S66 • Volume 29 • Number S6 • November/December 2009

Botulinum neurotoxin type A (BoNT-A) injectionsare the most frequently sought nonsurgical aes-thetic procedure, accounting for almost 2.5 million

procedures in 2008 alone.1 BoNT-A also has importanttherapeutic uses, including the treatment of movementdisorders such as cervical dystonia and tremors. Bothaesthetic and therapeutic uses of BoNT-A require repeatedinjections to maintain the desired effect(s). Accordingly,the potential of BoNT-A formulations for inducing neu-tralizing antibodies is of critical importance.

As with other antigens, BoNT-A immunogenicity isinfluenced by the specific formulation and by theextent of antigenic exposure, including specific activity,frequency of treatment, and dose.2-5 The purity and

Aesthetic Surgery Journal

Background: The induction of neutralizing antibodies during the aesthetic application of botulinum neuro-toxin type A is rare, but of potential clinical concern. Phase III studies of a new US formulation of botulinumneurotoxin type A, Dysport (BoNTA-ABO [abobotulinumtoxinA]; Medicis Aesthetics, Scottsdale, AZ), have notidentified any cases of neutralizing antibody formation during the treatment of glabellar lines in patients whoreceived up to nine treatments.Objective: To provide an in-depth analysis of the potential for induction of neutralizing antibodies in the studypopulation enrolled in phase III trials of BoNTA-ABO in the treatment of glabellar lines.Methods: First and last available serum samples from patients in the BoNTA-ABO Glabellar LinesDevelopment Program were screened for BoNTA-ABO antibodies with a radioimmunoprecipitation assay(RIPA), followed by a confirmatory competitive assay (RIPA-C). Confirmed RIPA-C–positive samples were fur-ther evaluated for the presence of neutralizing antibodies using a mouse protection assay (MPA), a highly spe-cific bioassay for neutralizing antibodies. We conducted safety and efficacy evaluations, including day 30responder rate (a rating of no or mild glabellar lines) and duration of response in the last treatment cycle.Results: Of 1554 patients who received at least one BoNTA-ABO treatment (10 units at five injection points,for a total dose of 50 units/treatment; range one to nine treatments), five (0.32%) were antibody positive onthe RIPA-C assay—two at baseline and three at the last treatment cycle. None of the RIPA-C–positive samples test-ed positive for neutralizing antibodies upon further evaluation using the highly specific MPA. Of note, the RIPA-C–positive group had a responder rate of 100% and a mean response of 103.3 days, while the RIPA-C–negative group had a responder rate of 90% and a mean response of 89.4 days. The safety of BoNTA-ABOdid not appear to be altered in the RIPA-C–positive group.Conclusions: At the dose and treatment interval used in the correction of glabellar lines, induction of neu-tralizing antibodies to BoNTA-ABO was not observed. None of the five samples that initially gave positiveresults in a RIPA-C assay were positive when further evaluated using the MPA. Clinically, RIPA-C–positive sta-tus did not correlate with any reduction in efficacy or an altered safety profile, although the small numbersprevent definitive conclusions. These data suggest that the five RIPA-C–positive samples represented false pos-itives. (Aesthetic Surg J 2009;29:S66–S71.)

Dr. Lawrence is Senior Vice President of Research andDevelopment and Chief Medical Officer of Medicis PharmaceuticalCorporation, Scottsdale, AZ. Dr. Moy is from the Fincher MedicalGroup, Veterans Affairs West Los Angeles Medical Center, andfrom the David Geffen School of Medicine, University ofCalifornia, Los Angeles, Los Angeles, CA.

