaine, benzyl alcohol bottom. Subjects were asked to evaluate the pain of infil- with epinephrine, and diphenhydramine. Outcomes mea- tration only, not the pain of the needle stick. The injection sured included pain of infiltration and duration of anes- site was then covered with a perforated gauze sponge (to thesia. Our null hypothesis was that there would be no blind the investigators to any changes in skin color result- differences among the 3 anesthetics in either outcome.
D E C E M B E R 1 9 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 A L T E R N A T I V E S T O L I D O C A I N E A N E S T H E S I A
Bartfield, Jandreau & Raccio-Robak Duration was assessed by a different investigator (JMB).
with the exact χ2 test. All tests were 2-tailed, with signifi- Duration of anesthesia was assessed for each injection site cance defined as a probability value less than .05. A sample at 5-minute intervals for a maximum of 45 minutes by the size calculation was performed by assuming a parametric presence or absence of pain to needle prick with a 20-gauge distribution of the data, using variance data from a previous needle. This was tested by inserting the needle through the study performed on volunteers15 and based on Student’s t gauze sponge at the injection point (which was marked test with an α-value of .05 and a β-value of .20. We esti- with a marker so that the same spot could be used for all mated that a sample size of 30 subjects would be required testing) and applying gentle pressure. Subjects were then to detect a difference of at least 15 mm.
asked whether they felt a sharp sensation; the degree of painexperienced if sharp sensation returned was not quantified.
The time at which sharp sensation returned was recorded.
If sharp sensation had not returned after 45 minutes, this A total of 30 subjects were enrolled, including 12 men and was also noted. Subjects were questioned the next day 18 women. Subjects ranged in age from 19 to 50 years, with regarding the development of serious adverse reactions (defined as skin breakdown or necrosis) to any of the injec- The median pain of infiltration for diphenhydramine was 55 mm (IQR, 35 to 78 mm), compared with 12.5 mm After all subjects were enrolled, pain was quantified by (IQR, 4 to 26 mm) for buffered lidocaine and 5 mm (IQR, one of the investigators (JMB) by measuring to the nearest 1 to 13 mm) for benzyl alcohol (P=.001). Pairwise com- millimeter from the point of origin to the point marked by parisons (Wilcoxon’s signed rank test) showed that all pos- the patient (maximum, 100 mm). Duration of anesthesia sible combinations were statistically significant (buffered (maximum, 45 minutes) and return of sharp sensation dur- lidocaine versus benzyl alcohol, P=.022; diphenhydramine ing the study period were recorded for each injection.
versus benzyl alcohol, P=.001; diphenhydramine versus Pain scores failed normality testing and are reported as buffered lidocaine, P=.001). This information is displayed in medians with interquartile ranges (IQR). Pain scores for the Figure 1. The order of injection did not significantly influ- 3 solutions were compared with the use of a Kruskal-Wallis ence this result (P=.5), as shown in Figure 2.
test. Duration of anesthesia was tested with the use of sur- Duration of anesthesia is displayed by survival analysis vival analysis techniques by a log-rank test. The potential in Figure 3. Significant differences were found among the effect of order of injection on pain score and duration ofanesthesia was tested by an analysis of covariance performedon ranked responses. The proportion of anesthetic injections Figure 2.
in which sensation returned within 45 minutes was tested Study solution pain scores with injection order. Median painscores and IQRs for the 3 study solutions for each of the 3possible injection orders. Figure 1.
Study solution pain scores. Median pain scores and IQRs forthe 3 study solutions (P=.001). Pairwise comparisons:buffered lidocaine versus benzyl alcohol, P=.022; diphenhy- dramine versus benzyl alcohol, P=.001; diphenhydramineversus buffered lidocaine, P=.001. 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 2 : 6 D E C E M B E R 1 9 9 8
A L T E R N A T I V E S T O L I D O C A I N E A N E S T H E S I A
Bartfield, Jandreau & Raccio-Robak 3 anesthetics (P<.001). Pairwise comparisons (log-rank patients who are allergic to lidocaine (or to methylparaben test) revealed a longer duration of anesthesia for buffered lidocaine, compared with benzyl alcohol (P=.02) or diphen- Anecdotal reports and research protocols have shown hydramine (P=.001). No significant difference in duration that benzyl alcohol (found as a preservative in multidose of anesthesia was found between benzyl alcohol and normal saline solutions) produces minimal pain of infil- diphenhydramine (P=.13). The order of injection did not tration and good anesthesia.10-13 Studies by Novak16 and significantly influence this result (P>.9).