An Evaluation of Neutralizing AntibodyInduction During Treatment of GlabellarLines with a New US Formulation ofBotulinum Neurotoxin Type AIra Lawrence, MD; and Ronald Moy, MD

Volume 29 • Number S6 • November/December 2009 • S67Neutralizing Antibody Induction During Treatment With BoNTA-ABO

specific activity of a formulation have been shown tohave a significant effect on the induction of BoNT-Aantibodies.4,6 Preparations with a high specific activityrequire less protein per injection and therefore mayreduce antigenic exposure and immunogenicity.7

Frequent injections and long-term therapy increase apatient’s total antigenic exposure and correlate withthe appearance of BoNT-A–blocking antibodies.5

Dosage is another important factor. One study foundthat only patients who received BoNT-A doses of morethan 600 units for the treatment of movement disor-ders developed neutralizing antibodies.2 The BoNT-Adose for aesthetic applications is much lower.

Antibody responses to currently available BoNT-Apreparations are generally rare in patients undergoingtreatment for movement disorders. Although the originalformulation of a US Food and Drug Administration(FDA)–approved BoNT-A preparation (BoNTA-ONA[onabotulinumtoxinA]; Botox, Allergan, Irvine, CA) wasassociated with a neutralizing antibody rate of 9.5% inpatients receiving large doses (mean dose of approxi-mately 200 units per injection) for the treatment of cervi-cal dystonia,3 a revised BoNTA-ONA formulationintroduced in the United States in 1998 was much lessimmunogenic. A study of 119 cervical dystonia patientsdid not identify any cases of blocking antibodies associ-ated with the new formulation.3 Recent studies suggest aneutralizing antibody rate of 1.2% in cervical dystoniapatients treated with BoNTA-ONA for up to four years.8

In 2009, a new US formulation of BoNT-A, Dysport(BoNTA-ABO [abobotulinumtoxinA]; Medicis Aesthetics,Scottsdale, AZ), was approved by the FDA for the treat-ment of moderate to severe glabellar lines on the basisof data obtained from large-scale clinical trials. This for-mulation has been available in countries outside ofNorth America since 1991. In a study conducted inEurope, Göschel et al2 evaluated the immunogenicity ofBoNTA-ABO and reported that in patients receiving theproduct for therapeutic purposes, the lowest dose thatinduced neutralizing antibodies was 15.5 ng (620units).2 None of the patients receiving lower dosesdeveloped neutralizing antibodies, while four of 40(10%) patients receiving doses of more than 600 unitshad detectable titers of BoNT-A–neutralizing antibodies.

The induction of neutralizing antibodies to BoNT-Aduring aesthetic applications has not been systemati-cally studied,6 but is likely to be a very rare eventbecause aesthetic procedures generally require muchlower doses and longer retreatment intervals.Nevertheless, there has been at least one case reportof a patient who developed neutralizing antibodies toBoNTA-ONA during the treatment of facial rhytides.9

Such cases are of potential clinical concern.9,10 Notonly do they result in nonresponse to aesthetic treat-ment, but they may also prevent the future therapeu-tic use of BoNT-A in that patient.

Studies of BoNT-A–neutralizing antibodies are com-plicated by difficulty in detecting these antibodies.6

The gold standard assay is the mouse protection assay(MPA), which tests for neutralizing BoNT-A antibodiesby determining the ability of sera to prevent the deathof mice given a lethal dose of botulinum toxin.Although this assay is highly specific (100%), it haslow sensitivity (range 30%-50%).11 Furthermore,because it relies on the use of live animals, both ethi-cal and cost issues dictate against its use in large-scale antibody screening studies. Enzyme-linkedimmunosorbent assays, Western blots, and other rapidtests for antibody identification are less cumbersomethan the MPA, but they detect both neutralizing andnonneutralizing antibodies and are therefore not assensitive or specific.

As a key component of the BoNTA-ABO USGlabellar Lines Development Program, the immuno-genicity of BoNTA-ABO during aesthetic applicationshas been closely evaluated. Individual phase III stud-ies of BoNTA-ABO failed to identify any cases of neu-tralizing antibody formation during treatment ofglabellar lines, but the methods used in these analyseshave not been fully described.12-14 Here we provide acomprehensive, in-depth evaluation of antibody for-mation in the study population enrolled in phase IIItrials of BoNTA-ABO for the treatment of glabellarlines. The data reported here indicate that patientstreated with BoNTA-ABO did not develop neutralizingantibodies during the course of these studies at theprescribed dose and treatment schedules.

METHODSThe serum samples for antibody testing and the efficacyand safety data analyzed as part of this study wereobtained from studies approved by the InstitutionalReview Boards of centers participating in a GlabellarLines Development Program. These studies were conduct-ed in accordance with ethical standards for biomedicalresearch, as established by the 18th World MedicalAssembly, Helsinki, Finland, 1946 and later revisions,and with US federal regulations and guidelines.

Patients and TreatmentThe study population consisted of individuals enrolled inthe BoNTA-ABO Glabellar Lines Development Program whohad serum antibodies available for testing. Subjects providedwritten informed consent before enrolling in the originalstudy. All patients in this study received at least one treat-ment consisting of 50 units of BoNTA-ABO (10 units/0.05mL at each of five separate injection points in the glabellarregion). Subsequent treatments (if any) were given approx-imately every 12 to 16 weeks, for up to nine retreatments.

Antibody AssessmentsFor each subject, the first and last available serum sam-ples were analyzed for the presence of BoNTA-ABOantibodies as described below. All samples were ana-lyzed at Ipsen Pharma AA, a subsidiary of IpsenBiopharm, in Barcelona, Spain.

S68 • Volume 29 • Number S6 • November/December 2009 Aesthetic Surgery Journal

Samples were initially screened with a radioimmuno-precipitation assay (RIPA) that used an 125I-labeledrecombinant C-terminal fragment of the botulinum toxinheavy chain (Hc) as the ligand,15 because this is the por-tion of the molecule that is usually recognized by neu-tralizing antibodies.16,17 Following incubation withserum samples, protein G-coated beads were used toprecipitate antibody-ligand complexes and radioactivitywas measured in a gamma counter.15 Positive sampleswere tested in a confirmatory competitive assay (RIPA-C).Each sample was analyzed twice, with competitor coldunlabeled Hc added to the second sample at 100 timesthe amount of 125I-labeled Hc.15 Samples with greaterthan 20% inhibition in the presence of competitorwere considered to be RIPA-C–positive. All samplesthat were positive in the RIPA-C assay were furtherevaluated in the MPA.

To characterize the RIPA used in this study, samplesfrom 197 BoNTA-ABO–treated subjects were analyzed byRIPA and the results were compared with MPA data. TheRIPA was found to be highly sensitive, but of low speci-ficity (100% sensitivity compared with the MPA; 66.5%specificity). Its negative predictive value was 100%,whereas its positive predictive value was 24.0%.Addition of the RIPA-C assay improved the specificityand predictive value of the assay. Compared with theMPA, the RIPA-C was 100% sensitive and 94.7% specific.Its negative predictive value was 100% and its positivepredictive value was 66.7%.

Efficacy and Safety EvaluationsEfficacy was assessed using a validated four-point photo-graphic scale, the glabellar line severity scale (GLSS;where 0 � no visible glabellar lines, 1=mild glabellarlines, 2 =moderate glabellar lines and 3 � severe glabel-lar lines).18 Responders were defined as subjects with abaseline GLSS score of 2 or 3 and a GLSS score of 0 or 1at maximum frown at day 30 following BoNTA-ABO treat-ment. The efficacy endpoints evaluated in this study werethe day 30 responder rate in RIPA-C–positive subjectscompared with RIPA-C–negative subjects and the durationof response (time to return to GLSS of 2 or 3) in the lasttreatment cycle in RIPA-C–positive subjects comparedwith RIPA-C–negative subjects.

All adverse events were recorded. To determinewhether the appearance of antibodies correlated withan altered safety profile, the incidences and types oftreatment-emergent adverse events (TEAE) in RIPA-C–positive subjects were compared with those inRIPA-C–negative subjects. Particular attention waspaid to inflammatory injection site reactions, reactionsconsistent with a systemic immunologic response, andto ocular events related to muscle paralysis.