Kimura17 found very low toxicity of benzyl alcohol in Pain sensation returned within the 45-minute study parenteral administration. Nuttal et al13 and Thomas12 period in only 3 (10%) of 30 lidocaine injections (95% reported that benzyl alcohol facilitates intravenous line confidence interval [CI], 2% to 26%), compared with 11 placement because of its anesthetic effect. Wightman and (37%) of 30 benzyl alcohol injections (95% CI, 20% to 56%) Vaughan10 compared benzyl alcohol and 5 other anesthet- and 19 (63%) of 30 diphenhydramine injections (95% CI, ics as intradermal injections in volunteers and found 44% to 80%)(P=.001). One patient reported return ofpain sensation after 5 minutes for the benzyl alcoholinjection. Pain sensation did not return for at least 15 Figure 3.
Duration of anesthesia. Survival curve for numbers of sub- minutes in all of the remaining injections. All subjects jects reporting lack of sharp sensation for each of the study were available for next day follow-up, and none reported solutions versus time (in 5-minute intervals, 45-minute maximum), P<.001. Pairwise comparisons: buffered lido-caine versus diphenhydramine, P=.001; buffered lidocaineversus benzyl alcohol, P=.02; diphenhydramine versus ben- Although anaphylactic reactions to local anesthetics arerare, they represent a potentially life-threatening complica- tion. When patients report an allergy to lidocaine, physiciansare forced to use alternative local anesthetics or no local anesthetics. Diphenhydramine has been studied by anumber of researchers over the last decade, using volun-teers and patients.3-9 It has been found that 1% diphenhy- dramine provides anesthesia comparable to 1% lidocaine,but the solution is considerably more painful on infiltra-tion than lidocaine.5-7 Ernst et al7 compared .5% diphen-hydramine with 1% lidocaine and found the former agent to be less effective. Singer and Hollander9 attempted toattenuate the pain of injection as well as the local irritanteffects of diphenhydramine by buffering the solution; they found no significant differences between plain and bufferedsolutions.
A number of adverse reactions have been reported when diphenhydramine is used as a local anesthetic. Sedation was reported in 2 of 48 subjects who received diphenhydraminein 1 study.6 Local erythema was reported at the injectionsite in approximately half of the subjects in 1 study5 and in all of the subjects in another study.9 Persistent sorenessfor up to 3 days was reported in 7 of 24 subjects receivinglocal injections of diphenhydramine in another study.5One subject was found to have skin sloughing at the site of a diphenhydramine injection with persistent anesthesia for 2 weeks.8 Given the relative discomfort of diphenhy- dramine infiltration and its potential side effects, an alternative non-“caine” anesthetic would be desirable for D E C E M B E R 1 9 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 A L T E R N A T I V E S T O L I D O C A I N E A N E S T H E S I A
Bartfield, Jandreau & Raccio-Robak that benzyl alcohol was the least painful. However, the may have gone unreported. It is also possible that adverse duration of anesthesia for benzyl alcohol was only a few effects would have been reported had a larger volume of minutes. By adding epinephrine to the solution, Martin and solution, similar to that required for most laceration repairs, Wilson11 were able to show that benzyl alcohol can pro- been injected. A prospective trial using patients with lac- vide long-term anesthesia, although less adequately than erations may be indicated to determine the generalizabil- lidocaine with epinephrine. They compared benzyl alcohol ity of our results and to study the side effects of the drugs.
with epinephrine, lidocaine with epinephrine, and placebo In instances in which an alternative local anesthetic agent and found that benzyl alcohol with epinephrine was the least is required, particularly in the patient with a reported lido- painful on administration and that its anesthetic potential was caine allergy, benzyl alcohol with epinephrine appears to be greater than that of placebo but not as great as that of lido- a better choice than diphenhydramine. Diphenhydramine caine with epinephrine.11 These results, which had been remains preferable for those cases in which epinephrine is published only in abstract form at the time our study was performed, represent the only report of benzyl alcohol with The authors gratefully thank Terry L Peters, MS, for her help in data analysis.
epinephrine as a local anesthetic that we were able to findin the literature.