Statistical AnalysesTwo-sided statistical tests were used to compare RIPA-C–positive and RIPA-C–negative subjects at the 0.050significance level. Two-sided 95% confidence intervals

(CI) were calculated using the normal approximationto the binomial distribution. Where relevant, two-sided 95% CI for the estimated differences betweenseroconversion groups were calculated with themethod described by Agresti and Caffo.19 The primaryP value for differences between RIPA-C–positive andRIPA-C–negative groups was determined by using theFisher exact test with the exception of duration ofresponse, which was evaluated by the Wilcoxon testfor treatment difference.

RESULTS

Antibody StatusAntibody screening was conducted on the 1554patients who had received at least one treatment withBoNTA-ABO, had baseline and follow-up blood sam-ples that could be evaluated for antibody status, andhad received BoNTA-ABO in their final treatmentcycle. A total of 847 (54.5%) patients were BoNT-Anaïve at the time of entry into the original study. Fiveof the 1554 subjects (0.32%) were antibody-positiveas assessed by the RIPA-C assay (Figure). Samplesfrom the remaining 1549 subjects tested negative onthe RIPA-C assay.

All RIPA-C–positive samples were tested in the MPAfor the presence of BoNT-A–neutralizing antibodies.None of the RIPA-C–positive samples tested positive inthe MPA (Figure).

Figure 1. BoNT-A antibody status of patients in the BoNTA-ABO/Dysport formulation Glabellar Lines Development Program.

Volume 29 • Number S6 • November/December 2009 • S69Neutralizing Antibody Induction During Treatment With BoNTA-ABO

Demographic and Baseline Characteristics byRIPA-C StatusDemographic and baseline characteristics of RIPA-C–positive subjects were compared with those of RIPA-C–negative subjects (Table 1). Although somedifferences were noted between RIPA-C–positive andRIPA-C–negative subjects, no definitive conclusions canbe drawn because of the small number of subjects in theRIPA-C–positive group.

Characterization of RIPA-C–Positive SamplesCharacteristics of the RIPA-C–positive samples aresummarized in Table 2. Of the five RIPA-C–positivesubjects, two had positive samples at baseline andthree had positive samples at the last treatment cycle.None of the patients had positive samples at bothbaseline and last cycle. Three of the RIPA-C–positivesubjects were BoNT-A–naïve and two had been previ-ously exposed to BoNT-A. Of the two individuals with

previous exposure to BoNT-A, one had a positivebaseline sample but a negative sample at the lasttreatment cycle, and one had a positive sample at thelast treatment cycle.

All five of the RIPA-C–positive subjects had positivesamples within the first three treatment cycles. TwoRIPA-C–positive subjects had positive samples at base-line, and one each had a positive sample after one,two, and three cycles of treatment (median of twotreatments; Table 2). In comparison, the median num-ber of treatments in the RIPA-C–negative group wasthree (range one to nine).

Efficacy and SafetyEfficacy evaluations were used to assess the impact ofRIPA-C–positive antibodies on treatment outcomes. Thenumber of evaluable patients with a day 30 response(GLSS grade of 0 or 1 at maximum frown) was five offive (100%) for RIPA-C–positive subjects and 1277 of1412 (90%) for RIPA-C–negative subjects (P � .999using the Fisher exact test). Day 30 responder data werenot available for 137 of 1549 patients in the RIPA-C–negative group. The duration of response (time toreturn to GLSS grade of 2 or 3 at maximum frown) at thelast treatment cycle was a mean (� standard deviation)of 103.3 (� 54.01) days for the four evaluable patients inthe RIPA-C–positive group and 89.4 days (� 54.85) forthe 1448 evaluable subjects in the RIPA-C–negativegroup. The difference in response duration between thegroups was not statistically significant (P � .271 usingthe Wilcoxon test for treatment difference).

Reports on the short- and long-term safety profiles ofBoNTA-ABO have been previously published12-14,20,21

and summaries are presented elsewhere in this supple-ment. To assess the impact of RIPA-C–positive status onsafety, TEAE occurring in the last cycle of treatment inRIPA-C–positive subjects were compared with thoseobserved in RIPA-C–negative subjects. There did notappear to be any clustering of TEAE in RIPA-C–positivesubjects. One individual in the RIPA-C–positive groupdeveloped eyelid edema. One subject became pregnantafter the initiation of treatment and had a spontaneous

Table 2. Characteristics of RIPA-C–positive samples by patient

Timing of No. of treatmentsPatient BoNT-A RIPA-C–positive before positiveno. naïve sample status

1 No Baseline 0

2 Yes Baseline 0

3 Yes Last treatment 1

4 Yes Last treatment 2

5 No Last treatment 3

BoNT-A, botulinum neurotoxin type A; RIPA, radioimmunoprecipitationassay; RIPA-C, RIPA followed by a confirmatory competitive assay; SD,standard deviation.

Table 1. Demographic and baseline characteristics by seraconversion group

RIPA-C–positive RIPA-C–negative Parameter (N = 5) (N = 1549)

Age (years)Mean (SD) 41.8 (1.8) 48.3 (9.9)Median 42.0 48.0

Age category (n [%])≤50 yrs 5 (100%) 932 (60.2%)>50 to <65 yrs 0 523 (33.8%)≥65 yrs 0 94 (6.1%)

Gender (n [%])Female 5 (100%) 1377 (88.9%)Male 0 172 (11.1%)

Race/ethnicity (n [%])White 5 (100%) 1367 (88.3%)Black or African 0 15 (1.0%)Asian 0 129 (8.3%)Hispanic or Latino 0 20 (1.3%)Other 0 16 (1.0%)

BoNT-A naïve (n [%])Yes 3 (60%) 844 (54.5%)No 2 (40%) 705 (45.5%)

Baseline assessment of glabellar lineseverity (n [%])

Moderate 3 (60%) 1124 (72.6%)Severe 2 (40%) 416 (26.9%)

BoNT-A, botulinum neurotoxin type A; RIPA, radioimmunoprecipitationassay; RIPA-C, RIPA followed by a confirmatory competitive assay; SD,standard deviation.

S70 • Volume 29 • Number S6 • November/December 2009 Aesthetic Surgery Journal

abortion. Because of the small number of patients in theRIPA–C positive group, it is not possible to draw anyclinically meaningful conclusions from these findings.There were no reports of eyelid ptosis or asthenopia inthe RIPA-C–positive group.

DISCUSSIONBecause both therapeutic and aesthetic applications ofBoNT-A require repeated administration, low antigenicity isan important characteristic of BoNT-A preparations.6

BoNTA-ABO has high specific activity (115 units/ng) andonly very small quantities of protein are required per dose(approximately 0.43 ng per 50 unit dose). It has been pro-posed that this high specific activity may be associatedwith low antigenicity.22 The data presented here supportthis hypothesis.

In this study, we investigated whether BoNT-A–neutralizing antibodies could be identified in sub-jects receiving BoNTA-ABO treatment for glabellarlines. Out of 1554 BoNTA-ABO–treated subjects, onlyfive (0.32%) showed evidence of antibodies in theRIPA-C screening assay. Although we cannot definitivelyrule out the presence of false negatives, in-depth analy-ses of the RIPA and RIPA-C assays determined a nega-tive predictive value of 100% for both assays.

It is highly likely that all five RIPA-C-positive samplesdetected in our study represented false positives. None ofthe five RIPA-C–positive samples tested positive in the MPA(the “gold standard” bioassay for neutralizing antibodies),and no alterations in efficacy or safety were observed inRIPA-C–positive subjects. All five of the subjects recorded aresponse at day 30. Two of the RIPA-C–positive sampleswere positive at baseline, even though one of the patientshad never been exposed to BoNT-A. Together with the rela-tively high false positive rate of the RIPA-C assay (33.3%),these observations strongly suggest that the RIPA-C samplesdid not contain BoNT-A–neutralizing antibodies and weretherefore false positives.