In this study, we compared benzyl alcohol with epinephrine, diphenhydramine, and buffered lidocaine 1. Orlinsky M, Dean E: Local and topical anesthesia and nerve blocks of the thorax and extremi- in an effort to evaluate the 2 alternatives to traditional ties, in Roberts JR, Hedges JR (eds): Clinical Procedures in Emergency Medicine, ed 2. Philadelphia: anesthetics in a single controlled trial. The results of this double-blind study showed that the median pain score for 2. Chandler MJ, Grammer LC, Patterson R: Provocative challenge with local anesthetics in patients with a prior history of reaction. J Allergy Clin Immunol 1987;79:883-886.
benzyl alcohol with epinephrine (5 mm) was significantly 3. Incaudo G, Schatz M, Patterson R, et al: Administration of local anesthetics to patients with a less than for either buffered lidocaine (12.5 mm) or diphen- history of prior adverse reaction. J Allergy Clin Immunol 1978;61:339-345.
hydramine (55 mm). Previous studies have suggested that 4. Adriani J, Zepernick R: Allergic reactions to local anesthetics. South Med J 1981;74:694-699.
a difference in visual analog pain scores of 13 mm18 to 18 mm19 5. Green S, Rothrock S, Gorchynski J: Validation of diphenhydramine as a local anesthetic. Ann (on a 100-mm scale) may be clinically significant. Therefore, although we report a statistically significant difference 6. Ernst A, Anand P, Nick T, et al: Lidocaine versus diphenhydramine for anesthesia in the repair of between benzyl alcohol with epinephrine and buffered minor lacerations. J Trauma 1993;34:354-357.
lidocaine, this difference may not be clinically significant.
7. Ernst AA, Marvez-Valls E, Mall G, et al: 1% Lidocaine versus 0.5% diphenhydramine for local anesthesia in minor laceration repair. Ann Emerg Med 1994;23:1328-1332.
In the present study, buffered lidocaine was shown to 8. Dire DJ, Hogan DE: Double-blinded comparison of diphenhydramine versus lidocaine as a local have a longer duration of action than either of the other anesthetic. Ann Emerg Med 1993;22:1419-1422.
agents. However, all but 1 patient (for whom sensation 9. Singer AJ, Hollander JE: Infiltration pain and local anesthetic effects of buffered vs plain 1% returned after 5 minutes for benzyl alcohol with epinephrine) diphenhydramine. Acad Emerg Med 1995;2:884-888.
reported at least 15 minutes of anesthesia with all agents.
10. Wightman MA, Vaughan RW: Comparison of compounds used for intradermal anesthesia.
Our study was limited by the facts that we used only Anesthesiology 1976;45:687-689.
healthy volunteers and that only .5-mL injections over 5 11. Martin S, Wilson L: Benzyl alcohol with epinephrine as an alternative local anesthetic[abstract]. Acad Emerg Med 1996;3:493-494.
seconds into intact skin were compared. These results 12. Thomas DV: Saline with benzyl alcohol prevents pain of needle insertion. Anesth Analg may not be generalizable to the infiltration of larger volumes at different rates in traumatic wounds. Additionally, sub- 13. Nuttall GA, Barnett MR, Smith RL, et al: Establishing intravenous access: A study of local anes- jects were asked to report whether they had return of sen- thetic efficacy. Anesth Analg 1993;77:950-953.
sation, but the extent of the anesthesia was not tested. It is 14. Scott J, Huskisson EC: Graphic representation of pain. Pain 1976;2:175-184.
possible that the anesthesia achieved was adequate to blunt 15. Bartfield JM, Crisafulli K, Raccio-Robak N, et al: The effects of warming and buffering on pain the pain of a gentle needle stick but not pain of suturing or of infiltration of lidocaine. Acad Emerg Med 1995;2:254-258.
more invasive emergency procedures. Finally, the number 16. Novak E: The tolerance and safety of intravenously administered benzyl alcohol methylpred-nisolone sodium succinate formulations in normal human subjects. Toxicol Appl Pharmacol of subjects used in this study may have been too small to detect any significant side effects from the drugs tested. In 17. Kimura ET: Parenteral toxicity studies with benzyl alcohol. Toxicol Appl Pharmacol 1971;18:54-61.
addition, because subjects were specifically questioned 18. Todd KH, Funk KG, Funk JP, et al: Clinical significance of reported changes in pain severity. Ann only about major adverse effects such as skin sloughing on the day after the injections, other, more minor problems 19. Funk JP, Todd K: The minimum clinically important difference in the physician-assigned visual and adverse effects that became apparent after that day analogue pain scores. Acad Emerg Med 1996;3:142-146.
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 2 : 6 D E C E M B E R 1 9 9 8

Source: http://www.sjem.org/files/39778867.pdf