BoNTA-ABO treatments for glabellar lines are normallyrepeated every 12 to 16 weeks. In this study, somepatients received up to nine cycles of treatment (medianof three). Repeated administration did not increase thenumber of patients with positive samples. In fact, allRIPA-C–positive samples occurred in patients who hadreceived three or fewer cycles of treatment (median oftwo), although two patients had previously receivedBoNT-A treatment.

BoNT-A–neutralizing antibodies are known to beassociated with a loss of clinical response.2,6,10 In thecase study reported by Borodic,6 the patient who devel-oped neutralizing antibodies was first identified by alack of reduction in dynamic facial lines in the glabellarregion following BoNTA-ONA injections. In this study,subjects with RIPA-C–positive samples did not showsigns of reduced efficacy, as assessed by the proportionof responders at day 30 and the duration of response.Similarly, no obvious alterations were observed in thesafety profile, including TEAE likely to be associated

with immune reactions. However, firm conclusions con-cerning the effect of RIPA-C–positive antibodies on effi-cacy and safety cannot be drawn because of the smallnumber of affected subjects.

On the basis of these data, we conclude that, at thedose and treatment interval used in the correction ofglabellar lines, the induction of neutralizing antibodiesto BoNTA-ABO was not observed and therefore must bea very rare event. It is likely that none of the RIPA-C–positive samples identified in BoNTA-ABO–treatedsubjects represented true positives, because all of thesamples tested negative in the definitive assay for neu-tralizing antibodies, the MPA. The low (or absent) inci-dence of neutralizing antibody induction in subjectstreated with BoNTA-ABO may relate to the high specificactivity of BoNTA-ABO, which allows minute doses tobe used to achieve the reduction of glabellar lines.

CONCLUSIONSAlthough long-term monitoring is advisable, our datasuggest that BoNTA-ABO has very low immunogenicityand is unlikely to induce neutralizing antibodies whenused for aesthetic purposes. ◗

ACKNOWLEDGMENTS

The authors acknowledge the editorial and writing assistance ofSharon L. Cross, PhD, a writer for Premier Healthcare Resource(Morristown, NJ).

DISCLOSURES

The authors were compensated for their contributions in preparingthis manuscript and as investigators for this study. The writing ofthis manuscript was funded by Medicis Aesthetics (Scottsdale, AZ).Dr. Lawrence is employed by Medicis Pharmaceutical, the distribu-tor of Dysport for aesthetic use in the United States. Dr. Moy was apaid investigator for Inamed (Santa Barbara, CA). He has nofinancial arrangements with Medicis Aesthetics.

REFERENCES1. American Society for Aesthetic Plastic Surgery website. Cosmetic

Surgery National Data Bank statistics 2008. (Accessed August 29, 2009,at http://www.surgery.org/sites/default/files/2008stats.pdf.)

2. Göschel H, Wohlfarth K, Frevert J, Dengler R, Bigalke H. Botulinum Atoxin therapy: neutralizing and nonneutralizing antibodies—therapeuticconsequences. Exp Neurol 1997;147:96–102.

3. Jankovic J, Vuong KD, Ahsan J. Comparison of efficacy and immuno-genicity of original versus current botulinum toxin in cervical dystonia.Neurology 2003;60:1186–1188.

4. Aoki KR. Pharmacology and immunology of botulinum neurotoxins.Int Ophthalmol Clin 2005;45:25–37.

5. Atassi MZ. Basic immunological aspects of botulinum toxin therapy.Mov Disord 2004;19(suppl 8):S68–S84.

6. Borodic G. Botulinum toxin, immunologic considerations with long-term repeated use, with an emphasis on cosmetic applications. FacialPlast Surg Clin North Am 2007;15:11–16.

7. Borodic G, Johnson E, Goodnough M, Schantz E. Botulinum toxin ther-apy, immunologic resistance, and problems with available materials.Neurology 1996;46:26–29.

8. Brin MF, Comella CL, Jankovic J, Lai F, Naumann M, CD-017 BoNTAStudy Group. Long-term treatment with botulinum toxin type A in cer-vical dystonia has low immunogenicity by mouse protection assay. MovDisord 2008;23:1353–1360.

Volume 29 • Number S6 • November/December 2009 • S71Neutralizing Antibody Induction During Treatment With BoNTA-ABO

9. Borodic G. Immunologic resistance after repeated botulinum toxin typeA injections for facial rhytides [letter]. Ophthal Plast Reconstr Surg2006;22:239–240.

10. Jankovic J, Schwartz K. Response and immunoresistance to botulinumtoxin injections. Neurology 1995;45:1743–1746.

11. Hanna PA, Jankovic J, Vincent A. Comparison of mouse bioassay andimmunoprecipitation assay for botulinum toxin antibodies. J NeurolNeurosurg Psychiatry 1999;66:612–616.

12. Moy R, Maas C, Monheit G, Huber B, Reloxin Investigational Group.Long-term safety and efficacy of a new botulinum toxin type A in treat-ing glabellar lines. Arch Facial Plast Surg 2009;11:77–83.

13. Brandt F, Swanson N, Baumann L, Huber B. Randomized, placebo-controlled study of a new botulinum toxin type A for treatment ofglabellar lines: efficacy and safety. Dermatol Surg 2009 June 22[Epub ahead of print].

14. Monheit GD, Cohen JL, Reloxin Investigational Group. Long-term safetyof repeated administrations of a new formulation of botulinum toxintype A in the treatment of glabellar lines: interim analysis from anopen-label extension study. J Am Acad Dermatol 2009;61:421–425.

15. Morte C, Pla F, Cedo M, et al. Development of a multi-step approach forthe detection of neutralizing antibodies to botulinum toxin type A.Presented at Basic and Therapeutic Aspects of Botulinum and TetanusToxins. Baveno, Italy, June 12–15, 2008.

16. Rosenberg JS, Middlebrook JL, Atassi MZ. Localization of the regionson the C-terminal domain of the heavy chain of botulinum A recog-nized by T lymphocytes and by antibodies after immunization of micewith pentavalent toxoid. Immunol Invest 1997;26:491–504.

17. Oshima M, Hayakari M, Middlebrook JL, Atassi MZ. Immune recogni-tion of botulinum neurotoxin type A: regions recognized by T cells andantibodies against the protective H(C) fragment (residues 855-1296) ofthe toxin. Mol Immunol 1997;34:1031–1040.

18. Ascher B, Zakine B, Kestemont P, Baspeyras M, Bougara A, Santini J. Amulticenter, randomized, double-blind, placebo-controlled study of effi-cacy and safety of 3 doses of botulinum toxin A in the treatment ofglabellar lines. J Am Acad Dermatol 2004;51:223–233.

19. Agresti A, Caffo B. Simple and effective confidence intervals for propor-tions and differences of proportions result from adding two successesand two failures. Am Stat 2000;54:280–288.

20. Rubin MG, Dover JS, Glogau RG, Goldberg DJ, Goldman MP,Schlessinger J. The efficacy and safety of a new US botulinum toxintype A in the retreatment of glabellar lines following open-label treat-ment. J Drugs Dermatol 2009;8:439–444.

21. Kane MA, Brandt F, Rohrich RJ, Narins RS, Monheit GD, Huber MB.Evaluation of variable-dose treatment with a new U.S. botulinum toxintype A (Dysport) for correction of moderate to severe glabellar lines:results from a phase 3, randomized, double-blind, placebo-controlledstudy. Plast Reconstr Surg 2009 Jul 28 [Epub ahead of print].

22. Dressler D, Hallett M. Immunological aspects of Botox, Dysport, andMyobloc/NeuroBloc. Eur J Neurol 2006;13(suppl 1):11–15.

Accepted for publication August 31, 2009.

Reprint requests: Ronald Moy, MD, Moy Fincher Medical Group, Universityof California, Los Angeles, 100 UCLA Medical Plaza, Suite 590, Los Angeles,CA 90024. E-mail: rmoy@ucla.edu.

Copyright © 2009 by The American Society for Aesthetic Plastic Surgery, Inc.

1090-820X/$36.00

doi:10.1016/j.asj.2009.09.009